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 (i)[k] = F (j)[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].shift_table = sp;
 160   }
 161
 162   {
 163     reductions *rp;
 164     for (rp = first_reduction; rp; rp = rp->next)
 165       state_table[rp->number].reduction_table = rp;
 166   }
 167
 168   /* Initializing the lookaheads members.  Please note that it must be
 169      performed after having set some of the other members which are
 170      used below.  Change with extreme caution.  */
 171   {
 172     int i;
 173     int count = 0;
 174     for (i = 0; i < nstates; i++)
 175       {
 176         int k;
 177         reductions *rp = state_table[i].reduction_table;
 178         shifts *sp = state_table[i].shift_table;
 179
 180         state_table[i].lookaheads = count;
 181
 182         if (rp
 183             && (rp->nreds > 1 || (sp && SHIFT_IS_SHIFT (sp, 0))))
 184           count += rp->nreds;
 185         else
 186           state_table[i].consistent = 1;
 187
 188         if (sp)
 189           for (k = 0; k < sp->nshifts; k++)
 190             if (SHIFT_IS_ERROR (sp, k))
 191               {
 192                 state_table[i].consistent = 0;
 193                 break;
 194               }
 195       }
 196      state_table[nstates].lookaheads = count;
 197   }
 198 }
 199
 200
 201 static void
 202 initialize_LA (void)
 203 {
 204   int i;
 205   int j;
 206   short *np;
 207   reductions *rp;
 208
 209   size_t nLA = state_table[nstates].lookaheads;
 210   if (!nLA)
 211     nLA = 1;
 212
 213   LA = XCALLOC (unsigned, nLA * tokensetsize);
 214   LAruleno = XCALLOC (short, nLA);
 215   lookback = XCALLOC (shorts *, nLA);
 216
 217   np = LAruleno;
 218   for (i = 0; i < nstates; i++)
 219     if (!state_table[i].consistent)
 220       if ((rp = state_table[i].reduction_table))
 221         for (j = 0; j < rp->nreds; j++)
 222           *np++ = rp->rules[j];
 223 }
 224
 225
 226 static void
 227 set_goto_map (void)
 228 {
 229   shifts *sp;
 230   int i;
 231   int symbol;
 232   int k;
 233   short *temp_map;
 234   int state2;
 235   int state1;
 236
 237   goto_map = XCALLOC (short, nvars + 1) - ntokens;
 238   temp_map = XCALLOC (short, nvars + 1) - ntokens;
 239
 240   ngotos = 0;
 241   for (sp = first_shift; sp; sp = sp->next)
 242     {
 243       for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
 244         {
 245           symbol = state_table[sp->shifts[i]].accessing_symbol;
 246
 247           if (ngotos == MAXSHORT)
 248             fatal (_("too many gotos (max %d)"), MAXSHORT);
 249
 250           ngotos++;
 251           goto_map[symbol]++;
 252         }
 253     }
 254
 255   k = 0;
 256   for (i = ntokens; i < nsyms; i++)
 257     {
 258       temp_map[i] = k;
 259       k += goto_map[i];
 260     }
 261
 262   for (i = ntokens; i < nsyms; i++)
 263     goto_map[i] = temp_map[i];
 264
 265   goto_map[nsyms] = ngotos;
 266   temp_map[nsyms] = ngotos;
 267
 268   from_state = XCALLOC (short, ngotos);
 269   to_state = XCALLOC (short, ngotos);
 270
 271   for (sp = first_shift; sp; sp = sp->next)
 272     {
 273       state1 = sp->number;
 274       for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
 275         {
 276           state2 = sp->shifts[i];
 277           symbol = state_table[state2].accessing_symbol;
 278
 279           k = temp_map[symbol]++;
 280           from_state[k] = state1;
 281           to_state[k] = state2;
 282         }
 283     }
 284
 285   XFREE (temp_map + ntokens);
 286 }
 287
 288
 289
 290 /*----------------------------------------------------------.
