]>
Commit | Line | Data |
---|---|---|
1 | /* Bison Action Scanner -*- C -*- | |
2 | ||
3 | Copyright (C) 2006-2012 Free Software Foundation, Inc. | |
4 | ||
5 | This file is part of Bison, the GNU Compiler Compiler. | |
6 | ||
7 | This program is free software: you can redistribute it and/or modify | |
8 | it under the terms of the GNU General Public License as published by | |
9 | the Free Software Foundation, either version 3 of the License, or | |
10 | (at your option) any later version. | |
11 | ||
12 | This program is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
18 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ | |
19 | ||
20 | %option debug nodefault noinput nounput noyywrap never-interactive | |
21 | %option prefix="code_" outfile="lex.yy.c" | |
22 | ||
23 | %{ | |
24 | /* Work around a bug in flex 2.5.31. See Debian bug 333231 | |
25 | <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=333231>. */ | |
26 | #undef code_wrap | |
27 | #define code_wrap() 1 | |
28 | ||
29 | #define FLEX_PREFIX(Id) code_ ## Id | |
30 | #include <src/flex-scanner.h> | |
31 | ||
32 | #include <src/complain.h> | |
33 | #include <src/reader.h> | |
34 | #include <src/getargs.h> | |
35 | #include <src/muscle-tab.h> | |
36 | #include <src/scan-code.h> | |
37 | #include <src/symlist.h> | |
38 | ||
39 | #include <c-ctype.h> | |
40 | #include <get-errno.h> | |
41 | #include <quote.h> | |
42 | ||
43 | /* The current calling start condition: SC_RULE_ACTION or | |
44 | SC_SYMBOL_ACTION. */ | |
45 | # define YY_DECL static char *code_lex (code_props *self, int sc_context) | |
46 | YY_DECL; | |
47 | ||
48 | #define YY_USER_ACTION location_compute (loc, &loc->end, yytext, yyleng); | |
49 | ||
50 | static char *fetch_type_name (char *cp, char const **type_name, | |
51 | location dollar_loc); | |
52 | ||
53 | static void handle_action_dollar (symbol_list *rule, char *cp, | |
54 | location dollar_loc); | |
55 | static void handle_action_at (symbol_list *rule, char *cp, location at_loc); | |
56 | ||
57 | /* A string to be pushed to obstack after dollar/at has been handled. */ | |
58 | static char *ref_tail_fields; | |
59 | ||
60 | static location the_location; | |
61 | static location *loc = &the_location; | |
62 | ||
63 | /* A string representing the most recent translation. */ | |
64 | static char *last_string; | |
65 | ||
66 | /* True if an untyped $$ or $n was seen. */ | |
67 | static bool untyped_var_seen; | |
68 | ||
69 | %} | |
70 | /* C and C++ comments in code. */ | |
71 | %x SC_COMMENT SC_LINE_COMMENT | |
72 | /* Strings and characters in code. */ | |
73 | %x SC_STRING SC_CHARACTER | |
74 | /* Whether in a rule or symbol action. Specifies the translation | |
75 | of $ and @. */ | |
76 | %x SC_RULE_ACTION SC_SYMBOL_ACTION | |
77 | ||
78 | ||
79 | /* POSIX says that a tag must be both an id and a C union member, but | |
80 | historically almost any character is allowed in a tag. We disallow | |
81 | NUL and newline, as this simplifies our implementation. */ | |
82 | tag [^\0\n>]+ | |
83 | ||
84 | /* Zero or more instances of backslash-newline. Following GCC, allow | |
85 | white space between the backslash and the newline. */ | |
86 | splice (\\[ \f\t\v]*\n)* | |
87 | ||
88 | /* C style identifier. Must start with letter. Will be used for | |
89 | named symbol references. Shall be kept synchronized with | |
90 | scan-gram.l "letter" and "id". */ | |
91 | letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] | |
92 | id {letter}({letter}|[-0-9])* | |
93 | ref -?[0-9]+|{id}|"["{id}"]"|"$" | |
94 | ||
95 | %% | |
96 | ||
97 | %{ | |
98 | /* Nesting level of the current code in braces. */ | |
99 | int braces_level = 0; | |
100 | ||
101 | /* Whether a semicolon is probably needed. | |
102 | ||
103 | The heuristic is that a semicolon is not needed after '{', '}', | |
104 | ';', or a C preprocessor directive, and that whitespaces and | |
