]>
Commit | Line | Data |
---|---|---|
1 | /* Bison Grammar Scanner -*- C -*- | |
2 | ||
3 | Copyright (C) 2002-2007, 2009-2010 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="gram_" 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 gram_wrap | |
27 | #define gram_wrap() 1 | |
28 | ||
29 | #define FLEX_PREFIX(Id) gram_ ## Id | |
30 | #include "flex-scanner.h" | |
31 | ||
32 | #include "complain.h" | |
33 | #include "files.h" | |
34 | #include "gram.h" | |
35 | #include "quotearg.h" | |
36 | #include "reader.h" | |
37 | #include "uniqstr.h" | |
38 | ||
39 | #include <ctype.h> | |
40 | #include <mbswidth.h> | |
41 | #include <quote.h> | |
42 | ||
43 | #include "scan-gram.h" | |
44 | ||
45 | #define YY_DECL GRAM_LEX_DECL | |
46 | ||
47 | #define YY_USER_INIT \ | |
48 | code_start = scanner_cursor = loc->start; \ | |
49 | ||
50 | /* Location of scanner cursor. */ | |
51 | static boundary scanner_cursor; | |
52 | ||
53 | #define YY_USER_ACTION location_compute (loc, &scanner_cursor, yytext, yyleng); | |
54 | ||
55 | static size_t no_cr_read (FILE *, char *, size_t); | |
56 | #define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size)) | |
57 | ||
58 | /* A string representing the most recently saved token. */ | |
59 | static char *last_string; | |
60 | ||
61 | void | |
62 | gram_scanner_last_string_free (void) | |
63 | { | |
64 | STRING_FREE; | |
65 | } | |
66 | ||
67 | static void handle_syncline (char *, location); | |
68 | static unsigned long int scan_integer (char const *p, int base, location loc); | |
69 | static int convert_ucn_to_byte (char const *hex_text); | |
70 | static void unexpected_eof (boundary, char const *); | |
71 | static void unexpected_newline (boundary, char const *); | |
72 | ||
73 | %} | |
74 | /* A C-like comment in directives/rules. */ | |
75 | %x SC_YACC_COMMENT | |
76 | /* Strings and characters in directives/rules. */ | |
77 | %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER | |
78 | /* A identifier was just read in directives/rules. Special state | |
79 | to capture the sequence `identifier :'. */ | |
80 | %x SC_AFTER_IDENTIFIER | |
81 | ||
82 | /* Three types of user code: | |
83 | - prologue (code between `%{' `%}' in the first section, before %%); | |
84 | - actions, printers, union, etc, (between braced in the middle section); | |
85 | - epilogue (everything after the second %%). */ | |
86 | %x SC_PROLOGUE SC_BRACED_CODE SC_EPILOGUE | |
87 | /* C and C++ comments in code. */ | |
88 | %x SC_COMMENT SC_LINE_COMMENT | |
89 | /* Strings and characters in code. */ | |
90 | %x SC_STRING SC_CHARACTER | |
91 | ||
92 | letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] | |
93 | id {letter}({letter}|[0-9])* | |
94 | directive %{letter}({letter}|[0-9]|-)* | |
95 | int [0-9]+ | |
96 | ||
97 | /* POSIX says that a tag must be both an id and a C union member, but | |
98 | historically almost any character is allowed in a tag. We disallow | |
99 | NUL and newline, as this simplifies our implementation. */ | |
100 | tag [^\0\n>]+ | |
101 | ||
102 | /* Zero or more instances of backslash-newline. Following GCC, allow | |
103 | white space between the backslash and the newline. */ | |
104 | splice (\\[ \f\t\v]*\n)* | |
105 | ||
106 | %% | |
107 | %{ | |
108 | /* Nesting level of the current code in braces. */ | |
109 | int braces_level IF_LINT (= 0); | |
110 | ||