 291 | Map a state/symbol pair into its numeric representation.  |
 292 `----------------------------------------------------------*/
 293
 294 static int
 295 map_goto (int state, int symbol)
 296 {
 297   int high;
 298   int low;
 299   int middle;
 300   int s;
 301
 302   low = goto_map[symbol];
 303   high = goto_map[symbol + 1] - 1;
 304
 305   while (low <= high)
 306     {
 307       middle = (low + high) / 2;
 308       s = from_state[middle];
 309       if (s == state)
 310         return middle;
 311       else if (s < state)
 312         low = middle + 1;
 313       else
 314         high = middle - 1;
 315     }
 316
 317   assert (0);
 318   /* NOTREACHED */
 319   return 0;
 320 }
 321
 322
 323 static void
 324 initialize_F (void)
 325 {
 326   short **reads = XCALLOC (short *, ngotos);
 327   short *edge = XCALLOC (short, ngotos + 1);
 328   int nedges = 0;
 329
 330   int i;
 331
 332   F = XCALLOC (unsigned, ngotos * tokensetsize);
 333
 334   for (i = 0; i < ngotos; i++)
 335     {
 336       int stateno = to_state[i];
 337       shifts *sp = state_table[stateno].shift_table;
 338
 339       if (sp)
 340         {
 341           int j;
 342           for (j = 0; j < sp->nshifts && SHIFT_IS_SHIFT (sp, j); j++)
 343             {
 344               int symbol = state_table[sp->shifts[j]].accessing_symbol;
 345               SETBIT (F + i * tokensetsize, symbol);
 346             }
 347
 348           for (; j < sp->nshifts; j++)
 349             {
 350               int symbol = state_table[sp->shifts[j]].accessing_symbol;
 351               if (nullable[symbol])
 352                 edge[nedges++] = map_goto (stateno, symbol);
 353             }
 354
 355           if (nedges)
 356             {
 357               reads[i] = XCALLOC (short, nedges + 1);
 358               shortcpy (reads[i], edge, nedges);
 359               reads[i][nedges] = -1;
 360               nedges = 0;
 361             }
 362         }
 363     }
 364
 365   digraph (reads);
 366
 367   for (i = 0; i < ngotos; i++)
 368     XFREE (reads[i]);
 369
 370   XFREE (reads);
 371   XFREE (edge);
 372 }
 373
 374
 375 static void
 376 add_lookback_edge (int stateno, int ruleno, int gotono)
 377 {
 378   int i;
 379   int k;
 380   int found;
 381   shorts *sp;
 382
 383   i = state_table[stateno].lookaheads;
 384   k = state_table[stateno + 1].lookaheads;
 385   found = 0;
 386   while (!found && i < k)
 387     {
 388       if (LAruleno[i] == ruleno)
 389         found = 1;
 390       else
 391         i++;
 392     }
 393
 394   assert (found);
 395
 396   sp = XCALLOC (shorts, 1);
 397   sp->next = lookback[i];
 398   sp->value = gotono;
 399   lookback[i] = sp;
 400 }
 401
 402
 403 static void
 404 matrix_print (FILE *out, short **matrix, int n)
 405 {
 406   int i, j;
 407
 408   for (i = 0; i < n; ++i)
 409     {
 410       fprintf (out, "%3d: ", i);
 411       if (matrix[i])
 412         for (j = 0; matrix[i][j] != -1; ++j)
 413           fprintf (out, "%3d ", matrix[i][j]);
 414       fputc ('\n', out);
 415     }
 416   fputc ('\n', out);
 417 }
 418
 419 /*-------------------------------------------------------------------.
 420 | Return the transpose of R_ARG, of size N.  Destroy R_ARG, as it is |
 421 | replaced with the result.                                          |
 422 |                                                                    |
 423 | R_ARG[I] is NULL or a -1 terminated list of numbers.               |
 424 |                                                                    |
 425 | RESULT[NUM] is NULL or the -1 terminated list of the I such as NUM |
 426 | is in R_ARG[I].                                                    |
 427 `-------------------------------------------------------------------*/
 428
 429 static short **
 430 transpose (short **R_arg, int n)
 431 {
 432   /* The result. */
 433   short **new_R = XCALLOC (short *, n);
 434   /* END_R[I] -- next entry of NEW_R[I]. */
 435   short **end_R = XCALLOC (short *, n);
 436   /* NEDGES[I] -- total size of NEW_R[I]. */
 437   short *nedges = XCALLOC (short, n);
 438   int i, j;
 439
 440   if (trace_flag)
 441     {
 442       fputs ("transpose: input\n", stderr);
 443       matrix_print (stderr, R_arg, n);
 444     }
 445
 446   /* Count. */
 447   for (i = 0; i < n; i++)