105 | comments do not affect this flag. Note that '{' does not need a | |
106 | semicolon because of '{}'. A semicolon may be needed before a | |
107 | cpp directive, but don't bother. | |
108 | ||
109 | While it is maintained in several start-conditions (factoring | |
110 | opportunities), it is meaningful only for SC_RULE_ACTION. */ | |
111 | bool need_semicolon = false; | |
112 | ||
113 | /* Whether in a C preprocessor directive. Don't use a start condition | |
114 | for this because, at the end of strings and comments, we still need | |
115 | to know whether we're in a directive. */ | |
116 | bool in_cpp = false; | |
117 | ||
118 | /* This scanner is special: it is invoked only once, henceforth | |
119 | is expected to return only once. This initialization is | |
120 | therefore done once per action to translate. */ | |
121 | aver (sc_context == SC_SYMBOL_ACTION | |
122 | || sc_context == SC_RULE_ACTION | |
123 | || sc_context == INITIAL); | |
124 | BEGIN sc_context; | |
125 | %} | |
126 | ||
127 | /*------------------------------------------------------------. | |
128 | | Scanning a C comment. The initial '/ *' is already eaten. | | |
129 | `------------------------------------------------------------*/ | |
130 | ||
131 | <SC_COMMENT> | |
132 | { | |
133 | "*"{splice}"/" STRING_GROW; BEGIN sc_context; | |
134 | } | |
135 | ||
136 | ||
137 | /*--------------------------------------------------------------. | |
138 | | Scanning a line comment. The initial '//' is already eaten. | | |
139 | `--------------------------------------------------------------*/ | |
140 | ||
141 | <SC_LINE_COMMENT> | |
142 | { | |
143 | "\n" STRING_GROW; BEGIN sc_context; | |
144 | {splice} STRING_GROW; | |
145 | } | |
146 | ||
147 | ||
148 | /*--------------------------------------------. | |
149 | | Scanning user-code characters and strings. | | |
150 | `--------------------------------------------*/ | |
151 | ||
152 | <SC_CHARACTER,SC_STRING> | |
153 | { | |
154 | {splice}|\\{splice}. STRING_GROW; | |
155 | } | |
156 | ||
157 | <SC_CHARACTER> | |
158 | { | |
159 | "'" STRING_GROW; BEGIN sc_context; | |
160 | } | |
161 | ||
162 | <SC_STRING> | |
163 | { | |
164 | "\"" STRING_GROW; BEGIN sc_context; | |
165 | } | |
166 | ||
167 | ||
168 | <SC_RULE_ACTION,SC_SYMBOL_ACTION> | |
169 | { | |
170 | "'" { | |
171 | STRING_GROW; | |
172 | BEGIN SC_CHARACTER; | |
173 | need_semicolon = true; | |
174 | } | |
175 | "\"" { | |
176 | STRING_GROW; | |
177 | BEGIN SC_STRING; | |
178 | need_semicolon = true; | |
179 | } | |
180 | "/"{splice}"*" { | |
181 | STRING_GROW; | |
182 | BEGIN SC_COMMENT; | |
183 | } | |
184 | "/"{splice}"/" { | |
185 | STRING_GROW; | |
186 | BEGIN SC_LINE_COMMENT; | |
187 | } | |
188 | [$@] { | |
189 | complain_at (*loc, Wother, _("stray '%s'"), yytext); | |
190 | obstack_escape (&obstack_for_string, yytext); | |
191 | need_semicolon = true; | |
192 | } | |
193 | [\[\]] { | |
194 | obstack_escape (&obstack_for_string, yytext); | |
195 | need_semicolon = true; | |
196 | } | |
197 | } | |
198 | ||
199 | <SC_RULE_ACTION> | |
200 | { | |
201 | "$"("<"{tag}">")?{ref} { | |
202 | ref_tail_fields = NULL; | |
203 | handle_action_dollar (self->rule, yytext, *loc); | |
204 | if (ref_tail_fields) | |
205 | obstack_sgrow (&obstack_for_string, ref_tail_fields); | |
206 | need_semicolon = true; | |
207 | } | |
208 | "@"{ref} { | |
209 | ref_tail_fields = NULL; | |
210 | handle_action_at (self->rule, yytext, *loc); | |
211 | if (ref_tail_fields) | |
212 | obstack_sgrow (&obstack_for_string, ref_tail_fields); | |
213 | need_semicolon = true; | |
214 | } | |
215 | ||
216 | ";" STRING_GROW; need_semicolon = false; | |
217 | "{" STRING_GROW; ++braces_level; need_semicolon = false; | |
218 | "}" { | |
219 | bool outer_brace = --braces_level == 0; | |
220 | ||
221 | /* As an undocumented Bison extension, append ';' before the last | |
