1: \input texinfo @c -*-texinfo-*-
2: @comment The source is gforth.ds, from which gforth.texi is generated
3: @comment %**start of header (This is for running Texinfo on a region.)
4: @setfilename gforth.info
5: @settitle Gforth Manual
6: @comment @setchapternewpage odd
7: @comment %**end of header (This is for running Texinfo on a region.)
8:
9: @ifinfo
10: This file documents Gforth 0.2
11:
12: Copyright @copyright{} 1995,1996 Free Software Foundation, Inc.
13:
14: Permission is granted to make and distribute verbatim copies of
15: this manual provided the copyright notice and this permission notice
16: are preserved on all copies.
17:
18: @ignore
19: Permission is granted to process this file through TeX and print the
20: results, provided the printed document carries a copying permission
21: notice identical to this one except for the removal of this paragraph
22: (this paragraph not being relevant to the printed manual).
23:
24: @end ignore
25: Permission is granted to copy and distribute modified versions of this
26: manual under the conditions for verbatim copying, provided also that the
27: sections entitled "Distribution" and "General Public License" are
28: included exactly as in the original, and provided that the entire
29: resulting derived work is distributed under the terms of a permission
30: notice identical to this one.
31:
32: Permission is granted to copy and distribute translations of this manual
33: into another language, under the above conditions for modified versions,
34: except that the sections entitled "Distribution" and "General Public
35: License" may be included in a translation approved by the author instead
36: of in the original English.
37: @end ifinfo
38:
39: @finalout
40: @titlepage
41: @sp 10
42: @center @titlefont{Gforth Manual}
43: @sp 2
44: @center for version 0.2
45: @sp 2
46: @center Anton Ertl
47: @center Bernd Paysan
48: @sp 3
49: @center This manual is under construction
50:
51: @comment The following two commands start the copyright page.
52: @page
53: @vskip 0pt plus 1filll
54: Copyright @copyright{} 1995,1996 Free Software Foundation, Inc.
55:
56: @comment !! Published by ... or You can get a copy of this manual ...
57:
58: Permission is granted to make and distribute verbatim copies of
59: this manual provided the copyright notice and this permission notice
60: are preserved on all copies.
61:
62: Permission is granted to copy and distribute modified versions of this
63: manual under the conditions for verbatim copying, provided also that the
64: sections entitled "Distribution" and "General Public License" are
65: included exactly as in the original, and provided that the entire
66: resulting derived work is distributed under the terms of a permission
67: notice identical to this one.
68:
69: Permission is granted to copy and distribute translations of this manual
70: into another language, under the above conditions for modified versions,
71: except that the sections entitled "Distribution" and "General Public
72: License" may be included in a translation approved by the author instead
73: of in the original English.
74: @end titlepage
75:
76:
77: @node Top, License, (dir), (dir)
78: @ifinfo
79: Gforth is a free implementation of ANS Forth available on many
80: personal machines. This manual corresponds to version 0.2.
81: @end ifinfo
82:
83: @menu
84: * License::
85: * Goals:: About the Gforth Project
86: * Other Books:: Things you might want to read
87: * Invocation:: Starting Gforth
88: * Words:: Forth words available in Gforth
89: * Tools:: Programming tools
90: * ANS conformance:: Implementation-defined options etc.
91: * Model:: The abstract machine of Gforth
92: * Integrating Gforth:: Forth as scripting language for applications.
93: * Emacs and Gforth:: The Gforth Mode
94: * Internals:: Implementation details
95: * Bugs:: How to report them
96: * Origin:: Authors and ancestors of Gforth
97: * Word Index:: An item for each Forth word
98: * Node Index:: An item for each node
99: @end menu
100:
101: @node License, Goals, Top, Top
102: @unnumbered GNU GENERAL PUBLIC LICENSE
103: @center Version 2, June 1991
104:
105: @display
106: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
107: 675 Mass Ave, Cambridge, MA 02139, USA
108:
109: Everyone is permitted to copy and distribute verbatim copies
110: of this license document, but changing it is not allowed.
111: @end display
112:
113: @unnumberedsec Preamble
114:
115: The licenses for most software are designed to take away your
116: freedom to share and change it. By contrast, the GNU General Public
117: License is intended to guarantee your freedom to share and change free
118: software---to make sure the software is free for all its users. This
119: General Public License applies to most of the Free Software
120: Foundation's software and to any other program whose authors commit to
121: using it. (Some other Free Software Foundation software is covered by
122: the GNU Library General Public License instead.) You can apply it to
123: your programs, too.
124:
125: When we speak of free software, we are referring to freedom, not
126: price. Our General Public Licenses are designed to make sure that you
127: have the freedom to distribute copies of free software (and charge for
128: this service if you wish), that you receive source code or can get it
129: if you want it, that you can change the software or use pieces of it
130: in new free programs; and that you know you can do these things.
131:
132: To protect your rights, we need to make restrictions that forbid
133: anyone to deny you these rights or to ask you to surrender the rights.
134: These restrictions translate to certain responsibilities for you if you
135: distribute copies of the software, or if you modify it.
136:
137: For example, if you distribute copies of such a program, whether
138: gratis or for a fee, you must give the recipients all the rights that
139: you have. You must make sure that they, too, receive or can get the
140: source code. And you must show them these terms so they know their
141: rights.
142:
143: We protect your rights with two steps: (1) copyright the software, and
144: (2) offer you this license which gives you legal permission to copy,
145: distribute and/or modify the software.
146:
147: Also, for each author's protection and ours, we want to make certain
148: that everyone understands that there is no warranty for this free
149: software. If the software is modified by someone else and passed on, we
150: want its recipients to know that what they have is not the original, so
151: that any problems introduced by others will not reflect on the original
152: authors' reputations.
153:
154: Finally, any free program is threatened constantly by software
155: patents. We wish to avoid the danger that redistributors of a free
156: program will individually obtain patent licenses, in effect making the
157: program proprietary. To prevent this, we have made it clear that any
158: patent must be licensed for everyone's free use or not licensed at all.
159:
160: The precise terms and conditions for copying, distribution and
161: modification follow.
162:
163: @iftex
164: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
165: @end iftex
166: @ifinfo
167: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
168: @end ifinfo
169:
170: @enumerate 0
171: @item
172: This License applies to any program or other work which contains
173: a notice placed by the copyright holder saying it may be distributed
174: under the terms of this General Public License. The ``Program'', below,
175: refers to any such program or work, and a ``work based on the Program''
176: means either the Program or any derivative work under copyright law:
177: that is to say, a work containing the Program or a portion of it,
178: either verbatim or with modifications and/or translated into another
179: language. (Hereinafter, translation is included without limitation in
180: the term ``modification''.) Each licensee is addressed as ``you''.
181:
182: Activities other than copying, distribution and modification are not
183: covered by this License; they are outside its scope. The act of
184: running the Program is not restricted, and the output from the Program
185: is covered only if its contents constitute a work based on the
186: Program (independent of having been made by running the Program).
187: Whether that is true depends on what the Program does.
188:
189: @item
190: You may copy and distribute verbatim copies of the Program's
191: source code as you receive it, in any medium, provided that you
192: conspicuously and appropriately publish on each copy an appropriate
193: copyright notice and disclaimer of warranty; keep intact all the
194: notices that refer to this License and to the absence of any warranty;
195: and give any other recipients of the Program a copy of this License
196: along with the Program.
197:
198: You may charge a fee for the physical act of transferring a copy, and
199: you may at your option offer warranty protection in exchange for a fee.
200:
201: @item
202: You may modify your copy or copies of the Program or any portion
203: of it, thus forming a work based on the Program, and copy and
204: distribute such modifications or work under the terms of Section 1
205: above, provided that you also meet all of these conditions:
206:
207: @enumerate a
208: @item
209: You must cause the modified files to carry prominent notices
210: stating that you changed the files and the date of any change.
211:
212: @item
213: You must cause any work that you distribute or publish, that in
214: whole or in part contains or is derived from the Program or any
215: part thereof, to be licensed as a whole at no charge to all third
216: parties under the terms of this License.
217:
218: @item
219: If the modified program normally reads commands interactively
220: when run, you must cause it, when started running for such
221: interactive use in the most ordinary way, to print or display an
222: announcement including an appropriate copyright notice and a
223: notice that there is no warranty (or else, saying that you provide
224: a warranty) and that users may redistribute the program under
225: these conditions, and telling the user how to view a copy of this
226: License. (Exception: if the Program itself is interactive but
227: does not normally print such an announcement, your work based on
228: the Program is not required to print an announcement.)
229: @end enumerate
230:
231: These requirements apply to the modified work as a whole. If
232: identifiable sections of that work are not derived from the Program,
233: and can be reasonably considered independent and separate works in
234: themselves, then this License, and its terms, do not apply to those
235: sections when you distribute them as separate works. But when you
236: distribute the same sections as part of a whole which is a work based
237: on the Program, the distribution of the whole must be on the terms of
238: this License, whose permissions for other licensees extend to the
239: entire whole, and thus to each and every part regardless of who wrote it.
240:
241: Thus, it is not the intent of this section to claim rights or contest
242: your rights to work written entirely by you; rather, the intent is to
243: exercise the right to control the distribution of derivative or
244: collective works based on the Program.
245:
246: In addition, mere aggregation of another work not based on the Program
247: with the Program (or with a work based on the Program) on a volume of
248: a storage or distribution medium does not bring the other work under
249: the scope of this License.
250:
251: @item
252: You may copy and distribute the Program (or a work based on it,
253: under Section 2) in object code or executable form under the terms of
254: Sections 1 and 2 above provided that you also do one of the following:
255:
256: @enumerate a
257: @item
258: Accompany it with the complete corresponding machine-readable
259: source code, which must be distributed under the terms of Sections
260: 1 and 2 above on a medium customarily used for software interchange; or,
261:
262: @item
263: Accompany it with a written offer, valid for at least three
264: years, to give any third party, for a charge no more than your
265: cost of physically performing source distribution, a complete
266: machine-readable copy of the corresponding source code, to be
267: distributed under the terms of Sections 1 and 2 above on a medium
268: customarily used for software interchange; or,
269:
270: @item
271: Accompany it with the information you received as to the offer
272: to distribute corresponding source code. (This alternative is
273: allowed only for noncommercial distribution and only if you
274: received the program in object code or executable form with such
275: an offer, in accord with Subsection b above.)
276: @end enumerate
277:
278: The source code for a work means the preferred form of the work for
279: making modifications to it. For an executable work, complete source
280: code means all the source code for all modules it contains, plus any
281: associated interface definition files, plus the scripts used to
282: control compilation and installation of the executable. However, as a
283: special exception, the source code distributed need not include
284: anything that is normally distributed (in either source or binary
285: form) with the major components (compiler, kernel, and so on) of the
286: operating system on which the executable runs, unless that component
287: itself accompanies the executable.
288:
289: If distribution of executable or object code is made by offering
290: access to copy from a designated place, then offering equivalent
291: access to copy the source code from the same place counts as
292: distribution of the source code, even though third parties are not
293: compelled to copy the source along with the object code.
294:
295: @item
296: You may not copy, modify, sublicense, or distribute the Program
297: except as expressly provided under this License. Any attempt
298: otherwise to copy, modify, sublicense or distribute the Program is
299: void, and will automatically terminate your rights under this License.
300: However, parties who have received copies, or rights, from you under
301: this License will not have their licenses terminated so long as such
302: parties remain in full compliance.
303:
304: @item
305: You are not required to accept this License, since you have not
306: signed it. However, nothing else grants you permission to modify or
307: distribute the Program or its derivative works. These actions are
308: prohibited by law if you do not accept this License. Therefore, by
309: modifying or distributing the Program (or any work based on the
310: Program), you indicate your acceptance of this License to do so, and
311: all its terms and conditions for copying, distributing or modifying
312: the Program or works based on it.
313:
314: @item
315: Each time you redistribute the Program (or any work based on the
316: Program), the recipient automatically receives a license from the
317: original licensor to copy, distribute or modify the Program subject to
318: these terms and conditions. You may not impose any further
319: restrictions on the recipients' exercise of the rights granted herein.
320: You are not responsible for enforcing compliance by third parties to
321: this License.
322:
323: @item
324: If, as a consequence of a court judgment or allegation of patent
325: infringement or for any other reason (not limited to patent issues),
326: conditions are imposed on you (whether by court order, agreement or
327: otherwise) that contradict the conditions of this License, they do not
328: excuse you from the conditions of this License. If you cannot
329: distribute so as to satisfy simultaneously your obligations under this
330: License and any other pertinent obligations, then as a consequence you
331: may not distribute the Program at all. For example, if a patent
332: license would not permit royalty-free redistribution of the Program by
333: all those who receive copies directly or indirectly through you, then
334: the only way you could satisfy both it and this License would be to
335: refrain entirely from distribution of the Program.
336:
337: If any portion of this section is held invalid or unenforceable under
338: any particular circumstance, the balance of the section is intended to
339: apply and the section as a whole is intended to apply in other
340: circumstances.
341:
342: It is not the purpose of this section to induce you to infringe any
343: patents or other property right claims or to contest validity of any
344: such claims; this section has the sole purpose of protecting the
345: integrity of the free software distribution system, which is
346: implemented by public license practices. Many people have made
347: generous contributions to the wide range of software distributed
348: through that system in reliance on consistent application of that
349: system; it is up to the author/donor to decide if he or she is willing
350: to distribute software through any other system and a licensee cannot
351: impose that choice.
352:
353: This section is intended to make thoroughly clear what is believed to
354: be a consequence of the rest of this License.
355:
356: @item
357: If the distribution and/or use of the Program is restricted in
358: certain countries either by patents or by copyrighted interfaces, the
359: original copyright holder who places the Program under this License
360: may add an explicit geographical distribution limitation excluding
361: those countries, so that distribution is permitted only in or among
362: countries not thus excluded. In such case, this License incorporates
363: the limitation as if written in the body of this License.
364:
365: @item
366: The Free Software Foundation may publish revised and/or new versions
367: of the General Public License from time to time. Such new versions will
368: be similar in spirit to the present version, but may differ in detail to
369: address new problems or concerns.
370:
371: Each version is given a distinguishing version number. If the Program
372: specifies a version number of this License which applies to it and ``any
373: later version'', you have the option of following the terms and conditions
374: either of that version or of any later version published by the Free
375: Software Foundation. If the Program does not specify a version number of
376: this License, you may choose any version ever published by the Free Software
377: Foundation.
378:
379: @item
380: If you wish to incorporate parts of the Program into other free
381: programs whose distribution conditions are different, write to the author
382: to ask for permission. For software which is copyrighted by the Free
383: Software Foundation, write to the Free Software Foundation; we sometimes
384: make exceptions for this. Our decision will be guided by the two goals
385: of preserving the free status of all derivatives of our free software and
386: of promoting the sharing and reuse of software generally.
387:
388: @iftex
389: @heading NO WARRANTY
390: @end iftex
391: @ifinfo
392: @center NO WARRANTY
393: @end ifinfo
394:
395: @item
396: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
397: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
398: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
399: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
400: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
401: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
402: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
403: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
404: REPAIR OR CORRECTION.
405:
406: @item
407: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
408: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
409: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
410: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
411: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
412: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
413: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
414: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
415: POSSIBILITY OF SUCH DAMAGES.
416: @end enumerate
417:
418: @iftex
419: @heading END OF TERMS AND CONDITIONS
420: @end iftex
421: @ifinfo
422: @center END OF TERMS AND CONDITIONS
423: @end ifinfo
424:
425: @page
426: @unnumberedsec How to Apply These Terms to Your New Programs
427:
428: If you develop a new program, and you want it to be of the greatest
429: possible use to the public, the best way to achieve this is to make it
430: free software which everyone can redistribute and change under these terms.
431:
432: To do so, attach the following notices to the program. It is safest
433: to attach them to the start of each source file to most effectively
434: convey the exclusion of warranty; and each file should have at least
435: the ``copyright'' line and a pointer to where the full notice is found.
436:
437: @smallexample
438: @var{one line to give the program's name and a brief idea of what it does.}
439: Copyright (C) 19@var{yy} @var{name of author}
440:
441: This program is free software; you can redistribute it and/or modify
442: it under the terms of the GNU General Public License as published by
443: the Free Software Foundation; either version 2 of the License, or
444: (at your option) any later version.
445:
446: This program is distributed in the hope that it will be useful,
447: but WITHOUT ANY WARRANTY; without even the implied warranty of
448: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
449: GNU General Public License for more details.
450:
451: You should have received a copy of the GNU General Public License
452: along with this program; if not, write to the Free Software
453: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
454: @end smallexample
455:
456: Also add information on how to contact you by electronic and paper mail.
457:
458: If the program is interactive, make it output a short notice like this
459: when it starts in an interactive mode:
460:
461: @smallexample
462: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
463: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
464: type `show w'.
465: This is free software, and you are welcome to redistribute it
466: under certain conditions; type `show c' for details.
467: @end smallexample
468:
469: The hypothetical commands @samp{show w} and @samp{show c} should show
470: the appropriate parts of the General Public License. Of course, the
471: commands you use may be called something other than @samp{show w} and
472: @samp{show c}; they could even be mouse-clicks or menu items---whatever
473: suits your program.
474:
475: You should also get your employer (if you work as a programmer) or your
476: school, if any, to sign a ``copyright disclaimer'' for the program, if
477: necessary. Here is a sample; alter the names:
478:
479: @smallexample
480: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
481: `Gnomovision' (which makes passes at compilers) written by James Hacker.
482:
483: @var{signature of Ty Coon}, 1 April 1989
484: Ty Coon, President of Vice
485: @end smallexample
486:
487: This General Public License does not permit incorporating your program into
488: proprietary programs. If your program is a subroutine library, you may
489: consider it more useful to permit linking proprietary applications with the
490: library. If this is what you want to do, use the GNU Library General
491: Public License instead of this License.
492:
493: @iftex
494: @node Preface
495: @comment node-name, next, previous, up
496: @unnumbered Preface
497: @cindex Preface
498: This manual documents Gforth. The reader is expected to know
499: Forth. This manual is primarily a reference manual. @xref{Other Books}
500: for introductory material.
501: @end iftex
502:
503: @node Goals, Other Books, License, Top
504: @comment node-name, next, previous, up
505: @chapter Goals of Gforth
506: @cindex Goals
507: The goal of the Gforth Project is to develop a standard model for
508: ANSI Forth. This can be split into several subgoals:
509:
510: @itemize @bullet
511: @item
512: Gforth should conform to the ANSI Forth standard.
513: @item
514: It should be a model, i.e. it should define all the
515: implementation-dependent things.
516: @item
517: It should become standard, i.e. widely accepted and used. This goal
518: is the most difficult one.
519: @end itemize
520:
521: To achieve these goals Gforth should be
522: @itemize @bullet
523: @item
524: Similar to previous models (fig-Forth, F83)
525: @item
526: Powerful. It should provide for all the things that are considered
527: necessary today and even some that are not yet considered necessary.
528: @item
529: Efficient. It should not get the reputation of being exceptionally
530: slow.
531: @item
532: Free.
533: @item
534: Available on many machines/easy to port.
535: @end itemize
536:
537: Have we achieved these goals? Gforth conforms to the ANS Forth
538: standard. It may be considered a model, but we have not yet documented
539: which parts of the model are stable and which parts we are likely to
540: change. It certainly has not yet become a de facto standard. It has some
541: similarities and some differences to previous models. It has some
542: powerful features, but not yet everything that we envisioned. We
543: certainly have achieved our execution speed goals (@pxref{Performance}).
544: It is free and available on many machines.
545:
546: @node Other Books, Invocation, Goals, Top
547: @chapter Other books on ANS Forth
548:
549: As the standard is relatively new, there are not many books out yet. It
550: is not recommended to learn Forth by using Gforth and a book that is
551: not written for ANS Forth, as you will not know your mistakes from the
552: deviations of the book.
