2.9BSD - /usr/src/ucb/eyacc/ey2.c

   1: /* (c) 1979 Regents of the University of California */
   2: # include "ey.h"
   3: # define IDENTIFIER 257
   4: # define MARK 258
   5: # define TERM 259
   6: # define LEFT 260
   7: # define BINARY 261
   8: # define RIGHT 262
   9: # define PREC 263
  10: # define LCURLY 264
  11: # define C_IDENTIFIER 265  /* name followed by colon */
  12: # define NUMBER 266
  13: 
  14: setup(argc,argv) int argc; char *argv[];
  15: {   int i,j,lev,t;
  16:     int c;
  17: 
  18:     foutput = stdout;
  19:     i = 1;
  20:     while( argc >= 2  && argv[1][0] == '-' ) {
  21:         while( *++(argv[1]) ){
  22:             switch( *argv[1] ){
  23:             case 'v':
  24:             case 'V':
  25:                 foutput = copen("y.output", 'w' );
  26:                 if( foutput < 0 ) error( "cannot open y.output");
  27:                 continue;
  28:             case 'o':
  29:             case 'O':
  30:                 oflag = 1;
  31:                 continue;
  32:             case 'r':
  33:             case 'R':
  34:                 oflag = 1;
  35:                 rflag = 1;
  36:                 continue;
  37:             default:  error( "illegal option: %c", *argv[1]);
  38:                 }
  39:             }
  40:         argv++;
  41:         argc--;
  42:         }
  43: 
  44:     ftable = copen( oflag ? "yacc.tmp" : "y.tab.c" , 'w' );
  45:     if( ftable<0 ) error( "cannot open table file" );
  46:     if( argc > 1 ) cin = copen( argv[1], 'r' );
  47:     if( cin < 0 ) error( "cannot open input" );
  48:     settab();
  49:     fprintf( cout , "#\n");
  50:     ctokn = "$end";
  51:     defin(0);  /* eof */
  52:     extval = 0400;  /* beginning of assigned values */
  53:     ctokn = "error";
  54:     defin(0);
  55:     ctokn = "$accept";
  56:     defin(1);
  57:     mem=mem0;
  58:     cnamp = cnames;
  59:     lev=0;
  60:     i=0;
  61: 
  62:     while( ( t = gettok() ) != EOF ) {
  63:         switch( t ){
  64:             case IDENTIFIER:    j = chfind(0);
  65:                     trmlev[j] = lev;
  66:                     continue;
  67:             case ',':
  68:             case ';':       continue;
  69:             case TERM:      lev=0; continue;
  70:             case LEFT:      lev=(++i<<3)|01; continue;
  71:             case BINARY:    lev=(++i<<3)|02; continue;
  72:             case RIGHT: lev=(++i<<3)|03; continue;
  73:             case MARK:
  74:                     defout();
  75:                     if( rflag ){ /* RATFOR */
  76:                         fprintf( cout ,  "define yyerrok yyerrf = 0\n" );
  77:                         fprintf( cout ,  "define yyclearin yychar = -1\n" );
  78:                         fprintf( cout ,  "subroutine yyactr(yyprdn)\n");
  79:                         fprintf( cout ,  "common/yycomn/yylval,yyval,yypv,yyvalv(150)\n" );
  80:                         fprintf( cout ,  "common/yylcom/yychar,yyerrf,yydebu\n" );
  81:                         fprintf( cout ,  "integer yychar, yyerrf, yydebu\n" );
  82:                         fprintf( cout ,  "integer yyprdn,yyval,yylval,yypv,yyvalv\n" );
  83:                         }
  84:                     else {
  85:                         fprintf( cout ,  "#define yyclearin yychar = -1\n" );
  86:                         fprintf( cout ,  "#define yyerrok yyerrflag = 0\n" );
  87:                         fprintf( cout ,  "extern int yychar, yyerrflag;\n" );
  88:                         fprintf( cout , "\nint yyval 0;\nint *yypv;\nint yylval 0;");
  89:                         fprintf( cout , "\nyyactr(__np__){\n");
  90:                         }
  91:                     break;
  92:             case LCURLY:    defout();
  93:                     cpycode();
  94:                     continue;
