File:  [gforth] / gforth / prim
Revision 1.74: download - view: text, annotated - select for diffs
Sat Feb 24 13:44:39 2001 UTC (23 years ago) by anton
Branches: MAIN
CVS tags: HEAD
added C and primitive support for peeophole optimization

    1: \ Gforth primitives
    2: 
    3: \ Copyright (C) 1995,1996,1997,1998,2000 Free Software Foundation, Inc.
    4: 
    5: \ This file is part of Gforth.
    6: 
    7: \ Gforth is free software; you can redistribute it and/or
    8: \ modify it under the terms of the GNU General Public License
    9: \ as published by the Free Software Foundation; either version 2
   10: \ of the License, or (at your option) any later version.
   11: 
   12: \ This program is distributed in the hope that it will be useful,
   13: \ but WITHOUT ANY WARRANTY; without even the implied warranty of
   14: \ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   15: \ GNU General Public License for more details.
   16: 
   17: \ You should have received a copy of the GNU General Public License
   18: \ along with this program; if not, write to the Free Software
   19: \ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
   20: 
   21: 
   22: \ WARNING: This file is processed by m4. Make sure your identifiers
   23: \ don't collide with m4's (e.g. by undefining them).
   24: \ 
   25: \ 
   26: \ 
   27: \ This file contains primitive specifications in the following format:
   28: \ 
   29: \ forth name	( stack effect )	category	[pronunciation]
   30: \ [""glossary entry""]
   31: \ C code
   32: \ [:
   33: \ Forth code]
   34: \ 
   35: \ Note: Fields in brackets are optional.  Word specifications have to
   36: \ be separated by at least one empty line
   37: \
   38: \ Both pronounciation and stack items (in the stack effect) must
   39: \ conform to the C identifier syntax or the C compiler will complain.
   40: \ If you don't have a pronounciation field, the Forth name is used,
   41: \ and has to conform to the C identifier syntax.
   42: \ 
   43: \ These specifications are automatically translated into C-code for the
   44: \ interpreter and into some other files. I hope that your C compiler has
   45: \ decent optimization, otherwise the automatically generated code will
   46: \ be somewhat slow. The Forth version of the code is included for manual
   47: \ compilers, so they will need to compile only the important words.
   48: \ 
   49: \ Note that stack pointer adjustment is performed according to stack
   50: \ effect by automatically generated code and NEXT is automatically
   51: \ appended to the C code. Also, you can use the names in the stack
   52: \ effect in the C code. Stack access is automatic. One exception: if
   53: \ your code does not fall through, the results are not stored into the
   54: \ stack. Use different names on both sides of the '--', if you change a
   55: \ value (some stores to the stack are optimized away).
   56: \ 
   57: \ 
   58: \ 
   59: \ The stack variables have the following types:
   60: \ 
   61: \ name matches	type
   62: \ f.*		Bool
   63: \ c.*		Char
   64: \ [nw].*		Cell
   65: \ u.*		UCell
   66: \ d.*		DCell
   67: \ ud.*		UDCell
   68: \ r.*		Float
   69: \ a_.*		Cell *
   70: \ c_.*		Char *
   71: \ f_.*		Float *
   72: \ df_.*		DFloat *
   73: \ sf_.*		SFloat *
   74: \ xt.*		XT
   75: \ f83name.*	F83Name *
   76: 
   77: \E get-current prefixes set-current
   78: \E 
   79: \E s" Bool"		single data-stack type-prefix f
   80: \E s" Char"		single data-stack type-prefix c
   81: \E s" Cell"		single data-stack type-prefix n
   82: \E s" Cell"		single data-stack type-prefix w
   83: \E s" UCell"		single data-stack type-prefix u
   84: \E s" DCell"		double data-stack type-prefix d
   85: \E s" UDCell"		double data-stack type-prefix ud
   86: \E s" Float"		single fp-stack   type-prefix r
   87: \E s" Cell *"		single data-stack type-prefix a_
   88: \E s" Char *"		single data-stack type-prefix c_
   89: \E s" Float *"		single data-stack type-prefix f_
   90: \E s" DFloat *"		single data-stack type-prefix df_
   91: \E s" SFloat *"		single data-stack type-prefix sf_
   92: \E s" Xt"		single data-stack type-prefix xt
   93: \E s" struct F83Name *"	single data-stack type-prefix f83name
   94: \E s" struct Longname *" single data-stack type-prefix longname
   95: \E 
   96: \E return-stack stack-prefix R:
   97: \E inst-stream  stack-prefix #
   98: \E 
   99: \E set-current
  100: 
  101: \ 
  102: \ 
  103: \ 
  104: \ In addition the following names can be used:
  105: \ ip	the instruction pointer
  106: \ sp	the data stack pointer
  107: \ rp	the parameter stack pointer
  108: \ lp	the locals stack pointer
  109: \ NEXT	executes NEXT
  110: \ cfa	
  111: \ NEXT1	executes NEXT1
  112: \ FLAG(x)	makes a Forth flag from a C flag
  113: \ 
  114: \ 
  115: \ 
  116: \ Percentages in comments are from Koopmans book: average/maximum use
  117: \ (taken from four, not very representative benchmarks)
  118: \ 
  119: \ 
  120: \ 
  121: \ To do:
  122: \ 
  123: \ throw execute, cfa and NEXT1 out?
  124: \ macroize *ip, ip++, *ip++ (pipelining)?
  125: 
  126: \ these m4 macros would collide with identifiers
  127: undefine(`index')
  128: undefine(`shift')
  129: 
  130: noop	( -- )		gforth
  131: :
  132:  ;
  133: 
  134: lit	( #w -- w )		gforth
  135: :
  136:  r> dup @ swap cell+ >r ;
  137: 
  138: execute	( xt -- )		core
  139: ""Perform the semantics represented by the execution token, @i{xt}.""
  140: ip=IP;
  141: IF_spTOS(spTOS = sp[0]);
  142: EXEC(xt);
  143: 
  144: perform	( a_addr -- )	gforth
  145: ""@code{@@ execute}.""
  146: /* and pfe */
  147: ip=IP;
  148: IF_spTOS(spTOS = sp[0]);
  149: EXEC(*(Xt *)a_addr);
  150: :
  151:  @ execute ;
  152: 
  153: \fhas? skipbranchprims 0= [IF]
  154: \+glocals
  155: 
  156: branch-lp+!#	( #ndisp #nlocals -- )	gforth	branch_lp_plus_store_number
  157: /* this will probably not be used */
  158: lp += nlocals;
  159: SET_IP((Xt *)(((Cell)(IP-2))+ndisp));
  160: 
  161: \+
  162: 
  163: branch	( #ndisp -- )		gforth
  164: SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  165: :
  166:  r> dup @ + >r ;
  167: 
  168: \ condbranch(forthname,stackeffect,restline,code,forthcode)
  169: \ this is non-syntactical: code must open a brace that is closed by the macro
  170: define(condbranch,
  171: $1 ( `#'ndisp $2 ) $3
  172: $4	SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  173: TAIL;
  174: }
  175: $5
  176: 
  177: \+glocals
  178: 
  179: $1-lp+!`#' ( `#'ndisp `#'nlocals $2 ) $3_lp_plus_store_number
  180: $4    lp += nlocals;
  181: SET_IP((Xt *)(((Cell)(IP-2))+ndisp));
  182: TAIL;
  183: }
  184: 
  185: \+
  186: )
  187: 
  188: condbranch(?branch,f --,f83	question_branch,
  189: if (f==0) {
  190: ,:
  191:  0= dup     \ !f !f
  192:  r> dup @   \ !f !f IP branchoffset
  193:  rot and +  \ !f IP|IP+branchoffset
  194:  swap 0= cell and + \ IP''
  195:  >r ;)
  196: 
  197: \ we don't need an lp_plus_store version of the ?dup-stuff, because it
  198: \ is only used in if's (yet)
  199: 
  200: \+xconds
  201: 
  202: ?dup-?branch	( #ndisp f -- f )	new	question_dupe_question_branch
  203: ""The run-time procedure compiled by @code{?DUP-IF}.""
  204: if (f==0) {
  205:   sp++;
  206:   IF_spTOS(spTOS = sp[0]);
  207:   SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  208:   TAIL;
  209: }
  210: 
  211: ?dup-0=-?branch	( #ndisp f -- )	new	question_dupe_zero_equals_question_branch
  212: ""The run-time procedure compiled by @code{?DUP-0=-IF}.""
  213: /* the approach taken here of declaring the word as having the stack
  214: effect ( f -- ) and correcting for it in the branch-taken case costs a
  215: few cycles in that case, but is easy to convert to a CONDBRANCH
  216: invocation */
  217: if (f!=0) {
  218:   sp--;
  219:   SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  220:   NEXT;
  221: }
  222: 
  223: \+
  224: \f[THEN]
  225: \fhas? skiploopprims 0= [IF]
  226: 
  227: condbranch((next),R:n1 -- R:n2,cmFORTH	paren_next,
  228: n2=n1-1;
  229: if (n1) {
  230: ,:
  231:  r> r> dup 1- >r
  232:  IF dup @ + >r ELSE cell+ >r THEN ;)
  233: 
  234: condbranch((loop),R:nlimit R:n1 -- R:nlimit R:n2,gforth	paren_loop,
  235: n2=n1+1;
  236: if (n2 != nlimit) {
  237: ,:
  238:  r> r> 1+ r> 2dup =
  239:  IF >r 1- >r cell+ >r
  240:  ELSE >r >r dup @ + >r THEN ;)
  241: 
  242: condbranch((+loop),n R:nlimit R:n1 -- R:nlimit R:n2,gforth paren_plus_loop,
  243: /* !! check this thoroughly */
  244: /* sign bit manipulation and test: (x^y)<0 is equivalent to (x<0) != (y<0) */
  245: /* dependent upon two's complement arithmetic */
  246: Cell olddiff = n1-nlimit;
  247: n2=n1+n;	
  248: if ((olddiff^(olddiff+n))>=0   /* the limit is not crossed */
  249:     || (olddiff^n)>=0          /* it is a wrap-around effect */) {
  250: ,:
  251:  r> swap
  252:  r> r> 2dup - >r
  253:  2 pick r@ + r@ xor 0< 0=
  254:  3 pick r> xor 0< 0= or
  255:  IF    >r + >r dup @ + >r
  256:  ELSE  >r >r drop cell+ >r THEN ;)
  257: 
  258: \+xconds
  259: 
  260: condbranch((-loop),u R:nlimit R:n1 -- R:nlimit R:n2,gforth paren_minus_loop,
  261: UCell olddiff = n1-nlimit;
  262: n2=n1-u;
  263: if (olddiff>u) {
  264: ,)
  265: 
  266: condbranch((s+loop),n R:nlimit R:n1 -- R:nlimit R:n2,gforth	paren_symmetric_plus_loop,
  267: ""The run-time procedure compiled by S+LOOP. It loops until the index
  268: crosses the boundary between limit and limit-sign(n). I.e. a symmetric
  269: version of (+LOOP).""
  270: /* !! check this thoroughly */
  271: Cell diff = n1-nlimit;
  272: Cell newdiff = diff+n;
  273: if (n<0) {
  274:     diff = -diff;
  275:     newdiff = -newdiff;
  276: }
  277: n2=n1+n;
  278: if (diff>=0 || newdiff<0) {
  279: ,)
  280: 
  281: \+
  282: 
  283: unloop	( R:w1 R:w2 -- )	core
  284: /* !! alias for 2rdrop */
  285: :
  286:  r> rdrop rdrop >r ;
  287: 
  288: (for)	( ncount -- R:nlimit R:ncount )		cmFORTH		paren_for
  289: /* or (for) = >r -- collides with unloop! */
  290: nlimit=0;
  291: :
  292:  r> swap 0 >r >r >r ;
  293: 
  294: (do)	( nlimit nstart -- R:nlimit R:nstart )	gforth		paren_do
  295: :
  296:  r> swap rot >r >r >r ;
  297: 
  298: (?do)	( #ndisp nlimit nstart -- R:nlimit R:nstart )	gforth	paren_question_do
  299: if (nstart == nlimit) {
  300:     SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  301:     TAIL;
  302: }
  303: :
  304:   2dup =
  305:   IF   r> swap rot >r >r
  306:        dup @ + >r
  307:   ELSE r> swap rot >r >r
  308:        cell+ >r
  309:   THEN ;				\ --> CORE-EXT
  310: 
  311: \+xconds
  312: 
  313: (+do)	( #ndisp nlimit nstart -- R:nlimit R:nstart )	gforth	paren_plus_do
  314: if (nstart >= nlimit) {
  315:     SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  316:     TAIL;
  317: }
  318: :
  319:  swap 2dup
  320:  r> swap >r swap >r
  321:  >=
  322:  IF
  323:      dup @ +
  324:  ELSE
  325:      cell+
  326:  THEN  >r ;
  327: 
  328: (u+do)	( #ndisp ulimit ustart -- R:ulimit R:ustart )	gforth	paren_u_plus_do
  329: if (ustart >= ulimit) {
  330:     SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  331:     TAIL;
  332: }
  333: :
  334:  swap 2dup
  335:  r> swap >r swap >r
  336:  u>=
  337:  IF
  338:      dup @ +
  339:  ELSE
  340:      cell+
  341:  THEN  >r ;
  342: 
  343: (-do)	( #ndisp nlimit nstart -- R:nlimit R:nstart )	gforth	paren_minus_do
  344: if (nstart <= nlimit) {
  345:     SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  346:     TAIL;
  347: }
  348: :
  349:  swap 2dup
  350:  r> swap >r swap >r
  351:  <=
  352:  IF
  353:      dup @ +
  354:  ELSE
  355:      cell+
  356:  THEN  >r ;
  357: 
  358: (u-do)	( #ndisp ulimit ustart -- R:ulimit R:ustart )	gforth	paren_u_minus_do
  359: if (ustart <= ulimit) {
  360:     SET_IP((Xt *)(((Cell)(IP-1))+ndisp));
  361:     TAIL;
  362: }
  363: :
  364:  swap 2dup
  365:  r> swap >r swap >r
  366:  u<=
  367:  IF
  368:      dup @ +
  369:  ELSE
  370:      cell+
  371:  THEN  >r ;
  372: 
  373: \+
  374: 
  375: \ don't make any assumptions where the return stack is!!
