]>
Commit | Line | Data |
---|---|---|
1 | /* Grammar reduction for Bison. | |
2 | Copyright (C) 1988, 1989, 2000, 2001, 2002 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 | /* Reduce the grammar: Find and eliminate unreachable terminals, | |
23 | nonterminals, and productions. David S. Bakin. */ | |
24 | ||
25 | /* Don't eliminate unreachable terminals: They may be used by the | |
26 | user's parser. */ | |
27 | ||
28 | #include "system.h" | |
29 | #include "quotearg.h" | |
30 | #include "getargs.h" | |
31 | #include "files.h" | |
32 | #include "symtab.h" | |
33 | #include "gram.h" | |
34 | #include "complain.h" | |
35 | #include "reduce.h" | |
36 | #include "reader.h" | |
37 | #include "getargs.h" | |
38 | #include "bitset.h" | |
39 | ||
40 | typedef short *rule; | |
41 | ||
42 | ||
43 | /* Set of all nonterminals which are not useless. */ | |
44 | static bitset N; | |
45 | ||
46 | /* Set of all rules which have no useless nonterminals in their RHS. */ | |
47 | static bitset P; | |
48 | ||
49 | /* Set of all accessible symbols. */ | |
50 | static bitset V; | |
51 | ||
52 | /* Set of symbols used to define rule precedence (so they are | |
53 | `useless', but no warning should be issued). */ | |
54 | static bitset V1; | |
55 | ||
56 | static int nuseful_productions; | |
57 | static int nuseless_productions; | |
58 | static int nuseful_nonterminals; | |
59 | int nuseless_nonterminals; | |
60 | \f | |
61 | /*-------------------------------------------------------------------. | |
62 | | Another way to do this would be with a set for each production and | | |
63 | | then do subset tests against N0, but even for the C grammar the | | |
64 | | whole reducing process takes only 2 seconds on my 8Mhz AT. | | |
65 | `-------------------------------------------------------------------*/ | |
66 | ||
67 | static bool | |
68 | useful_production (int i, bitset N0) | |
69 | { | |
70 | rule r; | |
71 | short n; | |
72 | ||
73 | /* A production is useful if all of the nonterminals in its appear | |
74 | in the set of useful nonterminals. */ | |
75 | ||
76 | for (r = rules[i].rhs; *r >= 0; r++) | |
77 | if (ISVAR (n = *r) && !bitset_test (N0, n - ntokens)) | |
78 | return FALSE; | |
79 | return TRUE; | |
80 | } | |
81 | ||
82 | ||
83 | /*---------------------------------------------------------. | |
84 | | Remember that rules are 1-origin, symbols are 0-origin. | | |
85 | `---------------------------------------------------------*/ | |
86 | ||
87 | static void | |
88 | useless_nonterminals (void) | |
89 | { | |
90 | bitset Np, Ns; | |
91 | int i; | |
92 | ||
93 | /* N is set as built. Np is set being built this iteration. P is | |
94 | set of all productions which have a RHS all in N. */ | |
95 | ||
96 | Np = bitset_create (nvars, BITSET_FIXED); | |
97 | ||
98 | ||
99 | /* The set being computed is a set of nonterminals which can derive | |
100 | the empty string or strings consisting of all terminals. At each | |
101 | iteration a nonterminal is added to the set if there is a | |
102 | production with that nonterminal as its LHS for which all the | |
103 | nonterminals in its RHS are already in the set. Iterate until | |
104 | the set being computed remains unchanged. Any nonterminals not | |
105 | in the set at that point are useless in that they will never be | |
106 | used in deriving a sentence of the language. | |
107 | ||
108 | This iteration doesn't use any special traversal over the | |
109 | productions. A set is kept of all productions for which all the | |
