Diff for /gforth/Attic/gforth.ds between versions 1.3 and 1.15

version 1.3, 1994/11/23 16:54:39 version 1.15, 1995/04/29 14:51:16
Line 1 Line 1
 \input texinfo   @c -*-texinfo-*-  \input texinfo   @c -*-texinfo-*-
 @comment The source is gforth.ds, from which gforth.texi is generated  @comment The source is gforth.ds, from which gforth.texi is generated
 @comment %**start of header (This is for running Texinfo on a region.)  @comment %**start of header (This is for running Texinfo on a region.)
 @setfilename gforth-info  @setfilename gforth.info
 @settitle GNU Forth Manual  @settitle GNU Forth Manual
 @setchapternewpage odd  @comment @setchapternewpage odd
 @comment %**end of header (This is for running Texinfo on a region.)  @comment %**end of header (This is for running Texinfo on a region.)
   
 @ifinfo  @ifinfo
Line 15  Copyright @copyright{} 1994 GNU Forth De Line 15  Copyright @copyright{} 1994 GNU Forth De
      this manual provided the copyright notice and this permission notice       this manual provided the copyright notice and this permission notice
      are preserved on all copies.       are preserved on all copies.
             
      @ignore  @ignore
      Permission is granted to process this file through TeX and print the       Permission is granted to process this file through TeX and print the
      results, provided the printed document carries a copying permission       results, provided the printed document carries a copying permission
      notice identical to this one except for the removal of this paragraph       notice identical to this one except for the removal of this paragraph
      (this paragraph not being relevant to the printed manual).       (this paragraph not being relevant to the printed manual).
             
      @end ignore  @end ignore
      Permission is granted to copy and distribute modified versions of this       Permission is granted to copy and distribute modified versions of this
      manual under the conditions for verbatim copying, provided also that the       manual under the conditions for verbatim copying, provided also that the
      sections entitled "Distribution" and "General Public License" are       sections entitled "Distribution" and "General Public License" are
Line 77  personal machines. This manual correspon Line 77  personal machines. This manual correspon
 @end ifinfo  @end ifinfo
   
 @menu  @menu
 * License::               * License::                     
 * Goals::               About the GNU Forth Project  * Goals::                       About the GNU Forth Project
 * Other Books::         Things you might want to read  * Other Books::                 Things you might want to read
 * Invocation::          Starting GNU Forth  * Invocation::                  Starting GNU Forth
 * Words::               Forth words available in GNU Forth  * Words::                       Forth words available in GNU Forth
 * ANS conformance::     Implementation-defined options etc.  * ANS conformance::             Implementation-defined options etc.
 * Model::               The abstract machine of GNU Forth  * Model::                       The abstract machine of GNU Forth
 * Emacs and GForth::    The GForth Mode  * Emacs and GForth::            The GForth Mode
 * Internals::           Implementation details  * Internals::                   Implementation details
 * Bugs::                How to report them  * Bugs::                        How to report them
 * Pedigree::            Ancestors of GNU Forth  * Pedigree::                    Ancestors of GNU Forth
 * Word Index::          An item for each Forth word  * Word Index::                  An item for each Forth word
 * Node Index::          An item for each node  * Node Index::                  An item for each node
 @end menu  @end menu
   
 @node License, Goals, Top, Top  @node License, Goals, Top, Top
Line 253  the user initialization file @file{.gfor Line 253  the user initialization file @file{.gfor
 option @code{--no-rc} is given; this file is first searched in @file{.},  option @code{--no-rc} is given; this file is first searched in @file{.},
 then in @file{~}, then in the normal path (see above).  then in @file{~}, then in the normal path (see above).
   
 @node Words,  , Invocation, Top  @node Words, ANS conformance, Invocation, Top
 @chapter Forth Words  @chapter Forth Words
   
 @menu  @menu
 * Notation::  * Notation::                    
 * Arithmetic::  * Arithmetic::                  
 * Stack Manipulation::  * Stack Manipulation::          
 * Memory access::  * Memory access::               
 * Control Structures::  * Control Structures::          
 * Local Variables::  * Locals::                      
 * Defining Words::  * Defining Words::              
 * Vocabularies::  * Wordlists::                   
 * Files::  * Files::                       
 * Blocks::  * Blocks::                      
 * Other I/O::  * Other I/O::                   
 * Programming Tools::  * Programming Tools::           
   * Threading Words::             
 @end menu  @end menu
   
 @node Notation, Arithmetic, Words, Words  @node Notation, Arithmetic, Words, Words
Line 277  then in @file{~}, then in the normal pat Line 278  then in @file{~}, then in the normal pat
 The Forth words are described in this section in the glossary notation  The Forth words are described in this section in the glossary notation
 that has become a de-facto standard for Forth texts, i.e.  that has become a de-facto standard for Forth texts, i.e.
   
 @quotation  @format
 @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}  @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}
   @end format
 @var{Description}  @var{Description}
 @end quotation  
   
 @table @var  @table @var
 @item word  @item word
 The name of the word. BTW, GNU Forth is case insensitive, so you can  The name of the word. BTW, GNU Forth is case insensitive, so you can
 type the words in in lower case.  type the words in in lower case (However, @pxref{core-idef}).
   
 @item Stack effect  @item Stack effect
 The stack effect is written in the notation @code{@var{before} --  The stack effect is written in the notation @code{@var{before} --
Line 314  wordsets. Words that are not defined in Line 315  wordsets. Words that are not defined in
 A description of the behaviour of the word.  A description of the behaviour of the word.
 @end table  @end table
   
 The name of a stack item corresponds in the following way with its type:  The type of a stack item is specified by the character(s) the name
   starts with:
   
 @table @code  @table @code
 @item name starts with  
 Type  
 @item f  @item f
 Bool, i.e. @code{false} or @code{true}.  Bool, i.e. @code{false} or @code{true}.
 @item c  @item c
Line 353  Wordlist ID, same size as Cell Line 353  Wordlist ID, same size as Cell
 Pointer to a name structure  Pointer to a name structure
 @end table  @end table
   
 @node Arithmetic,  , Notation, Words  @node Arithmetic, Stack Manipulation, Notation, Words
 @section Arithmetic  @section Arithmetic
 Forth arithmetic is not checked, i.e., you will not hear about integer  Forth arithmetic is not checked, i.e., you will not hear about integer
 overflow on addition or multiplication, you may hear about division by  overflow on addition or multiplication, you may hear about division by
Line 363  corresponds to @code{2 1 -}. Forth offer Line 363  corresponds to @code{2 1 -}. Forth offer
 operators. If you perform division with potentially negative operands,  operators. If you perform division with potentially negative operands,
 you do not want to use @code{/} or @code{/mod} with its undefined  you do not want to use @code{/} or @code{/mod} with its undefined
 behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the  behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
 former).  former, @pxref{Mixed precision}).
   
   @menu
   * Single precision::            
   * Bitwise operations::          
   * Mixed precision::             operations with single and double-cell integers
   * Double precision::            Double-cell integer arithmetic
   * Floating Point::              
   @end menu
   
   @node Single precision, Bitwise operations, Arithmetic, Arithmetic
 @subsection Single precision  @subsection Single precision
 doc-+  doc-+
 doc--  doc--
Line 377  doc-abs Line 386  doc-abs
 doc-min  doc-min
 doc-max  doc-max
   
   @node Bitwise operations, Mixed precision, Single precision, Arithmetic
 @subsection Bitwise operations  @subsection Bitwise operations
 doc-and  doc-and
 doc-or  doc-or
Line 385  doc-invert Line 395  doc-invert
 doc-2*  doc-2*
 doc-2/  doc-2/
   
   @node Mixed precision, Double precision, Bitwise operations, Arithmetic
 @subsection Mixed precision  @subsection Mixed precision
 doc-m+  doc-m+
 doc-*/  doc-*/
Line 396  doc-um/mod Line 407  doc-um/mod
 doc-fm/mod  doc-fm/mod
 doc-sm/rem  doc-sm/rem
   
   @node Double precision, Floating Point, Mixed precision, Arithmetic
 @subsection Double precision  @subsection Double precision
 doc-d+  doc-d+
 doc-d-  doc-d-
Line 404  doc-dabs Line 416  doc-dabs
 doc-dmin  doc-dmin
 doc-dmax  doc-dmax
   
 @node Stack Manipulation,,,  @node Floating Point,  , Double precision, Arithmetic
   @subsection Floating Point
   
   Angles in floating point operations are given in radians (a full circle
   has 2 pi radians). Note, that gforth has a separate floating point
   stack, but we use the unified notation.
   
