]> git.saurik.com Git - bison.git/blame_incremental - src/lalr.c
bison: catch bad symbol names.
[bison.git] / src / lalr.c
... / ...
CommitLineData
1/* Compute lookahead criteria for Bison.
2
3 Copyright (C) 1984, 1986, 1989, 2000, 2001, 2002, 2003, 2004, 2005,
4 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
5
6 This file is part of Bison, the GNU Compiler Compiler.
7
8 This program is free software: you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation, either version 3 of the License, or
11 (at your option) any later version.
12
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
17
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
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 <config.h>
27#include "system.h"
28
29#include <bitset.h>
30#include <bitsetv.h>
31#include <quotearg.h>
32
33#include "LR0.h"
34#include "complain.h"
35#include "derives.h"
36#include "getargs.h"
37#include "gram.h"
38#include "lalr.h"
39#include "muscle_tab.h"
40#include "nullable.h"
41#include "reader.h"
42#include "relation.h"
43#include "symtab.h"
44
45goto_number *goto_map;
46goto_number ngotos;
47state_number *from_state;
48state_number *to_state;
49bitsetv goto_follows = NULL;
50
51/* Linked list of goto numbers. */
52typedef struct goto_list
53{
54 struct goto_list *next;
55 goto_number value;
56} goto_list;
57
58
59/* LA is an NLA by NTOKENS matrix of bits. LA[l, i] is 1 if the rule
60 LArule[l] is applicable in the appropriate state when the next
61 token is symbol i. If LA[l, i] and LA[l, j] are both 1 for i != j,
62 it is a conflict. */
63
64static bitsetv LA = NULL;
65size_t nLA;
66
67
68static goto_number **includes;
69static goto_list **lookback;
70
71
72
73
74void
75set_goto_map (void)
76{
77 state_number s;
78 goto_number *temp_map;
79
80 goto_map = xcalloc (nvars + 1, sizeof *goto_map);
81 temp_map = xnmalloc (nvars + 1, sizeof *temp_map);
82
83 ngotos = 0;
84 for (s = 0; s < nstates; ++s)
85 {
86 transitions *sp = states[s]->transitions;
87 int i;
88 for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i)
89 {
90 ngotos++;
91
92 /* Abort if (ngotos + 1) would overflow. */
93 aver (ngotos != GOTO_NUMBER_MAXIMUM);
94
95 goto_map[TRANSITION_SYMBOL (sp, i) - ntokens]++;
96 }
97 }
98
99 {
100 goto_number k = 0;
101 int i;
102 for (i = ntokens; i < nsyms; i++)
103 {
104 temp_map[i - ntokens] = k;
105 k += goto_map[i - ntokens];
106 }
107
108 for (i = ntokens; i < nsyms; i++)
109 goto_map[i - ntokens] = temp_map[i - ntokens];
110
111 goto_map[nsyms - ntokens] = ngotos;
112 temp_map[nsyms - ntokens] = ngotos;
113 }
114
115 from_state = xcalloc (ngotos, sizeof *from_state);
116 to_state = xcalloc (ngotos, sizeof *to_state);
117
118 for (s = 0; s < nstates; ++s)
119 {
120 transitions *sp = states[s]->transitions;
121 int i;
122 for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i)
123 {
124 goto_number k = temp_map[TRANSITION_SYMBOL (sp, i) - ntokens]++;
125 from_state[k] = s;
126 to_state[k] = sp->states[i]->number;
127 }
128 }
129
130 free (temp_map);
131}
132
133
134goto_number
135map_goto (state_number s0, symbol_number sym)
136{
137 goto_number high;
138 goto_number low;
139 goto_number middle;
140 state_number s;
141
142 low = goto_map[sym - ntokens];
143 high = goto_map[sym - ntokens + 1] - 1;
144
145 for (;;)
146 {
147 aver (low <= high);
148 middle = (low + high) / 2;
149 s = from_state[middle];
150 if (s == s0)
151 return middle;
152 else if (s < s0)
153 low = middle + 1;
154 else
155 high = middle - 1;
156 }
157}
158
159
160static void
161initialize_F (void)
162{
163 goto_number **reads = xnmalloc (ngotos, sizeof *reads);
164 goto_number *edge = xnmalloc (ngotos + 1, sizeof *edge);
165 goto_number nedges = 0;
166
167 goto_number i;
168
169 goto_follows = bitsetv_create (ngotos, ntokens, BITSET_FIXED);
170
171 for (i = 0; i < ngotos; i++)
172 {
173 state_number stateno = to_state[i];
174 transitions *sp = states[stateno]->transitions;
175
