INFO-DIR-SECTION Programming START-INFO-DIR-ENTRY * Gambit: (gambit). A portable implementation of Scheme. * gsi: (gambit) interpreter. Gambit interpreter. * gsc: (gambit) compiler. Gambit compiler. END-INFO-DIR-ENTRY This file documents Gambit, a portable implementation of Scheme. Copyright (C) 1994-2016 Marc Feeley. Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies. Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one. Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that this permission notice may be stated in a translation approved by the copyright holder. Gambit ****** This manual documents Gambit. It covers release v4.9.3. 1 The Gambit 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. The Gambit compiler generates portable C code, making the whole Gambit 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://gambitscheme.org'. 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' under UNIX, `/Library/Gambit' under Mac OS X and `C:/Program Files/Gambit' under Microsoft Windows. This may have been overridden when the system was built with the command `configure --prefix=/my/Gambit'. If the system was built with the command `configure --enable-multiple-versions' then the central installation directory is `prefix/version', where `version' is the system version string (e.g. `v4.9.3' for Gambit v4.9.3). 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, 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/lib gsi A similar problem exists with the Gambit header file `gambit.h', located in the `include' installation directory. This header file is needed for compiling Scheme programs with the Gambit compiler. When the C compiler is being called explicitly it may be necessary to use a `-I' command line option to indicate where to find header files and a `-L' command line option to indicate where to find libraries. Access to both of these files can be simplified by creating a link to them in the appropriate system directories (special privileges may however be required): $ ln -s /usr/local/Gambit/lib/libgambit.a /usr/lib # name may vary $ ln -s /usr/local/Gambit/include/gambit.h /usr/include Alternatively these files can be copied or linked in the directory where the C compiler is invoked (this requires no special privileges). Another approach is to set some environment variables which are used to tell the C compiler where to find header files and libraries. For example, the following settings can be used for the `gcc' C compiler: $ export LIBRARY_PATH=/usr/local/Gambit/lib $ export CPATH=/usr/local/Gambit/include Note that this may have been done by the installation process. In particular, the 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 (*note GSI 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 *Note Runtime options::. 2.1 Interactive mode ==================== In interactive mode a read-eval-print loop (REPL) is started for the user to interact with the interpreter. At each iteration of this loop the interpreter displays a prompt, reads a command and executes it. The commands can be expressions to evaluate (the typical case) or special commands related to debugging, for example `,q' to terminate the process (for a complete list of commands see *Note Debugging::). 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 *Note 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 v4.9.3 > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1))))) > (map fact '(1 2 3 4 5 6)) (1 2 6 24 120 720) > (values (fact 10) (fact 40)) 3628800 815915283247897734345611269596115894272000000000 > ,q What happens when errors occur is explained in *Note Debugging::. 2.2 Batch mode ============== In batch mode the command line arguments denote 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 *Note 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/gambext")' (for an explanation of how file names are interpreted see *Note Host environment::). An error is not signaled when the file does not exist. Interpreter extensions and patches that are meant to apply to all users and all modes should go in that file. Extensions which are meant to apply to a single user or to a specific working directory are best placed in the "initialization file", which is a file containing Scheme code. In all modes, the interpreter first tries to locate the initialization file by searching the following locations: `.gambini' and `~/.gambini' (with no extension, a `.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 `GAMBOPT' contained a syntax error or were invalid. `70' This normally indicates that an exception was raised in the primordial thread and the exception was not handled. `71' There was a problem initializing the runtime system, for example insufficient memory to allocate critical tables. For example, if the shell is `sh': $ gsi -: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 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 installation process creates the executable `gsi' and also the executables `six', `gsi-script', `six-script', `scheme-r5rs', `scheme-srfi-0', etc as links to `gsi'. A Scheme script need only start with the name of the desired Scheme language variant prefixed with `#!' and the directory where the Gambit executables are stored. This script should be made executable by setting the execute permission bits (with a `chmod +x SCRIPT'). Here is an example of a script which lists on standard output the files in the current directory: #!/usr/local/Gambit/bin/gsi-script (for-each pretty-print (directory-files)) Here is another UNIX script, using the Scheme infix syntax extension, which takes a single integer argument and prints on standard output the numbers from 1 to that integer: #!/usr/local/Gambit/bin/six-script 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 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 (*note GSC::) into a standalone executable. The script line will provide information to the compiler on which language to use. The script line also provides information on which runtime options to use when executing the compiled script. This is useful to set the default runtime options of an executable program. The compiled script will be executed similarly to an interpreted script (i.e. the list of command line arguments returned by the `command-line' procedure and the invocation of the `main' procedure). For example: $ cat square.scm #!/usr/local/Gambit/bin/gsi-script -: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] [-cfg] [-dg] [-debug] [-debug-location] [-debug-source] [-debug-environments] [-track-scheme] [-o OUTPUT] [-c] [-keep-c] [-link] [-flat] [-l BASE] [-module-name MODULE-NAME] [-linker-name LINKER-NAME] [[-] [-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 *Note Scheme infix syntax extension::). Otherwise, `gsc' will parse the Scheme file using the normal Scheme prefix syntax. Files with a C file extension must have been previously produced by `gsc', with the `-c' 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 (*note Runtime options::). If present, the `-i', `-f', and `-v' compiler options must come first. The available options are: `-i' Force interpreter mode. `-f' Do not examine the initialization file. `-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. `-cfg' Generate a control flow graph of the GVM code. `-dg' Generate a dependency graph. `-debug' Include all debugging information in the code generated. `-debug-location' Include source code location debugging information in the code generated. `-debug-source' Include the source code debugging information in the code generated. `-debug-environments' Include environment debugging information in the code generated. `-track-scheme' Generate `#line' directives referring back to the Scheme code. `-o OUTPUT' Set name of output file or directory where output file(s) are written. `-dynamic' Compile Scheme source files to dynamically loadable object files (this is the default). `-exe' Compile Scheme source files 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. `-module-name MODULE-NAME' Specify the name of the generated module. `-linker-name LINKER-NAME' Specify the name of the low-level initialization function exported by the module. `-' 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 `-cfg' option generates a visual representation of the control flow graph of the intermediate code for the "Gambit Virtual Machine" (GVM) of each Scheme file on `FILE.cfg'. The file is suitable for processing with the "dot" program. For example, to generate the PDF file `FILE.cfg.pdf' from `FILE.cfg' the following command can be used: $ dot -O -Tpdf FILE.cfg The `-dg' option generates a visual representation of the dependency graph of each Scheme file on `FILE.dg'. The file is suitable for processing with the "dot" program. For example, to generate the PDF file `FILE.dg.pdf' from `FILE.dg' the following command can be used: $ dot -O -Tpdf FILE.dg The `-debug' option causes all kinds of debugging information to be saved in the code generated. See the documentation of the `debug' declaration for details. The `-debug-location' option causes source code location debugging information to be saved in the code generated. See the documentation of the `debug-location' declaration for details. The `-debug-source' option causes source code debugging information to be saved in the code generated. See the documentation of the `debug-source' declaration for details. The `-debug-environments' option causes environment debugging information to be saved in the code generated. See the documentation of the `debug-environments' declaration for details. The `-track-scheme' 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/_gambit.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 `GAMBCOMP_VERBOSE' environment variable to a nonnull value. For example: $ export GAMBCOMP_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 -I"/usr/local/Gambit/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 -I"/usr/local/Gambit/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 -I"/usr/local/Gambit/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 -I"/usr/local/Gambit/include" -o "hello.exe" h.o w.o w_.o "/usr/local/Gambit/lib/libgambit.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< /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 #< (compile-file-to-target "h") "/Users/feeley/gambit/doc/h.c" -- procedure: compile-file FILE [`options:' OPTIONS] [`output:' OUTPUT] [`module-name:' MODULE-NAME] [`linker-name:' LINKER-NAME] [`expression:' EXPRESSION] [`cc-options:' CC-OPTIONS] [`ld-options-prelude:' LD-OPTIONS-PRELUDE] [`ld-options:' LD-OPTIONS] The FILE, OPTIONS, OUTPUT, MODULE-NAME, LINKER-NAME, and EXPRESSION parameters have the same meaning as for the `compile-file-to-target' procedure, except that FILE may be a Scheme source file or a file possibly generated by the Gambit Scheme compiler (for example with the `compile-file-to-target' procedure). The CC-OPTIONS parameter is a string containing the options to pass to the C compiler and the LD-OPTIONS-PRELUDE and LD-OPTIONS parameters are strings containing the options to pass to the C linker (the options in LD-OPTIONS-PRELUDE are passed to the C linker before the input file and the options in LD-OPTIONS are passed after). The `compile-file' procedure compiles the source file FILE into an object file, which is either a file dynamically loadable using the `load' procedure, or a C linkable object file destined to be linked with the C linker (for example to create a standalone executable program). The presence of the `obj' option in OPTIONS will cause the creation of a C linkable object file and therefore the options LD-OPTIONS-PRELUDE and LD-OPTIONS are ignored, otherwise a dynamically loadable file is created. In both cases, if FILE is a Scheme source file, the compiler first compiles FILE to a C file which is created in the same directory as FILE regardless of the OUTPUT parameter. Then the C file is compiled with the C compiler. When the compilation is successful, `compile-file' returns the name of the object file generated. When there is a compilation error, `#f' is returned. The name of the object file can be specified with the OUTPUT parameter. If OUTPUT is a string naming a directory then the object file is created in that directory. Otherwise the name of the object file is OUTPUT. In the case of a dynamically loadable object file, by default the object file is named after FILE with the extension replaced with `.oN', where N is a positive integer that acts as a version number. The next available version number is generated automatically by `compile-file'. When dynamically loaded object files are loaded using the `load' procedure, the `.oN' extension can be specified (to select a particular version) or omitted (to load the file with a `.oN' extension with the highest N consecutively from 1). When the `.oN' extension is not specified and older versions are no longer needed, all versions must be deleted and the compilation must be repeated (this is necessary because the file name, including the extension, is used to name some of the exported symbols of the object file). Note that dynamically loadable object files can only be generated on host operating systems that support dynamic loading. $ cat h.scm (display "hello") (newline) $ gsc Gambit v4.9.3 > (compile-file "h") "/Users/feeley/gambit/doc/h.o1" > (load "h") hello "/Users/feeley/gambit/doc/h.o1" > (compile-file-to-target "h" output: "h.o99.c") "/Users/feeley/gambit/doc/h.o99.c" > (compile-file "h.o99.c") "/Users/feeley/gambit/doc/h.o99" > (load "h.o99") hello "/Users/feeley/gambit/doc/h.o99" > (compile-file-to-target "h") "/Users/feeley/gambit/doc/h.c" > (compile-file "h.c" options: '(obj)) "/Users/feeley/gambit/doc/h.o" -- procedure: link-incremental MODULE-LIST [`output:' OUTPUT] [`linker-name:' LINKER-NAME] [`base:' BASE] [`warnings?:' WARNINGS?] The first parameter must be a non empty list of strings naming Scheme modules to link (the file extension may be omitted). An incremental link file is generated for the modules specified in MODULE-LIST. By default the link file generated is named `LAST_.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/_gambit.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 v4.9.3 > (compile-file-to-target "h") "/Users/feeley/gambit/doc/h.c" > (compile-file-to-target "w") "/Users/feeley/gambit/doc/w.c" > (link-incremental '("h" "w") output: "hello.c") "/Users/feeley/gambit/doc/hello_.c" > ,q $ gsc -obj h.c w.c hello.c h.c: w.c: hello.c: $ gcc h.o w.o hello.o -lgambit -o hello $ ./hello hello world -- procedure: link-flat MODULE-LIST [`output:' OUTPUT] [`linker-name:' LINKER-NAME] [`warnings?:' WARNINGS?] The first parameter must be a non empty list of strings naming Scheme modules to link (the file extension may be omitted). The first string must be the name of a Scheme module or the name of a link file and the remaining strings must name Scheme modules. A flat link file is generated for the modules specified in MODULE-LIST. By default the link file generated is named `LAST_.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 v4.9.3 > (compile-file-to-target "m6") "/Users/feeley/gambit/doc/m6.c" > (compile-file-to-target "m7") "/Users/feeley/gambit/doc/m7.c" > (link-flat '("m6" "m7") output: "lib.o1.c") *** WARNING -- "*" is not defined, *** referenced in: ("m6.c") *** WARNING -- "+" is not defined, *** referenced in: ("m7.c") *** WARNING -- "n" is not defined, *** referenced in: ("m7.c") "/Users/feeley/gambit/doc/lib.o1.c" > ,q $ gcc -bundle -D___DYNAMIC m6.c m7.c lib.o1.c -o lib.o1 $ gsc Gambit v4.9.3 > (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 various sources of runtime options: the defaults, the environment variable `GAMBOPT', the script line of the source code, and, unless the program is an interpreted script, the first command line argument of the program. Any runtime option can be overriden by a subsequent source of runtime options. It is sometimes useful to prevent overriding the runtime options of the script line. This can be achieved by starting the script line runtime options with `-::'. In this case the environment variable `GAMBOPT' is ignored, and the first command line argument of the program is not used for runtime options (it is treated like a normal command line argument even if it starts with `-:'). For example: $ GAMBOPT=d0,=~/my-gambit2 $ export GAMBOPT $ 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: a. An evaluation error has occured, such as attempting to divide by zero. b. The user has interrupted the evaluation (usually by typing <^C>). c. A breakpoint has been reached or `(step)' was evaluated. d. Single-stepping mode is enabled. When an evaluation stops, a message is displayed indicating the reason and location where the evaluation was stopped. The location information includes, if known, the name of the procedure where the evaluation was stopped and the source code location in the format `STREAM@LINE.COLUMN', where STREAM is either a string naming a file or a symbol within parentheses, such as `(console)'. A "nested REPL" is then initiated in the context of the point of execution where the evaluation was stopped. The nested REPL's continuation and evaluation environment are the same as the point where the evaluation was stopped. For example when evaluating the expression `(let ((y (- 1 1))) (* (/ x y) 2))', a "divide by zero" error is reported and the nested REPL's continuation is the one that takes the result and multiplies it by two. The REPL's lexical environment includes the lexical variable `y'. This allows the inspection of the evaluation context (i.e. the lexical and dynamic environments and continuation), which is particularly useful to determine the exact location and cause of an error. The prompt of nested REPLs includes the nesting level; `1>' is the prompt at the first nesting level, `2>' at the second nesting level, and so on. An end of file (usually <^D>) will cause the current REPL to be terminated and the enclosing REPL (one nesting level less) to be resumed. At any time the user can examine the frames in the REPL's continuation, which is useful to determine which chain of procedure calls lead to an error. A backtrace that lists the chain of active continuation frames in the REPL's continuation can be obtained with the `,b' command. The frames are numbered from 0, that is frame 0 is the most recent frame of the continuation where execution stopped, frame 1 is the parent frame of frame 0, and so on. It is also possible to move the REPL to a specific parent continuation (i.e. a specific frame of the continuation where execution stopped) with the `,N', `,N+', `,N-', `,+', `,-', `,++', and `,--' commands. When the frame number of the frame being examined is not zero, it is shown in the prompt after the nesting level, for example `1\5>' is the prompt when the REPL nesting level is 1 and the frame number is 5. Expressions entered at a nested REPL are evaluated in the environment (both lexical and dynamic) of the continuation frame currently being examined if that frame was created by interpreted Scheme code. If the frame was created by compiled Scheme code then expressions get evaluated in the global interaction environment. This feature may be used in interpreted code to fetch the value of a variable in the current frame or to change its value with `set!'. Note that some special forms (`define' in particular) can only be evaluated in the global interaction environment. 5.2 Debugging commands ====================== In addition to expressions, the REPL accepts the following special "comma" commands: `,?' Give a summary of the REPL commands. `,(h SUBJECT)' This command will show the section of the Gambit manual with the definition of the procedure or special form SUBJECT, which must be a symbol. For example `,(h time)' will show the section documenting the `time' special form. Please see the `help' procedure for additional information. `,h' This command will show the section of the Gambit manual with the definition of the procedure which raised the exception for which this REPL was started. `,q' Terminate the process with exit status 0. This is equivalent to calling `(exit 0)'. `,qt' Terminate the current thread (note that terminating the primordial thread terminates the process). `,t' Return to the outermost REPL, also known as the "top-level REPL". `,d' Leave the current REPL and resume the enclosing REPL. This command does nothing in the top-level REPL. `,(c EXPR)' Leave the current REPL and continue the computation that initiated the REPL with a specific value. This command can only be used to continue a computation that signaled an error. The expression EXPR is evaluated in the current context and the resulting value is returned as the value of the expression which signaled the error. For example, if the evaluation of the expression `(* (/ x y) 2)' signaled an error because `y' is zero, then in the nested REPL a `,(c (+ 4 y))' will resume the computation of `(* (/ x y) 2)' as though the value of `(/ x y)' was 4. This command must be used carefully because the context where the error occured may rely on the result being of a particular type. For instance a `,(c #f)' in the previous example will cause `*' to signal a type error (this problem is the most troublesome when debugging Scheme code that was compiled with type checking turned off so be careful). `,c' Leave the current REPL and continue the computation that initiated the REPL. This command can only be used to continue a computation that was stopped due to a user interrupt, breakpoint or a single-step. `,s' Leave the current REPL and continue the computation that initiated the REPL in single-stepping mode. The computation will perform an evaluation step (as defined by `step-level-set!') and then stop, causing a nested REPL to be entered. Just before the evaluation step is performed, a line is displayed (in the same format as `trace') which indicates the expression that is being evaluated. If the evaluation step produces a result, the result is also displayed on another line. A nested REPL is then entered after displaying a message which describes the next step of the computation. This command can only be used to continue a computation that was stopped due to a user interrupt, breakpoint or a single-step. `,l' This command is similar to `,s' except that it "leaps" over procedure calls, that is procedure calls are treated like a single step. Single-stepping mode will resume when the procedure call returns, or if and when the execution of the called procedure encounters a breakpoint. `,N' Move to frame number N of the continuation. After changing the current frame, a one-line summary of the frame is displayed as if the `,y' command was entered. `,N+' Move forward by N frames in the chain of continuation frames (i.e. towards older continuation frames). After changing the current frame, a one-line summary of the frame is displayed as if the `,y' command was entered. `,N-' Move backward by N frames in the chain of continuation frames (i.e. towards more recent continuation frames). After changing the current frame, a one-line summary of the frame is displayed as if the `,y' command was entered. `,+' Equivalent to `,1+'. `,-' Equivalent to `,1-'. `,++' Equivalent to `,N+' where N is the number of continuation frames displayed at the head of a backtrace. `,--' Equivalent to `,N-' where N is the number of continuation frames displayed at the head of a backtrace. `,y' Display a one-line summary of the current frame. The information is displayed in four fields. The first field is the frame number. The second field is the procedure that created the frame or `(interaction)' if the frame was created by an expression entered at the REPL. The remaining fields describe the subproblem associated with the frame, that is the expression whose value is being computed. The third field is the location of the subproblem's source code and the fourth field is a reproduction of the source code, possibly truncated to fit on the line. The last two fields may be missing if that information is not available. In particular, the third field is missing when the frame was created by a user call to the `eval' procedure or by a compiled procedure not compiled with the declaration `debug-location', and the last field is missing when the frame was created by a compiled procedure not compiled with the declaration `debug-source'. `,b' Display a backtrace summarizing each frame in the chain of continuation frames starting with the current frame. For each frame, the same information as for the `,y' command is displayed (except that location information is displayed in the format `STREAM@LINE:COLUMN'). If there are more than 15 frames in the chain of continuation frames, some of the middle frames will be omitted. `,be' Like the `,b' command but also display the environment. `,bed' Like the `,be' command but also display the dynamic environment. `,(b EXPR)' Display the backtrace of EXPR's value, X, which is obtained by evaluating EXPR in the current frame. X must be a continuation or a thread. When X is a continuation, the