aboutsummaryrefslogtreecommitdiffstats
path: root/contrib/tools/bison/src/reduce.c
blob: eea11b1e20345158c318f2403b2a33f039ce20a1 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
/* Grammar reduction for Bison.

   Copyright (C) 1988-1989, 2000-2003, 2005-2013 Free Software
   Foundation, Inc.

   This file is part of Bison, the GNU Compiler Compiler.

   This program is free software: you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation, either version 3 of the License, or
   (at your option) any later version.

   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License
   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */


/* Reduce the grammar: Find and eliminate unreachable terminals,
   nonterminals, and productions.  David S. Bakin.  */

/* Don't eliminate unreachable terminals: They may be used by the
   user's parser.  */

#include <config.h>
#include "system.h"

#include <bitset.h>

#include "complain.h"
#include "files.h"
#include "getargs.h"
#include "gram.h"
#include "print-xml.h"
#include "reader.h"
#include "reduce.h"
#include "symtab.h"

/* Set of all nonterminals which are not useless.  */
static bitset N;

/* Set of all rules which have no useless nonterminals in their RHS.  */
static bitset P;

/* Set of all accessible symbols.  */
static bitset V;

/* Set of symbols used to define rule precedence (so they are
   'useless', but no warning should be issued).  */
static bitset V1;

static rule_number nuseful_productions;
rule_number nuseless_productions;
static int nuseful_nonterminals;
symbol_number nuseless_nonterminals;

/*-------------------------------------------------------------------.
| Another way to do this would be with a set for each production and |
| then do subset tests against N0, but even for the C grammar the    |
| whole reducing process takes only 2 seconds on my 8Mhz AT.         |
`-------------------------------------------------------------------*/

static bool
useful_production (rule_number r, bitset N0)
{
  item_number *rhsp;

  /* A production is useful if all of the nonterminals in its appear
     in the set of useful nonterminals.  */

  for (rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp)
    if (ISVAR (*rhsp) && !bitset_test (N0, *rhsp - ntokens))
      return false;
  return true;
}


/*---------------------------------------------------------.
| Remember that rules are 1-origin, symbols are 0-origin.  |
`---------------------------------------------------------*/

static void
useless_nonterminals (void)
{
  bitset Np, Ns;
  rule_number r;

  /* N is set as built.  Np is set being built this iteration. P is
     set of all productions which have a RHS all in N.  */

  Np = bitset_create (nvars, BITSET_FIXED);


  /* The set being computed is a set of nonterminals which can derive
     the empty string or strings consisting of all terminals. At each
     iteration a nonterminal is added to the set if there is a
     production with that nonterminal as its LHS for which all the
     nonterminals in its RHS are already in the set.  Iterate until
     the set being computed remains unchanged.  Any nonterminals not
     in the set at that point are useless in that they will never be
     used in deriving a sentence of the language.

     This iteration doesn't use any special traversal over the
     productions.  A set is kept of all productions for which all the
     nonterminals in the RHS are in useful.  Only productions not in
     this set are scanned on each iteration.  At the end, this set is
     saved to be used when finding useful productions: only
     productions in this set will appear in the final grammar.  */

  while (1)
    {
      bitset_copy (Np, N);
      for (r = 0; r < nrules; r++)
        if (!bitset_test (P, r)
            && useful_production (r, N))
          {
            bitset_set (Np, rules[r].lhs->number - ntokens);
            bitset_set (P, r);
          }
      if (bitset_equal_p (N, Np))
        break;
      Ns = Np;
      Np = N;
      N = Ns;
    }
  bitset_free (N);
  N = Np;
}


static void
inaccessable_symbols (void)
{
  bitset Vp, Vs, Pp;

  /* Find out which productions are reachable and which symbols are
     used.  Starting with an empty set of productions and a set of
     symbols which only has the start symbol in it, iterate over all
     productions until the set of productions remains unchanged for an
     iteration.  For each production which has a LHS in the set of
     reachable symbols, add the production to the set of reachable
     productions, and add all of the nonterminals in the RHS of the
     production to the set of reachable symbols.

     Consider only the (partially) reduced grammar which has only
     nonterminals in N and productions in P.

     The result is the set P of productions in the reduced grammar,
     and the set V of symbols in the reduced grammar.