176 int j;
177 FOR_EACH_SHIFT (sp, j)
178 bitset_set (goto_follows[i], TRANSITION_SYMBOL (sp, j));
179
180 for (; j < sp->num; j++)
181 {
182 symbol_number sym = TRANSITION_SYMBOL (sp, j);
183 if (nullable[sym - ntokens])
184 edge[nedges++] = map_goto (stateno, sym);
185 }
186
187 if (nedges == 0)
188 reads[i] = NULL;
189 else
190 {
191 reads[i] = xnmalloc (nedges + 1, sizeof reads[i][0]);
192 memcpy (reads[i], edge, nedges * sizeof edge[0]);
193 reads[i][nedges] = END_NODE;
194 nedges = 0;
195 }
196 }
197
198 relation_digraph (reads, ngotos, &goto_follows);
199
200 for (i = 0; i < ngotos; i++)
201 free (reads[i]);
202
203 free (reads);
204 free (edge);
205}
206
207
208static void
209add_lookback_edge (state *s, rule *r, goto_number gotono)
210{
211 int ri = state_reduction_find (s, r);
212 goto_list *sp = xmalloc (sizeof *sp);
213 sp->next = lookback[(s->reductions->lookahead_tokens - LA) + ri];
214 sp->value = gotono;
215 lookback[(s->reductions->lookahead_tokens - LA) + ri] = sp;
216}
217
218
219
220static void
221build_relations (void)
222{
223 goto_number *edge = xnmalloc (ngotos + 1, sizeof *edge);
224 state_number *states1 = xnmalloc (ritem_longest_rhs () + 1, sizeof *states1);
225 goto_number i;
226
227 includes = xnmalloc (ngotos, sizeof *includes);
228
229 for (i = 0; i < ngotos; i++)
230 {
231 int nedges = 0;
232 symbol_number symbol1 = states[to_state[i]]->accessing_symbol;
233 rule **rulep;
234
235 for (rulep = derives[symbol1 - ntokens]; *rulep; rulep++)
236 {
237 bool done;
238 int length = 1;
239 item_number const *rp;
240 state *s = states[from_state[i]];
241 states1[0] = s->number;
242
243 for (rp = (*rulep)->rhs; ! item_number_is_rule_number (*rp); rp++)
244 {
245 s = transitions_to (s->transitions,
246 item_number_as_symbol_number (*rp));
247 states1[length++] = s->number;
248 }
249
250 if (!s->consistent)
251 add_lookback_edge (s, *rulep, i);
252
253 length--;
254 done = false;
255 while (!done)
256 {
257 done = true;
258 /* Each rhs ends in a rule number, and there is a
259 sentinel (ritem[-1]=0) before the first rhs, so it is safe to
260 decrement RP here. */
261 rp--;
262 if (ISVAR (*rp))
263 {
264 /* Downcasting from item_number to symbol_number. */
265 edge[nedges++] = map_goto (states1[--length],
266 item_number_as_symbol_number (*rp));
267 if (nullable[*rp - ntokens])
268 done = false;
269 }
270 }
271 }
272
273 if (nedges == 0)
274 includes[i] = NULL;
275 else
276 {
277 int j;
278 includes[i] = xnmalloc (nedges + 1, sizeof includes[i][0]);
279 for (j = 0; j < nedges; j++)
280 includes[i][j] = edge[j];
281 includes[i][nedges] = END_NODE;
282 }
283 }
284
285 free (edge);
286 free (states1);
287
288 relation_transpose (&includes, ngotos);
289}
290
291
292
293static void
294compute_FOLLOWS (void)
295{
296 goto_number i;
297
298 relation_digraph (includes, ngotos, &goto_follows);
299
300 for (i = 0; i < ngotos; i++)
301 free (includes[i]);
302
303 free (includes);
304}
305
306
307static void
308compute_lookahead_tokens (void)
309{
310 size_t i;
311 goto_list *sp;
312
313 for (i = 0; i < nLA; i++)
314 for (sp = lookback[i]; sp; sp = sp->next)
315 bitset_or (LA[i], LA[i], goto_follows[sp->value]);
316
317 /* Free LOOKBACK. */
318 for (i = 0; i < nLA; i++)
319 LIST_FREE (goto_list, lookback[i]);
320
321 free (lookback);
322}
323
324
325/*----------------------------------------------------.
326| Count the number of lookahead tokens required for S |
327| (N_LOOKAHEAD_TOKENS member). |
328`----------------------------------------------------*/
329
330static int
331state_lookahead_tokens_count (state *s, bool default_reduction_only_for_accept)
332{
333 int n_lookahead_tokens = 0;
334 reductions *rp = s->reductions;
335 transitions *sp = s->transitions;
336
337 /* Transitions are only disabled during conflict resolution, and that
338 hasn't happened yet, so there should be no need to check that
339 transition 0 hasn't been disabled before checking if it is a shift.
