Annotation of gforth/gforth.ds, revision 1.16

1.1       anton       1: \input texinfo   @c -*-texinfo-*-
                      2: @comment The source is gforth.ds, from which gforth.texi is generated
                      3: @comment %**start of header (This is for running Texinfo on a region.)
1.4       anton       4: @setfilename
1.1       anton       5: @settitle GNU Forth Manual
1.4       anton       6: @comment @setchapternewpage odd
1.1       anton       7: @comment %**end of header (This is for running Texinfo on a region.)
                      9: @ifinfo
                     10: This file documents GNU Forth 0.0
                     12: Copyright @copyright{} 1994 GNU Forth Development Group
                     14:      Permission is granted to make and distribute verbatim copies of
                     15:      this manual provided the copyright notice and this permission notice
                     16:      are preserved on all copies.
1.4       anton      18: @ignore
1.1       anton      19:      Permission is granted to process this file through TeX and print the
                     20:      results, provided the printed document carries a copying permission
                     21:      notice identical to this one except for the removal of this paragraph
                     22:      (this paragraph not being relevant to the printed manual).
1.4       anton      24: @end ignore
1.1       anton      25:      Permission is granted to copy and distribute modified versions of this
                     26:      manual under the conditions for verbatim copying, provided also that the
                     27:      sections entitled "Distribution" and "General Public License" are
                     28:      included exactly as in the original, and provided that the entire
                     29:      resulting derived work is distributed under the terms of a permission
                     30:      notice identical to this one.
                     32:      Permission is granted to copy and distribute translations of this manual
                     33:      into another language, under the above conditions for modified versions,
                     34:      except that the sections entitled "Distribution" and "General Public
                     35:      License" may be included in a translation approved by the author instead
                     36:      of in the original English.
                     37: @end ifinfo
                     39: @titlepage
                     40: @sp 10
                     41: @center @titlefont{GNU Forth Manual}
                     42: @sp 2
                     43: @center for version 0.0
                     44: @sp 2
                     45: @center Anton Ertl
                     47: @comment  The following two commands start the copyright page.
                     48: @page
                     49: @vskip 0pt plus 1filll
                     50: Copyright @copyright{} 1994 GNU Forth Development Group
                     52: @comment !! Published by ... or You can get a copy of this manual ...
                     54:      Permission is granted to make and distribute verbatim copies of
                     55:      this manual provided the copyright notice and this permission notice
                     56:      are preserved on all copies.
                     58:      Permission is granted to copy and distribute modified versions of this
                     59:      manual under the conditions for verbatim copying, provided also that the
                     60:      sections entitled "Distribution" and "General Public License" are
                     61:      included exactly as in the original, and provided that the entire
                     62:      resulting derived work is distributed under the terms of a permission
                     63:      notice identical to this one.
                     65:      Permission is granted to copy and distribute translations of this manual
                     66:      into another language, under the above conditions for modified versions,
                     67:      except that the sections entitled "Distribution" and "General Public
                     68:      License" may be included in a translation approved by the author instead
                     69:      of in the original English.
                     70: @end titlepage
                     73: @node Top, License, (dir), (dir)
                     74: @ifinfo
                     75: GNU Forth is a free implementation of ANS Forth available on many
                     76: personal machines. This manual corresponds to version 0.0.
                     77: @end ifinfo
                     79: @menu
1.4       anton      80: * License::                     
                     81: * Goals::                       About the GNU Forth Project
                     82: * Other Books::                 Things you might want to read
                     83: * Invocation::                  Starting GNU Forth
                     84: * Words::                       Forth words available in GNU Forth
                     85: * ANS conformance::             Implementation-defined options etc.
                     86: * Model::                       The abstract machine of GNU Forth
                     87: * Emacs and GForth::            The GForth Mode
                     88: * Internals::                   Implementation details
                     89: * Bugs::                        How to report them
                     90: * Pedigree::                    Ancestors of GNU Forth
                     91: * Word Index::                  An item for each Forth word
                     92: * Node Index::                  An item for each node
1.1       anton      93: @end menu
                     95: @node License, Goals, Top, Top
                     96: @unnumbered License
                     97: !! Insert GPL here
                     99: @iftex
                    100: @unnumbered Preface
                    101: This manual documents GNU Forth. The reader is expected to know
                    102: Forth. This manual is primarily a reference manual. @xref{Other Books}
                    103: for introductory material.
                    104: @end iftex
                    106: @node    Goals, Other Books, License, Top
                    107: @comment node-name,     next,           previous, up
                    108: @chapter Goals of GNU Forth
                    109: @cindex Goals
                    110: The goal of the GNU Forth Project is to develop a standard model for
                    111: ANSI Forth. This can be split into several subgoals:
                    113: @itemize @bullet
                    114: @item
                    115: GNU Forth should conform to the ANSI Forth standard.
                    116: @item
                    117: It should be a model, i.e. it should define all the
                    118: implementation-dependent things.
                    119: @item
                    120: It should become standard, i.e. widely accepted and used. This goal
                    121: is the most difficult one.
                    122: @end itemize
                    124: To achieve these goals GNU Forth should be
                    125: @itemize @bullet
                    126: @item
                    127: Similar to previous models (fig-Forth, F83)
                    128: @item
                    129: Powerful. It should provide for all the things that are considered
                    130: necessary today and even some that are not yet considered necessary.
                    131: @item
                    132: Efficient. It should not get the reputation of being exceptionally
                    133: slow.
                    134: @item
                    135: Free.
                    136: @item
                    137: Available on many machines/easy to port.
                    138: @end itemize
                    140: Have we achieved these goals? GNU Forth conforms to the ANS Forth
                    141: standard; it may be considered a model, but we have not yet documented
                    142: which parts of the model are stable and which parts we are likely to
                    143: change; it certainly has not yet become a de facto standard. It has some
                    144: similarities and some differences to previous models; It has some
                    145: powerful features, but not yet everything that we envisioned; on RISCs
                    146: it is as fast as interpreters programmed in assembly, on
                    147: register-starved machines it is not so fast, but still faster than any
                    148: other C-based interpretive implementation; it is free and available on
                    149: many machines.
                    151: @node Other Books, Invocation, Goals, Top
                    152: @chapter Other books on ANS Forth
                    154: As the standard is relatively new, there are not many books out yet. It
                    155: is not recommended to learn Forth by using GNU Forth and a book that is
                    156: not written for ANS Forth, as you will not know your mistakes from the
                    157: deviations of the book.
                    159: There is, of course, the standard, the definite reference if you want to
                    160: write ANS Forth programs. It will be available in printed form from
                    161: Global Engineering Documents !! somtime in spring or summer 1994. If you
                    162: are lucky, you can still get dpANS6 (the draft that was approved as
                    163: standard) by aftp from
                    165: @cite{Forth: The new model} by Jack Woehr (!! Publisher) is an
                    166: introductory book based on a draft version of the standard. It does not
                    167: cover the whole standard. It also contains interesting background
                    168: information (Jack Woehr was in the ANS Forth Technical Committe). It is
                    169: not appropriate for complete newbies, but programmers experienced in
                    170: other languages should find it ok.
                    172: @node Invocation, Words, Other Books, Top
                    173: @chapter Invocation
                    175: You will usually just say @code{gforth}. In many other cases the default
                    176: GNU Forth image will be invoked like this:
                    178: @example
                    179: gforth [files] [-e forth-code]
                    180: @end example
                    182: executing the contents of the files and the Forth code in the order they
                    183: are given.
                    185: In general, the command line looks like this:
                    187: @example
                    188: gforth [initialization options] [image-specific options]
                    189: @end example
                    191: The initialization options must come before the rest of the command
                    192: line. They are:
                    194: @table @code
                    195: @item --image-file @var{file}
                    196: Loads the Forth image @var{file} instead of the default
                    197: @file{}.
                    199: @item --path @var{path}
                    200: Uses @var{path} for searching the image file and Forth source code
                    201: files instead of the default in the environment variable
                    202: @code{GFORTHPATH} or the path specified at installation time (typically
                    203: @file{/usr/local/lib/gforth:.}). A path is given as a @code{:}-separated
                    204: list.
                    206: @item --dictionary-size @var{size}
                    207: @item -m @var{size}
                    208: Allocate @var{size} space for the Forth dictionary space instead of
                    209: using the default specified in the image (typically 256K). The
                    210: @var{size} specification consists of an integer and a unit (e.g.,
                    211: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
                    212: size, in this case Cells), @code{k} (kilobytes), and @code{M}
                    213: (Megabytes). If no unit is specified, @code{e} is used.
                    215: @item --data-stack-size @var{size}
                    216: @item -d @var{size}
                    217: Allocate @var{size} space for the data stack instead of using the
                    218: default specified in the image (typically 16K).
                    220: @item --return-stack-size @var{size}
                    221: @item -r @var{size}
                    222: Allocate @var{size} space for the return stack instead of using the
                    223: default specified in the image (typically 16K).
                    225: @item --fp-stack-size @var{size}
                    226: @item -f @var{size}
                    227: Allocate @var{size} space for the floating point stack instead of
                    228: using the default specified in the image (typically 16K). In this case
                    229: the unit specifier @code{e} refers to floating point numbers.
                    231: @item --locals-stack-size @var{size}
                    232: @item -l @var{size}
                    233: Allocate @var{size} space for the locals stack instead of using the
                    234: default specified in the image (typically 16K).
                    236: @end table
                    238: As explained above, the image-specific command-line arguments for the
                    239: default image @file{} consist of a sequence of filenames and
                    240: @code{-e @var{forth-code}} options that are interpreted in the seqence
                    241: in which they are given. The @code{-e @var{forth-code}} or
                    242: @code{--evaluate @var{forth-code}} option evaluates the forth
                    243: code. This option takes only one argument; if you want to evaluate more
                    244: Forth words, you have to quote them or use several @code{-e}s. To exit
                    245: after processing the command line (instead of entering interactive mode)
                    246: append @code{-e bye} to the command line.
                    248: Not yet implemented:
                    249: On startup the system first executes the system initialization file
                    250: (unless the option @code{--no-init-file} is given; note that the system
                    251: resulting from using this option may not be ANS Forth conformant). Then
                    252: the user initialization file @file{.gforth.fs} is executed, unless the
                    253: option @code{--no-rc} is given; this file is first searched in @file{.},
                    254: then in @file{~}, then in the normal path (see above).
1.4       anton     256: @node Words, ANS conformance, Invocation, Top
1.1       anton     257: @chapter Forth Words
                    259: @menu
1.4       anton     260: * Notation::                    
                    261: * Arithmetic::                  
                    262: * Stack Manipulation::          
                    263: * Memory access::               
                    264: * Control Structures::          
                    265: * Locals::                      
                    266: * Defining Words::              
                    267: * Wordlists::                   
                    268: * Files::                       
                    269: * Blocks::                      
                    270: * Other I/O::                   
                    271: * Programming Tools::           
                    272: * Threading Words::             
1.1       anton     273: @end menu
                    275: @node Notation, Arithmetic, Words, Words
                    276: @section Notation
                    278: The Forth words are described in this section in the glossary notation
                    279: that has become a de-facto standard for Forth texts, i.e.
1.4       anton     281: @format
1.1       anton     282: @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}
1.4       anton     283: @end format
1.1       anton     284: @var{Description}
                    286: @table @var
                    287: @item word
                    288: The name of the word. BTW, GNU Forth is case insensitive, so you can
1.14      anton     289: type the words in in lower case (However, @pxref{core-idef}).
1.1       anton     290: 
                    291: @item Stack effect
                    292: The stack effect is written in the notation @code{@var{before} --
                    293: @var{after}}, where @var{before} and @var{after} describe the top of
                    294: stack entries before and after the execution of the word. The rest of
                    295: the stack is not touched by the word. The top of stack is rightmost,
                    296: i.e., a stack sequence is written as it is typed in. Note that GNU Forth
                    297: uses a separate floating point stack, but a unified stack
                    298: notation. Also, return stack effects are not shown in @var{stack
                    299: effect}, but in @var{Description}. The name of a stack item describes
                    300: the type and/or the function of the item. See below for a discussion of
                    301: the types.
                    303: @item pronunciation
                    304: How the word is pronounced
                    306: @item wordset
                    307: The ANS Forth standard is divided into several wordsets. A standard
                    308: system need not support all of them. So, the fewer wordsets your program
                    309: uses the more portable it will be in theory. However, we suspect that
                    310: most ANS Forth systems on personal machines will feature all
                    311: wordsets. Words that are not defined in the ANS standard have
                    312: @code{gforth} as wordset.
                    314: @item Description
                    315: A description of the behaviour of the word.
                    316: @end table
1.4       anton     318: The type of a stack item is specified by the character(s) the name
                    319: starts with:
1.1       anton     320: 
                    321: @table @code
                    322: @item f
                    323: Bool, i.e. @code{false} or @code{true}.
                    324: @item c
                    325: Char
                    326: @item w
                    327: Cell, can contain an integer or an address
                    328: @item n
                    329: signed integer
                    330: @item u
                    331: unsigned integer
                    332: @item d
                    333: double sized signed integer
                    334: @item ud
                    335: double sized unsigned integer
                    336: @item r
                    337: Float
                    338: @item a_
                    339: Cell-aligned address
                    340: @item c_
                    341: Char-aligned address (note that a Char is two bytes in Windows NT)
                    342: @item f_
                    343: Float-aligned address
                    344: @item df_
                    345: Address aligned for IEEE double precision float
                    346: @item sf_
                    347: Address aligned for IEEE single precision float
                    348: @item xt
                    349: Execution token, same size as Cell
                    350: @item wid
                    351: Wordlist ID, same size as Cell
                    352: @item f83name
                    353: Pointer to a name structure
                    354: @end table
1.4       anton     356: @node Arithmetic, Stack Manipulation, Notation, Words
1.1       anton     357: @section Arithmetic
                    358: Forth arithmetic is not checked, i.e., you will not hear about integer
                    359: overflow on addition or multiplication, you may hear about division by
                    360: zero if you are lucky. The operator is written after the operands, but
                    361: the operands are still in the original order. I.e., the infix @code{2-1}
                    362: corresponds to @code{2 1 -}. Forth offers a variety of division
                    363: operators. If you perform division with potentially negative operands,
                    364: you do not want to use @code{/} or @code{/mod} with its undefined
                    365: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
1.4       anton     366: former, @pxref{Mixed precision}).
                    368: @menu
                    369: * Single precision::            
                    370: * Bitwise operations::          
                    371: * Mixed precision::             operations with single and double-cell integers
                    372: * Double precision::            Double-cell integer arithmetic
                    373: * Floating Point::              
                    374: @end menu
1.1       anton     375: 
1.4       anton     376: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
1.1       anton     377: @subsection Single precision
                    378: doc-+
                    379: doc--
                    380: doc-*
                    381: doc-/
                    382: doc-mod
                    383: doc-/mod
                    384: doc-negate
                    385: doc-abs
                    386: doc-min
                    387: doc-max
1.4       anton     389: @node Bitwise operations, Mixed precision, Single precision, Arithmetic
1.1       anton     390: @subsection Bitwise operations
                    391: doc-and
                    392: doc-or
                    393: doc-xor
                    394: doc-invert
                    395: doc-2*
                    396: doc-2/
1.4       anton     398: @node Mixed precision, Double precision, Bitwise operations, Arithmetic
1.1       anton     399: @subsection Mixed precision
                    400: doc-m+
                    401: doc-*/
                    402: doc-*/mod
                    403: doc-m*
                    404: doc-um*
                    405: doc-m*/
                    406: doc-um/mod
                    407: doc-fm/mod
                    408: doc-sm/rem
1.4       anton     410: @node Double precision, Floating Point, Mixed precision, Arithmetic
1.1       anton     411: @subsection Double precision
1.16    ! anton     412: 
        !           413: The outer (aka text) interpreter converts numbers containing a dot into
        !           414: a double precision number. Note that only numbers with the dot as last
        !           415: character are standard-conforming.