222 | brace in braced code, so that the user code can omit trailing | |
223 | ';'. But do not append ';' if emulating Yacc, since Yacc does | |
224 | not append one. */ | |
225 | if (outer_brace && !yacc_flag && language_prio == default_prio | |
226 | && skeleton_prio == default_prio && need_semicolon && ! in_cpp) | |
227 | { | |
228 | complain_at (*loc, Wother, | |
229 | _("a ';' might be needed at the end of action code")); | |
230 | complain_at (*loc, Wother, | |
231 | _("future versions of Bison will not add the ';'")); | |
232 | obstack_1grow (&obstack_for_string, ';'); | |
233 | } | |
234 | ||
235 | STRING_GROW; | |
236 | need_semicolon = false; | |
237 | } | |
238 | ||
239 | /* Preprocessing directives should only be recognized at the beginning | |
240 | of lines, allowing whitespace including comments, but in C/C++, | |
241 | '#' can only be the start of preprocessor directives or within | |
242 | '#define' directives anyway, so don't bother with begin of line. */ | |
243 | "#" STRING_GROW; in_cpp = true; | |
244 | ||
245 | {splice} STRING_GROW; | |
246 | [\n\r] STRING_GROW; if (in_cpp) in_cpp = need_semicolon = false; | |
247 | [ \t\f] STRING_GROW; | |
248 | . STRING_GROW; need_semicolon = true; | |
249 | } | |
250 | ||
251 | <SC_SYMBOL_ACTION> | |
252 | { | |
253 | "$"("<"{tag}">")?"$" { | |
254 | const char *type_name = NULL; | |
255 | fetch_type_name (yytext + 1, &type_name, *loc)[-1] = 0; | |
256 | obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar("); | |
257 | obstack_quote (&obstack_for_string, type_name); | |
258 | obstack_sgrow (&obstack_for_string, ")["); | |
259 | self->is_value_used = true; | |
260 | } | |
261 | "@$" { | |
262 | obstack_sgrow (&obstack_for_string, "]b4_at_dollar["); | |
263 | muscle_percent_define_ensure("locations", the_location, true); | |
264 | } | |
265 | } | |
266 | ||
267 | ||
268 | <*> | |
269 | { | |
270 | /* Escape M4 quoting characters in C code. */ | |
271 | [$@\[\]] obstack_escape (&obstack_for_string, yytext); | |
272 | ||
273 | /* By default, grow the string obstack with the input. */ | |
274 | .|\n STRING_GROW; | |
275 | ||
276 | /* End of processing. */ | |
277 | <<EOF>> STRING_FINISH; return last_string; | |
278 | } | |
279 | ||
280 | %% | |
281 | ||
282 | static inline bool | |
283 | is_dot_or_dash (char ch) | |
284 | { | |
285 | return ch == '.' || ch == '-'; | |
286 | } | |
287 | ||
288 | static inline bool | |
289 | contains_dot_or_dash (const char* p) | |
290 | { | |
291 | for (; *p; ++p) | |
292 | if (is_dot_or_dash (*p)) | |
293 | return true; | |
294 | return false; | |
295 | } | |
296 | ||
297 | /* Defines a variant of a symbolic name resolution. */ | |
298 | typedef struct | |
299 | { | |
300 | /* Index in symbol list. */ | |
301 | unsigned symbol_index; | |
302 | ||
303 | /* Matched symbol id and loc. */ | |
304 | uniqstr id; | |
305 | location loc; | |
306 | ||
307 | /* Hiding named reference. */ | |
308 | named_ref* hidden_by; | |
309 | ||
310 | /* Error flags. May contain zero (no errors) or | |
311 | a combination of VARIANT_* values. */ | |
312 | unsigned err; | |
313 | } variant; | |
314 | ||
315 | /* Set when the variant refers to a symbol hidden | |
316 | by an explicit symbol reference. */ | |
317 | #define VARIANT_HIDDEN (1 << 0) | |
318 | ||
319 | /* Set when the variant refers to a symbol containing | |
320 | dots or dashes. Will require explicit bracketing. */ | |
321 | #define VARIANT_BAD_BRACKETING (1 << 1) | |
322 | ||
323 | /* Set when the variant refers to a symbol which is | |
324 | not visible from current midrule. */ | |
325 | #define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 2) | |
326 | ||
327 | static variant *variant_table = NULL; | |
328 | static unsigned variant_table_size = 0; | |
329 | static unsigned variant_count = 0; | |
330 | ||
331 | static variant * | |
332 | variant_table_grow (void) | |
333 | { | |
334 | ++variant_count; | |
