INFO-DIR-SECTION Programming
START-INFO-DIR-ENTRY
* Gambit: (gambit).		A portable implementation of Scheme.
* gsi: (gambit) interpreter.	Gambit interpreter.
* gsc: (gambit) compiler.	Gambit compiler.
END-INFO-DIR-ENTRY

   This file documents Gambit, a portable implementation of Scheme.

   Copyright (C) 1994-2022 Marc Feeley.

   Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

   Permission is granted to copy and distribute modified versions of
this manual under the conditions for verbatim copying, provided that
the entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

   Permission is granted to copy and distribute translations of this
manual into another language, under the above conditions for modified
versions, except that this permission notice may be stated in a
translation approved by the copyright holder.

Gambit
******

This manual documents Gambit.  It covers release v4.9.4.

1 The Gambit system
*******************

The Gambit programming system is a full implementation of the Scheme
language which conforms to the R4RS, R5RS, R7RS and IEEE Scheme
standards.  It consists of two main programs: `gsi', the Gambit Scheme
interpreter, and `gsc', the Gambit Scheme compiler.

   The Gambit Scheme compiler translates Scheme code to another target
language, currently C or JavaScript.  The C target is the most mature
and it offers portability and fast execution.  The JavaScript target
allows writing web apps in Scheme.

   Most of the Gambit system, including the interpreter and compiler, is
written in Scheme and compiled to portable C code using the compiler.
The high portability of the generated C code allows the interpreter,
compiler and user programs to be easily compiled and executed on any
platform for which a C compiler is available.  With appropriate
declarations in the source code the compiled Scheme programs run
roughly as fast as equivalent C programs.

   For the most up to date information on Gambit and related resources
please visit the Gambit web page at `https://gambitscheme.org'.  Issues
should be reported on the github source code repository
`https://github.com/gambit/gambit'.

1.1 Accessing the system files
==============================

Files related to Gambit, such as executables, libraries and header
files, are stored in multiple "Gambit installation directories".
Gambit may be installed on a system according to two different
installation models.

   In the first model there is a single directory where all the Gambit
installation directories are stored.  This "central installation
directory" is typically `/usr/local/Gambit' under UNIX,
`/Library/Gambit' under macOS and `C:/Program Files/Gambit' under
Microsoft Windows.  This may have been overridden when the system was
built with the command `configure --prefix=/my/Gambit'.  If the system
was built with the command `configure --enable-multiple-versions' then
the central installation directory is `prefix/version', where `version'
is the system version string (e.g. `v4.9.4' for Gambit v4.9.4).
Moreover, `prefix/current' will be a symbolic link which points to the
central installation directory.  In this model, the Gambit installation
directory named X is simply the subdirectory X of the central
installation directory.

   In the second model some or all of the Gambit installation
directories are stored in installation specific directories.  The
location of these directories is assigned when the system is built
using the command `configure --bindir=/my/bin --includedir=/my/include
--libdir=/my/lib'.

   The advantage of the first model is that it is easy to have multiple
versions of Gambit coexist and to remove all the files of a given
version.  However, the second model may be necessary to conform to the
package installation conventions of some operating systems.

   Executable programs such as the interpreter `gsi' and compiler `gsc'
can be found in the `bin' installation directory.  Adding this
directory to the `PATH' environment variable allows these programs to
be started by simply entering their name.  This is done automatically
by the macOS and Microsoft Windows installers.

   The runtime library is located in the `lib' installation directory.
When the system's runtime library is built as a shared-library (with
the command `configure --enable-shared') all programs built with
Gambit, including the interpreter and compiler, need to find this
library when they are executed and consequently this directory must be
in the path searched by the system for shared-libraries.  This path is
normally specified through an environment variable which is
`LD_LIBRARY_PATH' on most versions of UNIX, `LIBPATH' on AIX,
`SHLIB_PATH' on HPUX, `DYLD_LIBRARY_PATH' on macOS, and `PATH' on
Microsoft Windows.  If the shell is `sh', the setting of the path can be
made for a single execution by prefixing the program name with the
environment variable assignment, as in:

     $ LD_LIBRARY_PATH=/usr/local/Gambit/lib gsi

   A similar problem exists with the Gambit header file `gambit.h',
located in the `include' installation directory.  This header file is
needed for compiling Scheme programs with the Gambit compiler.  When
the C compiler is being called explicitly it may be necessary to use a
`-I<DIR>' command line option to indicate where to find header files
and a `-L<DIR>' command line option to indicate where to find libraries.

   Access to both of these files can be simplified by creating a link to
them in the appropriate system directories (special privileges may
however be required):

     $ ln -s /usr/local/Gambit/lib/libgambit.a /usr/lib # name may vary
     $ ln -s /usr/local/Gambit/include/gambit.h /usr/include

   Alternatively these files can be copied or linked in the directory
where the C compiler is invoked (this requires no special privileges).

   Another approach is to set some environment variables which are used
to tell the C compiler where to find header files and libraries.  For
example, the following settings can be used for the `gcc' C compiler:

     $ export LIBRARY_PATH=/usr/local/Gambit/lib
     $ export CPATH=/usr/local/Gambit/include

   Note that this may have been done by the installation process.  In
particular, the macOS and Microsoft Windows prebuilt installers set up
the environment so that the `gcc' compiler finds these files
automatically.

2 The Gambit Scheme interpreter
*******************************

Synopsis:

     gsi [-:RUNTIMEOPTION,...] [-i] [-f] [-h] [-help] [-v]
         [[-] [-e EXPRESSIONS] [-install] [-uninstall] [-update]
          [SEARCH-DIRECTORY-OR-MODULE-OR-FILE]]...

   The interpreter is executed in "batch mode" when the command line
contains a module or file or a `-', or `-e' option.  The interpreter is
executed in "module management mode" when the command line contains the
`-install', `-uninstall', or `-update' option.  Otherwise the
interpreter is executed in "interactive mode".  The `-i' option is
ignored by the interpreter.  The initialization file will be examined
unless the `-f' option is present (*note GSI customization::).  The
`-h' and `-help' options print brief usage information on standard
output and exit.  The `-v' option prints the system version string,
system time stamp, operating system type, and configure script options
on standard output and exits.  Runtime options are explained in *Note
Runtime options::.

2.1 Interactive mode
====================

In interactive mode a read-eval-print loop (REPL) is started for the
user to interact with the interpreter.  At each iteration of this loop
the interpreter displays a prompt, reads a command and executes it.
The commands can be expressions to evaluate (the typical case) or
special commands related to debugging, for example `,q' to terminate
the process (for a complete list of commands see *Note Debugging:: or
use the `,help' command).  Most commands produce some output, such as
the value or error message resulting from an evaluation.

   The input and output of the interaction is done on the "interaction
channel".  The interaction channel can be specified through the runtime
options but if none is specified the system uses a reasonable default
that depends on the system's configuration.  Typically the program's
standard input and output are used as the interaction channel.  When
using the runtime option `-:debug=c', the interaction channel is the
user's "console", also known as the "controlling terminal" in the UNIX
world.

   When the REPL starts, the ports associated with
`(current-input-port)', `(current-output-port)' and
`(current-error-port)' all refer to the interaction channel.

   Expressions are evaluated in the global "interaction environment".
The interpreter adds to this environment any definition entered using
the `define' and `define-macro' special forms.  Once the evaluation of
an expression is completed, the value or values resulting from the
evaluation are output to the interaction channel by the pretty printer.
The special "void object" is not output.  This object is returned by
most procedures and special forms which are defined as returning an
unspecified value (e.g. `write', `set!', `define').

   Here is a sample interaction with `gsi':

     $ gsi
     Gambit v4.9.4

     > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1)))))
     > (map fact '(1 2 3 4 5 6))
     (1 2 6 24 120 720)
     > (values (fact 10) (fact 40))
     3628800
     815915283247897734345611269596115894272000000000
     > ,q

   What happens when errors occur is explained in *Note Debugging::.

2.2 Batch mode
==============

In batch mode the command line arguments denote modules and files to
execute, REPL interactions to start (`-' option), and expressions to be
evaluated (`-e' option).  Those options can be interspersed with the
search directories, modules, and files on the command line and can
occur multiple times.

   In addition to these options the command line may contain 3 types of
non-options: "search directories", "modules", and "files".

"Search directories"
     Search directories are locations in the file system that are
     searched to resolve references to modules.  Any command line
     argument that ends with a path separator or a `.' is treated as a
     search directory.  By default the module search order is initially
     `~~lib' (which contains builtin modules) followed by `~~userlib'
     (which contains user installed modules and is typically the
     directory `.gambit_userlib' in the user's home directory).  Search
     directories on the command line are added to the front of the
     search order, and thus take precedence over the default module
     search order.

"Modules"
     Modules are either "unversioned" or "versioned" (managed by the
     `git' version-control system).  There are two flavors of versioned
     modules: "hosted" modules have a `git' repository on a network
     accessible repository manager site such as `github.com' and
     `gitlab.com', and "local" modules have a `git' repository on the
     local file system.  Module names have a syntax similar to the
     paths used to identify files.  They consist of one or more
     non-empty parts separated by `/'.  The last part may end with a
     suffix of the form `@VERSION'.  Only the first part and version
     may contain `.', otherwise only the characters a-z, A-Z, 0-9, `-',
     and `_' are permitted.  If there are at least 3 parts and the
     first part contains at least one `.'  and no `_', then it refers
     to a hosted module (1st part = host, 2nd part = account, 3rd part
     = repository name).  For example `github.com/gambit/hello@1.0' is
     a hosted module reference.  Otherwise it refers to a local
     versioned module or an unversioned module, for example `foobar' or
     `A/B/C/D'.

"Files"
     Files are simple code containers located on the local file system.
     They are also identified by a path.  If a path is a valid module
     or file, it is interpreted as a module.  Note that a path with a
     last component containing an extension, such as `.scm', and no
     `@', is always interpreted as a file.


   The interpreter processes the command line arguments from left to
right.  Search directories are added to the head of the module search
order.  Files are executed using the `load' procedure.  Modules are
requested using the `##demand-module' special form (this form is
explained in *Note Modules::, but essentially it causes that module to
be searched in the module search order and executed once).  The `-e'
option uses the `eval' procedure to evaluate expressions in the global
interaction environment.  After this processing the interpreter exits.

   The ports associated with `(current-input-port)',
`(current-output-port)' and `(current-error-port)' initially refer
respectively to the standard input (`stdin'), standard output
(`stdout') and the standard error (`stderr') of the interpreter.  This
is true even in REPLs started with the `-' option.  The usual
interaction channel is still used to read expressions and commands and
to display results.  This makes it possible to use REPLs to debug
programs which read the standard input and write to the standard
output, even when these have been redirected.

   Here is a sample use of the interpreter in batch mode, under UNIX:

     $ cat h.scm
     (display "hello") (newline)
     $ cat w.six
     display("world"); newline();
     $ gsi h.scm - w.six -e "(pretty-print 1)(pretty-print 2)"
     hello
     > (define (display x) (write (reverse (string->list x))))
     > ,c
     (#\d #\l #\r #\o #\w)
     1
     2
     $ gsi . h w   # add . to search order to load modules h and w
     hello
     world

2.3 Module management mode
==========================

Package management operations are executed using the command line
options `-install', `-uninstall', and `-update' which respectively
install, uninstall and update packages.  Package installation is
explained in detail in *Note Modules::, but here are a few examples:

     $ gsi -install github.com/gambit/hello
     installing github.com/gambit/hello to /Users/feeley/.gambit_userlib/
     $ gsi github.com/gambit/hello@1.0
     hello world!
     $ gsi -uninstall github.com/gambit/hello
     uninstalling github.com/gambit/hello from /Users/feeley/.gambit_userlib/

2.4 Customization
=================

There are two ways to customize the interpreter.  When the interpreter
starts off it tries to execute a `(load "~~lib/gambext")' (for an
explanation of how file names are interpreted see *Note Host
environment::).  An error is not signaled when the file does not exist.
Interpreter extensions and patches that are meant to apply to all
users and all modes should go in that file.

   Extensions which are meant to apply to a single user or to a specific
working directory are best placed in the "initialization file", which
is a file containing Scheme code.  In all modes, the interpreter first
tries to locate the initialization file by searching the following
locations: `.gambini' and `~/.gambini' (with no extension, a `.sld'
extension, a `.scm' extension, and a `.six' extension in that order).
The first file that is found is examined as though the expression
`(include INITIALIZATION-FILE)' had been entered at the read-eval-print
loop where INITIALIZATION-FILE is the file that was found.  Note that
by using an `include' the macros defined in the initialization file
will be visible from the read-eval-print loop (this would not have been
the case if `load' had been used).  The initialization file is not
searched for or examined when the `-f' option is specified.

2.5 Process exit status
=======================

The status is zero when the interpreter exits normally and is nonzero
when the interpreter exits due to an error.  Here is the meaning of the
exit statuses:

`0'
     The execution of the primordial thread (i.e. the main thread) did
     not encounter any error.  It is however possible that other threads
     terminated abnormally (by default threads other than the primordial
     thread terminate silently when they raise an exception that is not
     handled).

`64'
     The runtime options or the environment variable `GAMBOPT'
     contained a syntax error or were invalid.

`70'
     This normally indicates that an exception was raised in the
     primordial thread and the exception was not handled.

`71'
     There was a problem initializing the runtime system, for example
     insufficient memory to allocate critical tables.


   For example, if the shell is `sh':

     $ gsi -e "(pretty-print (expt 2 100))"
     1267650600228229401496703205376
     $ echo $?
     0
     $ gsi -e "(pretty-print (expo 2 100))"
     *** ERROR IN (string)@1.16 -- Unbound variable: expo
     $ echo $?
     70
     $ gsi -:debug=0 -e "(pretty-print (expo 2 100))"
     $ echo $?
     70
     $ gsi -:debug=0,unknown # try to use an unknown runtime option
     $ echo $?
     64
     $ gsi -:debug=0 nonexistent.scm # try to load a file that does not exist
     $ echo $?
     70
     $ gsi nonexistent.scm
     *** ERROR IN ##load-module-or-file -- No such file or directory
     (load "nonexistent.scm")
     $ echo $?
     70

   Note the use of the runtime option `-:debug=0' that prevents error
messages from being output.

2.6 Scheme scripts
==================

The `load' procedure treats specially files that begin with the two
characters `#!' and `@;'.  Such files are called "script files" and the
first line is called the "script line".  In addition to indicating that
the file is a script, the script line provides information about the
source code language to be used by the `load' procedure.  After the two
characters `#!' and `@;' the system will search for the first substring
matching one of the following language specifying tokens:

`scheme-r4rs'
     R4RS language with prefix syntax, case-insensitivity, keyword
     syntax not supported

`scheme-r5rs'
     R5RS language with prefix syntax, case-insensitivity, keyword
     syntax not supported

`scheme-ieee-1178-1990'
     IEEE 1178-1990 language with prefix syntax, case-insensitivity,
     keyword syntax not supported

`scheme-srfi-0'
     R5RS language with prefix syntax and SRFI 0 support (i.e.
     `cond-expand' special form), case-insensitivity, keyword syntax
     not supported

`gsi-script'
     Full Gambit Scheme language with prefix syntax, case-sensitivity,
     keyword syntax supported

`gsc-script'
     Full Gambit Scheme language with prefix syntax, case-sensitivity,
     keyword syntax supported

`six-script'
     Full Gambit Scheme language with infix syntax, case-sensitivity,
     keyword syntax supported


   If a language specifying token is not found, `load' will use the
same language as a nonscript file (i.e. it uses the file extension and
runtime system options to determine the language).

   After processing the script line, `load' will parse the rest of the
file (using the syntax of the language indicated) and then execute it.
When the file is being loaded because it is an argument on the
interpreter's command line, the interpreter will:

   * Setup the `command-line' procedure so that it returns a list
     containing the expanded file name of the script file and the
     arguments following the script file on the command line.  This is
     done before the script is executed.  The expanded file name of the
     script file can be used to determine the directory that contains
     the script (i.e. `(path-directory (car (command-line)))').

   * After the script is loaded the procedure `main' is called with the
     command line arguments.  The way this is done depends on the
     language specifying token.  For `scheme-r4rs', `scheme-r5rs',
     `scheme-ieee-1178-1990', and `scheme-srfi-0', the `main' procedure
     is called with the equivalent of `(main (cdr (command-line)))' and
     `main' is expected to return a process exit status code in the
     range 0 to 255.  This conforms to the "Running Scheme Scripts on
     Unix SRFI" (SRFI 22).  For `gsi-script' and `six-script' the `main'
     procedure is called with the equivalent of `(apply main (cdr
     (command-line)))' and the process exit status code is 0 (`main''s
     result is ignored).  The Gambit system has a predefined `main'
     procedure which accepts any number of arguments and returns 0, so
     it is perfectly valid for a script to not define `main' and to do
     all its processing with top-level expressions (examples are given
     in the next section).

   * When `main' returns, the interpreter exits.  The command line
     arguments after a script file are consequently not processed
     (however they do appear in the list returned by the `command-line'
     procedure, after the script file's expanded file name, so it is up
     to the script to process them).


2.6.1 Scripts under UNIX and macOS
----------------------------------

Under UNIX and macOS, the Gambit installation process creates the
executable `gsi' and also the executables `six', `gsi-script',
`six-script', `scheme-r5rs', `scheme-srfi-0', etc as links to `gsi'.  A
Scheme script need only start with the name of the desired Scheme
language variant prefixed with `#!' and the directory where the Gambit
executables are stored.  This script should be made executable by
setting the execute permission bits (with a `chmod +x SCRIPT').  Here
is an example of a script which lists on standard output the files in
the current directory:

     #!/usr/local/Gambit/bin/gsi-script
     (for-each pretty-print (directory-files))

   Here is another UNIX script, using the Scheme infix syntax extension,
which takes a single integer argument and prints on standard output the
numbers from 1 to that integer:

     #!/usr/local/Gambit/bin/six-script

     function main(n_str)
     {
       scmobj n = \string->number(n_str);
       for (scmobj i=1; i<=n; i++)
         \pretty-print(i);
     }

   For maximal portability it is a good idea to start scripts indirectly
through the `/usr/bin/env' program, so that the executable of the
interpreter will be searched in the user's `PATH'.  This is what SRFI
22 recommends.  For example here is a script that mimics the UNIX `cat'
utility for text files:

     #!/usr/bin/env gsi-script

     (define (display-file filename)
       (display (call-with-input-file filename
                  (lambda (port)
                    (read-line port #f)))))

     (for-each display-file (cdr (command-line)))

2.6.2 Scripts under Microsoft Windows
-------------------------------------

Under Microsoft Windows, the Gambit installation process creates the
executable `gsi.exe' and `six.exe' and also the batch files
`gsi-script.bat', `six-script.bat', `scheme-r5rs.bat',
`scheme-srfi-0.bat', etc which simply invoke `gsi.exe' with the same
command line arguments.  A Scheme script need only start with the name
of the desired Scheme language variant prefixed with `@;'.  A UNIX
script can be converted to a Microsoft Windows script simply by
changing the script line and storing the script in a file whose name
has a `.bat' or `.cmd' extension:

     @;gsi-script %~f0 %*
     (display "files:\n")
     (pretty-print (directory-files))

   Note that Microsoft Windows always searches executables in the user's
`PATH', so there is no need for an indirection such as the UNIX
`/usr/bin/env'.  However the script line must end with `%~f0 %*' to
pass the expanded filename of the script and command line arguments to
the interpreter.

2.6.3 Compiling scripts
-----------------------

A script file can be compiled using the Gambit Scheme compiler (*note
GSC::) into a standalone executable.  The script line will provide
information to the compiler on which language to use.  The script line
also provides information on which runtime options to use when
executing the compiled script.  This is useful to set the default
runtime options of an executable program.

   The compiled script will be executed similarly to an interpreted
script (i.e. the list of command line arguments returned by the
`command-line' procedure and the invocation of the `main' procedure).

   For example:

     $ cat square.scm
     #!/usr/local/Gambit/bin/gsi-script -:debug=0
     (define (main arg)
       (pretty-print (expt (string->number arg) 2)))
     $ gsi square 30        # gsi will load square.scm
     900
     $ gsc -exe square      # compile the script to a standalone program
     $ ./square 30
     900
     $ ./square 1 2 3       # too many arguments to main
     $ echo $?
     70
     $ ./square -:debug=1 1 2 3  # ask for error message
     *** ERROR IN ##start-main -- Wrong number of arguments passed to procedure
     (main "1" "2" "3")

3 The Gambit Scheme compiler
****************************

Synopsis:

     gsc [-:RUNTIMEOPTION,...] [-i] [-f] [-h] [-help] [-v]
         [-target TARGET]
         [-prelude EXPRESSIONS] [-postlude EXPRESSIONS]
         [-dynamic] [-exe] [-obj]
         [-nb-gvm-regs N] [-nb-arg-regs N] [-compactness LEVEL]
         [-cc COMPILER] [-cc-options OPTIONS]
         [-ld-options-prelude OPTIONS] [-ld-options OPTIONS]
         [-pkg-config PKG-CONFIG-ARGS] [-pkg-config-path PKG-CONFIG-PATH]
         [-warnings] [-verbose] [-report] [-expansion] [-gvm] [-cfg] [-dg]
         [-debug] [-debug-location] [-debug-source]
         [-debug-environments] [-track-scheme]
         [-o OUTPUT] [-c] [-keep-temp] [-link] [-flat] [-l BASE]
         [-module-ref MODULE-REF] [-linker-name LINKER-NAME]
         [[-] [-e EXPRESSIONS] [-preload] [-nopreload]
          [SEARCH-DIRECTORY-OR-MODULE-OR-FILE]]...

   The `-h' and `-help' options print brief usage information on
standard output and exit.  The `-v' option prints the system version
string, system time stamp, operating system type, and configure script
options on standard output and exits.

   The `-i' option can be used to force `gsc' to process the command
line like the interpreter.  The only difference with the interpreter is
that the compilation related procedures listed in this chapter are also
available (i.e. `compile-file', `compile-file-to-target', etc).

3.1 Interactive mode
====================

When no command line argument is present other than options `gsc'
behaves like `gsi' in interactive mode.

3.2 Customization
=================

Like the interpreter, the compiler will examine the initialization file
unless the `-f' option is specified.  Runtime options are explained in
*Note Runtime options::.

3.3 Batch mode
==============

In batch mode `gsc' accepts on the command line 3 types of non-options
which are processed from left to right: "search directories",
"modules", and "files".  Search directories are added to the list of
module search order directories.  Every command line argument that is
the name of a module that is found in the list of module search order
directories will cause that module to be compiled.  Similarly, file
names (with either no extension, or a C file extension, or some other
extension) on the command line will cause that file to be compiled.
The compilation is done for the target language specified with the
`-target' TARGET option.  TARGET is either `js', for JavaScript, or
`C', which is the default if no target language is specified.

   The recognized C file extensions are `.c', `.C', `.cc', `.cp',
`.cpp', `.CPP', `.cxx', `.c++', `.m', `.M', and `.mm'.

   The extension can be omitted from a file name when the Scheme file
has a `.scm', `.sld' or `.six' extension.  When the extension of the
Scheme file is `.six' the content of the file will be parsed using the
Scheme infix syntax extension (see *Note Scheme infix syntax
extension::). Otherwise, `gsc' will parse the Scheme file using the
normal Scheme prefix syntax.  Files with a C file extension must have
been previously produced by `gsc' with the `C' target and the `-c'
option, and are used by the `C' target Gambit linker.

   For each Scheme file the compiler creates a file of target code,
either `FILE.c' or `FILE.js' for the `C' and `js' targets respectively.
The file's name is the same as the Scheme file, but the extension is
changed to `.c' or `.js' as appropriate.  By default the file is
created in the same directory as the Scheme file.  This default can be
overridden with the compiler's `-o' option.

   The files of target code produced by the compiler serve two purposes.
They will be processed by a C compiler or JavaScript VM, and they also
contain information to be read by Gambit's linker to generate a "link
file".  The link file is a file of target code that collects various
linking information for a group of modules, such as the set of all
symbols and global variables used by the modules.  The linker is only
invoked when the `-link' or `-exe' options appear on the command line.

   Compiler options must be specified before the first file name and
after the `-:' runtime option (*note Runtime options::).  If present,
the `-i', `-f', and `-v' compiler options must come first.  The
available options are:

`-i'
     Force interpreter mode.

`-f'
     Do not examine the initialization file.

`-h / -help'
     Print brief usage information on standard output and exit.

`-v'
     Print the system version string, system time stamp, operating
     system type, and configure script options on standard output and
     exit.

`-target TARGET'
     Select the target language.

`-prelude EXPRESSIONS'
     Add expressions to the top of the source code being compiled.

`-postlude EXPRESSIONS'
     Add expressions to the bottom of the source code being compiled.

`-cc COMPILER'
     Select specific C compiler.

`-cc-options OPTIONS'
     Add options to the command that invokes the C compiler.

`-ld-options-prelude OPTIONS'
     Add options to the command that invokes the C linker.

`-ld-options OPTIONS'
     Add options to the command that invokes the C linker.

`-pkg-config PKG-CONFIG-ARGS'
     Use the `pkg-config' program to determine options for the C
     compiler and C linker.

`-pkg-config-path PKG-CONFIG-PATH'
     Add a path to the `PKG_CONFIG_PATH' environment variable.

`-warnings'
     Display warnings.

`-verbose'
     Display a trace of the compiler's activity.

`-report'
     Display a global variable usage report.

`-expansion'
     Display the source code after expansion.

`-gvm'
     Generate a listing of the GVM code.

`-cfg'
     Generate a control flow graph of the GVM code.

`-dg'
     Generate a dependency graph.

`-debug'
     Include all debugging information in the code generated.

`-debug-location'
     Include source code location debugging information in the code
     generated.

`-debug-source'
     Include the source code debugging information in the code
     generated.

`-debug-environments'
     Include environment debugging information in the code generated.

`-track-scheme'
     Generate `#line' directives referring back to the Scheme code.

`-o OUTPUT'
     Set name of output file or directory where output file(s) are
     written.

`-dynamic'
     Compile Scheme source files to dynamically loadable object files
     (this is the default).

`-exe'
     Compile Scheme source files to an executable program (machine code
     or script).

`-obj'
     Compile Scheme source files to object files by invoking the C
     compiler.

`-keep-temp'
     Keep any intermediate files that are generated.

`-c'
     Compile Scheme source files to target code without generating a
     link file.

`-link'
     Compile Scheme source files to target code and generate a link
     file.

`-flat'
     Generate a flat link file instead of the default incremental link
     file.

`-l BASE'
     Specify the link file of the base library to use for the link.

`-module-ref MODULE-REF'
     Specify the reference of the generated module.

`-linker-name LINKER-NAME'
     Specify the name of the low-level initialization function exported
     by the module.

`-preload'
     Turn on `preload' linker bit.

`-nopreload'
     Turn off `preload' linker bit.  Start REPL interaction.

`-e EXPRESSIONS'
     Evaluate expressions in the interaction environment.

`-nb-gvm-regs N'
     Specify the number of available Gambit virtual machine registers.

`-nb-arg-regs N'
     Specify the number of procedure call parameters passed in Gambit
     virtual machine registers.

`-compactness LEVEL'
     Specify the compactness of the generated code.

   The `-i' option forces the compiler to process the remaining command
line arguments like the interpreter.

   The `-target' option selects the target language of the compilation.
It is either `js' for JavaScript, or `C' for C (which is the default).

   The `-prelude' option adds the specified expressions to the top of
the source code being compiled.  It can appear multiple times.  The
main use of this option is to supply declarations on the command line.
For example the following invocation of the compiler will compile the
file `bench.scm' in unsafe mode:

     $ gsc -prelude "(declare (not safe))" bench.scm

   The `-postlude' option adds the specified expressions to the bottom
of the source code being compiled.  It can appear multiple times.  The
main use of this option is to supply the expression that will start the
execution of the program.  For example:

     $ gsc -postlude "(start-bench)" bench.scm

   The `-cc' option is only meaningful when the `C' target is selected.
The `-cc' option selects the specified C compiler for compiling the
generated C code.  When this option is used, the default C compiler
options that were determined to be needed by the configure script are
nullified because they are very likely to be invalid for the specified
C compiler.  Any options needed for this C compiler should be specified
explicitly using the `-cc-options', `-ld-options-prelude', and
`-ld-options' options.  For example:

     $ gsc -cc clang -cc-options "-O0 -bundle" bench.scm # clang on macOS
     $ gsc -cc tcc -cc-options -shared bench.scm         # tcc on linux

   The `-cc-options' option is only meaningful when the `C' target is
selected and a dynamically loadable object file is being generated
(neither the `-c' or `-link' options are used).  It can appear multiple
times.  The `-cc-options' option adds the specified options to the
command that invokes the C compiler.  The main use of this option is to
specify the include path, some symbols to define or undefine, the
optimization level, or any C compiler option that is different from the
default.  For example:

     $ gsc -cc-options "-U___SINGLE_HOST -O2 -I../include" bench.scm

   The `-ld-options-prelude' and `-ld-options' options are only
meaningful when the `C' target is selected and a dynamically loadable
object file is being generated (neither the `-c' or `-link' options are
used).  They can appear multiple times.  The `-ld-options-prelude' and
`-ld-options' options add the specified options to the command that
invokes the C linker (the options in LD-OPTIONS-PRELUDE are passed to
the C linker before the input file and the options in LD-OPTIONS are
passed after).  The main use of this option is to specify additional
object files or libraries that need to be linked, or any C linker
option that is different from the default (such as the library search
path and flags to select between static and dynamic linking).  For
example:

     $ gsc -ld-options "-L/usr/X11R6/lib -lX11 -dynamic" app.scm

   The `-pkg-config' is only meaningful when the `C' target is
selected.  The `-pkg-config' option will cause the `pkg-config' program
to be invoked to determine the options to add to the command that
invokes the C compiler and C linker.  It can appear multiple times.
The `pkg-config' program is passed the arguments in the string
PKG-CONFIG-ARGS in addition to either `--cflags' or `--libs'.  It is
typical for PKG-CONFIG-ARGS to be the name of a system library, such as
`"sqlite3"', but other `pkg-config' options can be specified, such as
`"--static sqlite3"'.  The `-pkg-config-path' option adds a path to the
`PKG_CONFIG_PATH' environment variable for use by the `pkg-config'
program to find `.pc' files.  For example:

     $ gsc -pkg-config "x11" -pkg-config-path "/usr/share/pkgconfig" app.scm

   The `-warnings' option displays on standard output all warnings that
the compiler may have.

   The `-verbose' option displays on standard output a trace of the
compiler's activity.

   The `-report' option displays on standard output a global variable
usage report.  Each global variable used in the program is listed with
4 flags that indicate whether the global variable is defined,
referenced, mutated and called.

   The `-expansion' option displays on standard output the source code
after expansion and inlining by the front end.

   The `-gvm' option generates a listing of the intermediate code for
the "Gambit Virtual Machine" (GVM) of each Scheme file on `FILE.gvm'.

   The `-cfg' option generates a visual representation of the control
flow graph of the intermediate code for the "Gambit Virtual Machine"
(GVM) of each Scheme file on `FILE.cfg'.  The file is suitable for
processing with the "dot" program.  For example, to generate the PDF
file `FILE.cfg.pdf' from `FILE.cfg' the following command can be used:

     $ dot -O -Tpdf FILE.cfg

   The `-dg' option generates a visual representation of the dependency
graph of each Scheme file on `FILE.dg'.  The file is suitable for
processing with the "dot" program.  For example, to generate the PDF
file `FILE.dg.pdf' from `FILE.dg' the following command can be used:

     $ dot -O -Tpdf FILE.dg

   The `-debug' option causes all kinds of debugging information to be
saved in the code generated.  See the documentation of the `debug'
declaration for details.

   The `-debug-location' option causes source code location debugging
information to be saved in the code generated.  See the documentation
of the `debug-location' declaration for details.

   The `-debug-source' option causes source code debugging information
to be saved in the code generated.  See the documentation of the
`debug-source' declaration for details.

   The `-debug-environments' option causes environment debugging
information to be saved in the code generated.  See the documentation
of the `debug-environments' declaration for details.

   The `-track-scheme' option is only meaningful when the `C' target is
selected.  The `-track-scheme' option causes the generation of `#line'
directives that refer back to the Scheme source code.  This allows the
use of a C debugger or profiler to debug Scheme code.

   The `-o' option sets the filename of the output file, or the
directory in which the output file(s) generated by the compiler are
written.

   If the `-link' or `-exe' options appear on the command line, the
Gambit linker is invoked to generate the link file from the set of
files specified on the command line or produced by the Gambit compiler.
By default the link file is named after the last file on the
compiler's command line.  If the last file stripped of it's extension
is `LAST' then the link file is `LAST_.c' for the `C' target and
`LAST_.js' for the `js' target.  When the `-c' option is specified, the
Scheme source files are compiled to target files without invoking the
linker, which is useful for separate compilation of modules.  When the
`-exe' option is specified, the generated target files and link file
are combined to produce an executable program whose name defaults to
`LAST' on Unix, and `LAST.exe' or `LAST.bat' on Windows depending on
whether a machine code executable or script is produced.  When the `C'
target is selected and the `-obj' option is specified, the generated C
files are compiled using the C compiler to produce object files (`.o'
or `.obj' extensions).  If neither the `-link', `-c', `-exe', or `-obj'
options appear on the command line, the Scheme source files are
compiled to dynamically loadable object files (`.oN' extension).  The
`-keep-temp' option will prevent the deletion of any intermediate files
that are generated.  Note that in this case the intermediate file will
be generated in the same directory as the Scheme source file even if
the `-o' option is used.

   The `-flat' option is only meaningful when a link file is being
generated (i.e. the `-link' or `-exe' options also appear on the
command line).  The `-flat' option directs the Gambit linker to
generate a flat link file.  By default, the linker generates an
incremental link file (see the next section for a description of the
two types of link files).

   The `-l' option is only meaningful when an incremental link file is
being generated (i.e. the `-link' or `-exe' options appear on the
command line and the `-flat' option is absent).  The `-l' option
specifies the link file (without the `.c' or `.js' extension) of the
base library to use for the incremental link.  By default the link file
of the Gambit runtime library is used (i.e. `~~lib/_gambit').

   The `-preload' and `-nopreload' options are only meaningful when a
link file is being generated.  The `-preload' option turns on the
`preload' linker bit for the modules that follow on the command line.
The following modules will be loaded unconditionally at program startup
and in command line order (this is the default for compatibility with
how legacy modules have been handled in the past).  The `-nopreload'
option turns off the `preload' linker bit.  The following modules will
be loaded only to satisfy the module dependencies of the
`##demand-module' form.

   The `-' option starts a REPL interaction.

   The `-e' option evaluates the specified expressions in the
interaction environment.

   The `-nb-gvm-regs' option specifies the number of Gambit virtual
machine registers that are available for the generated code.  The
default number depends on configuration options and the target but it
is typically 5.  All modules and the runtime library must be compiled
with the same setting.  This option exists mainly for experimentation
by the developers.  For example:

     $ gsc -nb-gvm-regs 10 -c bench.scm

   The `-nb-arg-regs' option specifies the number of procedure call
parameters passed in Gambit virtual machine registers.  The default
number depends on configuration options and the target but it is
typically 3.  All modules and the runtime library must be compiled with
the same setting.  This option exists mainly for experimentation by the
developers.  For example:

     $ gsc -nb-arg-regs 2 -c bench.scm

   The `-compactness' option selects the level of compactness of the
generated code.  The default level depends on configuration options and
the target but it is typically 5.  Levels from 0 to 5 cause the
generation of increasingly compact code with little or no impact on
execution speed.  Lower values tend to make the generated code more
humanly readable.  Above a level of 5 the compiler will trade execution
speed for saving code space.  The detailed meaning of this option
depends on the target, some targets may ignore it and some targets may
require all modules and the runtime library to be compiled with the
same compactness level.  For example:

     $ gsc -target js -compactness 0 -c bench.scm

3.4 Link files
==============

Gambit can be used to create programs and libraries of Scheme modules.
This section explains the steps required to do so and the role played
by the link files.

   In general, a program is composed of a set of Scheme modules and
modules in the target language.  Some of the modules are part of the
Gambit runtime library and the other modules are supplied by the user.
When the program is started it must setup various global tables
(including the symbol table and the global variable table) and then
sequentially execute the Scheme modules (more or less as though they
were being loaded one after another).  The information required for
this is contained in one or more "link files" generated by the Gambit
linker from the target files produced by the Gambit compiler.

   The order of execution of the Scheme modules corresponds to the
order of the modules on the command line which produced the link file.
The order is usually important because most modules define variables and
procedures which are used by other modules (for this reason the
program's main computation is normally started by the last module).

   When a single link file is used to contain the linking information of
all the Scheme modules it is called a "flat link file".  Thus a program
built with a flat link file contains in its link file both information
on the user modules and on the runtime library.  This is fine if the
program is to be statically linked but is wasteful in a shared-library
context because the linking information of the runtime library can't be
shared and will be duplicated in all programs (this linking information
typically takes hundreds of kilobytes).

   Flat link files are mainly useful to bundle multiple Scheme modules
to make a runtime library (such as the Gambit runtime library) or to
make a single file that can be loaded with the `load' procedure.

   An "incremental link file" contains only the linking information
that is not already contained in a second link file (the "base" link
file).  Assuming that a flat link file was produced when the runtime
library was linked, a program can be built by linking the user modules
with the runtime library's link file, producing an incremental link
file.  This allows the creation of a shared-library which contains the
modules of the runtime library and its flat link file.  The program is
dynamically linked with this shared-library and only contains the user
modules and the incremental link file.  For small programs this
approach greatly reduces the size of the program because the
incremental link file is small.  A "hello world" program built this way
can be as small as 5 Kbytes.  Note that it is perfectly fine to use an
incremental link file for statically linked programs (there is very
little loss compared to a single flat link file).

   Incremental link files may be built from other incremental link
files.  This allows the creation of shared-libraries which extend the
functionality of the Gambit runtime library.

3.4.1 Building an executable program
------------------------------------

The simplest way to create an executable program is to invoke `gsc'
with the `-exe' option.  The compiler will transparently perform all
the steps necessary, including compiling Scheme source files to target
files, generating the link file, and (when the `C' target is selected)
compiling the C files generated to object files and creating the final
executable file using the C linker.  The following example shows how to
use the `C' target to build the executable program `hello.exe' which
contains the two Scheme modules `h.scm' and `w.six'.

     $ cat h.scm
     (display "hello") (newline)
     $ cat w.six
     display("world"); newline();
     $ gsc -o hello.exe -exe h.scm w.six
     h.scm:
     /Users/feeley/gambit/doc/h.c:
     w.six:
     /Users/feeley/gambit/doc/w.c:
     /Users/feeley/gambit/doc/w_.c:
     $ ./hello.exe
     hello
     world

   The detailed steps which are performed can be viewed by setting the
`GAMBUILD_VERBOSE' environment variable to a nonnull value.
Alternatively, `gsc''s `-verbose' option can be used (it implicitly
sets the `GAMBUILD_VERBOSE' environment variable).  For example:

     $ export GAMBUILD_VERBOSE=yes
     $ gsc -o hello.exe -exe h.scm w.six
     h.scm:
     /Users/feeley/gambit/doc/h.c:
     gcc  -O1    -Wno-unused -Wno-write-strings -Wdisabled-optimization
     fwrapv -fno-strict-aliasing -fno-trapping-math -fno-math-errno
     -fschedule-insns2 -foptimize-sibling-calls -fomit-frame-pointer -fPIC
     -fno-common -mpc64   -D___SINGLE_HOST  -I"/usr/local/Gambit/include"
     -c -o 'h.o'  'h.c'
     w.six:
     /Users/feeley/gambit/doc/w.c:
     gcc  -O1    -Wno-unused -Wno-write-strings -Wdisabled-optimization
     -fwrapv -fno-strict-aliasing -fno-trapping-math -fno-math-errno
     -fschedule-insns2 -foptimize-sibling-calls -fomit-frame-pointer -fPIC
     -fno-common -mpc64   -D___SINGLE_HOST  -I"/usr/local/Gambit/include"
     -c -o 'w.o'  'w.c'
     /Users/feeley/gambit/doc/w_.c:
     gcc  -O1    -Wno-unused -Wno-write-strings -Wdisabled-optimization
     -fwrapv -fno-strict-aliasing -fno-trapping-math -fno-math-errno
     -fschedule-insns2 -foptimize-sibling-calls -fomit-frame-pointer -fPIC
     -fno-common -mpc64   -D___SINGLE_HOST  -I"/usr/local/Gambit/include"
     -c -o 'w_.o'  'w_.c'
     gcc     -Wno-unused -Wno-write-strings -Wdisabled-optimization
     -fwrapv -fno-strict-aliasing -fno-trapping-math -fno-math-errno
     -fschedule-insns2 -foptimize-sibling-calls -fomit-frame-pointer -fPIC
     -fno-common -mpc64    -D___SINGLE_HOST  -I"/usr/local/Gambit/include"
     -o 'hello.exe'   'w_.o' 'h.o' 'w.o' "/usr/local/Gambit/lib/libgambit.a"

   Here is the same example using the `js' target showing the creation
of a shell script invoking nodejs:

     $ export GAMBUILD_VERBOSE=yes
     $ gsc -target js -o hello.exe -exe h.scm w.six
     h.scm:
     /Users/feeley/gambit/doc/h.js:
     cat h.js > "h.o"
     w.six:
     /Users/feeley/gambit/doc/w.js:
     cat w.js > "w.o"
     /Users/feeley/gambit/doc/w_.js:
     cat w_.js > "w_.o"
     echo "#! /usr/bin/env node" > "hello.exe"
     cat w_.o h.o w.o "/usr/local/Gambit/lib/_gambit.js" >> "hello.exe"
     chmod +x "hello.exe"

   Using a single invocation of `gsc' with the `-exe' option is
sometimes inappropriate when the build process is more complex, for
example when the program is composed of several separately compiled
modules.  In such a case it is useful to decompose the build process
into smaller compilation steps.  The `hello.exe' executable program
could have been built with the `C' target by separating the generation
of C files from the C compilation and linking:

     $ gsc -c h.scm
     $ gsc -c w.six
     $ gsc -o hello.exe -exe h.c w.c

   When even finer control is desired the `C' target's build process
can be decomposed into smaller steps that invoke the C compiler and
linker explicitly.  This is described in the rest of this section.

   The `gsc' compiler can be invoked to compile each Scheme module into
a C file and to create an incremental link file.  The C files and the
link file must then be compiled with a C compiler and linked (at the
object file level) with the Gambit runtime library and possibly other
libraries (such as the math library and the dynamic loading library).

   Here is for example how a program with three modules (one in C and
two in Scheme) can be built.  The content of the three source files
(`m1.c', `m2.scm' and `m3.scm') is:

     /* File: "m1.c" */
     int power_of_2 (int x) { return 1<<x; }

     ; File: "m2.scm"
     (c-declare "extern int power_of_2 ();")
     (define pow2 (c-lambda (int) int "power_of_2"))
     (define (twice x) (cons x x))

     ; File: "m3.scm"
     (write (map twice (map pow2 '(1 2 3 4)))) (newline)

   The compilation of the two Scheme source files can be done with
three invocations of `gsc':

     $ gsc -c m2.scm        # create m2.c (note: .scm is optional)
     $ gsc -c m3.scm        # create m3.c (note: .scm is optional)
     $ gsc -link m2.c m3.c  # create the incremental link file m3_.c

   Alternatively, the three invocations of `gsc' can be replaced by a
single invocation:

     $ gsc -link m2 m3
     m2:
     m3:

   At this point there will be 4 C files: `m1.c', `m2.c', `m3.c', and
`m3_.c'.  To produce an executable program these files must be compiled
with a C compiler and linked with the Gambit runtime library.  The C
compiler options needed will depend on the C compiler and the operating
system (in particular it may be necessary to add the options
`-I/usr/local/Gambit/include -L/usr/local/Gambit/lib' to access the
`gambit.h' header file and the Gambit runtime library).

   Here is an example under macOS:

     $ uname -srmp
     Darwin 20.6.0 x86_64 i386
     $ gsc -obj m1.c m2.c m3.c m3_.c
     m1.c:
     m2.c:
     m3.c:
     m3_.c:
     $ gcc m1.o m2.o m3.o m3_.o -lgambit
     $ ./a.out
     ((2 . 2) (4 . 4) (8 . 8) (16 . 16))

   Here is an example under Linux:

     $ uname -srmp
     Linux 5.10.0-9-amd64 x86_64 unknown
     $ gsc -obj m1.c m2.c m3.c m3_.c
     m1.c:
     m2.c:
     m3.c:
     m3_.c:
     $ gcc m1.o m2.o m3.o m3_.o -lgambit -lm -ldl -lutil
     $ ./a.out
     ((2 . 2) (4 . 4) (8 . 8) (16 . 16))

3.4.2 Building a loadable library
---------------------------------

To bundle multiple modules into a single object file that can be
dynamically loaded with the `load' procedure, a flat link file is
needed.  The compiler's `-o' option must be used to name the C file
generated as follows.  If the dynamically loadable object file is to be
named `MYFILE.oN' then the `-o' option must set the name of the link
file generated to `MYFILE.oN.c' (note that the `.c' extension could
also be `.cc', `.cpp' or whatever extension is appropriate for C/C++
source files).  The three modules of the previous example can be
bundled by generating a link file in this way:

     $ gsc -link -flat -o foo.o1.c m2 m3
     m2:
     m3:
     *** WARNING -- "cons" is not defined,
     ***            referenced in: ("m2.c")
     *** WARNING -- "map" is not defined,
     ***            referenced in: ("m3.c")
     *** WARNING -- "newline" is not defined,
     ***            referenced in: ("m3.c")
     *** WARNING -- "write" is not defined,
     ***            referenced in: ("m3.c")

   The warnings indicate that there are no definitions (`define's or
`set!'s) of the variables `cons', `map', `newline' and `write' in the
set of modules being linked.  Before `foo.o1' is loaded, these
variables will have to be bound; either implicitly (by the runtime
library) or explicitly.

   When compiling the C files and link file generated, the flag
`-D___DYNAMIC' must be passed to the C compiler and the C compiler and
linker must be told to generate a dynamically loadable shared library.

   Here is an example under macOS:

     $ uname -srmp
     Darwin 20.6.0 x86_64 i386
     $ gsc -link -flat -o foo.o1.c m2 m3 > /dev/null
     m2:
     m3:
     $ gsc -cc-options "-D___DYNAMIC" -obj m1.c m2.c m3.c foo.o1.c
     m1.c:
     m2.c:
     m3.c:
     foo.o1.c:
     $ gcc -bundle m1.o m2.o m3.o foo.o1.o -o foo.o1
     $ gsi foo.o1
     ((2 . 2) (4 . 4) (8 . 8) (16 . 16))

   Here is an example under Linux:

     $ uname -srmp
     Linux 5.10.0-9-amd64 x86_64 unknown
     $ gsc -link -flat -o foo.o1.c m2 m3 > /dev/null
     m2:
     m3:
     $ gsc -cc-options "-D___DYNAMIC" -obj m1.c m2.c m3.c foo.o1.c
     m1.c:
     m2.c:
     m3.c:
     foo.o1.c:
     $ gcc -shared m1.o m2.o m3.o foo.o1.o -o foo.o1
     $ gsi foo.o1
     ((2 . 2) (4 . 4) (8 . 8) (16 . 16))

   Here is a more complex example, under Solaris, which shows how to
build a loadable library `mymod.o1' composed of the files `m4.scm',
`m5.scm' and `x.c' that links to system shared libraries (for
X-windows):

     $ uname -srmp
     SunOS ungava 5.6 Generic_105181-05 sun4m sparc SUNW,SPARCstation-20
     $ gsc -link -flat -o mymod.o1.c m4 m5
     m4:
     m5:
     *** WARNING -- "*" is not defined,
     ***            referenced in: ("m4.c")
     *** WARNING -- "+" is not defined,
     ***            referenced in: ("m5.c")
     *** WARNING -- "display" is not defined,
     ***            referenced in: ("m5.c" "m4.c")
     *** WARNING -- "newline" is not defined,
     ***            referenced in: ("m5.c" "m4.c")
     *** WARNING -- "write" is not defined,
     ***            referenced in: ("m5.c")
     $ gsc -cc-options "-D___DYNAMIC" -obj m4.c m5.c x.c mymod.o1.c
     m4.c:
     m5.c:
     x.c:
     mymod.o1.c:
     $ /usr/ccs/bin/ld -G -o mymod.o1 mymod.o1.o m4.o m5.o x.o -lX11 -lsocket
     $ gsi mymod.o1
     hello from m4
     hello from m5
     (f1 10) = 22
     $ cat m4.scm
     (define (f1 x) (* 2 (f2 x)))
     (display "hello from m4")
     (newline)

     (c-declare #<<c-declare-end
     #include "x.h"
     c-declare-end
     )
     (define x-initialize (c-lambda (char-string) bool "x_initialize"))
     (define x-display-name (c-lambda () char-string "x_display_name"))
     (define x-bell (c-lambda (int) void "x_bell"))
     $ cat m5.scm
     (define (f2 x) (+ x 1))
     (display "hello from m5")
     (newline)

     (display "(f1 10) = ")
     (write (f1 10))
     (newline)

     (x-initialize (x-display-name))
     (x-bell 50) ; sound the bell at 50%
     $ cat x.c
     #include <X11/Xlib.h>

     static Display *display;

     int x_initialize (char *display_name)
     {
       display = XOpenDisplay (display_name);
       return display != NULL;
     }

     char *x_display_name (void)
     {
       return XDisplayName (NULL);
     }

     void x_bell (int volume)
     {
       XBell (display, volume);
       XFlush (display);
     }
     $ cat x.h
     int x_initialize (char *display_name);
     char *x_display_name (void);
     void x_bell (int);

3.4.3 Building a shared-library
-------------------------------

A shared-library can be built using an incremental link file or a flat
link file.  An incremental link file is normally used when the Gambit
runtime library (or some other library) is to be extended with new
procedures.  A flat link file is mainly useful when building a "primal"
runtime library, which is a library (such as the Gambit runtime
library) that does not extend another library.  When compiling the C
files and link file generated, the flags `-D___LIBRARY' and
`-D___SHARED' must be passed to the C compiler.  The flag `-D___PRIMAL'
must also be passed to the C compiler when a primal library is being
built.

   A shared-library `mylib.so' containing the two first modules of the
previous example can be built this way:

     $ uname -srmp
     Linux 5.10.0-9-amd64 x86_64 unknown
     $ gsc -link -o mylib.c m2
     $ gsc -obj -cc-options "-D___SHARED" m1.c m2.c mylib.c
     m1.c:
     m2.c:
     mylib.c:
     $ gcc -shared  m1.o m2.o mylib.o -o mylib.so

   Note that this shared-library is built using an incremental link file
(it extends the Gambit runtime library with the procedures `pow2' and
`twice').  This shared-library can in turn be used to build an
executable program from the third module of the previous example:

     $ gsc -link -l mylib m3
     $ gsc -obj m3.c m3_.c
     m3.c:
     m3_.c:
     $ gcc m3.o m3_.o mylib.so -lgambit
     $ LD_LIBRARY_PATH=.:/usr/local/lib ./a.out
     ((2 . 2) (4 . 4) (8 . 8) (16 . 16))

3.4.4 Other compilation options
-------------------------------

The performance of the code can be increased by passing the
`-D___SINGLE_HOST' flag to the C compiler.  This will merge all the
procedures of a module into a single C procedure, which reduces the
cost of intra-module procedure calls.  In addition the `-O2' option can
be passed to the C compiler.  For large modules, it will not be
practical to specify both `-O2' and `-D___SINGLE_HOST' for typical C
compilers because the compile time will be high and the C compiler
might even fail to compile the program for lack of memory.  It has been
observed that lower levels of optimization (e.g. `-O1') often give
faster compilation and also generate faster code.  It is a good idea to
experiment.

   Normally C compilers will not automatically search
`/usr/local/Gambit/include' for header files so the flag
`-I/usr/local/Gambit/include' should be passed to the C compiler.
Similarly, C compilers/linkers will not automatically search
`/usr/local/Gambit/lib' for libraries so the flag
`-L/usr/local/Gambit/lib' should be passed to the C compiler/linker.
Alternatives are given in *Note Accessing the system files::.

   A variety of flags are needed by some C compilers when compiling a
shared-library or a dynamically loadable library.  Some of these flags
are: `-shared', `-call_shared', `-rdynamic', `-fpic', `-fPIC', `-Kpic',
`-KPIC', `-pic', `+z', `-G'.  Check your compiler's documentation to see
which flag you need.

3.5 Procedures specific to compiler
===================================

The Gambit Scheme compiler features the following procedures that are
not available in the Gambit Scheme interpreter.

 -- procedure: compile-file-to-target FILE [`options:' OPTIONS]
          [`output:' OUTPUT] [`expression:' EXPRESSION]
     The FILE parameter must be a string.  If EXPRESSION is not
     specified, FILE must name an existing file containing Scheme
     source code.  The extension can be omitted from FILE when the
     Scheme file has a `.scm', `.sld' or `.six' extension.  By default,
     this procedure compiles the source file into a file containing C
     code.  A different target language can be selected in the OPTIONS.
     The generated file is named after FILE with the extension
     replaced with `.c' or `.js', as appropriate for the target
     selected.  The name of the generated file can also be specified
     directly with the OUTPUT parameter.  If OUTPUT is a string naming
     a directory then the generated file is created in that directory.
     Otherwise the name of the generated file is OUTPUT.

     Compilation options are specified through the OPTIONS parameter
     which must be an association list.  Any combination of the
     following options can be used: `target', `verbose', `report',
     `expansion', `gvm', `debug', `module-ref', and `linker-name'.

     When EXPRESSION is specified, the FILE parameter is not open or
     read.  Instead, EXPRESSION is used as though it was the content of
     the file.  This makes it possible to compile source code without
     having to create a file to contain the code.  Note that FILE is
     used in error messages and to determine the output file name if
     OUTPUT is not specified.

     When the compilation is successful, `compile-file-to-target'
     returns the name of the file generated.  When there is a
     compilation error, `#f' is returned.

          $ cat h.scm
          (display "hello") (newline)
          $ gsc
          Gambit v4.9.4

          > (compile-file-to-target "h")
          "/Users/feeley/gambit/doc/h.c"

 -- procedure: compile-file FILE [`options:' OPTIONS] [`output:'
          OUTPUT] [`base:' BASE] [`expression:' EXPRESSION]
          [`cc-options:' CC-OPTIONS] [`ld-options-prelude:'
          LD-OPTIONS-PRELUDE] [`ld-options:' LD-OPTIONS]
     The FILE, OPTIONS, OUTPUT, and EXPRESSION parameters have the same
     meaning as for the `compile-file-to-target' procedure, except that
     FILE may be a Scheme source file or a file possibly generated by
     the Gambit Scheme compiler (for example with the
     `compile-file-to-target' procedure).  The CC-OPTIONS parameter is
     a string containing the options to pass to the C compiler and the
     LD-OPTIONS-PRELUDE and LD-OPTIONS parameters are strings
     containing the options to pass to the C linker (the options in
     LD-OPTIONS-PRELUDE are passed to the C linker before the input
     file and the options in LD-OPTIONS are passed after).

     The `compile-file' procedure compiles the source file FILE into an
     object file, which is either a file dynamically loadable using the
     `load' procedure, or a C linkable object file destined to be
     linked with the C linker (for example to create a standalone
     executable program).  The presence of the `obj' option in OPTIONS
     will cause the creation of a C linkable object file and therefore
     the options LD-OPTIONS-PRELUDE and LD-OPTIONS are ignored,
     otherwise a dynamically loadable file is created.  In both cases,
     if FILE is a Scheme source file, the compiler first compiles FILE
     to a C file which is created in the same directory as FILE
     regardless of the OUTPUT parameter.  Then the C file is compiled
     with the C compiler.

     When the compilation is successful, `compile-file' returns the
     name of the object file generated.  When there is a compilation
     error, `#f' is returned.

     The name of the object file can be specified with the OUTPUT
     parameter.  If OUTPUT is a string naming a directory then the
     object file is created in that directory.  Otherwise the name of
     the object file is OUTPUT.

     In the case of a dynamically loadable object file, by default the
     object file is named after FILE with the extension replaced with
     `.oN', where N is a positive integer that acts as a version
     number.  The next available version number is generated
     automatically by `compile-file'.

     When dynamically loaded object files are loaded using the `load'
     procedure, the `.oN' extension can be specified (to select a
     particular version) or omitted (to load the file with a `.oN'
     extension with the highest N consecutively from 1).  When the
     `.oN' extension is not specified and older versions are no longer
     needed, all versions must be deleted and the compilation must be
     repeated (this is necessary because the file name, including the
     extension, is used to name some of the exported symbols of the
     object file).

     Note that dynamically loadable object files can only be generated
     on host operating systems that support dynamic loading.

          $ cat h.scm
          (display "hello") (newline)
          $ gsc
          Gambit v4.9.4

          > (compile-file "h")
          "/Users/feeley/gambit/doc/h.o1"
          > (load "h")
          hello
          "/Users/feeley/gambit/doc/h.o1"
          > (compile-file-to-target "h" output: "h.o99.c")
          "/Users/feeley/gambit/doc/h.o99.c"
          > (compile-file "h.o99.c")
          "/Users/feeley/gambit/doc/h.o99"
          > (load "h.o99")
          hello
          "/Users/feeley/gambit/doc/h.o99"
          > (compile-file-to-target "h")
          "/Users/feeley/gambit/doc/h.c"
          > (compile-file "h.c" options: '(obj))
          "/Users/feeley/gambit/doc/h.o"


 -- procedure: link-incremental MODULE-LIST [`output:' OUTPUT]
          [`linker-name:' LINKER-NAME] [`base:' BASE] [`warnings?:'
          WARNINGS?]
     The first parameter must be a non empty list of strings naming
     Scheme modules to link (the file extension may be omitted).  An
     incremental link file is generated for the modules specified in
     MODULE-LIST.  By default the link file generated is named
     `LAST_.EXT', where LAST is the name of the last module, without
     the file extension, and EXT is the appropriate extension for the
     target.  The name of the generated link file can be specified with
     the OUTPUT parameter.  If OUTPUT is a string naming a directory
     then the link file is created in that directory.  Otherwise the
     name of the link file is OUTPUT.

     The base link file is specified by the BASE parameter, which must
     be a string.  By default the base link file is the Gambit runtime
     library link file `~~lib/_gambit' (with extension appropriate for
     the target).  However, when BASE is supplied it is the name of the
     base link file (the file extension may be omitted).

     The WARNINGS? parameter controls whether warnings are generated
     for undefined references.

     The following example shows how to build the executable program
     `hello' which contains the two Scheme modules `h.scm' and `w.six'.

          $ uname -srmp
          Darwin 8.1.0 Power Macintosh powerpc
          $ cat h.scm
          (display "hello") (newline)
          $ cat w.six
          display("world"); newline();
          $ gsc
          Gambit v4.9.4

          > (compile-file-to-target "h")
          "/Users/feeley/gambit/doc/h.c"
          > (compile-file-to-target "w")
          "/Users/feeley/gambit/doc/w.c"
          > (link-incremental '("h" "w") output: "hello.c")
          "/Users/feeley/gambit/doc/hello_.c"
          > ,q
          $ gsc -obj h.c w.c hello.c
          h.c:
          w.c:
          hello.c:
          $ gcc h.o w.o hello.o -lgambit -o hello
          $ ./hello
          hello
          world


 -- procedure: link-flat MODULE-LIST [`output:' OUTPUT] [`linker-name:'
          LINKER-NAME] [`warnings?:' WARNINGS?]
     The first parameter must be a non empty list of strings naming
     Scheme modules to link (the file extension may be omitted).  The
     first string must be the name of a Scheme module or the name of a
     link file and the remaining strings must name Scheme modules.  A
     flat link file is generated for the modules specified in
     MODULE-LIST.  By default the link file generated is named
     `LAST_.EXT', where LAST is the name of the last module, without
     the file extension, and EXT is the appropriate extension for the
     target.  The name of the generated link file can be specified with
     the OUTPUT parameter.  If OUTPUT is a string naming a directory
     then the link file is created in that directory.  Otherwise the
     name of the link file is OUTPUT.  If a dynamically loadable object
     file is produced from the link file `OUTPUT', then the name of the
     dynamically loadable object file must be `OUTPUT' stripped of its
     file extension.

     The WARNINGS? parameter controls whether warnings are generated
     for undefined references.

     The following example shows how to build the dynamically loadable
     object file `lib.o1' which contains the two Scheme modules
     `m6.scm' and `m7.scm'.

          $ uname -srmp
          Darwin 8.1.0 Power Macintosh powerpc
          $ cat m6.scm
          (define (f x) (g (* x x)))
          $ cat m7.scm
          (define (g y) (+ n y))
          $ gsc
          Gambit v4.9.4

          > (compile-file-to-target "m6")
          "/Users/feeley/gambit/doc/m6.c"
          > (compile-file-to-target "m7")
          "/Users/feeley/gambit/doc/m7.c"
          > (link-flat '("m6" "m7") output: "lib.o1.c")
          *** WARNING -- "*" is not defined,
          ***            referenced in: ("m6.c")
          *** WARNING -- "+" is not defined,
          ***            referenced in: ("m7.c")
          *** WARNING -- "n" is not defined,
          ***            referenced in: ("m7.c")
          "/Users/feeley/gambit/doc/lib.o1.c"
          > ,q
          $ gcc -bundle -D___DYNAMIC m6.c m7.c lib.o1.c -o lib.o1
          $ gsc
          Gambit v4.9.4

          > (load "lib")
          *** WARNING -- Variable "n" used in module "m7" is undefined
          "/Users/feeley/gambit/doc/lib.o1"
          > (define n 10)
          > (f 5)
          35
          > ,q

     The warnings indicate that there are no definitions (`define's or
     `set!'s) of the variables `*', `+' and `n' in the modules
     contained in the library.  Before the library is used, these
     variables will have to be bound; either implicitly (by the runtime
     library) or explicitly.


4 Runtime options
*****************

Both `gsi' and `gsc' as well as executable programs compiled and linked
using `gsc' take a `-:' option which supplies parameters to the runtime
system.  This option must appear first on the command line.  The colon
is followed by a comma separated list of options with no intervening
spaces.  The available options are:

`min-heap='SIZE or the shorthand `m'SIZE
     Set minimum heap size.

`max-heap='SIZE or the shorthand `h'SIZE
     Set maximum heap size.

`live-ratio='RATIO or the shorthand `l'RATIO
     Set the ratio of heap that is live after a garbage collection.

`gambit' or the (deprecated) shorthand `S'
     Select Gambit Scheme mode. This is the default mode.

`r5rs' or the (deprecated) shorthand `s'
     Select R5RS Scheme mode.

`r7rs'
     Select R7RS Scheme mode.

`debug'[`='[OPT...]] or the shorthand `d'[OPT...]
     Set debugging options.

`~~'NAME`='DIRECTORY
     Override the NAME installation directory.

`add-arg='ARGUMENT or the shorthand `+'ARGUMENT
     Add ARGUMENT to the command line before other arguments.

`io-settings='[IO...] or the shorthand `i'[IO...]
     Set general I/O settings.

`file-settings='[IO...] or the shorthand `f'[IO...]
     Set general file I/O settings.

`stdio-settings='[IO...] or the shorthand `-'[IO...]
     Set general stdio settings.

`0'[IO...]
     Set stdin settings.

`1'[IO...]
     Set stdout settings.

`2'[IO...]
     Set stderr settings.

`terminal-settings='[IO...] or the shorthand `t'[IO...]
     Set terminal I/O settings.

`search='[DIR]
     Set or reset module search order.

`whitelist='[SOURCE]
     Set or reset the whitelist of trusted sources for automatic
     installation of hosted modules.

`ask-install='WHEN
     Set automatic installation confirmation mode.


   The `min-heap='SIZE and `max-heap='SIZE options set limits on the
size of the heap.  The SIZE is an integer that may be followed by `G'
(gigabytes), `M' (megabytes), or `K' or nothing (kilobytes).  The heap
will not shrink lower than the minimum heap size which defaults to 0.
The heap will not grow larger than the maximum heap size if it is set
(by default the heap may grow until the virtual memory is exhausted).

   The `live-ratio='RATIO option sets the percentage of the heap that
will be occupied with live objects after the heap is resized at the end
of a garbage collection.  RATIO is an integer between 1 and 100
inclusively indicating the desired percentage.  The garbage collector
resizes the heap to reach this percentage occupation (roughly), within
the limits of the `min-heap' and `max-heap' options.  By default, the
percentage is 50.

   The `gambit', `r5rs' and `r7rs' options configure the runtime system
to conform to Gambit Scheme, R5RS Scheme and R7RS Scheme respectively.
The reader is case-insensitive in `r5rs' mode, and is case-sensitive in
`r7rs' and `gambit' modes.  The reader supports keywords only in
`gambit' mode, which is the default mode.

   The `debug='OPT,... option sets various debugging options.  The
equal sign is followed by a sequence of letters indicating suboptions.

`p'
     Uncaught exceptions will be treated as "errors" in the primordial
     thread only.

`a'
     Uncaught exceptions will be treated as "errors" in all threads.

`r'
     When an "error" occurs a new REPL will be started.

`s'
     When an "error" occurs a new REPL will be started.  Moreover the
     program starts in single-stepping mode.

`q'
     When an "error" occurs the program will terminate with a nonzero
     exit status.

`R'
     When a user interrupt occurs a new REPL will be started.  User
     interrupts are typically obtained by typing <^C>.  Note that with
     some system configurations <^C> abruptly terminates the process.
     For example, under Microsoft Windows, <^C> works fine with the
     standard console but with the MSYS terminal window it terminates
     the process.

`D'
     When a user interrupt occurs it will be deferred until the
     parameter object `current-user-interrupt-handler' is set or bound.

`Q'
     When a user interrupt occurs the program will terminate with a
     nonzero exit status.

LEVEL
     The verbosity level is set to LEVEL, a digit from 0 to 9.  At
     level 0 the runtime system will not display error messages and
     warnings.  At level 1 and above error messages and warnings are
     displayed.  At level 2 and above a backtrace is displayed.  At
     level 3 and above variable bindings are displayed in the backtrace.
     At level 5 and above garbage collection reports are displayed
     during program execution.

`c'
     The REPL interaction channel will be the console.

`-'
     The REPL interaction channel will be standard input and standard
     output.

`+'
     The REPL interaction channel will be standard input and standard
     output and standard error.

`@'[HOST][`:'PORT]
     When a REPL is started by a thread a connection will be established
     with the address HOST:PORT and that will be the REPL's interaction
     channel.  The default HOST is 127.0.0.1 and the default PORT is
     44556.

`$'[INTF][`:'PORT]
     The runtime system will open a socket to listen on port number
     PORT for incoming connections on the network interface with
     address INTF.  The default INTF is 127.0.0.1 and the default PORT
     is 44555.


   The default debugging options are equivalent to `debug=pqQ1-' (i.e.
an uncaught exception in the primordial thread terminates the program
after displaying an error message).  When the option `debug' is used
without suboptions it is equivalent to `debug=prR1-' (i.e. a new REPL
is started only when an uncaught exception occurs in the primordial
thread).  When `gsi' and `gsc' are running the main REPL, the debugging
options are changed to cause errors in the primordial thread and user
interrupts to start a nested REPL.

   The `~~'NAME`='DIRECTORY option overrides the setting of the NAME
installation directory.  If NAME is empty, it will override the central
installation directory.

   The `add-arg='ARGUMENT option adds the text that follows to the
command line before other arguments.

   The option `io-settings='[IO...] sets the default I/O settings of
all types of ports.  The option `file-settings='[IO...] sets the
default I/O settings for ports associated to files.  The option
`stdio-settings='[IO...] sets the default I/O settings for ports
associated to stdio (but finer control is possible with `0'[IO...],
`1'[IO...], and `2'[IO...] that set the I/O settings of stdin, stdout,
and stderr respectively).  The option `terminal-settings='[IO...]
overrides the default I/O settings for ports associated to terminals.
The default character encoding, end-of-line encoding and buffering can
be set.  Moreover, for terminals the line-editing feature can be
enabled or disabled.  Each IO is a one or two letter code as follows:

`A'
     ASCII character encoding.

`1'
     ISO-8859-1 character encoding.

`2'
     UCS-2 character encoding.

`4'
     UCS-4 character encoding.

`6'
     UTF-16 character encoding.

`8'
     UTF-8 character encoding.

`U'
     UTF character encoding with fallback to UTF-8 on input if no BOM
     is present.

`UA'
     UTF character encoding with fallback to ASCII on input if no BOM
     is present.

`U1'
     UTF character encoding with fallback to ISO-8859-1 on input if no
     BOM is present.

`U6'
     UTF character encoding with fallback to UTF-16 on input if no BOM
     is present.

`U8'
     UTF character encoding with fallback to UTF-8 on input if no BOM
     is present.

`L'
     If the `LC_ALL' or `LC_CTYPE' or `LANG' environment variables end
     with `.UTF-8' or `.ISO-8859-1' or `.LATIN-1' (or a variation) set
     the character encoding accordingly.

`c'
     End-of-line is encoded as CR (carriage-return).

`l'
     End-of-line is encoded as LF (linefeed)

`cl'
     End-of-line is encoded as CR-LF.

`u'
     Unbuffered I/O.

`n'
     Line buffered I/O (`n' for "at newline").

`f'
     Fully buffered I/O.

`r'
     Illegal character encoding is treated as an error (exception
     raised).

`R'
     Silently replace illegal character encodings with Unicode
     character #xfffd (replacement character).

`e'
     Enable line-editing (applies to terminals only).

`E'
     Disable line-editing (applies to terminals only).

   The `search='[DIR] option adds DIR to the head of the list of module
search order directories, unless DIR is empty, in which case it is set
to the empty list.  The initial setting of the list of module search
order directories is `~~lib' followed by `~~userlib'.

   When a hosted module can't be found in the directories on the list of
module search order directories it will be automatically installed if
it is from a source on the whitelist of trusted sources, which
initially contains only `github.com/gambit'.  The `whitelist='[SOURCE]
option adds SOURCE to the whitelist, unless SOURCE is empty in which
case the whitelist will be set to the empty list (no source is trusted).

   The `ask-install='WHEN option sets the automatic installation mode
confirmation mode to WHEN, which is one of `always', `repl', and
`never'.  When a hosted module can't be found in the directories on the
list of module search order directories and it is from a source not on
the whitelist the runtime system will ask for installation confirmation
when WHEN is `always', or when a REPL has already been started for the
current thread and WHEN is `repl'.  In the `never' mode the runtime
system will not install the module automatically.  The default mode is
`repl'.

   When a program's execution starts, the runtime system obtains the
runtime options by processing in turn various sources of runtime
options: the defaults, the environment variable `GAMBOPT', the script
line of the source code, and, unless the program is an interpreted
script, the first command line argument of the program.  Any runtime
option can be overriden by a subsequent source of runtime options.  It
is sometimes useful to prevent overriding the runtime options of the
script line.  This can be achieved by starting the script line runtime
options with `-:,'.  In this case the environment variable `GAMBOPT' is
ignored, and the first command line argument of the program is not used
for runtime options (it is treated like a normal command line argument
even if it starts with `-:').

   For example:

     $ export GAMBOPT=debug=0,~~=~/my-gambit2
     $ gsi -e '(pretty-print (path-expand "~~")) (/ 1 0)'
     "/Users/feeley/my-gambit2/"
     $ echo $?
     70
     $ gsi -:debug=1 -e '(pretty-print (path-expand "~~")) (/ 1 0)'
     "/Users/feeley/my-gambit2/"
     *** ERROR IN (string)@1.35 -- Divide by zero
     (/ 1 0)

5 Debugging
***********

5.1 Debugging model
===================

The evaluation of an expression may stop before it is completed for the
following reasons:

  a. An evaluation error has occured, such as attempting to divide by
     zero.

  b. The user has interrupted the evaluation (usually by typing <^C>).

  c. A breakpoint has been reached or `(step)' was evaluated.

  d. Single-stepping mode is enabled.


   When an evaluation stops, a message is displayed indicating the
reason and location where the evaluation was stopped.  The location
information includes, if known, the name of the procedure where the
evaluation was stopped and the source code location in the format
`STREAM@LINE.COLUMN', where STREAM is either a string naming a file or
a symbol within parentheses, such as `(console)'.

   A "nested REPL" is then initiated in the context of the point of
execution where the evaluation was stopped.  The nested REPL's
continuation and evaluation environment are the same as the point where
the evaluation was stopped.  For example when evaluating the expression
`(let ((y (- 1 1))) (* (/ x y) 2))', a "divide by zero" error is
reported and the nested REPL's continuation is the one that takes the
result and multiplies it by two.  The REPL's lexical environment
includes the lexical variable `y'.  This allows the inspection of the
evaluation context (i.e. the lexical and dynamic environments and
continuation), which is particularly useful to determine the exact
location and cause of an error.

   The prompt of nested REPLs includes the nesting level; `1>' is the
prompt at the first nesting level, `2>' at the second nesting level,
and so on.  An end of file (usually <^D>) will cause the current REPL
to be terminated and the enclosing REPL (one nesting level less) to be
resumed.

   At any time the user can examine the frames in the REPL's
continuation, which is useful to determine which chain of procedure
calls lead to an error.  A backtrace that lists the chain of active
continuation frames in the REPL's continuation can be obtained with the
`,b' command.  The frames are numbered from 0, that is frame 0 is the
most recent frame of the continuation where execution stopped, frame 1
is the parent frame of frame 0, and so on.  It is also possible to move
the REPL to a specific parent continuation (i.e. a specific frame of
the continuation where execution stopped) with the `,N', `,N+', `,N-',
`,+', `,-', `,++', and `,--' commands.  When the frame number of the
frame being examined is not zero, it is shown in the prompt after the
nesting level, for example `1\5>' is the prompt when the REPL nesting
level is 1 and the frame number is 5.

   Expressions entered at a nested REPL are evaluated in the environment
(both lexical and dynamic) of the continuation frame currently being
examined if that frame was created by interpreted Scheme code.  If the
frame was created by compiled Scheme code then expressions get evaluated
in the global interaction environment.  This feature may be used in
interpreted code to fetch the value of a variable in the current frame
or to change its value with `set!'.  Note that some special forms
(`define' in particular) can only be evaluated in the global
interaction environment.

5.2 Debugging commands
======================

In addition to expressions, the REPL accepts the following special
"comma" commands:

`,? and ,help'
     Give a summary of the REPL commands.

`,(h SUBJECT)'
     This command will show the section of the Gambit manual with the
     definition of the procedure or special form SUBJECT, which must be
     a symbol.  For example `,(h time)' will show the section
     documenting the `time' special form.  Please see the `help'
     procedure for additional information.

`,h'
     This command will show the section of the Gambit manual with the
     definition of the procedure which raised the exception for which
     this REPL was started.

`,q'
     Terminate the process with exit status 0.  This is equivalent to
     calling `(exit 0)'.

`,qt'
     Terminate the current thread (note that terminating the primordial
     thread terminates the process).

`,t'
     Return to the outermost REPL, also known as the "top-level REPL".

`,d'
     Leave the current REPL and resume the enclosing REPL.  This
     command does nothing in the top-level REPL.

`,(c EXPR)'
     Leave the current REPL and continue the computation that initiated
     the REPL with a specific value.  This command can only be used to
     continue a computation that signaled an error.  The expression
     EXPR is evaluated in the current context and the resulting value
     is returned as the value of the expression which signaled the
     error.  For example, if the evaluation of the expression `(* (/ x
     y) 2)' signaled an error because `y' is zero, then in the nested
     REPL a `,(c (+ 4 y))' will resume the computation of `(* (/ x y)
     2)' as though the value of `(/ x y)' was 4.  This command must be
     used carefully because the context where the error occured may
     rely on the result being of a particular type.  For instance a
     `,(c #f)' in the previous example will cause `*' to signal a type
     error (this problem is the most troublesome when debugging Scheme
     code that was compiled with type checking turned off so be
     careful).

`,c'
     Leave the current REPL and continue the computation that initiated
     the REPL.  This command can only be used to continue a computation
     that was stopped due to a user interrupt, breakpoint or a
     single-step.

`,s'
     Leave the current REPL and continue the computation that initiated
     the REPL in single-stepping mode.  The computation will perform an
     evaluation step (as defined by `step-level-set!') and then stop,
     causing a nested REPL to be entered.  Just before the evaluation
     step is performed, a line is displayed (in the same format as
     `trace') which indicates the expression that is being evaluated.
     If the evaluation step produces a result, the result is also
     displayed on another line.  A nested REPL is then entered after
     displaying a message which describes the next step of the
     computation.  This command can only be used to continue a
     computation that was stopped due to a user interrupt, breakpoint
     or a single-step.

`,l'
     This command is similar to `,s' except that it "leaps" over
     procedure calls, that is procedure calls are treated like a single
     step.  Single-stepping mode will resume when the procedure call
     returns, or if and when the execution of the called procedure
     encounters a breakpoint.

`,N'
     Move to frame number N of the continuation.  After changing the
     current frame, a one-line summary of the frame is displayed as if
     the `,y' command was entered.

`,N+'
     Move forward by N frames in the chain of continuation frames (i.e.
     towards older continuation frames).  After changing the current
     frame, a one-line summary of the frame is displayed as if the `,y'
     command was entered.

`,N-'
     Move backward by N frames in the chain of continuation frames
     (i.e.  towards more recent continuation frames).  After changing
     the current frame, a one-line summary of the frame is displayed as
     if the `,y' command was entered.

`,+'
     Equivalent to `,1+'.

`,-'
     Equivalent to `,1-'.

`,++'
     Equivalent to `,N+' where N is the number of continuation frames
     displayed at the head of a backtrace.

`,--'
     Equivalent to `,N-' where N is the number of continuation frames
     displayed at the head of a backtrace.

`,y'
     Display a one-line summary of the current frame.  The information
     is displayed in four fields.  The first field is the frame number.
     The second field is the procedure that created the frame or
     `(interaction)' if the frame was created by an expression entered
     at the REPL.  The remaining fields describe the subproblem
     associated with the frame, that is the expression whose value is
     being computed.  The third field is the location of the
     subproblem's source code and the fourth field is a reproduction of
     the source code, possibly truncated to fit on the line.  The last
     two fields may be missing if that information is not available.
     In particular, the third field is missing when the frame was
     created by a user call to the `eval' procedure or by a compiled
     procedure not compiled with the declaration `debug-location', and
     the last field is missing when the frame was created by a compiled
     procedure not compiled with the declaration `debug-source'.

`,b'
     Display a backtrace summarizing each frame in the chain of
     continuation frames starting with the current frame.  For each
     frame, the same information as for the `,y' command is displayed
     (except that location information is displayed in the format
     `STREAM@LINE:COLUMN').  If there are more than 15 frames in the
     chain of continuation frames, some of the middle frames will be
     omitted.

`,be'
     Like the `,b' command but also display the environment.

`,bed'
     Like the `,be' command but also display the dynamic environment.

`,(b EXPR)'
     Display the backtrace of EXPR's value, X, which is obtained by
     evaluating EXPR in the current frame.  X must be a continuation or
     a thread.  When X is a continuation, the frames in that
     continuation are displayed.  When X is a thread, the backtrace of
     the current continuation of that thread is displayed.

`,(be EXPR)'
     Like the `,(b EXPR)' command but also display the environment.

`,(bed EXPR)'
     Like the `,(be EXPR)' command but also display the dynamic
     environment.

`,i'
     Pretty print the procedure that created the current frame or
     `(interaction)' if the frame was created by an expression entered
     at the REPL.  Compiled procedures will only be pretty printed when
     they are compiled with the declaration `debug-source'.

`,e'
     Display the environment which is accessible from the current frame.
     The lexical environment is displayed, followed by the dynamic
     environment if the parameter object
     `repl-display-dynamic-environment?' is not false.  Global lexical
     variables are not displayed.  Moreover the frame must have been
     created by interpreted code or code compiled with the declaration
     `debug-environments'.  Due to space safety considerations and
     compiler optimizations, some of the lexical variable bindings may
     be missing.  Lexical variable bindings are displayed using the
     format `VARIABLE = EXPRESSION' (when VARIABLE is mutable) or
     `VARIABLE == EXPRESSION' (when VARIABLE is immutable, which may
     happen in compiled code due to compiler optimization) and
     dynamically-bound parameter bindings are displayed using the
     format `(PARAMETER) = EXPRESSION'.  Note that EXPRESSION can be a
     self-evaluating expression (number, string, boolean, character,
     ...), a quoted expression, a lambda expression or a global
     variable (the last two cases, which are only used when the value
     of the variable or parameter is a procedure, simplifies the
     debugging of higher-order procedures).  A PARAMETER can be a
     quoted expression or a global variable.  Lexical bindings are
     displayed in inverse binding order (most deeply nested first) and
     shadowed variables are included in the list.

`,ed'
     Like the `,e' command but the dynamic environment is always
     displayed.

`,(e EXPR)'
     Display the environment of EXPR's value, X, which is obtained by
     evaluating EXPR in the current frame.  X must be a continuation, a
     thread, a procedure, or a nonnegative integer.  When X is a
     continuation, the environment at that point in the code is
     displayed.  When X is a thread, the environment of the current
     continuation of that thread is displayed. When X is a procedure,
     the lexical environment where X was created is combined with the
     current continuation and this combined environment is displayed.
     When X is an integer, the environment at frame number X of the
     continuation is displayed.

`,(ed EXPR)'
     Like the `,(e EXPR)' command but the dynamic environment is always
     displayed.

`,st'
     Display the state of the threads in the current thread's thread
     group.  A thread can be: uninitialized, initialized, active, and
     terminated (normally or abnormally).  Active threads can be
     running, sleeping and waiting on a synchronization object (mutex,
     condition variable or port) possibly with a timeout.

`,(st EXPR)'
     Display the state of a specific thread or thread group.  The value
     of EXPR must be a thread or thread group.

`,(v EXPR)'
     Start a new REPL visiting EXPR's value, X, which is obtained by
     evaluating EXPR in the current frame.  X must be a continuation, a
     thread, a procedure, or a nonnegative integer.  When X is a
     continuation, the new REPL's continuation is X and evaluations are
     done in the environment at that point in the code.  When X is a
     thread, the thread is interrupted and the new REPL's continuation
     is the point where the thread was interrupted.  When X is a
     procedure, the lexical environment where X was created is combined
     with the current continuation and evaluations are done in this
     combined environment.  When X is an integer, the REPL is started
     in frame number X of the continuation.


5.3 Debugging example
=====================

Here is a sample interaction with `gsi':

     $ gsi
     Gambit v4.9.4

     > (define (invsqr x) (/ 1 (expt x 2)))
     > (define (mymap fn lst)
         (define (mm in)
           (if (null? in)
               '()
               (cons (fn (car in)) (mm (cdr in)))))
         (mm lst))
     > (mymap invsqr '(5 2 hello 9 1))
     *** ERROR IN invsqr, (console)@1.25 -- (Argument 1) NUMBER expected
     (expt 'hello 2)
     1> ,i
     #<procedure #2 invsqr> =
     (lambda (x) (/ 1 (expt x 2)))
     1> ,e
     x = 'hello
     1> ,b
     0  invsqr                    (console)@1:25          (expt x 2)
     1  #<procedure #4>           (console)@6:17          (fn (car in))
     2  #<procedure #4>           (console)@6:31          (mm (cdr in))
     3  #<procedure #4>           (console)@6:31          (mm (cdr in))
     4  (interaction)             (console)@8:1           (mymap invsqr '(5 2 hel...
     1> ,+
     1  #<procedure #4>           (console)@6.17          (fn (car in))
     1\1> (pp #4)
     (lambda (in) (if (null? in) '() (cons (fn (car in)) (mm (cdr in)))))
     1\1> ,e
     in = '(hello 9 1)
     mm = (lambda (in) (if (null? in) '() (cons (fn (car in)) (mm (cdr in)))))
     fn = invsqr
     lst = '(5 2 hello 9 1)
     1\1> ,(e mm)
     mm = (lambda (in) (if (null? in) '() (cons (fn (car in)) (mm (cdr in)))))
     fn = invsqr
     lst = '(5 2 hello 9 1)
     1\1> fn
     #<procedure #2 invsqr>
     1\1> (pp fn)
     (lambda (x) (/ 1 (expt x 2)))
     1\1> ,+
     2  #<procedure #4>           (console)@6.31          (mm (cdr in))
     1\2> ,e
     in = '(2 hello 9 1)
     mm = (lambda (in) (if (null? in) '() (cons (fn (car in)) (mm (cdr in)))))
     fn = invsqr
     lst = '(5 2 hello 9 1)
     1\2> ,(c (list 3 4 5))
     (1/25 1/4 3 4 5)
     > ,q

5.4 Procedures related to debugging
===================================

 -- procedure: help [SUBJECT]
 -- procedure: help-browser [NEW-VALUE]
     The `help' procedure displays the section of the Gambit manual
     with the definition of the procedure or special form SUBJECT,
     which must be a procedure or symbol.  For example the call `(help
     gensym)' will show the section documenting the `gensym' procedure
     and the call `(help 'time)' will show the section documenting the
     `time' special form.  When the SUBJECT is absent, the
     documentation of the `help' procedure is shown. The `help'
     procedure returns the void object.