        !           416: 
1.1       anton     417: doc-d+
                    418: doc-d-
                    419: doc-dnegate
                    420: doc-dabs
                    421: doc-dmin
                    422: doc-dmax
1.4       anton     424: @node Floating Point,  , Double precision, Arithmetic
                    425: @subsection Floating Point
1.16    ! anton     426: 
        !           427: The format of floating point numbers recognized by the outer (aka text)
        !           428: interpreter is: a signed decimal number, possibly containing a decimal
        !           429: point (@code{.}), followed by @code{E} or @code{e}, optionally followed
        !           430: by a signed integer (the exponent). E.g., @code{1e} ist the same as
        !           431: @code{+1.0e+1}. Note that a number without @code{e}
        !           432: is not interpreted as floating-point number, but as double (if the
        !           433: number contains a @code{.}) or single precision integer. Also,
        !           434: conversions between string and floating point numbers always use base
        !           435: 10, irrespective of the value of @code{BASE}. If @code{BASE} contains a
        !           436: value greater then 14, the @code{E} may be interpreted as digit and the
        !           437: number will be interpreted as integer, unless it has a signed exponent
        !           438: (both @code{+} and @code{-} are allowed as signs).
1.4       anton     439: 
                    440: Angles in floating point operations are given in radians (a full circle
                    441: has 2 pi radians). Note, that gforth has a separate floating point
                    442: stack, but we use the unified notation.
                    444: Floating point numbers have a number of unpleasant surprises for the
                    445: unwary (e.g., floating point addition is not associative) and even a few
                    446: for the wary. You should not use them unless you know what you are doing
                    447: or you don't care that the results you get are totally bogus. If you
                    448: want to learn about the problems of floating point numbers (and how to
1.11      anton     449: avoid them), you might start with @cite{David Goldberg, What Every
1.6       anton     450: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
                    451: Computing Surveys 23(1):5@minus{}48, March 1991}.
1.4       anton     452: 
                    453: doc-f+
                    454: doc-f-
                    455: doc-f*
                    456: doc-f/
                    457: doc-fnegate
                    458: doc-fabs
                    459: doc-fmax
                    460: doc-fmin
                    461: doc-floor
                    462: doc-fround
                    463: doc-f**
                    464: doc-fsqrt
                    465: doc-fexp
                    466: doc-fexpm1
                    467: doc-fln
                    468: doc-flnp1
                    469: doc-flog
1.6       anton     470: doc-falog
1.4       anton     471: doc-fsin
                    472: doc-fcos
                    473: doc-fsincos
                    474: doc-ftan
                    475: doc-fasin
                    476: doc-facos
                    477: doc-fatan
                    478: doc-fatan2
                    479: doc-fsinh
                    480: doc-fcosh
                    481: doc-ftanh
                    482: doc-fasinh
                    483: doc-facosh
                    484: doc-fatanh
                    486: @node Stack Manipulation, Memory access, Arithmetic, Words
1.1       anton     487: @section Stack Manipulation
                    489: gforth has a data stack (aka parameter stack) for characters, cells,
                    490: addresses, and double cells, a floating point stack for floating point
                    491: numbers, a return stack for storing the return addresses of colon
                    492: definitions and other data, and a locals stack for storing local
                    493: variables. Note that while every sane Forth has a separate floating
                    494: point stack, this is not strictly required; an ANS Forth system could
                    495: theoretically keep floating point numbers on the data stack. As an
                    496: additional difficulty, you don't know how many cells a floating point
                    497: number takes. It is reportedly possible to write words in a way that
                    498: they work also for a unified stack model, but we do not recommend trying
1.4       anton     499: it. Instead, just say that your program has an environmental dependency
                    500: on a separate FP stack.
                    502: Also, a Forth system is allowed to keep the local variables on the
1.1       anton     503: return stack. This is reasonable, as local variables usually eliminate
                    504: the need to use the return stack explicitly. So, if you want to produce
                    505: a standard complying program and if you are using local variables in a
                    506: word, forget about return stack manipulations in that word (see the
                    507: standard document for the exact rules).
1.4       anton     509: @menu
                    510: * Data stack::                  
                    511: * Floating point stack::        
                    512: * Return stack::                
                    513: * Locals stack::                
                    514: * Stack pointer manipulation::  
                    515: @end menu
                    517: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
1.1       anton     518: @subsection Data stack
                    519: doc-drop
                    520: doc-nip
                    521: doc-dup
                    522: doc-over
                    523: doc-tuck
                    524: doc-swap
                    525: doc-rot
                    526: doc--rot
                    527: doc-?dup
                    528: doc-pick
                    529: doc-roll
                    530: doc-2drop
                    531: doc-2nip
                    532: doc-2dup
                    533: doc-2over
                    534: doc-2tuck
                    535: doc-2swap
                    536: doc-2rot
1.4       anton     538: @node Floating point stack, Return stack, Data stack, Stack Manipulation
1.1       anton     539: @subsection Floating point stack
                    540: doc-fdrop
                    541: doc-fnip
                    542: doc-fdup
                    543: doc-fover
                    544: doc-ftuck
                    545: doc-fswap
                    546: doc-frot
1.4       anton     548: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
1.1       anton     549: @subsection Return stack
                    550: doc->r
                    551: doc-r>
                    552: doc-r@
                    553: doc-rdrop
                    554: doc-2>r
                    555: doc-2r>
                    556: doc-2r@
                    557: doc-2rdrop
1.4       anton     559: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
1.1       anton     560: @subsection Locals stack
1.4       anton     562: @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
1.1       anton     563: @subsection Stack pointer manipulation
                    564: doc-sp@
                    565: doc-sp!
                    566: doc-fp@
                    567: doc-fp!
                    568: doc-rp@
                    569: doc-rp!
                    570: doc-lp@
                    571: doc-lp!
1.4       anton     573: @node Memory access, Control Structures, Stack Manipulation, Words
1.1       anton     574: @section Memory access
1.4       anton     576: @menu
                    577: * Stack-Memory transfers::      
                    578: * Address arithmetic::          
                    579: * Memory block access::         
                    580: @end menu
                    582: @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
1.1       anton     583: @subsection Stack-Memory transfers
                    585: doc-@
                    586: doc-!
                    587: doc-+!
                    588: doc-c@
                    589: doc-c!
                    590: doc-2@
                    591: doc-2!
                    592: doc-f@
                    593: doc-f!
                    594: doc-sf@
                    595: doc-sf!
                    596: doc-df@
                    597: doc-df!
1.4       anton     599: @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
1.1       anton     600: @subsection Address arithmetic
                    602: ANS Forth does not specify the sizes of the data types. Instead, it
                    603: offers a number of words for computing sizes and doing address
                    604: arithmetic. Basically, address arithmetic is performed in terms of
                    605: address units (aus); on most systems the address unit is one byte. Note
                    606: that a character may have more than one au, so @code{chars} is no noop
                    607: (on systems where it is a noop, it compiles to nothing).
                    609: ANS Forth also defines words for aligning addresses for specific
                    610: addresses. Many computers require that accesses to specific data types
                    611: must only occur at specific addresses; e.g., that cells may only be
                    612: accessed at addresses divisible by 4. Even if a machine allows unaligned
                    613: accesses, it can usually perform aligned accesses faster. 
                    615: For the performance-concious: alignment operations are usually only
                    616: necessary during the definition of a data structure, not during the
                    617: (more frequent) accesses to it.
                    619: ANS Forth defines no words for character-aligning addresses. This is not
                    620: an oversight, but reflects the fact that addresses that are not
                    621: char-aligned have no use in the standard and therefore will not be
                    622: created.
                    624: The standard guarantees that addresses returned by @code{CREATE}d words
                    625: are cell-aligned; in addition, gforth guarantees that these addresses
                    626: are aligned for all purposes.
1.9       anton     628: Note that the standard defines a word @code{char}, which has nothing to
                    629: do with address arithmetic.
1.1       anton     631: doc-chars
                    632: doc-char+
                    633: doc-cells
                    634: doc-cell+
                    635: doc-align
                    636: doc-aligned
                    637: doc-floats
                    638: doc-float+
                    639: doc-falign
                    640: doc-faligned
                    641: doc-sfloats
                    642: doc-sfloat+
                    643: doc-sfalign
                    644: doc-sfaligned
                    645: doc-dfloats
                    646: doc-dfloat+
                    647: doc-dfalign
                    648: doc-dfaligned
1.10      anton     649: doc-maxalign
                    650: doc-maxaligned
                    651: doc-cfalign
                    652: doc-cfaligned
1.1       anton     653: doc-address-unit-bits
1.4       anton     655: @node Memory block access,  , Address arithmetic, Memory access
1.1       anton     656: @subsection Memory block access
                    658: doc-move
                    659: doc-erase
                    661: While the previous words work on address units, the rest works on
                    662: characters.
                    664: doc-cmove
                    665: doc-cmove>
                    666: doc-fill
                    667: doc-blank
1.4       anton     669: @node Control Structures, Locals, Memory access, Words
1.1       anton     670: @section Control Structures
                    672: Control structures in Forth cannot be used in interpret state, only in
                    673: compile state, i.e., in a colon definition. We do not like this
                    674: limitation, but have not seen a satisfying way around it yet, although
                    675: many schemes have been proposed.
1.4       anton     677: @menu
                    678: * Selection::                   
                    679: * Simple Loops::                
                    680: * Counted Loops::               
                    681: * Arbitrary control structures::  
                    682: * Calls and returns::           
                    683: * Exception Handling::          
                    684: @end menu
                    686: @node Selection, Simple Loops, Control Structures, Control Structures
1.1       anton     687: @subsection Selection
                    689: @example
                    690: @var{flag}
                    691: IF
                    692:   @var{code}
                    693: ENDIF
                    694: @end example
                    695: or
                    696: @example
                    697: @var{flag}
                    698: IF
                    699:   @var{code1}
                    700: ELSE
                    701:   @var{code2}
                    702: ENDIF
                    703: @end example
1.4       anton     705: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
1.1       anton     706: standard, and @code{ENDIF} is not, although it is quite popular. We
                    707: recommend using @code{ENDIF}, because it is less confusing for people
                    708: who also know other languages (and is not prone to reinforcing negative
                    709: prejudices against Forth in these people). Adding @code{ENDIF} to a
                    710: system that only supplies @code{THEN} is simple:
                    711: @example
                    712: : endif   POSTPONE then ; immediate
                    713: @end example
                    715: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
                    716: (adv.)}  has the following meanings:
                    717: @quotation
                    718: ... 2b: following next after in order ... 3d: as a necessary consequence
                    719: (if you were there, then you saw them).
                    720: @end quotation
                    721: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
                    722: and many other programming languages has the meaning 3d.]
                    724: We also provide the words @code{?dup-if} and @code{?dup-0=-if}, so you
                    725: can avoid using @code{?dup}.
                    727: @example
                    728: @var{n}
                    729: CASE
                    730:   @var{n1} OF @var{code1} ENDOF
                    731:   @var{n2} OF @var{code2} ENDOF
1.4       anton     732:   @dots{}
1.1       anton     733: ENDCASE
                    734: @end example
                    736: Executes the first @var{codei}, where the @var{ni} is equal to
                    737: @var{n}. A default case can be added by simply writing the code after
                    738: the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
                    739: but must not consume it.
1.4       anton     741: @node Simple Loops, Counted Loops, Selection, Control Structures
1.1       anton     742: @subsection Simple Loops
                    744: @example
                    745: BEGIN
                    746:   @var{code1}
                    747:   @var{flag}
                    748: WHILE
                    749:   @var{code2}
                    750: REPEAT
                    751: @end example
                    753: @var{code1} is executed and @var{flag} is computed. If it is true,
                    754: @var{code2} is executed and the loop is restarted; If @var{flag} is false, execution continues after the @code{REPEAT}.
                    756: @example
                    757: BEGIN
                    758:   @var{code}
                    759:   @var{flag}
                    760: UNTIL
                    761: @end example
                    763: @var{code} is executed. The loop is restarted if @code{flag} is false.
                    765: @example
                    766: BEGIN
                    767:   @var{code}
                    768: AGAIN
                    769: @end example
                    771: This is an endless loop.
1.4       anton     773: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
1.1       anton     774: @subsection Counted Loops
                    776: The basic counted loop is:
                    777: @example
                    778: @var{limit} @var{start}
                    779: ?DO
                    780:   @var{body}
                    781: LOOP
                    782: @end example
                    784: This performs one iteration for every integer, starting from @var{start}
                    785: and up to, but excluding @var{limit}. The counter, aka index, can be
                    786: accessed with @code{i}. E.g., the loop
                    787: @example
                    788: 10 0 ?DO
                    789:   i .
                    790: LOOP
                    791: @end example
                    792: prints
                    793: @example
                    794: 0 1 2 3 4 5 6 7 8 9
                    795: @end example
                    796: The index of the innermost loop can be accessed with @code{i}, the index
                    797: of the next loop with @code{j}, and the index of the third loop with
                    798: @code{k}.
                    800: The loop control data are kept on the return stack, so there are some
                    801: restrictions on mixing return stack accesses and counted loop
                    802: words. E.g., if you put values on the return stack outside the loop, you
                    803: cannot read them inside the loop. If you put values on the return stack
                    804: within a loop, you have to remove them before the end of the loop and
                    805: before accessing the index of the loop.
                    807: There are several variations on the counted loop:
                    809: @code{LEAVE} leaves the innermost counted loop immediately.
                    811: @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the
                    812: index by @var{n} instead of by 1. The loop is terminated when the border
                    813: between @var{limit-1} and @var{limit} is crossed. E.g.:
1.2       anton     815: @code{4 0 ?DO  i .  2 +LOOP}   prints @code{0 2}
1.1       anton     816: 
1.2       anton     817: @code{4 1 ?DO  i .  2 +LOOP}   prints @code{1 3}
1.1       anton     818: 
                    819: The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:
1.2       anton     821: @code{-1 0 ?DO  i .  -1 +LOOP}  prints @code{0 -1}
1.1       anton     822: 
1.2       anton     823: @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing
1.1       anton     824: 
                    825: Therefore we recommend avoiding using @code{@var{n} +LOOP} with negative
                    826: @var{n}. One alternative is @code{@var{n} S+LOOP}, where the negative
                    827: case behaves symmetrical to the positive case:
1.7       pazsan    829: @code{-2 0 ?DO  i .  -1 S+LOOP}  prints @code{0 -1}
1.1       anton     830: 
1.7       pazsan    831: @code{-1 0 ?DO  i .  -1 S+LOOP}  prints @code{0}
1.1       anton     832: 
1.7       pazsan    833: @code{ 0 0 ?DO  i .  -1 S+LOOP}  prints nothing
1.1       anton     834: 
1.2       anton     835: The loop is terminated when the border between @var{limit@minus{}sgn(n)} and
1.1       anton     836: @var{limit} is crossed. However, @code{S+LOOP} is not part of the ANS
                    837: Forth standard.
