src/lalr.c

   1 /* Compute look-ahead criteria for bison,
   2    Copyright 1984, 1986, 1989, 2000, 2001  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 /* Compute how to make the finite state machine deterministic; find
  23    which rules need lookahead in each state, and which lookahead
  24    tokens they accept.  */
  25
  26 #include "system.h"
  27 #include "types.h"
  28 #include "LR0.h"
  29 #include "gram.h"
  30 #include "complain.h"
  31 #include "lalr.h"
  32 #include "nullable.h"
  33 #include "derives.h"
  34 #include "getargs.h"
  35
  36 /* All the decorated states, indexed by the state number.  Warning:
  37    there is a state_TABLE in LR0.c, but it is different and static.
  38    */
  39 state_t *state_table = NULL;
  40
  41 int tokensetsize;
  42 short *LAruleno;
  43 unsigned *LA;
  44
  45 static int ngotos;
  46 short *goto_map;
  47 short *from_state;
  48 short *to_state;
  49
  50 /* And for the famous F variable, which name is so descriptive that a
  51    comment is hardly needed.  <grin>.  */
  52 static unsigned *F = NULL;
  53 #define F(Rule)  (F + (Rule) * tokensetsize)
  54
  55 static short **includes;
  56 static shorts **lookback;
  57
  58
  59 /*---------------------------------------------------------------.
  60 | digraph & traverse.                                            |
  61 |                                                                |
  62 | The following variables are used as common storage between the |
  63 | two.                                                           |
  64 `---------------------------------------------------------------*/
  65
  66 static short **R;
  67 static short *INDEX;
  68 static short *VERTICES;
  69 static int top;
  70 static int infinity;
  71
  72 static void
  73 traverse (int i)
  74 {
  75   int j;
  76   size_t k;
  77   int height;
  78   size_t size = F (i + 1) - F(i);
  79
  80   VERTICES[++top] = i;
  81   INDEX[i] = height = top;
  82
  83   if (R[i])
  84     for (j = 0; R[i][j] >= 0; ++j)
  85       {
  86         if (INDEX[R[i][j]] == 0)
  87           traverse (R[i][j]);
  88
  89         if (INDEX[i] > INDEX[R[i][j]])
  90           INDEX[i] = INDEX[R[i][j]];
  91
  92         for (k = 0; k < size; ++k)
  93           F (i)[k] |= F (R[i][j])[k];
  94       }
  95
  96   if (INDEX[i] == height)
  97     for (;;)
  98       {
  99         j = VERTICES[top--];
 100         INDEX[j] = infinity;
 101
 102         if (i == j)
 103           break;
 104
 105         for (k = 0; k < size; ++k)
 106           F (j)[k] = F (i)[k];
 107       }
 108 }
 109
 110
 111 static void
 112 digraph (short **relation)
 113 {
 114   int i;
 115
 116   infinity = ngotos + 2;
 117   INDEX = XCALLOC (short, ngotos + 1);
 118   VERTICES = XCALLOC (short, ngotos + 1);
 119   top = 0;
 120
 121   R = relation;
 122
 123   for (i = 0; i < ngotos; i++)
 124     INDEX[i] = 0;
 125
 126   for (i = 0; i < ngotos; i++)
 127     if (INDEX[i] == 0 && R[i])
 128       traverse (i);
 129
 130   XFREE (INDEX);
 131   XFREE (VERTICES);
 132 }
 133
 134
 135 /*--------------------.