340 However, this check was performed at one time, so we leave it as an
341 aver. */
342 aver (sp->num == 0 || !TRANSITION_IS_DISABLED (sp, 0));
343
344 /* We need a lookahead either to distinguish different reductions
345 (i.e., there are two or more), or to distinguish a reduction from a
346 shift. Otherwise, it is straightforward, and the state is
347 `consistent'. However, treat only the accepting state as
348 consistent (because there is never a lookahead token that makes
349 sense there, and so no lookahead token should be read) if the user
350 has otherwise disabled default reductions. */
351 if (rp->num > 1
352 || (rp->num == 1 && sp->num && TRANSITION_IS_SHIFT (sp, 0))
353 || (rp->num == 1 && rp->rules[0]->number != 0
354 && default_reduction_only_for_accept))
355 n_lookahead_tokens += rp->num;
356 else
357 s->consistent = 1;
358
359 return n_lookahead_tokens;
360}
361
362
363/*----------------------------------------------------.
364| Compute LA, NLA, and the lookahead_tokens members. |
365`----------------------------------------------------*/
366
367void
368initialize_LA (void)
369{
370 state_number i;
371 bitsetv pLA;
372 bool default_reduction_only_for_accept;
373 {
374 char *default_reductions =
375 muscle_percent_define_get ("lr.default-reductions");
376 default_reduction_only_for_accept =
377 0 == strcmp (default_reductions, "accepting");
378 free (default_reductions);
379 }
380
381 /* Compute the total number of reductions requiring a lookahead. */
382 nLA = 0;
383 for (i = 0; i < nstates; i++)
384 nLA +=
385 state_lookahead_tokens_count (states[i],
386 default_reduction_only_for_accept);
387 /* Avoid having to special case 0. */
388 if (!nLA)
389 nLA = 1;
390
391 pLA = LA = bitsetv_create (nLA, ntokens, BITSET_FIXED);
392
393 /* Initialize the members LOOKAHEAD_TOKENS for each state whose reductions
394 require lookahead tokens. */
395 for (i = 0; i < nstates; i++)
396 {
397 int count =
398 state_lookahead_tokens_count (states[i],
399 default_reduction_only_for_accept);
400 if (count)
401 {
402 states[i]->reductions->lookahead_tokens = pLA;
403 pLA += count;
404 }
405 }
406}
407
408
409/*---------------------------------------------.
410| Output the lookahead tokens for each state. |
411`---------------------------------------------*/
412
413static void
414lookahead_tokens_print (FILE *out)
415{
416 state_number i;
417 int j, k;
418 fprintf (out, "Lookahead tokens: BEGIN\n");
419 for (i = 0; i < nstates; ++i)
420 {
421 reductions *reds = states[i]->reductions;
422 bitset_iterator iter;
423 int n_lookahead_tokens = 0;
424
425 if (reds->lookahead_tokens)
426 for (k = 0; k < reds->num; ++k)
427 if (reds->lookahead_tokens[k])
428 ++n_lookahead_tokens;
429
430 fprintf (out, "State %d: %d lookahead tokens\n",
431 i, n_lookahead_tokens);
432
433 if (reds->lookahead_tokens)
434 for (j = 0; j < reds->num; ++j)
435 BITSET_FOR_EACH (iter, reds->lookahead_tokens[j], k, 0)
436 {
437 fprintf (out, " on %d (%s) -> rule %d\n",
438 k, symbols[k]->tag,
439 reds->rules[j]->number);
440 };
441 }
442 fprintf (out, "Lookahead tokens: END\n");
443}
444
445void
446lalr (void)
447{
448 initialize_LA ();
449 set_goto_map ();
450 initialize_F ();
451 lookback = xcalloc (nLA, sizeof *lookback);
452 build_relations ();
453 compute_FOLLOWS ();
454 compute_lookahead_tokens ();
455
456 if (trace_flag & trace_sets)
457 lookahead_tokens_print (stderr);
458}
459
460
461void
462lalr_update_state_numbers (state_number old_to_new[], state_number nstates_old)
463{
464 goto_number ngotos_reachable = 0;
465 symbol_number nonterminal = 0;
466 aver (nsyms == nvars + ntokens);
467 {
468 goto_number i;
469 for (i = 0; i < ngotos; ++i)
470 {
471 while (i == goto_map[nonterminal])
472 goto_map[nonterminal++] = ngotos_reachable;
473 /* If old_to_new[from_state[i]] = nstates_old, remove this goto
474 entry. */
475 if (old_to_new[from_state[i]] != nstates_old)
476 {
477 /* from_state[i] is not removed, so it and thus to_state[i] are
478 reachable, so to_state[i] != nstates_old. */
479 aver (old_to_new[to_state[i]] != nstates_old);
480 from_state[ngotos_reachable] = old_to_new[from_state[i]];
481 to_state[ngotos_reachable] = old_to_new[to_state[i]];
482 ++ngotos_reachable;
483 }
484 }
485 }
486 while (nonterminal <= nvars)
487 {
488 aver (ngotos == goto_map[nonterminal]);
489 goto_map[nonterminal++] = ngotos_reachable;
490 }
491 ngotos = ngotos_reachable;
492}
493
494
495void
496lalr_free (void)
497{
498 state_number s;
499 for (s = 0; s < nstates; ++s)
500 states[s]->reductions->lookahead_tokens = NULL;
501 bitsetv_free (LA);
502}