|
|
1.1 root 1: %token CHAR CCL NCCL STR DELIM SCON ITER NEWE NULLS
2: %left SCON '/' NEWE
3: %left '|'
4: %left '$' '^'
5: %left CHAR CCL NCCL '(' '.' STR NULLS
6: %left ITER
7: %left CAT
8: %left '*' '+' '?'
9:
10: %{
11: #define YYSTYPE union _yystype_
12: union _yystype_
13: {
14: int i;
15: char *cp;
16: };
17: # include "ldefs.c"
18: %}
19: %%
20: %{
21: int i;
22: int j,k;
23: int g;
24: char *p;
25: %}
26: acc : lexinput
27: ={
28: # ifdef DEBUG
29: if(debug) sect2dump();
30: # endif
31: }
32: ;
33: lexinput: defns delim prods end
34: | defns delim end
35: ={
36: if(!funcflag)phead2();
37: funcflag = TRUE;
38: }
39: | error
40: ={
41: # ifdef DEBUG
42: if(debug) {
43: sect1dump();
44: sect2dump();
45: }
46: # endif
47: }
48: ;
49: end: delim | ;
50: defns: defns STR STR
51: ={ scopy($2,dp);
52: def[dptr] = dp;
53: dp += slength($2) + 1;
54: scopy($3,dp);
55: subs[dptr++] = dp;
56: if(dptr >= DEFSIZE)
57: error("Too many definitions");
58: dp += slength($3) + 1;
59: if(dp >= dchar+DEFCHAR)
60: error("Definitions too long");
61: subs[dptr]=def[dptr]=0; /* for lookup - require ending null */
62: }
63: |
64: ;
65: delim: DELIM
66: ={
67: # ifdef DEBUG
68: if(sect == DEFSECTION && debug) sect1dump();
69: # endif
70: sect++;
71: }
72: ;
73: prods: prods pr
74: ={ $$.i = mn2(RNEWE,$1.i,$2.i);
75: }
76: | pr
77: ={ $$.i = $1.i;}
78: ;
79: pr: r NEWE
80: ={
81: if(divflg == TRUE)
82: i = mn1(S1FINAL,casecount);
83: else i = mn1(FINAL,casecount);
84: $$.i = mn2(RCAT,$1.i,i);
85: divflg = FALSE;
86: casecount++;
87: }
88: | error NEWE
89: ={
90: # ifdef DEBUG
91: if(debug) sect2dump();
92: # endif
93: }
94: r: CHAR
95: ={ $$.i = mn0($1.i); }
96: | STR
97: ={
98: p = $1.cp;
99: i = mn0(*p++);
100: while(*p)
101: i = mn2(RSTR,i,*p++);
102: $$.i = i;
103: }
104: | '.'
105: ={ symbol['\n'] = 0;
106: if(psave == FALSE){
107: p = ccptr;
108: psave = ccptr;
109: for(i=1;i<'\n';i++){
110: symbol[i] = 1;
111: *ccptr++ = i;
112: }
113: for(i='\n'+1;i<NCH;i++){
114: symbol[i] = 1;
115: *ccptr++ = i;
116: }
117: *ccptr++ = 0;
118: if(ccptr > ccl+CCLSIZE)
119: error("Too many large character classes");
120: }
121: else
122: p = psave;
123: $$.i = mn1(RCCL,p);
124: cclinter(1);
125: }
126: | CCL
127: ={ $$.i = mn1(RCCL,$1.i); }
128: | NCCL
129: ={ $$.i = mn1(RNCCL,$1.i); }
130: | r '*'
131: ={ $$.i = mn1(STAR,$1.i); }
132: | r '+'
133: ={ $$.i = mn1(PLUS,$1.i); }
134: | r '?'