335 | if (variant_count > variant_table_size) | |
336 | { | |
337 | while (variant_count > variant_table_size) | |
338 | variant_table_size = 2 * variant_table_size + 3; | |
339 | variant_table = xnrealloc (variant_table, variant_table_size, | |
340 | sizeof *variant_table); | |
341 | } | |
342 | return &variant_table[variant_count - 1]; | |
343 | } | |
344 | ||
345 | static void | |
346 | variant_table_free (void) | |
347 | { | |
348 | free (variant_table); | |
349 | variant_table = NULL; | |
350 | variant_table_size = variant_count = 0; | |
351 | } | |
352 | ||
353 | static char * | |
354 | find_prefix_end (const char *prefix, char *begin, char *end) | |
355 | { | |
356 | char *ptr = begin; | |
357 | ||
358 | for (; *prefix && ptr != end; ++prefix, ++ptr) | |
359 | if (*prefix != *ptr) | |
360 | return 0; | |
361 | ||
362 | if (*prefix) | |
363 | return 0; | |
364 | ||
365 | return ptr; | |
366 | } | |
367 | ||
368 | static variant * | |
369 | variant_add (uniqstr id, location id_loc, unsigned symbol_index, | |
370 | char *cp, char *cp_end, bool explicit_bracketing) | |
371 | { | |
372 | char *prefix_end; | |
373 | ||
374 | prefix_end = find_prefix_end (id, cp, cp_end); | |
375 | if (prefix_end && | |
376 | (prefix_end == cp_end || | |
377 | (!explicit_bracketing && is_dot_or_dash (*prefix_end)))) | |
378 | { | |
379 | variant *r = variant_table_grow (); | |
380 | r->symbol_index = symbol_index; | |
381 | r->id = id; | |
382 | r->loc = id_loc; | |
383 | r->hidden_by = NULL; | |
384 | r->err = 0; | |
385 | return r; | |
386 | } | |
387 | else | |
388 | return NULL; | |
389 | } | |
390 | ||
391 | static const char * | |
392 | get_at_spec(unsigned symbol_index) | |
393 | { | |
394 | static char at_buf[20]; | |
395 | if (symbol_index == 0) | |
396 | strcpy (at_buf, "$$"); | |
397 | else | |
398 | snprintf (at_buf, sizeof at_buf, "$%u", symbol_index); | |
399 | return at_buf; | |
400 | } | |
401 | ||
402 | static void | |
403 | show_sub_message (const char* cp, bool explicit_bracketing, | |
404 | int midrule_rhs_index, char dollar_or_at, | |
405 | warnings wflags, unsigned indent, | |
406 | const variant *var) | |
407 | { | |
408 | const char *at_spec = get_at_spec (var->symbol_index); | |
409 | ||
410 | if (var->err == 0) | |
411 | complain_at_indent (var->loc, wflags, &indent, | |
412 | _("refers to: %c%s at %s"), dollar_or_at, | |
413 | var->id, at_spec); | |
414 | else | |
415 | { | |
416 | static struct obstack msg_buf; | |
417 | const char *tail = explicit_bracketing ? "" : cp + strlen (var->id); | |
418 | const char *id = var->hidden_by ? var->hidden_by->id : var->id; | |
419 | location id_loc = var->hidden_by ? var->hidden_by->loc : var->loc; | |
420 | ||
421 | /* Create the explanation message. */ | |
422 | obstack_init (&msg_buf); | |
423 | ||
424 | obstack_printf (&msg_buf, _("possibly meant: %c"), dollar_or_at); | |
425 | if (contains_dot_or_dash (id)) | |
426 | obstack_printf (&msg_buf, "[%s]", id); | |
427 | else | |
428 | obstack_sgrow (&msg_buf, id); | |
429 | obstack_sgrow (&msg_buf, tail); | |
430 | ||
431 | if (var->err & VARIANT_HIDDEN) | |
432 | { | |
433 | obstack_printf (&msg_buf, _(", hiding %c"), dollar_or_at); | |
434 | if (contains_dot_or_dash (var->id)) | |
435 | obstack_printf (&msg_buf, "[%s]", var->id); | |
436 | else | |
437 | obstack_sgrow (&msg_buf, var->id); | |
438 | obstack_sgrow (&msg_buf, tail); | |
439 | } | |
440 | ||
441 | obstack_printf (&msg_buf, _(" at %s"), at_spec); | |
442 | ||
443 | if (var->err & VARIANT_NOT_VISIBLE_FROM_MIDRULE) | |
444 | obstack_printf (&msg_buf, | |
445 | _(", cannot be accessed from mid-rule action at $%d"), | |
446 | midrule_rhs_index); | |
447 | ||
448 | obstack_1grow (&msg_buf, '\0'); | |
449 | complain_at_indent (id_loc, wflags, &indent, "%s", | |
450 | (char *) obstack_finish (&msg_buf)); | |
451 | obstack_free (&msg_buf, 0); | |
452 | } | |
