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
|
/* Generate the LR(0) parser states for Bison.
Copyright (C) 1984, 1986, 1989, 2000-2002, 2004-2015, 2018-2019 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/>. */
/* See comments in state.h for the data structures that represent it.
The entry point is generate_states. */
#include <config.h>
#include "system.h"
#include <bitset.h>
#include "LR0.h"
#include "closure.h"
#include "complain.h"
#include "getargs.h"
#include "gram.h"
#include "lalr.h"
#include "reader.h"
#include "reduce.h"
#include "state.h"
#include "symtab.h"
typedef struct state_list
{
struct state_list *next;
state *state;
} state_list;
static state_list *first_state = NULL;
static state_list *last_state = NULL;
/*------------------------------------------------------------------.
| A state was just discovered from another state. Queue it for |
| later examination, in order to find its transitions. Return it. |
`------------------------------------------------------------------*/
static state *
state_list_append (symbol_number sym, size_t core_size, item_number *core)
{
state_list *node = xmalloc (sizeof *node);
state *s = state_new (sym, core_size, core);
if (trace_flag & trace_automaton)
fprintf (stderr, "state_list_append (state = %d, symbol = %d (%s))\n",
nstates, sym, symbols[sym]->tag);
node->next = NULL;
node->state = s;
if (!first_state)
first_state = node;
if (last_state)
last_state->next = node;
last_state = node;
return s;
}
static int nshifts;
static symbol_number *shift_symbol;
static rule **redset;
static state **shiftset;
static item_number **kernel_base;
static int *kernel_size;
static item_number *kernel_items;
static void
allocate_itemsets (void)
{
/* Count the number of occurrences of all the symbols in RITEMS.
Note that useless productions (hence useless nonterminals) are
browsed too, hence we need to allocate room for _all_ the
symbols. */
size_t count = 0;
size_t *symbol_count = xcalloc (nsyms + nuseless_nonterminals,
sizeof *symbol_count);
for (rule_number r = 0; r < nrules; ++r)
for (item_number *rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp)
{
count++;
symbol_count[*rhsp]++;
}
/* See comments before new_itemsets. All the vectors of items
live inside KERNEL_ITEMS. The number of active items after
some symbol S cannot be more than the number of times that S
appears as an item, which is SYMBOL_COUNT[S].
We allocate that much space for each symbol. */
kernel_base = xnmalloc (nsyms, sizeof *kernel_base);
kernel_items = xnmalloc (count, sizeof *kernel_items);
count = 0;
for (symbol_number i = 0; i < nsyms; i++)
{
kernel_base[i] = kernel_items + count;
count += symbol_count[i];
}
free (symbol_count);
kernel_size = xnmalloc (nsyms, sizeof *kernel_size);
}
static void
allocate_storage (void)
{
allocate_itemsets ();
shiftset = xnmalloc (nsyms, sizeof *shiftset);
redset = xnmalloc (nrules, sizeof *redset);
state_hash_new ();
shift_symbol = xnmalloc (nsyms, sizeof *shift_symbol);
}
static void
free_storage (void)
{
free (shift_symbol);
free (redset);
free (shiftset);
free (kernel_base);
free (kernel_size);
free (kernel_items);
state_hash_free ();
}
/*---------------------------------------------------------------.
| Find which symbols can be shifted in S, and for each one |
| record which items would be active after that shift. Uses the |
| contents of itemset. |
| |
| shift_symbol is set to a vector of the symbols that can be |
| shifted. For each symbol in the grammar, kernel_base[symbol] |
| points to a vector of item numbers activated if that symbol is |
| shifted, and kernel_size[symbol] is their numbers. |
| |
| itemset is sorted on item index in ritem, which is sorted on |
| rule number. Compute each kernel_base[symbol] with the same |
| sort. |
`---------------------------------------------------------------*/
static void
new_itemsets (state *s)
{
if (trace_flag & trace_automaton)
fprintf (stderr, "Entering new_itemsets, state = %d\n", s->number);
memset (kernel_size, 0, nsyms * sizeof *kernel_size);
nshifts = 0;
for (size_t i = 0; i < nitemset; ++i)
if (item_number_is_symbol_number (ritem[itemset[i]]))
{
symbol_number sym = item_number_as_symbol_number (ritem[itemset[i]]);
if (!kernel_size[sym])
{
shift_symbol[nshifts] = sym;
nshifts++;
}
kernel_base[sym][kernel_size[sym]] = itemset[i] + 1;
kernel_size[sym]++;
}
}
/*--------------------------------------------------------------.