  95:             case NUMBER:
  96:                 trmset[j].value = numbval;
  97:                 if( j < ndefout && j>2 )
  98:                     error("please define type # of %s earlier", trmset[j].name );
  99:                 continue;
 100:             default:    error("bad precedence syntax, input %d", t );
 101:             }
 102:         break;
 103:         }
 104:     prdptr[0]=mem;
 105:     /* added production */
 106:     *mem++ = NTBASE;
 107:     *mem++ = NTBASE+1;
 108:     *mem++ = 1;
 109:     *mem++ = 0;
 110:     prdptr[1]=mem;
 111:     i=0;
 112: 
 113:     /* i is 0 when a rule can begin, 1 otherwise */
 114: 
 115:     for(;;) switch( t=gettok() ) {
 116:     case C_IDENTIFIER:      if( mem == prdptr[1] ) {  /* first time */
 117:                         if( rflag ){
 118:                             fprintf( cout ,  "goto 1000\n" );
 119:                             }
 120:                         else fprintf( cout , "\nswitch(__np__){\n");
 121:                         }
 122:                 if( i != 0 ) error( "previous rule not terminated" );
 123:                 *mem = chfind(1);
 124:                 if( *mem < NTBASE )error( "token illegal on lhs of grammar rule" );
 125:                 i=1;
 126:                 ++mem;
 127:                 continue;
 128:     case IDENTIFIER:
 129:             *mem=chfind(1);
 130:             if(*mem < NTBASE)levprd[nprod]=trmlev[*mem];
 131:             mem++;
 132:             if(i==0) error("missing :");
 133:             continue;
 134:     case '=':       levprd[nprod] |= 04;
 135:                 if( i==0 ) error("semicolon preceeds action");
 136:             fprintf( cout ,  rflag?"\n%d ":"\ncase %d:", nprod );
 137:             cpyact();
 138:             fprintf( cout ,  rflag ? " return" : " break;" );
 139:     case '|':
 140:     case ';':       if(i){
 141:                 *mem++ = -nprod;
 142:                 prdptr[++nprod] = mem;
 143:                 levprd[nprod]=0;
 144:                 i=0;}
 145:             if (t=='|'){i=1;*mem++ = *prdptr[nprod-1];}
 146:             continue;
 147:     case 0:     /* End Of File */
 148:     case EOF:
 149:     case MARK:  if( i != 0 ) error( "rule not terminated before %%%% or EOF" );
 150:             settab();
 151:             finact();
 152:             /* copy the programs which follow the rules */
 153:             if( t == MARK ){
 154:                 while (( c=fgetc( cin)) != EOF ) fputc(c,cout);
 155:                 }
 156:             return;
 157:     case PREC:
 158:         if( i==0 ) error( "%%prec must appear inside rule" );
 159:         if( gettok()!=IDENTIFIER)error("illegal %%prec syntax" );
 160:         j=chfind(2);
 161:         if(j>=NTBASE)error("nonterminal %s illegal after %%prec", nontrst[j-NTBASE].name);
 162:         levprd[nprod]=trmlev[j];
 163:         continue;
 164:     case LCURLY:
 165:         if( i!=0 ) error( "%%{ appears within a rule" );
 166:         cpycode();
 167:         continue;
 168:     default: error( "syntax error, input %d", t  );
 169:     }
 170: }
 171: 
 172: finact(){
 173:     /* finish action routine */
 174:     register i;
 175: 
 176:     if( rflag ){
 177: 
 178:         fprintf( cout ,  "\n1000 goto(" );
 179:         for( i=1; i<nprod; ++i ){
 180:             fprintf( cout ,  "%d,", (levprd[i]&04)==0?999:i );
 181:             }
 182:         fprintf( cout ,  "999),yyprdn\n" );
 183:         fprintf( cout ,  "999 return\nend\n" );
 184:         fprintf( cout ,  "define YYERRCODE %d\n", trmset[2].value );
 185:         }
 186:     else {
 187:         fprintf( cout ,  "\n}\n}\n" );
 188:         fprintf( cout ,  "int yyerrval %d;\n", trmset[2].value );
 189:         }
 190:     }
 191: defin(t) {
 192: /*	define ctokn to be a terminal if t=0