135: ={ $$.i = mn1(QUEST,$1.i); }
136: | r '|' r
137: ={ $$.i = mn2(BAR,$1.i,$3.i); }
138: | r r %prec CAT
139: ={ $$.i = mn2(RCAT,$1.i,$2.i); }
140: | r '/' r
141: ={ if(!divflg){
142: j = mn1(S2FINAL,-casecount);
143: i = mn2(RCAT,$1.i,j);
144: $$.i = mn2(DIV,i,$3.i);
145: }
146: else {
147: $$.i = mn2(RCAT,$1.i,$3.i);
148: warning("Extra slash removed");
149: }
150: divflg = TRUE;
151: }
152: | r ITER ',' ITER '}'
153: ={ if($2.i > $4.i){
154: i = $2.i;
155: $2.i = $4.i;
156: $4.i = i;
157: }
158: if($4.i <= 0)
159: warning("Iteration range must be positive");
160: else {
161: j = $1.i;
162: for(k = 2; k<=$2.i;k++)
163: j = mn2(RCAT,j,dupl($1.i));
164: for(i = $2.i+1; i<=$4.i; i++){
165: g = dupl($1.i);
166: for(k=2;k<=i;k++)
167: g = mn2(RCAT,g,dupl($1.i));
168: j = mn2(BAR,j,g);
169: }
170: $$.i = j;
171: }
172: }
173: | r ITER '}'
174: ={
175: if($2.i < 0)warning("Can't have negative iteration");
176: else if($2.i == 0) $$.i = mn0(RNULLS);
177: else {
178: j = $1.i;
179: for(k=2;k<=$2.i;k++)
180: j = mn2(RCAT,j,dupl($1.i));
181: $$.i = j;
182: }
183: }
184: | r ITER ',' '}'
185: ={
186: /* from n to infinity */
187: if($2.i < 0)warning("Can't have negative iteration");
188: else if($2.i == 0) $$.i = mn1(STAR,$1.i);
189: else if($2.i == 1)$$.i = mn1(PLUS,$1.i);
190: else { /* >= 2 iterations minimum */
191: j = $1.i;
192: for(k=2;k<$2.i;k++)
193: j = mn2(RCAT,j,dupl($1.i));
194: k = mn1(PLUS,dupl($1.i));
195: $$.i = mn2(RCAT,j,k);
196: }
197: }
198: | SCON r
199: ={ $$.i = mn2(RSCON,$2.i,$1.i); }
200: | '^' r
201: ={ $$.i = mn1(CARAT,$2.i); }
202: | r '$'
203: ={ i = mn0('\n');
204: if(!divflg){
205: j = mn1(S2FINAL,-casecount);
206: k = mn2(RCAT,$1.i,j);
207: $$.i = mn2(DIV,k,i);
208: }
209: else $$.i = mn2(RCAT,$1.i,i);
210: divflg = TRUE;
211: }
212: | '(' r ')'
213: ={ $$.i = $2.i; }
214: | NULLS
215: ={ $$.i = mn0(RNULLS); }
216: ;
217: %%
218: yylex(){
219: register char *p;
220: register int c, i;
221: char *t, *xp;
222: int n, j, k, x;
223: static int sectbegin;
224: static char token[TOKENSIZE];
225: static int iter;
226:
227: # ifdef DEBUG
228: yylval.i = 0;
229: # endif
230:
231: if(sect == DEFSECTION) { /* definitions section */
232: while(!eof) {
233: if(prev == '\n'){ /* next char is at beginning of line */
234: getl(p=buf);
235: switch(*p){
236: case '%':
237: switch(c= *(p+1)){
238: case '%':
239: lgate();
240: if(!ratfor)fprintf(fout,"# ");
241: fprintf(fout,"define YYNEWLINE %d\n",ctable['\n']);
242: if(!ratfor)fprintf(fout,"yylex(){\nint nstr; extern int yyprevious;\n");
243: sectbegin = TRUE;
244: i = treesize*(sizeof(*name)+sizeof(*left)+
245: sizeof(*right)+sizeof(*nullstr)+sizeof(*parent))+ALITTLEEXTRA;
246: c = (int)myalloc(i,1);
247: if(c == 0)
248: error("Too little core for parse tree");
249: p = (char *)c;
250: cfree((char *)p,i,1);
251: name = (int *)myalloc(treesize,sizeof(*name));
252: left = (int *)myalloc(treesize,sizeof(*left));
253: right = (int *)myalloc(treesize,sizeof(*right));
254: nullstr = myalloc(treesize,sizeof(*nullstr));
255: parent = (int *)myalloc(treesize,sizeof(*parent));
256: if(name == 0 || left == 0 || right == 0 || parent == 0 || nullstr == 0)
257: error("Too little core for parse tree");
258: return(freturn(DELIM));
259: case 'p': case 'P': /* has overridden number of positions */
260: while(*p && !digit(*p))p++;
261: maxpos = siconv(p);
262: # ifdef DEBUG
263: if (debug) printf("positions (%%p) now %d\n",maxpos);
264: # endif
265: if(report == 2)report = 1;
266: continue;
267: case 'n': case 'N': /* has overridden number of states */
