]>
Commit | Line | Data |
---|---|---|
e9955c83 AD |
1 | /* Bison Grammar Scanner -*- C -*- |
2 | Copyright (C) 2002 Free Software Foundation, Inc. | |
3 | ||
4 | This file is part of Bison, the GNU Compiler Compiler. | |
5 | ||
6 | This program 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 of the License, or | |
9 | (at your option) any later version. | |
10 | ||
11 | This program 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 this program; if not, write to the Free Software | |
18 | Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA | |
19 | 02111-1307 USA | |
20 | */ | |
21 | ||
22 | %option debug nodefault noyywrap nounput never-interactive stack | |
23 | %option prefix="gram_" outfile="lex.yy.c" | |
24 | ||
25 | %{ | |
26 | #include "system.h" | |
d8d3f94a | 27 | #include "mbswidth.h" |
e9955c83 AD |
28 | #include "complain.h" |
29 | #include "quote.h" | |
30 | #include "getargs.h" | |
31 | #include "gram.h" | |
32 | #include "reader.h" | |
33 | ||
34 | /* Each time we match a string, move the end cursor to its end. */ | |
8efe435c AD |
35 | #define YY_USER_INIT \ |
36 | do { \ | |
37 | LOCATION_RESET (*yylloc); \ | |
1a715ef2 | 38 | yylloc->file = infile; \ |
8efe435c AD |
39 | /* This is only to avoid GCC warnings. */ \ |
40 | if (yycontrol) {;}; \ | |
41 | } while (0) | |
42 | ||
d8d3f94a PE |
43 | #define YY_USER_ACTION extend_location (yylloc, yytext, yyleng); |
44 | #define YY_STEP LOCATION_STEP (*yylloc) | |
45 | ||
46 | #define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size)) | |
47 | ||
48 | ||
49 | /* Read bytes from FP into buffer BUF of size SIZE. Return the | |
50 | number of bytes read. Remove '\r' from input, treating \r\n | |
51 | and isolated \r as \n. */ | |
52 | ||
53 | static size_t | |
54 | no_cr_read (FILE *fp, char *buf, size_t size) | |
55 | { | |
56 | size_t s = fread (buf, 1, size, fp); | |
57 | if (s) | |
58 | { | |
59 | char *w = memchr (buf, '\r', s); | |
60 | if (w) | |
61 | { | |
62 | char const *r = ++w; | |
63 | char const *lim = buf + s; | |
64 | ||
65 | for (;;) | |
66 | { | |
67 | /* Found an '\r'. Treat it like '\n', but ignore any | |
68 | '\n' that immediately follows. */ | |
69 | w[-1] = '\n'; | |
70 | if (r == lim) | |
71 | { | |
72 | int ch = getc (fp); | |
73 | if (ch != '\n' && ungetc (ch, fp) != ch) | |
74 | break; | |
75 | } | |
76 | else if (*r == '\n') | |
77 | r++; | |
78 | ||
79 | /* Copy until the next '\r'. */ | |
80 | do | |
81 | { | |
82 | if (r == lim) | |
83 | return w - buf; | |
84 | } | |
85 | while ((*w++ = *r++) != '\r'); | |
86 | } | |
87 | ||
88 | return w - buf; | |
89 | } | |
90 | } | |
91 | ||
92 | return s; | |
93 | } | |
94 | ||
95 | ||
96 | /* Extend *LOC to account for token TOKEN of size SIZE. */ | |
97 | ||
98 | static void | |
99 | extend_location (location_t *loc, char const *token, int size) | |
100 | { | |
101 | int line = loc->last_line; | |
102 | int column = loc->last_column; | |
103 | char const *p0 = token; | |
104 | char const *p = token; | |
105 | char const *lim = token + size; | |
106 | ||
107 | for (p = token; p < lim; p++) | |
108 | switch (*p) | |
109 | { | |
110 | case '\r': | |
111 | /* \r shouldn't survive no_cr_read. */ | |
112 | abort (); | |
113 | ||
114 | case '\n': | |
115 | line++; | |
116 | column = 1; | |
117 | p0 = p + 1; | |
118 | break; | |
119 | ||
120 | case '\t': | |
121 | column += mbsnwidth (p0, p - p0, 0); | |
122 | column += 8 - ((column - 1) & 7); | |
123 | p0 = p + 1; | |
124 | break; | |
125 | } | |
126 | ||
127 | loc->last_line = line; | |
128 | loc->last_column = column + mbsnwidth (p0, p - p0, 0); | |
129 | } | |
130 | ||
131 | ||
8e6ef483 PE |
132 | /* Report an unexpected end of file at LOC. A token or comment began |
133 | with TOKEN_START, but an end of file was encountered and the | |
134 | expected TOKEN_END was missing. */ | |
135 | ||
136 | static void | |
137 | unexpected_end_of_file (location_t loc, | |
138 | char const *token_start, char const *token_end) | |
139 | { | |
140 | complain_at (loc, _("unexpected end of file in `%s ... %s'"), | |
141 | token_start, token_end); | |
142 | } | |
143 | ||
144 | ||
e9955c83 | 145 | |
44995b2e AD |
146 | /* STRING_OBSTACK -- Used to store all the characters that we need to |
147 | keep (to construct ID, STRINGS etc.). Use the following macros to | |
148 | use it. | |
149 | ||
1d6412ad AD |