     The parameter object `help-browser' is bound to a string naming
     the external program that is used by the `help' procedure to view
     the documentation.  Initially it is bound to the empty string.  In
     normal circumstances when `help-browser' is bound to an empty
     string the `help' procedure runs the script `~~bin/gambdoc.bat'
     which searches for a suitable web browser to open the
     documentation in HTML format.  Unless the system was built with
     the command `configure --enable-help-browser=...', the text-only
     browser `lynx' (see `http://lynx.isc.org/') will be used by
     default if it is available.  We highly recommend that you install
     this browser if you are interested in viewing the documentation
     within the console in which the REPL is running.  You can exit
     `lynx' conveniently by typing an end of file (usually <^D>).

     For example:

          > (help-browser "firefox") ; use firefox instead of lynx
          > (help 'gensym)
          > (help gensym) ; OK because gensym is a procedure
          > (help 'time)
          > (help time) ; not OK because time is a special form
          *** ERROR IN (console)@5.7 -- Macro name can't be used as a variable: time
          >


 -- procedure: apropos [SUBSTRING [PORT]]
     The `apropos' procedure writes to the port PORT a report of all
     the global variables whose name contains SUBSTRING, a string or
     symbol.  If SUBSTRING is not specified the report contains all the
     global variables.  If it is not specified, PORT defaults to the
     interaction channel (i.e. the output will appear at the REPL).
     The `apropos' procedure returns the void object.

     The global variables are grouped into namespaces.  The empty
     namespace, if it is relevant, is last.  This reduces the
     likelihood it will scroll off the screen if there are several
     global variables in other namespaces, which are typically less
     interesting.

     Note that with the `apropos' procedure it is possible to reveal
     the existence of procedures of the runtime system and modules that
     are not intended to be called by user code.  These procedures
     often avoid type checking their arguments or must be called in a
     specific context, so calling them incorrectly may crash the
     system.  On the other hand it also allows discovering the
     existence of certain functionalities that may have gone unnoticed.

     For example:

          > (apropos "cons")
          "##" namespace:
            10^-constants, cons, cons*, cons*-aux, console-port,
            constant-expression-value, constant-expression?,
            cprc-quasi-cons, deconstruct-call,
            define-type-construct-constant, degen-quasi-cons,
            gen-quasi-cons, quasi-cons, stdio/console-repl-channel,
            void-constant?, xcons
          empty namespace:
            cons, cons*, console-port, xcons
          > (import (srfi 69))
          > (apropos "table?")
          "##" namespace:
            gc-hash-table?, mutable?, readtable?, table?
          "srfi/69#" namespace:
            hash-table?
          empty namespace:
            readtable?, table?
          > (apropos "srfi/69#")
          "srfi/69#" namespace:
            ||, alist->hash-table, hash, hash-by-identity,
            hash-table->alist, hash-table-copy, hash-table-delete!,
            hash-table-equivalence-function, hash-table-exists?,
            hash-table-fold, hash-table-hash-function,
            hash-table-keys, hash-table-merge!, hash-table-ref,
            hash-table-ref/default, hash-table-set!, hash-table-size,
            hash-table-update!, hash-table-update!/default,
            hash-table-values, hash-table-walk, hash-table?,
            make-hash-table, string-ci-hash, string-hash


 -- procedure: repl-result-history-ref I
 -- procedure: repl-result-history-max-length-set! N
     The REPL keeps a history of the last few results printed by the
     REPL. The call `(repl-result-history-ref I)' returns the Ith
     previous result (the last for I=0, the next to last for I=1, etc).
     By default the REPL result history remembers up to 3 results.
     The maximal length of the history can be set to N between 0 and 10
     by a call to `(repl-result-history-max-length-set! N)'.

     For convenience the reader defines an abbreviation for calling
     `repl-result-history-ref'.  Tokens formed by a sequence of one or
     more hash signs, such as ``#'', ``##'', etc, are expanded by the
     reader into the list `(repl-result-history-ref I)', where I is the
     number of hash signs minus 1.  In other words, ``#'' will return
     the last result printed by the REPL, ``##'' will return the next
     to last, etc.

     For example:

          > (map (lambda (x) (* x x)) '(1 2 3))
          (1 4 9)
          > (reverse #)
          (9 4 1)
          > (append # ##)
          (9 4 1 1 4 9)
          > 1
          1
          > 1
          1
          > (+ # ##)
          2
          > (+ # ##)
          3
          > (+ # ##)
          5
          > ####
          *** ERROR IN (console)@9.1 -- (Argument 1) Out of range
          (repl-result-history-ref 3)
          1>


 -- procedure: trace PROC...
 -- procedure: untrace PROC...
     The `trace' procedure starts tracing calls to the specified
     procedures.  When a traced procedure is called, a line containing
     the procedure and its arguments is displayed (using the procedure
     call expression syntax).  The line is indented with a sequence of
     vertical bars which indicate the nesting depth of the procedure's
     continuation.  After the vertical bars is a greater-than sign
     which indicates that the evaluation of the call is starting.

     When a traced procedure returns a result, it is displayed with the
     same indentation as the call but without the greater-than sign.
     This makes it easy to match calls and results (the result of a
     given call is the value at the same indentation as the
     greater-than sign).  If a traced procedure P1 performs a tail call
     to a traced procedure P2, then P2 will use the same indentation as
     P1.  This makes it easy to spot tail calls.  The special handling
     for tail calls is needed to preserve the space complexity of the
     program (i.e. tail calls are implemented as required by Scheme
     even when they involve traced procedures).

     The `untrace' procedure stops tracing calls to the specified
     procedures.  When no argument is passed to the `trace' procedure,
     the list of procedures currently being traced is returned.  The
     void object is returned by the `trace' procedure when it is passed
     one or more arguments.  When no argument is passed to the
     `untrace' procedure stops all tracing and returns the void object.
     A compiled procedure may be traced but only if it is bound to a
     global variable.

     For example:

          > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1)))))
          > (trace fact)
          > (fact 5)
          | > (fact 5)
          | | > (fact 4)
          | | | > (fact 3)
          | | | | > (fact 2)
          | | | | | > (fact 1)
          | | | | | 1
          | | | | 2
          | | | 6
          | | 24
          | 120
          120
          > (trace -)
          *** WARNING -- Rebinding global variable "-" to an interpreted procedure
          > (define (fact-iter n r) (if (< n 2) r (fact-iter (- n 1) (* n r))))
          > (trace fact-iter)
          > (fact-iter 5 1)
          | > (fact-iter 5 1)
          | | > (- 5 1)
          | | 4
          | > (fact-iter 4 5)
          | | > (- 4 1)
          | | 3
          | > (fact-iter 3 20)
          | | > (- 3 1)
          | | 2
          | > (fact-iter 2 60)
          | | > (- 2 1)
          | | 1
          | > (fact-iter 1 120)
          | 120
          120
          > (trace)
          (#<procedure #2 fact-iter> #<procedure #3 -> #<procedure #4 fact>)
          > (untrace)
          > (fact 5)
          120


 -- procedure: step
 -- procedure: step-level-set! LEVEL
     The `step' procedure enables single-stepping mode.  After the call
     to `step' the computation will stop just before the interpreter
     executes the next evaluation step (as defined by
     `step-level-set!').  A nested REPL is then started.  Note that
     because single-stepping is stopped by the REPL whenever the prompt
     is displayed it is pointless to enter `(step)' by itself.  On the
     other hand entering `(begin (step) EXPR)' will evaluate EXPR in
     single-stepping mode.

     The procedure `step-level-set!' sets the stepping level which
     determines the granularity of the evaluation steps when
     single-stepping is enabled.  The stepping level LEVEL must be an
     exact integer in the range 0 to 7.  At a level of 0, the
     interpreter ignores single-stepping mode.  At higher levels the
     interpreter stops the computation just before it performs the
     following operations, depending on the stepping level:

       1. procedure call

       2. `delay' special form and operations at lower levels

       3. `lambda' special form and operations at lower levels

       4. `define' special form and operations at lower levels

       5. `set!' special form and operations at lower levels

       6. variable reference and operations at lower levels

       7. constant reference and operations at lower levels


     The default stepping level is 7.

     For example:

          > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1)))))
          > (step-level-set! 1)
          > (begin (step) (fact 5))
          *** STOPPED IN (console)@3.15
          1> ,s
          | > (fact 5)
          *** STOPPED IN fact, (console)@1.22
          1> ,s
          | | > (< n 2)
          | | #f
          *** STOPPED IN fact, (console)@1.43
          1> ,s
          | | > (- n 1)
          | | 4
          *** STOPPED IN fact, (console)@1.37
          1> ,s
          | | > (fact (- n 1))
          *** STOPPED IN fact, (console)@1.22
          1> ,s
          | | | > (< n 2)
          | | | #f
          *** STOPPED IN fact, (console)@1.43
          1> ,s
          | | | > (- n 1)
          | | | 3
          *** STOPPED IN fact, (console)@1.37
          1> ,l
          | | | > (fact (- n 1))
          *** STOPPED IN fact, (console)@1.22
          1> ,l
          | | > (* n (fact (- n 1)))
          | | 24
          *** STOPPED IN fact, (console)@1.32
          1> ,l
          | > (* n (fact (- n 1)))
          | 120
          120


 -- procedure: break PROC...
 -- procedure: unbreak PROC...
     The `break' procedure places a breakpoint on each of the specified
     procedures.  When a procedure is called that has a breakpoint, the
     interpreter will enable single-stepping mode (as if `step' had
     been called).  This typically causes the computation to stop soon
     inside the procedure if the stepping level is high enough.

     The `unbreak' procedure removes the breakpoints on the specified
     procedures.  With no argument, `break' returns the list of
     procedures currently containing breakpoints.  The void object is
     returned by `break' if it is passed one or more arguments.  With
     no argument `unbreak' removes all the breakpoints and returns the
     void object.  A breakpoint can be placed on a compiled procedure
     but only if it is bound to a global variable.

     For example:

          > (define (double x) (+ x x))
          > (define (triple y) (- (double (double y)) y))
          > (define (f z) (* (triple z) 10))
          > (break double)
          > (break -)
          *** WARNING -- Rebinding global variable "-" to an interpreted procedure
          > (f 5)
          *** STOPPED IN double, (console)@1.21
          1> ,b
          0  double                    (console)@1:21          +
          1  triple                    (console)@2:31          (double y)
          2  f                         (console)@3:18          (triple z)
          3  (interaction)             (console)@6:1           (f 5)
          1> ,e
          x = 5
          1> ,c
          *** STOPPED IN double, (console)@1.21
          1> ,c
          *** STOPPED IN f, (console)@3.29
          1> ,c
          150
          > (break)
          (#<procedure #3 -> #<procedure #4 double>)
          > (unbreak)
          > (f 5)
          150


 -- procedure: generate-proper-tail-calls [NEW-VALUE]
     [Note: this procedure is DEPRECATED and will be removed in a
     future version of Gambit.  Use the `proper-tail-calls' declaration
     instead.]

     The parameter object `generate-proper-tail-calls' is bound to a
     boolean value controlling how the interpreter handles tail calls.
     When it is bound to `#f' the interpreter will treat tail calls
     like nontail calls, that is a new continuation will be created for
     the call.  This setting is useful for debugging, because when a
     primitive signals an error the location information will point to
     the call site of the primitive even if this primitive was called
     with a tail call.  The initial value of this parameter object is
     `#t', which means that a tail call will reuse the continuation of
     the calling function.

     This parameter object only affects code that is subsequently
     processed by `load' or `eval', or entered at the REPL.

     For example:

          > (generate-proper-tail-calls)
          #t
          > (let loop ((i 1)) (if (< i 10) (loop (* i 2)) oops))
          *** ERROR IN #<procedure #2>, (console)@2.47 -- Unbound variable: oops
          1> ,b
          0  #<procedure #2>           (console)@2:47          oops
          1  (interaction)             (console)@2:1           ((letrec ((loop (lambda...
          1> ,t
          > (generate-proper-tail-calls #f)
          > (let loop ((i 1)) (if (< i 10) (loop (* i 2)) oops))
          *** ERROR IN #<procedure #3>, (console)@6.47 -- Unbound variable: oops
          1> ,b
          0  #<procedure #3>           (console)@6:47          oops
          1  #<procedure #3>           (console)@6:32          (loop (* i 2))
          2  #<procedure #3>           (console)@6:32          (loop (* i 2))
          3  #<procedure #3>           (console)@6:32          (loop (* i 2))
          4  #<procedure #3>           (console)@6:32          (loop (* i 2))
          5  (interaction)             (console)@6:1           ((letrec ((loop (lambda...


 -- procedure: display-environment-set! DISPLAY?
     [Note: this procedure is DEPRECATED and will be removed in a
     future version of Gambit.  Use the parameter object
     `repl-display-environment?' instead.]

     This procedure sets a flag that controls the automatic display of
     the environment by the REPL.  If DISPLAY? is true, the environment
     is displayed by the REPL before the prompt.  The default setting is
     not to display the environment.


 -- procedure: repl-display-environment? DISPLAY?
     The parameter object `repl-display-environment?' is bound to a
     boolean value that controls the automatic display of the
     environment by the REPL.  If DISPLAY? is true, the environment is
     displayed by the REPL before the prompt.  This is particularly
     useful in single-stepping mode.  The default setting is not to
     display the environment.


 -- procedure: display-dynamic-environment? DISPLAY?
     The parameter object `display-dynamic-environment?' is bound to a
     boolean value that controls wether the dynamic environment is
     displayed when the environment is displayed.  The default setting
     is not to display the dynamic environment.


 -- procedure: pretty-print OBJ [PORT]
     This procedure pretty-prints OBJ on the port PORT.  If it is not
     specified, PORT defaults to the current output-port.

     For example:

          > (pretty-print
              (let* ((x '(1 2 3 4)) (y (list x x x))) (list y y y)))
          (((1 2 3 4) (1 2 3 4) (1 2 3 4))
           ((1 2 3 4) (1 2 3 4) (1 2 3 4))
           ((1 2 3 4) (1 2 3 4) (1 2 3 4)))


 -- procedure: pp OBJ [PORT]
     This procedure pretty-prints OBJ on the port PORT.  When OBJ is a
     procedure created by the interpreter or a procedure created by
     code compiled with the declaration `debug-source', the procedure's
     source code is displayed.  If it is not specified, PORT defaults
     to the interaction channel (i.e. the output will appear at the
     REPL).

     For example:

          > (define (f g) (+ (time (g 100)) (time (g 1000))))
          > (pp f)
          (lambda (g)
            (+ (##time (lambda () (g 100)) '(g 100))
               (##time (lambda () (g 1000)) '(g 1000))))


 -- procedure: gc-report-set! REPORT?
     This procedure controls the generation of reports during garbage
     collections.  If the argument is true, a brief report of memory
     usage is generated after every garbage collection.  It contains:
     the time taken for this garbage collection, the amount of memory
     allocated in megabytes since the program was started, the size of
     the heap in megabytes, the heap memory in megabytes occupied by
     live data, the proportion of the heap occupied by live data, and
     the number of bytes occupied by movable and nonmovable objects.


5.5 Console line-editing
========================

The console implements a simple Scheme-friendly line-editing
user-interface that is enabled by default.  It offers parentheses
balancing, a history of previous commands, symbol completion, and
several emacs-compatible keyboard commands.  The user's input is
displayed in a bold font and the output produced by the system is in a
plain font.  The history of previous commands is saved in the file
`~/.gambit_history'.  It is restored when a REPL is started.

   Symbol completion is triggered with the tab key.  When the cursor is
after a sequence of characters that can form a symbol, typing the tab
key will search the symbol table for the first symbol (in alphabetical
order) that begins with that sequence and insert that symbol.  Typing
the tab key in succession will cycle through all symbols with that
prefix.  When all possible symbols have been shown or there are no
possible completions, the text reverts to the uncompleted symbol and
the bell is rung.

   Here are the keyboard commands available (where the ``M-'' prefix
means the escape key is typed and the ``C-'' prefix means the control
key is pressed):

`C-d'
     Generate an end-of-file when the line is empty, otherwise delete
     character at cursor.

`delete or backspace'
     Delete character before cursor.

`M-C-d'
     Delete word forward and keep a copy of this text on the clipboard.

`M-delete'
     Delete word backward and keep a copy of this text on the clipboard.

`M-backspace'
     Delete S-expression backward and keep a copy of this text on the
     clipboard.

`C-a'
     Move cursor to beginning of line.

`C-e'
     Move cursor to end of line.

`C-b or left-arrow'
     Move cursor left one character.

`M-b'
     Move cursor left one word.

`M-C-b or `M-'left-arrow'
     Move cursor left one S-expression.

`C-f or right-arrow'
     Move cursor right one character.

`M-f'
     Move cursor right one word.

`M-C-f or `M-'right-arrow'
     Move cursor right one S-expression.

`C-p or `M-p' or up-arrow'
     Move to previous line in history.

`C-n or `M-n' or down-arrow'
     Move to next line in history.

`C-t'
     Transpose character at cursor with previous character.

`M-t'
     Transpose word after cursor with previous word.

`M-C-t'
     Transpose S-expression after cursor with previous S-expression.

`C-l'
     Clear console and redraw line being edited.

`C-nul'
     Set the mark to the cursor.

`C-w'
     Delete the text between the cursor and the mark and keep a copy of
     this text on the clipboard.

`C-k'
     Delete the text from the cursor to the end of the line and keep a
     copy of this text on the clipboard.

`C-y'
     Paste the text that is on the clipboard.

`F8'
     Same as typing `#||#,c;' (REPL command to continue the
     computation).

`F9'
     Same as typing `#||#,-;' (REPL command to move to newer frame).

`F10'
     Same as typing `#||#,+;' (REPL command to move to older frame).

`F11'
     Same as typing `#||#,s;' (REPL command to step the computation).

`F12'
     Same as typing `#||#,l;' (REPL command to leap the computation).


   On macOS, depending on your configuration, you may have to press the
`fn' key to access the function key `F12' and the `option' key to
access the other function keys.

   On Microsoft Windows the clipboard is the system clipboard.  This
allows text to be copied and pasted between the program and other
applications.  On other operating systems the clipboard is internal to
the program (it is not integrated with the operating system).

5.6 Emacs interface
===================

Gambit comes with the Emacs package `gambit.el' which provides a nice
environment for running Gambit from within the Emacs editor.  This
package filters the standard output of the Gambit process and when it
intercepts a location information (in the format `STREAM@LINE.COLUMN'
where STREAM is either `(stdin)' when the expression was obtained from
standard input, `(console)' when the expression was obtained from the
console, or a string naming a file) it opens a window to highlight the
corresponding expression.

   To use this package, make sure the file `gambit.el' is accessible
from your load-path and that the following lines are in your `.emacs'
file:

     (autoload 'gambit-inferior-mode "gambit" "Hook Gambit mode into cmuscheme.")
     (autoload 'gambit-mode "gambit" "Hook Gambit mode into scheme.")
     (add-hook 'inferior-scheme-mode-hook (function gambit-inferior-mode))
     (add-hook 'scheme-mode-hook (function gambit-mode))
     (setq scheme-program-name "gsi -:debug=-")

   Alternatively, if you don't mind always loading this package, you
can simply add this line to your `.emacs' file:

     (require 'gambit)

   You can then start an inferior Gambit process by typing `M-x
run-scheme'.  The commands provided in `cmuscheme' mode will be
available in the Gambit interaction buffer (i.e. `*scheme*') and in
buffers attached to Scheme source files.  Here is a list of the most
useful commands (for a complete list type `C-h m' in the Gambit
interaction buffer):
`C-x C-e'
     Evaluate the expression which is before the cursor (the expression
     will be copied to the Gambit interaction buffer).

`C-c C-z'
     Switch to Gambit interaction buffer.

`C-c C-l'
     Load a file (file attached to current buffer is default) using
     `(load FILE)'.

`C-c C-k'
     Compile a file (file attached to current buffer is default) using
     `(compile-file FILE)'.

   The file `gambit.el' provides these additional commands:

`F8 or C-c c'
     Continue the computation (same as typing `#||#,c;' to the REPL).

`F9 or C-c ]'
     Move to newer frame (same as typing `#||#,-;' to the REPL).

`F10 or C-c ['
     Move to older frame (same as typing `#||#,+;' to the REPL).

`F11 or C-c s'
     Step the computation (same as typing `#||#,s;' to the REPL).

`F12 or C-c l'
     Leap the computation (same as typing `#||#,l;' to the REPL).

`C-c _'
     Removes the last window that was opened to highlight an expression.

   The two keystroke version of these commands can be shortened to
`M-c', `M-[', `M-]', `M-s', `M-l', and `M-_' respectively by adding
this line to your `.emacs' file:

     (setq gambit-repl-command-prefix "\e")

   This is more convenient to type than the two keystroke `C-c' based
sequences but the purist may not like this because it does not follow
normal Emacs conventions.

   Here is what a typical `.emacs' file will look like:

     (setq load-path ; add directory containing gambit.el
       (cons "/usr/local/Gambit/share/emacs/site-lisp"
             load-path))
     (setq scheme-program-name "/tmp/gsi -:debug=-") ; if gsi not in executable path
     (setq gambit-highlight-color "gray") ; if you don't like the default
     (setq gambit-repl-command-prefix "\e") ; if you want M-c, M-s, etc
     (require 'gambit)

5.7 GUIDE
=========

The implementation and documentation for GUIDE, the Gambit Universal
IDE, are not yet complete.

6 Scheme extensions
*******************

6.1 Extensions to standard procedures
=====================================

 -- procedure: transcript-on FILE
 -- procedure: transcript-off
     These procedures do nothing.


 -- procedure: call-with-current-continuation PROC
 -- procedure: call/cc PROC
     The procedure `call-with-current-continuation' is bound to the
     global variables `call-with-current-continuation' and `call/cc'.


6.2 Extensions to standard special forms
========================================

 -- special form: lambda lambda-formals body
 -- special form: define (variable define-formals) body
          lambda-formals = `(' formal-argument-list `)' |
          r4rs-lambda-formals

          define-formals = formal-argument-list | r4rs-define-formals

          formal-argument-list = dsssl-formal-argument-list |
          rest-at-end-formal-argument-list

          dsssl-formal-argument-list = reqs opts rest keys

          rest-at-end-formal-argument-list = reqs opts keys rest | reqs
          opts keys `.' rest-formal-argument

          reqs = required-formal-argument*

          required-formal-argument = variable

          opts = `#!optional' optional-formal-argument* | empty

          optional-formal-argument = variable | `(' variable
          initializer `)'

          rest = `#!rest' rest-formal-argument | empty

          rest-formal-argument = variable

          keys = `#!key' keyword-formal-argument* | empty

          keyword-formal-argument = variable | `(' variable initializer
          `)'

          initializer = expression

          r4rs-lambda-formals = `(' variable* `)' | `(' variable+ `.'
          variable `)' | variable

          r4rs-define-formals = variable* | variable* `.' variable

     These forms are extended versions of the `lambda' and `define'
     special forms of standard Scheme.  They allow the use of optional
     formal arguments, either positional or named, and support the
     syntax and semantics of the DSSSL standard.

     When the procedure introduced by a `lambda' (or `define') is
     applied to a list of actual arguments, the formal and actual
     arguments are processed as specified in the R4RS if the
     lambda-formals (or define-formals) is a r4rs-lambda-formals (or
     r4rs-define-formals).

     If the formal-argument-list matches dsssl-formal-argument-list or
     extended-formal-argument-list they are processed as follows:

       a. Variables in required-formal-arguments are bound to
          successive actual arguments starting with the first actual
          argument.  It shall be an error if there are fewer actual
          arguments than required-formal-arguments.

       b. Next variables in optional-formal-arguments are bound to
          remaining actual arguments.  If there are fewer remaining
          actual arguments than optional-formal-arguments, then the
          variables are bound to the result of evaluating initializer,
          if one was specified, and otherwise to `#f'.  The initializer
          is evaluated in an environment in which all previous formal
          arguments have been bound.

       c. If `#!key' does not appear in the formal-argument-list and
          there is no rest-formal-argument then it shall be an error if
          there are any remaining actual arguments.

       d. If `#!key' does not appear in the formal-argument-list and
          there is a rest-formal-argument then the rest-formal-argument
          is bound to a list of all remaining actual arguments.

       e. If `#!key' appears in the formal-argument-list and there is
          no rest-formal-argument then there shall be an even number of
          remaining actual arguments.  These are interpreted as a
          series of pairs, where the first member of each pair is a
          keyword specifying the argument name, and the second is the
          corresponding value.  It shall be an error if the first
          member of a pair is not a keyword.  It shall be an error if
          the argument name is not the same as a variable in a
          keyword-formal-argument.  If the same argument name occurs
          more than once in the list of actual arguments, then the
          first value is used.  If there is no actual argument for a
          particular keyword-formal-argument, then the variable is
          bound to the result of evaluating initializer if one was
          specified, and otherwise to `#f'.  The initializer is
          evaluated in an environment in which all previous formal
          arguments have been bound.

       f. If `#!key' appears in the formal-argument-list and there is a
          rest-formal-argument before the `#!key' then there may be an
          even or odd number of remaining actual arguments and the
          rest-formal-argument is bound to a list of all remaining
          actual arguments.  Then, these remaining actual arguments are
          scanned from left to right in pairs, stopping at the first
          pair whose first element is not a keyword.  Each pair whose
          first element is a keyword matching the name of a
          keyword-formal-argument gives the value (i.e. the second
          element of the pair) of the corresponding formal argument.  If
          the same argument name occurs more than once in the list of
          actual arguments, then the first value is used.  If there is
          no actual argument for a particular keyword-formal-argument,
          then the variable is bound to the result of evaluating
          initializer if one was specified, and otherwise to `#f'.  The
          initializer is evaluated in an environment in which all
          previous formal arguments have been bound.

       g. If `#!key' appears in the formal-argument-list and there is a
          rest-formal-argument after the `#!key' then there may be an
          even or odd number of remaining actual arguments.  The
          remaining actual arguments are scanned from left to right in
          pairs, stopping at the first pair whose first element is not
          a keyword.  Each pair shall have as its first element a
          keyword matching the name of a keyword-formal-argument; the
          second element gives the value of the corresponding formal
          argument.  If the same argument name occurs more than once in
          the list of actual arguments, then the first value is used.
          If there is no actual argument for a particular
          keyword-formal-argument, then the variable is bound to the
          result of evaluating initializer if one was specified, and
          otherwise to `#f'.  The initializer is evaluated in an
          environment in which all previous formal arguments have been
          bound.  Finally, the rest-formal-argument is bound to the
          list of the actual arguments that were not scanned (i.e.
          after the last keyword/value pair).

     In all cases it is an error for a variable to appear more than
     once in a formal-argument-list.

     Note that this specification is compatible with the DSSSL language
     standard (i.e. a correct DSSSL program will have the same semantics
     when run with Gambit).

     It is unspecified whether variables receive their value by binding
     or by assignment.  Currently the compiler and interpreter use
     different methods, which can lead to different semantics if
     `call-with-current-continuation' is used in an initializer.  Note
     that this is irrelevant for DSSSL programs because
     `call-with-current-continuation' does not exist in DSSSL.

     For example:

          > ((lambda (#!rest x) x) 1 2 3)
          (1 2 3)
          > (define (f a #!optional b) (list a b))
          > (define (g a #!optional (b a) #!key (k (* a b))) (list a b k))
          > (define (h1 a #!rest r #!key k) (list a k r))
          > (define (h2 a #!key k #!rest r) (list a k r))
          > (f 1)
          (1 #f)
          > (f 1 2)
          (1 2)
          > (g 3)
          (3 3 9)
          > (g 3 4)
          (3 4 12)
          > (g 3 4 k: 5)
          (3 4 5)
          > (g 3 4 k: 5 k: 6)
          (3 4 5)
          > (h1 7)
          (7 #f ())
          > (h1 7 k: 8 9)
          (7 8 (k: 8 9))
          > (h1 7 k: 8 z: 9)
          (7 8 (k: 8 z: 9))
          > (h2 7)
          (7 #f ())
          > (h2 7 k: 8 9)
          (7 8 (9))
          > (h2 7 k: 8 z: 9)
          *** ERROR IN (console)@17.1 -- Unknown keyword argument passed to procedure
          (h2 7 k: 8 z: 9)


6.3 Miscellaneous extensions
============================

 -- procedure: subvector VECTOR START END
     This procedure is the vector analog of the `substring' procedure.
     It returns a newly allocated vector formed from the elements of
     the vector VECTOR beginning with index START (inclusive) and
     ending with index END (exclusive).

     For example:

          > (subvector '#(a b c d e f) 3 5)
          #(d e)


 -- procedure: vector-copy VECTOR [START [END]]
     This procedure is like the procedure `subvector' except the
     parameter START defaults to 0 and the parameter END defaults to
     the length of the vector VECTOR.  Note that the elements are not
     recursively copied.

     For example:

          > (define v1 '#(a b c d e f))
          > (define v2 (vector-copy v1))
          > v2
          #(a b c d e f)
          > (eq? v1 v2)
          #f
          > (vector-copy v1 3)
          #(d e f)
          > (vector-copy v1 3 5)
          #(d e)


 -- procedure: vector-copy! DEST-VECTOR DEST-START VECTOR [START [END]]
     This procedure mutates the vector DEST-VECTOR.  It copies the
     elements of the vector VECTOR beginning with index START
     (inclusive) and ending with index END (exclusive) to the vector
     DEST-VECTOR at index DEST-START.  The parameters START and END
     default respectively to 0 and the length of the vector VECTOR.  It
     is an error to copy more elements than will fit in the tail of the
     vector DEST-VECTOR starting at index DEST-START.  Note that the
     elements are not recursively copied.

     For example:

          > (define v1 (vector 10 11 12 13 14 15))
          > (define v2 (vector 20 21 22 23))
          > (vector-copy! v1 1 v2)
          > v1
          #(10 20 21 22 23 15)
          > (vector-copy! v1 1 v2 3)
          > v1
          #(10 23 21 22 23 15)
          > (vector-copy! v1 1 v2 1 3)
          > v1
          #(10 21 22 22 23 15)


 -- procedure: vector-append VECTOR...
     This procedure is the vector analog of the `string-append'
     procedure.  It returns a newly allocated vector whose elements
     form the concatenation of the given vectors.

     For example:

          > (define v '#(1 2 3))
          > (vector-append v v v)
          #(1 2 3 1 2 3 1 2 3)


 -- procedure: vector-concatenate LST [SEPARATOR]
     This procedure returns a newly allocated vector whose elements form
     the concatenation of all the vectors in the list LST. If the
     optional vector SEPARATOR argument is specified, it will be added
     between all the elements of LST.  Without the SEPARATOR argument
     the result is the same as `(apply vector-append LST)'.

     For example:

          > (define v '#(1 2 3))
          > (vector-concatenate (list v v v))
          #(1 2 3 1 2 3 1 2 3)
          > (vector-concatenate (list v v v) '#(88 99))
          #(1 2 3 88 99 1 2 3 88 99 1 2 3)


 -- procedure: subvector-fill! VECTOR START END FILL
     This procedure is like `vector-fill!', but fills a selected part
     of the given vector. It sets the elements of the vector VECTOR,
     beginning with index START (inclusive) and ending with index END
     (exclusive) to FILL.  The value returned is unspecified.

     For example:

          > (define v (vector 'a 'b 'c 'd 'e 'f))
          > (subvector-fill! v 3 5 'x)
          > v
          #(a b c x x f)


 -- procedure: subvector-move! SRC-VECTOR SRC-START SRC-END DST-VECTOR
          DST-START
     This procedure replaces part of the contents of vector DST-VECTOR
     with part of the contents of vector SRC-VECTOR. It copies elements
     from SRC-VECTOR, beginning with index SRC-START (inclusive) and
     ending with index SRC-END (exclusive) to DST-VECTOR beginning with
     index DST-START (inclusive).  The value returned is unspecified.

     For example:

          > (define v1 '#(1 2 3 4 5 6))
          > (define v2 (vector 'a 'b 'c 'd 'e 'f))
          > (subvector-move! v1 3 5 v2 1)
          > v2
          #(a 4 5 d e f)


 -- procedure: vector-shrink! VECTOR K
     This procedure shortens the vector VECTOR so that its new size is
     K.  The value returned is unspecified.

     For example:

          > (define v (vector 'a 'b 'c 'd 'e 'f))
          > v
          #(a b c d e f)
          > (vector-shrink! v 3)
          > v
          #(a b c)


 -- procedure: vector-cas! VECTOR K NEW-VALUE OLD-VALUE
     The procedure `vector-cas!' performs an atomic compare-and-swap
     operation on the element of vector VECTOR at index K.  If the
     element's value is `eq?' to OLD-VALUE then the element is changed
     to NEW-VALUE, otherwise the value does not change.  Regardless
     what happened, the element's value prior to any change is
     returned.  It is thus possible to detect a change by an explicit
     `eq?'  test of the result.

     For example:

          > (define v (vector 'a))
          > (eq? 'foo (vector-cas! v 0 'b 'foo))
          #f
          > v
          #(a)
          > (eq? 'a (vector-cas! v 0 'b 'a))
          #t
          > v
          #(b)


 -- procedure: vector-inc! VECTOR K [STEP]
     The procedure `vector-inc!' performs an atomic incrementation on
     the element of vector VECTOR at index K, which must be a fixnum.
     The parameter STEP defaults to 1 and it is the fixnum value that
     is added (with wraparound) to the element.  The procedure returns
     the value of the element prior to the incrementation.

     For example:

          > (define v (vector 100))
          > (vector-inc! v 0)
          100
          > (vector-inc! v 0)
          101
          > (vector-inc! v 0 5)
          102
          > v
          #(107)


 -- procedure: vector-set VECTOR K OBJ
     The procedure `vector-set' returns a new copy of the vector VECTOR
     with the element at index K replaced with OBJ.

     For example:

          > (define v1 (vector 10 11 12 13))
          > (define v2 (vector-set v1 2 99))
          > v2
          #(10 11 99 13)
          > (eq? v1 v2)
          #f


 -- procedure: string-set STRING K CHAR
     The procedure `string-set' returns a new copy of the string STRING
     with the character at index K replaced with CHAR.

     For example:

          > (define s1 (string #\a #\b #\c #\d))
          > (define s2 (string-set s1 2 #\.))
          > s2
          "ab.d"
          > (eq? s1 s2)
          #f


 -- procedure: string-concatenate LST [SEPARATOR]
     This procedure returns a newly allocated string which is the
     concatenation of all the strings in the list LST. If the optional
     string SEPARATOR argument is specified, it will be added between
     all the elements of LST.  Without the SEPARATOR argument the
     result is the same as `(apply string-append LST)'.

     For example:

          > (define s "abc")
          > (string-concatenate (list s s s))
          "abcabcabc"
          > (string-concatenate (list s s s) ",")
          "abc,abc,abc"


 -- procedure: substring-fill! STRING START END FILL
     This procedure is like `string-fill!', but fills a selected part
     of the given string. It sets the elements of the string STRING,
     beginning with index START (inclusive) and ending with index END
     (exclusive) to FILL.  The value returned is unspecified.

     For example:

          > (define s (string #\a #\b #\c #\d #\e #\f))
          > (substring-fill! s 3 5 #\x)
          > s
          "abcxxf"


 -- procedure: substring-move! SRC-STRING SRC-START SRC-END DST-STRING
          DST-START
     This procedure replaces part of the contents of string DST-STRING
     with part of the contents of string SRC-STRING. It copies elements
     from SRC-STRING, beginning with index SRC-START (inclusive) and
     ending with index SRC-END (exclusive) to DST-STRING beginning with
     index DST-START (inclusive).  The value returned is unspecified.

     For example:

          > (define s1 "123456")
          > (define s2 (string #\a #\b #\c #\d #\e #\f))
          > (substring-move! s1 3 5 s2 1)
          > s2
          "a45def"


 -- procedure: string-shrink! STRING K
     This procedure shortens the string STRING so that its new size is
     K.  The value returned is unspecified.

     For example:

          > (define s (string #\a #\b #\c #\d #\e #\f))
          > s
          "abcdef"
          > (string-shrink! s 3)
          > s
          "abc"


 -- procedure: box OBJ
 -- procedure: box? OBJ
 -- procedure: unbox BOX
 -- procedure: set-box! BOX OBJ
     These procedures implement the "box" data type.  A box is a cell
     containing a single mutable field.  The lexical syntax of a box
     containing the object OBJ is `#&OBJ' (*note Box syntax::).

     The procedure `box' returns a new box object whose content is
     initialized to OBJ.  The procedure `box?' returns `#t' if OBJ is a
     box, and otherwise returns `#f'.  The procedure `unbox' returns
     the content of the box BOX.  The procedure `set-box!' changes the
     content of the box BOX to OBJ.  The procedure `set-box!' returns
     an unspecified value.

     For example:

          > (define b (box 0))
          > b
          #&0
          > (define (inc!) (set-box! b (+ (unbox b) 1)))
          > (inc!)
          > b
          #&1
          > (unbox b)
          1


 -- procedure: keyword? OBJ
 -- procedure: keyword->string KEYWORD
 -- procedure: string->keyword STRING
     These procedures implement the "keyword" data type.  Keywords are
     similar to symbols but are self evaluating and distinct from the
     symbol data type.  The lexical syntax of keywords is specified in
     *Note Keyword syntax::.

     The procedure `keyword?' returns `#t' if OBJ is a keyword, and
     otherwise returns `#f'.  The procedure `keyword->string' returns
     the name of KEYWORD as a string.  The procedure `string->keyword'
     returns the keyword whose name is STRING.

     For example:

          > (keyword? 'color)
          #f
          > (keyword? color:)
          #t
          > (keyword->string color:)
          "color"
          > (string->keyword "color")
          color:


 -- procedure: gensym [PREFIX]
     This procedure returns a new "uninterned symbol".  Uninterned
     symbols are guaranteed to be distinct from the symbols generated
     by the procedures `read' and `string->symbol'.  The symbol PREFIX
     is the prefix used to generate the new symbol's name.  If it is
     not specified, the prefix defaults to `g'.

     For example:

          > (gensym)
          #:g0
          > (gensym)
          #:g1
          > (gensym 'star-trek-)
          #:star-trek-2


 -- procedure: string->uninterned-symbol NAME [HASH]
 -- procedure: uninterned-symbol? OBJ
     The procedure `string->uninterned-symbol' returns a new uninterned
     symbol whose name is NAME and hash is HASH.  The name must be a
     string and the hash must be a nonnegative fixnum.

     The procedure `uninterned-symbol?' returns `#t' when OBJ is a
     symbol that is uninterned and `#f' otherwise.

     For example:

          > (uninterned-symbol? (gensym))
          #t
          > (string->uninterned-symbol "foo")
          #:foo:
          > (uninterned-symbol? (string->uninterned-symbol "foo"))
          #t
          > (uninterned-symbol? 'hello)
          #f
          > (uninterned-symbol? 123)
          #f


 -- procedure: string->uninterned-keyword NAME [HASH]
 -- procedure: uninterned-keyword? OBJ
     The procedure `string->uninterned-keyword' returns a new uninterned
     keyword whose name is NAME and hash is HASH.  The name must be a
     string and the hash must be a nonnegative fixnum.

     The procedure `uninterned-keyword?' returns `#t' when OBJ is a
     keyword that is uninterned and `#f' otherwise.

     For example:

          > (string->uninterned-keyword "foo")
          #:foo:
          > (uninterned-keyword? (string->uninterned-keyword "foo"))
          #t
          > (uninterned-keyword? hello:)
          #f
          > (uninterned-keyword? 123)
          #f


 -- procedure: identity OBJ
     This procedure returns OBJ.


 -- procedure: void
     This procedure returns the void object.  The read-eval-print loop
     prints nothing when the result is the void object.


 -- procedure: eval EXPR [ENV]
     The first parameter is a datum representing an expression.  The
     `eval' procedure evaluates this expression in the global
     interaction environment and returns the result.  If present, the
     second parameter is ignored (it is provided for compatibility with
     R5RS).

     For example:

          > (eval '(+ 1 2))
          3
          > ((eval 'car) '(1 2))
          1
          > (eval '(define x 5))
          > x
          5


 -- special form: define-macro (name define-formals) body
     Define name as a macro special form which expands into body.  This
     form can only appear where a `define' form is acceptable.  Macros
     are lexically scoped.  The scope of a local macro definition
     extends from the definition to the end of the body of the
     surrounding binding construct.  Macros defined at the top level of
     a Scheme module are only visible in that module.  To have access
     to the macro definitions contained in a file, that file must be
     included either directly using the `include' special form or
     indirectly with the `import' special form.  Macros which are
     visible from the REPL are also visible during the compilation of
     Scheme source files.

     For example:

          (define-macro (unless test . body)
            `(if ,test #f (begin ,@body)))

          (define-macro (push var #!optional val)
            `(set! ,var (cons ,val ,var)))

     To examine the code into which a macro expands you can use the
     compiler's `-expansion' option or the `pp' procedure.  For example:

          > (define-macro (push var #!optional val)
              `(set! ,var (cons ,val ,var)))
          > (pp (lambda () (push stack 1) (push stack) (push stack 3)))
          (lambda ()
            (set! stack (cons 1 stack))
            (set! stack (cons #f stack))
            (set! stack (cons 3 stack)))


 -- special form: define-syntax name expander
     Define name as a macro special form whose expansion is specified
     by expander.  This form is available only when the runtime option
     `-:s' is used.  This option causes the loading of the
     `~~lib/syntax-case' support library, which is the Hieb and Dybvig
     portable `syntax-case' implementation which has been ported to the
     Gambit interpreter and compiler.  Note that this implementation of
     `syntax-case' does not support special forms that are specific to
     Gambit.

     For example:

          $ gsi -:s
          Gambit v4.9.4

          > (define-syntax unless
              (syntax-rules ()
                ((unless test body ...)
                 (if test #f (begin body ...)))))
          > (let ((test 111)) (unless (= 1 2) (list test test)))
          (111 111)
          > (pp (lambda () (let ((test 111)) (unless (= 1 2) (list test test)))))
          (lambda () ((lambda (%%test14) (if (= 1 2) #f (list %%test14 %%test14))) 111))
          > (unless #f (pp xxx))
          *** ERROR IN (console)@7.16 -- Unbound variable: xxx


 -- procedure: compilation-target
     This procedure can only be executed during the phase of the Scheme
     code's processing (compilation) that corresponds to macro
     expansion.  Calls to this procedure are typically contained in
     macro definitions but they can also be contained in procedures
     that are called from a macro definition's body directly or
     indirectly.

     The result returned by the `compilation-target' procedure gives an
     indication of the target language of the compilation.  This can be
     used to write macros that depend on the type of compilation and
     the target language.

     When the result is the symbol `T' the macro expansion is in the
     context of compiling to the target language `T', e.g. `C', `js',
     etc.  When the result is a single element list `(T)' the macro
     expansion is for the interpreter which itself was compiled for the
     target language `T', e.g. `(C)', `(js)', etc.

     For example:

          $ cat ct.scm
          (define (level-0)
            (string-append "0: " (object->string (compilation-target))))

          (define-macro (test)

            (define (level-1)
              (string-append "1: " (object->string (compilation-target))))

            (define-macro (level-2)
              (string-append "2: " (object->string (compilation-target))))

            `(list ,(level-1) ,(level-2)))

          (pp (test))

          (pp (level-0)) ;; run time exception
          $ gsi ct.scm
          ("1: (C)" "2: (C)")
          *** ERROR IN level-0, "ct.scm"@2.40 -- Not in compilation context
          (compilation-target)
          $ gsc -target js -exe ct.scm
          $ ./ct
          ("1: js" "2: (C)")
          *** ERROR IN level-0 -- Not in compilation context
          (compilation-target)

     Regardless of whether `ct.scm' is being processed by the
     interpreter or the compiler, the body of the `level-0' procedure
     is not in a compilation context and in the body of the `level-2'
     macro the compilation target is `(C)' indicating that the macro
     expansion is being done for interpretation.

     During the execution of the `level-1' procedure, the compilation
     target will correspond to what is processing `ct.scm' (interpreter
     or compiler).

     Note that the compilation target can also be tested by the
     `cond-expand' special form.


 -- special form: cond-expand CE-CLAUSE ...
     The `cond-expand' expression type provides a way to statically
     expand different expressions depending on the presence or absence
     of a set of features.  A CE-CLAUSE takes the following form:

          (FEATURE-REQUIREMENT EXPRESSION ...)

     The last clause can be an "else clause," which has the form

          (else EXPRESSION)

     A FEATURE-REQUIREMENT takes one of the following forms:

        * `FEATURE-IDENTIFIER'

        * `(library LIBRARY-NAME)'

        * `(and FEATURE-REQUIREMENT ...)'

        * `(or FEATURE-REQUIREMENT ...)'

        * `(not FEATURE-REQUIREMENT)'

        * `(compilation-target TARGET ...)'


     The runtime system maintains a list of feature identifiers which
     are present, as well as a list of libraries which can be imported.
     The value of a FEATURE-REQUIREMENT is determined by replacing each
     `FEATURE-IDENTIFIER' and `(library LIBRARY-NAME)' on the runtime
     system's lists with `#t'.  Similarly, `#t' replaces each
     `(compilation-target TARGET ...)'  for which one of the TARGET
     matches the expansion time value of `(compilation-target)', with a
     `TARGET' of `(_)' matching any single element list (i.e. the
     interpreter).  All other `FEATURE-IDENTIFIER', `(library
     LIBRARY-NAME)', and `(compilation-target TARGET ...)' are replaced
     with `#f'.  The resulting expression is then evaluated as a Scheme
     boolean expression under the normal interpretation of `and', `or',
     and `not'.

     A `cond-expand' is then expanded by evaluating the
     FEATURE-REQUIREMENTs of successive CE-CLAUSEs in order until one
     of them returns `#t'. When a true clause is found, the
     corresponding EXPRESSIONs are expanded to a `begin', and the
     remaining clauses are ignored. If none of the FEATURE-REQUIREMENTs
     evaluate to `#t', then if there is an else clause, its EXPRESSIONs
     are included. Otherwise, an expansion time error is raised. Unlike
     `cond', `cond-expand' does not depend on the value of any
     variables.

     The feature identifier `gambit' is always true when the
     `cond-expand' is expanded by the Gambit interpreter or compiler.

     For example:

          > (cond-expand (foobar 111) (gambit 222) (else 333))
          222
          > (cond-expand ((compilation-target js) 111) (else 222))
          222
          > (cond-expand ((compilation-target (_)) 111) (else 222))
          111


 -- special form: define-cond-expand-feature FEATURE-IDENTIFIER ...
     The `define-cond-expand-feature' form can be used to add the
     feature identifiers FEATURE-IDENTIFIER ... to the list of features
     maintained by the runtime system.  These features are usable for
     the expansion of following `cond-expand' forms in the same file of
     source code, and the processing of other files and REPL
     interactions.

     For example:

          > (cond-expand (foobar 111) (gambit 222) (else 333))
          222
          > (define-cond-expand-feature foobar)
          > (cond-expand (foobar 111) (gambit 222) (else 333))
          111


 -- special form: declare declaration...
     This form introduces declarations to be used by the compiler
     (currently the interpreter ignores the declarations).  This form
     can only appear where a `define' form is acceptable.  Declarations
     are lexically scoped in the same way as macros.  The following
     declarations are accepted by the compiler:

    `(DIALECT)'
          Use the given dialect's semantics.  DIALECT can be:
          `ieee-scheme', `r4rs-scheme', `r5rs-scheme' or
          `gambit-scheme'.

    `(STRATEGY)'
          Select block compilation or separate compilation.  In block
          compilation, the compiler assumes that global variables
          defined in the current file that are not mutated in the file
          will never be mutated.  STRATEGY can be: `block' or
          `separate'.

    `([not] inline)'
          Allow (or disallow) inlining of user procedures.

    `([not] inline-primitives PRIMITIVE...)'
          The given primitives should (or should not) be inlined if
          possible (all primitives if none specified).

    `(inlining-limit N)'
          Select the degree to which the compiler inlines user
          procedures.  N is the upper-bound, in percent, on code
          expansion that will result from inlining.  Thus, a value of
          300 indicates that the size of the program will not grow by
          more than 300 percent (i.e. it will be at most 4 times the
          size of the original).  A value of 0 disables inlining.  The
          size of a program is the total number of subexpressions it
          contains (i.e. the size of an expression is one plus the size
          of its immediate subexpressions).  The following conditions
          must hold for a procedure to be inlined: inlining the
          procedure must not cause the size of the call site to grow
          more than specified by the inlining limit, the site of
          definition (the `define' or `lambda') and the call site must
          be declared as `(inline)', and the compiler must be able to
          find the definition of the procedure referred to at the call
          site (if the procedure is bound to a global variable, the
          definition site must have a `(block)' declaration).  Note
          that inlining usually causes much less code expansion than
          specified by the inlining limit (an expansion around 10% is
          common for N=370).

    `(allocation-limit N)'
          Indicate the maximum size of objects allocated with
          `make-vector', `make-string', `make-u8vector', etc.  Knowing
          the maximum size allows the compiler to inline calls to these
          allocators for small allocations.  This is only supported by
          the C target and only up to a size that is allowed for
          _movable objects_ (typically on the order of 1-2 KB).  When N
          is an exact nonnegative integer it is the upper-bound on the
          number of elements of the allocated objects.  When N is `#t'
          a dynamic test of the size is done.  When N is `#f' the
          allocation operation is not inlined.

    `([not] lambda-lift)'
          Lambda-lift (or don't lambda-lift) locally defined procedures.

    `([not] constant-fold)'
          Allow (or disallow) constant-folding of primitive procedures.

    `([not] standard-bindings VAR...)'
          The given global variables are known (or not known) to be
          equal to the value defined for them in the dialect (all
          variables defined in the standard if none specified).

    `([not] extended-bindings VAR...)'
          The given global variables are known (or not known) to be
          equal to the value defined for them in the runtime system
          (all variables defined in the runtime if none specified).

    `([not] run-time-bindings VAR...)'
          The given global variables will be tested at run time to see
          if they are equal to the value defined for them in the
          runtime system (all variables defined in the runtime if none
          specified).

    `([not] safe)'
          Generate (or don't generate) code that will prevent fatal
          errors at run time.  Note that in `safe' mode certain
          semantic errors will not be checked as long as they can't
          crash the system.  For example the primitive `char=?' may
          disregard the type of its arguments in `safe' as well as `not
          safe' mode.

    `([not] interrupts-enabled)'
          Generate (or don't generate) interrupt checks.  Interrupt
          checks are used to detect user interrupts and also to check
          for stack overflows.  Interrupt checking should not be turned
          off casually.

    `([not] poll-on-return)'
          Generate (or don't generate) interrupt checks on procedure
          returns (when interrupt checking is enabled).  This
          declaration has no effect on the behavior of interrupt
          checking on procedure calls, which is needed to guarantee
          that stack overflows are handled properly.

    `([not] debug)'
          Enable (or disable) the generation of debugging information.
          The kind of debugging information that is generated depends
          on the declarations `debug-location', `debug-source', and
          `debug-environments'.  If any of the command line options
          `-debug', `-debug-location', `-debug-source' and
          `-debug-environments' are present, the `debug' declaration is
          initially enabled, otherwise it is initially disabled.  When
          all kinds of debugging information are generated there is a
          substantial increase in the C compilation time and the size
          of the generated code.  When compiling a 3000 line Scheme
          file it was observed that the total compilation time is 500%
          longer and the executable code is 150% bigger.

    `([not] debug-location)'
          Select (or deselect) source code location debugging
          information.  When this declaration and the `debug'
          declaration are in effect, run time error messages indicate
          the location of the error in the source code file.  If any of
          the command line options `-debug-source' and
          `-debug-environments' are present and `-debug-location' is
          absent, the `debug-location' declaration is initially
          disabled, otherwise it is initially enabled.  When compiling
          a 3000 line Scheme file it was observed that the total
          compilation time is 200% longer and the executable code is
          60% bigger.

    `([not] debug-source)'
          Select (or deselect) source code debugging information.  When
          this declaration and the `debug' declaration are in effect,
          run time error messages indicate the source code, the
          backtraces are more precise, and the `pp' procedure will
          display the source code of compiled procedures.  If any of
          the command line options `-debug-location' and
          `-debug-environments' are present and `-debug-source' is
          absent, the `debug-source' declaration is initially disabled,
          otherwise it is initially enabled.  When compiling a 3000
          line Scheme file it was observed that the total compilation
          time is 90% longer and the executable code is 90% bigger.

    `([not] debug-environments)'
          Select (or deselect) environment debugging information.  When
          this declaration and the `debug' declaration are in effect,
          the debugger will have access to the environments of the
          continuations.  In other words the local variables defined in
          compiled procedures (and not optimized away by the compiler)
          will be shown by the `,e' REPL command.  If any of the
          command line options `-debug-location' and `-debug-source'
          are present and `-debug-environments' is absent, the
          `debug-environments' declaration is initially disabled,
          otherwise it is initially enabled.  When compiling a 3000
          line Scheme file it was observed that the total compilation
          time is 70% longer and the executable code is 40% bigger.

    `([not] proper-tail-calls)'
          Generate (or don't generate) proper tail calls.  When proper
          tail calls are turned off, tail calls are handled like
          non-tail calls, that is a continuation frame will be created
          for all calls regardless of their kind.  This is useful for
          debugging because the caller of a procedure will be visible
          in the backtrace produced by the REPL's `,b' command even
          when the call is a tail call.  Be advised that this does
          cause stack space to be consumed for tail calls which may
          cause the stack to overflow when performing long iterations
          with tail calls (whether they are expressed with a `letrec',
          named `let', `do', or other form).

    `([not] generative-lambda)'
          Force (or don't force) the creation of fresh closures when
          evaluating lambda-expressions.  A fresh closure is always
          created when a lambda-expression has at least one free
          variable (that has not been eliminated by dead-code
          elimination or other compiler optimization) or when the
          generative-lambda declaration is turned on.  When a
          lambda-expression has no free variables and the
          generative-lambda declaration is turned off, the value of the
          lambda-expression may be the same procedure (in the sense of
          `eq?').

    `([not] optimize-dead-local-variables)'
          Remove (or preserve) the dead local variables in the
          environment.  Preserving the dead local variables is useful
          for debugging because continuations will contain the dead
          variables.  Thus, if the code is also compiled with the
          declaration `debug-environments' the `,e', `,ed', `,be', and
          `,bed' REPL commands will display the dead variables.  On the
          other hand, preserving the dead local variables may change
          the space complexity of the program (i.e. some of the data
          that would normally be reclaimed by the garbage collector
          will not be).  Note that due to other compiler optimizations
          some dead local variables may be removed regardless of this
          declaration.

    `([not] optimize-dead-definitions VAR...)'
          Remove (or preserve) the dead toplevel definitions of the
          given global variables (all global variables if none
          specified).  A toplevel definition is dead if it is not
          referenced by toplevel expressions of the program or toplevel
          definitions that aren't dead (regardless of the evaluation of
          its expression causing a side-effect).  When a module is
          separately compiled and some of its definitions are only used
          by other modules, this declaration must be used with care to
          keep definitions that are used by other modules, for example
          if `foo' is referenced in another module the following
          declaration should be used: `(declare (not
          optimize-dead-definitions foo))'.

    `(NUMBER-TYPE PRIMITIVE...)'
          Numeric arguments and result of the specified primitives are
          known to be of the given type (all primitives if none
          specified).  NUMBER-TYPE can be: `generic', `fixnum', or
          `flonum'.

    `(MOSTLY-NUMBER-TYPE PRIMITIVE...)'
          Numeric arguments and result of the specified primitives are
          expected to be most often of the given type (all primitives
          if none specified).  MOSTLY-NUMBER-TYPE can be:
          `mostly-generic', `mostly-fixnum', `mostly-fixnum-flonum',
          `mostly-flonum', or `mostly-flonum-fixnum'.


     The default declarations used by the compiler are equivalent to:

          (declare
            (gambit-scheme)
            (separate)
            (inline)
            (inline-primitives)
            (inlining-limit 370)
            (allocation-limit #t)
            (constant-fold)
            (lambda-lift)
            (not standard-bindings)
            (not extended-bindings)
            (run-time-bindings)
            (safe)
            (interrupts-enabled)
            (not poll-on-return)
            (not debug)           ;; depends on debugging command line options
            (debug-location)      ;; depends on debugging command line options
            (debug-source)        ;; depends on debugging command line options
            (debug-environments)  ;; depends on debugging command line options
            (proper-tail-calls)
            (not generative-lambda)
            (optimize-dead-local-variables)
            (not optimize-dead-definitions)
            (generic)
            (mostly-fixnum-flonum)
          )

     These declarations are compatible with the semantics of R5RS Scheme
     and includes a few procedures from R6RS (mainly fixnum specific and
     flonum specific procedures).  Typically used declarations that
     enhance performance, at the cost of violating the R5RS Scheme
     semantics, are: `(standard-bindings)', `(block)', `(not safe)' and
     `(fixnum)'.


 -- procedure: continuation? OBJ
 -- procedure: continuation-capture PROC
 -- procedure: continuation-graft CONT PROC OBJ...
 -- procedure: continuation-return CONT OBJ...
     These procedures provide access to internal first-class
     continuations which are represented using continuation objects
     distinct from procedures.

     The procedure `continuation?' returns `#t' when OBJ is a
     continuation object and `#f' otherwise.

     The procedure `continuation-capture' is similar to the `call/cc'
     procedure but it represents the continuation with a continuation
     object.  The PROC parameter must be a procedure accepting a single
     argument.  The procedure `continuation-capture' reifies its
     continuation and calls PROC with the corresponding continuation
     object as its sole argument.  Like for `call/cc', the implicit
     continuation of the call to PROC is the implicit continuation of
     the call to `continuation-capture'.

     The procedure `continuation-graft' performs a procedure call to
     the procedure PROC with arguments OBJ... and the implicit
     continuation corresponding to the continuation object CONT.  The
     current continuation of the call to procedure `continuation-graft'
     is ignored.

     The procedure `continuation-return' invokes the implicit
     continuation corresponding to the continuation object CONT with
     the result(s) OBJ....  This procedure can be easily defined in
     terms of `continuation-graft':

          (define (continuation-return cont . objs)
            (continuation-graft cont apply values objs))

     For example:

          > (define x #f)
          > (define p (make-parameter 11))
          > (pp (parameterize ((p 22))
                  (cons 33 (continuation-capture
                            (lambda (c) (set! x c) 44)))))
          (33 . 44)
          > x
          #<continuation #2>
          > (continuation-return x 55)
          (33 . 55)
          > (continuation-graft x (lambda () (expt 2 10)))
          (33 . 1024)
          > (continuation-graft x expt 2 10)
          (33 . 1024)
          > (continuation-graft x (lambda () (p)))
          (33 . 22)
          > (define (map-sqrt1 lst)
              (call/cc
               (lambda (k)
                 (map (lambda (x)
                        (if (< x 0)
                            (k 'error)
                            (sqrt x)))
                      lst))))
          > (map-sqrt1 '(1 4 9))
          (1 2 3)
          > (map-sqrt1 '(1 -1 9))
          error
          > (define (map-sqrt2 lst)
              (continuation-capture
               (lambda (c)
                 (map (lambda (x)
                        (if (< x 0)
                            (continuation-return c 'error)
                            (sqrt x)))
                      lst))))
          > (map-sqrt2 '(1 4 9))
          (1 2 3)
          > (map-sqrt2 '(1 -1 9))
          error


 -- procedure: display-exception EXC [PORT]
 -- procedure: display-exception-in-context EXC CONT [PORT]
 -- procedure: display-procedure-environment PROC [PORT]
 -- procedure: display-continuation-environment CONT [PORT]
 -- procedure: display-continuation-dynamic-environment CONT [PORT]

 -- procedure: display-continuation-backtrace CONT [PORT [DISPLAY-ENV?
          [ALL-FRAMES? [MAX-HEAD [MAX-TAIL [DEPTH]]]]]]
     The procedure `display-continuation-backtrace' displays the frames
     of the continuation corresponding to the continuation object CONT
     on the port PORT.  If it is not specified, PORT defaults to the
     current output-port.  The frames are displayed in the same format
     as the REPL's `,b' command.

     The parameter DISPLAY-ENV?, which defaults to `#f', controls if
     the frames are displayed with its environment (the variables
     accessible and their bindings).

     The parameter ALL-FRAMES?, which defaults to `#f', controls which
     frames are displayed.  Some frames of ancillary importance, such
     as internal frames created by the interpreter, are not displayed
     when ALL-FRAMES? is `#f'.  Otherwise all frames are displayed.

     The parameters MAX-HEAD and MAX-TAIL, which default to 10 and 4
     respectively, control how many frames are displayed at the head
     and tail of the continuation.

     The parameter DEPTH, which defaults to 0, causes the frame numbers
     to be offset by that value.

     For example:

          > (define x #f)
          > (define (fib n)
              (if (< n 2)
                  (continuation-capture
                   (lambda (c) (set! x c) 1))
                  (+ (fib (- n 1))
                     (fib (- n 2)))))
          > (fib 10)
          89
          > (display-continuation-backtrace x)
          0  fib             (console)@7:12     (fib (- n 2))
          1  fib             (console)@7:12     (fib (- n 2))
          2  fib             (console)@7:12     (fib (- n 2))
          3  fib             (console)@7:12     (fib (- n 2))
          4  fib             (console)@7:12     (fib (- n 2))
          5  (interaction)   (console)@8:1      (fib 10)
          #f
          > (display-continuation-backtrace x (current-output-port) #f #t)
          0  fib             (console)@7:12     (fib (- n 2))
          1  fib             (console)@6:9      (+ (fib (- n 1)) (fib (- ...
          2  fib             (console)@7:12     (fib (- n 2))
          3  fib             (console)@6:9      (+ (fib (- n 1)) (fib (- ...
          4  fib             (console)@7:12     (fib (- n 2))
          5  fib             (console)@6:9      (+ (fib (- n 1)) (fib (- ...
          6  fib             (console)@7:12     (fib (- n 2))
          7  fib             (console)@6:9      (+ (fib (- n 1)) (fib (- ...
          8  fib             (console)@7:12     (fib (- n 2))
          9  fib             (console)@6:9      (+ (fib (- n 1)) (fib (- ...
          ...
          13 ##with-no-result-expected-toplevel
          14 ##repl-debug
          15 ##repl-debug-main
          16 ##kernel-handlers
          #f
          > (display-continuation-backtrace x (current-output-port) #t #f)
          0  fib             (console)@7:12     (fib (- n 2))
                  n = 2
          1  fib             (console)@7:12     (fib (- n 2))
                  n = 4
          2  fib             (console)@7:12     (fib (- n 2))
                  n = 6
          3  fib             (console)@7:12     (fib (- n 2))
                  n = 8
          4  fib             (console)@7:12     (fib (- n 2))
                  n = 10
          5  (interaction)   (console)@8:1      (fib 10)
          #f
          > (display-continuation-backtrace x (current-output-port) #f #f 2 1 100)
          100 fib            (console)@7:12     (fib (- n 2))
          101 fib            (console)@7:12     (fib (- n 2))
          ...
          105 (interaction)  (console)@8:1      (fib 10)
          #f


6.4 Undocumented extensions
===========================

The procedures in this section are not yet documented.

 -- procedure: make-thread-group [NAME [THREAD-GROUP]]
 -- procedure: thread-group? OBJ
 -- procedure: thread-group-name THREAD-GROUP
 -- procedure: thread-group-parent THREAD-GROUP
 -- procedure: thread-group-resume! THREAD-GROUP
 -- procedure: thread-group-suspend! THREAD-GROUP
 -- procedure: thread-group-terminate! THREAD-GROUP
 -- procedure: thread-group->thread-group-list THREAD-GROUP
 -- procedure: thread-group->thread-group-vector THREAD-GROUP
 -- procedure: thread-group->thread-list THREAD-GROUP
 -- procedure: thread-group->thread-vector THREAD-GROUP
 -- procedure: thread-group-specific THREAD-GROUP
 -- procedure: thread-group-specific-set! THREAD-GROUP OBJ

 -- procedure: thread-state THREAD
 -- procedure: thread-state-uninitialized? THREAD-STATE
 -- procedure: thread-state-initialized? THREAD-STATE
 -- procedure: thread-state-running? THREAD-STATE
 -- procedure: thread-state-running-processor THREAD-STATE
 -- procedure: thread-state-waiting? THREAD-STATE
 -- procedure: thread-state-waiting-for THREAD-STATE
 -- procedure: thread-state-waiting-timeout THREAD-STATE
 -- procedure: thread-state-normally-terminated? THREAD-STATE
 -- procedure: thread-state-normally-terminated-result THREAD-STATE
 -- procedure: thread-state-abnormally-terminated? THREAD-STATE
 -- procedure: thread-state-abnormally-terminated-reason THREAD-STATE
 -- procedure: top [TIMEOUT [THREAD-GROUP [PORT]]]

 -- procedure: thread-interrupt! THREAD [THUNK]

 -- procedure: thread-suspend! THREAD
 -- procedure: thread-resume! THREAD

 -- procedure: thread-thread-group THREAD

 -- special form: define-type-of-thread name field...

 -- procedure: thread-init! THREAD THUNK [NAME [THREAD-GROUP]]

 -- procedure: initialized-thread-exception? OBJ
 -- procedure: initialized-thread-exception-procedure EXC
 -- procedure: initialized-thread-exception-arguments EXC

 -- procedure: uninitialized-thread-exception? OBJ
 -- procedure: uninitialized-thread-exception-procedure EXC
 -- procedure: uninitialized-thread-exception-arguments EXC

 -- procedure: inactive-thread-exception? OBJ
 -- procedure: inactive-thread-exception-procedure EXC
 -- procedure: inactive-thread-exception-arguments EXC

 -- procedure: rpc-remote-error-exception? OBJ
 -- procedure: rpc-remote-error-exception-procedure EXC
 -- procedure: rpc-remote-error-exception-arguments EXC
 -- procedure: rpc-remote-error-exception-message EXC

 -- procedure: invalid-utf8-encoding-exception? OBJ
 -- procedure: invalid-utf8-encoding-exception-procedure EXC
 -- procedure: invalid-utf8-encoding-exception-arguments EXC

 -- procedure: module-not-found-exception? OBJ
 -- procedure: module-not-found-exception-procedure EXC
 -- procedure: module-not-found-exception-arguments EXC

 -- procedure: processor? OBJ
 -- procedure: current-processor
 -- procedure: processor-id PROCESSOR

 -- procedure: timeout->time TIMEOUT

 -- procedure: current-second

 -- procedure: current-jiffy

 -- procedure: jiffies-per-second

 -- procedure: get-environment-variable NAME

 -- procedure: get-environment-variables

 -- procedure: executable-path

 -- procedure: command-name
 -- procedure: command-args

 -- procedure: script-file
 -- procedure: script-directory

 -- procedure: open-dummy

 -- procedure: port-settings-set! PORT SETTINGS

 -- procedure: port-io-exception-handler-set! PORT HANDLER

 -- procedure: input-port-bytes-buffered PORT

 -- procedure: input-port-characters-buffered PORT

 -- procedure: nonempty-input-port-character-buffer-exception? OBJ
 -- procedure: nonempty-input-port-character-buffer-exception-arguments
          EXC
 -- procedure: nonempty-input-port-character-buffer-exception-procedure
          EXC

 -- procedure: repl-input-port
 -- procedure: repl-output-port
 -- procedure: repl-error-port
 -- procedure: console-port

 -- procedure: current-user-interrupt-handler [HANDLER]
 -- procedure: default-user-interrupt-handler
 -- procedure: defer-user-interrupts

 -- procedure: primordial-exception-handler EXC

 -- procedure: err-code->string CODE

 -- procedure: foreign? OBJ
 -- procedure: foreign-tags FOREIGN
 -- procedure: foreign-address FOREIGN
 -- procedure: foreign-release! FOREIGN
 -- procedure: foreign-released? FOREIGN

 -- procedure: invalid-hash-number-exception? OBJ
 -- procedure: invalid-hash-number-exception-procedure EXC
 -- procedure: invalid-hash-number-exception-arguments EXC

 -- procedure: tcp-client-local-socket-info TCP-CLIENT-PORT
 -- procedure: tcp-client-peer-socket-info TCP-CLIENT-PORT
 -- procedure: tcp-client-self-socket-info TCP-CLIENT-PORT

 -- procedure: tcp-server-socket-info TCP-SERVER-PORT

 -- procedure: socket-info? OBJ
 -- procedure: socket-info-address SOCKET-INFO
 -- procedure: socket-info-family SOCKET-INFO
 -- procedure: socket-info-port-number SOCKET-INFO

 -- procedure: system-version
 -- procedure: system-version-string

 -- procedure: system-type
 -- procedure: system-type-string
 -- procedure: configure-command-string

 -- procedure: system-stamp

 -- special form: future EXPR
 -- procedure: touch OBJ

 -- procedure: tty? OBJ
 -- procedure: tty-history TTY
 -- procedure: tty-history-set! TTY HISTORY
 -- procedure: tty-history-max-length-set! TTY N
 -- procedure: tty-paren-balance-duration-set! TTY DURATION
 -- procedure: tty-text-attributes-set! TTY ATTRIBUTES
 -- procedure: tty-mode-set! TTY MODE
 -- procedure: tty-type-set! TTY TYPE

 -- procedure: with-input-from-port PORT THUNK
 -- procedure: with-output-to-port PORT THUNK

 -- procedure: input-port-char-position PORT
 -- procedure: output-port-char-position PORT

 -- procedure: open-event-queue SELECTOR

 -- procedure: main ...

 -- procedure: dead-end

 -- procedure: poll-point

 -- special form: define-record-type ...
 -- special form: define-type ...

 -- special form: this-source-file

 -- special form: receive ...

 -- special form: define-values ...

 -- special form: define-module-alias ...

 -- special form: r7rs-guard ...

 -- special form: case-lambda ...
 -- special form: syntax-case ...
 -- special form: syntax ...

 -- procedure: datum->syntax OBJ
 -- procedure: syntax->datum STX
 -- procedure: syntax->list STX
 -- procedure: syntax->vector STX

 -- procedure: length+ CLIST

 -- procedure: car+cdr PAIR

 -- procedure: first PAIR
 -- procedure: second PAIR
 -- procedure: third PAIR
 -- procedure: fourth PAIR
 -- procedure: fifth PAIR
 -- procedure: sixth PAIR
 -- procedure: seventh PAIR
 -- procedure: eighth PAIR
 -- procedure: ninth PAIR
 -- procedure: tenth PAIR

 -- procedure: not-pair? X

 -- procedure: null-list? LIST

 -- procedure: proper-list? S

 -- procedure: circular-list? S

 -- procedure: dotted-list? S

 -- procedure: filter PRED LIST
 -- procedure: remove PRED LIST
 -- procedure: remq ELEM LIST

 -- procedure: concatenate LIST-OF-LISTS [SEPARATOR]
 -- procedure: concatenate! LIST-OF-LISTS

 -- procedure: list= ELT= LIST ...

 -- procedure: list-set LIST K VAL
 -- procedure: list-set! LIST K VAL

 -- procedure: fold PROC BASE LIST ...
 -- procedure: fold-right PROC BASE LIST ...

 -- procedure: iota COUNT [START [STEP]]

 -- procedure: circular-list X Y...

 -- procedure: cons* X Y...

 -- procedure: list-copy LIST

 -- procedure: list-tabulate N INIT-PROC

 -- procedure: make-list N [FILL]

 -- procedure: reverse! LIST

 -- procedure: append-reverse LIST TAIL
 -- procedure: append-reverse! LIST TAIL

 -- procedure: xcons D A

 -- procedure: take X I
 -- procedure: drop X I

 -- procedure: last PAIR
 -- procedure: last-pair PAIR

 -- procedure: list-sort PROC LIST
 -- procedure: list-sort! PROC LIST

 -- procedure: finite? X
 -- procedure: infinite? X
 -- procedure: nan? X

 -- procedure: acosh X

 -- procedure: asinh X

 -- procedure: atanh X

 -- procedure: cosh X

 -- procedure: sinh X

 -- procedure: tanh X

 -- procedure: conjugate X

 -- procedure: bits BOOL...
 -- procedure: bits->list I [LEN]
 -- procedure: list->bits LIST
 -- procedure: bits->vector I [LEN]
 -- procedure: vector->bits VECTOR

 -- special form: six.infix DATUM
 -- undefined: six.!
 -- special form: six.!x X
 -- special form: six.&x X
 -- special form: six.**x X
 -- special form: six.*x X
 -- special form: six.++x X
 -- special form: six.+x X
 -- special form: six.-x X
 -- special form: six.-x X
 -- special form: six.arrow EXPR IDENT
 -- undefined: six.asyncx X
 -- undefined: six.awaitx X
 -- undefined: six.break
 -- special form: six.call FUNC ARG...
 -- undefined: six.case
 -- undefined: six.clause
 -- special form: six.compound STATEMENT...
 -- special form: six.cons X Y
 -- undefined: six.continue
 -- special form: six.define-procedure IDENT PROC
 -- special form: six.define-variable IDENT TYPE DIMS INIT
 -- special form: six.do-while STAT EXPR
 -- special form: six.dot EXPR IDENT
 -- special form: six.for STAT1 EXPR2 EXPR3 STAT2
 -- undefined: six.goto EXPR
 -- undefined: six.from-import EXPR1 EXPR2
 -- undefined: six.from-import-* EXPR
 -- special form: six.identifier IDENT
 -- special form: six.if EXPR STAT1 [STAT2]
 -- undefined: six.import EXPR
 -- special form: six.index EXPR1 EXPR2
 -- undefined: six.label IDENT STAT
 -- special form: six.list X Y
 -- special form: six.literal VALUE
 -- procedure: six.make-array INIT DIM...
 -- special form: six.new IDENT ARG...
 -- special form: six.null
 -- special form: six.procedure TYPE PARAMS STAT
 -- special form: six.procedure-body STAT...
 -- undefined: six.return
 -- undefined: six.switch
 -- undefined: six.typeofx X
 -- special form: six.while EXPR STAT...
 -- special form: six.x!==y X Y
 -- special form: six.x!=y X Y
 -- special form: six.x%=y X Y
 -- special form: six.x%y X Y
 -- special form: six.x&&y X Y
 -- special form: six.x&=y X Y
 -- special form: six.x&y X Y
 -- special form: six.x**=y X Y
 -- special form: six.x**y X Y
 -- special form: six.x*=y X Y
 -- special form: six.x*y X Y
 -- special form: six.x@=y X Y
 -- special form: six.x@y X Y
 -- special form: six.x++ X
 -- special form: six.x+=y X Y
 -- special form: six.x+y X Y
 -- special form: |six.x,y| X Y
 -- special form: six.x- X
 -- special form: six.x-=y X Y
 -- special form: six.x-y X Y
 -- special form: six.x//=y X Y
 -- special form: six.x//y X Y
 -- special form: six.x/=y X Y
 -- special form: six.x/y X Y
 -- undefined: six.x:-y X Y
 -- special form: six.x:=y X Y
 -- special form: six.x:y X Y
 -- special form: six.x<<=y X Y
 -- special form: six.x<<y X Y
 -- special form: six.x<=y X Y
 -- special form: six.x<y X Y
 -- special form: six.x===y X Y
 -- special form: six.x==y X Y
 -- special form: six.x=y X Y
 -- special form: six.x>=y X Y
 -- special form: six.x>>>=y X Y
 -- special form: six.x>>>y X Y
 -- special form: six.x>>=y X Y
 -- special form: six.x>>y X Y
 -- special form: six.x>y X Y
 -- special form: six.x?y:z X Y Z
 -- special form: six.x^=y X Y
 -- special form: six.x^y X Y
 -- special form: |six.x\|=y| X Y
 -- special form: |six.x\|y| X Y
 -- special form: |six.x\|\|y| X Y
 -- special form: six.xandy X Y
 -- undefined: six.xinstanceofy X Y
 -- special form: six.xiny X Y
 -- special form: six.xisy X Y
 -- special form: six.notx X Y
 -- special form: six.xory X Y
 -- special form: six.~x X
 -- undefined: six.yieldx X

 -- procedure: readtable-comment-handler READTABLE
 -- procedure: readtable-comment-handler-set READTABLE NEW-VALUE

 -- procedure: open-output-bytevector [U8VECTOR-OR-SETTINGS]

7 Modules
*********

Gambit supports multiple modularization approaches and constructs:
legacy modules, primitive modules and R7RS compatible modules.  These
are described in that order, which corresponds to increased abstraction
level.  Unless there is a need for detailed control over the modules,
it is best to use the R7RS compatible module system for the development
of new code.

7.1 Legacy Modules
==================

The legacy way of modularizing code, which was popular up to R5RS, is
still supported by Gambit.  It consists of using the `load' procedure
and the `include' form.  We discuss it first to introduce some useful
terms and explain the shortcomings of this modularization approach.

 -- procedure: load PATH
     The `load' procedure's PATH argument, a string, specifies the
     location in the file system of a file to load.  Loading a file
     executes the code contained in the file, which is either source
     code or compiled code (a dynamically loadable "object file"
     created by the Gambit Scheme compiler, see the procedure
     `compile-file').  When PATH has no extension the `load' procedure
     first attempts to load the file with no extension as a Scheme
     source file.  If that file doesn't exist it will search for both a
     source file and an object file.  The object file's path is
     obtained by adding to PATH a `.oN' extension with the highest
     consecutive version number starting with 1.  The source file's
     path is obtained by adding to PATH the file extensions `.sld',
     `.scm' and `.six' (the first found is the source file).  If both a
     source file and an object file exist, then the one with the latest
     modification time is loaded.  Otherwise the file that is found is
     loaded.  When PATH has an extension, the `load' procedure will
     only attempt to load the file with that specific extension.  After
     executing the code contained in the file, the `load' procedure
     returns the path of the file that was loaded.

     When a source code file is loaded its extension is used to
     determine how it is parsed, unless the file's first line is a
     special script line (see *Note Scheme scripts::).  When the
     extension is different from `.six' the content of the file is
     parsed using the normal Scheme prefix syntax.  When the extension
     is `.six' the content of the file is parsed using the Scheme infix
     syntax extension (see *Note Scheme infix syntax extension::).

     Due to operating system limitations, loading a given `.oN' object
     file more than once in the same process is not supported.  It is
     possible however to recompile the source code file to create a new
     `.oM' object file with M > N and load that object file.

     For example:

          $ cat my-mod.scm
          (define (double x) (* x 2))
          (display "my-mod has finished loading!!!\n")
          $ gsi
          Gambit v4.9.4

          > (load "my-mod")
          my-mod has finished loading!!!
          "/Users/feeley/gambit/doc/my-mod.scm"
          > (double 21)
          42
          > (load "my-mod.scm")
          my-mod has finished loading!!!
          "/Users/feeley/gambit/doc/my-mod.scm"
          > ,q
          $ gsc my-mod
          $ gsi
          Gambit v4.9.4

          > (load "my-mod")
          my-mod has finished loading!!!
          "/Users/feeley/gambit/doc/my-mod.o1"
          > (double 21)
          42
          > (load "my-mod")
          *** ERROR IN (console)@3.1 -- Can't load a given object file more than once
          (load "my-mod")
          1>

     Note that any macro definition in the loaded file is local to the
     file and is not visible from the REPL or other files that loaded
     this file.  The `include' form can be used to access the macros
     defined in another file.


 -- special form: include PATH
 -- special form: ##include PATH
     The path argument must be a string specifying the location of an
     existing file containing Scheme source code.  Relative paths are
     relative to the file that contains the `include' form.  The
     `include' special form splices the content of the specified source
     file.  This form can only appear where a `define' form is
     acceptable, i.e. at top level or in the body of a binding form.

     For example:

          $ cat my-defs.scm
          (define-macro (double x) `(* ,x 2))
          (define (quad y) (double (double y)))
          (display "howdy!\n")
          $ cat my-includer.scm
          (define (f x)
            (include "my-defs.scm")
            (+ 1 (quad x)))
          $ gsi
          Gambit v4.9.4

          > (load "my-includer")
          "/Users/feeley/udem-dlteam/gambit/my-includer.scm"
          > (f 10)
          howdy!
          41
          > (f 20)
          howdy!
          81


   With legacy modularization, the code that implements the module's
functionality is put in a source code file and this module is accessed
by other code by using a `load' or `include' of that file.  Here is an
example of an `angle0' module that is used by an `app0' main program:

     ;;;---------------------------------------- file: angle0/angle0.scm
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (* x factor))
     (define (rad->deg x) (/ x factor))

     ;;;---------------------------------------- file: app0.scm
     (load "angle0/angle0.scm")   ;; or (include "angle0/angle0.scm")
     (println "90 degrees is " (deg->rad 90) " radians")

     ;; run with:  gsi app0.scm

   This modularization approach has a number of issues:

   * It hinders code sharing among different programs and users because
     a shared module's location in the filesystem must be known to all
     modules loading or including it.  In the above example the path
     `"angle0/angle0.scm"' is relative so the `load' procedure will
     resolve the path incorrectly if the program executes
     `(current-directory "...")'  before calling `load'.

   * When a module is needed by more than one other module there will be
     code duplication, redundant evaluation/compilation, and probably
     incorrect execution if the module has side effects that should only
     happen once (displaying a message, opening a database on the
     filesystem, initializing the module's state, etc).  Moreover, when
     the module has been compiled to an object file it can't be loaded
     more than once.

   * All the definitions of a module will be put in the global
     environment (including top level macro definitions when using a
     top level `include' but not when using `load').  This pollutes the
     global environment with definitions that were not intended to be
     exported by the module's designer, such as the variable `factor'
     in the above example that is only meant to be used by the
     `deg->rad' and `rad->deg' procedures.  Other modules may also need
     a `factor' variable internally, for instance to convert distances
     from the metric to the english system. Nothing prevents such
     accidental clashes.


7.2 Primitive Modules
=====================

7.2.1 `##demand-module' and `##supply-module' forms
---------------------------------------------------

The `##demand-module' form offers a way to avoid the issues of multiple
loading and filesystem localization of modules.  The sole parameter of
this form is an (unevaluated) symbol that identifies the module on
which the module containing the `##demand-module' depends.  When a
module `A' contains a `(##demand-module B)', Gambit's runtime system
will ensure that module `B' is loaded before module `A' is loaded.  It
also registers the module in a cache when it is loaded so that it is
loaded exactly once.  In other words the `##demand-module' form
expresses the requirement that the current module needs the
functionality of another module.  A module can contain multiple uses of
`##demand-module' and possibly more than once for a given module.  The
`##demand-module' form can appear anywhere a `define' can appear.
There is also a related `##supply-module' form that should appear in
the module to declare the module's identity.

   Gambit's runtime system searches for modules in various directories,
by default in `~~lib' then in `~~userlib' (which maps to
`~/.gambit_userlib' by default).  These two directories are where
builtin modules and user installed modules are located respectively.
The source code for a module `M' is searched, in each of the "module
search order" directories, first in `M/M.EXT' and then in `M.EXT',
where .EXT is one of the acceptable Scheme source code file extensions
(`.sld', `.scm', `.six', etc).  The list of module search order
directories can be extended with the `-:search='DIR runtime option or
by a command line argument to `gsi' and `gsc' that ends with a path
separator or a `.'.

   With `##demand-module' and `##supply-module' the previous example
can be rewritten like this:

     ;;;---------------------------------------- file: angle1/angle1.scm
     (##supply-module angle1) ;; declare that this is the module angle1
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (* x factor))
     (define (rad->deg x) (/ x factor))

     ;;;---------------------------------------- file: app1.scm
     (##demand-module angle1) ;; declare dependency on module angle1
     (println "90 degrees is " (deg->rad 90) " radians")

     ;; run with either:  gsi . app1.scm
     ;;              or:  gsi -:search=. app1.scm
     ;;
     ;; or install the angle1 module to avoid the . and -:search=.

7.2.2 `##namespace' and `##import' forms
----------------------------------------

 -- special form: namespace ...
 -- special form: ##namespace ...
 -- special form: import MODULE-REF
 -- special form: ##import MODULE-REF

   The `##namespace' form offers a way to avoid name clashes by
specifying a mapping between identifiers.  The mapping it specifies has
the same scope as a macro definition: it applies to the rest of a
source code file if it is at top level, or applies to the rest of the
body of a binding form if it is used in the body of a binding form.
The call `(##namespace ("foo#" a b))' specifies that a reference to `a'
becomes `foo#a' and a reference to `b' becomes `foo#b'.  Here `foo#' is
the namespace.  Finer control over the mapping is possible by using
aliases as in `(##namespace ("foo#" (a bar) b))' which maps `a' to
`foo#bar' and `b' to `foo#b'.  Multiple namespace specifications can
appear in the body of the `##namespace' form.  When no identifiers are
specified, the mapping maps all identifiers not containing `#' to the
namespace.  For example in the scope of `(##namespace ("foo#"))' the
reference `x' maps to `foo#x' and the reference `bar#x' remains
unchanged.

   Given that modules are identified by a unique symbol, the global
names defined by a module `M' can be put in the namespace `M#' to avoid
name clashes with other modules.  The source code of module `M' and the
modules depending on `M' can explicitly prefix the global names defined
by `M' with `M#' or use a `##namespace' form to make this prefixing
implicit.  By convention the namespace definition for the identifiers
exported by module `M' is specified in the source code file `M#.scm' in
the same directory as the `M.scm' file.