   Floating point numbers have a number of unpleasant surprises for the
   unwary (e.g., floating point addition is not associative) and even a few
   for the wary. You should not use them unless you know what you are doing
   or you don't care that the results you get are totally bogus. If you
   want to learn about the problems of floating point numbers (and how to
   avoid them), you might start with @cite{David Goldberg, What Every
   Computer Scientist Should Know About Floating-Point Arithmetic, ACM
   Computing Surveys 23(1):5@minus{}48, March 1991}.
   
   doc-f+
   doc-f-
   doc-f*
   doc-f/
   doc-fnegate
   doc-fabs
   doc-fmax
   doc-fmin
   doc-floor
   doc-fround
   doc-f**
   doc-fsqrt
   doc-fexp
   doc-fexpm1
   doc-fln
   doc-flnp1
   doc-flog
   doc-falog
   doc-fsin
   doc-fcos
   doc-fsincos
   doc-ftan
   doc-fasin
   doc-facos
   doc-fatan
   doc-fatan2
   doc-fsinh
   doc-fcosh
   doc-ftanh
   doc-fasinh
   doc-facosh
   doc-fatanh
   
   @node Stack Manipulation, Memory access, Arithmetic, Words
 @section Stack Manipulation  @section Stack Manipulation
   
 gforth has a data stack (aka parameter stack) for characters, cells,  gforth has a data stack (aka parameter stack) for characters, cells,
Line 417  theoretically keep floating point number Line 478  theoretically keep floating point number
 additional difficulty, you don't know how many cells a floating point  additional difficulty, you don't know how many cells a floating point
 number takes. It is reportedly possible to write words in a way that  number takes. It is reportedly possible to write words in a way that
 they work also for a unified stack model, but we do not recommend trying  they work also for a unified stack model, but we do not recommend trying
 it. Also, a Forth system is allowed to keep the local variables on the  it. Instead, just say that your program has an environmental dependency
   on a separate FP stack.
   
   Also, a Forth system is allowed to keep the local variables on the
 return stack. This is reasonable, as local variables usually eliminate  return stack. This is reasonable, as local variables usually eliminate
 the need to use the return stack explicitly. So, if you want to produce  the need to use the return stack explicitly. So, if you want to produce
 a standard complying program and if you are using local variables in a  a standard complying program and if you are using local variables in a
 word, forget about return stack manipulations in that word (see the  word, forget about return stack manipulations in that word (see the
 standard document for the exact rules).  standard document for the exact rules).
   
   @menu
   * Data stack::                  
   * Floating point stack::        
   * Return stack::                
   * Locals stack::                
   * Stack pointer manipulation::  
   @end menu
   
   @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
 @subsection Data stack  @subsection Data stack
 doc-drop  doc-drop
 doc-nip  doc-nip
Line 444  doc-2tuck Line 517  doc-2tuck
 doc-2swap  doc-2swap
 doc-2rot  doc-2rot
   
   @node Floating point stack, Return stack, Data stack, Stack Manipulation
 @subsection Floating point stack  @subsection Floating point stack
 doc-fdrop  doc-fdrop
 doc-fnip  doc-fnip
Line 453  doc-ftuck Line 527  doc-ftuck
 doc-fswap  doc-fswap
 doc-frot  doc-frot
   
   @node Return stack, Locals stack, Floating point stack, Stack Manipulation
 @subsection Return stack  @subsection Return stack
 doc->r  doc->r
 doc-r>  doc-r>
Line 463  doc-2r> Line 538  doc-2r>
 doc-2r@  doc-2r@
 doc-2rdrop  doc-2rdrop
   
   @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
 @subsection Locals stack  @subsection Locals stack
   
   @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
 @subsection Stack pointer manipulation  @subsection Stack pointer manipulation
 doc-sp@  doc-sp@
 doc-sp!  doc-sp!
Line 475  doc-rp! Line 552  doc-rp!
 doc-lp@  doc-lp@
 doc-lp!  doc-lp!
   
 @node Memory access  @node Memory access, Control Structures, Stack Manipulation, Words
 @section Memory access  @section Memory access
   
   @menu
   * Stack-Memory transfers::      
   * Address arithmetic::          
   * Memory block access::         
   @end menu
   
   @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
 @subsection Stack-Memory transfers  @subsection Stack-Memory transfers
   
 doc-@  doc-@
Line 494  doc-sf! Line 578  doc-sf!
 doc-df@  doc-df@
 doc-df!  doc-df!
   
   @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
 @subsection Address arithmetic  @subsection Address arithmetic
   
 ANS Forth does not specify the sizes of the data types. Instead, it  ANS Forth does not specify the sizes of the data types. Instead, it
Line 522  The standard guarantees that addresses r Line 607  The standard guarantees that addresses r
 are cell-aligned; in addition, gforth guarantees that these addresses  are cell-aligned; in addition, gforth guarantees that these addresses
 are aligned for all purposes.  are aligned for all purposes.
   
   Note that the standard defines a word @code{char}, which has nothing to
   do with address arithmetic.
   
 doc-chars  doc-chars
 doc-char+  doc-char+
 doc-cells  doc-cells
Line 540  doc-dfloats Line 628  doc-dfloats
 doc-dfloat+  doc-dfloat+
 doc-dfalign  doc-dfalign
 doc-dfaligned  doc-dfaligned
   doc-maxalign
   doc-maxaligned
   doc-cfalign
   doc-cfaligned
 doc-address-unit-bits  doc-address-unit-bits
   
   @node Memory block access,  , Address arithmetic, Memory access
 @subsection Memory block access  @subsection Memory block access
   
 doc-move  doc-move
Line 555  doc-cmove> Line 648  doc-cmove>
 doc-fill  doc-fill
 doc-blank  doc-blank
   
 @node Control Structures  @node Control Structures, Locals, Memory access, Words
 @section Control Structures  @section Control Structures
   
 Control structures in Forth cannot be used in interpret state, only in  Control structures in Forth cannot be used in interpret state, only in
Line 563  compile state, i.e., in a colon definiti Line 656  compile state, i.e., in a colon definiti
 limitation, but have not seen a satisfying way around it yet, although  limitation, but have not seen a satisfying way around it yet, although
 many schemes have been proposed.  many schemes have been proposed.
   
   @menu
   * Selection::                   
   * Simple Loops::                
   * Counted Loops::               
   * Arbitrary control structures::  
   * Calls and returns::           
   * Exception Handling::          
   @end menu
   
   @node Selection, Simple Loops, Control Structures, Control Structures
 @subsection Selection  @subsection Selection
   
 @example  @example
Line 581  ELSE Line 684  ELSE
 ENDIF  ENDIF
 @end example  @end example
   
 You can use @code{THEN} instead of {ENDIF}. Indeed, @code{THEN} is  You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
 standard, and @code{ENDIF} is not, although it is quite popular. We  standard, and @code{ENDIF} is not, although it is quite popular. We
 recommend using @code{ENDIF}, because it is less confusing for people  recommend using @code{ENDIF}, because it is less confusing for people
 who also know other languages (and is not prone to reinforcing negative  who also know other languages (and is not prone to reinforcing negative
Line 608  can avoid using @code{?dup}. Line 711  can avoid using @code{?dup}.
 CASE  CASE
   @var{n1} OF @var{code1} ENDOF    @var{n1} OF @var{code1} ENDOF
   @var{n2} OF @var{code2} ENDOF    @var{n2} OF @var{code2} ENDOF
   @dots    @dots{}
 ENDCASE  ENDCASE
 @end example  @end example
   
Line 617  Executes the first @var{codei}, where th Line 720  Executes the first @var{codei}, where th
 the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,  the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
 but must not consume it.  but must not consume it.
   
   @node Simple Loops, Counted Loops, Selection, Control Structures
 @subsection Simple Loops  @subsection Simple Loops
   
 @example  @example
Line 648  AGAIN Line 752  AGAIN
   
 This is an endless loop.  This is an endless loop.
   
   @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
 @subsection Counted Loops  @subsection Counted Loops
   
 The basic counted loop is:  The basic counted loop is:
Line 703  Therefore we recommend avoiding using @c Line 808  Therefore we recommend avoiding using @c
 @var{n}. One alternative is @code{@var{n} S+LOOP}, where the negative  @var{n}. One alternative is @code{@var{n} S+LOOP}, where the negative
 case behaves symmetrical to the positive case:  case behaves symmetrical to the positive case:
   
 @code{-2 0 ?DO  i .  -1 +LOOP}  prints @code{0 -1}  @code{-2 0 ?DO  i .  -1 S+LOOP}  prints @code{0 -1}
   
 @code{-1 0 ?DO  i .  -1 +LOOP}  prints @code{0}  @code{-1 0 ?DO  i .  -1 S+LOOP}  prints @code{0}
   
 @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing  @code{ 0 0 ?DO  i .  -1 S+LOOP}  prints nothing
   
 The loop is terminated when the border between @var{limit@minus{}sgn(n)} and  The loop is terminated when the border between @var{limit@minus{}sgn(n)} and
 @var{limit} is crossed. However, @code{S+LOOP} is not part of the ANS  @var{limit} is crossed. However, @code{S+LOOP} is not part of the ANS
Line 734  iterates @var{n+1} times; @code{i} produ Line 839  iterates @var{n+1} times; @code{i} produ
 and ending with 0. Other Forth systems may behave differently, even if  and ending with 0. Other Forth systems may behave differently, even if
 they support @code{FOR} loops.  they support @code{FOR} loops.
   