553:
554: There is, of course, the standard, the definite reference if you want to
555: write ANS Forth programs. It is available in printed form from the
556: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
557: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about $200. You
558: can also get it from Global Engineering Documents (Tel.: USA (800)
559: 854-7179; Fax.: (303) 843-9880) for about $300.
560:
561: @cite{dpANS6}, the last draft of the standard, which was then submitted to ANSI
562: for publication is available electronically and for free in some MS Word
563: format, and it has been converted to HTML. Some pointers to these
564: versions can be found through
565: @*@file{http://www.complang.tuwien.ac.at/projects/forth.html}.
566:
567: @cite{Forth: The new model} by Jack Woehr (Prentice-Hall, 1993) is an
568: introductory book based on a draft version of the standard. It does not
569: cover the whole standard. It also contains interesting background
570: information (Jack Woehr was in the ANS Forth Technical Committe). It is
571: not appropriate for complete newbies, but programmers experienced in
572: other languages should find it ok.
573:
574: @node Invocation, Words, Other Books, Top
575: @chapter Invocation
576:
577: You will usually just say @code{gforth}. In many other cases the default
578: Gforth image will be invoked like this:
579:
580: @example
581: gforth [files] [-e forth-code]
582: @end example
583:
584: executing the contents of the files and the Forth code in the order they
585: are given.
586:
587: In general, the command line looks like this:
588:
589: @example
590: gforth [initialization options] [image-specific options]
591: @end example
592:
593: The initialization options must come before the rest of the command
594: line. They are:
595:
596: @table @code
597: @item --image-file @var{file}
598: @item -i @var{file}
599: Loads the Forth image @var{file} instead of the default
600: @file{gforth.fi}.
601:
602: @item --path @var{path}
603: @item -p @var{path}
604: Uses @var{path} for searching the image file and Forth source code files
605: instead of the default in the environment variable @code{GFORTHPATH} or
606: the path specified at installation time (e.g.,
607: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
608: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
609:
610: @item --dictionary-size @var{size}
611: @item -m @var{size}
612: Allocate @var{size} space for the Forth dictionary space instead of
613: using the default specified in the image (typically 256K). The
614: @var{size} specification consists of an integer and a unit (e.g.,
615: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
616: size, in this case Cells), @code{k} (kilobytes), and @code{M}
617: (Megabytes). If no unit is specified, @code{e} is used.
618:
619: @item --data-stack-size @var{size}
620: @item -d @var{size}
621: Allocate @var{size} space for the data stack instead of using the
622: default specified in the image (typically 16K).
623:
624: @item --return-stack-size @var{size}
625: @item -r @var{size}
626: Allocate @var{size} space for the return stack instead of using the
627: default specified in the image (typically 16K).
628:
629: @item --fp-stack-size @var{size}
630: @item -f @var{size}
631: Allocate @var{size} space for the floating point stack instead of
632: using the default specified in the image (typically 16K). In this case
633: the unit specifier @code{e} refers to floating point numbers.
634:
635: @item --locals-stack-size @var{size}
636: @item -l @var{size}
637: Allocate @var{size} space for the locals stack instead of using the
638: default specified in the image (typically 16K).
639:
640: @end table
641:
642: As explained above, the image-specific command-line arguments for the
643: default image @file{gforth.fi} consist of a sequence of filenames and
644: @code{-e @var{forth-code}} options that are interpreted in the seqence
645: in which they are given. The @code{-e @var{forth-code}} or
646: @code{--evaluate @var{forth-code}} option evaluates the forth
647: code. This option takes only one argument; if you want to evaluate more
648: Forth words, you have to quote them or use several @code{-e}s. To exit
649: after processing the command line (instead of entering interactive mode)
650: append @code{-e bye} to the command line.
651:
652: If you have several versions of Gforth installed, @code{gforth} will
653: invoke the version that was installed last. @code{gforth-@var{version}}
654: invokes a specific version. You may want to use the option
655: @code{--path}, if your environment contains the variable
656: @code{GFORTHPATH}.
657:
658: Not yet implemented:
659: On startup the system first executes the system initialization file
660: (unless the option @code{--no-init-file} is given; note that the system
661: resulting from using this option may not be ANS Forth conformant). Then
662: the user initialization file @file{.gforth.fs} is executed, unless the
663: option @code{--no-rc} is given; this file is first searched in @file{.},
664: then in @file{~}, then in the normal path (see above).
665:
666: @node Words, Tools, Invocation, Top
667: @chapter Forth Words
668:
669: @menu
670: * Notation::
671: * Arithmetic::
672: * Stack Manipulation::
673: * Memory access::
674: * Control Structures::
675: * Locals::
676: * Defining Words::
677: * Tokens for Words::
678: * Wordlists::
679: * Files::
680: * Blocks::
681: * Other I/O::
682: * Programming Tools::
683: * Assembler and Code words::
684: * Threading Words::
685: @end menu
686:
687: @node Notation, Arithmetic, Words, Words
688: @section Notation
689:
690: The Forth words are described in this section in the glossary notation
691: that has become a de-facto standard for Forth texts, i.e.
692:
693: @format
694: @var{word} @var{Stack effect} @var{wordset} @var{pronunciation}
695: @end format
696: @var{Description}
697:
698: @table @var
699: @item word
700: The name of the word. BTW, Gforth is case insensitive, so you can
701: type the words in in lower case (However, @pxref{core-idef}).
702:
703: @item Stack effect
704: The stack effect is written in the notation @code{@var{before} --
705: @var{after}}, where @var{before} and @var{after} describe the top of
706: stack entries before and after the execution of the word. The rest of
707: the stack is not touched by the word. The top of stack is rightmost,
708: i.e., a stack sequence is written as it is typed in. Note that Gforth
709: uses a separate floating point stack, but a unified stack
710: notation. Also, return stack effects are not shown in @var{stack
711: effect}, but in @var{Description}. The name of a stack item describes
712: the type and/or the function of the item. See below for a discussion of
713: the types.
714:
715: All words have two stack effects: A compile-time stack effect and a
716: run-time stack effect. The compile-time stack-effect of most words is
717: @var{ -- }. If the compile-time stack-effect of a word deviates from
718: this standard behaviour, or the word does other unusual things at
719: compile time, both stack effects are shown; otherwise only the run-time
720: stack effect is shown.
721:
722: @item pronunciation
723: How the word is pronounced
724:
725: @item wordset
726: The ANS Forth standard is divided into several wordsets. A standard
727: system need not support all of them. So, the fewer wordsets your program
728: uses the more portable it will be in theory. However, we suspect that
729: most ANS Forth systems on personal machines will feature all
730: wordsets. Words that are not defined in the ANS standard have
731: @code{gforth} or @code{gforth-internal} as wordset. @code{gforth}
732: describes words that will work in future releases of Gforth;
733: @code{gforth-internal} words are more volatile. Environmental query
734: strings are also displayed like words; you can recognize them by the
735: @code{environment} in the wordset field.
736:
737: @item Description
738: A description of the behaviour of the word.
739: @end table
740:
741: The type of a stack item is specified by the character(s) the name
742: starts with:
743:
744: @table @code
745: @item f
746: Bool, i.e. @code{false} or @code{true}.
747: @item c
748: Char
749: @item w
750: Cell, can contain an integer or an address
751: @item n
752: signed integer
753: @item u
754: unsigned integer
755: @item d
756: double sized signed integer
757: @item ud
758: double sized unsigned integer
759: @item r
760: Float (on the FP stack)
761: @item a_
762: Cell-aligned address
763: @item c_
764: Char-aligned address (note that a Char may have two bytes in Windows NT)
765: @item f_
766: Float-aligned address
767: @item df_
768: Address aligned for IEEE double precision float
769: @item sf_
770: Address aligned for IEEE single precision float
771: @item xt
772: Execution token, same size as Cell
773: @item wid
774: Wordlist ID, same size as Cell
775: @item f83name
776: Pointer to a name structure
777: @item "
778: string in the input stream (not the stack). The terminating character is
779: a blank by default. If it is not a blank, it is shown in @code{<>}
780: quotes.
781:
782: @end table
783:
784: @node Arithmetic, Stack Manipulation, Notation, Words
785: @section Arithmetic
786: Forth arithmetic is not checked, i.e., you will not hear about integer
787: overflow on addition or multiplication, you may hear about division by
788: zero if you are lucky. The operator is written after the operands, but
789: the operands are still in the original order. I.e., the infix @code{2-1}
790: corresponds to @code{2 1 -}. Forth offers a variety of division
791: operators. If you perform division with potentially negative operands,
792: you do not want to use @code{/} or @code{/mod} with its undefined
793: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
794: former, @pxref{Mixed precision}).
795:
796: @menu
797: * Single precision::
798: * Bitwise operations::
799: * Mixed precision:: operations with single and double-cell integers
800: * Double precision:: Double-cell integer arithmetic
801: * Floating Point::
802: @end menu
803:
804: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
805: @subsection Single precision
806: doc-+
807: doc--
808: doc-*
809: doc-/
810: doc-mod
811: doc-/mod
812: doc-negate
813: doc-abs
814: doc-min
815: doc-max
816:
817: @node Bitwise operations, Mixed precision, Single precision, Arithmetic
818: @subsection Bitwise operations
819: doc-and
820: doc-or
821: doc-xor
822: doc-invert
823: doc-2*
824: doc-2/
825:
826: @node Mixed precision, Double precision, Bitwise operations, Arithmetic
827: @subsection Mixed precision
828: doc-m+
829: doc-*/
830: doc-*/mod
831: doc-m*
832: doc-um*
833: doc-m*/
834: doc-um/mod
835: doc-fm/mod
836: doc-sm/rem
837:
838: @node Double precision, Floating Point, Mixed precision, Arithmetic
839: @subsection Double precision
840:
841: The outer (aka text) interpreter converts numbers containing a dot into
842: a double precision number. Note that only numbers with the dot as last
843: character are standard-conforming.
844:
845: doc-d+
846: doc-d-
847: doc-dnegate
848: doc-dabs
849: doc-dmin
850: doc-dmax
851:
852: @node Floating Point, , Double precision, Arithmetic
853: @subsection Floating Point
854:
855: The format of floating point numbers recognized by the outer (aka text)
856: interpreter is: a signed decimal number, possibly containing a decimal
857: point (@code{.}), followed by @code{E} or @code{e}, optionally followed
858: by a signed integer (the exponent). E.g., @code{1e} ist the same as
859: @code{+1.0e+0}. Note that a number without @code{e}
860: is not interpreted as floating-point number, but as double (if the
861: number contains a @code{.}) or single precision integer. Also,
862: conversions between string and floating point numbers always use base
863: 10, irrespective of the value of @code{BASE}. If @code{BASE} contains a
864: value greater then 14, the @code{E} may be interpreted as digit and the
865: number will be interpreted as integer, unless it has a signed exponent
866: (both @code{+} and @code{-} are allowed as signs).
867:
868: Angles in floating point operations are given in radians (a full circle
869: has 2 pi radians). Note, that Gforth has a separate floating point
870: stack, but we use the unified notation.
871:
872: Floating point numbers have a number of unpleasant surprises for the
873: unwary (e.g., floating point addition is not associative) and even a few
874: for the wary. You should not use them unless you know what you are doing
875: or you don't care that the results you get are totally bogus. If you
876: want to learn about the problems of floating point numbers (and how to
877: avoid them), you might start with @cite{David Goldberg, What Every
878: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
879: Computing Surveys 23(1):5@minus{}48, March 1991}.
880:
881: doc-f+
882: doc-f-
883: doc-f*
884: doc-f/
885: doc-fnegate
886: doc-fabs
887: doc-fmax
888: doc-fmin
889: doc-floor
890: doc-fround
891: doc-f**
892: doc-fsqrt
893: doc-fexp
894: doc-fexpm1
895: doc-fln
896: doc-flnp1
897: doc-flog
898: doc-falog
899: doc-fsin
900: doc-fcos
901: doc-fsincos
902: doc-ftan
903: doc-fasin
904: doc-facos
905: doc-fatan
906: doc-fatan2
907: doc-fsinh
908: doc-fcosh
909: doc-ftanh
910: doc-fasinh
911: doc-facosh
912: doc-fatanh
913:
914: @node Stack Manipulation, Memory access, Arithmetic, Words
915: @section Stack Manipulation
916:
917: Gforth has a data stack (aka parameter stack) for characters, cells,
918: addresses, and double cells, a floating point stack for floating point
919: numbers, a return stack for storing the return addresses of colon
920: definitions and other data, and a locals stack for storing local
921: variables. Note that while every sane Forth has a separate floating
922: point stack, this is not strictly required; an ANS Forth system could
923: theoretically keep floating point numbers on the data stack. As an
924: additional difficulty, you don't know how many cells a floating point
925: number takes. It is reportedly possible to write words in a way that
926: they work also for a unified stack model, but we do not recommend trying
927: it. Instead, just say that your program has an environmental dependency
928: on a separate FP stack.
929:
930: Also, a Forth system is allowed to keep the local variables on the
931: return stack. This is reasonable, as local variables usually eliminate
932: the need to use the return stack explicitly. So, if you want to produce
933: a standard complying program and if you are using local variables in a
934: word, forget about return stack manipulations in that word (see the
935: standard document for the exact rules).
936:
937: @menu
938: * Data stack::
939: * Floating point stack::
940: * Return stack::
941: * Locals stack::
942: * Stack pointer manipulation::
943: @end menu
944:
945: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
946: @subsection Data stack
947: doc-drop
948: doc-nip
949: doc-dup
950: doc-over
951: doc-tuck
952: doc-swap
953: doc-rot
954: doc--rot
955: doc-?dup
956: doc-pick
957: doc-roll
958: doc-2drop
959: doc-2nip
960: doc-2dup
961: doc-2over
962: doc-2tuck
963: doc-2swap
964: doc-2rot
965:
966: @node Floating point stack, Return stack, Data stack, Stack Manipulation
967: @subsection Floating point stack
968: doc-fdrop
969: doc-fnip
970: doc-fdup
971: doc-fover
972: doc-ftuck
973: doc-fswap
974: doc-frot
975:
976: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
977: @subsection Return stack
978: doc->r
979: doc-r>
980: doc-r@
981: doc-rdrop
982: doc-2>r
983: doc-2r>
984: doc-2r@
985: doc-2rdrop
986:
987: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
988: @subsection Locals stack
989:
990: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
991: @subsection Stack pointer manipulation
992: doc-sp@
993: doc-sp!
994: doc-fp@
995: doc-fp!
996: doc-rp@
997: doc-rp!
998: doc-lp@
999: doc-lp!
1000:
1001: @node Memory access, Control Structures, Stack Manipulation, Words
1002: @section Memory access
1003:
1004: @menu
1005: * Stack-Memory transfers::
1006: * Address arithmetic::
1007: * Memory block access::
1008: @end menu
1009:
1010: @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
1011: @subsection Stack-Memory transfers
1012:
1013: doc-@
1014: doc-!
1015: doc-+!
1016: doc-c@
1017: doc-c!
1018: doc-2@
1019: doc-2!
1020: doc-f@
1021: doc-f!
1022: doc-sf@
1023: doc-sf!
1024: doc-df@
1025: doc-df!
1026:
1027: @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
1028: @subsection Address arithmetic
1029:
1030: ANS Forth does not specify the sizes of the data types. Instead, it
1031: offers a number of words for computing sizes and doing address
1032: arithmetic. Basically, address arithmetic is performed in terms of
1033: address units (aus); on most systems the address unit is one byte. Note
1034: that a character may have more than one au, so @code{chars} is no noop
1035: (on systems where it is a noop, it compiles to nothing).
1036:
1037: ANS Forth also defines words for aligning addresses for specific
1038: addresses. Many computers require that accesses to specific data types
1039: must only occur at specific addresses; e.g., that cells may only be
1040: accessed at addresses divisible by 4. Even if a machine allows unaligned
1041: accesses, it can usually perform aligned accesses faster.
1042:
1043: For the performance-conscious: alignment operations are usually only
1044: necessary during the definition of a data structure, not during the
1045: (more frequent) accesses to it.
1046:
1047: ANS Forth defines no words for character-aligning addresses. This is not
1048: an oversight, but reflects the fact that addresses that are not
1049: char-aligned have no use in the standard and therefore will not be
1050: created.
1051:
1052: The standard guarantees that addresses returned by @code{CREATE}d words
1053: are cell-aligned; in addition, Gforth guarantees that these addresses
1054: are aligned for all purposes.
1055:
1056: Note that the standard defines a word @code{char}, which has nothing to
1057: do with address arithmetic.
1058:
1059: doc-chars
1060: doc-char+
1061: doc-cells
1062: doc-cell+
1063: doc-align
1064: doc-aligned
1065: doc-floats
1066: doc-float+
1067: doc-falign
1068: doc-faligned
1069: doc-sfloats
1070: doc-sfloat+
1071: doc-sfalign
1072: doc-sfaligned
1073: doc-dfloats
1074: doc-dfloat+
1075: doc-dfalign
1076: doc-dfaligned
1077: doc-maxalign
1078: doc-maxaligned
1079: doc-cfalign
1080: doc-cfaligned
1081: doc-address-unit-bits
1082:
1083: @node Memory block access, , Address arithmetic, Memory access
1084: @subsection Memory block access
1085:
1086: doc-move
1087: doc-erase
1088:
1089: While the previous words work on address units, the rest works on
1090: characters.
1091:
1092: doc-cmove
1093: doc-cmove>
1094: doc-fill
1095: doc-blank
1096:
1097: @node Control Structures, Locals, Memory access, Words
1098: @section Control Structures
1099:
1100: Control structures in Forth cannot be used in interpret state, only in
1101: compile state, i.e., in a colon definition. We do not like this
1102: limitation, but have not seen a satisfying way around it yet, although
1103: many schemes have been proposed.
1104:
1105: @menu
1106: * Selection::
1107: * Simple Loops::
1108: * Counted Loops::
1109: * Arbitrary control structures::
1110: * Calls and returns::
1111: * Exception Handling::
1112: @end menu
1113:
1114: @node Selection, Simple Loops, Control Structures, Control Structures
1115: @subsection Selection
1116:
1117: @example
1118: @var{flag}
1119: IF
1120: @var{code}
1121: ENDIF
1122: @end example
1123: or
1124: @example
1125: @var{flag}
1126: IF
1127: @var{code1}
1128: ELSE
1129: @var{code2}
1130: ENDIF
1131: @end example
1132:
1133: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
1134: standard, and @code{ENDIF} is not, although it is quite popular. We
1135: recommend using @code{ENDIF}, because it is less confusing for people
1136: who also know other languages (and is not prone to reinforcing negative
1137: prejudices against Forth in these people). Adding @code{ENDIF} to a
1138: system that only supplies @code{THEN} is simple:
1139: @example
1140: : endif POSTPONE then ; immediate
1141: @end example
1142:
1143: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
1144: (adv.)} has the following meanings:
1145: @quotation
1146: ... 2b: following next after in order ... 3d: as a necessary consequence
1147: (if you were there, then you saw them).
1148: @end quotation
1149: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
1150: and many other programming languages has the meaning 3d.]
1151:
1152: Gforth also provides the words @code{?dup-if} and @code{?dup-0=-if}, so
1153: you can avoid using @code{?dup}. Using these alternatives is also more
1154: efficient than using @code{?dup}. Definitions in plain standard Forth
1155: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
1156: @file{compat/control.fs}.
1157:
1158: @example
1159: @var{n}
1160: CASE
1161: @var{n1} OF @var{code1} ENDOF
1162: @var{n2} OF @var{code2} ENDOF
1163: @dots{}
1164: ENDCASE
1165: @end example
1166:
1167: Executes the first @var{codei}, where the @var{ni} is equal to
1168: @var{n}. A default case can be added by simply writing the code after
1169: the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
1170: but must not consume it.
