Gambit-C
This manual documents Gambit-C. It covers release v4.6.6.
1. The Gambit-C system
The Gambit programming system is a full implementation of the Scheme
language which conforms to the R4RS, R5RS and IEEE Scheme standards. It
consists of two main programs: gsi
, the Gambit Scheme
interpreter, and gsc
, the Gambit Scheme compiler.
Gambit-C is a version of the Gambit programming system in which the
compiler generates portable C code, making the whole Gambit-C system and
the programs compiled with it easily portable to many computer
architectures for which a C compiler is available. With appropriate
declarations in the source code the executable programs generated by
the compiler run roughly as fast as equivalent C programs.
For the most up to date information on Gambit and add-on packages
please check the Gambit web page at
http://gambit.iro.umontreal.ca. The web page has links to the
Gambit mailing list, the bug reporting system, and the source code
repository.
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-C
under UNIX,
/Library/Gambit-C
under Mac OS X and C:/Program
Files/Gambit-C
under Microsoft Windows. This may have been
overridden when the system was built with the command ‘configure
--prefix=/my/Gambit-C’. 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. v4.6.6
for Gambit v4.6.6). 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 Mac OS X 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-C, 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 Mac OS X, 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-C/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-C
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-C/lib/libgambc.a /usr/lib # name may vary
$ ln -s /usr/local/Gambit-C/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-C/lib
$ export CPATH=/usr/local/Gambit-C/include
|
Note that this may have been done by the installation process. In
particular, the Mac OS X 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] [-v] [[-] [-e expressions] [file]]…
|
The interpreter is executed in interactive mode when no file or
‘-’ or ‘-e’ option is given on the command line. Otherwise
the interpreter is executed in batch mode. The ‘-i’ option
is ignored by the interpreter. The initialization file will be
examined unless the ‘-f’ option is present (see section Customization). 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 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
Debugging). 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. When
the system’s runtime library was built with support for GUIDE, the
Gambit Universal IDE (with the command ‘configure
--enable-guide’) the interaction channel corresponds to the
console window of the primordial thread (for details see
GUIDE), otherwise 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 the Scheme standard defines as returning an
unspecified value (e.g. write
, set!
, define
).
Here is a sample interaction with gsi
:
| $ gsi
Gambit v4.6.6
> (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 Debugging.
2.2 Batch mode
In batch mode the command line arguments denote files to be loaded,
REPL interactions to start (‘-’ option), and expressions to be
evaluated (‘-e’ option). Note that the ‘-’ and ‘-e’
options can be interspersed with the files on the command line and can
occur multiple times. The interpreter processes the command line
arguments from left to right, loading files with the load
procedure and evaluating expressions with the eval
procedure in
the global interaction environment. After this processing the
interpreter exits.
When the file name 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 name is obtained by adding to the
file name a ‘.on’ extension with the highest consecutive
version number starting with 1. The source file’s name is obtained by
adding to the file name the file extensions ‘.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 the file name has an extension, the load
procedure will only attempt to load the file with that specific name.
When the extension of the file loaded is ‘.scm’ the content of
the file will be parsed using the normal Scheme prefix syntax. When
the extension of the file loaded is ‘.six’ the content of the
file will be parsed using the Scheme infix syntax extension (see
Scheme infix syntax extension). Otherwise, gsi
will
parse the file using the normal Scheme prefix syntax.
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 (console or IDE’s console
window) 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 0)
(#\d #\l #\r #\o #\w)
1
2
|
2.3 Customization
There are two ways to customize the interpreter. When the interpreter
starts off it tries to execute a ‘(load "~~lib/gambcext")’ (for an
explanation of how file names are interpreted see 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: ‘.gambcini’ and ‘~/.gambcini’ (with no 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.
2.4 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 ‘GAMBCOPT’
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 -:d0 -e "(pretty-print (expt 2 100))"
1267650600228229401496703205376
$ echo $?
0
$ gsi -:d0,unknown # try to use an unknown runtime option
$ echo $?
64
$ gsi -:d0 nonexistent.scm # try to load a file that does not exist
$ echo $?
70
$ gsi nonexistent.scm
*** ERROR IN ##main -- No such file or directory
(load "nonexistent.scm")
$ echo $?
70
|
| $ gsi -:m4000000 # ask for a 4 gigabyte heap
*** malloc: vm_allocate(size=528384) failed (error code=3)
*** malloc[15068]: error: Can't allocate region
$ echo $?
71
|
Note the use of the runtime option ‘-:d0’ that prevents error
messages from being output, and the runtime option ‘-:m4000000’
which sets the minimum heap size to 4 gigabytes.
2.5 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-C 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).