   @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
 @subsection Arbitrary control structures  @subsection Arbitrary control structures
   
 ANS Forth permits and supports using control structures in a non-nested  ANS Forth permits and supports using control structures in a non-nested
Line 779  doc-next Line 885  doc-next
 doc-leave  doc-leave
 doc-?leave  doc-?leave
 doc-unloop  doc-unloop
 doc-undo  doc-done
   
 The standard does not allow using @code{cs-pick} and @code{cs-roll} on  The standard does not allow using @code{cs-pick} and @code{cs-roll} on
 @i{do-sys}. Our system allows it, but it's your job to ensure that for  @i{do-sys}. Our system allows it, but it's your job to ensure that for
 every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path  every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
 through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the  through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
 fall-through path). Also, you have to ensure that all @code{LEAVE}s are  fall-through path). Also, you have to ensure that all @code{LEAVE}s are
 resolved (by using one of the loop-ending words or @code{UNDO}).  resolved (by using one of the loop-ending words or @code{DONE}).
   
 Another group of control structure words are  Another group of control structure words are
   
Line 841  That's much easier to read, isn't it? Of Line 947  That's much easier to read, isn't it? Of
 @code{WHILE} are predefined, so in this example it would not be  @code{WHILE} are predefined, so in this example it would not be
 necessary to define them.  necessary to define them.
   
   @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
 @subsection Calls and returns  @subsection Calls and returns
   
 A definition can be called simply be writing the name of the  A definition can be called simply be writing the name of the
Line 854  primitive compiled by @code{EXIT} is Line 961  primitive compiled by @code{EXIT} is
   
 doc-;s  doc-;s
   
   @node Exception Handling,  , Calls and returns, Control Structures
 @subsection Exception Handling  @subsection Exception Handling
   
 doc-catch  doc-catch
 doc-throw  doc-throw
   
 @node Locals  @node Locals, Defining Words, Control Structures, Words
 @section Locals  @section Locals
   
 Local variables can make Forth programming more enjoyable and Forth  Local variables can make Forth programming more enjoyable and Forth
Line 869  locals wordset, but also our own, more p Line 977  locals wordset, but also our own, more p
 implemented the ANS Forth locals wordset through our locals wordset).  implemented the ANS Forth locals wordset through our locals wordset).
   
 @menu  @menu
   * gforth locals::               
   * ANS Forth locals::            
 @end menu  @end menu
   
   @node gforth locals, ANS Forth locals, Locals, Locals
 @subsection gforth locals  @subsection gforth locals
   
 Locals can be defined with  Locals can be defined with
Line 934  locals are initialized with values from Line 1045  locals are initialized with values from
 Currently there is no way to define locals with user-defined data  Currently there is no way to define locals with user-defined data
 structures, but we are working on it.  structures, but we are working on it.
   
 GNU Forth allows defining locals everywhere in a colon definition. This poses the following questions:  GNU Forth allows defining locals everywhere in a colon definition. This
   poses the following questions:
   
   @menu
   * Where are locals visible by name?::  
   * How long do locals live?::    
   * Programming Style::           
   * Implementation::              
   @end menu
   
   @node Where are locals visible by name?, How long do locals live?, gforth locals, gforth locals
 @subsubsection Where are locals visible by name?  @subsubsection Where are locals visible by name?
   
 Basically, the answer is that locals are visible where you would expect  Basically, the answer is that locals are visible where you would expect
Line 994  AHEAD Line 1114  AHEAD
 BEGIN  BEGIN
   x    x
 [ 1 CS-ROLL ] THEN  [ 1 CS-ROLL ] THEN
   { x }    @{ x @}
   ...    ...
 UNTIL  UNTIL
 @end example  @end example
Line 1026  compiler. When the branch to the @code{B Line 1146  compiler. When the branch to the @code{B
 warns the user if it was too optimisitic:  warns the user if it was too optimisitic:
 @example  @example
 IF  IF
   { x }    @{ x @}
 BEGIN  BEGIN
   \ x ?     \ x ? 
 [ 1 cs-roll ] THEN  [ 1 cs-roll ] THEN
Line 1042  is not used in the wrong area by using e Line 1162  is not used in the wrong area by using e
 @example  @example
 IF  IF
   SCOPE    SCOPE
   { x }    @{ x @}
   ENDSCOPE    ENDSCOPE
 BEGIN  BEGIN
 [ 1 cs-roll ] THEN  [ 1 cs-roll ] THEN
Line 1065  doc-assume-live Line 1185  doc-assume-live
   
 E.g.,  E.g.,
 @example  @example
 { x }  @{ x @}
 AHEAD  AHEAD
 ASSUME-LIVE  ASSUME-LIVE
 BEGIN  BEGIN
Line 1086  rearranging the loop. E.g., the ``most i Line 1206  rearranging the loop. E.g., the ``most i
 arranged into:  arranged into:
 @example  @example
 BEGIN  BEGIN
   { x }    @{ x @}
   ... 0=    ... 0=
 WHILE  WHILE
   x    x
 REPEAT  REPEAT
 @end example  @end example
   
   @node How long do locals live?, Programming Style, Where are locals visible by name?, gforth locals
 @subsubsection How long do locals live?  @subsubsection How long do locals live?
   
 The right answer for the lifetime question would be: A local lives at  The right answer for the lifetime question would be: A local lives at
Line 1106  languages (e.g., C): The local lives onl Line 1227  languages (e.g., C): The local lives onl
 afterwards its address is invalid (and programs that access it  afterwards its address is invalid (and programs that access it
 afterwards are erroneous).  afterwards are erroneous).
   
   @node Programming Style, Implementation, How long do locals live?, gforth locals
 @subsubsection Programming Style  @subsubsection Programming Style
   
 The freedom to define locals anywhere has the potential to change  The freedom to define locals anywhere has the potential to change
Line 1117  wrong order, just write a locals definit Line 1239  wrong order, just write a locals definit
 write the items in the order you want.  write the items in the order you want.
   
 This seems a little far-fetched and eliminating stack manipulations is  This seems a little far-fetched and eliminating stack manipulations is
 unlikely to become a conscious programming objective. Still, the  unlikely to become a conscious programming objective. Still, the number
 number of stack manipulations will be reduced dramatically if local  of stack manipulations will be reduced dramatically if local variables
 variables are used liberally (e.g., compare @code{max} in \sect{misc}  are used liberally (e.g., compare @code{max} in @ref{gforth locals} with
 with a traditional implementation of @code{max}).  a traditional implementation of @code{max}).
   
 This shows one potential benefit of locals: making Forth programs more  This shows one potential benefit of locals: making Forth programs more
 readable. Of course, this benefit will only be realized if the  readable. Of course, this benefit will only be realized if the
Line 1174  are initialized with the right value for Line 1296  are initialized with the right value for
 Here it is clear from the start that @code{s1} has a different value  Here it is clear from the start that @code{s1} has a different value
 in every loop iteration.  in every loop iteration.
   
   @node Implementation,  , Programming Style, gforth locals
 @subsubsection Implementation  @subsubsection Implementation
   
 GNU Forth uses an extra locals stack. The most compelling reason for  GNU Forth uses an extra locals stack. The most compelling reason for
Line 1197 local0 Line 1320 local0
 compile the right specialized version, or the general version, as  compile the right specialized version, or the general version, as
 appropriate:  appropriate:
   
 doc-compile-@@local  doc-compile-@local
 doc-compile-f@@local  doc-compile-f@local
 doc-compile-lp+!  doc-compile-lp+!
   
 Combinations of conditional branches and @code{lp+!#} like  Combinations of conditional branches and @code{lp+!#} like
Line 1214  special area is cleared at the start of Line 1337  special area is cleared at the start of
 A special feature of GNU Forths dictionary is used to implement the  A special feature of GNU Forths dictionary is used to implement the
 definition of locals without type specifiers: every wordlist (aka  definition of locals without type specifiers: every wordlist (aka
 vocabulary) has its own methods for searching  vocabulary) has its own methods for searching
 etc. (@xref{dictionary}). For the present purpose we defined a wordlist  etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
 with a special search method: When it is searched for a word, it  with a special search method: When it is searched for a word, it
 actually creates that word using @code{W:}. @code{@{} changes the search  actually creates that word using @code{W:}. @code{@{} changes the search
 order to first search the wordlist containing @code{@}}, @code{W:} etc.,  order to first search the wordlist containing @code{@}}, @code{W:} etc.,
Line 1251  The locals stack pointer is only adjuste Line 1374  The locals stack pointer is only adjuste
 @code{lp+!#} orig-locals-size @minus{} new-locals-size  @code{lp+!#} orig-locals-size @minus{} new-locals-size
 @end format  @end format
 The second @code{lp+!#} adjusts the locals stack pointer from the  The second @code{lp+!#} adjusts the locals stack pointer from the
 level at the {\em orig} point to the level after the @code{THEN}. The  level at the @var{orig} point to the level after the @code{THEN}. The
 first @code{lp+!#} adjusts the locals stack pointer from the current  first @code{lp+!#} adjusts the locals stack pointer from the current
 level to the level at the orig point, so the complete effect is an  level to the level at the orig point, so the complete effect is an
 adjustment from the current level to the right level after the  adjustment from the current level to the right level after the
Line 1301  this may lead to increased space needs f Line 1424  this may lead to increased space needs f
 usually less than reclaiming this space would cost in code size.  usually less than reclaiming this space would cost in code size.
   