 448     if (R_arg[i])
 449       for (j = 0; R_arg[i][j] >= 0; ++j)
 450         ++nedges[R_arg[i][j]];
 451
 452   /* Allocate. */
 453   for (i = 0; i < n; i++)
 454     if (nedges[i] > 0)
 455       {
 456         short *sp = XCALLOC (short, nedges[i] + 1);
 457         sp[nedges[i]] = -1;
 458         new_R[i] = sp;
 459         end_R[i] = sp;
 460       }
 461
 462   /* Store. */
 463   for (i = 0; i < n; i++)
 464     if (R_arg[i])
 465       for (j = 0; R_arg[i][j] >= 0; ++j)
 466         {
 467           *end_R[R_arg[i][j]] = i;
 468           ++end_R[R_arg[i][j]];
 469         }
 470
 471   free (nedges);
 472   free (end_R);
 473
 474   /* Free the input: it is replaced with the result. */
 475   for (i = 0; i < n; i++)
 476     XFREE (R_arg[i]);
 477   free (R_arg);
 478
 479   if (trace_flag)
 480     {
 481       fputs ("transpose: output\n", stderr);
 482       matrix_print (stderr, new_R, n);
 483     }
 484
 485   return new_R;
 486 }
 487
 488
 489 static void
 490 build_relations (void)
 491 {
 492   short *edge = XCALLOC (short, ngotos + 1);
 493   short *states = XCALLOC (short, ritem_longest_rhs () + 1);
 494   int i;
 495
 496   includes = XCALLOC (short *, ngotos);
 497
 498   for (i = 0; i < ngotos; i++)
 499     {
 500       int nedges = 0;
 501       int state1 = from_state[i];
 502       int symbol1 = state_table[to_state[i]].accessing_symbol;
 503       short *rulep;
 504
 505       for (rulep = derives[symbol1]; *rulep > 0; rulep++)
 506         {
 507           int done;
 508           int length = 1;
 509           int stateno = state1;
 510           short *rp;
 511           states[0] = state1;
 512
 513           for (rp = ritem + rule_table[*rulep].rhs; *rp > 0; rp++)
 514             {
 515               shifts *sp = state_table[stateno].shift_table;
 516               int j;
 517               for (j = 0; j < sp->nshifts; j++)
 518                 {
 519                   stateno = sp->shifts[j];
 520                   if (state_table[stateno].accessing_symbol == *rp)
 521                     break;
 522                 }
 523
 524               states[length++] = stateno;
 525             }
 526
 527           if (!state_table[stateno].consistent)
 528             add_lookback_edge (stateno, *rulep, i);
 529
 530           length--;
 531           done = 0;
 532           while (!done)
 533             {
 534               done = 1;
 535               rp--;
 536               /* JF added rp>=ritem &&   I hope to god its right! */
 537               if (rp >= ritem && ISVAR (*rp))
 538                 {
 539                   stateno = states[--length];
 540                   edge[nedges++] = map_goto (stateno, *rp);
 541                   if (nullable[*rp])
 542                     done = 0;
 543                 }
 544             }
 545         }
 546
 547       if (nedges)
 548         {
 549           int j;
 550           includes[i] = XCALLOC (short, nedges + 1);
 551           for (j = 0; j < nedges; j++)
 552             includes[i][j] = edge[j];
 553           includes[i][nedges] = -1;
 554         }
 555     }
 556
 557   XFREE (edge);
 558   XFREE (states);
 559
 560   includes = transpose (includes, ngotos);
 561 }
 562
 563
 564
 565 static void
 566 compute_FOLLOWS (void)
 567 {
 568   int i;
 569
 570   digraph (includes);
 571
 572   for (i = 0; i < ngotos; i++)
 573     XFREE (includes[i]);
 574
 575   XFREE (includes);
 576 }
 577
 578
 579 static void
 580 compute_lookaheads (void)
 581 {
 582   int i;
 583   shorts *sp;
 584
 585   for (i = 0; i < state_table[nstates].lookaheads; i++)
 586     for (sp = lookback[i]; sp; sp = sp->next)
 587       {
 588         int size = LA (i + 1) - LA (i);
 589         int j;
 590         for (j = 0; j < size; ++j)
 591           LA (i)[j] |= F (sp->value)[j];
 592       }
 593
 594   /* Free LOOKBACK. */
 595   for (i = 0; i < state_table[nstates].lookaheads; i++)
 596     LIST_FREE (shorts, lookback[i]);
 597
 598   XFREE (lookback);
 599   XFREE (F);
 600 }
 601
 602
 603 void
 604 lalr (void)
 605 {
 606   tokensetsize = WORDSIZE (ntokens);
 607
 608   set_state_table ();
 609   initialize_LA ();
 610   set_goto_map ();
 611   initialize_F ();
 612   build_relations ();
 613   compute_FOLLOWS ();
 614   compute_lookaheads ();
 615 }