110 | nonterminals in the RHS are in useful. Only productions not in | |
111 | this set are scanned on each iteration. At the end, this set is | |
112 | saved to be used when finding useful productions: only | |
113 | productions in this set will appear in the final grammar. */ | |
114 | ||
115 | while (1) | |
116 | { | |
117 | bitset_copy (Np, N); | |
118 | for (i = 1; i < nrules + 1; i++) | |
119 | if (!bitset_test (P, i) | |
120 | && useful_production (i, N)) | |
121 | { | |
122 | bitset_set (Np, rules[i].lhs->number - ntokens); | |
123 | bitset_set (P, i); | |
124 | } | |
125 | if (bitset_equal_p (N, Np)) | |
126 | break; | |
127 | Ns = Np; | |
128 | Np = N; | |
129 | N = Ns; | |
130 | } | |
131 | bitset_free (N); | |
132 | N = Np; | |
133 | } | |
134 | ||
135 | ||
136 | static void | |
137 | inaccessable_symbols (void) | |
138 | { | |
139 | bitset Vp, Vs, Pp; | |
140 | int i; | |
141 | short t; | |
142 | rule r; | |
143 | ||
144 | /* Find out which productions are reachable and which symbols are | |
145 | used. Starting with an empty set of productions and a set of | |
146 | symbols which only has the start symbol in it, iterate over all | |
147 | productions until the set of productions remains unchanged for an | |
148 | iteration. For each production which has a LHS in the set of | |
149 | reachable symbols, add the production to the set of reachable | |
150 | productions, and add all of the nonterminals in the RHS of the | |
151 | production to the set of reachable symbols. | |
152 | ||
153 | Consider only the (partially) reduced grammar which has only | |
154 | nonterminals in N and productions in P. | |
155 | ||
156 | The result is the set P of productions in the reduced grammar, | |
157 | and the set V of symbols in the reduced grammar. | |
158 | ||
159 | Although this algorithm also computes the set of terminals which | |
160 | are reachable, no terminal will be deleted from the grammar. Some | |
161 | terminals might not be in the grammar but might be generated by | |
162 | semantic routines, and so the user might want them available with | |
163 | specified numbers. (Is this true?) However, the nonreachable | |
164 | terminals are printed (if running in verbose mode) so that the | |
165 | user can know. */ | |
166 | ||
167 | Vp = bitset_create (nsyms, BITSET_FIXED); | |
168 | Pp = bitset_create (nrules + 1, BITSET_FIXED); | |
169 | ||
170 | /* If the start symbol isn't useful, then nothing will be useful. */ | |
171 | if (bitset_test (N, start_symbol - ntokens)) | |
172 | { | |
173 | bitset_set (V, start_symbol); | |
174 | ||
175 | while (1) | |
176 | { | |
177 | bitset_copy (Vp, V); | |
178 | for (i = 1; i < nrules + 1; i++) | |
179 | { | |
180 | if (!bitset_test (Pp, i) | |
181 | && bitset_test (P, i) | |
182 | && bitset_test (V, rules[i].lhs->number)) | |
183 | { | |
184 | for (r = rules[i].rhs; *r >= 0; r++) | |
185 | if (ISTOKEN (t = *r) || bitset_test (N, t - ntokens)) | |
186 | bitset_set (Vp, t); | |
187 | bitset_set (Pp, i); | |
188 | } | |
189 | } | |
190 | if (bitset_equal_p (V, Vp)) | |
191 | break; | |
192 | Vs = Vp; | |
193 | Vp = V; | |
194 | V = Vs; | |
195 | } | |
196 | } | |
197 | ||
198 | bitset_free (V); | |
199 | V = Vp; | |
200 | ||
201 | /* Tokens 0, 1, and 2 are internal to Bison. Consider them useful. */ | |
202 | bitset_set (V, eoftoken->number); /* end-of-input token */ | |
203 | bitset_set (V, errtoken->number); /* error token */ | |
204 | bitset_set (V, undeftoken->number); /* some undefined token */ | |