2.5.1 Scripts under UNIX and Mac OS X
Under UNIX and Mac OS X, the Gambit-C 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-C
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-C/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-C/bin/six-script
void main (obj n_str)
{
int n = \string->number(n_str);
for (int 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)))
|
2.5.2 Scripts under Microsoft Windows
Under Microsoft Windows, the Gambit-C 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.
2.5.3 Compiling scripts
A script file can be compiled using the Gambit Scheme compiler
(see section The Gambit Scheme compiler) 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-C/bin/gsi-script -:d0
(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 -:d1 1 2 3 # ask for error message
*** ERROR -- Wrong number of arguments passed to procedure
(main "1" "2" "3")
|
3. The Gambit Scheme compiler
Synopsis:
| gsc [-:runtimeoption,…] [-i] [-f] [-v]
[-prelude expressions] [-postlude expressions]
[-dynamic] [-exe] [-obj] [-cc-options options]
[-ld-options-prelude options] [-ld-options options]
[-warnings] [-verbose] [-report] [-expansion] [-gvm]
[-debug] [-debug-location] [-debug-source]
[-debug-environments] [-track-scheme]
[-o output] [-c] [-keep-c] [-link] [-flat] [-l base]
[[-] [-e expressions] [file]]…
|
3.1 Interactive mode
When no command line argument is present other than options the
compiler behaves like the interpreter in interactive mode. 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).
3.2 Customization
Like the interpreter, the compiler will examine the initialization
file unless the ‘-f’ option is specified.
3.3 Batch mode
In batch mode gsc
takes a set of file names (with either no
extension, or a C file extension, or some other extension) on the
command line and compiles each Scheme file into a C file.
The recognized C file extensions are ‘.c’, ‘.C’, ‘.cc’,
‘.cp’, ‘.cpp’, ‘.CPP’, ‘.cxx’, ‘.c++’,
‘.m’, ‘.M’, and ‘.mm’.
The extension can be omitted from file when the Scheme file has a
‘.scm’ 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 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’ option,
and are used by Gambit’s linker.
For each Scheme file a C file ‘file.c’ will be produced.
The C file’s name is the same as the Scheme file, but the extension is
changed to ‘.c’. By default the C file is created in the same
directory as the Scheme file. This default can be overridden with the
compiler’s ‘-o’ option.
The C files produced by the compiler serve two purposes. They will be
processed by a C compiler to generate object files, and they also
contain information to be read by Gambit’s linker to generate a
link file. The link file is a C file 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 (see section 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.
-
-v
Print the system version string, system time stamp, operating system
type, and configure script options on standard output and exit.
-
-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-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.
-
-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.
-
-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 into an executable program.
-
-obj
Compile Scheme source files to object files.
-
-keep-c
Keep any intermediate ‘.c’ files that are generated.
-
-c
Compile Scheme source files to C without generating link file.
-
-link
Compile Scheme source files to C 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.
-
-
Start REPL interaction.
-
-e expressions
Evaluate expressions in the interaction environment.
The ‘-i’ option forces the compiler to process the remaining
command line arguments like the interpreter.
The ‘-prelude’ option adds the specified expressions to the top of
the source code being compiled. 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. 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-options’ option is only meaningful when a dynamically
loadable object file is being generated (neither the ‘-c’ or
‘-link’ options are used). 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 a dynamically loadable object file is being generated
(neither the ‘-c’ or ‘-link’ options are used). 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" bench.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 ‘-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’ options 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
C files specified on the command line or produced by the Gambit
compiler. By default the link file is ‘last_.c’, where
‘last.c’ is the last file in the set of C files. When the
‘-c’ option is specified, the Scheme source files are compiled to
C files. When the ‘-exe’ option is specified, the generated C
files and link file are compiled and linked using the C compiler to
produce an executable program whose name defaults to
‘last.exe’. When 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’, ‘-obj’ options appear on
the command line, the Scheme source files are compiled to dynamically
loadable object files (‘.on’ extension). The
‘-keep-c’ option will prevent the deletion of any intermediate
‘.c’ file that is generated. Note that in this case the
intermediate ‘.c’ 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’
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/_gambc.c’).
The ‘-’ option starts a REPL interaction.
The ‘-e’ option evaluates the specified expressions in the
interaction environment.