   Using this convention and the `include' and `##namespace' forms, the
previous example can be rewritten like this:

     ;;;---------------------------------------- file: angle2/angle2#.scm
     (##namespace ("angle2#" deg->rad rad->deg))

     ;;;---------------------------------------- file: angle2/angle2.scm
     (include "angle2#.scm")
     (##namespace ("angle2#" factor))
     (##supply-module angle2)
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (* x factor))
     (define (rad->deg x) (/ x factor))

     ;;;---------------------------------------- file: app2.scm
     (include "angle2/angle2#.scm")
     (##demand-module angle2)
     (println "90 degrees is " (deg->rad 90) " radians")

   Note that the parameters of the two `include' forms are different,
but this is correct because the paths are relative to the file
containing the `include' form.  However the module localization problem
has been reintroduced for the file `angle2/angle2#.scm'.

   This problem can be solved using the `##import' form that combines
the semantics of the `include' and `##demand-module' forms.  The call
`(##import M)' will use the module search order directories to locate
the source code file of module `M' and will expand to an `include' of
the "hash" file `M#.EXT' if it exists in the same directory, and a
`(##demand-module M)'.

   In addition, a builtin module `gambit' exists that contains all the
global names exported by the runtime library.  The `gambit' module's
"hash" file `gambit#.scm' contains a `##namespace' form that lists all
the names exported by the runtime library in an empty namespace:

     ;;;---------------------------------------- file: ~~lib/gambit#.scm
     (##namespace ("" define if quote set! cons car cdr + - * / ;; etc

   Using the `gambit' module and the `##import' form, the previous
example can be rewritten like this:

     ;;;---------------------------------------- file: angle3/angle3#.scm
     (##namespace ("angle3#" deg->rad rad->deg))

     ;;;---------------------------------------- file: angle3/angle3.scm
     (##namespace ("angle3#")) ;; map all identifiers to angle3# namespace
     (##import gambit)         ;; except those defined by Gambit's RTS
     (##supply-module angle3)
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (* x factor))
     (define (rad->deg x) (/ x factor))

     ;;;---------------------------------------- file: app3.scm
     (##import angle3)
     (println "90 degrees is " (deg->rad 90) " radians")

   In this example the `(##import angle3)' takes care of the namespace
mapping and the loading of `angle3.scm' because it is equivalent to:

     (begin
      (##include "angle3/angle3#.scm")
      (##demand-module angle3))

7.2.3 Macros
------------

In addition to procedures, a module `M' may export macros.  The file
`M#.scm' is the designated place to put exported macro definitions.
These macro definitions will essentially be copied at the point where
the `##import' of the module is done.  Macros that are needed strictly
for the implementation of a module may be defined in the file `M.scm'
and these macro definitions will not be visible elsewhere.  Note that
the macros defined with `define-macro' are not hygienic, so the macro
definition writer should take care to explicitly indicate what
identifiers resolve to using fully qualified identifiers (i.e.
containing a `#' sign).

   To explain these issues, lets extend our example module in the
following ways.  First we want the module to export the macros `sind'
and `asind' that are like the `sin' and `asin' procedures but use
degrees instead of radians.  Note that it would be a better design for
`sind' and `asind' to be procedures, but we'll implement them as macros
for the sake of the example.  Second we want the procedures `deg->rad'
and `rad->deg' to check that their argument is a real number using a
`check-real' macro.

   In a setting where name clashes are not an issue these macros can be
defined as follows:

     (define-macro (sind x) `(sin (deg->rad ,x)))
     (define-macro (asind x) `(rad->deg (asin ,x)))
     (define-macro (check-real x y)
       `(if (real? ,x) ,y (error "not real!")))

   Name clashes will occur when the locations where these macros are
called are in the scope of new bindings for `sin', `deg->rad', `if',
`error', etc which are identifiers used in the expanded code.  A name
clash can also happen for the name `define-macro' itself.  To remove
the possibility of name clashes the `##namespace' form and fully
qualified identifiers can be used.  All the Gambit special forms, such
as `let', `if', and `define-macro', have a fully qualified version
(`##let', `##if', and `##define-macro'). Gambit predefined procedures,
such as `sin', `real?', and `error', don't necessarily have a fully
qualified version (some do and some don't) but an empty namespace
definition in a `##let' form can be used to avoid the clash, i.e.
`(##let () (##namespace ("") sin))' refers to the global variable `sin'
whatever scope it is in.  With these forms our example can be written
like this:

     ;;;---------------------------------------- file: angle4/angle4#.scm
     (##namespace ("angle4#" deg->rad rad->deg))
     (##define-macro (sind x) `((##let () (##namespace ("")) sin)
                                (angle4#deg->rad ,x)))
     (##define-macro (asind x) `(angle4#rad->deg
                                 ((##let () (##namespace ("")) asin) ,x)))

     ;;;---------------------------------------- file: angle4/angle4.scm
     (##namespace ("angle4#")) ;; map all identifiers to angle4# namespace
     (##import gambit)         ;; except those defined by Gambit's RTS
     (##supply-module angle4)
     (##define-macro (check-real x y)
       `(##if ((##let () (##namespace ("")) real?) ,x)
              ,y
              ((##let () (##namespace ("")) error) "not real!")))
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (check-real x (* x factor)))
     (define (rad->deg x) (check-real x (/ x factor)))

     ;;;---------------------------------------- file: app4.scm
     (##import angle4)
     (println "90 degrees is " (deg->rad 90) " radians")
     (println "sind(90) is " (sind 90))

7.3 Primitive Procedures
========================

Identifiers with a `##' prefix are not valid identifiers according to
RnRS. This means that code containing `##' prefixed identifiers cannot
be processed by and shared with other Scheme implementations.  They are
hard to read by people that aren't used to that extension.  Moreover
the code lacks abstraction and safety because using `##car' rather than
`car' has a specific meaning: avoiding type checks. Consequently it is
hard to "turn on" safe execution of the code when it needs to be
debugged. Many parts of the runtime library are expressed at a low
level of abstraction (with `##' prefixed identifiers) even when not
required.

   For those reasons `##' prefixed identifiers should be used sparingly
in new code, and existing code should gradually be rewritten to avoid
them. The primitive operations which are used to build higher-level
operations are all defined as procedures with a `##' prefix.

   The file `~~lib/_gambit#.scm' contains the definition of the
`primitive' macro whose purpose is to abstract from the `##' prefix.
The call `(primitive foo)' is equivalent to `##foo' and `(primitive
(foo a b))' is equivalent to `(##foo a b)'.  The file
`~~lib/_gambit#.scm' also contains the definition of the `standard'
macro whose purpose is similar, but forces the use of the empty
namespace.  The call `(standard +)' is equivalent to `(##let ()
(##namespace ("" +)) +)' and `(standard (+ a b))' is equivalent to
`((##let () (##namespace ("" +)) +) a b)'.  Code that uses the
`primitive' and `standard' macros can be ported to other Scheme
implementations by defining implementation specific `primitive' and
`standard' macros that implement the appropriate mapping for that
implementation.

   The file `~~lib/_gambit#.scm' also contains definitions for the
`define-procedure' and `define-primitive' macros.  The `primitive' and
`standard' macros work in tandem with the `define-procedure' and
`define-primitive' macros and the `~~lib/gambit/prim/prim#.scm' file
and `(gambit prim)' library. The file `~~lib/gambit/prim/prim#.scm'
contains namespace declarations that map operations exported by the
runtime library without a `##' prefix to their `##' prefixed names if
this preserves the meaning of the operation but possibly (and usually)
with no type checking. The `(gambit prim)' library is similar but in
the form of a R7RS library. For example the following code:

     (include "~~lib/gambit/prim/prim#.scm")
     (define (foo x) (square (car x)))
     (println (foo (bar 0.5)))
     (pp "hello")

   is equivalent to this code:

     (##define (foo x) (##square (##car x)))
     (##println (foo (bar 0.5)))
     (##unimplemented#pp "hello")

   The namespace declarations in `~~lib/gambit/prim/prim#.scm' have
caused a mapping of `square' to `##square', `car' to `##car' and
`println' to `##println' because those primitives perform the same
operations (when the code has no errors). Note that `foo' and `bar'
have remained the same, because they are not procedures exported by the
runtime library, and `pp' has been mapped to `##unimplemented#pp'
because `pp' is a procedure exported by the runtime library but `##pp'
is not defined. Having `unimplemented' in the name helps catch
situations where the programmer expected a primitive operation to exist
but this isn't the case.

   The `define-procedure' macro does two things. It supports type
annotations in the parameter list and it inserts a `(include
"~~lib/gambit/prim/prim#.scm")' in the body so that primitive operations
can be used without the `##' prefix. Type checking and automatic
forcing of promise arguments are also added implicitly. The macro
`define-primitive' is similar, but the procedure defined is implicitly
prefixed with `##'.

   So all of these things work together to abstract away from the
concept of _primitive_ operations. _Primitives_ are implemented using
procedures with a `##' prefix, but other Scheme implementations could
do it differently.

   Finally, there's the `(declare-safe-define-procedure <bool>)' macro
that can be used to enable/disable the mapping of names exported by the
runtime library to the corresponding primitives. This is useful to
enable type checks in the code. For example the following definition:

     (define-procedure (foo (x vector))
       (vector-ref x 5))

   which expands to

     (define (foo x)
       (macro-check-vector x '(1 . x) (foo x)
         (##vector-ref x 5)))

   which expands to

     (define (foo x)
         (if (##vector? x)
             (##vector-ref x 5)
             (##fail-check-vector '(1 . x) foo x)))

   If the code is in the scope of a `(declare-safe-define-procedure
#t)' then it is `vector-ref' that is called instead of `##vector-ref'
which will both check that X is a vector (redundantly) and that the
index is in range. However, the use of `##vector-ref' can be forced by
writing the code with an explicit use of the `primitive' macro:

     (define-procedure (foo (x vector))
       (primitive (vector-ref x 5)))

   The expectation is that the `primitive' special form will be used
sparingly. Searching the source code for the pattern `"(primitive"' is
a good way to find potentially unsafe code.

7.3.1 Type specifiers
---------------------

Here is a list of the available type specifiers for a
`define-procedure' parameter X and the associated constraint on the
value of X.

   Note that there is no direct way for checking for a "list" or "list
of elements of type T". A procedure taking a list parameter will likely
iterate on the list's pairs going from cdr to cdr until a non-pair is
found. Then a check for the empty list with
`(macro-check-proper-list-null lst <arg-id> (<procedure-name>
<args>...) <body>)' will check that the parameter is a proper list
(i.e. that it ends with the empty list).

7.3.1.1 Basic types (other than numbers)
........................................

`boolean'
     X is a boolean

`char'
     X is a character

`pair'
     X is a pair

`procedure'
     X is a procedure

`string'
     X is a string

`symbol'
     X is a symbol

`vector'
     X is a vector

7.3.1.2 Numbers
...............

`number'
     X is a number (possibly complex, rational, etc)

`real'
     X is a real number (any number except complex)

`fixnum'
     X is a fixnum and -2^(W-3) <= X <= 2^(W-3) - 1

`(fixnum-range lo hi)'
     X is a fixnum and lo <= X < hi

`(fixnum-range-incl lo hi)'
     X is a fixnum and lo <= X <= hi

`index'
     X is a fixnum and 0 <= X

`(index-range lo hi)'
     X is a fixnum and 0 <= lo <= X < hi

`(index-range-incl lo hi)'
     X is a fixnum and 0 <= lo <= X <= hi

`exact-signed-int8'
     X is an exact integer, -128 <= X <= 127

`exact-signed-int16'
     X is an exact integer n, -32768 <= X <= 32767

`exact-signed-int32'
     X is an exact integer n, -2^31 <= X <= 2^31 - 1

`exact-signed-int64'
     X is an exact integer n, -2^63 <= X <= 2^63 - 1

`exact-unsigned-int8'
     X is an exact integer n, 0 <= X <= 255

`exact-unsigned-int16'
     X is an exact integer n, 0 <= X <= 65535

`exact-unsigned-int32'
     X is an exact integer n, 0 <= X <= 2^32 - 1

`exact-unsigned-int64'
     X is an exact integer n, 0 <= X <= 2^64 - 1

`flonum'
     X is a flonum, exception mentions FLONUM

`inexact-real'
     X is a flonum, exception mentions Inexact REAL

`inexact-real-list'
     X is a flonum, exception mentions Inexact REAL LIST

7.3.1.3 Time types
..................

`time'
     X is a time object

`absrel-time'
     X is a real or a time object

`absrel-time-or-false'
     X is `#f' or a real or a time object

7.3.1.4 Ports
.............

`port'
     X is a port (input, output, or input-output)

`input-port'
     X is an input port

`output-port'
     X is an output port

`object-input-port'
     X is an object input port

`object-output-port'
     X is an object output port

`vector-input-port'
     X is a vector input port

`vector-output-port'
     X is a vector output port

`character-input-port'
     X is a character input port

`character-output-port'
     X is a character output port

`string-input-port'
     X is a string input port

`string-output-port'
     X is a string output port

`byte-port'
     X is a byte port (input, output, or input-output)

`byte-input-port'
     X is a byte input port

`byte-output-port'
     X is a byte output port

`u8vector-input-port'
     X is a u8vector input port u8vector-output-port       x is a
     u8vector output port

`device-input-port'
     X is a device intput port

`device-output-port'
     X is a device output port

`process-port'
     X is a process port

`tcp-client-port'
     X is a tcp-client port

`tcp-server-port'
     X is a tcp-server port

`udp-port'
     X is a udp port

`udp-input-port'
     X is a udp input port

`udp-output-port'
     X is a udp output port

`tty-port'
     X is a tty port

7.3.1.5 List and vector variants of above
.........................................

`list'
     no type checking (a non-null non-pair object is in fact a
     degenerate dotted list), exception mentions LIST

`proper-list'
     no type checking (code traversing the list must check for a
     proper-list), exception mentions PROPER LIST

`proper-list-null'
     X is the empty list, exception mentions PROPER LIST

`proper-or-circular-list'
     no type checking (code traversing the list must check for a
     proper-list or circular-list), exception mentions PROPER or
     CIRCULAR LIST

`proper-or-circular-list-null'
     X is the empty list, exception mentions PROPER LIST

`char-list'
     X is a character, exception mentions CHARACTER LIST

`char-vector'
     X is a character, exception mentions CHARACTER VECTOR

`pair-list'
     X is a pair, exception mentions PAIR LIST

`exact-unsigned-int8-list-exact-unsigned-int8'
     X is an exact-unsigned-int8, exception mentions INTEGER LIST

`exact-unsigned-int16-list-exact-unsigned-int16'
     X is an exact-unsigned-int16, exception mentions INTEGER LIST

`exact-unsigned-int32-list-exact-unsigned-int32'
     X is an exact-unsigned-int32, exception mentions INTEGER LIST

`exact-unsigned-int64-list-exact-unsigned-int64'
     X is an exact-unsigned-int64, exception mentions INTEGER LIST

`exact-signed-int8-list-exact-signed-int8'
     X is an exact-signed-int8, exception mentions INTEGER LIST

`exact-signed-int16-list-exact-signed-int16'
     X is an exact-signed-int16, exception mentions INTEGER LIST

`exact-signed-int32-list-exact-signed-int32'
     X is an exact-signed-int32, exception mentions INTEGER LIST

`exact-signed-int64-list-exact-signed-int64'
     X is an exact-signed-int64, exception mentions INTEGER LIST

7.3.1.6 Gambit types
....................

`error-exception'
     X is an error-exception object

`box'
     X is a box

`condvar'
     X is a condition variable

`f32vector'
     X is a f32vector

`f64vector'
     X is a f64vector

`foreign'
     X is a foreign object

`keyword'
     X is a keyword

`mutex'
     X is a mutex

`processor'
     X is a processor object

`s16vector'
     X is a s16vector

`s32vector'
     X is a s32vector

`s64vector'
     X is a s64vector

`s8vector'
     X is a s8vector

`table'
     X is a table

`tgroup'
     X is a thread group

`thread'
     X is a thread

`u16vector'
     X is a u16vector

`u32vector'
     X is a u32vector

`u64vector'
     X is a u64vector

`u8vector'
     X is a u8vector

`will'
     X is a will

`continuation'
     X is a continuation object

`random-source'
     X is a random-source object

`readtable'
     X is a readtable

`type'
     X is a structure type descriptor

`mutable'
     X is a mutable object

7.3.1.7 Others
..............

`initialized-thread'

`not-initialized-thread'

`not-started-thread'

`not-started-thread-given-initialized'

`string-or-ip-address'

`string-or-nonnegative-fixnum'

7.4 R7RS Compatible Modules
===========================

The R7RS Scheme standard specifies a modularization approach based on
the concept of library.  A library is defined using the
`define-library' form.  This form is implemented as a macro that
expands into the constructs used by primitive modules, in particular a
`##namespace' declaration with a namespace derived from the library's
name so that all variables defined by the library are in that
namespace.  With the `define-library' form the `angle3' module example
given previously can be written like this:

     ;;;---------------------------------------- file: angle3.sld
     (define-library (angle3)

       (export deg->rad rad->deg)

       (import (scheme base)
               (scheme inexact))

       (begin
         (define factor (/ (atan 1) 45))
         (define (deg->rad x) (* x factor))
         (define (rad->deg x) (/ x factor))))

   For this library the expansion of the `define-library' form will
contain a `##namespace' declaration that causes the definition of the
global variables `angle3#factor', `angle3#deg->rad', and
`angle3#rad->deg'.  Meanwhile an `(import (angle3))' in another library
will generate a `##namespace' declaration that maps uses of `deg->rad'
and `rad->deg' to the global variables `angle3#deg->rad' and
`angle3#rad->deg' respectively (note that the unexported global
variable `factor' is not included in the generated `##namespace'
declaration).

   For more complex libraries whose code is split into multiple files it
is convenient to put all the files in a dedicated subdirectory.  This
is the preferred filesystem structure for a library but the runtime
system supports both styles.  The previous module could be structured
like this instead:

     ;;;---------------------------------------- file: angle3/angle3.sld
     (define-library (angle3)

       (export deg->rad rad->deg)

       (import (scheme base)
               (scheme inexact))

       (include "angle3.scm")) ;; path is relative to angle3.sld file

     ;;;---------------------------------------- file: angle3/angle3.scm
     (define factor (/ (atan 1) 45))
     (define (deg->rad x) (* x factor))
     (define (rad->deg x) (/ x factor))

7.4.1 Identifying libraries
---------------------------

Each library is given a name so that it can be referred to in various
contexts, most notably in `import' forms and the interpreter's and
compiler's command line.  The R7RS defines a library name as a list
whose members are identifiers and exact non-negative integers, for
example `(widget)', `(_hamt)', `(scheme base)', and `(srfi 64)'.

   The system maps these R7RS library names to module identifiers that
are symbols formed by concatenating the parts of the library name
separated with `/'.  The library name and module name are
interchangeable.  Consequently, `(import srfi/64)' and `(import _hamt)'
are respectively equivalent to `(import (srfi 64))' and `(import
(_hamt))'.  Using the module name to identify libraries on the command
line is convenient as it avoids having to escape parentheses and spaces.

7.4.2 The `define-library' form
-------------------------------

 -- special form: define-library NAME DECLARATION ...

   In a library definition NAME specifies the name of the library and
DECLARATION is one of:

     (export <EXPORT SPEC> ...)
     (import <IMPORT SET> ...)
     (begin <COMMAND OR DEFINITION> ...)
     (include <FILENAME> ...)
     (include-ci <FILENAME> ...)
     (include-library-declarations <FILENAME> ...)
     (cond-expand <COND EXPAND FEATURES> ...)
     (namespace <NAMESPACE>)
     (cc-options <OPTIONS> ...)
     (ld-options <OPTIONS> ...)
     (ld-options-prelude <OPTIONS> ...)
     (pkg-config <OPTIONS> ...)
     (pkg-config-path <PATH> ...)

7.4.3 `(export <EXPORT SPEC> ...)'
----------------------------------

An `export' declaration specifies a list of identifiers which can be
made visible to other libraries or programs.  An <EXPORT SPEC> takes
one of the following forms:

     <IDENTIFIER>
     (rename <IDENTIFIER>1 <IDENTIFIER>2)

   In an <EXPORT SPEC>, an <IDENTIFIER> names a single binding
(variable or macro) defined within or imported into the library, where
the external name for the export is the same as the name of the binding
within the library. A `rename' spec exports the binding defined within
or imported into the library and named by <IDENTIFIER>1 in each
`(<IDENTIFIER>1 <IDENTIFIER>2)' pairing, using <IDENTIFIER>2 as the
external name.

7.4.4 `(import <IMPORT SET> ...)'
---------------------------------

A library declares a dependency to another library with the `import'
declaration.  The `(import <IMPORT SET> ...)'  form identifies the
imported library or libraries.

   Each <IMPORT SET> names a set of bindings from a library and
possibly specifies local names for the imported bindings. An <IMPORT
SET> takes one of the following forms:

     <LIBRARY NAME>
     (only <IMPORT SET> <IDENTIFIER> ...)
     (except <IMPORT SET> <IDENTIFIER> ...)
     (prefix <IMPORT SET> <IDENTIFIER>)
     (rename <IMPORT SET> (<IDENTIFIER>1 <IDENTIFIER>2) ...)

   In the first form, all of the identifiers in the named library's
`export' clauses are imported with the same names (or the exported
names if exported with `rename').  The additional <IMPORT SET> forms
modify this set as follows:

   * `only' produces a subset of the given <IMPORT SET> including only
     the listed identifiers (after any renaming).  It is an error if
     any of the listed identifiers are not found in the original set.

   * `except' produces a subset of the given <IMPORT SET>, excluding
     the listed identifiers (after any renaming). It is an error if any
     of the listed identifiers are not found in the original set.

   * `rename' modifies the given <IMPORT SET>, replacing each instance
     of <IDENTIFIER>1 with <IDENTIFIER>2. It is an error if any of the
     listed <IDENTIFIER>1s are not found in the original set.

   * `prefix' automatically renames all identifiers in the given
     <IMPORT SET>, prefixing each with the specified <IDENTIFIER>.


   It is an error to import the same identifier more than once with
different bindings, or to redefine or mutate an imported binding with a
definition or with `set!', or to refer to an identifier before it is
imported.

   As an extension to the R7RS syntax it is allowed for a <LIBRARY
NAME> to contain a trailing `@VERSION' when the library is hosted in a
`git' repository.  The VERSION must match a tag of that repository and
it indicates the specific library version required.  For example,
`(import (github.com/gambit hello @1.0))' or equivalently `(import
github.com/gambit/hello@1.0)'.  Note that the version specifier is not
separated with a `/' in the module name.

   Another extension to the R7RS syntax when the library is hosted in a
`git' repository is the use of dots before the name of the library to
indicate a relative reference within the repository.  The number of
dots indicates the number of parent hops.  For example, in the library
`(github.com/gambit hello demo)' an `(import (.. hello))' will resolve
to the `(github.com/gambit hello)' library.  A relative library
reference should not contain an explicit `@VERSION' because the VERSION
is implicitly the same as the referring module.

7.4.5 `(begin <COMMAND OR DEFINITION> ...)', `(include <FILENAME> ...)', and `(include-ci <FILENAME> ...)'
----------------------------------------------------------------------------------------------------------

The `begin', `include', and `include-ci' declarations are used to
specify the body of the library.  They have the same syntax and
semantics as the corresponding expression types.  This form of `begin'
is analogous to, but not the same as, the two types of `begin'
expressions.

7.4.6 `(include-library-declarations <FILENAME> ...)'
-----------------------------------------------------

The `include-library-declarations' declaration is similar to `include'
except that the contents of the file are spliced directly into the
current library definition.  This can be used, for example, to share the
same `export' declaration among multiple libraries as a simple form of
library interface.

7.4.7 `(cond-expand <COND EXPAND FEATURES> ...)'
------------------------------------------------

The `cond-expand' declaration has the same syntax and semantics as the
`cond-expand' expression type, except that it expands to spliced-in
library declarations rather than expressions enclosed in `begin'.

7.4.8 Extensions to the R7RS library declarations
-------------------------------------------------

The `(namespace <NAMESPACE>)' declaration allows overriding the
namespace used for the library.  This is mainly useful for system
libraries to prevent namespace prefixing using a `(namespace "")'
declaration.

   The remaining declarations are relevant to the `C' target and
ignored otherwise.  They provide information, in the form of strings,
to be passed to the compiler options of the same name when this library
is compiled:

   * `(cc-options <OPTIONS> ...)'

   * `(ld-options <OPTIONS> ...)'

   * `(ld-options-prelude <OPTIONS> ...)'

   * `(pkg-config <OPTIONS> ...)'

   * `(pkg-config-path <PATH> ...)'


   For example, a library could force the C compiler to generate
machine code for i386 with:

     (define-library (foo)
       (export bar)
       (import (scheme base))
       (cc-options "-march=i386") ;; request compilation for i386
       (begin (define (bar) 42)))

7.5 Installing Modules
======================

When a module is imported, the processing of the `import' form must
locate and read the source code of the module at macro expansion time
to determine which names are exported and to what they are mapped.  The
list of module search directories (`~~lib' followed by `~~userlib' by
default) is searched to find the module's source code.  At execution
time the same search algorithm is used to locate and load the module,
either in source code form or compiled form.  The `~~lib' directory is
where the system's builtin modules are put when Gambit is installed.
The `~~userlib' directory is a convenient place where other modules can
be installed by the user because locating them does not require
extending the list of module search directories.

 -- procedure: module-search-order-reset!
 -- procedure: module-search-order-add! DIR
     The list of module search directories can be modified using the
     procedures `module-search-order-reset!' and
     `module-search-order-add!' that respectively clear the list and
     extend the list with the directory DIR which must be a string. The
     list can also be extended by using the `-:search='DIR runtime
     option or by a command line argument to `gsi' and `gsc' that ends
     with a path separator or a `.' (*note ##demand-module and
     ##supply-module forms::).

     For example:

          $ cat foobar.sld
          (define-library (foobar)
            (import (scheme write))
            (begin (display "foobar library has executed\n")))
          $ gsi
          Gambit v4.9.4

          > (import (foobar))
          *** ERROR IN (stdin)@1.9 -- Cannot find library (foobar)
          > (module-search-order-add! ".")
          > (import (foobar))
          foobar library has executed
          > ,q
          $ gsi -:search=. foobar
          foobar library has executed
          $ gsi . foobar
          foobar library has executed


 -- procedure: module-whitelist-reset!
 -- procedure: module-whitelist-add! SOURCE
     When modules are installed it is done at the granularity of a
     "package", which is defined as a `git' repository possibly
     containing more than one module.  For example, if the hosted module
     `github.com/gambit/hello/demo' needs to be installed it is all of
     the code at `github.com/gambit/hello' that is installed (this
     includes the three modules `github.com/gambit/hello',
     `github.com/gambit/hello/demo', and
     `github.com/gambit/hello/test').

     For convenience the runtime system will automatically install in
     the `~~userlib' directory any hosted module that is from a trusted
     source.  The whitelist of trusted sources, which initially contains
     only `github.com/gambit', can be modified using the procedures
     `module-whitelist-reset!' and `module-whitelist-add!' that
     respectively clear the list and extend the list with the source
     SOURCE which must be a string. The list can also be extended with
     the `-:whitelist='SOURCE runtime option.

     For example:

          $ gsi github.com/gambit/hello/demo  # auto-install of github.com/gambit/hello package
          People customarily greet each other when they meet.
          In English you can say: hello Bob, nice to see you!
          In French you can say: bonjour Bob, je suis enchant!
          Demo source code: /Users/feeley/.gambit_userlib/github.com/gambit/hello/@/demo.scm
          $ gsi github.com/feeley/roman/demo  # no auto-install because not on whitelist
          *** ERROR IN ##main -- No such file or directory
          (load "github.com/feeley/roman/demo")
          $ gsi -:whitelist=github.com/feeley github.com/feeley/roman/demo
          1 is I in roman numerals
          2 is II in roman numerals
          4 is IV in roman numerals
          8 is VIII in roman numerals
          16 is XVI in roman numerals
          32 is XXXII in roman numerals
          64 is LXIV in roman numerals
          $ gsi github.com/feeley/roman/demo   # OK because module is now installed
          1 is I in roman numerals
          2 is II in roman numerals
          4 is IV in roman numerals
          8 is VIII in roman numerals
          16 is XVI in roman numerals
          32 is XXXII in roman numerals
          64 is LXIV in roman numerals
          $ gsi github.com/feeley/roman/test   # the test module was also installed
          *** all tests passed out of a total of 19 tests

     The use of the runtime option `-:whitelist=' (with no SOURCE) will
     disable the automatic installation of modules, even from
     `github.com/gambit'.  For example:

          $ gsi -:whitelist= github.com/gambit/hello/demo
          *** ERROR IN ##main -- No such file or directory
          (load "github.com/gambit/hello/demo")

     A manual management of packages is nevertheless possible with the
     `gsi' package management operations.  These are invoked with the
     command line options `-install', `-uninstall', and `-update' which
     respectively install, uninstall and update packages.  The package
     management operations accept a list of packages.  Packages are
     installed in `~~userlib' which is mapped to `~/.gambit_userlib' by
     default.  An optional `-dir DIR' option can be used to install the
     package in some other directory.

     For example:

          $ gsi -install github.com/feeley/roman
          installing github.com/feeley/roman to /Users/feeley/.gambit_userlib/
          $ gsi github.com/feeley/roman/demo
          1 is I in roman numerals
          2 is II in roman numerals
          4 is IV in roman numerals
          8 is VIII in roman numerals
          16 is XVI in roman numerals
          32 is XXXII in roman numerals
          64 is LXIV in roman numerals
          $ gsi -uninstall github.com/feeley/roman
          uninstalling github.com/feeley/roman from /Users/feeley/.gambit_userlib/
          $ gsi -install -dir ~/mylibs github.com/feeley/roman
          installing github.com/feeley/roman to /Users/feeley/mylibs
          $ gsi ~/mylibs/ github.com/feeley/roman/demo
          1 is I in roman numerals
          2 is II in roman numerals
          4 is IV in roman numerals
          8 is VIII in roman numerals
          16 is XVI in roman numerals
          32 is XXXII in roman numerals
          64 is LXIV in roman numerals

     Local `git' repositories can also be installed manually with the
     package management operations using a path to the local repository.
     This can be useful during the development phase before a library
     becomes hosted.

     For example:

          $ mkdir some
          $ mkdir some/dir
          $ mkdir some/dir/mylib
          $ cd some/dir/mylib
          $ cat > mylib.sld
          (define-library (mylib)
            (import (scheme write))
            (begin (display "mylib library has executed\n")))
          $ git init
          Initialized empty Git repository in /Users/feeley/doc/some/dir/mylib/.git/
          $ git add mylib.sld
          $ git commit -m "first commit"
          [master (root-commit) c3f6aff] first commit
           1 file changed, 3 insertions(+)
           create mode 100644 mylib.sld
          $ cd ../../..
          $ gsi some/dir/ mylib   # execution of mylib without installation
          mylib library has executed
          $ gsi -install some/dir/mylib
          installing some/dir/mylib to /Users/feeley/.gambit_userlib/
          $ gsi mylib   # execution of mylib after installation
          mylib library has executed


7.6 Compiling Modules
=====================

When `gsc' finds a command line argument that is the name of a module
found on the list of module search order directories (after an
automatic installation if that is appropriate) that module's main
source code file will be compiled.

   When a dynamic compilation is requested (which is the default
compilation mode and when the command line option `-dynamic' is used)
the compiler will compile for the selected target the main source code
file to a target file and a dynamically loadable object file with a
`.o1' extension.  These files will be created in a directory next to
the module's main source code file, with the same name stripped of it's
extension and suffixed with the Gambit version and the target name.
This naming strategy aims to avoid loading the compiled file in an
inappropriate context.  The module loading algorithm knows it should
check this directory to find a compiled version of the module.

   For example:

     $ mkdir lib1 lib2
     $ cat > lib1/lib1.sld
     (define-library (lib1)
       (export fact)
       (import (scheme base) (scheme write))
       (begin
         (define (fact n) (if (<= n 1) 1 (* n (fact (- n 1)))))
         (display "lib1 loaded\n")))
     $ cat > lib2/lib2.sld
     (define-library (lib2)
       (import (lib1) (scheme base) (scheme write))
       (begin
         (display
          (cond-expand
            ((compilation-target C)   "lib2 compiled to C\n")
            ((compilation-target (_)) "lib2 interpreted\n")
            (else                     "lib2 compiled to other\n")))
         (display (fact 10))
         (newline)))
     $ gsi . lib2        # loads lib1.sld and lib2.sld
     lib1 loaded
     lib2 interpreted
     3628800
     lib1
     $ tree -charset=ascii -noreport lib1 lib2
     `-- lib1.sld
     lib2
     `-- lib2.sld
     $ gsc . lib1 lib2   # compile lib1.sld and lib2.sld using C target
     $ gsi . lib2        # loads generated lib1.o1 and lib2.o1
     lib1 loaded
     lib2 compiled to C
     3628800
     $ gsc -target js . lib1 lib2   # also compile them for js target
     $ tree -charset=ascii -noreport lib1 lib2
     lib1
     |-- lib1.sld
     |-- lib1@gambit409003@C
     |   |-- lib1.c
     |   `-- lib1.o1
     `-- lib1@gambit409003@js
         |-- lib1.js
         `-- lib1.o1
     lib2
     |-- lib2.sld
     |-- lib2@gambit409003@C
     |   |-- lib2.c
     |   `-- lib2.o1
     `-- lib2@gambit409003@js
         |-- lib2.js
         `-- lib2.o1

   To create an executable program from a set of non-legacy modules it
is important to use the `-nopreload' linking option when linking so
that the modules will be initialized in an order that is consistent
with the module dependencies.  If the default `-preload' linking option
is used some modules may be initialized out of order, leading to
incorrect execution.

   Here is an example that extends the previous example:

     $ gsc -exe -nopreload . lib1/lib1.sld lib2/lib2.sld
     $ lib2/lib2
     lib1 loaded
     lib2 compiled to C
     3628800
     $ gsc -target js -exe -nopreload . lib1/lib1.sld lib2/lib2.sld
     $ lib2/lib2
     lib1 loaded
     lib2 compiled to other
     3628800

8 Built-in data types
*********************

8.1 Numbers
===========

8.1.1 Extensions to numeric procedures
--------------------------------------

 -- procedure: = Z1...
 -- procedure: < X1...
 -- procedure: > X1...
 -- procedure: <= X1...
 -- procedure: >= X1...
     These procedures take any number of arguments including no
     argument.  This is useful to test if the elements of a list are
     sorted in a particular order.  For example, testing that the list
     of numbers `lst' is sorted in nondecreasing order can be done with
     the call `(apply < lst)'.


8.1.2 IEEE floating point arithmetic
------------------------------------

To better conform to IEEE floating point arithmetic the standard
numeric tower is extended with these special inexact reals:

`+inf.0'
     positive infinity

`-inf.0'
     negative infinity

`+nan.0'
     "not a number"

`-0.'
     negative zero (`0.' is the positive zero)

   The infinities and "not a number" are reals (i.e. `(real?  +inf.0)'
is `#t') but are not rational (i.e. `(rational?  +inf.0)' is `#f').

   Both zeros are numerically equal (i.e. `(= -0. 0.)' is `#t') but are
not equivalent (i.e. `(eqv? -0. 0.)' and `(equal?  -0. 0.)' are `#f').
All numerical comparisons with "not a number", including `(= +nan.0
+nan.0)', are `#f'.

8.1.3 Integer square root and nth root
--------------------------------------

 -- procedure: integer-sqrt N
     This procedure returns the integer part of the square root of the
     nonnegative exact integer N.

     For example:

          > (integer-sqrt 123)
          11


 -- procedure: integer-nth-root N1 N2
     This procedure returns the integer part of N1 raised to the power
     1/N2, where N1 is a nonnegative exact integer and N2 is a positive
     exact integer.

     For example:

          > (integer-nth-root 100 3)
          4


8.1.4 Bitwise-operations on exact integers
------------------------------------------

The procedures defined in this section are compatible with the
withdrawn "Integer Bitwise-operation Library SRFI" (SRFI 33).  Note
that some of the procedures specified in SRFI 33 are not provided.

   Most procedures in this section are specified in terms of the binary
representation of exact integers.  The two's complement representation
is assumed where an integer is composed of an infinite number of bits.
The upper section of an integer (the most significant bits) are either
an infinite sequence of ones when the integer is negative, or they are
an infinite sequence of zeros when the integer is nonnegative.

 -- procedure: arithmetic-shift N1 N2
     This procedure returns N1 shifted to the left by N2 bits, that is
     `(floor (* N1 (expt 2 N2)))'.  Both N1 and N2 must be exact
     integers.

     For example:

          > (arithmetic-shift 1000 7)  ; n1=...0000001111101000
          128000
          > (arithmetic-shift 1000 -6) ; n1=...0000001111101000
          15
          > (arithmetic-shift -23 -3)  ; n1=...1111111111101001
          -3


 -- procedure: bitwise-merge N1 N2 N3
     This procedure returns an exact integer whose bits combine the bits
     from N2 and N3 depending on N1.  The bit at index I of the result
     depends only on the bits at index I in N1, N2 and N3: it is equal
     to the bit in N2 when the bit in N1 is 0 and it is equal to the
     bit in N3 when the bit in N1 is 1.  All arguments must be exact
     integers.

     For example:

          > (bitwise-merge -4 -11 10) ; ...11111100 ...11110101 ...00001010
          9
          > (bitwise-merge 12 -11 10) ; ...00001100 ...11110101 ...00001010
          -7


 -- procedure: bitwise-and N...
     This procedure returns the bitwise "and" of the exact integers
     N....  The value -1 is returned when there are no arguments.

     For example:

          > (bitwise-and 6 12)  ; ...00000110 ...00001100
          4
          > (bitwise-and 6 -4)  ; ...00000110 ...11111100
          4
          > (bitwise-and -6 -4) ; ...11111010 ...11111100
          -8
          > (bitwise-and)
          -1


 -- procedure: bitwise-andc1 N1 N2
     This procedure returns the bitwise "and" of the bitwise complement
     of the exact integer N1 and the exact integer N2.

     For example:

          > (bitwise-andc1 11 26)  ; ...00001011 ...00011010
          16
          > (bitwise-andc1 -12 26) ; ...11110100 ...00011010
          10


 -- procedure: bitwise-andc2 N1 N2
     This procedure returns the bitwise "and" of the exact integer N1
     and the bitwise complement of the exact integer N2.

     For example:

          > (bitwise-andc2 11 26)  ; ...00001011 ...00011010
          1
          > (bitwise-andc2 11 -27) ; ...00001011 ...11100101
          10


 -- procedure: bitwise-eqv N...
     This procedure returns the bitwise complement of the bitwise
     "exclusive-or" of the exact integers N....  The value -1 is
     returned when there are no arguments.

     For example:

          > (bitwise-eqv 6 12)  ; ...00000110 ...00001100
          -11
          > (bitwise-eqv 6 -4)  ; ...00000110 ...11111100
          5
          > (bitwise-eqv -6 -4) ; ...11111010 ...11111100
          -7
          > (bitwise-eqv)
          -1


 -- procedure: bitwise-ior N...
     This procedure returns the bitwise "inclusive-or" of the exact
     integers N....  The value 0 is returned when there are no
     arguments.

     For example:

          > (bitwise-ior 6 12)  ; ...00000110 ...00001100
          14
          > (bitwise-ior 6 -4)  ; ...00000110 ...11111100
          -2
          > (bitwise-ior -6 -4) ; ...11111010 ...11111100
          -2
          > (bitwise-ior)
          0


 -- procedure: bitwise-nand N1 N2
     This procedure returns the bitwise complement of the bitwise "and"
     of the exact integer N1 and the exact integer N2.

     For example:

          > (bitwise-nand 11 26)  ; ...00001011 ...00011010
          -11
          > (bitwise-nand 11 -27) ; ...00001011 ...11100101
          -2


 -- procedure: bitwise-nor N1 N2
     This procedure returns the bitwise complement of the bitwise
     "inclusive-or" of the exact integer N1 and the exact integer N2.

     For example:

          > (bitwise-nor 11 26)  ; ...00001011 ...00011010
          -28
          > (bitwise-nor 11 -27) ; ...00001011 ...11100101
          16


 -- procedure: bitwise-not N
     This procedure returns the bitwise complement of the exact integer
     N.

     For example:

          > (bitwise-not 3)  ; ...00000011
          -4
          > (bitwise-not -1) ; ...11111111
          0


 -- procedure: bitwise-orc1 N1 N2
     This procedure returns the bitwise "inclusive-or" of the bitwise
     complement of the exact integer N1 and the exact integer N2.

     For example:

          > (bitwise-orc1 11 26)  ; ...00001011 ...00011010
          -2
          > (bitwise-orc1 -12 26) ; ...11110100 ...00011010
          27


 -- procedure: bitwise-orc2 N1 N2
     This procedure returns the bitwise "inclusive-or" of the exact
     integer N1 and the bitwise complement of the exact integer N2.

     For example:

          > (bitwise-orc2 11 26)  ; ...00001011 ...00011010
          -17
          > (bitwise-orc2 11 -27) ; ...00001011 ...11100101
          27


 -- procedure: bitwise-xor N...
     This procedure returns the bitwise "exclusive-or" of the exact
     integers N....  The value 0 is returned when there are no
     arguments.

     For example:

          > (bitwise-xor 6 12)  ; ...00000110 ...00001100
          10
          > (bitwise-xor 6 -4)  ; ...00000110 ...11111100
          -6
          > (bitwise-xor -6 -4) ; ...11111010 ...11111100
          6
          > (bitwise-xor)
          0


 -- procedure: bit-count N
     This procedure returns the bit count of the exact integer N.  If N
     is nonnegative, the bit count is the number of 1 bits in the two's
     complement representation of N.  If N is negative, the bit count
     is the number of 0 bits in the two's complement representation of
     N.

     For example:

          > (bit-count 0)   ; ...00000000
          0
          > (bit-count 1)   ; ...00000001
          1
          > (bit-count 2)   ; ...00000010
          1
          > (bit-count 3)   ; ...00000011
          2
          > (bit-count 4)   ; ...00000100
          1
          > (bit-count -23) ; ...11101001
          3


 -- procedure: integer-length N
     This procedure returns the bit length of the exact integer N.  If
     N is a positive integer the bit length is one more than the index
     of the highest 1 bit (the least significant bit is at index 0).
     If N is a negative integer the bit length is one more than the
     index of the highest 0 bit.  If N is zero, the bit length is 0.

     For example:

          > (integer-length 0)   ; ...00000000
          0
          > (integer-length 1)   ; ...00000001
          1
          > (integer-length 2)   ; ...00000010
          2
          > (integer-length 3)   ; ...00000011
          2
          > (integer-length 4)   ; ...00000100
          3
          > (integer-length -23) ; ...11101001
          5


 -- procedure: bit-set? N1 N2
     This procedure returns a boolean indicating if the bit at index N1
     of N2 is set (i.e. equal to 1) or not.  Both N1 and N2 must be
     exact integers, and N1 must be nonnegative.

     For example:

          > (map (lambda (i) (bit-set? i -23)) ; ...11101001
                 '(7 6 5 4 3 2 1 0))
          (#t #t #t #f #t #f #f #t)


 -- procedure: any-bits-set? N1 N2
     This procedure returns a boolean indicating if the bitwise and of
     N1 and N2 is different from zero or not.  This procedure is
     implemented more efficiently than the naive definition:

          (define (any-bits-set? n1 n2) (not (zero? (bitwise-and n1 n2))))

     For example:

          > (any-bits-set? 5 10)   ; ...00000101 ...00001010
          #f
          > (any-bits-set? -23 32) ; ...11101001 ...00100000
          #t


 -- procedure: all-bits-set? N1 N2
     This procedure returns a boolean indicating if the bitwise and of
     N1 and N2 is equal to N1 or not.  This procedure is implemented
     more efficiently than the naive definition:

          (define (all-bits-set? n1 n2) (= n1 (bitwise-and n1 n2)))

     For example:

          > (all-bits-set? 1 3) ; ...00000001 ...00000011
          #t
          > (all-bits-set? 7 3) ; ...00000111 ...00000011
          #f


 -- procedure: first-set-bit N
     This procedure returns the bit index of the least significant bit
     of N equal to 1 (which is also the number of 0 bits that are below
     the least significant 1 bit).  This procedure returns `-1' when N
     is zero.

     For example:

          > (first-set-bit 24) ; ...00011000
          3
          > (first-set-bit 0)  ; ...00000000
          -1


 -- procedure: extract-bit-field N1 N2 N3
 -- procedure: test-bit-field? N1 N2 N3
 -- procedure: clear-bit-field N1 N2 N3
 -- procedure: replace-bit-field N1 N2 N3 N4
 -- procedure: copy-bit-field N1 N2 N3 N4
     These procedures operate on a bit-field which is N1 bits wide
     starting at bit index N2.  All arguments must be exact integers
     and N1 and N2 must be nonnegative.

     The procedure `extract-bit-field' returns the bit-field of N3
     shifted to the right so that the least significant bit of the
     bit-field is the least significant bit of the result.

     The procedure `test-bit-field?' returns `#t' if any bit in the
     bit-field of N3 is equal to 1, otherwise `#f' is returned.

     The procedure `clear-bit-field' returns N3 with all bits in the
     bit-field replaced with 0.

     The procedure `replace-bit-field' returns N4 with the bit-field
     replaced with the least-significant N1 bits of N3.

     The procedure `copy-bit-field' returns N4 with the bit-field
     replaced with the (same index and size) bit-field in N3.

     For example:

          > (extract-bit-field 5 2 -37)    ; ...11011011
          22
          > (test-bit-field? 5 2 -37)      ; ...11011011
          #t
          > (test-bit-field? 1 2 -37)      ; ...11011011
          #f
          > (clear-bit-field 5 2 -37)      ; ...11011011
          -125
          > (replace-bit-field 5 2 -6 -37) ; ...11111010 ...11011011
          -21
          > (copy-bit-field 5 2 -6 -37)    ; ...11111010 ...11011011
          -5


8.1.5 Fixnum specific operations
--------------------------------

 -- procedure: fixnum? OBJ

 -- procedure: fx* N1...

 -- procedure: fx+ N1...

 -- procedure: fx- N1 N2...

 -- procedure: fx< N1...

 -- procedure: fx<= N1...

 -- procedure: fx= N1...

 -- procedure: fx> N1...

 -- procedure: fx>= N1...

 -- procedure: fxabs N

 -- procedure: fxand N1...

 -- procedure: fxandc1 N1 N2

 -- procedure: fxandc2 N1 N2

 -- procedure: fxarithmetic-shift N1 N2

 -- procedure: fxarithmetic-shift-left N1 N2

 -- procedure: fxarithmetic-shift-right N1 N2

 -- procedure: fxbit-count N

 -- procedure: fxbit-set? N1 N2

 -- procedure: fxeqv N1...

 -- procedure: fxeven? N

 -- procedure: fxfirst-set-bit N

 -- procedure: fxif N1 N2 N3

 -- procedure: fxior N1...

 -- procedure: fxlength N

 -- procedure: fxmax N1 N2...

 -- procedure: fxmin N1 N2...

 -- procedure: fxmodulo N1 N2

 -- procedure: fxnegative? N

 -- procedure: fxnand N1 N2

 -- procedure: fxnor N1 N2

 -- procedure: fxnot N

 -- procedure: fxodd? N

 -- procedure: fxorc1 N1 N2

 -- procedure: fxorc2 N1 N2

 -- procedure: fxpositive? N

 -- procedure: fxquotient N1 N2

 -- procedure: fxremainder N1 N2

 -- procedure: fxwrap* N1...

 -- procedure: fxwrap+ N1...

 -- procedure: fxwrap- N1 N2...

 -- procedure: fxwrapabs N

 -- procedure: fxwraparithmetic-shift N1 N2

 -- procedure: fxwraparithmetic-shift-left N1 N2

 -- procedure: fxwraplogical-shift-right N1 N2

 -- procedure: fxwrapquotient N1 N2

 -- procedure: fxxor N1...

 -- procedure: fxzero? N

 -- procedure: fxsquare N

 -- procedure: fxwrapsquare N

 -- procedure: fixnum-overflow-exception? OBJ
 -- procedure: fixnum-overflow-exception-procedure EXC
 -- procedure: fixnum-overflow-exception-arguments EXC
     Fixnum-overflow-exception objects are raised by some of the fixnum
     specific procedures when the result is larger than can fit in a
     fixnum.  The parameter EXC must be a fixnum-overflow-exception
     object.

     The procedure `fixnum-overflow-exception?' returns `#t' when OBJ
     is a fixnum-overflow-exception object and `#f' otherwise.

     The procedure `fixnum-overflow-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `fixnum-overflow-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (fixnum-overflow-exception? exc)
                  (list (fixnum-overflow-exception-procedure exc)
                        (fixnum-overflow-exception-arguments exc))
                  'not-fixnum-overflow-exception))
          > (with-exception-catcher
              handler
              (lambda () (fx* 100000 100000)))
          (#<procedure #2 fx*> (100000 100000))


8.1.6 Flonum specific operations
--------------------------------

 -- procedure: flonum? OBJ

 -- procedure: fixnum->flonum N

 -- procedure: fl* X1...

 -- procedure: fl+ X1...

 -- procedure: fl- X1 X2...

 -- procedure: fl/ X1 X2

 -- procedure: fl< X1...

 -- procedure: fl<= X1...

 -- procedure: fl= X1...

 -- procedure: fl> X1...

 -- procedure: fl>= X1...

 -- procedure: flabs X

 -- procedure: flacos X

 -- procedure: flasin X

 -- procedure: flatan X
 -- procedure: flatan Y X

 -- procedure: flceiling X

 -- procedure: flcos X

 -- procedure: fldenominator X

 -- procedure: fleven? X

 -- procedure: flexp X

 -- procedure: flexpt X Y

 -- procedure: flhypot X Y

 -- procedure: flfinite? X

 -- procedure: flfloor X

 -- procedure: flinfinite? X

 -- procedure: flinteger? X

 -- procedure: fllog X

 -- procedure: flmax X1 X2...

 -- procedure: flmin X1 X2...

 -- procedure: flnan? X

 -- procedure: flnegative? X

 -- procedure: flnumerator X

 -- procedure: flodd? X

 -- procedure: flpositive? X

 -- procedure: flround X

 -- procedure: flsin X

 -- procedure: flsqrt X

 -- procedure: fltan X

 -- procedure: fltruncate X

 -- procedure: flzero? X

 -- procedure: fl+* X1 X2 X3

 -- procedure: flacosh X

 -- procedure: flasinh X

 -- procedure: flatanh X

 -- procedure: flcosh X

 -- procedure: flexpm1 X

 -- procedure: flilogb X

 -- procedure: fllog1p X

 -- procedure: flscalbn X

 -- procedure: flsinh X

 -- procedure: flsquare X

 -- procedure: fltanh X

8.1.7 Pseudo random numbers
---------------------------

The procedures and variables defined in this section are compatible
with the "Sources of Random Bits SRFI" (SRFI 27).  The implementation
is based on Pierre L'Ecuyer's MRG32k3a pseudo random number generator.
At the heart of SRFI 27's interface is the random source type which
encapsulates the state of a pseudo random number generator.  The state
of a random source object changes every time a pseudo random number is
generated from this random source object.

 -- variable: default-random-source
     The global variable `default-random-source' is bound to the random
     source object which is used by the `random-integer',
     `random-real', `random-u8vector' and `random-f64vector' procedures.


 -- procedure: random-integer N
     This procedure returns a pseudo random exact integer in the range
     0 to N-1.  The random source object in the global variable
     `default-random-source' is used to generate this number.  The
     parameter N must be a positive exact integer.

     For example:

          > (random-integer 100)
          24
          > (random-integer 100)
          2
          > (random-integer 10000000000000000000000000000000000000000)
          6143360270902284438072426748425263488507


 -- procedure: random-real
     This procedure returns a pseudo random inexact real between, but
     not including, 0 and 1.  The random source object in the global
     variable `default-random-source' is used to generate this number.

     For example:

          > (random-real)
          .24230672079133753
          > (random-real)
          .02317001922506932


 -- procedure: random-u8vector N
     This procedure returns a u8vector of length N containing pseudo
     random exact integers in the range 0 to 255.  The random source
     object in the global variable `default-random-source' is used to
     generate these numbers.  The parameter N must be a nonnegative
     exact integer.

     For example:

          > (random-u8vector 10)
          #u8(200 53 29 202 3 85 208 187 73 219)


 -- procedure: random-f64vector N
     This procedure returns a f64vector of length N containing pseudo
     random inexact reals between, but not including, 0 and 1.  The
     random source object in the global variable
     `default-random-source' is used to generate these numbers.  The
     parameter N must be a nonnegative exact integer.

     For example:

          > (random-f64vector 3)
          #f64(.7145854494613069 .47089632669147946 .5400124875182746)


 -- procedure: make-random-source
     This procedure returns a new random source object initialized to a
     predetermined state (to initialize to a pseudo random state the
     procedure `random-source-randomize!' should be called).

     For example:

          > (define rs (make-random-source))
          > ((random-source-make-integers rs) 10000000)
          8583952


 -- procedure: random-source? OBJ
     This procedure returns `#t' when OBJ is a random source object and
     `#f' otherwise.

     For example:

          > (random-source? default-random-source)
          #t
          > (random-source? 123)
          #f


 -- procedure: random-source-state-ref RANDOM-SOURCE
 -- procedure: random-source-state-set! RANDOM-SOURCE STATE
     The procedure `random-source-state-ref' extracts the state of the
     random source object RANDOM-SOURCE and returns a vector containing
     the state.

     The procedure `random-source-state-set!' restores the state of the
     random source object RANDOM-SOURCE to STATE which must be a vector
     returned from a call to the procedure `random-source-state-ref'.

     For example:

          > (define s (random-source-state-ref default-random-source))
          > (random-integer 10000000000000000000000000000000000000000)
          7583880188903074396261960585615270693321
          > (random-source-state-set! default-random-source s)
          > (random-integer 10000000000000000000000000000000000000000)
          7583880188903074396261960585615270693321


 -- procedure: random-source-randomize! RANDOM-SOURCE
 -- procedure: random-source-pseudo-randomize! RANDOM-SOURCE I J
     These procedures change the state of the random source object
     RANDOM-SOURCE.  The procedure `random-source-randomize!' sets the
     random source object to a state that depends on the current time
     (which for typical uses can be considered to randomly initialize
     the state).  The procedure `random-source-pseudo-randomize!' sets
     the random source object to a state that is determined only by the
     current state and the nonnegative exact integers I and J.  For
     both procedures the value returned is unspecified.

     For example:

          > (define s (random-source-state-ref default-random-source))
          > (random-source-pseudo-randomize! default-random-source 5 99)
          > (random-integer 10000000000000000000000000000000000000000)
          9816755163910623041601722050112674079767
          > (random-source-state-set! default-random-source s)
          > (random-source-pseudo-randomize! default-random-source 5 99)
          > (random-integer 10000000000000000000000000000000000000000)
          9816755163910623041601722050112674079767
          > (random-source-pseudo-randomize! default-random-source 5 99)
          > (random-integer 10000000000000000000000000000000000000000)
          9816755163910623041601722050112674079767
          > (random-source-state-set! default-random-source s)
          > (random-source-randomize! default-random-source)
          > (random-integer 10000000000000000000000000000000000000000)
          2271441220851914333384493143687768110622
          > (random-source-state-set! default-random-source s)
          > (random-source-randomize! default-random-source)
          > (random-integer 10000000000000000000000000000000000000000)
          6247966138948323029033944059178072366895


 -- procedure: random-source-make-integers RANDOM-SOURCE
     This procedure returns a procedure for generating pseudo random
     exact integers using the random source object RANDOM-SOURCE.  The
     returned procedure accepts a single parameter N, a positive exact
     integer, and returns a pseudo random exact integer in the range 0
     to N-1.

     For example:

          > (define rs (make-random-source))
          > (define ri (random-source-make-integers rs))
          > (ri 10000000)
          8583952
          > (ri 10000000)
          2879793


 -- procedure: random-source-make-reals RANDOM-SOURCE [PRECISION]
     This procedure returns a procedure for generating pseudo random
     inexact reals using the random source object RANDOM-SOURCE.  The
     returned procedure accepts no parameters and returns a pseudo
     random inexact real between, but not including, 0 and 1.  The
     optional parameter PRECISION specifies an upper bound on the
     minimum amount by which two generated pseudo-random numbers can be
     separated.

     For example:

          > (define rs (make-random-source))
          > (define rr (random-source-make-reals rs))
          > (rr)
          .857402537562821
          > (rr)
          .2876463473845367


 -- procedure: random-source-make-u8vectors RANDOM-SOURCE
     This procedure returns a procedure for generating pseudo random
     u8vectors using the random source object RANDOM-SOURCE.  The
     returned procedure accepts a single parameter N, a nonnegative
     exact integer, and returns a u8vector of length N containing
     pseudo random exact integers in the range 0 to 255.

     For example:

          > (define rs (make-random-source))
          > (define rv (random-source-make-u8vectors rs))
          > (rv 10)
          #u8(200 53 29 202 3 85 208 187 73 219)
          > (rv 10)
          #u8(113 8 182 120 138 103 53 192 40 176)


 -- procedure: random-source-make-f64vectors RANDOM-SOURCE [PRECISION]
     This procedure returns a procedure for generating pseudo random
     f64vectors using the random source object RANDOM-SOURCE.  The
     returned procedure accepts a single parameter N, a nonnegative
     exact integer, and returns an f64vector of length N containing
     pseudo random inexact reals between, but not including, 0 and 1.
     The optional parameter PRECISION specifies an upper bound on the
     minimum amount by which two generated pseudo-random numbers can be
     separated.

     For example:

          > (define rs (make-random-source))
          > (define rv (random-source-make-f64vectors rs))
          > (rv 3)
          #f64(.7342236104231586 .2876463473845367 .8574025375628211)
          > (rv 3)
          #f64(.013863292728449427 .33449296573515447 .8162050798467028)


8.2 Booleans
============

8.3 Pairs and lists
===================

8.4 Symbols and keywords
========================

8.5 Characters and strings
==========================

Gambit supports the Unicode character encoding standard.  Scheme
characters can be any of the characters whose Unicode encoding is in
the range 0 to #x10ffff (inclusive) but not in the range #xd800 to
#xdfff.  Source code can also contain any Unicode character, however to
read such source code properly `gsi' and `gsc' must be told which
character encoding to use for reading the source code (i.e. ISO-8859-1,
UTF-8, UTF-16, etc).  This can be done by specifying the runtime option
`-:file-settings=...' or `-:io-settings=...' when `gsi' and `gsc' are
started.

8.6 Extensions to character procedures
======================================

 -- procedure: char->integer CHAR
 -- procedure: integer->char N
     The procedure `char->integer' returns the Unicode encoding of the
     character CHAR.

     The procedure `integer->char' returns the character whose Unicode
     encoding is the exact integer N.

     For example:

          > (char->integer #\!)
          33
          > (integer->char 65)
          #\A
          > (integer->char (char->integer #\u1234))
          #\u1234
          > (integer->char #xd800)
          *** ERROR IN (console)@4.1 -- (Argument 1) Out of range
          (integer->char 55296)


 -- procedure: char=? CHAR1...
 -- procedure: char<? CHAR1...
 -- procedure: char>? CHAR1...
 -- procedure: char<=? CHAR1...
 -- procedure: char>=? CHAR1...
 -- procedure: char-ci=? CHAR1...
 -- procedure: char-ci<? CHAR1...
 -- procedure: char-ci>? CHAR1...
 -- procedure: char-ci<=? CHAR1...
 -- procedure: char-ci>=? CHAR1...
     These procedures take any number of arguments including no
     argument.  This is useful to test if the elements of a list are
     sorted in a particular order.  For example, testing that the list
     of characters `lst' is sorted in nondecreasing order can be done
     with the call `(apply char<? lst)'.


8.7 Extensions to string procedures
===================================

 -- procedure: string=? STRING1...
 -- procedure: string<? STRING1...
 -- procedure: string>? STRING1...
 -- procedure: string<=? STRING1...
 -- procedure: string>=? STRING1...
 -- procedure: string-ci=? STRING1...
 -- procedure: string-ci<? STRING1...
 -- procedure: string-ci>? STRING1...
 -- procedure: string-ci<=? STRING1...
 -- procedure: string-ci>=? STRING1...
     These procedures take any number of arguments including no
     argument.  This is useful to test if the elements of a list are
     sorted in a particular order.  For example, testing that the list
     of strings `lst' is sorted in nondecreasing order can be done with
     the call `(apply string<? lst)'.


8.8 Vectors
===========

8.9 Homogeneous numeric vectors
===============================

Homogeneous vectors are vectors containing raw numbers of the same type
(signed or unsigned exact integers or inexact reals).  There are 10
types of homogeneous vectors: `s8vector' (vector of exact integers in
the range -2^7 to 2^7-1), `u8vector' (vector of exact integers in the
range 0 to 2^8-1), `s16vector' (vector of exact integers in the range
-2^15 to 2^15-1), `u16vector' (vector of exact integers in the range 0
to 2^16-1), `s32vector' (vector of exact integers in the range -2^31 to
2^31-1), `u32vector' (vector of exact integers in the range 0 to
2^32-1), `s64vector' (vector of exact integers in the range -2^63 to
2^63-1), `u64vector' (vector of exact integers in the range 0 to
2^64-1), `f32vector' (vector of 32 bit floating point numbers), and
`f64vector' (vector of 64 bit floating point numbers).

   The lexical syntax of homogeneous vectors is specified in *Note
Homogeneous vector syntax::.

   The procedures available for homogeneous vectors, listed below, are
the analog of the normal vector/string procedures for each of the
homogeneous vector types.

 -- procedure: s8vector? OBJ
 -- procedure: make-s8vector K [FILL]
 -- procedure: s8vector EXACT-INT8...
 -- procedure: s8vector-length S8VECTOR
 -- procedure: s8vector-ref S8VECTOR K
 -- procedure: s8vector-set S8VECTOR K EXACT-INT8
 -- procedure: s8vector-set! S8VECTOR K EXACT-INT8
 -- procedure: s8vector->list S8VECTOR
 -- procedure: list->s8vector LIST-OF-EXACT-INT8
 -- procedure: s8vector-fill! S8VECTOR FILL [START [END]]
 -- procedure: subs8vector-fill! VECTOR START END FILL
 -- procedure: s8vector-concatenate LST [SEPARATOR]
 -- procedure: s8vector-copy S8VECTOR [START [END]]
 -- procedure: s8vector-copy! DEST-S8VECTOR DEST-START S8VECTOR [START
          [END]]
 -- procedure: s8vector-append S8VECTOR...
 -- procedure: subs8vector S8VECTOR START END
 -- procedure: subs8vector-move! SRC-S8VECTOR SRC-START SRC-END
          DST-S8VECTOR DST-START
 -- procedure: s8vector-shrink! S8VECTOR K

 -- procedure: u8vector? OBJ
 -- procedure: make-u8vector K [FILL]
 -- procedure: u8vector EXACT-INT8...
 -- procedure: u8vector-length U8VECTOR
 -- procedure: u8vector-ref U8VECTOR K
 -- procedure: u8vector-set U8VECTOR K EXACT-INT8
 -- procedure: u8vector-set! U8VECTOR K EXACT-INT8
 -- procedure: u8vector->list U8VECTOR
 -- procedure: list->u8vector LIST-OF-EXACT-INT8
 -- procedure: u8vector-fill! U8VECTOR FILL [START [END]]
 -- procedure: subu8vector-fill! VECTOR START END FILL
 -- procedure: u8vector-concatenate LST [SEPARATOR]
 -- procedure: u8vector-copy U8VECTOR [START [END]]
 -- procedure: u8vector-copy! DEST-U8VECTOR DEST-START U8VECTOR [START
          [END]]
 -- procedure: u8vector-append U8VECTOR...
 -- procedure: subu8vector U8VECTOR START END
 -- procedure: subu8vector-move! SRC-U8VECTOR SRC-START SRC-END
          DST-U8VECTOR DST-START
 -- procedure: u8vector-shrink! U8VECTOR K

 -- procedure: s16vector? OBJ
 -- procedure: make-s16vector K [FILL]
 -- procedure: s16vector EXACT-INT16...
 -- procedure: s16vector-length S16VECTOR
 -- procedure: s16vector-ref S16VECTOR K
 -- procedure: s16vector-set S16VECTOR K EXACT-INT16
 -- procedure: s16vector-set! S16VECTOR K EXACT-INT16
 -- procedure: s16vector->list S16VECTOR
 -- procedure: list->s16vector LIST-OF-EXACT-INT16
 -- procedure: s16vector-fill! S16VECTOR FILL [START [END]]
 -- procedure: subs16vector-fill! VECTOR START END FILL
 -- procedure: s16vector-concatenate LST [SEPARATOR]
 -- procedure: s16vector-copy S16VECTOR [START [END]]
 -- procedure: s16vector-copy! DEST-S16VECTOR DEST-START S16VECTOR
          [START [END]]
 -- procedure: s16vector-append S16VECTOR...
 -- procedure: subs16vector S16VECTOR START END
 -- procedure: subs16vector-move! SRC-S16VECTOR SRC-START SRC-END
          DST-S16VECTOR DST-START
 -- procedure: s16vector-shrink! S16VECTOR K

 -- procedure: u16vector? OBJ
 -- procedure: make-u16vector K [FILL]
 -- procedure: u16vector EXACT-INT16...
 -- procedure: u16vector-length U16VECTOR
 -- procedure: u16vector-ref U16VECTOR K
 -- procedure: u16vector-set U16VECTOR K EXACT-INT16
 -- procedure: u16vector-set! U16VECTOR K EXACT-INT16
 -- procedure: u16vector->list U16VECTOR
 -- procedure: list->u16vector LIST-OF-EXACT-INT16
 -- procedure: u16vector-fill! U16VECTOR FILL [START [END]]
 -- procedure: subu16vector-fill! VECTOR START END FILL
 -- procedure: u16vector-concatenate LST [SEPARATOR]
 -- procedure: u16vector-copy U16VECTOR [START [END]]
 -- procedure: u16vector-copy! DEST-U16VECTOR DEST-START U16VECTOR
          [START [END]]
 -- procedure: u16vector-append U16VECTOR...
 -- procedure: subu16vector U16VECTOR START END
 -- procedure: subu16vector-move! SRC-U16VECTOR SRC-START SRC-END
          DST-U16VECTOR DST-START
 -- procedure: u16vector-shrink! U16VECTOR K

 -- procedure: s32vector? OBJ
 -- procedure: make-s32vector K [FILL]
 -- procedure: s32vector EXACT-INT32...
 -- procedure: s32vector-length S32VECTOR
 -- procedure: s32vector-ref S32VECTOR K
 -- procedure: s32vector-set S32VECTOR K EXACT-INT32
 -- procedure: s32vector-set! S32VECTOR K EXACT-INT32
 -- procedure: s32vector->list S32VECTOR
 -- procedure: list->s32vector LIST-OF-EXACT-INT32
 -- procedure: s32vector-fill! S32VECTOR FILL [START [END]]
 -- procedure: subs32vector-fill! VECTOR START END FILL
 -- procedure: s32vector-concatenate LST [SEPARATOR]
 -- procedure: s32vector-copy S32VECTOR [START [END]]
 -- procedure: s32vector-copy! DEST-S32VECTOR DEST-START S32VECTOR
          [START [END]]
 -- procedure: s32vector-append S32VECTOR...
 -- procedure: subs32vector S32VECTOR START END
 -- procedure: subs32vector-move! SRC-S32VECTOR SRC-START SRC-END
          DST-S32VECTOR DST-START
 -- procedure: s32vector-shrink! S32VECTOR K

 -- procedure: u32vector? OBJ
 -- procedure: make-u32vector K [FILL]
 -- procedure: u32vector EXACT-INT32...
 -- procedure: u32vector-length U32VECTOR
 -- procedure: u32vector-ref U32VECTOR K
 -- procedure: u32vector-set U32VECTOR K EXACT-INT32
 -- procedure: u32vector-set! U32VECTOR K EXACT-INT32
 -- procedure: u32vector->list U32VECTOR
 -- procedure: list->u32vector LIST-OF-EXACT-INT32
 -- procedure: u32vector-fill! U32VECTOR FILL [START [END]]
 -- procedure: subu32vector-fill! VECTOR START END FILL
 -- procedure: u32vector-concatenate LST [SEPARATOR]
 -- procedure: u32vector-copy U32VECTOR [START [END]]
 -- procedure: u32vector-copy! DEST-U32VECTOR DEST-START U32VECTOR
          [START [END]]
 -- procedure: u32vector-append U32VECTOR...
 -- procedure: subu32vector U32VECTOR START END
 -- procedure: subu32vector-move! SRC-U32VECTOR SRC-START SRC-END
          DST-U32VECTOR DST-START
 -- procedure: u32vector-shrink! U32VECTOR K

 -- procedure: s64vector? OBJ
 -- procedure: make-s64vector K [FILL]
 -- procedure: s64vector EXACT-INT64...
 -- procedure: s64vector-length S64VECTOR
 -- procedure: s64vector-ref S64VECTOR K
 -- procedure: s64vector-set S64VECTOR K EXACT-INT64
 -- procedure: s64vector-set! S64VECTOR K EXACT-INT64
 -- procedure: s64vector->list S64VECTOR
 -- procedure: list->s64vector LIST-OF-EXACT-INT64
 -- procedure: s64vector-fill! S64VECTOR FILL [START [END]]
 -- procedure: subs64vector-fill! VECTOR START END FILL
 -- procedure: s64vector-concatenate LST [SEPARATOR]
 -- procedure: s64vector-copy S64VECTOR [START [END]]
 -- procedure: s64vector-copy! DEST-S64VECTOR DEST-START S64VECTOR
          [START [END]]
 -- procedure: s64vector-append S64VECTOR...
 -- procedure: subs64vector S64VECTOR START END
 -- procedure: subs64vector-move! SRC-S64VECTOR SRC-START SRC-END
          DST-S64VECTOR DST-START
 -- procedure: s64vector-shrink! S64VECTOR K

 -- procedure: u64vector? OBJ
 -- procedure: make-u64vector K [FILL]
 -- procedure: u64vector EXACT-INT64...
 -- procedure: u64vector-length U64VECTOR
 -- procedure: u64vector-ref U64VECTOR K
 -- procedure: u64vector-set U64VECTOR K EXACT-INT64
 -- procedure: u64vector-set! U64VECTOR K EXACT-INT64
 -- procedure: u64vector->list U64VECTOR
 -- procedure: list->u64vector LIST-OF-EXACT-INT64
 -- procedure: u64vector-fill! U64VECTOR FILL [START [END]]
 -- procedure: subu64vector-fill! VECTOR START END FILL
 -- procedure: u64vector-concatenate LST [SEPARATOR]
 -- procedure: u64vector-copy U64VECTOR [START [END]]
 -- procedure: u64vector-copy! DEST-U64VECTOR DEST-START U64VECTOR
          [START [END]]
 -- procedure: u64vector-append U64VECTOR...
 -- procedure: subu64vector U64VECTOR START END
 -- procedure: subu64vector-move! SRC-U64VECTOR SRC-START SRC-END
          DST-U64VECTOR DST-START
 -- procedure: u64vector-shrink! U64VECTOR K

 -- procedure: f32vector? OBJ
 -- procedure: make-f32vector K [FILL]
 -- procedure: f32vector INEXACT-REAL...
 -- procedure: f32vector-length F32VECTOR
 -- procedure: f32vector-ref F32VECTOR K
 -- procedure: f32vector-set F32VECTOR K INEXACT-REAL
 -- procedure: f32vector-set! F32VECTOR K INEXACT-REAL
 -- procedure: f32vector->list F32VECTOR
 -- procedure: list->f32vector LIST-OF-INEXACT-REAL
 -- procedure: f32vector-fill! F32VECTOR FILL [START [END]]
 -- procedure: subf32vector-fill! VECTOR START END FILL
 -- procedure: f32vector-concatenate LST [SEPARATOR]
 -- procedure: f32vector-copy F32VECTOR [START [END]]
 -- procedure: f32vector-copy! DEST-F32VECTOR DEST-START F32VECTOR
          [START [END]]
 -- procedure: f32vector-append F32VECTOR...
 -- procedure: subf32vector F32VECTOR START END
 -- procedure: subf32vector-move! SRC-F32VECTOR SRC-START SRC-END
          DST-F32VECTOR DST-START
 -- procedure: f32vector-shrink! F32VECTOR K

 -- procedure: f64vector? OBJ
 -- procedure: make-f64vector K [FILL]
 -- procedure: f64vector INEXACT-REAL...
 -- procedure: f64vector-length F64VECTOR
 -- procedure: f64vector-ref F64VECTOR K
 -- procedure: f64vector-set F64VECTOR K INEXACT-REAL
 -- procedure: f64vector-set! F64VECTOR K INEXACT-REAL
 -- procedure: f64vector->list F64VECTOR
 -- procedure: list->f64vector LIST-OF-INEXACT-REAL
 -- procedure: f64vector-fill! F64VECTOR FILL [START [END]]
 -- procedure: subf64vector-fill! VECTOR START END FILL
 -- procedure: f64vector-concatenate LST [SEPARATOR]
 -- procedure: f64vector-copy F64VECTOR [START [END]]
 -- procedure: f64vector-copy! DEST-F64VECTOR DEST-START F64VECTOR
          [START [END]]
 -- procedure: f64vector-append F64VECTOR...
 -- procedure: subf64vector F64VECTOR START END
 -- procedure: subf64vector-move! SRC-F64VECTOR SRC-START SRC-END
          DST-F64VECTOR DST-START
 -- procedure: f64vector-shrink! F64VECTOR K

   For example:

     > (define v (u8vector 10 255 13))
     > (u8vector-set! v 2 99)
     > v
     #u8(10 255 99)
     > (u8vector-ref v 1)
     255
     > (u8vector->list v)
     (10 255 99)
     > (u8vector-shrink! v 2)
     > (v)
     #u8(10 255)

 -- procedure: object->u8vector OBJ [ENCODER]
 -- procedure: u8vector->object U8VECTOR [DECODER]
     The procedure `object->u8vector' returns a u8vector that contains
     the sequence of bytes that encodes the object OBJ.  The procedure
     `u8vector->object' decodes the sequence of bytes contained in the
     u8vector U8VECTOR, which was produced by the procedure
     `object->u8vector', and reconstructs an object structurally equal
     to the original object.  In other words the procedures
     `object->u8vector' and `u8vector->object' respectively perform
     serialization and deserialization of Scheme objects.  Note that
     some objects are non-serializable (e.g. threads, wills, some types
     of ports, and any object containing a non-serializable object).

     The optional ENCODER and DECODER parameters are single parameter
     procedures which default to the identity function.  The ENCODER
     procedure is called during serialization.  As the serializer walks
     through OBJ, it calls the ENCODER procedure on each sub-object X
     that is encountered.  The ENCODER transforms the object X into an
     object Y that will be serialized instead of X.  Similarly the
     DECODER procedure is called during deserialization.  When an
     object Y is encountered, the DECODER procedure is called to
     transform it into the object X that is the result of
     deserialization.

     The ENCODER and DECODER procedures are useful to customize the
     serialized representation of objects.  In particular, it can be
     used to define the semantics of serializing objects, such as
     threads and ports, that would otherwise not be serializable.  The
     DECODER procedure is typically the inverse of the ENCODER
     procedure, i.e. `(DECODER (ENCODER X))' = `X'.

     For example:

          > (define (make-adder x) (lambda (y) (+ x y)))
          > (define f (make-adder 10))
          > (define a (object->u8vector f))
          > (define b (u8vector->object a))
          > (u8vector-length a)
          1639
          > (f 5)
          15
          > (b 5)
          15
          > (pp b)
          (lambda (y) (+ x y))


8.10 Hashing and weak references
================================

8.10.1 Hashing
--------------

 -- procedure: object->serial-number OBJ
 -- procedure: serial-number->object N [DEFAULT]
     All Scheme objects are uniquely identified with a serial number
     which is a nonnegative exact integer.  The `object->serial-number'
     procedure returns the serial number of object OBJ.  This serial
     number is only allocated the first time the `object->serial-number'
     procedure is called on that object.  Objects which do not have an
     external textual representation that can be read by the `read'
     procedure, use an external textual representation that includes a
     serial number of the form `#N'.  Consequently, the procedures
     `write', `pretty-print', etc will call the `object->serial-number'
     procedure to get the serial number, and this may cause the serial
     number to be allocated.

     The `serial-number->object' procedure takes an exact integer
     parameter N and returns the object whose serial number is N.  If
     no object currently exists with that serial number, DEFAULT is
     returned if it is specified, otherwise an
     unbound-serial-number-exception object is raised.  The reader
     defines the following abbreviation for calling
     `serial-number->object': the syntax `#N', where N is a sequence of
     decimal digits and it is not followed by ``='' or ``#'', is
     equivalent to the list `(serial-number->object N)'.

     For example:

          > (define z (list (lambda (x) (* x x)) (lambda (y) (/ 1 y))))
          > z
          (#<procedure #2> #<procedure #3>)
          > (#3 10)
          1/10
          > '(#3 10)
          ((serial-number->object 3) 10)
          > car
          #<procedure #4 car>
          > (#4 z)
          #<procedure #2>


 -- procedure: unbound-serial-number-exception? OBJ
 -- procedure: unbound-serial-number-exception-procedure EXC
 -- procedure: unbound-serial-number-exception-arguments EXC
     Unbound-serial-number-exception objects are raised by the procedure
     `serial-number->object' when no object currently exists with that
     serial number.  The parameter EXC must be an
     unbound-serial-number-exception object.

     The procedure `unbound-serial-number-exception?' returns `#t' when
     OBJ is a unbound-serial-number-exception object and `#f' otherwise.

     The procedure `unbound-serial-number-exception-procedure' returns
     the procedure that raised EXC.

     The procedure `unbound-serial-number-exception-arguments' returns
     the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (unbound-serial-number-exception? exc)
                  (list (unbound-serial-number-exception-procedure exc)
                        (unbound-serial-number-exception-arguments exc))
                  'not-unbound-serial-number-exception))
          > (with-exception-catcher
              handler
              (lambda () (serial-number->object 1000)))
          (#<procedure #2 serial-number->object> (1000))


 -- procedure: symbol-hash SYMBOL
     The `symbol-hash' procedure returns the hash number of the symbol
     SYMBOL.  The hash number is a small exact integer (fixnum).  When
     SYMBOL is an interned symbol the value returned is the same as
     `(string=?-hash (symbol->string SYMBOL))'.

     For example:

          > (symbol-hash 'car)
          444471047


 -- procedure: keyword-hash KEYWORD
     The `keyword-hash' procedure returns the hash number of the
     keyword KEYWORD.  The hash number is a small exact integer
     (fixnum).  When KEYWORD is an interned keyword the value returned
     is the same as `(string=?-hash (keyword->string KEYWORD))'.

     For example:

          > (keyword-hash car:)
          444471047


 -- procedure: string=?-hash STRING
     The `string=?-hash' procedure returns the hash number of the
     string STRING.  The hash number is a small exact integer (fixnum).
     For any two strings S1 and S2, `(string=?  S1 S2)' implies `(=
     (string=?-hash S1) (string=?-hash S2))'.

     For example:

          > (string=?-hash "car")
          444471047


 -- procedure: string-ci=?-hash STRING
     The `string-ci=?-hash' procedure returns the hash number of the
     string STRING.  The hash number is a small exact integer (fixnum).
     For any two strings S1 and S2, `(string-ci=?  S1 S2)' implies `(=
     (string-ci=?-hash S1) (string-ci=?-hash S2))'.

     For example:

          > (string-ci=?-hash "CaR")
          444471047


 -- procedure: eq?-hash OBJ
     The `eq?-hash' procedure returns the hash number of the object
     OBJ.  The hash number is a small exact integer (fixnum).  For any
     two objects O1 and O2, `(eq? O1 O2)' implies `(= (eq?-hash O1)
     (eq?-hash O2))'.

     For example:

          > (eq?-hash #t)
          536870910


 -- procedure: eqv?-hash OBJ
     The `eqv?-hash' procedure returns the hash number of the object
     OBJ.  The hash number is a small exact integer (fixnum).  For any
     two objects O1 and O2, `(eqv? O1 O2)' implies `(= (eqv?-hash O1)
     (eqv?-hash O2))'.

     For example:

          > (eqv?-hash 1.5)
          496387656


 -- procedure: equal?-hash OBJ
     The `equal?-hash' procedure returns the hash number of the object
     OBJ.  The hash number is a small exact integer (fixnum).  For any
     two objects O1 and O2, `(equal? O1 O2)' implies `(= (equal?-hash
     O1) (equal?-hash O2))'.

     For example:

          > (equal?-hash (list 1 2 3))
          442438567


8.10.2 Weak references
----------------------

The garbage collector is responsible for reclaiming objects that are no
longer needed by the program.  This is done by analyzing the
reachability graph of all objects from the roots (i.e. the global
variables, the runnable threads, permanently allocated objects such as
procedures defined in a compiled file, nonexecutable wills, etc).  If a
root or a reachable object X contains a reference to an object Y then Y
is reachable.  As a general rule, unreachable objects are reclaimed by
the garbage collector.

   There are two types of references: strong references and weak
references.  Most objects, including pairs, vectors, records and
closures, contain strong references.  An object X is "strongly
reachable" if there is a path from the roots to X that traverses only
strong references.  Weak references only occur in wills and tables.
There are two types of weak references: will-weak references and
table-weak references.  If all paths from the roots to an object Y
traverse at least one table-weak reference, then Y will be reclaimed by
the garbage collector.  The will-weak references are used for
finalization and are explained in the next section.

8.10.2.1 Wills
..............

The following procedures implement the "will" data type.  Will objects
provide support for finalization.  A will is an object that contains a
will-weak reference to a TESTATOR object (the object attached to the
will), and a strong reference to an ACTION procedure which is a one
parameter procedure which is called when the will is executed.

 -- procedure: make-will TESTATOR ACTION
 -- procedure: will? OBJ
 -- procedure: will-testator WILL
 -- procedure: will-execute! WILL
     The `make-will' procedure creates a will object with the given
     TESTATOR object and ACTION procedure.  The `will?' procedure tests
     if OBJ is a will object.  The `will-testator' procedure gets the
     testator object attached to the WILL.  The `will-execute!'
     procedure executes WILL.

     A will becomes "executable" when its TESTATOR object is not
     strongly reachable (i.e. the TESTATOR object is either unreachable
     or only reachable using paths from the roots that traverse at
     least one weak reference).  Some objects, including symbols, small
     exact integers (fixnums), booleans and characters, are considered
     to be always strongly reachable.

     When the runtime system detects that a will has become executable
     the current computation is interrupted, the will's testator is set
     to `#f' and the will's action procedure is called with the will's
     testator as the sole argument.  Currently only the garbage
     collector detects when wills become executable but this may change
     in future versions of Gambit (for example the compiler could
     perform an analysis to infer will executability at compile time).
     The garbage collector builds a list of all executable wills.
     Shortly after a garbage collection, the action procedures of these
     wills will be called.  The link from the will to the action
     procedure is severed when the action procedure is called.

     Note that the testator object will not be reclaimed during the
     garbage collection that determined executability of the will.  It
     is only when an object is not reachable from the roots that it is
     reclaimed by the garbage collector.

     A remarkable feature of wills is that an action procedure can
     "resurrect" an object.  An action procedure could for example
     assign the testator object to a global variable or create a new
     will with the same testator object.

     For example:

          > (define a (list 123))
          > (set-cdr! a a) ; create a circular list
          > (define b (vector a))
          > (define c #f)
          > (define w
              (let ((obj a))
                (make-will obj
                           (lambda (x) ; x will be eq? to obj
                             (display "executing action procedure")
                             (newline)
                             (set! c x)))))
          > (will? w)
          #t
          > (car (will-testator w))
          123
          > (##gc)
          > (set! a #f)
          > (##gc)
          > (set! b #f)
          > (##gc)
          executing action procedure
          > (will-testator w)
          #f
          > (car c)
          123


8.10.3 Tables
-------------

The following procedures implement the "table" data type.  Tables are
heterogenous structures whose elements are indexed by keys which are
arbitrary objects.  Tables are similar to association lists but are
abstract and the access time for large tables is typically smaller.
Each key contained in the table is bound to a value.  The length of the
table is the number of key/value bindings it contains.  New key/value
bindings can be added to a table, the value bound to a key can be
changed, and existing key/value bindings can be removed.

   The references to the keys can either be all strong or all table-weak
and the references to the values can either be all strong or all
table-weak.  The garbage collector removes key/value bindings from a
table when 1) the key is a table-weak reference and the key is
unreachable or only reachable using paths from the roots that traverse
at least one table-weak reference, or 2) the value is a table-weak
reference and the value is unreachable or only reachable using paths
from the roots that traverse at least one table-weak reference.
Key/value bindings that are removed by the garbage collector are
reclaimed immediately.

   Although there are several possible ways of implementing tables, the
current implementation uses hashing with open-addressing.  This is
space efficient and provides constant-time access.  Hash tables are
automatically resized to maintain the load within specified bounds.
The load is the number of active entries (the length of the table)
divided by the total number of entries in the hash table.

