]>
Commit | Line | Data |
---|---|---|
e9071366 AD |
1 | /* Bison Action Scanner -*- C -*- |
2 | ||
7d424de1 PE |
3 | Copyright (C) 2006, 2007, 2008, 2009, 2010 Free Software |
4 | Foundation, Inc. | |
e9071366 AD |
5 | |
6 | This file is part of Bison, the GNU Compiler Compiler. | |
7 | ||
f16b0819 | 8 | This program is free software: you can redistribute it and/or modify |
e9071366 | 9 | it under the terms of the GNU General Public License as published by |
f16b0819 | 10 | the Free Software Foundation, either version 3 of the License, or |
e9071366 AD |
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 | |
f16b0819 | 19 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
e9071366 | 20 | |
8d90395d | 21 | %option debug nodefault noinput nounput noyywrap never-interactive |
e9071366 AD |
22 | %option prefix="code_" outfile="lex.yy.c" |
23 | ||
24 | %{ | |
25 | /* Work around a bug in flex 2.5.31. See Debian bug 333231 | |
26 | <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=333231>. */ | |
27 | #undef code_wrap | |
28 | #define code_wrap() 1 | |
29 | ||
30 | #define FLEX_PREFIX(Id) code_ ## Id | |
0305d25e | 31 | #include <src/flex-scanner.h> |
f9bfc42a | 32 | |
0305d25e AD |
33 | #include <src/complain.h> |
34 | #include <src/reader.h> | |
35 | #include <src/getargs.h> | |
00f5d575 | 36 | #include <src/muscle-tab.h> |
0305d25e AD |
37 | #include <src/scan-code.h> |
38 | #include <src/symlist.h> | |
e9071366 | 39 | |
872b52bc AR |
40 | #include <c-ctype.h> |
41 | #include <get-errno.h> | |
42 | #include <quote.h> | |
43 | ||
e9071366 AD |
44 | /* The current calling start condition: SC_RULE_ACTION or |
45 | SC_SYMBOL_ACTION. */ | |
28e52c0d | 46 | # define YY_DECL static char *code_lex (code_props *self, int sc_context) |
e9071366 AD |
47 | YY_DECL; |
48 | ||
49 | #define YY_USER_ACTION location_compute (loc, &loc->end, yytext, yyleng); | |
50 | ||
4210cd0b JD |
51 | static void handle_action_dollar (symbol_list *rule, char *cp, |
52 | location dollar_loc); | |
53 | static void handle_action_at (symbol_list *rule, char *cp, location at_loc); | |
b9f1d9a4 | 54 | |
872b52bc | 55 | /* A string to be pushed to obstack after dollar/at has been handled. */ |
b9f1d9a4 AR |
56 | static char *ref_tail_fields; |
57 | ||
e9071366 AD |
58 | static location the_location; |
59 | static location *loc = &the_location; | |
ddc8ede1 | 60 | |
28e52c0d JD |
61 | /* A string representing the most recent translation. */ |
62 | static char *last_string; | |
63 | ||
ddc8ede1 PE |
64 | /* True if an untyped $$ or $n was seen. */ |
65 | static bool untyped_var_seen; | |
b9f1d9a4 | 66 | |
e9071366 AD |
67 | %} |
68 | /* C and C++ comments in code. */ | |
69 | %x SC_COMMENT SC_LINE_COMMENT | |
70 | /* Strings and characters in code. */ | |
71 | %x SC_STRING SC_CHARACTER | |
72 | /* Whether in a rule or symbol action. Specifies the translation | |
73 | of $ and @. */ | |
74 | %x SC_RULE_ACTION SC_SYMBOL_ACTION | |
75 | ||
76 | ||
77 | /* POSIX says that a tag must be both an id and a C union member, but | |
78 | historically almost any character is allowed in a tag. We disallow | |
79 | NUL and newline, as this simplifies our implementation. */ | |
80 | tag [^\0\n>]+ | |
81 | ||
82 | /* Zero or more instances of backslash-newline. Following GCC, allow | |
83 | white space between the backslash and the newline. */ | |
84 | splice (\\[ \f\t\v]*\n)* | |
85 | ||
b9f1d9a4 | 86 | /* C style identifier. Must start with letter. Will be used for |
872b52bc AR |
87 | named symbol references. Shall be kept synchronized with |
88 | scan-gram.l "letter" and "id". */ | |
b9f1d9a4 AR |
89 | letter [-.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] |
90 | id {letter}({letter}|[0-9])* | |
91 | ref -?[0-9]+|{id}|"["{id}"]"|"$" | |
92 | ||
e9071366 AD |
93 | %% |
94 | ||
95 | %{ | |
2346344a | 96 | /* Nesting level of the current code in braces. */ |
4f82b42a | 97 | int braces_level = 0; |
2346344a | 98 | |
e8cd1ad6 DJ |
99 | /* Whether a semicolon is probably needed. |
100 | The heuristic is that a semicolon is not needed after `{', `}', `;', | |
101 | or a C preprocessor directive, and that whitespaces and comments | |
102 | do not affect this flag. | |
103 | Note that `{' does not need a semicolon because of `{}'. | |
104 | A semicolon may be needed before a cpp direcive, but don't bother. */ | |
