Annotation of gforth/doc/gforth.ds, revision 1.91
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:
1.66 anton 17: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
1.28 crook 18:
1.1 anton 19: @comment %**start of header (This is for running Texinfo on a region.)
20: @setfilename gforth.info
21: @settitle Gforth Manual
22: @dircategory GNU programming tools
23: @direntry
24: * Gforth: (gforth). A fast interpreter for the Forth language.
25: @end direntry
1.49 anton 26: @c The Texinfo manual also recommends doing this, but for Gforth it may
27: @c not make much sense
28: @c @dircategory Individual utilities
29: @c @direntry
30: @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
31: @c @end direntry
32:
1.1 anton 33: @comment @setchapternewpage odd
1.29 crook 34: @comment TODO this gets left in by HTML converter
1.12 anton 35: @macro progstyle {}
36: Programming style note:
1.3 anton 37: @end macro
1.48 anton 38:
39: @macro assignment {}
40: @table @i
41: @item Assignment:
42: @end macro
43: @macro endassignment {}
44: @end table
45: @end macro
46:
1.1 anton 47: @comment %**end of header (This is for running Texinfo on a region.)
48:
1.29 crook 49:
50: @comment ----------------------------------------------------------
51: @comment macros for beautifying glossary entries
52: @comment if these are used, need to strip them out for HTML converter
53: @comment else they get repeated verbatim in HTML output.
54: @comment .. not working yet.
55:
56: @macro GLOSS-START {}
57: @iftex
58: @ninerm
59: @end iftex
60: @end macro
61:
62: @macro GLOSS-END {}
63: @iftex
64: @rm
65: @end iftex
66: @end macro
67:
68: @comment ----------------------------------------------------------
69:
70:
1.10 anton 71: @include version.texi
72:
1.49 anton 73: @ifnottex
1.11 anton 74: This file documents Gforth @value{VERSION}
1.1 anton 75:
1.62 crook 76: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1 anton 77:
78: Permission is granted to make and distribute verbatim copies of
79: this manual provided the copyright notice and this permission notice
80: are preserved on all copies.
81:
82: @ignore
83: Permission is granted to process this file through TeX and print the
84: results, provided the printed document carries a copying permission
85: notice identical to this one except for the removal of this paragraph
86: (this paragraph not being relevant to the printed manual).
87:
88: @end ignore
89: Permission is granted to copy and distribute modified versions of this
90: manual under the conditions for verbatim copying, provided also that the
91: sections entitled "Distribution" and "General Public License" are
92: included exactly as in the original, and provided that the entire
93: resulting derived work is distributed under the terms of a permission
94: notice identical to this one.
95:
96: Permission is granted to copy and distribute translations of this manual
97: into another language, under the above conditions for modified versions,
98: except that the sections entitled "Distribution" and "General Public
99: License" may be included in a translation approved by the author instead
100: of in the original English.
1.49 anton 101: @end ifnottex
1.1 anton 102:
103: @finalout
104: @titlepage
105: @sp 10
106: @center @titlefont{Gforth Manual}
107: @sp 2
1.11 anton 108: @center for version @value{VERSION}
1.1 anton 109: @sp 2
1.34 anton 110: @center Neal Crook
1.1 anton 111: @center Anton Ertl
1.6 pazsan 112: @center Bernd Paysan
1.5 anton 113: @center Jens Wilke
1.1 anton 114: @sp 3
1.47 crook 115: @center This manual is permanently under construction and was last updated on 15-Mar-2000
1.1 anton 116:
117: @comment The following two commands start the copyright page.
118: @page
119: @vskip 0pt plus 1filll
1.62 crook 120: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1 anton 121:
122: @comment !! Published by ... or You can get a copy of this manual ...
123:
124: Permission is granted to make and distribute verbatim copies of
125: this manual provided the copyright notice and this permission notice
126: are preserved on all copies.
127:
128: Permission is granted to copy and distribute modified versions of this
129: manual under the conditions for verbatim copying, provided also that the
130: sections entitled "Distribution" and "General Public License" are
131: included exactly as in the original, and provided that the entire
132: resulting derived work is distributed under the terms of a permission
133: notice identical to this one.
134:
135: Permission is granted to copy and distribute translations of this manual
136: into another language, under the above conditions for modified versions,
137: except that the sections entitled "Distribution" and "General Public
138: License" may be included in a translation approved by the author instead
139: of in the original English.
140: @end titlepage
141:
142: @node Top, License, (dir), (dir)
1.49 anton 143: @ifnottex
1.1 anton 144: Gforth is a free implementation of ANS Forth available on many
1.11 anton 145: personal machines. This manual corresponds to version @value{VERSION}.
1.49 anton 146: @end ifnottex
1.1 anton 147:
148: @menu
1.21 crook 149: * License:: The GPL
1.26 crook 150: * Goals:: About the Gforth Project
1.29 crook 151: * Gforth Environment:: Starting (and exiting) Gforth
1.48 anton 152: * Tutorial:: Hands-on Forth Tutorial
1.21 crook 153: * Introduction:: An introduction to ANS Forth
1.1 anton 154: * Words:: Forth words available in Gforth
1.24 anton 155: * Error messages:: How to interpret them
1.1 anton 156: * Tools:: Programming tools
157: * ANS conformance:: Implementation-defined options etc.
1.65 anton 158: * Standard vs Extensions:: Should I use extensions?
1.1 anton 159: * Model:: The abstract machine of Gforth
160: * Integrating Gforth:: Forth as scripting language for applications
161: * Emacs and Gforth:: The Gforth Mode
162: * Image Files:: @code{.fi} files contain compiled code
163: * Engine:: The inner interpreter and the primitives
1.24 anton 164: * Binding to System Library::
1.13 pazsan 165: * Cross Compiler:: The Cross Compiler
1.1 anton 166: * Bugs:: How to report them
167: * Origin:: Authors and ancestors of Gforth
1.21 crook 168: * Forth-related information:: Books and places to look on the WWW
1.1 anton 169: * Word Index:: An item for each Forth word
170: * Concept Index:: A menu covering many topics
1.12 anton 171:
1.91 ! anton 172: @detailmenu
! 173: --- The Detailed Node Listing ---
1.12 anton 174:
1.29 crook 175: Gforth Environment
176:
1.32 anton 177: * Invoking Gforth:: Getting in
178: * Leaving Gforth:: Getting out
179: * Command-line editing::
1.48 anton 180: * Environment variables:: that affect how Gforth starts up
1.32 anton 181: * Gforth Files:: What gets installed and where
1.48 anton 182: * Startup speed:: When 35ms is not fast enough ...
183:
184: Forth Tutorial
185:
186: * Starting Gforth Tutorial::
187: * Syntax Tutorial::
188: * Crash Course Tutorial::
189: * Stack Tutorial::
190: * Arithmetics Tutorial::
191: * Stack Manipulation Tutorial::
192: * Using files for Forth code Tutorial::
193: * Comments Tutorial::
194: * Colon Definitions Tutorial::
195: * Decompilation Tutorial::
196: * Stack-Effect Comments Tutorial::
197: * Types Tutorial::
198: * Factoring Tutorial::
199: * Designing the stack effect Tutorial::
200: * Local Variables Tutorial::
201: * Conditional execution Tutorial::
202: * Flags and Comparisons Tutorial::
203: * General Loops Tutorial::
204: * Counted loops Tutorial::
205: * Recursion Tutorial::
206: * Leaving definitions or loops Tutorial::
207: * Return Stack Tutorial::
208: * Memory Tutorial::
209: * Characters and Strings Tutorial::
210: * Alignment Tutorial::
1.87 anton 211: * Files Tutorial::
1.48 anton 212: * Interpretation and Compilation Semantics and Immediacy Tutorial::
213: * Execution Tokens Tutorial::
214: * Exceptions Tutorial::
215: * Defining Words Tutorial::
216: * Arrays and Records Tutorial::
217: * POSTPONE Tutorial::
218: * Literal Tutorial::
219: * Advanced macros Tutorial::
220: * Compilation Tokens Tutorial::
221: * Wordlists and Search Order Tutorial::
1.29 crook 222:
1.24 anton 223: An Introduction to ANS Forth
224:
1.67 anton 225: * Introducing the Text Interpreter::
226: * Stacks and Postfix notation::
227: * Your first definition::
228: * How does that work?::
229: * Forth is written in Forth::
230: * Review - elements of a Forth system::
231: * Where to go next::
232: * Exercises::
1.24 anton 233:
1.12 anton 234: Forth Words
235:
236: * Notation::
1.65 anton 237: * Case insensitivity::
238: * Comments::
239: * Boolean Flags::
1.12 anton 240: * Arithmetic::
241: * Stack Manipulation::
242: * Memory::
243: * Control Structures::
244: * Defining Words::
1.65 anton 245: * Interpretation and Compilation Semantics::
1.47 crook 246: * Tokens for Words::
1.81 anton 247: * Compiling words::
1.65 anton 248: * The Text Interpreter::
249: * Word Lists::
250: * Environmental Queries::
1.12 anton 251: * Files::
252: * Blocks::
253: * Other I/O::
1.78 anton 254: * Locals::
255: * Structures::
256: * Object-oriented Forth::
1.12 anton 257: * Programming Tools::
258: * Assembler and Code Words::
259: * Threading Words::
1.65 anton 260: * Passing Commands to the OS::
261: * Keeping track of Time::
262: * Miscellaneous Words::
1.12 anton 263:
264: Arithmetic
265:
266: * Single precision::
1.67 anton 267: * Double precision:: Double-cell integer arithmetic
1.12 anton 268: * Bitwise operations::
1.67 anton 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
1.67 anton 295: * Arbitrary control structures::
296: * Calls and returns::
1.12 anton 297: * Exception Handling::
298:
299: Defining Words
300:
1.67 anton 301: * CREATE::
1.44 crook 302: * Variables:: Variables and user variables
1.67 anton 303: * Constants::
1.44 crook 304: * Values:: Initialised variables
1.67 anton 305: * Colon Definitions::
1.44 crook 306: * Anonymous Definitions:: Definitions without names
1.71 anton 307: * Supplying names:: Passing definition names as strings
1.67 anton 308: * User-defined Defining Words::
1.44 crook 309: * Deferred words:: Allow forward references
1.67 anton 310: * Aliases::
1.47 crook 311:
1.63 anton 312: User-defined Defining Words
313:
314: * CREATE..DOES> applications::
315: * CREATE..DOES> details::
316: * Advanced does> usage example::
1.91 ! anton 317: * @code{Const-does>}::
1.63 anton 318:
1.47 crook 319: Interpretation and Compilation Semantics
320:
1.67 anton 321: * Combined words::
1.12 anton 322:
1.71 anton 323: Tokens for Words
324:
325: * Execution token:: represents execution/interpretation semantics
326: * Compilation token:: represents compilation semantics
327: * Name token:: represents named words
328:
1.82 anton 329: Compiling words
330:
331: * Literals:: Compiling data values
332: * Macros:: Compiling words
333:
1.21 crook 334: The Text Interpreter
335:
1.67 anton 336: * Input Sources::
337: * Number Conversion::
338: * Interpret/Compile states::
339: * Interpreter Directives::
1.21 crook 340:
1.26 crook 341: Word Lists
342:
1.75 anton 343: * Vocabularies::
1.67 anton 344: * Why use word lists?::
1.75 anton 345: * Word list example::
1.26 crook 346:
347: Files
348:
1.48 anton 349: * Forth source files::
350: * General files::
351: * Search Paths::
352:
353: Search Paths
354:
1.75 anton 355: * Source Search Paths::
1.26 crook 356: * General Search Paths::
357:
358: Other I/O
359:
1.32 anton 360: * Simple numeric output:: Predefined formats
361: * Formatted numeric output:: Formatted (pictured) output
362: * String Formats:: How Forth stores strings in memory
1.67 anton 363: * Displaying characters and strings:: Other stuff
1.32 anton 364: * Input:: Input
1.26 crook 365:
366: Locals
367:
368: * Gforth locals::
369: * ANS Forth locals::
370:
371: Gforth locals
372:
373: * Where are locals visible by name?::
374: * How long do locals live?::
1.78 anton 375: * Locals programming style::
376: * Locals implementation::
1.26 crook 377:
1.12 anton 378: Structures
379:
380: * Why explicit structure support?::
381: * Structure Usage::
382: * Structure Naming Convention::
383: * Structure Implementation::
384: * Structure Glossary::
385:
386: Object-oriented Forth
387:
1.48 anton 388: * Why object-oriented programming?::
389: * Object-Oriented Terminology::
390: * Objects::
391: * OOF::
392: * Mini-OOF::
1.23 crook 393: * Comparison with other object models::
1.12 anton 394:
1.24 anton 395: The @file{objects.fs} model
1.12 anton 396:
397: * Properties of the Objects model::
398: * Basic Objects Usage::
1.41 anton 399: * The Objects base class::
1.12 anton 400: * Creating objects::
401: * Object-Oriented Programming Style::
402: * Class Binding::
403: * Method conveniences::
404: * Classes and Scoping::
1.41 anton 405: * Dividing classes::
1.12 anton 406: * Object Interfaces::
407: * Objects Implementation::
408: * Objects Glossary::
409:
1.24 anton 410: The @file{oof.fs} model
1.12 anton 411:
1.67 anton 412: * Properties of the OOF model::
413: * Basic OOF Usage::
414: * The OOF base class::
415: * Class Declaration::
416: * Class Implementation::
1.12 anton 417:
1.24 anton 418: The @file{mini-oof.fs} model
1.23 crook 419:
1.48 anton 420: * Basic Mini-OOF Usage::
421: * Mini-OOF Example::
422: * Mini-OOF Implementation::
1.23 crook 423:
1.78 anton 424: Programming Tools
425:
426: * Examining::
427: * Forgetting words::
428: * Debugging:: Simple and quick.
429: * Assertions:: Making your programs self-checking.
430: * Singlestep Debugger:: Executing your program word by word.
431:
432: Assembler and Code Words
433:
434: * Code and ;code::
435: * Common Assembler:: Assembler Syntax
436: * Common Disassembler::
437: * 386 Assembler:: Deviations and special cases
438: * Alpha Assembler:: Deviations and special cases
439: * MIPS assembler:: Deviations and special cases
440: * Other assemblers:: How to write them
441:
1.12 anton 442: Tools
443:
444: * ANS Report:: Report the words used, sorted by wordset.
445:
446: ANS conformance
447:
448: * The Core Words::
449: * The optional Block word set::
450: * The optional Double Number word set::
451: * The optional Exception word set::
452: * The optional Facility word set::
453: * The optional File-Access word set::
454: * The optional Floating-Point word set::
455: * The optional Locals word set::
456: * The optional Memory-Allocation word set::
457: * The optional Programming-Tools word set::
458: * The optional Search-Order word set::
459:
460: The Core Words
461:
462: * core-idef:: Implementation Defined Options
463: * core-ambcond:: Ambiguous Conditions
464: * core-other:: Other System Documentation
465:
466: The optional Block word set
467:
468: * block-idef:: Implementation Defined Options
469: * block-ambcond:: Ambiguous Conditions
470: * block-other:: Other System Documentation
471:
472: The optional Double Number word set
473:
474: * double-ambcond:: Ambiguous Conditions
475:
476: The optional Exception word set
477:
478: * exception-idef:: Implementation Defined Options
479:
480: The optional Facility word set
481:
482: * facility-idef:: Implementation Defined Options
483: * facility-ambcond:: Ambiguous Conditions
484:
485: The optional File-Access word set
486:
487: * file-idef:: Implementation Defined Options
488: * file-ambcond:: Ambiguous Conditions
489:
490: The optional Floating-Point word set
491:
492: * floating-idef:: Implementation Defined Options
493: * floating-ambcond:: Ambiguous Conditions
494:
495: The optional Locals word set
496:
497: * locals-idef:: Implementation Defined Options
498: * locals-ambcond:: Ambiguous Conditions
499:
500: The optional Memory-Allocation word set
501:
502: * memory-idef:: Implementation Defined Options
503:
504: The optional Programming-Tools word set
505:
506: * programming-idef:: Implementation Defined Options
507: * programming-ambcond:: Ambiguous Conditions
508:
509: The optional Search-Order word set
510:
511: * search-idef:: Implementation Defined Options
512: * search-ambcond:: Ambiguous Conditions
513:
514: Image Files
515:
1.24 anton 516: * Image Licensing Issues:: Distribution terms for images.
517: * Image File Background:: Why have image files?
1.67 anton 518: * Non-Relocatable Image Files:: don't always work.
1.24 anton 519: * Data-Relocatable Image Files:: are better.
1.67 anton 520: * Fully Relocatable Image Files:: better yet.
1.24 anton 521: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.32 anton 522: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.24 anton 523: * Modifying the Startup Sequence:: and turnkey applications.
1.12 anton 524:
525: Fully Relocatable Image Files
526:
1.27 crook 527: * gforthmi:: The normal way
1.12 anton 528: * cross.fs:: The hard way
529:
530: Engine
531:
532: * Portability::
533: * Threading::
534: * Primitives::
535: * Performance::
536:
537: Threading
538:
539: * Scheduling::
540: * Direct or Indirect Threaded?::
541: * DOES>::
542:
543: Primitives
544:
545: * Automatic Generation::
546: * TOS Optimization::
547: * Produced code::
1.13 pazsan 548:
549: Cross Compiler
550:
1.67 anton 551: * Using the Cross Compiler::
552: * How the Cross Compiler Works::
1.13 pazsan 553:
1.24 anton 554: @end detailmenu
1.1 anton 555: @end menu
556:
1.26 crook 557: @node License, Goals, Top, Top
1.1 anton 558: @unnumbered GNU GENERAL PUBLIC LICENSE
559: @center Version 2, June 1991
560:
561: @display
562: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
1.88 anton 563: 59 Temple Place, Suite 330, Boston, MA 02111, USA
1.1 anton 564:
565: Everyone is permitted to copy and distribute verbatim copies
566: of this license document, but changing it is not allowed.
567: @end display
568:
569: @unnumberedsec Preamble
570:
571: The licenses for most software are designed to take away your
572: freedom to share and change it. By contrast, the GNU General Public
573: License is intended to guarantee your freedom to share and change free
574: software---to make sure the software is free for all its users. This
575: General Public License applies to most of the Free Software
576: Foundation's software and to any other program whose authors commit to
577: using it. (Some other Free Software Foundation software is covered by
578: the GNU Library General Public License instead.) You can apply it to
579: your programs, too.
580:
581: When we speak of free software, we are referring to freedom, not
582: price. Our General Public Licenses are designed to make sure that you
583: have the freedom to distribute copies of free software (and charge for
584: this service if you wish), that you receive source code or can get it
585: if you want it, that you can change the software or use pieces of it
586: in new free programs; and that you know you can do these things.
587:
588: To protect your rights, we need to make restrictions that forbid
589: anyone to deny you these rights or to ask you to surrender the rights.
590: These restrictions translate to certain responsibilities for you if you
591: distribute copies of the software, or if you modify it.
592:
593: For example, if you distribute copies of such a program, whether
594: gratis or for a fee, you must give the recipients all the rights that
595: you have. You must make sure that they, too, receive or can get the
596: source code. And you must show them these terms so they know their
597: rights.
598:
599: We protect your rights with two steps: (1) copyright the software, and
600: (2) offer you this license which gives you legal permission to copy,
601: distribute and/or modify the software.
602:
603: Also, for each author's protection and ours, we want to make certain
604: that everyone understands that there is no warranty for this free
605: software. If the software is modified by someone else and passed on, we
606: want its recipients to know that what they have is not the original, so
607: that any problems introduced by others will not reflect on the original
608: authors' reputations.
609:
610: Finally, any free program is threatened constantly by software
611: patents. We wish to avoid the danger that redistributors of a free
612: program will individually obtain patent licenses, in effect making the
613: program proprietary. To prevent this, we have made it clear that any
614: patent must be licensed for everyone's free use or not licensed at all.
615:
616: The precise terms and conditions for copying, distribution and
617: modification follow.
618:
619: @iftex
620: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
621: @end iftex
1.49 anton 622: @ifnottex
1.1 anton 623: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1.49 anton 624: @end ifnottex
1.1 anton 625:
626: @enumerate 0
627: @item
628: This License applies to any program or other work which contains
629: a notice placed by the copyright holder saying it may be distributed
630: under the terms of this General Public License. The ``Program'', below,
631: refers to any such program or work, and a ``work based on the Program''
632: means either the Program or any derivative work under copyright law:
633: that is to say, a work containing the Program or a portion of it,
634: either verbatim or with modifications and/or translated into another
635: language. (Hereinafter, translation is included without limitation in
636: the term ``modification''.) Each licensee is addressed as ``you''.
637:
638: Activities other than copying, distribution and modification are not
639: covered by this License; they are outside its scope. The act of
640: running the Program is not restricted, and the output from the Program
641: is covered only if its contents constitute a work based on the
642: Program (independent of having been made by running the Program).
643: Whether that is true depends on what the Program does.
644:
645: @item
646: You may copy and distribute verbatim copies of the Program's
647: source code as you receive it, in any medium, provided that you
648: conspicuously and appropriately publish on each copy an appropriate
649: copyright notice and disclaimer of warranty; keep intact all the
650: notices that refer to this License and to the absence of any warranty;
651: and give any other recipients of the Program a copy of this License
652: along with the Program.
653:
654: You may charge a fee for the physical act of transferring a copy, and
655: you may at your option offer warranty protection in exchange for a fee.
656:
657: @item
658: You may modify your copy or copies of the Program or any portion
659: of it, thus forming a work based on the Program, and copy and
660: distribute such modifications or work under the terms of Section 1
661: above, provided that you also meet all of these conditions:
662:
663: @enumerate a
664: @item
665: You must cause the modified files to carry prominent notices
666: stating that you changed the files and the date of any change.
667:
668: @item
669: You must cause any work that you distribute or publish, that in
670: whole or in part contains or is derived from the Program or any
671: part thereof, to be licensed as a whole at no charge to all third
672: parties under the terms of this License.
673:
674: @item
675: If the modified program normally reads commands interactively
676: when run, you must cause it, when started running for such
677: interactive use in the most ordinary way, to print or display an
678: announcement including an appropriate copyright notice and a
679: notice that there is no warranty (or else, saying that you provide
680: a warranty) and that users may redistribute the program under
681: these conditions, and telling the user how to view a copy of this
682: License. (Exception: if the Program itself is interactive but
683: does not normally print such an announcement, your work based on
684: the Program is not required to print an announcement.)
685: @end enumerate
686:
687: These requirements apply to the modified work as a whole. If
688: identifiable sections of that work are not derived from the Program,
689: and can be reasonably considered independent and separate works in
690: themselves, then this License, and its terms, do not apply to those
691: sections when you distribute them as separate works. But when you
692: distribute the same sections as part of a whole which is a work based
693: on the Program, the distribution of the whole must be on the terms of
694: this License, whose permissions for other licensees extend to the
695: entire whole, and thus to each and every part regardless of who wrote it.
696:
697: Thus, it is not the intent of this section to claim rights or contest
698: your rights to work written entirely by you; rather, the intent is to
699: exercise the right to control the distribution of derivative or
700: collective works based on the Program.
701:
702: In addition, mere aggregation of another work not based on the Program
703: with the Program (or with a work based on the Program) on a volume of
704: a storage or distribution medium does not bring the other work under
705: the scope of this License.
706:
707: @item
708: You may copy and distribute the Program (or a work based on it,
709: under Section 2) in object code or executable form under the terms of
710: Sections 1 and 2 above provided that you also do one of the following:
711:
712: @enumerate a
713: @item
714: Accompany it with the complete corresponding machine-readable
715: source code, which must be distributed under the terms of Sections
716: 1 and 2 above on a medium customarily used for software interchange; or,
717:
718: @item
719: Accompany it with a written offer, valid for at least three
720: years, to give any third party, for a charge no more than your
721: cost of physically performing source distribution, a complete
722: machine-readable copy of the corresponding source code, to be
723: distributed under the terms of Sections 1 and 2 above on a medium
724: customarily used for software interchange; or,
725:
726: @item
727: Accompany it with the information you received as to the offer
728: to distribute corresponding source code. (This alternative is
729: allowed only for noncommercial distribution and only if you
730: received the program in object code or executable form with such
731: an offer, in accord with Subsection b above.)
732: @end enumerate
733:
734: The source code for a work means the preferred form of the work for
735: making modifications to it. For an executable work, complete source
736: code means all the source code for all modules it contains, plus any
737: associated interface definition files, plus the scripts used to
738: control compilation and installation of the executable. However, as a
739: special exception, the source code distributed need not include
740: anything that is normally distributed (in either source or binary
741: form) with the major components (compiler, kernel, and so on) of the
742: operating system on which the executable runs, unless that component
743: itself accompanies the executable.
744:
745: If distribution of executable or object code is made by offering
746: access to copy from a designated place, then offering equivalent
747: access to copy the source code from the same place counts as
748: distribution of the source code, even though third parties are not
749: compelled to copy the source along with the object code.
750:
751: @item
752: You may not copy, modify, sublicense, or distribute the Program
753: except as expressly provided under this License. Any attempt
754: otherwise to copy, modify, sublicense or distribute the Program is
755: void, and will automatically terminate your rights under this License.
756: However, parties who have received copies, or rights, from you under
757: this License will not have their licenses terminated so long as such
758: parties remain in full compliance.
759:
760: @item
761: You are not required to accept this License, since you have not
762: signed it. However, nothing else grants you permission to modify or
763: distribute the Program or its derivative works. These actions are
764: prohibited by law if you do not accept this License. Therefore, by
765: modifying or distributing the Program (or any work based on the
766: Program), you indicate your acceptance of this License to do so, and
767: all its terms and conditions for copying, distributing or modifying
768: the Program or works based on it.
769:
770: @item
771: Each time you redistribute the Program (or any work based on the
772: Program), the recipient automatically receives a license from the
773: original licensor to copy, distribute or modify the Program subject to
774: these terms and conditions. You may not impose any further
775: restrictions on the recipients' exercise of the rights granted herein.
776: You are not responsible for enforcing compliance by third parties to
777: this License.
778:
779: @item
780: If, as a consequence of a court judgment or allegation of patent
781: infringement or for any other reason (not limited to patent issues),
782: conditions are imposed on you (whether by court order, agreement or
783: otherwise) that contradict the conditions of this License, they do not
784: excuse you from the conditions of this License. If you cannot
785: distribute so as to satisfy simultaneously your obligations under this
786: License and any other pertinent obligations, then as a consequence you
787: may not distribute the Program at all. For example, if a patent
788: license would not permit royalty-free redistribution of the Program by
789: all those who receive copies directly or indirectly through you, then
790: the only way you could satisfy both it and this License would be to
791: refrain entirely from distribution of the Program.
792:
793: If any portion of this section is held invalid or unenforceable under
794: any particular circumstance, the balance of the section is intended to
795: apply and the section as a whole is intended to apply in other
796: circumstances.
797:
798: It is not the purpose of this section to induce you to infringe any
799: patents or other property right claims or to contest validity of any
800: such claims; this section has the sole purpose of protecting the
801: integrity of the free software distribution system, which is
802: implemented by public license practices. Many people have made
803: generous contributions to the wide range of software distributed
804: through that system in reliance on consistent application of that
805: system; it is up to the author/donor to decide if he or she is willing
806: to distribute software through any other system and a licensee cannot
807: impose that choice.
808:
809: This section is intended to make thoroughly clear what is believed to
810: be a consequence of the rest of this License.
811:
812: @item
813: If the distribution and/or use of the Program is restricted in
814: certain countries either by patents or by copyrighted interfaces, the
815: original copyright holder who places the Program under this License
816: may add an explicit geographical distribution limitation excluding
817: those countries, so that distribution is permitted only in or among
818: countries not thus excluded. In such case, this License incorporates
819: the limitation as if written in the body of this License.
820:
821: @item
822: The Free Software Foundation may publish revised and/or new versions
823: of the General Public License from time to time. Such new versions will
824: be similar in spirit to the present version, but may differ in detail to
825: address new problems or concerns.
826:
827: Each version is given a distinguishing version number. If the Program
828: specifies a version number of this License which applies to it and ``any
829: later version'', you have the option of following the terms and conditions
830: either of that version or of any later version published by the Free
831: Software Foundation. If the Program does not specify a version number of
832: this License, you may choose any version ever published by the Free Software
833: Foundation.
834:
835: @item
836: If you wish to incorporate parts of the Program into other free
837: programs whose distribution conditions are different, write to the author
838: to ask for permission. For software which is copyrighted by the Free
839: Software Foundation, write to the Free Software Foundation; we sometimes
840: make exceptions for this. Our decision will be guided by the two goals
841: of preserving the free status of all derivatives of our free software and
842: of promoting the sharing and reuse of software generally.
843:
844: @iftex
845: @heading NO WARRANTY
846: @end iftex
1.49 anton 847: @ifnottex
1.1 anton 848: @center NO WARRANTY
1.49 anton 849: @end ifnottex
1.1 anton 850:
851: @item
852: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
853: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
854: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
855: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
856: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
857: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
858: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
859: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
860: REPAIR OR CORRECTION.
861:
862: @item
863: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
864: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
865: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
866: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
867: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
868: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
869: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
870: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
871: POSSIBILITY OF SUCH DAMAGES.
872: @end enumerate
873:
874: @iftex
875: @heading END OF TERMS AND CONDITIONS
876: @end iftex
1.49 anton 877: @ifnottex
1.1 anton 878: @center END OF TERMS AND CONDITIONS
1.49 anton 879: @end ifnottex
1.1 anton 880:
881: @page
882: @unnumberedsec How to Apply These Terms to Your New Programs
883:
884: If you develop a new program, and you want it to be of the greatest
885: possible use to the public, the best way to achieve this is to make it
886: free software which everyone can redistribute and change under these terms.
887:
888: To do so, attach the following notices to the program. It is safest
889: to attach them to the start of each source file to most effectively
890: convey the exclusion of warranty; and each file should have at least
891: the ``copyright'' line and a pointer to where the full notice is found.
892:
893: @smallexample
894: @var{one line to give the program's name and a brief idea of what it does.}
895: Copyright (C) 19@var{yy} @var{name of author}
896:
897: This program is free software; you can redistribute it and/or modify
898: it under the terms of the GNU General Public License as published by
899: the Free Software Foundation; either version 2 of the License, or
900: (at your option) any later version.
901:
902: This program is distributed in the hope that it will be useful,
903: but WITHOUT ANY WARRANTY; without even the implied warranty of
904: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
905: GNU General Public License for more details.
906:
907: You should have received a copy of the GNU General Public License
908: along with this program; if not, write to the Free Software
1.88 anton 909: Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
1.1 anton 910: @end smallexample
911:
912: Also add information on how to contact you by electronic and paper mail.
913:
914: If the program is interactive, make it output a short notice like this
915: when it starts in an interactive mode:
916:
917: @smallexample
918: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
919: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
920: type `show w'.
921: This is free software, and you are welcome to redistribute it
922: under certain conditions; type `show c' for details.
923: @end smallexample
924:
925: The hypothetical commands @samp{show w} and @samp{show c} should show
926: the appropriate parts of the General Public License. Of course, the
927: commands you use may be called something other than @samp{show w} and
928: @samp{show c}; they could even be mouse-clicks or menu items---whatever
929: suits your program.
930:
931: You should also get your employer (if you work as a programmer) or your
932: school, if any, to sign a ``copyright disclaimer'' for the program, if
933: necessary. Here is a sample; alter the names:
934:
935: @smallexample
936: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
937: `Gnomovision' (which makes passes at compilers) written by James Hacker.
938:
939: @var{signature of Ty Coon}, 1 April 1989
940: Ty Coon, President of Vice
941: @end smallexample
942:
943: This General Public License does not permit incorporating your program into
944: proprietary programs. If your program is a subroutine library, you may
945: consider it more useful to permit linking proprietary applications with the
946: library. If this is what you want to do, use the GNU Library General
947: Public License instead of this License.
948:
949: @iftex
950: @unnumbered Preface
951: @cindex Preface
1.21 crook 952: This manual documents Gforth. Some introductory material is provided for
953: readers who are unfamiliar with Forth or who are migrating to Gforth
954: from other Forth compilers. However, this manual is primarily a
955: reference manual.
1.1 anton 956: @end iftex
957:
1.28 crook 958: @comment TODO much more blurb here.
1.26 crook 959:
960: @c ******************************************************************
1.29 crook 961: @node Goals, Gforth Environment, License, Top
1.26 crook 962: @comment node-name, next, previous, up
963: @chapter Goals of Gforth
964: @cindex goals of the Gforth project
965: The goal of the Gforth Project is to develop a standard model for
966: ANS Forth. This can be split into several subgoals:
967:
968: @itemize @bullet
969: @item
970: Gforth should conform to the ANS Forth Standard.
971: @item
972: It should be a model, i.e. it should define all the
973: implementation-dependent things.
974: @item
975: It should become standard, i.e. widely accepted and used. This goal
976: is the most difficult one.
977: @end itemize
978:
979: To achieve these goals Gforth should be
980: @itemize @bullet
981: @item
982: Similar to previous models (fig-Forth, F83)
983: @item
984: Powerful. It should provide for all the things that are considered
985: necessary today and even some that are not yet considered necessary.
986: @item
987: Efficient. It should not get the reputation of being exceptionally
988: slow.
989: @item
990: Free.
991: @item
992: Available on many machines/easy to port.
993: @end itemize
994:
995: Have we achieved these goals? Gforth conforms to the ANS Forth
996: standard. It may be considered a model, but we have not yet documented
997: which parts of the model are stable and which parts we are likely to
998: change. It certainly has not yet become a de facto standard, but it
999: appears to be quite popular. It has some similarities to and some
1000: differences from previous models. It has some powerful features, but not
1001: yet everything that we envisioned. We certainly have achieved our
1.65 anton 1002: execution speed goals (@pxref{Performance})@footnote{However, in 1998
1003: the bar was raised when the major commercial Forth vendors switched to
1004: native code compilers.}. It is free and available on many machines.
1.29 crook 1005:
1.26 crook 1006: @c ******************************************************************
1.48 anton 1007: @node Gforth Environment, Tutorial, Goals, Top
1.29 crook 1008: @chapter Gforth Environment
1009: @cindex Gforth environment
1.21 crook 1010:
1.45 crook 1011: Note: ultimately, the Gforth man page will be auto-generated from the
1.29 crook 1012: material in this chapter.
1.21 crook 1013:
1014: @menu
1.29 crook 1015: * Invoking Gforth:: Getting in
1016: * Leaving Gforth:: Getting out
1017: * Command-line editing::
1.48 anton 1018: * Environment variables:: that affect how Gforth starts up
1.29 crook 1019: * Gforth Files:: What gets installed and where
1.48 anton 1020: * Startup speed:: When 35ms is not fast enough ...
1.21 crook 1021: @end menu
1022:
1.49 anton 1023: For related information about the creation of images see @ref{Image Files}.
1.29 crook 1024:
1.21 crook 1025: @comment ----------------------------------------------
1.48 anton 1026: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29 crook 1027: @section Invoking Gforth
1028: @cindex invoking Gforth
1029: @cindex running Gforth
1030: @cindex command-line options
1031: @cindex options on the command line
1032: @cindex flags on the command line
1.21 crook 1033:
1.30 anton 1034: Gforth is made up of two parts; an executable ``engine'' (named
1035: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1036: will usually just say @code{gforth} -- this automatically loads the
1037: default image file @file{gforth.fi}. In many other cases the default
1038: Gforth image will be invoked like this:
1.21 crook 1039: @example
1.30 anton 1040: gforth [file | -e forth-code] ...
1.21 crook 1041: @end example
1.29 crook 1042: @noindent
1043: This interprets the contents of the files and the Forth code in the order they
1044: are given.
1.21 crook 1045:
1.30 anton 1046: In addition to the @file{gforth} engine, there is also an engine called
1047: @file{gforth-fast}, which is faster, but gives less informative error
1.89 anton 1048: messages (@pxref{Error messages}). You should use it for debugged,
1049: performance-critical programs.
1.30 anton 1050:
1.29 crook 1051: In general, the command line looks like this:
1.21 crook 1052:
1053: @example
1.30 anton 1054: gforth[-fast] [engine options] [image options]
1.21 crook 1055: @end example
1056:
1.30 anton 1057: The engine options must come before the rest of the command
1.29 crook 1058: line. They are:
1.26 crook 1059:
1.29 crook 1060: @table @code
1061: @cindex -i, command-line option
1062: @cindex --image-file, command-line option
1063: @item --image-file @i{file}
1064: @itemx -i @i{file}
1065: Loads the Forth image @i{file} instead of the default
1066: @file{gforth.fi} (@pxref{Image Files}).
1.21 crook 1067:
1.39 anton 1068: @cindex --appl-image, command-line option
1069: @item --appl-image @i{file}
1070: Loads the image @i{file} and leaves all further command-line arguments
1.65 anton 1071: to the image (instead of processing them as engine options). This is
1072: useful for building executable application images on Unix, built with
1.39 anton 1073: @code{gforthmi --application ...}.
1074:
1.29 crook 1075: @cindex --path, command-line option
1076: @cindex -p, command-line option
1077: @item --path @i{path}
1078: @itemx -p @i{path}
1079: Uses @i{path} for searching the image file and Forth source code files
1080: instead of the default in the environment variable @code{GFORTHPATH} or
1081: the path specified at installation time (e.g.,
1082: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1083: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21 crook 1084:
1.29 crook 1085: @cindex --dictionary-size, command-line option
1086: @cindex -m, command-line option
1087: @cindex @i{size} parameters for command-line options
1088: @cindex size of the dictionary and the stacks
1089: @item --dictionary-size @i{size}
1090: @itemx -m @i{size}
1091: Allocate @i{size} space for the Forth dictionary space instead of
1092: using the default specified in the image (typically 256K). The
1093: @i{size} specification for this and subsequent options consists of
1094: an integer and a unit (e.g.,
1095: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1096: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1097: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1098: @code{e} is used.
1.21 crook 1099:
1.29 crook 1100: @cindex --data-stack-size, command-line option
1101: @cindex -d, command-line option
1102: @item --data-stack-size @i{size}
1103: @itemx -d @i{size}
1104: Allocate @i{size} space for the data stack instead of using the
1105: default specified in the image (typically 16K).
1.21 crook 1106:
1.29 crook 1107: @cindex --return-stack-size, command-line option
1108: @cindex -r, command-line option
1109: @item --return-stack-size @i{size}
1110: @itemx -r @i{size}
1111: Allocate @i{size} space for the return stack instead of using the
1112: default specified in the image (typically 15K).
1.21 crook 1113:
1.29 crook 1114: @cindex --fp-stack-size, command-line option
1115: @cindex -f, command-line option
1116: @item --fp-stack-size @i{size}
1117: @itemx -f @i{size}
1118: Allocate @i{size} space for the floating point stack instead of
1119: using the default specified in the image (typically 15.5K). In this case
1120: the unit specifier @code{e} refers to floating point numbers.
1.21 crook 1121:
1.48 anton 1122: @cindex --locals-stack-size, command-line option
1123: @cindex -l, command-line option
1124: @item --locals-stack-size @i{size}
1125: @itemx -l @i{size}
1126: Allocate @i{size} space for the locals stack instead of using the
1127: default specified in the image (typically 14.5K).
1128:
1129: @cindex -h, command-line option
1130: @cindex --help, command-line option
1131: @item --help
1132: @itemx -h
1133: Print a message about the command-line options
1134:
1135: @cindex -v, command-line option
1136: @cindex --version, command-line option
1137: @item --version
1138: @itemx -v
1139: Print version and exit
1140:
1141: @cindex --debug, command-line option
1142: @item --debug
1143: Print some information useful for debugging on startup.
1144:
1145: @cindex --offset-image, command-line option
1146: @item --offset-image
1147: Start the dictionary at a slightly different position than would be used
1148: otherwise (useful for creating data-relocatable images,
1149: @pxref{Data-Relocatable Image Files}).
1150:
1151: @cindex --no-offset-im, command-line option
1152: @item --no-offset-im
1153: Start the dictionary at the normal position.
1154:
1155: @cindex --clear-dictionary, command-line option
1156: @item --clear-dictionary
1157: Initialize all bytes in the dictionary to 0 before loading the image
1158: (@pxref{Data-Relocatable Image Files}).
1159:
1160: @cindex --die-on-signal, command-line-option
1161: @item --die-on-signal
1162: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1163: or the segmentation violation SIGSEGV) by translating it into a Forth
1164: @code{THROW}. With this option, Gforth exits if it receives such a
1165: signal. This option is useful when the engine and/or the image might be
1166: severely broken (such that it causes another signal before recovering
1167: from the first); this option avoids endless loops in such cases.
1168: @end table
1169:
1170: @cindex loading files at startup
1171: @cindex executing code on startup
1172: @cindex batch processing with Gforth
1173: As explained above, the image-specific command-line arguments for the
1174: default image @file{gforth.fi} consist of a sequence of filenames and
1175: @code{-e @var{forth-code}} options that are interpreted in the sequence
1176: in which they are given. The @code{-e @var{forth-code}} or
1177: @code{--evaluate @var{forth-code}} option evaluates the Forth
1178: code. This option takes only one argument; if you want to evaluate more
1179: Forth words, you have to quote them or use @code{-e} several times. To exit
1180: after processing the command line (instead of entering interactive mode)
1181: append @code{-e bye} to the command line.
1182:
1183: @cindex versions, invoking other versions of Gforth
1184: If you have several versions of Gforth installed, @code{gforth} will
1185: invoke the version that was installed last. @code{gforth-@i{version}}
1186: invokes a specific version. If your environment contains the variable
1187: @code{GFORTHPATH}, you may want to override it by using the
1188: @code{--path} option.
1189:
1190: Not yet implemented:
1191: On startup the system first executes the system initialization file
1192: (unless the option @code{--no-init-file} is given; note that the system
1193: resulting from using this option may not be ANS Forth conformant). Then
1194: the user initialization file @file{.gforth.fs} is executed, unless the
1.62 crook 1195: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48 anton 1196: then in @file{~}, then in the normal path (see above).
1197:
1198:
1199:
1200: @comment ----------------------------------------------
1201: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1202: @section Leaving Gforth
1203: @cindex Gforth - leaving
1204: @cindex leaving Gforth
1205:
1206: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1207: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1208: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49 anton 1209: data are discarded. For ways of saving the state of the system before
1210: leaving Gforth see @ref{Image Files}.
1.48 anton 1211:
1212: doc-bye
1213:
1214:
1215: @comment ----------------------------------------------
1.65 anton 1216: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48 anton 1217: @section Command-line editing
1218: @cindex command-line editing
1219:
1220: Gforth maintains a history file that records every line that you type to
1221: the text interpreter. This file is preserved between sessions, and is
1222: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1223: repeatedly you can recall successively older commands from this (or
1224: previous) session(s). The full list of command-line editing facilities is:
1225:
1226: @itemize @bullet
1227: @item
1228: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1229: commands from the history buffer.
1230: @item
1231: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1232: from the history buffer.
1233: @item
1234: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1235: @item
1236: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1237: @item
1238: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1239: closing up the line.
1240: @item
1241: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1242: @item
1243: @kbd{Ctrl-a} to move the cursor to the start of the line.
1244: @item
1245: @kbd{Ctrl-e} to move the cursor to the end of the line.
1246: @item
1247: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1248: line.
1249: @item
1250: @key{TAB} to step through all possible full-word completions of the word
1251: currently being typed.
1252: @item
1.65 anton 1253: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1254: using @code{bye}).
1255: @item
1256: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1257: character under the cursor.
1.48 anton 1258: @end itemize
1259:
1260: When editing, displayable characters are inserted to the left of the
1261: cursor position; the line is always in ``insert'' (as opposed to
1262: ``overstrike'') mode.
1263:
1264: @cindex history file
1265: @cindex @file{.gforth-history}
1266: On Unix systems, the history file is @file{~/.gforth-history} by
1267: default@footnote{i.e. it is stored in the user's home directory.}. You
1268: can find out the name and location of your history file using:
1269:
1270: @example
1271: history-file type \ Unix-class systems
1272:
1273: history-file type \ Other systems
1274: history-dir type
1275: @end example
1276:
1277: If you enter long definitions by hand, you can use a text editor to
1278: paste them out of the history file into a Forth source file for reuse at
1279: a later time.
1280:
1281: Gforth never trims the size of the history file, so you should do this
1282: periodically, if necessary.
1283:
1284: @comment this is all defined in history.fs
1285: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1286: @comment chosen?
1287:
1288:
1289: @comment ----------------------------------------------
1.65 anton 1290: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48 anton 1291: @section Environment variables
1292: @cindex environment variables
1293:
1294: Gforth uses these environment variables:
1295:
1296: @itemize @bullet
1297: @item
1298: @cindex @code{GFORTHHIST} -- environment variable
1299: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1300: open/create the history file, @file{.gforth-history}. Default:
1301: @code{$HOME}.
1302:
1303: @item
1304: @cindex @code{GFORTHPATH} -- environment variable
1305: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1306: for Forth source-code files.
1307:
1308: @item
1309: @cindex @code{GFORTH} -- environment variable
1.49 anton 1310: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1311:
1312: @item
1313: @cindex @code{GFORTHD} -- environment variable
1.62 crook 1314: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1315:
1316: @item
1317: @cindex @code{TMP}, @code{TEMP} - environment variable
1318: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1319: location for the history file.
1320: @end itemize
1321:
1322: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1323: @comment mentioning these.
1324:
1325: All the Gforth environment variables default to sensible values if they
1326: are not set.
1327:
1328:
1329: @comment ----------------------------------------------
1330: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1331: @section Gforth files
1332: @cindex Gforth files
1333:
1334: When you install Gforth on a Unix system, it installs files in these
1335: locations by default:
1336:
1337: @itemize @bullet
1338: @item
1339: @file{/usr/local/bin/gforth}
1340: @item
1341: @file{/usr/local/bin/gforthmi}
1342: @item
1343: @file{/usr/local/man/man1/gforth.1} - man page.
1344: @item
1345: @file{/usr/local/info} - the Info version of this manual.
1346: @item
1347: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1348: @item
1349: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1350: @item
1351: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1352: @item
1353: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1354: @end itemize
1355:
1356: You can select different places for installation by using
1357: @code{configure} options (listed with @code{configure --help}).
1358:
1359: @comment ----------------------------------------------
1360: @node Startup speed, , Gforth Files, Gforth Environment
1361: @section Startup speed
1362: @cindex Startup speed
1363: @cindex speed, startup
1364:
1365: If Gforth is used for CGI scripts or in shell scripts, its startup
1366: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1367: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1368: system time.
1369:
1370: If startup speed is a problem, you may consider the following ways to
1371: improve it; or you may consider ways to reduce the number of startups
1.62 crook 1372: (for example, by using Fast-CGI).
1.48 anton 1373:
1374: The first step to improve startup speed is to statically link Gforth, by
1375: building it with @code{XLDFLAGS=-static}. This requires more memory for
1376: the code and will therefore slow down the first invocation, but
1377: subsequent invocations avoid the dynamic linking overhead. Another
1378: disadvantage is that Gforth won't profit from library upgrades. As a
1379: result, @code{gforth-static -e bye} takes about 17.1ms user and
1380: 8.2ms system time.
1381:
1382: The next step to improve startup speed is to use a non-relocatable image
1.65 anton 1383: (@pxref{Non-Relocatable Image Files}). You can create this image with
1.48 anton 1384: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1385: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1386: and a part of the copy-on-write overhead. The disadvantage is that the
1.62 crook 1387: non-relocatable image does not work if the OS gives Gforth a different
1.48 anton 1388: address for the dictionary, for whatever reason; so you better provide a
1389: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1390: bye} takes about 15.3ms user and 7.5ms system time.
1391:
1392: The final step is to disable dictionary hashing in Gforth. Gforth
1393: builds the hash table on startup, which takes much of the startup
1394: overhead. You can do this by commenting out the @code{include hash.fs}
1395: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1396: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1397: The disadvantages are that functionality like @code{table} and
1398: @code{ekey} is missing and that text interpretation (e.g., compiling)
1399: now takes much longer. So, you should only use this method if there is
1400: no significant text interpretation to perform (the script should be
1.62 crook 1401: compiled into the image, amongst other things). @code{gforth-static -i
1.48 anton 1402: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1403:
1404: @c ******************************************************************
1405: @node Tutorial, Introduction, Gforth Environment, Top
1406: @chapter Forth Tutorial
1407: @cindex Tutorial
1408: @cindex Forth Tutorial
1409:
1.67 anton 1410: @c Topics from nac's Introduction that could be mentioned:
1411: @c press <ret> after each line
1412: @c Prompt
1413: @c numbers vs. words in dictionary on text interpretation
1414: @c what happens on redefinition
1415: @c parsing words (in particular, defining words)
1416:
1.83 anton 1417: The difference of this chapter from the Introduction
1418: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
1419: be used while sitting in front of a computer, and covers much more
1420: material, but does not explain how the Forth system works.
1421:
1.62 crook 1422: This tutorial can be used with any ANS-compliant Forth; any
1423: Gforth-specific features are marked as such and you can skip them if you
1424: work with another Forth. This tutorial does not explain all features of
1425: Forth, just enough to get you started and give you some ideas about the
1426: facilities available in Forth. Read the rest of the manual and the
1427: standard when you are through this.
1.48 anton 1428:
1429: The intended way to use this tutorial is that you work through it while
1430: sitting in front of the console, take a look at the examples and predict
1431: what they will do, then try them out; if the outcome is not as expected,
1432: find out why (e.g., by trying out variations of the example), so you
1433: understand what's going on. There are also some assignments that you
1434: should solve.
1435:
1436: This tutorial assumes that you have programmed before and know what,
1437: e.g., a loop is.
1438:
1439: @c !! explain compat library
1440:
1441: @menu
1442: * Starting Gforth Tutorial::
1443: * Syntax Tutorial::
1444: * Crash Course Tutorial::
1445: * Stack Tutorial::
1446: * Arithmetics Tutorial::
1447: * Stack Manipulation Tutorial::
1448: * Using files for Forth code Tutorial::
1449: * Comments Tutorial::
1450: * Colon Definitions Tutorial::
1451: * Decompilation Tutorial::
1452: * Stack-Effect Comments Tutorial::
1453: * Types Tutorial::
1454: * Factoring Tutorial::
1455: * Designing the stack effect Tutorial::
1456: * Local Variables Tutorial::
1457: * Conditional execution Tutorial::
1458: * Flags and Comparisons Tutorial::
1459: * General Loops Tutorial::
1460: * Counted loops Tutorial::
1461: * Recursion Tutorial::
1462: * Leaving definitions or loops Tutorial::
1463: * Return Stack Tutorial::
1464: * Memory Tutorial::
1465: * Characters and Strings Tutorial::
1466: * Alignment Tutorial::
1.87 anton 1467: * Files Tutorial::
1.48 anton 1468: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1469: * Execution Tokens Tutorial::
1470: * Exceptions Tutorial::
1471: * Defining Words Tutorial::
1472: * Arrays and Records Tutorial::
1473: * POSTPONE Tutorial::
1474: * Literal Tutorial::
1475: * Advanced macros Tutorial::
1476: * Compilation Tokens Tutorial::
1477: * Wordlists and Search Order Tutorial::
1478: @end menu
1479:
1480: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1481: @section Starting Gforth
1.66 anton 1482: @cindex starting Gforth tutorial
1.48 anton 1483: You can start Gforth by typing its name:
1484:
1485: @example
1486: gforth
1487: @end example
1488:
1489: That puts you into interactive mode; you can leave Gforth by typing
1490: @code{bye}. While in Gforth, you can edit the command line and access
1491: the command line history with cursor keys, similar to bash.
1492:
1493:
1494: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1495: @section Syntax
1.66 anton 1496: @cindex syntax tutorial
1.48 anton 1497:
1498: A @dfn{word} is a sequence of arbitrary characters (expcept white
1499: space). Words are separated by white space. E.g., each of the
1500: following lines contains exactly one word:
1501:
1502: @example
1503: word
1504: !@@#$%^&*()
1505: 1234567890
1506: 5!a
1507: @end example
1508:
1509: A frequent beginner's error is to leave away necessary white space,
1510: resulting in an error like @samp{Undefined word}; so if you see such an
1511: error, check if you have put spaces wherever necessary.
1512:
1513: @example
1514: ." hello, world" \ correct
1515: ."hello, world" \ gives an "Undefined word" error
1516: @end example
1517:
1.65 anton 1518: Gforth and most other Forth systems ignore differences in case (they are
1.48 anton 1519: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1520: your system is case-sensitive, you may have to type all the examples
1521: given here in upper case.
1522:
1523:
1524: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1525: @section Crash Course
1526:
1527: Type
1528:
1529: @example
1530: 0 0 !
1531: here execute
1532: ' catch >body 20 erase abort
1533: ' (quit) >body 20 erase
1534: @end example
1535:
1536: The last two examples are guaranteed to destroy parts of Gforth (and
1537: most other systems), so you better leave Gforth afterwards (if it has
1538: not finished by itself). On some systems you may have to kill gforth
1539: from outside (e.g., in Unix with @code{kill}).
1540:
1541: Now that you know how to produce crashes (and that there's not much to
1542: them), let's learn how to produce meaningful programs.
1543:
1544:
1545: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1546: @section Stack
1.66 anton 1547: @cindex stack tutorial
1.48 anton 1548:
1549: The most obvious feature of Forth is the stack. When you type in a
1550: number, it is pushed on the stack. You can display the content of the
1551: stack with @code{.s}.
1552:
1553: @example
1554: 1 2 .s
1555: 3 .s
1556: @end example
1557:
1558: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1559: appear in @code{.s} output as they appeared in the input.
1560:
1561: You can print the top of stack element with @code{.}.
1562:
1563: @example
1564: 1 2 3 . . .
1565: @end example
1566:
1567: In general, words consume their stack arguments (@code{.s} is an
1568: exception).
1569:
1570: @assignment
1571: What does the stack contain after @code{5 6 7 .}?
1572: @endassignment
1573:
1574:
1575: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1576: @section Arithmetics
1.66 anton 1577: @cindex arithmetics tutorial
1.48 anton 1578:
1579: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1580: operate on the top two stack items:
1581:
1582: @example
1.67 anton 1583: 2 2 .s
1584: + .s
1585: .
1.48 anton 1586: 2 1 - .
1587: 7 3 mod .
1588: @end example
1589:
1590: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1591: as in the corresponding infix expression (this is generally the case in
1592: Forth).
1593:
1594: Parentheses are superfluous (and not available), because the order of
1595: the words unambiguously determines the order of evaluation and the
1596: operands:
1597:
1598: @example
1599: 3 4 + 5 * .
1600: 3 4 5 * + .
1601: @end example
1602:
1603: @assignment
1604: What are the infix expressions corresponding to the Forth code above?
1605: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1606: known as Postfix or RPN (Reverse Polish Notation).}.
1607: @endassignment
1608:
1609: To change the sign, use @code{negate}:
1610:
1611: @example
1612: 2 negate .
1613: @end example
1614:
1615: @assignment
1616: Convert -(-3)*4-5 to Forth.
1617: @endassignment
1618:
1619: @code{/mod} performs both @code{/} and @code{mod}.
1620:
1621: @example
1622: 7 3 /mod . .
1623: @end example
1624:
1.66 anton 1625: Reference: @ref{Arithmetic}.
1626:
1627:
1.48 anton 1628: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1629: @section Stack Manipulation
1.66 anton 1630: @cindex stack manipulation tutorial
1.48 anton 1631:
1632: Stack manipulation words rearrange the data on the stack.
1633:
1634: @example
1635: 1 .s drop .s
1636: 1 .s dup .s drop drop .s
1637: 1 2 .s over .s drop drop drop
1638: 1 2 .s swap .s drop drop
1639: 1 2 3 .s rot .s drop drop drop
1640: @end example
1641:
1642: These are the most important stack manipulation words. There are also
1643: variants that manipulate twice as many stack items:
1644:
1645: @example
1646: 1 2 3 4 .s 2swap .s 2drop 2drop
1647: @end example
1648:
1649: Two more stack manipulation words are:
1650:
1651: @example
1652: 1 2 .s nip .s drop
1653: 1 2 .s tuck .s 2drop drop
1654: @end example
1655:
1656: @assignment
1657: Replace @code{nip} and @code{tuck} with combinations of other stack
1658: manipulation words.
1659:
1660: @example
1661: Given: How do you get:
1662: 1 2 3 3 2 1
1663: 1 2 3 1 2 3 2
1664: 1 2 3 1 2 3 3
1665: 1 2 3 1 3 3
1666: 1 2 3 2 1 3
1667: 1 2 3 4 4 3 2 1
1668: 1 2 3 1 2 3 1 2 3
1669: 1 2 3 4 1 2 3 4 1 2
1670: 1 2 3
1671: 1 2 3 1 2 3 4
1672: 1 2 3 1 3
1673: @end example
1674: @endassignment
1675:
1676: @example
1677: 5 dup * .
1678: @end example
1679:
1680: @assignment
1681: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1682: Write a piece of Forth code that expects two numbers on the stack
1683: (@var{a} and @var{b}, with @var{b} on top) and computes
1684: @code{(a-b)(a+1)}.
1685: @endassignment
1686:
1.66 anton 1687: Reference: @ref{Stack Manipulation}.
1688:
1689:
1.48 anton 1690: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1691: @section Using files for Forth code
1.66 anton 1692: @cindex loading Forth code, tutorial
1693: @cindex files containing Forth code, tutorial
1.48 anton 1694:
1695: While working at the Forth command line is convenient for one-line
1696: examples and short one-off code, you probably want to store your source
1697: code in files for convenient editing and persistence. You can use your
1698: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1699: Gforth}) to create @var{file} and use
1700:
1701: @example
1702: s" @var{file}" included
1703: @end example
1704:
1705: to load it into your Forth system. The file name extension I use for
1706: Forth files is @samp{.fs}.
1707:
1708: You can easily start Gforth with some files loaded like this:
1709:
1710: @example
1711: gforth @var{file1} @var{file2}
1712: @end example
1713:
1714: If an error occurs during loading these files, Gforth terminates,
1715: whereas an error during @code{INCLUDED} within Gforth usually gives you
1716: a Gforth command line. Starting the Forth system every time gives you a
1717: clean start every time, without interference from the results of earlier
1718: tries.
1719:
1720: I often put all the tests in a file, then load the code and run the
1721: tests with
1722:
1723: @example
1724: gforth @var{code} @var{tests} -e bye
1725: @end example
1726:
1727: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1728: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1729: restart this command without ado.
1730:
1731: The advantage of this approach is that the tests can be repeated easily
1732: every time the program ist changed, making it easy to catch bugs
1733: introduced by the change.
1734:
1.66 anton 1735: Reference: @ref{Forth source files}.
1736:
1.48 anton 1737:
1738: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1739: @section Comments
1.66 anton 1740: @cindex comments tutorial
1.48 anton 1741:
1742: @example
1743: \ That's a comment; it ends at the end of the line
1744: ( Another comment; it ends here: ) .s
1745: @end example
1746:
1747: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1748: separated with white space from the following text.
1749:
1750: @example
1751: \This gives an "Undefined word" error
1752: @end example
1753:
1754: The first @code{)} ends a comment started with @code{(}, so you cannot
1755: nest @code{(}-comments; and you cannot comment out text containing a
1756: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1757: avoid @code{)} in word names.}.
1758:
1759: I use @code{\}-comments for descriptive text and for commenting out code
1760: of one or more line; I use @code{(}-comments for describing the stack
1761: effect, the stack contents, or for commenting out sub-line pieces of
1762: code.
1763:
1764: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1765: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1766: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1767: with @kbd{M-q}.
1768:
1.66 anton 1769: Reference: @ref{Comments}.
1770:
1.48 anton 1771:
1772: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1773: @section Colon Definitions
1.66 anton 1774: @cindex colon definitions, tutorial
1775: @cindex definitions, tutorial
1776: @cindex procedures, tutorial
1777: @cindex functions, tutorial
1.48 anton 1778:
1779: are similar to procedures and functions in other programming languages.
1780:
1781: @example
1782: : squared ( n -- n^2 )
1783: dup * ;
1784: 5 squared .
1785: 7 squared .
1786: @end example
1787:
1788: @code{:} starts the colon definition; its name is @code{squared}. The
1789: following comment describes its stack effect. The words @code{dup *}
1790: are not executed, but compiled into the definition. @code{;} ends the
1791: colon definition.
1792:
1793: The newly-defined word can be used like any other word, including using
1794: it in other definitions:
1795:
1796: @example
1797: : cubed ( n -- n^3 )
1798: dup squared * ;
1799: -5 cubed .
1800: : fourth-power ( n -- n^4 )
1801: squared squared ;
1802: 3 fourth-power .
1803: @end example
1804:
1805: @assignment
1806: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1807: @code{/mod} in terms of other Forth words, and check if they work (hint:
1808: test your tests on the originals first). Don't let the
1809: @samp{redefined}-Messages spook you, they are just warnings.
1810: @endassignment
1811:
1.66 anton 1812: Reference: @ref{Colon Definitions}.
1813:
1.48 anton 1814:
1815: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1816: @section Decompilation
1.66 anton 1817: @cindex decompilation tutorial
1818: @cindex see tutorial
1.48 anton 1819:
1820: You can decompile colon definitions with @code{see}:
1821:
1822: @example
1823: see squared
1824: see cubed
1825: @end example
1826:
1827: In Gforth @code{see} shows you a reconstruction of the source code from
1828: the executable code. Informations that were present in the source, but
1829: not in the executable code, are lost (e.g., comments).
1830:
1.65 anton 1831: You can also decompile the predefined words:
1832:
1833: @example
1834: see .
1835: see +
1836: @end example
1837:
1838:
1.48 anton 1839: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1840: @section Stack-Effect Comments
1.66 anton 1841: @cindex stack-effect comments, tutorial
1842: @cindex --, tutorial
1.48 anton 1843: By convention the comment after the name of a definition describes the
1844: stack effect: The part in from of the @samp{--} describes the state of
1845: the stack before the execution of the definition, i.e., the parameters
1846: that are passed into the colon definition; the part behind the @samp{--}
1847: is the state of the stack after the execution of the definition, i.e.,
1848: the results of the definition. The stack comment only shows the top
1849: stack items that the definition accesses and/or changes.
1850:
1851: You should put a correct stack effect on every definition, even if it is
1852: just @code{( -- )}. You should also add some descriptive comment to
1853: more complicated words (I usually do this in the lines following
1854: @code{:}). If you don't do this, your code becomes unreadable (because
1855: you have to work through every definition before you can undertsand
1856: any).
1857:
1858: @assignment
1859: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1860: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1861: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1.65 anton 1862: are done, you can compare your stack effects to those in this manual
1.48 anton 1863: (@pxref{Word Index}).
1864: @endassignment
1865:
1866: Sometimes programmers put comments at various places in colon
1867: definitions that describe the contents of the stack at that place (stack
1868: comments); i.e., they are like the first part of a stack-effect
1869: comment. E.g.,
1870:
1871: @example
1872: : cubed ( n -- n^3 )
1873: dup squared ( n n^2 ) * ;
1874: @end example
1875:
1876: In this case the stack comment is pretty superfluous, because the word
1877: is simple enough. If you think it would be a good idea to add such a
1878: comment to increase readability, you should also consider factoring the
1879: word into several simpler words (@pxref{Factoring Tutorial,,
1.60 anton 1880: Factoring}), which typically eliminates the need for the stack comment;
1.48 anton 1881: however, if you decide not to refactor it, then having such a comment is
1882: better than not having it.
1883:
1884: The names of the stack items in stack-effect and stack comments in the
1885: standard, in this manual, and in many programs specify the type through
1886: a type prefix, similar to Fortran and Hungarian notation. The most
1887: frequent prefixes are:
1888:
1889: @table @code
1890: @item n
1891: signed integer
1892: @item u
1893: unsigned integer
1894: @item c
1895: character
1896: @item f
1897: Boolean flags, i.e. @code{false} or @code{true}.
1898: @item a-addr,a-
1899: Cell-aligned address
1900: @item c-addr,c-
1901: Char-aligned address (note that a Char may have two bytes in Windows NT)
1902: @item xt
1903: Execution token, same size as Cell
1904: @item w,x
1905: Cell, can contain an integer or an address. It usually takes 32, 64 or
1906: 16 bits (depending on your platform and Forth system). A cell is more
1907: commonly known as machine word, but the term @emph{word} already means
1908: something different in Forth.
1909: @item d
1910: signed double-cell integer
1911: @item ud
1912: unsigned double-cell integer
1913: @item r
1914: Float (on the FP stack)
1915: @end table
1916:
1917: You can find a more complete list in @ref{Notation}.
1918:
1919: @assignment
1920: Write stack-effect comments for all definitions you have written up to
1921: now.
1922: @endassignment
1923:
1924:
1925: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1926: @section Types
1.66 anton 1927: @cindex types tutorial
1.48 anton 1928:
1929: In Forth the names of the operations are not overloaded; so similar
1930: operations on different types need different names; e.g., @code{+} adds
1931: integers, and you have to use @code{f+} to add floating-point numbers.
1932: The following prefixes are often used for related operations on
1933: different types:
1934:
1935: @table @code
1936: @item (none)
1937: signed integer
1938: @item u
1939: unsigned integer
1940: @item c
1941: character
1942: @item d
1943: signed double-cell integer
1944: @item ud, du
1945: unsigned double-cell integer
1946: @item 2
1947: two cells (not-necessarily double-cell numbers)
1948: @item m, um
1949: mixed single-cell and double-cell operations
1950: @item f
1951: floating-point (note that in stack comments @samp{f} represents flags,
1.66 anton 1952: and @samp{r} represents FP numbers).
1.48 anton 1953: @end table
1954:
1955: If there are no differences between the signed and the unsigned variant
1956: (e.g., for @code{+}), there is only the prefix-less variant.
1957:
1958: Forth does not perform type checking, neither at compile time, nor at
1959: run time. If you use the wrong oeration, the data are interpreted
1960: incorrectly:
1961:
1962: @example
1963: -1 u.
1964: @end example
1965:
1966: If you have only experience with type-checked languages until now, and
1967: have heard how important type-checking is, don't panic! In my
1968: experience (and that of other Forthers), type errors in Forth code are
1969: usually easy to find (once you get used to it), the increased vigilance
1970: of the programmer tends to catch some harder errors in addition to most
1971: type errors, and you never have to work around the type system, so in
1972: most situations the lack of type-checking seems to be a win (projects to
1973: add type checking to Forth have not caught on).
1974:
1975:
1976: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1977: @section Factoring
1.66 anton 1978: @cindex factoring tutorial
1.48 anton 1979:
1980: If you try to write longer definitions, you will soon find it hard to
1981: keep track of the stack contents. Therefore, good Forth programmers
1982: tend to write only short definitions (e.g., three lines). The art of
1983: finding meaningful short definitions is known as factoring (as in
1984: factoring polynomials).
1985:
1986: Well-factored programs offer additional advantages: smaller, more
1987: general words, are easier to test and debug and can be reused more and
1988: better than larger, specialized words.
1989:
1990: So, if you run into difficulties with stack management, when writing
1991: code, try to define meaningful factors for the word, and define the word
1992: in terms of those. Even if a factor contains only two words, it is
1993: often helpful.
1994:
1.65 anton 1995: Good factoring is not easy, and it takes some practice to get the knack
1996: for it; but even experienced Forth programmers often don't find the
1997: right solution right away, but only when rewriting the program. So, if
1998: you don't come up with a good solution immediately, keep trying, don't
1999: despair.
1.48 anton 2000:
2001: @c example !!
2002:
2003:
2004: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
2005: @section Designing the stack effect
1.66 anton 2006: @cindex Stack effect design, tutorial
2007: @cindex design of stack effects, tutorial
1.48 anton 2008:
2009: In other languages you can use an arbitrary order of parameters for a
1.65 anton 2010: function; and since there is only one result, you don't have to deal with
1.48 anton 2011: the order of results, either.
2012:
2013: In Forth (and other stack-based languages, e.g., Postscript) the
2014: parameter and result order of a definition is important and should be
2015: designed well. The general guideline is to design the stack effect such
2016: that the word is simple to use in most cases, even if that complicates
2017: the implementation of the word. Some concrete rules are:
2018:
2019: @itemize @bullet
2020:
2021: @item
2022: Words consume all of their parameters (e.g., @code{.}).
2023:
2024: @item
2025: If there is a convention on the order of parameters (e.g., from
2026: mathematics or another programming language), stick with it (e.g.,
2027: @code{-}).
2028:
2029: @item
2030: If one parameter usually requires only a short computation (e.g., it is
2031: a constant), pass it on the top of the stack. Conversely, parameters
2032: that usually require a long sequence of code to compute should be passed
2033: as the bottom (i.e., first) parameter. This makes the code easier to
2034: read, because reader does not need to keep track of the bottom item
2035: through a long sequence of code (or, alternatively, through stack
1.49 anton 2036: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48 anton 2037: address on top of the stack because it is usually simpler to compute
2038: than the stored value (often the address is just a variable).
2039:
2040: @item
2041: Similarly, results that are usually consumed quickly should be returned
2042: on the top of stack, whereas a result that is often used in long
2043: computations should be passed as bottom result. E.g., the file words
2044: like @code{open-file} return the error code on the top of stack, because
2045: it is usually consumed quickly by @code{throw}; moreover, the error code
2046: has to be checked before doing anything with the other results.
2047:
2048: @end itemize
2049:
2050: These rules are just general guidelines, don't lose sight of the overall
2051: goal to make the words easy to use. E.g., if the convention rule
2052: conflicts with the computation-length rule, you might decide in favour
2053: of the convention if the word will be used rarely, and in favour of the
2054: computation-length rule if the word will be used frequently (because
2055: with frequent use the cost of breaking the computation-length rule would
2056: be quite high, and frequent use makes it easier to remember an
2057: unconventional order).
2058:
2059: @c example !! structure package
2060:
1.65 anton 2061:
1.48 anton 2062: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2063: @section Local Variables
1.66 anton 2064: @cindex local variables, tutorial
1.48 anton 2065:
2066: You can define local variables (@emph{locals}) in a colon definition:
2067:
2068: @example
2069: : swap @{ a b -- b a @}
2070: b a ;
2071: 1 2 swap .s 2drop
2072: @end example
2073:
2074: (If your Forth system does not support this syntax, include
2075: @file{compat/anslocals.fs} first).
2076:
2077: In this example @code{@{ a b -- b a @}} is the locals definition; it
2078: takes two cells from the stack, puts the top of stack in @code{b} and
2079: the next stack element in @code{a}. @code{--} starts a comment ending
2080: with @code{@}}. After the locals definition, using the name of the
2081: local will push its value on the stack. You can leave the comment
2082: part (@code{-- b a}) away:
2083:
2084: @example
2085: : swap ( x1 x2 -- x2 x1 )
2086: @{ a b @} b a ;
2087: @end example
2088:
2089: In Gforth you can have several locals definitions, anywhere in a colon
2090: definition; in contrast, in a standard program you can have only one
2091: locals definition per colon definition, and that locals definition must
2092: be outside any controll structure.
2093:
2094: With locals you can write slightly longer definitions without running
2095: into stack trouble. However, I recommend trying to write colon
2096: definitions without locals for exercise purposes to help you gain the
2097: essential factoring skills.
