]>
Commit | Line | Data |
---|---|---|
45b966db | 1 | /* CPP Library - lexical analysis. |
78b8811a | 2 | Copyright (C) 2000, 2001, 2002, 2003 Free Software Foundation, Inc. |
45b966db ZW |
3 | Contributed by Per Bothner, 1994-95. |
4 | Based on CCCP program by Paul Rubin, June 1986 | |
5 | Adapted to ANSI C, Richard Stallman, Jan 1987 | |
6 | Broken out to separate file, Zack Weinberg, Mar 2000 | |
7 | ||
8 | This program is free software; you can redistribute it and/or modify it | |
9 | under the terms of the GNU General Public License as published by the | |
10 | Free Software Foundation; either version 2, or (at your option) any | |
11 | 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 | |
19 | along with this program; if not, write to the Free Software | |
20 | Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ | |
21 | ||
22 | #include "config.h" | |
23 | #include "system.h" | |
45b966db ZW |
24 | #include "cpplib.h" |
25 | #include "cpphash.h" | |
26 | ||
93c80368 | 27 | enum spell_type |
f9a0e96c | 28 | { |
93c80368 | 29 | SPELL_OPERATOR = 0, |
93c80368 | 30 | SPELL_IDENT, |
6338b358 | 31 | SPELL_LITERAL, |
93c80368 | 32 | SPELL_NONE |
f9a0e96c ZW |
33 | }; |
34 | ||
93c80368 | 35 | struct token_spelling |
f9a0e96c | 36 | { |
93c80368 NB |
37 | enum spell_type category; |
38 | const unsigned char *name; | |
f9a0e96c ZW |
39 | }; |
40 | ||
8206c799 ZW |
41 | static const unsigned char *const digraph_spellings[] = |
42 | { U"%:", U"%:%:", U"<:", U":>", U"<%", U"%>" }; | |
93c80368 NB |
43 | |
44 | #define OP(e, s) { SPELL_OPERATOR, U s }, | |
9a238586 | 45 | #define TK(e, s) { s, U #e }, |
8206c799 | 46 | static const struct token_spelling token_spellings[N_TTYPES] = { TTYPE_TABLE }; |
93c80368 NB |
47 | #undef OP |
48 | #undef TK | |
49 | ||
50 | #define TOKEN_SPELL(token) (token_spellings[(token)->type].category) | |
51 | #define TOKEN_NAME(token) (token_spellings[(token)->type].name) | |
f2d5f0cc | 52 | |
6cf87ca4 ZW |
53 | static void add_line_note (cpp_buffer *, const uchar *, unsigned int); |
54 | static int skip_line_comment (cpp_reader *); | |
55 | static void skip_whitespace (cpp_reader *, cppchar_t); | |
56 | static cpp_hashnode *lex_identifier (cpp_reader *, const uchar *); | |
57 | static void lex_number (cpp_reader *, cpp_string *); | |
58 | static bool forms_identifier_p (cpp_reader *, int); | |
59 | static void lex_string (cpp_reader *, cpp_token *, const uchar *); | |
60 | static void save_comment (cpp_reader *, cpp_token *, const uchar *, cppchar_t); | |
61 | static void create_literal (cpp_reader *, cpp_token *, const uchar *, | |
62 | unsigned int, enum cpp_ttype); | |
63 | static bool warn_in_comment (cpp_reader *, _cpp_line_note *); | |
64 | static int name_p (cpp_reader *, const cpp_string *); | |
6cf87ca4 ZW |
65 | static tokenrun *next_tokenrun (tokenrun *); |
66 | ||
6cf87ca4 | 67 | static _cpp_buff *new_buff (size_t); |
15dad1d9 | 68 | |
9d10c9a9 | 69 | |
041c3194 | 70 | /* Utility routine: |
9e62c811 | 71 | |
bfb9dc7f ZW |
72 | Compares, the token TOKEN to the NUL-terminated string STRING. |
73 | TOKEN must be a CPP_NAME. Returns 1 for equal, 0 for unequal. */ | |
041c3194 | 74 | int |
6cf87ca4 | 75 | cpp_ideq (const cpp_token *token, const char *string) |
041c3194 | 76 | { |
bfb9dc7f | 77 | if (token->type != CPP_NAME) |
041c3194 | 78 | return 0; |
bfb9dc7f | 79 | |
562a5c27 | 80 | return !ustrcmp (NODE_NAME (token->val.node), (const uchar *) string); |
15dad1d9 | 81 | } |
1368ee70 | 82 | |
26aea073 NB |
83 | /* Record a note TYPE at byte POS into the current cleaned logical |
84 | line. */ | |
87062813 | 85 | static void |
6cf87ca4 | 86 | add_line_note (cpp_buffer *buffer, const uchar *pos, unsigned int type) |
0d9f234d | 87 | { |
26aea073 NB |
88 | if (buffer->notes_used == buffer->notes_cap) |
89 | { | |
90 | buffer->notes_cap = buffer->notes_cap * 2 + 200; | |
703ad42b KG |
91 | buffer->notes = xrealloc (buffer->notes, |
92 | buffer->notes_cap * sizeof (_cpp_line_note)); | |
26aea073 | 93 | } |
0d9f234d | 94 | |
26aea073 NB |
95 | buffer->notes[buffer->notes_used].pos = pos; |
96 | buffer->notes[buffer->notes_used].type = type; | |
97 | buffer->notes_used++; | |
0d9f234d NB |
98 | } |
99 | ||
26aea073 NB |
100 | /* Returns with a logical line that contains no escaped newlines or |
101 | trigraphs. This is a time-critical inner loop. */ | |
102 | void | |
6cf87ca4 | 103 | _cpp_clean_line (cpp_reader *pfile) |
45b966db | 104 | { |
26aea073 NB |
105 | cpp_buffer *buffer; |
106 | const uchar *s; | |
107 | uchar c, *d, *p; | |
87062813 | 108 | |
26aea073 NB |
109 | buffer = pfile->buffer; |
110 | buffer->cur_note = buffer->notes_used = 0; | |
111 | buffer->cur = buffer->line_base = buffer->next_line; | |
112 | buffer->need_line = false; | |
113 | s = buffer->next_line - 1; | |
87062813 | 114 | |
26aea073 | 115 | if (!buffer->from_stage3) |
45b966db | 116 | { |
d08dcf87 ZW |
117 | /* Short circuit for the common case of an un-escaped line with |
118 | no trigraphs. The primary win here is by not writing any | |
119 | data back to memory until we have to. */ | |
120 | for (;;) | |
121 | { | |
122 | c = *++s; | |
123 | if (c == '\n' || c == '\r') | |
124 | { | |
125 | d = (uchar *) s; | |
126 | ||
127 | if (s == buffer->rlimit) | |
128 | goto done; | |
129 | ||
130 | /* DOS line ending? */ | |
131 | if (c == '\r' && s[1] == '\n') | |
132 | s++; | |
133 | ||
134 | if (s == buffer->rlimit) | |
135 | goto done; | |
136 | ||
137 | /* check for escaped newline */ | |
138 | p = d; | |
139 | while (p != buffer->next_line && is_nvspace (p[-1])) | |
140 | p--; | |
141 | if (p == buffer->next_line || p[-1] != '\\') | |
142 | goto done; | |
143 | ||
144 | /* Have an escaped newline; process it and proceed to | |
145 | the slow path. */ | |
146 | add_line_note (buffer, p - 1, p != d ? ' ' : '\\'); | |
147 | d = p - 2; | |
148 | buffer->next_line = p - 1; | |
149 | break; | |
150 | } | |
151 | if (c == '?' && s[1] == '?' && _cpp_trigraph_map[s[2]]) | |
152 | { | |
153 | /* Have a trigraph. We may or may not have to convert | |
154 | it. Add a line note regardless, for -Wtrigraphs. */ | |
155 | add_line_note (buffer, s, s[2]); | |
156 | if (CPP_OPTION (pfile, trigraphs)) | |
157 | { | |
158 | /* We do, and that means we have to switch to the | |
159 | slow path. */ | |
160 | d = (uchar *) s; | |
161 | *d = _cpp_trigraph_map[s[2]]; | |
162 | s += 2; | |
163 | break; | |
164 | } | |
165 | } | |
166 | } | |
167 | ||
26aea073 NB |
168 | |
169 | for (;;) | |
4a5b68a2 | 170 | { |
26aea073 NB |
171 | c = *++s; |
172 | *++d = c; | |
173 | ||
174 | if (c == '\n' || c == '\r') | |
175 | { | |
176 | /* Handle DOS line endings. */ | |
177 | if (c == '\r' && s != buffer->rlimit && s[1] == '\n') | |
178 | s++; | |
179 | if (s == buffer->rlimit) | |
180 | break; | |
181 | ||
182 | /* Escaped? */ | |
183 | p = d; | |
184 | while (p != buffer->next_line && is_nvspace (p[-1])) | |
185 | p--; | |
186 | if (p == buffer->next_line || p[-1] != '\\') | |
187 | break; | |
188 | ||
41c32c98 | 189 | add_line_note (buffer, p - 1, p != d ? ' ': '\\'); |
26aea073 NB |
190 | d = p - 2; |
191 | buffer->next_line = p - 1; | |
192 | } | |
193 | else if (c == '?' && s[1] == '?' && _cpp_trigraph_map[s[2]]) | |
194 | { | |
195 | /* Add a note regardless, for the benefit of -Wtrigraphs. */ | |
41c32c98 | 196 | add_line_note (buffer, d, s[2]); |
26aea073 NB |
197 | if (CPP_OPTION (pfile, trigraphs)) |
198 | { | |
199 | *d = _cpp_trigraph_map[s[2]]; | |
200 | s += 2; | |
201 | } | |
202 | } | |
