[gforth] / gforth / doc / gforth.ds  

gforth: gforth/doc/gforth.ds


1 : anton 1.1 \input texinfo @c -*-texinfo-*-
2 :     @comment The source is gforth.ds, from which gforth.texi is generated
3 : crook 1.28
4 : crook 1.21 @comment TODO: nac29jan99 - a list of things to add in the next edit:
5 : crook 1.28 @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.
11 : anton 1.36 @c Not an improvement IMO - anton
12 :     @c and anyway, this should be taken up
13 :     @c with Karl Berry (the texinfo guy) - anton
14 : crook 1.29 @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 :    
17 : crook 1.28
18 : anton 1.1 @comment %**start of header (This is for running Texinfo on a region.)
19 :     @setfilename gforth.info
20 :     @settitle Gforth Manual
21 :     @dircategory GNU programming tools
22 :     @direntry
23 :     * Gforth: (gforth). A fast interpreter for the Forth language.
24 :     @end direntry
25 : anton 1.49 @c The Texinfo manual also recommends doing this, but for Gforth it may
26 :     @c not make much sense
27 :     @c @dircategory Individual utilities
28 :     @c @direntry
29 :     @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
30 :     @c @end direntry
31 :    
32 : anton 1.1 @comment @setchapternewpage odd
33 : crook 1.29 @comment TODO this gets left in by HTML converter
34 : anton 1.12 @macro progstyle {}
35 :     Programming style note:
36 : anton 1.3 @end macro
37 : anton 1.48
38 :     @macro assignment {}
39 :     @table @i
40 :     @item Assignment:
41 :     @end macro
42 :     @macro endassignment {}
43 :     @end table
44 :     @end macro
45 :    
46 : anton 1.1 @comment %**end of header (This is for running Texinfo on a region.)
47 :    
48 : crook 1.29
49 :     @comment ----------------------------------------------------------
50 :     @comment macros for beautifying glossary entries
51 :     @comment if these are used, need to strip them out for HTML converter
52 :     @comment else they get repeated verbatim in HTML output.
53 :     @comment .. not working yet.
54 :    
55 :     @macro GLOSS-START {}
56 :     @iftex
57 :     @ninerm
58 :     @end iftex
59 :     @end macro
60 :    
61 :     @macro GLOSS-END {}
62 :     @iftex
63 :     @rm
64 :     @end iftex
65 :     @end macro
66 :    
67 :     @comment ----------------------------------------------------------
68 :    
69 :    
70 : anton 1.10 @include version.texi
71 :    
72 : anton 1.49 @ifnottex
73 : anton 1.11 This file documents Gforth @value{VERSION}
74 : anton 1.1
75 : crook 1.62 Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
76 : anton 1.1
77 :     Permission is granted to make and distribute verbatim copies of
78 :     this manual provided the copyright notice and this permission notice
79 :     are preserved on all copies.
80 :    
81 :     @ignore
82 :     Permission is granted to process this file through TeX and print the
83 :     results, provided the printed document carries a copying permission
84 :     notice identical to this one except for the removal of this paragraph
85 :     (this paragraph not being relevant to the printed manual).
86 :    
87 :     @end ignore
88 :     Permission is granted to copy and distribute modified versions of this
89 :     manual under the conditions for verbatim copying, provided also that the
90 :     sections entitled "Distribution" and "General Public License" are
91 :     included exactly as in the original, and provided that the entire
92 :     resulting derived work is distributed under the terms of a permission
93 :     notice identical to this one.
94 :    
95 :     Permission is granted to copy and distribute translations of this manual
96 :     into another language, under the above conditions for modified versions,
97 :     except that the sections entitled "Distribution" and "General Public
98 :     License" may be included in a translation approved by the author instead
99 :     of in the original English.
100 : anton 1.49 @end ifnottex
101 : anton 1.1
102 :     @finalout
103 :     @titlepage
104 :     @sp 10
105 :     @center @titlefont{Gforth Manual}
106 :     @sp 2
107 : anton 1.11 @center for version @value{VERSION}
108 : anton 1.1 @sp 2
109 : anton 1.34 @center Neal Crook
110 : anton 1.1 @center Anton Ertl
111 : pazsan 1.6 @center Bernd Paysan
112 : anton 1.5 @center Jens Wilke
113 : anton 1.1 @sp 3
114 : crook 1.47 @center This manual is permanently under construction and was last updated on 15-Mar-2000
115 : anton 1.1
116 :     @comment The following two commands start the copyright page.
117 :     @page
118 :     @vskip 0pt plus 1filll
119 : crook 1.62 Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
120 : anton 1.1
121 :     @comment !! Published by ... or You can get a copy of this manual ...
122 :    
123 :     Permission is granted to make and distribute verbatim copies of
124 :     this manual provided the copyright notice and this permission notice
125 :     are preserved on all copies.
126 :    
127 :     Permission is granted to copy and distribute modified versions of this
128 :     manual under the conditions for verbatim copying, provided also that the
129 :     sections entitled "Distribution" and "General Public License" are
130 :     included exactly as in the original, and provided that the entire
131 :     resulting derived work is distributed under the terms of a permission
132 :     notice identical to this one.
133 :    
134 :     Permission is granted to copy and distribute translations of this manual
135 :     into another language, under the above conditions for modified versions,
136 :     except that the sections entitled "Distribution" and "General Public
137 :     License" may be included in a translation approved by the author instead
138 :     of in the original English.
139 :     @end titlepage
140 :    
141 :     @node Top, License, (dir), (dir)
142 : anton 1.49 @ifnottex
143 : anton 1.1 Gforth is a free implementation of ANS Forth available on many
144 : anton 1.11 personal machines. This manual corresponds to version @value{VERSION}.
145 : anton 1.49 @end ifnottex
146 : anton 1.1
147 :     @menu
148 : crook 1.21 * License:: The GPL
149 : crook 1.26 * Goals:: About the Gforth Project
150 : crook 1.29 * Gforth Environment:: Starting (and exiting) Gforth
151 : anton 1.48 * Tutorial:: Hands-on Forth Tutorial
152 : crook 1.21 * Introduction:: An introduction to ANS Forth
153 : anton 1.1 * Words:: Forth words available in Gforth
154 : anton 1.24 * Error messages:: How to interpret them
155 : anton 1.1 * Tools:: Programming tools
156 :     * ANS conformance:: Implementation-defined options etc.
157 : anton 1.65 * Standard vs Extensions:: Should I use extensions?
158 : anton 1.1 * Model:: The abstract machine of Gforth
159 :     * Integrating Gforth:: Forth as scripting language for applications
160 :     * Emacs and Gforth:: The Gforth Mode
161 :     * Image Files:: @code{.fi} files contain compiled code
162 :     * Engine:: The inner interpreter and the primitives
163 : anton 1.24 * Binding to System Library::
164 : pazsan 1.13 * Cross Compiler:: The Cross Compiler
165 : anton 1.1 * Bugs:: How to report them
166 :     * Origin:: Authors and ancestors of Gforth
167 : crook 1.21 * Forth-related information:: Books and places to look on the WWW
168 : anton 1.1 * Word Index:: An item for each Forth word
169 : anton 1.41 * Name Index:: Forth words, only names listed
170 : anton 1.1 * Concept Index:: A menu covering many topics
171 : anton 1.12
172 : anton 1.48 @detailmenu --- The Detailed Node Listing ---
173 : anton 1.12
174 : crook 1.29 Gforth Environment
175 :    
176 : anton 1.32 * Invoking Gforth:: Getting in
177 :     * Leaving Gforth:: Getting out
178 :     * Command-line editing::
179 : anton 1.48 * Environment variables:: that affect how Gforth starts up
180 : anton 1.32 * Gforth Files:: What gets installed and where
181 : anton 1.48 * Startup speed:: When 35ms is not fast enough ...
182 :    
183 :     Forth Tutorial
184 :    
185 :     * Starting Gforth Tutorial::
186 :     * Syntax Tutorial::
187 :     * Crash Course Tutorial::
188 :     * Stack Tutorial::
189 :     * Arithmetics Tutorial::
190 :     * Stack Manipulation Tutorial::
191 :     * Using files for Forth code Tutorial::
192 :     * Comments Tutorial::
193 :     * Colon Definitions Tutorial::
194 :     * Decompilation Tutorial::
195 :     * Stack-Effect Comments Tutorial::
196 :     * Types Tutorial::
197 :     * Factoring Tutorial::
198 :     * Designing the stack effect Tutorial::
199 :     * Local Variables Tutorial::
200 :     * Conditional execution Tutorial::
201 :     * Flags and Comparisons Tutorial::
202 :     * General Loops Tutorial::
203 :     * Counted loops Tutorial::
204 :     * Recursion Tutorial::
205 :     * Leaving definitions or loops Tutorial::
206 :     * Return Stack Tutorial::
207 :     * Memory Tutorial::
208 :     * Characters and Strings Tutorial::
209 :     * Alignment Tutorial::
210 :     * Interpretation and Compilation Semantics and Immediacy Tutorial::
211 :     * Execution Tokens Tutorial::
212 :     * Exceptions Tutorial::
213 :     * Defining Words Tutorial::
214 :     * Arrays and Records Tutorial::
215 :     * POSTPONE Tutorial::
216 :     * Literal Tutorial::
217 :     * Advanced macros Tutorial::
218 :     * Compilation Tokens Tutorial::
219 :     * Wordlists and Search Order Tutorial::
220 : crook 1.29
221 : anton 1.24 An Introduction to ANS Forth
222 :    
223 :     * Introducing the Text Interpreter::
224 :     * Stacks and Postfix notation::
225 :     * Your first definition::
226 :     * How does that work?::
227 :     * Forth is written in Forth::
228 :     * Review - elements of a Forth system::
229 : crook 1.29 * Where to go next::
230 : anton 1.24 * Exercises::
231 :    
232 : anton 1.12 Forth Words
233 :    
234 :     * Notation::
235 : anton 1.65 * Case insensitivity::
236 :     * Comments::
237 :     * Boolean Flags::
238 : anton 1.12 * Arithmetic::
239 :     * Stack Manipulation::
240 :     * Memory::
241 :     * Control Structures::
242 :     * Defining Words::
243 : anton 1.65 * Interpretation and Compilation Semantics::
244 : crook 1.47 * Tokens for Words::
245 : anton 1.65 * The Text Interpreter::
246 :     * Word Lists::
247 :     * Environmental Queries::
248 : anton 1.12 * Files::
249 :     * Blocks::
250 :     * Other I/O::
251 :     * Programming Tools::
252 :     * Assembler and Code Words::
253 :     * Threading Words::
254 : crook 1.26 * Locals::
255 :     * Structures::
256 :     * Object-oriented Forth::
257 : anton 1.65 * Passing Commands to the OS::
258 :     * Keeping track of Time::
259 :     * Miscellaneous Words::
260 : anton 1.12
261 :     Arithmetic
262 :    
263 :     * Single precision::
264 :     * Bitwise operations::
265 : crook 1.21 * Double precision:: Double-cell integer arithmetic
266 :     * Numeric comparison::
267 : anton 1.32 * Mixed precision:: Operations with single and double-cell integers
268 : anton 1.12 * Floating Point::
269 :    
270 :     Stack Manipulation
271 :    
272 :     * Data stack::
273 :     * Floating point stack::
274 :     * Return stack::
275 :     * Locals stack::
276 :     * Stack pointer manipulation::
277 :    
278 :     Memory
279 :    
280 : anton 1.32 * Memory model::
281 :     * Dictionary allocation::
282 :     * Heap Allocation::
283 :     * Memory Access::
284 :     * Address arithmetic::
285 :     * Memory Blocks::
286 : anton 1.12
287 :     Control Structures
288 :    
289 : anton 1.41 * Selection:: IF ... ELSE ... ENDIF
290 :     * Simple Loops:: BEGIN ...
291 : anton 1.32 * Counted Loops:: DO
292 :     * Arbitrary control structures::
293 :     * Calls and returns::
294 : anton 1.12 * Exception Handling::
295 :    
296 :     Defining Words
297 :    
298 : crook 1.45 * CREATE::
299 : crook 1.44 * Variables:: Variables and user variables
300 :     * Constants::
301 :     * Values:: Initialised variables
302 : anton 1.32 * Colon Definitions::
303 : crook 1.44 * Anonymous Definitions:: Definitions without names
304 : anton 1.32 * User-defined Defining Words::
305 : crook 1.44 * Deferred words:: Allow forward references
306 :     * Aliases::
307 : anton 1.32 * Supplying names::
308 : crook 1.47
309 : anton 1.63 User-defined Defining Words
310 :    
311 :     * CREATE..DOES> applications::
312 :     * CREATE..DOES> details::
313 :     * Advanced does> usage example::
314 :    
315 : crook 1.47 Interpretation and Compilation Semantics
316 :    
317 : crook 1.44 * Combined words::
318 : anton 1.12
319 : crook 1.21 The Text Interpreter
320 :    
321 : crook 1.29 * Input Sources::
322 : crook 1.21 * Number Conversion::
323 :     * Interpret/Compile states::
324 :     * Literals::
325 :     * Interpreter Directives::
326 :    
327 : crook 1.26 Word Lists
328 :    
329 :     * Why use word lists?::
330 :     * Word list examples::
331 :    
332 :     Files
333 :    
334 : anton 1.48 * Forth source files::
335 :     * General files::
336 :     * Search Paths::
337 :    
338 :     Search Paths
339 :    
340 :     * Forth Search Paths::
341 : crook 1.26 * General Search Paths::
342 :    
343 :     Other I/O
344 :    
345 : anton 1.32 * Simple numeric output:: Predefined formats
346 :     * Formatted numeric output:: Formatted (pictured) output
347 :     * String Formats:: How Forth stores strings in memory
348 :     * Displaying characters and strings:: Other stuff
349 :     * Input:: Input
350 : crook 1.26
351 :     Programming Tools
352 :    
353 :     * Debugging:: Simple and quick.
354 :     * Assertions:: Making your programs self-checking.
355 : pazsan 1.46 * Singlestep Debugger:: Executing your program word by word.
356 : crook 1.26
357 : anton 1.63 Assembler and Code Words
358 :    
359 :     * Code and ;code::
360 :     * Common Assembler:: Assembler Syntax
361 :     * Common Disassembler::
362 :     * 386 Assembler:: Deviations and special cases
363 :     * Alpha Assembler:: Deviations and special cases
364 :     * MIPS assembler:: Deviations and special cases
365 :     * Other assemblers:: How to write them
366 :    
367 : crook 1.26 Locals
368 :    
369 :     * Gforth locals::
370 :     * ANS Forth locals::
371 :    
372 :     Gforth locals
373 :    
374 :     * Where are locals visible by name?::
375 :     * How long do locals live?::
376 :     * Programming Style::
377 :     * Implementation::
378 :    
379 : anton 1.12 Structures
380 :    
381 :     * Why explicit structure support?::
382 :     * Structure Usage::
383 :     * Structure Naming Convention::
384 :     * Structure Implementation::
385 :     * Structure Glossary::
386 :    
387 :     Object-oriented Forth
388 :    
389 : anton 1.48 * Why object-oriented programming?::
390 :     * Object-Oriented Terminology::
391 :     * Objects::
392 :     * OOF::
393 :     * Mini-OOF::
394 : crook 1.23 * Comparison with other object models::
395 : anton 1.12
396 : anton 1.24 The @file{objects.fs} model
397 : anton 1.12
398 :     * Properties of the Objects model::
399 :     * Basic Objects Usage::
400 : anton 1.41 * The Objects base class::
401 : anton 1.12 * Creating objects::
402 :     * Object-Oriented Programming Style::
403 :     * Class Binding::
404 :     * Method conveniences::
405 :     * Classes and Scoping::
406 : anton 1.41 * Dividing classes::
407 : anton 1.12 * Object Interfaces::
408 :     * Objects Implementation::
409 :     * Objects Glossary::
410 :    
411 : anton 1.24 The @file{oof.fs} model
412 : anton 1.12
413 :     * Properties of the OOF model::
414 :     * Basic OOF Usage::
415 : crook 1.23 * The OOF base class::
416 : anton 1.12 * Class Declaration::
417 :     * Class Implementation::
418 :    
419 : anton 1.24 The @file{mini-oof.fs} model
420 : crook 1.23
421 : anton 1.48 * Basic Mini-OOF Usage::
422 :     * Mini-OOF Example::
423 :     * Mini-OOF Implementation::
424 :     * Comparison with other object models::
425 : crook 1.23
426 : anton 1.12 Tools
427 :    
428 :     * ANS Report:: Report the words used, sorted by wordset.
429 :    
430 :     ANS conformance
431 :    
432 :     * The Core Words::
433 :     * The optional Block word set::
434 :     * The optional Double Number word set::
435 :     * The optional Exception word set::
436 :     * The optional Facility word set::
437 :     * The optional File-Access word set::
438 :     * The optional Floating-Point word set::
439 :     * The optional Locals word set::
440 :     * The optional Memory-Allocation word set::
441 :     * The optional Programming-Tools word set::
442 :     * The optional Search-Order word set::
443 :    
444 :     The Core Words
445 :    
446 :     * core-idef:: Implementation Defined Options
447 :     * core-ambcond:: Ambiguous Conditions
448 :     * core-other:: Other System Documentation
449 :    
450 :     The optional Block word set
451 :    
452 :     * block-idef:: Implementation Defined Options
453 :     * block-ambcond:: Ambiguous Conditions
454 :     * block-other:: Other System Documentation
455 :    
456 :     The optional Double Number word set
457 :    
458 :     * double-ambcond:: Ambiguous Conditions
459 :    
460 :     The optional Exception word set
461 :    
462 :     * exception-idef:: Implementation Defined Options
463 :    
464 :     The optional Facility word set
465 :    
466 :     * facility-idef:: Implementation Defined Options
467 :     * facility-ambcond:: Ambiguous Conditions
468 :    
469 :     The optional File-Access word set
470 :    
471 :     * file-idef:: Implementation Defined Options
472 :     * file-ambcond:: Ambiguous Conditions
473 :    
474 :     The optional Floating-Point word set
475 :    
476 :     * floating-idef:: Implementation Defined Options
477 :     * floating-ambcond:: Ambiguous Conditions
478 :    
479 :     The optional Locals word set
480 :    
481 :     * locals-idef:: Implementation Defined Options
482 :     * locals-ambcond:: Ambiguous Conditions
483 :    
484 :     The optional Memory-Allocation word set
485 :    
486 :     * memory-idef:: Implementation Defined Options
487 :    
488 :     The optional Programming-Tools word set
489 :    
490 :     * programming-idef:: Implementation Defined Options
491 :     * programming-ambcond:: Ambiguous Conditions
492 :    
493 :     The optional Search-Order word set
494 :    
495 :     * search-idef:: Implementation Defined Options
496 :     * search-ambcond:: Ambiguous Conditions
497 :    
498 :     Image Files
499 :    
500 : anton 1.24 * Image Licensing Issues:: Distribution terms for images.
501 :     * Image File Background:: Why have image files?
502 : anton 1.32 * Non-Relocatable Image Files:: don't always work.
503 : anton 1.24 * Data-Relocatable Image Files:: are better.
504 : anton 1.32 * Fully Relocatable Image Files:: better yet.
505 : anton 1.24 * Stack and Dictionary Sizes:: Setting the default sizes for an image.
506 : anton 1.32 * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
507 : anton 1.24 * Modifying the Startup Sequence:: and turnkey applications.
508 : anton 1.12
509 :     Fully Relocatable Image Files
510 :    
511 : crook 1.27 * gforthmi:: The normal way
512 : anton 1.12 * cross.fs:: The hard way
513 :    
514 :     Engine
515 :    
516 :     * Portability::
517 :     * Threading::
518 :     * Primitives::
519 :     * Performance::
520 :    
521 :     Threading
522 :    
523 :     * Scheduling::
524 :     * Direct or Indirect Threaded?::
525 :     * DOES>::
526 :    
527 :     Primitives
528 :    
529 :     * Automatic Generation::
530 :     * TOS Optimization::
531 :     * Produced code::
532 : pazsan 1.13
533 :     Cross Compiler
534 :    
535 :     * Using the Cross Compiler::
536 :     * How the Cross Compiler Works::
537 :    
538 : anton 1.24 Other Forth-related information
539 : crook 1.21
540 :     * Internet resources::
541 :     * Books::
542 :     * The Forth Interest Group::
543 :     * Conferences::
544 :    
545 : anton 1.24 @end detailmenu
546 : anton 1.1 @end menu
547 :    
548 : crook 1.26 @node License, Goals, Top, Top
549 : anton 1.1 @unnumbered GNU GENERAL PUBLIC LICENSE
550 :     @center Version 2, June 1991
551 :    
552 :     @display
553 :     Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
554 :     675 Mass Ave, Cambridge, MA 02139, USA
555 :    
556 :     Everyone is permitted to copy and distribute verbatim copies
557 :     of this license document, but changing it is not allowed.
558 :     @end display
559 :    
560 :     @unnumberedsec Preamble
561 :    
562 :     The licenses for most software are designed to take away your
563 :     freedom to share and change it. By contrast, the GNU General Public
564 :     License is intended to guarantee your freedom to share and change free
565 :     software---to make sure the software is free for all its users. This
566 :     General Public License applies to most of the Free Software
567 :     Foundation's software and to any other program whose authors commit to
568 :     using it. (Some other Free Software Foundation software is covered by
569 :     the GNU Library General Public License instead.) You can apply it to
570 :     your programs, too.
571 :    
572 :     When we speak of free software, we are referring to freedom, not
573 :     price. Our General Public Licenses are designed to make sure that you
574 :     have the freedom to distribute copies of free software (and charge for
575 :     this service if you wish), that you receive source code or can get it
576 :     if you want it, that you can change the software or use pieces of it
577 :     in new free programs; and that you know you can do these things.
578 :    
579 :     To protect your rights, we need to make restrictions that forbid
580 :     anyone to deny you these rights or to ask you to surrender the rights.
581 :     These restrictions translate to certain responsibilities for you if you
582 :     distribute copies of the software, or if you modify it.
583 :    
584 :     For example, if you distribute copies of such a program, whether
585 :     gratis or for a fee, you must give the recipients all the rights that
586 :     you have. You must make sure that they, too, receive or can get the
587 :     source code. And you must show them these terms so they know their
588 :     rights.
589 :    
590 :     We protect your rights with two steps: (1) copyright the software, and
591 :     (2) offer you this license which gives you legal permission to copy,
592 :     distribute and/or modify the software.
593 :    
594 :     Also, for each author's protection and ours, we want to make certain
595 :     that everyone understands that there is no warranty for this free
596 :     software. If the software is modified by someone else and passed on, we
597 :     want its recipients to know that what they have is not the original, so
598 :     that any problems introduced by others will not reflect on the original
599 :     authors' reputations.
600 :    
601 :     Finally, any free program is threatened constantly by software
602 :     patents. We wish to avoid the danger that redistributors of a free
603 :     program will individually obtain patent licenses, in effect making the
604 :     program proprietary. To prevent this, we have made it clear that any
605 :     patent must be licensed for everyone's free use or not licensed at all.
606 :    
607 :     The precise terms and conditions for copying, distribution and
608 :     modification follow.
