src/reduce.c

   1 /* Grammar reduction for Bison.
   2    Copyright (C) 1988, 1989 Free Software Foundation, Inc.
   3
   4 This file is part of Bison, the GNU Compiler Compiler.
   5
   6 Bison is free software; you can redistribute it and/or modify
   7 it under the terms of the GNU General Public License as published by
   8 the Free Software Foundation; either version 2, or (at your option)
   9 any later version.
  10
  11 Bison is distributed in the hope that it will be useful,
  12 but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 GNU General Public License for more details.
  15
  16 You should have received a copy of the GNU General Public License
  17 along with Bison; see the file COPYING.  If not, write to
  18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19 Boston, MA 02111-1307, USA.  */
  20
  21
  22 /*
  23  * Reduce the grammar:  Find and eliminate unreachable terminals,
  24  * nonterminals, and productions.  David S. Bakin.
  25  */
  26
  27 /*
  28  * Don't eliminate unreachable terminals:  They may be used by the user's
  29  * parser.
  30  */
  31
  32 #include <stdio.h>
  33 #include "system.h"
  34 #include "files.h"
  35 #include "gram.h"
  36 #include "machine.h"
  37 #include "alloc.h"
  38
  39
  40 extern char   **tags;           /* reader.c */
  41 extern int      verboseflag;    /* getargs.c */
  42 static int      statisticsflag; /* XXXXXXX */
  43 extern int      fixed_outfiles;
  44
  45 #ifndef TRUE
  46 #define TRUE    (1)
  47 #define FALSE   (0)
  48 #endif
  49 typedef int bool;
  50 typedef unsigned *BSet;
  51 typedef short  *rule;
  52
  53
  54 /*
  55  * N is set of all nonterminals which are not useless.  P is set of all rules
  56  * which have no useless nonterminals in their RHS.  V is the set of all
  57  * accessible symbols.
  58  */
  59
  60 static BSet     N, P, V, V1;
  61
  62 static int      nuseful_productions, nuseless_productions,
  63                 nuseful_nonterminals, nuseless_nonterminals;
  64
  65
  66 bool bits_equal PARAMS((BSet, BSet, int));
  67 int nbits PARAMS((unsigned));
  68 int bits_size PARAMS((BSet, int));
  69 void reduce_grammar PARAMS((void));
  70 static void useless_nonterminals PARAMS((void));
  71 static void inaccessable_symbols PARAMS((void));
  72 static void reduce_grammar_tables PARAMS((void));
  73 static void print_results PARAMS((void));
  74 static void print_notices PARAMS((void));
  75 void dump_grammar PARAMS((void));
  76
  77 extern void fatals PARAMS((char *, char *));
  78 \f
  79
  80 bool
  81 bits_equal (BSet L, BSet R, int n)
  82 {
  83   int i;
  84
  85   for (i = n - 1; i >= 0; i--)
  86     if (L[i] != R[i])
  87       return FALSE;
  88   return TRUE;
  89 }
  90
  91
  92 int
  93 nbits (unsigned i)
  94 {
  95   int count = 0;
  96
  97   while (i != 0) {
  98     i ^= (i & ((unsigned) (- (int) i)));
  99     ++count;
 100   }
 101   return count;
 102 }
 103
 104
 105 int
 106 bits_size (BSet S, int n)
 107 {
 108   int i, count = 0;
 109
 110   for (i = n - 1; i >= 0; i--)
 111     count += nbits(S[i]);
 112   return count;
 113 }
 114 \f
 115 void
 116 reduce_grammar (void)
 117 {
 118   bool reduced;
 119
 120   /* Allocate the global sets used to compute the reduced grammar */
 121
 122   N = NEW2(WORDSIZE(nvars), unsigned);
 123   P = NEW2(WORDSIZE(nrules + 1), unsigned);
 124   V = NEW2(WORDSIZE(nsyms), unsigned);
 125   V1 = NEW2(WORDSIZE(nsyms), unsigned);
 126
 127   useless_nonterminals();
 128   inaccessable_symbols();
 129
 130   reduced = (bool) (nuseless_nonterminals + nuseless_productions > 0);
 131
 132   if (verboseflag)
 133     print_results();
 134
 135   if (reduced == FALSE)
 136     goto done_reducing;
 137
 138   print_notices();
 139
 140   if (!BITISSET(N, start_symbol - ntokens))
 141     fatals(_("Start symbol %s does not derive any sentence"),
 142            tags[start_symbol]);
 143
 144   reduce_grammar_tables();
 145   /* if (verboseflag) {
 146      fprintf(foutput, "REDUCED GRAMMAR\n\n");
 147      dump_grammar();