1171:
1172: @node Simple Loops, Counted Loops, Selection, Control Structures
1173: @subsection Simple Loops
1174:
1175: @example
1176: BEGIN
1177: @var{code1}
1178: @var{flag}
1179: WHILE
1180: @var{code2}
1181: REPEAT
1182: @end example
1183:
1184: @var{code1} is executed and @var{flag} is computed. If it is true,
1185: @var{code2} is executed and the loop is restarted; If @var{flag} is false, execution continues after the @code{REPEAT}.
1186:
1187: @example
1188: BEGIN
1189: @var{code}
1190: @var{flag}
1191: UNTIL
1192: @end example
1193:
1194: @var{code} is executed. The loop is restarted if @code{flag} is false.
1195:
1196: @example
1197: BEGIN
1198: @var{code}
1199: AGAIN
1200: @end example
1201:
1202: This is an endless loop.
1203:
1204: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
1205: @subsection Counted Loops
1206:
1207: The basic counted loop is:
1208: @example
1209: @var{limit} @var{start}
1210: ?DO
1211: @var{body}
1212: LOOP
1213: @end example
1214:
1215: This performs one iteration for every integer, starting from @var{start}
1216: and up to, but excluding @var{limit}. The counter, aka index, can be
1217: accessed with @code{i}. E.g., the loop
1218: @example
1219: 10 0 ?DO
1220: i .
1221: LOOP
1222: @end example
1223: prints
1224: @example
1225: 0 1 2 3 4 5 6 7 8 9
1226: @end example
1227: The index of the innermost loop can be accessed with @code{i}, the index
1228: of the next loop with @code{j}, and the index of the third loop with
1229: @code{k}.
1230:
1231: The loop control data are kept on the return stack, so there are some
1232: restrictions on mixing return stack accesses and counted loop
1233: words. E.g., if you put values on the return stack outside the loop, you
1234: cannot read them inside the loop. If you put values on the return stack
1235: within a loop, you have to remove them before the end of the loop and
1236: before accessing the index of the loop.
1237:
1238: There are several variations on the counted loop:
1239:
1240: @code{LEAVE} leaves the innermost counted loop immediately.
1241:
1242: If @var{start} is greater than @var{limit}, a @code{?DO} loop is entered
1243: (and @code{LOOP} iterates until they become equal by wrap-around
1244: arithmetic). This behaviour is usually not what you want. Therefore,
1245: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1246: @code{?DO}), which do not enter the loop if @var{start} is greater than
1247: @var{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1248: unsigned loop parameters.
1249:
1250: @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the
1251: index by @var{n} instead of by 1. The loop is terminated when the border
1252: between @var{limit-1} and @var{limit} is crossed. E.g.:
1253:
1254: @code{4 0 +DO i . 2 +LOOP} prints @code{0 2}
1255:
1256: @code{4 1 +DO i . 2 +LOOP} prints @code{1 3}
1257:
1258: The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:
1259:
1260: @code{-1 0 ?DO i . -1 +LOOP} prints @code{0 -1}
1261:
1262: @code{ 0 0 ?DO i . -1 +LOOP} prints nothing
1263:
1264: Therefore we recommend avoiding @code{@var{n} +LOOP} with negative
1265: @var{n}. One alternative is @code{@var{u} -LOOP}, which reduces the
1266: index by @var{u} each iteration. The loop is terminated when the border
1267: between @var{limit+1} and @var{limit} is crossed. Gforth also provides
1268: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
1269:
1270: @code{-2 0 -DO i . 1 -LOOP} prints @code{0 -1}
1271:
1272: @code{-1 0 -DO i . 1 -LOOP} prints @code{0}
1273:
1274: @code{ 0 0 -DO i . 1 -LOOP} prints nothing
1275:
1276: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1277: @code{-LOOP} are not in the ANS Forth standard. However, an
1278: implementation for these words that uses only standard words is provided
1279: in @file{compat/loops.fs}.
1280:
1281: @code{?DO} can also be replaced by @code{DO}. @code{DO} always enters
1282: the loop, independent of the loop parameters. Do not use @code{DO}, even
1283: if you know that the loop is entered in any case. Such knowledge tends
1284: to become invalid during maintenance of a program, and then the
1285: @code{DO} will make trouble.
1286:
1287: @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via
1288: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
1289: return stack so @code{EXIT} can get to its return address.
1290:
1291: Another counted loop is
1292: @example
1293: @var{n}
1294: FOR
1295: @var{body}
1296: NEXT
1297: @end example
1298: This is the preferred loop of native code compiler writers who are too
1299: lazy to optimize @code{?DO} loops properly. In Gforth, this loop
1300: iterates @var{n+1} times; @code{i} produces values starting with @var{n}
1301: and ending with 0. Other Forth systems may behave differently, even if
1302: they support @code{FOR} loops. To avoid problems, don't use @code{FOR}
1303: loops.
1304:
1305: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
1306: @subsection Arbitrary control structures
1307:
1308: ANS Forth permits and supports using control structures in a non-nested
1309: way. Information about incomplete control structures is stored on the
1310: control-flow stack. This stack may be implemented on the Forth data
1311: stack, and this is what we have done in Gforth.
1312:
1313: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
1314: entry represents a backward branch target. A few words are the basis for
1315: building any control structure possible (except control structures that
1316: need storage, like calls, coroutines, and backtracking).
1317:
1318: doc-if
1319: doc-ahead
1320: doc-then
1321: doc-begin
1322: doc-until
1323: doc-again
1324: doc-cs-pick
1325: doc-cs-roll
1326:
1327: On many systems control-flow stack items take one word, in Gforth they
1328: currently take three (this may change in the future). Therefore it is a
1329: really good idea to manipulate the control flow stack with
1330: @code{cs-pick} and @code{cs-roll}, not with data stack manipulation
1331: words.
1332:
1333: Some standard control structure words are built from these words:
1334:
1335: doc-else
1336: doc-while
1337: doc-repeat
1338:
1339: Gforth adds some more control-structure words:
1340:
1341: doc-endif
1342: doc-?dup-if
1343: doc-?dup-0=-if
1344:
1345: Counted loop words constitute a separate group of words:
1346:
1347: doc-?do
1348: doc-+do
1349: doc-u+do
1350: doc--do
1351: doc-u-do
1352: doc-do
1353: doc-for
1354: doc-loop
1355: doc-+loop
1356: doc--loop
1357: doc-next
1358: doc-leave
1359: doc-?leave
1360: doc-unloop
1361: doc-done
1362:
1363: The standard does not allow using @code{cs-pick} and @code{cs-roll} on
1364: @i{do-sys}. Our system allows it, but it's your job to ensure that for
1365: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
1366: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
1367: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
1368: resolved (by using one of the loop-ending words or @code{DONE}).
1369:
1370: Another group of control structure words are
1371:
1372: doc-case
1373: doc-endcase
1374: doc-of
1375: doc-endof
1376:
1377: @i{case-sys} and @i{of-sys} cannot be processed using @code{cs-pick} and
1378: @code{cs-roll}.
1379:
1380: @subsubsection Programming Style
1381:
1382: In order to ensure readability we recommend that you do not create
1383: arbitrary control structures directly, but define new control structure
1384: words for the control structure you want and use these words in your
1385: program.
1386:
1387: E.g., instead of writing
1388:
1389: @example
1390: begin
1391: ...
1392: if [ 1 cs-roll ]
1393: ...
1394: again then
1395: @end example
1396:
1397: we recommend defining control structure words, e.g.,
1398:
1399: @example
1400: : while ( dest -- orig dest )
1401: POSTPONE if
1402: 1 cs-roll ; immediate
1403:
1404: : repeat ( orig dest -- )
1405: POSTPONE again
1406: POSTPONE then ; immediate
1407: @end example
1408:
1409: and then using these to create the control structure:
1410:
1411: @example
1412: begin
1413: ...
1414: while
1415: ...
1416: repeat
1417: @end example
1418:
1419: That's much easier to read, isn't it? Of course, @code{REPEAT} and
1420: @code{WHILE} are predefined, so in this example it would not be
1421: necessary to define them.
1422:
1423: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
1424: @subsection Calls and returns
1425:
1426: A definition can be called simply be writing the name of the
1427: definition. When the end of the definition is reached, it returns. An
1428: earlier return can be forced using
1429:
1430: doc-exit
1431:
1432: Don't forget to clean up the return stack and @code{UNLOOP} any
1433: outstanding @code{?DO}...@code{LOOP}s before @code{EXIT}ing. The
1434: primitive compiled by @code{EXIT} is
1435:
1436: doc-;s
1437:
1438: @node Exception Handling, , Calls and returns, Control Structures
1439: @subsection Exception Handling
1440:
1441: doc-catch
1442: doc-throw
1443:
1444: @node Locals, Defining Words, Control Structures, Words
1445: @section Locals
1446:
1447: Local variables can make Forth programming more enjoyable and Forth
1448: programs easier to read. Unfortunately, the locals of ANS Forth are
1449: laden with restrictions. Therefore, we provide not only the ANS Forth
1450: locals wordset, but also our own, more powerful locals wordset (we
1451: implemented the ANS Forth locals wordset through our locals wordset).
1452:
1453: The ideas in this section have also been published in the paper
1454: @cite{Automatic Scoping of Local Variables} by M. Anton Ertl, presented
1455: at EuroForth '94; it is available at
1456: @*@file{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz}.
1457:
1458: @menu
1459: * Gforth locals::
1460: * ANS Forth locals::
1461: @end menu
1462:
1463: @node Gforth locals, ANS Forth locals, Locals, Locals
1464: @subsection Gforth locals
1465:
1466: Locals can be defined with
1467:
1468: @example
1469: @{ local1 local2 ... -- comment @}
1470: @end example
1471: or
1472: @example
1473: @{ local1 local2 ... @}
1474: @end example
1475:
1476: E.g.,
1477: @example
1478: : max @{ n1 n2 -- n3 @}
1479: n1 n2 > if
1480: n1
1481: else
1482: n2
1483: endif ;
1484: @end example
1485:
1486: The similarity of locals definitions with stack comments is intended. A
1487: locals definition often replaces the stack comment of a word. The order
1488: of the locals corresponds to the order in a stack comment and everything
1489: after the @code{--} is really a comment.
1490:
1491: This similarity has one disadvantage: It is too easy to confuse locals
1492: declarations with stack comments, causing bugs and making them hard to
1493: find. However, this problem can be avoided by appropriate coding
1494: conventions: Do not use both notations in the same program. If you do,
1495: they should be distinguished using additional means, e.g. by position.
1496:
1497: The name of the local may be preceded by a type specifier, e.g.,
1498: @code{F:} for a floating point value:
1499:
1500: @example
1501: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
1502: \ complex multiplication
1503: Ar Br f* Ai Bi f* f-
1504: Ar Bi f* Ai Br f* f+ ;
1505: @end example
1506:
1507: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
1508: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
1509: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
1510: with @code{W:}, @code{D:} etc.) produces its value and can be changed
1511: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
1512: produces its address (which becomes invalid when the variable's scope is
1513: left). E.g., the standard word @code{emit} can be defined in therms of
1514: @code{type} like this:
1515:
1516: @example
1517: : emit @{ C^ char* -- @}
1518: char* 1 type ;
1519: @end example
1520:
1521: A local without type specifier is a @code{W:} local. Both flavours of
1522: locals are initialized with values from the data or FP stack.
1523:
1524: Currently there is no way to define locals with user-defined data
1525: structures, but we are working on it.
1526:
1527: Gforth allows defining locals everywhere in a colon definition. This
1528: poses the following questions:
1529:
1530: @menu
1531: * Where are locals visible by name?::
1532: * How long do locals live?::
1533: * Programming Style::
1534: * Implementation::
1535: @end menu
1536:
1537: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
1538: @subsubsection Where are locals visible by name?
1539:
1540: Basically, the answer is that locals are visible where you would expect
1541: it in block-structured languages, and sometimes a little longer. If you
1542: want to restrict the scope of a local, enclose its definition in
1543: @code{SCOPE}...@code{ENDSCOPE}.
1544:
1545: doc-scope
1546: doc-endscope
1547:
1548: These words behave like control structure words, so you can use them
1549: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
1550: arbitrary ways.
1551:
1552: If you want a more exact answer to the visibility question, here's the
1553: basic principle: A local is visible in all places that can only be
1554: reached through the definition of the local@footnote{In compiler
1555: construction terminology, all places dominated by the definition of the
1556: local.}. In other words, it is not visible in places that can be reached
1557: without going through the definition of the local. E.g., locals defined
1558: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
1559: defined in @code{BEGIN}...@code{UNTIL} are visible after the
1560: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1561:
1562: The reasoning behind this solution is: We want to have the locals
1563: visible as long as it is meaningful. The user can always make the
1564: visibility shorter by using explicit scoping. In a place that can
1565: only be reached through the definition of a local, the meaning of a
1566: local name is clear. In other places it is not: How is the local
1567: initialized at the control flow path that does not contain the
1568: definition? Which local is meant, if the same name is defined twice in
1569: two independent control flow paths?
1570:
1571: This should be enough detail for nearly all users, so you can skip the
1572: rest of this section. If you relly must know all the gory details and
1573: options, read on.
1574:
1575: In order to implement this rule, the compiler has to know which places
1576: are unreachable. It knows this automatically after @code{AHEAD},
1577: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
1578: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
1579: compiler that the control flow never reaches that place. If
1580: @code{UNREACHABLE} is not used where it could, the only consequence is
1581: that the visibility of some locals is more limited than the rule above
1582: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
1583: lie to the compiler), buggy code will be produced.
1584:
1585: Another problem with this rule is that at @code{BEGIN}, the compiler
1586: does not know which locals will be visible on the incoming
1587: back-edge. All problems discussed in the following are due to this
1588: ignorance of the compiler (we discuss the problems using @code{BEGIN}
1589: loops as examples; the discussion also applies to @code{?DO} and other
1590: loops). Perhaps the most insidious example is:
1591: @example
1592: AHEAD
1593: BEGIN
1594: x
1595: [ 1 CS-ROLL ] THEN
1596: @{ x @}
1597: ...
1598: UNTIL
1599: @end example
1600:
1601: This should be legal according to the visibility rule. The use of
1602: @code{x} can only be reached through the definition; but that appears
1603: textually below the use.
1604:
1605: From this example it is clear that the visibility rules cannot be fully
1606: implemented without major headaches. Our implementation treats common
1607: cases as advertised and the exceptions are treated in a safe way: The
1608: compiler makes a reasonable guess about the locals visible after a
1609: @code{BEGIN}; if it is too pessimistic, the
1610: user will get a spurious error about the local not being defined; if the
1611: compiler is too optimistic, it will notice this later and issue a
1612: warning. In the case above the compiler would complain about @code{x}
1613: being undefined at its use. You can see from the obscure examples in
1614: this section that it takes quite unusual control structures to get the
1615: compiler into trouble, and even then it will often do fine.
1616:
1617: If the @code{BEGIN} is reachable from above, the most optimistic guess
1618: is that all locals visible before the @code{BEGIN} will also be
1619: visible after the @code{BEGIN}. This guess is valid for all loops that
1620: are entered only through the @code{BEGIN}, in particular, for normal
1621: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
1622: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
1623: compiler. When the branch to the @code{BEGIN} is finally generated by
1624: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
1625: warns the user if it was too optimisitic:
1626: @example
1627: IF
1628: @{ x @}
1629: BEGIN
1630: \ x ?
1631: [ 1 cs-roll ] THEN
1632: ...
1633: UNTIL
1634: @end example
1635:
1636: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
1637: optimistically assumes that it lives until the @code{THEN}. It notices
1638: this difference when it compiles the @code{UNTIL} and issues a
1639: warning. The user can avoid the warning, and make sure that @code{x}
1640: is not used in the wrong area by using explicit scoping:
1641: @example
1642: IF
1643: SCOPE
1644: @{ x @}
1645: ENDSCOPE
1646: BEGIN
1647: [ 1 cs-roll ] THEN
1648: ...
1649: UNTIL
1650: @end example
1651:
1652: Since the guess is optimistic, there will be no spurious error messages
1653: about undefined locals.
1654:
1655: If the @code{BEGIN} is not reachable from above (e.g., after
1656: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
1657: optimistic guess, as the locals visible after the @code{BEGIN} may be
1658: defined later. Therefore, the compiler assumes that no locals are
1659: visible after the @code{BEGIN}. However, the user can use
1660: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
1661: visible at the BEGIN as at the point where the top control-flow stack
1662: item was created.
1663:
1664: doc-assume-live
1665:
1666: E.g.,
1667: @example
1668: @{ x @}
1669: AHEAD
1670: ASSUME-LIVE
1671: BEGIN
1672: x
1673: [ 1 CS-ROLL ] THEN
1674: ...
1675: UNTIL
1676: @end example
1677:
1678: Other cases where the locals are defined before the @code{BEGIN} can be
1679: handled by inserting an appropriate @code{CS-ROLL} before the
1680: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
1681: behind the @code{ASSUME-LIVE}).
1682:
1683: Cases where locals are defined after the @code{BEGIN} (but should be
1684: visible immediately after the @code{BEGIN}) can only be handled by
1685: rearranging the loop. E.g., the ``most insidious'' example above can be
1686: arranged into:
1687: @example
1688: BEGIN
1689: @{ x @}
1690: ... 0=
1691: WHILE
1692: x
1693: REPEAT
1694: @end example
1695:
1696: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
1697: @subsubsection How long do locals live?
1698:
1699: The right answer for the lifetime question would be: A local lives at
1700: least as long as it can be accessed. For a value-flavoured local this
1701: means: until the end of its visibility. However, a variable-flavoured
1702: local could be accessed through its address far beyond its visibility
1703: scope. Ultimately, this would mean that such locals would have to be
1704: garbage collected. Since this entails un-Forth-like implementation
1705: complexities, I adopted the same cowardly solution as some other
1706: languages (e.g., C): The local lives only as long as it is visible;
1707: afterwards its address is invalid (and programs that access it
1708: afterwards are erroneous).
1709:
1710: @node Programming Style, Implementation, How long do locals live?, Gforth locals
1711: @subsubsection Programming Style
1712:
1713: The freedom to define locals anywhere has the potential to change
1714: programming styles dramatically. In particular, the need to use the
1715: return stack for intermediate storage vanishes. Moreover, all stack
1716: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
1717: determined arguments) can be eliminated: If the stack items are in the
1718: wrong order, just write a locals definition for all of them; then
1719: write the items in the order you want.
1720:
1721: This seems a little far-fetched and eliminating stack manipulations is
1722: unlikely to become a conscious programming objective. Still, the number
1723: of stack manipulations will be reduced dramatically if local variables
1724: are used liberally (e.g., compare @code{max} in @ref{Gforth locals} with
1725: a traditional implementation of @code{max}).
1726:
1727: This shows one potential benefit of locals: making Forth programs more
1728: readable. Of course, this benefit will only be realized if the
1729: programmers continue to honour the principle of factoring instead of
1730: using the added latitude to make the words longer.
1731:
1732: Using @code{TO} can and should be avoided. Without @code{TO},
1733: every value-flavoured local has only a single assignment and many
1734: advantages of functional languages apply to Forth. I.e., programs are
1735: easier to analyse, to optimize and to read: It is clear from the
1736: definition what the local stands for, it does not turn into something
1737: different later.
1738:
1739: E.g., a definition using @code{TO} might look like this:
1740: @example
1741: : strcmp @{ addr1 u1 addr2 u2 -- n @}
1742: u1 u2 min 0
1743: ?do
1744: addr1 c@@ addr2 c@@ -
1745: ?dup-if
1746: unloop exit
1747: then
1748: addr1 char+ TO addr1
1749: addr2 char+ TO addr2
1750: loop
1751: u1 u2 - ;
1752: @end example
1753: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
1754: every loop iteration. @code{strcmp} is a typical example of the
1755: readability problems of using @code{TO}. When you start reading
1756: @code{strcmp}, you think that @code{addr1} refers to the start of the
1757: string. Only near the end of the loop you realize that it is something
1758: else.