268: while(*p && !digit(*p))p++;
269: nstates = siconv(p);
270: # ifdef DEBUG
271: if(debug)printf( " no. states (%%n) now %d\n",nstates);
272: # endif
273: if(report == 2)report = 1;
274: continue;
275: case 'e': case 'E': /* has overridden number of tree nodes */
276: while(*p && !digit(*p))p++;
277: treesize = siconv(p);
278: # ifdef DEBUG
279: if (debug) printf("treesize (%%e) now %d\n",treesize);
280: # endif
281: if(report == 2)report = 1;
282: continue;
283: case 'o': case 'O':
284: while (*p && !digit(*p))p++;
285: outsize = siconv(p);
286: if (report ==2) report=1;
287: continue;
288: case 'a': case 'A': /* has overridden number of transitions */
289: while(*p && !digit(*p))p++;
290: if(report == 2)report = 1;
291: ntrans = siconv(p);
292: # ifdef DEBUG
293: if (debug)printf("N. trans (%%a) now %d\n",ntrans);
294: # endif
295: continue;
296: case 'k': case 'K': /* overriden packed char classes */
297: while (*p && !digit(*p))p++;
298: if (report==2) report=1;
299: cfree((char *)pchar, pchlen, sizeof(*pchar));
300: pchlen = siconv(p);
301: # ifdef DEBUG
302: if (debug) printf( "Size classes (%%k) now %d\n",pchlen);
303: # endif
304: pchar=pcptr=myalloc(pchlen, sizeof(*pchar));
305: continue;
306: case 't': case 'T': /* character set specifier */
307: ZCH = atoi(p+2);
308: if (ZCH < NCH) ZCH = NCH;
309: if (ZCH > 2*NCH) error("ch table needs redeclaration");
310: chset = TRUE;
311: for(i = 0; i<ZCH; i++)
312: ctable[i] = 0;
313: while(getl(p) && scomp(p,"%T") != 0 && scomp(p,"%t") != 0){
314: if((n = siconv(p)) <= 0 || n > ZCH){
315: warning("Character value %d out of range",n);
316: continue;
317: }
318: while(!space(*p) && *p) p++;
319: while(space(*p)) p++;
320: t = p;
321: while(*t){
322: c = ctrans(&t);
323: if(ctable[c]){
324: if (printable(c))
325: warning("Character '%c' used twice",c);
326: else
327: warning("Character %o used twice",c);
328: }
329: else ctable[c] = n;
330: t++;
331: }
332: p = buf;
333: }
334: {
335: char chused[2*NCH]; int kr;
336: for(i=0; i<ZCH; i++)
337: chused[i]=0;
338: for(i=0; i<NCH; i++)
339: chused[ctable[i]]=1;
340: for(kr=i=1; i<NCH; i++)
341: if (ctable[i]==0)
342: {
343: while (chused[kr] == 0)
344: kr++;
345: ctable[i]=kr;
346: chused[kr]=1;
347: }
348: }
349: lgate();
350: continue;
351: case 'r': case 'R':
352: c = 'r';
353: case 'c': case 'C':
354: if(lgatflg)
355: error("Too late for language specifier");
356: ratfor = (c == 'r');
357: continue;
358: case '{':
359: lgate();
360: while(getl(p) && scomp(p,"%}") != 0)
361: fprintf(fout, "%s\n",p);
362: if(p[0] == '%') continue;
363: error("Premature eof");
364: case 's': case 'S': /* start conditions */
365: lgate();
366: while(*p && index(*p," \t,") < 0) p++;
367: n = TRUE;
368: while(n){
369: while(*p && index(*p," \t,") >= 0) p++;
370: t = p;
371: while(*p && index(*p," \t,") < 0)p++;
372: if(!*p) n = FALSE;
373: *p++ = 0;
374: if (*t == 0) continue;
375: i = sptr*2;
376: if(!ratfor)fprintf(fout,"# ");
377: fprintf(fout,"define %s %d\n",t,i);
378: scopy(t,sp);
379: sname[sptr++] = sp;
380: sname[sptr] = 0; /* required by lookup */
381: if(sptr >= STARTSIZE)
382: error("Too many start conditions");
383: sp += slength(sp) + 1;
384: if(sp >= schar+STARTCHAR)
385: error("Start conditions too long");
386: }
387: continue;
388: default:
389: warning("Invalid request %s",p);
390: continue;
391: } /* end of switch after seeing '%' */
392: case ' ': case '\t': /* must be code */
393: lgate();
394: fprintf(fout, "%s\n",p);
395: continue;
396: default: /* definition */