150 | Use YY_OBS_GROW to append what has just been matched, and |
151 | YY_OBS_FINISH to end the string (it puts the ending 0). | |
152 | YY_OBS_FINISH also stores this string in LAST_STRING, which can be | |
153 | used, and which is used by YY_OBS_FREE to free the last string. */ | |
44995b2e AD |
154 | |
155 | static struct obstack string_obstack; | |
156 | char *last_string; | |
157 | ||
44995b2e AD |
158 | #define YY_OBS_GROW \ |
159 | obstack_grow (&string_obstack, yytext, yyleng) | |
160 | ||
161 | #define YY_OBS_FINISH \ | |
162 | do { \ | |
163 | obstack_1grow (&string_obstack, '\0'); \ | |
164 | last_string = obstack_finish (&string_obstack); \ | |
44995b2e AD |
165 | } while (0) |
166 | ||
167 | #define YY_OBS_FREE \ | |
168 | do { \ | |
169 | obstack_free (&string_obstack, last_string); \ | |
170 | } while (0) | |
e9955c83 | 171 | |
4cdb01db AD |
172 | void |
173 | scanner_last_string_free (void) | |
174 | { | |
175 | YY_OBS_FREE; | |
176 | } | |
177 | ||
178 | ||
e9955c83 AD |
179 | static int percent_percent_count = 0; |
180 | ||
efcb44dd PE |
181 | /* Within well-formed rules, RULE_LENGTH is the number of values in |
182 | the current rule so far, which says where to find `$0' with respect | |
183 | to the top of the stack. It is not the same as the rule->length in | |
184 | the case of mid rule actions. | |
185 | ||
186 | Outside of well-formed rules, RULE_LENGTH has an undefined value. */ | |
187 | static int rule_length; | |
188 | ||
d33cb3ae PE |
189 | static void handle_dollar (braced_code_t code_kind, |
190 | char *cp, location_t location); | |
191 | static void handle_at (braced_code_t code_kind, | |
192 | char *cp, location_t location); | |
900c5db5 | 193 | static void handle_syncline (char *args, location_t *location); |
d8d3f94a | 194 | static int convert_ucn_to_byte (char const *hex_text); |
e9955c83 AD |
195 | |
196 | %} | |
d8d3f94a | 197 | %x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT |
e9955c83 AD |
198 | %x SC_STRING SC_CHARACTER |
199 | %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER | |
200 | %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE | |
201 | ||
29c01725 AD |
202 | letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] |
203 | id {letter}({letter}|[0-9])* | |
204 | directive %{letter}({letter}|[0-9]|-)* | |
205 | int [0-9]+ | |
d8d3f94a PE |
206 | |
207 | /* POSIX says that a tag must be both an id and a C union member, but | |
208 | historically almost any character is allowed in a tag. We disallow | |
209 | NUL and newline, as this simplifies our implementation. */ | |
210 | tag [^\0\n>]+ | |
211 | ||
212 | /* Zero or more instances of backslash-newline. Following GCC, allow | |
213 | white space between the backslash and the newline. */ | |
214 | splice (\\[ \f\t\v]*\n)* | |
e9955c83 AD |
215 | |
216 | %% | |
217 | %{ | |
1a9e39f1 PE |
218 | int braces_level IF_LINT (= 0); |
219 | ||
e9955c83 AD |
220 | /* At each yylex invocation, mark the current position as the |
221 | start of the next token. */ | |
e9955c83 | 222 | YY_STEP; |
e9955c83 AD |
223 | %} |
224 | ||
225 | ||
226 | /*----------------------------. | |
227 | | Scanning Bison directives. | | |
228 | `----------------------------*/ | |
229 | <INITIAL> | |
230 | { | |
231 | "%binary" return PERCENT_NONASSOC; | |
232 | "%debug" return PERCENT_DEBUG; | |
233 | "%define" return PERCENT_DEFINE; | |
234 | "%defines" return PERCENT_DEFINES; | |
9280d3ef | 235 | "%destructor" return PERCENT_DESTRUCTOR; |
676385e2 | 236 | "%dprec" return PERCENT_DPREC; |
e9955c83 AD |
237 | "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE; |
238 | "%expect" return PERCENT_EXPECT; | |
239 | "%file-prefix" return PERCENT_FILE_PREFIX; | |
240 | "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC; | |
ae7453f2 | 241 | "%glr-parser" return PERCENT_GLR_PARSER; |
e9955c83 AD |
242 | "%left" return PERCENT_LEFT; |
243 | "%locations" return PERCENT_LOCATIONS; | |
676385e2 | 244 | "%merge" return PERCENT_MERGE; |
e9955c83 AD |
245 | "%name"[-_]"prefix" return PERCENT_NAME_PREFIX; |
246 | "%no"[-_]"lines" return PERCENT_NO_LINES; | |
247 | "%nonassoc" return PERCENT_NONASSOC; | |
248 | "%nterm" return PERCENT_NTERM; | |
249 | "%output" return PERCENT_OUTPUT; | |
ae7453f2 | 250 | "%parse-param" return PERCENT_PARSE_PARAM; |
d8d3f94a | 251 | "%prec" rule_length--; return PERCENT_PREC; |