4a5b68a2 | 203 | } |
45b966db | 204 | } |
26aea073 NB |
205 | else |
206 | { | |
207 | do | |
208 | s++; | |
209 | while (*s != '\n' && *s != '\r'); | |
210 | d = (uchar *) s; | |
211 | ||
212 | /* Handle DOS line endings. */ | |
213 | if (*s == '\r' && s != buffer->rlimit && s[1] == '\n') | |
214 | s++; | |
215 | } | |
0d9f234d | 216 | |
d08dcf87 | 217 | done: |
26aea073 | 218 | *d = '\n'; |
41c32c98 NB |
219 | /* A sentinel note that should never be processed. */ |
220 | add_line_note (buffer, d + 1, '\n'); | |
26aea073 | 221 | buffer->next_line = s + 1; |
45b966db ZW |
222 | } |
223 | ||
a8eb6044 NB |
224 | /* Return true if the trigraph indicated by NOTE should be warned |
225 | about in a comment. */ | |
226 | static bool | |
6cf87ca4 | 227 | warn_in_comment (cpp_reader *pfile, _cpp_line_note *note) |
a8eb6044 NB |
228 | { |
229 | const uchar *p; | |
230 | ||
231 | /* Within comments we don't warn about trigraphs, unless the | |
232 | trigraph forms an escaped newline, as that may change | |
6356f892 | 233 | behavior. */ |
a8eb6044 NB |
234 | if (note->type != '/') |
235 | return false; | |
236 | ||
237 | /* If -trigraphs, then this was an escaped newline iff the next note | |
238 | is coincident. */ | |
239 | if (CPP_OPTION (pfile, trigraphs)) | |
240 | return note[1].pos == note->pos; | |
241 | ||
242 | /* Otherwise, see if this forms an escaped newline. */ | |
243 | p = note->pos + 3; | |
244 | while (is_nvspace (*p)) | |
245 | p++; | |
246 | ||
247 | /* There might have been escaped newlines between the trigraph and the | |
248 | newline we found. Hence the position test. */ | |
249 | return (*p == '\n' && p < note[1].pos); | |
250 | } | |
251 | ||
26aea073 NB |
252 | /* Process the notes created by add_line_note as far as the current |
253 | location. */ | |
254 | void | |
6cf87ca4 | 255 | _cpp_process_line_notes (cpp_reader *pfile, int in_comment) |
45b966db | 256 | { |
29401c30 NB |
257 | cpp_buffer *buffer = pfile->buffer; |
258 | ||
26aea073 | 259 | for (;;) |
041c3194 | 260 | { |
26aea073 NB |
261 | _cpp_line_note *note = &buffer->notes[buffer->cur_note]; |
262 | unsigned int col; | |
a5c3cccd | 263 | |
26aea073 NB |
264 | if (note->pos > buffer->cur) |
265 | break; | |
a5c3cccd | 266 | |
26aea073 NB |
267 | buffer->cur_note++; |
268 | col = CPP_BUF_COLUMN (buffer, note->pos + 1); | |
4d6baafa | 269 | |
41c32c98 | 270 | if (note->type == '\\' || note->type == ' ') |
26aea073 | 271 | { |
41c32c98 | 272 | if (note->type == ' ' && !in_comment) |
0527bc4e | 273 | cpp_error_with_line (pfile, CPP_DL_WARNING, pfile->line, col, |
26aea073 | 274 | "backslash and newline separated by space"); |
41c32c98 | 275 | |
26aea073 | 276 | if (buffer->next_line > buffer->rlimit) |
87062813 | 277 | { |
0527bc4e | 278 | cpp_error_with_line (pfile, CPP_DL_PEDWARN, pfile->line, col, |
26aea073 NB |
279 | "backslash-newline at end of file"); |
280 | /* Prevent "no newline at end of file" warning. */ | |
281 | buffer->next_line = buffer->rlimit; | |
87062813 | 282 | } |
26aea073 NB |
283 | |
284 | buffer->line_base = note->pos; | |
285 | pfile->line++; | |
0d9f234d | 286 | } |
41c32c98 NB |
287 | else if (_cpp_trigraph_map[note->type]) |
288 | { | |
a8eb6044 NB |
289 | if (CPP_OPTION (pfile, warn_trigraphs) |
290 | && (!in_comment || warn_in_comment (pfile, note))) | |
41c32c98 NB |
291 | { |
292 | if (CPP_OPTION (pfile, trigraphs)) | |
0527bc4e | 293 | cpp_error_with_line (pfile, CPP_DL_WARNING, pfile->line, col, |
41c32c98 NB |
294 | "trigraph ??%c converted to %c", |
295 | note->type, | |
296 | (int) _cpp_trigraph_map[note->type]); | |
297 | else | |
905bd7b5 GK |
298 | { |
299 | cpp_error_with_line | |
0527bc4e | 300 | (pfile, CPP_DL_WARNING, pfile->line, col, |
905bd7b5 GK |
301 | "trigraph ??%c ignored, use -trigraphs to enable", |
302 | note->type); | |
303 | } | |
41c32c98 NB |
304 | } |
305 | } | |
306 | else | |
307 | abort (); | |
041c3194 | 308 | } |
45b966db ZW |
309 | } |
310 | ||
0d9f234d NB |
311 | /* Skip a C-style block comment. We find the end of the comment by |
312 | seeing if an asterisk is before every '/' we encounter. Returns | |
6f572ac2 NB |
313 | nonzero if comment terminated by EOF, zero otherwise. |
314 | ||
315 | Buffer->cur points to the initial asterisk of the comment. */ | |
26aea073 | 316 | bool |
6cf87ca4 | 317 | _cpp_skip_block_comment (cpp_reader *pfile) |
45b966db | 318 | { |
041c3194 | 319 | cpp_buffer *buffer = pfile->buffer; |
d08dcf87 ZW |
320 | const uchar *cur = buffer->cur; |
321 | uchar c; | |
0d9f234d | 322 | |
d08dcf87 ZW |
323 | cur++; |
324 | if (*cur == '/') | |
325 | cur++; | |
0d9f234d | 326 | |
26aea073 NB |
327 | for (;;) |
328 | { | |
0d9f234d NB |
329 | /* People like decorating comments with '*', so check for '/' |
330 | instead for efficiency. */ | |
d08dcf87 ZW |
331 | c = *cur++; |
332 | ||
041c3194 | 333 | if (c == '/') |
45b966db | 334 | { |
d08dcf87 | 335 | if (cur[-2] == '*') |
0d9f234d | 336 | break; |
041c3194 | 337 | |
0d9f234d | 338 | /* Warn about potential nested comments, but not if the '/' |
a1f300c0 | 339 | comes immediately before the true comment delimiter. |
041c3194 | 340 | Don't bother to get it right across escaped newlines. */ |
0d9f234d | 341 | if (CPP_OPTION (pfile, warn_comments) |
d08dcf87 ZW |
342 | && cur[0] == '*' && cur[1] != '/') |
343 | { | |
344 | buffer->cur = cur; | |
0527bc4e | 345 | cpp_error_with_line (pfile, CPP_DL_WARNING, |
d08dcf87 ZW |
346 | pfile->line, CPP_BUF_COL (buffer), |
347 | "\"/*\" within comment"); | |
348 | } | |
45b966db | 349 | } |
26aea073 NB |
350 | else if (c == '\n') |
351 | { | |
d08dcf87 | 352 | buffer->cur = cur - 1; |
26aea073 NB |
353 | _cpp_process_line_notes (pfile, true); |
354 | if (buffer->next_line >= buffer->rlimit) | |
355 | return true; | |
356 | _cpp_clean_line (pfile); | |
357 | pfile->line++; | |
d08dcf87 | 358 | cur = buffer->cur; |
26aea073 | 359 | } |
45b966db | 360 | } |
041c3194 | 361 | |
d08dcf87 | 362 | buffer->cur = cur; |
a8eb6044 | 363 | _cpp_process_line_notes (pfile, true); |
26aea073 | 364 | return false; |
45b966db ZW |
365 | } |
366 | ||
480709cc | 367 | /* Skip a C++ line comment, leaving buffer->cur pointing to the |
da7d8304 | 368 | terminating newline. Handles escaped newlines. Returns nonzero |
480709cc | 369 | if a multiline comment. */ |
041c3194 | 370 | static int |
6cf87ca4 | 371 | skip_line_comment (cpp_reader *pfile) |
45b966db | 372 | { |
cbcff6df | 373 | cpp_buffer *buffer = pfile->buffer; |
67821e3a | 374 | unsigned int orig_line = pfile->line; |
041c3194 | 375 | |
26aea073 NB |
376 | while (*buffer->cur != '\n') |
377 | buffer->cur++; | |
480709cc | 378 | |
26aea073 | 379 | _cpp_process_line_notes (pfile, true); |
67821e3a | 380 | return orig_line != pfile->line; |
041c3194 | 381 | } |
45b966db | 382 | |
26aea073 | 383 | /* Skips whitespace, saving the next non-whitespace character. */ |
52fadca8 | 384 | static void |
6cf87ca4 | 385 | skip_whitespace (cpp_reader *pfile, cppchar_t c) |
041c3194 ZW |
386 | { |
387 | cpp_buffer *buffer = pfile->buffer; | |
f7d151fb | 388 | bool saw_NUL = false; |
45b966db | 389 | |
0d9f234d | 390 | do |
041c3194 | 391 | { |
91fcd158 | 392 | /* Horizontal space always OK. */ |
26aea073 | 393 | if (c == ' ' || c == '\t') |
0d9f234d | 394 | ; |
0d9f234d | 395 | /* Just \f \v or \0 left. */ |
91fcd158 | 396 | else if (c == '\0') |
f7d151fb | 397 | saw_NUL = true; |
93c80368 | 398 | else if (pfile->state.in_directive && CPP_PEDANTIC (pfile)) |
0527bc4e | 399 | cpp_error_with_line (pfile, CPP_DL_PEDWARN, pfile->line, |