   Tables are parameterized with a key comparison procedure.  By default
the `equal?' procedure is used, but `eq?', `eqv?', `string=?',
`string-ci=?', or a user defined procedure can also be used.  To
support arbitrary key comparison procedures, tables are also
parameterized with a hashing procedure accepting a key as its single
parameter and returning a fixnum result.  The hashing procedure HASH
must be consistent with the key comparison procedure TEST, that is, for
any two keys K1 and K2 in the table, `(TEST K1 K2)' implies `(= (HASH
K1) (HASH K2))'.  A default hashing procedure consistent with the key
comparison procedure is provided by the system.  The default hashing
procedure generally gives good performance when the key comparison
procedure is `eq?', `eqv?', `equal?', `string=?', and `string-ci=?'.
However, for user defined key comparison procedures, the default
hashing procedure always returns 0.  This degrades the performance of
the table to a linear search.

   Tables can be compared for equality using the `equal?' procedure.
Two tables `X' and `Y' are considered equal by `equal?' when they have
the same weakness attributes, the same key comparison procedure, the
same hashing procedure, the same length, and for all the keys `K' in
`X', `(equal?  (table-ref X K) (table-ref Y K))'.

 -- procedure: make-table [`size:' SIZE] [`init:' INIT] [`weak-keys:'
          WEAK-KEYS] [`weak-values:' WEAK-VALUES] [`test:' TEST]
          [`hash:' HASH] [`min-load:' MIN-LOAD] [`max-load:' MAX-LOAD]
     The procedure `make-table' returns a new table.  The optional
     keyword parameters specify various parameters of the table.

     The SIZE parameter is a nonnegative exact integer indicating the
     expected length of the table.  The system uses SIZE to choose an
     appropriate initial size of the hash table so that it does not
     need to be resized too often.

     The INIT parameter indicates a value that is associated to keys
     that are not in the table.  When INIT is not specified, no value
     is associated to keys that are not in the table.

     The WEAK-KEYS and WEAK-VALUES parameters are extended booleans
     indicating respectively whether the keys and values are table-weak
     references (true) or strong references (false).  By default the
     keys and values are strong references.

     The TEST parameter indicates the key comparison procedure.  The
     default key comparison procedure is `equal?'.  The key comparison
     procedures `eq?', `eqv?', `equal?', `string=?', and `string-ci=?'
     are special because the system will use a reasonably good hash
     procedure when none is specified.

     The HASH parameter indicates the hash procedure.  This procedure
     must accept a single key parameter, return a fixnum, and be
     consistent with the key comparison procedure.  When HASH is not
     specified, a default hash procedure is used.  The default hash
     procedure is reasonably good when the key comparison procedure is
     `eq?', `eqv?', `equal?', `string=?', or `string-ci=?'.

     The MIN-LOAD and MAX-LOAD parameters are real numbers that
     indicate the minimum and maximum load of the table respectively.
     The table is resized when adding or deleting a key/value binding
     would bring the table's load outside of this range.  The MIN-LOAD
     parameter must be no less than 0.05 and the MAX-LOAD parameter
     must be no greater than 0.95.  Moreover the difference between
     MIN-LOAD and MAX-LOAD must be at least 0.20.  When MIN-LOAD is not
     specified, the value 0.45 is used.  When MAX-LOAD is not
     specified, the value 0.90 is used.

     For example:

          > (define t (make-table))
          > (table? t)
          #t
          > (table-length t)
          0
          > (table-set! t (list 1 2) 3)
          > (table-set! t (list 4 5) 6)
          > (table-ref t (list 1 2))
          3
          > (table-length t)
          2


 -- procedure: table? OBJ
     The procedure `table?' returns `#t' when OBJ is a table and `#f'
     otherwise.

     For example:

          > (table? (make-table))
          #t
          > (table? 123)
          #f


 -- procedure: table-length TABLE
     The procedure `table-length' returns the number of key/value
     bindings contained in the table TABLE.

     For example:

          > (define t (make-table weak-keys: #t))
          > (define x (list 1 2))
          > (define y (list 3 4))
          > (table-set! t x 111)
          > (table-set! t y 222)
          > (table-length t)
          2
          > (table-set! t x)
          > (table-length t)
          1
          > (##gc)
          > (table-length t)
          1
          > (set! y #f)
          > (##gc)
          > (table-length t)
          0


 -- procedure: table-ref TABLE KEY [DEFAULT]
     The procedure `table-ref' returns the value bound to the object
     KEY in the table TABLE.  When KEY is not bound and DEFAULT is
     specified, DEFAULT is returned.  When DEFAULT is not specified but
     an INIT parameter was specified when TABLE was created, INIT is
     returned.  Otherwise an unbound-key-exception object is raised.

     For example:

          > (define t1 (make-table init: 999))
          > (table-set! t1 (list 1 2) 3)
          > (table-ref t1 (list 1 2))
          3
          > (table-ref t1 (list 4 5))
          999
          > (table-ref t1 (list 4 5) #f)
          #f
          > (define t2 (make-table))
          > (table-ref t2 (list 4 5))
          *** ERROR IN (console)@7.1 -- Unbound key
          (table-ref '#<table #2> '(4 5))


 -- procedure: table-set! TABLE KEY [VALUE]
     The procedure `table-set!' binds the object KEY to VALUE in the
     table TABLE.  When VALUE is not specified, if TABLE contains a
     binding for KEY then the binding is removed from TABLE.  The
     procedure `table-set!' returns an unspecified value.

     For example:

          > (define t (make-table))
          > (table-set! t (list 1 2) 3)
          > (table-set! t (list 4 5) 6)
          > (table-set! t (list 4 5))
          > (table-set! t (list 7 8))
          > (table-ref t (list 1 2))
          3
          > (table-ref t (list 4 5))
          *** ERROR IN (console)@7.1 -- Unbound key
          (table-ref '#<table #2> '(4 5))


 -- procedure: table-search PROC TABLE
     The procedure `table-search' searches the table TABLE for a
     key/value binding for which the two parameter procedure PROC
     returns a non false result.  For each key/value binding visited by
     `table-search' the procedure PROC is called with the key as the
     first parameter and the value as the second parameter.  The
     procedure `table-search' returns the first non false value
     returned by PROC, or `#f' if PROC returned `#f' for all key/value
     bindings in TABLE.

     The order in which the key/value bindings are visited is
     unspecified and may vary from one call of `table-search' to the
     next.  While a call to `table-search' is being performed on TABLE,
     it is an error to call any of the following procedures on TABLE:
     `table-ref', `table-set!', `table-search', `table-for-each',
     `table-copy', `table-merge', `table-merge!', and `table->list'.
     It is also an error to compare with `equal?' (directly or
     indirectly with `member', `assoc', `table-ref', etc.) an object
     that contains TABLE.  All these procedures may cause TABLE to be
     reordered and resized.  This restriction allows a more efficient
     iteration over the key/value bindings.

     For example:

          > (define square (make-table))
          > (table-set! square 2 4)
          > (table-set! square 3 9)
          > (table-search (lambda (k v) (and (odd? k) v)) square)
          9


 -- procedure: table-for-each PROC TABLE
     The procedure `table-for-each' calls the two parameter procedure
     PROC for each key/value binding in the table TABLE.  The procedure
     PROC is called with the key as the first parameter and the value
     as the second parameter.  The procedure `table-for-each' returns
     an unspecified value.

     The order in which the key/value bindings are visited is
     unspecified and may vary from one call of `table-for-each' to the
     next.  While a call to `table-for-each' is being performed on
     TABLE, it is an error to call any of the following procedures on
     TABLE: `table-ref', `table-set!', `table-search',
     `table-for-each', and `table->list'.  It is also an error to
     compare with `equal?' (directly or indirectly with `member',
     `assoc', `table-ref', etc.) an object that contains TABLE.  All
     these procedures may cause TABLE to be reordered and resized.
     This restriction allows a more efficient iteration over the
     key/value bindings.

     For example:

          > (define square (make-table))
          > (table-set! square 2 4)
          > (table-set! square 3 9)
          > (table-for-each (lambda (k v) (write (list k v)) (newline)) square)
          (2 4)
          (3 9)


 -- procedure: table->list TABLE
     The procedure `table->list' returns an association list containing
     the key/value bindings in the table TABLE.  Each key/value binding
     yields a pair whose car field is the key and whose cdr field is
     the value bound to that key.  The order of the bindings in the
     list is unspecified.

     For example:

          > (define square (make-table))
          > (table-set! square 2 4)
          > (table-set! square 3 9)
          > (table->list square)
          ((3 . 9) (2 . 4))


 -- procedure: list->table LIST [`size:' SIZE] [`init:' INIT]
          [`weak-keys:' WEAK-KEYS] [`weak-values:' WEAK-VALUES]
          [`test:' TEST] [`hash:' HASH] [`min-load:' MIN-LOAD]
          [`max-load:' MAX-LOAD]
     The procedure `list->table' returns a new table containing the
     key/value bindings in the association list LIST.  The optional
     keyword parameters specify various parameters of the table and have
     the same meaning as for the `make-table' procedure.

     Each element of LIST is a pair whose car field is a key and whose
     cdr field is the value bound to that key.  If a key appears more
     than once in LIST (tested using the table's key comparison
     procedure) it is the first key/value binding in LIST that has
     precedence.

     For example:

          > (define t (list->table '((b . 2) (a . 1) (c . 3) (a . 4))))
          > (table->list t)
          ((a . 1) (b . 2) (c . 3))


 -- procedure: unbound-key-exception? OBJ
 -- procedure: unbound-key-exception-procedure EXC
 -- procedure: unbound-key-exception-arguments EXC
     Unbound-key-exception objects are raised by the procedure
     `table-ref' when the key does not have a binding in the table.
     The parameter EXC must be an unbound-key-exception object.

     The procedure `unbound-key-exception?' returns `#t' when OBJ is a
     unbound-key-exception object and `#f' otherwise.

     The procedure `unbound-key-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `unbound-key-exception-arguments' returns the list
     of arguments of the procedure that raised EXC.

     For example:

          > (define t (make-table))
          > (define (handler exc)
              (if (unbound-key-exception? exc)
                  (list (unbound-key-exception-procedure exc)
                        (unbound-key-exception-arguments exc))
                  'not-unbound-key-exception))
          > (with-exception-catcher
              handler
              (lambda () (table-ref t '(1 2))))
          (#<procedure #2 table-ref> (#<table #3> (1 2)))


 -- procedure: table-copy TABLE
     The procedure `table-copy' returns a new table containing the same
     key/value bindings as TABLE and the same table parameters (i.e.
     hash procedure, key comparison procedure, key and value weakness,
     etc).

     For example:

          > (define t (list->table '((b . 2) (a . 1) (c . 3))))
          > (define x (table-copy t))
          > (table-set! t 'b 99)
          > (table->list t)
          ((a . 1) (b . 99) (c . 3))
          > (table->list x)
          ((a . 1) (b . 2) (c . 3))


 -- procedure: table-merge! TABLE1 TABLE2 [TABLE2-TAKES-PRECEDENCE?]
     The procedure `table-merge!' returns TABLE1 after the key/value
     bindings contained in TABLE2 have been added to it.  When a key
     exists both in TABLE1 and TABLE2, then the parameter
     TABLE2-TAKES-PRECEDENCE? indicates which binding will be kept (the
     one in TABLE1 if TABLE2-TAKES-PRECEDENCE?  is false, and the one
     in TABLE2 otherwise).  If TABLE2-TAKES-PRECEDENCE? is not
     specified the binding in TABLE1 is kept.

     For example:

          > (define t1 (list->table '((a . 1) (b . 2) (c . 3))))
          > (define t2 (list->table '((a . 4) (b . 5) (z . 6))))
          > (table->list (table-merge! t1 t2))
          ((a . 1) (b . 2) (c . 3) (z . 6))
          > (define t1 (list->table '((a . 1) (b . 2) (c . 3))))
          > (define t2 (list->table '((a . 4) (b . 5) (z . 6))))
          > (table->list (table-merge! t1 t2 #t))
          ((a . 4) (b . 5) (c . 3) (z . 6))


 -- procedure: table-merge TABLE1 TABLE2 [TABLE2-TAKES-PRECEDENCE?]
     The procedure `table-merge' returns a copy of TABLE1 (created with
     `table-copy') to which the key/value bindings contained in TABLE2
     have been added using `table-merge!'.  When a key exists both in
     TABLE1 and TABLE2, then the parameter TABLE2-TAKES-PRECEDENCE?
     indicates which binding will be kept (the one in TABLE1 if
     TABLE2-TAKES-PRECEDENCE?  is false, and the one in TABLE2
     otherwise).  If TABLE2-TAKES-PRECEDENCE? is not specified the
     binding in TABLE1 is kept.

     For example:

          > (define t1 (list->table '((a . 1) (b . 2) (c . 3))))
          > (define t2 (list->table '((a . 4) (b . 5) (z . 6))))
          > (table->list (table-merge t1 t2))
          ((a . 1) (b . 2) (c . 3) (z . 6))
          > (table->list (table-merge t1 t2 #t))
          ((a . 4) (b . 5) (c . 3) (z . 6))


9 Records
*********

 -- special form: define-structure name field...
     Record data types similar to Pascal records and C `struct' types
     can be defined using the `define-structure' special form.  The
     identifier name specifies the name of the new data type.  The
     structure name is followed by K identifiers naming each field of
     the record.  The `define-structure' expands into a set of
     definitions of the following procedures:

        * `make-name' - A K argument procedure which constructs a new
          record from the value of its K fields.

        * `name?' - A procedure which tests if its single argument is
          of the given record type.

        * `name-field' - For each field, a procedure taking as its
          single argument a value of the given record type and returning
          the content of the corresponding field of the record.

        * `name-field-set!' - For each field, a two argument procedure
          taking as its first argument a value of the given record
          type.  The second argument gets assigned to the corresponding
          field of the record and the void object is returned.


     Record data types have a printed representation that includes the
     name of the type and the name and value of each field.  Record
     data types can not be read by the `read' procedure.

     For example:

          > (define-structure point x y color)
          > (define p (make-point 3 5 'red))
          > p
          #<point #2 x: 3 y: 5 color: red>
          > (point-x p)
          3
          > (point-color p)
          red
          > (point-color-set! p 'black)
          > p
          #<point #2 x: 3 y: 5 color: black>


10 Threads
**********

Gambit supports the execution of multiple Scheme threads.  These
threads are managed entirely by Gambit's runtime and are not related to
the host operating system's threads.  Gambit's runtime does not
currently take advantage of multiprocessors (i.e. at most one thread is
running).

10.1 Introduction
=================

Multithreading is a paradigm that is well suited for building complex
systems such as: servers, GUIs, and high-level operating systems.
Gambit's thread system offers mechanisms for creating threads of
execution and for synchronizing them.  The thread system also supports
features which are useful in a real-time context, such as priorities,
priority inheritance and timeouts.

   The thread system provides the following data types:

   * Thread (a virtual processor which shares object space with all
     other threads)

   * Mutex (a mutual exclusion device, also known as a lock and binary
     semaphore)

   * Condition variable (a set of blocked threads)


10.2 Thread objects
===================

A "running thread" is a thread that is currently executing.  A
"runnable thread" is a thread that is ready to execute or running.  A
thread is "blocked" if it is waiting for a mutex to become unlocked, an
I/O operation to become possible, the end of a "sleep" period, etc.  A
"new thread" is a thread that has been allocated but has not yet been
initialized.  An "initialized thread" is a thread that can be made
runnable.  A new thread becomes runnable when it is started by calling
`thread-start!'.  A "terminated thread" is a thread that can no longer
become runnable (but "deadlocked threads" are not considered
terminated).  The only valid transitions between the thread states are
from new to initialized, from initialized to runnable, between runnable
and blocked, and from any state except new to terminated as indicated in
the following diagram:

                                                 unblock
                               start            <-------
     NEW -------> INITIALIZED -------> RUNNABLE -------> BLOCKED
                            \             |      block  /
                             \            v            /
                              +-----> TERMINATED <----+

   Each thread has a "base priority", which is a real number (where a
higher numerical value means a higher priority), a "priority boost",
which is a nonnegative real number representing the priority increase
applied to a thread when it blocks, and a "quantum", which is a
nonnegative real number representing a duration in seconds.

   Each thread has a "specific field" which can be used in an
application specific way to associate data with the thread (some thread
systems call this "thread local storage").

   Each thread has a "mailbox" which is used for inter-thread
communication.

10.3 Mutex objects
==================

A mutex can be in one of four states: "locked" (either "owned" or "not
owned") and "unlocked" (either "abandoned" or "not abandoned").

   An attempt to lock a mutex only succeeds if the mutex is in an
unlocked state, otherwise the current thread will wait.  A mutex in the
locked/owned state has an associated "owner thread", which by
convention is the thread that is responsible for unlocking the mutex
(this case is typical of critical sections implemented as "lock mutex,
perform operation, unlock mutex").  A mutex in the locked/not-owned
state is not linked to a particular thread.

   A mutex becomes locked when a thread locks it using the
`mutex-lock!' primitive.  A mutex becomes unlocked/abandoned when the
owner of a locked/owned mutex terminates.  A mutex becomes
unlocked/not-abandoned when a thread unlocks it using the
`mutex-unlock!' primitive.

   The mutex primitives do not implement "recursive mutex semantics".
An attempt to lock a mutex that is locked implies that the current
thread waits even if the mutex is owned by the current thread (this can
lead to a deadlock if no other thread unlocks the mutex).

   Each mutex has a "specific field" which can be used in an
application specific way to associate data with the mutex.

10.4 Condition variable objects
===============================

A condition variable represents a set of blocked threads.  These blocked
threads are waiting for a certain condition to become true.  When a
thread modifies some program state that might make the condition true,
the thread unblocks some number of threads (one or all depending on the
primitive used) so they can check if the condition is now true.  This
allows complex forms of interthread synchronization to be expressed more
conveniently than with mutexes alone.

   Each condition variable has a "specific field" which can be used in
an application specific way to associate data with the condition
variable.

10.5 Fairness
=============

In various situations the scheduler must select one thread from a set of
threads (e.g. which thread to run when a running thread blocks or
expires its quantum, which thread to unblock when a mutex becomes
unlocked or a condition variable is signaled).  The constraints on the
selection process determine the scheduler's "fairness".  The selection
depends on the order in which threads become runnable or blocked and on
the "priority" attached to the threads.

   The definition of fairness requires the notion of time ordering,
i.e. "event A occured before event B".  For the purpose of establishing
time ordering, the scheduler uses a clock with a discrete, usually
variable, resolution (a "tick").  Events occuring in a given tick can
be considered to be simultaneous (i.e. if event A occured before event
B in real time, then the scheduler will claim that event A occured
before event B unless both events fall within the same tick, in which
case the scheduler arbitrarily chooses a time ordering).

   Each thread T has three priorities which affect fairness; the "base
priority", the "boosted priority", and the "effective priority".

   * The "base priority" is the value contained in T's "base priority"
     field (which is set with the `thread-base-priority-set!'
     primitive).

   * T's "boosted flag" field contains a boolean that affects T's
     "boosted priority".  When the boosted flag field is false, the
     boosted priority is equal to the base priority, otherwise the
     boosted priority is equal to the base priority plus the value
     contained in T's "priority boost" field (which is set with the
     `thread-priority-boost-set!' primitive).  The boosted flag field is
     set to false when a thread is created, when its quantum expires,
     and when "thread-yield!" is called.  The boosted flag field is set
     to true when a thread blocks.  By carefully choosing the base
     priority and priority boost, relatively to the other threads, it
     is possible to set up an interactive thread so that it has good
     I/O response time without being a CPU hog when it performs long
     computations.

   * The "effective priority" is equal to the maximum of T's boosted
     priority and the effective priority of all the threads that are
     blocked on a mutex owned by T.  This "priority inheritance" avoids
     priority inversion problems that would prevent a high priority
     thread blocked at the entry of a critical section to progress
     because a low priority thread inside the critical section is
     preempted for an arbitrary long time by a medium priority thread.


   Let P(T) be the effective priority of thread T and let R(T) be the
most recent time when one of the following events occurred for thread
T, thus making it runnable: T was started by calling `thread-start!', T
called `thread-yield!', T expired its quantum, or T became unblocked.
Let the relation NL(T1,T2), "T1 no later than T2", be true if
P(T1)<P(T2) or P(T1)=P(T2) and R(T1)>R(T2), and false otherwise.  The
scheduler will schedule the execution of threads in such a way that
whenever there is at least one runnable thread, 1) within a finite time
at least one thread will be running, and 2) there is never a pair of
runnable threads T1 and T2 for which NL(T1,T2) is true and T1 is not
running and T2 is running.

   A thread T expires its quantum when an amount of time equal to T's
quantum has elapsed since T entered the running state and T did not
block, terminate or call `thread-yield!'.  At that point T exits the
running state to allow other threads to run.  A thread's quantum is
thus an indication of the rate of progress of the thread relative to
the other threads of the same priority.  Moreover, the resolution of
the timer measuring the running time may cause a certain deviation from
the quantum, so a thread's quantum should only be viewed as an
approximation of the time it can run before yielding to another thread.

   Threads blocked on a given mutex or condition variable will unblock
in an order which is consistent with decreasing priority and increasing
blocking time (i.e. the highest priority thread unblocks first, and
among equal priority threads the one that blocked first unblocks first).

10.6 Memory coherency
=====================

Read and write operations on the store (such as reading and writing a
variable, an element of a vector or a string) are not atomic.  It is an
error for a thread to write a location in the store while some other
thread reads or writes that same location.  It is the responsibility of
the application to avoid write/read and write/write races through
appropriate uses of the synchronization primitives.

   Concurrent reads and writes to ports are allowed.  It is the
responsibility of the implementation to serialize accesses to a given
port using the appropriate synchronization primitives.

10.7 Timeouts
=============

All synchronization primitives which take a timeout parameter accept
three types of values as a timeout, with the following meaning:

   * a time object represents an absolute point in time

   * an exact or inexact real number represents a relative time in
     seconds from the moment the primitive was called

   * `#f' means that there is no timeout


   When a timeout denotes the current time or a time in the past, the
synchronization primitive claims that the timeout has been reached only
after the other synchronization conditions have been checked.  Moreover
the thread remains running (it does not enter the blocked state).  For
example, `(mutex-lock! m 0)' will lock mutex `m' and return `#t' if `m'
is currently unlocked, otherwise `#f' is returned because the timeout
is reached.

10.8 Primordial thread
======================

The execution of a program is initially under the control of a single
thread known as the "primordial thread".  The primordial thread has an
unspecified base priority, priority boost, boosted flag, quantum, name,
specific field, dynamic environment, `dynamic-wind' stack, and
exception-handler.  All threads are terminated when the primordial
thread terminates (normally or not).

10.9 Procedures
===============

 -- procedure: current-thread
     This procedure returns the current thread.  For example:

          > (current-thread)
          #<thread #1 primordial>
          > (eq? (current-thread) (current-thread))
          #t


 -- procedure: thread? OBJ
     This procedure returns `#t' when OBJ is a thread object and `#f'
     otherwise.

     For example:

          > (thread? (current-thread))
          #t
          > (thread? 'foo)
          #f


 -- procedure: make-thread THUNK [NAME [THREAD-GROUP]]
 -- procedure: make-root-thread THUNK [NAME [THREAD-GROUP [INPUT-PORT
          [OUTPUT-PORT]]]]
     The `make-thread' procedure creates and returns an initialized
     thread.  This thread is not automatically made runnable (the
     procedure `thread-start!' must be used for this).  A thread has the
     following fields: base priority, priority boost, boosted flag,
     quantum, name, specific, end-result, end-exception, and a list of
     locked/owned mutexes it owns.  The thread's execution consists of a
     call to THUNK with the "initial continuation".  This continuation
     causes the (then) current thread to store the result in its
     end-result field, abandon all mutexes it owns, and finally
     terminate.  The `dynamic-wind' stack of the initial continuation
     is empty.  The optional NAME is an arbitrary Scheme object which
     identifies the thread (useful for debugging); it defaults to an
     unspecified value.  The specific field is set to an unspecified
     value.  The optional THREAD-GROUP indicates which thread group this
     thread belongs to; it defaults to the thread group of the current
     thread.  The base priority, priority boost, and quantum of the
     thread are set to the same value as the current thread and the
     boosted flag is set to false.  The thread's mailbox is initially
     empty.  The thread inherits the dynamic environment from the
     current thread. Moreover, in this dynamic environment the
     exception-handler is bound to the "initial exception-handler"
     which is a unary procedure which causes the (then) current thread
     to store in its end-exception field an uncaught-exception object
     whose "reason" is the argument of the handler, abandon all mutexes
     it owns, and finally terminate.

     The `make-root-thread' procedure behaves like the `make-thread'
     procedure except the created thread does not inherit the dynamic
     environment from the current thread and the base priority is set
     to 0, the priority boost is set to 1.0e-6, and the quantum is set
     to 0.02.  The dynamic environment of the thread has the global
     bindings of the parameter objects, except `current-input-port'
     which is bound to INPUT-PORT, `current-output-port' which is bound
     to OUTPUT-PORT, and `current-directory' which is bound to the
     initial current working directory of the current process.  If
     INPUT-PORT is not specified it defaults to a port corresponding to
     the standard input (`stdin').  If OUTPUT-PORT is not specified it
     defaults to a port corresponding to the standard output (`stdout').

     For example:

          > (make-thread (lambda () (write 'hello)))
          #<thread #2>
          > (make-root-thread (lambda () (write 'world)) 'a-name)
          #<thread #3 a-name>


 -- procedure: thread THUNK
     The `thread' procedure creates, starts and returns a new thread.
     The call `(thread THUNK)' is equivalent to `(thread-start!
     (make-thread THUNK))'.

     For example:

          > (define a (thread (lambda () (expt 2 1005))))
          > (define b (thread (lambda () (expt 2 1000))))
          > (/ (thread-join! a) (thread-join! b))
          32


 -- procedure: thread-name THREAD
     This procedure returns the name of the THREAD.  For example:

          > (thread-name (make-thread (lambda () #f) 'foo))
          foo


 -- procedure: thread-specific THREAD
 -- procedure: thread-specific-set! THREAD OBJ
     The `thread-specific' procedure returns the content of the
     THREAD's specific field.

     The `thread-specific-set!' procedure stores OBJ into the THREAD's
     specific field and returns an unspecified value.

     For example:

          > (thread-specific-set! (current-thread) "hello")
          > (thread-specific (current-thread))
          "hello"


 -- procedure: thread-base-priority THREAD
 -- procedure: thread-base-priority-set! THREAD PRIORITY
     The procedure `thread-base-priority' returns a real number which
     corresponds to the base priority of the THREAD.

     The procedure `thread-base-priority-set!' changes the base
     priority of the THREAD to PRIORITY and returns an unspecified
     value.  The PRIORITY must be a real number.

     For example:

          > (thread-base-priority-set! (current-thread) 12.3)
          > (thread-base-priority (current-thread))
          12.3


 -- procedure: thread-priority-boost THREAD
 -- procedure: thread-priority-boost-set! THREAD PRIORITY-BOOST
     The procedure `thread-priority-boost' returns a real number which
     corresponds to the priority boost of the THREAD.

     The procedure `thread-priority-boost-set!' changes the priority
     boost of the THREAD to PRIORITY-BOOST and returns an unspecified
     value.  The PRIORITY-BOOST must be a nonnegative real.

     For example:

          > (thread-priority-boost-set! (current-thread) 2.5)
          > (thread-priority-boost (current-thread))
          2.5


 -- procedure: thread-quantum THREAD
 -- procedure: thread-quantum-set! THREAD QUANTUM
     The procedure `thread-quantum' returns a real number which
     corresponds to the quantum of the THREAD.

     The procedure `thread-quantum-set!' changes the quantum of the
     THREAD to QUANTUM and returns an unspecified value.  The QUANTUM
     must be a nonnegative real.  A value of zero selects the smallest
     quantum supported by the implementation.

     For example:

          > (thread-quantum-set! (current-thread) 1.5)
          > (thread-quantum (current-thread))
          1.5
          > (thread-quantum-set! (current-thread) 0)
          > (thread-quantum (current-thread))
          0.


 -- procedure: thread-start! THREAD
     This procedure makes THREAD runnable and returns the THREAD.  The
     THREAD must be an initialized thread.

     For example:

          > (let ((t (thread-start! (make-thread (lambda () (write 'a))))))
              (write 'b)
              (thread-join! t))
          ab> or ba>

     NOTE: It is useful to separate thread creation and thread
     activation to avoid the race condition that would occur if the
     created thread tries to examine a table in which the current
     thread stores the created thread.  See the last example of the
     `thread-terminate!' procedure which contains mutually recursive
     threads.


 -- procedure: thread-yield!
     This procedure causes the current thread to exit the running state
     as if its quantum had expired and returns an unspecified value.

     For example:

          ; a busy loop that avoids being too wasteful of the CPU

          (let loop ()
            (if (mutex-lock! m 0) ; try to lock m but don't block
                (begin
                  (display "locked mutex m")
                  (mutex-unlock! m))
                (begin
                  (do-something-else)
                  (thread-yield!) ; relinquish rest of quantum
                  (loop))))


 -- procedure: thread-sleep! TIMEOUT
     This procedure causes the current thread to wait until the timeout
     is reached and returns an unspecified value.  This blocks the
     thread only if TIMEOUT represents a point in the future.  It is an
     error for TIMEOUT to be `#f'.

     For example:

          ; a clock with a gradual drift:

          (let loop ((x 1))
            (thread-sleep! 1)
            (write x)
            (loop (+ x 1)))

          ; a clock with no drift:

          (let ((start (time->seconds (current-time)))
            (let loop ((x 1))
              (thread-sleep! (seconds->time (+ x start)))
              (write x)
              (loop (+ x 1))))


 -- procedure: thread-terminate! THREAD
     This procedure causes an abnormal termination of the THREAD.  If
     the THREAD is not already terminated, all mutexes owned by the
     THREAD become unlocked/abandoned and a terminated-thread-exception
     object is stored in the THREAD's end-exception field.  If THREAD
     is the current thread, `thread-terminate!' does not return.
     Otherwise `thread-terminate!' returns an unspecified value; the
     termination of the THREAD will occur at some point between the
     calling of `thread-terminate!'  and a finite time in the future
     (an explicit thread synchronization is needed to detect
     termination, see `thread-join!').

     For example:

          (define (amb thunk1 thunk2)
            (let ((result #f)
                  (result-mutex (make-mutex))
                  (done-mutex (make-mutex)))
              (letrec ((child1
                        (make-thread
                          (lambda ()
                            (let ((x (thunk1)))
                              (mutex-lock! result-mutex #f #f)
                              (set! result x)
                              (thread-terminate! child2)
                              (mutex-unlock! done-mutex)))))
                       (child2
                        (make-thread
                          (lambda ()
                            (let ((x (thunk2)))
                              (mutex-lock! result-mutex #f #f)
                              (set! result x)
                              (thread-terminate! child1)
                              (mutex-unlock! done-mutex))))))
                (mutex-lock! done-mutex #f #f)
                (thread-start! child1)
                (thread-start! child2)
                (mutex-lock! done-mutex #f #f)
                result)))

     NOTE: This operation must be used carefully because it terminates a
     thread abruptly and it is impossible for that thread to perform any
     kind of cleanup.  This may be a problem if the thread is in the
     middle of a critical section where some structure has been put in
     an inconsistent state.  However, another thread attempting to
     enter this critical section will raise an
     abandoned-mutex-exception object because the mutex is
     unlocked/abandoned.  This helps avoid observing an inconsistent
     state.  Clean termination can be obtained by polling, as shown in
     the example below.

     For example:

          (define (spawn thunk)
            (let ((t (make-thread thunk)))
              (thread-specific-set! t #t)
              (thread-start! t)
              t))

          (define (stop! thread)
            (thread-specific-set! thread #f)
            (thread-join! thread))

          (define (keep-going?)
            (thread-specific (current-thread)))

          (define count!
            (let ((m (make-mutex))
                  (i 0))
              (lambda ()
                (mutex-lock! m)
                (let ((x (+ i 1)))
                  (set! i x)
                  (mutex-unlock! m)
                  x))))

          (define (increment-forever!)
            (let loop () (count!) (if (keep-going?) (loop))))

          (let ((t1 (spawn increment-forever!))
                (t2 (spawn increment-forever!)))
            (thread-sleep! 1)
            (stop! t1)
            (stop! t2)
            (count!))  ==>  377290


 -- procedure: thread-join! thread [TIMEOUT [TIMEOUT-VAL]]
     This procedure causes the current thread to wait until the THREAD
     terminates (normally or not) or until the timeout is reached if
     TIMEOUT is supplied.  If the timeout is reached, THREAD-JOIN!
     returns TIMEOUT-VAL if it is supplied, otherwise a
     join-timeout-exception object is raised.  If the THREAD terminated
     normally, the content of the end-result field is returned,
     otherwise the content of the end-exception field is raised.

     For example:

          (let ((t (thread-start! (make-thread (lambda () (expt 2 100))))))
            (do-something-else)
            (thread-join! t))  ==>  1267650600228229401496703205376

          (let ((t (thread-start! (make-thread (lambda () (raise 123))))))
            (do-something-else)
            (with-exception-handler
              (lambda (exc)
                (if (uncaught-exception? exc)
                    (* 10 (uncaught-exception-reason exc))
                    99999))
              (lambda ()
                (+ 1 (thread-join! t)))))  ==>  1231

          (define thread-alive?
            (let ((unique (list 'unique)))
              (lambda (thread)
                ; Note: this procedure raises an exception if
                ; the thread terminated abnormally.
                (eq? (thread-join! thread 0 unique) unique))))

          (define (wait-for-termination! thread)
            (let ((eh (current-exception-handler)))
              (with-exception-handler
                (lambda (exc)
                  (if (not (or (terminated-thread-exception? exc)
                               (uncaught-exception? exc)))
                      (eh exc))) ; unexpected exceptions are handled by eh
                (lambda ()
                  ; The following call to thread-join! will wait until the
                  ; thread terminates.  If the thread terminated normally
                  ; thread-join! will return normally.  If the thread
                  ; terminated abnormally then one of these two exception
                  ; objects is raised by thread-join!:
                  ;   - terminated-thread-exception object
                  ;   - uncaught-exception object
                  (thread-join! thread)
                  #f)))) ; ignore result of thread-join!


 -- procedure: thread-send THREAD MSG
     Each thread has a mailbox which stores messages delivered to the
     thread in the order delivered.

     The procedure `thread-send' adds the message MSG at the end of the
     mailbox of thread THREAD and returns an unspecified value.

     For example:

          > (thread-send (current-thread) 111)
          > (thread-send (current-thread) 222)
          > (thread-receive)
          111
          > (thread-receive)
          222


 -- procedure: thread-receive [TIMEOUT [DEFAULT]]
 -- procedure: thread-mailbox-next [TIMEOUT [DEFAULT]]
 -- procedure: thread-mailbox-rewind
 -- procedure: thread-mailbox-extract-and-rewind
     To allow a thread to examine the messages in its mailbox without
     removing them from the mailbox, each thread has a "mailbox cursor"
     which normally points to the last message accessed in the mailbox.
     When a mailbox cursor is rewound using the procedure
     `thread-mailbox-rewind' or `thread-mailbox-extract-and-rewind' or
     `thread-receive', the cursor does not point to a message, but the
     next call to `thread-receive' and `thread-mailbox-next' will set
     the cursor to the oldest message in the mailbox.

     The procedure `thread-receive' advances the mailbox cursor of the
     current thread to the next message, removes that message from the
     mailbox, rewinds the mailbox cursor, and returns the message.  When
     TIMEOUT is not specified, the current thread will wait until a
     message is available in the mailbox.  When TIMEOUT is specified
     and DEFAULT is not specified, a mailbox-receive-timeout-exception
     object is raised if the timeout is reached before a message is
     available.  When TIMEOUT is specified and DEFAULT is specified,
     DEFAULT is returned if the timeout is reached before a message is
     available.

     The procedure `thread-mailbox-next' behaves like `thread-receive'
     except that the message remains in the mailbox and the mailbox
     cursor is not rewound.

     The procedures `thread-mailbox-rewind' or
     `thread-mailbox-extract-and-rewind' rewind the mailbox cursor of
     the current thread so that the next call to `thread-mailbox-next'
     and `thread-receive' will access the oldest message in the
     mailbox.  Additionally the procedure
     `thread-mailbox-extract-and-rewind' will remove from the mailbox
     the message most recently accessed by a call to
     `thread-mailbox-next'.  When `thread-mailbox-next' has not been
     called since the last call to `thread-receive' or
     `thread-mailbox-rewind' or `thread-mailbox-extract-and-rewind', a
     call to `thread-mailbox-extract-and-rewind' only resets the mailbox
     cursor (no message is removed).

     For example:

          > (thread-send (current-thread) 111)
          > (thread-receive 1 999)
          111
          > (thread-send (current-thread) 222)
          > (thread-send (current-thread) 333)
          > (thread-mailbox-next 1 999)
          222
          > (thread-mailbox-next 1 999)
          333
          > (thread-mailbox-next 1 999)
          999
          > (thread-mailbox-extract-and-rewind)
          > (thread-receive 1 999)
          222
          > (thread-receive 1 999)
          999


 -- procedure: mailbox-receive-timeout-exception? OBJ
 -- procedure: mailbox-receive-timeout-exception-procedure EXC
 -- procedure: mailbox-receive-timeout-exception-arguments EXC
     Mailbox-receive-timeout-exception objects are raised by the
     procedures `thread-receive' and `thread-mailbox-next' when a
     timeout expires before a message is available and no default value
     is specified.  The parameter EXC must be a
     mailbox-receive-timeout-exception object.

     The procedure `mailbox-receive-timeout-exception?' returns `#t'
     when OBJ is a mailbox-receive-timeout-exception object and `#f'
     otherwise.

     The procedure `mailbox-receive-timeout-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `mailbox-receive-timeout-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (mailbox-receive-timeout-exception? exc)
                  (list (mailbox-receive-timeout-exception-procedure exc)
                        (mailbox-receive-timeout-exception-arguments exc))
                  'not-mailbox-receive-timeout-exception))
          > (with-exception-catcher
              handler
              (lambda () (thread-receive 1)))
          (#<procedure #2 thread-receive> (1))


 -- procedure: mutex? OBJ
     This procedure returns `#t' when OBJ is a mutex object and `#f'
     otherwise.

     For example:

          > (mutex? (make-mutex))
          #t
          > (mutex? 'foo)
          #f


 -- procedure: make-mutex [NAME]
     This procedure returns a new mutex in the unlocked/not-abandoned
     state.  The optional NAME is an arbitrary Scheme object which
     identifies the mutex (useful for debugging); it defaults to an
     unspecified value.  The mutex's specific field is set to an
     unspecified value.

     For example:

          > (make-mutex)
          #<mutex #2>
          > (make-mutex 'foo)
          #<mutex #3 foo>


 -- procedure: mutex-name MUTEX
     Returns the name of the MUTEX.  For example:

          > (mutex-name (make-mutex 'foo))
          foo


 -- procedure: mutex-specific MUTEX
 -- procedure: mutex-specific-set! MUTEX OBJ
     The `mutex-specific' procedure returns the content of the MUTEX's
     specific field.

     The `mutex-specific-set!' procedure stores OBJ into the MUTEX's
     specific field and returns an unspecified value.

     For example:

          > (define m (make-mutex))
          > (mutex-specific-set! m "hello")
          > (mutex-specific m)
          "hello"
          > (define (mutex-lock-recursively! mutex)
              (if (eq? (mutex-state mutex) (current-thread))
                  (let ((n (mutex-specific mutex)))
                    (mutex-specific-set! mutex (+ n 1)))
                  (begin
                    (mutex-lock! mutex)
                    (mutex-specific-set! mutex 0))))
          > (define (mutex-unlock-recursively! mutex)
              (let ((n (mutex-specific mutex)))
                (if (= n 0)
                    (mutex-unlock! mutex)
                    (mutex-specific-set! mutex (- n 1)))))
          > (mutex-lock-recursively! m)
          > (mutex-lock-recursively! m)
          > (mutex-lock-recursively! m)
          > (mutex-specific m)
          2


 -- procedure: mutex-state MUTEX
     Thos procedure returns information about the state of the MUTEX.
     The possible results are:

        * thread T: the MUTEX is in the locked/owned state and thread T
          is the owner of the MUTEX

        * symbol `not-owned': the MUTEX is in the locked/not-owned state

        * symbol `abandoned': the MUTEX is in the unlocked/abandoned
          state

        * symbol `not-abandoned': the MUTEX is in the
          unlocked/not-abandoned state


     For example:

          (mutex-state (make-mutex))  ==>  not-abandoned

          (define (thread-alive? thread)
            (let ((mutex (make-mutex)))
              (mutex-lock! mutex #f thread)
              (let ((state (mutex-state mutex)))
                (mutex-unlock! mutex) ; avoid space leak
                (eq? state thread))))


 -- procedure: mutex-lock! MUTEX [TIMEOUT [THREAD]]
     This procedure locks MUTEX.  If the MUTEX is currently locked, the
     current thread waits until the MUTEX is unlocked, or until the
     timeout is reached if TIMEOUT is supplied.  If the timeout is
     reached, `mutex-lock!' returns `#f'.  Otherwise, the state of the
     MUTEX is changed as follows:

        * if THREAD is `#f' the MUTEX becomes locked/not-owned,

        * otherwise, let T be THREAD (or the current thread if THREAD
          is not supplied),

             * if T is terminated the MUTEX becomes unlocked/abandoned,

             * otherwise MUTEX becomes locked/owned with T as the owner.



     After changing the state of the MUTEX, an
     abandoned-mutex-exception object is raised if the MUTEX was
     unlocked/abandoned before the state change, otherwise
     `mutex-lock!' returns `#t'.  It is not an error if the MUTEX is
     owned by the current thread (but the current thread will have to
     wait).

     For example:

          ; an implementation of a mailbox object of depth one; this
          ; implementation does not behave well in the presence of forced
          ; thread terminations using thread-terminate! (deadlock can occur
          ; if a thread is terminated in the middle of a put! or get! operation)

          (define (make-empty-mailbox)
            (let ((put-mutex (make-mutex)) ; allow put! operation
                  (get-mutex (make-mutex))
                  (cell #f))

              (define (put! obj)
                (mutex-lock! put-mutex #f #f) ; prevent put! operation
                (set! cell obj)
                (mutex-unlock! get-mutex)) ; allow get! operation

              (define (get!)
                (mutex-lock! get-mutex #f #f) ; wait until object in mailbox
                (let ((result cell))
                  (set! cell #f) ; prevent space leaks
                  (mutex-unlock! put-mutex) ; allow put! operation
                  result))

              (mutex-lock! get-mutex #f #f) ; prevent get! operation

              (lambda (msg)
                (case msg
                  ((put!) put!)
                  ((get!) get!)
                  (else (error "unknown message"))))))

          (define (mailbox-put! m obj) ((m 'put!) obj))
          (define (mailbox-get! m) ((m 'get!)))

          ; an alternate implementation of thread-sleep!

          (define (sleep! timeout)
            (let ((m (make-mutex)))
              (mutex-lock! m #f #f)
              (mutex-lock! m timeout #f)))

          ; a procedure that waits for one of two mutexes to unlock

          (define (lock-one-of! mutex1 mutex2)
            ; this procedure assumes that neither mutex1 or mutex2
            ; are owned by the current thread
            (let ((ct (current-thread))
                  (done-mutex (make-mutex)))
              (mutex-lock! done-mutex #f #f)
              (let ((t1 (thread-start!
                         (make-thread
                          (lambda ()
                            (mutex-lock! mutex1 #f ct)
                            (mutex-unlock! done-mutex)))))
                    (t2 (thread-start!
                         (make-thread
                          (lambda ()
                            (mutex-lock! mutex2 #f ct)
                            (mutex-unlock! done-mutex))))))
                (mutex-lock! done-mutex #f #f)
                (thread-terminate! t1)
                (thread-terminate! t2)
                (if (eq? (mutex-state mutex1) ct)
                    (begin
                      (if (eq? (mutex-state mutex2) ct)
                          (mutex-unlock! mutex2)) ; don't lock both
                      mutex1)
                    mutex2))))


 -- procedure: mutex-unlock! MUTEX [CONDITION-VARIABLE [TIMEOUT]]
     This procedure unlocks the MUTEX by making it
     unlocked/not-abandoned.  It is not an error to unlock an unlocked
     mutex and a mutex that is owned by any thread.  If
     CONDITION-VARIABLE is supplied, the current thread is blocked and
     added to the CONDITION-VARIABLE before unlocking MUTEX; the thread
     can unblock at any time but no later than when an appropriate call
     to `condition-variable-signal!' or `condition-variable-broadcast!'
     is performed (see below), and no later than the timeout (if
     TIMEOUT is supplied).  If there are threads waiting to lock this
     MUTEX, the scheduler selects a thread, the mutex becomes
     locked/owned or locked/not-owned, and the thread is unblocked.
     `mutex-unlock!' returns `#f' when the timeout is reached,
     otherwise it returns `#t'.

     NOTE: The reason the thread can unblock at any time (when
     CONDITION-VARIABLE is supplied) is that the scheduler, when it
     detects a serious problem such as a deadlock, must interrupt one of
     the blocked threads (such as the primordial thread) so that it can
     perform some appropriate action.  After a thread blocked on a
     condition-variable has handled such an interrupt it would be wrong
     for the scheduler to return the thread to the blocked state,
     because any calls to `condition-variable-broadcast!' during the
     interrupt will have gone unnoticed.  It is necessary for the
     thread to remain runnable and return from the call to
     `mutex-unlock!' with a result of `#t'.

     NOTE: `mutex-unlock!' is related to the "wait" operation on
     condition variables available in other thread systems.  The main
     difference is that "wait" automatically locks MUTEX just after the
     thread is unblocked.  This operation is not performed by
     `mutex-unlock!' and so must be done by an explicit call to
     `mutex-lock!'.  This has the advantages that a different timeout
     and exception-handler can be specified on the `mutex-lock!' and
     `mutex-unlock!' and the location of all the mutex operations is
     clearly apparent.

     For example:

          (let loop ()
            (mutex-lock! m)
            (if (condition-is-true?)
                (begin
                  (do-something-when-condition-is-true)
                  (mutex-unlock! m))
                (begin
                  (mutex-unlock! m cv)
                  (loop))))


 -- procedure: condition-variable? OBJ
     This procedure returns `#t' when OBJ is a condition-variable
     object and `#f' otherwise.

     For example:

          > (condition-variable? (make-condition-variable))
          #t
          > (condition-variable? 'foo)
          #f


 -- procedure: make-condition-variable [NAME]
     This procedure returns a new empty condition variable.  The
     optional NAME is an arbitrary Scheme object which identifies the
     condition variable (useful for debugging); it defaults to an
     unspecified value.  The condition variable's specific field is set
     to an unspecified value.

     For example:

          > (make-condition-variable)
          #<condition-variable #2>


 -- procedure: condition-variable-name CONDITION-VARIABLE
     This procedure returns the name of the CONDITION-VARIABLE.  For
     example:

          > (condition-variable-name (make-condition-variable 'foo))
          foo


 -- procedure: condition-variable-specific CONDITION-VARIABLE
 -- procedure: condition-variable-specific-set! CONDITION-VARIABLE OBJ
     The `condition-variable-specific' procedure returns the content of
     the CONDITION-VARIABLE's specific field.

     The `condition-variable-specific-set!' procedure stores OBJ into
     the CONDITION-VARIABLE's specific field and returns an unspecified
     value.

     For example:

          > (define cv (make-condition-variable))
          > (condition-variable-specific-set! cv "hello")
          > (condition-variable-specific cv)
          "hello"


 -- procedure: condition-variable-signal! CONDITION-VARIABLE
     This procedure unblocks a thread blocked on the CONDITION-VARIABLE
     (if there is at least one) and returns an unspecified value.

     For example:

          ; an implementation of a mailbox object of depth one; this
          ; implementation behaves gracefully when threads are forcibly
          ; terminated using thread-terminate! (an abandoned-mutex-exception
          ; object will be raised when a put! or get! operation is attempted
          ; after a thread is terminated in the middle of a put! or get!
          ; operation)

          (define (make-empty-mailbox)
            (let ((mutex (make-mutex))
                  (put-condvar (make-condition-variable))
                  (get-condvar (make-condition-variable))
                  (full? #f)
                  (cell #f))

              (define (put! obj)
                (mutex-lock! mutex)
                (if full?
                    (begin
                      (mutex-unlock! mutex put-condvar)
                      (put! obj))
                    (begin
                      (set! cell obj)
                      (set! full? #t)
                      (condition-variable-signal! get-condvar)
                      (mutex-unlock! mutex))))

              (define (get!)
                (mutex-lock! mutex)
                (if (not full?)
                    (begin
                      (mutex-unlock! mutex get-condvar)
                      (get!))
                    (let ((result cell))
                      (set! cell #f) ; avoid space leaks
                      (set! full? #f)
                      (condition-variable-signal! put-condvar)
                      (mutex-unlock! mutex)
                      result)))

              (lambda (msg)
                (case msg
                  ((put!) put!)
                  ((get!) get!)
                  (else (error "unknown message"))))))

          (define (mailbox-put! m obj) ((m 'put!) obj))
          (define (mailbox-get! m) ((m 'get!)))


 -- procedure: condition-variable-broadcast! CONDITION-VARIABLE
     This procedure unblocks all the thread blocked on the
     CONDITION-VARIABLE and returns an unspecified value.

     For example:

          (define (make-semaphore n)
            (vector n (make-mutex) (make-condition-variable)))

          (define (semaphore-wait! sema)
            (mutex-lock! (vector-ref sema 1))
            (let ((n (vector-ref sema 0)))
              (if (> n 0)
                  (begin
                    (vector-set! sema 0 (- n 1))
                    (mutex-unlock! (vector-ref sema 1)))
                  (begin
                    (mutex-unlock! (vector-ref sema 1) (vector-ref sema 2))
                    (semaphore-wait! sema))))

          (define (semaphore-signal-by! sema increment)
            (mutex-lock! (vector-ref sema 1))
            (let ((n (+ (vector-ref sema 0) increment)))
              (vector-set! sema 0 n)
              (if (> n 0)
                  (condition-variable-broadcast! (vector-ref sema 2)))
              (mutex-unlock! (vector-ref sema 1))))


11 Dynamic environment
**********************

The "dynamic environment" is the structure which allows the system to
find the value returned by the standard procedures `current-input-port'
and `current-output-port'.  The standard procedures
`with-input-from-file' and `with-output-to-file' extend the dynamic
environment to produce a new dynamic environment which is in effect for
the dynamic extent of the call to the thunk passed as their last
argument.  These procedures are essentially special purpose dynamic
binding operations on hidden dynamic variables (one for
`current-input-port' and one for `current-output-port').  Gambit
generalizes this dynamic binding mechanism to allow the user to
introduce new dynamic variables, called "parameter objects", and
dynamically bind them.  The parameter objects implemented by Gambit are
compatible with the specification of the "Parameter objects SRFI" (SRFI
39).

   One important issue is the relationship between the dynamic
environments of the parent and child threads when a thread is created.
Each thread has its own dynamic environment that is accessed when
looking up the value bound to a parameter object by that thread.  When
a thread's dynamic environment is extended it does not affect the
dynamic environment of other threads.  When a thread is created it is
given a dynamic environment whose bindings are inherited from the
parent thread.  In this inherited dynamic environment the parameter
objects are bound to the same cells as the parent's dynamic environment
(in other words an assignment of a new value to a parameter object is
visible in the other thread).

   Another important issue is the interaction between the
`dynamic-wind' procedure and dynamic environments.  When a thread
creates a continuation, the thread's dynamic environment and the
`dynamic-wind' stack are saved within the continuation (an alternate
but equivalent point of view is that the `dynamic-wind' stack is part
of the dynamic environment).  When this continuation is invoked the
required `dynamic-wind' before and after thunks are called and the
saved dynamic environment is reinstated as the dynamic environment of
the current thread.  During the call to each required `dynamic-wind'
before and after thunk, the dynamic environment and the `dynamic-wind'
stack in effect when the corresponding `dynamic-wind' was executed are
reinstated.  Note that this specification precisely defines the
semantics of calling `call-with-current-continuation' or invoking a
continuation within a before or after thunk.  The semantics are well
defined even when a continuation created by another thread is invoked.
Below is an example exercising the subtleties of this semantics.

     (with-output-to-file
      "foo"
      (lambda ()
        (let ((k (call-with-current-continuation
                  (lambda (exit)
                    (with-output-to-file
                     "bar"
                     (lambda ()
                       (dynamic-wind
                        (lambda ()
                          (write '(b1))
                          (force-output))
                        (lambda ()
                          (let ((x (call-with-current-continuation
                                    (lambda (cont) (exit cont)))))
                            (write '(t1))
                            (force-output)
                            x))
                        (lambda ()
                          (write '(a1))
                          (force-output)))))))))
          (if k
              (dynamic-wind
               (lambda ()
                 (write '(b2))
                 (force-output))
               (lambda ()
                 (with-output-to-file
                  "baz"
                  (lambda ()
                    (write '(t2))
                    (force-output)
                    ; go back inside (with-output-to-file "bar" ...)
                    (k #f))))
               (lambda ()
                 (write '(a2))
                 (force-output)))))))

   The following actions will occur when this code is executed:
`(b1)(a1)' is written to "bar", `(b2)' is then written to "foo", `(t2)'
is then written to "baz", `(a2)' is then written to "foo", and finally
`(b1)(t1)(a1)' is written to "bar".

 -- procedure: make-parameter OBJ [SET-FILTER [GET-FILTER]]
     The dynamic environment is composed of two parts: the "local
     dynamic environment" and the "global dynamic environment".  There
     is a single global dynamic environment, and it is used to lookup
     parameter objects that can't be found in the local dynamic
     environment.

     The `make-parameter' procedure returns a new "parameter object".
     The SET-FILTER argument is a one argument "set" conversion
     procedure.  The GET-FILTER argument is a one argument "get"
     conversion procedure.  If they are not specified the conversion
     procedures default to the identity function.

     The global dynamic environment is updated to associate the
     parameter object to a new cell.  The initial content of the cell
     is the result of applying the "set" conversion procedure to OBJ.

     A parameter object is a procedure which accepts zero or one
     argument.  The cell bound to a particular parameter object in the
     dynamic environment is accessed by calling the parameter object.
     When no argument is passed, the value returned is the result of
     applying the "get" conversion procedure to the content of the cell.
     When one argument is passed the content of the cell is updated
     with the result of applying the parameter object's "set"
     conversion procedure to the argument.  Note that the conversion
     procedures can be used for guaranteeing the type of the parameter
     object's binding and/or to perform some conversion of the value.

     For example:

          > (define radix (make-parameter 10))
          > (radix)
          10
          > (radix 2)
          > (radix)
          2
          > (define prompt
              (make-parameter
                123
                (lambda (x)
                  (if (string? x)
                      x
                      (object->string x)))))
          > (prompt)
          "123"
          > (prompt "$")
          > (prompt)
          "$"
          > (define p
              (make-parameter
                100
                (lambda (val) ;; set filter
                  (pp (list val: val))
                  (list 0 val))
                (lambda (state) ;; get filter
                  (pp (list state: state))
                  (set-car! state (+ 1 (car state)))
                  (+ (car state) (cadr state)))))
          (val: 100)
          > (p)
          (state: (0 100))
          101
          > (p)
          (state: (1 100))
          102
          > (p)
          (state: (2 100))
          103
          > (p 555)
          (val: 555)
          > (p)
          (state: (0 555))
          556
          > (p)
          (state: (1 555))
          557
          > (define write-shared
              (make-parameter
                #f
                (lambda (x)
                  (if (boolean? x)
                      x
                      (error "only booleans are accepted by write-shared")))))
          > (write-shared 123)
          *** ERROR IN ##make-parameter -- only booleans are accepted by write-shared


 -- special form: parameterize ((procedure value)...) body
     The `parameterize' form, evaluates all procedure and value
     expressions in an unspecified order.  All the procedure
     expressions must evaluate to procedures, either parameter objects
     or procedures accepting zero and one argument.  Then, for each
     procedure p and in an unspecified order:

        * If p is a parameter object a new cell is created,
          initialized, and bound to the parameter object in the local
          dynamic environment.  The value contained in the cell is the
          result of applying the parameter object's "set" conversion
          procedure to value.  The resulting dynamic environment is
          then used for processing the remaining bindings (or the
          evaluation of body if there are no other bindings).

        * Otherwise p will be used according to the following protocol:
          we say that the call `(p)' "gets p's value" and that the call
          `(p x)' "sets p's value to x".  First, the `parameterize'
          form gets p's value and saves it in a local variable.  It
          then sets p's value to value.  It then processes the
          remaining bindings (or evaluates body if there are no other
          bindings).  Then it sets p's value to the saved value.  These
          steps are performed in a `dynamic-wind' so that it is
          possible to use continuations to jump into and out of the body
          (i.e. the `dynamic-wind''s before thunk sets p's value to
          value and the after thunk sets p's value to the saved value).


     The result(s) of the `parameterize' form are the result(s) of the
     body.

     Note that using procedures instead of parameter objects may lead to
     unexpected results in multithreaded programs because the before and
     after thunks of the `dynamic-wind' are not called when control
     switches between threads.

     For example:

          > (define radix (make-parameter 2))
          > (define prompt
              (make-parameter
                123
                (lambda (x)
                  (if (string? x)
                      x
                      (object->string x)))))
          > (radix)
          2
          > (parameterize ((radix 16)) (radix))
          16
          > (radix)
          2
          > (define (f n) (number->string n (radix)))
          > (f 10)
          "1010"
          > (parameterize ((radix 8)) (f 10))
          "12"
          > (parameterize ((radix 8) (prompt (f 10))) (prompt))
          "1010"
          > (define p
              (let ((x 1))
                (lambda args
                  (if (null? args) x (set! x (car args))))))
          > (let* ((a (p))
                   (b (parameterize ((p 2)) (list (p))))
                   (c (p)))
              (list a b c))
          (1 (2) 1)


12 Exceptions
*************

12.1 Exception-handling
=======================

Gambit's exception-handling model is inspired from the withdrawn
"Exception Handling SRFI" (SRFI 12), the "Multithreading support SRFI"
(SRFI 18), and the "Exception Handling for Programs SRFI" (SRFI 34).
The two fundamental operations are the dynamic binding of an exception
handler (i.e. the procedure `with-exception-handler') and the
invocation of the exception handler (i.e. the procedure `raise').

   All predefined procedures which check for errors (including type
errors, memory allocation errors, host operating-system errors, etc)
report these errors using the exception-handling system (i.e. they
"raise" an exception that can be handled in a user-defined exception
handler).  When an exception is raised and the exception is not handled
by a user-defined exception handler, the predefined exception handler
will display an error message (if the primordial thread raised the
exception) or the thread will silently terminate with no error message
(if it is not the primordial thread that raised the exception).  This
default behavior can be changed through the `-:debug=...' runtime
option (*note Runtime options::).

   Predefined procedures normally raise exceptions by performing a
tail-call to the exception handler (the exceptions are "complex"
procedures such as `eval', `compile-file', `read', `write', etc).  This
means that the continuation of the exception handler and of the REPL
that may be started due to this is normally the continuation of the
predefined procedure that raised the exception.  By exiting the REPL
with the `,(c EXPRESSION)' command it is thus possible to resume the
program as though the call to the predefined procedure returned the
value of EXPRESSION.  For example:

     > (define (f x) (+ (car x) 1))
     > (f 2) ; typo... we meant to say (f '(2))
     *** ERROR IN f, (console)@1.18 -- (Argument 1) PAIR expected
     (car 2)
     1> ,(c 2)
     3

 -- procedure: current-exception-handler [NEW-EXCEPTION-HANDLER]
     The parameter object `current-exception-handler' is bound to the
     current exception-handler.  Calling this procedure with no argument
     returns the current exception-handler and calling this procedure
     with one argument sets the current exception-handler to
     NEW-EXCEPTION-HANDLER.

     For example:

          > (current-exception-handler)
          #<procedure #2 primordial-exception-handler>
          > (current-exception-handler (lambda (exc) (pp exc) 999))
          > (/ 1 0)
          #<divide-by-zero-exception #3>
          999


 -- procedure: with-exception-handler HANDLER THUNK
     Returns the result(s) of calling THUNK with no arguments.  The
     HANDLER, which must be a procedure, is installed as the current
     exception-handler in the dynamic environment in effect during the
     call to THUNK.  Note that the dynamic environment in effect during
     the call to HANDLER has HANDLER as the exception-handler.
     Consequently, an exception raised during the call to HANDLER may
     lead to an infinite loop.

     For example:

          > (with-exception-handler
              (lambda (e) (write e) 5)
              (lambda () (+ 1 (* 2 3) 4)))
          11
          > (with-exception-handler
              (lambda (e) (write e) 5)
              (lambda () (+ 1 (* 'foo 3) 4)))
          #<type-exception #2>10
          > (with-exception-handler
              (lambda (e) (write e 9))
              (lambda () (+ 1 (* 'foo 3) 4)))
          INFINITE LOOP


 -- procedure: with-exception-catcher HANDLER THUNK
     Returns the result(s) of calling THUNK with no arguments.  A new
     exception-handler is installed as the current exception-handler in
     the dynamic environment in effect during the call to THUNK.  This
     new exception-handler will call the HANDLER, which must be a
     procedure, with the exception object as an argument and with the
     same continuation as the call to `with-exception-catcher'.  This
     implies that the dynamic environment in effect during the call to
     HANDLER is the same as the one in effect at the call to
     `with-exception-catcher'.  Consequently, an exception raised
     during the call to HANDLER will not lead to an infinite loop.

     For example:

          > (with-exception-catcher
              (lambda (e) (write e) 5)
              (lambda () (+ 1 (* 2 3) 4)))
          11
          > (with-exception-catcher
              (lambda (e) (write e) 5)
              (lambda () (+ 1 (* 'foo 3) 4)))
          #<type-exception #2>5
          > (with-exception-catcher
              (lambda (e) (write e 9))
              (lambda () (+ 1 (* 'foo 3) 4)))
          *** ERROR IN (console)@7.1 -- (Argument 2) OUTPUT PORT expected
          (write '#<type-exception #3> 9)


 -- procedure: raise OBJ
     This procedure tail-calls the current exception-handler with OBJ
     as the sole argument.  If the exception-handler returns, the
     continuation of the call to `raise' is invoked.

     For example:

          > (with-exception-handler
              (lambda (exc)
                (pp exc)
                100)
              (lambda ()
                (+ 1 (raise "hello"))))
          "hello"
          101


 -- procedure: abort OBJ
 -- procedure: noncontinuable-exception? OBJ
 -- procedure: noncontinuable-exception-reason EXC
     The procedure `abort' calls the current exception-handler with OBJ
     as the sole argument.  If the exception-handler returns, the
     procedure `abort' will be tail-called with a
     noncontinuable-exception object, whose reason field is OBJ, as
     sole argument.

     Noncontinuable-exception objects are raised by the `abort'
     procedure when the exception-handler returns.  The parameter EXC
     must be a noncontinuable-exception object.

     The procedure `noncontinuable-exception?' returns `#t' when OBJ is
     a noncontinuable-exception object and `#f' otherwise.

     The procedure `noncontinuable-exception-reason' returns the
     argument of the call to `abort' that raised EXC.

     For example:

          > (call-with-current-continuation
              (lambda (k)
                (with-exception-handler
                  (lambda (exc)
                    (pp exc)
                    (if (noncontinuable-exception? exc)
                        (k (list (noncontinuable-exception-reason exc)))
                        100))
                  (lambda ()
                    (+ 1 (abort "hello"))))))
          "hello"
          #<noncontinuable-exception #2>
          ("hello")


12.2 Exception objects related to memory management
===================================================

 -- procedure: heap-overflow-exception? OBJ
     Heap-overflow-exception objects are raised when the allocation of
     an object would cause the heap to use more memory space than is
     available.

     The procedure `heap-overflow-exception?' returns `#t' when OBJ is
     a heap-overflow-exception object and `#f' otherwise.

     For example:

          > (define (handler exc)
              (if (heap-overflow-exception? exc)
                  exc
                  'not-heap-overflow-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (define (f x) (f (cons 1 x)))
                (f '())))
          #<heap-overflow-exception #2>


 -- procedure: stack-overflow-exception? OBJ
     Stack-overflow-exception objects are raised when the allocation of
     a continuation frame would cause the heap to use more memory space
     than is available.

     The procedure `stack-overflow-exception?' returns `#t' when OBJ is
     a stack-overflow-exception object and `#f' otherwise.

     For example:

          > (define (handler exc)
              (if (stack-overflow-exception? exc)
                  exc
                  'not-stack-overflow-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (define (f) (+ 1 (f)))
                (f)))
          #<stack-overflow-exception #2>


12.3 Exception objects related to the host environment
======================================================

 -- procedure: os-exception? OBJ
 -- procedure: os-exception-procedure EXC
 -- procedure: os-exception-arguments EXC
 -- procedure: os-exception-code EXC
 -- procedure: os-exception-message EXC
     Os-exception objects are raised by procedures which access the host
     operating-system's services when the requested operation fails.
     The parameter EXC must be a os-exception object.

     The procedure `os-exception?' returns `#t' when OBJ is a
     os-exception object and `#f' otherwise.

     The procedure `os-exception-procedure' returns the procedure that
     raised EXC.

     The procedure `os-exception-arguments' returns the list of
     arguments of the procedure that raised EXC.

     The procedure `os-exception-code' returns an exact integer error
     code that can be converted to a string by the `err-code->string'
     procedure.  Note that the error code is operating-system dependent.

     The procedure `os-exception-message' returns `#f' or a string
     giving details of the exception in a human-readable form.

     For example:

          > (define (handler exc)
              (if (os-exception? exc)
                  (list (os-exception-procedure exc)
                        (os-exception-arguments exc)
                        (err-code->string (os-exception-code exc))
                        (os-exception-message exc))
                  'not-os-exception))
          > (with-exception-catcher
              handler
              (lambda () (host-info "x.y.z")))
          (#<procedure #2 host-info> ("x.y.z") "Unknown host" #f)


 -- procedure: no-such-file-or-directory-exception? OBJ
 -- procedure: no-such-file-or-directory-exception-procedure EXC
 -- procedure: no-such-file-or-directory-exception-arguments EXC
     No-such-file-or-directory-exception objects are raised by
     procedures which access the filesystem (such as `open-input-file'
     and `directory-files') when the path specified can't be found on
     the filesystem.  The parameter EXC must be a
     no-such-file-or-directory-exception object.

     The procedure `no-such-file-or-directory-exception?' returns `#t'
     when OBJ is a no-such-file-or-directory-exception object and `#f'
     otherwise.

     The procedure `no-such-file-or-directory-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `no-such-file-or-directory-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (no-such-file-or-directory-exception? exc)
                  (list (no-such-file-or-directory-exception-procedure exc)
                        (no-such-file-or-directory-exception-arguments exc))
                  'not-no-such-file-or-directory-exception))
          > (with-exception-catcher
              handler
              (lambda () (with-input-from-file "nofile" read)))
          (#<procedure #2 with-input-from-file> ("nofile" #<procedure #3 read>))


 -- procedure: file-exists-exception? OBJ
 -- procedure: file-exists-exception-procedure EXC
 -- procedure: file-exists-exception-arguments EXC
     File-exists-exception objects are raised by procedures which
     access the filesystem (such as `open-output-file' and
     `create-directory') when the path specified is an existing file on
     the filesystem.  The parameter EXC must be a file-exists-exception
     object.

     The procedure `file-exists-exception?' returns `#t' when OBJ is a
     file-exists-exception object and `#f' otherwise.

     The procedure `file-exists-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `file-exists-exception-arguments' returns the list
     of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (file-exists-exception? exc)
                  (list (file-exists-exception-procedure exc)
                        (file-exists-exception-arguments exc))
                  'not-file-exists-exception))
          > (with-exception-catcher
              handler
              (lambda () (with-output-to-file '(path: "foo" create: #t) newline)))
          > (with-exception-catcher
              handler
              (lambda () (with-output-to-file '(path: "foo" create: #t) newline)))
          (#<procedure #2 with-output-to-file>
           ((path: "foo" create: #t) #<procedure #3 newline>))


 -- procedure: permission-denied-exception? OBJ
 -- procedure: permission-denied-exception-procedure EXC
 -- procedure: permission-denied-exception-arguments EXC
     Permission-denied-exception objects are raised by procedures which
     access the filesystem (such as `open-file' and `open-directory')
     when the access to the specified path is not allowed, or search
     permission is denied for a directory in the path prefix, or write
     access to the parent directory isn't allowed for a file that
     doesn't exist yet on the filesystem.  The parameter EXC must be a
     permission-denied-exception object.

     The procedure `permission-denied-exception?' returns `#t' when OBJ
     is a permission-denied-exception object and `#f' otherwise.

     The procedure `permission-denied-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `permission-denied-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (permission-denied-exception? exc)
                  (list (permission-denied-exception-procedure exc)
                        (permission-denied-exception-arguments exc))
                  'not-permission-denied-exception))
          > (with-exception-catcher
              handler
              (lambda () (with-input-from-file "empty" read)))
          #!eof
          > (with-exception-catcher
              handler
              (lambda () (with-input-from-file "noperm" read)))
          (#<procedure #2 with-input-from-file> ("noperm" #<procedure #3 read>))


 -- procedure: unbound-os-environment-variable-exception? OBJ
 -- procedure: unbound-os-environment-variable-exception-procedure EXC
 -- procedure: unbound-os-environment-variable-exception-arguments EXC
     Unbound-os-environment-variable-exception objects are raised when
     an unbound operating-system environment variable is accessed by the
     procedures `getenv' and `setenv'.  The parameter EXC must be an
     unbound-os-environment-variable-exception object.

     The procedure `unbound-os-environment-variable-exception?' returns
     `#t' when OBJ is an unbound-os-environment-variable-exception
     object and `#f' otherwise.

     The procedure `unbound-os-environment-variable-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `unbound-os-environment-variable-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (unbound-os-environment-variable-exception? exc)
                  (list (unbound-os-environment-variable-exception-procedure exc)
                        (unbound-os-environment-variable-exception-arguments exc))
                  'not-unbound-os-environment-variable-exception))
          > (with-exception-catcher
              handler
              (lambda () (getenv "DOES_NOT_EXIST")))
          (#<procedure #2 getenv> ("DOES_NOT_EXIST"))


12.4 Exception objects related to threads
=========================================

 -- procedure: scheduler-exception? OBJ
 -- procedure: scheduler-exception-reason EXC
     Scheduler-exception objects are raised by the scheduler when some
     operation requested from the host operating system failed (e.g.
     checking the status of the devices in order to wake up threads
     waiting to perform I/O on these devices).  The parameter EXC must
     be a scheduler-exception object.

     The procedure `scheduler-exception?' returns `#t' when OBJ is a
     scheduler-exception object and `#f' otherwise.

     The procedure `scheduler-exception-reason' returns the
     os-exception object that describes the failure detected by the
     scheduler.


 -- procedure: deadlock-exception? OBJ
     Deadlock-exception objects are raised when the scheduler discovers
     that all threads are blocked and can make no further progress.  In
     that case the scheduler unblocks the primordial-thread and forces
     it to raise a deadlock-exception object.

     The procedure `deadlock-exception?' returns `#t' when OBJ is a
     deadlock-exception object and `#f' otherwise.

     For example:

          > (define (handler exc)
              (if (deadlock-exception? exc)
                  exc
                  'not-deadlock-exception))
          > (with-exception-catcher
              handler
              (lambda () (read (open-vector))))
          #<deadlock-exception #2>


 -- procedure: abandoned-mutex-exception? OBJ
     Abandoned-mutex-exception objects are raised when the current
     thread locks a mutex that was owned by a thread which terminated
     (see `mutex-lock!').

     The procedure `abandoned-mutex-exception?' returns `#t' when OBJ
     is a abandoned-mutex-exception object and `#f' otherwise.

     For example:

          > (define (handler exc)
              (if (abandoned-mutex-exception? exc)
                  exc
                  'not-abandoned-mutex-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (let ((m (make-mutex)))
                  (thread-join!
                    (thread-start!
                      (make-thread
                        (lambda () (mutex-lock! m)))))
                  (mutex-lock! m))))
          #<abandoned-mutex-exception #2>


 -- procedure: join-timeout-exception? OBJ
 -- procedure: join-timeout-exception-procedure EXC
 -- procedure: join-timeout-exception-arguments EXC
     Join-timeout-exception objects are raised when a call to the
     `thread-join!' procedure reaches its timeout before the target
     thread terminates and a timeout-value parameter is not specified.
     The parameter EXC must be a join-timeout-exception object.

     The procedure `join-timeout-exception?' returns `#t' when OBJ is a
     join-timeout-exception object and `#f' otherwise.

     The procedure `join-timeout-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `join-timeout-exception-arguments' returns the list
     of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (join-timeout-exception? exc)
                  (list (join-timeout-exception-procedure exc)
                        (join-timeout-exception-arguments exc))
                  'not-join-timeout-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (thread-join!
                  (thread-start!
                    (make-thread
                      (lambda () (thread-sleep! 10))))
                  5)))
          (#<procedure #2 thread-join!> (#<thread #3> 5))


 -- procedure: started-thread-exception? OBJ
 -- procedure: started-thread-exception-procedure EXC
 -- procedure: started-thread-exception-arguments EXC
     Started-thread-exception objects are raised when the target thread
     of a call to the procedure `thread-start!' is already started.  The
     parameter EXC must be a started-thread-exception object.

     The procedure `started-thread-exception?' returns `#t' when OBJ is
     a started-thread-exception object and `#f' otherwise.

     The procedure `started-thread-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `started-thread-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (started-thread-exception? exc)
                  (list (started-thread-exception-procedure exc)
                        (started-thread-exception-arguments exc))
                  'not-started-thread-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (let ((t (make-thread (lambda () (expt 2 1000)))))
                  (thread-start! t)
                  (thread-start! t))))
          (#<procedure #2 thread-start!> (#<thread #3>))


 -- procedure: terminated-thread-exception? OBJ
 -- procedure: terminated-thread-exception-procedure EXC
 -- procedure: terminated-thread-exception-arguments EXC
     Terminated-thread-exception objects are raised when the
     `thread-join!' procedure is called and the target thread has
     terminated as a result of a call to the `thread-terminate!'
     procedure.  The parameter EXC must be a
     terminated-thread-exception object.

     The procedure `terminated-thread-exception?' returns `#t' when OBJ
     is a terminated-thread-exception object and `#f' otherwise.

     The procedure `terminated-thread-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `terminated-thread-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (terminated-thread-exception? exc)
                  (list (terminated-thread-exception-procedure exc)
                        (terminated-thread-exception-arguments exc))
                  'not-terminated-thread-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (thread-join!
                  (thread-start!
                    (make-thread
                      (lambda () (thread-terminate! (current-thread))))))))
          (#<procedure #2 thread-join!> (#<thread #3>))


 -- procedure: uncaught-exception? OBJ
 -- procedure: uncaught-exception-procedure EXC
 -- procedure: uncaught-exception-arguments EXC
 -- procedure: uncaught-exception-reason EXC
     Uncaught-exception objects are raised when an object is raised in a
     thread and that thread does not handle it (i.e. the thread
     terminated because it did not catch an exception it raised).  The
     parameter EXC must be an uncaught-exception object.

     The procedure `uncaught-exception?' returns `#t' when OBJ is an
     uncaught-exception object and `#f' otherwise.

     The procedure `uncaught-exception-procedure' returns the procedure
     that raised EXC.

     The procedure `uncaught-exception-arguments' returns the list of
     arguments of the procedure that raised EXC.

     The procedure `uncaught-exception-reason' returns the object that
     was raised by the thread and not handled by that thread.

     For example:

          > (define (handler exc)
              (if (uncaught-exception? exc)
                  (list (uncaught-exception-procedure exc)
                        (uncaught-exception-arguments exc)
                        (uncaught-exception-reason exc))
                  'not-uncaught-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (thread-join!
                  (thread-start!
                    (make-thread
                      (lambda () (open-input-file "data" 99)))))))
          (#<procedure #2 thread-join!>
           (#<thread #3>)
           #<wrong-number-of-arguments-exception #4>)


12.5 Exception objects related to C-interface
=============================================

 -- procedure: cfun-conversion-exception? OBJ
 -- procedure: cfun-conversion-exception-procedure EXC
 -- procedure: cfun-conversion-exception-arguments EXC
 -- procedure: cfun-conversion-exception-code EXC
 -- procedure: cfun-conversion-exception-message EXC
     Cfun-conversion-exception objects are raised by the C-interface
     when converting between the Scheme representation and the C
     representation of a value during a call from Scheme to C.  The
     parameter EXC must be a cfun-conversion-exception object.

     The procedure `cfun-conversion-exception?' returns `#t' when OBJ
     is a cfun-conversion-exception object and `#f' otherwise.

     The procedure `cfun-conversion-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `cfun-conversion-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     The procedure `cfun-conversion-exception-code' returns an exact
     integer error code that can be converted to a string by the
     `err-code->string' procedure.

     The procedure `cfun-conversion-exception-message' returns `#f' or
     a string giving details of the exception in a human-readable form.

     For example:

          $ cat test1.scm
          (define weird
            (c-lambda (char-string) nonnull-char-string
              "___return(___arg1);"))
          $ gsc test1.scm
          $ gsi
          Gambit v4.9.4

          > (load "test1")
          "/Users/feeley/gambit/doc/test1.o1"
          > (weird "hello")
          "hello"
          > (define (handler exc)
              (if (cfun-conversion-exception? exc)
                  (list (cfun-conversion-exception-procedure exc)
                        (cfun-conversion-exception-arguments exc)
                        (err-code->string (cfun-conversion-exception-code exc))
                        (cfun-conversion-exception-message exc))
                  'not-cfun-conversion-exception))
          > (with-exception-catcher
              handler
              (lambda () (weird 'not-a-string)))
          (#<procedure #2 weird>
           (not-a-string)
           "(Argument 1) Can't convert to C char-string"
           #f)
          > (with-exception-catcher
              handler
              (lambda () (weird #f)))
          (#<procedure #2 weird>
           (#f)
           "Can't convert result from C nonnull-char-string"
           #f)


 -- procedure: sfun-conversion-exception? OBJ
 -- procedure: sfun-conversion-exception-procedure EXC
 -- procedure: sfun-conversion-exception-arguments EXC
 -- procedure: sfun-conversion-exception-code EXC
 -- procedure: sfun-conversion-exception-message EXC
     Sfun-conversion-exception objects are raised by the C-interface
     when converting between the Scheme representation and the C
     representation of a value during a call from C to Scheme.  The
     parameter EXC must be a sfun-conversion-exception object.

     The procedure `sfun-conversion-exception?' returns `#t' when OBJ
     is a sfun-conversion-exception object and `#f' otherwise.

     The procedure `sfun-conversion-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `sfun-conversion-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     The procedure `sfun-conversion-exception-code' returns an exact
     integer error code that can be converted to a string by the
     `err-code->string' procedure.

     The procedure `sfun-conversion-exception-message' returns `#f' or
     a string giving details of the exception in a human-readable form.

     For example:

          $ cat test2.scm
          (c-define (f str) (nonnull-char-string) int "f" ""
            (string->number str))
          (define t1 (c-lambda () int "___return(f (\"123\"));"))
          (define t2 (c-lambda () int "___return(f (0));"))
          (define t3 (c-lambda () int "___return(f (\"1.5\"));"))
          $ gsc test2.scm
          $ gsi
          Gambit v4.9.4

          > (load "test2")
          "/u/feeley/test2.o1"
          > (t1)
          123
          > (define (handler exc)
              (if (sfun-conversion-exception? exc)
                  (list (sfun-conversion-exception-procedure exc)
                        (sfun-conversion-exception-arguments exc)
                        (err-code->string (sfun-conversion-exception-code exc))
                        (sfun-conversion-exception-message exc))
                  'not-sfun-conversion-exception))
          > (with-exception-catcher handler t2)
          (#<procedure #2 f>
           ()
           "(Argument 1) Can't convert from C nonnull-char-string"
           #f)
          > (with-exception-catcher handler t3)
          (#<procedure #2 f> () "Can't convert result to C int" #f)


 -- procedure: multiple-c-return-exception? OBJ
     Multiple-c-return-exception objects are raised by the C-interface
     when a C to Scheme procedure call returns and that call's stack
     frame is no longer on the C stack because the call has already
     returned, or has been removed from the C stack by a `longjump'.

     The procedure `multiple-c-return-exception?' returns `#t' when OBJ
     is a multiple-c-return-exception object and `#f' otherwise.

     For example:

          $ cat test3.scm
          (c-define (f str) (char-string) scheme-object "f" ""
            (pp (list 'entry 'str= str))
            (let ((k (call-with-current-continuation (lambda (k) k))))
              (pp (list 'exit 'k= k))
              k))
          (define scheme-to-c-to-scheme-and-back
            (c-lambda (char-string) scheme-object
              "___return(f (___arg1));"))
          $ gsc test3.scm
          $ gsi
          Gambit v4.9.4

          > (load "test3")
          "/Users/feeley/gambit/doc/test3.o1"
          > (define (handler exc)
              (if (multiple-c-return-exception? exc)
                  exc
                  'not-multiple-c-return-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (let ((c (scheme-to-c-to-scheme-and-back "hello")))
                  (pp c)
                  (c 999))))
          (entry str= "hello")
          (exit k= #<procedure #2>)
          #<procedure #2>
          (exit k= 999)
          #<multiple-c-return-exception #3>


 -- procedure: wrong-processor-c-return-exception? OBJ
     Wrong-processor-c-return-exception objects are raised by the
     runtime system when a C to Scheme procedure call returns and that
     call's stack frame was created by another processor.

     The procedure `wrong-processor-c-return-exception?' returns `#t'
     when OBJ is a wrong-processor-c-return-exception object and `#f'
     otherwise.


12.6 Exception objects related to the reader
============================================

 -- procedure: datum-parsing-exception? OBJ
 -- procedure: datum-parsing-exception-kind EXC
 -- procedure: datum-parsing-exception-parameters EXC
 -- procedure: datum-parsing-exception-readenv EXC
     Datum-parsing-exception objects are raised by the reader (i.e. the
     `read' procedure) when the input does not conform to the grammar
     for datum.  The parameter EXC must be a datum-parsing-exception
     object.

     The procedure `datum-parsing-exception?' returns `#t' when OBJ is
     a datum-parsing-exception object and `#f' otherwise.

     The procedure `datum-parsing-exception-kind' returns a symbol
     denoting the kind of parsing error that was encountered by the
     reader when it raised EXC.  Here is a table of the possible return
     values:

     `datum-or-eof-expected'            Datum or EOF expected
     `datum-expected'                   Datum expected
     `improperly-placed-dot'            Improperly placed dot
     `incomplete-form-eof-reached'      Incomplete form, EOF reached
     `incomplete-form'                  Incomplete form
     `character-out-of-range'           Character out of range
     `invalid-character-name'           Invalid '#\' name
     `illegal-character'                Illegal character
     `s8-expected'                      Signed 8 bit exact integer
                                        expected
     `u8-expected'                      Unsigned 8 bit exact integer
                                        expected
     `s16-expected'                     Signed 16 bit exact integer
                                        expected
     `u16-expected'                     Unsigned 16 bit exact integer
                                        expected
     `s32-expected'                     Signed 32 bit exact integer
                                        expected
     `u32-expected'                     Unsigned 32 bit exact integer
                                        expected
     `s64-expected'                     Signed 64 bit exact integer
                                        expected
     `u64-expected'                     Unsigned 64 bit exact integer
                                        expected
     `inexact-real-expected'            Inexact real expected
     `invalid-hex-escape'               Invalid hexadecimal escape
     `invalid-escaped-character'        Invalid escaped character
     `open-paren-expected'              '(' expected
     `invalid-token'                    Invalid token
     `invalid-sharp-bang-name'          Invalid '#!' name
     `duplicate-label-definition'       Duplicate definition for label
     `missing-label-definition'         Missing definition for label
     `illegal-label-definition'         Illegal definition of label
     `invalid-infix-syntax-character'   Invalid infix syntax character
     `invalid-infix-syntax-number'      Invalid infix syntax number
     `invalid-infix-syntax'             Invalid infix syntax

     The procedure `datum-parsing-exception-parameters' returns a list
     of the parameters associated with the parsing error that was
     encountered by the reader when it raised EXC.

     For example:

          > (define (handler exc)
              (if (datum-parsing-exception? exc)
                  (list (datum-parsing-exception-kind exc)
                        (datum-parsing-exception-parameters exc))
                  'not-datum-parsing-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (with-input-from-string "(s #\\pace)" read)))
          (invalid-character-name ("pace"))


12.7 Exception objects related to evaluation and compilation
============================================================

 -- procedure: expression-parsing-exception? OBJ
 -- procedure: expression-parsing-exception-kind EXC
 -- procedure: expression-parsing-exception-parameters EXC
 -- procedure: expression-parsing-exception-source EXC
     Expression-parsing-exception objects are raised by the evaluator
     and compiler (i.e. the procedures `eval', `compile-file', etc)
     when the input does not conform to the grammar for expression.  The
     parameter EXC must be a expression-parsing-exception object.