366eea36 | 252 | "%printer" return PERCENT_PRINTER; |
e9955c83 AD |
253 | "%pure"[-_]"parser" return PERCENT_PURE_PARSER; |
254 | "%right" return PERCENT_RIGHT; | |
ae7453f2 | 255 | "%lex-param" return PERCENT_LEX_PARAM; |
e9955c83 AD |
256 | "%skeleton" return PERCENT_SKELETON; |
257 | "%start" return PERCENT_START; | |
258 | "%term" return PERCENT_TOKEN; | |
259 | "%token" return PERCENT_TOKEN; | |
260 | "%token"[-_]"table" return PERCENT_TOKEN_TABLE; | |
261 | "%type" return PERCENT_TYPE; | |
262 | "%union" return PERCENT_UNION; | |
263 | "%verbose" return PERCENT_VERBOSE; | |
264 | "%yacc" return PERCENT_YACC; | |
265 | ||
29c01725 AD |
266 | {directive} { |
267 | complain_at (*yylloc, _("invalid directive: %s"), quote (yytext)); | |
268 | YY_STEP; | |
269 | } | |
270 | ||
900c5db5 AD |
271 | ^"#line "{int}" \""[^\"]*"\"\n" handle_syncline (yytext + strlen ("#line "), yylloc); YY_STEP; |
272 | ||
e9955c83 | 273 | "=" return EQUAL; |
d8d3f94a PE |
274 | ":" rule_length = 0; return COLON; |
275 | "|" rule_length = 0; return PIPE; | |
ae7453f2 | 276 | "," return COMMA; |
e9955c83 AD |
277 | ";" return SEMICOLON; |
278 | ||
d8d3f94a PE |
279 | [ \f\n\t\v]+ YY_STEP; |
280 | ||
e9955c83 | 281 | {id} { |
39f41916 | 282 | yylval->symbol = symbol_get (yytext, *yylloc); |
efcb44dd | 283 | rule_length++; |
e9955c83 AD |
284 | return ID; |
285 | } | |
286 | ||
d8d3f94a PE |
287 | {int} { |
288 | unsigned long num; | |
289 | errno = 0; | |
290 | num = strtoul (yytext, 0, 10); | |
291 | if (INT_MAX < num || errno) | |
292 | { | |
98f2caaa | 293 | complain_at (*yylloc, _("integer out of range: %s"), quote (yytext)); |
d8d3f94a PE |
294 | num = INT_MAX; |
295 | } | |
296 | yylval->integer = num; | |
297 | return INT; | |
298 | } | |
e9955c83 AD |
299 | |
300 | /* Characters. We don't check there is only one. */ | |
db2cc12f | 301 | "'" YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER); |
e9955c83 AD |
302 | |
303 | /* Strings. */ | |
db2cc12f | 304 | "\"" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING); |
e9955c83 AD |
305 | |
306 | /* Comments. */ | |
d8d3f94a | 307 | "/*" BEGIN SC_YACC_COMMENT; |
e9955c83 AD |
308 | "//".* YY_STEP; |
309 | ||
310 | /* Prologue. */ | |
1d6412ad | 311 | "%{" yy_push_state (SC_PROLOGUE); |
e9955c83 AD |
312 | |
313 | /* Code in between braces. */ | |
1a9e39f1 | 314 | "{" YY_OBS_GROW; braces_level = 0; yy_push_state (SC_BRACED_CODE); |
e9955c83 AD |
315 | |
316 | /* A type. */ | |
d8d3f94a | 317 | "<"{tag}">" { |
4cdb01db AD |
318 | obstack_grow (&string_obstack, yytext + 1, yyleng - 2); |
319 | YY_OBS_FINISH; | |
320 | yylval->string = last_string; | |
321 | return TYPE; | |
322 | } | |
323 | ||
e9955c83 AD |
324 | |
325 | "%%" { | |
326 | if (++percent_percent_count == 2) | |
327 | yy_push_state (SC_EPILOGUE); | |
328 | return PERCENT_PERCENT; | |
329 | } | |
330 | ||
331 | . { | |
c4d720cd | 332 | complain_at (*yylloc, _("invalid character: %s"), quote (yytext)); |
e9955c83 AD |
333 | YY_STEP; |
334 | } | |
335 | } | |
336 | ||
337 | ||
d8d3f94a PE |
338 | /*-------------------------------------------------------------------. |
339 | | Whatever the start condition (but those which correspond to | | |
340 | | entities `swallowed' by Bison: SC_YACC_COMMENT, SC_ESCAPED_STRING, | | |
341 | | and SC_ESCAPED_CHARACTER), no M4 character must escape as is. | | |
342 | `-------------------------------------------------------------------*/ | |
e9955c83 | 343 | |
d8d3f94a | 344 | <SC_COMMENT,SC_LINE_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE> |
e9955c83 | 345 | { |
d8d3f94a PE |
346 | \[ obstack_sgrow (&string_obstack, "@<:@"); |
347 | \] obstack_sgrow (&string_obstack, "@:>@"); | |
e9955c83 AD |
348 | } |
349 | ||
350 | ||
d8d3f94a PE |
351 | /*---------------------------------------------------------------. |
352 | | Scanning a Yacc comment. The initial `/ *' is already eaten. | | |
353 | `---------------------------------------------------------------*/ | |
e9955c83 | 354 | |
d8d3f94a | 355 | <SC_YACC_COMMENT> |
e9955c83 | 356 | { |
d8d3f94a PE |
357 | "*/" { |
358 | YY_STEP; | |
359 | BEGIN INITIAL; | |
e9955c83 AD |
360 | } |
361 | ||
d8d3f94a PE |
362 | [^*]+|"*" ; |
363 | ||
364 | <<EOF>> { | |
8e6ef483 | 365 | unexpected_end_of_file (*yylloc, "/*", "*/"); |
d8d3f94a PE |
366 | BEGIN INITIAL; |
367 | } | |
368 | } | |
369 | ||
370 | ||
