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 0.1
11:
12: Copyright @copyright{} 1995 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: @titlepage
40: @sp 10
41: @center @titlefont{Gforth Manual}
42: @sp 2
43: @center for version 0.1
44: @sp 2
45: @center Anton Ertl
46: @sp 3
47: @center This manual is under construction
48:
49: @comment The following two commands start the copyright page.
50: @page
51: @vskip 0pt plus 1filll
52: Copyright @copyright{} 1995 Free Software Foundation, Inc.
53:
54: @comment !! Published by ... or You can get a copy of this manual ...
55:
56: Permission is granted to make and distribute verbatim copies of
57: this manual provided the copyright notice and this permission notice
58: are preserved on all copies.
59:
60: Permission is granted to copy and distribute modified versions of this
61: manual under the conditions for verbatim copying, provided also that the
62: sections entitled "Distribution" and "General Public License" are
63: included exactly as in the original, and provided that the entire
64: resulting derived work is distributed under the terms of a permission
65: notice identical to this one.
66:
67: Permission is granted to copy and distribute translations of this manual
68: into another language, under the above conditions for modified versions,
69: except that the sections entitled "Distribution" and "General Public
70: License" may be included in a translation approved by the author instead
71: of in the original English.
72: @end titlepage
73:
74:
75: @node Top, License, (dir), (dir)
76: @ifinfo
77: Gforth is a free implementation of ANS Forth available on many
78: personal machines. This manual corresponds to version 0.1.
79: @end ifinfo
80:
81: @menu
82: * License::
83: * Goals:: About the Gforth Project
84: * Other Books:: Things you might want to read
85: * Invocation:: Starting Gforth
86: * Words:: Forth words available in Gforth
87: * ANS conformance:: Implementation-defined options etc.
88: * Model:: The abstract machine of Gforth
89: * Emacs and Gforth:: The Gforth Mode
90: * Internals:: Implementation details
91: * Bugs:: How to report them
92: * Pedigree:: Ancestors of Gforth
93: * Word Index:: An item for each Forth word
94: * Node Index:: An item for each node
95: @end menu
96:
97: @node License, Goals, Top, Top
98: @unnumbered GNU GENERAL PUBLIC LICENSE
99: @center Version 2, June 1991
100:
101: @display
102: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
103: 675 Mass Ave, Cambridge, MA 02139, USA
104:
105: Everyone is permitted to copy and distribute verbatim copies
106: of this license document, but changing it is not allowed.
107: @end display
108:
109: @unnumberedsec Preamble
110:
111: The licenses for most software are designed to take away your
112: freedom to share and change it. By contrast, the GNU General Public
113: License is intended to guarantee your freedom to share and change free
114: software---to make sure the software is free for all its users. This
115: General Public License applies to most of the Free Software
116: Foundation's software and to any other program whose authors commit to
117: using it. (Some other Free Software Foundation software is covered by
118: the GNU Library General Public License instead.) You can apply it to
119: your programs, too.
120:
121: When we speak of free software, we are referring to freedom, not
122: price. Our General Public Licenses are designed to make sure that you
123: have the freedom to distribute copies of free software (and charge for
124: this service if you wish), that you receive source code or can get it
125: if you want it, that you can change the software or use pieces of it
126: in new free programs; and that you know you can do these things.
127:
128: To protect your rights, we need to make restrictions that forbid
129: anyone to deny you these rights or to ask you to surrender the rights.
130: These restrictions translate to certain responsibilities for you if you
131: distribute copies of the software, or if you modify it.
132:
133: For example, if you distribute copies of such a program, whether
134: gratis or for a fee, you must give the recipients all the rights that
135: you have. You must make sure that they, too, receive or can get the
136: source code. And you must show them these terms so they know their
137: rights.
138:
139: We protect your rights with two steps: (1) copyright the software, and
140: (2) offer you this license which gives you legal permission to copy,
141: distribute and/or modify the software.
142:
143: Also, for each author's protection and ours, we want to make certain
144: that everyone understands that there is no warranty for this free
145: software. If the software is modified by someone else and passed on, we
146: want its recipients to know that what they have is not the original, so
147: that any problems introduced by others will not reflect on the original
148: authors' reputations.
149:
150: Finally, any free program is threatened constantly by software
151: patents. We wish to avoid the danger that redistributors of a free
152: program will individually obtain patent licenses, in effect making the
153: program proprietary. To prevent this, we have made it clear that any
154: patent must be licensed for everyone's free use or not licensed at all.
155:
156: The precise terms and conditions for copying, distribution and
157: modification follow.
158:
159: @iftex
160: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
161: @end iftex
162: @ifinfo
163: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
164: @end ifinfo
165:
166: @enumerate 0
167: @item
168: This License applies to any program or other work which contains
169: a notice placed by the copyright holder saying it may be distributed
170: under the terms of this General Public License. The ``Program'', below,
171: refers to any such program or work, and a ``work based on the Program''
172: means either the Program or any derivative work under copyright law:
173: that is to say, a work containing the Program or a portion of it,
174: either verbatim or with modifications and/or translated into another
175: language. (Hereinafter, translation is included without limitation in
176: the term ``modification''.) Each licensee is addressed as ``you''.
177:
178: Activities other than copying, distribution and modification are not
179: covered by this License; they are outside its scope. The act of
180: running the Program is not restricted, and the output from the Program
181: is covered only if its contents constitute a work based on the
182: Program (independent of having been made by running the Program).
183: Whether that is true depends on what the Program does.
184:
185: @item
186: You may copy and distribute verbatim copies of the Program's
187: source code as you receive it, in any medium, provided that you
188: conspicuously and appropriately publish on each copy an appropriate
189: copyright notice and disclaimer of warranty; keep intact all the
190: notices that refer to this License and to the absence of any warranty;
191: and give any other recipients of the Program a copy of this License
192: along with the Program.
193:
194: You may charge a fee for the physical act of transferring a copy, and
195: you may at your option offer warranty protection in exchange for a fee.
196:
197: @item
198: You may modify your copy or copies of the Program or any portion
199: of it, thus forming a work based on the Program, and copy and
200: distribute such modifications or work under the terms of Section 1
201: above, provided that you also meet all of these conditions:
202:
203: @enumerate a
204: @item
205: You must cause the modified files to carry prominent notices
206: stating that you changed the files and the date of any change.
207:
208: @item
209: You must cause any work that you distribute or publish, that in
210: whole or in part contains or is derived from the Program or any
211: part thereof, to be licensed as a whole at no charge to all third
212: parties under the terms of this License.
213:
214: @item
215: If the modified program normally reads commands interactively
216: when run, you must cause it, when started running for such
217: interactive use in the most ordinary way, to print or display an
218: announcement including an appropriate copyright notice and a
219: notice that there is no warranty (or else, saying that you provide
220: a warranty) and that users may redistribute the program under
221: these conditions, and telling the user how to view a copy of this
222: License. (Exception: if the Program itself is interactive but
223: does not normally print such an announcement, your work based on
224: the Program is not required to print an announcement.)
225: @end enumerate
226:
227: These requirements apply to the modified work as a whole. If
228: identifiable sections of that work are not derived from the Program,
229: and can be reasonably considered independent and separate works in
230: themselves, then this License, and its terms, do not apply to those
231: sections when you distribute them as separate works. But when you
232: distribute the same sections as part of a whole which is a work based
233: on the Program, the distribution of the whole must be on the terms of
234: this License, whose permissions for other licensees extend to the
235: entire whole, and thus to each and every part regardless of who wrote it.
236:
237: Thus, it is not the intent of this section to claim rights or contest
238: your rights to work written entirely by you; rather, the intent is to
239: exercise the right to control the distribution of derivative or
240: collective works based on the Program.
241:
242: In addition, mere aggregation of another work not based on the Program
243: with the Program (or with a work based on the Program) on a volume of
244: a storage or distribution medium does not bring the other work under
245: the scope of this License.
246:
247: @item
248: You may copy and distribute the Program (or a work based on it,
249: under Section 2) in object code or executable form under the terms of
250: Sections 1 and 2 above provided that you also do one of the following:
251:
252: @enumerate a
253: @item
254: Accompany it with the complete corresponding machine-readable
255: source code, which must be distributed under the terms of Sections
256: 1 and 2 above on a medium customarily used for software interchange; or,
257:
258: @item
259: Accompany it with a written offer, valid for at least three
260: years, to give any third party, for a charge no more than your
261: cost of physically performing source distribution, a complete
262: machine-readable copy of the corresponding source code, to be
263: distributed under the terms of Sections 1 and 2 above on a medium
264: customarily used for software interchange; or,
265:
266: @item
267: Accompany it with the information you received as to the offer
268: to distribute corresponding source code. (This alternative is
269: allowed only for noncommercial distribution and only if you
270: received the program in object code or executable form with such
271: an offer, in accord with Subsection b above.)
272: @end enumerate
273:
274: The source code for a work means the preferred form of the work for
275: making modifications to it. For an executable work, complete source
276: code means all the source code for all modules it contains, plus any
277: associated interface definition files, plus the scripts used to
278: control compilation and installation of the executable. However, as a
279: special exception, the source code distributed need not include
280: anything that is normally distributed (in either source or binary
281: form) with the major components (compiler, kernel, and so on) of the
282: operating system on which the executable runs, unless that component
283: itself accompanies the executable.
284:
285: If distribution of executable or object code is made by offering
286: access to copy from a designated place, then offering equivalent
287: access to copy the source code from the same place counts as
288: distribution of the source code, even though third parties are not
289: compelled to copy the source along with the object code.
290:
291: @item
292: You may not copy, modify, sublicense, or distribute the Program
293: except as expressly provided under this License. Any attempt
294: otherwise to copy, modify, sublicense or distribute the Program is
295: void, and will automatically terminate your rights under this License.
296: However, parties who have received copies, or rights, from you under
297: this License will not have their licenses terminated so long as such
298: parties remain in full compliance.
299:
300: @item
301: You are not required to accept this License, since you have not
302: signed it. However, nothing else grants you permission to modify or
303: distribute the Program or its derivative works. These actions are
304: prohibited by law if you do not accept this License. Therefore, by
305: modifying or distributing the Program (or any work based on the
306: Program), you indicate your acceptance of this License to do so, and
307: all its terms and conditions for copying, distributing or modifying
308: the Program or works based on it.
309:
310: @item
311: Each time you redistribute the Program (or any work based on the
312: Program), the recipient automatically receives a license from the
313: original licensor to copy, distribute or modify the Program subject to
314: these terms and conditions. You may not impose any further
315: restrictions on the recipients' exercise of the rights granted herein.
316: You are not responsible for enforcing compliance by third parties to
317: this License.
318:
319: @item
320: If, as a consequence of a court judgment or allegation of patent
321: infringement or for any other reason (not limited to patent issues),
322: conditions are imposed on you (whether by court order, agreement or
323: otherwise) that contradict the conditions of this License, they do not
324: excuse you from the conditions of this License. If you cannot
325: distribute so as to satisfy simultaneously your obligations under this
326: License and any other pertinent obligations, then as a consequence you
327: may not distribute the Program at all. For example, if a patent
328: license would not permit royalty-free redistribution of the Program by
329: all those who receive copies directly or indirectly through you, then
330: the only way you could satisfy both it and this License would be to
331: refrain entirely from distribution of the Program.
332:
333: If any portion of this section is held invalid or unenforceable under
334: any particular circumstance, the balance of the section is intended to
335: apply and the section as a whole is intended to apply in other
336: circumstances.
337:
338: It is not the purpose of this section to induce you to infringe any
339: patents or other property right claims or to contest validity of any
340: such claims; this section has the sole purpose of protecting the
341: integrity of the free software distribution system, which is
342: implemented by public license practices. Many people have made
343: generous contributions to the wide range of software distributed
344: through that system in reliance on consistent application of that
345: system; it is up to the author/donor to decide if he or she is willing
346: to distribute software through any other system and a licensee cannot
347: impose that choice.
348:
349: This section is intended to make thoroughly clear what is believed to
350: be a consequence of the rest of this License.
351:
352: @item
353: If the distribution and/or use of the Program is restricted in
354: certain countries either by patents or by copyrighted interfaces, the
355: original copyright holder who places the Program under this License
356: may add an explicit geographical distribution limitation excluding
357: those countries, so that distribution is permitted only in or among
358: countries not thus excluded. In such case, this License incorporates
359: the limitation as if written in the body of this License.
360:
361: @item
362: The Free Software Foundation may publish revised and/or new versions
363: of the General Public License from time to time. Such new versions will
364: be similar in spirit to the present version, but may differ in detail to
365: address new problems or concerns.
366:
367: Each version is given a distinguishing version number. If the Program
368: specifies a version number of this License which applies to it and ``any
369: later version'', you have the option of following the terms and conditions
370: either of that version or of any later version published by the Free
371: Software Foundation. If the Program does not specify a version number of
372: this License, you may choose any version ever published by the Free Software
373: Foundation.
374:
375: @item
376: If you wish to incorporate parts of the Program into other free
377: programs whose distribution conditions are different, write to the author
378: to ask for permission. For software which is copyrighted by the Free
379: Software Foundation, write to the Free Software Foundation; we sometimes
380: make exceptions for this. Our decision will be guided by the two goals
381: of preserving the free status of all derivatives of our free software and
382: of promoting the sharing and reuse of software generally.
383:
384: @iftex
385: @heading NO WARRANTY
386: @end iftex
387: @ifinfo
388: @center NO WARRANTY
389: @end ifinfo
390:
391: @item
392: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
393: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
394: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
395: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
396: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
397: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
398: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
399: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
400: REPAIR OR CORRECTION.
401:
402: @item
403: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
404: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
405: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
406: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
407: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
408: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
409: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
410: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
411: POSSIBILITY OF SUCH DAMAGES.
412: @end enumerate
413:
414: @iftex
415: @heading END OF TERMS AND CONDITIONS
416: @end iftex
417: @ifinfo
418: @center END OF TERMS AND CONDITIONS
419: @end ifinfo
420:
421: @page
422: @unnumberedsec How to Apply These Terms to Your New Programs
423:
424: If you develop a new program, and you want it to be of the greatest
425: possible use to the public, the best way to achieve this is to make it
426: free software which everyone can redistribute and change under these terms.
427:
428: To do so, attach the following notices to the program. It is safest
429: to attach them to the start of each source file to most effectively
430: convey the exclusion of warranty; and each file should have at least
431: the ``copyright'' line and a pointer to where the full notice is found.
432:
433: @smallexample
434: @var{one line to give the program's name and a brief idea of what it does.}
435: Copyright (C) 19@var{yy} @var{name of author}
436:
437: This program is free software; you can redistribute it and/or modify
438: it under the terms of the GNU General Public License as published by
439: the Free Software Foundation; either version 2 of the License, or
440: (at your option) any later version.
441:
442: This program is distributed in the hope that it will be useful,
443: but WITHOUT ANY WARRANTY; without even the implied warranty of
444: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
445: GNU General Public License for more details.
446:
447: You should have received a copy of the GNU General Public License
448: along with this program; if not, write to the Free Software
449: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
450: @end smallexample
451:
452: Also add information on how to contact you by electronic and paper mail.
453:
454: If the program is interactive, make it output a short notice like this
455: when it starts in an interactive mode:
456:
457: @smallexample
458: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
459: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
460: type `show w'.
461: This is free software, and you are welcome to redistribute it
462: under certain conditions; type `show c' for details.
463: @end smallexample
464:
465: The hypothetical commands @samp{show w} and @samp{show c} should show
466: the appropriate parts of the General Public License. Of course, the
467: commands you use may be called something other than @samp{show w} and
468: @samp{show c}; they could even be mouse-clicks or menu items---whatever
469: suits your program.
470:
471: You should also get your employer (if you work as a programmer) or your
472: school, if any, to sign a ``copyright disclaimer'' for the program, if
473: necessary. Here is a sample; alter the names:
474:
475: @smallexample
476: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
477: `Gnomovision' (which makes passes at compilers) written by James Hacker.
478:
479: @var{signature of Ty Coon}, 1 April 1989
480: Ty Coon, President of Vice
481: @end smallexample
482:
483: This General Public License does not permit incorporating your program into
484: proprietary programs. If your program is a subroutine library, you may
485: consider it more useful to permit linking proprietary applications with the
486: library. If this is what you want to do, use the GNU Library General
487: Public License instead of this License.
488:
489: @iftex
490: @node Preface
491: @comment node-name, next, previous, up
492: @unnumbered Preface
493: @cindex Preface
494: This manual documents Gforth. The reader is expected to know
495: Forth. This manual is primarily a reference manual. @xref{Other Books}
496: for introductory material.
