pcrecpp.cc

php-4.4.7学习linux时下载的源代码

  return Rewrite(out, rewrite, text, vec, matches);
}

/*static*/ string RE::QuoteMeta(const StringPiece& unquoted) {
  string result;

  // Escape any ascii character not in [A-Za-z_0-9].
  //
  // Note that it's legal to escape a character even if it has no
  // special meaning in a regular expression -- so this function does
  // that.  (This also makes it identical to the perl function of the
  // same name; see `perldoc -f quotemeta`.)
  for (int ii = 0; ii < unquoted.size(); ++ii) {
    // Note that using 'isalnum' here raises the benchmark time from
    // 32ns to 58ns:
    if ((unquoted[ii] < 'a' || unquoted[ii] > 'z') &&
        (unquoted[ii] < 'A' || unquoted[ii] > 'Z') &&
        (unquoted[ii] < '0' || unquoted[ii] > '9') &&
        unquoted[ii] != '_' &&
        // If this is the part of a UTF8 or Latin1 character, we need
        // to copy this byte without escaping.  Experimentally this is
        // what works correctly with the regexp library.
        !(unquoted[ii] & 128)) {
      result += '\\';
    }
    result += unquoted[ii];
  }

  return result;
}

/***** Actual matching and rewriting code *****/

int RE::TryMatch(const StringPiece& text,
                 int startpos,
                 Anchor anchor,
                 int *vec,
                 int vecsize) const {
  pcre* re = (anchor == ANCHOR_BOTH) ? re_full_ : re_partial_;
  if (re == NULL) {
    //fprintf(stderr, "Matching against invalid re: %s\n", error_->c_str());
    return 0;
  }

  pcre_extra extra = { 0 };
  if (options_.match_limit() > 0) {
    extra.flags |= PCRE_EXTRA_MATCH_LIMIT;
    extra.match_limit = options_.match_limit();
  }
  if (options_.match_limit_recursion() > 0) {
    extra.flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
    extra.match_limit_recursion = options_.match_limit_recursion();
  }
  int rc = pcre_exec(re,              // The regular expression object
                     &extra,
                     (text.data() == NULL) ? "" : text.data(),
                     text.size(),
                     startpos,
                     (anchor == UNANCHORED) ? 0 : PCRE_ANCHORED,
                     vec,
                     vecsize);

  // Handle errors
  if (rc == PCRE_ERROR_NOMATCH) {
    return 0;
  } else if (rc < 0) {
    //fprintf(stderr, "Unexpected return code: %d when matching '%s'\n",
    //        re, pattern_.c_str());
    return 0;
  } else if (rc == 0) {
    // pcre_exec() returns 0 as a special case when the number of
    // capturing subpatterns exceeds the size of the vector.
    // When this happens, there is a match and the output vector
    // is filled, but we miss out on the positions of the extra subpatterns.
    rc = vecsize / 2;
  }

  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
  }

  return rc;
}

bool RE::DoMatchImpl(const StringPiece& text,
                     Anchor anchor,
                     int* consumed,
                     const Arg* const* args,
                     int n,
                     int* vec,
                     int vecsize) const {
  assert((1 + n) * 3 <= vecsize);  // results + PCRE workspace
  int matches = TryMatch(text, 0, anchor, vec, vecsize);
  assert(matches >= 0);  // TryMatch never returns negatives
  if (matches == 0)
    return false;

  *consumed = vec[1];

  if (n == 0 || args == NULL) {
    // We are not interested in results
    return true;
  }

  if (NumberOfCapturingGroups() < n) {
    // RE has fewer capturing groups than number of arg pointers passed in
    return false;
  }

  // If we got here, we must have matched the whole pattern.
  // We do not need (can not do) any more checks on the value of 'matches' here
  // -- see the comment for TryMatch.
  for (int i = 0; i < n; i++) {
    const int start = vec[2*(i+1)];
    const int limit = vec[2*(i+1)+1];
    if (!args[i]->Parse(text.data() + start, limit-start)) {
      // TODO: Should we indicate what the error was?
      return false;
    }
  }

  return true;
}

bool RE::DoMatch(const StringPiece& text,
                 Anchor anchor,
                 int* consumed,
                 const Arg* const args[],
                 int n) const {
  assert(n >= 0);
  size_t const vecsize = (1 + n) * 3;  // results + PCRE workspace
                                       // (as for kVecSize)
  int space[21];   // use stack allocation for small vecsize (common case)
  int* vec = vecsize <= 21 ? space : new int[vecsize];
  bool retval = DoMatchImpl(text, anchor, consumed, args, n, vec, vecsize);
  if (vec != space) delete [] vec;
  return retval;
}