371 | /*------------------------------------------------------------. | |
372 | | Scanning a C comment. The initial `/ *' is already eaten. | | |
373 | `------------------------------------------------------------*/ | |
374 | ||
375 | <SC_COMMENT> | |
376 | { | |
377 | "*"{splice}"/" YY_OBS_GROW; yy_pop_state (); | |
378 | [^*\[\]]+|"*" YY_OBS_GROW; | |
e9955c83 AD |
379 | |
380 | <<EOF>> { | |
8e6ef483 | 381 | unexpected_end_of_file (*yylloc, "/*", "*/"); |
e9955c83 AD |
382 | yy_pop_state (); |
383 | } | |
384 | } | |
385 | ||
386 | ||
d8d3f94a PE |
387 | /*--------------------------------------------------------------. |
388 | | Scanning a line comment. The initial `//' is already eaten. | | |
389 | `--------------------------------------------------------------*/ | |
390 | ||
391 | <SC_LINE_COMMENT> | |
392 | { | |
393 | "\n" YY_OBS_GROW; yy_pop_state (); | |
394 | ([^\n\[\]]|{splice})+ YY_OBS_GROW; | |
395 | <<EOF>> yy_pop_state (); | |
396 | } | |
397 | ||
398 | ||
e9955c83 AD |
399 | /*----------------------------------------------------------------. |
400 | | Scanning a C string, including its escapes. The initial `"' is | | |
401 | | already eaten. | | |
402 | `----------------------------------------------------------------*/ | |
403 | ||
404 | <SC_ESCAPED_STRING> | |
405 | { | |
db2cc12f | 406 | "\"" { |
e9955c83 | 407 | assert (yy_top_state () == INITIAL); |
44995b2e AD |
408 | YY_OBS_GROW; |
409 | YY_OBS_FINISH; | |
4cdb01db | 410 | yylval->string = last_string; |
e9955c83 | 411 | yy_pop_state (); |
efcb44dd | 412 | rule_length++; |
e9955c83 AD |
413 | return STRING; |
414 | } | |
415 | ||
d8d3f94a | 416 | [^\"\\]+ YY_OBS_GROW; |
e9955c83 AD |
417 | |
418 | <<EOF>> { | |
8e6ef483 | 419 | unexpected_end_of_file (*yylloc, "\"", "\""); |
e9955c83 | 420 | assert (yy_top_state () == INITIAL); |
44995b2e | 421 | YY_OBS_FINISH; |
4cdb01db | 422 | yylval->string = last_string; |
e9955c83 AD |
423 | yy_pop_state (); |
424 | return STRING; | |
425 | } | |
426 | } | |
427 | ||
428 | /*---------------------------------------------------------------. | |
429 | | Scanning a C character, decoding its escapes. The initial "'" | | |
430 | | is already eaten. | | |
431 | `---------------------------------------------------------------*/ | |
432 | ||
433 | <SC_ESCAPED_CHARACTER> | |
434 | { | |
db2cc12f | 435 | "'" { |
44995b2e | 436 | YY_OBS_GROW; |
e9955c83 AD |
437 | assert (yy_top_state () == INITIAL); |
438 | { | |
44995b2e | 439 | YY_OBS_FINISH; |
39f41916 | 440 | yylval->symbol = symbol_get (last_string, *yylloc); |
e776192e | 441 | symbol_class_set (yylval->symbol, token_sym, *yylloc); |
e68d4575 PE |
442 | symbol_user_token_number_set (yylval->symbol, |
443 | (unsigned char) last_string[1], *yylloc); | |
44995b2e | 444 | YY_OBS_FREE; |
e9955c83 | 445 | yy_pop_state (); |
efcb44dd | 446 | rule_length++; |
e9955c83 AD |
447 | return ID; |
448 | } | |
449 | } | |
450 | ||
c4d720cd | 451 | [^\'\\]+ YY_OBS_GROW; |
e9955c83 AD |
452 | |
453 | <<EOF>> { | |
8e6ef483 | 454 | unexpected_end_of_file (*yylloc, "'", "'"); |
e9955c83 | 455 | assert (yy_top_state () == INITIAL); |
44995b2e | 456 | YY_OBS_FINISH; |
4cdb01db | 457 | yylval->string = last_string; |
e9955c83 AD |
458 | yy_pop_state (); |
459 | return CHARACTER; | |
460 | } | |
461 | } | |
462 | ||
463 | ||
464 | /*----------------------------. | |
465 | | Decode escaped characters. | | |
466 | `----------------------------*/ | |
467 | ||
468 | <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER> | |
469 | { | |
d8d3f94a PE |
470 | \\[0-7]{1,3} { |
471 | unsigned long c = strtoul (yytext + 1, 0, 8); | |
472 | if (UCHAR_MAX < c) | |
e9955c83 | 473 | { |
98f2caaa PE |
474 | complain_at (*yylloc, _("invalid escape sequence: %s"), |
475 | quote (yytext)); | |
e9955c83 AD |
476 | YY_STEP; |
477 | } | |
478 | else | |
479 | obstack_1grow (&string_obstack, c); | |
480 | } | |
481 | ||
d8d3f94a PE |
482 | \\x[0-9a-fA-F]+ { |
483 | unsigned long c; | |
484 | errno = 0; | |
485 | c = strtoul (yytext + 2, 0, 16); | |
486 | if (UCHAR_MAX < c || errno) | |
487 | { | |
98f2caaa PE |
488 | complain_at (*yylloc, _("invalid escape sequence: %s"), |
489 | quote (yytext)); | |
d8d3f94a PE |
490 | YY_STEP; |
491 | } | |
492 | else | |
493 | obstack_1grow (&string_obstack, c); | |
e9955c83 AD |
494 | } |
495 | ||
496 | \\a obstack_1grow (&string_obstack, '\a'); | |