                    839: @code{?DO} can be replaced by @code{DO}. @code{DO} enters the loop even
                    840: when the start and the limit value are equal. We do not recommend using
                    841: @code{DO}. It will just give you maintenance troubles.
                    843: @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via
                    844: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
                    845: return stack so @code{EXIT} can get to its return address.
                    847: Another counted loop is
                    848: @example
                    849: @var{n}
                    850: FOR
                    851:   @var{body}
                    852: NEXT
                    853: @end example
                    854: This is the preferred loop of native code compiler writers who are too
                    855: lazy to optimize @code{?DO} loops properly. In GNU Forth, this loop
                    856: iterates @var{n+1} times; @code{i} produces values starting with @var{n}
                    857: and ending with 0. Other Forth systems may behave differently, even if
                    858: they support @code{FOR} loops.
1.4       anton     860: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
1.2       anton     861: @subsection Arbitrary control structures
                    863: ANS Forth permits and supports using control structures in a non-nested
                    864: way. Information about incomplete control structures is stored on the
                    865: control-flow stack. This stack may be implemented on the Forth data
                    866: stack, and this is what we have done in gforth.
                    868: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
                    869: entry represents a backward branch target. A few words are the basis for
                    870: building any control structure possible (except control structures that
                    871: need storage, like calls, coroutines, and backtracking).
1.3       anton     873: doc-if
                    874: doc-ahead
                    875: doc-then
                    876: doc-begin
                    877: doc-until
                    878: doc-again
                    879: doc-cs-pick
                    880: doc-cs-roll
1.2       anton     881: 
                    882: On many systems control-flow stack items take one word, in gforth they
                    883: currently take three (this may change in the future). Therefore it is a
                    884: really good idea to manipulate the control flow stack with
                    885: @code{cs-pick} and @code{cs-roll}, not with data stack manipulation
                    886: words.
                    888: Some standard control structure words are built from these words:
1.3       anton     890: doc-else
                    891: doc-while
                    892: doc-repeat
1.2       anton     893: 
                    894: Counted loop words constitute a separate group of words:
1.3       anton     896: doc-?do
                    897: doc-do
                    898: doc-for
                    899: doc-loop
                    900: doc-s+loop
                    901: doc-+loop
                    902: doc-next
                    903: doc-leave
                    904: doc-?leave
                    905: doc-unloop
1.10      anton     906: doc-done
1.2       anton     907: 
                    908: The standard does not allow using @code{cs-pick} and @code{cs-roll} on
                    909: @i{do-sys}. Our system allows it, but it's your job to ensure that for
                    910: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
1.3       anton     911: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
                    912: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
1.7       pazsan    913: resolved (by using one of the loop-ending words or @code{DONE}).
1.2       anton     914: 
                    915: Another group of control structure words are
1.3       anton     917: doc-case
                    918: doc-endcase
                    919: doc-of
                    920: doc-endof
1.2       anton     921: 
                    922: @i{case-sys} and @i{of-sys} cannot be processed using @code{cs-pick} and
                    923: @code{cs-roll}.
1.3       anton     925: @subsubsection Programming Style
                    927: In order to ensure readability we recommend that you do not create
                    928: arbitrary control structures directly, but define new control structure
                    929: words for the control structure you want and use these words in your
                    930: program.
                    932: E.g., instead of writing
                    934: @example
                    935: begin
                    936:   ...
                    937: if [ 1 cs-roll ]
                    938:   ...
                    939: again then
                    940: @end example
                    942: we recommend defining control structure words, e.g.,
                    944: @example
                    945: : while ( dest -- orig dest )
                    946:  POSTPONE if
                    947:  1 cs-roll ; immediate
                    949: : repeat ( orig dest -- )
                    950:  POSTPONE again
                    951:  POSTPONE then ; immediate
                    952: @end example
                    954: and then using these to create the control structure:
                    956: @example
                    957: begin
                    958:   ...
                    959: while
                    960:   ...
                    961: repeat
                    962: @end example
                    964: That's much easier to read, isn't it? Of course, @code{BEGIN} and
                    965: @code{WHILE} are predefined, so in this example it would not be
                    966: necessary to define them.
1.4       anton     968: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
1.3       anton     969: @subsection Calls and returns
                    971: A definition can be called simply be writing the name of the
                    972: definition. When the end of the definition is reached, it returns. An earlier return can be forced using
                    974: doc-exit
                    976: Don't forget to clean up the return stack and @code{UNLOOP} any
                    977: outstanding @code{?DO}...@code{LOOP}s before @code{EXIT}ing. The
                    978: primitive compiled by @code{EXIT} is
                    980: doc-;s
1.4       anton     982: @node Exception Handling,  , Calls and returns, Control Structures
1.3       anton     983: @subsection Exception Handling
                    985: doc-catch
                    986: doc-throw
1.4       anton     988: @node Locals, Defining Words, Control Structures, Words
1.1       anton     989: @section Locals
1.2       anton     991: Local variables can make Forth programming more enjoyable and Forth
                    992: programs easier to read. Unfortunately, the locals of ANS Forth are
                    993: laden with restrictions. Therefore, we provide not only the ANS Forth
                    994: locals wordset, but also our own, more powerful locals wordset (we
                    995: implemented the ANS Forth locals wordset through our locals wordset).
                    997: @menu
1.4       anton     998: * gforth locals::               
                    999: * ANS Forth locals::            
1.2       anton    1000: @end menu
1.4       anton    1002: @node gforth locals, ANS Forth locals, Locals, Locals
1.2       anton    1003: @subsection gforth locals
                   1005: Locals can be defined with
                   1007: @example
                   1008: @{ local1 local2 ... -- comment @}
                   1009: @end example
                   1010: or
                   1011: @example
                   1012: @{ local1 local2 ... @}
                   1013: @end example
                   1015: E.g.,
                   1016: @example
                   1017: : max @{ n1 n2 -- n3 @}
                   1018:  n1 n2 > if
                   1019:    n1
                   1020:  else
                   1021:    n2
                   1022:  endif ;
                   1023: @end example
                   1025: The similarity of locals definitions with stack comments is intended. A
                   1026: locals definition often replaces the stack comment of a word. The order
                   1027: of the locals corresponds to the order in a stack comment and everything
                   1028: after the @code{--} is really a comment.
                   1030: This similarity has one disadvantage: It is too easy to confuse locals
                   1031: declarations with stack comments, causing bugs and making them hard to
                   1032: find. However, this problem can be avoided by appropriate coding
                   1033: conventions: Do not use both notations in the same program. If you do,
                   1034: they should be distinguished using additional means, e.g. by position.
                   1036: The name of the local may be preceded by a type specifier, e.g.,
                   1037: @code{F:} for a floating point value:
                   1039: @example
                   1040: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
                   1041: \ complex multiplication
                   1042:  Ar Br f* Ai Bi f* f-
                   1043:  Ar Bi f* Ai Br f* f+ ;
                   1044: @end example
                   1046: GNU Forth currently supports cells (@code{W:}, @code{W^}), doubles
                   1047: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
                   1048: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
                   1049: with @code{W:}, @code{D:} etc.) produces its value and can be changed
                   1050: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
                   1051: produces its address (which becomes invalid when the variable's scope is
                   1052: left). E.g., the standard word @code{emit} can be defined in therms of
                   1053: @code{type} like this:
                   1055: @example
                   1056: : emit @{ C^ char* -- @}
                   1057:     char* 1 type ;
                   1058: @end example
                   1060: A local without type specifier is a @code{W:} local. Both flavours of
                   1061: locals are initialized with values from the data or FP stack.
                   1063: Currently there is no way to define locals with user-defined data
                   1064: structures, but we are working on it.
1.7       pazsan   1066: GNU Forth allows defining locals everywhere in a colon definition. This
                   1067: poses the following questions:
1.2       anton    1068: 
1.4       anton    1069: @menu
                   1070: * Where are locals visible by name?::  
1.14      anton    1071: * How long do locals live?::    
1.4       anton    1072: * Programming Style::           
                   1073: * Implementation::              
                   1074: @end menu
                   1076: @node Where are locals visible by name?, How long do locals live?, gforth locals, gforth locals
1.2       anton    1077: @subsubsection Where are locals visible by name?
                   1079: Basically, the answer is that locals are visible where you would expect
                   1080: it in block-structured languages, and sometimes a little longer. If you
                   1081: want to restrict the scope of a local, enclose its definition in
                   1082: @code{SCOPE}...@code{ENDSCOPE}.
                   1084: doc-scope
                   1085: doc-endscope
                   1087: These words behave like control structure words, so you can use them
                   1088: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
                   1089: arbitrary ways.
                   1091: If you want a more exact answer to the visibility question, here's the
                   1092: basic principle: A local is visible in all places that can only be
                   1093: reached through the definition of the local@footnote{In compiler
                   1094: construction terminology, all places dominated by the definition of the
                   1095: local.}. In other words, it is not visible in places that can be reached
                   1096: without going through the definition of the local. E.g., locals defined
                   1097: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
                   1098: defined in @code{BEGIN}...@code{UNTIL} are visible after the
                   1099: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
                   1101: The reasoning behind this solution is: We want to have the locals
                   1102: visible as long as it is meaningful. The user can always make the
                   1103: visibility shorter by using explicit scoping. In a place that can
                   1104: only be reached through the definition of a local, the meaning of a
                   1105: local name is clear. In other places it is not: How is the local
                   1106: initialized at the control flow path that does not contain the
                   1107: definition? Which local is meant, if the same name is defined twice in
                   1108: two independent control flow paths?
                   1110: This should be enough detail for nearly all users, so you can skip the
                   1111: rest of this section. If you relly must know all the gory details and
                   1112: options, read on.
                   1114: In order to implement this rule, the compiler has to know which places
                   1115: are unreachable. It knows this automatically after @code{AHEAD},
                   1116: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
                   1117: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
                   1118: compiler that the control flow never reaches that place. If
                   1119: @code{UNREACHABLE} is not used where it could, the only consequence is
                   1120: that the visibility of some locals is more limited than the rule above
                   1121: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
                   1122: lie to the compiler), buggy code will be produced.
                   1124: Another problem with this rule is that at @code{BEGIN}, the compiler
1.3       anton    1125: does not know which locals will be visible on the incoming
                   1126: back-edge. All problems discussed in the following are due to this
                   1127: ignorance of the compiler (we discuss the problems using @code{BEGIN}
                   1128: loops as examples; the discussion also applies to @code{?DO} and other
1.2       anton    1129: loops). Perhaps the most insidious example is:
                   1130: @example
                   1131: AHEAD
                   1132: BEGIN
                   1133:   x
                   1134: [ 1 CS-ROLL ] THEN
1.4       anton    1135:   @{ x @}
1.2       anton    1136:   ...
                   1137: UNTIL
                   1138: @end example
                   1140: This should be legal according to the visibility rule. The use of
                   1141: @code{x} can only be reached through the definition; but that appears
                   1142: textually below the use.
                   1144: From this example it is clear that the visibility rules cannot be fully
                   1145: implemented without major headaches. Our implementation treats common
                   1146: cases as advertised and the exceptions are treated in a safe way: The
                   1147: compiler makes a reasonable guess about the locals visible after a
                   1148: @code{BEGIN}; if it is too pessimistic, the
                   1149: user will get a spurious error about the local not being defined; if the
                   1150: compiler is too optimistic, it will notice this later and issue a
                   1151: warning. In the case above the compiler would complain about @code{x}
                   1152: being undefined at its use. You can see from the obscure examples in
                   1153: this section that it takes quite unusual control structures to get the
                   1154: compiler into trouble, and even then it will often do fine.
                   1156: If the @code{BEGIN} is reachable from above, the most optimistic guess
                   1157: is that all locals visible before the @code{BEGIN} will also be
                   1158: visible after the @code{BEGIN}. This guess is valid for all loops that
                   1159: are entered only through the @code{BEGIN}, in particular, for normal
                   1160: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
                   1161: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
                   1162: compiler. When the branch to the @code{BEGIN} is finally generated by
                   1163: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
                   1164: warns the user if it was too optimisitic:
                   1165: @example
                   1166: IF
1.4       anton    1167:   @{ x @}
1.2       anton    1168: BEGIN
                   1169:   \ x ? 
                   1170: [ 1 cs-roll ] THEN
                   1171:   ...
                   1172: UNTIL
                   1173: @end example
                   1175: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
                   1176: optimistically assumes that it lives until the @code{THEN}. It notices
                   1177: this difference when it compiles the @code{UNTIL} and issues a
                   1178: warning. The user can avoid the warning, and make sure that @code{x}
                   1179: is not used in the wrong area by using explicit scoping:
                   1180: @example
                   1181: IF
                   1182:   SCOPE
1.4       anton    1183:   @{ x @}
1.2       anton    1184:   ENDSCOPE
                   1185: BEGIN
                   1186: [ 1 cs-roll ] THEN
                   1187:   ...
                   1188: UNTIL
                   1189: @end example
                   1191: Since the guess is optimistic, there will be no spurious error messages
                   1192: about undefined locals.
                   1194: If the @code{BEGIN} is not reachable from above (e.g., after
                   1195: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
                   1196: optimistic guess, as the locals visible after the @code{BEGIN} may be
                   1197: defined later. Therefore, the compiler assumes that no locals are
                   1198: visible after the @code{BEGIN}. However, the useer can use
                   1199: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
                   1200: visible at the BEGIN as at the point where the item was created.
                   1202: doc-assume-live
                   1204: E.g.,
                   1205: @example
1.4       anton    1206: @{ x @}
1.2       anton    1207: AHEAD
                   1208: ASSUME-LIVE
                   1209: BEGIN
                   1210:   x
                   1211: [ 1 CS-ROLL ] THEN
                   1212:   ...
                   1213: UNTIL
                   1214: @end example
                   1216: Other cases where the locals are defined before the @code{BEGIN} can be
                   1217: handled by inserting an appropriate @code{CS-ROLL} before the
                   1218: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
                   1219: behind the @code{ASSUME-LIVE}).
                   1221: Cases where locals are defined after the @code{BEGIN} (but should be
                   1222: visible immediately after the @code{BEGIN}) can only be handled by
                   1223: rearranging the loop. E.g., the ``most insidious'' example above can be
                   1224: arranged into:
                   1225: @example
                   1226: BEGIN
1.4       anton    1227:   @{ x @}
1.2       anton    1228:   ... 0=
                   1229: WHILE
                   1230:   x
                   1231: REPEAT
                   1232: @end example
1.4       anton    1234: @node How long do locals live?, Programming Style, Where are locals visible by name?, gforth locals
1.2       anton    1235: @subsubsection How long do locals live?
                   1237: The right answer for the lifetime question would be: A local lives at
                   1238: least as long as it can be accessed. For a value-flavoured local this
                   1239: means: until the end of its visibility. However, a variable-flavoured
                   1240: local could be accessed through its address far beyond its visibility
                   1241: scope. Ultimately, this would mean that such locals would have to be
                   1242: garbage collected. Since this entails un-Forth-like implementation
                   1243: complexities, I adopted the same cowardly solution as some other
                   1244: languages (e.g., C): The local lives only as long as it is visible;
                   1245: afterwards its address is invalid (and programs that access it
                   1246: afterwards are erroneous).