1759:
1760: This can be avoided by defining two locals at the start of the loop that
1761: are initialized with the right value for the current iteration.
1762: @example
1763: : strcmp @{ addr1 u1 addr2 u2 -- n @}
1764: addr1 addr2
1765: u1 u2 min 0
1766: ?do @{ s1 s2 @}
1767: s1 c@@ s2 c@@ -
1768: ?dup-if
1769: unloop exit
1770: then
1771: s1 char+ s2 char+
1772: loop
1773: 2drop
1774: u1 u2 - ;
1775: @end example
1776: Here it is clear from the start that @code{s1} has a different value
1777: in every loop iteration.
1778:
1779: @node Implementation, , Programming Style, Gforth locals
1780: @subsubsection Implementation
1781:
1782: Gforth uses an extra locals stack. The most compelling reason for
1783: this is that the return stack is not float-aligned; using an extra stack
1784: also eliminates the problems and restrictions of using the return stack
1785: as locals stack. Like the other stacks, the locals stack grows toward
1786: lower addresses. A few primitives allow an efficient implementation:
1787:
1788: doc-@local#
1789: doc-f@local#
1790: doc-laddr#
1791: doc-lp+!#
1792: doc-lp!
1793: doc->l
1794: doc-f>l
1795:
1796: In addition to these primitives, some specializations of these
1797: primitives for commonly occurring inline arguments are provided for
1798: efficiency reasons, e.g., @code{@@local0} as specialization of
1799: @code{@@local#} for the inline argument 0. The following compiling words
1800: compile the right specialized version, or the general version, as
1801: appropriate:
1802:
1803: doc-compile-@local
1804: doc-compile-f@local
1805: doc-compile-lp+!
1806:
1807: Combinations of conditional branches and @code{lp+!#} like
1808: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
1809: is taken) are provided for efficiency and correctness in loops.
1810:
1811: A special area in the dictionary space is reserved for keeping the
1812: local variable names. @code{@{} switches the dictionary pointer to this
1813: area and @code{@}} switches it back and generates the locals
1814: initializing code. @code{W:} etc.@ are normal defining words. This
1815: special area is cleared at the start of every colon definition.
1816:
1817: A special feature of Gforth's dictionary is used to implement the
1818: definition of locals without type specifiers: every wordlist (aka
1819: vocabulary) has its own methods for searching
1820: etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
1821: with a special search method: When it is searched for a word, it
1822: actually creates that word using @code{W:}. @code{@{} changes the search
1823: order to first search the wordlist containing @code{@}}, @code{W:} etc.,
1824: and then the wordlist for defining locals without type specifiers.
1825:
1826: The lifetime rules support a stack discipline within a colon
1827: definition: The lifetime of a local is either nested with other locals
1828: lifetimes or it does not overlap them.
1829:
1830: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
1831: pointer manipulation is generated. Between control structure words
1832: locals definitions can push locals onto the locals stack. @code{AGAIN}
1833: is the simplest of the other three control flow words. It has to
1834: restore the locals stack depth of the corresponding @code{BEGIN}
1835: before branching. The code looks like this:
1836: @format
1837: @code{lp+!#} current-locals-size @minus{} dest-locals-size
1838: @code{branch} <begin>
1839: @end format
1840:
1841: @code{UNTIL} is a little more complicated: If it branches back, it
1842: must adjust the stack just like @code{AGAIN}. But if it falls through,
1843: the locals stack must not be changed. The compiler generates the
1844: following code:
1845: @format
1846: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
1847: @end format
1848: The locals stack pointer is only adjusted if the branch is taken.
1849:
1850: @code{THEN} can produce somewhat inefficient code:
1851: @format
1852: @code{lp+!#} current-locals-size @minus{} orig-locals-size
1853: <orig target>:
1854: @code{lp+!#} orig-locals-size @minus{} new-locals-size
1855: @end format
1856: The second @code{lp+!#} adjusts the locals stack pointer from the
1857: level at the @var{orig} point to the level after the @code{THEN}. The
1858: first @code{lp+!#} adjusts the locals stack pointer from the current
1859: level to the level at the orig point, so the complete effect is an
1860: adjustment from the current level to the right level after the
1861: @code{THEN}.
1862:
1863: In a conventional Forth implementation a dest control-flow stack entry
1864: is just the target address and an orig entry is just the address to be
1865: patched. Our locals implementation adds a wordlist to every orig or dest
1866: item. It is the list of locals visible (or assumed visible) at the point
1867: described by the entry. Our implementation also adds a tag to identify
1868: the kind of entry, in particular to differentiate between live and dead
1869: (reachable and unreachable) orig entries.
1870:
1871: A few unusual operations have to be performed on locals wordlists:
1872:
1873: doc-common-list
1874: doc-sub-list?
1875: doc-list-size
1876:
1877: Several features of our locals wordlist implementation make these
1878: operations easy to implement: The locals wordlists are organised as
1879: linked lists; the tails of these lists are shared, if the lists
1880: contain some of the same locals; and the address of a name is greater
1881: than the address of the names behind it in the list.
1882:
1883: Another important implementation detail is the variable
1884: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
1885: determine if they can be reached directly or only through the branch
1886: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
1887: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
1888: definition, by @code{BEGIN} and usually by @code{THEN}.
1889:
1890: Counted loops are similar to other loops in most respects, but
1891: @code{LEAVE} requires special attention: It performs basically the same
1892: service as @code{AHEAD}, but it does not create a control-flow stack
1893: entry. Therefore the information has to be stored elsewhere;
1894: traditionally, the information was stored in the target fields of the
1895: branches created by the @code{LEAVE}s, by organizing these fields into a
1896: linked list. Unfortunately, this clever trick does not provide enough
1897: space for storing our extended control flow information. Therefore, we
1898: introduce another stack, the leave stack. It contains the control-flow
1899: stack entries for all unresolved @code{LEAVE}s.
1900:
1901: Local names are kept until the end of the colon definition, even if
1902: they are no longer visible in any control-flow path. In a few cases
1903: this may lead to increased space needs for the locals name area, but
1904: usually less than reclaiming this space would cost in code size.
1905:
1906:
1907: @node ANS Forth locals, , Gforth locals, Locals
1908: @subsection ANS Forth locals
1909:
1910: The ANS Forth locals wordset does not define a syntax for locals, but
1911: words that make it possible to define various syntaxes. One of the
1912: possible syntaxes is a subset of the syntax we used in the Gforth locals
1913: wordset, i.e.:
1914:
1915: @example
1916: @{ local1 local2 ... -- comment @}
1917: @end example
1918: or
1919: @example
1920: @{ local1 local2 ... @}
1921: @end example
1922:
1923: The order of the locals corresponds to the order in a stack comment. The
1924: restrictions are:
1925:
1926: @itemize @bullet
1927: @item
1928: Locals can only be cell-sized values (no type specifiers are allowed).
1929: @item
1930: Locals can be defined only outside control structures.
1931: @item
1932: Locals can interfere with explicit usage of the return stack. For the
1933: exact (and long) rules, see the standard. If you don't use return stack
1934: accessing words in a definition using locals, you will be all right. The
1935: purpose of this rule is to make locals implementation on the return
1936: stack easier.
1937: @item
1938: The whole definition must be in one line.
1939: @end itemize
1940:
1941: Locals defined in this way behave like @code{VALUE}s (@xref{Simple
1942: Defining Words}). I.e., they are initialized from the stack. Using their
1943: name produces their value. Their value can be changed using @code{TO}.
1944:
1945: Since this syntax is supported by Gforth directly, you need not do
1946: anything to use it. If you want to port a program using this syntax to
1947: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
1948: syntax on the other system.
1949:
1950: Note that a syntax shown in the standard, section A.13 looks
1951: similar, but is quite different in having the order of locals
1952: reversed. Beware!
1953:
1954: The ANS Forth locals wordset itself consists of the following word
1955:
1956: doc-(local)
1957:
1958: The ANS Forth locals extension wordset defines a syntax, but it is so
1959: awful that we strongly recommend not to use it. We have implemented this
1960: syntax to make porting to Gforth easy, but do not document it here. The
1961: problem with this syntax is that the locals are defined in an order
1962: reversed with respect to the standard stack comment notation, making
1963: programs harder to read, and easier to misread and miswrite. The only
1964: merit of this syntax is that it is easy to implement using the ANS Forth
1965: locals wordset.
1966:
1967: @node Defining Words, Tokens for Words, Locals, Words
1968: @section Defining Words
1969:
1970: @menu
1971: * Simple Defining Words::
1972: * Colon Definitions::
1973: * User-defined Defining Words::
1974: * Supplying names::
1975: * Interpretation and Compilation Semantics::
1976: @end menu
1977:
1978: @node Simple Defining Words, Colon Definitions, Defining Words, Defining Words
1979: @subsection Simple Defining Words
1980:
1981: doc-constant
1982: doc-2constant
1983: doc-fconstant
1984: doc-variable
1985: doc-2variable
1986: doc-fvariable
1987: doc-create
1988: doc-user
1989: doc-value
1990: doc-to
1991: doc-defer
1992: doc-is
1993:
1994: @node Colon Definitions, User-defined Defining Words, Simple Defining Words, Defining Words
1995: @subsection Colon Definitions
1996:
1997: @example
1998: : name ( ... -- ... )
1999: word1 word2 word3 ;
2000: @end example
2001:
2002: creates a word called @code{name}, that, upon execution, executes
2003: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
2004:
2005: The explanation above is somewhat superficial. @xref{Interpretation and
2006: Compilation Semantics} for an in-depth discussion of some of the issues
2007: involved.
2008:
2009: doc-:
2010: doc-;
2011:
2012: @node User-defined Defining Words, Supplying names, Colon Definitions, Defining Words
2013: @subsection User-defined Defining Words
2014:
2015: You can create new defining words simply by wrapping defining-time code
2016: around existing defining words and putting the sequence in a colon
2017: definition.
2018:
2019: If you want the words defined with your defining words to behave
2020: differently from words defined with standard defining words, you can
2021: write your defining word like this:
2022:
2023: @example
2024: : def-word ( "name" -- )
2025: Create @var{code1}
2026: DOES> ( ... -- ... )
2027: @var{code2} ;
2028:
2029: def-word name
2030: @end example
2031:
2032: Technically, this fragment defines a defining word @code{def-word}, and
2033: a word @code{name}; when you execute @code{name}, the address of the
2034: body of @code{name} is put on the data stack and @var{code2} is executed
2035: (the address of the body of @code{name} is the address @code{HERE}
2036: returns immediately after the @code{CREATE}).
2037:
2038: In other words, if you make the following definitions:
2039:
2040: @example
2041: : def-word1 ( "name" -- )
2042: Create @var{code1} ;
2043:
2044: : action1 ( ... -- ... )
2045: @var{code2} ;
2046:
2047: def-word name1
2048: @end example
2049:
2050: Using @code{name1 action1} is equivalent to using @code{name}.
2051:
2052: E.g., you can implement @code{Constant} in this way:
2053:
2054: @example
2055: : constant ( w "name" -- )
2056: create ,
2057: DOES> ( -- w )
2058: @@ ;
2059: @end example
2060:
2061: When you create a constant with @code{5 constant five}, first a new word
2062: @code{five} is created, then the value 5 is laid down in the body of
2063: @code{five} with @code{,}. When @code{five} is invoked, the address of
2064: the body is put on the stack, and @code{@@} retrieves the value 5.
2065:
2066: In the example above the stack comment after the @code{DOES>} specifies
2067: the stack effect of the defined words, not the stack effect of the
2068: following code (the following code expects the address of the body on
2069: the top of stack, which is not reflected in the stack comment). This is
2070: the convention that I use and recommend (it clashes a bit with using
2071: locals declarations for stack effect specification, though).
2072:
2073: @subsubsection Applications of @code{CREATE..DOES>}
2074:
2075: You may wonder how to use this feature. Here are some usage patterns:
2076:
2077: When you see a sequence of code occurring several times, and you can
2078: identify a meaning, you will factor it out as a colon definition. When
2079: you see similar colon definitions, you can factor them using
2080: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
2081: that look very similar:
2082: @example
2083: : ori, ( reg-taget reg-source n -- )
2084: 0 asm-reg-reg-imm ;
2085: : andi, ( reg-taget reg-source n -- )
2086: 1 asm-reg-reg-imm ;
2087: @end example
2088:
2089: This could be factored with:
2090: @example
2091: : reg-reg-imm ( op-code -- )
2092: create ,
2093: DOES> ( reg-taget reg-source n -- )
2094: @@ asm-reg-reg-imm ;
2095:
2096: 0 reg-reg-imm ori,
2097: 1 reg-reg-imm andi,
2098: @end example
2099:
2100: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
2101: supply a part of the parameters for a word (known as @dfn{currying} in
2102: the functional language community). E.g., @code{+} needs two
2103: parameters. Creating versions of @code{+} with one parameter fixed can
2104: be done like this:
2105: @example
2106: : curry+ ( n1 -- )
2107: create ,
2108: DOES> ( n2 -- n1+n2 )
2109: @@ + ;
2110:
2111: 3 curry+ 3+
2112: -2 curry+ 2-
2113: @end example
2114:
2115: @subsubsection The gory details of @code{CREATE..DOES>}
2116:
2117: doc-does>
2118:
2119: This means that you need not use @code{CREATE} and @code{DOES>} in the
2120: same definition; E.g., you can put the @code{DOES>}-part in a separate
2121: definition. This allows us to, e.g., select among different DOES>-parts:
2122: @example
2123: : does1
2124: DOES> ( ... -- ... )
2125: ... ;
2126:
2127: : does2
2128: DOES> ( ... -- ... )
2129: ... ;
2130:
2131: : def-word ( ... -- ... )
2132: create ...
2133: IF
2134: does1
2135: ELSE
2136: does2
2137: ENDIF ;
2138: @end example
2139:
2140: In a standard program you can apply a @code{DOES>}-part only if the last
2141: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
2142: will override the behaviour of the last word defined in any case. In a
2143: standard program, you can use @code{DOES>} only in a colon
2144: definition. In Gforth, you can also use it in interpretation state, in a
2145: kind of one-shot mode:
2146: @example
2147: CREATE name ( ... -- ... )
2148: @var{initialization}
2149: DOES>
2150: @var{code} ;
2151: @end example
2152: This is equivalwent to the standard
2153: @example
2154: :noname
2155: DOES>
2156: @var{code} ;
2157: CREATE name EXECUTE ( ... -- ... )
2158: @var{initialization}
2159: @end example
2160:
2161: You can get the address of the body of a word with
2162:
2163: doc->body
2164:
2165: @node Supplying names, Interpretation and Compilation Semantics, User-defined Defining Words, Defining Words
2166: @subsection Supplying names for the defined words
2167:
2168: By default, defining words take the names for the defined words from the
2169: input stream. Sometimes you want to supply the name from a string. You
2170: can do this with
2171:
2172: doc-nextname
2173:
2174: E.g.,
2175:
2176: @example
2177: s" foo" nextname create
2178: @end example
2179: is equivalent to
2180: @example
2181: create foo
2182: @end example
2183:
2184: Sometimes you want to define a word without a name. You can do this with
2185:
2186: doc-noname
2187:
2188: To make any use of the newly defined word, you need its execution
2189: token. You can get it with
2190:
2191: doc-lastxt
2192:
2193: E.g., you can initialize a deferred word with an anonymous colon
2194: definition:
2195: @example
2196: Defer deferred
2197: noname : ( ... -- ... )
2198: ... ;
2199: lastxt IS deferred
2200: @end example
2201:
2202: @code{lastxt} also works when the last word was not defined as
2203: @code{noname}.
2204:
2205: The standard has also recognized the need for anonymous words and
2206: provides
2207:
2208: doc-:noname
2209:
2210: This leaves the execution token for the word on the stack after the
2211: closing @code{;}. You can rewrite the last example with @code{:noname}:
2212: @example
2213: Defer deferred
2214: :noname ( ... -- ... )
2215: ... ;
2216: IS deferred
2217: @end example
2218:
2219: @node Interpretation and Compilation Semantics, , Supplying names, Defining Words
2220: @subsection Interpretation and Compilation Semantics
2221:
2222: The @dfn{interpretation semantics} of a word are what the text
2223: interpreter does when it encounters the word in interpret state. It also
2224: appears in some other contexts, e.g., the execution token returned by
2225: @code{' @var{word}} identifies the interpretation semantics of
2226: @var{word} (in other words, @code{' @var{word} execute} is equivalent to
2227: interpret-state text interpretation of @code{@var{word}}).
2228:
2229: The @dfn{compilation semantics} of a word are what the text interpreter
2230: does when it encounters the word in compile state. It also appears in
2231: other contexts, e.g, @code{POSTPONE @var{word}} compiles@footnote{In
2232: standard terminology, ``appends to the current definition''.} the
2233: compilation semantics of @var{word}.
2234:
2235: The standard also talks about @dfn{execution semantics}. They are used
2236: only for defining the interpretation and compilation semantics of many
2237: words. By default, the interpretation semantics of a word are to
2238: @code{execute} its execution semantics, and the compilation semantics of
2239: a word are to @code{compile,} its execution semantics.@footnote{In
2240: standard terminology: The default interpretation semantics are its
2241: execution semantics; the default compilation semantics are to append its
2242: execution semantics to the execution semantics of the current
2243: definition.}
2244:
2245: You can change the compilation semantics into @code{execute}ing the
2246: execution semantics with
2247:
2248: doc-immediate
2249:
2250: You can remove the interpretation semantics of a word with
2251:
2252: doc-compile-only
2253: doc-restrict
2254:
2255: Note that ticking (@code{'}) compile-only words gives an error
2256: (``Interpreting a compile-only word'').
2257:
2258: Gforth also allows you to define words with arbitrary combinations of
2259: interpretation and compilation semantics.
2260:
2261: doc-interpret/compile:
2262:
2263: This feature was introduced for implementing @code{TO} and @code{S"}. I
2264: recommend that you do not define such words, as cute as they may be:
2265: they make it hard to get at both parts of the word in some contexts.
2266: E.g., assume you want to get an execution token for the compilation
2267: part. Instead, define two words, one that embodies the interpretation
2268: part, and one that embodies the compilation part.
2269:
2270: There is, however, a potentially useful application of this feature:
2271: Providing differing implementations for the default semantics. While
2272: this introduces redundancy and is therefore usually a bad idea, a
2273: performance improvement may be worth the trouble. E.g., consider the
2274: word @code{foobar}:
2275:
2276: @example
2277: : foobar
2278: foo bar ;
2279: @end example
2280:
2281: Let us assume that @code{foobar} is called so frequently that the
2282: calling overhead would take a significant amount of the run-time. We can
2283: optimize it with @code{interpret/compile:}:
2284:
2285: @example
2286: :noname
2287: foo bar ;
2288: :noname
2289: POSTPONE foo POSTPONE bar ;
2290: interpret/compile: foobar
2291: @end example
2292:
2293: This definition has the same interpretation semantics and essentially
2294: the same compilation semantics as the simple definition of
2295: @code{foobar}, but the implementation of the compilation semantics is
2296: more efficient with respect to run-time.
2297:
2298: Some people try to use state-smart words to emulate the feature provided
2299: by @code{interpret/compile:} (words are state-smart if they check
2300: @code{STATE} during execution). E.g., they would try to code
2301: @code{foobar} like this:
2302:
2303: @example
2304: : foobar
2305: STATE @@
2306: IF ( compilation state )
2307: POSTPONE foo POSTPONE bar
2308: ELSE
2309: foo bar
2310: ENDIF ; immediate
2311: @end example
2312:
2313: While this works if @code{foobar} is processed only by the text
2314: interpreter, it does not work in other contexts (like @code{'} or
2315: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
2316: for a state-smart word, not for the interpretation semantics of the
2317: original @code{foobar}; when you execute this execution token (directly
2318: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
2319: state, the result will not be what you expected (i.e., it will not
2320: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
2321: write them!