  376: \ implement this in machine code if it should run quickly!
  377: 
  378: i	( R:n -- R:n n )		core
  379: :
  380: \ rp@ cell+ @ ;
  381:   r> r> tuck >r >r ;
  382: 
  383: i'	( R:w R:w2 -- R:w R:w2 w )		gforth		i_tick
  384: :
  385: \ rp@ cell+ cell+ @ ;
  386:   r> r> r> dup itmp ! >r >r >r itmp @ ;
  387: variable itmp
  388: 
  389: j	( R:n R:d1 -- n R:n R:d1 )		core
  390: :
  391: \ rp@ cell+ cell+ cell+ @ ;
  392:   r> r> r> r> dup itmp ! >r >r >r >r itmp @ ;
  393: [IFUNDEF] itmp variable itmp [THEN]
  394: 
  395: k	( R:n R:d1 R:d2 -- n R:n R:d1 R:d2 )		gforth
  396: :
  397: \ rp@ [ 5 cells ] Literal + @ ;
  398:   r> r> r> r> r> r> dup itmp ! >r >r >r >r >r >r itmp @ ;
  399: [IFUNDEF] itmp variable itmp [THEN]
  400: 
  401: \f[THEN]
  402: 
  403: \ digit is high-level: 0/0%
  404: 
  405: move	( c_from c_to ucount -- )		core
  406: ""Copy the contents of @i{ucount} aus at @i{c-from} to
  407: @i{c-to}. @code{move} works correctly even if the two areas overlap.""
  408: /* !! note that the standard specifies addr, not c-addr */
  409: memmove(c_to,c_from,ucount);
  410: /* make an Ifdef for bsd and others? */
  411: :
  412:  >r 2dup u< IF r> cmove> ELSE r> cmove THEN ;
  413: 
  414: cmove	( c_from c_to u -- )	string	c_move
  415: ""Copy the contents of @i{ucount} characters from data space at
  416: @i{c-from} to @i{c-to}. The copy proceeds @code{char}-by-@code{char}
  417: from low address to high address; i.e., for overlapping areas it is
  418: safe if @i{c-to}=<@i{c-from}.""
  419: while (u-- > 0)
  420:   *c_to++ = *c_from++;
  421: :
  422:  bounds ?DO  dup c@ I c! 1+  LOOP  drop ;
  423: 
  424: cmove>	( c_from c_to u -- )	string	c_move_up
  425: ""Copy the contents of @i{ucount} characters from data space at
  426: @i{c-from} to @i{c-to}. The copy proceeds @code{char}-by-@code{char}
  427: from high address to low address; i.e., for overlapping areas it is
  428: safe if @i{c-to}>=@i{c-from}.""
  429: while (u-- > 0)
  430:   c_to[u] = c_from[u];
  431: :
  432:  dup 0= IF  drop 2drop exit  THEN
  433:  rot over + -rot bounds swap 1-
  434:  DO  1- dup c@ I c!  -1 +LOOP  drop ;
  435: 
  436: fill	( c_addr u c -- )	core
  437: ""Store @i{c} in @i{u} chars starting at @i{c-addr}.""
  438: memset(c_addr,c,u);
  439: :
  440:  -rot bounds
  441:  ?DO  dup I c!  LOOP  drop ;
  442: 
  443: compare	( c_addr1 u1 c_addr2 u2 -- n )	string
  444: ""Compare two strings lexicographically. If they are equal, @i{n} is 0; if
  445: the first string is smaller, @i{n} is -1; if the first string is larger, @i{n}
  446: is 1. Currently this is based on the machine's character
  447: comparison. In the future, this may change to consider the current
  448: locale and its collation order.""
  449: /* close ' to keep fontify happy */ 
  450: n = memcmp(c_addr1, c_addr2, u1<u2 ? u1 : u2);
  451: if (n==0)
  452:   n = u1-u2;
  453: if (n<0)
  454:   n = -1;
  455: else if (n>0)
  456:   n = 1;
  457: :
  458:  rot 2dup swap - >r min swap -text dup
  459:  IF  rdrop  ELSE  drop r> sgn  THEN ;
  460: : sgn ( n -- -1/0/1 )
  461:  dup 0= IF EXIT THEN  0< 2* 1+ ;
  462: 
  463: -text	( c_addr1 u c_addr2 -- n )	new	dash_text
  464: n = memcmp(c_addr1, c_addr2, u);
  465: if (n<0)
  466:   n = -1;
  467: else if (n>0)
  468:   n = 1;
  469: :
  470:  swap bounds
  471:  ?DO  dup c@ I c@ = WHILE  1+  LOOP  drop 0
  472:  ELSE  c@ I c@ - unloop  THEN  sgn ;
  473: : sgn ( n -- -1/0/1 )
  474:  dup 0= IF EXIT THEN  0< 2* 1+ ;
  475: 
  476: toupper	( c1 -- c2 )	gforth
  477: ""If @i{c1} is a lower-case character (in the current locale), @i{c2}
  478: is the equivalent upper-case character. All other characters are unchanged.""
  479: c2 = toupper(c1);
  480: :
  481:  dup [char] a - [ char z char a - 1 + ] Literal u<  bl and - ;
  482: 
  483: capscomp	( c_addr1 u c_addr2 -- n )	new
  484: n = memcasecmp(c_addr1, c_addr2, u); /* !! use something that works in all locales */
  485: if (n<0)
  486:   n = -1;
  487: else if (n>0)
  488:   n = 1;
  489: :
  490:  swap bounds
  491:  ?DO  dup c@ I c@ <>
  492:      IF  dup c@ toupper I c@ toupper =
  493:      ELSE  true  THEN  WHILE  1+  LOOP  drop 0
  494:  ELSE  c@ toupper I c@ toupper - unloop  THEN  sgn ;
  495: 
  496: -trailing	( c_addr u1 -- c_addr u2 )		string	dash_trailing
  497: ""Adjust the string specified by @i{c-addr, u1} to remove all trailing
  498: spaces. @i{u2} is the length of the modified string.""
  499: u2 = u1;
  500: while (u2>0 && c_addr[u2-1] == ' ')
  501:   u2--;
  502: :
  503:  BEGIN  1- 2dup + c@ bl =  WHILE
  504:         dup  0= UNTIL  ELSE  1+  THEN ;
  505: 
  506: /string	( c_addr1 u1 n -- c_addr2 u2 )	string	slash_string
  507: ""Adjust the string specified by @i{c-addr1, u1} to remove @i{n}
  508: characters from the start of the string.""
  509: c_addr2 = c_addr1+n;
  510: u2 = u1-n;
  511: :
  512:  tuck - >r + r> dup 0< IF  - 0  THEN ;
  513: 
  514: +	( n1 n2 -- n )		core	plus
  515: n = n1+n2;
  516: 
  517: \ PFE-0.9.14 has it differently, but the next release will have it as follows
  518: under+	( n1 n2 n3 -- n n2 )	gforth	under_plus
  519: ""add @i{n3} to @i{n1} (giving @i{n})""
  520: n = n1+n3;
  521: :
  522:  rot + swap ;
  523: 
  524: -	( n1 n2 -- n )		core	minus
  525: n = n1-n2;
  526: :
  527:  negate + ;
  528: 
  529: negate	( n1 -- n2 )		core
  530: /* use minus as alias */
  531: n2 = -n1;
  532: :
  533:  invert 1+ ;
  534: 
  535: 1+	( n1 -- n2 )		core		one_plus
  536: n2 = n1+1;
  537: :
  538:  1 + ;
  539: 
  540: 1-	( n1 -- n2 )		core		one_minus
  541: n2 = n1-1;
  542: :
  543:  1 - ;
  544: 
  545: max	( n1 n2 -- n )	core
  546: if (n1<n2)
  547:   n = n2;
  548: else
  549:   n = n1;
  550: :
  551:  2dup < IF swap THEN drop ;
  552: 
  553: min	( n1 n2 -- n )	core
  554: if (n1<n2)
  555:   n = n1;
  556: else
  557:   n = n2;
  558: :
  559:  2dup > IF swap THEN drop ;
  560: 
  561: abs	( n -- u )	core
  562: if (n<0)
  563:   u = -n;
  564: else
  565:   u = n;
  566: :
  567:  dup 0< IF negate THEN ;
  568: 
  569: *	( n1 n2 -- n )		core	star
  570: n = n1*n2;
  571: :
  572:  um* drop ;
  573: 
  574: /	( n1 n2 -- n )		core	slash
  575: n = n1/n2;
  576: :
  577:  /mod nip ;
  578: 
  579: mod	( n1 n2 -- n )		core
  580: n = n1%n2;
  581: :
  582:  /mod drop ;
  583: 
  584: /mod	( n1 n2 -- n3 n4 )		core		slash_mod
  585: n4 = n1/n2;
  586: n3 = n1%n2; /* !! is this correct? look into C standard! */
  587: :
  588:  >r s>d r> fm/mod ;
  589: 
  590: 2*	( n1 -- n2 )		core		two_star
  591: ""Shift left by 1; also works on unsigned numbers""
  592: n2 = 2*n1;
  593: :
  594:  dup + ;
  595: 
  596: 2/	( n1 -- n2 )		core		two_slash
  597: ""Arithmetic shift right by 1.  For signed numbers this is a floored
  598: division by 2 (note that @code{/} not necessarily floors).""
  599: n2 = n1>>1;
  600: :
  601:  dup MINI and IF 1 ELSE 0 THEN
  602:  [ bits/byte cell * 1- ] literal 
  603:  0 DO 2* swap dup 2* >r MINI and 
  604:      IF 1 ELSE 0 THEN or r> swap
  605:  LOOP nip ;
  606: 
  607: fm/mod	( d1 n1 -- n2 n3 )		core		f_m_slash_mod
  608: ""Floored division: @i{d1} = @i{n3}*@i{n1}+@i{n2}, @i{n1}>@i{n2}>=0 or 0>=@i{n2}>@i{n1}.""
  609: #ifdef BUGGY_LONG_LONG
  610: DCell r = fmdiv(d1,n1);
  611: n2=r.hi;
  612: n3=r.lo;
  613: #else
  614: /* assumes that the processor uses either floored or symmetric division */
  615: n3 = d1/n1;
  616: n2 = d1%n1;
  617: /* note that this 1%-3>0 is optimized by the compiler */
  618: if (1%-3>0 && (d1<0) != (n1<0) && n2!=0) {
  619:   n3--;
  620:   n2+=n1;
  621: }
  622: #endif
  623: :
  624:  dup >r dup 0< IF  negate >r dnegate r>  THEN
  625:  over       0< IF  tuck + swap  THEN
  626:  um/mod
  627:  r> 0< IF  swap negate swap  THEN ;
  628: 
  629: sm/rem	( d1 n1 -- n2 n3 )		core		s_m_slash_rem
  630: ""Symmetric division: @i{d1} = @i{n3}*@i{n1}+@i{n2}, sign(@i{n2})=sign(@i{d1}) or 0.""