609 :    
610 :     @iftex
611 :     @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
612 :     @end iftex
613 : anton 1.49 @ifnottex
614 : anton 1.1 @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
615 : anton 1.49 @end ifnottex
616 : anton 1.1
617 :     @enumerate 0
618 :     @item
619 :     This License applies to any program or other work which contains
620 :     a notice placed by the copyright holder saying it may be distributed
621 :     under the terms of this General Public License. The ``Program'', below,
622 :     refers to any such program or work, and a ``work based on the Program''
623 :     means either the Program or any derivative work under copyright law:
624 :     that is to say, a work containing the Program or a portion of it,
625 :     either verbatim or with modifications and/or translated into another
626 :     language. (Hereinafter, translation is included without limitation in
627 :     the term ``modification''.) Each licensee is addressed as ``you''.
628 :    
629 :     Activities other than copying, distribution and modification are not
630 :     covered by this License; they are outside its scope. The act of
631 :     running the Program is not restricted, and the output from the Program
632 :     is covered only if its contents constitute a work based on the
633 :     Program (independent of having been made by running the Program).
634 :     Whether that is true depends on what the Program does.
635 :    
636 :     @item
637 :     You may copy and distribute verbatim copies of the Program's
638 :     source code as you receive it, in any medium, provided that you
639 :     conspicuously and appropriately publish on each copy an appropriate
640 :     copyright notice and disclaimer of warranty; keep intact all the
641 :     notices that refer to this License and to the absence of any warranty;
642 :     and give any other recipients of the Program a copy of this License
643 :     along with the Program.
644 :    
645 :     You may charge a fee for the physical act of transferring a copy, and
646 :     you may at your option offer warranty protection in exchange for a fee.
647 :    
648 :     @item
649 :     You may modify your copy or copies of the Program or any portion
650 :     of it, thus forming a work based on the Program, and copy and
651 :     distribute such modifications or work under the terms of Section 1
652 :     above, provided that you also meet all of these conditions:
653 :    
654 :     @enumerate a
655 :     @item
656 :     You must cause the modified files to carry prominent notices
657 :     stating that you changed the files and the date of any change.
658 :    
659 :     @item
660 :     You must cause any work that you distribute or publish, that in
661 :     whole or in part contains or is derived from the Program or any
662 :     part thereof, to be licensed as a whole at no charge to all third
663 :     parties under the terms of this License.
664 :    
665 :     @item
666 :     If the modified program normally reads commands interactively
667 :     when run, you must cause it, when started running for such
668 :     interactive use in the most ordinary way, to print or display an
669 :     announcement including an appropriate copyright notice and a
670 :     notice that there is no warranty (or else, saying that you provide
671 :     a warranty) and that users may redistribute the program under
672 :     these conditions, and telling the user how to view a copy of this
673 :     License. (Exception: if the Program itself is interactive but
674 :     does not normally print such an announcement, your work based on
675 :     the Program is not required to print an announcement.)
676 :     @end enumerate
677 :    
678 :     These requirements apply to the modified work as a whole. If
679 :     identifiable sections of that work are not derived from the Program,
680 :     and can be reasonably considered independent and separate works in
681 :     themselves, then this License, and its terms, do not apply to those
682 :     sections when you distribute them as separate works. But when you
683 :     distribute the same sections as part of a whole which is a work based
684 :     on the Program, the distribution of the whole must be on the terms of
685 :     this License, whose permissions for other licensees extend to the
686 :     entire whole, and thus to each and every part regardless of who wrote it.
687 :    
688 :     Thus, it is not the intent of this section to claim rights or contest
689 :     your rights to work written entirely by you; rather, the intent is to
690 :     exercise the right to control the distribution of derivative or
691 :     collective works based on the Program.
692 :    
693 :     In addition, mere aggregation of another work not based on the Program
694 :     with the Program (or with a work based on the Program) on a volume of
695 :     a storage or distribution medium does not bring the other work under
696 :     the scope of this License.
697 :    
698 :     @item
699 :     You may copy and distribute the Program (or a work based on it,
700 :     under Section 2) in object code or executable form under the terms of
701 :     Sections 1 and 2 above provided that you also do one of the following:
702 :    
703 :     @enumerate a
704 :     @item
705 :     Accompany it with the complete corresponding machine-readable
706 :     source code, which must be distributed under the terms of Sections
707 :     1 and 2 above on a medium customarily used for software interchange; or,
708 :    
709 :     @item
710 :     Accompany it with a written offer, valid for at least three
711 :     years, to give any third party, for a charge no more than your
712 :     cost of physically performing source distribution, a complete
713 :     machine-readable copy of the corresponding source code, to be
714 :     distributed under the terms of Sections 1 and 2 above on a medium
715 :     customarily used for software interchange; or,
716 :    
717 :     @item
718 :     Accompany it with the information you received as to the offer
719 :     to distribute corresponding source code. (This alternative is
720 :     allowed only for noncommercial distribution and only if you
721 :     received the program in object code or executable form with such
722 :     an offer, in accord with Subsection b above.)
723 :     @end enumerate
724 :    
725 :     The source code for a work means the preferred form of the work for
726 :     making modifications to it. For an executable work, complete source
727 :     code means all the source code for all modules it contains, plus any
728 :     associated interface definition files, plus the scripts used to
729 :     control compilation and installation of the executable. However, as a
730 :     special exception, the source code distributed need not include
731 :     anything that is normally distributed (in either source or binary
732 :     form) with the major components (compiler, kernel, and so on) of the
733 :     operating system on which the executable runs, unless that component
734 :     itself accompanies the executable.
735 :    
736 :     If distribution of executable or object code is made by offering
737 :     access to copy from a designated place, then offering equivalent
738 :     access to copy the source code from the same place counts as
739 :     distribution of the source code, even though third parties are not
740 :     compelled to copy the source along with the object code.
741 :    
742 :     @item
743 :     You may not copy, modify, sublicense, or distribute the Program
744 :     except as expressly provided under this License. Any attempt
745 :     otherwise to copy, modify, sublicense or distribute the Program is
746 :     void, and will automatically terminate your rights under this License.
747 :     However, parties who have received copies, or rights, from you under
748 :     this License will not have their licenses terminated so long as such
749 :     parties remain in full compliance.
750 :    
751 :     @item
752 :     You are not required to accept this License, since you have not
753 :     signed it. However, nothing else grants you permission to modify or
754 :     distribute the Program or its derivative works. These actions are
755 :     prohibited by law if you do not accept this License. Therefore, by
756 :     modifying or distributing the Program (or any work based on the
757 :     Program), you indicate your acceptance of this License to do so, and
758 :     all its terms and conditions for copying, distributing or modifying
759 :     the Program or works based on it.
760 :    
761 :     @item
762 :     Each time you redistribute the Program (or any work based on the
763 :     Program), the recipient automatically receives a license from the
764 :     original licensor to copy, distribute or modify the Program subject to
765 :     these terms and conditions. You may not impose any further
766 :     restrictions on the recipients' exercise of the rights granted herein.
767 :     You are not responsible for enforcing compliance by third parties to
768 :     this License.
769 :    
770 :     @item
771 :     If, as a consequence of a court judgment or allegation of patent
772 :     infringement or for any other reason (not limited to patent issues),
773 :     conditions are imposed on you (whether by court order, agreement or
774 :     otherwise) that contradict the conditions of this License, they do not
775 :     excuse you from the conditions of this License. If you cannot
776 :     distribute so as to satisfy simultaneously your obligations under this
777 :     License and any other pertinent obligations, then as a consequence you
778 :     may not distribute the Program at all. For example, if a patent
779 :     license would not permit royalty-free redistribution of the Program by
780 :     all those who receive copies directly or indirectly through you, then
781 :     the only way you could satisfy both it and this License would be to
782 :     refrain entirely from distribution of the Program.
783 :    
784 :     If any portion of this section is held invalid or unenforceable under
785 :     any particular circumstance, the balance of the section is intended to
786 :     apply and the section as a whole is intended to apply in other
787 :     circumstances.
788 :    
789 :     It is not the purpose of this section to induce you to infringe any
790 :     patents or other property right claims or to contest validity of any
791 :     such claims; this section has the sole purpose of protecting the
792 :     integrity of the free software distribution system, which is
793 :     implemented by public license practices. Many people have made
794 :     generous contributions to the wide range of software distributed
795 :     through that system in reliance on consistent application of that
796 :     system; it is up to the author/donor to decide if he or she is willing
797 :     to distribute software through any other system and a licensee cannot
798 :     impose that choice.
799 :    
800 :     This section is intended to make thoroughly clear what is believed to
801 :     be a consequence of the rest of this License.
802 :    
803 :     @item
804 :     If the distribution and/or use of the Program is restricted in
805 :     certain countries either by patents or by copyrighted interfaces, the
806 :     original copyright holder who places the Program under this License
807 :     may add an explicit geographical distribution limitation excluding
808 :     those countries, so that distribution is permitted only in or among
809 :     countries not thus excluded. In such case, this License incorporates
810 :     the limitation as if written in the body of this License.
811 :    
812 :     @item
813 :     The Free Software Foundation may publish revised and/or new versions
814 :     of the General Public License from time to time. Such new versions will
815 :     be similar in spirit to the present version, but may differ in detail to
816 :     address new problems or concerns.
817 :    
818 :     Each version is given a distinguishing version number. If the Program
819 :     specifies a version number of this License which applies to it and ``any
820 :     later version'', you have the option of following the terms and conditions
821 :     either of that version or of any later version published by the Free
822 :     Software Foundation. If the Program does not specify a version number of
823 :     this License, you may choose any version ever published by the Free Software
824 :     Foundation.
825 :    
826 :     @item
827 :     If you wish to incorporate parts of the Program into other free
828 :     programs whose distribution conditions are different, write to the author
829 :     to ask for permission. For software which is copyrighted by the Free
830 :     Software Foundation, write to the Free Software Foundation; we sometimes
831 :     make exceptions for this. Our decision will be guided by the two goals
832 :     of preserving the free status of all derivatives of our free software and
833 :     of promoting the sharing and reuse of software generally.
834 :    
835 :     @iftex
836 :     @heading NO WARRANTY
837 :     @end iftex
838 : anton 1.49 @ifnottex
839 : anton 1.1 @center NO WARRANTY
840 : anton 1.49 @end ifnottex
841 : anton 1.1
842 :     @item
843 :     BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
844 :     FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
845 :     OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
846 :     PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
847 :     OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
848 :     MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
849 :     TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
850 :     PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
851 :     REPAIR OR CORRECTION.
852 :    
853 :     @item
854 :     IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
855 :     WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
856 :     REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
857 :     INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
858 :     OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
859 :     TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
860 :     YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
861 :     PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
862 :     POSSIBILITY OF SUCH DAMAGES.
863 :     @end enumerate
864 :    
865 :     @iftex
866 :     @heading END OF TERMS AND CONDITIONS
867 :     @end iftex
868 : anton 1.49 @ifnottex
869 : anton 1.1 @center END OF TERMS AND CONDITIONS
870 : anton 1.49 @end ifnottex
871 : anton 1.1
872 :     @page
873 :     @unnumberedsec How to Apply These Terms to Your New Programs
874 :    
875 :     If you develop a new program, and you want it to be of the greatest
876 :     possible use to the public, the best way to achieve this is to make it
877 :     free software which everyone can redistribute and change under these terms.
878 :    
879 :     To do so, attach the following notices to the program. It is safest
880 :     to attach them to the start of each source file to most effectively
881 :     convey the exclusion of warranty; and each file should have at least
882 :     the ``copyright'' line and a pointer to where the full notice is found.
883 :    
884 :     @smallexample
885 :     @var{one line to give the program's name and a brief idea of what it does.}
886 :     Copyright (C) 19@var{yy} @var{name of author}
887 :    
888 :     This program is free software; you can redistribute it and/or modify
889 :     it under the terms of the GNU General Public License as published by
890 :     the Free Software Foundation; either version 2 of the License, or
891 :     (at your option) any later version.
892 :    
893 :     This program is distributed in the hope that it will be useful,
894 :     but WITHOUT ANY WARRANTY; without even the implied warranty of
895 :     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
896 :     GNU General Public License for more details.
897 :    
898 :     You should have received a copy of the GNU General Public License
899 :     along with this program; if not, write to the Free Software
900 :     Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
901 :     @end smallexample
902 :    
903 :     Also add information on how to contact you by electronic and paper mail.
904 :    
905 :     If the program is interactive, make it output a short notice like this
906 :     when it starts in an interactive mode:
907 :    
908 :     @smallexample
909 :     Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
910 :     Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
911 :     type `show w'.
912 :     This is free software, and you are welcome to redistribute it
913 :     under certain conditions; type `show c' for details.
914 :     @end smallexample
915 :    
916 :     The hypothetical commands @samp{show w} and @samp{show c} should show
917 :     the appropriate parts of the General Public License. Of course, the
918 :     commands you use may be called something other than @samp{show w} and
919 :     @samp{show c}; they could even be mouse-clicks or menu items---whatever
920 :     suits your program.
921 :    
922 :     You should also get your employer (if you work as a programmer) or your
923 :     school, if any, to sign a ``copyright disclaimer'' for the program, if
924 :     necessary. Here is a sample; alter the names:
925 :    
926 :     @smallexample
927 :     Yoyodyne, Inc., hereby disclaims all copyright interest in the program
928 :     `Gnomovision' (which makes passes at compilers) written by James Hacker.
929 :    
930 :     @var{signature of Ty Coon}, 1 April 1989
931 :     Ty Coon, President of Vice
932 :     @end smallexample
933 :    
934 :     This General Public License does not permit incorporating your program into
935 :     proprietary programs. If your program is a subroutine library, you may
936 :     consider it more useful to permit linking proprietary applications with the
937 :     library. If this is what you want to do, use the GNU Library General
938 :     Public License instead of this License.
939 :    
940 :     @iftex
941 :     @unnumbered Preface
942 :     @cindex Preface
943 : crook 1.21 This manual documents Gforth. Some introductory material is provided for
944 :     readers who are unfamiliar with Forth or who are migrating to Gforth
945 :     from other Forth compilers. However, this manual is primarily a
946 :     reference manual.
947 : anton 1.1 @end iftex
948 :    
949 : crook 1.28 @comment TODO much more blurb here.
950 : crook 1.26
951 :     @c ******************************************************************
952 : crook 1.29 @node Goals, Gforth Environment, License, Top
953 : crook 1.26 @comment node-name, next, previous, up
954 :     @chapter Goals of Gforth
955 :     @cindex goals of the Gforth project
956 :     The goal of the Gforth Project is to develop a standard model for
957 :     ANS Forth. This can be split into several subgoals:
958 :    
959 :     @itemize @bullet
960 :     @item
961 :     Gforth should conform to the ANS Forth Standard.
962 :     @item
963 :     It should be a model, i.e. it should define all the
964 :     implementation-dependent things.
965 :     @item
966 :     It should become standard, i.e. widely accepted and used. This goal
967 :     is the most difficult one.
968 :     @end itemize
969 :    
970 :     To achieve these goals Gforth should be
971 :     @itemize @bullet
972 :     @item
973 :     Similar to previous models (fig-Forth, F83)
974 :     @item
975 :     Powerful. It should provide for all the things that are considered
976 :     necessary today and even some that are not yet considered necessary.
977 :     @item
978 :     Efficient. It should not get the reputation of being exceptionally
979 :     slow.
980 :     @item
981 :     Free.
982 :     @item
983 :     Available on many machines/easy to port.
984 :     @end itemize
985 :    
986 :     Have we achieved these goals? Gforth conforms to the ANS Forth
987 :     standard. It may be considered a model, but we have not yet documented
988 :     which parts of the model are stable and which parts we are likely to
989 :     change. It certainly has not yet become a de facto standard, but it
990 :     appears to be quite popular. It has some similarities to and some
991 :     differences from previous models. It has some powerful features, but not
992 :     yet everything that we envisioned. We certainly have achieved our
993 : anton 1.65 execution speed goals (@pxref{Performance})@footnote{However, in 1998
994 :     the bar was raised when the major commercial Forth vendors switched to
995 :     native code compilers.}. It is free and available on many machines.
996 : crook 1.29
997 : crook 1.26 @c ******************************************************************
998 : anton 1.48 @node Gforth Environment, Tutorial, Goals, Top
999 : crook 1.29 @chapter Gforth Environment
1000 :     @cindex Gforth environment
1001 : crook 1.21
1002 : crook 1.45 Note: ultimately, the Gforth man page will be auto-generated from the
1003 : crook 1.29 material in this chapter.
1004 : crook 1.21
1005 :     @menu
1006 : crook 1.29 * Invoking Gforth:: Getting in
1007 :     * Leaving Gforth:: Getting out
1008 :     * Command-line editing::
1009 : anton 1.48 * Environment variables:: that affect how Gforth starts up
1010 : crook 1.29 * Gforth Files:: What gets installed and where
1011 : anton 1.48 * Startup speed:: When 35ms is not fast enough ...
1012 : crook 1.21 @end menu
1013 :    
1014 : anton 1.49 For related information about the creation of images see @ref{Image Files}.
1015 : crook 1.29
1016 : crook 1.21 @comment ----------------------------------------------
1017 : anton 1.48 @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1018 : crook 1.29 @section Invoking Gforth
1019 :     @cindex invoking Gforth
1020 :     @cindex running Gforth
1021 :     @cindex command-line options
1022 :     @cindex options on the command line
1023 :     @cindex flags on the command line
1024 : crook 1.21
1025 : anton 1.30 Gforth is made up of two parts; an executable ``engine'' (named
1026 :     @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1027 :     will usually just say @code{gforth} -- this automatically loads the
1028 :     default image file @file{gforth.fi}. In many other cases the default
1029 :     Gforth image will be invoked like this:
1030 : crook 1.21 @example
1031 : anton 1.30 gforth [file | -e forth-code] ...
1032 : crook 1.21 @end example
1033 : crook 1.29 @noindent
1034 :     This interprets the contents of the files and the Forth code in the order they
1035 :     are given.
1036 : crook 1.21
1037 : anton 1.30 In addition to the @file{gforth} engine, there is also an engine called
1038 :     @file{gforth-fast}, which is faster, but gives less informative error
1039 :     messages (@pxref{Error messages}).
1040 :    
1041 : crook 1.29 In general, the command line looks like this:
1042 : crook 1.21
1043 :     @example
1044 : anton 1.30 gforth[-fast] [engine options] [image options]
1045 : crook 1.21 @end example
1046 :    
1047 : anton 1.30 The engine options must come before the rest of the command
1048 : crook 1.29 line. They are:
1049 : crook 1.26
1050 : crook 1.29 @table @code
1051 :     @cindex -i, command-line option
1052 :     @cindex --image-file, command-line option
1053 :     @item --image-file @i{file}
1054 :     @itemx -i @i{file}
1055 :     Loads the Forth image @i{file} instead of the default
1056 :     @file{gforth.fi} (@pxref{Image Files}).
1057 : crook 1.21
1058 : anton 1.39 @cindex --appl-image, command-line option
1059 :     @item --appl-image @i{file}
1060 :     Loads the image @i{file} and leaves all further command-line arguments
1061 : anton 1.65 to the image (instead of processing them as engine options). This is
1062 :     useful for building executable application images on Unix, built with
1063 : anton 1.39 @code{gforthmi --application ...}.
1064 :    
1065 : crook 1.29 @cindex --path, command-line option
1066 :     @cindex -p, command-line option
1067 :     @item --path @i{path}
1068 :     @itemx -p @i{path}
1069 :     Uses @i{path} for searching the image file and Forth source code files
1070 :     instead of the default in the environment variable @code{GFORTHPATH} or
1071 :     the path specified at installation time (e.g.,
1072 :     @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1073 :     directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1074 : crook 1.21
1075 : crook 1.29 @cindex --dictionary-size, command-line option
1076 :     @cindex -m, command-line option
1077 :     @cindex @i{size} parameters for command-line options
1078 :     @cindex size of the dictionary and the stacks
1079 :     @item --dictionary-size @i{size}
1080 :     @itemx -m @i{size}
1081 :     Allocate @i{size} space for the Forth dictionary space instead of
1082 :     using the default specified in the image (typically 256K). The
1083 :     @i{size} specification for this and subsequent options consists of
1084 :     an integer and a unit (e.g.,
1085 :     @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1086 :     size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1087 :     @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1088 :     @code{e} is used.
1089 : crook 1.21
1090 : crook 1.29 @cindex --data-stack-size, command-line option
1091 :     @cindex -d, command-line option
1092 :     @item --data-stack-size @i{size}
1093 :     @itemx -d @i{size}
1094 :     Allocate @i{size} space for the data stack instead of using the
1095 :     default specified in the image (typically 16K).
1096 : crook 1.21
1097 : crook 1.29 @cindex --return-stack-size, command-line option
1098 :     @cindex -r, command-line option
1099 :     @item --return-stack-size @i{size}
1100 :     @itemx -r @i{size}
1101 :     Allocate @i{size} space for the return stack instead of using the
1102 :     default specified in the image (typically 15K).
1103 : crook 1.21
1104 : crook 1.29 @cindex --fp-stack-size, command-line option
1105 :     @cindex -f, command-line option
1106 :     @item --fp-stack-size @i{size}
1107 :     @itemx -f @i{size}
1108 :     Allocate @i{size} space for the floating point stack instead of
1109 :     using the default specified in the image (typically 15.5K). In this case
1110 :     the unit specifier @code{e} refers to floating point numbers.
1111 : crook 1.21
1112 : anton 1.48 @cindex --locals-stack-size, command-line option
1113 :     @cindex -l, command-line option
1114 :     @item --locals-stack-size @i{size}
1115 :     @itemx -l @i{size}
1116 :     Allocate @i{size} space for the locals stack instead of using the
1117 :     default specified in the image (typically 14.5K).
1118 :    
1119 :     @cindex -h, command-line option
1120 :     @cindex --help, command-line option
1121 :     @item --help
1122 :     @itemx -h
1123 :     Print a message about the command-line options
1124 :    
1125 :     @cindex -v, command-line option
1126 :     @cindex --version, command-line option
1127 :     @item --version
1128 :     @itemx -v
1129 :     Print version and exit
1130 :    
1131 :     @cindex --debug, command-line option
1132 :     @item --debug
1133 :     Print some information useful for debugging on startup.
1134 :    
1135 :     @cindex --offset-image, command-line option
1136 :     @item --offset-image
1137 :     Start the dictionary at a slightly different position than would be used
1138 :     otherwise (useful for creating data-relocatable images,
1139 :     @pxref{Data-Relocatable Image Files}).
1140 :    
1141 :     @cindex --no-offset-im, command-line option
1142 :     @item --no-offset-im
1143 :     Start the dictionary at the normal position.
1144 :    
1145 :     @cindex --clear-dictionary, command-line option
1146 :     @item --clear-dictionary
1147 :     Initialize all bytes in the dictionary to 0 before loading the image
1148 :     (@pxref{Data-Relocatable Image Files}).
1149 :    
1150 :     @cindex --die-on-signal, command-line-option
1151 :     @item --die-on-signal
1152 :     Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1153 :     or the segmentation violation SIGSEGV) by translating it into a Forth
1154 :     @code{THROW}. With this option, Gforth exits if it receives such a
1155 :     signal. This option is useful when the engine and/or the image might be
1156 :     severely broken (such that it causes another signal before recovering
1157 :     from the first); this option avoids endless loops in such cases.
