1/*  This is the parser for the dlg
2 *  This is a part of the Purdue Compiler Construction Tool Set
3 *
4 * SOFTWARE RIGHTS
5 *
6 * We reserve no LEGAL rights to the Purdue Compiler Construction Tool
7 * Set (PCCTS) -- PCCTS is in the public domain.  An individual or
8 * company may do whatever they wish with source code distributed with
9 * PCCTS or the code generated by PCCTS, including the incorporation of
10 * PCCTS, or its output, into commerical software.
11 *
12 * We encourage users to develop software with PCCTS.  However, we do ask
13 * that credit is given to us for developing PCCTS.  By "credit",
14 * we mean that if you incorporate our source code into one of your
15 * programs (commercial product, research project, or otherwise) that you
16 * acknowledge this fact somewhere in the documentation, research report,
17 * etc...  If you like PCCTS and have developed a nice tool with the
18 * output, please mention that you developed it using PCCTS.  In
19 * addition, we ask that this header remain intact in our source code.
20 * As long as these guidelines are kept, we expect to continue enhancing
21 * this system and expect to make other tools available as they are
22 * completed.
23 *
24 * DLG 1.33
25 * Will Cohen
26 * With mods by Terence Parr; AHPCRC, University of Minnesota
27 * 1989-1995
28 */
29
30#header	<<
31#include <ctype.h>
32#include "dlg.h"
33>>
34
35<<
36
37/* MR20 G. Hobbelt
38   Fix for Borland C++ 4.x & 5.x compiling with ALL warnings enabled
39*/
40
41#ifdef __TURBOC__
42#pragma warn -aus  /* unused assignment of 'xxx' */
43#endif
44
45int	action_no = 0;	   /* keep track of actions outputed */
46int	nfa_allocated = 0; /* keeps track of number of nfa nodes */
47nfa_node **nfa_array = NULL;/* root of binary tree that stores nfa array */
48nfa_node nfa_model_node;   /* model to initialize new nodes */
49set	used_chars;	   /* used to label trans. arcs */
50set	used_classes;	   /* classes or chars used to label trans. arcs */
51set	normal_chars;	   /* mask to get rid elements that aren't used
52			      in set */
53int	flag_paren = FALSE;
54int	flag_brace = FALSE;
55int	mode_counter = 0;  /* keep track of number of %%names */
56
57>>
58
59#lexaction <<
60int	func_action;		/* should actions be turned into functions?*/
61int	lex_mode_counter = 0;	/* keeps track of the number of %%names */
62/* MR1									    */
63/* MR1  11-Apr-97	Provide mechanism for inserting code into DLG class */
64/* MR1				via <<%%lexmember...>>			    */
65/* MR1									    */
66int	lexMember = 0;		/* <<%%lexmemeber ...>>	   		MR1 */
67int	lexAction = 0;		/* <<%%lexaction ...>>			MR1 */
68int	parserClass = 0;	/* <<%%parserclass ...>>        MR1 */
69int	lexPrefix = 0;		/* <<%%lexprefix ...>>			MR1 */
70char	theClassName[100];						     /* MR11 */
71char	*pClassName=theClassName;					 /* MR11 */
72int	firstLexMember=1;					             /* MR1 */
73
74#ifdef __USE_PROTOS
75void  xxputc(int c) {						/* MR1 */
76#else
77void xxputc(c)							/* MR1 */
78  int	c;							/* MR1 */
79{								/* MR1 */
80#endif
81  if (parserClass) {						/* MR1 */
82    *pClassName++=c;						/* MR1 */
83    *pClassName=0;						/* MR1 */