397: while(*p && !space(*p)) p++;
398: if(*p == 0)
399: continue;
400: prev = *p;
401: *p = 0;
402: bptr = p+1;
403: yylval.cp = buf;
404: if(digit(buf[0]))
405: warning("Substitution strings may not begin with digits");
406: return(freturn(STR));
407: }
408: }
409: /* still sect 1, but prev != '\n' */
410: else {
411: p = bptr;
412: while(*p && space(*p)) p++;
413: if(*p == 0)
414: warning("No translation given - null string assumed");
415: scopy(p,token);
416: yylval.cp = token;
417: prev = '\n';
418: return(freturn(STR));
419: }
420: }
421: /* end of section one processing */
422: }
423: else if(sect == RULESECTION){ /* rules and actions */
424: while(!eof){
425: switch(c=gch()){
426: case '\0':
427: return(freturn(0));
428: case '\n':
429: if(prev == '\n') continue;
430: x = NEWE;
431: break;
432: case ' ':
433: case '\t':
434: if(sectbegin == TRUE){
435: cpyact();
436: while((c=gch()) && c != '\n');
437: continue;
438: }
439: if(!funcflag)phead2();
440: funcflag = TRUE;
441: if(ratfor)fprintf(fout,"%d\n",30000+casecount);
442: else fprintf(fout,"case %d:\n",casecount);
443: if(cpyact()){
444: if(ratfor)fprintf(fout,"goto 30997\n");
445: else fprintf(fout,"break;\n");
446: }
447: while((c=gch()) && c != '\n');
448: if(peek == ' ' || peek == '\t' || sectbegin == TRUE){
449: warning("Executable statements should occur right after %%");
450: continue;
451: }
452: x = NEWE;
453: break;
454: case '%':
455: if(prev != '\n') goto character;
456: if(peek == '{'){ /* included code */
457: getl(buf);
458: while(!eof && getl(buf) && scomp("%}",buf) != 0)
459: fprintf(fout,"%s\n",buf);
460: continue;
461: }
462: if(peek == '%'){
463: c = gch();
464: c = gch();
465: x = DELIM;
466: break;
467: }
468: goto character;
469: case '|':
470: if(peek == ' ' || peek == '\t' || peek == '\n'){
471: if(ratfor)fprintf(fout,"%d\n",30000+casecount++);
472: else fprintf(fout,"case %d:\n",casecount++);
473: continue;
474: }
475: x = '|';
476: break;
477: case '$':
478: if(peek == '\n' || peek == ' ' || peek == '\t' || peek == '|' || peek == '/'){
479: x = c;
480: break;
481: }
482: goto character;
483: case '^':
484: if(prev != '\n' && scon != TRUE) goto character; /* valid only at line begin */
485: x = c;
486: break;
487: case '?':
488: case '+':
489: case '.':
490: case '*':
491: case '(':
492: case ')':
493: case ',':
494: case '/':
495: x = c;
496: break;
497: case '}':
498: iter = FALSE;
499: x = c;
500: break;
501: case '{': /* either iteration or definition */
502: if(digit(c=gch())){ /* iteration */
503: iter = TRUE;
504: ieval:
505: i = 0;
506: while(digit(c)){
507: token[i++] = c;
508: c = gch();
509: }
510: token[i] = 0;
511: yylval.i = siconv(token);
512: munput('c',c);
513: x = ITER;
514: break;
515: }
516: else { /* definition */
517: i = 0;
518: while(c && c!='}'){
519: token[i++] = c;
520: c = gch();
521: }
522: token[i] = 0;
523: i = lookup(token,def);
524: if(i < 0)
525: warning("Definition %s not found",token);
526: else
527: munput('s',subs[i]);
528: continue;
529: }
530: case '<': /* start condition ? */
531: if(prev != '\n') /* not at line begin, not start */
532: goto character;
533: t = slptr;
534: do {
535: i = 0;
536: c = gch();
537: while(c != ',' && c && c != '>'){
538: token[i++] = c;
539: c = gch();
540: }
541: token[i] = 0;
542: if(i == 0)
543: goto character;
544: i = lookup(token,sname);
545: if(i < 0) {
546: warning("Undefined start condition %s",token);
547: continue;
548: }
549: *slptr++ = i+1;
550: } while(c && c != '>');
551: *slptr++ = 0;