453 | } | |
454 | ||
455 | static void | |
456 | show_sub_messages (const char* cp, bool explicit_bracketing, | |
457 | int midrule_rhs_index, char dollar_or_at, | |
458 | warnings wflags, unsigned indent) | |
459 | { | |
460 | unsigned i; | |
461 | ||
462 | for (i = 0; i < variant_count; ++i) | |
463 | show_sub_message (cp, explicit_bracketing, | |
464 | midrule_rhs_index, dollar_or_at, | |
465 | wflags, indent, &variant_table[i]); | |
466 | } | |
467 | ||
468 | /* Returned from "parse_ref" when the reference | |
469 | is inappropriate. */ | |
470 | #define INVALID_REF (INT_MIN) | |
471 | ||
472 | /* Returned from "parse_ref" when the reference | |
473 | points to LHS ($$) of the current rule or midrule. */ | |
474 | #define LHS_REF (INT_MIN + 1) | |
475 | ||
476 | /* Sub-messages indent. */ | |
477 | #define SUB_INDENT (4) | |
478 | ||
479 | /* Parse named or positional reference. In case of positional | |
480 | references, can return negative values for $-n "deep" stack | |
481 | accesses. */ | |
482 | static long int | |
483 | parse_ref (char *cp, symbol_list *rule, int rule_length, | |
484 | int midrule_rhs_index, char *text, location text_loc, | |
485 | char dollar_or_at) | |
486 | { | |
487 | symbol_list *l; | |
488 | char *cp_end; | |
489 | bool explicit_bracketing; | |
490 | unsigned i; | |
491 | unsigned valid_variants = 0; | |
492 | unsigned valid_variant_index = 0; | |
493 | ||
494 | if ('$' == *cp) | |
495 | return LHS_REF; | |
496 | ||
497 | if (c_isdigit (*cp) || (*cp == '-' && c_isdigit (* (cp + 1)))) | |
498 | { | |
499 | long int num = strtol (cp, &cp, 10); | |
500 | if (1 - INT_MAX + rule_length <= num && num <= rule_length) | |
501 | return num; | |
502 | else | |
503 | { | |
504 | complain_at (text_loc, complaint, _("integer out of range: %s"), | |
505 | quote (text)); | |
506 | return INVALID_REF; | |
507 | } | |
508 | } | |
509 | ||
510 | if ('[' == *cp) | |
511 | { | |
512 | /* Ignore the brackets. */ | |
513 | char *p; | |
514 | for (p = ++cp; *p != ']'; ++p) | |
515 | continue; | |
516 | cp_end = p; | |
517 | ||
518 | explicit_bracketing = true; | |
519 | } | |
520 | else | |
521 | { | |
522 | /* Take all characters of the name. */ | |
523 | char* p; | |
524 | for (p = cp; *p; ++p) | |
525 | if (is_dot_or_dash (*p)) | |
526 | { | |
527 | ref_tail_fields = p; | |
528 | break; | |
529 | } | |
530 | for (p = cp; *p; ++p) | |
531 | continue; | |
532 | cp_end = p; | |
533 | ||
534 | explicit_bracketing = false; | |
535 | } | |
536 | ||
537 | /* Add all relevant variants. */ | |
538 | { | |
539 | unsigned symbol_index; | |
540 | variant_count = 0; | |
541 | for (symbol_index = 0, l = rule; !symbol_list_null (l); | |
542 | ++symbol_index, l = l->next) | |
543 | { | |
544 | variant *var; | |
545 | if (l->content_type != SYMLIST_SYMBOL) | |
546 | continue; | |
547 | ||
548 | var = variant_add (l->content.sym->tag, l->sym_loc, | |
549 | symbol_index, cp, cp_end, explicit_bracketing); | |
550 | if (var && l->named_ref) | |
551 | var->hidden_by = l->named_ref; | |
552 | ||
553 | if (l->named_ref) | |
554 | variant_add (l->named_ref->id, l->named_ref->loc, | |
555 | symbol_index, cp, cp_end, explicit_bracketing); | |
556 | } | |
557 | } | |
558 | ||
559 | /* Check errors. */ | |
560 | for (i = 0; i < variant_count; ++i) | |
561 | { | |
562 | variant *var = &variant_table[i]; | |
563 | unsigned symbol_index = var->symbol_index; | |
564 | ||
565 | /* Check visibility from mid-rule actions. */ | |
566 | if (midrule_rhs_index != 0 | |
567 | && (symbol_index == 0 || midrule_rhs_index < symbol_index)) | |
568 | var->err |= VARIANT_NOT_VISIBLE_FROM_MIDRULE; | |
569 | ||
570 | /* Check correct bracketing. */ | |
571 | if (!explicit_bracketing && contains_dot_or_dash (var->id)) | |
572 | var->err |= VARIANT_BAD_BRACKETING; | |
573 | ||
574 | /* Check using of hidden symbols. */ | |
575 | if (var->hidden_by) | |
