/[pcre]/code/trunk/pcrecpp.cc
ViewVC logotype

Contents of /code/trunk/pcrecpp.cc

Parent Directory Parent Directory | Revision Log Revision Log


Revision 179 - (show annotations) (download)
Wed Jun 13 08:53:45 2007 UTC (7 years, 4 months ago) by ph10
File size: 29240 byte(s)
Apply C++ patch to fix a bad optimization.

1 // Copyright (c) 2005, Google Inc.
2 // All rights reserved.
3 //
4 // Redistribution and use in source and binary forms, with or without
5 // modification, are permitted provided that the following conditions are
6 // met:
7 //
8 // * Redistributions of source code must retain the above copyright
9 // notice, this list of conditions and the following disclaimer.
10 // * Redistributions in binary form must reproduce the above
11 // copyright notice, this list of conditions and the following disclaimer
12 // in the documentation and/or other materials provided with the
13 // distribution.
14 // * Neither the name of Google Inc. nor the names of its
15 // contributors may be used to endorse or promote products derived from
16 // this software without specific prior written permission.
17 //
18 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29 //
30 // Author: Sanjay Ghemawat
31
32 #ifdef HAVE_CONFIG_H
33 # include <config.h>
34 #endif
35
36 #include <stdlib.h>
37 #include <stdio.h>
38 #include <ctype.h>
39 #include <limits.h> /* for SHRT_MIN, USHRT_MAX, etc */
40 #include <assert.h>
41 #include <errno.h>
42 #include <string>
43 #include <algorithm>
44 // We need this to compile the proper dll on windows/msys. This is copied
45 // from pcre_internal.h. It would probably be better just to include that.
46 #define PCRE_DEFINITION /* Win32 __declspec(export) trigger for .dll */
47 #include <pcre.h>
48 #include "pcre_stringpiece.h"
49 #include "pcrecpp.h"
50
51
52 namespace pcrecpp {
53
54 // Maximum number of args we can set
55 static const int kMaxArgs = 16;
56 static const int kVecSize = (1 + kMaxArgs) * 3; // results + PCRE workspace
57
58 // Special object that stands-in for no argument
59 Arg no_arg((void*)NULL);
60
61 // If a regular expression has no error, its error_ field points here
62 static const string empty_string;
63
64 // If the user doesn't ask for any options, we just use this one
65 static RE_Options default_options;
66
67 void RE::Init(const string& pat, const RE_Options* options) {
68 pattern_ = pat;
69 if (options == NULL) {
70 options_ = default_options;
71 } else {
72 options_ = *options;
73 }
74 error_ = &empty_string;
75 re_full_ = NULL;
76 re_partial_ = NULL;
77
78 re_partial_ = Compile(UNANCHORED);
79 if (re_partial_ != NULL) {
80 re_full_ = Compile(ANCHOR_BOTH);
81 }
82 }
83
84 void RE::Cleanup() {
85 if (re_full_ != NULL) (*pcre_free)(re_full_);
86 if (re_partial_ != NULL) (*pcre_free)(re_partial_);
87 if (error_ != &empty_string) delete error_;
88 }
89
90
91 RE::~RE() {
92 Cleanup();
93 }
94
95
96 pcre* RE::Compile(Anchor anchor) {
97 // First, convert RE_Options into pcre options
98 int pcre_options = 0;
99 pcre_options = options_.all_options();
100
101 // Special treatment for anchoring. This is needed because at
102 // runtime pcre only provides an option for anchoring at the
103 // beginning of a string (unless you use offset).
104 //
105 // There are three types of anchoring we want:
106 // UNANCHORED Compile the original pattern, and use
107 // a pcre unanchored match.
108 // ANCHOR_START Compile the original pattern, and use
109 // a pcre anchored match.
110 // ANCHOR_BOTH Tack a "\z" to the end of the original pattern
111 // and use a pcre anchored match.
