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

Contents of /code/trunk/pcrecpp.cc

Parent Directory Parent Directory | Revision Log Revision Log


Revision 87 - (hide annotations) (download)
Sat Feb 24 21:41:21 2007 UTC (7 years, 4 months ago) by nigel
File size: 26943 byte(s)
Load pcre-6.5 into code/trunk.

1 nigel 77 // 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     #include <stdlib.h>
33     #include <stdio.h>
34     #include <ctype.h>
35     #include <limits.h> /* for SHRT_MIN, USHRT_MAX, etc */
36     #include <assert.h>
37     #include <errno.h>
38     #include <string>
39 nigel 81 #include <algorithm>
40 nigel 77 #include "config.h"
41     // We need this to compile the proper dll on windows/msys. This is copied
42     // from pcre_internal.h. It would probably be better just to include that.
43     #define PCRE_DEFINITION /* Win32 __declspec(export) trigger for .dll */
44     #include "pcre.h"
45     #include "pcre_stringpiece.h"
46     #include "pcrecpp.h"
47    
48    
49     namespace pcrecpp {
50    
51     // Maximum number of args we can set
52     static const int kMaxArgs = 16;
53     static const int kVecSize = (1 + kMaxArgs) * 3; // results + PCRE workspace
54    
55     // Special object that stands-in for no argument
56     Arg no_arg((void*)NULL);
57    
58     // If a regular expression has no error, its error_ field points here
59     static const string empty_string;
60    
61     // If the user doesn't ask for any options, we just use this one
62     static RE_Options default_options;
63    
64     void RE::Init(const char* pat, const RE_Options* options) {
65     pattern_ = pat;
66     if (options == NULL) {
67     options_ = default_options;
68     } else {
69     options_ = *options;
70     }
71     error_ = &empty_string;
72     re_full_ = NULL;
73     re_partial_ = NULL;
74    
75     re_partial_ = Compile(UNANCHORED);
76     if (re_partial_ != NULL) {
77     // Check for complicated patterns. The following change is
78     // conservative in that it may treat some "simple" patterns
79     // as "complex" (e.g., if the vertical bar is in a character
80     // class or is escaped). But it seems good enough.
81     if (strchr(pat, '|') == NULL) {
82     // Simple pattern: we can use position-based checks to perform
83     // fully anchored matches
84     re_full_ = re_partial_;
85     } else {
86     // We need a special pattern for anchored matches
87     re_full_ = Compile(ANCHOR_BOTH);
88     }
89     }
90     }
91    
92     RE::~RE() {
93     if (re_full_ != NULL && re_full_ != re_partial_) (*pcre_free)(re_full_);
94     if (re_partial_ != NULL) (*pcre_free)(re_partial_);
95     if (error_ != &empty_string) delete error_;
96     }
97    
98     pcre* RE::Compile(Anchor anchor) {
99     // First, convert RE_Options into pcre options
100     int pcre_options = 0;
101 nigel 81 pcre_options = options_.all_options();
102 nigel 77
103     // Special treatment for anchoring. This is needed because at
104     // runtime pcre only provides an option for anchoring at the
105     // beginning of a string (unless you use offset).
106     //
107     // There are three types of anchoring we want:
108     // UNANCHORED Compile the original pattern, and use
109     // a pcre unanchored match.
110     // ANCHOR_START Compile the original pattern, and use
111     // a pcre anchored match.
112     // ANCHOR_BOTH Tack a "\z" to the end of the original pattern
113     // and use a pcre anchored match.
114    
115     const char* compile_error;
116     int eoffset;
117     pcre* re;
118     if (anchor != ANCHOR_BOTH) {
119     re = pcre_compile(pattern_.c_str(), pcre_options,
120     &compile_error, &eoffset, NULL);
121     } else {
122     // Tack a '\z' at the end of RE. Parenthesize it first so that
123     // the '\z' applies to all top-level alternatives in the regexp.