105 | bool need_semicolon = false; | |
106 | ||
107 | /* Whether in a C preprocessor directive. Don't use a start condition | |
108 | for this because, at the end of strings and comments, we still need | |
109 | to know whether we're in a directive. */ | |
110 | bool in_cpp = false; | |
111 | ||
e9071366 AD |
112 | /* This scanner is special: it is invoked only once, henceforth |
113 | is expected to return only once. This initialization is | |
114 | therefore done once per action to translate. */ | |
4f82b42a PE |
115 | aver (sc_context == SC_SYMBOL_ACTION |
116 | || sc_context == SC_RULE_ACTION | |
117 | || sc_context == INITIAL); | |
e9071366 AD |
118 | BEGIN sc_context; |
119 | %} | |
120 | ||
121 | /*------------------------------------------------------------. | |
122 | | Scanning a C comment. The initial `/ *' is already eaten. | | |
123 | `------------------------------------------------------------*/ | |
124 | ||
125 | <SC_COMMENT> | |
126 | { | |
127 | "*"{splice}"/" STRING_GROW; BEGIN sc_context; | |
128 | } | |
129 | ||
130 | ||
131 | /*--------------------------------------------------------------. | |
132 | | Scanning a line comment. The initial `//' is already eaten. | | |
133 | `--------------------------------------------------------------*/ | |
134 | ||
135 | <SC_LINE_COMMENT> | |
136 | { | |
137 | "\n" STRING_GROW; BEGIN sc_context; | |
138 | {splice} STRING_GROW; | |
139 | } | |
140 | ||
141 | ||
142 | /*--------------------------------------------. | |
143 | | Scanning user-code characters and strings. | | |
144 | `--------------------------------------------*/ | |
145 | ||
146 | <SC_CHARACTER,SC_STRING> | |
147 | { | |
148 | {splice}|\\{splice}. STRING_GROW; | |
149 | } | |
150 | ||
151 | <SC_CHARACTER> | |
152 | { | |
153 | "'" STRING_GROW; BEGIN sc_context; | |
154 | } | |
155 | ||
156 | <SC_STRING> | |
157 | { | |
158 | "\"" STRING_GROW; BEGIN sc_context; | |
159 | } | |
160 | ||
161 | ||
162 | <SC_RULE_ACTION,SC_SYMBOL_ACTION>{ | |
163 | "'" { | |
164 | STRING_GROW; | |
165 | BEGIN SC_CHARACTER; | |
e8cd1ad6 | 166 | need_semicolon = true; |
e9071366 AD |
167 | } |
168 | "\"" { | |
169 | STRING_GROW; | |
170 | BEGIN SC_STRING; | |
e8cd1ad6 | 171 | need_semicolon = true; |
e9071366 AD |
172 | } |
173 | "/"{splice}"*" { | |
174 | STRING_GROW; | |
175 | BEGIN SC_COMMENT; | |
176 | } | |
177 | "/"{splice}"/" { | |
178 | STRING_GROW; | |
179 | BEGIN SC_LINE_COMMENT; | |
180 | } | |
181 | } | |
182 | ||
183 | <SC_RULE_ACTION> | |
184 | { | |
b9f1d9a4 AR |
185 | "$"("<"{tag}">")?{ref} { |
186 | ref_tail_fields = 0; | |
28e52c0d | 187 | handle_action_dollar (self->rule, yytext, *loc); |
872b52bc | 188 | if (ref_tail_fields) { |
b9f1d9a4 AR |
189 | obstack_sgrow (&obstack_for_string, ref_tail_fields); |
190 | } | |
e8cd1ad6 | 191 | need_semicolon = true; |
28e52c0d | 192 | } |
b9f1d9a4 AR |
193 | "@"{ref} { |
194 | ref_tail_fields = 0; | |
28e52c0d | 195 | handle_action_at (self->rule, yytext, *loc); |
872b52bc | 196 | if (ref_tail_fields) { |
b9f1d9a4 AR |
197 | obstack_sgrow (&obstack_for_string, ref_tail_fields); |
198 | } | |
e8cd1ad6 | 199 | need_semicolon = true; |
28e52c0d | 200 | } |
e9071366 AD |
201 | "$" { |
202 | warn_at (*loc, _("stray `$'")); | |
203 | obstack_sgrow (&obstack_for_string, "$]["); | |
e8cd1ad6 | 204 | need_semicolon = true; |
e9071366 AD |
205 | } |
206 | "@" { | |
207 | warn_at (*loc, _("stray `@'")); | |
208 | obstack_sgrow (&obstack_for_string, "@@"); | |
e8cd1ad6 DJ |
209 | need_semicolon = true; |
210 | } | |
211 | "[" { | |
212 | obstack_sgrow (&obstack_for_string, "@{"); | |
213 | need_semicolon = true; | |
214 | } | |
215 | "]" { | |
216 | obstack_sgrow (&obstack_for_string, "@}"); | |
217 | need_semicolon = true; | |
e9071366 | 218 | } |
2346344a | 219 | |
e8cd1ad6 DJ |
220 | ";" STRING_GROW; need_semicolon = false; |
221 | "{" STRING_GROW; ++braces_level; need_semicolon = false; | |
2346344a | 222 | "}" { |
58bd33b7 | 223 | bool outer_brace = --braces_level == 0; |
2346344a AD |
224 | |
225 | /* As an undocumented Bison extension, append `;' before the last | |
226 | brace in braced code, so that the user code can omit trailing | |
227 | `;'. But do not append `;' if emulating Yacc, since Yacc does | |
e8cd1ad6 | 228 | not append one. */ |
d9a9b96b | 229 | if (outer_brace && !yacc_flag && language_prio == default_prio |