ebef4e8c NB |
400 | CPP_BUF_COL (buffer), |
401 | "%s in preprocessing directive", | |
402 | c == '\f' ? "form feed" : "vertical tab"); | |
0d9f234d | 403 | |
0d9f234d | 404 | c = *buffer->cur++; |
45b966db | 405 | } |
ec5c56db | 406 | /* We only want non-vertical space, i.e. ' ' \t \f \v \0. */ |
0d9f234d NB |
407 | while (is_nvspace (c)); |
408 | ||
f7d151fb | 409 | if (saw_NUL) |
0527bc4e | 410 | cpp_error (pfile, CPP_DL_WARNING, "null character(s) ignored"); |
f7d151fb | 411 | |
480709cc | 412 | buffer->cur--; |
041c3194 | 413 | } |
45b966db | 414 | |
93c80368 NB |
415 | /* See if the characters of a number token are valid in a name (no |
416 | '.', '+' or '-'). */ | |
417 | static int | |
6cf87ca4 | 418 | name_p (cpp_reader *pfile, const cpp_string *string) |
93c80368 NB |
419 | { |
420 | unsigned int i; | |
421 | ||
422 | for (i = 0; i < string->len; i++) | |
423 | if (!is_idchar (string->text[i])) | |
424 | return 0; | |
425 | ||
df383483 | 426 | return 1; |
93c80368 NB |
427 | } |
428 | ||
bced6edf | 429 | /* Returns TRUE if the sequence starting at buffer->cur is invalid in |
1613e52b | 430 | an identifier. FIRST is TRUE if this starts an identifier. */ |
bced6edf | 431 | static bool |
6cf87ca4 | 432 | forms_identifier_p (cpp_reader *pfile, int first) |
bced6edf | 433 | { |
1613e52b NB |
434 | cpp_buffer *buffer = pfile->buffer; |
435 | ||
436 | if (*buffer->cur == '$') | |
437 | { | |
438 | if (!CPP_OPTION (pfile, dollars_in_ident)) | |
439 | return false; | |
440 | ||
441 | buffer->cur++; | |
78b8811a | 442 | if (CPP_OPTION (pfile, warn_dollars) && !pfile->state.skipping) |
1613e52b | 443 | { |
78b8811a | 444 | CPP_OPTION (pfile, warn_dollars) = 0; |
0527bc4e | 445 | cpp_error (pfile, CPP_DL_PEDWARN, "'$' in identifier or number"); |
1613e52b NB |
446 | } |
447 | ||
448 | return true; | |
449 | } | |
bced6edf | 450 | |
1613e52b NB |
451 | /* Is this a syntactically valid UCN? */ |
452 | if (0 && *buffer->cur == '\\' | |
453 | && (buffer->cur[1] == 'u' || buffer->cur[1] == 'U')) | |
bced6edf | 454 | { |
1613e52b | 455 | buffer->cur += 2; |
e6cc3a24 | 456 | if (_cpp_valid_ucn (pfile, &buffer->cur, buffer->rlimit, 1 + !first)) |
1613e52b NB |
457 | return true; |
458 | buffer->cur -= 2; | |
bced6edf | 459 | } |
bced6edf | 460 | |
1613e52b | 461 | return false; |
bced6edf NB |
462 | } |
463 | ||
464 | /* Lex an identifier starting at BUFFER->CUR - 1. */ | |
0d9f234d | 465 | static cpp_hashnode * |
6cf87ca4 | 466 | lex_identifier (cpp_reader *pfile, const uchar *base) |
45b966db | 467 | { |
93c80368 | 468 | cpp_hashnode *result; |
1613e52b | 469 | const uchar *cur; |
2c3fcba6 | 470 | |
bced6edf | 471 | do |
10cf9bde | 472 | { |
bced6edf NB |
473 | cur = pfile->buffer->cur; |
474 | ||
475 | /* N.B. ISIDNUM does not include $. */ | |
476 | while (ISIDNUM (*cur)) | |
477 | cur++; | |
10cf9bde | 478 | |
10cf9bde | 479 | pfile->buffer->cur = cur; |
2c3fcba6 | 480 | } |
1613e52b | 481 | while (forms_identifier_p (pfile, false)); |
bced6edf NB |
482 | |
483 | result = (cpp_hashnode *) | |
484 | ht_lookup (pfile->hash_table, base, cur - base, HT_ALLOC); | |
2c3fcba6 | 485 | |
bced6edf | 486 | /* Rarely, identifiers require diagnostics when lexed. */ |
2c3fcba6 ZW |
487 | if (__builtin_expect ((result->flags & NODE_DIAGNOSTIC) |
488 | && !pfile->state.skipping, 0)) | |
489 | { | |
490 | /* It is allowed to poison the same identifier twice. */ | |
491 | if ((result->flags & NODE_POISONED) && !pfile->state.poisoned_ok) | |
0527bc4e | 492 | cpp_error (pfile, CPP_DL_ERROR, "attempt to use poisoned \"%s\"", |
2c3fcba6 ZW |
493 | NODE_NAME (result)); |
494 | ||
495 | /* Constraint 6.10.3.5: __VA_ARGS__ should only appear in the | |
496 | replacement list of a variadic macro. */ | |
497 | if (result == pfile->spec_nodes.n__VA_ARGS__ | |
498 | && !pfile->state.va_args_ok) | |
0527bc4e | 499 | cpp_error (pfile, CPP_DL_PEDWARN, |
6cf87ca4 ZW |
500 | "__VA_ARGS__ can only appear in the expansion" |
501 | " of a C99 variadic macro"); | |
2c3fcba6 ZW |
502 | } |
503 | ||
504 | return result; | |
505 | } | |
506 | ||
bced6edf | 507 | /* Lex a number to NUMBER starting at BUFFER->CUR - 1. */ |
45b966db | 508 | static void |
6cf87ca4 | 509 | lex_number (cpp_reader *pfile, cpp_string *number) |
45b966db | 510 | { |
562a5c27 | 511 | const uchar *cur; |
bced6edf NB |
512 | const uchar *base; |
513 | uchar *dest; | |
45b966db | 514 | |
bced6edf NB |
515 | base = pfile->buffer->cur - 1; |
516 | do | |
041c3194 | 517 | { |
bced6edf | 518 | cur = pfile->buffer->cur; |
0d9f234d | 519 | |
bced6edf NB |
520 | /* N.B. ISIDNUM does not include $. */ |
521 | while (ISIDNUM (*cur) || *cur == '.' || VALID_SIGN (*cur, cur[-1])) | |
522 | cur++; | |
45b966db | 523 | |
10cf9bde | 524 | pfile->buffer->cur = cur; |
45b966db | 525 | } |
1613e52b | 526 | while (forms_identifier_p (pfile, false)); |
93c80368 | 527 | |
bced6edf NB |
528 | number->len = cur - base; |
529 | dest = _cpp_unaligned_alloc (pfile, number->len + 1); | |
530 | memcpy (dest, base, number->len); | |
531 | dest[number->len] = '\0'; | |
532 | number->text = dest; | |
93c80368 NB |
533 | } |
534 | ||
6338b358 NB |
535 | /* Create a token of type TYPE with a literal spelling. */ |
536 | static void | |
6cf87ca4 ZW |
537 | create_literal (cpp_reader *pfile, cpp_token *token, const uchar *base, |
538 | unsigned int len, enum cpp_ttype type) | |
6338b358 NB |
539 | { |
540 | uchar *dest = _cpp_unaligned_alloc (pfile, len + 1); | |
541 | ||
542 | memcpy (dest, base, len); | |
543 | dest[len] = '\0'; | |
544 | token->type = type; | |
545 | token->val.str.len = len; | |
546 | token->val.str.text = dest; | |
547 | } | |
548 | ||
bced6edf | 549 | /* Lexes a string, character constant, or angle-bracketed header file |
6338b358 NB |
550 | name. The stored string contains the spelling, including opening |
551 | quote and leading any leading 'L'. It returns the type of the | |
552 | literal, or CPP_OTHER if it was not properly terminated. | |
553 | ||
554 | The spelling is NUL-terminated, but it is not guaranteed that this | |
555 | is the first NUL since embedded NULs are preserved. */ | |
041c3194 | 556 | static void |
6cf87ca4 | 557 | lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base) |
45b966db | 558 | { |
6338b358 NB |
559 | bool saw_NUL = false; |
560 | const uchar *cur; | |
bced6edf | 561 | cppchar_t terminator; |
6338b358 NB |
562 | enum cpp_ttype type; |
563 | ||
564 | cur = base; | |
565 | terminator = *cur++; | |
566 | if (terminator == 'L') | |
567 | terminator = *cur++; | |
568 | if (terminator == '\"') | |
569 | type = *base == 'L' ? CPP_WSTRING: CPP_STRING; | |
570 | else if (terminator == '\'') | |
571 | type = *base == 'L' ? CPP_WCHAR: CPP_CHAR; | |
572 | else | |
573 | terminator = '>', type = CPP_HEADER_NAME; | |
93c80368 | 574 | |
0d9f234d | 575 | for (;;) |
45b966db | 576 | { |
6338b358 | 577 | cppchar_t c = *cur++; |
7868b4a2 | 578 | |
6f572ac2 | 579 | /* In #include-style directives, terminators are not escapable. */ |
6338b358 NB |
580 | if (c == '\\' && !pfile->state.angled_headers && *cur != '\n') |
581 | cur++; | |
582 | else if (c == terminator) | |
bced6edf | 583 | break; |
6338b358 | 584 | else if (c == '\n') |
0d9f234d | 585 | { |
6338b358 NB |
586 | cur--; |
587 | type = CPP_OTHER; | |
588 | break; | |
45b966db | 589 | } |
6338b358 NB |
590 | else if (c == '\0') |
591 | saw_NUL = true; | |
45b966db ZW |
592 | } |
593 | ||
6338b358 | 594 | if (saw_NUL && !pfile->state.skipping) |
0527bc4e JDA |
595 | cpp_error (pfile, CPP_DL_WARNING, |