112
113 const char* compile_error;
114 int eoffset;
115 pcre* re;
116 if (anchor != ANCHOR_BOTH) {
117 re = pcre_compile(pattern_.c_str(), pcre_options,
118 &compile_error, &eoffset, NULL);
119 } else {
120 // Tack a '\z' at the end of RE. Parenthesize it first so that
121 // the '\z' applies to all top-level alternatives in the regexp.
122 string wrapped = "(?:"; // A non-counting grouping operator
123 wrapped += pattern_;
124 wrapped += ")\\z";
125 re = pcre_compile(wrapped.c_str(), pcre_options,
126 &compile_error, &eoffset, NULL);
127 }
128 if (re == NULL) {
129 if (error_ == &empty_string) error_ = new string(compile_error);
130 }
131 return re;
132 }
133
134 /***** Matching interfaces *****/
135
136 bool RE::FullMatch(const StringPiece& text,
137 const Arg& ptr1,
138 const Arg& ptr2,
139 const Arg& ptr3,
140 const Arg& ptr4,
141 const Arg& ptr5,
142 const Arg& ptr6,
143 const Arg& ptr7,
144 const Arg& ptr8,
145 const Arg& ptr9,
146 const Arg& ptr10,
147 const Arg& ptr11,
148 const Arg& ptr12,
149 const Arg& ptr13,
150 const Arg& ptr14,
151 const Arg& ptr15,
152 const Arg& ptr16) const {
153 const Arg* args[kMaxArgs];
154 int n = 0;
155 if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
156 if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
157 if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
158 if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
159 if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
160 if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
161 if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
162 if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
163 if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
164 if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
165 if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
166 if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
167 if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
168 if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
169 if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
170 if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
171 done:
172
173 int consumed;
174 int vec[kVecSize];
175 return DoMatchImpl(text, ANCHOR_BOTH, &consumed, args, n, vec, kVecSize);
176 }
177
178 bool RE::PartialMatch(const StringPiece& text,
179 const Arg& ptr1,
180 const Arg& ptr2,
181 const Arg& ptr3,
182 const Arg& ptr4,
183 const Arg& ptr5,
184 const Arg& ptr6,
185 const Arg& ptr7,
186 const Arg& ptr8,
187 const Arg& ptr9,
188 const Arg& ptr10,
189 const Arg& ptr11,
190 const Arg& ptr12,
191 const Arg& ptr13,
192 const Arg& ptr14,
193 const Arg& ptr15,
194 const Arg& ptr16) const {
195 const Arg* args[kMaxArgs];
196 int n = 0;
197 if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
198 if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
199 if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
200 if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
201 if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
202 if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
203 if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
204 if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
205 if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
206 if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
207 if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
208 if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
209 if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
210 if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
211 if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
212 if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
213 done:
214
215 int consumed;
216 int vec[kVecSize];
217 return DoMatchImpl(text, UNANCHORED, &consumed, args, n, vec, kVecSize);
218 }
219
220 bool RE::Consume(StringPiece* input,
221 const Arg& ptr1,
222 const Arg& ptr2,
223 const Arg& ptr3,
224 const Arg& ptr4,
225 const Arg& ptr5,
226 const Arg& ptr6,
227 const Arg& ptr7,
228 const Arg& ptr8,
229 const Arg& ptr9,
230 const Arg& ptr10,
231 const Arg& ptr11,
232 const Arg& ptr12,
233 const Arg& ptr13,
234 const Arg& ptr14,
235 const Arg& ptr15,
236 const Arg& ptr16) const {
237 const Arg* args[kMaxArgs];
238 int n = 0;
239 if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
240 if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
241 if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
242 if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
243 if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
244 if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
245 if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
246 if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
247 if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
248 if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