84  } else if (lexMember || lexPrefix) {				/* MR1 */
85    if (class_stream != NULL) fputc(c,class_stream);		/* MR1 */
86  } else {							/* MR1 */
87    fputc(c,OUT);						/* MR1 */
88  };								/* MR1 */
89}  								/* MR1 */
90
91#ifdef __USE_PROTOS
92void xxprintf(char *format,char *string) {			/* MR1 */
93#else
94void xxprintf(format,string) 					/* MR1 */
95  char *format;							/* MR1 */
96  char *string;							/* MR1 */
97{								/* MR1 */
98#endif
99  if (lexMember || lexPrefix || parserClass) {			/* MR1 */
100    if (class_stream != NULL)					/* MR1 */
101	 fprintf(class_stream,format,string);			/* MR1 */
102  } else {							/* MR1 */
103    fprintf(OUT,format,string);					/* MR1 */
104  };								/* MR1 */
105}  								/* MR1 */
106>>
107
108#token "[\r\t\ ]+"	<< zzskip(); >>						/* Ignore white */
109#token "\n"			<< zzline++; zzskip(); DAWDLE; >>	/* Track Line # */
110#token L_EOF		"\@"
111#token PER_PER		"\%\%"
112#token NAME_PER_PER	"\%\%[a-zA-Z_][a-zA-Z0-9_]*"
113		<< p_mode_def(&zzlextext[2],lex_mode_counter++); >>
114
115#token LEXMEMBER	"\<\<\%\%lexmember"			/* MR1 */
116		<<lexMember=1;					/* MR1 */
117	          if (firstLexMember != 0) {			/* MR1 */
118	            firstLexMember=0;				/* MR1 */
119	            p_class_def1();				/* MR1 */
120		  };						/* MR1 */
121	          zzmode(ACT);					/* MR1 */
122                >>						/* MR1 */
123#token LEXACTION	"\<\<\%\%lexaction"			/* MR1 */
124		<<lexAction=1;zzmode(ACT);>>			/* MR1 */
125#token PARSERCLASS	"\<\<\%\%parserclass"			/* MR1 */
126		<<parserClass=1;				/* MR1 */
127		  zzmode(ACT);					/* MR1 */
128		>>						/* MR1 */
129#token LEXPREFIX	"\<\<\%\%lexprefix"			/* MR1 */
130		<<lexPrefix=1;zzmode(ACT);>>			/* MR1 */
131
132#token ACTION		"\<\<"
133		<< if (func_action)
134			fprintf(OUT,"\n%s %sact%d()\n{ ",
135					gen_cpp?"ANTLRTokenType":"static void",
136					gen_cpp?ClassName("::"):"", ++action_no);
137		   zzmode(ACT); zzskip();
138		>>
139#token GREAT_GREAT	"\>\>"
140#token L_BRACE		"\{"
141#token R_BRACE		"\}"
142#token L_PAR		"\("
143#token R_PAR		"\)"
144#token L_BRACK		"\["
145#token R_BRACK		"\]"
146#token ZERO_MORE	"\*"
147#token ONE_MORE		"\+"
148#token OR		"\|"
149#token RANGE		"\-"
150#token NOT		"\~"
151#token OCTAL_VALUE "\\0[0-7]*"
152	<< {int t; sscanf(&zzlextext[1],"%o",&t); zzlextext[0] = t;}>>
153#token HEX_VALUE   "\\0[Xx][0-9a-fA-F]+"
154	<< {int t; sscanf(&zzlextext[3],"%x",&t); zzlextext[0] = t;}>>
155#token DEC_VALUE   "\\[1-9][0-9]*"
156	<< {int t; sscanf(&zzlextext[1],"%d",&t); zzlextext[0] = t;}>>
157#token TAB		"\\t"		<< zzlextext[0] = '\t';>>
158#token NL		"\\n"		<< zzlextext[0] = '\n';>>
159#token CR		"\\r"		<< zzlextext[0] = '\r';>>
160#token BS		"\\b"		<< zzlextext[0] = '\b';>>
161
162/* MR1									*/
163/* MR1 10-Apr-97 MR1	Allow #token regular expressions to cross lines	*/
164/* MR1									*/
165#token CONTINUATION	"\\ \n"		<< zzline++; zzskip();>> /* MR1 */
166
167/* NOTE: this takes ANYTHING after the \ */
168#token LIT		"\\~[tnrb]"	<< zzlextext[0] = zzlextext[1];>>
169
170/* NOTE: this takes ANYTHING that doesn't match the other tokens */
171#token REGCHAR		"~[\\]"