frames in that continuation are displayed. When X is a thread, the backtrace of the current continuation of that thread is displayed. `,(be EXPR)' Like the `,(b EXPR)' command but also display the environment. `,(bed EXPR)' Like the `,(be EXPR)' command but also display the dynamic environment. `,i' Pretty print the procedure that created the current frame or `(interaction)' if the frame was created by an expression entered at the REPL. Compiled procedures will only be pretty printed when they are compiled with the declaration `debug-source'. `,e' Display the environment which is accessible from the current frame. The lexical environment is displayed, followed by the dynamic environment if the parameter object `repl-display-dynamic-environment?' is not false. Global lexical variables are not displayed. Moreover the frame must have been created by interpreted code or code compiled with the declaration `debug-environments'. Due to space safety considerations and compiler optimizations, some of the lexical variable bindings may be missing. Lexical variable bindings are displayed using the format `VARIABLE = EXPRESSION' (when VARIABLE is mutable) or `VARIABLE == EXPRESSION' (when VARIABLE is immutable, which may happen in compiled code due to compiler optimization) and dynamically-bound parameter bindings are displayed using the format `(PARAMETER) = EXPRESSION'. Note that EXPRESSION can be a self-evaluating expression (number, string, boolean, character, ...), a quoted expression, a lambda expression or a global variable (the last two cases, which are only used when the value of the variable or parameter is a procedure, simplifies the debugging of higher-order procedures). A PARAMETER can be a quoted expression or a global variable. Lexical bindings are displayed in inverse binding order (most deeply nested first) and shadowed variables are included in the list. `,ed' Like the `,e' command but the dynamic environment is always displayed. `,(e EXPR)' Display the environment of EXPR's value, X, which is obtained by evaluating EXPR in the current frame. X must be a continuation, a thread, a procedure, or a nonnegative integer. When X is a continuation, the environment at that point in the code is displayed. When X is a thread, the environment of the current continuation of that thread is displayed. When X is a procedure, the lexical environment where X was created is combined with the current continuation and this combined environment is displayed. When X is an integer, the environment at frame number X of the continuation is displayed. `,(ed EXPR)' Like the `,(e EXPR)' command but the dynamic environment is always displayed. `,st' Display the state of the threads in the current thread's thread group. A thread can be: uninitialized, initialized, active, and terminated (normally or abnormally). Active threads can be running, sleeping and waiting on a synchronization object (mutex, condition variable or port) possibly with a timeout. `,(st EXPR)' Display the state of a specific thread or thread group. The value of EXPR must be a thread or thread group. `,(v EXPR)' Start a new REPL visiting EXPR's value, X, which is obtained by evaluating EXPR in the current frame. X must be a continuation, a thread, a procedure, or a nonnegative integer. When X is a continuation, the new REPL's continuation is X and evaluations are done in the environment at that point in the code. When X is a thread, the thread is interrupted and the new REPL's continuation is the point where the thread was interrupted. When X is a procedure, the lexical environment where X was created is combined with the current continuation and evaluations are done in this combined environment. When X is an integer, the REPL is started in frame number X of the continuation. 5.3 Debugging example ===================== Here is a sample interaction with `gsi': $ gsi Gambit v4.9.3 > (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 # = (lambda (x) (/ 1 (expt x 2))) 1> ,e x = 'hello 1> ,b 0 invsqr (console)@1:25 (expt x 2) 1 # (console)@6:17 (fn (car in)) 2 # (console)@6:31 (mm (cdr in)) 3 # (console)@6:31 (mm (cdr in)) 4 (interaction) (console)@8:1 (mymap invsqr '(5 2 hel... 1> ,+ 1 # (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 # 1\1> (pp fn) (lambda (x) (/ 1 (expt x 2))) 1\1> ,+ 2 # (console)@6.31 (mm (cdr in)) 1\2> ,e in = '(2 hello 9 1) mm = (lambda (in) (if (null? in) '() (cons (fn (car in)) (mm (cdr in))))) fn = invsqr lst = '(5 2 hello 9 1) 1\2> ,(c (list 3 4 5)) (1/25 1/4 3 4 5) > ,q 5.4 Procedures related to debugging =================================== -- procedure: help SUBJECT -- procedure: help-browser [NEW-VALUE] The `help' procedure displays the section of the Gambit manual with the definition of the procedure or special form SUBJECT, which must be a procedure or symbol. For example the call `(help gensym)' will show the section documenting the `gensym' procedure and the call `(help 'time)' will show the section documenting the `time' special form. The `help' procedure returns the void object. The parameter object `help-browser' is bound to a string naming the external program that is used by the `help' procedure to view the documentation. Initially it is bound to the empty string. In normal circumstances when `help-browser' is bound to an empty string the `help' procedure runs the script `~~bin/gambdoc.bat' which searches for a suitable web browser to open the documentation in HTML format. Unless the system was built with the command `configure --enable-help-browser=...', the text-only browser `lynx' (see `http://lynx.isc.org/') will be used by default if it is available. We highly recommend that you install this browser if you are interested in viewing the documentation within the console in which the REPL is running. You can exit `lynx' conveniently by typing an end of file (usually <^D>). For example: > (help-browser "firefox") ; use firefox instead of lynx > (help 'gensym) > (help gensym) ; OK because gensym is a procedure > (help 'time) > (help time) ; not OK because time is a special form *** ERROR IN (console)@5.7 -- Macro name can't be used as a variable: time > -- procedure: repl-result-history-ref I -- procedure: repl-result-history-max-length-set! N The REPL keeps a history of the last few results printed by the REPL. The call `(repl-result-history-ref I)' returns the Ith previous result (the last for I=0, the next to last for I=1, etc). By default the REPL result history remembers up to 3 results. The maximal length of the history can be set to N between 0 and 10 by a call to `(repl-result-history-max-length-set! N)'. For convenience the reader defines an abbreviation for calling `repl-result-history-ref'. Tokens formed by a sequence of one or more hash signs, such as ``#'', ``##'', etc, are expanded by the reader into the list `(repl-result-history-ref I)', where I is the number of hash signs minus 1. In other words, ``#'' will return the last result printed by the REPL, ``##'' will return the next to last, etc. For example: > (map (lambda (x) (* x x)) '(1 2 3)) (1 4 9) > (reverse #) (9 4 1) > (append # ##) (9 4 1 1 4 9) > 1 1 > 1 1 > (+ # ##) 2 > (+ # ##) 3 > (+ # ##) 5 > #### *** ERROR IN (console)@9.1 -- (Argument 1) Out of range (repl-result-history-ref 3) 1> -- procedure: trace PROC... -- procedure: untrace PROC... The `trace' procedure starts tracing calls to the specified procedures. When a traced procedure is called, a line containing the procedure and its arguments is displayed (using the procedure call expression syntax). The line is indented with a sequence of vertical bars which indicate the nesting depth of the procedure's continuation. After the vertical bars is a greater-than sign which indicates that the evaluation of the call is starting. When a traced procedure returns a result, it is displayed with the same indentation as the call but without the greater-than sign. This makes it easy to match calls and results (the result of a given call is the value at the same indentation as the greater-than sign). If a traced procedure P1 performs a tail call to a traced procedure P2, then P2 will use the same indentation as P1. This makes it easy to spot tail calls. The special handling for tail calls is needed to preserve the space complexity of the program (i.e. tail calls are implemented as required by Scheme even when they involve traced procedures). The `untrace' procedure stops tracing calls to the specified procedures. When no argument is passed to the `trace' procedure, the list of procedures currently being traced is returned. The void object is returned by the `trace' procedure when it is passed one or more arguments. When no argument is passed to the `untrace' procedure stops all tracing and returns the void object. A compiled procedure may be traced but only if it is bound to a global variable. For example: > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1))))) > (trace fact) > (fact 5) | > (fact 5) | | > (fact 4) | | | > (fact 3) | | | | > (fact 2) | | | | | > (fact 1) | | | | | 1 | | | | 2 | | | 6 | | 24 | 120 120 > (trace -) *** WARNING -- Rebinding global variable "-" to an interpreted procedure > (define (fact-iter n r) (if (< n 2) r (fact-iter (- n 1) (* n r)))) > (trace fact-iter) > (fact-iter 5 1) | > (fact-iter 5 1) | | > (- 5 1) | | 4 | > (fact-iter 4 5) | | > (- 4 1) | | 3 | > (fact-iter 3 20) | | > (- 3 1) | | 2 | > (fact-iter 2 60) | | > (- 2 1) | | 1 | > (fact-iter 1 120) | 120 120 > (trace) (# # #) > (untrace) > (fact 5) 120 -- procedure: step -- procedure: step-level-set! LEVEL The `step' procedure enables single-stepping mode. After the call to `step' the computation will stop just before the interpreter executes the next evaluation step (as defined by `step-level-set!'). A nested REPL is then started. Note that because single-stepping is stopped by the REPL whenever the prompt is displayed it is pointless to enter `(step)' by itself. On the other hand entering `(begin (step) EXPR)' will evaluate EXPR in single-stepping mode. The procedure `step-level-set!' sets the stepping level which determines the granularity of the evaluation steps when single-stepping is enabled. The stepping level LEVEL must be an exact integer in the range 0 to 7. At a level of 0, the interpreter ignores single-stepping mode. At higher levels the interpreter stops the computation just before it performs the following operations, depending on the stepping level: 1. procedure call 2. `delay' special form and operations at lower levels 3. `lambda' special form and operations at lower levels 4. `define' special form and operations at lower levels 5. `set!' special form and operations at lower levels 6. variable reference and operations at lower levels 7. constant reference and operations at lower levels The default stepping level is 7. For example: > (define (fact n) (if (< n 2) 1 (* n (fact (- n 1))))) > (step-level-set! 1) > (begin (step) (fact 5)) *** STOPPED IN (console)@3.15 1> ,s | > (fact 5) *** STOPPED IN fact, (console)@1.22 1> ,s | | > (< n 2) | | #f *** STOPPED IN fact, (console)@1.43 1> ,s | | > (- n 1) | | 4 *** STOPPED IN fact, (console)@1.37 1> ,s | | > (fact (- n 1)) *** STOPPED IN fact, (console)@1.22 1> ,s | | | > (< n 2) | | | #f *** STOPPED IN fact, (console)@1.43 1> ,s | | | > (- n 1) | | | 3 *** STOPPED IN fact, (console)@1.37 1> ,l | | | > (fact (- n 1)) *** STOPPED IN fact, (console)@1.22 1> ,l | | > (* n (fact (- n 1))) | | 24 *** STOPPED IN fact, (console)@1.32 1> ,l | > (* n (fact (- n 1))) | 120 120 -- procedure: break PROC... -- procedure: unbreak PROC... The `break' procedure places a breakpoint on each of the specified procedures. When a procedure is called that has a breakpoint, the interpreter will enable single-stepping mode (as if `step' had been called). This typically causes the computation to stop soon inside the procedure if the stepping level is high enough. The `unbreak' procedure removes the breakpoints on the specified procedures. With no argument, `break' returns the list of procedures currently containing breakpoints. The void object is returned by `break' if it is passed one or more arguments. With no argument `unbreak' removes all the breakpoints and returns the void object. A breakpoint can be placed on a compiled procedure but only if it is bound to a global variable. For example: > (define (double x) (+ x x)) > (define (triple y) (- (double (double y)) y)) > (define (f z) (* (triple z) 10)) > (break double) > (break -) *** WARNING -- Rebinding global variable "-" to an interpreted procedure > (f 5) *** STOPPED IN double, (console)@1.21 1> ,b 0 double (console)@1:21 + 1 triple (console)@2:31 (double y) 2 f (console)@3:18 (triple z) 3 (interaction) (console)@6:1 (f 5) 1> ,e x = 5 1> ,c *** STOPPED IN double, (console)@1.21 1> ,c *** STOPPED IN f, (console)@3.29 1> ,c 150 > (break) (# #) > (unbreak) > (f 5) 150 -- procedure: generate-proper-tail-calls [NEW-VALUE] [Note: this procedure is DEPRECATED and will be removed in a future version of Gambit. Use the `proper-tail-calls' declaration instead.] The parameter object `generate-proper-tail-calls' is bound to a boolean value controlling how the interpreter handles tail calls. When it is bound to `#f' the interpreter will treat tail calls like nontail calls, that is a new continuation will be created for the call. This setting is useful for debugging, because when a primitive signals an error the location information will point to the call site of the primitive even if this primitive was called with a tail call. The initial value of this parameter object is `#t', which means that a tail call will reuse the continuation of the calling function. This parameter object only affects code that is subsequently processed by `load' or `eval', or entered at the REPL. For example: > (generate-proper-tail-calls) #t > (let loop ((i 1)) (if (< i 10) (loop (* i 2)) oops)) *** ERROR IN #, (console)@2.47 -- Unbound variable: oops 1> ,b 0 # (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 #, (console)@6.47 -- Unbound variable: oops 1> ,b 0 # (console)@6:47 oops 1 # (console)@6:32 (loop (* i 2)) 2 # (console)@6:32 (loop (* i 2)) 3 # (console)@6:32 (loop (* i 2)) 4 # (console)@6:32 (loop (* i 2)) 5 (interaction) (console)@6:1 ((letrec ((loop (lambda... -- procedure: display-environment-set! DISPLAY? [Note: this procedure is DEPRECATED and will be removed in a future version of Gambit. Use the parameter object `repl-display-environment?' instead.] This procedure sets a flag that controls the automatic display of the environment by the REPL. If DISPLAY? is true, the environment is displayed by the REPL before the prompt. The default setting is not to display the environment. -- procedure: repl-display-environment? DISPLAY? The parameter object `repl-display-environment?' is bound to a boolean value that controls the automatic display of the environment by the REPL. If DISPLAY? is true, the environment is displayed by the REPL before the prompt. This is particularly useful in single-stepping mode. The default setting is not to display the environment. -- procedure: display-dynamic-environment? DISPLAY? The parameter object `display-dynamic-environment?' is bound to a boolean value that controls wether the dynamic environment is displayed when the environment is displayed. The default setting is not to display the dynamic environment. -- procedure: pretty-print OBJ [PORT] This procedure pretty-prints OBJ on the port PORT. If it is not specified, PORT defaults to the current output-port. For example: > (pretty-print (let* ((x '(1 2 3 4)) (y (list x x x))) (list y y y))) (((1 2 3 4) (1 2 3 4) (1 2 3 4)) ((1 2 3 4) (1 2 3 4) (1 2 3 4)) ((1 2 3 4) (1 2 3 4) (1 2 3 4))) -- procedure: pp OBJ [PORT] This procedure pretty-prints OBJ on the port PORT. When OBJ is a procedure created by the interpreter or a procedure created by code compiled with the declaration `debug-source', the procedure's source code is displayed. If it is not specified, PORT defaults to the interaction channel (i.e. the output will appear at the REPL). For example: > (define (f g) (+ (time (g 100)) (time (g 1000)))) > (pp f) (lambda (g) (+ (##time (lambda () (g 100)) '(g 100)) (##time (lambda () (g 1000)) '(g 1000)))) -- procedure: gc-report-set! REPORT? This procedure controls the generation of reports during garbage collections. If the argument is true, a brief report of memory usage is generated after every garbage collection. It contains: the time taken for this garbage collection, the amount of memory allocated in megabytes since the program was started, the size of the heap in megabytes, the heap memory in megabytes occupied by live data, the proportion of the heap occupied by live data, and the number of bytes occupied by movable and nonmovable objects. 5.5 Console line-editing ======================== The console implements a simple Scheme-friendly line-editing user-interface that is enabled by default. It offers parentheses balancing, a history of previous commands, symbol completion, and several emacs-compatible keyboard commands. The user's input is displayed in a bold font and the output produced by the system is in a plain font. The history of previous commands is saved in the file `~/.gambit_history'. It is restored when a REPL is started. Symbol completion is triggered with the tab key. When the cursor is after a sequence of characters that can form a symbol, typing the tab key will search the symbol table for the first symbol (in alphabetical order) that begins with that sequence and insert that symbol. Typing the tab key in succession will cycle through all symbols with that prefix. When all possible symbols have been shown or there are no possible completions, the text reverts to the uncompleted symbol and the bell is rung. Here are the keyboard commands available (where the ``M-'' prefix means the escape key is typed and the ``C-'' prefix means the control key is pressed): `C-d' Generate an end-of-file when the line is empty, otherwise delete character at cursor. `delete or backspace' Delete character before cursor. `M-C-d' Delete word forward and keep a copy of this text on the clipboard. `M-delete' Delete word backward and keep a copy of this text on the clipboard. `M-backspace' Delete S-expression backward and keep a copy of this text on the clipboard. `C-a' Move cursor to beginning of line. `C-e' Move cursor to end of line. `C-b or left-arrow' Move cursor left one character. `M-b' Move cursor left one word. `M-C-b or `M-'left-arrow' Move cursor left one S-expression. `C-f or right-arrow' Move cursor right one character. `M-f' Move cursor right one word. `M-C-f or `M-'right-arrow' Move cursor right one S-expression. `C-p or `M-p' or up-arrow' Move to previous line in history. `C-n or `M-n' or down-arrow' Move to next line in history. `C-t' Transpose character at cursor with previous character. `M-t' Transpose word after cursor with previous word. `M-C-t' Transpose S-expression after cursor with previous S-expression. `C-l' Clear console and redraw line being edited. `C-nul' Set the mark to the cursor. `C-w' Delete the text between the cursor and the mark and keep a copy of this text on the clipboard. `C-k' Delete the text from the cursor to the end of the line and keep a copy of this text on the clipboard. `C-y' Paste the text that is on the clipboard. `F8' Same as typing `#||#,c;' (REPL command to continue the computation). `F9' Same as typing `#||#,-;' (REPL command to move to newer frame). `F10' Same as typing `#||#,+;' (REPL command to move to older frame). `F11' Same as typing `#||#,s;' (REPL command to step the computation). `F12' Same as typing `#||#,l;' (REPL command to leap the computation). On 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: (require 'gambit) You can then start an inferior Gambit process by typing `M-x run-scheme'. The commands provided in `cmuscheme' mode will be available in the Gambit interaction buffer (i.e. `*scheme*') and in buffers attached to Scheme source files. Here is a list of the most useful commands (for a complete list type `C-h m' in the Gambit interaction buffer): `C-x C-e' Evaluate the expression which is before the cursor (the expression will be copied to the Gambit interaction buffer). `C-c C-z' Switch to Gambit interaction buffer. `C-c C-l' Load a file (file attached to current buffer is default) using `(load FILE)'. `C-c C-k' Compile a file (file attached to current buffer is default) using `(compile-file FILE)'. The file `gambit.el' provides these additional commands: `F8 or C-c c' Continue the computation (same as typing `#||#,c;' to the REPL). `F9 or C-c ]' Move to newer frame (same as typing `#||#,-;' to the REPL). `F10 or C-c [' Move to older frame (same as typing `#||#,+;' to the REPL). `F11 or C-c s' Step the computation (same as typing `#||#,s;' to the REPL). `F12 or C-c l' Leap the computation (same as typing `#||#,l;' to the REPL). `C-c _' Removes the last window that was opened to highlight an expression. The two keystroke version of these commands can be shortened to `M-c', `M-[', `M-]', `M-s', `M-l', and `M-_' respectively by adding this line to your `.emacs' file: (setq gambit-repl-command-prefix "\e") This is more convenient to type than the two keystroke `C-c' based sequences but the purist may not like this because it does not follow normal Emacs conventions. Here is what a typical `.emacs' file will look like: (setq load-path ; add directory containing gambit.el (cons "/usr/local/Gambit/share/emacs/site-lisp" load-path)) (setq scheme-program-name "/tmp/gsi -: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 ===================================== -- procedure: transcript-on FILE -- procedure: transcript-off These procedures do nothing. -- procedure: call-with-current-continuation PROC -- procedure: call/cc PROC The procedure `call-with-current-continuation' is bound to the global variables `call-with-current-continuation' and `call/cc'. 