124     string wrapped = "(?:"; // A non-counting grouping operator
125     wrapped += pattern_;
126     wrapped += ")\\z";
127     re = pcre_compile(wrapped.c_str(), pcre_options,
128     &compile_error, &eoffset, NULL);
129     }
130     if (re == NULL) {
131     if (error_ == &empty_string) error_ = new string(compile_error);
132     }
133     return re;
134     }
135    
136     /***** Matching interfaces *****/
137    
138     bool RE::FullMatch(const StringPiece& text,
139     const Arg& ptr1,
140     const Arg& ptr2,
141     const Arg& ptr3,
142     const Arg& ptr4,
143     const Arg& ptr5,
144     const Arg& ptr6,
145     const Arg& ptr7,
146     const Arg& ptr8,
147     const Arg& ptr9,
148     const Arg& ptr10,
149     const Arg& ptr11,
150     const Arg& ptr12,
151     const Arg& ptr13,
152     const Arg& ptr14,
153     const Arg& ptr15,
154     const Arg& ptr16) const {
155     const Arg* args[kMaxArgs];
156     int n = 0;
157     if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
158     if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
159     if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
160     if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
161     if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
162     if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
163     if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
164     if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
165     if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
166     if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
167     if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
168     if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
169     if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
170     if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
171     if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
172     if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
173     done:
174    
175     int consumed;
176     int vec[kVecSize];
177     return DoMatchImpl(text, ANCHOR_BOTH, &consumed, args, n, vec, kVecSize);
178     }
179    
180     bool RE::PartialMatch(const StringPiece& text,
181     const Arg& ptr1,
182     const Arg& ptr2,
183     const Arg& ptr3,
184     const Arg& ptr4,
185     const Arg& ptr5,
186     const Arg& ptr6,
187     const Arg& ptr7,
188     const Arg& ptr8,
189     const Arg& ptr9,
190     const Arg& ptr10,
191     const Arg& ptr11,
192     const Arg& ptr12,
193     const Arg& ptr13,
194     const Arg& ptr14,
195     const Arg& ptr15,
196     const Arg& ptr16) const {
197     const Arg* args[kMaxArgs];
198     int n = 0;
199     if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
200     if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
201     if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
202     if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
203     if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
204     if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
205     if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
206     if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
207     if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
208     if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
209     if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
210     if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
211     if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
212     if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
213     if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
214     if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
215     done:
216    
217     int consumed;
218     int vec[kVecSize];
219     return DoMatchImpl(text, UNANCHORED, &consumed, args, n, vec, kVecSize);
220     }
221    
222     bool RE::Consume(StringPiece* input,
223     const Arg& ptr1,
224     const Arg& ptr2,
225     const Arg& ptr3,
226     const Arg& ptr4,
227     const Arg& ptr5,
228     const Arg& ptr6,
229     const Arg& ptr7,
230     const Arg& ptr8,
231     const Arg& ptr9,
232     const Arg& ptr10,
233     const Arg& ptr11,
234     const Arg& ptr12,
235     const Arg& ptr13,
236     const Arg& ptr14,
237     const Arg& ptr15,
238     const Arg& ptr16) const {
239     const Arg* args[kMaxArgs];
240     int n = 0;
241     if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
242     if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
243     if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
244     if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
245     if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
246     if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
247     if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
248     if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
249     if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
250     if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
251     if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
252     if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
253     if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
254     if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
255     if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
256     if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
257     done:
258    
259     int consumed;
260     int vec[kVecSize];
261     if (DoMatchImpl(*input, ANCHOR_START, &consumed,