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 C
modules. 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 C 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 C files, generating the link file, compiling
the C files generated to object files, and creating the final
executable file using the C linker. The following example shows how
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
‘GAMBC_CC_VERBOSE’ environment variable to a nonnull value. For
example:
| $ export GAMBC_CC_VERBOSE=yes
$ gsc -o hello.exe -exe h.scm w.six
h.scm:
/Users/feeley/gambit/doc/h.c:
gcc -no-cpp-precomp -Wno-unused -O1 -fno-math-errno -fschedule-insns2
-fno-trapping-math -fno-strict-aliasing -fwrapv -fomit-frame-pointer
-fPIC -fno-common -mieee-fp -I"/usr/local/Gambit-C/include" -c -o "h.o" h.c
w.six:
/Users/feeley/gambit/doc/w.c:
gcc -no-cpp-precomp -Wno-unused -O1 -fno-math-errno -fschedule-insns2
-fno-trapping-math -fno-strict-aliasing -fwrapv -fomit-frame-pointer
-fPIC -fno-common -mieee-fp -I"/usr/local/Gambit-C/include" -c -o "w.o" w.c
/Users/feeley/gambit/doc/w_.c:
gcc -no-cpp-precomp -Wno-unused -O1 -fno-math-errno -fschedule-insns2
-fno-trapping-math -fno-strict-aliasing -fwrapv -fomit-frame-pointer
-fPIC -fno-common -mieee-fp -I"/usr/local/Gambit-C/include" -c -o "w_.o" w_.c
gcc -no-cpp-precomp -Wno-unused -O1 -fno-math-errno -fschedule-insns2
-fno-trapping-math -fno-strict-aliasing -fwrapv -fomit-frame-pointer
-fPIC -fno-common -mieee-fp -I"/usr/local/Gambit-C/include"
-o "hello.exe" h.o w.o w_.o "/usr/local/Gambit-C/lib/libgambc.a"
|
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 seperately 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 by seperating 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 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-C
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-C/include
-L/usr/local/Gambit-C/lib’ to access the ‘gambit.h’ header file
and the Gambit-C runtime library).
Here is an example under Mac OS X:
| $ uname -srmp
Darwin 8.1.0 Power Macintosh powerpc
$ 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 -lgambc
$ ./a.out
((2 . 2) (4 . 4) (8 . 8) (16 . 16))
|
Here is an example under Linux:
| $ uname -srmp
Linux 2.6.8-1.521 i686 athlon
$ 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 -lgambc -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 Mac OS X:
| $ uname -srmp
Darwin 10.5.0 i386 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 2.6.8-1.521 i686 athlon
$ 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 bailey 1.2.13 #2 Wed Aug 28 16:29:41 GMT 1996 i586
$ 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 -lgambc
$ 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 ‘-O’
option can be passed to the C compiler. For large modules, it will
not be practical to specify both ‘-O’ 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-C/include’ for header files so the flag
‘-I/usr/local/Gambit-C/include’ should be passed to the C
compiler. Similarly, C compilers/linkers will not automatically
search ‘/usr/local/Gambit-C/lib’ for libraries so the flag
‘-L/usr/local/Gambit-C/lib’ should be passed to the C
compiler/linker. Alternatives are given in 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.
The file parameter must be a string naming an existing file
containing Scheme source code. The extension can be omitted from
file when the Scheme file has a ‘.scm’ or ‘.six’
extension. This procedure compiles the source file into a file
containing C code. By default, this file is named after file
with the extension replaced with ‘.c’. The name of the generated
file can be specified with the output parameter. If
output is a string naming a directory then the C file is created
in that directory. Otherwise the name of the C file is output.
The name of the generated module can be specified with the
module-name parameter. If module-name is #f
or is
not specified, then the name of the module is derived from the name of
the C file generated, without the extension.
Compilation options are specified through the options parameter
which must be a list of symbols. Any combination of the following
options can be used: ‘verbose’, ‘report’, ‘expansion’,
‘gvm’, and ‘debug’.
When the compilation is successful, compile-file-to-target
returns
the name of the C file generated. When there is a compilation error,
#f
is returned.
| $ cat h.scm
(display "hello") (newline)
$ gsc
Gambit v4.6.6
> (compile-file-to-target "h")
"/Users/feeley/gambit/doc/h.c"
|
( compile-file file [options: options] [output: output] [cc-options: cc-options] [ld-options-prelude: ld-options-prelude] [ld-options: ld-options]) | procedure |
The file, options, and output parameters have the
same meaning as for the compile-file-to-target
procedure, except that
file may be a Scheme source file or a
C 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 v4.6.6
> (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"
|
( link-incremental module-list [output: output] [base: base] [warnings?: warnings?]) | procedure |
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_.c’, where
last is the name of the last module, without the file extension.
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/_gambc.c’. 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 v4.6.6
> (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 -lgambc -o hello
$ ./hello
hello
world
|
( link-flat module-list [output: output] [warnings?: warnings?]) | procedure |
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_.c’, where
last is the name of the last module. 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 v4.6.6
> (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 v4.6.6
> (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:
-
mHEAPSIZE
Set minimum heap size in kilobytes.
-
hHEAPSIZE
Set maximum heap size in kilobytes.
-
lLIVEPERCENT
Set heap occupation after garbage collection.
-
s
Select standard Scheme mode.