e8cd1ad6 DJ |
230 | && skeleton_prio == default_prio && need_semicolon && ! in_cpp) |
231 | { | |
232 | warn_at (*loc, _("a `;' might be needed at the end of action code")); | |
233 | warn_at (*loc, _("future versions of Bison will not add the `;'")); | |
234 | obstack_1grow (&obstack_for_string, ';'); | |
235 | } | |
2346344a AD |
236 | |
237 | STRING_GROW; | |
e8cd1ad6 | 238 | need_semicolon = false; |
2346344a | 239 | } |
e8cd1ad6 DJ |
240 | |
241 | /* Preprocessing directives should only be recognized at the beginning | |
242 | of lines, allowing whitespace including comments, but in C/C++, | |
243 | `#' can only be the start of preprocessor directives or within | |
244 | `#define' directives anyway, so don't bother with begin of line. */ | |
245 | "#" STRING_GROW; in_cpp = true; | |
246 | ||
247 | {splice} STRING_GROW; | |
bc0f5737 | 248 | [\n\r] STRING_GROW; if (in_cpp) in_cpp = need_semicolon = false; |
e8cd1ad6 | 249 | [ \t\f] STRING_GROW; |
4395a9ff JD |
250 | |
251 | /* YYFAIL is undocumented and was formally deprecated in Bison | |
252 | 2.4.2. */ | |
253 | YYFAIL { | |
254 | STRING_GROW; need_semicolon = true; | |
255 | warn_at (*loc, _("use of YYFAIL, which is deprecated and will be" | |
256 | " removed")); | |
257 | } | |
258 | ||
259 | /* The sole purpose of this is to make sure identifiers that merely | |
260 | contain YYFAIL don't produce the above warning. */ | |
261 | [A-Za-z_][0-9A-Za-z_]* STRING_GROW; need_semicolon = true; | |
262 | ||
263 | . STRING_GROW; need_semicolon = true; | |
e9071366 AD |
264 | } |
265 | ||
266 | <SC_SYMBOL_ACTION> | |
267 | { | |
28e52c0d JD |
268 | "$$" { |
269 | obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar["); | |
270 | self->is_value_used = true; | |
271 | } | |
e785ccf7 JD |
272 | "@$" { |
273 | obstack_sgrow (&obstack_for_string, "]b4_at_dollar["); | |
bc0f5737 | 274 | muscle_percent_define_ensure("locations", the_location, true); |
e785ccf7 | 275 | } |
e9071366 AD |
276 | } |
277 | ||
278 | ||
279 | /*-----------------------------------------. | |
280 | | Escape M4 quoting characters in C code. | | |
281 | `-----------------------------------------*/ | |
282 | ||
283 | <*> | |
284 | { | |
285 | \$ obstack_sgrow (&obstack_for_string, "$]["); | |
286 | \@ obstack_sgrow (&obstack_for_string, "@@"); | |
287 | \[ obstack_sgrow (&obstack_for_string, "@{"); | |
288 | \] obstack_sgrow (&obstack_for_string, "@}"); | |
289 | } | |
290 | ||
291 | /*-----------------------------------------------------. | |
292 | | By default, grow the string obstack with the input. | | |
293 | `-----------------------------------------------------*/ | |
294 | ||
295 | <*>.|\n STRING_GROW; | |
296 | ||
297 | /* End of processing. */ | |
298 | <*><<EOF>> { | |
28e52c0d JD |
299 | STRING_FINISH; |
300 | return last_string; | |
e9071366 AD |
301 | } |
302 | ||
303 | %% | |
304 | ||
b9f1d9a4 | 305 | static inline bool |
872b52bc | 306 | is_dot_or_dash (char ch) |
b9f1d9a4 AR |
307 | { |
308 | return ch == '.' || ch == '-'; | |
309 | } | |
310 | ||
311 | static inline bool | |
872b52bc | 312 | contains_dot_or_dash (const char* p) |
b9f1d9a4 | 313 | { |
872b52bc AR |
314 | for (; *p; ++p) |
315 | if (is_dot_or_dash (*p)) | |
316 | return true; | |
317 | return false; | |
b9f1d9a4 AR |
318 | } |
319 | ||
872b52bc | 320 | /* Defines a variant of a symbolic name resolution. */ |
b9f1d9a4 AR |
321 | typedef struct |
322 | { | |
323 | /* Index in symbol list. */ | |
ce268795 | 324 | unsigned symbol_index; |
b9f1d9a4 AR |
325 | |
326 | /* Matched symbol id and loc. */ | |
327 | uniqstr id; | |
328 | location loc; | |
329 | ||
872b52bc | 330 | /* Hiding named reference. */ |
b9f1d9a4 AR |
331 | named_ref* hidden_by; |
332 | ||
872b52bc AR |
333 | /* Error flags. May contain zero (no errors) or |
334 | a combination of VARIANT_* values. */ | |
b9f1d9a4 AR |
335 | unsigned err; |
336 | } variant; | |
337 | ||
872b52bc AR |
338 | /* Set when the variant refers to a symbol hidden |
339 | by an explicit symbol reference. */ | |
340 | #define VARIANT_HIDDEN (1 << 0) | |
341 | ||
342 | /* Set when the variant refers to a symbol containing | |
343 | dots or dashes. Will require explicit bracketing. */ | |
344 | #define VARIANT_BAD_BRACKETING (1 << 1) | |
345 | ||
346 | /* Set when the variant refers to a symbol which is | |