 148      }
 149      */
 150
 151   /**/ statisticsflag = FALSE; /* someday getopts should handle this */
 152   if (statisticsflag == TRUE)
 153     fprintf(stderr,
 154             _("reduced %s defines %d terminal%s, %d nonterminal%s\
 155 , and %d production%s.\n"), infile,
 156             ntokens, (ntokens == 1 ? "" : "s"),
 157             nvars,   (nvars   == 1 ? "" : "s"),
 158             nrules,  (nrules  == 1 ? "" : "s"));
 159
 160  done_reducing:
 161
 162   /* Free the global sets used to compute the reduced grammar */
 163
 164   FREE(N);
 165   FREE(V);
 166   FREE(P);
 167
 168 }
 169 \f
 170 /*
 171  * Another way to do this would be with a set for each production and then do
 172  * subset tests against N0, but even for the C grammar the whole reducing
 173  * process takes only 2 seconds on my 8Mhz AT.
 174  */
 175
 176 static bool
 177 useful_production (int i, BSet N0)
 178 {
 179   rule  r;
 180   short n;
 181
 182   /*
 183    * A production is useful if all of the nonterminals in its RHS
 184    * appear in the set of useful nonterminals.
 185    */
 186
 187   for (r = &ritem[rrhs[i]]; *r > 0; r++)
 188     if (ISVAR(n = *r))
 189       if (!BITISSET(N0, n - ntokens))
 190         return FALSE;
 191   return TRUE;
 192 }
 193
 194
 195 /* Remember that rules are 1-origin, symbols are 0-origin. */
 196
 197 static void
 198 useless_nonterminals (void)
 199 {
 200   BSet Np, Ns;
 201   int  i, n;
 202
 203   /*
 204    * N is set as built.  Np is set being built this iteration. P is set
 205    * of all productions which have a RHS all in N.
 206    */
 207
 208   Np = NEW2(WORDSIZE(nvars), unsigned);
 209
 210   /*
 211    * The set being computed is a set of nonterminals which can derive
 212    * the empty string or strings consisting of all terminals. At each
 213    * iteration a nonterminal is added to the set if there is a
 214    * production with that nonterminal as its LHS for which all the
 215    * nonterminals in its RHS are already in the set.  Iterate until the
 216    * set being computed remains unchanged.  Any nonterminals not in the
 217    * set at that point are useless in that they will never be used in
 218    * deriving a sentence of the language.
 219    *
 220    * This iteration doesn't use any special traversal over the
 221    * productions.  A set is kept of all productions for which all the
 222    * nonterminals in the RHS are in useful.  Only productions not in
 223    * this set are scanned on each iteration.  At the end, this set is
 224    * saved to be used when finding useful productions: only productions
 225    * in this set will appear in the final grammar.
 226    */
 227
 228   n = 0;
 229   while (1)
 230     {
 231       for (i = WORDSIZE(nvars) - 1; i >= 0; i--)
 232         Np[i] = N[i];
 233       for (i = 1; i <= nrules; i++)
 234         {
 235           if (!BITISSET(P, i))
 236             {
 237               if (useful_production(i, N))
 238                 {
 239                   SETBIT(Np, rlhs[i] - ntokens);
 240                   SETBIT(P, i);
 241                 }
 242             }
 243         }
 244       if (bits_equal(N, Np, WORDSIZE(nvars)))
 245         break;
 246       Ns = Np;
 247       Np = N;
 248       N = Ns;
 249     }
 250   FREE(N);
 251   N = Np;
 252 }
 253 \f
 254 static void
 255 inaccessable_symbols (void)
 256 {
 257   BSet  Vp, Vs, Pp;
 258   int   i, n;
 259   short t;
 260   rule  r;
 261
 262   /*
 263    * Find out which productions are reachable and which symbols are
 264    * used.  Starting with an empty set of productions and a set of
 265    * symbols which only has the start symbol in it, iterate over all
 266    * productions until the set of productions remains unchanged for an
 267    * iteration.  For each production which has a LHS in the set of
 268    * reachable symbols, add the production to the set of reachable
 269    * productions, and add all of the nonterminals in the RHS of the
 270    * production to the set of reachable symbols.