1.4       anton    1248: @node Programming Style, Implementation, How long do locals live?, gforth locals
1.2       anton    1249: @subsubsection Programming Style
                   1251: The freedom to define locals anywhere has the potential to change
                   1252: programming styles dramatically. In particular, the need to use the
                   1253: return stack for intermediate storage vanishes. Moreover, all stack
                   1254: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
                   1255: determined arguments) can be eliminated: If the stack items are in the
                   1256: wrong order, just write a locals definition for all of them; then
                   1257: write the items in the order you want.
                   1259: This seems a little far-fetched and eliminating stack manipulations is
1.4       anton    1260: unlikely to become a conscious programming objective. Still, the number
                   1261: of stack manipulations will be reduced dramatically if local variables
                   1262: are used liberally (e.g., compare @code{max} in @ref{gforth locals} with
                   1263: a traditional implementation of @code{max}).
1.2       anton    1264: 
                   1265: This shows one potential benefit of locals: making Forth programs more
                   1266: readable. Of course, this benefit will only be realized if the
                   1267: programmers continue to honour the principle of factoring instead of
                   1268: using the added latitude to make the words longer.
                   1270: Using @code{TO} can and should be avoided.  Without @code{TO},
                   1271: every value-flavoured local has only a single assignment and many
                   1272: advantages of functional languages apply to Forth. I.e., programs are
                   1273: easier to analyse, to optimize and to read: It is clear from the
                   1274: definition what the local stands for, it does not turn into something
                   1275: different later.
                   1277: E.g., a definition using @code{TO} might look like this:
                   1278: @example
                   1279: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   1280:  u1 u2 min 0
                   1281:  ?do
                   1282:    addr1 c@ addr2 c@ - ?dup
                   1283:    if
                   1284:      unloop exit
                   1285:    then
                   1286:    addr1 char+ TO addr1
                   1287:    addr2 char+ TO addr2
                   1288:  loop
                   1289:  u1 u2 - ;
                   1290: @end example
                   1291: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
                   1292: every loop iteration. @code{strcmp} is a typical example of the
                   1293: readability problems of using @code{TO}. When you start reading
                   1294: @code{strcmp}, you think that @code{addr1} refers to the start of the
                   1295: string. Only near the end of the loop you realize that it is something
                   1296: else.
                   1298: This can be avoided by defining two locals at the start of the loop that
                   1299: are initialized with the right value for the current iteration.
                   1300: @example
                   1301: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   1302:  addr1 addr2
                   1303:  u1 u2 min 0 
                   1304:  ?do @{ s1 s2 @}
                   1305:    s1 c@ s2 c@ - ?dup 
                   1306:    if
                   1307:      unloop exit
                   1308:    then
                   1309:    s1 char+ s2 char+
                   1310:  loop
                   1311:  2drop
                   1312:  u1 u2 - ;
                   1313: @end example
                   1314: Here it is clear from the start that @code{s1} has a different value
                   1315: in every loop iteration.
1.4       anton    1317: @node Implementation,  , Programming Style, gforth locals
1.2       anton    1318: @subsubsection Implementation
                   1320: GNU Forth uses an extra locals stack. The most compelling reason for
                   1321: this is that the return stack is not float-aligned; using an extra stack
                   1322: also eliminates the problems and restrictions of using the return stack
                   1323: as locals stack. Like the other stacks, the locals stack grows toward
                   1324: lower addresses. A few primitives allow an efficient implementation:
                   1326: doc-@local#
                   1327: doc-f@local#
                   1328: doc-laddr#
                   1329: doc-lp+!#
                   1330: doc-lp!
                   1331: doc->l
                   1332: doc-f>l
                   1334: In addition to these primitives, some specializations of these
                   1335: primitives for commonly occurring inline arguments are provided for
                   1336: efficiency reasons, e.g., @code{@@local0} as specialization of
                   1337: @code{@@local#} for the inline argument 0. The following compiling words
                   1338: compile the right specialized version, or the general version, as
                   1339: appropriate:
1.12      anton    1341: doc-compile-@local
                   1342: doc-compile-f@local
1.2       anton    1343: doc-compile-lp+!
                   1345: Combinations of conditional branches and @code{lp+!#} like
                   1346: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
                   1347: is taken) are provided for efficiency and correctness in loops.
                   1349: A special area in the dictionary space is reserved for keeping the
                   1350: local variable names. @code{@{} switches the dictionary pointer to this
                   1351: area and @code{@}} switches it back and generates the locals
                   1352: initializing code. @code{W:} etc.@ are normal defining words. This
                   1353: special area is cleared at the start of every colon definition.
                   1355: A special feature of GNU Forths dictionary is used to implement the
                   1356: definition of locals without type specifiers: every wordlist (aka
                   1357: vocabulary) has its own methods for searching
1.4       anton    1358: etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
1.2       anton    1359: with a special search method: When it is searched for a word, it
                   1360: actually creates that word using @code{W:}. @code{@{} changes the search
                   1361: order to first search the wordlist containing @code{@}}, @code{W:} etc.,
                   1362: and then the wordlist for defining locals without type specifiers.
                   1364: The lifetime rules support a stack discipline within a colon
                   1365: definition: The lifetime of a local is either nested with other locals
                   1366: lifetimes or it does not overlap them.
                   1368: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
                   1369: pointer manipulation is generated. Between control structure words
                   1370: locals definitions can push locals onto the locals stack. @code{AGAIN}
                   1371: is the simplest of the other three control flow words. It has to
                   1372: restore the locals stack depth of the corresponding @code{BEGIN}
                   1373: before branching. The code looks like this:
                   1374: @format
                   1375: @code{lp+!#} current-locals-size @minus{} dest-locals-size
                   1376: @code{branch} <begin>
                   1377: @end format
                   1379: @code{UNTIL} is a little more complicated: If it branches back, it
                   1380: must adjust the stack just like @code{AGAIN}. But if it falls through,
                   1381: the locals stack must not be changed. The compiler generates the
                   1382: following code:
                   1383: @format
                   1384: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
                   1385: @end format
                   1386: The locals stack pointer is only adjusted if the branch is taken.
                   1388: @code{THEN} can produce somewhat inefficient code:
                   1389: @format
                   1390: @code{lp+!#} current-locals-size @minus{} orig-locals-size
                   1391: <orig target>:
                   1392: @code{lp+!#} orig-locals-size @minus{} new-locals-size
                   1393: @end format
                   1394: The second @code{lp+!#} adjusts the locals stack pointer from the
1.4       anton    1395: level at the @var{orig} point to the level after the @code{THEN}. The
1.2       anton    1396: first @code{lp+!#} adjusts the locals stack pointer from the current
                   1397: level to the level at the orig point, so the complete effect is an
                   1398: adjustment from the current level to the right level after the
                   1399: @code{THEN}.
                   1401: In a conventional Forth implementation a dest control-flow stack entry
                   1402: is just the target address and an orig entry is just the address to be
                   1403: patched. Our locals implementation adds a wordlist to every orig or dest
                   1404: item. It is the list of locals visible (or assumed visible) at the point
                   1405: described by the entry. Our implementation also adds a tag to identify
                   1406: the kind of entry, in particular to differentiate between live and dead
                   1407: (reachable and unreachable) orig entries.
                   1409: A few unusual operations have to be performed on locals wordlists:
                   1411: doc-common-list
                   1412: doc-sub-list?
                   1413: doc-list-size
                   1415: Several features of our locals wordlist implementation make these
                   1416: operations easy to implement: The locals wordlists are organised as
                   1417: linked lists; the tails of these lists are shared, if the lists
                   1418: contain some of the same locals; and the address of a name is greater
                   1419: than the address of the names behind it in the list.
                   1421: Another important implementation detail is the variable
                   1422: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
                   1423: determine if they can be reached directly or only through the branch
                   1424: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
                   1425: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
                   1426: definition, by @code{BEGIN} and usually by @code{THEN}.
                   1428: Counted loops are similar to other loops in most respects, but
                   1429: @code{LEAVE} requires special attention: It performs basically the same
                   1430: service as @code{AHEAD}, but it does not create a control-flow stack
                   1431: entry. Therefore the information has to be stored elsewhere;
                   1432: traditionally, the information was stored in the target fields of the
                   1433: branches created by the @code{LEAVE}s, by organizing these fields into a
                   1434: linked list. Unfortunately, this clever trick does not provide enough
                   1435: space for storing our extended control flow information. Therefore, we
                   1436: introduce another stack, the leave stack. It contains the control-flow
                   1437: stack entries for all unresolved @code{LEAVE}s.
                   1439: Local names are kept until the end of the colon definition, even if
                   1440: they are no longer visible in any control-flow path. In a few cases
                   1441: this may lead to increased space needs for the locals name area, but
                   1442: usually less than reclaiming this space would cost in code size.
1.4       anton    1445: @node ANS Forth locals,  , gforth locals, Locals
1.2       anton    1446: @subsection ANS Forth locals
                   1448: The ANS Forth locals wordset does not define a syntax for locals, but
                   1449: words that make it possible to define various syntaxes. One of the
                   1450: possible syntaxes is a subset of the syntax we used in the gforth locals
                   1451: wordset, i.e.:
                   1453: @example
                   1454: @{ local1 local2 ... -- comment @}
                   1455: @end example
                   1456: or
                   1457: @example
                   1458: @{ local1 local2 ... @}
                   1459: @end example
                   1461: The order of the locals corresponds to the order in a stack comment. The
                   1462: restrictions are:
1.1       anton    1463: 
1.2       anton    1464: @itemize @bullet
                   1465: @item
                   1466: Locals can only be cell-sized values (no type specifers are allowed).
                   1467: @item
                   1468: Locals can be defined only outside control structures.
                   1469: @item
                   1470: Locals can interfere with explicit usage of the return stack. For the
                   1471: exact (and long) rules, see the standard. If you don't use return stack
                   1472: accessing words in a definition using locals, you will we all right. The
                   1473: purpose of this rule is to make locals implementation on the return
                   1474: stack easier.
                   1475: @item
                   1476: The whole definition must be in one line.
                   1477: @end itemize
                   1479: Locals defined in this way behave like @code{VALUE}s
1.4       anton    1480: (@xref{Values}). I.e., they are initialized from the stack. Using their
1.2       anton    1481: name produces their value. Their value can be changed using @code{TO}.
                   1483: Since this syntax is supported by gforth directly, you need not do
                   1484: anything to use it. If you want to port a program using this syntax to
                   1485: another ANS Forth system, use @file{anslocal.fs} to implement the syntax
                   1486: on the other system.
                   1488: Note that a syntax shown in the standard, section A.13 looks
                   1489: similar, but is quite different in having the order of locals
                   1490: reversed. Beware!
                   1492: The ANS Forth locals wordset itself consists of the following word
                   1494: doc-(local)
                   1496: The ANS Forth locals extension wordset defines a syntax, but it is so
                   1497: awful that we strongly recommend not to use it. We have implemented this
                   1498: syntax to make porting to gforth easy, but do not document it here. The
                   1499: problem with this syntax is that the locals are defined in an order
                   1500: reversed with respect to the standard stack comment notation, making
                   1501: programs harder to read, and easier to misread and miswrite. The only
                   1502: merit of this syntax is that it is easy to implement using the ANS Forth
                   1503: locals wordset.
1.3       anton    1504: 
1.4       anton    1505: @node Defining Words, Wordlists, Locals, Words
                   1506: @section Defining Words
1.14      anton    1508: @menu
                   1509: * Values::                      
                   1510: @end menu
1.4       anton    1512: @node Values,  , Defining Words, Defining Words
                   1513: @subsection Values
                   1515: @node Wordlists, Files, Defining Words, Words
                   1516: @section Wordlists
                   1518: @node Files, Blocks, Wordlists, Words
                   1519: @section Files
                   1521: @node Blocks, Other I/O, Files, Words
                   1522: @section Blocks
                   1524: @node Other I/O, Programming Tools, Blocks, Words
                   1525: @section Other I/O
                   1527: @node Programming Tools, Threading Words, Other I/O, Words
                   1528: @section Programming Tools
1.5       anton    1530: @menu
                   1531: * Debugging::                   Simple and quick.
                   1532: * Assertions::                  Making your programs self-checking.
                   1533: @end menu
                   1535: @node Debugging, Assertions, Programming Tools, Programming Tools
1.4       anton    1536: @subsection Debugging
                   1538: The simple debugging aids provided in @file{debugging.fs}
                   1539: are meant to support a different style of debugging than the
                   1540: tracing/stepping debuggers used in languages with long turn-around
                   1541: times.
                   1543: A much better (faster) way in fast-compilig languages is to add
                   1544: printing code at well-selected places, let the program run, look at
                   1545: the output, see where things went wrong, add more printing code, etc.,
                   1546: until the bug is found.
                   1548: The word @code{~~} is easy to insert. It just prints debugging
                   1549: information (by default the source location and the stack contents). It
                   1550: is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
                   1551: query-replace them with nothing). The deferred words
                   1552: @code{printdebugdata} and @code{printdebugline} control the output of
                   1553: @code{~~}. The default source location output format works well with
                   1554: Emacs' compilation mode, so you can step through the program at the
1.5       anton    1555: source level using @kbd{C-x `} (the advantage over a stepping debugger
                   1556: is that you can step in any direction and you know where the crash has
                   1557: happened or where the strange data has occurred).
1.4       anton    1558: 
                   1559: Note that the default actions clobber the contents of the pictured
                   1560: numeric output string, so you should not use @code{~~}, e.g., between
                   1561: @code{<#} and @code{#>}.
                   1563: doc-~~
                   1564: doc-printdebugdata
                   1565: doc-printdebugline
1.5       anton    1567: @node Assertions,  , Debugging, Programming Tools
1.4       anton    1568: @subsection Assertions
1.5       anton    1570: It is a good idea to make your programs self-checking, in particular, if
                   1571: you use an assumption (e.g., that a certain field of a data structure is
                   1572: never zero) that may become wrong during maintenance. GForth supports
                   1573: assertions for this purpose. They are used like this:
                   1575: @example
                   1576: assert( @var{flag} )
                   1577: @end example
                   1579: The code between @code{assert(} and @code{)} should compute a flag, that
                   1580: should be true if everything is alright and false otherwise. It should
                   1581: not change anything else on the stack. The overall stack effect of the
                   1582: assertion is @code{( -- )}. E.g.
                   1584: @example
                   1585: assert( 1 1 + 2 = ) \ what we learn in school
                   1586: assert( dup 0<> ) \ assert that the top of stack is not zero
                   1587: assert( false ) \ this code should not be reached
                   1588: @end example
                   1590: The need for assertions is different at different times. During
                   1591: debugging, we want more checking, in production we sometimes care more
                   1592: for speed. Therefore, assertions can be turned off, i.e., the assertion
                   1593: becomes a comment. Depending on the importance of an assertion and the
                   1594: time it takes to check it, you may want to turn off some assertions and
                   1595: keep others turned on. GForth provides several levels of assertions for
                   1596: this purpose:
                   1598: doc-assert0(
                   1599: doc-assert1(
                   1600: doc-assert2(
                   1601: doc-assert3(
                   1602: doc-assert(
                   1603: doc-)
                   1605: @code{Assert(} is the same as @code{assert1(}. The variable
                   1606: @code{assert-level} specifies the highest assertions that are turned
                   1607: on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
                   1608: @code{assert1(} assertions perform checking, while @code{assert2(} and
                   1609: @code{assert3(} assertions are treated as comments.
