| 1 |
nigel |
29 |
/************************************************* |
| 2 |
|
|
* Perl-Compatible Regular Expressions * |
| 3 |
|
|
*************************************************/ |
| 4 |
|
|
|
| 5 |
|
|
/* |
| 6 |
|
|
This is a library of functions to support regular expressions whose syntax |
| 7 |
|
|
and semantics are as close as possible to those of the Perl 5 language. See |
| 8 |
|
|
the file Tech.Notes for some information on the internals. |
| 9 |
|
|
|
| 10 |
|
|
Written by: Philip Hazel <ph10@cam.ac.uk> |
| 11 |
|
|
|
| 12 |
|
|
Copyright (c) 1997-1999 University of Cambridge |
| 13 |
|
|
|
| 14 |
|
|
----------------------------------------------------------------------------- |
| 15 |
|
|
Permission is granted to anyone to use this software for any purpose on any |
| 16 |
|
|
computer system, and to redistribute it freely, subject to the following |
| 17 |
|
|
restrictions: |
| 18 |
|
|
|
| 19 |
|
|
1. This software is distributed in the hope that it will be useful, |
| 20 |
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 21 |
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. |
| 22 |
|
|
|
| 23 |
|
|
2. The origin of this software must not be misrepresented, either by |
| 24 |
|
|
explicit claim or by omission. |
| 25 |
|
|
|
| 26 |
|
|
3. Altered versions must be plainly marked as such, and must not be |
| 27 |
|
|
misrepresented as being the original software. |
| 28 |
|
|
|
| 29 |
|
|
4. If PCRE is embedded in any software that is released under the GNU |
| 30 |
|
|
General Purpose Licence (GPL), then the terms of that licence shall |
| 31 |
|
|
supersede any condition above with which it is incompatible. |
| 32 |
|
|
----------------------------------------------------------------------------- |
| 33 |
|
|
*/ |
| 34 |
|
|
|
| 35 |
|
|
/* This module contains some convenience functions for extracting substrings |
| 36 |
|
|
from the subject string after a regex match has succeeded. The original idea |
| 37 |
|
|
for these functions came from Scott Wimer <scottw@cgibuilder.com>. */ |
| 38 |
|
|
|
| 39 |
|
|
|
| 40 |
|
|
/* Include the internals header, which itself includes Standard C headers plus |
| 41 |
|
|
the external pcre header. */ |
| 42 |
|
|
|
| 43 |
|
|
#include "internal.h" |
| 44 |
|
|
|
| 45 |
|
|
|
| 46 |
|
|
|
| 47 |
|
|
/************************************************* |
| 48 |
|
|
* Copy captured string to given buffer * |
| 49 |
|
|
*************************************************/ |
| 50 |
|
|
|
| 51 |
|
|
/* This function copies a single captured substring into a given buffer. |
| 52 |
|
|
Note that we use memcpy() rather than strncpy() in case there are binary zeros |
| 53 |
|
|
in the string. |
| 54 |
|
|
|
| 55 |
|
|
Arguments: |
| 56 |
|
|
subject the subject string that was matched |
| 57 |
|
|
ovector pointer to the offsets table |
| 58 |
|
|
stringcount the number of substrings that were captured |
| 59 |
|
|
(i.e. the yield of the pcre_exec call, unless |
| 60 |
|
|
that was zero, in which case it should be 1/3 |
| 61 |
|
|
of the offset table size) |
| 62 |
|
|
stringnumber the number of the required substring |
| 63 |
|
|
buffer where to put the substring |
| 64 |
|
|
size the size of the buffer |
| 65 |
|
|
|
| 66 |
|
|
Returns: if successful: |
| 67 |
|
|
the length of the copied string, not including the zero |
| 68 |
|
|
that is put on the end; can be zero |
| 69 |
|
|
if not successful: |
| 70 |
|
|
PCRE_ERROR_NOMEMORY (-6) buffer too small |
| 71 |
|
|
PCRE_ERROR_NOSUBSTRING (-7) no such captured substring |
| 72 |
|
|
*/ |
| 73 |
|
|
|
| 74 |
|
|
int |
| 75 |
|
|
pcre_copy_substring(const char *subject, int *ovector, int stringcount, |
| 76 |
|
|
int stringnumber, char *buffer, int size) |
| 77 |
|
|
{ |
| 78 |
|
|
int yield; |
| 79 |
|
|
if (stringnumber < 0 || stringnumber >= stringcount) |
| 80 |
|
|
return PCRE_ERROR_NOSUBSTRING; |
| 81 |
|
|
stringnumber *= 2; |
| 82 |
|
|
yield = ovector[stringnumber+1] - ovector[stringnumber]; |
| 83 |
|
|
if (size < yield + 1) return PCRE_ERROR_NOMEMORY; |
| 84 |
|
|
memcpy(buffer, subject + ovector[stringnumber], yield); |
| 85 |
|
|
buffer[yield] = 0; |
| 86 |
|
|