 136 | Build STATE_TABLE.  |
 137 `--------------------*/
 138
 139 static void
 140 set_state_table (void)
 141 {
 142   /* NSTATES + 1 because lookahead for the pseudo state number NSTATES
 143      might be used (see conflicts.c).  It is too opaque for me to
 144      provide a probably less hacky implementation. --akim */
 145   state_table = XCALLOC (state_t, nstates + 1);
 146
 147   {
 148     core *sp;
 149     for (sp = first_state; sp; sp = sp->next)
 150       {
 151         state_table[sp->number].state = sp;
 152         state_table[sp->number].accessing_symbol = sp->accessing_symbol;
 153       }
 154   }
 155
 156   {
 157     shifts *sp;
 158     for (sp = first_shift; sp; sp = sp->next)
 159       state_table[sp->number].shifts = sp;
 160   }
 161
 162   {
 163     reductions *rp;
 164     for (rp = first_reduction; rp; rp = rp->next)
 165       state_table[rp->number].reductions = rp;
 166   }
 167
 168   /* Pessimization, but simplification of the code: make sense all the
 169      states have a shifts, even if reduced to 0 shifts.  */
 170   {
 171     int i;
 172     for (i = 0; i < nstates; i++)
 173       if (!state_table[i].shifts)
 174         state_table[i].shifts = shifts_new (0);
 175   }
 176
 177   /* Initializing the lookaheads members.  Please note that it must be
 178      performed after having set some of the other members which are
 179      used below.  Change with extreme caution.  */
 180   {
 181     int i;
 182     int count = 0;
 183     for (i = 0; i < nstates; i++)
 184       {
 185         int k;
 186         reductions *rp = state_table[i].reductions;
 187         shifts *sp = state_table[i].shifts;
 188
 189         state_table[i].lookaheads = count;
 190
 191         if (rp
 192             && (rp->nreds > 1 || (sp->nshifts && SHIFT_IS_SHIFT (sp, 0))))
 193           count += rp->nreds;
 194         else
 195           state_table[i].consistent = 1;
 196
 197         for (k = 0; k < sp->nshifts; k++)
 198           if (SHIFT_IS_ERROR (sp, k))
 199             {
 200               state_table[i].consistent = 0;
 201               break;
 202             }
 203       }
 204      state_table[nstates].lookaheads = count;
 205   }
 206 }
 207
 208
 209 static void
 210 initialize_LA (void)
 211 {
 212   int i;
 213   int j;
 214   short *np;
 215   reductions *rp;
 216
 217   size_t nLA = state_table[nstates].lookaheads;
 218   if (!nLA)
 219     nLA = 1;
 220
 221   LA = XCALLOC (unsigned, nLA * tokensetsize);
 222   LAruleno = XCALLOC (short, nLA);
 223   lookback = XCALLOC (shorts *, nLA);
 224
 225   np = LAruleno;
 226   for (i = 0; i < nstates; i++)
 227     if (!state_table[i].consistent)
 228       if ((rp = state_table[i].reductions))
 229         for (j = 0; j < rp->nreds; j++)
 230           *np++ = rp->rules[j];
 231 }
 232
 233
 234 static void
 235 set_goto_map (void)
 236 {
 237   shifts *sp;
 238   int i;
 239   int symbol;
 240   int k;
 241   short *temp_map;
 242   int state2;
 243   int state1;
 244
 245   goto_map = XCALLOC (short, nvars + 1) - ntokens;
 246   temp_map = XCALLOC (short, nvars + 1) - ntokens;
 247
 248   ngotos = 0;
 249   for (sp = first_shift; sp; sp = sp->next)
 250     if (sp->nshifts)
 251       for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
 252         {
 253           symbol = state_table[sp->shifts[i]].accessing_symbol;
 254
 255           if (ngotos == MAXSHORT)
 256             fatal (_("too many gotos (max %d)"), MAXSHORT);
 257
 258           ngotos++;
 259           goto_map[symbol]++;
 260         }
 261
 262   k = 0;
 263   for (i = ntokens; i < nsyms; i++)
 264     {
 265       temp_map[i] = k;
 266       k += goto_map[i];
 267     }
 268
 269   for (i = ntokens; i < nsyms; i++)
 270     goto_map[i] = temp_map[i];
 271
 272   goto_map[nsyms] = ngotos;
 273   temp_map[nsyms] = ngotos;
 274
 275   from_state = XCALLOC (short, ngotos);
 276   to_state = XCALLOC (short, ngotos);
 277
 278   for (sp = first_shift; sp; sp = sp->next)
 279     {
 280       state1 = sp->number;
 281       if (sp->nshifts)
 282         for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
 283           {
 284             state2 = sp->shifts[i];
 285             symbol = state_table[state2].accessing_symbol;
 286
 287             k = temp_map[symbol]++;
 288             from_state[k] = state1;
 289             to_state[k] = state2;
 290           }
 291     }
 292
 293   XFREE (temp_map + ntokens);
 294 }
 295
 296
 297
 298 /*----------------------------------------------------------.