2322:
2323: It is also possible to write defining words that define words with
2324: arbitrary combinations of interpretation and compilation semantics (or,
2325: preferably, arbitrary combinations of implementations of the default
2326: semantics). In general, this looks like:
2327:
2328: @example
2329: : def-word
2330: create-interpret/compile
2331: @var{code1}
2332: interpretation>
2333: @var{code2}
2334: <interpretation
2335: compilation>
2336: @var{code3}
2337: <compilation ;
2338: @end example
2339:
2340: For a @var{word} defined with @code{def-word}, the interpretation
2341: semantics are to push the address of the body of @var{word} and perform
2342: @var{code2}, and the compilation semantics are to push the address of
2343: the body of @var{word} and perform @var{code3}. E.g., @code{constant}
2344: can also be defined like this:
2345:
2346: @example
2347: : constant ( n "name" -- )
2348: create-interpret/compile
2349: ,
2350: interpretation> ( -- n )
2351: @@
2352: <interpretation
2353: compilation> ( compilation. -- ; run-time. -- n )
2354: @@ postpone literal
2355: <compilation ;
2356: @end example
2357:
2358: doc-create-interpret/compile
2359: doc-interpretation>
2360: doc-<interpretation
2361: doc-compilation>
2362: doc-<compilation
2363:
2364: Note that words defined with @code{interpret/compile:} and
2365: @code{create-interpret/compile} have an extended header structure that
2366: differs from other words; however, unless you try to access them with
2367: plain address arithmetic, you should not notice this. Words for
2368: accessing the header structure usually know how to deal with this; e.g.,
2369: @code{' word >body} also gives you the body of a word created with
2370: @code{create-interpret/compile}.
2371:
2372: @node Tokens for Words, Wordlists, Defining Words, Words
2373: @section Tokens for Words
2374:
2375: This chapter describes the creation and use of tokens that represent
2376: words on the stack (and in data space).
2377:
2378: Named words have interpretation and compilation semantics. Unnamed words
2379: just have execution semantics.
2380:
2381: An @dfn{execution token} represents the execution semantics of an
2382: unnamed word. An execution token occupies one cell. As explained in
2383: section @ref{Supplying names}, the execution token of the last words
2384: defined can be produced with
2385:
2386: short-lastxt
2387:
2388: You can perform the semantics represented by an execution token with
2389: doc-execute
2390: You can compile the word with
2391: doc-compile,
2392:
2393: In Gforth, the abstract data type @emph{execution token} is implemented
2394: as CFA (code field address).
2395:
2396: The interpretation semantics of a named word are also represented by an
2397: execution token. You can get it with
2398:
2399: doc-[']
2400: doc-'
2401:
2402: For literals, you use @code{'} in interpreted code and @code{[']} in
2403: compiled code. Gforth's @code{'} and @code{[']} behave somewhat unusual
2404: by complaining about compile-only words. To get an execution token for a
2405: compiling word @var{X}, use @code{COMP' @var{X} drop} or @code{[COMP']
2406: @var{X} drop}.
2407:
2408: The compilation semantics are represented by a @dfn{compilation token}
2409: consisting of two cells: @var{w xt}. The top cell @var{xt} is an
2410: execution token. The compilation semantics represented by the
2411: compilation token can be performed with @code{execute}, which consumes
2412: the whole compilation token, with an additional stack effect determined
2413: by the represented compilation semantics.
2414:
2415: doc-[comp']
2416: doc-comp'
2417:
2418: You can compile the compilation semantics with @code{postpone,}. I.e.,
2419: @code{COMP' @var{word} POSTPONE,} is equivalent to @code{POSTPONE
2420: @var{word}}.
2421:
2422: doc-postpone,
2423:
2424: At present, the @var{w} part of a compilation token is an execution
2425: token, and the @var{xt} part represents either @code{execute} or
2426: @code{compile,}. However, don't rely on that kowledge, unless necessary;
2427: we may introduce unusual compilation tokens in the future (e.g.,
2428: compilation tokens representing the compilation semantics of literals).
2429:
2430: Named words are also represented by the @dfn{name token}. The abstract
2431: data type @emph{name token} is implemented as NFA (name field address).
2432:
2433: doc-find-name
2434: doc-name>int
2435: doc-name?int
2436: doc-name>comp
2437: doc-name>string
2438:
2439: @node Wordlists, Files, Tokens for Words, Words
2440: @section Wordlists
2441:
2442: @node Files, Blocks, Wordlists, Words
2443: @section Files
2444:
2445: @node Blocks, Other I/O, Files, Words
2446: @section Blocks
2447:
2448: @node Other I/O, Programming Tools, Blocks, Words
2449: @section Other I/O
2450:
2451: @node Programming Tools, Assembler and Code words, Other I/O, Words
2452: @section Programming Tools
2453:
2454: @menu
2455: * Debugging:: Simple and quick.
2456: * Assertions:: Making your programs self-checking.
2457: @end menu
2458:
2459: @node Debugging, Assertions, Programming Tools, Programming Tools
2460: @subsection Debugging
2461:
2462: The simple debugging aids provided in @file{debugging.fs}
2463: are meant to support a different style of debugging than the
2464: tracing/stepping debuggers used in languages with long turn-around
2465: times.
2466:
2467: A much better (faster) way in fast-compilig languages is to add
2468: printing code at well-selected places, let the program run, look at
2469: the output, see where things went wrong, add more printing code, etc.,
2470: until the bug is found.
2471:
2472: The word @code{~~} is easy to insert. It just prints debugging
2473: information (by default the source location and the stack contents). It
2474: is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
2475: query-replace them with nothing). The deferred words
2476: @code{printdebugdata} and @code{printdebugline} control the output of
2477: @code{~~}. The default source location output format works well with
2478: Emacs' compilation mode, so you can step through the program at the
2479: source level using @kbd{C-x `} (the advantage over a stepping debugger
2480: is that you can step in any direction and you know where the crash has
2481: happened or where the strange data has occurred).
2482:
2483: Note that the default actions clobber the contents of the pictured
2484: numeric output string, so you should not use @code{~~}, e.g., between
2485: @code{<#} and @code{#>}.
2486:
2487: doc-~~
2488: doc-printdebugdata
2489: doc-printdebugline
2490:
2491: @node Assertions, , Debugging, Programming Tools
2492: @subsection Assertions
2493:
2494: It is a good idea to make your programs self-checking, in particular, if
2495: you use an assumption (e.g., that a certain field of a data structure is
2496: never zero) that may become wrong during maintenance. Gforth supports
2497: assertions for this purpose. They are used like this:
2498:
2499: @example
2500: assert( @var{flag} )
2501: @end example
2502:
2503: The code between @code{assert(} and @code{)} should compute a flag, that
2504: should be true if everything is alright and false otherwise. It should
2505: not change anything else on the stack. The overall stack effect of the
2506: assertion is @code{( -- )}. E.g.
2507:
2508: @example
2509: assert( 1 1 + 2 = ) \ what we learn in school
2510: assert( dup 0<> ) \ assert that the top of stack is not zero
2511: assert( false ) \ this code should not be reached
2512: @end example
2513:
2514: The need for assertions is different at different times. During
2515: debugging, we want more checking, in production we sometimes care more
2516: for speed. Therefore, assertions can be turned off, i.e., the assertion
2517: becomes a comment. Depending on the importance of an assertion and the
2518: time it takes to check it, you may want to turn off some assertions and
2519: keep others turned on. Gforth provides several levels of assertions for
2520: this purpose:
2521:
2522: doc-assert0(
2523: doc-assert1(
2524: doc-assert2(
2525: doc-assert3(
2526: doc-assert(
2527: doc-)
2528:
2529: @code{Assert(} is the same as @code{assert1(}. The variable
2530: @code{assert-level} specifies the highest assertions that are turned
2531: on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
2532: @code{assert1(} assertions perform checking, while @code{assert2(} and
2533: @code{assert3(} assertions are treated as comments.
2534:
2535: Note that the @code{assert-level} is evaluated at compile-time, not at
2536: run-time. I.e., you cannot turn assertions on or off at run-time, you
2537: have to set the @code{assert-level} appropriately before compiling a
2538: piece of code. You can compile several pieces of code at several
2539: @code{assert-level}s (e.g., a trusted library at level 1 and newly
2540: written code at level 3).
2541:
2542: doc-assert-level
2543:
2544: If an assertion fails, a message compatible with Emacs' compilation mode
2545: is produced and the execution is aborted (currently with @code{ABORT"}.
2546: If there is interest, we will introduce a special throw code. But if you
2547: intend to @code{catch} a specific condition, using @code{throw} is
2548: probably more appropriate than an assertion).
2549:
2550: @node Assembler and Code words, Threading Words, Programming Tools, Words
2551: @section Assembler and Code words
2552:
2553: Gforth provides some words for defining primitives (words written in
2554: machine code), and for defining the the machine-code equivalent of
2555: @code{DOES>}-based defining words. However, the machine-independent
2556: nature of Gforth poses a few problems: First of all, Gforth runs on
2557: several architectures, so it can provide no standard assembler. What's
2558: worse is that the register allocation not only depends on the processor,
2559: but also on the @code{gcc} version and options used.
2560:
2561: The words that Gforth offers encapsulate some system dependences (e.g., the
2562: header structure), so a system-independent assembler may be used in
2563: Gforth. If you do not have an assembler, you can compile machine code
2564: directly with @code{,} and @code{c,}.
2565:
2566: doc-assembler
2567: doc-code
2568: doc-end-code
2569: doc-;code
2570: doc-flush-icache
2571:
2572: If @code{flush-icache} does not work correctly, @code{code} words
2573: etc. will not work (reliably), either.
2574:
2575: These words are rarely used. Therefore they reside in @code{code.fs},
2576: which is usually not loaded (except @code{flush-icache}, which is always
2577: present). You can load them with @code{require code.fs}.
2578:
2579: In the assembly code you will want to refer to the inner interpreter's
2580: registers (e.g., the data stack pointer) and you may want to use other
2581: registers for temporary storage. Unfortunately, the register allocation
2582: is installation-dependent.
2583:
2584: The easiest solution is to use explicit register declarations
2585: (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
2586: GNU C Manual}) for all of the inner interpreter's registers: You have to
2587: compile Gforth with @code{-DFORCE_REG} (configure option
2588: @code{--enable-force-reg}) and the appropriate declarations must be
2589: present in the @code{machine.h} file (see @code{mips.h} for an example;
2590: you can find a full list of all declarable register symbols with
2591: @code{grep register engine.c}). If you give explicit registers to all
2592: variables that are declared at the beginning of @code{engine()}, you
2593: should be able to use the other caller-saved registers for temporary
2594: storage. Alternatively, you can use the @code{gcc} option
2595: @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
2596: Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
2597: (however, this restriction on register allocation may slow Gforth
2598: significantly).
2599:
2600: If this solution is not viable (e.g., because @code{gcc} does not allow
2601: you to explicitly declare all the registers you need), you have to find
2602: out by looking at the code where the inner interpreter's registers
2603: reside and which registers can be used for temporary storage. You can
2604: get an assembly listing of the engine's code with @code{make engine.s}.
2605:
2606: In any case, it is good practice to abstract your assembly code from the
2607: actual register allocation. E.g., if the data stack pointer resides in
2608: register @code{$17}, create an alias for this register called @code{sp},
2609: and use that in your assembly code.
2610:
2611: Another option for implementing normal and defining words efficiently
2612: is: adding the wanted functionality to the source of Gforth. For normal
2613: words you just have to edit @file{primitives} (@pxref{Automatic
2614: Generation}), defining words (equivalent to @code{;CODE} words, for fast
2615: defined words) may require changes in @file{engine.c}, @file{kernal.fs},
2616: @file{prims2x.fs}, and possibly @file{cross.fs}.
2617:
2618:
2619: @node Threading Words, , Assembler and Code words, Words
2620: @section Threading Words
2621:
2622: These words provide access to code addresses and other threading stuff
2623: in Gforth (and, possibly, other interpretive Forths). It more or less
2624: abstracts away the differences between direct and indirect threading
2625: (and, for direct threading, the machine dependences). However, at
2626: present this wordset is still inclomplete. It is also pretty low-level;
2627: some day it will hopefully be made unnecessary by an internals words set
2628: that abstracts implementation details away completely.
2629:
2630: doc->code-address
2631: doc->does-code
2632: doc-code-address!
2633: doc-does-code!
2634: doc-does-handler!
2635: doc-/does-handler
2636:
2637: The code addresses produced by various defining words are produced by
2638: the following words:
2639:
2640: doc-docol:
2641: doc-docon:
2642: doc-dovar:
2643: doc-douser:
2644: doc-dodefer:
2645: doc-dofield:
2646:
2647: You can recognize words defined by a @code{CREATE}...@code{DOES>} word
2648: with @code{>DOES-CODE}. If the word was defined in that way, the value
2649: returned is different from 0 and identifies the @code{DOES>} used by the
2650: defining word.
2651:
2652: @node Tools, ANS conformance, Words, Top
2653: @chapter Tools
2654:
2655: @menu
2656: * ANS Report:: Report the words used, sorted by wordset
2657: @end menu
2658:
2659: See also @ref{Emacs and Gforth}.
2660:
2661: @node ANS Report, , Tools, Tools
2662: @section @file{ans-report.fs}: Report the words used, sorted by wordset
2663:
2664: If you want to label a Forth program as ANS Forth Program, you must
2665: document which wordsets the program uses; for extension wordsets, it is
2666: helpful to list the words the program requires from these wordsets
2667: (because Forth systems are allowed to provide only some words of them).
2668:
2669: The @file{ans-report.fs} tool makes it easy for you to determine which
2670: words from which wordset and which non-ANS words your application
2671: uses. You simply have to include @file{ans-report.fs} before loading the
2672: program you want to check. After loading your program, you can get the
2673: report with @code{print-ans-report}. A typical use is to run this as
2674: batch job like this:
2675: @example
2676: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
2677: @end example
2678:
2679: The output looks like this (for @file{compat/control.fs}):
2680: @example
2681: The program uses the following words
2682: from CORE :
2683: : POSTPONE THEN ; immediate ?dup IF 0=
2684: from BLOCK-EXT :
2685: \
2686: from FILE :
2687: (
2688: @end example
2689:
2690: @subsection Caveats
2691:
2692: Note that @file{ans-report.fs} just checks which words are used, not whether
2693: they are used in an ANS Forth conforming way!
2694:
2695: Some words are defined in several wordsets in the
2696: standard. @file{ans-report.fs} reports them for only one of the
2697: wordsets, and not necessarily the one you expect. It depends on usage
2698: which wordset is the right one to specify. E.g., if you only use the
2699: compilation semantics of @code{S"}, it is a Core word; if you also use
2700: its interpretation semantics, it is a File word.
2701:
2702:
2703: @node ANS conformance, Model, Tools, Top
2704: @chapter ANS conformance
2705:
2706: To the best of our knowledge, Gforth is an
2707:
2708: ANS Forth System
2709: @itemize @bullet
2710: @item providing the Core Extensions word set
2711: @item providing the Block word set
2712: @item providing the Block Extensions word set
2713: @item providing the Double-Number word set
2714: @item providing the Double-Number Extensions word set
2715: @item providing the Exception word set
2716: @item providing the Exception Extensions word set
2717: @item providing the Facility word set
2718: @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
2719: @item providing the File Access word set
2720: @item providing the File Access Extensions word set
2721: @item providing the Floating-Point word set
2722: @item providing the Floating-Point Extensions word set
2723: @item providing the Locals word set
2724: @item providing the Locals Extensions word set
2725: @item providing the Memory-Allocation word set
2726: @item providing the Memory-Allocation Extensions word set (that one's easy)
2727: @item providing the Programming-Tools word set
2728: @item providing @code{;CODE}, @code{AHEAD}, @code{ASSEMBLER}, @code{BYE}, @code{CODE}, @code{CS-PICK}, @code{CS-ROLL}, @code{STATE}, @code{[ELSE]}, @code{[IF]}, @code{[THEN]} from the Programming-Tools Extensions word set
2729: @item providing the Search-Order word set
2730: @item providing the Search-Order Extensions word set
2731: @item providing the String word set
2732: @item providing the String Extensions word set (another easy one)
2733: @end itemize
2734:
2735: In addition, ANS Forth systems are required to document certain
2736: implementation choices. This chapter tries to meet these
2737: requirements. In many cases it gives a way to ask the system for the
2738: information instead of providing the information directly, in
2739: particular, if the information depends on the processor, the operating
2740: system or the installation options chosen, or if they are likely to
2741: change during the maintenance of Gforth.
2742:
2743: @comment The framework for the rest has been taken from pfe.
2744:
2745: @menu
2746: * The Core Words::
2747: * The optional Block word set::
2748: * The optional Double Number word set::
2749: * The optional Exception word set::
2750: * The optional Facility word set::
2751: * The optional File-Access word set::
2752: * The optional Floating-Point word set::
2753: * The optional Locals word set::
2754: * The optional Memory-Allocation word set::
2755: * The optional Programming-Tools word set::
2756: * The optional Search-Order word set::
2757: @end menu
2758:
2759:
2760: @c =====================================================================
2761: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
2762: @comment node-name, next, previous, up
2763: @section The Core Words
2764: @c =====================================================================
2765:
2766: @menu
2767: * core-idef:: Implementation Defined Options
2768: * core-ambcond:: Ambiguous Conditions
2769: * core-other:: Other System Documentation
2770: @end menu
2771:
2772: @c ---------------------------------------------------------------------
2773: @node core-idef, core-ambcond, The Core Words, The Core Words
2774: @subsection Implementation Defined Options
2775: @c ---------------------------------------------------------------------
2776:
2777: @table @i
2778:
2779: @item (Cell) aligned addresses:
2780: processor-dependent. Gforth's alignment words perform natural alignment
2781: (e.g., an address aligned for a datum of size 8 is divisible by
2782: 8). Unaligned accesses usually result in a @code{-23 THROW}.
2783:
2784: @item @code{EMIT} and non-graphic characters:
2785: The character is output using the C library function (actually, macro)
2786: @code{putc}.
2787:
2788: @item character editing of @code{ACCEPT} and @code{EXPECT}:
2789: This is modeled on the GNU readline library (@pxref{Readline
2790: Interaction, , Command Line Editing, readline, The GNU Readline
2791: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
2792: producing a full word completion every time you type it (instead of
2793: producing the common prefix of all completions).
2794:
2795: @item character set:
2796: The character set of your computer and display device. Gforth is
2797: 8-bit-clean (but some other component in your system may make trouble).
2798:
2799: @item Character-aligned address requirements:
2800: installation-dependent. Currently a character is represented by a C
2801: @code{unsigned char}; in the future we might switch to @code{wchar_t}
2802: (Comments on that requested).
2803:
2804: @item character-set extensions and matching of names:
2805: Any character except the ASCII NUL charcter can be used in a
2806: name. Matching is case-insensitive (except in @code{TABLE}s. The
2807: matching is performed using the C function @code{strncasecmp}, whose
2808: function is probably influenced by the locale. E.g., the @code{C} locale
2809: does not know about accents and umlauts, so they are matched
2810: case-sensitively in that locale. For portability reasons it is best to
2811: write programs such that they work in the @code{C} locale. Then one can
2812: use libraries written by a Polish programmer (who might use words
2813: containing ISO Latin-2 encoded characters) and by a French programmer
2814: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
2815: funny results for some of the words (which ones, depends on the font you
2816: are using)). Also, the locale you prefer may not be available in other
2817: operating systems. Hopefully, Unicode will solve these problems one day.