249 if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
250 if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
251 if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
252 if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
253 if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
254 if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
255 done:
256
257 int consumed;
258 int vec[kVecSize];
259 if (DoMatchImpl(*input, ANCHOR_START, &consumed,
260 args, n, vec, kVecSize)) {
261 input->remove_prefix(consumed);
262 return true;
263 } else {
264 return false;
265 }
266 }
267
268 bool RE::FindAndConsume(StringPiece* input,
269 const Arg& ptr1,
270 const Arg& ptr2,
271 const Arg& ptr3,
272 const Arg& ptr4,
273 const Arg& ptr5,
274 const Arg& ptr6,
275 const Arg& ptr7,
276 const Arg& ptr8,
277 const Arg& ptr9,
278 const Arg& ptr10,
279 const Arg& ptr11,
280 const Arg& ptr12,
281 const Arg& ptr13,
282 const Arg& ptr14,
283 const Arg& ptr15,
284 const Arg& ptr16) const {
285 const Arg* args[kMaxArgs];
286 int n = 0;
287 if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
288 if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
289 if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
290 if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
291 if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
292 if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
293 if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
294 if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
295 if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
296 if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
297 if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
298 if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
299 if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
300 if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
301 if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
302 if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
303 done:
304
305 int consumed;
306 int vec[kVecSize];
307 if (DoMatchImpl(*input, UNANCHORED, &consumed,
308 args, n, vec, kVecSize)) {
309 input->remove_prefix(consumed);
310 return true;
311 } else {
312 return false;
313 }
314 }
315
316 bool RE::Replace(const StringPiece& rewrite,
317 string *str) const {
318 int vec[kVecSize];
319 int matches = TryMatch(*str, 0, UNANCHORED, vec, kVecSize);
320 if (matches == 0)
321 return false;
322
323 string s;
324 if (!Rewrite(&s, rewrite, *str, vec, matches))
325 return false;
326
327 assert(vec[0] >= 0);
328 assert(vec[1] >= 0);
329 str->replace(vec[0], vec[1] - vec[0], s);
330 return true;
331 }
332
333 // Returns PCRE_NEWLINE_CRLF, PCRE_NEWLINE_CR, or PCRE_NEWLINE_LF.
334 // Note that PCRE_NEWLINE_CRLF is defined to be P_N_CR | P_N_LF.
335 static int NewlineMode(int pcre_options) {
336 // TODO: if we can make it threadsafe, cache this var
337 int newline_mode = 0;
338 /* if (newline_mode) return newline_mode; */ // do this once it's cached
339 if (pcre_options & (PCRE_NEWLINE_CRLF|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF)) {
340 newline_mode = (pcre_options &
341 (PCRE_NEWLINE_CRLF|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF));
342 } else {
343 int newline;
344 pcre_config(PCRE_CONFIG_NEWLINE, &newline);
345 if (newline == 10)
346 newline_mode = PCRE_NEWLINE_LF;
347 else if (newline == 13)
348 newline_mode = PCRE_NEWLINE_CR;
349 else if (newline == 3338)
350 newline_mode = PCRE_NEWLINE_CRLF;
351 else
352 assert("" == "Unexpected return value from pcre_config(NEWLINE)");
353 }
354 return newline_mode;
355 }
356
357 int RE::GlobalReplace(const StringPiece& rewrite,
358 string *str) const {
359 int count = 0;
360 int vec[kVecSize];
361 string out;
362 int start = 0;
363 int lastend = -1;
364
365 for (; start <= static_cast<int>(str->length()); count++) {
366 int matches = TryMatch(*str, start, UNANCHORED, vec, kVecSize);
367 if (matches <= 0)
368 break;
369 int matchstart = vec[0], matchend = vec[1];
370 assert(matchstart >= start);
371 assert(matchend >= matchstart);
372 if (matchstart == matchend && matchstart == lastend) {
373 // advance one character if we matched an empty string at the same
374 // place as the last match occurred
375 matchend = start + 1;
376 // If the current char is CR and we're in CRLF mode, skip LF too.
377 // Note it's better to call pcre_fullinfo() than to examine
378 // all_options(), since options_ could have changed bewteen
379 // compile-time and now, but this is simpler and safe enough.
380 if (start+1 < static_cast<int>(str->length()) &&
381 (*str)[start] == '\r' && (*str)[start+1] == '\n' &&
382 NewlineMode(options_.all_options()) == PCRE_NEWLINE_CRLF) {
383 matchend++;
384 }
385 // We also need to advance more than one char if we're in utf8 mode.