-
S
Select Gambit Scheme mode.
-
d[OPT...]
Set debugging options.
-
@[INTF][:PORT]
Override the configuration of the main RPC server.
-
=DIRECTORY
Override the central installation directory.
-
~~DIR=DIRECTORY
Override the DIR installation directory.
-
+ARGUMENT
Add ARGUMENT to the command line before other arguments.
-
f[OPT...]
Set file options.
-
t[OPT...]
Set terminal options.
-
-[OPT...]
Set standard input and output options.
The ‘m’ option specifies the minimum size of the heap. The
‘m’ is immediately followed by an integer indicating the number
of kilobytes of memory. The heap will not shrink lower than this
size. By default, the minimum size is 0.
The ‘h’ option specifies the maximum size of the heap. The
‘h’ is immediately followed by an integer indicating the number
of kilobytes of memory. The heap will not grow larger than this size.
By default, there is no limit (i.e. the heap will grow until the
virtual memory is exhausted).
The ‘l’ option specifies the percentage of the heap that will be
occupied with live objects after the heap is resized at the end of a
garbage collection. The ‘l’ is immediately followed by an
integer between 1 and 100 inclusively indicating the desired
percentage. The garbage collector resizes the heap to reach this
percentage occupation. By default, the percentage is 50.
The ‘s’ option selects standard Scheme mode. In this mode the
reader is case-insensitive and keywords are not recognized. The
‘S’ option selects Gambit Scheme mode (the reader is case-sensitive
and recognizes keywords which end with a colon). By default Gambit
Scheme mode is used.
The ‘d’ option sets various debugging options. The letter
‘d’ 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
current-user-interrupt-handler
is 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.
-
i
-
The REPL interaction channel will be the IDE REPL window (if the IDE
is available).
-
c
-
The REPL interaction channel will be the console.
-
-
-
The REPL interaction channel will be standard input and standard output.
-
@[HOST][:PORT]
-
The REPL interaction channel will be connected to the remote debugger
at address HOST:PORT (if there is a remote debugger at
that address). The default HOST is 127.0.0.1 and the default
PORT is 44555.
THIS OPTION IS NOT YET IMPLEMENTED!
The default debugging options are equivalent to -:dpqQ1i
(i.e. an uncaught exception in the primordial thread terminates the
program after displaying an error message). When the letter ‘d’
is not followed by suboptions, it is equivalent to -:dprR1i
(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 ‘@[INTF][:PORT]’ option
overrides the configuration of the main RPC server. The default
INTF is 127.0.0.1 and the default PORT is 44556.
THIS OPTION IS NOT YET IMPLEMENTED!
The ‘=DIRECTORY’ option overrides the setting of the
central installation directory.
The ‘~~DIR=DIRECTORY’ option overrides the setting of
the DIR installation directory.
The ‘+’ option adds the text that follows to the command line
before other arguments.
The ‘f’, ‘t’ and ‘-’ options specify the default
settings of the ports created for files, terminals and standard input
and output respectively. 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. The ‘f’,
‘t’ and ‘-’ must be followed by a sequence of these options:
-
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.
-
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).
When a program’s execution starts, the runtime system obtains the
runtime options by processing in turn four sources of runtime options:
the defaults, the environment variable ‘GAMBCOPT’, the script
line of the source code, and 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 ‘GAMBCOPT’ 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).
For example:
| $ GAMBCOPT=d0,=~/my-gambit2
$ export GAMBCOPT
$ gsi -e '(pretty-print (path-expand "~~")) (/ 1 0)'
"/Users/feeley/my-gambit2/"
$ echo $?
70
$ gsi -:d1 -e '(pretty-print (path-expand "~~")) (/ 1 0)'
"/Users/feeley/my-gambit2/"
*** ERROR IN (string)@1.3 -- 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:
- An evaluation error has occured, such as attempting to
divide by zero.
- The user has interrupted the evaluation (usually by typing <^C>).
- A breakpoint has been reached or
(step)
was evaluated.
- 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:
-
,?
-
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 v4.6.6
> (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
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. 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/gambc-doc.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
>
|
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>
|
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
|
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:
-
procedure call
-
delay
special form and operations at lower levels
-
lambda
special form and operations at lower levels
-
define
special form and operations at lower levels
-
set!
special form and operations at lower levels
-
variable reference and operations at lower levels
-
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
|
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
|
[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...
|
[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.
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.
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.
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)))
|
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))))
|
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
‘~/.gambc_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 Mac OS X, 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 -:d-")
|
Alternatively, if you don’t mind always loading this package,
you can simply add this line to your ‘.emacs’ file:
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-C/share/emacs/site-lisp"
load-path))
(setq scheme-program-name "/tmp/gsi -:d-") ; 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
These procedures do nothing.