  631: #ifdef BUGGY_LONG_LONG
  632: DCell r = smdiv(d1,n1);
  633: n2=r.hi;
  634: n3=r.lo;
  635: #else
  636: /* assumes that the processor uses either floored or symmetric division */
  637: n3 = d1/n1;
  638: n2 = d1%n1;
  639: /* note that this 1%-3<0 is optimized by the compiler */
  640: if (1%-3<0 && (d1<0) != (n1<0) && n2!=0) {
  641:   n3++;
  642:   n2-=n1;
  643: }
  644: #endif
  645: :
  646:  over >r dup >r abs -rot
  647:  dabs rot um/mod
  648:  r> r@ xor 0< IF       negate       THEN
  649:  r>        0< IF  swap negate swap  THEN ;
  650: 
  651: m*	( n1 n2 -- d )		core	m_star
  652: #ifdef BUGGY_LONG_LONG
  653: d = mmul(n1,n2);
  654: #else
  655: d = (DCell)n1 * (DCell)n2;
  656: #endif
  657: :
  658:  2dup      0< and >r
  659:  2dup swap 0< and >r
  660:  um* r> - r> - ;
  661: 
  662: um*	( u1 u2 -- ud )		core	u_m_star
  663: /* use u* as alias */
  664: #ifdef BUGGY_LONG_LONG
  665: ud = ummul(u1,u2);
  666: #else
  667: ud = (UDCell)u1 * (UDCell)u2;
  668: #endif
  669: :
  670:    >r >r 0 0 r> r> [ 8 cells ] literal 0
  671:    DO
  672:        over >r dup >r 0< and d2*+ drop
  673:        r> 2* r> swap
  674:    LOOP 2drop ;
  675: : d2*+ ( ud n -- ud+n c )
  676:    over MINI
  677:    and >r >r 2dup d+ swap r> + swap r> ;
  678: 
  679: um/mod	( ud u1 -- u2 u3 )		core	u_m_slash_mod
  680: ""ud=u3*u1+u2, u1>u2>=0""
  681: #ifdef BUGGY_LONG_LONG
  682: UDCell r = umdiv(ud,u1);
  683: u2=r.hi;
  684: u3=r.lo;
  685: #else
  686: u3 = ud/u1;
  687: u2 = ud%u1;
  688: #endif
  689: :
  690:    0 swap [ 8 cells 1 + ] literal 0
  691:    ?DO /modstep
  692:    LOOP drop swap 1 rshift or swap ;
  693: : /modstep ( ud c R: u -- ud-?u c R: u )
  694:    >r over r@ u< 0= or IF r@ - 1 ELSE 0 THEN  d2*+ r> ;
  695: : d2*+ ( ud n -- ud+n c )
  696:    over MINI
  697:    and >r >r 2dup d+ swap r> + swap r> ;
  698: 
  699: m+	( d1 n -- d2 )		double		m_plus
  700: #ifdef BUGGY_LONG_LONG
  701: d2.lo = d1.lo+n;
  702: d2.hi = d1.hi - (n<0) + (d2.lo<d1.lo);
  703: #else
  704: d2 = d1+n;
  705: #endif
  706: :
  707:  s>d d+ ;
  708: 
  709: d+	( d1 d2 -- d )		double	d_plus
  710: #ifdef BUGGY_LONG_LONG
  711: d.lo = d1.lo+d2.lo;
  712: d.hi = d1.hi + d2.hi + (d.lo<d1.lo);
  713: #else
  714: d = d1+d2;
  715: #endif
  716: :
  717:  rot + >r tuck + swap over u> r> swap - ;
  718: 
  719: d-	( d1 d2 -- d )		double		d_minus
  720: #ifdef BUGGY_LONG_LONG
  721: d.lo = d1.lo - d2.lo;
  722: d.hi = d1.hi-d2.hi-(d1.lo<d2.lo);
  723: #else
  724: d = d1-d2;
  725: #endif
  726: :
  727:  dnegate d+ ;
  728: 
  729: dnegate	( d1 -- d2 )		double	d_negate
  730: /* use dminus as alias */
  731: #ifdef BUGGY_LONG_LONG
  732: d2 = dnegate(d1);
  733: #else
  734: d2 = -d1;
  735: #endif
  736: :
  737:  invert swap negate tuck 0= - ;
  738: 
  739: d2*	( d1 -- d2 )		double		d_two_star
  740: ""Shift left by 1; also works on unsigned numbers""
  741: #ifdef BUGGY_LONG_LONG
  742: d2.lo = d1.lo<<1;
  743: d2.hi = (d1.hi<<1) | (d1.lo>>(CELL_BITS-1));
  744: #else
  745: d2 = 2*d1;
  746: #endif
  747: :
  748:  2dup d+ ;
  749: 
  750: d2/	( d1 -- d2 )		double		d_two_slash
  751: ""Arithmetic shift right by 1.  For signed numbers this is a floored
  752: division by 2.""
  753: #ifdef BUGGY_LONG_LONG
  754: d2.hi = d1.hi>>1;
  755: d2.lo= (d1.lo>>1) | (d1.hi<<(CELL_BITS-1));
  756: #else
  757: d2 = d1>>1;
  758: #endif
  759: :
  760:  dup 1 and >r 2/ swap 2/ [ 1 8 cells 1- lshift 1- ] Literal and
  761:  r> IF  [ 1 8 cells 1- lshift ] Literal + THEN  swap ;
  762: 
  763: and	( w1 w2 -- w )		core
  764: w = w1&w2;
  765: 
  766: or	( w1 w2 -- w )		core
  767: w = w1|w2;
  768: :
  769:  invert swap invert and invert ;
  770: 
  771: xor	( w1 w2 -- w )		core	x_or
  772: w = w1^w2;
  773: 
  774: invert	( w1 -- w2 )		core
  775: w2 = ~w1;
  776: :
  777:  MAXU xor ;
  778: 
  779: rshift	( u1 n -- u2 )		core	r_shift
  780: ""Logical shift right by @i{n} bits.""
  781:   u2 = u1>>n;
  782: :
  783:     0 ?DO 2/ MAXI and LOOP ;
  784: 
  785: lshift	( u1 n -- u2 )		core	l_shift
  786:   u2 = u1<<n;
  787: :
  788:     0 ?DO 2* LOOP ;
  789: 
  790: \ comparisons(prefix, args, prefix, arg1, arg2, wordsets...)
  791: define(comparisons,
  792: $1=	( $2 -- f )		$6	$3equals
  793: f = FLAG($4==$5);
  794: :
  795:     [ char $1x char 0 = [IF]
  796: 	] IF false ELSE true THEN [
  797:     [ELSE]
  798: 	] xor 0= [
  799:     [THEN] ] ;
  800: 
  801: $1<>	( $2 -- f )		$7	$3not_equals
  802: f = FLAG($4!=$5);
  803: :
  804:     [ char $1x char 0 = [IF]
  805: 	] IF true ELSE false THEN [
  806:     [ELSE]
  807: 	] xor 0<> [
  808:     [THEN] ] ;
  809: 
  810: $1<	( $2 -- f )		$8	$3less_than
  811: f = FLAG($4<$5);
  812: :
  813:     [ char $1x char 0 = [IF]
  814: 	] MINI and 0<> [
  815:     [ELSE] char $1x char u = [IF]
  816: 	]   2dup xor 0<  IF nip ELSE - THEN 0<  [
  817: 	[ELSE]
  818: 	    ] MINI xor >r MINI xor r> u< [
  819: 	[THEN]
  820:     [THEN] ] ;
  821: 
  822: $1>	( $2 -- f )		$9	$3greater_than
  823: f = FLAG($4>$5);
  824: :
  825:     [ char $1x char 0 = [IF] ] negate [ [ELSE] ] swap [ [THEN] ]
  826:     $1< ;
  827: 
  828: $1<=	( $2 -- f )		gforth	$3less_or_equal
  829: f = FLAG($4<=$5);
  830: :
  831:     $1> 0= ;
  832: 
  833: $1>=	( $2 -- f )		gforth	$3greater_or_equal
  834: f = FLAG($4>=$5);
  835: :
  836:     [ char $1x char 0 = [IF] ] negate [ [ELSE] ] swap [ [THEN] ]
  837:     $1<= ;
  838: 
  839: )
  840: 
  841: comparisons(0, n, zero_, n, 0, core, core-ext, core, core-ext)
  842: comparisons(, n1 n2, , n1, n2, core, core-ext, core, core)
  843: comparisons(u, u1 u2, u_, u1, u2, gforth, gforth, core, core-ext)
  844: 
  845: \ dcomparisons(prefix, args, prefix, arg1, arg2, wordsets...)
  846: define(dcomparisons,
  847: $1=	( $2 -- f )		$6	$3equals
  848: #ifdef BUGGY_LONG_LONG
  849: f = FLAG($4.lo==$5.lo && $4.hi==$5.hi);
  850: #else
  851: f = FLAG($4==$5);
  852: #endif
  853: 
  854: $1<>	( $2 -- f )		$7	$3not_equals
  855: #ifdef BUGGY_LONG_LONG
  856: f = FLAG($4.lo!=$5.lo || $4.hi!=$5.hi);
  857: #else
  858: f = FLAG($4!=$5);
  859: #endif
  860: 
  861: $1<	( $2 -- f )		$8	$3less_than
  862: #ifdef BUGGY_LONG_LONG
  863: f = FLAG($4.hi==$5.hi ? $4.lo<$5.lo : $4.hi<$5.hi);
  864: #else
  865: f = FLAG($4<$5);
  866: #endif
  867: 
  868: $1>	( $2 -- f )		$9	$3greater_than
  869: #ifdef BUGGY_LONG_LONG
  870: f = FLAG($4.hi==$5.hi ? $4.lo>$5.lo : $4.hi>$5.hi);
  871: #else
  872: f = FLAG($4>$5);
  873: #endif
  874: 
  875: $1<=	( $2 -- f )		gforth	$3less_or_equal
  876: #ifdef BUGGY_LONG_LONG
  877: f = FLAG($4.hi==$5.hi ? $4.lo<=$5.lo : $4.hi<=$5.hi);
  878: #else
  879: f = FLAG($4<=$5);
  880: #endif
  881: 
  882: $1>=	( $2 -- f )		gforth	$3greater_or_equal
  883: #ifdef BUGGY_LONG_LONG
  884: f = FLAG($4.hi==$5.hi ? $4.lo>=$5.lo : $4.hi>=$5.hi);
  885: #else
  886: f = FLAG($4>=$5);
  887: #endif
  888: 
  889: )
  890: 
  891: \+dcomps
  892: 
  893: dcomparisons(d, d1 d2, d_, d1, d2, double, gforth, double, gforth)
  894: dcomparisons(d0, d, d_zero_, d, DZERO, double, gforth, double, gforth)
  895: dcomparisons(du, ud1 ud2, d_u_, ud1, ud2, gforth, gforth, double-ext, gforth)
  896: 
  897: \+
  898: 
  899: within	( u1 u2 u3 -- f )		core-ext
  900: ""u2=<u1<u3 or: u3=<u2 and u1 is not in [u3,u2).  This works for
  901: unsigned and signed numbers (but not a mixture).  Another way to think
  902: about this word is to consider the numbers as a circle (wrapping
  903: around from @code{max-u} to 0 for unsigned, and from @code{max-n} to
  904: min-n for signed numbers); now consider the range from u2 towards
  905: increasing numbers up to and excluding u3 (giving an empty range if
  906: u2=u3); if u1 is in this range, @code{within} returns true.""
  907: f = FLAG(u1-u2 < u3-u2);
  908: :
  909:  over - >r - r> u< ;
  910: 
  911: sp@	( -- a_addr )		gforth		sp_fetch
  912: a_addr = sp+1;
  913: 
  914: sp!	( a_addr -- )		gforth		sp_store
  915: sp = a_addr;
  916: /* works with and without spTOS caching */
  917: 
  918: rp@	( -- a_addr )		gforth		rp_fetch
  919: a_addr = rp;
  920: 
  921: rp!	( a_addr -- )		gforth		rp_store
  922: rp = a_addr;
  923: 
  924: \+floating
  925: 
  926: fp@	( -- f_addr )	gforth	fp_fetch
  927: f_addr = fp;
  928: 
  929: fp!	( f_addr -- )	gforth	fp_store
  930: fp = f_addr;
  931: 
  932: \+
  933: 
  934: ;s	( R:w -- )		gforth	semis
  935: ""The primitive compiled by @code{EXIT}.""