   
   @node ANS Forth locals,  , gforth locals, Locals
 @subsection ANS Forth locals  @subsection ANS Forth locals
   
 The ANS Forth locals wordset does not define a syntax for locals, but  The ANS Forth locals wordset does not define a syntax for locals, but
Line 1335  The whole definition must be in one line Line 1459  The whole definition must be in one line
 @end itemize  @end itemize
   
 Locals defined in this way behave like @code{VALUE}s  Locals defined in this way behave like @code{VALUE}s
 (@xref{values}). I.e., they are initialized from the stack. Using their  (@xref{Values}). I.e., they are initialized from the stack. Using their
 name produces their value. Their value can be changed using @code{TO}.  name produces their value. Their value can be changed using @code{TO}.
   
 Since this syntax is supported by gforth directly, you need not do  Since this syntax is supported by gforth directly, you need not do
Line 1360  programs harder to read, and easier to m Line 1484  programs harder to read, and easier to m
 merit of this syntax is that it is easy to implement using the ANS Forth  merit of this syntax is that it is easy to implement using the ANS Forth
 locals wordset.  locals wordset.
   
 @node Internals  @node Defining Words, Wordlists, Locals, Words
   @section Defining Words
   
   @menu
   * Values::                      
   @end menu
   
   @node Values,  , Defining Words, Defining Words
   @subsection Values
   
   @node Wordlists, Files, Defining Words, Words
   @section Wordlists
   
   @node Files, Blocks, Wordlists, Words
   @section Files
   
   @node Blocks, Other I/O, Files, Words
   @section Blocks
   
   @node Other I/O, Programming Tools, Blocks, Words
   @section Other I/O
   
   @node Programming Tools, Threading Words, Other I/O, Words
   @section Programming Tools
   
   @menu
   * Debugging::                   Simple and quick.
   * Assertions::                  Making your programs self-checking.
   @end menu
   
   @node Debugging, Assertions, Programming Tools, Programming Tools
   @subsection Debugging
   
   The simple debugging aids provided in @file{debugging.fs}
   are meant to support a different style of debugging than the
   tracing/stepping debuggers used in languages with long turn-around
   times.
   
   A much better (faster) way in fast-compilig languages is to add
   printing code at well-selected places, let the program run, look at
   the output, see where things went wrong, add more printing code, etc.,
   until the bug is found.
   
   The word @code{~~} is easy to insert. It just prints debugging
   information (by default the source location and the stack contents). It
   is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
   query-replace them with nothing). The deferred words
   @code{printdebugdata} and @code{printdebugline} control the output of
   @code{~~}. The default source location output format works well with
   Emacs' compilation mode, so you can step through the program at the
   source level using @kbd{C-x `} (the advantage over a stepping debugger
   is that you can step in any direction and you know where the crash has
   happened or where the strange data has occurred).
   
   Note that the default actions clobber the contents of the pictured
   numeric output string, so you should not use @code{~~}, e.g., between
   @code{<#} and @code{#>}.
   
   doc-~~
   doc-printdebugdata
   doc-printdebugline
   
   @node Assertions,  , Debugging, Programming Tools
   @subsection Assertions
   
   It is a good idea to make your programs self-checking, in particular, if
   you use an assumption (e.g., that a certain field of a data structure is
   never zero) that may become wrong during maintenance. GForth supports
   assertions for this purpose. They are used like this:
   
   @example
   assert( @var{flag} )
   @end example
   
   The code between @code{assert(} and @code{)} should compute a flag, that
   should be true if everything is alright and false otherwise. It should
   not change anything else on the stack. The overall stack effect of the
   assertion is @code{( -- )}. E.g.
   
   @example
   assert( 1 1 + 2 = ) \ what we learn in school
   assert( dup 0<> ) \ assert that the top of stack is not zero
   assert( false ) \ this code should not be reached
   @end example
   
   The need for assertions is different at different times. During
   debugging, we want more checking, in production we sometimes care more
   for speed. Therefore, assertions can be turned off, i.e., the assertion
   becomes a comment. Depending on the importance of an assertion and the
   time it takes to check it, you may want to turn off some assertions and
   keep others turned on. GForth provides several levels of assertions for
   this purpose:
   
   doc-assert0(
   doc-assert1(
   doc-assert2(
   doc-assert3(
   doc-assert(
   doc-)
   
   @code{Assert(} is the same as @code{assert1(}. The variable
   @code{assert-level} specifies the highest assertions that are turned
   on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
   @code{assert1(} assertions perform checking, while @code{assert2(} and
   @code{assert3(} assertions are treated as comments.
   
   Note that the @code{assert-level} is evaluated at compile-time, not at
   run-time. I.e., you cannot turn assertions on or off at run-time, you
   have to set the @code{assert-level} appropriately before compiling a
   piece of code. You can compile several pieces of code at several
   @code{assert-level}s (e.g., a trusted library at level 1 and newly
   written code at level 3).
   
   doc-assert-level
   
   If an assertion fails, a message compatible with Emacs' compilation mode
   is produced and the execution is aborted (currently with @code{ABORT"}.
   If there is interest, we will introduce a special throw code. But if you
   intend to @code{catch} a specific condition, using @code{throw} is
   probably more appropriate than an assertion).
   
   @node Threading Words,  , Programming Tools, Words
   @section Threading Words
   
   These words provide access to code addresses and other threading stuff
   in gforth (and, possibly, other interpretive Forths). It more or less
   abstracts away the differences between direct and indirect threading
   (and, for direct threading, the machine dependences). However, at
   present this wordset is still inclomplete. It is also pretty low-level;
   some day it will hopefully be made unnecessary by an internals words set
   that abstracts implementation details away completely.
   
   doc->code-address
   doc->does-code
   doc-code-address!
   doc-does-code!
   doc-does-handler!
   doc-/does-handler
   
   
   
   @node ANS conformance, Model, Words, Top
   @chapter ANS conformance
   
   To the best of our knowledge, gforth is an
   
   ANS Forth System
   @itemize
   @item providing the Core Extensions word set
   @item providing the Block word set
   @item providing the Block Extensions word set
   @item providing the Double-Number word set
   @item providing the Double-Number Extensions word set
   @item providing the Exception word set
   @item providing the Exception Extensions word set
   @item providing the Facility word set
   @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
   @item providing the File Access word set
   @item providing the File Access Extensions word set
   @item providing the Floating-Point word set
   @item providing the Floating-Point Extensions word set
   @item providing the Locals word set
   @item providing the Locals Extensions word set
   @item providing the Memory-Allocation word set
   @item providing the Memory-Allocation Extensions word set (that one's easy)
   @item providing the Programming-Tools word set
   @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
   @item providing the Search-Order word set
   @item providing the Search-Order Extensions word set
   @item providing the String word set
   @item providing the String Extensions word set (another easy one)
   @end itemize
   
   In addition, ANS Forth systems are required to document certain
   implementation choices. This chapter tries to meet these
   requirements. In many cases it gives a way to ask the system for the
   information instead of providing the information directly, in
   particular, if the information depends on the processor, the operating
   system or the installation options chosen, or if they are likely to
   change during the maintenance of gforth.
   
   @comment The framework for the rest has been taken from pfe.
   
   @menu
   * The Core Words::              
   * The optional Block word set::  
   * The optional Double Number word set::  
   * The optional Exception word set::  
   * The optional Facility word set::  
   * The optional File-Access word set::  
   * The optional Floating-Point word set::  
   * The optional Locals word set::  
   * The optional Memory-Allocation word set::  
   * The optional Programming-Tools word set::  
   * The optional Search-Order word set::  
   @end menu
   
   
   @c =====================================================================
   @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
   @comment  node-name,  next,  previous,  up
   @section The Core Words
   @c =====================================================================
   
   @menu
   * core-idef::                   Implementation Defined Options                   
   * core-ambcond::                Ambiguous Conditions                
   * core-other::                  Other System Documentation                  
   @end menu
   
   @c ---------------------------------------------------------------------
   @node core-idef, core-ambcond, The Core Words, The Core Words
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item (Cell) aligned addresses:
   processor-dependent. Gforths alignment words perform natural alignment
   (e.g., an address aligned for a datum of size 8 is divisible by
   8). Unaligned accesses usually result in a @code{-23 THROW}.
   
   @item @code{EMIT} and non-graphic characters:
   The character is output using the C library function (actually, macro)
   @code{putchar}.
   
   @item character editing of @code{ACCEPT} and @code{EXPECT}:
   This is modeled on the GNU readline library (@pxref{Readline
   Interaction, , Command Line Editing, readline, The GNU Readline
   Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
   producing a full word completion every time you type it (instead of
   producing the common prefix of all completions).
   
   @item character set:
   The character set of your computer and display device. Gforth is
   8-bit-clean (but some other component in your system may make trouble).
   
   @item Character-aligned address requirements:
   installation-dependent. Currently a character is represented by a C
   @code{unsigned char}; in the future we might switch to @code{wchar_t}
   (Comments on that requested).
   