552: /* check if previous value re-usable */
553: for (xp=slist; xp<t; )
554: {
555: if (strcmp(xp, t)==0)
556: break;
557: while (*xp++);
558: }
559: if (xp<t)
560: {
561: /* re-use previous pointer to string */
562: slptr=t;
563: t=xp;
564: }
565: if(slptr > slist+STARTSIZE) /* note not packed ! */
566: error("Too many start conditions used");
567: yylval.cp = t;
568: x = SCON;
569: break;
570: case '"':
571: i = 0;
572: while((c=gch()) && c != '"' && c != '\n'){
573: if(c == '\\') c = usescape(c=gch());
574: token[i++] = c;
575: if(i > TOKENSIZE){
576: warning("String too long");
577: i = TOKENSIZE-1;
578: break;
579: }
580: }
581: if(c == '\n') {
582: yyline--;
583: warning("Non-terminated string");
584: yyline++;
585: }
586: token[i] = 0;
587: if(i == 0)x = NULLS;
588: else if(i == 1){
589: yylval.i = token[0];
590: x = CHAR;
591: }
592: else {
593: yylval.cp = token;
594: x = STR;
595: }
596: break;
597: case '[':
598: for(i=1;i<NCH;i++) symbol[i] = 0;
599: x = CCL;
600: if((c = gch()) == '^'){
601: x = NCCL;
602: c = gch();
603: }
604: while(c != ']' && c){
605: if(c == '\\') c = usescape(c=gch());
606: symbol[c] = 1;
607: j = c;
608: if((c=gch()) == '-' && peek != ']'){ /* range specified */
609: c = gch();
610: if(c == '\\') c = usescape(c=gch());
611: k = c;
612: if(j > k) {
613: n = j;
614: j = k;
615: k = n;
616: }
617: if(!(('A' <= j && k <= 'Z') ||
618: ('a' <= j && k <= 'z') ||
619: ('0' <= j && k <= '9')))
620: warning("Non-portable Character Class");
621: for(n=j+1;n<=k;n++)
622: symbol[n] = 1; /* implementation dependent */
623: c = gch();
624: }
625: }
626: /* try to pack ccl's */
627: i = 0;
628: for(j=0;j<NCH;j++)
629: if(symbol[j])token[i++] = j;
630: token[i] = 0;
631: p = ccptr;
632: if(optim){
633: p = ccl;
634: while(p <ccptr && scomp(token,p) != 0)p++;
635: }
636: if(p < ccptr) /* found it */
637: yylval.cp = p;
638: else {
639: yylval.cp = ccptr;
640: scopy(token,ccptr);
641: ccptr += slength(token) + 1;
642: if(ccptr >= ccl+CCLSIZE)
643: error("Too many large character classes");
644: }
645: cclinter(x==CCL);
646: break;
647: case '\\':
648: c = usescape(c=gch());
649: default:
650: character:
651: if(iter){ /* second part of an iteration */
652: iter = FALSE;
653: if('0' <= c && c <= '9')
654: goto ieval;
655: }
656: if(alpha(peek)){
657: i = 0;
658: yylval.cp = token;
659: token[i++] = c;
660: while(alpha(peek))
661: token[i++] = gch();
662: if(peek == '?' || peek == '*' || peek == '+')
663: munput('c',token[--i]);
664: token[i] = 0;
665: if(i == 1){
666: yylval.i = token[0];
667: x = CHAR;
668: }
669: else x = STR;
670: }
671: else {
672: yylval.i = c;
673: x = CHAR;
674: }
675: }
676: scon = FALSE;
677: if(x == SCON)scon = TRUE;
678: sectbegin = FALSE;
679: return(freturn(x));
680: }
681: }
682: /* section three */
683: ptail();
684: # ifdef DEBUG
685: if(debug)
686: fprintf(fout,"\n/*this comes from section three - debug */\n");
687: # endif
688: while(getl(buf) && !eof)
689: fprintf(fout,"%s\n",buf);
690: return(freturn(0));
691: }
692: /* end of yylex */
693: # ifdef DEBUG
694: freturn(i)
695: int i; {
696: if(yydebug) {
697: printf("now return ");
698: if(i < NCH) allprint(i);
699: else printf("%d",i);
700: printf(" yylval = ");
701: switch(i){
702: case STR: case CCL: case NCCL:
703: strpt(yylval.cp);
704: break;
705: case CHAR:
706: allprint(yylval.i);
707: break;
708: default:
709: printf("%d",yylval.i);
710: break;
711: }
712: putchar('\n');
713: }
714: return(i);
715: }
716: # endif
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.