2098:
2099: @assignment
2100: Rewrite your definitions until now with locals
2101: @endassignment
2102:
1.66 anton 2103: Reference: @ref{Locals}.
2104:
1.48 anton 2105:
2106: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2107: @section Conditional execution
1.66 anton 2108: @cindex conditionals, tutorial
2109: @cindex if, tutorial
1.48 anton 2110:
2111: In Forth you can use control structures only inside colon definitions.
2112: An @code{if}-structure looks like this:
2113:
2114: @example
2115: : abs ( n1 -- +n2 )
2116: dup 0 < if
2117: negate
2118: endif ;
2119: 5 abs .
2120: -5 abs .
2121: @end example
2122:
2123: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2124: the following code is performed, otherwise execution continues after the
1.51 pazsan 2125: @code{endif} (or @code{else}). @code{<} compares the top two stack
1.48 anton 2126: elements and prioduces a flag:
2127:
2128: @example
2129: 1 2 < .
2130: 2 1 < .
2131: 1 1 < .
2132: @end example
2133:
2134: Actually the standard name for @code{endif} is @code{then}. This
2135: tutorial presents the examples using @code{endif}, because this is often
2136: less confusing for people familiar with other programming languages
2137: where @code{then} has a different meaning. If your system does not have
2138: @code{endif}, define it with
2139:
2140: @example
2141: : endif postpone then ; immediate
2142: @end example
2143:
2144: You can optionally use an @code{else}-part:
2145:
2146: @example
2147: : min ( n1 n2 -- n )
2148: 2dup < if
2149: drop
2150: else
2151: nip
2152: endif ;
2153: 2 3 min .
2154: 3 2 min .
2155: @end example
2156:
2157: @assignment
2158: Write @code{min} without @code{else}-part (hint: what's the definition
2159: of @code{nip}?).
2160: @endassignment
2161:
1.66 anton 2162: Reference: @ref{Selection}.
2163:
1.48 anton 2164:
2165: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2166: @section Flags and Comparisons
1.66 anton 2167: @cindex flags tutorial
2168: @cindex comparison tutorial
1.48 anton 2169:
2170: In a false-flag all bits are clear (0 when interpreted as integer). In
2171: a canonical true-flag all bits are set (-1 as a twos-complement signed
2172: integer); in many contexts (e.g., @code{if}) any non-zero value is
2173: treated as true flag.
2174:
2175: @example
2176: false .
2177: true .
2178: true hex u. decimal
2179: @end example
2180:
2181: Comparison words produce canonical flags:
2182:
2183: @example
2184: 1 1 = .
2185: 1 0= .
2186: 0 1 < .
2187: 0 0 < .
2188: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2189: -1 1 < .
2190: @end example
2191:
1.66 anton 2192: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2193: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2194: these combinations are standard (for details see the standard,
2195: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48 anton 2196:
2197: You can use @code{and or xor invert} can be used as operations on
2198: canonical flags. Actually they are bitwise operations:
2199:
2200: @example
2201: 1 2 and .
2202: 1 2 or .
2203: 1 3 xor .
2204: 1 invert .
2205: @end example
2206:
2207: You can convert a zero/non-zero flag into a canonical flag with
2208: @code{0<>} (and complement it on the way with @code{0=}).
2209:
2210: @example
2211: 1 0= .
2212: 1 0<> .
2213: @end example
2214:
1.65 anton 2215: You can use the all-bits-set feature of canonical flags and the bitwise
1.48 anton 2216: operation of the Boolean operations to avoid @code{if}s:
2217:
2218: @example
2219: : foo ( n1 -- n2 )
2220: 0= if
2221: 14
2222: else
2223: 0
2224: endif ;
2225: 0 foo .
2226: 1 foo .
2227:
2228: : foo ( n1 -- n2 )
2229: 0= 14 and ;
2230: 0 foo .
2231: 1 foo .
2232: @end example
2233:
2234: @assignment
2235: Write @code{min} without @code{if}.
2236: @endassignment
2237:
1.66 anton 2238: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2239: @ref{Bitwise operations}.
2240:
1.48 anton 2241:
2242: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2243: @section General Loops
1.66 anton 2244: @cindex loops, indefinite, tutorial
1.48 anton 2245:
2246: The endless loop is the most simple one:
2247:
2248: @example
2249: : endless ( -- )
2250: 0 begin
2251: dup . 1+
2252: again ;
2253: endless
2254: @end example
2255:
2256: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2257: does nothing at run-time, @code{again} jumps back to @code{begin}.
2258:
2259: A loop with one exit at any place looks like this:
2260:
2261: @example
2262: : log2 ( +n1 -- n2 )
2263: \ logarithmus dualis of n1>0, rounded down to the next integer
2264: assert( dup 0> )
2265: 2/ 0 begin
2266: over 0> while
2267: 1+ swap 2/ swap
2268: repeat
2269: nip ;
2270: 7 log2 .
2271: 8 log2 .
2272: @end example
2273:
2274: At run-time @code{while} consumes a flag; if it is 0, execution
1.51 pazsan 2275: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48 anton 2276: continues behind the @code{while}. @code{Repeat} jumps back to
2277: @code{begin}, just like @code{again}.
2278:
2279: In Forth there are many combinations/abbreviations, like @code{1+}.
1.90 anton 2280: However, @code{2/} is not one of them; it shifts its argument right by
1.48 anton 2281: one bit (arithmetic shift right):
2282:
2283: @example
2284: -5 2 / .
2285: -5 2/ .
2286: @end example
2287:
2288: @code{assert(} is no standard word, but you can get it on systems other
2289: then Gforth by including @file{compat/assert.fs}. You can see what it
2290: does by trying
2291:
2292: @example
2293: 0 log2 .
2294: @end example
2295:
2296: Here's a loop with an exit at the end:
2297:
2298: @example
2299: : log2 ( +n1 -- n2 )
2300: \ logarithmus dualis of n1>0, rounded down to the next integer
2301: assert( dup 0 > )
2302: -1 begin
2303: 1+ swap 2/ swap
2304: over 0 <=
2305: until
2306: nip ;
2307: @end example
2308:
2309: @code{Until} consumes a flag; if it is non-zero, execution continues at
2310: the @code{begin}, otherwise after the @code{until}.
2311:
2312: @assignment
2313: Write a definition for computing the greatest common divisor.
2314: @endassignment
2315:
1.66 anton 2316: Reference: @ref{Simple Loops}.
2317:
1.48 anton 2318:
2319: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2320: @section Counted loops
1.66 anton 2321: @cindex loops, counted, tutorial
1.48 anton 2322:
2323: @example
2324: : ^ ( n1 u -- n )
2325: \ n = the uth power of u1
2326: 1 swap 0 u+do
2327: over *
2328: loop
2329: nip ;
2330: 3 2 ^ .
2331: 4 3 ^ .
2332: @end example
2333:
2334: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2335: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2336: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2337: times (or not at all, if @code{u3-u4<0}).
2338:
2339: You can see the stack effect design rules at work in the stack effect of
2340: the loop start words: Since the start value of the loop is more
2341: frequently constant than the end value, the start value is passed on
2342: the top-of-stack.
2343:
2344: You can access the counter of a counted loop with @code{i}:
2345:
2346: @example
2347: : fac ( u -- u! )
2348: 1 swap 1+ 1 u+do
2349: i *
2350: loop ;
2351: 5 fac .
2352: 7 fac .
2353: @end example
2354:
2355: There is also @code{+do}, which expects signed numbers (important for
2356: deciding whether to enter the loop).
2357:
2358: @assignment
2359: Write a definition for computing the nth Fibonacci number.
2360: @endassignment
2361:
1.65 anton 2362: You can also use increments other than 1:
2363:
2364: @example
2365: : up2 ( n1 n2 -- )
2366: +do
2367: i .
2368: 2 +loop ;
2369: 10 0 up2
2370:
2371: : down2 ( n1 n2 -- )
2372: -do
2373: i .
2374: 2 -loop ;
2375: 0 10 down2
2376: @end example
1.48 anton 2377:
1.66 anton 2378: Reference: @ref{Counted Loops}.
2379:
1.48 anton 2380:
2381: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2382: @section Recursion
1.66 anton 2383: @cindex recursion tutorial
1.48 anton 2384:
2385: Usually the name of a definition is not visible in the definition; but
2386: earlier definitions are usually visible:
2387:
2388: @example
2389: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2390: : / ( n1 n2 -- n )
2391: dup 0= if
2392: -10 throw \ report division by zero
2393: endif
2394: / \ old version
2395: ;
2396: 1 0 /
2397: @end example
2398:
2399: For recursive definitions you can use @code{recursive} (non-standard) or
2400: @code{recurse}:
2401:
2402: @example
2403: : fac1 ( n -- n! ) recursive
2404: dup 0> if
2405: dup 1- fac1 *
2406: else
2407: drop 1
2408: endif ;
2409: 7 fac1 .
2410:
2411: : fac2 ( n -- n! )
2412: dup 0> if
2413: dup 1- recurse *
2414: else
2415: drop 1
2416: endif ;
2417: 8 fac2 .
2418: @end example
2419:
2420: @assignment
2421: Write a recursive definition for computing the nth Fibonacci number.
2422: @endassignment
2423:
1.66 anton 2424: Reference (including indirect recursion): @xref{Calls and returns}.
2425:
1.48 anton 2426:
2427: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2428: @section Leaving definitions or loops
1.66 anton 2429: @cindex leaving definitions, tutorial
2430: @cindex leaving loops, tutorial
1.48 anton 2431:
2432: @code{EXIT} exits the current definition right away. For every counted
2433: loop that is left in this way, an @code{UNLOOP} has to be performed
2434: before the @code{EXIT}:
2435:
2436: @c !! real examples
2437: @example
2438: : ...
2439: ... u+do
2440: ... if
2441: ... unloop exit
2442: endif
2443: ...
2444: loop
2445: ... ;
2446: @end example
2447:
2448: @code{LEAVE} leaves the innermost counted loop right away:
2449:
2450: @example
2451: : ...
2452: ... u+do
2453: ... if
2454: ... leave
2455: endif
2456: ...
2457: loop
2458: ... ;
2459: @end example
2460:
1.65 anton 2461: @c !! example
1.48 anton 2462:
1.66 anton 2463: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2464:
2465:
1.48 anton 2466: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2467: @section Return Stack
1.66 anton 2468: @cindex return stack tutorial
1.48 anton 2469:
2470: In addition to the data stack Forth also has a second stack, the return
2471: stack; most Forth systems store the return addresses of procedure calls
2472: there (thus its name). Programmers can also use this stack:
2473:
2474: @example
2475: : foo ( n1 n2 -- )
2476: .s
2477: >r .s
1.50 anton 2478: r@@ .
1.48 anton 2479: >r .s
1.50 anton 2480: r@@ .
1.48 anton 2481: r> .
1.50 anton 2482: r@@ .
1.48 anton 2483: r> . ;
2484: 1 2 foo
2485: @end example
2486:
2487: @code{>r} takes an element from the data stack and pushes it onto the
2488: return stack; conversely, @code{r>} moves an elementm from the return to
2489: the data stack; @code{r@@} pushes a copy of the top of the return stack
2490: on the return stack.
2491:
2492: Forth programmers usually use the return stack for storing data
2493: temporarily, if using the data stack alone would be too complex, and
2494: factoring and locals are not an option:
2495:
2496: @example
2497: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2498: rot >r rot r> ;
2499: @end example
2500:
2501: The return address of the definition and the loop control parameters of
2502: counted loops usually reside on the return stack, so you have to take
2503: all items, that you have pushed on the return stack in a colon
2504: definition or counted loop, from the return stack before the definition
2505: or loop ends. You cannot access items that you pushed on the return
2506: stack outside some definition or loop within the definition of loop.
2507:
2508: If you miscount the return stack items, this usually ends in a crash:
2509:
2510: @example
2511: : crash ( n -- )
2512: >r ;
2513: 5 crash
2514: @end example
2515:
2516: You cannot mix using locals and using the return stack (according to the
2517: standard; Gforth has no problem). However, they solve the same
2518: problems, so this shouldn't be an issue.
2519:
2520: @assignment
2521: Can you rewrite any of the definitions you wrote until now in a better
2522: way using the return stack?
2523: @endassignment
2524:
1.66 anton 2525: Reference: @ref{Return stack}.
2526:
1.48 anton 2527:
2528: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2529: @section Memory
1.66 anton 2530: @cindex memory access/allocation tutorial
1.48 anton 2531:
2532: You can create a global variable @code{v} with
2533:
2534: @example
2535: variable v ( -- addr )
2536: @end example
2537:
2538: @code{v} pushes the address of a cell in memory on the stack. This cell
2539: was reserved by @code{variable}. You can use @code{!} (store) to store
2540: values into this cell and @code{@@} (fetch) to load the value from the
2541: stack into memory:
2542:
2543: @example
2544: v .
2545: 5 v ! .s
1.50 anton 2546: v @@ .
1.48 anton 2547: @end example
2548:
1.65 anton 2549: You can see a raw dump of memory with @code{dump}:
2550:
2551: @example
2552: v 1 cells .s dump
2553: @end example
2554:
2555: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2556: generally, address units (aus)) that @code{n1 cells} occupy. You can
2557: also reserve more memory:
1.48 anton 2558:
2559: @example
2560: create v2 20 cells allot
1.65 anton 2561: v2 20 cells dump
1.48 anton 2562: @end example
2563:
1.65 anton 2564: creates a word @code{v2} and reserves 20 uninitialized cells; the
2565: address pushed by @code{v2} points to the start of these 20 cells. You
2566: can use address arithmetic to access these cells:
1.48 anton 2567:
2568: @example
2569: 3 v2 5 cells + !
1.65 anton 2570: v2 20 cells dump
1.48 anton 2571: @end example
2572:
2573: You can reserve and initialize memory with @code{,}:
2574:
2575: @example
2576: create v3
2577: 5 , 4 , 3 , 2 , 1 ,
1.50 anton 2578: v3 @@ .
2579: v3 cell+ @@ .
2580: v3 2 cells + @@ .
1.65 anton 2581: v3 5 cells dump
1.48 anton 2582: @end example
2583:
2584: @assignment
2585: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2586: @code{u} cells, with the first of these cells at @code{addr}, the next
2587: one at @code{addr cell+} etc.
2588: @endassignment
2589:
2590: You can also reserve memory without creating a new word:
2591:
2592: @example
1.60 anton 2593: here 10 cells allot .
2594: here .
1.48 anton 2595: @end example
2596:
2597: @code{Here} pushes the start address of the memory area. You should
2598: store it somewhere, or you will have a hard time finding the memory area
2599: again.
2600:
2601: @code{Allot} manages dictionary memory. The dictionary memory contains
2602: the system's data structures for words etc. on Gforth and most other
2603: Forth systems. It is managed like a stack: You can free the memory that
2604: you have just @code{allot}ed with
2605:
2606: @example
2607: -10 cells allot
1.60 anton 2608: here .
1.48 anton 2609: @end example
2610:
2611: Note that you cannot do this if you have created a new word in the
2612: meantime (because then your @code{allot}ed memory is no longer on the
2613: top of the dictionary ``stack'').
2614:
2615: Alternatively, you can use @code{allocate} and @code{free} which allow
2616: freeing memory in any order:
2617:
2618: @example
2619: 10 cells allocate throw .s
2620: 20 cells allocate throw .s
2621: swap
2622: free throw
2623: free throw
2624: @end example
2625:
2626: The @code{throw}s deal with errors (e.g., out of memory).
2627:
1.65 anton 2628: And there is also a
2629: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2630: garbage collector}, which eliminates the need to @code{free} memory
2631: explicitly.
1.48 anton 2632:
1.66 anton 2633: Reference: @ref{Memory}.
2634:
1.48 anton 2635:
2636: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2637: @section Characters and Strings
1.66 anton 2638: @cindex strings tutorial
2639: @cindex characters tutorial
1.48 anton 2640:
2641: On the stack characters take up a cell, like numbers. In memory they
2642: have their own size (one 8-bit byte on most systems), and therefore
2643: require their own words for memory access:
2644:
2645: @example
2646: create v4
2647: 104 c, 97 c, 108 c, 108 c, 111 c,
1.50 anton 2648: v4 4 chars + c@@ .
1.65 anton 2649: v4 5 chars dump
1.48 anton 2650: @end example
2651:
2652: The preferred representation of strings on the stack is @code{addr
2653: u-count}, where @code{addr} is the address of the first character and
2654: @code{u-count} is the number of characters in the string.
2655:
2656: @example
2657: v4 5 type
2658: @end example
2659:
2660: You get a string constant with
2661:
2662: @example
2663: s" hello, world" .s
2664: type
2665: @end example
2666:
2667: Make sure you have a space between @code{s"} and the string; @code{s"}
2668: is a normal Forth word and must be delimited with white space (try what
2669: happens when you remove the space).
2670:
2671: However, this interpretive use of @code{s"} is quite restricted: the
2672: string exists only until the next call of @code{s"} (some Forth systems
2673: keep more than one of these strings, but usually they still have a
1.62 crook 2674: limited lifetime).
1.48 anton 2675:
2676: @example
2677: s" hello," s" world" .s
2678: type
2679: type
2680: @end example
2681:
1.62 crook 2682: You can also use @code{s"} in a definition, and the resulting
2683: strings then live forever (well, for as long as the definition):
1.48 anton 2684:
2685: @example
2686: : foo s" hello," s" world" ;
2687: foo .s
2688: type
2689: type
2690: @end example
2691:
2692: @assignment
2693: @code{Emit ( c -- )} types @code{c} as character (not a number).
2694: Implement @code{type ( addr u -- )}.
2695: @endassignment
2696:
1.66 anton 2697: Reference: @ref{Memory Blocks}.
2698:
2699:
1.84 pazsan 2700: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
1.48 anton 2701: @section Alignment
1.66 anton 2702: @cindex alignment tutorial
2703: @cindex memory alignment tutorial
1.48 anton 2704:
2705: On many processors cells have to be aligned in memory, if you want to
2706: access them with @code{@@} and @code{!} (and even if the processor does
1.62 crook 2707: not require alignment, access to aligned cells is faster).
1.48 anton 2708:
2709: @code{Create} aligns @code{here} (i.e., the place where the next
2710: allocation will occur, and that the @code{create}d word points to).
2711: Likewise, the memory produced by @code{allocate} starts at an aligned
2712: address. Adding a number of @code{cells} to an aligned address produces
2713: another aligned address.
2714:
2715: However, address arithmetic involving @code{char+} and @code{chars} can
2716: create an address that is not cell-aligned. @code{Aligned ( addr --
2717: a-addr )} produces the next aligned address:
2718:
2719: @example
1.50 anton 2720: v3 char+ aligned .s @@ .
2721: v3 char+ .s @@ .
1.48 anton 2722: @end example
2723:
2724: Similarly, @code{align} advances @code{here} to the next aligned
2725: address:
2726:
2727: @example
2728: create v5 97 c,
2729: here .
2730: align here .
2731: 1000 ,
2732: @end example
2733:
2734: Note that you should use aligned addresses even if your processor does
2735: not require them, if you want your program to be portable.
2736:
1.66 anton 2737: Reference: @ref{Address arithmetic}.
2738:
1.48 anton 2739:
1.84 pazsan 2740: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
2741: @section Files
2742: @cindex files tutorial
2743:
2744: This section gives a short introduction into how to use files inside
2745: Forth. It's broken up into five easy steps:
2746:
2747: @enumerate 1
2748: @item Opened an ASCII text file for input
2749: @item Opened a file for output
2750: @item Read input file until string matched (or some other condition matched)
2751: @item Wrote some lines from input ( modified or not) to output
2752: @item Closed the files.
2753: @end enumerate
2754:
2755: @subsection Open file for input
2756:
2757: @example
2758: s" foo.in" r/o open-file throw Value fd-in
2759: @end example
2760:
2761: @subsection Create file for output
2762:
2763: @example
2764: s" foo.out" w/o create-file throw Value fd-out
2765: @end example
2766:
2767: The available file modes are r/o for read-only access, r/w for
2768: read-write access, and w/o for write-only access. You could open both
2769: files with r/w, too, if you like. All file words return error codes; for
2770: most applications, it's best to pass there error codes with @code{throw}
2771: to the outer error handler.
2772:
2773: If you want words for opening and assigning, define them as follows:
2774:
2775: @example
2776: 0 Value fd-in
2777: 0 Value fd-out
2778: : open-input ( addr u -- ) r/o open-file throw to fd-in ;
2779: : open-output ( addr u -- ) w/o create-file throw to fd-out ;
2780: @end example
2781:
2782: Usage example:
2783:
2784: @example
2785: s" foo.in" open-input
2786: s" foo.out" open-output
2787: @end example
2788:
2789: @subsection Scan file for a particular line
2790:
2791: @example
2792: 256 Constant max-line
2793: Create line-buffer max-line 2 + allot
2794:
2795: : scan-file ( addr u -- )
2796: begin
2797: line-buffer max-line fd-in read-line throw
2798: while
2799: >r 2dup line-buffer r> compare 0=
2800: until
2801: else
2802: drop
2803: then
2804: 2drop ;
2805: @end example
2806:
2807: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
2808: the buffer at addr, and returns the number of bytes read, a flag that's
2809: true when the end of file is reached, and an error code.
2810:
2811: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
2812: returns zero if both strings are equal. It returns a positive number if
2813: the first string is lexically greater, a negative if the second string
2814: is lexically greater.
2815:
2816: We haven't seen this loop here; it has two exits. Since the @code{while}
2817: exits with the number of bytes read on the stack, we have to clean up
2818: that separately; that's after the @code{else}.
2819:
2820: Usage example:
2821:
2822: @example
2823: s" The text I search is here" scan-file
2824: @end example
2825:
2826: @subsection Copy input to output
2827:
2828: @example
2829: : copy-file ( -- )
2830: begin
2831: line-buffer max-line fd-in read-line throw
2832: while
2833: line-buffer swap fd-out write-file throw
2834: repeat ;
2835: @end example
2836:
2837: @subsection Close files
2838:
2839: @example
2840: fd-in close-file throw
2841: fd-out close-file throw
2842: @end example
2843:
2844: Likewise, you can put that into definitions, too:
2845:
2846: @example
2847: : close-input ( -- ) fd-in close-file throw ;
2848: : close-output ( -- ) fd-out close-file throw ;
2849: @end example
2850:
2851: @assignment
2852: How could you modify @code{copy-file} so that it copies until a second line is
2853: matched? Can you write a program that extracts a section of a text file,
2854: given the line that starts and the line that terminates that section?
2855: @endassignment
2856:
2857: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
1.48 anton 2858: @section Interpretation and Compilation Semantics and Immediacy
1.66 anton 2859: @cindex semantics tutorial
2860: @cindex interpretation semantics tutorial
2861: @cindex compilation semantics tutorial
2862: @cindex immediate, tutorial
1.48 anton 2863:
2864: When a word is compiled, it behaves differently from being interpreted.
2865: E.g., consider @code{+}:
2866:
2867: @example
2868: 1 2 + .
2869: : foo + ;
2870: @end example
2871:
2872: These two behaviours are known as compilation and interpretation
2873: semantics. For normal words (e.g., @code{+}), the compilation semantics
2874: is to append the interpretation semantics to the currently defined word
2875: (@code{foo} in the example above). I.e., when @code{foo} is executed
2876: later, the interpretation semantics of @code{+} (i.e., adding two
2877: numbers) will be performed.
2878:
2879: However, there are words with non-default compilation semantics, e.g.,
2880: the control-flow words like @code{if}. You can use @code{immediate} to
2881: change the compilation semantics of the last defined word to be equal to
2882: the interpretation semantics:
2883:
2884: @example
2885: : [FOO] ( -- )
2886: 5 . ; immediate
2887:
2888: [FOO]
2889: : bar ( -- )
2890: [FOO] ;
2891: bar
2892: see bar
2893: @end example
2894:
2895: Two conventions to mark words with non-default compilation semnatics are
2896: names with brackets (more frequently used) and to write them all in
2897: upper case (less frequently used).
2898:
2899: In Gforth (and many other systems) you can also remove the
2900: interpretation semantics with @code{compile-only} (the compilation
2901: semantics is derived from the original interpretation semantics):
2902:
2903: @example
2904: : flip ( -- )
2905: 6 . ; compile-only \ but not immediate
2906: flip
2907:
2908: : flop ( -- )
2909: flip ;
2910: flop
2911: @end example
2912:
2913: In this example the interpretation semantics of @code{flop} is equal to
2914: the original interpretation semantics of @code{flip}.
2915:
2916: The text interpreter has two states: in interpret state, it performs the
2917: interpretation semantics of words it encounters; in compile state, it
2918: performs the compilation semantics of these words.
2919:
2920: Among other things, @code{:} switches into compile state, and @code{;}
2921: switches back to interpret state. They contain the factors @code{]}
2922: (switch to compile state) and @code{[} (switch to interpret state), that
2923: do nothing but switch the state.
2924:
2925: @example
2926: : xxx ( -- )
2927: [ 5 . ]
2928: ;
2929:
2930: xxx
2931: see xxx
2932: @end example
2933:
2934: These brackets are also the source of the naming convention mentioned
2935: above.
2936:
1.66 anton 2937: Reference: @ref{Interpretation and Compilation Semantics}.
2938:
1.48 anton 2939:
2940: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2941: @section Execution Tokens
1.66 anton 2942: @cindex execution tokens tutorial
2943: @cindex XT tutorial
1.48 anton 2944:
2945: @code{' word} gives you the execution token (XT) of a word. The XT is a
2946: cell representing the interpretation semantics of a word. You can
2947: execute this semantics with @code{execute}:
2948:
2949: @example
2950: ' + .s
2951: 1 2 rot execute .
2952: @end example
2953:
2954: The XT is similar to a function pointer in C. However, parameter
2955: passing through the stack makes it a little more flexible:
2956:
2957: @example
2958: : map-array ( ... addr u xt -- ... )
1.50 anton 2959: \ executes xt ( ... x -- ... ) for every element of the array starting
2960: \ at addr and containing u elements
1.48 anton 2961: @{ xt @}
2962: cells over + swap ?do
1.50 anton 2963: i @@ xt execute
1.48 anton 2964: 1 cells +loop ;
2965:
2966: create a 3 , 4 , 2 , -1 , 4 ,
2967: a 5 ' . map-array .s
2968: 0 a 5 ' + map-array .
2969: s" max-n" environment? drop .s
2970: a 5 ' min map-array .
2971: @end example
2972:
2973: You can use map-array with the XTs of words that consume one element
2974: more than they produce. In theory you can also use it with other XTs,
2975: but the stack effect then depends on the size of the array, which is
2976: hard to understand.
2977:
1.51 pazsan 2978: Since XTs are cell-sized, you can store them in memory and manipulate
2979: them on the stack like other cells. You can also compile the XT into a
1.48 anton 2980: word with @code{compile,}:
2981:
2982: @example
2983: : foo1 ( n1 n2 -- n )
2984: [ ' + compile, ] ;
2985: see foo
2986: @end example
2987:
2988: This is non-standard, because @code{compile,} has no compilation
2989: semantics in the standard, but it works in good Forth systems. For the
2990: broken ones, use
2991:
2992: @example
2993: : [compile,] compile, ; immediate
2994:
2995: : foo1 ( n1 n2 -- n )
2996: [ ' + ] [compile,] ;
2997: see foo
2998: @end example
2999:
3000: @code{'} is a word with default compilation semantics; it parses the
3001: next word when its interpretation semantics are executed, not during
3002: compilation:
3003:
3004: @example
3005: : foo ( -- xt )
3006: ' ;
3007: see foo
3008: : bar ( ... "word" -- ... )
3009: ' execute ;
3010: see bar
1.60 anton 3011: 1 2 bar + .
1.48 anton 3012: @end example
3013:
3014: You often want to parse a word during compilation and compile its XT so
3015: it will be pushed on the stack at run-time. @code{[']} does this:
3016:
3017: @example
3018: : xt-+ ( -- xt )
3019: ['] + ;
3020: see xt-+
3021: 1 2 xt-+ execute .
3022: @end example
3023:
3024: Many programmers tend to see @code{'} and the word it parses as one
3025: unit, and expect it to behave like @code{[']} when compiled, and are
3026: confused by the actual behaviour. If you are, just remember that the
3027: Forth system just takes @code{'} as one unit and has no idea that it is
3028: a parsing word (attempts to convenience programmers in this issue have
3029: usually resulted in even worse pitfalls, see
1.66 anton 3030: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
3031: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48 anton 3032:
3033: Note that the state of the interpreter does not come into play when
1.51 pazsan 3034: creating and executing XTs. I.e., even when you execute @code{'} in
1.48 anton 3035: compile state, it still gives you the interpretation semantics. And
3036: whatever that state is, @code{execute} performs the semantics
1.66 anton 3037: represented by the XT (i.e., for XTs produced with @code{'} the
3038: interpretation semantics).
3039:
3040: Reference: @ref{Tokens for Words}.
1.48 anton 3041:
3042:
3043: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
3044: @section Exceptions
1.66 anton 3045: @cindex exceptions tutorial
1.48 anton 3046:
3047: @code{throw ( n -- )} causes an exception unless n is zero.
3048:
3049: @example
3050: 100 throw .s
3051: 0 throw .s
3052: @end example
3053:
3054: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
3055: it catches exceptions and pushes the number of the exception on the
3056: stack (or 0, if the xt executed without exception). If there was an
3057: exception, the stacks have the same depth as when entering @code{catch}:
3058:
3059: @example
3060: .s
3061: 3 0 ' / catch .s
3062: 3 2 ' / catch .s
3063: @end example
3064:
3065: @assignment
3066: Try the same with @code{execute} instead of @code{catch}.
3067: @endassignment
3068:
3069: @code{Throw} always jumps to the dynamically next enclosing
3070: @code{catch}, even if it has to leave several call levels to achieve
3071: this:
3072:
3073: @example
3074: : foo 100 throw ;
3075: : foo1 foo ." after foo" ;
1.51 pazsan 3076: : bar ['] foo1 catch ;
1.60 anton 3077: bar .
1.48 anton 3078: @end example
3079:
3080: It is often important to restore a value upon leaving a definition, even
3081: if the definition is left through an exception. You can ensure this
3082: like this:
3083:
3084: @example
3085: : ...
3086: save-x
1.51 pazsan 3087: ['] word-changing-x catch ( ... n )
1.48 anton 3088: restore-x
3089: ( ... n ) throw ;
3090: @end example
3091:
1.55 anton 3092: Gforth provides an alternative syntax in addition to @code{catch}:
1.48 anton 3093: @code{try ... recover ... endtry}. If the code between @code{try} and
3094: @code{recover} has an exception, the stack depths are restored, the
3095: exception number is pushed on the stack, and the code between
3096: @code{recover} and @code{endtry} is performed. E.g., the definition for
3097: @code{catch} is
3098:
3099: @example
3100: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
3101: try
3102: execute 0
3103: recover
3104: nip
3105: endtry ;
3106: @end example
3107:
3108: The equivalent to the restoration code above is
3109:
3110: @example
3111: : ...
3112: save-x
3113: try
3114: word-changing-x
3115: end-try
3116: restore-x
3117: throw ;
3118: @end example
3119:
3120: As you can see, the @code{recover} part is optional.
3121:
1.66 anton 3122: Reference: @ref{Exception Handling}.
3123:
1.48 anton 3124:
3125: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
3126: @section Defining Words
1.66 anton 3127: @cindex defining words tutorial
3128: @cindex does> tutorial
3129: @cindex create...does> tutorial
3130:
3131: @c before semantics?
1.48 anton 3132:
3133: @code{:}, @code{create}, and @code{variable} are definition words: They
3134: define other words. @code{Constant} is another definition word:
3135:
3136: @example
3137: 5 constant foo
3138: foo .
3139: @end example
3140:
3141: You can also use the prefixes @code{2} (double-cell) and @code{f}
3142: (floating point) with @code{variable} and @code{constant}.
3143:
3144: You can also define your own defining words. E.g.:
3145:
3146: @example
3147: : variable ( "name" -- )
3148: create 0 , ;
3149: @end example
3150:
3151: You can also define defining words that create words that do something
3152: other than just producing their address:
3153:
3154: @example
3155: : constant ( n "name" -- )
3156: create ,
3157: does> ( -- n )
1.50 anton 3158: ( addr ) @@ ;
1.48 anton 3159:
3160: 5 constant foo
3161: foo .
3162: @end example
3163:
3164: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3165: @code{does>} replaces @code{;}, but it also does something else: It
3166: changes the last defined word such that it pushes the address of the
3167: body of the word and then performs the code after the @code{does>}
3168: whenever it is called.
3169:
3170: In the example above, @code{constant} uses @code{,} to store 5 into the
3171: body of @code{foo}. When @code{foo} executes, it pushes the address of
3172: the body onto the stack, then (in the code after the @code{does>})
3173: fetches the 5 from there.
3174:
3175: The stack comment near the @code{does>} reflects the stack effect of the
3176: defined word, not the stack effect of the code after the @code{does>}
3177: (the difference is that the code expects the address of the body that
3178: the stack comment does not show).
3179:
3180: You can use these definition words to do factoring in cases that involve
3181: (other) definition words. E.g., a field offset is always added to an
3182: address. Instead of defining
3183:
3184: @example
3185: 2 cells constant offset-field1
3186: @end example
3187:
3188: and using this like
3189:
3190: @example
3191: ( addr ) offset-field1 +
3192: @end example
3193:
3194: you can define a definition word
3195:
3196: @example
3197: : simple-field ( n "name" -- )
3198: create ,
3199: does> ( n1 -- n1+n )
1.50 anton 3200: ( addr ) @@ + ;
1.48 anton 3201: @end example
1.21 crook 3202:
1.48 anton 3203: Definition and use of field offsets now look like this:
1.21 crook 3204:
1.48 anton 3205: @example
3206: 2 cells simple-field field1
1.60 anton 3207: create mystruct 4 cells allot
3208: mystruct .s field1 .s drop
1.48 anton 3209: @end example
1.21 crook 3210:
1.48 anton 3211: If you want to do something with the word without performing the code
3212: after the @code{does>}, you can access the body of a @code{create}d word
3213: with @code{>body ( xt -- addr )}:
1.21 crook 3214:
1.48 anton 3215: @example
3216: : value ( n "name" -- )
3217: create ,
3218: does> ( -- n1 )
1.50 anton 3219: @@ ;
1.48 anton 3220: : to ( n "name" -- )
3221: ' >body ! ;
1.21 crook 3222:
1.48 anton 3223: 5 value foo
3224: foo .
3225: 7 to foo
3226: foo .
3227: @end example
1.21 crook 3228:
1.48 anton 3229: @assignment
3230: Define @code{defer ( "name" -- )}, which creates a word that stores an
3231: XT (at the start the XT of @code{abort}), and upon execution
3232: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3233: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3234: recursion is one application of @code{defer}.
3235: @endassignment
1.29 crook 3236:
1.66 anton 3237: Reference: @ref{User-defined Defining Words}.
3238:
3239:
1.48 anton 3240: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3241: @section Arrays and Records
1.66 anton 3242: @cindex arrays tutorial
3243: @cindex records tutorial
3244: @cindex structs tutorial
1.29 crook 3245:
1.48 anton 3246: Forth has no standard words for defining data structures such as arrays
3247: and records (structs in C terminology), but you can build them yourself
3248: based on address arithmetic. You can also define words for defining
3249: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29 crook 3250:
1.48 anton 3251: One of the first projects a Forth newcomer sets out upon when learning
3252: about defining words is an array defining word (possibly for
3253: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3254: learn something from it. However, don't be disappointed when you later
3255: learn that you have little use for these words (inappropriate use would
3256: be even worse). I have not yet found a set of useful array words yet;
3257: the needs are just too diverse, and named, global arrays (the result of
3258: naive use of defining words) are often not flexible enough (e.g.,
1.66 anton 3259: consider how to pass them as parameters). Another such project is a set
3260: of words to help dealing with strings.
1.29 crook 3261:
1.48 anton 3262: On the other hand, there is a useful set of record words, and it has
3263: been defined in @file{compat/struct.fs}; these words are predefined in
3264: Gforth. They are explained in depth elsewhere in this manual (see
3265: @pxref{Structures}). The @code{simple-field} example above is
3266: simplified variant of fields in this package.
1.21 crook 3267:
3268:
1.48 anton 3269: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3270: @section @code{POSTPONE}
1.66 anton 3271: @cindex postpone tutorial
1.21 crook 3272:
1.48 anton 3273: You can compile the compilation semantics (instead of compiling the
3274: interpretation semantics) of a word with @code{POSTPONE}:
1.21 crook 3275:
1.48 anton 3276: @example
3277: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51 pazsan 3278: POSTPONE + ; immediate
1.48 anton 3279: : foo ( n1 n2 -- n )
3280: MY-+ ;
3281: 1 2 foo .
3282: see foo
3283: @end example
1.21 crook 3284:
1.48 anton 3285: During the definition of @code{foo} the text interpreter performs the
3286: compilation semantics of @code{MY-+}, which performs the compilation
3287: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3288:
3289: This example also displays separate stack comments for the compilation
3290: semantics and for the stack effect of the compiled code. For words with
3291: default compilation semantics these stack effects are usually not
3292: displayed; the stack effect of the compilation semantics is always
3293: @code{( -- )} for these words, the stack effect for the compiled code is
3294: the stack effect of the interpretation semantics.
3295:
3296: Note that the state of the interpreter does not come into play when
3297: performing the compilation semantics in this way. You can also perform
3298: it interpretively, e.g.:
3299:
3300: @example
3301: : foo2 ( n1 n2 -- n )
3302: [ MY-+ ] ;
3303: 1 2 foo .
3304: see foo
3305: @end example
1.21 crook 3306:
1.48 anton 3307: However, there are some broken Forth systems where this does not always
1.62 crook 3308: work, and therefore this practice was been declared non-standard in
1.48 anton 3309: 1999.
3310: @c !! repair.fs
3311:
3312: Here is another example for using @code{POSTPONE}:
1.44 crook 3313:
1.48 anton 3314: @example
3315: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3316: POSTPONE negate POSTPONE + ; immediate compile-only
3317: : bar ( n1 n2 -- n )
3318: MY-- ;
3319: 2 1 bar .
3320: see bar
3321: @end example
1.21 crook 3322:
1.48 anton 3323: You can define @code{ENDIF} in this way:
1.21 crook 3324:
1.48 anton 3325: @example
3326: : ENDIF ( Compilation: orig -- )
3327: POSTPONE then ; immediate
3328: @end example
1.21 crook 3329:
1.48 anton 3330: @assignment
3331: Write @code{MY-2DUP} that has compilation semantics equivalent to
3332: @code{2dup}, but compiles @code{over over}.
3333: @endassignment
1.29 crook 3334:
1.66 anton 3335: @c !! @xref{Macros} for reference
3336:
3337:
1.48 anton 3338: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3339: @section @code{Literal}
1.66 anton 3340: @cindex literal tutorial
1.29 crook 3341:
1.48 anton 3342: You cannot @code{POSTPONE} numbers:
1.21 crook 3343:
1.48 anton 3344: @example
3345: : [FOO] POSTPONE 500 ; immediate
1.21 crook 3346: @end example
3347:
1.48 anton 3348: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29 crook 3349:
1.48 anton 3350: @example
3351: : [FOO] ( compilation: --; run-time: -- n )
3352: 500 POSTPONE literal ; immediate
1.29 crook 3353:
1.60 anton 3354: : flip [FOO] ;
1.48 anton 3355: flip .
3356: see flip
3357: @end example
1.29 crook 3358:
1.48 anton 3359: @code{LITERAL} consumes a number at compile-time (when it's compilation
3360: semantics are executed) and pushes it at run-time (when the code it
3361: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3362: number computed at compile time into the current word:
1.29 crook 3363:
1.48 anton 3364: @example
3365: : bar ( -- n )
3366: [ 2 2 + ] literal ;
3367: see bar
3368: @end example
1.29 crook 3369:
1.48 anton 3370: @assignment
3371: Write @code{]L} which allows writing the example above as @code{: bar (
3372: -- n ) [ 2 2 + ]L ;}
3373: @endassignment
3374:
1.66 anton 3375: @c !! @xref{Macros} for reference
3376:
1.48 anton 3377:
3378: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3379: @section Advanced macros
1.66 anton 3380: @cindex macros, advanced tutorial
3381: @cindex run-time code generation, tutorial
1.48 anton 3382:
1.66 anton 3383: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3384: Execution Tokens}. It frequently performs @code{execute}, a relatively
3385: expensive operation in some Forth implementations. You can use
1.48 anton 3386: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3387: and produce a word that contains the word to be performed directly:
3388:
3389: @c use ]] ... [[
3390: @example
3391: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3392: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3393: \ array beginning at addr and containing u elements
3394: @{ xt @}
3395: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50 anton 3396: POSTPONE i POSTPONE @@ xt compile,
1.48 anton 3397: 1 cells POSTPONE literal POSTPONE +loop ;
3398:
3399: : sum-array ( addr u -- n )
3400: 0 rot rot [ ' + compile-map-array ] ;
3401: see sum-array
3402: a 5 sum-array .
3403: @end example
3404:
3405: You can use the full power of Forth for generating the code; here's an
3406: example where the code is generated in a loop:
3407:
3408: @example
3409: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3410: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50 anton 3411: POSTPONE tuck POSTPONE @@
1.48 anton 3412: POSTPONE literal POSTPONE * POSTPONE +
3413: POSTPONE swap POSTPONE cell+ ;
3414:
3415: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51 pazsan 3416: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48 anton 3417: 0 postpone literal postpone swap
3418: [ ' compile-vmul-step compile-map-array ]
3419: postpone drop ;
3420: see compile-vmul
3421:
3422: : a-vmul ( addr -- n )
1.51 pazsan 3423: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48 anton 3424: [ a 5 compile-vmul ] ;
3425: see a-vmul
3426: a a-vmul .
3427: @end example
3428:
3429: This example uses @code{compile-map-array} to show off, but you could
1.66 anton 3430: also use @code{map-array} instead (try it now!).
1.48 anton 3431:
3432: You can use this technique for efficient multiplication of large
3433: matrices. In matrix multiplication, you multiply every line of one
3434: matrix with every column of the other matrix. You can generate the code
3435: for one line once, and use it for every column. The only downside of
3436: this technique is that it is cumbersome to recover the memory consumed
3437: by the generated code when you are done (and in more complicated cases
3438: it is not possible portably).
3439:
1.66 anton 3440: @c !! @xref{Macros} for reference
3441:
3442:
1.48 anton 3443: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3444: @section Compilation Tokens
1.66 anton 3445: @cindex compilation tokens, tutorial
3446: @cindex CT, tutorial
1.48 anton 3447:
3448: This section is Gforth-specific. You can skip it.
3449:
3450: @code{' word compile,} compiles the interpretation semantics. For words
3451: with default compilation semantics this is the same as performing the
3452: compilation semantics. To represent the compilation semantics of other
3453: words (e.g., words like @code{if} that have no interpretation
3454: semantics), Gforth has the concept of a compilation token (CT,
3455: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3456: You can perform the compilation semantics represented by a CT with
3457: @code{execute}:
1.29 crook 3458:
1.48 anton 3459: @example
3460: : foo2 ( n1 n2 -- n )
3461: [ comp' + execute ] ;
3462: see foo
3463: @end example
1.29 crook 3464:
1.48 anton 3465: You can compile the compilation semantics represented by a CT with
3466: @code{postpone,}:
1.30 anton 3467:
1.48 anton 3468: @example
3469: : foo3 ( -- )
3470: [ comp' + postpone, ] ;
3471: see foo3
3472: @end example
1.30 anton 3473:
1.51 pazsan 3474: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48 anton 3475: @code{comp'} is particularly useful for words that have no
3476: interpretation semantics:
1.29 crook 3477:
1.30 anton 3478: @example
1.48 anton 3479: ' if
1.60 anton 3480: comp' if .s 2drop
1.30 anton 3481: @end example
3482:
1.66 anton 3483: Reference: @ref{Tokens for Words}.
3484:
1.29 crook 3485:
1.48 anton 3486: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3487: @section Wordlists and Search Order
1.66 anton 3488: @cindex wordlists tutorial
3489: @cindex search order, tutorial
1.48 anton 3490:
3491: The dictionary is not just a memory area that allows you to allocate
3492: memory with @code{allot}, it also contains the Forth words, arranged in
3493: several wordlists. When searching for a word in a wordlist,
3494: conceptually you start searching at the youngest and proceed towards
3495: older words (in reality most systems nowadays use hash-tables); i.e., if
3496: you define a word with the same name as an older word, the new word
3497: shadows the older word.
3498:
3499: Which wordlists are searched in which order is determined by the search
3500: order. You can display the search order with @code{order}. It displays
3501: first the search order, starting with the wordlist searched first, then
3502: it displays the wordlist that will contain newly defined words.
1.21 crook 3503:
1.48 anton 3504: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21 crook 3505:
1.48 anton 3506: @example
3507: wordlist constant mywords
3508: @end example
1.21 crook 3509:
1.48 anton 3510: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3511: defined words (the @emph{current} wordlist):
1.21 crook 3512:
1.48 anton 3513: @example
3514: mywords set-current
3515: order
3516: @end example
1.26 crook 3517:
1.48 anton 3518: Gforth does not display a name for the wordlist in @code{mywords}
3519: because this wordlist was created anonymously with @code{wordlist}.
1.21 crook 3520:
1.48 anton 3521: You can get the current wordlist with @code{get-current ( -- wid)}. If
3522: you want to put something into a specific wordlist without overall
3523: effect on the current wordlist, this typically looks like this:
1.21 crook 3524:
1.48 anton 3525: @example
3526: get-current mywords set-current ( wid )
3527: create someword
3528: ( wid ) set-current
3529: @end example
1.21 crook 3530:
1.48 anton 3531: You can write the search order with @code{set-order ( wid1 .. widn n --
3532: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3533: searched wordlist is topmost.
1.21 crook 3534:
1.48 anton 3535: @example
3536: get-order mywords swap 1+ set-order
3537: order
3538: @end example
1.21 crook 3539:
1.48 anton 3540: Yes, the order of wordlists in the output of @code{order} is reversed
3541: from stack comments and the output of @code{.s} and thus unintuitive.
1.21 crook 3542:
1.48 anton 3543: @assignment
3544: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3545: wordlist to the search order. Define @code{previous ( -- )}, which
3546: removes the first searched wordlist from the search order. Experiment
3547: with boundary conditions (you will see some crashes or situations that
3548: are hard or impossible to leave).
3549: @endassignment
1.21 crook 3550:
1.48 anton 3551: The search order is a powerful foundation for providing features similar
3552: to Modula-2 modules and C++ namespaces. However, trying to modularize
3553: programs in this way has disadvantages for debugging and reuse/factoring
3554: that overcome the advantages in my experience (I don't do huge projects,
1.55 anton 3555: though). These disadvantages are not so clear in other
1.82 anton 3556: languages/programming environments, because these languages are not so
1.48 anton 3557: strong in debugging and reuse.
1.21 crook 3558:
1.66 anton 3559: @c !! example
3560:
3561: Reference: @ref{Word Lists}.
1.21 crook 3562:
1.29 crook 3563: @c ******************************************************************
1.48 anton 3564: @node Introduction, Words, Tutorial, Top
1.29 crook 3565: @comment node-name, next, previous, up
3566: @chapter An Introduction to ANS Forth
3567: @cindex Forth - an introduction
1.21 crook 3568:
1.83 anton 3569: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
3570: that it is slower-paced in its examples, but uses them to dive deep into
3571: explaining Forth internals (not covered by the Tutorial). Apart from
3572: that, this chapter covers far less material. It is suitable for reading
3573: without using a computer.
3574:
1.29 crook 3575: The primary purpose of this manual is to document Gforth. However, since
3576: Forth is not a widely-known language and there is a lack of up-to-date
3577: teaching material, it seems worthwhile to provide some introductory
1.49 anton 3578: material. For other sources of Forth-related
3579: information, see @ref{Forth-related information}.
1.21 crook 3580:
1.29 crook 3581: The examples in this section should work on any ANS Forth; the
3582: output shown was produced using Gforth. Each example attempts to
3583: reproduce the exact output that Gforth produces. If you try out the
3584: examples (and you should), what you should type is shown @kbd{like this}
3585: and Gforth's response is shown @code{like this}. The single exception is
1.30 anton 3586: that, where the example shows @key{RET} it means that you should
1.29 crook 3587: press the ``carriage return'' key. Unfortunately, some output formats for
3588: this manual cannot show the difference between @kbd{this} and
3589: @code{this} which will make trying out the examples harder (but not
3590: impossible).
1.21 crook 3591:
1.29 crook 3592: Forth is an unusual language. It provides an interactive development
3593: environment which includes both an interpreter and compiler. Forth
3594: programming style encourages you to break a problem down into many
3595: @cindex factoring
3596: small fragments (@dfn{factoring}), and then to develop and test each
3597: fragment interactively. Forth advocates assert that breaking the
3598: edit-compile-test cycle used by conventional programming languages can
3599: lead to great productivity improvements.
1.21 crook 3600:
1.29 crook 3601: @menu
1.67 anton 3602: * Introducing the Text Interpreter::
3603: * Stacks and Postfix notation::
3604: * Your first definition::
3605: * How does that work?::
3606: * Forth is written in Forth::
3607: * Review - elements of a Forth system::
3608: * Where to go next::
3609: * Exercises::
1.29 crook 3610: @end menu
1.21 crook 3611:
1.29 crook 3612: @comment ----------------------------------------------
3613: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3614: @section Introducing the Text Interpreter
3615: @cindex text interpreter
3616: @cindex outer interpreter
1.21 crook 3617:
1.30 anton 3618: @c IMO this is too detailed and the pace is too slow for
3619: @c an introduction. If you know German, take a look at
3620: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3621: @c to see how I do it - anton
3622:
1.44 crook 3623: @c nac-> Where I have accepted your comments 100% and modified the text
3624: @c accordingly, I have deleted your comments. Elsewhere I have added a
3625: @c response like this to attempt to rationalise what I have done. Of
3626: @c course, this is a very clumsy mechanism for something that would be
3627: @c done far more efficiently over a beer. Please delete any dialogue
3628: @c you consider closed.
3629:
1.29 crook 3630: When you invoke the Forth image, you will see a startup banner printed
3631: and nothing else (if you have Gforth installed on your system, try
1.30 anton 3632: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29 crook 3633: its command line interpreter, which is called the @dfn{Text Interpreter}
3634: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
1.49 anton 3635: about the text interpreter as you read through this chapter, for more
3636: detail @pxref{The Text Interpreter}).
1.21 crook 3637:
1.29 crook 3638: Although it's not obvious, Forth is actually waiting for your
1.30 anton 3639: input. Type a number and press the @key{RET} key:
1.21 crook 3640:
1.26 crook 3641: @example
1.30 anton 3642: @kbd{45@key{RET}} ok
1.26 crook 3643: @end example
1.21 crook 3644:
1.29 crook 3645: Rather than give you a prompt to invite you to input something, the text
3646: interpreter prints a status message @i{after} it has processed a line
3647: of input. The status message in this case (``@code{ ok}'' followed by
3648: carriage-return) indicates that the text interpreter was able to process
3649: all of your input successfully. Now type something illegal:
3650:
3651: @example
1.30 anton 3652: @kbd{qwer341@key{RET}}
1.29 crook 3653: :1: Undefined word
3654: qwer341
3655: ^^^^^^^
3656: $400D2BA8 Bounce
3657: $400DBDA8 no.extensions
3658: @end example
1.23 crook 3659:
1.29 crook 3660: The exact text, other than the ``Undefined word'' may differ slightly on
3661: your system, but the effect is the same; when the text interpreter
3662: detects an error, it discards any remaining text on a line, resets
1.49 anton 3663: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3664: messages}.
1.23 crook 3665:
1.29 crook 3666: The text interpreter waits for you to press carriage-return, and then
3667: processes your input line. Starting at the beginning of the line, it
3668: breaks the line into groups of characters separated by spaces. For each
3669: group of characters in turn, it makes two attempts to do something:
1.23 crook 3670:
1.29 crook 3671: @itemize @bullet
3672: @item
1.44 crook 3673: @cindex name dictionary
1.29 crook 3674: It tries to treat it as a command. It does this by searching a @dfn{name
3675: dictionary}. If the group of characters matches an entry in the name
3676: dictionary, the name dictionary provides the text interpreter with
3677: information that allows the text interpreter perform some actions. In
3678: Forth jargon, we say that the group
3679: @cindex word
3680: @cindex definition
3681: @cindex execution token
3682: @cindex xt
3683: of characters names a @dfn{word}, that the dictionary search returns an
3684: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3685: word, and that the text interpreter executes the xt. Often, the terms
3686: @dfn{word} and @dfn{definition} are used interchangeably.
3687: @item
3688: If the text interpreter fails to find a match in the name dictionary, it
3689: tries to treat the group of characters as a number in the current number
3690: base (when you start up Forth, the current number base is base 10). If
3691: the group of characters legitimately represents a number, the text
3692: interpreter pushes the number onto a stack (we'll learn more about that
3693: in the next section).
3694: @end itemize
1.23 crook 3695:
1.29 crook 3696: If the text interpreter is unable to do either of these things with any
3697: group of characters, it discards the group of characters and the rest of
3698: the line, then prints an error message. If the text interpreter reaches
3699: the end of the line without error, it prints the status message ``@code{ ok}''
3700: followed by carriage-return.
1.21 crook 3701:
1.29 crook 3702: This is the simplest command we can give to the text interpreter:
1.23 crook 3703:
3704: @example
1.30 anton 3705: @key{RET} ok
1.23 crook 3706: @end example
1.21 crook 3707:
1.29 crook 3708: The text interpreter did everything we asked it to do (nothing) without
3709: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3710: command:
1.21 crook 3711:
1.23 crook 3712: @example
1.30 anton 3713: @kbd{12 dup fred dup@key{RET}}
1.29 crook 3714: :1: Undefined word
3715: 12 dup fred dup
3716: ^^^^
3717: $400D2BA8 Bounce
3718: $400DBDA8 no.extensions
1.23 crook 3719: @end example
1.21 crook 3720:
1.29 crook 3721: When you press the carriage-return key, the text interpreter starts to
3722: work its way along the line:
1.21 crook 3723:
1.29 crook 3724: @itemize @bullet
3725: @item
3726: When it gets to the space after the @code{2}, it takes the group of
3727: characters @code{12} and looks them up in the name
3728: dictionary@footnote{We can't tell if it found them or not, but assume
3729: for now that it did not}. There is no match for this group of characters
3730: in the name dictionary, so it tries to treat them as a number. It is
3731: able to do this successfully, so it puts the number, 12, ``on the stack''
3732: (whatever that means).
3733: @item
3734: The text interpreter resumes scanning the line and gets the next group
3735: of characters, @code{dup}. It looks it up in the name dictionary and
3736: (you'll have to take my word for this) finds it, and executes the word
3737: @code{dup} (whatever that means).
3738: @item
3739: Once again, the text interpreter resumes scanning the line and gets the
3740: group of characters @code{fred}. It looks them up in the name
3741: dictionary, but can't find them. It tries to treat them as a number, but
3742: they don't represent any legal number.
3743: @end itemize
1.21 crook 3744:
1.29 crook 3745: At this point, the text interpreter gives up and prints an error
3746: message. The error message shows exactly how far the text interpreter
3747: got in processing the line. In particular, it shows that the text
3748: interpreter made no attempt to do anything with the final character
3749: group, @code{dup}, even though we have good reason to believe that the
3750: text interpreter would have no problem looking that word up and
3751: executing it a second time.
1.21 crook 3752:
3753:
1.29 crook 3754: @comment ----------------------------------------------
3755: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3756: @section Stacks, postfix notation and parameter passing
3757: @cindex text interpreter
3758: @cindex outer interpreter
1.21 crook 3759:
1.29 crook 3760: In procedural programming languages (like C and Pascal), the
3761: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3762: functions or procedures are called with @dfn{explicit parameters}. For
3763: example, in C we might write:
1.21 crook 3764:
1.23 crook 3765: @example
1.29 crook 3766: total = total + new_volume(length,height,depth);
1.23 crook 3767: @end example
1.21 crook 3768:
1.23 crook 3769: @noindent
1.29 crook 3770: where new_volume is a function-call to another piece of code, and total,
3771: length, height and depth are all variables. length, height and depth are
3772: parameters to the function-call.
1.21 crook 3773:
1.29 crook 3774: In Forth, the equivalent of the function or procedure is the
3775: @dfn{definition} and parameters are implicitly passed between
3776: definitions using a shared stack that is visible to the
3777: programmer. Although Forth does support variables, the existence of the
3778: stack means that they are used far less often than in most other
3779: programming languages. When the text interpreter encounters a number, it
3780: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30 anton 3781: actual number is implementation-dependent ...) and the particular stack
1.29 crook 3782: used for any operation is implied unambiguously by the operation being
3783: performed. The stack used for all integer operations is called the @dfn{data
3784: stack} and, since this is the stack used most commonly, references to
3785: ``the data stack'' are often abbreviated to ``the stack''.
1.21 crook 3786:
1.29 crook 3787: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21 crook 3788:
1.23 crook 3789: @example
1.30 anton 3790: @kbd{1 2 3@key{RET}} ok
1.23 crook 3791: @end example
1.21 crook 3792:
1.29 crook 3793: Then this instructs the text interpreter to placed three numbers on the
3794: (data) stack. An analogy for the behaviour of the stack is to take a
3795: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3796: the table. The 3 was the last card onto the pile (``last-in'') and if
3797: you take a card off the pile then, unless you're prepared to fiddle a
3798: bit, the card that you take off will be the 3 (``first-out''). The
3799: number that will be first-out of the stack is called the @dfn{top of
3800: stack}, which
3801: @cindex TOS definition
3802: is often abbreviated to @dfn{TOS}.
1.21 crook 3803:
1.29 crook 3804: To understand how parameters are passed in Forth, consider the
3805: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3806: be surprised to learn that this definition performs addition. More
3807: precisely, it adds two number together and produces a result. Where does
3808: it get the two numbers from? It takes the top two numbers off the
3809: stack. Where does it place the result? On the stack. You can act-out the
3810: behaviour of @code{+} with your playing cards like this:
1.21 crook 3811:
3812: @itemize @bullet
3813: @item
1.29 crook 3814: Pick up two cards from the stack on the table
1.21 crook 3815: @item
1.29 crook 3816: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3817: numbers''
1.21 crook 3818: @item
1.29 crook 3819: Decide that the answer is 5
1.21 crook 3820: @item
1.29 crook 3821: Shuffle the two cards back into the pack and find a 5
1.21 crook 3822: @item
1.29 crook 3823: Put a 5 on the remaining ace that's on the table.
1.21 crook 3824: @end itemize
3825:
1.29 crook 3826: If you don't have a pack of cards handy but you do have Forth running,
3827: you can use the definition @code{.s} to show the current state of the stack,
3828: without affecting the stack. Type:
1.21 crook 3829:
3830: @example
1.30 anton 3831: @kbd{clearstack 1 2 3@key{RET}} ok
3832: @kbd{.s@key{RET}} <3> 1 2 3 ok
1.23 crook 3833: @end example
3834:
1.29 crook 3835: The text interpreter looks up the word @code{clearstack} and executes
3836: it; it tidies up the stack and removes any entries that may have been
3837: left on it by earlier examples. The text interpreter pushes each of the
3838: three numbers in turn onto the stack. Finally, the text interpreter
3839: looks up the word @code{.s} and executes it. The effect of executing
3840: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3841: followed by a list of all the items on the stack; the item on the far
3842: right-hand side is the TOS.
1.21 crook 3843:
1.29 crook 3844: You can now type:
1.21 crook 3845:
1.29 crook 3846: @example
1.30 anton 3847: @kbd{+ .s@key{RET}} <2> 1 5 ok
1.29 crook 3848: @end example
1.21 crook 3849:
1.29 crook 3850: @noindent
3851: which is correct; there are now 2 items on the stack and the result of
3852: the addition is 5.
1.23 crook 3853:
1.29 crook 3854: If you're playing with cards, try doing a second addition: pick up the
3855: two cards, work out that their sum is 6, shuffle them into the pack,
3856: look for a 6 and place that on the table. You now have just one item on
3857: the stack. What happens if you try to do a third addition? Pick up the
3858: first card, pick up the second card -- ah! There is no second card. This
3859: is called a @dfn{stack underflow} and consitutes an error. If you try to
3860: do the same thing with Forth it will report an error (probably a Stack
3861: Underflow or an Invalid Memory Address error).
1.23 crook 3862:
1.29 crook 3863: The opposite situation to a stack underflow is a @dfn{stack overflow},
3864: which simply accepts that there is a finite amount of storage space
3865: reserved for the stack. To stretch the playing card analogy, if you had
3866: enough packs of cards and you piled the cards up on the table, you would
3867: eventually be unable to add another card; you'd hit the ceiling. Gforth
3868: allows you to set the maximum size of the stacks. In general, the only
3869: time that you will get a stack overflow is because a definition has a
3870: bug in it and is generating data on the stack uncontrollably.
1.23 crook 3871:
1.29 crook 3872: There's one final use for the playing card analogy. If you model your
3873: stack using a pack of playing cards, the maximum number of items on
3874: your stack will be 52 (I assume you didn't use the Joker). The maximum
3875: @i{value} of any item on the stack is 13 (the King). In fact, the only
3876: possible numbers are positive integer numbers 1 through 13; you can't
3877: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3878: think about some of the cards, you can accommodate different
3879: numbers. For example, you could think of the Jack as representing 0,
3880: the Queen as representing -1 and the King as representing -2. Your
1.45 crook 3881: @i{range} remains unchanged (you can still only represent a total of 13
1.29 crook 3882: numbers) but the numbers that you can represent are -2 through 10.
1.28 crook 3883:
1.29 crook 3884: In that analogy, the limit was the amount of information that a single
3885: stack entry could hold, and Forth has a similar limit. In Forth, the
3886: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3887: implementation dependent and affects the maximum value that a stack
3888: entry can hold. A Standard Forth provides a cell size of at least
3889: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21 crook 3890:
1.29 crook 3891: Forth does not do any type checking for you, so you are free to
3892: manipulate and combine stack items in any way you wish. A convenient way
3893: of treating stack items is as 2's complement signed integers, and that
3894: is what Standard words like @code{+} do. Therefore you can type:
1.21 crook 3895:
1.29 crook 3896: @example
1.30 anton 3897: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
1.29 crook 3898: @end example
1.21 crook 3899:
1.29 crook 3900: If you use numbers and definitions like @code{+} in order to turn Forth
3901: into a great big pocket calculator, you will realise that it's rather
3902: different from a normal calculator. Rather than typing 2 + 3 = you had
3903: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3904: result). The terminology used to describe this difference is to say that
3905: your calculator uses @dfn{Infix Notation} (parameters and operators are
3906: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3907: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21 crook 3908:
1.29 crook 3909: Whilst postfix notation might look confusing to begin with, it has
3910: several important advantages:
1.21 crook 3911:
1.23 crook 3912: @itemize @bullet
3913: @item
1.29 crook 3914: it is unambiguous
1.23 crook 3915: @item
1.29 crook 3916: it is more concise
1.23 crook 3917: @item
1.29 crook 3918: it fits naturally with a stack-based system
1.23 crook 3919: @end itemize
1.21 crook 3920:
1.29 crook 3921: To examine these claims in more detail, consider these sums:
1.21 crook 3922:
1.29 crook 3923: @example
3924: 6 + 5 * 4 =
3925: 4 * 5 + 6 =
3926: @end example
1.21 crook 3927:
1.29 crook 3928: If you're just learning maths or your maths is very rusty, you will
3929: probably come up with the answer 44 for the first and 26 for the
3930: second. If you are a bit of a whizz at maths you will remember the
3931: @i{convention} that multiplication takes precendence over addition, and
3932: you'd come up with the answer 26 both times. To explain the answer 26
3933: to someone who got the answer 44, you'd probably rewrite the first sum
3934: like this:
1.21 crook 3935:
1.29 crook 3936: @example
3937: 6 + (5 * 4) =
3938: @end example
1.21 crook 3939:
1.29 crook 3940: If what you really wanted was to perform the addition before the
3941: multiplication, you would have to use parentheses to force it.
1.21 crook 3942:
1.29 crook 3943: If you did the first two sums on a pocket calculator you would probably
3944: get the right answers, unless you were very cautious and entered them using
3945: these keystroke sequences:
1.21 crook 3946:
1.29 crook 3947: 6 + 5 = * 4 =
3948: 4 * 5 = + 6 =
1.21 crook 3949:
1.29 crook 3950: Postfix notation is unambiguous because the order that the operators
3951: are applied is always explicit; that also means that parentheses are
3952: never required. The operators are @i{active} (the act of quoting the
3953: operator makes the operation occur) which removes the need for ``=''.
1.28 crook 3954:
1.29 crook 3955: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3956: equivalent ways:
1.26 crook 3957:
3958: @example
1.29 crook 3959: 6 5 4 * + or:
3960: 5 4 * 6 +
1.26 crook 3961: @end example
1.23 crook 3962:
1.29 crook 3963: An important thing that you should notice about this notation is that
3964: the @i{order} of the numbers does not change; if you want to subtract
3965: 2 from 10 you type @code{10 2 -}.
1.1 anton 3966:
1.29 crook 3967: The reason that Forth uses postfix notation is very simple to explain: it
3968: makes the implementation extremely simple, and it follows naturally from
3969: using the stack as a mechanism for passing parameters. Another way of
3970: thinking about this is to realise that all Forth definitions are
3971: @i{active}; they execute as they are encountered by the text
3972: interpreter. The result of this is that the syntax of Forth is trivially
3973: simple.
1.1 anton 3974:
3975:
3976:
1.29 crook 3977: @comment ----------------------------------------------
3978: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3979: @section Your first Forth definition
3980: @cindex first definition
1.1 anton 3981:
1.29 crook 3982: Until now, the examples we've seen have been trivial; we've just been
3983: using Forth as a bigger-than-pocket calculator. Also, each calculation
3984: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3985: again@footnote{That's not quite true. If you press the up-arrow key on
3986: your keyboard you should be able to scroll back to any earlier command,
3987: edit it and re-enter it.} In this section we'll see how to add new
3988: words to Forth's vocabulary.
1.1 anton 3989:
1.29 crook 3990: The easiest way to create a new word is to use a @dfn{colon
3991: definition}. We'll define a few and try them out before worrying too
3992: much about how they work. Try typing in these examples; be careful to
3993: copy the spaces accurately:
1.1 anton 3994:
1.29 crook 3995: @example
3996: : add-two 2 + . ;
3997: : greet ." Hello and welcome" ;
3998: : demo 5 add-two ;
3999: @end example
1.1 anton 4000:
1.29 crook 4001: @noindent
4002: Now try them out:
1.1 anton 4003:
1.29 crook 4004: @example
1.30 anton 4005: @kbd{greet@key{RET}} Hello and welcome ok
4006: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
4007: @kbd{4 add-two@key{RET}} 6 ok
4008: @kbd{demo@key{RET}} 7 ok
4009: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
1.29 crook 4010: @end example
1.1 anton 4011:
1.29 crook 4012: The first new thing that we've introduced here is the pair of words
4013: @code{:} and @code{;}. These are used to start and terminate a new
4014: definition, respectively. The first word after the @code{:} is the name
4015: for the new definition.
1.1 anton 4016:
1.29 crook 4017: As you can see from the examples, a definition is built up of words that
4018: have already been defined; Forth makes no distinction between
4019: definitions that existed when you started the system up, and those that
4020: you define yourself.
1.1 anton 4021:
1.29 crook 4022: The examples also introduce the words @code{.} (dot), @code{."}
4023: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
4024: the stack and displays it. It's like @code{.s} except that it only
4025: displays the top item of the stack and it is destructive; after it has
4026: executed, the number is no longer on the stack. There is always one
4027: space printed after the number, and no spaces before it. Dot-quote
4028: defines a string (a sequence of characters) that will be printed when
4029: the word is executed. The string can contain any printable characters
4030: except @code{"}. A @code{"} has a special function; it is not a Forth
4031: word but it acts as a delimiter (the way that delimiters work is
4032: described in the next section). Finally, @code{dup} duplicates the value
4033: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1 anton 4034:
1.29 crook 4035: We already know that the text interpreter searches through the
4036: dictionary to locate names. If you've followed the examples earlier, you
4037: will already have a definition called @code{add-two}. Lets try modifying
4038: it by typing in a new definition:
1.1 anton 4039:
1.29 crook 4040: @example
1.30 anton 4041: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
1.29 crook 4042: @end example
1.5 anton 4043:
1.29 crook 4044: Forth recognised that we were defining a word that already exists, and
4045: printed a message to warn us of that fact. Let's try out the new
4046: definition:
1.5 anton 4047:
1.29 crook 4048: @example
1.30 anton 4049: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
1.29 crook 4050: @end example
1.1 anton 4051:
1.29 crook 4052: @noindent
4053: All that we've actually done here, though, is to create a new
4054: definition, with a particular name. The fact that there was already a
4055: definition with the same name did not make any difference to the way
4056: that the new definition was created (except that Forth printed a warning
4057: message). The old definition of add-two still exists (try @code{demo}
4058: again to see that this is true). Any new definition will use the new
4059: definition of @code{add-two}, but old definitions continue to use the
4060: version that already existed at the time that they were @code{compiled}.
1.1 anton 4061:
1.29 crook 4062: Before you go on to the next section, try defining and redefining some
4063: words of your own.
1.1 anton 4064:
1.29 crook 4065: @comment ----------------------------------------------
4066: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
4067: @section How does that work?
4068: @cindex parsing words
1.1 anton 4069:
1.30 anton 4070: @c That's pretty deep (IMO way too deep) for an introduction. - anton
4071:
4072: @c Is it a good idea to talk about the interpretation semantics of a
4073: @c number? We don't have an xt to go along with it. - anton
4074:
4075: @c Now that I have eliminated execution semantics, I wonder if it would not
4076: @c be better to keep them (or add run-time semantics), to make it easier to
4077: @c explain what compilation semantics usually does. - anton
4078:
1.44 crook 4079: @c nac-> I removed the term ``default compilation sematics'' from the
4080: @c introductory chapter. Removing ``execution semantics'' was making
4081: @c everything simpler to explain, then I think the use of this term made
4082: @c everything more complex again. I replaced it with ``default
4083: @c semantics'' (which is used elsewhere in the manual) by which I mean
4084: @c ``a definition that has neither the immediate nor the compile-only
1.83 anton 4085: @c flag set''.
4086:
4087: @c anton: I have eliminated default semantics (except in one place where it
4088: @c means "default interpretation and compilation semantics"), because it
4089: @c makes no sense in the presence of combined words. I reverted to
4090: @c "execution semantics" where necessary.
4091:
4092: @c nac-> I reworded big chunks of the ``how does that work''
1.44 crook 4093: @c section (and, unusually for me, I think I even made it shorter!). See
4094: @c what you think -- I know I have not addressed your primary concern
4095: @c that it is too heavy-going for an introduction. From what I understood
4096: @c of your course notes it looks as though they might be a good framework.