172
173
174grammar		:   << p_head(); p_class_hdr(); func_action = FALSE;>>
175		 ( {LEXACTION | LEXMEMBER | LEXPREFIX | PARSERCLASS } ACTION)* /* MR1 */
176		    <<if ( gen_cpp ) p_includes();>>
177		    start_states
178		    << func_action = FALSE; p_tables(); p_tail(); >>
179		    (ACTION)* "@"
180			<< if (firstLexMember != 0) p_class_def1(); >> 		/* MR1 */
181		;
182
183start_states	: ( PER_PER do_conversion
184		  | NAME_PER_PER do_conversion (NAME_PER_PER do_conversion)*)
185		    PER_PER
186		;
187
188do_conversion	: <<new_automaton_mode(); func_action = TRUE;>>
189			rule_list
190			<<
191				dfa_class_nop[mode_counter] =
192					relabel($1.l,comp_level);
193				if (comp_level)
194					p_shift_table(mode_counter);
195				dfa_basep[mode_counter] = dfa_allocated+1;
196				make_dfa_model_node(dfa_class_nop[mode_counter]);
197				nfa_to_dfa($1.l);
198				++mode_counter;
199		    		func_action = FALSE;
200#ifdef HASH_STAT
201				fprint_hash_stats(stderr);
202#endif
203			>>
204		;
205
206rule_list	: rule <<$$.l=$1.l; $$.r=$1.r;>>
207			(rule
208				<<{nfa_node *t1;
209				   t1 = new_nfa_node();
210				   (t1)->trans[0]=$$.l;
211				   (t1)->trans[1]=$1.l;
212				   /* all accept nodes "dead ends" */
213				   $$.l=t1; $$.r=NULL;
214				   }
215				>>
216			)*
217		| /* empty */
218			<<$$.l = new_nfa_node(); $$.r = NULL;
219			   warning("no regular expressions", zzline);
220			>>
221		;
222
223rule	: reg_expr ACTION
224/* MR23 */		<< if ($1.r != NULL) {
225					$$.l=$1.l; $$.r=$1.r; ($1.r)->accept=action_no;
226				   }
227				>>
228		| ACTION
229			<<$$.l = NULL; $$.r = NULL;
230			  error("no expression for action  ", zzline);
231			>>
232		;
233
234reg_expr	: and_expr <<$$.l=$1.l; $$.r=$1.r;>>
235			(OR and_expr
236				<<{nfa_node *t1, *t2;
237				   t1 = new_nfa_node(); t2 = new_nfa_node();
238				   (t1)->trans[0]=$$.l;
239				   (t1)->trans[1]=$2.l;
240/* MR23 */		   if ($$.r != NULL) ($$.r)->trans[1]=t2;
241                   if ($2.r) {
242    				   ($2.r)->trans[1]=t2;     /* MR20 */
243                   }
244				   $$.l=t1; $$.r=t2;
245				  }
246				>>
247			)*
248		;
249
250and_expr	: repeat_expr
251					<<
252						$$.l=$1.l; $$.r=$1.r;
253				    >>
254			(repeat_expr
255/* MR23 */				<< if ($$.r != NULL) {
256							($$.r)->trans[1]=$1.l;
257							$$.r=$1.r;
258						   }
259						>>
260			)*
261		;
262
263repeat_expr	: expr <<$$.l=$1.l; $$.r=$1.r;>>
264			{ ZERO_MORE
265			<<{	nfa_node *t1,*t2;
266/* MR23 */		if ($$.r != NULL) ($$.r)->trans[0] = $$.l;
267				t1 = new_nfa_node(); t2 = new_nfa_node();
268				t1->trans[0]=$$.l;
269				t1->trans[1]=t2;
270/* MR23 */		if ($$.r != NULL) ($$.r)->trans[1]=t2;
271				$$.l=t1;$$.r=t2;
272			  }
273			>>
274			| ONE_MORE
275/* MR23 */		<<if ($$.r != NULL) ($$.r)->trans[0] = $$.l;>>
276			}
277		| ZERO_MORE
278			<< error("no expression for *", zzline);>>
279		| ONE_MORE
280			<< error("no expression for +", zzline);>>
281		;
282
283expr	: << $$.l = new_nfa_node();
284			 $$.r = new_nfa_node();
285		  >>
286		  L_BRACK atom_list R_BRACK
287			<<
288/* MR23 */		if ($$.l != NULL) {
289					($$.l)->trans[0] = $$.r;
290					($$.l)->label = set_dup($2.label);