2818:
2819: @item conditions under which control characters match a space delimiter:
2820: If @code{WORD} is called with the space character as a delimiter, all
2821: white-space characters (as identified by the C macro @code{isspace()})
2822: are delimiters. @code{PARSE}, on the other hand, treats space like other
2823: delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
2824: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
2825: interpreter (aka text interpreter) by default, treats all white-space
2826: characters as delimiters.
2827:
2828: @item format of the control flow stack:
2829: The data stack is used as control flow stack. The size of a control flow
2830: stack item in cells is given by the constant @code{cs-item-size}. At the
2831: time of this writing, an item consists of a (pointer to a) locals list
2832: (third), an address in the code (second), and a tag for identifying the
2833: item (TOS). The following tags are used: @code{defstart},
2834: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
2835: @code{scopestart}.
2836:
2837: @item conversion of digits > 35
2838: The characters @code{[\]^_'} are the digits with the decimal value
2839: 36@minus{}41. There is no way to input many of the larger digits.
2840:
2841: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
2842: The cursor is moved to the end of the entered string. If the input is
2843: terminated using the @kbd{Return} key, a space is typed.
2844:
2845: @item exception abort sequence of @code{ABORT"}:
2846: The error string is stored into the variable @code{"error} and a
2847: @code{-2 throw} is performed.
2848:
2849: @item input line terminator:
2850: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
2851: lines. One of these characters is typically produced when you type the
2852: @kbd{Enter} or @kbd{Return} key.
2853:
2854: @item maximum size of a counted string:
2855: @code{s" /counted-string" environment? drop .}. Currently 255 characters
2856: on all ports, but this may change.
2857:
2858: @item maximum size of a parsed string:
2859: Given by the constant @code{/line}. Currently 255 characters.
2860:
2861: @item maximum size of a definition name, in characters:
2862: 31
2863:
2864: @item maximum string length for @code{ENVIRONMENT?}, in characters:
2865: 31
2866:
2867: @item method of selecting the user input device:
2868: The user input device is the standard input. There is currently no way to
2869: change it from within Gforth. However, the input can typically be
2870: redirected in the command line that starts Gforth.
2871:
2872: @item method of selecting the user output device:
2873: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
2874: @code{outfile-id} (@code{stdout} by default). Gforth uses buffered
2875: output, so output on a terminal does not become visible before the next
2876: newline or buffer overflow. Output on non-terminals is invisible until
2877: the buffer overflows.
2878:
2879: @item methods of dictionary compilation:
2880: What are we expected to document here?
2881:
2882: @item number of bits in one address unit:
2883: @code{s" address-units-bits" environment? drop .}. 8 in all current
2884: ports.
2885:
2886: @item number representation and arithmetic:
2887: Processor-dependent. Binary two's complement on all current ports.
2888:
2889: @item ranges for integer types:
2890: Installation-dependent. Make environmental queries for @code{MAX-N},
2891: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
2892: unsigned (and positive) types is 0. The lower bound for signed types on
2893: two's complement and one's complement machines machines can be computed
2894: by adding 1 to the upper bound.
2895:
2896: @item read-only data space regions:
2897: The whole Forth data space is writable.
2898:
2899: @item size of buffer at @code{WORD}:
2900: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
2901: shared with the pictured numeric output string. If overwriting
2902: @code{PAD} is acceptable, it is as large as the remaining dictionary
2903: space, although only as much can be sensibly used as fits in a counted
2904: string.
2905:
2906: @item size of one cell in address units:
2907: @code{1 cells .}.
2908:
2909: @item size of one character in address units:
2910: @code{1 chars .}. 1 on all current ports.
2911:
2912: @item size of the keyboard terminal buffer:
2913: Varies. You can determine the size at a specific time using @code{lp@@
2914: tib - .}. It is shared with the locals stack and TIBs of files that
2915: include the current file. You can change the amount of space for TIBs
2916: and locals stack at Gforth startup with the command line option
2917: @code{-l}.
2918:
2919: @item size of the pictured numeric output buffer:
2920: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
2921: shared with @code{WORD}.
2922:
2923: @item size of the scratch area returned by @code{PAD}:
2924: The remainder of dictionary space. You can even use the unused part of
2925: the data stack space. The current size can be computed with @code{sp@@
2926: pad - .}.
2927:
2928: @item system case-sensitivity characteristics:
2929: Dictionary searches are case insensitive (except in
2930: @code{TABLE}s). However, as explained above under @i{character-set
2931: extensions}, the matching for non-ASCII characters is determined by the
2932: locale you are using. In the default @code{C} locale all non-ASCII
2933: characters are matched case-sensitively.
2934:
2935: @item system prompt:
2936: @code{ ok} in interpret state, @code{ compiled} in compile state.
2937:
2938: @item division rounding:
2939: installation dependent. @code{s" floored" environment? drop .}. We leave
2940: the choice to @code{gcc} (what to use for @code{/}) and to you (whether to use
2941: @code{fm/mod}, @code{sm/rem} or simply @code{/}).
2942:
2943: @item values of @code{STATE} when true:
2944: -1.
2945:
2946: @item values returned after arithmetic overflow:
2947: On two's complement machines, arithmetic is performed modulo
2948: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
2949: arithmetic (with appropriate mapping for signed types). Division by zero
2950: typically results in a @code{-55 throw} (Floating-point unidentified
2951: fault), although a @code{-10 throw} (divide by zero) would be more
2952: appropriate.
2953:
2954: @item whether the current definition can be found after @t{DOES>}:
2955: No.
2956:
2957: @end table
2958:
2959: @c ---------------------------------------------------------------------
2960: @node core-ambcond, core-other, core-idef, The Core Words
2961: @subsection Ambiguous conditions
2962: @c ---------------------------------------------------------------------
2963:
2964: @table @i
2965:
2966: @item a name is neither a word nor a number:
2967: @code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
2968: preserves the data and FP stack, so you don't lose more work than
2969: necessary.
2970:
2971: @item a definition name exceeds the maximum length allowed:
2972: @code{-19 throw} (Word name too long)
2973:
2974: @item addressing a region not inside the various data spaces of the forth system:
2975: The stacks, code space and name space are accessible. Machine code space is
2976: typically readable. Accessing other addresses gives results dependent on
2977: the operating system. On decent systems: @code{-9 throw} (Invalid memory
2978: address).
2979:
2980: @item argument type incompatible with parameter:
2981: This is usually not caught. Some words perform checks, e.g., the control
2982: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
2983: mismatch).
2984:
2985: @item attempting to obtain the execution token of a word with undefined execution semantics:
2986: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
2987: get an execution token for @code{compile-only-error} (which performs a
2988: @code{-14 throw} when executed).
2989:
2990: @item dividing by zero:
2991: typically results in a @code{-55 throw} (floating point unidentified
2992: fault), although a @code{-10 throw} (divide by zero) would be more
2993: appropriate.
2994:
2995: @item insufficient data stack or return stack space:
2996: Not checked. This typically results in mysterious illegal memory
2997: accesses, producing @code{-9 throw} (Invalid memory address) or
2998: @code{-23 throw} (Address alignment exception).
2999:
3000: @item insufficient space for loop control parameters:
3001: like other return stack overflows.
3002:
3003: @item insufficient space in the dictionary:
3004: Not checked. Similar results as stack overflows. However, typically the
3005: error appears at a different place when one inserts or removes code.
3006:
3007: @item interpreting a word with undefined interpretation semantics:
3008: For some words, we defined interpretation semantics. For the others:
3009: @code{-14 throw} (Interpreting a compile-only word).
3010:
3011: @item modifying the contents of the input buffer or a string literal:
3012: These are located in writable memory and can be modified.
3013:
3014: @item overflow of the pictured numeric output string:
3015: Not checked.
3016:
3017: @item parsed string overflow:
3018: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
3019:
3020: @item producing a result out of range:
3021: On two's complement machines, arithmetic is performed modulo
3022: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
3023: arithmetic (with appropriate mapping for signed types). Division by zero
3024: typically results in a @code{-55 throw} (floatingpoint unidentified
3025: fault), although a @code{-10 throw} (divide by zero) would be more
3026: appropriate. @code{convert} and @code{>number} currently overflow
3027: silently.
3028:
3029: @item reading from an empty data or return stack:
3030: The data stack is checked by the outer (aka text) interpreter after
3031: every word executed. If it has underflowed, a @code{-4 throw} (Stack
3032: underflow) is performed. Apart from that, the stacks are not checked and
3033: underflows can result in similar behaviour as overflows (of adjacent
3034: stacks).
3035:
3036: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
3037: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
3038: use zero-length string as a name). Words like @code{'} probably will not
3039: find what they search. Note that it is possible to create zero-length
3040: names with @code{nextname} (should it not?).
3041:
3042: @item @code{>IN} greater than input buffer:
3043: The next invocation of a parsing word returns a string wih length 0.
3044:
3045: @item @code{RECURSE} appears after @code{DOES>}:
3046: Compiles a recursive call to the defining word, not to the defined word.
3047:
3048: @item argument input source different than current input source for @code{RESTORE-INPUT}:
3049: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
3050: the end of the file was reached), its source-id may be
3051: reused. Therefore, restoring an input source specification referencing a
3052: closed file may lead to unpredictable results instead of a @code{-12
3053: THROW}.
3054:
3055: In the future, Gforth may be able to restore input source specifications
3056: from other than the current input soruce.
3057:
3058: @item data space containing definitions gets de-allocated:
3059: Deallocation with @code{allot} is not checked. This typically resuls in
3060: memory access faults or execution of illegal instructions.
3061:
3062: @item data space read/write with incorrect alignment:
3063: Processor-dependent. Typically results in a @code{-23 throw} (Address
3064: alignment exception). Under Linux on a 486 or later processor with
3065: alignment turned on, incorrect alignment results in a @code{-9 throw}
3066: (Invalid memory address). There are reportedly some processors with
3067: alignment restrictions that do not report them.
3068:
3069: @item data space pointer not properly aligned, @code{,}, @code{C,}:
3070: Like other alignment errors.
3071:
3072: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
3073: Not checked. May cause an illegal memory access.
3074:
3075: @item loop control parameters not available:
3076: Not checked. The counted loop words simply assume that the top of return
3077: stack items are loop control parameters and behave accordingly.
3078:
3079: @item most recent definition does not have a name (@code{IMMEDIATE}):
3080: @code{abort" last word was headerless"}.
3081:
3082: @item name not defined by @code{VALUE} used by @code{TO}:
3083: @code{-32 throw} (Invalid name argument) (unless name was defined by
3084: @code{CONSTANT}; then it just changes the constant).
3085:
3086: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
3087: @code{-13 throw} (Undefined word)
3088:
3089: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
3090: Gforth behaves as if they were of the same type. I.e., you can predict
3091: the behaviour by interpreting all parameters as, e.g., signed.
3092:
3093: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
3094: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
3095: compilation semantics of @code{TO}.
3096:
3097: @item String longer than a counted string returned by @code{WORD}:
3098: Not checked. The string will be ok, but the count will, of course,
3099: contain only the least significant bits of the length.
3100:
3101: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
3102: Processor-dependent. Typical behaviours are returning 0 and using only
3103: the low bits of the shift count.
3104:
3105: @item word not defined via @code{CREATE}:
3106: @code{>BODY} produces the PFA of the word no matter how it was defined.
3107:
3108: @code{DOES>} changes the execution semantics of the last defined word no
3109: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
3110: @code{CREATE , DOES>}.
3111:
3112: @item words improperly used outside @code{<#} and @code{#>}:
3113: Not checked. As usual, you can expect memory faults.
3114:
3115: @end table
3116:
3117:
3118: @c ---------------------------------------------------------------------
3119: @node core-other, , core-ambcond, The Core Words
3120: @subsection Other system documentation
3121: @c ---------------------------------------------------------------------
3122:
3123: @table @i
3124:
3125: @item nonstandard words using @code{PAD}:
3126: None.
3127:
3128: @item operator's terminal facilities available:
3129: After processing the command line, Gforth goes into interactive mode,
3130: and you can give commands to Gforth interactively. The actual facilities
3131: available depend on how you invoke Gforth.
3132:
3133: @item program data space available:
3134: @code{sp@@ here - .} gives the space remaining for dictionary and data
3135: stack together.
3136:
3137: @item return stack space available:
3138: By default 16 KBytes. The default can be overridden with the @code{-r}
3139: switch (@pxref{Invocation}) when Gforth starts up.
3140:
3141: @item stack space available:
3142: @code{sp@@ here - .} gives the space remaining for dictionary and data
3143: stack together.
3144:
3145: @item system dictionary space required, in address units:
3146: Type @code{here forthstart - .} after startup. At the time of this
3147: writing, this gives 70108 (bytes) on a 32-bit system.
3148: @end table
3149:
3150:
3151: @c =====================================================================
3152: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
3153: @section The optional Block word set
3154: @c =====================================================================
3155:
3156: @menu
3157: * block-idef:: Implementation Defined Options
3158: * block-ambcond:: Ambiguous Conditions
3159: * block-other:: Other System Documentation
3160: @end menu
3161:
3162:
3163: @c ---------------------------------------------------------------------
3164: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
3165: @subsection Implementation Defined Options
3166: @c ---------------------------------------------------------------------
3167:
3168: @table @i
3169:
3170: @item the format for display by @code{LIST}:
3171: First the screen number is displayed, then 16 lines of 64 characters,
3172: each line preceded by the line number.
3173:
3174: @item the length of a line affected by @code{\}:
3175: 64 characters.
3176: @end table
3177:
3178:
3179: @c ---------------------------------------------------------------------
3180: @node block-ambcond, block-other, block-idef, The optional Block word set
3181: @subsection Ambiguous conditions
3182: @c ---------------------------------------------------------------------
3183:
3184: @table @i
3185:
3186: @item correct block read was not possible:
3187: Typically results in a @code{throw} of some OS-derived value (between
3188: -512 and -2048). If the blocks file was just not long enough, blanks are
3189: supplied for the missing portion.
3190:
3191: @item I/O exception in block transfer:
3192: Typically results in a @code{throw} of some OS-derived value (between
3193: -512 and -2048).
3194:
3195: @item invalid block number:
3196: @code{-35 throw} (Invalid block number)
3197:
3198: @item a program directly alters the contents of @code{BLK}:
3199: The input stream is switched to that other block, at the same
3200: position. If the storing to @code{BLK} happens when interpreting
3201: non-block input, the system will get quite confused when the block ends.
3202:
3203: @item no current block buffer for @code{UPDATE}:
3204: @code{UPDATE} has no effect.
3205:
3206: @end table
3207:
3208:
3209: @c ---------------------------------------------------------------------
3210: @node block-other, , block-ambcond, The optional Block word set
3211: @subsection Other system documentation
3212: @c ---------------------------------------------------------------------
3213:
3214: @table @i
3215:
3216: @item any restrictions a multiprogramming system places on the use of buffer addresses:
3217: No restrictions (yet).
3218:
3219: @item the number of blocks available for source and data:
3220: depends on your disk space.
3221:
3222: @end table
3223:
3224:
3225: @c =====================================================================
3226: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
3227: @section The optional Double Number word set
3228: @c =====================================================================
3229:
3230: @menu
3231: * double-ambcond:: Ambiguous Conditions
3232: @end menu
3233:
3234:
3235: @c ---------------------------------------------------------------------
3236: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
3237: @subsection Ambiguous conditions
3238: @c ---------------------------------------------------------------------
3239:
3240: @table @i
3241:
3242: @item @var{d} outside of range of @var{n} in @code{D>S}:
3243: The least significant cell of @var{d} is produced.
3244:
3245: @end table
3246:
3247:
3248: @c =====================================================================
3249: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
3250: @section The optional Exception word set
3251: @c =====================================================================
3252:
3253: @menu
3254: * exception-idef:: Implementation Defined Options
3255: @end menu
3256:
3257:
3258: @c ---------------------------------------------------------------------
3259: @node exception-idef, , The optional Exception word set, The optional Exception word set
3260: @subsection Implementation Defined Options
3261: @c ---------------------------------------------------------------------
3262:
3263: @table @i
3264: @item @code{THROW}-codes used in the system:
3265: The codes -256@minus{}-511 are used for reporting signals (see
3266: @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
3267: (for file and memory allocation operations). The mapping from OS error
3268: numbers to throw code is -512@minus{}@code{errno}. One side effect of
3269: this mapping is that undefined OS errors produce a message with a
3270: strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
3271: 488} on my system.
3272: @end table
3273:
3274: @c =====================================================================
3275: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
3276: @section The optional Facility word set
3277: @c =====================================================================
3278:
3279: @menu
3280: * facility-idef:: Implementation Defined Options
3281: * facility-ambcond:: Ambiguous Conditions
3282: @end menu
3283:
3284:
3285: @c ---------------------------------------------------------------------
3286: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
3287: @subsection Implementation Defined Options
3288: @c ---------------------------------------------------------------------
3289:
3290: @table @i
3291:
3292: @item encoding of keyboard events (@code{EKEY}):
3293: Not yet implemeted.
3294:
3295: @item duration of a system clock tick
3296: System dependent. With respect to @code{MS}, the time is specified in
3297: microseconds. How well the OS and the hardware implement this, is
3298: another question.
3299:
3300: @item repeatability to be expected from the execution of @code{MS}:
3301: System dependent. On Unix, a lot depends on load. If the system is
3302: lightly loaded, and the delay is short enough that Gforth does not get
3303: swapped out, the performance should be acceptable. Under MS-DOS and
3304: other single-tasking systems, it should be good.
3305:
3306: @end table
3307:
3308:
3309: @c ---------------------------------------------------------------------
3310: @node facility-ambcond, , facility-idef, The optional Facility word set
3311: @subsection Ambiguous conditions
3312: @c ---------------------------------------------------------------------
3313:
3314: @table @i
3315:
3316: @item @code{AT-XY} can't be performed on user output device:
3317: Largely terminal dependant. No range checks are done on the arguments.
3318: No errors are reported. You may see some garbage appearing, you may see
3319: simply nothing happen.
3320:
3321: @end table
3322:
3323:
3324: @c =====================================================================
3325: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
3326: @section The optional File-Access word set
3327: @c =====================================================================
3328:
3329: @menu
3330: * file-idef:: Implementation Defined Options
3331: * file-ambcond:: Ambiguous Conditions
3332: @end menu
3333:
3334:
3335: @c ---------------------------------------------------------------------
3336: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
3337: @subsection Implementation Defined Options
3338: @c ---------------------------------------------------------------------
3339:
3340: @table @i
3341:
3342: @item File access methods used:
3343: @code{R/O}, @code{R/W} and @code{BIN} work as you would
3344: expect. @code{W/O} translates into the C file opening mode @code{w} (or
3345: @code{wb}): The file is cleared, if it exists, and created, if it does
3346: not (both with @code{open-file} and @code{create-file}). Under Unix
3347: @code{create-file} creates a file with 666 permissions modified by your
3348: umask.
3349:
3350: @item file exceptions:
3351: The file words do not raise exceptions (except, perhaps, memory access
3352: faults when you pass illegal addresses or file-ids).
3353:
3354: @item file line terminator:
3355: System-dependent. Gforth uses C's newline character as line
3356: terminator. What the actual character code(s) of this are is
3357: system-dependent.
3358:
3359: @item file name format
3360: System dependent. Gforth just uses the file name format of your OS.
3361:
3362: @item information returned by @code{FILE-STATUS}:
3363: @code{FILE-STATUS} returns the most powerful file access mode allowed
3364: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
3365: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
3366: along with the retured mode.
3367:
3368: @item input file state after an exception when including source:
3369: All files that are left via the exception are closed.
3370:
3371: @item @var{ior} values and meaning:
3372: The @var{ior}s returned by the file and memory allocation words are
3373: intended as throw codes. They typically are in the range
3374: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
3375: @var{ior}s is -512@minus{}@var{errno}.