4097: @c Things that I've tried to capture here are some things that came as a
4098: @c great revelation here when I first understood them. Also, I like the
4099: @c fact that a very simple code example shows up almost all of the issues
4100: @c that you need to understand to see how Forth works. That's unique and
4101: @c worthwhile to emphasise.
4102:
1.83 anton 4103: @c anton: I think it's a good idea to present the details, especially those
4104: @c that you found to be a revelation, and probably the tutorial tries to be
4105: @c too superficial and does not get some of the things across that make
4106: @c Forth special. I do believe that most of the time these things should
4107: @c be discussed at the end of a section or in separate sections instead of
4108: @c in the middle of a section (e.g., the stuff you added in "User-defined
4109: @c defining words" leads in a completely different direction from the rest
4110: @c of the section).
4111:
1.29 crook 4112: Now we're going to take another look at the definition of @code{add-two}
4113: from the previous section. From our knowledge of the way that the text
4114: interpreter works, we would have expected this result when we tried to
4115: define @code{add-two}:
1.21 crook 4116:
1.29 crook 4117: @example
1.44 crook 4118: @kbd{: add-two 2 + . ;@key{RET}}
1.29 crook 4119: ^^^^^^^
4120: Error: Undefined word
4121: @end example
1.28 crook 4122:
1.29 crook 4123: The reason that this didn't happen is bound up in the way that @code{:}
4124: works. The word @code{:} does two special things. The first special
4125: thing that it does prevents the text interpreter from ever seeing the
4126: characters @code{add-two}. The text interpreter uses a variable called
4127: @cindex modifying >IN
1.44 crook 4128: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29 crook 4129: input line. When it encounters the word @code{:} it behaves in exactly
4130: the same way as it does for any other word; it looks it up in the name
4131: dictionary, finds its xt and executes it. When @code{:} executes, it
4132: looks at the input buffer, finds the word @code{add-two} and advances the
4133: value of @code{>IN} to point past it. It then does some other stuff
4134: associated with creating the new definition (including creating an entry
4135: for @code{add-two} in the name dictionary). When the execution of @code{:}
4136: completes, control returns to the text interpreter, which is oblivious
4137: to the fact that it has been tricked into ignoring part of the input
4138: line.
1.21 crook 4139:
1.29 crook 4140: @cindex parsing words
4141: Words like @code{:} -- words that advance the value of @code{>IN} and so
4142: prevent the text interpreter from acting on the whole of the input line
4143: -- are called @dfn{parsing words}.
1.21 crook 4144:
1.29 crook 4145: @cindex @code{state} - effect on the text interpreter
4146: @cindex text interpreter - effect of state
4147: The second special thing that @code{:} does is change the value of a
4148: variable called @code{state}, which affects the way that the text
4149: interpreter behaves. When Gforth starts up, @code{state} has the value
4150: 0, and the text interpreter is said to be @dfn{interpreting}. During a
4151: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44 crook 4152: the text interpreter is said to be @dfn{compiling}.
4153:
4154: In this example, the text interpreter is compiling when it processes the
4155: string ``@code{2 + . ;}''. It still breaks the string down into
4156: character sequences in the same way. However, instead of pushing the
4157: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
4158: into the definition of @code{add-two} that will make the number @code{2} get
4159: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
4160: the behaviours of @code{+} and @code{.} are also compiled into the
4161: definition.
4162:
4163: One category of words don't get compiled. These so-called @dfn{immediate
4164: words} get executed (performed @i{now}) regardless of whether the text
4165: interpreter is interpreting or compiling. The word @code{;} is an
4166: immediate word. Rather than being compiled into the definition, it
4167: executes. Its effect is to terminate the current definition, which
4168: includes changing the value of @code{state} back to 0.
4169:
4170: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4171: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4172: definition.
1.28 crook 4173:
1.30 anton 4174: In Forth, every word or number can be described in terms of two
1.29 crook 4175: properties:
1.28 crook 4176:
4177: @itemize @bullet
4178: @item
1.29 crook 4179: @cindex interpretation semantics
1.44 crook 4180: Its @dfn{interpretation semantics} describe how it will behave when the
4181: text interpreter encounters it in @dfn{interpret} state. The
4182: interpretation semantics of a word are represented by an @dfn{execution
4183: token}.
1.28 crook 4184: @item
1.29 crook 4185: @cindex compilation semantics
1.44 crook 4186: Its @dfn{compilation semantics} describe how it will behave when the
4187: text interpreter encounters it in @dfn{compile} state. The compilation
4188: semantics of a word are represented in an implementation-dependent way;
4189: Gforth uses a @dfn{compilation token}.
1.29 crook 4190: @end itemize
4191:
4192: @noindent
4193: Numbers are always treated in a fixed way:
4194:
4195: @itemize @bullet
1.28 crook 4196: @item
1.44 crook 4197: When the number is @dfn{interpreted}, its behaviour is to push the
4198: number onto the stack.
1.28 crook 4199: @item
1.30 anton 4200: When the number is @dfn{compiled}, a piece of code is appended to the
4201: current definition that pushes the number when it runs. (In other words,
4202: the compilation semantics of a number are to postpone its interpretation
4203: semantics until the run-time of the definition that it is being compiled
4204: into.)
1.29 crook 4205: @end itemize
4206:
1.44 crook 4207: Words don't behave in such a regular way, but most have @i{default
4208: semantics} which means that they behave like this:
1.29 crook 4209:
4210: @itemize @bullet
1.28 crook 4211: @item
1.30 anton 4212: The @dfn{interpretation semantics} of the word are to do something useful.
4213: @item
1.29 crook 4214: The @dfn{compilation semantics} of the word are to append its
1.30 anton 4215: @dfn{interpretation semantics} to the current definition (so that its
4216: run-time behaviour is to do something useful).
1.28 crook 4217: @end itemize
4218:
1.30 anton 4219: @cindex immediate words
1.44 crook 4220: The actual behaviour of any particular word can be controlled by using
4221: the words @code{immediate} and @code{compile-only} when the word is
4222: defined. These words set flags in the name dictionary entry of the most
4223: recently defined word, and these flags are retrieved by the text
4224: interpreter when it finds the word in the name dictionary.
4225:
4226: A word that is marked as @dfn{immediate} has compilation semantics that
4227: are identical to its interpretation semantics. In other words, it
4228: behaves like this:
1.29 crook 4229:
4230: @itemize @bullet
4231: @item
1.30 anton 4232: The @dfn{interpretation semantics} of the word are to do something useful.
1.29 crook 4233: @item
1.30 anton 4234: The @dfn{compilation semantics} of the word are to do something useful
4235: (and actually the same thing); i.e., it is executed during compilation.
1.29 crook 4236: @end itemize
1.28 crook 4237:
1.44 crook 4238: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4239: performing the interpretation semantics of the word directly; an attempt
4240: to do so will generate an error. It is never necessary to use
4241: @code{compile-only} (and it is not even part of ANS Forth, though it is
4242: provided by many implementations) but it is good etiquette to apply it
4243: to a word that will not behave correctly (and might have unexpected
4244: side-effects) in interpret state. For example, it is only legal to use
4245: the conditional word @code{IF} within a definition. If you forget this
4246: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4247: @code{compile-only} allows the text interpreter to generate a helpful
4248: error message rather than subjecting you to the consequences of your
4249: folly.
4250:
1.29 crook 4251: This example shows the difference between an immediate and a
4252: non-immediate word:
1.28 crook 4253:
1.29 crook 4254: @example
4255: : show-state state @@ . ;
4256: : show-state-now show-state ; immediate
4257: : word1 show-state ;
4258: : word2 show-state-now ;
1.28 crook 4259: @end example
1.23 crook 4260:
1.29 crook 4261: The word @code{immediate} after the definition of @code{show-state-now}
4262: makes that word an immediate word. These definitions introduce a new
4263: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4264: variable, and leaves it on the stack. Therefore, the behaviour of
4265: @code{show-state} is to print a number that represents the current value
4266: of @code{state}.
1.28 crook 4267:
1.29 crook 4268: When you execute @code{word1}, it prints the number 0, indicating that
4269: the system is interpreting. When the text interpreter compiled the
4270: definition of @code{word1}, it encountered @code{show-state} whose
1.30 anton 4271: compilation semantics are to append its interpretation semantics to the
1.29 crook 4272: current definition. When you execute @code{word1}, it performs the
1.30 anton 4273: interpretation semantics of @code{show-state}. At the time that @code{word1}
1.29 crook 4274: (and therefore @code{show-state}) are executed, the system is
4275: interpreting.
1.28 crook 4276:
1.30 anton 4277: When you pressed @key{RET} after entering the definition of @code{word2},
1.29 crook 4278: you should have seen the number -1 printed, followed by ``@code{
4279: ok}''. When the text interpreter compiled the definition of
4280: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30 anton 4281: whose compilation semantics are therefore to perform its interpretation
1.29 crook 4282: semantics. It is executed straight away (even before the text
4283: interpreter has moved on to process another group of characters; the
4284: @code{;} in this example). The effect of executing it are to display the
4285: value of @code{state} @i{at the time that the definition of}
4286: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4287: system is compiling at this time. If you execute @code{word2} it does
4288: nothing at all.
1.28 crook 4289:
1.29 crook 4290: @cindex @code{."}, how it works
4291: Before leaving the subject of immediate words, consider the behaviour of
4292: @code{."} in the definition of @code{greet}, in the previous
4293: section. This word is both a parsing word and an immediate word. Notice
4294: that there is a space between @code{."} and the start of the text
4295: @code{Hello and welcome}, but that there is no space between the last
4296: letter of @code{welcome} and the @code{"} character. The reason for this
4297: is that @code{."} is a Forth word; it must have a space after it so that
4298: the text interpreter can identify it. The @code{"} is not a Forth word;
4299: it is a @dfn{delimiter}. The examples earlier show that, when the string
4300: is displayed, there is neither a space before the @code{H} nor after the
4301: @code{e}. Since @code{."} is an immediate word, it executes at the time
4302: that @code{greet} is defined. When it executes, its behaviour is to
4303: search forward in the input line looking for the delimiter. When it
4304: finds the delimiter, it updates @code{>IN} to point past the
4305: delimiter. It also compiles some magic code into the definition of
4306: @code{greet}; the xt of a run-time routine that prints a text string. It
4307: compiles the string @code{Hello and welcome} into memory so that it is
4308: available to be printed later. When the text interpreter gains control,
4309: the next word it finds in the input stream is @code{;} and so it
4310: terminates the definition of @code{greet}.
1.28 crook 4311:
4312:
4313: @comment ----------------------------------------------
1.29 crook 4314: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4315: @section Forth is written in Forth
4316: @cindex structure of Forth programs
4317:
4318: When you start up a Forth compiler, a large number of definitions
4319: already exist. In Forth, you develop a new application using bottom-up
4320: programming techniques to create new definitions that are defined in
4321: terms of existing definitions. As you create each definition you can
4322: test and debug it interactively.
4323:
4324: If you have tried out the examples in this section, you will probably
4325: have typed them in by hand; when you leave Gforth, your definitions will
4326: be lost. You can avoid this by using a text editor to enter Forth source
4327: code into a file, and then loading code from the file using
1.49 anton 4328: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29 crook 4329: processed by the text interpreter, just as though you had typed it in by
4330: hand@footnote{Actually, there are some subtle differences -- see
4331: @ref{The Text Interpreter}.}.
4332:
4333: Gforth also supports the traditional Forth alternative to using text
1.49 anton 4334: files for program entry (@pxref{Blocks}).
1.28 crook 4335:
1.29 crook 4336: In common with many, if not most, Forth compilers, most of Gforth is
4337: actually written in Forth. All of the @file{.fs} files in the
4338: installation directory@footnote{For example,
1.30 anton 4339: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29 crook 4340: study to see examples of Forth programming.
1.28 crook 4341:
1.29 crook 4342: Gforth maintains a history file that records every line that you type to
4343: the text interpreter. This file is preserved between sessions, and is
4344: used to provide a command-line recall facility. If you enter long
4345: definitions by hand, you can use a text editor to paste them out of the
4346: history file into a Forth source file for reuse at a later time
1.49 anton 4347: (for more information @pxref{Command-line editing}).
1.28 crook 4348:
4349:
4350: @comment ----------------------------------------------
1.29 crook 4351: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4352: @section Review - elements of a Forth system
4353: @cindex elements of a Forth system
1.28 crook 4354:
1.29 crook 4355: To summarise this chapter:
1.28 crook 4356:
4357: @itemize @bullet
4358: @item
1.29 crook 4359: Forth programs use @dfn{factoring} to break a problem down into small
4360: fragments called @dfn{words} or @dfn{definitions}.
4361: @item
4362: Forth program development is an interactive process.
4363: @item
4364: The main command loop that accepts input, and controls both
4365: interpretation and compilation, is called the @dfn{text interpreter}
4366: (also known as the @dfn{outer interpreter}).
4367: @item
4368: Forth has a very simple syntax, consisting of words and numbers
4369: separated by spaces or carriage-return characters. Any additional syntax
4370: is imposed by @dfn{parsing words}.
4371: @item
4372: Forth uses a stack to pass parameters between words. As a result, it
4373: uses postfix notation.
4374: @item
4375: To use a word that has previously been defined, the text interpreter
4376: searches for the word in the @dfn{name dictionary}.
4377: @item
1.30 anton 4378: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28 crook 4379: @item
1.29 crook 4380: The text interpreter uses the value of @code{state} to select between
4381: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4382: semantics} of a word that it encounters.
1.28 crook 4383: @item
1.30 anton 4384: The relationship between the @dfn{interpretation semantics} and
4385: @dfn{compilation semantics} for a word
1.29 crook 4386: depend upon the way in which the word was defined (for example, whether
4387: it is an @dfn{immediate} word).
1.28 crook 4388: @item
1.29 crook 4389: Forth definitions can be implemented in Forth (called @dfn{high-level
4390: definitions}) or in some other way (usually a lower-level language and
4391: as a result often called @dfn{low-level definitions}, @dfn{code
4392: definitions} or @dfn{primitives}).
1.28 crook 4393: @item
1.29 crook 4394: Many Forth systems are implemented mainly in Forth.
1.28 crook 4395: @end itemize
4396:
4397:
1.29 crook 4398: @comment ----------------------------------------------
1.48 anton 4399: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29 crook 4400: @section Where To Go Next
4401: @cindex where to go next
1.28 crook 4402:
1.29 crook 4403: Amazing as it may seem, if you have read (and understood) this far, you
4404: know almost all the fundamentals about the inner workings of a Forth
4405: system. You certainly know enough to be able to read and understand the
4406: rest of this manual and the ANS Forth document, to learn more about the
4407: facilities that Forth in general and Gforth in particular provide. Even
4408: scarier, you know almost enough to implement your own Forth system.
1.30 anton 4409: However, that's not a good idea just yet... better to try writing some
1.29 crook 4410: programs in Gforth.
1.28 crook 4411:
1.29 crook 4412: Forth has such a rich vocabulary that it can be hard to know where to
4413: start in learning it. This section suggests a few sets of words that are
4414: enough to write small but useful programs. Use the word index in this
4415: document to learn more about each word, then try it out and try to write
4416: small definitions using it. Start by experimenting with these words:
1.28 crook 4417:
4418: @itemize @bullet
4419: @item
1.29 crook 4420: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4421: @item
4422: Comparison: @code{MIN MAX =}
4423: @item
4424: Logic: @code{AND OR XOR NOT}
4425: @item
4426: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28 crook 4427: @item
1.29 crook 4428: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28 crook 4429: @item
1.29 crook 4430: Input/Output: @code{. ." EMIT CR KEY}
1.28 crook 4431: @item
1.29 crook 4432: Defining words: @code{: ; CREATE}
1.28 crook 4433: @item
1.29 crook 4434: Memory allocation words: @code{ALLOT ,}
1.28 crook 4435: @item
1.29 crook 4436: Tools: @code{SEE WORDS .S MARKER}
4437: @end itemize
4438:
4439: When you have mastered those, go on to:
4440:
4441: @itemize @bullet
1.28 crook 4442: @item
1.29 crook 4443: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28 crook 4444: @item
1.29 crook 4445: Memory access: @code{@@ !}
1.28 crook 4446: @end itemize
1.23 crook 4447:
1.29 crook 4448: When you have mastered these, there's nothing for it but to read through
4449: the whole of this manual and find out what you've missed.
4450:
4451: @comment ----------------------------------------------
1.48 anton 4452: @node Exercises, , Where to go next, Introduction
1.29 crook 4453: @section Exercises
4454: @cindex exercises
4455:
4456: TODO: provide a set of programming excercises linked into the stuff done
4457: already and into other sections of the manual. Provide solutions to all
4458: the exercises in a .fs file in the distribution.
4459:
4460: @c Get some inspiration from Starting Forth and Kelly&Spies.
4461:
4462: @c excercises:
4463: @c 1. take inches and convert to feet and inches.
4464: @c 2. take temperature and convert from fahrenheight to celcius;
4465: @c may need to care about symmetric vs floored??
4466: @c 3. take input line and do character substitution
4467: @c to encipher or decipher
4468: @c 4. as above but work on a file for in and out
4469: @c 5. take input line and convert to pig-latin
4470: @c
4471: @c thing of sets of things to exercise then come up with
4472: @c problems that need those things.
4473:
4474:
1.26 crook 4475: @c ******************************************************************
1.29 crook 4476: @node Words, Error messages, Introduction, Top
1.1 anton 4477: @chapter Forth Words
1.26 crook 4478: @cindex words
1.1 anton 4479:
4480: @menu
4481: * Notation::
1.65 anton 4482: * Case insensitivity::
4483: * Comments::
4484: * Boolean Flags::
1.1 anton 4485: * Arithmetic::
4486: * Stack Manipulation::
1.5 anton 4487: * Memory::
1.1 anton 4488: * Control Structures::
4489: * Defining Words::
1.65 anton 4490: * Interpretation and Compilation Semantics::
1.47 crook 4491: * Tokens for Words::
1.81 anton 4492: * Compiling words::
1.65 anton 4493: * The Text Interpreter::
4494: * Word Lists::
4495: * Environmental Queries::
1.12 anton 4496: * Files::
4497: * Blocks::
4498: * Other I/O::
1.78 anton 4499: * Locals::
4500: * Structures::
4501: * Object-oriented Forth::
1.12 anton 4502: * Programming Tools::
4503: * Assembler and Code Words::
4504: * Threading Words::
1.65 anton 4505: * Passing Commands to the OS::
4506: * Keeping track of Time::
4507: * Miscellaneous Words::
1.1 anton 4508: @end menu
4509:
1.65 anton 4510: @node Notation, Case insensitivity, Words, Words
1.1 anton 4511: @section Notation
4512: @cindex notation of glossary entries
4513: @cindex format of glossary entries
4514: @cindex glossary notation format
4515: @cindex word glossary entry format
4516:
4517: The Forth words are described in this section in the glossary notation
1.67 anton 4518: that has become a de-facto standard for Forth texts:
1.1 anton 4519:
4520: @format
1.29 crook 4521: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
1.1 anton 4522: @end format
1.29 crook 4523: @i{Description}
1.1 anton 4524:
4525: @table @var
4526: @item word
1.28 crook 4527: The name of the word.
1.1 anton 4528:
4529: @item Stack effect
4530: @cindex stack effect
1.29 crook 4531: The stack effect is written in the notation @code{@i{before} --
4532: @i{after}}, where @i{before} and @i{after} describe the top of
1.1 anton 4533: stack entries before and after the execution of the word. The rest of
4534: the stack is not touched by the word. The top of stack is rightmost,
4535: i.e., a stack sequence is written as it is typed in. Note that Gforth
4536: uses a separate floating point stack, but a unified stack
1.29 crook 4537: notation. Also, return stack effects are not shown in @i{stack
4538: effect}, but in @i{Description}. The name of a stack item describes
1.1 anton 4539: the type and/or the function of the item. See below for a discussion of
4540: the types.
4541:
4542: All words have two stack effects: A compile-time stack effect and a
4543: run-time stack effect. The compile-time stack-effect of most words is
1.29 crook 4544: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1 anton 4545: this standard behaviour, or the word does other unusual things at
4546: compile time, both stack effects are shown; otherwise only the run-time
4547: stack effect is shown.
4548:
4549: @cindex pronounciation of words
4550: @item pronunciation
4551: How the word is pronounced.
4552:
4553: @cindex wordset
1.67 anton 4554: @cindex environment wordset
1.1 anton 4555: @item wordset
1.21 crook 4556: The ANS Forth standard is divided into several word sets. A standard
4557: system need not support all of them. Therefore, in theory, the fewer
4558: word sets your program uses the more portable it will be. However, we
4559: suspect that most ANS Forth systems on personal machines will feature
1.26 crook 4560: all word sets. Words that are not defined in ANS Forth have
1.21 crook 4561: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1 anton 4562: describes words that will work in future releases of Gforth;
4563: @code{gforth-internal} words are more volatile. Environmental query
4564: strings are also displayed like words; you can recognize them by the
1.21 crook 4565: @code{environment} in the word set field.
1.1 anton 4566:
4567: @item Description
4568: A description of the behaviour of the word.
4569: @end table
4570:
4571: @cindex types of stack items
4572: @cindex stack item types
4573: The type of a stack item is specified by the character(s) the name
4574: starts with:
4575:
4576: @table @code
4577: @item f
4578: @cindex @code{f}, stack item type
4579: Boolean flags, i.e. @code{false} or @code{true}.
4580: @item c
4581: @cindex @code{c}, stack item type
4582: Char
4583: @item w
4584: @cindex @code{w}, stack item type
4585: Cell, can contain an integer or an address
4586: @item n
4587: @cindex @code{n}, stack item type
4588: signed integer
4589: @item u
4590: @cindex @code{u}, stack item type
4591: unsigned integer
4592: @item d
4593: @cindex @code{d}, stack item type
4594: double sized signed integer
4595: @item ud
4596: @cindex @code{ud}, stack item type
4597: double sized unsigned integer
4598: @item r
4599: @cindex @code{r}, stack item type
4600: Float (on the FP stack)
1.21 crook 4601: @item a-
1.1 anton 4602: @cindex @code{a_}, stack item type
4603: Cell-aligned address
1.21 crook 4604: @item c-
1.1 anton 4605: @cindex @code{c_}, stack item type
4606: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21 crook 4607: @item f-
1.1 anton 4608: @cindex @code{f_}, stack item type
4609: Float-aligned address
1.21 crook 4610: @item df-
1.1 anton 4611: @cindex @code{df_}, stack item type
4612: Address aligned for IEEE double precision float
1.21 crook 4613: @item sf-
1.1 anton 4614: @cindex @code{sf_}, stack item type
4615: Address aligned for IEEE single precision float
4616: @item xt
4617: @cindex @code{xt}, stack item type
4618: Execution token, same size as Cell
4619: @item wid
4620: @cindex @code{wid}, stack item type
1.21 crook 4621: Word list ID, same size as Cell
1.68 anton 4622: @item ior, wior
4623: @cindex ior type description
4624: @cindex wior type description
4625: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
1.1 anton 4626: @item f83name
4627: @cindex @code{f83name}, stack item type
4628: Pointer to a name structure
4629: @item "
4630: @cindex @code{"}, stack item type
1.12 anton 4631: string in the input stream (not on the stack). The terminating character
4632: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1 anton 4633: quotes.
4634: @end table
4635:
1.65 anton 4636: @comment ----------------------------------------------
4637: @node Case insensitivity, Comments, Notation, Words
4638: @section Case insensitivity
4639: @cindex case sensitivity
4640: @cindex upper and lower case
4641:
4642: Gforth is case-insensitive; you can enter definitions and invoke
4643: Standard words using upper, lower or mixed case (however,
4644: @pxref{core-idef, Implementation-defined options, Implementation-defined
4645: options}).
4646:
4647: ANS Forth only @i{requires} implementations to recognise Standard words
4648: when they are typed entirely in upper case. Therefore, a Standard
4649: program must use upper case for all Standard words. You can use whatever
4650: case you like for words that you define, but in a Standard program you
4651: have to use the words in the same case that you defined them.
4652:
4653: Gforth supports case sensitivity through @code{table}s (case-sensitive
4654: wordlists, @pxref{Word Lists}).
4655:
4656: Two people have asked how to convert Gforth to be case-sensitive; while
4657: we think this is a bad idea, you can change all wordlists into tables
4658: like this:
4659:
4660: @example
4661: ' table-find forth-wordlist wordlist-map @ !
4662: @end example
4663:
4664: Note that you now have to type the predefined words in the same case
4665: that we defined them, which are varying. You may want to convert them
4666: to your favourite case before doing this operation (I won't explain how,
4667: because if you are even contemplating doing this, you'd better have
4668: enough knowledge of Forth systems to know this already).
4669:
4670: @node Comments, Boolean Flags, Case insensitivity, Words
1.21 crook 4671: @section Comments
1.26 crook 4672: @cindex comments
1.21 crook 4673:
1.29 crook 4674: Forth supports two styles of comment; the traditional @i{in-line} comment,
4675: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21 crook 4676:
1.44 crook 4677:
1.23 crook 4678: doc-(
1.21 crook 4679: doc-\
1.23 crook 4680: doc-\G
1.21 crook 4681:
1.44 crook 4682:
1.21 crook 4683: @node Boolean Flags, Arithmetic, Comments, Words
4684: @section Boolean Flags
1.26 crook 4685: @cindex Boolean flags
1.21 crook 4686:
4687: A Boolean flag is cell-sized. A cell with all bits clear represents the
4688: flag @code{false} and a flag with all bits set represents the flag
1.26 crook 4689: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29 crook 4690: a cell that has @i{any} bit set as @code{true}.
1.67 anton 4691: @c on and off to Memory?
4692: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44 crook 4693:
1.21 crook 4694: doc-true
4695: doc-false
1.29 crook 4696: doc-on
4697: doc-off
1.21 crook 4698:
1.44 crook 4699:
1.21 crook 4700: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1 anton 4701: @section Arithmetic
4702: @cindex arithmetic words
4703:
4704: @cindex division with potentially negative operands
4705: Forth arithmetic is not checked, i.e., you will not hear about integer
4706: overflow on addition or multiplication, you may hear about division by
4707: zero if you are lucky. The operator is written after the operands, but
4708: the operands are still in the original order. I.e., the infix @code{2-1}
4709: corresponds to @code{2 1 -}. Forth offers a variety of division
4710: operators. If you perform division with potentially negative operands,
4711: you do not want to use @code{/} or @code{/mod} with its undefined
4712: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4713: former, @pxref{Mixed precision}).
1.26 crook 4714: @comment TODO discuss the different division forms and the std approach
1.1 anton 4715:
4716: @menu
4717: * Single precision::
1.67 anton 4718: * Double precision:: Double-cell integer arithmetic
1.1 anton 4719: * Bitwise operations::
1.67 anton 4720: * Numeric comparison::
1.29 crook 4721: * Mixed precision:: Operations with single and double-cell integers
1.1 anton 4722: * Floating Point::
4723: @end menu
4724:
1.67 anton 4725: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1 anton 4726: @subsection Single precision
4727: @cindex single precision arithmetic words
4728:
1.67 anton 4729: @c !! cell undefined
4730:
4731: By default, numbers in Forth are single-precision integers that are one
1.26 crook 4732: cell in size. They can be signed or unsigned, depending upon how you
1.49 anton 4733: treat them. For the rules used by the text interpreter for recognising
4734: single-precision integers see @ref{Number Conversion}.
1.21 crook 4735:
1.67 anton 4736: These words are all defined for signed operands, but some of them also
4737: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4738: @code{*}.
1.44 crook 4739:
1.1 anton 4740: doc-+
1.21 crook 4741: doc-1+
1.1 anton 4742: doc--
1.21 crook 4743: doc-1-
1.1 anton 4744: doc-*
4745: doc-/
4746: doc-mod
4747: doc-/mod
4748: doc-negate
4749: doc-abs
4750: doc-min
4751: doc-max
1.27 crook 4752: doc-floored
1.1 anton 4753:
1.44 crook 4754:
1.67 anton 4755: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21 crook 4756: @subsection Double precision
4757: @cindex double precision arithmetic words
4758:
1.49 anton 4759: For the rules used by the text interpreter for
4760: recognising double-precision integers, see @ref{Number Conversion}.
1.21 crook 4761:
4762: A double precision number is represented by a cell pair, with the most
1.67 anton 4763: significant cell at the TOS. It is trivial to convert an unsigned single
4764: to a double: simply push a @code{0} onto the TOS. Since numbers are
4765: represented by Gforth using 2's complement arithmetic, converting a
4766: signed single to a (signed) double requires sign-extension across the
4767: most significant cell. This can be achieved using @code{s>d}. The moral
4768: of the story is that you cannot convert a number without knowing whether
4769: it represents an unsigned or a signed number.
1.21 crook 4770:
1.67 anton 4771: These words are all defined for signed operands, but some of them also
4772: work for unsigned numbers: @code{d+}, @code{d-}.
1.44 crook 4773:
1.21 crook 4774: doc-s>d
1.67 anton 4775: doc-d>s
1.21 crook 4776: doc-d+
4777: doc-d-
4778: doc-dnegate
4779: doc-dabs
4780: doc-dmin
4781: doc-dmax
4782:
1.44 crook 4783:
1.67 anton 4784: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4785: @subsection Bitwise operations
4786: @cindex bitwise operation words
4787:
4788:
4789: doc-and
4790: doc-or
4791: doc-xor
4792: doc-invert
4793: doc-lshift
4794: doc-rshift
4795: doc-2*
4796: doc-d2*
4797: doc-2/
4798: doc-d2/
4799:
4800:
4801: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21 crook 4802: @subsection Numeric comparison
4803: @cindex numeric comparison words
4804:
1.67 anton 4805: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4806: d0= d0<>}) work for for both signed and unsigned numbers.
1.44 crook 4807:
1.28 crook 4808: doc-<
4809: doc-<=
4810: doc-<>
4811: doc-=
4812: doc->
4813: doc->=
4814:
1.21 crook 4815: doc-0<
1.23 crook 4816: doc-0<=
1.21 crook 4817: doc-0<>
4818: doc-0=
1.23 crook 4819: doc-0>
4820: doc-0>=
1.28 crook 4821:
4822: doc-u<
4823: doc-u<=
1.44 crook 4824: @c u<> and u= exist but are the same as <> and =
1.31 anton 4825: @c doc-u<>
4826: @c doc-u=
1.28 crook 4827: doc-u>
4828: doc-u>=
4829:
4830: doc-within
4831:
4832: doc-d<
4833: doc-d<=
4834: doc-d<>
4835: doc-d=
4836: doc-d>
4837: doc-d>=
1.23 crook 4838:
1.21 crook 4839: doc-d0<
1.23 crook 4840: doc-d0<=
4841: doc-d0<>
1.21 crook 4842: doc-d0=
1.23 crook 4843: doc-d0>
4844: doc-d0>=
4845:
1.21 crook 4846: doc-du<
1.28 crook 4847: doc-du<=
1.44 crook 4848: @c du<> and du= exist but are the same as d<> and d=
1.31 anton 4849: @c doc-du<>
4850: @c doc-du=
1.28 crook 4851: doc-du>
4852: doc-du>=
1.1 anton 4853:
1.44 crook 4854:
1.21 crook 4855: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1 anton 4856: @subsection Mixed precision
4857: @cindex mixed precision arithmetic words
4858:
1.44 crook 4859:
1.1 anton 4860: doc-m+
4861: doc-*/
4862: doc-*/mod
4863: doc-m*
4864: doc-um*
4865: doc-m*/
4866: doc-um/mod
4867: doc-fm/mod
4868: doc-sm/rem
4869:
1.44 crook 4870:
1.21 crook 4871: @node Floating Point, , Mixed precision, Arithmetic
1.1 anton 4872: @subsection Floating Point
4873: @cindex floating point arithmetic words
4874:
1.49 anton 4875: For the rules used by the text interpreter for
4876: recognising floating-point numbers see @ref{Number Conversion}.
1.1 anton 4877:
1.67 anton 4878: Gforth has a separate floating point stack, but the documentation uses
4879: the unified notation.@footnote{It's easy to generate the separate
4880: notation from that by just separating the floating-point numbers out:
4881: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4882: r3 )}.}
1.1 anton 4883:
4884: @cindex floating-point arithmetic, pitfalls
4885: Floating point numbers have a number of unpleasant surprises for the
4886: unwary (e.g., floating point addition is not associative) and even a few
4887: for the wary. You should not use them unless you know what you are doing
4888: or you don't care that the results you get are totally bogus. If you
4889: want to learn about the problems of floating point numbers (and how to
1.66 anton 4890: avoid them), you might start with @cite{David Goldberg,
4891: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4892: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4893: Surveys 23(1):5@minus{}48, March 1991}.
1.1 anton 4894:
1.44 crook 4895:
1.21 crook 4896: doc-d>f
4897: doc-f>d
1.1 anton 4898: doc-f+
4899: doc-f-
4900: doc-f*
4901: doc-f/
4902: doc-fnegate
4903: doc-fabs
4904: doc-fmax
4905: doc-fmin
4906: doc-floor
4907: doc-fround
4908: doc-f**
4909: doc-fsqrt
4910: doc-fexp
4911: doc-fexpm1
4912: doc-fln
4913: doc-flnp1
4914: doc-flog
4915: doc-falog
1.32 anton 4916: doc-f2*
4917: doc-f2/
4918: doc-1/f
4919: doc-precision
4920: doc-set-precision
4921:
4922: @cindex angles in trigonometric operations
4923: @cindex trigonometric operations
4924: Angles in floating point operations are given in radians (a full circle
4925: has 2 pi radians).
4926:
1.1 anton 4927: doc-fsin
4928: doc-fcos
4929: doc-fsincos
4930: doc-ftan
4931: doc-fasin
4932: doc-facos
4933: doc-fatan
4934: doc-fatan2
4935: doc-fsinh
4936: doc-fcosh
4937: doc-ftanh
4938: doc-fasinh
4939: doc-facosh
4940: doc-fatanh
1.21 crook 4941: doc-pi
1.28 crook 4942:
1.32 anton 4943: @cindex equality of floats
4944: @cindex floating-point comparisons
1.31 anton 4945: One particular problem with floating-point arithmetic is that comparison
4946: for equality often fails when you would expect it to succeed. For this
4947: reason approximate equality is often preferred (but you still have to
1.67 anton 4948: know what you are doing). Also note that IEEE NaNs may compare
1.68 anton 4949: differently from what you might expect. The comparison words are:
1.31 anton 4950:
4951: doc-f~rel
4952: doc-f~abs
1.68 anton 4953: doc-f~
1.31 anton 4954: doc-f=
4955: doc-f<>
4956:
4957: doc-f<
4958: doc-f<=
4959: doc-f>
4960: doc-f>=
4961:
1.21 crook 4962: doc-f0<
1.28 crook 4963: doc-f0<=
4964: doc-f0<>
1.21 crook 4965: doc-f0=
1.28 crook 4966: doc-f0>
4967: doc-f0>=
4968:
1.1 anton 4969:
4970: @node Stack Manipulation, Memory, Arithmetic, Words
4971: @section Stack Manipulation
4972: @cindex stack manipulation words
4973:
4974: @cindex floating-point stack in the standard
1.21 crook 4975: Gforth maintains a number of separate stacks:
4976:
1.29 crook 4977: @cindex data stack
4978: @cindex parameter stack
1.21 crook 4979: @itemize @bullet
4980: @item
1.29 crook 4981: A data stack (also known as the @dfn{parameter stack}) -- for
4982: characters, cells, addresses, and double cells.
1.21 crook 4983:
1.29 crook 4984: @cindex floating-point stack
1.21 crook 4985: @item
1.44 crook 4986: A floating point stack -- for holding floating point (FP) numbers.
1.21 crook 4987:
1.29 crook 4988: @cindex return stack
1.21 crook 4989: @item
1.44 crook 4990: A return stack -- for holding the return addresses of colon
1.32 anton 4991: definitions and other (non-FP) data.
1.21 crook 4992:
1.29 crook 4993: @cindex locals stack
1.21 crook 4994: @item
1.44 crook 4995: A locals stack -- for holding local variables.
1.21 crook 4996: @end itemize
4997:
1.1 anton 4998: @menu
4999: * Data stack::
5000: * Floating point stack::
5001: * Return stack::
5002: * Locals stack::
5003: * Stack pointer manipulation::
5004: @end menu
5005:
5006: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
5007: @subsection Data stack
5008: @cindex data stack manipulation words
5009: @cindex stack manipulations words, data stack
5010:
1.44 crook 5011:
1.1 anton 5012: doc-drop
5013: doc-nip
5014: doc-dup
5015: doc-over
5016: doc-tuck
5017: doc-swap
1.21 crook 5018: doc-pick
1.1 anton 5019: doc-rot
5020: doc--rot
5021: doc-?dup
5022: doc-roll
5023: doc-2drop
5024: doc-2nip
5025: doc-2dup
5026: doc-2over
5027: doc-2tuck
5028: doc-2swap
5029: doc-2rot
5030:
1.44 crook 5031:
1.1 anton 5032: @node Floating point stack, Return stack, Data stack, Stack Manipulation
5033: @subsection Floating point stack
5034: @cindex floating-point stack manipulation words
5035: @cindex stack manipulation words, floating-point stack
5036:
1.32 anton 5037: Whilst every sane Forth has a separate floating-point stack, it is not
5038: strictly required; an ANS Forth system could theoretically keep
5039: floating-point numbers on the data stack. As an additional difficulty,
5040: you don't know how many cells a floating-point number takes. It is
5041: reportedly possible to write words in a way that they work also for a
5042: unified stack model, but we do not recommend trying it. Instead, just
5043: say that your program has an environmental dependency on a separate
5044: floating-point stack.
5045:
5046: doc-floating-stack
5047:
1.1 anton 5048: doc-fdrop
5049: doc-fnip
5050: doc-fdup
5051: doc-fover
5052: doc-ftuck
5053: doc-fswap
1.21 crook 5054: doc-fpick
1.1 anton 5055: doc-frot
5056:
1.44 crook 5057:
1.1 anton 5058: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
5059: @subsection Return stack
5060: @cindex return stack manipulation words
5061: @cindex stack manipulation words, return stack
5062:
1.32 anton 5063: @cindex return stack and locals
5064: @cindex locals and return stack
5065: A Forth system is allowed to keep local variables on the
5066: return stack. This is reasonable, as local variables usually eliminate
5067: the need to use the return stack explicitly. So, if you want to produce
5068: a standard compliant program and you are using local variables in a
5069: word, forget about return stack manipulations in that word (refer to the
5070: standard document for the exact rules).
5071:
1.1 anton 5072: doc->r
5073: doc-r>
5074: doc-r@
5075: doc-rdrop
5076: doc-2>r
5077: doc-2r>
5078: doc-2r@
5079: doc-2rdrop
5080:
1.44 crook 5081:
1.1 anton 5082: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
5083: @subsection Locals stack
5084:
1.78 anton 5085: Gforth uses an extra locals stack. It is described, along with the
5086: reasons for its existence, in @ref{Locals implementation}.
1.21 crook 5087:
1.1 anton 5088: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
5089: @subsection Stack pointer manipulation
5090: @cindex stack pointer manipulation words
5091:
1.44 crook 5092: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21 crook 5093: doc-sp0
1.1 anton 5094: doc-sp@
5095: doc-sp!
1.21 crook 5096: doc-fp0
1.1 anton 5097: doc-fp@
5098: doc-fp!
1.21 crook 5099: doc-rp0
1.1 anton 5100: doc-rp@
5101: doc-rp!
1.21 crook 5102: doc-lp0
1.1 anton 5103: doc-lp@
5104: doc-lp!
5105:
1.44 crook 5106:
1.1 anton 5107: @node Memory, Control Structures, Stack Manipulation, Words
5108: @section Memory
1.26 crook 5109: @cindex memory words
1.1 anton 5110:
1.32 anton 5111: @menu
5112: * Memory model::
5113: * Dictionary allocation::
5114: * Heap Allocation::
5115: * Memory Access::
5116: * Address arithmetic::
5117: * Memory Blocks::
5118: @end menu
5119:
1.67 anton 5120: In addition to the standard Forth memory allocation words, there is also
5121: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
5122: garbage collector}.
5123:
1.32 anton 5124: @node Memory model, Dictionary allocation, Memory, Memory
5125: @subsection ANS Forth and Gforth memory models
5126:
5127: @c The ANS Forth description is a mess (e.g., is the heap part of
5128: @c the dictionary?), so let's not stick to closely with it.
5129:
1.67 anton 5130: ANS Forth considers a Forth system as consisting of several address
5131: spaces, of which only @dfn{data space} is managed and accessible with
5132: the memory words. Memory not necessarily in data space includes the
5133: stacks, the code (called code space) and the headers (called name
5134: space). In Gforth everything is in data space, but the code for the
5135: primitives is usually read-only.
1.32 anton 5136:
5137: Data space is divided into a number of areas: The (data space portion of
5138: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
5139: refer to the search data structure embodied in word lists and headers,
5140: because it is used for looking up names, just as you would in a
5141: conventional dictionary.}, the heap, and a number of system-allocated
5142: buffers.
5143:
1.68 anton 5144: @cindex address arithmetic restrictions, ANS vs. Gforth
5145: @cindex contiguous regions, ANS vs. Gforth
1.32 anton 5146: In ANS Forth data space is also divided into contiguous regions. You
5147: can only use address arithmetic within a contiguous region, not between
5148: them. Usually each allocation gives you one contiguous region, but the
1.33 anton 5149: dictionary allocation words have additional rules (@pxref{Dictionary
1.32 anton 5150: allocation}).
5151:
5152: Gforth provides one big address space, and address arithmetic can be
5153: performed between any addresses. However, in the dictionary headers or
5154: code are interleaved with data, so almost the only contiguous data space
5155: regions there are those described by ANS Forth as contiguous; but you
5156: can be sure that the dictionary is allocated towards increasing
5157: addresses even between contiguous regions. The memory order of
5158: allocations in the heap is platform-dependent (and possibly different
5159: from one run to the next).
5160:
1.27 crook 5161:
1.32 anton 5162: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5163: @subsection Dictionary allocation
1.27 crook 5164: @cindex reserving data space
5165: @cindex data space - reserving some
5166:
1.32 anton 5167: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5168: you want to deallocate X, you also deallocate everything
5169: allocated after X.
5170:
1.68 anton 5171: @cindex contiguous regions in dictionary allocation
1.32 anton 5172: The allocations using the words below are contiguous and grow the region
5173: towards increasing addresses. Other words that allocate dictionary
5174: memory of any kind (i.e., defining words including @code{:noname}) end
5175: the contiguous region and start a new one.
5176:
5177: In ANS Forth only @code{create}d words are guaranteed to produce an
5178: address that is the start of the following contiguous region. In
5179: particular, the cell allocated by @code{variable} is not guaranteed to
5180: be contiguous with following @code{allot}ed memory.
5181:
5182: You can deallocate memory by using @code{allot} with a negative argument
5183: (with some restrictions, see @code{allot}). For larger deallocations use
5184: @code{marker}.
1.27 crook 5185:
1.29 crook 5186:
1.27 crook 5187: doc-here
5188: doc-unused
5189: doc-allot
5190: doc-c,
1.29 crook 5191: doc-f,
1.27 crook 5192: doc-,
5193: doc-2,
5194:
1.32 anton 5195: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5196: course you should allocate memory in an aligned way, too. I.e., before
5197: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5198: The words below align @code{here} if it is not already. Basically it is
5199: only already aligned for a type, if the last allocation was a multiple
5200: of the size of this type and if @code{here} was aligned for this type
5201: before.
5202:
5203: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5204: ANS Forth (@code{maxalign}ed in Gforth).
5205:
5206: doc-align
5207: doc-falign
5208: doc-sfalign
5209: doc-dfalign
5210: doc-maxalign
5211: doc-cfalign
5212:
5213:
5214: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5215: @subsection Heap allocation
5216: @cindex heap allocation
5217: @cindex dynamic allocation of memory
5218: @cindex memory-allocation word set
5219:
1.68 anton 5220: @cindex contiguous regions and heap allocation
1.32 anton 5221: Heap allocation supports deallocation of allocated memory in any
5222: order. Dictionary allocation is not affected by it (i.e., it does not
5223: end a contiguous region). In Gforth, these words are implemented using
5224: the standard C library calls malloc(), free() and resize().
5225:
1.68 anton 5226: The memory region produced by one invocation of @code{allocate} or
5227: @code{resize} is internally contiguous. There is no contiguity between
5228: such a region and any other region (including others allocated from the
5229: heap).
5230:
1.32 anton 5231: doc-allocate
5232: doc-free
5233: doc-resize
5234:
1.27 crook 5235:
1.32 anton 5236: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1 anton 5237: @subsection Memory Access
5238: @cindex memory access words
5239:
5240: doc-@
5241: doc-!
5242: doc-+!
5243: doc-c@
5244: doc-c!
5245: doc-2@
5246: doc-2!
5247: doc-f@
5248: doc-f!
5249: doc-sf@
5250: doc-sf!
5251: doc-df@
5252: doc-df!
5253:
1.68 anton 5254:
1.32 anton 5255: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5256: @subsection Address arithmetic
1.1 anton 5257: @cindex address arithmetic words
5258:
1.67 anton 5259: Address arithmetic is the foundation on which you can build data
5260: structures like arrays, records (@pxref{Structures}) and objects
5261: (@pxref{Object-oriented Forth}).
1.32 anton 5262:
1.68 anton 5263: @cindex address unit
5264: @cindex au (address unit)
1.1 anton 5265: ANS Forth does not specify the sizes of the data types. Instead, it
5266: offers a number of words for computing sizes and doing address
1.29 crook 5267: arithmetic. Address arithmetic is performed in terms of address units
5268: (aus); on most systems the address unit is one byte. Note that a
5269: character may have more than one au, so @code{chars} is no noop (on
1.68 anton 5270: platforms where it is a noop, it compiles to nothing).
1.1 anton 5271:
1.67 anton 5272: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5273: you have the address of a cell, perform @code{1 cells +}, and you will
5274: have the address of the next cell.
5275:
1.68 anton 5276: @cindex contiguous regions and address arithmetic
1.67 anton 5277: In ANS Forth you can perform address arithmetic only within a contiguous
5278: region, i.e., if you have an address into one region, you can only add
5279: and subtract such that the result is still within the region; you can
5280: only subtract or compare addresses from within the same contiguous
5281: region. Reasons: several contiguous regions can be arranged in memory
5282: in any way; on segmented systems addresses may have unusual
5283: representations, such that address arithmetic only works within a
5284: region. Gforth provides a few more guarantees (linear address space,
5285: dictionary grows upwards), but in general I have found it easy to stay
5286: within contiguous regions (exception: computing and comparing to the
5287: address just beyond the end of an array).
5288:
1.1 anton 5289: @cindex alignment of addresses for types
5290: ANS Forth also defines words for aligning addresses for specific
5291: types. Many computers require that accesses to specific data types
5292: must only occur at specific addresses; e.g., that cells may only be
5293: accessed at addresses divisible by 4. Even if a machine allows unaligned
5294: accesses, it can usually perform aligned accesses faster.
5295:
5296: For the performance-conscious: alignment operations are usually only
5297: necessary during the definition of a data structure, not during the
5298: (more frequent) accesses to it.
5299:
5300: ANS Forth defines no words for character-aligning addresses. This is not
5301: an oversight, but reflects the fact that addresses that are not
5302: char-aligned have no use in the standard and therefore will not be
5303: created.
5304:
5305: @cindex @code{CREATE} and alignment
1.29 crook 5306: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1 anton 5307: are cell-aligned; in addition, Gforth guarantees that these addresses
5308: are aligned for all purposes.
5309:
1.26 crook 5310: Note that the ANS Forth word @code{char} has nothing to do with address
5311: arithmetic.
1.1 anton 5312:
1.44 crook 5313:
1.1 anton 5314: doc-chars
5315: doc-char+
5316: doc-cells
5317: doc-cell+
5318: doc-cell
5319: doc-aligned
5320: doc-floats
5321: doc-float+
5322: doc-float
5323: doc-faligned
5324: doc-sfloats
5325: doc-sfloat+
5326: doc-sfaligned
5327: doc-dfloats
5328: doc-dfloat+
5329: doc-dfaligned
5330: doc-maxaligned
5331: doc-cfaligned
5332: doc-address-unit-bits
5333:
1.44 crook 5334:
1.32 anton 5335: @node Memory Blocks, , Address arithmetic, Memory
1.1 anton 5336: @subsection Memory Blocks
5337: @cindex memory block words
1.27 crook 5338: @cindex character strings - moving and copying
5339:
1.49 anton 5340: Memory blocks often represent character strings; For ways of storing
5341: character strings in memory see @ref{String Formats}. For other
5342: string-processing words see @ref{Displaying characters and strings}.
1.1 anton 5343:
1.67 anton 5344: A few of these words work on address unit blocks. In that case, you
5345: usually have to insert @code{CHARS} before the word when working on
5346: character strings. Most words work on character blocks, and expect a
5347: char-aligned address.
5348:
5349: When copying characters between overlapping memory regions, use
5350: @code{chars move} or choose carefully between @code{cmove} and
5351: @code{cmove>}.
1.44 crook 5352:
1.1 anton 5353: doc-move
5354: doc-erase
5355: doc-cmove
5356: doc-cmove>
5357: doc-fill
5358: doc-blank
1.21 crook 5359: doc-compare
5360: doc-search
1.27 crook 5361: doc--trailing
5362: doc-/string
1.82 anton 5363: doc-bounds
1.44 crook 5364:
1.27 crook 5365: @comment TODO examples
5366:
1.1 anton 5367:
1.26 crook 5368: @node Control Structures, Defining Words, Memory, Words
1.1 anton 5369: @section Control Structures
5370: @cindex control structures
5371:
1.33 anton 5372: Control structures in Forth cannot be used interpretively, only in a
5373: colon definition@footnote{To be precise, they have no interpretation
5374: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5375: not like this limitation, but have not seen a satisfying way around it
5376: yet, although many schemes have been proposed.
1.1 anton 5377:
5378: @menu
1.33 anton 5379: * Selection:: IF ... ELSE ... ENDIF
5380: * Simple Loops:: BEGIN ...
1.29 crook 5381: * Counted Loops:: DO
1.67 anton 5382: * Arbitrary control structures::
5383: * Calls and returns::
1.1 anton 5384: * Exception Handling::
5385: @end menu
5386:
5387: @node Selection, Simple Loops, Control Structures, Control Structures
5388: @subsection Selection
5389: @cindex selection control structures
5390: @cindex control structures for selection
5391:
5392: @cindex @code{IF} control structure
5393: @example
1.29 crook 5394: @i{flag}
1.1 anton 5395: IF
1.29 crook 5396: @i{code}
1.1 anton 5397: ENDIF
5398: @end example
1.21 crook 5399: @noindent
1.33 anton 5400:
1.44 crook 5401: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5402: with any bit set represents truth) @i{code} is executed.
1.33 anton 5403:
1.1 anton 5404: @example
1.29 crook 5405: @i{flag}
1.1 anton 5406: IF
1.29 crook 5407: @i{code1}
1.1 anton 5408: ELSE
1.29 crook 5409: @i{code2}
1.1 anton 5410: ENDIF
5411: @end example
5412:
1.44 crook 5413: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5414: executed.
1.33 anton 5415:
1.1 anton 5416: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5417: standard, and @code{ENDIF} is not, although it is quite popular. We
5418: recommend using @code{ENDIF}, because it is less confusing for people
5419: who also know other languages (and is not prone to reinforcing negative
5420: prejudices against Forth in these people). Adding @code{ENDIF} to a
5421: system that only supplies @code{THEN} is simple:
5422: @example
1.82 anton 5423: : ENDIF POSTPONE then ; immediate
1.1 anton 5424: @end example
5425:
5426: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5427: (adv.)} has the following meanings:
5428: @quotation
5429: ... 2b: following next after in order ... 3d: as a necessary consequence
5430: (if you were there, then you saw them).
5431: @end quotation
5432: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5433: and many other programming languages has the meaning 3d.]
5434:
1.21 crook 5435: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1 anton 5436: you can avoid using @code{?dup}. Using these alternatives is also more
1.26 crook 5437: efficient than using @code{?dup}. Definitions in ANS Forth
1.1 anton 5438: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5439: @file{compat/control.fs}.
5440:
5441: @cindex @code{CASE} control structure
5442: @example
1.29 crook 5443: @i{n}
1.1 anton 5444: CASE
1.29 crook 5445: @i{n1} OF @i{code1} ENDOF
5446: @i{n2} OF @i{code2} ENDOF
1.1 anton 5447: @dots{}
1.68 anton 5448: ( n ) @i{default-code} ( n )
1.1 anton 5449: ENDCASE
5450: @end example
5451:
1.68 anton 5452: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If no
5453: @i{ni} matches, the optional @i{default-code} is executed. The optional
5454: default case can be added by simply writing the code after the last
5455: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
5456: not consume it.
1.1 anton 5457:
1.69 anton 5458: @progstyle
5459: To keep the code understandable, you should ensure that on all paths
5460: through a selection construct the stack is changed in the same way
5461: (wrt. number and types of stack items consumed and pushed).
5462:
1.1 anton 5463: @node Simple Loops, Counted Loops, Selection, Control Structures
5464: @subsection Simple Loops
5465: @cindex simple loops
5466: @cindex loops without count
5467:
5468: @cindex @code{WHILE} loop
5469: @example
5470: BEGIN
1.29 crook 5471: @i{code1}
5472: @i{flag}
1.1 anton 5473: WHILE
1.29 crook 5474: @i{code2}
1.1 anton 5475: REPEAT
5476: @end example
5477:
1.29 crook 5478: @i{code1} is executed and @i{flag} is computed. If it is true,
5479: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1 anton 5480: false, execution continues after the @code{REPEAT}.
5481:
5482: @cindex @code{UNTIL} loop
5483: @example
5484: BEGIN
1.29 crook 5485: @i{code}
5486: @i{flag}
1.1 anton 5487: UNTIL
5488: @end example
5489:
1.29 crook 5490: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1 anton 5491:
1.69 anton 5492: @progstyle
5493: To keep the code understandable, a complete iteration of the loop should
5494: not change the number and types of the items on the stacks.
5495:
1.1 anton 5496: @cindex endless loop
5497: @cindex loops, endless
5498: @example
5499: BEGIN
1.29 crook 5500: @i{code}
1.1 anton 5501: AGAIN
5502: @end example
5503:
5504: This is an endless loop.
5505:
5506: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5507: @subsection Counted Loops
5508: @cindex counted loops
5509: @cindex loops, counted
5510: @cindex @code{DO} loops
5511:
5512: The basic counted loop is:
5513: @example
1.29 crook 5514: @i{limit} @i{start}
1.1 anton 5515: ?DO
1.29 crook 5516: @i{body}
1.1 anton 5517: LOOP
5518: @end example
5519:
1.29 crook 5520: This performs one iteration for every integer, starting from @i{start}
5521: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21 crook 5522: accessed with @code{i}. For example, the loop:
1.1 anton 5523: @example
5524: 10 0 ?DO
5525: i .
5526: LOOP
5527: @end example
1.21 crook 5528: @noindent
5529: prints @code{0 1 2 3 4 5 6 7 8 9}
5530:
1.1 anton 5531: The index of the innermost loop can be accessed with @code{i}, the index
5532: of the next loop with @code{j}, and the index of the third loop with
5533: @code{k}.
5534:
1.44 crook 5535:
1.1 anton 5536: doc-i
5537: doc-j
5538: doc-k
5539:
1.44 crook 5540:
1.1 anton 5541: The loop control data are kept on the return stack, so there are some
1.21 crook 5542: restrictions on mixing return stack accesses and counted loop words. In
5543: particuler, if you put values on the return stack outside the loop, you
5544: cannot read them inside the loop@footnote{well, not in a way that is
5545: portable.}. If you put values on the return stack within a loop, you
5546: have to remove them before the end of the loop and before accessing the
5547: index of the loop.
1.1 anton 5548:
5549: There are several variations on the counted loop:
5550:
1.21 crook 5551: @itemize @bullet
5552: @item
5553: @code{LEAVE} leaves the innermost counted loop immediately; execution
5554: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5555:
5556: @example
5557: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5558: @end example
5559: prints @code{0 1 2 3}
5560:
1.1 anton 5561:
1.21 crook 5562: @item
5563: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5564: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5565: return stack so @code{EXIT} can get to its return address. For example:
5566:
5567: @example
5568: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5569: @end example
5570: prints @code{0 1 2 3}
5571:
5572:
5573: @item
1.29 crook 5574: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1 anton 5575: (and @code{LOOP} iterates until they become equal by wrap-around
5576: arithmetic). This behaviour is usually not what you want. Therefore,
5577: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29 crook 5578: @code{?DO}), which do not enter the loop if @i{start} is greater than
5579: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1 anton 5580: unsigned loop parameters.
5581:
1.21 crook 5582: @item
5583: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5584: the loop, independent of the loop parameters. Do not use @code{DO}, even
5585: if you know that the loop is entered in any case. Such knowledge tends
5586: to become invalid during maintenance of a program, and then the
5587: @code{DO} will make trouble.
5588:
5589: @item
1.29 crook 5590: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5591: index by @i{n} instead of by 1. The loop is terminated when the border
5592: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1 anton 5593:
1.21 crook 5594: @example
5595: 4 0 +DO i . 2 +LOOP
5596: @end example
5597: @noindent
5598: prints @code{0 2}
5599:
5600: @example
5601: 4 1 +DO i . 2 +LOOP
5602: @end example
5603: @noindent
5604: prints @code{1 3}
1.1 anton 5605:
1.68 anton 5606: @item
1.1 anton 5607: @cindex negative increment for counted loops
5608: @cindex counted loops with negative increment
1.29 crook 5609: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1 anton 5610:
1.21 crook 5611: @example
5612: -1 0 ?DO i . -1 +LOOP
5613: @end example
5614: @noindent
5615: prints @code{0 -1}
1.1 anton 5616:
1.21 crook 5617: @example
5618: 0 0 ?DO i . -1 +LOOP
5619: @end example
5620: prints nothing.
1.1 anton 5621:
1.29 crook 5622: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5623: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5624: index by @i{u} each iteration. The loop is terminated when the border
5625: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1 anton 5626: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5627:
1.21 crook 5628: @example
5629: -2 0 -DO i . 1 -LOOP
5630: @end example
5631: @noindent
5632: prints @code{0 -1}
1.1 anton 5633:
1.21 crook 5634: @example
5635: -1 0 -DO i . 1 -LOOP
5636: @end example
5637: @noindent
5638: prints @code{0}
5639:
5640: @example
5641: 0 0 -DO i . 1 -LOOP
5642: @end example
5643: @noindent
5644: prints nothing.
1.1 anton 5645:
1.21 crook 5646: @end itemize
1.1 anton 5647:
5648: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26 crook 5649: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5650: for these words that uses only standard words is provided in
5651: @file{compat/loops.fs}.
1.1 anton 5652:
5653:
5654: @cindex @code{FOR} loops
1.26 crook 5655: Another counted loop is:
1.1 anton 5656: @example
1.29 crook 5657: @i{n}
1.1 anton 5658: FOR
1.29 crook 5659: @i{body}
1.1 anton 5660: NEXT
5661: @end example
5662: This is the preferred loop of native code compiler writers who are too
1.26 crook 5663: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29 crook 5664: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5665: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26 crook 5666: Forth systems may behave differently, even if they support @code{FOR}
5667: loops. To avoid problems, don't use @code{FOR} loops.
1.1 anton 5668:
5669: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5670: @subsection Arbitrary control structures
5671: @cindex control structures, user-defined
5672:
5673: @cindex control-flow stack
5674: ANS Forth permits and supports using control structures in a non-nested
5675: way. Information about incomplete control structures is stored on the
5676: control-flow stack. This stack may be implemented on the Forth data
5677: stack, and this is what we have done in Gforth.
5678:
5679: @cindex @code{orig}, control-flow stack item
5680: @cindex @code{dest}, control-flow stack item
5681: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5682: entry represents a backward branch target. A few words are the basis for
5683: building any control structure possible (except control structures that
5684: need storage, like calls, coroutines, and backtracking).
5685:
1.44 crook 5686:
1.1 anton 5687: doc-if
5688: doc-ahead
5689: doc-then
5690: doc-begin
5691: doc-until
5692: doc-again
5693: doc-cs-pick
5694: doc-cs-roll
5695:
1.44 crook 5696:
1.21 crook 5697: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5698: manipulate the control-flow stack in a portable way. Without them, you
5699: would need to know how many stack items are occupied by a control-flow
5700: entry (many systems use one cell. In Gforth they currently take three,
5701: but this may change in the future).
5702:
1.1 anton 5703: Some standard control structure words are built from these words:
5704:
1.44 crook 5705:
1.1 anton 5706: doc-else
5707: doc-while
5708: doc-repeat
5709:
1.44 crook 5710:
5711: @noindent
1.1 anton 5712: Gforth adds some more control-structure words:
5713:
1.44 crook 5714:
1.1 anton 5715: doc-endif
5716: doc-?dup-if
5717: doc-?dup-0=-if
5718:
1.44 crook 5719:
5720: @noindent
1.1 anton 5721: Counted loop words constitute a separate group of words:
5722:
1.44 crook 5723:
1.1 anton 5724: doc-?do
5725: doc-+do
5726: doc-u+do
5727: doc--do
5728: doc-u-do
5729: doc-do
5730: doc-for
5731: doc-loop
5732: doc-+loop
5733: doc--loop
5734: doc-next
5735: doc-leave
5736: doc-?leave
5737: doc-unloop
5738: doc-done
5739:
1.44 crook 5740:
1.21 crook 5741: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5742: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1 anton 5743: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5744: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5745: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5746: resolved (by using one of the loop-ending words or @code{DONE}).
5747:
1.44 crook 5748: @noindent
1.26 crook 5749: Another group of control structure words are:
1.1 anton 5750:
1.44 crook 5751:
1.1 anton 5752: doc-case
5753: doc-endcase
5754: doc-of
5755: doc-endof
5756:
1.44 crook 5757:
1.21 crook 5758: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5759: @code{CS-ROLL}.
1.1 anton 5760:
5761: @subsubsection Programming Style
1.47 crook 5762: @cindex control structures programming style
5763: @cindex programming style, arbitrary control structures
1.1 anton 5764:
5765: In order to ensure readability we recommend that you do not create
5766: arbitrary control structures directly, but define new control structure
5767: words for the control structure you want and use these words in your
1.26 crook 5768: program. For example, instead of writing:
1.1 anton 5769:
5770: @example
1.26 crook 5771: BEGIN
1.1 anton 5772: ...
1.26 crook 5773: IF [ 1 CS-ROLL ]
1.1 anton 5774: ...
1.26 crook 5775: AGAIN THEN
1.1 anton 5776: @end example
5777:
1.21 crook 5778: @noindent
1.1 anton 5779: we recommend defining control structure words, e.g.,
5780:
5781: @example
1.26 crook 5782: : WHILE ( DEST -- ORIG DEST )
5783: POSTPONE IF
5784: 1 CS-ROLL ; immediate
5785:
5786: : REPEAT ( orig dest -- )
5787: POSTPONE AGAIN
5788: POSTPONE THEN ; immediate
1.1 anton 5789: @end example
5790:
1.21 crook 5791: @noindent
1.1 anton 5792: and then using these to create the control structure:
5793:
5794: @example
1.26 crook 5795: BEGIN
1.1 anton 5796: ...
1.26 crook 5797: WHILE
1.1 anton 5798: ...
1.26 crook 5799: REPEAT
1.1 anton 5800: @end example
5801:
5802: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5803: @code{WHILE} are predefined, so in this example it would not be
5804: necessary to define them.
5805:
5806: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5807: @subsection Calls and returns
5808: @cindex calling a definition
5809: @cindex returning from a definition
5810:
1.3 anton 5811: @cindex recursive definitions
5812: A definition can be called simply be writing the name of the definition
1.26 crook 5813: to be called. Normally a definition is invisible during its own
1.3 anton 5814: definition. If you want to write a directly recursive definition, you
1.26 crook 5815: can use @code{recursive} to make the current definition visible, or
5816: @code{recurse} to call the current definition directly.
1.3 anton 5817:
1.44 crook 5818:
1.3 anton 5819: doc-recursive
5820: doc-recurse
5821:
1.44 crook 5822:
1.21 crook 5823: @comment TODO add example of the two recursion methods
1.12 anton 5824: @quotation
5825: @progstyle
5826: I prefer using @code{recursive} to @code{recurse}, because calling the
5827: definition by name is more descriptive (if the name is well-chosen) than
5828: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5829: implementation, it is much better to read (and think) ``now sort the
5830: partitions'' than to read ``now do a recursive call''.
5831: @end quotation
1.3 anton 5832:
1.29 crook 5833: For mutual recursion, use @code{Defer}red words, like this:
1.3 anton 5834:
5835: @example
1.28 crook 5836: Defer foo
1.3 anton 5837:
5838: : bar ( ... -- ... )
5839: ... foo ... ;
5840:
5841: :noname ( ... -- ... )
5842: ... bar ... ;
5843: IS foo
5844: @end example
5845:
1.44 crook 5846: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33 anton 5847:
1.26 crook 5848: The current definition returns control to the calling definition when
1.33 anton 5849: the end of the definition is reached or @code{EXIT} is encountered.
1.1 anton 5850:
5851: doc-exit
5852: doc-;s
5853:
1.44 crook 5854:
1.1 anton 5855: @node Exception Handling, , Calls and returns, Control Structures
5856: @subsection Exception Handling
1.26 crook 5857: @cindex exceptions
1.1 anton 5858:
1.68 anton 5859: @c quit is a very bad idea for error handling,
5860: @c because it does not translate into a THROW
5861: @c it also does not belong into this chapter
5862:
5863: If a word detects an error condition that it cannot handle, it can
5864: @code{throw} an exception. In the simplest case, this will terminate
5865: your program, and report an appropriate error.
1.21 crook 5866:
1.68 anton 5867: doc-throw
1.1 anton 5868:
1.69 anton 5869: @code{Throw} consumes a cell-sized error number on the stack. There are
5870: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5871: Gforth (and most other systems) you can use the iors produced by various
5872: words as error numbers (e.g., a typical use of @code{allocate} is
5873: @code{allocate throw}). Gforth also provides the word @code{exception}
5874: to define your own error numbers (with decent error reporting); an ANS
5875: Forth version of this word (but without the error messages) is available
5876: in @code{compat/except.fs}. And finally, you can use your own error
1.68 anton 5877: numbers (anything outside the range -4095..0), but won't get nice error
5878: messages, only numbers. For example, try:
5879:
5880: @example
1.69 anton 5881: -10 throw \ ANS defined
5882: -267 throw \ system defined
5883: s" my error" exception throw \ user defined
5884: 7 throw \ arbitrary number
1.68 anton 5885: @end example
5886:
5887: doc---exception-exception
1.1 anton 5888:
1.69 anton 5889: A common idiom to @code{THROW} a specific error if a flag is true is
5890: this:
5891:
5892: @example
5893: @code{( flag ) 0<> @i{errno} and throw}
5894: @end example
5895:
5896: Your program can provide exception handlers to catch exceptions. An
5897: exception handler can be used to correct the problem, or to clean up
5898: some data structures and just throw the exception to the next exception
5899: handler. Note that @code{throw} jumps to the dynamically innermost
5900: exception handler. The system's exception handler is outermost, and just
5901: prints an error and restarts command-line interpretation (or, in batch
5902: mode (i.e., while processing the shell command line), leaves Gforth).
1.1 anton 5903:
1.68 anton 5904: The ANS Forth way to catch exceptions is @code{catch}:
1.1 anton 5905:
1.68 anton 5906: doc-catch
5907:
5908: The most common use of exception handlers is to clean up the state when
5909: an error happens. E.g.,
1.1 anton 5910:
1.26 crook 5911: @example
1.68 anton 5912: base @ >r hex \ actually the hex should be inside foo, or we h
5913: ['] foo catch ( nerror|0 )
5914: r> base !
1.69 anton 5915: ( nerror|0 ) throw \ pass it on
1.26 crook 5916: @end example
1.1 anton 5917:
1.69 anton 5918: A use of @code{catch} for handling the error @code{myerror} might look
5919: like this:
1.44 crook 5920:
1.68 anton 5921: @example
1.69 anton 5922: ['] foo catch
5923: CASE
5924: myerror OF ... ( do something about it ) ENDOF
5925: dup throw \ default: pass other errors on, do nothing on non-errors
5926: ENDCASE
1.68 anton 5927: @end example
1.44 crook 5928:
1.68 anton 5929: Having to wrap the code into a separate word is often cumbersome,
5930: therefore Gforth provides an alternative syntax:
1.1 anton 5931:
5932: @example
1.69 anton 5933: TRY
1.68 anton 5934: @i{code1}
1.69 anton 5935: RECOVER \ optional
1.68 anton 5936: @i{code2} \ optional
1.69 anton 5937: ENDTRY
1.1 anton 5938: @end example
5939:
1.68 anton 5940: This performs @i{Code1}. If @i{code1} completes normally, execution
5941: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5942: reset to the state during @code{try}, the throw value is pushed on the
5943: data stack, and execution constinues at @i{code2}, and finally falls
5944: through the @code{endtry} into the following code. If there is no
5945: @code{recover} clause, this works like an empty recover clause.
1.26 crook 5946:
1.68 anton 5947: doc-try
5948: doc-recover
5949: doc-endtry
1.26 crook 5950:
1.69 anton 5951: The cleanup example from above in this syntax:
1.26 crook 5952:
1.68 anton 5953: @example
1.69 anton 5954: base @ >r TRY
1.68 anton 5955: hex foo \ now the hex is placed correctly
1.69 anton 5956: 0 \ value for throw
5957: ENDTRY
1.68 anton 5958: r> base ! throw
1.1 anton 5959: @end example
5960:
1.69 anton 5961: And here's the error handling example:
1.1 anton 5962:
1.68 anton 5963: @example
1.69 anton 5964: TRY
1.68 anton 5965: foo
1.69 anton 5966: RECOVER
5967: CASE
5968: myerror OF ... ( do something about it ) ENDOF
5969: throw \ pass other errors on
5970: ENDCASE
5971: ENDTRY
1.68 anton 5972: @end example
1.1 anton 5973:
1.69 anton 5974: @progstyle
5975: As usual, you should ensure that the stack depth is statically known at
5976: the end: either after the @code{throw} for passing on errors, or after
5977: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5978: selection construct for handling the error).
5979:
1.68 anton 5980: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5981: and you can provide an error message. @code{Abort} just produces an
5982: ``Aborted'' error.
1.1 anton 5983:
1.68 anton 5984: The problem with these words is that exception handlers cannot
5985: differentiate between different @code{abort"}s; they just look like
5986: @code{-2 throw} to them (the error message cannot be accessed by
5987: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5988: exception handlers.
1.44 crook 5989:
1.68 anton 5990: doc-abort"
1.26 crook 5991: doc-abort
1.29 crook 5992:
5993:
1.44 crook 5994:
1.29 crook 5995: @c -------------------------------------------------------------
1.47 crook 5996: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29 crook 5997: @section Defining Words
5998: @cindex defining words
5999:
1.47 crook 6000: Defining words are used to extend Forth by creating new entries in the dictionary.