262     args, n, vec, kVecSize)) {
263     input->remove_prefix(consumed);
264     return true;
265     } else {
266     return false;
267     }
268     }
269    
270     bool RE::FindAndConsume(StringPiece* input,
271     const Arg& ptr1,
272     const Arg& ptr2,
273     const Arg& ptr3,
274     const Arg& ptr4,
275     const Arg& ptr5,
276     const Arg& ptr6,
277     const Arg& ptr7,
278     const Arg& ptr8,
279     const Arg& ptr9,
280     const Arg& ptr10,
281     const Arg& ptr11,
282     const Arg& ptr12,
283     const Arg& ptr13,
284     const Arg& ptr14,
285     const Arg& ptr15,
286     const Arg& ptr16) const {
287     const Arg* args[kMaxArgs];
288     int n = 0;
289     if (&ptr1 == &no_arg) goto done; args[n++] = &ptr1;
290     if (&ptr2 == &no_arg) goto done; args[n++] = &ptr2;
291     if (&ptr3 == &no_arg) goto done; args[n++] = &ptr3;
292     if (&ptr4 == &no_arg) goto done; args[n++] = &ptr4;
293     if (&ptr5 == &no_arg) goto done; args[n++] = &ptr5;
294     if (&ptr6 == &no_arg) goto done; args[n++] = &ptr6;
295     if (&ptr7 == &no_arg) goto done; args[n++] = &ptr7;
296     if (&ptr8 == &no_arg) goto done; args[n++] = &ptr8;
297     if (&ptr9 == &no_arg) goto done; args[n++] = &ptr9;
298     if (&ptr10 == &no_arg) goto done; args[n++] = &ptr10;
299     if (&ptr11 == &no_arg) goto done; args[n++] = &ptr11;
300     if (&ptr12 == &no_arg) goto done; args[n++] = &ptr12;
301     if (&ptr13 == &no_arg) goto done; args[n++] = &ptr13;
302     if (&ptr14 == &no_arg) goto done; args[n++] = &ptr14;
303     if (&ptr15 == &no_arg) goto done; args[n++] = &ptr15;
304     if (&ptr16 == &no_arg) goto done; args[n++] = &ptr16;
305     done:
306    
307     int consumed;
308     int vec[kVecSize];
309     if (DoMatchImpl(*input, UNANCHORED, &consumed,
310     args, n, vec, kVecSize)) {
311     input->remove_prefix(consumed);
312     return true;
313     } else {
314     return false;
315     }
316     }
317    
318     bool RE::Replace(const StringPiece& rewrite,
319     string *str) const {
320     int vec[kVecSize];
321     int matches = TryMatch(*str, 0, UNANCHORED, vec, kVecSize);
322     if (matches == 0)
323     return false;
324    
325     string s;
326     if (!Rewrite(&s, rewrite, *str, vec, matches))
327     return false;
328    
329     assert(vec[0] >= 0);
330     assert(vec[1] >= 0);
331     str->replace(vec[0], vec[1] - vec[0], s);
332     return true;
333     }
334    
335     int RE::GlobalReplace(const StringPiece& rewrite,
336     string *str) const {
337     int count = 0;
338     int vec[kVecSize];
339     string out;
340     int start = 0;
341     int lastend = -1;
342    
343     for (; start <= static_cast<int>(str->length()); count++) {
344     int matches = TryMatch(*str, start, UNANCHORED, vec, kVecSize);
345     if (matches <= 0)
346     break;
347     int matchstart = vec[0], matchend = vec[1];
348     assert(matchstart >= start);
349     assert(matchend >= matchstart);
350     if (matchstart == matchend && matchstart == lastend) {
351     // advance one character if we matched an empty string at the same
352     // place as the last match occurred
353     if (start < static_cast<int>(str->length()))
354     out.push_back((*str)[start]);
355     start++;
356     } else {
357     out.append(*str, start, matchstart - start);
358     Rewrite(&out, rewrite, *str, vec, matches);
359     start = matchend;
360     lastend = matchend;
361     count++;
362     }
363     }
364    
365     if (count == 0)
366     return 0;
367    
368     if (start < static_cast<int>(str->length()))
369     out.append(*str, start, str->length() - start);
370     swap(out, *str);
371     return count;
372     }
373    
374     bool RE::Extract(const StringPiece& rewrite,
375     const StringPiece& text,
376     string *out) const {
377     int vec[kVecSize];
378     int matches = TryMatch(text, 0, UNANCHORED, vec, kVecSize);
379     if (matches == 0)
380     return false;
381 nigel 81 out->erase();
382 nigel 77 return Rewrite(out, rewrite, text, vec, matches);
383     }
384    
385     /***** Actual matching and rewriting code *****/
386    
387     int RE::TryMatch(const StringPiece& text,
388     int startpos,
389     Anchor anchor,
390     int *vec,
391     int vecsize) const {
392     pcre* re = (anchor == ANCHOR_BOTH) ? re_full_ : re_partial_;
393     if (re == NULL) {
394     //fprintf(stderr, "Matching against invalid re: %s\n", error_->c_str());
395     return 0;
396     }
397    
398     pcre_extra extra = { 0 };
399     if (options_.match_limit() > 0) {
400 nigel 87 extra.flags |= PCRE_EXTRA_MATCH_LIMIT;
401 nigel 77 extra.match_limit = options_.match_limit();
402     }
403 nigel 87 if (options_.match_limit_recursion() > 0) {
404     extra.flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
405     extra.match_limit_recursion = options_.match_limit_recursion();
406     }
407 nigel 77 int rc = pcre_exec(re, // The regular expression object
408     &extra,
409 nigel 87 (text.data() == NULL) ? "" : text.data(),
410 nigel 77 text.size(),
411     startpos,
412     (anchor == UNANCHORED) ? 0 : PCRE_ANCHORED,
413     vec,
414     vecsize);
415    
416     // Handle errors
417     if (rc == PCRE_ERROR_NOMATCH) {
418     return 0;
419     } else if (rc < 0) {
420     //fprintf(stderr, "Unexpected return code: %d when matching '%s'\n",
421     // re, pattern_.c_str());
422     return 0;
423     } else if (rc == 0) {
424     // pcre_exec() returns 0 as a special case when the number of
425     // capturing subpatterns exceeds the size of the vector.