3376:
3377: @item maximum depth of file input nesting:
3378: limited by the amount of return stack, locals/TIB stack, and the number
3379: of open files available. This should not give you troubles.
3380:
3381: @item maximum size of input line:
3382: @code{/line}. Currently 255.
3383:
3384: @item methods of mapping block ranges to files:
3385: By default, blocks are accessed in the file @file{blocks.fb} in the
3386: current working directory. The file can be switched with @code{USE}.
3387:
3388: @item number of string buffers provided by @code{S"}:
3389: 1
3390:
3391: @item size of string buffer used by @code{S"}:
3392: @code{/line}. currently 255.
3393:
3394: @end table
3395:
3396: @c ---------------------------------------------------------------------
3397: @node file-ambcond, , file-idef, The optional File-Access word set
3398: @subsection Ambiguous conditions
3399: @c ---------------------------------------------------------------------
3400:
3401: @table @i
3402:
3403: @item attempting to position a file outside it's boundaries:
3404: @code{REPOSITION-FILE} is performed as usual: Afterwards,
3405: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
3406:
3407: @item attempting to read from file positions not yet written:
3408: End-of-file, i.e., zero characters are read and no error is reported.
3409:
3410: @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
3411: An appropriate exception may be thrown, but a memory fault or other
3412: problem is more probable.
3413:
3414: @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
3415: The @var{ior} produced by the operation, that discovered the problem, is
3416: thrown.
3417:
3418: @item named file cannot be opened (@code{included}):
3419: The @var{ior} produced by @code{open-file} is thrown.
3420:
3421: @item requesting an unmapped block number:
3422: There are no unmapped legal block numbers. On some operating systems,
3423: writing a block with a large number may overflow the file system and
3424: have an error message as consequence.
3425:
3426: @item using @code{source-id} when @code{blk} is non-zero:
3427: @code{source-id} performs its function. Typically it will give the id of
3428: the source which loaded the block. (Better ideas?)
3429:
3430: @end table
3431:
3432:
3433: @c =====================================================================
3434: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
3435: @section The optional Floating-Point word set
3436: @c =====================================================================
3437:
3438: @menu
3439: * floating-idef:: Implementation Defined Options
3440: * floating-ambcond:: Ambiguous Conditions
3441: @end menu
3442:
3443:
3444: @c ---------------------------------------------------------------------
3445: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
3446: @subsection Implementation Defined Options
3447: @c ---------------------------------------------------------------------
3448:
3449: @table @i
3450:
3451: @item format and range of floating point numbers:
3452: System-dependent; the @code{double} type of C.
3453:
3454: @item results of @code{REPRESENT} when @var{float} is out of range:
3455: System dependent; @code{REPRESENT} is implemented using the C library
3456: function @code{ecvt()} and inherits its behaviour in this respect.
3457:
3458: @item rounding or truncation of floating-point numbers:
3459: System dependent; the rounding behaviour is inherited from the hosting C
3460: compiler. IEEE-FP-based (i.e., most) systems by default round to
3461: nearest, and break ties by rounding to even (i.e., such that the last
3462: bit of the mantissa is 0).
3463:
3464: @item size of floating-point stack:
3465: @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
3466: with the command-line option @code{-f}.
3467:
3468: @item width of floating-point stack:
3469: @code{1 floats}.
3470:
3471: @end table
3472:
3473:
3474: @c ---------------------------------------------------------------------
3475: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
3476: @subsection Ambiguous conditions
3477: @c ---------------------------------------------------------------------
3478:
3479: @table @i
3480:
3481: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
3482: System-dependent. Typically results in a @code{-23 THROW} like other
3483: alignment violations.
3484:
3485: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
3486: System-dependent. Typically results in a @code{-23 THROW} like other
3487: alignment violations.
3488:
3489: @item Floating-point result out of range:
3490: System-dependent. Can result in a @code{-55 THROW} (Floating-point
3491: unidentified fault), or can produce a special value representing, e.g.,
3492: Infinity.
3493:
3494: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
3495: System-dependent. Typically results in an alignment fault like other
3496: alignment violations.
3497:
3498: @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
3499: The floating-point number is converted into decimal nonetheless.
3500:
3501: @item Both arguments are equal to zero (@code{FATAN2}):
3502: System-dependent. @code{FATAN2} is implemented using the C library
3503: function @code{atan2()}.
3504:
3505: @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
3506: System-dependent. Anyway, typically the cos of @var{r1} will not be zero
3507: because of small errors and the tan will be a very large (or very small)
3508: but finite number.
3509:
3510: @item @var{d} cannot be presented precisely as a float in @code{D>F}:
3511: The result is rounded to the nearest float.
3512:
3513: @item dividing by zero:
3514: @code{-55 throw} (Floating-point unidentified fault)
3515:
3516: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
3517: System dependent. On IEEE-FP based systems the number is converted into
3518: an infinity.
3519:
3520: @item @var{float}<1 (@code{facosh}):
3521: @code{-55 throw} (Floating-point unidentified fault)
3522:
3523: @item @var{float}=<-1 (@code{flnp1}):
3524: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
3525: negative infinity is typically produced for @var{float}=-1.
3526:
3527: @item @var{float}=<0 (@code{fln}, @code{flog}):
3528: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
3529: negative infinity is typically produced for @var{float}=0.
3530:
3531: @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
3532: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
3533: produces values for these inputs on my Linux box (Bug in the C library?)
3534:
3535: @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
3536: @code{-55 throw} (Floating-point unidentified fault).
3537:
3538: @item integer part of float cannot be represented by @var{d} in @code{f>d}:
3539: @code{-55 throw} (Floating-point unidentified fault).
3540:
3541: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
3542: This does not happen.
3543: @end table
3544:
3545:
3546:
3547: @c =====================================================================
3548: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
3549: @section The optional Locals word set
3550: @c =====================================================================
3551:
3552: @menu
3553: * locals-idef:: Implementation Defined Options
3554: * locals-ambcond:: Ambiguous Conditions
3555: @end menu
3556:
3557:
3558: @c ---------------------------------------------------------------------
3559: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
3560: @subsection Implementation Defined Options
3561: @c ---------------------------------------------------------------------
3562:
3563: @table @i
3564:
3565: @item maximum number of locals in a definition:
3566: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
3567: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
3568: characters. The number of locals in a definition is bounded by the size
3569: of locals-buffer, which contains the names of the locals.
3570:
3571: @end table
3572:
3573:
3574: @c ---------------------------------------------------------------------
3575: @node locals-ambcond, , locals-idef, The optional Locals word set
3576: @subsection Ambiguous conditions
3577: @c ---------------------------------------------------------------------
3578:
3579: @table @i
3580:
3581: @item executing a named local in interpretation state:
3582: @code{-14 throw} (Interpreting a compile-only word).
3583:
3584: @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
3585: @code{-32 throw} (Invalid name argument)
3586:
3587: @end table
3588:
3589:
3590: @c =====================================================================
3591: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
3592: @section The optional Memory-Allocation word set
3593: @c =====================================================================
3594:
3595: @menu
3596: * memory-idef:: Implementation Defined Options
3597: @end menu
3598:
3599:
3600: @c ---------------------------------------------------------------------
3601: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
3602: @subsection Implementation Defined Options
3603: @c ---------------------------------------------------------------------
3604:
3605: @table @i
3606:
3607: @item values and meaning of @var{ior}:
3608: The @var{ior}s returned by the file and memory allocation words are
3609: intended as throw codes. They typically are in the range
3610: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
3611: @var{ior}s is -512@minus{}@var{errno}.
3612:
3613: @end table
3614:
3615: @c =====================================================================
3616: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
3617: @section The optional Programming-Tools word set
3618: @c =====================================================================
3619:
3620: @menu
3621: * programming-idef:: Implementation Defined Options
3622: * programming-ambcond:: Ambiguous Conditions
3623: @end menu
3624:
3625:
3626: @c ---------------------------------------------------------------------
3627: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
3628: @subsection Implementation Defined Options
3629: @c ---------------------------------------------------------------------
3630:
3631: @table @i
3632:
3633: @item ending sequence for input following @code{;code} and @code{code}:
3634: @code{end-code}
3635:
3636: @item manner of processing input following @code{;code} and @code{code}:
3637: The @code{assembler} vocabulary is pushed on the search order stack, and
3638: the input is processed by the text interpreter, (starting) in interpret
3639: state.
3640:
3641: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
3642: The ANS Forth search order word set.
3643:
3644: @item source and format of display by @code{SEE}:
3645: The source for @code{see} is the intermediate code used by the inner
3646: interpreter. The current @code{see} tries to output Forth source code
3647: as well as possible.
3648:
3649: @end table
3650:
3651: @c ---------------------------------------------------------------------
3652: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
3653: @subsection Ambiguous conditions
3654: @c ---------------------------------------------------------------------
3655:
3656: @table @i
3657:
3658: @item deleting the compilation wordlist (@code{FORGET}):
3659: Not implemented (yet).
3660:
3661: @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
3662: This typically results in an @code{abort"} with a descriptive error
3663: message (may change into a @code{-22 throw} (Control structure mismatch)
3664: in the future). You may also get a memory access error. If you are
3665: unlucky, this ambiguous condition is not caught.
3666:
3667: @item @var{name} can't be found (@code{forget}):
3668: Not implemented (yet).
3669:
3670: @item @var{name} not defined via @code{CREATE}:
3671: @code{;code} behaves like @code{DOES>} in this respect, i.e., it changes
3672: the execution semantics of the last defined word no matter how it was
3673: defined.
3674:
3675: @item @code{POSTPONE} applied to @code{[IF]}:
3676: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
3677: equivalent to @code{[IF]}.
3678:
3679: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
3680: Continue in the same state of conditional compilation in the next outer
3681: input source. Currently there is no warning to the user about this.
3682:
3683: @item removing a needed definition (@code{FORGET}):
3684: Not implemented (yet).
3685:
3686: @end table
3687:
3688:
3689: @c =====================================================================
3690: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
3691: @section The optional Search-Order word set
3692: @c =====================================================================
3693:
3694: @menu
3695: * search-idef:: Implementation Defined Options
3696: * search-ambcond:: Ambiguous Conditions
3697: @end menu
3698:
3699:
3700: @c ---------------------------------------------------------------------
3701: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
3702: @subsection Implementation Defined Options
3703: @c ---------------------------------------------------------------------
3704:
3705: @table @i
3706:
3707: @item maximum number of word lists in search order:
3708: @code{s" wordlists" environment? drop .}. Currently 16.
3709:
3710: @item minimum search order:
3711: @code{root root}.
3712:
3713: @end table
3714:
3715: @c ---------------------------------------------------------------------
3716: @node search-ambcond, , search-idef, The optional Search-Order word set
3717: @subsection Ambiguous conditions
3718: @c ---------------------------------------------------------------------
3719:
3720: @table @i
3721:
3722: @item changing the compilation wordlist (during compilation):
3723: The word is entered into the wordlist that was the compilation wordlist
3724: at the start of the definition. Any changes to the name field (e.g.,
3725: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
3726: are applied to the latest defined word (as reported by @code{last} or
3727: @code{lastxt}), if possible, irrespective of the compilation wordlist.
3728:
3729: @item search order empty (@code{previous}):
3730: @code{abort" Vocstack empty"}.
3731:
3732: @item too many word lists in search order (@code{also}):
3733: @code{abort" Vocstack full"}.
3734:
3735: @end table
3736:
3737: @node Model, Integrating Gforth, ANS conformance, Top
3738: @chapter Model
3739:
3740: This chapter has yet to be written. It will contain information, on
3741: which internal structures you can rely.
3742:
3743: @node Integrating Gforth, Emacs and Gforth, Model, Top
3744: @chapter Integrating Gforth into C programs
3745:
3746: This is not yet implemented.
3747:
3748: Several people like to use Forth as scripting language for applications
3749: that are otherwise written in C, C++, or some other language.
3750:
3751: The Forth system ATLAST provides facilities for embedding it into
3752: applications; unfortunately it has several disadvantages: most
3753: importantly, it is not based on ANS Forth, and it is apparently dead
3754: (i.e., not developed further and not supported). The facilities
3755: provided by Gforth in this area are inspired by ATLASTs facilities, so
3756: making the switch should not be hard.
3757:
3758: We also tried to design the interface such that it can easily be
3759: implemented by other Forth systems, so that we may one day arrive at a
3760: standardized interface. Such a standard interface would allow you to
3761: replace the Forth system without having to rewrite C code.
3762:
3763: You embed the Gforth interpreter by linking with the library
3764: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
3765: global symbols in this library that belong to the interface, have the
3766: prefix @code{forth_}. (Global symbols that are used internally have the
3767: prefix @code{gforth_}).
3768:
3769: You can include the declarations of Forth types and the functions and
3770: variables of the interface with @code{#include <forth.h>}.
3771:
3772: Types.
3773:
3774: Variables.
3775:
3776: Data and FP Stack pointer. Area sizes.
3777:
3778: functions.
3779:
3780: forth_init(imagefile)
3781: forth_evaluate(string) exceptions?
3782: forth_goto(address) (or forth_execute(xt)?)
3783: forth_continue() (a corountining mechanism)
3784:
3785: Adding primitives.
3786:
3787: No checking.
3788:
3789: Signals?
3790:
3791: Accessing the Stacks
3792:
3793: @node Emacs and Gforth, Internals, Integrating Gforth, Top
3794: @chapter Emacs and Gforth
3795:
3796: Gforth comes with @file{gforth.el}, an improved version of
3797: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
3798: improvements are a better (but still not perfect) handling of
3799: indentation. I have also added comment paragraph filling (@kbd{M-q}),
3800: commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
3801: removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
3802: stuff I do not use alone, even though some of it only makes sense for
3803: TILE. To get a description of these features, enter Forth mode and type
3804: @kbd{C-h m}.
3805:
3806: In addition, Gforth supports Emacs quite well: The source code locations
3807: given in error messages, debugging output (from @code{~~}) and failed
3808: assertion messages are in the right format for Emacs' compilation mode
3809: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
3810: Manual}) so the source location corresponding to an error or other
3811: message is only a few keystrokes away (@kbd{C-x `} for the next error,
3812: @kbd{C-c C-c} for the error under the cursor).
3813:
3814: Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
3815: (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
3816: contains the definitions of all words defined afterwards. You can then
3817: find the source for a word using @kbd{M-.}. Note that emacs can use
3818: several tags files at the same time (e.g., one for the Gforth sources
3819: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
3820: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
3821: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
3822: @file{/usr/local/share/gforth/0.2.0/TAGS}).
3823:
3824: To get all these benefits, add the following lines to your @file{.emacs}
3825: file:
3826:
3827: @example
3828: (autoload 'forth-mode "gforth.el")
3829: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
3830: @end example
3831:
3832: @node Internals, Bugs, Emacs and Gforth, Top
3833: @chapter Internals
3834:
3835: Reading this section is not necessary for programming with Gforth. It
3836: should be helpful for finding your way in the Gforth sources.
3837:
3838: The ideas in this section have also been published in the papers
3839: @cite{ANS fig/GNU/??? Forth} (in German) by Bernd Paysan, presented at
3840: the Forth-Tagung '93 and @cite{A Portable Forth Engine} by M. Anton
3841: Ertl, presented at EuroForth '93; the latter is available at
3842: @*@file{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z}.
3843:
3844: @menu
3845: * Portability::
3846: * Threading::
3847: * Primitives::
3848: * System Architecture::
3849: * Performance::
3850: @end menu
3851:
3852: @node Portability, Threading, Internals, Internals
3853: @section Portability
3854:
3855: One of the main goals of the effort is availability across a wide range
3856: of personal machines. fig-Forth, and, to a lesser extent, F83, achieved
3857: this goal by manually coding the engine in assembly language for several
3858: then-popular processors. This approach is very labor-intensive and the
3859: results are short-lived due to progress in computer architecture.
3860:
3861: Others have avoided this problem by coding in C, e.g., Mitch Bradley
3862: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
3863: particularly popular for UNIX-based Forths due to the large variety of
3864: architectures of UNIX machines. Unfortunately an implementation in C
3865: does not mix well with the goals of efficiency and with using
3866: traditional techniques: Indirect or direct threading cannot be expressed
3867: in C, and switch threading, the fastest technique available in C, is
3868: significantly slower. Another problem with C is that it's very
3869: cumbersome to express double integer arithmetic.
3870:
3871: Fortunately, there is a portable language that does not have these
3872: limitations: GNU C, the version of C processed by the GNU C compiler
3873: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
3874: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
3875: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
3876: threading possible, its @code{long long} type (@pxref{Long Long, ,
3877: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
3878: double numbers@footnote{Unfortunately, long longs are not implemented
3879: properly on all machines (e.g., on alpha-osf1, long longs are only 64
3880: bits, the same size as longs (and pointers), but they should be twice as
3881: long according to @ref{Long Long, , Double-Word Integers, gcc.info, GNU
3882: C Manual}). So, we had to implement doubles in C after all. Still, on
3883: most machines we can use long longs and achieve better performance than
3884: with the emulation package.}. GNU C is available for free on all
3885: important (and many unimportant) UNIX machines, VMS, 80386s running
3886: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
3887: on all these machines.
3888:
3889: Writing in a portable language has the reputation of producing code that
3890: is slower than assembly. For our Forth engine we repeatedly looked at
3891: the code produced by the compiler and eliminated most compiler-induced
3892: inefficiencies by appropriate changes in the source-code.
3893:
3894: However, register allocation cannot be portably influenced by the
3895: programmer, leading to some inefficiencies on register-starved
3896: machines. We use explicit register declarations (@pxref{Explicit Reg
3897: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
3898: improve the speed on some machines. They are turned on by using the
3899: @code{gcc} switch @code{-DFORCE_REG}. Unfortunately, this feature not
3900: only depends on the machine, but also on the compiler version: On some
3901: machines some compiler versions produce incorrect code when certain
3902: explicit register declarations are used. So by default
3903: @code{-DFORCE_REG} is not used.
3904:
3905: @node Threading, Primitives, Portability, Internals
3906: @section Threading
3907:
3908: GNU C's labels as values extension (available since @code{gcc-2.0},
3909: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
3910: makes it possible to take the address of @var{label} by writing
3911: @code{&&@var{label}}. This address can then be used in a statement like
3912: @code{goto *@var{address}}. I.e., @code{goto *&&x} is the same as
3913: @code{goto x}.
3914:
3915: With this feature an indirect threaded NEXT looks like:
3916: @example
3917: cfa = *ip++;
3918: ca = *cfa;
3919: goto *ca;
3920: @end example
3921: For those unfamiliar with the names: @code{ip} is the Forth instruction
3922: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
3923: execution token and points to the code field of the next word to be
3924: executed; The @code{ca} (code address) fetched from there points to some
3925: executable code, e.g., a primitive or the colon definition handler
3926: @code{docol}.
3927:
3928: Direct threading is even simpler:
3929: @example
3930: ca = *ip++;
3931: goto *ca;
3932: @end example
3933:
3934: Of course we have packaged the whole thing neatly in macros called
3935: @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
3936:
3937: @menu
3938: * Scheduling::
3939: * Direct or Indirect Threaded?::
3940: * DOES>::
3941: @end menu
3942:
3943: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
3944: @subsection Scheduling
3945:
3946: There is a little complication: Pipelined and superscalar processors,
3947: i.e., RISC and some modern CISC machines can process independent
3948: instructions while waiting for the results of an instruction. The
3949: compiler usually reorders (schedules) the instructions in a way that
3950: achieves good usage of these delay slots. However, on our first tries
3951: the compiler did not do well on scheduling primitives. E.g., for
3952: @code{+} implemented as
3953: @example
3954: n=sp[0]+sp[1];
3955: sp++;
3956: sp[0]=n;
3957: NEXT;
3958: @end example
3959: the NEXT comes strictly after the other code, i.e., there is nearly no
3960: scheduling. After a little thought the problem becomes clear: The
3961: compiler cannot know that sp and ip point to different addresses (and
3962: the version of @code{gcc} we used would not know it even if it was
3963: possible), so it could not move the load of the cfa above the store to
3964: the TOS. Indeed the pointers could be the same, if code on or very near
3965: the top of stack were executed. In the interest of speed we chose to
3966: forbid this probably unused ``feature'' and helped the compiler in
3967: scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
3968: the goto part (@code{NEXT_P2}). @code{+} now looks like:
3969: @example
3970: n=sp[0]+sp[1];
3971: sp++;
3972: NEXT_P1;
3973: sp[0]=n;
3974: NEXT_P2;
3975: @end example
3976: This can be scheduled optimally by the compiler.