 193: 	or a nonterminal if t=1		*/
 194:     char *cp,*p;
 195:     int c;
 196: 
 197: 
 198:         if (t) {
 199:           if( ++nnonter >= ntlim ) error("too many nonterminals, limit %d",ntlim);
 200:       nontrst[nnonter].name = ctokn;
 201:       return( NTBASE + nnonter );
 202:           }
 203:         else {
 204:           if( ++nterms >= tlim ) error("too many terminals, limit %d",tlim );
 205:           trmset[nterms].name = ctokn;
 206:     if( ctokn[0]==' ' && ctokn[2]=='\0' ) /* single character literal */
 207:         trmset[nterms].value = ctokn[1];
 208:     else if ( ctokn[0]==' ' && ctokn[1]=='\\' ) { /* escape sequence */
 209:         if( ctokn[3] == '\0' ){ /* single character escape sequence */
 210:             switch ( ctokn[2] ){
 211:                  /* character which is escaped */
 212:             case 'n': trmset[nterms].value = '\n'; break;
 213:             case 'r': trmset[nterms].value = '\r'; break;
 214:             case 'b': trmset[nterms].value = '\b'; break;
 215:             case 't': trmset[nterms].value = '\t'; break;
 216:             case '\'': trmset[nterms].value = '\''; break;
 217:             case '"': trmset[nterms].value = '"'; break;
 218:             case '\\': trmset[nterms].value = '\\'; break;
 219:             default: error( "invalid escape" );
 220:                 }
 221:             }
 222:         else if( ctokn[2] <= '7' && ctokn[2]>='0' ){ /* \nnn sequence */
 223:             if( ctokn[3]<'0' || ctokn[3] > '7' || ctokn[4]<'0' ||
 224:                 ctokn[4]>'7' || ctokn[5] != '\0' ) error("illegal \\nnn construction" );
 225:             trmset[nterms].value = 64*(ctokn[2]-'0')+8*(ctokn[3]-'0')+ctokn[4]-'0';
 226:             if( trmset[nterms].value == 0 ) error( "'\\000' is illegal" );
 227:             }
 228:         }
 229:     else {
 230:         trmset[nterms].value = extval++;
 231: 
 232:         }
 233:     trmlev[nterms] = 0;
 234:     return( nterms );
 235:           }
 236: }
 237: 
 238: defout(){ /* write out the defines (at the end of the declaration section) */
 239: 
 240:     _REGISTER int i, c;
 241:     _REGISTER char *cp;
 242: 
 243:     for( i=ndefout; i<=nterms; ++i ){
 244: 
 245:         cp = trmset[i].name;
 246:         if( *cp == ' ' ) ++cp;  /* literals */
 247: 
 248:         for( ; (c= *cp)!='\0'; ++cp ){
 249: 
 250:             if( c>='a' && c<='z' ||
 251:                 c>='A' && c<='Z' ||
 252:                 c>='0' && c<='9' ||
 253:                 c=='_' )  ; /* VOID */
 254:             else goto nodef;
 255:             }
 256: 
 257:         /* define it */
 258: 
 259:         fprintf( cout ,  "%c define %s %d\n", rflag?' ':'#', trmset[i].name, trmset[i].value );
 260: 
 261:     nodef:  ;
 262:         }
 263: 
 264:     ndefout = nterms+1;
 265: 
 266:     }
 267: 
 268: chstash( c ){
 269:   /* put character away into cnames */
 270:   if( cnamp >= &cnames[cnamsz] ) error("too many characters in id's and literals" );
 271:   else *cnamp++ = c;
 272:   }
 273: 
 274: int gettok() {
 275:     int j, base;
 276:     static int peekline; /* number of '\n' seen in lookahead */
 277:     auto int c, match, reserve;
 278: 
 279: begin:
 280:     reserve = 0;
 281:         if( peekc>=0 ) {
 282:         c = peekc;
 283:         lineno += peekline;
 284:         peekc = -1;
 285:         peekline = 0;
 286:         }
 287:         else c = fgetc( cin);
 288:         while( c==' ' || c=='\n' || c=='\t' || c == '\014'){
 289:           if( c == '\n' ) ++lineno;
 290:           c=fgetc( cin);
 291:           }
 292:     if (c=='/')
 293:         {if (fgetc( cin)!='*')error("illegal /");
 294:         c=fgetc( cin);
 295:         while(c != EOF) {
 296:             if( c == '\n' ) ++lineno;
 297:             if (c=='*')
 298:                 {if((c=fgetc( cin))=='/')break;}