6.2 Extensions to standard special forms ======================================== -- special form: lambda lambda-formals body -- special form: define (variable define-formals) body lambda-formals = `(' formal-argument-list `)' | r4rs-lambda-formals define-formals = formal-argument-list | r4rs-define-formals formal-argument-list = dsssl-formal-argument-list | rest-at-end-formal-argument-list dsssl-formal-argument-list = reqs opts rest keys rest-at-end-formal-argument-list = reqs opts keys rest | reqs opts keys `.' rest-formal-argument reqs = required-formal-argument* required-formal-argument = variable opts = `#!optional' optional-formal-argument* | empty optional-formal-argument = variable | `(' variable initializer `)' rest = `#!rest' rest-formal-argument | empty rest-formal-argument = variable keys = `#!key' keyword-formal-argument* | empty keyword-formal-argument = variable | `(' variable initializer `)' initializer = expression r4rs-lambda-formals = `(' variable* `)' | `(' variable+ `.' variable `)' | variable r4rs-define-formals = variable* | variable* `.' variable These forms are extended versions of the `lambda' and `define' special forms of standard Scheme. They allow the use of optional formal arguments, either positional or named, and support the syntax and semantics of the DSSSL standard. When the procedure introduced by a `lambda' (or `define') is applied to a list of actual arguments, the formal and actual arguments are processed as specified in the R4RS if the lambda-formals (or define-formals) is a r4rs-lambda-formals (or r4rs-define-formals). If the formal-argument-list matches dsssl-formal-argument-list or extended-formal-argument-list they are processed as follows: a. Variables in required-formal-arguments are bound to successive actual arguments starting with the first actual argument. It shall be an error if there are fewer actual arguments than required-formal-arguments. b. Next variables in optional-formal-arguments are bound to remaining actual arguments. If there are fewer remaining actual arguments than optional-formal-arguments, then the variables are bound to the result of evaluating initializer, if one was specified, and otherwise to `#f'. The initializer is evaluated in an environment in which all previous formal arguments have been bound. c. If `#!key' does not appear in the formal-argument-list and there is no rest-formal-argument then it shall be an error if there are any remaining actual arguments. d. If `#!key' does not appear in the formal-argument-list and there is a rest-formal-argument then the rest-formal-argument is bound to a list of all remaining actual arguments. e. If `#!key' appears in the formal-argument-list and there is no rest-formal-argument then there shall be an even number of remaining actual arguments. These are interpreted as a series of pairs, where the first member of each pair is a keyword specifying the argument name, and the second is the corresponding value. It shall be an error if the first member of a pair is not a keyword. It shall be an error if the argument name is not the same as a variable in a keyword-formal-argument. If the same argument name occurs more than once in the list of actual arguments, then the first value is used. If there is no actual argument for a particular keyword-formal-argument, then the variable is bound to the result of evaluating initializer if one was specified, and otherwise to `#f'. The initializer is evaluated in an environment in which all previous formal arguments have been bound. f. If `#!key' appears in the formal-argument-list and there is a rest-formal-argument before the `#!key' then there may be an even or odd number of remaining actual arguments and the rest-formal-argument is bound to a list of all remaining actual arguments. Then, these remaining actual arguments are scanned from left to right in pairs, stopping at the first pair whose first element is not a keyword. Each pair whose first element is a keyword matching the name of a keyword-formal-argument gives the value (i.e. the second element of the pair) of the corresponding formal argument. If the same argument name occurs more than once in the list of actual arguments, then the first value is used. If there is no actual argument for a particular keyword-formal-argument, then the variable is bound to the result of evaluating initializer if one was specified, and otherwise to `#f'. The initializer is evaluated in an environment in which all previous formal arguments have been bound. g. If `#!key' appears in the formal-argument-list and there is a rest-formal-argument after the `#!key' then there may be an even or odd number of remaining actual arguments. The remaining actual arguments are scanned from left to right in pairs, stopping at the first pair whose first element is not a keyword. Each pair shall have as its first element a keyword matching the name of a keyword-formal-argument; the second element gives the value of the corresponding formal argument. If the same argument name occurs more than once in the list of actual arguments, then the first value is used. If there is no actual argument for a particular keyword-formal-argument, then the variable is bound to the result of evaluating initializer if one was specified, and otherwise to `#f'. The initializer is evaluated in an environment in which all previous formal arguments have been bound. Finally, the rest-formal-argument is bound to the list of the actual arguments that were not scanned (i.e. after the last keyword/value pair). In all cases it is an error for a variable to appear more than once in a formal-argument-list. Note that this specification is compatible with the DSSSL language standard (i.e. a correct DSSSL program will have the same semantics when run with Gambit). It is unspecified whether variables receive their value by binding or by assignment. Currently the compiler and interpreter use different methods, which can lead to different semantics if `call-with-current-continuation' is used in an initializer. Note that this is irrelevant for DSSSL programs because `call-with-current-continuation' does not exist in DSSSL. For example: > ((lambda (#!rest x) x) 1 2 3) (1 2 3) > (define (f a #!optional b) (list a b)) > (define (g a #!optional (b a) #!key (k (* a b))) (list a b k)) > (define (h1 a #!rest r #!key k) (list a k r)) > (define (h2 a #!key k #!rest r) (list a k r)) > (f 1) (1 #f) > (f 1 2) (1 2) > (g 3) (3 3 9) > (g 3 4) (3 4 12) > (g 3 4 k: 5) (3 4 5) > (g 3 4 k: 5 k: 6) (3 4 5) > (h1 7) (7 #f ()) > (h1 7 k: 8 9) (7 8 (k: 8 9)) > (h1 7 k: 8 z: 9) (7 8 (k: 8 z: 9)) > (h2 7) (7 #f ()) > (h2 7 k: 8 9) (7 8 (9)) > (h2 7 k: 8 z: 9) *** ERROR IN (console)@17.1 -- Unknown keyword argument passed to procedure (h2 7 k: 8 z: 9) 6.3 Miscellaneous extensions ============================ -- procedure: vector-copy VECTOR 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 -- procedure: subvector VECTOR START END This procedure is the vector analog of the `substring' procedure. It returns a newly allocated vector formed from the elements of the vector VECTOR beginning with index START (inclusive) and ending with index END (exclusive). For example: > (subvector '#(a b c d e f) 3 5) #(d e) -- procedure: vector-append VECTOR... This procedure is the vector analog of the `string-append' procedure. It returns a newly allocated vector whose elements form the concatenation of the given vectors. For example: > (define v '#(1 2 3)) > (vector-append v v v) #(1 2 3 1 2 3 1 2 3) -- procedure: append-vectors LST 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) -- procedure: subvector-fill! VECTOR START END FILL This procedure is like `vector-fill!', but fills a selected part of the given vector. It sets the elements of the vector VECTOR, beginning with index START (inclusive) and ending with index END (exclusive) to FILL. The value returned is unspecified. For example: > (define v (vector 'a 'b 'c 'd 'e 'f)) > (subvector-fill! v 3 5 'x) > v #(a b c x x f) -- procedure: subvector-move! SRC-VECTOR SRC-START SRC-END DST-VECTOR DST-START This procedure replaces part of the contents of vector DST-VECTOR with part of the contents of vector SRC-VECTOR. It copies elements from SRC-VECTOR, beginning with index SRC-START (inclusive) and ending with index SRC-END (exclusive) to DST-VECTOR beginning with index DST-START (inclusive). The value returned is unspecified. For example: > (define v1 '#(1 2 3 4 5 6)) > (define v2 (vector 'a 'b 'c 'd 'e 'f)) > (subvector-move! v1 3 5 v2 1) > v2 #(a 4 5 d e f) -- procedure: vector-shrink! VECTOR K This procedure shortens the vector VECTOR so that its new size is K. The value returned is unspecified. For example: > (define v (vector 'a 'b 'c 'd 'e 'f)) > v #(a b c d e f) > (vector-shrink! v 3) > v #(a b c) -- procedure: append-strings LST 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" -- procedure: substring-fill! STRING START END FILL This procedure is like `string-fill!', but fills a selected part of the given string. It sets the elements of the string STRING, beginning with index START (inclusive) and ending with index END (exclusive) to FILL. The value returned is unspecified. For example: > (define s (string #\a #\b #\c #\d #\e #\f)) > (substring-fill! s 3 5 #\x) > s "abcxxf" -- procedure: substring-move! SRC-STRING SRC-START SRC-END DST-STRING DST-START This procedure replaces part of the contents of string DST-STRING with part of the contents of string SRC-STRING. It copies elements from SRC-STRING, beginning with index SRC-START (inclusive) and ending with index SRC-END (exclusive) to DST-STRING beginning with index DST-START (inclusive). The value returned is unspecified. For example: > (define s1 "123456") > (define s2 (string #\a #\b #\c #\d #\e #\f)) > (substring-move! s1 3 5 s2 1) > s2 "a45def" -- procedure: string-shrink! STRING K This procedure shortens the string STRING so that its new size is K. The value returned is unspecified. For example: > (define s (string #\a #\b #\c #\d #\e #\f)) > s "abcdef" > (string-shrink! s 3) > s "abc" -- procedure: box OBJ -- procedure: box? OBJ -- procedure: unbox BOX -- procedure: set-box! BOX OBJ These procedures implement the "box" data type. A box is a cell containing a single mutable field. The lexical syntax of a box containing the object OBJ is `#&OBJ' (*note Box syntax::). The procedure `box' returns a new box object whose content is initialized to OBJ. The procedure `box?' returns `#t' if OBJ is a box, and otherwise returns `#f'. The procedure `unbox' returns the content of the box BOX. The procedure `set-box!' changes the content of the box BOX to OBJ. The procedure `set-box!' returns an unspecified value. For example: > (define b (box 0)) > b #&0 > (define (inc!) (set-box! b (+ (unbox b) 1))) > (inc!) > b #&1 > (unbox b) 1 -- procedure: keyword? OBJ -- procedure: keyword->string KEYWORD -- procedure: string->keyword STRING These procedures implement the "keyword" data type. Keywords are similar to symbols but are self evaluating and distinct from the symbol data type. The lexical syntax of keywords is specified in *Note Keyword syntax::. The procedure `keyword?' returns `#t' if OBJ is a keyword, and otherwise returns `#f'. The procedure `keyword->string' returns the name of KEYWORD as a string. The procedure `string->keyword' returns the keyword whose name is STRING. For example: > (keyword? 'color) #f > (keyword? color:) #t > (keyword->string color:) "color" > (string->keyword "color") color: -- procedure: gensym [PREFIX] This procedure returns a new "uninterned symbol". Uninterned symbols are guaranteed to be distinct from the symbols generated by the procedures `read' and `string->symbol'. The symbol PREFIX is the prefix used to generate the new symbol's name. If it is not specified, the prefix defaults to `g'. For example: > (gensym) #:g0 > (gensym) #:g1 > (gensym 'star-trek-) #:star-trek-2 -- procedure: string->uninterned-symbol NAME [HASH] -- procedure: uninterned-symbol? OBJ The procedure `string->uninterned-symbol' returns a new uninterned symbol whose name is NAME and hash is HASH. The name must be a string and the hash must be a nonnegative fixnum. The procedure `uninterned-symbol?' returns `#t' when OBJ is a symbol that is uninterned and `#f' otherwise. For example: > (uninterned-symbol? (gensym)) #t > (string->uninterned-symbol "foo") #:foo: > (uninterned-symbol? (string->uninterned-symbol "foo")) #t > (uninterned-symbol? 'hello) #f > (uninterned-symbol? 123) #f -- procedure: string->uninterned-keyword NAME [HASH] -- procedure: uninterned-keyword? OBJ The procedure `string->uninterned-keyword' returns a new uninterned keyword whose name is NAME and hash is HASH. The name must be a string and the hash must be a nonnegative fixnum. The procedure `uninterned-keyword?' returns `#t' when OBJ is a keyword that is uninterned and `#f' otherwise. For example: > (string->uninterned-keyword "foo") #:foo: > (uninterned-keyword? (string->uninterned-keyword "foo")) #t > (uninterned-keyword? hello:) #f > (uninterned-keyword? 123) #f -- procedure: identity OBJ This procedure returns OBJ. -- procedure: void This procedure returns the void object. The read-eval-print loop prints nothing when the result is the void object. -- procedure: eval EXPR [ENV] The first parameter is a datum representing an expression. The `eval' procedure evaluates this expression in the global interaction environment and returns the result. If present, the second parameter is ignored (it is provided for compatibility with R5RS). For example: > (eval '(+ 1 2)) 3 > ((eval 'car) '(1 2)) 1 > (eval '(define x 5)) > x 5 -- special form: include file 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))) -- special form: define-macro (name define-formals) body Define name as a macro special form which expands into body. This form can only appear where a `define' form is acceptable. Macros are lexically scoped. The scope of a local macro definition extends from the definition to the end of the body of the surrounding binding construct. Macros defined at the top level of a Scheme module are only visible in that module. To have access to the macro definitions contained in a file, that file must be included 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))) -- special form: define-syntax name expander Define name as a macro special form whose expansion is specified by expander. This form is available only when the runtime option `-:s' is used. This option causes the loading of the `~~lib/syntax-case' support library, which is the Hieb and Dybvig portable `syntax-case' implementation which has been ported to the Gambit interpreter and compiler. Note that this implementation of `syntax-case' does not support special forms that are specific to Gambit. For example: $ gsi -:s Gambit v4.9.3 > (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 -- special form: declare declaration... This form introduces declarations to be used by the compiler (currently the interpreter ignores the declarations). This form can only appear where a `define' form is acceptable. Declarations are lexically scoped in the same way as macros. The following declarations are accepted by the compiler: `(DIALECT)' Use the given dialect's semantics. DIALECT can be: `ieee-scheme', `r4rs-scheme', `r5rs-scheme' or `gambit-scheme'. `(STRATEGY)' Select block compilation or separate compilation. In block compilation, the compiler assumes that global variables defined in the current file that are not mutated in the file will never be mutated. STRATEGY can be: `block' or `separate'. `([not] inline)' Allow (or disallow) inlining of user procedures. `([not] inline-primitives PRIMITIVE...)' The given primitives should (or should not) be inlined if possible (all primitives if none specified). `(inlining-limit N)' Select the degree to which the compiler inlines user procedures. N is the upper-bound, in percent, on code expansion that will result from inlining. Thus, a value of 300 indicates that the size of the program will not grow by more than 300 percent (i.e. it will be at most 4 times the size of the original). A value of 0 disables inlining. The size of a program is the total number of subexpressions it contains (i.e. the size of an expression is one plus the size of its immediate subexpressions). The following conditions must hold for a procedure to be inlined: inlining the procedure must not cause the size of the call site to grow more than specified by the inlining limit, the site of definition (the `define' or `lambda') and the call site must be declared as `(inline)', and the compiler must be able to find the definition of the procedure referred to at the call site (if the procedure is bound to a global variable, the definition site must have a `(block)' declaration). Note that inlining usually causes much less code expansion than specified by the inlining limit (an expansion around 10% is common for N=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] poll-on-return)' Generate (or don't generate) interrupt checks on procedure returns (when interrupt checking is enabled). This declaration has no effect on the behavior of interrupt checking on procedure calls, which is needed to guarantee that stack overflows are handled properly. `([not] debug)' Enable (or disable) the generation of debugging information. The kind of debugging information that is generated depends on the declarations `debug-location', `debug-source', and `debug-environments'. If any of the command line options `-debug', `-debug-location', `-debug-source' and `-debug-environments' are present, the `debug' declaration is initially enabled, otherwise it is initially disabled. When all kinds of debugging information are generated there is a substantial increase in the C compilation time and the size of the generated code. When compiling a 3000 line Scheme file it was observed that the total compilation time is 500% longer and the executable code is 150% bigger. `([not] debug-location)' Select (or deselect) source code location debugging information. When this declaration and the `debug' declaration are in effect, run time error messages indicate the location of the error in the source code file. If any of the command line options `-debug-source' and `-debug-environments' are present and `-debug-location' is absent, the `debug-location' declaration is initially disabled, otherwise it is initially enabled. When compiling a 3000 line Scheme file it was observed that the total compilation time is 200% longer and the executable code is 60% bigger. `([not] debug-source)' Select (or deselect) source code debugging information. When this declaration and the `debug' declaration are in effect, run time error messages indicate the source code, the backtraces are more precise, and the `pp' procedure will display the source code of compiled procedures. If any of the command line options `-debug-location' and `-debug-environments' are present and `-debug-source' is absent, the `debug-source' declaration is initially disabled, otherwise it is initially enabled. When compiling a 3000 line Scheme file it was observed that the total compilation time is 90% longer and the executable code is 90% bigger. `([not] debug-environments)' Select (or deselect) environment debugging information. When this declaration and the `debug' declaration are in effect, the debugger will have access to the environments of the continuations. In other words the local variables defined in compiled procedures (and not optimized away by the compiler) will be shown by the `,e' REPL command. If any of the command line options `-debug-location' and `-debug-source' are present and `-debug-environments' is absent, the `debug-environments' declaration is initially disabled, otherwise it is initially enabled. When compiling a 3000 line Scheme file it was observed that the total compilation time is 70% longer and the executable code is 40% bigger. `([not] proper-tail-calls)' Generate (or don't generate) proper tail calls. When proper tail calls are turned off, tail calls are handled like non-tail calls, that is a continuation frame will be created for all calls regardless of their kind. This is useful for debugging because the caller of a procedure will be visible in the backtrace produced by the REPL's `,b' command even when the call is a tail call. Be advised that this does cause stack space to be consumed for tail calls which may cause the stack to overflow when performing long iterations with tail calls (whether they are expressed with a `letrec', named `let', `do', or other form). `([not] generative-lambda)' Force (or don't force) the creation of fresh closures when evaluating lambda-expressions. A fresh closure is always created when a lambda-expression has at least one free variable (that has not been eliminated by dead-code elimination or other compiler optimization) or when the generative-lambda declaration is turned on. When a lambda-expression has no free variables and the generative-lambda declaration is turned off, the value of the lambda-expression may be the same procedure (in the sense of `eq?'). `([not] optimize-dead-local-variables)' Remove (or preserve) the dead local variables in the environment. Preserving the dead local variables is useful for debugging because continuations will contain the dead variables. Thus, if the code is also compiled with the declaration `debug-environments' the `,e', `,ed', `,be', and `,bed' REPL commands will display the dead variables. On the other hand, preserving the dead local variables may change the space complexity of the program (i.e. some of the data that would normally be reclaimed by the garbage collector will not be). Note that due to other compiler optimizations some dead local variables may be removed regardless of this declaration. `([not] optimize-dead-definitions VAR...)' Remove (or preserve) the dead toplevel definitions of the given global variables (all global variables if none specified). A toplevel definition is dead if it is not referenced by toplevel expressions of the program or toplevel definitions that aren't dead (regardless of the evaluation of its expression causing a side-effect). When a module is separately compiled and some of its definitions are only used by other modules, this declaration must be used with care to keep definitions that are used by other modules, for example if `foo' is referenced in another module the following declaration should be used: `(declare (not optimize-dead-definitions foo))'. `(NUMBER-TYPE PRIMITIVE...)' Numeric arguments and result of the specified primitives are known to be of the given type (all primitives if none specified). NUMBER-TYPE can be: `generic', `fixnum', or `flonum'. `(MOSTLY-NUMBER-TYPE PRIMITIVE...)' Numeric arguments and result of the specified primitives are expected to be most often of the given type (all primitives if none specified). MOSTLY-NUMBER-TYPE can be: `mostly-generic', `mostly-fixnum', `mostly-fixnum-flonum', `mostly-flonum', or `mostly-flonum-fixnum'. The default declarations used by the compiler are equivalent to: (declare (gambit-scheme) (separate) (inline) (inline-primitives) (inlining-limit 370) (constant-fold) (lambda-lift) (not standard-bindings) (not extended-bindings) (run-time-bindings) (safe) (interrupts-enabled) (not poll-on-return) (not debug) ;; depends on debugging command line options (debug-location) ;; depends on debugging command line options (debug-source) ;; depends on debugging command line options (debug-environments) ;; depends on debugging command line options (proper-tail-calls) (not generative-lambda) (optimize-dead-local-variables) (not optimize-dead-definitions) (generic) (mostly-fixnum-flonum) ) These declarations are compatible with the semantics of R5RS Scheme and includes a few procedures from R6RS (mainly fixnum specific and flonum specific procedures). Typically used declarations that enhance performance, at the cost of violating the R5RS Scheme semantics, are: `(standard-bindings)', `(block)', `(not safe)' and `(fixnum)'. 6.4 Undocumented extensions =========================== The procedures in this section are not yet documented. -- procedure: continuation? OBJ -- procedure: continuation-capture PROC -- procedure: continuation-graft CONT PROC OBJ... -- procedure: continuation-return CONT OBJ... These procedures provide access to internal first-class continuations which are represented using continuation objects distinct from procedures. The procedure `continuation?' returns `#t' when OBJ is a continuation object and `#f' otherwise. The procedure `continuation-capture' is similar to the `call/cc' procedure but it represents the continuation with a continuation object. The PROC parameter must be a procedure accepting a single argument. The procedure `continuation-capture' reifies its continuation and calls PROC with the corresponding continuation object as its sole argument. Like for `call/cc', the implicit continuation of the call to PROC is the implicit continuation of the call to `continuation-capture'. The procedure `continuation-graft' performs a procedure call to the procedure PROC with arguments OBJ... and the implicit continuation corresponding to the continuation object CONT. The current continuation of the call to procedure `continuation-graft' is ignored. The procedure `continuation-return' invokes the implicit continuation corresponding to the continuation object CONT with the result(s) OBJ.... This procedure can be easily defined in terms of `continuation-graft': (define (continuation-return cont . objs) (continuation-graft (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-return x 55) (33 . 55) > (continuation-graft x (lambda () (expt 2 10))) (33 . 1024) > (continuation-graft x expt 2 10) (33 . 1024) > (continuation-graft x (lambda () (p))) (33 . 