386 #ifdef SUPPORT_UTF8
387 if (options_.utf8()) {
388 while (matchend < static_cast<int>(str->length()) &&
389 ((*str)[matchend] & 0xc0) == 0x80)
390 matchend++;
391 }
392 #endif
393 if (matchend <= static_cast<int>(str->length()))
394 out.append(*str, start, matchend - start);
395 start = matchend;
396 } else {
397 out.append(*str, start, matchstart - start);
398 Rewrite(&out, rewrite, *str, vec, matches);
399 start = matchend;
400 lastend = matchend;
401 count++;
402 }
403 }
404
405 if (count == 0)
406 return 0;
407
408 if (start < static_cast<int>(str->length()))
409 out.append(*str, start, str->length() - start);
410 swap(out, *str);
411 return count;
412 }
413
414 bool RE::Extract(const StringPiece& rewrite,
415 const StringPiece& text,
416 string *out) const {
417 int vec[kVecSize];
418 int matches = TryMatch(text, 0, UNANCHORED, vec, kVecSize);
419 if (matches == 0)
420 return false;
421 out->erase();
422 return Rewrite(out, rewrite, text, vec, matches);
423 }
424
425 /*static*/ string RE::QuoteMeta(const StringPiece& unquoted) {
426 string result;
427
428 // Escape any ascii character not in [A-Za-z_0-9].
429 //
430 // Note that it's legal to escape a character even if it has no
431 // special meaning in a regular expression -- so this function does
432 // that. (This also makes it identical to the perl function of the
433 // same name; see `perldoc -f quotemeta`.)
434 for (int ii = 0; ii < unquoted.size(); ++ii) {
435 // Note that using 'isalnum' here raises the benchmark time from
436 // 32ns to 58ns:
437 if ((unquoted[ii] < 'a' || unquoted[ii] > 'z') &&
438 (unquoted[ii] < 'A' || unquoted[ii] > 'Z') &&
439 (unquoted[ii] < '0' || unquoted[ii] > '9') &&
440 unquoted[ii] != '_' &&
441 // If this is the part of a UTF8 or Latin1 character, we need
442 // to copy this byte without escaping. Experimentally this is
443 // what works correctly with the regexp library.
444 !(unquoted[ii] & 128)) {
445 result += '\\';
446 }
447 result += unquoted[ii];
448 }
449
450 return result;
451 }
452
453 /***** Actual matching and rewriting code *****/
454
455 int RE::TryMatch(const StringPiece& text,
456 int startpos,
457 Anchor anchor,
458 int *vec,
459 int vecsize) const {
460 pcre* re = (anchor == ANCHOR_BOTH) ? re_full_ : re_partial_;
461 if (re == NULL) {
462 //fprintf(stderr, "Matching against invalid re: %s\n", error_->c_str());
463 return 0;
464 }
465
466 pcre_extra extra = { 0 };
467 if (options_.match_limit() > 0) {
468 extra.flags |= PCRE_EXTRA_MATCH_LIMIT;
469 extra.match_limit = options_.match_limit();
470 }
471 if (options_.match_limit_recursion() > 0) {
472 extra.flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
473 extra.match_limit_recursion = options_.match_limit_recursion();
474 }
475 int rc = pcre_exec(re, // The regular expression object
476 &extra,
477 (text.data() == NULL) ? "" : text.data(),
478 text.size(),
479 startpos,
480 (anchor == UNANCHORED) ? 0 : PCRE_ANCHORED,
481 vec,
482 vecsize);
483
484 // Handle errors
485 if (rc == PCRE_ERROR_NOMATCH) {
486 return 0;
487 } else if (rc < 0) {
488 //fprintf(stderr, "Unexpected return code: %d when matching '%s'\n",
489 // re, pattern_.c_str());
490 return 0;
491 } else if (rc == 0) {
492 // pcre_exec() returns 0 as a special case when the number of
493 // capturing subpatterns exceeds the size of the vector.
494 // When this happens, there is a match and the output vector
495 // is filled, but we miss out on the positions of the extra subpatterns.