                   1611: Note that the @code{assert-level} is evaluated at compile-time, not at
                   1612: run-time. I.e., you cannot turn assertions on or off at run-time, you
                   1613: have to set the @code{assert-level} appropriately before compiling a
                   1614: piece of code. You can compile several pieces of code at several
                   1615: @code{assert-level}s (e.g., a trusted library at level 1 and newly
                   1616: written code at level 3).
                   1618: doc-assert-level
                   1620: If an assertion fails, a message compatible with Emacs' compilation mode
                   1621: is produced and the execution is aborted (currently with @code{ABORT"}.
                   1622: If there is interest, we will introduce a special throw code. But if you
                   1623: intend to @code{catch} a specific condition, using @code{throw} is
                   1624: probably more appropriate than an assertion).
1.4       anton    1626: @node Threading Words,  , Programming Tools, Words
                   1627: @section Threading Words
                   1629: These words provide access to code addresses and other threading stuff
                   1630: in gforth (and, possibly, other interpretive Forths). It more or less
                   1631: abstracts away the differences between direct and indirect threading
                   1632: (and, for direct threading, the machine dependences). However, at
                   1633: present this wordset is still inclomplete. It is also pretty low-level;
                   1634: some day it will hopefully be made unnecessary by an internals words set
                   1635: that abstracts implementation details away completely.
                   1637: doc->code-address
                   1638: doc->does-code
                   1639: doc-code-address!
                   1640: doc-does-code!
                   1641: doc-does-handler!
                   1642: doc-/does-handler
1.14      anton    1644: 
1.4       anton    1646: @node ANS conformance, Model, Words, Top
                   1647: @chapter ANS conformance
1.15      anton    1649: To the best of our knowledge, gforth is an
1.14      anton    1650: 
1.15      anton    1651: ANS Forth System
                   1652: @itemize
                   1653: @item providing the Core Extensions word set
                   1654: @item providing the Block word set
                   1655: @item providing the Block Extensions word set
                   1656: @item providing the Double-Number word set
                   1657: @item providing the Double-Number Extensions word set
                   1658: @item providing the Exception word set
                   1659: @item providing the Exception Extensions word set
                   1660: @item providing the Facility word set
                   1661: @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
                   1662: @item providing the File Access word set
                   1663: @item providing the File Access Extensions word set
                   1664: @item providing the Floating-Point word set
                   1665: @item providing the Floating-Point Extensions word set
                   1666: @item providing the Locals word set
                   1667: @item providing the Locals Extensions word set
                   1668: @item providing the Memory-Allocation word set
                   1669: @item providing the Memory-Allocation Extensions word set (that one's easy)
                   1670: @item providing the Programming-Tools word set
                   1671: @item providing @code{AHEAD}, @code{BYE}, @code{CS-PICK}, @code{CS-ROLL}, @code{STATE}, @code{[ELSE]}, @code{[IF]}, @code{[THEN]} from the Programming-Tools Extensions word set
                   1672: @item providing the Search-Order word set
                   1673: @item providing the Search-Order Extensions word set
                   1674: @item providing the String word set
                   1675: @item providing the String Extensions word set (another easy one)
                   1676: @end itemize
                   1678: In addition, ANS Forth systems are required to document certain
                   1679: implementation choices. This chapter tries to meet these
                   1680: requirements. In many cases it gives a way to ask the system for the
                   1681: information instead of providing the information directly, in
                   1682: particular, if the information depends on the processor, the operating
                   1683: system or the installation options chosen, or if they are likely to
                   1684: change during the maintenance of gforth.
1.14      anton    1686: @comment The framework for the rest has been taken from pfe.
                   1688: @menu
                   1689: * The Core Words::              
                   1690: * The optional Block word set::  
                   1691: * The optional Double Number word set::  
                   1692: * The optional Exception word set::  
                   1693: * The optional Facility word set::  
                   1694: * The optional File-Access word set::  
                   1695: * The optional Floating-Point word set::  
                   1696: * The optional Locals word set::  
                   1697: * The optional Memory-Allocation word set::  
                   1698: * The optional Programming-Tools word set::  
                   1699: * The optional Search-Order word set::  
                   1700: @end menu
                   1703: @c =====================================================================
                   1704: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
                   1705: @comment  node-name,  next,  previous,  up
                   1706: @section The Core Words
                   1707: @c =====================================================================
                   1709: @menu
1.15      anton    1710: * core-idef::                   Implementation Defined Options                   
                   1711: * core-ambcond::                Ambiguous Conditions                
                   1712: * core-other::                  Other System Documentation                  
1.14      anton    1713: @end menu
                   1715: @c ---------------------------------------------------------------------
                   1716: @node core-idef, core-ambcond, The Core Words, The Core Words
                   1717: @subsection Implementation Defined Options
                   1718: @c ---------------------------------------------------------------------
                   1720: @table @i
                   1722: @item (Cell) aligned addresses:
                   1723: processor-dependent. Gforths alignment words perform natural alignment
                   1724: (e.g., an address aligned for a datum of size 8 is divisible by
                   1725: 8). Unaligned accesses usually result in a @code{-23 THROW}.
                   1727: @item @code{EMIT} and non-graphic characters:
                   1728: The character is output using the C library function (actually, macro)
                   1729: @code{putchar}.
                   1731: @item character editing of @code{ACCEPT} and @code{EXPECT}:
                   1732: This is modeled on the GNU readline library (@pxref{Readline
                   1733: Interaction, , Command Line Editing, readline, The GNU Readline
                   1734: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
                   1735: producing a full word completion every time you type it (instead of
                   1736: producing the common prefix of all completions).
                   1738: @item character set:
                   1739: The character set of your computer and display device. Gforth is
                   1740: 8-bit-clean (but some other component in your system may make trouble).
                   1742: @item Character-aligned address requirements:
                   1743: installation-dependent. Currently a character is represented by a C
                   1744: @code{unsigned char}; in the future we might switch to @code{wchar_t}
                   1745: (Comments on that requested).
                   1747: @item character-set extensions and matching of names:
                   1748: Any character except 0 can be used in a name. Matching is
                   1749: case-insensitive. The matching is performed using the C function
                   1750: @code{strncasecmp}, whose function is probably influenced by the
                   1751: locale. E.g., the @code{C} locale does not know about accents and
                   1752: umlauts, so they are matched case-sensitively in that locale. For
                   1753: portability reasons it is best to write programs such that they work in
                   1754: the @code{C} locale. Then one can use libraries written by a Polish
                   1755: programmer (who might use words containing ISO Latin-2 encoded
                   1756: characters) and by a French programmer (ISO Latin-1) in the same program
                   1757: (of course, @code{WORDS} will produce funny results for some of the
                   1758: words (which ones, depends on the font you are using)). Also, the locale
                   1759: you prefer may not be available in other operating systems. Hopefully,
                   1760: Unicode will solve these problems one day.
                   1762: @item conditions under which control characters match a space delimiter:
                   1763: If @code{WORD} is called with the space character as a delimiter, all
                   1764: white-space characters (as identified by the C macro @code{isspace()})
                   1765: are delimiters. @code{PARSE}, on the other hand, treats space like other
                   1766: delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
                   1767: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
                   1768: interpreter (aka text interpreter) by default, treats all white-space
                   1769: characters as delimiters.
                   1771: @item format of the control flow stack:
                   1772: The data stack is used as control flow stack. The size of a control flow
                   1773: stack item in cells is given by the constant @code{cs-item-size}. At the
                   1774: time of this writing, an item consists of a (pointer to a) locals list
                   1775: (third), an address in the code (second), and a tag for identifying the
                   1776: item (TOS). The following tags are used: @code{defstart},
                   1777: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
                   1778: @code{scopestart}.
                   1780: @item conversion of digits > 35
                   1781: The characters @code{[\]^_'} are the digits with the decimal value
                   1782: 36@minus{}41. There is no way to input many of the larger digits.
                   1784: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
                   1785: The cursor is moved to the end of the entered string. If the input is
                   1786: terminated using the @kbd{Return} key, a space is typed.
                   1788: @item exception abort sequence of @code{ABORT"}:
                   1789: The error string is stored into the variable @code{"error} and a
                   1790: @code{-2 throw} is performed.
                   1792: @item input line terminator:
                   1793: For interactive input, @kbd{C-m} and @kbd{C-j} terminate lines. One of
                   1794: these characters is typically produced when you type the @kbd{Enter} or
                   1795: @kbd{Return} key.
                   1797: @item maximum size of a counted string:
                   1798: @code{s" /counted-string" environment? drop .}. Currently 255 characters
                   1799: on all ports, but this may change.
                   1801: @item maximum size of a parsed string:
                   1802: Given by the constant @code{/line}. Currently 255 characters.
                   1804: @item maximum size of a definition name, in characters:
                   1805: 31
                   1807: @item maximum string length for @code{ENVIRONMENT?}, in characters:
                   1808: 31
                   1810: @item method of selecting the user input device:
                   1811: The user input device is the standard input. There is current no way to
                   1812: change it from within gforth. However, the input can typically be
                   1813: redirected in the command line that starts gforth.
                   1815: @item method of selecting the user output device:
                   1816: The user output device is the standard output. It cannot be redirected
                   1817: from within gforth, but typically from the command line that starts
                   1818: gforth. Gforth uses buffered output, so output on a terminal does not
                   1819: become visible before the next newline or buffer overflow. Output on
                   1820: non-terminals is invisible until the buffer overflows.
                   1822: @item methods of dictionary compilation:
                   1823: Waht are we expected to document here?
                   1825: @item number of bits in one address unit:
                   1826: @code{s" address-units-bits" environment? drop .}. 8 in all current
                   1827: ports.
                   1829: @item number representation and arithmetic:
                   1830: Processor-dependent. Binary two's complement on all current ports.
                   1832: @item ranges for integer types:
                   1833: Installation-dependent. Make environmental queries for @code{MAX-N},
                   1834: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
                   1835: unsigned (and positive) types is 0. The lower bound for signed types on
                   1836: two's complement and one's complement machines machines can be computed
                   1837: by adding 1 to the upper bound.
                   1839: @item read-only data space regions:
                   1840: The whole Forth data space is writable.
                   1842: @item size of buffer at @code{WORD}:
                   1843: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   1844: shared with the pictured numeric output string. If overwriting
                   1845: @code{PAD} is acceptable, it is as large as the remaining dictionary
                   1846: space, although only as much can be sensibly used as fits in a counted
                   1847: string.
                   1849: @item size of one cell in address units:
                   1850: @code{1 cells .}.
                   1852: @item size of one character in address units:
                   1853: @code{1 chars .}. 1 on all current ports.
                   1855: @item size of the keyboard terminal buffer:
                   1856: Varies. You can determine the size at a specific time using @code{lp@
                   1857: tib - .}. It is shared with the locals stack and TIBs of files that
                   1858: include the current file. You can change the amount of space for TIBs
                   1859: and locals stack at gforth startup with the command line option
                   1860: @code{-l}.
                   1862: @item size of the pictured numeric output buffer:
                   1863: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   1864: shared with @code{WORD}.
                   1866: @item size of the scratch area returned by @code{PAD}:
                   1867: The remainder of dictionary space. You can even use the unused part of
                   1868: the data stack space. The current size can be computed with @code{sp@
                   1869: pad - .}.
                   1871: @item system case-sensitivity characteristics:
                   1872: Dictionary searches are case insensitive. However, as explained above
                   1873: under @i{character-set extensions}, the matching for non-ASCII
                   1874: characters is determined by the locale you are using. In the default
                   1875: @code{C} locale all non-ASCII characters are matched case-sensitively.
                   1877: @item system prompt:
                   1878: @code{ ok} in interpret state, @code{ compiled} in compile state.
                   1880: @item division rounding:
                   1881: installation dependent. @code{s" floored" environment? drop .}. We leave
                   1882: the choice to gcc (what to use for @code{/}) and to you (whether to use
                   1883: @code{fm/mod}, @code{sm/rem} or simply @code{/}).
                   1885: @item values of @code{STATE} when true:
                   1886: -1.
                   1888: @item values returned after arithmetic overflow:
                   1889: On two's complement machines, arithmetic is performed modulo
                   1890: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   1891: arithmetic (with appropriate mapping for signed types). Division by zero
                   1892: typically results in a @code{-55 throw} (floatingpoint unidentified
                   1893: fault), although a @code{-10 throw} (divide by zero) would be more
                   1894: appropriate.
                   1896: @item whether the current definition can be found after @t{DOES>}:
                   1897: No.
                   1899: @end table
                   1901: @c ---------------------------------------------------------------------
                   1902: @node core-ambcond, core-other, core-idef, The Core Words
                   1903: @subsection Ambiguous conditions
                   1904: @c ---------------------------------------------------------------------
                   1906: @table @i
                   1908: @item a name is neither a word nor a number:
                   1909: @code{-13 throw} (Undefined word)
                   1911: @item a definition name exceeds the maximum length allowed:
                   1912: @code{-19 throw} (Word name too long)
                   1914: @item addressing a region not inside the various data spaces of the forth system:
                   1915: The stacks, code space and name space are accessible. Machine code space is
                   1916: typically readable. Accessing other addresses gives results dependent on
                   1917: the operating system. On decent systems: @code{-9 throw} (Invalid memory
                   1918: address).
                   1920: @item argument type incompatible with parameter:
                   1921: This is usually not caught. Some words perform checks, e.g., the control
                   1922: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
                   1923: mismatch).
                   1925: @item attempting to obtain the execution token of a word with undefined execution semantics:
                   1926: You get an execution token representing the compilation semantics
                   1927: instead.
                   1929: @item dividing by zero:
                   1930: typically results in a @code{-55 throw} (floating point unidentified
                   1931: fault), although a @code{-10 throw} (divide by zero) would be more
                   1932: appropriate.
                   1934: @item insufficient data stack or return stack space:
                   1935: Not checked. This typically results in mysterious illegal memory
                   1936: accesses, producing @code{-9 throw} (Invalid memory address) or
                   1937: @code{-23 throw} (Address alignment exception).
                   1939: @item insufficient space for loop control parameters:
                   1940: like other return stack overflows.
                   1942: @item insufficient space in the dictionary:
                   1943: Not checked. Similar results as stack overflows. However, typically the
                   1944: error appears at a different place when one inserts or removes code.
                   1946: @item interpreting a word with undefined interpretation semantics:
                   1947: For some words, we defined interpretation semantics. For the others:
                   1948: @code{-14 throw} (Interpreting a compile-only word). Note that this is
                   1949: checked only by the outer (aka text) interpreter; if the word is
                   1950: @code{execute}d in some other way, it will typically perform it's
                   1951: compilation semantics even in interpret state. (We could change @code{'}
                   1952: and relatives not to give the xt of such words, but we think that would
                   1953: be too restrictive).
                   1955: @item modifying the contents of the input buffer or a string literal:
                   1956: These are located in writable memory and can be modified.
                   1958: @item overflow of the pictured numeric output string:
                   1959: Not checked.
                   1961: @item parsed string overflow:
                   1962: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
                   1964: @item producing a result out of range:
                   1965: On two's complement machines, arithmetic is performed modulo
                   1966: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   1967: arithmetic (with appropriate mapping for signed types). Division by zero
                   1968: typically results in a @code{-55 throw} (floatingpoint unidentified
                   1969: fault), although a @code{-10 throw} (divide by zero) would be more
                   1970: appropriate. @code{convert} and @code{>number} currently overflow
                   1971: silently.