22) > (define (map-sqrt1 lst) (call/cc (lambda (k) (map (lambda (x) (if (< x 0) (k 'error) (sqrt x))) lst)))) > (map-sqrt1 '(1 4 9)) (1 2 3) > (map-sqrt1 '(1 -1 9)) error > (define (map-sqrt2 lst) (continuation-capture (lambda (c) (map (lambda (x) (if (< x 0) (continuation-return c 'error) (sqrt x))) lst)))) > (map-sqrt2 '(1 4 9)) (1 2 3) > (map-sqrt2 '(1 -1 9)) error -- procedure: display-exception EXC [PORT] -- procedure: display-exception-in-context EXC CONT [PORT] -- procedure: display-procedure-environment PROC [PORT] -- procedure: display-continuation-environment CONT [PORT] -- procedure: display-continuation-dynamic-environment CONT [PORT] -- procedure: display-continuation-backtrace CONT [PORT [ALL-FRAMES? [DISPLAY-ENV? [MAX-HEAD [MAX-TAIL [DEPTH]]]]]] The procedure `display-continuation-backtrace' displays the frames of the continuation corresponding to the continuation object CONT on the port PORT. If it is not specified, PORT defaults to the current output-port. The frames are displayed in the same format as the REPL's `,b' command. The parameter 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 -- procedure: make-thread-group [NAME [THREAD-GROUP]] -- procedure: thread-group? OBJ -- procedure: thread-group-name THREAD-GROUP -- procedure: thread-group-parent THREAD-GROUP -- procedure: thread-group-resume! THREAD-GROUP -- procedure: thread-group-suspend! THREAD-GROUP -- procedure: thread-group-terminate! THREAD-GROUP -- procedure: thread-group->thread-group-list THREAD-GROUP -- procedure: thread-group->thread-group-vector THREAD-GROUP -- procedure: thread-group->thread-list THREAD-GROUP -- procedure: thread-group->thread-vector THREAD-GROUP -- procedure: thread-state THREAD -- procedure: thread-state-uninitialized? THREAD-STATE -- procedure: thread-state-initialized? THREAD-STATE -- procedure: thread-state-running? THREAD-STATE -- procedure: thread-state-running-processor THREAD-STATE -- procedure: thread-state-waiting? THREAD-STATE -- procedure: thread-state-waiting-for THREAD-STATE -- procedure: thread-state-waiting-timeout THREAD-STATE -- procedure: thread-state-normally-terminated? THREAD-STATE -- procedure: thread-state-normally-terminated-result THREAD-STATE -- procedure: thread-state-abnormally-terminated? THREAD-STATE -- procedure: thread-state-abnormally-terminated-reason THREAD-STATE -- procedure: top [TIMEOUT [THREAD-GROUP [PORT]]] -- procedure: thread-interrupt! THREAD [THUNK] -- procedure: thread-suspend! THREAD -- procedure: thread-resume! THREAD -- procedure: thread-thread-group THREAD -- special form: define-type-of-thread name field... -- procedure: thread-init! THREAD THUNK [NAME [THREAD-GROUP]] -- procedure: initialized-thread-exception? OBJ -- procedure: initialized-thread-exception-procedure EXC -- procedure: initialized-thread-exception-arguments EXC -- procedure: uninitialized-thread-exception? OBJ -- procedure: uninitialized-thread-exception-procedure EXC -- procedure: uninitialized-thread-exception-arguments EXC -- procedure: inactive-thread-exception? OBJ -- procedure: inactive-thread-exception-procedure EXC -- procedure: inactive-thread-exception-arguments EXC -- procedure: rpc-remote-error-exception? OBJ -- procedure: rpc-remote-error-exception-procedure EXC -- procedure: rpc-remote-error-exception-arguments EXC -- procedure: rpc-remote-error-exception-message EXC -- procedure: processor? OBJ -- procedure: current-processor -- procedure: processor-id PROCESSOR -- procedure: timeout->time TIMEOUT -- procedure: current-second -- procedure: current-jiffy -- procedure: jiffies-per-second -- procedure: get-environment-variable NAME -- procedure: get-environment-variables -- procedure: open-dummy -- procedure: port-settings-set! PORT SETTINGS -- procedure: port-io-exception-handler-set! PORT HANDLER -- procedure: input-port-bytes-buffered PORT -- procedure: input-port-characters-buffered PORT -- procedure: nonempty-input-port-character-buffer-exception? OBJ -- procedure: nonempty-input-port-character-buffer-exception-arguments EXC -- procedure: nonempty-input-port-character-buffer-exception-procedure EXC -- procedure: repl-input-port -- procedure: repl-output-port -- procedure: console-port -- procedure: current-user-interrupt-handler [HANDLER] -- procedure: defer-user-interrupts -- procedure: primordial-exception-handler EXC -- procedure: err-code->string CODE -- procedure: foreign? OBJ -- procedure: foreign-tags FOREIGN -- procedure: foreign-address FOREIGN -- procedure: foreign-release! FOREIGN -- procedure: foreign-released? FOREIGN -- procedure: invalid-hash-number-exception? OBJ -- procedure: invalid-hash-number-exception-procedure EXC -- procedure: invalid-hash-number-exception-arguments EXC -- procedure: tcp-client-local-socket-info TCP-CLIENT-PORT -- procedure: tcp-client-peer-socket-info TCP-CLIENT-PORT -- procedure: tcp-server-socket-info TCP-SERVER-PORT -- procedure: socket-info? OBJ -- procedure: socket-info-address SOCKET-INFO -- procedure: socket-info-family SOCKET-INFO -- procedure: socket-info-port-number SOCKET-INFO -- procedure: system-version -- procedure: system-version-string -- procedure: system-type -- procedure: system-type-string -- procedure: configure-command-string -- procedure: system-stamp -- special form: future EXPR -- procedure: touch OBJ -- procedure: tty? OBJ -- procedure: tty-history TTY -- procedure: tty-history-set! TTY HISTORY -- procedure: tty-history-max-length-set! TTY N -- procedure: tty-paren-balance-duration-set! TTY DURATION -- procedure: tty-text-attributes-set! TTY ATTRIBUTES -- procedure: tty-mode-set! TTY MODE -- procedure: tty-type-set! TTY TYPE -- procedure: with-input-from-port PORT THUNK -- procedure: with-output-to-port PORT THUNK -- procedure: input-port-char-position PORT -- procedure: output-port-char-position PORT -- procedure: open-event-queue N -- procedure: main ... -- special form: define-record-type ... -- special form: define-type ... -- special form: namespace ... -- special form: this-source-file -- special form: receive ... -- special form: cond-expand ... -- special form: define-cond-expand-feature IDENT ... -- special form: case-lambda ... -- procedure: list-set LIST K VAL -- procedure: list-set! LIST K VAL -- procedure: fold PROC BASE LIST ... -- procedure: fold-right PROC BASE LIST ... -- procedure: iota COUNT [START [STEP]] -- procedure: circular-list X Y... -- procedure: cons* X Y... -- procedure: list-copy LIST -- procedure: list-tabulate N INIT-PROC -- procedure: make-list N [FILL] -- procedure: reverse! LIST -- procedure: xcons D A -- procedure: take X I -- procedure: drop X I -- procedure: last PAIR -- procedure: last-pair PAIR -- procedure: finite? X -- procedure: infinite? X -- procedure: nan? X -- undefined: six.! -- special form: six.!x X -- special form: six.&x X -- special form: six.*x X -- special form: six.++x X -- special form: six.+x X -- special form: six.-x X -- special form: six.-x X -- special form: six.arrow EXPR IDENT -- undefined: six.break -- special form: six.call FUNC ARG... -- undefined: six.case -- undefined: six.clause -- special form: six.compound STATEMENT... -- special form: six.cons X Y -- undefined: six.continue -- special form: six.define-procedure IDENT PROC -- special form: six.define-variable IDENT TYPE DIMS INIT -- special form: six.do-while STAT EXPR -- special form: six.dot EXPR IDENT -- special form: six.for STAT1 EXPR2 EXPR3 STAT2 -- undefined: six.goto -- special form: six.identifier IDENT -- special form: six.if EXPR STAT1 [STAT2] -- special form: six.index EXPR1 EXPR2 -- undefined: six.label -- special form: six.list X Y -- special form: six.literal VALUE -- procedure: six.make-array INIT DIM... -- special form: six.new IDENT ARG... -- special form: six.null -- special form: six.prefix DATUM -- special form: six.procedure TYPE PARAMS STAT -- special form: six.procedure-body STAT... -- undefined: six.return -- undefined: six.switch -- special form: six.while EXPR STAT... -- special form: six.x!=y X Y -- special form: six.x%=y X Y -- special form: six.x%y X Y -- special form: six.x&&y X Y -- special form: six.x&=y X Y -- special form: six.x&y X Y -- special form: six.x*=y X Y -- special form: six.x*y X Y -- special form: six.x++ X -- special form: six.x+=y X Y -- special form: six.x+y X Y -- special form: |six.x,y| X Y -- special form: six.x- X -- special form: six.x-=y X Y -- special form: six.x-y X Y -- special form: six.x/=y X Y -- special form: six.x/y X Y -- undefined: six.x:-y X Y -- special form: six.x:=y X Y -- special form: six.x:y X Y -- special form: six.x<<=y X Y -- special form: six.x<=y X Y -- special form: six.x>>=y X Y -- special form: six.x>>y X Y -- special form: six.x>y X Y -- special form: six.x?y:z X Y Z -- special form: six.x^=y X Y -- special form: six.x^y X Y -- special form: |six.x\|=y| X Y -- special form: |six.x\|y| X Y -- special form: |six.x\|\|y| X Y -- special form: six.~x X 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 ====================================== -- procedure: char->integer CHAR -- procedure: integer->char N The procedure `char->integer' returns the Unicode encoding of the character CHAR. The procedure `integer->char' returns the character whose Unicode encoding is the exact integer N. For example: > (char->integer #\!) 33 > (integer->char 65) #\A > (integer->char (char->integer #\u1234)) #\u1234 > (integer->char #xd800) *** ERROR IN (console)@4.1 -- (Argument 1) Out of range (integer->char 55296) -- procedure: char=? CHAR1... -- procedure: char? CHAR1... -- procedure: char<=? CHAR1... -- procedure: char>=? CHAR1... -- procedure: char-ci=? CHAR1... -- procedure: char-ci? CHAR1... -- procedure: char-ci<=? CHAR1... -- procedure: char-ci>=? CHAR1... These procedures take any number of arguments including no argument. This is useful to test if the elements of a list are sorted in a particular order. For example, testing that the list of characters `lst' is sorted in nondecreasing order can be done with the call `(apply char? STRING1... -- procedure: string<=? STRING1... -- procedure: string>=? STRING1... -- procedure: string-ci=? STRING1... -- procedure: string-ci? STRING1... -- procedure: string-ci<=? STRING1... -- procedure: string-ci>=? STRING1... These procedures take any number of arguments including no argument. This is useful to test if the elements of a list are sorted in a particular order. For example, testing that the list of strings `lst' is sorted in nondecreasing order can be done with the call `(apply string X1... -- procedure: <= X1... -- procedure: >= X1... These procedures take any number of arguments including no argument. This is useful to test if the elements of a list are sorted in a particular order. For example, testing that the list of numbers `lst' is sorted in nondecreasing order can be done with the call `(apply < lst)'. 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 ==================================== -- procedure: integer-sqrt N This procedure returns the integer part of the square root of the nonnegative exact integer N. For example: > (integer-sqrt 123) 11 -- procedure: integer-nth-root N1 N2 This procedure returns the integer part of N1 raised to the power 1/N2, where N1 is a nonnegative exact integer and N2 is a positive exact integer. For example: > (integer-nth-root 100 3) 4 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. -- procedure: arithmetic-shift N1 N2 This procedure returns N1 shifted to the left by N2 bits, that is `(floor (* N1 (expt 2 N2)))'. Both N1 and N2 must be exact integers. For example: > (arithmetic-shift 1000 7) ; n1=...0000001111101000 128000 > (arithmetic-shift 1000 -6) ; n1=...0000001111101000 15 > (arithmetic-shift -23 -3) ; n1=...1111111111101001 -3 -- procedure: bitwise-merge N1 N2 N3 This procedure returns an exact integer whose bits combine the bits from N2 and N3 depending on N1. The bit at index I of the result depends only on the bits at index I in N1, N2 and N3: it is equal to the bit in N2 when the bit in N1 is 0 and it is equal to the bit in N3 when the bit in N1 is 1. All arguments must be exact integers. For example: > (bitwise-merge -4 -11 10) ; ...11111100 ...11110101 ...00001010 9 > (bitwise-merge 12 -11 10) ; ...00001100 ...11110101 ...00001010 -7 -- procedure: bitwise-and N... This procedure returns the bitwise "and" of the exact integers N.... The value -1 is returned when there are no arguments. For example: > (bitwise-and 6 12) ; ...00000110 ...00001100 4 > (bitwise-and 6 -4) ; ...00000110 ...11111100 4 > (bitwise-and -6 -4) ; ...11111010 ...11111100 -8 > (bitwise-and) -1 -- procedure: bitwise-ior N... This procedure returns the bitwise "inclusive-or" of the exact integers N.... The value 0 is returned when there are no arguments. For example: > (bitwise-ior 6 12) ; ...00000110 ...00001100 14 > (bitwise-ior 6 -4) ; ...00000110 ...11111100 -2 > (bitwise-ior -6 -4) ; ...11111010 ...11111100 -2 > (bitwise-ior) 0 -- procedure: bitwise-xor N... This procedure returns the bitwise "exclusive-or" of the exact integers N.... The value 0 is returned when there are no arguments. For example: > (bitwise-xor 6 12) ; ...00000110 ...00001100 10 > (bitwise-xor 6 -4) ; ...00000110 ...11111100 -6 > (bitwise-xor -6 -4) ; ...11111010 ...11111100 6 > (bitwise-xor) 0 -- procedure: bitwise-not N This procedure returns the bitwise complement of the exact integer N. For example: > (bitwise-not 3) ; ...00000011 -4 > (bitwise-not -1) ; ...11111111 0 -- procedure: bit-count N This procedure returns the bit count of the exact integer N. If N is nonnegative, the bit count is the number of 1 bits in the two's complement representation of N. If N is negative, the bit count is the number of 0 bits in the two's complement representation of N. For example: > (bit-count 0) ; ...00000000 0 > (bit-count 1) ; ...00000001 1 > (bit-count 2) ; ...00000010 1 > (bit-count 3) ; ...00000011 2 > (bit-count 4) ; ...00000100 1 > (bit-count -23) ; ...11101001 3 -- procedure: integer-length N This procedure returns the bit length of the exact integer N. If N is a positive integer the bit length is one more than the index of the highest 1 bit (the least significant bit is at index 0). If N is a negative integer the bit length is one more than the index of the highest 0 bit. If N is zero, the bit length is 0. For example: > (integer-length 0) ; ...00000000 0 > (integer-length 1) ; ...00000001 1 > (integer-length 2) ; ...00000010 2 > (integer-length 3) ; ...00000011 2 > (integer-length 4) ; ...00000100 3 > (integer-length -23) ; ...11101001 5 -- procedure: bit-set? N1 N2 This procedure returns a boolean indicating if the bit at index N1 of N2 is set (i.e. equal to 1) or not. Both N1 and N2 must be exact integers, and N1 must be nonnegative. For example: > (map (lambda (i) (bit-set? i -23)) ; ...11101001 '(7 6 5 4 3 2 1 0)) (#t #t #t #f #t #f #f #t) -- procedure: any-bits-set? N1 N2 This procedure returns a boolean indicating if the bitwise and of N1 and N2 is different from zero or not. This procedure is implemented more efficiently than the naive definition: (define (any-bits-set? n1 n2) (not (zero? (bitwise-and n1 n2)))) For example: > (any-bits-set? 5 10) ; ...00000101 ...00001010 #f > (any-bits-set? -23 32) ; ...11101001 ...00100000 #t -- procedure: all-bits-set? N1 N2 This procedure returns a boolean indicating if the bitwise and of N1 and N2 is equal to N1 or not. This procedure is implemented more efficiently than the naive definition: (define (all-bits-set? n1 n2) (= n1 (bitwise-and n1 n2))) For example: > (all-bits-set? 1 3) ; ...00000001 ...00000011 #t > (all-bits-set? 7 3) ; ...00000111 ...00000011 #f -- procedure: first-bit-set N This procedure returns the bit index of the least significant bit of N equal to 1 (which is also the number of 0 bits that are below the least significant 1 bit). This procedure returns `-1' when N is zero. For example: > (first-bit-set 24) ; ...00011000 3 > (first-bit-set 0) ; ...00000000 -1 -- procedure: extract-bit-field N1 N2 N3 -- procedure: test-bit-field? N1 N2 N3 -- procedure: clear-bit-field N1 N2 N3 -- procedure: replace-bit-field N1 N2 N3 N4 -- procedure: copy-bit-field N1 N2 N3 N4 These procedures operate on a bit-field which is N1 bits wide starting at bit index N2. All arguments must be exact integers and N1 and N2 must be nonnegative. The procedure `extract-bit-field' returns the bit-field of N3 shifted to the right so that the least significant bit of the bit-field is the least significant bit of the result. The procedure `test-bit-field?' returns `#t' if any bit in the bit-field of N3 is equal to 1, otherwise `#f' is returned. The procedure `clear-bit-field' returns N3 with all bits in the bit-field replaced with 0. The procedure `replace-bit-field' returns N4 with the bit-field replaced with the least-significant N1 bits of N3. The procedure `copy-bit-field' returns N4 with the bit-field replaced with the (same index and size) bit-field in N3. For example: > (extract-bit-field 5 2 -37) ; ...11011011 22 > (test-bit-field? 5 2 -37) ; ...11011011 #t > (test-bit-field? 1 2 -37) ; ...11011011 #f > (clear-bit-field 5 2 -37) ; ...11011011 -125 > (replace-bit-field 5 2 -6 -37) ; ...11111010 ...11011011 -21 > (copy-bit-field 5 2 -6 -37) ; ...11111010 ...11011011 -5 9.5 Fixnum specific operations ============================== -- procedure: fixnum? OBJ -- procedure: fx* N1... -- procedure: fx+ N1... -- procedure: fx- N1 N2... -- procedure: fx< N1... -- procedure: fx<= N1... -- procedure: fx= N1... -- procedure: fx> N1... -- procedure: fx>= N1... -- procedure: fxabs N -- procedure: fxand N1... -- procedure: fxarithmetic-shift N1 N2 -- procedure: fxarithmetic-shift-left N1 N2 -- procedure: fxarithmetic-shift-right N1 N2 -- procedure: fxbit-count N -- procedure: fxbit-set? N1 N2 -- procedure: fxeven? N -- procedure: fxfirst-bit-set N -- procedure: fxif N1 N2 N3 -- procedure: fxior N1... -- procedure: fxlength N -- procedure: fxmax N1 N2... -- procedure: fxmin N1 N2... -- procedure: fxmodulo N1 N2 -- procedure: fxnegative? N -- procedure: fxnot N -- procedure: fxodd? N -- procedure: fxpositive? N -- procedure: fxquotient N1 N2 -- procedure: fxremainder N1 N2 -- procedure: fxwrap* N1... -- procedure: fxwrap+ N1... -- procedure: fxwrap- N1 N2... -- procedure: fxwrapabs N -- procedure: fxwraparithmetic-shift N1 N2 -- procedure: fxwraparithmetic-shift-left N1 N2 -- procedure: fxwraplogical-shift-right N1 N2 -- procedure: fxwrapquotient N1 N2 -- procedure: fxxor N1... -- procedure: fxzero? N -- procedure: fixnum-overflow-exception? OBJ -- procedure: fixnum-overflow-exception-procedure EXC -- procedure: fixnum-overflow-exception-arguments EXC Fixnum-overflow-exception objects are raised by some of the fixnum specific procedures when the result is larger than can fit in a fixnum. The parameter EXC must be a fixnum-overflow-exception object. The procedure `fixnum-overflow-exception?' returns `#t' when OBJ is a fixnum-overflow-exception object and `#f' otherwise. The procedure `fixnum-overflow-exception-procedure' returns the procedure that raised EXC. The procedure `fixnum-overflow-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (fixnum-overflow-exception? exc) (list (fixnum-overflow-exception-procedure exc) (fixnum-overflow-exception-arguments exc)) 'not-fixnum-overflow-exception)) > (with-exception-catcher handler (lambda () (fx* 100000 100000))) (# (100000 100000)) 9.6 Flonum specific operations ============================== -- procedure: flonum? OBJ -- procedure: fixnum->flonum N -- procedure: fl* X1... -- procedure: fl+ X1... -- procedure: fl- X1 X2... -- procedure: fl/ X1 X2 -- procedure: fl< X1... -- procedure: fl<= X1... -- procedure: fl= X1... -- procedure: fl> X1... -- procedure: fl>= X1... -- procedure: flabs X -- procedure: flacos X -- procedure: flasin X -- procedure: flatan X -- procedure: flatan Y X -- procedure: flceiling X -- procedure: flcos X -- procedure: fldenominator X -- procedure: fleven? X -- procedure: flexp X -- procedure: flexpt X Y -- procedure: flfinite? X -- procedure: flfloor X -- procedure: flinfinite? X -- procedure: flinteger? X -- procedure: fllog X -- procedure: flmax X1 X2... -- procedure: flmin X1 X2... -- procedure: flnan? X -- procedure: flnegative? X -- procedure: flnumerator X -- procedure: flodd? X -- procedure: flpositive? X -- procedure: flround X -- procedure: flsin X -- procedure: flsqrt X -- procedure: fltan X -- procedure: fltruncate X -- procedure: flzero? X 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. -- variable: default-random-source The global variable `default-random-source' is bound to the random source object which is used by the `random-integer', `random-real', `random-u8vector' and `random-f64vector' procedures. -- procedure: random-integer N This procedure returns a pseudo random exact integer in the range 0 to N-1. The random source object in the global variable `default-random-source' is used to generate this number. The parameter N must be a positive exact integer. For example: > (random-integer 100) 24 > (random-integer 100) 2 > (random-integer 10000000000000000000000000000000000000000) 6143360270902284438072426748425263488507 -- procedure: random-real This procedure returns a pseudo random inexact real between, but not including, 0 and 1. The random source object in the global variable `default-random-source' is used to generate this number. For example: > (random-real) .24230672079133753 > (random-real) .02317001922506932 -- procedure: random-u8vector N This procedure returns a u8vector of length N containing pseudo random exact integers in the range 0 to 255. The random source object in the global variable `default-random-source' is used to generate these numbers. The parameter N must be a nonnegative exact integer. For example: > (random-u8vector 10) #u8(200 53 29 202 3 85 208 187 73 219) -- procedure: random-f64vector N This procedure returns a f64vector of length N containing pseudo random inexact reals between, but not including, 0 and 1. The random source object in the global variable `default-random-source' is used to generate these numbers. The parameter N must be a nonnegative exact integer. For example: > (random-f64vector 3) #f64(.7145854494613069 .47089632669147946 .5400124875182746) -- procedure: make-random-source This procedure returns a new random source object initialized to a predetermined state (to initialize to a pseudo random state the procedure `random-source-randomize!' should be called). For example: > (define rs (make-random-source)) > ((random-source-make-integers rs) 10000000) 8583952 -- procedure: random-source? OBJ This procedure returns `#t' when OBJ is a random source object and `#f' otherwise. For example: > (random-source? default-random-source) #t > (random-source? 123) #f -- procedure: random-source-state-ref RANDOM-SOURCE -- procedure: random-source-state-set! RANDOM-SOURCE STATE The procedure `random-source-state-ref' extracts the state of the random source object RANDOM-SOURCE and returns a vector containing the state. The procedure `random-source-state-set!' restores the state of the random source object RANDOM-SOURCE to STATE which must be a vector returned from a call to the procedure `random-source-state-ref'. For example: > (define s (random-source-state-ref default-random-source)) > (random-integer 10000000000000000000000000000000000000000) 7583880188903074396261960585615270693321 > (random-source-state-set! default-random-source s) > (random-integer 10000000000000000000000000000000000000000) 7583880188903074396261960585615270693321 -- procedure: random-source-randomize! RANDOM-SOURCE -- procedure: random-source-pseudo-randomize! RANDOM-SOURCE I J These procedures change the state of the random source object RANDOM-SOURCE. The procedure `random-source-randomize!' sets the random source object to a state that depends on the current time (which for typical uses can be considered to randomly initialize the state). The procedure `random-source-pseudo-randomize!' sets the random source object to a state that is determined only by the current state and the nonnegative exact integers I and J. For both procedures the value returned is unspecified. For example: > (define s (random-source-state-ref default-random-source)) > (random-source-pseudo-randomize! default-random-source 5 99) > (random-integer 10000000000000000000000000000000000000000) 9816755163910623041601722050112674079767 > (random-source-state-set! default-random-source s) > (random-source-pseudo-randomize! default-random-source 5 99) > (random-integer 10000000000000000000000000000000000000000) 9816755163910623041601722050112674079767 > (random-source-pseudo-randomize! default-random-source 5 99) > (random-integer 10000000000000000000000000000000000000000) 9816755163910623041601722050112674079767 > (random-source-state-set! default-random-source s) > (random-source-randomize! default-random-source) > (random-integer 10000000000000000000000000000000000000000) 2271441220851914333384493143687768110622 > (random-source-state-set! default-random-source s) > (random-source-randomize! default-random-source) > (random-integer 10000000000000000000000000000000000000000) 6247966138948323029033944059178072366895 -- procedure: random-source-make-integers RANDOM-SOURCE This procedure returns a procedure for generating pseudo random exact integers using the random source object RANDOM-SOURCE. The returned procedure accepts a single parameter N, a positive exact integer, and returns a pseudo random exact integer in the range 0 to N-1. For example: > (define rs (make-random-source)) > (define ri (random-source-make-integers rs)) > (ri 10000000) 8583952 > (ri 10000000) 2879793 -- procedure: random-source-make-reals RANDOM-SOURCE [PRECISION] This procedure returns a procedure for generating pseudo random inexact reals using the random source object RANDOM-SOURCE. The returned procedure accepts no parameters and returns a pseudo random inexact real between, but not including, 0 and 1. The optional parameter PRECISION specifies an upper bound on the minimum amount by which two generated pseudo-random numbers can be separated. For example: > (define rs (make-random-source)) > (define rr (random-source-make-reals rs)) > (rr) .857402537562821 > (rr) .2876463473845367 -- procedure: random-source-make-u8vectors RANDOM-SOURCE This procedure returns a procedure for generating pseudo random u8vectors using the random source object RANDOM-SOURCE. The returned procedure accepts a single parameter N, a nonnegative exact integer, and returns a u8vector of length N containing pseudo random exact integers in the range 0 to 255. For example: > (define rs (make-random-source)) > (define rv (random-source-make-u8vectors rs)) > (rv 10) #u8(200 53 29 202 3 85 208 187 73 219) > (rv 10) #u8(113 8 182 120 138 103 53 192 40 176) -- procedure: random-source-make-f64vectors RANDOM-SOURCE [PRECISION] This procedure returns a procedure for generating pseudo random f64vectors using the random source object RANDOM-SOURCE. The returned procedure accepts a single parameter N, a nonnegative exact integer, and returns an f64vector of length N containing pseudo random inexact reals between, but not including, 0 and 1. The optional parameter PRECISION specifies an upper bound on the minimum amount by which two generated pseudo-random numbers can be separated. For example: > (define rs (make-random-source)) > (define rv (random-source-make-f64vectors rs)) > (rv 3) #f64(.7342236104231586 .2876463473845367 .8574025375628211) > (rv 3) #f64(.013863292728449427 .33449296573515447 .8162050798467028) 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 *Note Homogeneous vector syntax::. The procedures available for homogeneous vectors, listed below, are the analog of the normal vector/string procedures for each of