1158 :     @end table
1159 :    
1160 :     @cindex loading files at startup
1161 :     @cindex executing code on startup
1162 :     @cindex batch processing with Gforth
1163 :     As explained above, the image-specific command-line arguments for the
1164 :     default image @file{gforth.fi} consist of a sequence of filenames and
1165 :     @code{-e @var{forth-code}} options that are interpreted in the sequence
1166 :     in which they are given. The @code{-e @var{forth-code}} or
1167 :     @code{--evaluate @var{forth-code}} option evaluates the Forth
1168 :     code. This option takes only one argument; if you want to evaluate more
1169 :     Forth words, you have to quote them or use @code{-e} several times. To exit
1170 :     after processing the command line (instead of entering interactive mode)
1171 :     append @code{-e bye} to the command line.
1172 :    
1173 :     @cindex versions, invoking other versions of Gforth
1174 :     If you have several versions of Gforth installed, @code{gforth} will
1175 :     invoke the version that was installed last. @code{gforth-@i{version}}
1176 :     invokes a specific version. If your environment contains the variable
1177 :     @code{GFORTHPATH}, you may want to override it by using the
1178 :     @code{--path} option.
1179 :    
1180 :     Not yet implemented:
1181 :     On startup the system first executes the system initialization file
1182 :     (unless the option @code{--no-init-file} is given; note that the system
1183 :     resulting from using this option may not be ANS Forth conformant). Then
1184 :     the user initialization file @file{.gforth.fs} is executed, unless the
1185 : crook 1.62 option @code{--no-rc} is given; this file is searched for in @file{.},
1186 : anton 1.48 then in @file{~}, then in the normal path (see above).
1187 :    
1188 :    
1189 :    
1190 :     @comment ----------------------------------------------
1191 :     @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1192 :     @section Leaving Gforth
1193 :     @cindex Gforth - leaving
1194 :     @cindex leaving Gforth
1195 :    
1196 :     You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1197 :     of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1198 :     option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1199 : anton 1.49 data are discarded. For ways of saving the state of the system before
1200 :     leaving Gforth see @ref{Image Files}.
1201 : anton 1.48
1202 :     doc-bye
1203 :    
1204 :    
1205 :     @comment ----------------------------------------------
1206 : anton 1.65 @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1207 : anton 1.48 @section Command-line editing
1208 :     @cindex command-line editing
1209 :    
1210 :     Gforth maintains a history file that records every line that you type to
1211 :     the text interpreter. This file is preserved between sessions, and is
1212 :     used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1213 :     repeatedly you can recall successively older commands from this (or
1214 :     previous) session(s). The full list of command-line editing facilities is:
1215 :    
1216 :     @itemize @bullet
1217 :     @item
1218 :     @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1219 :     commands from the history buffer.
1220 :     @item
1221 :     @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1222 :     from the history buffer.
1223 :     @item
1224 :     @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1225 :     @item
1226 :     @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1227 :     @item
1228 :     @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1229 :     closing up the line.
1230 :     @item
1231 :     @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1232 :     @item
1233 :     @kbd{Ctrl-a} to move the cursor to the start of the line.
1234 :     @item
1235 :     @kbd{Ctrl-e} to move the cursor to the end of the line.
1236 :     @item
1237 :     @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1238 :     line.
1239 :     @item
1240 :     @key{TAB} to step through all possible full-word completions of the word
1241 :     currently being typed.
1242 :     @item
1243 : anton 1.65 @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1244 :     using @code{bye}).
1245 :     @item
1246 :     @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1247 :     character under the cursor.
1248 : anton 1.48 @end itemize
1249 :    
1250 :     When editing, displayable characters are inserted to the left of the
1251 :     cursor position; the line is always in ``insert'' (as opposed to
1252 :     ``overstrike'') mode.
1253 :    
1254 :     @cindex history file
1255 :     @cindex @file{.gforth-history}
1256 :     On Unix systems, the history file is @file{~/.gforth-history} by
1257 :     default@footnote{i.e. it is stored in the user's home directory.}. You
1258 :     can find out the name and location of your history file using:
1259 :    
1260 :     @example
1261 :     history-file type \ Unix-class systems
1262 :    
1263 :     history-file type \ Other systems
1264 :     history-dir type
1265 :     @end example
1266 :    
1267 :     If you enter long definitions by hand, you can use a text editor to
1268 :     paste them out of the history file into a Forth source file for reuse at
1269 :     a later time.
1270 :    
1271 :     Gforth never trims the size of the history file, so you should do this
1272 :     periodically, if necessary.
1273 :    
1274 :     @comment this is all defined in history.fs
1275 :     @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1276 :     @comment chosen?
1277 :    
1278 :    
1279 :     @comment ----------------------------------------------
1280 : anton 1.65 @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1281 : anton 1.48 @section Environment variables
1282 :     @cindex environment variables
1283 :    
1284 :     Gforth uses these environment variables:
1285 :    
1286 :     @itemize @bullet
1287 :     @item
1288 :     @cindex @code{GFORTHHIST} -- environment variable
1289 :     @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1290 :     open/create the history file, @file{.gforth-history}. Default:
1291 :     @code{$HOME}.
1292 :    
1293 :     @item
1294 :     @cindex @code{GFORTHPATH} -- environment variable
1295 :     @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1296 :     for Forth source-code files.
1297 :    
1298 :     @item
1299 :     @cindex @code{GFORTH} -- environment variable
1300 : anton 1.49 @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1301 : anton 1.48
1302 :     @item
1303 :     @cindex @code{GFORTHD} -- environment variable
1304 : crook 1.62 @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1305 : anton 1.48
1306 :     @item
1307 :     @cindex @code{TMP}, @code{TEMP} - environment variable
1308 :     @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1309 :     location for the history file.
1310 :     @end itemize
1311 :    
1312 :     @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1313 :     @comment mentioning these.
1314 :    
1315 :     All the Gforth environment variables default to sensible values if they
1316 :     are not set.
1317 :    
1318 :    
1319 :     @comment ----------------------------------------------
1320 :     @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1321 :     @section Gforth files
1322 :     @cindex Gforth files
1323 :    
1324 :     When you install Gforth on a Unix system, it installs files in these
1325 :     locations by default:
1326 :    
1327 :     @itemize @bullet
1328 :     @item
1329 :     @file{/usr/local/bin/gforth}
1330 :     @item
1331 :     @file{/usr/local/bin/gforthmi}
1332 :     @item
1333 :     @file{/usr/local/man/man1/gforth.1} - man page.
1334 :     @item
1335 :     @file{/usr/local/info} - the Info version of this manual.
1336 :     @item
1337 :     @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1338 :     @item
1339 :     @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1340 :     @item
1341 :     @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1342 :     @item
1343 :     @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1344 :     @end itemize
1345 :    
1346 :     You can select different places for installation by using
1347 :     @code{configure} options (listed with @code{configure --help}).
1348 :    
1349 :     @comment ----------------------------------------------
1350 :     @node Startup speed, , Gforth Files, Gforth Environment
1351 :     @section Startup speed
1352 :     @cindex Startup speed
1353 :     @cindex speed, startup
1354 :    
1355 :     If Gforth is used for CGI scripts or in shell scripts, its startup
1356 :     speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1357 :     glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1358 :     system time.
1359 :    
1360 :     If startup speed is a problem, you may consider the following ways to
1361 :     improve it; or you may consider ways to reduce the number of startups
1362 : crook 1.62 (for example, by using Fast-CGI).
1363 : anton 1.48
1364 :     The first step to improve startup speed is to statically link Gforth, by
1365 :     building it with @code{XLDFLAGS=-static}. This requires more memory for
1366 :     the code and will therefore slow down the first invocation, but
1367 :     subsequent invocations avoid the dynamic linking overhead. Another
1368 :     disadvantage is that Gforth won't profit from library upgrades. As a
1369 :     result, @code{gforth-static -e bye} takes about 17.1ms user and
1370 :     8.2ms system time.
1371 :    
1372 :     The next step to improve startup speed is to use a non-relocatable image
1373 : anton 1.65 (@pxref{Non-Relocatable Image Files}). You can create this image with
1374 : anton 1.48 @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1375 :     @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1376 :     and a part of the copy-on-write overhead. The disadvantage is that the
1377 : crook 1.62 non-relocatable image does not work if the OS gives Gforth a different
1378 : anton 1.48 address for the dictionary, for whatever reason; so you better provide a
1379 :     fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1380 :     bye} takes about 15.3ms user and 7.5ms system time.
1381 :    
1382 :     The final step is to disable dictionary hashing in Gforth. Gforth
1383 :     builds the hash table on startup, which takes much of the startup
1384 :     overhead. You can do this by commenting out the @code{include hash.fs}
1385 :     in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1386 :     moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1387 :     The disadvantages are that functionality like @code{table} and
1388 :     @code{ekey} is missing and that text interpretation (e.g., compiling)
1389 :     now takes much longer. So, you should only use this method if there is
1390 :     no significant text interpretation to perform (the script should be
1391 : crook 1.62 compiled into the image, amongst other things). @code{gforth-static -i
1392 : anton 1.48 gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1393 :    
1394 :     @c ******************************************************************
1395 :     @node Tutorial, Introduction, Gforth Environment, Top
1396 :     @chapter Forth Tutorial
1397 :     @cindex Tutorial
1398 :     @cindex Forth Tutorial
1399 :    
1400 : crook 1.62 This tutorial can be used with any ANS-compliant Forth; any
1401 :     Gforth-specific features are marked as such and you can skip them if you
1402 :     work with another Forth. This tutorial does not explain all features of
1403 :     Forth, just enough to get you started and give you some ideas about the
1404 :     facilities available in Forth. Read the rest of the manual and the
1405 :     standard when you are through this.
1406 : anton 1.48
1407 :     The intended way to use this tutorial is that you work through it while
1408 :     sitting in front of the console, take a look at the examples and predict
1409 :     what they will do, then try them out; if the outcome is not as expected,
1410 :     find out why (e.g., by trying out variations of the example), so you
1411 :     understand what's going on. There are also some assignments that you
1412 :     should solve.
1413 :    
1414 :     This tutorial assumes that you have programmed before and know what,
1415 :     e.g., a loop is.
1416 :    
1417 :     @c !! explain compat library
1418 :    
1419 :     @menu
1420 :     * Starting Gforth Tutorial::
1421 :     * Syntax Tutorial::
1422 :     * Crash Course Tutorial::
1423 :     * Stack Tutorial::
1424 :     * Arithmetics Tutorial::
1425 :     * Stack Manipulation Tutorial::
1426 :     * Using files for Forth code Tutorial::
1427 :     * Comments Tutorial::
1428 :     * Colon Definitions Tutorial::
1429 :     * Decompilation Tutorial::
1430 :     * Stack-Effect Comments Tutorial::
1431 :     * Types Tutorial::
1432 :     * Factoring Tutorial::
1433 :     * Designing the stack effect Tutorial::
1434 :     * Local Variables Tutorial::
1435 :     * Conditional execution Tutorial::
1436 :     * Flags and Comparisons Tutorial::
1437 :     * General Loops Tutorial::
1438 :     * Counted loops Tutorial::
1439 :     * Recursion Tutorial::
1440 :     * Leaving definitions or loops Tutorial::
1441 :     * Return Stack Tutorial::
1442 :     * Memory Tutorial::
1443 :     * Characters and Strings Tutorial::
1444 :     * Alignment Tutorial::
1445 :     * Interpretation and Compilation Semantics and Immediacy Tutorial::
1446 :     * Execution Tokens Tutorial::
1447 :     * Exceptions Tutorial::
1448 :     * Defining Words Tutorial::
1449 :     * Arrays and Records Tutorial::
1450 :     * POSTPONE Tutorial::
1451 :     * Literal Tutorial::
1452 :     * Advanced macros Tutorial::
1453 :     * Compilation Tokens Tutorial::
1454 :     * Wordlists and Search Order Tutorial::
1455 :     @end menu
1456 :    
1457 :     @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1458 :     @section Starting Gforth
1459 :    
1460 :     You can start Gforth by typing its name:
1461 :    
1462 :     @example
1463 :     gforth
1464 :     @end example
1465 :    
1466 :     That puts you into interactive mode; you can leave Gforth by typing
1467 :     @code{bye}. While in Gforth, you can edit the command line and access
1468 :     the command line history with cursor keys, similar to bash.
1469 :    
1470 :    
1471 :     @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1472 :     @section Syntax
1473 :    
1474 :     A @dfn{word} is a sequence of arbitrary characters (expcept white
1475 :     space). Words are separated by white space. E.g., each of the
1476 :     following lines contains exactly one word:
1477 :    
1478 :     @example
1479 :     word
1480 :     !@@#$%^&*()
1481 :     1234567890
1482 :     5!a
1483 :     @end example
1484 :    
1485 :     A frequent beginner's error is to leave away necessary white space,
1486 :     resulting in an error like @samp{Undefined word}; so if you see such an
1487 :     error, check if you have put spaces wherever necessary.
1488 :    
1489 :     @example
1490 :     ." hello, world" \ correct
1491 :     ."hello, world" \ gives an "Undefined word" error
1492 :     @end example
1493 :    
1494 : anton 1.65 Gforth and most other Forth systems ignore differences in case (they are
1495 : anton 1.48 case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1496 :     your system is case-sensitive, you may have to type all the examples
1497 :     given here in upper case.
1498 :    
1499 :    
1500 :     @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1501 :     @section Crash Course
1502 :    
1503 :     Type
1504 :    
1505 :     @example
1506 :     0 0 !
1507 :     here execute
1508 :     ' catch >body 20 erase abort
1509 :     ' (quit) >body 20 erase
1510 :     @end example
1511 :    
1512 :     The last two examples are guaranteed to destroy parts of Gforth (and
1513 :     most other systems), so you better leave Gforth afterwards (if it has
1514 :     not finished by itself). On some systems you may have to kill gforth
1515 :     from outside (e.g., in Unix with @code{kill}).
1516 :    
1517 :     Now that you know how to produce crashes (and that there's not much to
1518 :     them), let's learn how to produce meaningful programs.
1519 :    
1520 :    
1521 :     @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1522 :     @section Stack
1523 :    
1524 :     The most obvious feature of Forth is the stack. When you type in a
1525 :     number, it is pushed on the stack. You can display the content of the
1526 :     stack with @code{.s}.
1527 :    
1528 :     @example
1529 :     1 2 .s
1530 :     3 .s
1531 :     @end example
1532 :    
1533 :     @code{.s} displays the top-of-stack to the right, i.e., the numbers
1534 :     appear in @code{.s} output as they appeared in the input.
1535 :    
1536 :     You can print the top of stack element with @code{.}.
1537 :    
1538 :     @example
1539 :     1 2 3 . . .
1540 :     @end example
1541 :    
1542 :     In general, words consume their stack arguments (@code{.s} is an
1543 :     exception).
1544 :    
1545 :     @assignment
1546 :     What does the stack contain after @code{5 6 7 .}?
1547 :     @endassignment
1548 :    
1549 :    
1550 :     @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1551 :     @section Arithmetics
1552 :    
1553 :     The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1554 :     operate on the top two stack items:
1555 :    
1556 :     @example
1557 :     2 2 + .
1558 :     2 1 - .
1559 :     7 3 mod .
1560 :     @end example
1561 :    
1562 :     The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1563 :     as in the corresponding infix expression (this is generally the case in
1564 :     Forth).
1565 :    
1566 :     Parentheses are superfluous (and not available), because the order of
1567 :     the words unambiguously determines the order of evaluation and the
1568 :     operands:
1569 :    
1570 :     @example
1571 :     3 4 + 5 * .
1572 :     3 4 5 * + .
1573 :     @end example
1574 :    
1575 :     @assignment
1576 :     What are the infix expressions corresponding to the Forth code above?
1577 :     Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1578 :     known as Postfix or RPN (Reverse Polish Notation).}.
1579 :     @endassignment
1580 :    
1581 :     To change the sign, use @code{negate}:
1582 :    
1583 :     @example
1584 :     2 negate .
1585 :     @end example
1586 :    
1587 :     @assignment
1588 :     Convert -(-3)*4-5 to Forth.
1589 :     @endassignment
1590 :    
1591 :     @code{/mod} performs both @code{/} and @code{mod}.
1592 :    
1593 :     @example
1594 :     7 3 /mod . .
1595 :     @end example
1596 :    
1597 :     @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1598 :     @section Stack Manipulation
1599 :    
1600 :     Stack manipulation words rearrange the data on the stack.
1601 :    
1602 :     @example
1603 :     1 .s drop .s
1604 :     1 .s dup .s drop drop .s
1605 :     1 2 .s over .s drop drop drop
1606 :     1 2 .s swap .s drop drop
1607 :     1 2 3 .s rot .s drop drop drop
1608 :     @end example
1609 :    
1610 :     These are the most important stack manipulation words. There are also
1611 :     variants that manipulate twice as many stack items:
1612 :    
1613 :     @example
1614 :     1 2 3 4 .s 2swap .s 2drop 2drop
1615 :     @end example
1616 :    
1617 :     Two more stack manipulation words are:
1618 :    
1619 :     @example
1620 :     1 2 .s nip .s drop
1621 :     1 2 .s tuck .s 2drop drop
1622 :     @end example
1623 :    
1624 :     @assignment
1625 :     Replace @code{nip} and @code{tuck} with combinations of other stack
1626 :     manipulation words.
1627 :    
1628 :     @example
1629 :     Given: How do you get:
1630 :     1 2 3 3 2 1
1631 :     1 2 3 1 2 3 2
1632 :     1 2 3 1 2 3 3
1633 :     1 2 3 1 3 3
1634 :     1 2 3 2 1 3
1635 :     1 2 3 4 4 3 2 1
1636 :     1 2 3 1 2 3 1 2 3
1637 :     1 2 3 4 1 2 3 4 1 2
1638 :     1 2 3
1639 :     1 2 3 1 2 3 4
1640 :     1 2 3 1 3
1641 :     @end example
1642 :     @endassignment
1643 :    
1644 :     @example
1645 :     5 dup * .
1646 :     @end example
1647 :    
1648 :     @assignment
1649 :     Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1650 :     Write a piece of Forth code that expects two numbers on the stack
1651 :     (@var{a} and @var{b}, with @var{b} on top) and computes
1652 :     @code{(a-b)(a+1)}.
1653 :     @endassignment
1654 :    
1655 :     @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1656 :     @section Using files for Forth code
1657 :    
1658 :     While working at the Forth command line is convenient for one-line
1659 :     examples and short one-off code, you probably want to store your source
1660 :     code in files for convenient editing and persistence. You can use your
1661 :     favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1662 :     Gforth}) to create @var{file} and use
1663 :    
1664 :     @example
1665 :     s" @var{file}" included
1666 :     @end example
1667 :    
1668 :     to load it into your Forth system. The file name extension I use for
1669 :     Forth files is @samp{.fs}.
1670 :    
1671 :     You can easily start Gforth with some files loaded like this:
1672 :    
1673 :     @example
1674 :     gforth @var{file1} @var{file2}
1675 :     @end example
1676 :    
1677 :     If an error occurs during loading these files, Gforth terminates,
1678 :     whereas an error during @code{INCLUDED} within Gforth usually gives you
1679 :     a Gforth command line. Starting the Forth system every time gives you a
1680 :     clean start every time, without interference from the results of earlier
1681 :     tries.
1682 :    
1683 :     I often put all the tests in a file, then load the code and run the
1684 :     tests with
1685 :    
1686 :     @example
1687 :     gforth @var{code} @var{tests} -e bye
1688 :     @end example
1689 :    
1690 :     (often by performing this command with @kbd{C-x C-e} in Emacs). The
1691 :     @code{-e bye} ensures that Gforth terminates afterwards so that I can
1692 :     restart this command without ado.
1693 :    
1694 :     The advantage of this approach is that the tests can be repeated easily
1695 :     every time the program ist changed, making it easy to catch bugs
1696 :     introduced by the change.
1697 :    
1698 :    
1699 :     @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1700 :     @section Comments
1701 :    
1702 :     @example
1703 :     \ That's a comment; it ends at the end of the line
1704 :     ( Another comment; it ends here: ) .s
1705 :     @end example
1706 :    
1707 :     @code{\} and @code{(} are ordinary Forth words and therefore have to be
1708 :     separated with white space from the following text.
1709 :    
1710 :     @example
1711 :     \This gives an "Undefined word" error
1712 :     @end example
1713 :    
1714 :     The first @code{)} ends a comment started with @code{(}, so you cannot
1715 :     nest @code{(}-comments; and you cannot comment out text containing a
1716 :     @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1717 :     avoid @code{)} in word names.}.
1718 :    
1719 :     I use @code{\}-comments for descriptive text and for commenting out code
1720 :     of one or more line; I use @code{(}-comments for describing the stack
1721 :     effect, the stack contents, or for commenting out sub-line pieces of
1722 :     code.
1723 :    
1724 :     The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1725 :     these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1726 :     region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1727 :     with @kbd{M-q}.
1728 :    
1729 :    
1730 :     @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1731 :     @section Colon Definitions
1732 :    
1733 :     are similar to procedures and functions in other programming languages.
1734 :    
1735 :     @example
1736 :     : squared ( n -- n^2 )
1737 :     dup * ;
1738 :     5 squared .
1739 :     7 squared .
1740 :     @end example
1741 :    
1742 :     @code{:} starts the colon definition; its name is @code{squared}. The
1743 :     following comment describes its stack effect. The words @code{dup *}
1744 :     are not executed, but compiled into the definition. @code{;} ends the
1745 :     colon definition.
1746 :    
1747 :     The newly-defined word can be used like any other word, including using
1748 :     it in other definitions:
1749 :    
1750 :     @example
1751 :     : cubed ( n -- n^3 )
1752 :     dup squared * ;
1753 :     -5 cubed .
1754 :     : fourth-power ( n -- n^4 )
1755 :     squared squared ;
1756 :     3 fourth-power .
1757 :     @end example
1758 :    
1759 :     @assignment
1760 :     Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1761 :     @code{/mod} in terms of other Forth words, and check if they work (hint:
1762 :     test your tests on the originals first). Don't let the
1763 :     @samp{redefined}-Messages spook you, they are just warnings.
1764 :     @endassignment
1765 :    
1766 :    
1767 :     @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1768 :     @section Decompilation
1769 :    
1770 :     You can decompile colon definitions with @code{see}:
1771 :    
1772 :     @example
1773 :     see squared
1774 :     see cubed
1775 :     @end example
1776 :    
1777 :     In Gforth @code{see} shows you a reconstruction of the source code from
1778 :     the executable code. Informations that were present in the source, but
1779 :     not in the executable code, are lost (e.g., comments).
1780 :    
1781 : anton 1.65 You can also decompile the predefined words:
1782 :    
1783 :     @example
1784 :     see .
1785 :     see +
1786 :     @end example
1787 :    
1788 :    
1789 : anton 1.48 @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1790 :     @section Stack-Effect Comments
1791 :    
1792 :     By convention the comment after the name of a definition describes the
1793 :     stack effect: The part in from of the @samp{--} describes the state of
1794 :     the stack before the execution of the definition, i.e., the parameters
1795 :     that are passed into the colon definition; the part behind the @samp{--}
1796 :     is the state of the stack after the execution of the definition, i.e.,
1797 :     the results of the definition. The stack comment only shows the top
1798 :     stack items that the definition accesses and/or changes.