bool RE::Rewrite(string *out, const StringPiece &rewrite,
                 const StringPiece &text, int *vec, int veclen) const {
  for (const char *s = rewrite.data(), *end = s + rewrite.size();
       s < end; s++) {
    int c = *s;
    if (c == '\\') {
      c = *++s;
      if (isdigit(c)) {
        int n = (c - '0');
        if (n >= veclen) {
          //fprintf(stderr, requested group %d in regexp %.*s\n",
          //        n, rewrite.size(), rewrite.data());
          return false;
        }
        int start = vec[2 * n];
        if (start >= 0)
          out->append(text.data() + start, vec[2 * n + 1] - start);
      } else if (c == '\\') {
        out->push_back('\\');
      } else {
        //fprintf(stderr, "invalid rewrite pattern: %.*s\n",
        //        rewrite.size(), rewrite.data());
        return false;
      }
    } else {
      out->push_back(c);
    }
  }
  return true;
}

// Return the number of capturing subpatterns, or -1 if the
// regexp wasn't valid on construction.
int RE::NumberOfCapturingGroups() const {
  if (re_partial_ == NULL) return -1;

  int result;
  int pcre_retval = pcre_fullinfo(re_partial_,  // The regular expression object
                                  NULL,         // We did not study the pattern
                                  PCRE_INFO_CAPTURECOUNT,
                                  &result);
  assert(pcre_retval == 0);
  return result;
}

/***** Parsers for various types *****/

bool Arg::parse_null(const char* str, int n, void* dest) {
  // We fail if somebody asked us to store into a non-NULL void* pointer
  return (dest == NULL);
}

bool Arg::parse_string(const char* str, int n, void* dest) {
  reinterpret_cast<string*>(dest)->assign(str, n);
  return true;
}

bool Arg::parse_stringpiece(const char* str, int n, void* dest) {
  reinterpret_cast<StringPiece*>(dest)->set(str, n);
  return true;
}

bool Arg::parse_char(const char* str, int n, void* dest) {
  if (n != 1) return false;
  *(reinterpret_cast<char*>(dest)) = str[0];
  return true;
}

bool Arg::parse_uchar(const char* str, int n, void* dest) {
  if (n != 1) return false;
  *(reinterpret_cast<unsigned char*>(dest)) = str[0];
  return true;
}

// Largest number spec that we are willing to parse
static const int kMaxNumberLength = 32;

// REQUIRES "buf" must have length at least kMaxNumberLength+1
// REQUIRES "n > 0"
// Copies "str" into "buf" and null-terminates if necessary.
// Returns one of:
//      a. "str" if no termination is needed
//      b. "buf" if the string was copied and null-terminated
//      c. "" if the input was invalid and has no hope of being parsed
static const char* TerminateNumber(char* buf, const char* str, int n) {
  if ((n > 0) && isspace(*str)) {
    // We are less forgiving than the strtoxxx() routines and do not
    // allow leading spaces.
    return "";
  }

  // See if the character right after the input text may potentially
  // look like a digit.
  if (isdigit(str[n]) ||
      ((str[n] >= 'a') && (str[n] <= 'f')) ||
      ((str[n] >= 'A') && (str[n] <= 'F'))) {
    if (n > kMaxNumberLength) return ""; // Input too big to be a valid number
    memcpy(buf, str, n);
    buf[n] = '\0';
    return buf;
  } else {
    // We can parse right out of the supplied string, so return it.
    return str;
  }
}

bool Arg::parse_long_radix(const char* str,
                           int n,
                           void* dest,
                           int radix) {
  if (n == 0) return false;
  char buf[kMaxNumberLength+1];
  str = TerminateNumber(buf, str, n);
  char* end;
  errno = 0;
  long r = strtol(str, &end, radix);
  if (end != str + n) return false;   // Leftover junk
  if (errno) return false;
  *(reinterpret_cast<long*>(dest)) = r;
  return true;
}

bool Arg::parse_ulong_radix(const char* str,
                            int n,
                            void* dest,
                            int radix) {
  if (n == 0) return false;
  char buf[kMaxNumberLength+1];
  str = TerminateNumber(buf, str, n);
  if (str[0] == '-') return false;    // strtoul() on a negative number?!
  char* end;
  errno = 0;
  unsigned long r = strtoul(str, &end, radix);
  if (end != str + n) return false;   // Leftover junk
  if (errno) return false;
  *(reinterpret_cast<unsigned long*>(dest)) = r;
  return true;
}

bool Arg::parse_short_radix(const char* str,
                            int n,
                            void* dest,
                            int radix) {
  long r;
  if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
  if (r < SHRT_MIN || r > SHRT_MAX) return false;       // Out of range
  *(reinterpret_cast<short*>(dest)) = r;
  return true;
}

bool Arg::parse_ushort_radix(const char* str,
                             int n,
                             void* dest,
                             int radix) {
  unsigned long r;
  if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
  if (r > USHRT_MAX) return false;                      // Out of range
  *(reinterpret_cast<unsigned short*>(dest)) = r;
  return true;
}