111 | /* Parent context state, when applicable. */ | |
112 | int context_state IF_LINT (= 0); | |
113 | ||
114 | /* Location of most recent identifier, when applicable. */ | |
115 | location id_loc IF_LINT (= empty_location); | |
116 | ||
117 | /* Where containing code started, when applicable. Its initial | |
118 | value is relevant only when yylex is invoked in the SC_EPILOGUE | |
119 | start condition. */ | |
120 | boundary code_start = scanner_cursor; | |
121 | ||
122 | /* Where containing comment or string or character literal started, | |
123 | when applicable. */ | |
124 | boundary token_start IF_LINT (= scanner_cursor); | |
125 | %} | |
126 | ||
127 | ||
128 | /*-----------------------. | |
129 | | Scanning white space. | | |
130 | `-----------------------*/ | |
131 | ||
132 | <INITIAL,SC_AFTER_IDENTIFIER> | |
133 | { | |
134 | /* Comments and white space. */ | |
135 | "," warn_at (*loc, _("stray `,' treated as white space")); | |
136 | [ \f\n\t\v] | | |
137 | "//".* ; | |
138 | "/*" { | |
139 | token_start = loc->start; | |
140 | context_state = YY_START; | |
141 | BEGIN SC_YACC_COMMENT; | |
142 | } | |
143 | ||
144 | /* #line directives are not documented, and may be withdrawn or | |
145 | modified in future versions of Bison. */ | |
146 | ^"#line "{int}" \"".*"\"\n" { | |
147 | handle_syncline (yytext + sizeof "#line " - 1, *loc); | |
148 | } | |
149 | } | |
150 | ||
151 | ||
152 | /*----------------------------. | |
153 | | Scanning Bison directives. | | |
154 | `----------------------------*/ | |
155 | <INITIAL> | |
156 | { | |
157 | "%binary" return PERCENT_NONASSOC; | |
158 | "%code" return PERCENT_CODE; | |
159 | "%debug" return PERCENT_DEBUG; | |
160 | "%default"[-_]"prec" return PERCENT_DEFAULT_PREC; | |
161 | "%define" return PERCENT_DEFINE; | |
162 | "%defines" return PERCENT_DEFINES; | |
163 | "%destructor" return PERCENT_DESTRUCTOR; | |
164 | "%dprec" return PERCENT_DPREC; | |
165 | "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE; | |
166 | "%expect" return PERCENT_EXPECT; | |
167 | "%expect"[-_]"rr" return PERCENT_EXPECT_RR; | |
168 | "%file-prefix" return PERCENT_FILE_PREFIX; | |
169 | "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC; | |
170 | "%initial-action" return PERCENT_INITIAL_ACTION; | |
171 | "%glr-parser" return PERCENT_GLR_PARSER; | |
172 | "%language" return PERCENT_LANGUAGE; | |
173 | "%left" return PERCENT_LEFT; | |
174 | "%lex-param" return PERCENT_LEX_PARAM; | |
175 | "%locations" return PERCENT_LOCATIONS; | |
176 | "%merge" return PERCENT_MERGE; | |
177 | "%name"[-_]"prefix" return PERCENT_NAME_PREFIX; | |
178 | "%no"[-_]"default"[-_]"prec" return PERCENT_NO_DEFAULT_PREC; | |
179 | "%no"[-_]"lines" return PERCENT_NO_LINES; | |
180 | "%nonassoc" return PERCENT_NONASSOC; | |
181 | "%nondeterministic-parser" return PERCENT_NONDETERMINISTIC_PARSER; | |
182 | "%nterm" return PERCENT_NTERM; | |
183 | "%output" return PERCENT_OUTPUT; | |
184 | "%parse-param" return PERCENT_PARSE_PARAM; | |
185 | "%prec" return PERCENT_PREC; | |
186 | "%printer" return PERCENT_PRINTER; | |
187 | "%pure"[-_]"parser" return PERCENT_PURE_PARSER; | |
188 | "%require" return PERCENT_REQUIRE; | |
189 | "%right" return PERCENT_RIGHT; | |
190 | "%skeleton" return PERCENT_SKELETON; | |
191 | "%start" return PERCENT_START; | |
192 | "%term" return PERCENT_TOKEN; | |
193 | "%token" return PERCENT_TOKEN; | |
