===================================================================== ---------- Mini-Scheme Interpreter Version 0.85 ---------- coded by Atsushi Moriwaki (11/5/1989) E-MAIL : moriwaki@kurims.kurims.kyoto-u.ac.jp MIX : riemann NIFTY : PBB01074 (Note that these addresses are now obsolete, see below) ===================================================================== Revised by Akira KIDA Version 0.85k4 (17 May 1994) Version 0.85k3 (30 Nov 1989) Version 0.85k2 (28 Nov 1989) Version 0.85k1 (14 Nov 1989) Mini-Scheme is now maintained by Akira KIDA. E-Mail : SDI00379@niftyserve.or.jp Most part of this document is written by Akira KIDA. Send comments/requests/bug reports to Akira KIDA at the above email address. ===================================================================== This Mini-Scheme Interpreter is based on "SCHEME Interpreter in Common Lisp" in Appendix of T.Matsuda & K.Saigo, Programming of LISP, archive No5 (1987) p6 - p42 (published in Japan). Copyright Notice: THIS SOFTWARE IS PLACED IN THE PUBLIC DOMAIN BY THE AUTHOR. This software is completely free to copy, modify and/or re-distribute. But I (Atsushi Moriwaki) would appreciate it if you left my name on the code as the author. DISCLAIMER: THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. Supported features (or, NOT supported features :-) 1) Lists, symbols, strings. However, strings have very limited capability. For instance, there is *NO* string-ref, string-set!, ... etc. 2) Numbers are limited to FIXNUM only. There is *NO* complex, real, rational and even bignum. 3) Macro feature is supported, though not the one defined in R4RS. Known problems: 1) Poor error recovery from illegal use of syntax and procedure. 2) Certain procedures do not check its argument type. Installation: 1) Select system declaration and configuration options by editing source file. You may choose one of the following systems by #define'ing the preprocessor symbol. Supported systems are: Macintosh: LSC -- LightSpeed C (3.0) for Macintosh LSC4 -- LightSpeed C (4.0) for Macintosh They are different in #include header only. I (kida) think THINK C 5.0, 6.0, 7.0 may be OK with LSC4 configuration, though not tested. MPW2 -- Macintosh Programmer's Workshop v2.0x I don't tested v3.x or later. DOS: MSC4 -- Microsoft C v4.0 (use /AL) MSC v5.1, v6.0, v7.0 are all OK. TURBO2 -- Bolarnd's Turbo C v2.0 (use -ml) Turbo C++ 1.0 is OK. UNIX: BSD -- UNIX of BSD flavor, ex. SuOS 4.x SYSV -- UNIX of System-V flavor, ex. Sun/Solaris 2.x VAX/VMS: VAXC -- VAX-C v3.x (this symbol may be defined by the compiler automatically). 2) Configure some preprocessor symbols by editing source files. Configurable #define's are: #define VERBOSE -- if defined, GC messages is verbose on default. #define AVOID_HACK_LOOP -- if defined, do _NOT_ use loop construction in the form do { ... } while (0) This form is used in macro expansion, since this is the best "safety" blocking construct when used in macro definition. However, some compiler (including SunPRO CC 2.0.1) is silly enough to warning this construct, like "warning: end-of-loop code not reached", etc. If you dislike such warning, please define this symbol. NOTE: You may get some "statement not reached" warning even if you have define this symbol. Please be patient, or, use more smart compiler. In short if you use GCC, undefine this and forget it at all. #define USE_SETJMP -- if defined, use setjmp to global jump on error. if not defined, avoid to use it. Compiled with this symbol defined, the interpreter issue fatal error and return to the operating system immediately when we run out of free cells. By default, i.e., compiled with this symbol is not defined, the interpreter will just return to the top level in such a case. May not be used except for compiling as Mac DA. #define USE_MACRO -- if defined, macro features are enabled. #define USE_QQUOTE -- if defined, you can use quasi-quote "`" in macro. You can use macro even if this symbol is undefined. 3) Compile! I think there is virtually no problem about how to compile. Since there is exactly one C source file. :-) If you are on UNIX boxes with some BSD flavors, instead of using make command, it's enough to type: cc -DBSD -O -o miniscm miniscm.c You may have additional warnings like 'function should return value'. This is due to omitting 'void' keyword from the source in order to get pre-ANSI compatibility. Usage : miniscm Sorry, no command line argnumet is allowed. Special procedures of this system: gc : (gc) -- force garbage collection gc-verbose : (gc-verbose bool) -- GC verbose on/off Argument #f turnes off the GC message. Enything else turn on the GC message. quit : (quit) -- quit to the operating system put : (put sym prop expr) -- set the value of a property of a symbol. get : (get sym prop) -- get the value of a property of a symbol. new-segment : (new-segment n) -- allocate n new cell segments. print-width : (print-width list) -- returns 'printed' width of list. closure? : (closure? obj) -- test if obj is a closure or not. macro? : (macro? obj) -- test if obj is a macro or not. note that a macro is also a closure. get-closure-code : (get-closure-code closure-obj) -- extract S-expr from closure-obj. Scheme files: init.scm -- Automatically loaded at invocation. Default setting assumes that this file is in the current working directory. Change #define InitFile if you don't like it. tools.scm -- This is a sample file. Contains very tiny pretty-print procedure. After invoking miniscm, type: (load "tools.scm") and try (pp getd) (pp do) Documents?: Sorry, there is no other documents. Do not ask one for me, please see the source instead. :-) But if you _absolutely_ need help, please email to me at: Enjoy! -- Akira KIDA Sysop for FPL in NIFTY-Serve in JAPAN. (FPL stands for 'Forum for Program-Language')