347 | not visible from current midrule. */ | |
348 | #define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 2) | |
349 | ||
b9f1d9a4 AR |
350 | static variant *variant_table = 0; |
351 | static unsigned variant_table_size = 0; | |
352 | static unsigned variant_count = 0; | |
353 | ||
354 | static variant * | |
ba4184ec | 355 | variant_table_grow (void) |
b9f1d9a4 AR |
356 | { |
357 | ++variant_count; | |
358 | if (variant_count > variant_table_size) | |
359 | { | |
360 | while (variant_count > variant_table_size) | |
361 | variant_table_size = 2 * variant_table_size + 3; | |
362 | variant_table = xnrealloc (variant_table, variant_table_size, | |
363 | sizeof *variant_table); | |
364 | } | |
365 | return &variant_table[variant_count - 1]; | |
366 | } | |
367 | ||
13cdf208 | 368 | static void |
ba4184ec | 369 | variant_table_free (void) |
13cdf208 | 370 | { |
ba4184ec | 371 | free (variant_table); |
13cdf208 AR |
372 | variant_table = 0; |
373 | variant_table_size = variant_count = 0; | |
374 | } | |
375 | ||
b9f1d9a4 | 376 | static char * |
872b52bc | 377 | find_prefix_end (const char *prefix, char *begin, char *end) |
b9f1d9a4 AR |
378 | { |
379 | char *ptr = begin; | |
380 | ||
872b52bc AR |
381 | for (; *prefix && ptr != end; ++prefix, ++ptr) |
382 | if (*prefix != *ptr) | |
383 | return 0; | |
b9f1d9a4 AR |
384 | |
385 | if (*prefix) | |
386 | return 0; | |
387 | ||
388 | return ptr; | |
389 | } | |
390 | ||
391 | static variant * | |
ba4184ec | 392 | variant_add (uniqstr id, location id_loc, unsigned symbol_index, |
61bc57e5 | 393 | char *cp, char *cp_end, bool explicit_bracketing) |
b9f1d9a4 AR |
394 | { |
395 | char *prefix_end; | |
396 | ||
872b52bc | 397 | prefix_end = find_prefix_end (id, cp, cp_end); |
b9f1d9a4 AR |
398 | if (prefix_end && |
399 | (prefix_end == cp_end || | |
61bc57e5 | 400 | (!explicit_bracketing && is_dot_or_dash (*prefix_end)))) |
b9f1d9a4 | 401 | { |
872b52bc | 402 | variant *r = variant_table_grow (); |
ce268795 | 403 | r->symbol_index = symbol_index; |
b9f1d9a4 | 404 | r->id = id; |
ba4184ec | 405 | r->loc = id_loc; |
b9f1d9a4 AR |
406 | r->hidden_by = NULL; |
407 | r->err = 0; | |
408 | return r; | |
409 | } | |
410 | else | |
411 | return NULL; | |
412 | } | |
413 | ||
872b52bc | 414 | static const char * |
ba4184ec | 415 | get_at_spec(unsigned symbol_index) |
872b52bc AR |
416 | { |
417 | static char at_buf[20]; | |
ba4184ec | 418 | if (symbol_index == 0) |
872b52bc AR |
419 | strcpy (at_buf, "$$"); |
420 | else | |
ba4184ec | 421 | snprintf (at_buf, sizeof at_buf, "$%u", symbol_index); |
872b52bc AR |
422 | return at_buf; |
423 | } | |
424 | ||
ce268795 | 425 | static void |
61bc57e5 | 426 | show_sub_messages (const char* cp, bool explicit_bracketing, |
66381412 AR |
427 | int midrule_rhs_index, char dollar_or_at, |
428 | bool is_warning, unsigned indent) | |
ce268795 AR |
429 | { |
430 | unsigned i; | |
431 | ||
432 | for (i = 0; i < variant_count; ++i) | |
433 | { | |
434 | const variant *var = &variant_table[i]; | |
435 | const char *at_spec = get_at_spec (var->symbol_index); | |
436 | ||
437 | if (var->err == 0) | |
438 | { | |
439 | if (is_warning) | |
66381412 AR |
440 | warn_at_indent (var->loc, &indent, _("refers to: %c%s at %s"), |
441 | dollar_or_at, var->id, at_spec); | |
ce268795 | 442 | else |
66381412 AR |
443 | complain_at_indent (var->loc, &indent, _("refers to: %c%s at %s"), |
444 | dollar_or_at, var->id, at_spec); | |
ce268795 AR |
445 | } |
446 | else | |
447 | { | |
448 | static struct obstack msg_buf; | |
61bc57e5 | 449 | const char *tail = explicit_bracketing ? "" : |
ce268795 AR |
450 | cp + strlen (var->id); |
451 | const char *id = var->hidden_by ? var->hidden_by->id : | |
452 | var->id; | |
453 | location id_loc = var->hidden_by ? var->hidden_by->loc : | |
454 | var->loc; | |
455 | ||
456 | /* Create the explanation message. */ | |
457 | obstack_init (&msg_buf); | |
458 | ||
66381412 | 459 | obstack_fgrow1 (&msg_buf, _("possibly meant: %c"), dollar_or_at); |
ce268795 AR |
460 | if (contains_dot_or_dash (id)) |
461 | obstack_fgrow1 (&msg_buf, "[%s]", id); | |
462 | else | |
463 | obstack_sgrow (&msg_buf, id); | |
464 | obstack_sgrow (&msg_buf, tail); | |
465 | ||
466 | if (var->err & VARIANT_HIDDEN) | |
467 | { | |
66381412 | 468 | obstack_fgrow1 (&msg_buf, _(", hiding %c"), dollar_or_at); |