596 | "null character(s) preserved in literal"); | |
45b966db | 597 | |
6338b358 NB |
598 | pfile->buffer->cur = cur; |
599 | create_literal (pfile, token, base, cur - base, type); | |
0d9f234d | 600 | } |
041c3194 | 601 | |
93c80368 | 602 | /* The stored comment includes the comment start and any terminator. */ |
9e62c811 | 603 | static void |
6cf87ca4 ZW |
604 | save_comment (cpp_reader *pfile, cpp_token *token, const unsigned char *from, |
605 | cppchar_t type) | |
9e62c811 | 606 | { |
041c3194 | 607 | unsigned char *buffer; |
477cdac7 | 608 | unsigned int len, clen; |
df383483 | 609 | |
1c6d33ef | 610 | len = pfile->buffer->cur - from + 1; /* + 1 for the initial '/'. */ |
480709cc | 611 | |
3542203b NB |
612 | /* C++ comments probably (not definitely) have moved past a new |
613 | line, which we don't want to save in the comment. */ | |
480709cc | 614 | if (is_vspace (pfile->buffer->cur[-1])) |
3542203b | 615 | len--; |
477cdac7 JT |
616 | |
617 | /* If we are currently in a directive, then we need to store all | |
618 | C++ comments as C comments internally, and so we need to | |
619 | allocate a little extra space in that case. | |
620 | ||
621 | Note that the only time we encounter a directive here is | |
622 | when we are saving comments in a "#define". */ | |
623 | clen = (pfile->state.in_directive && type == '/') ? len + 2 : len; | |
624 | ||
625 | buffer = _cpp_unaligned_alloc (pfile, clen); | |
df383483 | 626 | |
041c3194 | 627 | token->type = CPP_COMMENT; |
477cdac7 | 628 | token->val.str.len = clen; |
0d9f234d | 629 | token->val.str.text = buffer; |
45b966db | 630 | |
1c6d33ef NB |
631 | buffer[0] = '/'; |
632 | memcpy (buffer + 1, from, len - 1); | |
477cdac7 | 633 | |
1eeeb6a4 | 634 | /* Finish conversion to a C comment, if necessary. */ |
477cdac7 JT |
635 | if (pfile->state.in_directive && type == '/') |
636 | { | |
637 | buffer[1] = '*'; | |
638 | buffer[clen - 2] = '*'; | |
639 | buffer[clen - 1] = '/'; | |
640 | } | |
0d9f234d | 641 | } |
45b966db | 642 | |
5fddcffc NB |
643 | /* Allocate COUNT tokens for RUN. */ |
644 | void | |
6cf87ca4 | 645 | _cpp_init_tokenrun (tokenrun *run, unsigned int count) |
5fddcffc NB |
646 | { |
647 | run->base = xnewvec (cpp_token, count); | |
648 | run->limit = run->base + count; | |
649 | run->next = NULL; | |
650 | } | |
651 | ||
652 | /* Returns the next tokenrun, or creates one if there is none. */ | |
653 | static tokenrun * | |
6cf87ca4 | 654 | next_tokenrun (tokenrun *run) |
5fddcffc NB |
655 | { |
656 | if (run->next == NULL) | |
657 | { | |
658 | run->next = xnew (tokenrun); | |
bdcbe496 | 659 | run->next->prev = run; |
5fddcffc NB |
660 | _cpp_init_tokenrun (run->next, 250); |
661 | } | |
662 | ||
663 | return run->next; | |
664 | } | |
665 | ||
4ed5bcfb NB |
666 | /* Allocate a single token that is invalidated at the same time as the |
667 | rest of the tokens on the line. Has its line and col set to the | |
668 | same as the last lexed token, so that diagnostics appear in the | |
669 | right place. */ | |
670 | cpp_token * | |
6cf87ca4 | 671 | _cpp_temp_token (cpp_reader *pfile) |
4ed5bcfb NB |
672 | { |
673 | cpp_token *old, *result; | |
674 | ||
675 | old = pfile->cur_token - 1; | |
676 | if (pfile->cur_token == pfile->cur_run->limit) | |
677 | { | |
678 | pfile->cur_run = next_tokenrun (pfile->cur_run); | |
679 | pfile->cur_token = pfile->cur_run->base; | |
680 | } | |
681 | ||
682 | result = pfile->cur_token++; | |
683 | result->line = old->line; | |
684 | result->col = old->col; | |
685 | return result; | |
686 | } | |
687 | ||
14baae01 NB |
688 | /* Lex a token into RESULT (external interface). Takes care of issues |
689 | like directive handling, token lookahead, multiple include | |
a1f300c0 | 690 | optimization and skipping. */ |
345894b4 | 691 | const cpp_token * |
6cf87ca4 | 692 | _cpp_lex_token (cpp_reader *pfile) |
5fddcffc | 693 | { |
bdcbe496 | 694 | cpp_token *result; |
5fddcffc | 695 | |
bdcbe496 | 696 | for (;;) |
5fddcffc | 697 | { |
bdcbe496 | 698 | if (pfile->cur_token == pfile->cur_run->limit) |
5fddcffc | 699 | { |
bdcbe496 NB |
700 | pfile->cur_run = next_tokenrun (pfile->cur_run); |
701 | pfile->cur_token = pfile->cur_run->base; | |
5fddcffc NB |
702 | } |
703 | ||
bdcbe496 | 704 | if (pfile->lookaheads) |
14baae01 NB |
705 | { |
706 | pfile->lookaheads--; | |
707 | result = pfile->cur_token++; | |
708 | } | |
bdcbe496 | 709 | else |
14baae01 | 710 | result = _cpp_lex_direct (pfile); |
bdcbe496 NB |
711 | |
712 | if (result->flags & BOL) | |
5fddcffc | 713 | { |
bdcbe496 NB |
714 | /* Is this a directive. If _cpp_handle_directive returns |
715 | false, it is an assembler #. */ | |
716 | if (result->type == CPP_HASH | |
e808ec9c NB |
717 | /* 6.10.3 p 11: Directives in a list of macro arguments |
718 | gives undefined behavior. This implementation | |
719 | handles the directive as normal. */ | |
720 | && pfile->state.parsing_args != 1 | |
bdcbe496 NB |
721 | && _cpp_handle_directive (pfile, result->flags & PREV_WHITE)) |
722 | continue; | |
97293897 | 723 | if (pfile->cb.line_change && !pfile->state.skipping) |
6cf87ca4 | 724 | pfile->cb.line_change (pfile, result, pfile->state.parsing_args); |
5fddcffc | 725 | } |
5fddcffc | 726 | |
bdcbe496 NB |
727 | /* We don't skip tokens in directives. */ |
728 | if (pfile->state.in_directive) | |
729 | break; | |
5fddcffc | 730 | |
bdcbe496 | 731 | /* Outside a directive, invalidate controlling macros. At file |
14baae01 | 732 | EOF, _cpp_lex_direct takes care of popping the buffer, so we never |
6356f892 | 733 | get here and MI optimization works. */ |
5fddcffc | 734 | pfile->mi_valid = false; |
bdcbe496 NB |
735 | |
736 | if (!pfile->state.skipping || result->type == CPP_EOF) | |
737 | break; | |
5fddcffc NB |
738 | } |
739 | ||
345894b4 | 740 | return result; |
5fddcffc NB |
741 | } |
742 | ||
26aea073 NB |
743 | /* Returns true if a fresh line has been loaded. */ |
744 | bool | |
6cf87ca4 | 745 | _cpp_get_fresh_line (cpp_reader *pfile) |
004cb263 | 746 | { |
26aea073 NB |
747 | /* We can't get a new line until we leave the current directive. */ |
748 | if (pfile->state.in_directive) | |
749 | return false; | |
df383483 | 750 | |
26aea073 | 751 | for (;;) |
1a76916c | 752 | { |
26aea073 | 753 | cpp_buffer *buffer = pfile->buffer; |
1a76916c | 754 | |
26aea073 NB |
755 | if (!buffer->need_line) |
756 | return true; | |
757 | ||
758 | if (buffer->next_line < buffer->rlimit) | |
004cb263 | 759 | { |
26aea073 NB |
760 | _cpp_clean_line (pfile); |
761 | return true; | |
762 | } | |
004cb263 | 763 | |
26aea073 NB |
764 | /* First, get out of parsing arguments state. */ |
765 | if (pfile->state.parsing_args) | |
766 | return false; | |
767 | ||
768 | /* End of buffer. Non-empty files should end in a newline. */ | |
769 | if (buffer->buf != buffer->rlimit | |
770 | && buffer->next_line > buffer->rlimit | |
771 | && !buffer->from_stage3) | |
772 | { | |
773 | /* Only warn once. */ | |
774 | buffer->next_line = buffer->rlimit; | |
0527bc4e | 775 | cpp_error_with_line (pfile, CPP_DL_PEDWARN, pfile->line - 1, |
26aea073 NB |
776 | CPP_BUF_COLUMN (buffer, buffer->cur), |
777 | "no newline at end of file"); | |
778 | } | |
779 | ||
26aea073 | 780 | _cpp_pop_buffer (pfile); |
a506c55c PB |
781 | if (pfile->buffer == NULL) |
782 | return false; | |
26aea073 | 783 | } |
004cb263 NB |
784 | } |
785 | ||
6f572ac2 NB |
786 | #define IF_NEXT_IS(CHAR, THEN_TYPE, ELSE_TYPE) \ |
787 | do \ | |
788 | { \ | |
789 | result->type = ELSE_TYPE; \ | |
790 | if (*buffer->cur == CHAR) \ | |