 271    *
 272    * Consider only the (partially) reduced grammar which has only
 273    * nonterminals in N and productions in P.
 274    *
 275    * The result is the set P of productions in the reduced grammar, and
 276    * the set V of symbols in the reduced grammar.
 277    *
 278    * Although this algorithm also computes the set of terminals which are
 279    * reachable, no terminal will be deleted from the grammar. Some
 280    * terminals might not be in the grammar but might be generated by
 281    * semantic routines, and so the user might want them available with
 282    * specified numbers.  (Is this true?)  However, the nonreachable
 283    * terminals are printed (if running in verbose mode) so that the user
 284    * can know.
 285    */
 286
 287   Vp = NEW2(WORDSIZE(nsyms), unsigned);
 288   Pp = NEW2(WORDSIZE(nrules + 1), unsigned);
 289
 290   /* If the start symbol isn't useful, then nothing will be useful. */
 291   if (!BITISSET(N, start_symbol - ntokens))
 292     goto end_iteration;
 293
 294   SETBIT(V, start_symbol);
 295
 296   n = 0;
 297   while (1)
 298     {
 299       for (i = WORDSIZE(nsyms) - 1; i >= 0; i--)
 300         Vp[i] = V[i];
 301       for (i = 1; i <= nrules; i++)
 302         {
 303           if (!BITISSET(Pp, i) && BITISSET(P, i) &&
 304               BITISSET(V, rlhs[i]))
 305             {
 306               for (r = &ritem[rrhs[i]]; *r >= 0; r++)
 307                 {
 308                   if (ISTOKEN(t = *r)
 309                       || BITISSET(N, t - ntokens))
 310                     {
 311                       SETBIT(Vp, t);
 312                     }
 313                 }
 314               SETBIT(Pp, i);
 315             }
 316         }
 317       if (bits_equal(V, Vp, WORDSIZE(nsyms)))
 318         {
 319           break;
 320         }
 321       Vs = Vp;
 322       Vp = V;
 323       V = Vs;
 324     }
 325  end_iteration:
 326
 327   FREE(V);
 328   V = Vp;
 329
 330   /* Tokens 0, 1, and 2 are internal to Bison.  Consider them useful. */
 331   SETBIT(V, 0);                 /* end-of-input token */
 332   SETBIT(V, 1);                 /* error token */
 333   SETBIT(V, 2);                 /* some undefined token */
 334
 335   FREE(P);
 336   P = Pp;
 337
 338   nuseful_productions = bits_size(P, WORDSIZE(nrules + 1));
 339   nuseless_productions = nrules - nuseful_productions;
 340
 341   nuseful_nonterminals = 0;
 342   for (i = ntokens; i < nsyms; i++)
 343     if (BITISSET(V, i))
 344       nuseful_nonterminals++;
 345   nuseless_nonterminals = nvars - nuseful_nonterminals;
 346
 347   /* A token that was used in %prec should not be warned about.  */
 348   for (i = 1; i < nrules; i++)
 349     if (rprecsym[i] != 0)
 350       SETBIT(V1, rprecsym[i]);
 351 }
 352 \f
 353 static void
 354 reduce_grammar_tables (void)
 355 {
 356 /* This is turned off because we would need to change the numbers
 357    in the case statements in the actions file.  */
 358 #if 0
 359   /* remove useless productions */
 360   if (nuseless_productions > 0)
 361     {
 362       short np, pn, ni, pi;
 363
 364       np = 0;
 365       ni = 0;
 366       for (pn = 1; pn <= nrules; pn++)
 367         {
 368           if (BITISSET(P, pn))
 369             {
 370               np++;
 371               if (pn != np)
 372                 {
 373                   rlhs[np] = rlhs[pn];
 374                   rline[np] = rline[pn];
 375                   rprec[np] = rprec[pn];
 376                   rassoc[np] = rassoc[pn];
 377                   rrhs[np] = rrhs[pn];
 378                   if (rrhs[np] != ni)
 379                     {
 380                       pi = rrhs[np];
 381                       rrhs[np] = ni;
 382                       while (ritem[pi] >= 0)
 383                         ritem[ni++] = ritem[pi++];
 384                       ritem[ni++] = -np;
 385                     }
 386                 } else {
 387                   while (ritem[ni++] >= 0);
 388                 }
 389             }
 390         }
 391       ritem[ni] = 0;
 392       nrules -= nuseless_productions;
 393       nitems = ni;
 394
 395       /*
 396        * Is it worth it to reduce the amount of memory for the
 397        * grammar? Probably not.