205 | ||
206 | bitset_free (P); | |
207 | P = Pp; | |
208 | ||
209 | nuseful_productions = bitset_count (P); | |
210 | nuseless_productions = nrules - nuseful_productions; | |
211 | ||
212 | nuseful_nonterminals = 0; | |
213 | for (i = ntokens; i < nsyms; i++) | |
214 | if (bitset_test (V, i)) | |
215 | nuseful_nonterminals++; | |
216 | nuseless_nonterminals = nvars - nuseful_nonterminals; | |
217 | ||
218 | /* A token that was used in %prec should not be warned about. */ | |
219 | for (i = 1; i < nrules + 1; i++) | |
220 | if (rules[i].precsym != 0) | |
221 | bitset_set (V1, rules[i].precsym->number); | |
222 | } | |
223 | ||
224 | ||
225 | /*-------------------------------------------------------------------. | |
226 | | Put the useless productions at the end of RULES, and adjust NRULES | | |
227 | | accordingly. | | |
228 | `-------------------------------------------------------------------*/ | |
229 | ||
230 | static void | |
231 | reduce_grammar_tables (void) | |
232 | { | |
233 | /* Flag useless productions. */ | |
234 | { | |
235 | int pn; | |
236 | for (pn = 1; pn < nrules + 1; pn++) | |
237 | rules[pn].useful = bitset_test (P, pn); | |
238 | } | |
239 | ||
240 | /* Map the nonterminals to their new index: useful first, useless | |
241 | afterwards. Kept for later report. */ | |
242 | { | |
243 | int useful = 1; | |
244 | int useless = nrules + 1 - nuseless_productions; | |
245 | rule_t *rules_sorted = XMALLOC (rule_t, nrules + 1) - 1; | |
246 | int i; | |
247 | for (i = 1; i < nrules + 1; ++i) | |
248 | rules_sorted[rules[i].useful ? useful++ : useless++] = rules[i]; | |
249 | free (rules + 1); | |
250 | rules = rules_sorted; | |
251 | ||
252 | /* Renumber the rules markers in RITEMS. */ | |
253 | for (i = 1; i < nrules + 1; ++i) | |
254 | { | |
255 | short *rhsp = rules[i].rhs; | |
256 | for (/* Nothing. */; *rhsp >= 0; ++rhsp) | |
257 | /* Nothing. */; | |
258 | *rhsp = -i; | |
259 | rules[i].number = i; | |
260 | } | |
261 | nrules -= nuseless_productions; | |
262 | } | |
263 | ||
264 | /* Adjust NRITEMS and NITEMS. */ | |
265 | { | |
266 | int r; | |
267 | int length; | |
268 | for (r = nrules + 1; r < nrules + 1 + nuseless_productions; ++r) | |
269 | { | |
270 | length = rule_rhs_length (&rules[r]); | |
271 | nritems -= length + 1; | |
272 | } | |
273 | } | |
274 | } | |
275 | ||
276 | ||
277 | /*------------------------------. | |
278 | | Remove useless nonterminals. | | |
279 | `------------------------------*/ | |
280 | ||
281 | static void | |
282 | nonterminals_reduce (void) | |
283 | { | |
284 | int i, n; | |
285 | ||
286 | /* Map the nonterminals to their new index: useful first, useless | |
287 | afterwards. Kept for later report. */ | |
288 | ||
289 | short *nontermmap = XCALLOC (short, nvars) - ntokens; | |
290 | n = ntokens; | |
291 | for (i = ntokens; i < nsyms; i++) | |
292 | if (bitset_test (V, i)) | |
293 | nontermmap[i] = n++; | |
294 | for (i = ntokens; i < nsyms; i++) | |
295 | if (!bitset_test (V, i)) | |
296 | nontermmap[i] = n++; | |
297 | ||
298 | ||
299 | /* Shuffle elements of tables indexed by symbol number. */ | |
300 | { | |
301 | symbol_t **symbols_sorted = XMALLOC (symbol_t *, nvars) - ntokens; | |
302 | ||
303 | for (i = ntokens; i < nsyms; i++) | |
304 | symbols[i]->number = nontermmap[i]; | |
305 | for (i = ntokens; i < nsyms; i++) | |
306 | symbols_sorted[nontermmap[i]] = symbols[i]; | |
307 | for (i = ntokens; i < nsyms; i++) | |
308 | symbols[i] = symbols_sorted[i]; | |