                   1973: @item reading from an empty data or return stack:
                   1974: The data stack is checked by the outer (aka text) interpreter after
                   1975: every word executed. If it has underflowed, a @code{-4 throw} (Stack
                   1976: underflow) is performed. Apart from that, the stacks are not checked and
                   1977: underflows can result in similar behaviour as overflows (of adjacent
                   1978: stacks).
                   1980: @item unexepected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
                   1981: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
                   1982: use zero-length string as a name). Words like @code{'} probably will not
                   1983: find what they search. Note that it is possible to create zero-length
                   1984: names with @code{nextname} (should it not?).
                   1986: @item @code{>IN} greater than input buffer:
                   1987: The next invocation of a parsing word returns a string wih length 0.
                   1989: @item @code{RECURSE} appears after @code{DOES>}:
                   1990: Compiles a recursive call to the defining word not to the defined word.
                   1992: @item argument input source different than current input source for @code{RESTORE-INPUT}:
                   1993: !!???If the argument input source is a valid input source then it gets
                   1994: restored. Otherwise causes @code{-12 THROW} which unless caught issues
                   1995: the message "argument type mismatch" and aborts.
                   1997: @item data space containing definitions gets de-allocated:
                   1998: Deallocation with @code{allot} is not checked. This typically resuls in
                   1999: memory access faults or execution of illegal instructions.
                   2001: @item data space read/write with incorrect alignment:
                   2002: Processor-dependent. Typically results in a @code{-23 throw} (Address
                   2003: alignment exception). Under Linux on a 486 or later processor with
                   2004: alignment turned on, incorrect alignment results in a @code{-9 throw}
                   2005: (Invalid memory address). There are reportedly some processors with
                   2006: alignment restrictions that do not report them.
                   2008: @item data space pointer not properly aligned, @code{,}, @code{C,}:
                   2009: Like other alignment errors.
                   2011: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
                   2012: Not checked. May cause an illegal memory access.
                   2014: @item loop control parameters not available:
                   2015: Not checked. The counted loop words simply assume that the top of return
                   2016: stack items are loop control parameters and behave accordingly.
                   2018: @item most recent definition does not have a name (@code{IMMEDIATE}):
                   2019: @code{abort" last word was headerless"}.
                   2021: @item name not defined by @code{VALUE} used by @code{TO}:
                   2022: @code{-32 throw} (Invalid name argument)
1.15      anton    2024: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
1.14      anton    2025: @code{-13 throw} (Undefined word)
                   2027: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
                   2028: Gforth behaves as if they were of the same type. I.e., you can predict
                   2029: the behaviour by interpreting all parameters as, e.g., signed.
                   2031: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
                   2032: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} is equivalent to
                   2033: @code{TO}.
                   2035: @item String longer than a counted string returned by @code{WORD}:
                   2036: Not checked. The string will be ok, but the count will, of course,
                   2037: contain only the least significant bits of the length.
1.15      anton    2039: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
1.14      anton    2040: Processor-dependent. Typical behaviours are returning 0 and using only
                   2041: the low bits of the shift count.
                   2043: @item word not defined via @code{CREATE}:
                   2044: @code{>BODY} produces the PFA of the word no matter how it was defined.
                   2046: @code{DOES>} changes the execution semantics of the last defined word no
                   2047: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
                   2048: @code{CREATE , DOES>}.
                   2050: @item words improperly used outside @code{<#} and @code{#>}:
                   2051: Not checked. As usual, you can expect memory faults.
                   2053: @end table
                   2056: @c ---------------------------------------------------------------------
                   2057: @node core-other,  , core-ambcond, The Core Words
                   2058: @subsection Other system documentation
                   2059: @c ---------------------------------------------------------------------
                   2061: @table @i
                   2063: @item nonstandard words using @code{PAD}:
                   2064: None.
                   2066: @item operator's terminal facilities available:
                   2067: !!??
                   2069: @item program data space available:
                   2070: @code{sp@ here - .} gives the space remaining for dictionary and data
                   2071: stack together.
                   2073: @item return stack space available:
                   2074: !!??
                   2076: @item stack space available:
                   2077: @code{sp@ here - .} gives the space remaining for dictionary and data
                   2078: stack together.
                   2080: @item system dictionary space required, in address units:
                   2081: Type @code{here forthstart - .} after startup. At the time of this
                   2082: writing, this gives 70108 (bytes) on a 32-bit system.
                   2083: @end table
                   2086: @c =====================================================================
                   2087: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
                   2088: @section The optional Block word set
                   2089: @c =====================================================================
                   2091: @menu
1.15      anton    2092: * block-idef::                  Implementation Defined Options                  
                   2093: * block-ambcond::               Ambiguous Conditions               
                   2094: * block-other::                 Other System Documentation                 
1.14      anton    2095: @end menu
                   2098: @c ---------------------------------------------------------------------
                   2099: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
                   2100: @subsection Implementation Defined Options
                   2101: @c ---------------------------------------------------------------------
                   2103: @table @i
                   2105: @item the format for display by @code{LIST}:
                   2106: First the screen number is displayed, then 16 lines of 64 characters,
                   2107: each line preceded by the line number.
                   2109: @item the length of a line affected by @code{\}:
                   2110: 64 characters.
                   2111: @end table
                   2114: @c ---------------------------------------------------------------------
                   2115: @node block-ambcond, block-other, block-idef, The optional Block word set
                   2116: @subsection Ambiguous conditions
                   2117: @c ---------------------------------------------------------------------
                   2119: @table @i
                   2121: @item correct block read was not possible:
                   2122: Typically results in a @code{throw} of some OS-derived value (between
                   2123: -512 and -2048). If the blocks file was just not long enough, blanks are
                   2124: supplied for the missing portion.
                   2126: @item I/O exception in block transfer:
                   2127: Typically results in a @code{throw} of some OS-derived value (between
                   2128: -512 and -2048).
                   2130: @item invalid block number:
                   2131: @code{-35 throw} (Invalid block number)
                   2133: @item a program directly alters the contents of @code{BLK}:
                   2134: The input stream is switched to that other block, at the same
                   2135: position. If the storing to @code{BLK} happens when interpreting
                   2136: non-block input, the system will get quite confused when the block ends.
                   2138: @item no current block buffer for @code{UPDATE}:
                   2139: @code{UPDATE} has no effect.
                   2141: @end table
                   2144: @c ---------------------------------------------------------------------
                   2145: @node block-other,  , block-ambcond, The optional Block word set
                   2146: @subsection Other system documentation
                   2147: @c ---------------------------------------------------------------------
                   2149: @table @i
                   2151: @item any restrictions a multiprogramming system places on the use of buffer addresses:
                   2152: No restrictions (yet).
                   2154: @item the number of blocks available for source and data:
                   2155: depends on your disk space.
                   2157: @end table
                   2160: @c =====================================================================
                   2161: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
                   2162: @section The optional Double Number word set
                   2163: @c =====================================================================
                   2165: @menu
1.15      anton    2166: * double-ambcond::              Ambiguous Conditions              
1.14      anton    2167: @end menu
                   2170: @c ---------------------------------------------------------------------
1.15      anton    2171: @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
1.14      anton    2172: @subsection Ambiguous conditions
                   2173: @c ---------------------------------------------------------------------
                   2175: @table @i
1.15      anton    2177: @item @var{d} outside of range of @var{n} in @code{D>S}:
1.14      anton    2178: The least significant cell of @var{d} is produced.
                   2180: @end table
                   2183: @c =====================================================================
                   2184: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
                   2185: @section The optional Exception word set
                   2186: @c =====================================================================
                   2188: @menu
1.15      anton    2189: * exception-idef::              Implementation Defined Options              
1.14      anton    2190: @end menu
                   2193: @c ---------------------------------------------------------------------
1.15      anton    2194: @node exception-idef,  , The optional Exception word set, The optional Exception word set
1.14      anton    2195: @subsection Implementation Defined Options
                   2196: @c ---------------------------------------------------------------------
                   2198: @table @i
                   2199: @item @code{THROW}-codes used in the system:
                   2200: The codes -256@minus{}-511 are used for reporting signals (see
                   2201: @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
                   2202: (for file and memory allocation operations). The mapping from OS error
                   2203: numbers to throw code is -512@minus{}@var{errno}. One side effect of
                   2204: this mapping is that undefined OS errors produce a message with a
                   2205: strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
                   2206: 488} on my system.
                   2207: @end table
                   2209: @c =====================================================================
                   2210: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
                   2211: @section The optional Facility word set
                   2212: @c =====================================================================
                   2214: @menu
1.15      anton    2215: * facility-idef::               Implementation Defined Options               
                   2216: * facility-ambcond::            Ambiguous Conditions            
1.14      anton    2217: @end menu
                   2220: @c ---------------------------------------------------------------------
                   2221: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
                   2222: @subsection Implementation Defined Options
                   2223: @c ---------------------------------------------------------------------
                   2225: @table @i
                   2227: @item encoding of keyboard events (@code{EKEY}):
                   2228: Not yet implemeted.
                   2230: @item duration of a system clock tick
                   2231: System dependent. With respect to @code{MS}, the time is specified in
                   2232: microseconds. How well the OS and the hardware implement this, is
                   2233: another question.
                   2235: @item repeatability to be expected from the execution of @code{MS}:
                   2236: System dependent. On Unix, a lot depends on load. If the system is
                   2237: lightly loaded, and the delay is short enough that gforth does not get
                   2238: swapped out, the performance should be acceptable. Under MS-DOS and
                   2239: other single-tasking systems, it should be good.
                   2241: @end table
                   2244: @c ---------------------------------------------------------------------
1.15      anton    2245: @node facility-ambcond,  , facility-idef, The optional Facility word set
1.14      anton    2246: @subsection Ambiguous conditions
                   2247: @c ---------------------------------------------------------------------
                   2249: @table @i
                   2251: @item @code{AT-XY} can't be performed on user output device:
                   2252: Largely terminal dependant. No range checks are done on the arguments.
                   2253: No errors are reported. You may see some garbage appearing, you may see
                   2254: simply nothing happen.
                   2256: @end table
                   2259: @c =====================================================================
                   2260: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
                   2261: @section The optional File-Access word set
                   2262: @c =====================================================================
                   2264: @menu
1.15      anton    2265: * file-idef::                   Implementation Defined Options                   
                   2266: * file-ambcond::                Ambiguous Conditions                
1.14      anton    2267: @end menu
                   2270: @c ---------------------------------------------------------------------
                   2271: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
                   2272: @subsection Implementation Defined Options
                   2273: @c ---------------------------------------------------------------------
                   2275: @table @i
                   2277: @item File access methods used:
                   2278: @code{R/O}, @code{R/W} and @code{BIN} work as you would
                   2279: expect. @code{W/O} translates into the C file opening mode @code{w} (or
                   2280: @code{wb}): The file is cleared, if it exists, and created, if it does
1.15      anton    2281: not (both with @code{open-file} and @code{create-file}).  Under Unix
1.14      anton    2282: @code{create-file} creates a file with 666 permissions modified by your
                   2283: umask.
                   2285: @item file exceptions:
                   2286: The file words do not raise exceptions (except, perhaps, memory access
                   2287: faults when you pass illegal addresses or file-ids).
                   2289: @item file line terminator:
                   2290: System-dependent. Gforth uses C's newline character as line
                   2291: terminator. What the actual character code(s) of this are is
                   2292: system-dependent.
                   2294: @item file name format
                   2295: System dependent. Gforth just uses the file name format of your OS.
                   2297: @item information returned by @code{FILE-STATUS}:
                   2298: @code{FILE-STATUS} returns the most powerful file access mode allowed
                   2299: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
                   2300: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
                   2301: along with the retured mode.
                   2303: @item input file state after an exception when including source:
                   2304: All files that are left via the exception are closed.
                   2306: @item @var{ior} values and meaning:
1.15      anton    2307: The @var{ior}s returned by the file and memory allocation words are
                   2308: intended as throw codes. They typically are in the range
                   2309: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
                   2310: @var{ior}s is -512@minus{}@var{errno}.
1.14      anton    2311: 
                   2312: @item maximum depth of file input nesting:
                   2313: limited by the amount of return stack, locals/TIB stack, and the number
                   2314: of open files available. This should not give you troubles.
                   2316: @item maximum size of input line:
                   2317: @code{/line}. Currently 255.
                   2319: @item methods of mapping block ranges to files:
                   2320: Currently, the block words automatically access the file
                   2321: @file{blocks.fb} in the currend working directory. More sophisticated
                   2322: methods could be implemented if there is demand (and a volunteer).
                   2324: @item number of string buffers provided by @code{S"}:
                   2325: 1
                   2327: @item size of string buffer used by @code{S"}:
                   2328: @code{/line}. currently 255.
                   2330: @end table
                   2332: @c ---------------------------------------------------------------------
1.15      anton    2333: @node file-ambcond,  , file-idef, The optional File-Access word set
1.14      anton    2334: @subsection Ambiguous conditions
                   2335: @c ---------------------------------------------------------------------
                   2337: @table @i
                   2339: @item attempting to position a file outside it's boundaries:
                   2340: @code{REPOSITION-FILE} is performed as usual: Afterwards,
                   2341: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
                   2343: @item attempting to read from file positions not yet written:
                   2344: End-of-file, i.e., zero characters are read and no error is reported.
                   2346: @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
                   2347: An appropriate exception may be thrown, but a memory fault or other
                   2348: problem is more probable.
                   2350: @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
                   2351: The @var{ior} produced by the operation, that discovered the problem, is
                   2352: thrown.
                   2354: @item named file cannot be opened (@code{included}):
                   2355: The @var{ior} produced by @code{open-file} is thrown.
                   2357: @item requesting an unmapped block number:
                   2358: There are no unmapped legal block numbers. On some operating systems,
                   2359: writing a block with a large number may overflow the file system and
                   2360: have an error message as consequence.
                   2362: @item using @code{source-id} when @code{blk} is non-zero:
                   2363: @code{source-id} performs its function. Typically it will give the id of
                   2364: the source which loaded the block. (Better ideas?)
                   2366: @end table
                   2369: @c =====================================================================
                   2370: @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
1.15      anton    2371: @section The optional Floating-Point word set
1.14      anton    2372: @c =====================================================================
                   2374: @menu
1.15      anton    2375: * floating-idef::               Implementation Defined Options
                   2376: * floating-ambcond::            Ambiguous Conditions            
1.14      anton    2377: @end menu
                   2380: @c ---------------------------------------------------------------------
                   2381: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
                   2382: @subsection Implementation Defined Options
                   2383: @c ---------------------------------------------------------------------
                   2385: @table @i
1.15      anton    2387: @item format and range of floating point numbers:
                   2388: System-dependent; the @code{double} type of C.
1.14      anton    2389: 
1.15      anton    2390: @item results of @code{REPRESENT} when @var{float} is out of range:
                   2391: System dependent; @code{REPRESENT} is implemented using the C library
                   2392: function @code{ecvt()} and inherits its behaviour in this respect.
1.14      anton    2393: 
1.15      anton    2394: @item rounding or truncation of floating-point numbers:
                   2395: What's the question?!!
1.14      anton    2396: 
1.15      anton    2397: @item size of floating-point stack:
                   2398: @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
                   2399: with the command-line option @code{-f}.
1.14      anton    2400: 
1.15      anton    2401: @item width of floating-point stack:
                   2402: @code{1 floats}.