426     // When this happens, there is a match and the output vector
427     // is filled, but we miss out on the positions of the extra subpatterns.
428     rc = vecsize / 2;
429     }
430    
431     if ((anchor == ANCHOR_BOTH) && (re_full_ == re_partial_)) {
432     // We need an extra check to make sure that the match extended
433     // to the end of the input string
434     assert(vec[0] == 0); // PCRE_ANCHORED forces starting match
435     if (vec[1] != text.size()) return 0; // Did not get ending match
436     }
437    
438     return rc;
439     }
440    
441     bool RE::DoMatchImpl(const StringPiece& text,
442     Anchor anchor,
443     int* consumed,
444     const Arg* const* args,
445     int n,
446     int* vec,
447     int vecsize) const {
448     assert((1 + n) * 3 <= vecsize); // results + PCRE workspace
449     int matches = TryMatch(text, 0, anchor, vec, vecsize);
450     assert(matches >= 0); // TryMatch never returns negatives
451     if (matches == 0)
452     return false;
453    
454     *consumed = vec[1];
455    
456 nigel 87 if (n == 0 || args == NULL) {
457 nigel 77 // We are not interested in results
458     return true;
459     }
460    
461 nigel 87 if (NumberOfCapturingGroups() < n) {
462     // RE has fewer capturing groups than number of arg pointers passed in
463     return false;
464     }
465    
466 nigel 77 // If we got here, we must have matched the whole pattern.
467     // We do not need (can not do) any more checks on the value of 'matches' here
468     // -- see the comment for TryMatch.
469     for (int i = 0; i < n; i++) {
470     const int start = vec[2*(i+1)];
471     const int limit = vec[2*(i+1)+1];
472     if (!args[i]->Parse(text.data() + start, limit-start)) {
473     // TODO: Should we indicate what the error was?
474     return false;
475     }
476     }
477    
478     return true;
479     }
480    
481     bool RE::DoMatch(const StringPiece& text,
482     Anchor anchor,
483     int* consumed,
484     const Arg* const args[],
485     int n) const {
486     assert(n >= 0);
487     size_t const vecsize = (1 + n) * 3; // results + PCRE workspace
488     // (as for kVecSize)
489     int space[21]; // use stack allocation for small vecsize (common case)
490     int* vec = vecsize <= 21 ? space : new int[vecsize];
491     bool retval = DoMatchImpl(text, anchor, consumed, args, n, vec, vecsize);
492     if (vec != space) delete [] vec;
493     return retval;
494     }
495    
496     bool RE::Rewrite(string *out, const StringPiece &rewrite,
497     const StringPiece &text, int *vec, int veclen) const {
498     for (const char *s = rewrite.data(), *end = s + rewrite.size();
499     s < end; s++) {
500     int c = *s;
501     if (c == '\\') {
502     c = *++s;
503     if (isdigit(c)) {
504     int n = (c - '0');
505     if (n >= veclen) {
506     //fprintf(stderr, requested group %d in regexp %.*s\n",
507     // n, rewrite.size(), rewrite.data());
508     return false;
509     }
510     int start = vec[2 * n];
511     if (start >= 0)
512     out->append(text.data() + start, vec[2 * n + 1] - start);
513     } else if (c == '\\') {
514     out->push_back('\\');
515     } else {
516     //fprintf(stderr, "invalid rewrite pattern: %.*s\n",
517     // rewrite.size(), rewrite.data());
518     return false;
519     }
520     } else {
521     out->push_back(c);
522     }
523     }
524     return true;
525     }
526    
527     // Return the number of capturing subpatterns, or -1 if the
528     // regexp wasn't valid on construction.