309 | free (symbols_sorted + ntokens); | |
310 | } | |
311 | ||
312 | for (i = 0; i < nritems; ++i) | |
313 | if (ISVAR (ritem[i])) | |
314 | ritem[i] = nontermmap[ritem[i]]; | |
315 | ||
316 | start_symbol = nontermmap[start_symbol]; | |
317 | ||
318 | nsyms -= nuseless_nonterminals; | |
319 | nvars -= nuseless_nonterminals; | |
320 | ||
321 | free (nontermmap + ntokens); | |
322 | } | |
323 | ||
324 | ||
325 | /*------------------------------------------------------------------. | |
326 | | Output the detailed results of the reductions. For FILE.output. | | |
327 | `------------------------------------------------------------------*/ | |
328 | ||
329 | void | |
330 | reduce_output (FILE *out) | |
331 | { | |
332 | if (nuseless_nonterminals > 0) | |
333 | { | |
334 | int i; | |
335 | fprintf (out, "%s\n\n", _("Useless nonterminals:")); | |
336 | for (i = 0; i < nuseless_nonterminals; ++i) | |
337 | fprintf (out, " %s\n", quotearg_style (escape_quoting_style, | |
338 | symbols[nsyms + i]->tag)); | |
339 | fputs ("\n\n", out); | |
340 | } | |
341 | ||
342 | { | |
343 | bool b = FALSE; | |
344 | int i; | |
345 | for (i = 0; i < ntokens; i++) | |
346 | if (!bitset_test (V, i) && !bitset_test (V1, i)) | |
347 | { | |
348 | if (!b) | |
349 | fprintf (out, "%s\n\n", _("Terminals which are not used:")); | |
350 | b = TRUE; | |
351 | fprintf (out, " %s\n", quotearg_style (escape_quoting_style, | |
352 | symbols[i]->tag)); | |
353 | } | |
354 | if (b) | |
355 | fputs ("\n\n", out); | |
356 | } | |
357 | ||
358 | if (nuseless_productions > 0) | |
359 | { | |
360 | int i; | |
361 | fprintf (out, "%s\n\n", _("Useless rules:")); | |
362 | for (i = nrules + 1; i < nuseless_productions + nrules + 1; i++) | |
363 | { | |
364 | rule r; | |
365 | fprintf (out, "#%-4d ", rules[i].user_number - 1); | |
366 | fprintf (out, "%s:", quotearg_style (escape_quoting_style, | |
367 | rules[i].lhs->tag)); | |
368 | for (r = rules[i].rhs; *r >= 0; r++) | |
369 | fprintf (out, " %s", quotearg_style (escape_quoting_style, | |
370 | symbols[*r]->tag)); | |
371 | fputs (";\n", out); | |
372 | } | |
373 | fputs ("\n\n", out); | |
374 | } | |
375 | } | |
376 | \f | |
377 | static void | |
378 | dump_grammar (FILE *out) | |
379 | { | |
380 | int i; | |
381 | rule r; | |
382 | ||
383 | fprintf (out, "REDUCED GRAMMAR\n\n"); | |
384 | fprintf (out, | |
385 | "ntokens = %d, nvars = %d, nsyms = %d, nrules = %d, nritems = %d\n\n", | |
386 | ntokens, nvars, nsyms, nrules, nritems); | |
387 | fprintf (out, "Variables\n---------\n\n"); | |
388 | fprintf (out, "Value Sprec Sassoc Tag\n"); | |
389 | for (i = ntokens; i < nsyms; i++) | |
390 | fprintf (out, "%5d %5d %5d %s\n", | |
391 | i, | |
392 | symbols[i]->prec, symbols[i]->assoc, | |
393 | quotearg_style (escape_quoting_style, symbols[i]->tag)); | |
394 | fprintf (out, "\n\n"); | |
395 | fprintf (out, "Rules\n-----\n\n"); | |
396 | fprintf (out, "Num (Prec, Assoc, Useful, Ritem Range) Lhs -> Rhs (Ritem range) [Num]\n"); | |
397 | for (i = 1; i < nrules + nuseless_productions + 1; i++) | |
398 | { | |
399 | int rhs_count = 0; | |
400 | /* Find the last RHS index in ritems. */ | |
401 | for (r = rules[i].rhs; *r >= 0; ++r) | |
402 | ++rhs_count; | |
403 | fprintf (out, "%3d (%2d, %2d, %2d, %2d-%2d) %2d ->", | |
404 | i - 1, | |
405 | rules[i].prec->prec, | |
406 | rules[i].prec->assoc, | |
407 | rules[i].useful, | |
408 | rules[i].rhs - ritem, | |
409 | rules[i].rhs - ritem + rhs_count - 1, | |
410 | rules[i].lhs->number); | |
411 | /* Dumped the RHS. */ | |