  936: SET_IP((Xt *)w);
  937: 
  938: >r	( w -- R:w )		core	to_r
  939: :
  940:  (>r) ;
  941: : (>r)  rp@ cell+ @ rp@ ! rp@ cell+ ! ;
  942: 
  943: r>	( R:w -- w )		core	r_from
  944: :
  945:  rp@ cell+ @ rp@ @ rp@ cell+ ! (rdrop) rp@ ! ;
  946: Create (rdrop) ' ;s A,
  947: 
  948: rdrop	( R:w -- )		gforth
  949: :
  950:  r> r> drop >r ;
  951: 
  952: 2>r	( w1 w2 -- R:w1 R:w2 )	core-ext	two_to_r
  953: :
  954:  swap r> swap >r swap >r >r ;
  955: 
  956: 2r>	( R:w1 R:w2 -- w1 w2 )	core-ext	two_r_from
  957: :
  958:  r> r> swap r> swap >r swap ;
  959: 
  960: 2r@	( R:w1 R:w2 -- R:w1 R:w2 w1 w2 )	core-ext	two_r_fetch
  961: :
  962:  i' j ;
  963: 
  964: 2rdrop	(  R:w1 R:w2 -- )		gforth	two_r_drop
  965: :
  966:  r> r> drop r> drop >r ;
  967: 
  968: over	( w1 w2 -- w1 w2 w1 )		core
  969: :
  970:  sp@ cell+ @ ;
  971: 
  972: drop	( w -- )		core
  973: :
  974:  IF THEN ;
  975: 
  976: swap	( w1 w2 -- w2 w1 )		core
  977: :
  978:  >r (swap) ! r> (swap) @ ;
  979: Variable (swap)
  980: 
  981: dup	( w -- w w )		core	dupe
  982: :
  983:  sp@ @ ;
  984: 
  985: rot	( w1 w2 w3 -- w2 w3 w1 )	core	rote
  986: :
  987: [ defined? (swap) [IF] ]
  988:     (swap) ! (rot) ! >r (rot) @ (swap) @ r> ;
  989: Variable (rot)
  990: [ELSE] ]
  991:     >r swap r> swap ;
  992: [THEN]
  993: 
  994: -rot	( w1 w2 w3 -- w3 w1 w2 )	gforth	not_rote
  995: :
  996:  rot rot ;
  997: 
  998: nip	( w1 w2 -- w2 )		core-ext
  999: :
 1000:  swap drop ;
 1001: 
 1002: tuck	( w1 w2 -- w2 w1 w2 )	core-ext
 1003: :
 1004:  swap over ;
 1005: 
 1006: ?dup	( w -- w )			core	question_dupe
 1007: ""Actually the stack effect is: @code{( w -- 0 | w w )}.  It performs a
 1008: @code{dup} if w is nonzero.""
 1009: if (w!=0) {
 1010:   IF_spTOS(*sp-- = w;)
 1011: #ifndef USE_TOS
 1012:   *--sp = w;
 1013: #endif
 1014: }
 1015: :
 1016:  dup IF dup THEN ;
 1017: 
 1018: pick	( u -- w )			core-ext
 1019: ""Actually the stack effect is @code{ x0 ... xu u -- x0 ... xu x0 }.""
 1020: w = sp[u+1];
 1021: :
 1022:  1+ cells sp@ + @ ;
 1023: 
 1024: 2drop	( w1 w2 -- )		core	two_drop
 1025: :
 1026:  drop drop ;
 1027: 
 1028: 2dup	( w1 w2 -- w1 w2 w1 w2 )	core	two_dupe
 1029: :
 1030:  over over ;
 1031: 
 1032: 2over	( w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2 )	core	two_over
 1033: :
 1034:  3 pick 3 pick ;
 1035: 
 1036: 2swap	( w1 w2 w3 w4 -- w3 w4 w1 w2 )	core	two_swap
 1037: :
 1038:  rot >r rot r> ;
 1039: 
 1040: 2rot	( w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2 )	double-ext	two_rote
 1041: :
 1042:  >r >r 2swap r> r> 2swap ;
 1043: 
 1044: 2nip	( w1 w2 w3 w4 -- w3 w4 )	gforth	two_nip
 1045: :
 1046:  2swap 2drop ;
 1047: 
 1048: 2tuck	( w1 w2 w3 w4 -- w3 w4 w1 w2 w3 w4 )	gforth	two_tuck
 1049: :
 1050:  2swap 2over ;
 1051: 
 1052: \ toggle is high-level: 0.11/0.42%
 1053: 
 1054: @	( a_addr -- w )		core	fetch
 1055: ""@i{w} is the cell stored at @i{a_addr}.""
 1056: w = *a_addr;
 1057: 
 1058: !	( w a_addr -- )		core	store
 1059: ""Store @i{w} into the cell at @i{a-addr}.""
 1060: *a_addr = w;
 1061: 
 1062: +!	( n a_addr -- )		core	plus_store
 1063: ""Add @i{n} to the cell at @i{a-addr}.""
 1064: *a_addr += n;
 1065: :
 1066:  tuck @ + swap ! ;
 1067: 
 1068: c@	( c_addr -- c )		core	c_fetch
 1069: ""@i{c} is the char stored at @i{c_addr}.""
 1070: c = *c_addr;
 1071: :
 1072: [ bigendian [IF] ]
 1073:     [ cell>bit 4 = [IF] ]
 1074: 	dup [ 0 cell - ] Literal and @ swap 1 and
 1075: 	IF  $FF and  ELSE  8>>  THEN  ;
 1076:     [ [ELSE] ]
 1077: 	dup [ cell 1- ] literal and
 1078: 	tuck - @ swap [ cell 1- ] literal xor
 1079:  	0 ?DO 8>> LOOP $FF and
 1080:     [ [THEN] ]
 1081: [ [ELSE] ]
 1082:     [ cell>bit 4 = [IF] ]
 1083: 	dup [ 0 cell - ] Literal and @ swap 1 and
 1084: 	IF  8>>  ELSE  $FF and  THEN
 1085:     [ [ELSE] ]
 1086: 	dup [ cell  1- ] literal and 
 1087: 	tuck - @ swap
 1088: 	0 ?DO 8>> LOOP 255 and
 1089:     [ [THEN] ]
 1090: [ [THEN] ]
 1091: ;
 1092: : 8>> 2/ 2/ 2/ 2/  2/ 2/ 2/ 2/ ;
 1093: 
 1094: c!	( c c_addr -- )		core	c_store
 1095: ""Store @i{c} into the char at @i{c-addr}.""
 1096: *c_addr = c;
 1097: :
 1098: [ bigendian [IF] ]
 1099:     [ cell>bit 4 = [IF] ]
 1100: 	tuck 1 and IF  $FF and  ELSE  8<<  THEN >r
 1101: 	dup -2 and @ over 1 and cells masks + @ and
 1102: 	r> or swap -2 and ! ;
 1103: 	Create masks $00FF , $FF00 ,
 1104:     [ELSE] ]
 1105: 	dup [ cell 1- ] literal and dup 
 1106: 	[ cell 1- ] literal xor >r
 1107: 	- dup @ $FF r@ 0 ?DO 8<< LOOP invert and
 1108: 	rot $FF and r> 0 ?DO 8<< LOOP or swap ! ;
 1109:     [THEN]
 1110: [ELSE] ]
 1111:     [ cell>bit 4 = [IF] ]
 1112: 	tuck 1 and IF  8<<  ELSE  $FF and  THEN >r
 1113: 	dup -2 and @ over 1 and cells masks + @ and
 1114: 	r> or swap -2 and ! ;
 1115: 	Create masks $FF00 , $00FF ,
 1116:     [ELSE] ]
 1117: 	dup [ cell 1- ] literal and dup >r
 1118: 	- dup @ $FF r@ 0 ?DO 8<< LOOP invert and
 1119: 	rot $FF and r> 0 ?DO 8<< LOOP or swap ! ;
 1120:     [THEN]
 1121: [THEN]
 1122: : 8<< 2* 2* 2* 2*  2* 2* 2* 2* ;
 1123: 
 1124: 2!	( w1 w2 a_addr -- )		core	two_store
 1125: ""Store @i{w2} into the cell at @i{c-addr} and @i{w1} into the next cell.""
 1126: a_addr[0] = w2;
 1127: a_addr[1] = w1;
 1128: :
 1129:  tuck ! cell+ ! ;
 1130: 
 1131: 2@	( a_addr -- w1 w2 )		core	two_fetch
 1132: ""@i{w2} is the content of the cell stored at @i{a-addr}, @i{w1} is
 1133: the content of the next cell.""
 1134: w2 = a_addr[0];
 1135: w1 = a_addr[1];
 1136: :
 1137:  dup cell+ @ swap @ ;
 1138: 
 1139: cell+	( a_addr1 -- a_addr2 )	core	cell_plus
 1140: ""@code{1 cells +}""
 1141: a_addr2 = a_addr1+1;
 1142: :
 1143:  cell + ;
 1144: 
 1145: cells	( n1 -- n2 )		core
 1146: "" @i{n2} is the number of address units of @i{n1} cells.""
 1147: n2 = n1 * sizeof(Cell);
 1148: :
 1149:  [ cell
 1150:  2/ dup [IF] ] 2* [ [THEN]
 1151:  2/ dup [IF] ] 2* [ [THEN]
 1152:  2/ dup [IF] ] 2* [ [THEN]
 1153:  2/ dup [IF] ] 2* [ [THEN]
 1154:  drop ] ;
 1155: 
 1156: char+	( c_addr1 -- c_addr2 )	core	char_plus
 1157: ""@code{1 chars +}.""
 1158: c_addr2 = c_addr1 + 1;
 1159: :
 1160:  1+ ;
 1161: 
 1162: (chars)	( n1 -- n2 )	gforth	paren_chars
 1163: n2 = n1 * sizeof(Char);
 1164: :
 1165:  ;
 1166: 
 1167: count	( c_addr1 -- c_addr2 u )	core
 1168: ""@i{c-addr2} is the first character and @i{u} the length of the
 1169: counted string at @i{c-addr1}.""
 1170: u = *c_addr1;
 1171: c_addr2 = c_addr1+1;
 1172: :
 1173:  dup 1+ swap c@ ;
 1174: 
 1175: (f83find)	( c_addr u f83name1 -- f83name2 )	new	paren_f83find
 1176: for (; f83name1 != NULL; f83name1 = (struct F83Name *)(f83name1->next))
 1177:   if ((UCell)F83NAME_COUNT(f83name1)==u &&
 1178:       memcasecmp(c_addr, f83name1->name, u)== 0 /* or inline? */)
 1179:     break;
 1180: f83name2=f83name1;
 1181: :
 1182:     BEGIN  dup WHILE  (find-samelen)  dup  WHILE
 1183: 	>r 2dup r@ cell+ char+ capscomp  0=
 1184: 	IF  2drop r>  EXIT  THEN
 1185: 	r> @
 1186:     REPEAT  THEN  nip nip ;
 1187: : (find-samelen) ( u f83name1 -- u f83name2/0 )
 1188:     BEGIN  2dup cell+ c@ $1F and <> WHILE  @  dup 0= UNTIL THEN ;
 1189: 
 1190: \+hash
 1191: 
 1192: (hashfind)	( c_addr u a_addr -- f83name2 )	new	paren_hashfind
 1193: struct F83Name *f83name1;
 1194: f83name2=NULL;
 1195: while(a_addr != NULL)
 1196: {
 1197:    f83name1=(struct F83Name *)(a_addr[1]);
 1198:    a_addr=(Cell *)(a_addr[0]);
 1199:    if ((UCell)F83NAME_COUNT(f83name1)==u &&
 1200:        memcasecmp(c_addr, f83name1->name, u)== 0 /* or inline? */)
 1201:      {
 1202: 	f83name2=f83name1;
 1203: 	break;
 1204:      }
 1205: }
 1206: :
 1207:  BEGIN  dup  WHILE
 1208:         2@ >r >r dup r@ cell+ c@ $1F and =
 1209:         IF  2dup r@ cell+ char+ capscomp 0=
 1210: 	    IF  2drop r> rdrop  EXIT  THEN  THEN
 1211: 	rdrop r>
 1212:  REPEAT nip nip ;
 1213: 
 1214: (tablefind)	( c_addr u a_addr -- f83name2 )	new	paren_tablefind
 1215: ""A case-sensitive variant of @code{(hashfind)}""
 1216: struct F83Name *f83name1;
 1217: f83name2=NULL;
 1218: while(a_addr != NULL)
 1219: {
 1220:    f83name1=(struct F83Name *)(a_addr[1]);
 1221:    a_addr=(Cell *)(a_addr[0]);
 1222:    if ((UCell)F83NAME_COUNT(f83name1)==u &&
 1223:        memcmp(c_addr, f83name1->name, u)== 0 /* or inline? */)
 1224:      {
 1225: 	f83name2=f83name1;
 1226: 	break;
 1227:      }
 1228: }
 1229: :
 1230:  BEGIN  dup  WHILE
 1231:         2@ >r >r dup r@ cell+ c@ $1F and =
 1232:         IF  2dup r@ cell+ char+ -text 0=
 1233: 	    IF  2drop r> rdrop  EXIT  THEN  THEN
 1234: 	rdrop r>
 1235:  REPEAT nip nip ;
 1236: 
 1237: (hashkey)	( c_addr u1 -- u2 )		gforth	paren_hashkey
 1238: u2=0;
 1239: while(u1--)
 1240:    u2+=(Cell)toupper(*c_addr++);
 1241: :
 1242:  0 -rot bounds ?DO  I c@ toupper +  LOOP ;
 1243: 
 1244: (hashkey1)	( c_addr u ubits -- ukey )		gforth	paren_hashkey1
 1245: ""ukey is the hash key for the string c_addr u fitting in ubits bits""
 1246: /* this hash function rotates the key at every step by rot bits within
 1247:    ubits bits and xors it with the character. This function does ok in
 1248:    the chi-sqare-test.  Rot should be <=7 (preferably <=5) for
 1249:    ASCII strings (larger if ubits is large), and should share no
 1250:    divisors with ubits.