496 rc = vecsize / 2;
497 }
498
499 return rc;
500 }
501
502 bool RE::DoMatchImpl(const StringPiece& text,
503 Anchor anchor,
504 int* consumed,
505 const Arg* const* args,
506 int n,
507 int* vec,
508 int vecsize) const {
509 assert((1 + n) * 3 <= vecsize); // results + PCRE workspace
510 int matches = TryMatch(text, 0, anchor, vec, vecsize);
511 assert(matches >= 0); // TryMatch never returns negatives
512 if (matches == 0)
513 return false;
514
515 *consumed = vec[1];
516
517 if (n == 0 || args == NULL) {
518 // We are not interested in results
519 return true;
520 }
521
522 if (NumberOfCapturingGroups() < n) {
523 // RE has fewer capturing groups than number of arg pointers passed in
524 return false;
525 }
526
527 // If we got here, we must have matched the whole pattern.
528 // We do not need (can not do) any more checks on the value of 'matches' here
529 // -- see the comment for TryMatch.
530 for (int i = 0; i < n; i++) {
531 const int start = vec[2*(i+1)];
532 const int limit = vec[2*(i+1)+1];
533 if (!args[i]->Parse(text.data() + start, limit-start)) {
534 // TODO: Should we indicate what the error was?
535 return false;
536 }
537 }
538
539 return true;
540 }
541
542 bool RE::DoMatch(const StringPiece& text,
543 Anchor anchor,
544 int* consumed,
545 const Arg* const args[],
546 int n) const {
547 assert(n >= 0);
548 size_t const vecsize = (1 + n) * 3; // results + PCRE workspace
549 // (as for kVecSize)
550 int space[21]; // use stack allocation for small vecsize (common case)
551 int* vec = vecsize <= 21 ? space : new int[vecsize];
552 bool retval = DoMatchImpl(text, anchor, consumed, args, n, vec, vecsize);
553 if (vec != space) delete [] vec;
554 return retval;
555 }
556
557 bool RE::Rewrite(string *out, const StringPiece &rewrite,
558 const StringPiece &text, int *vec, int veclen) const {
559 for (const char *s = rewrite.data(), *end = s + rewrite.size();
560 s < end; s++) {
561 int c = *s;
562 if (c == '\\') {
563 c = *++s;
564 if (isdigit(c)) {
565 int n = (c - '0');
566 if (n >= veclen) {
567 //fprintf(stderr, requested group %d in regexp %.*s\n",
568 // n, rewrite.size(), rewrite.data());
569 return false;
570 }
571 int start = vec[2 * n];
572 if (start >= 0)
573 out->append(text.data() + start, vec[2 * n + 1] - start);
574 } else if (c == '\\') {
575 out->push_back('\\');
576 } else {
577 //fprintf(stderr, "invalid rewrite pattern: %.*s\n",
578 // rewrite.size(), rewrite.data());
579 return false;
580 }
581 } else {
582 out->push_back(c);
583 }
584 }
585 return true;
586 }
587
588 // Return the number of capturing subpatterns, or -1 if the
589 // regexp wasn't valid on construction.
590 int RE::NumberOfCapturingGroups() const {
591 if (re_partial_ == NULL) return -1;
592
593 int result;
594 int pcre_retval = pcre_fullinfo(re_partial_, // The regular expression object
595 NULL, // We did not study the pattern
596 PCRE_INFO_CAPTURECOUNT,
597 &result);
598 assert(pcre_retval == 0);
599 return result;
600 }
601
602 /***** Parsers for various types *****/
603
604 bool Arg::parse_null(const char* str, int n, void* dest) {
605 // We fail if somebody asked us to store into a non-NULL void* pointer
606 return (dest == NULL);
607 }
608
609 bool Arg::parse_string(const char* str, int n, void* dest) {
610 reinterpret_cast<string*>(dest)->assign(str, n);
611 return true;
612 }
613
614 bool Arg::parse_stringpiece(const char* str, int n, void* dest) {
615 reinterpret_cast<StringPiece*>(dest)->set(str, n);
616 return true;
617 }
618
619 bool Arg::parse_char(const char* str, int n, void* dest) {
620 if (n != 1) return false;
621 *(reinterpret_cast<char*>(dest)) = str[0];
622 return true;
623 }
624
625 bool Arg::parse_uchar(const char* str, int n, void* dest) {
626 if (n != 1) return false;
627 *(reinterpret_cast<unsigned char*>(dest)) = str[0];
628 return true;
629 }
630
631 // Largest number spec that we are willing to parse
632 static const int kMaxNumberLength = 32;
633
634 // REQUIRES "buf" must have length at least kMaxNumberLength+1
635 // REQUIRES "n > 0"
636 // Copies "str" into "buf" and null-terminates if necessary.