     The procedure `expression-parsing-exception?' returns `#t' when
     OBJ is a expression-parsing-exception object and `#f' otherwise.

     The procedure `expression-parsing-exception-kind' returns a symbol
     denoting the kind of parsing error that was encountered by the
     evaluator or compiler when it raised EXC.  Here is a table of the
     possible return values:

     `id-expected'                      Identifier expected
     `ill-formed-namespace'             Ill-formed namespace
     `ill-formed-namespace-prefix'      Ill-formed namespace prefix
     `namespace-prefix-must-be-string'  Namespace prefix must be a string
     `macro-used-as-variable'           Macro name can't be used as a
                                        variable
     `variable-is-immutable'            Variable is immutable
     `ill-formed-macro-transformer'     Macro transformer must be a
                                        lambda expression
     `reserved-used-as-variable'        Reserved identifier can't be used
                                        as a variable
     `ill-formed-special-form'          Ill-formed special form
     `cannot-open-file'                 Can't open file
     `filename-expected'                Filename expected
     `ill-placed-define'                Ill-placed 'define'
     `ill-placed-**include'             Ill-placed '##include'
     `ill-placed-**define-macro'        Ill-placed '##define-macro'
     `ill-placed-**declare'             Ill-placed '##declare'
     `ill-placed-**namespace'           Ill-placed '##namespace'
     `ill-formed-expression'            Ill-formed expression
     `unsupported-special-form'         Interpreter does not support
     `ill-placed-unquote'               Ill-placed 'unquote'
     `ill-placed-unquote-splicing'      Ill-placed 'unquote-splicing'
     `parameter-must-be-id'             Parameter must be an identifier
     `parameter-must-be-id-or-default'  Parameter must be an identifier
                                        or default binding
     `duplicate-parameter'              Duplicate parameter in parameter
                                        list
     `ill-placed-dotted-rest-parameter' Ill-placed dotted rest parameter
     `parameter-expected-after-rest'    #!rest must be followed by a
                                        parameter
     `ill-formed-default'               Ill-formed default binding
     `ill-placed-optional'              Ill-placed #!optional
     `ill-placed-rest'                  Ill-placed #!rest
     `ill-placed-key'                   Ill-placed #!key
     `key-expected-after-rest'          #!key expected after rest
                                        parameter
     `ill-placed-default'               Ill-placed default binding
     `duplicate-variable-definition'    Duplicate definition of a variable
     `empty-body'                       Body must contain at least one
                                        expression
     `variable-must-be-id'              Defined variable must be an
                                        identifier
     `else-clause-not-last'             Else clause must be last
     `ill-formed-selector-list'         Ill-formed selector list
     `duplicate-variable-binding'       Duplicate variable in bindings
     `ill-formed-binding-list'          Ill-formed binding list
     `ill-formed-call'                  Ill-formed procedure call
     `ill-formed-cond-expand'           Ill-formed 'cond-expand'
     `unfulfilled-cond-expand'          Unfulfilled 'cond-expand'

     The procedure `expression-parsing-exception-parameters' returns a
     list of the parameters associated with the parsing error that was
     encountered by the evaluator or compiler when it raised EXC.

     For example:

          > (define (handler exc)
              (if (expression-parsing-exception? exc)
                  (list (expression-parsing-exception-kind exc)
                        (expression-parsing-exception-parameters exc))
                  'not-expression-parsing-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (eval '(+ do 1))))
          (reserved-used-as-variable (do))


 -- procedure: unbound-global-exception? OBJ
 -- procedure: unbound-global-exception-variable EXC
 -- procedure: unbound-global-exception-code EXC
 -- procedure: unbound-global-exception-rte EXC
     Unbound-global-exception objects are raised when an unbound global
     variable is accessed.  The parameter EXC must be an
     unbound-global-exception object.

     The procedure `unbound-global-exception?' returns `#t' when OBJ is
     an unbound-global-exception object and `#f' otherwise.

     The procedure `unbound-global-exception-variable' returns a symbol
     identifying the unbound global variable.

     For example:

          > (define (handler exc)
              (if (unbound-global-exception? exc)
                  (list 'variable= (unbound-global-exception-variable exc))
                  'not-unbound-global-exception))
          > (with-exception-catcher
              handler
              (lambda () foo))
          (variable= foo)


 -- procedure: not-in-compilation-context-exception? OBJ
 -- procedure: not-in-compilation-context-exception-procedure EXC
 -- procedure: not-in-compilation-context-exception-arguments EXC
     Not-in-compilation-context-exception objects are raised by the
     procedure `compilation-target' when it is executed outside of a
     compilation context.  The parameter EXC must be a
     not-in-compilation-context-exception object.

     The procedure `not-in-compilation-context-exception?' returns `#t'
     when OBJ is a not-in-compilation-context-exception object and `#f'
     otherwise.

     The procedure `not-in-compilation-context-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `not-in-compilation-context-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (not-in-compilation-context-exception? exc)
                  (list (not-in-compilation-context-exception-procedure exc)
                        (not-in-compilation-context-exception-arguments exc))
                  'not-not-in-compilation-context-exception))
          > (with-exception-catcher
              handler
              (lambda () (compilation-target)))
          (#<procedure #2 compilation-target> ())


12.8 Exception objects related to type checking
===============================================

 -- procedure: type-exception? OBJ
 -- procedure: type-exception-procedure EXC
 -- procedure: type-exception-arguments EXC
 -- procedure: type-exception-arg-id EXC
 -- procedure: type-exception-type-id EXC
     Type-exception objects are raised when a primitive procedure is
     called with an argument of incorrect type (i.e. when a run time
     type-check fails).  The parameter EXC must be a type-exception
     object.

     The procedure `type-exception?' returns `#t' when OBJ is a
     type-exception object and `#f' otherwise.

     The procedure `type-exception-procedure' returns the procedure
     that raised EXC.

     The procedure `type-exception-arguments' returns the list of
     arguments of the procedure that raised EXC.

     The procedure `type-exception-arg-id' returns the identity of the
     argument whose type is incorrect, which can be an exact integer
     position (1 for the first argument) or a pair whose `car' is the
     position and the `cdr' is the parameter name as a symbol.

     The procedure `type-exception-type-id' returns an identifier of
     the type expected.  The type-id can be a symbol, such as `number'
     and `string-or-nonnegative-fixnum', or a record type descriptor.

     For example:

          > (define (handler exc)
              (if (type-exception? exc)
                  (list (type-exception-procedure exc)
                        (type-exception-arguments exc)
                        (type-exception-arg-id exc)
                        (type-exception-type-id exc))
                  'not-type-exception))
          > (with-exception-catcher
              handler
              (lambda () (vector-ref '#(a b c) 'foo)))
          (#<procedure #2 vector-ref> (#(a b c) foo) 2 exact-integer)
          > (with-exception-catcher
              handler
              (lambda () (time->seconds 'foo)))
          (#<procedure #3 time->seconds> (foo) 1 #<type #4 time>)


 -- procedure: range-exception? OBJ
 -- procedure: range-exception-procedure EXC
 -- procedure: range-exception-arguments EXC
 -- procedure: range-exception-arg-id EXC
     Range-exception objects are raised when a numeric parameter is not
     in the allowed range.  The parameter EXC must be a range-exception
     object.

     The procedure `range-exception?' returns `#t' when OBJ is a
     range-exception object and `#f' otherwise.

     The procedure `range-exception-procedure' returns the procedure
     that raised EXC.

     The procedure `range-exception-arguments' returns the list of
     arguments of the procedure that raised EXC.

     The procedure `range-exception-arg-id' returns the identity of the
     argument which is not in the allowed range, which can be an exact
     integer position (1 for the first argument) or a pair whose `car'
     is the position and the `cdr' is the parameter name as a symbol.

     For example:

          > (define (handler exc)
              (if (range-exception? exc)
                  (list (range-exception-procedure exc)
                        (range-exception-arguments exc)
                        (range-exception-arg-id exc))
                  'not-range-exception))
          > (with-exception-catcher
              handler
              (lambda () (string-ref "abcde" 10)))
          (#<procedure #2 string-ref> ("abcde" 10) 2)


 -- procedure: divide-by-zero-exception? OBJ
 -- procedure: divide-by-zero-exception-procedure EXC
 -- procedure: divide-by-zero-exception-arguments EXC
     Divide-by-zero-exception objects are raised when a division by
     zero is attempted.  The parameter EXC must be a
     divide-by-zero-exception object.

     The procedure `divide-by-zero-exception?' returns `#t' when OBJ is
     a divide-by-zero-exception object and `#f' otherwise.

     The procedure `divide-by-zero-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `divide-by-zero-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (divide-by-zero-exception? exc)
                  (list (divide-by-zero-exception-procedure exc)
                        (divide-by-zero-exception-arguments exc))
                  'not-divide-by-zero-exception))
          > (with-exception-catcher
              handler
              (lambda () (/ 5 0 7)))
          (#<procedure #2 /> (5 0 7))


 -- procedure: length-mismatch-exception? OBJ
 -- procedure: length-mismatch-exception-procedure EXC
 -- procedure: length-mismatch-exception-arguments EXC
 -- procedure: length-mismatch-exception-arg-id EXC
     Length-mismatch-exception objects are raised by some procedures
     when they are called with two or more list arguments and the lists
     are not of the same length.  The parameter EXC must be a
     length-mismatch-exception object.

     The procedure `length-mismatch-exception?' returns `#t' when OBJ
     is an length-mismatch-exception object and `#f' otherwise.

     The procedure `length-mismatch-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `length-mismatch-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     The procedure `length-mismatch-exception-arg-id' returns the
     identity of the argument whose length is the shortest, which can
     be an exact integer position (1 for the first argument) or a pair
     whose `car' is the position and the `cdr' is the parameter name as
     a symbol.


12.9 Exception objects related to procedure call
================================================

 -- procedure: wrong-number-of-arguments-exception? OBJ
 -- procedure: wrong-number-of-arguments-exception-procedure EXC
 -- procedure: wrong-number-of-arguments-exception-arguments EXC
     Wrong-number-of-arguments-exception objects are raised when a
     procedure is called with the wrong number of arguments.  The
     parameter EXC must be a wrong-number-of-arguments-exception object.

     The procedure `wrong-number-of-arguments-exception?' returns `#t'
     when OBJ is a wrong-number-of-arguments-exception object and `#f'
     otherwise.

     The procedure `wrong-number-of-arguments-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `wrong-number-of-arguments-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (wrong-number-of-arguments-exception? exc)
                  (list (wrong-number-of-arguments-exception-procedure exc)
                        (wrong-number-of-arguments-exception-arguments exc))
                  'not-wrong-number-of-arguments-exception))
          > (with-exception-catcher
              handler
              (lambda () (open-input-file "data" 99)))
          (#<procedure #2 open-input-file> ("data" 99))


 -- procedure: number-of-arguments-limit-exception? OBJ
 -- procedure: number-of-arguments-limit-exception-procedure EXC
 -- procedure: number-of-arguments-limit-exception-arguments EXC
     Number-of-arguments-limit-exception objects are raised by the
     `apply' procedure when the procedure being called is passed more
     than 8192 arguments.  The parameter EXC must be a
     number-of-arguments-limit-exception object.

     The procedure `number-of-arguments-limit-exception?' returns `#t'
     when OBJ is a number-of-arguments-limit-exception object and `#f'
     otherwise.

     The procedure `number-of-arguments-limit-exception-procedure'
     returns the target procedure of the call to `apply' that raised
     EXC.

     The procedure `number-of-arguments-limit-exception-arguments'
     returns the list of arguments of the target procedure of the call
     to `apply' that raised EXC.

     For example:

          > (define (iota n) (if (= n 0) '() (cons n (iota (- n 1)))))
          > (define (handler exc)
              (if (number-of-arguments-limit-exception? exc)
                  (list (number-of-arguments-limit-exception-procedure exc)
                        (length (number-of-arguments-limit-exception-arguments exc)))
                  'not-number-of-arguments-limit-exception))
          > (with-exception-catcher
              handler
              (lambda () (apply + 1 2 3 (iota 8190))))
          (#<procedure #2 +> 8193)


 -- procedure: nonprocedure-operator-exception? OBJ
 -- procedure: nonprocedure-operator-exception-operator EXC
 -- procedure: nonprocedure-operator-exception-arguments EXC
 -- procedure: nonprocedure-operator-exception-code EXC
 -- procedure: nonprocedure-operator-exception-rte EXC
     Nonprocedure-operator-exception objects are raised when a procedure
     call is executed and the operator position is not a procedure.  The
     parameter EXC must be an nonprocedure-operator-exception object.

     The procedure `nonprocedure-operator-exception?' returns `#t' when
     OBJ is an nonprocedure-operator-exception object and `#f'
     otherwise.

     The procedure `nonprocedure-operator-exception-operator' returns
     the value in operator position of the procedure call that raised
     EXC.

     The procedure `nonprocedure-operator-exception-arguments' returns
     the list of arguments of the procedure call that raised EXC.

     For example:

          > (define (handler exc)
              (if (nonprocedure-operator-exception? exc)
                  (list (nonprocedure-operator-exception-operator exc)
                        (nonprocedure-operator-exception-arguments exc))
                  'not-nonprocedure-operator-exception))
          > (with-exception-catcher
              handler
              (lambda () (11 22 33)))
          (11 (22 33))


 -- procedure: wrong-number-of-values-exception? OBJ
 -- procedure: wrong-number-of-values-exception-vals EXC
 -- procedure: wrong-number-of-values-exception-code EXC
 -- procedure: wrong-number-of-values-exception-rte EXC
     Wrong-number-of-values-exception objects are raised by the
     `let-values' and `define-values' forms when the number of values
     does not conform to the number of variables to be bound.  The
     parameter EXC must be an wrong-number-of-values-exception object.

     The procedure `wrong-number-of-values-exception?' returns `#t'
     when OBJ is an wrong-number-of-values-exception object and `#f'
     otherwise.

     The procedure `wrong-number-of-values-exception-vals' returns the
     values that were to be bound.

     For example:

          > (define (handler exc)
              (if (wrong-number-of-values-exception? exc)
                  (call-with-values
                   (lambda () (wrong-number-of-values-exception-vals exc))
                   list)
                  'not-wrong-number-of-values-exception))
          > (with-exception-catcher
              handler
              (lambda () (let-values (((a b) (values 11 22 33))) (* a b))))
          (11 22 33)


 -- procedure: unknown-keyword-argument-exception? OBJ
 -- procedure: unknown-keyword-argument-exception-procedure EXC
 -- procedure: unknown-keyword-argument-exception-arguments EXC
     Unknown-keyword-argument-exception objects are raised when a
     procedure accepting keyword arguments is called and one of the
     keywords supplied is not among those that are expected.  The
     parameter EXC must be an unknown-keyword-argument-exception object.

     The procedure `unknown-keyword-argument-exception?' returns `#t'
     when OBJ is an unknown-keyword-argument-exception object and `#f'
     otherwise.

     The procedure `unknown-keyword-argument-exception-procedure'
     returns the procedure that raised EXC.

     The procedure `unknown-keyword-argument-exception-arguments'
     returns the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (unknown-keyword-argument-exception? exc)
                  (list (unknown-keyword-argument-exception-procedure exc)
                        (unknown-keyword-argument-exception-arguments exc))
                  'not-unknown-keyword-argument-exception))
          > (with-exception-catcher
              handler
              (lambda () ((lambda (#!key (foo 5)) foo) bar: 11)))
          (#<procedure #2> (bar: 11))


 -- procedure: keyword-expected-exception? OBJ
 -- procedure: keyword-expected-exception-procedure EXC
 -- procedure: keyword-expected-exception-arguments EXC
     Keyword-expected-exception objects are raised when a procedure
     accepting keyword arguments is called and a nonkeyword was supplied
     where a keyword was expected.  The parameter EXC must be an
     keyword-expected-exception object.

     The procedure `keyword-expected-exception?' returns `#t' when OBJ
     is an keyword-expected-exception object and `#f' otherwise.

     The procedure `keyword-expected-exception-procedure' returns the
     procedure that raised EXC.

     The procedure `keyword-expected-exception-arguments' returns the
     list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (keyword-expected-exception? exc)
                  (list (keyword-expected-exception-procedure exc)
                        (keyword-expected-exception-arguments exc))
                  'not-keyword-expected-exception))
          > (with-exception-catcher
              handler
              (lambda () ((lambda (#!key (foo 5)) foo) 11 22)))
          (#<procedure #2> (11 22))


12.10 Other exception objects
=============================

 -- procedure: error-exception? OBJ
 -- procedure: error-exception-message EXC
 -- procedure: error-exception-parameters EXC
 -- procedure: error MESSAGE OBJ...
     Error-exception objects are raised when the `error' procedure is
     called.  The parameter EXC must be an error-exception object.

     The procedure `error-exception?' returns `#t' when OBJ is an
     error-exception object and `#f' otherwise.

     The procedure `error-exception-message' returns the first argument
     of the call to `error' that raised EXC.

     The procedure `error-exception-parameters' returns the list of
     arguments, starting with the second argument, of the call to
     `error' that raised EXC.

     The `error' procedure raises an error-exception object whose
     message field is MESSAGE and parameters field is the list of
     values OBJ....

     For example:

          > (define (handler exc)
              (if (error-exception? exc)
                  (list (error-exception-message exc)
                        (error-exception-parameters exc))
                  'not-error-exception))
          > (with-exception-catcher
              handler
              (lambda () (error "unexpected object:" 123)))
          ("unexpected object:" (123))


13 Host environment
*******************

The host environment is the set of resources, such as the filesystem,
network and processes, that are managed by the operating system within
which the Scheme program is executing.  This chapter specifies how the
host environment can be accessed from within the Scheme program.

   In this chapter we say that the Scheme program being executed is a
process, even though the concept of process does not exist in some
operating systems supported by Gambit (e.g. MSDOS).

13.1 Handling of file names
===========================

Gambit uses a naming convention for files that is compatible with the
one used by the host environment but extended to allow referring to the
"home directory" of the current user or some specific user and the
"installation directories".

   A "path" is a string that denotes a file, for example
`"src/readme.txt"'.  Each component of a path is separated by a `/'
under UNIX and macOS and by a `/' or `\' under MSDOS and Microsoft
Windows.  A leading separator indicates an absolute path under UNIX,
macOS, MSDOS and Microsoft Windows.  A path which does not contain a
path separator is relative to the "current working directory" on all
operating systems.  A volume specifier such as `C:' may prefix a file
name under MSDOS and Microsoft Windows.

   A path which starts with the characters `~~' denotes a file in an
installation directory.  If nothing follows the `~~' then the directory
denoted is the central installation directory.  Otherwise what follows
the `~~' is the name of the installation directory, for example `~~lib'
denotes the `lib' installation directory.  Note that the location of
the installation directories may be overridden by using the
`-:~~NAME=DIRECTORY' runtime option or by defining the `GAMBOPT'
environment variable.  Unless explicitly overridden, `~~execdir'
denotes the directory containing the current executable program.

   A path which starts with the character `~' not followed by `~'
denotes a file in the user's home directory.  The user's home directory
is contained in the `HOME' environment variable under UNIX, macOS,
MSDOS and Microsoft Windows.  Under MSDOS and Microsoft Windows, if the
`HOME' environment variable is not defined, the environment variables
`HOMEDRIVE' and `HOMEPATH' are concatenated if they are defined.  If
this fails to yield a home directory, the central installation
directory is used instead.

   A path which starts with the characters `~USERNAME' denotes a file
in the home directory of the given user.  Under UNIX and macOS this is
found using the password file.  There is no equivalent under MSDOS and
Microsoft Windows.

 -- procedure: initial-current-directory
 -- procedure: current-directory [NEW-CURRENT-DIRECTORY]
     The procedure `initial-current-directory' returns the absolute
     "normalized path" of the current working directory of the current
     process when it was started.

     The parameter object `current-directory' is bound to the current
     working directory.  Calling this procedure with no argument returns
     the absolute "normalized path" of the directory and calling this
     procedure with one argument sets the directory to
     NEW-CURRENT-DIRECTORY.  The initial binding of this parameter
     object is the path returned by `initial-current-directory'.  The
     path returned by `current-directory' always contains a trailing
     directory separator.  Modifications of the parameter object do not
     change the current working directory of the current process (i.e.
     that is accessible with the UNIX `getcwd()' function and the
     Microsoft Windows `GetCurrentDirectory' function).  It is an error
     to mutate the string returned by `current-directory'.

     For example under UNIX:

          > (current-directory)
          "/Users/feeley/gambit/doc/"
          > (current-directory "..")
          > (current-directory)
          "/Users/feeley/gambit/"
          > (initial-current-directory)
          "/Users/feeley/gambit/doc/"
          > (path-expand "foo" "~~")
          "/usr/local/Gambit/foo"
          > (parameterize ((current-directory "~~")) (path-expand "foo"))
          "/usr/local/Gambit/foo"


 -- procedure: path-expand PATH [ORIGIN-DIRECTORY]
     The procedure `path-expand' takes the path of a file or directory
     and returns an expanded path, which is an absolute path when PATH
     or ORIGIN-DIRECTORY are absolute paths.  The optional
     ORIGIN-DIRECTORY parameter, which defaults to the current working
     directory, is the directory used to resolve relative paths.
     Components of the paths PATH and ORIGIN-DIRECTORY need not exist.

     For example under UNIX:

          > (path-expand "foo")
          "/Users/feeley/gambit/doc/foo"
          > (path-expand "~/foo")
          "/Users/feeley/foo"
          > (path-expand "~~lib/foo")
          "/usr/local/Gambit/lib/foo"
          > (path-expand "../foo")
          "/Users/feeley/gambit/doc/../foo"
          > (path-expand "foo" "")
          "foo"
          > (path-expand "foo" "/tmp")
          "/tmp/foo"
          > (path-expand "this/file/does/not/exist")
          "/Users/feeley/gambit/doc/this/file/does/not/exist"
          > (path-expand "")
          "/Users/feeley/gambit/doc/"


 -- procedure: path-normalize PATH [ALLOW-RELATIVE? [ORIGIN-DIRECTORY]]
     The procedure `path-normalize' takes a path of a file or directory
     and returns its normalized path.  The optional ORIGIN-DIRECTORY
     parameter, which defaults to the current working directory, is the
     directory used to resolve relative paths.  All components of the
     paths PATH and ORIGIN-DIRECTORY must exist, except possibly the
     last component of PATH.  A normalized path is a path containing no
     redundant parts and which is consistent with the current structure
     of the filesystem.  A normalized path of a directory will always
     end with a path separator (i.e. `/', `\', or `:' depending on the
     operating system).  The optional ALLOW-RELATIVE? parameter, which
     defaults to `#f', indicates if the path returned can be expressed
     relatively to ORIGIN-DIRECTORY: a `#f' requests an absolute path,
     the symbol `shortest' requests the shortest of the absolute and
     relative paths, and any other value requests the relative path.
     The shortest path is useful for interaction with the user because
     short relative paths are typically easier to read than long
     absolute paths.

     For example under UNIX:

          > (path-expand "../foo")
          "/Users/feeley/gambit/doc/../foo"
          > (path-normalize "../foo")
          "/Users/feeley/gambit/foo"
          > (path-normalize "this/file/does/not/exist")
          *** ERROR IN (console)@3.1 -- No such file or directory
          (path-normalize "this/file/does/not/exist")


 -- procedure: path-extension PATH
 -- procedure: path-strip-extension PATH
 -- procedure: path-directory PATH
 -- procedure: path-strip-directory PATH
 -- procedure: path-strip-trailing-directory-separator PATH
 -- procedure: path-volume PATH
 -- procedure: path-strip-volume PATH
     These procedures extract various parts of a path, which need not
     be a normalized path.  The procedure `path-extension' returns the
     file extension (including the period) or the empty string if there
     is no extension.  The procedure `path-strip-extension' returns the
     path with the extension stripped off.  The procedure
     `path-directory' returns the file's directory (including the last
     path separator) or the empty string if no directory is specified
     in the path.  The procedure `path-strip-directory' returns the
     path with the directory stripped off.  The procedure
     `path-strip-trailing-directory-separator' returns the path with
     the directory separator stripped off if one is at the end of the
     path.  The procedure `path-volume' returns the file's volume
     (including the last path separator) or the empty string if no
     volume is specified in the path.  The procedure
     `path-strip-volume' returns the path with the volume stripped off.

     For example under UNIX:

          > (path-extension "/tmp/foo")
          ""
          > (path-extension "/tmp/foo.txt")
          ".txt"
          > (path-strip-extension "/tmp/foo.txt")
          "/tmp/foo"
          > (path-directory "/tmp/foo.txt")
          "/tmp/"
          > (path-strip-directory "/tmp/foo.txt")
          "foo.txt"
          > (path-strip-trailing-directory-separator "/usr/local/bin/")
          "/usr/local/bin"
          > (path-strip-trailing-directory-separator "/usr/local/bin")
          "/usr/local/bin"
          > (path-volume "/tmp/foo.txt")
          ""
          > (path-volume "C:/tmp/foo.txt")
          "" ; result is "C:" under Microsoft Windows
          > (path-strip-volume "C:/tmp/foo.txt")
          "C:/tmp/foo.txt" ; result is "/tmp/foo.txt" under Microsoft Windows


13.2 Filesystem operations
==========================

 -- procedure: create-directory PATH-OR-SETTINGS
 -- procedure: create-temporary-directory PATH-OR-SETTINGS
     These procedures create directories.  The argument
     PATH-OR-SETTINGS is either a string denoting a filesystem path or
     a list of port settings which must contain a `path:' setting.  The
     procedure `create-directory' returns an unspecified value.  In the
     case of `create-temporary-directory' the path is used as a prefix
     to generate new directory paths until the path of a directory not
     currently existing is generated and that path is returned.  Here
     are the settings allowed:

        * `path:' STRING

          This setting indicates the location of the directory to
          create in the filesystem.  There is no default value for this
          setting.

        * `permissions:' 12-BIT-EXACT-INTEGER

          This setting controls the UNIX permissions that will be
          attached to the file if it is created.  The default value of
          this setting is `#o777'.


     For example:

          > (create-directory "newdir")
          > (create-temporary-directory "/tmp/foo.")
          "/tmp/foo.85812"
          > (create-directory "newdir")
          *** ERROR IN (console)@2.1 -- File exists
          (create-directory "newdir")


 -- procedure: create-fifo PATH-OR-SETTINGS
     This procedure creates a FIFO.  The argument PATH-OR-SETTINGS is
     either a string denoting a filesystem path or a list of port
     settings which must contain a `path:' setting.  Here are the
     settings allowed:

        * `path:' STRING

          This setting indicates the location of the FIFO to create in
          the filesystem.  There is no default value for this setting.

        * `permissions:' 12-BIT-EXACT-INTEGER

          This setting controls the UNIX permissions that will be
          attached to the file if it is created.  The default value of
          this setting is `#o666'.


     For example:

          > (create-fifo "fifo")
          > (define a (open-input-file "fifo"))
          > (define b (open-output-file "fifo"))
          > (display "1 22 333" b)
          > (force-output b)
          > (read a)
          1
          > (read a)
          22


 -- procedure: create-link SOURCE-PATH DESTINATION-PATH
     This procedure creates a hard link between SOURCE-PATH and
     DESTINATION-PATH.  The argument SOURCE-PATH must be a string
     denoting the path of an existing file.  The argument
     DESTINATION-PATH must be a string denoting the path of the link to
     create.


 -- procedure: create-symbolic-link SOURCE-PATH DESTINATION-PATH
     This procedure creates a symbolic link between SOURCE-PATH and
     DESTINATION-PATH.  The argument SOURCE-PATH must be a string
     denoting the path of an existing file.  The argument
     DESTINATION-PATH must be a string denoting the path of the
     symbolic link to create.


 -- procedure: rename-file SOURCE-PATH DESTINATION-PATH [REPLACE?]
     This procedure renames the file SOURCE-PATH to DESTINATION-PATH.
     The argument SOURCE-PATH must be a string denoting the path of an
     existing file.  The argument DESTINATION-PATH must be a string
     denoting the new path of the file.  If REPLACE? is absent or true,
     an existing DESTINATION-PATH will be replaced by SOURCE-PATH.
     Otherwise, the rename operation will fail if DESTINATION-PATH
     exists.  Not all filesystems support atomic renaming and existence
     testing.


 -- procedure: copy-file SOURCE-PATH DESTINATION-PATH
     This procedure copies the file SOURCE-PATH to DESTINATION-PATH.
     The argument SOURCE-PATH must be a string denoting the path of an
     existing file.  The argument DESTINATION-PATH must be a string
     denoting the path of the file to create.


 -- procedure: delete-file PATH
     This procedure deletes the file PATH.  The argument PATH must be a
     string denoting the path of an existing file.


 -- procedure: delete-directory PATH
     This procedure deletes the directory PATH.  The argument PATH must
     be a string denoting the path of an existing empty directory.


 -- procedure: delete-file-or-directory PATH [RECURSIVE?]
     This procedure deletes the file or directory PATH.  The argument
     PATH must be a string denoting the path of an existing file or
     directory.  If RECURSIVE? is specified and is true, directories
     are recursively deleted.  Otherwise only empty directories can be
     deleted.


 -- procedure: directory-files [PATH-OR-SETTINGS]
     This procedure returns the list of the files in a directory.  The
     argument PATH-OR-SETTINGS is either a string denoting a filesystem
     path to a directory or a list of settings which must contain a
     `path:' setting.  If it is not specified, PATH-OR-SETTINGS
     defaults to the current directory (the value bound to the
     `current-directory' parameter object).  Here are the settings
     allowed:

        * `path:' STRING

          This setting indicates the location of the directory in the
          filesystem.  There is no default value for this setting.

        * `ignore-hidden:' ( `#f' | `#t' | `dot-and-dot-dot' )

          This setting controls whether hidden-files will be returned.
          Under UNIX and macOS hidden-files are those that start with a
          period (such as `.', `..', and `.profile').  Under Microsoft
          Windows hidden files are the `.' and `..' entries and the
          files whose "hidden file" attribute is set.  A setting of `#f'
          will enumerate all the files.  A setting of `#t' will only
          enumerate the files that are not hidden.  A setting of
          `dot-and-dot-dot' will enumerate all the files except for the
          `.' and `..' hidden files.  The default value of this setting
          is `#t'.


     For example:

          > (directory-files)
          ("complex" "README" "simple")
          > (directory-files "../include")
          ("config.h" "config.h.in" "gambit.h" "makefile" "makefile.in")
          > (directory-files (list path: "../include" ignore-hidden: #f))
          ("." ".." "config.h" "config.h.in" "gambit.h" "makefile" "makefile.in")


13.3 Shell command execution
============================

 -- procedure: shell-command COMMAND [CAPTURE?]
     The procedure `shell-command' calls up the shell to execute
     COMMAND which must be a string.  The argument CAPTURE?, which
     defaults to `#f', indicates if the output of the command is
     captured as a string.  If CAPTURE? is `#f', this procedure returns
     the exit status of the shell in the form that the C library's
     `system' routine returns.  If CAPTURE? is not `#f', this procedure
     returns a pair consisting of the exit status of the shell in the
     `car' field, and the captured output in the `cdr' field.  Be
     advised that the shell that is used, and consequently the syntax
     of COMMAND, depends on the operating system.  On Unix, the shell
     `/bin/sh' is usually invoked.  On Windows, the shell `cmd.exe' is
     usually invoked.

     For example under UNIX:

          > (shell-command "ls -sk f*.scm")
          4 fact.scm   4 fib.scm
          0
          > (shell-command "ls -sk f*.scm" #t)
          (0 . "4 fact.scm   4 fib.scm\n")
          > (shell-command "echo x\\\\\\\\y $HOME" #t)
          (0 . "x\\y /Users/feeley\n")

     For example under Windows:

          > (shell-command "echo x\\\\\\\\y %HOME%" #t)
          (0 . "x\\\\\\\\y C:\\Users\\feeley\r\n")


13.4 Process termination
========================

 -- procedure: exit [STATUS]
     The procedure `exit' causes the process to terminate with the
     status STATUS which must be an exact integer in the range 0 to 255
     or `#f'.  If it is not specified, STATUS defaults to 0.  When
     STATUS is `#f' the process terminates with an error status.

     For example under UNIX:

          $ gsi
          Gambit v4.9.4

          > (exit #f)
          $ echo $?
          70


13.5 Command line arguments
===========================

 -- procedure: command-line
     This procedure returns a list of strings corresponding to the
     command line arguments, including the program file name as the
     first element of the list.  When the interpreter executes a Scheme
     script, the list returned by `command-line' contains the script's
     absolute path followed by the remaining command line arguments.

     For example under UNIX:

          $ gsi -:debug -e "(pretty-print (command-line))"
          ("gsi" "-e" "(pretty-print (command-line))")
          $ cat foo
          #!/usr/local/Gambit/bin/gsi-script
          (pretty-print (command-line))
          $ ./foo 1 2 "3 4"
          ("/u/feeley/./foo" "1" "2" "3 4")


13.6 Environment variables
==========================

 -- procedure: getenv NAME [DEFAULT]
 -- procedure: setenv NAME [NEW-VALUE]
     The procedure `getenv' returns the value of the environment
     variable NAME of the current process.  Variable names are denoted
     with strings.  A string is returned if the environment variable is
     bound, otherwise DEFAULT is returned if it is specified, otherwise
     an exception is raised.

     The procedure `setenv' changes the binding of the environment
     variable NAME to NEW-VALUE which must be a string.  If NEW-VALUE
     is not specified the binding is removed.

     For example under UNIX:

          > (getenv "HOME")
          "/Users/feeley"
          > (getenv "DOES_NOT_EXIST" #f)
          #f
          > (setenv "DOES_NOT_EXIST" "it does now")
          > (getenv "DOES_NOT_EXIST" #f)
          "it does now"
          > (setenv "DOES_NOT_EXIST")
          > (getenv "DOES_NOT_EXIST" #f)
          #f
          > (getenv "DOES_NOT_EXIST")
          *** ERROR IN (console)@7.1 -- Unbound OS environment variable
          (getenv "DOES_NOT_EXIST")


13.7 Measuring time
===================

Procedures are available for measuring real time (aka "wall" time) and
cpu time (the amount of time the cpu has been executing the process).
The resolution of the real time and cpu time clock is operating system
dependent.  Typically the resolution of the cpu time clock is rather
coarse (measured in "ticks" of 1/60th or 1/100th of a second).  Real
time is internally computed relative to some arbitrary point in time
using floating point numbers, which means that there is a gradual loss
of resolution as time elapses.  Moreover, some operating systems report
time in number of ticks using a 32 bit integer so the value returned by
the time related procedures may wraparound much before any significant
loss of resolution occurs (for example 2.7 years if ticks are 1/50th of
a second).

 -- procedure: current-time
 -- procedure: time? OBJ
 -- procedure: time->seconds TIME
 -- procedure: seconds->time X
     The procedure `current-time' returns a "time object" representing
     the current point in real time.

     The procedure `time?' returns `#t' when OBJ is a time object and
     `#f' otherwise.

     The procedure `time->seconds' converts the time object TIME into
     an inexact real number representing the number of seconds elapsed
     since the "epoch" (which is 00:00:00 Coordinated Universal Time
     01-01-1970).

     The procedure `seconds->time' converts the real number X
     representing the number of seconds elapsed since the "epoch" into a
     time object.

     For example:

          > (current-time)
          #<time #2>
          > (time? (current-time))
          #t
          > (time? 123)
          #f
          > (time->seconds (current-time))
          1083118758.63973
          > (time->seconds (current-time))
          1083118759.909163
          > (seconds->time (+ 10 (time->seconds (current-time))
          #<time #3>  ; a time object representing 10 seconds in the future


 -- procedure: process-times
 -- procedure: cpu-time
 -- procedure: real-time
     The procedure `process-times' returns a three element f64vector
     containing the cpu time that has been used by the program and the
     real time that has elapsed since it was started.  The first element
     corresponds to "user" time in seconds, the second element
     corresponds to "system" time in seconds and the third element is
     the elapsed real time in seconds.  On operating systems that can't
     differentiate user and system time, the system time is zero.  On
     operating systems that can't measure cpu time, the user time is
     equal to the elapsed real time and the system time is zero.

     The procedure `cpu-time' returns the cpu time in seconds that has
     been used by the program (user time plus system time).

     The procedure `real-time' returns the real time that has elapsed
     since the program was started.

     For example:

          > (process-times)
          #f64(.02794 .021754 .159926176071167)
          > (cpu-time)
          .051223
          > (real-time)
          .40660619735717773


 -- special form: time expr [PORT]
     The `time' special form evaluates expr and returns the result.  As
     a side effect it displays a message on the port PORT which
     indicates various statistics about the evaluation of expr
     including how long the evaluation took (in real time and cpu time),
     how much time was spent in the garbage collector, how much memory
     was allocated during the evaluation and how many minor and major
     page faults occured (0 is reported if not running under UNIX).  If
     it is not specified, PORT defaults to the interaction channel
     (i.e. the output will appear at the REPL).

     For example:

          > (define (f x)
              (let loop ((x x) (lst '()))
                (if (= x 0)
                    lst
                    (loop (- x 1) (cons x lst)))))
          > (length (time (f 100000)))
          (time (f 100000))
              683 ms real time
              558 ms cpu time (535 user, 23 system)
              8 collections accounting for 102 ms real time (70 user, 5 system)
              6400160 bytes allocated
              no minor faults
              no major faults
          100000


13.8 File information
=====================

 -- procedure: file-exists? PATH [CHASE?]
     The PATH argument must be a string.  This procedure returns `#t'
     when a file by that name exists, and returns `#f' otherwise.

     When CHASE? is present and `#f', symbolic links will not be
     chased, in other words if PATH refers to a symbolic link,
     `file-exists?' will return `#t' whether or not it points to an
     existing file.

     For example:

          > (file-exists? "nofile")
          #f


 -- procedure: file-info PATH [CHASE?]
     This procedure accesses the filesystem to get information about the
     file whose location is given by the string PATH.  A
     file-information record is returned that contains the file's type,
     the device number, the inode number, the mode (permission bits),
     the number of links, the file's user id, the file's group id, the
     file's size in bytes, the times of last-access, last-modification
     and last-change, the attributes, and the creation time.

     When CHASE? is present and `#f', symbolic links will not be
     chased, in other words if PATH refers to a symbolic link the
     `file-info' procedure will return information about the link
     rather than the file it links to.

     For example:

          > (file-info "/dev/tty")
          #<file-info #2
             type: character-special
             device: 19513156
             inode: 20728196
             mode: 438
             number-of-links: 1
             owner: 0
             group: 0
             size: 0
             last-access-time: #<time #3>
             last-modification-time: #<time #4>
             last-change-time: #<time #5>
             attributes: 128
             creation-time: #<time #6>>


 -- procedure: file-info? OBJ
     This procedure returns `#t' when OBJ is a file-information record
     and `#f' otherwise.

     For example:

          > (file-info? (file-info "/dev/tty"))
          #t
          > (file-info? 123)
          #f


 -- procedure: file-info-type FILE-INFO
     Returns the type field of the file-information record FILE-INFO.
     The type is denoted by a symbol.  The following types are possible:

    `regular'
          Regular file

    `directory'
          Directory

    `character-special'
          Character special device

    `block-special'
          Block special device

    `fifo'
          FIFO

    `symbolic-link'
          Symbolic link

    `socket'
          Socket

    `unknown'
          File is of an unknown type

     For example:

          > (file-info-type (file-info "/dev/tty"))
          character-special
          > (file-info-type (file-info "/dev"))
          directory


 -- procedure: file-info-device FILE-INFO
     Returns the device field of the file-information record FILE-INFO.

     For example:

          > (file-info-device (file-info "/dev/tty"))
          19513156


 -- procedure: file-info-inode FILE-INFO
     Returns the inode field of the file-information record FILE-INFO.

     For example:

          > (file-info-inode (file-info "/dev/tty"))
          20728196


 -- procedure: file-info-mode FILE-INFO
     Returns the mode field of the file-information record FILE-INFO.

     For example:

          > (file-info-mode (file-info "/dev/tty"))
          438


 -- procedure: file-info-number-of-links FILE-INFO
     Returns the number-of-links field of the file-information record
     FILE-INFO.

     For example:

          > (file-info-number-of-links (file-info "/dev/tty"))
          1


 -- procedure: file-info-owner FILE-INFO
     Returns the owner field of the file-information record FILE-INFO.

     For example:

          > (file-info-owner (file-info "/dev/tty"))
          0


 -- procedure: file-info-group FILE-INFO
     Returns the group field of the file-information record FILE-INFO.

     For example:

          > (file-info-group (file-info "/dev/tty"))
          0


 -- procedure: file-info-size FILE-INFO
     Returns the size field of the file-information record FILE-INFO.

     For example:

          > (file-info-size (file-info "/dev/tty"))
          0


 -- procedure: file-info-last-access-time FILE-INFO
     Returns the last-access-time field of the file-information record
     FILE-INFO.

     For example:

          > (file-info-last-access-time (file-info "/dev/tty"))
          #<time #2>


 -- procedure: file-info-last-modification-time FILE-INFO
     Returns the last-modification-time field of the file-information
     record FILE-INFO.

     For example:

          > (file-info-last-modification-time (file-info "/dev/tty"))
          #<time #2>


 -- procedure: file-info-last-change-time FILE-INFO
     Returns the last-change-time field of the file-information record
     FILE-INFO.

     For example:

          > (file-info-last-change-time (file-info "/dev/tty"))
          #<time #2>


 -- procedure: file-info-attributes FILE-INFO
     Returns the attributes field of the file-information record
     FILE-INFO.

     For example:

          > (file-info-attributes (file-info "/dev/tty"))
          128


 -- procedure: file-info-creation-time FILE-INFO
     Returns the creation-time field of the file-information record
     FILE-INFO.

     For example:

          > (file-info-creation-time (file-info "/dev/tty"))
          #<time #2>


 -- procedure: file-type PATH
 -- procedure: file-device PATH
 -- procedure: file-inode PATH
 -- procedure: file-mode PATH
 -- procedure: file-number-of-links PATH
 -- procedure: file-owner PATH
 -- procedure: file-group PATH
 -- procedure: file-size PATH
 -- procedure: file-last-access-time PATH
 -- procedure: file-last-modification-time PATH
 -- procedure: file-last-change-time PATH
 -- procedure: file-attributes PATH
 -- procedure: file-creation-time PATH
     These procedures combine a call to the `file-info' procedure and a
     call to a file-information record field accessor.  For instance
     `(file-type PATH)' is equivalent to `(file-info-type (file-info
     PATH))'.


 -- procedure: file-last-access-and-modification-times-set! PATH [ATIME
          [MTIME]]
     This procedure changes the last-access and last-modification times
     of the file whose location is given by the string PATH.  Time is
     specified either with a time object indicating an absolute point in
     time or a real number indicating the number of seconds relative to
     the moment the procedure is called.  When ATIME and MTIME are not
     specified, the last-access and last-modification times are set to
     the current time.  When MTIME is not specified, the last-access
     and last-modification times are set to ATIME.  Otherwise the
     last-access time is set to ATIME and the last-modification time is
     set to MTIME.

     For example:

          > (define (t path)
              (list (time->seconds (file-last-access-time path))
                    (time->seconds (file-last-modification-time path))))
          > (with-output-to-file "nl.txt" newline)
          > (t "nl.txt")
          (1429547027. 1429547027.)
          > (t "nl.txt")
          (1429547027. 1429547027.)
          > (file-last-access-and-modification-times-set! "nl.txt")
          > (t "nl.txt")
          (1429547039. 1429547039.)
          > (file-last-access-and-modification-times-set! "nl.txt" -60)
          > (t "nl.txt")
          (1429547006. 1429547006.)
          > (file-last-access-and-modification-times-set! "nl.txt" -60 0)
          > (t "nl.txt")
          (1429547049. 1429547109.)


13.9 Group information
======================

 -- procedure: group-info GROUP-NAME-OR-ID
     This procedure accesses the group database to get information
     about the group identified by GROUP-NAME-OR-ID, which is the
     group's symbolic name (string) or the group's GID (exact integer).
     A group-information record is returned that contains the group's
     symbolic name, the group's id (GID), and the group's members (list
     of symbolic user names).

     For example:

          > (group-info "staff")
          #<group-info #2 name: "staff" gid: 20 members: ("root")>
          > (group-info 29)
          #<group-info #3
             name: "certusers"
             gid: 29
             members: ("root" "jabber" "postfix" "cyrusimap")>
          > (group-info 5000)
          *** ERROR IN (console)@3.1 -- Resource temporarily unavailable
          (group-info 5000)


 -- procedure: group-info? OBJ
     This procedure returns `#t' when OBJ is a group-information record
     and `#f' otherwise.

     For example:

          > (group-info? (group-info "daemon"))
          #t
          > (group-info? 123)
          #f


 -- procedure: group-info-name GROUP-INFO
     Returns the symbolic name field of the group-information record
     GROUP-INFO.

     For example:

          > (group-info-name (group-info 29))
          "certusers"


 -- procedure: group-info-gid GROUP-INFO
     Returns the group id field of the group-information record
     GROUP-INFO.

     For example:

          > (group-info-gid (group-info "staff"))
          20


 -- procedure: group-info-members GROUP-INFO
     Returns the members field of the group-information record
     GROUP-INFO.

     For example:

          > (group-info-members (group-info "staff"))
          ("root")


13.10 User information
======================

 -- procedure: user-name
     This procedure returns the user's name as a string.

     For example:

          > (user-name)
          "feeley"


 -- procedure: user-info USER-NAME-OR-ID
     This procedure accesses the user database to get information about
     the user identified by USER-NAME-OR-ID, which is the user's
     symbolic name (string) or the user's UID (exact integer).  A
     user-information record is returned that contains the user's
     symbolic name, the user's id (UID), the user's group id (GID), the
     path to the user's home directory, and the user's login shell.

     For example:

          > (user-info "feeley")
          #<user-info #2
             name: "feeley"
             uid: 506
             gid: 506
             home: "/Users/feeley"
             shell: "/bin/bash">
          > (user-info 0)
          #<user-info #3 name: "root" uid: 0 gid: 0 home: "/var/root" shell: "/bin/sh">
          > (user-info 5000)
          *** ERROR IN (console)@3.1 -- Resource temporarily unavailable
          (user-info 5000)


 -- procedure: user-info? OBJ
     This procedure returns `#t' when OBJ is a user-information record
     and `#f' otherwise.

     For example:

          > (user-info? (user-info "feeley"))
          #t
          > (user-info? 123)
          #f


 -- procedure: user-info-name USER-INFO
     Returns the symbolic name field of the user-information record
     USER-INFO.

     For example:

          > (user-info-name (user-info 0))
          "root"


 -- procedure: user-info-uid USER-INFO
     Returns the user id field of the user-information record USER-INFO.

     For example:

          > (user-info-uid (user-info "feeley"))
          506


 -- procedure: user-info-gid USER-INFO
     Returns the group id field of the user-information record
     USER-INFO.

     For example:

          > (user-info-gid (user-info "feeley"))
          506


 -- procedure: user-info-home USER-INFO
     Returns the home directory field of the user-information record
     USER-INFO.

     For example:

          > (user-info-home (user-info 0))
          "/var/root"


 -- procedure: user-info-shell USER-INFO
     Returns the shell field of the user-information record USER-INFO.

     For example:

          > (user-info-shell (user-info 0))
          "/bin/sh"


13.11 Host information
======================

 -- procedure: host-name
     This procedure returns the machine's host name as a string.

     For example:

          > (host-name)
          "mega.iro.umontreal.ca"


 -- procedure: host-info HOST-NAME
     This procedure accesses the internet host database to get
     information about the machine whose name is denoted by the string
     HOST-NAME.  A host-information record is returned that contains
     the official name of the machine, a list of aliases (alternative
     names), and a non-empty list of IP addresses for this machine.  An
     exception is raised when HOST-NAME does not appear in the database.

     For example:

          > (host-info "www.google.com")
          #<host-info #2
             name: "www.l.google.com"
             aliases: ("www.google.com")
             addresses: (#u8(66 249 85 99) #u8(66 249 85 104))>
          > (host-info "unknown.domain")
          *** ERROR IN (console)@2.1 -- Unknown host
          (host-info "unknown.domain")


 -- procedure: host-info? OBJ
     This procedure returns `#t' when OBJ is a host-information record
     and `#f' otherwise.

     For example:

          > (host-info? (host-info "www.google.com"))
          #t
          > (host-info? 123)
          #f


 -- procedure: host-info-name HOST-INFO
     Returns the official name field of the host-information record
     HOST-INFO.

     For example:

          > (host-info-name (host-info "www.google.com"))
          "www.l.google.com"


 -- procedure: host-info-aliases HOST-INFO
     Returns the aliases field of the host-information record
     HOST-INFO.  This field is a possibly empty list of strings.

     For example:

          > (host-info-aliases (host-info "www.google.com"))
          ("www.google.com")


 -- procedure: host-info-addresses HOST-INFO
     Returns the addresses field of the host-information record
     HOST-INFO.  This field is a non-empty list of u8vectors denoting
     IP addresses.

     For example:

          > (host-info-addresses (host-info "www.google.com"))
          (#u8(66 249 85 99) #u8(66 249 85 104))


 -- procedure: address-infos [`host:' HOST] [`service:' SERVICE]
          [`family:' FAMILY] [`socket-type:' SOCKET-TYPE] [`protocol:'
          PROTOCOL]
     This procedure is an interface to the `getaddrinfo' system call.
     It accesses the internet host database to get information about the
     machine whose name is denoted by the string HOST and service is
     denoted by the string SERVICE and network address family is FAMILY
     (`INET' or `INET6') and network socket-type is SOCKET-TYPE
     (`STREAM' or `DGRAM' or `RAW') and network protocol is SOCKET-TYPE
     (`TCP' or `UDP').  A list of address-information records is
     returned.

     For example:

          > (address-infos host: "ftp.at.debian.org")
          (#<address-info #2
              family: INET6
              socket-type: DGRAM
              protocol: UDP
              socket-info:
               #<socket-info #3
                  family: INET6
                  port-number: 0
                  address: #u16(8193 2136 2 1 0 0 0 16)>>
           #<address-info #4
              family: INET6
              socket-type: STREAM
              protocol: TCP
              socket-info:
               #<socket-info #5
                  family: INET6
                  port-number: 0
                  address: #u16(8193 2136 2 1 0 0 0 16)>>
           #<address-info #6
              family: INET
              socket-type: DGRAM
              protocol: UDP
              socket-info:
               #<socket-info #7
                  family: INET
                  port-number: 0
                  address: #u8(213 129 232 18)>>
           #<address-info #8
              family: INET
              socket-type: STREAM
              protocol: TCP
              socket-info:
               #<socket-info #9
                  family: INET
                  port-number: 0
                  address: #u8(213 129 232 18)>>)
          > (address-infos host: "ftp.at.debian.org"
                           family: 'INET
                           protocol: 'TCP)
          (#<address-info #10
              family: INET
              socket-type: STREAM
              protocol: TCP
              socket-info:
               #<socket-info #11
                  family: INET
                  port-number: 0
                  address: #u8(213 129 232 18)>>)
          > (address-infos host: "unknown.domain")
          *** ERROR IN (console)@5.1 -- nodename nor servname provided, or not known
          (address-infos host: "unknown.domain")


 -- procedure: address-info? OBJ
     This procedure returns `#t' when OBJ is an address-information
     record and `#f' otherwise.

     For example:

          > (map address-info?
                 (address-infos host: "ftp.at.debian.org"))
          (#t #t #t #t)
          > (address-info? 123)
          #f


 -- procedure: address-info-family ADDRESS-INFO
     Returns the family field of the address-information record
     ADDRESS-INFO.

     For example:

          > (map address-info-family
                 (address-infos host: "ftp.at.debian.org"))
          (INET6 INET6 INET INET)


 -- procedure: address-info-socket-type ADDRESS-INFO
     Returns the socket-type field of the address-information record
     ADDRESS-INFO.

     For example:

          > (map address-info-socket-type
                 (address-infos host: "ftp.at.debian.org"))
          (DGRAM STREAM DGRAM STREAM)


 -- procedure: address-info-protocol ADDRESS-INFO
     Returns the protocol field of the address-information record
     ADDRESS-INFO.

     For example:

          > (map address-info-protocol
                 (address-infos host: "ftp.at.debian.org"))
          (UDP TCP UDP TCP)


 -- procedure: address-info-socket-info ADDRESS-INFO
     Returns the socket-info field of the address-information record
     ADDRESS-INFO.

     For example:

          > (map address-info-socket-info
                 (address-infos host: "ftp.at.debian.org"))
          (#<socket-info #2
              family: INET6
              port-number: 0
              address: #u16(8193 2136 2 1 0 0 0 16)>
           #<socket-info #3
              family: INET6
              port-number: 0
              address: #u16(8193 2136 2 1 0 0 0 16)>
           #<socket-info #4
              family: INET
              port-number: 0
              address: #u8(213 129 232 18)>
           #<socket-info #5
              family: INET
              port-number: 0
              address: #u8(213 129 232 18)>)


13.12 Service information
=========================

 -- procedure: service-info SERVICE-NAME-OR-ID
     This procedure accesses the service database to get information
     about the service identified by SERVICE-NAME-OR-ID, which is the
     service's symbolic name (string) or the service's port number
     (exact integer).  A service-information record is returned that
     contains the service's symbolic name, a list of aliases
     (alternative names), the port number (exact integer), and the
     protocol name (string).  An exception is raised when
     SERVICE-NAME-OR-ID does not appear in the database.

     For example:

          > (service-info "http")
          #<service-info #2
             name: "http"
             aliases: ("www" "www-http")
             port-number: 80
             protocol: "udp">
          > (service-info 80)
          #<service-info #3
             name: "http"
             aliases: ("www" "www-http")
             port-number: 80
             protocol: "udp">


 -- procedure: service-info? OBJ
     This procedure returns `#t' when OBJ is a service-information
     record and `#f' otherwise.

     For example:

          > (service-info? (service-info "http"))
          #t
          > (service-info? 123)
          #f


 -- procedure: service-info-name SERVICE-INFO
     Returns the symbolic name field of the service-information record
     SERVICE-INFO.

     For example:

          > (service-info-name (service-info 80))
          "http"


 -- procedure: service-info-aliases SERVICE-INFO
     Returns the aliases field of the service-information record
     SERVICE-INFO.  This field is a possibly empty list of strings.

     For example:

          > (service-info-aliases (service-info "http"))
          ("www" "www-http")


 -- procedure: service-info-port-number SERVICE-INFO
     Returns the service port number field of the service-information
     record SERVICE-INFO.

     For example:

          > (service-info-port-number (service-info "http"))
          80


 -- procedure: service-info-protocol SERVICE-INFO
     Returns the service protocol name field of the service-information
     record SERVICE-INFO.

     For example:

          > (service-info-protocol (service-info "http"))
          "udp"


13.13 Protocol information
==========================

 -- procedure: protocol-info PROTOCOL-NAME-OR-ID
     This procedure accesses the protocol database to get information
     about the protocol identified by PROTOCOL-NAME-OR-ID, which is the
     protocol's symbolic name (string) or the protocol's number (exact
     integer).  A protocol-information record is returned that contains
     the protocol's symbolic name, a list of aliases (alternative
     names), and the protocol number (32 bit unsigned exact integer).
     An exception is raised when PROTOCOL-NAME-OR-ID does not appear in
     the database.

     For example:

          > (protocol-info "tcp")
          #<protocol-info #2 name: "tcp" aliases: ("TCP") number: 6>
          > (protocol-info 6)
          #<protocol-info #2 name: "tcp" aliases: ("TCP") number: 6>


 -- procedure: protocol-info? OBJ
     This procedure returns `#t' when OBJ is a protocol-information
     record and `#f' otherwise.

     For example:

          > (protocol-info? (protocol-info "tcp"))
          #t
          > (protocol-info? 123)
          #f


 -- procedure: protocol-info-name PROTOCOL-INFO
     Returns the symbolic name field of the protocol-information record
     PROTOCOL-INFO.

     For example:

          > (protocol-info-name (protocol-info 6))
          "tcp"


 -- procedure: protocol-info-aliases PROTOCOL-INFO
     Returns the aliases field of the protocol-information record
     PROTOCOL-INFO.  This field is a possibly empty list of strings.

     For example:

          > (protocol-info-aliases (protocol-info "tcp"))
          ("TCP")


 -- procedure: protocol-info-number PROTOCOL-INFO
     Returns the protocol number field of the protocol-information
     record PROTOCOL-INFO.

     For example:

          > (protocol-info-number (protocol-info "tcp"))
          6


13.14 Network information
=========================

 -- procedure: network-info NETWORK-NAME-OR-ID
     This procedure accesses the network database to get information
     about the network identified by NETWORK-NAME-OR-ID, which is the
     network's symbolic name (string) or the network's number (exact
     integer).  A network-information record is returned that contains
     the network's symbolic name, a list of aliases (alternative
     names), and the network number (32 bit unsigned exact integer).
     An exception is raised when NETWORK-NAME-OR-ID does not appear in
     the database.

     For example:

          > (network-info "loopback")
          #<network-info #2
             name: "loopback"
             aliases: ("loopback-net")
             number: 127>
          > (network-info 127)
          #<network-info #3
             name: "loopback"
             aliases: ("loopback-net")
             number: 127>


 -- procedure: network-info? OBJ
     This procedure returns `#t' when OBJ is a network-information
     record and `#f' otherwise.

     For example:

          > (network-info? (network-info "loopback"))
          #t
          > (network-info? 123)
          #f


 -- procedure: network-info-name NETWORK-INFO
     Returns the symbolic name field of the network-information record
     NETWORK-INFO.

     For example:

          > (network-info-name (network-info 127))
          "loopback"


 -- procedure: network-info-aliases NETWORK-INFO
     Returns the aliases field of the network-information record
     NETWORK-INFO.  This field is a possibly empty list of strings.

     For example:

          > (network-info-aliases (network-info "loopback"))
          ("loopback-net")


 -- procedure: network-info-number NETWORK-INFO
     Returns the network number field of the network-information record
     NETWORK-INFO.

     For example:

          > (network-info-number (network-info "loopback"))
          127


14 I/O and ports
****************

14.1 Unidirectional and bidirectional ports
===========================================

Unidirectional ports allow communication between a producer of
information and a consumer.  An input-port's producer is typically a
resource managed by the operating system (such as a file, a process or
a network connection) and the consumer is the Scheme program.  The
roles are reversed for an output-port.

   Associated with each port are settings that affect I/O operations on
that port (encoding of characters to bytes, end-of-line encoding, type
of buffering, etc).  Port settings are specified when the port is
created.  Some port settings can be changed after a port is created.

   Bidirectional ports, also called input-output-ports, allow
communication in both directions.  They are best viewed as an object
that groups two separate unidirectional ports (one in each direction).
Each direction has its own port settings and can be closed
independently from the other direction.

14.2 Port classes
=================

The four classes of ports listed below form an inheritance hierarchy.
Operations possible for a certain class of port are also possible for
the subclasses.  Only device-ports are connected to a device managed by
the operating system.  For instance it is possible to create ports that
behave as a FIFO where the Scheme program is both the producer and
consumer of information (possibly one thread is the producer and
another thread is the consumer).

  1. An "object-port" (or simply a port) provides operations to read
     and write Scheme data (i.e. any Scheme object) to/from the port.
     It also provides operations to force output to occur, to change
     the way threads block on the port, and to close the port.  Note
     that the class of objects for which write/read invariance is
     guaranteed depends on the particular class of port.

  2. A "character-port" provides all the operations of an object-port,
     and also operations to read and write individual characters to/from
     the port.  When a Scheme object is written to a character-port, it
     is converted into the sequence of characters that corresponds to
     its external-representation.  When reading a Scheme object, an
     inverse conversion occurs.  Note that some Scheme objects do not
     have an external textual representation that can be read back.

  3. A "byte-port" provides all the operations of a character-port, and
     also operations to read and write individual bytes to/from the
     port.  When a character is written to a byte-port, some encoding
     of that character into a sequence of bytes will occur (for example,
     `#\newline' will be encoded as the 2 bytes CR-LF when using
     ISO-8859-1 character encoding and `cr-lf' end-of-line encoding, and
     a non-ASCII character will generate more than 1 byte when using
     UTF-8 character encoding).  When reading a character, a similar
     decoding occurs.

  4. A "device-port" provides all the operations of a byte-port, and
     also operations to control the operating system managed device
     (file, network connection, terminal, etc) that is connected to the
     port.


14.3 Port settings
==================

Some port settings are only valid for specific port classes whereas
some others are valid for all ports.  Port settings are specified when
a port is created.  The settings that are not specified will default to
some reasonable values.  Keyword objects are used to name the settings
to be set.  As a simple example, a device-port connected to the file
`"foo"' can be created using the call

     (open-input-file "foo")

   This will use default settings for the character encoding, buffering,
etc.  When a specific character encoding is desired, such as UTF-16BE,
the port can be opened using the call

     (open-input-file (list path: "foo" char-encoding: 'UTF-16BE))

   Here the argument of the procedure `open-input-file' has been
replaced by a "port settings list" which specifies the value of each
port setting that should not be set to the default value.  Note that
some port settings have no useful default and it is therefore required
to specify a value for them, such as the `path:' in the case of the
file opening procedures.  All port creation procedures (i.e. named
`open-...') take a single argument that can either be a port settings
list or a value of a type that depends on the kind of port being
created (a path string for files, an IP port number for socket servers,
etc).

14.4 Object-ports
=================

14.4.1 Object-port settings
---------------------------

The following is a list of port settings that are valid for all types
of ports.

   * `direction:' ( `input' | `output' | `input-output' )

     This setting controls the direction of the port.  The symbol
     `input' indicates a unidirectional input-port, the symbol `output'
     indicates a unidirectional output-port, and the symbol
     `input-output' indicates a bidirectional port.  The default value
     of this setting depends on the port creation procedure.

   * `buffering:' ( `#f' | `#t' | `line' )

     This setting controls the buffering of the port.  To set each
     direction separately the keywords `input-buffering:' and
     `output-buffering:' must be used instead of `buffering:'.  The
     value `#f' selects unbuffered I/O, the value `#t' selects fully
     buffered I/O, and the symbol `line' selects line buffered I/O (the
     output buffer is drained when a `#\newline' character is written).
     Line buffered I/O only applies to character-ports.  The default
     value of this setting is operating system dependent except
     consoles which are unbuffered.


14.4.2 Object-port operations
-----------------------------

 -- procedure: input-port? OBJ
 -- procedure: output-port? OBJ
 -- procedure: port? OBJ
     The procedure `input-port?' returns `#t' when OBJ is a
     unidirectional input-port or a bidirectional port and `#f'
     otherwise.

     The procedure `output-port?' returns `#t' when OBJ is a
     unidirectional output-port or a bidirectional port and `#f'
     otherwise.

     The procedure `port?' returns `#t' when OBJ is a port (either
     unidirectional or bidirectional) and `#f' otherwise.

     For example:

          > (input-port? (current-input-port))
          #t
          > (call-with-input-string "some text" output-port?)
          #f
          > (port? (current-output-port))
          #t


 -- procedure: read [PORT]
     This procedure reads and returns the next Scheme datum from the
     input-port PORT.  The end-of-file object is returned when the end
     of the stream is reached.  If it is not specified, PORT defaults
     to the current input-port.

     For example:

          > (call-with-input-string "some text" read)
          some
          > (call-with-input-string "" read)
          #!eof


 -- procedure: read-all [PORT [READER]]
     This procedure repeatedly calls the procedure READER with PORT as
     the sole argument and accumulates a list of each value returned up
     to the end-of-file object.  The procedure `read-all' returns the
     accumulated list without the end-of-file object.  If it is not
     specified, PORT defaults to the current input-port.  If it is not
     specified, READER defaults to the procedure `read'.

     For example:

          > (call-with-input-string "3,2,1\ngo!" read-all)
          (3 ,2 ,1 go!)
          > (call-with-input-string "3,2,1\ngo!"
                                    (lambda (p) (read-all p read-char)))
          (#\3 #\, #\2 #\, #\1 #\newline #\g #\o #\!)
          > (call-with-input-string "3,2,1\ngo!"
                                    (lambda (p) (read-all p read-line)))
          ("3,2,1" "go!")


 -- procedure: write OBJ [PORT]
     This procedure writes the Scheme datum OBJ to the output-port PORT
     and the value returned is unspecified.  If it is not specified,
     PORT defaults to the current output-port.

     For example:

          > (write (list 'compare (list 'quote '@x) 'and (list 'unquote '@x)))
          (compare '@x and , @x)>


 -- procedure: newline [PORT]
     This procedure writes an "object separator" to the output-port
     PORT and the value returned is unspecified.  The separator ensures
     that the next Scheme datum written with the `write' procedure will
     not be confused with the latest datum that was written.  On
     character-ports this is done by writing the character `#\newline'.
     On ports where successive objects are implicitly distinct (such
     as "vector ports") this procedure does nothing.

     Regardless of the class of a port P and assuming that the external
     textual representation of the object X is readable, the expression
     `(begin (write X P) (newline P))' will write to P a representation
     of X that can be read back with the procedure `read'.  If it is
     not specified, PORT defaults to the current output-port.

     For example:

          > (begin (write 123) (newline) (write 456) (newline))
          123
          456


 -- procedure: force-output [PORT [LEVEL]]
     The procedure `force-output' causes the data that was written to
     the output-port PORT to be moved closer to its destination
     according to LEVEL, an exact integer in the range 0 to 2.  If PORT
     is not specified, the current output-port is used.  If LEVEL is
     not specified, it defaults to 0.  Values of LEVEL above 0 are
     equivalent to LEVEL = 0 except for device ports as explained below.

     When LEVEL is 0, the output buffers of PORT which are managed in
     the Scheme process are drained (i.e.  the output operation that
     was delayed due to buffering is actually performed).  In the case
     of a device port the data is passed to the operating system and it
     becomes its responsibility to transmit the data to the device.  The
     operating system may implement its own buffering approach which
     delays the transmission of the data to the device.

     When LEVEL is 1, in addition to the operations for LEVEL = 0 and
     if the operating system supports the functionality, the operating
     system is requested to transmit the data to the device.  On UNIX
     this corresponds to a `fsync' system call.

     When LEVEL is 2, in addition to the operations for LEVEL = 1 and
     if the operating system supports the functionality, the operating
     system is requested to wait until the device reports that the data
     was saved by the device (e.g. actually written to disk in the case
     of a file).  This operation can take a long time on some operating
     systems.  On macOS this corresponds to a `fcntl' system call with
     operation `F_FULLFSYNC'.

     For example:

          > (define p (open-tcp-client "www.iro.umontreal.ca:80"))
          > (display "GET /\n" p)
          > (force-output p)
          > (read-line p)
          "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01 Transitional//EN\""


 -- procedure: close-input-port PORT
 -- procedure: close-output-port PORT
 -- procedure: close-port PORT
     The PORT argument of these procedures must be a unidirectional or
     a bidirectional port.  For all three procedures the value returned
     is unspecified.

     The procedure `close-input-port' closes the input-port side of
     PORT, which must not be a unidirectional output-port.

     The procedure `close-output-port' closes the output-port side of
     PORT, which must not be a unidirectional input-port.  The ouput
     buffers are drained before PORT is closed.

     The procedure `close-port' closes all sides of the PORT.  Unless
     PORT is a unidirectional input-port, the output buffers are
     drained before PORT is closed.

     For example:

          > (define p (open-tcp-client "www.iro.umontreal.ca:80"))
          > (display "GET /\n" p)
          > (close-output-port p)
          > (read-line p)
          "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01 Transitional//EN\""


 -- procedure: input-port-timeout-set! PORT TIMEOUT [THUNK]
 -- procedure: output-port-timeout-set! PORT TIMEOUT [THUNK]
     When a thread tries to perform an I/O operation on a port, the
     requested operation may not be immediately possible and the thread
     must wait.  For example, the thread may be trying to read a line of
     text from the console and the user has not typed anything yet, or
     the thread may be trying to write to a network connection faster
     than the network can handle.  In such situations the thread
     normally blocks until the operation becomes possible.

     It is sometimes necessary to guarantee that the thread will not
     block too long.  For this purpose, to each input-port and
     output-port is attached a "timeout" and "timeout-thunk".  The
     timeout indicates the point in time beyond which the thread should
     stop waiting on an input and output operation respectively.  When
     the timeout is reached, the thread calls the port's timeout-thunk.
     If the timeout-thunk returns `#f' the thread abandons trying to
     perform the operation (in the case of an input operation an
     end-of-file is read and in the case of an output operation an
     exception is raised).  Otherwise, the thread will block again
     waiting for the operation to become possible (note that if the
     port's timeout has not changed the thread will immediately call
     the timeout-thunk again).

     The procedure `input-port-timeout-set!' sets the timeout of the
     input-port PORT to TIMEOUT and the timeout-thunk to THUNK.  The
     procedure `output-port-timeout-set!' sets the timeout of the
     output-port PORT to TIMEOUT and the timeout-thunk to THUNK.  If it
     is not specified, the THUNK defaults to a thunk that returns `#f'.
     The TIMEOUT is either a time object indicating an absolute point
     in time, or it is a real number which indicates the number of
     seconds relative to the moment the procedure is called.  For both
     procedures the value returned is unspecified.

     When a port is created the timeout is set to infinity (`+inf.0').
     This causes the thread to wait as long as needed for the operation
     to become possible.  Setting the timeout to a point in the past
     (`-inf.0') will cause the thread to attempt the I/O operation and
     never block (i.e. the timeout-thunk is called if the operation is
     not immediately possible).

     The following example shows how to cause the REPL to terminate
     when the user does not enter an expression within the next 60
     seconds.

          > (input-port-timeout-set! (repl-input-port) 60)
          >
          *** EOF again to exit


14.5 Character-ports
====================

14.5.1 Character-port settings
------------------------------

The following is a list of port settings that are valid for
character-ports.

   * `readtable:' READTABLE

     This setting determines the readtable attached to the
     character-port.  To set each direction separately the keywords
     `input-readtable:' and `output-readtable:' must be used instead of
     `readtable:'.  Readtables control the external textual
     representation of Scheme objects, that is the encoding of Scheme
     objects using characters.  The behavior of the `read' procedure
     depends on the port's input-readtable and the behavior of the
     procedures `write', `pretty-print', and related procedures is
     affected by the port's output-readtable.  The default value of this
     setting is the value bound to the parameter object
     `current-readtable'.

   * `output-width:' POSITIVE-INTEGER

     This setting indicates the width of the character output-port in
     number of characters.  This information is used by the
     pretty-printer.  The default value of this setting is 80.


14.5.2 Character-port operations
--------------------------------

 -- procedure: input-port-line PORT
 -- procedure: input-port-column PORT
 -- procedure: output-port-line PORT
 -- procedure: output-port-column PORT
     The current character location of a character input-port is the
     location of the next character to read.  The current character
     location of a character output-port is the location of the next
     character to write.  Location is denoted by a line number (the
     first line is line 1) and a column number, that is the location on
     the current line (the first column is column 1).  The procedures
     `input-port-line' and `input-port-column' return the line location
     and the column location respectively of the character input-port
     PORT.  The procedures `output-port-line' and `output-port-column'
     return the line location and the column location respectively of
     the character output-port PORT.

     For example:

          > (call-with-output-string
              (lambda (p)
                (display "abc\n123def" p)
                (write (list (output-port-line p) (output-port-column p))
                       p)))
          "abc\n123def(2 7)"


 -- procedure: output-port-width PORT
     This procedure returns the width, in characters, of the character
     output-port PORT.  The value returned is the port's output-width
     setting.

     For example:

          > (output-port-width (repl-output-port))
          80


 -- procedure: read-char [PORT]
     This procedure reads the character input-port PORT and returns the
     character at the current character location and advances the
     current character location to the next character, unless the PORT
     is already at end-of-file in which case `read-char' returns the
     end-of-file object.  If it is not specified, PORT defaults to the
     current input-port.

     For example:

          > (call-with-input-string
              "some text"
              (lambda (p)
                (let ((a (read-char p))) (list a (read-char p)))))
          (#\s #\o)
          > (call-with-input-string "" read-char)
          #!eof


 -- procedure: peek-char [PORT]
     This procedure returns the same result as `read-char' but it does
     not advance the current character location of the input-port PORT.
     If it is not specified, PORT defaults to the current input-port.

     For example:

          > (call-with-input-string
              "some text"
              (lambda (p)
                (let ((a (peek-char p))) (list a (read-char p)))))
          (#\s #\s)
          > (call-with-input-string "" peek-char)
          #!eof


 -- procedure: write-char CHAR [PORT]
     This procedure writes the character CHAR to the character
     output-port PORT and advances the current character location of
     that output-port.  The value returned is unspecified.  If it is not
     specified, PORT defaults to the current output-port.

     For example:

          > (write-char #\=)
          =>


 -- procedure: read-line [PORT [SEPARATOR [INCLUDE-SEPARATOR?
          [MAX-LENGTH]]]]
     This procedure reads characters from the character input-port PORT
     until a specific SEPARATOR or the end-of-file is encountered and
     returns a string containing the sequence of characters read.  If
     it is specified, MAX-LENGTH must be a nonnegative exact integer
     and it places an upper limit on the number of characters that are
     read.

     The SEPARATOR is included at the end of the string only if it was
     the last character read and INCLUDE-SEPARATOR? is not `#f'.  The
     SEPARATOR must be a character or `#f' (in which case all the
     characters until the end-of-file are read).  If it is not
     specified, PORT defaults to the current input-port.  If it is not
     specified, SEPARATOR defaults to `#\newline'.  If it is not
     specified, INCLUDE-SEPARATOR? defaults to `#f'.

     For example:

          > (define (split sep)
              (lambda (str)
                (call-with-input-string
                  str
                  (lambda (p)
                    (read-all p (lambda (p) (read-line p sep)))))))
          > ((split #\,) "a,b,c")
          ("a" "b" "c")
          > (map (split #\,)
                 (call-with-input-string "1,2,3\n4,5"
                                         (lambda (p) (read-all p read-line))))
          (("1" "2" "3") ("4" "5"))
          > (read-line (current-input-port) #\newline #f 2)1234
          "12"
          > 34


 -- procedure: read-substring STRING START END [PORT [NEED]]
 -- procedure: write-substring STRING START END [PORT]
     These procedures support bulk character I/O.  The part of the
     string STRING starting at index START and ending just before index
     END is used as a character buffer that will be the target of
     `read-substring' or the source of the `write-substring'.  The
     `read-substring' also accepts a NEED parameter which must be a
     nonnegative fixnum.  Up to END-START characters will be
     transferred.  The number of characters transferred, possibly zero,
     is returned by these procedures.  Fewer characters will be read by
     `read-substring' if an end-of-file is read, or a timeout occurs
     before all the requested characters are transferred and the
     timeout thunk returns `#f' (see the procedure
     `input-port-timeout-set!'), or NEED is specified and at least that
     many characters have been read (in other words the procedure does
     not block for more characters but may transfer more characters if
     they are immediately available).  Fewer characters will be written
     by `write-substring' if a timeout occurs before all the requested
     characters are transferred and the timeout thunk returns `#f' (see
     the procedure `output-port-timeout-set!').  If it is not
     specified, PORT defaults to the current input-port and current
     output-port respectively.

     For example:

          > (define s (make-string 10 #\x))
          > (read-substring s 2 5)123456789
          3
          > 456789
          > s
          "xx123xxxxx"
          > (read-substring s 2 10 (current-input-port) 3)abcd
          5
          > s
          "xxabcd\nxxx"


 -- procedure: input-port-readtable PORT
 -- procedure: output-port-readtable PORT
     These procedures return the readtable attached to the
     character-port PORT.  The PORT parameter of `input-port-readtable'
     must be an input-port.  The PORT parameter of
     `output-port-readtable' must be an output-port.


 -- procedure: input-port-readtable-set! PORT READTABLE
 -- procedure: output-port-readtable-set! PORT READTABLE
     These procedures change the readtable attached to the
     character-port PORT to the readtable READTABLE.  The PORT parameter
     of `input-port-readtable-set!'  must be an input-port.  The PORT
     parameter of `output-port-readtable-set!' must be an output-port.
     The value returned is unspecified.


14.6 Byte-ports
===============

14.6.1 Byte-port settings
-------------------------

The following is a list of port settings that are valid for byte-ports.

   * `char-encoding:' ENCODING

     This setting controls the character encoding of the byte-port.  For
     bidirectional byte-ports, the character encoding for input and
     output is set.  To set each direction separately the keywords
     `input-char-encoding:' and `output-char-encoding:' must be used
     instead of `char-encoding:'.  The default value of this setting
     depends on how the runtime system was configured but typically
     UTF-8 is used.  The default can be overridden through various
     runtime options (*note Runtime options::), such as
     `-:file-settings=...' and `-:io-settings=...'.  The following
     encodings are supported:

    `ISO-8859-1'
          ISO-8859-1 character encoding.  Each character is encoded by
          a single byte.  Only Unicode characters with a code in the
          range 0 to 255 are allowed.

    `ASCII'
          ASCII character encoding.  Each character is encoded by a
          single byte.  In principle only Unicode characters with a
          code in the range 0 to 127 are allowed but most types of
          ports treat this exactly like `ISO-8859-1'.

    `UTF-8'
          UTF-8 character encoding.  Each character is encoded by a
          sequence of one to four bytes.  The minimum length UTF-8
          encoding is used.  If a BOM is needed at the beginning of the
          stream then it must be explicitly written.

    `UTF-16'
          UTF-16 character encoding.  Each character is encoded by one
          or two 16 bit integers (2 or 4 bytes).  The 16 bit integers
          may be encoded using little-endian encoding or big-endian
          encoding.  If the port is an input-port and the first two
          bytes read are a BOM ("Byte Order Mark" character with
          hexadecimal code FEFF) then the BOM will be discarded and the
          endianness will be set accordingly, otherwise the endianness
          depends on the operating system and how the Gambit runtime was
          compiled.  If the port is an output-port then a BOM will be
          output at the beginning of the stream and the endianness
          depends on the operating system and how the Gambit runtime
          was compiled.

    `UTF-16LE'
          UTF-16 character encoding with little-endian endianness.  It
          is like `UTF-16' except the endianness is set to
          little-endian and there is no BOM processing.  If a BOM is
          needed at the beginning of the stream then it must be
          explicitly written.

    `UTF-16BE'
          UTF-16 character encoding with big-endian endianness.  It is
          like `UTF-16LE' except the endianness is set to big-endian.

    `UTF / UTF-fallback-ASCII / UTF-fallback-ISO-8859-1 / UTF-fallback-UTF-16 / UTF-fallback-UTF-16LE / UTF-fallback-UTF-16BE'
          These encodings combine the UTF-8 and UTF-16 encodings.  When
          one of these character encodings is used for an output port,
          characters will be encoded using the UTF-8 encoding.  The
          first character, if there is one, is prefixed with a UTF-8
          BOM (the three byte sequence EF BB BF in hexadecimal).  When
          one of these character encodings is used for an input port,
          the character encoding depends on the first few bytes.  If
          the first bytes of the stream are a UTF-16LE BOM (FF FE in
          hexadecimal), or a UTF-16BE BOM (FE FF in hexadecimal), or a
          UTF-8 BOM (EF BB BF in hexadecimal), then the BOM is
          discarded and the remaining bytes of the stream are decoded
          using the corresponding character encoding.  If a BOM is not
          present, then the stream is decoded using the fallback
          encoding specified.  The encoding `UTF' is a synonym for
          `UTF-fallback-UTF-8'.  Note that the `UTF' character encoding
          for input will correctly handle streams produced using the
          encodings `UTF', `UTF-8', `UTF-16', `ASCII', and if an
          explicit BOM is output, the encodings `UTF-16LE', and
          `UTF-16BE'.

    `UCS-2'
          UCS-2 character encoding.  Each character is encoded by a 16
          bit integer (2 bytes).  The 16 bit integers may be encoded
          using little-endian encoding or big-endian encoding.  If the
          port is an input-port and the first two bytes read are a BOM
          ("Byte Order Mark" character with hexadecimal code FEFF) then
          the BOM will be discarded and the endianness will be set
          accordingly, otherwise the endianness depends on the
          operating system and how the Gambit runtime was compiled.  If
          the port is an output-port then a BOM will be output at the
          beginning of the stream and the endianness depends on the
          operating system and how the Gambit runtime was compiled.

    `UCS-2LE'
          UCS-2 character encoding with little-endian endianness.  It
          is like `UCS-2' except the endianness is set to little-endian
          and there is no BOM processing.  If a BOM is needed at the
          beginning of the stream then it must be explicitly written.

    `UCS-2BE'
          UCS-2 character encoding with big-endian endianness.  It is
          like `UCS-2LE' except the endianness is set to big-endian.

    `UCS-4'
          UCS-4 character encoding.  Each character is encoded by a 32
          integer (4 bytes).  The 32 bit integers may be encoded using
          little-endian encoding or big-endian encoding.  If the port
          is an input-port and the first four bytes read are a BOM
          ("Byte Order Mark" character with hexadecimal code 0000FEFF)
          then the BOM will be discarded and the endianness will be set
          accordingly, otherwise the endianness depends on the
          operating system and how the Gambit runtime was compiled.  If
          the port is an output-port then a BOM will be output at the
          beginning of the stream and the endianness depends on the
          operating system and how the Gambit runtime was compiled.

    `UCS-4LE'
          UCS-4 character encoding with little-endian endianness.  It
          is like `UCS-4' except the endianness is set to little-endian
          and there is no BOM processing.  If a BOM is needed at the
          beginning of the stream then it must be explicitly written.

    `UCS-4BE'
          UCS-4 character encoding with big-endian endianness.  It is
          like `UCS-4LE' except the endianness is set to big-endian.


   * `char-encoding-errors:' ( `#f' | `#t' )

     This setting controls whether illegal character encodings are
     silently replaced with the Unicode character #xfffd (replacement
     character) or raise an error.  To set each direction separately
     the keywords `input-char-encoding-errors:' and
     `output-char-encoding-errors:' must be used instead of
     `char-encoding-errors:'.  The default value of this setting is
     `#t'.

   * `eol-encoding:' ENCODING

     This setting controls the end-of-line encoding of the byte-port.
     To set each direction separately the keywords `input-eol-encoding:'
     and `output-eol-encoding:' must be used instead of
     `eol-encoding:'.  The default value of this setting is operating
     system dependent, but this can be overridden through the runtime
     options (*note Runtime options::).  Note that for output-ports the
     end-of-line encoding is applied before the character encoding, and
     for input-ports it is applied after.  The following encodings are
     supported:

    `lf'
          For an output-port, writing a `#\newline' character outputs a
          `#\linefeed' character to the stream (Unicode character code
          10).  For an input-port, a `#\newline' character is read when
          a `#\linefeed' character is encountered on the stream.  Note
          that `#\linefeed' and `#\newline' are two names for the same
          character, so this end-of-line encoding is actually the
          identity function.  Text files created by UNIX applications
          typically use this end-of-line encoding.

    `cr'
          For an output-port, writing a `#\newline' character outputs a
          `#\return' character to the stream (Unicode character code
          13).  For an input-port, a `#\newline' character is read when
          a `#\linefeed' character or a `#\return' character is
          encountered on the stream.  Text files created by Classic Mac
          OS applications typically use this end-of-line encoding.

    `cr-lf'
          For an output-port, writing a `#\newline' character outputs to
          the stream a `#\return' character followed by a `#\linefeed'
          character.  For an input-port, a `#\newline' character is read
          when a `#\linefeed' character or a `#\return' character is
          encountered on the stream.  Moreover, if this character is
          immediately followed by the opposite character (`#\linefeed'
          followed by `#\return' or `#\return' followed by
          `#\linefeed') then the second character is ignored.  In other
          words, all four possible end-of-line encodings are read as a
          single `#\newline' character.  Text files created by DOS and
          Microsoft Windows applications typically use this end-of-line
          encoding.



14.6.2 Byte-port operations
---------------------------

 -- procedure: read-u8 [PORT]
 -- procedure: peek-u8 [PORT]
     These procedures read the byte input-port PORT and return the byte
     at the current byte location unless the PORT is already at
     end-of-file in which case the end-of-file object is returned.  If
     the end-of-file is not reached then the procedure `read-u8'
     advances the current byte location to the next byte.  The procedure
     `peek-u8' does not advance the port's current byte location.  If
     it is not specified, PORT defaults to the current input-port.

     One way to ensure that the port's input character buffer is empty
     is to call `peek-u8' strictly before any use of the port in a
     character input operation (i.e. a call to the procedures `read',
     `read-char', `peek-char', etc).  Alternatively
     `input-port-characters-buffered' can be used to get the number of
     characters in the port's input character buffer, and to empty the
     buffer with calls to `read-char' or `read-substring'.

     For example:

          > (call-with-input-u8vector
              '#u8(11 22 33 44)
              (lambda (p)
                (let ((a (read-u8 p))) (list a (read-u8 p)))))
          (11 22)
          > (call-with-input-u8vector '#u8() read-u8)
          #!eof
          > (with-input-from-u8vector '#u8(1 5) (lambda () (+ (peek-u8) (peek-u8))))
          2


 -- procedure: write-u8 N [PORT]
     This procedure writes the byte N to the byte output-port PORT and
     advances the current byte location of that output-port.  The value
     returned is unspecified.  If it is not specified, PORT defaults to
     the current output-port.