   @item character-set extensions and matching of names:
   Any character except 0 can be used in a name. Matching is
   case-insensitive. The matching is performed using the C function
   @code{strncasecmp}, whose function is probably influenced by the
   locale. E.g., the @code{C} locale does not know about accents and
   umlauts, so they are matched case-sensitively in that locale. For
   portability reasons it is best to write programs such that they work in
   the @code{C} locale. Then one can use libraries written by a Polish
   programmer (who might use words containing ISO Latin-2 encoded
   characters) and by a French programmer (ISO Latin-1) in the same program
   (of course, @code{WORDS} will produce funny results for some of the
   words (which ones, depends on the font you are using)). Also, the locale
   you prefer may not be available in other operating systems. Hopefully,
   Unicode will solve these problems one day.
   
   @item conditions under which control characters match a space delimiter:
   If @code{WORD} is called with the space character as a delimiter, all
   white-space characters (as identified by the C macro @code{isspace()})
   are delimiters. @code{PARSE}, on the other hand, treats space like other
   delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
   like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
   interpreter (aka text interpreter) by default, treats all white-space
   characters as delimiters.
   
   @item format of the control flow stack:
   The data stack is used as control flow stack. The size of a control flow
   stack item in cells is given by the constant @code{cs-item-size}. At the
   time of this writing, an item consists of a (pointer to a) locals list
   (third), an address in the code (second), and a tag for identifying the
   item (TOS). The following tags are used: @code{defstart},
   @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
   @code{scopestart}.
   
   @item conversion of digits > 35
   The characters @code{[\]^_'} are the digits with the decimal value
   36@minus{}41. There is no way to input many of the larger digits.
   
   @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
   The cursor is moved to the end of the entered string. If the input is
   terminated using the @kbd{Return} key, a space is typed.
   
   @item exception abort sequence of @code{ABORT"}:
   The error string is stored into the variable @code{"error} and a
   @code{-2 throw} is performed.
   
   @item input line terminator:
   For interactive input, @kbd{C-m} and @kbd{C-j} terminate lines. One of
   these characters is typically produced when you type the @kbd{Enter} or
   @kbd{Return} key.
   
   @item maximum size of a counted string:
   @code{s" /counted-string" environment? drop .}. Currently 255 characters
   on all ports, but this may change.
   
   @item maximum size of a parsed string:
   Given by the constant @code{/line}. Currently 255 characters.
   
   @item maximum size of a definition name, in characters:
   31
   
   @item maximum string length for @code{ENVIRONMENT?}, in characters:
   31
   
   @item method of selecting the user input device:
   The user input device is the standard input. There is current no way to
   change it from within gforth. However, the input can typically be
   redirected in the command line that starts gforth.
   
   @item method of selecting the user output device:
   The user output device is the standard output. It cannot be redirected
   from within gforth, but typically from the command line that starts
   gforth. Gforth uses buffered output, so output on a terminal does not
   become visible before the next newline or buffer overflow. Output on
   non-terminals is invisible until the buffer overflows.
   
   @item methods of dictionary compilation:
   Waht are we expected to document here?
   
   @item number of bits in one address unit:
   @code{s" address-units-bits" environment? drop .}. 8 in all current
   ports.
   
   @item number representation and arithmetic:
   Processor-dependent. Binary two's complement on all current ports.
   
   @item ranges for integer types:
   Installation-dependent. Make environmental queries for @code{MAX-N},
   @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
   unsigned (and positive) types is 0. The lower bound for signed types on
   two's complement and one's complement machines machines can be computed
   by adding 1 to the upper bound.
   
   @item read-only data space regions:
   The whole Forth data space is writable.
   
   @item size of buffer at @code{WORD}:
   @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
   shared with the pictured numeric output string. If overwriting
   @code{PAD} is acceptable, it is as large as the remaining dictionary
   space, although only as much can be sensibly used as fits in a counted
   string.
   
   @item size of one cell in address units:
   @code{1 cells .}.
   
   @item size of one character in address units:
   @code{1 chars .}. 1 on all current ports.
   
   @item size of the keyboard terminal buffer:
   Varies. You can determine the size at a specific time using @code{lp@
   tib - .}. It is shared with the locals stack and TIBs of files that
   include the current file. You can change the amount of space for TIBs
   and locals stack at gforth startup with the command line option
   @code{-l}.
   
   @item size of the pictured numeric output buffer:
   @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
   shared with @code{WORD}.
   
   @item size of the scratch area returned by @code{PAD}:
   The remainder of dictionary space. You can even use the unused part of
   the data stack space. The current size can be computed with @code{sp@
   pad - .}.
   
   @item system case-sensitivity characteristics:
   Dictionary searches are case insensitive. However, as explained above
   under @i{character-set extensions}, the matching for non-ASCII
   characters is determined by the locale you are using. In the default
   @code{C} locale all non-ASCII characters are matched case-sensitively.
   
   @item system prompt:
   @code{ ok} in interpret state, @code{ compiled} in compile state.
   
   @item division rounding:
   installation dependent. @code{s" floored" environment? drop .}. We leave
   the choice to gcc (what to use for @code{/}) and to you (whether to use
   @code{fm/mod}, @code{sm/rem} or simply @code{/}).
   
   @item values of @code{STATE} when true:
   -1.
   
   @item values returned after arithmetic overflow:
   On two's complement machines, arithmetic is performed modulo
   2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
   arithmetic (with appropriate mapping for signed types). Division by zero
   typically results in a @code{-55 throw} (floatingpoint unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate.
   
   @item whether the current definition can be found after @t{DOES>}:
   No.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node core-ambcond, core-other, core-idef, The Core Words
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item a name is neither a word nor a number:
   @code{-13 throw} (Undefined word)
   
   @item a definition name exceeds the maximum length allowed:
   @code{-19 throw} (Word name too long)
   
   @item addressing a region not inside the various data spaces of the forth system:
   The stacks, code space and name space are accessible. Machine code space is
   typically readable. Accessing other addresses gives results dependent on
   the operating system. On decent systems: @code{-9 throw} (Invalid memory
   address).
   
   @item argument type incompatible with parameter:
   This is usually not caught. Some words perform checks, e.g., the control
   flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
   mismatch).
   
   @item attempting to obtain the execution token of a word with undefined execution semantics:
   You get an execution token representing the compilation semantics
   instead.
   
   @item dividing by zero:
   typically results in a @code{-55 throw} (floating point unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate.
   
   @item insufficient data stack or return stack space:
   Not checked. This typically results in mysterious illegal memory
   accesses, producing @code{-9 throw} (Invalid memory address) or
   @code{-23 throw} (Address alignment exception).
   
   @item insufficient space for loop control parameters:
   like other return stack overflows.
   
   @item insufficient space in the dictionary:
   Not checked. Similar results as stack overflows. However, typically the
   error appears at a different place when one inserts or removes code.
   
   @item interpreting a word with undefined interpretation semantics:
   For some words, we defined interpretation semantics. For the others:
   @code{-14 throw} (Interpreting a compile-only word). Note that this is
   checked only by the outer (aka text) interpreter; if the word is
   @code{execute}d in some other way, it will typically perform it's
   compilation semantics even in interpret state. (We could change @code{'}
   and relatives not to give the xt of such words, but we think that would
   be too restrictive).
   
   @item modifying the contents of the input buffer or a string literal:
   These are located in writable memory and can be modified.
   
   @item overflow of the pictured numeric output string:
   Not checked.
   
   @item parsed string overflow:
   @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
   
   @item producing a result out of range:
   On two's complement machines, arithmetic is performed modulo
   2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
   arithmetic (with appropriate mapping for signed types). Division by zero
   typically results in a @code{-55 throw} (floatingpoint unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate. @code{convert} and @code{>number} currently overflow
   silently.
   
   @item reading from an empty data or return stack:
   The data stack is checked by the outer (aka text) interpreter after
   every word executed. If it has underflowed, a @code{-4 throw} (Stack
   underflow) is performed. Apart from that, the stacks are not checked and
   underflows can result in similar behaviour as overflows (of adjacent
   stacks).
   
   @item unexepected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
   @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
   use zero-length string as a name). Words like @code{'} probably will not
   find what they search. Note that it is possible to create zero-length
   names with @code{nextname} (should it not?).
   
   @item @code{>IN} greater than input buffer:
   The next invocation of a parsing word returns a string wih length 0.
   
   @item @code{RECURSE} appears after @code{DOES>}:
   Compiles a recursive call to the defining word not to the defined word.
   
   @item argument input source different than current input source for @code{RESTORE-INPUT}:
   !!???If the argument input source is a valid input source then it gets
   restored. Otherwise causes @code{-12 THROW} which unless caught issues
   the message "argument type mismatch" and aborts.
   
   @item data space containing definitions gets de-allocated:
   Deallocation with @code{allot} is not checked. This typically resuls in
   memory access faults or execution of illegal instructions.
   
   @item data space read/write with incorrect alignment:
   Processor-dependent. Typically results in a @code{-23 throw} (Address
   alignment exception). Under Linux on a 486 or later processor with
   alignment turned on, incorrect alignment results in a @code{-9 throw}
   (Invalid memory address). There are reportedly some processors with
   alignment restrictions that do not report them.
   
   @item data space pointer not properly aligned, @code{,}, @code{C,}:
   Like other alignment errors.
   
   @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
   Not checked. May cause an illegal memory access.
   
   @item loop control parameters not available:
   Not checked. The counted loop words simply assume that the top of return
   stack items are loop control parameters and behave accordingly.
   