529 nigel 87 int RE::NumberOfCapturingGroups() const {
530 nigel 77 if (re_partial_ == NULL) return -1;
531    
532     int result;
533     int pcre_retval = pcre_fullinfo(re_partial_, // The regular expression object
534     NULL, // We did not study the pattern
535     PCRE_INFO_CAPTURECOUNT,
536     &result);
537     assert(pcre_retval == 0);
538     return result;
539     }
540    
541     /***** Parsers for various types *****/
542    
543     bool Arg::parse_null(const char* str, int n, void* dest) {
544     // We fail if somebody asked us to store into a non-NULL void* pointer
545     return (dest == NULL);
546     }
547    
548     bool Arg::parse_string(const char* str, int n, void* dest) {
549     reinterpret_cast<string*>(dest)->assign(str, n);
550     return true;
551     }
552    
553     bool Arg::parse_stringpiece(const char* str, int n, void* dest) {
554     reinterpret_cast<StringPiece*>(dest)->set(str, n);
555     return true;
556     }
557    
558     bool Arg::parse_char(const char* str, int n, void* dest) {
559     if (n != 1) return false;
560     *(reinterpret_cast<char*>(dest)) = str[0];
561     return true;
562     }
563    
564     bool Arg::parse_uchar(const char* str, int n, void* dest) {
565     if (n != 1) return false;
566     *(reinterpret_cast<unsigned char*>(dest)) = str[0];
567     return true;
568     }
569    
570     // Largest number spec that we are willing to parse
571     static const int kMaxNumberLength = 32;
572    
573     // REQUIRES "buf" must have length at least kMaxNumberLength+1
574     // REQUIRES "n > 0"
575     // Copies "str" into "buf" and null-terminates if necessary.
576     // Returns one of:
577     // a. "str" if no termination is needed
578     // b. "buf" if the string was copied and null-terminated
579     // c. "" if the input was invalid and has no hope of being parsed
580     static const char* TerminateNumber(char* buf, const char* str, int n) {
581     if ((n > 0) && isspace(*str)) {
582     // We are less forgiving than the strtoxxx() routines and do not
583     // allow leading spaces.
584     return "";
585     }
586    
587     // See if the character right after the input text may potentially
588     // look like a digit.
589     if (isdigit(str[n]) ||
590     ((str[n] >= 'a') && (str[n] <= 'f')) ||
591     ((str[n] >= 'A') && (str[n] <= 'F'))) {
592     if (n > kMaxNumberLength) return ""; // Input too big to be a valid number
593     memcpy(buf, str, n);
594     buf[n] = '\0';
595     return buf;
596     } else {
597     // We can parse right out of the supplied string, so return it.
598     return str;
599     }
600     }
601    
602     bool Arg::parse_long_radix(const char* str,
603     int n,
604     void* dest,
605     int radix) {
606     if (n == 0) return false;
607     char buf[kMaxNumberLength+1];
608     str = TerminateNumber(buf, str, n);
609     char* end;
610     errno = 0;
611     long r = strtol(str, &end, radix);
612     if (end != str + n) return false; // Leftover junk
613     if (errno) return false;
614     *(reinterpret_cast<long*>(dest)) = r;
615     return true;
616     }
617    
618     bool Arg::parse_ulong_radix(const char* str,
619     int n,
620     void* dest,
621     int radix) {
622     if (n == 0) return false;
623     char buf[kMaxNumberLength+1];
624     str = TerminateNumber(buf, str, n);
625 nigel 87 if (str[0] == '-') return false; // strtoul() on a negative number?!