1799 :    
1800 :     You should put a correct stack effect on every definition, even if it is
1801 :     just @code{( -- )}. You should also add some descriptive comment to
1802 :     more complicated words (I usually do this in the lines following
1803 :     @code{:}). If you don't do this, your code becomes unreadable (because
1804 :     you have to work through every definition before you can undertsand
1805 :     any).
1806 :    
1807 :     @assignment
1808 :     The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1809 :     x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1810 :     @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1811 : anton 1.65 are done, you can compare your stack effects to those in this manual
1812 : anton 1.48 (@pxref{Word Index}).
1813 :     @endassignment
1814 :    
1815 :     Sometimes programmers put comments at various places in colon
1816 :     definitions that describe the contents of the stack at that place (stack
1817 :     comments); i.e., they are like the first part of a stack-effect
1818 :     comment. E.g.,
1819 :    
1820 :     @example
1821 :     : cubed ( n -- n^3 )
1822 :     dup squared ( n n^2 ) * ;
1823 :     @end example
1824 :    
1825 :     In this case the stack comment is pretty superfluous, because the word
1826 :     is simple enough. If you think it would be a good idea to add such a
1827 :     comment to increase readability, you should also consider factoring the
1828 :     word into several simpler words (@pxref{Factoring Tutorial,,
1829 : anton 1.60 Factoring}), which typically eliminates the need for the stack comment;
1830 : anton 1.48 however, if you decide not to refactor it, then having such a comment is
1831 :     better than not having it.
1832 :    
1833 :     The names of the stack items in stack-effect and stack comments in the
1834 :     standard, in this manual, and in many programs specify the type through
1835 :     a type prefix, similar to Fortran and Hungarian notation. The most
1836 :     frequent prefixes are:
1837 :    
1838 :     @table @code
1839 :     @item n
1840 :     signed integer
1841 :     @item u
1842 :     unsigned integer
1843 :     @item c
1844 :     character
1845 :     @item f
1846 :     Boolean flags, i.e. @code{false} or @code{true}.
1847 :     @item a-addr,a-
1848 :     Cell-aligned address
1849 :     @item c-addr,c-
1850 :     Char-aligned address (note that a Char may have two bytes in Windows NT)
1851 :     @item xt
1852 :     Execution token, same size as Cell
1853 :     @item w,x
1854 :     Cell, can contain an integer or an address. It usually takes 32, 64 or
1855 :     16 bits (depending on your platform and Forth system). A cell is more
1856 :     commonly known as machine word, but the term @emph{word} already means
1857 :     something different in Forth.
1858 :     @item d
1859 :     signed double-cell integer
1860 :     @item ud
1861 :     unsigned double-cell integer
1862 :     @item r
1863 :     Float (on the FP stack)
1864 :     @end table
1865 :    
1866 :     You can find a more complete list in @ref{Notation}.
1867 :    
1868 :     @assignment
1869 :     Write stack-effect comments for all definitions you have written up to
1870 :     now.
1871 :     @endassignment
1872 :    
1873 :    
1874 :     @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1875 :     @section Types
1876 :    
1877 :     In Forth the names of the operations are not overloaded; so similar
1878 :     operations on different types need different names; e.g., @code{+} adds
1879 :     integers, and you have to use @code{f+} to add floating-point numbers.
1880 :     The following prefixes are often used for related operations on
1881 :     different types:
1882 :    
1883 :     @table @code
1884 :     @item (none)
1885 :     signed integer
1886 :     @item u
1887 :     unsigned integer
1888 :     @item c
1889 :     character
1890 :     @item d
1891 :     signed double-cell integer
1892 :     @item ud, du
1893 :     unsigned double-cell integer
1894 :     @item 2
1895 :     two cells (not-necessarily double-cell numbers)
1896 :     @item m, um
1897 :     mixed single-cell and double-cell operations
1898 :     @item f
1899 :     floating-point (note that in stack comments @samp{f} represents flags,
1900 :     and @samp{r} represents FP number).
1901 :     @end table
1902 :    
1903 :     If there are no differences between the signed and the unsigned variant
1904 :     (e.g., for @code{+}), there is only the prefix-less variant.
1905 :    
1906 :     Forth does not perform type checking, neither at compile time, nor at
1907 :     run time. If you use the wrong oeration, the data are interpreted
1908 :     incorrectly:
1909 :    
1910 :     @example
1911 :     -1 u.
1912 :     @end example
1913 :    
1914 :     If you have only experience with type-checked languages until now, and
1915 :     have heard how important type-checking is, don't panic! In my
1916 :     experience (and that of other Forthers), type errors in Forth code are
1917 :     usually easy to find (once you get used to it), the increased vigilance
1918 :     of the programmer tends to catch some harder errors in addition to most
1919 :     type errors, and you never have to work around the type system, so in
1920 :     most situations the lack of type-checking seems to be a win (projects to
1921 :     add type checking to Forth have not caught on).
1922 :    
1923 :    
1924 :     @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1925 :     @section Factoring
1926 :    
1927 :     If you try to write longer definitions, you will soon find it hard to
1928 :     keep track of the stack contents. Therefore, good Forth programmers
1929 :     tend to write only short definitions (e.g., three lines). The art of
1930 :     finding meaningful short definitions is known as factoring (as in
1931 :     factoring polynomials).
1932 :    
1933 :     Well-factored programs offer additional advantages: smaller, more
1934 :     general words, are easier to test and debug and can be reused more and
1935 :     better than larger, specialized words.
1936 :    
1937 :     So, if you run into difficulties with stack management, when writing
1938 :     code, try to define meaningful factors for the word, and define the word
1939 :     in terms of those. Even if a factor contains only two words, it is
1940 :     often helpful.
1941 :    
1942 : anton 1.65 Good factoring is not easy, and it takes some practice to get the knack
1943 :     for it; but even experienced Forth programmers often don't find the
1944 :     right solution right away, but only when rewriting the program. So, if
1945 :     you don't come up with a good solution immediately, keep trying, don't
1946 :     despair.
1947 : anton 1.48
1948 :     @c example !!
1949 :    
1950 :    
1951 :     @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1952 :     @section Designing the stack effect
1953 :    
1954 :     In other languages you can use an arbitrary order of parameters for a
1955 : anton 1.65 function; and since there is only one result, you don't have to deal with
1956 : anton 1.48 the order of results, either.
1957 :    
1958 :     In Forth (and other stack-based languages, e.g., Postscript) the
1959 :     parameter and result order of a definition is important and should be
1960 :     designed well. The general guideline is to design the stack effect such
1961 :     that the word is simple to use in most cases, even if that complicates
1962 :     the implementation of the word. Some concrete rules are:
1963 :    
1964 :     @itemize @bullet
1965 :    
1966 :     @item
1967 :     Words consume all of their parameters (e.g., @code{.}).
1968 :    
1969 :     @item
1970 :     If there is a convention on the order of parameters (e.g., from
1971 :     mathematics or another programming language), stick with it (e.g.,
1972 :     @code{-}).
1973 :    
1974 :     @item
1975 :     If one parameter usually requires only a short computation (e.g., it is
1976 :     a constant), pass it on the top of the stack. Conversely, parameters
1977 :     that usually require a long sequence of code to compute should be passed
1978 :     as the bottom (i.e., first) parameter. This makes the code easier to
1979 :     read, because reader does not need to keep track of the bottom item
1980 :     through a long sequence of code (or, alternatively, through stack
1981 : anton 1.49 manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1982 : anton 1.48 address on top of the stack because it is usually simpler to compute
1983 :     than the stored value (often the address is just a variable).
1984 :    
1985 :     @item
1986 :     Similarly, results that are usually consumed quickly should be returned
1987 :     on the top of stack, whereas a result that is often used in long
1988 :     computations should be passed as bottom result. E.g., the file words
1989 :     like @code{open-file} return the error code on the top of stack, because
1990 :     it is usually consumed quickly by @code{throw}; moreover, the error code
1991 :     has to be checked before doing anything with the other results.
1992 :    
1993 :     @end itemize
1994 :    
1995 :     These rules are just general guidelines, don't lose sight of the overall
1996 :     goal to make the words easy to use. E.g., if the convention rule
1997 :     conflicts with the computation-length rule, you might decide in favour
1998 :     of the convention if the word will be used rarely, and in favour of the
1999 :     computation-length rule if the word will be used frequently (because
2000 :     with frequent use the cost of breaking the computation-length rule would
2001 :     be quite high, and frequent use makes it easier to remember an
2002 :     unconventional order).
2003 :    
2004 :     @c example !! structure package
2005 :    
2006 : anton 1.65
2007 : anton 1.48 @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2008 :     @section Local Variables
2009 :    
2010 :     You can define local variables (@emph{locals}) in a colon definition:
2011 :    
2012 :     @example
2013 :     : swap @{ a b -- b a @}
2014 :     b a ;
2015 :     1 2 swap .s 2drop
2016 :     @end example
2017 :    
2018 :     (If your Forth system does not support this syntax, include
2019 :     @file{compat/anslocals.fs} first).
2020 :    
2021 :     In this example @code{@{ a b -- b a @}} is the locals definition; it
2022 :     takes two cells from the stack, puts the top of stack in @code{b} and
2023 :     the next stack element in @code{a}. @code{--} starts a comment ending
2024 :     with @code{@}}. After the locals definition, using the name of the
2025 :     local will push its value on the stack. You can leave the comment
2026 :     part (@code{-- b a}) away:
2027 :    
2028 :     @example
2029 :     : swap ( x1 x2 -- x2 x1 )
2030 :     @{ a b @} b a ;
2031 :     @end example
2032 :    
2033 :     In Gforth you can have several locals definitions, anywhere in a colon
2034 :     definition; in contrast, in a standard program you can have only one
2035 :     locals definition per colon definition, and that locals definition must
2036 :     be outside any controll structure.
2037 :    
2038 :     With locals you can write slightly longer definitions without running
2039 :     into stack trouble. However, I recommend trying to write colon
2040 :     definitions without locals for exercise purposes to help you gain the
2041 :     essential factoring skills.
2042 :    
2043 :     @assignment
2044 :     Rewrite your definitions until now with locals
2045 :     @endassignment
2046 :    
2047 :    
2048 :     @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2049 :     @section Conditional execution
2050 :    
2051 :     In Forth you can use control structures only inside colon definitions.
2052 :     An @code{if}-structure looks like this:
2053 :    
2054 :     @example
2055 :     : abs ( n1 -- +n2 )
2056 :     dup 0 < if
2057 :     negate
2058 :     endif ;
2059 :     5 abs .
2060 :     -5 abs .
2061 :     @end example
2062 :    
2063 :     @code{if} takes a flag from the stack. If the flag is non-zero (true),
2064 :     the following code is performed, otherwise execution continues after the
2065 : pazsan 1.51 @code{endif} (or @code{else}). @code{<} compares the top two stack
2066 : anton 1.48 elements and prioduces a flag:
2067 :    
2068 :     @example
2069 :     1 2 < .
2070 :     2 1 < .
2071 :     1 1 < .
2072 :     @end example
2073 :    
2074 :     Actually the standard name for @code{endif} is @code{then}. This
2075 :     tutorial presents the examples using @code{endif}, because this is often
2076 :     less confusing for people familiar with other programming languages
2077 :     where @code{then} has a different meaning. If your system does not have
2078 :     @code{endif}, define it with
2079 :    
2080 :     @example
2081 :     : endif postpone then ; immediate
2082 :     @end example
2083 :    
2084 :     You can optionally use an @code{else}-part:
2085 :    
2086 :     @example
2087 :     : min ( n1 n2 -- n )
2088 :     2dup < if
2089 :     drop
2090 :     else
2091 :     nip
2092 :     endif ;
2093 :     2 3 min .
2094 :     3 2 min .
2095 :     @end example
2096 :    
2097 :     @assignment
2098 :     Write @code{min} without @code{else}-part (hint: what's the definition
2099 :     of @code{nip}?).
2100 :     @endassignment
2101 :    
2102 :    
2103 :     @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2104 :     @section Flags and Comparisons
2105 :    
2106 :     In a false-flag all bits are clear (0 when interpreted as integer). In
2107 :     a canonical true-flag all bits are set (-1 as a twos-complement signed
2108 :     integer); in many contexts (e.g., @code{if}) any non-zero value is
2109 :     treated as true flag.
2110 :    
2111 :     @example
2112 :     false .
2113 :     true .
2114 :     true hex u. decimal
2115 :     @end example
2116 :    
2117 :     Comparison words produce canonical flags:
2118 :    
2119 :     @example
2120 :     1 1 = .
2121 :     1 0= .
2122 :     0 1 < .
2123 :     0 0 < .
2124 :     -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2125 :     -1 1 < .
2126 :     @end example
2127 :    
2128 :     Gforth supports all combinations of the prefixes @code{0 u d d0 du} (or
2129 :     none) and the comparisons @code{= <> < > <= >=}. Only a part of these
2130 : anton 1.60 combinations are standard (for details see the standard or @ref{Word
2131 : anton 1.61 Index}).
2132 : anton 1.48
2133 :     You can use @code{and or xor invert} can be used as operations on
2134 :     canonical flags. Actually they are bitwise operations:
2135 :    
2136 :     @example
2137 :     1 2 and .
2138 :     1 2 or .
2139 :     1 3 xor .
2140 :     1 invert .
2141 :     @end example
2142 :    
2143 :     You can convert a zero/non-zero flag into a canonical flag with
2144 :     @code{0<>} (and complement it on the way with @code{0=}).
2145 :    
2146 :     @example
2147 :     1 0= .
2148 :     1 0<> .
2149 :     @end example
2150 :    
2151 : anton 1.65 You can use the all-bits-set feature of canonical flags and the bitwise
2152 : anton 1.48 operation of the Boolean operations to avoid @code{if}s:
2153 :    
2154 :     @example
2155 :     : foo ( n1 -- n2 )
2156 :     0= if
2157 :     14
2158 :     else
2159 :     0
2160 :     endif ;
2161 :     0 foo .
2162 :     1 foo .
2163 :    
2164 :     : foo ( n1 -- n2 )
2165 :     0= 14 and ;
2166 :     0 foo .
2167 :     1 foo .
2168 :     @end example
2169 :    
2170 :     @assignment
2171 :     Write @code{min} without @code{if}.
2172 :     @endassignment
2173 :    
2174 :    
2175 :     @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2176 :     @section General Loops
2177 :    
2178 :     The endless loop is the most simple one:
2179 :    
2180 :     @example
2181 :     : endless ( -- )
2182 :     0 begin
2183 :     dup . 1+
2184 :     again ;
2185 :     endless
2186 :     @end example
2187 :    
2188 :     Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2189 :     does nothing at run-time, @code{again} jumps back to @code{begin}.
2190 :    
2191 :     A loop with one exit at any place looks like this:
2192 :    
2193 :     @example
2194 :     : log2 ( +n1 -- n2 )
2195 :     \ logarithmus dualis of n1>0, rounded down to the next integer
2196 :     assert( dup 0> )
2197 :     2/ 0 begin
2198 :     over 0> while
2199 :     1+ swap 2/ swap
2200 :     repeat
2201 :     nip ;
2202 :     7 log2 .
2203 :     8 log2 .
2204 :     @end example
2205 :    
2206 :     At run-time @code{while} consumes a flag; if it is 0, execution
2207 : pazsan 1.51 continues behind the @code{repeat}; if the flag is non-zero, execution
2208 : anton 1.48 continues behind the @code{while}. @code{Repeat} jumps back to
2209 :     @code{begin}, just like @code{again}.
2210 :    
2211 :     In Forth there are many combinations/abbreviations, like @code{1+}.
2212 :     However, @code{2/} is not one of them; it shifts it's argument right by
2213 :     one bit (arithmetic shift right):
2214 :    
2215 :     @example
2216 :     -5 2 / .
2217 :     -5 2/ .
2218 :     @end example
2219 :    
2220 :     @code{assert(} is no standard word, but you can get it on systems other
2221 :     then Gforth by including @file{compat/assert.fs}. You can see what it
2222 :     does by trying
2223 :    
2224 :     @example
2225 :     0 log2 .
2226 :     @end example
2227 :    
2228 :     Here's a loop with an exit at the end:
2229 :    
2230 :     @example
2231 :     : log2 ( +n1 -- n2 )
2232 :     \ logarithmus dualis of n1>0, rounded down to the next integer
2233 :     assert( dup 0 > )
2234 :     -1 begin
2235 :     1+ swap 2/ swap
2236 :     over 0 <=
2237 :     until
2238 :     nip ;
2239 :     @end example
2240 :    
2241 :     @code{Until} consumes a flag; if it is non-zero, execution continues at
2242 :     the @code{begin}, otherwise after the @code{until}.
2243 :    
2244 :     @assignment
2245 :     Write a definition for computing the greatest common divisor.
2246 :     @endassignment
2247 :    
2248 :    
2249 :     @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2250 :     @section Counted loops
2251 :    
2252 :     @example
2253 :     : ^ ( n1 u -- n )
2254 :     \ n = the uth power of u1
2255 :     1 swap 0 u+do
2256 :     over *
2257 :     loop
2258 :     nip ;
2259 :     3 2 ^ .
2260 :     4 3 ^ .
2261 :     @end example
2262 :    
2263 :     @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2264 :     have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2265 :     performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2266 :     times (or not at all, if @code{u3-u4<0}).
2267 :    
2268 :     You can see the stack effect design rules at work in the stack effect of
2269 :     the loop start words: Since the start value of the loop is more
2270 :     frequently constant than the end value, the start value is passed on
2271 :     the top-of-stack.
2272 :    
2273 :     You can access the counter of a counted loop with @code{i}:
2274 :    
2275 :     @example
2276 :     : fac ( u -- u! )
2277 :     1 swap 1+ 1 u+do
2278 :     i *
2279 :     loop ;
2280 :     5 fac .
2281 :     7 fac .
2282 :     @end example
2283 :    
2284 :     There is also @code{+do}, which expects signed numbers (important for
2285 :     deciding whether to enter the loop).
2286 :    
2287 :     @assignment
2288 :     Write a definition for computing the nth Fibonacci number.
2289 :     @endassignment
2290 :    
2291 : anton 1.65 You can also use increments other than 1:
2292 :    
2293 :     @example
2294 :     : up2 ( n1 n2 -- )
2295 :     +do
2296 :     i .
2297 :     2 +loop ;
2298 :     10 0 up2
2299 :    
2300 :     : down2 ( n1 n2 -- )
2301 :     -do
2302 :     i .
2303 :     2 -loop ;
2304 :     0 10 down2
2305 :     @end example
2306 : anton 1.48
2307 :    
2308 :     @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2309 :     @section Recursion
2310 :    
2311 :     Usually the name of a definition is not visible in the definition; but
2312 :     earlier definitions are usually visible:
2313 :    
2314 :     @example
2315 :     1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2316 :     : / ( n1 n2 -- n )
2317 :     dup 0= if
2318 :     -10 throw \ report division by zero
2319 :     endif
2320 :     / \ old version
2321 :     ;
2322 :     1 0 /
2323 :     @end example
2324 :    
2325 :     For recursive definitions you can use @code{recursive} (non-standard) or
2326 :     @code{recurse}:
2327 :    
2328 :     @example
2329 :     : fac1 ( n -- n! ) recursive
2330 :     dup 0> if
2331 :     dup 1- fac1 *
2332 :     else
2333 :     drop 1
2334 :     endif ;
2335 :     7 fac1 .
2336 :    
2337 :     : fac2 ( n -- n! )
2338 :     dup 0> if
2339 :     dup 1- recurse *
2340 :     else
2341 :     drop 1
2342 :     endif ;
2343 :     8 fac2 .
2344 :     @end example
2345 :    
2346 :     @assignment
2347 :     Write a recursive definition for computing the nth Fibonacci number.
2348 :     @endassignment
2349 :    
2350 :    
2351 :     @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2352 :     @section Leaving definitions or loops
2353 :    
2354 :     @code{EXIT} exits the current definition right away. For every counted
2355 :     loop that is left in this way, an @code{UNLOOP} has to be performed
2356 :     before the @code{EXIT}:
2357 :    
2358 :     @c !! real examples
2359 :     @example
2360 :     : ...
2361 :     ... u+do
2362 :     ... if
2363 :     ... unloop exit
2364 :     endif
2365 :     ...
2366 :     loop
2367 :     ... ;
2368 :     @end example
2369 :    
2370 :     @code{LEAVE} leaves the innermost counted loop right away:
2371 :    
2372 :     @example
2373 :     : ...
2374 :     ... u+do
2375 :     ... if
2376 :     ... leave
2377 :     endif
2378 :     ...
2379 :     loop
2380 :     ... ;
2381 :     @end example
2382 :    
2383 : anton 1.65 @c !! example
2384 : anton 1.48
2385 :     @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2386 :     @section Return Stack
2387 :    
2388 :     In addition to the data stack Forth also has a second stack, the return
2389 :     stack; most Forth systems store the return addresses of procedure calls
2390 :     there (thus its name). Programmers can also use this stack:
2391 :    
2392 :     @example
2393 :     : foo ( n1 n2 -- )
2394 :     .s
2395 :     >r .s
2396 : anton 1.50 r@@ .
2397 : anton 1.48 >r .s
2398 : anton 1.50 r@@ .
2399 : anton 1.48 r> .
2400 : anton 1.50 r@@ .
2401 : anton 1.48 r> . ;
2402 :     1 2 foo
2403 :     @end example
2404 :    
2405 :     @code{>r} takes an element from the data stack and pushes it onto the
2406 :     return stack; conversely, @code{r>} moves an elementm from the return to
2407 :     the data stack; @code{r@@} pushes a copy of the top of the return stack
2408 :     on the return stack.
2409 :    
2410 :     Forth programmers usually use the return stack for storing data
2411 :     temporarily, if using the data stack alone would be too complex, and
2412 :     factoring and locals are not an option:
2413 :    
2414 :     @example
2415 :     : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2416 :     rot >r rot r> ;
2417 :     @end example
2418 :    
2419 :     The return address of the definition and the loop control parameters of
2420 :     counted loops usually reside on the return stack, so you have to take
2421 :     all items, that you have pushed on the return stack in a colon
2422 :     definition or counted loop, from the return stack before the definition
2423 :     or loop ends. You cannot access items that you pushed on the return
2424 :     stack outside some definition or loop within the definition of loop.
2425 :    
2426 :     If you miscount the return stack items, this usually ends in a crash:
2427 :    
2428 :     @example
2429 :     : crash ( n -- )
2430 :     >r ;
2431 :     5 crash
2432 :     @end example
2433 :    
2434 :     You cannot mix using locals and using the return stack (according to the
2435 :     standard; Gforth has no problem). However, they solve the same
2436 :     problems, so this shouldn't be an issue.
2437 :    
2438 :     @assignment
2439 :     Can you rewrite any of the definitions you wrote until now in a better
2440 :     way using the return stack?