791 | buffer->cur++, result->type = THEN_TYPE; \ | |
792 | } \ | |
793 | while (0) | |
480709cc | 794 | |
14baae01 NB |
795 | /* Lex a token into pfile->cur_token, which is also incremented, to |
796 | get diagnostics pointing to the correct location. | |
797 | ||
798 | Does not handle issues such as token lookahead, multiple-include | |
f1ba665b | 799 | optimization, directives, skipping etc. This function is only |
14baae01 NB |
800 | suitable for use by _cpp_lex_token, and in special cases like |
801 | lex_expansion_token which doesn't care for any of these issues. | |
802 | ||
803 | When meeting a newline, returns CPP_EOF if parsing a directive, | |
804 | otherwise returns to the start of the token buffer if permissible. | |
805 | Returns the location of the lexed token. */ | |
806 | cpp_token * | |
6cf87ca4 | 807 | _cpp_lex_direct (cpp_reader *pfile) |
45b966db | 808 | { |
0d9f234d | 809 | cppchar_t c; |
adb84b42 | 810 | cpp_buffer *buffer; |
0d9f234d | 811 | const unsigned char *comment_start; |
14baae01 | 812 | cpp_token *result = pfile->cur_token++; |
9ec7291f | 813 | |
5fddcffc | 814 | fresh_line: |
26aea073 | 815 | result->flags = 0; |
2be570f9 | 816 | buffer = pfile->buffer; |
a506c55c | 817 | if (buffer->need_line) |
26aea073 NB |
818 | { |
819 | if (!_cpp_get_fresh_line (pfile)) | |
820 | { | |
821 | result->type = CPP_EOF; | |
9ff7868d NB |
822 | if (!pfile->state.in_directive) |
823 | { | |
824 | /* Tell the compiler the line number of the EOF token. */ | |
825 | result->line = pfile->line; | |
826 | result->flags = BOL; | |
827 | } | |
26aea073 NB |
828 | return result; |
829 | } | |
830 | if (!pfile->keep_tokens) | |
831 | { | |
832 | pfile->cur_run = &pfile->base_run; | |
833 | result = pfile->base_run.base; | |
834 | pfile->cur_token = result + 1; | |
835 | } | |
836 | result->flags = BOL; | |
837 | if (pfile->state.parsing_args == 2) | |
838 | result->flags |= PREV_WHITE; | |
839 | } | |
a506c55c | 840 | buffer = pfile->buffer; |
5fddcffc | 841 | update_tokens_line: |
1444f2ed | 842 | result->line = pfile->line; |
041c3194 | 843 | |
5fddcffc | 844 | skipped_white: |
26aea073 NB |
845 | if (buffer->cur >= buffer->notes[buffer->cur_note].pos |
846 | && !pfile->overlaid_buffer) | |
847 | { | |
848 | _cpp_process_line_notes (pfile, false); | |
849 | result->line = pfile->line; | |
850 | } | |
480709cc | 851 | c = *buffer->cur++; |
5fddcffc | 852 | result->col = CPP_BUF_COLUMN (buffer, buffer->cur); |
5fddcffc | 853 | |
0d9f234d | 854 | switch (c) |
45b966db | 855 | { |
4d6baafa NB |
856 | case ' ': case '\t': case '\f': case '\v': case '\0': |
857 | result->flags |= PREV_WHITE; | |
26aea073 NB |
858 | skip_whitespace (pfile, c); |
859 | goto skipped_white; | |
0d9f234d | 860 | |
26aea073 NB |
861 | case '\n': |
862 | pfile->line++; | |
863 | buffer->need_line = true; | |
864 | goto fresh_line; | |
46d07497 | 865 | |
0d9f234d NB |
866 | case '0': case '1': case '2': case '3': case '4': |
867 | case '5': case '6': case '7': case '8': case '9': | |
868 | result->type = CPP_NUMBER; | |
bced6edf | 869 | lex_number (pfile, &result->val.str); |
0d9f234d | 870 | break; |
46d07497 | 871 | |
0abc6a6a NB |
872 | case 'L': |
873 | /* 'L' may introduce wide characters or strings. */ | |
bced6edf NB |
874 | if (*buffer->cur == '\'' || *buffer->cur == '"') |
875 | { | |
6338b358 | 876 | lex_string (pfile, result, buffer->cur - 1); |
bced6edf NB |
877 | break; |
878 | } | |
df383483 | 879 | /* Fall through. */ |
0abc6a6a | 880 | |
0d9f234d NB |
881 | case '_': |
882 | case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': | |
883 | case 'g': case 'h': case 'i': case 'j': case 'k': case 'l': | |
884 | case 'm': case 'n': case 'o': case 'p': case 'q': case 'r': | |
885 | case 's': case 't': case 'u': case 'v': case 'w': case 'x': | |
886 | case 'y': case 'z': | |
887 | case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': | |
0abc6a6a | 888 | case 'G': case 'H': case 'I': case 'J': case 'K': |
0d9f234d NB |
889 | case 'M': case 'N': case 'O': case 'P': case 'Q': case 'R': |
890 | case 'S': case 'T': case 'U': case 'V': case 'W': case 'X': | |
891 | case 'Y': case 'Z': | |
892 | result->type = CPP_NAME; | |
1613e52b | 893 | result->val.node = lex_identifier (pfile, buffer->cur - 1); |
0d9f234d | 894 | |
0d9f234d | 895 | /* Convert named operators to their proper types. */ |
0abc6a6a | 896 | if (result->val.node->flags & NODE_OPERATOR) |
0d9f234d NB |
897 | { |
898 | result->flags |= NAMED_OP; | |
4977bab6 | 899 | result->type = result->val.node->directive_index; |
0d9f234d NB |
900 | } |
901 | break; | |
902 | ||
903 | case '\'': | |
904 | case '"': | |
6338b358 | 905 | lex_string (pfile, result, buffer->cur - 1); |
0d9f234d | 906 | break; |
041c3194 | 907 | |
0d9f234d | 908 | case '/': |
1c6d33ef NB |
909 | /* A potential block or line comment. */ |
910 | comment_start = buffer->cur; | |
6f572ac2 NB |
911 | c = *buffer->cur; |
912 | ||
1c6d33ef NB |
913 | if (c == '*') |
914 | { | |
26aea073 | 915 | if (_cpp_skip_block_comment (pfile)) |
0527bc4e | 916 | cpp_error (pfile, CPP_DL_ERROR, "unterminated comment"); |
0d9f234d | 917 | } |
480709cc NB |
918 | else if (c == '/' && (CPP_OPTION (pfile, cplusplus_comments) |
919 | || CPP_IN_SYSTEM_HEADER (pfile))) | |
0d9f234d | 920 | { |
bdb05a7b NB |
921 | /* Warn about comments only if pedantically GNUC89, and not |
922 | in system headers. */ | |
923 | if (CPP_OPTION (pfile, lang) == CLK_GNUC89 && CPP_PEDANTIC (pfile) | |
a94c1199 | 924 | && ! buffer->warned_cplusplus_comments) |
041c3194 | 925 | { |
0527bc4e | 926 | cpp_error (pfile, CPP_DL_PEDWARN, |
56508306 | 927 | "C++ style comments are not allowed in ISO C90"); |
0527bc4e | 928 | cpp_error (pfile, CPP_DL_PEDWARN, |
ebef4e8c | 929 | "(this will be reported only once per input file)"); |
1c6d33ef NB |
930 | buffer->warned_cplusplus_comments = 1; |
931 | } | |
0d9f234d | 932 | |
01ef6563 | 933 | if (skip_line_comment (pfile) && CPP_OPTION (pfile, warn_comments)) |
0527bc4e | 934 | cpp_error (pfile, CPP_DL_WARNING, "multi-line comment"); |
1c6d33ef | 935 | } |
480709cc NB |
936 | else if (c == '=') |
937 | { | |
6f572ac2 | 938 | buffer->cur++; |
480709cc NB |
939 | result->type = CPP_DIV_EQ; |
940 | break; | |
941 | } | |
942 | else | |
943 | { | |
480709cc NB |
944 | result->type = CPP_DIV; |
945 | break; | |
946 | } | |
0d9f234d | 947 | |
1c6d33ef NB |
948 | if (!pfile->state.save_comments) |
949 | { | |
950 | result->flags |= PREV_WHITE; | |
5fddcffc | 951 | goto update_tokens_line; |
0d9f234d | 952 | } |
1c6d33ef NB |
953 | |
954 | /* Save the comment as a token in its own right. */ | |
477cdac7 | 955 | save_comment (pfile, result, comment_start, c); |
bdcbe496 | 956 | break; |
0d9f234d NB |
957 | |
958 | case '<': | |
959 | if (pfile->state.angled_headers) | |
960 | { | |
6338b358 | 961 | lex_string (pfile, result, buffer->cur - 1); |
480709cc | 962 | break; |
0d9f234d | 963 | } |
45b966db | 964 | |
6f572ac2 NB |
965 | result->type = CPP_LESS; |
966 | if (*buffer->cur == '=') | |
967 | buffer->cur++, result->type = CPP_LESS_EQ; | |
968 | else if (*buffer->cur == '<') | |
0d9f234d | 969 | { |
6f572ac2 NB |
970 | buffer->cur++; |
971 | IF_NEXT_IS ('=', CPP_LSHIFT_EQ, CPP_LSHIFT); | |
0d9f234d | 972 | } |
6f572ac2 | 973 | else if (*buffer->cur == '?' && CPP_OPTION (pfile, cplusplus)) |
0d9f234d | 974 | { |
6f572ac2 NB |
975 | buffer->cur++; |
976 | IF_NEXT_IS ('=', CPP_MIN_EQ, CPP_MIN); | |
0d9f234d | 977 | } |
6f572ac2 | 978 | else if (CPP_OPTION (pfile, digraphs)) |
480709cc | 979 | { |