     For example:

          > (call-with-output-u8vector (lambda (p) (write-u8 33 p)))
          #u8(33)


 -- procedure: read-subu8vector U8VECTOR START END [PORT [NEED]]
 -- procedure: write-subu8vector U8VECTOR START END [PORT]
     These procedures support bulk byte I/O.  The part of the u8vector
     U8VECTOR starting at index START and ending just before index END
     is used as a byte buffer that will be the target of
     `read-subu8vector' or the source of the `write-subu8vector'.  The
     `read-subu8vector' also accepts a NEED parameter which must be a
     nonnegative fixnum.  Up to END-START bytes will be transferred.
     The number of bytes transferred, possibly zero, is returned by
     these procedures.  Fewer bytes will be read by `read-subu8vector'
     if an end-of-file is read, or a timeout occurs before all the
     requested bytes are transferred and the timeout thunk returns `#f'
     (see the procedure `input-port-timeout-set!'), or NEED is
     specified and at least that many bytes have been read (in other
     words the procedure does not block for more bytes but may transfer
     more bytes if they are immediately available).  Fewer bytes will
     be written by `write-subu8vector' if a timeout occurs before all
     the requested bytes are transferred and the timeout thunk returns
     `#f' (see the procedure `output-port-timeout-set!').  If it is not
     specified, PORT defaults to the current input-port and current
     output-port respectively.

     The procedure `read-subu8vector' must be called before any use of
     the port in a character input operation (i.e. a call to the
     procedures `read', `read-char', `peek-char', etc) because
     otherwise the character-stream and byte-stream may be out of sync
     due to the port buffering.

     For example:

          > (define v (make-u8vector 10))
          > (read-subu8vector v 2 5)123456789
          3
          > 456789
          > v
          #u8(0 0 49 50 51 0 0 0 0 0)
          > (read-subu8vector v 2 10 (current-input-port) 3)abcd
          5
          > v
          #u8(0 0 97 98 99 100 10 0 0 0)


14.7 Device-ports
=================

14.7.1 Filesystem devices
-------------------------

 -- procedure: open-file PATH-OR-SETTINGS
 -- procedure: open-input-file PATH-OR-SETTINGS
 -- procedure: open-output-file PATH-OR-SETTINGS
 -- procedure: call-with-input-file PATH-OR-SETTINGS PROC
 -- procedure: call-with-output-file PATH-OR-SETTINGS PROC
 -- procedure: with-input-from-file PATH-OR-SETTINGS THUNK
 -- procedure: with-output-to-file PATH-OR-SETTINGS THUNK
     All of these procedures create a port to interface to a byte-stream
     device (such as a file, console, serial port, named pipe, etc)
     whose name is given by a path of the filesystem.  The `direction:'
     setting will default to the value `input' for the procedures
     `open-input-file', `call-with-input-file' and
     `with-input-from-file', to the value `output' for the procedures
     `open-output-file', `call-with-output-file' and
     `with-output-to-file', and to the value `input-output' for the
     procedure `open-file'.

     The procedures `open-file', `open-input-file' and
     `open-output-file' return the port that is created.  The
     procedures `call-with-input-file' and `call-with-output-file' call
     the procedure PROC with the port as single argument, and then
     return the value(s) of this call after closing the port.  The
     procedures `with-input-from-file' and `with-output-to-file'
     dynamically bind the current input-port and current output-port
     respectively to the port created for the duration of a call to the
     procedure THUNK with no argument.  The value(s) of the call to
     THUNK are returned after closing the port.

     The first argument of these procedures is either a string denoting
     a filesystem path or a list of port settings which must contain a
     `path:' setting.  Here are the settings allowed in addition to the
     generic settings of byte-ports:

        * `path:' STRING

          This setting indicates the location of the file in the
          filesystem.  There is no default value for this setting.

        * `append:' ( `#f' | `#t' )

          This setting controls whether output will be added to the end
          of the file.  This is useful for writing to log files that
          might be open by more than one process.  The default value of
          this setting is `#f'.

        * `create:' ( `#f' | `#t' | `maybe' )

          This setting controls whether the file will be created when
          it is opened.  A setting of `#f' requires that the file exist
          (otherwise an exception is raised).  A setting of `#t'
          requires that the file does not exist (otherwise an exception
          is raised).  A setting of `maybe' will create the file if it
          does not exist.  The default value of this setting is `maybe'
          for output-ports and `#f' for input-ports and bidirectional
          ports.

        * `permissions:' 12-BIT-EXACT-INTEGER

          This setting controls the UNIX permissions that will be
          attached to the file if it is created.  The default value of
          this setting is `#o666'.

        * `truncate:' ( `#f' | `#t' )

          This setting controls whether the file will be truncated when
          it is opened.  For input-ports and bidirectional ports, the
          default value of this setting is `#f'.  For output-ports, the
          default value of this setting is `#t' when the `append:'
          setting is `#f', and `#f' otherwise.


     For example:

          > (with-output-to-file
              (list path: "nofile"
                    create: #f)
              (lambda ()
                (display "hello world!\n")))
          *** ERROR IN (console)@1.1 -- No such file or directory
          (with-output-to-file '(path: "nofile" create: #f) '#<procedure #2>)


 -- procedure: input-port-byte-position PORT [POSITION [WHENCE]]
 -- procedure: output-port-byte-position PORT [POSITION [WHENCE]]
     When called with a single argument these procedures return the byte
     position where the next I/O operation would take place in the file
     attached to the given PORT (relative to the beginning of the
     file).  When called with two or three arguments, the byte position
     for subsequent I/O operations on the given PORT is changed to
     POSITION, which must be an exact integer.  When WHENCE is omitted
     or is 0, the POSITION is relative to the beginning of the file.
     When WHENCE is 1, the POSITION is relative to the current byte
     position of the file.  When WHENCE is 2, the POSITION is relative
     to the end of the file.  The return value is the new byte
     position.  On most operating systems the byte position for reading
     and writing of a given bidirectional port are the same.

     When `input-port-byte-position' is called to change the byte
     position of an input-port, all input buffers will be flushed so
     that the next byte read will be the one at the given position.

     When `output-port-byte-position' is called to change the byte
     position of an output-port, there is an implicit call to
     `force-output' before the position is changed.

     For example:

          > (define p  ; p is an input-output-port
              (open-file '(path: "test" char-encoding: ISO-8859-1 create: maybe)))
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (0 0)
          > (display "abcdefghij\n" p)
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (0 0)
          > (force-output p)
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (11 11)
          > (input-port-byte-position p 2)
          2
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (2 2)
          > (peek-char p)
          #\c
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (11 11)
          > (output-port-byte-position p -7 2)
          4
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (4 4)
          > (write-char #\! p)
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (4 4)
          > (force-output p)
          > (list (input-port-byte-position p) (output-port-byte-position p))
          (5 5)
          > (input-port-byte-position p 1)
          1
          > (read p)
          bcd!fghij


14.7.2 Process devices
----------------------

 -- procedure: open-process PATH-OR-SETTINGS
 -- procedure: open-input-process PATH-OR-SETTINGS
 -- procedure: open-output-process PATH-OR-SETTINGS
 -- procedure: call-with-input-process PATH-OR-SETTINGS PROC
 -- procedure: call-with-output-process PATH-OR-SETTINGS PROC
 -- procedure: with-input-from-process PATH-OR-SETTINGS THUNK
 -- procedure: with-output-to-process PATH-OR-SETTINGS THUNK
     All of these procedures start a new operating system process and
     create a bidirectional port which allows communication with that
     process on its standard input and standard output.  The
     `direction:' setting will default to the value `input' for the
     procedures `open-input-process', `call-with-input-process' and
     `with-input-from-process', to the value `output' for the
     procedures `open-output-process', `call-with-output-process' and
     `with-output-to-process', and to the value `input-output' for the
     procedure `open-process'.  If the `direction:' setting is `input',
     the output-port side is closed.  If the `direction:' setting is
     `output', the input-port side is closed.

     The procedures `open-process', `open-input-process' and
     `open-output-process' return the port that is created.  The
     procedures `call-with-input-process' and
     `call-with-output-process' call the procedure PROC with the port
     as single argument, and then return the value(s) of this call
     after closing the port and waiting for the process to terminate.
     The procedures `with-input-from-process' and
     `with-output-to-process' dynamically bind the current input-port
     and current output-port respectively to the port created for the
     duration of a call to the procedure THUNK with no argument.  The
     value(s) of the call to THUNK are returned after closing the port
     and waiting for the process to terminate.

     The first argument of this procedure is either a string denoting a
     filesystem path of an executable program or a list of port settings
     which must contain a `path:' setting.  Here are the settings
     allowed in addition to the generic settings of byte-ports:

        * `path:' STRING

          This setting indicates the location of the executable program
          in the filesystem.  There is no default value for this
          setting.

        * `arguments:' LIST-OF-STRINGS

          This setting indicates the string arguments that are passed
          to the program.  The default value of this setting is the
          empty list (i.e. no arguments).

        * `environment:' LIST-OF-STRINGS

          This setting indicates the set of environment variable
          bindings that the process receives.  Each element of the list
          is a string of the form "`VAR=VALUE'", where `VAR' is the
          name of the variable and `VALUE' is its binding.  When
          LIST-OF-STRINGS is `#f', the process inherits the environment
          variable bindings of the Scheme program.  The default value
          of this setting is `#f'.

        * `directory:' DIR

          This setting indicates the current working directory of the
          process.  When DIR is `#f', the process uses the value of
          `(current-directory)'.  The default value of this setting is
          `#f'.

        * `stdin-redirection:' ( `#f' | `#t' )

          This setting indicates how the standard input of the process
          is redirected.  A setting of `#t' will redirect the standard
          input from the process-port (i.e. what is written to the
          process-port will be available on the standard input).  A
          setting of `#f' will leave the standard input as-is, which
          typically results in input coming from the console.  The
          default value of this setting is `#t'.

        * `stdout-redirection:' ( `#f' | `#t' )

          This setting indicates how the standard output of the process
          is redirected.  A setting of `#t' will redirect the standard
          output to the process-port (i.e. all output to standard
          output can be read from the process-port).  A setting of `#f'
          will leave the standard output as-is, which typically results
          in the output going to the console.  The default value of
          this setting is `#t'.

        * `stderr-redirection:' ( `#f' | `#t' )

          This setting indicates how the standard error of the process
          is redirected.  A setting of `#t' will redirect the standard
          error to the process-port (i.e. all output to standard error
          can be read from the process-port).  A setting of `#f' will
          leave the standard error as-is, which typically results in
          error messages being output to the console.  The default
          value of this setting is `#f'.

        * `pseudo-terminal:' ( `#f' | `#t' )

          This setting applies to UNIX.  It indicates what type of
          device will be bound to the process' standard input and
          standard output.  A setting of `#t' will use a
          pseudo-terminal device (this is a device that behaves like a
          tty device even though there is no real terminal or user
          directly involved).  A setting of `#f' will use a pair of
          pipes.  The difference is important for programs which behave
          differently when they are used interactively, for example
          shells.  The default value of this setting is `#f'.

        * `show-console:' ( `#f' | `#t' )

          This setting applies to Microsoft Windows.  It controls
          whether the process' console window will be hidden or
          visible.  The default value of this setting is `#t' (i.e.
          show the console window).


     For example:

          > (with-input-from-process "date" read-line)
          "Sun Jun 14 15:06:41 EDT 2009"
          > (define p (open-process (list path: "ls"
                                          arguments: '("../examples"))))
          > (read-line p)
          "README"
          > (read-line p)
          "Xlib-simple"
          > (close-port p)
          > (define p (open-process "/usr/bin/dc"))
          > (display "2 100 ^ p\n" p)
          > (force-output p)
          > (read-line p)
          "1267650600228229401496703205376"


 -- procedure: process-pid PROCESS-PORT
     This procedure returns the PID (Process Identifier) of the process
     of PROCESS-PORT.  The PID is a small exact integer.

     For example:

          > (let ((p (open-process "sort")))
              (process-pid p))
          318


 -- procedure: process-status PROCESS-PORT [TIMEOUT [TIMEOUT-VAL]]
     This procedure causes the current thread to wait until the process
     of PROCESS-PORT terminates (normally or not) or until the timeout
     is reached if TIMEOUT is supplied.  If the timeout is reached,
     PROCESS-STATUS returns TIMEOUT-VAL if it is supplied, otherwise an
     unterminated-process-exception object is raised.  The procedure
     returns the process exit status as encoded by the operating
     system.  Typically, if the process exited normally the return
     value is the process exit status multiplied by 256.

     For example:

          > (let ((p (open-process "sort")))
              (for-each (lambda (x) (pretty-print x p))
                        '(22 11 33))
              (close-output-port p)
              (let ((r (read-all p)))
                (close-input-port p)
                (list (process-status p) r)))
          (0 (11 22 33))


 -- procedure: unterminated-process-exception? OBJ
 -- procedure: unterminated-process-exception-procedure EXC
 -- procedure: unterminated-process-exception-arguments EXC
     Unterminated-process-exception objects are raised when a call to
     the `process-status' procedure reaches its timeout before the
     target process terminates and a timeout-value parameter is not
     specified.  The parameter EXC must be an
     unterminated-process-exception object.

     The procedure `unterminated-process-exception?' returns `#t' when
     OBJ is an unterminated-process-exception object and `#f' otherwise.

     The procedure `unterminated-process-exception-procedure' returns
     the procedure that raised EXC.

     The procedure `unterminated-process-exception-arguments' returns
     the list of arguments of the procedure that raised EXC.

     For example:

          > (define (handler exc)
              (if (unterminated-process-exception? exc)
                  (list (unterminated-process-exception-procedure exc)
                        (unterminated-process-exception-arguments exc))
                  'not-unterminated-process-exception))
          > (with-exception-catcher
              handler
              (lambda ()
                (let ((p (open-process "sort")))
                  (process-status p 1))))
          (#<procedure #2 process-status> (#<input-output-port #3 (process "sort")>))


14.7.3 Network devices
----------------------

 -- procedure: open-tcp-client PORT-NUMBER-OR-ADDRESS-OR-SETTINGS
     This procedure opens a network connection to a socket server and
     returns a tcp-client-port (a subtype of device-port) that
     represents this connection and allows communication with that
     server.  The default value of the `direction:' setting is
     `input-output', i.e. the Scheme program can send information to
     the server and receive information from the server.  The sending
     direction can be "shutdown" using the `close-output-port'
     procedure and the receiving direction can be "shutdown" using the
     `close-input-port' procedure.  The `close-port' procedure closes
     both directions of the connection.

     The parameter of this procedure is an IP port number (16-bit
     nonnegative exact integer), a string of the form `"HOST:PORT"' or
     a list of port settings.  When the parameter is the number PORT it
     is handled as if it was the setting `port-number:' PORT.  When the
     parameter is the string `"HOST:PORT"' it is handled as if it was
     the setting `address:' `"HOST:PORT"'.

     Here are the settings allowed in addition to the generic settings
     of byte-ports:

        * `address:' STRING-OR-IP-ADDRESS

          This setting indicates the internet address of the server, and
          possibly the IP port number.  When this parameter is not
          specified or is `""', the connection requests are sent to the
          loopback interface (with IP address 127.0.0.1).  The
          parameter can be a string denoting a host name, which will be
          translated to an IP address by the `host-info' procedure, or
          a 4 element u8vector which contains the 32-bit IPv4 address
          or an 8 element u16vector which contains the 128-bit IPv6
          address.  A string of the form `"HOST:PORT"' is handled as if
          it was the combination of settings `address:' `"HOST"'
          `port-number:' PORT.

        * `port-number:' 16-BIT-EXACT-INTEGER

          This setting indicates the IP port number of the server to
          connect to (e.g. 80 for the standard HTTP server, 23 for the
          standard telnet server).  There is no default value for this
          setting.

        * `local-address:' STRING-OR-IP-ADDRESS

          This setting indicates the internet address of the local
          network interface on which connections requests are
          initiated, and possibly the IP port number.  When this
          parameter is not specified or is `"*"', the connection
          requests are initiated on any network interface (i.e. address
          INADDR_ANY).  When this parameter is `""', the connection
          requests are initiated only on the loopback interface (with
          IP address 127.0.0.1).  The parameter can be a string
          denoting a host name, which will be translated to an IP
          address by the `host-info' procedure, or a 4 element u8vector
          which contains the 32-bit IPv4 address or an 8 element
          u16vector which contains the 128-bit IPv6 address.  A string
          of the form `"INTF:PORT"' is handled as if it was the
          combination of settings `local-address:' `"INTF"'
          `local-port-number:' PORT.

        * `local-port-number:' 16-BIT-EXACT-INTEGER

          This setting indicates the IP port number assigned to the
          socket which initiates connection requests.  The special
          value 0 requests that a currently unused port number be
          assigned to the socket.  This is the default value for this
          setting.

        * `keep-alive:' ( `#f' | `#t' )

          This setting controls the use of the "keep alive" option on
          the connection.  The "keep alive" option will periodically
          send control packets on otherwise idle network connections to
          ensure that the server host is active and reachable.  The
          default value of this setting is `#f'.

        * `coalesce:' ( `#f' | `#t' )

          This setting controls the use of TCP's "Nagle algorithm" which
          reduces the number of small packets by delaying their
          transmission and coalescing them into larger packets.  A
          setting of `#t' will coalesce small packets into larger ones.
          A setting of `#f' will transmit packets as soon as possible.
          The default value of this setting is `#t'.  Note that this
          setting does not affect the buffering of the port.

        * `tls-context:' ( `#f' | TLS-CONTEXT )

          This setting controls the use of TLS encryption.  If
          provided, the client will use this configuration for setting
          up a TCP connection with TLS encryption, otherwise it will
          use a plain TCP connection as usual.  Please note that Gambit
          must be compiled with TLS support for this option to be
          implemented.  See `make-tls-context' for futher information.
          The default value of this setting is `#f'.


     Below is an example of the client-side code that opens a
     connection to an HTTP server on port 8080 of the loopback
     interface (with IP address 127.0.0.1).  For the server-side code
     see the example for the procedure `open-tcp-server'.

          > (define p (open-tcp-client (list port-number: 8080
                                             eol-encoding: 'cr-lf)))
          > p
          #<input-output-port #2 (tcp-client #u8(127 0 0 1) 8080)>
          > (display "GET /\n" p)
          > (force-output p)
          > (read-line p)
          "<HTML>"


 -- procedure: open-tcp-server PORT-NUMBER-OR-ADDRESS-OR-SETTINGS
     This procedure sets up a socket to accept network connection
     requests from clients and returns a tcp-server-port from which
     network connections to clients are obtained.  Tcp-server-ports are
     a direct subtype of object-ports (i.e. they are not
     character-ports) and are input-ports.  Reading from a
     tcp-server-port with the `read' procedure will block until a
     network connection request is received from a client.  The `read'
     procedure will then return a tcp-client-port (a subtype of
     device-port) that represents this connection and allows
     communication with that client.  Closing a tcp-server-port with
     either the `close-input-port' or `close-port' procedures will
     cause the network subsystem to stop accepting connections on that
     socket.

     The parameter of this procedure is an IP port number (16-bit
     nonnegative exact integer), a string of the form `"INTF:PORT"' or
     a list of port settings which must contain a `local-port-number:'
     setting.  When the parameter is the number PORT it is handled as
     if it was the setting `local-port-number:' PORT.  When the
     parameter is the string `"INTF:PORT"' it is handled as if it was
     the setting `local-address:' `"INTF:PORT"'.

     Below is a list of the settings allowed in addition to the settings
     `keep-alive:' and `coalesce:' allowed by the `open-tcp-client'
     procedure and the generic settings of byte-ports.  The settings
     which are not listed below apply to the tcp-client-port that is
     returned by `read' when a connection is accepted and have the same
     meaning as if they were used in a call to the `open-tcp-client'
     procedure.

        * `local-address:' STRING-OR-IP-ADDRESS

          This setting indicates the internet address of the local
          network interface on which connections requests are accepted,
          and possibly the IP port number.  When this parameter is not
          specified or is `""', the connection requests are accepted
          only on the loopback interface (with IP address 127.0.0.1).
          When this parameter is `"*"', the connection requests are
          accepted on all network interfaces (i.e. address INADDR_ANY).
          The parameter can be a string denoting a host name, which
          will be translated to an IP address by the `host-info'
          procedure, or a 4 element u8vector which contains the 32-bit
          IPv4 address or an 8 element u16vector which contains the
          128-bit IPv6 address.  A string of the form `"INTF:PORT"' is
          handled as if it was the combination of settings
          `local-address:' `"INTF"' `local-port-number:' PORT.

        * `local-port-number:' 16-BIT-EXACT-INTEGER

          This setting indicates the IP port number assigned to the
          socket which accepts connection requests from clients.  So
          called "well-known ports", which are reserved for standard
          services, have a port number below 1024 and can only be
          assigned to a socket by a process with superuser priviledges
          (e.g. 80 for the HTTP service, 23 for the telnet service).
          No special priviledges are needed to assign higher port
          numbers to a socket.  The special value 0 requests that a
          currently unused port number be assigned to the socket (the
          port number assigned can be retrieved using the procedure
          `tcp-server-socket-info').  There is no default value for this
          setting.

        * `backlog:' POSITIVE-EXACT-INTEGER

          This setting indicates the maximum number of connection
          requests that can be waiting to be accepted by a call to
          `read' (technically it is the value passed as the second
          argument of the UNIX `listen()' function).  The default value
          of this setting is 128.

        * `reuse-address:' ( `#f' | `#t' )

          This setting controls whether it is possible to assign a port
          number that is currently active.  Note that when a server
          process terminates, the socket it was using to accept
          connection requests does not become inactive immediately.
          Instead it remains active for a few minutes to ensure clean
          termination of the connections.  A setting of `#f' will cause
          an exception to be raised in that case.  A setting of `#t'
          will allow a port number to be used even if it is active.
          The default value of this setting is `#t'.

        * `tls-context:' ( `#f' | TLS-CONTEXT )

          This setting controls the use of TLS encryption.  If
          provided, the server will use this configuration for
          accepting TCP connections with TLS encryption, otherwise it
          will accept plain TCP connections as usual.  Please note that
          Gambit must be compiled with TLS support for this option to be
          implemented.  See `make-tls-context' for futher information.
          The default value of this setting is `#f'.


     Below is an example of the server-side code that accepts
     connections on port 8080 of any network interface.  For the
     client-side code see the example for the procedure
     `open-tcp-client'.

          > (define s (open-tcp-server (list local-address: "*"
                                             local-port-number: 8080
                                             eol-encoding: 'cr-lf)))
          > (define p (read s))  ; blocks until client connects
          > p
          #<input-output-port #2 (tcp-client 8080)>
          > (read-line p)
          "GET /"
          > (display "<HTML>\n" p)
          > (force-output p)


 -- procedure: tcp-service-register! PORT-NUMBER-OR-ADDRESS-OR-SETTINGS
          THUNK [THREAD-GROUP]
 -- procedure: tcp-service-unregister!
          PORT-NUMBER-OR-ADDRESS-OR-SETTINGS
     The procedure `tcp-service-register!' sets up a socket to accept
     network connection requests from clients and creates a "service"
     thread which processes the incoming connections and returns this
     thread.  The parameter PORT-NUMBER-OR-ADDRESS-OR-SETTINGS has the
     same meaning as for the procedure `open-tcp-server'.

     For each connection established the service thread creates a
     "handler" thread which executes a call to the procedure THUNK with
     no argument.  The handler thread's current input-port and current
     output-port are both set to the tcp-client-port created for the
     connection.  There is no need for the THUNK to close the
     tcp-client-port, as this is done by the handler thread when the
     THUNK returns normally.

     The procedure `tcp-service-unregister!' terminates the service
     thread which was registered by `tcp-service-register!' with the
     same network interface and port number (if a service thread is
     still registered).  The procedure `tcp-service-register!'
     implicitly calls `tcp-service-unregister!' before registering the
     new service thread.

          > (tcp-service-register!
             8000
             (lambda () (display "hello\n")))
          > (define p (open-tcp-client 8000))
          > (read-line p)
          "hello"
          > (tcp-service-unregister! 8000)


 -- procedure: make-tls-context [OPTIONS]
     This procedure requires Gambit to be compiled with TLS support,
     which is currently provided by OpenSSL. The `--enable-openssl'
     flag of the configure script will activate it, provided that you
     have the OpenSSL library and headers installed. It is strongly
     recommended that versions above 1.x are used.  On OSX, this means
     updating the OpenSSL bundled by default. This can be achieved
     using Homebrew, but manual installation or any other package
     manager will do.  Some notes on Windows with MinGW are also
     relevant here. Once you have a sane MinGW environment, remember to
     decompress the OpenSSL tarball with the tar utility, otherwise
     links to files won't work during the compilation process. The
     recommended build procedure for MinGW is as follows.

     Configure OpenSSL on MinGW 32 bits:

          perl Configure mingw no-asm --prefix=/usr/local --openssldir=/usr/local/openssl

     Configure OpenSSL on MinGW 64 bits:

          perl Configure mingw64 no-asm --prefix=/usr/local --openssldir=/usr/local/openssl

     Build and install with the following commands:

          make depend
          make
          make install

     A TLS context describes the options that will be used for setting
     up a TLS connection. If no TLS context is provided to
     `open-tcp-client' or `open-tcp-server', regular TCP connections
     without encryption will be used instead. The result of this
     procedure is a `SSL_CTX' pointer, which can be further manipulated
     with custom OpenSSL bindings. The configuration options are:

        * `min-version:' SYMBOL

          Establish a minimum TLS version for the connection. If the
          other peer doesn't support or agree with it, the connection
          will fail. Possible options (support depends on linked
          OpenSSL version): `ssl-v2', `ssl-v3', `tls-v1', `tls-v1.1',
          `tls-v1.2'.

        * `options:' LIST-OF-SYMBOLS

          A list of flags enabling/disabling TLS options. `server-mode'
          is required for using the TLS context with `open-tcp-server'.
          `use-diffie-hellman' enables the Diffie-Hellman key exchange.
          `use-elliptic-curves' enables Elliptic Curves. If no curve
          name is provided (with `elliptic-curve:'), `prime256v1' will
          be used. `request-client-authentication' is used by a server
          to enable request of authentication to clients.
          `insert-empty-fragments' enables a countermeasure against a
          SSL 3.0/TLS 1.0 protocol vulnerability affecting CBC ciphers.
          If used, the resulting connection may not be handled by some
          broken SSL implementations. This option has no effect for
          connections using other ciphers.

        * `certificate:' PATH

          Path to PEM Certificate file. This is a recommended option.
          If not provided OpenSSL will try to use anonymous cipher
          suites.

        * `private-key:' PATH

          Path to PEM Private Key file. If not provided, the
          Certificate path will be used instead.

        * `client-ca:' PATH

          Path to PEM file containing Certificate Authorities allowed
          for client authentication. Used only if
          `request-client-authentication' option is enabled.

        * `elliptic-curve:' STRING

          Name of the Elliptic Curve to use, according to RFC 4492.
          Used only if `use-elliptic-curves' option is enabled.


     TCP Client example with TLS encryption.

          (define ctx (make-tls-context))
          (define c (open-tcp-client (list address: "twitter.com"
                                           port-number: 443
                                           tls-context: ctx)))
          (display "GET / HTTP/1.1\nHost: twitter.com\n\n" c)
          (force-output c)
          (read-line c)

     TCP Server example with several options. These are not mandatory,
     except for `server-mode'.

          (define ctx (make-tls-context options: '(server-mode
                                                   use-diffie-hellman
                                                   use-elliptic-curves)
                                        certificate: "server.pem"
                                        diffie-hellman-parameters: "dh_param_1024.pem"
                                        elliptic-curve: "prime256v1"))

          (define s (open-tcp-server (list local-address: "localhost"
                                           local-port-number: 1443
                                           tls-context: ctx)))
          (define p (read s))
          (display "<HTML></HTML>\n" p)
          (force-output p)

     A practical way of testing TLS options are the `s_server' and
     `s_client' commands of the `openssl' tool.


 -- procedure: open-udp PORT-NUMBER-OR-ADDRESS-OR-SETTINGS
     This procedure opens a socket for doing network communication with
     the UDP protocol.  The default value of the `direction:' setting is
     `input-output', i.e. the Scheme program can send information and
     receive information on the socket.  The sending direction can be
     closed using the `close-output-port' procedure and the receiving
     direction can be closed using the `close-input-port' procedure.
     The `close-port' procedure closes both directions.

     The resulting port designates a UDP socket.  Each call to `read'
     and `udp-read-subu8vector' causes the reception of a single
     datagram on the designated UDP socket, and each call to `write'
     and `udp-write-subu8vector' sends a single datagram.  UDP ports
     are a direct subtype of object-ports (i.e. they are not
     character-ports) and `read' and `write' transfer u8vectors.  If
     `read' is called and a timeout occurs before a datagram is
     transferred and the timeout thunk returns `#f' (see the procedure
     `input-port-timeout-set!') then the end-of-file object is returned.

     The parameter of this procedure is an IP port number (16-bit
     nonnegative exact integer), a string of the form `"HOST:PORT"' or
     a list of port settings.  When the parameter is the number PORT it
     is handled as if it was the setting `local-port-number:' PORT.
     When the parameter is the string `"HOST:PORT"' it is handled as if
     it was the setting `local-address:' `"HOST:PORT"'.

     Here are the settings allowed:

        * `direction:' ( `input' | `output' | `input-output' )

          This setting controls the direction of the port.  The symbol
          `input' indicates a unidirectional input-port, the symbol
          `output' indicates a unidirectional output-port, and the
          symbol `input-output' indicates a bidirectional port.  The
          default value of this setting is `input-output'.

        * `local-address:' STRING-OR-IP-ADDRESS

          This setting indicates the internet address of the local
          network interface on which the socket is open and possibly
          the IP port number.  When this parameter is not specified or
          is `""', the socket is open on the loopback interface (with
          IP address 127.0.0.1).  When this parameter is `"*"' the
          socket is open on all network interfaces (i.e. address
          INADDR_ANY).  The parameter can be a string denoting a host
          name, which will be translated to an IP address by the
          `host-info' procedure, or a 4 element u8vector which contains
          the 32-bit IPv4 address or an 8 element u16vector which
          contains the 128-bit IPv6 address.  A string of the form
          `"INTF:PORT"' is handled as if it was the combination of
          settings `local-address:' `"INTF"' `local-port-number:' PORT.

        * `local-port-number:' 16-BIT-EXACT-INTEGER

          This setting indicates the IP port number assigned to the
          socket.  The special value 0 requests that a currently unused
          port number be assigned to the socket.  This is the default
          value for this setting.

        * `address:' STRING-OR-IP-ADDRESS

          This setting indicates the initial destination internet
          address of the datagrams, and possibly the IP port number.
          When this parameter is not specified the destination is set
          to the local address if it is not all network interfaces
          (i.e. `"*"' = address INADDR_ANY).  When this parameter is
          `""' or this parameter is not specified and the local address
          is all network interfaces, the destination is set to the
          loopback interface (with IP address 127.0.0.1).  The
          parameter can be a string denoting a host name, which will be
          translated to an IP address by the `host-info' procedure, or
          a 4 element u8vector which contains the 32-bit IPv4 address
          or an 8 element u16vector which contains the 128-bit IPv6
          address.  A string of the form `"HOST:PORT"' is handled as if
          it was the combination of settings `address:' `"HOST"'
          `port-number:' PORT.

        * `port-number:' 16-BIT-EXACT-INTEGER

          This setting indicates the initial destination IP port number
          of the datagrams.  It defaults to the local port number.



 -- procedure: udp-destination-set! ADDRESS PORT-NUMBER [UDP-PORT]
     This procedure sets the destination address and port-number for the
     next datagram sent on the UDP socket designated by UDP-PORT,
     obtained with a call to `open-udp'.  If it is not specified,
     UDP-PORT defaults to the current output-port.


 -- procedure: udp-read-u8vector [UDP-PORT]
 -- procedure: udp-write-u8vector U8VECTOR [UDP-PORT]
 -- procedure: udp-read-subu8vector U8VECTOR START END [UDP-PORT]
 -- procedure: udp-write-subu8vector U8VECTOR START END [UDP-PORT]
     These procedures receive and send datagrams on the UDP socket
     designated by UDP-PORT, obtained with a call to `open-udp'.  If it
     is not specified, UDP-PORT defaults to the current input-port for
     `udp-read-u8vector' and `udp-read-subu8vector' and to the current
     output-port for `udp-write-u8vector' and `udp-write-subu8vector'.

     These procedures are similar in function to `read-subu8vector' and
     `write-subu8vector', but because they read or write a group of
     bytes at a time rather than a stream of bytes, they are distinct
     procedures with slightly different APIs.

     The procedure `udp-read-u8vector' receives the next datagram and
     returns it in a fresh u8vector.  If a timeout occurs before a
     datagram is transferred and the timeout thunk returns `#f' (see the
     procedure `input-port-timeout-set!') then `#f' is returned.

     The procedure `udp-write-u8vector' sends as a datagram the
     u8vector U8VECTOR.

     For the procedures `udp-read-subu8vector' and
     `udp-write-subu8vector', the part of the u8vector U8VECTOR
     starting at index START and ending just before index END is used
     as a byte buffer that will be the target of `udp-read-subu8vector'
     or the source of the `udp-write-subu8vector'.  Up to END-START
     bytes will be transferred.  The number of bytes transferred,
     possibly zero, is returned by these procedures, unless a timeout
     occurs (see below).  Fewer bytes will be read by
     `udp-read-subu8vector' if the received datagram's length is less
     than END-START.  `udp-write-subu8vector' always transfers END-START
     bytes, but note that this must be less than 65536 bytes, and some
     operating systems have a lower limit (for example macOS limits the
     number of bytes to 9216 by default).  If a timeout occurs before a
     datagram is transferred and the timeout thunk returns `#f' (see
     the procedure `input-port-timeout-set!') then `#f' is returned by
     these procedures (this is different from the procedures
     `read-subu8vector' and `write-subu8vector' which return 0).

     For `udp-write-u8vector' and `udp-write-subu8vector' the
     datagram's destination is the address initially supplied when
     `open-udp' was called, or the latest address set when
     `udp-destination-set!'  was called.

     Here is an example of sending a 3 byte datagram to port 5678 of the
     loopback interface (with IP address 127.0.0.1):

          > (define p (open-udp (list address: '#u8(127 0 0 1) port-number: 5678)))
          > (write '#u8(11 22 33) p)

     An alternative approach is to use `udp-destination-set!':

          > (define p (open-udp))
          > (udp-destination-set! '#u8(127 0 0 1) 5678 p)
          > (write '#u8(11 22 33) p)

     Another approach is to use `udp-write-subu8vector':

          > (define p (open-udp))
          > (udp-destination-set! '#u8(127 0 0 1) 5678 p)
          > (define v '#u8(11 22 33))
          > (udp-write-subu8vector v 0 3 p)
          3

     Note that by default the internet address of the local network
     interface is the loopback interface, which is not connected to the
     internet.  To contact an external socket the address of the local
     network interface must be specified, for example `"*"' will select
     all interfaces.  The following example shows how to connect to a
     Time Protocol server to get the current time:

          > (define p (open-udp (list local-address: "*" address: "time.nist.gov:37")))
          > (write '#u8() p)
          > (read p)
          #u8(222 27 158 226)

     Here is an example of receiving a 3 byte datagram on port 5678 of
     the loopback interface:

          > (define p (open-udp 5678))
          > (read p)
          #u8(11 22 33)

     An alternative approach is to use `udp-read-subu8vector':

          > (define p (open-udp 5678))
          > (define v (make-u8vector 10000))
          > (udp-read-subu8vector v 0 10000 p)
          3
          > (subu8vector v 0 3)
          #u8(11 22 33)

     Note that using `udp-read-subu8vector' and `udp-write-subu8vector'
     is typically more efficient than `read' and `write' because it
     avoids having to construct a new u8vector for each datagram
     transferred.


 -- procedure: udp-local-socket-info UDP-PORT
 -- procedure: udp-source-socket-info UDP-PORT
     The procedure `udp-local-socket-info' returns the local
     socket-info of the UDP socket designated by UDP-PORT.

     The procedure `udp-source-socket-info' returns the socket-info of
     the source of the latest datagram received on the UDP socket
     designated by UDP-PORT.  When a datagram hasn't been received yet,
     `#f' is returned.

     For example:

          > (define p (open-udp (list local-address: "*" address: "time.nist.gov:37")))
          > (udp-local-socket-info p)
          #<socket-info #2 family: INET port-number: 64716 address: #f>
          > (udp-source-socket-info p)
          #f
          > (write '#u8() p)
          > (read p)
          #u8(222 27 162 109)
          > (udp-source-socket-info p)
          #<socket-info #3 family: INET port-number: 37 address: #u8(132 163 97 4)>


14.8 Directory-ports
====================

 -- procedure: open-directory PATH-OR-SETTINGS
     This procedure opens a directory of the filesystem for reading its
     entries and returns a directory-port from which the entries can be
     enumerated.  Directory-ports are a direct subtype of object-ports
     (i.e. they are not character-ports) and are input-ports.  Reading
     from a directory-port with the `read' procedure returns the next
     file name in the directory as a string.  The end-of-file object is
     returned when all the file names have been enumerated.  Another
     way to get the list of all files in a directory is the
     `directory-files' procedure which returns a list of the files in
     the directory.  The advantage of using directory-ports is that it
     allows iterating over the files in a directory in constant space,
     which is interesting when the number of files in the directory is
     not known in advance and may be large.  Note that the order in
     which the names are returned is operating-system dependent.

     The parameter of this procedure is either a string denoting a
     filesystem path to a directory or a list of port settings which
     must contain a `path:' setting.  Here are the settings allowed in
     addition to the generic settings of object-ports:

        * `path:' STRING

          This setting indicates the location of the directory in the
          filesystem.  There is no default value for this setting.

        * `ignore-hidden:' ( `#f' | `#t' | `dot-and-dot-dot' )

          This setting controls whether hidden-files will be returned.
          Under UNIX and macOS hidden-files are those that start with a
          period (such as `.', `..', and `.profile').  Under Microsoft
          Windows hidden files are the `.' and `..' entries and the
          files whose "hidden file" attribute is set.  A setting of `#f'
          will enumerate all the files.  A setting of `#t' will only
          enumerate the files that are not hidden.  A setting of
          `dot-and-dot-dot' will enumerate all the files except for the
          `.' and `..' hidden files.  The default value of this setting
          is `#t'.


     For example:

          > (let ((p (open-directory (list path: "../examples"
                                           ignore-hidden: #f))))
              (let loop ()
                (let ((fn (read p)))
                  (if (string? fn)
                      (begin
                        (pp (path-expand fn))
                        (loop)))))
              (close-input-port p))
          "/u/feeley/examples/."
          "/u/feeley/examples/.."
          "/u/feeley/examples/complex"
          "/u/feeley/examples/README"
          "/u/feeley/examples/simple"
          > (define x (open-directory "../examples"))
          > (read-all x)
          ("complex" "README" "simple")


14.9 Vector-ports
=================

 -- procedure: open-vector [VECTOR-OR-SETTINGS]
 -- procedure: open-input-vector [VECTOR-OR-SETTINGS]
 -- procedure: open-output-vector [VECTOR-OR-SETTINGS]
 -- procedure: call-with-input-vector VECTOR-OR-SETTINGS PROC
 -- procedure: call-with-output-vector [VECTOR-OR-SETTINGS] PROC
 -- procedure: with-input-from-vector VECTOR-OR-SETTINGS THUNK
 -- procedure: with-output-to-vector [VECTOR-OR-SETTINGS] THUNK
     Vector-ports represent streams of Scheme objects.  They are a
     direct subtype of object-ports (i.e. they are not
     character-ports).  All of these procedures create vector-ports
     that are either unidirectional or bidirectional.  The `direction:'
     setting will default to the value `input' for the procedures
     `open-input-vector', `call-with-input-vector' and
     `with-input-from-vector', to the value `output' for the procedures
     `open-output-vector', `call-with-output-vector' and
     `with-output-to-vector', and to the value `input-output' for the
     procedure `open-vector'.  Bidirectional vector-ports behave like
     FIFOs: data written to the port is added to the end of the stream
     that is read.  It is only when a bidirectional vector-port's
     output-side is closed with a call to the `close-output-port'
     procedure that the stream's end is known (when the stream's end is
     reached, reading the port returns the end-of-file object).

     The procedures `open-vector', `open-input-vector' and
     `open-output-vector' return the port that is created.  The
     procedures `call-with-input-vector' and `call-with-output-vector'
     create a vector port, call the procedure PROC with the port as
     single argument and then close the port.  The procedures
     `with-input-from-vector' and `with-output-to-vector' create a
     vector port, dynamically bind the current input-port and current
     output-port respectively to the port created for the duration of a
     call to the procedure THUNK with no argument, and then close the
     port.  The procedures `call-with-input-vector' and
     `with-input-from-vector' return the value returned by the
     procedures PROC and THUNK respectively.  The procedures
     `call-with-output-vector' and `with-output-to-vector' return the
     vector accumulated in the port (see `get-output-vector').

     The VECTOR-OR-SETTINGS parameter of these procedures is either a
     vector of the elements used to initialize the stream or a list of
     port settings.  If it is not specified, the parameter of the
     `open-vector', `open-input-vector', `open-output-vector',
     `with-output-to-vector', and `call-with-output-vector' procedures
     defaults to an empty list of port settings.  Here are the settings
     allowed in addition to the generic settings of object-ports:

        * `init:' VECTOR

          This setting indicates the initial content of the stream.
          The default value of this setting is an empty vector.

        * `permanent-close:' ( `#f' | `#t' )

          This setting controls whether a call to the procedures
          `close-output-port' will close the output-side of a
          bidirectional vector-port permanently or not.  A permanently
          closed bidirectional vector-port whose end-of-file has been
          reached on the input-side will return the end-of-file object
          for all subsequent calls to the `read' procedure.  A
          non-permanently closed bidirectional vector-port will return
          to its opened state when its end-of-file is read.  The
          default value of this setting is `#t'.


     For example:

          > (define p (open-vector))
          > (write 1 p)
          > (write 2 p)
          > (write 3 p)
          > (force-output p)
          > (read p)
          1
          > (read p)
          2
          > (close-output-port p)
          > (read p)
          3
          > (read p)
          #!eof
          > (with-output-to-vector (lambda () (write 1) (write 2)))
          #(1 2)


 -- procedure: open-vector-pipe [VECTOR-OR-SETTINGS1
          [VECTOR-OR-SETTINGS2]]
     The procedure `open-vector-pipe' creates two vector-ports and
     returns these two ports.  The two ports are interrelated as
     follows: the first port's output-side is connected to the second
     port's input-side and the first port's input-side is connected to
     the second port's output-side.  The value VECTOR-OR-SETTINGS1 is
     used to setup the first vector-port and VECTOR-OR-SETTINGS2 is
     used to setup the second vector-port.  The same settings as for
     `open-vector' are allowed.  The default `direction:' setting is
     `input-output' (i.e. a bidirectional port is created).  If it is
     not specified VECTOR-OR-SETTINGS1 defaults to the empty list.  If
     it is not specified VECTOR-OR-SETTINGS2 defaults to
     VECTOR-OR-SETTINGS1 but with the `init:' setting set to the empty
     vector and with the input and output settings exchanged (e.g. if
     the first port is an input-port then the second port is an
     output-port, if the first port's input-side is non-buffered then
     the second port's output-side is non-buffered).

     For example:

          > (define (server op)
              (receive (c s) (open-vector-pipe)  ; client-side and server-side ports
                (thread-start!
                  (make-thread
                    (lambda ()
                      (let loop ()
                        (let ((request (read s)))
                          (if (not (eof-object? request))
                              (begin
                                (write (op request) s)
                                (newline s)
                                (force-output s)
                                (loop))))))))
                c))
          > (define a (server (lambda (x) (expt 2 x))))
          > (define b (server (lambda (x) (expt 10 x))))
          > (write 100 a)
          > (write 30 b)
          > (read a)
          1267650600228229401496703205376
          > (read b)
          1000000000000000000000000000000


 -- procedure: get-output-vector VECTOR-PORT
     The procedure `get-output-vector' takes an output vector-port or a
     bidirectional vector-port as parameter and removes all the objects
     currently on the output-side, returning them in a vector.  The port
     remains open and subsequent output to the port and calls to the
     procedure `get-output-vector' are possible.

     For example:

          > (define p (open-vector '#(1 2 3)))
          > (write 4 p)
          > (get-output-vector p)
          #(1 2 3 4)
          > (write 5 p)
          > (write 6 p)
          > (get-output-vector p)
          #(5 6)


14.10 String-ports
==================

 -- procedure: open-string [STRING-OR-SETTINGS]
 -- procedure: open-input-string [STRING-OR-SETTINGS]
 -- procedure: open-output-string [STRING-OR-SETTINGS]
 -- procedure: call-with-input-string STRING-OR-SETTINGS PROC
 -- procedure: call-with-output-string [STRING-OR-SETTINGS] PROC
 -- procedure: with-input-from-string STRING-OR-SETTINGS THUNK
 -- procedure: with-output-to-string [STRING-OR-SETTINGS] THUNK
 -- procedure: open-string-pipe [STRING-OR-SETTINGS1
          [STRING-OR-SETTINGS2]]
 -- procedure: get-output-string STRING-PORT
     String-ports represent streams of characters.  They are a direct
     subtype of character-ports.  These procedures are the string-port
     analog of the procedures specified in the vector-ports section.
     Note that these procedures are a superset of the procedures
     specified in the "Basic String Ports SRFI" (SRFI 6).

     For example:

          > (define p (open-string))
          > (write 1 p)
          > (write 2 p)
          > (write 3 p)
          > (force-output p)
          > (read-char p)
          #\1
          > (read-char p)
          #\2
          > (close-output-port p)
          > (read-char p)
          #\3
          > (read-char p)
          #!eof
          > (with-output-to-string (lambda () (write 1) (write 2)))
          "12"


 -- procedure: object->string OBJ [N]
     This procedure converts the object OBJ to its external
     representation and returns it in a string.  The parameter N
     specifies the maximal width of the resulting string.  If the
     external representation is wider than N, the resulting string will
     be truncated to N characters and the last 3 characters will be set
     to periods.  Note that the current readtable is used.

     For example:

          > (object->string (expt 2 100))
          "1267650600228229401496703205376"
          > (object->string (expt 2 100) 30)
          "126765060022822940149670320..."
          > (object->string (cons car cdr))
          "(#<procedure #2 car> . #<procedure #3 cdr>)"


14.11 U8vector-ports
====================

 -- procedure: open-u8vector [U8VECTOR-OR-SETTINGS]
 -- procedure: open-input-u8vector [U8VECTOR-OR-SETTINGS]
 -- procedure: open-output-u8vector [U8VECTOR-OR-SETTINGS]
 -- procedure: call-with-input-u8vector U8VECTOR-OR-SETTINGS PROC
 -- procedure: call-with-output-u8vector [U8VECTOR-OR-SETTINGS] PROC
 -- procedure: with-input-from-u8vector U8VECTOR-OR-SETTINGS THUNK
 -- procedure: with-output-to-u8vector [U8VECTOR-OR-SETTINGS] THUNK
 -- procedure: open-u8vector-pipe [U8VECTOR-OR-SETTINGS1
          [U8VECTOR-OR-SETTINGS2]]
 -- procedure: get-output-u8vector U8VECTOR-PORT
     U8vector-ports represent streams of bytes.  They are a direct
     subtype of byte-ports.  These procedures are the u8vector-port
     analog of the procedures specified in the vector-ports section.

     For example:

          > (define p (open-u8vector))
          > (write 1 p)
          > (write 2 p)
          > (write 3 p)
          > (force-output p)
          > (read-u8 p)
          49
          > (read-u8 p)
          50
          > (close-output-port p)
          > (read-u8 p)
          51
          > (read-u8 p)
          #!eof
          > (with-output-to-u8vector (lambda () (write 1) (write 2)))
          #u8(49 50)


14.12 Other procedures related to I/O
=====================================

 -- procedure: current-input-port [NEW-VALUE]
 -- procedure: current-output-port [NEW-VALUE]
 -- procedure: current-error-port [NEW-VALUE]
 -- procedure: current-readtable [NEW-VALUE]
     These procedures are parameter objects which represent
     respectively: the current input-port, the current output-port, the
     current error-port, and the current readtable.


 -- procedure: print [`port:' PORT] OBJ...
 -- procedure: println [`port:' PORT] OBJ...
     The `print' procedure writes a representation of each OBJ, from
     left to right, to PORT.  When a compound object is encountered
     (pair, list, vector, box) the elements of that object are
     recursively written without the surrounding tokens (parentheses,
     spaces, dots, etc).  Strings, symbols, keywords and characters are
     written like the `display' procedure.  If there are more than one
     OBJ, the first OBJ must not be a keyword object.  If it is not
     specified, PORT defaults to the current output-port.  The
     procedure `print' returns an unspecified value.

     The `println' procedure does the same thing as the `print'
     procedure and then writes an end of line to PORT.

     For example:

          > (println "2*2 is " (* 2 2) " and 2+2 is " (+ 2 2))
          2*2 is 4 and 2+2 is 4
          > (define x (list "<i>" (list "<tt>" 123 "</tt>") "</i>"))
          > (println x)
          <i><tt>123</tt></i>
          > (define p (open-output-string))
          > (print port: p 1 #\2 "345")
          > (get-output-string p)
          "12345"


 -- procedure: read-file-string PATH-OR-SETTINGS
 -- procedure: read-file-string-list PATH-OR-SETTINGS
 -- procedure: read-file-u8vector PATH-OR-SETTINGS
     These procedures open the file designated by PATH-OR-SETTINGS and
     read the whole content.  They respectively return a string (the
     characters in the file), a list of strings (the lines in the file),
     and a u8vector (the bytes in the file).

     The PATH-OR-SETTINGS parameter is either a string denoting a
     filesystem path or a list of port settings which must contain a
     `path:' setting.  The same settings as `open-input-file' are
     allowed, and the same default settings are used.  The default
     value of the `char-encoding:' setting (which is relevant for
     `read-file-string' and `read-file-string-list') depends on how the
     runtime system was configured but typically UTF-8 is used.  The
     default can be overridden through various runtime options (*note
     Runtime options::), such as `-:file-settings=...' and
     `-:io-settings=...'.

     For example:

          > (with-output-to-file
              "test"
              (lambda () (for-each pretty-print (map square (iota 5)))))
          > (read-file-string "test")
          "0\n1\n4\n9\n16\n"
          > (read-file-string-list "test")
          ("0" "1" "4" "9" "16")
          > (read-file-u8vector "test")
          #u8(48 10 49 10 52 10 57 10 49 54 10)
          > (utf8->string (read-file-u8vector "test"))
          "0\n1\n4\n9\n16\n"


 -- procedure: write-file-string PATH-OR-SETTINGS STRING
 -- procedure: write-file-string-list PATH-OR-SETTINGS STRING-LIST
 -- procedure: write-file-u8vector PATH-OR-SETTINGS U8VECTOR
     These procedures open the file designated by PATH-OR-SETTINGS and
     write the data specified by the second parameter.  They
     respectively write a string (the characters to write to the file),
     a list of strings (the lines to write to the file), and a u8vector
     (the bytes to write to the file).  These procedures return the
     void object.

     The PATH-OR-SETTINGS parameter is either a string denoting a
     filesystem path or a list of port settings which must contain a
     `path:' setting.  The same settings as `open-output-file' are
     allowed, and the same default settings are used.  The default
     value of the `char-encoding:' setting (which is relevant for
     `write-file-string' and `write-file-string-list') depends on how
     the runtime system was configured but typically UTF-8 is used.
     The default can be overridden through various runtime options
     (*note Runtime options::), such as `-:file-settings=...' and
     `-:io-settings=...'.

     For example:

          > (write-file-string "test" "1\n2\n3\n")
          > (read-file-u8vector "test")
          #u8(49 10 50 10 51 10)
          > (write-file-string-list "test" (list "1" "2" "3"))
          > (read-file-u8vector "test")
          #u8(49 10 50 10 51 10)
          > (write-file-u8vector "test" (u8vector 97 98 99))
          > (read-file-string "test")
          "abc"


15 Lexical syntax and readtables
********************************

15.1 Readtables
===============

Readtables control the external textual representation of Scheme
objects, that is the encoding of Scheme objects using characters.
Readtables affect the behavior of the reader (i.e. the `read' procedure
and the parser used by the `load' procedure and the interpreter and
compiler) and the printer (i.e. the procedures `write', `display',
`print', `println', `pretty-print', and `pp', and the procedure used by
the REPL to print results).  To preserve write/read invariance the
printer and reader must be using compatible readtables.  For example a
symbol which contains upper case letters will be printed with special
escapes if the readtable indicates that the reader is case-insensitive.

   Readtables are immutable records whose fields specify various textual
representation aspects.  There are accessor procedures to retrieve the
content of specific fields.  There are also functional update
procedures that create a copy of a readtable, with a specific field set
to a new value.

 -- procedure: readtable? OBJ
     This procedure returns `#t' when OBJ is a readtable and `#f'
     otherwise.

     For example:

          > (readtable? (current-readtable))
          #t
          > (readtable? 123)
          #f


 -- procedure: readtable-case-conversion? READTABLE
 -- procedure: readtable-case-conversion?-set READTABLE NEW-VALUE
     The procedure `readtable-case-conversion?' returns the content of
     the `case-conversion?' field of READTABLE.  When the content of
     this field is `#f', the reader preserves the case of symbols,
     keyword and named characters that are read (i.e. `Ice' and `ice'
     are distinct symbols).  When the content of this field is the
     symbol `upcase', the reader converts to uppercase (i.e. `Ice' is
     read as the symbol `(string->symbol "ICE")').  Otherwise the reader
     converts using `string-foldcase', which for many letters converts
     them to lowercase (i.e. `Ice' is read as the symbol
     `(string->symbol "ice")').

     The procedure `readtable-case-conversion?-set' returns a copy of
     READTABLE where only the `case-conversion?' field has been changed
     to NEW-VALUE.

     For example:

          > (output-port-readtable-set!
              (repl-output-port)
              (readtable-case-conversion?-set
                (output-port-readtable (repl-output-port))
                #f))
          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-case-conversion?-set
                (input-port-readtable (repl-input-port))
                #f))
          > 'Ice
          Ice
          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-case-conversion?-set
                (input-port-readtable (repl-input-port))
                #t))
          > 'Ice
          ice
          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-case-conversion?-set
                (input-port-readtable (repl-input-port))
                'upcase))
          > 'Ice
          ICE


 -- procedure: readtable-keywords-allowed? READTABLE
 -- procedure: readtable-keywords-allowed?-set READTABLE NEW-VALUE
     The procedure `readtable-keywords-allowed?' returns the content of
     the `keywords-allowed?' field of READTABLE.  When the content of
     this field is `#f', the reader does not recognize keyword objects
     (i.e. `:foo' and `foo:' are read as the symbols `(string->symbol
     ":foo")' and `(string->symbol "foo:")' respectively).  When the
     content of this field is the symbol `prefix', the reader
     recognizes keyword objects that start with a colon, as in Common
     Lisp (i.e. `:foo' is read as the keyword `(string->keyword
     "foo")').  Otherwise the reader recognizes keyword objects that
     end with a colon, as in DSSSL (i.e. `foo:' is read as the symbol
     `(string->symbol "foo")').

     The procedure `readtable-keywords-allowed?-set' returns a copy of
     READTABLE where only the `keywords-allowed?' field has been
     changed to NEW-VALUE.

     For example:

          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-keywords-allowed?-set
                (input-port-readtable (repl-input-port))
                #f))
          > (map keyword? '(foo :foo foo:))
          (#f #f #f)
          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-keywords-allowed?-set
                (input-port-readtable (repl-input-port))
                #t))
          > (map keyword? '(foo :foo foo:))
          (#f #f #t)
          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-keywords-allowed?-set
                (input-port-readtable (repl-input-port))
                'prefix))
          > (map keyword? '(foo :foo foo:))
          (#f #t #f)


 -- procedure: readtable-sharing-allowed? READTABLE
 -- procedure: readtable-sharing-allowed?-set READTABLE NEW-VALUE
     The procedure `readtable-sharing-allowed?' returns the content of
     the `sharing-allowed?' field of READTABLE.  The reader recognizes
     the `#N#' and `#N=DATUM' notation for shared and circular
     structures and the printer uses this notation depending on the
     content of the `sharing-allowed?' field and the printing primitive
     used.  When the content of the `sharing-allowed?' field is the
     symbol `default', the procedure `write-shared' will use this
     notation for shared and circular structures, the procedures
     `write', `display', `pp', `pretty-print', `print', and `println'
     will use this notation for circular structures, and the procedure
     `write-simple' does not use this notation.  When the content of
     the `sharing-allowed?' field is `#f', the printing procedures will
     not use this notation. When the content of the `sharing-allowed?'
     field is `#t', the printing procedures will use this notation for
     shared and circular structures.  Finally, when the content of the
     `sharing-allowed?' field is the symbol `serialize', the printer
     uses a special external representation that the reader understands
     and that extends write/read invariance to the following types:
     records, procedures and continuations.  Note that an object can be
     serialized and deserialized if and only if all of its components
     are serializable.

     The procedure `readtable-sharing-allowed?-set' returns a copy of
     READTABLE where only the `sharing-allowed?' field has been changed
     to NEW-VALUE.

     Here is a simple example:

          > (define (wr obj allow?)
              (call-with-output-string
                (lambda (p)
                  (output-port-readtable-set!
                    p
                    (readtable-sharing-allowed?-set
                      (output-port-readtable p)
                      allow?))
                  (write obj p))))
          > (define (rd str allow?)
              (call-with-input-string
                str
                (lambda (p)
                  (input-port-readtable-set!
                    p
                    (readtable-sharing-allowed?-set
                      (input-port-readtable p)
                      allow?))
                  (read p))))
          > (define x (list 1 2 3))
          > (set-car! (cdr x) (cddr x))
          > (wr x #f)
          "(1 (3) 3)"
          > (wr x #t)
          "(1 #0=(3) . #0#)"
          > (define y (rd (wr x #t) #t))
          > y
          (1 (3) 3)
          > (eq? (cadr y) (cddr y))
          #t
          > (define f #f)
          > (let ((free (expt 2 10)))
            (set! f (lambda (x) (+ x free))))
          > (define s (wr f 'serialize))
          > (string-length s)
          4196
          > (define g (rd s 'serialize))
          > (eq? f g)
          #f
          > (g 4)
          1028

     Continuations are tricky to serialize because they contain a
     dynamic environment and this dynamic environment may contain
     non-serializable objects, in particular ports attached to
     operating-system streams such as files, the console or standard
     input/output.  Indeed, all dynamic environments contain a binding
     for the `current-input-port' and `current-output-port'.  Moreover,
     any thread that has started a REPL has a continuation which refers
     to the "repl-context" object in its dynamic environment.  A
     repl-context object contains the interaction channel, which is
     typically connected to a non-serializable port, such as the
     console.  Another problem is that the `parameterize' form saves
     the old binding of the parameter in the continuation, so it is not
     possible to eliminate the references to these ports in the
     continuation by using the `parameterize' form alone.

     Serialization of continuations can be achieved dependably by taking
     advantage of string-ports, which are serializable objects (unless
     there is a blocked thread), and the following features of threads:
     they inherit the dynamic environment of the parent thread and they
     start with an initial continuation that contains only serializable
     objects.  So a thread created in a dynamic environment where
     `current-input-port' and `current-output-port' are bound to a
     dummy string-port has a serializable continuation.

     Here is an example where continuations are serialized:

          > (define (wr obj)
              (call-with-output-string
               (lambda (p)
                 (output-port-readtable-set!
                  p
                  (readtable-sharing-allowed?-set
                   (output-port-readtable p)
                   'serialize))
                 (write obj p))))
          > (define (rd str)
              (call-with-input-string
               str
               (lambda (p)
                 (input-port-readtable-set!
                  p
                  (readtable-sharing-allowed?-set
                   (input-port-readtable p)
                   'serialize))
                 (read p))))
          > (define fifo (open-vector))
          > (define (suspend-and-die!)
              (call-with-current-continuation
               (lambda (k)
                 (write (wr k) fifo)
                 (newline fifo)
                 (force-output fifo)
                 (thread-terminate! (current-thread)))))
          > (let ((dummy-port (open-string)))
              (parameterize ((current-input-port dummy-port)
                             (current-output-port dummy-port))
                (thread-start!
                 (make-thread
                  (lambda ()
                    (* 100
                       (suspend-and-die!)))))))
          #<thread #2>
          > (define s (read fifo))
          > (thread-join!
              (thread-start!
                (make-thread
                  (lambda ()
                    ((rd s) 111)))))
          11100
          > (thread-join!
              (thread-start!
                (make-thread
                  (lambda ()
                    ((rd s) 222)))))
          22200
          > (string-length s)
          13114


 -- procedure: readtable-eval-allowed? READTABLE
 -- procedure: readtable-eval-allowed?-set READTABLE NEW-VALUE
     The procedure `readtable-eval-allowed?' returns the content of the
     `eval-allowed?' field of READTABLE.  The reader recognizes the
     `#.EXPRESSION' notation for read-time evaluation if and only if
     the content of the `eval-allowed?' field is not `#f'.

     The procedure `readtable-eval-allowed?-set' returns a copy of
     READTABLE where only the `eval-allowed?' field has been changed to
     NEW-VALUE.

     For example:

          > (input-port-readtable-set!
              (repl-input-port)
              (readtable-eval-allowed?-set
                (input-port-readtable (repl-input-port))
                #t))
          > '(5 plus 7 is #.(+ 5 7))
          (5 plus 7 is 12)
          > '(buf = #.(make-u8vector 25))
          (buf = #u8(0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0))


 -- procedure: readtable-write-cdr-read-macros? READTABLE
 -- procedure: readtable-write-cdr-read-macros?-set READTABLE NEW-VALUE
 -- procedure: readtable-write-extended-read-macros? READTABLE
 -- procedure: readtable-write-extended-read-macros?-set READTABLE
          NEW-VALUE
     The procedure `readtable-write-cdr-read-macros?' returns the
     content of the `write-cdr-read-macros?' field of READTABLE.  The
     procedure `readtable-write-extended-read-macros?' returns the
     content of the `write-extended-read-macros?' field of READTABLE.

     At all times the printer uses read-macros in its output for datums
     of the form `(quote DATUM)', `(quasiquote DATUM)', `(unquote
     DATUM)', and `(unquote-splicing DATUM)'.  That is the following
     read-macro notations will be used respectively: `'DATUM', ``DATUM',
     `,DATUM', and `,@DATUM'.  Moreover, normally the read-macros will
     not be used when the form appears in the cdr of a list, for
     example `(foo quote bar)', `(foo . (quote bar))' and `(foo .
     'bar)' will all be printed as `(foo quote bar)'.

     When the content of the `write-cdr-read-macros?' field is not
     `#f', the printer will use read-macros when the forms appear in
     the cdr of a list.  For example `(foo quote bar)' will be printed
     as `(foo . 'bar)'.  When the content of the
     `write-extended-read-macros?' field is not `#f', the printer will
     also use extended read-macros, for example `#'DATUM' in place of
     `(syntax DATUM)'.

     The procedure `readtable-write-cdr-read-macros?-set' returns a
     copy of READTABLE where only the `write-cdr-read-macros?' field
     has been changed to NEW-VALUE.  The procedure
     `readtable-write-extended-read-macros?-set' returns a copy of
     READTABLE where only the `write-extended-read-macros?' field has
     been changed to NEW-VALUE.

     For example:

          > (output-port-readtable-set!
              (repl-output-port)
              (readtable-write-extended-read-macros?-set
                (output-port-readtable (repl-output-port))
                #t))
          > '(foo (syntax bar))
          (foo #'bar)
          > '(foo syntax bar)
          (foo syntax bar)
          > (output-port-readtable-set!
              (repl-output-port)
              (readtable-write-cdr-read-macros?-set
                (output-port-readtable (repl-output-port))
                #t))
          > '(foo syntax bar)
          (foo . #'bar)


 -- procedure: readtable-max-write-level READTABLE
 -- procedure: readtable-max-write-level-set READTABLE NEW-VALUE
     The procedure `readtable-max-write-level' returns the content of
     the `max-write-level' field of READTABLE.  The printer will
     display an ellipsis for the elements of lists and vectors that are
     nested deeper than this level.

     The procedure `readtable-max-write-level-set' returns a copy of
     READTABLE where only the `max-write-level' field has been changed
     to NEW-VALUE, which must be an nonnegative fixnum.

     For example:

          > (define (wr obj n)
              (call-with-output-string
                (lambda (p)
                  (output-port-readtable-set!
                    p
                    (readtable-max-write-level-set
                      (output-port-readtable p)
                      n))
                  (write obj p))))
          > (wr '(a #(b (c c) #u8(9 9 9) b) a) 3)
          "(a #(b (c c) #u8(9 9 9) b) a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) a) 2)
          "(a #(b (...) #u8(...) b) a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) a) 1)
          "(a #(...) a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) a) 0)
          "(...)"
          > (wr 'hello 0)
          "hello"


 -- procedure: readtable-max-write-length READTABLE
 -- procedure: readtable-max-write-length-set READTABLE NEW-VALUE
     The procedure `readtable-max-write-length' returns the content of
     the `max-write-length' field of READTABLE.  The printer will
     display an ellipsis for the elements of lists and vectors that are
     at an index beyond that length.

     The procedure `readtable-max-write-length-set' returns a copy of
     READTABLE where only the `max-write-length' field has been changed
     to NEW-VALUE, which must be an nonnegative fixnum.

     For example:

          > (define (wr obj n)
              (call-with-output-string
                (lambda (p)
                  (output-port-readtable-set!
                    p
                    (readtable-max-write-length-set
                      (output-port-readtable p)
                      n))
                  (write obj p))))
          > (wr '(a #(b (c c) #u8(9 9 9) b) . a) 4)
          "(a #(b (c c) #u8(9 9 9) b) . a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) . a) 3)
          "(a #(b (c c) #u8(9 9 9) ...) . a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) . a) 2)
          "(a #(b (c c) ...) . a)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) . a) 1)
          "(a ...)"
          > (wr '(a #(b (c c) #u8(9 9 9) b) . a) 0)
          "(...)"


 -- procedure: readtable-max-unescaped-char READTABLE
 -- procedure: readtable-max-unescaped-char-set READTABLE NEW-VALUE
     The procedure `readtable-max-unescaped-char' returns the content
     of the `max-unescaped-char' field of READTABLE.  The printer will
     display using an escape sequence any character within symbols,
     strings and character objects greater than `max-unescaped-char'.
     When `max-unescaped-char' is `#f', the default value, the printer
     will take into account the output port and use an escape sequence
     for any character that isn't supported by the port's character
     encoding.

     The procedure `readtable-max-unescaped-char-set' returns a copy of
     READTABLE where only the `max-unescaped-char' field has been
     changed to NEW-VALUE, which must be a character or `#f'.

     For example:

          > (define rt (output-port-readtable (repl-output-port)))
          > (readtable-max-unescaped-char rt)
          #\delete
          > (string (integer->char 233))
          "\351"
          > (define (f c)
              (with-output-to-string
               (list readtable: (readtable-max-unescaped-char-set rt c))
               (lambda () (write (string (integer->char 233))))))
          > (f #\delete)
          "\"\\351\""
          > (string-length (f #\delete))
          6
          > (f #\U0010ffff)
          "\"\351\""
          > (string-length (f #\U0010ffff))
          3
          > (output-port-readtable-set!
             (repl-output-port)
             (readtable-max-unescaped-char-set rt #\U0010ffff))
          > (string (integer->char 233))
          "e'"


 -- procedure: readtable-start-syntax READTABLE
 -- procedure: readtable-start-syntax-set READTABLE NEW-VALUE
     The procedure `readtable-start-syntax' returns the content of the
     `start-syntax' field of READTABLE.  The reader uses this field to
     determine in which syntax to start parsing the input.  When the
     content of this field is the symbol `six', the reader starts in
     the infix syntax.  Otherwise the reader starts in the prefix
     syntax.

     The procedure `readtable-start-syntax-set' returns a copy of
     READTABLE where only the `start-syntax' field has been changed to
     NEW-VALUE.

     For example:

          > (+ 2 3)
          5
          > (input-port-readtable-set!
             (repl-input-port)
             (readtable-start-syntax-set
               (input-port-readtable (repl-input-port))
               'six))
          > 2+3;
          5
          > exit();


15.2 Boolean syntax
===================

Booleans are required to be followed by a delimiter (i.e. `#f64()' is
not the boolean `#f' followed by the number `64' and the empty list).

15.3 Character syntax
=====================

Characters are required to be followed by a delimiter (i.e.
`#\spaceballs' is not the character `#\space' followed by the symbol
`balls').  The lexical syntax of characters is extended to allow the
following:

`#\nul'
     Unicode character 0

`#\alarm'
     Unicode character 7

`#\backspace'
     Unicode character 8

`#\tab'
     Unicode character 9

`#\newline'
     Unicode character 10 (newline character)

`#\linefeed'
     Unicode character 10

`#\vtab'
     Unicode character 11

`#\page'
     Unicode character 12

`#\return'
     Unicode character 13

`#\esc'
     Unicode character 27

`#\space'
     Unicode character 32 (space character)

`#\delete'
     Unicode character 127

`#\xHH'
     character encoded in hexadecimal (>= 1 hexadecimal digit)

`#\uHHHH'
     character encoded in hexadecimal (exactly 4 hexadecimal digits)

`#\UHHHHHHHH'
     character encoded in hexadecimal (exactly 8 hexadecimal digits)

15.4 String syntax
==================

The lexical syntax of quoted strings is extended to allow the following
escape codes:

`\a'
     Unicode character 7

`\b'
     Unicode character 8

`\t'
     Unicode character 9

`\n'
     Unicode character 10 (newline character)

`\v'
     Unicode character 11

`\f'
     Unicode character 12

`\r'
     Unicode character 13

`\"'
     `"'

`\\'
     `\'

`\|'
     `|'

`\?'
     `?'

`\OOO'
     character encoded in octal (1 to 3 octal digits, first digit must
     be less than 4 when there are 3 octal digits)

`\xHH'
     character encoded in hexadecimal (>= 1 hexadecimal digit)

`\uHHHH'
     character encoded in hexadecimal (exactly 4 hexadecimal digits)

`\UHHHHHHHH'
     character encoded in hexadecimal (exactly 8 hexadecimal digits)

`\<space>'
     Unicode character 32 (space character)

`\<newline><whitespace-except-newline>*'
     This sequence expands to nothing (it is useful for splitting a long
     string literal on multiple lines while respecting proper
     indentation of the source code)

   Gambit also supports a "here string" syntax that is similar to shell
"here documents".  For example:

     > (pretty-print #<<THE-END
     hello
     world
     THE-END
     )
     "hello\nworld"

   The here string starts with the sequence `#<<'.  The part of the
line after the `#<<' up to and including the newline character is the
key. The first line afterward that matches the key marks the end of the
here string.  The string contains all the characters between the start
key and the end key, with the exception of the newline character before
the end key.

15.5 Symbol syntax
==================

The lexical syntax of symbols is extended to allow a leading and
trailing vertical bar (e.g. `|a\|b"c:|').  The symbol's name
corresponds verbatim to the characters between the vertical bars except
for escaped characters.  The same escape sequences as for strings are
permitted except that `"' does not need to be escaped and `|' needs to
be escaped.

   For example:

     > (symbol->string '|a\|b"c:|)
     "a|b\"c:"

15.6 Keyword syntax
===================

The lexical syntax of keywords is like symbols, but with a colon at the
end (note that this can be changed to a leading colon by setting the
`keywords-allowed?' field of the readtable to the symbol `prefix').  A
colon by itself is not a keyword, it is a symbol.  Vertical bars can be
used like symbols but the colon must be outside the vertical bars.
Note that the string returned by the `keyword->string' procedure does
not include the colon.

   For example:

     > (keyword->string 'foo:)
     "foo"
     > (map keyword? '(|ab()cd:| |ab()cd|: : ||:))
     (#f #t #f #t)

15.7 Box syntax
===============

The lexical syntax of boxes is `#&OBJ' where OBJ is the content of the
box.

   For example:

     > (list '#&"hello" '#&123)
     (#&"hello" #&123)
     > (box (box (+ 10 20)))
     #&#&30

15.8 Number syntax
==================

The lexical syntax of the special inexact real numbers is as follows:

`+inf.0'
     positive infinity

`-inf.0'
     negative infinity

`+nan.0'
     "not a number"

`-0.'
     negative zero (`0.' is the positive zero)

15.9 Homogeneous vector syntax
==============================

Homogeneous vectors are vectors containing raw numbers of the same type
(signed or unsigned exact integers or inexact reals).  There are 10
types of homogeneous vectors: `s8vector' (vector of 8 bit signed
integers), `u8vector' (vector of 8 bit unsigned integers), `s16vector'
(vector of 16 bit signed integers), `u16vector' (vector of 16 bit
unsigned integers), `s32vector' (vector of 32 bit signed integers),
`u32vector' (vector of 32 bit unsigned integers), `s64vector' (vector
of 64 bit signed integers), `u64vector' (vector of 64 bit unsigned
integers), `f32vector' (vector of 32 bit floating point numbers), and
`f64vector' (vector of 64 bit floating point numbers).

   The external representation of homogeneous vectors is similar to
normal vectors but with the `#(' prefix replaced respectively with
`#s8(', `#u8(', `#s16(', `#u16(', `#s32(', `#u32(', `#s64(', `#u64(',
`#f32(', and `#f64('.

   The elements of the integer homogeneous vectors must be exact
integers fitting in the given precision.  The elements of the floating
point homogeneous vectors must be inexact reals.

15.10 Special `#!' syntax
=========================

The lexical syntax of the special `#!' objects is as follows:

`#!eof'
     end-of-file object

`#!void'
     void object

`#!optional'
     optional object

`#!rest'
     rest object

`#!key'
     key object

15.11 Multiline comment syntax
==============================

Multiline comments are delimited by the tokens `#|' and `|#'.  These
comments can be nested.

15.12 Scheme infix syntax extension
===================================

The reader supports an infix syntax extension which is called SIX
(Scheme Infix eXtension).  This extension is both supported by the
`read' procedure and in program source code.

   The backslash character is a delimiter that marks the beginning of a
single datum expressed in the infix syntax (the details are given
below).  One way to think about it is that the backslash character
escapes the prefix syntax temporarily to use the infix syntax.  For
example a three element list could be written as `(X \Y Z)', the
elements X and Z are expressed using the normal prefix syntax and Y is
expressed using the infix syntax.

   When the reader encounters an infix datum, it constructs a syntax
tree for that particular datum.  Each node of this tree is represented
with a list whose first element is a symbol indicating the type of node.
For example, `(six.identifier abc)' is the representation of the infix
identifier `abc' and `(six.index (six.identifier abc) (six.identifier
i))' is the representation of the infix datum `abc[i];'.  The reader
will return this representation wrapped with a `(six.infix ...)' form.

15.12.1 SIX grammar
-------------------

The SIX grammar is given below.  On the left hand side are the
production rules.  On the right hand side is the datum that is
constructed by the reader.  The notation $I denotes the datum that is
constructed by the reader for the Ith part of the production rule.

   In this grammar most statements end with a semicolon.  When the
<infix datum> is immediately following the backslash character that
indicates the start of an infix datum, the ending semicolon is optional
(a semicolon is automatically inserted when the datum could be complete
and a whitespace or inappropriate character is encountered such as a
closing parenthesis).  For example `(f \2*n (list) \5)' is equivalent
to `(f \2*n; (list) \5;)'.

<infix datum> ::=                           
<stat>                                      $1
<stat> ::=                                  
<if stat>                                   $1
| <for stat>                                $1
| <while stat>                              $1
| <do stat>                                 $1
| <switch stat>                             $1
| <case stat>                               $1
| <break stat>                              $1
| <continue stat>                           $1
| <label stat>                              $1
| <goto stat>                               $1
| <return stat>                             $1
| <import stat>                             $1
| <from stat>                               $1
| <expression stat>                         $1
| <procedure definition>                    $1
| <variable definition> `;'                 $1
| <clause stat>                             $1
| <compound stat>                           $1
| `;'                                       `(six.compound)'
<if stat> ::=                               
`if' `(' <pexpr> `)' <stat>                 `(six.if $3 $5)'
| `if' `(' <pexpr> `)' <stat> `else'        `(six.if $3 $5 $7)'
<stat>                                      
<for stat> ::=                              
`for' `(' <stat> <oexpr> `;' <oexpr> `)'    `(six.for $3 $4 $6 $8)'
<stat>                                      
<while stat> ::=                            
`while' `(' <pexpr> `)' <stat>              `(six.while $3 $5)'
<do stat> ::=                               
`do' <stat> `while' `(' <pexpr> `)' `;'     `(six.do-while $2 $5)'
<switch stat> ::=                           
`switch' `(' <pexpr> `)' <stat>             `(six.switch $3 $5)'
<case stat> ::=                             
`case' <expr> `:' <stat>                    `(six.case $2 $4)'
<break stat> ::=                            
`break' `;'                                 `(six.break)'
<continue stat> ::=                         
`continue' `;'                              `(six.continue)'
<label stat> ::=                            
<identifier> `:' <stat>                     `(six.label $1 $3)'
<goto stat> ::=                             
`goto' <expr> `;'                           `(six.goto $2)'
<return stat> ::=                           
`return' `;'                                `(six.return)'
| `return' <expr> `;'                       `(six.return $2)'
<import stat> ::=                           
`import' <expr> `;'                         `(six.import $2)'
<from stat> ::=                             
`from' <expr> `import' <expr> `;'           `(six.from-import $2 $4)'
`from' <expr> `import' `*' `;'              `(six.from-import-* $2)'
<expression stat> ::=                       
<expr> `;'                                  $1
<clause stat> ::=                           
<expr> `.'                                  `(six.clause $1)'
<pexpr> ::=                                 
<procedure definition>                      $1
| <variable definition>                     $1
| <expr>                                    $1
<procedure definition> ::=                  
<type> <id-or-prefix> `(' <parameters> `)'  `(six.define-procedure $2
<body>                                      (six.procedure $1 $4 $6))'
| `function' <id-or-prefix> `('             `(six.define-procedure $2
<parameters> `)' <body>                     (six.procedure #f $4 $6))'
<variable definition> ::=                   
<type> <id-or-prefix> <dimensions> <iexpr>  `(six.define-variable $2 $1
                                            $3 $4)'
<iexpr> ::=                                 
`=' <expr>                                  $2
|                                           `#f'
<dimensions> ::=                            
| `[' <expr> `]' <dimensions>               `($2 . $4)'
|                                           `()'
<oexpr> ::=                                 
<expr>                                      $1
|                                           `#f'
<expr> ::=                                  
<expr22>                                    $1
<expr22> ::=                                
<expr21> `:-' <expr22>                      `(six.x:-y $1 $3)'
| <expr21>                                  $1
<expr21> ::=                                
<expr21> `,' <expr20>                       `(|six.x,y| $1 $3)'
| <expr20>                                  $1
<expr20> ::=                                
`yield' <expr20>                            `(six.yieldx $2)'
| <expr19>                                  $1
<expr19> ::=                                
<expr18> `:=' <expr19>                      `(six.x:=y $1 $3)'
| <expr18>                                  $1
<expr18> ::=                                
<expr17> `%=' <expr18>                      `(six.x%=y $1 $3)'
| <expr17> `&=' <expr18>                    `(six.x&=y $1 $3)'
| <expr17> `**=' <expr18>                   `(six.x**=y $1 $3)'
| <expr17> `*=' <expr18>                    `(six.x*=y $1 $3)'
| <expr17> `@=' <expr18>                    `(six.x@=y $1 $3)'
| <expr17> `+=' <expr18>                    `(six.x+=y $1 $3)'
| <expr17> `-=' <expr18>                    `(six.x-=y $1 $3)'
| <expr17> `//=' <expr18>                   `(six.x//=y $1 $3)'
| <expr17> `/=' <expr18>                    `(six.x/=y $1 $3)'
| <expr17> `<<=' <expr18>                   `(six.x<<=y $1 $3)'
| <expr17> `=' <expr18>                     `(six.x=y $1 $3)'
| <expr17> `>>>=' <expr18>                  `(six.x>>>=y $1 $3)'
| <expr17> `>>=' <expr18>                   `(six.x>>=y $1 $3)'
| <expr17> `^=' <expr18>                    `(six.x^=y $1 $3)'
| <expr17> `|=' <expr18>                    `(|six.x\|=y| $1 $3)'
| <expr17>                                  $1
<expr17> ::=                                
<expr16> `:' <expr17>                       `(six.x:y $1 $3)'
| <expr16>                                  $1
<expr16> ::=                                
<expr15> `?' <expr> `:' <expr16>            `(six.x?y:z $1 $3 $5)'
| <expr15>                                  $1
<expr15> ::=                                
<expr15> `or' <expr14>                      `(six.xory $1 $3)'
| <expr14>                                  $1
<expr14> ::=                                
<expr14> `and' <expr13>                     `(six.xandy $1 $3)'
| <expr13>                                  $1
<expr13> ::=                                
`not' <expr13>                              `(six.notx $1)'
| <expr12>                                  $1
<expr12> ::=                                
<expr12> `||' <expr11>                      `(|six.x\|\|y| $1 $3)'
| <expr11>                                  $1
<expr11> ::=                                
<expr11> `&&' <expr10>                      `(six.x&&y $1 $3)'
| <expr10>                                  $1
<expr10> ::=                                
<expr10> `|' <expr9>                        `(|six.x\|y| $1 $3)'
| <expr9>                                   $1
<expr9> ::=                                 
<expr9> `^' <expr8>                         `(six.x^y $1 $3)'
| <expr8>                                   $1
<expr8> ::=                                 
<expr8> `&' <expr7>                         `(six.x&y $1 $3)'
| <expr7>                                   $1
<expr7> ::=                                 
<expr7> `!=' <expr6>                        `(six.x!=y $1 $3)'
| <expr7> `==' <expr6>                      `(six.x==y $1 $3)'
| <expr7> `!==' <expr6>                     `(six.x!==y $1 $3)'
| <expr7> `===' <expr6>                     `(six.x===y $1 $3)'
| <expr6>                                   $1
<expr6> ::=                                 
<expr6> `<' <expr5>                         `(six.x<y $1 $3)'
| <expr6> `<=' <expr5>                      `(six.x<=y $1 $3)'
| <expr6> `>' <expr5>                       `(six.x>y $1 $3)'
| <expr6> `>=' <expr5>                      `(six.x>=y $1 $3)'
| <expr6> `in' <expr5>                      `(six.xiny $1 $3)'
| <expr6> `is' <expr5>                      `(six.xisy $1 $3)'
| <expr6> `instanceof' <expr5>              `(six.xinstanceofy $1 $3)'
| <expr5>                                   $1
<expr5> ::=                                 
<expr5> `<<' <expr4>                        `(six.x<<y $1 $3)'
| <expr5> `>>' <expr4>                      `(six.x>>y $1 $3)'
| <expr5> `>>>' <expr4>                     `(six.x>>>y $1 $3)'
| <expr4>                                   $1
<expr4> ::=                                 
<expr4> `+' <expr3>                         `(six.x+y $1 $3)'
| <expr4> `-' <expr3>                       `(six.x-y $1 $3)'
| <expr3>                                   $1
<expr3> ::=                                 
<expr3> `%' <expr2>                         `(six.x%y $1 $3)'
| <expr3> `*' <expr2>                       `(six.x*y $1 $3)'
| <expr3> `@' <expr2>                       `(six.x@y $1 $3)'
| <expr3> `/' <expr2>                       `(six.x/y $1 $3)'
| <expr3> `//' <expr2>                      `(six.x//y $1 $3)'
| <expr2>                                   $1
<expr2> ::=                                 
`&' <expr2>                                 `(six.&x $2)'
| `+' <expr2>                               `(six.+x $2)'
| `-' <expr2>                               `(six.-x $2)'
| `*' <expr2>                               `(six.*x $2)'
| `**' <expr2>                              `(six.**x $2)'
| `!' <expr2>                               `(six.!x $2)'
| `!'                                       `(six.!)'
| `++' <expr2>                              `(six.++x $2)'
| `--' <expr2>                              `(six.--x $2)'
| `~' <expr2>                               `(six.~x $2)'
| <expr2> `**' <expr1>                      `(six.x**y $1 $3)'
| `await' <expr2>                           `(six.awaitx $2)'
| `typeof' <expr2>                          `(six.typeofx $2)'
| <expr1>                                   $1
<expr1> ::=                                 
<expr1> `++'                                `(six.x++ $1)'
| <expr1> `--'                              `(six.x-- $1)'
| <expr1> `(' <arguments> `)'               `(six.call $1 . $3)'
| <expr1> `[' <expr> `]'                    `(six.index $1 $3)'
| <expr1> `->' <id-or-prefix>               `(six.arrow $1 $3)'
| <expr1> `.' <id-or-prefix>                `(six.dot $1 $3)'
| <expr0>                                   $1
<expr0> ::=                                 
<id-or-prefix>                              $1
| <string>                                  `(six.literal $1)'
| <char>                                    `(six.literal $1)'
| <number>                                  `(six.literal $1)'
| `(' <expr> `)'                            $2
| `(' <block stat> `)'                      $2
| <datum-starting-with-#-or-backquote>      $1
| `[' <elements> `]'                        $2
| <type> `(' <parameters> `)' <body>        `(six.procedure $1 $3 $5)'
| `function' `(' <parameters> `)' <body>    `(six.procedure #f $3 $5)'
| `new' <id-or-prefix> `(' <arguments> `)'  `(six.new $2 . $4)'
| `async' <expr0>                           `(six.asyncx $2)'
<block stat> ::=                            
`{' <stat list> `}'                         `(six.compound . $2)'
<body> ::=                                  
`{' <stat list> `}'                         `(six.procedure-body . $2)'
<stat list> ::=                             
<stat> <stat list>                          `($1 . $2)'
|                                           `()'
<parameters> ::=                            
<nonempty parameters>                       $1
|                                           `()'
<nonempty parameters> ::=                   
<parameter> `,' <nonempty parameters>       `($1 . $3)'
| <parameter>                               `($1)'
<parameter> ::=                             
<type> <id-or-prefix>                       `($2 $1)'
<arguments> ::=                             
<nonempty arguments>                        $1
|                                           `()'
<nonempty arguments> ::=                    
<expr> `,' <nonempty arguments>             `($1 . $3)'
| <expr>                                    `($1)'
<elements> ::=                              
<nonempty elements>                         $1
|                                           `(six.null)'
<nonempty elements> ::=                     
<expr>                                      `(six.list $1 (six.null))'
| <expr> `,' <nonempty elements>            `(six.list $1 $3)'
| <expr> `|' <expr>                         `(six.cons $1 $3)'
<id-or-prefix> ::=                          
<identifier>                                `(six.identifier $1)'
| `\' <datum>                               $2
<type> ::=                                  
`scmobj'                                    `scmobj'

15.12.2 SIX semantics
---------------------

The semantics of SIX depends on the definition of the `six.XXX'
identifiers (as procedures and macros).  Many of these identifiers are
predefined macros which give SIX a semantics that is close to C's while
also borrowing some semantics from JavaScript and Python (such as `==='
`in', `**', `//').  The programmer may override these definitions to
change the semantics either globally or locally.  For example,
`six.x^y' is a predefined macro that expands `(six.x^y x y)' into
`(bitwise-xor x y)'.  If the programmer prefers the `^' operator to
express exponentiation in a specific procedure, then in that procedure
`six.x^y' can be redefined as a macro that expands `(six.x^y x y)' into
`(expt x y)'.  Note that the associativity and precedence of operators
cannot be changed as that is a syntactic issue, so this will give a
left associative exponentiation operator with an unusual precedence.