ce268795 AR |
469 | if (contains_dot_or_dash (var->id)) |
470 | obstack_fgrow1 (&msg_buf, "[%s]", var->id); | |
471 | else | |
472 | obstack_sgrow (&msg_buf, var->id); | |
473 | obstack_sgrow (&msg_buf, tail); | |
474 | } | |
475 | ||
66381412 | 476 | obstack_fgrow1 (&msg_buf, _(" at %s"), at_spec); |
ce268795 AR |
477 | |
478 | if (var->err & VARIANT_NOT_VISIBLE_FROM_MIDRULE) | |
66381412 AR |
479 | { |
480 | const char *format = | |
481 | _(", cannot be accessed from mid-rule action at $%d"); | |
482 | obstack_fgrow1 (&msg_buf, format, midrule_rhs_index); | |
483 | } | |
ce268795 AR |
484 | |
485 | obstack_1grow (&msg_buf, '\0'); | |
486 | if (is_warning) | |
66381412 AR |
487 | warn_at_indent (id_loc, &indent, "%s", |
488 | (char *) obstack_finish (&msg_buf)); | |
ce268795 | 489 | else |
66381412 AR |
490 | complain_at_indent (id_loc, &indent, "%s", |
491 | (char *) obstack_finish (&msg_buf)); | |
ce268795 AR |
492 | obstack_free (&msg_buf, 0); |
493 | } | |
494 | } | |
495 | } | |
496 | ||
872b52bc AR |
497 | /* Returned from "parse_ref" when the reference |
498 | is inappropriate. */ | |
b9f1d9a4 | 499 | #define INVALID_REF (INT_MIN) |
872b52bc AR |
500 | |
501 | /* Returned from "parse_ref" when the reference | |
ce268795 | 502 | points to LHS ($$) of the current rule or midrule. */ |
b9f1d9a4 AR |
503 | #define LHS_REF (INT_MIN + 1) |
504 | ||
66381412 AR |
505 | /* Sub-messages indent. */ |
506 | #define SUB_INDENT (4) | |
507 | ||
872b52bc AR |
508 | /* Parse named or positional reference. In case of positional |
509 | references, can return negative values for $-n "deep" stack | |
510 | accesses. */ | |
b9f1d9a4 | 511 | static long int |
872b52bc | 512 | parse_ref (char *cp, symbol_list *rule, int rule_length, |
ba4184ec | 513 | int midrule_rhs_index, char *text, location text_loc, |
872b52bc | 514 | char dollar_or_at) |
b9f1d9a4 AR |
515 | { |
516 | symbol_list *l; | |
517 | char *cp_end; | |
61bc57e5 | 518 | bool explicit_bracketing; |
872b52bc | 519 | unsigned i; |
ce268795 AR |
520 | unsigned valid_variants = 0; |
521 | unsigned valid_variant_index = 0; | |
b9f1d9a4 AR |
522 | |
523 | if ('$' == *cp) | |
524 | return LHS_REF; | |
525 | ||
872b52bc | 526 | if (c_isdigit (*cp) || (*cp == '-' && c_isdigit (* (cp + 1)))) |
b9f1d9a4 AR |
527 | { |
528 | long int num = strtol (cp, &cp, 10); | |
529 | if (1 - INT_MAX + rule_length <= num && num <= rule_length) | |
530 | return num; | |
531 | else | |
532 | { | |
ba4184ec AR |
533 | complain_at (text_loc, _("integer out of range: %s"), |
534 | quote (text)); | |
b9f1d9a4 AR |
535 | return INVALID_REF; |
536 | } | |
537 | } | |
538 | ||
539 | if ('[' == *cp) | |
540 | { | |
b9f1d9a4 | 541 | /* Ignore the brackets. */ |
872b52bc AR |
542 | char *p; |
543 | for (p = ++cp; *p != ']'; ++p) | |
544 | continue; | |
b9f1d9a4 | 545 | cp_end = p; |
872b52bc | 546 | |
61bc57e5 | 547 | explicit_bracketing = true; |
b9f1d9a4 AR |
548 | } |
549 | else | |
550 | { | |
b9f1d9a4 | 551 | /* Take all characters of the name. */ |
872b52bc | 552 | char* p; |
b9f1d9a4 | 553 | for (p = cp; *p; ++p) |
872b52bc | 554 | if (is_dot_or_dash (*p)) |
b9f1d9a4 AR |
555 | { |
556 | ref_tail_fields = p; | |
557 | break; | |
558 | } | |
872b52bc AR |
559 | for (p = cp; *p; ++p) |
560 | continue; | |
b9f1d9a4 | 561 | cp_end = p; |
872b52bc | 562 | |
61bc57e5 | 563 | explicit_bracketing = false; |
b9f1d9a4 AR |
564 | } |
565 | ||
566 | /* Add all relevant variants. */ | |
872b52bc | 567 | { |
ba4184ec | 568 | unsigned symbol_index; |
872b52bc | 569 | variant_count = 0; |
ba4184ec AR |
570 | for (symbol_index = 0, l = rule; !symbol_list_null (l); |
571 | ++symbol_index, l = l->next) | |
872b52bc | 572 | { |
ba4184ec | 573 | variant *var; |
872b52bc AR |
574 | if (l->content_type != SYMLIST_SYMBOL) |
575 | continue; | |
b9f1d9a4 | 576 | |
ba4184ec | 577 | var = variant_add (l->content.sym->tag, l->sym_loc, |
61bc57e5 | 578 | symbol_index, cp, cp_end, explicit_bracketing); |
ba4184ec AR |
579 | if (var && l->named_ref) |
580 | var->hidden_by = l->named_ref; | |
b9f1d9a4 | 581 | |
872b52bc | 582 | if (l->named_ref) |
ba4184ec | 583 | variant_add (l->named_ref->id, l->named_ref->loc, |
61bc57e5 | 584 | symbol_index, cp, cp_end, explicit_bracketing); |
872b52bc AR |
585 | } |
586 | } | |
b9f1d9a4 AR |
587 | |