497 | \\b obstack_1grow (&string_obstack, '\b'); | |
498 | \\f obstack_1grow (&string_obstack, '\f'); | |
499 | \\n obstack_1grow (&string_obstack, '\n'); | |
500 | \\r obstack_1grow (&string_obstack, '\r'); | |
501 | \\t obstack_1grow (&string_obstack, '\t'); | |
502 | \\v obstack_1grow (&string_obstack, '\v'); | |
c4d720cd | 503 | \\[\"\'?\\] obstack_1grow (&string_obstack, yytext[1]); |
d8d3f94a PE |
504 | \\(u|U[0-9a-fA-F]{4})[0-9a-fA-F]{4} { |
505 | int c = convert_ucn_to_byte (yytext); | |
506 | if (c < 0) | |
507 | { | |
98f2caaa PE |
508 | complain_at (*yylloc, _("invalid escape sequence: %s"), |
509 | quote (yytext)); | |
d8d3f94a PE |
510 | YY_STEP; |
511 | } | |
512 | else | |
513 | obstack_1grow (&string_obstack, c); | |
514 | } | |
4f25ebb0 | 515 | \\(.|\n) { |
98f2caaa PE |
516 | complain_at (*yylloc, _("unrecognized escape sequence: %s"), |
517 | quote (yytext)); | |
44995b2e | 518 | YY_OBS_GROW; |
e9955c83 | 519 | } |
4f25ebb0 AD |
520 | /* FLex wants this rule, in case of a `\<<EOF>>'. */ |
521 | \\ YY_OBS_GROW; | |
e9955c83 AD |
522 | } |
523 | ||
524 | ||
525 | /*----------------------------------------------------------. | |
526 | | Scanning a C character without decoding its escapes. The | | |
527 | | initial "'" is already eaten. | | |
528 | `----------------------------------------------------------*/ | |
529 | ||
530 | <SC_CHARACTER> | |
531 | { | |
db2cc12f | 532 | "'" { |
44995b2e | 533 | YY_OBS_GROW; |
e9955c83 AD |
534 | assert (yy_top_state () != INITIAL); |
535 | yy_pop_state (); | |
536 | } | |
537 | ||
d8d3f94a PE |
538 | [^'\[\]\\]+ YY_OBS_GROW; |
539 | \\{splice}[^\[\]] YY_OBS_GROW; | |
540 | {splice} YY_OBS_GROW; | |
541 | /* Needed for `\<<EOF>>', `\\<<newline>>[', and `\\<<newline>>]'. */ | |
4f25ebb0 | 542 | \\ YY_OBS_GROW; |
e9955c83 | 543 | |
e9955c83 | 544 | <<EOF>> { |
8e6ef483 | 545 | unexpected_end_of_file (*yylloc, "'", "'"); |
e9955c83 AD |
546 | assert (yy_top_state () != INITIAL); |
547 | yy_pop_state (); | |
548 | } | |
549 | } | |
550 | ||
551 | ||
552 | /*----------------------------------------------------------------. | |
553 | | Scanning a C string, without decoding its escapes. The initial | | |
554 | | `"' is already eaten. | | |
555 | `----------------------------------------------------------------*/ | |
556 | ||
557 | <SC_STRING> | |
558 | { | |
db2cc12f | 559 | "\"" { |
e9955c83 | 560 | assert (yy_top_state () != INITIAL); |
44995b2e | 561 | YY_OBS_GROW; |
e9955c83 AD |
562 | yy_pop_state (); |
563 | } | |
564 | ||
d8d3f94a PE |
565 | [^\"\[\]\\]+ YY_OBS_GROW; |
566 | \\{splice}[^\[\]] YY_OBS_GROW; | |
567 | {splice} YY_OBS_GROW; | |
568 | /* Needed for `\<<EOF>>', `\\<<newline>>[', and `\\<<newline>>]'. */ | |
4f25ebb0 | 569 | \\ YY_OBS_GROW; |
e9955c83 | 570 | |
e9955c83 | 571 | <<EOF>> { |
8e6ef483 | 572 | unexpected_end_of_file (*yylloc, "\"", "\""); |
e9955c83 AD |
573 | assert (yy_top_state () != INITIAL); |
574 | yy_pop_state (); | |
575 | } | |
576 | } | |
577 | ||
578 | ||
579 | /*---------------------------------------------------. | |
580 | | Strings, comments etc. can be found in user code. | | |
581 | `---------------------------------------------------*/ | |
582 | ||
583 | <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE> | |
584 | { | |
585 | /* Characters. We don't check there is only one. */ | |
db2cc12f | 586 | "'" YY_OBS_GROW; yy_push_state (SC_CHARACTER); |
e9955c83 AD |
587 | |
588 | /* Strings. */ | |
db2cc12f | 589 | "\"" YY_OBS_GROW; yy_push_state (SC_STRING); |
e9955c83 AD |
590 | |
591 | /* Comments. */ | |
d8d3f94a PE |
592 | "/"{splice}"*" YY_OBS_GROW; yy_push_state (SC_COMMENT); |
593 | "/"{splice}"/" YY_OBS_GROW; yy_push_state (SC_LINE_COMMENT); | |
4f25ebb0 AD |
594 | |
595 | /* Not comments. */ | |
596 | "/" YY_OBS_GROW; | |
e9955c83 AD |
597 | } |
598 | ||
599 | ||
600 | /*---------------------------------------------------------------. | |
601 | | Scanning some code in braces (%union and actions). The initial | | |
602 | | "{" is already eaten. | | |
603 | `---------------------------------------------------------------*/ | |
604 | ||
605 | <SC_BRACED_CODE> | |
606 | { | |
1a9e39f1 PE |
607 | "{"|"<"{splice}"%" YY_OBS_GROW; braces_level++; |
608 | "%"{splice}">" YY_OBS_GROW; braces_level--; | |
e9955c83 | 609 | "}" { |
44995b2e | 610 | YY_OBS_GROW; |
1a9e39f1 PE |
611 | braces_level--; |
612 | if (braces_level < 0) | |