1.14      anton    2403: 
                   2404: @end table
                   2407: @c ---------------------------------------------------------------------
1.15      anton    2408: @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
                   2409: @subsection Ambiguous conditions
1.14      anton    2410: @c ---------------------------------------------------------------------
                   2412: @table @i
1.15      anton    2414: @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
                   2415: System-dependent. Typically results in an alignment fault like other
                   2416: alignment violations.
1.14      anton    2417: 
1.15      anton    2418: @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
                   2419: System-dependent. Typically results in an alignment fault like other
                   2420: alignment violations.
1.14      anton    2421: 
1.15      anton    2422: @item Floating-point result out of range:
                   2423: System-dependent. Can result in a @code{-55 THROW} (Floating-point
                   2424: unidentified fault), or can produce a special value representing, e.g.,
                   2425: Infinity.
1.14      anton    2426: 
1.15      anton    2427: @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
                   2428: System-dependent. Typically results in an alignment fault like other
                   2429: alignment violations.
1.14      anton    2430: 
1.15      anton    2431: @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
                   2432: The floating-point number is converted into decimal nonetheless.
1.14      anton    2433: 
1.15      anton    2434: @item Both arguments are equal to zero (@code{FATAN2}):
                   2435: System-dependent. @code{FATAN2} is implemented using the C library
                   2436: function @code{atan2()}.
1.14      anton    2437: 
1.15      anton    2438: @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
                   2439: System-dependent. Anyway, typically the cos of @var{r1} will not be zero
                   2440: because of small errors and the tan will be a very large (or very small)
                   2441: but finite number.
1.14      anton    2442: 
1.15      anton    2443: @item @var{d} cannot be presented precisely as a float in @code{D>F}:
                   2444: The result is rounded to the nearest float.
1.14      anton    2445: 
1.15      anton    2446: @item dividing by zero:
                   2447: @code{-55 throw} (Floating-point unidentified fault)
1.14      anton    2448: 
1.15      anton    2449: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
                   2450: System dependent. On IEEE-FP based systems the number is converted into
                   2451: an infinity.
1.14      anton    2452: 
1.15      anton    2453: @item @var{float}<1 (@code{facosh}):
                   2454: @code{-55 throw} (Floating-point unidentified fault)
1.14      anton    2455: 
1.15      anton    2456: @item @var{float}=<-1 (@code{flnp1}):
                   2457: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
                   2458: negative infinity is typically produced for @var{float}=-1.
1.14      anton    2459: 
1.15      anton    2460: @item @var{float}=<0 (@code{fln}, @code{flog}):
                   2461: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
                   2462: negative infinity is typically produced for @var{float}=0.
1.14      anton    2463: 
1.15      anton    2464: @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
                   2465: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
                   2466: produces values for these inputs on my Linux box (Bug in the C library?)
1.14      anton    2467: 
1.15      anton    2468: @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
                   2469: @code{-55 throw} (Floating-point unidentified fault).
1.14      anton    2470: 
1.15      anton    2471: @item integer part of float cannot be represented by @var{d} in @code{f>d}:
                   2472: @code{-55 throw} (Floating-point unidentified fault).
1.14      anton    2473: 
1.15      anton    2474: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
                   2475: This does not happen.
                   2476: @end table
1.14      anton    2477: 
                   2480: @c =====================================================================
1.15      anton    2481: @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
                   2482: @section The optional Locals word set
1.14      anton    2483: @c =====================================================================
                   2485: @menu
1.15      anton    2486: * locals-idef::                 Implementation Defined Options                 
                   2487: * locals-ambcond::              Ambiguous Conditions              
1.14      anton    2488: @end menu
                   2491: @c ---------------------------------------------------------------------
1.15      anton    2492: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
1.14      anton    2493: @subsection Implementation Defined Options
                   2494: @c ---------------------------------------------------------------------
                   2496: @table @i
1.15      anton    2498: @item maximum number of locals in a definition:
                   2499: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
                   2500: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
                   2501: characters. The number of locals in a definition is bounded by the size
                   2502: of locals-buffer, which contains the names of the locals.
1.14      anton    2503: 
                   2504: @end table
                   2507: @c ---------------------------------------------------------------------
1.15      anton    2508: @node locals-ambcond,  , locals-idef, The optional Locals word set
1.14      anton    2509: @subsection Ambiguous conditions
                   2510: @c ---------------------------------------------------------------------
                   2512: @table @i
1.15      anton    2514: @item executing a named local in interpretation state:
                   2515: @code{-14 throw} (Interpreting a compile-only word).
1.14      anton    2516: 
1.15      anton    2517: @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
                   2518: @code{-32 throw} (Invalid name argument)
1.14      anton    2519: 
                   2520: @end table
                   2523: @c =====================================================================
1.15      anton    2524: @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
                   2525: @section The optional Memory-Allocation word set
1.14      anton    2526: @c =====================================================================
                   2528: @menu
1.15      anton    2529: * memory-idef::                 Implementation Defined Options                 
1.14      anton    2530: @end menu
                   2533: @c ---------------------------------------------------------------------
1.15      anton    2534: @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
1.14      anton    2535: @subsection Implementation Defined Options
                   2536: @c ---------------------------------------------------------------------
                   2538: @table @i
1.15      anton    2540: @item values and meaning of @var{ior}:
                   2541: The @var{ior}s returned by the file and memory allocation words are
                   2542: intended as throw codes. They typically are in the range
                   2543: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
                   2544: @var{ior}s is -512@minus{}@var{errno}.
1.14      anton    2545: 
                   2546: @end table
                   2548: @c =====================================================================
1.15      anton    2549: @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
                   2550: @section The optional Programming-Tools word set
1.14      anton    2551: @c =====================================================================
                   2553: @menu
1.15      anton    2554: * programming-idef::            Implementation Defined Options            
                   2555: * programming-ambcond::         Ambiguous Conditions         
1.14      anton    2556: @end menu
                   2559: @c ---------------------------------------------------------------------
1.15      anton    2560: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
1.14      anton    2561: @subsection Implementation Defined Options
                   2562: @c ---------------------------------------------------------------------
                   2564: @table @i
1.15      anton    2566: @item ending sequence for input following @code{;code} and @code{code}:
                   2567: Not implemented (yet).
1.14      anton    2568: 
1.15      anton    2569: @item manner of processing input following @code{;code} and @code{code}:
                   2570: Not implemented (yet).
                   2572: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
                   2573: Not implemented (yet). If they were implemented, they would use the
                   2574: search order wordset.
                   2576: @item source and format of display by @code{SEE}:
                   2577: The source for @code{see} is the intermediate code used by the inner
                   2578: interpreter.  The current @code{see} tries to output Forth source code
                   2579: as well as possible.
1.14      anton    2581: @end table
                   2583: @c ---------------------------------------------------------------------
1.15      anton    2584: @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
1.14      anton    2585: @subsection Ambiguous conditions
                   2586: @c ---------------------------------------------------------------------
                   2588: @table @i
1.15      anton    2590: @item deleting the compilation wordlist (@code{FORGET}):
                   2591: Not implemented (yet).
1.14      anton    2592: 
1.15      anton    2593: @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
                   2594: This typically results in an @code{abort"} with a descriptive error
                   2595: message (may change into a @code{-22 throw} (Control structure mismatch)
                   2596: in the future). You may also get a memory access error. If you are
                   2597: unlucky, this ambiguous condition is not caught.
                   2599: @item @var{name} can't be found (@code{forget}):
                   2600: Not implemented (yet).
1.14      anton    2601: 
1.15      anton    2602: @item @var{name} not defined via @code{CREATE}:
                   2603: @code{;code} is not implemented (yet). If it were, it would behave like
                   2604: @code{DOES>} in this respect, i.e., change the execution semantics of
                   2605: the last defined word no matter how it was defined.
1.14      anton    2606: 
1.15      anton    2607: @item @code{POSTPONE} applied to @code{[IF]}:
                   2608: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
                   2609: equivalent to @code{[IF]}.
1.14      anton    2610: 
1.15      anton    2611: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
                   2612: Continue in the same state of conditional compilation in the next outer
                   2613: input source. Currently there is no warning to the user about this.
1.14      anton    2614: 
1.15      anton    2615: @item removing a needed definition (@code{FORGET}):
                   2616: Not implemented (yet).
1.14      anton    2617: 
                   2618: @end table
                   2621: @c =====================================================================
1.15      anton    2622: @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
                   2623: @section The optional Search-Order word set
1.14      anton    2624: @c =====================================================================
                   2626: @menu
1.15      anton    2627: * search-idef::                 Implementation Defined Options                 
                   2628: * search-ambcond::              Ambiguous Conditions              
1.14      anton    2629: @end menu
                   2632: @c ---------------------------------------------------------------------
1.15      anton    2633: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
1.14      anton    2634: @subsection Implementation Defined Options
                   2635: @c ---------------------------------------------------------------------
                   2637: @table @i
1.15      anton    2639: @item maximum number of word lists in search order:
                   2640: @code{s" wordlists" environment? drop .}. Currently 16.
                   2642: @item minimum search order:
                   2643: @code{root root}.
1.14      anton    2644: 
                   2645: @end table
                   2647: @c ---------------------------------------------------------------------
1.15      anton    2648: @node search-ambcond,  , search-idef, The optional Search-Order word set
1.14      anton    2649: @subsection Ambiguous conditions
                   2650: @c ---------------------------------------------------------------------
                   2652: @table @i
1.15      anton    2654: @item changing the compilation wordlist (during compilation):
                   2655: The definition is put into the wordlist that is the compilation wordlist
                   2656: when @code{REVEAL} is executed (by @code{;}, @code{DOES>},
                   2657: @code{RECURSIVE}, etc.).
1.14      anton    2658: 
1.15      anton    2659: @item search order empty (@code{previous}):
                   2660: @code{abort" Vocstack empty"}.
1.14      anton    2661: 
1.15      anton    2662: @item too many word lists in search order (@code{also}):
                   2663: @code{abort" Vocstack full"}.
1.14      anton    2664: 
                   2665: @end table
1.13      anton    2666: 
1.4       anton    2668: @node Model, Emacs and GForth, ANS conformance, Top
                   2669: @chapter Model
                   2671: @node Emacs and GForth, Internals, Model, Top
                   2672: @chapter Emacs and GForth
                   2674: GForth comes with @file{gforth.el}, an improved version of
                   2675: @file{forth.el} by Goran Rydqvist (icluded in the TILE package). The
                   2676: improvements are a better (but still not perfect) handling of
                   2677: indentation. I have also added comment paragraph filling (@kbd{M-q}),
1.8       anton    2678: commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
                   2679: removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
                   2680: stuff I do not use alone, even though some of it only makes sense for
                   2681: TILE. To get a description of these features, enter Forth mode and type
                   2682: @kbd{C-h m}.
1.4       anton    2683: 
                   2684: In addition, GForth supports Emacs quite well: The source code locations
                   2685: given in error messages, debugging output (from @code{~~}) and failed
                   2686: assertion messages are in the right format for Emacs' compilation mode
                   2687: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
                   2688: Manual}) so the source location corresponding to an error or other
                   2689: message is only a few keystrokes away (@kbd{C-x `} for the next error,
                   2690: @kbd{C-c C-c} for the error under the cursor).
                   2692: Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
                   2693: (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
                   2694: contains the definitions of all words defined afterwards. You can then
                   2695: find the source for a word using @kbd{M-.}. Note that emacs can use
                   2696: several tags files at the same time (e.g., one for the gforth sources
                   2697: and one for your program).
                   2699: To get all these benefits, add the following lines to your @file{.emacs}
                   2700: file:
                   2702: @example
                   2703: (autoload 'forth-mode "gforth.el")
                   2704: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
                   2705: @end example
                   2707: @node Internals, Bugs, Emacs and GForth, Top
1.3       anton    2708: @chapter Internals
                   2710: Reading this section is not necessary for programming with gforth. It
                   2711: should be helpful for finding your way in the gforth sources.
1.4       anton    2713: @menu
                   2714: * Portability::                 
                   2715: * Threading::                   
                   2716: * Primitives::                  
                   2717: * System Architecture::         
                   2718: @end menu
                   2720: @node Portability, Threading, Internals, Internals
1.3       anton    2721: @section Portability
                   2723: One of the main goals of the effort is availability across a wide range
                   2724: of personal machines. fig-Forth, and, to a lesser extent, F83, achieved
                   2725: this goal by manually coding the engine in assembly language for several
                   2726: then-popular processors. This approach is very labor-intensive and the
                   2727: results are short-lived due to progress in computer architecture.
                   2729: Others have avoided this problem by coding in C, e.g., Mitch Bradley
                   2730: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
                   2731: particularly popular for UNIX-based Forths due to the large variety of
                   2732: architectures of UNIX machines. Unfortunately an implementation in C
                   2733: does not mix well with the goals of efficiency and with using
                   2734: traditional techniques: Indirect or direct threading cannot be expressed
                   2735: in C, and switch threading, the fastest technique available in C, is
                   2736: significantly slower. Another problem with C is that it's very
                   2737: cumbersome to express double integer arithmetic.
                   2739: Fortunately, there is a portable language that does not have these
                   2740: limitations: GNU C, the version of C processed by the GNU C compiler
                   2741: (@pxref{C Extensions, , Extensions to the C Language Family,,
                   2742: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
                   2743: Labels as Values,, GNU C Manual}) makes direct and indirect
                   2744: threading possible, its @code{long long} type (@pxref{Long Long, ,
                   2745: Double-Word Integers,, GNU C Manual}) corresponds to Forths
                   2746: double numbers. GNU C is available for free on all important (and many
                   2747: unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and
                   2748: the Atari ST, so a Forth written in GNU C can run on all these
                   2749: machines@footnote{Due to Apple's look-and-feel lawsuit it is not
1.5       anton    2750: available on the Mac (@pxref{Boycott, , Protect Your Freedom---Fight
1.3       anton    2751: ``Look And Feel'',, GNU C Manual}).}.
                   2753: Writing in a portable language has the reputation of producing code that
                   2754: is slower than assembly. For our Forth engine we repeatedly looked at
                   2755: the code produced by the compiler and eliminated most compiler-induced
                   2756: inefficiencies by appropriate changes in the source-code.
                   2758: However, register allocation cannot be portably influenced by the
                   2759: programmer, leading to some inefficiencies on register-starved
                   2760: machines. We use explicit register declarations (@pxref{Explicit Reg
                   2761: Vars, , Variables in Specified Registers,, GNU C Manual}) to
                   2762: improve the speed on some machines. They are turned on by using the
                   2763: @code{gcc} switch @code{-DFORCE_REG}. Unfortunately, this feature not
                   2764: only depends on the machine, but also on the compiler version: On some
                   2765: machines some compiler versions produce incorrect code when certain
                   2766: explicit register declarations are used. So by default
                   2767: @code{-DFORCE_REG} is not used.
1.4       anton    2769: @node Threading, Primitives, Portability, Internals
1.3       anton    2770: @section Threading
                   2772: GNU C's labels as values extension (available since @code{gcc-2.0},
                   2773: @pxref{Labels as Values, , Labels as Values,, GNU C Manual})
                   2774: makes it possible to take the address of @var{label} by writing
                   2775: @code{&&@var{label}}.  This address can then be used in a statement like
                   2776: @code{goto *@var{address}}. I.e., @code{goto *&&x} is the same as
                   2777: @code{goto x}.