   The following identifiers do not have a predefined semantics (they
are undefined identifiers): `six.label', `six.goto', `six.switch',
`six.case', `six.break', `six.continue', `six.return', `six.clause',
`six.x:-y', `six.@', `six.@=', `six.import', `six.from-import',
`six.from-import-*', and `six.!'.

   Here is an example showing some of the predefined syntax and
semantics of SIX:

     > (list (+ 1 2) \3+4 (+ 5 6))
     (3 7 11)
     > \[ 1+2, \(+ 3 4), 5+6 ]
     (3 7 11)
     > (map (lambda (x) \(x*x-1)/log(x+1)) '(1 2 3 4))
     (0 2.730717679880512 5.7707801635558535 9.320024018394177)
     > (map \function (x) { return (x*x-1)/log(x+1); } '(1 2 3 4))
     (0. 2.730717679880512 5.7707801635558535 9.320024018394177)
     > \map(function (x) { return (x*x-1)/log(x+1); }, [1, 2, 3, 4])
     (0. 2.730717679880512 5.7707801635558535 9.320024018394177)
     > \scmobj n=expt(10,5)
     > n
     100000
     > \scmobj t[3][10]=88
     > \t[0][9]=t[2][1]=11
     11
     > t
     #(#(88 88 88 88 88 88 88 88 88 11)
       #(88 88 88 88 88 88 88 88 88 88)
       #(88 11 88 88 88 88 88 88 88 88))
     > \scmobj radix=new parameter(10)
     > \radix(2)
     > \radix()
     2
     > \for (scmobj i=0; i<5; i++) pp(1<<i*8)
     1
     256
     65536
     16777216
     4294967296
     > \function \make-adder(x) { return function (y) { x+y; }; }
     > \map(new adder(100), [1,2,3,4])
     (101 102 103 104)
     > (map (make-adder 100) (list 1 2 3 4))
     (101 102 103 104)

16 C-interface
**************

The Gambit Scheme system offers a mechanism for interfacing Scheme code
and C code called the "C-interface".  A Scheme program indicates which
C functions it needs to have access to and which Scheme procedures can
be called from C, and the C interface automatically constructs the
corresponding Scheme procedures and C functions.  The conversions needed
to transform data from the Scheme representation to the C representation
(and back), are generated automatically in accordance with the argument
and result types of the C function or Scheme procedure.

   The C-interface places some restrictions on the types of data that
can be exchanged between C and Scheme.  The mapping of data types
between C and Scheme is discussed in the next section.  The remaining
sections of this chapter describe each special form of the C-interface.

16.1 The mapping of types between C and Scheme
==============================================

Scheme and C do not provide the same set of built-in data types so it is
important to understand which Scheme type is compatible with which C
type and how values get mapped from one environment to the other.  To
improve compatibility a new type is added to Scheme, the `foreign'
object type, and the following data types are added to C:

`scheme-object'
     denotes the universal type of Scheme objects (type `___SCMOBJ'
     defined in `gambit.h')

`bool'
     denotes the C++ `bool' type or the C `int' type (type `___BOOL'
     defined in `gambit.h')

`int8'
     8 bit signed integer (type `___S8' defined in `gambit.h')

`unsigned-int8'
     8 bit unsigned integer (type `___U8' defined in `gambit.h')

`int16'
     16 bit signed integer (type `___S16' defined in `gambit.h')

`unsigned-int16'
     16 bit unsigned integer (type `___U16' defined in `gambit.h')

`int32'
     32 bit signed integer (type `___S32' defined in `gambit.h')

`unsigned-int32'
     32 bit unsigned integer (type `___U32' defined in `gambit.h')

`int64'
     64 bit signed integer (type `___S64' defined in `gambit.h')

`unsigned-int64'
     64 bit unsigned integer (type `___U64' defined in `gambit.h')

`float32'
     32 bit floating point number (type `___F32' defined in `gambit.h')

`float64'
     64 bit floating point number (type `___F64' defined in `gambit.h')

`ISO-8859-1'
     denotes ISO-8859-1 encoded characters (8 bit unsigned integer,
     type `___ISO_8859_1' defined in `gambit.h')

`UCS-2'
     denotes UCS-2 encoded characters (16 bit unsigned integer, type
     `___UCS_2' defined in `gambit.h')

`UCS-4'
     denotes UCS-4 encoded characters (32 bit unsigned integer, type
     `___UCS_4' defined in `gambit.h')

`char-string'
     denotes the C `char*' type when used as a null terminated string

`nonnull-char-string'
     denotes the nonnull C `char*' type when used as a null terminated
     string

`nonnull-char-string-list'
     denotes an array of nonnull C `char*' terminated with a null
     pointer

`ISO-8859-1-string'
     denotes ISO-8859-1 encoded strings (null terminated string of 8
     bit unsigned integers, i.e. `___ISO_8859_1*')

`nonnull-ISO-8859-1-string'
     denotes nonnull ISO-8859-1 encoded strings (null terminated string
     of 8 bit unsigned integers, i.e. `___ISO_8859_1*')

`nonnull-ISO-8859-1-stringlist'
     denotes an array of nonnull ISO-8859-1 encoded strings terminated
     with a null pointer

`UTF-8-string'
     denotes UTF-8 encoded strings (null terminated string of `char',
     i.e. `char*')

`nonnull-UTF-8-string'
     denotes nonnull UTF-8 encoded strings (null terminated string of
     `char', i.e. `char*')

`nonnull-UTF-8-string-list'
     denotes an array of nonnull UTF-8 encoded strings terminated with
     a null pointer

`UTF-16-string'
     denotes UTF-16 encoded strings (null terminated string of `char',
     i.e. `char*')

`nonnull-UTF-16-string'
     denotes nonnull UTF-16 encoded strings (null terminated string of
     `char', i.e. `char*')

`nonnull-UTF-16-string-list'
     denotes an array of nonnull UTF-16 encoded strings terminated with
     a null pointer

`UCS-2-string'
     denotes UCS-2 encoded strings (null terminated string of 16 bit
     unsigned integers, i.e. `___UCS_2*')

`nonnull-UCS-2-string'
     denotes nonnull UCS-2 encoded strings (null terminated string of
     16 bit unsigned integers, i.e. `___UCS_2*')

`nonnull-UCS-2-string-list'
     denotes an array of nonnull UCS-2 encoded strings terminated with
     a null pointer

`UCS-4-string'
     denotes UCS-4 encoded strings (null terminated string of 32 bit
     unsigned integers, i.e. `___UCS_4*')

`nonnull-UCS-4-string'
     denotes nonnull UCS-4 encoded strings (null terminated string of
     32 bit unsigned integers, i.e. `___UCS_4*')

`nonnull-UCS-4-string-list'
     denotes an array of nonnull UCS-4 encoded strings terminated with
     a null pointer

`wchar_t-string'
     denotes `wchar_t' encoded strings (null terminated string of
     `wchar_t', i.e. `wchar_t*')

`nonnull-wchar_t-string'
     denotes nonnull `wchar_t' encoded strings (null terminated string
     of `wchar_t', i.e. `wchar_t*')

`nonnull-wchar_t-string-list'
     denotes an array of nonnull `wchar_t' encoded strings terminated
     with a null pointer

   To specify a particular C type inside the `c-lambda', `c-define' and
`c-define-type' forms, the following "Scheme notation" is used:

`Scheme notation'
     C type

`void'
     `void'

`bool'
     `bool'

`char'
     `char'  (may be signed or unsigned depending on the C compiler)

`signed-char'
     `signed char'

`unsigned-char'
     `unsigned char'

`ISO-8859-1'
     `ISO-8859-1'

`UCS-2'
     `UCS-2'

`UCS-4'
     `UCS-4'

`wchar_t'
     `wchar_t'

`size_t'
     `size_t' (type `___SIZE_T' defined in `gambit.h')

`ssize_t'
     `ssize_t' (type `___SSIZE_T' defined in `gambit.h')

`ptrdiff_t'
     `ptrdiff_t' (type `___PTRDIFF_T' defined in `gambit.h')

`short'
     `short'

`unsigned-short'
     `unsigned short'

`int'
     `int'

`unsigned-int'
     `unsigned int'

`long'
     `long'

`unsigned-long'
     `unsigned long'

`long-long'
     `long long'

`unsigned-long-long'
     `unsigned long long'

`float'
     `float'

`double'
     `double'

`int8'
     `int8'

`unsigned-int8'
     `unsigned-int8'

`int16'
     `int16'

`unsigned-int16'
     `unsigned-int16'

`int32'
     `int32'

`unsigned-int32'
     `unsigned-int32'

`int64'
     `int64'

`unsigned-int64'
     `unsigned-int64'

`float32'
     `float32'

`float64'
     `float64'

`(struct "C-STRUCT-ID" [TAGS [RELEASE-FUNCTION]])'
     `struct C-STRUCT-ID'  (where C-STRUCT-ID is the name of a C
     structure; see below for the meaning of TAGS and RELEASE-FUNCTION)

`(union "C-UNION-ID" [TAGS [RELEASE-FUNCTION]])'
     `union C-UNION-ID'  (where C-UNION-ID is the name of a C union;
     see below for the meaning of TAGS and RELEASE-FUNCTION)

`(type "C-TYPE-ID" [TAGS [RELEASE-FUNCTION]])'
     `C-TYPE-ID'  (where C-TYPE-ID is an identifier naming a C type;
     see below for the meaning of TAGS and RELEASE-FUNCTION)

`(pointer TYPE [TAGS [RELEASE-FUNCTION]])'
     `T*'  (where T is the C equivalent of TYPE which must be the
     Scheme notation of a C type; see below for the meaning of TAGS and
     RELEASE-FUNCTION)

`(nonnull-pointer TYPE [TAGS [RELEASE-FUNCTION]])'
     same as `(pointer TYPE [TAGS [RELEASE-FUNCTION]])' except the
     `NULL' pointer is not allowed

`(function (TYPE1...) RESULT-TYPE)'
     function with the given argument types and result type

`(nonnull-function (TYPE1...) RESULT-TYPE)'
     same as `(function (TYPE1...) RESULT-TYPE)' except the `NULL'
     pointer is not allowed

`char-string'
     `char-string'

`nonnull-char-string'
     `nonnull-char-string'

`nonnull-char-string-list'
     `nonnull-char-string-list'

`ISO-8859-1-string'
     `ISO-8859-1-string'

`nonnull-ISO-8859-1-string'
     `nonnull-ISO-8859-1-string'

`nonnull-ISO-8859-1-string-list'
     `nonnull-ISO-8859-1-string-list'

`UTF-8-string'
     `UTF-8-string'

`nonnull-UTF-8-string'
     `nonnull-UTF-8-string'

`nonnull-UTF-8-string-list'
     `nonnull-UTF-8-string-list'

`UTF-16-string'
     `UTF-16-string'

`nonnull-UTF-16-string'
     `nonnull-UTF-16-string'

`nonnull-UTF-16-string-list'
     `nonnull-UTF-16-string-list'

`UCS-2-string'
     `UCS-2-string'

`nonnull-UCS-2-string'
     `nonnull-UCS-2-string'

`nonnull-UCS-2-string-list'
     `nonnull-UCS-2-string-list'

`UCS-4-string'
     `UCS-4-string'

`nonnull-UCS-4-string'
     `nonnull-UCS-4-string'

`nonnull-UCS-4-string-list'
     `nonnull-UCS-4-string-list'

`wchar_t-string'
     `wchar_t-string'

`nonnull-wchar_t-string'
     `nonnull-wchar_t-string'

`nonnull-wchar_t-string-list'
     `nonnull-wchar_t-string-list'

`scheme-object'
     `scheme-object'

`NAME'
     appropriate translation of NAME (where NAME is a C type defined
     with `c-define-type')

`"C-TYPE-ID"'
     C-TYPE-ID (this form is equivalent to `(type "C-TYPE-ID")')

   The `struct', `union', `type', `pointer' and `nonnull-pointer' types
are "foreign types" and they are represented on the Scheme side as
"foreign objects".  A foreign object is internally represented as a
pointer.  This internal pointer is identical to the C pointer being
represented in the case of the `pointer' and `nonnull-pointer' types.

   In the case of the `struct', `union' and `type' types, the internal
pointer points to a copy of the C data type being represented.  When an
instance of one of these types is converted from C to Scheme, a block
of memory is allocated from the C heap and initialized with the
instance and then a foreign object is allocated from the Scheme heap
and initialized with the pointer to this copy.  This approach may
appear overly complex, but it allows the conversion of C++ classes that
do not have a zero parameter constructor or an assignment method (i.e.
when compiling with a C++ compiler an instance is copied using `new
TYPE (INSTANCE)', which calls the copy-constructor of TYPE if it is a
class; TYPE's assignment operator is never used).  Conversion from
Scheme to C simply dereferences the internal pointer (no allocation
from the C heap is performed).  Deallocation of the copy on the C heap
is under the control of the release function attached to the foreign
object (see below).

   The optional TAGS field of foreign type specifications is used for
type checking on the Scheme side.  The TAGS field must be `#f', a
symbol or a non-empty list of symbols.  When it is not specified the
TAGS field defaults to a symbol whose name, as returned by
`symbol->string', is the C type declaration for that type.  For example
the symbol `char**' is the default for the type `(pointer (pointer
char))'.  A TAGS field that is a single symbol is equivalent to a list
containing only that symbol.  The first symbol in the list of tags is
the primary tag.  For example the primary tag of the type `(pointer
char)' is `char*' and the primary tag of the type `(pointer char (foo
bar))' is `foo'.

   Type compatibility between two foreign types depends on their tags.
An instance of a foreign type T can be used where a foreign type E is
expected if and only if

   * T's TAGS field is `#f', or

   * E's TAGS field is `#f', or

   * T's primary tag is a member of E's tags.


   For the safest code a TAGS field of `#f' should be used sparingly,
as it completely bypasses type checking.  The external representation
of Scheme foreign objects (used by the `write' procedure) contains the
primary tag (if the TAGS field is not `#f'), and the hexadecimal
address denoted by the internal pointer, for example `#<char** #2
0x2AAC535C>'.  Note that the hexadecimal address is in C notation,
which can be easily transferred to a C debugger with a "cut-and-paste".

   A RELEASE-FUNCTION can also be specified within a foreign type
specification.  The RELEASE-FUNCTION must be `#f' or a string naming a
C function with a single parameter of type `void*' (in which the
internal pointer is passed) and with a result of type `___SCMOBJ' (for
returning an error code).  When the RELEASE-FUNCTION is not specified
or is `#f' a default function is constructed by the C-interface.  This
default function does nothing in the case of the `pointer' and
`nonnull-pointer' types (deallocation is not the responsibility of the
C-interface) and returns the fixnum `___FIX(___NO_ERR)' to indicate no
error.  In the case of the `struct', `union' and `type' types, the
default function reclaims the copy on the C heap referenced by the
internal pointer (when using a C++ compiler this is done using `delete
(TYPE*)INTERNAL-POINTER', which calls the destructor of TYPE if it is a
class) and returns `___FIX(___NO_ERR)'.  In many situations the default
RELEASE-FUNCTION will perform the appropriate cleanup for the foreign
type.  However, in certain cases special operations (such as
decrementing a reference count, removing the object from a table, etc)
must be performed.  For such cases a user supplied RELEASE-FUNCTION is
needed.

   The RELEASE-FUNCTION is invoked at most once for any foreign object.
After the RELEASE-FUNCTION is invoked, the foreign object is
considered "released" and can no longer be used in a foreign type
conversion.  When the garbage collector detects that a foreign object
is no longer reachable by the program, it will invoke the
RELEASE-FUNCTION if the foreign object is not yet released.  When there
is a need to release the foreign object promptly, the program can
explicitly call `(foreign-release! OBJ)' which invokes the
RELEASE-FUNCTION if the foreign object is not yet released, and does
nothing otherwise.  The call `(foreign-released? OBJ)' returns a
boolean indicating whether the foreign object OBJ has been released yet
or not.  The call `(foreign-address OBJ)' returns the address denoted
by the internal pointer of foreign object OBJ or 0 if it has been
released.  The call `(foreign? OBJ)' tests that OBJ is a foreign
object.  Finally the call `(foreign-tags OBJ)' returns the list of tags
of foreign object OBJ, or `#f'.

   The following table gives the C types to which each Scheme type can
be converted:

Scheme type
     Allowed target C types

boolean `#f'
     `scheme-object'; `bool'; `pointer'; `function'; `char-string';
     `ISO-8859-1-string'; `UTF-8-string'; `UTF-16-string';
     `UCS-2-string'; `UCS-4-string'; `wchar_t-string'

boolean `#t'
     `scheme-object'; `bool'

character
     `scheme-object'; `bool'; [`[un]signed'] `char'; `ISO-8859-1';
     `UCS-2'; `UCS-4'; `wchar_t'

exact integer
     `scheme-object'; `bool'; [`unsigned-']
     `int8'/`int16'/`int32'/`int64'; [`unsigned'] `short'/`int'/`long';
     `size_t'/`ssize_t'/`ptrdiff_t'

inexact real
     `scheme-object'; `bool'; `float'; `double'; `float32'; `float64'

string
     `scheme-object'; `bool'; `char-string'; `nonnull-char-string';
     `ISO-8859-1-string'; `nonnull-ISO-8859-1-string'; `UTF-8-string';
     `nonnull-UTF-8-string'; `UTF-16-string'; `nonnull-UTF-16-string';
     `UCS-2-string'; `nonnull-UCS-2-string'; `UCS-4-string';
     `nonnull-UCS-4-string'; `wchar_t-string'; `nonnull-wchar_t-string'

foreign object
     `scheme-object'; `bool';
     `struct'/`union'/`type'/`pointer'/`nonnull-pointer' with the
     appropriate tags

vector
     `scheme-object'; `bool'

symbol
     `scheme-object'; `bool'

procedure
     `scheme-object'; `bool'; `function'; `nonnull-function'

other objects
     `scheme-object'; `bool'

   The following table gives the Scheme types to which each C type will
be converted:

C type
     Resulting Scheme type

scheme-object
     the Scheme object encoded

bool
     boolean

[`[un]signed'] `char'; `ISO-8859-1'; `UCS-2'; `UCS-4'; `wchar_t'
     character

[`unsigned-'] `int8'/`int16'/`int32'/`int64'; [`unsigned'] `short'/`int'/`long'; `size_t'/`ssize_t'/`ptrdiff_t'
     exact integer

`float'; `double'; `float32'; `float64'
     inexact real

`char-string'; `ISO-8859-1-string'; `UTF-8-string'; `UTF-16-string'; `UCS-2-string'; `UCS-4-string'; `wchar_t-string'
     string or `#f' if it is equal to `NULL'

`nonnull-char-string'; `nonnull-ISO-8859-1-string'; `nonnull-UTF-8-string'; `nonnull-UTF-16-string'; `nonnull-UCS-2-string'; `nonnull-UCS-4-string'; `nonnull-wchar_t-string'
     string

`struct'/`union'/`type'/`pointer'/`nonnull-pointer'
     foreign object with the appropriate tags or `#f' in the case of a
     `pointer' equal to `NULL'

`function'
     procedure or `#f' if it is equal to `NULL'

`nonnull-function'
     procedure

`void'
     void object

   All Scheme types are compatible with the C types `scheme-object' and
`bool'.  Conversion to and from the C type `scheme-object' is the
identity function on the object encoding.  This provides a low-level
mechanism for accessing Scheme's object representation from C (with the
help of the macros in the `gambit.h' header file).  When a C `bool'
type is expected, an extended Scheme boolean can be passed (`#f' is
converted to 0 and all other values are converted to 1).

   The Scheme boolean `#f' can be passed to the C environment where a
`char-string', `ISO-8859-1-string', `UTF-8-string', `UTF-16-string',
`UCS-2-string', `UCS-4-string', `wchar_t-string', `pointer' or
`function' type is expected.  In this case, `#f' is converted to the
`NULL' pointer.  C `bool's are extended booleans so any value different
from 0 represents true.  Thus, a C `bool' passed to the Scheme
environment is mapped as follows: 0 to `#f' and all other values to
`#t'.

   A Scheme character passed to the C environment where any C character
type is expected is converted to the corresponding character in the C
environment.  An error is signaled if the Scheme character does not fit
in the C character.  Any C character type passed to Scheme is converted
to the corresponding Scheme character.  An error is signaled if the C
character does not fit in the Scheme character.

   A Scheme exact integer passed to the C environment where a C integer
type (other than `char') is expected is converted to the corresponding
integral value.  An error is signaled if the value falls outside of the
range representable by that integral type.  C integer values passed to
the Scheme environment are mapped to the same Scheme exact integer.  If
the value is outside the fixnum range, a bignum is created.

   A Scheme inexact real passed to the C environment is converted to the
corresponding `float', `double', `float32' or `float64' value.  C
`float', `double', `float32' and `float64' values passed to the Scheme
environment are mapped to the closest Scheme inexact real.

   Scheme's rational numbers and complex numbers are not compatible with
any C numeric type.

   A Scheme string passed to the C environment where any C string type
is expected is converted to a null terminated string using the
appropriate encoding.  The C string is a fresh copy of the Scheme
string.  If the C string was created for an argument of a `c-lambda',
the C string will be reclaimed when the `c-lambda' returns.  If the C
string was created for returning the result of a `c-define' to C, the
caller is responsible for reclaiming the C string with a call to the
`___release_string' function (see below for an example).  Any C string
type passed to the Scheme environment causes the creation of a fresh
Scheme string containing a copy of the C string (unless the C string is
equal to `NULL', in which case it is converted to `#f').

   A foreign type passed to the Scheme environment causes the creation
and initialization of a Scheme foreign object with the appropriate tags
(except for the case of a `pointer' equal to `NULL' which is converted
to `#f').  A Scheme foreign object can be passed where a foreign type
is expected, on the condition that the tags are compatible and the
Scheme foreign object is not yet released.  The value `#f' is also
acceptable for a `pointer' type, and is converted to `NULL'.

   Scheme procedures defined with the `c-define' special form can be
passed where the `function' and `nonnull-function' types are expected.
The value `#f' is also acceptable for a `function' type, and is
converted to `NULL'.  No other Scheme procedures are acceptable.
Conversion from the `function' and `nonnull-function' types to Scheme
procedures is not currently implemented.

16.2 The `c-declare' special form
=================================

 -- special form: c-declare c-declaration
     Initially, the C file produced by `gsc' contains only an
     `#include' of `gambit.h'.  This header file provides a number of
     macro and procedure declarations to access the Scheme object
     representation.  The special form `c-declare' adds c-declaration
     (which must be a string containing the C declarations) to the C
     file.  This string is copied to the C file on a new line so it can
     start with preprocessor directives.  All types of C declarations
     are allowed (including type declarations, variable declarations,
     function declarations, `#include' directives, `#define's, and so
     on).  These declarations are visible to subsequent `c-declare's,
     `c-initialize's, and `c-lambda's, and `c-define's in the same
     module.  The most common use of this special form is to declare
     the external functions that are referenced in `c-lambda' special
     forms.  Such functions must either be declared explicitly or by
     including a header file which contains the appropriate C
     declarations.

     The `c-declare' special form does not return a value.  This form
     can only appear where a `define' form is acceptable.

     For example:

          (c-declare #<<c-declare-end

          #include <stdio.h>

          extern char *getlogin ();

          #ifdef sparc
          char *host = "sparc";
          #else
          char *host = "unknown";
          #endif

          FILE *tfile;

          c-declare-end
          )


16.3 The `c-initialize' special form
====================================

 -- special form: c-initialize c-code
     Just after the program is loaded and before control is passed to
     the Scheme code, each C file is initialized by calling its
     associated initialization function.  The body of this function is
     normally empty but it can be extended by using the `c-initialize'
     form.  Each occurence of the `c-initialize' form adds code to the
     body of the initialization function in the order of appearance in
     the source file.  c-code must be a string containing the C code to
     execute.  This string is copied to the C file on a new line so it
     can start with preprocessor directives.

     The `c-initialize' special form does not return a value.  This
     form can only appear where a `define' form is acceptable.

     For example:

          (c-initialize "tfile = tmpfile ();")


16.4 The `c-lambda' special form
================================

 -- special form: c-lambda (type1...) result-type c-name-or-code
     The `c-lambda' special form makes it possible to create a Scheme
     procedure that will act as a representative of some C function or
     C code sequence.  The first subform is a list containing the type
     of each argument.  The type of the function's result is given
     next.  Finally, the last subform is a string that either contains
     the name of the C function to call or some sequence of C code to
     execute.  Variadic C functions are not supported.  The resulting
     Scheme procedure takes exactly the number of arguments specified
     and delivers them in the same order to the C function.  When the
     Scheme procedure is called, the arguments will be converted to
     their C representation and then the C function will be called.
     The result returned by the C function will be converted to its
     Scheme representation and this value will be returned from the
     Scheme procedure call.  An error will be signaled if some
     conversion is not possible.  The temporary memory allocated from
     the C heap for the conversion of the arguments and result will be
     reclaimed whether there is an error or not.

     When c-name-or-code is not a valid C identifier, it is treated as
     an arbitrary piece of C code.  Within the C code the variables
     `___arg1', `___arg2', etc. can be referenced to access the
     converted arguments.  Note that the C `return' statement can't be
     used to return from the procedure.  Instead, the `___return' macro
     must be used.  A procedure whose result-type is not `void' must
     pass the procedure's result as the single argument to the
     `___return' macro, for example `___return(123);' to return the
     value 123.  When result-type is `void', the `___return' macro must
     be called without a parameter list, for example `___return;'.

     The C code is copied to the C file on a new line so it can start
     with preprocessor directives.  Moreover the C code is always
     placed at the head of a compound statement whose lifetime encloses
     the C to Scheme conversion of the procedure's result.
     Consequently, temporary storage (strings in particular) declared
     at the head of the C code can be returned with the `___return'
     macro.

     In the c-name-or-code, the macro `___AT_END' may be defined as the
     piece of C code to execute before control is returned to Scheme
     but after the procedure's result is converted to its Scheme
     representation.  This is mainly useful to deallocate temporary
     storage contained in the result.

     When passed to the Scheme environment, the C `void' type is
     converted to the void object.

     For example:

          (define fopen
            (c-lambda (nonnull-char-string nonnull-char-string)
                      (pointer "FILE")
             "fopen"))

          (define fgetc
            (c-lambda ((pointer "FILE"))
                      int
             "fgetc"))

          (let ((f (fopen "datafile" "r")))
            (if f (write (fgetc f))))

          (define char-code
            (c-lambda (char) int "___return(___arg1);"))

          (define host
            ((c-lambda () nonnull-char-string "___return(host);")))

          (define stdin
            ((c-lambda () (pointer "FILE") "___return(stdin);")))

          ((c-lambda () void
          #<<c-lambda-end
            printf( "hello\n" );
            printf( "world\n" );
          c-lambda-end
          ))

          (define pack-1-char
            (c-lambda (char)
                      nonnull-char-string
          #<<c-lambda-end
             char *s = (char *)malloc (2);
             if (s != NULL) { s[0] = ___arg1; s[1] = 0; }
             ___return(s);
             #define ___AT_END if (s != NULL) free (s);
          c-lambda-end
          ))

          (define pack-2-chars
            (c-lambda (char char)
                      nonnull-char-string
          #<<c-lambda-end
             char s[3];
             s[0] = ___arg1;
             s[1] = ___arg2;
             s[2] = 0;
             ___return(s);
          c-lambda-end
          ))


16.5 The `c-define' special form
================================

 -- special form: c-define (variable define-formals) (type1...)
          result-type c-name scope body
     The `c-define' special form makes it possible to create a C
     function that will act as a representative of some Scheme
     procedure.  A C function named c-name as well as a Scheme
     procedure bound to the variable variable are defined.  The
     parameters of the Scheme procedure are define-formals and its body
     is at the end of the form.  The type of each argument of the C
     function, its result type and c-name (which must be a string) are
     specified after the parameter specification of the Scheme
     procedure.  When the C function c-name is called from C, its
     arguments are converted to their Scheme representation and passed
     to the Scheme procedure.  The result of the Scheme procedure is
     then converted to its C representation and the C function c-name
     returns it to its caller.

     The scope of the C function can be changed with the scope
     parameter, which must be a string.  This string is placed
     immediately before the declaration of the C function.  So if scope
     is the string `"static"', the scope of c-name is local to the
     module it is in, whereas if scope is the empty string, c-name is
     visible from other modules.

     The `c-define' special form does not return a value.  It can only
     appear at top level.

     For example:

          (c-define (proc x #!optional (y x) #!rest z) (int int char float) int "f" ""
            (write (cons x (cons y z)))
            (newline)
            (+ x y))

          (proc 1 2 #\x 1.5) => 3 and prints (1 2 #\x 1.5)
          (proc 1)           => 2 and prints (1 1)

          ; if f is called from C with the call  f (1, 2, 'x', 1.5)
          ; the value 3 is returned and (1 2 #\x 1.5) is printed.
          ; f has to be called with 4 arguments.

     The `c-define' special form is particularly useful when the
     driving part of an application is written in C and Scheme
     procedures are called directly from C.  The Scheme part of the
     application is in a sense a "server" that is providing services to
     the C part.  The Scheme procedures that are to be called from C
     need to be defined using the `c-define' special form.  Before it
     can be used, the Scheme part must be initialized with a call to
     the function `___setup'.  Before the program terminates, it must
     call the function `___cleanup' so that the Scheme part may do final
     cleanup.  A sample application is given in the file
     `tests/server.scm'.


16.6 The `c-define-type' special form
=====================================

 -- special form: c-define-type name type [c-to-scheme scheme-to-c
          [cleanup]]
     This form associates the type identifier name to the C type type.
     The name must not clash with predefined types (e.g. `char-string',
     `ISO-8859-1', etc.) or with types previously defined with
     `c-define-type' in the same file.  The `c-define-type' special
     form does not return a value.  It can only appear at top level.

     If only the two parameters name and type are supplied then after
     this definition, the use of name in a type specification is
     synonymous to type.

     For example:

          (c-define-type FILE "FILE")
          (c-define-type FILE* (pointer FILE))
          (c-define-type time-struct-ptr (pointer (struct "tms")))
          (define fopen (c-lambda (char-string char-string) FILE* "fopen"))
          (define fgetc (c-lambda (FILE*) int "fgetc"))

     Note that identifiers are not case-sensitive in standard Scheme
     but it is good programming practice to use a name with the same
     case as in C.

     If four or more parameters are supplied, then type must be a
     string naming the C type, c-to-scheme and scheme-to-c must be
     strings suffixing the C macros that convert data of that type
     between C and Scheme.  If cleanup is supplied it must be a boolean
     indicating whether it is necessary to perform a cleanup operation
     (such as freeing memory) when data of that type is converted from
     Scheme to C (it defaults to `#t').  The cleanup information is
     used when the C stack is unwound due to a continuation invocation
     (see *Note continuations::).  Although it is safe to always
     specify `#t', it is more efficient in time and space to specify
     `#f' because the unwinding mechanism can skip C-interface frames
     which only contain conversions of data types requiring no cleanup.
     Two pairs of C macros need to be defined for conversions
     performed by `c-lambda' forms and two pairs for conversions
     performed by `c-define' forms:

          ___BEGIN_CFUN_scheme-to-c(___SCMOBJ, type, int)
          ___END_CFUN_scheme-to-c(___SCMOBJ, type, int)

          ___BEGIN_CFUN_c-to-scheme(type, ___SCMOBJ)
          ___END_CFUN_c-to-scheme(type, ___SCMOBJ)

          ___BEGIN_SFUN_c-to-scheme(type, ___SCMOBJ, int)
          ___END_SFUN_c-to-scheme(type, ___SCMOBJ, int)

          ___BEGIN_SFUN_scheme-to-c(___SCMOBJ, type)
          ___END_SFUN_scheme-to-c(___SCMOBJ, type)

     The macros prefixed with `___BEGIN' perform the conversion and
     those prefixed with `___END' perform any cleanup necessary (such as
     freeing memory temporarily allocated for the conversion).  The
     macro `___END_CFUN_scheme-to-c' must free the result of the
     conversion if it is memory allocated, and
     `___END_SFUN_scheme-to-c' must not (i.e. it is the responsibility
     of the caller to free the result).

     The first parameter of these macros is the C variable that
     contains the value to be converted, and the second parameter is
     the C variable in which to store the converted value.  The third
     parameter, when present, is the index (starting at 1) of the
     parameter of the `c-lambda' or `c-define' form that is being
     converted (this is useful for reporting precise error information
     when a conversion is impossible).

     To allow for type checking, the first three `___BEGIN' macros must
     expand to an unterminated compound statement prefixed by an `if',
     conditional on the absence of type check error:

          if ((___err = CONVERSION_OPERATION) == ___FIX(___NO_ERR)) {

     The last `___BEGIN' macro must expand to an unterminated compound
     statement:

          { ___err = CONVERSION_OPERATION;

     If type check errors are impossible then a `___BEGIN' macro can
     simply expand to an unterminated compound statement performing the
     conversion:

          { CONVERSION_OPERATION;

     The `___END' macros must expand to a statement, or to nothing if no
     cleanup is required, followed by a closing brace (to terminate the
     compound statement started at the corresponding `___BEGIN' macro).

     The CONVERSION_OPERATION is typically a function call that returns
     an error code value of type `___SCMOBJ' (the error codes are
     defined in `gambit.h', and the error code `___FIX(___UNKNOWN_ERR)'
     is available for generic errors).  CONVERSION_OPERATION can also
     set the variable `___errdata' of type `___SCMOBJ' to a specific
     Scheme string error message.

     Below is a simple example showing how to interface to an `EBCDIC'
     character type.  Memory allocation is not needed for conversion
     and type check errors are impossible when converting EBCDIC to
     Scheme characters, but they are possible when converting from
     Scheme characters to EBCDIC since Gambit supports Unicode
     characters.

          (c-declare #<<c-declare-end

          typedef char EBCDIC; /* EBCDIC encoded characters */

          void put_char (EBCDIC c) { ... } /* EBCDIC I/O functions */
          EBCDIC get_char (void) { ... }

          char EBCDIC_to_ISO_8859_1[256] = { ... }; /* conversion tables */
          char ISO_8859_1_to_EBCDIC[256] = { ... };

          ___SCMOBJ SCMOBJ_to_EBCDIC (___SCMOBJ src, EBCDIC *dst)
          {
            int x = ___INT(src); /* convert from Scheme character to int */
            if (x > 255) return ___FIX(___UNKNOWN_ERR);
            *dst = ISO_8859_1_to_EBCDIC[x];
            return ___FIX(___NO_ERR);
          }

          #define ___BEGIN_CFUN_SCMOBJ_to_EBCDIC(src,dst,i) \
          if ((___err = SCMOBJ_to_EBCDIC (src, &dst)) == ___FIX(___NO_ERR)) {
          #define ___END_CFUN_SCMOBJ_to_EBCDIC(src,dst,i) }

          #define ___BEGIN_CFUN_EBCDIC_to_SCMOBJ(src,dst) \
          { dst = ___CHR(EBCDIC_to_ISO_8859_1[src]);
          #define ___END_CFUN_EBCDIC_to_SCMOBJ(src,dst) }

          #define ___BEGIN_SFUN_EBCDIC_to_SCMOBJ(src,dst,i) \
          { dst = ___CHR(EBCDIC_to_ISO_8859_1[src]);
          #define ___END_SFUN_EBCDIC_to_SCMOBJ(src,dst,i) }

          #define ___BEGIN_SFUN_SCMOBJ_to_EBCDIC(src,dst) \
          { ___err = SCMOBJ_to_EBCDIC (src, &dst);
          #define ___END_SFUN_SCMOBJ_to_EBCDIC(src,dst) }

          c-declare-end
          )

          (c-define-type EBCDIC "EBCDIC" "EBCDIC_to_SCMOBJ" "SCMOBJ_to_EBCDIC" #f)

          (define put-char (c-lambda (EBCDIC) void "put_char"))
          (define get-char (c-lambda () EBCDIC "get_char"))

          (c-define (write-EBCDIC c) (EBCDIC) void "write_EBCDIC" ""
            (write-char c))

          (c-define (read-EBCDIC) () EBCDIC "read_EBCDIC" ""
            (read-char))

     Below is a more complex example that requires memory allocation
     when converting from C to Scheme.  It is an interface to a 2D
     `point' type which is represented in Scheme by a pair of integers.
     The conversion of the `x' and `y' components is done by calls to
     the conversion macros for the `int' type (defined in `gambit.h').
     Note that no cleanup is necessary when converting from Scheme to C
     (i.e. the last parameter of the `c-define-type' is `#f').

          (c-declare #<<c-declare-end

          typedef struct { int x, y; } point;

          void line_to (point p) { ... }
          point get_mouse (void) { ... }
          point add_points (point p1, point p2) { ... }

          ___SCMOBJ SCMOBJ_to_POINT (___PSD ___SCMOBJ src, point *dst, int arg_num)
          {
            ___SCMOBJ ___err = ___FIX(___NO_ERR);
            if (!___PAIRP(src))
              ___err = ___FIX(___UNKNOWN_ERR);
            else
              {
                ___SCMOBJ car = ___CAR(src);
                ___SCMOBJ cdr = ___CDR(src);
                ___BEGIN_CFUN_SCMOBJ_TO_INT(car,dst->x,arg_num)
                ___BEGIN_CFUN_SCMOBJ_TO_INT(cdr,dst->y,arg_num)
                ___END_CFUN_SCMOBJ_TO_INT(cdr,dst->y,arg_num)
                ___END_CFUN_SCMOBJ_TO_INT(car,dst->x,arg_num)
              }
            return ___err;
          }

          ___SCMOBJ POINT_to_SCMOBJ (___processor_state ___ps, point src, ___SCMOBJ *dst, int arg_num)
          {
            ___SCMOBJ ___err = ___FIX(___NO_ERR);
            ___SCMOBJ x_scmobj;
            ___SCMOBJ y_scmobj;
            ___BEGIN_SFUN_INT_TO_SCMOBJ(src.x,x_scmobj,arg_num)
            ___BEGIN_SFUN_INT_TO_SCMOBJ(src.y,y_scmobj,arg_num)
            *dst = ___EXT(___make_pair) (___ps, x_scmobj, y_scmobj);
            if (___FIXNUMP(*dst))
              ___err = *dst; /* return allocation error */
            ___END_SFUN_INT_TO_SCMOBJ(src.y,y_scmobj,arg_num)
            ___END_SFUN_INT_TO_SCMOBJ(src.x,x_scmobj,arg_num)
            return ___err;
          }

          #define ___BEGIN_CFUN_SCMOBJ_to_POINT(src,dst,i) \
          if ((___err = SCMOBJ_to_POINT (___PSP src, &dst, i)) == ___FIX(___NO_ERR)) {
          #define ___END_CFUN_SCMOBJ_to_POINT(src,dst,i) }

          #define ___BEGIN_CFUN_POINT_to_SCMOBJ(src,dst) \
          if ((___err = POINT_to_SCMOBJ (___ps, src, &dst, ___RETURN_POS)) == ___FIX(___NO_ERR)) {
          #define ___END_CFUN_POINT_to_SCMOBJ(src,dst) \
          ___EXT(___release_scmobj) (dst); }

          #define ___BEGIN_SFUN_POINT_to_SCMOBJ(src,dst,i) \
          if ((___err = POINT_to_SCMOBJ (___ps, src, &dst, i)) == ___FIX(___NO_ERR)) {
          #define ___END_SFUN_POINT_to_SCMOBJ(src,dst,i) \
          ___EXT(___release_scmobj) (dst); }

          #define ___BEGIN_SFUN_SCMOBJ_to_POINT(src,dst) \
          { ___err = SCMOBJ_to_POINT (___PSP src, &dst, ___RETURN_POS);
          #define ___END_SFUN_SCMOBJ_to_POINT(src,dst) }

          c-declare-end
          )

          (c-define-type point "point" "POINT_to_SCMOBJ" "SCMOBJ_to_POINT" #f)

          (define line-to (c-lambda (point) void "line_to"))
          (define get-mouse (c-lambda () point "get_mouse"))
          (define add-points (c-lambda (point point) point "add_points"))

          (c-define (write-point p) (point) void "write_point" ""
            (write p))

          (c-define (read-point) () point "read_point" ""
            (read))

     Note that the pair is allocated using the `___make_pair' runtime
     library function.  The prototype of this function is

          ___SCMOBJ ___make_pair(___processor_state ___ps, ___SCMOBJ car, ___SCMOBJ cdr);

     The fields of the pair are initialized to the `car' and `cdr'
     parameters.  The `___ps' parameter indicates how the pair is
     allocated.  A `NULL' `___ps' parameter will allocate the pair
     permanently (i.e. the pair will only be deallocated when
     `___cleanup' is called).  Otherwise a "still" object is allocated
     and the `___ps' parameter indicates the processor in whose heap
     the object is allocated (this is to support multithreaded
     execution).  Still objects are reference counted and initially
     have a reference count equal to 1.  The call to
     `___release_scmobj' in the macros `___END_CFUN_POINT_to_SCMOBJ' and
     `___END_SFUN_POINT_to_SCMOBJ' decrement this reference count.  A
     still object whose reference count is zero will be deallocated
     when a garbage collection is performed and there are no references
     to it from the Scheme world.  Note that the use of the `___PSD'
     macro in the parameter list of `SCMOBJ_to_POINT' and the `___PSP'
     macro in the calls of `SCMOBJ_to_POINT', are necessary to
     propagate the current processor state to that function.

     An example that requires memory allocation when converting from C
     to Scheme and Scheme to C is shown below.  It is an interface to a
     "null-terminated array of strings" type which is represented in
     Scheme by a list of strings.  Note that some cleanup is necessary
     when converting from Scheme to C.

          (c-declare #<<c-declare-end

          #include <stdlib.h>
          #include <unistd.h>

          extern char **environ;

          char **get_environ (void) { return environ; }

          void free_strings (char **strings)
          {
            char **ptr = strings;
            while (*ptr != NULL)
              {
                ___EXT(___release_string) (*ptr);
                ptr++;
              }
            free (strings);
          }

          ___SCMOBJ SCMOBJ_to_STRINGS (___PSD ___SCMOBJ src, char ***dst, int arg_num)
          {
            /*
             * Src is a list of Scheme strings.  Dst will be a null terminated
             * array of C strings.
             */

            int i;
            ___SCMOBJ lst = src;
            int len = 4; /* start with a small result array */
            char **result = (char**) malloc (len * sizeof (char*));

            if (result == NULL)
              return ___FIX(___HEAP_OVERFLOW_ERR);

            i = 0;
            result[i] = NULL; /* always keep array null terminated */

            while (___PAIRP(lst))
              {
                ___SCMOBJ scm_str = ___CAR(lst);
                char *c_str;
                ___SCMOBJ ___err;

                if (i >= len-1) /* need to grow the result array? */
                  {
                    char **new_result;
                    int j;

                    len = len * 3 / 2;
                    new_result = (char**) malloc (len * sizeof (char*));
                    if (new_result == NULL)
                      {
                        free_strings (result);
                        return ___FIX(___HEAP_OVERFLOW_ERR);
                      }
                    for (j=i; j>=0; j--)
                      new_result[j] = result[j];
                    free (result);
                    result = new_result;
                  }

                ___err = ___EXT(___SCMOBJ_to_CHARSTRING) (___PSP scm_str, &c_str, arg_num);

                if (___err != ___FIX(___NO_ERR))
                  {
                    free_strings (result);
                    return ___err;
                  }

                result[i++] = c_str;
                result[i] = NULL;
                lst = ___CDR(lst);
              }

            if (!___NULLP(lst))
              {
                free_strings (result);
                return ___FIX(___UNKNOWN_ERR);
              }

            /*
             * Note that the caller is responsible for calling free_strings
             * when it is done with the result.
             */

            *dst = result;
            return ___FIX(___NO_ERR);
          }

          ___SCMOBJ STRINGS_to_SCMOBJ (___processor_state ___ps, char **src, ___SCMOBJ *dst, int arg_num)
          {
            ___SCMOBJ ___err = ___FIX(___NO_ERR);
            ___SCMOBJ result = ___NUL; /* start with the empty list */
            int i = 0;

            while (src[i] != NULL)
              i++;

            /* build the list of strings starting at the tail */

            while (--i >= 0)
              {
                ___SCMOBJ scm_str;
                ___SCMOBJ new_result;

                /*
                 * Invariant: result is either the empty list or a ___STILL pair
                 * with reference count equal to 1.  This is important because
                 * it is possible that ___CHARSTRING_to_SCMOBJ and ___make_pair
                 * will invoke the garbage collector and we don't want the
                 * reference in result to become invalid (which would be the
                 * case if result was a ___MOVABLE pair or if it had a zero
                 * reference count).
                 */

                ___err = ___EXT(___CHARSTRING_to_SCMOBJ) (___ps, src[i], &scm_str, arg_num);

                if (___err != ___FIX(___NO_ERR))
                  {
                    ___EXT(___release_scmobj) (result); /* allow GC to reclaim result */
                    return ___FIX(___UNKNOWN_ERR);
                  }

                /*
                 * Note that scm_str will be a ___STILL object with reference
                 * count equal to 1, so there is no risk that it will be
                 * reclaimed or moved if ___make_pair invokes the garbage
                 * collector.
                 */

                new_result = ___EXT(___make_pair) (___ps, scm_str, result);

                /*
                 * We can zero the reference count of scm_str and result (if
                 * not the empty list) because the pair now references these
                 * objects and the pair is reachable (it can't be reclaimed
                 * or moved by the garbage collector).
                 */

                ___EXT(___release_scmobj) (scm_str);
                ___EXT(___release_scmobj) (result);

                result = new_result;

                if (___FIXNUMP(result))
                  return result; /* allocation failed */
              }

            /*
             * Note that result is either the empty list or a ___STILL pair
             * with a reference count equal to 1.  There will be a call to
             * ___release_scmobj later on (in ___END_CFUN_STRINGS_to_SCMOBJ
             * or ___END_SFUN_STRINGS_to_SCMOBJ) that will allow the garbage
             * collector to reclaim the whole list of strings when the Scheme
             * world no longer references it.
             */

            *dst = result;
            return ___FIX(___NO_ERR);
          }

          #define ___BEGIN_CFUN_SCMOBJ_to_STRINGS(src,dst,i) \
          if ((___err = SCMOBJ_to_STRINGS (___PSP src, &dst, i)) == ___FIX(___NO_ERR)) {
          #define ___END_CFUN_SCMOBJ_to_STRINGS(src,dst,i) \
          free_strings (dst); }

          #define ___BEGIN_CFUN_STRINGS_to_SCMOBJ(src,dst) \
          if ((___err = STRINGS_to_SCMOBJ (___ps, src, &dst, ___RETURN_POS)) == ___FIX(___NO_ERR)) {
          #define ___END_CFUN_STRINGS_to_SCMOBJ(src,dst) \
          ___EXT(___release_scmobj) (dst); }

          #define ___BEGIN_SFUN_STRINGS_to_SCMOBJ(src,dst,i) \
          if ((___err = STRINGS_to_SCMOBJ (___ps, src, &dst, i)) == ___FIX(___NO_ERR)) {
          #define ___END_SFUN_STRINGS_to_SCMOBJ(src,dst,i) \
          ___EXT(___release_scmobj) (dst); }

          #define ___BEGIN_SFUN_SCMOBJ_to_STRINGS(src,dst) \
          { ___err = SCMOBJ_to_STRINGS (___PSP src, &dst, ___RETURN_POS);
          #define ___END_SFUN_SCMOBJ_to_STRINGS(src,dst) }

          c-declare-end
          )

          (c-define-type char** "char**" "STRINGS_to_SCMOBJ" "SCMOBJ_to_STRINGS" #t)

          (define execv (c-lambda (char-string char**) int "execv"))
          (define get-environ (c-lambda () char** "get_environ"))

          (c-define (write-strings x) (char**) void "write_strings" ""
            (write x))

          (c-define (read-strings) () char** "read_strings" ""
            (read))


16.7 Continuations, the C-interface and threads
===============================================

The C-interface allows C to Scheme calls to be nested.  This means that
during a call from C to Scheme another call from C to Scheme can be
performed.  This case occurs in the following program:

     (c-declare #<<c-declare-end

     int p (char *); /* forward declarations */
     int q (void);

     int a (char *x) { return 2 * p (x+1); }
     int b (short y) { return y + q (); }

     c-declare-end
     )

     (define a (c-lambda (char-string) int "a"))
     (define b (c-lambda (short) int "b"))

     (c-define (p z) (char-string) int "p" ""
       (+ (b 10) (string-length z)))

     (c-define (q) () int "q" ""
       123)

     (write (a "hello"))

   In this example, the main Scheme program calls the C function `a'
which calls the Scheme procedure `p' which in turn calls the C function
`b' which finally calls the Scheme procedure `q'.

   Gambit maintains the Scheme continuation separately from the C stack,
thus allowing the Scheme continuation to be unwound independently from
the C stack.  The C stack frame created for the C function `f' is only
removed from the C stack when control returns from `f' or when control
returns to a C function "above" `f'.  Special care is required for
programs which escape to Scheme (using first-class continuations) from
a Scheme to C (to Scheme) call because the C stack frame will remain on
the stack.  The C stack may overflow if this happens in a loop with no
intervening return to a C function.  To avoid this problem make sure
the C stack gets cleaned up by executing a normal return from a Scheme
to C call.

   This approach to manage Scheme to C to Scheme calls may cause
problems when used with Scheme threads because context switching is
implemented with continuations.  If a Scheme thread T1 is in the middle
of a Scheme to C to Scheme call and a second thread T2 does a Scheme to
C to Scheme call and there is a Scheme thread context switch back to T1
which completes its call, the C stack frames of T2 will get removed,
preventing T2 (when it gets resumed) to complete its call correctly.
This situation can be avoided by having only one Scheme thread that
does Scheme to C to Scheme calls.  Other Scheme threads are limited to
simple Scheme to C calls that don't call back to Scheme.

17 System limitations
*********************

   * On some systems floating point overflows will cause the program to
     terminate with a floating point exception.

   * On some systems floating point operations involving `+nan.0'
     `+inf.0', `-inf.0', or `-0.' do not return the value required by
     the IEEE 754 floating point standard.

   * The maximum number of arguments that can be passed to a procedure
     by the `apply' procedure is 8192.


18 Copyright and license
************************

The Gambit system release v4.9.4 is Copyright (C) 1994-2022 by Marc
Feeley, all rights reserved.  The Gambit system release v4.9.4 is
licensed under two licenses: the Apache License, Version 2.0, and the
GNU LESSER GENERAL PUBLIC LICENSE, Version 2.1.  You have the option to
choose which of these two licenses to abide by.  The licenses are
copied below.


                                   Apache License
                             Version 2.0, January 2004
                          http://www.apache.org/licenses/

     TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

     1. Definitions.

        "License" shall mean the terms and conditions for use, reproduction,
        and distribution as defined by Sections 1 through 9 of this document.

        "Licensor" shall mean the copyright owner or entity authorized by
        the copyright owner that is granting the License.

        "Legal Entity" shall mean the union of the acting entity and all
        other entities that control, are controlled by, or are under common
        control with that entity. For the purposes of this definition,
        "control" means (i) the power, direct or indirect, to cause the
        direction or management of such entity, whether by contract or
        otherwise, or (ii) ownership of fifty percent (50%) or more of the
        outstanding shares, or (iii) beneficial ownership of such entity.

        "You" (or "Your") shall mean an individual or Legal Entity
        exercising permissions granted by this License.

        "Source" form shall mean the preferred form for making modifications,
        including but not limited to software source code, documentation
        source, and configuration files.

        "Object" form shall mean any form resulting from mechanical
        transformation or translation of a Source form, including but
        not limited to compiled object code, generated documentation,
        and conversions to other media types.

        "Work" shall mean the work of authorship, whether in Source or
        Object form, made available under the License, as indicated by a
        copyright notice that is included in or attached to the work
        (an example is provided in the Appendix below).

        "Derivative Works" shall mean any work, whether in Source or Object
        form, that is based on (or derived from) the Work and for which the
        editorial revisions, annotations, elaborations, or other modifications
        represent, as a whole, an original work of authorship. For the purposes
        of this License, Derivative Works shall not include works that remain
        separable from, or merely link (or bind by name) to the interfaces of,
        the Work and Derivative Works thereof.

        "Contribution" shall mean any work of authorship, including
        the original version of the Work and any modifications or additions
        to that Work or Derivative Works thereof, that is intentionally
        submitted to Licensor for inclusion in the Work by the copyright owner
        or by an individual or Legal Entity authorized to submit on behalf of
        the copyright owner. For the purposes of this definition, "submitted"
        means any form of electronic, verbal, or written communication sent
        to the Licensor or its representatives, including but not limited to
        communication on electronic mailing lists, source code control systems,
        and issue tracking systems that are managed by, or on behalf of, the
        Licensor for the purpose of discussing and improving the Work, but
        excluding communication that is conspicuously marked or otherwise
        designated in writing by the copyright owner as "Not a Contribution."

        "Contributor" shall mean Licensor and any individual or Legal Entity
        on behalf of whom a Contribution has been received by Licensor and
        subsequently incorporated within the Work.

     2. Grant of Copyright License. Subject to the terms and conditions of
        this License, each Contributor hereby grants to You a perpetual,
        worldwide, non-exclusive, no-charge, royalty-free, irrevocable
        copyright license to reproduce, prepare Derivative Works of,
        publicly display, publicly perform, sublicense, and distribute the
        Work and such Derivative Works in Source or Object form.

     3. Grant of Patent License. Subject to the terms and conditions of
        this License, each Contributor hereby grants to You a perpetual,
        worldwide, non-exclusive, no-charge, royalty-free, irrevocable
        (except as stated in this section) patent license to make, have made,
        use, offer to sell, sell, import, and otherwise transfer the Work,
        where such license applies only to those patent claims licensable
        by such Contributor that are necessarily infringed by their
        Contribution(s) alone or by combination of their Contribution(s)
        with the Work to which such Contribution(s) was submitted. If You
        institute patent litigation against any entity (including a
        cross-claim or counterclaim in a lawsuit) alleging that the Work
        or a Contribution incorporated within the Work constitutes direct
        or contributory patent infringement, then any patent licenses
        granted to You under this License for that Work shall terminate
        as of the date such litigation is filed.

     4. Redistribution. You may reproduce and distribute copies of the
        Work or Derivative Works thereof in any medium, with or without
        modifications, and in Source or Object form, provided that You
        meet the following conditions:

        (a) You must give any other recipients of the Work or
            Derivative Works a copy of this License; and

        (b) You must cause any modified files to carry prominent notices
            stating that You changed the files; and

        (c) You must retain, in the Source form of any Derivative Works
            that You distribute, all copyright, patent, trademark, and
            attribution notices from the Source form of the Work,
            excluding those notices that do not pertain to any part of
            the Derivative Works; and

        (d) If the Work includes a "NOTICE" text file as part of its
            distribution, then any Derivative Works that You distribute must
            include a readable copy of the attribution notices contained
            within such NOTICE file, excluding those notices that do not
            pertain to any part of the Derivative Works, in at least one
            of the following places: within a NOTICE text file distributed
            as part of the Derivative Works; within the Source form or
            documentation, if provided along with the Derivative Works; or,
            within a display generated by the Derivative Works, if and
            wherever such third-party notices normally appear. The contents
            of the NOTICE file are for informational purposes only and
            do not modify the License. You may add Your own attribution
            notices within Derivative Works that You distribute, alongside
            or as an addendum to the NOTICE text from the Work, provided
            that such additional attribution notices cannot be construed
            as modifying the License.

        You may add Your own copyright statement to Your modifications and
        may provide additional or different license terms and conditions
        for use, reproduction, or distribution of Your modifications, or
        for any such Derivative Works as a whole, provided Your use,
        reproduction, and distribution of the Work otherwise complies with
        the conditions stated in this License.

     5. Submission of Contributions. Unless You explicitly state otherwise,
        any Contribution intentionally submitted for inclusion in the Work
        by You to the Licensor shall be under the terms and conditions of
        this License, without any additional terms or conditions.
        Notwithstanding the above, nothing herein shall supersede or modify
        the terms of any separate license agreement you may have executed
        with Licensor regarding such Contributions.

     6. Trademarks. This License does not grant permission to use the trade
        names, trademarks, service marks, or product names of the Licensor,
        except as required for reasonable and customary use in describing the
        origin of the Work and reproducing the content of the NOTICE file.

     7. Disclaimer of Warranty. Unless required by applicable law or
        agreed to in writing, Licensor provides the Work (and each
        Contributor provides its Contributions) on an "AS IS" BASIS,
        WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
        implied, including, without limitation, any warranties or conditions
        of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
        PARTICULAR PURPOSE. You are solely responsible for determining the
        appropriateness of using or redistributing the Work and assume any
        risks associated with Your exercise of permissions under this License.

     8. Limitation of Liability. In no event and under no legal theory,
        whether in tort (including negligence), contract, or otherwise,
        unless required by applicable law (such as deliberate and grossly
        negligent acts) or agreed to in writing, shall any Contributor be
        liable to You for damages, including any direct, indirect, special,
        incidental, or consequential damages of any character arising as a
        result of this License or out of the use or inability to use the
        Work (including but not limited to damages for loss of goodwill,
        work stoppage, computer failure or malfunction, or any and all
        other commercial damages or losses), even if such Contributor
        has been advised of the possibility of such damages.

     9. Accepting Warranty or Additional Liability. While redistributing
        the Work or Derivative Works thereof, You may choose to offer,
        and charge a fee for, acceptance of support, warranty, indemnity,
        or other liability obligations and/or rights consistent with this
        License. However, in accepting such obligations, You may act only
        on Your own behalf and on Your sole responsibility, not on behalf
        of any other Contributor, and only if You agree to indemnify,
        defend, and hold each Contributor harmless for any liability
        incurred by, or claims asserted against, such Contributor by reason
        of your accepting any such warranty or additional liability.

     END OF TERMS AND CONDITIONS

     APPENDIX: How to apply the Apache License to your work.

        To apply the Apache License to your work, attach the following
        boilerplate notice, with the fields enclosed by brackets "[]"
        replaced with your own identifying information. (Don't include
        the brackets!)  The text should be enclosed in the appropriate
        comment syntax for the file format. We also recommend that a
        file or class name and description of purpose be included on the
        same "printed page" as the copyright notice for easier
        identification within third-party archives.

     Copyright [yyyy] [name of copyright owner]

     Licensed under the Apache License, Version 2.0 (the "License");
     you may not use this file except in compliance with the License.
     You may obtain a copy of the License at

         http://www.apache.org/licenses/LICENSE-2.0

     Unless required by applicable law or agreed to in writing, software
     distributed under the License is distributed on an "AS IS" BASIS,
     WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
     See the License for the specific language governing permissions and
     limitations under the License.

                       GNU LESSER GENERAL PUBLIC LICENSE
                            Version 2.1, February 1999

      Copyright (C) 1991, 1999 Free Software Foundation, Inc.
          51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
      Everyone is permitted to copy and distribute verbatim copies
      of this license document, but changing it is not allowed.

     [This is the first released version of the Lesser GPL.  It also counts
      as the successor of the GNU Library Public License, version 2, hence
      the version number 2.1.]

                                 Preamble

       The licenses for most software are designed to take away your
     freedom to share and change it.  By contrast, the GNU General Public
     Licenses are intended to guarantee your freedom to share and change
     free software--to make sure the software is free for all its users.

       This license, the Lesser General Public License, applies to some
     specially designated software packages--typically libraries--of the
     Free Software Foundation and other authors who decide to use it.  You
     can use it too, but we suggest you first think carefully about whether
     this license or the ordinary General Public License is the better
     strategy to use in any particular case, based on the explanations below.

       When we speak of free software, we are referring to freedom of use,
     not price.  Our General Public Licenses are designed to make sure that
     you have the freedom to distribute copies of free software (and charge
     for this service if you wish); that you receive source code or can get
     it if you want it; that you can change the software and use pieces of
     it in new free programs; and that you are informed that you can do
     these things.

       To protect your rights, we need to make restrictions that forbid
     distributors to deny you these rights or to ask you to surrender these
     rights.  These restrictions translate to certain responsibilities for
     you if you distribute copies of the library or if you modify it.

       For example, if you distribute copies of the library, whether gratis
     or for a fee, you must give the recipients all the rights that we gave
     you.  You must make sure that they, too, receive or can get the source
     code.  If you link other code with the library, you must provide
     complete object files to the recipients, so that they can relink them
     with the library after making changes to the library and recompiling
     it.  And you must show them these terms so they know their rights.

       We protect your rights with a two-step method: (1) we copyright the
     library, and (2) we offer you this license, which gives you legal
     permission to copy, distribute and/or modify the library.

       To protect each distributor, we want to make it very clear that
     there is no warranty for the free library.  Also, if the library is
     modified by someone else and passed on, the recipients should know
     that what they have is not the original version, so that the original
     author's reputation will not be affected by problems that might be
     introduced by others.

       Finally, software patents pose a constant threat to the existence of
     any free program.  We wish to make sure that a company cannot
     effectively restrict the users of a free program by obtaining a
     restrictive license from a patent holder.  Therefore, we insist that
     any patent license obtained for a version of the library must be
     consistent with the full freedom of use specified in this license.

       Most GNU software, including some libraries, is covered by the
     ordinary GNU General Public License.  This license, the GNU Lesser
     General Public License, applies to certain designated libraries, and
     is quite different from the ordinary General Public License.  We use
     this license for certain libraries in order to permit linking those
     libraries into non-free programs.

       When a program is linked with a library, whether statically or using
     a shared library, the combination of the two is legally speaking a
     combined work, a derivative of the original library.  The ordinary
     General Public License therefore permits such linking only if the
     entire combination fits its criteria of freedom.  The Lesser General
     Public License permits more lax criteria for linking other code with
     the library.

       We call this license the "Lesser" General Public License because it
     does Less to protect the user's freedom than the ordinary General
     Public License.  It also provides other free software developers Less
     of an advantage over competing non-free programs.  These disadvantages
     are the reason we use the ordinary General Public License for many
     libraries.  However, the Lesser license provides advantages in certain
     special circumstances.

       For example, on rare occasions, there may be a special need to
     encourage the widest possible use of a certain library, so that it becomes
     a de-facto standard.  To achieve this, non-free programs must be
     allowed to use the library.  A more frequent case is that a free
     library does the same job as widely used non-free libraries.  In this
     case, there is little to gain by limiting the free library to free
     software only, so we use the Lesser General Public License.

       In other cases, permission to use a particular library in non-free
     programs enables a greater number of people to use a large body of
     free software.  For example, permission to use the GNU C Library in
     non-free programs enables many more people to use the whole GNU
     operating system, as well as its variant, the GNU/Linux operating
     system.

       Although the Lesser General Public License is Less protective of the
     users' freedom, it does ensure that the user of a program that is
     linked with the Library has the freedom and the wherewithal to run
     that program using a modified version of the Library.

       The precise terms and conditions for copying, distribution and
     modification follow.  Pay close attention to the difference between a
     "work based on the library" and a "work that uses the library".  The
     former contains code derived from the library, whereas the latter must
     be combined with the library in order to run.

                       GNU LESSER GENERAL PUBLIC LICENSE
        TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION

       0. This License Agreement applies to any software library or other
     program which contains a notice placed by the copyright holder or
     other authorized party saying it may be distributed under the terms of
     this Lesser General Public License (also called "this License").
     Each licensee is addressed as "you".

       A "library" means a collection of software functions and/or data
     prepared so as to be conveniently linked with application programs
     (which use some of those functions and data) to form executables.

       The "Library", below, refers to any such software library or work
     which has been distributed under these terms.  A "work based on the
     Library" means either the Library or any derivative work under
     copyright law: that is to say, a work containing the Library or a
     portion of it, either verbatim or with modifications and/or translated
     straightforwardly into another language.  (Hereinafter, translation is
     included without limitation in the term "modification".)

       "Source code" for a work means the preferred form of the work for
     making modifications to it.  For a library, complete source code means
     all the source code for all modules it contains, plus any associated
     interface definition files, plus the scripts used to control compilation
     and installation of the library.

       Activities other than copying, distribution and modification are not
     covered by this License; they are outside its scope.  The act of
     running a program using the Library is not restricted, and output from
     such a program is covered only if its contents constitute a work based
     on the Library (independent of the use of the Library in a tool for
     writing it).  Whether that is true depends on what the Library does
     and what the program that uses the Library does.

       1. You may copy and distribute verbatim copies of the Library's
     complete source code as you receive it, in any medium, provided that
     you conspicuously and appropriately publish on each copy an
     appropriate copyright notice and disclaimer of warranty; keep intact
     all the notices that refer to this License and to the absence of any
     warranty; and distribute a copy of this License along with the
     Library.

       You may charge a fee for the physical act of transferring a copy,
     and you may at your option offer warranty protection in exchange for a
     fee.

       2. You may modify your copy or copies of the Library or any portion
     of it, thus forming a work based on the Library, and copy and
     distribute such modifications or work under the terms of Section 1
     above, provided that you also meet all of these conditions:

         a) The modified work must itself be a software library.

         b) You must cause the files modified to carry prominent notices
         stating that you changed the files and the date of any change.

         c) You must cause the whole of the work to be licensed at no
         charge to all third parties under the terms of this License.

         d) If a facility in the modified Library refers to a function or a
         table of data to be supplied by an application program that uses
         the facility, other than as an argument passed when the facility
         is invoked, then you must make a good faith effort to ensure that,
         in the event an application does not supply such function or
         table, the facility still operates, and performs whatever part of
         its purpose remains meaningful.

         (For example, a function in a library to compute square roots has
         a purpose that is entirely well-defined independent of the
         application.  Therefore, Subsection 2d requires that any
         application-supplied function or table used by this function must
         be optional: if the application does not supply it, the square
         root function must still compute square roots.)

     These requirements apply to the modified work as a whole.  If
     identifiable sections of that work are not derived from the Library,
     and can be reasonably considered independent and separate works in
     themselves, then this License, and its terms, do not apply to those
     sections when you distribute them as separate works.  But when you
     distribute the same sections as part of a whole which is a work based
     on the Library, the distribution of the whole must be on the terms of
     this License, whose permissions for other licensees extend to the
     entire whole, and thus to each and every part regardless of who wrote
     it.

     Thus, it is not the intent of this section to claim rights or contest
     your rights to work written entirely by you; rather, the intent is to
     exercise the right to control the distribution of derivative or
     collective works based on the Library.

     In addition, mere aggregation of another work not based on the Library
     with the Library (or with a work based on the Library) on a volume of
     a storage or distribution medium does not bring the other work under
     the scope of this License.

       3. You may opt to apply the terms of the ordinary GNU General Public
     License instead of this License to a given copy of the Library.  To do
     this, you must alter all the notices that refer to this License, so
     that they refer to the ordinary GNU General Public License, version 2,
     instead of to this License.  (If a newer version than version 2 of the
     ordinary GNU General Public License has appeared, then you can specify
     that version instead if you wish.)  Do not make any other change in
     these notices.

       Once this change is made in a given copy, it is irreversible for
     that copy, so the ordinary GNU General Public License applies to all
     subsequent copies and derivative works made from that copy.

       This option is useful when you wish to copy part of the code of
     the Library into a program that is not a library.

       4. You may copy and distribute the Library (or a portion or
     derivative of it, under Section 2) in object code or executable form
     under the terms of Sections 1 and 2 above provided that you accompany
     it with the complete corresponding machine-readable source code, which
     must be distributed under the terms of Sections 1 and 2 above on a
     medium customarily used for software interchange.

       If distribution of object code is made by offering access to copy
     from a designated place, then offering equivalent access to copy the
     source code from the same place satisfies the requirement to
     distribute the source code, even though third parties are not
     compelled to copy the source along with the object code.

       5. A program that contains no derivative of any portion of the
     Library, but is designed to work with the Library by being compiled or
     linked with it, is called a "work that uses the Library".  Such a
     work, in isolation, is not a derivative work of the Library, and
     therefore falls outside the scope of this License.

       However, linking a "work that uses the Library" with the Library
     creates an executable that is a derivative of the Library (because it
     contains portions of the Library), rather than a "work that uses the
     library".  The executable is therefore covered by this License.
     Section 6 states terms for distribution of such executables.

       When a "work that uses the Library" uses material from a header file
     that is part of the Library, the object code for the work may be a
     derivative work of the Library even though the source code is not.
     Whether this is true is especially significant if the work can be
     linked without the Library, or if the work is itself a library.  The
     threshold for this to be true is not precisely defined by law.

       If such an object file uses only numerical parameters, data
     structure layouts and accessors, and small macros and small inline
     functions (ten lines or less in length), then the use of the object
     file is unrestricted, regardless of whether it is legally a derivative
     work.  (Executables containing this object code plus portions of the
     Library will still fall under Section 6.)

       Otherwise, if the work is a derivative of the Library, you may
     distribute the object code for the work under the terms of Section 6.
     Any executables containing that work also fall under Section 6,
     whether or not they are linked directly with the Library itself.

       6. As an exception to the Sections above, you may also combine or
     link a "work that uses the Library" with the Library to produce a
     work containing portions of the Library, and distribute that work
     under terms of your choice, provided that the terms permit
     modification of the work for the customer's own use and reverse
     engineering for debugging such modifications.

       You must give prominent notice with each copy of the work that the
     Library is used in it and that the Library and its use are covered by
     this License.  You must supply a copy of this License.  If the work
     during execution displays copyright notices, you must include the
     copyright notice for the Library among them, as well as a reference
     directing the user to the copy of this License.  Also, you must do one
     of these things:

         a) Accompany the work with the complete corresponding
         machine-readable source code for the Library including whatever
         changes were used in the work (which must be distributed under
         Sections 1 and 2 above); and, if the work is an executable linked
         with the Library, with the complete machine-readable "work that
         uses the Library", as object code and/or source code, so that the
         user can modify the Library and then relink to produce a modified
         executable containing the modified Library.  (It is understood
         that the user who changes the contents of definitions files in the
         Library will not necessarily be able to recompile the application
         to use the modified definitions.)

         b) Use a suitable shared library mechanism for linking with the
         Library.  A suitable mechanism is one that (1) uses at run time a
         copy of the library already present on the user's computer system,
         rather than copying library functions into the executable, and (2)
         will operate properly with a modified version of the library, if
         the user installs one, as long as the modified version is
         interface-compatible with the version that the work was made with.

         c) Accompany the work with a written offer, valid for at
         least three years, to give the same user the materials
         specified in Subsection 6a, above, for a charge no more
         than the cost of performing this distribution.

         d) If distribution of the work is made by offering access to copy
         from a designated place, offer equivalent access to copy the above
         specified materials from the same place.

         e) Verify that the user has already received a copy of these
         materials or that you have already sent this user a copy.

       For an executable, the required form of the "work that uses the
     Library" must include any data and utility programs needed for
     reproducing the executable from it.  However, as a special exception,
     the materials to be distributed need not include anything that is
     normally distributed (in either source or binary form) with the major
     components (compiler, kernel, and so on) of the operating system on
     which the executable runs, unless that component itself accompanies
     the executable.

       It may happen that this requirement contradicts the license
     restrictions of other proprietary libraries that do not normally
     accompany the operating system.  Such a contradiction means you cannot
     use both them and the Library together in an executable that you
     distribute.

       7. You may place library facilities that are a work based on the
     Library side-by-side in a single library together with other library
     facilities not covered by this License, and distribute such a combined
     library, provided that the separate distribution of the work based on
     the Library and of the other library facilities is otherwise
     permitted, and provided that you do these two things:

         a) Accompany the combined library with a copy of the same work
         based on the Library, uncombined with any other library
         facilities.  This must be distributed under the terms of the
         Sections above.

         b) Give prominent notice with the combined library of the fact
         that part of it is a work based on the Library, and explaining
         where to find the accompanying uncombined form of the same work.

       8. You may not copy, modify, sublicense, link with, or distribute
     the Library except as expressly provided under this License.  Any
     attempt otherwise to copy, modify, sublicense, link with, or
     distribute the Library is void, and will automatically terminate your
     rights under this License.  However, parties who have received copies,
     or rights, from you under this License will not have their licenses
     terminated so long as such parties remain in full compliance.

       9. You are not required to accept this License, since you have not
     signed it.  However, nothing else grants you permission to modify or
     distribute the Library or its derivative works.  These actions are
     prohibited by law if you do not accept this License.  Therefore, by
     modifying or distributing the Library (or any work based on the
     Library), you indicate your acceptance of this License to do so, and
     all its terms and conditions for copying, distributing or modifying
     the Library or works based on it.

       10. Each time you redistribute the Library (or any work based on the
     Library), the recipient automatically receives a license from the
     original licensor to copy, distribute, link with or modify the Library
     subject to these terms and conditions.  You may not impose any further
     restrictions on the recipients' exercise of the rights granted herein.
     You are not responsible for enforcing compliance by third parties with
     this License.

       11. If, as a consequence of a court judgment or allegation of patent
     infringement or for any other reason (not limited to patent issues),
     conditions are imposed on you (whether by court order, agreement or
     otherwise) that contradict the conditions of this License, they do not
     excuse you from the conditions of this License.  If you cannot
     distribute so as to satisfy simultaneously your obligations under this
     License and any other pertinent obligations, then as a consequence you
     may not distribute the Library at all.  For example, if a patent
     license would not permit royalty-free redistribution of the Library by
     all those who receive copies directly or indirectly through you, then
     the only way you could satisfy both it and this License would be to
     refrain entirely from distribution of the Library.

     If any portion of this section is held invalid or unenforceable under any
     particular circumstance, the balance of the section is intended to apply,
     and the section as a whole is intended to apply in other circumstances.

     It is not the purpose of this section to induce you to infringe any
     patents or other property right claims or to contest validity of any
     such claims; this section has the sole purpose of protecting the
     integrity of the free software distribution system which is
     implemented by public license practices.  Many people have made
     generous contributions to the wide range of software distributed
     through that system in reliance on consistent application of that
     system; it is up to the author/donor to decide if he or she is willing
     to distribute software through any other system and a licensee cannot
     impose that choice.

     This section is intended to make thoroughly clear what is believed to
     be a consequence of the rest of this License.

       12. If the distribution and/or use of the Library is restricted in
     certain countries either by patents or by copyrighted interfaces, the
     original copyright holder who places the Library under this License may add
     an explicit geographical distribution limitation excluding those countries,
     so that distribution is permitted only in or among countries not thus
     excluded.  In such case, this License incorporates the limitation as if
     written in the body of this License.

       13. The Free Software Foundation may publish revised and/or new
     versions of the Lesser General Public License from time to time.
     Such new versions will be similar in spirit to the present version,
     but may differ in detail to address new problems or concerns.

     Each version is given a distinguishing version number.  If the Library
     specifies a version number of this License which applies to it and
     "any later version", you have the option of following the terms and
     conditions either of that version or of any later version published by
     the Free Software Foundation.  If the Library does not specify a
     license version number, you may choose any version ever published by
     the Free Software Foundation.

       14. If you wish to incorporate parts of the Library into other free
     programs whose distribution conditions are incompatible with these,
     write to the author to ask for permission.  For software which is
     copyrighted by the Free Software Foundation, write to the Free
     Software Foundation; we sometimes make exceptions for this.  Our
     decision will be guided by the two goals of preserving the free status
     of all derivatives of our free software and of promoting the sharing
     and reuse of software generally.

                                 NO WARRANTY

       15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO
     WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW.
     EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR
     OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY
     KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE
     IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE
     LIBRARY IS WITH YOU.  SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME
     THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

       16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
     WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY
     AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU
     FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR
     CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE
     LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING
     RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A
     FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF
     SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
     DAMAGES.

                          END OF TERMS AND CONDITIONS

                How to Apply These Terms to Your New Libraries

       If you develop a new library, and you want it to be of the greatest
     possible use to the public, we recommend making it free software that
     everyone can redistribute and change.  You can do so by permitting
     redistribution under these terms (or, alternatively, under the terms of the
     ordinary General Public License).

       To apply these terms, attach the following notices to the library.  It is
     safest to attach them to the start of each source file to most effectively
     convey the exclusion of warranty; and each file should have at least the
     "copyright" line and a pointer to where the full notice is found.

         <one line to give the library's name and a brief idea of what it does.>
         Copyright (C) <year>  <name of author>

         This library is free software; you can redistribute it and/or
         modify it under the terms of the GNU Lesser General Public
         License as published by the Free Software Foundation; either
         version 2.1 of the License, or (at your option) any later version.

         This library is distributed in the hope that it will be useful,
         but WITHOUT ANY WARRANTY; without even the implied warranty of
         MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
         Lesser General Public License for more details.

         You should have received a copy of the GNU Lesser General Public
         License along with this library; if not, write to the Free Software
         Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

     Also add information on how to contact you by electronic and paper mail.

     You should also get your employer (if you work as a programmer) or your
     school, if any, to sign a "copyright disclaimer" for the library, if
     necessary.  Here is a sample; alter the names:

       Yoyodyne, Inc., hereby disclaims all copyright interest in the
       library `Frob' (a library for tweaking knobs) written by James Random Hacker.

       <signature of Ty Coon>, 1 April 1990
       Ty Coon, President of Vice

     That's all there is to it!