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
( lambda lambda-formals body) | special form |
( define (variable define-formals) body) | special form |
-
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:
-
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.
-
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.
-
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.
-
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.
-
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.
-
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.
-
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
This procedure returns a newly allocated vector with the same content
as the vector vector. Note that the elements are not
recursively copied.
For example:
| > (define v1 '#(1 2 3))
> (define v2 (vector-copy v1))
> v2
#(1 2 3)
> (eq? v1 v2)
#f
|
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)
|
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)
|
This procedure returns a newly allocated vector whose elements form
the concatenation of all the vectors in the list lst. It is
equivalent to (apply vector-append lst)
.
For example:
| > (define v '#(1 2 3))
> (append-vectors (list v v v))
#(1 2 3 1 2 3 1 2 3)
|
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)
|
( subvector-move! src-vector src-start src-end dst-vector dst-start) | procedure |
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)
|
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)
|
This procedure returns a newly allocated string whose elements form
the concatenation of all the strings in the list lst. It is
equivalent to (apply string-append lst)
.
For example:
| > (define s "abc")
> (append-strings (list s s s))
"abcabcabc"
|
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"
|
( substring-move! src-string src-start src-end dst-string dst-start) | procedure |
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"
|
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"
|
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
(see section 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
|
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
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:
|
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
|
The procedure make-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
> (make-uninterned-symbol "foo")
#:foo:
> (uninterned-symbol? (make-uninterned-symbol "foo"))
#t
> (uninterned-symbol? 'hello)
#f
> (uninterned-symbol? 123)
#f
|
The procedure make-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:
| > (make-uninterned-keyword "foo")
#:foo:
> (uninterned-keyword? (make-uninterned-keyword "foo"))
#t
> (uninterned-keyword? hello:)
#f
> (uninterned-keyword? 123)
#f
|
This procedure returns the void object. The read-eval-print loop
prints nothing when the result is the void object.
( eval expr [env]) | procedure |
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
|
The file parameter must be a string naming an existing file
containing Scheme source code. The include
special form
splices the content of the specified source file. This form can only
appear where a define
form is acceptable.
For example:
| (include "macros.scm")
(define (f lst)
(include "sort.scm")
(map sqrt (sort lst)))
|
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 using the
include
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)))
|
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 v4.6.6
> (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
|
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=350).
-
([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] 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] 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.
-
(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 350)
(constant-fold)
(lambda-lift)
(not standard-bindings)
(not extended-bindings)
(run-time-bindings)
(safe)
(interrupts-enabled)
(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)
(optimize-dead-local-variables)
(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)
.
6.4 Undocumented extensions
The procedures in this section are not yet documented.
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 (lambda () (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
|
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 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 parameter display-env?, which defaults to #f
, controls
if the frames are displayed with its environment (the variables
accessible and their bindings).
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) #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) #f #t)
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
|
( top [thread-group [port]]) | procedure |
( six.for stat1 expr2 expr3 stat2) | special form |
( six.if expr stat1 [stat2]) | special form |
7. Namespaces
TO DO!
8. 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. ASCII, ISO-8859-1, UTF-8, etc). This can be done by specifying
the runtime option ‘-:f’ when gsi
and gsc
are
started.
8.1 Extensions to character procedures
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)
|
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)
.
8.2 Extensions to string procedures
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)
.
9. Numbers
9.1 Extensions to numeric procedures
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)
.
9.2 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
.
9.3 Integer square root and nth root
This procedure returns the integer part of the square root of the
nonnegative exact integer n.
For example:
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
|
9.4 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.
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
|
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
|
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
|
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
|
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
|
This procedure returns the bitwise complement of the exact integer
n.
For example:
| > (bitwise-not 3) ; ...00000011
-4
> (bitwise-not -1) ; ...11111111
0
|
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
|
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
|
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)
|
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
|
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
|
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-bit-set 24) ; ...00011000
3
> (first-bit-set 0) ; ...00000000
-1
|
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
|
9.5 Fixnum specific operations
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))
|
9.6 Flonum specific operations
9.7 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.
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.
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
|
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
|
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)
|
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)
|
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
|
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
|
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
|
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
|
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
|
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
|
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)
|
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)
|
10. Homogeneous 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
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.
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)
|
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))
|
11. Hashing and weak references
11.1 Hashing
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>
|
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))
|
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
|
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
|
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
|
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
|
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
|
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
|
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
|
11.2 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.
11.2.1 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.
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
|
11.2.2 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))
.
( 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]) | procedure |
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
|
The procedure table?
returns #t
when obj is a
table and #f
otherwise.