576 | var->err |= VARIANT_HIDDEN; | |
577 | ||
578 | if (!var->err) | |
579 | { | |
580 | valid_variant_index = i; | |
581 | ++valid_variants; | |
582 | } | |
583 | } | |
584 | ||
585 | switch (valid_variants) | |
586 | { | |
587 | case 0: | |
588 | { | |
589 | unsigned len = (explicit_bracketing || !ref_tail_fields) ? | |
590 | cp_end - cp : ref_tail_fields - cp; | |
591 | unsigned indent = 0; | |
592 | ||
593 | complain_at_indent (text_loc, complaint, &indent, | |
594 | _("invalid reference: %s"), quote (text)); | |
595 | indent += SUB_INDENT; | |
596 | if (len == 0) | |
597 | { | |
598 | location sym_loc = text_loc; | |
599 | sym_loc.start.column += 1; | |
600 | sym_loc.end = sym_loc.start; | |
601 | complain_at_indent (sym_loc, complaint, &indent, | |
602 | _("syntax error after '%c', expecting integer, " | |
603 | "letter, '_', '[', or '$'"), | |
604 | dollar_or_at); | |
605 | } | |
606 | else if (midrule_rhs_index) | |
607 | complain_at_indent (rule->location, complaint, &indent, | |
608 | _("symbol not found in production before $%d: " | |
609 | "%.*s"), | |
610 | midrule_rhs_index, len, cp); | |
611 | else | |
612 | complain_at_indent (rule->location, complaint, &indent, | |
613 | _("symbol not found in production: %.*s"), | |
614 | len, cp); | |
615 | ||
616 | if (variant_count > 0) | |
617 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, | |
618 | dollar_or_at, complaint, indent); | |
619 | return INVALID_REF; | |
620 | } | |
621 | case 1: | |
622 | { | |
623 | unsigned indent = 0; | |
624 | if (variant_count > 1) | |
625 | { | |
626 | complain_at_indent (text_loc, Wother, &indent, | |
627 | _("misleading reference: %s"), quote (text)); | |
628 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, | |
629 | dollar_or_at, Wother, indent + SUB_INDENT); | |
630 | } | |
631 | { | |
632 | unsigned symbol_index = | |
633 | variant_table[valid_variant_index].symbol_index; | |
634 | return (symbol_index == midrule_rhs_index) ? LHS_REF : symbol_index; | |
635 | } | |
636 | } | |
637 | case 2: | |
638 | default: | |
639 | { | |
640 | unsigned indent = 0; | |
641 | complain_at_indent (text_loc, complaint, &indent, | |
642 | _("ambiguous reference: %s"), quote (text)); | |
643 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, | |
644 | dollar_or_at, complaint, indent + SUB_INDENT); | |
645 | return INVALID_REF; | |
646 | } | |
647 | } | |
648 | ||
649 | /* Not reachable. */ | |
650 | return INVALID_REF; | |
651 | } | |
652 | ||
653 | /* Keeps track of the maximum number of semantic values to the left of | |
654 | a handle (those referenced by $0, $-1, etc.) are required by the | |
655 | semantic actions of this grammar. */ | |
656 | int max_left_semantic_context = 0; | |
657 | ||
658 | ||
659 | /* If CP points to a typename (i.e., <.*?>), set TYPE_NAME to its | |
660 | beginning (i.e., after the opening "<", and return the pointer | |
661 | immediately after it. */ | |
662 | ||
663 | static | |
664 | char * | |
665 | fetch_type_name (char *cp, char const **type_name, | |
666 | location dollar_loc) | |
667 | { | |
668 | if (*cp == '<') | |
669 | { | |
670 | *type_name = ++cp; | |
671 | while (*cp != '>') | |
672 | ++cp; | |
673 | ||
674 | /* The '>' symbol will be later replaced by '\0'. Original | |
675 | 'text' is needed for error messages. */ | |
676 | ++cp; | |
677 | if (untyped_var_seen) | |
678 | complain_at (dollar_loc, complaint, | |
679 | _("explicit type given in untyped grammar")); | |
680 | tag_seen = true; | |
681 | } | |
682 | return cp; | |
683 | } | |
684 | ||
685 | /*------------------------------------------------------------------. | |
686 | | TEXT is pointing to a wannabee semantic value (i.e., a '$'). | | |
687 | | | | |
688 | | Possible inputs: $[<TYPENAME>]($|integer) | | |
689 | | | | |
690 | | Output to OBSTACK_FOR_STRING a reference to this semantic value. | | |
691 | `------------------------------------------------------------------*/ | |