bool Arg::parse_int_radix(const char* str,
                          int n,
                          void* dest,
                          int radix) {
  long r;
  if (!parse_long_radix(str, n, &r, radix)) return false; // Could not parse
  if (r < INT_MIN || r > INT_MAX) return false;         // Out of range
  *(reinterpret_cast<int*>(dest)) = r;
  return true;
}

bool Arg::parse_uint_radix(const char* str,
                           int n,
                           void* dest,
                           int radix) {
  unsigned long r;
  if (!parse_ulong_radix(str, n, &r, radix)) return false; // Could not parse
  if (r > UINT_MAX) return false;                       // Out of range
  *(reinterpret_cast<unsigned int*>(dest)) = r;
  return true;
}

bool Arg::parse_longlong_radix(const char* str,
                               int n,
                               void* dest,
                               int radix) {
#ifndef HAVE_LONG_LONG
  return false;
#else
  if (n == 0) return false;
  char buf[kMaxNumberLength+1];
  str = TerminateNumber(buf, str, n);
  char* end;
  errno = 0;
#if defined HAVE_STRTOQ
  long long r = strtoq(str, &end, radix);
#elif defined HAVE_STRTOLL
  long long r = strtoll(str, &end, radix);
#else
#error parse_longlong_radix: cannot convert input to a long-long
#endif
  if (end != str + n) return false;   // Leftover junk
  if (errno) return false;
  *(reinterpret_cast<long long*>(dest)) = r;
  return true;
#endif   /* HAVE_LONG_LONG */
}

bool Arg::parse_ulonglong_radix(const char* str,
                                int n,
                                void* dest,
                                int radix) {
#ifndef HAVE_UNSIGNED_LONG_LONG
  return false;
#else
  if (n == 0) return false;
  char buf[kMaxNumberLength+1];
  str = TerminateNumber(buf, str, n);
  if (str[0] == '-') return false;    // strtoull() on a negative number?!
  char* end;
  errno = 0;
#if defined HAVE_STRTOQ
  unsigned long long r = strtouq(str, &end, radix);
#elif defined HAVE_STRTOLL
  unsigned long long r = strtoull(str, &end, radix);
#else
#error parse_ulonglong_radix: cannot convert input to a long-long
#endif
  if (end != str + n) return false;   // Leftover junk
  if (errno) return false;
  *(reinterpret_cast<unsigned long long*>(dest)) = r;
  return true;
#endif   /* HAVE_UNSIGNED_LONG_LONG */
}

bool Arg::parse_double(const char* str, int n, void* dest) {
  if (n == 0) return false;
  static const int kMaxLength = 200;
  char buf[kMaxLength];
  if (n >= kMaxLength) return false;
  memcpy(buf, str, n);
  buf[n] = '\0';
  errno = 0;
  char* end;
  double r = strtod(buf, &end);
  if (end != buf + n) return false;   // Leftover junk
  if (errno) return false;
  *(reinterpret_cast<double*>(dest)) = r;
  return true;
}

bool Arg::parse_float(const char* str, int n, void* dest) {
  double r;
  if (!parse_double(str, n, &r)) return false;
  *(reinterpret_cast<float*>(dest)) = static_cast<float>(r);
  return true;
}


#define DEFINE_INTEGER_PARSERS(name)                                    \
  bool Arg::parse_##name(const char* str, int n, void* dest) {          \
    return parse_##name##_radix(str, n, dest, 10);                      \
  }                                                                     \
  bool Arg::parse_##name##_hex(const char* str, int n, void* dest) {    \
    return parse_##name##_radix(str, n, dest, 16);                      \
  }                                                                     \
  bool Arg::parse_##name##_octal(const char* str, int n, void* dest) {  \
    return parse_##name##_radix(str, n, dest, 8);                       \
  }                                                                     \
  bool Arg::parse_##name##_cradix(const char* str, int n, void* dest) { \
    return parse_##name##_radix(str, n, dest, 0);                       \
  }

DEFINE_INTEGER_PARSERS(short)      /*                                   */
DEFINE_INTEGER_PARSERS(ushort)     /*                                   */
DEFINE_INTEGER_PARSERS(int)        /* Don't use semicolons after these  */
DEFINE_INTEGER_PARSERS(uint)       /* statements because they can cause */
DEFINE_INTEGER_PARSERS(long)       /* compiler warnings if the checking */
DEFINE_INTEGER_PARSERS(ulong)      /* level is turned up high enough.   */
DEFINE_INTEGER_PARSERS(longlong)   /*                                   */
DEFINE_INTEGER_PARSERS(ulonglong)  /*                                   */

#undef DEFINE_INTEGER_PARSERS

}   // namespace pcrecpp