6f572ac2 NB |
980 | if (*buffer->cur == ':') |
981 | { | |
982 | buffer->cur++; | |
983 | result->flags |= DIGRAPH; | |
984 | result->type = CPP_OPEN_SQUARE; | |
985 | } | |
986 | else if (*buffer->cur == '%') | |
987 | { | |
988 | buffer->cur++; | |
989 | result->flags |= DIGRAPH; | |
990 | result->type = CPP_OPEN_BRACE; | |
991 | } | |
480709cc | 992 | } |
0d9f234d NB |
993 | break; |
994 | ||
995 | case '>': | |
6f572ac2 NB |
996 | result->type = CPP_GREATER; |
997 | if (*buffer->cur == '=') | |
998 | buffer->cur++, result->type = CPP_GREATER_EQ; | |
999 | else if (*buffer->cur == '>') | |
0d9f234d | 1000 | { |
6f572ac2 NB |
1001 | buffer->cur++; |
1002 | IF_NEXT_IS ('=', CPP_RSHIFT_EQ, CPP_RSHIFT); | |
1003 | } | |
1004 | else if (*buffer->cur == '?' && CPP_OPTION (pfile, cplusplus)) | |
1005 | { | |
1006 | buffer->cur++; | |
1007 | IF_NEXT_IS ('=', CPP_MAX_EQ, CPP_MAX); | |
0d9f234d NB |
1008 | } |
1009 | break; | |
1010 | ||
cbcff6df | 1011 | case '%': |
6f572ac2 NB |
1012 | result->type = CPP_MOD; |
1013 | if (*buffer->cur == '=') | |
1014 | buffer->cur++, result->type = CPP_MOD_EQ; | |
1015 | else if (CPP_OPTION (pfile, digraphs)) | |
480709cc | 1016 | { |
6f572ac2 | 1017 | if (*buffer->cur == ':') |
480709cc | 1018 | { |
6f572ac2 NB |
1019 | buffer->cur++; |
1020 | result->flags |= DIGRAPH; | |
1021 | result->type = CPP_HASH; | |
1022 | if (*buffer->cur == '%' && buffer->cur[1] == ':') | |
1023 | buffer->cur += 2, result->type = CPP_PASTE; | |
1024 | } | |
1025 | else if (*buffer->cur == '>') | |
1026 | { | |
1027 | buffer->cur++; | |
1028 | result->flags |= DIGRAPH; | |
1029 | result->type = CPP_CLOSE_BRACE; | |
480709cc | 1030 | } |
480709cc | 1031 | } |
0d9f234d NB |
1032 | break; |
1033 | ||
cbcff6df | 1034 | case '.': |
480709cc | 1035 | result->type = CPP_DOT; |
6f572ac2 | 1036 | if (ISDIGIT (*buffer->cur)) |
480709cc NB |
1037 | { |
1038 | result->type = CPP_NUMBER; | |
bced6edf | 1039 | lex_number (pfile, &result->val.str); |
480709cc | 1040 | } |
6f572ac2 NB |
1041 | else if (*buffer->cur == '.' && buffer->cur[1] == '.') |
1042 | buffer->cur += 2, result->type = CPP_ELLIPSIS; | |
1043 | else if (*buffer->cur == '*' && CPP_OPTION (pfile, cplusplus)) | |
1044 | buffer->cur++, result->type = CPP_DOT_STAR; | |
0d9f234d | 1045 | break; |
45b966db | 1046 | |
0d9f234d | 1047 | case '+': |
6f572ac2 NB |
1048 | result->type = CPP_PLUS; |
1049 | if (*buffer->cur == '+') | |
1050 | buffer->cur++, result->type = CPP_PLUS_PLUS; | |
1051 | else if (*buffer->cur == '=') | |
1052 | buffer->cur++, result->type = CPP_PLUS_EQ; | |
0d9f234d | 1053 | break; |
04e3ec78 | 1054 | |
0d9f234d | 1055 | case '-': |
6f572ac2 NB |
1056 | result->type = CPP_MINUS; |
1057 | if (*buffer->cur == '>') | |
0d9f234d | 1058 | { |
6f572ac2 | 1059 | buffer->cur++; |
480709cc | 1060 | result->type = CPP_DEREF; |
6f572ac2 NB |
1061 | if (*buffer->cur == '*' && CPP_OPTION (pfile, cplusplus)) |
1062 | buffer->cur++, result->type = CPP_DEREF_STAR; | |
480709cc | 1063 | } |
6f572ac2 NB |
1064 | else if (*buffer->cur == '-') |
1065 | buffer->cur++, result->type = CPP_MINUS_MINUS; | |
1066 | else if (*buffer->cur == '=') | |
1067 | buffer->cur++, result->type = CPP_MINUS_EQ; | |
0d9f234d | 1068 | break; |
45b966db | 1069 | |
0d9f234d | 1070 | case '&': |
6f572ac2 NB |
1071 | result->type = CPP_AND; |
1072 | if (*buffer->cur == '&') | |
1073 | buffer->cur++, result->type = CPP_AND_AND; | |
1074 | else if (*buffer->cur == '=') | |
1075 | buffer->cur++, result->type = CPP_AND_EQ; | |
0d9f234d | 1076 | break; |
df383483 | 1077 | |
0d9f234d | 1078 | case '|': |
6f572ac2 NB |
1079 | result->type = CPP_OR; |
1080 | if (*buffer->cur == '|') | |
1081 | buffer->cur++, result->type = CPP_OR_OR; | |
1082 | else if (*buffer->cur == '=') | |
1083 | buffer->cur++, result->type = CPP_OR_EQ; | |
0d9f234d | 1084 | break; |
45b966db | 1085 | |
0d9f234d | 1086 | case ':': |
6f572ac2 NB |
1087 | result->type = CPP_COLON; |
1088 | if (*buffer->cur == ':' && CPP_OPTION (pfile, cplusplus)) | |
1089 | buffer->cur++, result->type = CPP_SCOPE; | |
1090 | else if (*buffer->cur == '>' && CPP_OPTION (pfile, digraphs)) | |
0d9f234d | 1091 | { |
6f572ac2 | 1092 | buffer->cur++; |
0d9f234d | 1093 | result->flags |= DIGRAPH; |
480709cc NB |
1094 | result->type = CPP_CLOSE_SQUARE; |
1095 | } | |
0d9f234d | 1096 | break; |
45b966db | 1097 | |
480709cc NB |
1098 | case '*': IF_NEXT_IS ('=', CPP_MULT_EQ, CPP_MULT); break; |
1099 | case '=': IF_NEXT_IS ('=', CPP_EQ_EQ, CPP_EQ); break; | |
1100 | case '!': IF_NEXT_IS ('=', CPP_NOT_EQ, CPP_NOT); break; | |
1101 | case '^': IF_NEXT_IS ('=', CPP_XOR_EQ, CPP_XOR); break; | |
1102 | case '#': IF_NEXT_IS ('#', CPP_PASTE, CPP_HASH); break; | |
1103 | ||
26aea073 | 1104 | case '?': result->type = CPP_QUERY; break; |
0d9f234d NB |
1105 | case '~': result->type = CPP_COMPL; break; |
1106 | case ',': result->type = CPP_COMMA; break; | |
1107 | case '(': result->type = CPP_OPEN_PAREN; break; | |
1108 | case ')': result->type = CPP_CLOSE_PAREN; break; | |
1109 | case '[': result->type = CPP_OPEN_SQUARE; break; | |
1110 | case ']': result->type = CPP_CLOSE_SQUARE; break; | |
1111 | case '{': result->type = CPP_OPEN_BRACE; break; | |
1112 | case '}': result->type = CPP_CLOSE_BRACE; break; | |
1113 | case ';': result->type = CPP_SEMICOLON; break; | |
1114 | ||
40f03658 | 1115 | /* @ is a punctuator in Objective-C. */ |
cc937581 | 1116 | case '@': result->type = CPP_ATSIGN; break; |
0d9f234d | 1117 | |
0abc6a6a | 1118 | case '$': |
1613e52b NB |
1119 | case '\\': |
1120 | { | |
1121 | const uchar *base = --buffer->cur; | |
0abc6a6a | 1122 | |
1613e52b NB |
1123 | if (forms_identifier_p (pfile, true)) |
1124 | { | |
1125 | result->type = CPP_NAME; | |
1126 | result->val.node = lex_identifier (pfile, base); | |
1127 | break; | |
1128 | } | |
1129 | buffer->cur++; | |
1067694a | 1130 | } |
1613e52b | 1131 | |
1067694a | 1132 | default: |
6338b358 NB |
1133 | create_literal (pfile, result, buffer->cur - 1, 1, CPP_OTHER); |
1134 | break; | |
0d9f234d | 1135 | } |
bdcbe496 NB |
1136 | |
1137 | return result; | |
0d9f234d NB |
1138 | } |
1139 | ||
59325650 NB |
1140 | /* An upper bound on the number of bytes needed to spell TOKEN. |
1141 | Does not include preceding whitespace. */ | |
93c80368 | 1142 | unsigned int |
6cf87ca4 | 1143 | cpp_token_len (const cpp_token *token) |
0d9f234d | 1144 | { |
93c80368 | 1145 | unsigned int len; |
6d2c2047 | 1146 | |
93c80368 | 1147 | switch (TOKEN_SPELL (token)) |
041c3194 | 1148 | { |
59325650 | 1149 | default: len = 4; break; |
6338b358 | 1150 | case SPELL_LITERAL: len = token->val.str.len; break; |
a28c5035 | 1151 | case SPELL_IDENT: len = NODE_LEN (token->val.node); break; |
041c3194 | 1152 | } |
59325650 NB |
1153 | |
1154 | return len; | |
6d2c2047 ZW |
1155 | } |
1156 | ||
041c3194 | 1157 | /* Write the spelling of a token TOKEN to BUFFER. The buffer must |
cf00a885 | 1158 | already contain the enough space to hold the token's spelling. |
6cf87ca4 ZW |
1159 | Returns a pointer to the character after the last character written. |
1160 | FIXME: Would be nice if we didn't need the PFILE argument. */ | |
93c80368 | 1161 | unsigned char * |
6cf87ca4 ZW |
1162 | cpp_spell_token (cpp_reader *pfile, const cpp_token *token, |
1163 | unsigned char *buffer) | |
041c3194 | 1164 | { |
96be6998 | 1165 | switch (TOKEN_SPELL (token)) |
041c3194 ZW |
1166 | { |
1167 | case SPELL_OPERATOR: | |
1168 | { | |
1169 | const unsigned char *spelling; | |
1170 | unsigned char c; | |
d6d5f795 | 1171 | |