497: @end iftex
498:
499: @node Goals, Other Books, License, Top
500: @comment node-name, next, previous, up
501: @chapter Goals of Gforth
502: @cindex Goals
503: The goal of the Gforth Project is to develop a standard model for
504: ANSI Forth. This can be split into several subgoals:
505:
506: @itemize @bullet
507: @item
508: Gforth should conform to the ANSI Forth standard.
509: @item
510: It should be a model, i.e. it should define all the
511: implementation-dependent things.
512: @item
513: It should become standard, i.e. widely accepted and used. This goal
514: is the most difficult one.
515: @end itemize
516:
517: To achieve these goals Gforth should be
518: @itemize @bullet
519: @item
520: Similar to previous models (fig-Forth, F83)
521: @item
522: Powerful. It should provide for all the things that are considered
523: necessary today and even some that are not yet considered necessary.
524: @item
525: Efficient. It should not get the reputation of being exceptionally
526: slow.
527: @item
528: Free.
529: @item
530: Available on many machines/easy to port.
531: @end itemize
532:
533: Have we achieved these goals? Gforth conforms to the ANS Forth
534: standard. It may be considered a model, but we have not yet documented
535: which parts of the model are stable and which parts we are likely to
536: change. It certainly has not yet become a de facto standard. It has some
537: similarities and some differences to previous models. It has some
538: powerful features, but not yet everything that we envisioned. We
539: certainly have achieved our execution speed goals (@pxref{Performance}).
540: It is free and available on many machines.
541:
542: @node Other Books, Invocation, Goals, Top
543: @chapter Other books on ANS Forth
544:
545: As the standard is relatively new, there are not many books out yet. It
546: is not recommended to learn Forth by using Gforth and a book that is
547: not written for ANS Forth, as you will not know your mistakes from the
548: deviations of the book.
549:
550: There is, of course, the standard, the definite reference if you want to
551: write ANS Forth programs. It is available in printed form from the
552: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
553: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about $200. You
554: can also get it from Global Engineering Documents (Tel.: USA (800)
555: 854-7179; Fax.: (303) 843-9880) for about $300.
556:
557: @cite{dpANS6}, the last draft of the standard, which was then submitted to ANSI
558: for publication is available electronically and for free in some MS Word
559: format, and it has been converted to HTML. Some pointers to these
560: versions can be found through
561: http://www.complang.tuwien.ac.at/projects/forth.html.
562:
563: @cite{Forth: The new model} by Jack Woehr (Prentice-Hall, 1993) is an
564: introductory book based on a draft version of the standard. It does not
565: cover the whole standard. It also contains interesting background
566: information (Jack Woehr was in the ANS Forth Technical Committe). It is
567: not appropriate for complete newbies, but programmers experienced in
568: other languages should find it ok.
569:
570: @node Invocation, Words, Other Books, Top
571: @chapter Invocation
572:
573: You will usually just say @code{gforth}. In many other cases the default
574: Gforth image will be invoked like this:
575:
576: @example
577: gforth [files] [-e forth-code]
578: @end example
579:
580: executing the contents of the files and the Forth code in the order they
581: are given.
582:
583: In general, the command line looks like this:
584:
585: @example
586: gforth [initialization options] [image-specific options]
587: @end example
588:
589: The initialization options must come before the rest of the command
590: line. They are:
591:
592: @table @code
593: @item --image-file @var{file}
594: @item -i @var{file}
595: Loads the Forth image @var{file} instead of the default
596: @file{gforth.fi}.
597:
598: @item --path @var{path}
599: @item -p @var{path}
600: Uses @var{path} for searching the image file and Forth source code
601: files instead of the default in the environment variable
602: @code{GFORTHPATH} or the path specified at installation time (typically
603: @file{/usr/local/lib/gforth:.}). A path is given as a @code{:}-separated
604: list.
605:
606: @item --dictionary-size @var{size}
607: @item -m @var{size}
608: Allocate @var{size} space for the Forth dictionary space instead of
609: using the default specified in the image (typically 256K). The
610: @var{size} specification consists of an integer and a unit (e.g.,
611: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
612: size, in this case Cells), @code{k} (kilobytes), and @code{M}
613: (Megabytes). If no unit is specified, @code{e} is used.
614:
615: @item --data-stack-size @var{size}
616: @item -d @var{size}
617: Allocate @var{size} space for the data stack instead of using the
618: default specified in the image (typically 16K).
619:
620: @item --return-stack-size @var{size}
621: @item -r @var{size}
622: Allocate @var{size} space for the return stack instead of using the
623: default specified in the image (typically 16K).
624:
625: @item --fp-stack-size @var{size}
626: @item -f @var{size}
627: Allocate @var{size} space for the floating point stack instead of
628: using the default specified in the image (typically 16K). In this case
629: the unit specifier @code{e} refers to floating point numbers.
630:
631: @item --locals-stack-size @var{size}
632: @item -l @var{size}
633: Allocate @var{size} space for the locals stack instead of using the
634: default specified in the image (typically 16K).
635:
636: @end table
637:
638: As explained above, the image-specific command-line arguments for the
639: default image @file{gforth.fi} consist of a sequence of filenames and
640: @code{-e @var{forth-code}} options that are interpreted in the seqence
641: in which they are given. The @code{-e @var{forth-code}} or
642: @code{--evaluate @var{forth-code}} option evaluates the forth
643: code. This option takes only one argument; if you want to evaluate more
644: Forth words, you have to quote them or use several @code{-e}s. To exit
645: after processing the command line (instead of entering interactive mode)
646: append @code{-e bye} to the command line.
647:
648: If you have several versions of Gforth installed, @code{gforth} will
649: invoke the version that was installed last. @code{gforth-@var{version}}
650: invokes a specific version. You may want to use the option
651: @code{--path}, if your environment contains the variable
652: @code{GFORTHPATH}.
653:
654: Not yet implemented:
655: On startup the system first executes the system initialization file
656: (unless the option @code{--no-init-file} is given; note that the system
657: resulting from using this option may not be ANS Forth conformant). Then
658: the user initialization file @file{.gforth.fs} is executed, unless the
659: option @code{--no-rc} is given; this file is first searched in @file{.},
660: then in @file{~}, then in the normal path (see above).
661:
662: @node Words, ANS conformance, Invocation, Top
663: @chapter Forth Words
664:
665: @menu
666: * Notation::
667: * Arithmetic::
668: * Stack Manipulation::
669: * Memory access::
670: * Control Structures::
671: * Locals::
672: * Defining Words::
673: * Wordlists::
674: * Files::
675: * Blocks::
676: * Other I/O::
677: * Programming Tools::
678: * Assembler and Code words::
679: * Threading Words::
680: @end menu
681:
682: @node Notation, Arithmetic, Words, Words
683: @section Notation
684:
685: The Forth words are described in this section in the glossary notation
686: that has become a de-facto standard for Forth texts, i.e.
687:
688: @format
689: @var{word} @var{Stack effect} @var{wordset} @var{pronunciation}
690: @end format
691: @var{Description}
692:
693: @table @var
694: @item word
695: The name of the word. BTW, Gforth is case insensitive, so you can
696: type the words in in lower case (However, @pxref{core-idef}).
697:
698: @item Stack effect
699: The stack effect is written in the notation @code{@var{before} --
700: @var{after}}, where @var{before} and @var{after} describe the top of
701: stack entries before and after the execution of the word. The rest of
702: the stack is not touched by the word. The top of stack is rightmost,
703: i.e., a stack sequence is written as it is typed in. Note that Gforth
704: uses a separate floating point stack, but a unified stack
705: notation. Also, return stack effects are not shown in @var{stack
706: effect}, but in @var{Description}. The name of a stack item describes
707: the type and/or the function of the item. See below for a discussion of
708: the types.
709:
710: All words have two stack effects: A compile-time stack effect and a
711: run-time stack effect. The compile-time stack-effect of most words is
712: @var{ -- }. If the compile-time stack-effect of a word deviates from
713: this standard behaviour, or the word does other unusual things at
714: compile time, both stack effects are shown; otherwise only the run-time
715: stack effect is shown.
716:
717: @item pronunciation
718: How the word is pronounced
719:
720: @item wordset
721: The ANS Forth standard is divided into several wordsets. A standard
722: system need not support all of them. So, the fewer wordsets your program
723: uses the more portable it will be in theory. However, we suspect that
724: most ANS Forth systems on personal machines will feature all
725: wordsets. Words that are not defined in the ANS standard have
726: @code{gforth} or @code{gforth-internal} as wordset. @code{gforth}
727: describes words that will work in future releases of Gforth;
728: @code{gforth-internal} words are more volatile. Environmental query
729: strings are also displayed like words; you can recognize them by the
730: @code{environment} in the wordset field.
731:
732: @item Description
733: A description of the behaviour of the word.
734: @end table
735:
736: The type of a stack item is specified by the character(s) the name
737: starts with:
738:
739: @table @code
740: @item f
741: Bool, i.e. @code{false} or @code{true}.
742: @item c
743: Char
744: @item w
745: Cell, can contain an integer or an address
746: @item n
747: signed integer
748: @item u
749: unsigned integer
750: @item d
751: double sized signed integer
752: @item ud
753: double sized unsigned integer
754: @item r
755: Float
756: @item a_
757: Cell-aligned address
758: @item c_
759: Char-aligned address (note that a Char is two bytes in Windows NT)
760: @item f_
761: Float-aligned address
762: @item df_
763: Address aligned for IEEE double precision float
764: @item sf_
765: Address aligned for IEEE single precision float
766: @item xt
767: Execution token, same size as Cell
768: @item wid
769: Wordlist ID, same size as Cell
770: @item f83name
771: Pointer to a name structure
772: @end table
773:
774: @node Arithmetic, Stack Manipulation, Notation, Words
775: @section Arithmetic
776: Forth arithmetic is not checked, i.e., you will not hear about integer
777: overflow on addition or multiplication, you may hear about division by
778: zero if you are lucky. The operator is written after the operands, but
779: the operands are still in the original order. I.e., the infix @code{2-1}
780: corresponds to @code{2 1 -}. Forth offers a variety of division
781: operators. If you perform division with potentially negative operands,
782: you do not want to use @code{/} or @code{/mod} with its undefined
783: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
784: former, @pxref{Mixed precision}).
785:
786: @menu
787: * Single precision::
788: * Bitwise operations::
789: * Mixed precision:: operations with single and double-cell integers
790: * Double precision:: Double-cell integer arithmetic
791: * Floating Point::
792: @end menu
793:
794: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
795: @subsection Single precision
796: doc-+
797: doc--
798: doc-*
799: doc-/
800: doc-mod
801: doc-/mod
802: doc-negate
803: doc-abs
804: doc-min
805: doc-max
806:
807: @node Bitwise operations, Mixed precision, Single precision, Arithmetic
808: @subsection Bitwise operations
809: doc-and
810: doc-or
811: doc-xor
812: doc-invert
813: doc-2*
814: doc-2/
815:
816: @node Mixed precision, Double precision, Bitwise operations, Arithmetic
817: @subsection Mixed precision
818: doc-m+
819: doc-*/
820: doc-*/mod
821: doc-m*
822: doc-um*
823: doc-m*/
824: doc-um/mod
825: doc-fm/mod
826: doc-sm/rem
827:
828: @node Double precision, Floating Point, Mixed precision, Arithmetic
829: @subsection Double precision
830:
831: The outer (aka text) interpreter converts numbers containing a dot into
832: a double precision number. Note that only numbers with the dot as last
833: character are standard-conforming.
834:
835: doc-d+
836: doc-d-
837: doc-dnegate
838: doc-dabs
839: doc-dmin
840: doc-dmax
841:
842: @node Floating Point, , Double precision, Arithmetic
843: @subsection Floating Point
844:
845: The format of floating point numbers recognized by the outer (aka text)
846: interpreter is: a signed decimal number, possibly containing a decimal
847: point (@code{.}), followed by @code{E} or @code{e}, optionally followed
848: by a signed integer (the exponent). E.g., @code{1e} ist the same as
849: @code{+1.0e+1}. Note that a number without @code{e}
850: is not interpreted as floating-point number, but as double (if the
851: number contains a @code{.}) or single precision integer. Also,
852: conversions between string and floating point numbers always use base
853: 10, irrespective of the value of @code{BASE}. If @code{BASE} contains a
854: value greater then 14, the @code{E} may be interpreted as digit and the
855: number will be interpreted as integer, unless it has a signed exponent
856: (both @code{+} and @code{-} are allowed as signs).
857:
858: Angles in floating point operations are given in radians (a full circle
859: has 2 pi radians). Note, that Gforth has a separate floating point
860: stack, but we use the unified notation.
861:
862: Floating point numbers have a number of unpleasant surprises for the
863: unwary (e.g., floating point addition is not associative) and even a few
864: for the wary. You should not use them unless you know what you are doing
865: or you don't care that the results you get are totally bogus. If you
866: want to learn about the problems of floating point numbers (and how to
867: avoid them), you might start with @cite{David Goldberg, What Every
868: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
869: Computing Surveys 23(1):5@minus{}48, March 1991}.
870:
871: doc-f+
872: doc-f-
873: doc-f*
874: doc-f/
875: doc-fnegate
876: doc-fabs
877: doc-fmax
878: doc-fmin
879: doc-floor
880: doc-fround
881: doc-f**
882: doc-fsqrt
883: doc-fexp
884: doc-fexpm1
885: doc-fln
886: doc-flnp1
887: doc-flog
888: doc-falog
889: doc-fsin
890: doc-fcos
891: doc-fsincos
892: doc-ftan
893: doc-fasin
894: doc-facos
895: doc-fatan
896: doc-fatan2
897: doc-fsinh
898: doc-fcosh
899: doc-ftanh
900: doc-fasinh
901: doc-facosh
902: doc-fatanh
903:
904: @node Stack Manipulation, Memory access, Arithmetic, Words
905: @section Stack Manipulation
906:
907: Gforth has a data stack (aka parameter stack) for characters, cells,
908: addresses, and double cells, a floating point stack for floating point
909: numbers, a return stack for storing the return addresses of colon
910: definitions and other data, and a locals stack for storing local
911: variables. Note that while every sane Forth has a separate floating
912: point stack, this is not strictly required; an ANS Forth system could
913: theoretically keep floating point numbers on the data stack. As an
914: additional difficulty, you don't know how many cells a floating point
915: number takes. It is reportedly possible to write words in a way that
916: they work also for a unified stack model, but we do not recommend trying
917: it. Instead, just say that your program has an environmental dependency
918: on a separate FP stack.
919:
920: Also, a Forth system is allowed to keep the local variables on the
921: return stack. This is reasonable, as local variables usually eliminate
922: the need to use the return stack explicitly. So, if you want to produce
923: a standard complying program and if you are using local variables in a
924: word, forget about return stack manipulations in that word (see the
925: standard document for the exact rules).
926:
927: @menu
928: * Data stack::
929: * Floating point stack::
930: * Return stack::
931: * Locals stack::
932: * Stack pointer manipulation::
933: @end menu
934:
935: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
936: @subsection Data stack
937: doc-drop
938: doc-nip
939: doc-dup
940: doc-over
941: doc-tuck
942: doc-swap
943: doc-rot
944: doc--rot
945: doc-?dup
946: doc-pick
947: doc-roll
948: doc-2drop
949: doc-2nip
950: doc-2dup
951: doc-2over
952: doc-2tuck
953: doc-2swap
954: doc-2rot
955:
956: @node Floating point stack, Return stack, Data stack, Stack Manipulation
957: @subsection Floating point stack
958: doc-fdrop
959: doc-fnip
960: doc-fdup
961: doc-fover
962: doc-ftuck
963: doc-fswap
964: doc-frot
965:
966: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
967: @subsection Return stack
968: doc->r
969: doc-r>
970: doc-r@
971: doc-rdrop
972: doc-2>r
973: doc-2r>
974: doc-2r@
975: doc-2rdrop
976:
977: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
978: @subsection Locals stack
979:
980: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
981: @subsection Stack pointer manipulation
982: doc-sp@
983: doc-sp!
984: doc-fp@
985: doc-fp!
986: doc-rp@
987: doc-rp!
988: doc-lp@
989: doc-lp!
990:
991: @node Memory access, Control Structures, Stack Manipulation, Words
992: @section Memory access
993:
994: @menu
995: * Stack-Memory transfers::
996: * Address arithmetic::
997: * Memory block access::
998: @end menu
999:
1000: @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
1001: @subsection Stack-Memory transfers
1002:
1003: doc-@
1004: doc-!
1005: doc-+!
1006: doc-c@
1007: doc-c!