194 | "%token"[-_]"table" return PERCENT_TOKEN_TABLE; | |
195 | "%type" return PERCENT_TYPE; | |
196 | "%union" return PERCENT_UNION; | |
197 | "%verbose" return PERCENT_VERBOSE; | |
198 | "%yacc" return PERCENT_YACC; | |
199 | ||
200 | {directive} { | |
201 | complain_at (*loc, _("invalid directive: %s"), quote (yytext)); | |
202 | } | |
203 | ||
204 | "=" return EQUAL; | |
205 | "|" return PIPE; | |
206 | ";" return SEMICOLON; | |
207 | "<*>" return TYPE_TAG_ANY; | |
208 | "<>" return TYPE_TAG_NONE; | |
209 | ||
210 | {id} { | |
211 | val->uniqstr = uniqstr_new (yytext); | |
212 | id_loc = *loc; | |
213 | BEGIN SC_AFTER_IDENTIFIER; | |
214 | } | |
215 | ||
216 | {int} { | |
217 | val->integer = scan_integer (yytext, 10, *loc); | |
218 | return INT; | |
219 | } | |
220 | 0[xX][0-9abcdefABCDEF]+ { | |
221 | val->integer = scan_integer (yytext, 16, *loc); | |
222 | return INT; | |
223 | } | |
224 | ||
225 | /* Characters. We don't check there is only one. */ | |
226 | "'" STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER; | |
227 | ||
228 | /* Strings. */ | |
229 | "\"" token_start = loc->start; BEGIN SC_ESCAPED_STRING; | |
230 | ||
231 | /* Prologue. */ | |
232 | "%{" code_start = loc->start; BEGIN SC_PROLOGUE; | |
233 | ||
234 | /* Code in between braces. */ | |
235 | "{" { | |
236 | STRING_GROW; | |
237 | braces_level = 0; | |
238 | code_start = loc->start; | |
239 | BEGIN SC_BRACED_CODE; | |
240 | } | |
241 | ||
242 | /* A type. */ | |
243 | "<"{tag}">" { | |
244 | obstack_grow (&obstack_for_string, yytext + 1, yyleng - 2); | |
245 | STRING_FINISH; | |
246 | val->uniqstr = uniqstr_new (last_string); | |
247 | STRING_FREE; | |
248 | return TYPE; | |
249 | } | |
250 | ||
251 | "%%" { | |
252 | static int percent_percent_count; | |
253 | if (++percent_percent_count == 2) | |
254 | BEGIN SC_EPILOGUE; | |
255 | return PERCENT_PERCENT; | |
256 | } | |
257 | ||
258 | . { | |
259 | complain_at (*loc, _("invalid character: %s"), quote (yytext)); | |
260 | } | |
261 | ||
262 | <<EOF>> { | |
263 | loc->start = loc->end = scanner_cursor; | |
264 | yyterminate (); | |
265 | } | |
266 | } | |
267 | ||
268 | ||
269 | /*-----------------------------------------------------------------. | |
270 | | Scanning after an identifier, checking whether a colon is next. | | |
271 | `-----------------------------------------------------------------*/ | |
272 | ||
273 | <SC_AFTER_IDENTIFIER> | |
274 | { | |
275 | ":" { | |
276 | *loc = id_loc; | |
277 | BEGIN INITIAL; | |
278 | return ID_COLON; | |
279 | } | |
280 | . { | |
281 | scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0); | |
282 | yyless (0); | |
283 | *loc = id_loc; | |
284 | BEGIN INITIAL; | |
285 | return ID; | |
286 | } | |
287 | <<EOF>> { | |
288 | *loc = id_loc; | |
289 | BEGIN INITIAL; | |
290 | return ID; | |
291 | } | |
292 | } | |
293 | ||
294 | ||
295 | /*---------------------------------------------------------------. | |
296 | | Scanning a Yacc comment. The initial `/ *' is already eaten. | | |
297 | `---------------------------------------------------------------*/ | |
298 | ||
299 | <SC_YACC_COMMENT> | |
300 | { | |
301 | "*/" BEGIN context_state; | |
302 | .|\n ; | |
303 | <<EOF>> unexpected_eof (token_start, "*/"); BEGIN context_state; | |
304 | } | |
305 | ||
306 | ||
307 | /*------------------------------------------------------------. | |