692 | ||
693 | static void | |
694 | handle_action_dollar (symbol_list *rule, char *text, location dollar_loc) | |
695 | { | |
696 | char const *type_name = NULL; | |
697 | char *cp = text + 1; | |
698 | symbol_list *effective_rule; | |
699 | int effective_rule_length; | |
700 | int n; | |
701 | ||
702 | if (rule->midrule_parent_rule) | |
703 | { | |
704 | effective_rule = rule->midrule_parent_rule; | |
705 | effective_rule_length = rule->midrule_parent_rhs_index - 1; | |
706 | } | |
707 | else | |
708 | { | |
709 | effective_rule = rule; | |
710 | effective_rule_length = symbol_list_length (rule->next); | |
711 | } | |
712 | ||
713 | /* Get the type name if explicit. */ | |
714 | cp = fetch_type_name (cp, &type_name, dollar_loc); | |
715 | ||
716 | n = parse_ref (cp, effective_rule, effective_rule_length, | |
717 | rule->midrule_parent_rhs_index, text, dollar_loc, '$'); | |
718 | ||
719 | /* End type_name. */ | |
720 | if (type_name) | |
721 | cp[-1] = '\0'; | |
722 | ||
723 | switch (n) | |
724 | { | |
725 | case INVALID_REF: | |
726 | break; | |
727 | ||
728 | case LHS_REF: | |
729 | if (!type_name) | |
730 | type_name = symbol_list_n_type_name_get (rule, dollar_loc, 0); | |
731 | ||
732 | if (!type_name) | |
733 | { | |
734 | if (union_seen | tag_seen) | |
735 | { | |
736 | if (rule->midrule_parent_rule) | |
737 | complain_at (dollar_loc, complaint, | |
738 | _("$$ for the midrule at $%d of %s" | |
739 | " has no declared type"), | |
740 | rule->midrule_parent_rhs_index, | |
741 | quote (effective_rule->content.sym->tag)); | |
742 | else | |
743 | complain_at (dollar_loc, complaint, | |
744 | _("$$ of %s has no declared type"), | |
745 | quote (rule->content.sym->tag)); | |
746 | } | |
747 | else | |
748 | untyped_var_seen = true; | |
749 | } | |
750 | ||
751 | obstack_sgrow (&obstack_for_string, "]b4_lhs_value("); | |
752 | obstack_quote (&obstack_for_string, type_name); | |
753 | obstack_sgrow (&obstack_for_string, ")["); | |
754 | rule->action_props.is_value_used = true; | |
755 | break; | |
756 | ||
757 | default: | |
758 | if (max_left_semantic_context < 1 - n) | |
759 | max_left_semantic_context = 1 - n; | |
760 | if (!type_name && 0 < n) | |
761 | type_name = | |
762 | symbol_list_n_type_name_get (effective_rule, dollar_loc, n); | |
763 | if (!type_name) | |
764 | { | |
765 | if (union_seen | tag_seen) | |
766 | complain_at (dollar_loc, complaint, | |
767 | _("$%s of %s has no declared type"), cp, | |
768 | quote (effective_rule->content.sym->tag)); | |
769 | else | |
770 | untyped_var_seen = true; | |
771 | } | |
772 | ||
773 | obstack_printf (&obstack_for_string, | |
774 | "]b4_rhs_value(%d, %d, ", effective_rule_length, n); | |
775 | obstack_quote (&obstack_for_string, type_name); | |
776 | obstack_sgrow (&obstack_for_string, ")["); | |
777 | if (n > 0) | |
778 | symbol_list_n_get (effective_rule, n)->action_props.is_value_used = | |
779 | true; | |
780 | break; | |
781 | } | |
782 | } | |
783 | ||
784 | ||
785 | /*------------------------------------------------------. | |
786 | | TEXT is a location token (i.e., a '@...'). Output to | | |
787 | | OBSTACK_FOR_STRING a reference to this location. | | |
788 | `------------------------------------------------------*/ | |
789 | ||
790 | static void | |
791 | handle_action_at (symbol_list *rule, char *text, location at_loc) | |
792 | { | |
793 | char *cp = text + 1; | |
794 | symbol_list *effective_rule; | |
795 | int effective_rule_length; | |
796 | int n; | |
797 | ||
798 | if (rule->midrule_parent_rule) | |
799 | { | |
800 | effective_rule = rule->midrule_parent_rule; | |
801 | effective_rule_length = rule->midrule_parent_rhs_index - 1; | |
802 | } | |
803 | else | |
804 | { | |
805 | effective_rule = rule; | |
806 | effective_rule_length = symbol_list_length (rule->next); | |
807 | } | |
808 | ||