 299:             else c=fgetc( cin);}
 300:         if (!c) return(0);
 301:         goto begin;}
 302:     j=0;
 303:     switch(c){
 304:     case '"':
 305:     case '\'':  match = c;
 306:             ctokn = cnamp;
 307:             chstash( ' ' );
 308:             while(1){
 309:                 c = fgetc( cin);
 310:                 if( c == '\n' || c == '\0' )
 311:                     error("illegal or missing ' or \"");
 312:                 if( c == '\\' ){
 313:                     c = fgetc( cin);
 314:                     chstash( '\\' );
 315:                     }
 316:                 else if( c == match ) break;
 317:                 chstash( c );
 318:                 }
 319:             break;
 320:     case '%':
 321:     case '\\':  switch(c=fgetc( cin))
 322:         {case '0':  return(TERM);
 323:         case '<':   return(LEFT);
 324:         case '2':   return(BINARY);
 325:         case '>':   return(RIGHT);
 326:         case '%':
 327:         case '\\':  return(MARK);
 328:         case '=':   return(PREC);
 329:         case '{':   return(LCURLY);
 330:         default:    reserve = 1;
 331:         }
 332:     default:    if( c >= '0' && c <= '9' ){ /* number */
 333:                 numbval = c-'0' ;
 334:                 base = (c=='0') ? 8 : 10 ;
 335:                 for( c=fgetc( cin); c>='0' && c<='9'; c=fgetc( cin) ){
 336:                     numbval = numbval*base + c - '0';
 337:                     }
 338:                 peekc = c;
 339:                 return(NUMBER);
 340:                 }
 341:             else if( (c>='a'&&c<='z')||(c>='A'&&c<='Z')||c=='_'||c=='.'||c=='$'){
 342:                 ctokn = cnamp;
 343:                 while(  (c>='a'&&c<='z') ||
 344:                     (c>='A'&&c<='Z') ||
 345:                     (c>='0'&&c<='9') ||
 346:                     c=='_' || c=='.' || c=='$' ) {
 347:                     chstash( c );
 348:                     if( peekc>=0 ) { c = peekc; peekc = -1; }
 349:                     else c = fgetc( cin);
 350:                     }
 351:                 }
 352:             else return(c);
 353: 
 354:             peekc=c;
 355:             }
 356:     chstash( '\0' );
 357: 
 358:     if( reserve ){ /* find a reserved word */
 359:         if( compare("term")) return( TERM );
 360:         if( compare("TERM")) return( TERM );
 361:         if( compare("token")) return( TERM );
 362:         if( compare("TOKEN")) return( TERM );
 363:         if( compare("left")) return( LEFT );
 364:         if( compare("LEFT")) return( LEFT );
 365:         if( compare("nonassoc")) return( BINARY );
 366:         if( compare("NONASSOC")) return( BINARY );
 367:         if( compare("binary")) return( BINARY );
 368:         if( compare("BINARY")) return( BINARY );
 369:         if( compare("right")) return( RIGHT );
 370:         if( compare("RIGHT")) return( RIGHT );
 371:         if( compare("prec")) return( PREC );
 372:         if( compare("PREC")) return( PREC );
 373:         error("invalid escape, or illegal reserved word: %s", ctokn );
 374:         }
 375: 
 376:     /* look ahead to distinguish IDENTIFIER from C_IDENTIFIER */
 377: 
 378:   look:
 379:     while( peekc==' ' || peekc=='\t' || peekc == '\n' || peekc == '\014' )
 380:     {
 381:         if( peekc == '\n' ) ++peekline;
 382:         peekc = fgetc( cin);
 383:     }
 384: 
 385:     if( peekc != ':' ) return( IDENTIFIER );
 386:     peekc = -1;
 387:     lineno += peekline;
 388:     peekline = 0;
 389:     return( C_IDENTIFIER );
 390: }
 391: chfind(t)
 392: 
 393: {   int i,j;
 394: 
 395:     if (ctokn[0]==' ')t=0;
 396:     for(i=1;i<=nterms;i++)
 397:         if(compare(trmset[i].name)){
 398:             cnamp = ctokn;
 399:             return( i );
 400:             }
 401:     for(i=1;i<=nnonter;i++)