6001:
1.29 crook 6002: @menu
1.67 anton 6003: * CREATE::
1.44 crook 6004: * Variables:: Variables and user variables
1.67 anton 6005: * Constants::
1.44 crook 6006: * Values:: Initialised variables
1.67 anton 6007: * Colon Definitions::
1.44 crook 6008: * Anonymous Definitions:: Definitions without names
1.69 anton 6009: * Supplying names:: Passing definition names as strings
1.67 anton 6010: * User-defined Defining Words::
1.44 crook 6011: * Deferred words:: Allow forward references
1.67 anton 6012: * Aliases::
1.29 crook 6013: @end menu
6014:
1.44 crook 6015: @node CREATE, Variables, Defining Words, Defining Words
6016: @subsection @code{CREATE}
1.29 crook 6017: @cindex simple defining words
6018: @cindex defining words, simple
6019:
6020: Defining words are used to create new entries in the dictionary. The
6021: simplest defining word is @code{CREATE}. @code{CREATE} is used like
6022: this:
6023:
6024: @example
6025: CREATE new-word1
6026: @end example
6027:
1.69 anton 6028: @code{CREATE} is a parsing word, i.e., it takes an argument from the
6029: input stream (@code{new-word1} in our example). It generates a
6030: dictionary entry for @code{new-word1}. When @code{new-word1} is
6031: executed, all that it does is leave an address on the stack. The address
6032: represents the value of the data space pointer (@code{HERE}) at the time
6033: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
6034: associating a name with the address of a region of memory.
1.29 crook 6035:
1.34 anton 6036: doc-create
6037:
1.69 anton 6038: Note that in ANS Forth guarantees only for @code{create} that its body
6039: is in dictionary data space (i.e., where @code{here}, @code{allot}
6040: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
6041: @code{create}d words can be modified with @code{does>}
6042: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
6043: can only be applied to @code{create}d words.
6044:
1.29 crook 6045: By extending this example to reserve some memory in data space, we end
1.69 anton 6046: up with something like a @i{variable}. Here are two different ways to do
6047: it:
1.29 crook 6048:
6049: @example
6050: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
6051: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
6052: @end example
6053:
6054: The variable can be examined and modified using @code{@@} (``fetch'') and
6055: @code{!} (``store'') like this:
6056:
6057: @example
6058: new-word2 @@ . \ get address, fetch from it and display
6059: 1234 new-word2 ! \ new value, get address, store to it
6060: @end example
6061:
1.44 crook 6062: @cindex arrays
6063: A similar mechanism can be used to create arrays. For example, an
6064: 80-character text input buffer:
1.29 crook 6065:
6066: @example
1.44 crook 6067: CREATE text-buf 80 chars allot
6068:
6069: text-buf 0 chars c@@ \ the 1st character (offset 0)
6070: text-buf 3 chars c@@ \ the 4th character (offset 3)
6071: @end example
1.29 crook 6072:
1.44 crook 6073: You can build arbitrarily complex data structures by allocating
1.49 anton 6074: appropriate areas of memory. For further discussions of this, and to
1.66 anton 6075: learn about some Gforth tools that make it easier,
1.49 anton 6076: @xref{Structures}.
1.44 crook 6077:
6078:
6079: @node Variables, Constants, CREATE, Defining Words
6080: @subsection Variables
6081: @cindex variables
6082:
6083: The previous section showed how a sequence of commands could be used to
6084: generate a variable. As a final refinement, the whole code sequence can
6085: be wrapped up in a defining word (pre-empting the subject of the next
6086: section), making it easier to create new variables:
6087:
6088: @example
6089: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
6090: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
6091:
6092: myvariableX foo \ variable foo starts off with an unknown value
6093: myvariable0 joe \ whilst joe is initialised to 0
1.29 crook 6094:
6095: 45 3 * foo ! \ set foo to 135
6096: 1234 joe ! \ set joe to 1234
6097: 3 joe +! \ increment joe by 3.. to 1237
6098: @end example
6099:
6100: Not surprisingly, there is no need to define @code{myvariable}, since
1.44 crook 6101: Forth already has a definition @code{Variable}. ANS Forth does not
1.69 anton 6102: guarantee that a @code{Variable} is initialised when it is created
6103: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
6104: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
6105: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47 crook 6106: @code{fvariable} for double and floating-point variables, respectively
1.69 anton 6107: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47 crook 6108: store a boolean, you can use @code{on} and @code{off} to toggle its
6109: state.
1.29 crook 6110:
1.34 anton 6111: doc-variable
6112: doc-2variable
6113: doc-fvariable
6114:
1.29 crook 6115: @cindex user variables
6116: @cindex user space
6117: The defining word @code{User} behaves in the same way as @code{Variable}.
6118: The difference is that it reserves space in @i{user (data) space} rather
6119: than normal data space. In a Forth system that has a multi-tasker, each
6120: task has its own set of user variables.
6121:
1.34 anton 6122: doc-user
1.67 anton 6123: @c doc-udp
6124: @c doc-uallot
1.34 anton 6125:
1.29 crook 6126: @comment TODO is that stuff about user variables strictly correct? Is it
6127: @comment just terminal tasks that have user variables?
6128: @comment should document tasker.fs (with some examples) elsewhere
6129: @comment in this manual, then expand on user space and user variables.
6130:
1.44 crook 6131: @node Constants, Values, Variables, Defining Words
6132: @subsection Constants
6133: @cindex constants
6134:
6135: @code{Constant} allows you to declare a fixed value and refer to it by
6136: name. For example:
1.29 crook 6137:
6138: @example
6139: 12 Constant INCHES-PER-FOOT
6140: 3E+08 fconstant SPEED-O-LIGHT
6141: @end example
6142:
6143: A @code{Variable} can be both read and written, so its run-time
6144: behaviour is to supply an address through which its current value can be
6145: manipulated. In contrast, the value of a @code{Constant} cannot be
6146: changed once it has been declared@footnote{Well, often it can be -- but
6147: not in a Standard, portable way. It's safer to use a @code{Value} (read
6148: on).} so it's not necessary to supply the address -- it is more
6149: efficient to return the value of the constant directly. That's exactly
6150: what happens; the run-time effect of a constant is to put its value on
1.49 anton 6151: the top of the stack (You can find one
6152: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29 crook 6153:
1.69 anton 6154: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29 crook 6155: double and floating-point constants, respectively.
6156:
1.34 anton 6157: doc-constant
6158: doc-2constant
6159: doc-fconstant
6160:
6161: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44 crook 6162: @c nac-> How could that not be true in an ANS Forth? You can't define a
6163: @c constant, use it and then delete the definition of the constant..
1.69 anton 6164:
6165: @c anton->An ANS Forth system can compile a constant to a literal; On
6166: @c decompilation you would see only the number, just as if it had been used
6167: @c in the first place. The word will stay, of course, but it will only be
6168: @c used by the text interpreter (no run-time duties, except when it is
6169: @c POSTPONEd or somesuch).
6170:
6171: @c nac:
1.44 crook 6172: @c I agree that it's rather deep, but IMO it is an important difference
6173: @c relative to other programming languages.. often it's annoying: it
6174: @c certainly changes my programming style relative to C.
6175:
1.69 anton 6176: @c anton: In what way?
6177:
1.29 crook 6178: Constants in Forth behave differently from their equivalents in other
6179: programming languages. In other languages, a constant (such as an EQU in
6180: assembler or a #define in C) only exists at compile-time; in the
6181: executable program the constant has been translated into an absolute
6182: number and, unless you are using a symbolic debugger, it's impossible to
6183: know what abstract thing that number represents. In Forth a constant has
1.44 crook 6184: an entry in the header space and remains there after the code that uses
6185: it has been defined. In fact, it must remain in the dictionary since it
6186: has run-time duties to perform. For example:
1.29 crook 6187:
6188: @example
6189: 12 Constant INCHES-PER-FOOT
6190: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6191: @end example
6192:
6193: @cindex in-lining of constants
6194: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6195: associated with the constant @code{INCHES-PER-FOOT}. If you use
6196: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6197: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6198: attempt to optimise constants by in-lining them where they are used. You
6199: can force Gforth to in-line a constant like this:
6200:
6201: @example
6202: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6203: @end example
6204:
6205: If you use @code{see} to decompile @i{this} version of
6206: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49 anton 6207: longer present. To understand how this works, read
6208: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29 crook 6209:
6210: In-lining constants in this way might improve execution time
6211: fractionally, and can ensure that a constant is now only referenced at
6212: compile-time. However, the definition of the constant still remains in
6213: the dictionary. Some Forth compilers provide a mechanism for controlling
6214: a second dictionary for holding transient words such that this second
6215: dictionary can be deleted later in order to recover memory
6216: space. However, there is no standard way of doing this.
6217:
6218:
1.44 crook 6219: @node Values, Colon Definitions, Constants, Defining Words
6220: @subsection Values
6221: @cindex values
1.34 anton 6222:
1.69 anton 6223: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6224: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6225: (not in ANS Forth) you can access (and change) a @code{value} also with
6226: @code{>body}.
6227:
6228: Here are some
6229: examples:
1.29 crook 6230:
6231: @example
1.69 anton 6232: 12 Value APPLES \ Define APPLES with an initial value of 12
6233: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6234: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6235: APPLES \ puts 35 on the top of the stack.
1.29 crook 6236: @end example
6237:
1.44 crook 6238: doc-value
6239: doc-to
1.29 crook 6240:
1.35 anton 6241:
1.69 anton 6242:
1.44 crook 6243: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6244: @subsection Colon Definitions
6245: @cindex colon definitions
1.35 anton 6246:
6247: @example
1.44 crook 6248: : name ( ... -- ... )
6249: word1 word2 word3 ;
1.29 crook 6250: @end example
6251:
1.44 crook 6252: @noindent
6253: Creates a word called @code{name} that, upon execution, executes
6254: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29 crook 6255:
1.49 anton 6256: The explanation above is somewhat superficial. For simple examples of
6257: colon definitions see @ref{Your first definition}. For an in-depth
1.66 anton 6258: discussion of some of the issues involved, @xref{Interpretation and
1.49 anton 6259: Compilation Semantics}.
1.29 crook 6260:
1.44 crook 6261: doc-:
6262: doc-;
1.1 anton 6263:
1.34 anton 6264:
1.69 anton 6265: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44 crook 6266: @subsection Anonymous Definitions
6267: @cindex colon definitions
6268: @cindex defining words without name
1.34 anton 6269:
1.44 crook 6270: Sometimes you want to define an @dfn{anonymous word}; a word without a
6271: name. You can do this with:
1.1 anton 6272:
1.44 crook 6273: doc-:noname
1.1 anton 6274:
1.44 crook 6275: This leaves the execution token for the word on the stack after the
6276: closing @code{;}. Here's an example in which a deferred word is
6277: initialised with an @code{xt} from an anonymous colon definition:
1.1 anton 6278:
1.29 crook 6279: @example
1.44 crook 6280: Defer deferred
6281: :noname ( ... -- ... )
6282: ... ;
6283: IS deferred
1.29 crook 6284: @end example
1.26 crook 6285:
1.44 crook 6286: @noindent
6287: Gforth provides an alternative way of doing this, using two separate
6288: words:
1.27 crook 6289:
1.44 crook 6290: doc-noname
6291: @cindex execution token of last defined word
6292: doc-lastxt
1.1 anton 6293:
1.44 crook 6294: @noindent
6295: The previous example can be rewritten using @code{noname} and
6296: @code{lastxt}:
1.1 anton 6297:
1.26 crook 6298: @example
1.44 crook 6299: Defer deferred
6300: noname : ( ... -- ... )
6301: ... ;
6302: lastxt IS deferred
1.26 crook 6303: @end example
1.1 anton 6304:
1.29 crook 6305: @noindent
1.44 crook 6306: @code{noname} works with any defining word, not just @code{:}.
6307:
6308: @code{lastxt} also works when the last word was not defined as
1.71 anton 6309: @code{noname}. It does not work for combined words, though. It also has
6310: the useful property that is is valid as soon as the header for a
6311: definition has been built. Thus:
1.44 crook 6312:
6313: @example
6314: lastxt . : foo [ lastxt . ] ; ' foo .
6315: @end example
1.1 anton 6316:
1.44 crook 6317: @noindent
6318: prints 3 numbers; the last two are the same.
1.26 crook 6319:
1.69 anton 6320: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6321: @subsection Supplying the name of a defined word
6322: @cindex names for defined words
6323: @cindex defining words, name given in a string
6324:
6325: By default, a defining word takes the name for the defined word from the
6326: input stream. Sometimes you want to supply the name from a string. You
6327: can do this with:
6328:
6329: doc-nextname
6330:
6331: For example:
6332:
6333: @example
6334: s" foo" nextname create
6335: @end example
6336:
6337: @noindent
6338: is equivalent to:
6339:
6340: @example
6341: create foo
6342: @end example
6343:
6344: @noindent
6345: @code{nextname} works with any defining word.
6346:
1.1 anton 6347:
1.69 anton 6348: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
1.26 crook 6349: @subsection User-defined Defining Words
6350: @cindex user-defined defining words
6351: @cindex defining words, user-defined
1.1 anton 6352:
1.29 crook 6353: You can create a new defining word by wrapping defining-time code around
6354: an existing defining word and putting the sequence in a colon
1.69 anton 6355: definition.
6356:
6357: @c anton: This example is very complex and leads in a quite different
6358: @c direction from the CREATE-DOES> stuff that follows. It should probably
6359: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6360: @c subsection of Defining Words)
6361:
6362: For example, suppose that you have a word @code{stats} that
1.29 crook 6363: gathers statistics about colon definitions given the @i{xt} of the
6364: definition, and you want every colon definition in your application to
6365: make a call to @code{stats}. You can define and use a new version of
6366: @code{:} like this:
6367:
6368: @example
6369: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6370: ... ; \ other code
6371:
6372: : my: : lastxt postpone literal ['] stats compile, ;
6373:
6374: my: foo + - ;
6375: @end example
6376:
6377: When @code{foo} is defined using @code{my:} these steps occur:
6378:
6379: @itemize @bullet
6380: @item
6381: @code{my:} is executed.
6382: @item
6383: The @code{:} within the definition (the one between @code{my:} and
6384: @code{lastxt}) is executed, and does just what it always does; it parses
6385: the input stream for a name, builds a dictionary header for the name
6386: @code{foo} and switches @code{state} from interpret to compile.
6387: @item
6388: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6389: being defined -- @code{foo} -- onto the stack.
6390: @item
6391: The code that was produced by @code{postpone literal} is executed; this
6392: causes the value on the stack to be compiled as a literal in the code
6393: area of @code{foo}.
6394: @item
6395: The code @code{['] stats} compiles a literal into the definition of
6396: @code{my:}. When @code{compile,} is executed, that literal -- the
6397: execution token for @code{stats} -- is layed down in the code area of
6398: @code{foo} , following the literal@footnote{Strictly speaking, the
6399: mechanism that @code{compile,} uses to convert an @i{xt} into something
6400: in the code area is implementation-dependent. A threaded implementation
6401: might spit out the execution token directly whilst another
6402: implementation might spit out a native code sequence.}.
6403: @item
6404: At this point, the execution of @code{my:} is complete, and control
6405: returns to the text interpreter. The text interpreter is in compile
6406: state, so subsequent text @code{+ -} is compiled into the definition of
6407: @code{foo} and the @code{;} terminates the definition as always.
6408: @end itemize
6409:
6410: You can use @code{see} to decompile a word that was defined using
6411: @code{my:} and see how it is different from a normal @code{:}
6412: definition. For example:
6413:
6414: @example
6415: : bar + - ; \ like foo but using : rather than my:
6416: see bar
6417: : bar
6418: + - ;
6419: see foo
6420: : foo
6421: 107645672 stats + - ;
6422:
6423: \ use ' stats . to show that 107645672 is the xt for stats
6424: @end example
6425:
6426: You can use techniques like this to make new defining words in terms of
6427: @i{any} existing defining word.
1.1 anton 6428:
6429:
1.29 crook 6430: @cindex defining defining words
1.26 crook 6431: @cindex @code{CREATE} ... @code{DOES>}
6432: If you want the words defined with your defining words to behave
6433: differently from words defined with standard defining words, you can
6434: write your defining word like this:
1.1 anton 6435:
6436: @example
1.26 crook 6437: : def-word ( "name" -- )
1.29 crook 6438: CREATE @i{code1}
1.26 crook 6439: DOES> ( ... -- ... )
1.29 crook 6440: @i{code2} ;
1.26 crook 6441:
6442: def-word name
1.1 anton 6443: @end example
6444:
1.29 crook 6445: @cindex child words
6446: This fragment defines a @dfn{defining word} @code{def-word} and then
6447: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6448: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6449: is not executed at this time. The word @code{name} is sometimes called a
6450: @dfn{child} of @code{def-word}.
6451:
6452: When you execute @code{name}, the address of the body of @code{name} is
6453: put on the data stack and @i{code2} is executed (the address of the body
6454: of @code{name} is the address @code{HERE} returns immediately after the
1.69 anton 6455: @code{CREATE}, i.e., the address a @code{create}d word returns by
6456: default).
6457:
6458: @c anton:
6459: @c www.dictionary.com says:
6460: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6461: @c several generations of absence, usually caused by the chance
6462: @c recombination of genes. 2.An individual or a part that exhibits
6463: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6464: @c of previous behavior after a period of absence.
6465: @c
6466: @c Doesn't seem to fit.
1.29 crook 6467:
1.69 anton 6468: @c @cindex atavism in child words
1.33 anton 6469: You can use @code{def-word} to define a set of child words that behave
1.69 anton 6470: similarly; they all have a common run-time behaviour determined by
6471: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6472: body of the child word. The structure of the data is common to all
6473: children of @code{def-word}, but the data values are specific -- and
6474: private -- to each child word. When a child word is executed, the
6475: address of its private data area is passed as a parameter on TOS to be
6476: used and manipulated@footnote{It is legitimate both to read and write to
6477: this data area.} by @i{code2}.
1.29 crook 6478:
6479: The two fragments of code that make up the defining words act (are
6480: executed) at two completely separate times:
1.1 anton 6481:
1.29 crook 6482: @itemize @bullet
6483: @item
6484: At @i{define time}, the defining word executes @i{code1} to generate a
6485: child word
6486: @item
6487: At @i{child execution time}, when a child word is invoked, @i{code2}
6488: is executed, using parameters (data) that are private and specific to
6489: the child word.
6490: @end itemize
6491:
1.44 crook 6492: Another way of understanding the behaviour of @code{def-word} and
6493: @code{name} is to say that, if you make the following definitions:
1.33 anton 6494: @example
6495: : def-word1 ( "name" -- )
6496: CREATE @i{code1} ;
6497:
6498: : action1 ( ... -- ... )
6499: @i{code2} ;
6500:
6501: def-word1 name1
6502: @end example
6503:
1.44 crook 6504: @noindent
6505: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1 anton 6506:
1.29 crook 6507: The classic example is that you can define @code{CONSTANT} in this way:
1.26 crook 6508:
1.1 anton 6509: @example
1.29 crook 6510: : CONSTANT ( w "name" -- )
6511: CREATE ,
1.26 crook 6512: DOES> ( -- w )
6513: @@ ;
1.1 anton 6514: @end example
6515:
1.29 crook 6516: @comment There is a beautiful description of how this works and what
6517: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6518: @comment commentary on the Counting Fruits problem.
6519:
6520: When you create a constant with @code{5 CONSTANT five}, a set of
6521: define-time actions take place; first a new word @code{five} is created,
6522: then the value 5 is laid down in the body of @code{five} with
1.44 crook 6523: @code{,}. When @code{five} is executed, the address of the body is put on
1.29 crook 6524: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6525: no code of its own; it simply contains a data field and a pointer to the
6526: code that follows @code{DOES>} in its defining word. That makes words
6527: created in this way very compact.
6528:
6529: The final example in this section is intended to remind you that space
6530: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6531: both read and written by a Standard program@footnote{Exercise: use this
6532: example as a starting point for your own implementation of @code{Value}
6533: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6534: @code{[']}.}:
6535:
6536: @example
6537: : foo ( "name" -- )
6538: CREATE -1 ,
6539: DOES> ( -- )
1.33 anton 6540: @@ . ;
1.29 crook 6541:
6542: foo first-word
6543: foo second-word
6544:
6545: 123 ' first-word >BODY !
6546: @end example
6547:
6548: If @code{first-word} had been a @code{CREATE}d word, we could simply
6549: have executed it to get the address of its data field. However, since it
6550: was defined to have @code{DOES>} actions, its execution semantics are to
6551: perform those @code{DOES>} actions. To get the address of its data field
6552: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6553: translate the xt into the address of the data field. When you execute
6554: @code{first-word}, it will display @code{123}. When you execute
6555: @code{second-word} it will display @code{-1}.
1.26 crook 6556:
6557: @cindex stack effect of @code{DOES>}-parts
6558: @cindex @code{DOES>}-parts, stack effect
1.29 crook 6559: In the examples above the stack comment after the @code{DOES>} specifies
1.26 crook 6560: the stack effect of the defined words, not the stack effect of the
6561: following code (the following code expects the address of the body on
6562: the top of stack, which is not reflected in the stack comment). This is
6563: the convention that I use and recommend (it clashes a bit with using
6564: locals declarations for stack effect specification, though).
1.1 anton 6565:
1.53 anton 6566: @menu
6567: * CREATE..DOES> applications::
6568: * CREATE..DOES> details::
1.63 anton 6569: * Advanced does> usage example::
1.91 ! anton 6570: * @code{Const-does>}::
1.53 anton 6571: @end menu
6572:
6573: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26 crook 6574: @subsubsection Applications of @code{CREATE..DOES>}
6575: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1 anton 6576:
1.26 crook 6577: You may wonder how to use this feature. Here are some usage patterns:
1.1 anton 6578:
1.26 crook 6579: @cindex factoring similar colon definitions
6580: When you see a sequence of code occurring several times, and you can
6581: identify a meaning, you will factor it out as a colon definition. When
6582: you see similar colon definitions, you can factor them using
6583: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6584: that look very similar:
1.1 anton 6585: @example
1.26 crook 6586: : ori, ( reg-target reg-source n -- )
6587: 0 asm-reg-reg-imm ;
6588: : andi, ( reg-target reg-source n -- )
6589: 1 asm-reg-reg-imm ;
1.1 anton 6590: @end example
6591:
1.26 crook 6592: @noindent
6593: This could be factored with:
6594: @example
6595: : reg-reg-imm ( op-code -- )
6596: CREATE ,
6597: DOES> ( reg-target reg-source n -- )
6598: @@ asm-reg-reg-imm ;
6599:
6600: 0 reg-reg-imm ori,
6601: 1 reg-reg-imm andi,
6602: @end example
1.1 anton 6603:
1.26 crook 6604: @cindex currying
6605: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6606: supply a part of the parameters for a word (known as @dfn{currying} in
6607: the functional language community). E.g., @code{+} needs two
6608: parameters. Creating versions of @code{+} with one parameter fixed can
6609: be done like this:
1.82 anton 6610:
1.1 anton 6611: @example
1.82 anton 6612: : curry+ ( n1 "name" -- )
1.26 crook 6613: CREATE ,
6614: DOES> ( n2 -- n1+n2 )
6615: @@ + ;
6616:
6617: 3 curry+ 3+
6618: -2 curry+ 2-
1.1 anton 6619: @end example
6620:
1.91 ! anton 6621:
1.63 anton 6622: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26 crook 6623: @subsubsection The gory details of @code{CREATE..DOES>}
6624: @cindex @code{CREATE} ... @code{DOES>}, details
1.1 anton 6625:
1.26 crook 6626: doc-does>
1.1 anton 6627:
1.26 crook 6628: @cindex @code{DOES>} in a separate definition
6629: This means that you need not use @code{CREATE} and @code{DOES>} in the
6630: same definition; you can put the @code{DOES>}-part in a separate
1.29 crook 6631: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26 crook 6632: @example
6633: : does1
6634: DOES> ( ... -- ... )
1.44 crook 6635: ... ;
6636:
6637: : does2
6638: DOES> ( ... -- ... )
6639: ... ;
6640:
6641: : def-word ( ... -- ... )
6642: create ...
6643: IF
6644: does1
6645: ELSE
6646: does2
6647: ENDIF ;
6648: @end example
6649:
6650: In this example, the selection of whether to use @code{does1} or
1.69 anton 6651: @code{does2} is made at definition-time; at the time that the child word is
1.44 crook 6652: @code{CREATE}d.
6653:
6654: @cindex @code{DOES>} in interpretation state
6655: In a standard program you can apply a @code{DOES>}-part only if the last
6656: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6657: will override the behaviour of the last word defined in any case. In a
6658: standard program, you can use @code{DOES>} only in a colon
6659: definition. In Gforth, you can also use it in interpretation state, in a
6660: kind of one-shot mode; for example:
6661: @example
6662: CREATE name ( ... -- ... )
6663: @i{initialization}
6664: DOES>
6665: @i{code} ;
6666: @end example
6667:
6668: @noindent
6669: is equivalent to the standard:
6670: @example
6671: :noname
6672: DOES>
6673: @i{code} ;
6674: CREATE name EXECUTE ( ... -- ... )
6675: @i{initialization}
6676: @end example
6677:
1.53 anton 6678: doc->body
6679:
1.91 ! anton 6680: @node Advanced does> usage example, @code{Const-does>}, CREATE..DOES> details, User-defined Defining Words
1.63 anton 6681: @subsubsection Advanced does> usage example
6682:
6683: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6684: for disassembling instructions, that follow a very repetetive scheme:
6685:
6686: @example
6687: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6688: @var{entry-num} cells @var{table} + !
6689: @end example
6690:
6691: Of course, this inspires the idea to factor out the commonalities to
6692: allow a definition like
6693:
6694: @example
6695: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6696: @end example
6697:
6698: The parameters @var{disasm-operands} and @var{table} are usually
1.69 anton 6699: correlated. Moreover, before I wrote the disassembler, there already
6700: existed code that defines instructions like this:
1.63 anton 6701:
6702: @example
6703: @var{entry-num} @var{inst-format} @var{inst-name}
6704: @end example
6705:
6706: This code comes from the assembler and resides in
6707: @file{arch/mips/insts.fs}.
6708:
6709: So I had to define the @var{inst-format} words that performed the scheme
6710: above when executed. At first I chose to use run-time code-generation:
6711:
6712: @example
6713: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6714: :noname Postpone @var{disasm-operands}
6715: name Postpone sliteral Postpone type Postpone ;
6716: swap cells @var{table} + ! ;
6717: @end example
6718:
6719: Note that this supplies the other two parameters of the scheme above.
1.44 crook 6720:
1.63 anton 6721: An alternative would have been to write this using
6722: @code{create}/@code{does>}:
6723:
6724: @example
6725: : @var{inst-format} ( entry-num "name" -- )
6726: here name string, ( entry-num c-addr ) \ parse and save "name"
6727: noname create , ( entry-num )
6728: lastxt swap cells @var{table} + !
6729: does> ( addr w -- )
6730: \ disassemble instruction w at addr
6731: @@ >r
6732: @var{disasm-operands}
6733: r> count type ;
6734: @end example
6735:
6736: Somehow the first solution is simpler, mainly because it's simpler to
6737: shift a string from definition-time to use-time with @code{sliteral}
6738: than with @code{string,} and friends.
6739:
6740: I wrote a lot of words following this scheme and soon thought about
6741: factoring out the commonalities among them. Note that this uses a
6742: two-level defining word, i.e., a word that defines ordinary defining
6743: words.
6744:
6745: This time a solution involving @code{postpone} and friends seemed more
6746: difficult (try it as an exercise), so I decided to use a
6747: @code{create}/@code{does>} word; since I was already at it, I also used
6748: @code{create}/@code{does>} for the lower level (try using
6749: @code{postpone} etc. as an exercise), resulting in the following
6750: definition:
6751:
6752: @example
6753: : define-format ( disasm-xt table-xt -- )
6754: \ define an instruction format that uses disasm-xt for
6755: \ disassembling and enters the defined instructions into table
6756: \ table-xt
6757: create 2,
6758: does> ( u "inst" -- )
6759: \ defines an anonymous word for disassembling instruction inst,
6760: \ and enters it as u-th entry into table-xt
6761: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6762: noname create 2, \ define anonymous word
6763: execute lastxt swap ! \ enter xt of defined word into table-xt
6764: does> ( addr w -- )
6765: \ disassemble instruction w at addr
6766: 2@@ >r ( addr w disasm-xt R: c-addr )
6767: execute ( R: c-addr ) \ disassemble operands
6768: r> count type ; \ print name
6769: @end example
6770:
6771: Note that the tables here (in contrast to above) do the @code{cells +}
6772: by themselves (that's why you have to pass an xt). This word is used in
6773: the following way:
6774:
6775: @example
6776: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6777: @end example
6778:
1.71 anton 6779: As shown above, the defined instruction format is then used like this:
6780:
6781: @example
6782: @var{entry-num} @var{inst-format} @var{inst-name}
6783: @end example
6784:
1.63 anton 6785: In terms of currying, this kind of two-level defining word provides the
6786: parameters in three stages: first @var{disasm-operands} and @var{table},
6787: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6788: the instruction to be disassembled.
6789:
6790: Of course this did not quite fit all the instruction format names used
6791: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6792: the parameters into the right form.
6793:
6794: If you have trouble following this section, don't worry. First, this is
6795: involved and takes time (and probably some playing around) to
6796: understand; second, this is the first two-level
6797: @code{create}/@code{does>} word I have written in seventeen years of
6798: Forth; and if I did not have @file{insts.fs} to start with, I may well
6799: have elected to use just a one-level defining word (with some repeating
6800: of parameters when using the defining word). So it is not necessary to
6801: understand this, but it may improve your understanding of Forth.
1.44 crook 6802:
6803:
1.91 ! anton 6804: @node @code{Const-does>}, , Advanced does> usage example, User-defined Defining Words
! 6805: @subsubsection @code{Const-does>}
! 6806:
! 6807: A frequent use of @code{create}...@code{does>} is for transferring some
! 6808: values from definition-time to run-time. Gforth supports this use with
! 6809:
! 6810: doc-const-does>
! 6811:
! 6812: A typical use of this word is:
! 6813:
! 6814: @example
! 6815: : curry+ ( n1 "name" -- )
! 6816: 1 0 CONST-DOES> ( n2 -- n1+n2 )
! 6817: + ;
! 6818:
! 6819: 3 curry+ 3+
! 6820: @end example
! 6821:
! 6822: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
! 6823: definition to run-time.
! 6824:
! 6825: The advantages of using @code{const-does>} are:
! 6826:
! 6827: @itemize
! 6828:
! 6829: @item
! 6830: You don't have to deal with storing and retrieving the values, i.e.,
! 6831: your program becomes more writable and readable.
! 6832:
! 6833: @item
! 6834: When using @code{does>}, you have to introduce a @code{@@} that cannot
! 6835: be optimized away (because you could change the data using
! 6836: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
! 6837:
! 6838: @end itemize
! 6839:
! 6840: An ANS Forth implementation of @code{const-does>} is available in
! 6841: @file{compat/const-does.fs}.
! 6842:
! 6843:
1.44 crook 6844: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6845: @subsection Deferred words
6846: @cindex deferred words
6847:
6848: The defining word @code{Defer} allows you to define a word by name
6849: without defining its behaviour; the definition of its behaviour is
6850: deferred. Here are two situation where this can be useful:
6851:
6852: @itemize @bullet
6853: @item
6854: Where you want to allow the behaviour of a word to be altered later, and
6855: for all precompiled references to the word to change when its behaviour
6856: is changed.
6857: @item
6858: For mutual recursion; @xref{Calls and returns}.
6859: @end itemize
6860:
6861: In the following example, @code{foo} always invokes the version of
6862: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6863: always invokes the version that prints ``@code{Hello}''. There is no way
6864: of getting @code{foo} to use the later version without re-ordering the
6865: source code and recompiling it.
6866:
6867: @example
6868: : greet ." Good morning" ;
6869: : foo ... greet ... ;
6870: : greet ." Hello" ;
6871: : bar ... greet ... ;
6872: @end example
6873:
6874: This problem can be solved by defining @code{greet} as a @code{Defer}red
6875: word. The behaviour of a @code{Defer}red word can be defined and
6876: redefined at any time by using @code{IS} to associate the xt of a
6877: previously-defined word with it. The previous example becomes:
6878:
6879: @example
1.69 anton 6880: Defer greet ( -- )
1.44 crook 6881: : foo ... greet ... ;
6882: : bar ... greet ... ;
1.69 anton 6883: : greet1 ( -- ) ." Good morning" ;
6884: : greet2 ( -- ) ." Hello" ;
1.44 crook 6885: ' greet2 <IS> greet \ make greet behave like greet2
6886: @end example
6887:
1.69 anton 6888: @progstyle
6889: You should write a stack comment for every deferred word, and put only
6890: XTs into deferred words that conform to this stack effect. Otherwise
6891: it's too difficult to use the deferred word.
6892:
1.44 crook 6893: A deferred word can be used to improve the statistics-gathering example
6894: from @ref{User-defined Defining Words}; rather than edit the
6895: application's source code to change every @code{:} to a @code{my:}, do
6896: this:
6897:
6898: @example
6899: : real: : ; \ retain access to the original
6900: defer : \ redefine as a deferred word
1.69 anton 6901: ' my: <IS> : \ use special version of :
1.44 crook 6902: \
6903: \ load application here
6904: \
1.69 anton 6905: ' real: <IS> : \ go back to the original
1.44 crook 6906: @end example
6907:
6908:
6909: One thing to note is that @code{<IS>} consumes its name when it is
6910: executed. If you want to specify the name at compile time, use
6911: @code{[IS]}:
6912:
6913: @example
6914: : set-greet ( xt -- )
6915: [IS] greet ;
6916:
6917: ' greet1 set-greet
6918: @end example
6919:
1.69 anton 6920: A deferred word can only inherit execution semantics from the xt
6921: (because that is all that an xt can represent -- for more discussion of
6922: this @pxref{Tokens for Words}); by default it will have default
6923: interpretation and compilation semantics deriving from this execution
6924: semantics. However, you can change the interpretation and compilation
6925: semantics of the deferred word in the usual ways:
1.44 crook 6926:
6927: @example
6928: : bar .... ; compile-only
6929: Defer fred immediate
6930: Defer jim
6931:
6932: ' bar <IS> jim \ jim has default semantics
6933: ' bar <IS> fred \ fred is immediate
6934: @end example
6935:
6936: doc-defer
6937: doc-<is>
6938: doc-[is]
6939: doc-is
6940: @comment TODO document these: what's defers [is]
6941: doc-what's
6942: doc-defers
6943:
6944: @c Use @code{words-deferred} to see a list of deferred words.
6945:
6946: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6947: are provided in @file{compat/defer.fs}.
6948:
6949:
1.69 anton 6950: @node Aliases, , Deferred words, Defining Words
1.44 crook 6951: @subsection Aliases
6952: @cindex aliases
1.1 anton 6953:
1.44 crook 6954: The defining word @code{Alias} allows you to define a word by name that
6955: has the same behaviour as some other word. Here are two situation where
6956: this can be useful:
1.1 anton 6957:
1.44 crook 6958: @itemize @bullet
6959: @item
6960: When you want access to a word's definition from a different word list
6961: (for an example of this, see the definition of the @code{Root} word list
6962: in the Gforth source).
6963: @item
6964: When you want to create a synonym; a definition that can be known by
6965: either of two names (for example, @code{THEN} and @code{ENDIF} are
6966: aliases).
6967: @end itemize
1.1 anton 6968:
1.69 anton 6969: Like deferred words, an alias has default compilation and interpretation
6970: semantics at the beginning (not the modifications of the other word),
6971: but you can change them in the usual ways (@code{immediate},
6972: @code{compile-only}). For example:
1.1 anton 6973:
6974: @example
1.44 crook 6975: : foo ... ; immediate
6976:
6977: ' foo Alias bar \ bar is not an immediate word
6978: ' foo Alias fooby immediate \ fooby is an immediate word
1.1 anton 6979: @end example
6980:
1.44 crook 6981: Words that are aliases have the same xt, different headers in the
6982: dictionary, and consequently different name tokens (@pxref{Tokens for
6983: Words}) and possibly different immediate flags. An alias can only have
6984: default or immediate compilation semantics; you can define aliases for
6985: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1 anton 6986:
1.44 crook 6987: doc-alias
1.1 anton 6988:
6989:
1.47 crook 6990: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6991: @section Interpretation and Compilation Semantics
1.26 crook 6992: @cindex semantics, interpretation and compilation
1.1 anton 6993:
1.71 anton 6994: @c !! state and ' are used without explanation
6995: @c example for immediate/compile-only? or is the tutorial enough
6996:
1.26 crook 6997: @cindex interpretation semantics
1.71 anton 6998: The @dfn{interpretation semantics} of a (named) word are what the text
1.26 crook 6999: interpreter does when it encounters the word in interpret state. It also
7000: appears in some other contexts, e.g., the execution token returned by
1.71 anton 7001: @code{' @i{word}} identifies the interpretation semantics of @i{word}
7002: (in other words, @code{' @i{word} execute} is equivalent to
1.29 crook 7003: interpret-state text interpretation of @code{@i{word}}).
1.1 anton 7004:
1.26 crook 7005: @cindex compilation semantics
1.71 anton 7006: The @dfn{compilation semantics} of a (named) word are what the text
7007: interpreter does when it encounters the word in compile state. It also
7008: appears in other contexts, e.g, @code{POSTPONE @i{word}}
7009: compiles@footnote{In standard terminology, ``appends to the current
7010: definition''.} the compilation semantics of @i{word}.
1.1 anton 7011:
1.26 crook 7012: @cindex execution semantics
7013: The standard also talks about @dfn{execution semantics}. They are used
7014: only for defining the interpretation and compilation semantics of many
7015: words. By default, the interpretation semantics of a word are to
7016: @code{execute} its execution semantics, and the compilation semantics of
7017: a word are to @code{compile,} its execution semantics.@footnote{In
7018: standard terminology: The default interpretation semantics are its
7019: execution semantics; the default compilation semantics are to append its
7020: execution semantics to the execution semantics of the current
7021: definition.}
7022:
1.71 anton 7023: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
7024: the text interpreter, ticked, or @code{postpone}d, so they have no
7025: interpretation or compilation semantics. Their behaviour is represented
7026: by their XT (@pxref{Tokens for Words}), and we call it execution
7027: semantics, too.
7028:
1.26 crook 7029: @comment TODO expand, make it co-operate with new sections on text interpreter.
7030:
7031: @cindex immediate words
7032: @cindex compile-only words
7033: You can change the semantics of the most-recently defined word:
7034:
1.44 crook 7035:
1.26 crook 7036: doc-immediate
7037: doc-compile-only
7038: doc-restrict
7039:
1.82 anton 7040: By convention, words with non-default compilation semantics (e.g.,
7041: immediate words) often have names surrounded with brackets (e.g.,
7042: @code{[']}, @pxref{Execution token}).
1.44 crook 7043:
1.26 crook 7044: Note that ticking (@code{'}) a compile-only word gives an error
7045: (``Interpreting a compile-only word'').
1.1 anton 7046:
1.47 crook 7047: @menu
1.67 anton 7048: * Combined words::
1.47 crook 7049: @end menu
1.44 crook 7050:
1.71 anton 7051:
1.48 anton 7052: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44 crook 7053: @subsection Combined Words
7054: @cindex combined words
7055:
7056: Gforth allows you to define @dfn{combined words} -- words that have an
7057: arbitrary combination of interpretation and compilation semantics.
7058:
1.26 crook 7059: doc-interpret/compile:
1.1 anton 7060:
1.26 crook 7061: This feature was introduced for implementing @code{TO} and @code{S"}. I
7062: recommend that you do not define such words, as cute as they may be:
7063: they make it hard to get at both parts of the word in some contexts.
7064: E.g., assume you want to get an execution token for the compilation
7065: part. Instead, define two words, one that embodies the interpretation
7066: part, and one that embodies the compilation part. Once you have done
7067: that, you can define a combined word with @code{interpret/compile:} for
7068: the convenience of your users.
1.1 anton 7069:
1.26 crook 7070: You might try to use this feature to provide an optimizing
7071: implementation of the default compilation semantics of a word. For
7072: example, by defining:
1.1 anton 7073: @example
1.26 crook 7074: :noname
7075: foo bar ;
7076: :noname
7077: POSTPONE foo POSTPONE bar ;
1.29 crook 7078: interpret/compile: opti-foobar
1.1 anton 7079: @end example
1.26 crook 7080:
1.23 crook 7081: @noindent
1.26 crook 7082: as an optimizing version of:
7083:
1.1 anton 7084: @example
1.26 crook 7085: : foobar
7086: foo bar ;
1.1 anton 7087: @end example
7088:
1.26 crook 7089: Unfortunately, this does not work correctly with @code{[compile]},
7090: because @code{[compile]} assumes that the compilation semantics of all
7091: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29 crook 7092: opti-foobar} would compile compilation semantics, whereas
7093: @code{[compile] foobar} would compile interpretation semantics.
1.1 anton 7094:
1.26 crook 7095: @cindex state-smart words (are a bad idea)
1.82 anton 7096: @anchor{state-smartness}
1.29 crook 7097: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26 crook 7098: by @code{interpret/compile:} (words are state-smart if they check
7099: @code{STATE} during execution). E.g., they would try to code
7100: @code{foobar} like this:
1.1 anton 7101:
1.26 crook 7102: @example
7103: : foobar
7104: STATE @@
7105: IF ( compilation state )
7106: POSTPONE foo POSTPONE bar
7107: ELSE
7108: foo bar
7109: ENDIF ; immediate
7110: @end example
1.1 anton 7111:
1.26 crook 7112: Although this works if @code{foobar} is only processed by the text
7113: interpreter, it does not work in other contexts (like @code{'} or
7114: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
7115: for a state-smart word, not for the interpretation semantics of the
7116: original @code{foobar}; when you execute this execution token (directly
7117: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
7118: state, the result will not be what you expected (i.e., it will not
7119: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
7120: write them@footnote{For a more detailed discussion of this topic, see
1.66 anton 7121: M. Anton Ertl,
7122: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
7123: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1 anton 7124:
1.26 crook 7125: @cindex defining words with arbitrary semantics combinations
7126: It is also possible to write defining words that define words with
7127: arbitrary combinations of interpretation and compilation semantics. In
7128: general, they look like this:
1.1 anton 7129:
1.26 crook 7130: @example
7131: : def-word
7132: create-interpret/compile
1.29 crook 7133: @i{code1}
1.26 crook 7134: interpretation>
1.29 crook 7135: @i{code2}
1.26 crook 7136: <interpretation
7137: compilation>
1.29 crook 7138: @i{code3}
1.26 crook 7139: <compilation ;
7140: @end example
1.1 anton 7141:
1.29 crook 7142: For a @i{word} defined with @code{def-word}, the interpretation
7143: semantics are to push the address of the body of @i{word} and perform
7144: @i{code2}, and the compilation semantics are to push the address of
7145: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26 crook 7146: can also be defined like this (except that the defined constants don't
7147: behave correctly when @code{[compile]}d):
1.1 anton 7148:
1.26 crook 7149: @example
7150: : constant ( n "name" -- )
7151: create-interpret/compile
7152: ,
7153: interpretation> ( -- n )
7154: @@
7155: <interpretation
7156: compilation> ( compilation. -- ; run-time. -- n )
7157: @@ postpone literal
7158: <compilation ;
7159: @end example
1.1 anton 7160:
1.44 crook 7161:
1.26 crook 7162: doc-create-interpret/compile
7163: doc-interpretation>
7164: doc-<interpretation
7165: doc-compilation>
7166: doc-<compilation
1.1 anton 7167:
1.44 crook 7168:
1.29 crook 7169: Words defined with @code{interpret/compile:} and
1.26 crook 7170: @code{create-interpret/compile} have an extended header structure that
7171: differs from other words; however, unless you try to access them with
7172: plain address arithmetic, you should not notice this. Words for
7173: accessing the header structure usually know how to deal with this; e.g.,
1.29 crook 7174: @code{'} @i{word} @code{>body} also gives you the body of a word created
7175: with @code{create-interpret/compile}.
1.1 anton 7176:
1.44 crook 7177:
1.47 crook 7178: @c -------------------------------------------------------------
1.81 anton 7179: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
1.47 crook 7180: @section Tokens for Words
7181: @cindex tokens for words
7182:
7183: This section describes the creation and use of tokens that represent
7184: words.
7185:
1.71 anton 7186: @menu
7187: * Execution token:: represents execution/interpretation semantics
7188: * Compilation token:: represents compilation semantics
7189: * Name token:: represents named words
7190: @end menu
1.47 crook 7191:
1.71 anton 7192: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
7193: @subsection Execution token
1.47 crook 7194:
7195: @cindex xt
7196: @cindex execution token
1.71 anton 7197: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
7198: You can use @code{execute} to invoke this behaviour.
1.47 crook 7199:
1.71 anton 7200: @cindex tick (')
7201: You can use @code{'} to get an execution token that represents the
7202: interpretation semantics of a named word:
1.47 crook 7203:
7204: @example
1.71 anton 7205: 5 ' .
7206: execute
7207: @end example
1.47 crook 7208:
1.71 anton 7209: doc-'
7210:
7211: @code{'} parses at run-time; there is also a word @code{[']} that parses
7212: when it is compiled, and compiles the resulting XT:
7213:
7214: @example
7215: : foo ['] . execute ;
7216: 5 foo
7217: : bar ' execute ; \ by contrast,
7218: 5 bar . \ ' parses "." when bar executes
7219: @end example
7220:
7221: doc-[']
7222:
7223: If you want the execution token of @i{word}, write @code{['] @i{word}}
7224: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7225: @code{'} and @code{[']} behave somewhat unusually by complaining about
7226: compile-only words (because these words have no interpretation
7227: semantics). You might get what you want by using @code{COMP' @i{word}
7228: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7229: token}).
7230:
7231: Another way to get an XT is @code{:noname} or @code{lastxt}
7232: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7233: for the only behaviour the word has (the execution semantics). For
7234: named words, @code{lastxt} produces an XT for the same behaviour it
7235: would produce if the word was defined anonymously.
7236:
7237: @example
7238: :noname ." hello" ;
7239: execute
1.47 crook 7240: @end example
7241:
1.71 anton 7242: An XT occupies one cell and can be manipulated like any other cell.
7243:
1.47 crook 7244: @cindex code field address
7245: @cindex CFA
1.71 anton 7246: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7247: operations that produce or consume it). For old hands: In Gforth, the
7248: XT is implemented as a code field address (CFA).
7249:
7250: doc-execute
7251: doc-perform
7252:
7253: @node Compilation token, Name token, Execution token, Tokens for Words
7254: @subsection Compilation token
1.47 crook 7255:
7256: @cindex compilation token
1.71 anton 7257: @cindex CT (compilation token)
7258: Gforth represents the compilation semantics of a named word by a
1.47 crook 7259: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7260: @i{xt} is an execution token. The compilation semantics represented by
7261: the compilation token can be performed with @code{execute}, which
7262: consumes the whole compilation token, with an additional stack effect
7263: determined by the represented compilation semantics.
7264:
7265: At present, the @i{w} part of a compilation token is an execution token,
7266: and the @i{xt} part represents either @code{execute} or
7267: @code{compile,}@footnote{Depending upon the compilation semantics of the
7268: word. If the word has default compilation semantics, the @i{xt} will
7269: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7270: @i{xt} will represent @code{execute}.}. However, don't rely on that
7271: knowledge, unless necessary; future versions of Gforth may introduce
7272: unusual compilation tokens (e.g., a compilation token that represents
7273: the compilation semantics of a literal).
7274:
1.71 anton 7275: You can perform the compilation semantics represented by the compilation
7276: token with @code{execute}. You can compile the compilation semantics
7277: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7278: equivalent to @code{postpone @i{word}}.
7279:
7280: doc-[comp']
7281: doc-comp'
7282: doc-postpone,
7283:
7284: @node Name token, , Compilation token, Tokens for Words
7285: @subsection Name token
1.47 crook 7286:
7287: @cindex name token
7288: @cindex name field address
7289: @cindex NFA
1.71 anton 7290: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
1.47 crook 7291: Gforth, the abstract data type @emph{name token} is implemented as a
7292: name field address (NFA).
7293:
7294: doc-find-name
7295: doc-name>int
7296: doc-name?int
7297: doc-name>comp
7298: doc-name>string
7299:
1.81 anton 7300: @c ----------------------------------------------------------
7301: @node Compiling words, The Text Interpreter, Tokens for Words, Words
7302: @section Compiling words
7303: @cindex compiling words
7304: @cindex macros
7305:
7306: In contrast to most other languages, Forth has no strict boundary
1.82 anton 7307: between compilation and run-time. E.g., you can run arbitrary code
7308: between defining words (or for computing data used by defining words
7309: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
7310: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
7311: running arbitrary code while compiling a colon definition (exception:
7312: you must not allot dictionary space).
7313:
7314: @menu
7315: * Literals:: Compiling data values
7316: * Macros:: Compiling words
7317: @end menu
7318:
7319: @node Literals, Macros, Compiling words, Compiling words
7320: @subsection Literals
7321: @cindex Literals
7322:
7323: The simplest and most frequent example is to compute a literal during
7324: compilation. E.g., the following definition prints an array of strings,
7325: one string per line:
7326:
7327: @example
7328: : .strings ( addr u -- ) \ gforth
7329: 2* cells bounds U+DO
7330: cr i 2@@ type
7331: 2 cells +LOOP ;
7332: @end example
1.81 anton 7333:
1.82 anton 7334: With a simple-minded compiler like Gforth's, this computes @code{2
7335: cells} on every loop iteration. You can compute this value once and for
7336: all at compile time and compile it into the definition like this:
7337:
7338: @example
7339: : .strings ( addr u -- ) \ gforth
7340: 2* cells bounds U+DO
7341: cr i 2@@ type
7342: [ 2 cells ] literal +LOOP ;
7343: @end example
7344:
7345: @code{[} switches the text interpreter to interpret state (you will get
7346: an @code{ok} prompt if you type this example interactively and insert a
7347: newline between @code{[} and @code{]}), so it performs the
7348: interpretation semantics of @code{2 cells}; this computes a number.
7349: @code{]} switches the text interpreter back into compile state. It then
7350: performs @code{Literal}'s compilation semantics, which are to compile
7351: this number into the current word. You can decompile the word with
7352: @code{see .strings} to see the effect on the compiled code.
1.81 anton 7353:
1.82 anton 7354: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
7355: *} in this way.
1.81 anton 7356:
1.82 anton 7357: doc-[
7358: doc-]
1.81 anton 7359: doc-literal
7360: doc-]L
1.82 anton 7361:
7362: There are also words for compiling other data types than single cells as
7363: literals:
7364:
1.81 anton 7365: doc-2literal
7366: doc-fliteral
1.82 anton 7367: doc-sliteral
7368:
7369: @cindex colon-sys, passing data across @code{:}
7370: @cindex @code{:}, passing data across
7371: You might be tempted to pass data from outside a colon definition to the
7372: inside on the data stack. This does not work, because @code{:} puhes a
7373: colon-sys, making stuff below unaccessible. E.g., this does not work:
7374:
7375: @example
7376: 5 : foo literal ; \ error: "unstructured"
7377: @end example
7378:
7379: Instead, you have to pass the value in some other way, e.g., through a
7380: variable:
7381:
7382: @example
7383: variable temp
7384: 5 temp !
7385: : foo [ temp @@ ] literal ;
7386: @end example
7387:
7388:
7389: @node Macros, , Literals, Compiling words
7390: @subsection Macros
7391: @cindex Macros
7392: @cindex compiling compilation semantics
7393:
7394: @code{Literal} and friends compile data values into the current
7395: definition. You can also write words that compile other words into the
7396: current definition. E.g.,
7397:
7398: @example
7399: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
7400: POSTPONE + ;
7401:
7402: : foo ( n1 n2 -- n )
7403: [ compile-+ ] ;
7404: 1 2 foo .
7405: @end example
7406:
7407: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
7408: What happens in this example? @code{Postpone} compiles the compilation
7409: semantics of @code{+} into @code{compile-+}; later the text interpreter
7410: executes @code{compile-+} and thus the compilation semantics of +, which
7411: compile (the execution semantics of) @code{+} into
7412: @code{foo}.@footnote{A recent RFI answer requires that compiling words
7413: should only be executed in compile state, so this example is not
7414: guaranteed to work on all standard systems, but on any decent system it
7415: will work.}
7416:
7417: doc-postpone
7418: doc-[compile]
7419:
7420: Compiling words like @code{compile-+} are usually immediate (or similar)
7421: so you do not have to switch to interpret state to execute them;
7422: mopifying the last example accordingly produces:
7423:
7424: @example
7425: : [compile-+] ( compilation: --; interpretation: -- )
7426: \ compiled code: ( n1 n2 -- n )
7427: POSTPONE + ; immediate
7428:
7429: : foo ( n1 n2 -- n )
7430: [compile-+] ;
7431: 1 2 foo .
7432: @end example
7433:
7434: Immediate compiling words are similar to macros in other languages (in
7435: particular, Lisp). The important differences to macros in, e.g., C are:
7436:
7437: @itemize @bullet
7438:
7439: @item
7440: You use the same language for defining and processing macros, not a
7441: separate preprocessing language and processor.
7442:
7443: @item
7444: Consequently, the full power of Forth is available in macro definitions.
7445: E.g., you can perform arbitrarily complex computations, or generate
7446: different code conditionally or in a loop (e.g., @pxref{Advanced macros
7447: Tutorial}). This power is very useful when writing a parser generators
7448: or other code-generating software.
7449:
7450: @item
7451: Macros defined using @code{postpone} etc. deal with the language at a
7452: higher level than strings; name binding happens at macro definition
7453: time, so you can avoid the pitfalls of name collisions that can happen
7454: in C macros. Of course, Forth is a liberal language and also allows to
7455: shoot yourself in the foot with text-interpreted macros like
7456:
7457: @example
7458: : [compile-+] s" +" evaluate ; immediate
7459: @end example
7460:
7461: Apart from binding the name at macro use time, using @code{evaluate}
7462: also makes your definition @code{state}-smart (@pxref{state-smartness}).
7463: @end itemize
7464:
7465: You may want the macro to compile a number into a word. The word to do
7466: it is @code{literal}, but you have to @code{postpone} it, so its
7467: compilation semantics take effect when the macro is executed, not when
7468: it is compiled:
7469:
7470: @example
7471: : [compile-5] ( -- ) \ compiled code: ( -- n )
7472: 5 POSTPONE literal ; immediate
7473:
7474: : foo [compile-5] ;
7475: foo .
7476: @end example
7477:
7478: You may want to pass parameters to a macro, that the macro should
7479: compile into the current definition. If the parameter is a number, then
7480: you can use @code{postpone literal} (similar for other values).
7481:
7482: If you want to pass a word that is to be compiled, the usual way is to
7483: pass an execution token and @code{compile,} it:
7484:
7485: @example
7486: : twice1 ( xt -- ) \ compiled code: ... -- ...
7487: dup compile, compile, ;
7488:
7489: : 2+ ( n1 -- n2 )
7490: [ ' 1+ twice1 ] ;
7491: @end example
7492:
7493: doc-compile,
7494:
7495: An alternative available in Gforth, that allows you to pass compile-only
7496: words as parameters is to use the compilation token (@pxref{Compilation
7497: token}). The same example in this technique:
7498:
7499: @example
7500: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
7501: 2dup 2>r execute 2r> execute ;
7502:
7503: : 2+ ( n1 -- n2 )
7504: [ comp' 1+ twice ] ;
7505: @end example
7506:
7507: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
7508: works even if the executed compilation semantics has an effect on the
7509: data stack.
7510:
7511: You can also define complete definitions with these words; this provides
7512: an alternative to using @code{does>} (@pxref{User-defined Defining
7513: Words}). E.g., instead of
7514:
7515: @example
7516: : curry+ ( n1 "name" -- )
7517: CREATE ,
7518: DOES> ( n2 -- n1+n2 )
7519: @@ + ;
7520: @end example
7521:
7522: you could define
7523:
7524: @example
7525: : curry+ ( n1 "name" -- )
7526: \ name execution: ( n2 -- n1+n2 )
7527: >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
1.81 anton 7528:
1.82 anton 7529: -3 curry+ 3-
7530: see 3-
7531: @end example
1.81 anton 7532:
1.82 anton 7533: The sequence @code{>r : r>} is necessary, because @code{:} puts a
7534: colon-sys on the data stack that makes everything below it unaccessible.
1.81 anton 7535:
1.82 anton 7536: This way of writing defining words is sometimes more, sometimes less
7537: convenient than using @code{does>} (@pxref{Advanced does> usage
7538: example}). One advantage of this method is that it can be optimized
7539: better, because the compiler knows that the value compiled with
7540: @code{literal} is fixed, whereas the data associated with a
7541: @code{create}d word can be changed.
1.47 crook 7542:
1.26 crook 7543: @c ----------------------------------------------------------
1.81 anton 7544: @node The Text Interpreter, Word Lists, Compiling words, Words
1.26 crook 7545: @section The Text Interpreter
7546: @cindex interpreter - outer
7547: @cindex text interpreter
7548: @cindex outer interpreter
1.1 anton 7549:
1.34 anton 7550: @c Should we really describe all these ugly details? IMO the text
7551: @c interpreter should be much cleaner, but that may not be possible within
7552: @c ANS Forth. - anton
1.44 crook 7553: @c nac-> I wanted to explain how it works to show how you can exploit
7554: @c it in your own programs. When I was writing a cross-compiler, figuring out
7555: @c some of these gory details was very helpful to me. None of the textbooks
7556: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7557: @c seems to positively avoid going into too much detail for some of
7558: @c the internals.
1.34 anton 7559:
1.71 anton 7560: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7561: @c it is; for the ugly details, I would prefer another place. I wonder
7562: @c whether we should have a chapter before "Words" that describes some
7563: @c basic concepts referred to in words, and a chapter after "Words" that
7564: @c describes implementation details.
7565:
1.29 crook 7566: The text interpreter@footnote{This is an expanded version of the
7567: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34 anton 7568: that processes input from the current input device. It is also called
7569: the outer interpreter, in contrast to the inner interpreter
7570: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7571: implementations.
1.27 crook 7572:
1.29 crook 7573: @cindex interpret state
7574: @cindex compile state
7575: The text interpreter operates in one of two states: @dfn{interpret
7576: state} and @dfn{compile state}. The current state is defined by the
1.71 anton 7577: aptly-named variable @code{state}.
1.29 crook 7578:
7579: This section starts by describing how the text interpreter behaves when
7580: it is in interpret state, processing input from the user input device --
7581: the keyboard. This is the mode that a Forth system is in after it starts
7582: up.
7583:
7584: @cindex input buffer
7585: @cindex terminal input buffer
7586: The text interpreter works from an area of memory called the @dfn{input
7587: buffer}@footnote{When the text interpreter is processing input from the
7588: keyboard, this area of memory is called the @dfn{terminal input buffer}
7589: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7590: @code{#TIB}.}, which stores your keyboard input when you press the
1.30 anton 7591: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29 crook 7592: leading spaces (called @dfn{delimiters}) then parses a string (a
7593: sequence of non-space characters) until it reaches either a space
7594: character or the end of the buffer. Having parsed a string, it makes two
7595: attempts to process it:
1.27 crook 7596:
1.29 crook 7597: @cindex dictionary
1.27 crook 7598: @itemize @bullet
7599: @item
1.29 crook 7600: It looks for the string in a @dfn{dictionary} of definitions. If the
7601: string is found, the string names a @dfn{definition} (also known as a
7602: @dfn{word}) and the dictionary search returns information that allows
7603: the text interpreter to perform the word's @dfn{interpretation
7604: semantics}. In most cases, this simply means that the word will be
7605: executed.
1.27 crook 7606: @item
7607: If the string is not found in the dictionary, the text interpreter
1.29 crook 7608: attempts to treat it as a number, using the rules described in
7609: @ref{Number Conversion}. If the string represents a legal number in the
7610: current radix, the number is pushed onto a parameter stack (the data
7611: stack for integers, the floating-point stack for floating-point
7612: numbers).
7613: @end itemize
7614:
7615: If both attempts fail, or if the word is found in the dictionary but has
7616: no interpretation semantics@footnote{This happens if the word was
7617: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7618: remainder of the input buffer, issues an error message and waits for
7619: more input. If one of the attempts succeeds, the text interpreter
7620: repeats the parsing process until the whole of the input buffer has been
7621: processed, at which point it prints the status message ``@code{ ok}''
7622: and waits for more input.
7623:
1.71 anton 7624: @c anton: this should be in the input stream subsection (or below it)
7625:
1.29 crook 7626: @cindex parse area
7627: The text interpreter keeps track of its position in the input buffer by
7628: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7629: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7630: of the input buffer. The region from offset @code{>IN @@} to the end of
7631: the input buffer is called the @dfn{parse area}@footnote{In other words,
7632: the text interpreter processes the contents of the input buffer by
7633: parsing strings from the parse area until the parse area is empty.}.
7634: This example shows how @code{>IN} changes as the text interpreter parses
7635: the input buffer:
7636:
7637: @example
7638: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7639: CR ." ->" TYPE ." <-" ; IMMEDIATE
7640:
7641: 1 2 3 remaining + remaining .
7642:
7643: : foo 1 2 3 remaining SWAP remaining ;
7644: @end example
7645:
7646: @noindent
7647: The result is:
7648:
7649: @example
7650: ->+ remaining .<-
7651: ->.<-5 ok
7652:
7653: ->SWAP remaining ;-<
7654: ->;<- ok
7655: @end example
7656:
7657: @cindex parsing words
7658: The value of @code{>IN} can also be modified by a word in the input
7659: buffer that is executed by the text interpreter. This means that a word
7660: can ``trick'' the text interpreter into either skipping a section of the
7661: input buffer@footnote{This is how parsing words work.} or into parsing a
7662: section twice. For example:
1.27 crook 7663:
1.29 crook 7664: @example
1.71 anton 7665: : lat ." <<foo>>" ;
7666: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29 crook 7667: @end example
7668:
7669: @noindent
7670: When @code{flat} is executed, this output is produced@footnote{Exercise
7671: for the reader: what would happen if the @code{3} were replaced with
7672: @code{4}?}:
7673:
7674: @example
1.71 anton 7675: <<bar>><<foo>>
1.29 crook 7676: @end example
7677:
1.71 anton 7678: This technique can be used to work around some of the interoperability
7679: problems of parsing words. Of course, it's better to avoid parsing
7680: words where possible.
7681:
1.29 crook 7682: @noindent
7683: Two important notes about the behaviour of the text interpreter:
1.27 crook 7684:
7685: @itemize @bullet
7686: @item
7687: It processes each input string to completion before parsing additional
1.29 crook 7688: characters from the input buffer.
7689: @item
7690: It treats the input buffer as a read-only region (and so must your code).
7691: @end itemize
7692:
7693: @noindent
7694: When the text interpreter is in compile state, its behaviour changes in
7695: these ways:
7696:
7697: @itemize @bullet
7698: @item
7699: If a parsed string is found in the dictionary, the text interpreter will
7700: perform the word's @dfn{compilation semantics}. In most cases, this
7701: simply means that the execution semantics of the word will be appended
7702: to the current definition.
1.27 crook 7703: @item
1.29 crook 7704: When a number is encountered, it is compiled into the current definition
7705: (as a literal) rather than being pushed onto a parameter stack.
7706: @item
7707: If an error occurs, @code{state} is modified to put the text interpreter
7708: back into interpret state.
7709: @item
7710: Each time a line is entered from the keyboard, Gforth prints
7711: ``@code{ compiled}'' rather than `` @code{ok}''.
7712: @end itemize
7713:
7714: @cindex text interpreter - input sources
7715: When the text interpreter is using an input device other than the
7716: keyboard, its behaviour changes in these ways:
7717:
7718: @itemize @bullet
7719: @item
7720: When the parse area is empty, the text interpreter attempts to refill
7721: the input buffer from the input source. When the input source is
1.71 anton 7722: exhausted, the input source is set back to the previous input source.
1.29 crook 7723: @item
7724: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7725: time the parse area is emptied.
7726: @item
7727: If an error occurs, the input source is set back to the user input
7728: device.
1.27 crook 7729: @end itemize
1.21 crook 7730:
1.49 anton 7731: You can read about this in more detail in @ref{Input Sources}.
1.44 crook 7732:
1.26 crook 7733: doc->in
1.27 crook 7734: doc-source
7735:
1.26 crook 7736: doc-tib
7737: doc-#tib
1.1 anton 7738:
1.44 crook 7739:
1.26 crook 7740: @menu
1.67 anton 7741: * Input Sources::
7742: * Number Conversion::
7743: * Interpret/Compile states::
7744: * Interpreter Directives::
1.26 crook 7745: @end menu
1.1 anton 7746:
1.29 crook 7747: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7748: @subsection Input Sources
7749: @cindex input sources
7750: @cindex text interpreter - input sources
7751:
1.44 crook 7752: By default, the text interpreter processes input from the user input
1.29 crook 7753: device (the keyboard) when Forth starts up. The text interpreter can
7754: process input from any of these sources:
7755:
7756: @itemize @bullet
7757: @item
7758: The user input device -- the keyboard.
7759: @item
7760: A file, using the words described in @ref{Forth source files}.
7761: @item
7762: A block, using the words described in @ref{Blocks}.
7763: @item
7764: A text string, using @code{evaluate}.
7765: @end itemize
7766:
7767: A program can identify the current input device from the values of
7768: @code{source-id} and @code{blk}.
7769:
1.44 crook 7770:
1.29 crook 7771: doc-source-id
7772: doc-blk
7773:
7774: doc-save-input
7775: doc-restore-input
7776:
7777: doc-evaluate
1.1 anton 7778:
1.29 crook 7779:
1.44 crook 7780:
1.29 crook 7781: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26 crook 7782: @subsection Number Conversion
7783: @cindex number conversion
7784: @cindex double-cell numbers, input format
7785: @cindex input format for double-cell numbers
7786: @cindex single-cell numbers, input format
7787: @cindex input format for single-cell numbers
7788: @cindex floating-point numbers, input format
7789: @cindex input format for floating-point numbers
1.1 anton 7790:
1.29 crook 7791: This section describes the rules that the text interpreter uses when it
7792: tries to convert a string into a number.
1.1 anton 7793:
1.26 crook 7794: Let <digit> represent any character that is a legal digit in the current
1.29 crook 7795: number base@footnote{For example, 0-9 when the number base is decimal or
7796: 0-9, A-F when the number base is hexadecimal.}.
1.1 anton 7797:
1.26 crook 7798: Let <decimal digit> represent any character in the range 0-9.
1.1 anton 7799:
1.29 crook 7800: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7801: in the braces (@i{a} or @i{b} or neither).
1.1 anton 7802:
1.26 crook 7803: Let * represent any number of instances of the previous character
7804: (including none).
1.1 anton 7805:
1.26 crook 7806: Let any other character represent itself.
1.1 anton 7807:
1.29 crook 7808: @noindent
1.26 crook 7809: Now, the conversion rules are:
1.21 crook 7810:
1.26 crook 7811: @itemize @bullet
7812: @item
7813: A string of the form <digit><digit>* is treated as a single-precision
1.29 crook 7814: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26 crook 7815: @item
7816: A string of the form -<digit><digit>* is treated as a single-precision
1.29 crook 7817: (cell-sized) negative integer, and is represented using 2's-complement
1.26 crook 7818: arithmetic. Examples are -45 -5681 -0
7819: @item
7820: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29 crook 7821: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7822: (all three of these represent the same number).
1.26 crook 7823: @item
7824: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29 crook 7825: double-precision (double-cell-sized) negative integer, and is
1.26 crook 7826: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29 crook 7827: -34.65 (all three of these represent the same number).
1.26 crook 7828: @item
1.29 crook 7829: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7830: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35 anton 7831: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29 crook 7832: number) +12.E-4
1.26 crook 7833: @end itemize
1.1 anton 7834:
1.26 crook 7835: By default, the number base used for integer number conversion is given
1.35 anton 7836: by the contents of the variable @code{base}. Note that a lot of
7837: confusion can result from unexpected values of @code{base}. If you
7838: change @code{base} anywhere, make sure to save the old value and restore
7839: it afterwards. In general I recommend keeping @code{base} decimal, and
7840: using the prefixes described below for the popular non-decimal bases.
1.1 anton 7841:
1.29 crook 7842: doc-dpl
1.26 crook 7843: doc-base
7844: doc-hex
7845: doc-decimal
1.1 anton 7846:
1.44 crook 7847:
1.26 crook 7848: @cindex '-prefix for character strings
7849: @cindex &-prefix for decimal numbers
7850: @cindex %-prefix for binary numbers
7851: @cindex $-prefix for hexadecimal numbers
1.35 anton 7852: Gforth allows you to override the value of @code{base} by using a
1.29 crook 7853: prefix@footnote{Some Forth implementations provide a similar scheme by
7854: implementing @code{$} etc. as parsing words that process the subsequent
7855: number in the input stream and push it onto the stack. For example, see
7856: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7857: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7858: is required between the prefix and the number.} before the first digit
7859: of an (integer) number. Four prefixes are supported:
1.1 anton 7860:
1.26 crook 7861: @itemize @bullet
7862: @item
1.35 anton 7863: @code{&} -- decimal
1.26 crook 7864: @item
1.35 anton 7865: @code{%} -- binary
1.26 crook 7866: @item
1.35 anton 7867: @code{$} -- hexadecimal
1.26 crook 7868: @item
1.35 anton 7869: @code{'} -- base @code{max-char+1}
1.26 crook 7870: @end itemize
1.1 anton 7871:
1.26 crook 7872: Here are some examples, with the equivalent decimal number shown after
7873: in braces:
1.1 anton 7874:
1.26 crook 7875: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7876: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7877: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7878: &905 (905), $abc (2478), $ABC (2478).
1.1 anton 7879:
1.26 crook 7880: @cindex number conversion - traps for the unwary
1.29 crook 7881: @noindent
1.26 crook 7882: Number conversion has a number of traps for the unwary:
1.1 anton 7883:
1.26 crook 7884: @itemize @bullet
7885: @item
7886: You cannot determine the current number base using the code sequence
1.35 anton 7887: @code{base @@ .} -- the number base is always 10 in the current number
7888: base. Instead, use something like @code{base @@ dec.}
1.26 crook 7889: @item
7890: If the number base is set to a value greater than 14 (for example,
7891: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7892: it to be intepreted as either a single-precision integer or a
7893: floating-point number (Gforth treats it as an integer). The ambiguity
7894: can be resolved by explicitly stating the sign of the mantissa and/or
7895: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7896: ambiguity arises; either representation will be treated as a
7897: floating-point number.
7898: @item
1.29 crook 7899: There is a word @code{bin} but it does @i{not} set the number base!
1.26 crook 7900: It is used to specify file types.
7901: @item
1.72 anton 7902: ANS Forth requires the @code{.} of a double-precision number to be the
7903: final character in the string. Gforth allows the @code{.} to be
7904: anywhere after the first digit.
1.26 crook 7905: @item
7906: The number conversion process does not check for overflow.