1008: doc-2@
1009: doc-2!
1010: doc-f@
1011: doc-f!
1012: doc-sf@
1013: doc-sf!
1014: doc-df@
1015: doc-df!
1016:
1017: @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
1018: @subsection Address arithmetic
1019:
1020: ANS Forth does not specify the sizes of the data types. Instead, it
1021: offers a number of words for computing sizes and doing address
1022: arithmetic. Basically, address arithmetic is performed in terms of
1023: address units (aus); on most systems the address unit is one byte. Note
1024: that a character may have more than one au, so @code{chars} is no noop
1025: (on systems where it is a noop, it compiles to nothing).
1026:
1027: ANS Forth also defines words for aligning addresses for specific
1028: addresses. Many computers require that accesses to specific data types
1029: must only occur at specific addresses; e.g., that cells may only be
1030: accessed at addresses divisible by 4. Even if a machine allows unaligned
1031: accesses, it can usually perform aligned accesses faster.
1032:
1033: For the performance-conscious: alignment operations are usually only
1034: necessary during the definition of a data structure, not during the
1035: (more frequent) accesses to it.
1036:
1037: ANS Forth defines no words for character-aligning addresses. This is not
1038: an oversight, but reflects the fact that addresses that are not
1039: char-aligned have no use in the standard and therefore will not be
1040: created.
1041:
1042: The standard guarantees that addresses returned by @code{CREATE}d words
1043: are cell-aligned; in addition, Gforth guarantees that these addresses
1044: are aligned for all purposes.
1045:
1046: Note that the standard defines a word @code{char}, which has nothing to
1047: do with address arithmetic.
1048:
1049: doc-chars
1050: doc-char+
1051: doc-cells
1052: doc-cell+
1053: doc-align
1054: doc-aligned
1055: doc-floats
1056: doc-float+
1057: doc-falign
1058: doc-faligned
1059: doc-sfloats
1060: doc-sfloat+
1061: doc-sfalign
1062: doc-sfaligned
1063: doc-dfloats
1064: doc-dfloat+
1065: doc-dfalign
1066: doc-dfaligned
1067: doc-maxalign
1068: doc-maxaligned
1069: doc-cfalign
1070: doc-cfaligned
1071: doc-address-unit-bits
1072:
1073: @node Memory block access, , Address arithmetic, Memory access
1074: @subsection Memory block access
1075:
1076: doc-move
1077: doc-erase
1078:
1079: While the previous words work on address units, the rest works on
1080: characters.
1081:
1082: doc-cmove
1083: doc-cmove>
1084: doc-fill
1085: doc-blank
1086:
1087: @node Control Structures, Locals, Memory access, Words
1088: @section Control Structures
1089:
1090: Control structures in Forth cannot be used in interpret state, only in
1091: compile state, i.e., in a colon definition. We do not like this
1092: limitation, but have not seen a satisfying way around it yet, although
1093: many schemes have been proposed.
1094:
1095: @menu
1096: * Selection::
1097: * Simple Loops::
1098: * Counted Loops::
1099: * Arbitrary control structures::
1100: * Calls and returns::
1101: * Exception Handling::
1102: @end menu
1103:
1104: @node Selection, Simple Loops, Control Structures, Control Structures
1105: @subsection Selection
1106:
1107: @example
1108: @var{flag}
1109: IF
1110: @var{code}
1111: ENDIF
1112: @end example
1113: or
1114: @example
1115: @var{flag}
1116: IF
1117: @var{code1}
1118: ELSE
1119: @var{code2}
1120: ENDIF
1121: @end example
1122:
1123: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
1124: standard, and @code{ENDIF} is not, although it is quite popular. We
1125: recommend using @code{ENDIF}, because it is less confusing for people
1126: who also know other languages (and is not prone to reinforcing negative
1127: prejudices against Forth in these people). Adding @code{ENDIF} to a
1128: system that only supplies @code{THEN} is simple:
1129: @example
1130: : endif POSTPONE then ; immediate
1131: @end example
1132:
1133: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
1134: (adv.)} has the following meanings:
1135: @quotation
1136: ... 2b: following next after in order ... 3d: as a necessary consequence
1137: (if you were there, then you saw them).
1138: @end quotation
1139: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
1140: and many other programming languages has the meaning 3d.]
1141:
1142: We also provide the words @code{?dup-if} and @code{?dup-0=-if}, so you
1143: can avoid using @code{?dup}.
1144:
1145: @example
1146: @var{n}
1147: CASE
1148: @var{n1} OF @var{code1} ENDOF
1149: @var{n2} OF @var{code2} ENDOF
1150: @dots{}
1151: ENDCASE
1152: @end example
1153:
1154: Executes the first @var{codei}, where the @var{ni} is equal to
1155: @var{n}. A default case can be added by simply writing the code after
1156: the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
1157: but must not consume it.
1158:
1159: @node Simple Loops, Counted Loops, Selection, Control Structures
1160: @subsection Simple Loops
1161:
1162: @example
1163: BEGIN
1164: @var{code1}
1165: @var{flag}
1166: WHILE
1167: @var{code2}
1168: REPEAT
1169: @end example
1170:
1171: @var{code1} is executed and @var{flag} is computed. If it is true,
1172: @var{code2} is executed and the loop is restarted; If @var{flag} is false, execution continues after the @code{REPEAT}.
1173:
1174: @example
1175: BEGIN
1176: @var{code}
1177: @var{flag}
1178: UNTIL
1179: @end example
1180:
1181: @var{code} is executed. The loop is restarted if @code{flag} is false.
1182:
1183: @example
1184: BEGIN
1185: @var{code}
1186: AGAIN
1187: @end example
1188:
1189: This is an endless loop.
1190:
1191: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
1192: @subsection Counted Loops
1193:
1194: The basic counted loop is:
1195: @example
1196: @var{limit} @var{start}
1197: ?DO
1198: @var{body}
1199: LOOP
1200: @end example
1201:
1202: This performs one iteration for every integer, starting from @var{start}
1203: and up to, but excluding @var{limit}. The counter, aka index, can be
1204: accessed with @code{i}. E.g., the loop
1205: @example
1206: 10 0 ?DO
1207: i .
1208: LOOP
1209: @end example
1210: prints
1211: @example
1212: 0 1 2 3 4 5 6 7 8 9
1213: @end example
1214: The index of the innermost loop can be accessed with @code{i}, the index
1215: of the next loop with @code{j}, and the index of the third loop with
1216: @code{k}.
1217:
1218: The loop control data are kept on the return stack, so there are some
1219: restrictions on mixing return stack accesses and counted loop
1220: words. E.g., if you put values on the return stack outside the loop, you
1221: cannot read them inside the loop. If you put values on the return stack
1222: within a loop, you have to remove them before the end of the loop and
1223: before accessing the index of the loop.
1224:
1225: There are several variations on the counted loop:
1226:
1227: @code{LEAVE} leaves the innermost counted loop immediately.
1228:
1229: If @var{start} is greater than @var{limit}, a @code{?DO} loop is entered
1230: (and @code{LOOP} iterates until they become equal by wrap-around
1231: arithmetic). This behaviour is usually not what you want. Therefore,
1232: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1233: @code{?DO}), which do not enter the loop if @var{start} is greater than
1234: @var{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1235: unsigned loop parameters. These words can be implemented easily on
1236: standard systems, so using them does not make your programs hard to
1237: port; e.g.:
1238: @example
1239: : +DO ( compile-time: -- do-sys; run-time: n1 n2 -- )
1240: POSTPONE over POSTPONE min POSTPONE ?DO ; immediate
1241: @end example
1242:
1243: @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the
1244: index by @var{n} instead of by 1. The loop is terminated when the border
1245: between @var{limit-1} and @var{limit} is crossed. E.g.:
1246:
1247: @code{4 0 +DO i . 2 +LOOP} prints @code{0 2}
1248:
1249: @code{4 1 +DO i . 2 +LOOP} prints @code{1 3}
1250:
1251: The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:
1252:
1253: @code{-1 0 ?DO i . -1 +LOOP} prints @code{0 -1}
1254:
1255: @code{ 0 0 ?DO i . -1 +LOOP} prints nothing
1256:
1257: Therefore we recommend avoiding @code{@var{n} +LOOP} with negative
1258: @var{n}. One alternative is @code{@var{u} -LOOP}, which reduces the
1259: index by @var{u} each iteration. The loop is terminated when the border
1260: between @var{limit+1} and @var{limit} is crossed. Gforth also provides
1261: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
1262:
1263: @code{-2 0 -DO i . 1 -LOOP} prints @code{0 -1}
1264:
1265: @code{-1 0 -DO i . 1 -LOOP} prints @code{0}
1266:
1267: @code{ 0 0 -DO i . 1 -LOOP} prints nothing
1268:
1269: Another alternative is @code{@var{n} S+LOOP}, where the negative
1270: case behaves symmetrical to the positive case:
1271:
1272: @code{-2 0 -DO i . -1 S+LOOP} prints @code{0 -1}
1273:
1274: The loop is terminated when the border between @var{limit@minus{}sgn(n)}
1275: and @var{limit} is crossed. Unfortunately, neither @code{-LOOP} nor
1276: @code{S+LOOP} are part of the ANS Forth standard, and they are not easy
1277: to implement using standard words. If you want to write standard
1278: programs, just avoid counting down.
1279:
1280: @code{?DO} can also be replaced by @code{DO}. @code{DO} always enters
1281: the loop, independent of the loop parameters. Do not use @code{DO}, even
1282: if you know that the loop is entered in any case. Such knowledge tends
1283: to become invalid during maintenance of a program, and then the
1284: @code{DO} will make trouble.
1285:
1286: @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via
1287: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
1288: return stack so @code{EXIT} can get to its return address.
1289:
1290: Another counted loop is
1291: @example
1292: @var{n}
1293: FOR
1294: @var{body}
1295: NEXT
1296: @end example
1297: This is the preferred loop of native code compiler writers who are too
1298: lazy to optimize @code{?DO} loops properly. In Gforth, this loop
1299: iterates @var{n+1} times; @code{i} produces values starting with @var{n}
1300: and ending with 0. Other Forth systems may behave differently, even if
1301: they support @code{FOR} loops.
1302:
1303: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
1304: @subsection Arbitrary control structures
1305:
1306: ANS Forth permits and supports using control structures in a non-nested
1307: way. Information about incomplete control structures is stored on the
1308: control-flow stack. This stack may be implemented on the Forth data
1309: stack, and this is what we have done in Gforth.
1310:
1311: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
1312: entry represents a backward branch target. A few words are the basis for
1313: building any control structure possible (except control structures that
1314: need storage, like calls, coroutines, and backtracking).
1315:
1316: doc-if
1317: doc-ahead
1318: doc-then
1319: doc-begin
1320: doc-until
1321: doc-again
1322: doc-cs-pick
1323: doc-cs-roll
1324:
1325: On many systems control-flow stack items take one word, in Gforth they
1326: currently take three (this may change in the future). Therefore it is a
1327: really good idea to manipulate the control flow stack with
1328: @code{cs-pick} and @code{cs-roll}, not with data stack manipulation
1329: words.
1330:
1331: Some standard control structure words are built from these words:
1332:
1333: doc-else
1334: doc-while
1335: doc-repeat
1336:
1337: Counted loop words constitute a separate group of words:
1338:
1339: doc-?do
1340: doc-+do
1341: doc-u+do
1342: doc--do
1343: doc-u-do
1344: doc-do
1345: doc-for
1346: doc-loop
1347: doc-s+loop
1348: doc-+loop
1349: doc--loop
1350: doc-next
1351: doc-leave
1352: doc-?leave
1353: doc-unloop
1354: doc-done
1355:
1356: The standard does not allow using @code{cs-pick} and @code{cs-roll} on
1357: @i{do-sys}. Our system allows it, but it's your job to ensure that for
1358: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
1359: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
1360: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
1361: resolved (by using one of the loop-ending words or @code{DONE}).
1362:
1363: Another group of control structure words are
1364:
1365: doc-case
1366: doc-endcase
1367: doc-of
1368: doc-endof
1369:
1370: @i{case-sys} and @i{of-sys} cannot be processed using @code{cs-pick} and
1371: @code{cs-roll}.
1372:
1373: @subsubsection Programming Style
1374:
1375: In order to ensure readability we recommend that you do not create
1376: arbitrary control structures directly, but define new control structure
1377: words for the control structure you want and use these words in your
1378: program.
1379:
1380: E.g., instead of writing
1381:
1382: @example
1383: begin
1384: ...
1385: if [ 1 cs-roll ]
1386: ...
1387: again then
1388: @end example
1389:
1390: we recommend defining control structure words, e.g.,
1391:
1392: @example
1393: : while ( dest -- orig dest )
1394: POSTPONE if
1395: 1 cs-roll ; immediate
1396:
1397: : repeat ( orig dest -- )
1398: POSTPONE again
1399: POSTPONE then ; immediate
1400: @end example
1401:
1402: and then using these to create the control structure:
1403:
1404: @example
1405: begin
1406: ...
1407: while
1408: ...
1409: repeat
1410: @end example
1411:
1412: That's much easier to read, isn't it? Of course, @code{BEGIN} and
1413: @code{WHILE} are predefined, so in this example it would not be
1414: necessary to define them.
1415:
1416: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
1417: @subsection Calls and returns
1418:
1419: A definition can be called simply be writing the name of the
1420: definition. When the end of the definition is reached, it returns. An
1421: earlier return can be forced using
1422:
1423: doc-exit
1424:
1425: Don't forget to clean up the return stack and @code{UNLOOP} any
1426: outstanding @code{?DO}...@code{LOOP}s before @code{EXIT}ing. The
1427: primitive compiled by @code{EXIT} is
1428:
1429: doc-;s
1430:
1431: @node Exception Handling, , Calls and returns, Control Structures
1432: @subsection Exception Handling
1433:
1434: doc-catch
1435: doc-throw
1436:
1437: @node Locals, Defining Words, Control Structures, Words
1438: @section Locals
1439:
1440: Local variables can make Forth programming more enjoyable and Forth
1441: programs easier to read. Unfortunately, the locals of ANS Forth are
1442: laden with restrictions. Therefore, we provide not only the ANS Forth
1443: locals wordset, but also our own, more powerful locals wordset (we
1444: implemented the ANS Forth locals wordset through our locals wordset).
1445:
1446: @menu
1447: * Gforth locals::
1448: * ANS Forth locals::
1449: @end menu
1450:
1451: @node Gforth locals, ANS Forth locals, Locals, Locals
1452: @subsection Gforth locals
1453:
1454: Locals can be defined with
1455:
1456: @example
1457: @{ local1 local2 ... -- comment @}
1458: @end example
1459: or
1460: @example
1461: @{ local1 local2 ... @}
1462: @end example
1463:
1464: E.g.,
1465: @example
1466: : max @{ n1 n2 -- n3 @}
1467: n1 n2 > if
1468: n1
1469: else
1470: n2
1471: endif ;
1472: @end example
1473:
1474: The similarity of locals definitions with stack comments is intended. A
1475: locals definition often replaces the stack comment of a word. The order
1476: of the locals corresponds to the order in a stack comment and everything
1477: after the @code{--} is really a comment.
1478:
1479: This similarity has one disadvantage: It is too easy to confuse locals
1480: declarations with stack comments, causing bugs and making them hard to
1481: find. However, this problem can be avoided by appropriate coding
1482: conventions: Do not use both notations in the same program. If you do,
1483: they should be distinguished using additional means, e.g. by position.
1484:
1485: The name of the local may be preceded by a type specifier, e.g.,
1486: @code{F:} for a floating point value:
1487:
1488: @example
1489: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
1490: \ complex multiplication
1491: Ar Br f* Ai Bi f* f-
1492: Ar Bi f* Ai Br f* f+ ;
1493: @end example
1494:
1495: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
1496: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
1497: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
1498: with @code{W:}, @code{D:} etc.) produces its value and can be changed
1499: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
1500: produces its address (which becomes invalid when the variable's scope is
1501: left). E.g., the standard word @code{emit} can be defined in therms of
1502: @code{type} like this:
1503:
1504: @example
1505: : emit @{ C^ char* -- @}
1506: char* 1 type ;
1507: @end example
1508:
1509: A local without type specifier is a @code{W:} local. Both flavours of
1510: locals are initialized with values from the data or FP stack.
1511:
1512: Currently there is no way to define locals with user-defined data
1513: structures, but we are working on it.
1514:
1515: Gforth allows defining locals everywhere in a colon definition. This
1516: poses the following questions:
1517:
1518: @menu
1519: * Where are locals visible by name?::
1520: * How long do locals live?::
1521: * Programming Style::
1522: * Implementation::
1523: @end menu
1524:
1525: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
1526: @subsubsection Where are locals visible by name?