308 | | Scanning a C comment. The initial `/ *' is already eaten. | | |
309 | `------------------------------------------------------------*/ | |
310 | ||
311 | <SC_COMMENT> | |
312 | { | |
313 | "*"{splice}"/" STRING_GROW; BEGIN context_state; | |
314 | <<EOF>> unexpected_eof (token_start, "*/"); BEGIN context_state; | |
315 | } | |
316 | ||
317 | ||
318 | /*--------------------------------------------------------------. | |
319 | | Scanning a line comment. The initial `//' is already eaten. | | |
320 | `--------------------------------------------------------------*/ | |
321 | ||
322 | <SC_LINE_COMMENT> | |
323 | { | |
324 | "\n" STRING_GROW; BEGIN context_state; | |
325 | {splice} STRING_GROW; | |
326 | <<EOF>> BEGIN context_state; | |
327 | } | |
328 | ||
329 | ||
330 | /*------------------------------------------------. | |
331 | | Scanning a Bison string, including its escapes. | | |
332 | | The initial quote is already eaten. | | |
333 | `------------------------------------------------*/ | |
334 | ||
335 | <SC_ESCAPED_STRING> | |
336 | { | |
337 | "\""|"\n" { | |
338 | if (yytext[0] == '\n') | |
339 | unexpected_newline (token_start, "\""); | |
340 | STRING_FINISH; | |
341 | loc->start = token_start; | |
342 | val->chars = last_string; | |
343 | BEGIN INITIAL; | |
344 | return STRING; | |
345 | } | |
346 | <<EOF>> { | |
347 | unexpected_eof (token_start, "\""); | |
348 | STRING_FINISH; | |
349 | loc->start = token_start; | |
350 | val->chars = last_string; | |
351 | BEGIN INITIAL; | |
352 | return STRING; | |
353 | } | |
354 | } | |
355 | ||
356 | /*----------------------------------------------------------. | |
357 | | Scanning a Bison character literal, decoding its escapes. | | |
358 | | The initial quote is already eaten. | | |
359 | `----------------------------------------------------------*/ | |
360 | ||
361 | <SC_ESCAPED_CHARACTER> | |
362 | { | |
363 | "'"|"\n" { | |
364 | if (yytext[0] == '\n') | |
365 | unexpected_newline (token_start, "'"); | |
366 | STRING_GROW; | |
367 | STRING_FINISH; | |
368 | loc->start = token_start; | |
369 | val->character = last_string[1]; | |
370 | STRING_FREE; | |
371 | BEGIN INITIAL; | |
372 | return CHAR; | |
373 | } | |
374 | <<EOF>> { | |
375 | unexpected_eof (token_start, "'"); | |
376 | STRING_FINISH; | |
377 | loc->start = token_start; | |
378 | if (strlen(last_string) > 1) | |
379 | val->character = last_string[1]; | |
380 | else | |
381 | val->character = last_string[0]; | |
382 | STRING_FREE; | |
383 | BEGIN INITIAL; | |
384 | return CHAR; | |
385 | } | |
386 | } | |
387 | ||
388 | <SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING> | |
389 | { | |
390 | \0 complain_at (*loc, _("invalid null character")); | |
391 | } | |
392 | ||
393 | ||
394 | /*----------------------------. | |
395 | | Decode escaped characters. | | |
396 | `----------------------------*/ | |
397 | ||
398 | <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER> | |
399 | { | |
400 | \\[0-7]{1,3} { | |
401 | unsigned long int c = strtoul (yytext + 1, NULL, 8); | |
402 | if (!c || UCHAR_MAX < c) | |
403 | complain_at (*loc, _("invalid number after \\-escape: %s"), | |
404 | yytext+1); | |
405 | else | |
406 | obstack_1grow (&obstack_for_string, c); | |
407 | } | |
408 | ||
409 | \\x[0-9abcdefABCDEF]+ { | |
410 | verify (UCHAR_MAX < ULONG_MAX); | |
411 | unsigned long int c = strtoul (yytext + 2, NULL, 16); | |