2441 :     @endassignment
2442 :    
2443 :    
2444 :     @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2445 :     @section Memory
2446 :    
2447 :     You can create a global variable @code{v} with
2448 :    
2449 :     @example
2450 :     variable v ( -- addr )
2451 :     @end example
2452 :    
2453 :     @code{v} pushes the address of a cell in memory on the stack. This cell
2454 :     was reserved by @code{variable}. You can use @code{!} (store) to store
2455 :     values into this cell and @code{@@} (fetch) to load the value from the
2456 :     stack into memory:
2457 :    
2458 :     @example
2459 :     v .
2460 :     5 v ! .s
2461 : anton 1.50 v @@ .
2462 : anton 1.48 @end example
2463 :    
2464 : anton 1.65 You can see a raw dump of memory with @code{dump}:
2465 :    
2466 :     @example
2467 :     v 1 cells .s dump
2468 :     @end example
2469 :    
2470 :     @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2471 :     generally, address units (aus)) that @code{n1 cells} occupy. You can
2472 :     also reserve more memory:
2473 : anton 1.48
2474 :     @example
2475 :     create v2 20 cells allot
2476 : anton 1.65 v2 20 cells dump
2477 : anton 1.48 @end example
2478 :    
2479 : anton 1.65 creates a word @code{v2} and reserves 20 uninitialized cells; the
2480 :     address pushed by @code{v2} points to the start of these 20 cells. You
2481 :     can use address arithmetic to access these cells:
2482 : anton 1.48
2483 :     @example
2484 :     3 v2 5 cells + !
2485 : anton 1.65 v2 20 cells dump
2486 : anton 1.48 @end example
2487 :    
2488 :     You can reserve and initialize memory with @code{,}:
2489 :    
2490 :     @example
2491 :     create v3
2492 :     5 , 4 , 3 , 2 , 1 ,
2493 : anton 1.50 v3 @@ .
2494 :     v3 cell+ @@ .
2495 :     v3 2 cells + @@ .
2496 : anton 1.65 v3 5 cells dump
2497 : anton 1.48 @end example
2498 :    
2499 :     @assignment
2500 :     Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2501 :     @code{u} cells, with the first of these cells at @code{addr}, the next
2502 :     one at @code{addr cell+} etc.
2503 :     @endassignment
2504 :    
2505 :     You can also reserve memory without creating a new word:
2506 :    
2507 :     @example
2508 : anton 1.60 here 10 cells allot .
2509 :     here .
2510 : anton 1.48 @end example
2511 :    
2512 :     @code{Here} pushes the start address of the memory area. You should
2513 :     store it somewhere, or you will have a hard time finding the memory area
2514 :     again.
2515 :    
2516 :     @code{Allot} manages dictionary memory. The dictionary memory contains
2517 :     the system's data structures for words etc. on Gforth and most other
2518 :     Forth systems. It is managed like a stack: You can free the memory that
2519 :     you have just @code{allot}ed with
2520 :    
2521 :     @example
2522 :     -10 cells allot
2523 : anton 1.60 here .
2524 : anton 1.48 @end example
2525 :    
2526 :     Note that you cannot do this if you have created a new word in the
2527 :     meantime (because then your @code{allot}ed memory is no longer on the
2528 :     top of the dictionary ``stack'').
2529 :    
2530 :     Alternatively, you can use @code{allocate} and @code{free} which allow
2531 :     freeing memory in any order:
2532 :    
2533 :     @example
2534 :     10 cells allocate throw .s
2535 :     20 cells allocate throw .s
2536 :     swap
2537 :     free throw
2538 :     free throw
2539 :     @end example
2540 :    
2541 :     The @code{throw}s deal with errors (e.g., out of memory).
2542 :    
2543 : anton 1.65 And there is also a
2544 :     @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2545 :     garbage collector}, which eliminates the need to @code{free} memory
2546 :     explicitly.
2547 : anton 1.48
2548 :    
2549 :     @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2550 :     @section Characters and Strings
2551 :    
2552 :     On the stack characters take up a cell, like numbers. In memory they
2553 :     have their own size (one 8-bit byte on most systems), and therefore
2554 :     require their own words for memory access:
2555 :    
2556 :     @example
2557 :     create v4
2558 :     104 c, 97 c, 108 c, 108 c, 111 c,
2559 : anton 1.50 v4 4 chars + c@@ .
2560 : anton 1.65 v4 5 chars dump
2561 : anton 1.48 @end example
2562 :    
2563 :     The preferred representation of strings on the stack is @code{addr
2564 :     u-count}, where @code{addr} is the address of the first character and
2565 :     @code{u-count} is the number of characters in the string.
2566 :    
2567 :     @example
2568 :     v4 5 type
2569 :     @end example
2570 :    
2571 :     You get a string constant with
2572 :    
2573 :     @example
2574 :     s" hello, world" .s
2575 :     type
2576 :     @end example
2577 :    
2578 :     Make sure you have a space between @code{s"} and the string; @code{s"}
2579 :     is a normal Forth word and must be delimited with white space (try what
2580 :     happens when you remove the space).
2581 :    
2582 :     However, this interpretive use of @code{s"} is quite restricted: the
2583 :     string exists only until the next call of @code{s"} (some Forth systems
2584 :     keep more than one of these strings, but usually they still have a
2585 : crook 1.62 limited lifetime).
2586 : anton 1.48
2587 :     @example
2588 :     s" hello," s" world" .s
2589 :     type
2590 :     type
2591 :     @end example
2592 :    
2593 : crook 1.62 You can also use @code{s"} in a definition, and the resulting
2594 :     strings then live forever (well, for as long as the definition):
2595 : anton 1.48
2596 :     @example
2597 :     : foo s" hello," s" world" ;
2598 :     foo .s
2599 :     type
2600 :     type
2601 :     @end example
2602 :    
2603 :     @assignment
2604 :     @code{Emit ( c -- )} types @code{c} as character (not a number).
2605 :     Implement @code{type ( addr u -- )}.
2606 :     @endassignment
2607 :    
2608 :     @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2609 :     @section Alignment
2610 :    
2611 :     On many processors cells have to be aligned in memory, if you want to
2612 :     access them with @code{@@} and @code{!} (and even if the processor does
2613 : crook 1.62 not require alignment, access to aligned cells is faster).
2614 : anton 1.48
2615 :     @code{Create} aligns @code{here} (i.e., the place where the next
2616 :     allocation will occur, and that the @code{create}d word points to).
2617 :     Likewise, the memory produced by @code{allocate} starts at an aligned
2618 :     address. Adding a number of @code{cells} to an aligned address produces
2619 :     another aligned address.
2620 :    
2621 :     However, address arithmetic involving @code{char+} and @code{chars} can
2622 :     create an address that is not cell-aligned. @code{Aligned ( addr --
2623 :     a-addr )} produces the next aligned address:
2624 :    
2625 :     @example
2626 : anton 1.50 v3 char+ aligned .s @@ .
2627 :     v3 char+ .s @@ .
2628 : anton 1.48 @end example
2629 :    
2630 :     Similarly, @code{align} advances @code{here} to the next aligned
2631 :     address:
2632 :    
2633 :     @example
2634 :     create v5 97 c,
2635 :     here .
2636 :     align here .
2637 :     1000 ,
2638 :     @end example
2639 :    
2640 :     Note that you should use aligned addresses even if your processor does
2641 :     not require them, if you want your program to be portable.
2642 :    
2643 :    
2644 :     @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2645 :     @section Interpretation and Compilation Semantics and Immediacy
2646 :    
2647 :     When a word is compiled, it behaves differently from being interpreted.
2648 :     E.g., consider @code{+}:
2649 :    
2650 :     @example
2651 :     1 2 + .
2652 :     : foo + ;
2653 :     @end example
2654 :    
2655 :     These two behaviours are known as compilation and interpretation
2656 :     semantics. For normal words (e.g., @code{+}), the compilation semantics
2657 :     is to append the interpretation semantics to the currently defined word
2658 :     (@code{foo} in the example above). I.e., when @code{foo} is executed
2659 :     later, the interpretation semantics of @code{+} (i.e., adding two
2660 :     numbers) will be performed.
2661 :    
2662 :     However, there are words with non-default compilation semantics, e.g.,
2663 :     the control-flow words like @code{if}. You can use @code{immediate} to
2664 :     change the compilation semantics of the last defined word to be equal to
2665 :     the interpretation semantics:
2666 :    
2667 :     @example
2668 :     : [FOO] ( -- )
2669 :     5 . ; immediate
2670 :    
2671 :     [FOO]
2672 :     : bar ( -- )
2673 :     [FOO] ;
2674 :     bar
2675 :     see bar
2676 :     @end example
2677 :    
2678 :     Two conventions to mark words with non-default compilation semnatics are
2679 :     names with brackets (more frequently used) and to write them all in
2680 :     upper case (less frequently used).
2681 :    
2682 :     In Gforth (and many other systems) you can also remove the
2683 :     interpretation semantics with @code{compile-only} (the compilation
2684 :     semantics is derived from the original interpretation semantics):
2685 :    
2686 :     @example
2687 :     : flip ( -- )
2688 :     6 . ; compile-only \ but not immediate
2689 :     flip
2690 :    
2691 :     : flop ( -- )
2692 :     flip ;
2693 :     flop
2694 :     @end example
2695 :    
2696 :     In this example the interpretation semantics of @code{flop} is equal to
2697 :     the original interpretation semantics of @code{flip}.
2698 :    
2699 :     The text interpreter has two states: in interpret state, it performs the
2700 :     interpretation semantics of words it encounters; in compile state, it
2701 :     performs the compilation semantics of these words.
2702 :    
2703 :     Among other things, @code{:} switches into compile state, and @code{;}
2704 :     switches back to interpret state. They contain the factors @code{]}
2705 :     (switch to compile state) and @code{[} (switch to interpret state), that
2706 :     do nothing but switch the state.
2707 :    
2708 :     @example
2709 :     : xxx ( -- )
2710 :     [ 5 . ]
2711 :     ;
2712 :    
2713 :     xxx
2714 :     see xxx
2715 :     @end example
2716 :    
2717 :     These brackets are also the source of the naming convention mentioned
2718 :     above.
2719 :    
2720 :    
2721 :     @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2722 :     @section Execution Tokens
2723 :    
2724 :     @code{' word} gives you the execution token (XT) of a word. The XT is a
2725 :     cell representing the interpretation semantics of a word. You can
2726 :     execute this semantics with @code{execute}:
2727 :    
2728 :     @example
2729 :     ' + .s
2730 :     1 2 rot execute .
2731 :     @end example
2732 :    
2733 :     The XT is similar to a function pointer in C. However, parameter
2734 :     passing through the stack makes it a little more flexible:
2735 :    
2736 :     @example
2737 :     : map-array ( ... addr u xt -- ... )
2738 : anton 1.50 \ executes xt ( ... x -- ... ) for every element of the array starting
2739 :     \ at addr and containing u elements
2740 : anton 1.48 @{ xt @}
2741 :     cells over + swap ?do
2742 : anton 1.50 i @@ xt execute
2743 : anton 1.48 1 cells +loop ;
2744 :    
2745 :     create a 3 , 4 , 2 , -1 , 4 ,
2746 :     a 5 ' . map-array .s
2747 :     0 a 5 ' + map-array .
2748 :     s" max-n" environment? drop .s
2749 :     a 5 ' min map-array .
2750 :     @end example
2751 :    
2752 :     You can use map-array with the XTs of words that consume one element
2753 :     more than they produce. In theory you can also use it with other XTs,
2754 :     but the stack effect then depends on the size of the array, which is
2755 :     hard to understand.
2756 :    
2757 : pazsan 1.51 Since XTs are cell-sized, you can store them in memory and manipulate
2758 :     them on the stack like other cells. You can also compile the XT into a
2759 : anton 1.48 word with @code{compile,}:
2760 :    
2761 :     @example
2762 :     : foo1 ( n1 n2 -- n )
2763 :     [ ' + compile, ] ;
2764 :     see foo
2765 :     @end example
2766 :    
2767 :     This is non-standard, because @code{compile,} has no compilation
2768 :     semantics in the standard, but it works in good Forth systems. For the
2769 :     broken ones, use
2770 :    
2771 :     @example
2772 :     : [compile,] compile, ; immediate
2773 :    
2774 :     : foo1 ( n1 n2 -- n )
2775 :     [ ' + ] [compile,] ;
2776 :     see foo
2777 :     @end example
2778 :    
2779 :     @code{'} is a word with default compilation semantics; it parses the
2780 :     next word when its interpretation semantics are executed, not during
2781 :     compilation:
2782 :    
2783 :     @example
2784 :     : foo ( -- xt )
2785 :     ' ;
2786 :     see foo
2787 :     : bar ( ... "word" -- ... )
2788 :     ' execute ;
2789 :     see bar
2790 : anton 1.60 1 2 bar + .
2791 : anton 1.48 @end example
2792 :    
2793 :     You often want to parse a word during compilation and compile its XT so
2794 :     it will be pushed on the stack at run-time. @code{[']} does this:
2795 :    
2796 :     @example
2797 :     : xt-+ ( -- xt )
2798 :     ['] + ;
2799 :     see xt-+
2800 :     1 2 xt-+ execute .
2801 :     @end example
2802 :    
2803 :     Many programmers tend to see @code{'} and the word it parses as one
2804 :     unit, and expect it to behave like @code{[']} when compiled, and are
2805 :     confused by the actual behaviour. If you are, just remember that the
2806 :     Forth system just takes @code{'} as one unit and has no idea that it is
2807 :     a parsing word (attempts to convenience programmers in this issue have
2808 :     usually resulted in even worse pitfalls, see
2809 :     @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}).
2810 :    
2811 :     Note that the state of the interpreter does not come into play when
2812 : pazsan 1.51 creating and executing XTs. I.e., even when you execute @code{'} in
2813 : anton 1.48 compile state, it still gives you the interpretation semantics. And
2814 :     whatever that state is, @code{execute} performs the semantics
2815 : pazsan 1.51 represented by the XT (i.e., the interpretation semantics).
2816 : anton 1.48
2817 :    
2818 :     @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2819 :     @section Exceptions
2820 :    
2821 :     @code{throw ( n -- )} causes an exception unless n is zero.
2822 :    
2823 :     @example
2824 :     100 throw .s
2825 :     0 throw .s
2826 :     @end example
2827 :    
2828 :     @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2829 :     it catches exceptions and pushes the number of the exception on the
2830 :     stack (or 0, if the xt executed without exception). If there was an
2831 :     exception, the stacks have the same depth as when entering @code{catch}:
2832 :    
2833 :     @example
2834 :     .s
2835 :     3 0 ' / catch .s
2836 :     3 2 ' / catch .s
2837 :     @end example
2838 :    
2839 :     @assignment
2840 :     Try the same with @code{execute} instead of @code{catch}.
2841 :     @endassignment
2842 :    
2843 :     @code{Throw} always jumps to the dynamically next enclosing
2844 :     @code{catch}, even if it has to leave several call levels to achieve
2845 :     this:
2846 :    
2847 :     @example
2848 :     : foo 100 throw ;
2849 :     : foo1 foo ." after foo" ;
2850 : pazsan 1.51 : bar ['] foo1 catch ;
2851 : anton 1.60 bar .
2852 : anton 1.48 @end example
2853 :    
2854 :     It is often important to restore a value upon leaving a definition, even
2855 :     if the definition is left through an exception. You can ensure this
2856 :     like this:
2857 :    
2858 :     @example
2859 :     : ...
2860 :     save-x
2861 : pazsan 1.51 ['] word-changing-x catch ( ... n )
2862 : anton 1.48 restore-x
2863 :     ( ... n ) throw ;
2864 :     @end example
2865 :    
2866 : anton 1.55 Gforth provides an alternative syntax in addition to @code{catch}:
2867 : anton 1.48 @code{try ... recover ... endtry}. If the code between @code{try} and
2868 :     @code{recover} has an exception, the stack depths are restored, the
2869 :     exception number is pushed on the stack, and the code between
2870 :     @code{recover} and @code{endtry} is performed. E.g., the definition for
2871 :     @code{catch} is
2872 :    
2873 :     @example
2874 :     : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2875 :     try
2876 :     execute 0
2877 :     recover
2878 :     nip
2879 :     endtry ;
2880 :     @end example
2881 :    
2882 :     The equivalent to the restoration code above is
2883 :    
2884 :     @example
2885 :     : ...
2886 :     save-x
2887 :     try
2888 :     word-changing-x
2889 :     end-try
2890 :     restore-x
2891 :     throw ;
2892 :     @end example
2893 :    
2894 :     As you can see, the @code{recover} part is optional.
2895 :    
2896 :    
2897 :     @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2898 :     @section Defining Words
2899 :    
2900 :     @code{:}, @code{create}, and @code{variable} are definition words: They
2901 :     define other words. @code{Constant} is another definition word:
2902 :    
2903 :     @example
2904 :     5 constant foo
2905 :     foo .
2906 :     @end example
2907 :    
2908 :     You can also use the prefixes @code{2} (double-cell) and @code{f}
2909 :     (floating point) with @code{variable} and @code{constant}.
2910 :    
2911 :     You can also define your own defining words. E.g.:
2912 :    
2913 :     @example
2914 :     : variable ( "name" -- )
2915 :     create 0 , ;
2916 :     @end example
2917 :    
2918 :     You can also define defining words that create words that do something
2919 :     other than just producing their address:
2920 :    
2921 :     @example
2922 :     : constant ( n "name" -- )
2923 :     create ,
2924 :     does> ( -- n )
2925 : anton 1.50 ( addr ) @@ ;
2926 : anton 1.48
2927 :     5 constant foo
2928 :     foo .
2929 :     @end example
2930 :    
2931 :     The definition of @code{constant} above ends at the @code{does>}; i.e.,
2932 :     @code{does>} replaces @code{;}, but it also does something else: It
2933 :     changes the last defined word such that it pushes the address of the
2934 :     body of the word and then performs the code after the @code{does>}
2935 :     whenever it is called.
2936 :    
2937 :     In the example above, @code{constant} uses @code{,} to store 5 into the
2938 :     body of @code{foo}. When @code{foo} executes, it pushes the address of
2939 :     the body onto the stack, then (in the code after the @code{does>})
2940 :     fetches the 5 from there.
2941 :    
2942 :     The stack comment near the @code{does>} reflects the stack effect of the
2943 :     defined word, not the stack effect of the code after the @code{does>}
2944 :     (the difference is that the code expects the address of the body that
2945 :     the stack comment does not show).
2946 :    
2947 :     You can use these definition words to do factoring in cases that involve
2948 :     (other) definition words. E.g., a field offset is always added to an
2949 :     address. Instead of defining
2950 :    
2951 :     @example
2952 :     2 cells constant offset-field1
2953 :     @end example
2954 :    
2955 :     and using this like
2956 :    
2957 :     @example
2958 :     ( addr ) offset-field1 +
2959 :     @end example
2960 :    
2961 :     you can define a definition word
2962 :    
2963 :     @example
2964 :     : simple-field ( n "name" -- )
2965 :     create ,
2966 :     does> ( n1 -- n1+n )
2967 : anton 1.50 ( addr ) @@ + ;
2968 : anton 1.48 @end example
2969 : crook 1.21
2970 : anton 1.48 Definition and use of field offsets now look like this:
2971 : crook 1.21
2972 : anton 1.48 @example
2973 :     2 cells simple-field field1
2974 : anton 1.60 create mystruct 4 cells allot
2975 :     mystruct .s field1 .s drop
2976 : anton 1.48 @end example
2977 : crook 1.21
2978 : anton 1.48 If you want to do something with the word without performing the code
2979 :     after the @code{does>}, you can access the body of a @code{create}d word
2980 :     with @code{>body ( xt -- addr )}:
2981 : crook 1.21
2982 : anton 1.48 @example
2983 :     : value ( n "name" -- )
2984 :     create ,
2985 :     does> ( -- n1 )
2986 : anton 1.50 @@ ;
2987 : anton 1.48 : to ( n "name" -- )
2988 :     ' >body ! ;
2989 : crook 1.21
2990 : anton 1.48 5 value foo
2991 :     foo .
2992 :     7 to foo
2993 :     foo .
2994 :     @end example
2995 : crook 1.21
2996 : anton 1.48 @assignment
2997 :     Define @code{defer ( "name" -- )}, which creates a word that stores an
2998 :     XT (at the start the XT of @code{abort}), and upon execution
2999 :     @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3000 :     @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3001 :     recursion is one application of @code{defer}.
3002 :     @endassignment
3003 : crook 1.29
3004 : anton 1.48 @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3005 :     @section Arrays and Records
3006 : crook 1.29
3007 : anton 1.48 Forth has no standard words for defining data structures such as arrays
3008 :     and records (structs in C terminology), but you can build them yourself
3009 :     based on address arithmetic. You can also define words for defining
3010 :     arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3011 : crook 1.29
3012 : anton 1.48 One of the first projects a Forth newcomer sets out upon when learning
3013 :     about defining words is an array defining word (possibly for
3014 :     n-dimensional arrays). Go ahead and do it, I did it, too; you will
3015 :     learn something from it. However, don't be disappointed when you later
3016 :     learn that you have little use for these words (inappropriate use would
3017 :     be even worse). I have not yet found a set of useful array words yet;
3018 :     the needs are just too diverse, and named, global arrays (the result of
3019 :     naive use of defining words) are often not flexible enough (e.g.,
3020 :     consider how to pass them as parameters).
3021 : crook 1.29
3022 : anton 1.48 On the other hand, there is a useful set of record words, and it has
3023 :     been defined in @file{compat/struct.fs}; these words are predefined in
3024 :     Gforth. They are explained in depth elsewhere in this manual (see
3025 :     @pxref{Structures}). The @code{simple-field} example above is
3026 :     simplified variant of fields in this package.
3027 : crook 1.21
3028 :    
3029 : anton 1.48 @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3030 :     @section @code{POSTPONE}
3031 : crook 1.21
3032 : anton 1.48 You can compile the compilation semantics (instead of compiling the
3033 :     interpretation semantics) of a word with @code{POSTPONE}:
3034 : crook 1.21
3035 : anton 1.48 @example
3036 :     : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3037 : pazsan 1.51 POSTPONE + ; immediate
3038 : anton 1.48 : foo ( n1 n2 -- n )
3039 :     MY-+ ;
3040 :     1 2 foo .
3041 :     see foo
3042 :     @end example
3043 : crook 1.21
3044 : anton 1.48 During the definition of @code{foo} the text interpreter performs the
3045 :     compilation semantics of @code{MY-+}, which performs the compilation
3046 :     semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3047 :    
3048 :     This example also displays separate stack comments for the compilation
3049 :     semantics and for the stack effect of the compiled code. For words with
3050 :     default compilation semantics these stack effects are usually not
3051 :     displayed; the stack effect of the compilation semantics is always
3052 :     @code{( -- )} for these words, the stack effect for the compiled code is
3053 :     the stack effect of the interpretation semantics.
3054 :    
3055 :     Note that the state of the interpreter does not come into play when
3056 :     performing the compilation semantics in this way. You can also perform
3057 :     it interpretively, e.g.:
3058 :    
3059 :     @example
3060 :     : foo2 ( n1 n2 -- n )
3061 :     [ MY-+ ] ;
3062 :     1 2 foo .
3063 :     see foo
3064 :     @end example
3065 : crook 1.21
3066 : anton 1.48 However, there are some broken Forth systems where this does not always
3067 : crook 1.62 work, and therefore this practice was been declared non-standard in
3068 : anton 1.48 1999.