 1251: */
 1252: unsigned rot = ((char []){5,0,1,2,3,4,5,5,5,5,3,5,5,5,5,7,5,5,5,5,7,5,5,5,5,6,5,5,5,5,7,5,5})[ubits];
 1253: Char *cp = c_addr;
 1254: for (ukey=0; cp<c_addr+u; cp++)
 1255:     ukey = ((((ukey<<rot) | (ukey>>(ubits-rot))) 
 1256: 	     ^ toupper(*cp))
 1257: 	    & ((1<<ubits)-1));
 1258: :
 1259:  dup rot-values + c@ over 1 swap lshift 1- >r
 1260:  tuck - 2swap r> 0 2swap bounds
 1261:  ?DO  dup 4 pick lshift swap 3 pick rshift or
 1262:       I c@ toupper xor
 1263:       over and  LOOP
 1264:  nip nip nip ;
 1265: Create rot-values
 1266:   5 c, 0 c, 1 c, 2 c, 3 c,  4 c, 5 c, 5 c, 5 c, 5 c,
 1267:   3 c, 5 c, 5 c, 5 c, 5 c,  7 c, 5 c, 5 c, 5 c, 5 c,
 1268:   7 c, 5 c, 5 c, 5 c, 5 c,  6 c, 5 c, 5 c, 5 c, 5 c,
 1269:   7 c, 5 c, 5 c,
 1270: 
 1271: \+
 1272: 
 1273: (parse-white)	( c_addr1 u1 -- c_addr2 u2 )	gforth	paren_parse_white
 1274: /* use !isgraph instead of isspace? */
 1275: Char *endp = c_addr1+u1;
 1276: while (c_addr1<endp && isspace(*c_addr1))
 1277:   c_addr1++;
 1278: if (c_addr1<endp) {
 1279:   for (c_addr2 = c_addr1; c_addr1<endp && !isspace(*c_addr1); c_addr1++)
 1280:     ;
 1281:   u2 = c_addr1-c_addr2;
 1282: }
 1283: else {
 1284:   c_addr2 = c_addr1;
 1285:   u2 = 0;
 1286: }
 1287: :
 1288:  BEGIN  dup  WHILE  over c@ bl <=  WHILE  1 /string
 1289:  REPEAT  THEN  2dup
 1290:  BEGIN  dup  WHILE  over c@ bl >   WHILE  1 /string
 1291:  REPEAT  THEN  nip - ;
 1292: 
 1293: aligned	( c_addr -- a_addr )	core
 1294: "" @i{a-addr} is the first aligned address greater than or equal to @i{c-addr}.""
 1295: a_addr = (Cell *)((((Cell)c_addr)+(sizeof(Cell)-1))&(-sizeof(Cell)));
 1296: :
 1297:  [ cell 1- ] Literal + [ -1 cells ] Literal and ;
 1298: 
 1299: faligned	( c_addr -- f_addr )	float	f_aligned
 1300: "" @i{f-addr} is the first float-aligned address greater than or equal to @i{c-addr}.""
 1301: f_addr = (Float *)((((Cell)c_addr)+(sizeof(Float)-1))&(-sizeof(Float)));
 1302: :
 1303:  [ 1 floats 1- ] Literal + [ -1 floats ] Literal and ;
 1304: 
 1305: >body	( xt -- a_addr )	core	to_body
 1306: "" Get the address of the body of the word represented by @i{xt} (the address
 1307: of the word's data field).""
 1308: a_addr = PFA(xt);
 1309: :
 1310:     2 cells + ;
 1311: 
 1312: \ threading stuff is currently only interesting if we have a compiler
 1313: \fhas? standardthreading has? compiler and [IF]
 1314: 
 1315: >code-address	( xt -- c_addr )		gforth	to_code_address
 1316: ""@i{c-addr} is the code address of the word @i{xt}.""
 1317: /* !! This behaves installation-dependently for DOES-words */
 1318: c_addr = (Address)CODE_ADDRESS(xt);
 1319: :
 1320:     @ ;
 1321: 
 1322: >does-code	( xt -- a_addr )		gforth	to_does_code
 1323: ""If @i{xt} is the execution token of a child of a @code{DOES>} word,
 1324: @i{a-addr} is the start of the Forth code after the @code{DOES>};
 1325: Otherwise @i{a-addr} is 0.""
 1326: a_addr = (Cell *)DOES_CODE(xt);
 1327: :
 1328:     cell+ @ ;
 1329: 
 1330: code-address!	( c_addr xt -- )		gforth	code_address_store
 1331: ""Create a code field with code address @i{c-addr} at @i{xt}.""
 1332: MAKE_CF(xt, c_addr);
 1333: CACHE_FLUSH(xt,(size_t)PFA(0));
 1334: :
 1335:     ! ;
 1336: 
 1337: does-code!	( a_addr xt -- )		gforth	does_code_store
 1338: ""Create a code field at @i{xt} for a child of a @code{DOES>}-word;
 1339: @i{a-addr} is the start of the Forth code after @code{DOES>}.""
 1340: MAKE_DOES_CF(xt, a_addr);
 1341: CACHE_FLUSH(xt,(size_t)PFA(0));
 1342: :
 1343:     dodoes: over ! cell+ ! ;
 1344: 
 1345: does-handler!	( a_addr -- )	gforth	does_handler_store
 1346: ""Create a @code{DOES>}-handler at address @i{a-addr}. Normally,
 1347: @i{a-addr} points just behind a @code{DOES>}.""
 1348: MAKE_DOES_HANDLER(a_addr);
 1349: CACHE_FLUSH((caddr_t)a_addr,DOES_HANDLER_SIZE);
 1350: :
 1351:     drop ;
 1352: 
 1353: /does-handler	( -- n )	gforth	slash_does_handler
 1354: ""The size of a @code{DOES>}-handler (includes possible padding).""
 1355: /* !! a constant or environmental query might be better */
 1356: n = DOES_HANDLER_SIZE;
 1357: :
 1358:     2 cells ;
 1359: 
 1360: threading-method	( -- n )	gforth	threading_method
 1361: ""0 if the engine is direct threaded. Note that this may change during
 1362: the lifetime of an image.""
 1363: #if defined(DOUBLY_INDIRECT)
 1364: n=2;
 1365: #else
 1366: # if defined(DIRECT_THREADED)
 1367: n=0;
 1368: # else
 1369: n=1;
 1370: # endif
 1371: #endif
 1372: :
 1373:  1 ;
 1374: 
 1375: \f[THEN]
 1376: 
 1377: key-file	( wfileid -- n )		gforth	paren_key_file
 1378: #ifdef HAS_FILE
 1379: fflush(stdout);
 1380: n = key((FILE*)wfileid);
 1381: #else
 1382: n = key(stdin);
 1383: #endif
 1384: 
 1385: key?-file	( wfileid -- n )		facility	key_q_file
 1386: #ifdef HAS_FILE
 1387: fflush(stdout);
 1388: n = key_query((FILE*)wfileid);
 1389: #else
 1390: n = key_query(stdin);
 1391: #endif
 1392: 
 1393: \+os
 1394: 
 1395: stdin	( -- wfileid )	gforth
 1396: wfileid = (Cell)stdin;
 1397: 
 1398: stdout	( -- wfileid )	gforth
 1399: wfileid = (Cell)stdout;
 1400: 
 1401: stderr	( -- wfileid )	gforth
 1402: wfileid = (Cell)stderr;
 1403: 
 1404: form	( -- urows ucols )	gforth
 1405: ""The number of lines and columns in the terminal. These numbers may change
 1406: with the window size.""
 1407: /* we could block SIGWINCH here to get a consistent size, but I don't
 1408:  think this is necessary or always beneficial */
 1409: urows=rows;
 1410: ucols=cols;
 1411: 
 1412: flush-icache	( c_addr u -- )	gforth	flush_icache
 1413: ""Make sure that the instruction cache of the processor (if there is
 1414: one) does not contain stale data at @i{c-addr} and @i{u} bytes
 1415: afterwards. @code{END-CODE} performs a @code{flush-icache}
 1416: automatically. Caveat: @code{flush-icache} might not work on your
 1417: installation; this is usually the case if direct threading is not
 1418: supported on your machine (take a look at your @file{machine.h}) and
 1419: your machine has a separate instruction cache. In such cases,
 1420: @code{flush-icache} does nothing instead of flushing the instruction
 1421: cache.""
 1422: FLUSH_ICACHE(c_addr,u);
 1423: 
 1424: (bye)	( n -- )	gforth	paren_bye
 1425: return (Label *)n;
 1426: 
 1427: (system)	( c_addr u -- wretval wior )	gforth	peren_system
 1428: #ifndef MSDOS
 1429: int old_tp=terminal_prepped;
 1430: deprep_terminal();
 1431: #endif
 1432: wretval=system(cstr(c_addr,u,1)); /* ~ expansion on first part of string? */
 1433: wior = IOR(wretval==-1 || (wretval==127 && errno != 0));
 1434: #ifndef MSDOS
 1435: if (old_tp)
 1436:   prep_terminal();
 1437: #endif
 1438: 
 1439: getenv	( c_addr1 u1 -- c_addr2 u2 )	gforth
 1440: ""The string @i{c-addr1 u1} specifies an environment variable. The string @i{c-addr2 u2}
 1441: is the host operating system's expansion of that environment variable. If the
 1442: environment variable does not exist, @i{c-addr2 u2} specifies a string 0 characters
 1443: in length.""
 1444: /* close ' to keep fontify happy */
 1445: c_addr2 = getenv(cstr(c_addr1,u1,1));
 1446: u2 = (c_addr2 == NULL ? 0 : strlen(c_addr2));
 1447: 
 1448: open-pipe	( c_addr u wfam -- wfileid wior )	gforth	open_pipe
 1449: wfileid=(Cell)popen(cstr(c_addr,u,1),fileattr[wfam]); /* ~ expansion of 1st arg? */
 1450: wior = IOR(wfileid==0); /* !! the man page says that errno is not set reliably */
 1451: 
 1452: close-pipe	( wfileid -- wretval wior )		gforth	close_pipe
 1453: wretval = pclose((FILE *)wfileid);
 1454: wior = IOR(wretval==-1);
 1455: 
 1456: time&date	( -- nsec nmin nhour nday nmonth nyear )	facility-ext	time_and_date
 1457: ""Report the current time of day. Seconds, minutes and hours are numbered from 0.
 1458: Months are numbered from 1.""
 1459: struct timeval time1;
 1460: struct timezone zone1;
 1461: struct tm *ltime;
 1462: gettimeofday(&time1,&zone1);
 1463: /* !! Single Unix specification: 
 1464:    If tzp is not a null pointer, the behaviour is unspecified. */
 1465: ltime=localtime((time_t *)&time1.tv_sec);
 1466: nyear =ltime->tm_year+1900;
 1467: nmonth=ltime->tm_mon+1;
 1468: nday  =ltime->tm_mday;
 1469: nhour =ltime->tm_hour;
 1470: nmin  =ltime->tm_min;
 1471: nsec  =ltime->tm_sec;
 1472: 
 1473: ms	( n -- )	facility-ext
 1474: ""Wait at least @i{n} milli-second.""
 1475: struct timeval timeout;
 1476: timeout.tv_sec=n/1000;
 1477: timeout.tv_usec=1000*(n%1000);
 1478: (void)select(0,0,0,0,&timeout);
 1479: 
 1480: allocate	( u -- a_addr wior )	memory
 1481: ""Allocate @i{u} address units of contiguous data space. The initial
 1482: contents of the data space is undefined. If the allocation is successful,
 1483: @i{a-addr} is the start address of the allocated region and @i{wior}
 1484: is 0. If the allocation fails, @i{a-addr} is undefined and @i{wior}
 1485: is a non-zero I/O result code.""
 1486: a_addr = (Cell *)malloc(u?u:1);
 1487: wior = IOR(a_addr==NULL);
 1488: 
 1489: free	( a_addr -- wior )		memory
 1490: ""Return the region of data space starting at @i{a-addr} to the system.
 1491: The region must originally have been obtained using @code{allocate} or
 1492: @code{resize}. If the operational is successful, @i{wior} is 0.
 1493: If the operation fails, @i{wior} is a non-zero I/O result code.""
 1494: free(a_addr);
 1495: wior = 0;
 1496: 
 1497: resize	( a_addr1 u -- a_addr2 wior )	memory
 1498: ""Change the size of the allocated area at @i{a-addr1} to @i{u}
 1499: address units, possibly moving the contents to a different
 1500: area. @i{a-addr2} is the address of the resulting area.