the homogeneous vector types. -- procedure: s8vector? OBJ -- procedure: make-s8vector K [FILL] -- procedure: s8vector EXACT-INT8... -- procedure: s8vector-length S8VECTOR -- procedure: s8vector-ref S8VECTOR K -- procedure: s8vector-set S8VECTOR K EXACT-INT8 -- procedure: s8vector-set! S8VECTOR K EXACT-INT8 -- procedure: s8vector->list S8VECTOR -- procedure: list->s8vector LIST-OF-EXACT-INT8 -- procedure: s8vector-fill! S8VECTOR FILL [START [END]] -- procedure: subs8vector-fill! VECTOR START END FILL -- procedure: append-s8vectors LST -- procedure: s8vector-copy S8VECTOR -- procedure: s8vector-append S8VECTOR... -- procedure: subs8vector S8VECTOR START END -- procedure: subs8vector-move! SRC-S8VECTOR SRC-START SRC-END DST-S8VECTOR DST-START -- procedure: s8vector-shrink! S8VECTOR K -- procedure: u8vector? OBJ -- procedure: make-u8vector K [FILL] -- procedure: u8vector EXACT-INT8... -- procedure: u8vector-length U8VECTOR -- procedure: u8vector-ref U8VECTOR K -- procedure: u8vector-set U8VECTOR K EXACT-INT8 -- procedure: u8vector-set! U8VECTOR K EXACT-INT8 -- procedure: u8vector->list U8VECTOR -- procedure: list->u8vector LIST-OF-EXACT-INT8 -- procedure: u8vector-fill! U8VECTOR FILL [START [END]] -- procedure: subu8vector-fill! VECTOR START END FILL -- procedure: append-u8vectors LST -- procedure: u8vector-copy U8VECTOR -- procedure: u8vector-append U8VECTOR... -- procedure: subu8vector U8VECTOR START END -- procedure: subu8vector-move! SRC-U8VECTOR SRC-START SRC-END DST-U8VECTOR DST-START -- procedure: u8vector-shrink! U8VECTOR K -- procedure: s16vector? OBJ -- procedure: make-s16vector K [FILL] -- procedure: s16vector EXACT-INT16... -- procedure: s16vector-length S16VECTOR -- procedure: s16vector-ref S16VECTOR K -- procedure: s16vector-set S16VECTOR K EXACT-INT16 -- procedure: s16vector-set! S16VECTOR K EXACT-INT16 -- procedure: s16vector->list S16VECTOR -- procedure: list->s16vector LIST-OF-EXACT-INT16 -- procedure: s16vector-fill! S16VECTOR FILL [START [END]] -- procedure: subs16vector-fill! VECTOR START END FILL -- procedure: append-s16vectors LST -- procedure: s16vector-copy S16VECTOR -- procedure: s16vector-append S16VECTOR... -- procedure: subs16vector S16VECTOR START END -- procedure: subs16vector-move! SRC-S16VECTOR SRC-START SRC-END DST-S16VECTOR DST-START -- procedure: s16vector-shrink! S16VECTOR K -- procedure: u16vector? OBJ -- procedure: make-u16vector K [FILL] -- procedure: u16vector EXACT-INT16... -- procedure: u16vector-length U16VECTOR -- procedure: u16vector-ref U16VECTOR K -- procedure: u16vector-set U16VECTOR K EXACT-INT16 -- procedure: u16vector-set! U16VECTOR K EXACT-INT16 -- procedure: u16vector->list U16VECTOR -- procedure: list->u16vector LIST-OF-EXACT-INT16 -- procedure: u16vector-fill! U16VECTOR FILL [START [END]] -- procedure: subu16vector-fill! VECTOR START END FILL -- procedure: append-u16vectors LST -- procedure: u16vector-copy U16VECTOR -- procedure: u16vector-append U16VECTOR... -- procedure: subu16vector U16VECTOR START END -- procedure: subu16vector-move! SRC-U16VECTOR SRC-START SRC-END DST-U16VECTOR DST-START -- procedure: u16vector-shrink! U16VECTOR K -- procedure: s32vector? OBJ -- procedure: make-s32vector K [FILL] -- procedure: s32vector EXACT-INT32... -- procedure: s32vector-length S32VECTOR -- procedure: s32vector-ref S32VECTOR K -- procedure: s32vector-set S32VECTOR K EXACT-INT32 -- procedure: s32vector-set! S32VECTOR K EXACT-INT32 -- procedure: s32vector->list S32VECTOR -- procedure: list->s32vector LIST-OF-EXACT-INT32 -- procedure: s32vector-fill! S32VECTOR FILL [START [END]] -- procedure: subs32vector-fill! VECTOR START END FILL -- procedure: append-s32vectors LST -- procedure: s32vector-copy S32VECTOR -- procedure: s32vector-append S32VECTOR... -- procedure: subs32vector S32VECTOR START END -- procedure: subs32vector-move! SRC-S32VECTOR SRC-START SRC-END DST-S32VECTOR DST-START -- procedure: s32vector-shrink! S32VECTOR K -- procedure: u32vector? OBJ -- procedure: make-u32vector K [FILL] -- procedure: u32vector EXACT-INT32... -- procedure: u32vector-length U32VECTOR -- procedure: u32vector-ref U32VECTOR K -- procedure: u32vector-set U32VECTOR K EXACT-INT32 -- procedure: u32vector-set! U32VECTOR K EXACT-INT32 -- procedure: u32vector->list U32VECTOR -- procedure: list->u32vector LIST-OF-EXACT-INT32 -- procedure: u32vector-fill! U32VECTOR FILL [START [END]] -- procedure: subu32vector-fill! VECTOR START END FILL -- procedure: append-u32vectors LST -- procedure: u32vector-copy U32VECTOR -- procedure: u32vector-append U32VECTOR... -- procedure: subu32vector U32VECTOR START END -- procedure: subu32vector-move! SRC-U32VECTOR SRC-START SRC-END DST-U32VECTOR DST-START -- procedure: u32vector-shrink! U32VECTOR K -- procedure: s64vector? OBJ -- procedure: make-s64vector K [FILL] -- procedure: s64vector EXACT-INT64... -- procedure: s64vector-length S64VECTOR -- procedure: s64vector-ref S64VECTOR K -- procedure: s64vector-set S64VECTOR K EXACT-INT64 -- procedure: s64vector-set! S64VECTOR K EXACT-INT64 -- procedure: s64vector->list S64VECTOR -- procedure: list->s64vector LIST-OF-EXACT-INT64 -- procedure: s64vector-fill! S64VECTOR FILL [START [END]] -- procedure: subs64vector-fill! VECTOR START END FILL -- procedure: append-s64vectors LST -- procedure: s64vector-copy S64VECTOR -- procedure: s64vector-append S64VECTOR... -- procedure: subs64vector S64VECTOR START END -- procedure: subs64vector-move! SRC-S64VECTOR SRC-START SRC-END DST-S64VECTOR DST-START -- procedure: s64vector-shrink! S64VECTOR K -- procedure: u64vector? OBJ -- procedure: make-u64vector K [FILL] -- procedure: u64vector EXACT-INT64... -- procedure: u64vector-length U64VECTOR -- procedure: u64vector-ref U64VECTOR K -- procedure: u64vector-set U64VECTOR K EXACT-INT64 -- procedure: u64vector-set! U64VECTOR K EXACT-INT64 -- procedure: u64vector->list U64VECTOR -- procedure: list->u64vector LIST-OF-EXACT-INT64 -- procedure: u64vector-fill! U64VECTOR FILL [START [END]] -- procedure: subu64vector-fill! VECTOR START END FILL -- procedure: append-u64vectors LST -- procedure: u64vector-copy U64VECTOR -- procedure: u64vector-append U64VECTOR... -- procedure: subu64vector U64VECTOR START END -- procedure: subu64vector-move! SRC-U64VECTOR SRC-START SRC-END DST-U64VECTOR DST-START -- procedure: u64vector-shrink! U64VECTOR K -- procedure: f32vector? OBJ -- procedure: make-f32vector K [FILL] -- procedure: f32vector INEXACT-REAL... -- procedure: f32vector-length F32VECTOR -- procedure: f32vector-ref F32VECTOR K -- procedure: f32vector-set F32VECTOR K INEXACT-REAL -- procedure: f32vector-set! F32VECTOR K INEXACT-REAL -- procedure: f32vector->list F32VECTOR -- procedure: list->f32vector LIST-OF-INEXACT-REAL -- procedure: f32vector-fill! F32VECTOR FILL [START [END]] -- procedure: subf32vector-fill! VECTOR START END FILL -- procedure: append-f32vectors LST -- procedure: f32vector-copy F32VECTOR -- procedure: f32vector-append F32VECTOR... -- procedure: subf32vector F32VECTOR START END -- procedure: subf32vector-move! SRC-F32VECTOR SRC-START SRC-END DST-F32VECTOR DST-START -- procedure: f32vector-shrink! F32VECTOR K -- procedure: f64vector? OBJ -- procedure: make-f64vector K [FILL] -- procedure: f64vector INEXACT-REAL... -- procedure: f64vector-length F64VECTOR -- procedure: f64vector-ref F64VECTOR K -- procedure: f64vector-set F64VECTOR K INEXACT-REAL -- procedure: f64vector-set! F64VECTOR K INEXACT-REAL -- procedure: f64vector->list F64VECTOR -- procedure: list->f64vector LIST-OF-INEXACT-REAL -- procedure: f64vector-fill! F64VECTOR FILL [START [END]] -- procedure: subf64vector-fill! VECTOR START END FILL -- procedure: append-f64vectors LST -- procedure: f64vector-copy F64VECTOR -- procedure: f64vector-append F64VECTOR... -- procedure: subf64vector F64VECTOR START END -- procedure: subf64vector-move! SRC-F64VECTOR SRC-START SRC-END DST-F64VECTOR DST-START -- procedure: f64vector-shrink! F64VECTOR K For example: > (define v (u8vector 10 255 13)) > (u8vector-set! v 2 99) > v #u8(10 255 99) > (u8vector-ref v 1) 255 > (u8vector->list v) (10 255 99) > (u8vector-shrink! v 2) > (v) #u8(10 255) -- procedure: object->u8vector OBJ [ENCODER] -- procedure: u8vector->object U8VECTOR [DECODER] The procedure `object->u8vector' returns a u8vector that contains the sequence of bytes that encodes the object OBJ. The procedure `u8vector->object' decodes the sequence of bytes contained in the u8vector U8VECTOR, which was produced by the procedure `object->u8vector', and reconstructs an object structurally equal to the original object. In other words the procedures `object->u8vector' and `u8vector->object' respectively perform serialization and deserialization of Scheme objects. Note that some objects are non-serializable (e.g. threads, wills, some types of ports, and any object containing a non-serializable object). The optional ENCODER and DECODER parameters are single parameter procedures which default to the identity function. The ENCODER procedure is called during serialization. As the serializer walks through OBJ, it calls the ENCODER procedure on each sub-object X that is encountered. The ENCODER transforms the object X into an object Y that will be serialized instead of X. Similarly the DECODER procedure is called during deserialization. When an object Y is encountered, the DECODER procedure is called to transform it into the object X that is the result of deserialization. The ENCODER and DECODER procedures are useful to customize the serialized representation of objects. In particular, it can be used to define the semantics of serializing objects, such as threads and ports, that would otherwise not be serializable. The DECODER procedure is typically the inverse of the ENCODER procedure, i.e. `(DECODER (ENCODER X))' = `X'. For example: > (define (make-adder x) (lambda (y) (+ x y))) > (define f (make-adder 10)) > (define a (object->u8vector f)) > (define b (u8vector->object a)) > (u8vector-length a) 1639 > (f 5) 15 > (b 5) 15 > (pp b) (lambda (y) (+ x y)) 11 Hashing and weak references ****************************** 11.1 Hashing ============ -- procedure: object->serial-number OBJ -- procedure: serial-number->object N [DEFAULT] All Scheme objects are uniquely identified with a serial number which is a nonnegative exact integer. The `object->serial-number' procedure returns the serial number of object OBJ. This serial number is only allocated the first time the `object->serial-number' procedure is called on that object. Objects which do not have an external textual representation that can be read by the `read' procedure, use an external textual representation that includes a serial number of the form `#N'. Consequently, the procedures `write', `pretty-print', etc will call the `object->serial-number' procedure to get the serial number, and this may cause the serial number to be allocated. The `serial-number->object' procedure takes an exact integer parameter N and returns the object whose serial number is N. If no object currently exists with that serial number, DEFAULT is returned if it is specified, otherwise an unbound-serial-number-exception object is raised. The reader defines the following abbreviation for calling `serial-number->object': the syntax `#N', where N is a sequence of decimal digits and it is not followed by ``='' or ``#'', is equivalent to the list `(serial-number->object N)'. For example: > (define z (list (lambda (x) (* x x)) (lambda (y) (/ 1 y)))) > z (# #) > (#3 10) 1/10 > '(#3 10) ((serial-number->object 3) 10) > car # > (#4 z) # -- procedure: unbound-serial-number-exception? OBJ -- procedure: unbound-serial-number-exception-procedure EXC -- procedure: unbound-serial-number-exception-arguments EXC Unbound-serial-number-exception objects are raised by the procedure `serial-number->object' when no object currently exists with that serial number. The parameter EXC must be an unbound-serial-number-exception object. The procedure `unbound-serial-number-exception?' returns `#t' when OBJ is a unbound-serial-number-exception object and `#f' otherwise. The procedure `unbound-serial-number-exception-procedure' returns the procedure that raised EXC. The procedure `unbound-serial-number-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (unbound-serial-number-exception? exc) (list (unbound-serial-number-exception-procedure exc) (unbound-serial-number-exception-arguments exc)) 'not-unbound-serial-number-exception)) > (with-exception-catcher handler (lambda () (serial-number->object 1000))) (#object> (1000)) -- procedure: symbol-hash SYMBOL The `symbol-hash' procedure returns the hash number of the symbol SYMBOL. The hash number is a small exact integer (fixnum). When SYMBOL is an interned symbol the value returned is the same as `(string=?-hash (symbol->string SYMBOL))'. For example: > (symbol-hash 'car) 444471047 -- procedure: keyword-hash KEYWORD The `keyword-hash' procedure returns the hash number of the keyword KEYWORD. The hash number is a small exact integer (fixnum). When KEYWORD is an interned keyword the value returned is the same as `(string=?-hash (keyword->string KEYWORD))'. For example: > (keyword-hash car:) 444471047 -- procedure: string=?-hash STRING The `string=?-hash' procedure returns the hash number of the string STRING. The hash number is a small exact integer (fixnum). For any two strings S1 and S2, `(string=? S1 S2)' implies `(= (string=?-hash S1) (string=?-hash S2))'. For example: > (string=?-hash "car") 444471047 -- procedure: string-ci=?-hash STRING The `string-ci=?-hash' procedure returns the hash number of the string STRING. The hash number is a small exact integer (fixnum). For any two strings S1 and S2, `(string-ci=? S1 S2)' implies `(= (string-ci=?-hash S1) (string-ci=?-hash S2))'. For example: > (string-ci=?-hash "CaR") 444471047 -- procedure: eq?-hash OBJ The `eq?-hash' procedure returns the hash number of the object OBJ. The hash number is a small exact integer (fixnum). For any two objects O1 and O2, `(eq? O1 O2)' implies `(= (eq?-hash O1) (eq?-hash O2))'. For example: > (eq?-hash #t) 536870910 -- procedure: eqv?-hash OBJ The `eqv?-hash' procedure returns the hash number of the object OBJ. The hash number is a small exact integer (fixnum). For any two objects O1 and O2, `(eqv? O1 O2)' implies `(= (eqv?-hash O1) (eqv?-hash O2))'. For example: > (eqv?-hash 1.5) 496387656 -- procedure: equal?-hash OBJ The `equal?-hash' procedure returns the hash number of the object OBJ. The hash number is a small exact integer (fixnum). For any two objects O1 and O2, `(equal? O1 O2)' implies `(= (equal?-hash O1) (equal?-hash O2))'. For example: > (equal?-hash (list 1 2 3)) 442438567 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. -- procedure: make-will TESTATOR ACTION -- procedure: will? OBJ -- procedure: will-testator WILL -- procedure: will-execute! WILL The `make-will' procedure creates a will object with the given TESTATOR object and ACTION procedure. The `will?' procedure tests if OBJ is a will object. The `will-testator' procedure gets the testator object attached to the WILL. The `will-execute!' procedure executes WILL. A will becomes "executable" when its TESTATOR object is not strongly reachable (i.e. the TESTATOR object is either unreachable or only reachable using paths from the roots that traverse at least one weak reference). Some objects, including symbols, small exact integers (fixnums), booleans and characters, are considered to be always strongly reachable. When the runtime system detects that a will has become executable the current computation is interrupted, the will's testator is set to `#f' and the will's action procedure is called with the will's testator as the sole argument. Currently only the garbage collector detects when wills become executable but this may change in future versions of Gambit (for example the compiler could perform an analysis to infer will executability at compile time). The garbage collector builds a list of all executable wills. Shortly after a garbage collection, the action procedures of these wills will be called. The link from the will to the action procedure is severed when the action procedure is called. Note that the testator object will not be reclaimed during the garbage collection that determined executability of the will. It is only when an object is not reachable from the roots that it is reclaimed by the garbage collector. A remarkable feature of wills is that an action procedure can "resurrect" an object. An action procedure could for example assign the testator object to a global variable or create a new will with the same testator object. For example: > (define a (list 123)) > (set-cdr! a a) ; create a circular list > (define b (vector a)) > (define c #f) > (define w (let ((obj a)) (make-will obj (lambda (x) ; x will be eq? to obj (display "executing action procedure") (newline) (set! c x))))) > (will? w) #t > (car (will-testator w)) 123 > (##gc) > (set! a #f) > (##gc) > (set! b #f) > (##gc) executing action procedure > (will-testator w) #f > (car c) 123 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))'. -- procedure: make-table [`size:' SIZE] [`init:' INIT] [`weak-keys:' WEAK-KEYS] [`weak-values:' WEAK-VALUES] [`test:' TEST] [`hash:' HASH] [`min-load:' MIN-LOAD] [`max-load:' MAX-LOAD] The procedure `make-table' returns a new table. The optional keyword parameters specify various parameters of the table. The SIZE parameter is a nonnegative exact integer indicating the expected length of the table. The system uses SIZE to choose an appropriate initial size of the hash table so that it does not need to be resized too often. The INIT parameter indicates a value that is associated to keys that are not in the table. When INIT is not specified, no value is associated to keys that are not in the table. The WEAK-KEYS and WEAK-VALUES parameters are extended booleans indicating respectively whether the keys and values are table-weak references (true) or strong references (false). By default the keys and values are strong references. The TEST parameter indicates the key comparison procedure. The default key comparison procedure is `equal?'. The key comparison procedures `eq?', `eqv?', `equal?', `string=?', and `string-ci=?' are special because the system will use a reasonably good hash procedure when none is specified. The HASH parameter indicates the hash procedure. This procedure must accept a single key parameter, return a fixnum, and be consistent with the key comparison procedure. When HASH is not specified, a default hash procedure is used. The default hash procedure is reasonably good when the key comparison procedure is `eq?', `eqv?', `equal?', `string=?', or `string-ci=?'. The MIN-LOAD and MAX-LOAD parameters are real numbers that indicate the minimum and maximum load of the table respectively. The table is resized when adding or deleting a key/value binding would bring the table's load outside of this range. The MIN-LOAD parameter must be no less than 0.05 and the MAX-LOAD parameter must be no greater than 0.95. Moreover the difference between MIN-LOAD and MAX-LOAD must be at least 0.20. When MIN-LOAD is not specified, the value 0.45 is used. When MAX-LOAD is not specified, the value 0.90 is used. For example: > (define t (make-table)) > (table? t) #t > (table-length t) 0 > (table-set! t (list 1 2) 3) > (table-set! t (list 4 5) 6) > (table-ref t (list 1 2)) 3 > (table-length t) 2 -- procedure: table? OBJ The procedure `table?' returns `#t' when OBJ is a table and `#f' otherwise. For example: > (table? (make-table)) #t > (table? 123) #f -- procedure: table-length TABLE The procedure `table-length' returns the number of key/value bindings contained in the table TABLE. For example: > (define t (make-table weak-keys: #t)) > (define x (list 1 2)) > (define y (list 3 4)) > (table-set! t x 111) > (table-set! t y 222) > (table-length t) 2 > (table-set! t x) > (table-length t) 1 > (##gc) > (table-length t) 1 > (set! y #f) > (##gc) > (table-length t) 0 -- procedure: table-ref TABLE KEY [DEFAULT] The procedure `table-ref' returns the value bound to the object KEY in the table TABLE. When KEY is not bound and DEFAULT is specified, DEFAULT is returned. When DEFAULT is not specified but an INIT parameter was specified when TABLE was created, INIT is returned. Otherwise an unbound-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 '# '(4 5)) -- procedure: table-set! TABLE KEY [VALUE] The procedure `table-set!' binds the object KEY to VALUE in the table TABLE. When VALUE is not specified, if TABLE contains a binding for KEY then the binding is removed from TABLE. The procedure `table-set!' returns an unspecified value. For example: > (define t (make-table)) > (table-set! t (list 1 2) 3) > (table-set! t (list 4 5) 6) > (table-set! t (list 4 5)) > (table-set! t (list 7 8)) > (table-ref t (list 1 2)) 3 > (table-ref t (list 4 5)) *** ERROR IN (console)@7.1 -- Unbound table key (table-ref '#
'(4 5)) -- procedure: table-search PROC TABLE The procedure `table-search' searches the table TABLE for a key/value binding for which the two parameter procedure PROC returns a non false result. For each key/value binding visited by `table-search' the procedure PROC is called with the key as the first parameter and the value as the second parameter. The procedure `table-search' returns the first non false value returned by PROC, or `#f' if PROC returned `#f' for all key/value bindings in TABLE. The order in which the key/value bindings are visited is unspecified and may vary from one call of `table-search' to the next. While a call to `table-search' is being performed on TABLE, it is an error to call any of the following procedures on TABLE: `table-ref', `table-set!', `table-search', `table-for-each', `table-copy', `table-merge', `table-merge!', and `table->list'. It is also an error to compare with `equal?' (directly or indirectly with `member', `assoc', `table-ref', etc.) an object that contains TABLE. All these procedures may cause TABLE to be reordered and resized. This restriction allows a more efficient iteration over the key/value bindings. For example: > (define square (make-table)) > (table-set! square 2 4) > (table-set! square 3 9) > (table-search (lambda (k v) (and (odd? k) v)) square) 9 -- procedure: table-for-each PROC TABLE The procedure `table-for-each' calls the two parameter procedure PROC for each key/value binding in the table TABLE. The procedure PROC is called with the key as the first parameter and the value as the second parameter. The procedure `table-for-each' returns an unspecified value. The order in which the key/value bindings are visited is unspecified and may vary from one call of `table-for-each' to the next. While a call to `table-for-each' is being performed on TABLE, it is an error to call any of the following procedures on TABLE: `table-ref', `table-set!', `table-search', `table-for-each', and `table->list'. It is also an error to compare with `equal?' (directly or indirectly with `member', `assoc', `table-ref', etc.) an object that contains TABLE. All these procedures may cause TABLE to be reordered and resized. This restriction allows a more efficient iteration over the key/value bindings. For example: > (define square (make-table)) > (table-set! square 2 4) > (table-set! square 3 9) > (table-for-each (lambda (k v) (write (list k v)) (newline)) square) (2 4) (3 9) -- procedure: table->list TABLE The procedure `table->list' returns an association list containing the key/value bindings in the table TABLE. Each key/value binding yields a pair whose car field is the key and whose cdr field is the value bound to that key. The order of the bindings in the list is unspecified. For example: > (define square (make-table)) > (table-set! square 2 4) > (table-set! square 3 9) > (table->list square) ((3 . 9) (2 . 4)) -- procedure: list->table LIST [`size:' SIZE] [`init:' INIT] [`weak-keys:' WEAK-KEYS] [`weak-values:' WEAK-VALUES] [`test:' TEST] [`hash:' HASH] [`min-load:' MIN-LOAD] [`max-load:' MAX-LOAD] The procedure `list->table' returns a new table containing the key/value bindings in the association list LIST. The optional keyword parameters specify various parameters of the table and have the same meaning as for the `make-table' procedure. Each element of LIST is a pair whose car field is a key and whose cdr field is the value bound to that key. If a key appears more than once in LIST (tested using the table's key comparison procedure) it is the first key/value binding in LIST that has precedence. For example: > (define t (list->table '((b . 2) (a . 1) (c . 3) (a . 4)))) > (table->list t) ((a . 1) (b . 2) (c . 3)) -- procedure: unbound-table-key-exception? OBJ -- procedure: unbound-table-key-exception-procedure EXC -- procedure: unbound-table-key-exception-arguments EXC 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)))) (# (#