   @item most recent definition does not have a name (@code{IMMEDIATE}):
   @code{abort" last word was headerless"}.
   
   @item name not defined by @code{VALUE} used by @code{TO}:
   @code{-32 throw} (Invalid name argument)
   
   @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
   @code{-13 throw} (Undefined word)
   
   @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
   Gforth behaves as if they were of the same type. I.e., you can predict
   the behaviour by interpreting all parameters as, e.g., signed.
   
   @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
   Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} is equivalent to
   @code{TO}.
   
   @item String longer than a counted string returned by @code{WORD}:
   Not checked. The string will be ok, but the count will, of course,
   contain only the least significant bits of the length.
   
   @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
   Processor-dependent. Typical behaviours are returning 0 and using only
   the low bits of the shift count.
   
   @item word not defined via @code{CREATE}:
   @code{>BODY} produces the PFA of the word no matter how it was defined.
   
   @code{DOES>} changes the execution semantics of the last defined word no
   matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
   @code{CREATE , DOES>}.
   
   @item words improperly used outside @code{<#} and @code{#>}:
   Not checked. As usual, you can expect memory faults.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node core-other,  , core-ambcond, The Core Words
   @subsection Other system documentation
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item nonstandard words using @code{PAD}:
   None.
   
   @item operator's terminal facilities available:
   !!??
   
   @item program data space available:
   @code{sp@ here - .} gives the space remaining for dictionary and data
   stack together.
   
   @item return stack space available:
   !!??
   
   @item stack space available:
   @code{sp@ here - .} gives the space remaining for dictionary and data
   stack together.
   
   @item system dictionary space required, in address units:
   Type @code{here forthstart - .} after startup. At the time of this
   writing, this gives 70108 (bytes) on a 32-bit system.
   @end table
   
   
   @c =====================================================================
   @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
   @section The optional Block word set
   @c =====================================================================
   
   @menu
   * block-idef::                  Implementation Defined Options                  
   * block-ambcond::               Ambiguous Conditions               
   * block-other::                 Other System Documentation                 
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item the format for display by @code{LIST}:
   First the screen number is displayed, then 16 lines of 64 characters,
   each line preceded by the line number.
   
   @item the length of a line affected by @code{\}:
   64 characters.
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node block-ambcond, block-other, block-idef, The optional Block word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item correct block read was not possible:
   Typically results in a @code{throw} of some OS-derived value (between
   -512 and -2048). If the blocks file was just not long enough, blanks are
   supplied for the missing portion.
   
   @item I/O exception in block transfer:
   Typically results in a @code{throw} of some OS-derived value (between
   -512 and -2048).
   
   @item invalid block number:
   @code{-35 throw} (Invalid block number)
   
   @item a program directly alters the contents of @code{BLK}:
   The input stream is switched to that other block, at the same
   position. If the storing to @code{BLK} happens when interpreting
   non-block input, the system will get quite confused when the block ends.
   
   @item no current block buffer for @code{UPDATE}:
   @code{UPDATE} has no effect.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node block-other,  , block-ambcond, The optional Block word set
   @subsection Other system documentation
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item any restrictions a multiprogramming system places on the use of buffer addresses:
   No restrictions (yet).
   
   @item the number of blocks available for source and data:
   depends on your disk space.
   
   @end table
   
   
   @c =====================================================================
   @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
   @section The optional Double Number word set
   @c =====================================================================
   
   @menu
   * double-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @var{d} outside of range of @var{n} in @code{D>S}:
   The least significant cell of @var{d} is produced.
   
   @end table
   
   
   @c =====================================================================
   @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
   @section The optional Exception word set
   @c =====================================================================
   
   @menu
   * exception-idef::              Implementation Defined Options              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node exception-idef,  , The optional Exception word set, The optional Exception word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   @item @code{THROW}-codes used in the system:
   The codes -256@minus{}-511 are used for reporting signals (see
   @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
   (for file and memory allocation operations). The mapping from OS error
   numbers to throw code is -512@minus{}@var{errno}. One side effect of
   this mapping is that undefined OS errors produce a message with a
   strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
   488} on my system.
   @end table
   
   @c =====================================================================
   @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
   @section The optional Facility word set
   @c =====================================================================
   
   @menu
   * facility-idef::               Implementation Defined Options               
   * facility-ambcond::            Ambiguous Conditions            
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item encoding of keyboard events (@code{EKEY}):
   Not yet implemeted.
   
   @item duration of a system clock tick
   System dependent. With respect to @code{MS}, the time is specified in
   microseconds. How well the OS and the hardware implement this, is
   another question.
   
   @item repeatability to be expected from the execution of @code{MS}:
   System dependent. On Unix, a lot depends on load. If the system is
   lightly loaded, and the delay is short enough that gforth does not get
   swapped out, the performance should be acceptable. Under MS-DOS and
   other single-tasking systems, it should be good.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node facility-ambcond,  , facility-idef, The optional Facility word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @code{AT-XY} can't be performed on user output device:
   Largely terminal dependant. No range checks are done on the arguments.
   No errors are reported. You may see some garbage appearing, you may see
   simply nothing happen.
   
   @end table
   
   
   @c =====================================================================
   @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
   @section The optional File-Access word set
   @c =====================================================================
   
   @menu
   * file-idef::                   Implementation Defined Options                   
   * file-ambcond::                Ambiguous Conditions                
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item File access methods used:
   @code{R/O}, @code{R/W} and @code{BIN} work as you would
   expect. @code{W/O} translates into the C file opening mode @code{w} (or
   @code{wb}): The file is cleared, if it exists, and created, if it does
   not (both with @code{open-file} and @code{create-file}).  Under Unix
   @code{create-file} creates a file with 666 permissions modified by your
   umask.
   
   @item file exceptions:
   The file words do not raise exceptions (except, perhaps, memory access
   faults when you pass illegal addresses or file-ids).
   
   @item file line terminator:
   System-dependent. Gforth uses C's newline character as line
   terminator. What the actual character code(s) of this are is
   system-dependent.
   
   @item file name format
   System dependent. Gforth just uses the file name format of your OS.
   
   @item information returned by @code{FILE-STATUS}:
   @code{FILE-STATUS} returns the most powerful file access mode allowed
   for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
   cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
   along with the retured mode.
   
   @item input file state after an exception when including source:
   All files that are left via the exception are closed.
   
   @item @var{ior} values and meaning:
   The @var{ior}s returned by the file and memory allocation words are
   intended as throw codes. They typically are in the range
   -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
   @var{ior}s is -512@minus{}@var{errno}.
   
   @item maximum depth of file input nesting:
   limited by the amount of return stack, locals/TIB stack, and the number
   of open files available. This should not give you troubles.
   
   @item maximum size of input line:
   @code{/line}. Currently 255.
   
   @item methods of mapping block ranges to files:
   Currently, the block words automatically access the file
   @file{blocks.fb} in the currend working directory. More sophisticated
   methods could be implemented if there is demand (and a volunteer).
   
   @item number of string buffers provided by @code{S"}:
   1
   
   @item size of string buffer used by @code{S"}:
   @code{/line}. currently 255.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node file-ambcond,  , file-idef, The optional File-Access word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item attempting to position a file outside it's boundaries:
   @code{REPOSITION-FILE} is performed as usual: Afterwards,
   @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
   
   @item attempting to read from file positions not yet written:
   End-of-file, i.e., zero characters are read and no error is reported.
   
   @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
   An appropriate exception may be thrown, but a memory fault or other
   problem is more probable.
   
   @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
   The @var{ior} produced by the operation, that discovered the problem, is
   thrown.
   
   @item named file cannot be opened (@code{included}):
   The @var{ior} produced by @code{open-file} is thrown.
   
   @item requesting an unmapped block number:
   There are no unmapped legal block numbers. On some operating systems,
   writing a block with a large number may overflow the file system and
   have an error message as consequence.
   
   @item using @code{source-id} when @code{blk} is non-zero:
   @code{source-id} performs its function. Typically it will give the id of
   the source which loaded the block. (Better ideas?)
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
   @section The optional Floating-Point word set
   @c =====================================================================
   
   @menu
   * floating-idef::               Implementation Defined Options
   * floating-ambcond::            Ambiguous Conditions            
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item format and range of floating point numbers:
   System-dependent; the @code{double} type of C.
   
   @item results of @code{REPRESENT} when @var{float} is out of range:
   System dependent; @code{REPRESENT} is implemented using the C library
   function @code{ecvt()} and inherits its behaviour in this respect.
   
   @item rounding or truncation of floating-point numbers:
   What's the question?!!
   
   @item size of floating-point stack:
   @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
   with the command-line option @code{-f}.
   
   @item width of floating-point stack:
   @code{1 floats}.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item Floating-point result out of range:
   System-dependent. Can result in a @code{-55 THROW} (Floating-point
   unidentified fault), or can produce a special value representing, e.g.,
   Infinity.
   
   @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
   The floating-point number is converted into decimal nonetheless.
   
   @item Both arguments are equal to zero (@code{FATAN2}):
   System-dependent. @code{FATAN2} is implemented using the C library
   function @code{atan2()}.
   
   @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
   System-dependent. Anyway, typically the cos of @var{r1} will not be zero
   because of small errors and the tan will be a very large (or very small)
   but finite number.
   
