Annotation of gforth/doc/gforth.ds, revision 1.60
1.1 anton 1: \input texinfo @c -*-texinfo-*-
2: @comment The source is gforth.ds, from which gforth.texi is generated
1.28 crook 3:
1.21 crook 4: @comment TODO: nac29jan99 - a list of things to add in the next edit:
1.28 crook 5: @comment 1. x-ref all ambiguous or implementation-defined features?
6: @comment 2. Describe the use of Auser Avariable AConstant A, etc.
7: @comment 3. words in miscellaneous section need a home.
8: @comment 4. search for TODO for other minor and major works required.
9: @comment 5. [rats] change all @var to @i in Forth source so that info
10: @comment file looks decent.
1.36 anton 11: @c Not an improvement IMO - anton
12: @c and anyway, this should be taken up
13: @c with Karl Berry (the texinfo guy) - anton
1.29 crook 14: @comment .. would be useful to have a word that identified all deferred words
15: @comment should semantics stuff in intro be moved to another section
16:
1.28 crook 17:
1.1 anton 18: @comment %**start of header (This is for running Texinfo on a region.)
19: @setfilename gforth.info
20: @settitle Gforth Manual
21: @dircategory GNU programming tools
22: @direntry
23: * Gforth: (gforth). A fast interpreter for the Forth language.
24: @end direntry
1.49 anton 25: @c The Texinfo manual also recommends doing this, but for Gforth it may
26: @c not make much sense
27: @c @dircategory Individual utilities
28: @c @direntry
29: @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
30: @c @end direntry
31:
1.1 anton 32: @comment @setchapternewpage odd
1.29 crook 33: @comment TODO this gets left in by HTML converter
1.12 anton 34: @macro progstyle {}
35: Programming style note:
1.3 anton 36: @end macro
1.48 anton 37:
38: @macro assignment {}
39: @table @i
40: @item Assignment:
41: @end macro
42: @macro endassignment {}
43: @end table
44: @end macro
45:
1.1 anton 46: @comment %**end of header (This is for running Texinfo on a region.)
47:
1.29 crook 48:
49: @comment ----------------------------------------------------------
50: @comment macros for beautifying glossary entries
51: @comment if these are used, need to strip them out for HTML converter
52: @comment else they get repeated verbatim in HTML output.
53: @comment .. not working yet.
54:
55: @macro GLOSS-START {}
56: @iftex
57: @ninerm
58: @end iftex
59: @end macro
60:
61: @macro GLOSS-END {}
62: @iftex
63: @rm
64: @end iftex
65: @end macro
66:
67: @comment ----------------------------------------------------------
68:
69:
1.10 anton 70: @include version.texi
71:
1.49 anton 72: @ifnottex
1.11 anton 73: This file documents Gforth @value{VERSION}
1.1 anton 74:
1.26 crook 75: Copyright @copyright{} 1995-1999 Free Software Foundation, Inc.
1.1 anton 76:
77: Permission is granted to make and distribute verbatim copies of
78: this manual provided the copyright notice and this permission notice
79: are preserved on all copies.
80:
81: @ignore
82: Permission is granted to process this file through TeX and print the
83: results, provided the printed document carries a copying permission
84: notice identical to this one except for the removal of this paragraph
85: (this paragraph not being relevant to the printed manual).
86:
87: @end ignore
88: Permission is granted to copy and distribute modified versions of this
89: manual under the conditions for verbatim copying, provided also that the
90: sections entitled "Distribution" and "General Public License" are
91: included exactly as in the original, and provided that the entire
92: resulting derived work is distributed under the terms of a permission
93: notice identical to this one.
94:
95: Permission is granted to copy and distribute translations of this manual
96: into another language, under the above conditions for modified versions,
97: except that the sections entitled "Distribution" and "General Public
98: License" may be included in a translation approved by the author instead
99: of in the original English.
1.49 anton 100: @end ifnottex
1.1 anton 101:
102: @finalout
103: @titlepage
104: @sp 10
105: @center @titlefont{Gforth Manual}
106: @sp 2
1.11 anton 107: @center for version @value{VERSION}
1.1 anton 108: @sp 2
1.34 anton 109: @center Neal Crook
1.1 anton 110: @center Anton Ertl
1.6 pazsan 111: @center Bernd Paysan
1.5 anton 112: @center Jens Wilke
1.1 anton 113: @sp 3
1.47 crook 114: @center This manual is permanently under construction and was last updated on 15-Mar-2000
1.1 anton 115:
116: @comment The following two commands start the copyright page.
117: @page
118: @vskip 0pt plus 1filll
1.29 crook 119: Copyright @copyright{} 1995--1999 Free Software Foundation, Inc.
1.1 anton 120:
121: @comment !! Published by ... or You can get a copy of this manual ...
122:
123: Permission is granted to make and distribute verbatim copies of
124: this manual provided the copyright notice and this permission notice
125: are preserved on all copies.
126:
127: Permission is granted to copy and distribute modified versions of this
128: manual under the conditions for verbatim copying, provided also that the
129: sections entitled "Distribution" and "General Public License" are
130: included exactly as in the original, and provided that the entire
131: resulting derived work is distributed under the terms of a permission
132: notice identical to this one.
133:
134: Permission is granted to copy and distribute translations of this manual
135: into another language, under the above conditions for modified versions,
136: except that the sections entitled "Distribution" and "General Public
137: License" may be included in a translation approved by the author instead
138: of in the original English.
139: @end titlepage
140:
141: @node Top, License, (dir), (dir)
1.49 anton 142: @ifnottex
1.1 anton 143: Gforth is a free implementation of ANS Forth available on many
1.11 anton 144: personal machines. This manual corresponds to version @value{VERSION}.
1.49 anton 145: @end ifnottex
1.1 anton 146:
147: @menu
1.21 crook 148: * License:: The GPL
1.26 crook 149: * Goals:: About the Gforth Project
1.29 crook 150: * Gforth Environment:: Starting (and exiting) Gforth
1.48 anton 151: * Tutorial:: Hands-on Forth Tutorial
1.21 crook 152: * Introduction:: An introduction to ANS Forth
1.1 anton 153: * Words:: Forth words available in Gforth
1.24 anton 154: * Error messages:: How to interpret them
1.1 anton 155: * Tools:: Programming tools
156: * ANS conformance:: Implementation-defined options etc.
157: * Model:: The abstract machine of Gforth
158: * Integrating Gforth:: Forth as scripting language for applications
159: * Emacs and Gforth:: The Gforth Mode
160: * Image Files:: @code{.fi} files contain compiled code
161: * Engine:: The inner interpreter and the primitives
1.24 anton 162: * Binding to System Library::
1.13 pazsan 163: * Cross Compiler:: The Cross Compiler
1.1 anton 164: * Bugs:: How to report them
165: * Origin:: Authors and ancestors of Gforth
1.21 crook 166: * Forth-related information:: Books and places to look on the WWW
1.1 anton 167: * Word Index:: An item for each Forth word
1.41 anton 168: * Name Index:: Forth words, only names listed
1.1 anton 169: * Concept Index:: A menu covering many topics
1.12 anton 170:
1.48 anton 171: @detailmenu --- The Detailed Node Listing ---
1.12 anton 172:
1.26 crook 173: Goals of Gforth
174:
175: * Gforth Extensions Sinful?::
176:
1.29 crook 177: Gforth Environment
178:
1.32 anton 179: * Invoking Gforth:: Getting in
180: * Leaving Gforth:: Getting out
181: * Command-line editing::
1.48 anton 182: * Upper and lower case::
183: * Environment variables:: that affect how Gforth starts up
1.32 anton 184: * Gforth Files:: What gets installed and where
1.48 anton 185: * Startup speed:: When 35ms is not fast enough ...
186:
187: Forth Tutorial
188:
189: * Starting Gforth Tutorial::
190: * Syntax Tutorial::
191: * Crash Course Tutorial::
192: * Stack Tutorial::
193: * Arithmetics Tutorial::
194: * Stack Manipulation Tutorial::
195: * Using files for Forth code Tutorial::
196: * Comments Tutorial::
197: * Colon Definitions Tutorial::
198: * Decompilation Tutorial::
199: * Stack-Effect Comments Tutorial::
200: * Types Tutorial::
201: * Factoring Tutorial::
202: * Designing the stack effect Tutorial::
203: * Local Variables Tutorial::
204: * Conditional execution Tutorial::
205: * Flags and Comparisons Tutorial::
206: * General Loops Tutorial::
207: * Counted loops Tutorial::
208: * Recursion Tutorial::
209: * Leaving definitions or loops Tutorial::
210: * Return Stack Tutorial::
211: * Memory Tutorial::
212: * Characters and Strings Tutorial::
213: * Alignment Tutorial::
214: * Interpretation and Compilation Semantics and Immediacy Tutorial::
215: * Execution Tokens Tutorial::
216: * Exceptions Tutorial::
217: * Defining Words Tutorial::
218: * Arrays and Records Tutorial::
219: * POSTPONE Tutorial::
220: * Literal Tutorial::
221: * Advanced macros Tutorial::
222: * Compilation Tokens Tutorial::
223: * Wordlists and Search Order Tutorial::
1.29 crook 224:
1.24 anton 225: An Introduction to ANS Forth
226:
227: * Introducing the Text Interpreter::
228: * Stacks and Postfix notation::
229: * Your first definition::
230: * How does that work?::
231: * Forth is written in Forth::
232: * Review - elements of a Forth system::
1.29 crook 233: * Where to go next::
1.24 anton 234: * Exercises::
235:
1.12 anton 236: Forth Words
237:
238: * Notation::
1.21 crook 239: * Comments::
240: * Boolean Flags::
1.12 anton 241: * Arithmetic::
242: * Stack Manipulation::
243: * Memory::
244: * Control Structures::
245: * Defining Words::
1.47 crook 246: * Interpretation and Compilation Semantics::
247: * Tokens for Words::
1.21 crook 248: * The Text Interpreter::
249: * Word Lists::
250: * Environmental Queries::
1.12 anton 251: * Files::
252: * Blocks::
253: * Other I/O::
254: * Programming Tools::
255: * Assembler and Code Words::
256: * Threading Words::
1.26 crook 257: * Locals::
258: * Structures::
259: * Object-oriented Forth::
1.21 crook 260: * Passing Commands to the OS::
1.47 crook 261: * Keeping track of Time::
1.21 crook 262: * Miscellaneous Words::
1.12 anton 263:
264: Arithmetic
265:
266: * Single precision::
267: * Bitwise operations::
1.21 crook 268: * Double precision:: Double-cell integer arithmetic
269: * Numeric comparison::
1.32 anton 270: * Mixed precision:: Operations with single and double-cell integers
1.12 anton 271: * Floating Point::
272:
273: Stack Manipulation
274:
275: * Data stack::
276: * Floating point stack::
277: * Return stack::
278: * Locals stack::
279: * Stack pointer manipulation::
280:
281: Memory
282:
1.32 anton 283: * Memory model::
284: * Dictionary allocation::
285: * Heap Allocation::
286: * Memory Access::
287: * Address arithmetic::
288: * Memory Blocks::
1.12 anton 289:
290: Control Structures
291:
1.41 anton 292: * Selection:: IF ... ELSE ... ENDIF
293: * Simple Loops:: BEGIN ...
1.32 anton 294: * Counted Loops:: DO
295: * Arbitrary control structures::
296: * Calls and returns::
1.12 anton 297: * Exception Handling::
298:
299: Defining Words
300:
1.45 crook 301: * CREATE::
1.44 crook 302: * Variables:: Variables and user variables
303: * Constants::
304: * Values:: Initialised variables
1.32 anton 305: * Colon Definitions::
1.44 crook 306: * Anonymous Definitions:: Definitions without names
1.32 anton 307: * User-defined Defining Words::
1.44 crook 308: * Deferred words:: Allow forward references
309: * Aliases::
1.32 anton 310: * Supplying names::
1.47 crook 311:
1.59 anton 312: User-defined Defining Words
313:
314: * CREATE..DOES> applications::
315: * CREATE..DOES> details::
316: * Advanced does> usage example::
317:
1.47 crook 318: Interpretation and Compilation Semantics
319:
1.44 crook 320: * Combined words::
1.12 anton 321:
1.21 crook 322: The Text Interpreter
323:
1.29 crook 324: * Input Sources::
1.21 crook 325: * Number Conversion::
326: * Interpret/Compile states::
327: * Literals::
328: * Interpreter Directives::
329:
1.26 crook 330: Word Lists
331:
332: * Why use word lists?::
333: * Word list examples::
334:
335: Files
336:
1.48 anton 337: * Forth source files::
338: * General files::
339: * Search Paths::
340:
341: Search Paths
342:
343: * Forth Search Paths::
1.26 crook 344: * General Search Paths::
345:
346: Other I/O
347:
1.32 anton 348: * Simple numeric output:: Predefined formats
349: * Formatted numeric output:: Formatted (pictured) output
350: * String Formats:: How Forth stores strings in memory
351: * Displaying characters and strings:: Other stuff
352: * Input:: Input
1.26 crook 353:
354: Programming Tools
355:
356: * Debugging:: Simple and quick.
357: * Assertions:: Making your programs self-checking.
1.46 pazsan 358: * Singlestep Debugger:: Executing your program word by word.
1.26 crook 359:
1.59 anton 360: Assembler and Code Words
361:
362: * Code and ;code::
363: * Common Assembler:: Assembler Syntax
364: * Common Disassembler::
365: * 386 Assembler:: Deviations and special cases
366: * Alpha Assembler:: Deviations and special cases
367: * MIPS assembler:: Deviations and special cases
368: * Other assemblers:: How to write them
369:
1.26 crook 370: Locals
371:
372: * Gforth locals::
373: * ANS Forth locals::
374:
375: Gforth locals
376:
377: * Where are locals visible by name?::
378: * How long do locals live?::
379: * Programming Style::
380: * Implementation::
381:
1.12 anton 382: Structures
383:
384: * Why explicit structure support?::
385: * Structure Usage::
386: * Structure Naming Convention::
387: * Structure Implementation::
388: * Structure Glossary::
389:
390: Object-oriented Forth
391:
1.48 anton 392: * Why object-oriented programming?::
393: * Object-Oriented Terminology::
394: * Objects::
395: * OOF::
396: * Mini-OOF::
1.23 crook 397: * Comparison with other object models::
1.12 anton 398:
1.24 anton 399: The @file{objects.fs} model
1.12 anton 400:
401: * Properties of the Objects model::
402: * Basic Objects Usage::
1.41 anton 403: * The Objects base class::
1.12 anton 404: * Creating objects::
405: * Object-Oriented Programming Style::
406: * Class Binding::
407: * Method conveniences::
408: * Classes and Scoping::
1.41 anton 409: * Dividing classes::
1.12 anton 410: * Object Interfaces::
411: * Objects Implementation::
412: * Objects Glossary::
413:
1.24 anton 414: The @file{oof.fs} model
1.12 anton 415:
416: * Properties of the OOF model::
417: * Basic OOF Usage::
1.23 crook 418: * The OOF base class::
1.12 anton 419: * Class Declaration::
420: * Class Implementation::
421:
1.24 anton 422: The @file{mini-oof.fs} model
1.23 crook 423:
1.48 anton 424: * Basic Mini-OOF Usage::
425: * Mini-OOF Example::
426: * Mini-OOF Implementation::
427: * Comparison with other object models::
1.23 crook 428:
1.12 anton 429: Tools
430:
431: * ANS Report:: Report the words used, sorted by wordset.
432:
433: ANS conformance
434:
435: * The Core Words::
436: * The optional Block word set::
437: * The optional Double Number word set::
438: * The optional Exception word set::
439: * The optional Facility word set::
440: * The optional File-Access word set::
441: * The optional Floating-Point word set::
442: * The optional Locals word set::
443: * The optional Memory-Allocation word set::
444: * The optional Programming-Tools word set::
445: * The optional Search-Order word set::
446:
447: The Core Words
448:
449: * core-idef:: Implementation Defined Options
450: * core-ambcond:: Ambiguous Conditions
451: * core-other:: Other System Documentation
452:
453: The optional Block word set
454:
455: * block-idef:: Implementation Defined Options
456: * block-ambcond:: Ambiguous Conditions
457: * block-other:: Other System Documentation
458:
459: The optional Double Number word set
460:
461: * double-ambcond:: Ambiguous Conditions
462:
463: The optional Exception word set
464:
465: * exception-idef:: Implementation Defined Options
466:
467: The optional Facility word set
468:
469: * facility-idef:: Implementation Defined Options
470: * facility-ambcond:: Ambiguous Conditions
471:
472: The optional File-Access word set
473:
474: * file-idef:: Implementation Defined Options
475: * file-ambcond:: Ambiguous Conditions
476:
477: The optional Floating-Point word set
478:
479: * floating-idef:: Implementation Defined Options
480: * floating-ambcond:: Ambiguous Conditions
481:
482: The optional Locals word set
483:
484: * locals-idef:: Implementation Defined Options
485: * locals-ambcond:: Ambiguous Conditions
486:
487: The optional Memory-Allocation word set
488:
489: * memory-idef:: Implementation Defined Options
490:
491: The optional Programming-Tools word set
492:
493: * programming-idef:: Implementation Defined Options
494: * programming-ambcond:: Ambiguous Conditions
495:
496: The optional Search-Order word set
497:
498: * search-idef:: Implementation Defined Options
499: * search-ambcond:: Ambiguous Conditions
500:
501: Image Files
502:
1.24 anton 503: * Image Licensing Issues:: Distribution terms for images.
504: * Image File Background:: Why have image files?
1.32 anton 505: * Non-Relocatable Image Files:: don't always work.
1.24 anton 506: * Data-Relocatable Image Files:: are better.
1.32 anton 507: * Fully Relocatable Image Files:: better yet.
1.24 anton 508: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.32 anton 509: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.24 anton 510: * Modifying the Startup Sequence:: and turnkey applications.
1.12 anton 511:
512: Fully Relocatable Image Files
513:
1.27 crook 514: * gforthmi:: The normal way
1.12 anton 515: * cross.fs:: The hard way
516:
517: Engine
518:
519: * Portability::
520: * Threading::
521: * Primitives::
522: * Performance::
523:
524: Threading
525:
526: * Scheduling::
527: * Direct or Indirect Threaded?::
528: * DOES>::
529:
530: Primitives
531:
532: * Automatic Generation::
533: * TOS Optimization::
534: * Produced code::
1.13 pazsan 535:
536: Cross Compiler
537:
538: * Using the Cross Compiler::
539: * How the Cross Compiler Works::
540:
1.24 anton 541: Other Forth-related information
1.21 crook 542:
543: * Internet resources::
544: * Books::
545: * The Forth Interest Group::
546: * Conferences::
547:
1.24 anton 548: @end detailmenu
1.1 anton 549: @end menu
550:
1.26 crook 551: @node License, Goals, Top, Top
1.1 anton 552: @unnumbered GNU GENERAL PUBLIC LICENSE
553: @center Version 2, June 1991
554:
555: @display
556: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
557: 675 Mass Ave, Cambridge, MA 02139, USA
558:
559: Everyone is permitted to copy and distribute verbatim copies
560: of this license document, but changing it is not allowed.
561: @end display
562:
563: @unnumberedsec Preamble
564:
565: The licenses for most software are designed to take away your
566: freedom to share and change it. By contrast, the GNU General Public
567: License is intended to guarantee your freedom to share and change free
568: software---to make sure the software is free for all its users. This
569: General Public License applies to most of the Free Software
570: Foundation's software and to any other program whose authors commit to
571: using it. (Some other Free Software Foundation software is covered by
572: the GNU Library General Public License instead.) You can apply it to
573: your programs, too.
574:
575: When we speak of free software, we are referring to freedom, not
576: price. Our General Public Licenses are designed to make sure that you
577: have the freedom to distribute copies of free software (and charge for
578: this service if you wish), that you receive source code or can get it
579: if you want it, that you can change the software or use pieces of it
580: in new free programs; and that you know you can do these things.
581:
582: To protect your rights, we need to make restrictions that forbid
583: anyone to deny you these rights or to ask you to surrender the rights.
584: These restrictions translate to certain responsibilities for you if you
585: distribute copies of the software, or if you modify it.
586:
587: For example, if you distribute copies of such a program, whether
588: gratis or for a fee, you must give the recipients all the rights that
589: you have. You must make sure that they, too, receive or can get the
590: source code. And you must show them these terms so they know their
591: rights.
592:
593: We protect your rights with two steps: (1) copyright the software, and
594: (2) offer you this license which gives you legal permission to copy,
595: distribute and/or modify the software.
596:
597: Also, for each author's protection and ours, we want to make certain
598: that everyone understands that there is no warranty for this free
599: software. If the software is modified by someone else and passed on, we
600: want its recipients to know that what they have is not the original, so
601: that any problems introduced by others will not reflect on the original
602: authors' reputations.
603:
604: Finally, any free program is threatened constantly by software
605: patents. We wish to avoid the danger that redistributors of a free
606: program will individually obtain patent licenses, in effect making the
607: program proprietary. To prevent this, we have made it clear that any
608: patent must be licensed for everyone's free use or not licensed at all.
609:
610: The precise terms and conditions for copying, distribution and
611: modification follow.
612:
613: @iftex
614: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
615: @end iftex
1.49 anton 616: @ifnottex
1.1 anton 617: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1.49 anton 618: @end ifnottex
1.1 anton 619:
620: @enumerate 0
621: @item
622: This License applies to any program or other work which contains
623: a notice placed by the copyright holder saying it may be distributed
624: under the terms of this General Public License. The ``Program'', below,
625: refers to any such program or work, and a ``work based on the Program''
626: means either the Program or any derivative work under copyright law:
627: that is to say, a work containing the Program or a portion of it,
628: either verbatim or with modifications and/or translated into another
629: language. (Hereinafter, translation is included without limitation in
630: the term ``modification''.) Each licensee is addressed as ``you''.
631:
632: Activities other than copying, distribution and modification are not
633: covered by this License; they are outside its scope. The act of
634: running the Program is not restricted, and the output from the Program
635: is covered only if its contents constitute a work based on the
636: Program (independent of having been made by running the Program).
637: Whether that is true depends on what the Program does.
638:
639: @item
640: You may copy and distribute verbatim copies of the Program's
641: source code as you receive it, in any medium, provided that you
642: conspicuously and appropriately publish on each copy an appropriate
643: copyright notice and disclaimer of warranty; keep intact all the
644: notices that refer to this License and to the absence of any warranty;
645: and give any other recipients of the Program a copy of this License
646: along with the Program.
647:
648: You may charge a fee for the physical act of transferring a copy, and
649: you may at your option offer warranty protection in exchange for a fee.
650:
651: @item
652: You may modify your copy or copies of the Program or any portion
653: of it, thus forming a work based on the Program, and copy and
654: distribute such modifications or work under the terms of Section 1
655: above, provided that you also meet all of these conditions:
656:
657: @enumerate a
658: @item
659: You must cause the modified files to carry prominent notices
660: stating that you changed the files and the date of any change.
661:
662: @item
663: You must cause any work that you distribute or publish, that in
664: whole or in part contains or is derived from the Program or any
665: part thereof, to be licensed as a whole at no charge to all third
666: parties under the terms of this License.
667:
668: @item
669: If the modified program normally reads commands interactively
670: when run, you must cause it, when started running for such
671: interactive use in the most ordinary way, to print or display an
672: announcement including an appropriate copyright notice and a
673: notice that there is no warranty (or else, saying that you provide
674: a warranty) and that users may redistribute the program under
675: these conditions, and telling the user how to view a copy of this
676: License. (Exception: if the Program itself is interactive but
677: does not normally print such an announcement, your work based on
678: the Program is not required to print an announcement.)
679: @end enumerate
680:
681: These requirements apply to the modified work as a whole. If
682: identifiable sections of that work are not derived from the Program,
683: and can be reasonably considered independent and separate works in
684: themselves, then this License, and its terms, do not apply to those
685: sections when you distribute them as separate works. But when you
686: distribute the same sections as part of a whole which is a work based
687: on the Program, the distribution of the whole must be on the terms of
688: this License, whose permissions for other licensees extend to the
689: entire whole, and thus to each and every part regardless of who wrote it.
690:
691: Thus, it is not the intent of this section to claim rights or contest
692: your rights to work written entirely by you; rather, the intent is to
693: exercise the right to control the distribution of derivative or
694: collective works based on the Program.
695:
696: In addition, mere aggregation of another work not based on the Program
697: with the Program (or with a work based on the Program) on a volume of
698: a storage or distribution medium does not bring the other work under
699: the scope of this License.
700:
701: @item
702: You may copy and distribute the Program (or a work based on it,
703: under Section 2) in object code or executable form under the terms of
704: Sections 1 and 2 above provided that you also do one of the following:
705:
706: @enumerate a
707: @item
708: Accompany it with the complete corresponding machine-readable
709: source code, which must be distributed under the terms of Sections
710: 1 and 2 above on a medium customarily used for software interchange; or,
711:
712: @item
713: Accompany it with a written offer, valid for at least three
714: years, to give any third party, for a charge no more than your
715: cost of physically performing source distribution, a complete
716: machine-readable copy of the corresponding source code, to be
717: distributed under the terms of Sections 1 and 2 above on a medium
718: customarily used for software interchange; or,
719:
720: @item
721: Accompany it with the information you received as to the offer
722: to distribute corresponding source code. (This alternative is
723: allowed only for noncommercial distribution and only if you
724: received the program in object code or executable form with such
725: an offer, in accord with Subsection b above.)
726: @end enumerate
727:
728: The source code for a work means the preferred form of the work for
729: making modifications to it. For an executable work, complete source
730: code means all the source code for all modules it contains, plus any
731: associated interface definition files, plus the scripts used to
732: control compilation and installation of the executable. However, as a
733: special exception, the source code distributed need not include
734: anything that is normally distributed (in either source or binary
735: form) with the major components (compiler, kernel, and so on) of the
736: operating system on which the executable runs, unless that component
737: itself accompanies the executable.
738:
739: If distribution of executable or object code is made by offering
740: access to copy from a designated place, then offering equivalent
741: access to copy the source code from the same place counts as
742: distribution of the source code, even though third parties are not
743: compelled to copy the source along with the object code.
744:
745: @item
746: You may not copy, modify, sublicense, or distribute the Program
747: except as expressly provided under this License. Any attempt
748: otherwise to copy, modify, sublicense or distribute the Program is
749: void, and will automatically terminate your rights under this License.
750: However, parties who have received copies, or rights, from you under
751: this License will not have their licenses terminated so long as such
752: parties remain in full compliance.
753:
754: @item
755: You are not required to accept this License, since you have not
756: signed it. However, nothing else grants you permission to modify or
757: distribute the Program or its derivative works. These actions are
758: prohibited by law if you do not accept this License. Therefore, by
759: modifying or distributing the Program (or any work based on the
760: Program), you indicate your acceptance of this License to do so, and
761: all its terms and conditions for copying, distributing or modifying
762: the Program or works based on it.
763:
764: @item
765: Each time you redistribute the Program (or any work based on the
766: Program), the recipient automatically receives a license from the
767: original licensor to copy, distribute or modify the Program subject to
768: these terms and conditions. You may not impose any further
769: restrictions on the recipients' exercise of the rights granted herein.
770: You are not responsible for enforcing compliance by third parties to
771: this License.
772:
773: @item
774: If, as a consequence of a court judgment or allegation of patent
775: infringement or for any other reason (not limited to patent issues),
776: conditions are imposed on you (whether by court order, agreement or
777: otherwise) that contradict the conditions of this License, they do not
778: excuse you from the conditions of this License. If you cannot
779: distribute so as to satisfy simultaneously your obligations under this
780: License and any other pertinent obligations, then as a consequence you
781: may not distribute the Program at all. For example, if a patent
782: license would not permit royalty-free redistribution of the Program by
783: all those who receive copies directly or indirectly through you, then
784: the only way you could satisfy both it and this License would be to
785: refrain entirely from distribution of the Program.
786:
787: If any portion of this section is held invalid or unenforceable under
788: any particular circumstance, the balance of the section is intended to
789: apply and the section as a whole is intended to apply in other
790: circumstances.
791:
792: It is not the purpose of this section to induce you to infringe any
793: patents or other property right claims or to contest validity of any
794: such claims; this section has the sole purpose of protecting the
795: integrity of the free software distribution system, which is
796: implemented by public license practices. Many people have made
797: generous contributions to the wide range of software distributed
798: through that system in reliance on consistent application of that
799: system; it is up to the author/donor to decide if he or she is willing
800: to distribute software through any other system and a licensee cannot
801: impose that choice.
802:
803: This section is intended to make thoroughly clear what is believed to
804: be a consequence of the rest of this License.
805:
806: @item
807: If the distribution and/or use of the Program is restricted in
808: certain countries either by patents or by copyrighted interfaces, the
809: original copyright holder who places the Program under this License
810: may add an explicit geographical distribution limitation excluding
811: those countries, so that distribution is permitted only in or among
812: countries not thus excluded. In such case, this License incorporates
813: the limitation as if written in the body of this License.
814:
815: @item
816: The Free Software Foundation may publish revised and/or new versions
817: of the General Public License from time to time. Such new versions will
818: be similar in spirit to the present version, but may differ in detail to
819: address new problems or concerns.
820:
821: Each version is given a distinguishing version number. If the Program
822: specifies a version number of this License which applies to it and ``any
823: later version'', you have the option of following the terms and conditions
824: either of that version or of any later version published by the Free
825: Software Foundation. If the Program does not specify a version number of
826: this License, you may choose any version ever published by the Free Software
827: Foundation.
828:
829: @item
830: If you wish to incorporate parts of the Program into other free
831: programs whose distribution conditions are different, write to the author
832: to ask for permission. For software which is copyrighted by the Free
833: Software Foundation, write to the Free Software Foundation; we sometimes
834: make exceptions for this. Our decision will be guided by the two goals
835: of preserving the free status of all derivatives of our free software and
836: of promoting the sharing and reuse of software generally.
837:
838: @iftex
839: @heading NO WARRANTY
840: @end iftex
1.49 anton 841: @ifnottex
1.1 anton 842: @center NO WARRANTY
1.49 anton 843: @end ifnottex
1.1 anton 844:
845: @item
846: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
847: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
848: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
849: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
850: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
851: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
852: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
853: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
854: REPAIR OR CORRECTION.
855:
856: @item
857: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
858: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
859: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
860: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
861: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
862: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
863: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
864: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
865: POSSIBILITY OF SUCH DAMAGES.
866: @end enumerate
867:
868: @iftex
869: @heading END OF TERMS AND CONDITIONS
870: @end iftex
1.49 anton 871: @ifnottex
1.1 anton 872: @center END OF TERMS AND CONDITIONS
1.49 anton 873: @end ifnottex
1.1 anton 874:
875: @page
876: @unnumberedsec How to Apply These Terms to Your New Programs
877:
878: If you develop a new program, and you want it to be of the greatest
879: possible use to the public, the best way to achieve this is to make it
880: free software which everyone can redistribute and change under these terms.
881:
882: To do so, attach the following notices to the program. It is safest
883: to attach them to the start of each source file to most effectively
884: convey the exclusion of warranty; and each file should have at least
885: the ``copyright'' line and a pointer to where the full notice is found.
886:
887: @smallexample
888: @var{one line to give the program's name and a brief idea of what it does.}
889: Copyright (C) 19@var{yy} @var{name of author}
890:
891: This program is free software; you can redistribute it and/or modify
892: it under the terms of the GNU General Public License as published by
893: the Free Software Foundation; either version 2 of the License, or
894: (at your option) any later version.
895:
896: This program is distributed in the hope that it will be useful,
897: but WITHOUT ANY WARRANTY; without even the implied warranty of
898: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
899: GNU General Public License for more details.
900:
901: You should have received a copy of the GNU General Public License
902: along with this program; if not, write to the Free Software
903: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
904: @end smallexample
905:
906: Also add information on how to contact you by electronic and paper mail.
907:
908: If the program is interactive, make it output a short notice like this
909: when it starts in an interactive mode:
910:
911: @smallexample
912: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
913: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
914: type `show w'.
915: This is free software, and you are welcome to redistribute it
916: under certain conditions; type `show c' for details.
917: @end smallexample
918:
919: The hypothetical commands @samp{show w} and @samp{show c} should show
920: the appropriate parts of the General Public License. Of course, the
921: commands you use may be called something other than @samp{show w} and
922: @samp{show c}; they could even be mouse-clicks or menu items---whatever
923: suits your program.
924:
925: You should also get your employer (if you work as a programmer) or your
926: school, if any, to sign a ``copyright disclaimer'' for the program, if
927: necessary. Here is a sample; alter the names:
928:
929: @smallexample
930: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
931: `Gnomovision' (which makes passes at compilers) written by James Hacker.
932:
933: @var{signature of Ty Coon}, 1 April 1989
934: Ty Coon, President of Vice
935: @end smallexample
936:
937: This General Public License does not permit incorporating your program into
938: proprietary programs. If your program is a subroutine library, you may
939: consider it more useful to permit linking proprietary applications with the
940: library. If this is what you want to do, use the GNU Library General
941: Public License instead of this License.
942:
943: @iftex
944: @unnumbered Preface
945: @cindex Preface
1.21 crook 946: This manual documents Gforth. Some introductory material is provided for
947: readers who are unfamiliar with Forth or who are migrating to Gforth
948: from other Forth compilers. However, this manual is primarily a
949: reference manual.
1.1 anton 950: @end iftex
951:
1.28 crook 952: @comment TODO much more blurb here.
1.26 crook 953:
954: @c ******************************************************************
1.29 crook 955: @node Goals, Gforth Environment, License, Top
1.26 crook 956: @comment node-name, next, previous, up
957: @chapter Goals of Gforth
958: @cindex goals of the Gforth project
959: The goal of the Gforth Project is to develop a standard model for
960: ANS Forth. This can be split into several subgoals:
961:
962: @itemize @bullet
963: @item
964: Gforth should conform to the ANS Forth Standard.
965: @item
966: It should be a model, i.e. it should define all the
967: implementation-dependent things.
968: @item
969: It should become standard, i.e. widely accepted and used. This goal
970: is the most difficult one.
971: @end itemize
972:
973: To achieve these goals Gforth should be
974: @itemize @bullet
975: @item
976: Similar to previous models (fig-Forth, F83)
977: @item
978: Powerful. It should provide for all the things that are considered
979: necessary today and even some that are not yet considered necessary.
980: @item
981: Efficient. It should not get the reputation of being exceptionally
982: slow.
983: @item
984: Free.
985: @item
986: Available on many machines/easy to port.
987: @end itemize
988:
989: Have we achieved these goals? Gforth conforms to the ANS Forth
990: standard. It may be considered a model, but we have not yet documented
991: which parts of the model are stable and which parts we are likely to
992: change. It certainly has not yet become a de facto standard, but it
993: appears to be quite popular. It has some similarities to and some
994: differences from previous models. It has some powerful features, but not
995: yet everything that we envisioned. We certainly have achieved our
996: execution speed goals (@pxref{Performance}). It is free and available
997: on many machines.
998:
999: @menu
1000: * Gforth Extensions Sinful?::
1001: @end menu
1002:
1.48 anton 1003: @node Gforth Extensions Sinful?, , Goals, Goals
1.26 crook 1004: @comment node-name, next, previous, up
1005: @section Is it a Sin to use Gforth Extensions?
1006: @cindex Gforth extensions
1007:
1008: If you've been paying attention, you will have realised that there is an
1009: ANS (American National Standard) for Forth. As you read through the rest
1.29 crook 1010: of this manual, you will see documentation for @i{Standard} words, and
1011: documentation for some appealing Gforth @i{extensions}. You might ask
1012: yourself the question: @i{``Given that there is a standard, would I be
1.45 crook 1013: committing a sin if I use (non-Standard) Gforth extensions?''}
1.26 crook 1014:
1015: The answer to that question is somewhat pragmatic and somewhat
1016: philosophical. Consider these points:
1017:
1018: @itemize @bullet
1019: @item
1020: A number of the Gforth extensions can be implemented in ANS Forth using
1021: files provided in the @file{compat/} directory. These are mentioned in
1022: the text in passing.
1023: @item
1024: Forth has a rich historical precedent for programmers taking advantage
1025: of implementation-dependent features of their tools (for example,
1026: relying on a knowledge of the dictionary structure). Sometimes these
1027: techniques are necessary to extract every last bit of performance from
1028: the hardware, sometimes they are just a programming shorthand.
1029: @item
1030: The best way to break the rules is to know what the rules are. To learn
1031: the rules, there is no substitute for studying the text of the Standard
1032: itself. In particular, Appendix A of the Standard (@var{Rationale})
1033: provides a valuable insight into the thought processes of the technical
1034: committee.
1035: @item
1036: The best reason to break a rule is because you have to; because it's
1037: more productive to do that, because it makes your code run fast enough
1038: or because you can see no Standard way to achieve what you want to
1039: achieve.
1040: @end itemize
1041:
1042: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
1043: analyse your program and determine what non-Standard definitions it
1044: relies upon.
1045:
1.29 crook 1046:
1.26 crook 1047: @c ******************************************************************
1.48 anton 1048: @node Gforth Environment, Tutorial, Goals, Top
1.29 crook 1049: @chapter Gforth Environment
1050: @cindex Gforth environment
1.21 crook 1051:
1.45 crook 1052: Note: ultimately, the Gforth man page will be auto-generated from the
1.29 crook 1053: material in this chapter.
1.21 crook 1054:
1055: @menu
1.29 crook 1056: * Invoking Gforth:: Getting in
1057: * Leaving Gforth:: Getting out
1058: * Command-line editing::
1.48 anton 1059: * Upper and lower case::
1060: * Environment variables:: that affect how Gforth starts up
1.29 crook 1061: * Gforth Files:: What gets installed and where
1.48 anton 1062: * Startup speed:: When 35ms is not fast enough ...
1.21 crook 1063: @end menu
1064:
1.49 anton 1065: For related information about the creation of images see @ref{Image Files}.
1.29 crook 1066:
1.21 crook 1067: @comment ----------------------------------------------
1.48 anton 1068: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29 crook 1069: @section Invoking Gforth
1070: @cindex invoking Gforth
1071: @cindex running Gforth
1072: @cindex command-line options
1073: @cindex options on the command line
1074: @cindex flags on the command line
1.21 crook 1075:
1.30 anton 1076: Gforth is made up of two parts; an executable ``engine'' (named
1077: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1078: will usually just say @code{gforth} -- this automatically loads the
1079: default image file @file{gforth.fi}. In many other cases the default
1080: Gforth image will be invoked like this:
1.21 crook 1081: @example
1.30 anton 1082: gforth [file | -e forth-code] ...
1.21 crook 1083: @end example
1.29 crook 1084: @noindent
1085: This interprets the contents of the files and the Forth code in the order they
1086: are given.
1.21 crook 1087:
1.30 anton 1088: In addition to the @file{gforth} engine, there is also an engine called
1089: @file{gforth-fast}, which is faster, but gives less informative error
1090: messages (@pxref{Error messages}).
1091:
1.29 crook 1092: In general, the command line looks like this:
1.21 crook 1093:
1094: @example
1.30 anton 1095: gforth[-fast] [engine options] [image options]
1.21 crook 1096: @end example
1097:
1.30 anton 1098: The engine options must come before the rest of the command
1.29 crook 1099: line. They are:
1.26 crook 1100:
1.29 crook 1101: @table @code
1102: @cindex -i, command-line option
1103: @cindex --image-file, command-line option
1104: @item --image-file @i{file}
1105: @itemx -i @i{file}
1106: Loads the Forth image @i{file} instead of the default
1107: @file{gforth.fi} (@pxref{Image Files}).
1.21 crook 1108:
1.39 anton 1109: @cindex --appl-image, command-line option
1110: @item --appl-image @i{file}
1111: Loads the image @i{file} and leaves all further command-line arguments
1112: to the image (instead of processing them as options). This is useful
1113: for building executable application images on Unix, built with
1114: @code{gforthmi --application ...}.
1115:
1.29 crook 1116: @cindex --path, command-line option
1117: @cindex -p, command-line option
1118: @item --path @i{path}
1119: @itemx -p @i{path}
1120: Uses @i{path} for searching the image file and Forth source code files
1121: instead of the default in the environment variable @code{GFORTHPATH} or
1122: the path specified at installation time (e.g.,
1123: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1124: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21 crook 1125:
1.29 crook 1126: @cindex --dictionary-size, command-line option
1127: @cindex -m, command-line option
1128: @cindex @i{size} parameters for command-line options
1129: @cindex size of the dictionary and the stacks
1130: @item --dictionary-size @i{size}
1131: @itemx -m @i{size}
1132: Allocate @i{size} space for the Forth dictionary space instead of
1133: using the default specified in the image (typically 256K). The
1134: @i{size} specification for this and subsequent options consists of
1135: an integer and a unit (e.g.,
1136: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1137: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1138: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1139: @code{e} is used.
1.21 crook 1140:
1.29 crook 1141: @cindex --data-stack-size, command-line option
1142: @cindex -d, command-line option
1143: @item --data-stack-size @i{size}
1144: @itemx -d @i{size}
1145: Allocate @i{size} space for the data stack instead of using the
1146: default specified in the image (typically 16K).
1.21 crook 1147:
1.29 crook 1148: @cindex --return-stack-size, command-line option
1149: @cindex -r, command-line option
1150: @item --return-stack-size @i{size}
1151: @itemx -r @i{size}
1152: Allocate @i{size} space for the return stack instead of using the
1153: default specified in the image (typically 15K).
1.21 crook 1154:
1.29 crook 1155: @cindex --fp-stack-size, command-line option
1156: @cindex -f, command-line option
1157: @item --fp-stack-size @i{size}
1158: @itemx -f @i{size}
1159: Allocate @i{size} space for the floating point stack instead of
1160: using the default specified in the image (typically 15.5K). In this case
1161: the unit specifier @code{e} refers to floating point numbers.
1.21 crook 1162:
1.48 anton 1163: @cindex --locals-stack-size, command-line option
1164: @cindex -l, command-line option
1165: @item --locals-stack-size @i{size}
1166: @itemx -l @i{size}
1167: Allocate @i{size} space for the locals stack instead of using the
1168: default specified in the image (typically 14.5K).
1169:
1170: @cindex -h, command-line option
1171: @cindex --help, command-line option
1172: @item --help
1173: @itemx -h
1174: Print a message about the command-line options
1175:
1176: @cindex -v, command-line option
1177: @cindex --version, command-line option
1178: @item --version
1179: @itemx -v
1180: Print version and exit
1181:
1182: @cindex --debug, command-line option
1183: @item --debug
1184: Print some information useful for debugging on startup.
1185:
1186: @cindex --offset-image, command-line option
1187: @item --offset-image
1188: Start the dictionary at a slightly different position than would be used
1189: otherwise (useful for creating data-relocatable images,
1190: @pxref{Data-Relocatable Image Files}).
1191:
1192: @cindex --no-offset-im, command-line option
1193: @item --no-offset-im
1194: Start the dictionary at the normal position.
1195:
1196: @cindex --clear-dictionary, command-line option
1197: @item --clear-dictionary
1198: Initialize all bytes in the dictionary to 0 before loading the image
1199: (@pxref{Data-Relocatable Image Files}).
1200:
1201: @cindex --die-on-signal, command-line-option
1202: @item --die-on-signal
1203: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1204: or the segmentation violation SIGSEGV) by translating it into a Forth
1205: @code{THROW}. With this option, Gforth exits if it receives such a
1206: signal. This option is useful when the engine and/or the image might be
1207: severely broken (such that it causes another signal before recovering
1208: from the first); this option avoids endless loops in such cases.
1209: @end table
1210:
1211: @cindex loading files at startup
1212: @cindex executing code on startup
1213: @cindex batch processing with Gforth
1214: As explained above, the image-specific command-line arguments for the
1215: default image @file{gforth.fi} consist of a sequence of filenames and
1216: @code{-e @var{forth-code}} options that are interpreted in the sequence
1217: in which they are given. The @code{-e @var{forth-code}} or
1218: @code{--evaluate @var{forth-code}} option evaluates the Forth
1219: code. This option takes only one argument; if you want to evaluate more
1220: Forth words, you have to quote them or use @code{-e} several times. To exit
1221: after processing the command line (instead of entering interactive mode)
1222: append @code{-e bye} to the command line.
1223:
1224: @cindex versions, invoking other versions of Gforth
1225: If you have several versions of Gforth installed, @code{gforth} will
1226: invoke the version that was installed last. @code{gforth-@i{version}}
1227: invokes a specific version. If your environment contains the variable
1228: @code{GFORTHPATH}, you may want to override it by using the
1229: @code{--path} option.
1230:
1231: Not yet implemented:
1232: On startup the system first executes the system initialization file
1233: (unless the option @code{--no-init-file} is given; note that the system
1234: resulting from using this option may not be ANS Forth conformant). Then
1235: the user initialization file @file{.gforth.fs} is executed, unless the
1236: option @code{--no-rc} is given; this file is first searched in @file{.},
1237: then in @file{~}, then in the normal path (see above).
1238:
1239:
1240:
1241: @comment ----------------------------------------------
1242: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1243: @section Leaving Gforth
1244: @cindex Gforth - leaving
1245: @cindex leaving Gforth
1246:
1247: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1248: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1249: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49 anton 1250: data are discarded. For ways of saving the state of the system before
1251: leaving Gforth see @ref{Image Files}.
1.48 anton 1252:
1253: doc-bye
1254:
1255:
1256: @comment ----------------------------------------------
1257: @node Command-line editing, Upper and lower case, Leaving Gforth, Gforth Environment
1258: @section Command-line editing
1259: @cindex command-line editing
1260:
1261: Gforth maintains a history file that records every line that you type to
1262: the text interpreter. This file is preserved between sessions, and is
1263: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1264: repeatedly you can recall successively older commands from this (or
1265: previous) session(s). The full list of command-line editing facilities is:
1266:
1267: @itemize @bullet
1268: @item
1269: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1270: commands from the history buffer.
1271: @item
1272: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1273: from the history buffer.
1274: @item
1275: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1276: @item
1277: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1278: @item
1279: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1280: closing up the line.
1281: @item
1282: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1283: @item
1284: @kbd{Ctrl-a} to move the cursor to the start of the line.
1285: @item
1286: @kbd{Ctrl-e} to move the cursor to the end of the line.
1287: @item
1288: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1289: line.
1290: @item
1291: @key{TAB} to step through all possible full-word completions of the word
1292: currently being typed.
1293: @item
1294: @kbd{Ctrl-d} at the start of the line to terminate Gforth (gracefully,
1295: using @code{bye}).
1296: @end itemize
1297:
1298: When editing, displayable characters are inserted to the left of the
1299: cursor position; the line is always in ``insert'' (as opposed to
1300: ``overstrike'') mode.
1301:
1302: @cindex history file
1303: @cindex @file{.gforth-history}
1304: On Unix systems, the history file is @file{~/.gforth-history} by
1305: default@footnote{i.e. it is stored in the user's home directory.}. You
1306: can find out the name and location of your history file using:
1307:
1308: @example
1309: history-file type \ Unix-class systems
1310:
1311: history-file type \ Other systems
1312: history-dir type
1313: @end example
1314:
1315: If you enter long definitions by hand, you can use a text editor to
1316: paste them out of the history file into a Forth source file for reuse at
1317: a later time.
1318:
1319: Gforth never trims the size of the history file, so you should do this
1320: periodically, if necessary.
1321:
1322: @comment this is all defined in history.fs
1323: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1324: @comment chosen?
1325:
1326:
1327:
1328: @comment ----------------------------------------------
1329: @node Upper and lower case, Environment variables, Command-line editing, Gforth Environment
1330: @section Upper and lower case
1331: @cindex case-sensitivity
1332: @cindex upper and lower case
1333:
1334: Gforth is case-insensitive; you can enter definitions and invoke
1335: Standard words using upper, lower or mixed case (however,
1336: @pxref{core-idef, Implementation-defined options, Implementation-defined
1337: options}).
1338:
1339: ANS Forth only @i{requires} implementations to recognise Standard words
1340: when they are typed entirely in upper case. Therefore, a Standard
1341: program must use upper case for all Standard words. You can use whatever
1342: case you like for words that you define, but in a standard program you
1343: have to use the words in the same case that you defined them.
1344:
1345: Gforth supports case sensitivity through @code{table}s (case-sensitive
1346: wordlists, @pxref{Word Lists}).
1347:
1348: Two people have asked how to convert Gforth to case sensitivity; while
1349: we think this is a bad idea, you can change all wordlists into tables
1350: like this:
1351:
1352: @example
1353: ' table-find forth-wordlist wordlist-map @ !
1354: @end example
1355:
1356: Note that you now have to type the predefined words in the same case
1357: that we defined them, which are varying. You may want to convert them
1358: to your favourite case before doing this operation (I won't explain how,
1359: because if you are even contemplating to do this, you'd better have
1360: enough knowledge of Forth systems to know this already).
1361:
1362: @comment ----------------------------------------------
1363: @node Environment variables, Gforth Files, Upper and lower case, Gforth Environment
1364: @section Environment variables
1365: @cindex environment variables
1366:
1367: Gforth uses these environment variables:
1368:
1369: @itemize @bullet
1370: @item
1371: @cindex @code{GFORTHHIST} -- environment variable
1372: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1373: open/create the history file, @file{.gforth-history}. Default:
1374: @code{$HOME}.
1375:
1376: @item
1377: @cindex @code{GFORTHPATH} -- environment variable
1378: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1379: for Forth source-code files.
1380:
1381: @item
1382: @cindex @code{GFORTH} -- environment variable
1.49 anton 1383: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1384:
1385: @item
1386: @cindex @code{GFORTHD} -- environment variable
1387: @code{GFORTHD} -- used by @file{gforthmi} @xref{gforthmi}.
1388:
1389: @item
1390: @cindex @code{TMP}, @code{TEMP} - environment variable
1391: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1392: location for the history file.
1393: @end itemize
1394:
1395: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1396: @comment mentioning these.
1397:
1398: All the Gforth environment variables default to sensible values if they
1399: are not set.
1400:
1401:
1402: @comment ----------------------------------------------
1403: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1404: @section Gforth files
1405: @cindex Gforth files
1406:
1407: When you install Gforth on a Unix system, it installs files in these
1408: locations by default:
1409:
1410: @itemize @bullet
1411: @item
1412: @file{/usr/local/bin/gforth}
1413: @item
1414: @file{/usr/local/bin/gforthmi}
1415: @item
1416: @file{/usr/local/man/man1/gforth.1} - man page.
1417: @item
1418: @file{/usr/local/info} - the Info version of this manual.
1419: @item
1420: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1421: @item
1422: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1423: @item
1424: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1425: @item
1426: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1427: @end itemize
1428:
1429: You can select different places for installation by using
1430: @code{configure} options (listed with @code{configure --help}).
1431:
1432: @comment ----------------------------------------------
1433: @node Startup speed, , Gforth Files, Gforth Environment
1434: @section Startup speed
1435: @cindex Startup speed
1436: @cindex speed, startup
1437:
1438: If Gforth is used for CGI scripts or in shell scripts, its startup
1439: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1440: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1441: system time.
1442:
1443: If startup speed is a problem, you may consider the following ways to
1444: improve it; or you may consider ways to reduce the number of startups
1445: (e.g., Fast-CGI).
1446:
1447: The first step to improve startup speed is to statically link Gforth, by
1448: building it with @code{XLDFLAGS=-static}. This requires more memory for
1449: the code and will therefore slow down the first invocation, but
1450: subsequent invocations avoid the dynamic linking overhead. Another
1451: disadvantage is that Gforth won't profit from library upgrades. As a
1452: result, @code{gforth-static -e bye} takes about 17.1ms user and
1453: 8.2ms system time.
1454:
1455: The next step to improve startup speed is to use a non-relocatable image
1456: @ref{Non-Relocatable Image Files}. You can create this image with
1457: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1458: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1459: and a part of the copy-on-write overhead. The disadvantage is that the
1460: nonrelocatable image does not work if the OS gives Gforth a different
1461: address for the dictionary, for whatever reason; so you better provide a
1462: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1463: bye} takes about 15.3ms user and 7.5ms system time.
1464:
1465: The final step is to disable dictionary hashing in Gforth. Gforth
1466: builds the hash table on startup, which takes much of the startup
1467: overhead. You can do this by commenting out the @code{include hash.fs}
1468: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1469: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1470: The disadvantages are that functionality like @code{table} and
1471: @code{ekey} is missing and that text interpretation (e.g., compiling)
1472: now takes much longer. So, you should only use this method if there is
1473: no significant text interpretation to perform (the script should be
1474: compiled into the image, among other things). @code{gforth-static -i
1475: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1476:
1477: @c ******************************************************************
1478: @node Tutorial, Introduction, Gforth Environment, Top
1479: @chapter Forth Tutorial
1480: @cindex Tutorial
1481: @cindex Forth Tutorial
1482:
1483: This tutorial can be used with any ANS-compliant Forth; any places that
1484: mention features specific to Gforth are marked as such and you can skip
1485: them, if you work with another Forth. This tutorial does not explain
1486: all features of Forth, just enough to get you started and give you some
1487: ideas about the facilities available in Forth. Read the rest of the
1488: manual and the standard when you are through this.
1489:
1490: The intended way to use this tutorial is that you work through it while
1491: sitting in front of the console, take a look at the examples and predict
1492: what they will do, then try them out; if the outcome is not as expected,
1493: find out why (e.g., by trying out variations of the example), so you
1494: understand what's going on. There are also some assignments that you
1495: should solve.
1496:
1497: This tutorial assumes that you have programmed before and know what,
1498: e.g., a loop is.
1499:
1500: @c !! explain compat library
1501:
1502: @menu
1503: * Starting Gforth Tutorial::
1504: * Syntax Tutorial::
1505: * Crash Course Tutorial::
1506: * Stack Tutorial::
1507: * Arithmetics Tutorial::
1508: * Stack Manipulation Tutorial::
1509: * Using files for Forth code Tutorial::
1510: * Comments Tutorial::
1511: * Colon Definitions Tutorial::
1512: * Decompilation Tutorial::
1513: * Stack-Effect Comments Tutorial::
1514: * Types Tutorial::
1515: * Factoring Tutorial::
1516: * Designing the stack effect Tutorial::
1517: * Local Variables Tutorial::
1518: * Conditional execution Tutorial::
1519: * Flags and Comparisons Tutorial::
1520: * General Loops Tutorial::
1521: * Counted loops Tutorial::
1522: * Recursion Tutorial::
1523: * Leaving definitions or loops Tutorial::
1524: * Return Stack Tutorial::
1525: * Memory Tutorial::
1526: * Characters and Strings Tutorial::
1527: * Alignment Tutorial::
1528: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1529: * Execution Tokens Tutorial::
1530: * Exceptions Tutorial::
1531: * Defining Words Tutorial::
1532: * Arrays and Records Tutorial::
1533: * POSTPONE Tutorial::
1534: * Literal Tutorial::
1535: * Advanced macros Tutorial::
1536: * Compilation Tokens Tutorial::
1537: * Wordlists and Search Order Tutorial::
1538: @end menu
1539:
1540: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1541: @section Starting Gforth
1542:
1543: You can start Gforth by typing its name:
1544:
1545: @example
1546: gforth
1547: @end example
1548:
1549: That puts you into interactive mode; you can leave Gforth by typing
1550: @code{bye}. While in Gforth, you can edit the command line and access
1551: the command line history with cursor keys, similar to bash.
1552:
1553:
1554: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1555: @section Syntax
1556:
1557: A @dfn{word} is a sequence of arbitrary characters (expcept white
1558: space). Words are separated by white space. E.g., each of the
1559: following lines contains exactly one word:
1560:
1561: @example
1562: word
1563: !@@#$%^&*()
1564: 1234567890
1565: 5!a
1566: @end example
1567:
1568: A frequent beginner's error is to leave away necessary white space,
1569: resulting in an error like @samp{Undefined word}; so if you see such an
1570: error, check if you have put spaces wherever necessary.
1571:
1572: @example
1573: ." hello, world" \ correct
1574: ."hello, world" \ gives an "Undefined word" error
1575: @end example
1576:
1577: Gforth and most other Forth systems ignores differences in case (it is
1578: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1579: your system is case-sensitive, you may have to type all the examples
1580: given here in upper case.
1581:
1582:
1583: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1584: @section Crash Course
1585:
1586: Type
1587:
1588: @example
1589: 0 0 !
1590: here execute
1591: ' catch >body 20 erase abort
1592: ' (quit) >body 20 erase
1593: @end example
1594:
1595: The last two examples are guaranteed to destroy parts of Gforth (and
1596: most other systems), so you better leave Gforth afterwards (if it has
1597: not finished by itself). On some systems you may have to kill gforth
1598: from outside (e.g., in Unix with @code{kill}).
1599:
1600: Now that you know how to produce crashes (and that there's not much to
1601: them), let's learn how to produce meaningful programs.
1602:
1603:
1604: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1605: @section Stack
1606:
1607: The most obvious feature of Forth is the stack. When you type in a
1608: number, it is pushed on the stack. You can display the content of the
1609: stack with @code{.s}.
1610:
1611: @example
1612: 1 2 .s
1613: 3 .s
1614: @end example
1615:
1616: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1617: appear in @code{.s} output as they appeared in the input.
1618:
1619: You can print the top of stack element with @code{.}.
1620:
1621: @example
1622: 1 2 3 . . .
1623: @end example
1624:
1625: In general, words consume their stack arguments (@code{.s} is an
1626: exception).
1627:
1628: @assignment
1629: What does the stack contain after @code{5 6 7 .}?
1630: @endassignment
1631:
1632:
1633: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1634: @section Arithmetics
1635:
1636: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1637: operate on the top two stack items:
1638:
1639: @example
1640: 2 2 + .
1641: 2 1 - .
1642: 7 3 mod .
1643: @end example
1644:
1645: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1646: as in the corresponding infix expression (this is generally the case in
1647: Forth).
1648:
1649: Parentheses are superfluous (and not available), because the order of
1650: the words unambiguously determines the order of evaluation and the
1651: operands:
1652:
1653: @example
1654: 3 4 + 5 * .
1655: 3 4 5 * + .
1656: @end example
1657:
1658: @assignment
1659: What are the infix expressions corresponding to the Forth code above?
1660: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1661: known as Postfix or RPN (Reverse Polish Notation).}.
1662: @endassignment
1663:
1664: To change the sign, use @code{negate}:
1665:
1666: @example
1667: 2 negate .
1668: @end example
1669:
1670: @assignment
1671: Convert -(-3)*4-5 to Forth.
1672: @endassignment
1673:
1674: @code{/mod} performs both @code{/} and @code{mod}.
1675:
1676: @example
1677: 7 3 /mod . .
1678: @end example
1679:
1680: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1681: @section Stack Manipulation
1682:
1683: Stack manipulation words rearrange the data on the stack.
1684:
1685: @example
1686: 1 .s drop .s
1687: 1 .s dup .s drop drop .s
1688: 1 2 .s over .s drop drop drop
1689: 1 2 .s swap .s drop drop
1690: 1 2 3 .s rot .s drop drop drop
1691: @end example
1692:
1693: These are the most important stack manipulation words. There are also
1694: variants that manipulate twice as many stack items:
1695:
1696: @example
1697: 1 2 3 4 .s 2swap .s 2drop 2drop
1698: @end example
1699:
1700: Two more stack manipulation words are:
1701:
1702: @example
1703: 1 2 .s nip .s drop
1704: 1 2 .s tuck .s 2drop drop
1705: @end example
1706:
1707: @assignment
1708: Replace @code{nip} and @code{tuck} with combinations of other stack
1709: manipulation words.
1710:
1711: @example
1712: Given: How do you get:
1713: 1 2 3 3 2 1
1714: 1 2 3 1 2 3 2
1715: 1 2 3 1 2 3 3
1716: 1 2 3 1 3 3
1717: 1 2 3 2 1 3
1718: 1 2 3 4 4 3 2 1
1719: 1 2 3 1 2 3 1 2 3
1720: 1 2 3 4 1 2 3 4 1 2
1721: 1 2 3
1722: 1 2 3 1 2 3 4
1723: 1 2 3 1 3
1724: @end example
1725: @endassignment
1726:
1727: @example
1728: 5 dup * .
1729: @end example
1730:
1731: @assignment
1732: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1733: Write a piece of Forth code that expects two numbers on the stack
1734: (@var{a} and @var{b}, with @var{b} on top) and computes
1735: @code{(a-b)(a+1)}.
1736: @endassignment
1737:
1738: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1739: @section Using files for Forth code
1740:
1741: While working at the Forth command line is convenient for one-line
1742: examples and short one-off code, you probably want to store your source
1743: code in files for convenient editing and persistence. You can use your
1744: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1745: Gforth}) to create @var{file} and use
1746:
1747: @example
1748: s" @var{file}" included
1749: @end example
1750:
1751: to load it into your Forth system. The file name extension I use for
1752: Forth files is @samp{.fs}.
1753:
1754: You can easily start Gforth with some files loaded like this:
1755:
1756: @example
1757: gforth @var{file1} @var{file2}
1758: @end example
1759:
1760: If an error occurs during loading these files, Gforth terminates,
1761: whereas an error during @code{INCLUDED} within Gforth usually gives you
1762: a Gforth command line. Starting the Forth system every time gives you a
1763: clean start every time, without interference from the results of earlier
1764: tries.
1765:
1766: I often put all the tests in a file, then load the code and run the
1767: tests with
1768:
1769: @example
1770: gforth @var{code} @var{tests} -e bye
1771: @end example
1772:
1773: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1774: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1775: restart this command without ado.
1776:
1777: The advantage of this approach is that the tests can be repeated easily
1778: every time the program ist changed, making it easy to catch bugs
1779: introduced by the change.
1780:
1781:
1782: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1783: @section Comments
1784:
1785: @example
1786: \ That's a comment; it ends at the end of the line
1787: ( Another comment; it ends here: ) .s
1788: @end example
1789:
1790: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1791: separated with white space from the following text.
1792:
1793: @example
1794: \This gives an "Undefined word" error
1795: @end example
1796:
1797: The first @code{)} ends a comment started with @code{(}, so you cannot
1798: nest @code{(}-comments; and you cannot comment out text containing a
1799: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1800: avoid @code{)} in word names.}.
1801:
1802: I use @code{\}-comments for descriptive text and for commenting out code
1803: of one or more line; I use @code{(}-comments for describing the stack
1804: effect, the stack contents, or for commenting out sub-line pieces of
1805: code.
1806:
1807: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1808: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1809: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1810: with @kbd{M-q}.
1811:
1812:
1813: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1814: @section Colon Definitions
1815:
1816: are similar to procedures and functions in other programming languages.
1817:
1818: @example
1819: : squared ( n -- n^2 )
1820: dup * ;
1821: 5 squared .
1822: 7 squared .
1823: @end example
1824:
1825: @code{:} starts the colon definition; its name is @code{squared}. The
1826: following comment describes its stack effect. The words @code{dup *}
1827: are not executed, but compiled into the definition. @code{;} ends the
1828: colon definition.
1829:
1830: The newly-defined word can be used like any other word, including using
1831: it in other definitions:
1832:
1833: @example
1834: : cubed ( n -- n^3 )
1835: dup squared * ;
1836: -5 cubed .
1837: : fourth-power ( n -- n^4 )
1838: squared squared ;
1839: 3 fourth-power .
1840: @end example
1841:
1842: @assignment
1843: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1844: @code{/mod} in terms of other Forth words, and check if they work (hint:
1845: test your tests on the originals first). Don't let the
1846: @samp{redefined}-Messages spook you, they are just warnings.
1847: @endassignment
1848:
1849:
1850: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1851: @section Decompilation
1852:
1853: You can decompile colon definitions with @code{see}:
1854:
1855: @example
1856: see squared
1857: see cubed
1858: @end example
1859:
1860: In Gforth @code{see} shows you a reconstruction of the source code from
1861: the executable code. Informations that were present in the source, but
1862: not in the executable code, are lost (e.g., comments).
1863:
1864: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1865: @section Stack-Effect Comments
1866:
1867: By convention the comment after the name of a definition describes the
1868: stack effect: The part in from of the @samp{--} describes the state of
1869: the stack before the execution of the definition, i.e., the parameters
1870: that are passed into the colon definition; the part behind the @samp{--}
1871: is the state of the stack after the execution of the definition, i.e.,
1872: the results of the definition. The stack comment only shows the top
1873: stack items that the definition accesses and/or changes.
1874:
1875: You should put a correct stack effect on every definition, even if it is
1876: just @code{( -- )}. You should also add some descriptive comment to
1877: more complicated words (I usually do this in the lines following
1878: @code{:}). If you don't do this, your code becomes unreadable (because
1879: you have to work through every definition before you can undertsand
1880: any).
1881:
1882: @assignment
1883: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1884: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1885: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1886: are done, you can compare your stack effects to this in this manual
1887: (@pxref{Word Index}).
1888: @endassignment
1889:
1890: Sometimes programmers put comments at various places in colon
1891: definitions that describe the contents of the stack at that place (stack
1892: comments); i.e., they are like the first part of a stack-effect
1893: comment. E.g.,
1894:
1895: @example
1896: : cubed ( n -- n^3 )
1897: dup squared ( n n^2 ) * ;
1898: @end example
1899:
1900: In this case the stack comment is pretty superfluous, because the word
1901: is simple enough. If you think it would be a good idea to add such a
1902: comment to increase readability, you should also consider factoring the
1903: word into several simpler words (@pxref{Factoring Tutorial,,
1.60 ! anton 1904: Factoring}), which typically eliminates the need for the stack comment;
1.48 anton 1905: however, if you decide not to refactor it, then having such a comment is
1906: better than not having it.
1907:
1908: The names of the stack items in stack-effect and stack comments in the
1909: standard, in this manual, and in many programs specify the type through
1910: a type prefix, similar to Fortran and Hungarian notation. The most
1911: frequent prefixes are:
1912:
1913: @table @code
1914: @item n
1915: signed integer
1916: @item u
1917: unsigned integer
1918: @item c
1919: character
1920: @item f
1921: Boolean flags, i.e. @code{false} or @code{true}.
1922: @item a-addr,a-
1923: Cell-aligned address
1924: @item c-addr,c-
1925: Char-aligned address (note that a Char may have two bytes in Windows NT)
1926: @item xt
1927: Execution token, same size as Cell
1928: @item w,x
1929: Cell, can contain an integer or an address. It usually takes 32, 64 or
1930: 16 bits (depending on your platform and Forth system). A cell is more
1931: commonly known as machine word, but the term @emph{word} already means
1932: something different in Forth.
1933: @item d
1934: signed double-cell integer
1935: @item ud
1936: unsigned double-cell integer
1937: @item r
1938: Float (on the FP stack)
1939: @end table
1940:
1941: You can find a more complete list in @ref{Notation}.
1942:
1943: @assignment
1944: Write stack-effect comments for all definitions you have written up to
1945: now.
1946: @endassignment
1947:
1948:
1949: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1950: @section Types
1951:
1952: In Forth the names of the operations are not overloaded; so similar
1953: operations on different types need different names; e.g., @code{+} adds
1954: integers, and you have to use @code{f+} to add floating-point numbers.
1955: The following prefixes are often used for related operations on
1956: different types:
1957:
1958: @table @code
1959: @item (none)
1960: signed integer
1961: @item u
1962: unsigned integer
1963: @item c
1964: character
1965: @item d
1966: signed double-cell integer
1967: @item ud, du
1968: unsigned double-cell integer
1969: @item 2
1970: two cells (not-necessarily double-cell numbers)
1971: @item m, um
1972: mixed single-cell and double-cell operations
1973: @item f
1974: floating-point (note that in stack comments @samp{f} represents flags,
1975: and @samp{r} represents FP number).
1976: @end table
1977:
1978: If there are no differences between the signed and the unsigned variant
1979: (e.g., for @code{+}), there is only the prefix-less variant.
1980:
1981: Forth does not perform type checking, neither at compile time, nor at
1982: run time. If you use the wrong oeration, the data are interpreted
1983: incorrectly:
1984:
1985: @example
1986: -1 u.
1987: @end example
1988:
1989: If you have only experience with type-checked languages until now, and
1990: have heard how important type-checking is, don't panic! In my
1991: experience (and that of other Forthers), type errors in Forth code are
1992: usually easy to find (once you get used to it), the increased vigilance
1993: of the programmer tends to catch some harder errors in addition to most
1994: type errors, and you never have to work around the type system, so in
1995: most situations the lack of type-checking seems to be a win (projects to
1996: add type checking to Forth have not caught on).