1527:
1528: Basically, the answer is that locals are visible where you would expect
1529: it in block-structured languages, and sometimes a little longer. If you
1530: want to restrict the scope of a local, enclose its definition in
1531: @code{SCOPE}...@code{ENDSCOPE}.
1532:
1533: doc-scope
1534: doc-endscope
1535:
1536: These words behave like control structure words, so you can use them
1537: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
1538: arbitrary ways.
1539:
1540: If you want a more exact answer to the visibility question, here's the
1541: basic principle: A local is visible in all places that can only be
1542: reached through the definition of the local@footnote{In compiler
1543: construction terminology, all places dominated by the definition of the
1544: local.}. In other words, it is not visible in places that can be reached
1545: without going through the definition of the local. E.g., locals defined
1546: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
1547: defined in @code{BEGIN}...@code{UNTIL} are visible after the
1548: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1549:
1550: The reasoning behind this solution is: We want to have the locals
1551: visible as long as it is meaningful. The user can always make the
1552: visibility shorter by using explicit scoping. In a place that can
1553: only be reached through the definition of a local, the meaning of a
1554: local name is clear. In other places it is not: How is the local
1555: initialized at the control flow path that does not contain the
1556: definition? Which local is meant, if the same name is defined twice in
1557: two independent control flow paths?
1558:
1559: This should be enough detail for nearly all users, so you can skip the
1560: rest of this section. If you relly must know all the gory details and
1561: options, read on.
1562:
1563: In order to implement this rule, the compiler has to know which places
1564: are unreachable. It knows this automatically after @code{AHEAD},
1565: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
1566: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
1567: compiler that the control flow never reaches that place. If
1568: @code{UNREACHABLE} is not used where it could, the only consequence is
1569: that the visibility of some locals is more limited than the rule above
1570: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
1571: lie to the compiler), buggy code will be produced.
1572:
1573: Another problem with this rule is that at @code{BEGIN}, the compiler
1574: does not know which locals will be visible on the incoming
1575: back-edge. All problems discussed in the following are due to this
1576: ignorance of the compiler (we discuss the problems using @code{BEGIN}
1577: loops as examples; the discussion also applies to @code{?DO} and other
1578: loops). Perhaps the most insidious example is:
1579: @example
1580: AHEAD
1581: BEGIN
1582: x
1583: [ 1 CS-ROLL ] THEN
1584: @{ x @}
1585: ...
1586: UNTIL
1587: @end example
1588:
1589: This should be legal according to the visibility rule. The use of
1590: @code{x} can only be reached through the definition; but that appears
1591: textually below the use.
1592:
1593: From this example it is clear that the visibility rules cannot be fully
1594: implemented without major headaches. Our implementation treats common
1595: cases as advertised and the exceptions are treated in a safe way: The
1596: compiler makes a reasonable guess about the locals visible after a
1597: @code{BEGIN}; if it is too pessimistic, the
1598: user will get a spurious error about the local not being defined; if the
1599: compiler is too optimistic, it will notice this later and issue a
1600: warning. In the case above the compiler would complain about @code{x}
1601: being undefined at its use. You can see from the obscure examples in
1602: this section that it takes quite unusual control structures to get the
1603: compiler into trouble, and even then it will often do fine.
1604:
1605: If the @code{BEGIN} is reachable from above, the most optimistic guess
1606: is that all locals visible before the @code{BEGIN} will also be
1607: visible after the @code{BEGIN}. This guess is valid for all loops that
1608: are entered only through the @code{BEGIN}, in particular, for normal
1609: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
1610: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
1611: compiler. When the branch to the @code{BEGIN} is finally generated by
1612: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
1613: warns the user if it was too optimisitic:
1614: @example
1615: IF
1616: @{ x @}
1617: BEGIN
1618: \ x ?
1619: [ 1 cs-roll ] THEN
1620: ...
1621: UNTIL
1622: @end example
1623:
1624: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
1625: optimistically assumes that it lives until the @code{THEN}. It notices
1626: this difference when it compiles the @code{UNTIL} and issues a
1627: warning. The user can avoid the warning, and make sure that @code{x}
1628: is not used in the wrong area by using explicit scoping:
1629: @example
1630: IF
1631: SCOPE
1632: @{ x @}
1633: ENDSCOPE
1634: BEGIN
1635: [ 1 cs-roll ] THEN
1636: ...
1637: UNTIL
1638: @end example
1639:
1640: Since the guess is optimistic, there will be no spurious error messages
1641: about undefined locals.
1642:
1643: If the @code{BEGIN} is not reachable from above (e.g., after
1644: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
1645: optimistic guess, as the locals visible after the @code{BEGIN} may be
1646: defined later. Therefore, the compiler assumes that no locals are
1647: visible after the @code{BEGIN}. However, the user can use
1648: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
1649: visible at the BEGIN as at the point where the top control-flow stack
1650: item was created.
1651:
1652: doc-assume-live
1653:
1654: E.g.,
1655: @example
1656: @{ x @}
1657: AHEAD
1658: ASSUME-LIVE
1659: BEGIN
1660: x
1661: [ 1 CS-ROLL ] THEN
1662: ...
1663: UNTIL
1664: @end example
1665:
1666: Other cases where the locals are defined before the @code{BEGIN} can be
1667: handled by inserting an appropriate @code{CS-ROLL} before the
1668: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
1669: behind the @code{ASSUME-LIVE}).
1670:
1671: Cases where locals are defined after the @code{BEGIN} (but should be
1672: visible immediately after the @code{BEGIN}) can only be handled by
1673: rearranging the loop. E.g., the ``most insidious'' example above can be
1674: arranged into:
1675: @example
1676: BEGIN
1677: @{ x @}
1678: ... 0=
1679: WHILE
1680: x
1681: REPEAT
1682: @end example
1683:
1684: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
1685: @subsubsection How long do locals live?
1686:
1687: The right answer for the lifetime question would be: A local lives at
1688: least as long as it can be accessed. For a value-flavoured local this
1689: means: until the end of its visibility. However, a variable-flavoured
1690: local could be accessed through its address far beyond its visibility
1691: scope. Ultimately, this would mean that such locals would have to be
1692: garbage collected. Since this entails un-Forth-like implementation
1693: complexities, I adopted the same cowardly solution as some other
1694: languages (e.g., C): The local lives only as long as it is visible;
1695: afterwards its address is invalid (and programs that access it
1696: afterwards are erroneous).
1697:
1698: @node Programming Style, Implementation, How long do locals live?, Gforth locals
1699: @subsubsection Programming Style
1700:
1701: The freedom to define locals anywhere has the potential to change
1702: programming styles dramatically. In particular, the need to use the
1703: return stack for intermediate storage vanishes. Moreover, all stack
1704: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
1705: determined arguments) can be eliminated: If the stack items are in the
1706: wrong order, just write a locals definition for all of them; then
1707: write the items in the order you want.
1708:
1709: This seems a little far-fetched and eliminating stack manipulations is
1710: unlikely to become a conscious programming objective. Still, the number
1711: of stack manipulations will be reduced dramatically if local variables
1712: are used liberally (e.g., compare @code{max} in @ref{Gforth locals} with
1713: a traditional implementation of @code{max}).
1714:
1715: This shows one potential benefit of locals: making Forth programs more
1716: readable. Of course, this benefit will only be realized if the
1717: programmers continue to honour the principle of factoring instead of
1718: using the added latitude to make the words longer.
1719:
1720: Using @code{TO} can and should be avoided. Without @code{TO},
1721: every value-flavoured local has only a single assignment and many
1722: advantages of functional languages apply to Forth. I.e., programs are
1723: easier to analyse, to optimize and to read: It is clear from the
1724: definition what the local stands for, it does not turn into something
1725: different later.
1726:
1727: E.g., a definition using @code{TO} might look like this:
1728: @example
1729: : strcmp @{ addr1 u1 addr2 u2 -- n @}
1730: u1 u2 min 0
1731: ?do
1732: addr1 c@ addr2 c@ - ?dup
1733: if
1734: unloop exit
1735: then
1736: addr1 char+ TO addr1
1737: addr2 char+ TO addr2
1738: loop
1739: u1 u2 - ;
1740: @end example
1741: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
1742: every loop iteration. @code{strcmp} is a typical example of the
1743: readability problems of using @code{TO}. When you start reading
1744: @code{strcmp}, you think that @code{addr1} refers to the start of the
1745: string. Only near the end of the loop you realize that it is something
1746: else.
1747:
1748: This can be avoided by defining two locals at the start of the loop that
1749: are initialized with the right value for the current iteration.
1750: @example
1751: : strcmp @{ addr1 u1 addr2 u2 -- n @}
1752: addr1 addr2
1753: u1 u2 min 0
1754: ?do @{ s1 s2 @}
1755: s1 c@ s2 c@ - ?dup
1756: if
1757: unloop exit
1758: then
1759: s1 char+ s2 char+
1760: loop
1761: 2drop
1762: u1 u2 - ;
1763: @end example
1764: Here it is clear from the start that @code{s1} has a different value
1765: in every loop iteration.
1766:
1767: @node Implementation, , Programming Style, Gforth locals
1768: @subsubsection Implementation
1769:
1770: Gforth uses an extra locals stack. The most compelling reason for
1771: this is that the return stack is not float-aligned; using an extra stack
1772: also eliminates the problems and restrictions of using the return stack
1773: as locals stack. Like the other stacks, the locals stack grows toward
1774: lower addresses. A few primitives allow an efficient implementation:
1775:
1776: doc-@local#
1777: doc-f@local#
1778: doc-laddr#
1779: doc-lp+!#
1780: doc-lp!
1781: doc->l
1782: doc-f>l
1783:
1784: In addition to these primitives, some specializations of these
1785: primitives for commonly occurring inline arguments are provided for
1786: efficiency reasons, e.g., @code{@@local0} as specialization of
1787: @code{@@local#} for the inline argument 0. The following compiling words
1788: compile the right specialized version, or the general version, as
1789: appropriate:
1790:
1791: doc-compile-@local
1792: doc-compile-f@local
1793: doc-compile-lp+!
1794:
1795: Combinations of conditional branches and @code{lp+!#} like
1796: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
1797: is taken) are provided for efficiency and correctness in loops.
1798:
1799: A special area in the dictionary space is reserved for keeping the
1800: local variable names. @code{@{} switches the dictionary pointer to this
1801: area and @code{@}} switches it back and generates the locals
1802: initializing code. @code{W:} etc.@ are normal defining words. This
1803: special area is cleared at the start of every colon definition.
1804:
1805: A special feature of Gforth's dictionary is used to implement the
1806: definition of locals without type specifiers: every wordlist (aka
1807: vocabulary) has its own methods for searching
1808: etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
1809: with a special search method: When it is searched for a word, it
1810: actually creates that word using @code{W:}. @code{@{} changes the search
1811: order to first search the wordlist containing @code{@}}, @code{W:} etc.,
1812: and then the wordlist for defining locals without type specifiers.
1813:
1814: The lifetime rules support a stack discipline within a colon
1815: definition: The lifetime of a local is either nested with other locals
1816: lifetimes or it does not overlap them.
1817:
1818: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
1819: pointer manipulation is generated. Between control structure words
1820: locals definitions can push locals onto the locals stack. @code{AGAIN}
1821: is the simplest of the other three control flow words. It has to
1822: restore the locals stack depth of the corresponding @code{BEGIN}
1823: before branching. The code looks like this:
1824: @format
1825: @code{lp+!#} current-locals-size @minus{} dest-locals-size
1826: @code{branch} <begin>
1827: @end format
1828:
1829: @code{UNTIL} is a little more complicated: If it branches back, it
1830: must adjust the stack just like @code{AGAIN}. But if it falls through,
1831: the locals stack must not be changed. The compiler generates the
1832: following code:
1833: @format
1834: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
1835: @end format
1836: The locals stack pointer is only adjusted if the branch is taken.
1837:
1838: @code{THEN} can produce somewhat inefficient code:
1839: @format
1840: @code{lp+!#} current-locals-size @minus{} orig-locals-size
1841: <orig target>:
1842: @code{lp+!#} orig-locals-size @minus{} new-locals-size
1843: @end format
1844: The second @code{lp+!#} adjusts the locals stack pointer from the
1845: level at the @var{orig} point to the level after the @code{THEN}. The
1846: first @code{lp+!#} adjusts the locals stack pointer from the current
1847: level to the level at the orig point, so the complete effect is an
1848: adjustment from the current level to the right level after the
1849: @code{THEN}.
1850:
1851: In a conventional Forth implementation a dest control-flow stack entry
1852: is just the target address and an orig entry is just the address to be
1853: patched. Our locals implementation adds a wordlist to every orig or dest
1854: item. It is the list of locals visible (or assumed visible) at the point
1855: described by the entry. Our implementation also adds a tag to identify
1856: the kind of entry, in particular to differentiate between live and dead
1857: (reachable and unreachable) orig entries.
1858:
1859: A few unusual operations have to be performed on locals wordlists:
1860:
1861: doc-common-list
1862: doc-sub-list?
1863: doc-list-size
1864:
1865: Several features of our locals wordlist implementation make these
1866: operations easy to implement: The locals wordlists are organised as
1867: linked lists; the tails of these lists are shared, if the lists
1868: contain some of the same locals; and the address of a name is greater
1869: than the address of the names behind it in the list.
1870:
1871: Another important implementation detail is the variable
1872: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
1873: determine if they can be reached directly or only through the branch
1874: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
1875: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
1876: definition, by @code{BEGIN} and usually by @code{THEN}.
1877:
1878: Counted loops are similar to other loops in most respects, but
1879: @code{LEAVE} requires special attention: It performs basically the same
1880: service as @code{AHEAD}, but it does not create a control-flow stack
1881: entry. Therefore the information has to be stored elsewhere;
1882: traditionally, the information was stored in the target fields of the
1883: branches created by the @code{LEAVE}s, by organizing these fields into a
1884: linked list. Unfortunately, this clever trick does not provide enough
1885: space for storing our extended control flow information. Therefore, we
1886: introduce another stack, the leave stack. It contains the control-flow
1887: stack entries for all unresolved @code{LEAVE}s.
1888:
1889: Local names are kept until the end of the colon definition, even if
1890: they are no longer visible in any control-flow path. In a few cases
1891: this may lead to increased space needs for the locals name area, but
1892: usually less than reclaiming this space would cost in code size.
1893:
1894:
1895: @node ANS Forth locals, , Gforth locals, Locals
1896: @subsection ANS Forth locals
1897:
1898: The ANS Forth locals wordset does not define a syntax for locals, but
1899: words that make it possible to define various syntaxes. One of the
1900: possible syntaxes is a subset of the syntax we used in the Gforth locals
1901: wordset, i.e.:
1902:
1903: @example
1904: @{ local1 local2 ... -- comment @}
1905: @end example
1906: or
1907: @example
1908: @{ local1 local2 ... @}
1909: @end example
1910:
1911: The order of the locals corresponds to the order in a stack comment. The
1912: restrictions are:
1913:
1914: @itemize @bullet
1915: @item
1916: Locals can only be cell-sized values (no type specifiers are allowed).
1917: @item
1918: Locals can be defined only outside control structures.
1919: @item
1920: Locals can interfere with explicit usage of the return stack. For the
1921: exact (and long) rules, see the standard. If you don't use return stack
1922: accessing words in a definition using locals, you will be all right. The
1923: purpose of this rule is to make locals implementation on the return
1924: stack easier.
1925: @item
1926: The whole definition must be in one line.
1927: @end itemize
1928:
1929: Locals defined in this way behave like @code{VALUE}s
1930: (@xref{Values}). I.e., they are initialized from the stack. Using their
1931: name produces their value. Their value can be changed using @code{TO}.
1932:
1933: Since this syntax is supported by Gforth directly, you need not do
1934: anything to use it. If you want to port a program using this syntax to
1935: another ANS Forth system, use @file{anslocal.fs} to implement the syntax
1936: on the other system.
1937:
1938: Note that a syntax shown in the standard, section A.13 looks
1939: similar, but is quite different in having the order of locals
1940: reversed. Beware!
1941:
1942: The ANS Forth locals wordset itself consists of the following word
1943:
1944: doc-(local)
1945:
1946: The ANS Forth locals extension wordset defines a syntax, but it is so
1947: awful that we strongly recommend not to use it. We have implemented this
1948: syntax to make porting to Gforth easy, but do not document it here. The
1949: problem with this syntax is that the locals are defined in an order
1950: reversed with respect to the standard stack comment notation, making
1951: programs harder to read, and easier to misread and miswrite. The only
1952: merit of this syntax is that it is easy to implement using the ANS Forth
1953: locals wordset.