588 | /* Check errors. */ | |
b9f1d9a4 AR |
589 | for (i = 0; i < variant_count; ++i) |
590 | { | |
ba4184ec | 591 | variant *var = &variant_table[i]; |
ce268795 | 592 | unsigned symbol_index = var->symbol_index; |
b9f1d9a4 AR |
593 | |
594 | /* Check visibility from mid-rule actions. */ | |
872b52bc | 595 | if (midrule_rhs_index != 0 |
ba4184ec | 596 | && (symbol_index == 0 || midrule_rhs_index < symbol_index)) |
ce268795 | 597 | var->err |= VARIANT_NOT_VISIBLE_FROM_MIDRULE; |
b9f1d9a4 AR |
598 | |
599 | /* Check correct bracketing. */ | |
61bc57e5 | 600 | if (!explicit_bracketing && contains_dot_or_dash (var->id)) |
ce268795 | 601 | var->err |= VARIANT_BAD_BRACKETING; |
b9f1d9a4 AR |
602 | |
603 | /* Check using of hidden symbols. */ | |
ba4184ec | 604 | if (var->hidden_by) |
ce268795 | 605 | var->err |= VARIANT_HIDDEN; |
b9f1d9a4 | 606 | |
ba4184ec | 607 | if (!var->err) |
ce268795 AR |
608 | { |
609 | valid_variant_index = i; | |
610 | ++valid_variants; | |
611 | } | |
b9f1d9a4 AR |
612 | } |
613 | ||
ce268795 | 614 | switch (valid_variants) |
b9f1d9a4 | 615 | { |
ce268795 | 616 | case 0: |
f64e406b AR |
617 | { |
618 | unsigned len = (explicit_bracketing || !ref_tail_fields) ? | |
619 | cp_end - cp : ref_tail_fields - cp; | |
66381412 | 620 | unsigned indent = 0; |
f64e406b | 621 | |
66381412 AR |
622 | complain_at_indent (text_loc, &indent, _("invalid reference: %s"), |
623 | quote (text)); | |
624 | indent += SUB_INDENT; | |
f64e406b | 625 | if (midrule_rhs_index) |
66381412 AR |
626 | { |
627 | const char *format = | |
628 | _("symbol not found in production before $%d: %.*s"); | |
629 | complain_at_indent (rule->location, &indent, format, | |
630 | midrule_rhs_index, len, cp); | |
631 | } | |
f64e406b | 632 | else |
66381412 AR |
633 | { |
634 | const char *format = | |
635 | _("symbol not found in production: %.*s"); | |
636 | complain_at_indent (rule->location, &indent, format, | |
637 | len, cp); | |
638 | } | |
f64e406b AR |
639 | |
640 | if (variant_count > 0) | |
61bc57e5 | 641 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, |
66381412 | 642 | dollar_or_at, false, indent); |
f64e406b AR |
643 | return INVALID_REF; |
644 | } | |
ce268795 AR |
645 | case 1: |
646 | { | |
66381412 | 647 | unsigned indent = 0; |
ce268795 AR |
648 | if (variant_count > 1) |
649 | { | |
66381412 AR |
650 | warn_at_indent (text_loc, &indent, _("misleading reference: %s"), |
651 | quote (text)); | |
61bc57e5 | 652 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, |
66381412 | 653 | dollar_or_at, true, indent + SUB_INDENT); |
ce268795 AR |
654 | } |
655 | { | |
656 | unsigned symbol_index = | |
657 | variant_table[valid_variant_index].symbol_index; | |
658 | return (symbol_index == midrule_rhs_index) ? LHS_REF : symbol_index; | |
659 | } | |
660 | } | |
661 | case 2: | |
662 | default: | |
66381412 AR |
663 | { |
664 | unsigned indent = 0; | |
665 | complain_at_indent (text_loc, &indent, _("ambiguous reference: %s"), | |
666 | quote (text)); | |
667 | show_sub_messages (cp, explicit_bracketing, midrule_rhs_index, | |
668 | dollar_or_at, false, indent + SUB_INDENT); | |
669 | return INVALID_REF; | |
670 | } | |
b9f1d9a4 AR |
671 | } |
672 | ||
ce268795 | 673 | /* Not reachable. */ |
b9f1d9a4 AR |
674 | return INVALID_REF; |
675 | } | |
676 | ||
e9071366 AD |
677 | /* Keeps track of the maximum number of semantic values to the left of |
678 | a handle (those referenced by $0, $-1, etc.) are required by the | |
679 | semantic actions of this grammar. */ | |
680 | int max_left_semantic_context = 0; | |
681 | ||
682 | ||
683 | /*------------------------------------------------------------------. | |
684 | | TEXT is pointing to a wannabee semantic value (i.e., a `$'). | | |
685 | | | | |
686 | | Possible inputs: $[<TYPENAME>]($|integer) | | |
687 | | | | |
688 | | Output to OBSTACK_FOR_STRING a reference to this semantic value. | | |
689 | `------------------------------------------------------------------*/ | |
690 | ||
691 | static void | |
4210cd0b | 692 | handle_action_dollar (symbol_list *rule, char *text, location dollar_loc) |
e9071366 | 693 | { |
28e52c0d | 694 | char const *type_name = NULL; |
e9071366 | 695 | char *cp = text + 1; |
b9f1d9a4 | 696 | char *gt_ptr = 0; |
ffa4ba3a | 697 | symbol_list *effective_rule; |
872b52bc AR |
698 | int effective_rule_length; |