(1 2))) -- procedure: table-copy TABLE The procedure `table-copy' returns a new table containing the same key/value bindings as TABLE and the same table parameters (i.e. hash procedure, key comparison procedure, key and value weakness, etc). For example: > (define t (list->table '((b . 2) (a . 1) (c . 3)))) > (define x (table-copy t)) > (table-set! t 'b 99) > (table->list t) ((a . 1) (b . 99) (c . 3)) > (table->list x) ((a . 1) (b . 2) (c . 3)) -- procedure: table-merge! TABLE1 TABLE2 [TABLE2-TAKES-PRECEDENCE?] The procedure `table-merge!' returns TABLE1 after the key/value bindings contained in TABLE2 have been added to it. When a key exists both in TABLE1 and TABLE2, then the parameter TABLE2-TAKES-PRECEDENCE? indicates which binding will be kept (the one in TABLE1 if TABLE2-TAKES-PRECEDENCE? is false, and the one in TABLE2 otherwise). If TABLE2-TAKES-PRECEDENCE? is not specified the binding in TABLE1 is kept. For example: > (define t1 (list->table '((a . 1) (b . 2) (c . 3)))) > (define t2 (list->table '((a . 4) (b . 5) (z . 6)))) > (table->list (table-merge! t1 t2)) ((a . 1) (b . 2) (c . 3) (z . 6)) > (define t1 (list->table '((a . 1) (b . 2) (c . 3)))) > (define t2 (list->table '((a . 4) (b . 5) (z . 6)))) > (table->list (table-merge! t1 t2 #t)) ((a . 4) (b . 5) (c . 3) (z . 6)) -- procedure: table-merge TABLE1 TABLE2 [TABLE2-TAKES-PRECEDENCE?] The procedure `table-merge' returns a copy of TABLE1 (created with `table-copy') to which the key/value bindings contained in TABLE2 have been added using `table-merge!'. When a key exists both in TABLE1 and TABLE2, then the parameter TABLE2-TAKES-PRECEDENCE? indicates which binding will be kept (the one in TABLE1 if TABLE2-TAKES-PRECEDENCE? is false, and the one in TABLE2 otherwise). If TABLE2-TAKES-PRECEDENCE? is not specified the binding in TABLE1 is kept. For example: > (define t1 (list->table '((a . 1) (b . 2) (c . 3)))) > (define t2 (list->table '((a . 4) (b . 5) (z . 6)))) > (table->list (table-merge t1 t2)) ((a . 1) (b . 2) (c . 3) (z . 6)) > (table->list (table-merge t1 t2 #t)) ((a . 4) (b . 5) (c . 3) (z . 6)) 12 Records ********** -- special form: define-structure name field... Record data types similar to Pascal records and C `struct' types can be defined using the `define-structure' special form. The identifier name specifies the name of the new data type. The structure name is followed by K identifiers naming each field of the record. The `define-structure' expands into a set of definitions of the following procedures: * `make-name' - A K argument procedure which constructs a new record from the value of its K fields. * `name?' - A procedure which tests if its single argument is of the given record type. * `name-field' - For each field, a procedure taking as its single argument a value of the given record type and returning the content of the corresponding field of the record. * `name-field-set!' - For each field, a two argument procedure taking as its first argument a value of the given record type. The second argument gets assigned to the corresponding field of the record and the void object is returned. Record data types have a printed representation that includes the name of the type and the name and value of each field. Record data types can not be read by the `read' procedure. For example: > (define-structure point x y color) > (define p (make-point 3 5 'red)) > p # > (point-x p) 3 > (point-color p) red > (point-color-set! p 'black) > p # 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)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 =============== -- procedure: current-thread This procedure returns the current thread. For example: > (current-thread) # > (eq? (current-thread) (current-thread)) #t -- procedure: thread? OBJ This procedure returns `#t' when OBJ is a thread object and `#f' otherwise. For example: > (thread? (current-thread)) #t > (thread? 'foo) #f -- procedure: make-thread THUNK [NAME [THREAD-GROUP]] -- procedure: make-root-thread THUNK [NAME [THREAD-GROUP [INPUT-PORT [OUTPUT-PORT]]]] The `make-thread' procedure creates and returns an initialized thread. This thread is not automatically made runnable (the procedure `thread-start!' must be used for this). A thread has the following fields: base priority, priority boost, boosted flag, quantum, name, specific, end-result, end-exception, and a list of locked/owned mutexes it owns. The thread's execution consists of a call to THUNK with the "initial continuation". This continuation causes the (then) current thread to store the result in its end-result field, abandon all mutexes it owns, and finally terminate. The `dynamic-wind' stack of the initial continuation is empty. The optional NAME is an arbitrary Scheme object which identifies the thread (useful for debugging); it defaults to an unspecified value. The specific field is set to an unspecified value. The optional THREAD-GROUP indicates which thread group this thread belongs to; it defaults to the thread group of the current thread. The base priority, priority boost, and quantum of the thread are set to the same value as the current thread and the boosted flag is set to false. The thread's mailbox is initially empty. The thread inherits the dynamic environment from the current thread. Moreover, in this dynamic environment the exception-handler is bound to the "initial exception-handler" which is a unary procedure which causes the (then) current thread to store in its end-exception field an uncaught-exception object whose "reason" is the argument of the handler, abandon all mutexes it owns, and finally terminate. The `make-root-thread' procedure behaves like the `make-thread' procedure except the created thread does not inherit the dynamic environment from the current thread and the base priority is set to 0, the priority boost is set to 1.0e-6, and the quantum is set to 0.02. The dynamic environment of the thread has the global bindings of the parameter objects, except `current-input-port' which is bound to INPUT-PORT, `current-output-port' which is bound to OUTPUT-PORT, and `current-directory' which is bound to the initial current working directory of the current process. If INPUT-PORT is not specified it defaults to a port corresponding to the standard input (`stdin'). If OUTPUT-PORT is not specified it defaults to a port corresponding to the standard output (`stdout'). For example: > (make-thread (lambda () (write 'hello))) # > (make-root-thread (lambda () (write 'world)) 'a-name) # -- procedure: thread-name THREAD This procedure returns the name of the THREAD. For example: > (thread-name (make-thread (lambda () #f) 'foo)) foo -- procedure: thread-specific THREAD -- procedure: thread-specific-set! THREAD OBJ The `thread-specific' procedure returns the content of the THREAD's specific field. The `thread-specific-set!' procedure stores OBJ into the THREAD's specific field and returns an unspecified value. For example: > (thread-specific-set! (current-thread) "hello") > (thread-specific (current-thread)) "hello" -- procedure: thread-base-priority THREAD -- procedure: thread-base-priority-set! THREAD PRIORITY The procedure `thread-base-priority' returns a real number which corresponds to the base priority of the THREAD. The procedure `thread-base-priority-set!' changes the base priority of the THREAD to PRIORITY and returns an unspecified value. The PRIORITY must be a real number. For example: > (thread-base-priority-set! (current-thread) 12.3) > (thread-base-priority (current-thread)) 12.3 -- procedure: thread-priority-boost THREAD -- procedure: thread-priority-boost-set! THREAD PRIORITY-BOOST The procedure `thread-priority-boost' returns a real number which corresponds to the priority boost of the THREAD. The procedure `thread-priority-boost-set!' changes the priority boost of the THREAD to PRIORITY-BOOST and returns an unspecified value. The PRIORITY-BOOST must be a nonnegative real. For example: > (thread-priority-boost-set! (current-thread) 2.5) > (thread-priority-boost (current-thread)) 2.5 -- procedure: thread-quantum THREAD -- procedure: thread-quantum-set! THREAD QUANTUM The procedure `thread-quantum' returns a real number which corresponds to the quantum of the THREAD. The procedure `thread-quantum-set!' changes the quantum of the THREAD to QUANTUM and returns an unspecified value. The QUANTUM must be a nonnegative real. A value of zero selects the smallest quantum supported by the implementation. For example: > (thread-quantum-set! (current-thread) 1.5) > (thread-quantum (current-thread)) 1.5 > (thread-quantum-set! (current-thread) 0) > (thread-quantum (current-thread)) 0. -- procedure: thread-start! THREAD This procedure makes THREAD runnable and returns the THREAD. The THREAD must be an initialized thread. For example: > (let ((t (thread-start! (make-thread (lambda () (write 'a)))))) (write 'b) (thread-join! t)) ab> or ba> NOTE: It is useful to separate thread creation and thread activation to avoid the race condition that would occur if the created thread tries to examine a table in which the current thread stores the created thread. See the last example of the `thread-terminate!' procedure which contains mutually recursive threads. -- procedure: thread-yield! This procedure causes the current thread to exit the running state as if its quantum had expired and returns an unspecified value. For example: ; a busy loop that avoids being too wasteful of the CPU (let loop () (if (mutex-lock! m 0) ; try to lock m but don't block (begin (display "locked mutex m") (mutex-unlock! m)) (begin (do-something-else) (thread-yield!) ; relinquish rest of quantum (loop)))) -- procedure: thread-sleep! TIMEOUT This procedure causes the current thread to wait until the timeout is reached and returns an unspecified value. This blocks the thread only if TIMEOUT represents a point in the future. It is an error for TIMEOUT to be `#f'. For example: ; a clock with a gradual drift: (let loop ((x 1)) (thread-sleep! 1) (write x) (loop (+ x 1))) ; a clock with no drift: (let ((start (time->seconds (current-time))) (let loop ((x 1)) (thread-sleep! (seconds->time (+ x start))) (write x) (loop (+ x 1)))) -- procedure: thread-terminate! THREAD This procedure causes an abnormal termination of the THREAD. If the THREAD is not already terminated, all mutexes owned by the THREAD become unlocked/abandoned and a terminated-thread-exception object is stored in the THREAD's end-exception field. If THREAD is the current thread, `thread-terminate!' does not return. Otherwise `thread-terminate!' returns an unspecified value; the termination of the THREAD will occur at some point between the calling of `thread-terminate!' and a finite time in the future (an explicit thread synchronization is needed to detect termination, see `thread-join!'). For example: (define (amb thunk1 thunk2) (let ((result #f) (result-mutex (make-mutex)) (done-mutex (make-mutex))) (letrec ((child1 (make-thread (lambda () (let ((x (thunk1))) (mutex-lock! result-mutex #f #f) (set! result x) (thread-terminate! child2) (mutex-unlock! done-mutex))))) (child2 (make-thread (lambda () (let ((x (thunk2))) (mutex-lock! result-mutex #f #f) (set! result x) (thread-terminate! child1) (mutex-unlock! done-mutex)))))) (mutex-lock! done-mutex #f #f) (thread-start! child1) (thread-start! child2) (mutex-lock! done-mutex #f #f) result))) NOTE: This operation must be used carefully because it terminates a thread abruptly and it is impossible for that thread to perform any kind of cleanup. This may be a problem if the thread is in the middle of a critical section where some structure has been put in an inconsistent state. However, another thread attempting to enter this critical section will raise an abandoned-mutex-exception object because the mutex is unlocked/abandoned. This helps avoid observing an inconsistent state. Clean termination can be obtained by polling, as shown in the example below. For example: (define (spawn thunk) (let ((t (make-thread thunk))) (thread-specific-set! t #t) (thread-start! t) t)) (define (stop! thread) (thread-specific-set! thread #f) (thread-join! thread)) (define (keep-going?) (thread-specific (current-thread))) (define count! (let ((m (make-mutex)) (i 0)) (lambda () (mutex-lock! m) (let ((x (+ i 1))) (set! i x) (mutex-unlock! m) x)))) (define (increment-forever!) (let loop () (count!) (if (keep-going?) (loop)))) (let ((t1 (spawn increment-forever!)) (t2 (spawn increment-forever!))) (thread-sleep! 1) (stop! t1) (stop! t2) (count!)) ==> 377290 -- procedure: thread-join! thread [TIMEOUT [TIMEOUT-VAL]] This procedure causes the current thread to wait until the THREAD terminates (normally or not) or until the timeout is reached if TIMEOUT is supplied. If the timeout is reached, THREAD-JOIN! returns TIMEOUT-VAL if it is supplied, otherwise a join-timeout-exception object is raised. If the THREAD terminated normally, the content of the end-result field is returned, otherwise the content of the end-exception field is raised. For example: (let ((t (thread-start! (make-thread (lambda () (expt 2 100)))))) (do-something-else) (thread-join! t)) ==> 1267650600228229401496703205376 (let ((t (thread-start! (make-thread (lambda () (raise 123)))))) (do-something-else) (with-exception-handler (lambda (exc) (if (uncaught-exception? exc) (* 10 (uncaught-exception-reason exc)) 99999)) (lambda () (+ 1 (thread-join! t))))) ==> 1231 (define thread-alive? (let ((unique (list 'unique))) (lambda (thread) ; Note: this procedure raises an exception if ; the thread terminated abnormally. (eq? (thread-join! thread 0 unique) unique)))) (define (wait-for-termination! thread) (let ((eh (current-exception-handler))) (with-exception-handler (lambda (exc) (if (not (or (terminated-thread-exception? exc) (uncaught-exception? exc))) (eh exc))) ; unexpected exceptions are handled by eh (lambda () ; The following call to thread-join! will wait until the ; thread terminates. If the thread terminated normally ; thread-join! will return normally. If the thread ; terminated abnormally then one of these two exception ; objects is raised by thread-join!: ; - terminated-thread-exception object ; - uncaught-exception object (thread-join! thread) #f)))) ; ignore result of thread-join! -- procedure: thread-send THREAD MSG Each thread has a mailbox which stores messages delivered to the thread in the order delivered. The procedure `thread-send' adds the message MSG at the end of the mailbox of thread THREAD and returns an unspecified value. For example: > (thread-send (current-thread) 111) > (thread-send (current-thread) 222) > (thread-receive) 111 > (thread-receive) 222 -- procedure: thread-receive [TIMEOUT [DEFAULT]] -- procedure: thread-mailbox-next [TIMEOUT [DEFAULT]] -- procedure: thread-mailbox-rewind -- procedure: thread-mailbox-extract-and-rewind To allow a thread to examine the messages in its mailbox without removing them from the mailbox, each thread has a "mailbox cursor" which normally points to the last message accessed in the mailbox. When a mailbox cursor is rewound using the procedure `thread-mailbox-rewind' or `thread-mailbox-extract-and-rewind' or `thread-receive', the cursor does not point to a message, but the next call to `thread-receive' and `thread-mailbox-next' will set the cursor to the oldest message in the mailbox. The procedure `thread-receive' advances the mailbox cursor of the current thread to the next message, removes that message from the mailbox, rewinds the mailbox cursor, and returns the message. When TIMEOUT is not specified, the current thread will wait until a message is available in the mailbox. When TIMEOUT is specified and DEFAULT is not specified, a mailbox-receive-timeout-exception object is raised if the timeout is reached before a message is available. When TIMEOUT is specified and DEFAULT is specified, DEFAULT is returned if the timeout is reached before a message is available. The procedure `thread-mailbox-next' behaves like `thread-receive' except that the message remains in the mailbox and the mailbox cursor is not rewound. The procedures `thread-mailbox-rewind' or `thread-mailbox-extract-and-rewind' rewind the mailbox cursor of the current thread so that the next call to `thread-mailbox-next' and `thread-receive' will access the oldest message in the mailbox. Additionally the procedure `thread-mailbox-extract-and-rewind' will remove from the mailbox the message most recently accessed by a call to `thread-mailbox-next'. When `thread-mailbox-next' has not been called since the last call to `thread-receive' or `thread-mailbox-rewind' or `thread-mailbox-extract-and-rewind', a call to `thread-mailbox-extract-and-rewind' only resets the mailbox cursor (no message is removed). For example: > (thread-send (current-thread) 111) > (thread-receive 1 999) 111 > (thread-send (current-thread) 222) > (thread-send (current-thread) 333) > (thread-mailbox-next 1 999) 222 > (thread-mailbox-next 1 999) 333 > (thread-mailbox-next 1 999) 999 > (thread-mailbox-extract-and-rewind) > (thread-receive 1 999) 222 > (thread-receive 1 999) 999 -- procedure: mailbox-receive-timeout-exception? OBJ -- procedure: mailbox-receive-timeout-exception-procedure EXC -- procedure: mailbox-receive-timeout-exception-arguments EXC Mailbox-receive-timeout-exception objects are raised by the procedures `thread-receive' and `thread-mailbox-next' when a timeout expires before a message is available and no default value is specified. The parameter EXC must be a mailbox-receive-timeout-exception object. The procedure `mailbox-receive-timeout-exception?' returns `#t' when OBJ is a mailbox-receive-timeout-exception object and `#f' otherwise. The procedure `mailbox-receive-timeout-exception-procedure' returns the procedure that raised EXC. The procedure `mailbox-receive-timeout-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (mailbox-receive-timeout-exception? exc) (list (mailbox-receive-timeout-exception-procedure exc) (mailbox-receive-timeout-exception-arguments exc)) 'not-mailbox-receive-timeout-exception)) > (with-exception-catcher handler (lambda () (thread-receive 1))) (# (1)) -- procedure: mutex? OBJ This procedure returns `#t' when OBJ is a mutex object and `#f' otherwise. For example: > (mutex? (make-mutex)) #t > (mutex? 'foo) #f -- procedure: make-mutex [NAME] This procedure returns a new mutex in the unlocked/not-abandoned state. The optional NAME is an arbitrary Scheme object which identifies the mutex (useful for debugging); it defaults to an unspecified value. The mutex's specific field is set to an unspecified value. For example: > (make-mutex) # > (make-mutex 'foo) # -- procedure: mutex-name MUTEX Returns the name of the MUTEX. For example: > (mutex-name (make-mutex 'foo)) foo -- procedure: mutex-specific MUTEX -- procedure: mutex-specific-set! MUTEX OBJ The `mutex-specific' procedure returns the content of the MUTEX's specific field. The `mutex-specific-set!' procedure stores OBJ into the MUTEX's specific field and returns an unspecified value. For example: > (define m (make-mutex)) > (mutex-specific-set! m "hello") > (mutex-specific m) "hello" > (define (mutex-lock-recursively! mutex) (if (eq? (mutex-state mutex) (current-thread)) (let ((n (mutex-specific mutex))) (mutex-specific-set! mutex (+ n 1))) (begin (mutex-lock! mutex) (mutex-specific-set! mutex 0)))) > (define (mutex-unlock-recursively! mutex) (let ((n (mutex-specific mutex))) (if (= n 0) (mutex-unlock! mutex) (mutex-specific-set! mutex (- n 1))))) > (mutex-lock-recursively! m) > (mutex-lock-recursively! m) > (mutex-lock-recursively! m) > (mutex-specific m) 2 -- procedure: mutex-state MUTEX Thos procedure returns information about the state of the MUTEX. The possible results are: * thread T: the MUTEX is in the locked/owned state and thread T is the owner of the MUTEX * symbol `not-owned': the MUTEX is in the locked/not-owned state * symbol `abandoned': the MUTEX is in the unlocked/abandoned state * symbol `not-abandoned': the MUTEX is in the unlocked/not-abandoned state For example: (mutex-state (make-mutex)) ==> not-abandoned (define (thread-alive? thread) (let ((mutex (make-mutex))) (mutex-lock! mutex #f thread) (let ((state (mutex-state mutex))) (mutex-unlock! mutex) ; avoid space leak (eq? state thread)))) -- procedure: mutex-lock! MUTEX [TIMEOUT [THREAD]] This procedure locks MUTEX. If the MUTEX is currently locked, the current thread waits until the MUTEX is unlocked, or until the timeout is reached if TIMEOUT is supplied. If the timeout is reached, `mutex-lock!' returns `#f'. Otherwise, the state of the MUTEX is changed as follows: * if THREAD is `#f' the MUTEX becomes locked/not-owned, * otherwise, let T be THREAD (or the current thread if THREAD is not supplied), * if T is terminated the MUTEX becomes unlocked/abandoned, * otherwise MUTEX becomes locked/owned with T as the owner. After changing the state of the MUTEX, an abandoned-mutex-exception object is raised if the MUTEX was unlocked/abandoned before the state change, otherwise `mutex-lock!' returns `#t'. It is not an error if the MUTEX is owned by the current thread (but the current thread will have to wait). For example: ; an implementation of a mailbox object of depth one; this ; implementation does not behave well in the presence of forced ; thread terminations using thread-terminate! (deadlock can occur ; if a thread is terminated in the middle of a put! or get! operation) (define (make-empty-mailbox) (let ((put-mutex (make-mutex)) ; allow put! operation (get-mutex (make-mutex)) (cell #f)) (define (put! obj) (mutex-lock! put-mutex #f #f) ; prevent put! operation (set! cell obj) (mutex-unlock! get-mutex)) ; allow get! operation (define (get!) (mutex-lock! get-mutex #f #f) ; wait until object in mailbox (let ((result cell)) (set! cell #f) ; prevent space leaks (mutex-unlock! put-mutex) ; allow put! operation result)) (mutex-lock! get-mutex #f #f) ; prevent get! operation (lambda (msg) (case msg ((put!) put!) ((get!) get!) (else (error "unknown message")))))) (define (mailbox-put! m obj) ((m 'put!) obj)) (define (mailbox-get! m) ((m 'get!))) ; an alternate implementation of thread-sleep! (define (sleep! timeout) (let ((m (make-mutex))) (mutex-lock! m #f #f) (mutex-lock! m timeout #f))) ; a procedure that waits for one of two mutexes to unlock (define (lock-one-of! mutex1 mutex2) ; this procedure assumes that neither mutex1 or mutex2 ; are owned by the current thread (let ((ct (current-thread)) (done-mutex (make-mutex))) (mutex-lock! done-mutex #f #f) (let ((t1 (thread-start! (make-thread (lambda () (mutex-lock! mutex1 #f ct) (mutex-unlock! done-mutex))))) (t2 (thread-start! (make-thread (lambda () (mutex-lock! mutex2 #f ct) (mutex-unlock! done-mutex)))))) (mutex-lock! done-mutex #f #f) (thread-terminate! t1) (thread-terminate! t2) (if (eq? (mutex-state mutex1) ct) (begin (if (eq? (mutex-state mutex2) ct) (mutex-unlock! mutex2)) ; don't lock both mutex1) mutex2)))) -- procedure: mutex-unlock! MUTEX [CONDITION-VARIABLE [TIMEOUT]] This procedure unlocks the MUTEX by making it unlocked/not-abandoned. It is not an error to unlock an unlocked mutex and a mutex that is owned by any thread. If CONDITION-VARIABLE is supplied, the current thread is blocked and added to the CONDITION-VARIABLE before unlocking MUTEX; the thread can unblock at any time but no later than when an appropriate call to `condition-variable-signal!' or `condition-variable-broadcast!' is performed (see below), and no later than the timeout (if TIMEOUT is supplied). If there are threads waiting to lock this MUTEX, the scheduler selects a thread, the mutex becomes locked/owned or locked/not-owned, and the thread is unblocked. `mutex-unlock!' returns `#f' when the timeout is reached, otherwise it returns `#t'. NOTE: The reason the thread can unblock at any time (when CONDITION-VARIABLE is supplied) is that the scheduler, when it detects a serious problem such as a deadlock, must interrupt one of the blocked threads (such as the primordial thread) so that it can perform some appropriate action. After a thread blocked on a condition-variable has handled such an interrupt it would be wrong for the scheduler to return the thread to the blocked state, because any calls to `condition-variable-broadcast!' during the interrupt will have gone unnoticed. It is necessary for the thread to remain runnable and return from the call to `mutex-unlock!' with a result of `#t'. NOTE: `mutex-unlock!' is related to the "wait" operation on condition variables available in other thread systems. The main difference is that "wait" automatically locks MUTEX just after the thread is unblocked. This operation is not performed by `mutex-unlock!' and so must be done by an explicit call to `mutex-lock!'. This has the advantages that a different timeout and exception-handler can be specified on the `mutex-lock!' and `mutex-unlock!' and the location of all the mutex operations is clearly apparent. For example: (let loop () (mutex-lock! m) (if (condition-is-true?) (begin (do-something-when-condition-is-true) (mutex-unlock! m)) (begin (mutex-unlock! m cv) (loop)))) -- procedure: condition-variable? OBJ This procedure returns `#t' when OBJ is a condition-variable object and `#f' otherwise. For example: > (condition-variable? (make-condition-variable)) #t > (condition-variable? 