     Although this algorithm also computes the set of terminals which
     are reachable, no terminal will be deleted from the grammar. Some
     terminals might not be in the grammar but might be generated by
     semantic routines, and so the user might want them available with
     specified numbers.  (Is this true?)  However, the nonreachable
     terminals are printed (if running in verbose mode) so that the
     user can know.  */

  Vp = bitset_create (nsyms, BITSET_FIXED);
  Pp = bitset_create (nrules, BITSET_FIXED);

  /* If the start symbol isn't useful, then nothing will be useful. */
  if (bitset_test (N, accept->number - ntokens))
    {
      bitset_set (V, accept->number);

      while (1)
        {
          rule_number r;
          bitset_copy (Vp, V);
          for (r = 0; r < nrules; r++)
            {
              if (!bitset_test (Pp, r)
                  && bitset_test (P, r)
                  && bitset_test (V, rules[r].lhs->number))
                {
                  item_number *rhsp;
                  for (rhsp = rules[r].rhs; *rhsp >= 0; rhsp++)
                    if (ISTOKEN (*rhsp) || bitset_test (N, *rhsp - ntokens))
                      bitset_set (Vp, *rhsp);
                  bitset_set (Pp, r);
                }
            }
          if (bitset_equal_p (V, Vp))
            break;
          Vs = Vp;
          Vp = V;
          V = Vs;
        }
    }

  bitset_free (V);
  V = Vp;

  /* Tokens 0, 1, and 2 are internal to Bison.  Consider them useful. */
  bitset_set (V, endtoken->number);             /* end-of-input token */
  bitset_set (V, errtoken->number);             /* error token */
  bitset_set (V, undeftoken->number);           /* some undefined token */

  bitset_free (P);
  P = Pp;

  nuseful_productions = bitset_count (P);
  nuseless_productions = nrules - nuseful_productions;

  nuseful_nonterminals = 0;
  {
    symbol_number i;
    for (i = ntokens; i < nsyms; i++)
      if (bitset_test (V, i))
        nuseful_nonterminals++;
  }
  nuseless_nonterminals = nvars - nuseful_nonterminals;

  /* A token that was used in %prec should not be warned about.  */
  {
    rule_number r;
    for (r = 0; r < nrules; ++r)
      if (rules[r].precsym != 0)
        bitset_set (V1, rules[r].precsym->number);
  }
}


/*-------------------------------------------------------------------.
| Put the useless productions at the end of RULES, and adjust NRULES |
| accordingly.                                                       |
`-------------------------------------------------------------------*/

static void
reduce_grammar_tables (void)
{
  /* Report and flag useless productions.  */
  {
    rule_number r;
    for (r = 0; r < nrules; r++)
      rules[r].useful = bitset_test (P, r);
    grammar_rules_useless_report (_("rule useless in grammar"));
  }

  /* Map the nonterminals to their new index: useful first, useless
     afterwards.  Kept for later report.  */
  {
    int useful = 0;
    int useless = nrules - nuseless_productions;
    rule *rules_sorted = xnmalloc (nrules, sizeof *rules_sorted);
    rule_number r;
    for (r = 0; r < nrules; ++r)
      rules_sorted[rules[r].useful ? useful++ : useless++] = rules[r];
    free (rules);
    rules = rules_sorted;

    /* Renumber the rules markers in RITEMS.  */
    for (r = 0; r < nrules; ++r)
      {
        item_number *rhsp = rules[r].rhs;
        for (/* Nothing. */; *rhsp >= 0; ++rhsp)
          /* Nothing. */;
        *rhsp = rule_number_as_item_number (r);
        rules[r].number = r;
      }
    nrules -= nuseless_productions;
  }

  /* Adjust NRITEMS.  */
  {
    rule_number r;
    int length;
    for (r = nrules; r < nrules + nuseless_productions; ++r)
      {
        length = rule_rhs_length (&rules[r]);
        nritems -= length + 1;
      }
  }
}


/*------------------------------.
| Remove useless nonterminals.  |
`------------------------------*/

static void
nonterminals_reduce (void)
{
  /* Map the nonterminals to their new index: useful first, useless
     afterwards.  Kept for later report.  */

  symbol_number *nontermmap = xnmalloc (nvars, sizeof *nontermmap);
  symbol_number n = ntokens;
  symbol_number i;
  for (i = ntokens; i < nsyms; i++)
    if (bitset_test (V, i))
      nontermmap[i - ntokens] = n++;
  for (i = ntokens; i < nsyms; i++)
    if (!bitset_test (V, i))
      {
        nontermmap[i - ntokens] = n++;
        if (symbols[i]->status != used)
          complain (&symbols[i]->location, Wother,
                    _("nonterminal useless in grammar: %s"),
                    symbols[i]->tag);
      }