7907: @item
1.72 anton 7908: In an ANS Forth program @code{base} is required to be decimal when
7909: converting floating-point numbers. In Gforth, number conversion to
7910: floating-point numbers always uses base &10, irrespective of the value
7911: of @code{base}.
1.26 crook 7912: @end itemize
1.1 anton 7913:
1.49 anton 7914: You can read numbers into your programs with the words described in
7915: @ref{Input}.
1.1 anton 7916:
1.82 anton 7917: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
1.26 crook 7918: @subsection Interpret/Compile states
7919: @cindex Interpret/Compile states
1.1 anton 7920:
1.29 crook 7921: A standard program is not permitted to change @code{state}
7922: explicitly. However, it can change @code{state} implicitly, using the
7923: words @code{[} and @code{]}. When @code{[} is executed it switches
7924: @code{state} to interpret state, and therefore the text interpreter
7925: starts interpreting. When @code{]} is executed it switches @code{state}
7926: to compile state and therefore the text interpreter starts
1.44 crook 7927: compiling. The most common usage for these words is for switching into
7928: interpret state and back from within a colon definition; this technique
1.49 anton 7929: can be used to compile a literal (for an example, @pxref{Literals}) or
7930: for conditional compilation (for an example, @pxref{Interpreter
7931: Directives}).
1.44 crook 7932:
1.35 anton 7933:
7934: @c This is a bad example: It's non-standard, and it's not necessary.
7935: @c However, I can't think of a good example for switching into compile
7936: @c state when there is no current word (@code{state}-smart words are not a
7937: @c good reason). So maybe we should use an example for switching into
7938: @c interpret @code{state} in a colon def. - anton
1.44 crook 7939: @c nac-> I agree. I started out by putting in the example, then realised
7940: @c that it was non-ANS, so wrote more words around it. I hope this
7941: @c re-written version is acceptable to you. I do want to keep the example
7942: @c as it is helpful for showing what is and what is not portable, particularly
7943: @c where it outlaws a style in common use.
7944:
1.72 anton 7945: @c anton: it's more important to show what's portable. After we have done
1.83 anton 7946: @c that, we can also show what's not. In any case, I have written a
7947: @c section Compiling Words which also deals with [ ].
1.35 anton 7948:
1.44 crook 7949: @code{[} and @code{]} also give you the ability to switch into compile
7950: state and back, but we cannot think of any useful Standard application
7951: for this ability. Pre-ANS Forth textbooks have examples like this:
1.29 crook 7952:
7953: @example
7954: : AA ." this is A" ;
7955: : BB ." this is B" ;
7956: : CC ." this is C" ;
7957:
1.44 crook 7958: create table ] aa bb cc [
7959:
1.29 crook 7960: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7961: cells table + @ execute ;
7962: @end example
7963:
1.44 crook 7964: This example builds a jump table; @code{0 go} will display ``@code{this
7965: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7966: defining @code{table} like this:
1.29 crook 7967:
7968: @example
1.44 crook 7969: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
1.29 crook 7970: @end example
7971:
1.44 crook 7972: The problem with this code is that the definition of @code{table} is not
7973: portable -- it @i{compile}s execution tokens into code space. Whilst it
7974: @i{may} work on systems where code space and data space co-incide, the
1.29 crook 7975: Standard only allows data space to be assigned for a @code{CREATE}d
7976: word. In addition, the Standard only allows @code{@@} to access data
7977: space, whilst this example is using it to access code space. The only
7978: portable, Standard way to build this table is to build it in data space,
7979: like this:
7980:
7981: @example
7982: create table ' aa , ' bb , ' cc ,
7983: @end example
7984:
1.26 crook 7985: doc-state
1.44 crook 7986:
1.29 crook 7987:
1.82 anton 7988: @node Interpreter Directives, , Interpret/Compile states, The Text Interpreter
1.26 crook 7989: @subsection Interpreter Directives
7990: @cindex interpreter directives
1.72 anton 7991: @cindex conditional compilation
1.1 anton 7992:
1.29 crook 7993: These words are usually used in interpret state; typically to control
7994: which parts of a source file are processed by the text
1.26 crook 7995: interpreter. There are only a few ANS Forth Standard words, but Gforth
7996: supplements these with a rich set of immediate control structure words
7997: to compensate for the fact that the non-immediate versions can only be
1.29 crook 7998: used in compile state (@pxref{Control Structures}). Typical usages:
7999:
8000: @example
1.72 anton 8001: FALSE Constant HAVE-ASSEMBLER
1.29 crook 8002: .
8003: .
1.72 anton 8004: HAVE-ASSEMBLER [IF]
1.29 crook 8005: : ASSEMBLER-FEATURE
8006: ...
8007: ;
8008: [ENDIF]
8009: .
8010: .
8011: : SEE
8012: ... \ general-purpose SEE code
1.72 anton 8013: [ HAVE-ASSEMBLER [IF] ]
1.29 crook 8014: ... \ assembler-specific SEE code
8015: [ [ENDIF] ]
8016: ;
8017: @end example
1.1 anton 8018:
1.44 crook 8019:
1.26 crook 8020: doc-[IF]
8021: doc-[ELSE]
8022: doc-[THEN]
8023: doc-[ENDIF]
1.1 anton 8024:
1.26 crook 8025: doc-[IFDEF]
8026: doc-[IFUNDEF]
1.1 anton 8027:
1.26 crook 8028: doc-[?DO]
8029: doc-[DO]
8030: doc-[FOR]
8031: doc-[LOOP]
8032: doc-[+LOOP]
8033: doc-[NEXT]
1.1 anton 8034:
1.26 crook 8035: doc-[BEGIN]
8036: doc-[UNTIL]
8037: doc-[AGAIN]
8038: doc-[WHILE]
8039: doc-[REPEAT]
1.1 anton 8040:
1.27 crook 8041:
1.26 crook 8042: @c -------------------------------------------------------------
1.47 crook 8043: @node Word Lists, Environmental Queries, The Text Interpreter, Words
1.26 crook 8044: @section Word Lists
8045: @cindex word lists
1.32 anton 8046: @cindex header space
1.1 anton 8047:
1.36 anton 8048: A wordlist is a list of named words; you can add new words and look up
8049: words by name (and you can remove words in a restricted way with
8050: markers). Every named (and @code{reveal}ed) word is in one wordlist.
8051:
8052: @cindex search order stack
8053: The text interpreter searches the wordlists present in the search order
8054: (a stack of wordlists), from the top to the bottom. Within each
8055: wordlist, the search starts conceptually at the newest word; i.e., if
8056: two words in a wordlist have the same name, the newer word is found.
1.1 anton 8057:
1.26 crook 8058: @cindex compilation word list
1.36 anton 8059: New words are added to the @dfn{compilation wordlist} (aka current
8060: wordlist).
1.1 anton 8061:
1.36 anton 8062: @cindex wid
8063: A word list is identified by a cell-sized word list identifier (@i{wid})
8064: in much the same way as a file is identified by a file handle. The
8065: numerical value of the wid has no (portable) meaning, and might change
8066: from session to session.
1.1 anton 8067:
1.29 crook 8068: The ANS Forth ``Search order'' word set is intended to provide a set of
8069: low-level tools that allow various different schemes to be
1.74 anton 8070: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26 crook 8071: word. @file{compat/vocabulary.fs} provides an implementation in ANS
1.45 crook 8072: Forth.
1.1 anton 8073:
1.27 crook 8074: @comment TODO: locals section refers to here, saying that every word list (aka
8075: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.78 anton 8076: @c anton: but better in a separate subsection on wordlist internals
1.1 anton 8077:
1.45 crook 8078: @comment TODO: document markers, reveal, tables, mappedwordlist
8079:
8080: @comment the gforthman- prefix is used to pick out the true definition of a
1.27 crook 8081: @comment word from the source files, rather than some alias.
1.44 crook 8082:
1.26 crook 8083: doc-forth-wordlist
8084: doc-definitions
8085: doc-get-current
8086: doc-set-current
8087: doc-get-order
1.45 crook 8088: doc---gforthman-set-order
1.26 crook 8089: doc-wordlist
1.30 anton 8090: doc-table
1.79 anton 8091: doc->order
1.36 anton 8092: doc-previous
1.26 crook 8093: doc-also
1.45 crook 8094: doc---gforthman-forth
1.26 crook 8095: doc-only
1.45 crook 8096: doc---gforthman-order
1.15 anton 8097:
1.26 crook 8098: doc-find
8099: doc-search-wordlist
1.15 anton 8100:
1.26 crook 8101: doc-words
8102: doc-vlist
1.44 crook 8103: @c doc-words-deferred
1.1 anton 8104:
1.74 anton 8105: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26 crook 8106: doc-root
8107: doc-vocabulary
8108: doc-seal
8109: doc-vocs
8110: doc-current
8111: doc-context
1.1 anton 8112:
1.44 crook 8113:
1.26 crook 8114: @menu
1.75 anton 8115: * Vocabularies::
1.67 anton 8116: * Why use word lists?::
1.75 anton 8117: * Word list example::
1.26 crook 8118: @end menu
8119:
1.75 anton 8120: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
8121: @subsection Vocabularies
8122: @cindex Vocabularies, detailed explanation
8123:
8124: Here is an example of creating and using a new wordlist using ANS
8125: Forth words:
8126:
8127: @example
8128: wordlist constant my-new-words-wordlist
8129: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
8130:
8131: \ add it to the search order
8132: also my-new-words
8133:
8134: \ alternatively, add it to the search order and make it
8135: \ the compilation word list
8136: also my-new-words definitions
8137: \ type "order" to see the problem
8138: @end example
8139:
8140: The problem with this example is that @code{order} has no way to
8141: associate the name @code{my-new-words} with the wid of the word list (in
8142: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
8143: that has no associated name). There is no Standard way of associating a
8144: name with a wid.
8145:
8146: In Gforth, this example can be re-coded using @code{vocabulary}, which
8147: associates a name with a wid:
8148:
8149: @example
8150: vocabulary my-new-words
8151:
8152: \ add it to the search order
8153: also my-new-words
8154:
8155: \ alternatively, add it to the search order and make it
8156: \ the compilation word list
8157: my-new-words definitions
8158: \ type "order" to see that the problem is solved
8159: @end example
8160:
8161:
8162: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26 crook 8163: @subsection Why use word lists?
8164: @cindex word lists - why use them?
8165:
1.74 anton 8166: Here are some reasons why people use wordlists:
1.26 crook 8167:
8168: @itemize @bullet
1.74 anton 8169:
8170: @c anton: Gforth's hashing implementation makes the search speed
8171: @c independent from the number of words. But it is linear with the number
8172: @c of wordlists that have to be searched, so in effect using more wordlists
8173: @c actually slows down compilation.
8174:
8175: @c @item
8176: @c To improve compilation speed by reducing the number of header space
8177: @c entries that must be searched. This is achieved by creating a new
8178: @c word list that contains all of the definitions that are used in the
8179: @c definition of a Forth system but which would not usually be used by
8180: @c programs running on that system. That word list would be on the search
8181: @c list when the Forth system was compiled but would be removed from the
8182: @c search list for normal operation. This can be a useful technique for
8183: @c low-performance systems (for example, 8-bit processors in embedded
8184: @c systems) but is unlikely to be necessary in high-performance desktop
8185: @c systems.
8186:
1.26 crook 8187: @item
8188: To prevent a set of words from being used outside the context in which
8189: they are valid. Two classic examples of this are an integrated editor
8190: (all of the edit commands are defined in a separate word list; the
8191: search order is set to the editor word list when the editor is invoked;
8192: the old search order is restored when the editor is terminated) and an
8193: integrated assembler (the op-codes for the machine are defined in a
8194: separate word list which is used when a @code{CODE} word is defined).
1.74 anton 8195:
8196: @item
8197: To organize the words of an application or library into a user-visible
8198: set (in @code{forth-wordlist} or some other common wordlist) and a set
8199: of helper words used just for the implementation (hidden in a separate
1.75 anton 8200: wordlist). This keeps @code{words}' output smaller, separates
8201: implementation and interface, and reduces the chance of name conflicts
8202: within the common wordlist.
1.74 anton 8203:
1.26 crook 8204: @item
8205: To prevent a name-space clash between multiple definitions with the same
8206: name. For example, when building a cross-compiler you might have a word
8207: @code{IF} that generates conditional code for your target system. By
8208: placing this definition in a different word list you can control whether
8209: the host system's @code{IF} or the target system's @code{IF} get used in
8210: any particular context by controlling the order of the word lists on the
8211: search order stack.
1.74 anton 8212:
1.26 crook 8213: @end itemize
1.1 anton 8214:
1.74 anton 8215: The downsides of using wordlists are:
8216:
8217: @itemize
8218:
8219: @item
8220: Debugging becomes more cumbersome.
8221:
8222: @item
8223: Name conflicts worked around with wordlists are still there, and you
8224: have to arrange the search order carefully to get the desired results;
8225: if you forget to do that, you get hard-to-find errors (as in any case
8226: where you read the code differently from the compiler; @code{see} can
1.75 anton 8227: help seeing which of several possible words the name resolves to in such
8228: cases). @code{See} displays just the name of the words, not what
8229: wordlist they belong to, so it might be misleading. Using unique names
8230: is a better approach to avoid name conflicts.
1.74 anton 8231:
8232: @item
8233: You have to explicitly undo any changes to the search order. In many
8234: cases it would be more convenient if this happened implicitly. Gforth
8235: currently does not provide such a feature, but it may do so in the
8236: future.
8237: @end itemize
8238:
8239:
1.75 anton 8240: @node Word list example, , Why use word lists?, Word Lists
8241: @subsection Word list example
8242: @cindex word lists - example
1.1 anton 8243:
1.74 anton 8244: The following example is from the
8245: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
8246: garbage collector} and uses wordlists to separate public words from
8247: helper words:
8248:
8249: @example
8250: get-current ( wid )
8251: vocabulary garbage-collector also garbage-collector definitions
8252: ... \ define helper words
8253: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
8254: ... \ define the public (i.e., API) words
8255: \ they can refer to the helper words
8256: previous \ restore original search order (helper words become invisible)
8257: @end example
8258:
1.26 crook 8259: @c -------------------------------------------------------------
8260: @node Environmental Queries, Files, Word Lists, Words
8261: @section Environmental Queries
8262: @cindex environmental queries
1.21 crook 8263:
1.26 crook 8264: ANS Forth introduced the idea of ``environmental queries'' as a way
8265: for a program running on a system to determine certain characteristics of the system.
8266: The Standard specifies a number of strings that might be recognised by a system.
1.21 crook 8267:
1.32 anton 8268: The Standard requires that the header space used for environmental queries
8269: be distinct from the header space used for definitions.
1.21 crook 8270:
1.26 crook 8271: Typically, environmental queries are supported by creating a set of
1.29 crook 8272: definitions in a word list that is @i{only} used during environmental
1.26 crook 8273: queries; that is what Gforth does. There is no Standard way of adding
8274: definitions to the set of recognised environmental queries, but any
8275: implementation that supports the loading of optional word sets must have
8276: some mechanism for doing this (after loading the word set, the
8277: associated environmental query string must return @code{true}). In
8278: Gforth, the word list used to honour environmental queries can be
8279: manipulated just like any other word list.
1.21 crook 8280:
1.44 crook 8281:
1.26 crook 8282: doc-environment?
8283: doc-environment-wordlist
1.21 crook 8284:
1.26 crook 8285: doc-gforth
8286: doc-os-class
1.21 crook 8287:
1.44 crook 8288:
1.26 crook 8289: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
8290: returning two items on the stack, querying it using @code{environment?}
8291: will return an additional item; the @code{true} flag that shows that the
8292: string was recognised.
1.21 crook 8293:
1.26 crook 8294: @comment TODO Document the standard strings or note where they are documented herein
1.21 crook 8295:
1.26 crook 8296: Here are some examples of using environmental queries:
1.21 crook 8297:
1.26 crook 8298: @example
8299: s" address-unit-bits" environment? 0=
8300: [IF]
8301: cr .( environmental attribute address-units-bits unknown... ) cr
1.75 anton 8302: [ELSE]
8303: drop \ ensure balanced stack effect
1.26 crook 8304: [THEN]
1.21 crook 8305:
1.75 anton 8306: \ this might occur in the prelude of a standard program that uses THROW
8307: s" exception" environment? [IF]
8308: 0= [IF]
8309: : throw abort" exception thrown" ;
8310: [THEN]
8311: [ELSE] \ we don't know, so make sure
8312: : throw abort" exception thrown" ;
8313: [THEN]
1.21 crook 8314:
1.26 crook 8315: s" gforth" environment? [IF] .( Gforth version ) TYPE
8316: [ELSE] .( Not Gforth..) [THEN]
1.75 anton 8317:
8318: \ a program using v*
8319: s" gforth" environment? [IF]
8320: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
8321: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8322: >r swap 2swap swap 0e r> 0 ?DO
8323: dup f@ over + 2swap dup f@ f* f+ over + 2swap
8324: LOOP
8325: 2drop 2drop ;
8326: [THEN]
8327: [ELSE] \
8328: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8329: ...
8330: [THEN]
1.26 crook 8331: @end example
1.21 crook 8332:
1.26 crook 8333: Here is an example of adding a definition to the environment word list:
1.21 crook 8334:
1.26 crook 8335: @example
8336: get-current environment-wordlist set-current
8337: true constant block
8338: true constant block-ext
8339: set-current
8340: @end example
1.21 crook 8341:
1.26 crook 8342: You can see what definitions are in the environment word list like this:
1.21 crook 8343:
1.26 crook 8344: @example
1.79 anton 8345: environment-wordlist >order words previous
1.26 crook 8346: @end example
1.21 crook 8347:
8348:
1.26 crook 8349: @c -------------------------------------------------------------
8350: @node Files, Blocks, Environmental Queries, Words
8351: @section Files
1.28 crook 8352: @cindex files
8353: @cindex I/O - file-handling
1.21 crook 8354:
1.26 crook 8355: Gforth provides facilities for accessing files that are stored in the
8356: host operating system's file-system. Files that are processed by Gforth
8357: can be divided into two categories:
1.21 crook 8358:
1.23 crook 8359: @itemize @bullet
8360: @item
1.29 crook 8361: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23 crook 8362: @item
1.29 crook 8363: Files that are processed by some other program (@dfn{general files}).
1.26 crook 8364: @end itemize
8365:
8366: @menu
1.48 anton 8367: * Forth source files::
8368: * General files::
8369: * Search Paths::
1.26 crook 8370: @end menu
8371:
8372: @c -------------------------------------------------------------
8373: @node Forth source files, General files, Files, Files
8374: @subsection Forth source files
8375: @cindex including files
8376: @cindex Forth source files
1.21 crook 8377:
1.26 crook 8378: The simplest way to interpret the contents of a file is to use one of
8379: these two formats:
1.21 crook 8380:
1.26 crook 8381: @example
8382: include mysource.fs
8383: s" mysource.fs" included
8384: @end example
1.21 crook 8385:
1.75 anton 8386: You usually want to include a file only if it is not included already
1.26 crook 8387: (by, say, another source file). In that case, you can use one of these
1.45 crook 8388: three formats:
1.21 crook 8389:
1.26 crook 8390: @example
8391: require mysource.fs
8392: needs mysource.fs
8393: s" mysource.fs" required
8394: @end example
1.21 crook 8395:
1.26 crook 8396: @cindex stack effect of included files
8397: @cindex including files, stack effect
1.45 crook 8398: It is good practice to write your source files such that interpreting them
8399: does not change the stack. Source files designed in this way can be used with
1.26 crook 8400: @code{required} and friends without complications. For example:
1.21 crook 8401:
1.26 crook 8402: @example
1.75 anton 8403: 1024 require foo.fs drop
1.26 crook 8404: @end example
1.21 crook 8405:
1.75 anton 8406: Here you want to pass the argument 1024 (e.g., a buffer size) to
8407: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8408: ), which allows its use with @code{require}. Of course with such
8409: parameters to required files, you have to ensure that the first
8410: @code{require} fits for all uses (i.e., @code{require} it early in the
8411: master load file).
1.44 crook 8412:
1.26 crook 8413: doc-include-file
8414: doc-included
1.28 crook 8415: doc-included?
1.26 crook 8416: doc-include
8417: doc-required
8418: doc-require
8419: doc-needs
1.75 anton 8420: @c doc-init-included-files @c internal
8421: doc-sourcefilename
8422: doc-sourceline#
1.44 crook 8423:
1.26 crook 8424: A definition in ANS Forth for @code{required} is provided in
8425: @file{compat/required.fs}.
1.21 crook 8426:
1.26 crook 8427: @c -------------------------------------------------------------
8428: @node General files, Search Paths, Forth source files, Files
8429: @subsection General files
8430: @cindex general files
8431: @cindex file-handling
1.21 crook 8432:
1.75 anton 8433: Files are opened/created by name and type. The following file access
8434: methods (FAMs) are recognised:
1.44 crook 8435:
1.75 anton 8436: @cindex fam (file access method)
1.26 crook 8437: doc-r/o
8438: doc-r/w
8439: doc-w/o
8440: doc-bin
1.1 anton 8441:
1.44 crook 8442:
1.26 crook 8443: When a file is opened/created, it returns a file identifier,
1.29 crook 8444: @i{wfileid} that is used for all other file commands. All file
8445: commands also return a status value, @i{wior}, that is 0 for a
1.26 crook 8446: successful operation and an implementation-defined non-zero value in the
8447: case of an error.
1.21 crook 8448:
1.44 crook 8449:
1.26 crook 8450: doc-open-file
8451: doc-create-file
1.21 crook 8452:
1.26 crook 8453: doc-close-file
8454: doc-delete-file
8455: doc-rename-file
8456: doc-read-file
8457: doc-read-line
8458: doc-write-file
8459: doc-write-line
8460: doc-emit-file
8461: doc-flush-file
1.21 crook 8462:
1.26 crook 8463: doc-file-status
8464: doc-file-position
8465: doc-reposition-file
8466: doc-file-size
8467: doc-resize-file
1.21 crook 8468:
1.44 crook 8469:
1.26 crook 8470: @c ---------------------------------------------------------
1.48 anton 8471: @node Search Paths, , General files, Files
1.26 crook 8472: @subsection Search Paths
8473: @cindex path for @code{included}
8474: @cindex file search path
8475: @cindex @code{include} search path
8476: @cindex search path for files
1.21 crook 8477:
1.26 crook 8478: If you specify an absolute filename (i.e., a filename starting with
8479: @file{/} or @file{~}, or with @file{:} in the second position (as in
8480: @samp{C:...})) for @code{included} and friends, that file is included
8481: just as you would expect.
1.21 crook 8482:
1.75 anton 8483: If the filename starts with @file{./}, this refers to the directory that
8484: the present file was @code{included} from. This allows files to include
8485: other files relative to their own position (irrespective of the current
8486: working directory or the absolute position). This feature is essential
8487: for libraries consisting of several files, where a file may include
8488: other files from the library. It corresponds to @code{#include "..."}
8489: in C. If the current input source is not a file, @file{.} refers to the
8490: directory of the innermost file being included, or, if there is no file
8491: being included, to the current working directory.
8492:
8493: For relative filenames (not starting with @file{./}), Gforth uses a
8494: search path similar to Forth's search order (@pxref{Word Lists}). It
8495: tries to find the given filename in the directories present in the path,
8496: and includes the first one it finds. There are separate search paths for
8497: Forth source files and general files. If the search path contains the
8498: directory @file{.}, this refers to the directory of the current file, or
8499: the working directory, as if the file had been specified with @file{./}.
1.21 crook 8500:
1.26 crook 8501: Use @file{~+} to refer to the current working directory (as in the
8502: @code{bash}).
1.1 anton 8503:
1.75 anton 8504: @c anton: fold the following subsubsections into this subsection?
1.1 anton 8505:
1.48 anton 8506: @menu
1.75 anton 8507: * Source Search Paths::
1.48 anton 8508: * General Search Paths::
8509: @end menu
8510:
1.26 crook 8511: @c ---------------------------------------------------------
1.75 anton 8512: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8513: @subsubsection Source Search Paths
8514: @cindex search path control, source files
1.5 anton 8515:
1.26 crook 8516: The search path is initialized when you start Gforth (@pxref{Invoking
1.75 anton 8517: Gforth}). You can display it and change it using @code{fpath} in
8518: combination with the general path handling words.
1.5 anton 8519:
1.75 anton 8520: doc-fpath
8521: @c the functionality of the following words is easily available through
8522: @c fpath and the general path words. The may go away.
8523: @c doc-.fpath
8524: @c doc-fpath+
8525: @c doc-fpath=
8526: @c doc-open-fpath-file
1.44 crook 8527:
8528: @noindent
1.26 crook 8529: Here is an example of using @code{fpath} and @code{require}:
1.5 anton 8530:
1.26 crook 8531: @example
1.75 anton 8532: fpath path= /usr/lib/forth/|./
1.26 crook 8533: require timer.fs
8534: @end example
1.5 anton 8535:
1.75 anton 8536:
1.26 crook 8537: @c ---------------------------------------------------------
1.75 anton 8538: @node General Search Paths, , Source Search Paths, Search Paths
1.26 crook 8539: @subsubsection General Search Paths
1.75 anton 8540: @cindex search path control, source files
1.5 anton 8541:
1.26 crook 8542: Your application may need to search files in several directories, like
8543: @code{included} does. To facilitate this, Gforth allows you to define
8544: and use your own search paths, by providing generic equivalents of the
8545: Forth search path words:
1.5 anton 8546:
1.75 anton 8547: doc-open-path-file
8548: doc-path-allot
8549: doc-clear-path
8550: doc-also-path
1.26 crook 8551: doc-.path
8552: doc-path+
8553: doc-path=
1.5 anton 8554:
1.75 anton 8555: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44 crook 8556:
1.75 anton 8557: Here's an example of creating an empty search path:
8558: @c
1.26 crook 8559: @example
1.75 anton 8560: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26 crook 8561: @end example
1.5 anton 8562:
1.26 crook 8563: @c -------------------------------------------------------------
8564: @node Blocks, Other I/O, Files, Words
8565: @section Blocks
1.28 crook 8566: @cindex I/O - blocks
8567: @cindex blocks
8568:
8569: When you run Gforth on a modern desk-top computer, it runs under the
8570: control of an operating system which provides certain services. One of
8571: these services is @var{file services}, which allows Forth source code
8572: and data to be stored in files and read into Gforth (@pxref{Files}).
8573:
8574: Traditionally, Forth has been an important programming language on
8575: systems where it has interfaced directly to the underlying hardware with
8576: no intervening operating system. Forth provides a mechanism, called
1.29 crook 8577: @dfn{blocks}, for accessing mass storage on such systems.
1.28 crook 8578:
8579: A block is a 1024-byte data area, which can be used to hold data or
8580: Forth source code. No structure is imposed on the contents of the
8581: block. A block is identified by its number; blocks are numbered
8582: contiguously from 1 to an implementation-defined maximum.
8583:
8584: A typical system that used blocks but no operating system might use a
8585: single floppy-disk drive for mass storage, with the disks formatted to
8586: provide 256-byte sectors. Blocks would be implemented by assigning the
8587: first four sectors of the disk to block 1, the second four sectors to
8588: block 2 and so on, up to the limit of the capacity of the disk. The disk
8589: would not contain any file system information, just the set of blocks.
8590:
1.29 crook 8591: @cindex blocks file
1.28 crook 8592: On systems that do provide file services, blocks are typically
1.29 crook 8593: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28 crook 8594: file}. The size of the blocks file will be an exact multiple of 1024
8595: bytes, corresponding to the number of blocks it contains. This is the
8596: mechanism that Gforth uses.
8597:
1.29 crook 8598: @cindex @file{blocks.fb}
1.75 anton 8599: Only one blocks file can be open at a time. If you use block words without
1.28 crook 8600: having specified a blocks file, Gforth defaults to the blocks file
8601: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75 anton 8602: locate a blocks file (@pxref{Source Search Paths}).
1.28 crook 8603:
1.29 crook 8604: @cindex block buffers
1.28 crook 8605: When you read and write blocks under program control, Gforth uses a
1.29 crook 8606: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28 crook 8607: not used when you use @code{load} to interpret the contents of a block.
8608:
1.75 anton 8609: The behaviour of the block buffers is analagous to that of a cache.
8610: Each block buffer has three states:
1.28 crook 8611:
8612: @itemize @bullet
8613: @item
8614: Unassigned
8615: @item
8616: Assigned-clean
8617: @item
8618: Assigned-dirty
8619: @end itemize
8620:
1.29 crook 8621: Initially, all block buffers are @i{unassigned}. In order to access a
1.28 crook 8622: block, the block (specified by its block number) must be assigned to a
8623: block buffer.
8624:
8625: The assignment of a block to a block buffer is performed by @code{block}
8626: or @code{buffer}. Use @code{block} when you wish to modify the existing
8627: contents of a block. Use @code{buffer} when you don't care about the
8628: existing contents of the block@footnote{The ANS Forth definition of
1.35 anton 8629: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28 crook 8630: with the particular block is already stored in a block buffer due to an
8631: earlier @code{block} command, @code{buffer} will return that block
8632: buffer and the existing contents of the block will be
8633: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29 crook 8634: block buffer for the block.}.
1.28 crook 8635:
1.47 crook 8636: Once a block has been assigned to a block buffer using @code{block} or
1.75 anton 8637: @code{buffer}, that block buffer becomes the @i{current block
8638: buffer}. Data may only be manipulated (read or written) within the
8639: current block buffer.
1.47 crook 8640:
8641: When the contents of the current block buffer has been modified it is
1.48 anton 8642: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75 anton 8643: either abandon the changes (by doing nothing) or mark the block as
8644: changed (assigned-dirty), using @code{update}. Using @code{update} does
8645: not change the blocks file; it simply changes a block buffer's state to
8646: @i{assigned-dirty}. The block will be written implicitly when it's
8647: buffer is needed for another block, or explicitly by @code{flush} or
8648: @code{save-buffers}.
8649:
8650: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8651: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8652: @code{flush}.
1.28 crook 8653:
1.29 crook 8654: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28 crook 8655: algorithm to assign a block buffer to a block. That means that any
8656: particular block can only be assigned to one specific block buffer,
1.29 crook 8657: called (for the particular operation) the @i{victim buffer}. If the
1.47 crook 8658: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8659: the new block immediately. If it is @i{assigned-dirty} its current
8660: contents are written back to the blocks file on disk before it is
1.28 crook 8661: allocated to the new block.
8662:
8663: Although no structure is imposed on the contents of a block, it is
8664: traditional to display the contents as 16 lines each of 64 characters. A
8665: block provides a single, continuous stream of input (for example, it
8666: acts as a single parse area) -- there are no end-of-line characters
8667: within a block, and no end-of-file character at the end of a
8668: block. There are two consequences of this:
1.26 crook 8669:
1.28 crook 8670: @itemize @bullet
8671: @item
8672: The last character of one line wraps straight into the first character
8673: of the following line
8674: @item
8675: The word @code{\} -- comment to end of line -- requires special
8676: treatment; in the context of a block it causes all characters until the
8677: end of the current 64-character ``line'' to be ignored.
8678: @end itemize
8679:
8680: In Gforth, when you use @code{block} with a non-existent block number,
1.45 crook 8681: the current blocks file will be extended to the appropriate size and the
1.28 crook 8682: block buffer will be initialised with spaces.
8683:
1.47 crook 8684: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8685: for details) but doesn't encourage the use of blocks; the mechanism is
8686: only provided for backward compatibility -- ANS Forth requires blocks to
8687: be available when files are.
1.28 crook 8688:
8689: Common techniques that are used when working with blocks include:
8690:
8691: @itemize @bullet
8692: @item
8693: A screen editor that allows you to edit blocks without leaving the Forth
8694: environment.
8695: @item
8696: Shadow screens; where every code block has an associated block
8697: containing comments (for example: code in odd block numbers, comments in
8698: even block numbers). Typically, the block editor provides a convenient
8699: mechanism to toggle between code and comments.
8700: @item
8701: Load blocks; a single block (typically block 1) contains a number of
8702: @code{thru} commands which @code{load} the whole of the application.
8703: @end itemize
1.26 crook 8704:
1.29 crook 8705: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8706: integrated into a Forth programming environment.
1.26 crook 8707:
8708: @comment TODO what about errors on open-blocks?
1.44 crook 8709:
1.26 crook 8710: doc-open-blocks
8711: doc-use
1.75 anton 8712: doc-block-offset
1.26 crook 8713: doc-get-block-fid
8714: doc-block-position
1.28 crook 8715:
1.75 anton 8716: doc-list
1.28 crook 8717: doc-scr
8718:
1.45 crook 8719: doc---gforthman-block
1.28 crook 8720: doc-buffer
8721:
1.75 anton 8722: doc-empty-buffers
8723: doc-empty-buffer
1.26 crook 8724: doc-update
1.28 crook 8725: doc-updated?
1.26 crook 8726: doc-save-buffers
1.75 anton 8727: doc-save-buffer
1.26 crook 8728: doc-flush
1.28 crook 8729:
1.26 crook 8730: doc-load
8731: doc-thru
8732: doc-+load
8733: doc-+thru
1.45 crook 8734: doc---gforthman--->
1.26 crook 8735: doc-block-included
8736:
1.44 crook 8737:
1.26 crook 8738: @c -------------------------------------------------------------
1.78 anton 8739: @node Other I/O, Locals, Blocks, Words
1.26 crook 8740: @section Other I/O
1.28 crook 8741: @cindex I/O - keyboard and display
1.26 crook 8742:
8743: @menu
8744: * Simple numeric output:: Predefined formats
8745: * Formatted numeric output:: Formatted (pictured) output
8746: * String Formats:: How Forth stores strings in memory
1.67 anton 8747: * Displaying characters and strings:: Other stuff
1.26 crook 8748: * Input:: Input
8749: @end menu
8750:
8751: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8752: @subsection Simple numeric output
1.28 crook 8753: @cindex numeric output - simple/free-format
1.5 anton 8754:
1.26 crook 8755: The simplest output functions are those that display numbers from the
8756: data or floating-point stacks. Floating-point output is always displayed
8757: using base 10. Numbers displayed from the data stack use the value stored
8758: in @code{base}.
1.5 anton 8759:
1.44 crook 8760:
1.26 crook 8761: doc-.
8762: doc-dec.
8763: doc-hex.
8764: doc-u.
8765: doc-.r
8766: doc-u.r
8767: doc-d.
8768: doc-ud.
8769: doc-d.r
8770: doc-ud.r
8771: doc-f.
8772: doc-fe.
8773: doc-fs.
1.5 anton 8774:
1.44 crook 8775:
1.26 crook 8776: Examples of printing the number 1234.5678E23 in the different floating-point output
8777: formats are shown below:
1.5 anton 8778:
8779: @example
1.26 crook 8780: f. 123456779999999000000000000.
8781: fe. 123.456779999999E24
8782: fs. 1.23456779999999E26
1.5 anton 8783: @end example
8784:
8785:
1.26 crook 8786: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8787: @subsection Formatted numeric output
1.28 crook 8788: @cindex formatted numeric output
1.26 crook 8789: @cindex pictured numeric output
1.28 crook 8790: @cindex numeric output - formatted
1.26 crook 8791:
1.29 crook 8792: Forth traditionally uses a technique called @dfn{pictured numeric
1.26 crook 8793: output} for formatted printing of integers. In this technique, digits
8794: are extracted from the number (using the current output radix defined by
8795: @code{base}), converted to ASCII codes and appended to a string that is
8796: built in a scratch-pad area of memory (@pxref{core-idef,
8797: Implementation-defined options, Implementation-defined
8798: options}). Arbitrary characters can be appended to the string during the
8799: extraction process. The completed string is specified by an address
8800: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8801: under program control.
1.5 anton 8802:
1.75 anton 8803: All of the integer output words described in the previous section
8804: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8805: numeric output.
1.5 anton 8806:
1.47 crook 8807: Three important things to remember about pictured numeric output:
1.5 anton 8808:
1.26 crook 8809: @itemize @bullet
8810: @item
1.28 crook 8811: It always operates on double-precision numbers; to display a
1.49 anton 8812: single-precision number, convert it first (for ways of doing this
8813: @pxref{Double precision}).
1.26 crook 8814: @item
1.28 crook 8815: It always treats the double-precision number as though it were
8816: unsigned. The examples below show ways of printing signed numbers.
1.26 crook 8817: @item
8818: The string is built up from right to left; least significant digit first.
8819: @end itemize
1.5 anton 8820:
1.44 crook 8821:
1.26 crook 8822: doc-<#
1.47 crook 8823: doc-<<#
1.26 crook 8824: doc-#
8825: doc-#s
8826: doc-hold
8827: doc-sign
8828: doc-#>
1.47 crook 8829: doc-#>>
1.5 anton 8830:
1.26 crook 8831: doc-represent
1.5 anton 8832:
1.44 crook 8833:
8834: @noindent
1.26 crook 8835: Here are some examples of using pictured numeric output:
1.5 anton 8836:
8837: @example
1.26 crook 8838: : my-u. ( u -- )
8839: \ Simplest use of pns.. behaves like Standard u.
8840: 0 \ convert to unsigned double
1.75 anton 8841: <<# \ start conversion
1.26 crook 8842: #s \ convert all digits
8843: #> \ complete conversion
1.75 anton 8844: TYPE SPACE \ display, with trailing space
8845: #>> ; \ release hold area
1.5 anton 8846:
1.26 crook 8847: : cents-only ( u -- )
8848: 0 \ convert to unsigned double
1.75 anton 8849: <<# \ start conversion
1.26 crook 8850: # # \ convert two least-significant digits
8851: #> \ complete conversion, discard other digits
1.75 anton 8852: TYPE SPACE \ display, with trailing space
8853: #>> ; \ release hold area
1.5 anton 8854:
1.26 crook 8855: : dollars-and-cents ( u -- )
8856: 0 \ convert to unsigned double
1.75 anton 8857: <<# \ start conversion
1.26 crook 8858: # # \ convert two least-significant digits
8859: [char] . hold \ insert decimal point
8860: #s \ convert remaining digits
8861: [char] $ hold \ append currency symbol
8862: #> \ complete conversion
1.75 anton 8863: TYPE SPACE \ display, with trailing space
8864: #>> ; \ release hold area
1.5 anton 8865:
1.26 crook 8866: : my-. ( n -- )
8867: \ handling negatives.. behaves like Standard .
8868: s>d \ convert to signed double
8869: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8870: <<# \ start conversion
1.26 crook 8871: #s \ convert all digits
8872: rot sign \ get at sign byte, append "-" if needed
8873: #> \ complete conversion
1.75 anton 8874: TYPE SPACE \ display, with trailing space
8875: #>> ; \ release hold area
1.5 anton 8876:
1.26 crook 8877: : account. ( n -- )
1.75 anton 8878: \ accountants don't like minus signs, they use parentheses
1.26 crook 8879: \ for negative numbers
8880: s>d \ convert to signed double
8881: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8882: <<# \ start conversion
1.26 crook 8883: 2 pick \ get copy of sign byte
8884: 0< IF [char] ) hold THEN \ right-most character of output
8885: #s \ convert all digits
8886: rot \ get at sign byte
8887: 0< IF [char] ( hold THEN
8888: #> \ complete conversion
1.75 anton 8889: TYPE SPACE \ display, with trailing space
8890: #>> ; \ release hold area
8891:
1.5 anton 8892: @end example
8893:
1.26 crook 8894: Here are some examples of using these words:
1.5 anton 8895:
8896: @example
1.26 crook 8897: 1 my-u. 1
8898: hex -1 my-u. decimal FFFFFFFF
8899: 1 cents-only 01
8900: 1234 cents-only 34
8901: 2 dollars-and-cents $0.02
8902: 1234 dollars-and-cents $12.34
8903: 123 my-. 123
8904: -123 my. -123
8905: 123 account. 123
8906: -456 account. (456)
1.5 anton 8907: @end example
8908:
8909:
1.26 crook 8910: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8911: @subsection String Formats
1.27 crook 8912: @cindex strings - see character strings
8913: @cindex character strings - formats
1.28 crook 8914: @cindex I/O - see character strings
1.75 anton 8915: @cindex counted strings
8916:
8917: @c anton: this does not really belong here; maybe the memory section,
8918: @c or the principles chapter
1.26 crook 8919:
1.27 crook 8920: Forth commonly uses two different methods for representing character
8921: strings:
1.26 crook 8922:
8923: @itemize @bullet
8924: @item
8925: @cindex address of counted string
1.45 crook 8926: @cindex counted string
1.29 crook 8927: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8928: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8929: string and the string occupies the subsequent @i{n} char addresses in
1.26 crook 8930: memory.
8931: @item
1.29 crook 8932: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8933: of the string in characters, and @i{c-addr} is the address of the
1.26 crook 8934: first byte of the string.
8935: @end itemize
8936:
8937: ANS Forth encourages the use of the second format when representing
1.75 anton 8938: strings.
1.26 crook 8939:
1.44 crook 8940:
1.26 crook 8941: doc-count
8942:
1.44 crook 8943:
1.49 anton 8944: For words that move, copy and search for strings see @ref{Memory
8945: Blocks}. For words that display characters and strings see
8946: @ref{Displaying characters and strings}.
1.26 crook 8947:
8948: @node Displaying characters and strings, Input, String Formats, Other I/O
8949: @subsection Displaying characters and strings
1.27 crook 8950: @cindex characters - compiling and displaying
8951: @cindex character strings - compiling and displaying
1.26 crook 8952:
8953: This section starts with a glossary of Forth words and ends with a set
8954: of examples.
8955:
1.44 crook 8956:
1.26 crook 8957: doc-bl
8958: doc-space
8959: doc-spaces
8960: doc-emit
8961: doc-toupper
8962: doc-."
8963: doc-.(
8964: doc-type
1.44 crook 8965: doc-typewhite
1.26 crook 8966: doc-cr
1.27 crook 8967: @cindex cursor control
1.26 crook 8968: doc-at-xy
8969: doc-page
8970: doc-s"
8971: doc-c"
8972: doc-char
8973: doc-[char]
8974:
1.44 crook 8975:
8976: @noindent
1.26 crook 8977: As an example, consider the following text, stored in a file @file{test.fs}:
1.5 anton 8978:
8979: @example
1.26 crook 8980: .( text-1)
8981: : my-word
8982: ." text-2" cr
8983: .( text-3)
8984: ;
8985:
8986: ." text-4"
8987:
8988: : my-char
8989: [char] ALPHABET emit
8990: char emit
8991: ;
1.5 anton 8992: @end example
8993:
1.26 crook 8994: When you load this code into Gforth, the following output is generated:
1.5 anton 8995:
1.26 crook 8996: @example
1.30 anton 8997: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26 crook 8998: @end example
1.5 anton 8999:
1.26 crook 9000: @itemize @bullet
9001: @item
9002: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
9003: is an immediate word; it behaves in the same way whether it is used inside
9004: or outside a colon definition.
9005: @item
9006: Message @code{text-4} is displayed because of Gforth's added interpretation
9007: semantics for @code{."}.
9008: @item
1.29 crook 9009: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26 crook 9010: performs the compilation semantics for @code{."} within the definition of
9011: @code{my-word}.
9012: @end itemize
1.5 anton 9013:
1.26 crook 9014: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5 anton 9015:
1.26 crook 9016: @example
1.30 anton 9017: @kbd{my-word @key{RET}} text-2
1.26 crook 9018: ok
1.30 anton 9019: @kbd{my-char fred @key{RET}} Af ok
9020: @kbd{my-char jim @key{RET}} Aj ok
1.26 crook 9021: @end example
1.5 anton 9022:
9023: @itemize @bullet
9024: @item
1.26 crook 9025: Message @code{text-2} is displayed because of the run-time behaviour of
9026: @code{."}.
9027: @item
9028: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
9029: on the stack at run-time. @code{emit} always displays the character
9030: when @code{my-char} is executed.
9031: @item
9032: @code{char} parses a string at run-time and the second @code{emit} displays
9033: the first character of the string.
1.5 anton 9034: @item
1.26 crook 9035: If you type @code{see my-char} you can see that @code{[char]} discarded
9036: the text ``LPHABET'' and only compiled the display code for ``A'' into the
9037: definition of @code{my-char}.
1.5 anton 9038: @end itemize
9039:
9040:
9041:
1.48 anton 9042: @node Input, , Displaying characters and strings, Other I/O
1.26 crook 9043: @subsection Input
9044: @cindex input
1.28 crook 9045: @cindex I/O - see input
9046: @cindex parsing a string
1.5 anton 9047:
1.49 anton 9048: For ways of storing character strings in memory see @ref{String Formats}.
1.5 anton 9049:
1.27 crook 9050: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29 crook 9051: @comment then index them
1.27 crook 9052:
1.44 crook 9053:
1.27 crook 9054: doc-key
9055: doc-key?
1.45 crook 9056: doc-ekey
9057: doc-ekey?
9058: doc-ekey>char
1.26 crook 9059: doc->number
9060: doc->float
9061: doc-accept
1.27 crook 9062: doc-pad
1.75 anton 9063: @c anton: these belong in the input stream section
1.27 crook 9064: doc-parse
9065: doc-word
9066: doc-sword
1.75 anton 9067: doc-name
1.27 crook 9068: doc-refill
9069: @comment obsolescent words..
9070: doc-convert
1.26 crook 9071: doc-query
9072: doc-expect
1.27 crook 9073: doc-span
1.5 anton 9074:
9075:
1.78 anton 9076: @c -------------------------------------------------------------
9077: @node Locals, Structures, Other I/O, Words
9078: @section Locals
9079: @cindex locals
9080:
9081: Local variables can make Forth programming more enjoyable and Forth
9082: programs easier to read. Unfortunately, the locals of ANS Forth are
9083: laden with restrictions. Therefore, we provide not only the ANS Forth
9084: locals wordset, but also our own, more powerful locals wordset (we
9085: implemented the ANS Forth locals wordset through our locals wordset).
1.44 crook 9086:
1.78 anton 9087: The ideas in this section have also been published in M. Anton Ertl,
9088: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9089: Automatic Scoping of Local Variables}}, EuroForth '94.
1.12 anton 9090:
9091: @menu
1.78 anton 9092: * Gforth locals::
9093: * ANS Forth locals::
1.5 anton 9094: @end menu
9095:
1.78 anton 9096: @node Gforth locals, ANS Forth locals, Locals, Locals
9097: @subsection Gforth locals
9098: @cindex Gforth locals
9099: @cindex locals, Gforth style
1.5 anton 9100:
1.78 anton 9101: Locals can be defined with
1.44 crook 9102:
1.78 anton 9103: @example
9104: @{ local1 local2 ... -- comment @}
9105: @end example
9106: or
9107: @example
9108: @{ local1 local2 ... @}
9109: @end example
1.5 anton 9110:
1.78 anton 9111: E.g.,
9112: @example
9113: : max @{ n1 n2 -- n3 @}
9114: n1 n2 > if
9115: n1
9116: else
9117: n2
9118: endif ;
9119: @end example
1.44 crook 9120:
1.78 anton 9121: The similarity of locals definitions with stack comments is intended. A
9122: locals definition often replaces the stack comment of a word. The order
9123: of the locals corresponds to the order in a stack comment and everything
9124: after the @code{--} is really a comment.
1.77 anton 9125:
1.78 anton 9126: This similarity has one disadvantage: It is too easy to confuse locals
9127: declarations with stack comments, causing bugs and making them hard to
9128: find. However, this problem can be avoided by appropriate coding
9129: conventions: Do not use both notations in the same program. If you do,
9130: they should be distinguished using additional means, e.g. by position.
1.77 anton 9131:
1.78 anton 9132: @cindex types of locals
9133: @cindex locals types
9134: The name of the local may be preceded by a type specifier, e.g.,
9135: @code{F:} for a floating point value:
1.5 anton 9136:
1.78 anton 9137: @example
9138: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9139: \ complex multiplication
9140: Ar Br f* Ai Bi f* f-
9141: Ar Bi f* Ai Br f* f+ ;
9142: @end example
1.44 crook 9143:
1.78 anton 9144: @cindex flavours of locals
9145: @cindex locals flavours
9146: @cindex value-flavoured locals
9147: @cindex variable-flavoured locals
9148: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9149: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9150: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9151: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9152: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9153: produces its address (which becomes invalid when the variable's scope is
9154: left). E.g., the standard word @code{emit} can be defined in terms of
9155: @code{type} like this:
1.5 anton 9156:
1.78 anton 9157: @example
9158: : emit @{ C^ char* -- @}
9159: char* 1 type ;
9160: @end example
1.5 anton 9161:
1.78 anton 9162: @cindex default type of locals
9163: @cindex locals, default type
9164: A local without type specifier is a @code{W:} local. Both flavours of
9165: locals are initialized with values from the data or FP stack.
1.44 crook 9166:
1.78 anton 9167: Currently there is no way to define locals with user-defined data
9168: structures, but we are working on it.
1.5 anton 9169:
1.78 anton 9170: Gforth allows defining locals everywhere in a colon definition. This
9171: poses the following questions:
1.5 anton 9172:
1.78 anton 9173: @menu
9174: * Where are locals visible by name?::
9175: * How long do locals live?::
9176: * Locals programming style::
9177: * Locals implementation::
9178: @end menu
1.44 crook 9179:
1.78 anton 9180: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9181: @subsubsection Where are locals visible by name?
9182: @cindex locals visibility
9183: @cindex visibility of locals
9184: @cindex scope of locals
1.5 anton 9185:
1.78 anton 9186: Basically, the answer is that locals are visible where you would expect
9187: it in block-structured languages, and sometimes a little longer. If you
9188: want to restrict the scope of a local, enclose its definition in
9189: @code{SCOPE}...@code{ENDSCOPE}.
1.5 anton 9190:
9191:
1.78 anton 9192: doc-scope
9193: doc-endscope
1.5 anton 9194:
9195:
1.78 anton 9196: These words behave like control structure words, so you can use them
9197: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9198: arbitrary ways.
1.77 anton 9199:
1.78 anton 9200: If you want a more exact answer to the visibility question, here's the
9201: basic principle: A local is visible in all places that can only be
9202: reached through the definition of the local@footnote{In compiler
9203: construction terminology, all places dominated by the definition of the
9204: local.}. In other words, it is not visible in places that can be reached
9205: without going through the definition of the local. E.g., locals defined
9206: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9207: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9208: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.77 anton 9209:
1.78 anton 9210: The reasoning behind this solution is: We want to have the locals
9211: visible as long as it is meaningful. The user can always make the
9212: visibility shorter by using explicit scoping. In a place that can
9213: only be reached through the definition of a local, the meaning of a
9214: local name is clear. In other places it is not: How is the local
9215: initialized at the control flow path that does not contain the
9216: definition? Which local is meant, if the same name is defined twice in
9217: two independent control flow paths?
1.77 anton 9218:
1.78 anton 9219: This should be enough detail for nearly all users, so you can skip the
9220: rest of this section. If you really must know all the gory details and
9221: options, read on.
1.77 anton 9222:
1.78 anton 9223: In order to implement this rule, the compiler has to know which places
9224: are unreachable. It knows this automatically after @code{AHEAD},
9225: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9226: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9227: compiler that the control flow never reaches that place. If
9228: @code{UNREACHABLE} is not used where it could, the only consequence is
9229: that the visibility of some locals is more limited than the rule above
9230: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9231: lie to the compiler), buggy code will be produced.
1.77 anton 9232:
1.5 anton 9233:
1.78 anton 9234: doc-unreachable
1.5 anton 9235:
1.23 crook 9236:
1.78 anton 9237: Another problem with this rule is that at @code{BEGIN}, the compiler
9238: does not know which locals will be visible on the incoming
9239: back-edge. All problems discussed in the following are due to this
9240: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9241: loops as examples; the discussion also applies to @code{?DO} and other
9242: loops). Perhaps the most insidious example is:
1.26 crook 9243: @example
1.78 anton 9244: AHEAD
9245: BEGIN
9246: x
9247: [ 1 CS-ROLL ] THEN
9248: @{ x @}
9249: ...
9250: UNTIL
1.26 crook 9251: @end example
1.23 crook 9252:
1.78 anton 9253: This should be legal according to the visibility rule. The use of
9254: @code{x} can only be reached through the definition; but that appears
9255: textually below the use.
9256:
9257: From this example it is clear that the visibility rules cannot be fully
9258: implemented without major headaches. Our implementation treats common
9259: cases as advertised and the exceptions are treated in a safe way: The
9260: compiler makes a reasonable guess about the locals visible after a
9261: @code{BEGIN}; if it is too pessimistic, the
9262: user will get a spurious error about the local not being defined; if the
9263: compiler is too optimistic, it will notice this later and issue a
9264: warning. In the case above the compiler would complain about @code{x}
9265: being undefined at its use. You can see from the obscure examples in
9266: this section that it takes quite unusual control structures to get the
9267: compiler into trouble, and even then it will often do fine.
1.23 crook 9268:
1.78 anton 9269: If the @code{BEGIN} is reachable from above, the most optimistic guess
9270: is that all locals visible before the @code{BEGIN} will also be
9271: visible after the @code{BEGIN}. This guess is valid for all loops that
9272: are entered only through the @code{BEGIN}, in particular, for normal
9273: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9274: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9275: compiler. When the branch to the @code{BEGIN} is finally generated by
9276: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9277: warns the user if it was too optimistic:
1.26 crook 9278: @example
1.78 anton 9279: IF
9280: @{ x @}
9281: BEGIN
9282: \ x ?
9283: [ 1 cs-roll ] THEN
9284: ...
9285: UNTIL
1.26 crook 9286: @end example
1.23 crook 9287:
1.78 anton 9288: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9289: optimistically assumes that it lives until the @code{THEN}. It notices
9290: this difference when it compiles the @code{UNTIL} and issues a
9291: warning. The user can avoid the warning, and make sure that @code{x}
9292: is not used in the wrong area by using explicit scoping:
9293: @example
9294: IF
9295: SCOPE
9296: @{ x @}
9297: ENDSCOPE
9298: BEGIN
9299: [ 1 cs-roll ] THEN
9300: ...
9301: UNTIL
9302: @end example
1.23 crook 9303:
1.78 anton 9304: Since the guess is optimistic, there will be no spurious error messages
9305: about undefined locals.
1.44 crook 9306:
1.78 anton 9307: If the @code{BEGIN} is not reachable from above (e.g., after
9308: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9309: optimistic guess, as the locals visible after the @code{BEGIN} may be
9310: defined later. Therefore, the compiler assumes that no locals are
9311: visible after the @code{BEGIN}. However, the user can use
9312: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9313: visible at the BEGIN as at the point where the top control-flow stack
9314: item was created.
1.23 crook 9315:
1.44 crook 9316:
1.78 anton 9317: doc-assume-live
1.26 crook 9318:
1.23 crook 9319:
1.78 anton 9320: @noindent
9321: E.g.,
9322: @example
9323: @{ x @}
9324: AHEAD
9325: ASSUME-LIVE
9326: BEGIN
9327: x
9328: [ 1 CS-ROLL ] THEN
9329: ...
9330: UNTIL
9331: @end example
1.44 crook 9332:
1.78 anton 9333: Other cases where the locals are defined before the @code{BEGIN} can be
9334: handled by inserting an appropriate @code{CS-ROLL} before the
9335: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9336: behind the @code{ASSUME-LIVE}).
1.23 crook 9337:
1.78 anton 9338: Cases where locals are defined after the @code{BEGIN} (but should be
9339: visible immediately after the @code{BEGIN}) can only be handled by
9340: rearranging the loop. E.g., the ``most insidious'' example above can be
9341: arranged into:
9342: @example
9343: BEGIN
9344: @{ x @}
9345: ... 0=
9346: WHILE
9347: x
9348: REPEAT
9349: @end example
1.44 crook 9350:
1.78 anton 9351: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
9352: @subsubsection How long do locals live?
9353: @cindex locals lifetime
9354: @cindex lifetime of locals
1.23 crook 9355:
1.78 anton 9356: The right answer for the lifetime question would be: A local lives at
9357: least as long as it can be accessed. For a value-flavoured local this
9358: means: until the end of its visibility. However, a variable-flavoured
9359: local could be accessed through its address far beyond its visibility
9360: scope. Ultimately, this would mean that such locals would have to be
9361: garbage collected. Since this entails un-Forth-like implementation
9362: complexities, I adopted the same cowardly solution as some other
9363: languages (e.g., C): The local lives only as long as it is visible;
9364: afterwards its address is invalid (and programs that access it
9365: afterwards are erroneous).
1.23 crook 9366:
1.78 anton 9367: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9368: @subsubsection Locals programming style
9369: @cindex locals programming style
9370: @cindex programming style, locals
1.23 crook 9371:
1.78 anton 9372: The freedom to define locals anywhere has the potential to change
9373: programming styles dramatically. In particular, the need to use the
9374: return stack for intermediate storage vanishes. Moreover, all stack
9375: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9376: determined arguments) can be eliminated: If the stack items are in the
9377: wrong order, just write a locals definition for all of them; then
9378: write the items in the order you want.
1.23 crook 9379:
1.78 anton 9380: This seems a little far-fetched and eliminating stack manipulations is
9381: unlikely to become a conscious programming objective. Still, the number
9382: of stack manipulations will be reduced dramatically if local variables
9383: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9384: a traditional implementation of @code{max}).
1.23 crook 9385:
1.78 anton 9386: This shows one potential benefit of locals: making Forth programs more
9387: readable. Of course, this benefit will only be realized if the
9388: programmers continue to honour the principle of factoring instead of
9389: using the added latitude to make the words longer.
1.23 crook 9390:
1.78 anton 9391: @cindex single-assignment style for locals
9392: Using @code{TO} can and should be avoided. Without @code{TO},
9393: every value-flavoured local has only a single assignment and many
9394: advantages of functional languages apply to Forth. I.e., programs are
9395: easier to analyse, to optimize and to read: It is clear from the
9396: definition what the local stands for, it does not turn into something
9397: different later.
1.23 crook 9398:
1.78 anton 9399: E.g., a definition using @code{TO} might look like this:
9400: @example
9401: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9402: u1 u2 min 0
9403: ?do
9404: addr1 c@@ addr2 c@@ -
9405: ?dup-if
9406: unloop exit
9407: then
9408: addr1 char+ TO addr1
9409: addr2 char+ TO addr2
9410: loop
9411: u1 u2 - ;
1.26 crook 9412: @end example
1.78 anton 9413: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9414: every loop iteration. @code{strcmp} is a typical example of the
9415: readability problems of using @code{TO}. When you start reading
9416: @code{strcmp}, you think that @code{addr1} refers to the start of the
9417: string. Only near the end of the loop you realize that it is something
9418: else.
1.23 crook 9419:
1.78 anton 9420: This can be avoided by defining two locals at the start of the loop that
9421: are initialized with the right value for the current iteration.
9422: @example
9423: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9424: addr1 addr2
9425: u1 u2 min 0
9426: ?do @{ s1 s2 @}
9427: s1 c@@ s2 c@@ -
9428: ?dup-if
9429: unloop exit
9430: then
9431: s1 char+ s2 char+
9432: loop
9433: 2drop
9434: u1 u2 - ;
9435: @end example
9436: Here it is clear from the start that @code{s1} has a different value
9437: in every loop iteration.
1.23 crook 9438:
1.78 anton 9439: @node Locals implementation, , Locals programming style, Gforth locals
9440: @subsubsection Locals implementation
9441: @cindex locals implementation
9442: @cindex implementation of locals
1.23 crook 9443:
1.78 anton 9444: @cindex locals stack
9445: Gforth uses an extra locals stack. The most compelling reason for
9446: this is that the return stack is not float-aligned; using an extra stack
9447: also eliminates the problems and restrictions of using the return stack
9448: as locals stack. Like the other stacks, the locals stack grows toward
9449: lower addresses. A few primitives allow an efficient implementation:
9450:
9451:
9452: doc-@local#
9453: doc-f@local#
9454: doc-laddr#
9455: doc-lp+!#
9456: doc-lp!
9457: doc->l
9458: doc-f>l
9459:
9460:
9461: In addition to these primitives, some specializations of these
9462: primitives for commonly occurring inline arguments are provided for
9463: efficiency reasons, e.g., @code{@@local0} as specialization of
9464: @code{@@local#} for the inline argument 0. The following compiling words
9465: compile the right specialized version, or the general version, as
9466: appropriate:
1.23 crook 9467:
1.5 anton 9468:
1.78 anton 9469: doc-compile-@local
9470: doc-compile-f@local
9471: doc-compile-lp+!
1.5 anton 9472:
9473:
1.78 anton 9474: Combinations of conditional branches and @code{lp+!#} like
9475: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9476: is taken) are provided for efficiency and correctness in loops.
1.5 anton 9477:
1.78 anton 9478: A special area in the dictionary space is reserved for keeping the
9479: local variable names. @code{@{} switches the dictionary pointer to this
9480: area and @code{@}} switches it back and generates the locals
9481: initializing code. @code{W:} etc.@ are normal defining words. This
9482: special area is cleared at the start of every colon definition.
1.5 anton 9483:
1.78 anton 9484: @cindex word list for defining locals
9485: A special feature of Gforth's dictionary is used to implement the
9486: definition of locals without type specifiers: every word list (aka
9487: vocabulary) has its own methods for searching
9488: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9489: with a special search method: When it is searched for a word, it
9490: actually creates that word using @code{W:}. @code{@{} changes the search
9491: order to first search the word list containing @code{@}}, @code{W:} etc.,
9492: and then the word list for defining locals without type specifiers.
1.5 anton 9493:
1.78 anton 9494: The lifetime rules support a stack discipline within a colon
9495: definition: The lifetime of a local is either nested with other locals
9496: lifetimes or it does not overlap them.
1.23 crook 9497:
1.78 anton 9498: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9499: pointer manipulation is generated. Between control structure words
9500: locals definitions can push locals onto the locals stack. @code{AGAIN}
9501: is the simplest of the other three control flow words. It has to
9502: restore the locals stack depth of the corresponding @code{BEGIN}
9503: before branching. The code looks like this:
9504: @format
9505: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9506: @code{branch} <begin>
9507: @end format
1.26 crook 9508:
1.78 anton 9509: @code{UNTIL} is a little more complicated: If it branches back, it
9510: must adjust the stack just like @code{AGAIN}. But if it falls through,
9511: the locals stack must not be changed. The compiler generates the
9512: following code:
9513: @format
9514: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9515: @end format
9516: The locals stack pointer is only adjusted if the branch is taken.
1.44 crook 9517:
1.78 anton 9518: @code{THEN} can produce somewhat inefficient code:
9519: @format
9520: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9521: <orig target>:
9522: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9523: @end format
9524: The second @code{lp+!#} adjusts the locals stack pointer from the
9525: level at the @i{orig} point to the level after the @code{THEN}. The
9526: first @code{lp+!#} adjusts the locals stack pointer from the current
9527: level to the level at the orig point, so the complete effect is an
9528: adjustment from the current level to the right level after the
9529: @code{THEN}.
1.26 crook 9530:
1.78 anton 9531: @cindex locals information on the control-flow stack
9532: @cindex control-flow stack items, locals information
9533: In a conventional Forth implementation a dest control-flow stack entry
9534: is just the target address and an orig entry is just the address to be
9535: patched. Our locals implementation adds a word list to every orig or dest
9536: item. It is the list of locals visible (or assumed visible) at the point
9537: described by the entry. Our implementation also adds a tag to identify
9538: the kind of entry, in particular to differentiate between live and dead
9539: (reachable and unreachable) orig entries.
1.26 crook 9540:
1.78 anton 9541: A few unusual operations have to be performed on locals word lists:
1.44 crook 9542:
1.5 anton 9543:
1.78 anton 9544: doc-common-list
9545: doc-sub-list?
9546: doc-list-size
1.52 anton 9547:
9548:
1.78 anton 9549: Several features of our locals word list implementation make these
9550: operations easy to implement: The locals word lists are organised as
9551: linked lists; the tails of these lists are shared, if the lists
9552: contain some of the same locals; and the address of a name is greater
9553: than the address of the names behind it in the list.
1.5 anton 9554:
1.78 anton 9555: Another important implementation detail is the variable
9556: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9557: determine if they can be reached directly or only through the branch
9558: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9559: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9560: definition, by @code{BEGIN} and usually by @code{THEN}.
1.5 anton 9561:
1.78 anton 9562: Counted loops are similar to other loops in most respects, but
9563: @code{LEAVE} requires special attention: It performs basically the same
9564: service as @code{AHEAD}, but it does not create a control-flow stack
9565: entry. Therefore the information has to be stored elsewhere;
9566: traditionally, the information was stored in the target fields of the
9567: branches created by the @code{LEAVE}s, by organizing these fields into a
9568: linked list. Unfortunately, this clever trick does not provide enough
9569: space for storing our extended control flow information. Therefore, we
9570: introduce another stack, the leave stack. It contains the control-flow
9571: stack entries for all unresolved @code{LEAVE}s.
1.44 crook 9572:
1.78 anton 9573: Local names are kept until the end of the colon definition, even if
9574: they are no longer visible in any control-flow path. In a few cases
9575: this may lead to increased space needs for the locals name area, but
9576: usually less than reclaiming this space would cost in code size.
1.5 anton 9577:
1.44 crook 9578:
1.78 anton 9579: @node ANS Forth locals, , Gforth locals, Locals
9580: @subsection ANS Forth locals
9581: @cindex locals, ANS Forth style
1.5 anton 9582:
1.78 anton 9583: The ANS Forth locals wordset does not define a syntax for locals, but
9584: words that make it possible to define various syntaxes. One of the
9585: possible syntaxes is a subset of the syntax we used in the Gforth locals
9586: wordset, i.e.:
1.29 crook 9587:
9588: @example
1.78 anton 9589: @{ local1 local2 ... -- comment @}
9590: @end example
9591: @noindent
9592: or
9593: @example
9594: @{ local1 local2 ... @}
1.29 crook 9595: @end example
9596:
1.78 anton 9597: The order of the locals corresponds to the order in a stack comment. The
9598: restrictions are:
1.5 anton 9599:
1.78 anton 9600: @itemize @bullet
9601: @item
9602: Locals can only be cell-sized values (no type specifiers are allowed).
9603: @item
9604: Locals can be defined only outside control structures.
9605: @item
9606: Locals can interfere with explicit usage of the return stack. For the
9607: exact (and long) rules, see the standard. If you don't use return stack
9608: accessing words in a definition using locals, you will be all right. The
9609: purpose of this rule is to make locals implementation on the return
9610: stack easier.
9611: @item
9612: The whole definition must be in one line.
9613: @end itemize
1.5 anton 9614:
1.78 anton 9615: Locals defined in ANS Forth behave like @code{VALUE}s
9616: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9617: name produces their value. Their value can be changed using @code{TO}.
1.77 anton 9618:
1.78 anton 9619: Since the syntax above is supported by Gforth directly, you need not do
9620: anything to use it. If you want to port a program using this syntax to
9621: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9622: syntax on the other system.
1.5 anton 9623:
1.78 anton 9624: Note that a syntax shown in the standard, section A.13 looks
9625: similar, but is quite different in having the order of locals
9626: reversed. Beware!
1.5 anton 9627:
1.78 anton 9628: The ANS Forth locals wordset itself consists of one word:
1.5 anton 9629:
1.78 anton 9630: doc-(local)
1.5 anton 9631:
1.78 anton 9632: The ANS Forth locals extension wordset defines a syntax using
9633: @code{locals|}, but it is so awful that we strongly recommend not to use
9634: it. We have implemented this syntax to make porting to Gforth easy, but
9635: do not document it here. The problem with this syntax is that the locals
9636: are defined in an order reversed with respect to the standard stack
9637: comment notation, making programs harder to read, and easier to misread
9638: and miswrite. The only merit of this syntax is that it is easy to
9639: implement using the ANS Forth locals wordset.
1.53 anton 9640:
9641:
1.78 anton 9642: @c ----------------------------------------------------------
9643: @node Structures, Object-oriented Forth, Locals, Words
9644: @section Structures
9645: @cindex structures
9646: @cindex records
1.53 anton 9647:
1.78 anton 9648: This section presents the structure package that comes with Gforth. A
9649: version of the package implemented in ANS Forth is available in
9650: @file{compat/struct.fs}. This package was inspired by a posting on
9651: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9652: possibly John Hayes). A version of this section has been published in
9653: M. Anton Ertl,
9654: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9655: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9656: 13--16. Marcel Hendrix provided helpful comments.
1.53 anton 9657:
1.78 anton 9658: @menu
9659: * Why explicit structure support?::
9660: * Structure Usage::
9661: * Structure Naming Convention::
9662: * Structure Implementation::
9663: * Structure Glossary::
9664: @end menu
1.55 anton 9665:
1.78 anton 9666: @node Why explicit structure support?, Structure Usage, Structures, Structures
9667: @subsection Why explicit structure support?
1.53 anton 9668:
1.78 anton 9669: @cindex address arithmetic for structures
9670: @cindex structures using address arithmetic
9671: If we want to use a structure containing several fields, we could simply
9672: reserve memory for it, and access the fields using address arithmetic
9673: (@pxref{Address arithmetic}). As an example, consider a structure with
9674: the following fields
1.57 anton 9675:
1.78 anton 9676: @table @code
9677: @item a
9678: is a float
9679: @item b
9680: is a cell
9681: @item c
9682: is a float
9683: @end table
1.57 anton 9684:
1.78 anton 9685: Given the (float-aligned) base address of the structure we get the
9686: address of the field
1.52 anton 9687:
1.78 anton 9688: @table @code
9689: @item a
9690: without doing anything further.
9691: @item b
9692: with @code{float+}
9693: @item c
9694: with @code{float+ cell+ faligned}
9695: @end table
1.52 anton 9696:
1.78 anton 9697: It is easy to see that this can become quite tiring.
1.52 anton 9698:
1.78 anton 9699: Moreover, it is not very readable, because seeing a
9700: @code{cell+} tells us neither which kind of structure is
9701: accessed nor what field is accessed; we have to somehow infer the kind
9702: of structure, and then look up in the documentation, which field of
9703: that structure corresponds to that offset.
1.53 anton 9704:
1.78 anton 9705: Finally, this kind of address arithmetic also causes maintenance
9706: troubles: If you add or delete a field somewhere in the middle of the
9707: structure, you have to find and change all computations for the fields
9708: afterwards.
1.52 anton 9709:
1.78 anton 9710: So, instead of using @code{cell+} and friends directly, how
9711: about storing the offsets in constants:
1.52 anton 9712:
1.78 anton 9713: @example
9714: 0 constant a-offset
9715: 0 float+ constant b-offset
9716: 0 float+ cell+ faligned c-offset
9717: @end example
1.64 pazsan 9718:
1.78 anton 9719: Now we can get the address of field @code{x} with @code{x-offset
9720: +}. This is much better in all respects. Of course, you still
9721: have to change all later offset definitions if you add a field. You can
9722: fix this by declaring the offsets in the following way:
1.57 anton 9723:
1.78 anton 9724: @example
9725: 0 constant a-offset
9726: a-offset float+ constant b-offset
9727: b-offset cell+ faligned constant c-offset
9728: @end example
1.57 anton 9729:
1.78 anton 9730: Since we always use the offsets with @code{+}, we could use a defining
9731: word @code{cfield} that includes the @code{+} in the action of the
9732: defined word:
1.64 pazsan 9733:
1.78 anton 9734: @example
9735: : cfield ( n "name" -- )
9736: create ,
9737: does> ( name execution: addr1 -- addr2 )
9738: @@ + ;
1.64 pazsan 9739:
1.78 anton 9740: 0 cfield a
9741: 0 a float+ cfield b
9742: 0 b cell+ faligned cfield c
9743: @end example
1.64 pazsan 9744:
1.78 anton 9745: Instead of @code{x-offset +}, we now simply write @code{x}.