e9955c83 AD |
613 | { |
614 | yy_pop_state (); | |
44995b2e | 615 | YY_OBS_FINISH; |
4cdb01db | 616 | yylval->string = last_string; |
efcb44dd | 617 | rule_length++; |
e9955c83 AD |
618 | return BRACED_CODE; |
619 | } | |
620 | } | |
621 | ||
d8d3f94a | 622 | "$"("<"{tag}">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code, |
f25bfb75 AD |
623 | yytext, *yylloc); } |
624 | "@"(-?[0-9]+|"$") { handle_at (current_braced_code, | |
625 | yytext, *yylloc); } | |
e9955c83 | 626 | |
1a9e39f1 PE |
627 | /* `"<"{splice}"<"' tokenizes `<<%' correctly (as `<<' `%') rather |
628 | than incorrrectly (as `<' `<%'). */ | |
629 | [^\"$%\'/<@\[\]\{\}]+|[$%/<@]|"<"{splice}"<" YY_OBS_GROW; | |
e9955c83 AD |
630 | |
631 | <<EOF>> { | |
8e6ef483 | 632 | unexpected_end_of_file (*yylloc, "{", "}"); |
e9955c83 | 633 | yy_pop_state (); |
44995b2e | 634 | YY_OBS_FINISH; |
4cdb01db AD |
635 | yylval->string = last_string; |
636 | return BRACED_CODE; | |
e9955c83 AD |
637 | } |
638 | ||
639 | } | |
640 | ||
641 | ||
642 | /*--------------------------------------------------------------. | |
643 | | Scanning some prologue: from "%{" (already scanned) to "%}". | | |
644 | `--------------------------------------------------------------*/ | |
645 | ||
646 | <SC_PROLOGUE> | |
647 | { | |
648 | "%}" { | |
649 | yy_pop_state (); | |
44995b2e | 650 | YY_OBS_FINISH; |
4cdb01db | 651 | yylval->string = last_string; |
e9955c83 AD |
652 | return PROLOGUE; |
653 | } | |
654 | ||
c4d720cd | 655 | [^%\[\]/\'\"]+ YY_OBS_GROW; |
4f25ebb0 | 656 | "%" YY_OBS_GROW; |
e9955c83 AD |
657 | |
658 | <<EOF>> { | |
8e6ef483 | 659 | unexpected_end_of_file (*yylloc, "%{", "%}"); |
e9955c83 | 660 | yy_pop_state (); |
44995b2e | 661 | YY_OBS_FINISH; |
4cdb01db | 662 | yylval->string = last_string; |
e9955c83 AD |
663 | return PROLOGUE; |
664 | } | |
e9955c83 AD |
665 | } |
666 | ||
667 | ||
668 | /*---------------------------------------------------------------. | |
669 | | Scanning the epilogue (everything after the second "%%", which | | |
d8d3f94a | 670 | | has already been eaten). | |
e9955c83 AD |
671 | `---------------------------------------------------------------*/ |
672 | ||
673 | <SC_EPILOGUE> | |
674 | { | |
d8d3f94a | 675 | [^\[\]]+ YY_OBS_GROW; |
e9955c83 AD |
676 | |
677 | <<EOF>> { | |
678 | yy_pop_state (); | |
44995b2e | 679 | YY_OBS_FINISH; |
4cdb01db | 680 | yylval->string = last_string; |
e9955c83 AD |
681 | return EPILOGUE; |
682 | } | |
683 | } | |
684 | ||
685 | ||
686 | %% | |
687 | ||
688 | /*------------------------------------------------------------------. | |
366eea36 | 689 | | TEXT is pointing to a wannabee semantic value (i.e., a `$'). | |
e9955c83 AD |
690 | | | |
691 | | Possible inputs: $[<TYPENAME>]($|integer) | | |
692 | | | | |
693 | | Output to the STRING_OBSTACK a reference to this semantic value. | | |
694 | `------------------------------------------------------------------*/ | |
695 | ||
f25bfb75 | 696 | static inline void |
366eea36 | 697 | handle_action_dollar (char *text, location_t location) |
e9955c83 AD |
698 | { |
699 | const char *type_name = NULL; | |
366eea36 | 700 | char *cp = text + 1; |
e9955c83 AD |
701 | |
702 | /* Get the type name if explicit. */ | |
703 | if (*cp == '<') | |
704 | { | |
705 | type_name = ++cp; | |
706 | while (*cp != '>') | |
707 | ++cp; | |
708 | *cp = '\0'; | |
709 | ++cp; | |
710 | } | |
711 | ||
712 | if (*cp == '$') | |
713 | { | |
714 | if (!type_name) | |
56c47203 | 715 | type_name = symbol_list_n_type_name_get (current_rule, location, 0); |
e9955c83 | 716 | if (!type_name && typed) |
56c47203 | 717 | complain_at (location, _("$$ of `%s' has no declared type"), |
97650f4e | 718 | current_rule->sym->tag); |
e9955c83 AD |
719 | if (!type_name) |
720 | type_name = ""; | |
721 | obstack_fgrow1 (&string_obstack, | |
722 | "]b4_lhs_value([%s])[", type_name); | |
723 | } | |
d8d3f94a | 724 | else |
e9955c83 | 725 | { |
d8d3f94a PE |
726 | long num; |
727 | errno = 0; | |
728 | num = strtol (cp, 0, 10); | |
e9955c83 | 729 | |
d8d3f94a | 730 | if (INT_MIN <= num && num <= rule_length && ! errno) |
e9955c83 | 731 | { |
d8d3f94a | 732 | int n = num; |
e9955c83 | 733 | if (!type_name && n > 0) |
56c47203 AD |
734 | type_name = symbol_list_n_type_name_get (current_rule, location, |
735 | n); | |
e9955c83 | 736 | if (!type_name && typed) |