| Find the state we would get to (from the current state) by |
| shifting SYM. Create a new state if no equivalent one exists |
| already. Used by append_states. |
`--------------------------------------------------------------*/
static state *
get_state (symbol_number sym, size_t core_size, item_number *core)
{
if (trace_flag & trace_automaton)
fprintf (stderr, "Entering get_state, symbol = %d (%s)\n",
sym, symbols[sym]->tag);
state *s = state_hash_lookup (core_size, core);
if (!s)
s = state_list_append (sym, core_size, core);
if (trace_flag & trace_automaton)
fprintf (stderr, "Exiting get_state => %d\n", s->number);
return s;
}
/*---------------------------------------------------------------.
| Use the information computed by new_itemsets to find the state |
| numbers reached by each shift transition from S. |
| |
| SHIFTSET is set up as a vector of those states. |
`---------------------------------------------------------------*/
static void
append_states (state *s)
{
if (trace_flag & trace_automaton)
fprintf (stderr, "Entering append_states, state = %d\n", s->number);
/* First sort shift_symbol into increasing order. */
for (int i = 1; i < nshifts; i++)
{
const symbol_number sym = shift_symbol[i];
int j = i;
for (; 0 < j && sym < shift_symbol[j - 1]; j--)
shift_symbol[j] = shift_symbol[j - 1];
shift_symbol[j] = sym;
}
for (int i = 0; i < nshifts; i++)
{
const symbol_number sym = shift_symbol[i];
shiftset[i] = get_state (sym, kernel_size[sym], kernel_base[sym]);
}
}
/*----------------------------------------------------------------.
| Find which rules can be used for reduction transitions from the |
| current state and make a reductions structure for the state to |
| record their rule numbers. |
`----------------------------------------------------------------*/
static void
save_reductions (state *s)
{
int count = 0;
/* Find and count the active items that represent ends of rules. */
for (size_t i = 0; i < nitemset; ++i)
{
item_number item = ritem[itemset[i]];
if (item_number_is_rule_number (item))
{
rule_number r = item_number_as_rule_number (item);
redset[count++] = &rules[r];
if (r == 0)
{
/* This is "reduce 0", i.e., accept. */
aver (!final_state);
final_state = s;
}
}
}
/* Make a reductions structure and copy the data into it. */
state_reductions_set (s, count, redset);
}
/*---------------.
| Build STATES. |
`---------------*/
static void
set_states (void)
{
states = xcalloc (nstates, sizeof *states);
while (first_state)
{
state_list *this = first_state;
/* Pessimization, but simplification of the code: make sure all
the states have valid transitions and reductions members,
even if reduced to 0. It is too soon for errs, which are
computed later, but set_conflicts. */
state *s = this->state;
if (!s->transitions)
state_transitions_set (s, 0, 0);
if (!s->reductions)
state_reductions_set (s, 0, 0);
states[s->number] = s;
first_state = this->next;
free (this);
}
first_state = NULL;
last_state = NULL;
}
/*-------------------------------------------------------------------.
| Compute the LR(0) parser states (see state.h for details) from the |
| grammar. |
`-------------------------------------------------------------------*/
void
generate_states (void)
{
allocate_storage ();
new_closure (nritems);
/* Create the initial state. The 0 at the lhs is the index of the
item of this initial rule. */
item_number initial_core = 0;
state_list_append (0, 1, &initial_core);
/* States are queued when they are created; process them all. */
for (state_list *list = first_state; list; list = list->next)
{
state *s = list->state;
if (trace_flag & trace_automaton)
fprintf (stderr, "Processing state %d (reached by %s)\n",
s->number,
symbols[s->accessing_symbol]->tag);
/* Set up itemset for the transitions out of this state. itemset gets a
vector of all the items that could be accepted next. */
closure (s->items, s->nitems);
/* Record the reductions allowed out of this state. */
save_reductions (s);
/* Find the itemsets of the states that shifts can reach. */
new_itemsets (s);
/* Find or create the core structures for those states. */
append_states (s);
/* Create the shifts structures for the shifts to those states,
now that the state numbers transitioning to are known. */
state_transitions_set (s, nshifts, shiftset);
}
/* discard various storage */
free_closure ();
free_storage ();
/* Set up STATES. */
set_states ();
}
|