412 | if (!c || UCHAR_MAX < c) | |
413 | complain_at (*loc, _("invalid number after \\-escape: %s"), | |
414 | yytext+1); | |
415 | else | |
416 | obstack_1grow (&obstack_for_string, c); | |
417 | } | |
418 | ||
419 | \\a obstack_1grow (&obstack_for_string, '\a'); | |
420 | \\b obstack_1grow (&obstack_for_string, '\b'); | |
421 | \\f obstack_1grow (&obstack_for_string, '\f'); | |
422 | \\n obstack_1grow (&obstack_for_string, '\n'); | |
423 | \\r obstack_1grow (&obstack_for_string, '\r'); | |
424 | \\t obstack_1grow (&obstack_for_string, '\t'); | |
425 | \\v obstack_1grow (&obstack_for_string, '\v'); | |
426 | ||
427 | /* \\[\"\'?\\] would be shorter, but it confuses xgettext. */ | |
428 | \\("\""|"'"|"?"|"\\") obstack_1grow (&obstack_for_string, yytext[1]); | |
429 | ||
430 | \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} { | |
431 | int c = convert_ucn_to_byte (yytext); | |
432 | if (c <= 0) | |
433 | complain_at (*loc, _("invalid number after \\-escape: %s"), | |
434 | yytext+1); | |
435 | else | |
436 | obstack_1grow (&obstack_for_string, c); | |
437 | } | |
438 | \\(.|\n) { | |
439 | char const *p = yytext + 1; | |
440 | /* Quote only if escaping won't make the character visible. */ | |
441 | if (isspace ((unsigned char) *p) && isprint ((unsigned char) *p)) | |
442 | p = quote (p); | |
443 | else | |
444 | p = quotearg_style_mem (escape_quoting_style, p, 1); | |
445 | complain_at (*loc, _("invalid character after \\-escape: %s"), p); | |
446 | } | |
447 | } | |
448 | ||
449 | /*--------------------------------------------. | |
450 | | Scanning user-code characters and strings. | | |
451 | `--------------------------------------------*/ | |
452 | ||
453 | <SC_CHARACTER,SC_STRING> | |
454 | { | |
455 | {splice}|\\{splice}[^\n\[\]] STRING_GROW; | |
456 | } | |
457 | ||
458 | <SC_CHARACTER> | |
459 | { | |
460 | "'" STRING_GROW; BEGIN context_state; | |
461 | \n unexpected_newline (token_start, "'"); BEGIN context_state; | |
462 | <<EOF>> unexpected_eof (token_start, "'"); BEGIN context_state; | |
463 | } | |
464 | ||
465 | <SC_STRING> | |
466 | { | |
467 | "\"" STRING_GROW; BEGIN context_state; | |
468 | \n unexpected_newline (token_start, "\""); BEGIN context_state; | |
469 | <<EOF>> unexpected_eof (token_start, "\""); BEGIN context_state; | |
470 | } | |
471 | ||
472 | ||
473 | /*---------------------------------------------------. | |
474 | | Strings, comments etc. can be found in user code. | | |
475 | `---------------------------------------------------*/ | |
476 | ||
477 | <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE> | |
478 | { | |
479 | "'" { | |
480 | STRING_GROW; | |
481 | context_state = YY_START; | |
482 | token_start = loc->start; | |
483 | BEGIN SC_CHARACTER; | |
484 | } | |
485 | "\"" { | |
486 | STRING_GROW; | |
487 | context_state = YY_START; | |
488 | token_start = loc->start; | |
489 | BEGIN SC_STRING; | |
490 | } | |
491 | "/"{splice}"*" { | |
492 | STRING_GROW; | |
493 | context_state = YY_START; | |
494 | token_start = loc->start; | |
495 | BEGIN SC_COMMENT; | |
496 | } | |
497 | "/"{splice}"/" { | |
498 | STRING_GROW; | |
499 | context_state = YY_START; | |
500 | BEGIN SC_LINE_COMMENT; | |
501 | } | |
502 | } | |
503 | ||
504 | ||
505 | ||
506 | /*-----------------------------------------------------------. | |
507 | | Scanning some code in braces (actions). The initial "{" is | | |