For example:
| > (table? (make-table))
#t
> (table? 123)
#f
|
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
|
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-table-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 table key
(table-ref '#<table #2> '(4 5))
|
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 table key
(table-ref '#<table #2> '(4 5))
|
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
|
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)
|
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))
|
( 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]) | procedure |
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))
|
Unbound-table-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-table-key-exception object.
The procedure unbound-table-key-exception?
returns
#t
when obj is a unbound-table-key-exception
object and #f
otherwise.
The procedure unbound-table-key-exception-procedure
returns the procedure that raised exc.
The procedure unbound-table-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-table-key-exception? exc)
(list (unbound-table-key-exception-procedure exc)
(unbound-table-key-exception-arguments exc))
'not-unbound-table-key-exception))
> (with-exception-catcher
handler
(lambda () (table-ref t '(1 2))))
(#<procedure #2 table-ref> (#<table #3> (1 2)))
|
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))
|
( table-merge! table1 table2 [table2-takes-precedence?]) | procedure |
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))
|
( table-merge table1 table2 [table2-takes-precedence?]) | procedure |
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))
|
12. Records
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>
|
13. 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).
13.1 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)
13.2 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.
13.3 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.
13.4 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.
13.5 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).
13.6 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.
13.7 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.
13.8 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).
13.9 Procedures
This procedure returns the current thread. For example:
| > (current-thread)
#<thread #1 primordial>
> (eq? (current-thread) (current-thread))
#t
|
This procedure returns #t
when obj is a thread object and
#f
otherwise.
For example:
| > (thread? (current-thread))
#t
> (thread? 'foo)
#f
|
( make-root-thread thunk [name [thread-group [input-port [output-port]]]]) | procedure |
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>
|
This procedure returns the name of the thread. For example:
| > (thread-name (make-thread (lambda () #f) 'foo))
foo
|
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"
|
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
|
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
|
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.
|
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.
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))))
|
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))))
|
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
|
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!
|
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
|
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
|
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))
|
This procedure returns #t
when obj is a mutex object and
#f
otherwise.
For example:
| > (mutex? (make-mutex))
#t
> (mutex? 'foo)
#f
|
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>
|
Returns the name of the mutex. For example:
| > (mutex-name (make-mutex 'foo))
foo
|
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
|
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))))
|
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))))
|
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))))
|
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
|
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>
|
This procedure returns the name of the condition-variable. For
example:
| > (condition-variable-name (make-condition-variable 'foo))
foo
|
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"
|
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!)))
|
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))))
|
14. 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”.
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 filter argument is a one argument conversion
procedure. If it is not specified, filter defaults 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 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 content of the cell is returned. When one
argument is passed the content of the cell is updated with the result
of applying the parameter object’s conversion procedure to the
argument. Note that the conversion procedure 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 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
|
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 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)
|
15. Exceptions
15.1 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
‘-:d’ runtime option (see section 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
|
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
|
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
|
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)
|
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
|
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")
|
15.2 Exception objects related to memory management
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>
|
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>
|
15.3 Exception objects related to the host environment
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)
|
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>))
|
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"))
|
15.4 Exception objects related to threads
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.
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>
|
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>
|
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))
|
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>))
|
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>))
|
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>)
|
15.5 Exception objects related to C-interface
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
"___result = ___arg1;"))
$ gsc test1.scm
$ gsi
Gambit v4.6.6
> (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)
|
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 "___result = f (\"123\");"))
(define t2 (c-lambda () int "___result = f (0);"))
(define t3 (c-lambda () int "___result = f (\"1.5\");"))
$ gsc test2.scm
$ gsi
Gambit v4.6.6
> (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)
|
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
"___result = f (___arg1);"))
$ gsc test3.scm
$ gsi
Gambit v4.6.6
> (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>
|
15.6 Exception objects related to the reader
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"))
|
15.7 Exception objects related to evaluation and compilation
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))
|
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)
|
15.8 Exception objects related to type checking
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-num
returns the position of the
argument whose type is incorrect. Position 1 is the first argument.
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-num 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>)
|
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-num
returns the position of
the argument which is not in the allowed range. Position 1 is the
first argument.
For example:
| > (define (handler exc)
(if (range-exception? exc)
(list (range-exception-procedure exc)
(range-exception-arguments exc)
(range-exception-arg-num exc))
'not-range-exception))
> (with-exception-catcher
handler
(lambda () (string-ref "abcde" 10)))
(#<procedure #2 string-ref> ("abcde" 10) 2)
|
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))
|
Improper-length-list-exception objects are raised by the map
and for-each
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 an improper-length-list-exception object.
The procedure improper-length-list-exception?
returns
#t
when obj is an improper-length-list-exception
object and #f
otherwise.
The procedure improper-length-list-exception-procedure
returns the procedure that raised exc.
The procedure improper-length-list-exception-arguments
returns the list of arguments of the procedure that raised exc.