 299 | Map a state/symbol pair into its numeric representation.  |
 300 `----------------------------------------------------------*/
 301
 302 static int
 303 map_goto (int state, int symbol)
 304 {
 305   int high;
 306   int low;
 307   int middle;
 308   int s;
 309
 310   low = goto_map[symbol];
 311   high = goto_map[symbol + 1] - 1;
 312
 313   while (low <= high)
 314     {
 315       middle = (low + high) / 2;
 316       s = from_state[middle];
 317       if (s == state)
 318         return middle;
 319       else if (s < state)
 320         low = middle + 1;
 321       else
 322         high = middle - 1;
 323     }
 324
 325   assert (0);
 326   /* NOTREACHED */
 327   return 0;
 328 }
 329
 330
 331 static void
 332 initialize_F (void)
 333 {
 334   short **reads = XCALLOC (short *, ngotos);
 335   short *edge = XCALLOC (short, ngotos + 1);
 336   int nedges = 0;
 337
 338   int i;
 339
 340   F = XCALLOC (unsigned, ngotos * tokensetsize);
 341
 342   for (i = 0; i < ngotos; i++)
 343     {
 344       int stateno = to_state[i];
 345       shifts *sp = state_table[stateno].shifts;
 346
 347       int j;
 348       for (j = 0; j < sp->nshifts && SHIFT_IS_SHIFT (sp, j); j++)
 349         {
 350           int symbol = state_table[sp->shifts[j]].accessing_symbol;
 351           SETBIT (F (i), symbol);
 352         }
 353
 354       for (; j < sp->nshifts; j++)
 355         {
 356           int symbol = state_table[sp->shifts[j]].accessing_symbol;
 357           if (nullable[symbol])
 358             edge[nedges++] = map_goto (stateno, symbol);
 359         }
 360
 361       if (nedges)
 362         {
 363           reads[i] = XCALLOC (short, nedges + 1);
 364           shortcpy (reads[i], edge, nedges);
 365           reads[i][nedges] = -1;
 366           nedges = 0;
 367         }
 368     }
 369
 370   digraph (reads);
 371
 372   for (i = 0; i < ngotos; i++)
 373     XFREE (reads[i]);
 374
 375   XFREE (reads);
 376   XFREE (edge);
 377 }
 378
 379
 380 static void
 381 add_lookback_edge (int stateno, int ruleno, int gotono)
 382 {
 383   int i;
 384   int k;
 385   int found;
 386   shorts *sp;
 387
 388   i = state_table[stateno].lookaheads;
 389   k = state_table[stateno + 1].lookaheads;
 390   found = 0;
 391   while (!found && i < k)
 392     {
 393       if (LAruleno[i] == ruleno)
 394         found = 1;
 395       else
 396         i++;
 397     }
 398
 399   assert (found);
 400
 401   sp = XCALLOC (shorts, 1);
 402   sp->next = lookback[i];
 403   sp->value = gotono;
 404   lookback[i] = sp;
 405 }
 406
 407
 408 static void
 409 matrix_print (FILE *out, short **matrix, int n)
 410 {
 411   int i, j;
 412
 413   for (i = 0; i < n; ++i)
 414     {
 415       fprintf (out, "%3d: ", i);
 416       if (matrix[i])
 417         for (j = 0; matrix[i][j] != -1; ++j)
 418           fprintf (out, "%3d ", matrix[i][j]);
 419       fputc ('\n', out);
 420     }
 421   fputc ('\n', out);
 422 }
 423
 424 /*-------------------------------------------------------------------.
 425 | Return the transpose of R_ARG, of size N.  Destroy R_ARG, as it is |
 426 | replaced with the result.                                          |
 427 |                                                                    |
 428 | R_ARG[I] is NULL or a -1 terminated list of numbers.               |
 429 |                                                                    |
 430 | RESULT[NUM] is NULL or the -1 terminated list of the I such as NUM |
 431 | is in R_ARG[I].                                                    |
 432 `-------------------------------------------------------------------*/
 433
 434 static short **
 435 transpose (short **R_arg, int n)
 436 {
 437   /* The result. */
 438   short **new_R = XCALLOC (short *, n);
 439   /* END_R[I] -- next entry of NEW_R[I]. */
 440   short **end_R = XCALLOC (short *, n);
 441   /* NEDGES[I] -- total size of NEW_R[I]. */
 442   short *nedges = XCALLOC (short, n);
 443   int i, j;
 444
 445   if (trace_flag)
 446     {
 447       fputs ("transpose: input\n", stderr);
 448       matrix_print (stderr, R_arg, n);
 449     }
 450
 451   /* Count. */
 452   for (i = 0; i < n; i++)
 453     if (R_arg[i])
 454       for (j = 0; R_arg[i][j] >= 0; ++j)
