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

Diff of /code/trunk/pcrecpp.cc

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 179 by ph10, Wed Jun 13 08:53:45 2007 UTC
# Line 29  Line 29 
29  //  //
30  // Author: Sanjay Ghemawat  // Author: Sanjay Ghemawat
31    
32    #ifdef HAVE_CONFIG_H
33    #  include <config.h>
34    #endif
35    
36  #include <stdlib.h>  #include <stdlib.h>
37  #include <stdio.h>  #include <stdio.h>
38  #include <ctype.h>  #include <ctype.h>
# Line 36  Line 40 
40  #include <assert.h>  #include <assert.h>
41  #include <errno.h>  #include <errno.h>
42  #include <string>  #include <string>
43  #include "config.h"  #include <algorithm>
44  // We need this to compile the proper dll on windows/msys.  This is copied  // 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.  // from pcre_internal.h.  It would probably be better just to include that.
46  #define PCRE_DEFINITION  /* Win32 __declspec(export) trigger for .dll */  #define PCRE_DEFINITION  /* Win32 __declspec(export) trigger for .dll */
47  #include "pcre.h"  #include <pcre.h>
48  #include "pcre_stringpiece.h"  #include "pcre_stringpiece.h"
49  #include "pcrecpp.h"  #include "pcrecpp.h"
50    
# Line 60  static const string empty_string; Line 64  static const string empty_string;
64  // If the user doesn't ask for any options, we just use this one  // If the user doesn't ask for any options, we just use this one
65  static RE_Options default_options;  static RE_Options default_options;
66    
67  void RE::Init(const char* pat, const RE_Options* options) {  void RE::Init(const string& pat, const RE_Options* options) {
68    pattern_ = pat;    pattern_ = pat;
69    if (options == NULL) {    if (options == NULL) {
70      options_ = default_options;      options_ = default_options;
# Line 73  void RE::Init(const char* pat, const RE_ Line 77  void RE::Init(const char* pat, const RE_
77    
78    re_partial_ = Compile(UNANCHORED);    re_partial_ = Compile(UNANCHORED);
79    if (re_partial_ != NULL) {    if (re_partial_ != NULL) {
80      // Check for complicated patterns.  The following change is      re_full_ = Compile(ANCHOR_BOTH);
     // conservative in that it may treat some "simple" patterns  
     // as "complex" (e.g., if the vertical bar is in a character  
     // class or is escaped).  But it seems good enough.  
     if (strchr(pat, '|') == NULL) {  
       // Simple pattern: we can use position-based checks to perform  
       // fully anchored matches  
       re_full_ = re_partial_;  
     } else {  
       // We need a special pattern for anchored matches  
       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() {  RE::~RE() {
92    if (re_full_ != NULL && re_full_ != re_partial_) (*pcre_free)(re_full_);    Cleanup();
   if (re_partial_ != NULL)                         (*pcre_free)(re_partial_);  
   if (error_ != &empty_string)                     delete error_;  
93  }  }
94    
95    
96  pcre* RE::Compile(Anchor anchor) {  pcre* RE::Compile(Anchor anchor) {
97    // First, convert RE_Options into pcre options    // First, convert RE_Options into pcre options
98    int pcre_options = 0;    int pcre_options = 0;
99    if (options_.utf8())    pcre_options = options_.all_options();
     pcre_options |= PCRE_UTF8;  
100    
101    // Special treatment for anchoring.  This is needed because at    // Special treatment for anchoring.  This is needed because at
102    // runtime pcre only provides an option for anchoring at the    // runtime pcre only provides an option for anchoring at the
# Line 332  bool RE::Replace(const StringPiece& rewr Line 330  bool RE::Replace(const StringPiece& rewr
330    return true;    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,  int RE::GlobalReplace(const StringPiece& rewrite,
358                        string *str) const {                        string *str) const {
359    int count = 0;    int count = 0;
# Line 350  int RE::GlobalReplace(const StringPiece& Line 372  int RE::GlobalReplace(const StringPiece&
372      if (matchstart == matchend && matchstart == lastend) {      if (matchstart == matchend && matchstart == lastend) {
373        // advance one character if we matched an empty string at the same        // advance one character if we matched an empty string at the same
374        // place as the last match occurred        // place as the last match occurred
375        if (start < static_cast<int>(str->length()))        matchend = start + 1;
376          out.push_back((*str)[start]);        // If the current char is CR and we're in CRLF mode, skip LF too.
377        start++;        // 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 {      } else {
397        out.append(*str, start, matchstart - start);        out.append(*str, start, matchstart - start);
398        Rewrite(&out, rewrite, *str, vec, matches);        Rewrite(&out, rewrite, *str, vec, matches);
# Line 378  bool RE::Extract(const StringPiece& rewr Line 418  bool RE::Extract(const StringPiece& rewr
418    int matches = TryMatch(text, 0, UNANCHORED, vec, kVecSize);    int matches = TryMatch(text, 0, UNANCHORED, vec, kVecSize);
419    if (matches == 0)    if (matches == 0)
420      return false;      return false;
421    out->clear();    out->erase();
422    return Rewrite(out, rewrite, text, vec, matches);    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 *****/  /***** Actual matching and rewriting code *****/
454    
455  int RE::TryMatch(const StringPiece& text,  int RE::TryMatch(const StringPiece& text,
# Line 397  int RE::TryMatch(const StringPiece& text Line 465  int RE::TryMatch(const StringPiece& text
465    
466    pcre_extra extra = { 0 };    pcre_extra extra = { 0 };
467    if (options_.match_limit() > 0) {    if (options_.match_limit() > 0) {
468      extra.flags = PCRE_EXTRA_MATCH_LIMIT;      extra.flags |= PCRE_EXTRA_MATCH_LIMIT;
469      extra.match_limit = options_.match_limit();      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    int rc = pcre_exec(re,              // The regular expression object
476                       &extra,                       &extra,
477                       text.data(),                       (text.data() == NULL) ? "" : text.data(),
478                       text.size(),                       text.size(),
479                       startpos,                       startpos,
480                       (anchor == UNANCHORED) ? 0 : PCRE_ANCHORED,                       (anchor == UNANCHORED) ? 0 : PCRE_ANCHORED,
# Line 424  int RE::TryMatch(const StringPiece& text Line 496  int RE::TryMatch(const StringPiece& text
496      rc = vecsize / 2;      rc = vecsize / 2;
497    }    }
498    
   if ((anchor == ANCHOR_BOTH) && (re_full_ == re_partial_)) {  
     // We need an extra check to make sure that the match extended  
     // to the end of the input string  
     assert(vec[0] == 0);                 // PCRE_ANCHORED forces starting match  
     if (vec[1] != text.size()) return 0; // Did not get ending match  
   }  
   