1997:
1998:
1999: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
2000: @section Factoring
2001:
2002: If you try to write longer definitions, you will soon find it hard to
2003: keep track of the stack contents. Therefore, good Forth programmers
2004: tend to write only short definitions (e.g., three lines). The art of
2005: finding meaningful short definitions is known as factoring (as in
2006: factoring polynomials).
2007:
2008: Well-factored programs offer additional advantages: smaller, more
2009: general words, are easier to test and debug and can be reused more and
2010: better than larger, specialized words.
2011:
2012: So, if you run into difficulties with stack management, when writing
2013: code, try to define meaningful factors for the word, and define the word
2014: in terms of those. Even if a factor contains only two words, it is
2015: often helpful.
2016:
2017: Good factoring is not easy, and even experienced Forth programmers often
2018: don't find the right solution right away, but only when rewriting the
2019: program. So, if you don't come up with a good solution immediately,
2020: keep trying, don't despair.
2021:
2022: @c example !!
2023:
2024:
2025: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
2026: @section Designing the stack effect
2027:
2028: In other languages you can use an arbitrary order of parameters for a
2029: function; and since ther is only one result, you don't have to deal with
2030: the order of results, either.
2031:
2032: In Forth (and other stack-based languages, e.g., Postscript) the
2033: parameter and result order of a definition is important and should be
2034: designed well. The general guideline is to design the stack effect such
2035: that the word is simple to use in most cases, even if that complicates
2036: the implementation of the word. Some concrete rules are:
2037:
2038: @itemize @bullet
2039:
2040: @item
2041: Words consume all of their parameters (e.g., @code{.}).
2042:
2043: @item
2044: If there is a convention on the order of parameters (e.g., from
2045: mathematics or another programming language), stick with it (e.g.,
2046: @code{-}).
2047:
2048: @item
2049: If one parameter usually requires only a short computation (e.g., it is
2050: a constant), pass it on the top of the stack. Conversely, parameters
2051: that usually require a long sequence of code to compute should be passed
2052: as the bottom (i.e., first) parameter. This makes the code easier to
2053: read, because reader does not need to keep track of the bottom item
2054: through a long sequence of code (or, alternatively, through stack
1.49 anton 2055: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48 anton 2056: address on top of the stack because it is usually simpler to compute
2057: than the stored value (often the address is just a variable).
2058:
2059: @item
2060: Similarly, results that are usually consumed quickly should be returned
2061: on the top of stack, whereas a result that is often used in long
2062: computations should be passed as bottom result. E.g., the file words
2063: like @code{open-file} return the error code on the top of stack, because
2064: it is usually consumed quickly by @code{throw}; moreover, the error code
2065: has to be checked before doing anything with the other results.
2066:
2067: @end itemize
2068:
2069: These rules are just general guidelines, don't lose sight of the overall
2070: goal to make the words easy to use. E.g., if the convention rule
2071: conflicts with the computation-length rule, you might decide in favour
2072: of the convention if the word will be used rarely, and in favour of the
2073: computation-length rule if the word will be used frequently (because
2074: with frequent use the cost of breaking the computation-length rule would
2075: be quite high, and frequent use makes it easier to remember an
2076: unconventional order).
2077:
2078: @c example !! structure package
2079:
2080: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2081: @section Local Variables
2082:
2083: You can define local variables (@emph{locals}) in a colon definition:
2084:
2085: @example
2086: : swap @{ a b -- b a @}
2087: b a ;
2088: 1 2 swap .s 2drop
2089: @end example
2090:
2091: (If your Forth system does not support this syntax, include
2092: @file{compat/anslocals.fs} first).
2093:
2094: In this example @code{@{ a b -- b a @}} is the locals definition; it
2095: takes two cells from the stack, puts the top of stack in @code{b} and
2096: the next stack element in @code{a}. @code{--} starts a comment ending
2097: with @code{@}}. After the locals definition, using the name of the
2098: local will push its value on the stack. You can leave the comment
2099: part (@code{-- b a}) away:
2100:
2101: @example
2102: : swap ( x1 x2 -- x2 x1 )
2103: @{ a b @} b a ;
2104: @end example
2105:
2106: In Gforth you can have several locals definitions, anywhere in a colon
2107: definition; in contrast, in a standard program you can have only one
2108: locals definition per colon definition, and that locals definition must
2109: be outside any controll structure.
2110:
2111: With locals you can write slightly longer definitions without running
2112: into stack trouble. However, I recommend trying to write colon
2113: definitions without locals for exercise purposes to help you gain the
2114: essential factoring skills.
2115:
2116: @assignment
2117: Rewrite your definitions until now with locals
2118: @endassignment
2119:
2120:
2121: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2122: @section Conditional execution
2123:
2124: In Forth you can use control structures only inside colon definitions.
2125: An @code{if}-structure looks like this:
2126:
2127: @example
2128: : abs ( n1 -- +n2 )
2129: dup 0 < if
2130: negate
2131: endif ;
2132: 5 abs .
2133: -5 abs .
2134: @end example
2135:
2136: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2137: the following code is performed, otherwise execution continues after the
1.51 pazsan 2138: @code{endif} (or @code{else}). @code{<} compares the top two stack
1.48 anton 2139: elements and prioduces a flag:
2140:
2141: @example
2142: 1 2 < .
2143: 2 1 < .
2144: 1 1 < .
2145: @end example
2146:
2147: Actually the standard name for @code{endif} is @code{then}. This
2148: tutorial presents the examples using @code{endif}, because this is often
2149: less confusing for people familiar with other programming languages
2150: where @code{then} has a different meaning. If your system does not have
2151: @code{endif}, define it with
2152:
2153: @example
2154: : endif postpone then ; immediate
2155: @end example
2156:
2157: You can optionally use an @code{else}-part:
2158:
2159: @example
2160: : min ( n1 n2 -- n )
2161: 2dup < if
2162: drop
2163: else
2164: nip
2165: endif ;
2166: 2 3 min .
2167: 3 2 min .
2168: @end example
2169:
2170: @assignment
2171: Write @code{min} without @code{else}-part (hint: what's the definition
2172: of @code{nip}?).
2173: @endassignment
2174:
2175:
2176: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2177: @section Flags and Comparisons
2178:
2179: In a false-flag all bits are clear (0 when interpreted as integer). In
2180: a canonical true-flag all bits are set (-1 as a twos-complement signed
2181: integer); in many contexts (e.g., @code{if}) any non-zero value is
2182: treated as true flag.
2183:
2184: @example
2185: false .
2186: true .
2187: true hex u. decimal
2188: @end example
2189:
2190: Comparison words produce canonical flags:
2191:
2192: @example
2193: 1 1 = .
2194: 1 0= .
2195: 0 1 < .
2196: 0 0 < .
2197: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2198: -1 1 < .
2199: @end example
2200:
2201: Gforth supports all combinations of the prefixes @code{0 u d d0 du} (or
2202: none) and the comparisons @code{= <> < > <= >=}. Only a part of these
1.60 ! anton 2203: combinations are standard (for details see the standard or @ref{Word
! 2204: Glossary}).
1.48 anton 2205:
2206: You can use @code{and or xor invert} can be used as operations on
2207: canonical flags. Actually they are bitwise operations:
2208:
2209: @example
2210: 1 2 and .
2211: 1 2 or .
2212: 1 3 xor .
2213: 1 invert .
2214: @end example
2215:
2216: You can convert a zero/non-zero flag into a canonical flag with
2217: @code{0<>} (and complement it on the way with @code{0=}).
2218:
2219: @example
2220: 1 0= .
2221: 1 0<> .
2222: @end example
2223:
2224: You can use the all-bits-set feature of canonicasl flags and the bitwise
2225: operation of the Boolean operations to avoid @code{if}s:
2226:
2227: @example
2228: : foo ( n1 -- n2 )
2229: 0= if
2230: 14
2231: else
2232: 0
2233: endif ;
2234: 0 foo .
2235: 1 foo .
2236:
2237: : foo ( n1 -- n2 )
2238: 0= 14 and ;
2239: 0 foo .
2240: 1 foo .
2241: @end example
2242:
2243: @assignment
2244: Write @code{min} without @code{if}.
2245: @endassignment
2246:
2247:
2248: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2249: @section General Loops
2250:
2251: The endless loop is the most simple one:
2252:
2253: @example
2254: : endless ( -- )
2255: 0 begin
2256: dup . 1+
2257: again ;
2258: endless
2259: @end example
2260:
2261: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2262: does nothing at run-time, @code{again} jumps back to @code{begin}.
2263:
2264: A loop with one exit at any place looks like this:
2265:
2266: @example
2267: : log2 ( +n1 -- n2 )
2268: \ logarithmus dualis of n1>0, rounded down to the next integer
2269: assert( dup 0> )
2270: 2/ 0 begin
2271: over 0> while
2272: 1+ swap 2/ swap
2273: repeat
2274: nip ;
2275: 7 log2 .
2276: 8 log2 .
2277: @end example
2278:
2279: At run-time @code{while} consumes a flag; if it is 0, execution
1.51 pazsan 2280: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48 anton 2281: continues behind the @code{while}. @code{Repeat} jumps back to
2282: @code{begin}, just like @code{again}.
2283:
2284: In Forth there are many combinations/abbreviations, like @code{1+}.
2285: However, @code{2/} is not one of them; it shifts it's argument right by
2286: one bit (arithmetic shift right):
2287:
2288: @example
2289: -5 2 / .
2290: -5 2/ .
2291: @end example
2292:
2293: @code{assert(} is no standard word, but you can get it on systems other
2294: then Gforth by including @file{compat/assert.fs}. You can see what it
2295: does by trying
2296:
2297: @example
2298: 0 log2 .
2299: @end example
2300:
2301: Here's a loop with an exit at the end:
2302:
2303: @example
2304: : log2 ( +n1 -- n2 )
2305: \ logarithmus dualis of n1>0, rounded down to the next integer
2306: assert( dup 0 > )
2307: -1 begin
2308: 1+ swap 2/ swap
2309: over 0 <=
2310: until
2311: nip ;
2312: @end example
2313:
2314: @code{Until} consumes a flag; if it is non-zero, execution continues at
2315: the @code{begin}, otherwise after the @code{until}.
2316:
2317: @assignment
2318: Write a definition for computing the greatest common divisor.
2319: @endassignment
2320:
2321:
2322: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2323: @section Counted loops
2324:
2325: @example
2326: : ^ ( n1 u -- n )
2327: \ n = the uth power of u1
2328: 1 swap 0 u+do
2329: over *
2330: loop
2331: nip ;
2332: 3 2 ^ .
2333: 4 3 ^ .
2334: @end example
2335:
2336: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2337: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2338: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2339: times (or not at all, if @code{u3-u4<0}).
2340:
2341: You can see the stack effect design rules at work in the stack effect of
2342: the loop start words: Since the start value of the loop is more
2343: frequently constant than the end value, the start value is passed on
2344: the top-of-stack.
2345:
2346: You can access the counter of a counted loop with @code{i}:
2347:
2348: @example
2349: : fac ( u -- u! )
2350: 1 swap 1+ 1 u+do
2351: i *
2352: loop ;
2353: 5 fac .
2354: 7 fac .
2355: @end example
2356:
2357: There is also @code{+do}, which expects signed numbers (important for
2358: deciding whether to enter the loop).
2359:
2360: @assignment
2361: Write a definition for computing the nth Fibonacci number.
2362: @endassignment
2363:
2364: !! +DO...+LOOP
2365: !! -DO...-LOOP
2366:
2367:
2368: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2369: @section Recursion
2370:
2371: Usually the name of a definition is not visible in the definition; but
2372: earlier definitions are usually visible:
2373:
2374: @example
2375: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2376: : / ( n1 n2 -- n )
2377: dup 0= if
2378: -10 throw \ report division by zero
2379: endif
2380: / \ old version
2381: ;
2382: 1 0 /
2383: @end example
2384:
2385: For recursive definitions you can use @code{recursive} (non-standard) or
2386: @code{recurse}:
2387:
2388: @example
2389: : fac1 ( n -- n! ) recursive
2390: dup 0> if
2391: dup 1- fac1 *
2392: else
2393: drop 1
2394: endif ;
2395: 7 fac1 .
2396:
2397: : fac2 ( n -- n! )
2398: dup 0> if
2399: dup 1- recurse *
2400: else
2401: drop 1
2402: endif ;
2403: 8 fac2 .
2404: @end example
2405:
2406: @assignment
2407: Write a recursive definition for computing the nth Fibonacci number.
2408: @endassignment
2409:
2410:
2411: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2412: @section Leaving definitions or loops
2413:
2414: @code{EXIT} exits the current definition right away. For every counted
2415: loop that is left in this way, an @code{UNLOOP} has to be performed
2416: before the @code{EXIT}:
2417:
2418: @c !! real examples
2419: @example
2420: : ...
2421: ... u+do
2422: ... if
2423: ... unloop exit
2424: endif
2425: ...
2426: loop
2427: ... ;
2428: @end example
2429:
2430: @code{LEAVE} leaves the innermost counted loop right away:
2431:
2432: @example
2433: : ...
2434: ... u+do
2435: ... if
2436: ... leave
2437: endif
2438: ...
2439: loop
2440: ... ;
2441: @end example
2442:
2443:
2444: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2445: @section Return Stack
2446:
2447: In addition to the data stack Forth also has a second stack, the return
2448: stack; most Forth systems store the return addresses of procedure calls
2449: there (thus its name). Programmers can also use this stack:
2450:
2451: @example
2452: : foo ( n1 n2 -- )
2453: .s
2454: >r .s
1.50 anton 2455: r@@ .
1.48 anton 2456: >r .s
1.50 anton 2457: r@@ .
1.48 anton 2458: r> .
1.50 anton 2459: r@@ .
1.48 anton 2460: r> . ;
2461: 1 2 foo
2462: @end example
2463:
2464: @code{>r} takes an element from the data stack and pushes it onto the
2465: return stack; conversely, @code{r>} moves an elementm from the return to
2466: the data stack; @code{r@@} pushes a copy of the top of the return stack
2467: on the return stack.
2468:
2469: Forth programmers usually use the return stack for storing data
2470: temporarily, if using the data stack alone would be too complex, and
2471: factoring and locals are not an option:
2472:
2473: @example
2474: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2475: rot >r rot r> ;
2476: @end example
2477:
2478: The return address of the definition and the loop control parameters of
2479: counted loops usually reside on the return stack, so you have to take
2480: all items, that you have pushed on the return stack in a colon
2481: definition or counted loop, from the return stack before the definition
2482: or loop ends. You cannot access items that you pushed on the return
2483: stack outside some definition or loop within the definition of loop.
2484:
2485: If you miscount the return stack items, this usually ends in a crash:
2486:
2487: @example
2488: : crash ( n -- )
2489: >r ;
2490: 5 crash
2491: @end example
2492:
2493: You cannot mix using locals and using the return stack (according to the
2494: standard; Gforth has no problem). However, they solve the same
2495: problems, so this shouldn't be an issue.
2496:
2497: @assignment
2498: Can you rewrite any of the definitions you wrote until now in a better
2499: way using the return stack?
2500: @endassignment
2501:
2502:
2503: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2504: @section Memory
2505:
2506: You can create a global variable @code{v} with
2507:
2508: @example
2509: variable v ( -- addr )
2510: @end example
2511:
2512: @code{v} pushes the address of a cell in memory on the stack. This cell
2513: was reserved by @code{variable}. You can use @code{!} (store) to store
2514: values into this cell and @code{@@} (fetch) to load the value from the
2515: stack into memory:
2516:
2517: @example
2518: v .
2519: 5 v ! .s
1.50 anton 2520: v @@ .
1.48 anton 2521: @end example
2522:
2523: You can also reserve more memory:
2524:
2525: @example
2526: create v2 20 cells allot
2527: @end example
2528:
2529: creates a word @code{v2} and reserves 20 cells; the address pushed by
2530: @code{v2} points to the start of these 20 cells. You can use address
2531: arithmetic to access these cells:
2532:
2533: @example
2534: 3 v2 5 cells + !
2535: @end example
2536:
2537: You can reserve and initialize memory with @code{,}:
2538:
2539: @example
2540: create v3
2541: 5 , 4 , 3 , 2 , 1 ,
1.50 anton 2542: v3 @@ .
2543: v3 cell+ @@ .
2544: v3 2 cells + @@ .
1.48 anton 2545: @end example
2546:
2547: @assignment
2548: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2549: @code{u} cells, with the first of these cells at @code{addr}, the next
2550: one at @code{addr cell+} etc.
2551: @endassignment
2552:
2553: You can also reserve memory without creating a new word:
2554:
2555: @example
1.60 ! anton 2556: here 10 cells allot .
! 2557: here .
1.48 anton 2558: @end example
2559:
2560: @code{Here} pushes the start address of the memory area. You should
2561: store it somewhere, or you will have a hard time finding the memory area
2562: again.
2563:
2564: @code{Allot} manages dictionary memory. The dictionary memory contains
2565: the system's data structures for words etc. on Gforth and most other
2566: Forth systems. It is managed like a stack: You can free the memory that
2567: you have just @code{allot}ed with
2568:
2569: @example
2570: -10 cells allot
1.60 ! anton 2571: here .
1.48 anton 2572: @end example
2573:
2574: Note that you cannot do this if you have created a new word in the
2575: meantime (because then your @code{allot}ed memory is no longer on the
2576: top of the dictionary ``stack'').
2577:
2578: Alternatively, you can use @code{allocate} and @code{free} which allow
2579: freeing memory in any order:
2580:
2581: @example
2582: 10 cells allocate throw .s
2583: 20 cells allocate throw .s
2584: swap
2585: free throw
2586: free throw
2587: @end example
2588:
2589: The @code{throw}s deal with errors (e.g., out of memory).
2590:
1.60 ! anton 2591: And there is also a garbage collector
! 2592: @url{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip},
! 2593: which eliminates the need to @code{free} memory explicitly.
1.48 anton 2594:
2595:
2596: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2597: @section Characters and Strings
2598:
2599: On the stack characters take up a cell, like numbers. In memory they
2600: have their own size (one 8-bit byte on most systems), and therefore
2601: require their own words for memory access:
2602:
2603: @example
2604: create v4
2605: 104 c, 97 c, 108 c, 108 c, 111 c,
1.50 anton 2606: v4 4 chars + c@@ .
1.48 anton 2607: @end example
2608:
2609: The preferred representation of strings on the stack is @code{addr
2610: u-count}, where @code{addr} is the address of the first character and
2611: @code{u-count} is the number of characters in the string.
2612:
2613: @example
2614: v4 5 type
2615: @end example
2616:
2617: You get a string constant with
2618:
2619: @example
2620: s" hello, world" .s
2621: type
2622: @end example
2623:
2624: Make sure you have a space between @code{s"} and the string; @code{s"}
2625: is a normal Forth word and must be delimited with white space (try what
2626: happens when you remove the space).
2627:
2628: However, this interpretive use of @code{s"} is quite restricted: the
2629: string exists only until the next call of @code{s"} (some Forth systems
2630: keep more than one of these strings, but usually they still have a
2631: limited lifetime.
2632:
2633: @example
2634: s" hello," s" world" .s
2635: type
2636: type
2637: @end example
2638:
2639: However, you can also use @code{s"} in a definition, and the resulting
2640: strings then live forever (well, as long as the definition):
2641:
2642: @example
2643: : foo s" hello," s" world" ;
2644: foo .s
2645: type
2646: type
2647: @end example
2648:
2649: @assignment
2650: @code{Emit ( c -- )} types @code{c} as character (not a number).
2651: Implement @code{type ( addr u -- )}.
2652: @endassignment
2653:
2654: @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2655: @section Alignment
2656:
2657: On many processors cells have to be aligned in memory, if you want to
2658: access them with @code{@@} and @code{!} (and even if the processor does
2659: not require alignment, access to aligned cells are faster).
2660:
2661: @code{Create} aligns @code{here} (i.e., the place where the next
2662: allocation will occur, and that the @code{create}d word points to).
2663: Likewise, the memory produced by @code{allocate} starts at an aligned
2664: address. Adding a number of @code{cells} to an aligned address produces
2665: another aligned address.
2666:
2667: However, address arithmetic involving @code{char+} and @code{chars} can
2668: create an address that is not cell-aligned. @code{Aligned ( addr --
2669: a-addr )} produces the next aligned address:
2670:
2671: @example
1.50 anton 2672: v3 char+ aligned .s @@ .
2673: v3 char+ .s @@ .
1.48 anton 2674: @end example
2675:
2676: Similarly, @code{align} advances @code{here} to the next aligned
2677: address:
2678:
2679: @example
2680: create v5 97 c,
2681: here .
2682: align here .
2683: 1000 ,
2684: @end example
2685:
2686: Note that you should use aligned addresses even if your processor does
2687: not require them, if you want your program to be portable.
2688:
2689:
2690: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2691: @section Interpretation and Compilation Semantics and Immediacy
2692:
2693: When a word is compiled, it behaves differently from being interpreted.
2694: E.g., consider @code{+}:
2695:
2696: @example
2697: 1 2 + .
2698: : foo + ;
2699: @end example
2700:
2701: These two behaviours are known as compilation and interpretation
2702: semantics. For normal words (e.g., @code{+}), the compilation semantics
2703: is to append the interpretation semantics to the currently defined word
2704: (@code{foo} in the example above). I.e., when @code{foo} is executed
2705: later, the interpretation semantics of @code{+} (i.e., adding two
2706: numbers) will be performed.
2707:
2708: However, there are words with non-default compilation semantics, e.g.,
2709: the control-flow words like @code{if}. You can use @code{immediate} to
2710: change the compilation semantics of the last defined word to be equal to
2711: the interpretation semantics:
2712:
2713: @example
2714: : [FOO] ( -- )
2715: 5 . ; immediate
2716:
2717: [FOO]
2718: : bar ( -- )
2719: [FOO] ;
2720: bar
2721: see bar
2722: @end example
2723:
2724: Two conventions to mark words with non-default compilation semnatics are
2725: names with brackets (more frequently used) and to write them all in
2726: upper case (less frequently used).
2727:
2728: In Gforth (and many other systems) you can also remove the
2729: interpretation semantics with @code{compile-only} (the compilation
2730: semantics is derived from the original interpretation semantics):
2731:
2732: @example
2733: : flip ( -- )
2734: 6 . ; compile-only \ but not immediate
2735: flip
2736:
2737: : flop ( -- )
2738: flip ;
2739: flop
2740: @end example
2741:
2742: In this example the interpretation semantics of @code{flop} is equal to
2743: the original interpretation semantics of @code{flip}.
2744:
2745: The text interpreter has two states: in interpret state, it performs the
2746: interpretation semantics of words it encounters; in compile state, it
2747: performs the compilation semantics of these words.
2748:
2749: Among other things, @code{:} switches into compile state, and @code{;}
2750: switches back to interpret state. They contain the factors @code{]}
2751: (switch to compile state) and @code{[} (switch to interpret state), that
2752: do nothing but switch the state.
2753:
2754: @example
2755: : xxx ( -- )
2756: [ 5 . ]
2757: ;
2758:
2759: xxx
2760: see xxx
2761: @end example
2762:
2763: These brackets are also the source of the naming convention mentioned
2764: above.
2765:
2766:
2767: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2768: @section Execution Tokens
2769:
2770: @code{' word} gives you the execution token (XT) of a word. The XT is a
2771: cell representing the interpretation semantics of a word. You can
2772: execute this semantics with @code{execute}:
2773:
2774: @example
2775: ' + .s
2776: 1 2 rot execute .
2777: @end example
2778:
2779: The XT is similar to a function pointer in C. However, parameter
2780: passing through the stack makes it a little more flexible:
2781:
2782: @example
2783: : map-array ( ... addr u xt -- ... )
1.50 anton 2784: \ executes xt ( ... x -- ... ) for every element of the array starting
2785: \ at addr and containing u elements
1.48 anton 2786: @{ xt @}
2787: cells over + swap ?do
1.50 anton 2788: i @@ xt execute
1.48 anton 2789: 1 cells +loop ;
2790:
2791: create a 3 , 4 , 2 , -1 , 4 ,
2792: a 5 ' . map-array .s
2793: 0 a 5 ' + map-array .
2794: s" max-n" environment? drop .s
2795: a 5 ' min map-array .
2796: @end example
2797:
2798: You can use map-array with the XTs of words that consume one element
2799: more than they produce. In theory you can also use it with other XTs,
2800: but the stack effect then depends on the size of the array, which is
2801: hard to understand.
2802:
1.51 pazsan 2803: Since XTs are cell-sized, you can store them in memory and manipulate
2804: them on the stack like other cells. You can also compile the XT into a
1.48 anton 2805: word with @code{compile,}:
2806:
2807: @example
2808: : foo1 ( n1 n2 -- n )
2809: [ ' + compile, ] ;
2810: see foo
2811: @end example
2812:
2813: This is non-standard, because @code{compile,} has no compilation
2814: semantics in the standard, but it works in good Forth systems. For the
2815: broken ones, use
2816:
2817: @example
2818: : [compile,] compile, ; immediate
2819:
2820: : foo1 ( n1 n2 -- n )
2821: [ ' + ] [compile,] ;
2822: see foo
2823: @end example
2824:
2825: @code{'} is a word with default compilation semantics; it parses the
2826: next word when its interpretation semantics are executed, not during
2827: compilation:
2828:
2829: @example
2830: : foo ( -- xt )
2831: ' ;
2832: see foo
2833: : bar ( ... "word" -- ... )
2834: ' execute ;
2835: see bar
1.60 ! anton 2836: 1 2 bar + .
1.48 anton 2837: @end example
2838:
2839: You often want to parse a word during compilation and compile its XT so
2840: it will be pushed on the stack at run-time. @code{[']} does this:
2841:
2842: @example
2843: : xt-+ ( -- xt )
2844: ['] + ;
2845: see xt-+
2846: 1 2 xt-+ execute .
2847: @end example
2848:
2849: Many programmers tend to see @code{'} and the word it parses as one
2850: unit, and expect it to behave like @code{[']} when compiled, and are
2851: confused by the actual behaviour. If you are, just remember that the
2852: Forth system just takes @code{'} as one unit and has no idea that it is
2853: a parsing word (attempts to convenience programmers in this issue have
2854: usually resulted in even worse pitfalls, see
2855: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}).
2856:
2857: Note that the state of the interpreter does not come into play when
1.51 pazsan 2858: creating and executing XTs. I.e., even when you execute @code{'} in
1.48 anton 2859: compile state, it still gives you the interpretation semantics. And
2860: whatever that state is, @code{execute} performs the semantics
1.51 pazsan 2861: represented by the XT (i.e., the interpretation semantics).
1.48 anton 2862:
2863:
2864: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2865: @section Exceptions
2866:
2867: @code{throw ( n -- )} causes an exception unless n is zero.
2868:
2869: @example
2870: 100 throw .s
2871: 0 throw .s
2872: @end example
2873:
2874: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2875: it catches exceptions and pushes the number of the exception on the
2876: stack (or 0, if the xt executed without exception). If there was an
2877: exception, the stacks have the same depth as when entering @code{catch}:
2878:
2879: @example
2880: .s
2881: 3 0 ' / catch .s
2882: 3 2 ' / catch .s
2883: @end example
2884:
2885: @assignment
2886: Try the same with @code{execute} instead of @code{catch}.
2887: @endassignment
2888:
2889: @code{Throw} always jumps to the dynamically next enclosing
2890: @code{catch}, even if it has to leave several call levels to achieve
2891: this:
2892:
2893: @example
2894: : foo 100 throw ;
2895: : foo1 foo ." after foo" ;
1.51 pazsan 2896: : bar ['] foo1 catch ;
1.60 ! anton 2897: bar .
1.48 anton 2898: @end example
2899:
2900: It is often important to restore a value upon leaving a definition, even
2901: if the definition is left through an exception. You can ensure this
2902: like this:
2903:
2904: @example
2905: : ...
2906: save-x
1.51 pazsan 2907: ['] word-changing-x catch ( ... n )
1.48 anton 2908: restore-x
2909: ( ... n ) throw ;
2910: @end example
2911:
1.55 anton 2912: Gforth provides an alternative syntax in addition to @code{catch}:
1.48 anton 2913: @code{try ... recover ... endtry}. If the code between @code{try} and
2914: @code{recover} has an exception, the stack depths are restored, the
2915: exception number is pushed on the stack, and the code between
2916: @code{recover} and @code{endtry} is performed. E.g., the definition for
2917: @code{catch} is
2918:
2919: @example
2920: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2921: try
2922: execute 0
2923: recover
2924: nip
2925: endtry ;
2926: @end example
2927:
2928: The equivalent to the restoration code above is
2929:
2930: @example
2931: : ...
2932: save-x
2933: try
2934: word-changing-x
2935: end-try
2936: restore-x
2937: throw ;
2938: @end example
2939:
2940: As you can see, the @code{recover} part is optional.
2941:
2942:
2943: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2944: @section Defining Words
2945:
2946: @code{:}, @code{create}, and @code{variable} are definition words: They
2947: define other words. @code{Constant} is another definition word:
2948:
2949: @example
2950: 5 constant foo
2951: foo .
2952: @end example
2953:
2954: You can also use the prefixes @code{2} (double-cell) and @code{f}
2955: (floating point) with @code{variable} and @code{constant}.
2956:
2957: You can also define your own defining words. E.g.:
2958:
2959: @example
2960: : variable ( "name" -- )
2961: create 0 , ;
2962: @end example
2963:
2964: You can also define defining words that create words that do something
2965: other than just producing their address:
2966:
2967: @example
2968: : constant ( n "name" -- )
2969: create ,
2970: does> ( -- n )
1.50 anton 2971: ( addr ) @@ ;
1.48 anton 2972:
2973: 5 constant foo
2974: foo .
2975: @end example
2976:
2977: The definition of @code{constant} above ends at the @code{does>}; i.e.,
2978: @code{does>} replaces @code{;}, but it also does something else: It
2979: changes the last defined word such that it pushes the address of the
2980: body of the word and then performs the code after the @code{does>}
2981: whenever it is called.
2982:
2983: In the example above, @code{constant} uses @code{,} to store 5 into the
2984: body of @code{foo}. When @code{foo} executes, it pushes the address of
2985: the body onto the stack, then (in the code after the @code{does>})
2986: fetches the 5 from there.
2987:
2988: The stack comment near the @code{does>} reflects the stack effect of the
2989: defined word, not the stack effect of the code after the @code{does>}
2990: (the difference is that the code expects the address of the body that
2991: the stack comment does not show).
2992:
2993: You can use these definition words to do factoring in cases that involve
2994: (other) definition words. E.g., a field offset is always added to an
2995: address. Instead of defining
2996:
2997: @example
2998: 2 cells constant offset-field1
2999: @end example
3000:
3001: and using this like
3002:
3003: @example
3004: ( addr ) offset-field1 +
3005: @end example
3006:
3007: you can define a definition word
3008:
3009: @example
3010: : simple-field ( n "name" -- )
3011: create ,
3012: does> ( n1 -- n1+n )
1.50 anton 3013: ( addr ) @@ + ;
1.48 anton 3014: @end example
1.21 crook 3015:
1.48 anton 3016: Definition and use of field offsets now look like this:
1.21 crook 3017:
1.48 anton 3018: @example
3019: 2 cells simple-field field1
1.60 ! anton 3020: create mystruct 4 cells allot
! 3021: mystruct .s field1 .s drop
1.48 anton 3022: @end example
1.21 crook 3023:
1.48 anton 3024: If you want to do something with the word without performing the code
3025: after the @code{does>}, you can access the body of a @code{create}d word
3026: with @code{>body ( xt -- addr )}:
1.21 crook 3027:
1.48 anton 3028: @example
3029: : value ( n "name" -- )
3030: create ,
3031: does> ( -- n1 )
1.50 anton 3032: @@ ;
1.48 anton 3033: : to ( n "name" -- )
3034: ' >body ! ;
1.21 crook 3035:
1.48 anton 3036: 5 value foo
3037: foo .
3038: 7 to foo
3039: foo .
3040: @end example
1.21 crook 3041:
1.48 anton 3042: @assignment
3043: Define @code{defer ( "name" -- )}, which creates a word that stores an
3044: XT (at the start the XT of @code{abort}), and upon execution
3045: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3046: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3047: recursion is one application of @code{defer}.
3048: @endassignment
1.29 crook 3049:
1.48 anton 3050: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3051: @section Arrays and Records
1.29 crook 3052:
1.48 anton 3053: Forth has no standard words for defining data structures such as arrays
3054: and records (structs in C terminology), but you can build them yourself
3055: based on address arithmetic. You can also define words for defining
3056: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29 crook 3057:
1.48 anton 3058: One of the first projects a Forth newcomer sets out upon when learning
3059: about defining words is an array defining word (possibly for
3060: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3061: learn something from it. However, don't be disappointed when you later
3062: learn that you have little use for these words (inappropriate use would
3063: be even worse). I have not yet found a set of useful array words yet;
3064: the needs are just too diverse, and named, global arrays (the result of
3065: naive use of defining words) are often not flexible enough (e.g.,
3066: consider how to pass them as parameters).
1.29 crook 3067:
1.48 anton 3068: On the other hand, there is a useful set of record words, and it has
3069: been defined in @file{compat/struct.fs}; these words are predefined in
3070: Gforth. They are explained in depth elsewhere in this manual (see
3071: @pxref{Structures}). The @code{simple-field} example above is
3072: simplified variant of fields in this package.
1.21 crook 3073:
3074:
1.48 anton 3075: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3076: @section @code{POSTPONE}
1.21 crook 3077:
1.48 anton 3078: You can compile the compilation semantics (instead of compiling the
3079: interpretation semantics) of a word with @code{POSTPONE}:
1.21 crook 3080:
1.48 anton 3081: @example
3082: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51 pazsan 3083: POSTPONE + ; immediate
1.48 anton 3084: : foo ( n1 n2 -- n )
3085: MY-+ ;
3086: 1 2 foo .
3087: see foo
3088: @end example
1.21 crook 3089:
1.48 anton 3090: During the definition of @code{foo} the text interpreter performs the
3091: compilation semantics of @code{MY-+}, which performs the compilation
3092: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3093:
3094: This example also displays separate stack comments for the compilation
3095: semantics and for the stack effect of the compiled code. For words with
3096: default compilation semantics these stack effects are usually not
3097: displayed; the stack effect of the compilation semantics is always
3098: @code{( -- )} for these words, the stack effect for the compiled code is
3099: the stack effect of the interpretation semantics.
3100:
3101: Note that the state of the interpreter does not come into play when
3102: performing the compilation semantics in this way. You can also perform
3103: it interpretively, e.g.:
3104:
3105: @example
3106: : foo2 ( n1 n2 -- n )
3107: [ MY-+ ] ;
3108: 1 2 foo .
3109: see foo
3110: @end example
1.21 crook 3111:
1.48 anton 3112: However, there are some broken Forth systems where this does not always
3113: work, and therefore this practice has been declared non-standard in
3114: 1999.
3115: @c !! repair.fs
3116:
3117: Here is another example for using @code{POSTPONE}:
1.44 crook 3118:
1.48 anton 3119: @example
3120: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3121: POSTPONE negate POSTPONE + ; immediate compile-only
3122: : bar ( n1 n2 -- n )
3123: MY-- ;
3124: 2 1 bar .
3125: see bar
3126: @end example
1.21 crook 3127:
1.48 anton 3128: You can define @code{ENDIF} in this way:
1.21 crook 3129:
1.48 anton 3130: @example
3131: : ENDIF ( Compilation: orig -- )
3132: POSTPONE then ; immediate
3133: @end example
1.21 crook 3134:
1.48 anton 3135: @assignment
3136: Write @code{MY-2DUP} that has compilation semantics equivalent to
3137: @code{2dup}, but compiles @code{over over}.
3138: @endassignment
1.29 crook 3139:
1.48 anton 3140: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3141: @section @code{Literal}
1.29 crook 3142:
1.48 anton 3143: You cannot @code{POSTPONE} numbers:
1.21 crook 3144:
1.48 anton 3145: @example
3146: : [FOO] POSTPONE 500 ; immediate
1.21 crook 3147: @end example
3148:
1.48 anton 3149: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29 crook 3150:
1.48 anton 3151: @example
3152: : [FOO] ( compilation: --; run-time: -- n )
3153: 500 POSTPONE literal ; immediate
1.29 crook 3154:
1.60 ! anton 3155: : flip [FOO] ;
1.48 anton 3156: flip .
3157: see flip
3158: @end example
1.29 crook 3159:
1.48 anton 3160: @code{LITERAL} consumes a number at compile-time (when it's compilation
3161: semantics are executed) and pushes it at run-time (when the code it
3162: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3163: number computed at compile time into the current word:
1.29 crook 3164:
1.48 anton 3165: @example
3166: : bar ( -- n )
3167: [ 2 2 + ] literal ;
3168: see bar
3169: @end example
1.29 crook 3170:
1.48 anton 3171: @assignment
3172: Write @code{]L} which allows writing the example above as @code{: bar (
3173: -- n ) [ 2 2 + ]L ;}
3174: @endassignment
3175:
3176:
3177: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3178: @section Advanced macros
3179:
3180: Reconsider @code{map-array} from @ref{Execution Tokens
3181: Tutorial,, Execution Tokens}. It frequently performs @code{execute}, a
3182: relatively expensive operation in some implementations. You can use
3183: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3184: and produce a word that contains the word to be performed directly:
3185:
3186: @c use ]] ... [[
3187: @example
3188: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3189: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3190: \ array beginning at addr and containing u elements
3191: @{ xt @}
3192: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50 anton 3193: POSTPONE i POSTPONE @@ xt compile,
1.48 anton 3194: 1 cells POSTPONE literal POSTPONE +loop ;
3195:
3196: : sum-array ( addr u -- n )
3197: 0 rot rot [ ' + compile-map-array ] ;
3198: see sum-array
3199: a 5 sum-array .
3200: @end example
3201:
3202: You can use the full power of Forth for generating the code; here's an
3203: example where the code is generated in a loop:
3204:
3205: @example
3206: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3207: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50 anton 3208: POSTPONE tuck POSTPONE @@
1.48 anton 3209: POSTPONE literal POSTPONE * POSTPONE +
3210: POSTPONE swap POSTPONE cell+ ;
3211:
3212: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51 pazsan 3213: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48 anton 3214: 0 postpone literal postpone swap
3215: [ ' compile-vmul-step compile-map-array ]
3216: postpone drop ;
3217: see compile-vmul
3218:
3219: : a-vmul ( addr -- n )
1.51 pazsan 3220: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48 anton 3221: [ a 5 compile-vmul ] ;
3222: see a-vmul
3223: a a-vmul .
3224: @end example
3225:
3226: This example uses @code{compile-map-array} to show off, but you could
3227: also use @code{map-array} instead (try it now).
3228:
3229: You can use this technique for efficient multiplication of large
3230: matrices. In matrix multiplication, you multiply every line of one
3231: matrix with every column of the other matrix. You can generate the code
3232: for one line once, and use it for every column. The only downside of
3233: this technique is that it is cumbersome to recover the memory consumed
3234: by the generated code when you are done (and in more complicated cases
3235: it is not possible portably).
3236:
3237: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3238: @section Compilation Tokens
3239:
3240: This section is Gforth-specific. You can skip it.
3241:
3242: @code{' word compile,} compiles the interpretation semantics. For words
3243: with default compilation semantics this is the same as performing the
3244: compilation semantics. To represent the compilation semantics of other
3245: words (e.g., words like @code{if} that have no interpretation
3246: semantics), Gforth has the concept of a compilation token (CT,
3247: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3248: You can perform the compilation semantics represented by a CT with
3249: @code{execute}:
1.29 crook 3250:
1.48 anton 3251: @example
3252: : foo2 ( n1 n2 -- n )
3253: [ comp' + execute ] ;
3254: see foo
3255: @end example
1.29 crook 3256:
1.48 anton 3257: You can compile the compilation semantics represented by a CT with
3258: @code{postpone,}:
1.30 anton 3259:
1.48 anton 3260: @example
3261: : foo3 ( -- )
3262: [ comp' + postpone, ] ;
3263: see foo3
3264: @end example
1.30 anton 3265:
1.51 pazsan 3266: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48 anton 3267: @code{comp'} is particularly useful for words that have no
3268: interpretation semantics:
1.29 crook 3269:
1.30 anton 3270: @example
1.48 anton 3271: ' if
1.60 ! anton 3272: comp' if .s 2drop
1.30 anton 3273: @end example
3274:
1.29 crook 3275:
1.48 anton 3276: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3277: @section Wordlists and Search Order
3278:
3279: The dictionary is not just a memory area that allows you to allocate
3280: memory with @code{allot}, it also contains the Forth words, arranged in
3281: several wordlists. When searching for a word in a wordlist,
3282: conceptually you start searching at the youngest and proceed towards
3283: older words (in reality most systems nowadays use hash-tables); i.e., if
3284: you define a word with the same name as an older word, the new word
3285: shadows the older word.
3286:
3287: Which wordlists are searched in which order is determined by the search
3288: order. You can display the search order with @code{order}. It displays
3289: first the search order, starting with the wordlist searched first, then
3290: it displays the wordlist that will contain newly defined words.
1.21 crook 3291:
1.48 anton 3292: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21 crook 3293:
1.48 anton 3294: @example
3295: wordlist constant mywords
3296: @end example
1.21 crook 3297:
1.48 anton 3298: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3299: defined words (the @emph{current} wordlist):
1.21 crook 3300:
1.48 anton 3301: @example
3302: mywords set-current
3303: order
3304: @end example
1.26 crook 3305:
1.48 anton 3306: Gforth does not display a name for the wordlist in @code{mywords}
3307: because this wordlist was created anonymously with @code{wordlist}.
1.21 crook 3308:
1.48 anton 3309: You can get the current wordlist with @code{get-current ( -- wid)}. If
3310: you want to put something into a specific wordlist without overall
3311: effect on the current wordlist, this typically looks like this:
1.21 crook 3312:
1.48 anton 3313: @example
3314: get-current mywords set-current ( wid )
3315: create someword
3316: ( wid ) set-current
3317: @end example
1.21 crook 3318:
1.48 anton 3319: You can write the search order with @code{set-order ( wid1 .. widn n --
3320: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3321: searched wordlist is topmost.
1.21 crook 3322:
1.48 anton 3323: @example
3324: get-order mywords swap 1+ set-order
3325: order
3326: @end example
1.21 crook 3327:
1.48 anton 3328: Yes, the order of wordlists in the output of @code{order} is reversed
3329: from stack comments and the output of @code{.s} and thus unintuitive.
1.21 crook 3330:
1.48 anton 3331: @assignment
3332: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3333: wordlist to the search order. Define @code{previous ( -- )}, which
3334: removes the first searched wordlist from the search order. Experiment
3335: with boundary conditions (you will see some crashes or situations that
3336: are hard or impossible to leave).
3337: @endassignment
1.21 crook 3338:
1.48 anton 3339: The search order is a powerful foundation for providing features similar
3340: to Modula-2 modules and C++ namespaces. However, trying to modularize
3341: programs in this way has disadvantages for debugging and reuse/factoring
3342: that overcome the advantages in my experience (I don't do huge projects,
1.55 anton 3343: though). These disadvantages are not so clear in other
1.48 anton 3344: languages/programming environments, because these langauges are not so
3345: strong in debugging and reuse.
1.21 crook 3346:
3347:
1.29 crook 3348: @c ******************************************************************
1.48 anton 3349: @node Introduction, Words, Tutorial, Top
1.29 crook 3350: @comment node-name, next, previous, up
3351: @chapter An Introduction to ANS Forth
3352: @cindex Forth - an introduction
1.21 crook 3353:
1.29 crook 3354: The primary purpose of this manual is to document Gforth. However, since
3355: Forth is not a widely-known language and there is a lack of up-to-date
3356: teaching material, it seems worthwhile to provide some introductory
1.49 anton 3357: material. For other sources of Forth-related
3358: information, see @ref{Forth-related information}.
1.21 crook 3359:
1.29 crook 3360: The examples in this section should work on any ANS Forth; the
3361: output shown was produced using Gforth. Each example attempts to
3362: reproduce the exact output that Gforth produces. If you try out the
3363: examples (and you should), what you should type is shown @kbd{like this}
3364: and Gforth's response is shown @code{like this}. The single exception is
1.30 anton 3365: that, where the example shows @key{RET} it means that you should
1.29 crook 3366: press the ``carriage return'' key. Unfortunately, some output formats for
3367: this manual cannot show the difference between @kbd{this} and
3368: @code{this} which will make trying out the examples harder (but not
3369: impossible).
1.21 crook 3370:
1.29 crook 3371: Forth is an unusual language. It provides an interactive development
3372: environment which includes both an interpreter and compiler. Forth
3373: programming style encourages you to break a problem down into many
3374: @cindex factoring
3375: small fragments (@dfn{factoring}), and then to develop and test each
3376: fragment interactively. Forth advocates assert that breaking the
3377: edit-compile-test cycle used by conventional programming languages can
3378: lead to great productivity improvements.
1.21 crook 3379:
1.29 crook 3380: @menu
3381: * Introducing the Text Interpreter::
3382: * Stacks and Postfix notation::
3383: * Your first definition::
3384: * How does that work?::
3385: * Forth is written in Forth::
3386: * Review - elements of a Forth system::
3387: * Where to go next::
3388: * Exercises::
3389: @end menu
1.21 crook 3390:
1.29 crook 3391: @comment ----------------------------------------------
3392: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3393: @section Introducing the Text Interpreter
3394: @cindex text interpreter
3395: @cindex outer interpreter
1.21 crook 3396:
1.30 anton 3397: @c IMO this is too detailed and the pace is too slow for
3398: @c an introduction. If you know German, take a look at
3399: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3400: @c to see how I do it - anton
3401:
1.44 crook 3402: @c nac-> Where I have accepted your comments 100% and modified the text
3403: @c accordingly, I have deleted your comments. Elsewhere I have added a
3404: @c response like this to attempt to rationalise what I have done. Of
3405: @c course, this is a very clumsy mechanism for something that would be
3406: @c done far more efficiently over a beer. Please delete any dialogue
3407: @c you consider closed.
3408:
1.29 crook 3409: When you invoke the Forth image, you will see a startup banner printed
3410: and nothing else (if you have Gforth installed on your system, try
1.30 anton 3411: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29 crook 3412: its command line interpreter, which is called the @dfn{Text Interpreter}
3413: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
1.49 anton 3414: about the text interpreter as you read through this chapter, for more
3415: detail @pxref{The Text Interpreter}).
1.21 crook 3416:
1.29 crook 3417: Although it's not obvious, Forth is actually waiting for your
1.30 anton 3418: input. Type a number and press the @key{RET} key:
1.21 crook 3419:
1.26 crook 3420: @example
1.30 anton 3421: @kbd{45@key{RET}} ok
1.26 crook 3422: @end example
1.21 crook 3423:
1.29 crook 3424: Rather than give you a prompt to invite you to input something, the text
3425: interpreter prints a status message @i{after} it has processed a line
3426: of input. The status message in this case (``@code{ ok}'' followed by
3427: carriage-return) indicates that the text interpreter was able to process
3428: all of your input successfully. Now type something illegal:
3429:
3430: @example
1.30 anton 3431: @kbd{qwer341@key{RET}}
1.29 crook 3432: :1: Undefined word
3433: qwer341
3434: ^^^^^^^
3435: $400D2BA8 Bounce
3436: $400DBDA8 no.extensions
3437: @end example
1.23 crook 3438:
1.29 crook 3439: The exact text, other than the ``Undefined word'' may differ slightly on
3440: your system, but the effect is the same; when the text interpreter
3441: detects an error, it discards any remaining text on a line, resets
1.49 anton 3442: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3443: messages}.
1.23 crook 3444:
1.29 crook 3445: The text interpreter waits for you to press carriage-return, and then
3446: processes your input line. Starting at the beginning of the line, it
3447: breaks the line into groups of characters separated by spaces. For each
3448: group of characters in turn, it makes two attempts to do something:
1.23 crook 3449:
1.29 crook 3450: @itemize @bullet
3451: @item
1.44 crook 3452: @cindex name dictionary
1.29 crook 3453: It tries to treat it as a command. It does this by searching a @dfn{name
3454: dictionary}. If the group of characters matches an entry in the name
3455: dictionary, the name dictionary provides the text interpreter with
3456: information that allows the text interpreter perform some actions. In
3457: Forth jargon, we say that the group
3458: @cindex word
3459: @cindex definition
3460: @cindex execution token
3461: @cindex xt
3462: of characters names a @dfn{word}, that the dictionary search returns an
3463: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3464: word, and that the text interpreter executes the xt. Often, the terms
3465: @dfn{word} and @dfn{definition} are used interchangeably.
3466: @item
3467: If the text interpreter fails to find a match in the name dictionary, it
3468: tries to treat the group of characters as a number in the current number
3469: base (when you start up Forth, the current number base is base 10). If
3470: the group of characters legitimately represents a number, the text
3471: interpreter pushes the number onto a stack (we'll learn more about that
3472: in the next section).
3473: @end itemize
1.23 crook 3474:
1.29 crook 3475: If the text interpreter is unable to do either of these things with any
3476: group of characters, it discards the group of characters and the rest of
3477: the line, then prints an error message. If the text interpreter reaches
3478: the end of the line without error, it prints the status message ``@code{ ok}''
3479: followed by carriage-return.
1.21 crook 3480:
1.29 crook 3481: This is the simplest command we can give to the text interpreter:
1.23 crook 3482:
3483: @example
1.30 anton 3484: @key{RET} ok
1.23 crook 3485: @end example
1.21 crook 3486:
1.29 crook 3487: The text interpreter did everything we asked it to do (nothing) without
3488: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3489: command:
1.21 crook 3490:
1.23 crook 3491: @example
1.30 anton 3492: @kbd{12 dup fred dup@key{RET}}
1.29 crook 3493: :1: Undefined word
3494: 12 dup fred dup
3495: ^^^^
3496: $400D2BA8 Bounce
3497: $400DBDA8 no.extensions
1.23 crook 3498: @end example
1.21 crook 3499:
1.29 crook 3500: When you press the carriage-return key, the text interpreter starts to
3501: work its way along the line:
1.21 crook 3502:
1.29 crook 3503: @itemize @bullet
3504: @item
3505: When it gets to the space after the @code{2}, it takes the group of
3506: characters @code{12} and looks them up in the name
3507: dictionary@footnote{We can't tell if it found them or not, but assume
3508: for now that it did not}. There is no match for this group of characters
3509: in the name dictionary, so it tries to treat them as a number. It is
3510: able to do this successfully, so it puts the number, 12, ``on the stack''
3511: (whatever that means).
3512: @item
3513: The text interpreter resumes scanning the line and gets the next group
3514: of characters, @code{dup}. It looks it up in the name dictionary and
3515: (you'll have to take my word for this) finds it, and executes the word
3516: @code{dup} (whatever that means).
3517: @item
3518: Once again, the text interpreter resumes scanning the line and gets the
3519: group of characters @code{fred}. It looks them up in the name
3520: dictionary, but can't find them. It tries to treat them as a number, but
3521: they don't represent any legal number.
3522: @end itemize
1.21 crook 3523:
1.29 crook 3524: At this point, the text interpreter gives up and prints an error
3525: message. The error message shows exactly how far the text interpreter
3526: got in processing the line. In particular, it shows that the text
3527: interpreter made no attempt to do anything with the final character
3528: group, @code{dup}, even though we have good reason to believe that the
3529: text interpreter would have no problem looking that word up and
3530: executing it a second time.
1.21 crook 3531:
3532:
1.29 crook 3533: @comment ----------------------------------------------
3534: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3535: @section Stacks, postfix notation and parameter passing
3536: @cindex text interpreter
3537: @cindex outer interpreter
1.21 crook 3538:
1.29 crook 3539: In procedural programming languages (like C and Pascal), the
3540: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3541: functions or procedures are called with @dfn{explicit parameters}. For
3542: example, in C we might write:
1.21 crook 3543:
1.23 crook 3544: @example
1.29 crook 3545: total = total + new_volume(length,height,depth);
1.23 crook 3546: @end example
1.21 crook 3547:
1.23 crook 3548: @noindent
1.29 crook 3549: where new_volume is a function-call to another piece of code, and total,
3550: length, height and depth are all variables. length, height and depth are
3551: parameters to the function-call.
1.21 crook 3552:
1.29 crook 3553: In Forth, the equivalent of the function or procedure is the
3554: @dfn{definition} and parameters are implicitly passed between
3555: definitions using a shared stack that is visible to the
3556: programmer. Although Forth does support variables, the existence of the
3557: stack means that they are used far less often than in most other
3558: programming languages. When the text interpreter encounters a number, it
3559: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30 anton 3560: actual number is implementation-dependent ...) and the particular stack
1.29 crook 3561: used for any operation is implied unambiguously by the operation being
3562: performed. The stack used for all integer operations is called the @dfn{data
3563: stack} and, since this is the stack used most commonly, references to
3564: ``the data stack'' are often abbreviated to ``the stack''.
1.21 crook 3565:
1.29 crook 3566: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21 crook 3567:
1.23 crook 3568: @example
1.30 anton 3569: @kbd{1 2 3@key{RET}} ok
1.23 crook 3570: @end example
1.21 crook 3571:
1.29 crook 3572: Then this instructs the text interpreter to placed three numbers on the
3573: (data) stack. An analogy for the behaviour of the stack is to take a
3574: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3575: the table. The 3 was the last card onto the pile (``last-in'') and if
3576: you take a card off the pile then, unless you're prepared to fiddle a
3577: bit, the card that you take off will be the 3 (``first-out''). The
3578: number that will be first-out of the stack is called the @dfn{top of
3579: stack}, which
3580: @cindex TOS definition
3581: is often abbreviated to @dfn{TOS}.
1.21 crook 3582:
1.29 crook 3583: To understand how parameters are passed in Forth, consider the
3584: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3585: be surprised to learn that this definition performs addition. More
3586: precisely, it adds two number together and produces a result. Where does
3587: it get the two numbers from? It takes the top two numbers off the
3588: stack. Where does it place the result? On the stack. You can act-out the
3589: behaviour of @code{+} with your playing cards like this:
1.21 crook 3590:
3591: @itemize @bullet
3592: @item
1.29 crook 3593: Pick up two cards from the stack on the table
1.21 crook 3594: @item
1.29 crook 3595: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3596: numbers''
1.21 crook 3597: @item
1.29 crook 3598: Decide that the answer is 5
1.21 crook 3599: @item
1.29 crook 3600: Shuffle the two cards back into the pack and find a 5
1.21 crook 3601: @item
1.29 crook 3602: Put a 5 on the remaining ace that's on the table.
1.21 crook 3603: @end itemize
3604:
1.29 crook 3605: If you don't have a pack of cards handy but you do have Forth running,
3606: you can use the definition @code{.s} to show the current state of the stack,
3607: without affecting the stack. Type:
1.21 crook 3608:
3609: @example
1.30 anton 3610: @kbd{clearstack 1 2 3@key{RET}} ok
3611: @kbd{.s@key{RET}} <3> 1 2 3 ok
1.23 crook 3612: @end example
3613:
1.29 crook 3614: The text interpreter looks up the word @code{clearstack} and executes
3615: it; it tidies up the stack and removes any entries that may have been
3616: left on it by earlier examples. The text interpreter pushes each of the
3617: three numbers in turn onto the stack. Finally, the text interpreter
3618: looks up the word @code{.s} and executes it. The effect of executing
3619: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3620: followed by a list of all the items on the stack; the item on the far
3621: right-hand side is the TOS.
1.21 crook 3622:
1.29 crook 3623: You can now type:
1.21 crook 3624:
1.29 crook 3625: @example
1.30 anton 3626: @kbd{+ .s@key{RET}} <2> 1 5 ok
1.29 crook 3627: @end example
1.21 crook 3628:
1.29 crook 3629: @noindent
3630: which is correct; there are now 2 items on the stack and the result of
3631: the addition is 5.
1.23 crook 3632:
1.29 crook 3633: If you're playing with cards, try doing a second addition: pick up the
3634: two cards, work out that their sum is 6, shuffle them into the pack,
3635: look for a 6 and place that on the table. You now have just one item on
3636: the stack. What happens if you try to do a third addition? Pick up the
3637: first card, pick up the second card -- ah! There is no second card. This
3638: is called a @dfn{stack underflow} and consitutes an error. If you try to
3639: do the same thing with Forth it will report an error (probably a Stack
3640: Underflow or an Invalid Memory Address error).
1.23 crook 3641:
1.29 crook 3642: The opposite situation to a stack underflow is a @dfn{stack overflow},
3643: which simply accepts that there is a finite amount of storage space
3644: reserved for the stack. To stretch the playing card analogy, if you had
3645: enough packs of cards and you piled the cards up on the table, you would
3646: eventually be unable to add another card; you'd hit the ceiling. Gforth
3647: allows you to set the maximum size of the stacks. In general, the only
3648: time that you will get a stack overflow is because a definition has a
3649: bug in it and is generating data on the stack uncontrollably.
1.23 crook 3650:
1.29 crook 3651: There's one final use for the playing card analogy. If you model your
3652: stack using a pack of playing cards, the maximum number of items on
3653: your stack will be 52 (I assume you didn't use the Joker). The maximum
3654: @i{value} of any item on the stack is 13 (the King). In fact, the only
3655: possible numbers are positive integer numbers 1 through 13; you can't
3656: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3657: think about some of the cards, you can accommodate different
3658: numbers. For example, you could think of the Jack as representing 0,
3659: the Queen as representing -1 and the King as representing -2. Your
1.45 crook 3660: @i{range} remains unchanged (you can still only represent a total of 13
1.29 crook 3661: numbers) but the numbers that you can represent are -2 through 10.
1.28 crook 3662:
1.29 crook 3663: In that analogy, the limit was the amount of information that a single
3664: stack entry could hold, and Forth has a similar limit. In Forth, the
3665: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3666: implementation dependent and affects the maximum value that a stack
3667: entry can hold. A Standard Forth provides a cell size of at least
3668: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21 crook 3669:
1.29 crook 3670: Forth does not do any type checking for you, so you are free to
3671: manipulate and combine stack items in any way you wish. A convenient way
3672: of treating stack items is as 2's complement signed integers, and that
3673: is what Standard words like @code{+} do. Therefore you can type:
1.21 crook 3674:
1.29 crook 3675: @example
1.30 anton 3676: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
1.29 crook 3677: @end example
1.21 crook 3678:
1.29 crook 3679: If you use numbers and definitions like @code{+} in order to turn Forth
3680: into a great big pocket calculator, you will realise that it's rather
3681: different from a normal calculator. Rather than typing 2 + 3 = you had
3682: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3683: result). The terminology used to describe this difference is to say that
3684: your calculator uses @dfn{Infix Notation} (parameters and operators are
3685: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3686: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21 crook 3687:
1.29 crook 3688: Whilst postfix notation might look confusing to begin with, it has
3689: several important advantages:
1.21 crook 3690:
1.23 crook 3691: @itemize @bullet
3692: @item
1.29 crook 3693: it is unambiguous
1.23 crook 3694: @item
1.29 crook 3695: it is more concise
1.23 crook 3696: @item
1.29 crook 3697: it fits naturally with a stack-based system
1.23 crook 3698: @end itemize
1.21 crook 3699:
1.29 crook 3700: To examine these claims in more detail, consider these sums:
1.21 crook 3701:
1.29 crook 3702: @example
3703: 6 + 5 * 4 =
3704: 4 * 5 + 6 =
3705: @end example
1.21 crook 3706:
1.29 crook 3707: If you're just learning maths or your maths is very rusty, you will
3708: probably come up with the answer 44 for the first and 26 for the
3709: second. If you are a bit of a whizz at maths you will remember the
3710: @i{convention} that multiplication takes precendence over addition, and
3711: you'd come up with the answer 26 both times. To explain the answer 26
3712: to someone who got the answer 44, you'd probably rewrite the first sum
3713: like this:
1.21 crook 3714:
1.29 crook 3715: @example
3716: 6 + (5 * 4) =
3717: @end example
1.21 crook 3718:
1.29 crook 3719: If what you really wanted was to perform the addition before the
3720: multiplication, you would have to use parentheses to force it.
1.21 crook 3721:
1.29 crook 3722: If you did the first two sums on a pocket calculator you would probably
3723: get the right answers, unless you were very cautious and entered them using
3724: these keystroke sequences:
1.21 crook 3725:
1.29 crook 3726: 6 + 5 = * 4 =
3727: 4 * 5 = + 6 =
1.21 crook 3728:
1.29 crook 3729: Postfix notation is unambiguous because the order that the operators
3730: are applied is always explicit; that also means that parentheses are
3731: never required. The operators are @i{active} (the act of quoting the
3732: operator makes the operation occur) which removes the need for ``=''.
1.28 crook 3733:
1.29 crook 3734: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3735: equivalent ways:
1.26 crook 3736:
3737: @example
1.29 crook 3738: 6 5 4 * + or:
3739: 5 4 * 6 +
1.26 crook 3740: @end example
1.23 crook 3741:
1.29 crook 3742: An important thing that you should notice about this notation is that
3743: the @i{order} of the numbers does not change; if you want to subtract
3744: 2 from 10 you type @code{10 2 -}.
1.1 anton 3745:
1.29 crook 3746: The reason that Forth uses postfix notation is very simple to explain: it
3747: makes the implementation extremely simple, and it follows naturally from
3748: using the stack as a mechanism for passing parameters. Another way of
3749: thinking about this is to realise that all Forth definitions are
3750: @i{active}; they execute as they are encountered by the text
3751: interpreter. The result of this is that the syntax of Forth is trivially
3752: simple.
1.1 anton 3753:
3754:
3755:
1.29 crook 3756: @comment ----------------------------------------------
3757: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3758: @section Your first Forth definition
3759: @cindex first definition
1.1 anton 3760:
1.29 crook 3761: Until now, the examples we've seen have been trivial; we've just been
3762: using Forth as a bigger-than-pocket calculator. Also, each calculation
3763: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3764: again@footnote{That's not quite true. If you press the up-arrow key on
3765: your keyboard you should be able to scroll back to any earlier command,
3766: edit it and re-enter it.} In this section we'll see how to add new
3767: words to Forth's vocabulary.
1.1 anton 3768:
1.29 crook 3769: The easiest way to create a new word is to use a @dfn{colon
3770: definition}. We'll define a few and try them out before worrying too
3771: much about how they work. Try typing in these examples; be careful to
3772: copy the spaces accurately:
1.1 anton 3773:
1.29 crook 3774: @example
3775: : add-two 2 + . ;
3776: : greet ." Hello and welcome" ;
3777: : demo 5 add-two ;
3778: @end example
1.1 anton 3779:
1.29 crook 3780: @noindent
3781: Now try them out:
1.1 anton 3782:
1.29 crook 3783: @example
1.30 anton 3784: @kbd{greet@key{RET}} Hello and welcome ok
3785: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3786: @kbd{4 add-two@key{RET}} 6 ok
3787: @kbd{demo@key{RET}} 7 ok
3788: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
1.29 crook 3789: @end example
1.1 anton 3790:
1.29 crook 3791: The first new thing that we've introduced here is the pair of words
3792: @code{:} and @code{;}. These are used to start and terminate a new
3793: definition, respectively. The first word after the @code{:} is the name
3794: for the new definition.
1.1 anton 3795:
1.29 crook 3796: As you can see from the examples, a definition is built up of words that
3797: have already been defined; Forth makes no distinction between
3798: definitions that existed when you started the system up, and those that
3799: you define yourself.
1.1 anton 3800:
1.29 crook 3801: The examples also introduce the words @code{.} (dot), @code{."}
3802: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3803: the stack and displays it. It's like @code{.s} except that it only
3804: displays the top item of the stack and it is destructive; after it has
3805: executed, the number is no longer on the stack. There is always one
3806: space printed after the number, and no spaces before it. Dot-quote
3807: defines a string (a sequence of characters) that will be printed when
3808: the word is executed. The string can contain any printable characters
3809: except @code{"}. A @code{"} has a special function; it is not a Forth
3810: word but it acts as a delimiter (the way that delimiters work is
3811: described in the next section). Finally, @code{dup} duplicates the value
3812: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1 anton 3813:
1.29 crook 3814: We already know that the text interpreter searches through the
3815: dictionary to locate names. If you've followed the examples earlier, you
3816: will already have a definition called @code{add-two}. Lets try modifying
3817: it by typing in a new definition:
1.1 anton 3818:
1.29 crook 3819: @example
1.30 anton 3820: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
1.29 crook 3821: @end example
1.5 anton 3822:
1.29 crook 3823: Forth recognised that we were defining a word that already exists, and
3824: printed a message to warn us of that fact. Let's try out the new
3825: definition:
1.5 anton 3826:
1.29 crook 3827: @example
1.30 anton 3828: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
1.29 crook 3829: @end example
1.1 anton 3830:
1.29 crook 3831: @noindent
3832: All that we've actually done here, though, is to create a new
3833: definition, with a particular name. The fact that there was already a
3834: definition with the same name did not make any difference to the way
3835: that the new definition was created (except that Forth printed a warning
3836: message). The old definition of add-two still exists (try @code{demo}
3837: again to see that this is true). Any new definition will use the new
3838: definition of @code{add-two}, but old definitions continue to use the
3839: version that already existed at the time that they were @code{compiled}.
1.1 anton 3840:
1.29 crook 3841: Before you go on to the next section, try defining and redefining some
3842: words of your own.
1.1 anton 3843:
1.29 crook 3844: @comment ----------------------------------------------
3845: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3846: @section How does that work?
3847: @cindex parsing words
1.1 anton 3848:
1.30 anton 3849: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3850:
3851: @c Is it a good idea to talk about the interpretation semantics of a
3852: @c number? We don't have an xt to go along with it. - anton
3853:
3854: @c Now that I have eliminated execution semantics, I wonder if it would not
3855: @c be better to keep them (or add run-time semantics), to make it easier to
3856: @c explain what compilation semantics usually does. - anton
3857:
1.44 crook 3858: @c nac-> I removed the term ``default compilation sematics'' from the
3859: @c introductory chapter. Removing ``execution semantics'' was making
3860: @c everything simpler to explain, then I think the use of this term made
3861: @c everything more complex again. I replaced it with ``default
3862: @c semantics'' (which is used elsewhere in the manual) by which I mean
3863: @c ``a definition that has neither the immediate nor the compile-only
3864: @c flag set''. I reworded big chunks of the ``how does that work''
3865: @c section (and, unusually for me, I think I even made it shorter!). See
3866: @c what you think -- I know I have not addressed your primary concern
3867: @c that it is too heavy-going for an introduction. From what I understood
3868: @c of your course notes it looks as though they might be a good framework.
3869: @c Things that I've tried to capture here are some things that came as a
3870: @c great revelation here when I first understood them. Also, I like the
3871: @c fact that a very simple code example shows up almost all of the issues
3872: @c that you need to understand to see how Forth works. That's unique and
3873: @c worthwhile to emphasise.
3874:
1.29 crook 3875: Now we're going to take another look at the definition of @code{add-two}
3876: from the previous section. From our knowledge of the way that the text
3877: interpreter works, we would have expected this result when we tried to
3878: define @code{add-two}:
1.21 crook 3879:
1.29 crook 3880: @example
1.44 crook 3881: @kbd{: add-two 2 + . ;@key{RET}}
1.29 crook 3882: ^^^^^^^
3883: Error: Undefined word
3884: @end example
1.28 crook 3885:
1.29 crook 3886: The reason that this didn't happen is bound up in the way that @code{:}
3887: works. The word @code{:} does two special things. The first special
3888: thing that it does prevents the text interpreter from ever seeing the
3889: characters @code{add-two}. The text interpreter uses a variable called
3890: @cindex modifying >IN
1.44 crook 3891: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29 crook 3892: input line. When it encounters the word @code{:} it behaves in exactly
3893: the same way as it does for any other word; it looks it up in the name
3894: dictionary, finds its xt and executes it. When @code{:} executes, it
3895: looks at the input buffer, finds the word @code{add-two} and advances the
3896: value of @code{>IN} to point past it. It then does some other stuff
3897: associated with creating the new definition (including creating an entry
3898: for @code{add-two} in the name dictionary). When the execution of @code{:}
3899: completes, control returns to the text interpreter, which is oblivious
3900: to the fact that it has been tricked into ignoring part of the input
3901: line.