The procedure improper-length-list-exception-arg-num
returns
the position of the argument whose length is the shortest. Position 1
is the first argument.
For example:
| > (define (handler exc)
(if (improper-length-list-exception? exc)
(list (improper-length-list-exception-procedure exc)
(improper-length-list-exception-arguments exc)
(improper-length-list-exception-arg-num exc))
'not-improper-length-list-exception))
> (with-exception-catcher
handler
(lambda () (map + '(1 2) '(3) '(4 5))))
(#<procedure #2 map> (#<procedure #3 +> (1 2) (3) (4 5)) 3)
|
15.9 Exception objects related to procedure call
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))
|
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)
|
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))
|
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))
|
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))
|
15.10 Other exception objects
( error message obj…) | procedure |
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))
|
16. 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).
16.1 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 Mac OS X and by a ‘/’ or ‘\’ under
MSDOS and Microsoft Windows. A leading separator indicates an
absolute path under UNIX, Mac OS X, 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 ‘-:=DIRECTORY’ and
‘-:~~DIR=DIRECTORY’ runtime options or by defining
the ‘GAMBCOPT’ environment variable.
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, Mac OS X, 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 Mac OS X this is found using the password file. There is no
equivalent under MSDOS and Microsoft Windows.
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 current working directory of the current process. 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/"
> (path-expand "foo" "~~")
"/usr/local/Gambit-C/foo"
> (parameterize ((current-directory "~~")) (path-expand "foo"))
"/usr/local/Gambit-C/foo"
|
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-C/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/"
|
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")
|
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
|
16.2 Filesystem operations
This procedure creates a directory. 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 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-directory "newdir")
*** ERROR IN (console)@2.1 -- File exists
(create-directory "newdir")
|
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
|
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.
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.
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.
( copy-file source-path destination-path) | procedure |
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.
This procedure deletes the file path. The argument path
must be a string denoting the path of an existing file.
This procedure deletes the directory path. The argument
path must be a string denoting the path of an existing
directory.
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 Mac OS X 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")
|
16.3 Shell command execution
The procedure shell-command
calls up the shell to execute
command which must be a string. This procedure returns
the exit status of the shell in the form that the C library’s
system
routine returns.
For example under UNIX:
| > (shell-command "ls -sk f*.scm")
4 fact.scm 4 fib.scm
0
|
16.4 Process termination
The procedure exit
causes the process to terminate with the
status status which must be an exact integer in the range 0 to
255. If it is not specified, status defaults to 0.
For example under UNIX:
| $ gsi
Gambit v4.6.6
> (exit 42)
$ echo $?
42
|
16.5 Command line arguments
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 -:d -e "(pretty-print (command-line))"
("gsi" "-e" "(pretty-print (command-line))")
$ cat foo
#!/usr/local/Gambit-C/bin/gsi-script
(pretty-print (command-line))
$ ./foo 1 2 "3 4"
("/u/feeley/./foo" "1" "2" "3 4")
|
16.6 Environment variables
( getenv name [default]) | procedure |
( setenv name [new-value]) | procedure |
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")
|
16.7 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).
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
|
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
|
The time
special form evaluates expr and returns the
result. As a side effect it displays a message on the interaction
channel which indicates 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).
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
|
16.8 File information
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
|
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>>
|
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
|
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
|
Returns the device field of the file-information record
file-info.
For example:
| > (file-info-device (file-info "/dev/tty"))
19513156
|
Returns the inode field of the file-information record
file-info.
For example:
| > (file-info-inode (file-info "/dev/tty"))
20728196
|
Returns the mode field of the file-information record
file-info.
For example:
| > (file-info-mode (file-info "/dev/tty"))
438
|
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
|
Returns the owner field of the file-information record
file-info.
For example:
| > (file-info-owner (file-info "/dev/tty"))
0
|
Returns the group field of the file-information record
file-info.
For example:
| > (file-info-group (file-info "/dev/tty"))
0
|
Returns the size field of the file-information record
file-info.
For example:
| > (file-info-size (file-info "/dev/tty"))
0
|
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>
|
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>
|
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>
|
Returns the attributes field of the file-information record
file-info.
For example:
| > (file-info-attributes (file-info "/dev/tty"))
128
|
Returns the creation-time field of the file-information record
file-info.
For example:
| > (file-info-creation-time (file-info "/dev/tty"))
#<time #2>
|
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))
.
16.9 Group information
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)
|
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
|
Returns the symbolic name field of the group-information record
group-info.
For example:
| > (group-info-name (group-info 29))
"certusers"
|
Returns the group id field of the group-information record
group-info.
For example:
| > (group-info-gid (group-info "staff"))
20
|
Returns the members field of the group-information record
group-info.
For example:
| > (group-info-members (group-info "staff"))
("root")
|
16.10 User information
This procedure returns the user’s name as a string.
