/[pcre]/code/trunk/pcreposix.c
ViewVC logotype

Diff of /code/trunk/pcreposix.c

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

revision 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 199 by ph10, Tue Jul 31 14:39:09 2007 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2005 University of Cambridge             Copyright (c) 1997-2007 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  functions. */  functions. */
43    
44    
45    #ifdef HAVE_CONFIG_H
46    #include <config.h>
47    #endif
48    
49    
50    /* Ensure that the PCREPOSIX_EXP_xxx macros are set appropriately for
51    compiling these functions. This must come before including pcreposix.h, where
52    they are set for an application (using these functions) if they have not
53    previously been set. */
54    
55    #if defined(_WIN32) && !defined(PCRE_STATIC)
56    #  define PCREPOSIX_EXP_DECL extern __declspec(dllexport)
57    #  define PCREPOSIX_EXP_DEFN __declspec(dllexport)
58    #endif
59    
60  #include "pcre_internal.h"  #include "pcre_internal.h"
61  #include "pcreposix.h"  #include "pcreposix.h"
 #include "stdlib.h"  
62    
63    
64    
# Line 77  static const int eint[] = { Line 91  static const int eint[] = {
91    REG_ASSERT,  /* internal error: code overflow */    REG_ASSERT,  /* internal error: code overflow */
92    REG_BADPAT,  /* unrecognized character after (?< */    REG_BADPAT,  /* unrecognized character after (?< */
93    REG_BADPAT,  /* lookbehind assertion is not fixed length */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
94    REG_BADPAT,  /* malformed number after (?( */    REG_BADPAT,  /* malformed number or name after (?( */
95    REG_BADPAT,  /* conditional group containe more than two branches */    REG_BADPAT,  /* conditional group contains more than two branches */
96    REG_BADPAT,  /* assertion expected after (?( */    REG_BADPAT,  /* assertion expected after (?( */
97    REG_BADPAT,  /* (?R or (?digits must be followed by ) */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
98    REG_ECTYPE,  /* unknown POSIX class name */    REG_ECTYPE,  /* unknown POSIX class name */
99    REG_BADPAT,  /* POSIX collating elements are not supported */    REG_BADPAT,  /* POSIX collating elements are not supported */
100    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */
# Line 93  static const int eint[] = { Line 107  static const int eint[] = {
107    REG_BADPAT,  /* closing ) for (?C expected */    REG_BADPAT,  /* closing ) for (?C expected */
108    REG_BADPAT,  /* recursive call could loop indefinitely */    REG_BADPAT,  /* recursive call could loop indefinitely */
109    REG_BADPAT,  /* unrecognized character after (?P */    REG_BADPAT,  /* unrecognized character after (?P */
110    REG_BADPAT,  /* syntax error after (?P */    REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
111    REG_BADPAT,  /* two named groups have the same name */    REG_BADPAT,  /* two named subpatterns have the same name */
112    REG_BADPAT,  /* invalid UTF-8 string */    REG_BADPAT,  /* invalid UTF-8 string */
113    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
114    REG_BADPAT,  /* malformed \P or \p sequence */    REG_BADPAT,  /* malformed \P or \p sequence */
115    REG_BADPAT   /* unknown property name after \P or \p */    REG_BADPAT,  /* unknown property name after \P or \p */
116      REG_BADPAT,  /* subpattern name is too long (maximum 32 characters) */
117      REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */
118      REG_BADPAT,  /* repeated subpattern is too long */
119      REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */
120      REG_BADPAT,  /* internal error: overran compiling workspace */
121      REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */
122      REG_BADPAT,  /* DEFINE group contains more than one branch */
123      REG_BADPAT,  /* repeating a DEFINE group is not allowed */
124      REG_INVARG,  /* inconsistent NEWLINE options */
125      REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */
126      REG_BADPAT   /* (?+ or (?- must be followed by a non-zero number */
127  };  };
128    
129  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 131  static const char *const pstring[] = { Line 156  static const char *const pstring[] = {
156  *          Translate error code to string        *  *          Translate error code to string        *
157  *************************************************/  *************************************************/
158    
159  PCRE_EXPORT size_t  PCREPOSIX_EXP_DEFN size_t
160  regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)  regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
161  {  {