 1501: If the operation is successful, @i{wior} is 0.
 1502: If the operation fails, @i{wior} is a non-zero
 1503: I/O result code. If @i{a-addr1} is 0, Gforth's (but not the Standard)
 1504: @code{resize} @code{allocate}s @i{u} address units.""
 1505: /* the following check is not necessary on most OSs, but it is needed
 1506:    on SunOS 4.1.2. */
 1507: /* close ' to keep fontify happy */
 1508: if (a_addr1==NULL)
 1509:   a_addr2 = (Cell *)malloc(u);
 1510: else
 1511:   a_addr2 = (Cell *)realloc(a_addr1, u);
 1512: wior = IOR(a_addr2==NULL);	/* !! Define a return code */
 1513: 
 1514: strerror	( n -- c_addr u )	gforth
 1515: c_addr = strerror(n);
 1516: u = strlen(c_addr);
 1517: 
 1518: strsignal	( n -- c_addr u )	gforth
 1519: c_addr = strsignal(n);
 1520: u = strlen(c_addr);
 1521: 
 1522: call-c	( w -- )	gforth	call_c
 1523: ""Call the C function pointed to by @i{w}. The C function has to
 1524: access the stack itself. The stack pointers are exported in the global
 1525: variables @code{SP} and @code{FP}.""
 1526: /* This is a first attempt at support for calls to C. This may change in
 1527:    the future */
 1528: IF_fpTOS(fp[0]=fpTOS);
 1529: FP=fp;
 1530: SP=sp;
 1531: ((void (*)())w)();
 1532: sp=SP;
 1533: fp=FP;
 1534: IF_spTOS(spTOS=sp[0]);
 1535: IF_fpTOS(fpTOS=fp[0]);
 1536: 
 1537: \+
 1538: \+file
 1539: 
 1540: close-file	( wfileid -- wior )		file	close_file
 1541: wior = IOR(fclose((FILE *)wfileid)==EOF);
 1542: 
 1543: open-file	( c_addr u wfam -- wfileid wior )	file	open_file
 1544: wfileid = (Cell)fopen(tilde_cstr(c_addr, u, 1), fileattr[wfam]);
 1545: wior =  IOR(wfileid == 0);
 1546: 
 1547: create-file	( c_addr u wfam -- wfileid wior )	file	create_file
 1548: Cell	fd;
 1549: fd = open(tilde_cstr(c_addr, u, 1), O_CREAT|O_TRUNC|ufileattr[wfam], 0666);
 1550: if (fd != -1) {
 1551:   wfileid = (Cell)fdopen(fd, fileattr[wfam]);
 1552:   wior = IOR(wfileid == 0);
 1553: } else {
 1554:   wfileid = 0;
 1555:   wior = IOR(1);
 1556: }
 1557: 
 1558: delete-file	( c_addr u -- wior )		file	delete_file
 1559: wior = IOR(unlink(tilde_cstr(c_addr, u, 1))==-1);
 1560: 
 1561: rename-file	( c_addr1 u1 c_addr2 u2 -- wior )	file-ext	rename_file
 1562: ""Rename file @i{c_addr1 u1} to new name @i{c_addr2 u2}""
 1563: char *s1=tilde_cstr(c_addr2, u2, 1);
 1564: wior = IOR(rename(tilde_cstr(c_addr1, u1, 0), s1)==-1);
 1565: 
 1566: file-position	( wfileid -- ud wior )	file	file_position
 1567: /* !! use tell and lseek? */
 1568: ud = LONG2UD(ftell((FILE *)wfileid));
 1569: wior = IOR(UD2LONG(ud)==-1);
 1570: 
 1571: reposition-file	( ud wfileid -- wior )	file	reposition_file
 1572: wior = IOR(fseek((FILE *)wfileid, UD2LONG(ud), SEEK_SET)==-1);
 1573: 
 1574: file-size	( wfileid -- ud wior )	file	file_size
 1575: struct stat buf;
 1576: wior = IOR(fstat(fileno((FILE *)wfileid), &buf)==-1);
 1577: ud = LONG2UD(buf.st_size);
 1578: 
 1579: resize-file	( ud wfileid -- wior )	file	resize_file
 1580: wior = IOR(ftruncate(fileno((FILE *)wfileid), UD2LONG(ud))==-1);
 1581: 
 1582: read-file	( c_addr u1 wfileid -- u2 wior )	file	read_file
 1583: /* !! fread does not guarantee enough */
 1584: u2 = fread(c_addr, sizeof(Char), u1, (FILE *)wfileid);
 1585: wior = FILEIO(u2<u1 && ferror((FILE *)wfileid));
 1586: /* !! is the value of ferror errno-compatible? */
 1587: if (wior)
 1588:   clearerr((FILE *)wfileid);
 1589: 
 1590: read-line	( c_addr u1 wfileid -- u2 flag wior )	file	read_line
 1591: ""this is only for backward compatibility""
 1592: Cell c;
 1593: flag=-1;
 1594: for(u2=0; u2<u1; u2++)
 1595: {
 1596:    c = getc((FILE *)wfileid);
 1597:    if (c=='\n') break;
 1598:    if (c=='\r') {
 1599:      if ((c = getc((FILE *)wfileid))!='\n')
 1600:        ungetc(c,(FILE *)wfileid);
 1601:      break;
 1602:    }
 1603:    if (c==EOF) {
 1604: 	flag=FLAG(u2!=0);
 1605: 	break;
 1606:      }
 1607:    c_addr[u2] = (Char)c;
 1608: }
 1609: wior=FILEIO(ferror((FILE *)wfileid));
 1610: 
 1611: \+
 1612: 
 1613: write-file	( c_addr u1 wfileid -- wior )	file	write_file
 1614: /* !! fwrite does not guarantee enough */
 1615: #ifdef HAS_FILE
 1616: {
 1617:   UCell u2 = fwrite(c_addr, sizeof(Char), u1, (FILE *)wfileid);
 1618:   wior = FILEIO(u2<u1 && ferror((FILE *)wfileid));
 1619:   if (wior)
 1620:     clearerr((FILE *)wfileid);
 1621: }
 1622: #else
 1623: TYPE(c_addr, u1);
 1624: #endif
 1625: 
 1626: emit-file	( c wfileid -- wior )	gforth	emit_file
 1627: #ifdef HAS_FILE
 1628: wior = FILEIO(putc(c, (FILE *)wfileid)==EOF);
 1629: if (wior)
 1630:   clearerr((FILE *)wfileid);
 1631: #else
 1632: PUTC(c);
 1633: #endif
 1634: 
 1635: \+file
 1636: 
 1637: flush-file	( wfileid -- wior )		file-ext	flush_file
 1638: wior = IOR(fflush((FILE *) wfileid)==EOF);
 1639: 
 1640: file-status	( c_addr u -- wfam wior )	file-ext	file_status
 1641: char *filename=tilde_cstr(c_addr, u, 1);
 1642: if (access (filename, F_OK) != 0) {
 1643:   wfam=0;
 1644:   wior=IOR(1);
 1645: }
 1646: else if (access (filename, R_OK | W_OK) == 0) {
 1647:   wfam=2; /* r/w */
 1648:   wior=0;
 1649: }
 1650: else if (access (filename, R_OK) == 0) {
 1651:   wfam=0; /* r/o */
 1652:   wior=0;
 1653: }
 1654: else if (access (filename, W_OK) == 0) {
 1655:   wfam=4; /* w/o */
 1656:   wior=0;
 1657: }
 1658: else {
 1659:   wfam=1; /* well, we cannot access the file, but better deliver a legal
 1660: 	    access mode (r/o bin), so we get a decent error later upon open. */
 1661:   wior=0;
 1662: }
 1663: 
 1664: \+
 1665: \+floating
 1666: 
 1667: comparisons(f, r1 r2, f_, r1, r2, gforth, gforth, float, gforth)
 1668: comparisons(f0, r, f_zero_, r, 0., float, gforth, float, gforth)
 1669: 
 1670: d>f	( d -- r )		float	d_to_f
 1671: #ifdef BUGGY_LONG_LONG
 1672: extern double ldexp(double x, int exp);
 1673: r = ldexp((Float)d.hi,CELL_BITS) + (Float)d.lo;
 1674: #else
 1675: r = d;
 1676: #endif
 1677: 
 1678: f>d	( r -- d )		float	f_to_d
 1679: #ifdef BUGGY_LONG_LONG
 1680: d.hi = ldexp(r,-(int)(CELL_BITS)) - (r<0);
 1681: d.lo = r-ldexp((Float)d.hi,CELL_BITS);
 1682: #else
 1683: d = r;
 1684: #endif
 1685: 
 1686: f!	( r f_addr -- )	float	f_store
 1687: ""Store @i{r} into the float at address @i{f-addr}.""
 1688: *f_addr = r;
 1689: 
 1690: f@	( f_addr -- r )	float	f_fetch
 1691: ""@i{r} is the float at address @i{f-addr}.""
 1692: r = *f_addr;
 1693: 
 1694: df@	( df_addr -- r )	float-ext	d_f_fetch
 1695: ""Fetch the double-precision IEEE floating-point value @i{r} from the address @i{df-addr}.""
 1696: #ifdef IEEE_FP
 1697: r = *df_addr;
 1698: #else
 1699: !! df@
 1700: #endif
 1701: 
 1702: df!	( r df_addr -- )	float-ext	d_f_store
 1703: ""Store @i{r} as double-precision IEEE floating-point value to the
 1704: address @i{df-addr}.""
 1705: #ifdef IEEE_FP
 1706: *df_addr = r;
 1707: #else
 1708: !! df!
 1709: #endif
 1710: 
 1711: sf@	( sf_addr -- r )	float-ext	s_f_fetch
 1712: ""Fetch the single-precision IEEE floating-point value @i{r} from the address @i{sf-addr}.""
 1713: #ifdef IEEE_FP
 1714: r = *sf_addr;
 1715: #else
 1716: !! sf@
 1717: #endif
 1718: 
 1719: sf!	( r sf_addr -- )	float-ext	s_f_store
 1720: ""Store @i{r} as single-precision IEEE floating-point value to the
 1721: address @i{sf-addr}.""
 1722: #ifdef IEEE_FP
 1723: *sf_addr = r;
 1724: #else
 1725: !! sf!
 1726: #endif
 1727: 
 1728: f+	( r1 r2 -- r3 )	float	f_plus
 1729: r3 = r1+r2;
 1730: 
 1731: f-	( r1 r2 -- r3 )	float	f_minus
 1732: r3 = r1-r2;
 1733: 
 1734: f*	( r1 r2 -- r3 )	float	f_star
 1735: r3 = r1*r2;
 1736: 
 1737: f/	( r1 r2 -- r3 )	float	f_slash
 1738: r3 = r1/r2;
 1739: 
 1740: f**	( r1 r2 -- r3 )	float-ext	f_star_star
 1741: ""@i{r3} is @i{r1} raised to the @i{r2}th power.""
 1742: r3 = pow(r1,r2);
 1743: 
 1744: fnegate	( r1 -- r2 )	float	f_negate
 1745: r2 = - r1;
 1746: 
 1747: fdrop	( r -- )		float	f_drop
 1748: 
 1749: fdup	( r -- r r )	float	f_dupe
 1750: 
 1751: fswap	( r1 r2 -- r2 r1 )	float	f_swap
 1752: 
 1753: fover	( r1 r2 -- r1 r2 r1 )	float	f_over
 1754: 
 1755: frot	( r1 r2 r3 -- r2 r3 r1 )	float	f_rote
 1756: 
 1757: fnip	( r1 r2 -- r2 )	gforth	f_nip
 1758: 
 1759: ftuck	( r1 r2 -- r2 r1 r2 )	gforth	f_tuck
 1760: 
 1761: float+	( f_addr1 -- f_addr2 )	float	float_plus
 1762: ""@code{1 floats +}.""
 1763: f_addr2 = f_addr1+1;
 1764: 
 1765: floats	( n1 -- n2 )	float
 1766: ""@i{n2} is the number of address units of @i{n1} floats.""
 1767: n2 = n1*sizeof(Float);
 1768: 
 1769: floor	( r1 -- r2 )	float
 1770: ""Round towards the next smaller integral value, i.e., round toward negative infinity.""
 1771: /* !! unclear wording */
 1772: r2 = floor(r1);
 1773: 
 1774: fround	( r1 -- r2 )	float	f_round
 1775: ""Round to the nearest integral value.""