return yield; |
| 87 |
|
|
} |
| 88 |
|
|
|
| 89 |
|
|
|
| 90 |
|
|
|
| 91 |
|
|
/************************************************* |
| 92 |
|
|
* Copy all captured strings to new store * |
| 93 |
|
|
*************************************************/ |
| 94 |
|
|
|
| 95 |
|
|
/* This function gets one chunk of store and builds a list of pointers and all |
| 96 |
|
|
of the captured substrings in it. A NULL pointer is put on the end of the list. |
| 97 |
|
|
|
| 98 |
|
|
Arguments: |
| 99 |
|
|
subject the subject string that was matched |
| 100 |
|
|
ovector pointer to the offsets table |
| 101 |
|
|
stringcount the number of substrings that were captured |
| 102 |
|
|
(i.e. the yield of the pcre_exec call, unless |
| 103 |
|
|
that was zero, in which case it should be 1/3 |
| 104 |
|
|
of the offset table size) |
| 105 |
|
|
listptr set to point to the list of pointers |
| 106 |
|
|
|
| 107 |
|
|
Returns: if successful: 0 |
| 108 |
|
|
if not successful: |
| 109 |
|
|
PCRE_ERROR_NOMEMORY (-6) failed to get store |
| 110 |
|
|
*/ |
| 111 |
|
|
|
| 112 |
|
|
int |
| 113 |
|
|
pcre_get_substring_list(const char *subject, int *ovector, int stringcount, |
| 114 |
|
|
const char ***listptr) |
| 115 |
|
|
{ |
| 116 |
|
|
int i; |
| 117 |
|
|
int size = sizeof(char *); |
| 118 |
|
|
int double_count = stringcount * 2; |
| 119 |
|
|
char **stringlist; |
| 120 |
|
|
char *p; |
| 121 |
|
|
|
| 122 |
|
|
for (i = 0; i < double_count; i += 2) |
| 123 |
|
|
size += sizeof(char *) + ovector[i+1] - ovector[i] + 1; |
| 124 |
|
|
|
| 125 |
|
|
stringlist = (char **)(pcre_malloc)(size); |
| 126 |
|
|
if (stringlist == NULL) return PCRE_ERROR_NOMEMORY; |
| 127 |
|
|
|
| 128 |
|
|
*listptr = (const char **)stringlist; |
| 129 |
|
|
p = (char *)(stringlist + stringcount + 1); |
| 130 |
|
|
|
| 131 |
|
|
for (i = 0; i < double_count; i += 2) |
| 132 |
|
|
{ |
| 133 |
|
|
int len = ovector[i+1] - ovector[i]; |
| 134 |
|
|
memcpy(p, subject + ovector[i], len); |
| 135 |
|
|
*stringlist++ = p; |
| 136 |
|
|
p += len; |
| 137 |
|
|
*p++ = 0; |
| 138 |
|
|
} |
| 139 |
|
|
|
| 140 |
|
|
*stringlist = NULL; |
| 141 |
|
|
return 0; |
| 142 |
|
|
} |
| 143 |
|
|
|
| 144 |
|
|
|
| 145 |
|
|
|
| 146 |
|
|
/************************************************* |
| 147 |
|
|
* Copy captured string to new store * |
| 148 |
|
|
*************************************************/ |
| 149 |
|
|
|
| 150 |
|
|
/* This function copies a single captured substring into a piece of new |
| 151 |
|
|
store |
| 152 |
|
|
|
| 153 |
|
|
Arguments: |
| 154 |
|
|
subject the subject string that was matched |
| 155 |
|
|
ovector pointer to the offsets table |
| 156 |
|
|
stringcount the number of substrings that were captured |
| 157 |
|
|
(i.e. the yield of the pcre_exec call, unless |
| 158 |
|
|
that was zero, in which case it should be 1/3 |
| 159 |
|
|
of the offset table size) |
| 160 |
|
|
stringnumber the number of the required substring |
| 161 |
|
|
stringptr where to put a pointer to the substring |
| 162 |
|
|
|
| 163 |
|
|
Returns: if successful: |
| 164 |
|
|
the length of the string, not including the zero that |
| 165 |
|
|
is put on the end; can be zero |
| 166 |
|
|
if not successful: |
| 167 |
|
|
PCRE_ERROR_NOMEMORY (-6) failed to get store |
| 168 |
|
|
PCRE_ERROR_NOSUBSTRING (-7) substring not present |
| 169 |
|
|
*/ |
| 170 |
|
|
|
| 171 |
|
|
int |
| 172 |
|
|
pcre_get_substring(const char *subject, int *ovector, int stringcount, |
| 173 |
|
|
int stringnumber, const char **stringptr) |
| 174 |
|
|
{ |
| 175 |
|
|
int yield; |
| 176 |
|
|
char *substring; |
| 177 |
|
|
if (stringnumber < 0 || stringnumber >= stringcount) |
| 178 |
|
|
return PCRE_ERROR_NOSUBSTRING; |
| 179 |
|
|
stringnumber *= 2; |
| 180 |
|
|
yield = ovector[stringnumber+1] - ovector[stringnumber]; |
| 181 |
|
|
substring = (char *)(pcre_malloc)(yield + 1); |
| 182 |
|
|
if (substring == NULL) return PCRE_ERROR_NOMEMORY; |
| 183 |
|
|
memcpy(substring, subject + ovector[stringnumber], yield); |
| 184 |
|
|
substring[yield] = 0; |
| 185 |
|
|
*stringptr = substring; |
| 186 |
|
|
return yield; |
| 187 |
|
|
} |
| 188 |
|
|
|
| 189 |
|
|
/* End of get.c */ |