 398        */
 399
 400     }
 401 #endif /* 0 */
 402   /* Disable useless productions,
 403      since they may contain useless nonterms
 404      that would get mapped below to -1 and confuse everyone.  */
 405   if (nuseless_productions > 0)
 406     {
 407       int pn;
 408
 409       for (pn = 1; pn <= nrules; pn++)
 410         {
 411           if (!BITISSET(P, pn))
 412             {
 413               rlhs[pn] = -1;
 414             }
 415         }
 416     }
 417
 418   /* remove useless symbols */
 419   if (nuseless_nonterminals > 0)
 420     {
 421
 422       int    i, n;
 423 /*      short  j; JF unused */
 424       short *nontermmap;
 425       rule   r;
 426
 427       /*
 428        * create a map of nonterminal number to new nonterminal
 429        * number. -1 in the map means it was useless and is being
 430        * eliminated.
 431        */
 432
 433       nontermmap = NEW2(nvars, short) - ntokens;
 434       for (i = ntokens; i < nsyms; i++)
 435         nontermmap[i] = -1;
 436
 437       n = ntokens;
 438       for (i = ntokens; i < nsyms; i++)
 439         if (BITISSET(V, i))
 440           nontermmap[i] = n++;
 441
 442       /* Shuffle elements of tables indexed by symbol number.  */
 443
 444       for (i = ntokens; i < nsyms; i++)
 445         {
 446           n = nontermmap[i];
 447           if (n >= 0)
 448             {
 449               sassoc[n] = sassoc[i];
 450               sprec[n] = sprec[i];
 451               tags[n] = tags[i];
 452             } else {
 453               free(tags[i]);
 454             }
 455         }
 456
 457       /* Replace all symbol numbers in valid data structures.  */
 458
 459       for (i = 1; i <= nrules; i++)
 460         {
 461           /* Ignore the rules disabled above.  */
 462           if (rlhs[i] >= 0)
 463             rlhs[i] = nontermmap[rlhs[i]];
 464           if (ISVAR (rprecsym[i]))
 465             /* Can this happen?  */
 466             rprecsym[i] = nontermmap[rprecsym[i]];
 467         }
 468
 469       for (r = ritem; *r; r++)
 470         if (ISVAR(*r))
 471           *r = nontermmap[*r];
 472
 473       start_symbol = nontermmap[start_symbol];
 474
 475       nsyms -= nuseless_nonterminals;
 476       nvars -= nuseless_nonterminals;
 477
 478       free(&nontermmap[ntokens]);
 479     }
 480 }
 481 \f
 482 static void
 483 print_results (void)
 484 {
 485   int   i;
 486 /*  short j; JF unused */
 487   rule  r;
 488   bool  b;
 489
 490   if (nuseless_nonterminals > 0)
 491     {
 492       fprintf(foutput, _("Useless nonterminals:\n\n"));
 493       for (i = ntokens; i < nsyms; i++)
 494         if (!BITISSET(V, i))
 495           fprintf(foutput, "   %s\n", tags[i]);
 496     }
 497   b = FALSE;
 498   for (i = 0; i < ntokens; i++)
 499     {
 500       if (!BITISSET(V, i) && !BITISSET(V1, i))
 501         {
 502           if (!b)
 503             {
 504               fprintf(foutput, _("\n\nTerminals which are not used:\n\n"));