 455         ++nedges[R_arg[i][j]];
 456
 457   /* Allocate. */
 458   for (i = 0; i < n; i++)
 459     if (nedges[i] > 0)
 460       {
 461         short *sp = XCALLOC (short, nedges[i] + 1);
 462         sp[nedges[i]] = -1;
 463         new_R[i] = sp;
 464         end_R[i] = sp;
 465       }
 466
 467   /* Store. */
 468   for (i = 0; i < n; i++)
 469     if (R_arg[i])
 470       for (j = 0; R_arg[i][j] >= 0; ++j)
 471         {
 472           *end_R[R_arg[i][j]] = i;
 473           ++end_R[R_arg[i][j]];
 474         }
 475
 476   free (nedges);
 477   free (end_R);
 478
 479   /* Free the input: it is replaced with the result. */
 480   for (i = 0; i < n; i++)
 481     XFREE (R_arg[i]);
 482   free (R_arg);
 483
 484   if (trace_flag)
 485     {
 486       fputs ("transpose: output\n", stderr);
 487       matrix_print (stderr, new_R, n);
 488     }
 489
 490   return new_R;
 491 }
 492
 493
 494 static void
 495 build_relations (void)
 496 {
 497   short *edge = XCALLOC (short, ngotos + 1);
 498   short *states = XCALLOC (short, ritem_longest_rhs () + 1);
 499   int i;
 500
 501   includes = XCALLOC (short *, ngotos);
 502
 503   for (i = 0; i < ngotos; i++)
 504     {
 505       int nedges = 0;
 506       int state1 = from_state[i];
 507       int symbol1 = state_table[to_state[i]].accessing_symbol;
 508       short *rulep;
 509
 510       for (rulep = derives[symbol1]; *rulep > 0; rulep++)
 511         {
 512           int done;
 513           int length = 1;
 514           int stateno = state1;
 515           short *rp;
 516           states[0] = state1;
 517
 518           for (rp = ritem + rule_table[*rulep].rhs; *rp > 0; rp++)
 519             {
 520               shifts *sp = state_table[stateno].shifts;
 521               int j;
 522               for (j = 0; j < sp->nshifts; j++)
 523                 {
 524                   stateno = sp->shifts[j];
 525                   if (state_table[stateno].accessing_symbol == *rp)
 526                     break;
 527                 }
 528
 529               states[length++] = stateno;
 530             }
 531
 532           if (!state_table[stateno].consistent)
 533             add_lookback_edge (stateno, *rulep, i);
 534
 535           length--;
 536           done = 0;
 537           while (!done)
 538             {
 539               done = 1;
 540               rp--;
 541               /* JF added rp>=ritem &&   I hope to god its right! */
 542               if (rp >= ritem && ISVAR (*rp))
 543                 {
 544                   stateno = states[--length];
 545                   edge[nedges++] = map_goto (stateno, *rp);
 546                   if (nullable[*rp])
 547                     done = 0;
 548                 }
 549             }
 550         }
 551
 552       if (nedges)
 553         {
 554           int j;
 555           includes[i] = XCALLOC (short, nedges + 1);
 556           for (j = 0; j < nedges; j++)
 557             includes[i][j] = edge[j];
 558           includes[i][nedges] = -1;
 559         }
 560     }
 561
 562   XFREE (edge);
 563   XFREE (states);
 564
 565   includes = transpose (includes, ngotos);
 566 }
 567
 568
 569
 570 static void
 571 compute_FOLLOWS (void)
 572 {
 573   int i;
 574
 575   digraph (includes);
 576
 577   for (i = 0; i < ngotos; i++)
 578     XFREE (includes[i]);
 579
 580   XFREE (includes);
 581 }
 582
 583
 584 static void
 585 compute_lookaheads (void)
 586 {
 587   int i;
 588   shorts *sp;
 589
 590   for (i = 0; i < state_table[nstates].lookaheads; i++)
 591     for (sp = lookback[i]; sp; sp = sp->next)
 592       {
 593         int size = LA (i + 1) - LA (i);
 594         int j;
 595         for (j = 0; j < size; ++j)
 596           LA (i)[j] |= F (sp->value)[j];
 597       }
 598
 599   /* Free LOOKBACK. */
 600   for (i = 0; i < state_table[nstates].lookaheads; i++)
 601     LIST_FREE (shorts, lookback[i]);
 602
 603   XFREE (lookback);
 604   XFREE (F);
 605 }
 606
 607
 608 void
 609 lalr (void)
 610 {
 611   tokensetsize = WORDSIZE (ntokens);
 612
 613   set_state_table ();
 614   initialize_LA ();
 615   set_goto_map ();
 616   initialize_F ();
 617   build_relations ();
 618   compute_FOLLOWS ();
 619   compute_lookaheads ();
 620 }