809 | muscle_percent_define_ensure("locations", at_loc, true); | |
810 | ||
811 | n = parse_ref (cp, effective_rule, effective_rule_length, | |
812 | rule->midrule_parent_rhs_index, text, at_loc, '@'); | |
813 | switch (n) | |
814 | { | |
815 | case INVALID_REF: | |
816 | break; | |
817 | ||
818 | case LHS_REF: | |
819 | obstack_sgrow (&obstack_for_string, "]b4_lhs_location["); | |
820 | break; | |
821 | ||
822 | default: | |
823 | obstack_printf (&obstack_for_string, "]b4_rhs_location(%d, %d)[", | |
824 | effective_rule_length, n); | |
825 | break; | |
826 | } | |
827 | } | |
828 | ||
829 | ||
830 | /*-------------------------. | |
831 | | Initialize the scanner. | | |
832 | `-------------------------*/ | |
833 | ||
834 | /* Translate the dollars and ats in \a self, in the context \a sc_context | |
835 | (SC_RULE_ACTION, SC_SYMBOL_ACTION, INITIAL). */ | |
836 | ||
837 | static char const * | |
838 | translate_action (code_props *self, int sc_context) | |
839 | { | |
840 | char *res; | |
841 | static bool initialized = false; | |
842 | if (!initialized) | |
843 | { | |
844 | obstack_init (&obstack_for_string); | |
845 | yy_flex_debug = 0; | |
846 | initialized = true; | |
847 | } | |
848 | ||
849 | loc->start = loc->end = self->location.start; | |
850 | yy_switch_to_buffer (yy_scan_string (self->code)); | |
851 | res = code_lex (self, sc_context); | |
852 | yy_delete_buffer (YY_CURRENT_BUFFER); | |
853 | ||
854 | return res; | |
855 | } | |
856 | ||
857 | /*------------------------------------------------------------------------. | |
858 | | Implementation of the public interface as documented in "scan-code.h". | | |
859 | `------------------------------------------------------------------------*/ | |
860 | ||
861 | void | |
862 | code_props_none_init (code_props *self) | |
863 | { | |
864 | *self = code_props_none; | |
865 | } | |
866 | ||
867 | code_props code_props_none = CODE_PROPS_NONE_INIT; | |
868 | ||
869 | void | |
870 | code_props_plain_init (code_props *self, char const *code, | |
871 | location code_loc) | |
872 | { | |
873 | code_props_none_init (self); | |
874 | self->kind = CODE_PROPS_PLAIN; | |
875 | self->code = code; | |
876 | self->location = code_loc; | |
877 | } | |
878 | ||
879 | void | |
880 | code_props_symbol_action_init (code_props *self, char const *code, | |
881 | location code_loc) | |
882 | { | |
883 | code_props_none_init (self); | |
884 | self->kind = CODE_PROPS_SYMBOL_ACTION; | |
885 | self->code = code; | |
886 | self->location = code_loc; | |
887 | } | |
888 | ||
889 | void | |
890 | code_props_rule_action_init (code_props *self, char const *code, | |
891 | location code_loc, symbol_list *rule, | |
892 | named_ref *name, bool is_predicate) | |
893 | { | |
894 | code_props_none_init (self); | |
895 | self->kind = CODE_PROPS_RULE_ACTION; | |
896 | self->code = code; | |
897 | self->location = code_loc; | |
898 | self->rule = rule; | |
899 | self->named_ref = name; | |
900 | self->is_predicate = is_predicate; | |
901 | } | |
902 | ||
903 | void | |
904 | code_props_translate_code (code_props *self) | |
905 | { | |
906 | switch (self->kind) | |
907 | { | |
908 | case CODE_PROPS_NONE: | |
909 | break; | |
910 | case CODE_PROPS_PLAIN: | |
911 | self->code = translate_action (self, INITIAL); | |
912 | break; | |
913 | case CODE_PROPS_SYMBOL_ACTION: | |
914 | self->code = translate_action (self, SC_SYMBOL_ACTION); | |
915 | break; | |
916 | case CODE_PROPS_RULE_ACTION: | |
917 | self->code = translate_action (self, SC_RULE_ACTION); | |
918 | break; | |
919 | } | |
920 | } | |
921 | ||
922 | void | |
923 | code_scanner_last_string_free (void) | |
924 | { | |
925 | STRING_FREE; | |
926 | } | |
927 | ||
928 | void | |
929 | code_scanner_free (void) | |
930 | { | |
931 | obstack_free (&obstack_for_string, 0); | |
932 | variant_table_free (); | |
933 | ||
934 | /* Reclaim Flex's buffers. */ | |
935 | yylex_destroy (); | |
936 | } |