 1776: /* !! unclear wording */
 1777: #ifdef HAVE_RINT
 1778: r2 = rint(r1);
 1779: #else
 1780: r2 = floor(r1+0.5);
 1781: /* !! This is not quite true to the rounding rules given in the standard */
 1782: #endif
 1783: 
 1784: fmax	( r1 r2 -- r3 )	float	f_max
 1785: if (r1<r2)
 1786:   r3 = r2;
 1787: else
 1788:   r3 = r1;
 1789: 
 1790: fmin	( r1 r2 -- r3 )	float	f_min
 1791: if (r1<r2)
 1792:   r3 = r1;
 1793: else
 1794:   r3 = r2;
 1795: 
 1796: represent	( r c_addr u -- n f1 f2 )	float
 1797: char *sig;
 1798: int flag;
 1799: int decpt;
 1800: sig=ecvt(r, u, &decpt, &flag);
 1801: n=(r==0 ? 1 : decpt);
 1802: f1=FLAG(flag!=0);
 1803: f2=FLAG(isdigit((unsigned)(sig[0]))!=0);
 1804: memmove(c_addr,sig,u);
 1805: 
 1806: >float	( c_addr u -- flag )	float	to_float
 1807: ""Actual stack effect: ( c_addr u -- r t | f ).  Attempt to convert the
 1808: character string @i{c-addr u} to internal floating-point
 1809: representation. If the string represents a valid floating-point number
 1810: @i{r} is placed on the floating-point stack and @i{flag} is
 1811: true. Otherwise, @i{flag} is false. A string of blanks is a special
 1812: case and represents the floating-point number 0.""
 1813: /* real signature: c_addr u -- r t / f */
 1814: Float r;
 1815: char *number=cstr(c_addr, u, 1);
 1816: char *endconv;
 1817: int sign = 0;
 1818: if(number[0]=='-') {
 1819:    sign = 1;
 1820:    number++;
 1821:    u--;
 1822: }
 1823: while(isspace((unsigned)(number[--u])) && u>0);
 1824: switch(number[u])
 1825: {
 1826:    case 'd':
 1827:    case 'D':
 1828:    case 'e':
 1829:    case 'E':  break;
 1830:    default :  u++; break;
 1831: }
 1832: number[u]='\0';
 1833: r=strtod(number,&endconv);
 1834: if((flag=FLAG(!(Cell)*endconv)))
 1835: {
 1836:    IF_fpTOS(fp[0] = fpTOS);
 1837:    fp += -1;
 1838:    fpTOS = sign ? -r : r;
 1839: }
 1840: else if(*endconv=='d' || *endconv=='D')
 1841: {
 1842:    *endconv='E';
 1843:    r=strtod(number,&endconv);
 1844:    if((flag=FLAG(!(Cell)*endconv)))
 1845:      {
 1846: 	IF_fpTOS(fp[0] = fpTOS);
 1847: 	fp += -1;
 1848: 	fpTOS = sign ? -r : r;
 1849:      }
 1850: }
 1851: 
 1852: fabs	( r1 -- r2 )	float-ext	f_abs
 1853: r2 = fabs(r1);
 1854: 
 1855: facos	( r1 -- r2 )	float-ext	f_a_cos
 1856: r2 = acos(r1);
 1857: 
 1858: fasin	( r1 -- r2 )	float-ext	f_a_sine
 1859: r2 = asin(r1);
 1860: 
 1861: fatan	( r1 -- r2 )	float-ext	f_a_tan
 1862: r2 = atan(r1);
 1863: 
 1864: fatan2	( r1 r2 -- r3 )	float-ext	f_a_tan_two
 1865: ""@i{r1/r2}=tan(@i{r3}). ANS Forth does not require, but probably
 1866: intends this to be the inverse of @code{fsincos}. In gforth it is.""
 1867: r3 = atan2(r1,r2);
 1868: 
 1869: fcos	( r1 -- r2 )	float-ext	f_cos
 1870: r2 = cos(r1);
 1871: 
 1872: fexp	( r1 -- r2 )	float-ext	f_e_x_p
 1873: r2 = exp(r1);
 1874: 
 1875: fexpm1	( r1 -- r2 )	float-ext	f_e_x_p_m_one
 1876: ""@i{r2}=@i{e}**@i{r1}@minus{}1""
 1877: #ifdef HAVE_EXPM1
 1878: extern double
 1879: #ifdef NeXT
 1880:               const
 1881: #endif
 1882:                     expm1(double);
 1883: r2 = expm1(r1);
 1884: #else
 1885: r2 = exp(r1)-1.;
 1886: #endif
 1887: 
 1888: fln	( r1 -- r2 )	float-ext	f_l_n
 1889: r2 = log(r1);
 1890: 
 1891: flnp1	( r1 -- r2 )	float-ext	f_l_n_p_one
 1892: ""@i{r2}=ln(@i{r1}+1)""
 1893: #ifdef HAVE_LOG1P
 1894: extern double
 1895: #ifdef NeXT
 1896:               const
 1897: #endif
 1898:                     log1p(double);
 1899: r2 = log1p(r1);
 1900: #else
 1901: r2 = log(r1+1.);
 1902: #endif
 1903: 
 1904: flog	( r1 -- r2 )	float-ext	f_log
 1905: ""The decimal logarithm.""
 1906: r2 = log10(r1);
 1907: 
 1908: falog	( r1 -- r2 )	float-ext	f_a_log
 1909: ""@i{r2}=10**@i{r1}""
 1910: extern double pow10(double);
 1911: r2 = pow10(r1);
 1912: 
 1913: fsin	( r1 -- r2 )	float-ext	f_sine
 1914: r2 = sin(r1);
 1915: 
 1916: fsincos	( r1 -- r2 r3 )	float-ext	f_sine_cos
 1917: ""@i{r2}=sin(@i{r1}), @i{r3}=cos(@i{r1})""
 1918: r2 = sin(r1);
 1919: r3 = cos(r1);
 1920: 
 1921: fsqrt	( r1 -- r2 )	float-ext	f_square_root
 1922: r2 = sqrt(r1);
 1923: 
 1924: ftan	( r1 -- r2 )	float-ext	f_tan
 1925: r2 = tan(r1);
 1926: :
 1927:  fsincos f/ ;
 1928: 
 1929: fsinh	( r1 -- r2 )	float-ext	f_cinch
 1930: r2 = sinh(r1);
 1931: :
 1932:  fexpm1 fdup fdup 1. d>f f+ f/ f+ f2/ ;
 1933: 
 1934: fcosh	( r1 -- r2 )	float-ext	f_cosh
 1935: r2 = cosh(r1);
 1936: :
 1937:  fexp fdup 1/f f+ f2/ ;
 1938: 
 1939: ftanh	( r1 -- r2 )	float-ext	f_tan_h
 1940: r2 = tanh(r1);
 1941: :
 1942:  f2* fexpm1 fdup 2. d>f f+ f/ ;
 1943: 
 1944: fasinh	( r1 -- r2 )	float-ext	f_a_cinch
 1945: r2 = asinh(r1);
 1946: :
 1947:  fdup fdup f* 1. d>f f+ fsqrt f/ fatanh ;
 1948: 
 1949: facosh	( r1 -- r2 )	float-ext	f_a_cosh
 1950: r2 = acosh(r1);
 1951: :
 1952:  fdup fdup f* 1. d>f f- fsqrt f+ fln ;
 1953: 
 1954: fatanh	( r1 -- r2 )	float-ext	f_a_tan_h
 1955: r2 = atanh(r1);
 1956: :
 1957:  fdup f0< >r fabs 1. d>f fover f- f/  f2* flnp1 f2/
 1958:  r> IF  fnegate  THEN ;
 1959: 
 1960: sfloats	( n1 -- n2 )	float-ext	s_floats
 1961: ""@i{n2} is the number of address units of @i{n1}
 1962: single-precision IEEE floating-point numbers.""
 1963: n2 = n1*sizeof(SFloat);
 1964: 
 1965: dfloats	( n1 -- n2 )	float-ext	d_floats
 1966: ""@i{n2} is the number of address units of @i{n1}
 1967: double-precision IEEE floating-point numbers.""
 1968: n2 = n1*sizeof(DFloat);
 1969: 
 1970: sfaligned	( c_addr -- sf_addr )	float-ext	s_f_aligned
 1971: ""@i{sf-addr} is the first single-float-aligned address greater
 1972: than or equal to @i{c-addr}.""
 1973: sf_addr = (SFloat *)((((Cell)c_addr)+(sizeof(SFloat)-1))&(-sizeof(SFloat)));
 1974: :
 1975:  [ 1 sfloats 1- ] Literal + [ -1 sfloats ] Literal and ;
 1976: 
 1977: dfaligned	( c_addr -- df_addr )	float-ext	d_f_aligned
 1978: ""@i{df-addr} is the first double-float-aligned address greater
 1979: than or equal to @i{c-addr}.""
 1980: df_addr = (DFloat *)((((Cell)c_addr)+(sizeof(DFloat)-1))&(-sizeof(DFloat)));
 1981: :
 1982:  [ 1 dfloats 1- ] Literal + [ -1 dfloats ] Literal and ;
 1983: 
 1984: \ The following words access machine/OS/installation-dependent
 1985: \   Gforth internals
 1986: \ !! how about environmental queries DIRECT-THREADED,
 1987: \   INDIRECT-THREADED, TOS-CACHED, FTOS-CACHED, CODEFIELD-DOES */
 1988: 
 1989: \ local variable implementation primitives
 1990: \+
 1991: \+glocals
 1992: 
 1993: @local#	( #noffset -- w )	gforth	fetch_local_number
 1994: w = *(Cell *)(lp+noffset);
 1995: 
 1996: @local0	( -- w )	new	fetch_local_zero
 1997: w = *(Cell *)(lp+0*sizeof(Cell));
 1998: 
 1999: @local1	( -- w )	new	fetch_local_four
 2000: w = *(Cell *)(lp+1*sizeof(Cell));
 2001: 
 2002: @local2	( -- w )	new	fetch_local_eight
 2003: w = *(Cell *)(lp+2*sizeof(Cell));
 2004: 
 2005: @local3	( -- w )	new	fetch_local_twelve
 2006: w = *(Cell *)(lp+3*sizeof(Cell));
 2007: 
 2008: \+floating
 2009: 
 2010: f@local#	( #noffset -- r )	gforth	f_fetch_local_number
 2011: r = *(Float *)(lp+noffset);
 2012: 
 2013: f@local0	( -- r )	new	f_fetch_local_zero
 2014: r = *(Float *)(lp+0*sizeof(Float));
 2015: 
 2016: f@local1	( -- r )	new	f_fetch_local_eight
 2017: r = *(Float *)(lp+1*sizeof(Float));
 2018: 
 2019: \+
 2020: 
 2021: laddr#	( #noffset -- c_addr )	gforth	laddr_number
 2022: /* this can also be used to implement lp@ */
 2023: c_addr = (Char *)(lp+noffset);
 2024: 
 2025: lp+!#	( #noffset -- )	gforth	lp_plus_store_number
 2026: ""used with negative immediate values it allocates memory on the
 2027: local stack, a positive immediate argument drops memory from the local
 2028: stack""
 2029: lp += noffset;
 2030: 
 2031: lp-	( -- )	new	minus_four_lp_plus_store
 2032: lp += -sizeof(Cell);
 2033: 
 2034: lp+	( -- )	new	eight_lp_plus_store
 2035: lp += sizeof(Float);
 2036: 
 2037: lp+2	( -- )	new	sixteen_lp_plus_store
 2038: lp += 2*sizeof(Float);
 2039: 
 2040: lp!	( c_addr -- )	gforth	lp_store
 2041: lp = (Address)c_addr;
 2042: 
 2043: >l	( w -- )	gforth	to_l
 2044: lp -= sizeof(Cell);
 2045: *(Cell *)lp = w;
 2046: 
 2047: \+floating
 2048: 
 2049: f>l	( r -- )	gforth	f_to_l
 2050: lp -= sizeof(Float);
 2051: *(Float *)lp = r;
 2052: 
 2053: fpick	( u -- r )		gforth
 2054: ""Actually the stack effect is @code{ r0 ... ru u -- r0 ... ru r0 }.""