                   2779: With this feature an indirect threaded NEXT looks like:
                   2780: @example
                   2781: cfa = *ip++;
                   2782: ca = *cfa;
                   2783: goto *ca;
                   2784: @end example
                   2785: For those unfamiliar with the names: @code{ip} is the Forth instruction
                   2786: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
                   2787: execution token and points to the code field of the next word to be
                   2788: executed; The @code{ca} (code address) fetched from there points to some
                   2789: executable code, e.g., a primitive or the colon definition handler
                   2790: @code{docol}.
                   2792: Direct threading is even simpler:
                   2793: @example
                   2794: ca = *ip++;
                   2795: goto *ca;
                   2796: @end example
                   2798: Of course we have packaged the whole thing neatly in macros called
                   2799: @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
1.4       anton    2801: @menu
                   2802: * Scheduling::                  
                   2803: * Direct or Indirect Threaded?::  
                   2804: * DOES>::                       
                   2805: @end menu
                   2807: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
1.3       anton    2808: @subsection Scheduling
                   2810: There is a little complication: Pipelined and superscalar processors,
                   2811: i.e., RISC and some modern CISC machines can process independent
                   2812: instructions while waiting for the results of an instruction. The
                   2813: compiler usually reorders (schedules) the instructions in a way that
                   2814: achieves good usage of these delay slots. However, on our first tries
                   2815: the compiler did not do well on scheduling primitives. E.g., for
                   2816: @code{+} implemented as
                   2817: @example
                   2818: n=sp[0]+sp[1];
                   2819: sp++;
                   2820: sp[0]=n;
                   2821: NEXT;
                   2822: @end example
                   2823: the NEXT comes strictly after the other code, i.e., there is nearly no
                   2824: scheduling. After a little thought the problem becomes clear: The
                   2825: compiler cannot know that sp and ip point to different addresses (and
1.4       anton    2826: the version of @code{gcc} we used would not know it even if it was
                   2827: possible), so it could not move the load of the cfa above the store to
                   2828: the TOS. Indeed the pointers could be the same, if code on or very near
                   2829: the top of stack were executed. In the interest of speed we chose to
                   2830: forbid this probably unused ``feature'' and helped the compiler in
                   2831: scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
                   2832: the goto part (@code{NEXT_P2}). @code{+} now looks like:
1.3       anton    2833: @example
                   2834: n=sp[0]+sp[1];
                   2835: sp++;
                   2836: NEXT_P1;
                   2837: sp[0]=n;
                   2838: NEXT_P2;
                   2839: @end example
1.4       anton    2840: This can be scheduled optimally by the compiler.
1.3       anton    2841: 
                   2842: This division can be turned off with the switch @code{-DCISC_NEXT}. This
                   2843: switch is on by default on machines that do not profit from scheduling
                   2844: (e.g., the 80386), in order to preserve registers.
1.4       anton    2846: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
1.3       anton    2847: @subsection Direct or Indirect Threaded?
                   2849: Both! After packaging the nasty details in macro definitions we
                   2850: realized that we could switch between direct and indirect threading by
                   2851: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
                   2852: defining a few machine-specific macros for the direct-threading case.
                   2853: On the Forth level we also offer access words that hide the
                   2854: differences between the threading methods (@pxref{Threading Words}).
                   2856: Indirect threading is implemented completely
                   2857: machine-independently. Direct threading needs routines for creating
                   2858: jumps to the executable code (e.g. to docol or dodoes). These routines
                   2859: are inherently machine-dependent, but they do not amount to many source
                   2860: lines. I.e., even porting direct threading to a new machine is a small
                   2861: effort.
1.4       anton    2863: @node DOES>,  , Direct or Indirect Threaded?, Threading
1.3       anton    2864: @subsection DOES>
                   2865: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
                   2866: the chunk of code executed by every word defined by a
                   2867: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
                   2868: the Forth code to be executed, i.e. the code after the @code{DOES>} (the
                   2869: DOES-code)? There are two solutions:
                   2871: In fig-Forth the code field points directly to the dodoes and the
                   2872: DOES-code address is stored in the cell after the code address
                   2873: (i.e. at cfa cell+). It may seem that this solution is illegal in the
                   2874: Forth-79 and all later standards, because in fig-Forth this address
                   2875: lies in the body (which is illegal in these standards). However, by
                   2876: making the code field larger for all words this solution becomes legal
                   2877: again. We use this approach for the indirect threaded version. Leaving
                   2878: a cell unused in most words is a bit wasteful, but on the machines we
                   2879: are targetting this is hardly a problem. The other reason for having a
                   2880: code field size of two cells is to avoid having different image files
1.4       anton    2881: for direct and indirect threaded systems (@pxref{System Architecture}).
1.3       anton    2882: 
                   2883: The other approach is that the code field points or jumps to the cell
                   2884: after @code{DOES}. In this variant there is a jump to @code{dodoes} at
                   2885: this address. @code{dodoes} can then get the DOES-code address by
                   2886: computing the code address, i.e., the address of the jump to dodoes,
                   2887: and add the length of that jump field. A variant of this is to have a
                   2888: call to @code{dodoes} after the @code{DOES>}; then the return address
                   2889: (which can be found in the return register on RISCs) is the DOES-code
                   2890: address. Since the two cells available in the code field are usually
                   2891: used up by the jump to the code address in direct threading, we use
                   2892: this approach for direct threading. We did not want to add another
                   2893: cell to the code field.
1.4       anton    2895: @node Primitives, System Architecture, Threading, Internals
1.3       anton    2896: @section Primitives
1.4       anton    2898: @menu
                   2899: * Automatic Generation::        
                   2900: * TOS Optimization::            
                   2901: * Produced code::               
                   2902: @end menu
                   2904: @node Automatic Generation, TOS Optimization, Primitives, Primitives
1.3       anton    2905: @subsection Automatic Generation
                   2907: Since the primitives are implemented in a portable language, there is no
                   2908: longer any need to minimize the number of primitives. On the contrary,
                   2909: having many primitives is an advantage: speed. In order to reduce the
                   2910: number of errors in primitives and to make programming them easier, we
                   2911: provide a tool, the primitive generator (@file{prims2x.fs}), that
                   2912: automatically generates most (and sometimes all) of the C code for a
                   2913: primitive from the stack effect notation.  The source for a primitive
                   2914: has the following form:
                   2916: @format
                   2917: @var{Forth-name}       @var{stack-effect}      @var{category}  [@var{pronounc.}]
                   2918: [@code{""}@var{glossary entry}@code{""}]
                   2919: @var{C code}
                   2920: [@code{:}
                   2921: @var{Forth code}]
                   2922: @end format
                   2924: The items in brackets are optional. The category and glossary fields
                   2925: are there for generating the documentation, the Forth code is there
                   2926: for manual implementations on machines without GNU C. E.g., the source
                   2927: for the primitive @code{+} is:
                   2928: @example
                   2929: +    n1 n2 -- n    core    plus
                   2930: n = n1+n2;
                   2931: @end example
                   2933: This looks like a specification, but in fact @code{n = n1+n2} is C
                   2934: code. Our primitive generation tool extracts a lot of information from
                   2935: the stack effect notations@footnote{We use a one-stack notation, even
                   2936: though we have separate data and floating-point stacks; The separate
                   2937: notation can be generated easily from the unified notation.}: The number
                   2938: of items popped from and pushed on the stack, their type, and by what
                   2939: name they are referred to in the C code. It then generates a C code
                   2940: prelude and postlude for each primitive. The final C code for @code{+}
                   2941: looks like this:
                   2943: @example
                   2944: I_plus:        /* + ( n1 n2 -- n ) */  /* label, stack effect */
                   2945: /*  */                          /* documentation */
1.4       anton    2946: @{
1.3       anton    2947: DEF_CA                          /* definition of variable ca (indirect threading) */
                   2948: Cell n1;                        /* definitions of variables */
                   2949: Cell n2;
                   2950: Cell n;
                   2951: n1 = (Cell) sp[1];              /* input */
                   2952: n2 = (Cell) TOS;
                   2953: sp += 1;                        /* stack adjustment */
                   2954: NAME("+")                       /* debugging output (with -DDEBUG) */
1.4       anton    2955: @{
1.3       anton    2956: n = n1+n2;                      /* C code taken from the source */
1.4       anton    2957: @}
1.3       anton    2958: NEXT_P1;                        /* NEXT part 1 */
                   2959: TOS = (Cell)n;                  /* output */
                   2960: NEXT_P2;                        /* NEXT part 2 */
1.4       anton    2961: @}
1.3       anton    2962: @end example
                   2964: This looks long and inefficient, but the GNU C compiler optimizes quite
                   2965: well and produces optimal code for @code{+} on, e.g., the R3000 and the
                   2966: HP RISC machines: Defining the @code{n}s does not produce any code, and
                   2967: using them as intermediate storage also adds no cost.
                   2969: There are also other optimizations, that are not illustrated by this
                   2970: example: Assignments between simple variables are usually for free (copy
                   2971: propagation). If one of the stack items is not used by the primitive
                   2972: (e.g.  in @code{drop}), the compiler eliminates the load from the stack
                   2973: (dead code elimination). On the other hand, there are some things that
                   2974: the compiler does not do, therefore they are performed by
                   2975: @file{prims2x.fs}: The compiler does not optimize code away that stores
                   2976: a stack item to the place where it just came from (e.g., @code{over}).
                   2978: While programming a primitive is usually easy, there are a few cases
                   2979: where the programmer has to take the actions of the generator into
                   2980: account, most notably @code{?dup}, but also words that do not (always)
                   2981: fall through to NEXT.
1.4       anton    2983: @node TOS Optimization, Produced code, Automatic Generation, Primitives
1.3       anton    2984: @subsection TOS Optimization
                   2986: An important optimization for stack machine emulators, e.g., Forth
                   2987: engines, is keeping  one or more of the top stack items in
1.4       anton    2988: registers.  If a word has the stack effect @var{in1}...@var{inx} @code{--}
                   2989: @var{out1}...@var{outy}, keeping the top @var{n} items in registers
1.3       anton    2990: @itemize
                   2991: @item
                   2992: is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
                   2993: due to fewer loads from and stores to the stack.
                   2994: @item is slower than keeping @var{n-1} items, if @var{x<>y} and @var{x<n} and
                   2995: @var{y<n}, due to additional moves between registers.
                   2996: @end itemize
                   2998: In particular, keeping one item in a register is never a disadvantage,
                   2999: if there are enough registers. Keeping two items in registers is a
                   3000: disadvantage for frequent words like @code{?branch}, constants,
                   3001: variables, literals and @code{i}. Therefore our generator only produces
                   3002: code that keeps zero or one items in registers. The generated C code
                   3003: covers both cases; the selection between these alternatives is made at
                   3004: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
                   3005: code for @code{+} is just a simple variable name in the one-item case,
                   3006: otherwise it is a macro that expands into @code{sp[0]}. Note that the
                   3007: GNU C compiler tries to keep simple variables like @code{TOS} in
                   3008: registers, and it usually succeeds, if there are enough registers.
                   3010: The primitive generator performs the TOS optimization for the
                   3011: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
                   3012: operations the benefit of this optimization is even larger:
                   3013: floating-point operations take quite long on most processors, but can be
                   3014: performed in parallel with other operations as long as their results are
                   3015: not used. If the FP-TOS is kept in a register, this works. If
                   3016: it is kept on the stack, i.e., in memory, the store into memory has to
                   3017: wait for the result of the floating-point operation, lengthening the
                   3018: execution time of the primitive considerably.
                   3020: The TOS optimization makes the automatic generation of primitives a
                   3021: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
                   3022: @code{TOS} is not sufficient. There are some special cases to
                   3023: consider:
                   3024: @itemize
                   3025: @item In the case of @code{dup ( w -- w w )} the generator must not
                   3026: eliminate the store to the original location of the item on the stack,
                   3027: if the TOS optimization is turned on.
1.4       anton    3028: @item Primitives with stack effects of the form @code{--}
                   3029: @var{out1}...@var{outy} must store the TOS to the stack at the start.
                   3030: Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
1.3       anton    3031: must load the TOS from the stack at the end. But for the null stack
                   3032: effect @code{--} no stores or loads should be generated.
                   3033: @end itemize
1.4       anton    3035: @node Produced code,  , TOS Optimization, Primitives
1.3       anton    3036: @subsection Produced code
                   3038: To see what assembly code is produced for the primitives on your machine
                   3039: with your compiler and your flag settings, type @code{make engine.s} and
1.4       anton    3040: look at the resulting file @file{engine.s}.
1.3       anton    3041: 
1.4       anton    3042: @node System Architecture,  , Primitives, Internals
1.3       anton    3043: @section System Architecture
                   3045: Our Forth system consists not only of primitives, but also of
                   3046: definitions written in Forth. Since the Forth compiler itself belongs
                   3047: to those definitions, it is not possible to start the system with the
                   3048: primitives and the Forth source alone. Therefore we provide the Forth
                   3049: code as an image file in nearly executable form. At the start of the
                   3050: system a C routine loads the image file into memory, sets up the
                   3051: memory (stacks etc.) according to information in the image file, and
                   3052: starts executing Forth code.
                   3054: The image file format is a compromise between the goals of making it
                   3055: easy to generate image files and making them portable. The easiest way
                   3056: to generate an image file is to just generate a memory dump. However,
                   3057: this kind of image file cannot be used on a different machine, or on
                   3058: the next version of the engine on the same machine, it even might not
                   3059: work with the same engine compiled by a different version of the C
                   3060: compiler. We would like to have as few versions of the image file as
                   3061: possible, because we do not want to distribute many versions of the
                   3062: same image file, and to make it easy for the users to use their image
                   3063: files on many machines. We currently need to create a different image
                   3064: file for machines with different cell sizes and different byte order
                   3065: (little- or big-endian)@footnote{We consider adding information to the
                   3066: image file that enables the loader to change the byte order.}.
                   3068: Forth code that is going to end up in a portable image file has to
1.4       anton    3069: comply to some restrictions: addresses have to be stored in memory with
                   3070: special words (@code{A!}, @code{A,}, etc.) in order to make the code
                   3071: relocatable. Cells, floats, etc., have to be stored at the natural
                   3072: alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
                   3073: dividable by~8. This happens automatically in our system when you use
                   3074: the ANS Forth alignment words.}, in order to avoid alignment faults on
                   3075: machines with stricter alignment. The image file is produced by a
                   3076: metacompiler (@file{cross.fs}).
1.3       anton    3077: 
                   3078: So, unlike the image file of Mitch Bradleys @code{cforth}, our image
                   3079: file is not directly executable, but has to undergo some manipulations
                   3080: during loading. Address relocation is performed at image load-time, not
                   3081: at run-time. The loader also has to replace tokens standing for
                   3082: primitive calls with the appropriate code-field addresses (or code
                   3083: addresses in the case of direct threading).
1.4       anton    3084: 
                   3085: @node Bugs, Pedigree, Internals, Top
                   3086: @chapter Bugs
                   3088: @node Pedigree, Word Index, Bugs, Top
                   3089: @chapter Pedigree
                   3091: @node Word Index, Node Index, Pedigree, Top
                   3092: @chapter Word Index
                   3094: @node Node Index,  , Word Index, Top
                   3095: @chapter Node Index
1.1       anton    3096: 
                   3097: @contents
                   3098: @bye

FreeBSD-CVSweb <>