162  const char *message, *addmessage;  const char *message, *addmessage;
# Line 166  return length + addlength; Line 191  return length + addlength;
191  *           Free store held by a regex           *  *           Free store held by a regex           *
192  *************************************************/  *************************************************/
193    
194  PCRE_EXPORT void  PCREPOSIX_EXP_DEFN void
195  regfree(regex_t *preg)  regfree(regex_t *preg)
196  {  {
197  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 189  Returns:      0 on success Line 214  Returns:      0 on success
214                various non-zero codes on failure                various non-zero codes on failure
215  */  */
216    
217  PCRE_EXPORT int  PCREPOSIX_EXP_DEFN int
218  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
219  {  {
220  const char *errorptr;  const char *errorptr;
# Line 197  int erroffset; Line 222  int erroffset;
222  int errorcode;  int errorcode;
223  int options = 0;  int options = 0;
224    
225  if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)   options |= PCRE_CASELESS;
226  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;
227  if ((cflags & REG_DOTALL) != 0) options |= PCRE_DOTALL;  if ((cflags & REG_DOTALL) != 0)  options |= PCRE_DOTALL;
228    if ((cflags & REG_NOSUB) != 0)   options |= PCRE_NO_AUTO_CAPTURE;
229    if ((cflags & REG_UTF8) != 0)    options |= PCRE_UTF8;
230    
231  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
232    &erroffset, NULL);    &erroffset, NULL);
# Line 223  substring, so we have to get and release Line 250  substring, so we have to get and release
250  the POSIX structures as was done in earlier releases when PCRE needed only 2  the POSIX structures as was done in earlier releases when PCRE needed only 2
251  ints. However, if the number of possible capturing brackets is small, use a  ints. However, if the number of possible capturing brackets is small, use a
252  block of store on the stack, to reduce the use of malloc/free. The threshold is  block of store on the stack, to reduce the use of malloc/free. The threshold is
253  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
254    
255  PCRE_EXPORT int  If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
256    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
257    the only result is yes/no/error. */
258    
259    PCREPOSIX_EXP_DEFN int
260  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
261    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
262  {  {
# Line 234  int options = 0; Line 265  int options = 0;
265  int *ovector = NULL;  int *ovector = NULL;
266  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
267  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
268    BOOL nosub =
269      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
270    
271  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
272  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
273    
274  ((regex_t *)preg)->re_erroffset = (size_t)(-1);  /* Only has meaning after compile */  ((regex_t *)preg)->re_erroffset = (size_t)(-1);  /* Only has meaning after compile */
275    
276  if (nmatch > 0)  /* When no string data is being returned, ensure that nmatch is zero.
277    Otherwise, ensure the vector for holding the return data is large enough. */
278    
279    if (nosub) nmatch = 0;
280    
281    else if (nmatch > 0)
282    {    {
283    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
284      {      {
# Line 248  if (nmatch > 0) Line 286  if (nmatch > 0)
286      }      }
287    else    else
288      {      {
289        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
290      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
291      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
292      allocated_ovector = TRUE;      allocated_ovector = TRUE;
# Line 262  if (rc == 0) rc = nmatch;    /* All capt Line 301  if (rc == 0) rc = nmatch;    /* All capt
301  if (rc >= 0)  if (rc >= 0)
302    {    {
303    size_t i;    size_t i;
304    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
305      {      {
306      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
307      pmatch[i].rm_eo = ovector[i*2+1];        {
308          pmatch[i].rm_so = ovector[i*2];
309          pmatch[i].rm_eo = ovector[i*2+1];
310          }
311        if (allocated_ovector) free(ovector);
312        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
313      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
314    return 0;    return 0;
315    }    }
316    

Legend:
Removed from v.85  
changed lines
  Added in v.199

  ViewVC Help
Powered by ViewVC 1.1.5