'foo) #f -- procedure: make-condition-variable [NAME] This procedure returns a new empty condition variable. The optional NAME is an arbitrary Scheme object which identifies the condition variable (useful for debugging); it defaults to an unspecified value. The condition variable's specific field is set to an unspecified value. For example: > (make-condition-variable) # -- procedure: condition-variable-name CONDITION-VARIABLE This procedure returns the name of the CONDITION-VARIABLE. For example: > (condition-variable-name (make-condition-variable 'foo)) foo -- procedure: condition-variable-specific CONDITION-VARIABLE -- procedure: condition-variable-specific-set! CONDITION-VARIABLE OBJ The `condition-variable-specific' procedure returns the content of the CONDITION-VARIABLE's specific field. The `condition-variable-specific-set!' procedure stores OBJ into the CONDITION-VARIABLE's specific field and returns an unspecified value. For example: > (define cv (make-condition-variable)) > (condition-variable-specific-set! cv "hello") > (condition-variable-specific cv) "hello" -- procedure: condition-variable-signal! CONDITION-VARIABLE This procedure unblocks a thread blocked on the CONDITION-VARIABLE (if there is at least one) and returns an unspecified value. For example: ; an implementation of a mailbox object of depth one; this ; implementation behaves gracefully when threads are forcibly ; terminated using thread-terminate! (an abandoned-mutex-exception ; object will be raised when a put! or get! operation is attempted ; after a thread is terminated in the middle of a put! or get! ; operation) (define (make-empty-mailbox) (let ((mutex (make-mutex)) (put-condvar (make-condition-variable)) (get-condvar (make-condition-variable)) (full? #f) (cell #f)) (define (put! obj) (mutex-lock! mutex) (if full? (begin (mutex-unlock! mutex put-condvar) (put! obj)) (begin (set! cell obj) (set! full? #t) (condition-variable-signal! get-condvar) (mutex-unlock! mutex)))) (define (get!) (mutex-lock! mutex) (if (not full?) (begin (mutex-unlock! mutex get-condvar) (get!)) (let ((result cell)) (set! cell #f) ; avoid space leaks (set! full? #f) (condition-variable-signal! put-condvar) (mutex-unlock! mutex) result))) (lambda (msg) (case msg ((put!) put!) ((get!) get!) (else (error "unknown message")))))) (define (mailbox-put! m obj) ((m 'put!) obj)) (define (mailbox-get! m) ((m 'get!))) -- procedure: condition-variable-broadcast! CONDITION-VARIABLE This procedure unblocks all the thread blocked on the CONDITION-VARIABLE and returns an unspecified value. For example: (define (make-semaphore n) (vector n (make-mutex) (make-condition-variable))) (define (semaphore-wait! sema) (mutex-lock! (vector-ref sema 1)) (let ((n (vector-ref sema 0))) (if (> n 0) (begin (vector-set! sema 0 (- n 1)) (mutex-unlock! (vector-ref sema 1))) (begin (mutex-unlock! (vector-ref sema 1) (vector-ref sema 2)) (semaphore-wait! sema)))) (define (semaphore-signal-by! sema increment) (mutex-lock! (vector-ref sema 1)) (let ((n (+ (vector-ref sema 0) increment))) (vector-set! sema 0 n) (if (> n 0) (condition-variable-broadcast! (vector-ref sema 2))) (mutex-unlock! (vector-ref sema 1)))) 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". -- procedure: make-parameter OBJ [FILTER] The dynamic environment is composed of two parts: the "local dynamic environment" and the "global dynamic environment". There is a single global dynamic environment, and it is used to lookup parameter objects that can't be found in the local dynamic environment. The `make-parameter' procedure returns a new "parameter object". The 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 -- special form: parameterize ((procedure value)...) body The `parameterize' form, evaluates all procedure and value expressions in an unspecified order. All the procedure expressions must evaluate to procedures, either parameter objects or procedures accepting zero and one argument. Then, for each procedure p and in an unspecified order: * If p is a parameter object a new cell is created, initialized, and bound to the parameter object in the local dynamic environment. The value contained in the cell is the result of applying the parameter object's 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 (*note Runtime options::). Predefined procedures normally raise exceptions by performing a tail-call to the exception handler (the exceptions are "complex" procedures such as `eval', `compile-file', `read', `write', etc). This means that the continuation of the exception handler and of the REPL that may be started due to this is normally the continuation of the predefined procedure that raised the exception. By exiting the REPL with the `,(c EXPRESSION)' command it is thus possible to resume the program as though the call to the predefined procedure returned the value of EXPRESSION. For example: > (define (f x) (+ (car x) 1)) > (f 2) ; typo... we meant to say (f '(2)) *** ERROR IN f, (console)@1.18 -- (Argument 1) PAIR expected (car 2) 1> ,(c 2) 3 -- procedure: current-exception-handler [NEW-EXCEPTION-HANDLER] The parameter object `current-exception-handler' is bound to the current exception-handler. Calling this procedure with no argument returns the current exception-handler and calling this procedure with one argument sets the current exception-handler to NEW-EXCEPTION-HANDLER. For example: > (current-exception-handler) # > (current-exception-handler (lambda (exc) (pp exc) 999)) > (/ 1 0) # 999 -- procedure: with-exception-handler HANDLER THUNK Returns the result(s) of calling THUNK with no arguments. The HANDLER, which must be a procedure, is installed as the current exception-handler in the dynamic environment in effect during the call to THUNK. Note that the dynamic environment in effect during the call to HANDLER has HANDLER as the exception-handler. Consequently, an exception raised during the call to HANDLER may lead to an infinite loop. For example: > (with-exception-handler (lambda (e) (write e) 5) (lambda () (+ 1 (* 2 3) 4))) 11 > (with-exception-handler (lambda (e) (write e) 5) (lambda () (+ 1 (* 'foo 3) 4))) #10 > (with-exception-handler (lambda (e) (write e 9)) (lambda () (+ 1 (* 'foo 3) 4))) INFINITE LOOP -- procedure: with-exception-catcher HANDLER THUNK Returns the result(s) of calling THUNK with no arguments. A new exception-handler is installed as the current exception-handler in the dynamic environment in effect during the call to THUNK. This new exception-handler will call the HANDLER, which must be a procedure, with the exception object as an argument and with the same continuation as the call to `with-exception-catcher'. This implies that the dynamic environment in effect during the call to HANDLER is the same as the one in effect at the call to `with-exception-catcher'. Consequently, an exception raised during the call to HANDLER will not lead to an infinite loop. For example: > (with-exception-catcher (lambda (e) (write e) 5) (lambda () (+ 1 (* 2 3) 4))) 11 > (with-exception-catcher (lambda (e) (write e) 5) (lambda () (+ 1 (* 'foo 3) 4))) #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 '# 9) -- procedure: raise OBJ This procedure tail-calls the current exception-handler with OBJ as the sole argument. If the exception-handler returns, the continuation of the call to `raise' is invoked. For example: > (with-exception-handler (lambda (exc) (pp exc) 100) (lambda () (+ 1 (raise "hello")))) "hello" 101 -- procedure: abort OBJ -- procedure: noncontinuable-exception? OBJ -- procedure: noncontinuable-exception-reason EXC The procedure `abort' calls the current exception-handler with OBJ as the sole argument. If the exception-handler returns, the procedure `abort' will be tail-called with a noncontinuable-exception object, whose reason field is OBJ, as sole argument. Noncontinuable-exception objects are raised by the `abort' procedure when the exception-handler returns. The parameter EXC must be a noncontinuable-exception object. The procedure `noncontinuable-exception?' returns `#t' when OBJ is a noncontinuable-exception object and `#f' otherwise. The procedure `noncontinuable-exception-reason' returns the argument of the call to `abort' that raised EXC. For example: > (call-with-current-continuation (lambda (k) (with-exception-handler (lambda (exc) (pp exc) (if (noncontinuable-exception? exc) (k (list (noncontinuable-exception-reason exc))) 100)) (lambda () (+ 1 (abort "hello")))))) "hello" # ("hello") 15.2 Exception objects related to memory management =================================================== -- procedure: heap-overflow-exception? OBJ Heap-overflow-exception objects are raised when the allocation of an object would cause the heap to use more memory space than is available. The procedure `heap-overflow-exception?' returns `#t' when OBJ is a heap-overflow-exception object and `#f' otherwise. For example: > (define (handler exc) (if (heap-overflow-exception? exc) exc 'not-heap-overflow-exception)) > (with-exception-catcher handler (lambda () (define (f x) (f (cons 1 x))) (f '()))) # -- procedure: stack-overflow-exception? OBJ Stack-overflow-exception objects are raised when the allocation of a continuation frame would cause the heap to use more memory space than is available. The procedure `stack-overflow-exception?' returns `#t' when OBJ is a stack-overflow-exception object and `#f' otherwise. For example: > (define (handler exc) (if (stack-overflow-exception? exc) exc 'not-stack-overflow-exception)) > (with-exception-catcher handler (lambda () (define (f) (+ 1 (f))) (f))) # 15.3 Exception objects related to the host environment ====================================================== -- procedure: os-exception? OBJ -- procedure: os-exception-procedure EXC -- procedure: os-exception-arguments EXC -- procedure: os-exception-code EXC -- procedure: os-exception-message EXC Os-exception objects are raised by procedures which access the host operating-system's services when the requested operation fails. The parameter EXC must be a os-exception object. The procedure `os-exception?' returns `#t' when OBJ is a os-exception object and `#f' otherwise. The procedure `os-exception-procedure' returns the procedure that raised EXC. The procedure `os-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `os-exception-code' returns an exact integer error code that can be converted to a string by the `err-code->string' procedure. Note that the error code is operating-system dependent. The procedure `os-exception-message' returns `#f' or a string giving details of the exception in a human-readable form. For example: > (define (handler exc) (if (os-exception? exc) (list (os-exception-procedure exc) (os-exception-arguments exc) (err-code->string (os-exception-code exc)) (os-exception-message exc)) 'not-os-exception)) > (with-exception-catcher handler (lambda () (host-info "x.y.z"))) (# ("x.y.z") "Unknown host" #f) -- procedure: no-such-file-or-directory-exception? OBJ -- procedure: no-such-file-or-directory-exception-procedure EXC -- procedure: no-such-file-or-directory-exception-arguments EXC No-such-file-or-directory-exception objects are raised by procedures which access the filesystem (such as `open-input-file' and `directory-files') when the path specified can't be found on the filesystem. The parameter EXC must be a no-such-file-or-directory-exception object. The procedure `no-such-file-or-directory-exception?' returns `#t' when OBJ is a no-such-file-or-directory-exception object and `#f' otherwise. The procedure `no-such-file-or-directory-exception-procedure' returns the procedure that raised EXC. The procedure `no-such-file-or-directory-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (no-such-file-or-directory-exception? exc) (list (no-such-file-or-directory-exception-procedure exc) (no-such-file-or-directory-exception-arguments exc)) 'not-no-such-file-or-directory-exception)) > (with-exception-catcher handler (lambda () (with-input-from-file "nofile" read))) (# ("nofile" #)) -- procedure: unbound-os-environment-variable-exception? OBJ -- procedure: unbound-os-environment-variable-exception-procedure EXC -- procedure: unbound-os-environment-variable-exception-arguments EXC Unbound-os-environment-variable-exception objects are raised when an unbound operating-system environment variable is accessed by the procedures `getenv' and `setenv'. The parameter EXC must be an unbound-os-environment-variable-exception object. The procedure `unbound-os-environment-variable-exception?' returns `#t' when OBJ is an unbound-os-environment-variable-exception object and `#f' otherwise. The procedure `unbound-os-environment-variable-exception-procedure' returns the procedure that raised EXC. The procedure `unbound-os-environment-variable-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (unbound-os-environment-variable-exception? exc) (list (unbound-os-environment-variable-exception-procedure exc) (unbound-os-environment-variable-exception-arguments exc)) 'not-unbound-os-environment-variable-exception)) > (with-exception-catcher handler (lambda () (getenv "DOES_NOT_EXIST"))) (# ("DOES_NOT_EXIST")) 15.4 Exception objects related to threads ========================================= -- procedure: scheduler-exception? OBJ -- procedure: scheduler-exception-reason EXC Scheduler-exception objects are raised by the scheduler when some operation requested from the host operating system failed (e.g. checking the status of the devices in order to wake up threads waiting to perform I/O on these devices). The parameter EXC must be a scheduler-exception object. The procedure `scheduler-exception?' returns `#t' when OBJ is a scheduler-exception object and `#f' otherwise. The procedure `scheduler-exception-reason' returns the os-exception object that describes the failure detected by the scheduler. -- procedure: deadlock-exception? OBJ Deadlock-exception objects are raised when the scheduler discovers that all threads are blocked and can make no further progress. In that case the scheduler unblocks the primordial-thread and forces it to raise a deadlock-exception object. The procedure `deadlock-exception?' returns `#t' when OBJ is a deadlock-exception object and `#f' otherwise. For example: > (define (handler exc) (if (deadlock-exception? exc) exc 'not-deadlock-exception)) > (with-exception-catcher handler (lambda () (read (open-vector)))) # -- procedure: abandoned-mutex-exception? OBJ Abandoned-mutex-exception objects are raised when the current thread locks a mutex that was owned by a thread which terminated (see `mutex-lock!'). The procedure `abandoned-mutex-exception?' returns `#t' when OBJ is a abandoned-mutex-exception object and `#f' otherwise. For example: > (define (handler exc) (if (abandoned-mutex-exception? exc) exc 'not-abandoned-mutex-exception)) > (with-exception-catcher handler (lambda () (let ((m (make-mutex))) (thread-join! (thread-start! (make-thread (lambda () (mutex-lock! m))))) (mutex-lock! m)))) # -- procedure: join-timeout-exception? OBJ -- procedure: join-timeout-exception-procedure EXC -- procedure: join-timeout-exception-arguments EXC Join-timeout-exception objects are raised when a call to the `thread-join!' procedure reaches its timeout before the target thread terminates and a timeout-value parameter is not specified. The parameter EXC must be a join-timeout-exception object. The procedure `join-timeout-exception?' returns `#t' when OBJ is a join-timeout-exception object and `#f' otherwise. The procedure `join-timeout-exception-procedure' returns the procedure that raised EXC. The procedure `join-timeout-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (join-timeout-exception? exc) (list (join-timeout-exception-procedure exc) (join-timeout-exception-arguments exc)) 'not-join-timeout-exception)) > (with-exception-catcher handler (lambda () (thread-join! (thread-start! (make-thread (lambda () (thread-sleep! 10)))) 5))) (# (# 5)) -- procedure: started-thread-exception? OBJ -- procedure: started-thread-exception-procedure EXC -- procedure: started-thread-exception-arguments EXC Started-thread-exception objects are raised when the target thread of a call to the procedure `thread-start!' is already started. The parameter EXC must be a started-thread-exception object. The procedure `started-thread-exception?' returns `#t' when OBJ is a started-thread-exception object and `#f' otherwise. The procedure `started-thread-exception-procedure' returns the procedure that raised EXC. The procedure `started-thread-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (started-thread-exception? exc) (list (started-thread-exception-procedure exc) (started-thread-exception-arguments exc)) 'not-started-thread-exception)) > (with-exception-catcher handler (lambda () (let ((t (make-thread (lambda () (expt 2 1000))))) (thread-start! t) (thread-start! t)))) (# (#)) -- procedure: terminated-thread-exception? OBJ -- procedure: terminated-thread-exception-procedure EXC -- procedure: terminated-thread-exception-arguments EXC Terminated-thread-exception objects are raised when the `thread-join!' procedure is called and the target thread has terminated as a result of a call to the `thread-terminate!' procedure. The parameter EXC must be a terminated-thread-exception object. The procedure `terminated-thread-exception?' returns `#t' when OBJ is a terminated-thread-exception object and `#f' otherwise. The procedure `terminated-thread-exception-procedure' returns the procedure that raised EXC. The procedure `terminated-thread-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (terminated-thread-exception? exc) (list (terminated-thread-exception-procedure exc) (terminated-thread-exception-arguments exc)) 'not-terminated-thread-exception)) > (with-exception-catcher handler (lambda () (thread-join! (thread-start! (make-thread (lambda () (thread-terminate! (current-thread)))))))) (# (#)) -- procedure: uncaught-exception? OBJ -- procedure: uncaught-exception-procedure EXC -- procedure: uncaught-exception-arguments EXC -- procedure: uncaught-exception-reason EXC Uncaught-exception objects are raised when an object is raised in a thread and that thread does not handle it (i.e. the thread terminated because it did not catch an exception it raised). The parameter EXC must be an uncaught-exception object. The procedure `uncaught-exception?' returns `#t' when OBJ is an uncaught-exception object and `#f' otherwise. The procedure `uncaught-exception-procedure' returns the procedure that raised EXC. The procedure `uncaught-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `uncaught-exception-reason' returns the object that was raised by the thread and not handled by that thread. For example: > (define (handler exc) (if (uncaught-exception? exc) (list (uncaught-exception-procedure exc) (uncaught-exception-arguments exc) (uncaught-exception-reason exc)) 'not-uncaught-exception)) > (with-exception-catcher handler (lambda () (thread-join! (thread-start! (make-thread (lambda () (open-input-file "data" 99))))))) (# (#) #) 15.5 Exception objects related to C-interface ============================================= -- procedure: cfun-conversion-exception? OBJ -- procedure: cfun-conversion-exception-procedure EXC -- procedure: cfun-conversion-exception-arguments EXC -- procedure: cfun-conversion-exception-code EXC -- procedure: cfun-conversion-exception-message EXC Cfun-conversion-exception objects are raised by the C-interface when converting between the Scheme representation and the C representation of a value during a call from Scheme to C. The parameter EXC must be a cfun-conversion-exception object. The procedure `cfun-conversion-exception?' returns `#t' when OBJ is a cfun-conversion-exception object and `#f' otherwise. The procedure `cfun-conversion-exception-procedure' returns the procedure that raised EXC. The procedure `cfun-conversion-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `cfun-conversion-exception-code' returns an exact integer error code that can be converted to a string by the `err-code->string' procedure. The procedure `cfun-conversion-exception-message' returns `#f' or a string giving details of the exception in a human-readable form. For example: $ cat test1.scm (define weird (c-lambda (char-string) nonnull-char-string "___return(___arg1);")) $ gsc test1.scm $ gsi Gambit v4.9.3 > (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))) (# (not-a-string) "(Argument 1) Can't convert to C char-string" #f) > (with-exception-catcher handler (lambda () (weird #f))) (# (#f) "Can't convert result from C nonnull-char-string" #f) -- procedure: sfun-conversion-exception? OBJ -- procedure: sfun-conversion-exception-procedure EXC -- procedure: sfun-conversion-exception-arguments EXC -- procedure: sfun-conversion-exception-code EXC -- procedure: sfun-conversion-exception-message EXC Sfun-conversion-exception objects are raised by the C-interface when converting between the Scheme representation and the C representation of a value during a call from C to Scheme. The parameter EXC must be a sfun-conversion-exception object. The procedure `sfun-conversion-exception?' returns `#t' when OBJ is a sfun-conversion-exception object and `#f' otherwise. The procedure `sfun-conversion-exception-procedure' returns the procedure that raised EXC. The procedure `sfun-conversion-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `sfun-conversion-exception-code' returns an exact integer error code that can be converted to a string by the `err-code->string' procedure. The procedure `sfun-conversion-exception-message' returns `#f' or a string giving details of the exception in a human-readable form. For example: $ cat test2.scm (c-define (f str) (nonnull-char-string) int "f" "" (string->number str)) (define t1 (c-lambda () int "___return(f (\"123\"));")) (define t2 (c-lambda () int "___return(f (0));")) (define t3 (c-lambda () int "___return(f (\"1.5\"));")) $ gsc test2.scm $ gsi Gambit v4.9.3 > (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) (# () "(Argument 1) Can't convert from C nonnull-char-string" #f) > (with-exception-catcher handler t3) (# () "Can't convert result to C int" #f) -- procedure: multiple-c-return-exception? OBJ Multiple-c-return-exception objects are raised by the C-interface when a C to Scheme procedure call returns and that call's stack frame is no longer on the C stack because the call has already returned, or has been removed from the C stack by a `longjump'. The procedure `multiple-c-return-exception?' returns `#t' when OBJ is a multiple-c-return-exception object and `#f' otherwise. For example: $ cat test3.scm (c-define (f str) (char-string) scheme-object "f" "" (pp (list 'entry 'str= str)) (let ((k (call-with-current-continuation (lambda (k) k)))) (pp (list 'exit 'k= k)) k)) (define scheme-to-c-to-scheme-and-back (c-lambda (char-string) scheme-object "___return(f (___arg1));")) $ gsc test3.scm $ gsi Gambit v4.9.3 > (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= #) # (exit k= 999) # 15.6 Exception objects related to the reader ============================================ -- procedure: datum-parsing-exception? OBJ -- procedure: datum-parsing-exception-kind EXC -- procedure: datum-parsing-exception-parameters EXC -- procedure: datum-parsing-exception-readenv EXC Datum-parsing-exception objects are raised by the reader (i.e. the `read' procedure) when the input does not conform to the grammar for datum. The parameter EXC must be a datum-parsing-exception object. The procedure `datum-parsing-exception?' returns `#t' when OBJ is a datum-parsing-exception object and `#f' otherwise. The procedure `datum-parsing-exception-kind' returns a symbol denoting the kind of parsing error that was encountered by the reader when it raised EXC. Here is a table of the possible return values: `datum-or-eof-expected' Datum or EOF expected `datum-expected' Datum expected `improperly-placed-dot' Improperly placed dot `incomplete-form-eof-reached' Incomplete form, EOF reached `incomplete-form' Incomplete form `character-out-of-range' Character out of range `invalid-character-name' Invalid '#\' name `illegal-character' Illegal character `s8-expected' Signed 8 bit exact integer expected `u8-expected' Unsigned 8 bit exact integer expected `s16-expected' Signed 16 bit exact integer expected `u16-expected' Unsigned 16 bit exact integer expected `s32-expected' Signed 32 bit exact integer expected `u32-expected' Unsigned 32 bit exact integer expected `s64-expected' Signed 64 bit exact integer expected `u64-expected' Unsigned 64 bit exact integer expected `inexact-real-expected' Inexact real expected `invalid-hex-escape' Invalid hexadecimal escape `invalid-escaped-character' Invalid escaped character `open-paren-expected' '(' expected `invalid-token' Invalid token `invalid-sharp-bang-name' Invalid '#!' name `duplicate-label-definition' Duplicate definition for label `missing-label-definition' Missing definition for label `illegal-label-definition' Illegal definition of label `invalid-infix-syntax-character' Invalid infix syntax character `invalid-infix-syntax-number' Invalid infix syntax number `invalid-infix-syntax' Invalid infix syntax The procedure `datum-parsing-exception-parameters' returns a list of the parameters associated with the parsing error that was encountered by the reader when it raised EXC. For example: > (define (handler exc) (if (datum-parsing-exception? exc) (list (datum-parsing-exception-kind exc) (datum-parsing-exception-parameters exc)) 