508 | | already eaten. | | |
509 | `-----------------------------------------------------------*/ | |
510 | ||
511 | <SC_BRACED_CODE> | |
512 | { | |
513 | "{"|"<"{splice}"%" STRING_GROW; braces_level++; | |
514 | "%"{splice}">" STRING_GROW; braces_level--; | |
515 | "}" { | |
516 | obstack_1grow (&obstack_for_string, '}'); | |
517 | ||
518 | --braces_level; | |
519 | if (braces_level < 0) | |
520 | { | |
521 | STRING_FINISH; | |
522 | loc->start = code_start; | |
523 | val->code = last_string; | |
524 | BEGIN INITIAL; | |
525 | return BRACED_CODE; | |
526 | } | |
527 | } | |
528 | ||
529 | /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly | |
530 | (as `<' `<%'). */ | |
531 | "<"{splice}"<" STRING_GROW; | |
532 | ||
533 | <<EOF>> { | |
534 | unexpected_eof (code_start, "}"); | |
535 | STRING_FINISH; | |
536 | loc->start = code_start; | |
537 | val->code = last_string; | |
538 | BEGIN INITIAL; | |
539 | return BRACED_CODE; | |
540 | } | |
541 | } | |
542 | ||
543 | ||
544 | /*--------------------------------------------------------------. | |
545 | | Scanning some prologue: from "%{" (already scanned) to "%}". | | |
546 | `--------------------------------------------------------------*/ | |
547 | ||
548 | <SC_PROLOGUE> | |
549 | { | |
550 | "%}" { | |
551 | STRING_FINISH; | |
552 | loc->start = code_start; | |
553 | val->chars = last_string; | |
554 | BEGIN INITIAL; | |
555 | return PROLOGUE; | |
556 | } | |
557 | ||
558 | <<EOF>> { | |
559 | unexpected_eof (code_start, "%}"); | |
560 | STRING_FINISH; | |
561 | loc->start = code_start; | |
562 | val->chars = last_string; | |
563 | BEGIN INITIAL; | |
564 | return PROLOGUE; | |
565 | } | |
566 | } | |
567 | ||
568 | ||
569 | /*---------------------------------------------------------------. | |
570 | | Scanning the epilogue (everything after the second "%%", which | | |
571 | | has already been eaten). | | |
572 | `---------------------------------------------------------------*/ | |
573 | ||
574 | <SC_EPILOGUE> | |
575 | { | |
576 | <<EOF>> { | |
577 | STRING_FINISH; | |
578 | loc->start = code_start; | |
579 | val->chars = last_string; | |
580 | BEGIN INITIAL; | |
581 | return EPILOGUE; | |
582 | } | |
583 | } | |
584 | ||
585 | ||
586 | /*-----------------------------------------------------. | |
587 | | By default, grow the string obstack with the input. | | |
588 | `-----------------------------------------------------*/ | |
589 | ||
590 | <SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE,SC_STRING,SC_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>. | | |
591 | <SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>\n STRING_GROW; | |
592 | ||
593 | %% | |
594 | ||
595 | /* Read bytes from FP into buffer BUF of size SIZE. Return the | |
596 | number of bytes read. Remove '\r' from input, treating \r\n | |
597 | and isolated \r as \n. */ | |
598 | ||
599 | static size_t | |
600 | no_cr_read (FILE *fp, char *buf, size_t size) | |
601 | { | |
602 | size_t bytes_read = fread (buf, 1, size, fp); | |
603 | if (bytes_read) | |
604 | { | |
605 | char *w = memchr (buf, '\r', bytes_read); | |
606 | if (w) | |
607 | { | |
608 | char const *r = ++w; | |
609 | char const *lim = buf + bytes_read; | |
610 | ||
611 | for (;;) | |
612 | { | |
613 | /* Found an '\r'. Treat it like '\n', but ignore any | |
614 | '\n' that immediately follows. */ | |