291					set_orin(&used_chars,($$.l)->label);
292				}
293			>>
294		| NOT L_BRACK atom_list R_BRACK
295			<<
296/* MR23 */		if ($$.l != NULL) {
297					($$.l)->trans[0] = $$.r;
298					($$.l)->label = set_dif(normal_chars,$3.label);
299					set_orin(&used_chars,($$.l)->label);
300				}
301			>>
302		| L_PAR reg_expr R_PAR
303			<<
304/* MR23 */		if ($$.l != NULL) {
305					($$.l)->trans[0] = $2.l;
306					if ($2.r) {
307    					($2.r)->trans[1] = $$.r;    /* MR20 */
308					}
309				}
310			>>
311		| L_BRACE reg_expr R_BRACE
312			<<
313/* MR23 */		if ($$.l != NULL) {
314					($$.l)->trans[0] = $2.l;
315					($$.l)->trans[1] = $$.r;
316			        if ($2.r) {
317    					($2.r)->trans[1] = $$.r;    /* MR20 */
318					}
319				}
320			>>
321		| atom
322			<<
323/* MR23 */		if ($$.l != NULL) {
324					($$.l)->trans[0] = $$.r;
325					($$.l)->label = set_dup($1.label);
326					set_orin(&used_chars,($$.l)->label);
327				}
328			>>
329		;
330
331atom_list	: << set_free($$.label); >>
332				(near_atom <<set_orin(&($$.label),$1.label);>>)*
333		;
334
335near_atom	: << register int i;
336		     register int i_prime;
337		  >>
338		  anychar
339			<<$$.letter=$1.letter; $$.label=set_of($1.letter);
340			i_prime = $1.letter + MIN_CHAR;
341			if (case_insensitive && islower(i_prime))
342				set_orel(toupper(i_prime)-MIN_CHAR,
343					&($$.label));
344			if (case_insensitive && isupper(i_prime))
345	 			set_orel(tolower(i_prime)-MIN_CHAR,
346					&($$.label));
347			>>
348			{ RANGE anychar
349				<< if (case_insensitive){
350					i_prime = $$.letter+MIN_CHAR;
351					$$.letter = (islower(i_prime) ?
352						toupper(i_prime) : i_prime)-MIN_CHAR;
353					i_prime = $2.letter+MIN_CHAR;
354					$2.letter = (islower(i_prime) ?
355						toupper(i_prime) : i_prime)-MIN_CHAR;
356				   }
357				   /* check to see if range okay */
358					{
359					    int debugLetter1 = $$.letter;
360						int debugLetter2 = $2.letter;
361					}
362				   if ($$.letter > $2.letter
363                                       && $2.letter != 0xff){       /* MR16 */
364					  error("invalid range  ", zzline);
365				   }
366				   for (i=$$.letter; i<= (int)$2.letter; ++i){
367					set_orel(i,&($$.label));
368					i_prime = i+MIN_CHAR;
369					if (case_insensitive && islower(i_prime))
370						set_orel(toupper(i_prime)-MIN_CHAR,
371							&($$.label));
372					if (case_insensitive && isupper(i_prime))
373		 				set_orel(tolower(i_prime)-MIN_CHAR,
374							&($$.label));
375					}
376				>>
377			}
378		;
379
380atom		: << register int i_prime;>>
381		  anychar
382		  <<$$.label = set_of($1.letter);
383		    i_prime = $1.letter + MIN_CHAR;
384		    if (case_insensitive && islower(i_prime))
385			set_orel(toupper(i_prime)-MIN_CHAR,
386				&($$.label));
387		    if (case_insensitive && isupper(i_prime))
388	 		set_orel(tolower(i_prime)-MIN_CHAR,
389				&($$.label));
390		  >>
391		;
392
393anychar		: REGCHAR	<<$$.letter = $1.letter - MIN_CHAR;>>
394		| OCTAL_VALUE	<<$$.letter = $1.letter - MIN_CHAR;>>
395		| HEX_VALUE	<<$$.letter = $1.letter - MIN_CHAR;>>
396		| DEC_VALUE	<<$$.letter = $1.letter - MIN_CHAR;>>
397		| TAB		<<$$.letter = $1.letter - MIN_CHAR;>>