1.21 crook 3902:
1.29 crook 3903: @cindex parsing words
3904: Words like @code{:} -- words that advance the value of @code{>IN} and so
3905: prevent the text interpreter from acting on the whole of the input line
3906: -- are called @dfn{parsing words}.
1.21 crook 3907:
1.29 crook 3908: @cindex @code{state} - effect on the text interpreter
3909: @cindex text interpreter - effect of state
3910: The second special thing that @code{:} does is change the value of a
3911: variable called @code{state}, which affects the way that the text
3912: interpreter behaves. When Gforth starts up, @code{state} has the value
3913: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3914: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44 crook 3915: the text interpreter is said to be @dfn{compiling}.
3916:
3917: In this example, the text interpreter is compiling when it processes the
3918: string ``@code{2 + . ;}''. It still breaks the string down into
3919: character sequences in the same way. However, instead of pushing the
3920: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3921: into the definition of @code{add-two} that will make the number @code{2} get
3922: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3923: the behaviours of @code{+} and @code{.} are also compiled into the
3924: definition.
3925:
3926: One category of words don't get compiled. These so-called @dfn{immediate
3927: words} get executed (performed @i{now}) regardless of whether the text
3928: interpreter is interpreting or compiling. The word @code{;} is an
3929: immediate word. Rather than being compiled into the definition, it
3930: executes. Its effect is to terminate the current definition, which
3931: includes changing the value of @code{state} back to 0.
3932:
3933: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
3934: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
3935: definition.
1.28 crook 3936:
1.30 anton 3937: In Forth, every word or number can be described in terms of two
1.29 crook 3938: properties:
1.28 crook 3939:
3940: @itemize @bullet
3941: @item
1.29 crook 3942: @cindex interpretation semantics
1.44 crook 3943: Its @dfn{interpretation semantics} describe how it will behave when the
3944: text interpreter encounters it in @dfn{interpret} state. The
3945: interpretation semantics of a word are represented by an @dfn{execution
3946: token}.
1.28 crook 3947: @item
1.29 crook 3948: @cindex compilation semantics
1.44 crook 3949: Its @dfn{compilation semantics} describe how it will behave when the
3950: text interpreter encounters it in @dfn{compile} state. The compilation
3951: semantics of a word are represented in an implementation-dependent way;
3952: Gforth uses a @dfn{compilation token}.
1.29 crook 3953: @end itemize
3954:
3955: @noindent
3956: Numbers are always treated in a fixed way:
3957:
3958: @itemize @bullet
1.28 crook 3959: @item
1.44 crook 3960: When the number is @dfn{interpreted}, its behaviour is to push the
3961: number onto the stack.
1.28 crook 3962: @item
1.30 anton 3963: When the number is @dfn{compiled}, a piece of code is appended to the
3964: current definition that pushes the number when it runs. (In other words,
3965: the compilation semantics of a number are to postpone its interpretation
3966: semantics until the run-time of the definition that it is being compiled
3967: into.)
1.29 crook 3968: @end itemize
3969:
1.44 crook 3970: Words don't behave in such a regular way, but most have @i{default
3971: semantics} which means that they behave like this:
1.29 crook 3972:
3973: @itemize @bullet
1.28 crook 3974: @item
1.30 anton 3975: The @dfn{interpretation semantics} of the word are to do something useful.
3976: @item
1.29 crook 3977: The @dfn{compilation semantics} of the word are to append its
1.30 anton 3978: @dfn{interpretation semantics} to the current definition (so that its
3979: run-time behaviour is to do something useful).
1.28 crook 3980: @end itemize
3981:
1.30 anton 3982: @cindex immediate words
1.44 crook 3983: The actual behaviour of any particular word can be controlled by using
3984: the words @code{immediate} and @code{compile-only} when the word is
3985: defined. These words set flags in the name dictionary entry of the most
3986: recently defined word, and these flags are retrieved by the text
3987: interpreter when it finds the word in the name dictionary.
3988:
3989: A word that is marked as @dfn{immediate} has compilation semantics that
3990: are identical to its interpretation semantics. In other words, it
3991: behaves like this:
1.29 crook 3992:
3993: @itemize @bullet
3994: @item
1.30 anton 3995: The @dfn{interpretation semantics} of the word are to do something useful.
1.29 crook 3996: @item
1.30 anton 3997: The @dfn{compilation semantics} of the word are to do something useful
3998: (and actually the same thing); i.e., it is executed during compilation.
1.29 crook 3999: @end itemize
1.28 crook 4000:
1.44 crook 4001: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4002: performing the interpretation semantics of the word directly; an attempt
4003: to do so will generate an error. It is never necessary to use
4004: @code{compile-only} (and it is not even part of ANS Forth, though it is
4005: provided by many implementations) but it is good etiquette to apply it
4006: to a word that will not behave correctly (and might have unexpected
4007: side-effects) in interpret state. For example, it is only legal to use
4008: the conditional word @code{IF} within a definition. If you forget this
4009: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4010: @code{compile-only} allows the text interpreter to generate a helpful
4011: error message rather than subjecting you to the consequences of your
4012: folly.
4013:
1.29 crook 4014: This example shows the difference between an immediate and a
4015: non-immediate word:
1.28 crook 4016:
1.29 crook 4017: @example
4018: : show-state state @@ . ;
4019: : show-state-now show-state ; immediate
4020: : word1 show-state ;
4021: : word2 show-state-now ;
1.28 crook 4022: @end example
1.23 crook 4023:
1.29 crook 4024: The word @code{immediate} after the definition of @code{show-state-now}
4025: makes that word an immediate word. These definitions introduce a new
4026: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4027: variable, and leaves it on the stack. Therefore, the behaviour of
4028: @code{show-state} is to print a number that represents the current value
4029: of @code{state}.
1.28 crook 4030:
1.29 crook 4031: When you execute @code{word1}, it prints the number 0, indicating that
4032: the system is interpreting. When the text interpreter compiled the
4033: definition of @code{word1}, it encountered @code{show-state} whose
1.30 anton 4034: compilation semantics are to append its interpretation semantics to the
1.29 crook 4035: current definition. When you execute @code{word1}, it performs the
1.30 anton 4036: interpretation semantics of @code{show-state}. At the time that @code{word1}
1.29 crook 4037: (and therefore @code{show-state}) are executed, the system is
4038: interpreting.
1.28 crook 4039:
1.30 anton 4040: When you pressed @key{RET} after entering the definition of @code{word2},
1.29 crook 4041: you should have seen the number -1 printed, followed by ``@code{
4042: ok}''. When the text interpreter compiled the definition of
4043: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30 anton 4044: whose compilation semantics are therefore to perform its interpretation
1.29 crook 4045: semantics. It is executed straight away (even before the text
4046: interpreter has moved on to process another group of characters; the
4047: @code{;} in this example). The effect of executing it are to display the
4048: value of @code{state} @i{at the time that the definition of}
4049: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4050: system is compiling at this time. If you execute @code{word2} it does
4051: nothing at all.
1.28 crook 4052:
1.29 crook 4053: @cindex @code{."}, how it works
4054: Before leaving the subject of immediate words, consider the behaviour of
4055: @code{."} in the definition of @code{greet}, in the previous
4056: section. This word is both a parsing word and an immediate word. Notice
4057: that there is a space between @code{."} and the start of the text
4058: @code{Hello and welcome}, but that there is no space between the last
4059: letter of @code{welcome} and the @code{"} character. The reason for this
4060: is that @code{."} is a Forth word; it must have a space after it so that
4061: the text interpreter can identify it. The @code{"} is not a Forth word;
4062: it is a @dfn{delimiter}. The examples earlier show that, when the string
4063: is displayed, there is neither a space before the @code{H} nor after the
4064: @code{e}. Since @code{."} is an immediate word, it executes at the time
4065: that @code{greet} is defined. When it executes, its behaviour is to
4066: search forward in the input line looking for the delimiter. When it
4067: finds the delimiter, it updates @code{>IN} to point past the
4068: delimiter. It also compiles some magic code into the definition of
4069: @code{greet}; the xt of a run-time routine that prints a text string. It
4070: compiles the string @code{Hello and welcome} into memory so that it is
4071: available to be printed later. When the text interpreter gains control,
4072: the next word it finds in the input stream is @code{;} and so it
4073: terminates the definition of @code{greet}.
1.28 crook 4074:
4075:
4076: @comment ----------------------------------------------
1.29 crook 4077: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4078: @section Forth is written in Forth
4079: @cindex structure of Forth programs
4080:
4081: When you start up a Forth compiler, a large number of definitions
4082: already exist. In Forth, you develop a new application using bottom-up
4083: programming techniques to create new definitions that are defined in
4084: terms of existing definitions. As you create each definition you can
4085: test and debug it interactively.
4086:
4087: If you have tried out the examples in this section, you will probably
4088: have typed them in by hand; when you leave Gforth, your definitions will
4089: be lost. You can avoid this by using a text editor to enter Forth source
4090: code into a file, and then loading code from the file using
1.49 anton 4091: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29 crook 4092: processed by the text interpreter, just as though you had typed it in by
4093: hand@footnote{Actually, there are some subtle differences -- see
4094: @ref{The Text Interpreter}.}.
4095:
4096: Gforth also supports the traditional Forth alternative to using text
1.49 anton 4097: files for program entry (@pxref{Blocks}).
1.28 crook 4098:
1.29 crook 4099: In common with many, if not most, Forth compilers, most of Gforth is
4100: actually written in Forth. All of the @file{.fs} files in the
4101: installation directory@footnote{For example,
1.30 anton 4102: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29 crook 4103: study to see examples of Forth programming.
1.28 crook 4104:
1.29 crook 4105: Gforth maintains a history file that records every line that you type to
4106: the text interpreter. This file is preserved between sessions, and is
4107: used to provide a command-line recall facility. If you enter long
4108: definitions by hand, you can use a text editor to paste them out of the
4109: history file into a Forth source file for reuse at a later time
1.49 anton 4110: (for more information @pxref{Command-line editing}).
1.28 crook 4111:
4112:
4113: @comment ----------------------------------------------
1.29 crook 4114: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4115: @section Review - elements of a Forth system
4116: @cindex elements of a Forth system
1.28 crook 4117:
1.29 crook 4118: To summarise this chapter:
1.28 crook 4119:
4120: @itemize @bullet
4121: @item
1.29 crook 4122: Forth programs use @dfn{factoring} to break a problem down into small
4123: fragments called @dfn{words} or @dfn{definitions}.
4124: @item
4125: Forth program development is an interactive process.
4126: @item
4127: The main command loop that accepts input, and controls both
4128: interpretation and compilation, is called the @dfn{text interpreter}
4129: (also known as the @dfn{outer interpreter}).
4130: @item
4131: Forth has a very simple syntax, consisting of words and numbers
4132: separated by spaces or carriage-return characters. Any additional syntax
4133: is imposed by @dfn{parsing words}.
4134: @item
4135: Forth uses a stack to pass parameters between words. As a result, it
4136: uses postfix notation.
4137: @item
4138: To use a word that has previously been defined, the text interpreter
4139: searches for the word in the @dfn{name dictionary}.
4140: @item
1.30 anton 4141: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28 crook 4142: @item
1.29 crook 4143: The text interpreter uses the value of @code{state} to select between
4144: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4145: semantics} of a word that it encounters.
1.28 crook 4146: @item
1.30 anton 4147: The relationship between the @dfn{interpretation semantics} and
4148: @dfn{compilation semantics} for a word
1.29 crook 4149: depend upon the way in which the word was defined (for example, whether
4150: it is an @dfn{immediate} word).
1.28 crook 4151: @item
1.29 crook 4152: Forth definitions can be implemented in Forth (called @dfn{high-level
4153: definitions}) or in some other way (usually a lower-level language and
4154: as a result often called @dfn{low-level definitions}, @dfn{code
4155: definitions} or @dfn{primitives}).
1.28 crook 4156: @item
1.29 crook 4157: Many Forth systems are implemented mainly in Forth.
1.28 crook 4158: @end itemize
4159:
4160:
1.29 crook 4161: @comment ----------------------------------------------
1.48 anton 4162: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29 crook 4163: @section Where To Go Next
4164: @cindex where to go next
1.28 crook 4165:
1.29 crook 4166: Amazing as it may seem, if you have read (and understood) this far, you
4167: know almost all the fundamentals about the inner workings of a Forth
4168: system. You certainly know enough to be able to read and understand the
4169: rest of this manual and the ANS Forth document, to learn more about the
4170: facilities that Forth in general and Gforth in particular provide. Even
4171: scarier, you know almost enough to implement your own Forth system.
1.30 anton 4172: However, that's not a good idea just yet... better to try writing some
1.29 crook 4173: programs in Gforth.
1.28 crook 4174:
1.29 crook 4175: Forth has such a rich vocabulary that it can be hard to know where to
4176: start in learning it. This section suggests a few sets of words that are
4177: enough to write small but useful programs. Use the word index in this
4178: document to learn more about each word, then try it out and try to write
4179: small definitions using it. Start by experimenting with these words:
1.28 crook 4180:
4181: @itemize @bullet
4182: @item
1.29 crook 4183: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4184: @item
4185: Comparison: @code{MIN MAX =}
4186: @item
4187: Logic: @code{AND OR XOR NOT}
4188: @item
4189: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28 crook 4190: @item
1.29 crook 4191: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28 crook 4192: @item
1.29 crook 4193: Input/Output: @code{. ." EMIT CR KEY}
1.28 crook 4194: @item
1.29 crook 4195: Defining words: @code{: ; CREATE}
1.28 crook 4196: @item
1.29 crook 4197: Memory allocation words: @code{ALLOT ,}
1.28 crook 4198: @item
1.29 crook 4199: Tools: @code{SEE WORDS .S MARKER}
4200: @end itemize
4201:
4202: When you have mastered those, go on to:
4203:
4204: @itemize @bullet
1.28 crook 4205: @item
1.29 crook 4206: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28 crook 4207: @item
1.29 crook 4208: Memory access: @code{@@ !}
1.28 crook 4209: @end itemize
1.23 crook 4210:
1.29 crook 4211: When you have mastered these, there's nothing for it but to read through
4212: the whole of this manual and find out what you've missed.
4213:
4214: @comment ----------------------------------------------
1.48 anton 4215: @node Exercises, , Where to go next, Introduction
1.29 crook 4216: @section Exercises
4217: @cindex exercises
4218:
4219: TODO: provide a set of programming excercises linked into the stuff done
4220: already and into other sections of the manual. Provide solutions to all
4221: the exercises in a .fs file in the distribution.
4222:
4223: @c Get some inspiration from Starting Forth and Kelly&Spies.
4224:
4225: @c excercises:
4226: @c 1. take inches and convert to feet and inches.
4227: @c 2. take temperature and convert from fahrenheight to celcius;
4228: @c may need to care about symmetric vs floored??
4229: @c 3. take input line and do character substitution
4230: @c to encipher or decipher
4231: @c 4. as above but work on a file for in and out
4232: @c 5. take input line and convert to pig-latin
4233: @c
4234: @c thing of sets of things to exercise then come up with
4235: @c problems that need those things.
4236:
4237:
1.26 crook 4238: @c ******************************************************************
1.29 crook 4239: @node Words, Error messages, Introduction, Top
1.1 anton 4240: @chapter Forth Words
1.26 crook 4241: @cindex words
1.1 anton 4242:
4243: @menu
4244: * Notation::
1.21 crook 4245: * Comments::
4246: * Boolean Flags::
1.1 anton 4247: * Arithmetic::
4248: * Stack Manipulation::
1.5 anton 4249: * Memory::
1.1 anton 4250: * Control Structures::
4251: * Defining Words::
1.47 crook 4252: * Interpretation and Compilation Semantics::
4253: * Tokens for Words::
1.21 crook 4254: * The Text Interpreter::
4255: * Word Lists::
4256: * Environmental Queries::
1.12 anton 4257: * Files::
4258: * Blocks::
4259: * Other I/O::
4260: * Programming Tools::
4261: * Assembler and Code Words::
4262: * Threading Words::
1.26 crook 4263: * Locals::
4264: * Structures::
4265: * Object-oriented Forth::
1.21 crook 4266: * Passing Commands to the OS::
1.47 crook 4267: * Keeping track of Time::
1.21 crook 4268: * Miscellaneous Words::
1.1 anton 4269: @end menu
4270:
1.21 crook 4271: @node Notation, Comments, Words, Words
1.1 anton 4272: @section Notation
4273: @cindex notation of glossary entries
4274: @cindex format of glossary entries
4275: @cindex glossary notation format
4276: @cindex word glossary entry format
4277:
4278: The Forth words are described in this section in the glossary notation
4279: that has become a de-facto standard for Forth texts, i.e.,
4280:
4281: @format
1.29 crook 4282: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
1.1 anton 4283: @end format
1.29 crook 4284: @i{Description}
1.1 anton 4285:
4286: @table @var
4287: @item word
1.28 crook 4288: The name of the word.
1.1 anton 4289:
4290: @item Stack effect
4291: @cindex stack effect
1.29 crook 4292: The stack effect is written in the notation @code{@i{before} --
4293: @i{after}}, where @i{before} and @i{after} describe the top of
1.1 anton 4294: stack entries before and after the execution of the word. The rest of
4295: the stack is not touched by the word. The top of stack is rightmost,
4296: i.e., a stack sequence is written as it is typed in. Note that Gforth
4297: uses a separate floating point stack, but a unified stack
1.29 crook 4298: notation. Also, return stack effects are not shown in @i{stack
4299: effect}, but in @i{Description}. The name of a stack item describes
1.1 anton 4300: the type and/or the function of the item. See below for a discussion of
4301: the types.
4302:
4303: All words have two stack effects: A compile-time stack effect and a
4304: run-time stack effect. The compile-time stack-effect of most words is
1.29 crook 4305: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1 anton 4306: this standard behaviour, or the word does other unusual things at
4307: compile time, both stack effects are shown; otherwise only the run-time
4308: stack effect is shown.
4309:
4310: @cindex pronounciation of words
4311: @item pronunciation
4312: How the word is pronounced.
4313:
4314: @cindex wordset
4315: @item wordset
1.21 crook 4316: The ANS Forth standard is divided into several word sets. A standard
4317: system need not support all of them. Therefore, in theory, the fewer
4318: word sets your program uses the more portable it will be. However, we
4319: suspect that most ANS Forth systems on personal machines will feature
1.26 crook 4320: all word sets. Words that are not defined in ANS Forth have
1.21 crook 4321: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1 anton 4322: describes words that will work in future releases of Gforth;
4323: @code{gforth-internal} words are more volatile. Environmental query
4324: strings are also displayed like words; you can recognize them by the
1.21 crook 4325: @code{environment} in the word set field.
1.1 anton 4326:
4327: @item Description
4328: A description of the behaviour of the word.
4329: @end table
4330:
4331: @cindex types of stack items
4332: @cindex stack item types
4333: The type of a stack item is specified by the character(s) the name
4334: starts with:
4335:
4336: @table @code
4337: @item f
4338: @cindex @code{f}, stack item type
4339: Boolean flags, i.e. @code{false} or @code{true}.
4340: @item c
4341: @cindex @code{c}, stack item type
4342: Char
4343: @item w
4344: @cindex @code{w}, stack item type
4345: Cell, can contain an integer or an address
4346: @item n
4347: @cindex @code{n}, stack item type
4348: signed integer
4349: @item u
4350: @cindex @code{u}, stack item type
4351: unsigned integer
4352: @item d
4353: @cindex @code{d}, stack item type
4354: double sized signed integer
4355: @item ud
4356: @cindex @code{ud}, stack item type
4357: double sized unsigned integer
4358: @item r
4359: @cindex @code{r}, stack item type
4360: Float (on the FP stack)
1.21 crook 4361: @item a-
1.1 anton 4362: @cindex @code{a_}, stack item type
4363: Cell-aligned address
1.21 crook 4364: @item c-
1.1 anton 4365: @cindex @code{c_}, stack item type
4366: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21 crook 4367: @item f-
1.1 anton 4368: @cindex @code{f_}, stack item type
4369: Float-aligned address
1.21 crook 4370: @item df-
1.1 anton 4371: @cindex @code{df_}, stack item type
4372: Address aligned for IEEE double precision float
1.21 crook 4373: @item sf-
1.1 anton 4374: @cindex @code{sf_}, stack item type
4375: Address aligned for IEEE single precision float
4376: @item xt
4377: @cindex @code{xt}, stack item type
4378: Execution token, same size as Cell
4379: @item wid
4380: @cindex @code{wid}, stack item type
1.21 crook 4381: Word list ID, same size as Cell
1.1 anton 4382: @item f83name
4383: @cindex @code{f83name}, stack item type
4384: Pointer to a name structure
4385: @item "
4386: @cindex @code{"}, stack item type
1.12 anton 4387: string in the input stream (not on the stack). The terminating character
4388: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1 anton 4389: quotes.
4390: @end table
4391:
1.21 crook 4392: @node Comments, Boolean Flags, Notation, Words
4393: @section Comments
1.26 crook 4394: @cindex comments
1.21 crook 4395:
1.29 crook 4396: Forth supports two styles of comment; the traditional @i{in-line} comment,
4397: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21 crook 4398:
1.44 crook 4399:
1.23 crook 4400: doc-(
1.21 crook 4401: doc-\
1.23 crook 4402: doc-\G
1.21 crook 4403:
1.44 crook 4404:
1.21 crook 4405: @node Boolean Flags, Arithmetic, Comments, Words
4406: @section Boolean Flags
1.26 crook 4407: @cindex Boolean flags
1.21 crook 4408:
4409: A Boolean flag is cell-sized. A cell with all bits clear represents the
4410: flag @code{false} and a flag with all bits set represents the flag
1.26 crook 4411: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29 crook 4412: a cell that has @i{any} bit set as @code{true}.
1.21 crook 4413:
1.44 crook 4414:
1.21 crook 4415: doc-true
4416: doc-false
1.29 crook 4417: doc-on
4418: doc-off
1.21 crook 4419:
1.44 crook 4420:
1.21 crook 4421: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1 anton 4422: @section Arithmetic
4423: @cindex arithmetic words
4424:
4425: @cindex division with potentially negative operands
4426: Forth arithmetic is not checked, i.e., you will not hear about integer
4427: overflow on addition or multiplication, you may hear about division by
4428: zero if you are lucky. The operator is written after the operands, but
4429: the operands are still in the original order. I.e., the infix @code{2-1}
4430: corresponds to @code{2 1 -}. Forth offers a variety of division
4431: operators. If you perform division with potentially negative operands,
4432: you do not want to use @code{/} or @code{/mod} with its undefined
4433: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4434: former, @pxref{Mixed precision}).
1.26 crook 4435: @comment TODO discuss the different division forms and the std approach
1.1 anton 4436:
4437: @menu
4438: * Single precision::
4439: * Bitwise operations::
1.21 crook 4440: * Double precision:: Double-cell integer arithmetic
4441: * Numeric comparison::
1.29 crook 4442: * Mixed precision:: Operations with single and double-cell integers
1.1 anton 4443: * Floating Point::
4444: @end menu
4445:
4446: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
4447: @subsection Single precision
4448: @cindex single precision arithmetic words
4449:
1.21 crook 4450: By default, numbers in Forth are single-precision integers that are 1
1.26 crook 4451: cell in size. They can be signed or unsigned, depending upon how you
1.49 anton 4452: treat them. For the rules used by the text interpreter for recognising
4453: single-precision integers see @ref{Number Conversion}.
1.21 crook 4454:
1.44 crook 4455:
1.1 anton 4456: doc-+
1.21 crook 4457: doc-1+
1.1 anton 4458: doc--
1.21 crook 4459: doc-1-
1.1 anton 4460: doc-*
4461: doc-/
4462: doc-mod
4463: doc-/mod
4464: doc-negate
4465: doc-abs
4466: doc-min
4467: doc-max
1.21 crook 4468: doc-d>s
1.27 crook 4469: doc-floored
1.1 anton 4470:
1.44 crook 4471:
1.21 crook 4472: @node Bitwise operations, Double precision, Single precision, Arithmetic
1.1 anton 4473: @subsection Bitwise operations
4474: @cindex bitwise operation words
4475:
1.44 crook 4476:
1.1 anton 4477: doc-and
4478: doc-or
4479: doc-xor
4480: doc-invert
1.21 crook 4481: doc-lshift
4482: doc-rshift
1.1 anton 4483: doc-2*
1.21 crook 4484: doc-d2*
1.1 anton 4485: doc-2/
1.21 crook 4486: doc-d2/
4487:
1.44 crook 4488:
1.21 crook 4489: @node Double precision, Numeric comparison, Bitwise operations, Arithmetic
4490: @subsection Double precision
4491: @cindex double precision arithmetic words
4492:
1.49 anton 4493: For the rules used by the text interpreter for
4494: recognising double-precision integers, see @ref{Number Conversion}.
1.21 crook 4495:
4496: A double precision number is represented by a cell pair, with the most
1.31 anton 4497: significant cell at the TOS. It is trivial to convert an unsigned
1.26 crook 4498: single to an (unsigned) double; simply push a @code{0} onto the
4499: TOS. Since numbers are represented by Gforth using 2's complement
4500: arithmetic, converting a signed single to a (signed) double requires
1.31 anton 4501: sign-extension across the most significant cell. This can be achieved
1.26 crook 4502: using @code{s>d}. The moral of the story is that you cannot convert a
4503: number without knowing whether it represents an unsigned or a
4504: signed number.
1.21 crook 4505:
1.44 crook 4506:
1.21 crook 4507: doc-s>d
4508: doc-d+
4509: doc-d-
4510: doc-dnegate
4511: doc-dabs
4512: doc-dmin
4513: doc-dmax
4514:
1.44 crook 4515:
1.21 crook 4516: @node Numeric comparison, Mixed precision, Double precision, Arithmetic
4517: @subsection Numeric comparison
4518: @cindex numeric comparison words
4519:
1.44 crook 4520:
1.28 crook 4521: doc-<
4522: doc-<=
4523: doc-<>
4524: doc-=
4525: doc->
4526: doc->=
4527:
1.21 crook 4528: doc-0<
1.23 crook 4529: doc-0<=
1.21 crook 4530: doc-0<>
4531: doc-0=
1.23 crook 4532: doc-0>
4533: doc-0>=
1.28 crook 4534:
4535: doc-u<
4536: doc-u<=
1.44 crook 4537: @c u<> and u= exist but are the same as <> and =
1.31 anton 4538: @c doc-u<>
4539: @c doc-u=
1.28 crook 4540: doc-u>
4541: doc-u>=
4542:
4543: doc-within
4544:
4545: doc-d<
4546: doc-d<=
4547: doc-d<>
4548: doc-d=
4549: doc-d>
4550: doc-d>=
1.23 crook 4551:
1.21 crook 4552: doc-d0<
1.23 crook 4553: doc-d0<=
4554: doc-d0<>
1.21 crook 4555: doc-d0=
1.23 crook 4556: doc-d0>
4557: doc-d0>=
4558:
1.21 crook 4559: doc-du<
1.28 crook 4560: doc-du<=
1.44 crook 4561: @c du<> and du= exist but are the same as d<> and d=
1.31 anton 4562: @c doc-du<>
4563: @c doc-du=
1.28 crook 4564: doc-du>
4565: doc-du>=
1.1 anton 4566:
1.44 crook 4567:
1.21 crook 4568: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1 anton 4569: @subsection Mixed precision
4570: @cindex mixed precision arithmetic words
4571:
1.44 crook 4572:
1.1 anton 4573: doc-m+
4574: doc-*/
4575: doc-*/mod
4576: doc-m*
4577: doc-um*
4578: doc-m*/
4579: doc-um/mod
4580: doc-fm/mod
4581: doc-sm/rem
4582:
1.44 crook 4583:
1.21 crook 4584: @node Floating Point, , Mixed precision, Arithmetic
1.1 anton 4585: @subsection Floating Point
4586: @cindex floating point arithmetic words
4587:
1.49 anton 4588: For the rules used by the text interpreter for
4589: recognising floating-point numbers see @ref{Number Conversion}.
1.1 anton 4590:
1.32 anton 4591: Gforth has a separate floating point
1.26 crook 4592: stack, but the documentation uses the unified notation.
1.1 anton 4593:
4594: @cindex floating-point arithmetic, pitfalls
4595: Floating point numbers have a number of unpleasant surprises for the
4596: unwary (e.g., floating point addition is not associative) and even a few
4597: for the wary. You should not use them unless you know what you are doing
4598: or you don't care that the results you get are totally bogus. If you
4599: want to learn about the problems of floating point numbers (and how to
4600: avoid them), you might start with @cite{David Goldberg, What Every
4601: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
1.17 anton 4602: Computing Surveys 23(1):5@minus{}48, March 1991}
1.47 crook 4603: (@uref{http://www.validgh.com/goldberg/paper.ps}).
1.1 anton 4604:
1.44 crook 4605:
1.21 crook 4606: doc-d>f
4607: doc-f>d
1.1 anton 4608: doc-f+
4609: doc-f-
4610: doc-f*
4611: doc-f/
4612: doc-fnegate
4613: doc-fabs
4614: doc-fmax
4615: doc-fmin
4616: doc-floor
4617: doc-fround
4618: doc-f**
4619: doc-fsqrt
4620: doc-fexp
4621: doc-fexpm1
4622: doc-fln
4623: doc-flnp1
4624: doc-flog
4625: doc-falog
1.32 anton 4626: doc-f2*
4627: doc-f2/
4628: doc-1/f
4629: doc-precision
4630: doc-set-precision
4631:
4632: @cindex angles in trigonometric operations
4633: @cindex trigonometric operations
4634: Angles in floating point operations are given in radians (a full circle
4635: has 2 pi radians).
4636:
1.1 anton 4637: doc-fsin
4638: doc-fcos
4639: doc-fsincos
4640: doc-ftan
4641: doc-fasin
4642: doc-facos
4643: doc-fatan
4644: doc-fatan2
4645: doc-fsinh
4646: doc-fcosh
4647: doc-ftanh
4648: doc-fasinh
4649: doc-facosh
4650: doc-fatanh
1.21 crook 4651: doc-pi
1.28 crook 4652:
1.32 anton 4653: @cindex equality of floats
4654: @cindex floating-point comparisons
1.31 anton 4655: One particular problem with floating-point arithmetic is that comparison
4656: for equality often fails when you would expect it to succeed. For this
4657: reason approximate equality is often preferred (but you still have to
4658: know what you are doing). The comparison words are:
4659:
4660: doc-f~rel
4661: doc-f~abs
4662: doc-f=
4663: doc-f~
4664: doc-f<>
4665:
4666: doc-f<
4667: doc-f<=
4668: doc-f>
4669: doc-f>=
4670:
1.21 crook 4671: doc-f0<
1.28 crook 4672: doc-f0<=
4673: doc-f0<>
1.21 crook 4674: doc-f0=
1.28 crook 4675: doc-f0>
4676: doc-f0>=
4677:
1.1 anton 4678:
4679: @node Stack Manipulation, Memory, Arithmetic, Words
4680: @section Stack Manipulation
4681: @cindex stack manipulation words
4682:
4683: @cindex floating-point stack in the standard
1.21 crook 4684: Gforth maintains a number of separate stacks:
4685:
1.29 crook 4686: @cindex data stack
4687: @cindex parameter stack
1.21 crook 4688: @itemize @bullet
4689: @item
1.29 crook 4690: A data stack (also known as the @dfn{parameter stack}) -- for
4691: characters, cells, addresses, and double cells.
1.21 crook 4692:
1.29 crook 4693: @cindex floating-point stack
1.21 crook 4694: @item
1.44 crook 4695: A floating point stack -- for holding floating point (FP) numbers.
1.21 crook 4696:
1.29 crook 4697: @cindex return stack
1.21 crook 4698: @item
1.44 crook 4699: A return stack -- for holding the return addresses of colon
1.32 anton 4700: definitions and other (non-FP) data.
1.21 crook 4701:
1.29 crook 4702: @cindex locals stack
1.21 crook 4703: @item
1.44 crook 4704: A locals stack -- for holding local variables.
1.21 crook 4705: @end itemize
4706:
1.1 anton 4707: @menu
4708: * Data stack::
4709: * Floating point stack::
4710: * Return stack::
4711: * Locals stack::
4712: * Stack pointer manipulation::
4713: @end menu
4714:
4715: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4716: @subsection Data stack
4717: @cindex data stack manipulation words
4718: @cindex stack manipulations words, data stack
4719:
1.44 crook 4720:
1.1 anton 4721: doc-drop
4722: doc-nip
4723: doc-dup
4724: doc-over
4725: doc-tuck
4726: doc-swap
1.21 crook 4727: doc-pick
1.1 anton 4728: doc-rot
4729: doc--rot
4730: doc-?dup
4731: doc-roll
4732: doc-2drop
4733: doc-2nip
4734: doc-2dup
4735: doc-2over
4736: doc-2tuck
4737: doc-2swap
4738: doc-2rot
4739:
1.44 crook 4740:
1.1 anton 4741: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4742: @subsection Floating point stack
4743: @cindex floating-point stack manipulation words
4744: @cindex stack manipulation words, floating-point stack
4745:
1.32 anton 4746: Whilst every sane Forth has a separate floating-point stack, it is not
4747: strictly required; an ANS Forth system could theoretically keep
4748: floating-point numbers on the data stack. As an additional difficulty,
4749: you don't know how many cells a floating-point number takes. It is
4750: reportedly possible to write words in a way that they work also for a
4751: unified stack model, but we do not recommend trying it. Instead, just
4752: say that your program has an environmental dependency on a separate
4753: floating-point stack.
4754:
4755: doc-floating-stack
4756:
1.1 anton 4757: doc-fdrop
4758: doc-fnip
4759: doc-fdup
4760: doc-fover
4761: doc-ftuck
4762: doc-fswap
1.21 crook 4763: doc-fpick
1.1 anton 4764: doc-frot
4765:
1.44 crook 4766:
1.1 anton 4767: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4768: @subsection Return stack
4769: @cindex return stack manipulation words
4770: @cindex stack manipulation words, return stack
4771:
1.32 anton 4772: @cindex return stack and locals
4773: @cindex locals and return stack
4774: A Forth system is allowed to keep local variables on the
4775: return stack. This is reasonable, as local variables usually eliminate
4776: the need to use the return stack explicitly. So, if you want to produce
4777: a standard compliant program and you are using local variables in a
4778: word, forget about return stack manipulations in that word (refer to the
4779: standard document for the exact rules).
4780:
1.1 anton 4781: doc->r
4782: doc-r>
4783: doc-r@
4784: doc-rdrop
4785: doc-2>r
4786: doc-2r>
4787: doc-2r@
4788: doc-2rdrop
4789:
1.44 crook 4790:
1.1 anton 4791: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4792: @subsection Locals stack
4793:
1.47 crook 4794: Gforth uses an extra locals stack. It is described, along with the
4795: reasons for its existence, in @ref{Implementation,Implementation of locals}.
1.21 crook 4796:
1.1 anton 4797: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4798: @subsection Stack pointer manipulation
4799: @cindex stack pointer manipulation words
4800:
1.44 crook 4801: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21 crook 4802: doc-sp0
1.1 anton 4803: doc-sp@
4804: doc-sp!
1.21 crook 4805: doc-fp0
1.1 anton 4806: doc-fp@
4807: doc-fp!
1.21 crook 4808: doc-rp0
1.1 anton 4809: doc-rp@
4810: doc-rp!
1.21 crook 4811: doc-lp0
1.1 anton 4812: doc-lp@
4813: doc-lp!
4814:
1.44 crook 4815:
1.1 anton 4816: @node Memory, Control Structures, Stack Manipulation, Words
4817: @section Memory
1.26 crook 4818: @cindex memory words
1.1 anton 4819:
1.32 anton 4820: @menu
4821: * Memory model::
4822: * Dictionary allocation::
4823: * Heap Allocation::
4824: * Memory Access::
4825: * Address arithmetic::
4826: * Memory Blocks::
4827: @end menu
4828:
4829: @node Memory model, Dictionary allocation, Memory, Memory
4830: @subsection ANS Forth and Gforth memory models
4831:
4832: @c The ANS Forth description is a mess (e.g., is the heap part of
4833: @c the dictionary?), so let's not stick to closely with it.
4834:
4835: ANS Forth considers a Forth system as consisting of several memories, of
4836: which only @dfn{data space} is managed and accessible with the memory
4837: words. Memory not necessarily in data space includes the stacks, the
4838: code (called code space) and the headers (called name space). In Gforth
4839: everything is in data space, but the code for the primitives is usually
4840: read-only.
4841:
4842: Data space is divided into a number of areas: The (data space portion of
4843: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4844: refer to the search data structure embodied in word lists and headers,
4845: because it is used for looking up names, just as you would in a
4846: conventional dictionary.}, the heap, and a number of system-allocated
4847: buffers.
4848:
4849: In ANS Forth data space is also divided into contiguous regions. You
4850: can only use address arithmetic within a contiguous region, not between
4851: them. Usually each allocation gives you one contiguous region, but the
1.33 anton 4852: dictionary allocation words have additional rules (@pxref{Dictionary
1.32 anton 4853: allocation}).
4854:
4855: Gforth provides one big address space, and address arithmetic can be
4856: performed between any addresses. However, in the dictionary headers or
4857: code are interleaved with data, so almost the only contiguous data space
4858: regions there are those described by ANS Forth as contiguous; but you
4859: can be sure that the dictionary is allocated towards increasing
4860: addresses even between contiguous regions. The memory order of
4861: allocations in the heap is platform-dependent (and possibly different
4862: from one run to the next).
4863:
4864: @subsubsection ANS Forth dictionary details
4865:
4866: This section is just informative, you can skip it if you are in a hurry.
1.27 crook 4867:
1.29 crook 4868: When you create a colon definition, the text interpreter compiles the
1.32 anton 4869: code for the definition into the code space and compiles the name
4870: of the definition into the header space, together with other
1.27 crook 4871: information about the definition (such as its execution token).
4872:
1.44 crook 4873: When you create a variable, the execution of @code{Variable} will
1.32 anton 4874: compile some code, assign one cell in data space, and compile the name
4875: of the variable into the header space.
1.27 crook 4876:
4877: @cindex memory regions - relationship between them
4878: ANS Forth does not specify the relationship between the three memory
4879: regions, and specifies that a Standard program must not access code or
4880: data space directly -- it may only access data space directly. In
4881: addition, the Standard defines what relationships you may and may not
4882: rely on when allocating regions in data space. These constraints are
4883: simply a reflection of the many diverse techniques that are used to
4884: implement Forth systems; understanding and following the requirements of
4885: the Standard allows you to write portable programs -- programs that run
4886: in the same way on any of these diverse systems. Another way of looking
4887: at this is to say that ANS Forth was designed to permit compliant Forth
4888: systems to be implemented in many diverse ways.
4889:
4890: @cindex memory regions - how they are assigned
1.29 crook 4891: Here are some examples of ways in which name, code and data spaces
4892: might be assigned in different Forth implementations:
1.27 crook 4893:
4894: @itemize @bullet
4895: @item
4896: For a Forth system that runs from RAM under a general-purpose operating
4897: system, it can be convenient to interleave name, code and data spaces in
4898: a single contiguous memory region. This organisation can be
4899: memory-efficient (for example, because the relationship between the name
1.32 anton 4900: dictionary entry and the associated code space entry can be
1.27 crook 4901: implicit, rather than requiring an explicit memory pointer to reference
1.32 anton 4902: from the header space and the code space). This is the
1.27 crook 4903: organisation used by Gforth, as this example@footnote{The addresses
4904: in the example have been truncated to fit it onto the page, and the
4905: addresses and data shown will not match the output from your system} shows:
4906: @example
4907: hex
4908: variable fred 123456 fred !
4909: variable jim abcd jim !
4910: : foo + / - ;
4911: ' fred 10 - 50 dump
4912: ..80: 5C 46 0E 40 84 66 72 65 - 64 20 20 20 20 20 20 20 \F.@.fred
1.50 anton 4913: ..90: D0 9B 04 08 00 00 00 00 - 56 34 12 00 80 46 0E 40 ........V4...F.@@
1.27 crook 4914: ..A0: 83 6A 69 6D 20 20 20 20 - D0 9B 04 08 00 00 00 00 .jim ........
4915: ..B0: CD AB 00 00 9C 46 0E 40 - 83 66 6F 6F 20 20 20 20 .....F.@.foo
4916: ..C0: 80 9B 04 08 00 00 00 00 - E4 2E 05 08 0C 2F 05 08 ............./..
4917: @end example
4918:
4919: @item
4920: For a high-performance system running on a modern RISC processor with a
4921: modified Harvard architecture (one that has a unified main memory but
4922: separate instruction and data caches), it is desirable to separate
4923: processor instructions from processor data. This encourages a high cache
1.32 anton 4924: density and therefore a high cache hit rate. The Forth code space
1.27 crook 4925: is not necessarily made up entirely of processor instructions; its
4926: nature is dependent upon the Forth implementation.
4927:
4928: @item
4929: A Forth compiler that runs on a segmented 8086 processor could be
4930: designed to interleave the name, code and data spaces within a single
4931: 64Kbyte segment. A more common implementation choice is to use a
4932: separate 64Kbyte segment for each region, which provides more memory
4933: overall but provides an address map in which only the data space is
4934: accessible.
4935:
4936: @item
4937: Microprocessors exist that run Forth (or many of the primitives required
4938: to implement the Forth virtual machine efficiently) directly. On these
4939: processors, the relationship between name, code and data spaces may be
1.32 anton 4940: imposed as a side-effect of the architecture of the processor.
1.27 crook 4941:
4942: @item
4943: A Forth compiler that executes from ROM on an embedded system needs its
4944: data space separated from the name and code spaces so that the data
4945: space can be mapped to a RAM area.
4946:
4947: @item
4948: A Forth compiler that runs on an embedded system may have a requirement
4949: for a small memory footprint. On such a system it can be useful to
1.32 anton 4950: separate the header space from the data and code spaces; once the
4951: application has been compiled, the header space is no longer
1.27 crook 4952: required@footnote{more strictly speaking, most applications can be
1.32 anton 4953: designed so that this is the case}. The header space can be deleted
1.29 crook 4954: entirely, or could be stored in memory on a remote @i{host} system for
1.27 crook 4955: debug and development purposes. In the latter case, the compiler running
1.29 crook 4956: on the @i{target} system could implement a protocol across a
1.32 anton 4957: communication link that would allow it to interrogate the header space.
1.27 crook 4958: @end itemize
4959:
1.32 anton 4960: @node Dictionary allocation, Heap Allocation, Memory model, Memory
4961: @subsection Dictionary allocation
1.27 crook 4962: @cindex reserving data space
4963: @cindex data space - reserving some
4964:
1.32 anton 4965: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
4966: you want to deallocate X, you also deallocate everything
4967: allocated after X.
4968:
4969: The allocations using the words below are contiguous and grow the region
4970: towards increasing addresses. Other words that allocate dictionary
4971: memory of any kind (i.e., defining words including @code{:noname}) end
4972: the contiguous region and start a new one.
4973:
4974: In ANS Forth only @code{create}d words are guaranteed to produce an
4975: address that is the start of the following contiguous region. In
4976: particular, the cell allocated by @code{variable} is not guaranteed to
4977: be contiguous with following @code{allot}ed memory.
4978:
4979: You can deallocate memory by using @code{allot} with a negative argument
4980: (with some restrictions, see @code{allot}). For larger deallocations use
4981: @code{marker}.
1.27 crook 4982:
1.29 crook 4983:
1.27 crook 4984: doc-here
4985: doc-unused
4986: doc-allot
4987: doc-c,
1.29 crook 4988: doc-f,
1.27 crook 4989: doc-,
4990: doc-2,
1.29 crook 4991: @cindex user space
4992: doc-udp
4993: doc-uallot
1.27 crook 4994:
1.32 anton 4995: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
4996: course you should allocate memory in an aligned way, too. I.e., before
4997: allocating allocating a cell, @code{here} must be cell-aligned, etc.
4998: The words below align @code{here} if it is not already. Basically it is
4999: only already aligned for a type, if the last allocation was a multiple
5000: of the size of this type and if @code{here} was aligned for this type
5001: before.
5002:
5003: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5004: ANS Forth (@code{maxalign}ed in Gforth).
5005:
5006: doc-align
5007: doc-falign
5008: doc-sfalign
5009: doc-dfalign
5010: doc-maxalign
5011: doc-cfalign
5012:
5013:
5014: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5015: @subsection Heap allocation
5016: @cindex heap allocation
5017: @cindex dynamic allocation of memory
5018: @cindex memory-allocation word set
5019:
5020: Heap allocation supports deallocation of allocated memory in any
5021: order. Dictionary allocation is not affected by it (i.e., it does not
5022: end a contiguous region). In Gforth, these words are implemented using
5023: the standard C library calls malloc(), free() and resize().
5024:
5025: doc-allocate
5026: doc-free
5027: doc-resize
5028:
1.27 crook 5029:
1.32 anton 5030: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1 anton 5031: @subsection Memory Access
5032: @cindex memory access words
5033:
1.44 crook 5034:
1.1 anton 5035: doc-@
5036: doc-!
5037: doc-+!
5038: doc-c@
5039: doc-c!
5040: doc-2@
5041: doc-2!
5042: doc-f@
5043: doc-f!
5044: doc-sf@
5045: doc-sf!
5046: doc-df@
5047: doc-df!
5048:
1.32 anton 5049: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5050: @subsection Address arithmetic
1.1 anton 5051: @cindex address arithmetic words
5052:
1.32 anton 5053: Address arithmetic is the foundation on which data structures like
5054: arrays, records (@pxref{Structures}) and objects (@pxref{Object-oriented
5055: Forth}) are built.
5056:
1.1 anton 5057: ANS Forth does not specify the sizes of the data types. Instead, it
5058: offers a number of words for computing sizes and doing address
1.29 crook 5059: arithmetic. Address arithmetic is performed in terms of address units
5060: (aus); on most systems the address unit is one byte. Note that a
5061: character may have more than one au, so @code{chars} is no noop (on
5062: systems where it is a noop, it compiles to nothing).
1.1 anton 5063:
5064: @cindex alignment of addresses for types
5065: ANS Forth also defines words for aligning addresses for specific
5066: types. Many computers require that accesses to specific data types
5067: must only occur at specific addresses; e.g., that cells may only be
5068: accessed at addresses divisible by 4. Even if a machine allows unaligned
5069: accesses, it can usually perform aligned accesses faster.
5070:
5071: For the performance-conscious: alignment operations are usually only
5072: necessary during the definition of a data structure, not during the
5073: (more frequent) accesses to it.
5074:
5075: ANS Forth defines no words for character-aligning addresses. This is not
5076: an oversight, but reflects the fact that addresses that are not
5077: char-aligned have no use in the standard and therefore will not be
5078: created.
5079:
5080: @cindex @code{CREATE} and alignment
1.29 crook 5081: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1 anton 5082: are cell-aligned; in addition, Gforth guarantees that these addresses
5083: are aligned for all purposes.
5084:
1.26 crook 5085: Note that the ANS Forth word @code{char} has nothing to do with address
5086: arithmetic.
1.1 anton 5087:
1.44 crook 5088:
1.1 anton 5089: doc-chars
5090: doc-char+
5091: doc-cells
5092: doc-cell+
5093: doc-cell
5094: doc-aligned
5095: doc-floats
5096: doc-float+
5097: doc-float
5098: doc-faligned
5099: doc-sfloats
5100: doc-sfloat+
5101: doc-sfaligned
5102: doc-dfloats
5103: doc-dfloat+
5104: doc-dfaligned
5105: doc-maxaligned
5106: doc-cfaligned
5107: doc-address-unit-bits
5108:
1.44 crook 5109:
1.32 anton 5110: @node Memory Blocks, , Address arithmetic, Memory
1.1 anton 5111: @subsection Memory Blocks
5112: @cindex memory block words
1.27 crook 5113: @cindex character strings - moving and copying
5114:
1.49 anton 5115: Memory blocks often represent character strings; For ways of storing
5116: character strings in memory see @ref{String Formats}. For other
5117: string-processing words see @ref{Displaying characters and strings}.
1.1 anton 5118:
1.32 anton 5119: Some of these words work on address units. Others work on character
5120: units (increments of @code{CHAR}), and expect a @code{CHAR}-aligned
5121: address. Choose the correct operation depending upon your data type.
1.21 crook 5122:
5123: When copying characters between overlapping memory regions, choose
5124: carefully between @code{cmove} and @code{cmove>}.
5125:
1.29 crook 5126: You can only use any of these words @i{portably} to access data space.
1.21 crook 5127:
1.27 crook 5128: @comment TODO - think the naming of the arguments is wrong for move
1.29 crook 5129: @comment well, really it seems to be the Standard that's wrong; it
5130: @comment describes MOVE as a word that requires a CELL-aligned source
5131: @comment and destination address but a xtranfer count that need not
5132: @comment be a multiple of CELL.
1.44 crook 5133:
1.1 anton 5134: doc-move
5135: doc-erase
5136: doc-cmove
5137: doc-cmove>
5138: doc-fill
5139: doc-blank
1.21 crook 5140: doc-compare
5141: doc-search
1.27 crook 5142: doc--trailing
5143: doc-/string
5144:
1.44 crook 5145:
1.27 crook 5146: @comment TODO examples
5147:
1.1 anton 5148:
1.26 crook 5149: @node Control Structures, Defining Words, Memory, Words
1.1 anton 5150: @section Control Structures
5151: @cindex control structures
5152:
1.33 anton 5153: Control structures in Forth cannot be used interpretively, only in a
5154: colon definition@footnote{To be precise, they have no interpretation
5155: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5156: not like this limitation, but have not seen a satisfying way around it
5157: yet, although many schemes have been proposed.
1.1 anton 5158:
5159: @menu
1.33 anton 5160: * Selection:: IF ... ELSE ... ENDIF
5161: * Simple Loops:: BEGIN ...
1.29 crook 5162: * Counted Loops:: DO
5163: * Arbitrary control structures::
5164: * Calls and returns::
1.1 anton 5165: * Exception Handling::
5166: @end menu
5167:
5168: @node Selection, Simple Loops, Control Structures, Control Structures
5169: @subsection Selection
5170: @cindex selection control structures
5171: @cindex control structures for selection
5172:
1.33 anton 5173: @c what's the purpose of all these @i? Maybe we should define a macro
5174: @c so we can produce logical markup. - anton
5175:
1.44 crook 5176: @c nac-> When I started working on the manual, a mixture of @i and @var
5177: @c were used inconsistently in code examples and \Glossary entries. These
5178: @c two behave differently in info format so I decided to standardize on @i.
5179: @c Logical markup would be better but texi isn't really upto it, and
5180: @c texi2html just ignores macros.
1.47 crook 5181: @c nac02dec1999-> update: the latest texinfo release can spit out html
5182: @c and it handles macros, so we could do some logical markup. Unfortunately
5183: @c texinfo will not split html output, which would be a big pain if you
5184: @c wanted to put the document on the web, which would be nice.
1.44 crook 5185:
1.1 anton 5186: @cindex @code{IF} control structure
5187: @example
1.29 crook 5188: @i{flag}
1.1 anton 5189: IF
1.29 crook 5190: @i{code}
1.1 anton 5191: ENDIF
5192: @end example
1.21 crook 5193: @noindent
1.33 anton 5194:
1.44 crook 5195: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5196: with any bit set represents truth) @i{code} is executed.
1.33 anton 5197:
1.1 anton 5198: @example
1.29 crook 5199: @i{flag}
1.1 anton 5200: IF
1.29 crook 5201: @i{code1}
1.1 anton 5202: ELSE
1.29 crook 5203: @i{code2}
1.1 anton 5204: ENDIF
5205: @end example
5206:
1.44 crook 5207: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5208: executed.
1.33 anton 5209:
1.1 anton 5210: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5211: standard, and @code{ENDIF} is not, although it is quite popular. We
5212: recommend using @code{ENDIF}, because it is less confusing for people
5213: who also know other languages (and is not prone to reinforcing negative
5214: prejudices against Forth in these people). Adding @code{ENDIF} to a
5215: system that only supplies @code{THEN} is simple:
5216: @example
1.21 crook 5217: : ENDIF POSTPONE THEN ; immediate
1.1 anton 5218: @end example
5219:
5220: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5221: (adv.)} has the following meanings:
5222: @quotation
5223: ... 2b: following next after in order ... 3d: as a necessary consequence
5224: (if you were there, then you saw them).
5225: @end quotation
5226: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5227: and many other programming languages has the meaning 3d.]
5228:
1.21 crook 5229: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1 anton 5230: you can avoid using @code{?dup}. Using these alternatives is also more
1.26 crook 5231: efficient than using @code{?dup}. Definitions in ANS Forth
1.1 anton 5232: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5233: @file{compat/control.fs}.
5234:
5235: @cindex @code{CASE} control structure
5236: @example
1.29 crook 5237: @i{n}
1.1 anton 5238: CASE
1.29 crook 5239: @i{n1} OF @i{code1} ENDOF
5240: @i{n2} OF @i{code2} ENDOF
1.1 anton 5241: @dots{}
5242: ENDCASE
5243: @end example
5244:
1.29 crook 5245: Executes the first @i{codei}, where the @i{ni} is equal to
5246: @i{n}. A default case can be added by simply writing the code after
5247: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
1.1 anton 5248: but must not consume it.
5249:
5250: @node Simple Loops, Counted Loops, Selection, Control Structures
5251: @subsection Simple Loops
5252: @cindex simple loops
5253: @cindex loops without count
5254:
5255: @cindex @code{WHILE} loop
5256: @example
5257: BEGIN
1.29 crook 5258: @i{code1}
5259: @i{flag}
1.1 anton 5260: WHILE
1.29 crook 5261: @i{code2}
1.1 anton 5262: REPEAT
5263: @end example
5264:
1.29 crook 5265: @i{code1} is executed and @i{flag} is computed. If it is true,
5266: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1 anton 5267: false, execution continues after the @code{REPEAT}.
5268:
5269: @cindex @code{UNTIL} loop
5270: @example
5271: BEGIN
1.29 crook 5272: @i{code}
5273: @i{flag}
1.1 anton 5274: UNTIL
5275: @end example
5276:
1.29 crook 5277: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1 anton 5278:
5279: @cindex endless loop
5280: @cindex loops, endless
5281: @example
5282: BEGIN
1.29 crook 5283: @i{code}
1.1 anton 5284: AGAIN
5285: @end example
5286:
5287: This is an endless loop.
5288:
5289: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5290: @subsection Counted Loops
5291: @cindex counted loops
5292: @cindex loops, counted
5293: @cindex @code{DO} loops
5294:
5295: The basic counted loop is:
5296: @example
1.29 crook 5297: @i{limit} @i{start}
1.1 anton 5298: ?DO
1.29 crook 5299: @i{body}
1.1 anton 5300: LOOP
5301: @end example
5302:
1.29 crook 5303: This performs one iteration for every integer, starting from @i{start}
5304: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21 crook 5305: accessed with @code{i}. For example, the loop:
1.1 anton 5306: @example
5307: 10 0 ?DO
5308: i .
5309: LOOP
5310: @end example
1.21 crook 5311: @noindent
5312: prints @code{0 1 2 3 4 5 6 7 8 9}
5313:
1.1 anton 5314: The index of the innermost loop can be accessed with @code{i}, the index
5315: of the next loop with @code{j}, and the index of the third loop with
5316: @code{k}.
5317:
1.44 crook 5318:
1.1 anton 5319: doc-i
5320: doc-j
5321: doc-k
5322:
1.44 crook 5323:
1.1 anton 5324: The loop control data are kept on the return stack, so there are some
1.21 crook 5325: restrictions on mixing return stack accesses and counted loop words. In
5326: particuler, if you put values on the return stack outside the loop, you
5327: cannot read them inside the loop@footnote{well, not in a way that is
5328: portable.}. If you put values on the return stack within a loop, you
5329: have to remove them before the end of the loop and before accessing the
5330: index of the loop.
1.1 anton 5331:
5332: There are several variations on the counted loop:
5333:
1.21 crook 5334: @itemize @bullet
5335: @item
5336: @code{LEAVE} leaves the innermost counted loop immediately; execution
5337: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5338:
5339: @example
5340: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5341: @end example
5342: prints @code{0 1 2 3}
5343:
1.1 anton 5344:
1.21 crook 5345: @item
5346: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5347: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5348: return stack so @code{EXIT} can get to its return address. For example:
5349:
5350: @example
5351: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5352: @end example
5353: prints @code{0 1 2 3}
5354:
5355:
5356: @item
1.29 crook 5357: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1 anton 5358: (and @code{LOOP} iterates until they become equal by wrap-around
5359: arithmetic). This behaviour is usually not what you want. Therefore,
5360: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29 crook 5361: @code{?DO}), which do not enter the loop if @i{start} is greater than
5362: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1 anton 5363: unsigned loop parameters.
5364:
1.21 crook 5365: @item
5366: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5367: the loop, independent of the loop parameters. Do not use @code{DO}, even
5368: if you know that the loop is entered in any case. Such knowledge tends
5369: to become invalid during maintenance of a program, and then the
5370: @code{DO} will make trouble.
5371:
5372: @item
1.29 crook 5373: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5374: index by @i{n} instead of by 1. The loop is terminated when the border
5375: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1 anton 5376:
1.21 crook 5377: @example
5378: 4 0 +DO i . 2 +LOOP
5379: @end example
5380: @noindent
5381: prints @code{0 2}
5382:
5383: @example
5384: 4 1 +DO i . 2 +LOOP
5385: @end example
5386: @noindent
5387: prints @code{1 3}
1.1 anton 5388:
5389:
5390: @cindex negative increment for counted loops
5391: @cindex counted loops with negative increment
1.29 crook 5392: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1 anton 5393:
1.21 crook 5394: @example
5395: -1 0 ?DO i . -1 +LOOP
5396: @end example
5397: @noindent
5398: prints @code{0 -1}
1.1 anton 5399:
1.21 crook 5400: @example
5401: 0 0 ?DO i . -1 +LOOP
5402: @end example
5403: prints nothing.
1.1 anton 5404:
1.29 crook 5405: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5406: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5407: index by @i{u} each iteration. The loop is terminated when the border
5408: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1 anton 5409: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5410:
1.21 crook 5411: @example
5412: -2 0 -DO i . 1 -LOOP
5413: @end example
5414: @noindent
5415: prints @code{0 -1}
1.1 anton 5416:
1.21 crook 5417: @example
5418: -1 0 -DO i . 1 -LOOP
5419: @end example
5420: @noindent
5421: prints @code{0}
5422:
5423: @example
5424: 0 0 -DO i . 1 -LOOP
5425: @end example
5426: @noindent
5427: prints nothing.
1.1 anton 5428:
1.21 crook 5429: @end itemize
1.1 anton 5430:
5431: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26 crook 5432: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5433: for these words that uses only standard words is provided in
5434: @file{compat/loops.fs}.
1.1 anton 5435:
5436:
5437: @cindex @code{FOR} loops
1.26 crook 5438: Another counted loop is:
1.1 anton 5439: @example
1.29 crook 5440: @i{n}
1.1 anton 5441: FOR
1.29 crook 5442: @i{body}
1.1 anton 5443: NEXT
5444: @end example
5445: This is the preferred loop of native code compiler writers who are too
1.26 crook 5446: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29 crook 5447: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5448: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26 crook 5449: Forth systems may behave differently, even if they support @code{FOR}
5450: loops. To avoid problems, don't use @code{FOR} loops.
1.1 anton 5451:
5452: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5453: @subsection Arbitrary control structures
5454: @cindex control structures, user-defined
5455:
5456: @cindex control-flow stack
5457: ANS Forth permits and supports using control structures in a non-nested
5458: way. Information about incomplete control structures is stored on the
5459: control-flow stack. This stack may be implemented on the Forth data
5460: stack, and this is what we have done in Gforth.
5461:
5462: @cindex @code{orig}, control-flow stack item
5463: @cindex @code{dest}, control-flow stack item
5464: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5465: entry represents a backward branch target. A few words are the basis for
5466: building any control structure possible (except control structures that
5467: need storage, like calls, coroutines, and backtracking).
5468:
1.44 crook 5469:
1.1 anton 5470: doc-if
5471: doc-ahead
5472: doc-then
5473: doc-begin
5474: doc-until
5475: doc-again
5476: doc-cs-pick
5477: doc-cs-roll
5478:
1.44 crook 5479:
1.21 crook 5480: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5481: manipulate the control-flow stack in a portable way. Without them, you
5482: would need to know how many stack items are occupied by a control-flow
5483: entry (many systems use one cell. In Gforth they currently take three,
5484: but this may change in the future).
5485:
1.1 anton 5486: Some standard control structure words are built from these words:
5487:
1.44 crook 5488:
1.1 anton 5489: doc-else
5490: doc-while
5491: doc-repeat
5492:
1.44 crook 5493:
5494: @noindent
1.1 anton 5495: Gforth adds some more control-structure words:
5496:
1.44 crook 5497:
1.1 anton 5498: doc-endif
5499: doc-?dup-if
5500: doc-?dup-0=-if
5501:
1.44 crook 5502:
5503: @noindent
1.1 anton 5504: Counted loop words constitute a separate group of words:
5505:
1.44 crook 5506:
1.1 anton 5507: doc-?do
5508: doc-+do
5509: doc-u+do
5510: doc--do
5511: doc-u-do
5512: doc-do
5513: doc-for
5514: doc-loop
5515: doc-+loop
5516: doc--loop
5517: doc-next
5518: doc-leave
5519: doc-?leave
5520: doc-unloop
5521: doc-done
5522:
1.44 crook 5523:
1.21 crook 5524: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5525: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1 anton 5526: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5527: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5528: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5529: resolved (by using one of the loop-ending words or @code{DONE}).
5530:
1.44 crook 5531: @noindent
1.26 crook 5532: Another group of control structure words are:
1.1 anton 5533:
1.44 crook 5534:
1.1 anton 5535: doc-case
5536: doc-endcase
5537: doc-of
5538: doc-endof
5539:
1.44 crook 5540:
1.21 crook 5541: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5542: @code{CS-ROLL}.
1.1 anton 5543:
5544: @subsubsection Programming Style
1.47 crook 5545: @cindex control structures programming style
5546: @cindex programming style, arbitrary control structures
1.1 anton 5547:
5548: In order to ensure readability we recommend that you do not create
5549: arbitrary control structures directly, but define new control structure
5550: words for the control structure you want and use these words in your
1.26 crook 5551: program. For example, instead of writing:
1.1 anton 5552:
5553: @example
1.26 crook 5554: BEGIN
1.1 anton 5555: ...
1.26 crook 5556: IF [ 1 CS-ROLL ]
1.1 anton 5557: ...
1.26 crook 5558: AGAIN THEN
1.1 anton 5559: @end example
5560:
1.21 crook 5561: @noindent
1.1 anton 5562: we recommend defining control structure words, e.g.,
5563:
5564: @example
1.26 crook 5565: : WHILE ( DEST -- ORIG DEST )
5566: POSTPONE IF
5567: 1 CS-ROLL ; immediate
5568:
5569: : REPEAT ( orig dest -- )
5570: POSTPONE AGAIN
5571: POSTPONE THEN ; immediate
1.1 anton 5572: @end example
5573:
1.21 crook 5574: @noindent
1.1 anton 5575: and then using these to create the control structure:
5576:
5577: @example
1.26 crook 5578: BEGIN
1.1 anton 5579: ...
1.26 crook 5580: WHILE
1.1 anton 5581: ...
1.26 crook 5582: REPEAT
1.1 anton 5583: @end example
5584:
5585: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5586: @code{WHILE} are predefined, so in this example it would not be
5587: necessary to define them.
5588:
5589: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5590: @subsection Calls and returns
5591: @cindex calling a definition
5592: @cindex returning from a definition
5593:
1.3 anton 5594: @cindex recursive definitions
5595: A definition can be called simply be writing the name of the definition
1.26 crook 5596: to be called. Normally a definition is invisible during its own
1.3 anton 5597: definition. If you want to write a directly recursive definition, you
1.26 crook 5598: can use @code{recursive} to make the current definition visible, or
5599: @code{recurse} to call the current definition directly.
1.3 anton 5600:
1.44 crook 5601:
1.3 anton 5602: doc-recursive
5603: doc-recurse
5604:
1.44 crook 5605:
1.21 crook 5606: @comment TODO add example of the two recursion methods
1.12 anton 5607: @quotation
5608: @progstyle
5609: I prefer using @code{recursive} to @code{recurse}, because calling the
5610: definition by name is more descriptive (if the name is well-chosen) than
5611: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5612: implementation, it is much better to read (and think) ``now sort the
5613: partitions'' than to read ``now do a recursive call''.
5614: @end quotation
1.3 anton 5615:
1.29 crook 5616: For mutual recursion, use @code{Defer}red words, like this:
1.3 anton 5617:
5618: @example
1.28 crook 5619: Defer foo
1.3 anton 5620:
5621: : bar ( ... -- ... )
5622: ... foo ... ;
5623:
5624: :noname ( ... -- ... )
5625: ... bar ... ;
5626: IS foo
5627: @end example
5628:
1.44 crook 5629: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33 anton 5630:
1.26 crook 5631: The current definition returns control to the calling definition when
1.33 anton 5632: the end of the definition is reached or @code{EXIT} is encountered.
1.1 anton 5633:
5634: doc-exit
5635: doc-;s
5636:
1.44 crook 5637:
1.1 anton 5638: @node Exception Handling, , Calls and returns, Control Structures
5639: @subsection Exception Handling
1.26 crook 5640: @cindex exceptions
1.1 anton 5641:
1.26 crook 5642: If your program detects a fatal error condition, the simplest action
5643: that it can take is to @code{quit}. This resets the return stack and
5644: restarts the text interpreter, but does not print any error message.
1.21 crook 5645:
1.26 crook 5646: The next stage in severity is to execute @code{abort}, which has the
5647: same effect as @code{quit}, with the addition that it resets the data
5648: stack.
1.1 anton 5649:
1.26 crook 5650: A slightly more sophisticated approach is use use @code{abort"}, which
5651: compiles a string to be used as an error message and does a conditional
5652: @code{abort} at run-time. For example:
1.1 anton 5653:
1.26 crook 5654: @example
1.30 anton 5655: @kbd{: checker abort" That flag was true" ." A false flag" ;@key{RET}} ok
5656: @kbd{0 checker@key{RET}} A false flag ok
5657: @kbd{1 checker@key{RET}}
1.26 crook 5658: :1: That flag was true
5659: 1 checker
5660: ^^^^^^^
5661: $400D1648 throw
5662: $400E4660
5663: @end example
1.1 anton 5664:
1.26 crook 5665: These simple techniques allow a program to react to a fatal error
5666: condition, but they are not exactly user-friendly. The ANS Forth
5667: Exception word set provides the pair of words @code{throw} and
5668: @code{catch}, which can be used to provide sophisticated error-handling.
1.1 anton 5669:
1.26 crook 5670: @code{catch} has a similar behaviour to @code{execute}, in that it takes
1.29 crook 5671: an @i{xt} as a parameter and starts execution of the xt. However,
1.26 crook 5672: before passing control to the xt, @code{catch} pushes an
1.29 crook 5673: @dfn{exception frame} onto the @dfn{exception stack}. This exception
1.26 crook 5674: frame is used to restore the system to a known state if a detected error
5675: occurs during the execution of the xt. A typical way to use @code{catch}
5676: would be:
1.1 anton 5677:
1.26 crook 5678: @example
5679: ... ['] foo catch IF ...
5680: @end example
1.1 anton 5681:
1.33 anton 5682: @c TOS is undefined. - anton
1.44 crook 5683:
5684: @c nac-> TODO -- I need to look at this example again.
5685:
1.26 crook 5686: Whilst @code{foo} executes, it can call other words to any level of
5687: nesting, as usual. If @code{foo} (and all the words that it calls)
1.33 anton 5688: execute successfully, control will ultimately pass to the word following
5689: the @code{catch}, and there will be a 0 at TOS. However, if any word
5690: detects an error, it can terminate the execution of @code{foo} by
5691: pushing a non-zero error code onto the stack and then performing a
5692: @code{throw}. The execution of @code{throw} will pass control to the
5693: word following the @code{catch}, but this time the TOS will hold the
5694: error code. Therefore, the @code{IF} in the example can be used to
5695: determine whether @code{foo} executed successfully.