'not-datum-parsing-exception)) > (with-exception-catcher handler (lambda () (with-input-from-string "(s #\\pace)" read))) (invalid-character-name ("pace")) 15.7 Exception objects related to evaluation and compilation ============================================================ -- procedure: expression-parsing-exception? OBJ -- procedure: expression-parsing-exception-kind EXC -- procedure: expression-parsing-exception-parameters EXC -- procedure: expression-parsing-exception-source EXC Expression-parsing-exception objects are raised by the evaluator and compiler (i.e. the procedures `eval', `compile-file', etc) when the input does not conform to the grammar for expression. The parameter EXC must be a expression-parsing-exception object. The procedure `expression-parsing-exception?' returns `#t' when OBJ is a expression-parsing-exception object and `#f' otherwise. The procedure `expression-parsing-exception-kind' returns a symbol denoting the kind of parsing error that was encountered by the evaluator or compiler when it raised EXC. Here is a table of the possible return values: `id-expected' Identifier expected `ill-formed-namespace' Ill-formed namespace `ill-formed-namespace-prefix' Ill-formed namespace prefix `namespace-prefix-must-be-string' Namespace prefix must be a string `macro-used-as-variable' Macro name can't be used as a variable `variable-is-immutable' Variable is immutable `ill-formed-macro-transformer' Macro transformer must be a lambda expression `reserved-used-as-variable' Reserved identifier can't be used as a variable `ill-formed-special-form' Ill-formed special form `cannot-open-file' Can't open file `filename-expected' Filename expected `ill-placed-define' Ill-placed 'define' `ill-placed-**include' Ill-placed '##include' `ill-placed-**define-macro' Ill-placed '##define-macro' `ill-placed-**declare' Ill-placed '##declare' `ill-placed-**namespace' Ill-placed '##namespace' `ill-formed-expression' Ill-formed expression `unsupported-special-form' Interpreter does not support `ill-placed-unquote' Ill-placed 'unquote' `ill-placed-unquote-splicing' Ill-placed 'unquote-splicing' `parameter-must-be-id' Parameter must be an identifier `parameter-must-be-id-or-default' Parameter must be an identifier or default binding `duplicate-parameter' Duplicate parameter in parameter list `ill-placed-dotted-rest-parameter' Ill-placed dotted rest parameter `parameter-expected-after-rest' #!rest must be followed by a parameter `ill-formed-default' Ill-formed default binding `ill-placed-optional' Ill-placed #!optional `ill-placed-rest' Ill-placed #!rest `ill-placed-key' Ill-placed #!key `key-expected-after-rest' #!key expected after rest parameter `ill-placed-default' Ill-placed default binding `duplicate-variable-definition' Duplicate definition of a variable `empty-body' Body must contain at least one expression `variable-must-be-id' Defined variable must be an identifier `else-clause-not-last' Else clause must be last `ill-formed-selector-list' Ill-formed selector list `duplicate-variable-binding' Duplicate variable in bindings `ill-formed-binding-list' Ill-formed binding list `ill-formed-call' Ill-formed procedure call `ill-formed-cond-expand' Ill-formed 'cond-expand' `unfulfilled-cond-expand' Unfulfilled 'cond-expand' The procedure `expression-parsing-exception-parameters' returns a list of the parameters associated with the parsing error that was encountered by the evaluator or compiler when it raised EXC. For example: > (define (handler exc) (if (expression-parsing-exception? exc) (list (expression-parsing-exception-kind exc) (expression-parsing-exception-parameters exc)) 'not-expression-parsing-exception)) > (with-exception-catcher handler (lambda () (eval '(+ do 1)))) (reserved-used-as-variable (do)) -- procedure: unbound-global-exception? OBJ -- procedure: unbound-global-exception-variable EXC -- procedure: unbound-global-exception-code EXC -- procedure: unbound-global-exception-rte EXC Unbound-global-exception objects are raised when an unbound global variable is accessed. The parameter EXC must be an unbound-global-exception object. The procedure `unbound-global-exception?' returns `#t' when OBJ is an unbound-global-exception object and `#f' otherwise. The procedure `unbound-global-exception-variable' returns a symbol identifying the unbound global variable. For example: > (define (handler exc) (if (unbound-global-exception? exc) (list 'variable= (unbound-global-exception-variable exc)) 'not-unbound-global-exception)) > (with-exception-catcher handler (lambda () foo)) (variable= foo) 15.8 Exception objects related to type checking =============================================== -- procedure: type-exception? OBJ -- procedure: type-exception-procedure EXC -- procedure: type-exception-arguments EXC -- procedure: type-exception-arg-num EXC -- procedure: type-exception-type-id EXC Type-exception objects are raised when a primitive procedure is called with an argument of incorrect type (i.e. when a run time type-check fails). The parameter EXC must be a type-exception object. The procedure `type-exception?' returns `#t' when OBJ is a type-exception object and `#f' otherwise. The procedure `type-exception-procedure' returns the procedure that raised EXC. The procedure `type-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `type-exception-arg-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))) (# (#(a b c) foo) 2 exact-integer) > (with-exception-catcher handler (lambda () (time->seconds 'foo))) (#seconds> (foo) 1 #) -- procedure: range-exception? OBJ -- procedure: range-exception-procedure EXC -- procedure: range-exception-arguments EXC -- procedure: range-exception-arg-num EXC Range-exception objects are raised when a numeric parameter is not in the allowed range. The parameter EXC must be a range-exception object. The procedure `range-exception?' returns `#t' when OBJ is a range-exception object and `#f' otherwise. The procedure `range-exception-procedure' returns the procedure that raised EXC. The procedure `range-exception-arguments' returns the list of arguments of the procedure that raised EXC. The procedure `range-exception-arg-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))) (# ("abcde" 10) 2) -- procedure: divide-by-zero-exception? OBJ -- procedure: divide-by-zero-exception-procedure EXC -- procedure: divide-by-zero-exception-arguments EXC Divide-by-zero-exception objects are raised when a division by zero is attempted. The parameter EXC must be a divide-by-zero-exception object. The procedure `divide-by-zero-exception?' returns `#t' when OBJ is a divide-by-zero-exception object and `#f' otherwise. The procedure `divide-by-zero-exception-procedure' returns the procedure that raised EXC. The procedure `divide-by-zero-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (divide-by-zero-exception? exc) (list (divide-by-zero-exception-procedure exc) (divide-by-zero-exception-arguments exc)) 'not-divide-by-zero-exception)) > (with-exception-catcher handler (lambda () (/ 5 0 7))) (# (5 0 7)) -- procedure: improper-length-list-exception? OBJ -- procedure: improper-length-list-exception-procedure EXC -- procedure: improper-length-list-exception-arguments EXC -- procedure: improper-length-list-exception-arg-num EXC 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)))) (# (# (1 2) (3) (4 5)) 3) 15.9 Exception objects related to procedure call ================================================ -- procedure: wrong-number-of-arguments-exception? OBJ -- procedure: wrong-number-of-arguments-exception-procedure EXC -- procedure: wrong-number-of-arguments-exception-arguments EXC Wrong-number-of-arguments-exception objects are raised when a procedure is called with the wrong number of arguments. The parameter EXC must be a wrong-number-of-arguments-exception object. The procedure `wrong-number-of-arguments-exception?' returns `#t' when OBJ is a wrong-number-of-arguments-exception object and `#f' otherwise. The procedure `wrong-number-of-arguments-exception-procedure' returns the procedure that raised EXC. The procedure `wrong-number-of-arguments-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (wrong-number-of-arguments-exception? exc) (list (wrong-number-of-arguments-exception-procedure exc) (wrong-number-of-arguments-exception-arguments exc)) 'not-wrong-number-of-arguments-exception)) > (with-exception-catcher handler (lambda () (open-input-file "data" 99))) (# ("data" 99)) -- procedure: number-of-arguments-limit-exception? OBJ -- procedure: number-of-arguments-limit-exception-procedure EXC -- procedure: number-of-arguments-limit-exception-arguments EXC Number-of-arguments-limit-exception objects are raised by the `apply' procedure when the procedure being called is passed more than 8192 arguments. The parameter EXC must be a number-of-arguments-limit-exception object. The procedure `number-of-arguments-limit-exception?' returns `#t' when OBJ is a number-of-arguments-limit-exception object and `#f' otherwise. The procedure `number-of-arguments-limit-exception-procedure' returns the target procedure of the call to `apply' that raised EXC. The procedure `number-of-arguments-limit-exception-arguments' returns the list of arguments of the target procedure of the call to `apply' that raised EXC. For example: > (define (iota n) (if (= n 0) '() (cons n (iota (- n 1))))) > (define (handler exc) (if (number-of-arguments-limit-exception? exc) (list (number-of-arguments-limit-exception-procedure exc) (length (number-of-arguments-limit-exception-arguments exc))) 'not-number-of-arguments-limit-exception)) > (with-exception-catcher handler (lambda () (apply + 1 2 3 (iota 8190)))) (# 8193) -- procedure: nonprocedure-operator-exception? OBJ -- procedure: nonprocedure-operator-exception-operator EXC -- procedure: nonprocedure-operator-exception-arguments EXC -- procedure: nonprocedure-operator-exception-code EXC -- procedure: nonprocedure-operator-exception-rte EXC Nonprocedure-operator-exception objects are raised when a procedure call is executed and the operator position is not a procedure. The parameter EXC must be an nonprocedure-operator-exception object. The procedure `nonprocedure-operator-exception?' returns `#t' when OBJ is an nonprocedure-operator-exception object and `#f' otherwise. The procedure `nonprocedure-operator-exception-operator' returns the value in operator position of the procedure call that raised EXC. The procedure `nonprocedure-operator-exception-arguments' returns the list of arguments of the procedure call that raised EXC. For example: > (define (handler exc) (if (nonprocedure-operator-exception? exc) (list (nonprocedure-operator-exception-operator exc) (nonprocedure-operator-exception-arguments exc)) 'not-nonprocedure-operator-exception)) > (with-exception-catcher handler (lambda () (11 22 33))) (11 (22 33)) -- procedure: unknown-keyword-argument-exception? OBJ -- procedure: unknown-keyword-argument-exception-procedure EXC -- procedure: unknown-keyword-argument-exception-arguments EXC Unknown-keyword-argument-exception objects are raised when a procedure accepting keyword arguments is called and one of the keywords supplied is not among those that are expected. The parameter EXC must be an unknown-keyword-argument-exception object. The procedure `unknown-keyword-argument-exception?' returns `#t' when OBJ is an unknown-keyword-argument-exception object and `#f' otherwise. The procedure `unknown-keyword-argument-exception-procedure' returns the procedure that raised EXC. The procedure `unknown-keyword-argument-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (unknown-keyword-argument-exception? exc) (list (unknown-keyword-argument-exception-procedure exc) (unknown-keyword-argument-exception-arguments exc)) 'not-unknown-keyword-argument-exception)) > (with-exception-catcher handler (lambda () ((lambda (#!key (foo 5)) foo) bar: 11))) (# (bar: 11)) -- procedure: keyword-expected-exception? OBJ -- procedure: keyword-expected-exception-procedure EXC -- procedure: keyword-expected-exception-arguments EXC Keyword-expected-exception objects are raised when a procedure accepting keyword arguments is called and a nonkeyword was supplied where a keyword was expected. The parameter EXC must be an keyword-expected-exception object. The procedure `keyword-expected-exception?' returns `#t' when OBJ is an keyword-expected-exception object and `#f' otherwise. The procedure `keyword-expected-exception-procedure' returns the procedure that raised EXC. The procedure `keyword-expected-exception-arguments' returns the list of arguments of the procedure that raised EXC. For example: > (define (handler exc) (if (keyword-expected-exception? exc) (list (keyword-expected-exception-procedure exc) (keyword-expected-exception-arguments exc)) 'not-keyword-expected-exception)) > (with-exception-catcher handler (lambda () ((lambda (#!key (foo 5)) foo) 11 22))) (# (11 22)) 15.10 Other exception objects ============================= -- procedure: error-exception? OBJ -- procedure: error-exception-message EXC -- procedure: error-exception-parameters EXC -- procedure: error MESSAGE OBJ... Error-exception objects are raised when the `error' procedure is called. The parameter EXC must be an error-exception object. The procedure `error-exception?' returns `#t' when OBJ is an error-exception object and `#f' otherwise. The procedure `error-exception-message' returns the first argument of the call to `error' that raised EXC. The procedure `error-exception-parameters' returns the list of arguments, starting with the second argument, of the call to `error' that raised EXC. The `error' procedure raises an error-exception object whose message field is MESSAGE and parameters field is the list of values OBJ.... For example: > (define (handler exc) (if (error-exception? exc) (list (error-exception-message exc) (error-exception-parameters exc)) 'not-error-exception)) > (with-exception-catcher handler (lambda () (error "unexpected object:" 123))) ("unexpected object:" (123)) 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 `GAMBOPT' 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. -- procedure: current-directory [NEW-CURRENT-DIRECTORY] 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/foo" > (parameterize ((current-directory "~~")) (path-expand "foo")) "/usr/local/Gambit/foo" -- procedure: path-expand PATH [ORIGIN-DIRECTORY] The procedure `path-expand' takes the path of a file or directory and returns an expanded path, which is an absolute path when PATH or ORIGIN-DIRECTORY are absolute paths. The optional ORIGIN-DIRECTORY parameter, which defaults to the current working directory, is the directory used to resolve relative paths. Components of the paths PATH and ORIGIN-DIRECTORY need not exist. For example under UNIX: > (path-expand "foo") "/Users/feeley/gambit/doc/foo" > (path-expand "~/foo") "/Users/feeley/foo" > (path-expand "~~lib/foo") "/usr/local/Gambit/lib/foo" > (path-expand "../foo") "/Users/feeley/gambit/doc/../foo" > (path-expand "foo" "") "foo" > (path-expand "foo" "/tmp") "/tmp/foo" > (path-expand "this/file/does/not/exist") "/Users/feeley/gambit/doc/this/file/does/not/exist" > (path-expand "") "/Users/feeley/gambit/doc/" -- procedure: path-normalize PATH [ALLOW-RELATIVE? [ORIGIN-DIRECTORY]] The procedure `path-normalize' takes a path of a file or directory and returns its normalized path. The optional ORIGIN-DIRECTORY parameter, which defaults to the current working directory, is the directory used to resolve relative paths. All components of the paths PATH and ORIGIN-DIRECTORY must exist, except possibly the last component of PATH. A normalized path is a path containing no redundant parts and which is consistent with the current structure of the filesystem. A normalized path of a directory will always end with a path separator (i.e. `/', `\', or `:' depending on the operating system). The optional ALLOW-RELATIVE? parameter, which defaults to `#f', indicates if the path returned can be expressed relatively to ORIGIN-DIRECTORY: a `#f' requests an absolute path, the symbol `shortest' requests the shortest of the absolute and relative paths, and any other value requests the relative path. The shortest path is useful for interaction with the user because short relative paths are typically easier to read than long absolute paths. For example under UNIX: > (path-expand "../foo") "/Users/feeley/gambit/doc/../foo" > (path-normalize "../foo") "/Users/feeley/gambit/foo" > (path-normalize "this/file/does/not/exist") *** ERROR IN (console)@3.1 -- No such file or directory (path-normalize "this/file/does/not/exist") -- procedure: path-extension PATH -- procedure: path-strip-extension PATH -- procedure: path-directory PATH -- procedure: path-strip-directory PATH -- procedure: path-strip-trailing-directory-separator PATH -- procedure: path-volume PATH -- procedure: path-strip-volume PATH These procedures extract various parts of a path, which need not be a normalized path. The procedure `path-extension' returns the file extension (including the period) or the empty string if there is no extension. The procedure `path-strip-extension' returns the path with the extension stripped off. The procedure `path-directory' returns the file's directory (including the last path separator) or the empty string if no directory is specified in the path. The procedure `path-strip-directory' returns the path with the directory stripped off. The procedure `path-strip-trailing-directory-separator' returns the path with the directory separator stripped off if one is at the end of the path. The procedure `path-volume' returns the file's volume (including the last path separator) or the empty string if no volume is specified in the path. The procedure `path-strip-volume' returns the path with the volume stripped off. For example under UNIX: > (path-extension "/tmp/foo") "" > (path-extension "/tmp/foo.txt") ".txt" > (path-strip-extension "/tmp/foo.txt") "/tmp/foo" > (path-directory "/tmp/foo.txt") "/tmp/" > (path-strip-directory "/tmp/foo.txt") "foo.txt" > (path-strip-trailing-directory-separator "/usr/local/bin/") "/usr/local/bin" > (path-strip-trailing-directory-separator "/usr/local/bin") "/usr/local/bin" > (path-volume "/tmp/foo.txt") "" > (path-volume "C:/tmp/foo.txt") "" ; result is "C:" under Microsoft Windows > (path-strip-volume "C:/tmp/foo.txt") "C:/tmp/foo.txt" ; result is "/tmp/foo.txt" under Microsoft Windows 16.2 Filesystem operations ========================== -- procedure: create-directory PATH-OR-SETTINGS 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") -- procedure: create-fifo PATH-OR-SETTINGS This procedure creates a FIFO. The argument PATH-OR-SETTINGS is either a string denoting a filesystem path or a list of port settings which must contain a `path:' setting. Here are the settings allowed: * `path:' STRING This setting indicates the location of the FIFO to create in the filesystem. There is no default value for this setting. * `permissions:' 12-BIT-EXACT-INTEGER This setting controls the UNIX permissions that will be attached to the file if it is created. The default value of this setting is `#o666'. For example: > (create-fifo "fifo") > (define a (open-input-file "fifo")) > (define b (open-output-file "fifo")) > (display "1 22 333" b) > (force-output b) > (read a) 1 > (read a) 22 -- procedure: create-link SOURCE-PATH DESTINATION-PATH This procedure creates a hard link between SOURCE-PATH and DESTINATION-PATH. The argument SOURCE-PATH must be a string denoting the path of an existing file. The argument DESTINATION-PATH must be a string denoting the path of the link to create. -- procedure: create-symbolic-link SOURCE-PATH DESTINATION-PATH This procedure creates a symbolic link between SOURCE-PATH and DESTINATION-PATH. The argument SOURCE-PATH must be a string denoting the path of an existing file. The argument DESTINATION-PATH must be a string denoting the path of the symbolic link to create. -- procedure: rename-file SOURCE-PATH DESTINATION-PATH 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. -- procedure: copy-file SOURCE-PATH DESTINATION-PATH This procedure copies the file SOURCE-PATH to DESTINATION-PATH. The argument SOURCE-PATH must be a string denoting the path of an existing file. The argument DESTINATION-PATH must be a string denoting the path of the file to create. -- procedure: delete-file PATH This procedure deletes the file PATH. The argument PATH must be a string denoting the path of an existing file. -- procedure: delete-directory PATH This procedure deletes the directory PATH. The argument PATH must be a string denoting the path of an existing directory. -- procedure: directory-files [PATH-OR-SETTINGS] This procedure returns the list of the files in a directory. The argument PATH-OR-SETTINGS is either a string denoting a filesystem path to a directory or a list of settings which must contain a `path:' setting. If it is not specified, PATH-OR-SETTINGS defaults to the current directory (the value bound to the `current-directory' parameter object). Here are the settings allowed: * `path:' STRING This setting indicates the location of the directory in the filesystem. There is no default value for this setting. * `ignore-hidden:' ( `#f' | `#t' | `dot-and-dot-dot' ) This setting controls whether hidden-files will be returned. Under UNIX and 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 ============================ -- procedure: shell-command COMMAND [CAPTURE?] The procedure `shell-command' calls up the shell to execute COMMAND which must be a string. The argument CAPTURE?, which defaults to `#f', indicates if the output of the command is captured as a string. If CAPTURE? is `#f', this procedure returns the exit status of the shell in the form that the C library's `system' routine returns. If CAPTURE? is not `#f', this procedure returns a pair consisting of the exit status of the shell in the `car' field, and the captured output in the `cdr' field. Be advised that the shell that is used, and consequently the syntax of COMMAND, depends on the operating system. On Unix, the shell `/bin/sh' is usually invoked. On Windows, the shell `cmd.exe' is usually invoked. For example under UNIX: > (shell-command "ls -sk f*.scm") 4 fact.scm 4 fib.scm 0 > (shell-command "ls -sk f*.scm" #t) (0 . "4 fact.scm 4 fib.scm\n") > (shell-command "echo x\\\\\\\\y $HOME" #t) (0 . "x\\y /Users/feeley\n") For example under Windows: > (shell-command "echo x\\\\\\\\y %HOME%" #t) (0 . "x\\\\\\\\y C:\\Users\\feeley\r\n") 16.4 Process termination ======================== -- procedure: exit [STATUS] The procedure `exit' causes the process to terminate with the status STATUS which must be an exact integer in the range 0 to 255 or `#f'. If it is not specified, STATUS defaults to 0. When STATUS is `#f' the process terminates with an error status. For example under UNIX: $ gsi Gambit v4.9.3 > (exit #f) $ echo $? 70 16.5 Command line arguments =========================== -- procedure: command-line This procedure returns a list of strings corresponding to the command line arguments, including the program file name as the first element of the list. When the interpreter executes a Scheme script, the list returned by `command-line' contains the script's absolute path followed by the remaining command line arguments. For example under UNIX: $ gsi -:d -e "(pretty-print (command-line))" ("gsi" "-e" "(pretty-print (command-line))") $ cat foo #!/usr/local/Gambit/bin/gsi-script (pretty-print (command-line)) $ ./foo 1 2 "3 4" ("/u/feeley/./foo" "1" "2" "3 4") 16.6 Environment variables ========================== -- procedure: getenv NAME [DEFAULT] -- procedure: setenv NAME [NEW-VALUE] The procedure `getenv' returns the value of the environment variable NAME of the current process. Variable names are denoted with strings. A string is returned if the environment variable is bound, otherwise DEFAULT is returned if it is specified, otherwise an exception is raised. The procedure `setenv' changes the binding of the environment variable NAME to NEW-VALUE which must be a string. If NEW-VALUE is not specified the binding is removed. For example under UNIX: > (getenv "HOME") "/Users/feeley" > (getenv "DOES_NOT_EXIST" #f) #f > (setenv "DOES_NOT_EXIST" "it does now") > (getenv "DOES_NOT_EXIST" #f) "it does now" > (setenv "DOES_NOT_EXIST") > (getenv "DOES_NOT_EXIST" #f) #f > (getenv "DOES_NOT_EXIST") *** ERROR IN (console)@7.1 -- Unbound OS environment variable (getenv "DOES_NOT_EXIST") 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). -- procedure: current-time -- procedure: time? OBJ -- procedure: time->seconds TIME -- procedure: seconds->time X The procedure `current-time' returns a "time object" representing the current point in real time. The procedure `time?' returns `#t' when OBJ is a time object and `#f' otherwise. The procedure `time->seconds' converts the time object TIME into an inexact real number representing the number of seconds elapsed since the "epoch" (which is 00:00:00 Coordinated Universal Time 01-01-1970). The procedure `seconds->time' converts the real number X representing the number of seconds elapsed since the "epoch" into a time object. For example: > (current-time) #