1.64 pazsan 9746:
1.78 anton 9747: The structure field words now can be used quite nicely. However,
9748: their definition is still a bit cumbersome: We have to repeat the
9749: name, the information about size and alignment is distributed before
9750: and after the field definitions etc. The structure package presented
9751: here addresses these problems.
1.64 pazsan 9752:
1.78 anton 9753: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9754: @subsection Structure Usage
9755: @cindex structure usage
1.57 anton 9756:
1.78 anton 9757: @cindex @code{field} usage
9758: @cindex @code{struct} usage
9759: @cindex @code{end-struct} usage
9760: You can define a structure for a (data-less) linked list with:
1.57 anton 9761: @example
1.78 anton 9762: struct
9763: cell% field list-next
9764: end-struct list%
1.57 anton 9765: @end example
9766:
1.78 anton 9767: With the address of the list node on the stack, you can compute the
9768: address of the field that contains the address of the next node with
9769: @code{list-next}. E.g., you can determine the length of a list
9770: with:
1.57 anton 9771:
9772: @example
1.78 anton 9773: : list-length ( list -- n )
9774: \ "list" is a pointer to the first element of a linked list
9775: \ "n" is the length of the list
9776: 0 BEGIN ( list1 n1 )
9777: over
9778: WHILE ( list1 n1 )
9779: 1+ swap list-next @@ swap
9780: REPEAT
9781: nip ;
1.57 anton 9782: @end example
9783:
1.78 anton 9784: You can reserve memory for a list node in the dictionary with
9785: @code{list% %allot}, which leaves the address of the list node on the
9786: stack. For the equivalent allocation on the heap you can use @code{list%
9787: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9788: use @code{list% %allocate}). You can get the the size of a list
9789: node with @code{list% %size} and its alignment with @code{list%
9790: %alignment}.
9791:
9792: Note that in ANS Forth the body of a @code{create}d word is
9793: @code{aligned} but not necessarily @code{faligned};
9794: therefore, if you do a:
1.57 anton 9795:
9796: @example
1.78 anton 9797: create @emph{name} foo% %allot drop
1.57 anton 9798: @end example
9799:
1.78 anton 9800: @noindent
9801: then the memory alloted for @code{foo%} is guaranteed to start at the
9802: body of @code{@emph{name}} only if @code{foo%} contains only character,
9803: cell and double fields. Therefore, if your structure contains floats,
9804: better use
1.57 anton 9805:
9806: @example
1.78 anton 9807: foo% %allot constant @emph{name}
1.57 anton 9808: @end example
9809:
1.78 anton 9810: @cindex structures containing structures
9811: You can include a structure @code{foo%} as a field of
9812: another structure, like this:
1.65 anton 9813: @example
1.78 anton 9814: struct
9815: ...
9816: foo% field ...
9817: ...
9818: end-struct ...
1.65 anton 9819: @end example
1.52 anton 9820:
1.78 anton 9821: @cindex structure extension
9822: @cindex extended records
9823: Instead of starting with an empty structure, you can extend an
9824: existing structure. E.g., a plain linked list without data, as defined
9825: above, is hardly useful; You can extend it to a linked list of integers,
9826: like this:@footnote{This feature is also known as @emph{extended
9827: records}. It is the main innovation in the Oberon language; in other
9828: words, adding this feature to Modula-2 led Wirth to create a new
9829: language, write a new compiler etc. Adding this feature to Forth just
9830: required a few lines of code.}
1.52 anton 9831:
1.78 anton 9832: @example
9833: list%
9834: cell% field intlist-int
9835: end-struct intlist%
9836: @end example
1.55 anton 9837:
1.78 anton 9838: @code{intlist%} is a structure with two fields:
9839: @code{list-next} and @code{intlist-int}.
1.55 anton 9840:
1.78 anton 9841: @cindex structures containing arrays
9842: You can specify an array type containing @emph{n} elements of
9843: type @code{foo%} like this:
1.55 anton 9844:
9845: @example
1.78 anton 9846: foo% @emph{n} *
1.56 anton 9847: @end example
1.55 anton 9848:
1.78 anton 9849: You can use this array type in any place where you can use a normal
9850: type, e.g., when defining a @code{field}, or with
9851: @code{%allot}.
9852:
9853: @cindex first field optimization
9854: The first field is at the base address of a structure and the word for
9855: this field (e.g., @code{list-next}) actually does not change the address
9856: on the stack. You may be tempted to leave it away in the interest of
9857: run-time and space efficiency. This is not necessary, because the
9858: structure package optimizes this case: If you compile a first-field
9859: words, no code is generated. So, in the interest of readability and
9860: maintainability you should include the word for the field when accessing
9861: the field.
1.52 anton 9862:
9863:
1.78 anton 9864: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9865: @subsection Structure Naming Convention
9866: @cindex structure naming convention
1.52 anton 9867:
1.78 anton 9868: The field names that come to (my) mind are often quite generic, and,
9869: if used, would cause frequent name clashes. E.g., many structures
9870: probably contain a @code{counter} field. The structure names
9871: that come to (my) mind are often also the logical choice for the names
9872: of words that create such a structure.
1.52 anton 9873:
1.78 anton 9874: Therefore, I have adopted the following naming conventions:
1.52 anton 9875:
1.78 anton 9876: @itemize @bullet
9877: @cindex field naming convention
9878: @item
9879: The names of fields are of the form
9880: @code{@emph{struct}-@emph{field}}, where
9881: @code{@emph{struct}} is the basic name of the structure, and
9882: @code{@emph{field}} is the basic name of the field. You can
9883: think of field words as converting the (address of the)
9884: structure into the (address of the) field.
1.52 anton 9885:
1.78 anton 9886: @cindex structure naming convention
9887: @item
9888: The names of structures are of the form
9889: @code{@emph{struct}%}, where
9890: @code{@emph{struct}} is the basic name of the structure.
9891: @end itemize
1.52 anton 9892:
1.78 anton 9893: This naming convention does not work that well for fields of extended
9894: structures; e.g., the integer list structure has a field
9895: @code{intlist-int}, but has @code{list-next}, not
9896: @code{intlist-next}.
1.53 anton 9897:
1.78 anton 9898: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9899: @subsection Structure Implementation
9900: @cindex structure implementation
9901: @cindex implementation of structures
1.52 anton 9902:
1.78 anton 9903: The central idea in the implementation is to pass the data about the
9904: structure being built on the stack, not in some global
9905: variable. Everything else falls into place naturally once this design
9906: decision is made.
1.53 anton 9907:
1.78 anton 9908: The type description on the stack is of the form @emph{align
9909: size}. Keeping the size on the top-of-stack makes dealing with arrays
9910: very simple.
1.53 anton 9911:
1.78 anton 9912: @code{field} is a defining word that uses @code{Create}
9913: and @code{DOES>}. The body of the field contains the offset
9914: of the field, and the normal @code{DOES>} action is simply:
1.53 anton 9915:
9916: @example
1.78 anton 9917: @@ +
1.53 anton 9918: @end example
9919:
1.78 anton 9920: @noindent
9921: i.e., add the offset to the address, giving the stack effect
9922: @i{addr1 -- addr2} for a field.
9923:
9924: @cindex first field optimization, implementation
9925: This simple structure is slightly complicated by the optimization
9926: for fields with offset 0, which requires a different
9927: @code{DOES>}-part (because we cannot rely on there being
9928: something on the stack if such a field is invoked during
9929: compilation). Therefore, we put the different @code{DOES>}-parts
9930: in separate words, and decide which one to invoke based on the
9931: offset. For a zero offset, the field is basically a noop; it is
9932: immediate, and therefore no code is generated when it is compiled.
1.53 anton 9933:
1.78 anton 9934: @node Structure Glossary, , Structure Implementation, Structures
9935: @subsection Structure Glossary
9936: @cindex structure glossary
1.53 anton 9937:
1.5 anton 9938:
1.78 anton 9939: doc-%align
9940: doc-%alignment
9941: doc-%alloc
9942: doc-%allocate
9943: doc-%allot
9944: doc-cell%
9945: doc-char%
9946: doc-dfloat%
9947: doc-double%
9948: doc-end-struct
9949: doc-field
9950: doc-float%
9951: doc-naligned
9952: doc-sfloat%
9953: doc-%size
9954: doc-struct
1.54 anton 9955:
9956:
1.26 crook 9957: @c -------------------------------------------------------------
1.78 anton 9958: @node Object-oriented Forth, Programming Tools, Structures, Words
9959: @section Object-oriented Forth
9960:
9961: Gforth comes with three packages for object-oriented programming:
9962: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9963: is preloaded, so you have to @code{include} them before use. The most
9964: important differences between these packages (and others) are discussed
9965: in @ref{Comparison with other object models}. All packages are written
9966: in ANS Forth and can be used with any other ANS Forth.
1.5 anton 9967:
1.78 anton 9968: @menu
9969: * Why object-oriented programming?::
9970: * Object-Oriented Terminology::
9971: * Objects::
9972: * OOF::
9973: * Mini-OOF::
9974: * Comparison with other object models::
9975: @end menu
1.5 anton 9976:
1.78 anton 9977: @c ----------------------------------------------------------------
9978: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9979: @subsection Why object-oriented programming?
9980: @cindex object-oriented programming motivation
9981: @cindex motivation for object-oriented programming
1.44 crook 9982:
1.78 anton 9983: Often we have to deal with several data structures (@emph{objects}),
9984: that have to be treated similarly in some respects, but differently in
9985: others. Graphical objects are the textbook example: circles, triangles,
9986: dinosaurs, icons, and others, and we may want to add more during program
9987: development. We want to apply some operations to any graphical object,
9988: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9989: has to do something different for every kind of object.
9990: @comment TODO add some other operations eg perimeter, area
9991: @comment and tie in to concrete examples later..
1.5 anton 9992:
1.78 anton 9993: We could implement @code{draw} as a big @code{CASE}
9994: control structure that executes the appropriate code depending on the
9995: kind of object to be drawn. This would be not be very elegant, and,
9996: moreover, we would have to change @code{draw} every time we add
9997: a new kind of graphical object (say, a spaceship).
1.44 crook 9998:
1.78 anton 9999: What we would rather do is: When defining spaceships, we would tell
10000: the system: ``Here's how you @code{draw} a spaceship; you figure
10001: out the rest''.
1.5 anton 10002:
1.78 anton 10003: This is the problem that all systems solve that (rightfully) call
10004: themselves object-oriented; the object-oriented packages presented here
10005: solve this problem (and not much else).
10006: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.44 crook 10007:
1.78 anton 10008: @c ------------------------------------------------------------------------
10009: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
10010: @subsection Object-Oriented Terminology
10011: @cindex object-oriented terminology
10012: @cindex terminology for object-oriented programming
1.5 anton 10013:
1.78 anton 10014: This section is mainly for reference, so you don't have to understand
10015: all of it right away. The terminology is mainly Smalltalk-inspired. In
10016: short:
1.44 crook 10017:
1.78 anton 10018: @table @emph
10019: @cindex class
10020: @item class
10021: a data structure definition with some extras.
1.5 anton 10022:
1.78 anton 10023: @cindex object
10024: @item object
10025: an instance of the data structure described by the class definition.
1.5 anton 10026:
1.78 anton 10027: @cindex instance variables
10028: @item instance variables
10029: fields of the data structure.
1.5 anton 10030:
1.78 anton 10031: @cindex selector
10032: @cindex method selector
10033: @cindex virtual function
10034: @item selector
10035: (or @emph{method selector}) a word (e.g.,
10036: @code{draw}) that performs an operation on a variety of data
10037: structures (classes). A selector describes @emph{what} operation to
10038: perform. In C++ terminology: a (pure) virtual function.
1.5 anton 10039:
1.78 anton 10040: @cindex method
10041: @item method
10042: the concrete definition that performs the operation
10043: described by the selector for a specific class. A method specifies
10044: @emph{how} the operation is performed for a specific class.
1.5 anton 10045:
1.78 anton 10046: @cindex selector invocation
10047: @cindex message send
10048: @cindex invoking a selector
10049: @item selector invocation
10050: a call of a selector. One argument of the call (the TOS (top-of-stack))
10051: is used for determining which method is used. In Smalltalk terminology:
10052: a message (consisting of the selector and the other arguments) is sent
10053: to the object.
1.5 anton 10054:
1.78 anton 10055: @cindex receiving object
10056: @item receiving object
10057: the object used for determining the method executed by a selector
10058: invocation. In the @file{objects.fs} model, it is the object that is on
10059: the TOS when the selector is invoked. (@emph{Receiving} comes from
10060: the Smalltalk @emph{message} terminology.)
1.5 anton 10061:
1.78 anton 10062: @cindex child class
10063: @cindex parent class
10064: @cindex inheritance
10065: @item child class
10066: a class that has (@emph{inherits}) all properties (instance variables,
10067: selectors, methods) from a @emph{parent class}. In Smalltalk
10068: terminology: The subclass inherits from the superclass. In C++
10069: terminology: The derived class inherits from the base class.
1.5 anton 10070:
1.78 anton 10071: @end table
1.5 anton 10072:
1.78 anton 10073: @c If you wonder about the message sending terminology, it comes from
10074: @c a time when each object had it's own task and objects communicated via
10075: @c message passing; eventually the Smalltalk developers realized that
10076: @c they can do most things through simple (indirect) calls. They kept the
10077: @c terminology.
1.5 anton 10078:
1.78 anton 10079: @c --------------------------------------------------------------
10080: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10081: @subsection The @file{objects.fs} model
10082: @cindex objects
10083: @cindex object-oriented programming
1.26 crook 10084:
1.78 anton 10085: @cindex @file{objects.fs}
10086: @cindex @file{oof.fs}
1.26 crook 10087:
1.78 anton 10088: This section describes the @file{objects.fs} package. This material also
10089: has been published in M. Anton Ertl,
10090: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10091: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10092: 37--43.
10093: @c McKewan's and Zsoter's packages
1.26 crook 10094:
1.78 anton 10095: This section assumes that you have read @ref{Structures}.
1.5 anton 10096:
1.78 anton 10097: The techniques on which this model is based have been used to implement
10098: the parser generator, Gray, and have also been used in Gforth for
10099: implementing the various flavours of word lists (hashed or not,
10100: case-sensitive or not, special-purpose word lists for locals etc.).
1.5 anton 10101:
10102:
1.26 crook 10103: @menu
1.78 anton 10104: * Properties of the Objects model::
10105: * Basic Objects Usage::
10106: * The Objects base class::
10107: * Creating objects::
10108: * Object-Oriented Programming Style::
10109: * Class Binding::
10110: * Method conveniences::
10111: * Classes and Scoping::
10112: * Dividing classes::
10113: * Object Interfaces::
10114: * Objects Implementation::
10115: * Objects Glossary::
1.26 crook 10116: @end menu
1.5 anton 10117:
1.78 anton 10118: Marcel Hendrix provided helpful comments on this section.
1.5 anton 10119:
1.78 anton 10120: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10121: @subsubsection Properties of the @file{objects.fs} model
10122: @cindex @file{objects.fs} properties
1.5 anton 10123:
1.78 anton 10124: @itemize @bullet
10125: @item
10126: It is straightforward to pass objects on the stack. Passing
10127: selectors on the stack is a little less convenient, but possible.
1.44 crook 10128:
1.78 anton 10129: @item
10130: Objects are just data structures in memory, and are referenced by their
10131: address. You can create words for objects with normal defining words
10132: like @code{constant}. Likewise, there is no difference between instance
10133: variables that contain objects and those that contain other data.
1.5 anton 10134:
1.78 anton 10135: @item
10136: Late binding is efficient and easy to use.
1.44 crook 10137:
1.78 anton 10138: @item
10139: It avoids parsing, and thus avoids problems with state-smartness
10140: and reduced extensibility; for convenience there are a few parsing
10141: words, but they have non-parsing counterparts. There are also a few
10142: defining words that parse. This is hard to avoid, because all standard
10143: defining words parse (except @code{:noname}); however, such
10144: words are not as bad as many other parsing words, because they are not
10145: state-smart.
1.5 anton 10146:
1.78 anton 10147: @item
10148: It does not try to incorporate everything. It does a few things and does
10149: them well (IMO). In particular, this model was not designed to support
10150: information hiding (although it has features that may help); you can use
10151: a separate package for achieving this.
1.5 anton 10152:
1.78 anton 10153: @item
10154: It is layered; you don't have to learn and use all features to use this
10155: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10156: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10157: are optional and independent of each other.
1.5 anton 10158:
1.78 anton 10159: @item
10160: An implementation in ANS Forth is available.
1.5 anton 10161:
1.78 anton 10162: @end itemize
1.5 anton 10163:
1.44 crook 10164:
1.78 anton 10165: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10166: @subsubsection Basic @file{objects.fs} Usage
10167: @cindex basic objects usage
10168: @cindex objects, basic usage
1.5 anton 10169:
1.78 anton 10170: You can define a class for graphical objects like this:
1.44 crook 10171:
1.78 anton 10172: @cindex @code{class} usage
10173: @cindex @code{end-class} usage
10174: @cindex @code{selector} usage
1.5 anton 10175: @example
1.78 anton 10176: object class \ "object" is the parent class
10177: selector draw ( x y graphical -- )
10178: end-class graphical
10179: @end example
10180:
10181: This code defines a class @code{graphical} with an
10182: operation @code{draw}. We can perform the operation
10183: @code{draw} on any @code{graphical} object, e.g.:
10184:
10185: @example
10186: 100 100 t-rex draw
1.26 crook 10187: @end example
1.5 anton 10188:
1.78 anton 10189: @noindent
10190: where @code{t-rex} is a word (say, a constant) that produces a
10191: graphical object.
10192:
10193: @comment TODO add a 2nd operation eg perimeter.. and use for
10194: @comment a concrete example
1.5 anton 10195:
1.78 anton 10196: @cindex abstract class
10197: How do we create a graphical object? With the present definitions,
10198: we cannot create a useful graphical object. The class
10199: @code{graphical} describes graphical objects in general, but not
10200: any concrete graphical object type (C++ users would call it an
10201: @emph{abstract class}); e.g., there is no method for the selector
10202: @code{draw} in the class @code{graphical}.
1.5 anton 10203:
1.78 anton 10204: For concrete graphical objects, we define child classes of the
10205: class @code{graphical}, e.g.:
1.5 anton 10206:
1.78 anton 10207: @cindex @code{overrides} usage
10208: @cindex @code{field} usage in class definition
1.26 crook 10209: @example
1.78 anton 10210: graphical class \ "graphical" is the parent class
10211: cell% field circle-radius
1.5 anton 10212:
1.78 anton 10213: :noname ( x y circle -- )
10214: circle-radius @@ draw-circle ;
10215: overrides draw
1.5 anton 10216:
1.78 anton 10217: :noname ( n-radius circle -- )
10218: circle-radius ! ;
10219: overrides construct
1.5 anton 10220:
1.78 anton 10221: end-class circle
10222: @end example
1.44 crook 10223:
1.78 anton 10224: Here we define a class @code{circle} as a child of @code{graphical},
10225: with field @code{circle-radius} (which behaves just like a field
10226: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10227: for the selectors @code{draw} and @code{construct} (@code{construct} is
10228: defined in @code{object}, the parent class of @code{graphical}).
1.5 anton 10229:
1.78 anton 10230: Now we can create a circle on the heap (i.e.,
10231: @code{allocate}d memory) with:
1.44 crook 10232:
1.78 anton 10233: @cindex @code{heap-new} usage
1.5 anton 10234: @example
1.78 anton 10235: 50 circle heap-new constant my-circle
1.5 anton 10236: @end example
10237:
1.78 anton 10238: @noindent
10239: @code{heap-new} invokes @code{construct}, thus
10240: initializing the field @code{circle-radius} with 50. We can draw
10241: this new circle at (100,100) with:
1.5 anton 10242:
10243: @example
1.78 anton 10244: 100 100 my-circle draw
1.5 anton 10245: @end example
10246:
1.78 anton 10247: @cindex selector invocation, restrictions
10248: @cindex class definition, restrictions
10249: Note: You can only invoke a selector if the object on the TOS
10250: (the receiving object) belongs to the class where the selector was
10251: defined or one of its descendents; e.g., you can invoke
10252: @code{draw} only for objects belonging to @code{graphical}
10253: or its descendents (e.g., @code{circle}). Immediately before
10254: @code{end-class}, the search order has to be the same as
10255: immediately after @code{class}.
10256:
10257: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10258: @subsubsection The @file{object.fs} base class
10259: @cindex @code{object} class
10260:
10261: When you define a class, you have to specify a parent class. So how do
10262: you start defining classes? There is one class available from the start:
10263: @code{object}. It is ancestor for all classes and so is the
10264: only class that has no parent. It has two selectors: @code{construct}
10265: and @code{print}.
10266:
10267: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10268: @subsubsection Creating objects
10269: @cindex creating objects
10270: @cindex object creation
10271: @cindex object allocation options
10272:
10273: @cindex @code{heap-new} discussion
10274: @cindex @code{dict-new} discussion
10275: @cindex @code{construct} discussion
10276: You can create and initialize an object of a class on the heap with
10277: @code{heap-new} ( ... class -- object ) and in the dictionary
10278: (allocation with @code{allot}) with @code{dict-new} (
10279: ... class -- object ). Both words invoke @code{construct}, which
10280: consumes the stack items indicated by "..." above.
10281:
10282: @cindex @code{init-object} discussion
10283: @cindex @code{class-inst-size} discussion
10284: If you want to allocate memory for an object yourself, you can get its
10285: alignment and size with @code{class-inst-size 2@@} ( class --
10286: align size ). Once you have memory for an object, you can initialize
10287: it with @code{init-object} ( ... class object -- );
10288: @code{construct} does only a part of the necessary work.
10289:
10290: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10291: @subsubsection Object-Oriented Programming Style
10292: @cindex object-oriented programming style
10293: @cindex programming style, object-oriented
1.5 anton 10294:
1.78 anton 10295: This section is not exhaustive.
1.5 anton 10296:
1.78 anton 10297: @cindex stack effects of selectors
10298: @cindex selectors and stack effects
10299: In general, it is a good idea to ensure that all methods for the
10300: same selector have the same stack effect: when you invoke a selector,
10301: you often have no idea which method will be invoked, so, unless all
10302: methods have the same stack effect, you will not know the stack effect
10303: of the selector invocation.
1.5 anton 10304:
1.78 anton 10305: One exception to this rule is methods for the selector
10306: @code{construct}. We know which method is invoked, because we
10307: specify the class to be constructed at the same place. Actually, I
10308: defined @code{construct} as a selector only to give the users a
10309: convenient way to specify initialization. The way it is used, a
10310: mechanism different from selector invocation would be more natural
10311: (but probably would take more code and more space to explain).
1.5 anton 10312:
1.78 anton 10313: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10314: @subsubsection Class Binding
10315: @cindex class binding
10316: @cindex early binding
1.5 anton 10317:
1.78 anton 10318: @cindex late binding
10319: Normal selector invocations determine the method at run-time depending
10320: on the class of the receiving object. This run-time selection is called
10321: @i{late binding}.
1.5 anton 10322:
1.78 anton 10323: Sometimes it's preferable to invoke a different method. For example,
10324: you might want to use the simple method for @code{print}ing
10325: @code{object}s instead of the possibly long-winded @code{print} method
10326: of the receiver class. You can achieve this by replacing the invocation
10327: of @code{print} with:
1.5 anton 10328:
1.78 anton 10329: @cindex @code{[bind]} usage
1.5 anton 10330: @example
1.78 anton 10331: [bind] object print
1.5 anton 10332: @end example
10333:
1.78 anton 10334: @noindent
10335: in compiled code or:
10336:
10337: @cindex @code{bind} usage
1.5 anton 10338: @example
1.78 anton 10339: bind object print
1.5 anton 10340: @end example
10341:
1.78 anton 10342: @cindex class binding, alternative to
10343: @noindent
10344: in interpreted code. Alternatively, you can define the method with a
10345: name (e.g., @code{print-object}), and then invoke it through the
10346: name. Class binding is just a (often more convenient) way to achieve
10347: the same effect; it avoids name clutter and allows you to invoke
10348: methods directly without naming them first.
1.5 anton 10349:
1.78 anton 10350: @cindex superclass binding
10351: @cindex parent class binding
10352: A frequent use of class binding is this: When we define a method
10353: for a selector, we often want the method to do what the selector does
10354: in the parent class, and a little more. There is a special word for
10355: this purpose: @code{[parent]}; @code{[parent]
10356: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10357: selector}}, where @code{@emph{parent}} is the parent
10358: class of the current class. E.g., a method definition might look like:
1.44 crook 10359:
1.78 anton 10360: @cindex @code{[parent]} usage
10361: @example
10362: :noname
10363: dup [parent] foo \ do parent's foo on the receiving object
10364: ... \ do some more
10365: ; overrides foo
10366: @end example
1.6 pazsan 10367:
1.78 anton 10368: @cindex class binding as optimization
10369: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10370: March 1997), Andrew McKewan presents class binding as an optimization
10371: technique. I recommend not using it for this purpose unless you are in
10372: an emergency. Late binding is pretty fast with this model anyway, so the
10373: benefit of using class binding is small; the cost of using class binding
10374: where it is not appropriate is reduced maintainability.
1.44 crook 10375:
1.78 anton 10376: While we are at programming style questions: You should bind
10377: selectors only to ancestor classes of the receiving object. E.g., say,
10378: you know that the receiving object is of class @code{foo} or its
10379: descendents; then you should bind only to @code{foo} and its
10380: ancestors.
1.12 anton 10381:
1.78 anton 10382: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10383: @subsubsection Method conveniences
10384: @cindex method conveniences
1.44 crook 10385:
1.78 anton 10386: In a method you usually access the receiving object pretty often. If
10387: you define the method as a plain colon definition (e.g., with
10388: @code{:noname}), you may have to do a lot of stack
10389: gymnastics. To avoid this, you can define the method with @code{m:
10390: ... ;m}. E.g., you could define the method for
10391: @code{draw}ing a @code{circle} with
1.6 pazsan 10392:
1.78 anton 10393: @cindex @code{this} usage
10394: @cindex @code{m:} usage
10395: @cindex @code{;m} usage
10396: @example
10397: m: ( x y circle -- )
10398: ( x y ) this circle-radius @@ draw-circle ;m
10399: @end example
1.6 pazsan 10400:
1.78 anton 10401: @cindex @code{exit} in @code{m: ... ;m}
10402: @cindex @code{exitm} discussion
10403: @cindex @code{catch} in @code{m: ... ;m}
10404: When this method is executed, the receiver object is removed from the
10405: stack; you can access it with @code{this} (admittedly, in this
10406: example the use of @code{m: ... ;m} offers no advantage). Note
10407: that I specify the stack effect for the whole method (i.e. including
10408: the receiver object), not just for the code between @code{m:}
10409: and @code{;m}. You cannot use @code{exit} in
10410: @code{m:...;m}; instead, use
10411: @code{exitm}.@footnote{Moreover, for any word that calls
10412: @code{catch} and was defined before loading
10413: @code{objects.fs}, you have to redefine it like I redefined
10414: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.12 anton 10415:
1.78 anton 10416: @cindex @code{inst-var} usage
10417: You will frequently use sequences of the form @code{this
10418: @emph{field}} (in the example above: @code{this
10419: circle-radius}). If you use the field only in this way, you can
10420: define it with @code{inst-var} and eliminate the
10421: @code{this} before the field name. E.g., the @code{circle}
10422: class above could also be defined with:
1.6 pazsan 10423:
1.78 anton 10424: @example
10425: graphical class
10426: cell% inst-var radius
1.6 pazsan 10427:
1.78 anton 10428: m: ( x y circle -- )
10429: radius @@ draw-circle ;m
10430: overrides draw
1.6 pazsan 10431:
1.78 anton 10432: m: ( n-radius circle -- )
10433: radius ! ;m
10434: overrides construct
1.6 pazsan 10435:
1.78 anton 10436: end-class circle
10437: @end example
1.6 pazsan 10438:
1.78 anton 10439: @code{radius} can only be used in @code{circle} and its
10440: descendent classes and inside @code{m:...;m}.
1.6 pazsan 10441:
1.78 anton 10442: @cindex @code{inst-value} usage
10443: You can also define fields with @code{inst-value}, which is
10444: to @code{inst-var} what @code{value} is to
10445: @code{variable}. You can change the value of such a field with
10446: @code{[to-inst]}. E.g., we could also define the class
10447: @code{circle} like this:
1.44 crook 10448:
1.78 anton 10449: @example
10450: graphical class
10451: inst-value radius
1.6 pazsan 10452:
1.78 anton 10453: m: ( x y circle -- )
10454: radius draw-circle ;m
10455: overrides draw
1.44 crook 10456:
1.78 anton 10457: m: ( n-radius circle -- )
10458: [to-inst] radius ;m
10459: overrides construct
1.6 pazsan 10460:
1.78 anton 10461: end-class circle
10462: @end example
1.6 pazsan 10463:
1.78 anton 10464: @c !! :m is easy to confuse with m:. Another name would be better.
1.6 pazsan 10465:
1.78 anton 10466: @c Finally, you can define named methods with @code{:m}. One use of this
10467: @c feature is the definition of words that occur only in one class and are
10468: @c not intended to be overridden, but which still need method context
10469: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10470: @c would be bound frequently, if defined anonymously.
1.6 pazsan 10471:
10472:
1.78 anton 10473: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10474: @subsubsection Classes and Scoping
10475: @cindex classes and scoping
10476: @cindex scoping and classes
1.6 pazsan 10477:
1.78 anton 10478: Inheritance is frequent, unlike structure extension. This exacerbates
10479: the problem with the field name convention (@pxref{Structure Naming
10480: Convention}): One always has to remember in which class the field was
10481: originally defined; changing a part of the class structure would require
10482: changes for renaming in otherwise unaffected code.
1.6 pazsan 10483:
1.78 anton 10484: @cindex @code{inst-var} visibility
10485: @cindex @code{inst-value} visibility
10486: To solve this problem, I added a scoping mechanism (which was not in my
10487: original charter): A field defined with @code{inst-var} (or
10488: @code{inst-value}) is visible only in the class where it is defined and in
10489: the descendent classes of this class. Using such fields only makes
10490: sense in @code{m:}-defined methods in these classes anyway.
1.6 pazsan 10491:
1.78 anton 10492: This scoping mechanism allows us to use the unadorned field name,
10493: because name clashes with unrelated words become much less likely.
1.6 pazsan 10494:
1.78 anton 10495: @cindex @code{protected} discussion
10496: @cindex @code{private} discussion
10497: Once we have this mechanism, we can also use it for controlling the
10498: visibility of other words: All words defined after
10499: @code{protected} are visible only in the current class and its
10500: descendents. @code{public} restores the compilation
10501: (i.e. @code{current}) word list that was in effect before. If you
10502: have several @code{protected}s without an intervening
10503: @code{public} or @code{set-current}, @code{public}
10504: will restore the compilation word list in effect before the first of
10505: these @code{protected}s.
1.6 pazsan 10506:
1.78 anton 10507: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10508: @subsubsection Dividing classes
10509: @cindex Dividing classes
10510: @cindex @code{methods}...@code{end-methods}
1.6 pazsan 10511:
1.78 anton 10512: You may want to do the definition of methods separate from the
10513: definition of the class, its selectors, fields, and instance variables,
10514: i.e., separate the implementation from the definition. You can do this
10515: in the following way:
1.6 pazsan 10516:
1.78 anton 10517: @example
10518: graphical class
10519: inst-value radius
10520: end-class circle
1.6 pazsan 10521:
1.78 anton 10522: ... \ do some other stuff
1.6 pazsan 10523:
1.78 anton 10524: circle methods \ now we are ready
1.44 crook 10525:
1.78 anton 10526: m: ( x y circle -- )
10527: radius draw-circle ;m
10528: overrides draw
1.6 pazsan 10529:
1.78 anton 10530: m: ( n-radius circle -- )
10531: [to-inst] radius ;m
10532: overrides construct
1.44 crook 10533:
1.78 anton 10534: end-methods
10535: @end example
1.7 pazsan 10536:
1.78 anton 10537: You can use several @code{methods}...@code{end-methods} sections. The
10538: only things you can do to the class in these sections are: defining
10539: methods, and overriding the class's selectors. You must not define new
10540: selectors or fields.
1.7 pazsan 10541:
1.78 anton 10542: Note that you often have to override a selector before using it. In
10543: particular, you usually have to override @code{construct} with a new
10544: method before you can invoke @code{heap-new} and friends. E.g., you
10545: must not create a circle before the @code{overrides construct} sequence
10546: in the example above.
1.7 pazsan 10547:
1.78 anton 10548: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10549: @subsubsection Object Interfaces
10550: @cindex object interfaces
10551: @cindex interfaces for objects
1.7 pazsan 10552:
1.78 anton 10553: In this model you can only call selectors defined in the class of the
10554: receiving objects or in one of its ancestors. If you call a selector
10555: with a receiving object that is not in one of these classes, the
10556: result is undefined; if you are lucky, the program crashes
10557: immediately.
1.7 pazsan 10558:
1.78 anton 10559: @cindex selectors common to hardly-related classes
10560: Now consider the case when you want to have a selector (or several)
10561: available in two classes: You would have to add the selector to a
10562: common ancestor class, in the worst case to @code{object}. You
10563: may not want to do this, e.g., because someone else is responsible for
10564: this ancestor class.
1.7 pazsan 10565:
1.78 anton 10566: The solution for this problem is interfaces. An interface is a
10567: collection of selectors. If a class implements an interface, the
10568: selectors become available to the class and its descendents. A class
10569: can implement an unlimited number of interfaces. For the problem
10570: discussed above, we would define an interface for the selector(s), and
10571: both classes would implement the interface.
1.7 pazsan 10572:
1.78 anton 10573: As an example, consider an interface @code{storage} for
10574: writing objects to disk and getting them back, and a class
10575: @code{foo} that implements it. The code would look like this:
1.7 pazsan 10576:
1.78 anton 10577: @cindex @code{interface} usage
10578: @cindex @code{end-interface} usage
10579: @cindex @code{implementation} usage
10580: @example
10581: interface
10582: selector write ( file object -- )
10583: selector read1 ( file object -- )
10584: end-interface storage
1.13 pazsan 10585:
1.78 anton 10586: bar class
10587: storage implementation
1.13 pazsan 10588:
1.78 anton 10589: ... overrides write
10590: ... overrides read1
10591: ...
10592: end-class foo
10593: @end example
1.13 pazsan 10594:
1.78 anton 10595: @noindent
10596: (I would add a word @code{read} @i{( file -- object )} that uses
10597: @code{read1} internally, but that's beyond the point illustrated
10598: here.)
1.13 pazsan 10599:
1.78 anton 10600: Note that you cannot use @code{protected} in an interface; and
10601: of course you cannot define fields.
1.13 pazsan 10602:
1.78 anton 10603: In the Neon model, all selectors are available for all classes;
10604: therefore it does not need interfaces. The price you pay in this model
10605: is slower late binding, and therefore, added complexity to avoid late
10606: binding.
1.13 pazsan 10607:
1.78 anton 10608: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10609: @subsubsection @file{objects.fs} Implementation
10610: @cindex @file{objects.fs} implementation
1.13 pazsan 10611:
1.78 anton 10612: @cindex @code{object-map} discussion
10613: An object is a piece of memory, like one of the data structures
10614: described with @code{struct...end-struct}. It has a field
10615: @code{object-map} that points to the method map for the object's
10616: class.
1.13 pazsan 10617:
1.78 anton 10618: @cindex method map
10619: @cindex virtual function table
10620: The @emph{method map}@footnote{This is Self terminology; in C++
10621: terminology: virtual function table.} is an array that contains the
10622: execution tokens (@i{xt}s) of the methods for the object's class. Each
10623: selector contains an offset into a method map.
1.13 pazsan 10624:
1.78 anton 10625: @cindex @code{selector} implementation, class
10626: @code{selector} is a defining word that uses
10627: @code{CREATE} and @code{DOES>}. The body of the
10628: selector contains the offset; the @code{DOES>} action for a
10629: class selector is, basically:
1.8 pazsan 10630:
10631: @example
1.78 anton 10632: ( object addr ) @@ over object-map @@ + @@ execute
1.13 pazsan 10633: @end example
10634:
1.78 anton 10635: Since @code{object-map} is the first field of the object, it
10636: does not generate any code. As you can see, calling a selector has a
10637: small, constant cost.
1.26 crook 10638:
1.78 anton 10639: @cindex @code{current-interface} discussion
10640: @cindex class implementation and representation
10641: A class is basically a @code{struct} combined with a method
10642: map. During the class definition the alignment and size of the class
10643: are passed on the stack, just as with @code{struct}s, so
10644: @code{field} can also be used for defining class
10645: fields. However, passing more items on the stack would be
10646: inconvenient, so @code{class} builds a data structure in memory,
10647: which is accessed through the variable
10648: @code{current-interface}. After its definition is complete, the
10649: class is represented on the stack by a pointer (e.g., as parameter for
10650: a child class definition).
1.26 crook 10651:
1.78 anton 10652: A new class starts off with the alignment and size of its parent,
10653: and a copy of the parent's method map. Defining new fields extends the
10654: size and alignment; likewise, defining new selectors extends the
10655: method map. @code{overrides} just stores a new @i{xt} in the method
10656: map at the offset given by the selector.
1.13 pazsan 10657:
1.78 anton 10658: @cindex class binding, implementation
10659: Class binding just gets the @i{xt} at the offset given by the selector
10660: from the class's method map and @code{compile,}s (in the case of
10661: @code{[bind]}) it.
1.13 pazsan 10662:
1.78 anton 10663: @cindex @code{this} implementation
10664: @cindex @code{catch} and @code{this}
10665: @cindex @code{this} and @code{catch}
10666: I implemented @code{this} as a @code{value}. At the
10667: start of an @code{m:...;m} method the old @code{this} is
10668: stored to the return stack and restored at the end; and the object on
10669: the TOS is stored @code{TO this}. This technique has one
10670: disadvantage: If the user does not leave the method via
10671: @code{;m}, but via @code{throw} or @code{exit},
10672: @code{this} is not restored (and @code{exit} may
10673: crash). To deal with the @code{throw} problem, I have redefined
10674: @code{catch} to save and restore @code{this}; the same
10675: should be done with any word that can catch an exception. As for
10676: @code{exit}, I simply forbid it (as a replacement, there is
10677: @code{exitm}).
1.13 pazsan 10678:
1.78 anton 10679: @cindex @code{inst-var} implementation
10680: @code{inst-var} is just the same as @code{field}, with
10681: a different @code{DOES>} action:
1.13 pazsan 10682: @example
1.78 anton 10683: @@ this +
1.8 pazsan 10684: @end example
1.78 anton 10685: Similar for @code{inst-value}.
1.8 pazsan 10686:
1.78 anton 10687: @cindex class scoping implementation
10688: Each class also has a word list that contains the words defined with
10689: @code{inst-var} and @code{inst-value}, and its protected
10690: words. It also has a pointer to its parent. @code{class} pushes
10691: the word lists of the class and all its ancestors onto the search order stack,
10692: and @code{end-class} drops them.
1.20 pazsan 10693:
1.78 anton 10694: @cindex interface implementation
10695: An interface is like a class without fields, parent and protected
10696: words; i.e., it just has a method map. If a class implements an
10697: interface, its method map contains a pointer to the method map of the
10698: interface. The positive offsets in the map are reserved for class
10699: methods, therefore interface map pointers have negative
10700: offsets. Interfaces have offsets that are unique throughout the
10701: system, unlike class selectors, whose offsets are only unique for the
10702: classes where the selector is available (invokable).
1.20 pazsan 10703:
1.78 anton 10704: This structure means that interface selectors have to perform one
10705: indirection more than class selectors to find their method. Their body
10706: contains the interface map pointer offset in the class method map, and
10707: the method offset in the interface method map. The
10708: @code{does>} action for an interface selector is, basically:
1.20 pazsan 10709:
10710: @example
1.78 anton 10711: ( object selector-body )
10712: 2dup selector-interface @@ ( object selector-body object interface-offset )
10713: swap object-map @@ + @@ ( object selector-body map )
10714: swap selector-offset @@ + @@ execute
1.20 pazsan 10715: @end example
10716:
1.78 anton 10717: where @code{object-map} and @code{selector-offset} are
10718: first fields and generate no code.
1.20 pazsan 10719:
1.78 anton 10720: As a concrete example, consider the following code:
1.20 pazsan 10721:
10722: @example
1.78 anton 10723: interface
10724: selector if1sel1
10725: selector if1sel2
10726: end-interface if1
1.20 pazsan 10727:
1.78 anton 10728: object class
10729: if1 implementation
10730: selector cl1sel1
10731: cell% inst-var cl1iv1
1.20 pazsan 10732:
1.78 anton 10733: ' m1 overrides construct
10734: ' m2 overrides if1sel1
10735: ' m3 overrides if1sel2
10736: ' m4 overrides cl1sel2
10737: end-class cl1
1.20 pazsan 10738:
1.78 anton 10739: create obj1 object dict-new drop
10740: create obj2 cl1 dict-new drop
10741: @end example
1.20 pazsan 10742:
1.78 anton 10743: The data structure created by this code (including the data structure
10744: for @code{object}) is shown in the
10745: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10746: @comment TODO add this diagram..
1.20 pazsan 10747:
1.78 anton 10748: @node Objects Glossary, , Objects Implementation, Objects
10749: @subsubsection @file{objects.fs} Glossary
10750: @cindex @file{objects.fs} Glossary
1.20 pazsan 10751:
10752:
1.78 anton 10753: doc---objects-bind
10754: doc---objects-<bind>
10755: doc---objects-bind'
10756: doc---objects-[bind]
10757: doc---objects-class
10758: doc---objects-class->map
10759: doc---objects-class-inst-size
10760: doc---objects-class-override!
1.79 anton 10761: doc---objects-class-previous
10762: doc---objects-class>order
1.78 anton 10763: doc---objects-construct
10764: doc---objects-current'
10765: doc---objects-[current]
10766: doc---objects-current-interface
10767: doc---objects-dict-new
10768: doc---objects-end-class
10769: doc---objects-end-class-noname
10770: doc---objects-end-interface
10771: doc---objects-end-interface-noname
10772: doc---objects-end-methods
10773: doc---objects-exitm
10774: doc---objects-heap-new
10775: doc---objects-implementation
10776: doc---objects-init-object
10777: doc---objects-inst-value
10778: doc---objects-inst-var
10779: doc---objects-interface
10780: doc---objects-m:
10781: doc---objects-:m
10782: doc---objects-;m
10783: doc---objects-method
10784: doc---objects-methods
10785: doc---objects-object
10786: doc---objects-overrides
10787: doc---objects-[parent]
10788: doc---objects-print
10789: doc---objects-protected
10790: doc---objects-public
10791: doc---objects-selector
10792: doc---objects-this
10793: doc---objects-<to-inst>
10794: doc---objects-[to-inst]
10795: doc---objects-to-this
10796: doc---objects-xt-new
1.20 pazsan 10797:
10798:
1.78 anton 10799: @c -------------------------------------------------------------
10800: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10801: @subsection The @file{oof.fs} model
10802: @cindex oof
10803: @cindex object-oriented programming
1.20 pazsan 10804:
1.78 anton 10805: @cindex @file{objects.fs}
10806: @cindex @file{oof.fs}
1.20 pazsan 10807:
1.78 anton 10808: This section describes the @file{oof.fs} package.
1.20 pazsan 10809:
1.78 anton 10810: The package described in this section has been used in bigFORTH since 1991, and
10811: used for two large applications: a chromatographic system used to
10812: create new medicaments, and a graphic user interface library (MINOS).
1.20 pazsan 10813:
1.78 anton 10814: You can find a description (in German) of @file{oof.fs} in @cite{Object
10815: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10816: 10(2), 1994.
1.20 pazsan 10817:
1.78 anton 10818: @menu
10819: * Properties of the OOF model::
10820: * Basic OOF Usage::
10821: * The OOF base class::
10822: * Class Declaration::
10823: * Class Implementation::
10824: @end menu
1.20 pazsan 10825:
1.78 anton 10826: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10827: @subsubsection Properties of the @file{oof.fs} model
10828: @cindex @file{oof.fs} properties
1.20 pazsan 10829:
1.78 anton 10830: @itemize @bullet
10831: @item
10832: This model combines object oriented programming with information
10833: hiding. It helps you writing large application, where scoping is
10834: necessary, because it provides class-oriented scoping.
1.20 pazsan 10835:
1.78 anton 10836: @item
10837: Named objects, object pointers, and object arrays can be created,
10838: selector invocation uses the ``object selector'' syntax. Selector invocation
10839: to objects and/or selectors on the stack is a bit less convenient, but
10840: possible.
1.44 crook 10841:
1.78 anton 10842: @item
10843: Selector invocation and instance variable usage of the active object is
10844: straightforward, since both make use of the active object.
1.44 crook 10845:
1.78 anton 10846: @item
10847: Late binding is efficient and easy to use.
1.20 pazsan 10848:
1.78 anton 10849: @item
10850: State-smart objects parse selectors. However, extensibility is provided
10851: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.20 pazsan 10852:
1.78 anton 10853: @item
10854: An implementation in ANS Forth is available.
1.20 pazsan 10855:
1.78 anton 10856: @end itemize
1.23 crook 10857:
10858:
1.78 anton 10859: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10860: @subsubsection Basic @file{oof.fs} Usage
10861: @cindex @file{oof.fs} usage
1.23 crook 10862:
1.78 anton 10863: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.23 crook 10864:
1.78 anton 10865: You can define a class for graphical objects like this:
1.23 crook 10866:
1.78 anton 10867: @cindex @code{class} usage
10868: @cindex @code{class;} usage
10869: @cindex @code{method} usage
10870: @example
10871: object class graphical \ "object" is the parent class
10872: method draw ( x y graphical -- )
10873: class;
10874: @end example
1.23 crook 10875:
1.78 anton 10876: This code defines a class @code{graphical} with an
10877: operation @code{draw}. We can perform the operation
10878: @code{draw} on any @code{graphical} object, e.g.:
1.23 crook 10879:
1.78 anton 10880: @example
10881: 100 100 t-rex draw
10882: @end example
1.23 crook 10883:
1.78 anton 10884: @noindent
10885: where @code{t-rex} is an object or object pointer, created with e.g.
10886: @code{graphical : t-rex}.
1.23 crook 10887:
1.78 anton 10888: @cindex abstract class
10889: How do we create a graphical object? With the present definitions,
10890: we cannot create a useful graphical object. The class
10891: @code{graphical} describes graphical objects in general, but not
10892: any concrete graphical object type (C++ users would call it an
10893: @emph{abstract class}); e.g., there is no method for the selector
10894: @code{draw} in the class @code{graphical}.
1.23 crook 10895:
1.78 anton 10896: For concrete graphical objects, we define child classes of the
10897: class @code{graphical}, e.g.:
1.23 crook 10898:
1.78 anton 10899: @example
10900: graphical class circle \ "graphical" is the parent class
10901: cell var circle-radius
10902: how:
10903: : draw ( x y -- )
10904: circle-radius @@ draw-circle ;
1.23 crook 10905:
1.78 anton 10906: : init ( n-radius -- (
10907: circle-radius ! ;
10908: class;
10909: @end example
1.1 anton 10910:
1.78 anton 10911: Here we define a class @code{circle} as a child of @code{graphical},
10912: with a field @code{circle-radius}; it defines new methods for the
10913: selectors @code{draw} and @code{init} (@code{init} is defined in
10914: @code{object}, the parent class of @code{graphical}).
1.1 anton 10915:
1.78 anton 10916: Now we can create a circle in the dictionary with:
1.1 anton 10917:
1.78 anton 10918: @example
10919: 50 circle : my-circle
10920: @end example
1.21 crook 10921:
1.78 anton 10922: @noindent
10923: @code{:} invokes @code{init}, thus initializing the field
10924: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10925: with:
1.1 anton 10926:
1.78 anton 10927: @example
10928: 100 100 my-circle draw
10929: @end example
1.1 anton 10930:
1.78 anton 10931: @cindex selector invocation, restrictions
10932: @cindex class definition, restrictions
10933: Note: You can only invoke a selector if the receiving object belongs to
10934: the class where the selector was defined or one of its descendents;
10935: e.g., you can invoke @code{draw} only for objects belonging to
10936: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10937: mechanism will check if you try to invoke a selector that is not
10938: defined in this class hierarchy, so you'll get an error at compilation
10939: time.
1.1 anton 10940:
10941:
1.78 anton 10942: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10943: @subsubsection The @file{oof.fs} base class
10944: @cindex @file{oof.fs} base class
1.1 anton 10945:
1.78 anton 10946: When you define a class, you have to specify a parent class. So how do
10947: you start defining classes? There is one class available from the start:
10948: @code{object}. You have to use it as ancestor for all classes. It is the
10949: only class that has no parent. Classes are also objects, except that
10950: they don't have instance variables; class manipulation such as
10951: inheritance or changing definitions of a class is handled through
10952: selectors of the class @code{object}.
1.1 anton 10953:
1.78 anton 10954: @code{object} provides a number of selectors:
1.1 anton 10955:
1.78 anton 10956: @itemize @bullet
10957: @item
10958: @code{class} for subclassing, @code{definitions} to add definitions
10959: later on, and @code{class?} to get type informations (is the class a
10960: subclass of the class passed on the stack?).
1.1 anton 10961:
1.78 anton 10962: doc---object-class
10963: doc---object-definitions
10964: doc---object-class?
1.1 anton 10965:
10966:
1.26 crook 10967: @item
1.78 anton 10968: @code{init} and @code{dispose} as constructor and destructor of the
10969: object. @code{init} is invocated after the object's memory is allocated,
10970: while @code{dispose} also handles deallocation. Thus if you redefine
10971: @code{dispose}, you have to call the parent's dispose with @code{super
10972: dispose}, too.
10973:
10974: doc---object-init
10975: doc---object-dispose
10976:
1.1 anton 10977:
1.26 crook 10978: @item
1.78 anton 10979: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10980: @code{[]} to create named and unnamed objects and object arrays or
10981: object pointers.
10982:
10983: doc---object-new
10984: doc---object-new[]
10985: doc---object-:
10986: doc---object-ptr
10987: doc---object-asptr
10988: doc---object-[]
10989:
1.1 anton 10990:
1.26 crook 10991: @item
1.78 anton 10992: @code{::} and @code{super} for explicit scoping. You should use explicit
10993: scoping only for super classes or classes with the same set of instance
10994: variables. Explicitly-scoped selectors use early binding.
1.21 crook 10995:
1.78 anton 10996: doc---object-::
10997: doc---object-super
1.21 crook 10998:
10999:
1.26 crook 11000: @item
1.78 anton 11001: @code{self} to get the address of the object
1.21 crook 11002:
1.78 anton 11003: doc---object-self
1.21 crook 11004:
11005:
1.78 anton 11006: @item
11007: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11008: pointers and instance defers.
1.21 crook 11009:
1.78 anton 11010: doc---object-bind
11011: doc---object-bound
11012: doc---object-link
11013: doc---object-is
1.21 crook 11014:
11015:
1.78 anton 11016: @item
11017: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11018: form the stack, and @code{postpone} to generate selector invocation code.
1.21 crook 11019:
1.78 anton 11020: doc---object-'
11021: doc---object-postpone
1.21 crook 11022:
11023:
1.78 anton 11024: @item
11025: @code{with} and @code{endwith} to select the active object from the
11026: stack, and enable its scope. Using @code{with} and @code{endwith}
11027: also allows you to create code using selector @code{postpone} without being
11028: trapped by the state-smart objects.
1.21 crook 11029:
1.78 anton 11030: doc---object-with
11031: doc---object-endwith
1.21 crook 11032:
11033:
1.78 anton 11034: @end itemize
1.21 crook 11035:
1.78 anton 11036: @node Class Declaration, Class Implementation, The OOF base class, OOF
11037: @subsubsection Class Declaration
11038: @cindex class declaration
1.21 crook 11039:
1.78 anton 11040: @itemize @bullet
11041: @item
11042: Instance variables
1.21 crook 11043:
1.78 anton 11044: doc---oof-var
1.21 crook 11045:
11046:
1.78 anton 11047: @item
11048: Object pointers
1.21 crook 11049:
1.78 anton 11050: doc---oof-ptr
11051: doc---oof-asptr
1.21 crook 11052:
11053:
1.78 anton 11054: @item
11055: Instance defers
1.21 crook 11056:
1.78 anton 11057: doc---oof-defer
1.21 crook 11058:
11059:
1.78 anton 11060: @item
11061: Method selectors
1.21 crook 11062:
1.78 anton 11063: doc---oof-early
11064: doc---oof-method
1.21 crook 11065:
11066:
1.78 anton 11067: @item
11068: Class-wide variables
1.21 crook 11069:
1.78 anton 11070: doc---oof-static
1.21 crook 11071:
11072:
1.78 anton 11073: @item
11074: End declaration
1.1 anton 11075:
1.78 anton 11076: doc---oof-how:
11077: doc---oof-class;
1.21 crook 11078:
11079:
1.78 anton 11080: @end itemize
1.21 crook 11081:
1.78 anton 11082: @c -------------------------------------------------------------
11083: @node Class Implementation, , Class Declaration, OOF
11084: @subsubsection Class Implementation
11085: @cindex class implementation
1.21 crook 11086:
1.78 anton 11087: @c -------------------------------------------------------------
11088: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11089: @subsection The @file{mini-oof.fs} model
11090: @cindex mini-oof
1.21 crook 11091:
1.78 anton 11092: Gforth's third object oriented Forth package is a 12-liner. It uses a
1.79 anton 11093: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
1.78 anton 11094: and reduces to the bare minimum of features. This is based on a posting
11095: of Bernd Paysan in comp.lang.forth.
1.21 crook 11096:
1.78 anton 11097: @menu
11098: * Basic Mini-OOF Usage::
11099: * Mini-OOF Example::
11100: * Mini-OOF Implementation::
11101: @end menu
1.21 crook 11102:
1.78 anton 11103: @c -------------------------------------------------------------
11104: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11105: @subsubsection Basic @file{mini-oof.fs} Usage
11106: @cindex mini-oof usage
1.21 crook 11107:
1.78 anton 11108: There is a base class (@code{class}, which allocates one cell for the
11109: object pointer) plus seven other words: to define a method, a variable,
11110: a class; to end a class, to resolve binding, to allocate an object and
11111: to compile a class method.
11112: @comment TODO better description of the last one
1.26 crook 11113:
1.21 crook 11114:
1.78 anton 11115: doc-object
11116: doc-method
11117: doc-var
11118: doc-class
11119: doc-end-class
11120: doc-defines
11121: doc-new
11122: doc-::
1.21 crook 11123:
11124:
11125:
1.78 anton 11126: @c -------------------------------------------------------------
11127: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11128: @subsubsection Mini-OOF Example
11129: @cindex mini-oof example
1.1 anton 11130:
1.78 anton 11131: A short example shows how to use this package. This example, in slightly
11132: extended form, is supplied as @file{moof-exm.fs}
11133: @comment TODO could flesh this out with some comments from the Forthwrite article
1.20 pazsan 11134:
1.26 crook 11135: @example
1.78 anton 11136: object class
11137: method init
11138: method draw
11139: end-class graphical
1.26 crook 11140: @end example
1.20 pazsan 11141:
1.78 anton 11142: This code defines a class @code{graphical} with an
11143: operation @code{draw}. We can perform the operation
11144: @code{draw} on any @code{graphical} object, e.g.:
1.20 pazsan 11145:
1.26 crook 11146: @example
1.78 anton 11147: 100 100 t-rex draw
1.26 crook 11148: @end example
1.12 anton 11149:
1.78 anton 11150: where @code{t-rex} is an object or object pointer, created with e.g.
11151: @code{graphical new Constant t-rex}.
1.12 anton 11152:
1.78 anton 11153: For concrete graphical objects, we define child classes of the
11154: class @code{graphical}, e.g.:
1.12 anton 11155:
1.26 crook 11156: @example
11157: graphical class
1.78 anton 11158: cell var circle-radius
11159: end-class circle \ "graphical" is the parent class
1.12 anton 11160:
1.78 anton 11161: :noname ( x y -- )
11162: circle-radius @@ draw-circle ; circle defines draw
11163: :noname ( r -- )
11164: circle-radius ! ; circle defines init
11165: @end example
1.12 anton 11166:
1.78 anton 11167: There is no implicit init method, so we have to define one. The creation
11168: code of the object now has to call init explicitely.
1.21 crook 11169:
1.78 anton 11170: @example
11171: circle new Constant my-circle
11172: 50 my-circle init
1.12 anton 11173: @end example
11174:
1.78 anton 11175: It is also possible to add a function to create named objects with
11176: automatic call of @code{init}, given that all objects have @code{init}
11177: on the same place:
1.38 anton 11178:
1.78 anton 11179: @example
11180: : new: ( .. o "name" -- )
11181: new dup Constant init ;
11182: 80 circle new: large-circle
11183: @end example
1.12 anton 11184:
1.78 anton 11185: We can draw this new circle at (100,100) with:
1.12 anton 11186:
1.78 anton 11187: @example
11188: 100 100 my-circle draw
11189: @end example
1.12 anton 11190:
1.78 anton 11191: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11192: @subsubsection @file{mini-oof.fs} Implementation
1.12 anton 11193:
1.78 anton 11194: Object-oriented systems with late binding typically use a
11195: ``vtable''-approach: the first variable in each object is a pointer to a
11196: table, which contains the methods as function pointers. The vtable
11197: may also contain other information.
1.12 anton 11198:
1.79 anton 11199: So first, let's declare selectors:
1.37 anton 11200:
11201: @example
1.79 anton 11202: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
1.78 anton 11203: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
11204: @end example
1.37 anton 11205:
1.79 anton 11206: During selector declaration, the number of selectors and instance
11207: variables is on the stack (in address units). @code{method} creates one
11208: selector and increments the selector number. To execute a selector, it
1.78 anton 11209: takes the object, fetches the vtable pointer, adds the offset, and
1.79 anton 11210: executes the method @i{xt} stored there. Each selector takes the object
11211: it is invoked with as top of stack parameter; it passes the parameters
11212: (including the object) unchanged to the appropriate method which should
1.78 anton 11213: consume that object.
1.37 anton 11214:
1.78 anton 11215: Now, we also have to declare instance variables
1.37 anton 11216:
1.78 anton 11217: @example
1.79 anton 11218: : var ( m v size "name" -- m v' ) Create over , +
1.78 anton 11219: DOES> ( o -- addr ) @@ + ;
1.37 anton 11220: @end example
11221:
1.78 anton 11222: As before, a word is created with the current offset. Instance
11223: variables can have different sizes (cells, floats, doubles, chars), so
11224: all we do is take the size and add it to the offset. If your machine
11225: has alignment restrictions, put the proper @code{aligned} or
11226: @code{faligned} before the variable, to adjust the variable
11227: offset. That's why it is on the top of stack.
1.37 anton 11228:
1.78 anton 11229: We need a starting point (the base object) and some syntactic sugar:
1.37 anton 11230:
1.78 anton 11231: @example
11232: Create object 1 cells , 2 cells ,
1.79 anton 11233: : class ( class -- class selectors vars ) dup 2@@ ;
1.78 anton 11234: @end example
1.12 anton 11235:
1.78 anton 11236: For inheritance, the vtable of the parent object has to be
11237: copied when a new, derived class is declared. This gives all the
11238: methods of the parent class, which can be overridden, though.
1.12 anton 11239:
1.78 anton 11240: @example
1.79 anton 11241: : end-class ( class selectors vars "name" -- )
1.78 anton 11242: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11243: cell+ dup cell+ r> rot @@ 2 cells /string move ;
11244: @end example
1.12 anton 11245:
1.78 anton 11246: The first line creates the vtable, initialized with
11247: @code{noop}s. The second line is the inheritance mechanism, it
11248: copies the xts from the parent vtable.
1.12 anton 11249:
1.78 anton 11250: We still have no way to define new methods, let's do that now:
1.12 anton 11251:
1.26 crook 11252: @example
1.79 anton 11253: : defines ( xt class "name" -- ) ' >body @@ + ! ;
1.78 anton 11254: @end example
1.12 anton 11255:
1.78 anton 11256: To allocate a new object, we need a word, too:
1.12 anton 11257:
1.78 anton 11258: @example
11259: : new ( class -- o ) here over @@ allot swap over ! ;
1.12 anton 11260: @end example
11261:
1.78 anton 11262: Sometimes derived classes want to access the method of the
11263: parent object. There are two ways to achieve this with Mini-OOF:
11264: first, you could use named words, and second, you could look up the
11265: vtable of the parent object.
1.12 anton 11266:
1.78 anton 11267: @example
11268: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
11269: @end example
1.12 anton 11270:
11271:
1.78 anton 11272: Nothing can be more confusing than a good example, so here is
11273: one. First let's declare a text object (called
11274: @code{button}), that stores text and position:
1.12 anton 11275:
1.78 anton 11276: @example
11277: object class
11278: cell var text
11279: cell var len
11280: cell var x
11281: cell var y
11282: method init
11283: method draw
11284: end-class button
11285: @end example
1.12 anton 11286:
1.78 anton 11287: @noindent
11288: Now, implement the two methods, @code{draw} and @code{init}:
1.21 crook 11289:
1.26 crook 11290: @example
1.78 anton 11291: :noname ( o -- )
11292: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
11293: button defines draw
11294: :noname ( addr u o -- )
11295: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
11296: button defines init
1.26 crook 11297: @end example
1.12 anton 11298:
1.78 anton 11299: @noindent
11300: To demonstrate inheritance, we define a class @code{bold-button}, with no
1.79 anton 11301: new data and no new selectors:
1.78 anton 11302:
11303: @example
11304: button class
11305: end-class bold-button
1.12 anton 11306:
1.78 anton 11307: : bold 27 emit ." [1m" ;
11308: : normal 27 emit ." [0m" ;
11309: @end example
1.1 anton 11310:
1.78 anton 11311: @noindent
11312: The class @code{bold-button} has a different draw method to
11313: @code{button}, but the new method is defined in terms of the draw method
11314: for @code{button}:
1.20 pazsan 11315:
1.78 anton 11316: @example
11317: :noname bold [ button :: draw ] normal ; bold-button defines draw
11318: @end example
1.21 crook 11319:
1.78 anton 11320: @noindent
1.79 anton 11321: Finally, create two objects and apply selectors:
1.21 crook 11322:
1.26 crook 11323: @example
1.78 anton 11324: button new Constant foo
11325: s" thin foo" foo init
11326: page
11327: foo draw
11328: bold-button new Constant bar
11329: s" fat bar" bar init
11330: 1 bar y !
11331: bar draw
1.26 crook 11332: @end example
1.21 crook 11333:
11334:
1.78 anton 11335: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11336: @subsection Comparison with other object models
11337: @cindex comparison of object models
11338: @cindex object models, comparison
11339:
11340: Many object-oriented Forth extensions have been proposed (@cite{A survey
11341: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11342: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11343: relation of the object models described here to two well-known and two
11344: closely-related (by the use of method maps) models. Andras Zsoter
11345: helped us with this section.
11346:
11347: @cindex Neon model
11348: The most popular model currently seems to be the Neon model (see
11349: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11350: 1997) by Andrew McKewan) but this model has a number of limitations
11351: @footnote{A longer version of this critique can be
11352: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11353: Dimensions, May 1997) by Anton Ertl.}:
11354:
11355: @itemize @bullet
11356: @item
11357: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11358: to pass objects on the stack.
1.21 crook 11359:
1.78 anton 11360: @item
11361: It requires that the selector parses the input stream (at
1.79 anton 11362: compile time); this leads to reduced extensibility and to bugs that are
1.78 anton 11363: hard to find.
1.21 crook 11364:
1.78 anton 11365: @item
1.79 anton 11366: It allows using every selector on every object; this eliminates the
11367: need for interfaces, but makes it harder to create efficient
11368: implementations.
1.78 anton 11369: @end itemize
1.21 crook 11370:
1.78 anton 11371: @cindex Pountain's object-oriented model
11372: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11373: Press, London, 1987) by Dick Pountain. However, it is not really about
11374: object-oriented programming, because it hardly deals with late
11375: binding. Instead, it focuses on features like information hiding and
11376: overloading that are characteristic of modular languages like Ada (83).
1.26 crook 11377:
1.78 anton 11378: @cindex Zsoter's object-oriented model
1.79 anton 11379: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11380: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11381: describes a model that makes heavy use of an active object (like
11382: @code{this} in @file{objects.fs}): The active object is not only used
11383: for accessing all fields, but also specifies the receiving object of
11384: every selector invocation; you have to change the active object
11385: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11386: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11387: the method entry point is unnecessary with Zsoter's model, because the
11388: receiving object is the active object already. On the other hand, the
11389: explicit change is absolutely necessary in that model, because otherwise
11390: no one could ever change the active object. An ANS Forth implementation
11391: of this model is available through
11392: @uref{http://www.forth.org/oopf.html}.
1.21 crook 11393:
1.78 anton 11394: @cindex @file{oof.fs}, differences to other models
11395: The @file{oof.fs} model combines information hiding and overloading
11396: resolution (by keeping names in various word lists) with object-oriented
11397: programming. It sets the active object implicitly on method entry, but
11398: also allows explicit changing (with @code{>o...o>} or with
11399: @code{with...endwith}). It uses parsing and state-smart objects and
11400: classes for resolving overloading and for early binding: the object or
11401: class parses the selector and determines the method from this. If the
11402: selector is not parsed by an object or class, it performs a call to the
11403: selector for the active object (late binding), like Zsoter's model.
11404: Fields are always accessed through the active object. The big
11405: disadvantage of this model is the parsing and the state-smartness, which
11406: reduces extensibility and increases the opportunities for subtle bugs;
11407: essentially, you are only safe if you never tick or @code{postpone} an
11408: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.21 crook 11409:
1.78 anton 11410: @cindex @file{mini-oof.fs}, differences to other models
11411: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11412: version of the @file{objects.fs} model, but syntactically it is a
11413: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.21 crook 11414:
11415:
1.78 anton 11416: @c -------------------------------------------------------------
11417: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
11418: @section Programming Tools
11419: @cindex programming tools
1.21 crook 11420:
1.78 anton 11421: @c !! move this and assembler down below OO stuff.
1.21 crook 11422:
1.78 anton 11423: @menu
11424: * Examining::
11425: * Forgetting words::
11426: * Debugging:: Simple and quick.
11427: * Assertions:: Making your programs self-checking.
11428: * Singlestep Debugger:: Executing your program word by word.
11429: @end menu
1.21 crook 11430:
1.78 anton 11431: @node Examining, Forgetting words, Programming Tools, Programming Tools
11432: @subsection Examining data and code
11433: @cindex examining data and code
11434: @cindex data examination
11435: @cindex code examination
1.44 crook 11436:
1.78 anton 11437: The following words inspect the stack non-destructively:
1.21 crook 11438:
1.78 anton 11439: doc-.s
11440: doc-f.s
1.44 crook 11441:
1.78 anton 11442: There is a word @code{.r} but it does @i{not} display the return stack!
11443: It is used for formatted numeric output (@pxref{Simple numeric output}).
1.21 crook 11444:
1.78 anton 11445: doc-depth
11446: doc-fdepth
11447: doc-clearstack
1.21 crook 11448:
1.78 anton 11449: The following words inspect memory.
1.21 crook 11450:
1.78 anton 11451: doc-?
11452: doc-dump
1.21 crook 11453:
1.78 anton 11454: And finally, @code{see} allows to inspect code:
1.21 crook 11455:
1.78 anton 11456: doc-see
11457: doc-xt-see
1.21 crook 11458:
1.78 anton 11459: @node Forgetting words, Debugging, Examining, Programming Tools
11460: @subsection Forgetting words
11461: @cindex words, forgetting
11462: @cindex forgeting words
1.21 crook 11463:
1.78 anton 11464: @c anton: other, maybe better places for this subsection: Defining Words;
11465: @c Dictionary allocation. At least a reference should be there.
1.21 crook 11466:
1.78 anton 11467: Forth allows you to forget words (and everything that was alloted in the
11468: dictonary after them) in a LIFO manner.
1.21 crook 11469:
1.78 anton 11470: doc-marker
1.21 crook 11471:
1.78 anton 11472: The most common use of this feature is during progam development: when
11473: you change a source file, forget all the words it defined and load it
11474: again (since you also forget everything defined after the source file
11475: was loaded, you have to reload that, too). Note that effects like
11476: storing to variables and destroyed system words are not undone when you
11477: forget words. With a system like Gforth, that is fast enough at
11478: starting up and compiling, I find it more convenient to exit and restart
11479: Gforth, as this gives me a clean slate.
1.21 crook 11480:
1.78 anton 11481: Here's an example of using @code{marker} at the start of a source file
11482: that you are debugging; it ensures that you only ever have one copy of
11483: the file's definitions compiled at any time:
1.21 crook 11484:
1.78 anton 11485: @example
11486: [IFDEF] my-code
11487: my-code
11488: [ENDIF]
1.26 crook 11489:
1.78 anton 11490: marker my-code
11491: init-included-files
1.21 crook 11492:
1.78 anton 11493: \ .. definitions start here
11494: \ .
11495: \ .
11496: \ end
11497: @end example
1.21 crook 11498:
1.26 crook 11499:
1.78 anton 11500: @node Debugging, Assertions, Forgetting words, Programming Tools
11501: @subsection Debugging
11502: @cindex debugging
1.21 crook 11503:
1.78 anton 11504: Languages with a slow edit/compile/link/test development loop tend to
11505: require sophisticated tracing/stepping debuggers to facilate debugging.
1.21 crook 11506:
1.78 anton 11507: A much better (faster) way in fast-compiling languages is to add
11508: printing code at well-selected places, let the program run, look at
11509: the output, see where things went wrong, add more printing code, etc.,
11510: until the bug is found.
1.21 crook 11511:
1.78 anton 11512: The simple debugging aids provided in @file{debugs.fs}
11513: are meant to support this style of debugging.
1.21 crook 11514:
1.78 anton 11515: The word @code{~~} prints debugging information (by default the source
11516: location and the stack contents). It is easy to insert. If you use Emacs
11517: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11518: query-replace them with nothing). The deferred words
11519: @code{printdebugdata} and @code{printdebugline} control the output of
11520: @code{~~}. The default source location output format works well with
11521: Emacs' compilation mode, so you can step through the program at the
11522: source level using @kbd{C-x `} (the advantage over a stepping debugger
11523: is that you can step in any direction and you know where the crash has
11524: happened or where the strange data has occurred).