3069 :     @c !! repair.fs
3070 :    
3071 :     Here is another example for using @code{POSTPONE}:
3072 : crook 1.44
3073 : anton 1.48 @example
3074 :     : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3075 :     POSTPONE negate POSTPONE + ; immediate compile-only
3076 :     : bar ( n1 n2 -- n )
3077 :     MY-- ;
3078 :     2 1 bar .
3079 :     see bar
3080 :     @end example
3081 : crook 1.21
3082 : anton 1.48 You can define @code{ENDIF} in this way:
3083 : crook 1.21
3084 : anton 1.48 @example
3085 :     : ENDIF ( Compilation: orig -- )
3086 :     POSTPONE then ; immediate
3087 :     @end example
3088 : crook 1.21
3089 : anton 1.48 @assignment
3090 :     Write @code{MY-2DUP} that has compilation semantics equivalent to
3091 :     @code{2dup}, but compiles @code{over over}.
3092 :     @endassignment
3093 : crook 1.29
3094 : anton 1.48 @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3095 :     @section @code{Literal}
3096 : crook 1.29
3097 : anton 1.48 You cannot @code{POSTPONE} numbers:
3098 : crook 1.21
3099 : anton 1.48 @example
3100 :     : [FOO] POSTPONE 500 ; immediate
3101 : crook 1.21 @end example
3102 :    
3103 : anton 1.48 Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3104 : crook 1.29
3105 : anton 1.48 @example
3106 :     : [FOO] ( compilation: --; run-time: -- n )
3107 :     500 POSTPONE literal ; immediate
3108 : crook 1.29
3109 : anton 1.60 : flip [FOO] ;
3110 : anton 1.48 flip .
3111 :     see flip
3112 :     @end example
3113 : crook 1.29
3114 : anton 1.48 @code{LITERAL} consumes a number at compile-time (when it's compilation
3115 :     semantics are executed) and pushes it at run-time (when the code it
3116 :     compiled is executed). A frequent use of @code{LITERAL} is to compile a
3117 :     number computed at compile time into the current word:
3118 : crook 1.29
3119 : anton 1.48 @example
3120 :     : bar ( -- n )
3121 :     [ 2 2 + ] literal ;
3122 :     see bar
3123 :     @end example
3124 : crook 1.29
3125 : anton 1.48 @assignment
3126 :     Write @code{]L} which allows writing the example above as @code{: bar (
3127 :     -- n ) [ 2 2 + ]L ;}
3128 :     @endassignment
3129 :    
3130 :    
3131 :     @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3132 :     @section Advanced macros
3133 :    
3134 :     Reconsider @code{map-array} from @ref{Execution Tokens
3135 :     Tutorial,, Execution Tokens}. It frequently performs @code{execute}, a
3136 :     relatively expensive operation in some implementations. You can use
3137 :     @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3138 :     and produce a word that contains the word to be performed directly:
3139 :    
3140 :     @c use ]] ... [[
3141 :     @example
3142 :     : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3143 :     \ at run-time, execute xt ( ... x -- ... ) for each element of the
3144 :     \ array beginning at addr and containing u elements
3145 :     @{ xt @}
3146 :     POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3147 : anton 1.50 POSTPONE i POSTPONE @@ xt compile,
3148 : anton 1.48 1 cells POSTPONE literal POSTPONE +loop ;
3149 :    
3150 :     : sum-array ( addr u -- n )
3151 :     0 rot rot [ ' + compile-map-array ] ;
3152 :     see sum-array
3153 :     a 5 sum-array .
3154 :     @end example
3155 :    
3156 :     You can use the full power of Forth for generating the code; here's an
3157 :     example where the code is generated in a loop:
3158 :    
3159 :     @example
3160 :     : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3161 :     \ n2=n1+(addr1)*n, addr2=addr1+cell
3162 : anton 1.50 POSTPONE tuck POSTPONE @@
3163 : anton 1.48 POSTPONE literal POSTPONE * POSTPONE +
3164 :     POSTPONE swap POSTPONE cell+ ;
3165 :    
3166 :     : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3167 : pazsan 1.51 \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3168 : anton 1.48 0 postpone literal postpone swap
3169 :     [ ' compile-vmul-step compile-map-array ]
3170 :     postpone drop ;
3171 :     see compile-vmul
3172 :    
3173 :     : a-vmul ( addr -- n )
3174 : pazsan 1.51 \ n=a*v, where v is a vector that's as long as a and starts at addr
3175 : anton 1.48 [ a 5 compile-vmul ] ;
3176 :     see a-vmul
3177 :     a a-vmul .
3178 :     @end example
3179 :    
3180 :     This example uses @code{compile-map-array} to show off, but you could
3181 :     also use @code{map-array} instead (try it now).
3182 :    
3183 :     You can use this technique for efficient multiplication of large
3184 :     matrices. In matrix multiplication, you multiply every line of one
3185 :     matrix with every column of the other matrix. You can generate the code
3186 :     for one line once, and use it for every column. The only downside of
3187 :     this technique is that it is cumbersome to recover the memory consumed
3188 :     by the generated code when you are done (and in more complicated cases
3189 :     it is not possible portably).
3190 :    
3191 :     @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3192 :     @section Compilation Tokens
3193 :    
3194 :     This section is Gforth-specific. You can skip it.
3195 :    
3196 :     @code{' word compile,} compiles the interpretation semantics. For words
3197 :     with default compilation semantics this is the same as performing the
3198 :     compilation semantics. To represent the compilation semantics of other
3199 :     words (e.g., words like @code{if} that have no interpretation
3200 :     semantics), Gforth has the concept of a compilation token (CT,
3201 :     consisting of two cells), and words @code{comp'} and @code{[comp']}.
3202 :     You can perform the compilation semantics represented by a CT with
3203 :     @code{execute}:
3204 : crook 1.29
3205 : anton 1.48 @example
3206 :     : foo2 ( n1 n2 -- n )
3207 :     [ comp' + execute ] ;
3208 :     see foo
3209 :     @end example
3210 : crook 1.29
3211 : anton 1.48 You can compile the compilation semantics represented by a CT with
3212 :     @code{postpone,}:
3213 : anton 1.30
3214 : anton 1.48 @example
3215 :     : foo3 ( -- )
3216 :     [ comp' + postpone, ] ;
3217 :     see foo3
3218 :     @end example
3219 : anton 1.30
3220 : pazsan 1.51 @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3221 : anton 1.48 @code{comp'} is particularly useful for words that have no
3222 :     interpretation semantics:
3223 : crook 1.29
3224 : anton 1.30 @example
3225 : anton 1.48 ' if
3226 : anton 1.60 comp' if .s 2drop
3227 : anton 1.30 @end example
3228 :    
3229 : crook 1.29
3230 : anton 1.48 @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3231 :     @section Wordlists and Search Order
3232 :    
3233 :     The dictionary is not just a memory area that allows you to allocate
3234 :     memory with @code{allot}, it also contains the Forth words, arranged in
3235 :     several wordlists. When searching for a word in a wordlist,
3236 :     conceptually you start searching at the youngest and proceed towards
3237 :     older words (in reality most systems nowadays use hash-tables); i.e., if
3238 :     you define a word with the same name as an older word, the new word
3239 :     shadows the older word.
3240 :    
3241 :     Which wordlists are searched in which order is determined by the search
3242 :     order. You can display the search order with @code{order}. It displays
3243 :     first the search order, starting with the wordlist searched first, then
3244 :     it displays the wordlist that will contain newly defined words.
3245 : crook 1.21
3246 : anton 1.48 You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3247 : crook 1.21
3248 : anton 1.48 @example
3249 :     wordlist constant mywords
3250 :     @end example
3251 : crook 1.21
3252 : anton 1.48 @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3253 :     defined words (the @emph{current} wordlist):
3254 : crook 1.21
3255 : anton 1.48 @example
3256 :     mywords set-current
3257 :     order
3258 :     @end example
3259 : crook 1.26
3260 : anton 1.48 Gforth does not display a name for the wordlist in @code{mywords}
3261 :     because this wordlist was created anonymously with @code{wordlist}.
3262 : crook 1.21
3263 : anton 1.48 You can get the current wordlist with @code{get-current ( -- wid)}. If
3264 :     you want to put something into a specific wordlist without overall
3265 :     effect on the current wordlist, this typically looks like this:
3266 : crook 1.21
3267 : anton 1.48 @example
3268 :     get-current mywords set-current ( wid )
3269 :     create someword
3270 :     ( wid ) set-current
3271 :     @end example
3272 : crook 1.21
3273 : anton 1.48 You can write the search order with @code{set-order ( wid1 .. widn n --
3274 :     )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3275 :     searched wordlist is topmost.
3276 : crook 1.21
3277 : anton 1.48 @example
3278 :     get-order mywords swap 1+ set-order
3279 :     order
3280 :     @end example
3281 : crook 1.21
3282 : anton 1.48 Yes, the order of wordlists in the output of @code{order} is reversed
3283 :     from stack comments and the output of @code{.s} and thus unintuitive.
3284 : crook 1.21
3285 : anton 1.48 @assignment
3286 :     Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3287 :     wordlist to the search order. Define @code{previous ( -- )}, which
3288 :     removes the first searched wordlist from the search order. Experiment
3289 :     with boundary conditions (you will see some crashes or situations that
3290 :     are hard or impossible to leave).
3291 :     @endassignment
3292 : crook 1.21
3293 : anton 1.48 The search order is a powerful foundation for providing features similar
3294 :     to Modula-2 modules and C++ namespaces. However, trying to modularize
3295 :     programs in this way has disadvantages for debugging and reuse/factoring
3296 :     that overcome the advantages in my experience (I don't do huge projects,
3297 : anton 1.55 though). These disadvantages are not so clear in other
3298 : anton 1.48 languages/programming environments, because these langauges are not so
3299 :     strong in debugging and reuse.
3300 : crook 1.21
3301 :    
3302 : crook 1.29 @c ******************************************************************
3303 : anton 1.48 @node Introduction, Words, Tutorial, Top
3304 : crook 1.29 @comment node-name, next, previous, up
3305 :     @chapter An Introduction to ANS Forth
3306 :     @cindex Forth - an introduction
3307 : crook 1.21
3308 : crook 1.29 The primary purpose of this manual is to document Gforth. However, since
3309 :     Forth is not a widely-known language and there is a lack of up-to-date
3310 :     teaching material, it seems worthwhile to provide some introductory
3311 : anton 1.49 material. For other sources of Forth-related
3312 :     information, see @ref{Forth-related information}.
3313 : crook 1.21
3314 : crook 1.29 The examples in this section should work on any ANS Forth; the
3315 :     output shown was produced using Gforth. Each example attempts to
3316 :     reproduce the exact output that Gforth produces. If you try out the
3317 :     examples (and you should), what you should type is shown @kbd{like this}
3318 :     and Gforth's response is shown @code{like this}. The single exception is
3319 : anton 1.30 that, where the example shows @key{RET} it means that you should
3320 : crook 1.29 press the ``carriage return'' key. Unfortunately, some output formats for
3321 :     this manual cannot show the difference between @kbd{this} and
3322 :     @code{this} which will make trying out the examples harder (but not
3323 :     impossible).
3324 : crook 1.21
3325 : crook 1.29 Forth is an unusual language. It provides an interactive development
3326 :     environment which includes both an interpreter and compiler. Forth
3327 :     programming style encourages you to break a problem down into many
3328 :     @cindex factoring
3329 :     small fragments (@dfn{factoring}), and then to develop and test each
3330 :     fragment interactively. Forth advocates assert that breaking the
3331 :     edit-compile-test cycle used by conventional programming languages can
3332 :     lead to great productivity improvements.
3333 : crook 1.21
3334 : crook 1.29 @menu
3335 :     * Introducing the Text Interpreter::
3336 :     * Stacks and Postfix notation::
3337 :     * Your first definition::
3338 :     * How does that work?::
3339 :     * Forth is written in Forth::
3340 :     * Review - elements of a Forth system::
3341 :     * Where to go next::
3342 :     * Exercises::
3343 :     @end menu
3344 : crook 1.21
3345 : crook 1.29 @comment ----------------------------------------------
3346 :     @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3347 :     @section Introducing the Text Interpreter
3348 :     @cindex text interpreter
3349 :     @cindex outer interpreter
3350 : crook 1.21
3351 : anton 1.30 @c IMO this is too detailed and the pace is too slow for
3352 :     @c an introduction. If you know German, take a look at
3353 :     @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3354 :     @c to see how I do it - anton
3355 :    
3356 : crook 1.44 @c nac-> Where I have accepted your comments 100% and modified the text
3357 :     @c accordingly, I have deleted your comments. Elsewhere I have added a
3358 :     @c response like this to attempt to rationalise what I have done. Of
3359 :     @c course, this is a very clumsy mechanism for something that would be
3360 :     @c done far more efficiently over a beer. Please delete any dialogue
3361 :     @c you consider closed.
3362 :    
3363 : crook 1.29 When you invoke the Forth image, you will see a startup banner printed
3364 :     and nothing else (if you have Gforth installed on your system, try
3365 : anton 1.30 invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
3366 : crook 1.29 its command line interpreter, which is called the @dfn{Text Interpreter}
3367 :     (also known as the @dfn{Outer Interpreter}). (You will learn a lot
3368 : anton 1.49 about the text interpreter as you read through this chapter, for more
3369 :     detail @pxref{The Text Interpreter}).
3370 : crook 1.21
3371 : crook 1.29 Although it's not obvious, Forth is actually waiting for your
3372 : anton 1.30 input. Type a number and press the @key{RET} key:
3373 : crook 1.21
3374 : crook 1.26 @example
3375 : anton 1.30 @kbd{45@key{RET}} ok
3376 : crook 1.26 @end example
3377 : crook 1.21
3378 : crook 1.29 Rather than give you a prompt to invite you to input something, the text
3379 :     interpreter prints a status message @i{after} it has processed a line
3380 :     of input. The status message in this case (``@code{ ok}'' followed by
3381 :     carriage-return) indicates that the text interpreter was able to process
3382 :     all of your input successfully. Now type something illegal:
3383 :    
3384 :     @example
3385 : anton 1.30 @kbd{qwer341@key{RET}}
3386 : crook 1.29 :1: Undefined word
3387 :     qwer341
3388 :     ^^^^^^^
3389 :     $400D2BA8 Bounce
3390 :     $400DBDA8 no.extensions
3391 :     @end example
3392 : crook 1.23
3393 : crook 1.29 The exact text, other than the ``Undefined word'' may differ slightly on
3394 :     your system, but the effect is the same; when the text interpreter
3395 :     detects an error, it discards any remaining text on a line, resets
3396 : anton 1.49 certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3397 :     messages}.
3398 : crook 1.23
3399 : crook 1.29 The text interpreter waits for you to press carriage-return, and then
3400 :     processes your input line. Starting at the beginning of the line, it
3401 :     breaks the line into groups of characters separated by spaces. For each
3402 :     group of characters in turn, it makes two attempts to do something:
3403 : crook 1.23
3404 : crook 1.29 @itemize @bullet
3405 :     @item
3406 : crook 1.44 @cindex name dictionary
3407 : crook 1.29 It tries to treat it as a command. It does this by searching a @dfn{name
3408 :     dictionary}. If the group of characters matches an entry in the name
3409 :     dictionary, the name dictionary provides the text interpreter with
3410 :     information that allows the text interpreter perform some actions. In
3411 :     Forth jargon, we say that the group
3412 :     @cindex word
3413 :     @cindex definition
3414 :     @cindex execution token
3415 :     @cindex xt
3416 :     of characters names a @dfn{word}, that the dictionary search returns an
3417 :     @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3418 :     word, and that the text interpreter executes the xt. Often, the terms
3419 :     @dfn{word} and @dfn{definition} are used interchangeably.
3420 :     @item
3421 :     If the text interpreter fails to find a match in the name dictionary, it
3422 :     tries to treat the group of characters as a number in the current number
3423 :     base (when you start up Forth, the current number base is base 10). If
3424 :     the group of characters legitimately represents a number, the text
3425 :     interpreter pushes the number onto a stack (we'll learn more about that
3426 :     in the next section).
3427 :     @end itemize
3428 : crook 1.23
3429 : crook 1.29 If the text interpreter is unable to do either of these things with any
3430 :     group of characters, it discards the group of characters and the rest of
3431 :     the line, then prints an error message. If the text interpreter reaches
3432 :     the end of the line without error, it prints the status message ``@code{ ok}''
3433 :     followed by carriage-return.
3434 : crook 1.21
3435 : crook 1.29 This is the simplest command we can give to the text interpreter:
3436 : crook 1.23
3437 :     @example
3438 : anton 1.30 @key{RET} ok
3439 : crook 1.23 @end example
3440 : crook 1.21
3441 : crook 1.29 The text interpreter did everything we asked it to do (nothing) without
3442 :     an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3443 :     command:
3444 : crook 1.21
3445 : crook 1.23 @example
3446 : anton 1.30 @kbd{12 dup fred dup@key{RET}}
3447 : crook 1.29 :1: Undefined word
3448 :     12 dup fred dup
3449 :     ^^^^
3450 :     $400D2BA8 Bounce
3451 :     $400DBDA8 no.extensions
3452 : crook 1.23 @end example
3453 : crook 1.21
3454 : crook 1.29 When you press the carriage-return key, the text interpreter starts to
3455 :     work its way along the line:
3456 : crook 1.21
3457 : crook 1.29 @itemize @bullet
3458 :     @item
3459 :     When it gets to the space after the @code{2}, it takes the group of
3460 :     characters @code{12} and looks them up in the name
3461 :     dictionary@footnote{We can't tell if it found them or not, but assume
3462 :     for now that it did not}. There is no match for this group of characters
3463 :     in the name dictionary, so it tries to treat them as a number. It is
3464 :     able to do this successfully, so it puts the number, 12, ``on the stack''
3465 :     (whatever that means).
3466 :     @item
3467 :     The text interpreter resumes scanning the line and gets the next group
3468 :     of characters, @code{dup}. It looks it up in the name dictionary and
3469 :     (you'll have to take my word for this) finds it, and executes the word
3470 :     @code{dup} (whatever that means).
3471 :     @item
3472 :     Once again, the text interpreter resumes scanning the line and gets the
3473 :     group of characters @code{fred}. It looks them up in the name
3474 :     dictionary, but can't find them. It tries to treat them as a number, but
3475 :     they don't represent any legal number.
3476 :     @end itemize
3477 : crook 1.21
3478 : crook 1.29 At this point, the text interpreter gives up and prints an error
3479 :     message. The error message shows exactly how far the text interpreter
3480 :     got in processing the line. In particular, it shows that the text
3481 :     interpreter made no attempt to do anything with the final character
3482 :     group, @code{dup}, even though we have good reason to believe that the
3483 :     text interpreter would have no problem looking that word up and
3484 :     executing it a second time.
3485 : crook 1.21
3486 :    
3487 : crook 1.29 @comment ----------------------------------------------
3488 :     @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3489 :     @section Stacks, postfix notation and parameter passing
3490 :     @cindex text interpreter
3491 :     @cindex outer interpreter
3492 : crook 1.21
3493 : crook 1.29 In procedural programming languages (like C and Pascal), the
3494 :     building-block of programs is the @dfn{function} or @dfn{procedure}. These
3495 :     functions or procedures are called with @dfn{explicit parameters}. For
3496 :     example, in C we might write:
3497 : crook 1.21
3498 : crook 1.23 @example
3499 : crook 1.29 total = total + new_volume(length,height,depth);
3500 : crook 1.23 @end example
3501 : crook 1.21
3502 : crook 1.23 @noindent
3503 : crook 1.29 where new_volume is a function-call to another piece of code, and total,
3504 :     length, height and depth are all variables. length, height and depth are
3505 :     parameters to the function-call.
3506 : crook 1.21
3507 : crook 1.29 In Forth, the equivalent of the function or procedure is the
3508 :     @dfn{definition} and parameters are implicitly passed between
3509 :     definitions using a shared stack that is visible to the
3510 :     programmer. Although Forth does support variables, the existence of the
3511 :     stack means that they are used far less often than in most other
3512 :     programming languages. When the text interpreter encounters a number, it
3513 :     will place (@dfn{push}) it on the stack. There are several stacks (the
3514 : anton 1.30 actual number is implementation-dependent ...) and the particular stack
3515 : crook 1.29 used for any operation is implied unambiguously by the operation being
3516 :     performed. The stack used for all integer operations is called the @dfn{data
3517 :     stack} and, since this is the stack used most commonly, references to
3518 :     ``the data stack'' are often abbreviated to ``the stack''.
3519 : crook 1.21
3520 : crook 1.29 The stacks have a last-in, first-out (LIFO) organisation. If you type:
3521 : crook 1.21
3522 : crook 1.23 @example
3523 : anton 1.30 @kbd{1 2 3@key{RET}} ok
3524 : crook 1.23 @end example
3525 : crook 1.21
3526 : crook 1.29 Then this instructs the text interpreter to placed three numbers on the
3527 :     (data) stack. An analogy for the behaviour of the stack is to take a
3528 :     pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3529 :     the table. The 3 was the last card onto the pile (``last-in'') and if
3530 :     you take a card off the pile then, unless you're prepared to fiddle a
3531 :     bit, the card that you take off will be the 3 (``first-out''). The
3532 :     number that will be first-out of the stack is called the @dfn{top of
3533 :     stack}, which
3534 :     @cindex TOS definition
3535 :     is often abbreviated to @dfn{TOS}.
3536 : crook 1.21
3537 : crook 1.29 To understand how parameters are passed in Forth, consider the
3538 :     behaviour of the definition @code{+} (pronounced ``plus''). You will not
3539 :     be surprised to learn that this definition performs addition. More
3540 :     precisely, it adds two number together and produces a result. Where does
3541 :     it get the two numbers from? It takes the top two numbers off the
3542 :     stack. Where does it place the result? On the stack. You can act-out the
3543 :     behaviour of @code{+} with your playing cards like this:
3544 : crook 1.21
3545 :     @itemize @bullet
3546 :     @item
3547 : crook 1.29 Pick up two cards from the stack on the table
3548 : crook 1.21 @item
3549 : crook 1.29 Stare at them intently and ask yourself ``what @i{is} the sum of these two
3550 :     numbers''
3551 : crook 1.21 @item
3552 : crook 1.29 Decide that the answer is 5
3553 : crook 1.21 @item
3554 : crook 1.29 Shuffle the two cards back into the pack and find a 5
3555 : crook 1.21 @item
3556 : crook 1.29 Put a 5 on the remaining ace that's on the table.
3557 : crook 1.21 @end itemize
3558 :    
3559 : crook 1.29 If you don't have a pack of cards handy but you do have Forth running,
3560 :     you can use the definition @code{.s} to show the current state of the stack,
3561 :     without affecting the stack. Type:
3562 : crook 1.21
3563 :     @example
3564 : anton 1.30 @kbd{clearstack 1 2 3@key{RET}} ok
3565 :     @kbd{.s@key{RET}} <3> 1 2 3 ok
3566 : crook 1.23 @end example
3567 :    
3568 : crook 1.29 The text interpreter looks up the word @code{clearstack} and executes
3569 :     it; it tidies up the stack and removes any entries that may have been
3570 :     left on it by earlier examples. The text interpreter pushes each of the
3571 :     three numbers in turn onto the stack. Finally, the text interpreter
3572 :     looks up the word @code{.s} and executes it. The effect of executing
3573 :     @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3574 :     followed by a list of all the items on the stack; the item on the far
3575 :     right-hand side is the TOS.