615 | w[-1] = '\n'; | |
616 | if (r == lim) | |
617 | { | |
618 | int ch = getc (fp); | |
619 | if (ch != '\n' && ungetc (ch, fp) != ch) | |
620 | break; | |
621 | } | |
622 | else if (*r == '\n') | |
623 | r++; | |
624 | ||
625 | /* Copy until the next '\r'. */ | |
626 | do | |
627 | { | |
628 | if (r == lim) | |
629 | return w - buf; | |
630 | } | |
631 | while ((*w++ = *r++) != '\r'); | |
632 | } | |
633 | ||
634 | return w - buf; | |
635 | } | |
636 | } | |
637 | ||
638 | return bytes_read; | |
639 | } | |
640 | ||
641 | ||
642 | ||
643 | /*------------------------------------------------------. | |
644 | | Scan NUMBER for a base-BASE integer at location LOC. | | |
645 | `------------------------------------------------------*/ | |
646 | ||
647 | static unsigned long int | |
648 | scan_integer (char const *number, int base, location loc) | |
649 | { | |
650 | verify (INT_MAX < ULONG_MAX); | |
651 | unsigned long int num = strtoul (number, NULL, base); | |
652 | ||
653 | if (INT_MAX < num) | |
654 | { | |
655 | complain_at (loc, _("integer out of range: %s"), quote (number)); | |
656 | num = INT_MAX; | |
657 | } | |
658 | ||
659 | return num; | |
660 | } | |
661 | ||
662 | ||
663 | /*------------------------------------------------------------------. | |
664 | | Convert universal character name UCN to a single-byte character, | | |
665 | | and return that character. Return -1 if UCN does not correspond | | |
666 | | to a single-byte character. | | |
667 | `------------------------------------------------------------------*/ | |
668 | ||
669 | static int | |
670 | convert_ucn_to_byte (char const *ucn) | |
671 | { | |
672 | verify (UCHAR_MAX <= INT_MAX); | |
673 | unsigned long int code = strtoul (ucn + 2, NULL, 16); | |
674 | ||
675 | /* FIXME: Currently we assume Unicode-compatible unibyte characters | |
676 | on ASCII hosts (i.e., Latin-1 on hosts with 8-bit bytes). On | |
677 | non-ASCII hosts we support only the portable C character set. | |
678 | These limitations should be removed once we add support for | |
679 | multibyte characters. */ | |
680 | ||
681 | if (UCHAR_MAX < code) | |
682 | return -1; | |
683 | ||
684 | #if ! ('$' == 0x24 && '@' == 0x40 && '`' == 0x60 && '~' == 0x7e) | |
685 | { | |
686 | /* A non-ASCII host. Use CODE to index into a table of the C | |
687 | basic execution character set, which is guaranteed to exist on | |
688 | all Standard C platforms. This table also includes '$', '@', | |
689 | and '`', which are not in the basic execution character set but | |
690 | which are unibyte characters on all the platforms that we know | |
691 | about. */ | |
692 | static signed char const table[] = | |
693 | { | |
694 | '\0', -1, -1, -1, -1, -1, -1, '\a', | |
695 | '\b', '\t', '\n', '\v', '\f', '\r', -1, -1, | |
696 | -1, -1, -1, -1, -1, -1, -1, -1, | |
697 | -1, -1, -1, -1, -1, -1, -1, -1, | |
698 | ' ', '!', '"', '#', '$', '%', '&', '\'', | |
699 | '(', ')', '*', '+', ',', '-', '.', '/', | |
700 | '0', '1', '2', '3', '4', '5', '6', '7', | |
701 | '8', '9', ':', ';', '<', '=', '>', '?', | |
702 | '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', | |
703 | 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', | |
704 | 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', | |
705 | 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', | |
706 | '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', | |