For example:
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)
|
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
|
Returns the symbolic name field of the user-information record
user-info.
For example:
| > (user-info-name (user-info 0))
"root"
|
Returns the user id field of the user-information record
user-info.
For example:
| > (user-info-uid (user-info "feeley"))
506
|
Returns the group id field of the user-information record
user-info.
For example:
| > (user-info-gid (user-info "feeley"))
506
|
Returns the home directory field of the user-information record
user-info.
For example:
| > (user-info-home (user-info 0))
"/var/root"
|
Returns the shell field of the user-information record
user-info.
For example:
| > (user-info-shell (user-info 0))
"/bin/sh"
|
16.11 Host information
This procedure returns the machine’s host name as a string.
For example:
| > (host-name)
"mega.iro.umontreal.ca"
|
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")
|
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
|
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"
|
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")
|
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))
|
( address-infos [host: host] [service: service] [family: family] [socket-type: socket-type] [protocol: protocol]) | procedure |
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")
|
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
|
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)
|
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)
|
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)
|
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)>)
|
16.12 Service information
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">
|
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
|
Returns the symbolic name field of the service-information record
service-info.
For example:
| > (service-info-name (service-info 80))
"http"
|
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")
|
Returns the service port number field of the service-information
record service-info.
For example:
| > (service-info-port-number (service-info "http"))
80
|
Returns the service protocol name field of the service-information
record service-info.
For example:
| > (service-info-protocol (service-info "http"))
"udp"
|
16.13 Protocol information
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>
|
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
|
Returns the symbolic name field of the protocol-information record
protocol-info.
For example:
| > (protocol-info-name (protocol-info 6))
"tcp"
|
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")
|
Returns the protocol number field of the protocol-information record
protocol-info.
For example:
| > (protocol-info-number (protocol-info "tcp"))
6
|
16.14 Network information
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>
|
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
|
Returns the symbolic name field of the network-information record
network-info.
For example:
| > (network-info-name (network-info 127))
"loopback"
|
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")
|
Returns the network number field of the network-information record
network-info.
For example:
| > (network-info-number (network-info "loopback"))
127
|
17. I/O and ports
17.1 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.
17.2 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).
-
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.
-
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.
-
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.
-
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.
17.3 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
This will use default settings for the character encoding, buffering,
etc. If the UTF-8 character encoding is desired, then the port could
be opened using the call
| (open-input-file (list path: "foo" char-encoding: 'UTF-8))
|
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).
17.4 Object-ports
17.4.1 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.
17.4.2 Object-port operations
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
|
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
|
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!")
|
( write obj [port]) | procedure |
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)>
|
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
|
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 Mac OS X 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\""
|
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\""
|
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
|
17.5 Character-ports
17.5.1 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.
17.5.2 Character-port operations
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)"
|
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
|
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
|
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
|
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:
( read-line [port [separator [include-separator? [max-length]]]]) | procedure |
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
|
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"
|
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.
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.
17.6 Byte-ports
17.6.1 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 is operating system dependent, but this can be overridden
through the runtime options (see section Runtime options). 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 (see section 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.
17.6.2 Byte-port operations
This procedure reads the byte input-port port and returns the
byte at the current byte location and advances the current byte
location to the next byte, unless the port is already at
end-of-file in which case read-u8
returns the end-of-file
object. If it is not specified, port defaults to the current
input-port.
This procedure must be called when the port’s input character buffer
is empty otherwise the character-stream and byte-stream may be out of
sync due to buffering. The input character buffer is used for bulk
decoding of encoded characters (i.e. to translate the byte-stream into
a character-stream). The input character buffer is initially empty.
It is only when characters are read that it is filled with characters
obtained by decoding the byte-stream.
One way to ensure that the port’s input character buffer is empty is
to call read-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
|
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)
|
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)
|
17.7 Device-ports
17.7.1 Filesystem devices
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, 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>)
|
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
|
17.7.2 Process devices
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"
|
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
|
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))
|
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")>))
|
17.7.3 Network devices
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 server-address:
"HOST:PORT"
.
Here are the settings allowed in addition to the generic settings of
byte-ports:
-
server-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 server-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.
-
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.
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>"
|
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 port-number:
setting. 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 "INTF:PORT"
it is handled
as if it was the setting server-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.
-
server-address:
string-or-ip-address
This setting indicates the internet address of the 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 server-address:
"INTF"
port-number:
PORT.
-
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
.
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 server-address: "*"
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)
|
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. 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)
|
17.8 Directory-ports
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 Mac OS X 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")
|
17.9 Vector-ports
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 first 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
, and open-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)
|
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
|
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)
|
17.10 String-ports
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"
|
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>)"
|
17.11 U8vector-ports