 505               b = TRUE;
 506             }
 507           fprintf(foutput, "   %s\n", tags[i]);
 508         }
 509     }
 510
 511   if (nuseless_productions > 0)
 512     {
 513       fprintf(foutput, _("\n\nUseless rules:\n\n"));
 514       for (i = 1; i <= nrules; i++)
 515         {
 516           if (!BITISSET(P, i))
 517             {
 518               fprintf(foutput, "#%-4d  ", i);
 519               fprintf(foutput, "%s :\t", tags[rlhs[i]]);
 520               for (r = &ritem[rrhs[i]]; *r >= 0; r++)
 521                 {
 522                   fprintf(foutput, " %s", tags[*r]);
 523                 }
 524               fprintf(foutput, ";\n");
 525             }
 526         }
 527     }
 528   if (nuseless_nonterminals > 0 || nuseless_productions > 0 || b)
 529     fprintf(foutput, "\n\n");
 530 }
 531 \f
 532 void
 533 dump_grammar (void)
 534 {
 535   int i;
 536   rule r;
 537
 538   fprintf(foutput,
 539           "ntokens = %d, nvars = %d, nsyms = %d, nrules = %d, nitems = %d\n\n",
 540           ntokens, nvars, nsyms, nrules, nitems);
 541   fprintf(foutput, _("Variables\n---------\n\n"));
 542   fprintf(foutput, _("Value  Sprec    Sassoc    Tag\n"));
 543   for (i = ntokens; i < nsyms; i++)
 544     fprintf(foutput, "%5d  %5d  %5d  %s\n",
 545             i, sprec[i], sassoc[i], tags[i]);
 546   fprintf(foutput, "\n\n");
 547   fprintf(foutput, _("Rules\n-----\n\n"));
 548   for (i = 1; i <= nrules; i++)
 549     {
 550       fprintf(foutput, "%-5d(%5d%5d)%5d : (@%-5d)",
 551               i, rprec[i], rassoc[i], rlhs[i], rrhs[i]);
 552       for (r = &ritem[rrhs[i]]; *r > 0; r++)
 553         fprintf(foutput, "%5d", *r);
 554       fprintf(foutput, " [%d]\n", -(*r));
 555     }
 556   fprintf(foutput, "\n\n");
 557   fprintf(foutput, _("Rules interpreted\n-----------------\n\n"));
 558   for (i = 1; i <= nrules; i++)
 559     {
 560       fprintf(foutput, "%-5d  %s :", i, tags[rlhs[i]]);
 561       for (r = &ritem[rrhs[i]]; *r > 0; r++)
 562         fprintf(foutput, " %s", tags[*r]);
 563       fprintf(foutput, "\n");
 564     }
 565   fprintf(foutput, "\n\n");
 566 }
 567
 568
 569 static void
 570 print_notices (void)
 571 {
 572   if (fixed_outfiles && nuseless_productions)
 573     fprintf(stderr, _("%d rules never reduced\n"), nuseless_productions);
 574
 575   fprintf(stderr, _("%s contains "), infile);
 576
 577   if (nuseless_nonterminals > 0)
 578     {
 579       fprintf(stderr, _("%d useless nonterminal%s"),
 580               nuseless_nonterminals,
 581               (nuseless_nonterminals == 1 ? "" : "s"));
 582     }
 583   if (nuseless_nonterminals > 0 && nuseless_productions > 0)
 584     fprintf(stderr, _(" and "));
 585
 586   if (nuseless_productions > 0)
 587     {
 588       fprintf(stderr, _("%d useless rule%s"),
 589               nuseless_productions,
 590               (nuseless_productions == 1 ? "" : "s"));
 591     }
 592   fprintf(stderr, "\n");
 593   fflush(stderr);
 594 }