637 // Returns one of:
638 // a. "str" if no termination is needed
639 // b. "buf" if the string was copied and null-terminated
640 // c. "" if the input was invalid and has no hope of being parsed
641 static const char* TerminateNumber(char* buf, const char* str, int n) {
642 if ((n > 0) && isspace(*str)) {
643 // We are less forgiving than the strtoxxx() routines and do not
644 // allow leading spaces.
645 return "";
646 }
647
648 // See if the character right after the input text may potentially
649 // look like a digit.
650 if (isdigit(str[n]) ||
651 ((str[n] >= 'a') && (str[n] <= 'f')) ||
652 ((str[n] >= 'A') && (str[n] <= 'F'))) {
653 if (n > kMaxNumberLength) return ""; // Input too big to be a valid number
654 memcpy(buf, str, n);
655 buf[n] = '\0';
656 return buf;
657 } else {
658 // We can parse right out of the supplied string, so return it.
659 return str;
660 }
661 }
662
663 bool Arg::parse_long_radix(const char* str,
664 int n,
665 void* dest,
666 int radix) {
667 if (n == 0) return false;
668 char buf[kMaxNumberLength+1];
669 str = TerminateNumber(buf, str, n);
670 char* end;
671 errno = 0;
672 long r = strtol(str, &end, radix);
673 if (end != str + n) return false; // Leftover junk
674 if (errno) return false;
675 *(reinterpret_cast<long*>(dest)) = r;
676 return true;
677 }
678
679 bool Arg::parse_ulong_radix(const char* str,
680 int n,
681 void* dest,
682 int radix) {
683 if (n == 0) return false;
684 char buf[kMaxNumberLength+1];
685 str = TerminateNumber(buf, str, n);
686 if (str[0] == '-') return false; // strtoul() on a negative number?!
687 char* end;
688 errno = 0;
689 unsigned long r = strtoul(str, &end, radix);
690 if (end != str + n) return false; // Leftover junk
691 if (errno) return false;
692 *(reinterpret_cast<unsigned long*>(dest)) = r;
693 return true;
694 }
695
696 bool Arg::parse_short_radix(const char* str,
697 int n,
698 void* dest,
699 int radix) {
700 long r;
701 if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
702 if (r < SHRT_MIN || r > SHRT_MAX) return false; // Out of range
703 *(reinterpret_cast<short*>(dest)) = r;
704 return true;
705 }
706
707 bool Arg::parse_ushort_radix(const char* str,
708 int n,
709 void* dest,
710 int radix) {
711 unsigned long r;
712 if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
713 if (r > USHRT_MAX) return false; // Out of range
714 *(reinterpret_cast<unsigned short*>(dest)) = r;
715 return true;
716 }
717
718 bool Arg::parse_int_radix(const char* str,
719 int n,
720 void* dest,
721 int radix) {
722 long r;
723 if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
724 if (r < INT_MIN || r > INT_MAX) return false; // Out of range
725 *(reinterpret_cast<int*>(dest)) = r;
726 return true;
727 }
728
729 bool Arg::parse_uint_radix(const char* str,
730 int n,
731 void* dest,
732 int radix) {
733 unsigned long r;
734 if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
735 if (r > UINT_MAX) return false; // Out of range
736 *(reinterpret_cast<unsigned int*>(dest)) = r;
737 return true;
738 }
739
740 bool Arg::parse_longlong_radix(const char* str,
741 int n,
742 void* dest,
743 int radix) {
744 #ifndef HAVE_LONG_LONG
745 return false;
746 #else
747 if (n == 0) return false;
748 char buf[kMaxNumberLength+1];
749 str = TerminateNumber(buf, str, n);
750 char* end;
751 errno = 0;
752 #if defined HAVE_STRTOQ
753 long long r = strtoq(str, &end, radix);
754 #elif defined HAVE_STRTOLL
755 long long r = strtoll(str, &end, radix);
756 #else
757 #error parse_longlong_radix: cannot convert input to a long-long
758 #endif
759 if (end != str + n) return false; // Leftover junk
760 if (errno) return false;
761 *(reinterpret_cast<long long*>(dest)) = r;
762 return true;
763 #endif /* HAVE_LONG_LONG */
764 }
765
766 bool Arg::parse_ulonglong_radix(const char* str,
767 int n,
768 void* dest,
769 int radix) {
770 #ifndef HAVE_UNSIGNED_LONG_LONG
771 return false;
772 #else
773 if (n == 0) return false;
774 char buf[kMaxNumberLength+1];
775 str = TerminateNumber(buf, str, n);
776 if (str[0] == '-') return false; // strtoull() on a negative number?!