1954:
1955: @node Defining Words, Wordlists, Locals, Words
1956: @section Defining Words
1957:
1958: @menu
1959: * Values::
1960: @end menu
1961:
1962: @node Values, , Defining Words, Defining Words
1963: @subsection Values
1964:
1965: @node Wordlists, Files, Defining Words, Words
1966: @section Wordlists
1967:
1968: @node Files, Blocks, Wordlists, Words
1969: @section Files
1970:
1971: @node Blocks, Other I/O, Files, Words
1972: @section Blocks
1973:
1974: @node Other I/O, Programming Tools, Blocks, Words
1975: @section Other I/O
1976:
1977: @node Programming Tools, Assembler and Code words, Other I/O, Words
1978: @section Programming Tools
1979:
1980: @menu
1981: * Debugging:: Simple and quick.
1982: * Assertions:: Making your programs self-checking.
1983: @end menu
1984:
1985: @node Debugging, Assertions, Programming Tools, Programming Tools
1986: @subsection Debugging
1987:
1988: The simple debugging aids provided in @file{debugging.fs}
1989: are meant to support a different style of debugging than the
1990: tracing/stepping debuggers used in languages with long turn-around
1991: times.
1992:
1993: A much better (faster) way in fast-compilig languages is to add
1994: printing code at well-selected places, let the program run, look at
1995: the output, see where things went wrong, add more printing code, etc.,
1996: until the bug is found.
1997:
1998: The word @code{~~} is easy to insert. It just prints debugging
1999: information (by default the source location and the stack contents). It
2000: is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
2001: query-replace them with nothing). The deferred words
2002: @code{printdebugdata} and @code{printdebugline} control the output of
2003: @code{~~}. The default source location output format works well with
2004: Emacs' compilation mode, so you can step through the program at the
2005: source level using @kbd{C-x `} (the advantage over a stepping debugger
2006: is that you can step in any direction and you know where the crash has
2007: happened or where the strange data has occurred).
2008:
2009: Note that the default actions clobber the contents of the pictured
2010: numeric output string, so you should not use @code{~~}, e.g., between
2011: @code{<#} and @code{#>}.
2012:
2013: doc-~~
2014: doc-printdebugdata
2015: doc-printdebugline
2016:
2017: @node Assertions, , Debugging, Programming Tools
2018: @subsection Assertions
2019:
2020: It is a good idea to make your programs self-checking, in particular, if
2021: you use an assumption (e.g., that a certain field of a data structure is
2022: never zero) that may become wrong during maintenance. Gforth supports
2023: assertions for this purpose. They are used like this:
2024:
2025: @example
2026: assert( @var{flag} )
2027: @end example
2028:
2029: The code between @code{assert(} and @code{)} should compute a flag, that
2030: should be true if everything is alright and false otherwise. It should
2031: not change anything else on the stack. The overall stack effect of the
2032: assertion is @code{( -- )}. E.g.
2033:
2034: @example
2035: assert( 1 1 + 2 = ) \ what we learn in school
2036: assert( dup 0<> ) \ assert that the top of stack is not zero
2037: assert( false ) \ this code should not be reached
2038: @end example
2039:
2040: The need for assertions is different at different times. During
2041: debugging, we want more checking, in production we sometimes care more
2042: for speed. Therefore, assertions can be turned off, i.e., the assertion
2043: becomes a comment. Depending on the importance of an assertion and the
2044: time it takes to check it, you may want to turn off some assertions and
2045: keep others turned on. Gforth provides several levels of assertions for
2046: this purpose:
2047:
2048: doc-assert0(
2049: doc-assert1(
2050: doc-assert2(
2051: doc-assert3(
2052: doc-assert(
2053: doc-)
2054:
2055: @code{Assert(} is the same as @code{assert1(}. The variable
2056: @code{assert-level} specifies the highest assertions that are turned
2057: on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
2058: @code{assert1(} assertions perform checking, while @code{assert2(} and
2059: @code{assert3(} assertions are treated as comments.
2060:
2061: Note that the @code{assert-level} is evaluated at compile-time, not at
2062: run-time. I.e., you cannot turn assertions on or off at run-time, you
2063: have to set the @code{assert-level} appropriately before compiling a
2064: piece of code. You can compile several pieces of code at several
2065: @code{assert-level}s (e.g., a trusted library at level 1 and newly
2066: written code at level 3).
2067:
2068: doc-assert-level
2069:
2070: If an assertion fails, a message compatible with Emacs' compilation mode
2071: is produced and the execution is aborted (currently with @code{ABORT"}.
2072: If there is interest, we will introduce a special throw code. But if you
2073: intend to @code{catch} a specific condition, using @code{throw} is
2074: probably more appropriate than an assertion).
2075:
2076: @node Assembler and Code words, Threading Words, Programming Tools, Words
2077: @section Assembler and Code words
2078:
2079: Gforth provides some words for defining primitives (words written in
2080: machine code), and for defining the the machine-code equivalent of
2081: @code{DOES>}-based defining words. However, the machine-independent
2082: nature of Gforth poses a few problems: First of all. Gforth runs on
2083: several architectures, so it can provide no standard assembler. What's
2084: worse is that the register allocation not only depends on the processor,
2085: but also on the gcc version and options used.
2086:
2087: The words Gforth offers encapsulate some system dependences (e.g., the
2088: header structure), so a system-independent assembler may be used in
2089: Gforth. If you do not have an assembler, you can compile machine code
2090: directly with @code{,} and @code{c,}.
2091:
2092: doc-assembler
2093: doc-code
2094: doc-end-code
2095: doc-;code
2096: doc-flush-icache
2097:
2098: If @code{flush-icache} does not work correctly, @code{code} words
2099: etc. will not work (reliably), either.
2100:
2101: These words are rarely used. Therefore they reside in @code{code.fs},
2102: which is usually not loaded (except @code{flush-icache}, which is always
2103: present). You can load them with @code{require code.fs}.
2104:
2105: Another option for implementing normal and defining words efficiently
2106: is: adding the wanted functionality to the source of Gforth. For normal
2107: words you just have to edit @file{primitives}, defining words (for fast
2108: defined words) probably require changes in @file{engine.c},
2109: @file{kernal.fs}, @file{prims2x.fs}, and possibly @file{cross.fs}.
2110:
2111:
2112: @node Threading Words, , Assembler and Code words, Words
2113: @section Threading Words
2114:
2115: These words provide access to code addresses and other threading stuff
2116: in Gforth (and, possibly, other interpretive Forths). It more or less
2117: abstracts away the differences between direct and indirect threading
2118: (and, for direct threading, the machine dependences). However, at
2119: present this wordset is still inclomplete. It is also pretty low-level;
2120: some day it will hopefully be made unnecessary by an internals words set
2121: that abstracts implementation details away completely.
2122:
2123: doc->code-address
2124: doc->does-code
2125: doc-code-address!
2126: doc-does-code!
2127: doc-does-handler!
2128: doc-/does-handler
2129:
2130: The code addresses produced by various defining words are produced by
2131: the following words:
2132:
2133: doc-docol:
2134: doc-docon:
2135: doc-dovar:
2136: doc-douser:
2137: doc-dodefer:
2138: doc-dofield:
2139:
2140: Currently there is no installation-independent way for recogizing words
2141: defined by a @code{CREATE}...@code{DOES>} word; however, once you know
2142: that a word is defined by a @code{CREATE}...@code{DOES>} word, you can
2143: use @code{>DOES-CODE}.
2144:
2145: @node ANS conformance, Model, Words, Top
2146: @chapter ANS conformance
2147:
2148: To the best of our knowledge, Gforth is an
2149:
2150: ANS Forth System
2151: @itemize
2152: @item providing the Core Extensions word set
2153: @item providing the Block word set
2154: @item providing the Block Extensions word set
2155: @item providing the Double-Number word set
2156: @item providing the Double-Number Extensions word set
2157: @item providing the Exception word set
2158: @item providing the Exception Extensions word set
2159: @item providing the Facility word set
2160: @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
2161: @item providing the File Access word set
2162: @item providing the File Access Extensions word set
2163: @item providing the Floating-Point word set
2164: @item providing the Floating-Point Extensions word set
2165: @item providing the Locals word set
2166: @item providing the Locals Extensions word set
2167: @item providing the Memory-Allocation word set
2168: @item providing the Memory-Allocation Extensions word set (that one's easy)
2169: @item providing the Programming-Tools word set
2170: @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
2171: @item providing the Search-Order word set
2172: @item providing the Search-Order Extensions word set
2173: @item providing the String word set
2174: @item providing the String Extensions word set (another easy one)
2175: @end itemize
2176:
2177: In addition, ANS Forth systems are required to document certain
2178: implementation choices. This chapter tries to meet these
2179: requirements. In many cases it gives a way to ask the system for the
2180: information instead of providing the information directly, in
2181: particular, if the information depends on the processor, the operating
2182: system or the installation options chosen, or if they are likely to
2183: change during the maintenance of Gforth.
2184:
2185: @comment The framework for the rest has been taken from pfe.
2186:
2187: @menu
2188: * The Core Words::
2189: * The optional Block word set::
2190: * The optional Double Number word set::
2191: * The optional Exception word set::
2192: * The optional Facility word set::
2193: * The optional File-Access word set::
2194: * The optional Floating-Point word set::
2195: * The optional Locals word set::
2196: * The optional Memory-Allocation word set::
2197: * The optional Programming-Tools word set::
2198: * The optional Search-Order word set::
2199: @end menu
2200:
2201:
2202: @c =====================================================================
2203: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
2204: @comment node-name, next, previous, up
2205: @section The Core Words
2206: @c =====================================================================
2207:
2208: @menu
2209: * core-idef:: Implementation Defined Options
2210: * core-ambcond:: Ambiguous Conditions
2211: * core-other:: Other System Documentation
2212: @end menu
2213:
2214: @c ---------------------------------------------------------------------
2215: @node core-idef, core-ambcond, The Core Words, The Core Words
2216: @subsection Implementation Defined Options
2217: @c ---------------------------------------------------------------------
2218:
2219: @table @i
2220:
2221: @item (Cell) aligned addresses:
2222: processor-dependent. Gforth's alignment words perform natural alignment
2223: (e.g., an address aligned for a datum of size 8 is divisible by
2224: 8). Unaligned accesses usually result in a @code{-23 THROW}.
2225:
2226: @item @code{EMIT} and non-graphic characters:
2227: The character is output using the C library function (actually, macro)
2228: @code{putchar}.
2229:
2230: @item character editing of @code{ACCEPT} and @code{EXPECT}:
2231: This is modeled on the GNU readline library (@pxref{Readline
2232: Interaction, , Command Line Editing, readline, The GNU Readline
2233: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
2234: producing a full word completion every time you type it (instead of
2235: producing the common prefix of all completions).
2236:
2237: @item character set:
2238: The character set of your computer and display device. Gforth is
2239: 8-bit-clean (but some other component in your system may make trouble).
2240:
2241: @item Character-aligned address requirements:
2242: installation-dependent. Currently a character is represented by a C
2243: @code{unsigned char}; in the future we might switch to @code{wchar_t}
2244: (Comments on that requested).
2245:
2246: @item character-set extensions and matching of names:
2247: Any character except the ASCII NUL charcter can be used in a
2248: name. Matching is case-insensitive. The matching is performed using the
2249: C function @code{strncasecmp}, whose function is probably influenced by
2250: the locale. E.g., the @code{C} locale does not know about accents and
2251: umlauts, so they are matched case-sensitively in that locale. For
2252: portability reasons it is best to write programs such that they work in
2253: the @code{C} locale. Then one can use libraries written by a Polish
2254: programmer (who might use words containing ISO Latin-2 encoded
2255: characters) and by a French programmer (ISO Latin-1) in the same program
2256: (of course, @code{WORDS} will produce funny results for some of the
2257: words (which ones, depends on the font you are using)). Also, the locale
2258: you prefer may not be available in other operating systems. Hopefully,
2259: Unicode will solve these problems one day.
2260:
2261: @item conditions under which control characters match a space delimiter:
2262: If @code{WORD} is called with the space character as a delimiter, all
2263: white-space characters (as identified by the C macro @code{isspace()})
2264: are delimiters. @code{PARSE}, on the other hand, treats space like other
2265: delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
2266: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
2267: interpreter (aka text interpreter) by default, treats all white-space
2268: characters as delimiters.
2269:
2270: @item format of the control flow stack:
2271: The data stack is used as control flow stack. The size of a control flow
2272: stack item in cells is given by the constant @code{cs-item-size}. At the
2273: time of this writing, an item consists of a (pointer to a) locals list
2274: (third), an address in the code (second), and a tag for identifying the
2275: item (TOS). The following tags are used: @code{defstart},
2276: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
2277: @code{scopestart}.
2278:
2279: @item conversion of digits > 35
2280: The characters @code{[\]^_'} are the digits with the decimal value
2281: 36@minus{}41. There is no way to input many of the larger digits.
2282:
2283: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
2284: The cursor is moved to the end of the entered string. If the input is
2285: terminated using the @kbd{Return} key, a space is typed.
2286:
2287: @item exception abort sequence of @code{ABORT"}:
2288: The error string is stored into the variable @code{"error} and a
2289: @code{-2 throw} is performed.
2290:
2291: @item input line terminator:
2292: For interactive input, @kbd{C-m} and @kbd{C-j} terminate lines. One of
2293: these characters is typically produced when you type the @kbd{Enter} or
2294: @kbd{Return} key.
2295:
2296: @item maximum size of a counted string:
2297: @code{s" /counted-string" environment? drop .}. Currently 255 characters
2298: on all ports, but this may change.
2299:
2300: @item maximum size of a parsed string:
2301: Given by the constant @code{/line}. Currently 255 characters.
2302:
2303: @item maximum size of a definition name, in characters:
2304: 31
2305:
2306: @item maximum string length for @code{ENVIRONMENT?}, in characters:
2307: 31
2308:
2309: @item method of selecting the user input device:
2310: The user input device is the standard input. There is currently no way to
2311: change it from within Gforth. However, the input can typically be
2312: redirected in the command line that starts Gforth.
2313:
2314: @item method of selecting the user output device:
2315: The user output device is the standard output. It cannot be redirected
2316: from within Gforth, but typically from the command line that starts
2317: Gforth. Gforth uses buffered output, so output on a terminal does not
2318: become visible before the next newline or buffer overflow. Output on
2319: non-terminals is invisible until the buffer overflows.
2320:
2321: @item methods of dictionary compilation:
2322: What are we expected to document here?
2323:
2324: @item number of bits in one address unit:
2325: @code{s" address-units-bits" environment? drop .}. 8 in all current
2326: ports.
2327:
2328: @item number representation and arithmetic:
2329: Processor-dependent. Binary two's complement on all current ports.
2330:
2331: @item ranges for integer types:
2332: Installation-dependent. Make environmental queries for @code{MAX-N},
2333: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
2334: unsigned (and positive) types is 0. The lower bound for signed types on
2335: two's complement and one's complement machines machines can be computed
2336: by adding 1 to the upper bound.
2337:
2338: @item read-only data space regions:
2339: The whole Forth data space is writable.
2340:
2341: @item size of buffer at @code{WORD}:
2342: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
2343: shared with the pictured numeric output string. If overwriting
2344: @code{PAD} is acceptable, it is as large as the remaining dictionary
2345: space, although only as much can be sensibly used as fits in a counted
2346: string.
2347:
2348: @item size of one cell in address units:
2349: @code{1 cells .}.
2350:
2351: @item size of one character in address units:
2352: @code{1 chars .}. 1 on all current ports.
2353:
2354: @item size of the keyboard terminal buffer:
2355: Varies. You can determine the size at a specific time using @code{lp@
2356: tib - .}. It is shared with the locals stack and TIBs of files that
2357: include the current file. You can change the amount of space for TIBs
2358: and locals stack at Gforth startup with the command line option
2359: @code{-l}.
2360:
2361: @item size of the pictured numeric output buffer:
2362: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
2363: shared with @code{WORD}.
2364:
2365: @item size of the scratch area returned by @code{PAD}:
2366: The remainder of dictionary space. You can even use the unused part of
2367: the data stack space. The current size can be computed with @code{sp@
2368: pad - .}.
2369:
2370: @item system case-sensitivity characteristics:
2371: Dictionary searches are case insensitive. However, as explained above
2372: under @i{character-set extensions}, the matching for non-ASCII
2373: characters is determined by the locale you are using. In the default
2374: @code{C} locale all non-ASCII characters are matched case-sensitively.