1.1 anton 5696:
1.26 crook 5697: This simple example shows how you can use @code{throw} and @code{catch}
5698: to ``take over'' exception handling from the system:
1.1 anton 5699: @example
1.26 crook 5700: : my-div ['] / catch if ." DIVIDE ERROR" else ." OK.. " . then ;
1.1 anton 5701: @end example
5702:
1.26 crook 5703: The next example is more sophisticated and shows a multi-level
5704: @code{throw} and @code{catch}. To understand this example, start at the
5705: definition of @code{top-level} and work backwards:
5706:
1.1 anton 5707: @example
1.26 crook 5708: : lowest-level ( -- c )
5709: key dup 27 = if
1.44 crook 5710: 1 throw \ ESCAPE key pressed
1.26 crook 5711: else
1.44 crook 5712: ." lowest-level successful" CR
1.26 crook 5713: then
5714: ;
5715:
5716: : lower-level ( -- c )
5717: lowest-level
5718: \ at this level consider a CTRL-U to be a fatal error
5719: dup 21 = if \ CTRL-U
1.44 crook 5720: 2 throw
1.26 crook 5721: else
1.44 crook 5722: ." lower-level successful" CR
1.26 crook 5723: then
5724: ;
5725:
5726: : low-level ( -- c )
5727: ['] lower-level catch
5728: ?dup if
1.44 crook 5729: \ error occurred - do we recognise it?
5730: dup 1 = if
5731: \ ESCAPE key pressed.. pretend it was an E
5732: [char] E
5733: else throw \ propogate the error upwards
5734: then
1.26 crook 5735: then
5736: ." low-level successfull" CR
5737: ;
5738:
5739: : top-level ( -- )
5740: CR ['] low-level catch \ CATCH is used like EXECUTE
5741: ?dup if \ error occurred..
1.44 crook 5742: ." Error " . ." occurred - contact your supplier"
1.26 crook 5743: else
1.44 crook 5744: ." The '" emit ." ' key was pressed" CR
1.26 crook 5745: then
5746: ;
1.1 anton 5747: @end example
5748:
1.26 crook 5749: The ANS Forth document assigns @code{throw} codes thus:
1.1 anton 5750:
1.26 crook 5751: @itemize @bullet
5752: @item
5753: codes in the range -1 -- -255 are reserved to be assigned by the
5754: Standard. Assignments for codes in the range -1 -- -58 are currently
5755: documented in the Standard. In particular, @code{-1 throw} is equivalent
5756: to @code{abort} and @code{-2 throw} is equivalent to @code{abort"}.
5757: @item
5758: codes in the range -256 -- -4095 are reserved to be assigned by the system.
5759: @item
5760: all other codes may be assigned by programs.
5761: @end itemize
1.1 anton 5762:
1.26 crook 5763: Gforth provides the word @code{exception} as a mechanism for assigning
5764: system throw codes to applications. This allows multiple applications to
5765: co-exist in memory without any clash of @code{throw} codes. A definition
5766: of @code{exception} in ANS Forth is provided in
5767: @file{compat/exception.fs}.
1.1 anton 5768:
1.44 crook 5769:
1.26 crook 5770: doc-quit
5771: doc-abort
5772: doc-abort"
1.1 anton 5773:
1.26 crook 5774: doc-catch
1.29 crook 5775: doc-throw
5776: doc---exception-exception
5777:
5778:
1.44 crook 5779:
1.29 crook 5780: @c -------------------------------------------------------------
1.47 crook 5781: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29 crook 5782: @section Defining Words
5783: @cindex defining words
5784:
1.47 crook 5785: Defining words are used to extend Forth by creating new entries in the dictionary.
5786:
1.29 crook 5787: @menu
1.44 crook 5788: * CREATE::
5789: * Variables:: Variables and user variables
5790: * Constants::
5791: * Values:: Initialised variables
1.29 crook 5792: * Colon Definitions::
1.44 crook 5793: * Anonymous Definitions:: Definitions without names
1.29 crook 5794: * User-defined Defining Words::
1.44 crook 5795: * Deferred words:: Allow forward references
5796: * Aliases::
1.29 crook 5797: * Supplying names::
5798: @end menu
5799:
1.44 crook 5800: @node CREATE, Variables, Defining Words, Defining Words
5801: @subsection @code{CREATE}
1.29 crook 5802: @cindex simple defining words
5803: @cindex defining words, simple
5804:
5805: Defining words are used to create new entries in the dictionary. The
5806: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5807: this:
5808:
5809: @example
5810: CREATE new-word1
5811: @end example
5812:
5813: @code{CREATE} is a parsing word that generates a dictionary entry for
5814: @code{new-word1}. When @code{new-word1} is executed, all that it does is
5815: leave an address on the stack. The address represents the value of
5816: the data space pointer (@code{HERE}) at the time that @code{new-word1}
5817: was defined. Therefore, @code{CREATE} is a way of associating a name
5818: with the address of a region of memory.
5819:
1.34 anton 5820: doc-create
5821:
1.29 crook 5822: By extending this example to reserve some memory in data space, we end
5823: up with a @i{variable}. Here are two different ways to do it:
5824:
5825: @example
5826: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5827: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5828: @end example
5829:
5830: The variable can be examined and modified using @code{@@} (``fetch'') and
5831: @code{!} (``store'') like this:
5832:
5833: @example
5834: new-word2 @@ . \ get address, fetch from it and display
5835: 1234 new-word2 ! \ new value, get address, store to it
5836: @end example
5837:
1.44 crook 5838: @cindex arrays
5839: A similar mechanism can be used to create arrays. For example, an
5840: 80-character text input buffer:
1.29 crook 5841:
5842: @example
1.44 crook 5843: CREATE text-buf 80 chars allot
5844:
5845: text-buf 0 chars c@@ \ the 1st character (offset 0)
5846: text-buf 3 chars c@@ \ the 4th character (offset 3)
5847: @end example
1.29 crook 5848:
1.44 crook 5849: You can build arbitrarily complex data structures by allocating
1.49 anton 5850: appropriate areas of memory. For further discussions of this, and to
5851: learn about some Gforth tools that make it easier, see
5852: @xref{Structures}.
1.44 crook 5853:
5854:
5855: @node Variables, Constants, CREATE, Defining Words
5856: @subsection Variables
5857: @cindex variables
5858:
5859: The previous section showed how a sequence of commands could be used to
5860: generate a variable. As a final refinement, the whole code sequence can
5861: be wrapped up in a defining word (pre-empting the subject of the next
5862: section), making it easier to create new variables:
5863:
5864: @example
5865: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5866: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5867:
5868: myvariableX foo \ variable foo starts off with an unknown value
5869: myvariable0 joe \ whilst joe is initialised to 0
1.29 crook 5870:
5871: 45 3 * foo ! \ set foo to 135
5872: 1234 joe ! \ set joe to 1234
5873: 3 joe +! \ increment joe by 3.. to 1237
5874: @end example
5875:
5876: Not surprisingly, there is no need to define @code{myvariable}, since
1.44 crook 5877: Forth already has a definition @code{Variable}. ANS Forth does not
5878: require a @code{Variable} to be initialised when it is created (i.e., it
5879: behaves like @code{myvariableX}). In contrast, Gforth's @code{Variable}
5880: initialises the variable to 0 (i.e., it behaves exactly like
5881: @code{myvariable0}). Forth also provides @code{2Variable} and
1.47 crook 5882: @code{fvariable} for double and floating-point variables, respectively
5883: -- both are initialised to 0 in Gforth. If you use a @code{Variable} to
5884: store a boolean, you can use @code{on} and @code{off} to toggle its
5885: state.
1.29 crook 5886:
1.34 anton 5887: doc-variable
5888: doc-2variable
5889: doc-fvariable
5890:
1.29 crook 5891: @cindex user variables
5892: @cindex user space
5893: The defining word @code{User} behaves in the same way as @code{Variable}.
5894: The difference is that it reserves space in @i{user (data) space} rather
5895: than normal data space. In a Forth system that has a multi-tasker, each
5896: task has its own set of user variables.
5897:
1.34 anton 5898: doc-user
5899:
1.29 crook 5900: @comment TODO is that stuff about user variables strictly correct? Is it
5901: @comment just terminal tasks that have user variables?
5902: @comment should document tasker.fs (with some examples) elsewhere
5903: @comment in this manual, then expand on user space and user variables.
5904:
1.44 crook 5905:
5906: @node Constants, Values, Variables, Defining Words
5907: @subsection Constants
5908: @cindex constants
5909:
5910: @code{Constant} allows you to declare a fixed value and refer to it by
5911: name. For example:
1.29 crook 5912:
5913: @example
5914: 12 Constant INCHES-PER-FOOT
5915: 3E+08 fconstant SPEED-O-LIGHT
5916: @end example
5917:
5918: A @code{Variable} can be both read and written, so its run-time
5919: behaviour is to supply an address through which its current value can be
5920: manipulated. In contrast, the value of a @code{Constant} cannot be
5921: changed once it has been declared@footnote{Well, often it can be -- but
5922: not in a Standard, portable way. It's safer to use a @code{Value} (read
5923: on).} so it's not necessary to supply the address -- it is more
5924: efficient to return the value of the constant directly. That's exactly
5925: what happens; the run-time effect of a constant is to put its value on
1.49 anton 5926: the top of the stack (You can find one
5927: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29 crook 5928:
5929: Gforth also provides @code{2Constant} and @code{fconstant} for defining
5930: double and floating-point constants, respectively.
5931:
1.34 anton 5932: doc-constant
5933: doc-2constant
5934: doc-fconstant
5935:
5936: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44 crook 5937: @c nac-> How could that not be true in an ANS Forth? You can't define a
5938: @c constant, use it and then delete the definition of the constant..
5939: @c I agree that it's rather deep, but IMO it is an important difference
5940: @c relative to other programming languages.. often it's annoying: it
5941: @c certainly changes my programming style relative to C.
5942:
1.29 crook 5943: Constants in Forth behave differently from their equivalents in other
5944: programming languages. In other languages, a constant (such as an EQU in
5945: assembler or a #define in C) only exists at compile-time; in the
5946: executable program the constant has been translated into an absolute
5947: number and, unless you are using a symbolic debugger, it's impossible to
5948: know what abstract thing that number represents. In Forth a constant has
1.44 crook 5949: an entry in the header space and remains there after the code that uses
5950: it has been defined. In fact, it must remain in the dictionary since it
5951: has run-time duties to perform. For example:
1.29 crook 5952:
5953: @example
5954: 12 Constant INCHES-PER-FOOT
5955: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
5956: @end example
5957:
5958: @cindex in-lining of constants
5959: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
5960: associated with the constant @code{INCHES-PER-FOOT}. If you use
5961: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
5962: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
5963: attempt to optimise constants by in-lining them where they are used. You
5964: can force Gforth to in-line a constant like this:
5965:
5966: @example
5967: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
5968: @end example
5969:
5970: If you use @code{see} to decompile @i{this} version of
5971: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49 anton 5972: longer present. To understand how this works, read
5973: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29 crook 5974:
5975: In-lining constants in this way might improve execution time
5976: fractionally, and can ensure that a constant is now only referenced at
5977: compile-time. However, the definition of the constant still remains in
5978: the dictionary. Some Forth compilers provide a mechanism for controlling
5979: a second dictionary for holding transient words such that this second
5980: dictionary can be deleted later in order to recover memory
5981: space. However, there is no standard way of doing this.
5982:
5983:
1.44 crook 5984: @node Values, Colon Definitions, Constants, Defining Words
5985: @subsection Values
5986: @cindex values
1.34 anton 5987:
1.44 crook 5988: A @code{Value} is like a @code{Variable} but with two important
5989: differences:
1.29 crook 5990:
5991: @itemize @bullet
5992: @item
1.44 crook 5993: A @code{Value} is initialised when it is declared; like a
5994: @code{Constant} but unlike a @code{Variable}.
1.29 crook 5995: @item
1.44 crook 5996: A @code{Value} returns its value rather than its address when it is
5997: executed; i.e., it has the same run-time behaviour as @code{Constant}.
1.29 crook 5998: @end itemize
5999:
1.44 crook 6000: A @code{Value} needs an additional word, @code{TO} to allow its value to
6001: be changed. Here are some examples:
1.29 crook 6002:
6003: @example
1.44 crook 6004: 12 Value APPLES \ Define APPLES with an initial value of 12
6005: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6006: APPLES \ puts 34 on the top of the stack.
1.29 crook 6007: @end example
6008:
1.44 crook 6009: doc-value
6010: doc-to
1.29 crook 6011:
1.35 anton 6012:
1.44 crook 6013: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6014: @subsection Colon Definitions
6015: @cindex colon definitions
1.35 anton 6016:
6017: @example
1.44 crook 6018: : name ( ... -- ... )
6019: word1 word2 word3 ;
1.29 crook 6020: @end example
6021:
1.44 crook 6022: @noindent
6023: Creates a word called @code{name} that, upon execution, executes
6024: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29 crook 6025:
1.49 anton 6026: The explanation above is somewhat superficial. For simple examples of
6027: colon definitions see @ref{Your first definition}. For an in-depth
6028: discussion of some of the issues involved, see @xref{Interpretation and
6029: Compilation Semantics}.
1.29 crook 6030:
1.44 crook 6031: doc-:
6032: doc-;
1.1 anton 6033:
1.34 anton 6034:
1.44 crook 6035: @node Anonymous Definitions, User-defined Defining Words, Colon Definitions, Defining Words
6036: @subsection Anonymous Definitions
6037: @cindex colon definitions
6038: @cindex defining words without name
1.34 anton 6039:
1.44 crook 6040: Sometimes you want to define an @dfn{anonymous word}; a word without a
6041: name. You can do this with:
1.1 anton 6042:
1.44 crook 6043: doc-:noname
1.1 anton 6044:
1.44 crook 6045: This leaves the execution token for the word on the stack after the
6046: closing @code{;}. Here's an example in which a deferred word is
6047: initialised with an @code{xt} from an anonymous colon definition:
1.1 anton 6048:
1.29 crook 6049: @example
1.44 crook 6050: Defer deferred
6051: :noname ( ... -- ... )
6052: ... ;
6053: IS deferred
1.29 crook 6054: @end example
1.26 crook 6055:
1.44 crook 6056: @noindent
6057: Gforth provides an alternative way of doing this, using two separate
6058: words:
1.27 crook 6059:
1.44 crook 6060: doc-noname
6061: @cindex execution token of last defined word
6062: doc-lastxt
1.1 anton 6063:
1.44 crook 6064: @noindent
6065: The previous example can be rewritten using @code{noname} and
6066: @code{lastxt}:
1.1 anton 6067:
1.26 crook 6068: @example
1.44 crook 6069: Defer deferred
6070: noname : ( ... -- ... )
6071: ... ;
6072: lastxt IS deferred
1.26 crook 6073: @end example
1.1 anton 6074:
1.29 crook 6075: @noindent
1.44 crook 6076: @code{noname} works with any defining word, not just @code{:}.
6077:
6078: @code{lastxt} also works when the last word was not defined as
6079: @code{noname}. It also has the useful property that is is valid as soon
6080: as the header for a definition has been built. Thus:
6081:
6082: @example
6083: lastxt . : foo [ lastxt . ] ; ' foo .
6084: @end example
1.1 anton 6085:
1.44 crook 6086: @noindent
6087: prints 3 numbers; the last two are the same.
1.26 crook 6088:
1.1 anton 6089:
1.44 crook 6090: @node User-defined Defining Words, Deferred words, Anonymous Definitions, Defining Words
1.26 crook 6091: @subsection User-defined Defining Words
6092: @cindex user-defined defining words
6093: @cindex defining words, user-defined
1.1 anton 6094:
1.29 crook 6095: You can create a new defining word by wrapping defining-time code around
6096: an existing defining word and putting the sequence in a colon
6097: definition. For example, suppose that you have a word @code{stats} that
6098: gathers statistics about colon definitions given the @i{xt} of the
6099: definition, and you want every colon definition in your application to
6100: make a call to @code{stats}. You can define and use a new version of
6101: @code{:} like this:
6102:
6103: @example
6104: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6105: ... ; \ other code
6106:
6107: : my: : lastxt postpone literal ['] stats compile, ;
6108:
6109: my: foo + - ;
6110: @end example
6111:
6112: When @code{foo} is defined using @code{my:} these steps occur:
6113:
6114: @itemize @bullet
6115: @item
6116: @code{my:} is executed.
6117: @item
6118: The @code{:} within the definition (the one between @code{my:} and
6119: @code{lastxt}) is executed, and does just what it always does; it parses
6120: the input stream for a name, builds a dictionary header for the name
6121: @code{foo} and switches @code{state} from interpret to compile.
6122: @item
6123: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6124: being defined -- @code{foo} -- onto the stack.
6125: @item
6126: The code that was produced by @code{postpone literal} is executed; this
6127: causes the value on the stack to be compiled as a literal in the code
6128: area of @code{foo}.
6129: @item
6130: The code @code{['] stats} compiles a literal into the definition of
6131: @code{my:}. When @code{compile,} is executed, that literal -- the
6132: execution token for @code{stats} -- is layed down in the code area of
6133: @code{foo} , following the literal@footnote{Strictly speaking, the
6134: mechanism that @code{compile,} uses to convert an @i{xt} into something
6135: in the code area is implementation-dependent. A threaded implementation
6136: might spit out the execution token directly whilst another
6137: implementation might spit out a native code sequence.}.
6138: @item
6139: At this point, the execution of @code{my:} is complete, and control
6140: returns to the text interpreter. The text interpreter is in compile
6141: state, so subsequent text @code{+ -} is compiled into the definition of
6142: @code{foo} and the @code{;} terminates the definition as always.
6143: @end itemize
6144:
6145: You can use @code{see} to decompile a word that was defined using
6146: @code{my:} and see how it is different from a normal @code{:}
6147: definition. For example:
6148:
6149: @example
6150: : bar + - ; \ like foo but using : rather than my:
6151: see bar
6152: : bar
6153: + - ;
6154: see foo
6155: : foo
6156: 107645672 stats + - ;
6157:
6158: \ use ' stats . to show that 107645672 is the xt for stats
6159: @end example
6160:
6161: You can use techniques like this to make new defining words in terms of
6162: @i{any} existing defining word.
1.1 anton 6163:
6164:
1.29 crook 6165: @cindex defining defining words
1.26 crook 6166: @cindex @code{CREATE} ... @code{DOES>}
6167: If you want the words defined with your defining words to behave
6168: differently from words defined with standard defining words, you can
6169: write your defining word like this:
1.1 anton 6170:
6171: @example
1.26 crook 6172: : def-word ( "name" -- )
1.29 crook 6173: CREATE @i{code1}
1.26 crook 6174: DOES> ( ... -- ... )
1.29 crook 6175: @i{code2} ;
1.26 crook 6176:
6177: def-word name
1.1 anton 6178: @end example
6179:
1.29 crook 6180: @cindex child words
6181: This fragment defines a @dfn{defining word} @code{def-word} and then
6182: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6183: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6184: is not executed at this time. The word @code{name} is sometimes called a
6185: @dfn{child} of @code{def-word}.
6186:
6187: When you execute @code{name}, the address of the body of @code{name} is
6188: put on the data stack and @i{code2} is executed (the address of the body
6189: of @code{name} is the address @code{HERE} returns immediately after the
6190: @code{CREATE}).
6191:
6192: @cindex atavism in child words
1.33 anton 6193: You can use @code{def-word} to define a set of child words that behave
1.29 crook 6194: differently, though atavistically; they all have a common run-time
6195: behaviour determined by @i{code2}. Typically, the @i{code1} sequence
6196: builds a data area in the body of the child word. The structure of the
6197: data is common to all children of @code{def-word}, but the data values
6198: are specific -- and private -- to each child word. When a child word is
6199: executed, the address of its private data area is passed as a parameter
6200: on TOS to be used and manipulated@footnote{It is legitimate both to read
6201: and write to this data area.} by @i{code2}.
6202:
6203: The two fragments of code that make up the defining words act (are
6204: executed) at two completely separate times:
1.1 anton 6205:
1.29 crook 6206: @itemize @bullet
6207: @item
6208: At @i{define time}, the defining word executes @i{code1} to generate a
6209: child word
6210: @item
6211: At @i{child execution time}, when a child word is invoked, @i{code2}
6212: is executed, using parameters (data) that are private and specific to
6213: the child word.
6214: @end itemize
6215:
1.44 crook 6216: Another way of understanding the behaviour of @code{def-word} and
6217: @code{name} is to say that, if you make the following definitions:
1.33 anton 6218: @example
6219: : def-word1 ( "name" -- )
6220: CREATE @i{code1} ;
6221:
6222: : action1 ( ... -- ... )
6223: @i{code2} ;
6224:
6225: def-word1 name1
6226: @end example
6227:
1.44 crook 6228: @noindent
6229: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1 anton 6230:
1.29 crook 6231: The classic example is that you can define @code{CONSTANT} in this way:
1.26 crook 6232:
1.1 anton 6233: @example
1.29 crook 6234: : CONSTANT ( w "name" -- )
6235: CREATE ,
1.26 crook 6236: DOES> ( -- w )
6237: @@ ;
1.1 anton 6238: @end example
6239:
1.29 crook 6240: @comment There is a beautiful description of how this works and what
6241: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6242: @comment commentary on the Counting Fruits problem.
6243:
6244: When you create a constant with @code{5 CONSTANT five}, a set of
6245: define-time actions take place; first a new word @code{five} is created,
6246: then the value 5 is laid down in the body of @code{five} with
1.44 crook 6247: @code{,}. When @code{five} is executed, the address of the body is put on
1.29 crook 6248: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6249: no code of its own; it simply contains a data field and a pointer to the
6250: code that follows @code{DOES>} in its defining word. That makes words
6251: created in this way very compact.
6252:
6253: The final example in this section is intended to remind you that space
6254: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6255: both read and written by a Standard program@footnote{Exercise: use this
6256: example as a starting point for your own implementation of @code{Value}
6257: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6258: @code{[']}.}:
6259:
6260: @example
6261: : foo ( "name" -- )
6262: CREATE -1 ,
6263: DOES> ( -- )
1.33 anton 6264: @@ . ;
1.29 crook 6265:
6266: foo first-word
6267: foo second-word
6268:
6269: 123 ' first-word >BODY !
6270: @end example
6271:
6272: If @code{first-word} had been a @code{CREATE}d word, we could simply
6273: have executed it to get the address of its data field. However, since it
6274: was defined to have @code{DOES>} actions, its execution semantics are to
6275: perform those @code{DOES>} actions. To get the address of its data field
6276: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6277: translate the xt into the address of the data field. When you execute
6278: @code{first-word}, it will display @code{123}. When you execute
6279: @code{second-word} it will display @code{-1}.
1.26 crook 6280:
6281: @cindex stack effect of @code{DOES>}-parts
6282: @cindex @code{DOES>}-parts, stack effect
1.29 crook 6283: In the examples above the stack comment after the @code{DOES>} specifies
1.26 crook 6284: the stack effect of the defined words, not the stack effect of the
6285: following code (the following code expects the address of the body on
6286: the top of stack, which is not reflected in the stack comment). This is
6287: the convention that I use and recommend (it clashes a bit with using
6288: locals declarations for stack effect specification, though).
1.1 anton 6289:
1.53 anton 6290: @menu
6291: * CREATE..DOES> applications::
6292: * CREATE..DOES> details::
1.59 anton 6293: * Advanced does> usage example::
1.53 anton 6294: @end menu
6295:
6296: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26 crook 6297: @subsubsection Applications of @code{CREATE..DOES>}
6298: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1 anton 6299:
1.26 crook 6300: You may wonder how to use this feature. Here are some usage patterns:
1.1 anton 6301:
1.26 crook 6302: @cindex factoring similar colon definitions
6303: When you see a sequence of code occurring several times, and you can
6304: identify a meaning, you will factor it out as a colon definition. When
6305: you see similar colon definitions, you can factor them using
6306: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6307: that look very similar:
1.1 anton 6308: @example
1.26 crook 6309: : ori, ( reg-target reg-source n -- )
6310: 0 asm-reg-reg-imm ;
6311: : andi, ( reg-target reg-source n -- )
6312: 1 asm-reg-reg-imm ;
1.1 anton 6313: @end example
6314:
1.26 crook 6315: @noindent
6316: This could be factored with:
6317: @example
6318: : reg-reg-imm ( op-code -- )
6319: CREATE ,
6320: DOES> ( reg-target reg-source n -- )
6321: @@ asm-reg-reg-imm ;
6322:
6323: 0 reg-reg-imm ori,
6324: 1 reg-reg-imm andi,
6325: @end example
1.1 anton 6326:
1.26 crook 6327: @cindex currying
6328: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6329: supply a part of the parameters for a word (known as @dfn{currying} in
6330: the functional language community). E.g., @code{+} needs two
6331: parameters. Creating versions of @code{+} with one parameter fixed can
6332: be done like this:
1.1 anton 6333: @example
1.26 crook 6334: : curry+ ( n1 -- )
6335: CREATE ,
6336: DOES> ( n2 -- n1+n2 )
6337: @@ + ;
6338:
6339: 3 curry+ 3+
6340: -2 curry+ 2-
1.1 anton 6341: @end example
6342:
1.59 anton 6343: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26 crook 6344: @subsubsection The gory details of @code{CREATE..DOES>}
6345: @cindex @code{CREATE} ... @code{DOES>}, details
1.1 anton 6346:
1.26 crook 6347: doc-does>
1.1 anton 6348:
1.26 crook 6349: @cindex @code{DOES>} in a separate definition
6350: This means that you need not use @code{CREATE} and @code{DOES>} in the
6351: same definition; you can put the @code{DOES>}-part in a separate
1.29 crook 6352: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26 crook 6353: @example
6354: : does1
6355: DOES> ( ... -- ... )
1.44 crook 6356: ... ;
6357:
6358: : does2
6359: DOES> ( ... -- ... )
6360: ... ;
6361:
6362: : def-word ( ... -- ... )
6363: create ...
6364: IF
6365: does1
6366: ELSE
6367: does2
6368: ENDIF ;
6369: @end example
6370:
6371: In this example, the selection of whether to use @code{does1} or
6372: @code{does2} is made at compile-time; at the time that the child word is
6373: @code{CREATE}d.
6374:
6375: @cindex @code{DOES>} in interpretation state
6376: In a standard program you can apply a @code{DOES>}-part only if the last
6377: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6378: will override the behaviour of the last word defined in any case. In a
6379: standard program, you can use @code{DOES>} only in a colon
6380: definition. In Gforth, you can also use it in interpretation state, in a
6381: kind of one-shot mode; for example:
6382: @example
6383: CREATE name ( ... -- ... )
6384: @i{initialization}
6385: DOES>
6386: @i{code} ;
6387: @end example
6388:
6389: @noindent
6390: is equivalent to the standard:
6391: @example
6392: :noname
6393: DOES>
6394: @i{code} ;
6395: CREATE name EXECUTE ( ... -- ... )
6396: @i{initialization}
6397: @end example
6398:
1.53 anton 6399: doc->body
6400:
1.59 anton 6401: @node Advanced does> usage example, , CREATE..DOES> details, User-defined Defining Words
6402: @subsubsection Advanced does> usage example
6403:
6404: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6405: for disassembling instructions, that follow a very repetetive scheme:
6406:
6407: @example
6408: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6409: @var{entry-num} cells @var{table} + !
6410: @end example
6411:
6412: Of course, this inspires the idea to factor out the commonalities to
6413: allow a definition like
6414:
6415: @example
6416: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6417: @end example
6418:
6419: The parameters @var{disasm-operands} and @var{table} are usually
6420: correlated. Moreover, there existed code defining instructions like
6421: this:
6422:
6423: @example
6424: @var{entry-num} @var{inst-format} @var{inst-name}
6425: @end example
6426:
6427: This code comes from the assembler and resides in
6428: @file{arch/mips/insts.fs}.
6429:
6430: So I had to define the @var{inst-format} words that performed the scheme
6431: above when executed. At first I chose to use run-time code-generation:
6432:
6433: @example
6434: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6435: :noname Postpone @var{disasm-operands}
6436: name Postpone sliteral Postpone type Postpone ;
6437: swap cells @var{table} + ! ;
6438: @end example
6439:
6440: Note that this supplies the other two parameters of the scheme above.
1.44 crook 6441:
1.59 anton 6442: An alternative would have been to write this using
6443: @code{create}/@code{does>}:
6444:
6445: @example
6446: : @var{inst-format} ( entry-num "name" -- )
6447: here name string, ( entry-num c-addr ) \ parse and save "name"
6448: noname create , ( entry-num )
6449: lastxt swap cells @var{table} + !
6450: does> ( addr w -- )
6451: \ disassemble instruction w at addr
6452: @@ >r
6453: @var{disasm-operands}
6454: r> count type ;
6455: @end example
6456:
6457: Somehow the first solution is simpler, mainly because it's simpler to
6458: shift a string from definition-time to use-time with @code{sliteral}
6459: than with @code{string,} and friends.
6460:
6461: I wrote a lot of words following this scheme and soon thought about
6462: factoring out the commonalities among them. Note that this uses a
6463: two-level defining word, i.e., a word that defines ordinary defining
6464: words.
6465:
6466: This time a solution involving @code{postpone} and friends seemed more
6467: difficult (try it as an exercise), so I decided to use a
6468: @code{create}/@code{does>} word; since I was already at it, I also used
6469: @code{create}/@code{does>} for the lower level (try using
6470: @code{postpone} etc. as an exercise), resulting in the following
6471: definition:
6472:
6473: @example
6474: : define-format ( disasm-xt table-xt -- )
6475: \ define an instruction format that uses disasm-xt for
6476: \ disassembling and enters the defined instructions into table
6477: \ table-xt
6478: create 2,
6479: does> ( u "inst" -- )
6480: \ defines an anonymous word for disassembling instruction inst,
6481: \ and enters it as u-th entry into table-xt
6482: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6483: noname create 2, \ define anonymous word
6484: execute lastxt swap ! \ enter xt of defined word into table-xt
6485: does> ( addr w -- )
6486: \ disassemble instruction w at addr
6487: 2@@ >r ( addr w disasm-xt R: c-addr )
6488: execute ( R: c-addr ) \ disassemble operands
6489: r> count type ; \ print name
6490: @end example
6491:
6492: Note that the tables here (in contrast to above) do the @code{cells +}
6493: by themselves (that's why you have to pass an xt). This word is used in
6494: the following way:
6495:
6496: @example
6497: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6498: @end example
6499:
6500: In terms of currying, this kind of two-level defining word provides the
6501: parameters in three stages: first @var{disasm-operands} and @var{table},
6502: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6503: the instruction to be disassembled.
6504:
6505: Of course this did not quite fit all the instruction format names used
6506: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6507: the parameters into the right form.
6508:
6509: If you have trouble following this section, don't worry. First, this is
6510: involved and takes time (and probably some playing around) to
6511: understand; second, this is the first two-level
6512: @code{create}/@code{does>} word I have written in seventeen years of
6513: Forth; and if I did not have @file{insts.fs} to start with, I may well
6514: have elected to use just a one-level defining word (with some repeating
6515: of parameters when using the defining word). So it is not necessary to
6516: understand this, but it may improve your understanding of Forth.
1.44 crook 6517:
6518:
6519: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6520: @subsection Deferred words
6521: @cindex deferred words
6522:
6523: The defining word @code{Defer} allows you to define a word by name
6524: without defining its behaviour; the definition of its behaviour is
6525: deferred. Here are two situation where this can be useful:
6526:
6527: @itemize @bullet
6528: @item
6529: Where you want to allow the behaviour of a word to be altered later, and
6530: for all precompiled references to the word to change when its behaviour
6531: is changed.
6532: @item
6533: For mutual recursion; @xref{Calls and returns}.
6534: @end itemize
6535:
6536: In the following example, @code{foo} always invokes the version of
6537: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6538: always invokes the version that prints ``@code{Hello}''. There is no way
6539: of getting @code{foo} to use the later version without re-ordering the
6540: source code and recompiling it.
6541:
6542: @example
6543: : greet ." Good morning" ;
6544: : foo ... greet ... ;
6545: : greet ." Hello" ;
6546: : bar ... greet ... ;
6547: @end example
6548:
6549: This problem can be solved by defining @code{greet} as a @code{Defer}red
6550: word. The behaviour of a @code{Defer}red word can be defined and
6551: redefined at any time by using @code{IS} to associate the xt of a
6552: previously-defined word with it. The previous example becomes:
6553:
6554: @example
6555: Defer greet
6556: : foo ... greet ... ;
6557: : bar ... greet ... ;
6558: : greet1 ." Good morning" ;
6559: : greet2 ." Hello" ;
6560: ' greet2 <IS> greet \ make greet behave like greet2
6561: @end example
6562:
6563: A deferred word can be used to improve the statistics-gathering example
6564: from @ref{User-defined Defining Words}; rather than edit the
6565: application's source code to change every @code{:} to a @code{my:}, do
6566: this:
6567:
6568: @example
6569: : real: : ; \ retain access to the original
6570: defer : \ redefine as a deferred word
6571: ' my: IS : \ use special version of :
6572: \
6573: \ load application here
6574: \
6575: ' real: IS : \ go back to the original
6576: @end example
6577:
6578:
6579: One thing to note is that @code{<IS>} consumes its name when it is
6580: executed. If you want to specify the name at compile time, use
6581: @code{[IS]}:
6582:
6583: @example
6584: : set-greet ( xt -- )
6585: [IS] greet ;
6586:
6587: ' greet1 set-greet
6588: @end example
6589:
6590: A deferred word can only inherit default semantics from the xt (because
1.49 anton 6591: that is all that an xt can represent -- for more discussion of this
6592: @pxref{Tokens for Words}). However, the semantics of the deferred word
1.44 crook 6593: itself can be modified at the time that it is defined. For example:
6594:
6595: @example
6596: : bar .... ; compile-only
6597: Defer fred immediate
6598: Defer jim
6599:
6600: ' bar <IS> jim \ jim has default semantics
6601: ' bar <IS> fred \ fred is immediate
6602: @end example
6603:
6604: doc-defer
6605: doc-<is>
6606: doc-[is]
6607: doc-is
6608: @comment TODO document these: what's defers [is]
6609: doc-what's
6610: doc-defers
6611:
6612: @c Use @code{words-deferred} to see a list of deferred words.
6613:
6614: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6615: are provided in @file{compat/defer.fs}.
6616:
6617:
6618: @node Aliases, Supplying names, Deferred words, Defining Words
6619: @subsection Aliases
6620: @cindex aliases
1.1 anton 6621:
1.44 crook 6622: The defining word @code{Alias} allows you to define a word by name that
6623: has the same behaviour as some other word. Here are two situation where
6624: this can be useful:
1.1 anton 6625:
1.44 crook 6626: @itemize @bullet
6627: @item
6628: When you want access to a word's definition from a different word list
6629: (for an example of this, see the definition of the @code{Root} word list
6630: in the Gforth source).
6631: @item
6632: When you want to create a synonym; a definition that can be known by
6633: either of two names (for example, @code{THEN} and @code{ENDIF} are
6634: aliases).
6635: @end itemize
1.1 anton 6636:
1.44 crook 6637: The word whose behaviour the alias is to inherit is represented by an
6638: xt. Therefore, the alias only inherits default semantics from its
6639: ancestor. The semantics of the alias itself can be modified at the time
6640: that it is defined. For example:
1.1 anton 6641:
6642: @example
1.44 crook 6643: : foo ... ; immediate
6644:
6645: ' foo Alias bar \ bar is not an immediate word
6646: ' foo Alias fooby immediate \ fooby is an immediate word
1.1 anton 6647: @end example
6648:
1.44 crook 6649: Words that are aliases have the same xt, different headers in the
6650: dictionary, and consequently different name tokens (@pxref{Tokens for
6651: Words}) and possibly different immediate flags. An alias can only have
6652: default or immediate compilation semantics; you can define aliases for
6653: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1 anton 6654:
1.44 crook 6655: doc-alias
1.26 crook 6656:
1.1 anton 6657:
1.52 anton 6658: @node Supplying names, , Aliases, Defining Words
1.29 crook 6659: @subsection Supplying the name of a defined word
1.26 crook 6660: @cindex names for defined words
1.44 crook 6661: @cindex defining words, name given in a string
1.1 anton 6662:
1.29 crook 6663: By default, a defining word takes the name for the defined word from the
1.26 crook 6664: input stream. Sometimes you want to supply the name from a string. You
6665: can do this with:
1.1 anton 6666:
1.26 crook 6667: doc-nextname
1.1 anton 6668:
1.26 crook 6669: For example:
1.1 anton 6670:
1.26 crook 6671: @example
6672: s" foo" nextname create
6673: @end example
1.44 crook 6674:
1.26 crook 6675: @noindent
6676: is equivalent to:
1.44 crook 6677:
1.26 crook 6678: @example
6679: create foo
6680: @end example
1.1 anton 6681:
1.29 crook 6682: @noindent
1.44 crook 6683: @code{nextname} works with any defining word, not just @code{:}.
1.1 anton 6684:
6685:
1.47 crook 6686: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6687: @section Interpretation and Compilation Semantics
1.26 crook 6688: @cindex semantics, interpretation and compilation
1.1 anton 6689:
1.26 crook 6690: @cindex interpretation semantics
6691: The @dfn{interpretation semantics} of a word are what the text
6692: interpreter does when it encounters the word in interpret state. It also
6693: appears in some other contexts, e.g., the execution token returned by
1.29 crook 6694: @code{' @i{word}} identifies the interpretation semantics of
6695: @i{word} (in other words, @code{' @i{word} execute} is equivalent to
6696: interpret-state text interpretation of @code{@i{word}}).
1.1 anton 6697:
1.26 crook 6698: @cindex compilation semantics
6699: The @dfn{compilation semantics} of a word are what the text interpreter
6700: does when it encounters the word in compile state. It also appears in
1.29 crook 6701: other contexts, e.g, @code{POSTPONE @i{word}} compiles@footnote{In
1.26 crook 6702: standard terminology, ``appends to the current definition''.} the
1.29 crook 6703: compilation semantics of @i{word}.
1.1 anton 6704:
1.26 crook 6705: @cindex execution semantics
6706: The standard also talks about @dfn{execution semantics}. They are used
6707: only for defining the interpretation and compilation semantics of many
6708: words. By default, the interpretation semantics of a word are to
6709: @code{execute} its execution semantics, and the compilation semantics of
6710: a word are to @code{compile,} its execution semantics.@footnote{In
6711: standard terminology: The default interpretation semantics are its
6712: execution semantics; the default compilation semantics are to append its
6713: execution semantics to the execution semantics of the current
6714: definition.}
6715:
6716: @comment TODO expand, make it co-operate with new sections on text interpreter.
6717:
6718: @cindex immediate words
6719: @cindex compile-only words
6720: You can change the semantics of the most-recently defined word:
6721:
1.44 crook 6722:
1.26 crook 6723: doc-immediate
6724: doc-compile-only
6725: doc-restrict
6726:
1.44 crook 6727:
1.26 crook 6728: Note that ticking (@code{'}) a compile-only word gives an error
6729: (``Interpreting a compile-only word'').
1.1 anton 6730:
1.47 crook 6731: @menu
6732: * Combined words::
6733: @end menu
1.44 crook 6734:
1.48 anton 6735: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44 crook 6736: @subsection Combined Words
6737: @cindex combined words
6738:
6739: Gforth allows you to define @dfn{combined words} -- words that have an
6740: arbitrary combination of interpretation and compilation semantics.
6741:
1.1 anton 6742:
1.26 crook 6743: doc-interpret/compile:
1.1 anton 6744:
1.44 crook 6745:
1.26 crook 6746: This feature was introduced for implementing @code{TO} and @code{S"}. I
6747: recommend that you do not define such words, as cute as they may be:
6748: they make it hard to get at both parts of the word in some contexts.
6749: E.g., assume you want to get an execution token for the compilation
6750: part. Instead, define two words, one that embodies the interpretation
6751: part, and one that embodies the compilation part. Once you have done
6752: that, you can define a combined word with @code{interpret/compile:} for
6753: the convenience of your users.
1.1 anton 6754:
1.26 crook 6755: You might try to use this feature to provide an optimizing
6756: implementation of the default compilation semantics of a word. For
6757: example, by defining:
1.1 anton 6758: @example
1.26 crook 6759: :noname
6760: foo bar ;
6761: :noname
6762: POSTPONE foo POSTPONE bar ;
1.29 crook 6763: interpret/compile: opti-foobar
1.1 anton 6764: @end example
1.26 crook 6765:
1.23 crook 6766: @noindent
1.26 crook 6767: as an optimizing version of:
6768:
1.1 anton 6769: @example
1.26 crook 6770: : foobar
6771: foo bar ;
1.1 anton 6772: @end example
6773:
1.26 crook 6774: Unfortunately, this does not work correctly with @code{[compile]},
6775: because @code{[compile]} assumes that the compilation semantics of all
6776: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29 crook 6777: opti-foobar} would compile compilation semantics, whereas
6778: @code{[compile] foobar} would compile interpretation semantics.
1.1 anton 6779:
1.26 crook 6780: @cindex state-smart words (are a bad idea)
1.29 crook 6781: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26 crook 6782: by @code{interpret/compile:} (words are state-smart if they check
6783: @code{STATE} during execution). E.g., they would try to code
6784: @code{foobar} like this:
1.1 anton 6785:
1.26 crook 6786: @example
6787: : foobar
6788: STATE @@
6789: IF ( compilation state )
6790: POSTPONE foo POSTPONE bar
6791: ELSE
6792: foo bar
6793: ENDIF ; immediate
6794: @end example
1.1 anton 6795:
1.26 crook 6796: Although this works if @code{foobar} is only processed by the text
6797: interpreter, it does not work in other contexts (like @code{'} or
6798: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6799: for a state-smart word, not for the interpretation semantics of the
6800: original @code{foobar}; when you execute this execution token (directly
6801: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6802: state, the result will not be what you expected (i.e., it will not
6803: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6804: write them@footnote{For a more detailed discussion of this topic, see
6805: @cite{@code{State}-smartness -- Why it is Evil and How to Exorcise it} by Anton
6806: Ertl; presented at EuroForth '98 and available from
1.47 crook 6807: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}}!
1.1 anton 6808:
1.26 crook 6809: @cindex defining words with arbitrary semantics combinations
6810: It is also possible to write defining words that define words with
6811: arbitrary combinations of interpretation and compilation semantics. In
6812: general, they look like this:
1.1 anton 6813:
1.26 crook 6814: @example
6815: : def-word
6816: create-interpret/compile
1.29 crook 6817: @i{code1}
1.26 crook 6818: interpretation>
1.29 crook 6819: @i{code2}
1.26 crook 6820: <interpretation
6821: compilation>
1.29 crook 6822: @i{code3}
1.26 crook 6823: <compilation ;
6824: @end example
1.1 anton 6825:
1.29 crook 6826: For a @i{word} defined with @code{def-word}, the interpretation
6827: semantics are to push the address of the body of @i{word} and perform
6828: @i{code2}, and the compilation semantics are to push the address of
6829: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26 crook 6830: can also be defined like this (except that the defined constants don't
6831: behave correctly when @code{[compile]}d):
1.1 anton 6832:
1.26 crook 6833: @example
6834: : constant ( n "name" -- )
6835: create-interpret/compile
6836: ,
6837: interpretation> ( -- n )
6838: @@
6839: <interpretation
6840: compilation> ( compilation. -- ; run-time. -- n )
6841: @@ postpone literal
6842: <compilation ;
6843: @end example
1.1 anton 6844:
1.44 crook 6845:
1.26 crook 6846: doc-create-interpret/compile
6847: doc-interpretation>
6848: doc-<interpretation
6849: doc-compilation>
6850: doc-<compilation
1.1 anton 6851:
1.44 crook 6852:
1.29 crook 6853: Words defined with @code{interpret/compile:} and
1.26 crook 6854: @code{create-interpret/compile} have an extended header structure that
6855: differs from other words; however, unless you try to access them with
6856: plain address arithmetic, you should not notice this. Words for
6857: accessing the header structure usually know how to deal with this; e.g.,
1.29 crook 6858: @code{'} @i{word} @code{>body} also gives you the body of a word created
6859: with @code{create-interpret/compile}.
1.1 anton 6860:
1.44 crook 6861:
1.27 crook 6862: doc-postpone
1.44 crook 6863:
1.29 crook 6864: @comment TODO -- expand glossary text for POSTPONE
1.27 crook 6865:
1.47 crook 6866:
6867: @c -------------------------------------------------------------
6868: @node Tokens for Words, The Text Interpreter, Interpretation and Compilation Semantics, Words
6869: @section Tokens for Words
6870: @cindex tokens for words
6871:
6872: This section describes the creation and use of tokens that represent
6873: words.
6874:
6875: Named words have information stored in their header space entries to
6876: indicate any non-default semantics (@pxref{Interpretation and
6877: Compilation Semantics}). The semantics can be modified, using
6878: @code{immediate} and/or @code{compile-only}, at the time that the words
6879: are defined. Unnamed words have (by definition) no header space
6880: entry, and therefore must have default semantics.
6881:
6882: Named words have interpretation and compilation semantics. Unnamed words
6883: just have execution semantics.
6884:
6885: @cindex xt
6886: @cindex execution token
6887: The execution semantics of an unnamed word are represented by an
6888: @dfn{execution token} (@i{xt}). As explained in @ref{Supplying names},
6889: the execution token of the last word defined can be produced with
6890: @code{lastxt}.
6891:
6892: The interpretation semantics of a named word are also represented by an
6893: execution token. You can produce the execution token using @code{'} or
6894: @code{[']}. A simple example shows the difference between the two:
6895:
6896: @example
6897: : greet ( -- ) ." Hello" ;
6898: : foo ( -- xt ) ['] greet execute ; \ ['] parses greet at compile-time
6899: : bar ( -- ) ' execute ; \ ' parses at run-time
6900:
6901: \ the next four lines all do the same thing
6902: foo
6903: bar greet
6904: greet
6905: ' greet EXECUTE
6906: @end example
6907:
6908: An execution token occupies one cell.
6909: @cindex code field address
6910: @cindex CFA
6911: In Gforth, the abstract data type @i{execution token} is implemented
6912: as a code field address (CFA).
6913: @comment TODO note that the standard does not say what it represents..
6914: @comment and you cannot necessarily compile it in all Forths (eg native
6915: @comment compilers?).
6916:
6917: For literals, use @code{'} in interpreted code and @code{[']} in
6918: compiled code. Gforth's @code{'} and @code{[']} behave somewhat
6919: unusually by complaining about compile-only words. To get the execution
6920: token for a compile-only word @i{name}, use @code{COMP' @i{name} DROP}
6921: or @code{[COMP'] @i{name} DROP}.
6922:
6923: @cindex compilation token
6924: The compilation semantics of a named word are represented by a
6925: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
6926: @i{xt} is an execution token. The compilation semantics represented by
6927: the compilation token can be performed with @code{execute}, which
6928: consumes the whole compilation token, with an additional stack effect
6929: determined by the represented compilation semantics.
6930:
6931: At present, the @i{w} part of a compilation token is an execution token,
6932: and the @i{xt} part represents either @code{execute} or
6933: @code{compile,}@footnote{Depending upon the compilation semantics of the
6934: word. If the word has default compilation semantics, the @i{xt} will
6935: represent @code{compile,}. Otherwise (e.g., for immediate words), the
6936: @i{xt} will represent @code{execute}.}. However, don't rely on that
6937: knowledge, unless necessary; future versions of Gforth may introduce
6938: unusual compilation tokens (e.g., a compilation token that represents
6939: the compilation semantics of a literal).
6940:
6941: You can compile the compilation semantics with @code{postpone,}. I.e.,
6942: @code{COMP' @i{word} postpone,} is equivalent to @code{postpone
6943: @i{word}}.
6944:
6945: @cindex name token
6946: @cindex name field address
6947: @cindex NFA
6948: Named words are also represented by the @dfn{name token}, (@i{nt}). In
6949: Gforth, the abstract data type @emph{name token} is implemented as a
6950: name field address (NFA).
6951:
6952:
6953: doc-execute
6954: doc-perform
6955: doc-compile,
6956: doc-[']
6957: doc-'
6958: doc-[comp']
6959: doc-comp'
6960: doc-postpone,
6961:
6962: doc-find-name
6963: doc-name>int
6964: doc-name?int
6965: doc-name>comp
6966: doc-name>string
6967:
6968:
1.26 crook 6969: @c ----------------------------------------------------------
1.47 crook 6970: @node The Text Interpreter, Word Lists, Tokens for Words, Words
1.26 crook 6971: @section The Text Interpreter
6972: @cindex interpreter - outer
6973: @cindex text interpreter
6974: @cindex outer interpreter
1.1 anton 6975:
1.34 anton 6976: @c Should we really describe all these ugly details? IMO the text
6977: @c interpreter should be much cleaner, but that may not be possible within
6978: @c ANS Forth. - anton
1.44 crook 6979: @c nac-> I wanted to explain how it works to show how you can exploit
6980: @c it in your own programs. When I was writing a cross-compiler, figuring out
6981: @c some of these gory details was very helpful to me. None of the textbooks
6982: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
6983: @c seems to positively avoid going into too much detail for some of
6984: @c the internals.
1.34 anton 6985:
1.29 crook 6986: The text interpreter@footnote{This is an expanded version of the
6987: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34 anton 6988: that processes input from the current input device. It is also called
6989: the outer interpreter, in contrast to the inner interpreter
6990: (@pxref{Engine}) which executes the compiled Forth code on interpretive
6991: implementations.
1.27 crook 6992:
1.29 crook 6993: @cindex interpret state
6994: @cindex compile state
6995: The text interpreter operates in one of two states: @dfn{interpret
6996: state} and @dfn{compile state}. The current state is defined by the
6997: aptly-named variable, @code{state}.
6998:
6999: This section starts by describing how the text interpreter behaves when
7000: it is in interpret state, processing input from the user input device --
7001: the keyboard. This is the mode that a Forth system is in after it starts
7002: up.
7003:
7004: @cindex input buffer
7005: @cindex terminal input buffer
7006: The text interpreter works from an area of memory called the @dfn{input
7007: buffer}@footnote{When the text interpreter is processing input from the
7008: keyboard, this area of memory is called the @dfn{terminal input buffer}
7009: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7010: @code{#TIB}.}, which stores your keyboard input when you press the
1.30 anton 7011: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29 crook 7012: leading spaces (called @dfn{delimiters}) then parses a string (a
7013: sequence of non-space characters) until it reaches either a space
7014: character or the end of the buffer. Having parsed a string, it makes two
7015: attempts to process it:
1.27 crook 7016:
1.29 crook 7017: @cindex dictionary
1.27 crook 7018: @itemize @bullet
7019: @item
1.29 crook 7020: It looks for the string in a @dfn{dictionary} of definitions. If the
7021: string is found, the string names a @dfn{definition} (also known as a
7022: @dfn{word}) and the dictionary search returns information that allows
7023: the text interpreter to perform the word's @dfn{interpretation
7024: semantics}. In most cases, this simply means that the word will be
7025: executed.
1.27 crook 7026: @item
7027: If the string is not found in the dictionary, the text interpreter
1.29 crook 7028: attempts to treat it as a number, using the rules described in
7029: @ref{Number Conversion}. If the string represents a legal number in the
7030: current radix, the number is pushed onto a parameter stack (the data
7031: stack for integers, the floating-point stack for floating-point
7032: numbers).
7033: @end itemize
7034:
7035: If both attempts fail, or if the word is found in the dictionary but has
7036: no interpretation semantics@footnote{This happens if the word was
7037: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7038: remainder of the input buffer, issues an error message and waits for
7039: more input. If one of the attempts succeeds, the text interpreter
7040: repeats the parsing process until the whole of the input buffer has been
7041: processed, at which point it prints the status message ``@code{ ok}''
7042: and waits for more input.
7043:
7044: @cindex parse area
7045: The text interpreter keeps track of its position in the input buffer by
7046: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7047: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7048: of the input buffer. The region from offset @code{>IN @@} to the end of
7049: the input buffer is called the @dfn{parse area}@footnote{In other words,
7050: the text interpreter processes the contents of the input buffer by
7051: parsing strings from the parse area until the parse area is empty.}.
7052: This example shows how @code{>IN} changes as the text interpreter parses
7053: the input buffer:
7054:
7055: @example
7056: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7057: CR ." ->" TYPE ." <-" ; IMMEDIATE
7058:
7059: 1 2 3 remaining + remaining .
7060:
7061: : foo 1 2 3 remaining SWAP remaining ;
7062: @end example
7063:
7064: @noindent
7065: The result is:
7066:
7067: @example
7068: ->+ remaining .<-
7069: ->.<-5 ok
7070:
7071: ->SWAP remaining ;-<
7072: ->;<- ok
7073: @end example
7074:
7075: @cindex parsing words
7076: The value of @code{>IN} can also be modified by a word in the input
7077: buffer that is executed by the text interpreter. This means that a word
7078: can ``trick'' the text interpreter into either skipping a section of the
7079: input buffer@footnote{This is how parsing words work.} or into parsing a
7080: section twice. For example:
1.27 crook 7081:
1.29 crook 7082: @example
7083: : lat ." <<lat>>" ;
7084: : flat ." <<flat>>" >IN DUP @@ 3 - SWAP ! ;
7085: @end example
7086:
7087: @noindent
7088: When @code{flat} is executed, this output is produced@footnote{Exercise
7089: for the reader: what would happen if the @code{3} were replaced with
7090: @code{4}?}:
7091:
7092: @example
7093: <<flat>><<lat>>
7094: @end example
7095:
7096: @noindent
7097: Two important notes about the behaviour of the text interpreter:
1.27 crook 7098:
7099: @itemize @bullet
7100: @item
7101: It processes each input string to completion before parsing additional
1.29 crook 7102: characters from the input buffer.
7103: @item
7104: It treats the input buffer as a read-only region (and so must your code).
7105: @end itemize
7106:
7107: @noindent
7108: When the text interpreter is in compile state, its behaviour changes in
7109: these ways:
7110:
7111: @itemize @bullet
7112: @item
7113: If a parsed string is found in the dictionary, the text interpreter will
7114: perform the word's @dfn{compilation semantics}. In most cases, this
7115: simply means that the execution semantics of the word will be appended
7116: to the current definition.
1.27 crook 7117: @item
1.29 crook 7118: When a number is encountered, it is compiled into the current definition
7119: (as a literal) rather than being pushed onto a parameter stack.
7120: @item
7121: If an error occurs, @code{state} is modified to put the text interpreter
7122: back into interpret state.
7123: @item
7124: Each time a line is entered from the keyboard, Gforth prints
7125: ``@code{ compiled}'' rather than `` @code{ok}''.
7126: @end itemize
7127:
7128: @cindex text interpreter - input sources
7129: When the text interpreter is using an input device other than the
7130: keyboard, its behaviour changes in these ways:
7131:
7132: @itemize @bullet
7133: @item
7134: When the parse area is empty, the text interpreter attempts to refill
7135: the input buffer from the input source. When the input source is
7136: exhausted, the input source is set back to the user input device.
7137: @item
7138: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7139: time the parse area is emptied.
7140: @item
7141: If an error occurs, the input source is set back to the user input
7142: device.
1.27 crook 7143: @end itemize
1.21 crook 7144:
1.49 anton 7145: You can read about this in more detail in @ref{Input Sources}.
1.44 crook 7146:
1.26 crook 7147: doc->in
1.27 crook 7148: doc-source
7149:
1.26 crook 7150: doc-tib
7151: doc-#tib
1.1 anton 7152:
1.44 crook 7153:
1.26 crook 7154: @menu
1.29 crook 7155: * Input Sources::
1.26 crook 7156: * Number Conversion::
7157: * Interpret/Compile states::
7158: * Literals::
7159: * Interpreter Directives::
7160: @end menu
1.1 anton 7161:
1.29 crook 7162: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7163: @subsection Input Sources
7164: @cindex input sources
7165: @cindex text interpreter - input sources
7166:
1.44 crook 7167: By default, the text interpreter processes input from the user input
1.29 crook 7168: device (the keyboard) when Forth starts up. The text interpreter can
7169: process input from any of these sources:
7170:
7171: @itemize @bullet
7172: @item
7173: The user input device -- the keyboard.
7174: @item
7175: A file, using the words described in @ref{Forth source files}.
7176: @item
7177: A block, using the words described in @ref{Blocks}.
7178: @item
7179: A text string, using @code{evaluate}.
7180: @end itemize
7181:
7182: A program can identify the current input device from the values of
7183: @code{source-id} and @code{blk}.
7184:
1.44 crook 7185:
1.29 crook 7186: doc-source-id
7187: doc-blk
7188:
7189: doc-save-input
7190: doc-restore-input
7191:
7192: doc-evaluate
1.1 anton 7193:
1.29 crook 7194:
1.44 crook 7195:
1.29 crook 7196: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26 crook 7197: @subsection Number Conversion
7198: @cindex number conversion
7199: @cindex double-cell numbers, input format
7200: @cindex input format for double-cell numbers
7201: @cindex single-cell numbers, input format
7202: @cindex input format for single-cell numbers
7203: @cindex floating-point numbers, input format
7204: @cindex input format for floating-point numbers
1.1 anton 7205:
1.29 crook 7206: This section describes the rules that the text interpreter uses when it
7207: tries to convert a string into a number.
1.1 anton 7208:
1.26 crook 7209: Let <digit> represent any character that is a legal digit in the current
1.29 crook 7210: number base@footnote{For example, 0-9 when the number base is decimal or
7211: 0-9, A-F when the number base is hexadecimal.}.
1.1 anton 7212:
1.26 crook 7213: Let <decimal digit> represent any character in the range 0-9.
1.1 anton 7214:
1.29 crook 7215: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7216: in the braces (@i{a} or @i{b} or neither).
1.1 anton 7217:
1.26 crook 7218: Let * represent any number of instances of the previous character
7219: (including none).
1.1 anton 7220:
1.26 crook 7221: Let any other character represent itself.
1.1 anton 7222:
1.29 crook 7223: @noindent
1.26 crook 7224: Now, the conversion rules are:
1.21 crook 7225:
1.26 crook 7226: @itemize @bullet
7227: @item
7228: A string of the form <digit><digit>* is treated as a single-precision
1.29 crook 7229: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26 crook 7230: @item
7231: A string of the form -<digit><digit>* is treated as a single-precision
1.29 crook 7232: (cell-sized) negative integer, and is represented using 2's-complement
1.26 crook 7233: arithmetic. Examples are -45 -5681 -0
7234: @item
7235: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29 crook 7236: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7237: (all three of these represent the same number).
1.26 crook 7238: @item
7239: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29 crook 7240: double-precision (double-cell-sized) negative integer, and is
1.26 crook 7241: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29 crook 7242: -34.65 (all three of these represent the same number).
1.26 crook 7243: @item
1.29 crook 7244: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7245: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35 anton 7246: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29 crook 7247: number) +12.E-4
1.26 crook 7248: @end itemize
1.1 anton 7249:
1.26 crook 7250: By default, the number base used for integer number conversion is given
1.35 anton 7251: by the contents of the variable @code{base}. Note that a lot of
7252: confusion can result from unexpected values of @code{base}. If you
7253: change @code{base} anywhere, make sure to save the old value and restore
7254: it afterwards. In general I recommend keeping @code{base} decimal, and
7255: using the prefixes described below for the popular non-decimal bases.
1.1 anton 7256:
1.29 crook 7257: doc-dpl
1.26 crook 7258: doc-base
7259: doc-hex
7260: doc-decimal
1.1 anton 7261:
1.44 crook 7262:
1.26 crook 7263: @cindex '-prefix for character strings
7264: @cindex &-prefix for decimal numbers
7265: @cindex %-prefix for binary numbers
7266: @cindex $-prefix for hexadecimal numbers
1.35 anton 7267: Gforth allows you to override the value of @code{base} by using a
1.29 crook 7268: prefix@footnote{Some Forth implementations provide a similar scheme by
7269: implementing @code{$} etc. as parsing words that process the subsequent
7270: number in the input stream and push it onto the stack. For example, see
7271: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7272: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7273: is required between the prefix and the number.} before the first digit
7274: of an (integer) number. Four prefixes are supported:
1.1 anton 7275:
1.26 crook 7276: @itemize @bullet
7277: @item
1.35 anton 7278: @code{&} -- decimal
1.26 crook 7279: @item
1.35 anton 7280: @code{%} -- binary
1.26 crook 7281: @item
1.35 anton 7282: @code{$} -- hexadecimal
1.26 crook 7283: @item
1.35 anton 7284: @code{'} -- base @code{max-char+1}
1.26 crook 7285: @end itemize
1.1 anton 7286:
1.26 crook 7287: Here are some examples, with the equivalent decimal number shown after
7288: in braces:
1.1 anton 7289:
1.26 crook 7290: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7291: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7292: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7293: &905 (905), $abc (2478), $ABC (2478).
1.1 anton 7294:
1.26 crook 7295: @cindex number conversion - traps for the unwary
1.29 crook 7296: @noindent
1.26 crook 7297: Number conversion has a number of traps for the unwary:
1.1 anton 7298:
1.26 crook 7299: @itemize @bullet
7300: @item
7301: You cannot determine the current number base using the code sequence
1.35 anton 7302: @code{base @@ .} -- the number base is always 10 in the current number
7303: base. Instead, use something like @code{base @@ dec.}
1.26 crook 7304: @item
7305: If the number base is set to a value greater than 14 (for example,
7306: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7307: it to be intepreted as either a single-precision integer or a
7308: floating-point number (Gforth treats it as an integer). The ambiguity
7309: can be resolved by explicitly stating the sign of the mantissa and/or
7310: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7311: ambiguity arises; either representation will be treated as a
7312: floating-point number.
7313: @item
1.29 crook 7314: There is a word @code{bin} but it does @i{not} set the number base!
1.26 crook 7315: It is used to specify file types.
7316: @item
7317: ANS Forth requires the @code{.} of a double-precision number to
7318: be the final character in the string. Allowing the @code{.} to be
7319: anywhere after the first digit is a Gforth extension.
7320: @item
7321: The number conversion process does not check for overflow.
7322: @item
7323: In Gforth, number conversion to floating-point numbers always use base
1.35 anton 7324: 10, irrespective of the value of @code{base}. In ANS Forth,
1.26 crook 7325: conversion to floating-point numbers whilst the value of
1.35 anton 7326: @code{base} is not 10 is an ambiguous condition.
1.26 crook 7327: @end itemize
1.1 anton 7328:
1.49 anton 7329: You can read numbers into your programs with the words described in
7330: @ref{Input}.
1.1 anton 7331:
1.26 crook 7332: @node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
7333: @subsection Interpret/Compile states
7334: @cindex Interpret/Compile states
1.1 anton 7335:
1.29 crook 7336: A standard program is not permitted to change @code{state}
7337: explicitly. However, it can change @code{state} implicitly, using the
7338: words @code{[} and @code{]}. When @code{[} is executed it switches
7339: @code{state} to interpret state, and therefore the text interpreter
7340: starts interpreting. When @code{]} is executed it switches @code{state}
7341: to compile state and therefore the text interpreter starts
1.44 crook 7342: compiling. The most common usage for these words is for switching into
7343: interpret state and back from within a colon definition; this technique
1.49 anton 7344: can be used to compile a literal (for an example, @pxref{Literals}) or
7345: for conditional compilation (for an example, @pxref{Interpreter
7346: Directives}).
1.44 crook 7347:
1.35 anton 7348:
7349: @c This is a bad example: It's non-standard, and it's not necessary.
7350: @c However, I can't think of a good example for switching into compile
7351: @c state when there is no current word (@code{state}-smart words are not a
7352: @c good reason). So maybe we should use an example for switching into
7353: @c interpret @code{state} in a colon def. - anton
1.44 crook 7354: @c nac-> I agree. I started out by putting in the example, then realised
7355: @c that it was non-ANS, so wrote more words around it. I hope this
7356: @c re-written version is acceptable to you. I do want to keep the example
7357: @c as it is helpful for showing what is and what is not portable, particularly
7358: @c where it outlaws a style in common use.
7359:
1.35 anton 7360:
1.44 crook 7361: @code{[} and @code{]} also give you the ability to switch into compile
7362: state and back, but we cannot think of any useful Standard application
7363: for this ability. Pre-ANS Forth textbooks have examples like this:
1.29 crook 7364:
7365: @example
7366: : AA ." this is A" ;
7367: : BB ." this is B" ;
7368: : CC ." this is C" ;
7369:
1.44 crook 7370: create table ] aa bb cc [
7371:
1.29 crook 7372: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7373: cells table + @ execute ;
7374: @end example
7375:
1.44 crook 7376: This example builds a jump table; @code{0 go} will display ``@code{this
7377: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7378: defining @code{table} like this:
1.29 crook 7379:
7380: @example
1.44 crook 7381: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
1.29 crook 7382: @end example
7383:
1.44 crook 7384: The problem with this code is that the definition of @code{table} is not
7385: portable -- it @i{compile}s execution tokens into code space. Whilst it
7386: @i{may} work on systems where code space and data space co-incide, the
1.29 crook 7387: Standard only allows data space to be assigned for a @code{CREATE}d
7388: word. In addition, the Standard only allows @code{@@} to access data
7389: space, whilst this example is using it to access code space. The only
7390: portable, Standard way to build this table is to build it in data space,
7391: like this:
7392:
7393: @example
7394: create table ' aa , ' bb , ' cc ,
7395: @end example
7396:
1.26 crook 7397: doc-state
7398: doc-[
7399: doc-]
1.1 anton 7400:
1.44 crook 7401:
1.26 crook 7402: @node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
7403: @subsection Literals
7404: @cindex Literals
1.21 crook 7405:
1.29 crook 7406: Often, you want to use a number within a colon definition. When you do
7407: this, the text interpreter automatically compiles the number as a
7408: @i{literal}. A literal is a number whose run-time effect is to be pushed
7409: onto the stack. If you had to do some maths to generate the number, you
7410: might write it like this:
7411:
7412: @example
7413: : HOUR-TO-SEC ( n1 -- n2 )
7414: 60 * \ to minutes
7415: 60 * ; \ to seconds
7416: @end example
7417:
7418: It is very clear what this definition is doing, but it's inefficient
7419: since it is performing 2 multiples at run-time. An alternative would be
7420: to write:
7421:
7422: @example
7423: : HOUR-TO-SEC ( n1 -- n2 )
7424: 3600 * ; \ to seconds
7425: @end example
7426:
7427: Which does the same thing, and has the advantage of using a single
7428: multiply. Ideally, we'd like the efficiency of the second with the
7429: readability of the first.
7430:
7431: @code{Literal} allows us to achieve that. It takes a number from the
7432: stack and lays it down in the current definition just as though the
7433: number had been typed directly into the definition. Our first attempt
7434: might look like this:
7435:
7436: @example
7437: 60 \ mins per hour
7438: 60 * \ seconds per minute
7439: : HOUR-TO-SEC ( n1 -- n2 )
7440: Literal * ; \ to seconds
7441: @end example
7442:
7443: But this produces the error message @code{unstructured}. What happened?
7444: The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
7445: @i{colon-sys} is implementation-defined. In other words, once we start a
7446: colon definition we can't portably access anything that was on the stack
7447: before the definition began@footnote{@cite{Two Problems in ANS Forth},
7448: by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
7449: some situations where you might want to access stack items above
7450: colon-sys, and provides a solution to the problem.}. The correct way of
7451: solving this problem in this instance is to use @code{[ ]} like this:
7452:
7453: @example
7454: : HOUR-TO-SEC ( n1 -- n2 )
7455: [ 60 \ minutes per hour
7456: 60 * ] \ seconds per minute
7457: LITERAL * ; \ to seconds
7458: @end example
1.23 crook 7459:
1.44 crook 7460:
1.26 crook 7461: doc-literal
7462: doc-]L
7463: doc-2literal
7464: doc-fliteral
1.1 anton 7465:
1.44 crook 7466:
1.48 anton 7467: @node Interpreter Directives, , Literals, The Text Interpreter
1.26 crook 7468: @subsection Interpreter Directives
7469: @cindex interpreter directives
1.1 anton 7470:
1.29 crook 7471: These words are usually used in interpret state; typically to control
7472: which parts of a source file are processed by the text
1.26 crook 7473: interpreter. There are only a few ANS Forth Standard words, but Gforth
7474: supplements these with a rich set of immediate control structure words
7475: to compensate for the fact that the non-immediate versions can only be
1.29 crook 7476: used in compile state (@pxref{Control Structures}). Typical usages:
7477:
7478: @example
7479: FALSE Constant ASSEMBLER
7480: .