56c47203 | 737 | complain_at (location, _("$%d of `%s' has no declared type"), |
97650f4e | 738 | n, current_rule->sym->tag); |
e9955c83 AD |
739 | if (!type_name) |
740 | type_name = ""; | |
741 | obstack_fgrow3 (&string_obstack, | |
742 | "]b4_rhs_value([%d], [%d], [%s])[", | |
743 | rule_length, n, type_name); | |
744 | } | |
d8d3f94a | 745 | else |
98f2caaa | 746 | complain_at (location, _("integer out of range: %s"), quote (text)); |
9280d3ef AD |
747 | } |
748 | } | |
749 | ||
750 | ||
366eea36 | 751 | /*---------------------------------------------------------------. |
d8d3f94a | 752 | | TEXT is expected to be $$ in some code associated to a symbol: | |
366eea36 AD |
753 | | destructor or printer. | |
754 | `---------------------------------------------------------------*/ | |
9280d3ef | 755 | |
f25bfb75 | 756 | static inline void |
366eea36 | 757 | handle_symbol_code_dollar (char *text, location_t location) |
9280d3ef | 758 | { |
366eea36 | 759 | char *cp = text + 1; |
9280d3ef | 760 | if (*cp == '$') |
366eea36 | 761 | obstack_sgrow (&string_obstack, "]b4_dollar_dollar["); |
9280d3ef | 762 | else |
c4d720cd | 763 | complain_at (location, _("invalid value: %s"), quote (text)); |
e9955c83 AD |
764 | } |
765 | ||
f25bfb75 AD |
766 | |
767 | /*-----------------------------------------------------------------. | |
768 | | Dispatch onto handle_action_dollar, or handle_destructor_dollar, | | |
769 | | depending upon CODE_KIND. | | |
770 | `-----------------------------------------------------------------*/ | |
e9955c83 AD |
771 | |
772 | static void | |
f25bfb75 AD |
773 | handle_dollar (braced_code_t braced_code_kind, |
774 | char *text, location_t location) | |
775 | { | |
776 | switch (braced_code_kind) | |
777 | { | |
778 | case action_braced_code: | |
779 | handle_action_dollar (text, location); | |
780 | break; | |
781 | ||
782 | case destructor_braced_code: | |
366eea36 AD |
783 | case printer_braced_code: |
784 | handle_symbol_code_dollar (text, location); | |
f25bfb75 AD |
785 | break; |
786 | } | |
787 | } | |
788 | ||
789 | ||
790 | /*------------------------------------------------------. | |
791 | | TEXT is a location token (i.e., a `@...'). Output to | | |
792 | | STRING_OBSTACK a reference to this location. | | |
793 | `------------------------------------------------------*/ | |
794 | ||
795 | static inline void | |
796 | handle_action_at (char *text, location_t location) | |
e9955c83 | 797 | { |
366eea36 | 798 | char *cp = text + 1; |
e9955c83 | 799 | locations_flag = 1; |
e9955c83 | 800 | |
366eea36 | 801 | if (*cp == '$') |
e9955c83 AD |
802 | { |
803 | obstack_sgrow (&string_obstack, "]b4_lhs_location["); | |
804 | } | |
d8d3f94a | 805 | else |
e9955c83 | 806 | { |
d8d3f94a PE |
807 | long num; |
808 | errno = 0; | |
809 | num = strtol (cp, 0, 10); | |
dafdc66f | 810 | |
d8d3f94a PE |
811 | if (INT_MIN <= num && num <= rule_length && ! errno) |
812 | { | |
813 | int n = num; | |
814 | obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[", | |
815 | rule_length, n); | |
816 | } | |
e9955c83 | 817 | else |
98f2caaa | 818 | complain_at (location, _("integer out of range: %s"), quote (text)); |
f25bfb75 AD |
819 | } |
820 | } | |
821 | ||
822 | ||
366eea36 | 823 | /*---------------------------------------------------------------. |
d8d3f94a | 824 | | TEXT is expected to be @$ in some code associated to a symbol: | |
366eea36 AD |
825 | | destructor or printer. | |
826 | `---------------------------------------------------------------*/ | |
f25bfb75 AD |
827 | |
828 | static inline void | |
366eea36 | 829 | handle_symbol_code_at (char *text, location_t location) |
f25bfb75 | 830 | { |
366eea36 AD |
831 | char *cp = text + 1; |
832 | if (*cp == '$') | |
833 | obstack_sgrow (&string_obstack, "]b4_at_dollar["); | |
f25bfb75 | 834 | else |
c4d720cd | 835 | complain_at (location, _("invalid value: %s"), quote (text)); |
e9955c83 | 836 | } |
4cdb01db | 837 | |
f25bfb75 AD |
838 | |
839 | /*-------------------------------------------------------------------. | |
840 | | Dispatch onto handle_action_at, or handle_destructor_at, depending | | |
841 | | upon CODE_KIND. | | |
842 | `-------------------------------------------------------------------*/ | |
843 | ||
844 | static void | |
845 | handle_at (braced_code_t braced_code_kind, | |
846 | char *text, location_t location) | |
847 | { | |