 402:         if(compare(nontrst[i].name)) {
 403:             cnamp = ctokn;
 404:             return( i+NTBASE );
 405:             }
 406:     /* cannot find name */
 407:     if( t>1 && ctokn[0] != ' ' )
 408:         error( "%s should have been defined earlier", ctokn );
 409:     return( defin( t ) );
 410:     }
 411: 
 412: cpycode(){ /* copies code between \{ and \} */
 413: 
 414:     int c;
 415:     c = fgetc( cin);
 416:     if( c == '\n' ) {
 417:         c = fgetc( cin);
 418:         lineno++;
 419:         }
 420:     while( c != EOF ){
 421:         if( c=='\\' )
 422:             if( (c=fgetc( cin)) == '}' ) return;
 423:             else fputc('\\',cout);
 424:         if( c=='%' )
 425:             if( (c=fgetc( cin)) == '}' ) return;
 426:             else fputc('%',cout);
 427:         fputc( c, cout );
 428:         if( c == '\n' ) ++lineno;
 429:         c = fgetc( cin);
 430:         }
 431:     error("eof before %%}");
 432:     }
 433: 
 434: cpyact(){ /* copy C action to the next ; or closing } */
 435:     int brac, c, match, *i, j, s;
 436: 
 437:     brac = 0;
 438: 
 439: loop:
 440:     c = fgetc( cin);
 441: swt:
 442:     switch( c ){
 443: 
 444: case ';':
 445:         if( brac == 0 ){
 446:             fputc( c, cout );
 447:             return;
 448:             }
 449:         goto lcopy;
 450: 
 451: case '{':
 452:         brac++;
 453:         goto lcopy;
 454: 
 455: case '$':
 456:         s = 1;
 457:         c = fgetc( cin);
 458:         if( c == '$' ){
 459:             fprintf( cout , "yyval");
 460:             goto loop;
 461:             }
 462:         if( c == '-' ){
 463:             s = -s;
 464:             c = fgetc( cin);
 465:             }
 466:         if( c>='0' && c <= '9' ){
 467:             j=0;
 468:             while( c>='0' && c<= '9' ){
 469:                 j= j*10+c-'0';
 470:                 c = fgetc( cin);
 471:                 }
 472:             if( rflag ) fprintf( cout ,  "yyvalv(yypv%c%d)", s==1?'+':'-', j );
 473:             else fprintf( cout , "yypv[%d]", s*j );
 474:             goto swt;
 475:             }
 476:         fputc( '$' , cout);
 477:         if( s<0 ) fputc('-', cout);
 478:         goto swt;
 479: 
 480: case '}':
 481:         brac--;
 482:         if( brac == 0 ){
 483:             fputc( c , cout);
 484:             return;
 485:             }
 486:         goto lcopy;
 487: 
 488: case '/':   /* look for comments */
 489:         fputc( c ,cout);
 490:         c = fgetc( cin);
 491:         if( c != '*' ) goto swt;
 492: 
 493:         /* it really is a comment */
 494: 
 495:         fputc( c , cout);
 496:         while( (c=fgetc( cin)) != EOF ){
 497:             if( c=='*' ){
 498:                 fputc( c , cout);
 499:                 if( (c=fgetc( cin)) == '/' ) goto lcopy;
 500:                 }
 501:             fputc( c , cout);
 502:             }
 503:         error( "EOF inside comment" );
 504: 
 505: case '\'':  /* character constant */
 506:         match = '\'';
 507:         goto string;
 508: 
 509: case '"':   /* character string */
 510:         match = '"';
 511: 
 512:     string:
 513: 
 514:         fputc( c , cout);
 515:         while( (c=fgetc( cin)) != EOF ){
 516: 
 517:             if( c=='\\' ){
 518:                 fputc( c , cout);
 519:                 c=fgetc( cin);
 520:                 }
 521:             else if( c==match ) goto lcopy;
 522:             fputc( c , cout);
 523:             }
 524:         error( "EOF in string or character constant" );
 525: 
 526: case '\0':
 527:         error("action does not terminate");
 528: case '\n':  ++lineno;
 529:         goto lcopy;
 530: 
 531:         }
 532: 
 533: lcopy:
 534:     fputc( c , cout);
 535:     goto loop;
 536:     }

Defined functions

Defined macros