7481: .
7482: ASSEMBLER [IF]
7483: : ASSEMBLER-FEATURE
7484: ...
7485: ;
7486: [ENDIF]
7487: .
7488: .
7489: : SEE
7490: ... \ general-purpose SEE code
7491: [ ASSEMBLER [IF] ]
7492: ... \ assembler-specific SEE code
7493: [ [ENDIF] ]
7494: ;
7495: @end example
1.1 anton 7496:
1.44 crook 7497:
1.26 crook 7498: doc-[IF]
7499: doc-[ELSE]
7500: doc-[THEN]
7501: doc-[ENDIF]
1.1 anton 7502:
1.26 crook 7503: doc-[IFDEF]
7504: doc-[IFUNDEF]
1.1 anton 7505:
1.26 crook 7506: doc-[?DO]
7507: doc-[DO]
7508: doc-[FOR]
7509: doc-[LOOP]
7510: doc-[+LOOP]
7511: doc-[NEXT]
1.1 anton 7512:
1.26 crook 7513: doc-[BEGIN]
7514: doc-[UNTIL]
7515: doc-[AGAIN]
7516: doc-[WHILE]
7517: doc-[REPEAT]
1.1 anton 7518:
1.27 crook 7519:
1.26 crook 7520: @c -------------------------------------------------------------
1.47 crook 7521: @node Word Lists, Environmental Queries, The Text Interpreter, Words
1.26 crook 7522: @section Word Lists
7523: @cindex word lists
1.32 anton 7524: @cindex header space
1.1 anton 7525:
1.36 anton 7526: A wordlist is a list of named words; you can add new words and look up
7527: words by name (and you can remove words in a restricted way with
7528: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7529:
7530: @cindex search order stack
7531: The text interpreter searches the wordlists present in the search order
7532: (a stack of wordlists), from the top to the bottom. Within each
7533: wordlist, the search starts conceptually at the newest word; i.e., if
7534: two words in a wordlist have the same name, the newer word is found.
1.1 anton 7535:
1.26 crook 7536: @cindex compilation word list
1.36 anton 7537: New words are added to the @dfn{compilation wordlist} (aka current
7538: wordlist).
1.1 anton 7539:
1.36 anton 7540: @cindex wid
7541: A word list is identified by a cell-sized word list identifier (@i{wid})
7542: in much the same way as a file is identified by a file handle. The
7543: numerical value of the wid has no (portable) meaning, and might change
7544: from session to session.
1.1 anton 7545:
1.29 crook 7546: The ANS Forth ``Search order'' word set is intended to provide a set of
7547: low-level tools that allow various different schemes to be
1.26 crook 7548: implemented. Gforth provides @code{vocabulary}, a traditional Forth
7549: word. @file{compat/vocabulary.fs} provides an implementation in ANS
1.45 crook 7550: Forth.
1.1 anton 7551:
1.27 crook 7552: @comment TODO: locals section refers to here, saying that every word list (aka
7553: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.1 anton 7554:
1.45 crook 7555: @comment TODO: document markers, reveal, tables, mappedwordlist
7556:
7557: @comment the gforthman- prefix is used to pick out the true definition of a
1.27 crook 7558: @comment word from the source files, rather than some alias.
1.44 crook 7559:
1.26 crook 7560: doc-forth-wordlist
7561: doc-definitions
7562: doc-get-current
7563: doc-set-current
7564: doc-get-order
1.45 crook 7565: doc---gforthman-set-order
1.26 crook 7566: doc-wordlist
1.30 anton 7567: doc-table
1.36 anton 7568: doc-push-order
7569: doc-previous
1.26 crook 7570: doc-also
1.45 crook 7571: doc---gforthman-forth
1.26 crook 7572: doc-only
1.45 crook 7573: doc---gforthman-order
1.15 anton 7574:
1.26 crook 7575: doc-find
7576: doc-search-wordlist
1.15 anton 7577:
1.26 crook 7578: doc-words
7579: doc-vlist
1.44 crook 7580: @c doc-words-deferred
1.1 anton 7581:
1.26 crook 7582: doc-mappedwordlist
7583: doc-root
7584: doc-vocabulary
7585: doc-seal
7586: doc-vocs
7587: doc-current
7588: doc-context
1.1 anton 7589:
1.44 crook 7590:
1.26 crook 7591: @menu
7592: * Why use word lists?::
7593: * Word list examples::
7594: @end menu
7595:
7596: @node Why use word lists?, Word list examples, Word Lists, Word Lists
7597: @subsection Why use word lists?
7598: @cindex word lists - why use them?
7599:
1.29 crook 7600: Here are some reasons for using multiple word lists:
1.26 crook 7601:
7602: @itemize @bullet
7603: @item
1.32 anton 7604: To improve compilation speed by reducing the number of header space
1.26 crook 7605: entries that must be searched. This is achieved by creating a new
7606: word list that contains all of the definitions that are used in the
7607: definition of a Forth system but which would not usually be used by
7608: programs running on that system. That word list would be on the search
7609: list when the Forth system was compiled but would be removed from the
7610: search list for normal operation. This can be a useful technique for
7611: low-performance systems (for example, 8-bit processors in embedded
7612: systems) but is unlikely to be necessary in high-performance desktop
7613: systems.
7614: @item
7615: To prevent a set of words from being used outside the context in which
7616: they are valid. Two classic examples of this are an integrated editor
7617: (all of the edit commands are defined in a separate word list; the
7618: search order is set to the editor word list when the editor is invoked;
7619: the old search order is restored when the editor is terminated) and an
7620: integrated assembler (the op-codes for the machine are defined in a
7621: separate word list which is used when a @code{CODE} word is defined).
7622: @item
7623: To prevent a name-space clash between multiple definitions with the same
7624: name. For example, when building a cross-compiler you might have a word
7625: @code{IF} that generates conditional code for your target system. By
7626: placing this definition in a different word list you can control whether
7627: the host system's @code{IF} or the target system's @code{IF} get used in
7628: any particular context by controlling the order of the word lists on the
7629: search order stack.
7630: @end itemize
1.1 anton 7631:
1.48 anton 7632: @node Word list examples, , Why use word lists?, Word Lists
1.26 crook 7633: @subsection Word list examples
7634: @cindex word lists - examples
1.1 anton 7635:
1.26 crook 7636: Here is an example of creating and using a new wordlist using ANS
7637: Forth Standard words:
1.1 anton 7638:
7639: @example
1.26 crook 7640: wordlist constant my-new-words-wordlist
7641: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
1.21 crook 7642:
1.26 crook 7643: \ add it to the search order
7644: also my-new-words
1.21 crook 7645:
1.26 crook 7646: \ alternatively, add it to the search order and make it
7647: \ the compilation word list
7648: also my-new-words definitions
7649: \ type "order" to see the problem
1.21 crook 7650: @end example
7651:
1.26 crook 7652: The problem with this example is that @code{order} has no way to
7653: associate the name @code{my-new-words} with the wid of the word list (in
7654: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
7655: that has no associated name). There is no Standard way of associating a
7656: name with a wid.
7657:
7658: In Gforth, this example can be re-coded using @code{vocabulary}, which
7659: associates a name with a wid:
1.21 crook 7660:
1.26 crook 7661: @example
7662: vocabulary my-new-words
1.21 crook 7663:
1.26 crook 7664: \ add it to the search order
1.45 crook 7665: also my-new-words
1.21 crook 7666:
1.26 crook 7667: \ alternatively, add it to the search order and make it
7668: \ the compilation word list
7669: my-new-words definitions
7670: \ type "order" to see that the problem is solved
7671: @end example
1.23 crook 7672:
1.26 crook 7673: @c -------------------------------------------------------------
7674: @node Environmental Queries, Files, Word Lists, Words
7675: @section Environmental Queries
7676: @cindex environmental queries
1.21 crook 7677:
1.26 crook 7678: ANS Forth introduced the idea of ``environmental queries'' as a way
7679: for a program running on a system to determine certain characteristics of the system.
7680: The Standard specifies a number of strings that might be recognised by a system.
1.21 crook 7681:
1.32 anton 7682: The Standard requires that the header space used for environmental queries
7683: be distinct from the header space used for definitions.
1.21 crook 7684:
1.26 crook 7685: Typically, environmental queries are supported by creating a set of
1.29 crook 7686: definitions in a word list that is @i{only} used during environmental
1.26 crook 7687: queries; that is what Gforth does. There is no Standard way of adding
7688: definitions to the set of recognised environmental queries, but any
7689: implementation that supports the loading of optional word sets must have
7690: some mechanism for doing this (after loading the word set, the
7691: associated environmental query string must return @code{true}). In
7692: Gforth, the word list used to honour environmental queries can be
7693: manipulated just like any other word list.
1.21 crook 7694:
1.44 crook 7695:
1.26 crook 7696: doc-environment?
7697: doc-environment-wordlist
1.21 crook 7698:
1.26 crook 7699: doc-gforth
7700: doc-os-class
1.21 crook 7701:
1.44 crook 7702:
1.26 crook 7703: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
7704: returning two items on the stack, querying it using @code{environment?}
7705: will return an additional item; the @code{true} flag that shows that the
7706: string was recognised.
1.21 crook 7707:
1.26 crook 7708: @comment TODO Document the standard strings or note where they are documented herein
1.21 crook 7709:
1.26 crook 7710: Here are some examples of using environmental queries:
1.21 crook 7711:
1.26 crook 7712: @example
7713: s" address-unit-bits" environment? 0=
7714: [IF]
7715: cr .( environmental attribute address-units-bits unknown... ) cr
7716: [THEN]
1.21 crook 7717:
1.26 crook 7718: s" block" environment? [IF] DROP include block.fs [THEN]
1.21 crook 7719:
1.26 crook 7720: s" gforth" environment? [IF] 2DROP include compat/vocabulary.fs [THEN]
1.21 crook 7721:
1.26 crook 7722: s" gforth" environment? [IF] .( Gforth version ) TYPE
7723: [ELSE] .( Not Gforth..) [THEN]
7724: @end example
1.21 crook 7725:
7726:
1.26 crook 7727: Here is an example of adding a definition to the environment word list:
1.21 crook 7728:
1.26 crook 7729: @example
7730: get-current environment-wordlist set-current
7731: true constant block
7732: true constant block-ext
7733: set-current
7734: @end example
1.21 crook 7735:
1.26 crook 7736: You can see what definitions are in the environment word list like this:
1.21 crook 7737:
1.26 crook 7738: @example
7739: get-order 1+ environment-wordlist swap set-order words previous
7740: @end example
1.21 crook 7741:
7742:
1.26 crook 7743: @c -------------------------------------------------------------
7744: @node Files, Blocks, Environmental Queries, Words
7745: @section Files
1.28 crook 7746: @cindex files
7747: @cindex I/O - file-handling
1.21 crook 7748:
1.26 crook 7749: Gforth provides facilities for accessing files that are stored in the
7750: host operating system's file-system. Files that are processed by Gforth
7751: can be divided into two categories:
1.21 crook 7752:
1.23 crook 7753: @itemize @bullet
7754: @item
1.29 crook 7755: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23 crook 7756: @item
1.29 crook 7757: Files that are processed by some other program (@dfn{general files}).
1.26 crook 7758: @end itemize
7759:
1.45 crook 7760: doc-loadfilename
7761: doc-sourcefilename
7762: doc-sourceline#
7763:
1.26 crook 7764: @menu
1.48 anton 7765: * Forth source files::
7766: * General files::
7767: * Search Paths::
1.26 crook 7768: @end menu
7769:
1.21 crook 7770:
1.26 crook 7771: @c -------------------------------------------------------------
7772: @node Forth source files, General files, Files, Files
7773: @subsection Forth source files
7774: @cindex including files
7775: @cindex Forth source files
1.21 crook 7776:
1.26 crook 7777: The simplest way to interpret the contents of a file is to use one of
7778: these two formats:
1.21 crook 7779:
1.26 crook 7780: @example
7781: include mysource.fs
7782: s" mysource.fs" included
7783: @end example
1.21 crook 7784:
1.26 crook 7785: Sometimes you want to include a file only if it is not included already
7786: (by, say, another source file). In that case, you can use one of these
1.45 crook 7787: three formats:
1.21 crook 7788:
1.26 crook 7789: @example
7790: require mysource.fs
7791: needs mysource.fs
7792: s" mysource.fs" required
7793: @end example
1.21 crook 7794:
1.26 crook 7795: @cindex stack effect of included files
7796: @cindex including files, stack effect
1.45 crook 7797: It is good practice to write your source files such that interpreting them
7798: does not change the stack. Source files designed in this way can be used with
1.26 crook 7799: @code{required} and friends without complications. For example:
1.21 crook 7800:
1.26 crook 7801: @example
7802: 1 require foo.fs drop
7803: @end example
1.21 crook 7804:
1.44 crook 7805:
1.26 crook 7806: doc-include-file
7807: doc-included
1.28 crook 7808: doc-included?
1.26 crook 7809: doc-include
7810: doc-required
7811: doc-require
7812: doc-needs
1.28 crook 7813: doc-init-included-files
1.21 crook 7814:
1.44 crook 7815:
1.26 crook 7816: A definition in ANS Forth for @code{required} is provided in
7817: @file{compat/required.fs}.
1.21 crook 7818:
1.26 crook 7819: @c -------------------------------------------------------------
7820: @node General files, Search Paths, Forth source files, Files
7821: @subsection General files
7822: @cindex general files
7823: @cindex file-handling
1.21 crook 7824:
1.26 crook 7825: Files are opened/created by name and type. The following types are
7826: recognised:
1.1 anton 7827:
1.44 crook 7828:
1.26 crook 7829: doc-r/o
7830: doc-r/w
7831: doc-w/o
7832: doc-bin
1.1 anton 7833:
1.44 crook 7834:
1.26 crook 7835: When a file is opened/created, it returns a file identifier,
1.29 crook 7836: @i{wfileid} that is used for all other file commands. All file
7837: commands also return a status value, @i{wior}, that is 0 for a
1.26 crook 7838: successful operation and an implementation-defined non-zero value in the
7839: case of an error.
1.21 crook 7840:
1.44 crook 7841:
1.26 crook 7842: doc-open-file
7843: doc-create-file
1.21 crook 7844:
1.26 crook 7845: doc-close-file
7846: doc-delete-file
7847: doc-rename-file
7848: doc-read-file
7849: doc-read-line
7850: doc-write-file
7851: doc-write-line
7852: doc-emit-file
7853: doc-flush-file
1.21 crook 7854:
1.26 crook 7855: doc-file-status
7856: doc-file-position
7857: doc-reposition-file
7858: doc-file-size
7859: doc-resize-file
1.21 crook 7860:
1.44 crook 7861:
1.26 crook 7862: @c ---------------------------------------------------------
1.48 anton 7863: @node Search Paths, , General files, Files
1.26 crook 7864: @subsection Search Paths
7865: @cindex path for @code{included}
7866: @cindex file search path
7867: @cindex @code{include} search path
7868: @cindex search path for files
1.21 crook 7869:
1.26 crook 7870: If you specify an absolute filename (i.e., a filename starting with
7871: @file{/} or @file{~}, or with @file{:} in the second position (as in
7872: @samp{C:...})) for @code{included} and friends, that file is included
7873: just as you would expect.
1.21 crook 7874:
1.26 crook 7875: For relative filenames, Gforth uses a search path similar to Forth's
7876: search order (@pxref{Word Lists}). It tries to find the given filename
7877: in the directories present in the path, and includes the first one it
7878: finds. There are separate search paths for Forth source files and
7879: general files.
1.21 crook 7880:
1.26 crook 7881: If the search path contains the directory @file{.} (as it should), this
7882: refers to the directory that the present file was @code{included}
7883: from. This allows files to include other files relative to their own
7884: position (irrespective of the current working directory or the absolute
7885: position). This feature is essential for libraries consisting of
7886: several files, where a file may include other files from the library.
7887: It corresponds to @code{#include "..."} in C. If the current input
7888: source is not a file, @file{.} refers to the directory of the innermost
7889: file being included, or, if there is no file being included, to the
7890: current working directory.
1.21 crook 7891:
1.26 crook 7892: Use @file{~+} to refer to the current working directory (as in the
7893: @code{bash}).
1.1 anton 7894:
1.26 crook 7895: If the filename starts with @file{./}, the search path is not searched
7896: (just as with absolute filenames), and the @file{.} has the same meaning
7897: as described above.
1.1 anton 7898:
1.48 anton 7899: @menu
7900: * Forth Search Paths::
7901: * General Search Paths::
7902: @end menu
7903:
1.26 crook 7904: @c ---------------------------------------------------------
1.48 anton 7905: @node Forth Search Paths, General Search Paths, Search Paths, Search Paths
1.26 crook 7906: @subsubsection Forth Search Paths
1.28 crook 7907: @cindex search path control - Forth
1.5 anton 7908:
1.26 crook 7909: The search path is initialized when you start Gforth (@pxref{Invoking
7910: Gforth}). You can display it and change it using these words:
1.5 anton 7911:
1.44 crook 7912:
1.26 crook 7913: doc-.fpath
7914: doc-fpath+
7915: doc-fpath=
7916: doc-open-fpath-file
1.5 anton 7917:
1.44 crook 7918:
7919: @noindent
1.26 crook 7920: Here is an example of using @code{fpath} and @code{require}:
1.5 anton 7921:
1.26 crook 7922: @example
7923: fpath= /usr/lib/forth/|./
7924: require timer.fs
7925: @end example
1.5 anton 7926:
1.26 crook 7927: @c ---------------------------------------------------------
1.48 anton 7928: @node General Search Paths, , Forth Search Paths, Search Paths
1.26 crook 7929: @subsubsection General Search Paths
7930: @cindex search path control - for user applications
1.5 anton 7931:
1.26 crook 7932: Your application may need to search files in several directories, like
7933: @code{included} does. To facilitate this, Gforth allows you to define
7934: and use your own search paths, by providing generic equivalents of the
7935: Forth search path words:
1.5 anton 7936:
1.44 crook 7937:
1.26 crook 7938: doc-.path
7939: doc-path+
7940: doc-path=
7941: doc-open-path-file
1.5 anton 7942:
1.44 crook 7943:
1.26 crook 7944: Here's an example of creating a search path:
1.5 anton 7945:
1.26 crook 7946: @example
7947: \ Make a buffer for the path:
7948: create mypath 100 chars , \ maximum length (is checked)
7949: 0 , \ real len
7950: 100 chars allot \ space for path
7951: @end example
1.5 anton 7952:
1.26 crook 7953: @c -------------------------------------------------------------
7954: @node Blocks, Other I/O, Files, Words
7955: @section Blocks
1.28 crook 7956: @cindex I/O - blocks
7957: @cindex blocks
7958:
7959: When you run Gforth on a modern desk-top computer, it runs under the
7960: control of an operating system which provides certain services. One of
7961: these services is @var{file services}, which allows Forth source code
7962: and data to be stored in files and read into Gforth (@pxref{Files}).
7963:
7964: Traditionally, Forth has been an important programming language on
7965: systems where it has interfaced directly to the underlying hardware with
7966: no intervening operating system. Forth provides a mechanism, called
1.29 crook 7967: @dfn{blocks}, for accessing mass storage on such systems.
1.28 crook 7968:
7969: A block is a 1024-byte data area, which can be used to hold data or
7970: Forth source code. No structure is imposed on the contents of the
7971: block. A block is identified by its number; blocks are numbered
7972: contiguously from 1 to an implementation-defined maximum.
7973:
7974: A typical system that used blocks but no operating system might use a
7975: single floppy-disk drive for mass storage, with the disks formatted to
7976: provide 256-byte sectors. Blocks would be implemented by assigning the
7977: first four sectors of the disk to block 1, the second four sectors to
7978: block 2 and so on, up to the limit of the capacity of the disk. The disk
7979: would not contain any file system information, just the set of blocks.
7980:
1.29 crook 7981: @cindex blocks file
1.28 crook 7982: On systems that do provide file services, blocks are typically
1.29 crook 7983: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28 crook 7984: file}. The size of the blocks file will be an exact multiple of 1024
7985: bytes, corresponding to the number of blocks it contains. This is the
7986: mechanism that Gforth uses.
7987:
1.29 crook 7988: @cindex @file{blocks.fb}
1.28 crook 7989: Only 1 blocks file can be open at a time. If you use block words without
7990: having specified a blocks file, Gforth defaults to the blocks file
7991: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
7992: locate a blocks file (@pxref{Forth Search Paths}).
7993:
1.29 crook 7994: @cindex block buffers
1.28 crook 7995: When you read and write blocks under program control, Gforth uses a
1.29 crook 7996: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28 crook 7997: not used when you use @code{load} to interpret the contents of a block.
7998:
7999: The behaviour of the block buffers is directly analagous to that of a
8000: cache. Each block buffer has three states:
8001:
8002: @itemize @bullet
8003: @item
8004: Unassigned
8005: @item
8006: Assigned-clean
8007: @item
8008: Assigned-dirty
8009: @end itemize
8010:
1.29 crook 8011: Initially, all block buffers are @i{unassigned}. In order to access a
1.28 crook 8012: block, the block (specified by its block number) must be assigned to a
8013: block buffer.
8014:
8015: The assignment of a block to a block buffer is performed by @code{block}
8016: or @code{buffer}. Use @code{block} when you wish to modify the existing
8017: contents of a block. Use @code{buffer} when you don't care about the
8018: existing contents of the block@footnote{The ANS Forth definition of
1.35 anton 8019: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28 crook 8020: with the particular block is already stored in a block buffer due to an
8021: earlier @code{block} command, @code{buffer} will return that block
8022: buffer and the existing contents of the block will be
8023: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29 crook 8024: block buffer for the block.}.
1.28 crook 8025:
1.47 crook 8026: Once a block has been assigned to a block buffer using @code{block} or
8027: @code{buffer}, that block buffer becomes the @i{current block buffer}
8028: and its state changes to @i{assigned-clean}. Data may only be
8029: manipulated (read or written) within the current block buffer.
8030:
8031: When the contents of the current block buffer has been modified it is
1.48 anton 8032: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
8033: either abandon the changes (by doing nothing) or commit the changes,
8034: using @code{update}. Using @code{update} does not change the blocks
8035: file; it simply changes a block buffer's state to @i{assigned-dirty}.
1.28 crook 8036:
1.29 crook 8037: The word @code{flush} causes all @i{assigned-dirty} blocks to be
1.28 crook 8038: written back to the blocks file on disk. Leaving Gforth using @code{bye}
8039: also causes a @code{flush} to be performed.
8040:
1.29 crook 8041: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28 crook 8042: algorithm to assign a block buffer to a block. That means that any
8043: particular block can only be assigned to one specific block buffer,
1.29 crook 8044: called (for the particular operation) the @i{victim buffer}. If the
1.47 crook 8045: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8046: the new block immediately. If it is @i{assigned-dirty} its current
8047: contents are written back to the blocks file on disk before it is
1.28 crook 8048: allocated to the new block.
8049:
8050: Although no structure is imposed on the contents of a block, it is
8051: traditional to display the contents as 16 lines each of 64 characters. A
8052: block provides a single, continuous stream of input (for example, it
8053: acts as a single parse area) -- there are no end-of-line characters
8054: within a block, and no end-of-file character at the end of a
8055: block. There are two consequences of this:
1.26 crook 8056:
1.28 crook 8057: @itemize @bullet
8058: @item
8059: The last character of one line wraps straight into the first character
8060: of the following line
8061: @item
8062: The word @code{\} -- comment to end of line -- requires special
8063: treatment; in the context of a block it causes all characters until the
8064: end of the current 64-character ``line'' to be ignored.
8065: @end itemize
8066:
8067: In Gforth, when you use @code{block} with a non-existent block number,
1.45 crook 8068: the current blocks file will be extended to the appropriate size and the
1.28 crook 8069: block buffer will be initialised with spaces.
8070:
1.47 crook 8071: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8072: for details) but doesn't encourage the use of blocks; the mechanism is
8073: only provided for backward compatibility -- ANS Forth requires blocks to
8074: be available when files are.
1.28 crook 8075:
8076: Common techniques that are used when working with blocks include:
8077:
8078: @itemize @bullet
8079: @item
8080: A screen editor that allows you to edit blocks without leaving the Forth
8081: environment.
8082: @item
8083: Shadow screens; where every code block has an associated block
8084: containing comments (for example: code in odd block numbers, comments in
8085: even block numbers). Typically, the block editor provides a convenient
8086: mechanism to toggle between code and comments.
8087: @item
8088: Load blocks; a single block (typically block 1) contains a number of
8089: @code{thru} commands which @code{load} the whole of the application.
8090: @end itemize
1.26 crook 8091:
1.29 crook 8092: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8093: integrated into a Forth programming environment.
1.26 crook 8094:
8095: @comment TODO what about errors on open-blocks?
1.44 crook 8096:
1.26 crook 8097: doc-open-blocks
8098: doc-use
8099: doc-get-block-fid
8100: doc-block-position
1.28 crook 8101:
8102: doc-scr
8103: doc-list
8104:
1.45 crook 8105: doc---gforthman-block
1.28 crook 8106: doc-buffer
8107:
1.26 crook 8108: doc-update
1.28 crook 8109: doc-updated?
1.26 crook 8110: doc-save-buffers
8111: doc-empty-buffers
8112: doc-empty-buffer
8113: doc-flush
1.28 crook 8114:
1.26 crook 8115: doc-load
8116: doc-thru
8117: doc-+load
8118: doc-+thru
1.45 crook 8119: doc---gforthman--->
1.26 crook 8120: doc-block-included
8121:
1.44 crook 8122:
1.26 crook 8123: @c -------------------------------------------------------------
8124: @node Other I/O, Programming Tools, Blocks, Words
8125: @section Other I/O
1.28 crook 8126: @cindex I/O - keyboard and display
1.26 crook 8127:
8128: @menu
8129: * Simple numeric output:: Predefined formats
8130: * Formatted numeric output:: Formatted (pictured) output
8131: * String Formats:: How Forth stores strings in memory
8132: * Displaying characters and strings:: Other stuff
8133: * Input:: Input
8134: @end menu
8135:
8136: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8137: @subsection Simple numeric output
1.28 crook 8138: @cindex numeric output - simple/free-format
1.5 anton 8139:
1.26 crook 8140: The simplest output functions are those that display numbers from the
8141: data or floating-point stacks. Floating-point output is always displayed
8142: using base 10. Numbers displayed from the data stack use the value stored
8143: in @code{base}.
1.5 anton 8144:
1.44 crook 8145:
1.26 crook 8146: doc-.
8147: doc-dec.
8148: doc-hex.
8149: doc-u.
8150: doc-.r
8151: doc-u.r
8152: doc-d.
8153: doc-ud.
8154: doc-d.r
8155: doc-ud.r
8156: doc-f.
8157: doc-fe.
8158: doc-fs.
1.5 anton 8159:
1.44 crook 8160:
1.26 crook 8161: Examples of printing the number 1234.5678E23 in the different floating-point output
8162: formats are shown below:
1.5 anton 8163:
8164: @example
1.26 crook 8165: f. 123456779999999000000000000.
8166: fe. 123.456779999999E24
8167: fs. 1.23456779999999E26
1.5 anton 8168: @end example
8169:
8170:
1.26 crook 8171: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8172: @subsection Formatted numeric output
1.28 crook 8173: @cindex formatted numeric output
1.26 crook 8174: @cindex pictured numeric output
1.28 crook 8175: @cindex numeric output - formatted
1.26 crook 8176:
1.29 crook 8177: Forth traditionally uses a technique called @dfn{pictured numeric
1.26 crook 8178: output} for formatted printing of integers. In this technique, digits
8179: are extracted from the number (using the current output radix defined by
8180: @code{base}), converted to ASCII codes and appended to a string that is
8181: built in a scratch-pad area of memory (@pxref{core-idef,
8182: Implementation-defined options, Implementation-defined
8183: options}). Arbitrary characters can be appended to the string during the
8184: extraction process. The completed string is specified by an address
8185: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8186: under program control.
1.5 anton 8187:
1.26 crook 8188: All of the words described in the previous section for simple numeric
8189: output are implemented in Gforth using pictured numeric output.
1.5 anton 8190:
1.47 crook 8191: Three important things to remember about pictured numeric output:
1.5 anton 8192:
1.26 crook 8193: @itemize @bullet
8194: @item
1.28 crook 8195: It always operates on double-precision numbers; to display a
1.49 anton 8196: single-precision number, convert it first (for ways of doing this
8197: @pxref{Double precision}).
1.26 crook 8198: @item
1.28 crook 8199: It always treats the double-precision number as though it were
8200: unsigned. The examples below show ways of printing signed numbers.
1.26 crook 8201: @item
8202: The string is built up from right to left; least significant digit first.
8203: @end itemize
1.5 anton 8204:
1.44 crook 8205:
1.26 crook 8206: doc-<#
1.47 crook 8207: doc-<<#
1.26 crook 8208: doc-#
8209: doc-#s
8210: doc-hold
8211: doc-sign
8212: doc-#>
1.47 crook 8213: doc-#>>
1.5 anton 8214:
1.26 crook 8215: doc-represent
1.5 anton 8216:
1.44 crook 8217:
8218: @noindent
1.26 crook 8219: Here are some examples of using pictured numeric output:
1.5 anton 8220:
8221: @example
1.26 crook 8222: : my-u. ( u -- )
8223: \ Simplest use of pns.. behaves like Standard u.
8224: 0 \ convert to unsigned double
8225: <# \ start conversion
8226: #s \ convert all digits
8227: #> \ complete conversion
8228: TYPE SPACE ; \ display, with trailing space
1.5 anton 8229:
1.26 crook 8230: : cents-only ( u -- )
8231: 0 \ convert to unsigned double
8232: <# \ start conversion
8233: # # \ convert two least-significant digits
8234: #> \ complete conversion, discard other digits
8235: TYPE SPACE ; \ display, with trailing space
1.5 anton 8236:
1.26 crook 8237: : dollars-and-cents ( u -- )
8238: 0 \ convert to unsigned double
8239: <# \ start conversion
8240: # # \ convert two least-significant digits
8241: [char] . hold \ insert decimal point
8242: #s \ convert remaining digits
8243: [char] $ hold \ append currency symbol
8244: #> \ complete conversion
8245: TYPE SPACE ; \ display, with trailing space
1.5 anton 8246:
1.26 crook 8247: : my-. ( n -- )
8248: \ handling negatives.. behaves like Standard .
8249: s>d \ convert to signed double
8250: swap over dabs \ leave sign byte followed by unsigned double
8251: <# \ start conversion
8252: #s \ convert all digits
8253: rot sign \ get at sign byte, append "-" if needed
8254: #> \ complete conversion
8255: TYPE SPACE ; \ display, with trailing space
1.5 anton 8256:
1.26 crook 8257: : account. ( n -- )
8258: \ accountants don't like minus signs, they use braces
8259: \ for negative numbers
8260: s>d \ convert to signed double
8261: swap over dabs \ leave sign byte followed by unsigned double
8262: <# \ start conversion
8263: 2 pick \ get copy of sign byte
8264: 0< IF [char] ) hold THEN \ right-most character of output
8265: #s \ convert all digits
8266: rot \ get at sign byte
8267: 0< IF [char] ( hold THEN
8268: #> \ complete conversion
8269: TYPE SPACE ; \ display, with trailing space
1.5 anton 8270: @end example
8271:
1.26 crook 8272: Here are some examples of using these words:
1.5 anton 8273:
8274: @example
1.26 crook 8275: 1 my-u. 1
8276: hex -1 my-u. decimal FFFFFFFF
8277: 1 cents-only 01
8278: 1234 cents-only 34
8279: 2 dollars-and-cents $0.02
8280: 1234 dollars-and-cents $12.34
8281: 123 my-. 123
8282: -123 my. -123
8283: 123 account. 123
8284: -456 account. (456)
1.5 anton 8285: @end example
8286:
8287:
1.26 crook 8288: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8289: @subsection String Formats
1.27 crook 8290: @cindex strings - see character strings
8291: @cindex character strings - formats
1.28 crook 8292: @cindex I/O - see character strings
1.26 crook 8293:
1.27 crook 8294: Forth commonly uses two different methods for representing character
8295: strings:
1.26 crook 8296:
8297: @itemize @bullet
8298: @item
8299: @cindex address of counted string
1.45 crook 8300: @cindex counted string
1.29 crook 8301: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8302: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8303: string and the string occupies the subsequent @i{n} char addresses in
1.26 crook 8304: memory.
8305: @item
1.29 crook 8306: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8307: of the string in characters, and @i{c-addr} is the address of the
1.26 crook 8308: first byte of the string.
8309: @end itemize
8310:
8311: ANS Forth encourages the use of the second format when representing
8312: strings on the stack, whilst conceeding that the counted string format
8313: remains useful as a way of storing strings in memory.
8314:
1.44 crook 8315:
1.26 crook 8316: doc-count
8317:
1.44 crook 8318:
1.49 anton 8319: For words that move, copy and search for strings see @ref{Memory
8320: Blocks}. For words that display characters and strings see
8321: @ref{Displaying characters and strings}.
1.26 crook 8322:
8323: @node Displaying characters and strings, Input, String Formats, Other I/O
8324: @subsection Displaying characters and strings
1.27 crook 8325: @cindex characters - compiling and displaying
8326: @cindex character strings - compiling and displaying
1.26 crook 8327:
8328: This section starts with a glossary of Forth words and ends with a set
8329: of examples.
8330:
1.44 crook 8331:
1.26 crook 8332: doc-bl
8333: doc-space
8334: doc-spaces
8335: doc-emit
8336: doc-toupper
8337: doc-."
8338: doc-.(
8339: doc-type
1.44 crook 8340: doc-typewhite
1.26 crook 8341: doc-cr
1.27 crook 8342: @cindex cursor control
1.26 crook 8343: doc-at-xy
8344: doc-page
8345: doc-s"
8346: doc-c"
8347: doc-char
8348: doc-[char]
8349: doc-sliteral
8350:
1.44 crook 8351:
8352: @noindent
1.26 crook 8353: As an example, consider the following text, stored in a file @file{test.fs}:
1.5 anton 8354:
8355: @example
1.26 crook 8356: .( text-1)
8357: : my-word
8358: ." text-2" cr
8359: .( text-3)
8360: ;
8361:
8362: ." text-4"
8363:
8364: : my-char
8365: [char] ALPHABET emit
8366: char emit
8367: ;
1.5 anton 8368: @end example
8369:
1.26 crook 8370: When you load this code into Gforth, the following output is generated:
1.5 anton 8371:
1.26 crook 8372: @example
1.30 anton 8373: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26 crook 8374: @end example
1.5 anton 8375:
1.26 crook 8376: @itemize @bullet
8377: @item
8378: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8379: is an immediate word; it behaves in the same way whether it is used inside
8380: or outside a colon definition.
8381: @item
8382: Message @code{text-4} is displayed because of Gforth's added interpretation
8383: semantics for @code{."}.
8384: @item
1.29 crook 8385: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26 crook 8386: performs the compilation semantics for @code{."} within the definition of
8387: @code{my-word}.
8388: @end itemize
1.5 anton 8389:
1.26 crook 8390: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5 anton 8391:
1.26 crook 8392: @example
1.30 anton 8393: @kbd{my-word @key{RET}} text-2
1.26 crook 8394: ok
1.30 anton 8395: @kbd{my-char fred @key{RET}} Af ok
8396: @kbd{my-char jim @key{RET}} Aj ok
1.26 crook 8397: @end example
1.5 anton 8398:
8399: @itemize @bullet
8400: @item
1.26 crook 8401: Message @code{text-2} is displayed because of the run-time behaviour of
8402: @code{."}.
8403: @item
8404: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8405: on the stack at run-time. @code{emit} always displays the character
8406: when @code{my-char} is executed.
8407: @item
8408: @code{char} parses a string at run-time and the second @code{emit} displays
8409: the first character of the string.
1.5 anton 8410: @item
1.26 crook 8411: If you type @code{see my-char} you can see that @code{[char]} discarded
8412: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8413: definition of @code{my-char}.
1.5 anton 8414: @end itemize
8415:
8416:
8417:
1.48 anton 8418: @node Input, , Displaying characters and strings, Other I/O
1.26 crook 8419: @subsection Input
8420: @cindex input
1.28 crook 8421: @cindex I/O - see input
8422: @cindex parsing a string
1.5 anton 8423:
1.49 anton 8424: For ways of storing character strings in memory see @ref{String Formats}.
1.5 anton 8425:
1.27 crook 8426: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29 crook 8427: @comment then index them
1.27 crook 8428:
1.44 crook 8429:
1.27 crook 8430: doc-key
8431: doc-key?
1.45 crook 8432: doc-ekey
8433: doc-ekey?
8434: doc-ekey>char
1.26 crook 8435: doc->number
8436: doc->float
8437: doc-accept
1.27 crook 8438: doc-pad
8439: doc-parse
8440: doc-word
8441: doc-sword
1.44 crook 8442: doc-(name)
1.27 crook 8443: doc-refill
8444: @comment obsolescent words..
8445: doc-convert
1.26 crook 8446: doc-query
8447: doc-expect
1.27 crook 8448: doc-span
1.5 anton 8449:
8450:
1.44 crook 8451:
1.5 anton 8452: @c -------------------------------------------------------------
1.26 crook 8453: @node Programming Tools, Assembler and Code Words, Other I/O, Words
8454: @section Programming Tools
8455: @cindex programming tools
1.12 anton 8456:
8457: @menu
1.26 crook 8458: * Debugging:: Simple and quick.
8459: * Assertions:: Making your programs self-checking.
1.46 pazsan 8460: * Singlestep Debugger:: Executing your program word by word.
1.5 anton 8461: @end menu
8462:
1.26 crook 8463: @node Debugging, Assertions, Programming Tools, Programming Tools
8464: @subsection Debugging
8465: @cindex debugging
1.5 anton 8466:
1.26 crook 8467: Languages with a slow edit/compile/link/test development loop tend to
8468: require sophisticated tracing/stepping debuggers to facilate
8469: productive debugging.
1.5 anton 8470:
1.26 crook 8471: A much better (faster) way in fast-compiling languages is to add
8472: printing code at well-selected places, let the program run, look at
8473: the output, see where things went wrong, add more printing code, etc.,
8474: until the bug is found.
1.5 anton 8475:
1.26 crook 8476: The simple debugging aids provided in @file{debugs.fs}
8477: are meant to support this style of debugging. In addition, there are
8478: words for non-destructively inspecting the stack and memory:
1.5 anton 8479:
1.44 crook 8480:
1.26 crook 8481: doc-.s
8482: doc-f.s
1.5 anton 8483:
1.44 crook 8484:
1.29 crook 8485: There is a word @code{.r} but it does @i{not} display the return
1.26 crook 8486: stack! It is used for formatted numeric output.
1.5 anton 8487:
1.44 crook 8488:
1.26 crook 8489: doc-depth
8490: doc-fdepth
8491: doc-clearstack
8492: doc-?
8493: doc-dump
1.5 anton 8494:
1.44 crook 8495:
1.26 crook 8496: The word @code{~~} prints debugging information (by default the source
8497: location and the stack contents). It is easy to insert. If you use Emacs
8498: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
8499: query-replace them with nothing). The deferred words
8500: @code{printdebugdata} and @code{printdebugline} control the output of
8501: @code{~~}. The default source location output format works well with
8502: Emacs' compilation mode, so you can step through the program at the
8503: source level using @kbd{C-x `} (the advantage over a stepping debugger
8504: is that you can step in any direction and you know where the crash has
8505: happened or where the strange data has occurred).
1.5 anton 8506:
1.26 crook 8507: The default actions of @code{~~} clobber the contents of the pictured
8508: numeric output string, so you should not use @code{~~}, e.g., between
8509: @code{<#} and @code{#>}.
1.5 anton 8510:
1.44 crook 8511:
1.26 crook 8512: doc-~~
8513: doc-printdebugdata
8514: doc-printdebugline
1.5 anton 8515:
1.26 crook 8516: doc-see
8517: doc-marker
1.5 anton 8518:
1.44 crook 8519:
1.26 crook 8520: Here's an example of using @code{marker} at the start of a source file
8521: that you are debugging; it ensures that you only ever have one copy of
8522: the file's definitions compiled at any time:
1.5 anton 8523:
1.26 crook 8524: @example
8525: [IFDEF] my-code
8526: my-code
8527: [ENDIF]
1.5 anton 8528:
1.26 crook 8529: marker my-code
1.28 crook 8530: init-included-files
1.5 anton 8531:
1.26 crook 8532: \ .. definitions start here
8533: \ .
8534: \ .
8535: \ end
8536: @end example
1.5 anton 8537:
8538:
8539:
1.26 crook 8540: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
8541: @subsection Assertions
8542: @cindex assertions
1.5 anton 8543:
1.26 crook 8544: It is a good idea to make your programs self-checking, especially if you
8545: make an assumption that may become invalid during maintenance (for
8546: example, that a certain field of a data structure is never zero). Gforth
1.29 crook 8547: supports @dfn{assertions} for this purpose. They are used like this:
1.23 crook 8548:
1.26 crook 8549: @example
1.29 crook 8550: assert( @i{flag} )
1.26 crook 8551: @end example
1.23 crook 8552:
1.26 crook 8553: The code between @code{assert(} and @code{)} should compute a flag, that
8554: should be true if everything is alright and false otherwise. It should
8555: not change anything else on the stack. The overall stack effect of the
8556: assertion is @code{( -- )}. E.g.
1.23 crook 8557:
1.26 crook 8558: @example
8559: assert( 1 1 + 2 = ) \ what we learn in school
8560: assert( dup 0<> ) \ assert that the top of stack is not zero
8561: assert( false ) \ this code should not be reached
8562: @end example
1.23 crook 8563:
1.26 crook 8564: The need for assertions is different at different times. During
8565: debugging, we want more checking, in production we sometimes care more
8566: for speed. Therefore, assertions can be turned off, i.e., the assertion
8567: becomes a comment. Depending on the importance of an assertion and the
8568: time it takes to check it, you may want to turn off some assertions and
8569: keep others turned on. Gforth provides several levels of assertions for
8570: this purpose:
1.23 crook 8571:
1.44 crook 8572:
1.26 crook 8573: doc-assert0(
8574: doc-assert1(
8575: doc-assert2(
8576: doc-assert3(
8577: doc-assert(
8578: doc-)
1.23 crook 8579:
1.44 crook 8580:
1.26 crook 8581: The variable @code{assert-level} specifies the highest assertions that
8582: are turned on. I.e., at the default @code{assert-level} of one,
8583: @code{assert0(} and @code{assert1(} assertions perform checking, while
8584: @code{assert2(} and @code{assert3(} assertions are treated as comments.
8585:
8586: The value of @code{assert-level} is evaluated at compile-time, not at
8587: run-time. Therefore you cannot turn assertions on or off at run-time;
8588: you have to set the @code{assert-level} appropriately before compiling a
8589: piece of code. You can compile different pieces of code at different
8590: @code{assert-level}s (e.g., a trusted library at level 1 and
8591: newly-written code at level 3).
1.23 crook 8592:
1.44 crook 8593:
1.26 crook 8594: doc-assert-level
1.23 crook 8595:
1.44 crook 8596:
1.26 crook 8597: If an assertion fails, a message compatible with Emacs' compilation mode
8598: is produced and the execution is aborted (currently with @code{ABORT"}.
8599: If there is interest, we will introduce a special throw code. But if you
8600: intend to @code{catch} a specific condition, using @code{throw} is
8601: probably more appropriate than an assertion).
1.23 crook 8602:
1.26 crook 8603: Definitions in ANS Forth for these assertion words are provided
8604: in @file{compat/assert.fs}.
1.23 crook 8605:
8606:
1.48 anton 8607: @node Singlestep Debugger, , Assertions, Programming Tools
1.26 crook 8608: @subsection Singlestep Debugger
8609: @cindex singlestep Debugger
8610: @cindex debugging Singlestep
1.23 crook 8611:
1.26 crook 8612: When you create a new word there's often the need to check whether it
8613: behaves correctly or not. You can do this by typing @code{dbg
8614: badword}. A debug session might look like this:
1.23 crook 8615:
1.26 crook 8616: @example
8617: : badword 0 DO i . LOOP ; ok
8618: 2 dbg badword
8619: : badword
8620: Scanning code...
1.23 crook 8621:
1.26 crook 8622: Nesting debugger ready!
1.23 crook 8623:
1.26 crook 8624: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
8625: 400D4740 8049F68 DO -> [ 0 ]
8626: 400D4744 804A0C8 i -> [ 1 ] 00000
8627: 400D4748 400C5E60 . -> 0 [ 0 ]
8628: 400D474C 8049D0C LOOP -> [ 0 ]
8629: 400D4744 804A0C8 i -> [ 1 ] 00001
8630: 400D4748 400C5E60 . -> 1 [ 0 ]
8631: 400D474C 8049D0C LOOP -> [ 0 ]
8632: 400D4758 804B384 ; -> ok
8633: @end example
1.23 crook 8634:
1.26 crook 8635: Each line displayed is one step. You always have to hit return to
8636: execute the next word that is displayed. If you don't want to execute
8637: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
8638: an overview what keys are available:
1.23 crook 8639:
1.26 crook 8640: @table @i
1.23 crook 8641:
1.30 anton 8642: @item @key{RET}
1.26 crook 8643: Next; Execute the next word.
1.23 crook 8644:
1.26 crook 8645: @item n
8646: Nest; Single step through next word.
1.5 anton 8647:
1.26 crook 8648: @item u
8649: Unnest; Stop debugging and execute rest of word. If we got to this word
8650: with nest, continue debugging with the calling word.
1.5 anton 8651:
1.26 crook 8652: @item d
8653: Done; Stop debugging and execute rest.
1.5 anton 8654:
1.26 crook 8655: @item s
8656: Stop; Abort immediately.
1.5 anton 8657:
1.26 crook 8658: @end table
1.5 anton 8659:
1.26 crook 8660: Debugging large application with this mechanism is very difficult, because
8661: you have to nest very deeply into the program before the interesting part
8662: begins. This takes a lot of time.
1.5 anton 8663:
1.26 crook 8664: To do it more directly put a @code{BREAK:} command into your source code.
8665: When program execution reaches @code{BREAK:} the single step debugger is
8666: invoked and you have all the features described above.
1.23 crook 8667:
1.26 crook 8668: If you have more than one part to debug it is useful to know where the
8669: program has stopped at the moment. You can do this by the
8670: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
8671: string is typed out when the ``breakpoint'' is reached.
8672:
1.44 crook 8673:
1.26 crook 8674: doc-dbg
1.45 crook 8675: doc-break:
8676: doc-break"
1.26 crook 8677:
8678:
1.44 crook 8679:
1.26 crook 8680: @c -------------------------------------------------------------
8681: @node Assembler and Code Words, Threading Words, Programming Tools, Words
8682: @section Assembler and Code Words
8683: @cindex assembler
8684: @cindex code words
1.5 anton 8685:
1.52 anton 8686: @menu
1.53 anton 8687: * Code and ;code::
8688: * Common Assembler:: Assembler Syntax
1.52 anton 8689: * Common Disassembler::
8690: * 386 Assembler:: Deviations and special cases
8691: * Alpha Assembler:: Deviations and special cases
8692: * MIPS assembler:: Deviations and special cases
1.53 anton 8693: * Other assemblers:: How to write them
1.52 anton 8694: @end menu
8695:
1.53 anton 8696: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
8697: @subsection @code{Code} and @code{;code}
1.52 anton 8698:
1.26 crook 8699: Gforth provides some words for defining primitives (words written in
1.29 crook 8700: machine code), and for defining the machine-code equivalent of
1.26 crook 8701: @code{DOES>}-based defining words. However, the machine-independent
8702: nature of Gforth poses a few problems: First of all, Gforth runs on
8703: several architectures, so it can provide no standard assembler. What's
8704: worse is that the register allocation not only depends on the processor,
8705: but also on the @code{gcc} version and options used.
1.5 anton 8706:
1.29 crook 8707: The words that Gforth offers encapsulate some system dependences (e.g.,
8708: the header structure), so a system-independent assembler may be used in
1.26 crook 8709: Gforth. If you do not have an assembler, you can compile machine code
1.29 crook 8710: directly with @code{,} and @code{c,}@footnote{This isn't portable,
8711: because these words emit stuff in @i{data} space; it works because
8712: Gforth has unified code/data spaces. Assembler isn't likely to be
8713: portable anyway.}.
1.5 anton 8714:
1.44 crook 8715:
1.26 crook 8716: doc-assembler
1.45 crook 8717: doc-init-asm
1.26 crook 8718: doc-code
8719: doc-end-code
8720: doc-;code
8721: doc-flush-icache
1.5 anton 8722:
1.44 crook 8723:
1.26 crook 8724: If @code{flush-icache} does not work correctly, @code{code} words
8725: etc. will not work (reliably), either.
1.5 anton 8726:
1.29 crook 8727: The typical usage of these @code{code} words can be shown most easily by
8728: analogy to the equivalent high-level defining words:
8729:
8730: @example
1.53 anton 8731: : foo code foo
8732: <high-level Forth words> <assembler>
8733: ; end-code
8734:
8735: : bar : bar
8736: <high-level Forth words> <high-level Forth words>
8737: CREATE CREATE
8738: <high-level Forth words> <high-level Forth words>
8739: DOES> ;code
8740: <high-level Forth words> <assembler>
8741: ; end-code
1.29 crook 8742: @end example
8743:
1.26 crook 8744: @code{flush-icache} is always present. The other words are rarely used
8745: and reside in @code{code.fs}, which is usually not loaded. You can load
8746: it with @code{require code.fs}.
1.5 anton 8747:
1.26 crook 8748: @cindex registers of the inner interpreter
8749: In the assembly code you will want to refer to the inner interpreter's
8750: registers (e.g., the data stack pointer) and you may want to use other
8751: registers for temporary storage. Unfortunately, the register allocation
8752: is installation-dependent.
1.5 anton 8753:
1.26 crook 8754: The easiest solution is to use explicit register declarations
8755: (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
8756: GNU C Manual}) for all of the inner interpreter's registers: You have to
8757: compile Gforth with @code{-DFORCE_REG} (configure option
8758: @code{--enable-force-reg}) and the appropriate declarations must be
8759: present in the @code{machine.h} file (see @code{mips.h} for an example;
8760: you can find a full list of all declarable register symbols with
8761: @code{grep register engine.c}). If you give explicit registers to all
8762: variables that are declared at the beginning of @code{engine()}, you
8763: should be able to use the other caller-saved registers for temporary
8764: storage. Alternatively, you can use the @code{gcc} option
8765: @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
8766: Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
8767: (however, this restriction on register allocation may slow Gforth
8768: significantly).
1.5 anton 8769:
1.26 crook 8770: If this solution is not viable (e.g., because @code{gcc} does not allow
8771: you to explicitly declare all the registers you need), you have to find
8772: out by looking at the code where the inner interpreter's registers
8773: reside and which registers can be used for temporary storage. You can
8774: get an assembly listing of the engine's code with @code{make engine.s}.
1.5 anton 8775:
1.26 crook 8776: In any case, it is good practice to abstract your assembly code from the
8777: actual register allocation. E.g., if the data stack pointer resides in
8778: register @code{$17}, create an alias for this register called @code{sp},
8779: and use that in your assembly code.
1.5 anton 8780:
1.26 crook 8781: @cindex code words, portable
8782: Another option for implementing normal and defining words efficiently
8783: is to add the desired functionality to the source of Gforth. For normal
8784: words you just have to edit @file{primitives} (@pxref{Automatic
8785: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
8786: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
8787: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.5 anton 8788:
1.53 anton 8789: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
8790: @subsection Common Assembler
8791:
8792: The assemblers in Gforth generally use a postfix syntax, i.e., the
8793: instruction name follows the operands.
8794:
8795: The operands are passed in the usual order (the same that is used in the
8796: manual of the architecture). Since they all are Forth words, they have
8797: to be separated by spaces; you can also use Forth words to compute the
8798: operands.
8799:
8800: The instruction names usually end with a @code{,}. This makes it easier
8801: to visually separate instructions if you put several of them on one
8802: line; it also avoids shadowing other Forth words (e.g., @code{and}).
8803:
1.55 anton 8804: Registers are usually specified by number; e.g., (decimal) @code{11}
8805: specifies registers R11 and F11 on the Alpha architecture (which one,
8806: depends on the instruction). The usual names are also available, e.g.,
8807: @code{s2} for R11 on Alpha.
8808:
1.53 anton 8809: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
8810: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
8811: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
8812: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
8813: conditions are specified in a way specific to each assembler.
8814:
1.57 anton 8815: Note that the register assignments of the Gforth engine can change
8816: between Gforth versions, or even between different compilations of the
8817: same Gforth version (e.g., if you use a different GCC version). So if
8818: you want to refer to Gforth's registers (e.g., the stack pointer or
8819: TOS), I recommend defining your own words for refering to these
8820: registers, and using them later on; then you can easily adapt to a
8821: changed register assignment. The stability of the register assignment
8822: is usually better if you build Gforth with @code{--enable-force-reg}.
8823:
8824: In particular, the resturn stack pointer and the instruction pointer are
8825: in memory in @code{gforth}, and usually in registers in
8826: @code{gforth-fast}. The most common use of these registers is to
8827: dispatch to the next word (the @code{next} routine). A portable way to
8828: do this is to jump to @code{' noop >code-address} (of course, this is
8829: less efficient than integrating the @code{next} code and scheduling it
8830: well).
8831:
1.52 anton 8832: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
8833: @subsection Common Disassembler
8834:
8835: You can disassemble a @code{code} word with @code{see}
8836: (@pxref{Debugging}). You can disassemble a section of memory with
8837:
8838: doc-disasm
8839:
8840: The disassembler generally produces output that can be fed into the
8841: assembler (i.e., same syntax, etc.). It also includes additional
1.53 anton 8842: information in comments. In particular, the address of the instruction
8843: is given in a comment before the instruction.
8844:
8845: @code{See} may display more or less than the actual code of the word,
8846: because the recognition of the end of the code is unreliable. You can
8847: use @code{disasm} if it did not display enough. It may display more, if
8848: the code word is not immediately followed by a named word. If you have
8849: something else there, you can follow the word with @code{align last @ ,}
8850: to ensure that the end is recognized.
1.52 anton 8851:
8852: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
8853: @subsection 386 Assembler
8854:
8855: The 386 assembler and disassembler included in Gforth was written by
8856: Andrew McKewan; it is in the public domain.
1.57 anton 8857:
8858: The disassembler displays code in prefix Intel syntax.
8859:
8860: The assembler uses an Intel-inspired postfix syntax with reversed
8861: parameters. As usual, a @code{,} is appended to the instruction names
8862: (including @code{rep,} etc.).
8863:
8864: The assembler is somewhat meager, missing a number of instructions
8865: (including FP) and absolute memory addressing modes.
8866:
8867: The registers have their usual names @code{eax} etc. Immediate values
8868: are indicated by postfixing them with @code{#}, e.g., @code{3 #}. Here
8869: are some examples of addressing modes:
8870:
8871: @example
8872: 3 #
8873: eax
8874: 100 [edi]
8875: 4 [ebx] [ecx]
8876: 0 [edi] [eax] *4 \ base register required!
8877: @end example
8878:
8879: Some example of instructions are:
8880:
8881: @example
8882: EAX EBX MOV, \ move ebx,eax
8883: 3 # EAX MOV, \ mov eax,3
8884: 100 [EDI] EAX MOV, \ mov eax,100[edi]
8885: 4 [EBX] [ECX] EAX MOV, \ mov eax,4[ebx][ecx]
8886: 16: EAX EBX MOV, \ mov bx,ax
8887: @end example
8888:
8889: You cannot use the prefix @code{16:} with immediate operands. The
8890: following forms are supported for binary instructions:
8891:
8892: @example
8893: <reg> <reg> <inst>
8894: <n> # <reg> <inst>
8895: <mem> <reg> <inst>
8896: <reg> <mem> <inst>
8897: @end example
8898:
8899: Immediate to memory is not supported. The shift/rotate syntax is:
8900:
8901: @example
8902: <reg/mem> shl,
8903: <reg/mem> 4 shl,
8904: <reg/mem> cl shl,
8905: @end example
8906:
8907: Precede string instructions (@code{movs,} etc.) with @code{byte} to get
8908: the byte version.
8909:
8910: The control structure words @code{if,} @code{until,} etc. must be
8911: preceded by one of these conditions: @code{0= 0< u< u> < > ov ecx0<>}.
8912: You can invert the condition with @code{not} (Note that most of these
8913: words shadow some Forth words when @code{assembler} is before
8914: @code{forth} in the search path, e.g., in code words). Currently the
8915: control structure words use one stack item, so you have to use
8916: @code{roll} instead of @code{cs-roll} to shuffle them (you can also use
8917: @code{swap} etc.).
1.52 anton 8918:
8919: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
8920: @subsection Alpha Assembler
8921:
1.55 anton 8922: The Alpha assembler and disassembler were originally written by Bernd
8923: Thallner.
8924:
8925: The register names @code{a0}--@code{a5} are not available to avoid
8926: shadowing hex numbers.
8927:
8928: Immediate forms of arithmetic instructions are distinguished by a
8929: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
8930: does not count as arithmetic instruction).
8931:
8932: You have to specify all operands to an instruction, even those that
8933: other assemblers consider optional, e.g., the destination register for
8934: @code{br,}, or the destination register and hint for @code{jmp,}.
8935:
8936: You can specify conditions for @code{if,} by removing the first @code{b}
8937: and the trailing @code{,} from a branch with a corresponding name; e.g.,
8938:
8939: @example
8940: 11 fgt if, \ if F11>0e
8941: ...
8942: endif,
1.56 anton 8943: @end example
1.55 anton 8944:
8945: @code{fbgt,} gives @code{fgt}.
1.52 anton 8946:
1.53 anton 8947: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
1.52 anton 8948: @subsection MIPS assembler
8949:
8950: The MIPS assembler was originally written by Christian Pirker.
8951:
8952: Currently the assembler and disassembler only cover the MIPS-I
8953: architecture (R3000), and don't support FP instructions.
8954:
1.55 anton 8955: The register names @code{$a0}--@code{$a3} are not available to avoid
8956: shadowing hex numbers.
1.52 anton 8957:
8958: Because there is no way to distinguish registers from immediate values,
8959: you have to explicitly use the immediate forms of instructions, i.e.,
8960: @code{addiu,}, not just @code{addu,} (@command{as} does this
8961: implicitly).
8962:
8963: If the architecture manual specifies several formats for the instruction
8964: (e.g., for @code{jalr,}), you usually have to use the one with more
8965: arguments (i.e., two for @code{jalr,}). When in doubt, see
8966: @code{arch/mips/testasm.fs} for an example of correct use.
8967:
1.53 anton 8968: Branches and jumps in the MIPS architecture have a delay slot. You have
8969: to fill it yourself (the simplest way is to use @code{nop,}), the
8970: assembler does not do it for you (unlike @command{as}). Even
8971: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
8972: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
8973: and @code{then,} just specify branch targets, they are not affected.
8974:
8975: Note that you must not put branches, jumps, or @code{li,} into the delay
8976: slot: @code{li,} may expand to several instructions, and control flow
8977: instructions may not be put into the branch delay slot in any case.
1.52 anton 8978:
8979: For branches the argument specifying the target is a relative address;
8980: You have to add the address of the delay slot to get the absolute
8981: address.
1.53 anton 8982:
8983: The MIPS architecture also has load delay slots and restrictions on
8984: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
8985: yourself to satisfy these restrictions, the assembler does not do it for
8986: you.
8987:
8988: You can specify the conditions for @code{if,} etc. by taking a
8989: conditional branch and leaving away the @code{b} at the start and the
8990: @code{,} at the end. E.g.,
8991:
8992: @example
8993: 4 5 eq if,
8994: ... \ do something if $4 equals $5
8995: then,
8996: @end example
8997:
8998: @node Other assemblers, , MIPS assembler, Assembler and Code Words
8999: @subsection Other assemblers
9000:
9001: If you want to contribute another assembler/disassembler, please contact
9002: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
9003: already. If you are writing them from scratch, please use a similar
9004: syntax style as the one we use (i.e., postfix, commas at the end of the
9005: instruction names, @pxref{Common Assembler}); make the output of the
9006: disassembler be valid input for the assembler, and keep the style
9007: similar to the style we used.
9008:
9009: Hints on implementation: The most important part is to have a good test
9010: suite that contains all instructions. Once you have that, the rest is
9011: easy. For actual coding you can take a look at
9012: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
9013: the assembler and disassembler, avoiding redundancy and some potential
1.59 anton 9014: bugs. You can also look at that file (and @pxref{Advanced does> usage
9015: example}) to get ideas how to factor a disassembler.
1.5 anton 9016:
1.54 anton 9017: Start with the disassembler, because it's easier to reuse data from the
9018: disassembler for the assembler than the other way round.
9019:
9020: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
9021: how simple it can be.
9022:
1.26 crook 9023: @c -------------------------------------------------------------
9024: @node Threading Words, Locals, Assembler and Code Words, Words
9025: @section Threading Words
9026: @cindex threading words
1.5 anton 9027:
1.26 crook 9028: @cindex code address
9029: These words provide access to code addresses and other threading stuff
9030: in Gforth (and, possibly, other interpretive Forths). It more or less
9031: abstracts away the differences between direct and indirect threading
9032: (and, for direct threading, the machine dependences). However, at
9033: present this wordset is still incomplete. It is also pretty low-level;
9034: some day it will hopefully be made unnecessary by an internals wordset
9035: that abstracts implementation details away completely.
1.5 anton 9036:
1.44 crook 9037:
1.26 crook 9038: doc-threading-method
9039: doc->code-address
9040: doc->does-code
9041: doc-code-address!
9042: doc-does-code!
9043: doc-does-handler!
9044: doc-/does-handler
1.5 anton 9045:
1.44 crook 9046:
1.26 crook 9047: The code addresses produced by various defining words are produced by
9048: the following words:
1.5 anton 9049:
1.44 crook 9050:
1.26 crook 9051: doc-docol:
9052: doc-docon:
9053: doc-dovar:
9054: doc-douser:
9055: doc-dodefer:
9056: doc-dofield:
1.5 anton 9057:
1.44 crook 9058:
1.26 crook 9059: You can recognize words defined by a @code{CREATE}...@code{DOES>} word
9060: with @code{>does-code}. If the word was defined in that way, the value
9061: returned is non-zero and identifies the @code{DOES>} used by the
9062: defining word.
9063: @comment TODO should that be ``identifies the xt of the DOES> ??''
1.5 anton 9064:
1.26 crook 9065: @c -------------------------------------------------------------
9066: @node Locals, Structures, Threading Words, Words
9067: @section Locals
9068: @cindex locals
1.5 anton 9069:
1.26 crook 9070: Local variables can make Forth programming more enjoyable and Forth
9071: programs easier to read. Unfortunately, the locals of ANS Forth are
9072: laden with restrictions. Therefore, we provide not only the ANS Forth
9073: locals wordset, but also our own, more powerful locals wordset (we
9074: implemented the ANS Forth locals wordset through our locals wordset).
1.5 anton 9075:
1.26 crook 9076: The ideas in this section have also been published in the paper
9077: @cite{Automatic Scoping of Local Variables} by M. Anton Ertl, presented
9078: at EuroForth '94; it is available at
1.47 crook 9079: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz}.
1.5 anton 9080:
1.26 crook 9081: @menu
9082: * Gforth locals::
9083: * ANS Forth locals::
9084: @end menu
1.5 anton 9085:
1.26 crook 9086: @node Gforth locals, ANS Forth locals, Locals, Locals
9087: @subsection Gforth locals
9088: @cindex Gforth locals
9089: @cindex locals, Gforth style
1.5 anton 9090:
1.26 crook 9091: Locals can be defined with
1.5 anton 9092:
9093: @example
1.26 crook 9094: @{ local1 local2 ... -- comment @}
9095: @end example
9096: or
9097: @example
9098: @{ local1 local2 ... @}
1.5 anton 9099: @end example
9100:
1.26 crook 9101: E.g.,
1.5 anton 9102: @example
1.26 crook 9103: : max @{ n1 n2 -- n3 @}
9104: n1 n2 > if
9105: n1
9106: else
9107: n2
9108: endif ;
1.5 anton 9109: @end example
9110:
1.26 crook 9111: The similarity of locals definitions with stack comments is intended. A
9112: locals definition often replaces the stack comment of a word. The order
9113: of the locals corresponds to the order in a stack comment and everything
9114: after the @code{--} is really a comment.
1.5 anton 9115:
1.26 crook 9116: This similarity has one disadvantage: It is too easy to confuse locals
9117: declarations with stack comments, causing bugs and making them hard to
9118: find. However, this problem can be avoided by appropriate coding
9119: conventions: Do not use both notations in the same program. If you do,
9120: they should be distinguished using additional means, e.g. by position.
9121:
9122: @cindex types of locals
9123: @cindex locals types
9124: The name of the local may be preceded by a type specifier, e.g.,
9125: @code{F:} for a floating point value:
9126:
9127: @example
9128: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9129: \ complex multiplication
9130: Ar Br f* Ai Bi f* f-
9131: Ar Bi f* Ai Br f* f+ ;
9132: @end example
9133:
9134: @cindex flavours of locals
9135: @cindex locals flavours
9136: @cindex value-flavoured locals
9137: @cindex variable-flavoured locals
9138: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9139: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9140: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9141: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9142: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9143: produces its address (which becomes invalid when the variable's scope is
9144: left). E.g., the standard word @code{emit} can be defined in terms of
9145: @code{type} like this:
1.5 anton 9146:
9147: @example
1.26 crook 9148: : emit @{ C^ char* -- @}
9149: char* 1 type ;
1.5 anton 9150: @end example
9151:
1.26 crook 9152: @cindex default type of locals
9153: @cindex locals, default type
9154: A local without type specifier is a @code{W:} local. Both flavours of
9155: locals are initialized with values from the data or FP stack.
1.5 anton 9156:
1.26 crook 9157: Currently there is no way to define locals with user-defined data
9158: structures, but we are working on it.
1.5 anton 9159:
1.26 crook 9160: Gforth allows defining locals everywhere in a colon definition. This
9161: poses the following questions:
1.5 anton 9162:
1.26 crook 9163: @menu
9164: * Where are locals visible by name?::
9165: * How long do locals live?::
9166: * Programming Style::
9167: * Implementation::
9168: @end menu
1.5 anton 9169:
1.26 crook 9170: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9171: @subsubsection Where are locals visible by name?
9172: @cindex locals visibility
9173: @cindex visibility of locals
9174: @cindex scope of locals
1.5 anton 9175:
1.26 crook 9176: Basically, the answer is that locals are visible where you would expect
9177: it in block-structured languages, and sometimes a little longer. If you
9178: want to restrict the scope of a local, enclose its definition in
9179: @code{SCOPE}...@code{ENDSCOPE}.
1.5 anton 9180:
1.44 crook 9181:
1.26 crook 9182: doc-scope
9183: doc-endscope
1.5 anton 9184:
1.44 crook 9185:
1.26 crook 9186: These words behave like control structure words, so you can use them
9187: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9188: arbitrary ways.
1.5 anton 9189:
1.26 crook 9190: If you want a more exact answer to the visibility question, here's the
9191: basic principle: A local is visible in all places that can only be
9192: reached through the definition of the local@footnote{In compiler
9193: construction terminology, all places dominated by the definition of the
9194: local.}. In other words, it is not visible in places that can be reached
9195: without going through the definition of the local. E.g., locals defined
9196: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9197: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9198: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.5 anton 9199:
1.26 crook 9200: The reasoning behind this solution is: We want to have the locals
9201: visible as long as it is meaningful. The user can always make the
9202: visibility shorter by using explicit scoping. In a place that can
9203: only be reached through the definition of a local, the meaning of a
9204: local name is clear. In other places it is not: How is the local
9205: initialized at the control flow path that does not contain the
9206: definition? Which local is meant, if the same name is defined twice in
9207: two independent control flow paths?
1.5 anton 9208:
1.26 crook 9209: This should be enough detail for nearly all users, so you can skip the
9210: rest of this section. If you really must know all the gory details and
9211: options, read on.