848 | switch (braced_code_kind) | |
849 | { | |
850 | case action_braced_code: | |
851 | handle_action_at (text, location); | |
852 | break; | |
853 | ||
854 | case destructor_braced_code: | |
366eea36 AD |
855 | case printer_braced_code: |
856 | handle_symbol_code_at (text, location); | |
f25bfb75 AD |
857 | break; |
858 | } | |
859 | } | |
860 | ||
861 | ||
d8d3f94a PE |
862 | /*------------------------------------------------------------------. |
863 | | Convert universal character name UCN to a single-byte character, | | |
864 | | and return that character. Return -1 if UCN does not correspond | | |
865 | | to a single-byte character. | | |
866 | `------------------------------------------------------------------*/ | |
867 | ||
868 | static int | |
869 | convert_ucn_to_byte (char const *ucn) | |
870 | { | |
871 | unsigned long code = strtoul (ucn + 2, 0, 16); | |
872 | ||
873 | /* FIXME: Currently we assume Unicode-compatible unibyte characters | |
874 | on ASCII hosts (i.e., Latin-1 on hosts with 8-bit bytes). On | |
875 | non-ASCII hosts we support only the portable C character set. | |
876 | These limitations should be removed once we add support for | |
877 | multibyte characters. */ | |
878 | ||
879 | if (UCHAR_MAX < code) | |
880 | return -1; | |
881 | ||
882 | #if ! ('$' == 0x24 && '@' == 0x40 && '`' == 0x60 && '~' == 0x7e) | |
883 | { | |
884 | /* A non-ASCII host. Use CODE to index into a table of the C | |
885 | basic execution character set, which is guaranteed to exist on | |
886 | all Standard C platforms. This table also includes '$', '@', | |
8e6ef483 | 887 | and '`', which are not in the basic execution character set but |
d8d3f94a PE |
888 | which are unibyte characters on all the platforms that we know |
889 | about. */ | |
890 | static signed char const table[] = | |
891 | { | |
892 | '\0', -1, -1, -1, -1, -1, -1, '\a', | |
893 | '\b', '\t', '\n', '\v', '\f', '\r', -1, -1, | |
894 | -1, -1, -1, -1, -1, -1, -1, -1, | |
895 | -1, -1, -1, -1, -1, -1, -1, -1, | |
896 | ' ', '!', '"', '#', '$', '%', '&', '\'', | |
897 | '(', ')', '*', '+', ',', '-', '.', '/', | |
898 | '0', '1', '2', '3', '4', '5', '6', '7', | |
899 | '8', '9', ':', ';', '<', '=', '>', '?', | |
900 | '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', | |
901 | 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', | |
902 | 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', | |
903 | 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', | |
904 | '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', | |
905 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', | |
906 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', | |
907 | 'x', 'y', 'z', '{', '|', '}', '~' | |
908 | }; | |
909 | ||
910 | code = code < sizeof table ? table[code] : -1; | |
911 | } | |
912 | #endif | |
c4d720cd | 913 | |
d8d3f94a PE |
914 | return code; |
915 | } | |
916 | ||
917 | ||
900c5db5 AD |
918 | /*----------------------------------------------------------------. |
919 | | Handle `#line INT "FILE"'. ARGS has already skipped `#line '. | | |
920 | `----------------------------------------------------------------*/ | |
921 | ||
922 | static void | |
923 | handle_syncline (char *args, location_t *location) | |
924 | { | |
925 | int lineno = strtol (args, &args, 10); | |
926 | const char *file = NULL; | |
927 | file = strchr (args, '"') + 1; | |
928 | *strchr (file, '"') = 0; | |
929 | /* FIXME: Leaking... Can't free, as some locations are still | |
930 | pointing to the old file name. */ | |
931 | infile = xstrdup (file); | |
932 | location->file = infile; | |
933 | location->last_line = lineno; | |
934 | } | |
935 | ||
f25bfb75 AD |
936 | /*-------------------------. |
937 | | Initialize the scanner. | | |
938 | `-------------------------*/ | |
939 | ||
1d6412ad AD |
940 | void |
941 | scanner_initialize (void) | |
942 | { | |
943 | obstack_init (&string_obstack); | |
944 | } | |
945 | ||
946 | ||
f25bfb75 AD |
947 | /*-----------------------------------------------. |
948 | | Free all the memory allocated to the scanner. | | |
949 | `-----------------------------------------------*/ | |
950 | ||
4cdb01db AD |
951 | void |
952 | scanner_free (void) | |
953 | { | |
954 | obstack_free (&string_obstack, 0); | |
536545f3 AD |
955 | /* Reclaim Flex's buffers. */ |
956 | yy_delete_buffer (YY_CURRENT_BUFFER); | |
4cdb01db | 957 | } |