1.21 crook 11525:
1.78 anton 11526: doc-~~
11527: doc-printdebugdata
11528: doc-printdebugline
1.21 crook 11529:
1.78 anton 11530: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11531: @subsection Assertions
11532: @cindex assertions
1.21 crook 11533:
1.78 anton 11534: It is a good idea to make your programs self-checking, especially if you
11535: make an assumption that may become invalid during maintenance (for
11536: example, that a certain field of a data structure is never zero). Gforth
11537: supports @dfn{assertions} for this purpose. They are used like this:
1.21 crook 11538:
11539: @example
1.78 anton 11540: assert( @i{flag} )
1.26 crook 11541: @end example
11542:
1.78 anton 11543: The code between @code{assert(} and @code{)} should compute a flag, that
11544: should be true if everything is alright and false otherwise. It should
11545: not change anything else on the stack. The overall stack effect of the
11546: assertion is @code{( -- )}. E.g.
1.21 crook 11547:
1.26 crook 11548: @example
1.78 anton 11549: assert( 1 1 + 2 = ) \ what we learn in school
11550: assert( dup 0<> ) \ assert that the top of stack is not zero
11551: assert( false ) \ this code should not be reached
1.21 crook 11552: @end example
11553:
1.78 anton 11554: The need for assertions is different at different times. During
11555: debugging, we want more checking, in production we sometimes care more
11556: for speed. Therefore, assertions can be turned off, i.e., the assertion
11557: becomes a comment. Depending on the importance of an assertion and the
11558: time it takes to check it, you may want to turn off some assertions and
11559: keep others turned on. Gforth provides several levels of assertions for
11560: this purpose:
11561:
11562:
11563: doc-assert0(
11564: doc-assert1(
11565: doc-assert2(
11566: doc-assert3(
11567: doc-assert(
11568: doc-)
1.21 crook 11569:
11570:
1.78 anton 11571: The variable @code{assert-level} specifies the highest assertions that
11572: are turned on. I.e., at the default @code{assert-level} of one,
11573: @code{assert0(} and @code{assert1(} assertions perform checking, while
11574: @code{assert2(} and @code{assert3(} assertions are treated as comments.
1.26 crook 11575:
1.78 anton 11576: The value of @code{assert-level} is evaluated at compile-time, not at
11577: run-time. Therefore you cannot turn assertions on or off at run-time;
11578: you have to set the @code{assert-level} appropriately before compiling a
11579: piece of code. You can compile different pieces of code at different
11580: @code{assert-level}s (e.g., a trusted library at level 1 and
11581: newly-written code at level 3).
1.26 crook 11582:
11583:
1.78 anton 11584: doc-assert-level
1.26 crook 11585:
11586:
1.78 anton 11587: If an assertion fails, a message compatible with Emacs' compilation mode
11588: is produced and the execution is aborted (currently with @code{ABORT"}.
11589: If there is interest, we will introduce a special throw code. But if you
11590: intend to @code{catch} a specific condition, using @code{throw} is
11591: probably more appropriate than an assertion).
1.44 crook 11592:
1.78 anton 11593: Definitions in ANS Forth for these assertion words are provided
11594: in @file{compat/assert.fs}.
1.26 crook 11595:
1.44 crook 11596:
1.78 anton 11597: @node Singlestep Debugger, , Assertions, Programming Tools
11598: @subsection Singlestep Debugger
11599: @cindex singlestep Debugger
11600: @cindex debugging Singlestep
1.44 crook 11601:
1.78 anton 11602: When you create a new word there's often the need to check whether it
11603: behaves correctly or not. You can do this by typing @code{dbg
11604: badword}. A debug session might look like this:
1.26 crook 11605:
1.78 anton 11606: @example
11607: : badword 0 DO i . LOOP ; ok
11608: 2 dbg badword
11609: : badword
11610: Scanning code...
1.44 crook 11611:
1.78 anton 11612: Nesting debugger ready!
1.44 crook 11613:
1.78 anton 11614: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11615: 400D4740 8049F68 DO -> [ 0 ]
11616: 400D4744 804A0C8 i -> [ 1 ] 00000
11617: 400D4748 400C5E60 . -> 0 [ 0 ]
11618: 400D474C 8049D0C LOOP -> [ 0 ]
11619: 400D4744 804A0C8 i -> [ 1 ] 00001
11620: 400D4748 400C5E60 . -> 1 [ 0 ]
11621: 400D474C 8049D0C LOOP -> [ 0 ]
11622: 400D4758 804B384 ; -> ok
11623: @end example
1.21 crook 11624:
1.78 anton 11625: Each line displayed is one step. You always have to hit return to
11626: execute the next word that is displayed. If you don't want to execute
11627: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11628: an overview what keys are available:
1.44 crook 11629:
1.78 anton 11630: @table @i
1.44 crook 11631:
1.78 anton 11632: @item @key{RET}
11633: Next; Execute the next word.
1.21 crook 11634:
1.78 anton 11635: @item n
11636: Nest; Single step through next word.
1.44 crook 11637:
1.78 anton 11638: @item u
11639: Unnest; Stop debugging and execute rest of word. If we got to this word
11640: with nest, continue debugging with the calling word.
1.44 crook 11641:
1.78 anton 11642: @item d
11643: Done; Stop debugging and execute rest.
1.21 crook 11644:
1.78 anton 11645: @item s
11646: Stop; Abort immediately.
1.44 crook 11647:
1.78 anton 11648: @end table
1.44 crook 11649:
1.78 anton 11650: Debugging large application with this mechanism is very difficult, because
11651: you have to nest very deeply into the program before the interesting part
11652: begins. This takes a lot of time.
1.26 crook 11653:
1.78 anton 11654: To do it more directly put a @code{BREAK:} command into your source code.
11655: When program execution reaches @code{BREAK:} the single step debugger is
11656: invoked and you have all the features described above.
1.44 crook 11657:
1.78 anton 11658: If you have more than one part to debug it is useful to know where the
11659: program has stopped at the moment. You can do this by the
11660: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11661: string is typed out when the ``breakpoint'' is reached.
1.44 crook 11662:
1.26 crook 11663:
1.78 anton 11664: doc-dbg
11665: doc-break:
11666: doc-break"
1.44 crook 11667:
11668:
1.26 crook 11669:
1.78 anton 11670: @c -------------------------------------------------------------
11671: @node Assembler and Code Words, Threading Words, Programming Tools, Words
11672: @section Assembler and Code Words
11673: @cindex assembler
11674: @cindex code words
1.44 crook 11675:
1.78 anton 11676: @menu
11677: * Code and ;code::
11678: * Common Assembler:: Assembler Syntax
11679: * Common Disassembler::
11680: * 386 Assembler:: Deviations and special cases
11681: * Alpha Assembler:: Deviations and special cases
11682: * MIPS assembler:: Deviations and special cases
11683: * Other assemblers:: How to write them
11684: @end menu
1.21 crook 11685:
1.78 anton 11686: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
11687: @subsection @code{Code} and @code{;code}
1.26 crook 11688:
1.78 anton 11689: Gforth provides some words for defining primitives (words written in
11690: machine code), and for defining the machine-code equivalent of
11691: @code{DOES>}-based defining words. However, the machine-independent
11692: nature of Gforth poses a few problems: First of all, Gforth runs on
11693: several architectures, so it can provide no standard assembler. What's
11694: worse is that the register allocation not only depends on the processor,
11695: but also on the @code{gcc} version and options used.
1.44 crook 11696:
1.78 anton 11697: The words that Gforth offers encapsulate some system dependences (e.g.,
11698: the header structure), so a system-independent assembler may be used in
11699: Gforth. If you do not have an assembler, you can compile machine code
11700: directly with @code{,} and @code{c,}@footnote{This isn't portable,
11701: because these words emit stuff in @i{data} space; it works because
11702: Gforth has unified code/data spaces. Assembler isn't likely to be
11703: portable anyway.}.
1.21 crook 11704:
1.44 crook 11705:
1.78 anton 11706: doc-assembler
11707: doc-init-asm
11708: doc-code
11709: doc-end-code
11710: doc-;code
11711: doc-flush-icache
1.44 crook 11712:
1.21 crook 11713:
1.78 anton 11714: If @code{flush-icache} does not work correctly, @code{code} words
11715: etc. will not work (reliably), either.
1.44 crook 11716:
1.78 anton 11717: The typical usage of these @code{code} words can be shown most easily by
11718: analogy to the equivalent high-level defining words:
1.44 crook 11719:
1.78 anton 11720: @example
11721: : foo code foo
11722: <high-level Forth words> <assembler>
11723: ; end-code
11724:
11725: : bar : bar
11726: <high-level Forth words> <high-level Forth words>
11727: CREATE CREATE
11728: <high-level Forth words> <high-level Forth words>
11729: DOES> ;code
11730: <high-level Forth words> <assembler>
11731: ; end-code
11732: @end example
1.21 crook 11733:
1.78 anton 11734: @c anton: the following stuff is also in "Common Assembler", in less detail.
1.44 crook 11735:
1.78 anton 11736: @cindex registers of the inner interpreter
11737: In the assembly code you will want to refer to the inner interpreter's
11738: registers (e.g., the data stack pointer) and you may want to use other
11739: registers for temporary storage. Unfortunately, the register allocation
11740: is installation-dependent.
1.44 crook 11741:
1.78 anton 11742: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
11743: (return stack pointer) are in different places in @code{gforth} and
11744: @code{gforth-fast}. This means that you cannot write a @code{NEXT}
11745: routine that works on both versions; so for doing @code{NEXT}, I
11746: recomment jumping to @code{' noop >code-address}, which contains nothing
11747: but a @code{NEXT}.
1.21 crook 11748:
1.78 anton 11749: For general accesses to the inner interpreter's registers, the easiest
11750: solution is to use explicit register declarations (@pxref{Explicit Reg
11751: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
11752: all of the inner interpreter's registers: You have to compile Gforth
11753: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
11754: the appropriate declarations must be present in the @code{machine.h}
11755: file (see @code{mips.h} for an example; you can find a full list of all
11756: declarable register symbols with @code{grep register engine.c}). If you
11757: give explicit registers to all variables that are declared at the
11758: beginning of @code{engine()}, you should be able to use the other
11759: caller-saved registers for temporary storage. Alternatively, you can use
11760: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
11761: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
11762: reserve a register (however, this restriction on register allocation may
11763: slow Gforth significantly).
1.44 crook 11764:
1.78 anton 11765: If this solution is not viable (e.g., because @code{gcc} does not allow
11766: you to explicitly declare all the registers you need), you have to find
11767: out by looking at the code where the inner interpreter's registers
11768: reside and which registers can be used for temporary storage. You can
11769: get an assembly listing of the engine's code with @code{make engine.s}.
1.44 crook 11770:
1.78 anton 11771: In any case, it is good practice to abstract your assembly code from the
11772: actual register allocation. E.g., if the data stack pointer resides in
11773: register @code{$17}, create an alias for this register called @code{sp},
11774: and use that in your assembly code.
1.21 crook 11775:
1.78 anton 11776: @cindex code words, portable
11777: Another option for implementing normal and defining words efficiently
11778: is to add the desired functionality to the source of Gforth. For normal
11779: words you just have to edit @file{primitives} (@pxref{Automatic
11780: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
11781: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
11782: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.44 crook 11783:
1.78 anton 11784: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
11785: @subsection Common Assembler
1.44 crook 11786:
1.78 anton 11787: The assemblers in Gforth generally use a postfix syntax, i.e., the
11788: instruction name follows the operands.
1.21 crook 11789:
1.78 anton 11790: The operands are passed in the usual order (the same that is used in the
11791: manual of the architecture). Since they all are Forth words, they have
11792: to be separated by spaces; you can also use Forth words to compute the
11793: operands.
1.44 crook 11794:
1.78 anton 11795: The instruction names usually end with a @code{,}. This makes it easier
11796: to visually separate instructions if you put several of them on one
11797: line; it also avoids shadowing other Forth words (e.g., @code{and}).
1.21 crook 11798:
1.78 anton 11799: Registers are usually specified by number; e.g., (decimal) @code{11}
11800: specifies registers R11 and F11 on the Alpha architecture (which one,
11801: depends on the instruction). The usual names are also available, e.g.,
11802: @code{s2} for R11 on Alpha.
1.21 crook 11803:
1.78 anton 11804: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
11805: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
11806: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
11807: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
11808: conditions are specified in a way specific to each assembler.
1.1 anton 11809:
1.78 anton 11810: Note that the register assignments of the Gforth engine can change
11811: between Gforth versions, or even between different compilations of the
11812: same Gforth version (e.g., if you use a different GCC version). So if
11813: you want to refer to Gforth's registers (e.g., the stack pointer or
11814: TOS), I recommend defining your own words for refering to these
11815: registers, and using them later on; then you can easily adapt to a
11816: changed register assignment. The stability of the register assignment
11817: is usually better if you build Gforth with @code{--enable-force-reg}.
1.1 anton 11818:
1.78 anton 11819: In particular, the return stack pointer and the instruction pointer are
11820: in memory in @code{gforth}, and usually in registers in
11821: @code{gforth-fast}. The most common use of these registers is to
11822: dispatch to the next word (the @code{next} routine). A portable way to
11823: do this is to jump to @code{' noop >code-address} (of course, this is
11824: less efficient than integrating the @code{next} code and scheduling it
11825: well).
1.1 anton 11826:
1.78 anton 11827: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
11828: @subsection Common Disassembler
1.1 anton 11829:
1.78 anton 11830: You can disassemble a @code{code} word with @code{see}
11831: (@pxref{Debugging}). You can disassemble a section of memory with
1.1 anton 11832:
1.78 anton 11833: doc-disasm
1.44 crook 11834:
1.78 anton 11835: The disassembler generally produces output that can be fed into the
11836: assembler (i.e., same syntax, etc.). It also includes additional
11837: information in comments. In particular, the address of the instruction
11838: is given in a comment before the instruction.
1.1 anton 11839:
1.78 anton 11840: @code{See} may display more or less than the actual code of the word,
11841: because the recognition of the end of the code is unreliable. You can
11842: use @code{disasm} if it did not display enough. It may display more, if
11843: the code word is not immediately followed by a named word. If you have
11844: something else there, you can follow the word with @code{align last @ ,}
11845: to ensure that the end is recognized.
1.21 crook 11846:
1.78 anton 11847: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
11848: @subsection 386 Assembler
1.44 crook 11849:
1.78 anton 11850: The 386 assembler included in Gforth was written by Bernd Paysan, it's
11851: available under GPL, and originally part of bigFORTH.
1.21 crook 11852:
1.78 anton 11853: The 386 disassembler included in Gforth was written by Andrew McKewan
11854: and is in the public domain.
1.21 crook 11855:
1.91 ! anton 11856: The disassembler displays code in an Intel-like prefix syntax.
1.21 crook 11857:
1.78 anton 11858: The assembler uses a postfix syntax with reversed parameters.
1.1 anton 11859:
1.78 anton 11860: The assembler includes all instruction of the Athlon, i.e. 486 core
11861: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
11862: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
11863: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
1.1 anton 11864:
1.78 anton 11865: There are several prefixes to switch between different operation sizes,
11866: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
11867: double-word accesses. Addressing modes can be switched with @code{.wa}
11868: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
11869: need a prefix for byte register names (@code{AL} et al).
1.1 anton 11870:
1.78 anton 11871: For floating point operations, the prefixes are @code{.fs} (IEEE
11872: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
11873: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
1.21 crook 11874:
1.78 anton 11875: The MMX opcodes don't have size prefixes, they are spelled out like in
11876: the Intel assembler. Instead of move from and to memory, there are
11877: PLDQ/PLDD and PSTQ/PSTD.
1.21 crook 11878:
1.78 anton 11879: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
11880: ax. Immediate values are indicated by postfixing them with @code{#},
1.91 ! anton 11881: e.g., @code{3 #}. Here are some examples of addressing modes in various
! 11882: syntaxes:
1.21 crook 11883:
1.26 crook 11884: @example
1.91 ! anton 11885: Gforth Intel (NASM) AT&T (gas) Name
! 11886: .w ax ax %ax register (16 bit)
! 11887: ax eax %eax register (32 bit)
! 11888: 3 # offset 3 $3 immediate
! 11889: 1000 #) byte ptr 1000 1000 displacement
! 11890: bx ) [ebx] (%ebx) base
! 11891: 100 di d) 100[edi] 100(%edi) base+displacement
! 11892: 20 ax *4 i#) 20[eax*4] 20(,%eax,4) (index*scale)+displacement
! 11893: di ax *4 i) [edi][eax*4] (%edi,%eax,4) base+(index*scale)
! 11894: 4 bx cx di) 4[ebx][ecx] 4(%ebx,%ecx) base+index+displacement
! 11895: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
! 11896: @end example
! 11897:
! 11898: You can use @code{L)} and @code{LI)} instead of @code{D)} and
! 11899: @code{DI)} to enforce 32-bit displacement fields (useful for
! 11900: later patching).
1.21 crook 11901:
1.78 anton 11902: Some example of instructions are:
1.1 anton 11903:
11904: @example
1.78 anton 11905: ax bx mov \ move ebx,eax
11906: 3 # ax mov \ mov eax,3
11907: 100 di ) ax mov \ mov eax,100[edi]
11908: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
11909: .w ax bx mov \ mov bx,ax
1.1 anton 11910: @end example
11911:
1.78 anton 11912: The following forms are supported for binary instructions:
1.1 anton 11913:
11914: @example
1.78 anton 11915: <reg> <reg> <inst>
11916: <n> # <reg> <inst>
11917: <mem> <reg> <inst>
11918: <reg> <mem> <inst>
1.1 anton 11919: @end example
11920:
1.78 anton 11921: Immediate to memory is not supported. The shift/rotate syntax is:
1.1 anton 11922:
1.26 crook 11923: @example
1.78 anton 11924: <reg/mem> 1 # shl \ shortens to shift without immediate
11925: <reg/mem> 4 # shl
11926: <reg/mem> cl shl
1.26 crook 11927: @end example
1.1 anton 11928:
1.78 anton 11929: Precede string instructions (@code{movs} etc.) with @code{.b} to get
11930: the byte version.
1.1 anton 11931:
1.78 anton 11932: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
11933: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
11934: pc < >= <= >}. (Note that most of these words shadow some Forth words
11935: when @code{assembler} is in front of @code{forth} in the search path,
11936: e.g., in @code{code} words). Currently the control structure words use
11937: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
11938: to shuffle them (you can also use @code{swap} etc.).
1.21 crook 11939:
1.78 anton 11940: Here is an example of a @code{code} word (assumes that the stack pointer
11941: is in esi and the TOS is in ebx):
1.21 crook 11942:
1.26 crook 11943: @example
1.78 anton 11944: code my+ ( n1 n2 -- n )
11945: 4 si D) bx add
11946: 4 # si add
11947: Next
11948: end-code
1.26 crook 11949: @end example
1.21 crook 11950:
1.78 anton 11951: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
11952: @subsection Alpha Assembler
1.21 crook 11953:
1.78 anton 11954: The Alpha assembler and disassembler were originally written by Bernd
11955: Thallner.
1.26 crook 11956:
1.78 anton 11957: The register names @code{a0}--@code{a5} are not available to avoid
11958: shadowing hex numbers.
1.2 jwilke 11959:
1.78 anton 11960: Immediate forms of arithmetic instructions are distinguished by a
11961: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
11962: does not count as arithmetic instruction).
1.2 jwilke 11963:
1.78 anton 11964: You have to specify all operands to an instruction, even those that
11965: other assemblers consider optional, e.g., the destination register for
11966: @code{br,}, or the destination register and hint for @code{jmp,}.
1.2 jwilke 11967:
1.78 anton 11968: You can specify conditions for @code{if,} by removing the first @code{b}
11969: and the trailing @code{,} from a branch with a corresponding name; e.g.,
1.2 jwilke 11970:
1.26 crook 11971: @example
1.78 anton 11972: 11 fgt if, \ if F11>0e
11973: ...
11974: endif,
1.26 crook 11975: @end example
1.2 jwilke 11976:
1.78 anton 11977: @code{fbgt,} gives @code{fgt}.
11978:
11979: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
11980: @subsection MIPS assembler
1.2 jwilke 11981:
1.78 anton 11982: The MIPS assembler was originally written by Christian Pirker.
1.2 jwilke 11983:
1.78 anton 11984: Currently the assembler and disassembler only cover the MIPS-I
11985: architecture (R3000), and don't support FP instructions.
1.2 jwilke 11986:
1.78 anton 11987: The register names @code{$a0}--@code{$a3} are not available to avoid
11988: shadowing hex numbers.
1.2 jwilke 11989:
1.78 anton 11990: Because there is no way to distinguish registers from immediate values,
11991: you have to explicitly use the immediate forms of instructions, i.e.,
11992: @code{addiu,}, not just @code{addu,} (@command{as} does this
11993: implicitly).
1.2 jwilke 11994:
1.78 anton 11995: If the architecture manual specifies several formats for the instruction
11996: (e.g., for @code{jalr,}), you usually have to use the one with more
11997: arguments (i.e., two for @code{jalr,}). When in doubt, see
11998: @code{arch/mips/testasm.fs} for an example of correct use.
1.2 jwilke 11999:
1.78 anton 12000: Branches and jumps in the MIPS architecture have a delay slot. You have
12001: to fill it yourself (the simplest way is to use @code{nop,}), the
12002: assembler does not do it for you (unlike @command{as}). Even
12003: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
12004: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
12005: and @code{then,} just specify branch targets, they are not affected.
1.2 jwilke 12006:
1.78 anton 12007: Note that you must not put branches, jumps, or @code{li,} into the delay
12008: slot: @code{li,} may expand to several instructions, and control flow
12009: instructions may not be put into the branch delay slot in any case.
1.2 jwilke 12010:
1.78 anton 12011: For branches the argument specifying the target is a relative address;
12012: You have to add the address of the delay slot to get the absolute
12013: address.
1.1 anton 12014:
1.78 anton 12015: The MIPS architecture also has load delay slots and restrictions on
12016: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
12017: yourself to satisfy these restrictions, the assembler does not do it for
12018: you.
1.1 anton 12019:
1.78 anton 12020: You can specify the conditions for @code{if,} etc. by taking a
12021: conditional branch and leaving away the @code{b} at the start and the
12022: @code{,} at the end. E.g.,
1.1 anton 12023:
1.26 crook 12024: @example
1.78 anton 12025: 4 5 eq if,
12026: ... \ do something if $4 equals $5
12027: then,
1.26 crook 12028: @end example
1.1 anton 12029:
1.78 anton 12030: @node Other assemblers, , MIPS assembler, Assembler and Code Words
12031: @subsection Other assemblers
12032:
12033: If you want to contribute another assembler/disassembler, please contact
12034: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
12035: already. If you are writing them from scratch, please use a similar
12036: syntax style as the one we use (i.e., postfix, commas at the end of the
12037: instruction names, @pxref{Common Assembler}); make the output of the
12038: disassembler be valid input for the assembler, and keep the style
12039: similar to the style we used.
12040:
12041: Hints on implementation: The most important part is to have a good test
12042: suite that contains all instructions. Once you have that, the rest is
12043: easy. For actual coding you can take a look at
12044: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
12045: the assembler and disassembler, avoiding redundancy and some potential
12046: bugs. You can also look at that file (and @pxref{Advanced does> usage
12047: example}) to get ideas how to factor a disassembler.
12048:
12049: Start with the disassembler, because it's easier to reuse data from the
12050: disassembler for the assembler than the other way round.
1.1 anton 12051:
1.78 anton 12052: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
12053: how simple it can be.
1.1 anton 12054:
1.78 anton 12055: @c -------------------------------------------------------------
12056: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
12057: @section Threading Words
12058: @cindex threading words
1.1 anton 12059:
1.78 anton 12060: @cindex code address
12061: These words provide access to code addresses and other threading stuff
12062: in Gforth (and, possibly, other interpretive Forths). It more or less
12063: abstracts away the differences between direct and indirect threading
12064: (and, for direct threading, the machine dependences). However, at
12065: present this wordset is still incomplete. It is also pretty low-level;
12066: some day it will hopefully be made unnecessary by an internals wordset
12067: that abstracts implementation details away completely.
1.1 anton 12068:
1.78 anton 12069: The terminology used here stems from indirect threaded Forth systems; in
12070: such a system, the XT of a word is represented by the CFA (code field
12071: address) of a word; the CFA points to a cell that contains the code
12072: address. The code address is the address of some machine code that
12073: performs the run-time action of invoking the word (e.g., the
12074: @code{dovar:} routine pushes the address of the body of the word (a
12075: variable) on the stack
12076: ).
1.1 anton 12077:
1.78 anton 12078: @cindex code address
12079: @cindex code field address
12080: In an indirect threaded Forth, you can get the code address of @i{name}
12081: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
12082: >code-address}, independent of the threading method.
1.1 anton 12083:
1.78 anton 12084: doc-threading-method
12085: doc->code-address
12086: doc-code-address!
1.1 anton 12087:
1.78 anton 12088: @cindex @code{does>}-handler
12089: @cindex @code{does>}-code
12090: For a word defined with @code{DOES>}, the code address usually points to
12091: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
12092: routine (in Gforth on some platforms, it can also point to the dodoes
12093: routine itself). What you are typically interested in, though, is
12094: whether a word is a @code{DOES>}-defined word, and what Forth code it
12095: executes; @code{>does-code} tells you that.
1.1 anton 12096:
1.78 anton 12097: doc->does-code
1.1 anton 12098:
1.78 anton 12099: To create a @code{DOES>}-defined word with the following basic words,
12100: you have to set up a @code{DOES>}-handler with @code{does-handler!};
12101: @code{/does-handler} aus behind you have to place your executable Forth
12102: code. Finally you have to create a word and modify its behaviour with
12103: @code{does-handler!}.
1.1 anton 12104:
1.78 anton 12105: doc-does-code!
12106: doc-does-handler!
12107: doc-/does-handler
1.1 anton 12108:
1.78 anton 12109: The code addresses produced by various defining words are produced by
12110: the following words:
1.1 anton 12111:
1.78 anton 12112: doc-docol:
12113: doc-docon:
12114: doc-dovar:
12115: doc-douser:
12116: doc-dodefer:
12117: doc-dofield:
1.1 anton 12118:
1.26 crook 12119: @c -------------------------------------------------------------
1.78 anton 12120: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
1.21 crook 12121: @section Passing Commands to the Operating System
12122: @cindex operating system - passing commands
12123: @cindex shell commands
12124:
12125: Gforth allows you to pass an arbitrary string to the host operating
12126: system shell (if such a thing exists) for execution.
12127:
1.44 crook 12128:
1.21 crook 12129: doc-sh
12130: doc-system
12131: doc-$?
1.23 crook 12132: doc-getenv
1.21 crook 12133:
1.44 crook 12134:
1.26 crook 12135: @c -------------------------------------------------------------
1.47 crook 12136: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
12137: @section Keeping track of Time
12138: @cindex time-related words
12139:
12140: doc-ms
12141: doc-time&date
1.79 anton 12142: doc-utime
12143: doc-cputime
1.47 crook 12144:
12145:
12146: @c -------------------------------------------------------------
12147: @node Miscellaneous Words, , Keeping track of Time, Words
1.21 crook 12148: @section Miscellaneous Words
12149: @cindex miscellaneous words
12150:
1.29 crook 12151: @comment TODO find homes for these
12152:
1.26 crook 12153: These section lists the ANS Forth words that are not documented
1.21 crook 12154: elsewhere in this manual. Ultimately, they all need proper homes.
12155:
1.68 anton 12156: doc-quit
1.44 crook 12157:
1.26 crook 12158: The following ANS Forth words are not currently supported by Gforth
1.27 crook 12159: (@pxref{ANS conformance}):
1.21 crook 12160:
12161: @code{EDITOR}
12162: @code{EMIT?}
12163: @code{FORGET}
12164:
1.24 anton 12165: @c ******************************************************************
12166: @node Error messages, Tools, Words, Top
12167: @chapter Error messages
12168: @cindex error messages
12169: @cindex backtrace
12170:
12171: A typical Gforth error message looks like this:
12172:
12173: @example
1.86 anton 12174: in file included from \evaluated string/:-1
1.24 anton 12175: in file included from ./yyy.fs:1
12176: ./xxx.fs:4: Invalid memory address
12177: bar
12178: ^^^
1.79 anton 12179: Backtrace:
1.25 anton 12180: $400E664C @@
12181: $400E6664 foo
1.24 anton 12182: @end example
12183:
12184: The message identifying the error is @code{Invalid memory address}. The
12185: error happened when text-interpreting line 4 of the file
12186: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
12187: word on the line where the error happened, is pointed out (with
12188: @code{^^^}).
12189:
12190: The file containing the error was included in line 1 of @file{./yyy.fs},
12191: and @file{yyy.fs} was included from a non-file (in this case, by giving
12192: @file{yyy.fs} as command-line parameter to Gforth).
12193:
12194: At the end of the error message you find a return stack dump that can be
12195: interpreted as a backtrace (possibly empty). On top you find the top of
12196: the return stack when the @code{throw} happened, and at the bottom you
12197: find the return stack entry just above the return stack of the topmost
12198: text interpreter.
12199:
12200: To the right of most return stack entries you see a guess for the word
12201: that pushed that return stack entry as its return address. This gives a
12202: backtrace. In our case we see that @code{bar} called @code{foo}, and
12203: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
12204: address} exception).
12205:
12206: Note that the backtrace is not perfect: We don't know which return stack
12207: entries are return addresses (so we may get false positives); and in
12208: some cases (e.g., for @code{abort"}) we cannot determine from the return
12209: address the word that pushed the return address, so for some return
12210: addresses you see no names in the return stack dump.
1.25 anton 12211:
12212: @cindex @code{catch} and backtraces
12213: The return stack dump represents the return stack at the time when a
12214: specific @code{throw} was executed. In programs that make use of
12215: @code{catch}, it is not necessarily clear which @code{throw} should be
12216: used for the return stack dump (e.g., consider one @code{throw} that
12217: indicates an error, which is caught, and during recovery another error
1.42 anton 12218: happens; which @code{throw} should be used for the stack dump?). Gforth
1.25 anton 12219: presents the return stack dump for the first @code{throw} after the last
12220: executed (not returned-to) @code{catch}; this works well in the usual
12221: case.
12222:
12223: @cindex @code{gforth-fast} and backtraces
12224: @cindex @code{gforth-fast}, difference from @code{gforth}
12225: @cindex backtraces with @code{gforth-fast}
12226: @cindex return stack dump with @code{gforth-fast}
1.79 anton 12227: @code{Gforth} is able to do a return stack dump for throws generated
1.25 anton 12228: from primitives (e.g., invalid memory address, stack empty etc.);
12229: @code{gforth-fast} is only able to do a return stack dump from a
12230: directly called @code{throw} (including @code{abort} etc.). This is the
1.30 anton 12231: only difference (apart from a speed factor of between 1.15 (K6-2) and
1.78 anton 12232: 2 (21264)) between @code{gforth} and @code{gforth-fast}. Given an
1.30 anton 12233: exception caused by a primitive in @code{gforth-fast}, you will
12234: typically see no return stack dump at all; however, if the exception is
12235: caught by @code{catch} (e.g., for restoring some state), and then
12236: @code{throw}n again, the return stack dump will be for the first such
12237: @code{throw}.
1.2 jwilke 12238:
1.5 anton 12239: @c ******************************************************************
1.24 anton 12240: @node Tools, ANS conformance, Error messages, Top
1.1 anton 12241: @chapter Tools
12242:
12243: @menu
12244: * ANS Report:: Report the words used, sorted by wordset.
12245: @end menu
12246:
12247: See also @ref{Emacs and Gforth}.
12248:
12249: @node ANS Report, , Tools, Tools
12250: @section @file{ans-report.fs}: Report the words used, sorted by wordset
12251: @cindex @file{ans-report.fs}
12252: @cindex report the words used in your program
12253: @cindex words used in your program
12254:
12255: If you want to label a Forth program as ANS Forth Program, you must
12256: document which wordsets the program uses; for extension wordsets, it is
12257: helpful to list the words the program requires from these wordsets
12258: (because Forth systems are allowed to provide only some words of them).
12259:
12260: The @file{ans-report.fs} tool makes it easy for you to determine which
12261: words from which wordset and which non-ANS words your application
12262: uses. You simply have to include @file{ans-report.fs} before loading the
12263: program you want to check. After loading your program, you can get the
12264: report with @code{print-ans-report}. A typical use is to run this as
12265: batch job like this:
12266: @example
12267: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
12268: @end example
12269:
12270: The output looks like this (for @file{compat/control.fs}):
12271: @example
12272: The program uses the following words
12273: from CORE :
12274: : POSTPONE THEN ; immediate ?dup IF 0=
12275: from BLOCK-EXT :
12276: \
12277: from FILE :
12278: (
12279: @end example
12280:
12281: @subsection Caveats
12282:
12283: Note that @file{ans-report.fs} just checks which words are used, not whether
12284: they are used in an ANS Forth conforming way!
12285:
12286: Some words are defined in several wordsets in the
12287: standard. @file{ans-report.fs} reports them for only one of the
12288: wordsets, and not necessarily the one you expect. It depends on usage
12289: which wordset is the right one to specify. E.g., if you only use the
12290: compilation semantics of @code{S"}, it is a Core word; if you also use
12291: its interpretation semantics, it is a File word.
12292:
12293: @c ******************************************************************
1.65 anton 12294: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1 anton 12295: @chapter ANS conformance
12296: @cindex ANS conformance of Gforth
12297:
12298: To the best of our knowledge, Gforth is an
12299:
12300: ANS Forth System
12301: @itemize @bullet
12302: @item providing the Core Extensions word set
12303: @item providing the Block word set
12304: @item providing the Block Extensions word set
12305: @item providing the Double-Number word set
12306: @item providing the Double-Number Extensions word set
12307: @item providing the Exception word set
12308: @item providing the Exception Extensions word set
12309: @item providing the Facility word set
1.40 anton 12310: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1 anton 12311: @item providing the File Access word set
12312: @item providing the File Access Extensions word set
12313: @item providing the Floating-Point word set
12314: @item providing the Floating-Point Extensions word set
12315: @item providing the Locals word set
12316: @item providing the Locals Extensions word set
12317: @item providing the Memory-Allocation word set
12318: @item providing the Memory-Allocation Extensions word set (that one's easy)
12319: @item providing the Programming-Tools word set
12320: @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
12321: @item providing the Search-Order word set
12322: @item providing the Search-Order Extensions word set
12323: @item providing the String word set
12324: @item providing the String Extensions word set (another easy one)
12325: @end itemize
12326:
12327: @cindex system documentation
12328: In addition, ANS Forth systems are required to document certain
12329: implementation choices. This chapter tries to meet these
12330: requirements. In many cases it gives a way to ask the system for the
12331: information instead of providing the information directly, in
12332: particular, if the information depends on the processor, the operating
12333: system or the installation options chosen, or if they are likely to
12334: change during the maintenance of Gforth.
12335:
12336: @comment The framework for the rest has been taken from pfe.
12337:
12338: @menu
12339: * The Core Words::
12340: * The optional Block word set::
12341: * The optional Double Number word set::
12342: * The optional Exception word set::
12343: * The optional Facility word set::
12344: * The optional File-Access word set::
12345: * The optional Floating-Point word set::
12346: * The optional Locals word set::
12347: * The optional Memory-Allocation word set::
12348: * The optional Programming-Tools word set::
12349: * The optional Search-Order word set::
12350: @end menu
12351:
12352:
12353: @c =====================================================================
12354: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
12355: @comment node-name, next, previous, up
12356: @section The Core Words
12357: @c =====================================================================
12358: @cindex core words, system documentation
12359: @cindex system documentation, core words
12360:
12361: @menu
12362: * core-idef:: Implementation Defined Options
12363: * core-ambcond:: Ambiguous Conditions
12364: * core-other:: Other System Documentation
12365: @end menu
12366:
12367: @c ---------------------------------------------------------------------
12368: @node core-idef, core-ambcond, The Core Words, The Core Words
12369: @subsection Implementation Defined Options
12370: @c ---------------------------------------------------------------------
12371: @cindex core words, implementation-defined options
12372: @cindex implementation-defined options, core words
12373:
12374:
12375: @table @i
12376: @item (Cell) aligned addresses:
12377: @cindex cell-aligned addresses
12378: @cindex aligned addresses
12379: processor-dependent. Gforth's alignment words perform natural alignment
12380: (e.g., an address aligned for a datum of size 8 is divisible by
12381: 8). Unaligned accesses usually result in a @code{-23 THROW}.
12382:
12383: @item @code{EMIT} and non-graphic characters:
12384: @cindex @code{EMIT} and non-graphic characters
12385: @cindex non-graphic characters and @code{EMIT}
12386: The character is output using the C library function (actually, macro)
12387: @code{putc}.
12388:
12389: @item character editing of @code{ACCEPT} and @code{EXPECT}:
12390: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
12391: @cindex editing in @code{ACCEPT} and @code{EXPECT}
12392: @cindex @code{ACCEPT}, editing
12393: @cindex @code{EXPECT}, editing
12394: This is modeled on the GNU readline library (@pxref{Readline
12395: Interaction, , Command Line Editing, readline, The GNU Readline
12396: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
12397: producing a full word completion every time you type it (instead of
1.28 crook 12398: producing the common prefix of all completions). @xref{Command-line editing}.
1.1 anton 12399:
12400: @item character set:
12401: @cindex character set
12402: The character set of your computer and display device. Gforth is
12403: 8-bit-clean (but some other component in your system may make trouble).
12404:
12405: @item Character-aligned address requirements:
12406: @cindex character-aligned address requirements
12407: installation-dependent. Currently a character is represented by a C
12408: @code{unsigned char}; in the future we might switch to @code{wchar_t}
12409: (Comments on that requested).
12410:
12411: @item character-set extensions and matching of names:
12412: @cindex character-set extensions and matching of names
1.26 crook 12413: @cindex case-sensitivity for name lookup
12414: @cindex name lookup, case-sensitivity
12415: @cindex locale and case-sensitivity
1.21 crook 12416: Any character except the ASCII NUL character can be used in a
1.1 anton 12417: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47 crook 12418: matching is performed using the C library function @code{strncasecmp}, whose
1.1 anton 12419: function is probably influenced by the locale. E.g., the @code{C} locale
12420: does not know about accents and umlauts, so they are matched
12421: case-sensitively in that locale. For portability reasons it is best to
12422: write programs such that they work in the @code{C} locale. Then one can
12423: use libraries written by a Polish programmer (who might use words
12424: containing ISO Latin-2 encoded characters) and by a French programmer
12425: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
12426: funny results for some of the words (which ones, depends on the font you
12427: are using)). Also, the locale you prefer may not be available in other
12428: operating systems. Hopefully, Unicode will solve these problems one day.
12429:
12430: @item conditions under which control characters match a space delimiter:
12431: @cindex space delimiters
12432: @cindex control characters as delimiters
12433: If @code{WORD} is called with the space character as a delimiter, all
12434: white-space characters (as identified by the C macro @code{isspace()})
12435: are delimiters. @code{PARSE}, on the other hand, treats space like other
1.44 crook 12436: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
1.79 anton 12437: like @code{PARSE} otherwise. @code{Name}, which is used by the outer
1.1 anton 12438: interpreter (aka text interpreter) by default, treats all white-space
12439: characters as delimiters.
12440:
1.26 crook 12441: @item format of the control-flow stack:
12442: @cindex control-flow stack, format
12443: The data stack is used as control-flow stack. The size of a control-flow
1.1 anton 12444: stack item in cells is given by the constant @code{cs-item-size}. At the
12445: time of this writing, an item consists of a (pointer to a) locals list
12446: (third), an address in the code (second), and a tag for identifying the
12447: item (TOS). The following tags are used: @code{defstart},
12448: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12449: @code{scopestart}.
12450:
12451: @item conversion of digits > 35
12452: @cindex digits > 35
12453: The characters @code{[\]^_'} are the digits with the decimal value
12454: 36@minus{}41. There is no way to input many of the larger digits.
12455:
12456: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12457: @cindex @code{EXPECT}, display after end of input
12458: @cindex @code{ACCEPT}, display after end of input
12459: The cursor is moved to the end of the entered string. If the input is
12460: terminated using the @kbd{Return} key, a space is typed.
12461:
12462: @item exception abort sequence of @code{ABORT"}:
12463: @cindex exception abort sequence of @code{ABORT"}
12464: @cindex @code{ABORT"}, exception abort sequence
12465: The error string is stored into the variable @code{"error} and a
12466: @code{-2 throw} is performed.
12467:
12468: @item input line terminator:
12469: @cindex input line terminator
12470: @cindex line terminator on input
1.26 crook 12471: @cindex newline character on input
1.1 anton 12472: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12473: lines. One of these characters is typically produced when you type the
12474: @kbd{Enter} or @kbd{Return} key.
12475:
12476: @item maximum size of a counted string:
12477: @cindex maximum size of a counted string
12478: @cindex counted string, maximum size
12479: @code{s" /counted-string" environment? drop .}. Currently 255 characters
1.79 anton 12480: on all platforms, but this may change.
1.1 anton 12481:
12482: @item maximum size of a parsed string:
12483: @cindex maximum size of a parsed string
12484: @cindex parsed string, maximum size
12485: Given by the constant @code{/line}. Currently 255 characters.
12486:
12487: @item maximum size of a definition name, in characters:
12488: @cindex maximum size of a definition name, in characters
12489: @cindex name, maximum length
12490: 31
12491:
12492: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12493: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12494: @cindex @code{ENVIRONMENT?} string length, maximum
12495: 31
12496:
12497: @item method of selecting the user input device:
12498: @cindex user input device, method of selecting
12499: The user input device is the standard input. There is currently no way to
12500: change it from within Gforth. However, the input can typically be
12501: redirected in the command line that starts Gforth.
12502:
12503: @item method of selecting the user output device:
12504: @cindex user output device, method of selecting
12505: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10 anton 12506: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12507: output when the user output device is a terminal, otherwise the output
12508: is buffered.
1.1 anton 12509:
12510: @item methods of dictionary compilation:
12511: What are we expected to document here?
12512:
12513: @item number of bits in one address unit:
12514: @cindex number of bits in one address unit
12515: @cindex address unit, size in bits
12516: @code{s" address-units-bits" environment? drop .}. 8 in all current
1.79 anton 12517: platforms.
1.1 anton 12518:
12519: @item number representation and arithmetic:
12520: @cindex number representation and arithmetic
1.79 anton 12521: Processor-dependent. Binary two's complement on all current platforms.
1.1 anton 12522:
12523: @item ranges for integer types:
12524: @cindex ranges for integer types
12525: @cindex integer types, ranges
12526: Installation-dependent. Make environmental queries for @code{MAX-N},
12527: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12528: unsigned (and positive) types is 0. The lower bound for signed types on
12529: two's complement and one's complement machines machines can be computed
12530: by adding 1 to the upper bound.
12531:
12532: @item read-only data space regions:
12533: @cindex read-only data space regions
12534: @cindex data-space, read-only regions
12535: The whole Forth data space is writable.
12536:
12537: @item size of buffer at @code{WORD}:
12538: @cindex size of buffer at @code{WORD}
12539: @cindex @code{WORD} buffer size
12540: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12541: shared with the pictured numeric output string. If overwriting
12542: @code{PAD} is acceptable, it is as large as the remaining dictionary
12543: space, although only as much can be sensibly used as fits in a counted
12544: string.
12545:
12546: @item size of one cell in address units:
12547: @cindex cell size
12548: @code{1 cells .}.
12549:
12550: @item size of one character in address units:
12551: @cindex char size
1.79 anton 12552: @code{1 chars .}. 1 on all current platforms.
1.1 anton 12553:
12554: @item size of the keyboard terminal buffer:
12555: @cindex size of the keyboard terminal buffer
12556: @cindex terminal buffer, size
12557: Varies. You can determine the size at a specific time using @code{lp@@
12558: tib - .}. It is shared with the locals stack and TIBs of files that
12559: include the current file. You can change the amount of space for TIBs
12560: and locals stack at Gforth startup with the command line option
12561: @code{-l}.
12562:
12563: @item size of the pictured numeric output buffer:
12564: @cindex size of the pictured numeric output buffer
12565: @cindex pictured numeric output buffer, size
12566: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12567: shared with @code{WORD}.
12568:
12569: @item size of the scratch area returned by @code{PAD}:
12570: @cindex size of the scratch area returned by @code{PAD}
12571: @cindex @code{PAD} size
12572: The remainder of dictionary space. @code{unused pad here - - .}.
12573:
12574: @item system case-sensitivity characteristics:
12575: @cindex case-sensitivity characteristics
1.26 crook 12576: Dictionary searches are case-insensitive (except in
1.1 anton 12577: @code{TABLE}s). However, as explained above under @i{character-set
12578: extensions}, the matching for non-ASCII characters is determined by the
12579: locale you are using. In the default @code{C} locale all non-ASCII
12580: characters are matched case-sensitively.
12581:
12582: @item system prompt:
12583: @cindex system prompt
12584: @cindex prompt
12585: @code{ ok} in interpret state, @code{ compiled} in compile state.
12586:
12587: @item division rounding:
12588: @cindex division rounding
12589: installation dependent. @code{s" floored" environment? drop .}. We leave
12590: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12591: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12592:
12593: @item values of @code{STATE} when true:
12594: @cindex @code{STATE} values
12595: -1.
12596:
12597: @item values returned after arithmetic overflow:
12598: On two's complement machines, arithmetic is performed modulo
12599: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12600: arithmetic (with appropriate mapping for signed types). Division by zero
12601: typically results in a @code{-55 throw} (Floating-point unidentified
1.80 anton 12602: fault) or @code{-10 throw} (divide by zero).
1.1 anton 12603:
12604: @item whether the current definition can be found after @t{DOES>}:
12605: @cindex @t{DOES>}, visibility of current definition
12606: No.
12607:
12608: @end table
12609:
12610: @c ---------------------------------------------------------------------
12611: @node core-ambcond, core-other, core-idef, The Core Words
12612: @subsection Ambiguous conditions
12613: @c ---------------------------------------------------------------------
12614: @cindex core words, ambiguous conditions
12615: @cindex ambiguous conditions, core words
12616:
12617: @table @i
12618:
12619: @item a name is neither a word nor a number:
12620: @cindex name not found
1.26 crook 12621: @cindex undefined word
1.80 anton 12622: @code{-13 throw} (Undefined word).
1.1 anton 12623:
12624: @item a definition name exceeds the maximum length allowed:
1.26 crook 12625: @cindex word name too long
1.1 anton 12626: @code{-19 throw} (Word name too long)
12627:
12628: @item addressing a region not inside the various data spaces of the forth system:
12629: @cindex Invalid memory address
1.32 anton 12630: The stacks, code space and header space are accessible. Machine code space is
1.1 anton 12631: typically readable. Accessing other addresses gives results dependent on
12632: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12633: address).
12634:
12635: @item argument type incompatible with parameter:
1.26 crook 12636: @cindex argument type mismatch
1.1 anton 12637: This is usually not caught. Some words perform checks, e.g., the control
12638: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12639: mismatch).
12640:
12641: @item attempting to obtain the execution token of a word with undefined execution semantics:
12642: @cindex Interpreting a compile-only word, for @code{'} etc.
12643: @cindex execution token of words with undefined execution semantics
12644: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12645: get an execution token for @code{compile-only-error} (which performs a
12646: @code{-14 throw} when executed).
12647:
12648: @item dividing by zero:
12649: @cindex dividing by zero
12650: @cindex floating point unidentified fault, integer division
1.80 anton 12651: On some platforms, this produces a @code{-10 throw} (Division by
1.24 anton 12652: zero); on other systems, this typically results in a @code{-55 throw}
12653: (Floating-point unidentified fault).
1.1 anton 12654:
12655: @item insufficient data stack or return stack space:
12656: @cindex insufficient data stack or return stack space
12657: @cindex stack overflow
1.26 crook 12658: @cindex address alignment exception, stack overflow
1.1 anton 12659: @cindex Invalid memory address, stack overflow
12660: Depending on the operating system, the installation, and the invocation
12661: of Gforth, this is either checked by the memory management hardware, or
1.24 anton 12662: it is not checked. If it is checked, you typically get a @code{-3 throw}
12663: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12664: throw} (Invalid memory address) (depending on the platform and how you
12665: achieved the overflow) as soon as the overflow happens. If it is not
12666: checked, overflows typically result in mysterious illegal memory
12667: accesses, producing @code{-9 throw} (Invalid memory address) or
12668: @code{-23 throw} (Address alignment exception); they might also destroy
12669: the internal data structure of @code{ALLOCATE} and friends, resulting in
12670: various errors in these words.
1.1 anton 12671:
12672: @item insufficient space for loop control parameters:
12673: @cindex insufficient space for loop control parameters
1.80 anton 12674: Like other return stack overflows.
1.1 anton 12675:
12676: @item insufficient space in the dictionary:
12677: @cindex insufficient space in the dictionary
12678: @cindex dictionary overflow
1.12 anton 12679: If you try to allot (either directly with @code{allot}, or indirectly
12680: with @code{,}, @code{create} etc.) more memory than available in the
12681: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12682: to access memory beyond the end of the dictionary, the results are
12683: similar to stack overflows.
1.1 anton 12684:
12685: @item interpreting a word with undefined interpretation semantics:
12686: @cindex interpreting a word with undefined interpretation semantics
12687: @cindex Interpreting a compile-only word
12688: For some words, we have defined interpretation semantics. For the
12689: others: @code{-14 throw} (Interpreting a compile-only word).
12690:
12691: @item modifying the contents of the input buffer or a string literal:
12692: @cindex modifying the contents of the input buffer or a string literal
12693: These are located in writable memory and can be modified.
12694:
12695: @item overflow of the pictured numeric output string:
12696: @cindex overflow of the pictured numeric output string
12697: @cindex pictured numeric output string, overflow
1.24 anton 12698: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1 anton 12699:
12700: @item parsed string overflow:
12701: @cindex parsed string overflow
12702: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12703:
12704: @item producing a result out of range:
12705: @cindex result out of range
12706: On two's complement machines, arithmetic is performed modulo
12707: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12708: arithmetic (with appropriate mapping for signed types). Division by zero
1.24 anton 12709: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12710: throw} (floating point unidentified fault). @code{convert} and
12711: @code{>number} currently overflow silently.
1.1 anton 12712:
12713: @item reading from an empty data or return stack:
12714: @cindex stack empty
12715: @cindex stack underflow
1.24 anton 12716: @cindex return stack underflow
1.1 anton 12717: The data stack is checked by the outer (aka text) interpreter after
12718: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12719: underflow) is performed. Apart from that, stacks may be checked or not,
1.24 anton 12720: depending on operating system, installation, and invocation. If they are
12721: caught by a check, they typically result in @code{-4 throw} (Stack
12722: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12723: (Invalid memory address), depending on the platform and which stack
12724: underflows and by how much. Note that even if the system uses checking
12725: (through the MMU), your program may have to underflow by a significant
12726: number of stack items to trigger the reaction (the reason for this is
12727: that the MMU, and therefore the checking, works with a page-size
12728: granularity). If there is no checking, the symptoms resulting from an
12729: underflow are similar to those from an overflow. Unbalanced return
1.80 anton 12730: stack errors can result in a variety of symptoms, including @code{-9 throw}
1.24 anton 12731: (Invalid memory address) and Illegal Instruction (typically @code{-260
12732: throw}).
1.1 anton 12733:
12734: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12735: @cindex unexpected end of the input buffer
12736: @cindex zero-length string as a name
12737: @cindex Attempt to use zero-length string as a name
12738: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12739: use zero-length string as a name). Words like @code{'} probably will not
12740: find what they search. Note that it is possible to create zero-length
12741: names with @code{nextname} (should it not?).
12742:
12743: @item @code{>IN} greater than input buffer:
12744: @cindex @code{>IN} greater than input buffer
12745: The next invocation of a parsing word returns a string with length 0.
12746:
12747: @item @code{RECURSE} appears after @code{DOES>}:
12748: @cindex @code{RECURSE} appears after @code{DOES>}
12749: Compiles a recursive call to the defining word, not to the defined word.
12750:
12751: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12752: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26 crook 12753: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1 anton 12754: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12755: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12756: the end of the file was reached), its source-id may be
12757: reused. Therefore, restoring an input source specification referencing a
12758: closed file may lead to unpredictable results instead of a @code{-12
12759: THROW}.
12760:
12761: In the future, Gforth may be able to restore input source specifications
12762: from other than the current input source.
12763:
12764: @item data space containing definitions gets de-allocated:
12765: @cindex data space containing definitions gets de-allocated
12766: Deallocation with @code{allot} is not checked. This typically results in
12767: memory access faults or execution of illegal instructions.
12768:
12769: @item data space read/write with incorrect alignment:
12770: @cindex data space read/write with incorrect alignment
12771: @cindex alignment faults
1.26 crook 12772: @cindex address alignment exception
1.1 anton 12773: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12 anton 12774: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1 anton 12775: alignment turned on, incorrect alignment results in a @code{-9 throw}
12776: (Invalid memory address). There are reportedly some processors with
1.12 anton 12777: alignment restrictions that do not report violations.
1.1 anton 12778:
12779: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12780: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12781: Like other alignment errors.
12782:
12783: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12784: Like other stack underflows.
12785:
12786: @item loop control parameters not available:
12787: @cindex loop control parameters not available
12788: Not checked. The counted loop words simply assume that the top of return
12789: stack items are loop control parameters and behave accordingly.
12790:
12791: @item most recent definition does not have a name (@code{IMMEDIATE}):
12792: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12793: @cindex last word was headerless
12794: @code{abort" last word was headerless"}.
12795:
12796: @item name not defined by @code{VALUE} used by @code{TO}:
12797: @cindex name not defined by @code{VALUE} used by @code{TO}
12798: @cindex @code{TO} on non-@code{VALUE}s
12799: @cindex Invalid name argument, @code{TO}
12800: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12801: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12802:
12803: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12804: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26 crook 12805: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1 anton 12806: @code{-13 throw} (Undefined word)
12807:
12808: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12809: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12810: Gforth behaves as if they were of the same type. I.e., you can predict
12811: the behaviour by interpreting all parameters as, e.g., signed.
12812:
12813: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12814: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12815: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12816: compilation semantics of @code{TO}.
12817:
12818: @item String longer than a counted string returned by @code{WORD}:
1.26 crook 12819: @cindex string longer than a counted string returned by @code{WORD}
1.1 anton 12820: @cindex @code{WORD}, string overflow
12821: Not checked. The string will be ok, but the count will, of course,
12822: contain only the least significant bits of the length.
12823:
12824: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12825: @cindex @code{LSHIFT}, large shift counts
12826: @cindex @code{RSHIFT}, large shift counts
12827: Processor-dependent. Typical behaviours are returning 0 and using only
12828: the low bits of the shift count.
12829:
12830: @item word not defined via @code{CREATE}:
12831: @cindex @code{>BODY} of non-@code{CREATE}d words
12832: @code{>BODY} produces the PFA of the word no matter how it was defined.
12833:
12834: @cindex @code{DOES>} of non-@code{CREATE}d words
12835: @code{DOES>} changes the execution semantics of the last defined word no
12836: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12837: @code{CREATE , DOES>}.
12838:
12839: @item words improperly used outside @code{<#} and @code{#>}:
12840: Not checked. As usual, you can expect memory faults.
12841:
12842: @end table
12843:
12844:
12845: @c ---------------------------------------------------------------------
12846: @node core-other, , core-ambcond, The Core Words
12847: @subsection Other system documentation
12848: @c ---------------------------------------------------------------------
12849: @cindex other system documentation, core words
12850: @cindex core words, other system documentation
12851:
12852: @table @i
12853: @item nonstandard words using @code{PAD}:
12854: @cindex @code{PAD} use by nonstandard words
12855: None.
12856:
12857: @item operator's terminal facilities available:
12858: @cindex operator's terminal facilities available
1.80 anton 12859: After processing the OS's command line, Gforth goes into interactive mode,
1.1 anton 12860: and you can give commands to Gforth interactively. The actual facilities
12861: available depend on how you invoke Gforth.
12862:
12863: @item program data space available:
12864: @cindex program data space available
12865: @cindex data space available
12866: @code{UNUSED .} gives the remaining dictionary space. The total
12867: dictionary space can be specified with the @code{-m} switch
12868: (@pxref{Invoking Gforth}) when Gforth starts up.
12869:
12870: @item return stack space available:
12871: @cindex return stack space available
12872: You can compute the total return stack space in cells with
12873: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12874: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12875:
12876: @item stack space available:
12877: @cindex stack space available
12878: You can compute the total data stack space in cells with
12879: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12880: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12881:
12882: @item system dictionary space required, in address units:
12883: @cindex system dictionary space required, in address units
12884: Type @code{here forthstart - .} after startup. At the time of this
12885: writing, this gives 80080 (bytes) on a 32-bit system.
12886: @end table
12887:
12888:
12889: @c =====================================================================
12890: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12891: @section The optional Block word set
12892: @c =====================================================================
12893: @cindex system documentation, block words
12894: @cindex block words, system documentation
12895:
12896: @menu
12897: * block-idef:: Implementation Defined Options
12898: * block-ambcond:: Ambiguous Conditions
12899: * block-other:: Other System Documentation
12900: @end menu
12901:
12902:
12903: @c ---------------------------------------------------------------------
12904: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12905: @subsection Implementation Defined Options
12906: @c ---------------------------------------------------------------------
12907: @cindex implementation-defined options, block words
12908: @cindex block words, implementation-defined options
12909:
12910: @table @i
12911: @item the format for display by @code{LIST}:
12912: @cindex @code{LIST} display format
12913: First the screen number is displayed, then 16 lines of 64 characters,
12914: each line preceded by the line number.
12915:
12916: @item the length of a line affected by @code{\}:
12917: @cindex length of a line affected by @code{\}
12918: @cindex @code{\}, line length in blocks
12919: 64 characters.
12920: @end table
12921:
12922:
12923: @c ---------------------------------------------------------------------
12924: @node block-ambcond, block-other, block-idef, The optional Block word set
12925: @subsection Ambiguous conditions
12926: @c ---------------------------------------------------------------------
12927: @cindex block words, ambiguous conditions
12928: @cindex ambiguous conditions, block words
12929:
12930: @table @i
12931: @item correct block read was not possible:
12932: @cindex block read not possible
12933: Typically results in a @code{throw} of some OS-derived value (between
12934: -512 and -2048). If the blocks file was just not long enough, blanks are
12935: supplied for the missing portion.
12936:
12937: @item I/O exception in block transfer:
12938: @cindex I/O exception in block transfer
12939: @cindex block transfer, I/O exception
12940: Typically results in a @code{throw} of some OS-derived value (between
12941: -512 and -2048).
12942:
12943: @item invalid block number:
12944: @cindex invalid block number
12945: @cindex block number invalid
12946: @code{-35 throw} (Invalid block number)
12947:
12948: @item a program directly alters the contents of @code{BLK}:
12949: @cindex @code{BLK}, altering @code{BLK}
12950: The input stream is switched to that other block, at the same
12951: position. If the storing to @code{BLK} happens when interpreting
12952: non-block input, the system will get quite confused when the block ends.
12953:
12954: @item no current block buffer for @code{UPDATE}:
12955: @cindex @code{UPDATE}, no current block buffer
12956: @code{UPDATE} has no effect.
12957:
12958: @end table
12959:
12960: @c ---------------------------------------------------------------------
12961: @node block-other, , block-ambcond, The optional Block word set
12962: @subsection Other system documentation
12963: @c ---------------------------------------------------------------------
12964: @cindex other system documentation, block words
12965: @cindex block words, other system documentation
12966:
12967: @table @i
12968: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12969: No restrictions (yet).
12970:
12971: @item the number of blocks available for source and data:
12972: depends on your disk space.
12973:
12974: @end table
12975:
12976:
12977: @c =====================================================================
12978: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12979: @section The optional Double Number word set
12980: @c =====================================================================
12981: @cindex system documentation, double words
12982: @cindex double words, system documentation
12983:
12984: @menu
12985: * double-ambcond:: Ambiguous Conditions
12986: @end menu
12987:
12988:
12989: @c ---------------------------------------------------------------------
12990: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12991: @subsection Ambiguous conditions
12992: @c ---------------------------------------------------------------------
12993: @cindex double words, ambiguous conditions
12994: @cindex ambiguous conditions, double words
12995:
12996: @table @i
1.29 crook 12997: @item @i{d} outside of range of @i{n} in @code{D>S}:
12998: @cindex @code{D>S}, @i{d} out of range of @i{n}
12999: The least significant cell of @i{d} is produced.
1.1 anton 13000:
13001: @end table
13002:
13003:
13004: @c =====================================================================
13005: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
13006: @section The optional Exception word set
13007: @c =====================================================================
13008: @cindex system documentation, exception words
13009: @cindex exception words, system documentation
13010:
13011: @menu
13012: * exception-idef:: Implementation Defined Options
13013: @end menu
13014:
13015:
13016: @c ---------------------------------------------------------------------
13017: @node exception-idef, , The optional Exception word set, The optional Exception word set
13018: @subsection Implementation Defined Options
13019: @c ---------------------------------------------------------------------
13020: @cindex implementation-defined options, exception words
13021: @cindex exception words, implementation-defined options
13022:
13023: @table @i
13024: @item @code{THROW}-codes used in the system:
13025: @cindex @code{THROW}-codes used in the system
13026: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29 crook 13027: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1 anton 13028: codes -512@minus{}-2047 are used for OS errors (for file and memory
13029: allocation operations). The mapping from OS error numbers to throw codes
13030: is -512@minus{}@code{errno}. One side effect of this mapping is that
13031: undefined OS errors produce a message with a strange number; e.g.,
13032: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
13033: @end table
13034:
13035: @c =====================================================================
13036: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
13037: @section The optional Facility word set
13038: @c =====================================================================
13039: @cindex system documentation, facility words
13040: @cindex facility words, system documentation
13041:
13042: @menu
13043: * facility-idef:: Implementation Defined Options
13044: * facility-ambcond:: Ambiguous Conditions
13045: @end menu
13046:
13047:
13048: @c ---------------------------------------------------------------------
13049: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
13050: @subsection Implementation Defined Options
13051: @c ---------------------------------------------------------------------
13052: @cindex implementation-defined options, facility words
13053: @cindex facility words, implementation-defined options
13054:
13055: @table @i
13056: @item encoding of keyboard events (@code{EKEY}):
13057: @cindex keyboard events, encoding in @code{EKEY}
13058: @cindex @code{EKEY}, encoding of keyboard events
1.40 anton 13059: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41 anton 13060: Other keys are encoded with the constants @code{k-left}, @code{k-right},
13061: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
13062: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
13063: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40 anton 13064:
1.1 anton 13065:
13066: @item duration of a system clock tick:
13067: @cindex duration of a system clock tick
13068: @cindex clock tick duration
13069: System dependent. With respect to @code{MS}, the time is specified in
13070: microseconds. How well the OS and the hardware implement this, is
13071: another question.
13072:
13073: @item repeatability to be expected from the execution of @code{MS}:
13074: @cindex repeatability to be expected from the execution of @code{MS}
13075: @cindex @code{MS}, repeatability to be expected
13076: System dependent. On Unix, a lot depends on load. If the system is
13077: lightly loaded, and the delay is short enough that Gforth does not get
13078: swapped out, the performance should be acceptable. Under MS-DOS and
13079: other single-tasking systems, it should be good.
13080:
13081: @end table
13082:
13083:
13084: @c ---------------------------------------------------------------------
13085: @node facility-ambcond, , facility-idef, The optional Facility word set
13086: @subsection Ambiguous conditions
13087: @c ---------------------------------------------------------------------
13088: @cindex facility words, ambiguous conditions
13089: @cindex ambiguous conditions, facility words
13090:
13091: @table @i
13092: @item @code{AT-XY} can't be performed on user output device:
13093: @cindex @code{AT-XY} can't be performed on user output device
13094: Largely terminal dependent. No range checks are done on the arguments.
13095: No errors are reported. You may see some garbage appearing, you may see
13096: simply nothing happen.
13097:
13098: @end table
13099:
13100:
13101: @c =====================================================================
13102: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
13103: @section The optional File-Access word set
13104: @c =====================================================================
13105: @cindex system documentation, file words
13106: @cindex file words, system documentation
13107:
13108: @menu
13109: * file-idef:: Implementation Defined Options
13110: * file-ambcond:: Ambiguous Conditions
13111: @end menu
13112:
13113: @c ---------------------------------------------------------------------
13114: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
13115: @subsection Implementation Defined Options
13116: @c ---------------------------------------------------------------------
13117: @cindex implementation-defined options, file words
13118: @cindex file words, implementation-defined options
13119:
13120: @table @i
13121: @item file access methods used:
13122: @cindex file access methods used
13123: @code{R/O}, @code{R/W} and @code{BIN} work as you would
13124: expect. @code{W/O} translates into the C file opening mode @code{w} (or
13125: @code{wb}): The file is cleared, if it exists, and created, if it does
13126: not (with both @code{open-file} and @code{create-file}). Under Unix
13127: @code{create-file} creates a file with 666 permissions modified by your
13128: umask.
13129:
13130: @item file exceptions:
13131: @cindex file exceptions
13132: The file words do not raise exceptions (except, perhaps, memory access
13133: faults when you pass illegal addresses or file-ids).
13134:
13135: @item file line terminator:
13136: @cindex file line terminator
13137: System-dependent. Gforth uses C's newline character as line
13138: terminator. What the actual character code(s) of this are is
13139: system-dependent.
13140:
13141: @item file name format:
13142: @cindex file name format
13143: System dependent. Gforth just uses the file name format of your OS.
13144:
13145: @item information returned by @code{FILE-STATUS}:
13146: @cindex @code{FILE-STATUS}, returned information
13147: @code{FILE-STATUS} returns the most powerful file access mode allowed
13148: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
13149: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
13150: along with the returned mode.
13151:
13152: @item input file state after an exception when including source:
13153: @cindex exception when including source
13154: All files that are left via the exception are closed.
13155:
1.29 crook 13156: @item @i{ior} values and meaning:
13157: @cindex @i{ior} values and meaning
1.68 anton 13158: @cindex @i{wior} values and meaning
1.29 crook 13159: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 13160: intended as throw codes. They typically are in the range
13161: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 13162: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 13163:
13164: @item maximum depth of file input nesting:
13165: @cindex maximum depth of file input nesting
13166: @cindex file input nesting, maximum depth
13167: limited by the amount of return stack, locals/TIB stack, and the number
13168: of open files available. This should not give you troubles.
13169:
13170: @item maximum size of input line:
13171: @cindex maximum size of input line
13172: @cindex input line size, maximum
13173: @code{/line}. Currently 255.
13174:
13175: @item methods of mapping block ranges to files:
13176: @cindex mapping block ranges to files
13177: @cindex files containing blocks
13178: @cindex blocks in files
13179: By default, blocks are accessed in the file @file{blocks.fb} in the
13180: current working directory. The file can be switched with @code{USE}.
13181:
13182: @item number of string buffers provided by @code{S"}:
13183: @cindex @code{S"}, number of string buffers
13184: 1
13185:
13186: @item size of string buffer used by @code{S"}:
13187: @cindex @code{S"}, size of string buffer
13188: @code{/line}. currently 255.
13189:
13190: @end table
13191:
13192: @c ---------------------------------------------------------------------
13193: @node file-ambcond, , file-idef, The optional File-Access word set
13194: @subsection Ambiguous conditions
13195: @c ---------------------------------------------------------------------
13196: @cindex file words, ambiguous conditions
13197: @cindex ambiguous conditions, file words
13198:
13199: @table @i
13200: @item attempting to position a file outside its boundaries:
13201: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
13202: @code{REPOSITION-FILE} is performed as usual: Afterwards,
13203: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
13204:
13205: @item attempting to read from file positions not yet written:
13206: @cindex reading from file positions not yet written
13207: End-of-file, i.e., zero characters are read and no error is reported.
13208:
1.29 crook 13209: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
13210: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
1.1 anton 13211: An appropriate exception may be thrown, but a memory fault or other
13212: problem is more probable.
13213:
1.29 crook 13214: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
13215: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
13216: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
13217: The @i{ior} produced by the operation, that discovered the problem, is
1.1 anton 13218: thrown.
13219:
13220: @item named file cannot be opened (@code{INCLUDED}):
13221: @cindex @code{INCLUDED}, named file cannot be opened
1.29 crook 13222: The @i{ior} produced by @code{open-file} is thrown.
1.1 anton 13223:
13224: @item requesting an unmapped block number:
13225: @cindex unmapped block numbers
13226: There are no unmapped legal block numbers. On some operating systems,
13227: writing a block with a large number may overflow the file system and
13228: have an error message as consequence.
13229:
13230: @item using @code{source-id} when @code{blk} is non-zero:
13231: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
13232: @code{source-id} performs its function. Typically it will give the id of
13233: the source which loaded the block. (Better ideas?)
13234:
13235: @end table
13236:
13237:
13238: @c =====================================================================
13239: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
13240: @section The optional Floating-Point word set
13241: @c =====================================================================
13242: @cindex system documentation, floating-point words
13243: @cindex floating-point words, system documentation
13244:
13245: @menu
13246: * floating-idef:: Implementation Defined Options
13247: * floating-ambcond:: Ambiguous Conditions
13248: @end menu
13249:
13250:
13251: @c ---------------------------------------------------------------------
13252: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
13253: @subsection Implementation Defined Options
13254: @c ---------------------------------------------------------------------
13255: @cindex implementation-defined options, floating-point words
13256: @cindex floating-point words, implementation-defined options
13257:
13258: @table @i
13259: @item format and range of floating point numbers:
13260: @cindex format and range of floating point numbers
13261: @cindex floating point numbers, format and range
13262: System-dependent; the @code{double} type of C.
13263:
1.29 crook 13264: @item results of @code{REPRESENT} when @i{float} is out of range:
13265: @cindex @code{REPRESENT}, results when @i{float} is out of range
1.1 anton 13266: System dependent; @code{REPRESENT} is implemented using the C library
13267: function @code{ecvt()} and inherits its behaviour in this respect.
13268:
13269: @item rounding or truncation of floating-point numbers:
13270: @cindex rounding of floating-point numbers
13271: @cindex truncation of floating-point numbers
13272: @cindex floating-point numbers, rounding or truncation
13273: System dependent; the rounding behaviour is inherited from the hosting C
13274: compiler. IEEE-FP-based (i.e., most) systems by default round to
13275: nearest, and break ties by rounding to even (i.e., such that the last
13276: bit of the mantissa is 0).
13277:
13278: @item size of floating-point stack:
13279: @cindex floating-point stack size
13280: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
13281: the floating-point stack (in floats). You can specify this on startup
13282: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
13283:
13284: @item width of floating-point stack:
13285: @cindex floating-point stack width
13286: @code{1 floats}.
13287:
13288: @end table
13289:
13290:
13291: @c ---------------------------------------------------------------------
13292: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
13293: @subsection Ambiguous conditions
13294: @c ---------------------------------------------------------------------
13295: @cindex floating-point words, ambiguous conditions
13296: @cindex ambiguous conditions, floating-point words
13297:
13298: @table @i
13299: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
13300: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
13301: System-dependent. Typically results in a @code{-23 THROW} like other
13302: alignment violations.
13303:
13304: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
13305: @cindex @code{f@@} used with an address that is not float aligned
13306: @cindex @code{f!} used with an address that is not float aligned
13307: System-dependent. Typically results in a @code{-23 THROW} like other
13308: alignment violations.