3576 : crook 1.21
3577 : crook 1.29 You can now type:
3578 : crook 1.21
3579 : crook 1.29 @example
3580 : anton 1.30 @kbd{+ .s@key{RET}} <2> 1 5 ok
3581 : crook 1.29 @end example
3582 : crook 1.21
3583 : crook 1.29 @noindent
3584 :     which is correct; there are now 2 items on the stack and the result of
3585 :     the addition is 5.
3586 : crook 1.23
3587 : crook 1.29 If you're playing with cards, try doing a second addition: pick up the
3588 :     two cards, work out that their sum is 6, shuffle them into the pack,
3589 :     look for a 6 and place that on the table. You now have just one item on
3590 :     the stack. What happens if you try to do a third addition? Pick up the
3591 :     first card, pick up the second card -- ah! There is no second card. This
3592 :     is called a @dfn{stack underflow} and consitutes an error. If you try to
3593 :     do the same thing with Forth it will report an error (probably a Stack
3594 :     Underflow or an Invalid Memory Address error).
3595 : crook 1.23
3596 : crook 1.29 The opposite situation to a stack underflow is a @dfn{stack overflow},
3597 :     which simply accepts that there is a finite amount of storage space
3598 :     reserved for the stack. To stretch the playing card analogy, if you had
3599 :     enough packs of cards and you piled the cards up on the table, you would
3600 :     eventually be unable to add another card; you'd hit the ceiling. Gforth
3601 :     allows you to set the maximum size of the stacks. In general, the only
3602 :     time that you will get a stack overflow is because a definition has a
3603 :     bug in it and is generating data on the stack uncontrollably.
3604 : crook 1.23
3605 : crook 1.29 There's one final use for the playing card analogy. If you model your
3606 :     stack using a pack of playing cards, the maximum number of items on
3607 :     your stack will be 52 (I assume you didn't use the Joker). The maximum
3608 :     @i{value} of any item on the stack is 13 (the King). In fact, the only
3609 :     possible numbers are positive integer numbers 1 through 13; you can't
3610 :     have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3611 :     think about some of the cards, you can accommodate different
3612 :     numbers. For example, you could think of the Jack as representing 0,
3613 :     the Queen as representing -1 and the King as representing -2. Your
3614 : crook 1.45 @i{range} remains unchanged (you can still only represent a total of 13
3615 : crook 1.29 numbers) but the numbers that you can represent are -2 through 10.
3616 : crook 1.28
3617 : crook 1.29 In that analogy, the limit was the amount of information that a single
3618 :     stack entry could hold, and Forth has a similar limit. In Forth, the
3619 :     size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3620 :     implementation dependent and affects the maximum value that a stack
3621 :     entry can hold. A Standard Forth provides a cell size of at least
3622 :     16-bits, and most desktop systems use a cell size of 32-bits.
3623 : crook 1.21
3624 : crook 1.29 Forth does not do any type checking for you, so you are free to
3625 :     manipulate and combine stack items in any way you wish. A convenient way
3626 :     of treating stack items is as 2's complement signed integers, and that
3627 :     is what Standard words like @code{+} do. Therefore you can type:
3628 : crook 1.21
3629 : crook 1.29 @example
3630 : anton 1.30 @kbd{-5 12 + .s@key{RET}} <1> 7 ok
3631 : crook 1.29 @end example
3632 : crook 1.21
3633 : crook 1.29 If you use numbers and definitions like @code{+} in order to turn Forth
3634 :     into a great big pocket calculator, you will realise that it's rather
3635 :     different from a normal calculator. Rather than typing 2 + 3 = you had
3636 :     to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3637 :     result). The terminology used to describe this difference is to say that
3638 :     your calculator uses @dfn{Infix Notation} (parameters and operators are
3639 :     mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3640 :     operators are separate), also called @dfn{Reverse Polish Notation}.
3641 : crook 1.21
3642 : crook 1.29 Whilst postfix notation might look confusing to begin with, it has
3643 :     several important advantages:
3644 : crook 1.21
3645 : crook 1.23 @itemize @bullet
3646 :     @item
3647 : crook 1.29 it is unambiguous
3648 : crook 1.23 @item
3649 : crook 1.29 it is more concise
3650 : crook 1.23 @item
3651 : crook 1.29 it fits naturally with a stack-based system
3652 : crook 1.23 @end itemize
3653 : crook 1.21
3654 : crook 1.29 To examine these claims in more detail, consider these sums:
3655 : crook 1.21
3656 : crook 1.29 @example
3657 :     6 + 5 * 4 =
3658 :     4 * 5 + 6 =
3659 :     @end example
3660 : crook 1.21
3661 : crook 1.29 If you're just learning maths or your maths is very rusty, you will
3662 :     probably come up with the answer 44 for the first and 26 for the
3663 :     second. If you are a bit of a whizz at maths you will remember the
3664 :     @i{convention} that multiplication takes precendence over addition, and
3665 :     you'd come up with the answer 26 both times. To explain the answer 26
3666 :     to someone who got the answer 44, you'd probably rewrite the first sum
3667 :     like this:
3668 : crook 1.21
3669 : crook 1.29 @example
3670 :     6 + (5 * 4) =
3671 :     @end example
3672 : crook 1.21
3673 : crook 1.29 If what you really wanted was to perform the addition before the
3674 :     multiplication, you would have to use parentheses to force it.
3675 : crook 1.21
3676 : crook 1.29 If you did the first two sums on a pocket calculator you would probably
3677 :     get the right answers, unless you were very cautious and entered them using
3678 :     these keystroke sequences:
3679 : crook 1.21
3680 : crook 1.29 6 + 5 = * 4 =
3681 :     4 * 5 = + 6 =
3682 : crook 1.21
3683 : crook 1.29 Postfix notation is unambiguous because the order that the operators
3684 :     are applied is always explicit; that also means that parentheses are
3685 :     never required. The operators are @i{active} (the act of quoting the
3686 :     operator makes the operation occur) which removes the need for ``=''.
3687 : crook 1.28
3688 : crook 1.29 The sum 6 + 5 * 4 can be written (in postfix notation) in two
3689 :     equivalent ways:
3690 : crook 1.26
3691 :     @example
3692 : crook 1.29 6 5 4 * + or:
3693 :     5 4 * 6 +
3694 : crook 1.26 @end example
3695 : crook 1.23
3696 : crook 1.29 An important thing that you should notice about this notation is that
3697 :     the @i{order} of the numbers does not change; if you want to subtract
3698 :     2 from 10 you type @code{10 2 -}.
3699 : anton 1.1
3700 : crook 1.29 The reason that Forth uses postfix notation is very simple to explain: it
3701 :     makes the implementation extremely simple, and it follows naturally from
3702 :     using the stack as a mechanism for passing parameters. Another way of
3703 :     thinking about this is to realise that all Forth definitions are
3704 :     @i{active}; they execute as they are encountered by the text
3705 :     interpreter. The result of this is that the syntax of Forth is trivially
3706 :     simple.
3707 : anton 1.1
3708 :    
3709 :    
3710 : crook 1.29 @comment ----------------------------------------------
3711 :     @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3712 :     @section Your first Forth definition
3713 :     @cindex first definition
3714 : anton 1.1
3715 : crook 1.29 Until now, the examples we've seen have been trivial; we've just been
3716 :     using Forth as a bigger-than-pocket calculator. Also, each calculation
3717 :     we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3718 :     again@footnote{That's not quite true. If you press the up-arrow key on
3719 :     your keyboard you should be able to scroll back to any earlier command,
3720 :     edit it and re-enter it.} In this section we'll see how to add new
3721 :     words to Forth's vocabulary.
3722 : anton 1.1
3723 : crook 1.29 The easiest way to create a new word is to use a @dfn{colon
3724 :     definition}. We'll define a few and try them out before worrying too
3725 :     much about how they work. Try typing in these examples; be careful to
3726 :     copy the spaces accurately:
3727 : anton 1.1
3728 : crook 1.29 @example
3729 :     : add-two 2 + . ;
3730 :     : greet ." Hello and welcome" ;
3731 :     : demo 5 add-two ;
3732 :     @end example
3733 : anton 1.1
3734 : crook 1.29 @noindent
3735 :     Now try them out:
3736 : anton 1.1
3737 : crook 1.29 @example
3738 : anton 1.30 @kbd{greet@key{RET}} Hello and welcome ok
3739 :     @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3740 :     @kbd{4 add-two@key{RET}} 6 ok
3741 :     @kbd{demo@key{RET}} 7 ok
3742 :     @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
3743 : crook 1.29 @end example
3744 : anton 1.1
3745 : crook 1.29 The first new thing that we've introduced here is the pair of words
3746 :     @code{:} and @code{;}. These are used to start and terminate a new
3747 :     definition, respectively. The first word after the @code{:} is the name
3748 :     for the new definition.
3749 : anton 1.1
3750 : crook 1.29 As you can see from the examples, a definition is built up of words that
3751 :     have already been defined; Forth makes no distinction between
3752 :     definitions that existed when you started the system up, and those that
3753 :     you define yourself.
3754 : anton 1.1
3755 : crook 1.29 The examples also introduce the words @code{.} (dot), @code{."}
3756 :     (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3757 :     the stack and displays it. It's like @code{.s} except that it only
3758 :     displays the top item of the stack and it is destructive; after it has
3759 :     executed, the number is no longer on the stack. There is always one
3760 :     space printed after the number, and no spaces before it. Dot-quote
3761 :     defines a string (a sequence of characters) that will be printed when
3762 :     the word is executed. The string can contain any printable characters
3763 :     except @code{"}. A @code{"} has a special function; it is not a Forth
3764 :     word but it acts as a delimiter (the way that delimiters work is
3765 :     described in the next section). Finally, @code{dup} duplicates the value
3766 :     at the top of the stack. Try typing @code{5 dup .s} to see what it does.
3767 : anton 1.1
3768 : crook 1.29 We already know that the text interpreter searches through the
3769 :     dictionary to locate names. If you've followed the examples earlier, you
3770 :     will already have a definition called @code{add-two}. Lets try modifying
3771 :     it by typing in a new definition:
3772 : anton 1.1
3773 : crook 1.29 @example
3774 : anton 1.30 @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
3775 : crook 1.29 @end example
3776 : anton 1.5
3777 : crook 1.29 Forth recognised that we were defining a word that already exists, and
3778 :     printed a message to warn us of that fact. Let's try out the new
3779 :     definition:
3780 : anton 1.5
3781 : crook 1.29 @example
3782 : anton 1.30 @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
3783 : crook 1.29 @end example
3784 : anton 1.1
3785 : crook 1.29 @noindent
3786 :     All that we've actually done here, though, is to create a new
3787 :     definition, with a particular name. The fact that there was already a
3788 :     definition with the same name did not make any difference to the way
3789 :     that the new definition was created (except that Forth printed a warning
3790 :     message). The old definition of add-two still exists (try @code{demo}
3791 :     again to see that this is true). Any new definition will use the new
3792 :     definition of @code{add-two}, but old definitions continue to use the
3793 :     version that already existed at the time that they were @code{compiled}.
3794 : anton 1.1
3795 : crook 1.29 Before you go on to the next section, try defining and redefining some
3796 :     words of your own.
3797 : anton 1.1
3798 : crook 1.29 @comment ----------------------------------------------
3799 :     @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3800 :     @section How does that work?
3801 :     @cindex parsing words
3802 : anton 1.1
3803 : anton 1.30 @c That's pretty deep (IMO way too deep) for an introduction. - anton
3804 :    
3805 :     @c Is it a good idea to talk about the interpretation semantics of a
3806 :     @c number? We don't have an xt to go along with it. - anton
3807 :    
3808 :     @c Now that I have eliminated execution semantics, I wonder if it would not
3809 :     @c be better to keep them (or add run-time semantics), to make it easier to
3810 :     @c explain what compilation semantics usually does. - anton
3811 :    
3812 : crook 1.44 @c nac-> I removed the term ``default compilation sematics'' from the
3813 :     @c introductory chapter. Removing ``execution semantics'' was making
3814 :     @c everything simpler to explain, then I think the use of this term made
3815 :     @c everything more complex again. I replaced it with ``default
3816 :     @c semantics'' (which is used elsewhere in the manual) by which I mean
3817 :     @c ``a definition that has neither the immediate nor the compile-only
3818 :     @c flag set''. I reworded big chunks of the ``how does that work''
3819 :     @c section (and, unusually for me, I think I even made it shorter!). See
3820 :     @c what you think -- I know I have not addressed your primary concern
3821 :     @c that it is too heavy-going for an introduction. From what I understood
3822 :     @c of your course notes it looks as though they might be a good framework.
3823 :     @c Things that I've tried to capture here are some things that came as a
3824 :     @c great revelation here when I first understood them. Also, I like the
3825 :     @c fact that a very simple code example shows up almost all of the issues
3826 :     @c that you need to understand to see how Forth works. That's unique and
3827 :     @c worthwhile to emphasise.
3828 :    
3829 : crook 1.29 Now we're going to take another look at the definition of @code{add-two}
3830 :     from the previous section. From our knowledge of the way that the text
3831 :     interpreter works, we would have expected this result when we tried to
3832 :     define @code{add-two}:
3833 : crook 1.21
3834 : crook 1.29 @example
3835 : crook 1.44 @kbd{: add-two 2 + . ;@key{RET}}
3836 : crook 1.29 ^^^^^^^
3837 :     Error: Undefined word
3838 :     @end example
3839 : crook 1.28
3840 : crook 1.29 The reason that this didn't happen is bound up in the way that @code{:}
3841 :     works. The word @code{:} does two special things. The first special
3842 :     thing that it does prevents the text interpreter from ever seeing the
3843 :     characters @code{add-two}. The text interpreter uses a variable called
3844 :     @cindex modifying >IN
3845 : crook 1.44 @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
3846 : crook 1.29 input line. When it encounters the word @code{:} it behaves in exactly
3847 :     the same way as it does for any other word; it looks it up in the name
3848 :     dictionary, finds its xt and executes it. When @code{:} executes, it
3849 :     looks at the input buffer, finds the word @code{add-two} and advances the
3850 :     value of @code{>IN} to point past it. It then does some other stuff
3851 :     associated with creating the new definition (including creating an entry
3852 :     for @code{add-two} in the name dictionary). When the execution of @code{:}
3853 :     completes, control returns to the text interpreter, which is oblivious
3854 :     to the fact that it has been tricked into ignoring part of the input
3855 :     line.
3856 : crook 1.21
3857 : crook 1.29 @cindex parsing words
3858 :     Words like @code{:} -- words that advance the value of @code{>IN} and so
3859 :     prevent the text interpreter from acting on the whole of the input line
3860 :     -- are called @dfn{parsing words}.
3861 : crook 1.21
3862 : crook 1.29 @cindex @code{state} - effect on the text interpreter
3863 :     @cindex text interpreter - effect of state
3864 :     The second special thing that @code{:} does is change the value of a
3865 :     variable called @code{state}, which affects the way that the text
3866 :     interpreter behaves. When Gforth starts up, @code{state} has the value
3867 :     0, and the text interpreter is said to be @dfn{interpreting}. During a
3868 :     colon definition (started with @code{:}), @code{state} is set to -1 and
3869 : crook 1.44 the text interpreter is said to be @dfn{compiling}.
3870 :    
3871 :     In this example, the text interpreter is compiling when it processes the
3872 :     string ``@code{2 + . ;}''. It still breaks the string down into
3873 :     character sequences in the same way. However, instead of pushing the
3874 :     number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3875 :     into the definition of @code{add-two} that will make the number @code{2} get
3876 :     pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3877 :     the behaviours of @code{+} and @code{.} are also compiled into the
3878 :     definition.
3879 :    
3880 :     One category of words don't get compiled. These so-called @dfn{immediate
3881 :     words} get executed (performed @i{now}) regardless of whether the text
3882 :     interpreter is interpreting or compiling. The word @code{;} is an
3883 :     immediate word. Rather than being compiled into the definition, it
3884 :     executes. Its effect is to terminate the current definition, which
3885 :     includes changing the value of @code{state} back to 0.
3886 :    
3887 :     When you execute @code{add-two}, it has a @dfn{run-time effect} that is
3888 :     exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
3889 :     definition.
3890 : crook 1.28
3891 : anton 1.30 In Forth, every word or number can be described in terms of two
3892 : crook 1.29 properties:
3893 : crook 1.28
3894 :     @itemize @bullet
3895 :     @item
3896 : crook 1.29 @cindex interpretation semantics
3897 : crook 1.44 Its @dfn{interpretation semantics} describe how it will behave when the
3898 :     text interpreter encounters it in @dfn{interpret} state. The
3899 :     interpretation semantics of a word are represented by an @dfn{execution
3900 :     token}.
3901 : crook 1.28 @item
3902 : crook 1.29 @cindex compilation semantics
3903 : crook 1.44 Its @dfn{compilation semantics} describe how it will behave when the
3904 :     text interpreter encounters it in @dfn{compile} state. The compilation
3905 :     semantics of a word are represented in an implementation-dependent way;
3906 :     Gforth uses a @dfn{compilation token}.
3907 : crook 1.29 @end itemize
3908 :    
3909 :     @noindent
3910 :     Numbers are always treated in a fixed way:
3911 :    
3912 :     @itemize @bullet
3913 : crook 1.28 @item
3914 : crook 1.44 When the number is @dfn{interpreted}, its behaviour is to push the
3915 :     number onto the stack.
3916 : crook 1.28 @item
3917 : anton 1.30 When the number is @dfn{compiled}, a piece of code is appended to the
3918 :     current definition that pushes the number when it runs. (In other words,
3919 :     the compilation semantics of a number are to postpone its interpretation
3920 :     semantics until the run-time of the definition that it is being compiled
3921 :     into.)
3922 : crook 1.29 @end itemize
3923 :    
3924 : crook 1.44 Words don't behave in such a regular way, but most have @i{default
3925 :     semantics} which means that they behave like this:
3926 : crook 1.29
3927 :     @itemize @bullet
3928 : crook 1.28 @item
3929 : anton 1.30 The @dfn{interpretation semantics} of the word are to do something useful.
3930 :     @item
3931 : crook 1.29 The @dfn{compilation semantics} of the word are to append its
3932 : anton 1.30 @dfn{interpretation semantics} to the current definition (so that its
3933 :     run-time behaviour is to do something useful).
3934 : crook 1.28 @end itemize
3935 :    
3936 : anton 1.30 @cindex immediate words
3937 : crook 1.44 The actual behaviour of any particular word can be controlled by using
3938 :     the words @code{immediate} and @code{compile-only} when the word is
3939 :     defined. These words set flags in the name dictionary entry of the most
3940 :     recently defined word, and these flags are retrieved by the text
3941 :     interpreter when it finds the word in the name dictionary.
3942 :    
3943 :     A word that is marked as @dfn{immediate} has compilation semantics that
3944 :     are identical to its interpretation semantics. In other words, it
3945 :     behaves like this:
3946 : crook 1.29
3947 :     @itemize @bullet
3948 :     @item
3949 : anton 1.30 The @dfn{interpretation semantics} of the word are to do something useful.
3950 : crook 1.29 @item
3951 : anton 1.30 The @dfn{compilation semantics} of the word are to do something useful
3952 :     (and actually the same thing); i.e., it is executed during compilation.
3953 : crook 1.29 @end itemize
3954 : crook 1.28
3955 : crook 1.44 Marking a word as @dfn{compile-only} prohibits the text interpreter from
3956 :     performing the interpretation semantics of the word directly; an attempt
3957 :     to do so will generate an error. It is never necessary to use
3958 :     @code{compile-only} (and it is not even part of ANS Forth, though it is
3959 :     provided by many implementations) but it is good etiquette to apply it
3960 :     to a word that will not behave correctly (and might have unexpected
3961 :     side-effects) in interpret state. For example, it is only legal to use
3962 :     the conditional word @code{IF} within a definition. If you forget this
3963 :     and try to use it elsewhere, the fact that (in Gforth) it is marked as
3964 :     @code{compile-only} allows the text interpreter to generate a helpful
3965 :     error message rather than subjecting you to the consequences of your
3966 :     folly.
3967 :    
3968 : crook 1.29 This example shows the difference between an immediate and a
3969 :     non-immediate word:
3970 : crook 1.28
3971 : crook 1.29 @example
3972 :     : show-state state @@ . ;
3973 :     : show-state-now show-state ; immediate
3974 :     : word1 show-state ;
3975 :     : word2 show-state-now ;
3976 : crook 1.28 @end example
3977 : crook 1.23
3978 : crook 1.29 The word @code{immediate} after the definition of @code{show-state-now}
3979 :     makes that word an immediate word. These definitions introduce a new
3980 :     word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
3981 :     variable, and leaves it on the stack. Therefore, the behaviour of
3982 :     @code{show-state} is to print a number that represents the current value
3983 :     of @code{state}.
3984 : crook 1.28
3985 : crook 1.29 When you execute @code{word1}, it prints the number 0, indicating that
3986 :     the system is interpreting. When the text interpreter compiled the
3987 :     definition of @code{word1}, it encountered @code{show-state} whose
3988 : anton 1.30 compilation semantics are to append its interpretation semantics to the
3989 : crook 1.29 current definition. When you execute @code{word1}, it performs the
3990 : anton 1.30 interpretation semantics of @code{show-state}. At the time that @code{word1}
3991 : crook 1.29 (and therefore @code{show-state}) are executed, the system is
3992 :     interpreting.
3993 : crook 1.28
3994 : anton 1.30 When you pressed @key{RET} after entering the definition of @code{word2},
3995 : crook 1.29 you should have seen the number -1 printed, followed by ``@code{
3996 :     ok}''. When the text interpreter compiled the definition of
3997 :     @code{word2}, it encountered @code{show-state-now}, an immediate word,
3998 : anton 1.30 whose compilation semantics are therefore to perform its interpretation
3999 : crook 1.29 semantics. It is executed straight away (even before the text
4000 :     interpreter has moved on to process another group of characters; the
4001 :     @code{;} in this example). The effect of executing it are to display the
4002 :     value of @code{state} @i{at the time that the definition of}
4003 :     @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4004 :     system is compiling at this time. If you execute @code{word2} it does
4005 :     nothing at all.
4006 : crook 1.28
4007 : crook 1.29 @cindex @code{."}, how it works
4008 :     Before leaving the subject of immediate words, consider the behaviour of
4009 :     @code{."} in the definition of @code{greet}, in the previous
4010 :     section. This word is both a parsing word and an immediate word. Notice
4011 :     that there is a space between @code{."} and the start of the text
4012 :     @code{Hello and welcome}, but that there is no space between the last
4013 :     letter of @code{welcome} and the @code{"} character. The reason for this
4014 :     is that @code{."} is a Forth word; it must have a space after it so that
4015 :     the text interpreter can identify it. The @code{"} is not a Forth word;
4016 :     it is a @dfn{delimiter}. The examples earlier show that, when the string
4017 :     is displayed, there is neither a space before the @code{H} nor after the
4018 :     @code{e}. Since @code{."} is an immediate word, it executes at the time
4019 :     that @code{greet} is defined. When it executes, its behaviour is to
4020 :     search forward in the input line looking for the delimiter. When it
4021 :     finds the delimiter, it updates @code{>IN} to point past the
4022 :     delimiter. It also compiles some magic code into the definition of
4023 :     @code{greet}; the xt of a run-time routine that prints a text string. It
4024 :     compiles the string @code{Hello and welcome} into memory so that it is
4025 :     available to be printed later. When the text interpreter gains control,
4026 :     the next word it finds in the input stream is @code{;} and so it
4027 :     terminates the definition of @code{greet}.