3977:
3978: This division can be turned off with the switch @code{-DCISC_NEXT}. This
3979: switch is on by default on machines that do not profit from scheduling
3980: (e.g., the 80386), in order to preserve registers.
3981:
3982: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
3983: @subsection Direct or Indirect Threaded?
3984:
3985: Both! After packaging the nasty details in macro definitions we
3986: realized that we could switch between direct and indirect threading by
3987: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
3988: defining a few machine-specific macros for the direct-threading case.
3989: On the Forth level we also offer access words that hide the
3990: differences between the threading methods (@pxref{Threading Words}).
3991:
3992: Indirect threading is implemented completely
3993: machine-independently. Direct threading needs routines for creating
3994: jumps to the executable code (e.g. to docol or dodoes). These routines
3995: are inherently machine-dependent, but they do not amount to many source
3996: lines. I.e., even porting direct threading to a new machine is a small
3997: effort.
3998:
3999: @node DOES>, , Direct or Indirect Threaded?, Threading
4000: @subsection DOES>
4001: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
4002: the chunk of code executed by every word defined by a
4003: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
4004: the Forth code to be executed, i.e. the code after the @code{DOES>} (the
4005: DOES-code)? There are two solutions:
4006:
4007: In fig-Forth the code field points directly to the dodoes and the
4008: DOES-code address is stored in the cell after the code address
4009: (i.e. at cfa cell+). It may seem that this solution is illegal in the
4010: Forth-79 and all later standards, because in fig-Forth this address
4011: lies in the body (which is illegal in these standards). However, by
4012: making the code field larger for all words this solution becomes legal
4013: again. We use this approach for the indirect threaded version. Leaving
4014: a cell unused in most words is a bit wasteful, but on the machines we
4015: are targetting this is hardly a problem. The other reason for having a
4016: code field size of two cells is to avoid having different image files
4017: for direct and indirect threaded systems (@pxref{System Architecture}).
4018:
4019: The other approach is that the code field points or jumps to the cell
4020: after @code{DOES}. In this variant there is a jump to @code{dodoes} at
4021: this address. @code{dodoes} can then get the DOES-code address by
4022: computing the code address, i.e., the address of the jump to dodoes,
4023: and add the length of that jump field. A variant of this is to have a
4024: call to @code{dodoes} after the @code{DOES>}; then the return address
4025: (which can be found in the return register on RISCs) is the DOES-code
4026: address. Since the two cells available in the code field are usually
4027: used up by the jump to the code address in direct threading, we use
4028: this approach for direct threading. We did not want to add another
4029: cell to the code field.
4030:
4031: @node Primitives, System Architecture, Threading, Internals
4032: @section Primitives
4033:
4034: @menu
4035: * Automatic Generation::
4036: * TOS Optimization::
4037: * Produced code::
4038: @end menu
4039:
4040: @node Automatic Generation, TOS Optimization, Primitives, Primitives
4041: @subsection Automatic Generation
4042:
4043: Since the primitives are implemented in a portable language, there is no
4044: longer any need to minimize the number of primitives. On the contrary,
4045: having many primitives is an advantage: speed. In order to reduce the
4046: number of errors in primitives and to make programming them easier, we
4047: provide a tool, the primitive generator (@file{prims2x.fs}), that
4048: automatically generates most (and sometimes all) of the C code for a
4049: primitive from the stack effect notation. The source for a primitive
4050: has the following form:
4051:
4052: @format
4053: @var{Forth-name} @var{stack-effect} @var{category} [@var{pronounc.}]
4054: [@code{""}@var{glossary entry}@code{""}]
4055: @var{C code}
4056: [@code{:}
4057: @var{Forth code}]
4058: @end format
4059:
4060: The items in brackets are optional. The category and glossary fields
4061: are there for generating the documentation, the Forth code is there
4062: for manual implementations on machines without GNU C. E.g., the source
4063: for the primitive @code{+} is:
4064: @example
4065: + n1 n2 -- n core plus
4066: n = n1+n2;
4067: @end example
4068:
4069: This looks like a specification, but in fact @code{n = n1+n2} is C
4070: code. Our primitive generation tool extracts a lot of information from
4071: the stack effect notations@footnote{We use a one-stack notation, even
4072: though we have separate data and floating-point stacks; The separate
4073: notation can be generated easily from the unified notation.}: The number
4074: of items popped from and pushed on the stack, their type, and by what
4075: name they are referred to in the C code. It then generates a C code
4076: prelude and postlude for each primitive. The final C code for @code{+}
4077: looks like this:
4078:
4079: @example
4080: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
4081: /* */ /* documentation */
4082: @{
4083: DEF_CA /* definition of variable ca (indirect threading) */
4084: Cell n1; /* definitions of variables */
4085: Cell n2;
4086: Cell n;
4087: n1 = (Cell) sp[1]; /* input */
4088: n2 = (Cell) TOS;
4089: sp += 1; /* stack adjustment */
4090: NAME("+") /* debugging output (with -DDEBUG) */
4091: @{
4092: n = n1+n2; /* C code taken from the source */
4093: @}
4094: NEXT_P1; /* NEXT part 1 */
4095: TOS = (Cell)n; /* output */
4096: NEXT_P2; /* NEXT part 2 */
4097: @}
4098: @end example
4099:
4100: This looks long and inefficient, but the GNU C compiler optimizes quite
4101: well and produces optimal code for @code{+} on, e.g., the R3000 and the
4102: HP RISC machines: Defining the @code{n}s does not produce any code, and
4103: using them as intermediate storage also adds no cost.
4104:
4105: There are also other optimizations, that are not illustrated by this
4106: example: Assignments between simple variables are usually for free (copy
4107: propagation). If one of the stack items is not used by the primitive
4108: (e.g. in @code{drop}), the compiler eliminates the load from the stack
4109: (dead code elimination). On the other hand, there are some things that
4110: the compiler does not do, therefore they are performed by
4111: @file{prims2x.fs}: The compiler does not optimize code away that stores
4112: a stack item to the place where it just came from (e.g., @code{over}).
4113:
4114: While programming a primitive is usually easy, there are a few cases
4115: where the programmer has to take the actions of the generator into
4116: account, most notably @code{?dup}, but also words that do not (always)
4117: fall through to NEXT.
4118:
4119: @node TOS Optimization, Produced code, Automatic Generation, Primitives
4120: @subsection TOS Optimization
4121:
4122: An important optimization for stack machine emulators, e.g., Forth
4123: engines, is keeping one or more of the top stack items in
4124: registers. If a word has the stack effect @var{in1}...@var{inx} @code{--}
4125: @var{out1}...@var{outy}, keeping the top @var{n} items in registers
4126: @itemize @bullet
4127: @item
4128: is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
4129: due to fewer loads from and stores to the stack.
4130: @item is slower than keeping @var{n-1} items, if @var{x<>y} and @var{x<n} and
4131: @var{y<n}, due to additional moves between registers.
4132: @end itemize
4133:
4134: In particular, keeping one item in a register is never a disadvantage,
4135: if there are enough registers. Keeping two items in registers is a
4136: disadvantage for frequent words like @code{?branch}, constants,
4137: variables, literals and @code{i}. Therefore our generator only produces
4138: code that keeps zero or one items in registers. The generated C code
4139: covers both cases; the selection between these alternatives is made at
4140: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
4141: code for @code{+} is just a simple variable name in the one-item case,
4142: otherwise it is a macro that expands into @code{sp[0]}. Note that the
4143: GNU C compiler tries to keep simple variables like @code{TOS} in
4144: registers, and it usually succeeds, if there are enough registers.
4145:
4146: The primitive generator performs the TOS optimization for the
4147: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
4148: operations the benefit of this optimization is even larger:
4149: floating-point operations take quite long on most processors, but can be
4150: performed in parallel with other operations as long as their results are
4151: not used. If the FP-TOS is kept in a register, this works. If
4152: it is kept on the stack, i.e., in memory, the store into memory has to
4153: wait for the result of the floating-point operation, lengthening the
4154: execution time of the primitive considerably.
4155:
4156: The TOS optimization makes the automatic generation of primitives a
4157: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
4158: @code{TOS} is not sufficient. There are some special cases to
4159: consider:
4160: @itemize @bullet
4161: @item In the case of @code{dup ( w -- w w )} the generator must not
4162: eliminate the store to the original location of the item on the stack,
4163: if the TOS optimization is turned on.
4164: @item Primitives with stack effects of the form @code{--}
4165: @var{out1}...@var{outy} must store the TOS to the stack at the start.
4166: Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
4167: must load the TOS from the stack at the end. But for the null stack
4168: effect @code{--} no stores or loads should be generated.
4169: @end itemize
4170:
4171: @node Produced code, , TOS Optimization, Primitives
4172: @subsection Produced code
4173:
4174: To see what assembly code is produced for the primitives on your machine
4175: with your compiler and your flag settings, type @code{make engine.s} and
4176: look at the resulting file @file{engine.s}.
4177:
4178: @node System Architecture, Performance, Primitives, Internals
4179: @section System Architecture
4180:
4181: Our Forth system consists not only of primitives, but also of
4182: definitions written in Forth. Since the Forth compiler itself belongs
4183: to those definitions, it is not possible to start the system with the
4184: primitives and the Forth source alone. Therefore we provide the Forth
4185: code as an image file in nearly executable form. At the start of the
4186: system a C routine loads the image file into memory, sets up the
4187: memory (stacks etc.) according to information in the image file, and
4188: starts executing Forth code.
4189:
4190: The image file format is a compromise between the goals of making it
4191: easy to generate image files and making them portable. The easiest way
4192: to generate an image file is to just generate a memory dump. However,
4193: this kind of image file cannot be used on a different machine, or on
4194: the next version of the engine on the same machine, it even might not
4195: work with the same engine compiled by a different version of the C
4196: compiler. We would like to have as few versions of the image file as
4197: possible, because we do not want to distribute many versions of the
4198: same image file, and to make it easy for the users to use their image
4199: files on many machines. We currently need to create a different image
4200: file for machines with different cell sizes and different byte order
4201: (little- or big-endian)@footnote{We are considering adding information to the
4202: image file that enables the loader to change the byte order.}.
4203:
4204: Forth code that is going to end up in a portable image file has to
4205: comply to some restrictions: addresses have to be stored in memory with
4206: special words (@code{A!}, @code{A,}, etc.) in order to make the code
4207: relocatable. Cells, floats, etc., have to be stored at the natural
4208: alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
4209: dividable by~8. This happens automatically in our system when you use
4210: the ANS Forth alignment words.}, in order to avoid alignment faults on
4211: machines with stricter alignment. The image file is produced by a
4212: metacompiler (@file{cross.fs}).
4213:
4214: So, unlike the image file of Mitch Bradleys @code{cforth}, our image
4215: file is not directly executable, but has to undergo some manipulations
4216: during loading. Address relocation is performed at image load-time, not
4217: at run-time. The loader also has to replace tokens standing for
4218: primitive calls with the appropriate code-field addresses (or code
4219: addresses in the case of direct threading).
4220:
4221: @node Performance, , System Architecture, Internals
4222: @section Performance
4223:
4224: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
4225: impossible to write a significantly faster engine.
4226:
4227: On register-starved machines like the 386 architecture processors
4228: improvements are possible, because @code{gcc} does not utilize the
4229: registers as well as a human, even with explicit register declarations;
4230: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
4231: and hand-tuned it for the 486; this system is 1.19 times faster on the
4232: Sieve benchmark on a 486DX2/66 than Gforth compiled with
4233: @code{gcc-2.6.3} with @code{-DFORCE_REG}.
4234:
4235: However, this potential advantage of assembly language implementations
4236: is not necessarily realized in complete Forth systems: We compared
4237: Gforth (direct threaded, compiled with @code{gcc-2.6.3} and
4238: @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
4239: 1994) and Eforth (with and without peephole (aka pinhole) optimization
4240: of the threaded code); all these systems were written in assembly
4241: language. We also compared Gforth with three systems written in C:
4242: PFE-0.9.14 (compiled with @code{gcc-2.6.3} with the default
4243: configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
4244: -DUNROLL_NEXT}), ThisForth Beta (compiled with gcc-2.6.3 -O3
4245: -fomit-frame-pointer; ThisForth employs peephole optimization of the
4246: threaded code) and TILE (compiled with @code{make opt}). We benchmarked
4247: Gforth, PFE, ThisForth and TILE on a 486DX2/66 under Linux. Kenneth
4248: O'Heskin kindly provided the results for Win32Forth and NT Forth on a
4249: 486DX2/66 with similar memory performance under Windows NT. Marcel
4250: Hendrix ported Eforth to Linux, then extended it to run the benchmarks,
4251: added the peephole optimizer, ran the benchmarks and reported the
4252: results.
4253:
4254: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
4255: matrix multiplication come from the Stanford integer benchmarks and have
4256: been translated into Forth by Martin Fraeman; we used the versions
4257: included in the TILE Forth package, but with bigger data set sizes; and
4258: a recursive Fibonacci number computation for benchmarking calling
4259: performance. The following table shows the time taken for the benchmarks
4260: scaled by the time taken by Gforth (in other words, it shows the speedup
4261: factor that Gforth achieved over the other systems).
4262:
4263: @example
4264: relative Win32- NT eforth This-
4265: time Gforth Forth Forth eforth +opt PFE Forth TILE
4266: sieve 1.00 1.39 1.14 1.39 0.85 1.58 3.18 8.58
4267: bubble 1.00 1.31 1.41 1.48 0.88 1.50 3.88
4268: matmul 1.00 1.47 1.35 1.46 0.74 1.58 4.09
4269: fib 1.00 1.52 1.34 1.22 0.86 1.74 2.99 4.30
4270: @end example
4271:
4272: You may find the good performance of Gforth compared with the systems
4273: written in assembly language quite surprising. One important reason for
4274: the disappointing performance of these systems is probably that they are
4275: not written optimally for the 486 (e.g., they use the @code{lods}
4276: instruction). In addition, Win32Forth uses a comfortable, but costly
4277: method for relocating the Forth image: like @code{cforth}, it computes
4278: the actual addresses at run time, resulting in two address computations
4279: per NEXT (@pxref{System Architecture}).
4280:
4281: Only Eforth with the peephole optimizer performs comparable to
4282: Gforth. The speedups achieved with peephole optimization of threaded
4283: code are quite remarkable. Adding a peephole optimizer to Gforth should
4284: cause similar speedups.
4285:
4286: The speedup of Gforth over PFE, ThisForth and TILE can be easily
4287: explained with the self-imposed restriction to standard C, which makes
4288: efficient threading impossible (however, the measured implementation of
4289: PFE uses a GNU C extension: @ref{Global Reg Vars, , Defining Global
4290: Register Variables, gcc.info, GNU C Manual}). Moreover, current C
4291: compilers have a hard time optimizing other aspects of the ThisForth
4292: and the TILE source.
4293:
4294: Note that the performance of Gforth on 386 architecture processors
4295: varies widely with the version of @code{gcc} used. E.g., @code{gcc-2.5.8}
4296: failed to allocate any of the virtual machine registers into real
4297: machine registers by itself and would not work correctly with explicit
4298: register declarations, giving a 1.3 times slower engine (on a 486DX2/66
4299: running the Sieve) than the one measured above.
4300:
4301: In @cite{Translating Forth to Efficient C} by M. Anton Ertl and Martin
4302: Maierhofer (presented at EuroForth '95), an indirect threaded version of
4303: Gforth is compared with Win32Forth, NT Forth, PFE, and ThisForth; that
4304: version of Gforth is 2\%@minus{}8\% slower on a 486 than the version
4305: used here. The paper available at
4306: @*@file{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz};
4307: it also contains numbers for some native code systems. You can find
4308: numbers for Gforth on various machines in @file{Benchres}.
4309:
4310: @node Bugs, Origin, Internals, Top
4311: @chapter Bugs
4312:
4313: Known bugs are described in the file BUGS in the Gforth distribution.
4314:
4315: If you find a bug, please send a bug report to
4316: @code{bug-gforth@@gnu.ai.mit.edu}. A bug report should
4317: describe the Gforth version used (it is announced at the start of an
4318: interactive Gforth session), the machine and operating system (on Unix
4319: systems you can use @code{uname -a} to produce this information), the
4320: installation options (send the @code{config.status} file), and a
4321: complete list of changes you (or your installer) have made to the Gforth
4322: sources (if any); it should contain a program (or a sequence of keyboard
4323: commands) that reproduces the bug and a description of what you think
4324: constitutes the buggy behaviour.
4325:
4326: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
4327: to Report Bugs, gcc.info, GNU C Manual}.
4328:
4329:
4330: @node Origin, Word Index, Bugs, Top
4331: @chapter Authors and Ancestors of Gforth
4332:
4333: @section Authors and Contributors
4334:
4335: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
4336: Ertl. The third major author was Jens Wilke. Lennart Benschop (who was
4337: one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
4338: with their continuous feedback. Lennart Benshop contributed
4339: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
4340: support for calling C libraries. Helpful comments also came from Paul
4341: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
4342: Wavrik, Barrie Stott and Marc de Groot.
4343:
4344: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
4345: and autoconf, among others), and to the creators of the Internet: Gforth
4346: was developed across the Internet, and its authors have not met
4347: physically yet.
4348:
4349: @section Pedigree
4350:
4351: Gforth descends from BigForth (1993) and fig-Forth. Gforth and PFE (by
4352: Dirk Zoller) will cross-fertilize each other. Of course, a significant
4353: part of the design of Gforth was prescribed by ANS Forth.
4354:
4355: Bernd Paysan wrote BigForth, a descendent from TurboForth, an unreleased
4356: 32 bit native code version of VolksForth for the Atari ST, written
4357: mostly by Dietrich Weineck.
4358:
4359: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
4360: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
4361: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
4362:
4363: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
4364: Forth-83 standard. !! Pedigree? When?
4365:
4366: A team led by Bill Ragsdale implemented fig-Forth on many processors in
4367: 1979. Robert Selzer and Bill Ragsdale developed the original
4368: implementation of fig-Forth for the 6502 based on microForth.
4369:
4370: The principal architect of microForth was Dean Sanderson. microForth was
4371: FORTH, Inc.'s first off-the-shelf product. It was developped in 1976 for
4372: the 1802, and subsequently implemented on the 8080, the 6800 and the
4373: Z80.
4374:
4375: All earlier Forth systems were custom-made, usually by Charles Moore,
4376: who discovered (as he puts it) Forth during the late 60s. The first full
4377: Forth existed in 1971.
4378:
4379: A part of the information in this section comes from @cite{The Evolution
4380: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
4381: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
4382: Notices 28(3), 1993. You can find more historical and genealogical
4383: information about Forth there.
4384:
4385: @node Word Index, Node Index, Origin, Top
4386: @chapter Word Index
4387:
4388: This index is as incomplete as the manual. Each word is listed with
4389: stack effect and wordset.
4390:
4391: @printindex fn
4392:
4393: @node Node Index, , Word Index, Top
4394: @chapter Node Index
4395:
4396: This index is even less complete than the manual.
4397:
4398: @contents
4399: @bye
4400:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>