2375:
2376: @item system prompt:
2377: @code{ ok} in interpret state, @code{ compiled} in compile state.
2378:
2379: @item division rounding:
2380: installation dependent. @code{s" floored" environment? drop .}. We leave
2381: the choice to gcc (what to use for @code{/}) and to you (whether to use
2382: @code{fm/mod}, @code{sm/rem} or simply @code{/}).
2383:
2384: @item values of @code{STATE} when true:
2385: -1.
2386:
2387: @item values returned after arithmetic overflow:
2388: On two's complement machines, arithmetic is performed modulo
2389: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
2390: arithmetic (with appropriate mapping for signed types). Division by zero
2391: typically results in a @code{-55 throw} (floatingpoint unidentified
2392: fault), although a @code{-10 throw} (divide by zero) would be more
2393: appropriate.
2394:
2395: @item whether the current definition can be found after @t{DOES>}:
2396: No.
2397:
2398: @end table
2399:
2400: @c ---------------------------------------------------------------------
2401: @node core-ambcond, core-other, core-idef, The Core Words
2402: @subsection Ambiguous conditions
2403: @c ---------------------------------------------------------------------
2404:
2405: @table @i
2406:
2407: @item a name is neither a word nor a number:
2408: @code{-13 throw} (Undefined word)
2409:
2410: @item a definition name exceeds the maximum length allowed:
2411: @code{-19 throw} (Word name too long)
2412:
2413: @item addressing a region not inside the various data spaces of the forth system:
2414: The stacks, code space and name space are accessible. Machine code space is
2415: typically readable. Accessing other addresses gives results dependent on
2416: the operating system. On decent systems: @code{-9 throw} (Invalid memory
2417: address).
2418:
2419: @item argument type incompatible with parameter:
2420: This is usually not caught. Some words perform checks, e.g., the control
2421: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
2422: mismatch).
2423:
2424: @item attempting to obtain the execution token of a word with undefined execution semantics:
2425: You get an execution token representing the compilation semantics
2426: instead.
2427:
2428: @item dividing by zero:
2429: typically results in a @code{-55 throw} (floating point unidentified
2430: fault), although a @code{-10 throw} (divide by zero) would be more
2431: appropriate.
2432:
2433: @item insufficient data stack or return stack space:
2434: Not checked. This typically results in mysterious illegal memory
2435: accesses, producing @code{-9 throw} (Invalid memory address) or
2436: @code{-23 throw} (Address alignment exception).
2437:
2438: @item insufficient space for loop control parameters:
2439: like other return stack overflows.
2440:
2441: @item insufficient space in the dictionary:
2442: Not checked. Similar results as stack overflows. However, typically the
2443: error appears at a different place when one inserts or removes code.
2444:
2445: @item interpreting a word with undefined interpretation semantics:
2446: For some words, we defined interpretation semantics. For the others:
2447: @code{-14 throw} (Interpreting a compile-only word). Note that this is
2448: checked only by the outer (aka text) interpreter; if the word is
2449: @code{execute}d in some other way, it will typically perform it's
2450: compilation semantics even in interpret state. (We could change @code{'}
2451: and relatives not to give the xt of such words, but we think that would
2452: be too restrictive).
2453:
2454: @item modifying the contents of the input buffer or a string literal:
2455: These are located in writable memory and can be modified.
2456:
2457: @item overflow of the pictured numeric output string:
2458: Not checked.
2459:
2460: @item parsed string overflow:
2461: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
2462:
2463: @item producing a result out of range:
2464: On two's complement machines, arithmetic is performed modulo
2465: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
2466: arithmetic (with appropriate mapping for signed types). Division by zero
2467: typically results in a @code{-55 throw} (floatingpoint unidentified
2468: fault), although a @code{-10 throw} (divide by zero) would be more
2469: appropriate. @code{convert} and @code{>number} currently overflow
2470: silently.
2471:
2472: @item reading from an empty data or return stack:
2473: The data stack is checked by the outer (aka text) interpreter after
2474: every word executed. If it has underflowed, a @code{-4 throw} (Stack
2475: underflow) is performed. Apart from that, the stacks are not checked and
2476: underflows can result in similar behaviour as overflows (of adjacent
2477: stacks).
2478:
2479: @item unexepected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
2480: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
2481: use zero-length string as a name). Words like @code{'} probably will not
2482: find what they search. Note that it is possible to create zero-length
2483: names with @code{nextname} (should it not?).
2484:
2485: @item @code{>IN} greater than input buffer:
2486: The next invocation of a parsing word returns a string wih length 0.
2487:
2488: @item @code{RECURSE} appears after @code{DOES>}:
2489: Compiles a recursive call to the defining word not to the defined word.
2490:
2491: @item argument input source different than current input source for @code{RESTORE-INPUT}:
2492: !!???If the argument input source is a valid input source then it gets
2493: restored. Otherwise causes @code{-12 THROW}, which, unless caught, issues
2494: the message "argument type mismatch" and aborts.
2495:
2496: @item data space containing definitions gets de-allocated:
2497: Deallocation with @code{allot} is not checked. This typically resuls in
2498: memory access faults or execution of illegal instructions.
2499:
2500: @item data space read/write with incorrect alignment:
2501: Processor-dependent. Typically results in a @code{-23 throw} (Address
2502: alignment exception). Under Linux on a 486 or later processor with
2503: alignment turned on, incorrect alignment results in a @code{-9 throw}
2504: (Invalid memory address). There are reportedly some processors with
2505: alignment restrictions that do not report them.
2506:
2507: @item data space pointer not properly aligned, @code{,}, @code{C,}:
2508: Like other alignment errors.
2509:
2510: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
2511: Not checked. May cause an illegal memory access.
2512:
2513: @item loop control parameters not available:
2514: Not checked. The counted loop words simply assume that the top of return
2515: stack items are loop control parameters and behave accordingly.
2516:
2517: @item most recent definition does not have a name (@code{IMMEDIATE}):
2518: @code{abort" last word was headerless"}.
2519:
2520: @item name not defined by @code{VALUE} used by @code{TO}:
2521: @code{-32 throw} (Invalid name argument)
2522:
2523: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
2524: @code{-13 throw} (Undefined word)
2525:
2526: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
2527: Gforth behaves as if they were of the same type. I.e., you can predict
2528: the behaviour by interpreting all parameters as, e.g., signed.
2529:
2530: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
2531: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} is equivalent to
2532: @code{TO}.
2533:
2534: @item String longer than a counted string returned by @code{WORD}:
2535: Not checked. The string will be ok, but the count will, of course,
2536: contain only the least significant bits of the length.
2537:
2538: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
2539: Processor-dependent. Typical behaviours are returning 0 and using only
2540: the low bits of the shift count.
2541:
2542: @item word not defined via @code{CREATE}:
2543: @code{>BODY} produces the PFA of the word no matter how it was defined.
2544:
2545: @code{DOES>} changes the execution semantics of the last defined word no
2546: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
2547: @code{CREATE , DOES>}.
2548:
2549: @item words improperly used outside @code{<#} and @code{#>}:
2550: Not checked. As usual, you can expect memory faults.
2551:
2552: @end table
2553:
2554:
2555: @c ---------------------------------------------------------------------
2556: @node core-other, , core-ambcond, The Core Words
2557: @subsection Other system documentation
2558: @c ---------------------------------------------------------------------
2559:
2560: @table @i
2561:
2562: @item nonstandard words using @code{PAD}:
2563: None.
2564:
2565: @item operator's terminal facilities available:
2566: !!??
2567:
2568: @item program data space available:
2569: @code{sp@ here - .} gives the space remaining for dictionary and data
2570: stack together.
2571:
2572: @item return stack space available:
2573: !!??
2574:
2575: @item stack space available:
2576: @code{sp@ here - .} gives the space remaining for dictionary and data
2577: stack together.
2578:
2579: @item system dictionary space required, in address units:
2580: Type @code{here forthstart - .} after startup. At the time of this
2581: writing, this gives 70108 (bytes) on a 32-bit system.
2582: @end table
2583:
2584:
2585: @c =====================================================================
2586: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
2587: @section The optional Block word set
2588: @c =====================================================================
2589:
2590: @menu
2591: * block-idef:: Implementation Defined Options
2592: * block-ambcond:: Ambiguous Conditions
2593: * block-other:: Other System Documentation
2594: @end menu
2595:
2596:
2597: @c ---------------------------------------------------------------------
2598: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
2599: @subsection Implementation Defined Options
2600: @c ---------------------------------------------------------------------
2601:
2602: @table @i
2603:
2604: @item the format for display by @code{LIST}:
2605: First the screen number is displayed, then 16 lines of 64 characters,
2606: each line preceded by the line number.
2607:
2608: @item the length of a line affected by @code{\}:
2609: 64 characters.
2610: @end table
2611:
2612:
2613: @c ---------------------------------------------------------------------
2614: @node block-ambcond, block-other, block-idef, The optional Block word set
2615: @subsection Ambiguous conditions
2616: @c ---------------------------------------------------------------------
2617:
2618: @table @i
2619:
2620: @item correct block read was not possible:
2621: Typically results in a @code{throw} of some OS-derived value (between
2622: -512 and -2048). If the blocks file was just not long enough, blanks are
2623: supplied for the missing portion.
2624:
2625: @item I/O exception in block transfer:
2626: Typically results in a @code{throw} of some OS-derived value (between
2627: -512 and -2048).
2628:
2629: @item invalid block number:
2630: @code{-35 throw} (Invalid block number)
2631:
2632: @item a program directly alters the contents of @code{BLK}:
2633: The input stream is switched to that other block, at the same
2634: position. If the storing to @code{BLK} happens when interpreting
2635: non-block input, the system will get quite confused when the block ends.
2636:
2637: @item no current block buffer for @code{UPDATE}:
2638: @code{UPDATE} has no effect.
2639:
2640: @end table
2641:
2642:
2643: @c ---------------------------------------------------------------------
2644: @node block-other, , block-ambcond, The optional Block word set
2645: @subsection Other system documentation
2646: @c ---------------------------------------------------------------------
2647:
2648: @table @i
2649:
2650: @item any restrictions a multiprogramming system places on the use of buffer addresses:
2651: No restrictions (yet).
2652:
2653: @item the number of blocks available for source and data:
2654: depends on your disk space.
2655:
2656: @end table
2657:
2658:
2659: @c =====================================================================
2660: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
2661: @section The optional Double Number word set
2662: @c =====================================================================
2663:
2664: @menu
2665: * double-ambcond:: Ambiguous Conditions
2666: @end menu
2667:
2668:
2669: @c ---------------------------------------------------------------------
2670: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
2671: @subsection Ambiguous conditions
2672: @c ---------------------------------------------------------------------
2673:
2674: @table @i
2675:
2676: @item @var{d} outside of range of @var{n} in @code{D>S}:
2677: The least significant cell of @var{d} is produced.
2678:
2679: @end table
2680:
2681:
2682: @c =====================================================================
2683: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
2684: @section The optional Exception word set
2685: @c =====================================================================
2686:
2687: @menu
2688: * exception-idef:: Implementation Defined Options
2689: @end menu
2690:
2691:
2692: @c ---------------------------------------------------------------------
2693: @node exception-idef, , The optional Exception word set, The optional Exception word set
2694: @subsection Implementation Defined Options
2695: @c ---------------------------------------------------------------------
2696:
2697: @table @i
2698: @item @code{THROW}-codes used in the system:
2699: The codes -256@minus{}-511 are used for reporting signals (see
2700: @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
2701: (for file and memory allocation operations). The mapping from OS error
2702: numbers to throw code is -512@minus{}@var{errno}. One side effect of
2703: this mapping is that undefined OS errors produce a message with a
2704: strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
2705: 488} on my system.
2706: @end table
2707:
2708: @c =====================================================================
2709: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
2710: @section The optional Facility word set
2711: @c =====================================================================
2712:
2713: @menu
2714: * facility-idef:: Implementation Defined Options
2715: * facility-ambcond:: Ambiguous Conditions
2716: @end menu
2717:
2718:
2719: @c ---------------------------------------------------------------------
2720: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
2721: @subsection Implementation Defined Options
2722: @c ---------------------------------------------------------------------
2723:
2724: @table @i
2725:
2726: @item encoding of keyboard events (@code{EKEY}):
2727: Not yet implemeted.
2728:
2729: @item duration of a system clock tick
2730: System dependent. With respect to @code{MS}, the time is specified in
2731: microseconds. How well the OS and the hardware implement this, is
2732: another question.
2733:
2734: @item repeatability to be expected from the execution of @code{MS}:
2735: System dependent. On Unix, a lot depends on load. If the system is
2736: lightly loaded, and the delay is short enough that Gforth does not get
2737: swapped out, the performance should be acceptable. Under MS-DOS and
2738: other single-tasking systems, it should be good.
2739:
2740: @end table
2741:
2742:
2743: @c ---------------------------------------------------------------------
2744: @node facility-ambcond, , facility-idef, The optional Facility word set
2745: @subsection Ambiguous conditions
2746: @c ---------------------------------------------------------------------
2747:
2748: @table @i
2749:
2750: @item @code{AT-XY} can't be performed on user output device:
2751: Largely terminal dependant. No range checks are done on the arguments.
2752: No errors are reported. You may see some garbage appearing, you may see
2753: simply nothing happen.
2754:
2755: @end table
2756:
2757:
2758: @c =====================================================================
2759: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
2760: @section The optional File-Access word set
2761: @c =====================================================================
2762:
2763: @menu
2764: * file-idef:: Implementation Defined Options
2765: * file-ambcond:: Ambiguous Conditions
2766: @end menu
2767:
2768:
2769: @c ---------------------------------------------------------------------
2770: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
2771: @subsection Implementation Defined Options
2772: @c ---------------------------------------------------------------------
2773:
2774: @table @i
2775:
2776: @item File access methods used:
2777: @code{R/O}, @code{R/W} and @code{BIN} work as you would
2778: expect. @code{W/O} translates into the C file opening mode @code{w} (or
2779: @code{wb}): The file is cleared, if it exists, and created, if it does
2780: not (both with @code{open-file} and @code{create-file}). Under Unix
2781: @code{create-file} creates a file with 666 permissions modified by your
2782: umask.
2783:
2784: @item file exceptions:
2785: The file words do not raise exceptions (except, perhaps, memory access
2786: faults when you pass illegal addresses or file-ids).
2787:
2788: @item file line terminator:
2789: System-dependent. Gforth uses C's newline character as line
2790: terminator. What the actual character code(s) of this are is
2791: system-dependent.
2792:
2793: @item file name format
2794: System dependent. Gforth just uses the file name format of your OS.
2795:
2796: @item information returned by @code{FILE-STATUS}:
2797: @code{FILE-STATUS} returns the most powerful file access mode allowed
2798: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
2799: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
2800: along with the retured mode.
2801:
2802: @item input file state after an exception when including source:
2803: All files that are left via the exception are closed.
2804:
2805: @item @var{ior} values and meaning:
2806: The @var{ior}s returned by the file and memory allocation words are
2807: intended as throw codes. They typically are in the range
2808: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
2809: @var{ior}s is -512@minus{}@var{errno}.
2810:
2811: @item maximum depth of file input nesting:
2812: limited by the amount of return stack, locals/TIB stack, and the number
2813: of open files available. This should not give you troubles.
2814:
2815: @item maximum size of input line:
2816: @code{/line}. Currently 255.
2817:
2818: @item methods of mapping block ranges to files:
2819: Currently, the block words automatically access the file
2820: @file{blocks.fb} in the currend working directory. More sophisticated
2821: methods could be implemented if there is demand (and a volunteer).
2822:
2823: @item number of string buffers provided by @code{S"}:
2824: 1
2825:
2826: @item size of string buffer used by @code{S"}:
2827: @code{/line}. currently 255.
2828:
2829: @end table
2830:
2831: @c ---------------------------------------------------------------------
2832: @node file-ambcond, , file-idef, The optional File-Access word set
2833: @subsection Ambiguous conditions
2834: @c ---------------------------------------------------------------------
2835:
2836: @table @i
2837:
2838: @item attempting to position a file outside it's boundaries:
2839: @code{REPOSITION-FILE} is performed as usual: Afterwards,
2840: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
2841:
2842: @item attempting to read from file positions not yet written:
2843: End-of-file, i.e., zero characters are read and no error is reported.
2844:
2845: @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
2846: An appropriate exception may be thrown, but a memory fault or other
2847: problem is more probable.
2848:
2849: @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
2850: The @var{ior} produced by the operation, that discovered the problem, is
2851: thrown.