13309:
13310: @item floating-point result out of range:
13311: @cindex floating-point result out of range
1.80 anton 13312: System-dependent. Can result in a @code{-43 throw} (floating point
13313: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
13314: (floating point inexact result), @code{-55 THROW} (Floating-point
1.1 anton 13315: unidentified fault), or can produce a special value representing, e.g.,
13316: Infinity.
13317:
13318: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
13319: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
13320: System-dependent. Typically results in an alignment fault like other
13321: alignment violations.
13322:
1.35 anton 13323: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
13324: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1 anton 13325: The floating-point number is converted into decimal nonetheless.
13326:
13327: @item Both arguments are equal to zero (@code{FATAN2}):
13328: @cindex @code{FATAN2}, both arguments are equal to zero
13329: System-dependent. @code{FATAN2} is implemented using the C library
13330: function @code{atan2()}.
13331:
1.29 crook 13332: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
13333: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
13334: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1 anton 13335: because of small errors and the tan will be a very large (or very small)
13336: but finite number.
13337:
1.29 crook 13338: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
13339: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1 anton 13340: The result is rounded to the nearest float.
13341:
13342: @item dividing by zero:
13343: @cindex dividing by zero, floating-point
13344: @cindex floating-point dividing by zero
13345: @cindex floating-point unidentified fault, FP divide-by-zero
1.80 anton 13346: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
13347: (floating point divide by zero) or @code{-55 throw} (Floating-point
13348: unidentified fault).
1.1 anton 13349:
13350: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
13351: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
13352: System dependent. On IEEE-FP based systems the number is converted into
13353: an infinity.
13354:
1.29 crook 13355: @item @i{float}<1 (@code{FACOSH}):
13356: @cindex @code{FACOSH}, @i{float}<1
1.1 anton 13357: @cindex floating-point unidentified fault, @code{FACOSH}
1.80 anton 13358: Platform-dependent; on IEEE-FP systems typically produces a NaN.
1.1 anton 13359:
1.29 crook 13360: @item @i{float}=<-1 (@code{FLNP1}):
13361: @cindex @code{FLNP1}, @i{float}=<-1
1.1 anton 13362: @cindex floating-point unidentified fault, @code{FLNP1}
1.80 anton 13363: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13364: negative infinity for @i{float}=-1).
1.1 anton 13365:
1.29 crook 13366: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13367: @cindex @code{FLN}, @i{float}=<0
13368: @cindex @code{FLOG}, @i{float}=<0
1.1 anton 13369: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
1.80 anton 13370: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13371: negative infinity for @i{float}=0).
1.1 anton 13372:
1.29 crook 13373: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
13374: @cindex @code{FASINH}, @i{float}<0
13375: @cindex @code{FSQRT}, @i{float}<0
1.1 anton 13376: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
1.80 anton 13377: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
13378: @code{fasinh} some platforms produce a NaN, others a number (bug in the
13379: C library?).
1.1 anton 13380:
1.29 crook 13381: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
13382: @cindex @code{FACOS}, |@i{float}|>1
13383: @cindex @code{FASIN}, |@i{float}|>1
13384: @cindex @code{FATANH}, |@i{float}|>1
1.1 anton 13385: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
1.80 anton 13386: Platform-dependent; IEEE-FP systems typically produce a NaN.
1.1 anton 13387:
1.29 crook 13388: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
13389: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1 anton 13390: @cindex floating-point unidentified fault, @code{F>D}
1.80 anton 13391: Platform-dependent; typically, some double number is produced and no
13392: error is reported.
1.1 anton 13393:
13394: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
13395: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
1.80 anton 13396: @code{Precision} characters of the numeric output area are used. If
13397: @code{precision} is too high, these words will smash the data or code
13398: close to @code{here}.
1.1 anton 13399: @end table
13400:
13401: @c =====================================================================
13402: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
13403: @section The optional Locals word set
13404: @c =====================================================================
13405: @cindex system documentation, locals words
13406: @cindex locals words, system documentation
13407:
13408: @menu
13409: * locals-idef:: Implementation Defined Options
13410: * locals-ambcond:: Ambiguous Conditions
13411: @end menu
13412:
13413:
13414: @c ---------------------------------------------------------------------
13415: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
13416: @subsection Implementation Defined Options
13417: @c ---------------------------------------------------------------------
13418: @cindex implementation-defined options, locals words
13419: @cindex locals words, implementation-defined options
13420:
13421: @table @i
13422: @item maximum number of locals in a definition:
13423: @cindex maximum number of locals in a definition
13424: @cindex locals, maximum number in a definition
13425: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
13426: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
13427: characters. The number of locals in a definition is bounded by the size
13428: of locals-buffer, which contains the names of the locals.
13429:
13430: @end table
13431:
13432:
13433: @c ---------------------------------------------------------------------
13434: @node locals-ambcond, , locals-idef, The optional Locals word set
13435: @subsection Ambiguous conditions
13436: @c ---------------------------------------------------------------------
13437: @cindex locals words, ambiguous conditions
13438: @cindex ambiguous conditions, locals words
13439:
13440: @table @i
13441: @item executing a named local in interpretation state:
13442: @cindex local in interpretation state
13443: @cindex Interpreting a compile-only word, for a local
13444: Locals have no interpretation semantics. If you try to perform the
13445: interpretation semantics, you will get a @code{-14 throw} somewhere
13446: (Interpreting a compile-only word). If you perform the compilation
13447: semantics, the locals access will be compiled (irrespective of state).
13448:
1.29 crook 13449: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1 anton 13450: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13451: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13452: @cindex Invalid name argument, @code{TO}
13453: @code{-32 throw} (Invalid name argument)
13454:
13455: @end table
13456:
13457:
13458: @c =====================================================================
13459: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13460: @section The optional Memory-Allocation word set
13461: @c =====================================================================
13462: @cindex system documentation, memory-allocation words
13463: @cindex memory-allocation words, system documentation
13464:
13465: @menu
13466: * memory-idef:: Implementation Defined Options
13467: @end menu
13468:
13469:
13470: @c ---------------------------------------------------------------------
13471: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13472: @subsection Implementation Defined Options
13473: @c ---------------------------------------------------------------------
13474: @cindex implementation-defined options, memory-allocation words
13475: @cindex memory-allocation words, implementation-defined options
13476:
13477: @table @i
1.29 crook 13478: @item values and meaning of @i{ior}:
13479: @cindex @i{ior} values and meaning
13480: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 13481: intended as throw codes. They typically are in the range
13482: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 13483: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 13484:
13485: @end table
13486:
13487: @c =====================================================================
13488: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13489: @section The optional Programming-Tools word set
13490: @c =====================================================================
13491: @cindex system documentation, programming-tools words
13492: @cindex programming-tools words, system documentation
13493:
13494: @menu
13495: * programming-idef:: Implementation Defined Options
13496: * programming-ambcond:: Ambiguous Conditions
13497: @end menu
13498:
13499:
13500: @c ---------------------------------------------------------------------
13501: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13502: @subsection Implementation Defined Options
13503: @c ---------------------------------------------------------------------
13504: @cindex implementation-defined options, programming-tools words
13505: @cindex programming-tools words, implementation-defined options
13506:
13507: @table @i
13508: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13509: @cindex @code{;CODE} ending sequence
13510: @cindex @code{CODE} ending sequence
13511: @code{END-CODE}
13512:
13513: @item manner of processing input following @code{;CODE} and @code{CODE}:
13514: @cindex @code{;CODE}, processing input
13515: @cindex @code{CODE}, processing input
13516: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13517: the input is processed by the text interpreter, (starting) in interpret
13518: state.
13519:
13520: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13521: @cindex @code{ASSEMBLER}, search order capability
13522: The ANS Forth search order word set.
13523:
13524: @item source and format of display by @code{SEE}:
13525: @cindex @code{SEE}, source and format of output
1.80 anton 13526: The source for @code{see} is the executable code used by the inner
1.1 anton 13527: interpreter. The current @code{see} tries to output Forth source code
1.80 anton 13528: (and on some platforms, assembly code for primitives) as well as
13529: possible.
1.1 anton 13530:
13531: @end table
13532:
13533: @c ---------------------------------------------------------------------
13534: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13535: @subsection Ambiguous conditions
13536: @c ---------------------------------------------------------------------
13537: @cindex programming-tools words, ambiguous conditions
13538: @cindex ambiguous conditions, programming-tools words
13539:
13540: @table @i
13541:
1.21 crook 13542: @item deleting the compilation word list (@code{FORGET}):
13543: @cindex @code{FORGET}, deleting the compilation word list
1.1 anton 13544: Not implemented (yet).
13545:
1.29 crook 13546: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13547: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13548: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1 anton 13549: @cindex control-flow stack underflow
13550: This typically results in an @code{abort"} with a descriptive error
13551: message (may change into a @code{-22 throw} (Control structure mismatch)
13552: in the future). You may also get a memory access error. If you are
13553: unlucky, this ambiguous condition is not caught.
13554:
1.29 crook 13555: @item @i{name} can't be found (@code{FORGET}):
13556: @cindex @code{FORGET}, @i{name} can't be found
1.1 anton 13557: Not implemented (yet).
13558:
1.29 crook 13559: @item @i{name} not defined via @code{CREATE}:
13560: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1 anton 13561: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13562: the execution semantics of the last defined word no matter how it was
13563: defined.
13564:
13565: @item @code{POSTPONE} applied to @code{[IF]}:
13566: @cindex @code{POSTPONE} applied to @code{[IF]}
13567: @cindex @code{[IF]} and @code{POSTPONE}
13568: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13569: equivalent to @code{[IF]}.
13570:
13571: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13572: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13573: Continue in the same state of conditional compilation in the next outer
13574: input source. Currently there is no warning to the user about this.
13575:
13576: @item removing a needed definition (@code{FORGET}):
13577: @cindex @code{FORGET}, removing a needed definition
13578: Not implemented (yet).
13579:
13580: @end table
13581:
13582:
13583: @c =====================================================================
13584: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13585: @section The optional Search-Order word set
13586: @c =====================================================================
13587: @cindex system documentation, search-order words
13588: @cindex search-order words, system documentation
13589:
13590: @menu
13591: * search-idef:: Implementation Defined Options
13592: * search-ambcond:: Ambiguous Conditions
13593: @end menu
13594:
13595:
13596: @c ---------------------------------------------------------------------
13597: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13598: @subsection Implementation Defined Options
13599: @c ---------------------------------------------------------------------
13600: @cindex implementation-defined options, search-order words
13601: @cindex search-order words, implementation-defined options
13602:
13603: @table @i
13604: @item maximum number of word lists in search order:
13605: @cindex maximum number of word lists in search order
13606: @cindex search order, maximum depth
13607: @code{s" wordlists" environment? drop .}. Currently 16.
13608:
13609: @item minimum search order:
13610: @cindex minimum search order
13611: @cindex search order, minimum
13612: @code{root root}.
13613:
13614: @end table
13615:
13616: @c ---------------------------------------------------------------------
13617: @node search-ambcond, , search-idef, The optional Search-Order word set
13618: @subsection Ambiguous conditions
13619: @c ---------------------------------------------------------------------
13620: @cindex search-order words, ambiguous conditions
13621: @cindex ambiguous conditions, search-order words
13622:
13623: @table @i
1.21 crook 13624: @item changing the compilation word list (during compilation):
13625: @cindex changing the compilation word list (during compilation)
13626: @cindex compilation word list, change before definition ends
13627: The word is entered into the word list that was the compilation word list
1.1 anton 13628: at the start of the definition. Any changes to the name field (e.g.,
13629: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13630: are applied to the latest defined word (as reported by @code{last} or
1.21 crook 13631: @code{lastxt}), if possible, irrespective of the compilation word list.
1.1 anton 13632:
13633: @item search order empty (@code{previous}):
13634: @cindex @code{previous}, search order empty
1.26 crook 13635: @cindex vocstack empty, @code{previous}
1.1 anton 13636: @code{abort" Vocstack empty"}.
13637:
13638: @item too many word lists in search order (@code{also}):
13639: @cindex @code{also}, too many word lists in search order
1.26 crook 13640: @cindex vocstack full, @code{also}
1.1 anton 13641: @code{abort" Vocstack full"}.
13642:
13643: @end table
13644:
13645: @c ***************************************************************
1.65 anton 13646: @node Standard vs Extensions, Model, ANS conformance, Top
13647: @chapter Should I use Gforth extensions?
13648: @cindex Gforth extensions
13649:
13650: As you read through the rest of this manual, you will see documentation
13651: for @i{Standard} words, and documentation for some appealing Gforth
13652: @i{extensions}. You might ask yourself the question: @i{``Should I
13653: restrict myself to the standard, or should I use the extensions?''}
13654:
13655: The answer depends on the goals you have for the program you are working
13656: on:
13657:
13658: @itemize @bullet
13659:
13660: @item Is it just for yourself or do you want to share it with others?
13661:
13662: @item
13663: If you want to share it, do the others all use Gforth?
13664:
13665: @item
13666: If it is just for yourself, do you want to restrict yourself to Gforth?
13667:
13668: @end itemize
13669:
13670: If restricting the program to Gforth is ok, then there is no reason not
13671: to use extensions. It is still a good idea to keep to the standard
13672: where it is easy, in case you want to reuse these parts in another
13673: program that you want to be portable.
13674:
13675: If you want to be able to port the program to other Forth systems, there
13676: are the following points to consider:
13677:
13678: @itemize @bullet
13679:
13680: @item
13681: Most Forth systems that are being maintained support the ANS Forth
13682: standard. So if your program complies with the standard, it will be
13683: portable among many systems.
13684:
13685: @item
13686: A number of the Gforth extensions can be implemented in ANS Forth using
13687: public-domain files provided in the @file{compat/} directory. These are
13688: mentioned in the text in passing. There is no reason not to use these
13689: extensions, your program will still be ANS Forth compliant; just include
13690: the appropriate compat files with your program.
13691:
13692: @item
13693: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13694: analyse your program and determine what non-Standard words it relies
13695: upon. However, it does not check whether you use standard words in a
13696: non-standard way.
13697:
13698: @item
13699: Some techniques are not standardized by ANS Forth, and are hard or
13700: impossible to implement in a standard way, but can be implemented in
13701: most Forth systems easily, and usually in similar ways (e.g., accessing
13702: word headers). Forth has a rich historical precedent for programmers
13703: taking advantage of implementation-dependent features of their tools
13704: (for example, relying on a knowledge of the dictionary
13705: structure). Sometimes these techniques are necessary to extract every
13706: last bit of performance from the hardware, sometimes they are just a
13707: programming shorthand.
13708:
13709: @item
13710: Does using a Gforth extension save more work than the porting this part
13711: to other Forth systems (if any) will cost?
13712:
13713: @item
13714: Is the additional functionality worth the reduction in portability and
13715: the additional porting problems?
13716:
13717: @end itemize
13718:
13719: In order to perform these consideratios, you need to know what's
13720: standard and what's not. This manual generally states if something is
1.81 anton 13721: non-standard, but the authoritative source is the
13722: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
1.65 anton 13723: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13724: into the thought processes of the technical committee.
13725:
13726: Note also that portability between Forth systems is not the only
13727: portability issue; there is also the issue of portability between
13728: different platforms (processor/OS combinations).
13729:
13730: @c ***************************************************************
13731: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1 anton 13732: @chapter Model
13733:
13734: This chapter has yet to be written. It will contain information, on
13735: which internal structures you can rely.
13736:
13737: @c ***************************************************************
13738: @node Integrating Gforth, Emacs and Gforth, Model, Top
13739: @chapter Integrating Gforth into C programs
13740:
13741: This is not yet implemented.
13742:
13743: Several people like to use Forth as scripting language for applications
13744: that are otherwise written in C, C++, or some other language.
13745:
13746: The Forth system ATLAST provides facilities for embedding it into
13747: applications; unfortunately it has several disadvantages: most
13748: importantly, it is not based on ANS Forth, and it is apparently dead
13749: (i.e., not developed further and not supported). The facilities
1.21 crook 13750: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1 anton 13751: making the switch should not be hard.
13752:
13753: We also tried to design the interface such that it can easily be
13754: implemented by other Forth systems, so that we may one day arrive at a
13755: standardized interface. Such a standard interface would allow you to
13756: replace the Forth system without having to rewrite C code.
13757:
13758: You embed the Gforth interpreter by linking with the library
13759: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13760: global symbols in this library that belong to the interface, have the
13761: prefix @code{forth_}. (Global symbols that are used internally have the
13762: prefix @code{gforth_}).
13763:
13764: You can include the declarations of Forth types and the functions and
13765: variables of the interface with @code{#include <forth.h>}.
13766:
13767: Types.
13768:
13769: Variables.
13770:
13771: Data and FP Stack pointer. Area sizes.
13772:
13773: functions.
13774:
13775: forth_init(imagefile)
13776: forth_evaluate(string) exceptions?
13777: forth_goto(address) (or forth_execute(xt)?)
13778: forth_continue() (a corountining mechanism)
13779:
13780: Adding primitives.
13781:
13782: No checking.
13783:
13784: Signals?
13785:
13786: Accessing the Stacks
13787:
1.26 crook 13788: @c ******************************************************************
1.1 anton 13789: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13790: @chapter Emacs and Gforth
13791: @cindex Emacs and Gforth
13792:
13793: @cindex @file{gforth.el}
13794: @cindex @file{forth.el}
13795: @cindex Rydqvist, Goran
13796: @cindex comment editing commands
13797: @cindex @code{\}, editing with Emacs
13798: @cindex debug tracer editing commands
13799: @cindex @code{~~}, removal with Emacs
13800: @cindex Forth mode in Emacs
13801: Gforth comes with @file{gforth.el}, an improved version of
13802: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26 crook 13803: improvements are:
13804:
13805: @itemize @bullet
13806: @item
13807: A better (but still not perfect) handling of indentation.
13808: @item
13809: Comment paragraph filling (@kbd{M-q})
13810: @item
13811: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13812: @item
13813: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41 anton 13814: @item
13815: Support of the @code{info-lookup} feature for looking up the
13816: documentation of a word.
1.26 crook 13817: @end itemize
13818:
13819: I left the stuff I do not use alone, even though some of it only makes
13820: sense for TILE. To get a description of these features, enter Forth mode
13821: and type @kbd{C-h m}.
1.1 anton 13822:
13823: @cindex source location of error or debugging output in Emacs
13824: @cindex error output, finding the source location in Emacs
13825: @cindex debugging output, finding the source location in Emacs
13826: In addition, Gforth supports Emacs quite well: The source code locations
13827: given in error messages, debugging output (from @code{~~}) and failed
13828: assertion messages are in the right format for Emacs' compilation mode
13829: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13830: Manual}) so the source location corresponding to an error or other
13831: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13832: @kbd{C-c C-c} for the error under the cursor).
13833:
13834: @cindex @file{TAGS} file
13835: @cindex @file{etags.fs}
13836: @cindex viewing the source of a word in Emacs
1.43 anton 13837: @cindex @code{require}, placement in files
13838: @cindex @code{include}, placement in files
13839: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
1.26 crook 13840: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1 anton 13841: contains the definitions of all words defined afterwards. You can then
13842: find the source for a word using @kbd{M-.}. Note that emacs can use
13843: several tags files at the same time (e.g., one for the Gforth sources
13844: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13845: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13846: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43 anton 13847: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13848: with @file{etags.fs}, you should avoid putting definitions both before
13849: and after @code{require} etc., otherwise you will see the same file
13850: visited several times by commands like @code{tags-search}.
1.1 anton 13851:
1.41 anton 13852: @cindex viewing the documentation of a word in Emacs
13853: @cindex context-sensitive help
13854: Moreover, for words documented in this manual, you can look up the
13855: glossary entry quickly by using @kbd{C-h TAB}
1.80 anton 13856: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
1.41 anton 13857: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
1.42 anton 13858: later and does not work for words containing @code{:}.
1.41 anton 13859:
13860:
1.1 anton 13861: @cindex @file{.emacs}
13862: To get all these benefits, add the following lines to your @file{.emacs}
13863: file:
13864:
13865: @example
13866: (autoload 'forth-mode "gforth.el")
13867: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13868: @end example
13869:
1.26 crook 13870: @c ******************************************************************
1.1 anton 13871: @node Image Files, Engine, Emacs and Gforth, Top
13872: @chapter Image Files
1.26 crook 13873: @cindex image file
13874: @cindex @file{.fi} files
1.1 anton 13875: @cindex precompiled Forth code
13876: @cindex dictionary in persistent form
13877: @cindex persistent form of dictionary
13878:
13879: An image file is a file containing an image of the Forth dictionary,
13880: i.e., compiled Forth code and data residing in the dictionary. By
13881: convention, we use the extension @code{.fi} for image files.
13882:
13883: @menu
1.18 anton 13884: * Image Licensing Issues:: Distribution terms for images.
13885: * Image File Background:: Why have image files?
1.67 anton 13886: * Non-Relocatable Image Files:: don't always work.
1.18 anton 13887: * Data-Relocatable Image Files:: are better.
1.67 anton 13888: * Fully Relocatable Image Files:: better yet.
1.18 anton 13889: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.29 crook 13890: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.18 anton 13891: * Modifying the Startup Sequence:: and turnkey applications.
1.1 anton 13892: @end menu
13893:
1.18 anton 13894: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13895: @section Image Licensing Issues
13896: @cindex license for images
13897: @cindex image license
13898:
13899: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13900: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13901: original image; i.e., according to copyright law it is a derived work of
13902: the original image.
13903:
13904: Since Gforth is distributed under the GNU GPL, the newly created image
13905: falls under the GNU GPL, too. In particular, this means that if you
13906: distribute the image, you have to make all of the sources for the image
13907: available, including those you wrote. For details see @ref{License, ,
13908: GNU General Public License (Section 3)}.
13909:
13910: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13911: contains only code compiled from the sources you gave it; if none of
13912: these sources is under the GPL, the terms discussed above do not apply
13913: to the image. However, if your image needs an engine (a gforth binary)
13914: that is under the GPL, you should make sure that you distribute both in
13915: a way that is at most a @emph{mere aggregation}, if you don't want the
13916: terms of the GPL to apply to the image.
13917:
13918: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1 anton 13919: @section Image File Background
13920: @cindex image file background
13921:
1.80 anton 13922: Gforth consists not only of primitives (in the engine), but also of
1.1 anton 13923: definitions written in Forth. Since the Forth compiler itself belongs to
13924: those definitions, it is not possible to start the system with the
1.80 anton 13925: engine and the Forth source alone. Therefore we provide the Forth
1.26 crook 13926: code as an image file in nearly executable form. When Gforth starts up,
13927: a C routine loads the image file into memory, optionally relocates the
13928: addresses, then sets up the memory (stacks etc.) according to
13929: information in the image file, and (finally) starts executing Forth
13930: code.
1.1 anton 13931:
13932: The image file variants represent different compromises between the
13933: goals of making it easy to generate image files and making them
13934: portable.
13935:
13936: @cindex relocation at run-time
1.26 crook 13937: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1 anton 13938: run-time. This avoids many of the complications discussed below (image
13939: files are data relocatable without further ado), but costs performance
13940: (one addition per memory access).
13941:
13942: @cindex relocation at load-time
1.26 crook 13943: By contrast, the Gforth loader performs relocation at image load time. The
13944: loader also has to replace tokens that represent primitive calls with the
1.1 anton 13945: appropriate code-field addresses (or code addresses in the case of
13946: direct threading).
13947:
13948: There are three kinds of image files, with different degrees of
13949: relocatability: non-relocatable, data-relocatable, and fully relocatable
13950: image files.
13951:
13952: @cindex image file loader
13953: @cindex relocating loader
13954: @cindex loader for image files
13955: These image file variants have several restrictions in common; they are
13956: caused by the design of the image file loader:
13957:
13958: @itemize @bullet
13959: @item
13960: There is only one segment; in particular, this means, that an image file
13961: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26 crook 13962: them). The contents of the stacks are not represented, either.
1.1 anton 13963:
13964: @item
13965: The only kinds of relocation supported are: adding the same offset to
13966: all cells that represent data addresses; and replacing special tokens
13967: with code addresses or with pieces of machine code.
13968:
13969: If any complex computations involving addresses are performed, the
13970: results cannot be represented in the image file. Several applications that
13971: use such computations come to mind:
13972: @itemize @minus
13973: @item
13974: Hashing addresses (or data structures which contain addresses) for table
13975: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13976: purpose, you will have no problem, because the hash tables are
13977: recomputed automatically when the system is started. If you use your own
13978: hash tables, you will have to do something similar.
13979:
13980: @item
13981: There's a cute implementation of doubly-linked lists that uses
13982: @code{XOR}ed addresses. You could represent such lists as singly-linked
13983: in the image file, and restore the doubly-linked representation on
13984: startup.@footnote{In my opinion, though, you should think thrice before
13985: using a doubly-linked list (whatever implementation).}
13986:
13987: @item
13988: The code addresses of run-time routines like @code{docol:} cannot be
13989: represented in the image file (because their tokens would be replaced by
13990: machine code in direct threaded implementations). As a workaround,
13991: compute these addresses at run-time with @code{>code-address} from the
13992: executions tokens of appropriate words (see the definitions of
1.80 anton 13993: @code{docol:} and friends in @file{kernel/getdoers.fs}).
1.1 anton 13994:
13995: @item
13996: On many architectures addresses are represented in machine code in some
13997: shifted or mangled form. You cannot put @code{CODE} words that contain
13998: absolute addresses in this form in a relocatable image file. Workarounds
13999: are representing the address in some relative form (e.g., relative to
14000: the CFA, which is present in some register), or loading the address from
14001: a place where it is stored in a non-mangled form.
14002: @end itemize
14003: @end itemize
14004:
14005: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
14006: @section Non-Relocatable Image Files
14007: @cindex non-relocatable image files
1.26 crook 14008: @cindex image file, non-relocatable
1.1 anton 14009:
14010: These files are simple memory dumps of the dictionary. They are specific
14011: to the executable (i.e., @file{gforth} file) they were created
14012: with. What's worse, they are specific to the place on which the
14013: dictionary resided when the image was created. Now, there is no
14014: guarantee that the dictionary will reside at the same place the next
14015: time you start Gforth, so there's no guarantee that a non-relocatable
14016: image will work the next time (Gforth will complain instead of crashing,
14017: though).
14018:
14019: You can create a non-relocatable image file with
14020:
1.44 crook 14021:
1.1 anton 14022: doc-savesystem
14023:
1.44 crook 14024:
1.1 anton 14025: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
14026: @section Data-Relocatable Image Files
14027: @cindex data-relocatable image files
1.26 crook 14028: @cindex image file, data-relocatable
1.1 anton 14029:
14030: These files contain relocatable data addresses, but fixed code addresses
14031: (instead of tokens). They are specific to the executable (i.e.,
14032: @file{gforth} file) they were created with. For direct threading on some
14033: architectures (e.g., the i386), data-relocatable images do not work. You
14034: get a data-relocatable image, if you use @file{gforthmi} with a
14035: Gforth binary that is not doubly indirect threaded (@pxref{Fully
14036: Relocatable Image Files}).
14037:
14038: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
14039: @section Fully Relocatable Image Files
14040: @cindex fully relocatable image files
1.26 crook 14041: @cindex image file, fully relocatable
1.1 anton 14042:
14043: @cindex @file{kern*.fi}, relocatability
14044: @cindex @file{gforth.fi}, relocatability
14045: These image files have relocatable data addresses, and tokens for code
14046: addresses. They can be used with different binaries (e.g., with and
14047: without debugging) on the same machine, and even across machines with
14048: the same data formats (byte order, cell size, floating point
14049: format). However, they are usually specific to the version of Gforth
14050: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
14051: are fully relocatable.
14052:
14053: There are two ways to create a fully relocatable image file:
14054:
14055: @menu
1.29 crook 14056: * gforthmi:: The normal way
1.1 anton 14057: * cross.fs:: The hard way
14058: @end menu
14059:
14060: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
14061: @subsection @file{gforthmi}
14062: @cindex @file{comp-i.fs}
14063: @cindex @file{gforthmi}
14064:
14065: You will usually use @file{gforthmi}. If you want to create an
1.29 crook 14066: image @i{file} that contains everything you would load by invoking
14067: Gforth with @code{gforth @i{options}}, you simply say:
1.1 anton 14068: @example
1.29 crook 14069: gforthmi @i{file} @i{options}
1.1 anton 14070: @end example
14071:
14072: E.g., if you want to create an image @file{asm.fi} that has the file
14073: @file{asm.fs} loaded in addition to the usual stuff, you could do it
14074: like this:
14075:
14076: @example
14077: gforthmi asm.fi asm.fs
14078: @end example
14079:
1.27 crook 14080: @file{gforthmi} is implemented as a sh script and works like this: It
14081: produces two non-relocatable images for different addresses and then
14082: compares them. Its output reflects this: first you see the output (if
1.62 crook 14083: any) of the two Gforth invocations that produce the non-relocatable image
1.27 crook 14084: files, then you see the output of the comparing program: It displays the
14085: offset used for data addresses and the offset used for code addresses;
1.1 anton 14086: moreover, for each cell that cannot be represented correctly in the
1.44 crook 14087: image files, it displays a line like this:
1.1 anton 14088:
14089: @example
14090: 78DC BFFFFA50 BFFFFA40
14091: @end example
14092:
14093: This means that at offset $78dc from @code{forthstart}, one input image
14094: contains $bffffa50, and the other contains $bffffa40. Since these cells
14095: cannot be represented correctly in the output image, you should examine
14096: these places in the dictionary and verify that these cells are dead
14097: (i.e., not read before they are written).
1.39 anton 14098:
14099: @cindex --application, @code{gforthmi} option
14100: If you insert the option @code{--application} in front of the image file
14101: name, you will get an image that uses the @code{--appl-image} option
14102: instead of the @code{--image-file} option (@pxref{Invoking
14103: Gforth}). When you execute such an image on Unix (by typing the image
14104: name as command), the Gforth engine will pass all options to the image
14105: instead of trying to interpret them as engine options.
1.1 anton 14106:
1.27 crook 14107: If you type @file{gforthmi} with no arguments, it prints some usage
14108: instructions.
14109:
1.1 anton 14110: @cindex @code{savesystem} during @file{gforthmi}
14111: @cindex @code{bye} during @file{gforthmi}
14112: @cindex doubly indirect threaded code
1.44 crook 14113: @cindex environment variables
14114: @cindex @code{GFORTHD} -- environment variable
14115: @cindex @code{GFORTH} -- environment variable
1.1 anton 14116: @cindex @code{gforth-ditc}
1.29 crook 14117: There are a few wrinkles: After processing the passed @i{options}, the
1.1 anton 14118: words @code{savesystem} and @code{bye} must be visible. A special doubly
14119: indirect threaded version of the @file{gforth} executable is used for
1.62 crook 14120: creating the non-relocatable images; you can pass the exact filename of
1.1 anton 14121: this executable through the environment variable @code{GFORTHD}
14122: (default: @file{gforth-ditc}); if you pass a version that is not doubly
14123: indirect threaded, you will not get a fully relocatable image, but a
1.27 crook 14124: data-relocatable image (because there is no code address offset). The
14125: normal @file{gforth} executable is used for creating the relocatable
14126: image; you can pass the exact filename of this executable through the
14127: environment variable @code{GFORTH}.
1.1 anton 14128:
14129: @node cross.fs, , gforthmi, Fully Relocatable Image Files
14130: @subsection @file{cross.fs}
14131: @cindex @file{cross.fs}
14132: @cindex cross-compiler
14133: @cindex metacompiler
1.47 crook 14134: @cindex target compiler
1.1 anton 14135:
14136: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47 crook 14137: programming language (@pxref{Cross Compiler}).
1.1 anton 14138:
1.47 crook 14139: @code{cross} allows you to create image files for machines with
1.1 anton 14140: different data sizes and data formats than the one used for generating
14141: the image file. You can also use it to create an application image that
14142: does not contain a Forth compiler. These features are bought with
14143: restrictions and inconveniences in programming. E.g., addresses have to
14144: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
14145: order to make the code relocatable.
14146:
14147:
14148: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
14149: @section Stack and Dictionary Sizes
14150: @cindex image file, stack and dictionary sizes
14151: @cindex dictionary size default
14152: @cindex stack size default
14153:
14154: If you invoke Gforth with a command line flag for the size
14155: (@pxref{Invoking Gforth}), the size you specify is stored in the
14156: dictionary. If you save the dictionary with @code{savesystem} or create
14157: an image with @file{gforthmi}, this size will become the default
14158: for the resulting image file. E.g., the following will create a
1.21 crook 14159: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1 anton 14160:
14161: @example
14162: gforthmi gforth.fi -m 1M
14163: @end example
14164:
14165: In other words, if you want to set the default size for the dictionary
14166: and the stacks of an image, just invoke @file{gforthmi} with the
14167: appropriate options when creating the image.
14168:
14169: @cindex stack size, cache-friendly
14170: Note: For cache-friendly behaviour (i.e., good performance), you should
14171: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
14172: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
14173: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
14174:
14175: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
14176: @section Running Image Files
14177: @cindex running image files
14178: @cindex invoking image files
14179: @cindex image file invocation
14180:
14181: @cindex -i, invoke image file
14182: @cindex --image file, invoke image file
1.29 crook 14183: You can invoke Gforth with an image file @i{image} instead of the
1.1 anton 14184: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
14185: @example
1.29 crook 14186: gforth -i @i{image}
1.1 anton 14187: @end example
14188:
14189: @cindex executable image file
1.26 crook 14190: @cindex image file, executable
1.1 anton 14191: If your operating system supports starting scripts with a line of the
14192: form @code{#! ...}, you just have to type the image file name to start
14193: Gforth with this image file (note that the file extension @code{.fi} is
1.29 crook 14194: just a convention). I.e., to run Gforth with the image file @i{image},
14195: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27 crook 14196: This works because every @code{.fi} file starts with a line of this
14197: format:
14198:
14199: @example
14200: #! /usr/local/bin/gforth-0.4.0 -i
14201: @end example
14202:
14203: The file and pathname for the Gforth engine specified on this line is
14204: the specific Gforth executable that it was built against; i.e. the value
14205: of the environment variable @code{GFORTH} at the time that
14206: @file{gforthmi} was executed.
1.1 anton 14207:
1.27 crook 14208: You can make use of the same shell capability to make a Forth source
14209: file into an executable. For example, if you place this text in a file:
1.26 crook 14210:
14211: @example
14212: #! /usr/local/bin/gforth
14213:
14214: ." Hello, world" CR
14215: bye
14216: @end example
14217:
14218: @noindent
1.27 crook 14219: and then make the file executable (chmod +x in Unix), you can run it
1.26 crook 14220: directly from the command line. The sequence @code{#!} is used in two
14221: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29 crook 14222: system@footnote{The Unix kernel actually recognises two types of files:
14223: executable files and files of data, where the data is processed by an
14224: interpreter that is specified on the ``interpreter line'' -- the first
14225: line of the file, starting with the sequence #!. There may be a small
14226: limit (e.g., 32) on the number of characters that may be specified on
14227: the interpreter line.} secondly it is treated as a comment character by
14228: Gforth. Because of the second usage, a space is required between
1.80 anton 14229: @code{#!} and the path to the executable (moreover, some Unixes
14230: require the sequence @code{#! /}).
1.27 crook 14231:
14232: The disadvantage of this latter technique, compared with using
1.80 anton 14233: @file{gforthmi}, is that it is slightly slower; the Forth source code is
14234: compiled on-the-fly, each time the program is invoked.
1.26 crook 14235:
1.1 anton 14236: doc-#!
14237:
1.44 crook 14238:
1.1 anton 14239: @node Modifying the Startup Sequence, , Running Image Files, Image Files
14240: @section Modifying the Startup Sequence
14241: @cindex startup sequence for image file
14242: @cindex image file initialization sequence
14243: @cindex initialization sequence of image file
14244:
14245: You can add your own initialization to the startup sequence through the
1.26 crook 14246: deferred word @code{'cold}. @code{'cold} is invoked just before the
1.80 anton 14247: image-specific command line processing (i.e., loading files and
1.26 crook 14248: evaluating (@code{-e}) strings) starts.
1.1 anton 14249:
14250: A sequence for adding your initialization usually looks like this:
14251:
14252: @example
14253: :noname
14254: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
14255: ... \ your stuff
14256: ; IS 'cold
14257: @end example
14258:
14259: @cindex turnkey image files
1.26 crook 14260: @cindex image file, turnkey applications
1.1 anton 14261: You can make a turnkey image by letting @code{'cold} execute a word
14262: (your turnkey application) that never returns; instead, it exits Gforth
14263: via @code{bye} or @code{throw}.
14264:
14265: @cindex command-line arguments, access
14266: @cindex arguments on the command line, access
14267: You can access the (image-specific) command-line arguments through the
1.26 crook 14268: variables @code{argc} and @code{argv}. @code{arg} provides convenient
1.1 anton 14269: access to @code{argv}.
14270:
1.26 crook 14271: If @code{'cold} exits normally, Gforth processes the command-line
14272: arguments as files to be loaded and strings to be evaluated. Therefore,
14273: @code{'cold} should remove the arguments it has used in this case.
14274:
1.44 crook 14275:
14276:
1.26 crook 14277: doc-'cold
1.1 anton 14278: doc-argc
14279: doc-argv
14280: doc-arg
14281:
14282:
1.44 crook 14283:
1.1 anton 14284: @c ******************************************************************
1.13 pazsan 14285: @node Engine, Binding to System Library, Image Files, Top
1.1 anton 14286: @chapter Engine
14287: @cindex engine
14288: @cindex virtual machine
14289:
1.26 crook 14290: Reading this chapter is not necessary for programming with Gforth. It
1.1 anton 14291: may be helpful for finding your way in the Gforth sources.
14292:
1.66 anton 14293: The ideas in this section have also been published in Bernd Paysan,
14294: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
14295: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
14296: Portable Forth Engine}}, EuroForth '93.
1.1 anton 14297:
14298: @menu
14299: * Portability::
14300: * Threading::
14301: * Primitives::
14302: * Performance::
14303: @end menu
14304:
14305: @node Portability, Threading, Engine, Engine
14306: @section Portability
14307: @cindex engine portability
14308:
1.26 crook 14309: An important goal of the Gforth Project is availability across a wide
14310: range of personal machines. fig-Forth, and, to a lesser extent, F83,
14311: achieved this goal by manually coding the engine in assembly language
14312: for several then-popular processors. This approach is very
14313: labor-intensive and the results are short-lived due to progress in
14314: computer architecture.
1.1 anton 14315:
14316: @cindex C, using C for the engine
14317: Others have avoided this problem by coding in C, e.g., Mitch Bradley
14318: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
14319: particularly popular for UNIX-based Forths due to the large variety of
14320: architectures of UNIX machines. Unfortunately an implementation in C
14321: does not mix well with the goals of efficiency and with using
14322: traditional techniques: Indirect or direct threading cannot be expressed
14323: in C, and switch threading, the fastest technique available in C, is
14324: significantly slower. Another problem with C is that it is very
14325: cumbersome to express double integer arithmetic.
14326:
14327: @cindex GNU C for the engine
14328: @cindex long long
14329: Fortunately, there is a portable language that does not have these
14330: limitations: GNU C, the version of C processed by the GNU C compiler
14331: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
14332: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
14333: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
14334: threading possible, its @code{long long} type (@pxref{Long Long, ,
14335: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
14336: double numbers@footnote{Unfortunately, long longs are not implemented
14337: properly on all machines (e.g., on alpha-osf1, long longs are only 64
14338: bits, the same size as longs (and pointers), but they should be twice as
1.4 anton 14339: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
1.1 anton 14340: C Manual}). So, we had to implement doubles in C after all. Still, on
14341: most machines we can use long longs and achieve better performance than
14342: with the emulation package.}. GNU C is available for free on all
14343: important (and many unimportant) UNIX machines, VMS, 80386s running
14344: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
14345: on all these machines.
14346:
14347: Writing in a portable language has the reputation of producing code that
14348: is slower than assembly. For our Forth engine we repeatedly looked at
14349: the code produced by the compiler and eliminated most compiler-induced
14350: inefficiencies by appropriate changes in the source code.
14351:
14352: @cindex explicit register declarations
14353: @cindex --enable-force-reg, configuration flag
14354: @cindex -DFORCE_REG
14355: However, register allocation cannot be portably influenced by the
14356: programmer, leading to some inefficiencies on register-starved
14357: machines. We use explicit register declarations (@pxref{Explicit Reg
14358: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
14359: improve the speed on some machines. They are turned on by using the
14360: configuration flag @code{--enable-force-reg} (@code{gcc} switch
14361: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
14362: machine, but also on the compiler version: On some machines some
14363: compiler versions produce incorrect code when certain explicit register
14364: declarations are used. So by default @code{-DFORCE_REG} is not used.
14365:
14366: @node Threading, Primitives, Portability, Engine
14367: @section Threading
14368: @cindex inner interpreter implementation
14369: @cindex threaded code implementation
14370:
14371: @cindex labels as values
14372: GNU C's labels as values extension (available since @code{gcc-2.0},
14373: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29 crook 14374: makes it possible to take the address of @i{label} by writing
14375: @code{&&@i{label}}. This address can then be used in a statement like
14376: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1 anton 14377: @code{goto x}.
14378:
1.26 crook 14379: @cindex @code{NEXT}, indirect threaded
1.1 anton 14380: @cindex indirect threaded inner interpreter
14381: @cindex inner interpreter, indirect threaded
1.26 crook 14382: With this feature an indirect threaded @code{NEXT} looks like:
1.1 anton 14383: @example
14384: cfa = *ip++;
14385: ca = *cfa;
14386: goto *ca;
14387: @end example
14388: @cindex instruction pointer
14389: For those unfamiliar with the names: @code{ip} is the Forth instruction
14390: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
14391: execution token and points to the code field of the next word to be
14392: executed; The @code{ca} (code address) fetched from there points to some
14393: executable code, e.g., a primitive or the colon definition handler
14394: @code{docol}.
14395:
1.26 crook 14396: @cindex @code{NEXT}, direct threaded
1.1 anton 14397: @cindex direct threaded inner interpreter
14398: @cindex inner interpreter, direct threaded
14399: Direct threading is even simpler:
14400: @example
14401: ca = *ip++;
14402: goto *ca;
14403: @end example
14404:
14405: Of course we have packaged the whole thing neatly in macros called
1.26 crook 14406: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1 anton 14407:
14408: @menu
14409: * Scheduling::
14410: * Direct or Indirect Threaded?::
14411: * DOES>::
14412: @end menu
14413:
14414: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
14415: @subsection Scheduling
14416: @cindex inner interpreter optimization
14417:
14418: There is a little complication: Pipelined and superscalar processors,
14419: i.e., RISC and some modern CISC machines can process independent
14420: instructions while waiting for the results of an instruction. The
14421: compiler usually reorders (schedules) the instructions in a way that
14422: achieves good usage of these delay slots. However, on our first tries
14423: the compiler did not do well on scheduling primitives. E.g., for
14424: @code{+} implemented as
14425: @example
14426: n=sp[0]+sp[1];
14427: sp++;
14428: sp[0]=n;
14429: NEXT;
14430: @end example
1.81 anton 14431: the @code{NEXT} comes strictly after the other code, i.e., there is
14432: nearly no scheduling. After a little thought the problem becomes clear:
14433: The compiler cannot know that @code{sp} and @code{ip} point to different
1.21 crook 14434: addresses (and the version of @code{gcc} we used would not know it even
14435: if it was possible), so it could not move the load of the cfa above the
14436: store to the TOS. Indeed the pointers could be the same, if code on or
14437: very near the top of stack were executed. In the interest of speed we
14438: chose to forbid this probably unused ``feature'' and helped the compiler
1.81 anton 14439: in scheduling: @code{NEXT} is divided into several parts:
14440: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
14441: like:
1.1 anton 14442: @example
1.81 anton 14443: NEXT_P0;
1.1 anton 14444: n=sp[0]+sp[1];
14445: sp++;
14446: NEXT_P1;
14447: sp[0]=n;
14448: NEXT_P2;
14449: @end example
14450:
1.81 anton 14451: There are various schemes that distribute the different operations of
14452: NEXT between these parts in several ways; in general, different schemes
14453: perform best on different processors. We use a scheme for most
14454: architectures that performs well for most processors of this
14455: architecture; in the furture we may switch to benchmarking and chosing
14456: the scheme on installation time.
14457:
1.1 anton 14458:
14459: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
14460: @subsection Direct or Indirect Threaded?
14461: @cindex threading, direct or indirect?
14462:
14463: @cindex -DDIRECT_THREADED
14464: Both! After packaging the nasty details in macro definitions we
14465: realized that we could switch between direct and indirect threading by
14466: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
14467: defining a few machine-specific macros for the direct-threading case.
14468: On the Forth level we also offer access words that hide the
14469: differences between the threading methods (@pxref{Threading Words}).
14470:
14471: Indirect threading is implemented completely machine-independently.
14472: Direct threading needs routines for creating jumps to the executable
1.21 crook 14473: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
14474: machine-dependent, but they do not amount to many source lines. Therefore,
14475: even porting direct threading to a new machine requires little effort.
1.1 anton 14476:
14477: @cindex --enable-indirect-threaded, configuration flag
14478: @cindex --enable-direct-threaded, configuration flag
14479: The default threading method is machine-dependent. You can enforce a
14480: specific threading method when building Gforth with the configuration
14481: flag @code{--enable-direct-threaded} or
14482: @code{--enable-indirect-threaded}. Note that direct threading is not
14483: supported on all machines.
14484:
14485: @node DOES>, , Direct or Indirect Threaded?, Threading
14486: @subsection DOES>
14487: @cindex @code{DOES>} implementation
14488:
1.26 crook 14489: @cindex @code{dodoes} routine
14490: @cindex @code{DOES>}-code
1.1 anton 14491: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
14492: the chunk of code executed by every word defined by a
14493: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
14494: the Forth code to be executed, i.e. the code after the
1.26 crook 14495: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
1.1 anton 14496:
1.21 crook 14497: In fig-Forth the code field points directly to the @code{dodoes} and the
1.45 crook 14498: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
1.29 crook 14499: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
1.1 anton 14500: the Forth-79 and all later standards, because in fig-Forth this address
14501: lies in the body (which is illegal in these standards). However, by
14502: making the code field larger for all words this solution becomes legal
14503: again. We use this approach for the indirect threaded version and for
14504: direct threading on some machines. Leaving a cell unused in most words
14505: is a bit wasteful, but on the machines we are targeting this is hardly a
14506: problem. The other reason for having a code field size of two cells is
14507: to avoid having different image files for direct and indirect threaded
14508: systems (direct threaded systems require two-cell code fields on many
14509: machines).
14510:
1.26 crook 14511: @cindex @code{DOES>}-handler
1.1 anton 14512: The other approach is that the code field points or jumps to the cell
1.26 crook 14513: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
14514: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
14515: @code{DOES>}-code address by computing the code address, i.e., the address of
1.45 crook 14516: the jump to @code{dodoes}, and add the length of that jump field. A variant of
1.1 anton 14517: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
14518: return address (which can be found in the return register on RISCs) is
1.26 crook 14519: the @code{DOES>}-code address. Since the two cells available in the code field
1.1 anton 14520: are used up by the jump to the code address in direct threading on many
14521: architectures, we use this approach for direct threading on these
14522: architectures. We did not want to add another cell to the code field.
14523:
14524: @node Primitives, Performance, Threading, Engine
14525: @section Primitives
14526: @cindex primitives, implementation
14527: @cindex virtual machine instructions, implementation
14528:
14529: @menu
14530: * Automatic Generation::
14531: * TOS Optimization::
14532: * Produced code::
14533: @end menu
14534:
14535: @node Automatic Generation, TOS Optimization, Primitives, Primitives
14536: @subsection Automatic Generation
14537: @cindex primitives, automatic generation
14538:
14539: @cindex @file{prims2x.fs}
14540: Since the primitives are implemented in a portable language, there is no
14541: longer any need to minimize the number of primitives. On the contrary,
14542: having many primitives has an advantage: speed. In order to reduce the
14543: number of errors in primitives and to make programming them easier, we
14544: provide a tool, the primitive generator (@file{prims2x.fs}), that
14545: automatically generates most (and sometimes all) of the C code for a
14546: primitive from the stack effect notation. The source for a primitive
14547: has the following form:
14548:
14549: @cindex primitive source format
14550: @format
1.58 anton 14551: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
1.29 crook 14552: [@code{""}@i{glossary entry}@code{""}]
14553: @i{C code}
1.1 anton 14554: [@code{:}
1.29 crook 14555: @i{Forth code}]
1.1 anton 14556: @end format
14557:
14558: The items in brackets are optional. The category and glossary fields
14559: are there for generating the documentation, the Forth code is there
14560: for manual implementations on machines without GNU C. E.g., the source
14561: for the primitive @code{+} is:
14562: @example
1.58 anton 14563: + ( n1 n2 -- n ) core plus
1.1 anton 14564: n = n1+n2;
14565: @end example
14566:
14567: This looks like a specification, but in fact @code{n = n1+n2} is C
14568: code. Our primitive generation tool extracts a lot of information from
14569: the stack effect notations@footnote{We use a one-stack notation, even
14570: though we have separate data and floating-point stacks; The separate
14571: notation can be generated easily from the unified notation.}: The number
14572: of items popped from and pushed on the stack, their type, and by what
14573: name they are referred to in the C code. It then generates a C code
14574: prelude and postlude for each primitive. The final C code for @code{+}
14575: looks like this:
14576:
14577: @example
1.46 pazsan 14578: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
1.1 anton 14579: /* */ /* documentation */
1.81 anton 14580: NAME("+") /* debugging output (with -DDEBUG) */
1.1 anton 14581: @{
14582: DEF_CA /* definition of variable ca (indirect threading) */
14583: Cell n1; /* definitions of variables */
14584: Cell n2;
14585: Cell n;
1.81 anton 14586: NEXT_P0; /* NEXT part 0 */
1.1 anton 14587: n1 = (Cell) sp[1]; /* input */
14588: n2 = (Cell) TOS;
14589: sp += 1; /* stack adjustment */
14590: @{
14591: n = n1+n2; /* C code taken from the source */
14592: @}
14593: NEXT_P1; /* NEXT part 1 */
14594: TOS = (Cell)n; /* output */
14595: NEXT_P2; /* NEXT part 2 */
14596: @}
14597: @end example
14598:
14599: This looks long and inefficient, but the GNU C compiler optimizes quite
14600: well and produces optimal code for @code{+} on, e.g., the R3000 and the
14601: HP RISC machines: Defining the @code{n}s does not produce any code, and
14602: using them as intermediate storage also adds no cost.
14603:
1.26 crook 14604: There are also other optimizations that are not illustrated by this
14605: example: assignments between simple variables are usually for free (copy
1.1 anton 14606: propagation). If one of the stack items is not used by the primitive
14607: (e.g. in @code{drop}), the compiler eliminates the load from the stack
14608: (dead code elimination). On the other hand, there are some things that
14609: the compiler does not do, therefore they are performed by
14610: @file{prims2x.fs}: The compiler does not optimize code away that stores
14611: a stack item to the place where it just came from (e.g., @code{over}).
14612:
14613: While programming a primitive is usually easy, there are a few cases
14614: where the programmer has to take the actions of the generator into
14615: account, most notably @code{?dup}, but also words that do not (always)
1.26 crook 14616: fall through to @code{NEXT}.
1.1 anton 14617:
14618: @node TOS Optimization, Produced code, Automatic Generation, Primitives
14619: @subsection TOS Optimization
14620: @cindex TOS optimization for primitives
14621: @cindex primitives, keeping the TOS in a register
14622:
14623: An important optimization for stack machine emulators, e.g., Forth
14624: engines, is keeping one or more of the top stack items in
1.29 crook 14625: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
14626: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1 anton 14627: @itemize @bullet
14628: @item
1.29 crook 14629: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1 anton 14630: due to fewer loads from and stores to the stack.
1.29 crook 14631: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
14632: @i{y<n}, due to additional moves between registers.
1.1 anton 14633: @end itemize
14634:
14635: @cindex -DUSE_TOS
14636: @cindex -DUSE_NO_TOS
14637: In particular, keeping one item in a register is never a disadvantage,
14638: if there are enough registers. Keeping two items in registers is a
14639: disadvantage for frequent words like @code{?branch}, constants,
14640: variables, literals and @code{i}. Therefore our generator only produces
14641: code that keeps zero or one items in registers. The generated C code
14642: covers both cases; the selection between these alternatives is made at
14643: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
14644: code for @code{+} is just a simple variable name in the one-item case,
14645: otherwise it is a macro that expands into @code{sp[0]}. Note that the
14646: GNU C compiler tries to keep simple variables like @code{TOS} in
14647: registers, and it usually succeeds, if there are enough registers.
14648:
14649: @cindex -DUSE_FTOS
14650: @cindex -DUSE_NO_FTOS
14651: The primitive generator performs the TOS optimization for the
14652: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
14653: operations the benefit of this optimization is even larger:
14654: floating-point operations take quite long on most processors, but can be
14655: performed in parallel with other operations as long as their results are
14656: not used. If the FP-TOS is kept in a register, this works. If
14657: it is kept on the stack, i.e., in memory, the store into memory has to
14658: wait for the result of the floating-point operation, lengthening the
14659: execution time of the primitive considerably.
14660:
14661: The TOS optimization makes the automatic generation of primitives a
14662: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
14663: @code{TOS} is not sufficient. There are some special cases to
14664: consider:
14665: @itemize @bullet
14666: @item In the case of @code{dup ( w -- w w )} the generator must not
14667: eliminate the store to the original location of the item on the stack,
14668: if the TOS optimization is turned on.
14669: @item Primitives with stack effects of the form @code{--}
1.29 crook 14670: @i{out1}...@i{outy} must store the TOS to the stack at the start.
14671: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1 anton 14672: must load the TOS from the stack at the end. But for the null stack
14673: effect @code{--} no stores or loads should be generated.
14674: @end itemize
14675:
14676: @node Produced code, , TOS Optimization, Primitives
14677: @subsection Produced code
14678: @cindex primitives, assembly code listing
14679:
14680: @cindex @file{engine.s}
14681: To see what assembly code is produced for the primitives on your machine
14682: with your compiler and your flag settings, type @code{make engine.s} and
1.81 anton 14683: look at the resulting file @file{engine.s}. Alternatively, you can also
14684: disassemble the code of primitives with @code{see} on some architectures.
1.1 anton 14685:
14686: @node Performance, , Primitives, Engine
14687: @section Performance
14688: @cindex performance of some Forth interpreters
14689: @cindex engine performance
14690: @cindex benchmarking Forth systems
14691: @cindex Gforth performance
14692:
14693: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
14694: impossible to write a significantly faster engine.
14695:
14696: On register-starved machines like the 386 architecture processors
14697: improvements are possible, because @code{gcc} does not utilize the
14698: registers as well as a human, even with explicit register declarations;
14699: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
14700: and hand-tuned it for the 486; this system is 1.19 times faster on the
14701: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40 anton 14702: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
14703: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
14704: registers fit in real registers (and we can even afford to use the TOS
14705: optimization), resulting in a speedup of 1.14 on the sieve over the
14706: earlier results.
1.1 anton 14707:
14708: @cindex Win32Forth performance
14709: @cindex NT Forth performance
14710: @cindex eforth performance
14711: @cindex ThisForth performance
14712: @cindex PFE performance
14713: @cindex TILE performance
1.81 anton 14714: The potential advantage of assembly language implementations is not
14715: necessarily realized in complete Forth systems: We compared Gforth-0.4.9
14716: (direct threaded, compiled with @code{gcc-2.95.1} and
14717: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
14718: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
14719: (with and without peephole (aka pinhole) optimization of the threaded
14720: code); all these systems were written in assembly language. We also
14721: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
14722: with @code{gcc-2.6.3} with the default configuration for Linux:
14723: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
14724: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
14725: employs peephole optimization of the threaded code) and TILE (compiled
14726: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
14727: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
14728: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
14729: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
14730: then extended it to run the benchmarks, added the peephole optimizer,
14731: ran the benchmarks and reported the results.
1.40 anton 14732:
1.1 anton 14733: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
14734: matrix multiplication come from the Stanford integer benchmarks and have
14735: been translated into Forth by Martin Fraeman; we used the versions
14736: included in the TILE Forth package, but with bigger data set sizes; and
14737: a recursive Fibonacci number computation for benchmarking calling
14738: performance. The following table shows the time taken for the benchmarks
14739: scaled by the time taken by Gforth (in other words, it shows the speedup
14740: factor that Gforth achieved over the other systems).
14741:
14742: @example
1.40 anton 14743: relative Win32- NT eforth This-
1.1 anton 14744: time Gforth Forth Forth eforth +opt PFE Forth TILE
1.81 anton 14745: sieve 1.00 1.60 1.32 1.60 0.98 1.82 3.67 9.91
14746: bubble 1.00 1.55 1.66 1.75 1.04 1.78 4.58
14747: matmul 1.00 1.71 1.57 1.69 0.86 1.83 4.74
14748: fib 1.00 1.76 1.54 1.41 1.00 2.01 3.45 4.96
1.1 anton 14749: @end example
14750:
1.26 crook 14751: You may be quite surprised by the good performance of Gforth when
14752: compared with systems written in assembly language. One important reason
14753: for the disappointing performance of these other systems is probably
14754: that they are not written optimally for the 486 (e.g., they use the
14755: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
14756: but costly method for relocating the Forth image: like @code{cforth}, it
14757: computes the actual addresses at run time, resulting in two address
14758: computations per @code{NEXT} (@pxref{Image File Background}).
14759:
1.40 anton 14760: Only Eforth with the peephole optimizer performs comparable to
14761: Gforth. The speedups achieved with peephole optimization of threaded
14762: code are quite remarkable. Adding a peephole optimizer to Gforth should
14763: cause similar speedups.
1.1 anton 14764:
14765: The speedup of Gforth over PFE, ThisForth and TILE can be easily
14766: explained with the self-imposed restriction of the latter systems to
14767: standard C, which makes efficient threading impossible (however, the
1.4 anton 14768: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1 anton 14769: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
14770: Moreover, current C compilers have a hard time optimizing other aspects
14771: of the ThisForth and the TILE source.
14772:
1.26 crook 14773: The performance of Gforth on 386 architecture processors varies widely
14774: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
14775: allocate any of the virtual machine registers into real machine
14776: registers by itself and would not work correctly with explicit register
1.40 anton 14777: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
1.26 crook 14778: the Sieve) than the one measured above.
1.1 anton 14779:
1.26 crook 14780: Note that there have been several releases of Win32Forth since the
14781: release presented here, so the results presented above may have little
1.40 anton 14782: predictive value for the performance of Win32Forth today (results for
14783: the current release on an i486DX2/66 are welcome).
1.1 anton 14784:
14785: @cindex @file{Benchres}
1.66 anton 14786: In
14787: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
14788: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1 anton 14789: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66 anton 14790: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
14791: several native code systems; that version of Gforth is slower on a 486
14792: than the direct threaded version used here. You can find a newer version
14793: of these measurements at
1.47 crook 14794: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1 anton 14795: find numbers for Gforth on various machines in @file{Benchres}.
14796:
1.26 crook 14797: @c ******************************************************************
1.13 pazsan 14798: @node Binding to System Library, Cross Compiler, Engine, Top
1.14 pazsan 14799: @chapter Binding to System Library
1.13 pazsan 14800:
14801: @node Cross Compiler, Bugs, Binding to System Library, Top
1.14 pazsan 14802: @chapter Cross Compiler
1.47 crook 14803: @cindex @file{cross.fs}
14804: @cindex cross-compiler
14805: @cindex metacompiler
14806: @cindex target compiler
1.13 pazsan 14807:
1.46 pazsan 14808: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
14809: mostly written in Forth, including crucial parts like the outer
14810: interpreter and compiler, it needs compiled Forth code to get
14811: started. The cross compiler allows to create new images for other
14812: architectures, even running under another Forth system.
1.13 pazsan 14813:
14814: @menu
1.67 anton 14815: * Using the Cross Compiler::
14816: * How the Cross Compiler Works::
1.13 pazsan 14817: @end menu
14818:
1.21 crook 14819: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14 pazsan 14820: @section Using the Cross Compiler
1.46 pazsan 14821:
14822: The cross compiler uses a language that resembles Forth, but isn't. The
14823: main difference is that you can execute Forth code after definition,
14824: while you usually can't execute the code compiled by cross, because the
14825: code you are compiling is typically for a different computer than the
14826: one you are compiling on.
14827:
1.81 anton 14828: @c anton: This chapter is somewhat different from waht I would expect: I
14829: @c would expect an explanation of the cross language and how to create an
14830: @c application image with it. The section explains some aspects of
14831: @c creating a Gforth kernel.
14832:
1.46 pazsan 14833: The Makefile is already set up to allow you to create kernels for new
14834: architectures with a simple make command. The generic kernels using the
14835: GCC compiled virtual machine are created in the normal build process
14836: with @code{make}. To create a embedded Gforth executable for e.g. the
14837: 8086 processor (running on a DOS machine), type
14838:
14839: @example
14840: make kernl-8086.fi
14841: @end example
14842:
14843: This will use the machine description from the @file{arch/8086}
14844: directory to create a new kernel. A machine file may look like that:
14845:
14846: @example
14847: \ Parameter for target systems 06oct92py
14848:
14849: 4 Constant cell \ cell size in bytes
14850: 2 Constant cell<< \ cell shift to bytes
14851: 5 Constant cell>bit \ cell shift to bits
14852: 8 Constant bits/char \ bits per character
14853: 8 Constant bits/byte \ bits per byte [default: 8]
14854: 8 Constant float \ bytes per float
14855: 8 Constant /maxalign \ maximum alignment in bytes
14856: false Constant bigendian \ byte order
14857: ( true=big, false=little )
14858:
14859: include machpc.fs \ feature list
14860: @end example
14861:
14862: This part is obligatory for the cross compiler itself, the feature list
14863: is used by the kernel to conditionally compile some features in and out,
14864: depending on whether the target supports these features.
14865:
14866: There are some optional features, if you define your own primitives,
14867: have an assembler, or need special, nonstandard preparation to make the
1.81 anton 14868: boot process work. @code{asm-include} includes an assembler,
1.46 pazsan 14869: @code{prims-include} includes primitives, and @code{>boot} prepares for
14870: booting.
14871:
14872: @example
14873: : asm-include ." Include assembler" cr
14874: s" arch/8086/asm.fs" included ;
14875:
14876: : prims-include ." Include primitives" cr
14877: s" arch/8086/prim.fs" included ;
14878:
14879: : >boot ." Prepare booting" cr
14880: s" ' boot >body into-forth 1+ !" evaluate ;
14881: @end example
14882:
14883: These words are used as sort of macro during the cross compilation in
1.81 anton 14884: the file @file{kernel/main.fs}. Instead of using these macros, it would
1.46 pazsan 14885: be possible --- but more complicated --- to write a new kernel project
14886: file, too.
14887:
14888: @file{kernel/main.fs} expects the machine description file name on the
14889: stack; the cross compiler itself (@file{cross.fs}) assumes that either
14890: @code{mach-file} leaves a counted string on the stack, or
14891: @code{machine-file} leaves an address, count pair of the filename on the
14892: stack.
14893:
14894: The feature list is typically controlled using @code{SetValue}, generic
14895: files that are used by several projects can use @code{DefaultValue}
14896: instead. Both functions work like @code{Value}, when the value isn't
14897: defined, but @code{SetValue} works like @code{to} if the value is
14898: defined, and @code{DefaultValue} doesn't set anything, if the value is
14899: defined.
14900:
14901: @example
14902: \ generic mach file for pc gforth 03sep97jaw
14903:
14904: true DefaultValue NIL \ relocating
14905:
14906: >ENVIRON
14907:
14908: true DefaultValue file \ controls the presence of the
14909: \ file access wordset
14910: true DefaultValue OS \ flag to indicate a operating system
14911:
14912: true DefaultValue prims \ true: primitives are c-code
14913:
14914: true DefaultValue floating \ floating point wordset is present
14915:
14916: true DefaultValue glocals \ gforth locals are present
14917: \ will be loaded
14918: true DefaultValue dcomps \ double number comparisons
14919:
14920: true DefaultValue hash \ hashing primitives are loaded/present
14921:
14922: true DefaultValue xconds \ used together with glocals,
14923: \ special conditionals supporting gforths'
14924: \ local variables
14925: true DefaultValue header \ save a header information
14926:
14927: true DefaultValue backtrace \ enables backtrace code
14928:
14929: false DefaultValue ec
14930: false DefaultValue crlf
14931:
14932: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
14933:
14934: &16 KB DefaultValue stack-size
14935: &15 KB &512 + DefaultValue fstack-size
14936: &15 KB DefaultValue rstack-size
14937: &14 KB &512 + DefaultValue lstack-size
14938: @end example
1.13 pazsan 14939:
1.48 anton 14940: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
1.14 pazsan 14941: @section How the Cross Compiler Works
1.13 pazsan 14942:
14943: @node Bugs, Origin, Cross Compiler, Top
1.21 crook 14944: @appendix Bugs
1.1 anton 14945: @cindex bug reporting
14946:
1.21 crook 14947: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1 anton 14948:
14949: If you find a bug, please send a bug report to
1.33 anton 14950: @email{bug-gforth@@gnu.org}. A bug report should include this
1.21 crook 14951: information:
14952:
14953: @itemize @bullet
14954: @item
1.81 anton 14955: A program (or a sequence of keyboard commands) that reproduces the bug.
14956: @item
14957: A description of what you think constitutes the buggy behaviour.
14958: @item
1.21 crook 14959: The Gforth version used (it is announced at the start of an
14960: interactive Gforth session).
14961: @item
14962: The machine and operating system (on Unix
14963: systems @code{uname -a} will report this information).
14964: @item
1.81 anton 14965: The installation options (you can find the configure options at the
14966: start of @file{config.status}) and configuration (@code{configure}
14967: output or @file{config.cache}).
1.21 crook 14968: @item
14969: A complete list of changes (if any) you (or your installer) have made to the
14970: Gforth sources.
14971: @end itemize
1.1 anton 14972:
14973: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14974: to Report Bugs, gcc.info, GNU C Manual}.
14975:
14976:
1.21 crook 14977: @node Origin, Forth-related information, Bugs, Top
14978: @appendix Authors and Ancestors of Gforth
1.1 anton 14979:
14980: @section Authors and Contributors
14981: @cindex authors of Gforth
14982: @cindex contributors to Gforth
14983:
14984: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
1.81 anton 14985: Ertl. The third major author was Jens Wilke. Neal Crook contributed a
14986: lot to the manual. Assemblers and disassemblers were contributed by
14987: Andrew McKewan, Christian Pirker, and Bernd Thallner. Lennart Benschop
14988: (who was one of Gforth's first users, in mid-1993) and Stuart Ramsden
14989: inspired us with their continuous feedback. Lennart Benshop contributed
1.1 anton 14990: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14991: support for calling C libraries. Helpful comments also came from Paul
14992: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.58 anton 14993: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14994: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14995: helpful comments from many others; thank you all, sorry for not listing
14996: you here (but digging through my mailbox to extract your names is on my
1.81 anton 14997: to-do list).
1.1 anton 14998:
14999: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
15000: and autoconf, among others), and to the creators of the Internet: Gforth
1.21 crook 15001: was developed across the Internet, and its authors did not meet
1.20 pazsan 15002: physically for the first 4 years of development.
1.1 anton 15003:
15004: @section Pedigree
1.26 crook 15005: @cindex pedigree of Gforth
1.1 anton 15006:
1.81 anton 15007: Gforth descends from bigFORTH (1993) and fig-Forth. Of course, a
15008: significant part of the design of Gforth was prescribed by ANS Forth.
1.1 anton 15009:
1.20 pazsan 15010: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1 anton 15011: 32 bit native code version of VolksForth for the Atari ST, written
15012: mostly by Dietrich Weineck.
15013:
1.81 anton 15014: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
15015: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
15016: the mid-80s and ported to the Atari ST in 1986. It descends from F83.
1.1 anton 15017:
15018: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
15019: Forth-83 standard. !! Pedigree? When?
15020:
15021: A team led by Bill Ragsdale implemented fig-Forth on many processors in
15022: 1979. Robert Selzer and Bill Ragsdale developed the original
15023: implementation of fig-Forth for the 6502 based on microForth.
15024:
15025: The principal architect of microForth was Dean Sanderson. microForth was
15026: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
15027: the 1802, and subsequently implemented on the 8080, the 6800 and the
15028: Z80.
15029:
15030: All earlier Forth systems were custom-made, usually by Charles Moore,
15031: who discovered (as he puts it) Forth during the late 60s. The first full
15032: Forth existed in 1971.
15033:
1.81 anton 15034: A part of the information in this section comes from
15035: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
15036: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
15037: Charles H. Moore, presented at the HOPL-II conference and preprinted in
15038: SIGPLAN Notices 28(3), 1993. You can find more historical and
15039: genealogical information about Forth there.
1.1 anton 15040:
1.81 anton 15041: @c ------------------------------------------------------------------
1.21 crook 15042: @node Forth-related information, Word Index, Origin, Top
15043: @appendix Other Forth-related information
15044: @cindex Forth-related information
15045:
1.81 anton 15046: @c anton: I threw most of this stuff out, because it can be found through
15047: @c the FAQ and the FAQ is more likely to be up-to-date.
1.21 crook 15048:
15049: @cindex comp.lang.forth
15050: @cindex frequently asked questions
1.81 anton 15051: There is an active news group (comp.lang.forth) discussing Forth
15052: (including Gforth) and Forth-related issues. Its
15053: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
15054: (frequently asked questions and their answers) contains a lot of
15055: information on Forth. You should read it before posting to
15056: comp.lang.forth.
1.21 crook 15057:
1.81 anton 15058: The ANS Forth standard is most usable in its
15059: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
1.21 crook 15060:
1.81 anton 15061: @c ------------------------------------------------------------------
15062: @node Word Index, Concept Index, Forth-related information, Top
1.1 anton 15063: @unnumbered Word Index
15064:
1.26 crook 15065: This index is a list of Forth words that have ``glossary'' entries
15066: within this manual. Each word is listed with its stack effect and
15067: wordset.
1.1 anton 15068:
15069: @printindex fn
15070:
1.81 anton 15071: @c anton: the name index seems superfluous given the word and concept indices.
15072:
15073: @c @node Name Index, Concept Index, Word Index, Top
15074: @c @unnumbered Name Index
1.41 anton 15075:
1.81 anton 15076: @c This index is a list of Forth words that have ``glossary'' entries
15077: @c within this manual.
1.41 anton 15078:
1.81 anton 15079: @c @printindex ky
1.41 anton 15080:
1.81 anton 15081: @node Concept Index, , Word Index, Top
1.1 anton 15082: @unnumbered Concept and Word Index
15083:
1.26 crook 15084: Not all entries listed in this index are present verbatim in the
15085: text. This index also duplicates, in abbreviated form, all of the words
15086: listed in the Word Index (only the names are listed for the words here).
1.1 anton 15087:
15088: @printindex cp
15089:
15090: @contents
15091: @bye
1.81 anton 15092:
15093:
1.1 anton 15094:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>