   @item @var{d} cannot be presented precisely as a float in @code{D>F}:
   The result is rounded to the nearest float.
   
   @item dividing by zero:
   @code{-55 throw} (Floating-point unidentified fault)
   
   @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
   System dependent. On IEEE-FP based systems the number is converted into
   an infinity.
   
   @item @var{float}<1 (@code{facosh}):
   @code{-55 throw} (Floating-point unidentified fault)
   
   @item @var{float}=<-1 (@code{flnp1}):
   @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
   negative infinity is typically produced for @var{float}=-1.
   
   @item @var{float}=<0 (@code{fln}, @code{flog}):
   @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
   negative infinity is typically produced for @var{float}=0.
   
   @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
   @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
   produces values for these inputs on my Linux box (Bug in the C library?)
   
   @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
   @code{-55 throw} (Floating-point unidentified fault).
   
   @item integer part of float cannot be represented by @var{d} in @code{f>d}:
   @code{-55 throw} (Floating-point unidentified fault).
   
   @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
   This does not happen.
   @end table
   
   
   
   @c =====================================================================
   @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
   @section The optional Locals word set
   @c =====================================================================
   
   @menu
   * locals-idef::                 Implementation Defined Options                 
   * locals-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item maximum number of locals in a definition:
   @code{s" #locals" environment? drop .}. Currently 15. This is a lower
   bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
   characters. The number of locals in a definition is bounded by the size
   of locals-buffer, which contains the names of the locals.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node locals-ambcond,  , locals-idef, The optional Locals word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item executing a named local in interpretation state:
   @code{-14 throw} (Interpreting a compile-only word).
   
   @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
   @code{-32 throw} (Invalid name argument)
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
   @section The optional Memory-Allocation word set
   @c =====================================================================
   
   @menu
   * memory-idef::                 Implementation Defined Options                 
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item values and meaning of @var{ior}:
   The @var{ior}s returned by the file and memory allocation words are
   intended as throw codes. They typically are in the range
   -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
   @var{ior}s is -512@minus{}@var{errno}.
   
   @end table
   
   @c =====================================================================
   @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
   @section The optional Programming-Tools word set
   @c =====================================================================
   
   @menu
   * programming-idef::            Implementation Defined Options            
   * programming-ambcond::         Ambiguous Conditions         
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item ending sequence for input following @code{;code} and @code{code}:
   Not implemented (yet).
   
   @item manner of processing input following @code{;code} and @code{code}:
   Not implemented (yet).
   
   @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
   Not implemented (yet). If they were implemented, they would use the
   search order wordset.
   
   @item source and format of display by @code{SEE}:
   The source for @code{see} is the intermediate code used by the inner
   interpreter.  The current @code{see} tries to output Forth source code
   as well as possible.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item deleting the compilation wordlist (@code{FORGET}):
   Not implemented (yet).
   
   @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
   This typically results in an @code{abort"} with a descriptive error
   message (may change into a @code{-22 throw} (Control structure mismatch)
   in the future). You may also get a memory access error. If you are
   unlucky, this ambiguous condition is not caught.
   
   @item @var{name} can't be found (@code{forget}):
   Not implemented (yet).
   
   @item @var{name} not defined via @code{CREATE}:
   @code{;code} is not implemented (yet). If it were, it would behave like
   @code{DOES>} in this respect, i.e., change the execution semantics of
   the last defined word no matter how it was defined.
   
   @item @code{POSTPONE} applied to @code{[IF]}:
   After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
   equivalent to @code{[IF]}.
   
   @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
   Continue in the same state of conditional compilation in the next outer
   input source. Currently there is no warning to the user about this.
   
   @item removing a needed definition (@code{FORGET}):
   Not implemented (yet).
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
   @section The optional Search-Order word set
   @c =====================================================================
   
   @menu
   * search-idef::                 Implementation Defined Options                 
   * search-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item maximum number of word lists in search order:
   @code{s" wordlists" environment? drop .}. Currently 16.
   
   @item minimum search order:
   @code{root root}.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node search-ambcond,  , search-idef, The optional Search-Order word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item changing the compilation wordlist (during compilation):
   The definition is put into the wordlist that is the compilation wordlist
   when @code{REVEAL} is executed (by @code{;}, @code{DOES>},
   @code{RECURSIVE}, etc.).
   
   @item search order empty (@code{previous}):
   @code{abort" Vocstack empty"}.
   
   @item too many word lists in search order (@code{also}):
   @code{abort" Vocstack full"}.
   
   @end table
   
   
   @node Model, Emacs and GForth, ANS conformance, Top
   @chapter Model
   
   @node Emacs and GForth, Internals, Model, Top
   @chapter Emacs and GForth
   
   GForth comes with @file{gforth.el}, an improved version of
   @file{forth.el} by Goran Rydqvist (icluded in the TILE package). The
   improvements are a better (but still not perfect) handling of
   indentation. I have also added comment paragraph filling (@kbd{M-q}),
   commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
   removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
   stuff I do not use alone, even though some of it only makes sense for
   TILE. To get a description of these features, enter Forth mode and type
   @kbd{C-h m}.
   
   In addition, GForth supports Emacs quite well: The source code locations
   given in error messages, debugging output (from @code{~~}) and failed
   assertion messages are in the right format for Emacs' compilation mode
   (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
   Manual}) so the source location corresponding to an error or other
   message is only a few keystrokes away (@kbd{C-x `} for the next error,
   @kbd{C-c C-c} for the error under the cursor).
   
   Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
   (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
   contains the definitions of all words defined afterwards. You can then
   find the source for a word using @kbd{M-.}. Note that emacs can use
   several tags files at the same time (e.g., one for the gforth sources
   and one for your program).
   
   To get all these benefits, add the following lines to your @file{.emacs}
   file:
   
   @example
   (autoload 'forth-mode "gforth.el")
   (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
   @end example
   
   @node Internals, Bugs, Emacs and GForth, Top
 @chapter Internals  @chapter Internals
   
 Reading this section is not necessary for programming with gforth. It  Reading this section is not necessary for programming with gforth. It
 should be helpful for finding your way in the gforth sources.  should be helpful for finding your way in the gforth sources.
   
   @menu
   * Portability::                 
   * Threading::                   
   * Primitives::                  
   * System Architecture::         
   @end menu
   
   @node Portability, Threading, Internals, Internals
 @section Portability  @section Portability
   
 One of the main goals of the effort is availability across a wide range  One of the main goals of the effort is availability across a wide range
Line 1395  double numbers. GNU C is available for f Line 2729  double numbers. GNU C is available for f
 unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and  unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and
 the Atari ST, so a Forth written in GNU C can run on all these  the Atari ST, so a Forth written in GNU C can run on all these
 machines@footnote{Due to Apple's look-and-feel lawsuit it is not  machines@footnote{Due to Apple's look-and-feel lawsuit it is not
 available on the Mac (@pxref{Boycott, , Protect Your Freedom--Fight  available on the Mac (@pxref{Boycott, , Protect Your Freedom---Fight
 ``Look And Feel'', gcc.info, GNU C Manual}).}.  ``Look And Feel'', gcc.info, GNU C Manual}).}.
   
 Writing in a portable language has the reputation of producing code that  Writing in a portable language has the reputation of producing code that
Line 1414  machines some compiler versions produce Line 2748  machines some compiler versions produce
 explicit register declarations are used. So by default  explicit register declarations are used. So by default
 @code{-DFORCE_REG} is not used.  @code{-DFORCE_REG} is not used.
   
   @node Threading, Primitives, Portability, Internals
 @section Threading  @section Threading
   
 GNU C's labels as values extension (available since @code{gcc-2.0},  GNU C's labels as values extension (available since @code{gcc-2.0},
Line 1445  goto *ca; Line 2780  goto *ca;
 Of course we have packaged the whole thing neatly in macros called  Of course we have packaged the whole thing neatly in macros called
 @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).  @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
   
   @menu
   * Scheduling::                  
   * Direct or Indirect Threaded?::  
   * DOES>::                       
   @end menu
   
   @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
 @subsection Scheduling  @subsection Scheduling
   
 There is a little complication: Pipelined and superscalar processors,  There is a little complication: Pipelined and superscalar processors,
Line 1463  NEXT; Line 2805  NEXT;
 the NEXT comes strictly after the other code, i.e., there is nearly no  the NEXT comes strictly after the other code, i.e., there is nearly no
 scheduling. After a little thought the problem becomes clear: The  scheduling. After a little thought the problem becomes clear: The
 compiler cannot know that sp and ip point to different addresses (and  compiler cannot know that sp and ip point to different addresses (and
 the version of @code{gcc} we used would not know it even if it could),  the version of @code{gcc} we used would not know it even if it was
 so it could not move the load of the cfa above the store to the  possible), so it could not move the load of the cfa above the store to
 TOS. Indeed the pointers could be the same, if code on or very near the  the TOS. Indeed the pointers could be the same, if code on or very near
 top of stack were executed. In the interest of speed we chose to forbid  the top of stack were executed. In the interest of speed we chose to
 this probably unused ``feature'' and helped the compiler in scheduling:  forbid this probably unused ``feature'' and helped the compiler in
 NEXT is divided into the loading part (@code{NEXT_P1}) and the goto part  scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
 (@code{NEXT_P2}). @code{+} now looks like:  the goto part (@code{NEXT_P2}). @code{+} now looks like:
 @example  @example
 n=sp[0]+sp[1];  n=sp[0]+sp[1];
 sp++;  sp++;
Line 1477  NEXT_P1; Line 2819  NEXT_P1;
 sp[0]=n;  sp[0]=n;
 NEXT_P2;  NEXT_P2;
 @end example  @end example
 This can be scheduled optimally by the compiler (see \sect{TOS}).  This can be scheduled optimally by the compiler.
   