4028 : crook 1.28
4029 :    
4030 :     @comment ----------------------------------------------
4031 : crook 1.29 @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4032 :     @section Forth is written in Forth
4033 :     @cindex structure of Forth programs
4034 :    
4035 :     When you start up a Forth compiler, a large number of definitions
4036 :     already exist. In Forth, you develop a new application using bottom-up
4037 :     programming techniques to create new definitions that are defined in
4038 :     terms of existing definitions. As you create each definition you can
4039 :     test and debug it interactively.
4040 :    
4041 :     If you have tried out the examples in this section, you will probably
4042 :     have typed them in by hand; when you leave Gforth, your definitions will
4043 :     be lost. You can avoid this by using a text editor to enter Forth source
4044 :     code into a file, and then loading code from the file using
4045 : anton 1.49 @code{include} (@pxref{Forth source files}). A Forth source file is
4046 : crook 1.29 processed by the text interpreter, just as though you had typed it in by
4047 :     hand@footnote{Actually, there are some subtle differences -- see
4048 :     @ref{The Text Interpreter}.}.
4049 :    
4050 :     Gforth also supports the traditional Forth alternative to using text
4051 : anton 1.49 files for program entry (@pxref{Blocks}).
4052 : crook 1.28
4053 : crook 1.29 In common with many, if not most, Forth compilers, most of Gforth is
4054 :     actually written in Forth. All of the @file{.fs} files in the
4055 :     installation directory@footnote{For example,
4056 : anton 1.30 @file{/usr/local/share/gforth...}} are Forth source files, which you can
4057 : crook 1.29 study to see examples of Forth programming.
4058 : crook 1.28
4059 : crook 1.29 Gforth maintains a history file that records every line that you type to
4060 :     the text interpreter. This file is preserved between sessions, and is
4061 :     used to provide a command-line recall facility. If you enter long
4062 :     definitions by hand, you can use a text editor to paste them out of the
4063 :     history file into a Forth source file for reuse at a later time
4064 : anton 1.49 (for more information @pxref{Command-line editing}).
4065 : crook 1.28
4066 :    
4067 :     @comment ----------------------------------------------
4068 : crook 1.29 @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4069 :     @section Review - elements of a Forth system
4070 :     @cindex elements of a Forth system
4071 : crook 1.28
4072 : crook 1.29 To summarise this chapter:
4073 : crook 1.28
4074 :     @itemize @bullet
4075 :     @item
4076 : crook 1.29 Forth programs use @dfn{factoring} to break a problem down into small
4077 :     fragments called @dfn{words} or @dfn{definitions}.
4078 :     @item
4079 :     Forth program development is an interactive process.
4080 :     @item
4081 :     The main command loop that accepts input, and controls both
4082 :     interpretation and compilation, is called the @dfn{text interpreter}
4083 :     (also known as the @dfn{outer interpreter}).
4084 :     @item
4085 :     Forth has a very simple syntax, consisting of words and numbers
4086 :     separated by spaces or carriage-return characters. Any additional syntax
4087 :     is imposed by @dfn{parsing words}.
4088 :     @item
4089 :     Forth uses a stack to pass parameters between words. As a result, it
4090 :     uses postfix notation.
4091 :     @item
4092 :     To use a word that has previously been defined, the text interpreter
4093 :     searches for the word in the @dfn{name dictionary}.
4094 :     @item
4095 : anton 1.30 Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4096 : crook 1.28 @item
4097 : crook 1.29 The text interpreter uses the value of @code{state} to select between
4098 :     the use of the @dfn{interpretation semantics} and the @dfn{compilation
4099 :     semantics} of a word that it encounters.
4100 : crook 1.28 @item
4101 : anton 1.30 The relationship between the @dfn{interpretation semantics} and
4102 :     @dfn{compilation semantics} for a word
4103 : crook 1.29 depend upon the way in which the word was defined (for example, whether
4104 :     it is an @dfn{immediate} word).
4105 : crook 1.28 @item
4106 : crook 1.29 Forth definitions can be implemented in Forth (called @dfn{high-level
4107 :     definitions}) or in some other way (usually a lower-level language and
4108 :     as a result often called @dfn{low-level definitions}, @dfn{code
4109 :     definitions} or @dfn{primitives}).
4110 : crook 1.28 @item
4111 : crook 1.29 Many Forth systems are implemented mainly in Forth.
4112 : crook 1.28 @end itemize
4113 :    
4114 :    
4115 : crook 1.29 @comment ----------------------------------------------
4116 : anton 1.48 @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4117 : crook 1.29 @section Where To Go Next
4118 :     @cindex where to go next
4119 : crook 1.28
4120 : crook 1.29 Amazing as it may seem, if you have read (and understood) this far, you
4121 :     know almost all the fundamentals about the inner workings of a Forth
4122 :     system. You certainly know enough to be able to read and understand the
4123 :     rest of this manual and the ANS Forth document, to learn more about the
4124 :     facilities that Forth in general and Gforth in particular provide. Even
4125 :     scarier, you know almost enough to implement your own Forth system.
4126 : anton 1.30 However, that's not a good idea just yet... better to try writing some
4127 : crook 1.29 programs in Gforth.
4128 : crook 1.28
4129 : crook 1.29 Forth has such a rich vocabulary that it can be hard to know where to
4130 :     start in learning it. This section suggests a few sets of words that are
4131 :     enough to write small but useful programs. Use the word index in this
4132 :     document to learn more about each word, then try it out and try to write
4133 :     small definitions using it. Start by experimenting with these words:
4134 : crook 1.28
4135 :     @itemize @bullet
4136 :     @item
4137 : crook 1.29 Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4138 :     @item
4139 :     Comparison: @code{MIN MAX =}
4140 :     @item
4141 :     Logic: @code{AND OR XOR NOT}
4142 :     @item
4143 :     Stack manipulation: @code{DUP DROP SWAP OVER}
4144 : crook 1.28 @item
4145 : crook 1.29 Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4146 : crook 1.28 @item
4147 : crook 1.29 Input/Output: @code{. ." EMIT CR KEY}
4148 : crook 1.28 @item
4149 : crook 1.29 Defining words: @code{: ; CREATE}
4150 : crook 1.28 @item
4151 : crook 1.29 Memory allocation words: @code{ALLOT ,}
4152 : crook 1.28 @item
4153 : crook 1.29 Tools: @code{SEE WORDS .S MARKER}
4154 :     @end itemize
4155 :    
4156 :     When you have mastered those, go on to:
4157 :    
4158 :     @itemize @bullet
4159 : crook 1.28 @item
4160 : crook 1.29 More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4161 : crook 1.28 @item
4162 : crook 1.29 Memory access: @code{@@ !}
4163 : crook 1.28 @end itemize
4164 : crook 1.23
4165 : crook 1.29 When you have mastered these, there's nothing for it but to read through
4166 :     the whole of this manual and find out what you've missed.
4167 :    
4168 :     @comment ----------------------------------------------
4169 : anton 1.48 @node Exercises, , Where to go next, Introduction
4170 : crook 1.29 @section Exercises
4171 :     @cindex exercises
4172 :    
4173 :     TODO: provide a set of programming excercises linked into the stuff done
4174 :     already and into other sections of the manual. Provide solutions to all
4175 :     the exercises in a .fs file in the distribution.
4176 :    
4177 :     @c Get some inspiration from Starting Forth and Kelly&Spies.
4178 :    
4179 :     @c excercises:
4180 :     @c 1. take inches and convert to feet and inches.
4181 :     @c 2. take temperature and convert from fahrenheight to celcius;
4182 :     @c may need to care about symmetric vs floored??
4183 :     @c 3. take input line and do character substitution
4184 :     @c to encipher or decipher
4185 :     @c 4. as above but work on a file for in and out
4186 :     @c 5. take input line and convert to pig-latin
4187 :     @c
4188 :     @c thing of sets of things to exercise then come up with
4189 :     @c problems that need those things.
4190 :    
4191 :    
4192 : crook 1.26 @c ******************************************************************
4193 : crook 1.29 @node Words, Error messages, Introduction, Top
4194 : anton 1.1 @chapter Forth Words
4195 : crook 1.26 @cindex words
4196 : anton 1.1
4197 :     @menu
4198 :     * Notation::
4199 : anton 1.65 * Case insensitivity::
4200 :     * Comments::
4201 :     * Boolean Flags::
4202 : anton 1.1 * Arithmetic::
4203 :     * Stack Manipulation::
4204 : anton 1.5 * Memory::
4205 : anton 1.1 * Control Structures::
4206 :     * Defining Words::
4207 : anton 1.65 * Interpretation and Compilation Semantics::
4208 : crook 1.47 * Tokens for Words::
4209 : anton 1.65 * The Text Interpreter::
4210 :     * Word Lists::
4211 :     * Environmental Queries::
4212 : anton 1.12 * Files::
4213 :     * Blocks::
4214 :     * Other I/O::
4215 :     * Programming Tools::
4216 :     * Assembler and Code Words::
4217 :     * Threading Words::
4218 : crook 1.26 * Locals::
4219 :     * Structures::
4220 :     * Object-oriented Forth::
4221 : anton 1.65 * Passing Commands to the OS::
4222 :     * Keeping track of Time::
4223 :     * Miscellaneous Words::
4224 : anton 1.1 @end menu
4225 :    
4226 : anton 1.65 @node Notation, Case insensitivity, Words, Words
4227 : anton 1.1 @section Notation
4228 :     @cindex notation of glossary entries
4229 :     @cindex format of glossary entries
4230 :     @cindex glossary notation format
4231 :     @cindex word glossary entry format
4232 :    
4233 :     The Forth words are described in this section in the glossary notation
4234 :     that has become a de-facto standard for Forth texts, i.e.,
4235 :    
4236 :     @format
4237 : crook 1.29 @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4238 : anton 1.1 @end format
4239 : crook 1.29 @i{Description}
4240 : anton 1.1
4241 :     @table @var
4242 :     @item word
4243 : crook 1.28 The name of the word.
4244 : anton 1.1
4245 :     @item Stack effect
4246 :     @cindex stack effect
4247 : crook 1.29 The stack effect is written in the notation @code{@i{before} --
4248 :     @i{after}}, where @i{before} and @i{after} describe the top of
4249 : anton 1.1 stack entries before and after the execution of the word. The rest of
4250 :     the stack is not touched by the word. The top of stack is rightmost,
4251 :     i.e., a stack sequence is written as it is typed in. Note that Gforth
4252 :     uses a separate floating point stack, but a unified stack
4253 : crook 1.29 notation. Also, return stack effects are not shown in @i{stack
4254 :     effect}, but in @i{Description}. The name of a stack item describes
4255 : anton 1.1 the type and/or the function of the item. See below for a discussion of
4256 :     the types.
4257 :    
4258 :     All words have two stack effects: A compile-time stack effect and a
4259 :     run-time stack effect. The compile-time stack-effect of most words is
4260 : crook 1.29 @i{ -- }. If the compile-time stack-effect of a word deviates from
4261 : anton 1.1 this standard behaviour, or the word does other unusual things at
4262 :     compile time, both stack effects are shown; otherwise only the run-time
4263 :     stack effect is shown.
4264 :    
4265 :     @cindex pronounciation of words
4266 :     @item pronunciation
4267 :     How the word is pronounced.
4268 :    
4269 :     @cindex wordset
4270 :     @item wordset
4271 : crook 1.21 The ANS Forth standard is divided into several word sets. A standard
4272 :     system need not support all of them. Therefore, in theory, the fewer
4273 :     word sets your program uses the more portable it will be. However, we
4274 :     suspect that most ANS Forth systems on personal machines will feature
4275 : crook 1.26 all word sets. Words that are not defined in ANS Forth have
4276 : crook 1.21 @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4277 : anton 1.1 describes words that will work in future releases of Gforth;
4278 :     @code{gforth-internal} words are more volatile. Environmental query
4279 :     strings are also displayed like words; you can recognize them by the
4280 : crook 1.21 @code{environment} in the word set field.
4281 : anton 1.1
4282 :     @item Description
4283 :     A description of the behaviour of the word.
4284 :     @end table
4285 :    
4286 :     @cindex types of stack items
4287 :     @cindex stack item types
4288 :     The type of a stack item is specified by the character(s) the name
4289 :     starts with:
4290 :    
4291 :     @table @code
4292 :     @item f
4293 :     @cindex @code{f}, stack item type
4294 :     Boolean flags, i.e. @code{false} or @code{true}.
4295 :     @item c
4296 :     @cindex @code{c}, stack item type
4297 :     Char
4298 :     @item w
4299 :     @cindex @code{w}, stack item type
4300 :     Cell, can contain an integer or an address
4301 :     @item n
4302 :     @cindex @code{n}, stack item type
4303 :     signed integer
4304 :     @item u
4305 :     @cindex @code{u}, stack item type
4306 :     unsigned integer
4307 :     @item d
4308 :     @cindex @code{d}, stack item type
4309 :     double sized signed integer
4310 :     @item ud
4311 :     @cindex @code{ud}, stack item type
4312 :     double sized unsigned integer
4313 :     @item r
4314 :     @cindex @code{r}, stack item type
4315 :     Float (on the FP stack)
4316 : crook 1.21 @item a-
4317 : anton 1.1 @cindex @code{a_}, stack item type
4318 :     Cell-aligned address
4319 : crook 1.21 @item c-
4320 : anton 1.1 @cindex @code{c_}, stack item type
4321 :     Char-aligned address (note that a Char may have two bytes in Windows NT)
4322 : crook 1.21 @item f-
4323 : anton 1.1 @cindex @code{f_}, stack item type
4324 :     Float-aligned address
4325 : crook 1.21 @item df-
4326 : anton 1.1 @cindex @code{df_}, stack item type
4327 :     Address aligned for IEEE double precision float
4328 : crook 1.21 @item sf-
4329 : anton 1.1 @cindex @code{sf_}, stack item type
4330 :     Address aligned for IEEE single precision float
4331 :     @item xt
4332 :     @cindex @code{xt}, stack item type
4333 :     Execution token, same size as Cell
4334 :     @item wid
4335 :     @cindex @code{wid}, stack item type
4336 : crook 1.21 Word list ID, same size as Cell
4337 : anton 1.1 @item f83name
4338 :     @cindex @code{f83name}, stack item type
4339 :     Pointer to a name structure
4340 :     @item "
4341 :     @cindex @code{"}, stack item type
4342 : anton 1.12 string in the input stream (not on the stack). The terminating character
4343 :     is a blank by default. If it is not a blank, it is shown in @code{<>}
4344 : anton 1.1 quotes.
4345 :     @end table
4346 :    
4347 : anton 1.65 @comment ----------------------------------------------
4348 :     @node Case insensitivity, Comments, Notation, Words
4349 :     @section Case insensitivity
4350 :     @cindex case sensitivity
4351 :     @cindex upper and lower case
4352 :    
4353 :     Gforth is case-insensitive; you can enter definitions and invoke
4354 :     Standard words using upper, lower or mixed case (however,
4355 :     @pxref{core-idef, Implementation-defined options, Implementation-defined
4356 :     options}).
4357 :    
4358 :     ANS Forth only @i{requires} implementations to recognise Standard words
4359 :     when they are typed entirely in upper case. Therefore, a Standard
4360 :     program must use upper case for all Standard words. You can use whatever
4361 :     case you like for words that you define, but in a Standard program you
4362 :     have to use the words in the same case that you defined them.
4363 :    
4364 :     Gforth supports case sensitivity through @code{table}s (case-sensitive
4365 :     wordlists, @pxref{Word Lists}).
4366 :    
4367 :     Two people have asked how to convert Gforth to be case-sensitive; while
4368 :     we think this is a bad idea, you can change all wordlists into tables
4369 :     like this:
4370 :    
4371 :     @example
4372 :     ' table-find forth-wordlist wordlist-map @ !
4373 :     @end example
4374 :    
4375 :     Note that you now have to type the predefined words in the same case
4376 :     that we defined them, which are varying. You may want to convert them
4377 :     to your favourite case before doing this operation (I won't explain how,
4378 :     because if you are even contemplating doing this, you'd better have
4379 :     enough knowledge of Forth systems to know this already).
4380 :    
4381 :     @node Comments, Boolean Flags, Case insensitivity, Words
4382 : crook 1.21 @section Comments
4383 : crook 1.26 @cindex comments
4384 : crook 1.21
4385 : crook 1.29 Forth supports two styles of comment; the traditional @i{in-line} comment,
4386 :     @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4387 : crook 1.21
4388 : crook 1.44
4389 : crook 1.23 doc-(
4390 : crook 1.21 doc-\
4391 : crook 1.23 doc-\G
4392 : crook 1.21
4393 : crook 1.44
4394 : crook 1.21 @node Boolean Flags, Arithmetic, Comments, Words
4395 :     @section Boolean Flags
4396 : crook 1.26 @cindex Boolean flags
4397 : crook 1.21
4398 :     A Boolean flag is cell-sized. A cell with all bits clear represents the
4399 :     flag @code{false} and a flag with all bits set represents the flag
4400 : crook 1.26 @code{true}. Words that check a flag (for example, @code{IF}) will treat
4401 : crook 1.29 a cell that has @i{any} bit set as @code{true}.
4402 : crook 1.21
4403 : crook 1.44
4404 : crook 1.21 doc-true
4405 :     doc-false
4406 : crook 1.29 doc-on
4407 :     doc-off
4408 : crook 1.21
4409 : crook 1.44
4410 : crook 1.21 @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4411 : anton 1.1 @section Arithmetic
4412 :     @cindex arithmetic words
4413 :    
4414 :     @cindex division with potentially negative operands
4415 :     Forth arithmetic is not checked, i.e., you will not hear about integer
4416 :     overflow on addition or multiplication, you may hear about division by
4417 :     zero if you are lucky. The operator is written after the operands, but
4418 :     the operands are still in the original order. I.e., the infix @code{2-1}
4419 :     corresponds to @code{2 1 -}. Forth offers a variety of division
4420 :     operators. If you perform division with potentially negative operands,
4421 :     you do not want to use @code{/} or @code{/mod} with its undefined
4422 :     behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4423 :     former, @pxref{Mixed precision}).
4424 : crook 1.26 @comment TODO discuss the different division forms and the std approach
4425 : anton 1.1
4426 :     @menu
4427 :     * Single precision::
4428 :     * Bitwise operations::
4429 : crook 1.21 * Double precision:: Double-cell integer arithmetic
4430 :     * Numeric comparison::
4431 : crook 1.29 * Mixed precision:: Operations with single and double-cell integers
4432 : anton 1.1 * Floating Point::
4433 :     @end menu
4434 :    
4435 :     @node Single precision, Bitwise operations, Arithmetic, Arithmetic
4436 :     @subsection Single precision
4437 :     @cindex single precision arithmetic words
4438 :    
4439 : crook 1.21 By default, numbers in Forth are single-precision integers that are 1
4440 : crook 1.26 cell in size. They can be signed or unsigned, depending upon how you
4441 : anton 1.49 treat them. For the rules used by the text interpreter for recognising
4442 :     single-precision integers see @ref{Number Conversion}.
4443 : crook 1.21
4444 : crook 1.44
4445 : anton 1.1 doc-+
4446 : crook 1.21 doc-1+
4447 : anton 1.1 doc--
4448 : crook 1.21 doc-1-
4449 : anton 1.1 doc-*
4450 :     doc-/
4451 :     doc-mod
4452 :     doc-/mod
4453 :     doc-negate
4454 :     doc-abs
4455 :     doc-min
4456 :     doc-max
4457 : crook 1.21 doc-d>s
4458 : crook 1.27 doc-floored
4459 : anton 1.1
4460 : crook 1.44
4461 : crook 1.21 @node Bitwise operations, Double precision, Single precision, Arithmetic
4462 : anton 1.1 @subsection Bitwise operations
4463 :     @cindex bitwise operation words
4464 :    
4465 : crook 1.44
4466 : anton 1.1 doc-and
4467 :     doc-or
4468 :     doc-xor
4469 :     doc-invert
4470 : crook 1.21 doc-lshift
4471 :     doc-rshift
4472 : anton 1.1 doc-2*
4473 : crook 1.21 doc-d2*
4474 : anton 1.1 doc-2/
4475 : crook 1.21 doc-d2/
4476 :    
4477 : crook 1.44
4478 : crook 1.21 @node Double precision, Numeric comparison, Bitwise operations, Arithmetic
4479 :     @subsection Double precision
4480 :     @cindex double precision arithmetic words
4481 :    
4482 : anton 1.49 For the rules used by the text interpreter for
4483 :     recognising double-precision integers, see @ref{Number Conversion}.
4484 : crook 1.21
4485 :     A double precision number is represented by a cell pair, with the most
4486 : anton 1.31 significant cell at the TOS. It is trivial to convert an unsigned
4487 : crook 1.26 single to an (unsigned) double; simply push a @code{0} onto the
4488 :     TOS. Since numbers are represented by Gforth using 2's complement
4489 :     arithmetic, converting a signed single to a (signed) double requires
4490 : anton 1.31 sign-extension across the most significant cell. This can be achieved
4491 : crook 1.26 using @code{s>d}. The moral of the story is that you cannot convert a
4492 :     number without knowing whether it represents an unsigned or a
4493 :     signed number.
4494 : crook 1.21
4495 : crook 1.44
4496 : crook 1.21 doc-s>d
4497 :     doc-d+
4498 :     doc-d-
4499 :     doc-dnegate
4500 :     doc-dabs
4501 :     doc-dmin
4502 :     doc-dmax
4503 :    
4504 : crook 1.44
4505 : crook 1.21 @node Numeric comparison, Mixed precision, Double precision, Arithmetic
4506 :     @subsection Numeric comparison
4507 :     @cindex numeric comparison words
4508 :    
4509 : crook 1.44
4510 : crook 1.28 doc-<
4511 :     doc-<=
4512 :     doc-<>
4513 :     doc-=
4514 :     doc->
4515 :     doc->=
4516 :    
4517 : crook 1.21 doc-0<
4518 : crook 1.23 doc-0<=
4519 : crook 1.21 doc-0<>
4520 :     doc-0=
4521 : crook 1.23 doc-0>
4522 :     doc-0>=
4523 : crook 1.28
4524 :     doc-u<
4525 :     doc-u<=
4526 : crook 1.44 @c u<> and u= exist but are the same as <> and =
4527 : anton 1.31 @c doc-u<>
4528 :     @c doc-u=
4529 : crook 1.28 doc-u>
4530 :     doc-u>=
4531 :    
4532 :     doc-within
4533 :    
4534 :     doc-d<
4535 :     doc-d<=
4536 :     doc-d<>
4537 :     doc-d=
4538 :     doc-d>
4539 :