626 nigel 77 char* end;
627     errno = 0;
628     unsigned long r = strtoul(str, &end, radix);
629     if (end != str + n) return false; // Leftover junk
630     if (errno) return false;
631     *(reinterpret_cast<unsigned long*>(dest)) = r;
632     return true;
633     }
634    
635     bool Arg::parse_short_radix(const char* str,
636     int n,
637     void* dest,
638     int radix) {
639     long r;
640     if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
641     if (r < SHRT_MIN || r > SHRT_MAX) return false; // Out of range
642     *(reinterpret_cast<short*>(dest)) = r;
643     return true;
644     }
645    
646     bool Arg::parse_ushort_radix(const char* str,
647     int n,
648     void* dest,
649     int radix) {
650     unsigned long r;
651     if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
652     if (r > USHRT_MAX) return false; // Out of range
653     *(reinterpret_cast<unsigned short*>(dest)) = r;
654     return true;
655     }
656    
657     bool Arg::parse_int_radix(const char* str,
658     int n,
659     void* dest,
660     int radix) {
661     long r;
662     if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
663     if (r < INT_MIN || r > INT_MAX) return false; // Out of range
664     *(reinterpret_cast<int*>(dest)) = r;
665     return true;
666     }
667    
668     bool Arg::parse_uint_radix(const char* str,
669     int n,
670     void* dest,
671     int radix) {
672     unsigned long r;
673     if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
674     if (r > UINT_MAX) return false; // Out of range
675     *(reinterpret_cast<unsigned int*>(dest)) = r;
676     return true;
677     }
678    
679     bool Arg::parse_longlong_radix(const char* str,
680     int n,
681     void* dest,
682     int radix) {
683     #ifndef HAVE_LONG_LONG
684     return false;
685     #else
686     if (n == 0) return false;
687     char buf[kMaxNumberLength+1];
688     str = TerminateNumber(buf, str, n);
689     char* end;
690     errno = 0;
691     #if defined HAVE_STRTOQ
692     long long r = strtoq(str, &end, radix);
693     #elif defined HAVE_STRTOLL
694     long long r = strtoll(str, &end, radix);
695     #else
696     #error parse_longlong_radix: cannot convert input to a long-long
697     #endif
698     if (end != str + n) return false; // Leftover junk
699     if (errno) return false;
700     *(reinterpret_cast<long long*>(dest)) = r;
701     return true;
702     #endif /* HAVE_LONG_LONG */
703     }
704    
705     bool Arg::parse_ulonglong_radix(const char* str,
706     int n,
707     void* dest,
708     int radix) {
709     #ifndef HAVE_UNSIGNED_LONG_LONG
710     return false;
711     #else
712     if (n == 0) return false;
713     char buf[kMaxNumberLength+1];
714     str = TerminateNumber(buf, str, n);
715 nigel 87 if (str[0] == '-') return false; // strtoull() on a negative number?!
716 nigel 77 char* end;
717     errno = 0;
718     #if defined HAVE_STRTOQ
719     unsigned long long r = strtouq(str, &end, radix);
720     #elif defined HAVE_STRTOLL
721     unsigned long long r = strtoull(str, &end, radix);
722     #else
723     #error parse_ulonglong_radix: cannot convert input to a long-long
724     #endif
725     if (end != str + n) return false; // Leftover junk
726     if (errno) return false;
727     *(reinterpret_cast<unsigned long long*>(dest)) = r;
728     return true;
729     #endif /* HAVE_UNSIGNED_LONG_LONG */
730     }
731    
732     bool Arg::parse_double(const char* str, int n, void* dest) {
733     if (n == 0) return false;
734     static const int kMaxLength = 200;
735     char buf[kMaxLength];
736     if (n >= kMaxLength) return false;
737     memcpy(buf, str, n);
738     buf[n] = '\0';
739     errno = 0;
740     char* end;
741     double r = strtod(buf, &end);
742     if (end != buf + n) return false; // Leftover junk
743     if (errno) return false;
744     *(reinterpret_cast<double*>(dest)) = r;
745     return true;
746     }
747    
748     bool Arg::parse_float(const char* str, int n, void* dest) {
749     double r;
750     if (!parse_double(str, n, &r)) return false;
751     *(reinterpret_cast<float*>(dest)) = static_cast<float>(r);
752     return true;
753     }
754    
755    
756     #define DEFINE_INTEGER_PARSERS(name) \
757     bool Arg::parse_##name(const char* str, int n, void* dest) { \
758     return parse_##name##_radix(str, n, dest, 10); \
759     } \
760     bool Arg::parse_##name##_hex(const char* str, int n, void* dest) { \
761     return parse_##name##_radix(str, n, dest, 16); \
762     } \
763     bool Arg::parse_##name##_octal(const char* str, int n, void* dest) { \
764     return parse_##name##_radix(str, n, dest, 8); \
765     } \
766     bool Arg::parse_##name##_cradix(const char* str, int n, void* dest) { \
767     return parse_##name##_radix(str, n, dest, 0); \
768     }
769    
770     DEFINE_INTEGER_PARSERS(short);
771     DEFINE_INTEGER_PARSERS(ushort);
772     DEFINE_INTEGER_PARSERS(int);
773     DEFINE_INTEGER_PARSERS(uint);
774     DEFINE_INTEGER_PARSERS(long);
775     DEFINE_INTEGER_PARSERS(ulong);
776     DEFINE_INTEGER_PARSERS(longlong);
777     DEFINE_INTEGER_PARSERS(ulonglong);
778    
779     #undef DEFINE_INTEGER_PARSERS
780    
781     } // namespace pcrecpp

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12