041c3194 | 1172 | if (token->flags & DIGRAPH) |
37b8524c JDA |
1173 | spelling |
1174 | = digraph_spellings[(int) token->type - (int) CPP_FIRST_DIGRAPH]; | |
92936ecf ZW |
1175 | else if (token->flags & NAMED_OP) |
1176 | goto spell_ident; | |
041c3194 | 1177 | else |
96be6998 | 1178 | spelling = TOKEN_NAME (token); |
df383483 | 1179 | |
041c3194 ZW |
1180 | while ((c = *spelling++) != '\0') |
1181 | *buffer++ = c; | |
1182 | } | |
1183 | break; | |
d6d5f795 | 1184 | |
47ad4138 | 1185 | spell_ident: |
041c3194 | 1186 | case SPELL_IDENT: |
a28c5035 NB |
1187 | memcpy (buffer, NODE_NAME (token->val.node), NODE_LEN (token->val.node)); |
1188 | buffer += NODE_LEN (token->val.node); | |
041c3194 | 1189 | break; |
d6d5f795 | 1190 | |
6338b358 | 1191 | case SPELL_LITERAL: |
47ad4138 ZW |
1192 | memcpy (buffer, token->val.str.text, token->val.str.len); |
1193 | buffer += token->val.str.len; | |
1194 | break; | |
1195 | ||
041c3194 | 1196 | case SPELL_NONE: |
0527bc4e JDA |
1197 | cpp_error (pfile, CPP_DL_ICE, |
1198 | "unspellable token %s", TOKEN_NAME (token)); | |
041c3194 ZW |
1199 | break; |
1200 | } | |
d6d5f795 | 1201 | |
041c3194 ZW |
1202 | return buffer; |
1203 | } | |
d6d5f795 | 1204 | |
5d8ebbd8 NB |
1205 | /* Returns TOKEN spelt as a null-terminated string. The string is |
1206 | freed when the reader is destroyed. Useful for diagnostics. */ | |
93c80368 | 1207 | unsigned char * |
6cf87ca4 | 1208 | cpp_token_as_text (cpp_reader *pfile, const cpp_token *token) |
59325650 NB |
1209 | { |
1210 | unsigned int len = cpp_token_len (token) + 1; | |
ece54d54 | 1211 | unsigned char *start = _cpp_unaligned_alloc (pfile, len), *end; |
c5a04734 | 1212 | |
93c80368 NB |
1213 | end = cpp_spell_token (pfile, token, start); |
1214 | end[0] = '\0'; | |
c5a04734 | 1215 | |
93c80368 NB |
1216 | return start; |
1217 | } | |
c5a04734 | 1218 | |
5d8ebbd8 NB |
1219 | /* Used by C front ends, which really should move to using |
1220 | cpp_token_as_text. */ | |
93c80368 | 1221 | const char * |
6cf87ca4 | 1222 | cpp_type2name (enum cpp_ttype type) |
93c80368 NB |
1223 | { |
1224 | return (const char *) token_spellings[type].name; | |
1225 | } | |
c5a04734 | 1226 | |
4ed5bcfb NB |
1227 | /* Writes the spelling of token to FP, without any preceding space. |
1228 | Separated from cpp_spell_token for efficiency - to avoid stdio | |
1229 | double-buffering. */ | |
93c80368 | 1230 | void |
6cf87ca4 | 1231 | cpp_output_token (const cpp_token *token, FILE *fp) |
93c80368 | 1232 | { |
93c80368 | 1233 | switch (TOKEN_SPELL (token)) |
c5a04734 | 1234 | { |
93c80368 NB |
1235 | case SPELL_OPERATOR: |
1236 | { | |
1237 | const unsigned char *spelling; | |
3b681e9d | 1238 | int c; |
c5a04734 | 1239 | |
93c80368 | 1240 | if (token->flags & DIGRAPH) |
37b8524c JDA |
1241 | spelling |
1242 | = digraph_spellings[(int) token->type - (int) CPP_FIRST_DIGRAPH]; | |
93c80368 NB |
1243 | else if (token->flags & NAMED_OP) |
1244 | goto spell_ident; | |
1245 | else | |
1246 | spelling = TOKEN_NAME (token); | |
041c3194 | 1247 | |
3b681e9d ZW |
1248 | c = *spelling; |
1249 | do | |
1250 | putc (c, fp); | |
1251 | while ((c = *++spelling) != '\0'); | |
93c80368 NB |
1252 | } |
1253 | break; | |
041c3194 | 1254 | |
93c80368 NB |
1255 | spell_ident: |
1256 | case SPELL_IDENT: | |
3b681e9d | 1257 | fwrite (NODE_NAME (token->val.node), 1, NODE_LEN (token->val.node), fp); |
93c80368 | 1258 | break; |
041c3194 | 1259 | |
6338b358 | 1260 | case SPELL_LITERAL: |
47ad4138 ZW |
1261 | fwrite (token->val.str.text, 1, token->val.str.len, fp); |
1262 | break; | |
1263 | ||
93c80368 NB |
1264 | case SPELL_NONE: |
1265 | /* An error, most probably. */ | |
1266 | break; | |
041c3194 | 1267 | } |
c5a04734 ZW |
1268 | } |
1269 | ||
93c80368 NB |
1270 | /* Compare two tokens. */ |
1271 | int | |
6cf87ca4 | 1272 | _cpp_equiv_tokens (const cpp_token *a, const cpp_token *b) |
c5a04734 | 1273 | { |
93c80368 NB |
1274 | if (a->type == b->type && a->flags == b->flags) |
1275 | switch (TOKEN_SPELL (a)) | |
1276 | { | |
1277 | default: /* Keep compiler happy. */ | |
1278 | case SPELL_OPERATOR: | |
1279 | return 1; | |
93c80368 | 1280 | case SPELL_NONE: |
56051c0a | 1281 | return (a->type != CPP_MACRO_ARG || a->val.arg_no == b->val.arg_no); |
93c80368 NB |
1282 | case SPELL_IDENT: |
1283 | return a->val.node == b->val.node; | |
6338b358 | 1284 | case SPELL_LITERAL: |
93c80368 NB |
1285 | return (a->val.str.len == b->val.str.len |
1286 | && !memcmp (a->val.str.text, b->val.str.text, | |
1287 | a->val.str.len)); | |
1288 | } | |
c5a04734 | 1289 | |
041c3194 ZW |
1290 | return 0; |
1291 | } | |
1292 | ||
93c80368 NB |
1293 | /* Returns nonzero if a space should be inserted to avoid an |
1294 | accidental token paste for output. For simplicity, it is | |
1295 | conservative, and occasionally advises a space where one is not | |
1296 | needed, e.g. "." and ".2". */ | |
93c80368 | 1297 | int |
6cf87ca4 ZW |
1298 | cpp_avoid_paste (cpp_reader *pfile, const cpp_token *token1, |
1299 | const cpp_token *token2) | |
c5a04734 | 1300 | { |
93c80368 NB |
1301 | enum cpp_ttype a = token1->type, b = token2->type; |
1302 | cppchar_t c; | |
c5a04734 | 1303 | |
93c80368 NB |
1304 | if (token1->flags & NAMED_OP) |
1305 | a = CPP_NAME; | |
1306 | if (token2->flags & NAMED_OP) | |
1307 | b = CPP_NAME; | |
c5a04734 | 1308 | |
93c80368 NB |
1309 | c = EOF; |
1310 | if (token2->flags & DIGRAPH) | |
37b8524c | 1311 | c = digraph_spellings[(int) b - (int) CPP_FIRST_DIGRAPH][0]; |
93c80368 NB |
1312 | else if (token_spellings[b].category == SPELL_OPERATOR) |
1313 | c = token_spellings[b].name[0]; | |
c5a04734 | 1314 | |
93c80368 | 1315 | /* Quickly get everything that can paste with an '='. */ |
37b8524c | 1316 | if ((int) a <= (int) CPP_LAST_EQ && c == '=') |
93c80368 | 1317 | return 1; |
c5a04734 | 1318 | |
93c80368 | 1319 | switch (a) |
c5a04734 | 1320 | { |
93c80368 NB |
1321 | case CPP_GREATER: return c == '>' || c == '?'; |
1322 | case CPP_LESS: return c == '<' || c == '?' || c == '%' || c == ':'; | |
1323 | case CPP_PLUS: return c == '+'; | |
1324 | case CPP_MINUS: return c == '-' || c == '>'; | |
1325 | case CPP_DIV: return c == '/' || c == '*'; /* Comments. */ | |
1326 | case CPP_MOD: return c == ':' || c == '>'; | |
1327 | case CPP_AND: return c == '&'; | |
1328 | case CPP_OR: return c == '|'; | |
1329 | case CPP_COLON: return c == ':' || c == '>'; | |
1330 | case CPP_DEREF: return c == '*'; | |
26ec42ee | 1331 | case CPP_DOT: return c == '.' || c == '%' || b == CPP_NUMBER; |
93c80368 NB |
1332 | case CPP_HASH: return c == '#' || c == '%'; /* Digraph form. */ |
1333 | case CPP_NAME: return ((b == CPP_NUMBER | |
1334 | && name_p (pfile, &token2->val.str)) | |
1335 | || b == CPP_NAME | |
1336 | || b == CPP_CHAR || b == CPP_STRING); /* L */ | |
1337 | case CPP_NUMBER: return (b == CPP_NUMBER || b == CPP_NAME | |
1338 | || c == '.' || c == '+' || c == '-'); | |
1613e52b | 1339 | /* UCNs */ |
1067694a NB |
1340 | case CPP_OTHER: return ((token1->val.str.text[0] == '\\' |
1341 | && b == CPP_NAME) | |
1613e52b | 1342 | || (CPP_OPTION (pfile, objc) |
1067694a | 1343 | && token1->val.str.text[0] == '@' |
1613e52b | 1344 | && (b == CPP_NAME || b == CPP_STRING))); |
93c80368 | 1345 | default: break; |
c5a04734 | 1346 | } |
c5a04734 | 1347 | |
417f3e3a | 1348 | return 0; |
c5a04734 ZW |
1349 | } |
1350 | ||
93c80368 | 1351 | /* Output all the remaining tokens on the current line, and a newline |
4ed5bcfb NB |
1352 | character, to FP. Leading whitespace is removed. If there are |