2852:
2853: @item named file cannot be opened (@code{included}):
2854: The @var{ior} produced by @code{open-file} is thrown.
2855:
2856: @item requesting an unmapped block number:
2857: There are no unmapped legal block numbers. On some operating systems,
2858: writing a block with a large number may overflow the file system and
2859: have an error message as consequence.
2860:
2861: @item using @code{source-id} when @code{blk} is non-zero:
2862: @code{source-id} performs its function. Typically it will give the id of
2863: the source which loaded the block. (Better ideas?)
2864:
2865: @end table
2866:
2867:
2868: @c =====================================================================
2869: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
2870: @section The optional Floating-Point word set
2871: @c =====================================================================
2872:
2873: @menu
2874: * floating-idef:: Implementation Defined Options
2875: * floating-ambcond:: Ambiguous Conditions
2876: @end menu
2877:
2878:
2879: @c ---------------------------------------------------------------------
2880: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
2881: @subsection Implementation Defined Options
2882: @c ---------------------------------------------------------------------
2883:
2884: @table @i
2885:
2886: @item format and range of floating point numbers:
2887: System-dependent; the @code{double} type of C.
2888:
2889: @item results of @code{REPRESENT} when @var{float} is out of range:
2890: System dependent; @code{REPRESENT} is implemented using the C library
2891: function @code{ecvt()} and inherits its behaviour in this respect.
2892:
2893: @item rounding or truncation of floating-point numbers:
2894: What's the question?!!
2895:
2896: @item size of floating-point stack:
2897: @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
2898: with the command-line option @code{-f}.
2899:
2900: @item width of floating-point stack:
2901: @code{1 floats}.
2902:
2903: @end table
2904:
2905:
2906: @c ---------------------------------------------------------------------
2907: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
2908: @subsection Ambiguous conditions
2909: @c ---------------------------------------------------------------------
2910:
2911: @table @i
2912:
2913: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
2914: System-dependent. Typically results in an alignment fault like other
2915: alignment violations.
2916:
2917: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
2918: System-dependent. Typically results in an alignment fault like other
2919: alignment violations.
2920:
2921: @item Floating-point result out of range:
2922: System-dependent. Can result in a @code{-55 THROW} (Floating-point
2923: unidentified fault), or can produce a special value representing, e.g.,
2924: Infinity.
2925:
2926: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
2927: System-dependent. Typically results in an alignment fault like other
2928: alignment violations.
2929:
2930: @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
2931: The floating-point number is converted into decimal nonetheless.
2932:
2933: @item Both arguments are equal to zero (@code{FATAN2}):
2934: System-dependent. @code{FATAN2} is implemented using the C library
2935: function @code{atan2()}.
2936:
2937: @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
2938: System-dependent. Anyway, typically the cos of @var{r1} will not be zero
2939: because of small errors and the tan will be a very large (or very small)
2940: but finite number.
2941:
2942: @item @var{d} cannot be presented precisely as a float in @code{D>F}:
2943: The result is rounded to the nearest float.
2944:
2945: @item dividing by zero:
2946: @code{-55 throw} (Floating-point unidentified fault)
2947:
2948: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
2949: System dependent. On IEEE-FP based systems the number is converted into
2950: an infinity.
2951:
2952: @item @var{float}<1 (@code{facosh}):
2953: @code{-55 throw} (Floating-point unidentified fault)
2954:
2955: @item @var{float}=<-1 (@code{flnp1}):
2956: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
2957: negative infinity is typically produced for @var{float}=-1.
2958:
2959: @item @var{float}=<0 (@code{fln}, @code{flog}):
2960: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
2961: negative infinity is typically produced for @var{float}=0.
2962:
2963: @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
2964: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
2965: produces values for these inputs on my Linux box (Bug in the C library?)
2966:
2967: @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
2968: @code{-55 throw} (Floating-point unidentified fault).
2969:
2970: @item integer part of float cannot be represented by @var{d} in @code{f>d}:
2971: @code{-55 throw} (Floating-point unidentified fault).
2972:
2973: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
2974: This does not happen.
2975: @end table
2976:
2977:
2978:
2979: @c =====================================================================
2980: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
2981: @section The optional Locals word set
2982: @c =====================================================================
2983:
2984: @menu
2985: * locals-idef:: Implementation Defined Options
2986: * locals-ambcond:: Ambiguous Conditions
2987: @end menu
2988:
2989:
2990: @c ---------------------------------------------------------------------
2991: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
2992: @subsection Implementation Defined Options
2993: @c ---------------------------------------------------------------------
2994:
2995: @table @i
2996:
2997: @item maximum number of locals in a definition:
2998: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
2999: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
3000: characters. The number of locals in a definition is bounded by the size
3001: of locals-buffer, which contains the names of the locals.
3002:
3003: @end table
3004:
3005:
3006: @c ---------------------------------------------------------------------
3007: @node locals-ambcond, , locals-idef, The optional Locals word set
3008: @subsection Ambiguous conditions
3009: @c ---------------------------------------------------------------------
3010:
3011: @table @i
3012:
3013: @item executing a named local in interpretation state:
3014: @code{-14 throw} (Interpreting a compile-only word).
3015:
3016: @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
3017: @code{-32 throw} (Invalid name argument)
3018:
3019: @end table
3020:
3021:
3022: @c =====================================================================
3023: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
3024: @section The optional Memory-Allocation word set
3025: @c =====================================================================
3026:
3027: @menu
3028: * memory-idef:: Implementation Defined Options
3029: @end menu
3030:
3031:
3032: @c ---------------------------------------------------------------------
3033: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
3034: @subsection Implementation Defined Options
3035: @c ---------------------------------------------------------------------
3036:
3037: @table @i
3038:
3039: @item values and meaning of @var{ior}:
3040: The @var{ior}s returned by the file and memory allocation words are
3041: intended as throw codes. They typically are in the range
3042: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
3043: @var{ior}s is -512@minus{}@var{errno}.
3044:
3045: @end table
3046:
3047: @c =====================================================================
3048: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
3049: @section The optional Programming-Tools word set
3050: @c =====================================================================
3051:
3052: @menu
3053: * programming-idef:: Implementation Defined Options
3054: * programming-ambcond:: Ambiguous Conditions
3055: @end menu
3056:
3057:
3058: @c ---------------------------------------------------------------------
3059: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
3060: @subsection Implementation Defined Options
3061: @c ---------------------------------------------------------------------
3062:
3063: @table @i
3064:
3065: @item ending sequence for input following @code{;code} and @code{code}:
3066: Not implemented (yet).
3067:
3068: @item manner of processing input following @code{;code} and @code{code}:
3069: Not implemented (yet).
3070:
3071: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
3072: Not implemented (yet). If they were implemented, they would use the
3073: search order wordset.
3074:
3075: @item source and format of display by @code{SEE}:
3076: The source for @code{see} is the intermediate code used by the inner
3077: interpreter. The current @code{see} tries to output Forth source code
3078: as well as possible.
3079:
3080: @end table
3081:
3082: @c ---------------------------------------------------------------------
3083: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
3084: @subsection Ambiguous conditions
3085: @c ---------------------------------------------------------------------
3086:
3087: @table @i
3088:
3089: @item deleting the compilation wordlist (@code{FORGET}):
3090: Not implemented (yet).
3091:
3092: @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
3093: This typically results in an @code{abort"} with a descriptive error
3094: message (may change into a @code{-22 throw} (Control structure mismatch)
3095: in the future). You may also get a memory access error. If you are
3096: unlucky, this ambiguous condition is not caught.
3097:
3098: @item @var{name} can't be found (@code{forget}):
3099: Not implemented (yet).
3100:
3101: @item @var{name} not defined via @code{CREATE}:
3102: @code{;code} is not implemented (yet). If it were, it would behave like
3103: @code{DOES>} in this respect, i.e., change the execution semantics of
3104: the last defined word no matter how it was defined.
3105:
3106: @item @code{POSTPONE} applied to @code{[IF]}:
3107: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
3108: equivalent to @code{[IF]}.
3109:
3110: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
3111: Continue in the same state of conditional compilation in the next outer
3112: input source. Currently there is no warning to the user about this.
3113:
3114: @item removing a needed definition (@code{FORGET}):
3115: Not implemented (yet).
3116:
3117: @end table
3118:
3119:
3120: @c =====================================================================
3121: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
3122: @section The optional Search-Order word set
3123: @c =====================================================================
3124:
3125: @menu
3126: * search-idef:: Implementation Defined Options
3127: * search-ambcond:: Ambiguous Conditions
3128: @end menu
3129:
3130:
3131: @c ---------------------------------------------------------------------
3132: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
3133: @subsection Implementation Defined Options
3134: @c ---------------------------------------------------------------------
3135:
3136: @table @i
3137:
3138: @item maximum number of word lists in search order:
3139: @code{s" wordlists" environment? drop .}. Currently 16.
3140:
3141: @item minimum search order:
3142: @code{root root}.
3143:
3144: @end table
3145:
3146: @c ---------------------------------------------------------------------
3147: @node search-ambcond, , search-idef, The optional Search-Order word set
3148: @subsection Ambiguous conditions
3149: @c ---------------------------------------------------------------------
3150:
3151: @table @i
3152:
3153: @item changing the compilation wordlist (during compilation):
3154: The definition is put into the wordlist that is the compilation wordlist
3155: when @code{REVEAL} is executed (by @code{;}, @code{DOES>},
3156: @code{RECURSIVE}, etc.).
3157:
3158: @item search order empty (@code{previous}):
3159: @code{abort" Vocstack empty"}.
3160:
3161: @item too many word lists in search order (@code{also}):
3162: @code{abort" Vocstack full"}.
3163:
3164: @end table
3165:
3166:
3167: @node Model, Emacs and Gforth, ANS conformance, Top
3168: @chapter Model
3169:
3170: @node Emacs and Gforth, Internals, Model, Top
3171: @chapter Emacs and Gforth
3172:
3173: Gforth comes with @file{gforth.el}, an improved version of
3174: @file{forth.el} by Goran Rydqvist (icluded in the TILE package). The
3175: improvements are a better (but still not perfect) handling of
3176: indentation. I have also added comment paragraph filling (@kbd{M-q}),
3177: commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
3178: removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
3179: stuff I do not use alone, even though some of it only makes sense for
3180: TILE. To get a description of these features, enter Forth mode and type
3181: @kbd{C-h m}.
3182:
3183: In addition, Gforth supports Emacs quite well: The source code locations
3184: given in error messages, debugging output (from @code{~~}) and failed
3185: assertion messages are in the right format for Emacs' compilation mode
3186: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
3187: Manual}) so the source location corresponding to an error or other
3188: message is only a few keystrokes away (@kbd{C-x `} for the next error,
3189: @kbd{C-c C-c} for the error under the cursor).
3190:
3191: Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
3192: (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
3193: contains the definitions of all words defined afterwards. You can then
3194: find the source for a word using @kbd{M-.}. Note that emacs can use
3195: several tags files at the same time (e.g., one for the Gforth sources
3196: and one for your program).
3197:
3198: To get all these benefits, add the following lines to your @file{.emacs}
3199: file:
3200:
3201: @example
3202: (autoload 'forth-mode "gforth.el")
3203: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
3204: @end example
3205:
3206: @node Internals, Bugs, Emacs and Gforth, Top
3207: @chapter Internals
3208:
3209: Reading this section is not necessary for programming with Gforth. It
3210: should be helpful for finding your way in the Gforth sources.
3211:
3212: @menu
3213: * Portability::
3214: * Threading::
3215: * Primitives::
3216: * System Architecture::
3217: * Performance::
3218: @end menu
3219:
3220: @node Portability, Threading, Internals, Internals
3221: @section Portability
3222:
3223: One of the main goals of the effort is availability across a wide range
3224: of personal machines. fig-Forth, and, to a lesser extent, F83, achieved
3225: this goal by manually coding the engine in assembly language for several
3226: then-popular processors. This approach is very labor-intensive and the
3227: results are short-lived due to progress in computer architecture.
3228:
3229: Others have avoided this problem by coding in C, e.g., Mitch Bradley
3230: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
3231: particularly popular for UNIX-based Forths due to the large variety of
3232: architectures of UNIX machines. Unfortunately an implementation in C
3233: does not mix well with the goals of efficiency and with using
3234: traditional techniques: Indirect or direct threading cannot be expressed
3235: in C, and switch threading, the fastest technique available in C, is
3236: significantly slower. Another problem with C is that it's very
3237: cumbersome to express double integer arithmetic.
3238:
3239: Fortunately, there is a portable language that does not have these
3240: limitations: GNU C, the version of C processed by the GNU C compiler
3241: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
3242: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
3243: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
3244: threading possible, its @code{long long} type (@pxref{Long Long, ,
3245: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forths
3246: double numbers. GNU C is available for free on all important (and many
3247: unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and
3248: the Atari ST, so a Forth written in GNU C can run on all these
3249: machines.
3250:
3251: Writing in a portable language has the reputation of producing code that
3252: is slower than assembly. For our Forth engine we repeatedly looked at
3253: the code produced by the compiler and eliminated most compiler-induced
3254: inefficiencies by appropriate changes in the source-code.
3255:
3256: However, register allocation cannot be portably influenced by the
3257: programmer, leading to some inefficiencies on register-starved
3258: machines. We use explicit register declarations (@pxref{Explicit Reg
3259: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
3260: improve the speed on some machines. They are turned on by using the
3261: @code{gcc} switch @code{-DFORCE_REG}. Unfortunately, this feature not
3262: only depends on the machine, but also on the compiler version: On some
3263: machines some compiler versions produce incorrect code when certain
3264: explicit register declarations are used. So by default
3265: @code{-DFORCE_REG} is not used.
3266:
3267: @node Threading, Primitives, Portability, Internals
3268: @section Threading
3269:
3270: GNU C's labels as values extension (available since @code{gcc-2.0},
3271: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
3272: makes it possible to take the address of @var{label} by writing
3273: @code{&&@var{label}}. This address can then be used in a statement like
3274: @code{goto *@var{address}}. I.e., @code{goto *&&x} is the same as
3275: @code{goto x}.
3276:
3277: With this feature an indirect threaded NEXT looks like:
3278: @example
3279: cfa = *ip++;
3280: ca = *cfa;
3281: goto *ca;
3282: @end example
3283: For those unfamiliar with the names: @code{ip} is the Forth instruction
3284: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
3285: execution token and points to the code field of the next word to be
3286: executed; The @code{ca} (code address) fetched from there points to some
3287: executable code, e.g., a primitive or the colon definition handler
3288: @code{docol}.
3289:
3290: Direct threading is even simpler:
3291: @example
3292: ca = *ip++;
3293: goto *ca;
3294: @end example
3295:
3296: Of course we have packaged the whole thing neatly in macros called
3297: @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
3298:
3299: @menu
3300: * Scheduling::
3301: * Direct or Indirect Threaded?::
3302: * DOES>::
3303: @end menu
3304:
3305: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
3306: @subsection Scheduling
3307:
3308: There is a little complication: Pipelined and superscalar processors,
3309: i.e., RISC and some modern CISC machines can process independent
3310: instructions while waiting for the results of an instruction. The
3311: compiler usually reorders (schedules) the instructions in a way that
3312: achieves good usage of these delay slots. However, on our first tries
3313: the compiler did not do well on scheduling primitives. E.g., for
3314: @code{+} implemented as
3315: @example
3316: n=sp[0]+sp[1];
3317: sp++;
3318: sp[0]=n;
3319: NEXT;
3320: @end example
3321: the NEXT comes strictly after the other code, i.e., there is nearly no
3322: scheduling. After a little thought the problem becomes clear: The
3323: compiler cannot know that sp and ip point to different addresses (and
3324: the version of @code{gcc} we used would not know it even if it was
3325: possible), so it could not move the load of the cfa above the store to
3326: the TOS. Indeed the pointers could be the same, if code on or very near
3327: the top of stack were executed. In the interest of speed we chose to
3328: forbid this probably unused ``feature'' and helped the compiler in
3329: scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
3330: the goto part (@code{NEXT_P2}). @code{+} now looks like:
3331: @example
3332: n=sp[0]+sp[1];
3333: sp++;
3334: NEXT_P1;
3335: sp[0]=n;
3336: NEXT_P2;
3337: @end example
3338: This can be scheduled optimally by the compiler.
3339:
3340: This division can be turned off with the switch @code{-DCISC_NEXT}. This
3341: switch is on by default on machines that do not profit from scheduling
3342: (e.g., the 80386), in order to preserve registers.
3343:
3344: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
3345: @subsection Direct or Indirect Threaded?