412 | for (r = rules[i].rhs; *r >= 0; r++) | |
413 | fprintf (out, "%3d", *r); | |
414 | fprintf (out, " [%d]\n", -(*r) - 1); | |
415 | } | |
416 | fprintf (out, "\n\n"); | |
417 | fprintf (out, "Rules interpreted\n-----------------\n\n"); | |
418 | for (i = 1; i < nrules + nuseless_productions + 1; i++) | |
419 | { | |
420 | fprintf (out, "%-5d %s :", | |
421 | i, quotearg_style (escape_quoting_style, rules[i].lhs->tag)); | |
422 | for (r = rules[i].rhs; *r >= 0; r++) | |
423 | fprintf (out, " %s", | |
424 | quotearg_style (escape_quoting_style, symbols[*r]->tag)); | |
425 | fputc ('\n', out); | |
426 | } | |
427 | fprintf (out, "\n\n"); | |
428 | } | |
429 | ||
430 | ||
431 | ||
432 | /*-------------------------------. | |
433 | | Report the results to STDERR. | | |
434 | `-------------------------------*/ | |
435 | ||
436 | static void | |
437 | reduce_print (void) | |
438 | { | |
439 | if (yacc_flag && nuseless_productions) | |
440 | fprintf (stderr, ngettext ("%d rule never reduced\n", | |
441 | "%d rules never reduced\n", | |
442 | nuseless_productions), | |
443 | nuseless_productions); | |
444 | ||
445 | fprintf (stderr, _("%s contains "), infile); | |
446 | ||
447 | if (nuseless_nonterminals > 0) | |
448 | fprintf (stderr, ngettext ("%d useless nonterminal", | |
449 | "%d useless nonterminals", | |
450 | nuseless_nonterminals), | |
451 | nuseless_nonterminals); | |
452 | ||
453 | if (nuseless_nonterminals > 0 && nuseless_productions > 0) | |
454 | fprintf (stderr, _(" and ")); | |
455 | ||
456 | if (nuseless_productions > 0) | |
457 | fprintf (stderr, ngettext ("%d useless rule", | |
458 | "%d useless rules", | |
459 | nuseless_productions), | |
460 | nuseless_productions); | |
461 | fprintf (stderr, "\n"); | |
462 | fflush (stderr); | |
463 | } | |
464 | \f | |
465 | void | |
466 | reduce_grammar (void) | |
467 | { | |
468 | bool reduced; | |
469 | ||
470 | /* Allocate the global sets used to compute the reduced grammar */ | |
471 | ||
472 | N = bitset_create (nvars, BITSET_FIXED); | |
473 | P = bitset_create (nrules + 1, BITSET_FIXED); | |
474 | V = bitset_create (nsyms, BITSET_FIXED); | |
475 | V1 = bitset_create (nsyms, BITSET_FIXED); | |
476 | ||
477 | useless_nonterminals (); | |
478 | inaccessable_symbols (); | |
479 | ||
480 | reduced = (bool) (nuseless_nonterminals + nuseless_productions > 0); | |
481 | if (!reduced) | |
482 | return; | |
483 | ||
484 | reduce_print (); | |
485 | ||
486 | if (!bitset_test (N, start_symbol - ntokens)) | |
487 | fatal (_("Start symbol %s does not derive any sentence"), | |
488 | quotearg_style (escape_quoting_style, symbols[start_symbol]->tag)); | |
489 | ||
490 | if (nuseless_productions > 0) | |
491 | reduce_grammar_tables (); | |
492 | if (nuseless_nonterminals > 0) | |
493 | nonterminals_reduce (); | |
494 | ||
495 | if (trace_flag) | |
496 | { | |
497 | dump_grammar (stderr); | |
498 | ||
499 | fprintf (stderr, "reduced %s defines %d terminals, %d nonterminals\ | |
500 | , and %d productions.\n", | |
501 | infile, ntokens, nvars, nrules); | |
502 | } | |
503 | } | |
504 | ||
505 | ||
506 | /*-----------------------------------------------------------. | |
507 | | Free the global sets used to compute the reduced grammar. | | |
508 | `-----------------------------------------------------------*/ | |
509 | ||
510 | void | |
511 | reduce_free (void) | |
512 | { | |
513 | bitset_free (N); | |
514 | bitset_free (V); | |
515 | bitset_free (V1); | |
516 | bitset_free (P); | |
517 | } |