398		| NL		<<$$.letter = $1.letter - MIN_CHAR;>>
399		| CR		<<$$.letter = $1.letter - MIN_CHAR;>>
400		| BS		<<$$.letter = $1.letter - MIN_CHAR;>>
401		| LIT		<<$$.letter = $1.letter - MIN_CHAR;>>
402		/* NOTE: LEX_EOF is ALWAYS shifted to 0 = MIN_CHAR - MIN_CHAR*/
403		| L_EOF		<<$$.letter = 0;>>
404		;
405
406<</* empty action */>>
407
408#lexclass ACT
409#token "@"	<< error("unterminated action", zzline); zzmode(START); >>
410#token ACTION "\>\>"
411		<< if (func_action) fprintf(OUT,"}\n\n");
412		   zzmode(START);
413/* MR1									    */
414/* MR1  11-Apr-97	Provide mechanism for inserting code into DLG class */
415/* MR1				via <<%%lexmember ...>>			    */
416/* MR1			This is a consequence of not saving actions         */
417/* MR1									    */
418/* MR1 */	   parserClass=0;
419/* MR1 */	   lexPrefix=0;
420/* MR1 */	   lexAction=0;
421/* MR1 */	   lexMember=0;
422		>>
423#token "\>"		<< xxputc(zzlextext[0]); zzskip(); >>		/* MR1 */
424#token "\\\>"		<< xxputc('>'); zzskip(); >>			/* MR1 */
425#token "\\"		<< xxputc('\\'); zzskip(); >>			/* MR1 */
426#token "\n"		<< xxputc(zzlextext[0]); ++zzline; zzskip(); >>	/* MR1 */
427#token "/\*"		<< zzmode(ACTION_COMMENTS);			/* MR1 */
428			   xxprintf("%s", &(zzlextext[0])); zzskip();	/* MR1 */
429			>>						/* MR1 */
430#token "//"		<< zzmode(ACTION_CPP_COMMENTS);			/* MR1 */
431			   xxprintf("%s", &(zzlextext[0])); zzskip();	/* MR1 */
432			>>						/* MR1 */
433#token "~[]"		<< xxputc(zzlextext[0]); zzskip(); >>		/* MR1 */
434									/* MR1 */
435#lexclass ACTION_COMMENTS						/* MR1 */
436#token "\*/"		<< zzmode(ACT);					/* MR1 */
437			   xxprintf("%s", &(zzlextext[0])); zzskip();	/* MR1 */
438			>>						/* MR1 */
439#token "[\n\r]"		<< zzline++; xxputc(zzlextext[0]); zzskip();>>	/* MR1 */
440#token "~[]"		<< xxputc(zzlextext[0]); zzskip();>>		/* MR1 */
441									/* MR1 */
442#lexclass ACTION_CPP_COMMENTS						/* MR1 */
443#token "[\n\r]"		<< zzmode(ACT); zzline++;			/* MR1 */
444			   xxprintf("%s", &(zzlextext[0])); zzskip();	/* MR1 */
445			>>						/* MR1 */
446#token "~[]"		<< xxputc(zzlextext[0]); zzskip();>>		/* MR1 */
447
448<<
449/* adds a new nfa to the binary tree and returns a pointer to it */
450nfa_node *
451#ifdef __USE_PROTOS
452new_nfa_node(void)
453#else
454new_nfa_node()
455#endif
456{
457	register nfa_node *t;
458	static int nfa_size=0;	/* elements nfa_array[] can hold */
459
460	++nfa_allocated;
461	if (nfa_size<=nfa_allocated){
462		/* need to redo array */
463		if (!nfa_array){
464			/* need some to do inital allocation */
465			nfa_size=nfa_allocated+NFA_MIN;
466			nfa_array=(nfa_node **) malloc(sizeof(nfa_node*)*
467				nfa_size);
468		}else{
469			/* need more space */
470			nfa_size=2*(nfa_allocated+1);
471			nfa_array=(nfa_node **) realloc(nfa_array,
472				sizeof(nfa_node*)*nfa_size);
473		}
474	}
475	/* fill out entry in array */
476	t = (nfa_node*) malloc(sizeof(nfa_node));
477	nfa_array[nfa_allocated] = t;
478	*t = nfa_model_node;
479	t->node_no = nfa_allocated;
480	return t;
481}
482
483
484/* initialize the model node used to fill in newly made nfa_nodes */
485void
486#ifdef __USE_PROTOS