 This division can be turned off with the switch @code{-DCISC_NEXT}. This  This division can be turned off with the switch @code{-DCISC_NEXT}. This
 switch is on by default on machines that do not profit from scheduling  switch is on by default on machines that do not profit from scheduling
 (e.g., the 80386), in order to preserve registers.  (e.g., the 80386), in order to preserve registers.
   
   @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
 @subsection Direct or Indirect Threaded?  @subsection Direct or Indirect Threaded?
   
 Both! After packaging the nasty details in macro definitions we  Both! After packaging the nasty details in macro definitions we
Line 1499  are inherently machine-dependent, but th Line 2842  are inherently machine-dependent, but th
 lines. I.e., even porting direct threading to a new machine is a small  lines. I.e., even porting direct threading to a new machine is a small
 effort.  effort.
   
   @node DOES>,  , Direct or Indirect Threaded?, Threading
 @subsection DOES>  @subsection DOES>
 One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,  One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
 the chunk of code executed by every word defined by a  the chunk of code executed by every word defined by a
Line 1516  again. We use this approach for the indi Line 2860  again. We use this approach for the indi
 a cell unused in most words is a bit wasteful, but on the machines we  a cell unused in most words is a bit wasteful, but on the machines we
 are targetting this is hardly a problem. The other reason for having a  are targetting this is hardly a problem. The other reason for having a
 code field size of two cells is to avoid having different image files  code field size of two cells is to avoid having different image files
 for direct and indirect threaded systems (@pxref{image-format}).  for direct and indirect threaded systems (@pxref{System Architecture}).
   
 The other approach is that the code field points or jumps to the cell  The other approach is that the code field points or jumps to the cell
 after @code{DOES}. In this variant there is a jump to @code{dodoes} at  after @code{DOES}. In this variant there is a jump to @code{dodoes} at
Line 1530  used up by the jump to the code address Line 2874  used up by the jump to the code address
 this approach for direct threading. We did not want to add another  this approach for direct threading. We did not want to add another
 cell to the code field.  cell to the code field.
   
   @node Primitives, System Architecture, Threading, Internals
 @section Primitives  @section Primitives
   
   @menu
   * Automatic Generation::        
   * TOS Optimization::            
   * Produced code::               
   @end menu
   
   @node Automatic Generation, TOS Optimization, Primitives, Primitives
 @subsection Automatic Generation  @subsection Automatic Generation
   
 Since the primitives are implemented in a portable language, there is no  Since the primitives are implemented in a portable language, there is no
Line 1573  looks like this: Line 2925  looks like this:
 @example  @example
 I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */  I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */
 /*  */                          /* documentation */  /*  */                          /* documentation */
 {  @{
 DEF_CA                          /* definition of variable ca (indirect threading) */  DEF_CA                          /* definition of variable ca (indirect threading) */
 Cell n1;                        /* definitions of variables */  Cell n1;                        /* definitions of variables */
 Cell n2;  Cell n2;
Line 1582  n1 = (Cell) sp[1];              /* input Line 2934  n1 = (Cell) sp[1];              /* input
 n2 = (Cell) TOS;  n2 = (Cell) TOS;
 sp += 1;                        /* stack adjustment */  sp += 1;                        /* stack adjustment */
 NAME("+")                       /* debugging output (with -DDEBUG) */  NAME("+")                       /* debugging output (with -DDEBUG) */
 {  @{
 n = n1+n2;                      /* C code taken from the source */  n = n1+n2;                      /* C code taken from the source */
 }  @}
 NEXT_P1;                        /* NEXT part 1 */  NEXT_P1;                        /* NEXT part 1 */
 TOS = (Cell)n;                  /* output */  TOS = (Cell)n;                  /* output */
 NEXT_P2;                        /* NEXT part 2 */  NEXT_P2;                        /* NEXT part 2 */
 }  @}
 @end example  @end example
   
 This looks long and inefficient, but the GNU C compiler optimizes quite  This looks long and inefficient, but the GNU C compiler optimizes quite
Line 1610  where the programmer has to take the act Line 2962  where the programmer has to take the act
 account, most notably @code{?dup}, but also words that do not (always)  account, most notably @code{?dup}, but also words that do not (always)
 fall through to NEXT.  fall through to NEXT.
   
   @node TOS Optimization, Produced code, Automatic Generation, Primitives
 @subsection TOS Optimization  @subsection TOS Optimization
   
 An important optimization for stack machine emulators, e.g., Forth  An important optimization for stack machine emulators, e.g., Forth
 engines, is keeping  one or more of the top stack items in  engines, is keeping  one or more of the top stack items in
 registers.  If a word has the stack effect {@var{in1}...@var{inx} @code{--}  registers.  If a word has the stack effect @var{in1}...@var{inx} @code{--}
 @var{out1}...@var{outy}}, keeping the top @var{n} items in registers  @var{out1}...@var{outy}, keeping the top @var{n} items in registers
 @itemize  @itemize
 @item  @item
 is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},  is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
Line 1654  consider: Line 3007  consider:
 @item In the case of @code{dup ( w -- w w )} the generator must not  @item In the case of @code{dup ( w -- w w )} the generator must not
 eliminate the store to the original location of the item on the stack,  eliminate the store to the original location of the item on the stack,
 if the TOS optimization is turned on.  if the TOS optimization is turned on.
 @item Primitives with stack effects of the form {@code{--}  @item Primitives with stack effects of the form @code{--}
 @var{out1}...@var{outy}} must store the TOS to the stack at the start.  @var{out1}...@var{outy} must store the TOS to the stack at the start.
 Likewise, primitives with the stack effect {@var{in1}...@var{inx} @code{--}}  Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
 must load the TOS from the stack at the end. But for the null stack  must load the TOS from the stack at the end. But for the null stack
 effect @code{--} no stores or loads should be generated.  effect @code{--} no stores or loads should be generated.
 @end itemize  @end itemize
   
   @node Produced code,  , TOS Optimization, Primitives
 @subsection Produced code  @subsection Produced code
   
 To see what assembly code is produced for the primitives on your machine  To see what assembly code is produced for the primitives on your machine
 with your compiler and your flag settings, type @code{make engine.s} and  with your compiler and your flag settings, type @code{make engine.s} and
 look at the resulting file @file{engine.c}.  look at the resulting file @file{engine.s}.
   
   @node System Architecture,  , Primitives, Internals
 @section System Architecture  @section System Architecture
   
 Our Forth system consists not only of primitives, but also of  Our Forth system consists not only of primitives, but also of
Line 1693  file for machines with different cell si Line 3048  file for machines with different cell si
 image file that enables the loader to change the byte order.}.  image file that enables the loader to change the byte order.}.
   
 Forth code that is going to end up in a portable image file has to  Forth code that is going to end up in a portable image file has to
 comply to some restrictions: addresses have to be stored in memory  comply to some restrictions: addresses have to be stored in memory with
 with special words (@code{A!}, @code{A,}, etc.) in order to make the  special words (@code{A!}, @code{A,}, etc.) in order to make the code
 code relocatable. Cells, floats, etc., have to be stored at the  relocatable. Cells, floats, etc., have to be stored at the natural
 natural alignment boundaries@footnote{E.g., store floats (8 bytes) at  alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
 an address dividable by~8. This happens automatically in our system  dividable by~8. This happens automatically in our system when you use
 when you use the ANSI alignment words.}, in order to avoid alignment  the ANS Forth alignment words.}, in order to avoid alignment faults on
 faults on machines with stricter alignment. The image file is produced  machines with stricter alignment. The image file is produced by a
 by a metacompiler (@file{cross.fs}).  metacompiler (@file{cross.fs}).
   
 So, unlike the image file of Mitch Bradleys @code{cforth}, our image  So, unlike the image file of Mitch Bradleys @code{cforth}, our image
 file is not directly executable, but has to undergo some manipulations  file is not directly executable, but has to undergo some manipulations
Line 1709  at run-time. The loader also has to repl Line 3064  at run-time. The loader also has to repl
 primitive calls with the appropriate code-field addresses (or code  primitive calls with the appropriate code-field addresses (or code
 addresses in the case of direct threading).  addresses in the case of direct threading).
   
   @node Bugs, Pedigree, Internals, Top
   @chapter Bugs
   
   @node Pedigree, Word Index, Bugs, Top
   @chapter Pedigree
   
   @node Word Index, Node Index, Pedigree, Top
   @chapter Word Index
   
   @node Node Index,  , Word Index, Top
   @chapter Node Index
   
 @contents  @contents
 @bye  @bye
   

Removed from v.1.3  
changed lines
  Added in v.1.15


FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>