707 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', | |
708 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', | |
709 | 'x', 'y', 'z', '{', '|', '}', '~' | |
710 | }; | |
711 | ||
712 | code = code < sizeof table ? table[code] : -1; | |
713 | } | |
714 | #endif | |
715 | ||
716 | return code; | |
717 | } | |
718 | ||
719 | ||
720 | /*----------------------------------------------------------------. | |
721 | | Handle `#line INT "FILE"'. ARGS has already skipped `#line '. | | |
722 | `----------------------------------------------------------------*/ | |
723 | ||
724 | static void | |
725 | handle_syncline (char *args, location loc) | |
726 | { | |
727 | char *after_num; | |
728 | unsigned long int lineno = strtoul (args, &after_num, 10); | |
729 | char *file = strchr (after_num, '"') + 1; | |
730 | *strchr (file, '"') = '\0'; | |
731 | if (INT_MAX <= lineno) | |
732 | { | |
733 | warn_at (loc, _("line number overflow")); | |
734 | lineno = INT_MAX; | |
735 | } | |
736 | current_file = uniqstr_new (file); | |
737 | boundary_set (&scanner_cursor, current_file, lineno, 1); | |
738 | } | |
739 | ||
740 | ||
741 | /*----------------------------------------------------------------. | |
742 | | For a token or comment starting at START, report message MSGID, | | |
743 | | which should say that an end marker was found before | | |
744 | | the expected TOKEN_END. | | |
745 | `----------------------------------------------------------------*/ | |
746 | ||
747 | static void | |
748 | unexpected_end (boundary start, char const *msgid, char const *token_end) | |
749 | { | |
750 | location loc; | |
751 | loc.start = start; | |
752 | loc.end = scanner_cursor; | |
753 | complain_at (loc, _(msgid), token_end); | |
754 | } | |
755 | ||
756 | ||
757 | /*------------------------------------------------------------------------. | |
758 | | Report an unexpected EOF in a token or comment starting at START. | | |
759 | | An end of file was encountered and the expected TOKEN_END was missing. | | |
760 | `------------------------------------------------------------------------*/ | |
761 | ||
762 | static void | |
763 | unexpected_eof (boundary start, char const *token_end) | |
764 | { | |
765 | unexpected_end (start, N_("missing `%s' at end of file"), token_end); | |
766 | } | |
767 | ||
768 | ||
769 | /*----------------------------------------. | |
770 | | Likewise, but for unexpected newlines. | | |
771 | `----------------------------------------*/ | |
772 | ||
773 | static void | |
774 | unexpected_newline (boundary start, char const *token_end) | |
775 | { | |
776 | unexpected_end (start, N_("missing `%s' at end of line"), token_end); | |
777 | } | |
778 | ||
779 | ||
780 | /*-------------------------. | |
781 | | Initialize the scanner. | | |
782 | `-------------------------*/ | |
783 | ||
784 | void | |
785 | gram_scanner_initialize (void) | |
786 | { | |
787 | obstack_init (&obstack_for_string); | |
788 | } | |
789 | ||
790 | ||
791 | /*-----------------------------------------------. | |
792 | | Free all the memory allocated to the scanner. | | |
793 | `-----------------------------------------------*/ | |
794 | ||
795 | void | |
796 | gram_scanner_free (void) | |
797 | { | |
798 | obstack_free (&obstack_for_string, 0); | |
799 | /* Reclaim Flex's buffers. */ | |
800 | yylex_destroy (); | |
801 | } |