487make_nfa_model_node(void)
488#else
489make_nfa_model_node()
490#endif
491{
492	nfa_model_node.node_no = -1; /* impossible value for real nfa node */
493	nfa_model_node.nfa_set = 0;
494	nfa_model_node.accept = 0;   /* error state default*/
495	nfa_model_node.trans[0] = NULL;
496	nfa_model_node.trans[1] = NULL;
497	nfa_model_node.label = empty;
498}
499>>
500
501<<
502#if defined(DEBUG) || defined(_DEBUG)
503
504/* print out the pointer value and the node_number */
505void
506#ifdef __USE_PROTOS
507fprint_dfa_pair(FILE *f, nfa_node *p)
508#else
509fprint_dfa_pair(f, p)
510FILE *f;
511nfa_node *p;
512#endif
513{
514	if (p){
515		fprintf(f, "%x (%d)", p, p->node_no);
516	}else{
517		fprintf(f, "(nil)");
518	}
519}
520
521/* print out interest information on a set */
522void
523#ifdef __USE_PROTOS
524fprint_set(FILE *f, set s)
525#else
526fprint_set(f,s)
527FILE *f;
528set s;
529#endif
530{
531	unsigned int *x;
532
533	fprintf(f, "n = %d,", s.n);
534	if (s.setword){
535		fprintf(f, "setword = %x,   ", s.setword);
536		/* print out all the elements in the set */
537		x = set_pdq(s);
538		while (*x!=nil){
539			fprintf(f, "%d ", *x);
540			++x;
541		}
542	}else{
543		fprintf(f, "setword = (nil)");
544	}
545}
546
547/* code to be able to dump out the nfas
548	return 0 if okay dump
549	return 1 if screwed up
550 */
551int
552#ifdef __USE_PROTOS
553dump_nfas(int first_node, int last_node)
554#else
555dump_nfas(first_node, last_node)
556int first_node;
557int last_node;
558#endif
559{
560	register int i;
561	nfa_node *t;
562
563	for (i=first_node; i<=last_node; ++i){
564		t = NFA(i);
565		if (!t) break;
566		fprintf(stderr, "nfa_node %d {\n", t->node_no);
567		fprintf(stderr, "\n\tnfa_set = %d\n", t->nfa_set);
568		fprintf(stderr, "\taccept\t=\t%d\n", t->accept);
569		fprintf(stderr, "\ttrans\t=\t(");
570		fprint_dfa_pair(stderr, t->trans[0]);
571		fprintf(stderr, ",");
572		fprint_dfa_pair(stderr, t->trans[1]);
573		fprintf(stderr, ")\n");
574		fprintf(stderr, "\tlabel\t=\t{ ");
575		fprint_set(stderr, t->label);
576		fprintf(stderr, "\t}\n");
577		fprintf(stderr, "}\n\n");
578	}
579	return 0;
580}
581#endif
582>>
583
584<<
585/* DLG-specific syntax error message generator
586 * (define USER_ZZSYN when compiling so don't get 2 definitions)
587 */
588void
589#ifdef __USE_PROTOS
590zzsyn(char *text, int tok, char *egroup, SetWordType *eset, int etok, int k, char *bad_text)
591#else
592zzsyn(text, tok, egroup, eset, etok, k, bad_text)
593char *text, *egroup, *bad_text;
594int tok;
595int etok;
596int k;
597SetWordType *eset;
598#endif
599{
600	fprintf(stderr, ErrHdr, file_str[0]!=NULL?file_str[0]:"stdin", zzline);
601	fprintf(stderr, " syntax error at \"%s\"", (tok==zzEOF_TOKEN)?"EOF":text);
602	if ( !etok && !eset ) {fprintf(stderr, "\n"); return;}
603	if ( k==1 ) fprintf(stderr, " missing");
604	else
605	{
606		fprintf(stderr, "; \"%s\" not", bad_text);
607		if ( zzset_deg(eset)>1 ) fprintf(stderr, " in");
608	}
609	if ( zzset_deg(eset)>0 ) zzedecode(eset);
610	else fprintf(stderr, " %s", zztokens[etok]);
611	if ( strlen(egroup) > (size_t)0 ) fprintf(stderr, " in %s", egroup);
612	fprintf(stderr, "\n");
613}
614>>
615