3346:
3347: Both! After packaging the nasty details in macro definitions we
3348: realized that we could switch between direct and indirect threading by
3349: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
3350: defining a few machine-specific macros for the direct-threading case.
3351: On the Forth level we also offer access words that hide the
3352: differences between the threading methods (@pxref{Threading Words}).
3353:
3354: Indirect threading is implemented completely
3355: machine-independently. Direct threading needs routines for creating
3356: jumps to the executable code (e.g. to docol or dodoes). These routines
3357: are inherently machine-dependent, but they do not amount to many source
3358: lines. I.e., even porting direct threading to a new machine is a small
3359: effort.
3360:
3361: @node DOES>, , Direct or Indirect Threaded?, Threading
3362: @subsection DOES>
3363: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
3364: the chunk of code executed by every word defined by a
3365: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
3366: the Forth code to be executed, i.e. the code after the @code{DOES>} (the
3367: DOES-code)? There are two solutions:
3368:
3369: In fig-Forth the code field points directly to the dodoes and the
3370: DOES-code address is stored in the cell after the code address
3371: (i.e. at cfa cell+). It may seem that this solution is illegal in the
3372: Forth-79 and all later standards, because in fig-Forth this address
3373: lies in the body (which is illegal in these standards). However, by
3374: making the code field larger for all words this solution becomes legal
3375: again. We use this approach for the indirect threaded version. Leaving
3376: a cell unused in most words is a bit wasteful, but on the machines we
3377: are targetting this is hardly a problem. The other reason for having a
3378: code field size of two cells is to avoid having different image files
3379: for direct and indirect threaded systems (@pxref{System Architecture}).
3380:
3381: The other approach is that the code field points or jumps to the cell
3382: after @code{DOES}. In this variant there is a jump to @code{dodoes} at
3383: this address. @code{dodoes} can then get the DOES-code address by
3384: computing the code address, i.e., the address of the jump to dodoes,
3385: and add the length of that jump field. A variant of this is to have a
3386: call to @code{dodoes} after the @code{DOES>}; then the return address
3387: (which can be found in the return register on RISCs) is the DOES-code
3388: address. Since the two cells available in the code field are usually
3389: used up by the jump to the code address in direct threading, we use
3390: this approach for direct threading. We did not want to add another
3391: cell to the code field.
3392:
3393: @node Primitives, System Architecture, Threading, Internals
3394: @section Primitives
3395:
3396: @menu
3397: * Automatic Generation::
3398: * TOS Optimization::
3399: * Produced code::
3400: @end menu
3401:
3402: @node Automatic Generation, TOS Optimization, Primitives, Primitives
3403: @subsection Automatic Generation
3404:
3405: Since the primitives are implemented in a portable language, there is no
3406: longer any need to minimize the number of primitives. On the contrary,
3407: having many primitives is an advantage: speed. In order to reduce the
3408: number of errors in primitives and to make programming them easier, we
3409: provide a tool, the primitive generator (@file{prims2x.fs}), that
3410: automatically generates most (and sometimes all) of the C code for a
3411: primitive from the stack effect notation. The source for a primitive
3412: has the following form:
3413:
3414: @format
3415: @var{Forth-name} @var{stack-effect} @var{category} [@var{pronounc.}]
3416: [@code{""}@var{glossary entry}@code{""}]
3417: @var{C code}
3418: [@code{:}
3419: @var{Forth code}]
3420: @end format
3421:
3422: The items in brackets are optional. The category and glossary fields
3423: are there for generating the documentation, the Forth code is there
3424: for manual implementations on machines without GNU C. E.g., the source
3425: for the primitive @code{+} is:
3426: @example
3427: + n1 n2 -- n core plus
3428: n = n1+n2;
3429: @end example
3430:
3431: This looks like a specification, but in fact @code{n = n1+n2} is C
3432: code. Our primitive generation tool extracts a lot of information from
3433: the stack effect notations@footnote{We use a one-stack notation, even
3434: though we have separate data and floating-point stacks; The separate
3435: notation can be generated easily from the unified notation.}: The number
3436: of items popped from and pushed on the stack, their type, and by what
3437: name they are referred to in the C code. It then generates a C code
3438: prelude and postlude for each primitive. The final C code for @code{+}
3439: looks like this:
3440:
3441: @example
3442: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
3443: /* */ /* documentation */
3444: @{
3445: DEF_CA /* definition of variable ca (indirect threading) */
3446: Cell n1; /* definitions of variables */
3447: Cell n2;
3448: Cell n;
3449: n1 = (Cell) sp[1]; /* input */
3450: n2 = (Cell) TOS;
3451: sp += 1; /* stack adjustment */
3452: NAME("+") /* debugging output (with -DDEBUG) */
3453: @{
3454: n = n1+n2; /* C code taken from the source */
3455: @}
3456: NEXT_P1; /* NEXT part 1 */
3457: TOS = (Cell)n; /* output */
3458: NEXT_P2; /* NEXT part 2 */
3459: @}
3460: @end example
3461:
3462: This looks long and inefficient, but the GNU C compiler optimizes quite
3463: well and produces optimal code for @code{+} on, e.g., the R3000 and the
3464: HP RISC machines: Defining the @code{n}s does not produce any code, and
3465: using them as intermediate storage also adds no cost.
3466:
3467: There are also other optimizations, that are not illustrated by this
3468: example: Assignments between simple variables are usually for free (copy
3469: propagation). If one of the stack items is not used by the primitive
3470: (e.g. in @code{drop}), the compiler eliminates the load from the stack
3471: (dead code elimination). On the other hand, there are some things that
3472: the compiler does not do, therefore they are performed by
3473: @file{prims2x.fs}: The compiler does not optimize code away that stores
3474: a stack item to the place where it just came from (e.g., @code{over}).
3475:
3476: While programming a primitive is usually easy, there are a few cases
3477: where the programmer has to take the actions of the generator into
3478: account, most notably @code{?dup}, but also words that do not (always)
3479: fall through to NEXT.
3480:
3481: @node TOS Optimization, Produced code, Automatic Generation, Primitives
3482: @subsection TOS Optimization
3483:
3484: An important optimization for stack machine emulators, e.g., Forth
3485: engines, is keeping one or more of the top stack items in
3486: registers. If a word has the stack effect @var{in1}...@var{inx} @code{--}
3487: @var{out1}...@var{outy}, keeping the top @var{n} items in registers
3488: @itemize
3489: @item
3490: is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
3491: due to fewer loads from and stores to the stack.
3492: @item is slower than keeping @var{n-1} items, if @var{x<>y} and @var{x<n} and
3493: @var{y<n}, due to additional moves between registers.
3494: @end itemize
3495:
3496: In particular, keeping one item in a register is never a disadvantage,
3497: if there are enough registers. Keeping two items in registers is a
3498: disadvantage for frequent words like @code{?branch}, constants,
3499: variables, literals and @code{i}. Therefore our generator only produces
3500: code that keeps zero or one items in registers. The generated C code
3501: covers both cases; the selection between these alternatives is made at
3502: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
3503: code for @code{+} is just a simple variable name in the one-item case,
3504: otherwise it is a macro that expands into @code{sp[0]}. Note that the
3505: GNU C compiler tries to keep simple variables like @code{TOS} in
3506: registers, and it usually succeeds, if there are enough registers.
3507:
3508: The primitive generator performs the TOS optimization for the
3509: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
3510: operations the benefit of this optimization is even larger:
3511: floating-point operations take quite long on most processors, but can be
3512: performed in parallel with other operations as long as their results are
3513: not used. If the FP-TOS is kept in a register, this works. If
3514: it is kept on the stack, i.e., in memory, the store into memory has to
3515: wait for the result of the floating-point operation, lengthening the
3516: execution time of the primitive considerably.
3517:
3518: The TOS optimization makes the automatic generation of primitives a
3519: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
3520: @code{TOS} is not sufficient. There are some special cases to
3521: consider:
3522: @itemize
3523: @item In the case of @code{dup ( w -- w w )} the generator must not
3524: eliminate the store to the original location of the item on the stack,
3525: if the TOS optimization is turned on.
3526: @item Primitives with stack effects of the form @code{--}
3527: @var{out1}...@var{outy} must store the TOS to the stack at the start.
3528: Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
3529: must load the TOS from the stack at the end. But for the null stack
3530: effect @code{--} no stores or loads should be generated.
3531: @end itemize
3532:
3533: @node Produced code, , TOS Optimization, Primitives
3534: @subsection Produced code
3535:
3536: To see what assembly code is produced for the primitives on your machine
3537: with your compiler and your flag settings, type @code{make engine.s} and
3538: look at the resulting file @file{engine.s}.
3539:
3540: @node System Architecture, Performance, Primitives, Internals
3541: @section System Architecture
3542:
3543: Our Forth system consists not only of primitives, but also of
3544: definitions written in Forth. Since the Forth compiler itself belongs
3545: to those definitions, it is not possible to start the system with the
3546: primitives and the Forth source alone. Therefore we provide the Forth
3547: code as an image file in nearly executable form. At the start of the
3548: system a C routine loads the image file into memory, sets up the
3549: memory (stacks etc.) according to information in the image file, and
3550: starts executing Forth code.
3551:
3552: The image file format is a compromise between the goals of making it
3553: easy to generate image files and making them portable. The easiest way
3554: to generate an image file is to just generate a memory dump. However,
3555: this kind of image file cannot be used on a different machine, or on
3556: the next version of the engine on the same machine, it even might not
3557: work with the same engine compiled by a different version of the C
3558: compiler. We would like to have as few versions of the image file as
3559: possible, because we do not want to distribute many versions of the
3560: same image file, and to make it easy for the users to use their image
3561: files on many machines. We currently need to create a different image
3562: file for machines with different cell sizes and different byte order
3563: (little- or big-endian)@footnote{We are considering adding information to the
3564: image file that enables the loader to change the byte order.}.
3565:
3566: Forth code that is going to end up in a portable image file has to
3567: comply to some restrictions: addresses have to be stored in memory with
3568: special words (@code{A!}, @code{A,}, etc.) in order to make the code
3569: relocatable. Cells, floats, etc., have to be stored at the natural
3570: alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
3571: dividable by~8. This happens automatically in our system when you use
3572: the ANS Forth alignment words.}, in order to avoid alignment faults on
3573: machines with stricter alignment. The image file is produced by a
3574: metacompiler (@file{cross.fs}).
3575:
3576: So, unlike the image file of Mitch Bradleys @code{cforth}, our image
3577: file is not directly executable, but has to undergo some manipulations
3578: during loading. Address relocation is performed at image load-time, not
3579: at run-time. The loader also has to replace tokens standing for
3580: primitive calls with the appropriate code-field addresses (or code
3581: addresses in the case of direct threading).
3582:
3583: @node Performance, , System Architecture, Internals
3584: @section Performance
3585:
3586: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
3587: impossible to write a significantly faster engine.
3588:
3589: On register-starved machines like the 386 architecture processors
3590: improvements are possible, because @code{gcc} does not utilize the
3591: registers as well as a human, even with explicit register declarations;
3592: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
3593: and hand-tuned it for the 486; this system is 1.19 times faster on the
3594: Sieve benchmark on a 486DX2/66 than Gforth compiled with
3595: @code{gcc-2.6.3} with @code{-DFORCE_REG}.
3596:
3597: However, this potential advantage of assembly language implementations
3598: is not necessarily realized in complete Forth systems: We compared
3599: Gforth (compiled with @code{gcc-2.6.3} and @code{-DFORCE_REG}) with
3600: Win32Forth 1.2093 and LMI's NT Forth (Beta, May 1994), two systems
3601: written in assembly, and with two systems written in C: PFE-0.9.11
3602: (compiled with @code{gcc-2.6.3} with the default configuration for
3603: Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS}) and ThisForth Beta
3604: (compiled with gcc-2.6.3 -O3 -fomit-frame-pointer). We benchmarked
3605: Gforth, PFE and ThisForth on a 486DX2/66 under Linux. Kenneth O'Heskin
3606: kindly provided the results for Win32Forth and NT Forth on a 486DX2/66
3607: with similar memory performance under Windows NT.
3608:
3609: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
3610: matrix multiplication come from the Stanford integer benchmarks and have
3611: been translated into Forth by Martin Fraeman; we used the versions
3612: included in the TILE Forth package; and a recursive Fibonacci number
3613: computation for benchmark calling performance. The following table shows
3614: the time taken for the benchmarks scaled by the time taken by Gforth (in
3615: other words, it shows the speedup factor that Gforth achieved over the
3616: other systems).
3617:
3618: @example
3619: relative Win32- NT This-
3620: time Gforth Forth Forth PFE Forth
3621: sieve 1.00 1.30 1.07 1.67 2.98
3622: bubble 1.00 1.30 1.40 1.66
3623: matmul 1.00 1.40 1.29 2.24
3624: fib 1.00 1.44 1.26 1.82 2.82
3625: @end example
3626:
3627: You may find the good performance of Gforth compared with the systems
3628: written in assembly language quite surprising. One important reason for
3629: the disappointing performance of these systems is probably that they are
3630: not written optimally for the 486 (e.g., they use the @code{lods}
3631: instruction). In addition, Win32Forth uses a comfortable, but costly
3632: method for relocating the Forth image: like @code{cforth}, it computes
3633: the actual addresses at run time, resulting in two address computations
3634: per NEXT (@pxref{System Architecture}).
3635:
3636: The speedup of Gforth over PFE and ThisForth can be easily explained
3637: with the self-imposed restriction to standard C (although the measured
3638: implementation of PFE uses a GNU C extension: global register
3639: variables), which makes efficient threading impossible. Moreover,
3640: current C compilers have a hard time optimizing other aspects of the
3641: ThisForth source.
3642:
3643: Note that the performance of Gforth on 386 architecture processors
3644: varies widely with the version of @code{gcc} used. E.g., @code{gcc-2.5.8}
3645: failed to allocate any of the virtual machine registers into real
3646: machine registers by itself and would not work correctly with explicit
3647: register declarations, giving a 1.3 times slower engine (on a 486DX2/66
3648: running the Sieve) than the one measured above.
3649:
3650: @node Bugs, Pedigree, Internals, Top
3651: @chapter Bugs
3652:
3653: Known bugs are described in the file BUGS in the Gforth distribution.
3654:
3655: If you find a bug, please send a bug report to !!. A bug report should
3656: describe the Gforth version used (it is announced at the start of an
3657: interactive Gforth session), the machine and operating system (on Unix
3658: systems you can use @code{uname -a} to produce this information), the
3659: installation options (!! a way to find them out), and a complete list of
3660: changes you (or your installer) have made to the Gforth sources (if
3661: any); it should contain a program (or a sequence of keyboard commands)
3662: that reproduces the bug and a description of what you think constitutes
3663: the buggy behaviour.
3664:
3665: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
3666: to Report Bugs, gcc.info, GNU C Manual}.
3667:
3668:
3669: @node Pedigree, Word Index, Bugs, Top
3670: @chapter Pedigree
3671:
3672: Gforth descends from BigForth (1993) and fig-Forth. Gforth and PFE (by
3673: Dirk Zoller) will cross-fertilize each other. Of course, a significant part of the design of Gforth was prescribed by ANS Forth.
3674:
3675: Bernd Paysan wrote BigForth, a descendent from TurboForth, an unreleased
3676: 32 bit native code version of VolksForth for the Atari ST, written
3677: mostly by Dietrich Weineck.
3678:
3679: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
3680: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
3681: UltraForth there) in the midth of the 80th and ported to Atari ST in
3682: 1986.
3683:
3684: Laxen and Perry wrote F83 as a model implementation of the
3685: Forth-83 standard. !! Pedigree? When?
3686:
3687: A team led by Bill Ragsdale implemented fig-Forth on many processors in
3688: 1979. Dean Sanderson and Bill Ragsdale developed the original
3689: implementation of fig-Forth based on microForth.
3690:
3691: microForth appears to be a downsized version of polyForth from Forth
3692: Inc. for the 8080 written in the midth of the 70th.
3693:
3694: A part of the information in this section comes from @cite{The Evolution
3695: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
3696: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
3697: Notices 28(3), 1993. You can find more historical and genealogical
3698: information about Forth there.
3699:
3700: @node Word Index, Node Index, Pedigree, Top
3701: @chapter Word Index
3702:
3703: This index is as incomplete as the manual. Each word is listed with
3704: stack effect and wordset.
3705:
3706: @printindex fn
3707:
3708: @node Node Index, , Word Index, Top
3709: @chapter Node Index
3710:
3711: This index is even less complete than the manual.
3712:
3713: @contents
3714: @bye
3715:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>