699 | int n; | |
ffa4ba3a JD |
700 | |
701 | if (rule->midrule_parent_rule) | |
702 | { | |
703 | effective_rule = rule->midrule_parent_rule; | |
704 | effective_rule_length = rule->midrule_parent_rhs_index - 1; | |
705 | } | |
706 | else | |
707 | { | |
708 | effective_rule = rule; | |
709 | effective_rule_length = symbol_list_length (rule->next); | |
710 | } | |
e9071366 AD |
711 | |
712 | /* Get the type name if explicit. */ | |
713 | if (*cp == '<') | |
714 | { | |
715 | type_name = ++cp; | |
716 | while (*cp != '>') | |
717 | ++cp; | |
b9f1d9a4 AR |
718 | |
719 | /* The '>' symbol will be later replaced by '\0'. Original | |
720 | 'text' is needed for error messages. */ | |
721 | gt_ptr = cp; | |
e9071366 | 722 | ++cp; |
ddc8ede1 PE |
723 | if (untyped_var_seen) |
724 | complain_at (dollar_loc, _("explicit type given in untyped grammar")); | |
725 | tag_seen = true; | |
e9071366 AD |
726 | } |
727 | ||
872b52bc AR |
728 | n = parse_ref (cp, effective_rule, effective_rule_length, |
729 | rule->midrule_parent_rhs_index, text, dollar_loc, '$'); | |
b9f1d9a4 AR |
730 | |
731 | if (gt_ptr) | |
732 | *gt_ptr = '\0'; | |
733 | ||
734 | switch (n) | |
e9071366 | 735 | { |
b9f1d9a4 AR |
736 | case INVALID_REF: |
737 | break; | |
738 | ||
739 | case LHS_REF: | |
e9071366 | 740 | if (!type_name) |
4210cd0b | 741 | type_name = symbol_list_n_type_name_get (rule, dollar_loc, 0); |
ddc8ede1 PE |
742 | |
743 | if (!type_name) | |
ad6b1efa | 744 | { |
ddc8ede1 PE |
745 | if (union_seen | tag_seen) |
746 | { | |
747 | if (rule->midrule_parent_rule) | |
748 | complain_at (dollar_loc, | |
749 | _("$$ for the midrule at $%d of `%s'" | |
750 | " has no declared type"), | |
751 | rule->midrule_parent_rhs_index, | |
3be03b13 | 752 | effective_rule->content.sym->tag); |
ddc8ede1 PE |
753 | else |
754 | complain_at (dollar_loc, _("$$ of `%s' has no declared type"), | |
3be03b13 | 755 | rule->content.sym->tag); |
ddc8ede1 | 756 | } |
ad6b1efa | 757 | else |
ddc8ede1 PE |
758 | untyped_var_seen = true; |
759 | type_name = ""; | |
ad6b1efa | 760 | } |
ddc8ede1 | 761 | |
e9071366 AD |
762 | obstack_fgrow1 (&obstack_for_string, |
763 | "]b4_lhs_value([%s])[", type_name); | |
f6857bbf | 764 | rule->action_props.is_value_used = true; |
b9f1d9a4 AR |
765 | break; |
766 | ||
767 | default: | |
768 | if (max_left_semantic_context < 1 - n) | |
769 | max_left_semantic_context = 1 - n; | |
770 | if (!type_name && 0 < n) | |
771 | type_name = | |
772 | symbol_list_n_type_name_get (effective_rule, dollar_loc, n); | |
773 | if (!type_name) | |
e9071366 | 774 | { |
b9f1d9a4 AR |
775 | if (union_seen | tag_seen) |
776 | complain_at (dollar_loc, _("$%s of `%s' has no declared type"), | |
777 | cp, effective_rule->content.sym->tag); | |
778 | else | |
779 | untyped_var_seen = true; | |
780 | type_name = ""; | |
e9071366 | 781 | } |
b9f1d9a4 AR |
782 | |
783 | obstack_fgrow3 (&obstack_for_string, | |
784 | "]b4_rhs_value(%d, %d, [%s])[", | |
785 | effective_rule_length, n, type_name); | |
786 | if (n > 0) | |
787 | symbol_list_n_get (effective_rule, n)->action_props.is_value_used = | |
788 | true; | |
789 | break; | |
e9071366 AD |
790 | } |
791 | } | |
792 | ||
793 | ||
794 | /*------------------------------------------------------. | |
795 | | TEXT is a location token (i.e., a `@...'). Output to | | |
796 | | OBSTACK_FOR_STRING a reference to this location. | | |
797 | `------------------------------------------------------*/ | |
798 | ||
799 | static void | |
4210cd0b | 800 | handle_action_at (symbol_list *rule, char *text, location at_loc) |
e9071366 AD |
801 | { |
802 | char *cp = text + 1; | |
b9f1d9a4 | 803 | symbol_list *effective_rule; |
872b52bc AR |
804 | int effective_rule_length; |
805 | int n; | |
b9f1d9a4 AR |
806 | |
807 | if (rule->midrule_parent_rule) | |
808 | { | |
809 | effective_rule = rule->midrule_parent_rule; | |
810 | effective_rule_length = rule->midrule_parent_rhs_index - 1; | |
811 | } | |
812 | else | |
813 | { | |
814 | effective_rule = rule; | |
815 | effective_rule_length = symbol_list_length (rule->next); | |
816 | } | |
ffa4ba3a | 817 | |
bc0f5737 | 818 | muscle_percent_define_ensure("locations", at_loc, true); |
e9071366 | 819 | |
872b52bc | 820 | n = parse_ref (cp, effective_rule, effective_rule_length, |
b9f1d9a4 AR |
821 | rule->midrule_parent_rhs_index, text, at_loc, '@'); |