General index
*************

#:                                             See 5.4.     (line  2449)
##:                                            See 5.4.     (line  2449)
##import:                                      See 7.2.2.   (line  4861)
##include:                                     See 7.1.     (line  4728)
##namespace:                                   See 7.2.2.   (line  4859)
+z:                                            See 3.4.4.   (line  1439)
,(b EXPR):                                     See 5.2.     (line  2208)
,(be EXPR):                                    See 5.2.     (line  2215)
,(bed EXPR):                                   See 5.2.     (line  2218)
,(c EXPR):                                     See 5.2.     (line  2101)
,(e EXPR):                                     See 5.2.     (line  2256)
,(ed EXPR):                                    See 5.2.     (line  2268)
,(h SUBJECT):                                  See 5.2.     (line  2074)
,(st EXPR):                                    See 5.2.     (line  2279)
,(v EXPR):                                     See 5.2.     (line  2283)
,+:                                            See 5.2.     (line  2162)
,++:                                           See 5.2.     (line  2168)
,-:                                            See 5.2.     (line  2165)
,--:                                           See 5.2.     (line  2172)
,?:                                            See 5.2.     (line  2071)
,b:                                            See 5.2.     (line  2193)
,be:                                           See 5.2.     (line  2202)
,bed:                                          See 5.2.     (line  2205)
,c:                                            See 5.2.     (line  2118)
,d:                                            See 5.2.     (line  2097)
,e:                                            See 5.2.     (line  2228)
,ed:                                           See 5.2.     (line  2252)
,h:                                            See 5.2.     (line  2081)
,help:                                         See 5.2.     (line  2071)
,i:                                            See 5.2.     (line  2222)
,l:                                            See 5.2.     (line  2138)
,N:                                            See 5.2.     (line  2145)
,N+:                                           See 5.2.     (line  2150)
,N-:                                           See 5.2.     (line  2156)
,q:                                            See 5.2.     (line  2086)
,qt:                                           See 5.2.     (line  2090)
,s:                                            See 5.2.     (line  2124)
,st:                                           See 5.2.     (line  2272)
,t:                                            See 5.2.     (line  2094)
,y:                                            See 5.2.     (line  2176)
- <1>:                                         See 3.3.     (line   979)
- <2>:                                         See 2.       (line   149)
-:                                             See 2.2.     (line   214)
-:+ARGUMENT:                                   See 4.       (line  1859)
-:,:                                           See 4.       (line  1971)
-:-[IO...]:                                    See 4.       (line  1862)
-:0[IO...]:                                    See 4.       (line  1862)
-:1[IO...]:                                    See 4.       (line  1862)
-:2[IO...]:                                    See 4.       (line  1862)
-:add-arg=ARGUMENT:                            See 4.       (line  1859)
-:ask-install=WHEN:                            See 4.       (line  1961)
-:d:                                           See 4.       (line  1776)
-:d$[INTF][:PORT]:                             See 4.       (line  1840)
-:d+:                                          See 4.       (line  1830)
-:d-:                                          See 4.       (line  1826)
-:d@[HOST][:PORT]:                             See 4.       (line  1834)
-:da:                                          See 4.       (line  1784)
-:dc:                                          See 4.       (line  1823)
-:dD:                                          See 4.       (line  1806)
-:debug:                                       See 4.       (line  1776)
-:debug=$[INTF][:PORT]:                        See 4.       (line  1840)
-:debug=+:                                     See 4.       (line  1830)
-:debug=-:                                     See 4.       (line  1826)
-:debug=@[HOST][:PORT]:                        See 4.       (line  1834)
-:debug=a:                                     See 4.       (line  1784)
-:debug=c:                                     See 4.       (line  1823)
-:debug=D:                                     See 4.       (line  1806)
-:debug=LEVEL:                                 See 4.       (line  1814)
-:debug=p:                                     See 4.       (line  1780)
-:debug=q:                                     See 4.       (line  1794)
-:debug=Q:                                     See 4.       (line  1810)
-:debug=r:                                     See 4.       (line  1787)
-:debug=R:                                     See 4.       (line  1798)
-:debug=s:                                     See 4.       (line  1790)
-:dLEVEL:                                      See 4.       (line  1814)
-:dp:                                          See 4.       (line  1780)
-:dq:                                          See 4.       (line  1794)
-:dQ:                                          See 4.       (line  1810)
-:dR:                                          See 4.       (line  1798)
-:dr:                                          See 4.       (line  1787)
-:ds:                                          See 4.       (line  1790)
-:f[IO...]:                                    See 4.       (line  1862)
-:file-settings=[IO...]:                       See 4.       (line  1862)
-:gambit:                                      See 4.       (line  1770)
-:hSIZE:                                       See 4.       (line  1755)
-:i[IO...]:                                    See 4.       (line  1862)
-:io-settings=[IO...]:                         See 4.       (line  1862)
-:live-ratio=RATIO:                            See 4.       (line  1762)
-:lRATIO:                                      See 4.       (line  1762)
-:max-heap=SIZE:                               See 4.       (line  1755)
-:min-heap=SIZE:                               See 4.       (line  1755)
-:mSIZE:                                       See 4.       (line  1755)
-:r5rs:                                        See 4.       (line  1770)
-:r7rs:                                        See 4.       (line  1770)
-:S:                                           See 4.       (line  1770)
-:s <1>:                                       See 6.3.     (line  3647)
-:s:                                           See 4.       (line  1770)
-:search=[DIR]:                                See 4.       (line  1949)
-:stdio-settings=[IO...]:                      See 4.       (line  1862)
-:t[IO...]:                                    See 4.       (line  1862)
-:terminal-settings=[IO...]:                   See 4.       (line  1862)
-:whitelist=[SOURCE]:                          See 4.       (line  1954)
-:~~NAME=DIRECTORY:                            See 4.       (line  1855)
-c:                                            See 3.3.     (line   932)
-call_shared:                                  See 3.4.4.   (line  1439)
-cc:                                           See 3.3.     (line   823)
-cc-options:                                   See 3.3.     (line   835)
-cfg:                                          See 3.3.     (line   892)
-compactness LEVEL:                            See 3.3.     (line  1002)
-D___DYNAMIC:                                  See 3.4.2.   (line  1234)
-D___LIBRARY:                                  See 3.4.3.   (line  1380)
-D___PRIMAL:                                   See 3.4.3.   (line  1380)
-D___SHARED:                                   See 3.4.3.   (line  1380)
-D___SINGLE_HOST:                              See 3.4.4.   (line  1419)
-debug <1>:                                    See 3.3.     (line   907)
-debug:                                        See 6.3.     (line  3919)
-debug-environments <1>:                       See 3.3.     (line   919)
-debug-environments:                           See 6.3.     (line  3959)
-debug-location <1>:                           See 6.3.     (line  3933)
-debug-location:                               See 3.3.     (line   911)
-debug-source <1>:                             See 3.3.     (line   915)
-debug-source:                                 See 6.3.     (line  3946)
-dg:                                           See 3.3.     (line   900)
-dynamic:                                      See 3.3.     (line   932)
-e <1>:                                        See 3.3.     (line   981)
-e <2>:                                        See 2.       (line   149)
-e:                                            See 2.2.     (line   214)
-exe:                                          See 3.3.     (line   932)
-expansion:                                    See 3.3.     (line   886)
-f <1>:                                        See 3.2.     (line   611)
-f:                                            See 2.       (line   149)
-flat:                                         See 3.3.     (line   955)
-fpic:                                         See 3.4.4.   (line  1439)
-fPIC:                                         See 3.4.4.   (line  1439)
-G:                                            See 3.4.4.   (line  1439)
-gvm:                                          See 3.3.     (line   889)
-h <1>:                                        See 3.       (line   592)
-h:                                            See 2.       (line   149)
-help <1>:                                     See 3.       (line   592)
-help:                                         See 2.       (line   149)
-i <1>:                                        See 3.       (line   597)
-i <2>:                                        See 2.       (line   149)
-i:                                            See 3.3.     (line   802)
-I/usr/local/Gambit/include:                   See 3.4.4.   (line  1431)
-install <1>:                                  See 2.3.     (line   301)
-install <2>:                                  See 7.5.     (line  5807)
-install:                                      See 2.       (line   149)
-keep-temp:                                    See 3.3.     (line   932)
-KPIC:                                         See 3.4.4.   (line  1439)
-Kpic:                                         See 3.4.4.   (line  1439)
-l BASE:                                       See 3.3.     (line   962)
-L/usr/local/Gambit/lib:                       See 3.4.4.   (line  1431)
-ld-options:                                   See 3.3.     (line   846)
-ld-options-prelude:                           See 3.3.     (line   846)
-link:                                         See 3.3.     (line   654)
-nb-arg-regs N:                                See 3.3.     (line   993)
-nb-gvm-regs N:                                See 3.3.     (line   984)
-nopreload:                                    See 3.3.     (line   969)
-O:                                            See 3.4.4.   (line  1419)
-o OUTPUT:                                     See 3.3.     (line   928)
-obj:                                          See 3.3.     (line   932)
-pic:                                          See 3.4.4.   (line  1439)
-pkg-config:                                   See 3.3.     (line   861)
-pkg-config-path:                              See 3.3.     (line   861)
-postlude:                                     See 3.3.     (line   816)
-preload:                                      See 3.3.     (line   969)
-prelude:                                      See 3.3.     (line   808)
-rdynamic:                                     See 3.4.4.   (line  1439)
-report:                                       See 3.3.     (line   881)
-shared:                                       See 3.4.4.   (line  1439)
-target:                                       See 3.3.     (line   805)
-track-scheme:                                 See 3.3.     (line   923)
-uninstall <1>:                                See 2.3.     (line   301)
-uninstall <2>:                                See 2.       (line   149)
-uninstall:                                    See 7.5.     (line  5807)
-update:                                       See 2.       (line   149)
-upgrade <1>:                                  See 7.5.     (line  5807)
-upgrade:                                      See 2.3.     (line   301)
-v <1>:                                        See 2.       (line   149)
-v:                                            See 3.       (line   592)
-verbose:                                      See 3.3.     (line   878)
-warnings:                                     See 3.3.     (line   875)
.c:                                            See 3.3.     (line   618)
.js:                                           See 3.3.     (line   618)
.scm:                                          See 3.3.     (line   618)
.six:                                          See 3.3.     (line   618)
.sld:                                          See 3.3.     (line   618)
<:                                             See 8.1.1.   (line  5976)
<=:                                            See 8.1.1.   (line  5978)
=:                                             See 8.1.1.   (line  5975)
>:                                             See 8.1.1.   (line  5977)
>=:                                            See 8.1.1.   (line  5979)
^C <1>:                                        See 5.1.     (line  2008)
^C:                                            See 4.       (line  1798)
^D:                                            See 5.1.     (line  2034)
___cleanup:                                    See 16.5.    (line 16082)
___setup:                                      See 16.5.    (line 16082)
abandoned-mutex-exception?:                    See 12.4.    (line  9656)
abort:                                         See 12.1.    (line  9355)
absolute path:                                 See 13.1.    (line 10605)
acosh:                                         See 6.4.     (line  4514)
address-info-family:                           See 13.11.   (line 11740)
address-info-protocol:                         See 13.11.   (line 11762)
address-info-socket-info:                      See 13.11.   (line 11773)
address-info-socket-type:                      See 13.11.   (line 11751)
address-info?:                                 See 13.11.   (line 11727)
address-infos:                                 See 13.11.   (line 11661)
all-bits-set?:                                 See 8.1.4.   (line  6306)
allocation-limit:                              See 6.3.     (line  3864)
any-bits-set?:                                 See 8.1.4.   (line  6291)
append-reverse:                                See 6.4.     (line  4496)
append-reverse!:                               See 6.4.     (line  4497)
apropos:                                       See 5.4.     (line  2393)
arithmetic-shift:                              See 8.1.4.   (line  6051)
asinh:                                         See 6.4.     (line  4516)
atanh:                                         See 6.4.     (line  4518)
bit-count:                                     See 8.1.4.   (line  6233)
bit-set?:                                      See 8.1.4.   (line  6279)
bits:                                          See 6.4.     (line  4528)
bits->list:                                    See 6.4.     (line  4529)
bits->vector:                                  See 6.4.     (line  4531)
bitwise-and:                                   See 8.1.4.   (line  6082)
bitwise-andc1:                                 See 8.1.4.   (line  6098)
bitwise-andc2:                                 See 8.1.4.   (line  6110)
bitwise-eqv:                                   See 8.1.4.   (line  6122)
bitwise-ior:                                   See 8.1.4.   (line  6139)
bitwise-merge:                                 See 8.1.4.   (line  6066)
bitwise-nand:                                  See 8.1.4.   (line  6156)
bitwise-nor:                                   See 8.1.4.   (line  6168)
bitwise-not:                                   See 8.1.4.   (line  6180)
bitwise-orc1:                                  See 8.1.4.   (line  6192)
bitwise-orc2:                                  See 8.1.4.   (line  6204)
bitwise-xor:                                   See 8.1.4.   (line  6216)
block:                                         See 6.3.     (line  3828)
box:                                           See 6.3.     (line  3473)
box?:                                          See 6.3.     (line  3474)
boxes:                                         See 6.3.     (line  3476)
break:                                         See 5.4.     (line  2639)
c-declare:                                     See 16.2.    (line 15872)
c-define:                                      See 16.5.    (line 16044)
c-define-type:                                 See 16.6.    (line 16098)
c-initialize:                                  See 16.3.    (line 15915)
c-lambda:                                      See 16.4.    (line 15937)
call-with-current-continuation:                See 6.1.     (line  3022)
call-with-input-file:                          See 14.7.1.  (line 12836)
call-with-input-process:                       See 14.7.2.  (line 12980)
call-with-input-string:                        See 14.10.   (line 13986)
call-with-input-u8vector:                      See 14.11.   (line 14043)
call-with-input-vector:                        See 14.9.    (line 13835)
call-with-output-file:                         See 14.7.1.  (line 12837)
call-with-output-process:                      See 14.7.2.  (line 12981)
call-with-output-string:                       See 14.10.   (line 13987)
call-with-output-u8vector:                     See 14.11.   (line 14044)
call-with-output-vector:                       See 14.9.    (line 13836)
call/cc:                                       See 6.1.     (line  3023)
car+cdr:                                       See 6.4.     (line  4444)
case-lambda:                                   See 6.4.     (line  4433)
central installation directory:                See 13.1.    (line 10605)
cfun-conversion-exception-arguments:           See 12.5.    (line  9830)
cfun-conversion-exception-code:                See 12.5.    (line  9831)
cfun-conversion-exception-message:             See 12.5.    (line  9832)
cfun-conversion-exception-procedure:           See 12.5.    (line  9829)
cfun-conversion-exception?:                    See 12.5.    (line  9828)
char->integer:                                 See 8.6.     (line  6863)
char-ci<=?:                                    See 8.6.     (line  6892)
char-ci<?:                                     See 8.6.     (line  6890)
char-ci=?:                                     See 8.6.     (line  6889)
char-ci>=?:                                    See 8.6.     (line  6893)
char-ci>?:                                     See 8.6.     (line  6891)
char<=?:                                       See 8.6.     (line  6887)
char<?:                                        See 8.6.     (line  6885)
char=?:                                        See 8.6.     (line  6884)
char>=?:                                       See 8.6.     (line  6888)
char>?:                                        See 8.6.     (line  6886)
circular-list:                                 See 6.4.     (line  4484)
circular-list?:                                See 6.4.     (line  4463)
clear-bit-field:                               See 8.1.4.   (line  6337)
close-input-port:                              See 14.4.2.  (line 12268)
close-output-port:                             See 14.4.2.  (line 12269)
close-port:                                    See 14.4.2.  (line 12270)
command-args:                                  See 6.4.     (line  4330)
command-line <1>:                              See 2.6.     (line   442)
command-line:                                  See 13.5.    (line 10996)
command-name:                                  See 6.4.     (line  4329)
compilation-target:                            See 6.3.     (line  3674)
compile-file:                                  See 3.5.     (line  1494)
compile-file-to-target:                        See 3.5.     (line  1453)
compiler:                                      See 3.       (line   574)
compiler options:                              See 3.3.     (line   662)
concatenate:                                   See 6.4.     (line  4471)
concatenate!:                                  See 6.4.     (line  4472)
cond-expand:                                   See 6.3.     (line  3736)
condition-variable-broadcast!:                 See 10.9.    (line  8949)
condition-variable-name:                       See 10.9.    (line  8869)
condition-variable-signal!:                    See 10.9.    (line  8894)
condition-variable-specific:                   See 10.9.    (line  8877)
condition-variable-specific-set!:              See 10.9.    (line  8878)
condition-variable?:                           See 10.9.    (line  8844)
configure-command-string:                      See 6.4.     (line  4390)
conjugate:                                     See 6.4.     (line  4526)
cons*:                                         See 6.4.     (line  4486)
console-port:                                  See 6.4.     (line  4354)
constant-fold:                                 See 6.3.     (line  3879)
continuation-capture:                          See 6.3.     (line  4078)
continuation-graft:                            See 6.3.     (line  4079)
continuation-return:                           See 6.3.     (line  4080)
continuation?:                                 See 6.3.     (line  4077)
continuations:                                 See 16.7.    (line 16558)
copy-bit-field:                                See 8.1.4.   (line  6339)
copy-file:                                     See 13.2.    (line 10878)
cosh:                                          See 6.4.     (line  4520)
cpu-time:                                      See 13.7.    (line 11098)
create-directory:                              See 13.2.    (line 10786)
create-fifo:                                   See 13.2.    (line 10820)
create-link:                                   See 13.2.    (line 10851)
create-symbolic-link:                          See 13.2.    (line 10859)
create-temporary-directory:                    See 13.2.    (line 10787)
current exception-handler:                     See 12.1.    (line  9267)
current working directory:                     See 13.1.    (line 10649)
current-directory:                             See 13.1.    (line 10645)
current-error-port:                            See 14.12.   (line 14079)
current-exception-handler:                     See 12.1.    (line  9267)
current-input-port:                            See 14.12.   (line 14077)
current-jiffy:                                 See 6.4.     (line  4319)
current-output-port:                           See 14.12.   (line 14078)
current-processor:                             See 6.4.     (line  4312)
current-readtable:                             See 14.12.   (line 14080)
current-second:                                See 6.4.     (line  4317)
current-thread:                                See 10.9.    (line  8135)
current-time:                                  See 13.7.    (line 11062)
current-user-interrupt-handler:                See 6.4.     (line  4356)
datum->syntax:                                 See 6.4.     (line  4437)
datum-parsing-exception-kind:                  See 12.6.    (line 10006)
datum-parsing-exception-parameters:            See 12.6.    (line 10007)
datum-parsing-exception-readenv:               See 12.6.    (line 10008)
datum-parsing-exception?:                      See 12.6.    (line 10005)
dead-end:                                      See 6.4.     (line  4416)
deadlock-exception?:                           See 12.4.    (line  9635)
debug <1>:                                     See 3.3.     (line   907)
debug:                                         See 6.3.     (line  3919)
debug-environments <1>:                        See 6.3.     (line  3959)
debug-environments:                            See 3.3.     (line   919)
debug-location <1>:                            See 6.3.     (line  3933)
debug-location:                                See 3.3.     (line   911)
debug-source <1>:                              See 6.3.     (line  3946)
debug-source:                                  See 3.3.     (line   915)
declare:                                       See 6.3.     (line  3816)
default-random-source:                         See 8.1.7.   (line  6626)
default-user-interrupt-handler:                See 6.4.     (line  4357)
defer-user-interrupts:                         See 6.4.     (line  4358)
define:                                        See 6.2.     (line  3032)
define-cond-expand-feature:                    See 6.3.     (line  3799)
define-library:                                See 7.4.2.   (line  5549)
define-macro:                                  See 6.3.     (line  3614)
define-module-alias:                           See 6.4.     (line  4429)
define-record-type:                            See 6.4.     (line  4420)
define-structure:                              See 9.       (line  7856)
define-syntax:                                 See 6.3.     (line  3647)
define-type:                                   See 6.4.     (line  4421)
define-type-of-thread:                         See 6.4.     (line  4282)
define-values:                                 See 6.4.     (line  4427)
delete-directory:                              See 13.2.    (line 10890)
delete-file:                                   See 13.2.    (line 10885)
delete-file-or-directory:                      See 13.2.    (line 10895)
deserialization <1>:                           See 15.1.    (line 14308)
deserialization:                               See 8.9.     (line  7172)
directory-files:                               See 13.2.    (line 10903)
display-continuation-backtrace:                See 6.3.     (line  4162)
display-continuation-dynamic-environment:      See 6.3.     (line  4159)
display-continuation-environment:              See 6.3.     (line  4158)
display-dynamic-environment?:                  See 5.4.     (line  2746)
display-environment-set!:                      See 5.4.     (line  2726)
display-exception:                             See 6.3.     (line  4155)
display-exception-in-context:                  See 6.3.     (line  4156)
display-procedure-environment:                 See 6.3.     (line  4157)
divide-by-zero-exception-arguments:            See 12.8.    (line 10311)
divide-by-zero-exception-procedure:            See 12.8.    (line 10310)
divide-by-zero-exception?:                     See 12.8.    (line 10309)
dotted-list?:                                  See 6.4.     (line  4465)
drop:                                          See 6.4.     (line  4502)
eighth:                                        See 6.4.     (line  4453)
Emacs:                                         See 5.6.     (line  2919)
eq?-hash:                                      See 8.10.1.  (line  7341)
equal?-hash:                                   See 8.10.1.  (line  7365)
eqv?-hash:                                     See 8.10.1.  (line  7353)
err-code->string:                              See 6.4.     (line  4362)
error:                                         See 12.10.   (line 10560)
error-exception-message:                       See 12.10.   (line 10558)
error-exception-parameters:                    See 12.10.   (line 10559)
error-exception?:                              See 12.10.   (line 10557)
eval:                                          See 6.3.     (line  3596)
executable-path:                               See 6.4.     (line  4327)
exit:                                          See 13.4.    (line 10977)
expression-parsing-exception-kind:             See 12.7.    (line 10081)
expression-parsing-exception-parameters:       See 12.7.    (line 10082)
expression-parsing-exception-source:           See 12.7.    (line 10083)
expression-parsing-exception?:                 See 12.7.    (line 10080)
extended-bindings:                             See 6.3.     (line  3887)
extract-bit-field:                             See 8.1.4.   (line  6335)
f32vector:                                     See 8.9.     (line  7117)
f32vector->list:                               See 8.9.     (line  7122)
f32vector-append:                              See 8.9.     (line  7130)
f32vector-concatenate:                         See 8.9.     (line  7126)
f32vector-copy:                                See 8.9.     (line  7127)
f32vector-copy!:                               See 8.9.     (line  7129)
f32vector-fill!:                               See 8.9.     (line  7124)
f32vector-length:                              See 8.9.     (line  7118)
f32vector-ref:                                 See 8.9.     (line  7119)
f32vector-set:                                 See 8.9.     (line  7120)
f32vector-set!:                                See 8.9.     (line  7121)
f32vector-shrink!:                             See 8.9.     (line  7134)
f32vector?:                                    See 8.9.     (line  7115)
f64vector:                                     See 8.9.     (line  7138)
f64vector->list:                               See 8.9.     (line  7143)
f64vector-append:                              See 8.9.     (line  7151)
f64vector-concatenate:                         See 8.9.     (line  7147)
f64vector-copy:                                See 8.9.     (line  7148)
f64vector-copy!:                               See 8.9.     (line  7150)
f64vector-fill!:                               See 8.9.     (line  7145)
f64vector-length:                              See 8.9.     (line  7139)
f64vector-ref:                                 See 8.9.     (line  7140)
f64vector-set:                                 See 8.9.     (line  7141)
f64vector-set!:                                See 8.9.     (line  7142)
f64vector-shrink!:                             See 8.9.     (line  7155)
f64vector?:                                    See 8.9.     (line  7136)
FFI:                                           See 16.      (line 15263)
fifth:                                         See 6.4.     (line  4450)
file names:                                    See 13.1.    (line 10605)
file-attributes:                               See 13.8.    (line 11379)
file-creation-time:                            See 13.8.    (line 11380)
file-device:                                   See 13.8.    (line 11369)
file-exists-exception-arguments:               See 12.3.    (line  9515)
file-exists-exception-procedure:               See 12.3.    (line  9514)
file-exists-exception?:                        See 12.3.    (line  9513)
file-exists?:                                  See 13.8.    (line 11158)
file-group:                                    See 13.8.    (line 11374)
file-info:                                     See 13.8.    (line 11173)
file-info-attributes:                          See 13.8.    (line 11348)
file-info-creation-time:                       See 13.8.    (line 11358)
file-info-device:                              See 13.8.    (line 11254)
file-info-group:                               See 13.8.    (line 11300)
file-info-inode:                               See 13.8.    (line 11263)
file-info-last-access-time:                    See 13.8.    (line 11318)
file-info-last-change-time:                    See 13.8.    (line 11338)
file-info-last-modification-time:              See 13.8.    (line 11328)
file-info-mode:                                See 13.8.    (line 11272)
file-info-number-of-links:                     See 13.8.    (line 11281)
file-info-owner:                               See 13.8.    (line 11291)
file-info-size:                                See 13.8.    (line 11309)
file-info-type:                                See 13.8.    (line 11218)
file-info?:                                    See 13.8.    (line 11206)
file-inode:                                    See 13.8.    (line 11370)
file-last-access-and-modification-times-set!:  See 13.8.    (line 11388)
file-last-access-time:                         See 13.8.    (line 11376)
file-last-change-time:                         See 13.8.    (line 11378)
file-last-modification-time:                   See 13.8.    (line 11377)
file-mode:                                     See 13.8.    (line 11371)
file-number-of-links:                          See 13.8.    (line 11372)
file-owner:                                    See 13.8.    (line 11373)
file-size:                                     See 13.8.    (line 11375)
file-type:                                     See 13.8.    (line 11368)
FILE.c:                                        See 3.3.     (line   618)
FILE.js:                                       See 3.3.     (line   618)
FILE.scm:                                      See 3.3.     (line   618)
FILE.six:                                      See 3.3.     (line   618)
FILE.sld:                                      See 3.3.     (line   618)
filter:                                        See 6.4.     (line  4467)
finite?:                                       See 6.4.     (line  4510)
first:                                         See 6.4.     (line  4446)
first-set-bit:                                 See 8.1.4.   (line  6321)
fixnum:                                        See 6.3.     (line  4026)
fixnum->flonum:                                See 8.1.6.   (line  6512)
fixnum-overflow-exception-arguments:           See 8.1.5.   (line  6479)
fixnum-overflow-exception-procedure:           See 8.1.5.   (line  6478)
fixnum-overflow-exception?:                    See 8.1.5.   (line  6477)
fixnum?:                                       See 8.1.5.   (line  6379)
fl*:                                           See 8.1.6.   (line  6514)
fl+:                                           See 8.1.6.   (line  6516)
fl+*:                                          See 8.1.6.   (line  6591)
fl-:                                           See 8.1.6.   (line  6518)
fl/:                                           See 8.1.6.   (line  6520)
fl<:                                           See 8.1.6.   (line  6522)
fl<=:                                          See 8.1.6.   (line  6524)
fl=:                                           See 8.1.6.   (line  6526)
fl>:                                           See 8.1.6.   (line  6528)
fl>=:                                          See 8.1.6.   (line  6530)
flabs:                                         See 8.1.6.   (line  6532)
flacos:                                        See 8.1.6.   (line  6534)
flacosh:                                       See 8.1.6.   (line  6593)
flasin:                                        See 8.1.6.   (line  6536)
flasinh:                                       See 8.1.6.   (line  6595)
flatan:                                        See 8.1.6.   (line  6538)
flatanh:                                       See 8.1.6.   (line  6597)
flceiling:                                     See 8.1.6.   (line  6541)
flcos:                                         See 8.1.6.   (line  6543)
flcosh:                                        See 8.1.6.   (line  6599)
fldenominator:                                 See 8.1.6.   (line  6545)
fleven?:                                       See 8.1.6.   (line  6547)
flexp:                                         See 8.1.6.   (line  6549)
flexpm1:                                       See 8.1.6.   (line  6601)
flexpt:                                        See 8.1.6.   (line  6551)
flfinite?:                                     See 8.1.6.   (line  6555)
flfloor:                                       See 8.1.6.   (line  6557)
flhypot:                                       See 8.1.6.   (line  6553)
flilogb:                                       See 8.1.6.   (line  6603)
flinfinite?:                                   See 8.1.6.   (line  6559)
flinteger?:                                    See 8.1.6.   (line  6561)
fllog:                                         See 8.1.6.   (line  6563)
fllog1p:                                       See 8.1.6.   (line  6605)
flmax:                                         See 8.1.6.   (line  6565)
flmin:                                         See 8.1.6.   (line  6567)
flnan?:                                        See 8.1.6.   (line  6569)
flnegative?:                                   See 8.1.6.   (line  6571)
flnumerator:                                   See 8.1.6.   (line  6573)
floating point overflow:                       See 17.      (line 16614)
flodd?:                                        See 8.1.6.   (line  6575)
flonum:                                        See 6.3.     (line  4026)
flonum?:                                       See 8.1.6.   (line  6510)
flpositive?:                                   See 8.1.6.   (line  6577)
flround:                                       See 8.1.6.   (line  6579)
flscalbn:                                      See 8.1.6.   (line  6607)
flsin:                                         See 8.1.6.   (line  6581)
flsinh:                                        See 8.1.6.   (line  6609)
flsqrt:                                        See 8.1.6.   (line  6583)
flsquare:                                      See 8.1.6.   (line  6611)
fltan:                                         See 8.1.6.   (line  6585)
fltanh:                                        See 8.1.6.   (line  6613)
fltruncate:                                    See 8.1.6.   (line  6587)
flzero?:                                       See 8.1.6.   (line  6589)
fold:                                          See 6.4.     (line  4479)
fold-right:                                    See 6.4.     (line  4480)
force-output:                                  See 14.4.2.  (line 12230)
foreign function interface:                    See 16.      (line 15263)
foreign-address:                               See 6.4.     (line  4366)
foreign-release!:                              See 6.4.     (line  4367)
foreign-released?:                             See 6.4.     (line  4368)
foreign-tags:                                  See 6.4.     (line  4365)
foreign?:                                      See 6.4.     (line  4364)
fourth:                                        See 6.4.     (line  4449)
future:                                        See 6.4.     (line  4394)
fx*:                                           See 8.1.5.   (line  6381)
fx+:                                           See 8.1.5.   (line  6383)
fx-:                                           See 8.1.5.   (line  6385)
fx<:                                           See 8.1.5.   (line  6387)
fx<=:                                          See 8.1.5.   (line  6389)
fx=:                                           See 8.1.5.   (line  6391)
fx>:                                           See 8.1.5.   (line  6393)
fx>=:                                          See 8.1.5.   (line  6395)
fxabs:                                         See 8.1.5.   (line  6397)
fxand:                                         See 8.1.5.   (line  6399)
fxandc1:                                       See 8.1.5.   (line  6401)
fxandc2:                                       See 8.1.5.   (line  6403)
fxarithmetic-shift:                            See 8.1.5.   (line  6405)
fxarithmetic-shift-left:                       See 8.1.5.   (line  6407)
fxarithmetic-shift-right:                      See 8.1.5.   (line  6409)
fxbit-count:                                   See 8.1.5.   (line  6411)
fxbit-set?:                                    See 8.1.5.   (line  6413)
fxeqv:                                         See 8.1.5.   (line  6415)
fxeven?:                                       See 8.1.5.   (line  6417)
fxfirst-set-bit:                               See 8.1.5.   (line  6419)
fxif:                                          See 8.1.5.   (line  6421)
fxior:                                         See 8.1.5.   (line  6423)
fxlength:                                      See 8.1.5.   (line  6425)
fxmax:                                         See 8.1.5.   (line  6427)
fxmin:                                         See 8.1.5.   (line  6429)
fxmodulo:                                      See 8.1.5.   (line  6431)
fxnand:                                        See 8.1.5.   (line  6435)
fxnegative?:                                   See 8.1.5.   (line  6433)
fxnor:                                         See 8.1.5.   (line  6437)
fxnot:                                         See 8.1.5.   (line  6439)
fxodd?:                                        See 8.1.5.   (line  6441)
fxorc1:                                        See 8.1.5.   (line  6443)
fxorc2:                                        See 8.1.5.   (line  6445)
fxpositive?:                                   See 8.1.5.   (line  6447)
fxquotient:                                    See 8.1.5.   (line  6449)
fxremainder:                                   See 8.1.5.   (line  6451)
fxsquare:                                      See 8.1.5.   (line  6473)
fxwrap*:                                       See 8.1.5.   (line  6453)
fxwrap+:                                       See 8.1.5.   (line  6455)
fxwrap-:                                       See 8.1.5.   (line  6457)
fxwrapabs:                                     See 8.1.5.   (line  6459)
fxwraparithmetic-shift:                        See 8.1.5.   (line  6461)
fxwraparithmetic-shift-left:                   See 8.1.5.   (line  6463)
fxwraplogical-shift-right:                     See 8.1.5.   (line  6465)
fxwrapquotient:                                See 8.1.5.   (line  6467)
fxwrapsquare:                                  See 8.1.5.   (line  6475)
fxxor:                                         See 8.1.5.   (line  6469)
fxzero?:                                       See 8.1.5.   (line  6471)
Gambit <1>:
          See ``Gambit''.                                   (line    29)
Gambit:                                        See 1.       (line    34)
gambit-scheme:                                 See 6.3.     (line  3823)
gambit.el:                                     See 5.6.     (line  2919)
GAMBOPT, environment variable:                 See 4.       (line  1971)
GC:                                            See 5.4.     (line  2783)
gc-report-set!:                                See 5.4.     (line  2783)
generate-proper-tail-calls:                    See 5.4.     (line  2685)
generative-lambda:                             See 6.3.     (line  3986)
generic:                                       See 6.3.     (line  4026)
gensym:                                        See 6.3.     (line  3526)
get-environment-variable:                      See 6.4.     (line  4323)
get-environment-variables:                     See 6.4.     (line  4325)
get-output-string:                             See 14.10.   (line 13992)
get-output-u8vector:                           See 14.11.   (line 14049)
get-output-vector:                             See 14.9.    (line 13961)
getenv:                                        See 13.6.    (line 11017)
group-info:                                    See 13.9.    (line 11424)
group-info-gid:                                See 13.9.    (line 11468)
group-info-members:                            See 13.9.    (line 11478)
group-info-name:                               See 13.9.    (line 11458)
group-info?:                                   See 13.9.    (line 11446)
gsc <1>:                                       See 4.       (line  1690)
gsc <2>:                                       See 3.5.     (line  1623)
gsc <3>:                                       See 3.3.     (line   618)
gsc <4>:                                       See 1.       (line    34)
gsc:                                           See 3.5.     (line  1572)
gsc-script:                                    See 2.6.     (line   425)
gsi <1>:                                       See 1.       (line    34)
gsi <2>:                                       See 2.       (line   149)
gsi:                                           See 4.       (line  1690)
gsi-script:                                    See 2.6.     (line   421)
hashing:                                       See 8.10.    (line  7220)
heap-overflow-exception?:                      See 12.2.    (line  9394)
help:                                          See 5.4.     (line  2357)
help-browser:                                  See 5.4.     (line  2358)
home directory:                                See 13.1.    (line 10605)
Homogeneous numeric vectors:                   See 8.9.     (line  6926)
homogeneous vectors:                           See 15.9.    (line 14873)
host-info:                                     See 13.11.   (line 11596)
host-info-addresses:                           See 13.11.   (line 11648)
host-info-aliases:                             See 13.11.   (line 11638)
host-info-name:                                See 13.11.   (line 11628)
host-info?:                                    See 13.11.   (line 11616)
host-name:                                     See 13.11.   (line 11587)
identity:                                      See 6.3.     (line  3587)
ieee-scheme:                                   See 6.3.     (line  3823)
import <1>:                                    See 6.3.     (line  3614)
import:                                        See 7.2.2.   (line  4860)
inactive-thread-exception-arguments:           See 6.4.     (line  4296)
inactive-thread-exception-procedure:           See 6.4.     (line  4295)
inactive-thread-exception?:                    See 6.4.     (line  4294)
include <1>:                                   See 6.3.     (line  3614)
include:                                       See 7.1.     (line  4727)
infinite?:                                     See 6.4.     (line  4511)
initial current working directory:             See 13.1.    (line 10645)
initial-current-directory:                     See 13.1.    (line 10644)
initialized-thread-exception-arguments:        See 6.4.     (line  4288)
initialized-thread-exception-procedure:        See 6.4.     (line  4287)
initialized-thread-exception?:                 See 6.4.     (line  4286)
inline:                                        See 6.3.     (line  3835)
inline-primitives:                             See 6.3.     (line  3838)
inlining-limit:                                See 6.3.     (line  3842)
input-port-byte-position:                      See 14.7.1.  (line 12915)
input-port-bytes-buffered:                     See 6.4.     (line  4341)
input-port-char-position:                      See 6.4.     (line  4409)
input-port-characters-buffered:                See 6.4.     (line  4343)
input-port-column:                             See 14.5.2.  (line 12379)
input-port-line:                               See 14.5.2.  (line 12378)
input-port-readtable:                          See 14.5.2.  (line 12535)
input-port-readtable-set!:                     See 14.5.2.  (line 12543)
input-port-timeout-set!:                       See 14.4.2.  (line 12295)
input-port?:                                   See 14.4.2.  (line 12139)
installation directories:                      See 13.1.    (line 10605)
integer->char:                                 See 8.6.     (line  6864)
integer-length:                                See 8.1.4.   (line  6256)
integer-nth-root:                              See 8.1.3.   (line  6026)
integer-sqrt:                                  See 8.1.3.   (line  6016)
interpreter <1>:                               See 2.       (line   143)
interpreter:                                   See 3.       (line   574)
interrupts-enabled:                            See 6.3.     (line  3906)
invalid-hash-number-exception-arguments:       See 6.4.     (line  4372)
invalid-hash-number-exception-procedure:       See 6.4.     (line  4371)
invalid-hash-number-exception?:                See 6.4.     (line  4370)
invalid-utf8-encoding-exception-arguments:     See 6.4.     (line  4305)
invalid-utf8-encoding-exception-procedure:     See 6.4.     (line  4304)
invalid-utf8-encoding-exception?:              See 6.4.     (line  4303)
iota:                                          See 6.4.     (line  4482)
jiffies-per-second:                            See 6.4.     (line  4321)
join-timeout-exception-arguments:              See 12.4.    (line  9684)
join-timeout-exception-procedure:              See 12.4.    (line  9683)
join-timeout-exception?:                       See 12.4.    (line  9682)
keyword->string:                               See 6.3.     (line  3502)
keyword-expected-exception-arguments:          See 12.9.    (line 10526)
keyword-expected-exception-procedure:          See 12.9.    (line 10525)
keyword-expected-exception?:                   See 12.9.    (line 10524)
keyword-hash:                                  See 8.10.1.  (line  7305)
keyword?:                                      See 6.3.     (line  3501)
keywords:                                      See 6.3.     (line  3503)
lambda:                                        See 6.2.     (line  3031)
lambda-lift:                                   See 6.3.     (line  3876)
last:                                          See 6.4.     (line  4504)
last-pair:                                     See 6.4.     (line  4505)
LAST_.c:                                       See 3.3.     (line   932)
LAST_.js:                                      See 3.3.     (line   932)
length+:                                       See 6.4.     (line  4442)
length-mismatch-exception-arg-id:              See 12.8.    (line 10341)
length-mismatch-exception-arguments:           See 12.8.    (line 10340)
length-mismatch-exception-procedure:           See 12.8.    (line 10339)
length-mismatch-exception?:                    See 12.8.    (line 10338)
limitations:                                   See 17.      (line 16614)
link-flat:                                     See 3.5.     (line  1623)
link-incremental:                              See 3.5.     (line  1572)
list->bits:                                    See 6.4.     (line  4530)
list->f32vector:                               See 8.9.     (line  7123)
list->f64vector:                               See 8.9.     (line  7144)
list->s16vector:                               See 8.9.     (line  6997)
list->s32vector:                               See 8.9.     (line  7039)
list->s64vector:                               See 8.9.     (line  7081)
list->s8vector:                                See 8.9.     (line  6955)
list->table:                                   See 8.10.3.  (line  7746)
list->u16vector:                               See 8.9.     (line  7018)
list->u32vector:                               See 8.9.     (line  7060)
list->u64vector:                               See 8.9.     (line  7102)
list->u8vector:                                See 8.9.     (line  6976)
list-copy:                                     See 6.4.     (line  4488)
list-set:                                      See 6.4.     (line  4476)
list-set!:                                     See 6.4.     (line  4477)
list-sort:                                     See 6.4.     (line  4507)
list-sort!:                                    See 6.4.     (line  4508)
list-tabulate:                                 See 6.4.     (line  4490)
list=:                                         See 6.4.     (line  4474)
load <1>:                                      See 7.1.     (line  4656)
load:                                          See 3.5.     (line  1494)
mailbox-receive-timeout-exception-arguments:   See 10.9.    (line  8573)
mailbox-receive-timeout-exception-procedure:   See 10.9.    (line  8572)
mailbox-receive-timeout-exception?:            See 10.9.    (line  8571)
main:                                          See 6.4.     (line  4414)
make-condition-variable:                       See 10.9.    (line  8856)
make-f32vector:                                See 8.9.     (line  7116)
make-f64vector:                                See 8.9.     (line  7137)
make-list:                                     See 6.4.     (line  4492)
make-mutex:                                    See 10.9.    (line  8615)
make-parameter:                                See 11.      (line  9068)
make-random-source:                            See 8.1.7.   (line  6687)
make-root-thread:                              See 10.9.    (line  8158)
make-s16vector:                                See 8.9.     (line  6990)
make-s32vector:                                See 8.9.     (line  7032)
make-s64vector:                                See 8.9.     (line  7074)
make-s8vector:                                 See 8.9.     (line  6948)
make-table:                                    See 8.10.3.  (line  7531)
make-thread:                                   See 10.9.    (line  8156)
make-thread-group:                             See 6.4.     (line  4247)
make-tls-context:                              See 14.7.3.  (line 13438)
make-u16vector:                                See 8.9.     (line  7011)
make-u32vector:                                See 8.9.     (line  7053)
make-u64vector:                                See 8.9.     (line  7095)
make-u8vector:                                 See 8.9.     (line  6969)
make-will:                                     See 8.10.2.1.
                                                            (line  7409)
module-not-found-exception-arguments:          See 6.4.     (line  4309)
module-not-found-exception-procedure:          See 6.4.     (line  4308)
module-not-found-exception?:                   See 6.4.     (line  4307)
module-search-order-add!:                      See 7.5.     (line  5720)
module-search-order-reset!:                    See 7.5.     (line  5719)
module-whitelist-add!:                         See 7.5.     (line  5752)
module-whitelist-reset!:                       See 7.5.     (line  5751)
mostly-fixnum:                                 See 6.3.     (line  4032)
mostly-fixnum-flonum:                          See 6.3.     (line  4032)
mostly-flonum:                                 See 6.3.     (line  4032)
mostly-flonum-fixnum:                          See 6.3.     (line  4032)
mostly-generic:                                See 6.3.     (line  4032)
multiple-c-return-exception?:                  See 12.5.    (line  9949)
mutex-lock!:                                   See 10.9.    (line  8698)
mutex-name:                                    See 10.9.    (line  8630)
mutex-specific:                                See 10.9.    (line  8637)
mutex-specific-set!:                           See 10.9.    (line  8638)
mutex-state:                                   See 10.9.    (line  8670)
mutex-unlock!:                                 See 10.9.    (line  8794)
mutex?:                                        See 10.9.    (line  8603)
namespace:                                     See 7.2.2.   (line  4858)
nan?:                                          See 6.4.     (line  4512)
network-info:                                  See 13.14.   (line 11946)
network-info-aliases:                          See 13.14.   (line 11992)
network-info-name:                             See 13.14.   (line 11982)
network-info-number:                           See 13.14.   (line 12002)
network-info?:                                 See 13.14.   (line 11970)
newline:                                       See 14.4.2.  (line 12208)
ninth:                                         See 6.4.     (line  4454)
no-such-file-or-directory-exception-arguments: See 12.3.    (line  9483)
no-such-file-or-directory-exception-procedure: See 12.3.    (line  9482)
no-such-file-or-directory-exception?:          See 12.3.    (line  9481)
noncontinuable-exception-reason:               See 12.1.    (line  9357)
noncontinuable-exception?:                     See 12.1.    (line  9356)
nonempty-input-port-character-buffer-exception-arguments:See 6.4.
                                                            (line  4347)
nonempty-input-port-character-buffer-exception-procedure:See 6.4.
                                                            (line  4349)
nonempty-input-port-character-buffer-exception?:See 6.4.    (line  4345)
nonprocedure-operator-exception-arguments:     See 12.9.    (line 10432)
nonprocedure-operator-exception-code:          See 12.9.    (line 10433)
nonprocedure-operator-exception-operator:      See 12.9.    (line 10431)
nonprocedure-operator-exception-rte:           See 12.9.    (line 10434)
nonprocedure-operator-exception?:              See 12.9.    (line 10430)
normalized path:                               See 13.1.    (line 10706)
not-in-compilation-context-exception-arguments:See 12.7.    (line 10194)
not-in-compilation-context-exception-procedure:See 12.7.    (line 10193)
not-in-compilation-context-exception?:         See 12.7.    (line 10192)
not-pair?:                                     See 6.4.     (line  4457)
null-list?:                                    See 6.4.     (line  4459)
number-of-arguments-limit-exception-arguments: See 12.9.    (line 10398)
number-of-arguments-limit-exception-procedure: See 12.9.    (line 10397)
number-of-arguments-limit-exception?:          See 12.9.    (line 10396)
object file:                                   See 3.5.     (line  1494)
object->serial-number:                         See 8.10.1.  (line  7224)
object->string:                                See 14.10.   (line 14019)
object->u8vector:                              See 8.9.     (line  7171)
open-directory:                                See 14.8.    (line 13768)
open-dummy:                                    See 6.4.     (line  4335)
open-event-queue:                              See 6.4.     (line  4412)
open-file:                                     See 14.7.1.  (line 12833)
open-input-file:                               See 14.7.1.  (line 12834)
open-input-process:                            See 14.7.2.  (line 12978)
open-input-string:                             See 14.10.   (line 13984)
open-input-u8vector:                           See 14.11.   (line 14041)
open-input-vector:                             See 14.9.    (line 13833)
open-output-bytevector:                        See 6.4.     (line  4636)
open-output-file:                              See 14.7.1.  (line 12835)
open-output-process:                           See 14.7.2.  (line 12979)
open-output-string:                            See 14.10.   (line 13985)
open-output-u8vector:                          See 14.11.   (line 14042)
open-output-vector:                            See 14.9.    (line 13834)
open-process:                                  See 14.7.2.  (line 12977)
open-string:                                   See 14.10.   (line 13983)
open-string-pipe:                              See 14.10.   (line 13991)
open-tcp-client:                               See 14.7.3.  (line 13180)
open-tcp-server:                               See 14.7.3.  (line 13293)
open-u8vector:                                 See 14.11.   (line 14040)
open-u8vector-pipe:                            See 14.11.   (line 14048)
open-udp:                                      See 14.7.3.  (line 13550)
open-vector:                                   See 14.9.    (line 13832)
open-vector-pipe:                              See 14.9.    (line 13917)
optimize-dead-definitions:                     See 6.3.     (line  4012)
optimize-dead-local-variables:                 See 6.3.     (line  3998)
options, compiler:                             See 3.3.     (line   662)
options, runtime:                              See 4.       (line  1690)
os-exception-arguments:                        See 12.3.    (line  9443)
os-exception-code:                             See 12.3.    (line  9444)
os-exception-message:                          See 12.3.    (line  9445)
os-exception-procedure:                        See 12.3.    (line  9442)
os-exception?:                                 See 12.3.    (line  9441)
output-port-byte-position:                     See 14.7.1.  (line 12916)
output-port-char-position:                     See 6.4.     (line  4410)
output-port-column:                            See 14.5.2.  (line 12381)
output-port-line:                              See 14.5.2.  (line 12380)
output-port-readtable:                         See 14.5.2.  (line 12536)
output-port-readtable-set!:                    See 14.5.2.  (line 12544)
output-port-timeout-set!:                      See 14.4.2.  (line 12296)
output-port-width:                             See 14.5.2.  (line 12404)
output-port?:                                  See 14.4.2.  (line 12140)
overflow, floating point:                      See 17.      (line 16614)
parameterize:                                  See 11.      (line  9155)
path-directory:                                See 13.1.    (line 10738)
path-expand:                                   See 13.1.    (line 10678)
path-extension:                                See 13.1.    (line 10736)
path-normalize:                                See 13.1.    (line 10706)
path-strip-directory:                          See 13.1.    (line 10739)
path-strip-extension:                          See 13.1.    (line 10737)
path-strip-trailing-directory-separator:       See 13.1.    (line 10740)
path-strip-volume:                             See 13.1.    (line 10742)
path-volume:                                   See 13.1.    (line 10741)
peek-char:                                     See 14.5.2.  (line 12434)
peek-u8:                                       See 14.6.2.  (line 12743)
permission-denied-exception-arguments:         See 12.3.    (line  9550)
permission-denied-exception-procedure:         See 12.3.    (line  9549)
permission-denied-exception?:                  See 12.3.    (line  9548)
poll-on-return:                                See 6.3.     (line  3912)
poll-point:                                    See 6.4.     (line  4418)
port-io-exception-handler-set!:                See 6.4.     (line  4339)
port-settings-set!:                            See 6.4.     (line  4337)
port?:                                         See 14.4.2.  (line 12141)
pp:                                            See 5.4.     (line  2766)
pretty-print:                                  See 5.4.     (line  2753)
primordial-exception-handler:                  See 6.4.     (line  4360)
print:                                         See 14.12.   (line 14086)
println:                                       See 14.12.   (line 14087)
process-pid:                                   See 14.7.2.  (line 13111)
process-status:                                See 14.7.2.  (line 13122)
process-times:                                 See 13.7.    (line 11097)
processor-id:                                  See 6.4.     (line  4313)
processor?:                                    See 6.4.     (line  4311)
proper tail-calls <1>:                         See 6.3.     (line  3973)
proper tail-calls:                             See 5.4.     (line  2685)
proper-list?:                                  See 6.4.     (line  4461)
proper-tail-calls:                             See 6.3.     (line  3973)
protocol-info:                                 See 13.13.   (line 11883)
protocol-info-aliases:                         See 13.13.   (line 11923)
protocol-info-name:                            See 13.13.   (line 11913)
protocol-info-number:                          See 13.13.   (line 11933)
protocol-info?:                                See 13.13.   (line 11901)
r4rs-scheme:                                   See 6.3.     (line  3823)
r5rs-scheme:                                   See 6.3.     (line  3823)
r7rs-guard:                                    See 6.4.     (line  4431)
raise:                                         See 12.1.    (line  9338)
random-f64vector:                              See 8.1.7.   (line  6674)
random-integer:                                See 8.1.7.   (line  6632)
random-real:                                   See 8.1.7.   (line  6648)
random-source-make-f64vectors:                 See 8.1.7.   (line  6818)
random-source-make-integers:                   See 8.1.7.   (line  6765)
random-source-make-reals:                      See 8.1.7.   (line  6782)
random-source-make-u8vectors:                  See 8.1.7.   (line  6801)
random-source-pseudo-randomize!:               See 8.1.7.   (line  6732)
random-source-randomize!:                      See 8.1.7.   (line  6731)
random-source-state-ref:                       See 8.1.7.   (line  6711)
random-source-state-set!:                      See 8.1.7.   (line  6712)
random-source?:                                See 8.1.7.   (line  6699)
random-u8vector:                               See 8.1.7.   (line  6661)
range-exception-arg-id:                        See 12.8.    (line 10276)
range-exception-arguments:                     See 12.8.    (line 10275)
range-exception-procedure:                     See 12.8.    (line 10274)
range-exception?:                              See 12.8.    (line 10273)
read:                                          See 14.4.2.  (line 12163)
read-all:                                      See 14.4.2.  (line 12177)
read-char:                                     See 14.5.2.  (line 12415)
read-file-string:                              See 14.12.   (line 14114)
read-file-string-list:                         See 14.12.   (line 14115)
read-file-u8vector:                            See 14.12.   (line 14116)
read-line:                                     See 14.5.2.  (line 12463)
read-substring:                                See 14.5.2.  (line 12498)
read-subu8vector:                              See 14.6.2.  (line 12785)
read-u8:                                       See 14.6.2.  (line 12742)
readtable-case-conversion?:                    See 15.1.    (line 14217)
readtable-case-conversion?-set:                See 15.1.    (line 14218)
readtable-comment-handler:                     See 6.4.     (line  4633)
readtable-comment-handler-set:                 See 6.4.     (line  4634)
readtable-eval-allowed?:                       See 15.1.    (line 14458)
readtable-eval-allowed?-set:                   See 15.1.    (line 14459)
readtable-keywords-allowed?:                   See 15.1.    (line 14264)
readtable-keywords-allowed?-set:               See 15.1.    (line 14265)
readtable-max-unescaped-char:                  See 15.1.    (line 14604)
readtable-max-unescaped-char-set:              See 15.1.    (line 14605)
readtable-max-write-length:                    See 15.1.    (line 14570)
readtable-max-write-length-set:                See 15.1.    (line 14571)
readtable-max-write-level:                     See 15.1.    (line 14536)
readtable-max-write-level-set:                 See 15.1.    (line 14537)
readtable-sharing-allowed?:                    See 15.1.    (line 14307)
readtable-sharing-allowed?-set:                See 15.1.    (line 14308)
readtable-start-syntax:                        See 15.1.    (line 14645)
readtable-start-syntax-set:                    See 15.1.    (line 14646)
readtable-write-cdr-read-macros?:              See 15.1.    (line 14482)
readtable-write-cdr-read-macros?-set:          See 15.1.    (line 14483)
readtable-write-extended-read-macros?:         See 15.1.    (line 14484)
readtable-write-extended-read-macros?-set:     See 15.1.    (line 14486)
readtable?:                                    See 15.1.    (line 14205)
real-time:                                     See 13.7.    (line 11099)
receive:                                       See 6.4.     (line  4425)
relative path:                                 See 13.1.    (line 10605)
remove:                                        See 6.4.     (line  4468)
remq:                                          See 6.4.     (line  4469)
rename-file:                                   See 13.2.    (line 10867)
repl-display-environment?:                     See 5.4.     (line  2737)
repl-error-port:                               See 6.4.     (line  4353)
repl-input-port:                               See 6.4.     (line  4351)
repl-output-port:                              See 6.4.     (line  4352)
repl-result-history-max-length-set!:           See 5.4.     (line  2449)
repl-result-history-ref:                       See 5.4.     (line  2448)
replace-bit-field:                             See 8.1.4.   (line  6338)
reverse!:                                      See 6.4.     (line  4494)
rpc-remote-error-exception-arguments:          See 6.4.     (line  4300)
rpc-remote-error-exception-message:            See 6.4.     (line  4301)
rpc-remote-error-exception-procedure:          See 6.4.     (line  4299)
rpc-remote-error-exception?:                   See 6.4.     (line  4298)
run-time-bindings:                             See 6.3.     (line  3892)
runtime options:                               See 4.       (line  1690)
s16vector:                                     See 8.9.     (line  6991)
s16vector->list:                               See 8.9.     (line  6996)
s16vector-append:                              See 8.9.     (line  7004)
s16vector-concatenate:                         See 8.9.     (line  7000)
s16vector-copy:                                See 8.9.     (line  7001)
s16vector-copy!:                               See 8.9.     (line  7003)
s16vector-fill!:                               See 8.9.     (line  6998)
s16vector-length:                              See 8.9.     (line  6992)
s16vector-ref:                                 See 8.9.     (line  6993)
s16vector-set:                                 See 8.9.     (line  6994)
s16vector-set!:                                See 8.9.     (line  6995)
s16vector-shrink!:                             See 8.9.     (line  7008)
s16vector?:                                    See 8.9.     (line  6989)
s32vector:                                     See 8.9.     (line  7033)
s32vector->list:                               See 8.9.     (line  7038)
s32vector-append:                              See 8.9.     (line  7046)
s32vector-concatenate:                         See 8.9.     (line  7042)
s32vector-copy:                                See 8.9.     (line  7043)
s32vector-copy!:                               See 8.9.     (line  7045)
s32vector-fill!:                               See 8.9.     (line  7040)
s32vector-length:                              See 8.9.     (line  7034)
s32vector-ref:                                 See 8.9.     (line  7035)
s32vector-set:                                 See 8.9.     (line  7036)
s32vector-set!:                                See 8.9.     (line  7037)
s32vector-shrink!:                             See 8.9.     (line  7050)
s32vector?:                                    See 8.9.     (line  7031)
s64vector:                                     See 8.9.     (line  7075)
s64vector->list:                               See 8.9.     (line  7080)
s64vector-append:                              See 8.9.     (line  7088)
s64vector-concatenate:                         See 8.9.     (line  7084)
s64vector-copy:                                See 8.9.     (line  7085)
s64vector-copy!:                               See 8.9.     (line  7087)
s64vector-fill!:                               See 8.9.     (line  7082)
s64vector-length:                              See 8.9.     (line  7076)
s64vector-ref:                                 See 8.9.     (line  7077)
s64vector-set:                                 See 8.9.     (line  7078)
s64vector-set!:                                See 8.9.     (line  7079)
s64vector-shrink!:                             See 8.9.     (line  7092)
s64vector?:                                    See 8.9.     (line  7073)
s8vector:                                      See 8.9.     (line  6949)
s8vector->list:                                See 8.9.     (line  6954)
s8vector-append:                               See 8.9.     (line  6962)
s8vector-concatenate:                          See 8.9.     (line  6958)
s8vector-copy:                                 See 8.9.     (line  6959)
s8vector-copy!:                                See 8.9.     (line  6961)
s8vector-fill!:                                See 8.9.     (line  6956)
s8vector-length:                               See 8.9.     (line  6950)
s8vector-ref:                                  See 8.9.     (line  6951)
s8vector-set:                                  See 8.9.     (line  6952)
s8vector-set!:                                 See 8.9.     (line  6953)
s8vector-shrink!:                              See 8.9.     (line  6966)
s8vector?:                                     See 8.9.     (line  6947)
safe:                                          See 6.3.     (line  3898)
scheduler-exception-reason:                    See 12.4.    (line  9620)
scheduler-exception?:                          See 12.4.    (line  9619)
Scheme:                                        See 1.       (line    34)
Scheme, implementation of:
          See ``Gambit''.                                   (line    29)
scheme-ieee-1178-1990:                         See 2.6.     (line   412)
scheme-r4rs:                                   See 2.6.     (line   404)
scheme-r5rs:                                   See 2.6.     (line   408)
scheme-srfi-0:                                 See 2.6.     (line   416)
script-directory:                              See 6.4.     (line  4333)
script-file:                                   See 6.4.     (line  4332)
second:                                        See 6.4.     (line  4447)
seconds->time:                                 See 13.7.    (line 11065)
separate:                                      See 6.3.     (line  3828)
serial-number->object:                         See 8.10.1.  (line  7225)
serialization <1>:                             See 15.1.    (line 14308)
serialization:                                 See 8.9.     (line  7172)
service-info:                                  See 13.12.   (line 11802)
service-info-aliases:                          See 13.12.   (line 11850)
service-info-name:                             See 13.12.   (line 11840)
service-info-port-number:                      See 13.12.   (line 11860)
service-info-protocol:                         See 13.12.   (line 11870)
service-info?:                                 See 13.12.   (line 11828)
set-box!:                                      See 6.3.     (line  3476)
setenv:                                        See 13.6.    (line 11018)
seventh:                                       See 6.4.     (line  4452)
sfun-conversion-exception-arguments:           See 12.5.    (line  9893)
sfun-conversion-exception-code:                See 12.5.    (line  9894)
sfun-conversion-exception-message:             See 12.5.    (line  9895)
sfun-conversion-exception-procedure:           See 12.5.    (line  9892)
sfun-conversion-exception?:                    See 12.5.    (line  9891)
shell-command:                                 See 13.3.    (line 10944)
sinh:                                          See 6.4.     (line  4522)
six-script:                                    See 2.6.     (line   429)
six.!:                                         See 6.4.     (line  4535)
six.!x:                                        See 6.4.     (line  4536)
six.&x:                                        See 6.4.     (line  4537)
six.**x:                                       See 6.4.     (line  4538)
six.*x:                                        See 6.4.     (line  4539)
six.++x:                                       See 6.4.     (line  4540)
six.+x:                                        See 6.4.     (line  4541)
six.--x:                                       See 6.4.     (line  4542)
six.-x:                                        See 6.4.     (line  4543)
six.arrow:                                     See 6.4.     (line  4544)
six.asyncx:                                    See 6.4.     (line  4545)
six.awaitx:                                    See 6.4.     (line  4546)
six.break:                                     See 6.4.     (line  4547)
six.call:                                      See 6.4.     (line  4548)
six.case:                                      See 6.4.     (line  4549)
six.clause:                                    See 6.4.     (line  4550)
six.compound:                                  See 6.4.     (line  4551)
six.cons:                                      See 6.4.     (line  4552)
six.continue:                                  See 6.4.     (line  4553)
six.define-procedure:                          See 6.4.     (line  4554)
six.define-variable:                           See 6.4.     (line  4555)
six.do-while:                                  See 6.4.     (line  4556)
six.dot:                                       See 6.4.     (line  4557)
six.for:                                       See 6.4.     (line  4558)
six.from-import:                               See 6.4.     (line  4560)
six.from-import-*:                             See 6.4.     (line  4561)
six.goto:                                      See 6.4.     (line  4559)
six.identifier:                                See 6.4.     (line  4562)
six.if:                                        See 6.4.     (line  4563)
six.import:                                    See 6.4.     (line  4564)
six.index:                                     See 6.4.     (line  4565)
six.infix:                                     See 6.4.     (line  4534)
six.label:                                     See 6.4.     (line  4566)
six.list:                                      See 6.4.     (line  4567)
six.literal:                                   See 6.4.     (line  4568)
six.make-array:                                See 6.4.     (line  4569)
six.new:                                       See 6.4.     (line  4570)
six.notx:                                      See 6.4.     (line  4628)
six.null:                                      See 6.4.     (line  4571)
six.procedure:                                 See 6.4.     (line  4572)
six.procedure-body:                            See 6.4.     (line  4573)
six.return:                                    See 6.4.     (line  4574)
six.switch:                                    See 6.4.     (line  4575)
six.typeofx:                                   See 6.4.     (line  4576)
six.while:                                     See 6.4.     (line  4577)
six.x:                                         See 6.4.     (line  4589)
six.x!==y:                                     See 6.4.     (line  4578)
six.x!=y:                                      See 6.4.     (line  4579)
six.x%=y:                                      See 6.4.     (line  4580)
six.x%y:                                       See 6.4.     (line  4581)
six.x&&y:                                      See 6.4.     (line  4582)
six.x&=y:                                      See 6.4.     (line  4583)
six.x&y:                                       See 6.4.     (line  4584)
six.x**=y:                                     See 6.4.     (line  4585)
six.x**y:                                      See 6.4.     (line  4586)
six.x*=y:                                      See 6.4.     (line  4587)
six.x*y:                                       See 6.4.     (line  4588)
six.x++:                                       See 6.4.     (line  4591)
six.x+=y:                                      See 6.4.     (line  4592)
six.x+y:                                       See 6.4.     (line  4593)
six.x--:                                       See 6.4.     (line  4595)
six.x-=y:                                      See 6.4.     (line  4596)
six.x-y:                                       See 6.4.     (line  4597)
six.x//=y:                                     See 6.4.     (line  4598)
six.x//y:                                      See 6.4.     (line  4599)
six.x/=y:                                      See 6.4.     (line  4600)
six.x/y:                                       See 6.4.     (line  4601)
six.x:-y:                                      See 6.4.     (line  4602)
six.x:=y:                                      See 6.4.     (line  4603)
six.x:y:                                       See 6.4.     (line  4604)
six.x<<=y:                                     See 6.4.     (line  4605)
six.x<<y:                                      See 6.4.     (line  4606)
six.x<=y:                                      See 6.4.     (line  4607)
six.x<y:                                       See 6.4.     (line  4608)
six.x===y:                                     See 6.4.     (line  4609)
six.x==y:                                      See 6.4.     (line  4610)
six.x=y:                                       See 6.4.     (line  4611)
six.x>=y:                                      See 6.4.     (line  4612)
six.x>>=y:                                     See 6.4.     (line  4615)
six.x>>>=y:                                    See 6.4.     (line  4613)
six.x>>>y:                                     See 6.4.     (line  4614)
six.x>>y:                                      See 6.4.     (line  4616)
six.x>y:                                       See 6.4.     (line  4617)
six.x?y:z:                                     See 6.4.     (line  4618)
six.x^=y:                                      See 6.4.     (line  4619)
six.x^y:                                       See 6.4.     (line  4620)
six.xandy:                                     See 6.4.     (line  4624)
six.xinstanceofy:                              See 6.4.     (line  4625)
six.xiny:                                      See 6.4.     (line  4626)
six.xisy:                                      See 6.4.     (line  4627)
six.xory:                                      See 6.4.     (line  4629)
six.yieldx:                                    See 6.4.     (line  4631)
six.~x:                                        See 6.4.     (line  4630)
sixth:                                         See 6.4.     (line  4451)
socket-info-address:                           See 6.4.     (line  4381)
socket-info-family:                            See 6.4.     (line  4382)
socket-info-port-number:                       See 6.4.     (line  4383)
socket-info?:                                  See 6.4.     (line  4380)
stack-overflow-exception?:                     See 12.2.    (line  9416)
standard-bindings:                             See 6.3.     (line  3882)
started-thread-exception-arguments:            See 12.4.    (line  9719)
started-thread-exception-procedure:            See 12.4.    (line  9718)
started-thread-exception?:                     See 12.4.    (line  9717)
step:                                          See 5.4.     (line  2562)
step-level-set!:                               See 5.4.     (line  2563)
string->keyword:                               See 6.3.     (line  3503)
string->uninterned-keyword:                    See 6.3.     (line  3566)
string->uninterned-symbol:                     See 6.3.     (line  3543)
string-ci<=?:                                  See 8.7.     (line  6912)
string-ci<?:                                   See 8.7.     (line  6910)
string-ci=?:                                   See 8.7.     (line  6909)
string-ci=?-hash:                              See 8.10.1.  (line  7329)
string-ci>=?:                                  See 8.7.     (line  6913)
string-ci>?:                                   See 8.7.     (line  6911)
string-concatenate:                            See 6.3.     (line  3412)
string-set:                                    See 6.3.     (line  3398)
string-shrink!:                                See 6.3.     (line  3459)
string<=?:                                     See 8.7.     (line  6907)
string<?:                                      See 8.7.     (line  6905)
string=?:                                      See 8.7.     (line  6904)
string=?-hash:                                 See 8.10.1.  (line  7317)
string>=?:                                     See 8.7.     (line  6908)
string>?:                                      See 8.7.     (line  6906)
subf32vector:                                  See 8.9.     (line  7131)
subf32vector-fill!:                            See 8.9.     (line  7125)
subf32vector-move!:                            See 8.9.     (line  7133)
subf64vector:                                  See 8.9.     (line  7152)
subf64vector-fill!:                            See 8.9.     (line  7146)
subf64vector-move!:                            See 8.9.     (line  7154)
subs16vector:                                  See 8.9.     (line  7005)
subs16vector-fill!:                            See 8.9.     (line  6999)
subs16vector-move!:                            See 8.9.     (line  7007)
subs32vector:                                  See 8.9.     (line  7047)
subs32vector-fill!:                            See 8.9.     (line  7041)
subs32vector-move!:                            See 8.9.     (line  7049)
subs64vector:                                  See 8.9.     (line  7089)
subs64vector-fill!:                            See 8.9.     (line  7083)
subs64vector-move!:                            See 8.9.     (line  7091)
subs8vector:                                   See 8.9.     (line  6963)
subs8vector-fill!:                             See 8.9.     (line  6957)
subs8vector-move!:                             See 8.9.     (line  6965)
substring-fill!:                               See 6.3.     (line  3428)
substring-move!:                               See 6.3.     (line  3443)
subu16vector:                                  See 8.9.     (line  7026)
subu16vector-fill!:                            See 8.9.     (line  7020)
subu16vector-move!:                            See 8.9.     (line  7028)
subu32vector:                                  See 8.9.     (line  7068)
subu32vector-fill!:                            See 8.9.     (line  7062)
subu32vector-move!:                            See 8.9.     (line  7070)
subu64vector:                                  See 8.9.     (line  7110)
subu64vector-fill!:                            See 8.9.     (line  7104)
subu64vector-move!:                            See 8.9.     (line  7112)
subu8vector:                                   See 8.9.     (line  6984)
subu8vector-fill!:                             See 8.9.     (line  6978)
subu8vector-move!:                             See 8.9.     (line  6986)
subvector:                                     See 6.3.     (line  3212)
subvector-fill!:                               See 6.3.     (line  3297)
subvector-move!:                               See 6.3.     (line  3312)
symbol-hash:                                   See 8.10.1.  (line  7293)
syntax:                                        See 6.4.     (line  4435)
syntax->datum:                                 See 6.4.     (line  4438)
syntax->list:                                  See 6.4.     (line  4439)
syntax->vector:                                See 6.4.     (line  4440)
syntax-case <1>:                               See 6.3.     (line  3647)
syntax-case:                                   See 6.4.     (line  4434)
syntax-rules:                                  See 6.3.     (line  3647)
system-stamp:                                  See 6.4.     (line  4392)
system-type:                                   See 6.4.     (line  4388)
system-type-string:                            See 6.4.     (line  4389)
system-version:                                See 6.4.     (line  4385)
system-version-string:                         See 6.4.     (line  4386)
table->list:                                   See 8.10.3.  (line  7727)
table-copy:                                    See 8.10.3.  (line  7795)
table-for-each:                                See 8.10.3.  (line  7698)
table-length:                                  See 8.10.3.  (line  7599)
table-merge:                                   See 8.10.3.  (line  7833)
table-merge!:                                  See 8.10.3.  (line  7812)
table-ref:                                     See 8.10.3.  (line  7624)
table-search:                                  See 8.10.3.  (line  7667)
table-set!:                                    See 8.10.3.  (line  7647)
table?:                                        See 8.10.3.  (line  7587)
tables:                                        See 8.10.    (line  7220)
tail-calls <1>:                                See 6.3.     (line  3973)
tail-calls:                                    See 5.4.     (line  2685)
take:                                          See 6.4.     (line  4501)
tanh:                                          See 6.4.     (line  4524)
tcp-client-local-socket-info:                  See 6.4.     (line  4374)
tcp-client-peer-socket-info:                   See 6.4.     (line  4375)
tcp-client-self-socket-info:                   See 6.4.     (line  4376)
tcp-server-socket-info:                        See 6.4.     (line  4378)
tcp-service-register!:                         See 14.7.3.  (line 13405)
tcp-service-unregister!:                       See 14.7.3.  (line 13407)
tenth:                                         See 6.4.     (line  4455)
terminated-thread-exception-arguments:         See 12.4.    (line  9751)
terminated-thread-exception-procedure:         See 12.4.    (line  9750)
terminated-thread-exception?:                  See 12.4.    (line  9749)
test-bit-field?:                               See 8.1.4.   (line  6336)
third:                                         See 6.4.     (line  4448)
this-source-file:                              See 6.4.     (line  4423)
thread:                                        See 10.9.    (line  8206)
thread-base-priority:                          See 10.9.    (line  8241)
thread-base-priority-set!:                     See 10.9.    (line  8242)
thread-group->thread-group-list:               See 6.4.     (line  4254)
thread-group->thread-group-vector:             See 6.4.     (line  4255)
thread-group->thread-list:                     See 6.4.     (line  4256)
thread-group->thread-vector:                   See 6.4.     (line  4257)
thread-group-name:                             See 6.4.     (line  4249)
thread-group-parent:                           See 6.4.     (line  4250)
thread-group-resume!:                          See 6.4.     (line  4251)
thread-group-specific:                         See 6.4.     (line  4258)
thread-group-specific-set!:                    See 6.4.     (line  4259)
thread-group-suspend!:                         See 6.4.     (line  4252)
thread-group-terminate!:                       See 6.4.     (line  4253)
thread-group?:                                 See 6.4.     (line  4248)
thread-init!:                                  See 6.4.     (line  4284)
thread-interrupt!:                             See 6.4.     (line  4275)
thread-join!:                                  See 10.9.    (line  8442)
thread-mailbox-extract-and-rewind:             See 10.9.    (line  8513)
thread-mailbox-next:                           See 10.9.    (line  8511)
thread-mailbox-rewind:                         See 10.9.    (line  8512)
thread-name:                                   See 10.9.    (line  8219)
thread-priority-boost:                         See 10.9.    (line  8257)
thread-priority-boost-set!:                    See 10.9.    (line  8258)
thread-quantum:                                See 10.9.    (line  8273)
thread-quantum-set!:                           See 10.9.    (line  8274)
thread-receive:                                See 10.9.    (line  8510)
thread-resume!:                                See 6.4.     (line  4278)
thread-send:                                   See 10.9.    (line  8493)
thread-sleep!:                                 See 10.9.    (line  8331)
thread-specific:                               See 10.9.    (line  8226)
thread-specific-set!:                          See 10.9.    (line  8227)
thread-start!:                                 See 10.9.    (line  8293)
thread-state:                                  See 6.4.     (line  4261)
thread-state-abnormally-terminated-reason:     See 6.4.     (line  4272)
thread-state-abnormally-terminated?:           See 6.4.     (line  4271)
thread-state-initialized?:                     See 6.4.     (line  4263)
thread-state-normally-terminated-result:       See 6.4.     (line  4270)
thread-state-normally-terminated?:             See 6.4.     (line  4269)
thread-state-running-processor:                See 6.4.     (line  4265)
thread-state-running?:                         See 6.4.     (line  4264)
thread-state-uninitialized?:                   See 6.4.     (line  4262)
thread-state-waiting-for:                      See 6.4.     (line  4267)
thread-state-waiting-timeout:                  See 6.4.     (line  4268)
thread-state-waiting?:                         See 6.4.     (line  4266)
thread-suspend!:                               See 6.4.     (line  4277)
thread-terminate!:                             See 10.9.    (line  8355)
thread-thread-group:                           See 6.4.     (line  4280)
thread-yield!:                                 See 10.9.    (line  8312)
thread?:                                       See 10.9.    (line  8144)
threads:                                       See 10.      (line  7901)
time:                                          See 13.7.    (line 11126)
time->seconds:                                 See 13.7.    (line 11064)
time?:                                         See 13.7.    (line 11063)
timeout->time:                                 See 6.4.     (line  4315)
top:                                           See 6.4.     (line  4273)
touch:                                         See 6.4.     (line  4395)
trace:                                         See 5.4.     (line  2489)
transcript-off:                                See 6.1.     (line  3018)
transcript-on:                                 See 6.1.     (line  3017)
tty-history:                                   See 6.4.     (line  4398)
tty-history-max-length-set!:                   See 6.4.     (line  4400)
tty-history-set!:                              See 6.4.     (line  4399)
tty-mode-set!:                                 See 6.4.     (line  4403)
tty-paren-balance-duration-set!:               See 6.4.     (line  4401)
tty-text-attributes-set!:                      See 6.4.     (line  4402)
tty-type-set!:                                 See 6.4.     (line  4404)
tty?:                                          See 6.4.     (line  4397)
type-exception-arg-id:                         See 12.8.    (line 10229)
type-exception-arguments:                      See 12.8.    (line 10228)
type-exception-procedure:                      See 12.8.    (line 10227)
type-exception-type-id:                        See 12.8.    (line 10230)
type-exception?:                               See 12.8.    (line 10226)
u16vector:                                     See 8.9.     (line  7012)
u16vector->list:                               See 8.9.     (line  7017)
u16vector-append:                              See 8.9.     (line  7025)
u16vector-concatenate:                         See 8.9.     (line  7021)
u16vector-copy:                                See 8.9.     (line  7022)
u16vector-copy!:                               See 8.9.     (line  7024)
u16vector-fill!:                               See 8.9.     (line  7019)
u16vector-length:                              See 8.9.     (line  7013)
u16vector-ref:                                 See 8.9.     (line  7014)
u16vector-set:                                 See 8.9.     (line  7015)
u16vector-set!:                                See 8.9.     (line  7016)
u16vector-shrink!:                             See 8.9.     (line  7029)
u16vector?:                                    See 8.9.     (line  7010)
u32vector:                                     See 8.9.     (line  7054)
u32vector->list:                               See 8.9.     (line  7059)
u32vector-append:                              See 8.9.     (line  7067)
u32vector-concatenate:                         See 8.9.     (line  7063)
u32vector-copy:                                See 8.9.     (line  7064)
u32vector-copy!:                               See 8.9.     (line  7066)
u32vector-fill!:                               See 8.9.     (line  7061)
u32vector-length:                              See 8.9.     (line  7055)
u32vector-ref:                                 See 8.9.     (line  7056)
u32vector-set:                                 See 8.9.     (line  7057)
u32vector-set!:                                See 8.9.     (line  7058)
u32vector-shrink!:                             See 8.9.     (line  7071)
u32vector?:                                    See 8.9.     (line  7052)
u64vector:                                     See 8.9.     (line  7096)
u64vector->list:                               See 8.9.     (line  7101)
u64vector-append:                              See 8.9.     (line  7109)
u64vector-concatenate:                         See 8.9.     (line  7105)
u64vector-copy:                                See 8.9.     (line  7106)
u64vector-copy!:                               See 8.9.     (line  7108)
u64vector-fill!:                               See 8.9.     (line  7103)
u64vector-length:                              See 8.9.     (line  7097)
u64vector-ref:                                 See 8.9.     (line  7098)
u64vector-set:                                 See 8.9.     (line  7099)
u64vector-set!:                                See 8.9.     (line  7100)
u64vector-shrink!:                             See 8.9.     (line  7113)
u64vector?:                                    See 8.9.     (line  7094)
u8vector:                                      See 8.9.     (line  6970)
u8vector->list:                                See 8.9.     (line  6975)
u8vector->object:                              See 8.9.     (line  7172)
u8vector-append:                               See 8.9.     (line  6983)
u8vector-concatenate:                          See 8.9.     (line  6979)
u8vector-copy:                                 See 8.9.     (line  6980)
u8vector-copy!:                                See 8.9.     (line  6982)
u8vector-fill!:                                See 8.9.     (line  6977)
u8vector-length:                               See 8.9.     (line  6971)
u8vector-ref:                                  See 8.9.     (line  6972)
u8vector-set:                                  See 8.9.     (line  6973)
u8vector-set!:                                 See 8.9.     (line  6974)
u8vector-shrink!:                              See 8.9.     (line  6987)
u8vector?:                                     See 8.9.     (line  6968)
udp-destination-set!:                          See 14.7.3.  (line 13634)
udp-local-socket-info:                         See 14.7.3.  (line 13741)
udp-read-subu8vector:                          See 14.7.3.  (line 13643)
udp-read-u8vector:                             See 14.7.3.  (line 13641)
udp-source-socket-info:                        See 14.7.3.  (line 13742)
udp-write-subu8vector:                         See 14.7.3.  (line 13644)
udp-write-u8vector:                            See 14.7.3.  (line 13642)
unbound-global-exception-code:                 See 12.7.    (line 10168)
unbound-global-exception-rte:                  See 12.7.    (line 10169)
unbound-global-exception-variable:             See 12.7.    (line 10167)
unbound-global-exception?:                     See 12.7.    (line 10166)
unbound-key-exception-arguments:               See 8.10.3.  (line  7767)
unbound-key-exception-procedure:               See 8.10.3.  (line  7766)
unbound-key-exception?:                        See 8.10.3.  (line  7765)
unbound-os-environment-variable-exception-arguments:See 12.3.
                                                            (line  9587)
unbound-os-environment-variable-exception-procedure:See 12.3.
                                                            (line  9586)
unbound-os-environment-variable-exception?:    See 12.3.    (line  9585)
unbound-serial-number-exception-arguments:     See 8.10.1.  (line  7265)
unbound-serial-number-exception-procedure:     See 8.10.1.  (line  7264)
unbound-serial-number-exception?:              See 8.10.1.  (line  7263)
unbox:                                         See 6.3.     (line  3475)
unbreak:                                       See 5.4.     (line  2640)
uncaught-exception-arguments:                  See 12.4.    (line  9786)
uncaught-exception-procedure:                  See 12.4.    (line  9785)
uncaught-exception-reason:                     See 12.4.    (line  9787)
uncaught-exception?:                           See 12.4.    (line  9784)
uninitialized-thread-exception-arguments:      See 6.4.     (line  4292)
uninitialized-thread-exception-procedure:      See 6.4.     (line  4291)
uninitialized-thread-exception?:               See 6.4.     (line  4290)
uninterned-keyword?:                           See 6.3.     (line  3567)
uninterned-symbol?:                            See 6.3.     (line  3544)
unknown-keyword-argument-exception-arguments:  See 12.9.    (line 10495)
unknown-keyword-argument-exception-procedure:  See 12.9.    (line 10494)
unknown-keyword-argument-exception?:           See 12.9.    (line 10493)
unterminated-process-exception-arguments:      See 14.7.2.  (line 13146)
unterminated-process-exception-procedure:      See 14.7.2.  (line 13145)
unterminated-process-exception?:               See 14.7.2.  (line 13144)
untrace:                                       See 5.4.     (line  2490)
user-info:                                     See 13.10.   (line 11500)
user-info-gid:                                 See 13.10.   (line 11555)
user-info-home:                                See 13.10.   (line 11565)
user-info-name:                                See 13.10.   (line 11536)
user-info-shell:                               See 13.10.   (line 11575)
user-info-uid:                                 See 13.10.   (line 11546)
user-info?:                                    See 13.10.   (line 11524)
user-name:                                     See 13.10.   (line 11491)
vector->bits:                                  See 6.4.     (line  4532)
vector-append:                                 See 6.3.     (line  3269)
vector-cas!:                                   See 6.3.     (line  3342)
vector-concatenate:                            See 6.3.     (line  3281)
vector-copy:                                   See 6.3.     (line  3224)
vector-copy!:                                  See 6.3.     (line  3244)
vector-inc!:                                   See 6.3.     (line  3364)
vector-set:                                    See 6.3.     (line  3384)
vector-shrink!:                                See 6.3.     (line  3328)
void:                                          See 6.3.     (line  3591)
weak references:                               See 8.10.    (line  7220)
will-execute!:                                 See 8.10.2.1.
                                                            (line  7412)
will-testator:                                 See 8.10.2.1.
                                                            (line  7411)
will?:                                         See 8.10.2.1.
                                                            (line  7410)
with-exception-catcher:                        See 12.1.    (line  9309)
with-exception-handler:                        See 12.1.    (line  9284)
with-input-from-file:                          See 14.7.1.  (line 12838)
with-input-from-port:                          See 6.4.     (line  4406)
with-input-from-process:                       See 14.7.2.  (line 12982)
with-input-from-string:                        See 14.10.   (line 13988)
with-input-from-u8vector:                      See 14.11.   (line 14045)
with-input-from-vector:                        See 14.9.    (line 13837)
with-output-to-file:                           See 14.7.1.  (line 12839)
with-output-to-port:                           See 6.4.     (line  4407)
with-output-to-process:                        See 14.7.2.  (line 12983)
with-output-to-string:                         See 14.10.   (line 13989)
with-output-to-u8vector:                       See 14.11.   (line 14046)
with-output-to-vector:                         See 14.9.    (line 13838)
write:                                         See 14.4.2.  (line 12197)
write-char:                                    See 14.5.2.  (line 12450)
write-file-string:                             See 14.12.   (line 14148)
write-file-string-list:                        See 14.12.   (line 14149)
write-file-u8vector:                           See 14.12.   (line 14150)
write-substring:                               See 14.5.2.  (line 12499)
write-subu8vector:                             See 14.6.2.  (line 12786)
write-u8:                                      See 14.6.2.  (line 12773)
wrong-number-of-arguments-exception-arguments: See 12.9.    (line 10368)
wrong-number-of-arguments-exception-procedure: See 12.9.    (line 10367)
wrong-number-of-arguments-exception?:          See 12.9.    (line 10366)
wrong-number-of-values-exception-code:         See 12.9.    (line 10465)
wrong-number-of-values-exception-rte:          See 12.9.    (line 10466)
wrong-number-of-values-exception-vals:         See 12.9.    (line 10464)
wrong-number-of-values-exception?:             See 12.9.    (line 10463)
wrong-processor-c-return-exception?:           See 12.5.    (line  9992)
xcons:                                         See 6.4.     (line  4499)
|six.x:                                        See 6.4.     (line  4594)
|six.x\|=y|:                                   See 6.4.     (line  4621)
|six.x\|\|y|:                                  See 6.4.     (line  4623)
|six.x\|y|:                                    See 6.4.     (line  4622)
~:                                             See 13.1.    (line 10629)
~username:                                     See 13.1.    (line 10638)
~~:                                            See 13.1.    (line 10619)
Table of Contents
*****************

Gambit
1 The Gambit system
  1.1 Accessing the system files
2 The Gambit Scheme interpreter
  2.1 Interactive mode
  2.2 Batch mode
  2.3 Module management mode
  2.4 Customization
  2.5 Process exit status
  2.6 Scheme scripts
    2.6.1 Scripts under UNIX and macOS
    2.6.2 Scripts under Microsoft Windows
    2.6.3 Compiling scripts
3 The Gambit Scheme compiler
  3.1 Interactive mode
  3.2 Customization
  3.3 Batch mode
  3.4 Link files
    3.4.1 Building an executable program
    3.4.2 Building a loadable library
    3.4.3 Building a shared-library
    3.4.4 Other compilation options
  3.5 Procedures specific to compiler
4 Runtime options
5 Debugging
  5.1 Debugging model
  5.2 Debugging commands
  5.3 Debugging example
  5.4 Procedures related to debugging
  5.5 Console line-editing
  5.6 Emacs interface
  5.7 GUIDE
6 Scheme extensions
  6.1 Extensions to standard procedures
  6.2 Extensions to standard special forms
  6.3 Miscellaneous extensions
  6.4 Undocumented extensions
7 Modules
  7.1 Legacy Modules
  7.2 Primitive Modules
    7.2.1 `##demand-module' and `##supply-module' forms
    7.2.2 `##namespace' and `##import' forms
    7.2.3 Macros
  7.3 Primitive Procedures
    7.3.1 Type specifiers
      7.3.1.1 Basic types (other than numbers)
      7.3.1.2 Numbers
      7.3.1.3 Time types
      7.3.1.4 Ports
      7.3.1.5 List and vector variants of above
      7.3.1.6 Gambit types
      7.3.1.7 Others
  7.4 R7RS Compatible Modules
    7.4.1 Identifying libraries
    7.4.2 The `define-library' form
    7.4.3 `(export <EXPORT SPEC> ...)'
    7.4.4 `(import <IMPORT SET> ...)'
    7.4.5 `(begin <COMMAND OR DEFINITION> ...)', `(include <FILENAME> ...)', and `(include-ci <FILENAME> ...)'
    7.4.6 `(include-library-declarations <FILENAME> ...)'
    7.4.7 `(cond-expand <COND EXPAND FEATURES> ...)'
    7.4.8 Extensions to the R7RS library declarations
  7.5 Installing Modules
  7.6 Compiling Modules
8 Built-in data types
  8.1 Numbers
    8.1.1 Extensions to numeric procedures
    8.1.2 IEEE floating point arithmetic
    8.1.3 Integer square root and nth root
    8.1.4 Bitwise-operations on exact integers
    8.1.5 Fixnum specific operations
    8.1.6 Flonum specific operations
    8.1.7 Pseudo random numbers
  8.2 Booleans
  8.3 Pairs and lists
  8.4 Symbols and keywords
  8.5 Characters and strings
  8.6 Extensions to character procedures
  8.7 Extensions to string procedures
  8.8 Vectors
  8.9 Homogeneous numeric vectors
  8.10 Hashing and weak references
    8.10.1 Hashing
    8.10.2 Weak references
      8.10.2.1 Wills
    8.10.3 Tables
9 Records
10 Threads
  10.1 Introduction
  10.2 Thread objects
  10.3 Mutex objects
  10.4 Condition variable objects
  10.5 Fairness
  10.6 Memory coherency
  10.7 Timeouts
  10.8 Primordial thread
  10.9 Procedures
11 Dynamic environment
12 Exceptions
  12.1 Exception-handling
  12.2 Exception objects related to memory management
  12.3 Exception objects related to the host environment
  12.4 Exception objects related to threads
  12.5 Exception objects related to C-interface
  12.6 Exception objects related to the reader
  12.7 Exception objects related to evaluation and compilation
  12.8 Exception objects related to type checking
  12.9 Exception objects related to procedure call
  12.10 Other exception objects
13 Host environment
  13.1 Handling of file names
  13.2 Filesystem operations
  13.3 Shell command execution
  13.4 Process termination
  13.5 Command line arguments
  13.6 Environment variables
  13.7 Measuring time
  13.8 File information
  13.9 Group information
  13.10 User information
  13.11 Host information
  13.12 Service information
  13.13 Protocol information
  13.14 Network information
14 I/O and ports
  14.1 Unidirectional and bidirectional ports
  14.2 Port classes
  14.3 Port settings
  14.4 Object-ports
    14.4.1 Object-port settings
    14.4.2 Object-port operations
  14.5 Character-ports
    14.5.1 Character-port settings
    14.5.2 Character-port operations
  14.6 Byte-ports
    14.6.1 Byte-port settings
    14.6.2 Byte-port operations
  14.7 Device-ports
    14.7.1 Filesystem devices
    14.7.2 Process devices
    14.7.3 Network devices
  14.8 Directory-ports
  14.9 Vector-ports
  14.10 String-ports
  14.11 U8vector-ports
  14.12 Other procedures related to I/O
15 Lexical syntax and readtables
  15.1 Readtables
  15.2 Boolean syntax
  15.3 Character syntax
  15.4 String syntax
  15.5 Symbol syntax
  15.6 Keyword syntax
  15.7 Box syntax
  15.8 Number syntax
  15.9 Homogeneous vector syntax
  15.10 Special `#!' syntax
  15.11 Multiline comment syntax
  15.12 Scheme infix syntax extension
    15.12.1 SIX grammar
    15.12.2 SIX semantics
16 C-interface
  16.1 The mapping of types between C and Scheme
  16.2 The `c-declare' special form
  16.3 The `c-initialize' special form
  16.4 The `c-lambda' special form
  16.5 The `c-define' special form
  16.6 The `c-define-type' special form
  16.7 Continuations, the C-interface and threads
17 System limitations
18 Copyright and license
General index