 2055: r = fp[u+1]; /* +1, because update of fp happens before this fragment */
 2056: :
 2057:  floats fp@ + f@ ;
 2058: 
 2059: \+
 2060: \+
 2061: 
 2062: \+OS
 2063: 
 2064: define(`uploop',
 2065:        `pushdef(`$1', `$2')_uploop(`$1', `$2', `$3', `$4', `$5')`'popdef(`$1')')
 2066: define(`_uploop',
 2067:        `ifelse($1, `$3', `$5',
 2068: 	       `$4`'define(`$1', incr($1))_uploop(`$1', `$2', `$3', `$4', `$5')')')
 2069: \ argflist(argnum): Forth argument list
 2070: define(argflist,
 2071:        `ifelse($1, 0, `',
 2072:                `uploop(`_i', 1, $1, `format(`u%d ', _i)', `format(`u%d ', _i)')')')
 2073: \ argdlist(argnum): declare C's arguments
 2074: define(argdlist,
 2075:        `ifelse($1, 0, `',
 2076:                `uploop(`_i', 1, $1, `Cell, ', `Cell')')')
 2077: \ argclist(argnum): pass C's arguments
 2078: define(argclist,
 2079:        `ifelse($1, 0, `',
 2080:                `uploop(`_i', 1, $1, `format(`u%d, ', _i)', `format(`u%d', _i)')')')
 2081: \ icall(argnum)
 2082: define(icall,
 2083: `icall$1	( argflist($1)u -- uret )	gforth
 2084: uret = (SYSCALL(Cell(*)(argdlist($1)))u)(argclist($1));
 2085: 
 2086: ')
 2087: define(fcall,
 2088: `fcall$1	( argflist($1)u -- rret )	gforth
 2089: rret = (SYSCALL(Float(*)(argdlist($1)))u)(argclist($1));
 2090: 
 2091: ')
 2092: 
 2093: \ close ' to keep fontify happy
 2094: 
 2095: open-lib	( c_addr1 u1 -- u2 )	gforth	open_lib
 2096: #if defined(HAVE_LIBDL) || defined(HAVE_DLOPEN)
 2097: #ifndef RTLD_GLOBAL
 2098: #define RTLD_GLOBAL 0
 2099: #endif
 2100: u2=(UCell) dlopen(cstr(c_addr1, u1, 1), RTLD_GLOBAL | RTLD_LAZY);
 2101: #else
 2102: #  ifdef _WIN32
 2103: u2 = (Cell) GetModuleHandle(cstr(c_addr1, u1, 1));
 2104: #  else
 2105: #warning Define open-lib!
 2106: u2 = 0;
 2107: #  endif
 2108: #endif
 2109: 
 2110: lib-sym	( c_addr1 u1 u2 -- u3 )	gforth	lib_sym
 2111: #if defined(HAVE_LIBDL) || defined(HAVE_DLOPEN)
 2112: u3 = (UCell) dlsym((void*)u2,cstr(c_addr1, u1, 1));
 2113: #else
 2114: #  ifdef _WIN32
 2115: u3 = (Cell) GetProcAddress((HMODULE)u2, cstr(c_addr1, u1, 1));
 2116: #  else
 2117: #warning Define lib-sym!
 2118: u3 = 0;
 2119: #  endif
 2120: #endif
 2121: 
 2122: uploop(i, 0, 7, `icall(i)')
 2123: icall(20)
 2124: uploop(i, 0, 7, `fcall(i)')
 2125: fcall(20)
 2126: 
 2127: \+
 2128: 
 2129: up!	( a_addr -- )	gforth	up_store
 2130: UP=up=(char *)a_addr;
 2131: :
 2132:  up ! ;
 2133: Variable UP
 2134: 
 2135: wcall	( u -- )	gforth
 2136: IF_fpTOS(fp[0]=fpTOS);
 2137: FP=fp;
 2138: sp=(SYSCALL(Cell(*)(Cell *, void *))u)(sp, &FP);
 2139: fp=FP;
 2140: IF_spTOS(spTOS=sp[0];)
 2141: IF_fpTOS(fpTOS=fp[0]);
 2142: 
 2143: \+file
 2144: 
 2145: open-dir	( c_addr u -- wdirid wior )	gforth	open_dir
 2146: wdirid = (Cell)opendir(tilde_cstr(c_addr, u, 1));
 2147: wior =  IOR(wdirid == 0);
 2148: 
 2149: read-dir	( c_addr u1 wdirid -- u2 flag wior )	gforth	read_dir
 2150: struct dirent * dent;
 2151: dent = readdir((DIR *)wdirid);
 2152: wior = 0;
 2153: flag = -1;
 2154: if(dent == NULL) {
 2155:   u2 = 0;
 2156:   flag = 0;
 2157: } else {
 2158:   u2 = strlen(dent->d_name);
 2159:   if(u2 > u1) {
 2160:     u2 = u1;
 2161:     wior = -512-ENAMETOOLONG;
 2162:   }
 2163:   memmove(c_addr, dent->d_name, u2);
 2164: }
 2165: 
 2166: close-dir	( wdirid -- wior )	gforth	close_dir
 2167: wior = IOR(closedir((DIR *)wdirid));
 2168: 
 2169: filename-match	( c_addr1 u1 c_addr2 u2 -- flag )	gforth	match_file
 2170: char * string = cstr(c_addr1, u1, 1);
 2171: char * pattern = cstr(c_addr2, u2, 0);
 2172: flag = FLAG(!fnmatch(pattern, string, 0));
 2173: 
 2174: \+
 2175: 
 2176: newline	( -- c_addr u )	gforth
 2177: ""String containing the newline sequence of the host OS""
 2178: char newline[] = {
 2179: #if defined(unix) || defined(__MACH__)
 2180: /* Darwin/MacOS X sets __MACH__, but not unix. */
 2181: '\n'
 2182: #else
 2183: '\r','\n'
 2184: #endif
 2185: };
 2186: c_addr=newline;
 2187: u=sizeof(newline);
 2188: :
 2189:  "newline count ;
 2190: Create "newline e? crlf [IF] 2 c, $0D c, [ELSE] 1 c, [THEN] $0A c,
 2191: 
 2192: \+os
 2193: 
 2194: utime	( -- dtime )	gforth
 2195: ""Report the current time in microseconds since some epoch.""
 2196: struct timeval time1;
 2197: gettimeofday(&time1,NULL);
 2198: dtime = timeval2us(&time1);
 2199: 
 2200: cputime ( -- duser dsystem ) gforth
 2201: ""duser and dsystem are the respective user- and system-level CPU
 2202: times used since the start of the Forth system (excluding child
 2203: processes), in microseconds (the granularity may be much larger,
 2204: however).  On platforms without the getrusage call, it reports elapsed
 2205: time (since some epoch) for duser and 0 for dsystem.""
 2206: #ifdef HAVE_GETRUSAGE
 2207: struct rusage usage;
 2208: getrusage(RUSAGE_SELF, &usage);
 2209: duser = timeval2us(&usage.ru_utime);
 2210: dsystem = timeval2us(&usage.ru_stime);
 2211: #else
 2212: struct timeval time1;
 2213: gettimeofday(&time1,NULL);
 2214: duser = timeval2us(&time1);
 2215: #ifndef BUGGY_LONG_LONG
 2216: dsystem = (DCell)0;
 2217: #else
 2218: dsystem=(DCell){0,0};
 2219: #endif
 2220: #endif
 2221: 
 2222: \+
 2223: 
 2224: \+floating
 2225: 
 2226: v*	( f_addr1 nstride1 f_addr2 nstride2 ucount -- r ) gforth v_star
 2227: ""dot-product: r=v1*v2.  The first element of v1 is at f_addr1, the
 2228: next at f_addr1+nstride1 and so on (similar for v2). Both vectors have
 2229: ucount elements.""
 2230: for (r=0.; ucount>0; ucount--) {
 2231:   r += *f_addr1 * *f_addr2;
 2232:   f_addr1 = (Float *)(((Address)f_addr1)+nstride1);
 2233:   f_addr2 = (Float *)(((Address)f_addr2)+nstride2);
 2234: }
 2235: :
 2236:  >r swap 2swap swap 0e r> 0 ?DO
 2237:      dup f@ over + 2swap dup f@ f* f+ over + 2swap
 2238:  LOOP 2drop 2drop ; 
 2239: 
 2240: faxpy	( ra f_x nstridex f_y nstridey ucount -- )	gforth
 2241: ""vy=ra*vx+vy""
 2242: for (; ucount>0; ucount--) {
 2243:   *f_y += ra * *f_x;
 2244:   f_x = (Float *)(((Address)f_x)+nstridex);
 2245:   f_y = (Float *)(((Address)f_y)+nstridey);
 2246: }
 2247: :
 2248:  >r swap 2swap swap r> 0 ?DO
 2249:      fdup dup f@ f* over + 2swap dup f@ f+ dup f! over + 2swap
 2250:  LOOP 2drop 2drop fdrop ;
 2251: 
 2252: \+
 2253: 
 2254: \+file
 2255: 
 2256: (read-line)	( c_addr u1 wfileid -- u2 flag u3 wior )	file	paren_read_line
 2257: Cell c;
 2258: flag=-1;
 2259: u3=0;
 2260: for(u2=0; u2<u1; u2++)
 2261: {
 2262:    c = getc((FILE *)wfileid);
 2263:    u3++;
 2264:    if (c=='\n') break;
 2265:    if (c=='\r') {
 2266:      if ((c = getc((FILE *)wfileid))!='\n')
 2267:        ungetc(c,(FILE *)wfileid);
 2268:      else
 2269:        u3++;
 2270:      break;
 2271:    }
 2272:    if (c==EOF) {
 2273: 	flag=FLAG(u2!=0);
 2274: 	break;
 2275:      }
 2276:    c_addr[u2] = (Char)c;
 2277: }
 2278: wior=FILEIO(ferror((FILE *)wfileid));
 2279: 
 2280: \+
 2281: 
 2282: (listlfind)	( c_addr u longname1 -- longname2 )	new	paren_listlfind
 2283: for (; longname1 != NULL; longname1 = (struct Longname *)(longname1->next))
 2284:   if ((UCell)LONGNAME_COUNT(longname1)==u &&
 2285:       memcasecmp(c_addr, longname1->name, u)== 0 /* or inline? */)
 2286:     break;
 2287: longname2=longname1;
 2288: :
 2289:     BEGIN  dup WHILE  (findl-samelen)  dup  WHILE
 2290: 	>r 2dup r@ cell+ cell+ capscomp  0=
 2291: 	IF  2drop r>  EXIT  THEN
 2292: 	r> @
 2293:     REPEAT  THEN  nip nip ;
 2294: : (findl-samelen) ( u longname1 -- u longname2/0 )
 2295:     BEGIN  2dup cell+ @ lcount-mask and <> WHILE  @  dup 0= UNTIL  THEN ;
 2296: 
 2297: \+hash
 2298: 
 2299: (hashlfind)	( c_addr u a_addr -- longname2 )	new	paren_hashlfind
 2300: struct Longname *longname1;
 2301: longname2=NULL;
 2302: while(a_addr != NULL)
 2303: {
 2304:    longname1=(struct Longname *)(a_addr[1]);
 2305:    a_addr=(Cell *)(a_addr[0]);
 2306:    if ((UCell)LONGNAME_COUNT(longname1)==u &&
 2307:        memcasecmp(c_addr, longname1->name, u)== 0 /* or inline? */)
 2308:      {
 2309: 	longname2=longname1;
 2310: 	break;
 2311:      }
 2312: }
 2313: :
 2314:  BEGIN  dup  WHILE
 2315:         2@ >r >r dup r@ cell+ @ lcount-mask and =
 2316:         IF  2dup r@ cell+ cell+ capscomp 0=
 2317: 	    IF  2drop r> rdrop  EXIT  THEN  THEN
 2318: 	rdrop r>
 2319:  REPEAT nip nip ;
 2320: 
 2321: (tablelfind)	( c_addr u a_addr -- longname2 )	new	paren_tablelfind
 2322: ""A case-sensitive variant of @code{(hashfind)}""
 2323: struct Longname *longname1;
 2324: longname2=NULL;
 2325: while(a_addr != NULL)
 2326: {
 2327:    longname1=(struct Longname *)(a_addr[1]);
 2328:    a_addr=(Cell *)(a_addr[0]);
 2329:    if ((UCell)LONGNAME_COUNT(longname1)==u &&
 2330:        memcmp(c_addr, longname1->name, u)== 0 /* or inline? */)
 2331:      {
 2332: 	longname2=longname1;
 2333: 	break;
 2334:      }
 2335: }
 2336: :
 2337:  BEGIN  dup  WHILE
 2338:         2@ >r >r dup r@ cell+ @ lcount-mask and =
 2339:         IF  2dup r@ cell+ cell+ -text 0=
 2340: 	    IF  2drop r> rdrop  EXIT  THEN  THEN
 2341: 	rdrop r>
 2342:  REPEAT nip nip ;
 2343: 
 2344: \+
 2345: 
 2346: primtable	( -- wprimtable )	new
 2347: ""wprimtable is a table containing the xts of the primitives indexed
 2348: by sequence-number in prim (for use in prepare-peephole-table).""
 2349: wprimtable = (Cell)primtable(symbols+DOESJUMP+1,MAX_SYMBOLS);
 2350: 
 2351: prepare-peephole-table	( wprimtable -- wpeeptable ) new prepare_peephole_opt
 2352: ""wpeeptable is a data structure used by @code{peephole-opt}; it is
 2353: constructed by combining a primitives table with a simple peephole
 2354: optimization table.""
 2355: wpeeptable = prepare_peephole_table((Xt *)wprimtable);
 2356: 
 2357: peephole-opt	( xt1 xt2 wpeeptable -- xt )	new	peephole_opt
 2358: ""xt is the combination of xt1 and xt2 (according to wpeeptable); if
 2359: they cannot be combined, xt is 0.""
 2360: xt = peephole_opt(xt1, xt2, wpeeptable);
 2361: 
 2362: lit_plus = lit +

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