1.5 anton 9212:
1.26 crook 9213: In order to implement this rule, the compiler has to know which places
9214: are unreachable. It knows this automatically after @code{AHEAD},
9215: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9216: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9217: compiler that the control flow never reaches that place. If
9218: @code{UNREACHABLE} is not used where it could, the only consequence is
9219: that the visibility of some locals is more limited than the rule above
9220: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9221: lie to the compiler), buggy code will be produced.
1.5 anton 9222:
1.44 crook 9223:
1.26 crook 9224: doc-unreachable
1.5 anton 9225:
1.44 crook 9226:
1.26 crook 9227: Another problem with this rule is that at @code{BEGIN}, the compiler
9228: does not know which locals will be visible on the incoming
9229: back-edge. All problems discussed in the following are due to this
9230: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9231: loops as examples; the discussion also applies to @code{?DO} and other
9232: loops). Perhaps the most insidious example is:
1.5 anton 9233: @example
1.26 crook 9234: AHEAD
9235: BEGIN
9236: x
9237: [ 1 CS-ROLL ] THEN
9238: @{ x @}
9239: ...
9240: UNTIL
9241: @end example
1.5 anton 9242:
1.26 crook 9243: This should be legal according to the visibility rule. The use of
9244: @code{x} can only be reached through the definition; but that appears
9245: textually below the use.
1.5 anton 9246:
1.26 crook 9247: From this example it is clear that the visibility rules cannot be fully
9248: implemented without major headaches. Our implementation treats common
9249: cases as advertised and the exceptions are treated in a safe way: The
9250: compiler makes a reasonable guess about the locals visible after a
9251: @code{BEGIN}; if it is too pessimistic, the
9252: user will get a spurious error about the local not being defined; if the
9253: compiler is too optimistic, it will notice this later and issue a
9254: warning. In the case above the compiler would complain about @code{x}
9255: being undefined at its use. You can see from the obscure examples in
9256: this section that it takes quite unusual control structures to get the
9257: compiler into trouble, and even then it will often do fine.
1.5 anton 9258:
1.26 crook 9259: If the @code{BEGIN} is reachable from above, the most optimistic guess
9260: is that all locals visible before the @code{BEGIN} will also be
9261: visible after the @code{BEGIN}. This guess is valid for all loops that
9262: are entered only through the @code{BEGIN}, in particular, for normal
9263: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9264: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9265: compiler. When the branch to the @code{BEGIN} is finally generated by
9266: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9267: warns the user if it was too optimistic:
9268: @example
9269: IF
9270: @{ x @}
9271: BEGIN
9272: \ x ?
9273: [ 1 cs-roll ] THEN
9274: ...
9275: UNTIL
1.5 anton 9276: @end example
9277:
1.26 crook 9278: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9279: optimistically assumes that it lives until the @code{THEN}. It notices
9280: this difference when it compiles the @code{UNTIL} and issues a
9281: warning. The user can avoid the warning, and make sure that @code{x}
9282: is not used in the wrong area by using explicit scoping:
9283: @example
9284: IF
9285: SCOPE
9286: @{ x @}
9287: ENDSCOPE
9288: BEGIN
9289: [ 1 cs-roll ] THEN
9290: ...
9291: UNTIL
9292: @end example
1.5 anton 9293:
1.26 crook 9294: Since the guess is optimistic, there will be no spurious error messages
9295: about undefined locals.
1.5 anton 9296:
1.26 crook 9297: If the @code{BEGIN} is not reachable from above (e.g., after
9298: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9299: optimistic guess, as the locals visible after the @code{BEGIN} may be
9300: defined later. Therefore, the compiler assumes that no locals are
9301: visible after the @code{BEGIN}. However, the user can use
9302: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9303: visible at the BEGIN as at the point where the top control-flow stack
9304: item was created.
1.5 anton 9305:
1.44 crook 9306:
1.26 crook 9307: doc-assume-live
1.5 anton 9308:
1.44 crook 9309:
9310: @noindent
1.26 crook 9311: E.g.,
1.5 anton 9312: @example
1.26 crook 9313: @{ x @}
9314: AHEAD
9315: ASSUME-LIVE
9316: BEGIN
9317: x
9318: [ 1 CS-ROLL ] THEN
9319: ...
9320: UNTIL
1.5 anton 9321: @end example
9322:
1.26 crook 9323: Other cases where the locals are defined before the @code{BEGIN} can be
9324: handled by inserting an appropriate @code{CS-ROLL} before the
9325: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9326: behind the @code{ASSUME-LIVE}).
1.5 anton 9327:
1.26 crook 9328: Cases where locals are defined after the @code{BEGIN} (but should be
9329: visible immediately after the @code{BEGIN}) can only be handled by
9330: rearranging the loop. E.g., the ``most insidious'' example above can be
9331: arranged into:
1.5 anton 9332: @example
1.26 crook 9333: BEGIN
9334: @{ x @}
9335: ... 0=
9336: WHILE
9337: x
9338: REPEAT
1.5 anton 9339: @end example
9340:
1.26 crook 9341: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
9342: @subsubsection How long do locals live?
9343: @cindex locals lifetime
9344: @cindex lifetime of locals
1.5 anton 9345:
1.26 crook 9346: The right answer for the lifetime question would be: A local lives at
9347: least as long as it can be accessed. For a value-flavoured local this
9348: means: until the end of its visibility. However, a variable-flavoured
9349: local could be accessed through its address far beyond its visibility
9350: scope. Ultimately, this would mean that such locals would have to be
9351: garbage collected. Since this entails un-Forth-like implementation
9352: complexities, I adopted the same cowardly solution as some other
9353: languages (e.g., C): The local lives only as long as it is visible;
9354: afterwards its address is invalid (and programs that access it
9355: afterwards are erroneous).
1.5 anton 9356:
1.26 crook 9357: @node Programming Style, Implementation, How long do locals live?, Gforth locals
9358: @subsubsection Programming Style
9359: @cindex locals programming style
9360: @cindex programming style, locals
1.5 anton 9361:
1.26 crook 9362: The freedom to define locals anywhere has the potential to change
9363: programming styles dramatically. In particular, the need to use the
9364: return stack for intermediate storage vanishes. Moreover, all stack
9365: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9366: determined arguments) can be eliminated: If the stack items are in the
9367: wrong order, just write a locals definition for all of them; then
9368: write the items in the order you want.
1.5 anton 9369:
1.26 crook 9370: This seems a little far-fetched and eliminating stack manipulations is
9371: unlikely to become a conscious programming objective. Still, the number
9372: of stack manipulations will be reduced dramatically if local variables
1.49 anton 9373: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
1.26 crook 9374: a traditional implementation of @code{max}).
1.5 anton 9375:
1.26 crook 9376: This shows one potential benefit of locals: making Forth programs more
9377: readable. Of course, this benefit will only be realized if the
9378: programmers continue to honour the principle of factoring instead of
9379: using the added latitude to make the words longer.
1.5 anton 9380:
1.26 crook 9381: @cindex single-assignment style for locals
9382: Using @code{TO} can and should be avoided. Without @code{TO},
9383: every value-flavoured local has only a single assignment and many
9384: advantages of functional languages apply to Forth. I.e., programs are
9385: easier to analyse, to optimize and to read: It is clear from the
9386: definition what the local stands for, it does not turn into something
9387: different later.
1.5 anton 9388:
1.26 crook 9389: E.g., a definition using @code{TO} might look like this:
1.5 anton 9390: @example
1.26 crook 9391: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9392: u1 u2 min 0
9393: ?do
9394: addr1 c@@ addr2 c@@ -
9395: ?dup-if
9396: unloop exit
9397: then
9398: addr1 char+ TO addr1
9399: addr2 char+ TO addr2
9400: loop
9401: u1 u2 - ;
1.5 anton 9402: @end example
1.26 crook 9403: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9404: every loop iteration. @code{strcmp} is a typical example of the
9405: readability problems of using @code{TO}. When you start reading
9406: @code{strcmp}, you think that @code{addr1} refers to the start of the
9407: string. Only near the end of the loop you realize that it is something
9408: else.
1.5 anton 9409:
1.26 crook 9410: This can be avoided by defining two locals at the start of the loop that
9411: are initialized with the right value for the current iteration.
1.5 anton 9412: @example
1.26 crook 9413: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9414: addr1 addr2
9415: u1 u2 min 0
9416: ?do @{ s1 s2 @}
9417: s1 c@@ s2 c@@ -
9418: ?dup-if
9419: unloop exit
9420: then
9421: s1 char+ s2 char+
9422: loop
9423: 2drop
9424: u1 u2 - ;
1.5 anton 9425: @end example
1.26 crook 9426: Here it is clear from the start that @code{s1} has a different value
9427: in every loop iteration.
1.5 anton 9428:
1.26 crook 9429: @node Implementation, , Programming Style, Gforth locals
9430: @subsubsection Implementation
9431: @cindex locals implementation
9432: @cindex implementation of locals
1.5 anton 9433:
1.26 crook 9434: @cindex locals stack
9435: Gforth uses an extra locals stack. The most compelling reason for
9436: this is that the return stack is not float-aligned; using an extra stack
9437: also eliminates the problems and restrictions of using the return stack
9438: as locals stack. Like the other stacks, the locals stack grows toward
9439: lower addresses. A few primitives allow an efficient implementation:
1.5 anton 9440:
1.44 crook 9441:
1.26 crook 9442: doc-@local#
9443: doc-f@local#
9444: doc-laddr#
9445: doc-lp+!#
9446: doc-lp!
9447: doc->l
9448: doc-f>l
1.5 anton 9449:
1.44 crook 9450:
1.26 crook 9451: In addition to these primitives, some specializations of these
9452: primitives for commonly occurring inline arguments are provided for
9453: efficiency reasons, e.g., @code{@@local0} as specialization of
9454: @code{@@local#} for the inline argument 0. The following compiling words
9455: compile the right specialized version, or the general version, as
9456: appropriate:
1.6 pazsan 9457:
1.44 crook 9458:
1.26 crook 9459: doc-compile-@local
9460: doc-compile-f@local
9461: doc-compile-lp+!
1.12 anton 9462:
1.44 crook 9463:
1.26 crook 9464: Combinations of conditional branches and @code{lp+!#} like
9465: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9466: is taken) are provided for efficiency and correctness in loops.
1.6 pazsan 9467:
1.26 crook 9468: A special area in the dictionary space is reserved for keeping the
9469: local variable names. @code{@{} switches the dictionary pointer to this
9470: area and @code{@}} switches it back and generates the locals
9471: initializing code. @code{W:} etc.@ are normal defining words. This
9472: special area is cleared at the start of every colon definition.
1.6 pazsan 9473:
1.26 crook 9474: @cindex word list for defining locals
9475: A special feature of Gforth's dictionary is used to implement the
9476: definition of locals without type specifiers: every word list (aka
9477: vocabulary) has its own methods for searching
9478: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9479: with a special search method: When it is searched for a word, it
9480: actually creates that word using @code{W:}. @code{@{} changes the search
9481: order to first search the word list containing @code{@}}, @code{W:} etc.,
9482: and then the word list for defining locals without type specifiers.
1.12 anton 9483:
1.26 crook 9484: The lifetime rules support a stack discipline within a colon
9485: definition: The lifetime of a local is either nested with other locals
9486: lifetimes or it does not overlap them.
1.6 pazsan 9487:
1.26 crook 9488: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9489: pointer manipulation is generated. Between control structure words
9490: locals definitions can push locals onto the locals stack. @code{AGAIN}
9491: is the simplest of the other three control flow words. It has to
9492: restore the locals stack depth of the corresponding @code{BEGIN}
9493: before branching. The code looks like this:
9494: @format
9495: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9496: @code{branch} <begin>
9497: @end format
1.6 pazsan 9498:
1.26 crook 9499: @code{UNTIL} is a little more complicated: If it branches back, it
9500: must adjust the stack just like @code{AGAIN}. But if it falls through,
9501: the locals stack must not be changed. The compiler generates the
9502: following code:
9503: @format
9504: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9505: @end format
9506: The locals stack pointer is only adjusted if the branch is taken.
1.6 pazsan 9507:
1.26 crook 9508: @code{THEN} can produce somewhat inefficient code:
9509: @format
9510: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9511: <orig target>:
9512: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9513: @end format
9514: The second @code{lp+!#} adjusts the locals stack pointer from the
1.29 crook 9515: level at the @i{orig} point to the level after the @code{THEN}. The
1.26 crook 9516: first @code{lp+!#} adjusts the locals stack pointer from the current
9517: level to the level at the orig point, so the complete effect is an
9518: adjustment from the current level to the right level after the
9519: @code{THEN}.
1.6 pazsan 9520:
1.26 crook 9521: @cindex locals information on the control-flow stack
9522: @cindex control-flow stack items, locals information
9523: In a conventional Forth implementation a dest control-flow stack entry
9524: is just the target address and an orig entry is just the address to be
9525: patched. Our locals implementation adds a word list to every orig or dest
9526: item. It is the list of locals visible (or assumed visible) at the point
9527: described by the entry. Our implementation also adds a tag to identify
9528: the kind of entry, in particular to differentiate between live and dead
9529: (reachable and unreachable) orig entries.
1.6 pazsan 9530:
1.26 crook 9531: A few unusual operations have to be performed on locals word lists:
1.6 pazsan 9532:
1.44 crook 9533:
1.26 crook 9534: doc-common-list
9535: doc-sub-list?
9536: doc-list-size
1.6 pazsan 9537:
1.44 crook 9538:
1.26 crook 9539: Several features of our locals word list implementation make these
9540: operations easy to implement: The locals word lists are organised as
9541: linked lists; the tails of these lists are shared, if the lists
9542: contain some of the same locals; and the address of a name is greater
9543: than the address of the names behind it in the list.
1.6 pazsan 9544:
1.26 crook 9545: Another important implementation detail is the variable
9546: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9547: determine if they can be reached directly or only through the branch
9548: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9549: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9550: definition, by @code{BEGIN} and usually by @code{THEN}.
1.6 pazsan 9551:
1.26 crook 9552: Counted loops are similar to other loops in most respects, but
9553: @code{LEAVE} requires special attention: It performs basically the same
9554: service as @code{AHEAD}, but it does not create a control-flow stack
9555: entry. Therefore the information has to be stored elsewhere;
9556: traditionally, the information was stored in the target fields of the
9557: branches created by the @code{LEAVE}s, by organizing these fields into a
9558: linked list. Unfortunately, this clever trick does not provide enough
9559: space for storing our extended control flow information. Therefore, we
9560: introduce another stack, the leave stack. It contains the control-flow
9561: stack entries for all unresolved @code{LEAVE}s.
1.6 pazsan 9562:
1.26 crook 9563: Local names are kept until the end of the colon definition, even if
9564: they are no longer visible in any control-flow path. In a few cases
9565: this may lead to increased space needs for the locals name area, but
9566: usually less than reclaiming this space would cost in code size.
1.6 pazsan 9567:
9568:
1.26 crook 9569: @node ANS Forth locals, , Gforth locals, Locals
9570: @subsection ANS Forth locals
9571: @cindex locals, ANS Forth style
1.6 pazsan 9572:
1.26 crook 9573: The ANS Forth locals wordset does not define a syntax for locals, but
9574: words that make it possible to define various syntaxes. One of the
9575: possible syntaxes is a subset of the syntax we used in the Gforth locals
9576: wordset, i.e.:
1.6 pazsan 9577:
9578: @example
1.26 crook 9579: @{ local1 local2 ... -- comment @}
1.6 pazsan 9580: @end example
1.23 crook 9581: @noindent
1.26 crook 9582: or
1.6 pazsan 9583: @example
1.26 crook 9584: @{ local1 local2 ... @}
1.6 pazsan 9585: @end example
9586:
1.26 crook 9587: The order of the locals corresponds to the order in a stack comment. The
9588: restrictions are:
1.6 pazsan 9589:
9590: @itemize @bullet
9591: @item
1.26 crook 9592: Locals can only be cell-sized values (no type specifiers are allowed).
1.6 pazsan 9593: @item
1.26 crook 9594: Locals can be defined only outside control structures.
1.6 pazsan 9595: @item
1.26 crook 9596: Locals can interfere with explicit usage of the return stack. For the
9597: exact (and long) rules, see the standard. If you don't use return stack
9598: accessing words in a definition using locals, you will be all right. The
9599: purpose of this rule is to make locals implementation on the return
9600: stack easier.
1.6 pazsan 9601: @item
1.26 crook 9602: The whole definition must be in one line.
9603: @end itemize
1.6 pazsan 9604:
1.44 crook 9605: Locals defined in this way behave like @code{VALUE}s
1.49 anton 9606: (@pxref{Values}). I.e., they are initialized from the stack. Using their
1.26 crook 9607: name produces their value. Their value can be changed using @code{TO}.
1.6 pazsan 9608:
1.26 crook 9609: Since this syntax is supported by Gforth directly, you need not do
9610: anything to use it. If you want to port a program using this syntax to
9611: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9612: syntax on the other system.
1.6 pazsan 9613:
1.26 crook 9614: Note that a syntax shown in the standard, section A.13 looks
9615: similar, but is quite different in having the order of locals
9616: reversed. Beware!
1.6 pazsan 9617:
1.26 crook 9618: The ANS Forth locals wordset itself consists of a word:
1.6 pazsan 9619:
1.44 crook 9620:
1.26 crook 9621: doc-(local)
1.6 pazsan 9622:
1.44 crook 9623:
1.26 crook 9624: The ANS Forth locals extension wordset defines a syntax using @code{locals|}, but it is so
9625: awful that we strongly recommend not to use it. We have implemented this
9626: syntax to make porting to Gforth easy, but do not document it here. The
9627: problem with this syntax is that the locals are defined in an order
9628: reversed with respect to the standard stack comment notation, making
9629: programs harder to read, and easier to misread and miswrite. The only
9630: merit of this syntax is that it is easy to implement using the ANS Forth
9631: locals wordset.
1.7 pazsan 9632:
9633:
1.26 crook 9634: @c ----------------------------------------------------------
9635: @node Structures, Object-oriented Forth, Locals, Words
9636: @section Structures
9637: @cindex structures
9638: @cindex records
1.7 pazsan 9639:
1.26 crook 9640: This section presents the structure package that comes with Gforth. A
9641: version of the package implemented in ANS Forth is available in
9642: @file{compat/struct.fs}. This package was inspired by a posting on
9643: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9644: possibly John Hayes). A version of this section has been published in
9645: ???. Marcel Hendrix provided helpful comments.
1.7 pazsan 9646:
1.26 crook 9647: @menu
9648: * Why explicit structure support?::
9649: * Structure Usage::
9650: * Structure Naming Convention::
9651: * Structure Implementation::
9652: * Structure Glossary::
9653: @end menu
1.7 pazsan 9654:
1.26 crook 9655: @node Why explicit structure support?, Structure Usage, Structures, Structures
9656: @subsection Why explicit structure support?
1.7 pazsan 9657:
1.26 crook 9658: @cindex address arithmetic for structures
9659: @cindex structures using address arithmetic
9660: If we want to use a structure containing several fields, we could simply
9661: reserve memory for it, and access the fields using address arithmetic
1.32 anton 9662: (@pxref{Address arithmetic}). As an example, consider a structure with
1.26 crook 9663: the following fields
1.7 pazsan 9664:
1.26 crook 9665: @table @code
9666: @item a
9667: is a float
9668: @item b
9669: is a cell
9670: @item c
9671: is a float
9672: @end table
1.7 pazsan 9673:
1.26 crook 9674: Given the (float-aligned) base address of the structure we get the
9675: address of the field
1.13 pazsan 9676:
1.26 crook 9677: @table @code
9678: @item a
9679: without doing anything further.
9680: @item b
9681: with @code{float+}
9682: @item c
9683: with @code{float+ cell+ faligned}
9684: @end table
1.13 pazsan 9685:
1.26 crook 9686: It is easy to see that this can become quite tiring.
1.13 pazsan 9687:
1.26 crook 9688: Moreover, it is not very readable, because seeing a
9689: @code{cell+} tells us neither which kind of structure is
9690: accessed nor what field is accessed; we have to somehow infer the kind
9691: of structure, and then look up in the documentation, which field of
9692: that structure corresponds to that offset.
1.13 pazsan 9693:
1.26 crook 9694: Finally, this kind of address arithmetic also causes maintenance
9695: troubles: If you add or delete a field somewhere in the middle of the
9696: structure, you have to find and change all computations for the fields
9697: afterwards.
1.13 pazsan 9698:
1.26 crook 9699: So, instead of using @code{cell+} and friends directly, how
9700: about storing the offsets in constants:
1.13 pazsan 9701:
9702: @example
1.26 crook 9703: 0 constant a-offset
9704: 0 float+ constant b-offset
9705: 0 float+ cell+ faligned c-offset
1.13 pazsan 9706: @end example
9707:
1.26 crook 9708: Now we can get the address of field @code{x} with @code{x-offset
9709: +}. This is much better in all respects. Of course, you still
9710: have to change all later offset definitions if you add a field. You can
9711: fix this by declaring the offsets in the following way:
1.13 pazsan 9712:
9713: @example
1.26 crook 9714: 0 constant a-offset
9715: a-offset float+ constant b-offset
9716: b-offset cell+ faligned constant c-offset
1.13 pazsan 9717: @end example
9718:
1.26 crook 9719: Since we always use the offsets with @code{+}, we could use a defining
9720: word @code{cfield} that includes the @code{+} in the action of the
9721: defined word:
1.8 pazsan 9722:
9723: @example
1.26 crook 9724: : cfield ( n "name" -- )
9725: create ,
9726: does> ( name execution: addr1 -- addr2 )
9727: @@ + ;
1.13 pazsan 9728:
1.26 crook 9729: 0 cfield a
9730: 0 a float+ cfield b
9731: 0 b cell+ faligned cfield c
1.13 pazsan 9732: @end example
9733:
1.26 crook 9734: Instead of @code{x-offset +}, we now simply write @code{x}.
9735:
9736: The structure field words now can be used quite nicely. However,
9737: their definition is still a bit cumbersome: We have to repeat the
9738: name, the information about size and alignment is distributed before
9739: and after the field definitions etc. The structure package presented
9740: here addresses these problems.
9741:
9742: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9743: @subsection Structure Usage
9744: @cindex structure usage
1.13 pazsan 9745:
1.26 crook 9746: @cindex @code{field} usage
9747: @cindex @code{struct} usage
9748: @cindex @code{end-struct} usage
9749: You can define a structure for a (data-less) linked list with:
1.13 pazsan 9750: @example
1.26 crook 9751: struct
9752: cell% field list-next
9753: end-struct list%
1.13 pazsan 9754: @end example
9755:
1.26 crook 9756: With the address of the list node on the stack, you can compute the
9757: address of the field that contains the address of the next node with
9758: @code{list-next}. E.g., you can determine the length of a list
9759: with:
1.13 pazsan 9760:
9761: @example
1.26 crook 9762: : list-length ( list -- n )
9763: \ "list" is a pointer to the first element of a linked list
9764: \ "n" is the length of the list
9765: 0 BEGIN ( list1 n1 )
9766: over
9767: WHILE ( list1 n1 )
9768: 1+ swap list-next @@ swap
9769: REPEAT
9770: nip ;
1.13 pazsan 9771: @end example
9772:
1.26 crook 9773: You can reserve memory for a list node in the dictionary with
9774: @code{list% %allot}, which leaves the address of the list node on the
9775: stack. For the equivalent allocation on the heap you can use @code{list%
9776: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9777: use @code{list% %allocate}). You can get the the size of a list
9778: node with @code{list% %size} and its alignment with @code{list%
9779: %alignment}.
1.13 pazsan 9780:
1.26 crook 9781: Note that in ANS Forth the body of a @code{create}d word is
9782: @code{aligned} but not necessarily @code{faligned};
9783: therefore, if you do a:
1.13 pazsan 9784: @example
1.26 crook 9785: create @emph{name} foo% %allot
1.8 pazsan 9786: @end example
9787:
1.26 crook 9788: @noindent
9789: then the memory alloted for @code{foo%} is
9790: guaranteed to start at the body of @code{@emph{name}} only if
9791: @code{foo%} contains only character, cell and double fields.
1.20 pazsan 9792:
1.45 crook 9793: @cindex structures containing structures
1.26 crook 9794: You can include a structure @code{foo%} as a field of
9795: another structure, like this:
1.20 pazsan 9796: @example
1.26 crook 9797: struct
9798: ...
9799: foo% field ...
9800: ...
9801: end-struct ...
1.20 pazsan 9802: @end example
9803:
1.26 crook 9804: @cindex structure extension
9805: @cindex extended records
9806: Instead of starting with an empty structure, you can extend an
9807: existing structure. E.g., a plain linked list without data, as defined
9808: above, is hardly useful; You can extend it to a linked list of integers,
9809: like this:@footnote{This feature is also known as @emph{extended
9810: records}. It is the main innovation in the Oberon language; in other
9811: words, adding this feature to Modula-2 led Wirth to create a new
9812: language, write a new compiler etc. Adding this feature to Forth just
9813: required a few lines of code.}
1.20 pazsan 9814:
9815: @example
1.26 crook 9816: list%
9817: cell% field intlist-int
9818: end-struct intlist%
1.20 pazsan 9819: @end example
9820:
1.26 crook 9821: @code{intlist%} is a structure with two fields:
9822: @code{list-next} and @code{intlist-int}.
1.20 pazsan 9823:
1.26 crook 9824: @cindex structures containing arrays
9825: You can specify an array type containing @emph{n} elements of
9826: type @code{foo%} like this:
1.20 pazsan 9827:
9828: @example
1.26 crook 9829: foo% @emph{n} *
1.20 pazsan 9830: @end example
9831:
1.26 crook 9832: You can use this array type in any place where you can use a normal
9833: type, e.g., when defining a @code{field}, or with
9834: @code{%allot}.
1.20 pazsan 9835:
1.26 crook 9836: @cindex first field optimization
9837: The first field is at the base address of a structure and the word
9838: for this field (e.g., @code{list-next}) actually does not change
9839: the address on the stack. You may be tempted to leave it away in the
9840: interest of run-time and space efficiency. This is not necessary,
9841: because the structure package optimizes this case and compiling such
9842: words does not generate any code. So, in the interest of readability
9843: and maintainability you should include the word for the field when
9844: accessing the field.
1.20 pazsan 9845:
1.26 crook 9846: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9847: @subsection Structure Naming Convention
9848: @cindex structure naming convention
1.20 pazsan 9849:
1.26 crook 9850: The field names that come to (my) mind are often quite generic, and,
9851: if used, would cause frequent name clashes. E.g., many structures
9852: probably contain a @code{counter} field. The structure names
9853: that come to (my) mind are often also the logical choice for the names
9854: of words that create such a structure.
1.20 pazsan 9855:
1.26 crook 9856: Therefore, I have adopted the following naming conventions:
1.20 pazsan 9857:
1.26 crook 9858: @itemize @bullet
9859: @cindex field naming convention
9860: @item
9861: The names of fields are of the form
9862: @code{@emph{struct}-@emph{field}}, where
9863: @code{@emph{struct}} is the basic name of the structure, and
9864: @code{@emph{field}} is the basic name of the field. You can
9865: think of field words as converting the (address of the)
9866: structure into the (address of the) field.
1.20 pazsan 9867:
1.26 crook 9868: @cindex structure naming convention
9869: @item
9870: The names of structures are of the form
9871: @code{@emph{struct}%}, where
9872: @code{@emph{struct}} is the basic name of the structure.
9873: @end itemize
1.20 pazsan 9874:
1.26 crook 9875: This naming convention does not work that well for fields of extended
9876: structures; e.g., the integer list structure has a field
9877: @code{intlist-int}, but has @code{list-next}, not
9878: @code{intlist-next}.
1.20 pazsan 9879:
1.26 crook 9880: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9881: @subsection Structure Implementation
9882: @cindex structure implementation
9883: @cindex implementation of structures
1.20 pazsan 9884:
1.26 crook 9885: The central idea in the implementation is to pass the data about the
9886: structure being built on the stack, not in some global
9887: variable. Everything else falls into place naturally once this design
9888: decision is made.
1.20 pazsan 9889:
1.26 crook 9890: The type description on the stack is of the form @emph{align
9891: size}. Keeping the size on the top-of-stack makes dealing with arrays
9892: very simple.
1.20 pazsan 9893:
1.26 crook 9894: @code{field} is a defining word that uses @code{Create}
9895: and @code{DOES>}. The body of the field contains the offset
9896: of the field, and the normal @code{DOES>} action is simply:
1.20 pazsan 9897:
9898: @example
1.48 anton 9899: @@ +
1.20 pazsan 9900: @end example
9901:
1.23 crook 9902: @noindent
1.26 crook 9903: i.e., add the offset to the address, giving the stack effect
1.29 crook 9904: @i{addr1 -- addr2} for a field.
1.20 pazsan 9905:
1.26 crook 9906: @cindex first field optimization, implementation
9907: This simple structure is slightly complicated by the optimization
9908: for fields with offset 0, which requires a different
9909: @code{DOES>}-part (because we cannot rely on there being
9910: something on the stack if such a field is invoked during
9911: compilation). Therefore, we put the different @code{DOES>}-parts
9912: in separate words, and decide which one to invoke based on the
9913: offset. For a zero offset, the field is basically a noop; it is
9914: immediate, and therefore no code is generated when it is compiled.
1.20 pazsan 9915:
1.26 crook 9916: @node Structure Glossary, , Structure Implementation, Structures
9917: @subsection Structure Glossary
9918: @cindex structure glossary
1.20 pazsan 9919:
1.44 crook 9920:
1.26 crook 9921: doc-%align
9922: doc-%alignment
9923: doc-%alloc
9924: doc-%allocate
9925: doc-%allot
9926: doc-cell%
9927: doc-char%
9928: doc-dfloat%
9929: doc-double%
9930: doc-end-struct
9931: doc-field
9932: doc-float%
9933: doc-naligned
9934: doc-sfloat%
9935: doc-%size
9936: doc-struct
1.23 crook 9937:
1.44 crook 9938:
1.26 crook 9939: @c -------------------------------------------------------------
9940: @node Object-oriented Forth, Passing Commands to the OS, Structures, Words
9941: @section Object-oriented Forth
1.20 pazsan 9942:
1.26 crook 9943: Gforth comes with three packages for object-oriented programming:
9944: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9945: is preloaded, so you have to @code{include} them before use. The most
9946: important differences between these packages (and others) are discussed
9947: in @ref{Comparison with other object models}. All packages are written
9948: in ANS Forth and can be used with any other ANS Forth.
1.20 pazsan 9949:
1.26 crook 9950: @menu
1.48 anton 9951: * Why object-oriented programming?::
9952: * Object-Oriented Terminology::
9953: * Objects::
9954: * OOF::
9955: * Mini-OOF::
1.26 crook 9956: * Comparison with other object models::
9957: @end menu
1.20 pazsan 9958:
1.48 anton 9959: @c ----------------------------------------------------------------
9960: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9961: @subsection Why object-oriented programming?
1.26 crook 9962: @cindex object-oriented programming motivation
9963: @cindex motivation for object-oriented programming
1.23 crook 9964:
1.26 crook 9965: Often we have to deal with several data structures (@emph{objects}),
9966: that have to be treated similarly in some respects, but differently in
9967: others. Graphical objects are the textbook example: circles, triangles,
9968: dinosaurs, icons, and others, and we may want to add more during program
9969: development. We want to apply some operations to any graphical object,
9970: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9971: has to do something different for every kind of object.
9972: @comment TODO add some other operations eg perimeter, area
9973: @comment and tie in to concrete examples later..
1.23 crook 9974:
1.26 crook 9975: We could implement @code{draw} as a big @code{CASE}
9976: control structure that executes the appropriate code depending on the
9977: kind of object to be drawn. This would be not be very elegant, and,
9978: moreover, we would have to change @code{draw} every time we add
9979: a new kind of graphical object (say, a spaceship).
1.23 crook 9980:
1.26 crook 9981: What we would rather do is: When defining spaceships, we would tell
9982: the system: ``Here's how you @code{draw} a spaceship; you figure
9983: out the rest''.
1.23 crook 9984:
1.26 crook 9985: This is the problem that all systems solve that (rightfully) call
9986: themselves object-oriented; the object-oriented packages presented here
9987: solve this problem (and not much else).
9988: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.23 crook 9989:
1.48 anton 9990: @c ------------------------------------------------------------------------
1.26 crook 9991: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
1.48 anton 9992: @subsection Object-Oriented Terminology
1.26 crook 9993: @cindex object-oriented terminology
9994: @cindex terminology for object-oriented programming
1.23 crook 9995:
1.26 crook 9996: This section is mainly for reference, so you don't have to understand
9997: all of it right away. The terminology is mainly Smalltalk-inspired. In
9998: short:
1.23 crook 9999:
1.26 crook 10000: @table @emph
10001: @cindex class
10002: @item class
10003: a data structure definition with some extras.
1.23 crook 10004:
1.26 crook 10005: @cindex object
10006: @item object
10007: an instance of the data structure described by the class definition.
1.23 crook 10008:
1.26 crook 10009: @cindex instance variables
10010: @item instance variables
10011: fields of the data structure.
1.23 crook 10012:
1.26 crook 10013: @cindex selector
10014: @cindex method selector
10015: @cindex virtual function
10016: @item selector
10017: (or @emph{method selector}) a word (e.g.,
10018: @code{draw}) that performs an operation on a variety of data
10019: structures (classes). A selector describes @emph{what} operation to
10020: perform. In C++ terminology: a (pure) virtual function.
1.23 crook 10021:
1.26 crook 10022: @cindex method
10023: @item method
10024: the concrete definition that performs the operation
10025: described by the selector for a specific class. A method specifies
10026: @emph{how} the operation is performed for a specific class.
1.23 crook 10027:
1.26 crook 10028: @cindex selector invocation
10029: @cindex message send
10030: @cindex invoking a selector
10031: @item selector invocation
10032: a call of a selector. One argument of the call (the TOS (top-of-stack))
10033: is used for determining which method is used. In Smalltalk terminology:
10034: a message (consisting of the selector and the other arguments) is sent
10035: to the object.
1.1 anton 10036:
1.26 crook 10037: @cindex receiving object
10038: @item receiving object
10039: the object used for determining the method executed by a selector
10040: invocation. In the @file{objects.fs} model, it is the object that is on
10041: the TOS when the selector is invoked. (@emph{Receiving} comes from
10042: the Smalltalk @emph{message} terminology.)
1.1 anton 10043:
1.26 crook 10044: @cindex child class
10045: @cindex parent class
10046: @cindex inheritance
10047: @item child class
10048: a class that has (@emph{inherits}) all properties (instance variables,
10049: selectors, methods) from a @emph{parent class}. In Smalltalk
10050: terminology: The subclass inherits from the superclass. In C++
10051: terminology: The derived class inherits from the base class.
1.1 anton 10052:
1.26 crook 10053: @end table
1.21 crook 10054:
1.26 crook 10055: @c If you wonder about the message sending terminology, it comes from
10056: @c a time when each object had it's own task and objects communicated via
10057: @c message passing; eventually the Smalltalk developers realized that
10058: @c they can do most things through simple (indirect) calls. They kept the
10059: @c terminology.
1.1 anton 10060:
1.48 anton 10061: @c --------------------------------------------------------------
1.26 crook 10062: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10063: @subsection The @file{objects.fs} model
10064: @cindex objects
10065: @cindex object-oriented programming
1.1 anton 10066:
1.26 crook 10067: @cindex @file{objects.fs}
10068: @cindex @file{oof.fs}
1.1 anton 10069:
1.37 anton 10070: This section describes the @file{objects.fs} package. This material also
10071: has been published in @cite{Yet Another Forth Objects Package} by Anton
10072: Ertl and appeared in Forth Dimensions 19(2), pages 37--43
1.47 crook 10073: (@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html}).
1.26 crook 10074: @c McKewan's and Zsoter's packages
1.1 anton 10075:
1.26 crook 10076: This section assumes that you have read @ref{Structures}.
1.1 anton 10077:
1.26 crook 10078: The techniques on which this model is based have been used to implement
10079: the parser generator, Gray, and have also been used in Gforth for
10080: implementing the various flavours of word lists (hashed or not,
10081: case-sensitive or not, special-purpose word lists for locals etc.).
1.1 anton 10082:
10083:
1.26 crook 10084: @menu
10085: * Properties of the Objects model::
10086: * Basic Objects Usage::
1.37 anton 10087: * The Objects base class::
1.26 crook 10088: * Creating objects::
10089: * Object-Oriented Programming Style::
10090: * Class Binding::
10091: * Method conveniences::
10092: * Classes and Scoping::
1.37 anton 10093: * Dividing classes::
1.26 crook 10094: * Object Interfaces::
10095: * Objects Implementation::
10096: * Objects Glossary::
10097: @end menu
1.1 anton 10098:
1.26 crook 10099: Marcel Hendrix provided helpful comments on this section. Andras Zsoter
10100: and Bernd Paysan helped me with the related works section.
1.1 anton 10101:
1.26 crook 10102: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10103: @subsubsection Properties of the @file{objects.fs} model
10104: @cindex @file{objects.fs} properties
1.1 anton 10105:
1.26 crook 10106: @itemize @bullet
10107: @item
10108: It is straightforward to pass objects on the stack. Passing
10109: selectors on the stack is a little less convenient, but possible.
1.1 anton 10110:
1.26 crook 10111: @item
10112: Objects are just data structures in memory, and are referenced by their
10113: address. You can create words for objects with normal defining words
10114: like @code{constant}. Likewise, there is no difference between instance
10115: variables that contain objects and those that contain other data.
1.1 anton 10116:
1.26 crook 10117: @item
10118: Late binding is efficient and easy to use.
1.21 crook 10119:
1.26 crook 10120: @item
10121: It avoids parsing, and thus avoids problems with state-smartness
10122: and reduced extensibility; for convenience there are a few parsing
10123: words, but they have non-parsing counterparts. There are also a few
10124: defining words that parse. This is hard to avoid, because all standard
10125: defining words parse (except @code{:noname}); however, such
10126: words are not as bad as many other parsing words, because they are not
10127: state-smart.
1.21 crook 10128:
1.26 crook 10129: @item
10130: It does not try to incorporate everything. It does a few things and does
10131: them well (IMO). In particular, this model was not designed to support
10132: information hiding (although it has features that may help); you can use
10133: a separate package for achieving this.
1.21 crook 10134:
1.26 crook 10135: @item
10136: It is layered; you don't have to learn and use all features to use this
1.49 anton 10137: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10138: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
1.26 crook 10139: are optional and independent of each other.
1.21 crook 10140:
1.26 crook 10141: @item
10142: An implementation in ANS Forth is available.
1.21 crook 10143:
1.26 crook 10144: @end itemize
1.21 crook 10145:
10146:
1.26 crook 10147: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10148: @subsubsection Basic @file{objects.fs} Usage
10149: @cindex basic objects usage
10150: @cindex objects, basic usage
1.21 crook 10151:
1.26 crook 10152: You can define a class for graphical objects like this:
1.21 crook 10153:
1.26 crook 10154: @cindex @code{class} usage
10155: @cindex @code{end-class} usage
10156: @cindex @code{selector} usage
10157: @example
10158: object class \ "object" is the parent class
10159: selector draw ( x y graphical -- )
10160: end-class graphical
10161: @end example
1.21 crook 10162:
1.26 crook 10163: This code defines a class @code{graphical} with an
10164: operation @code{draw}. We can perform the operation
10165: @code{draw} on any @code{graphical} object, e.g.:
1.21 crook 10166:
1.26 crook 10167: @example
10168: 100 100 t-rex draw
10169: @end example
1.21 crook 10170:
1.26 crook 10171: @noindent
10172: where @code{t-rex} is a word (say, a constant) that produces a
10173: graphical object.
1.21 crook 10174:
1.29 crook 10175: @comment TODO add a 2nd operation eg perimeter.. and use for
1.26 crook 10176: @comment a concrete example
1.21 crook 10177:
1.26 crook 10178: @cindex abstract class
10179: How do we create a graphical object? With the present definitions,
10180: we cannot create a useful graphical object. The class
10181: @code{graphical} describes graphical objects in general, but not
10182: any concrete graphical object type (C++ users would call it an
10183: @emph{abstract class}); e.g., there is no method for the selector
10184: @code{draw} in the class @code{graphical}.
1.21 crook 10185:
1.26 crook 10186: For concrete graphical objects, we define child classes of the
10187: class @code{graphical}, e.g.:
1.21 crook 10188:
1.26 crook 10189: @cindex @code{overrides} usage
10190: @cindex @code{field} usage in class definition
10191: @example
10192: graphical class \ "graphical" is the parent class
10193: cell% field circle-radius
1.21 crook 10194:
1.26 crook 10195: :noname ( x y circle -- )
10196: circle-radius @@ draw-circle ;
10197: overrides draw
1.21 crook 10198:
1.26 crook 10199: :noname ( n-radius circle -- )
10200: circle-radius ! ;
10201: overrides construct
1.21 crook 10202:
1.26 crook 10203: end-class circle
1.21 crook 10204: @end example
10205:
1.26 crook 10206: Here we define a class @code{circle} as a child of @code{graphical},
10207: with field @code{circle-radius} (which behaves just like a field
10208: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10209: for the selectors @code{draw} and @code{construct} (@code{construct} is
10210: defined in @code{object}, the parent class of @code{graphical}).
1.21 crook 10211:
1.26 crook 10212: Now we can create a circle on the heap (i.e.,
10213: @code{allocate}d memory) with:
1.21 crook 10214:
1.26 crook 10215: @cindex @code{heap-new} usage
1.21 crook 10216: @example
1.26 crook 10217: 50 circle heap-new constant my-circle
10218: @end example
1.21 crook 10219:
1.26 crook 10220: @noindent
10221: @code{heap-new} invokes @code{construct}, thus
10222: initializing the field @code{circle-radius} with 50. We can draw
10223: this new circle at (100,100) with:
1.21 crook 10224:
1.26 crook 10225: @example
10226: 100 100 my-circle draw
1.21 crook 10227: @end example
10228:
1.26 crook 10229: @cindex selector invocation, restrictions
10230: @cindex class definition, restrictions
10231: Note: You can only invoke a selector if the object on the TOS
10232: (the receiving object) belongs to the class where the selector was
10233: defined or one of its descendents; e.g., you can invoke
10234: @code{draw} only for objects belonging to @code{graphical}
10235: or its descendents (e.g., @code{circle}). Immediately before
10236: @code{end-class}, the search order has to be the same as
10237: immediately after @code{class}.
1.21 crook 10238:
1.26 crook 10239: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10240: @subsubsection The @file{object.fs} base class
10241: @cindex @code{object} class
1.21 crook 10242:
1.26 crook 10243: When you define a class, you have to specify a parent class. So how do
10244: you start defining classes? There is one class available from the start:
10245: @code{object}. It is ancestor for all classes and so is the
10246: only class that has no parent. It has two selectors: @code{construct}
10247: and @code{print}.
1.21 crook 10248:
1.26 crook 10249: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10250: @subsubsection Creating objects
10251: @cindex creating objects
10252: @cindex object creation
10253: @cindex object allocation options
1.21 crook 10254:
1.26 crook 10255: @cindex @code{heap-new} discussion
10256: @cindex @code{dict-new} discussion
10257: @cindex @code{construct} discussion
10258: You can create and initialize an object of a class on the heap with
10259: @code{heap-new} ( ... class -- object ) and in the dictionary
10260: (allocation with @code{allot}) with @code{dict-new} (
10261: ... class -- object ). Both words invoke @code{construct}, which
10262: consumes the stack items indicated by "..." above.
1.21 crook 10263:
1.26 crook 10264: @cindex @code{init-object} discussion
10265: @cindex @code{class-inst-size} discussion
10266: If you want to allocate memory for an object yourself, you can get its
10267: alignment and size with @code{class-inst-size 2@@} ( class --
10268: align size ). Once you have memory for an object, you can initialize
10269: it with @code{init-object} ( ... class object -- );
10270: @code{construct} does only a part of the necessary work.
1.21 crook 10271:
1.26 crook 10272: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10273: @subsubsection Object-Oriented Programming Style
10274: @cindex object-oriented programming style
1.47 crook 10275: @cindex programming style, object-oriented
1.21 crook 10276:
1.26 crook 10277: This section is not exhaustive.
1.1 anton 10278:
1.26 crook 10279: @cindex stack effects of selectors
10280: @cindex selectors and stack effects
10281: In general, it is a good idea to ensure that all methods for the
10282: same selector have the same stack effect: when you invoke a selector,
10283: you often have no idea which method will be invoked, so, unless all
10284: methods have the same stack effect, you will not know the stack effect
10285: of the selector invocation.
1.21 crook 10286:
1.26 crook 10287: One exception to this rule is methods for the selector
10288: @code{construct}. We know which method is invoked, because we
10289: specify the class to be constructed at the same place. Actually, I
10290: defined @code{construct} as a selector only to give the users a
10291: convenient way to specify initialization. The way it is used, a
10292: mechanism different from selector invocation would be more natural
10293: (but probably would take more code and more space to explain).
1.21 crook 10294:
1.26 crook 10295: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10296: @subsubsection Class Binding
10297: @cindex class binding
10298: @cindex early binding
1.21 crook 10299:
1.26 crook 10300: @cindex late binding
10301: Normal selector invocations determine the method at run-time depending
10302: on the class of the receiving object. This run-time selection is called
1.29 crook 10303: @i{late binding}.
1.21 crook 10304:
1.26 crook 10305: Sometimes it's preferable to invoke a different method. For example,
10306: you might want to use the simple method for @code{print}ing
10307: @code{object}s instead of the possibly long-winded @code{print} method
10308: of the receiver class. You can achieve this by replacing the invocation
10309: of @code{print} with:
1.21 crook 10310:
1.26 crook 10311: @cindex @code{[bind]} usage
10312: @example
10313: [bind] object print
1.21 crook 10314: @end example
10315:
1.26 crook 10316: @noindent
10317: in compiled code or:
1.21 crook 10318:
1.26 crook 10319: @cindex @code{bind} usage
1.21 crook 10320: @example
1.26 crook 10321: bind object print
1.21 crook 10322: @end example
10323:
1.26 crook 10324: @cindex class binding, alternative to
10325: @noindent
10326: in interpreted code. Alternatively, you can define the method with a
10327: name (e.g., @code{print-object}), and then invoke it through the
10328: name. Class binding is just a (often more convenient) way to achieve
10329: the same effect; it avoids name clutter and allows you to invoke
10330: methods directly without naming them first.
10331:
10332: @cindex superclass binding
10333: @cindex parent class binding
10334: A frequent use of class binding is this: When we define a method
10335: for a selector, we often want the method to do what the selector does
10336: in the parent class, and a little more. There is a special word for
10337: this purpose: @code{[parent]}; @code{[parent]
10338: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10339: selector}}, where @code{@emph{parent}} is the parent
10340: class of the current class. E.g., a method definition might look like:
1.21 crook 10341:
1.26 crook 10342: @cindex @code{[parent]} usage
1.21 crook 10343: @example
1.26 crook 10344: :noname
10345: dup [parent] foo \ do parent's foo on the receiving object
10346: ... \ do some more
10347: ; overrides foo
1.21 crook 10348: @end example
10349:
1.26 crook 10350: @cindex class binding as optimization
10351: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10352: March 1997), Andrew McKewan presents class binding as an optimization
10353: technique. I recommend not using it for this purpose unless you are in
10354: an emergency. Late binding is pretty fast with this model anyway, so the
10355: benefit of using class binding is small; the cost of using class binding
10356: where it is not appropriate is reduced maintainability.
1.21 crook 10357:
1.26 crook 10358: While we are at programming style questions: You should bind
10359: selectors only to ancestor classes of the receiving object. E.g., say,
10360: you know that the receiving object is of class @code{foo} or its
10361: descendents; then you should bind only to @code{foo} and its
10362: ancestors.
1.21 crook 10363:
1.26 crook 10364: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10365: @subsubsection Method conveniences
10366: @cindex method conveniences
1.1 anton 10367:
1.26 crook 10368: In a method you usually access the receiving object pretty often. If
10369: you define the method as a plain colon definition (e.g., with
10370: @code{:noname}), you may have to do a lot of stack
10371: gymnastics. To avoid this, you can define the method with @code{m:
10372: ... ;m}. E.g., you could define the method for
10373: @code{draw}ing a @code{circle} with
1.20 pazsan 10374:
1.26 crook 10375: @cindex @code{this} usage
10376: @cindex @code{m:} usage
10377: @cindex @code{;m} usage
10378: @example
10379: m: ( x y circle -- )
10380: ( x y ) this circle-radius @@ draw-circle ;m
10381: @end example
1.20 pazsan 10382:
1.26 crook 10383: @cindex @code{exit} in @code{m: ... ;m}
10384: @cindex @code{exitm} discussion
10385: @cindex @code{catch} in @code{m: ... ;m}
10386: When this method is executed, the receiver object is removed from the
10387: stack; you can access it with @code{this} (admittedly, in this
10388: example the use of @code{m: ... ;m} offers no advantage). Note
10389: that I specify the stack effect for the whole method (i.e. including
10390: the receiver object), not just for the code between @code{m:}
10391: and @code{;m}. You cannot use @code{exit} in
10392: @code{m:...;m}; instead, use
10393: @code{exitm}.@footnote{Moreover, for any word that calls
10394: @code{catch} and was defined before loading
10395: @code{objects.fs}, you have to redefine it like I redefined
10396: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.20 pazsan 10397:
1.26 crook 10398: @cindex @code{inst-var} usage
10399: You will frequently use sequences of the form @code{this
10400: @emph{field}} (in the example above: @code{this
10401: circle-radius}). If you use the field only in this way, you can
10402: define it with @code{inst-var} and eliminate the
10403: @code{this} before the field name. E.g., the @code{circle}
10404: class above could also be defined with:
1.20 pazsan 10405:
1.26 crook 10406: @example
10407: graphical class
10408: cell% inst-var radius
1.20 pazsan 10409:
1.26 crook 10410: m: ( x y circle -- )
10411: radius @@ draw-circle ;m
10412: overrides draw
1.20 pazsan 10413:
1.26 crook 10414: m: ( n-radius circle -- )
10415: radius ! ;m
10416: overrides construct
1.12 anton 10417:
1.26 crook 10418: end-class circle
10419: @end example
1.12 anton 10420:
1.26 crook 10421: @code{radius} can only be used in @code{circle} and its
10422: descendent classes and inside @code{m:...;m}.
1.12 anton 10423:
1.26 crook 10424: @cindex @code{inst-value} usage
10425: You can also define fields with @code{inst-value}, which is
10426: to @code{inst-var} what @code{value} is to
10427: @code{variable}. You can change the value of such a field with
10428: @code{[to-inst]}. E.g., we could also define the class
10429: @code{circle} like this:
1.12 anton 10430:
1.26 crook 10431: @example
10432: graphical class
10433: inst-value radius
1.12 anton 10434:
1.26 crook 10435: m: ( x y circle -- )
10436: radius draw-circle ;m
10437: overrides draw
1.12 anton 10438:
1.26 crook 10439: m: ( n-radius circle -- )
10440: [to-inst] radius ;m
10441: overrides construct
1.21 crook 10442:
1.26 crook 10443: end-class circle
1.12 anton 10444: @end example
10445:
1.38 anton 10446: Finally, you can define named methods with @code{:m}. One use of this
10447: feature is the definition of words that occur only in one class and are
10448: not intended to be overridden, but which still need method context
10449: (e.g., for accessing @code{inst-var}s). Another use is for methods that
10450: would be bound frequently, if defined anonymously.
10451:
1.12 anton 10452:
1.37 anton 10453: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
1.26 crook 10454: @subsubsection Classes and Scoping
10455: @cindex classes and scoping
10456: @cindex scoping and classes
1.12 anton 10457:
1.26 crook 10458: Inheritance is frequent, unlike structure extension. This exacerbates
10459: the problem with the field name convention (@pxref{Structure Naming
10460: Convention}): One always has to remember in which class the field was
10461: originally defined; changing a part of the class structure would require
10462: changes for renaming in otherwise unaffected code.
1.12 anton 10463:
1.26 crook 10464: @cindex @code{inst-var} visibility
10465: @cindex @code{inst-value} visibility
10466: To solve this problem, I added a scoping mechanism (which was not in my
10467: original charter): A field defined with @code{inst-var} (or
10468: @code{inst-value}) is visible only in the class where it is defined and in
10469: the descendent classes of this class. Using such fields only makes
10470: sense in @code{m:}-defined methods in these classes anyway.
1.12 anton 10471:
1.26 crook 10472: This scoping mechanism allows us to use the unadorned field name,
10473: because name clashes with unrelated words become much less likely.
1.12 anton 10474:
1.26 crook 10475: @cindex @code{protected} discussion
10476: @cindex @code{private} discussion
10477: Once we have this mechanism, we can also use it for controlling the
10478: visibility of other words: All words defined after
10479: @code{protected} are visible only in the current class and its
10480: descendents. @code{public} restores the compilation
10481: (i.e. @code{current}) word list that was in effect before. If you
10482: have several @code{protected}s without an intervening
10483: @code{public} or @code{set-current}, @code{public}
10484: will restore the compilation word list in effect before the first of
10485: these @code{protected}s.
1.12 anton 10486:
1.37 anton 10487: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10488: @subsubsection Dividing classes
10489: @cindex Dividing classes
10490: @cindex @code{methods}...@code{end-methods}
10491:
10492: You may want to do the definition of methods separate from the
10493: definition of the class, its selectors, fields, and instance variables,
10494: i.e., separate the implementation from the definition. You can do this
10495: in the following way:
10496:
10497: @example
10498: graphical class
10499: inst-value radius
10500: end-class circle
10501:
10502: ... \ do some other stuff
10503:
10504: circle methods \ now we are ready
10505:
10506: m: ( x y circle -- )
10507: radius draw-circle ;m
10508: overrides draw
10509:
10510: m: ( n-radius circle -- )
10511: [to-inst] radius ;m
10512: overrides construct
10513:
10514: end-methods
10515: @end example
10516:
10517: You can use several @code{methods}...@code{end-methods} sections. The
10518: only things you can do to the class in these sections are: defining
10519: methods, and overriding the class's selectors. You must not define new
10520: selectors or fields.
10521:
10522: Note that you often have to override a selector before using it. In
10523: particular, you usually have to override @code{construct} with a new
10524: method before you can invoke @code{heap-new} and friends. E.g., you
10525: must not create a circle before the @code{overrides construct} sequence
10526: in the example above.
10527:
10528: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
1.26 crook 10529: @subsubsection Object Interfaces
10530: @cindex object interfaces
10531: @cindex interfaces for objects
1.12 anton 10532:
1.26 crook 10533: In this model you can only call selectors defined in the class of the
10534: receiving objects or in one of its ancestors. If you call a selector
10535: with a receiving object that is not in one of these classes, the
10536: result is undefined; if you are lucky, the program crashes
10537: immediately.
1.12 anton 10538:
1.26 crook 10539: @cindex selectors common to hardly-related classes
10540: Now consider the case when you want to have a selector (or several)
10541: available in two classes: You would have to add the selector to a
10542: common ancestor class, in the worst case to @code{object}. You
10543: may not want to do this, e.g., because someone else is responsible for
10544: this ancestor class.
1.12 anton 10545:
1.26 crook 10546: The solution for this problem is interfaces. An interface is a
10547: collection of selectors. If a class implements an interface, the
10548: selectors become available to the class and its descendents. A class
10549: can implement an unlimited number of interfaces. For the problem
10550: discussed above, we would define an interface for the selector(s), and
10551: both classes would implement the interface.
1.12 anton 10552:
1.26 crook 10553: As an example, consider an interface @code{storage} for
10554: writing objects to disk and getting them back, and a class
10555: @code{foo} that implements it. The code would look like this:
1.12 anton 10556:
1.26 crook 10557: @cindex @code{interface} usage
10558: @cindex @code{end-interface} usage
10559: @cindex @code{implementation} usage
10560: @example
10561: interface
10562: selector write ( file object -- )
10563: selector read1 ( file object -- )
10564: end-interface storage
1.12 anton 10565:
1.26 crook 10566: bar class
10567: storage implementation
1.12 anton 10568:
1.26 crook 10569: ... overrides write
1.37 anton 10570: ... overrides read1
1.26 crook 10571: ...
10572: end-class foo
1.12 anton 10573: @end example
10574:
1.26 crook 10575: @noindent
1.29 crook 10576: (I would add a word @code{read} @i{( file -- object )} that uses
1.26 crook 10577: @code{read1} internally, but that's beyond the point illustrated
10578: here.)
1.12 anton 10579:
1.26 crook 10580: Note that you cannot use @code{protected} in an interface; and
10581: of course you cannot define fields.
1.12 anton 10582:
1.26 crook 10583: In the Neon model, all selectors are available for all classes;
10584: therefore it does not need interfaces. The price you pay in this model
10585: is slower late binding, and therefore, added complexity to avoid late
10586: binding.
1.12 anton 10587:
1.26 crook 10588: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10589: @subsubsection @file{objects.fs} Implementation
10590: @cindex @file{objects.fs} implementation
1.12 anton 10591:
1.26 crook 10592: @cindex @code{object-map} discussion
10593: An object is a piece of memory, like one of the data structures
10594: described with @code{struct...end-struct}. It has a field
10595: @code{object-map} that points to the method map for the object's
10596: class.
1.12 anton 10597:
1.26 crook 10598: @cindex method map
10599: @cindex virtual function table
10600: The @emph{method map}@footnote{This is Self terminology; in C++
10601: terminology: virtual function table.} is an array that contains the
1.29 crook 10602: execution tokens (@i{xt}s) of the methods for the object's class. Each
1.26 crook 10603: selector contains an offset into a method map.
1.12 anton 10604:
1.26 crook 10605: @cindex @code{selector} implementation, class
10606: @code{selector} is a defining word that uses
10607: @code{CREATE} and @code{DOES>}. The body of the
1.44 crook 10608: selector contains the offset; the @code{DOES>} action for a
1.26 crook 10609: class selector is, basically:
1.21 crook 10610:
1.26 crook 10611: @example
10612: ( object addr ) @@ over object-map @@ + @@ execute
10613: @end example
1.12 anton 10614:
1.26 crook 10615: Since @code{object-map} is the first field of the object, it
10616: does not generate any code. As you can see, calling a selector has a
10617: small, constant cost.
1.12 anton 10618:
1.26 crook 10619: @cindex @code{current-interface} discussion
10620: @cindex class implementation and representation
10621: A class is basically a @code{struct} combined with a method
10622: map. During the class definition the alignment and size of the class
10623: are passed on the stack, just as with @code{struct}s, so
10624: @code{field} can also be used for defining class
10625: fields. However, passing more items on the stack would be
10626: inconvenient, so @code{class} builds a data structure in memory,
10627: which is accessed through the variable
10628: @code{current-interface}. After its definition is complete, the
10629: class is represented on the stack by a pointer (e.g., as parameter for
10630: a child class definition).
1.1 anton 10631:
1.26 crook 10632: A new class starts off with the alignment and size of its parent,
10633: and a copy of the parent's method map. Defining new fields extends the
10634: size and alignment; likewise, defining new selectors extends the
1.29 crook 10635: method map. @code{overrides} just stores a new @i{xt} in the method
1.26 crook 10636: map at the offset given by the selector.
1.20 pazsan 10637:
1.26 crook 10638: @cindex class binding, implementation
1.29 crook 10639: Class binding just gets the @i{xt} at the offset given by the selector
1.26 crook 10640: from the class's method map and @code{compile,}s (in the case of
10641: @code{[bind]}) it.
1.21 crook 10642:
1.26 crook 10643: @cindex @code{this} implementation
10644: @cindex @code{catch} and @code{this}
10645: @cindex @code{this} and @code{catch}
10646: I implemented @code{this} as a @code{value}. At the
10647: start of an @code{m:...;m} method the old @code{this} is
10648: stored to the return stack and restored at the end; and the object on
10649: the TOS is stored @code{TO this}. This technique has one
10650: disadvantage: If the user does not leave the method via
10651: @code{;m}, but via @code{throw} or @code{exit},
10652: @code{this} is not restored (and @code{exit} may
10653: crash). To deal with the @code{throw} problem, I have redefined
10654: @code{catch} to save and restore @code{this}; the same
10655: should be done with any word that can catch an exception. As for
10656: @code{exit}, I simply forbid it (as a replacement, there is
10657: @code{exitm}).
1.21 crook 10658:
1.26 crook 10659: @cindex @code{inst-var} implementation
10660: @code{inst-var} is just the same as @code{field}, with
10661: a different @code{DOES>} action:
10662: @example
10663: @@ this +
10664: @end example
10665: Similar for @code{inst-value}.
1.21 crook 10666:
1.26 crook 10667: @cindex class scoping implementation
10668: Each class also has a word list that contains the words defined with
10669: @code{inst-var} and @code{inst-value}, and its protected
10670: words. It also has a pointer to its parent. @code{class} pushes
10671: the word lists of the class and all its ancestors onto the search order stack,
10672: and @code{end-class} drops them.
1.21 crook 10673:
1.26 crook 10674: @cindex interface implementation
10675: An interface is like a class without fields, parent and protected
10676: words; i.e., it just has a method map. If a class implements an
10677: interface, its method map contains a pointer to the method map of the
10678: interface. The positive offsets in the map are reserved for class
10679: methods, therefore interface map pointers have negative
10680: offsets. Interfaces have offsets that are unique throughout the
10681: system, unlike class selectors, whose offsets are only unique for the
10682: classes where the selector is available (invokable).
1.21 crook 10683:
1.26 crook 10684: This structure means that interface selectors have to perform one
10685: indirection more than class selectors to find their method. Their body
10686: contains the interface map pointer offset in the class method map, and
10687: the method offset in the interface method map. The
10688: @code{does>} action for an interface selector is, basically:
1.21 crook 10689:
10690: @example
1.26 crook 10691: ( object selector-body )
10692: 2dup selector-interface @@ ( object selector-body object interface-offset )
10693: swap object-map @@ + @@ ( object selector-body map )
10694: swap selector-offset @@ + @@ execute
1.21 crook 10695: @end example
10696:
1.26 crook 10697: where @code{object-map} and @code{selector-offset} are
10698: first fields and generate no code.
10699:
10700: As a concrete example, consider the following code:
1.21 crook 10701:
1.26 crook 10702: @example
10703: interface
10704: selector if1sel1
10705: selector if1sel2
10706: end-interface if1
1.21 crook 10707:
1.26 crook 10708: object class
10709: if1 implementation
10710: selector cl1sel1
10711: cell% inst-var cl1iv1
1.21 crook 10712:
1.26 crook 10713: ' m1 overrides construct
10714: ' m2 overrides if1sel1
10715: ' m3 overrides if1sel2
10716: ' m4 overrides cl1sel2
10717: end-class cl1
1.21 crook 10718:
1.26 crook 10719: create obj1 object dict-new drop
10720: create obj2 cl1 dict-new drop
10721: @end example
1.21 crook 10722:
1.26 crook 10723: The data structure created by this code (including the data structure
10724: for @code{object}) is shown in the <a
10725: href="objects-implementation.eps">figure</a>, assuming a cell size of 4.
1.29 crook 10726: @comment TODO add this diagram..
1.21 crook 10727:
1.26 crook 10728: @node Objects Glossary, , Objects Implementation, Objects
10729: @subsubsection @file{objects.fs} Glossary
10730: @cindex @file{objects.fs} Glossary
1.21 crook 10731:
1.44 crook 10732:
1.26 crook 10733: doc---objects-bind
10734: doc---objects-<bind>
10735: doc---objects-bind'
10736: doc---objects-[bind]
10737: doc---objects-class
10738: doc---objects-class->map
10739: doc---objects-class-inst-size
10740: doc---objects-class-override!
10741: doc---objects-construct
10742: doc---objects-current'
10743: doc---objects-[current]
10744: doc---objects-current-interface
10745: doc---objects-dict-new
10746: doc---objects-drop-order
10747: doc---objects-end-class
10748: doc---objects-end-class-noname
10749: doc---objects-end-interface
10750: doc---objects-end-interface-noname
1.37 anton 10751: doc---objects-end-methods
1.26 crook 10752: doc---objects-exitm
10753: doc---objects-heap-new
10754: doc---objects-implementation
10755: doc---objects-init-object
10756: doc---objects-inst-value
10757: doc---objects-inst-var
10758: doc---objects-interface
1.38 anton 10759: doc---objects-m:
10760: doc---objects-:m
1.26 crook 10761: doc---objects-;m
10762: doc---objects-method
1.37 anton 10763: doc---objects-methods
1.26 crook 10764: doc---objects-object
10765: doc---objects-overrides
10766: doc---objects-[parent]
10767: doc---objects-print
10768: doc---objects-protected
10769: doc---objects-public
10770: doc---objects-push-order
10771: doc---objects-selector
10772: doc---objects-this
10773: doc---objects-<to-inst>
10774: doc---objects-[to-inst]
10775: doc---objects-to-this
10776: doc---objects-xt-new
1.21 crook 10777:
1.44 crook 10778:
1.26 crook 10779: @c -------------------------------------------------------------
10780: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10781: @subsection The @file{oof.fs} model
10782: @cindex oof
10783: @cindex object-oriented programming
1.21 crook 10784:
1.26 crook 10785: @cindex @file{objects.fs}
10786: @cindex @file{oof.fs}
1.21 crook 10787:
1.26 crook 10788: This section describes the @file{oof.fs} package.
1.21 crook 10789:
1.26 crook 10790: The package described in this section has been used in bigFORTH since 1991, and
10791: used for two large applications: a chromatographic system used to
10792: create new medicaments, and a graphic user interface library (MINOS).
1.21 crook 10793:
1.26 crook 10794: You can find a description (in German) of @file{oof.fs} in @cite{Object
10795: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10796: 10(2), 1994.
1.21 crook 10797:
1.26 crook 10798: @menu
10799: * Properties of the OOF model::
10800: * Basic OOF Usage::
10801: * The OOF base class::
10802: * Class Declaration::
10803: * Class Implementation::
10804: @end menu
1.21 crook 10805:
1.26 crook 10806: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10807: @subsubsection Properties of the @file{oof.fs} model
10808: @cindex @file{oof.fs} properties
1.21 crook 10809:
1.26 crook 10810: @itemize @bullet
10811: @item
10812: This model combines object oriented programming with information
10813: hiding. It helps you writing large application, where scoping is
10814: necessary, because it provides class-oriented scoping.
1.21 crook 10815:
1.26 crook 10816: @item
10817: Named objects, object pointers, and object arrays can be created,
10818: selector invocation uses the ``object selector'' syntax. Selector invocation
10819: to objects and/or selectors on the stack is a bit less convenient, but
10820: possible.
1.21 crook 10821:
1.26 crook 10822: @item
10823: Selector invocation and instance variable usage of the active object is
10824: straightforward, since both make use of the active object.
1.21 crook 10825:
1.26 crook 10826: @item
10827: Late binding is efficient and easy to use.
1.21 crook 10828:
1.26 crook 10829: @item
10830: State-smart objects parse selectors. However, extensibility is provided
10831: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.21 crook 10832:
10833: @item
1.26 crook 10834: An implementation in ANS Forth is available.
10835:
1.21 crook 10836: @end itemize
10837:
10838:
1.26 crook 10839: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10840: @subsubsection Basic @file{oof.fs} Usage
10841: @cindex @file{oof.fs} usage
10842:
10843: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.21 crook 10844:
1.26 crook 10845: You can define a class for graphical objects like this:
1.21 crook 10846:
1.26 crook 10847: @cindex @code{class} usage
10848: @cindex @code{class;} usage
10849: @cindex @code{method} usage
10850: @example
10851: object class graphical \ "object" is the parent class
10852: method draw ( x y graphical -- )
10853: class;
10854: @end example
1.21 crook 10855:
1.26 crook 10856: This code defines a class @code{graphical} with an
10857: operation @code{draw}. We can perform the operation
10858: @code{draw} on any @code{graphical} object, e.g.:
1.21 crook 10859:
1.26 crook 10860: @example
10861: 100 100 t-rex draw
10862: @end example
1.21 crook 10863:
1.26 crook 10864: @noindent
10865: where @code{t-rex} is an object or object pointer, created with e.g.
10866: @code{graphical : t-rex}.