1353 | macros, special token padding is not performed. */ | |
c5a04734 | 1354 | void |
6cf87ca4 | 1355 | cpp_output_line (cpp_reader *pfile, FILE *fp) |
c5a04734 | 1356 | { |
4ed5bcfb | 1357 | const cpp_token *token; |
96be6998 | 1358 | |
4ed5bcfb NB |
1359 | token = cpp_get_token (pfile); |
1360 | while (token->type != CPP_EOF) | |
96be6998 | 1361 | { |
4ed5bcfb NB |
1362 | cpp_output_token (token, fp); |
1363 | token = cpp_get_token (pfile); | |
1364 | if (token->flags & PREV_WHITE) | |
1365 | putc (' ', fp); | |
96be6998 ZW |
1366 | } |
1367 | ||
93c80368 | 1368 | putc ('\n', fp); |
041c3194 | 1369 | } |
c5a04734 | 1370 | |
1e013d2e NB |
1371 | /* Memory buffers. Changing these three constants can have a dramatic |
1372 | effect on performance. The values here are reasonable defaults, | |
1373 | but might be tuned. If you adjust them, be sure to test across a | |
1374 | range of uses of cpplib, including heavy nested function-like macro | |
1375 | expansion. Also check the change in peak memory usage (NJAMD is a | |
1376 | good tool for this). */ | |
1377 | #define MIN_BUFF_SIZE 8000 | |
87062813 | 1378 | #define BUFF_SIZE_UPPER_BOUND(MIN_SIZE) (MIN_BUFF_SIZE + (MIN_SIZE) * 3 / 2) |
1e013d2e NB |
1379 | #define EXTENDED_BUFF_SIZE(BUFF, MIN_EXTRA) \ |
1380 | (MIN_EXTRA + ((BUFF)->limit - (BUFF)->cur) * 2) | |
417f3e3a | 1381 | |
87062813 NB |
1382 | #if MIN_BUFF_SIZE > BUFF_SIZE_UPPER_BOUND (0) |
1383 | #error BUFF_SIZE_UPPER_BOUND must be at least as large as MIN_BUFF_SIZE! | |
1384 | #endif | |
1385 | ||
c9e7a609 NB |
1386 | /* Create a new allocation buffer. Place the control block at the end |
1387 | of the buffer, so that buffer overflows will cause immediate chaos. */ | |
b8af0ca5 | 1388 | static _cpp_buff * |
6cf87ca4 | 1389 | new_buff (size_t len) |
b8af0ca5 NB |
1390 | { |
1391 | _cpp_buff *result; | |
ece54d54 | 1392 | unsigned char *base; |
b8af0ca5 | 1393 | |
1e013d2e NB |
1394 | if (len < MIN_BUFF_SIZE) |
1395 | len = MIN_BUFF_SIZE; | |
c70f6ed3 | 1396 | len = CPP_ALIGN (len); |
b8af0ca5 NB |
1397 | |
1398 | base = xmalloc (len + sizeof (_cpp_buff)); | |
1399 | result = (_cpp_buff *) (base + len); | |
1400 | result->base = base; | |
1401 | result->cur = base; | |
1402 | result->limit = base + len; | |
1403 | result->next = NULL; | |
1404 | return result; | |
1405 | } | |
1406 | ||
1407 | /* Place a chain of unwanted allocation buffers on the free list. */ | |
1408 | void | |
6cf87ca4 | 1409 | _cpp_release_buff (cpp_reader *pfile, _cpp_buff *buff) |
b8af0ca5 NB |
1410 | { |
1411 | _cpp_buff *end = buff; | |
1412 | ||
1413 | while (end->next) | |
1414 | end = end->next; | |
1415 | end->next = pfile->free_buffs; | |
1416 | pfile->free_buffs = buff; | |
1417 | } | |
1418 | ||
1419 | /* Return a free buffer of size at least MIN_SIZE. */ | |
1420 | _cpp_buff * | |
6cf87ca4 | 1421 | _cpp_get_buff (cpp_reader *pfile, size_t min_size) |
b8af0ca5 NB |
1422 | { |
1423 | _cpp_buff *result, **p; | |
1424 | ||
1425 | for (p = &pfile->free_buffs;; p = &(*p)->next) | |
1426 | { | |
6142088c | 1427 | size_t size; |
1e013d2e NB |
1428 | |
1429 | if (*p == NULL) | |
b8af0ca5 | 1430 | return new_buff (min_size); |
1e013d2e NB |
1431 | result = *p; |
1432 | size = result->limit - result->base; | |
1433 | /* Return a buffer that's big enough, but don't waste one that's | |
1434 | way too big. */ | |
34f5271d | 1435 | if (size >= min_size && size <= BUFF_SIZE_UPPER_BOUND (min_size)) |
b8af0ca5 NB |
1436 | break; |
1437 | } | |
1438 | ||
1439 | *p = result->next; | |
1440 | result->next = NULL; | |
1441 | result->cur = result->base; | |
1442 | return result; | |
1443 | } | |
1444 | ||
4fe9b91c | 1445 | /* Creates a new buffer with enough space to hold the uncommitted |
8c3b2693 NB |
1446 | remaining bytes of BUFF, and at least MIN_EXTRA more bytes. Copies |
1447 | the excess bytes to the new buffer. Chains the new buffer after | |
1448 | BUFF, and returns the new buffer. */ | |
b8af0ca5 | 1449 | _cpp_buff * |
6cf87ca4 | 1450 | _cpp_append_extend_buff (cpp_reader *pfile, _cpp_buff *buff, size_t min_extra) |
b8af0ca5 | 1451 | { |
6142088c | 1452 | size_t size = EXTENDED_BUFF_SIZE (buff, min_extra); |
8c3b2693 | 1453 | _cpp_buff *new_buff = _cpp_get_buff (pfile, size); |
b8af0ca5 | 1454 | |
8c3b2693 NB |
1455 | buff->next = new_buff; |
1456 | memcpy (new_buff->base, buff->cur, BUFF_ROOM (buff)); | |
1457 | return new_buff; | |
1458 | } | |
1459 | ||
4fe9b91c | 1460 | /* Creates a new buffer with enough space to hold the uncommitted |
8c3b2693 NB |
1461 | remaining bytes of the buffer pointed to by BUFF, and at least |
1462 | MIN_EXTRA more bytes. Copies the excess bytes to the new buffer. | |
1463 | Chains the new buffer before the buffer pointed to by BUFF, and | |
1464 | updates the pointer to point to the new buffer. */ | |
1465 | void | |
6cf87ca4 | 1466 | _cpp_extend_buff (cpp_reader *pfile, _cpp_buff **pbuff, size_t min_extra) |
8c3b2693 NB |
1467 | { |
1468 | _cpp_buff *new_buff, *old_buff = *pbuff; | |
1469 | size_t size = EXTENDED_BUFF_SIZE (old_buff, min_extra); | |
1470 | ||
1471 | new_buff = _cpp_get_buff (pfile, size); | |
1472 | memcpy (new_buff->base, old_buff->cur, BUFF_ROOM (old_buff)); | |
1473 | new_buff->next = old_buff; | |
1474 | *pbuff = new_buff; | |
b8af0ca5 NB |
1475 | } |
1476 | ||
1477 | /* Free a chain of buffers starting at BUFF. */ | |
1478 | void | |
5671bf27 | 1479 | _cpp_free_buff (_cpp_buff *buff) |
b8af0ca5 NB |
1480 | { |
1481 | _cpp_buff *next; | |
1482 | ||
1483 | for (; buff; buff = next) | |
1484 | { | |
1485 | next = buff->next; | |
1486 | free (buff->base); | |
1487 | } | |
1488 | } | |
417f3e3a | 1489 | |
ece54d54 NB |
1490 | /* Allocate permanent, unaligned storage of length LEN. */ |
1491 | unsigned char * | |
6cf87ca4 | 1492 | _cpp_unaligned_alloc (cpp_reader *pfile, size_t len) |
ece54d54 NB |
1493 | { |
1494 | _cpp_buff *buff = pfile->u_buff; | |
1495 | unsigned char *result = buff->cur; | |
1496 | ||
1497 | if (len > (size_t) (buff->limit - result)) | |
1498 | { | |
1499 | buff = _cpp_get_buff (pfile, len); | |
1500 | buff->next = pfile->u_buff; | |
1501 | pfile->u_buff = buff; | |
1502 | result = buff->cur; | |
1503 | } | |
1504 | ||
1505 | buff->cur = result + len; | |
1506 | return result; | |
1507 | } | |
1508 | ||
87062813 NB |
1509 | /* Allocate permanent, unaligned storage of length LEN from a_buff. |
1510 | That buffer is used for growing allocations when saving macro | |
1511 | replacement lists in a #define, and when parsing an answer to an | |
1512 | assertion in #assert, #unassert or #if (and therefore possibly | |
1513 | whilst expanding macros). It therefore must not be used by any | |
1514 | code that they might call: specifically the lexer and the guts of | |
1515 | the macro expander. | |
1516 | ||
1517 | All existing other uses clearly fit this restriction: storing | |
1518 | registered pragmas during initialization. */ | |
93c80368 | 1519 | unsigned char * |
6cf87ca4 | 1520 | _cpp_aligned_alloc (cpp_reader *pfile, size_t len) |
3fef5b2b | 1521 | { |
8c3b2693 NB |
1522 | _cpp_buff *buff = pfile->a_buff; |
1523 | unsigned char *result = buff->cur; | |
3fef5b2b | 1524 | |
8c3b2693 | 1525 | if (len > (size_t) (buff->limit - result)) |
3fef5b2b | 1526 | { |
8c3b2693 NB |
1527 | buff = _cpp_get_buff (pfile, len); |
1528 | buff->next = pfile->a_buff; | |
1529 | pfile->a_buff = buff; | |
1530 | result = buff->cur; | |
3fef5b2b | 1531 | } |
041c3194 | 1532 | |
8c3b2693 | 1533 | buff->cur = result + len; |
93c80368 | 1534 | return result; |
041c3194 | 1535 | } |