499    return rc;    return rc;
500  }  }
501    
# Line 449  bool RE::DoMatchImpl(const StringPiece& Line 514  bool RE::DoMatchImpl(const StringPiece&
514    
515    *consumed = vec[1];    *consumed = vec[1];
516    
517    if (args == NULL) {    if (n == 0 || args == NULL) {
518      // We are not interested in results      // We are not interested in results
519      return true;      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.    // 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    // We do not need (can not do) any more checks on the value of 'matches' here
529    // -- see the comment for TryMatch.    // -- see the comment for TryMatch.
# Line 517  bool RE::Rewrite(string *out, const Stri Line 587  bool RE::Rewrite(string *out, const Stri
587    
588  // Return the number of capturing subpatterns, or -1 if the  // Return the number of capturing subpatterns, or -1 if the
589  // regexp wasn't valid on construction.  // regexp wasn't valid on construction.
590  int RE::NumberOfCapturingGroups() {  int RE::NumberOfCapturingGroups() const {
591    if (re_partial_ == NULL) return -1;    if (re_partial_ == NULL) return -1;
592    
593    int result;    int result;
# Line 613  bool Arg::parse_ulong_radix(const char* Line 683  bool Arg::parse_ulong_radix(const char*
683    if (n == 0) return false;    if (n == 0) return false;
684    char buf[kMaxNumberLength+1];    char buf[kMaxNumberLength+1];
685    str = TerminateNumber(buf, str, n);    str = TerminateNumber(buf, str, n);
686      if (str[0] == '-') return false;    // strtoul() on a negative number?!
687    char* end;    char* end;
688    errno = 0;    errno = 0;
689    unsigned long r = strtoul(str, &end, radix);    unsigned long r = strtoul(str, &end, radix);
# Line 702  bool Arg::parse_ulonglong_radix(const ch Line 773  bool Arg::parse_ulonglong_radix(const ch
773    if (n == 0) return false;    if (n == 0) return false;
774    char buf[kMaxNumberLength+1];    char buf[kMaxNumberLength+1];
775    str = TerminateNumber(buf, str, n);    str = TerminateNumber(buf, str, n);
776      if (str[0] == '-') return false;    // strtoull() on a negative number?!
777    char* end;    char* end;
778    errno = 0;    errno = 0;
779  #if defined HAVE_STRTOQ  #if defined HAVE_STRTOQ
# Line 756  bool Arg::parse_float(const char* str, i Line 828  bool Arg::parse_float(const char* str, i
828      return parse_##name##_radix(str, n, dest, 0);                       \      return parse_##name##_radix(str, n, dest, 0);                       \
829    }    }
830    
831  DEFINE_INTEGER_PARSERS(short);  DEFINE_INTEGER_PARSERS(short)      /*                                   */
832  DEFINE_INTEGER_PARSERS(ushort);  DEFINE_INTEGER_PARSERS(ushort)     /*                                   */
833  DEFINE_INTEGER_PARSERS(int);  DEFINE_INTEGER_PARSERS(int)        /* Don't use semicolons after these  */
834  DEFINE_INTEGER_PARSERS(uint);  DEFINE_INTEGER_PARSERS(uint)       /* statements because they can cause */
835  DEFINE_INTEGER_PARSERS(long);  DEFINE_INTEGER_PARSERS(long)       /* compiler warnings if the checking */
836  DEFINE_INTEGER_PARSERS(ulong);  DEFINE_INTEGER_PARSERS(ulong)      /* level is turned up high enough.   */
837  DEFINE_INTEGER_PARSERS(longlong);  DEFINE_INTEGER_PARSERS(longlong)   /*                                   */
838  DEFINE_INTEGER_PARSERS(ulonglong);  DEFINE_INTEGER_PARSERS(ulonglong)  /*                                   */
839    
840  #undef DEFINE_INTEGER_PARSERS  #undef DEFINE_INTEGER_PARSERS
841    

Legend:
Removed from v.77  
changed lines
  Added in v.179

  ViewVC Help
Powered by ViewVC 1.1.5