822 | switch (n) | |
e9071366 | 823 | { |
b9f1d9a4 AR |
824 | case INVALID_REF: |
825 | break; | |
e9071366 | 826 | |
b9f1d9a4 AR |
827 | case LHS_REF: |
828 | obstack_sgrow (&obstack_for_string, "]b4_lhs_location["); | |
829 | break; | |
830 | ||
831 | default: | |
832 | obstack_fgrow2 (&obstack_for_string, "]b4_rhs_location(%d, %d)[", | |
833 | effective_rule_length, n); | |
834 | break; | |
e9071366 AD |
835 | } |
836 | } | |
837 | ||
838 | ||
839 | /*-------------------------. | |
840 | | Initialize the scanner. | | |
841 | `-------------------------*/ | |
842 | ||
28e52c0d | 843 | /* Translate the dollars and ats in \a self, in the context \a sc_context |
ddc8ede1 | 844 | (SC_RULE_ACTION, SC_SYMBOL_ACTION, INITIAL). */ |
e9071366 | 845 | |
eb095650 | 846 | static char const * |
28e52c0d | 847 | translate_action (code_props *self, int sc_context) |
e9071366 | 848 | { |
2ce4ed68 | 849 | char *res; |
e9071366 AD |
850 | static bool initialized = false; |
851 | if (!initialized) | |
852 | { | |
853 | obstack_init (&obstack_for_string); | |
e9071366 AD |
854 | yy_flex_debug = 0; |
855 | initialized = true; | |
856 | } | |
857 | ||
28e52c0d JD |
858 | loc->start = loc->end = self->location.start; |
859 | yy_switch_to_buffer (yy_scan_string (self->code)); | |
860 | res = code_lex (self, sc_context); | |
e9071366 AD |
861 | yy_delete_buffer (YY_CURRENT_BUFFER); |
862 | ||
863 | return res; | |
864 | } | |
865 | ||
28e52c0d JD |
866 | /*------------------------------------------------------------------------. |
867 | | Implementation of the public interface as documented in "scan-code.h". | | |
868 | `------------------------------------------------------------------------*/ | |
869 | ||
870 | void | |
871 | code_props_none_init (code_props *self) | |
76290368 | 872 | { |
28e52c0d | 873 | *self = code_props_none; |
76290368 JD |
874 | } |
875 | ||
28e52c0d JD |
876 | code_props const code_props_none = CODE_PROPS_NONE_INIT; |
877 | ||
878 | void | |
872b52bc AR |
879 | code_props_plain_init (code_props *self, char const *code, |
880 | location code_loc) | |
76290368 | 881 | { |
28e52c0d JD |
882 | self->kind = CODE_PROPS_PLAIN; |
883 | self->code = code; | |
884 | self->location = code_loc; | |
885 | self->is_value_used = false; | |
886 | self->rule = NULL; | |
b9f1d9a4 | 887 | self->named_ref = NULL; |
76290368 JD |
888 | } |
889 | ||
28e52c0d JD |
890 | void |
891 | code_props_symbol_action_init (code_props *self, char const *code, | |
892 | location code_loc) | |
e9071366 | 893 | { |
28e52c0d JD |
894 | self->kind = CODE_PROPS_SYMBOL_ACTION; |
895 | self->code = code; | |
896 | self->location = code_loc; | |
897 | self->is_value_used = false; | |
898 | self->rule = NULL; | |
b9f1d9a4 | 899 | self->named_ref = NULL; |
e9071366 AD |
900 | } |
901 | ||
28e52c0d JD |
902 | void |
903 | code_props_rule_action_init (code_props *self, char const *code, | |
b9f1d9a4 | 904 | location code_loc, symbol_list *rule, |
ca2a6d15 | 905 | named_ref *name, bool is_predicate) |
28e52c0d JD |
906 | { |
907 | self->kind = CODE_PROPS_RULE_ACTION; | |
908 | self->code = code; | |
909 | self->location = code_loc; | |
910 | self->is_value_used = false; | |
911 | self->rule = rule; | |
ba4184ec | 912 | self->named_ref = name; |
ca2a6d15 | 913 | self->is_predicate = is_predicate; |
28e52c0d JD |
914 | } |
915 | ||
916 | void | |
917 | code_props_translate_code (code_props *self) | |
918 | { | |
919 | switch (self->kind) | |
920 | { | |
921 | case CODE_PROPS_NONE: | |
922 | break; | |
923 | case CODE_PROPS_PLAIN: | |
924 | self->code = translate_action (self, INITIAL); | |
925 | break; | |
926 | case CODE_PROPS_SYMBOL_ACTION: | |
927 | self->code = translate_action (self, SC_SYMBOL_ACTION); | |
928 | break; | |
929 | case CODE_PROPS_RULE_ACTION: | |
930 | self->code = translate_action (self, SC_RULE_ACTION); | |
931 | break; | |
932 | } | |
933 | } | |
934 | ||
935 | void | |
936 | code_scanner_last_string_free (void) | |
937 | { | |
938 | STRING_FREE; | |
939 | } | |
e9071366 AD |
940 | |
941 | void | |
942 | code_scanner_free (void) | |
943 | { | |
944 | obstack_free (&obstack_for_string, 0); | |
13cdf208 AR |
945 | variant_table_free (); |
946 | ||
e9071366 | 947 | /* Reclaim Flex's buffers. */ |
580b8926 | 948 | yylex_destroy (); |
e9071366 | 949 | } |