  /* Shuffle elements of tables indexed by symbol number.  */
  {
    symbol **symbols_sorted = xnmalloc (nvars, sizeof *symbols_sorted);

    for (i = ntokens; i < nsyms; i++)
      symbols[i]->number = nontermmap[i - ntokens];
    for (i = ntokens; i < nsyms; i++)
      symbols_sorted[nontermmap[i - ntokens] - ntokens] = symbols[i];
    for (i = ntokens; i < nsyms; i++)
      symbols[i] = symbols_sorted[i - ntokens];
    free (symbols_sorted);
  }

  {
    rule_number r;
    for (r = 0; r < nrules; ++r)
      {
        item_number *rhsp;
        for (rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp)
          if (ISVAR (*rhsp))
            *rhsp =  symbol_number_as_item_number (nontermmap[*rhsp
                                                              - ntokens]);
      }
    accept->number = nontermmap[accept->number - ntokens];
  }

  nsyms -= nuseless_nonterminals;
  nvars -= nuseless_nonterminals;

  free (nontermmap);
}


/*------------------------------------------------------------------.
| Output the detailed results of the reductions.  For FILE.output.  |
`------------------------------------------------------------------*/

void
reduce_output (FILE *out)
{
  if (nuseless_nonterminals > 0)
    {
      int i;
      fprintf (out, "%s\n\n", _("Nonterminals useless in grammar"));
      for (i = 0; i < nuseless_nonterminals; ++i)
        fprintf (out, "   %s\n", symbols[nsyms + i]->tag);
      fputs ("\n\n", out);
    }

  {
    bool b = false;
    int i;
    for (i = 0; i < ntokens; i++)
      if (reduce_token_unused_in_grammar (i))
        {
          if (!b)
            fprintf (out, "%s\n\n", _("Terminals unused in grammar"));
          b = true;
          fprintf (out, "   %s\n", symbols[i]->tag);
        }
    if (b)
      fputs ("\n\n", out);
  }

  if (nuseless_productions > 0)
    grammar_rules_partial_print (out, _("Rules useless in grammar"),
                                 rule_useless_in_grammar_p);
}


/*-------------------------------.
| Report the results to STDERR.  |
`-------------------------------*/

static void
reduce_print (void)
{
  if (nuseless_nonterminals > 0)
    complain (NULL, Wother, ngettext ("%d nonterminal useless in grammar",
                                      "%d nonterminals useless in grammar",
                                      nuseless_nonterminals),
              nuseless_nonterminals);
  if (nuseless_productions > 0)
    complain (NULL, Wother, ngettext ("%d rule useless in grammar",
                                      "%d rules useless in grammar",
                                      nuseless_productions),
              nuseless_productions);
}

void
reduce_grammar (void)
{
  bool reduced;

  /* Allocate the global sets used to compute the reduced grammar */

  N = bitset_create (nvars, BITSET_FIXED);
  P =  bitset_create (nrules, BITSET_FIXED);
  V = bitset_create (nsyms, BITSET_FIXED);
  V1 = bitset_create (nsyms, BITSET_FIXED);

  useless_nonterminals ();
  inaccessable_symbols ();

  reduced = (nuseless_nonterminals + nuseless_productions > 0);
  if (!reduced)
    return;

  reduce_print ();

  if (!bitset_test (N, accept->number - ntokens))
    complain (&startsymbol_location, fatal,
              _("start symbol %s does not derive any sentence"),
              startsymbol->tag);

  /* First reduce the nonterminals, as they renumber themselves in the
     whole grammar.  If you change the order, nonterms would be
     renumbered only in the reduced grammar.  */
  if (nuseless_nonterminals > 0)
    nonterminals_reduce ();
  if (nuseless_productions > 0)
    reduce_grammar_tables ();

  if (trace_flag & trace_grammar)
    {
      grammar_dump (stderr, "Reduced Grammar");

      fprintf (stderr, "reduced %s defines %d terminals, %d nonterminals"
               ", and %d productions.\n",
               grammar_file, ntokens, nvars, nrules);
    }
}

bool
reduce_token_unused_in_grammar (symbol_number i)
{
  aver (i < ntokens);
  return !bitset_test (V, i) && !bitset_test (V1, i);
}

bool
reduce_nonterminal_useless_in_grammar (symbol_number i)
{
  aver (ntokens <= i && i < nsyms + nuseless_nonterminals);
  return nsyms <= i;
}

/*-----------------------------------------------------------.
| Free the global sets used to compute the reduced grammar.  |
`-----------------------------------------------------------*/

void
reduce_free (void)
{
  bitset_free (N);
  bitset_free (V);
  bitset_free (V1);
  bitset_free (P);
}