1.21 crook 10867:
1.26 crook 10868: @cindex abstract class
10869: How do we create a graphical object? With the present definitions,
10870: we cannot create a useful graphical object. The class
10871: @code{graphical} describes graphical objects in general, but not
10872: any concrete graphical object type (C++ users would call it an
10873: @emph{abstract class}); e.g., there is no method for the selector
10874: @code{draw} in the class @code{graphical}.
1.21 crook 10875:
1.26 crook 10876: For concrete graphical objects, we define child classes of the
10877: class @code{graphical}, e.g.:
1.21 crook 10878:
10879: @example
1.26 crook 10880: graphical class circle \ "graphical" is the parent class
10881: cell var circle-radius
10882: how:
10883: : draw ( x y -- )
10884: circle-radius @@ draw-circle ;
10885:
10886: : init ( n-radius -- (
10887: circle-radius ! ;
10888: class;
10889: @end example
10890:
10891: Here we define a class @code{circle} as a child of @code{graphical},
10892: with a field @code{circle-radius}; it defines new methods for the
10893: selectors @code{draw} and @code{init} (@code{init} is defined in
10894: @code{object}, the parent class of @code{graphical}).
1.21 crook 10895:
1.26 crook 10896: Now we can create a circle in the dictionary with:
1.21 crook 10897:
1.26 crook 10898: @example
10899: 50 circle : my-circle
1.21 crook 10900: @end example
10901:
1.26 crook 10902: @noindent
10903: @code{:} invokes @code{init}, thus initializing the field
10904: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10905: with:
1.21 crook 10906:
10907: @example
1.26 crook 10908: 100 100 my-circle draw
1.21 crook 10909: @end example
10910:
1.26 crook 10911: @cindex selector invocation, restrictions
10912: @cindex class definition, restrictions
10913: Note: You can only invoke a selector if the receiving object belongs to
10914: the class where the selector was defined or one of its descendents;
10915: e.g., you can invoke @code{draw} only for objects belonging to
10916: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10917: mechanism will check if you try to invoke a selector that is not
10918: defined in this class hierarchy, so you'll get an error at compilation
10919: time.
10920:
10921:
10922: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10923: @subsubsection The @file{oof.fs} base class
10924: @cindex @file{oof.fs} base class
10925:
10926: When you define a class, you have to specify a parent class. So how do
10927: you start defining classes? There is one class available from the start:
10928: @code{object}. You have to use it as ancestor for all classes. It is the
10929: only class that has no parent. Classes are also objects, except that
10930: they don't have instance variables; class manipulation such as
10931: inheritance or changing definitions of a class is handled through
10932: selectors of the class @code{object}.
10933:
10934: @code{object} provides a number of selectors:
10935:
1.21 crook 10936: @itemize @bullet
10937: @item
1.26 crook 10938: @code{class} for subclassing, @code{definitions} to add definitions
10939: later on, and @code{class?} to get type informations (is the class a
10940: subclass of the class passed on the stack?).
1.44 crook 10941:
1.26 crook 10942: doc---object-class
10943: doc---object-definitions
10944: doc---object-class?
10945:
1.44 crook 10946:
1.21 crook 10947: @item
1.26 crook 10948: @code{init} and @code{dispose} as constructor and destructor of the
10949: object. @code{init} is invocated after the object's memory is allocated,
10950: while @code{dispose} also handles deallocation. Thus if you redefine
10951: @code{dispose}, you have to call the parent's dispose with @code{super
10952: dispose}, too.
1.44 crook 10953:
1.26 crook 10954: doc---object-init
10955: doc---object-dispose
10956:
1.44 crook 10957:
1.21 crook 10958: @item
1.26 crook 10959: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10960: @code{[]} to create named and unnamed objects and object arrays or
10961: object pointers.
1.44 crook 10962:
1.26 crook 10963: doc---object-new
10964: doc---object-new[]
10965: doc---object-:
10966: doc---object-ptr
10967: doc---object-asptr
10968: doc---object-[]
1.21 crook 10969:
1.44 crook 10970:
1.26 crook 10971: @item
10972: @code{::} and @code{super} for explicit scoping. You should use explicit
10973: scoping only for super classes or classes with the same set of instance
10974: variables. Explicitly-scoped selectors use early binding.
1.44 crook 10975:
1.26 crook 10976: doc---object-::
10977: doc---object-super
1.21 crook 10978:
1.44 crook 10979:
1.26 crook 10980: @item
10981: @code{self} to get the address of the object
1.44 crook 10982:
1.26 crook 10983: doc---object-self
1.21 crook 10984:
1.44 crook 10985:
1.21 crook 10986: @item
1.26 crook 10987: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
10988: pointers and instance defers.
1.44 crook 10989:
1.26 crook 10990: doc---object-bind
10991: doc---object-bound
10992: doc---object-link
10993: doc---object-is
10994:
1.44 crook 10995:
1.21 crook 10996: @item
1.26 crook 10997: @code{'} to obtain selector tokens, @code{send} to invocate selectors
10998: form the stack, and @code{postpone} to generate selector invocation code.
1.44 crook 10999:
1.26 crook 11000: doc---object-'
11001: doc---object-postpone
11002:
1.44 crook 11003:
1.21 crook 11004: @item
1.26 crook 11005: @code{with} and @code{endwith} to select the active object from the
11006: stack, and enable its scope. Using @code{with} and @code{endwith}
11007: also allows you to create code using selector @code{postpone} without being
11008: trapped by the state-smart objects.
1.44 crook 11009:
1.26 crook 11010: doc---object-with
11011: doc---object-endwith
11012:
1.44 crook 11013:
1.21 crook 11014: @end itemize
11015:
1.26 crook 11016: @node Class Declaration, Class Implementation, The OOF base class, OOF
11017: @subsubsection Class Declaration
11018: @cindex class declaration
11019:
11020: @itemize @bullet
11021: @item
11022: Instance variables
1.44 crook 11023:
1.26 crook 11024: doc---oof-var
1.21 crook 11025:
1.44 crook 11026:
1.26 crook 11027: @item
11028: Object pointers
1.44 crook 11029:
1.26 crook 11030: doc---oof-ptr
11031: doc---oof-asptr
1.21 crook 11032:
1.44 crook 11033:
1.26 crook 11034: @item
11035: Instance defers
1.44 crook 11036:
1.26 crook 11037: doc---oof-defer
1.21 crook 11038:
1.44 crook 11039:
1.26 crook 11040: @item
11041: Method selectors
1.44 crook 11042:
1.26 crook 11043: doc---oof-early
11044: doc---oof-method
1.21 crook 11045:
1.44 crook 11046:
1.26 crook 11047: @item
11048: Class-wide variables
1.44 crook 11049:
1.26 crook 11050: doc---oof-static
1.21 crook 11051:
1.44 crook 11052:
1.26 crook 11053: @item
11054: End declaration
1.44 crook 11055:
1.26 crook 11056: doc---oof-how:
11057: doc---oof-class;
1.21 crook 11058:
1.44 crook 11059:
1.26 crook 11060: @end itemize
1.21 crook 11061:
1.26 crook 11062: @c -------------------------------------------------------------
11063: @node Class Implementation, , Class Declaration, OOF
11064: @subsubsection Class Implementation
11065: @cindex class implementation
1.21 crook 11066:
1.26 crook 11067: @c -------------------------------------------------------------
11068: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11069: @subsection The @file{mini-oof.fs} model
11070: @cindex mini-oof
1.1 anton 11071:
1.26 crook 11072: Gforth's third object oriented Forth package is a 12-liner. It uses a
11073: mixture of the @file{object.fs} and the @file{oof.fs} syntax,
11074: and reduces to the bare minimum of features. This is based on a posting
11075: of Bernd Paysan in comp.arch.
1.1 anton 11076:
11077: @menu
1.48 anton 11078: * Basic Mini-OOF Usage::
11079: * Mini-OOF Example::
11080: * Mini-OOF Implementation::
11081: * Comparison with other object models::
1.1 anton 11082: @end menu
11083:
1.26 crook 11084: @c -------------------------------------------------------------
1.48 anton 11085: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
1.26 crook 11086: @subsubsection Basic @file{mini-oof.fs} Usage
11087: @cindex mini-oof usage
1.1 anton 11088:
1.28 crook 11089: There is a base class (@code{class}, which allocates one cell for the
11090: object pointer) plus seven other words: to define a method, a variable,
11091: a class; to end a class, to resolve binding, to allocate an object and
11092: to compile a class method.
1.26 crook 11093: @comment TODO better description of the last one
1.1 anton 11094:
1.44 crook 11095:
1.26 crook 11096: doc-object
11097: doc-method
11098: doc-var
11099: doc-class
11100: doc-end-class
11101: doc-defines
11102: doc-new
11103: doc-::
1.1 anton 11104:
1.21 crook 11105:
1.44 crook 11106:
1.26 crook 11107: @c -------------------------------------------------------------
11108: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11109: @subsubsection Mini-OOF Example
11110: @cindex mini-oof example
1.21 crook 11111:
1.26 crook 11112: A short example shows how to use this package. This example, in slightly
11113: extended form, is supplied as @file{moof-exm.fs}
1.29 crook 11114: @comment TODO could flesh this out with some comments from the Forthwrite article
1.21 crook 11115:
1.26 crook 11116: @example
11117: object class
11118: method init
11119: method draw
11120: end-class graphical
11121: @end example
1.21 crook 11122:
1.26 crook 11123: This code defines a class @code{graphical} with an
11124: operation @code{draw}. We can perform the operation
11125: @code{draw} on any @code{graphical} object, e.g.:
1.1 anton 11126:
1.26 crook 11127: @example
11128: 100 100 t-rex draw
11129: @end example
1.1 anton 11130:
1.26 crook 11131: where @code{t-rex} is an object or object pointer, created with e.g.
11132: @code{graphical new Constant t-rex}.
1.1 anton 11133:
1.26 crook 11134: For concrete graphical objects, we define child classes of the
11135: class @code{graphical}, e.g.:
1.21 crook 11136:
11137: @example
1.26 crook 11138: graphical class
11139: cell var circle-radius
11140: end-class circle \ "graphical" is the parent class
1.21 crook 11141:
1.26 crook 11142: :noname ( x y -- )
11143: circle-radius @@ draw-circle ; circle defines draw
11144: :noname ( r -- )
11145: circle-radius ! ; circle defines init
1.21 crook 11146: @end example
11147:
1.26 crook 11148: There is no implicit init method, so we have to define one. The creation
11149: code of the object now has to call init explicitely.
1.21 crook 11150:
1.26 crook 11151: @example
11152: circle new Constant my-circle
11153: 50 my-circle init
11154: @end example
1.21 crook 11155:
1.26 crook 11156: It is also possible to add a function to create named objects with
11157: automatic call of @code{init}, given that all objects have @code{init}
11158: on the same place:
1.1 anton 11159:
11160: @example
1.26 crook 11161: : new: ( .. o "name" -- )
11162: new dup Constant init ;
11163: 80 circle new: large-circle
1.1 anton 11164: @end example
11165:
1.26 crook 11166: We can draw this new circle at (100,100) with:
1.1 anton 11167:
11168: @example
1.26 crook 11169: 100 100 my-circle draw
1.1 anton 11170: @end example
11171:
1.48 anton 11172: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
1.26 crook 11173: @subsubsection @file{mini-oof.fs} Implementation
1.1 anton 11174:
1.26 crook 11175: Object-oriented systems with late binding typically use a
11176: ``vtable''-approach: the first variable in each object is a pointer to a
11177: table, which contains the methods as function pointers. The vtable
11178: may also contain other information.
1.1 anton 11179:
1.26 crook 11180: So first, let's declare methods:
1.1 anton 11181:
1.26 crook 11182: @example
11183: : method ( m v -- m' v ) Create over , swap cell+ swap
11184: DOES> ( ... o -- ... ) @ over @ + @ execute ;
11185: @end example
1.1 anton 11186:
1.26 crook 11187: During method declaration, the number of methods and instance
11188: variables is on the stack (in address units). @code{method} creates
11189: one method and increments the method number. To execute a method, it
11190: takes the object, fetches the vtable pointer, adds the offset, and
1.29 crook 11191: executes the @i{xt} stored there. Each method takes the object it is
1.26 crook 11192: invoked from as top of stack parameter. The method itself should
11193: consume that object.
1.1 anton 11194:
1.26 crook 11195: Now, we also have to declare instance variables
1.21 crook 11196:
1.26 crook 11197: @example
11198: : var ( m v size -- m v' ) Create over , +
11199: DOES> ( o -- addr ) @ + ;
11200: @end example
1.21 crook 11201:
1.26 crook 11202: As before, a word is created with the current offset. Instance
11203: variables can have different sizes (cells, floats, doubles, chars), so
11204: all we do is take the size and add it to the offset. If your machine
11205: has alignment restrictions, put the proper @code{aligned} or
11206: @code{faligned} before the variable, to adjust the variable
11207: offset. That's why it is on the top of stack.
1.2 jwilke 11208:
1.26 crook 11209: We need a starting point (the base object) and some syntactic sugar:
1.21 crook 11210:
1.26 crook 11211: @example
11212: Create object 1 cells , 2 cells ,
11213: : class ( class -- class methods vars ) dup 2@ ;
11214: @end example
1.21 crook 11215:
1.26 crook 11216: For inheritance, the vtable of the parent object has to be
11217: copied when a new, derived class is declared. This gives all the
11218: methods of the parent class, which can be overridden, though.
1.21 crook 11219:
1.2 jwilke 11220: @example
1.26 crook 11221: : end-class ( class methods vars -- )
11222: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11223: cell+ dup cell+ r> rot @ 2 cells /string move ;
11224: @end example
11225:
11226: The first line creates the vtable, initialized with
11227: @code{noop}s. The second line is the inheritance mechanism, it
11228: copies the xts from the parent vtable.
1.2 jwilke 11229:
1.26 crook 11230: We still have no way to define new methods, let's do that now:
1.2 jwilke 11231:
1.26 crook 11232: @example
11233: : defines ( xt class -- ) ' >body @ + ! ;
1.2 jwilke 11234: @end example
11235:
1.26 crook 11236: To allocate a new object, we need a word, too:
1.2 jwilke 11237:
1.26 crook 11238: @example
11239: : new ( class -- o ) here over @ allot swap over ! ;
11240: @end example
1.2 jwilke 11241:
1.26 crook 11242: Sometimes derived classes want to access the method of the
11243: parent object. There are two ways to achieve this with Mini-OOF:
11244: first, you could use named words, and second, you could look up the
11245: vtable of the parent object.
1.2 jwilke 11246:
1.26 crook 11247: @example
11248: : :: ( class "name" -- ) ' >body @ + @ compile, ;
11249: @end example
1.2 jwilke 11250:
11251:
1.26 crook 11252: Nothing can be more confusing than a good example, so here is
11253: one. First let's declare a text object (called
11254: @code{button}), that stores text and position:
1.2 jwilke 11255:
1.26 crook 11256: @example
11257: object class
11258: cell var text
11259: cell var len
11260: cell var x
11261: cell var y
11262: method init
11263: method draw
11264: end-class button
11265: @end example
1.2 jwilke 11266:
1.26 crook 11267: @noindent
11268: Now, implement the two methods, @code{draw} and @code{init}:
1.2 jwilke 11269:
1.26 crook 11270: @example
11271: :noname ( o -- )
11272: >r r@ x @ r@ y @ at-xy r@ text @ r> len @ type ;
11273: button defines draw
11274: :noname ( addr u o -- )
11275: >r 0 r@ x ! 0 r@ y ! r@ len ! r> text ! ;
11276: button defines init
11277: @end example
1.2 jwilke 11278:
1.26 crook 11279: @noindent
11280: To demonstrate inheritance, we define a class @code{bold-button}, with no
11281: new data and no new methods:
1.2 jwilke 11282:
1.26 crook 11283: @example
11284: button class
11285: end-class bold-button
1.1 anton 11286:
1.26 crook 11287: : bold 27 emit ." [1m" ;
11288: : normal 27 emit ." [0m" ;
11289: @end example
1.1 anton 11290:
1.26 crook 11291: @noindent
11292: The class @code{bold-button} has a different draw method to
11293: @code{button}, but the new method is defined in terms of the draw method
11294: for @code{button}:
1.1 anton 11295:
1.26 crook 11296: @example
11297: :noname bold [ button :: draw ] normal ; bold-button defines draw
11298: @end example
1.1 anton 11299:
1.26 crook 11300: @noindent
11301: Finally, create two objects and apply methods:
1.1 anton 11302:
1.26 crook 11303: @example
11304: button new Constant foo
11305: s" thin foo" foo init
11306: page
11307: foo draw
11308: bold-button new Constant bar
11309: s" fat bar" bar init
11310: 1 bar y !
11311: bar draw
11312: @end example
1.1 anton 11313:
11314:
1.48 anton 11315: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11316: @subsection Comparison with other object models
1.26 crook 11317: @cindex comparison of object models
11318: @cindex object models, comparison
1.1 anton 11319:
1.26 crook 11320: Many object-oriented Forth extensions have been proposed (@cite{A survey
11321: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11322: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11323: relation of the object models described here to two well-known and two
11324: closely-related (by the use of method maps) models.
1.1 anton 11325:
1.26 crook 11326: @cindex Neon model
11327: The most popular model currently seems to be the Neon model (see
11328: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11329: 1997) by Andrew McKewan) but this model has a number of limitations
11330: @footnote{A longer version of this critique can be
11331: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11332: Dimensions, May 1997) by Anton Ertl.}:
1.1 anton 11333:
1.26 crook 11334: @itemize @bullet
11335: @item
1.48 anton 11336: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11337: to pass objects on the stack.
1.1 anton 11338:
1.26 crook 11339: @item
11340: It requires that the selector parses the input stream (at
11341: compile time); this leads to reduced extensibility and to bugs that are+
11342: hard to find.
1.1 anton 11343:
1.26 crook 11344: @item
11345: It allows using every selector to every object;
11346: this eliminates the need for classes, but makes it harder to create
11347: efficient implementations.
11348: @end itemize
1.1 anton 11349:
1.26 crook 11350: @cindex Pountain's object-oriented model
11351: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11352: Press, London, 1987) by Dick Pountain. However, it is not really about
11353: object-oriented programming, because it hardly deals with late
11354: binding. Instead, it focuses on features like information hiding and
11355: overloading that are characteristic of modular languages like Ada (83).
1.1 anton 11356:
1.26 crook 11357: @cindex Zsoter's object-oriented model
1.48 anton 11358: In @cite{Does late binding have to be slow?} (Forth Dimensions 18(1)
11359: 1996, pages 31-35) Andras Zsoter describes a model that makes heavy use
11360: of an active object (like @code{this} in @file{objects.fs}): The active
11361: object is not only used for accessing all fields, but also specifies the
11362: receiving object of every selector invocation; you have to change the
11363: active object explicitly with @code{@{ ... @}}, whereas in
11364: @file{objects.fs} it changes more or less implicitly at @code{m:
11365: ... ;m}. Such a change at the method entry point is unnecessary with the
11366: Zsoter's model, because the receiving object is the active object
11367: already. On the other hand, the explicit change is absolutely necessary
11368: in that model, because otherwise no one could ever change the active
11369: object. An ANS Forth implementation of this model is available at
11370: @uref{http://www.forth.org/fig/oopf.html}.
1.1 anton 11371:
1.26 crook 11372: @cindex @file{oof.fs}, differences to other models
11373: The @file{oof.fs} model combines information hiding and overloading
11374: resolution (by keeping names in various word lists) with object-oriented
11375: programming. It sets the active object implicitly on method entry, but
11376: also allows explicit changing (with @code{>o...o>} or with
11377: @code{with...endwith}). It uses parsing and state-smart objects and
11378: classes for resolving overloading and for early binding: the object or
11379: class parses the selector and determines the method from this. If the
11380: selector is not parsed by an object or class, it performs a call to the
11381: selector for the active object (late binding), like Zsoter's model.
11382: Fields are always accessed through the active object. The big
11383: disadvantage of this model is the parsing and the state-smartness, which
11384: reduces extensibility and increases the opportunities for subtle bugs;
11385: essentially, you are only safe if you never tick or @code{postpone} an
11386: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.1 anton 11387:
1.26 crook 11388: @cindex @file{mini-oof.fs}, differences to other models
1.48 anton 11389: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11390: version of the @file{objects.fs} model, but syntactically it is a
11391: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.1 anton 11392:
1.26 crook 11393: @c -------------------------------------------------------------
1.47 crook 11394: @node Passing Commands to the OS, Keeping track of Time, Object-oriented Forth, Words
1.21 crook 11395: @section Passing Commands to the Operating System
11396: @cindex operating system - passing commands
11397: @cindex shell commands
11398:
11399: Gforth allows you to pass an arbitrary string to the host operating
11400: system shell (if such a thing exists) for execution.
11401:
1.44 crook 11402:
1.21 crook 11403: doc-sh
11404: doc-system
11405: doc-$?
1.23 crook 11406: doc-getenv
1.21 crook 11407:
1.44 crook 11408:
1.26 crook 11409: @c -------------------------------------------------------------
1.47 crook 11410: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
11411: @section Keeping track of Time
11412: @cindex time-related words
11413:
11414: Gforth implements time-related operations by making calls to the C
11415: library function, @code{gettimeofday}.
11416:
11417: doc-ms
11418: doc-time&date
11419:
11420:
11421:
11422: @c -------------------------------------------------------------
11423: @node Miscellaneous Words, , Keeping track of Time, Words
1.21 crook 11424: @section Miscellaneous Words
11425: @cindex miscellaneous words
11426:
1.29 crook 11427: @comment TODO find homes for these
11428:
1.26 crook 11429: These section lists the ANS Forth words that are not documented
1.21 crook 11430: elsewhere in this manual. Ultimately, they all need proper homes.
11431:
11432: doc-[compile]
11433:
1.44 crook 11434:
1.26 crook 11435: The following ANS Forth words are not currently supported by Gforth
1.27 crook 11436: (@pxref{ANS conformance}):
1.21 crook 11437:
11438: @code{EDITOR}
11439: @code{EMIT?}
11440: @code{FORGET}
11441:
1.24 anton 11442: @c ******************************************************************
11443: @node Error messages, Tools, Words, Top
11444: @chapter Error messages
11445: @cindex error messages
11446: @cindex backtrace
11447:
11448: A typical Gforth error message looks like this:
11449:
11450: @example
11451: in file included from :-1
11452: in file included from ./yyy.fs:1
11453: ./xxx.fs:4: Invalid memory address
11454: bar
11455: ^^^
1.25 anton 11456: $400E664C @@
11457: $400E6664 foo
1.24 anton 11458: @end example
11459:
11460: The message identifying the error is @code{Invalid memory address}. The
11461: error happened when text-interpreting line 4 of the file
11462: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
11463: word on the line where the error happened, is pointed out (with
11464: @code{^^^}).
11465:
11466: The file containing the error was included in line 1 of @file{./yyy.fs},
11467: and @file{yyy.fs} was included from a non-file (in this case, by giving
11468: @file{yyy.fs} as command-line parameter to Gforth).
11469:
11470: At the end of the error message you find a return stack dump that can be
11471: interpreted as a backtrace (possibly empty). On top you find the top of
11472: the return stack when the @code{throw} happened, and at the bottom you
11473: find the return stack entry just above the return stack of the topmost
11474: text interpreter.
11475:
11476: To the right of most return stack entries you see a guess for the word
11477: that pushed that return stack entry as its return address. This gives a
11478: backtrace. In our case we see that @code{bar} called @code{foo}, and
11479: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
11480: address} exception).
11481:
11482: Note that the backtrace is not perfect: We don't know which return stack
11483: entries are return addresses (so we may get false positives); and in
11484: some cases (e.g., for @code{abort"}) we cannot determine from the return
11485: address the word that pushed the return address, so for some return
11486: addresses you see no names in the return stack dump.
1.25 anton 11487:
11488: @cindex @code{catch} and backtraces
11489: The return stack dump represents the return stack at the time when a
11490: specific @code{throw} was executed. In programs that make use of
11491: @code{catch}, it is not necessarily clear which @code{throw} should be
11492: used for the return stack dump (e.g., consider one @code{throw} that
11493: indicates an error, which is caught, and during recovery another error
1.42 anton 11494: happens; which @code{throw} should be used for the stack dump?). Gforth
1.25 anton 11495: presents the return stack dump for the first @code{throw} after the last
11496: executed (not returned-to) @code{catch}; this works well in the usual
11497: case.
11498:
11499: @cindex @code{gforth-fast} and backtraces
11500: @cindex @code{gforth-fast}, difference from @code{gforth}
11501: @cindex backtraces with @code{gforth-fast}
11502: @cindex return stack dump with @code{gforth-fast}
11503: @code{gforth} is able to do a return stack dump for throws generated
11504: from primitives (e.g., invalid memory address, stack empty etc.);
11505: @code{gforth-fast} is only able to do a return stack dump from a
11506: directly called @code{throw} (including @code{abort} etc.). This is the
1.30 anton 11507: only difference (apart from a speed factor of between 1.15 (K6-2) and
11508: 1.6 (21164A)) between @code{gforth} and @code{gforth-fast}. Given an
11509: exception caused by a primitive in @code{gforth-fast}, you will
11510: typically see no return stack dump at all; however, if the exception is
11511: caught by @code{catch} (e.g., for restoring some state), and then
11512: @code{throw}n again, the return stack dump will be for the first such
11513: @code{throw}.
1.2 jwilke 11514:
1.5 anton 11515: @c ******************************************************************
1.24 anton 11516: @node Tools, ANS conformance, Error messages, Top
1.1 anton 11517: @chapter Tools
11518:
11519: @menu
11520: * ANS Report:: Report the words used, sorted by wordset.
11521: @end menu
11522:
11523: See also @ref{Emacs and Gforth}.
11524:
11525: @node ANS Report, , Tools, Tools
11526: @section @file{ans-report.fs}: Report the words used, sorted by wordset
11527: @cindex @file{ans-report.fs}
11528: @cindex report the words used in your program
11529: @cindex words used in your program
11530:
11531: If you want to label a Forth program as ANS Forth Program, you must
11532: document which wordsets the program uses; for extension wordsets, it is
11533: helpful to list the words the program requires from these wordsets
11534: (because Forth systems are allowed to provide only some words of them).
11535:
11536: The @file{ans-report.fs} tool makes it easy for you to determine which
11537: words from which wordset and which non-ANS words your application
11538: uses. You simply have to include @file{ans-report.fs} before loading the
11539: program you want to check. After loading your program, you can get the
11540: report with @code{print-ans-report}. A typical use is to run this as
11541: batch job like this:
11542: @example
11543: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
11544: @end example
11545:
11546: The output looks like this (for @file{compat/control.fs}):
11547: @example
11548: The program uses the following words
11549: from CORE :
11550: : POSTPONE THEN ; immediate ?dup IF 0=
11551: from BLOCK-EXT :
11552: \
11553: from FILE :
11554: (
11555: @end example
11556:
11557: @subsection Caveats
11558:
11559: Note that @file{ans-report.fs} just checks which words are used, not whether
11560: they are used in an ANS Forth conforming way!
11561:
11562: Some words are defined in several wordsets in the
11563: standard. @file{ans-report.fs} reports them for only one of the
11564: wordsets, and not necessarily the one you expect. It depends on usage
11565: which wordset is the right one to specify. E.g., if you only use the
11566: compilation semantics of @code{S"}, it is a Core word; if you also use
11567: its interpretation semantics, it is a File word.
11568:
11569: @c ******************************************************************
11570: @node ANS conformance, Model, Tools, Top
11571: @chapter ANS conformance
11572: @cindex ANS conformance of Gforth
11573:
11574: To the best of our knowledge, Gforth is an
11575:
11576: ANS Forth System
11577: @itemize @bullet
11578: @item providing the Core Extensions word set
11579: @item providing the Block word set
11580: @item providing the Block Extensions word set
11581: @item providing the Double-Number word set
11582: @item providing the Double-Number Extensions word set
11583: @item providing the Exception word set
11584: @item providing the Exception Extensions word set
11585: @item providing the Facility word set
1.40 anton 11586: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1 anton 11587: @item providing the File Access word set
11588: @item providing the File Access Extensions word set
11589: @item providing the Floating-Point word set
11590: @item providing the Floating-Point Extensions word set
11591: @item providing the Locals word set
11592: @item providing the Locals Extensions word set
11593: @item providing the Memory-Allocation word set
11594: @item providing the Memory-Allocation Extensions word set (that one's easy)
11595: @item providing the Programming-Tools word set
11596: @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
11597: @item providing the Search-Order word set
11598: @item providing the Search-Order Extensions word set
11599: @item providing the String word set
11600: @item providing the String Extensions word set (another easy one)
11601: @end itemize
11602:
11603: @cindex system documentation
11604: In addition, ANS Forth systems are required to document certain
11605: implementation choices. This chapter tries to meet these
11606: requirements. In many cases it gives a way to ask the system for the
11607: information instead of providing the information directly, in
11608: particular, if the information depends on the processor, the operating
11609: system or the installation options chosen, or if they are likely to
11610: change during the maintenance of Gforth.
11611:
11612: @comment The framework for the rest has been taken from pfe.
11613:
11614: @menu
11615: * The Core Words::
11616: * The optional Block word set::
11617: * The optional Double Number word set::
11618: * The optional Exception word set::
11619: * The optional Facility word set::
11620: * The optional File-Access word set::
11621: * The optional Floating-Point word set::
11622: * The optional Locals word set::
11623: * The optional Memory-Allocation word set::
11624: * The optional Programming-Tools word set::
11625: * The optional Search-Order word set::
11626: @end menu
11627:
11628:
11629: @c =====================================================================
11630: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
11631: @comment node-name, next, previous, up
11632: @section The Core Words
11633: @c =====================================================================
11634: @cindex core words, system documentation
11635: @cindex system documentation, core words
11636:
11637: @menu
11638: * core-idef:: Implementation Defined Options
11639: * core-ambcond:: Ambiguous Conditions
11640: * core-other:: Other System Documentation
11641: @end menu
11642:
11643: @c ---------------------------------------------------------------------
11644: @node core-idef, core-ambcond, The Core Words, The Core Words
11645: @subsection Implementation Defined Options
11646: @c ---------------------------------------------------------------------
11647: @cindex core words, implementation-defined options
11648: @cindex implementation-defined options, core words
11649:
11650:
11651: @table @i
11652: @item (Cell) aligned addresses:
11653: @cindex cell-aligned addresses
11654: @cindex aligned addresses
11655: processor-dependent. Gforth's alignment words perform natural alignment
11656: (e.g., an address aligned for a datum of size 8 is divisible by
11657: 8). Unaligned accesses usually result in a @code{-23 THROW}.
11658:
11659: @item @code{EMIT} and non-graphic characters:
11660: @cindex @code{EMIT} and non-graphic characters
11661: @cindex non-graphic characters and @code{EMIT}
11662: The character is output using the C library function (actually, macro)
11663: @code{putc}.
11664:
11665: @item character editing of @code{ACCEPT} and @code{EXPECT}:
11666: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
11667: @cindex editing in @code{ACCEPT} and @code{EXPECT}
11668: @cindex @code{ACCEPT}, editing
11669: @cindex @code{EXPECT}, editing
11670: This is modeled on the GNU readline library (@pxref{Readline
11671: Interaction, , Command Line Editing, readline, The GNU Readline
11672: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
11673: producing a full word completion every time you type it (instead of
1.28 crook 11674: producing the common prefix of all completions). @xref{Command-line editing}.
1.1 anton 11675:
11676: @item character set:
11677: @cindex character set
11678: The character set of your computer and display device. Gforth is
11679: 8-bit-clean (but some other component in your system may make trouble).
11680:
11681: @item Character-aligned address requirements:
11682: @cindex character-aligned address requirements
11683: installation-dependent. Currently a character is represented by a C
11684: @code{unsigned char}; in the future we might switch to @code{wchar_t}
11685: (Comments on that requested).
11686:
11687: @item character-set extensions and matching of names:
11688: @cindex character-set extensions and matching of names
1.26 crook 11689: @cindex case-sensitivity for name lookup
11690: @cindex name lookup, case-sensitivity
11691: @cindex locale and case-sensitivity
1.21 crook 11692: Any character except the ASCII NUL character can be used in a
1.1 anton 11693: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47 crook 11694: matching is performed using the C library function @code{strncasecmp}, whose
1.1 anton 11695: function is probably influenced by the locale. E.g., the @code{C} locale
11696: does not know about accents and umlauts, so they are matched
11697: case-sensitively in that locale. For portability reasons it is best to
11698: write programs such that they work in the @code{C} locale. Then one can
11699: use libraries written by a Polish programmer (who might use words
11700: containing ISO Latin-2 encoded characters) and by a French programmer
11701: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
11702: funny results for some of the words (which ones, depends on the font you
11703: are using)). Also, the locale you prefer may not be available in other
11704: operating systems. Hopefully, Unicode will solve these problems one day.
11705:
11706: @item conditions under which control characters match a space delimiter:
11707: @cindex space delimiters
11708: @cindex control characters as delimiters
11709: If @code{WORD} is called with the space character as a delimiter, all
11710: white-space characters (as identified by the C macro @code{isspace()})
11711: are delimiters. @code{PARSE}, on the other hand, treats space like other
1.44 crook 11712: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
1.1 anton 11713: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
11714: interpreter (aka text interpreter) by default, treats all white-space
11715: characters as delimiters.
11716:
1.26 crook 11717: @item format of the control-flow stack:
11718: @cindex control-flow stack, format
11719: The data stack is used as control-flow stack. The size of a control-flow
1.1 anton 11720: stack item in cells is given by the constant @code{cs-item-size}. At the
11721: time of this writing, an item consists of a (pointer to a) locals list
11722: (third), an address in the code (second), and a tag for identifying the
11723: item (TOS). The following tags are used: @code{defstart},
11724: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
11725: @code{scopestart}.
11726:
11727: @item conversion of digits > 35
11728: @cindex digits > 35
11729: The characters @code{[\]^_'} are the digits with the decimal value
11730: 36@minus{}41. There is no way to input many of the larger digits.
11731:
11732: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
11733: @cindex @code{EXPECT}, display after end of input
11734: @cindex @code{ACCEPT}, display after end of input
11735: The cursor is moved to the end of the entered string. If the input is
11736: terminated using the @kbd{Return} key, a space is typed.
11737:
11738: @item exception abort sequence of @code{ABORT"}:
11739: @cindex exception abort sequence of @code{ABORT"}
11740: @cindex @code{ABORT"}, exception abort sequence
11741: The error string is stored into the variable @code{"error} and a
11742: @code{-2 throw} is performed.
11743:
11744: @item input line terminator:
11745: @cindex input line terminator
11746: @cindex line terminator on input
1.26 crook 11747: @cindex newline character on input
1.1 anton 11748: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
11749: lines. One of these characters is typically produced when you type the
11750: @kbd{Enter} or @kbd{Return} key.
11751:
11752: @item maximum size of a counted string:
11753: @cindex maximum size of a counted string
11754: @cindex counted string, maximum size
11755: @code{s" /counted-string" environment? drop .}. Currently 255 characters
11756: on all ports, but this may change.
11757:
11758: @item maximum size of a parsed string:
11759: @cindex maximum size of a parsed string
11760: @cindex parsed string, maximum size
11761: Given by the constant @code{/line}. Currently 255 characters.
11762:
11763: @item maximum size of a definition name, in characters:
11764: @cindex maximum size of a definition name, in characters
11765: @cindex name, maximum length
11766: 31
11767:
11768: @item maximum string length for @code{ENVIRONMENT?}, in characters:
11769: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
11770: @cindex @code{ENVIRONMENT?} string length, maximum
11771: 31
11772:
11773: @item method of selecting the user input device:
11774: @cindex user input device, method of selecting
11775: The user input device is the standard input. There is currently no way to
11776: change it from within Gforth. However, the input can typically be
11777: redirected in the command line that starts Gforth.
11778:
11779: @item method of selecting the user output device:
11780: @cindex user output device, method of selecting
11781: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10 anton 11782: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
11783: output when the user output device is a terminal, otherwise the output
11784: is buffered.
1.1 anton 11785:
11786: @item methods of dictionary compilation:
11787: What are we expected to document here?
11788:
11789: @item number of bits in one address unit:
11790: @cindex number of bits in one address unit
11791: @cindex address unit, size in bits
11792: @code{s" address-units-bits" environment? drop .}. 8 in all current
11793: ports.
11794:
11795: @item number representation and arithmetic:
11796: @cindex number representation and arithmetic
11797: Processor-dependent. Binary two's complement on all current ports.
11798:
11799: @item ranges for integer types:
11800: @cindex ranges for integer types
11801: @cindex integer types, ranges
11802: Installation-dependent. Make environmental queries for @code{MAX-N},
11803: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
11804: unsigned (and positive) types is 0. The lower bound for signed types on
11805: two's complement and one's complement machines machines can be computed
11806: by adding 1 to the upper bound.
11807:
11808: @item read-only data space regions:
11809: @cindex read-only data space regions
11810: @cindex data-space, read-only regions
11811: The whole Forth data space is writable.
11812:
11813: @item size of buffer at @code{WORD}:
11814: @cindex size of buffer at @code{WORD}
11815: @cindex @code{WORD} buffer size
11816: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11817: shared with the pictured numeric output string. If overwriting
11818: @code{PAD} is acceptable, it is as large as the remaining dictionary
11819: space, although only as much can be sensibly used as fits in a counted
11820: string.
11821:
11822: @item size of one cell in address units:
11823: @cindex cell size
11824: @code{1 cells .}.
11825:
11826: @item size of one character in address units:
11827: @cindex char size
11828: @code{1 chars .}. 1 on all current ports.
11829:
11830: @item size of the keyboard terminal buffer:
11831: @cindex size of the keyboard terminal buffer
11832: @cindex terminal buffer, size
11833: Varies. You can determine the size at a specific time using @code{lp@@
11834: tib - .}. It is shared with the locals stack and TIBs of files that
11835: include the current file. You can change the amount of space for TIBs
11836: and locals stack at Gforth startup with the command line option
11837: @code{-l}.
11838:
11839: @item size of the pictured numeric output buffer:
11840: @cindex size of the pictured numeric output buffer
11841: @cindex pictured numeric output buffer, size
11842: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11843: shared with @code{WORD}.
11844:
11845: @item size of the scratch area returned by @code{PAD}:
11846: @cindex size of the scratch area returned by @code{PAD}
11847: @cindex @code{PAD} size
11848: The remainder of dictionary space. @code{unused pad here - - .}.
11849:
11850: @item system case-sensitivity characteristics:
11851: @cindex case-sensitivity characteristics
1.26 crook 11852: Dictionary searches are case-insensitive (except in
1.1 anton 11853: @code{TABLE}s). However, as explained above under @i{character-set
11854: extensions}, the matching for non-ASCII characters is determined by the
11855: locale you are using. In the default @code{C} locale all non-ASCII
11856: characters are matched case-sensitively.
11857:
11858: @item system prompt:
11859: @cindex system prompt
11860: @cindex prompt
11861: @code{ ok} in interpret state, @code{ compiled} in compile state.
11862:
11863: @item division rounding:
11864: @cindex division rounding
11865: installation dependent. @code{s" floored" environment? drop .}. We leave
11866: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
11867: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
11868:
11869: @item values of @code{STATE} when true:
11870: @cindex @code{STATE} values
11871: -1.
11872:
11873: @item values returned after arithmetic overflow:
11874: On two's complement machines, arithmetic is performed modulo
11875: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
11876: arithmetic (with appropriate mapping for signed types). Division by zero
11877: typically results in a @code{-55 throw} (Floating-point unidentified
11878: fault), although a @code{-10 throw} (divide by zero) would be more
11879: appropriate.
11880:
11881: @item whether the current definition can be found after @t{DOES>}:
11882: @cindex @t{DOES>}, visibility of current definition
11883: No.
11884:
11885: @end table
11886:
11887: @c ---------------------------------------------------------------------
11888: @node core-ambcond, core-other, core-idef, The Core Words
11889: @subsection Ambiguous conditions
11890: @c ---------------------------------------------------------------------
11891: @cindex core words, ambiguous conditions
11892: @cindex ambiguous conditions, core words
11893:
11894: @table @i
11895:
11896: @item a name is neither a word nor a number:
11897: @cindex name not found
1.26 crook 11898: @cindex undefined word
1.1 anton 11899: @code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
11900: preserves the data and FP stack, so you don't lose more work than
11901: necessary.
11902:
11903: @item a definition name exceeds the maximum length allowed:
1.26 crook 11904: @cindex word name too long
1.1 anton 11905: @code{-19 throw} (Word name too long)
11906:
11907: @item addressing a region not inside the various data spaces of the forth system:
11908: @cindex Invalid memory address
1.32 anton 11909: The stacks, code space and header space are accessible. Machine code space is
1.1 anton 11910: typically readable. Accessing other addresses gives results dependent on
11911: the operating system. On decent systems: @code{-9 throw} (Invalid memory
11912: address).
11913:
11914: @item argument type incompatible with parameter:
1.26 crook 11915: @cindex argument type mismatch
1.1 anton 11916: This is usually not caught. Some words perform checks, e.g., the control
11917: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
11918: mismatch).
11919:
11920: @item attempting to obtain the execution token of a word with undefined execution semantics:
11921: @cindex Interpreting a compile-only word, for @code{'} etc.
11922: @cindex execution token of words with undefined execution semantics
11923: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
11924: get an execution token for @code{compile-only-error} (which performs a
11925: @code{-14 throw} when executed).
11926:
11927: @item dividing by zero:
11928: @cindex dividing by zero
11929: @cindex floating point unidentified fault, integer division
1.24 anton 11930: On better platforms, this produces a @code{-10 throw} (Division by
11931: zero); on other systems, this typically results in a @code{-55 throw}
11932: (Floating-point unidentified fault).
1.1 anton 11933:
11934: @item insufficient data stack or return stack space:
11935: @cindex insufficient data stack or return stack space
11936: @cindex stack overflow
1.26 crook 11937: @cindex address alignment exception, stack overflow
1.1 anton 11938: @cindex Invalid memory address, stack overflow
11939: Depending on the operating system, the installation, and the invocation
11940: of Gforth, this is either checked by the memory management hardware, or
1.24 anton 11941: it is not checked. If it is checked, you typically get a @code{-3 throw}
11942: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
11943: throw} (Invalid memory address) (depending on the platform and how you
11944: achieved the overflow) as soon as the overflow happens. If it is not
11945: checked, overflows typically result in mysterious illegal memory
11946: accesses, producing @code{-9 throw} (Invalid memory address) or
11947: @code{-23 throw} (Address alignment exception); they might also destroy
11948: the internal data structure of @code{ALLOCATE} and friends, resulting in
11949: various errors in these words.
1.1 anton 11950:
11951: @item insufficient space for loop control parameters:
11952: @cindex insufficient space for loop control parameters
11953: like other return stack overflows.
11954:
11955: @item insufficient space in the dictionary:
11956: @cindex insufficient space in the dictionary
11957: @cindex dictionary overflow
1.12 anton 11958: If you try to allot (either directly with @code{allot}, or indirectly
11959: with @code{,}, @code{create} etc.) more memory than available in the
11960: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
11961: to access memory beyond the end of the dictionary, the results are
11962: similar to stack overflows.
1.1 anton 11963:
11964: @item interpreting a word with undefined interpretation semantics:
11965: @cindex interpreting a word with undefined interpretation semantics
11966: @cindex Interpreting a compile-only word
11967: For some words, we have defined interpretation semantics. For the
11968: others: @code{-14 throw} (Interpreting a compile-only word).
11969:
11970: @item modifying the contents of the input buffer or a string literal:
11971: @cindex modifying the contents of the input buffer or a string literal
11972: These are located in writable memory and can be modified.
11973:
11974: @item overflow of the pictured numeric output string:
11975: @cindex overflow of the pictured numeric output string
11976: @cindex pictured numeric output string, overflow
1.24 anton 11977: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1 anton 11978:
11979: @item parsed string overflow:
11980: @cindex parsed string overflow
11981: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
11982:
11983: @item producing a result out of range:
11984: @cindex result out of range
11985: On two's complement machines, arithmetic is performed modulo
11986: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
11987: arithmetic (with appropriate mapping for signed types). Division by zero
1.24 anton 11988: typically results in a @code{-10 throw} (divide by zero) or @code{-55
11989: throw} (floating point unidentified fault). @code{convert} and
11990: @code{>number} currently overflow silently.
1.1 anton 11991:
11992: @item reading from an empty data or return stack:
11993: @cindex stack empty
11994: @cindex stack underflow
1.24 anton 11995: @cindex return stack underflow
1.1 anton 11996: The data stack is checked by the outer (aka text) interpreter after
11997: every word executed. If it has underflowed, a @code{-4 throw} (Stack
11998: underflow) is performed. Apart from that, stacks may be checked or not,
1.24 anton 11999: depending on operating system, installation, and invocation. If they are
12000: caught by a check, they typically result in @code{-4 throw} (Stack
12001: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12002: (Invalid memory address), depending on the platform and which stack
12003: underflows and by how much. Note that even if the system uses checking
12004: (through the MMU), your program may have to underflow by a significant
12005: number of stack items to trigger the reaction (the reason for this is
12006: that the MMU, and therefore the checking, works with a page-size
12007: granularity). If there is no checking, the symptoms resulting from an
12008: underflow are similar to those from an overflow. Unbalanced return
12009: stack errors result in a variaty of symptoms, including @code{-9 throw}
12010: (Invalid memory address) and Illegal Instruction (typically @code{-260
12011: throw}).
1.1 anton 12012:
12013: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12014: @cindex unexpected end of the input buffer
12015: @cindex zero-length string as a name
12016: @cindex Attempt to use zero-length string as a name
12017: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12018: use zero-length string as a name). Words like @code{'} probably will not
12019: find what they search. Note that it is possible to create zero-length
12020: names with @code{nextname} (should it not?).
12021:
12022: @item @code{>IN} greater than input buffer:
12023: @cindex @code{>IN} greater than input buffer
12024: The next invocation of a parsing word returns a string with length 0.
12025:
12026: @item @code{RECURSE} appears after @code{DOES>}:
12027: @cindex @code{RECURSE} appears after @code{DOES>}
12028: Compiles a recursive call to the defining word, not to the defined word.
12029:
12030: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12031: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26 crook 12032: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1 anton 12033: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12034: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12035: the end of the file was reached), its source-id may be
12036: reused. Therefore, restoring an input source specification referencing a
12037: closed file may lead to unpredictable results instead of a @code{-12
12038: THROW}.
12039:
12040: In the future, Gforth may be able to restore input source specifications
12041: from other than the current input source.
12042:
12043: @item data space containing definitions gets de-allocated:
12044: @cindex data space containing definitions gets de-allocated
12045: Deallocation with @code{allot} is not checked. This typically results in
12046: memory access faults or execution of illegal instructions.
12047:
12048: @item data space read/write with incorrect alignment:
12049: @cindex data space read/write with incorrect alignment
12050: @cindex alignment faults
1.26 crook 12051: @cindex address alignment exception
1.1 anton 12052: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12 anton 12053: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1 anton 12054: alignment turned on, incorrect alignment results in a @code{-9 throw}
12055: (Invalid memory address). There are reportedly some processors with
1.12 anton 12056: alignment restrictions that do not report violations.
1.1 anton 12057:
12058: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12059: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12060: Like other alignment errors.
12061:
12062: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12063: Like other stack underflows.
12064:
12065: @item loop control parameters not available:
12066: @cindex loop control parameters not available
12067: Not checked. The counted loop words simply assume that the top of return
12068: stack items are loop control parameters and behave accordingly.
12069:
12070: @item most recent definition does not have a name (@code{IMMEDIATE}):
12071: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12072: @cindex last word was headerless
12073: @code{abort" last word was headerless"}.
12074:
12075: @item name not defined by @code{VALUE} used by @code{TO}:
12076: @cindex name not defined by @code{VALUE} used by @code{TO}
12077: @cindex @code{TO} on non-@code{VALUE}s
12078: @cindex Invalid name argument, @code{TO}
12079: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12080: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12081:
12082: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12083: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26 crook 12084: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1 anton 12085: @code{-13 throw} (Undefined word)
12086:
12087: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12088: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12089: Gforth behaves as if they were of the same type. I.e., you can predict
12090: the behaviour by interpreting all parameters as, e.g., signed.
12091:
12092: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12093: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12094: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12095: compilation semantics of @code{TO}.
12096:
12097: @item String longer than a counted string returned by @code{WORD}:
1.26 crook 12098: @cindex string longer than a counted string returned by @code{WORD}
1.1 anton 12099: @cindex @code{WORD}, string overflow
12100: Not checked. The string will be ok, but the count will, of course,
12101: contain only the least significant bits of the length.
12102:
12103: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12104: @cindex @code{LSHIFT}, large shift counts
12105: @cindex @code{RSHIFT}, large shift counts
12106: Processor-dependent. Typical behaviours are returning 0 and using only
12107: the low bits of the shift count.
12108:
12109: @item word not defined via @code{CREATE}:
12110: @cindex @code{>BODY} of non-@code{CREATE}d words
12111: @code{>BODY} produces the PFA of the word no matter how it was defined.
12112:
12113: @cindex @code{DOES>} of non-@code{CREATE}d words
12114: @code{DOES>} changes the execution semantics of the last defined word no
12115: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12116: @code{CREATE , DOES>}.
12117:
12118: @item words improperly used outside @code{<#} and @code{#>}:
12119: Not checked. As usual, you can expect memory faults.
12120:
12121: @end table
12122:
12123:
12124: @c ---------------------------------------------------------------------
12125: @node core-other, , core-ambcond, The Core Words
12126: @subsection Other system documentation
12127: @c ---------------------------------------------------------------------
12128: @cindex other system documentation, core words
12129: @cindex core words, other system documentation
12130:
12131: @table @i
12132: @item nonstandard words using @code{PAD}:
12133: @cindex @code{PAD} use by nonstandard words
12134: None.
12135:
12136: @item operator's terminal facilities available:
12137: @cindex operator's terminal facilities available
12138: After processing the command line, Gforth goes into interactive mode,
12139: and you can give commands to Gforth interactively. The actual facilities
12140: available depend on how you invoke Gforth.
12141:
12142: @item program data space available:
12143: @cindex program data space available
12144: @cindex data space available
12145: @code{UNUSED .} gives the remaining dictionary space. The total
12146: dictionary space can be specified with the @code{-m} switch
12147: (@pxref{Invoking Gforth}) when Gforth starts up.
12148:
12149: @item return stack space available:
12150: @cindex return stack space available
12151: You can compute the total return stack space in cells with
12152: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12153: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12154:
12155: @item stack space available:
12156: @cindex stack space available
12157: You can compute the total data stack space in cells with
12158: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12159: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12160:
12161: @item system dictionary space required, in address units:
12162: @cindex system dictionary space required, in address units
12163: Type @code{here forthstart - .} after startup. At the time of this
12164: writing, this gives 80080 (bytes) on a 32-bit system.
12165: @end table
12166:
12167:
12168: @c =====================================================================
12169: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12170: @section The optional Block word set
12171: @c =====================================================================
12172: @cindex system documentation, block words
12173: @cindex block words, system documentation
12174:
12175: @menu
12176: * block-idef:: Implementation Defined Options
12177: * block-ambcond:: Ambiguous Conditions
12178: * block-other:: Other System Documentation
12179: @end menu
12180:
12181:
12182: @c ---------------------------------------------------------------------
12183: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12184: @subsection Implementation Defined Options
12185: @c ---------------------------------------------------------------------
12186: @cindex implementation-defined options, block words
12187: @cindex block words, implementation-defined options
12188:
12189: @table @i
12190: @item the format for display by @code{LIST}:
12191: @cindex @code{LIST} display format
12192: First the screen number is displayed, then 16 lines of 64 characters,
12193: each line preceded by the line number.
12194:
12195: @item the length of a line affected by @code{\}:
12196: @cindex length of a line affected by @code{\}
12197: @cindex @code{\}, line length in blocks
12198: 64 characters.
12199: @end table
12200:
12201:
12202: @c ---------------------------------------------------------------------
12203: @node block-ambcond, block-other, block-idef, The optional Block word set
12204: @subsection Ambiguous conditions
12205: @c ---------------------------------------------------------------------
12206: @cindex block words, ambiguous conditions
12207: @cindex ambiguous conditions, block words
12208:
12209: @table @i
12210: @item correct block read was not possible:
12211: @cindex block read not possible
12212: Typically results in a @code{throw} of some OS-derived value (between
12213: -512 and -2048). If the blocks file was just not long enough, blanks are
12214: supplied for the missing portion.
12215:
12216: @item I/O exception in block transfer:
12217: @cindex I/O exception in block transfer
12218: @cindex block transfer, I/O exception
12219: Typically results in a @code{throw} of some OS-derived value (between
12220: -512 and -2048).
12221:
12222: @item invalid block number:
12223: @cindex invalid block number
12224: @cindex block number invalid
12225: @code{-35 throw} (Invalid block number)
12226:
12227: @item a program directly alters the contents of @code{BLK}:
12228: @cindex @code{BLK}, altering @code{BLK}
12229: The input stream is switched to that other block, at the same
12230: position. If the storing to @code{BLK} happens when interpreting
12231: non-block input, the system will get quite confused when the block ends.
12232:
12233: @item no current block buffer for @code{UPDATE}:
12234: @cindex @code{UPDATE}, no current block buffer
12235: @code{UPDATE} has no effect.
12236:
12237: @end table
12238:
12239: @c ---------------------------------------------------------------------
12240: @node block-other, , block-ambcond, The optional Block word set
12241: @subsection Other system documentation
12242: @c ---------------------------------------------------------------------
12243: @cindex other system documentation, block words
12244: @cindex block words, other system documentation
12245:
12246: @table @i
12247: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12248: No restrictions (yet).
12249:
12250: @item the number of blocks available for source and data:
12251: depends on your disk space.
12252:
12253: @end table
12254:
12255:
12256: @c =====================================================================
12257: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12258: @section The optional Double Number word set
12259: @c =====================================================================
12260: @cindex system documentation, double words
12261: @cindex double words, system documentation
12262:
12263: @menu
12264: * double-ambcond:: Ambiguous Conditions
12265: @end menu
12266:
12267:
12268: @c ---------------------------------------------------------------------
12269: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12270: @subsection Ambiguous conditions
12271: @c ---------------------------------------------------------------------
12272: @cindex double words, ambiguous conditions
12273: @cindex ambiguous conditions, double words
12274:
12275: @table @i
1.29 crook 12276: @item @i{d} outside of range of @i{n} in @code{D>S}:
12277: @cindex @code{D>S}, @i{d} out of range of @i{n}
12278: The least significant cell of @i{d} is produced.
1.1 anton 12279:
12280: @end table
12281:
12282:
12283: @c =====================================================================
12284: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12285: @section The optional Exception word set
12286: @c =====================================================================
12287: @cindex system documentation, exception words
12288: @cindex exception words, system documentation
12289:
12290: @menu
12291: * exception-idef:: Implementation Defined Options
12292: @end menu
12293:
12294:
12295: @c ---------------------------------------------------------------------
12296: @node exception-idef, , The optional Exception word set, The optional Exception word set
12297: @subsection Implementation Defined Options
12298: @c ---------------------------------------------------------------------
12299: @cindex implementation-defined options, exception words
12300: @cindex exception words, implementation-defined options
12301:
12302: @table @i
12303: @item @code{THROW}-codes used in the system:
12304: @cindex @code{THROW}-codes used in the system
12305: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29 crook 12306: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1 anton 12307: codes -512@minus{}-2047 are used for OS errors (for file and memory
12308: allocation operations). The mapping from OS error numbers to throw codes
12309: is -512@minus{}@code{errno}. One side effect of this mapping is that
12310: undefined OS errors produce a message with a strange number; e.g.,
12311: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12312: @end table
12313:
12314: @c =====================================================================
12315: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12316: @section The optional Facility word set
12317: @c =====================================================================
12318: @cindex system documentation, facility words
12319: @cindex facility words, system documentation
12320:
12321: @menu
12322: * facility-idef:: Implementation Defined Options
12323: * facility-ambcond:: Ambiguous Conditions
12324: @end menu
12325:
12326:
12327: @c ---------------------------------------------------------------------
12328: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12329: @subsection Implementation Defined Options
12330: @c ---------------------------------------------------------------------
12331: @cindex implementation-defined options, facility words
12332: @cindex facility words, implementation-defined options
12333:
12334: @table @i
12335: @item encoding of keyboard events (@code{EKEY}):
12336: @cindex keyboard events, encoding in @code{EKEY}
12337: @cindex @code{EKEY}, encoding of keyboard events
1.40 anton 12338: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41 anton 12339: Other keys are encoded with the constants @code{k-left}, @code{k-right},
12340: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
12341: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
12342: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40 anton 12343:
1.1 anton 12344:
12345: @item duration of a system clock tick:
12346: @cindex duration of a system clock tick
12347: @cindex clock tick duration
12348: System dependent. With respect to @code{MS}, the time is specified in
12349: microseconds. How well the OS and the hardware implement this, is
12350: another question.
12351:
12352: @item repeatability to be expected from the execution of @code{MS}:
12353: @cindex repeatability to be expected from the execution of @code{MS}
12354: @cindex @code{MS}, repeatability to be expected
12355: System dependent. On Unix, a lot depends on load. If the system is
12356: lightly loaded, and the delay is short enough that Gforth does not get
12357: swapped out, the performance should be acceptable. Under MS-DOS and
12358: other single-tasking systems, it should be good.
12359:
12360: @end table
12361:
12362:
12363: @c ---------------------------------------------------------------------
12364: @node facility-ambcond, , facility-idef, The optional Facility word set
12365: @subsection Ambiguous conditions
12366: @c ---------------------------------------------------------------------
12367: @cindex facility words, ambiguous conditions
12368: @cindex ambiguous conditions, facility words
12369:
12370: @table @i
12371: @item @code{AT-XY} can't be performed on user output device:
12372: @cindex @code{AT-XY} can't be performed on user output device
12373: Largely terminal dependent. No range checks are done on the arguments.
12374: No errors are reported. You may see some garbage appearing, you may see
12375: simply nothing happen.
12376:
12377: @end table
12378:
12379:
12380: @c =====================================================================
12381: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
12382: @section The optional File-Access word set
12383: @c =====================================================================
12384: @cindex system documentation, file words
12385: @cindex file words, system documentation
12386:
12387: @menu
12388: * file-idef:: Implementation Defined Options
12389: * file-ambcond:: Ambiguous Conditions
12390: @end menu
12391:
12392: @c ---------------------------------------------------------------------
12393: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
12394: @subsection Implementation Defined Options
12395: @c ---------------------------------------------------------------------
12396: @cindex implementation-defined options, file words
12397: @cindex file words, implementation-defined options
12398:
12399: @table @i
12400: @item file access methods used:
12401: @cindex file access methods used
12402: @code{R/O}, @code{R/W} and @code{BIN} work as you would
12403: expect. @code{W/O} translates into the C file opening mode @code{w} (or
12404: @code{wb}): The file is cleared, if it exists, and created, if it does
12405: not (with both @code{open-file} and @code{create-file}). Under Unix
12406: @code{create-file} creates a file with 666 permissions modified by your
12407: umask.
12408:
12409: @item file exceptions:
12410: @cindex file exceptions
12411: The file words do not raise exceptions (except, perhaps, memory access
12412: faults when you pass illegal addresses or file-ids).
12413:
12414: @item file line terminator:
12415: @cindex file line terminator
12416: System-dependent. Gforth uses C's newline character as line
12417: terminator. What the actual character code(s) of this are is
12418: system-dependent.
12419:
12420: @item file name format:
12421: @cindex file name format
12422: System dependent. Gforth just uses the file name format of your OS.
12423:
12424: @item information returned by @code{FILE-STATUS}:
12425: @cindex @code{FILE-STATUS}, returned information
12426: @code{FILE-STATUS} returns the most powerful file access mode allowed
12427: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
12428: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
12429: along with the returned mode.
12430:
12431: @item input file state after an exception when including source:
12432: @cindex exception when including source
12433: All files that are left via the exception are closed.
12434:
1.29 crook 12435: @item @i{ior} values and meaning:
12436: @cindex @i{ior} values and meaning
12437: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 12438: intended as throw codes. They typically are in the range
12439: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 12440: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 12441:
12442: @item maximum depth of file input nesting:
12443: @cindex maximum depth of file input nesting
12444: @cindex file input nesting, maximum depth
12445: limited by the amount of return stack, locals/TIB stack, and the number
12446: of open files available. This should not give you troubles.
12447:
12448: @item maximum size of input line:
12449: @cindex maximum size of input line
12450: @cindex input line size, maximum
12451: @code{/line}. Currently 255.
12452:
12453: @item methods of mapping block ranges to files:
12454: @cindex mapping block ranges to files
12455: @cindex files containing blocks
12456: @cindex blocks in files
12457: By default, blocks are accessed in the file @file{blocks.fb} in the
12458: current working directory. The file can be switched with @code{USE}.
12459:
12460: @item number of string buffers provided by @code{S"}:
12461: @cindex @code{S"}, number of string buffers
12462: 1
12463:
12464: @item size of string buffer used by @code{S"}:
12465: @cindex @code{S"}, size of string buffer
12466: @code{/line}. currently 255.
12467:
12468: @end table
12469:
12470: @c ---------------------------------------------------------------------
12471: @node file-ambcond, , file-idef, The optional File-Access word set
12472: @subsection Ambiguous conditions
12473: @c ---------------------------------------------------------------------
12474: @cindex file words, ambiguous conditions
12475: @cindex ambiguous conditions, file words
12476:
12477: @table @i
12478: @item attempting to position a file outside its boundaries:
12479: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
12480: @code{REPOSITION-FILE} is performed as usual: Afterwards,
12481: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
12482:
12483: @item attempting to read from file positions not yet written:
12484: @cindex reading from file positions not yet written
12485: End-of-file, i.e., zero characters are read and no error is reported.
12486:
1.29 crook 12487: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
12488: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
1.1 anton 12489: An appropriate exception may be thrown, but a memory fault or other
12490: problem is more probable.
12491:
1.29 crook 12492: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
12493: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
12494: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
12495: The @i{ior} produced by the operation, that discovered the problem, is
1.1 anton 12496: thrown.
12497:
12498: @item named file cannot be opened (@code{INCLUDED}):
12499: @cindex @code{INCLUDED}, named file cannot be opened
1.29 crook 12500: The @i{ior} produced by @code{open-file} is thrown.
1.1 anton 12501:
12502: @item requesting an unmapped block number:
12503: @cindex unmapped block numbers
12504: There are no unmapped legal block numbers. On some operating systems,
12505: writing a block with a large number may overflow the file system and
12506: have an error message as consequence.
12507:
12508: @item using @code{source-id} when @code{blk} is non-zero:
12509: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
12510: @code{source-id} performs its function. Typically it will give the id of
12511: the source which loaded the block. (Better ideas?)
12512:
12513: @end table
12514:
12515:
12516: @c =====================================================================
12517: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
12518: @section The optional Floating-Point word set
12519: @c =====================================================================
12520: @cindex system documentation, floating-point words
12521: @cindex floating-point words, system documentation
12522:
12523: @menu
12524: * floating-idef:: Implementation Defined Options
12525: * floating-ambcond:: Ambiguous Conditions
12526: @end menu
12527:
12528:
12529: @c ---------------------------------------------------------------------
12530: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
12531: @subsection Implementation Defined Options
12532: @c ---------------------------------------------------------------------
12533: @cindex implementation-defined options, floating-point words
12534: @cindex floating-point words, implementation-defined options
12535:
12536: @table @i
12537: @item format and range of floating point numbers:
12538: @cindex format and range of floating point numbers
12539: @cindex floating point numbers, format and range
12540: System-dependent; the @code{double} type of C.
12541:
1.29 crook 12542: @item results of @code{REPRESENT} when @i{float} is out of range:
12543: @cindex @code{REPRESENT}, results when @i{float} is out of range
1.1 anton 12544: System dependent; @code{REPRESENT} is implemented using the C library
12545: function @code{ecvt()} and inherits its behaviour in this respect.
12546:
12547: @item rounding or truncation of floating-point numbers:
12548: @cindex rounding of floating-point numbers
12549: @cindex truncation of floating-point numbers
12550: @cindex floating-point numbers, rounding or truncation
12551: System dependent; the rounding behaviour is inherited from the hosting C
12552: compiler. IEEE-FP-based (i.e., most) systems by default round to
12553: nearest, and break ties by rounding to even (i.e., such that the last
12554: bit of the mantissa is 0).
12555:
12556: @item size of floating-point stack:
12557: @cindex floating-point stack size
12558: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
12559: the floating-point stack (in floats). You can specify this on startup
12560: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
12561:
12562: @item width of floating-point stack:
12563: @cindex floating-point stack width
12564: @code{1 floats}.
12565:
12566: @end table
12567:
12568:
12569: @c ---------------------------------------------------------------------
12570: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
12571: @subsection Ambiguous conditions
12572: @c ---------------------------------------------------------------------
12573: @cindex floating-point words, ambiguous conditions
12574: @cindex ambiguous conditions, floating-point words
12575:
12576: @table @i
12577: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
12578: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
12579: System-dependent. Typically results in a @code{-23 THROW} like other
12580: alignment violations.
12581:
12582: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
12583: @cindex @code{f@@} used with an address that is not float aligned
12584: @cindex @code{f!} used with an address that is not float aligned
12585: System-dependent. Typically results in a @code{-23 THROW} like other
12586: alignment violations.
12587:
12588: @item floating-point result out of range:
12589: @cindex floating-point result out of range
12590: System-dependent. Can result in a @code{-55 THROW} (Floating-point
12591: unidentified fault), or can produce a special value representing, e.g.,
12592: Infinity.
12593:
12594: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
12595: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
12596: System-dependent. Typically results in an alignment fault like other
12597: alignment violations.
12598:
1.35 anton 12599: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
12600: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1 anton 12601: The floating-point number is converted into decimal nonetheless.
12602:
12603: @item Both arguments are equal to zero (@code{FATAN2}):
12604: @cindex @code{FATAN2}, both arguments are equal to zero
12605: System-dependent. @code{FATAN2} is implemented using the C library
12606: function @code{atan2()}.
12607:
1.29 crook 12608: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
12609: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
12610: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1 anton 12611: because of small errors and the tan will be a very large (or very small)
12612: but finite number.
12613:
1.29 crook 12614: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
12615: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1 anton 12616: The result is rounded to the nearest float.
12617:
12618: @item dividing by zero:
12619: @cindex dividing by zero, floating-point
12620: @cindex floating-point dividing by zero
12621: @cindex floating-point unidentified fault, FP divide-by-zero
12622: @code{-55 throw} (Floating-point unidentified fault)
12623:
12624: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
12625: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
12626: System dependent. On IEEE-FP based systems the number is converted into
12627: an infinity.
12628:
1.29 crook 12629: @item @i{float}<1 (@code{FACOSH}):
12630: @cindex @code{FACOSH}, @i{float}<1
1.1 anton 12631: @cindex floating-point unidentified fault, @code{FACOSH}
12632: @code{-55 throw} (Floating-point unidentified fault)
12633:
1.29 crook 12634: @item @i{float}=<-1 (@code{FLNP1}):
12635: @cindex @code{FLNP1}, @i{float}=<-1
1.1 anton 12636: @cindex floating-point unidentified fault, @code{FLNP1}
12637: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
1.29 crook 12638: negative infinity is typically produced for @i{float}=-1.
1.1 anton 12639:
1.29 crook 12640: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
12641: @cindex @code{FLN}, @i{float}=<0
12642: @cindex @code{FLOG}, @i{float}=<0
1.1 anton 12643: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
12644: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
1.29 crook 12645: negative infinity is typically produced for @i{float}=0.
1.1 anton 12646:
1.29 crook 12647: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
12648: @cindex @code{FASINH}, @i{float}<0
12649: @cindex @code{FSQRT}, @i{float}<0
1.1 anton 12650: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
12651: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
12652: produces values for these inputs on my Linux box (Bug in the C library?)
12653:
1.29 crook 12654: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
12655: @cindex @code{FACOS}, |@i{float}|>1
12656: @cindex @code{FASIN}, |@i{float}|>1
12657: @cindex @code{FATANH}, |@i{float}|>1
1.1 anton 12658: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
12659: @code{-55 throw} (Floating-point unidentified fault).