777 char* end;
778 errno = 0;
779 #if defined HAVE_STRTOQ
780 unsigned long long r = strtouq(str, &end, radix);
781 #elif defined HAVE_STRTOLL
782 unsigned long long r = strtoull(str, &end, radix);
783 #else
784 #error parse_ulonglong_radix: cannot convert input to a long-long
785 #endif
786 if (end != str + n) return false; // Leftover junk
787 if (errno) return false;
788 *(reinterpret_cast<unsigned long long*>(dest)) = r;
789 return true;
790 #endif /* HAVE_UNSIGNED_LONG_LONG */
791 }
792
793 bool Arg::parse_double(const char* str, int n, void* dest) {
794 if (n == 0) return false;
795 static const int kMaxLength = 200;
796 char buf[kMaxLength];
797 if (n >= kMaxLength) return false;
798 memcpy(buf, str, n);
799 buf[n] = '\0';
800 errno = 0;
801 char* end;
802 double r = strtod(buf, &end);
803 if (end != buf + n) return false; // Leftover junk
804 if (errno) return false;
805 *(reinterpret_cast<double*>(dest)) = r;
806 return true;
807 }
808
809 bool Arg::parse_float(const char* str, int n, void* dest) {
810 double r;
811 if (!parse_double(str, n, &r)) return false;
812 *(reinterpret_cast<float*>(dest)) = static_cast<float>(r);
813 return true;
814 }
815
816
817 #define DEFINE_INTEGER_PARSERS(name) \
818 bool Arg::parse_##name(const char* str, int n, void* dest) { \
819 return parse_##name##_radix(str, n, dest, 10); \
820 } \
821 bool Arg::parse_##name##_hex(const char* str, int n, void* dest) { \
822 return parse_##name##_radix(str, n, dest, 16); \
823 } \
824 bool Arg::parse_##name##_octal(const char* str, int n, void* dest) { \
825 return parse_##name##_radix(str, n, dest, 8); \
826 } \
827 bool Arg::parse_##name##_cradix(const char* str, int n, void* dest) { \
828 return parse_##name##_radix(str, n, dest, 0); \
829 }
830
831 DEFINE_INTEGER_PARSERS(short) /* */
832 DEFINE_INTEGER_PARSERS(ushort) /* */
833 DEFINE_INTEGER_PARSERS(int) /* Don't use semicolons after these */
834 DEFINE_INTEGER_PARSERS(uint) /* statements because they can cause */
835 DEFINE_INTEGER_PARSERS(long) /* compiler warnings if the checking */
836 DEFINE_INTEGER_PARSERS(ulong) /* level is turned up high enough. */
837 DEFINE_INTEGER_PARSERS(longlong) /* */
838 DEFINE_INTEGER_PARSERS(ulonglong) /* */
839
840 #undef DEFINE_INTEGER_PARSERS
841
842 } // namespace pcrecpp

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12