12660:
1.29 crook 12661: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
12662: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1 anton 12663: @cindex floating-point unidentified fault, @code{F>D}
12664: @code{-55 throw} (Floating-point unidentified fault).
12665:
12666: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
12667: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
12668: This does not happen.
12669: @end table
12670:
12671: @c =====================================================================
12672: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
12673: @section The optional Locals word set
12674: @c =====================================================================
12675: @cindex system documentation, locals words
12676: @cindex locals words, system documentation
12677:
12678: @menu
12679: * locals-idef:: Implementation Defined Options
12680: * locals-ambcond:: Ambiguous Conditions
12681: @end menu
12682:
12683:
12684: @c ---------------------------------------------------------------------
12685: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
12686: @subsection Implementation Defined Options
12687: @c ---------------------------------------------------------------------
12688: @cindex implementation-defined options, locals words
12689: @cindex locals words, implementation-defined options
12690:
12691: @table @i
12692: @item maximum number of locals in a definition:
12693: @cindex maximum number of locals in a definition
12694: @cindex locals, maximum number in a definition
12695: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
12696: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
12697: characters. The number of locals in a definition is bounded by the size
12698: of locals-buffer, which contains the names of the locals.
12699:
12700: @end table
12701:
12702:
12703: @c ---------------------------------------------------------------------
12704: @node locals-ambcond, , locals-idef, The optional Locals word set
12705: @subsection Ambiguous conditions
12706: @c ---------------------------------------------------------------------
12707: @cindex locals words, ambiguous conditions
12708: @cindex ambiguous conditions, locals words
12709:
12710: @table @i
12711: @item executing a named local in interpretation state:
12712: @cindex local in interpretation state
12713: @cindex Interpreting a compile-only word, for a local
12714: Locals have no interpretation semantics. If you try to perform the
12715: interpretation semantics, you will get a @code{-14 throw} somewhere
12716: (Interpreting a compile-only word). If you perform the compilation
12717: semantics, the locals access will be compiled (irrespective of state).
12718:
1.29 crook 12719: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1 anton 12720: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
12721: @cindex @code{TO} on non-@code{VALUE}s and non-locals
12722: @cindex Invalid name argument, @code{TO}
12723: @code{-32 throw} (Invalid name argument)
12724:
12725: @end table
12726:
12727:
12728: @c =====================================================================
12729: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
12730: @section The optional Memory-Allocation word set
12731: @c =====================================================================
12732: @cindex system documentation, memory-allocation words
12733: @cindex memory-allocation words, system documentation
12734:
12735: @menu
12736: * memory-idef:: Implementation Defined Options
12737: @end menu
12738:
12739:
12740: @c ---------------------------------------------------------------------
12741: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
12742: @subsection Implementation Defined Options
12743: @c ---------------------------------------------------------------------
12744: @cindex implementation-defined options, memory-allocation words
12745: @cindex memory-allocation words, implementation-defined options
12746:
12747: @table @i
1.29 crook 12748: @item values and meaning of @i{ior}:
12749: @cindex @i{ior} values and meaning
12750: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 12751: intended as throw codes. They typically are in the range
12752: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 12753: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 12754:
12755: @end table
12756:
12757: @c =====================================================================
12758: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
12759: @section The optional Programming-Tools word set
12760: @c =====================================================================
12761: @cindex system documentation, programming-tools words
12762: @cindex programming-tools words, system documentation
12763:
12764: @menu
12765: * programming-idef:: Implementation Defined Options
12766: * programming-ambcond:: Ambiguous Conditions
12767: @end menu
12768:
12769:
12770: @c ---------------------------------------------------------------------
12771: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
12772: @subsection Implementation Defined Options
12773: @c ---------------------------------------------------------------------
12774: @cindex implementation-defined options, programming-tools words
12775: @cindex programming-tools words, implementation-defined options
12776:
12777: @table @i
12778: @item ending sequence for input following @code{;CODE} and @code{CODE}:
12779: @cindex @code{;CODE} ending sequence
12780: @cindex @code{CODE} ending sequence
12781: @code{END-CODE}
12782:
12783: @item manner of processing input following @code{;CODE} and @code{CODE}:
12784: @cindex @code{;CODE}, processing input
12785: @cindex @code{CODE}, processing input
12786: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
12787: the input is processed by the text interpreter, (starting) in interpret
12788: state.
12789:
12790: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
12791: @cindex @code{ASSEMBLER}, search order capability
12792: The ANS Forth search order word set.
12793:
12794: @item source and format of display by @code{SEE}:
12795: @cindex @code{SEE}, source and format of output
12796: The source for @code{see} is the intermediate code used by the inner
12797: interpreter. The current @code{see} tries to output Forth source code
12798: as well as possible.
12799:
12800: @end table
12801:
12802: @c ---------------------------------------------------------------------
12803: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
12804: @subsection Ambiguous conditions
12805: @c ---------------------------------------------------------------------
12806: @cindex programming-tools words, ambiguous conditions
12807: @cindex ambiguous conditions, programming-tools words
12808:
12809: @table @i
12810:
1.21 crook 12811: @item deleting the compilation word list (@code{FORGET}):
12812: @cindex @code{FORGET}, deleting the compilation word list
1.1 anton 12813: Not implemented (yet).
12814:
1.29 crook 12815: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
12816: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
12817: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1 anton 12818: @cindex control-flow stack underflow
12819: This typically results in an @code{abort"} with a descriptive error
12820: message (may change into a @code{-22 throw} (Control structure mismatch)
12821: in the future). You may also get a memory access error. If you are
12822: unlucky, this ambiguous condition is not caught.
12823:
1.29 crook 12824: @item @i{name} can't be found (@code{FORGET}):
12825: @cindex @code{FORGET}, @i{name} can't be found
1.1 anton 12826: Not implemented (yet).
12827:
1.29 crook 12828: @item @i{name} not defined via @code{CREATE}:
12829: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1 anton 12830: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
12831: the execution semantics of the last defined word no matter how it was
12832: defined.
12833:
12834: @item @code{POSTPONE} applied to @code{[IF]}:
12835: @cindex @code{POSTPONE} applied to @code{[IF]}
12836: @cindex @code{[IF]} and @code{POSTPONE}
12837: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
12838: equivalent to @code{[IF]}.
12839:
12840: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
12841: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
12842: Continue in the same state of conditional compilation in the next outer
12843: input source. Currently there is no warning to the user about this.
12844:
12845: @item removing a needed definition (@code{FORGET}):
12846: @cindex @code{FORGET}, removing a needed definition
12847: Not implemented (yet).
12848:
12849: @end table
12850:
12851:
12852: @c =====================================================================
12853: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
12854: @section The optional Search-Order word set
12855: @c =====================================================================
12856: @cindex system documentation, search-order words
12857: @cindex search-order words, system documentation
12858:
12859: @menu
12860: * search-idef:: Implementation Defined Options
12861: * search-ambcond:: Ambiguous Conditions
12862: @end menu
12863:
12864:
12865: @c ---------------------------------------------------------------------
12866: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
12867: @subsection Implementation Defined Options
12868: @c ---------------------------------------------------------------------
12869: @cindex implementation-defined options, search-order words
12870: @cindex search-order words, implementation-defined options
12871:
12872: @table @i
12873: @item maximum number of word lists in search order:
12874: @cindex maximum number of word lists in search order
12875: @cindex search order, maximum depth
12876: @code{s" wordlists" environment? drop .}. Currently 16.
12877:
12878: @item minimum search order:
12879: @cindex minimum search order
12880: @cindex search order, minimum
12881: @code{root root}.
12882:
12883: @end table
12884:
12885: @c ---------------------------------------------------------------------
12886: @node search-ambcond, , search-idef, The optional Search-Order word set
12887: @subsection Ambiguous conditions
12888: @c ---------------------------------------------------------------------
12889: @cindex search-order words, ambiguous conditions
12890: @cindex ambiguous conditions, search-order words
12891:
12892: @table @i
1.21 crook 12893: @item changing the compilation word list (during compilation):
12894: @cindex changing the compilation word list (during compilation)
12895: @cindex compilation word list, change before definition ends
12896: The word is entered into the word list that was the compilation word list
1.1 anton 12897: at the start of the definition. Any changes to the name field (e.g.,
12898: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
12899: are applied to the latest defined word (as reported by @code{last} or
1.21 crook 12900: @code{lastxt}), if possible, irrespective of the compilation word list.
1.1 anton 12901:
12902: @item search order empty (@code{previous}):
12903: @cindex @code{previous}, search order empty
1.26 crook 12904: @cindex vocstack empty, @code{previous}
1.1 anton 12905: @code{abort" Vocstack empty"}.
12906:
12907: @item too many word lists in search order (@code{also}):
12908: @cindex @code{also}, too many word lists in search order
1.26 crook 12909: @cindex vocstack full, @code{also}
1.1 anton 12910: @code{abort" Vocstack full"}.
12911:
12912: @end table
12913:
12914: @c ***************************************************************
12915: @node Model, Integrating Gforth, ANS conformance, Top
12916: @chapter Model
12917:
12918: This chapter has yet to be written. It will contain information, on
12919: which internal structures you can rely.
12920:
12921: @c ***************************************************************
12922: @node Integrating Gforth, Emacs and Gforth, Model, Top
12923: @chapter Integrating Gforth into C programs
12924:
12925: This is not yet implemented.
12926:
12927: Several people like to use Forth as scripting language for applications
12928: that are otherwise written in C, C++, or some other language.
12929:
12930: The Forth system ATLAST provides facilities for embedding it into
12931: applications; unfortunately it has several disadvantages: most
12932: importantly, it is not based on ANS Forth, and it is apparently dead
12933: (i.e., not developed further and not supported). The facilities
1.21 crook 12934: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1 anton 12935: making the switch should not be hard.
12936:
12937: We also tried to design the interface such that it can easily be
12938: implemented by other Forth systems, so that we may one day arrive at a
12939: standardized interface. Such a standard interface would allow you to
12940: replace the Forth system without having to rewrite C code.
12941:
12942: You embed the Gforth interpreter by linking with the library
12943: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
12944: global symbols in this library that belong to the interface, have the
12945: prefix @code{forth_}. (Global symbols that are used internally have the
12946: prefix @code{gforth_}).
12947:
12948: You can include the declarations of Forth types and the functions and
12949: variables of the interface with @code{#include <forth.h>}.
12950:
12951: Types.
12952:
12953: Variables.
12954:
12955: Data and FP Stack pointer. Area sizes.
12956:
12957: functions.
12958:
12959: forth_init(imagefile)
12960: forth_evaluate(string) exceptions?
12961: forth_goto(address) (or forth_execute(xt)?)
12962: forth_continue() (a corountining mechanism)
12963:
12964: Adding primitives.
12965:
12966: No checking.
12967:
12968: Signals?
12969:
12970: Accessing the Stacks
12971:
1.26 crook 12972: @c ******************************************************************
1.1 anton 12973: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
12974: @chapter Emacs and Gforth
12975: @cindex Emacs and Gforth
12976:
12977: @cindex @file{gforth.el}
12978: @cindex @file{forth.el}
12979: @cindex Rydqvist, Goran
12980: @cindex comment editing commands
12981: @cindex @code{\}, editing with Emacs
12982: @cindex debug tracer editing commands
12983: @cindex @code{~~}, removal with Emacs
12984: @cindex Forth mode in Emacs
12985: Gforth comes with @file{gforth.el}, an improved version of
12986: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26 crook 12987: improvements are:
12988:
12989: @itemize @bullet
12990: @item
12991: A better (but still not perfect) handling of indentation.
12992: @item
12993: Comment paragraph filling (@kbd{M-q})
12994: @item
12995: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
12996: @item
12997: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41 anton 12998: @item
12999: Support of the @code{info-lookup} feature for looking up the
13000: documentation of a word.
1.26 crook 13001: @end itemize
13002:
13003: I left the stuff I do not use alone, even though some of it only makes
13004: sense for TILE. To get a description of these features, enter Forth mode
13005: and type @kbd{C-h m}.
1.1 anton 13006:
13007: @cindex source location of error or debugging output in Emacs
13008: @cindex error output, finding the source location in Emacs
13009: @cindex debugging output, finding the source location in Emacs
13010: In addition, Gforth supports Emacs quite well: The source code locations
13011: given in error messages, debugging output (from @code{~~}) and failed
13012: assertion messages are in the right format for Emacs' compilation mode
13013: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13014: Manual}) so the source location corresponding to an error or other
13015: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13016: @kbd{C-c C-c} for the error under the cursor).
13017:
13018: @cindex @file{TAGS} file
13019: @cindex @file{etags.fs}
13020: @cindex viewing the source of a word in Emacs
1.43 anton 13021: @cindex @code{require}, placement in files
13022: @cindex @code{include}, placement in files
13023: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
1.26 crook 13024: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1 anton 13025: contains the definitions of all words defined afterwards. You can then
13026: find the source for a word using @kbd{M-.}. Note that emacs can use
13027: several tags files at the same time (e.g., one for the Gforth sources
13028: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13029: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13030: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43 anton 13031: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13032: with @file{etags.fs}, you should avoid putting definitions both before
13033: and after @code{require} etc., otherwise you will see the same file
13034: visited several times by commands like @code{tags-search}.
1.1 anton 13035:
1.41 anton 13036: @cindex viewing the documentation of a word in Emacs
13037: @cindex context-sensitive help
13038: Moreover, for words documented in this manual, you can look up the
13039: glossary entry quickly by using @kbd{C-h TAB}
13040: (@code{info-lookup-symbol}, see @pxref{Documentation, ,Documentation
13041: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
1.42 anton 13042: later and does not work for words containing @code{:}.
1.41 anton 13043:
13044:
1.1 anton 13045: @cindex @file{.emacs}
13046: To get all these benefits, add the following lines to your @file{.emacs}
13047: file:
13048:
13049: @example
13050: (autoload 'forth-mode "gforth.el")
13051: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13052: @end example
13053:
1.26 crook 13054: @c ******************************************************************
1.1 anton 13055: @node Image Files, Engine, Emacs and Gforth, Top
13056: @chapter Image Files
1.26 crook 13057: @cindex image file
13058: @cindex @file{.fi} files
1.1 anton 13059: @cindex precompiled Forth code
13060: @cindex dictionary in persistent form
13061: @cindex persistent form of dictionary
13062:
13063: An image file is a file containing an image of the Forth dictionary,
13064: i.e., compiled Forth code and data residing in the dictionary. By
13065: convention, we use the extension @code{.fi} for image files.
13066:
13067: @menu
1.18 anton 13068: * Image Licensing Issues:: Distribution terms for images.
13069: * Image File Background:: Why have image files?
1.29 crook 13070: * Non-Relocatable Image Files:: don't always work.
1.18 anton 13071: * Data-Relocatable Image Files:: are better.
1.29 crook 13072: * Fully Relocatable Image Files:: better yet.
1.18 anton 13073: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.29 crook 13074: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.18 anton 13075: * Modifying the Startup Sequence:: and turnkey applications.
1.1 anton 13076: @end menu
13077:
1.18 anton 13078: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13079: @section Image Licensing Issues
13080: @cindex license for images
13081: @cindex image license
13082:
13083: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13084: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13085: original image; i.e., according to copyright law it is a derived work of
13086: the original image.
13087:
13088: Since Gforth is distributed under the GNU GPL, the newly created image
13089: falls under the GNU GPL, too. In particular, this means that if you
13090: distribute the image, you have to make all of the sources for the image
13091: available, including those you wrote. For details see @ref{License, ,
13092: GNU General Public License (Section 3)}.
13093:
13094: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13095: contains only code compiled from the sources you gave it; if none of
13096: these sources is under the GPL, the terms discussed above do not apply
13097: to the image. However, if your image needs an engine (a gforth binary)
13098: that is under the GPL, you should make sure that you distribute both in
13099: a way that is at most a @emph{mere aggregation}, if you don't want the
13100: terms of the GPL to apply to the image.
13101:
13102: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1 anton 13103: @section Image File Background
13104: @cindex image file background
13105:
13106: Our Forth system consists not only of primitives, but also of
13107: definitions written in Forth. Since the Forth compiler itself belongs to
13108: those definitions, it is not possible to start the system with the
13109: primitives and the Forth source alone. Therefore we provide the Forth
1.26 crook 13110: code as an image file in nearly executable form. When Gforth starts up,
13111: a C routine loads the image file into memory, optionally relocates the
13112: addresses, then sets up the memory (stacks etc.) according to
13113: information in the image file, and (finally) starts executing Forth
13114: code.
1.1 anton 13115:
13116: The image file variants represent different compromises between the
13117: goals of making it easy to generate image files and making them
13118: portable.
13119:
13120: @cindex relocation at run-time
1.26 crook 13121: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1 anton 13122: run-time. This avoids many of the complications discussed below (image
13123: files are data relocatable without further ado), but costs performance
13124: (one addition per memory access).
13125:
13126: @cindex relocation at load-time
1.26 crook 13127: By contrast, the Gforth loader performs relocation at image load time. The
13128: loader also has to replace tokens that represent primitive calls with the
1.1 anton 13129: appropriate code-field addresses (or code addresses in the case of
13130: direct threading).
13131:
13132: There are three kinds of image files, with different degrees of
13133: relocatability: non-relocatable, data-relocatable, and fully relocatable
13134: image files.
13135:
13136: @cindex image file loader
13137: @cindex relocating loader
13138: @cindex loader for image files
13139: These image file variants have several restrictions in common; they are
13140: caused by the design of the image file loader:
13141:
13142: @itemize @bullet
13143: @item
13144: There is only one segment; in particular, this means, that an image file
13145: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26 crook 13146: them). The contents of the stacks are not represented, either.
1.1 anton 13147:
13148: @item
13149: The only kinds of relocation supported are: adding the same offset to
13150: all cells that represent data addresses; and replacing special tokens
13151: with code addresses or with pieces of machine code.
13152:
13153: If any complex computations involving addresses are performed, the
13154: results cannot be represented in the image file. Several applications that
13155: use such computations come to mind:
13156: @itemize @minus
13157: @item
13158: Hashing addresses (or data structures which contain addresses) for table
13159: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13160: purpose, you will have no problem, because the hash tables are
13161: recomputed automatically when the system is started. If you use your own
13162: hash tables, you will have to do something similar.
13163:
13164: @item
13165: There's a cute implementation of doubly-linked lists that uses
13166: @code{XOR}ed addresses. You could represent such lists as singly-linked
13167: in the image file, and restore the doubly-linked representation on
13168: startup.@footnote{In my opinion, though, you should think thrice before
13169: using a doubly-linked list (whatever implementation).}
13170:
13171: @item
13172: The code addresses of run-time routines like @code{docol:} cannot be
13173: represented in the image file (because their tokens would be replaced by
13174: machine code in direct threaded implementations). As a workaround,
13175: compute these addresses at run-time with @code{>code-address} from the
13176: executions tokens of appropriate words (see the definitions of
13177: @code{docol:} and friends in @file{kernel.fs}).
13178:
13179: @item
13180: On many architectures addresses are represented in machine code in some
13181: shifted or mangled form. You cannot put @code{CODE} words that contain
13182: absolute addresses in this form in a relocatable image file. Workarounds
13183: are representing the address in some relative form (e.g., relative to
13184: the CFA, which is present in some register), or loading the address from
13185: a place where it is stored in a non-mangled form.
13186: @end itemize
13187: @end itemize
13188:
13189: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13190: @section Non-Relocatable Image Files
13191: @cindex non-relocatable image files
1.26 crook 13192: @cindex image file, non-relocatable
1.1 anton 13193:
13194: These files are simple memory dumps of the dictionary. They are specific
13195: to the executable (i.e., @file{gforth} file) they were created
13196: with. What's worse, they are specific to the place on which the
13197: dictionary resided when the image was created. Now, there is no
13198: guarantee that the dictionary will reside at the same place the next
13199: time you start Gforth, so there's no guarantee that a non-relocatable
13200: image will work the next time (Gforth will complain instead of crashing,
13201: though).
13202:
13203: You can create a non-relocatable image file with
13204:
1.44 crook 13205:
1.1 anton 13206: doc-savesystem
13207:
1.44 crook 13208:
1.1 anton 13209: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13210: @section Data-Relocatable Image Files
13211: @cindex data-relocatable image files
1.26 crook 13212: @cindex image file, data-relocatable
1.1 anton 13213:
13214: These files contain relocatable data addresses, but fixed code addresses
13215: (instead of tokens). They are specific to the executable (i.e.,
13216: @file{gforth} file) they were created with. For direct threading on some
13217: architectures (e.g., the i386), data-relocatable images do not work. You
13218: get a data-relocatable image, if you use @file{gforthmi} with a
13219: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13220: Relocatable Image Files}).
13221:
13222: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13223: @section Fully Relocatable Image Files
13224: @cindex fully relocatable image files
1.26 crook 13225: @cindex image file, fully relocatable
1.1 anton 13226:
13227: @cindex @file{kern*.fi}, relocatability
13228: @cindex @file{gforth.fi}, relocatability
13229: These image files have relocatable data addresses, and tokens for code
13230: addresses. They can be used with different binaries (e.g., with and
13231: without debugging) on the same machine, and even across machines with
13232: the same data formats (byte order, cell size, floating point
13233: format). However, they are usually specific to the version of Gforth
13234: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13235: are fully relocatable.
13236:
13237: There are two ways to create a fully relocatable image file:
13238:
13239: @menu
1.29 crook 13240: * gforthmi:: The normal way
1.1 anton 13241: * cross.fs:: The hard way
13242: @end menu
13243:
13244: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
13245: @subsection @file{gforthmi}
13246: @cindex @file{comp-i.fs}
13247: @cindex @file{gforthmi}
13248:
13249: You will usually use @file{gforthmi}. If you want to create an
1.29 crook 13250: image @i{file} that contains everything you would load by invoking
13251: Gforth with @code{gforth @i{options}}, you simply say:
1.1 anton 13252: @example
1.29 crook 13253: gforthmi @i{file} @i{options}
1.1 anton 13254: @end example
13255:
13256: E.g., if you want to create an image @file{asm.fi} that has the file
13257: @file{asm.fs} loaded in addition to the usual stuff, you could do it
13258: like this:
13259:
13260: @example
13261: gforthmi asm.fi asm.fs
13262: @end example
13263:
1.27 crook 13264: @file{gforthmi} is implemented as a sh script and works like this: It
13265: produces two non-relocatable images for different addresses and then
13266: compares them. Its output reflects this: first you see the output (if
13267: any) of the two Gforth invocations that produce the nonrelocatable image
13268: files, then you see the output of the comparing program: It displays the
13269: offset used for data addresses and the offset used for code addresses;
1.1 anton 13270: moreover, for each cell that cannot be represented correctly in the
1.44 crook 13271: image files, it displays a line like this:
1.1 anton 13272:
13273: @example
13274: 78DC BFFFFA50 BFFFFA40
13275: @end example
13276:
13277: This means that at offset $78dc from @code{forthstart}, one input image
13278: contains $bffffa50, and the other contains $bffffa40. Since these cells
13279: cannot be represented correctly in the output image, you should examine
13280: these places in the dictionary and verify that these cells are dead
13281: (i.e., not read before they are written).
1.39 anton 13282:
13283: @cindex --application, @code{gforthmi} option
13284: If you insert the option @code{--application} in front of the image file
13285: name, you will get an image that uses the @code{--appl-image} option
13286: instead of the @code{--image-file} option (@pxref{Invoking
13287: Gforth}). When you execute such an image on Unix (by typing the image
13288: name as command), the Gforth engine will pass all options to the image
13289: instead of trying to interpret them as engine options.
1.1 anton 13290:
1.27 crook 13291: If you type @file{gforthmi} with no arguments, it prints some usage
13292: instructions.
13293:
1.1 anton 13294: @cindex @code{savesystem} during @file{gforthmi}
13295: @cindex @code{bye} during @file{gforthmi}
13296: @cindex doubly indirect threaded code
1.44 crook 13297: @cindex environment variables
13298: @cindex @code{GFORTHD} -- environment variable
13299: @cindex @code{GFORTH} -- environment variable
1.1 anton 13300: @cindex @code{gforth-ditc}
1.29 crook 13301: There are a few wrinkles: After processing the passed @i{options}, the
1.1 anton 13302: words @code{savesystem} and @code{bye} must be visible. A special doubly
13303: indirect threaded version of the @file{gforth} executable is used for
13304: creating the nonrelocatable images; you can pass the exact filename of
13305: this executable through the environment variable @code{GFORTHD}
13306: (default: @file{gforth-ditc}); if you pass a version that is not doubly
13307: indirect threaded, you will not get a fully relocatable image, but a
1.27 crook 13308: data-relocatable image (because there is no code address offset). The
13309: normal @file{gforth} executable is used for creating the relocatable
13310: image; you can pass the exact filename of this executable through the
13311: environment variable @code{GFORTH}.
1.1 anton 13312:
13313: @node cross.fs, , gforthmi, Fully Relocatable Image Files
13314: @subsection @file{cross.fs}
13315: @cindex @file{cross.fs}
13316: @cindex cross-compiler
13317: @cindex metacompiler
1.47 crook 13318: @cindex target compiler
1.1 anton 13319:
13320: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47 crook 13321: programming language (@pxref{Cross Compiler}).
1.1 anton 13322:
1.47 crook 13323: @code{cross} allows you to create image files for machines with
1.1 anton 13324: different data sizes and data formats than the one used for generating
13325: the image file. You can also use it to create an application image that
13326: does not contain a Forth compiler. These features are bought with
13327: restrictions and inconveniences in programming. E.g., addresses have to
13328: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
13329: order to make the code relocatable.
13330:
13331:
13332: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
13333: @section Stack and Dictionary Sizes
13334: @cindex image file, stack and dictionary sizes
13335: @cindex dictionary size default
13336: @cindex stack size default
13337:
13338: If you invoke Gforth with a command line flag for the size
13339: (@pxref{Invoking Gforth}), the size you specify is stored in the
13340: dictionary. If you save the dictionary with @code{savesystem} or create
13341: an image with @file{gforthmi}, this size will become the default
13342: for the resulting image file. E.g., the following will create a
1.21 crook 13343: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1 anton 13344:
13345: @example
13346: gforthmi gforth.fi -m 1M
13347: @end example
13348:
13349: In other words, if you want to set the default size for the dictionary
13350: and the stacks of an image, just invoke @file{gforthmi} with the
13351: appropriate options when creating the image.
13352:
13353: @cindex stack size, cache-friendly
13354: Note: For cache-friendly behaviour (i.e., good performance), you should
13355: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
13356: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
13357: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
13358:
13359: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
13360: @section Running Image Files
13361: @cindex running image files
13362: @cindex invoking image files
13363: @cindex image file invocation
13364:
13365: @cindex -i, invoke image file
13366: @cindex --image file, invoke image file
1.29 crook 13367: You can invoke Gforth with an image file @i{image} instead of the
1.1 anton 13368: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
13369: @example
1.29 crook 13370: gforth -i @i{image}
1.1 anton 13371: @end example
13372:
13373: @cindex executable image file
1.26 crook 13374: @cindex image file, executable
1.1 anton 13375: If your operating system supports starting scripts with a line of the
13376: form @code{#! ...}, you just have to type the image file name to start
13377: Gforth with this image file (note that the file extension @code{.fi} is
1.29 crook 13378: just a convention). I.e., to run Gforth with the image file @i{image},
13379: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27 crook 13380: This works because every @code{.fi} file starts with a line of this
13381: format:
13382:
13383: @example
13384: #! /usr/local/bin/gforth-0.4.0 -i
13385: @end example
13386:
13387: The file and pathname for the Gforth engine specified on this line is
13388: the specific Gforth executable that it was built against; i.e. the value
13389: of the environment variable @code{GFORTH} at the time that
13390: @file{gforthmi} was executed.
1.1 anton 13391:
1.27 crook 13392: You can make use of the same shell capability to make a Forth source
13393: file into an executable. For example, if you place this text in a file:
1.26 crook 13394:
13395: @example
13396: #! /usr/local/bin/gforth
13397:
13398: ." Hello, world" CR
13399: bye
13400: @end example
13401:
13402: @noindent
1.27 crook 13403: and then make the file executable (chmod +x in Unix), you can run it
1.26 crook 13404: directly from the command line. The sequence @code{#!} is used in two
13405: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29 crook 13406: system@footnote{The Unix kernel actually recognises two types of files:
13407: executable files and files of data, where the data is processed by an
13408: interpreter that is specified on the ``interpreter line'' -- the first
13409: line of the file, starting with the sequence #!. There may be a small
13410: limit (e.g., 32) on the number of characters that may be specified on
13411: the interpreter line.} secondly it is treated as a comment character by
13412: Gforth. Because of the second usage, a space is required between
13413: @code{#!} and the path to the executable.
1.27 crook 13414:
13415: The disadvantage of this latter technique, compared with using
13416: @file{gforthmi}, is that it is slower; the Forth source code is compiled
13417: on-the-fly, each time the program is invoked.
13418:
1.26 crook 13419:
1.1 anton 13420: doc-#!
13421:
1.44 crook 13422:
1.1 anton 13423: @node Modifying the Startup Sequence, , Running Image Files, Image Files
13424: @section Modifying the Startup Sequence
13425: @cindex startup sequence for image file
13426: @cindex image file initialization sequence
13427: @cindex initialization sequence of image file
13428:
13429: You can add your own initialization to the startup sequence through the
1.26 crook 13430: deferred word @code{'cold}. @code{'cold} is invoked just before the
13431: image-specific command line processing (by default, loading files and
13432: evaluating (@code{-e}) strings) starts.
1.1 anton 13433:
13434: A sequence for adding your initialization usually looks like this:
13435:
13436: @example
13437: :noname
13438: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
13439: ... \ your stuff
13440: ; IS 'cold
13441: @end example
13442:
13443: @cindex turnkey image files
1.26 crook 13444: @cindex image file, turnkey applications
1.1 anton 13445: You can make a turnkey image by letting @code{'cold} execute a word
13446: (your turnkey application) that never returns; instead, it exits Gforth
13447: via @code{bye} or @code{throw}.
13448:
13449: @cindex command-line arguments, access
13450: @cindex arguments on the command line, access
13451: You can access the (image-specific) command-line arguments through the
1.26 crook 13452: variables @code{argc} and @code{argv}. @code{arg} provides convenient
1.1 anton 13453: access to @code{argv}.
13454:
1.26 crook 13455: If @code{'cold} exits normally, Gforth processes the command-line
13456: arguments as files to be loaded and strings to be evaluated. Therefore,
13457: @code{'cold} should remove the arguments it has used in this case.
13458:
1.44 crook 13459:
13460:
1.26 crook 13461: doc-'cold
1.1 anton 13462: doc-argc
13463: doc-argv
13464: doc-arg
13465:
13466:
1.44 crook 13467:
1.1 anton 13468: @c ******************************************************************
1.13 pazsan 13469: @node Engine, Binding to System Library, Image Files, Top
1.1 anton 13470: @chapter Engine
13471: @cindex engine
13472: @cindex virtual machine
13473:
1.26 crook 13474: Reading this chapter is not necessary for programming with Gforth. It
1.1 anton 13475: may be helpful for finding your way in the Gforth sources.
13476:
13477: The ideas in this section have also been published in the papers
13478: @cite{ANS fig/GNU/??? Forth} (in German) by Bernd Paysan, presented at
13479: the Forth-Tagung '93 and @cite{A Portable Forth Engine} by M. Anton
13480: Ertl, presented at EuroForth '93; the latter is available at
1.47 crook 13481: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z}.
1.1 anton 13482:
13483: @menu
13484: * Portability::
13485: * Threading::
13486: * Primitives::
13487: * Performance::
13488: @end menu
13489:
13490: @node Portability, Threading, Engine, Engine
13491: @section Portability
13492: @cindex engine portability
13493:
1.26 crook 13494: An important goal of the Gforth Project is availability across a wide
13495: range of personal machines. fig-Forth, and, to a lesser extent, F83,
13496: achieved this goal by manually coding the engine in assembly language
13497: for several then-popular processors. This approach is very
13498: labor-intensive and the results are short-lived due to progress in
13499: computer architecture.
1.1 anton 13500:
13501: @cindex C, using C for the engine
13502: Others have avoided this problem by coding in C, e.g., Mitch Bradley
13503: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
13504: particularly popular for UNIX-based Forths due to the large variety of
13505: architectures of UNIX machines. Unfortunately an implementation in C
13506: does not mix well with the goals of efficiency and with using
13507: traditional techniques: Indirect or direct threading cannot be expressed
13508: in C, and switch threading, the fastest technique available in C, is
13509: significantly slower. Another problem with C is that it is very
13510: cumbersome to express double integer arithmetic.
13511:
13512: @cindex GNU C for the engine
13513: @cindex long long
13514: Fortunately, there is a portable language that does not have these
13515: limitations: GNU C, the version of C processed by the GNU C compiler
13516: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
13517: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
13518: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
13519: threading possible, its @code{long long} type (@pxref{Long Long, ,
13520: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
13521: double numbers@footnote{Unfortunately, long longs are not implemented
13522: properly on all machines (e.g., on alpha-osf1, long longs are only 64
13523: bits, the same size as longs (and pointers), but they should be twice as
1.4 anton 13524: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
1.1 anton 13525: C Manual}). So, we had to implement doubles in C after all. Still, on
13526: most machines we can use long longs and achieve better performance than
13527: with the emulation package.}. GNU C is available for free on all
13528: important (and many unimportant) UNIX machines, VMS, 80386s running
13529: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
13530: on all these machines.
13531:
13532: Writing in a portable language has the reputation of producing code that
13533: is slower than assembly. For our Forth engine we repeatedly looked at
13534: the code produced by the compiler and eliminated most compiler-induced
13535: inefficiencies by appropriate changes in the source code.
13536:
13537: @cindex explicit register declarations
13538: @cindex --enable-force-reg, configuration flag
13539: @cindex -DFORCE_REG
13540: However, register allocation cannot be portably influenced by the
13541: programmer, leading to some inefficiencies on register-starved
13542: machines. We use explicit register declarations (@pxref{Explicit Reg
13543: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
13544: improve the speed on some machines. They are turned on by using the
13545: configuration flag @code{--enable-force-reg} (@code{gcc} switch
13546: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
13547: machine, but also on the compiler version: On some machines some
13548: compiler versions produce incorrect code when certain explicit register
13549: declarations are used. So by default @code{-DFORCE_REG} is not used.
13550:
13551: @node Threading, Primitives, Portability, Engine
13552: @section Threading
13553: @cindex inner interpreter implementation
13554: @cindex threaded code implementation
13555:
13556: @cindex labels as values
13557: GNU C's labels as values extension (available since @code{gcc-2.0},
13558: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29 crook 13559: makes it possible to take the address of @i{label} by writing
13560: @code{&&@i{label}}. This address can then be used in a statement like
13561: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1 anton 13562: @code{goto x}.
13563:
1.26 crook 13564: @cindex @code{NEXT}, indirect threaded
1.1 anton 13565: @cindex indirect threaded inner interpreter
13566: @cindex inner interpreter, indirect threaded
1.26 crook 13567: With this feature an indirect threaded @code{NEXT} looks like:
1.1 anton 13568: @example
13569: cfa = *ip++;
13570: ca = *cfa;
13571: goto *ca;
13572: @end example
13573: @cindex instruction pointer
13574: For those unfamiliar with the names: @code{ip} is the Forth instruction
13575: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
13576: execution token and points to the code field of the next word to be
13577: executed; The @code{ca} (code address) fetched from there points to some
13578: executable code, e.g., a primitive or the colon definition handler
13579: @code{docol}.
13580:
1.26 crook 13581: @cindex @code{NEXT}, direct threaded
1.1 anton 13582: @cindex direct threaded inner interpreter
13583: @cindex inner interpreter, direct threaded
13584: Direct threading is even simpler:
13585: @example
13586: ca = *ip++;
13587: goto *ca;
13588: @end example
13589:
13590: Of course we have packaged the whole thing neatly in macros called
1.26 crook 13591: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1 anton 13592:
13593: @menu
13594: * Scheduling::
13595: * Direct or Indirect Threaded?::
13596: * DOES>::
13597: @end menu
13598:
13599: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
13600: @subsection Scheduling
13601: @cindex inner interpreter optimization
13602:
13603: There is a little complication: Pipelined and superscalar processors,
13604: i.e., RISC and some modern CISC machines can process independent
13605: instructions while waiting for the results of an instruction. The
13606: compiler usually reorders (schedules) the instructions in a way that
13607: achieves good usage of these delay slots. However, on our first tries
13608: the compiler did not do well on scheduling primitives. E.g., for
13609: @code{+} implemented as
13610: @example
13611: n=sp[0]+sp[1];
13612: sp++;
13613: sp[0]=n;
13614: NEXT;
13615: @end example
1.26 crook 13616: the @code{NEXT} comes strictly after the other code, i.e., there is nearly no
1.1 anton 13617: scheduling. After a little thought the problem becomes clear: The
1.21 crook 13618: compiler cannot know that @code{sp} and @code{ip} point to different
13619: addresses (and the version of @code{gcc} we used would not know it even
13620: if it was possible), so it could not move the load of the cfa above the
13621: store to the TOS. Indeed the pointers could be the same, if code on or
13622: very near the top of stack were executed. In the interest of speed we
13623: chose to forbid this probably unused ``feature'' and helped the compiler
1.26 crook 13624: in scheduling: @code{NEXT} is divided into the loading part (@code{NEXT_P1})
1.21 crook 13625: and the goto part (@code{NEXT_P2}). @code{+} now looks like:
1.1 anton 13626: @example
13627: n=sp[0]+sp[1];
13628: sp++;
13629: NEXT_P1;
13630: sp[0]=n;
13631: NEXT_P2;
13632: @end example
13633: This can be scheduled optimally by the compiler.
13634:
13635: This division can be turned off with the switch @code{-DCISC_NEXT}. This
13636: switch is on by default on machines that do not profit from scheduling
13637: (e.g., the 80386), in order to preserve registers.
13638:
13639: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
13640: @subsection Direct or Indirect Threaded?
13641: @cindex threading, direct or indirect?
13642:
13643: @cindex -DDIRECT_THREADED
13644: Both! After packaging the nasty details in macro definitions we
13645: realized that we could switch between direct and indirect threading by
13646: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
13647: defining a few machine-specific macros for the direct-threading case.
13648: On the Forth level we also offer access words that hide the
13649: differences between the threading methods (@pxref{Threading Words}).
13650:
13651: Indirect threading is implemented completely machine-independently.
13652: Direct threading needs routines for creating jumps to the executable
1.21 crook 13653: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
13654: machine-dependent, but they do not amount to many source lines. Therefore,
13655: even porting direct threading to a new machine requires little effort.
1.1 anton 13656:
13657: @cindex --enable-indirect-threaded, configuration flag
13658: @cindex --enable-direct-threaded, configuration flag
13659: The default threading method is machine-dependent. You can enforce a
13660: specific threading method when building Gforth with the configuration
13661: flag @code{--enable-direct-threaded} or
13662: @code{--enable-indirect-threaded}. Note that direct threading is not
13663: supported on all machines.
13664:
13665: @node DOES>, , Direct or Indirect Threaded?, Threading
13666: @subsection DOES>
13667: @cindex @code{DOES>} implementation
13668:
1.26 crook 13669: @cindex @code{dodoes} routine
13670: @cindex @code{DOES>}-code
1.1 anton 13671: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
13672: the chunk of code executed by every word defined by a
13673: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
13674: the Forth code to be executed, i.e. the code after the
1.26 crook 13675: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
1.1 anton 13676:
1.21 crook 13677: In fig-Forth the code field points directly to the @code{dodoes} and the
1.45 crook 13678: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
1.29 crook 13679: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
1.1 anton 13680: the Forth-79 and all later standards, because in fig-Forth this address
13681: lies in the body (which is illegal in these standards). However, by
13682: making the code field larger for all words this solution becomes legal
13683: again. We use this approach for the indirect threaded version and for
13684: direct threading on some machines. Leaving a cell unused in most words
13685: is a bit wasteful, but on the machines we are targeting this is hardly a
13686: problem. The other reason for having a code field size of two cells is
13687: to avoid having different image files for direct and indirect threaded
13688: systems (direct threaded systems require two-cell code fields on many
13689: machines).
13690:
1.26 crook 13691: @cindex @code{DOES>}-handler
1.1 anton 13692: The other approach is that the code field points or jumps to the cell
1.26 crook 13693: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
13694: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
13695: @code{DOES>}-code address by computing the code address, i.e., the address of
1.45 crook 13696: the jump to @code{dodoes}, and add the length of that jump field. A variant of
1.1 anton 13697: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
13698: return address (which can be found in the return register on RISCs) is
1.26 crook 13699: the @code{DOES>}-code address. Since the two cells available in the code field
1.1 anton 13700: are used up by the jump to the code address in direct threading on many
13701: architectures, we use this approach for direct threading on these
13702: architectures. We did not want to add another cell to the code field.
13703:
13704: @node Primitives, Performance, Threading, Engine
13705: @section Primitives
13706: @cindex primitives, implementation
13707: @cindex virtual machine instructions, implementation
13708:
13709: @menu
13710: * Automatic Generation::
13711: * TOS Optimization::
13712: * Produced code::
13713: @end menu
13714:
13715: @node Automatic Generation, TOS Optimization, Primitives, Primitives
13716: @subsection Automatic Generation
13717: @cindex primitives, automatic generation
13718:
13719: @cindex @file{prims2x.fs}
13720: Since the primitives are implemented in a portable language, there is no
13721: longer any need to minimize the number of primitives. On the contrary,
13722: having many primitives has an advantage: speed. In order to reduce the
13723: number of errors in primitives and to make programming them easier, we
13724: provide a tool, the primitive generator (@file{prims2x.fs}), that
13725: automatically generates most (and sometimes all) of the C code for a
13726: primitive from the stack effect notation. The source for a primitive
13727: has the following form:
13728:
13729: @cindex primitive source format
13730: @format
1.58 anton 13731: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
1.29 crook 13732: [@code{""}@i{glossary entry}@code{""}]
13733: @i{C code}
1.1 anton 13734: [@code{:}
1.29 crook 13735: @i{Forth code}]
1.1 anton 13736: @end format
13737:
13738: The items in brackets are optional. The category and glossary fields
13739: are there for generating the documentation, the Forth code is there
13740: for manual implementations on machines without GNU C. E.g., the source
13741: for the primitive @code{+} is:
13742: @example
1.58 anton 13743: + ( n1 n2 -- n ) core plus
1.1 anton 13744: n = n1+n2;
13745: @end example
13746:
13747: This looks like a specification, but in fact @code{n = n1+n2} is C
13748: code. Our primitive generation tool extracts a lot of information from
13749: the stack effect notations@footnote{We use a one-stack notation, even
13750: though we have separate data and floating-point stacks; The separate
13751: notation can be generated easily from the unified notation.}: The number
13752: of items popped from and pushed on the stack, their type, and by what
13753: name they are referred to in the C code. It then generates a C code
13754: prelude and postlude for each primitive. The final C code for @code{+}
13755: looks like this:
13756:
13757: @example
1.46 pazsan 13758: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
1.1 anton 13759: /* */ /* documentation */
13760: @{
13761: DEF_CA /* definition of variable ca (indirect threading) */
13762: Cell n1; /* definitions of variables */
13763: Cell n2;
13764: Cell n;
13765: n1 = (Cell) sp[1]; /* input */
13766: n2 = (Cell) TOS;
13767: sp += 1; /* stack adjustment */
13768: NAME("+") /* debugging output (with -DDEBUG) */
13769: @{
13770: n = n1+n2; /* C code taken from the source */
13771: @}
13772: NEXT_P1; /* NEXT part 1 */
13773: TOS = (Cell)n; /* output */
13774: NEXT_P2; /* NEXT part 2 */
13775: @}
13776: @end example
13777:
13778: This looks long and inefficient, but the GNU C compiler optimizes quite
13779: well and produces optimal code for @code{+} on, e.g., the R3000 and the
13780: HP RISC machines: Defining the @code{n}s does not produce any code, and
13781: using them as intermediate storage also adds no cost.
13782:
1.26 crook 13783: There are also other optimizations that are not illustrated by this
13784: example: assignments between simple variables are usually for free (copy
1.1 anton 13785: propagation). If one of the stack items is not used by the primitive
13786: (e.g. in @code{drop}), the compiler eliminates the load from the stack
13787: (dead code elimination). On the other hand, there are some things that
13788: the compiler does not do, therefore they are performed by
13789: @file{prims2x.fs}: The compiler does not optimize code away that stores
13790: a stack item to the place where it just came from (e.g., @code{over}).
13791:
13792: While programming a primitive is usually easy, there are a few cases
13793: where the programmer has to take the actions of the generator into
13794: account, most notably @code{?dup}, but also words that do not (always)
1.26 crook 13795: fall through to @code{NEXT}.
1.1 anton 13796:
13797: @node TOS Optimization, Produced code, Automatic Generation, Primitives
13798: @subsection TOS Optimization
13799: @cindex TOS optimization for primitives
13800: @cindex primitives, keeping the TOS in a register
13801:
13802: An important optimization for stack machine emulators, e.g., Forth
13803: engines, is keeping one or more of the top stack items in
1.29 crook 13804: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
13805: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1 anton 13806: @itemize @bullet
13807: @item
1.29 crook 13808: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1 anton 13809: due to fewer loads from and stores to the stack.
1.29 crook 13810: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
13811: @i{y<n}, due to additional moves between registers.
1.1 anton 13812: @end itemize
13813:
13814: @cindex -DUSE_TOS
13815: @cindex -DUSE_NO_TOS
13816: In particular, keeping one item in a register is never a disadvantage,
13817: if there are enough registers. Keeping two items in registers is a
13818: disadvantage for frequent words like @code{?branch}, constants,
13819: variables, literals and @code{i}. Therefore our generator only produces
13820: code that keeps zero or one items in registers. The generated C code
13821: covers both cases; the selection between these alternatives is made at
13822: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
13823: code for @code{+} is just a simple variable name in the one-item case,
13824: otherwise it is a macro that expands into @code{sp[0]}. Note that the
13825: GNU C compiler tries to keep simple variables like @code{TOS} in
13826: registers, and it usually succeeds, if there are enough registers.
13827:
13828: @cindex -DUSE_FTOS
13829: @cindex -DUSE_NO_FTOS
13830: The primitive generator performs the TOS optimization for the
13831: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
13832: operations the benefit of this optimization is even larger:
13833: floating-point operations take quite long on most processors, but can be
13834: performed in parallel with other operations as long as their results are
13835: not used. If the FP-TOS is kept in a register, this works. If
13836: it is kept on the stack, i.e., in memory, the store into memory has to
13837: wait for the result of the floating-point operation, lengthening the
13838: execution time of the primitive considerably.
13839:
13840: The TOS optimization makes the automatic generation of primitives a
13841: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
13842: @code{TOS} is not sufficient. There are some special cases to
13843: consider:
13844: @itemize @bullet
13845: @item In the case of @code{dup ( w -- w w )} the generator must not
13846: eliminate the store to the original location of the item on the stack,
13847: if the TOS optimization is turned on.
13848: @item Primitives with stack effects of the form @code{--}
1.29 crook 13849: @i{out1}...@i{outy} must store the TOS to the stack at the start.
13850: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1 anton 13851: must load the TOS from the stack at the end. But for the null stack
13852: effect @code{--} no stores or loads should be generated.
13853: @end itemize
13854:
13855: @node Produced code, , TOS Optimization, Primitives
13856: @subsection Produced code
13857: @cindex primitives, assembly code listing
13858:
13859: @cindex @file{engine.s}
13860: To see what assembly code is produced for the primitives on your machine
13861: with your compiler and your flag settings, type @code{make engine.s} and
13862: look at the resulting file @file{engine.s}.
13863:
13864: @node Performance, , Primitives, Engine
13865: @section Performance
13866: @cindex performance of some Forth interpreters
13867: @cindex engine performance
13868: @cindex benchmarking Forth systems
13869: @cindex Gforth performance
13870:
13871: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
13872: impossible to write a significantly faster engine.
13873:
13874: On register-starved machines like the 386 architecture processors
13875: improvements are possible, because @code{gcc} does not utilize the
13876: registers as well as a human, even with explicit register declarations;
13877: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
13878: and hand-tuned it for the 486; this system is 1.19 times faster on the
13879: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40 anton 13880: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
13881: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
13882: registers fit in real registers (and we can even afford to use the TOS
13883: optimization), resulting in a speedup of 1.14 on the sieve over the
13884: earlier results.
1.1 anton 13885:
13886: @cindex Win32Forth performance
13887: @cindex NT Forth performance
13888: @cindex eforth performance
13889: @cindex ThisForth performance
13890: @cindex PFE performance
13891: @cindex TILE performance
1.40 anton 13892: The potential advantage of assembly language implementations
1.1 anton 13893: is not necessarily realized in complete Forth systems: We compared
1.40 anton 13894: Gforth-0.4.9 (direct threaded, compiled with @code{gcc-2.95.1} and
1.1 anton 13895: @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
13896: 1994) and Eforth (with and without peephole (aka pinhole) optimization
13897: of the threaded code); all these systems were written in assembly
13898: language. We also compared Gforth with three systems written in C:
13899: PFE-0.9.14 (compiled with @code{gcc-2.6.3} with the default
13900: configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
1.21 crook 13901: -DUNROLL_NEXT}), ThisForth Beta (compiled with @code{gcc-2.6.3 -O3
13902: -fomit-frame-pointer}; ThisForth employs peephole optimization of the
1.1 anton 13903: threaded code) and TILE (compiled with @code{make opt}). We benchmarked
13904: Gforth, PFE, ThisForth and TILE on a 486DX2/66 under Linux. Kenneth
13905: O'Heskin kindly provided the results for Win32Forth and NT Forth on a
13906: 486DX2/66 with similar memory performance under Windows NT. Marcel
13907: Hendrix ported Eforth to Linux, then extended it to run the benchmarks,
13908: added the peephole optimizer, ran the benchmarks and reported the
13909: results.
1.40 anton 13910:
1.1 anton 13911: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
13912: matrix multiplication come from the Stanford integer benchmarks and have
13913: been translated into Forth by Martin Fraeman; we used the versions
13914: included in the TILE Forth package, but with bigger data set sizes; and
13915: a recursive Fibonacci number computation for benchmarking calling
13916: performance. The following table shows the time taken for the benchmarks
13917: scaled by the time taken by Gforth (in other words, it shows the speedup
13918: factor that Gforth achieved over the other systems).
13919:
13920: @example
1.40 anton 13921: relative Win32- NT eforth This-
1.1 anton 13922: time Gforth Forth Forth eforth +opt PFE Forth TILE
1.40 anton 13923: sieve 1.00 1.58 1.30 1.58 0.97 1.80 3.63 9.79
13924: bubble 1.00 1.55 1.67 1.75 1.04 1.78 4.59
13925: matmul 1.00 1.67 1.53 1.66 0.84 1.79 4.63
13926: fib 1.00 1.75 1.53 1.40 0.99 1.99 3.43 4.93
1.1 anton 13927: @end example
13928:
1.26 crook 13929: You may be quite surprised by the good performance of Gforth when
13930: compared with systems written in assembly language. One important reason
13931: for the disappointing performance of these other systems is probably
13932: that they are not written optimally for the 486 (e.g., they use the
13933: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
13934: but costly method for relocating the Forth image: like @code{cforth}, it
13935: computes the actual addresses at run time, resulting in two address
13936: computations per @code{NEXT} (@pxref{Image File Background}).
13937:
1.40 anton 13938: Only Eforth with the peephole optimizer performs comparable to
13939: Gforth. The speedups achieved with peephole optimization of threaded
13940: code are quite remarkable. Adding a peephole optimizer to Gforth should
13941: cause similar speedups.
1.1 anton 13942:
13943: The speedup of Gforth over PFE, ThisForth and TILE can be easily
13944: explained with the self-imposed restriction of the latter systems to
13945: standard C, which makes efficient threading impossible (however, the
1.4 anton 13946: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1 anton 13947: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
13948: Moreover, current C compilers have a hard time optimizing other aspects
13949: of the ThisForth and the TILE source.
13950:
1.26 crook 13951: The performance of Gforth on 386 architecture processors varies widely
13952: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
13953: allocate any of the virtual machine registers into real machine
13954: registers by itself and would not work correctly with explicit register
1.40 anton 13955: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
1.26 crook 13956: the Sieve) than the one measured above.
1.1 anton 13957:
1.26 crook 13958: Note that there have been several releases of Win32Forth since the
13959: release presented here, so the results presented above may have little
1.40 anton 13960: predictive value for the performance of Win32Forth today (results for
13961: the current release on an i486DX2/66 are welcome).
1.1 anton 13962:
13963: @cindex @file{Benchres}
13964: In @cite{Translating Forth to Efficient C} by M. Anton Ertl and Martin
13965: Maierhofer (presented at EuroForth '95), an indirect threaded version of
13966: Gforth is compared with Win32Forth, NT Forth, PFE, and ThisForth; that
1.40 anton 13967: version of Gforth is slower on a 486 than the direct threaded version
13968: used here. The paper available at
1.47 crook 13969: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz};
1.1 anton 13970: it also contains numbers for some native code systems. You can find a
13971: newer version of these measurements at
1.47 crook 13972: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1 anton 13973: find numbers for Gforth on various machines in @file{Benchres}.
13974:
1.26 crook 13975: @c ******************************************************************
1.13 pazsan 13976: @node Binding to System Library, Cross Compiler, Engine, Top
1.14 pazsan 13977: @chapter Binding to System Library
1.13 pazsan 13978:
13979: @node Cross Compiler, Bugs, Binding to System Library, Top
1.14 pazsan 13980: @chapter Cross Compiler
1.47 crook 13981: @cindex @file{cross.fs}
13982: @cindex cross-compiler
13983: @cindex metacompiler
13984: @cindex target compiler
1.13 pazsan 13985:
1.46 pazsan 13986: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
13987: mostly written in Forth, including crucial parts like the outer
13988: interpreter and compiler, it needs compiled Forth code to get
13989: started. The cross compiler allows to create new images for other
13990: architectures, even running under another Forth system.
1.13 pazsan 13991:
13992: @menu
13993: * Using the Cross Compiler::
13994: * How the Cross Compiler Works::
13995: @end menu
13996:
1.21 crook 13997: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14 pazsan 13998: @section Using the Cross Compiler
1.46 pazsan 13999:
14000: The cross compiler uses a language that resembles Forth, but isn't. The
14001: main difference is that you can execute Forth code after definition,
14002: while you usually can't execute the code compiled by cross, because the
14003: code you are compiling is typically for a different computer than the
14004: one you are compiling on.
14005:
14006: The Makefile is already set up to allow you to create kernels for new
14007: architectures with a simple make command. The generic kernels using the
14008: GCC compiled virtual machine are created in the normal build process
14009: with @code{make}. To create a embedded Gforth executable for e.g. the
14010: 8086 processor (running on a DOS machine), type
14011:
14012: @example
14013: make kernl-8086.fi
14014: @end example
14015:
14016: This will use the machine description from the @file{arch/8086}
14017: directory to create a new kernel. A machine file may look like that:
14018:
14019: @example
14020: \ Parameter for target systems 06oct92py
14021:
14022: 4 Constant cell \ cell size in bytes
14023: 2 Constant cell<< \ cell shift to bytes
14024: 5 Constant cell>bit \ cell shift to bits
14025: 8 Constant bits/char \ bits per character
14026: 8 Constant bits/byte \ bits per byte [default: 8]
14027: 8 Constant float \ bytes per float
14028: 8 Constant /maxalign \ maximum alignment in bytes
14029: false Constant bigendian \ byte order
14030: ( true=big, false=little )
14031:
14032: include machpc.fs \ feature list
14033: @end example
14034:
14035: This part is obligatory for the cross compiler itself, the feature list
14036: is used by the kernel to conditionally compile some features in and out,
14037: depending on whether the target supports these features.
14038:
14039: There are some optional features, if you define your own primitives,
14040: have an assembler, or need special, nonstandard preparation to make the
14041: boot process work. @code{asm-include} include an assembler,
14042: @code{prims-include} includes primitives, and @code{>boot} prepares for
14043: booting.
14044:
14045: @example
14046: : asm-include ." Include assembler" cr
14047: s" arch/8086/asm.fs" included ;
14048:
14049: : prims-include ." Include primitives" cr
14050: s" arch/8086/prim.fs" included ;
14051:
14052: : >boot ." Prepare booting" cr
14053: s" ' boot >body into-forth 1+ !" evaluate ;
14054: @end example
14055:
14056: These words are used as sort of macro during the cross compilation in
14057: the file @file{kernel/main.fs}. Instead of using this macros, it would
14058: be possible --- but more complicated --- to write a new kernel project
14059: file, too.
14060:
14061: @file{kernel/main.fs} expects the machine description file name on the
14062: stack; the cross compiler itself (@file{cross.fs}) assumes that either
14063: @code{mach-file} leaves a counted string on the stack, or
14064: @code{machine-file} leaves an address, count pair of the filename on the
14065: stack.
14066:
14067: The feature list is typically controlled using @code{SetValue}, generic
14068: files that are used by several projects can use @code{DefaultValue}
14069: instead. Both functions work like @code{Value}, when the value isn't
14070: defined, but @code{SetValue} works like @code{to} if the value is
14071: defined, and @code{DefaultValue} doesn't set anything, if the value is
14072: defined.
14073:
14074: @example
14075: \ generic mach file for pc gforth 03sep97jaw
14076:
14077: true DefaultValue NIL \ relocating
14078:
14079: >ENVIRON
14080:
14081: true DefaultValue file \ controls the presence of the
14082: \ file access wordset
14083: true DefaultValue OS \ flag to indicate a operating system
14084:
14085: true DefaultValue prims \ true: primitives are c-code
14086:
14087: true DefaultValue floating \ floating point wordset is present
14088:
14089: true DefaultValue glocals \ gforth locals are present
14090: \ will be loaded
14091: true DefaultValue dcomps \ double number comparisons
14092:
14093: true DefaultValue hash \ hashing primitives are loaded/present
14094:
14095: true DefaultValue xconds \ used together with glocals,
14096: \ special conditionals supporting gforths'
14097: \ local variables
14098: true DefaultValue header \ save a header information
14099:
14100: true DefaultValue backtrace \ enables backtrace code
14101:
14102: false DefaultValue ec
14103: false DefaultValue crlf
14104:
14105: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
14106:
14107: &16 KB DefaultValue stack-size
14108: &15 KB &512 + DefaultValue fstack-size
14109: &15 KB DefaultValue rstack-size
14110: &14 KB &512 + DefaultValue lstack-size
14111: @end example
1.13 pazsan 14112:
1.48 anton 14113: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
1.14 pazsan 14114: @section How the Cross Compiler Works
1.13 pazsan 14115:
14116: @node Bugs, Origin, Cross Compiler, Top
1.21 crook 14117: @appendix Bugs
1.1 anton 14118: @cindex bug reporting
14119:
1.21 crook 14120: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1 anton 14121:
14122: If you find a bug, please send a bug report to
1.33 anton 14123: @email{bug-gforth@@gnu.org}. A bug report should include this
1.21 crook 14124: information:
14125:
14126: @itemize @bullet
14127: @item
14128: The Gforth version used (it is announced at the start of an
14129: interactive Gforth session).
14130: @item
14131: The machine and operating system (on Unix
14132: systems @code{uname -a} will report this information).
14133: @item
14134: The installation options (send the file @file{config.status}).
14135: @item
14136: A complete list of changes (if any) you (or your installer) have made to the
14137: Gforth sources.
14138: @item
14139: A program (or a sequence of keyboard commands) that reproduces the bug.
14140: @item
14141: A description of what you think constitutes the buggy behaviour.
14142: @end itemize
1.1 anton 14143:
14144: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14145: to Report Bugs, gcc.info, GNU C Manual}.
14146:
14147:
1.21 crook 14148: @node Origin, Forth-related information, Bugs, Top
14149: @appendix Authors and Ancestors of Gforth
1.1 anton 14150:
14151: @section Authors and Contributors
14152: @cindex authors of Gforth
14153: @cindex contributors to Gforth
14154:
14155: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
14156: Ertl. The third major author was Jens Wilke. Lennart Benschop (who was
14157: one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
14158: with their continuous feedback. Lennart Benshop contributed
14159: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14160: support for calling C libraries. Helpful comments also came from Paul
14161: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.58 anton 14162: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14163: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14164: helpful comments from many others; thank you all, sorry for not listing
14165: you here (but digging through my mailbox to extract your names is on my
14166: to-do list). Since the release of Gforth-0.4.0 Neal Crook worked on the
14167: manual.
1.1 anton 14168:
14169: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
14170: and autoconf, among others), and to the creators of the Internet: Gforth
1.21 crook 14171: was developed across the Internet, and its authors did not meet
1.20 pazsan 14172: physically for the first 4 years of development.
1.1 anton 14173:
14174: @section Pedigree
1.26 crook 14175: @cindex pedigree of Gforth
1.1 anton 14176:
1.20 pazsan 14177: Gforth descends from bigFORTH (1993) and fig-Forth. Gforth and PFE (by
1.1 anton 14178: Dirk Zoller) will cross-fertilize each other. Of course, a significant
14179: part of the design of Gforth was prescribed by ANS Forth.
14180:
1.20 pazsan 14181: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1 anton 14182: 32 bit native code version of VolksForth for the Atari ST, written
14183: mostly by Dietrich Weineck.
14184:
14185: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
14186: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
14187: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
14188:
14189: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
14190: Forth-83 standard. !! Pedigree? When?
14191:
14192: A team led by Bill Ragsdale implemented fig-Forth on many processors in
14193: 1979. Robert Selzer and Bill Ragsdale developed the original
14194: implementation of fig-Forth for the 6502 based on microForth.
14195:
14196: The principal architect of microForth was Dean Sanderson. microForth was
14197: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
14198: the 1802, and subsequently implemented on the 8080, the 6800 and the
14199: Z80.
14200:
14201: All earlier Forth systems were custom-made, usually by Charles Moore,
14202: who discovered (as he puts it) Forth during the late 60s. The first full
14203: Forth existed in 1971.
14204:
14205: A part of the information in this section comes from @cite{The Evolution
14206: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
14207: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
14208: Notices 28(3), 1993. You can find more historical and genealogical
14209: information about Forth there.
14210:
1.21 crook 14211: @node Forth-related information, Word Index, Origin, Top
14212: @appendix Other Forth-related information
14213: @cindex Forth-related information
14214:
14215: @menu
14216: * Internet resources::
14217: * Books::
14218: * The Forth Interest Group::
14219: * Conferences::
14220: @end menu
14221:
14222:
14223: @node Internet resources, Books, Forth-related information, Forth-related information
14224: @section Internet resources
1.26 crook 14225: @cindex internet resources
1.21 crook 14226:
14227: @cindex comp.lang.forth
14228: @cindex frequently asked questions
1.45 crook 14229: There is an active news group (comp.lang.forth) discussing Forth and
1.21 crook 14230: Forth-related issues. A frequently-asked-questions (FAQ) list
1.45 crook 14231: is posted to the news group regularly, and archived at these sites:
1.21 crook 14232:
14233: @itemize @bullet
14234: @item
1.47 crook 14235: @uref{ftp://rtfm.mit.edu/pub/usenet-by-group/comp.lang.forth/}
1.21 crook 14236: @item
1.47 crook 14237: @uref{ftp://ftp.forth.org/pub/Forth/FAQ/}
1.21 crook 14238: @end itemize
14239:
14240: The FAQ list should be considered mandatory reading before posting to
1.45 crook 14241: the news group.
1.21 crook 14242:
14243: Here are some other web sites holding Forth-related material:
14244:
14245: @itemize @bullet
14246: @item
1.47 crook 14247: @uref{http://www.taygeta.com/forth.html} -- Skip Carter's Forth pages.
1.21 crook 14248: @item
1.47 crook 14249: @uref{http://www.jwdt.com/~paysan/gforth.html} -- the Gforth home page.
1.21 crook 14250: @item
1.47 crook 14251: @uref{http://www.minerva.com/uathena.htm} -- home of ANS Forth Standard.
1.21 crook 14252: @item
1.47 crook 14253: @uref{http://dec.bournemouth.ac.uk/forth/index.html} -- the Forth
1.21 crook 14254: Research page, including links to the Journal of Forth Application and
14255: Research (JFAR) and a searchable Forth bibliography.
14256: @end itemize
14257:
14258:
14259: @node Books, The Forth Interest Group, Internet resources, Forth-related information
14260: @section Books
1.26 crook 14261: @cindex books on Forth
1.21 crook 14262:
14263: As the Standard is relatively new, there are not many books out yet. It
14264: is not recommended to learn Forth by using Gforth and a book that is not
14265: written for ANS Forth, as you will not know your mistakes from the
14266: deviations of the book. However, books based on the Forth-83 standard
14267: should be ok, because ANS Forth is primarily an extension of Forth-83.
1.44 crook 14268: Refer to the Forth FAQ for details of Forth-related books.
1.21 crook 14269:
14270: @cindex standard document for ANS Forth
14271: @cindex ANS Forth document
14272: The definite reference if you want to write ANS Forth programs is, of
1.26 crook 14273: course, the ANS Forth document. It is available in printed form from the
1.21 crook 14274: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
14275: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about
14276: $200. You can also get it from Global Engineering Documents (Tel.: USA
14277: (800) 854-7179; Fax.: (303) 843-9880) for about $300.
14278:
14279: @cite{dpANS6}, the last draft of the standard, which was then submitted
14280: to ANSI for publication is available electronically and for free in some
14281: MS Word format, and it has been converted to HTML
1.47 crook 14282: (@uref{http://www.taygeta.com/forth/dpans.html}; this HTML version also
1.44 crook 14283: includes the answers to Requests for Interpretation (RFIs). Some
14284: pointers to these versions can be found through
1.47 crook 14285: @*@uref{http://www.complang.tuwien.ac.at/projects/forth.html}.
1.44 crook 14286:
1.21 crook 14287:
14288: @node The Forth Interest Group, Conferences, Books, Forth-related information
14289: @section The Forth Interest Group
14290: @cindex Forth interest group (FIG)
14291:
14292: The Forth Interest Group (FIG) is a world-wide, non-profit,
1.26 crook 14293: member-supported organisation. It publishes a regular magazine,
14294: @var{FORTH Dimensions}, and offers other benefits of membership. You can
14295: contact the FIG through their office email address:
14296: @email{office@@forth.org} or by visiting their web site at
1.47 crook 14297: @uref{http://www.forth.org/}. This web site also includes links to FIG
1.26 crook 14298: chapters in other countries and American cities
1.47 crook 14299: (@uref{http://www.forth.org/chapters.html}).
1.21 crook 14300:
1.48 anton 14301: @node Conferences, , The Forth Interest Group, Forth-related information
1.21 crook 14302: @section Conferences
14303: @cindex Conferences
14304:
14305: There are several regular conferences related to Forth. They are all
1.26 crook 14306: well-publicised in @var{FORTH Dimensions} and on the comp.lang.forth
1.45 crook 14307: news group:
1.21 crook 14308:
14309: @itemize @bullet
14310: @item
14311: FORML -- the Forth modification laboratory convenes every year near
14312: Monterey, California.
14313: @item
14314: The Rochester Forth Conference -- an annual conference traditionally
14315: held in Rochester, New York.
14316: @item
14317: EuroForth -- this European conference takes place annually.
14318: @end itemize
14319:
14320:
1.41 anton 14321: @node Word Index, Name Index, Forth-related information, Top
1.1 anton 14322: @unnumbered Word Index
14323:
1.26 crook 14324: This index is a list of Forth words that have ``glossary'' entries
14325: within this manual. Each word is listed with its stack effect and
14326: wordset.
1.1 anton 14327:
14328: @printindex fn
14329:
1.41 anton 14330: @node Name Index, Concept Index, Word Index, Top
14331: @unnumbered Name Index
14332:
14333: This index is a list of Forth words that have ``glossary'' entries
14334: within this manual.
14335:
14336: @printindex ky
14337:
14338: @node Concept Index, , Name Index, Top
1.1 anton 14339: @unnumbered Concept and Word Index
14340:
1.26 crook 14341: Not all entries listed in this index are present verbatim in the
14342: text. This index also duplicates, in abbreviated form, all of the words
14343: listed in the Word Index (only the names are listed for the words here).
1.1 anton 14344:
14345: @printindex cp
14346:
14347: @contents
14348: @bye
14349:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>