/[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 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 432 by ph10, Wed Sep 2 16:02:56 2009 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) 1997-2009 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.h"
61  #include "pcre_internal.h"  #include "pcre_internal.h"
62  #include "pcreposix.h"  #include "pcreposix.h"
 #include "stdlib.h"  
   
63    
64    
65  /* Table to translate PCRE compile time error codes into POSIX error codes. */  /* Table to translate PCRE compile time error codes into POSIX error codes. */
# Line 56  static const int eint[] = { Line 70  static const int eint[] = {
70    REG_EESCAPE, /* \c at end of pattern */    REG_EESCAPE, /* \c at end of pattern */
71    REG_EESCAPE, /* unrecognized character follows \ */    REG_EESCAPE, /* unrecognized character follows \ */
72    REG_BADBR,   /* numbers out of order in {} quantifier */    REG_BADBR,   /* numbers out of order in {} quantifier */
73      /* 5 */
74    REG_BADBR,   /* number too big in {} quantifier */    REG_BADBR,   /* number too big in {} quantifier */
75    REG_EBRACK,  /* missing terminating ] for character class */    REG_EBRACK,  /* missing terminating ] for character class */
76    REG_ECTYPE,  /* invalid escape sequence in character class */    REG_ECTYPE,  /* invalid escape sequence in character class */
77    REG_ERANGE,  /* range out of order in character class */    REG_ERANGE,  /* range out of order in character class */
78    REG_BADRPT,  /* nothing to repeat */    REG_BADRPT,  /* nothing to repeat */
79      /* 10 */
80    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */
81    REG_ASSERT,  /* internal error: unexpected repeat */    REG_ASSERT,  /* internal error: unexpected repeat */
82    REG_BADPAT,  /* unrecognized character after (? */    REG_BADPAT,  /* unrecognized character after (? */
83    REG_BADPAT,  /* POSIX named classes are supported only within a class */    REG_BADPAT,  /* POSIX named classes are supported only within a class */
84    REG_EPAREN,  /* missing ) */    REG_EPAREN,  /* missing ) */
85      /* 15 */
86    REG_ESUBREG, /* reference to non-existent subpattern */    REG_ESUBREG, /* reference to non-existent subpattern */
87    REG_INVARG,  /* erroffset passed as NULL */    REG_INVARG,  /* erroffset passed as NULL */
88    REG_INVARG,  /* unknown option bit(s) set */    REG_INVARG,  /* unknown option bit(s) set */
89    REG_EPAREN,  /* missing ) after comment */    REG_EPAREN,  /* missing ) after comment */
90    REG_ESIZE,   /* parentheses nested too deeply */    REG_ESIZE,   /* parentheses nested too deeply */
91      /* 20 */
92    REG_ESIZE,   /* regular expression too large */    REG_ESIZE,   /* regular expression too large */
93    REG_ESPACE,  /* failed to get memory */    REG_ESPACE,  /* failed to get memory */
94    REG_EPAREN,  /* unmatched brackets */    REG_EPAREN,  /* unmatched parentheses */
95    REG_ASSERT,  /* internal error: code overflow */    REG_ASSERT,  /* internal error: code overflow */
96    REG_BADPAT,  /* unrecognized character after (?< */    REG_BADPAT,  /* unrecognized character after (?< */
97      /* 25 */
98    REG_BADPAT,  /* lookbehind assertion is not fixed length */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
99    REG_BADPAT,  /* malformed number after (?( */    REG_BADPAT,  /* malformed number or name after (?( */
100    REG_BADPAT,  /* conditional group containe more than two branches */    REG_BADPAT,  /* conditional group contains more than two branches */
101    REG_BADPAT,  /* assertion expected after (?( */    REG_BADPAT,  /* assertion expected after (?( */
102    REG_BADPAT,  /* (?R or (?digits must be followed by ) */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
103      /* 30 */
104    REG_ECTYPE,  /* unknown POSIX class name */    REG_ECTYPE,  /* unknown POSIX class name */
105    REG_BADPAT,  /* POSIX collating elements are not supported */    REG_BADPAT,  /* POSIX collating elements are not supported */
106    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 */
107    REG_BADPAT,  /* spare error */    REG_BADPAT,  /* spare error */
108    REG_BADPAT,  /* character value in \x{...} sequence is too large */    REG_BADPAT,  /* character value in \x{...} sequence is too large */
109      /* 35 */
110    REG_BADPAT,  /* invalid condition (?(0) */    REG_BADPAT,  /* invalid condition (?(0) */
111    REG_BADPAT,  /* \C not allowed in lookbehind assertion */    REG_BADPAT,  /* \C not allowed in lookbehind assertion */
112    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
113    REG_BADPAT,  /* number after (?C is > 255 */    REG_BADPAT,  /* number after (?C is > 255 */
114    REG_BADPAT,  /* closing ) for (?C expected */    REG_BADPAT,  /* closing ) for (?C expected */
115      /* 40 */
116    REG_BADPAT,  /* recursive call could loop indefinitely */    REG_BADPAT,  /* recursive call could loop indefinitely */
117    REG_BADPAT,  /* unrecognized character after (?P */    REG_BADPAT,  /* unrecognized character after (?P */
118    REG_BADPAT,  /* syntax error after (?P */    REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
119    REG_BADPAT,  /* two named groups have the same name */    REG_BADPAT,  /* two named subpatterns have the same name */
120    REG_BADPAT,  /* invalid UTF-8 string */    REG_BADPAT,  /* invalid UTF-8 string */
121      /* 45 */
122    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
123    REG_BADPAT,  /* malformed \P or \p sequence */    REG_BADPAT,  /* malformed \P or \p sequence */
124    REG_BADPAT   /* unknown property name after \P or \p */    REG_BADPAT,  /* unknown property name after \P or \p */
125      REG_BADPAT,  /* subpattern name is too long (maximum 32 characters) */
126      REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */
127      /* 50 */
128      REG_BADPAT,  /* repeated subpattern is too long */
129      REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */
130      REG_BADPAT,  /* internal error: overran compiling workspace */
131      REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */
132      REG_BADPAT,  /* DEFINE group contains more than one branch */
133      /* 55 */
134      REG_BADPAT,  /* repeating a DEFINE group is not allowed */
135      REG_INVARG,  /* inconsistent NEWLINE options */
136      REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */
137      REG_BADPAT,  /* a numbered reference must not be zero */
138      REG_BADPAT,  /* (*VERB) with an argument is not supported */
139      /* 60 */
140      REG_BADPAT,  /* (*VERB) not recognized */
141      REG_BADPAT,  /* number is too big */
142      REG_BADPAT,  /* subpattern name expected */
143      REG_BADPAT,  /* digit expected after (?+ */
144      REG_BADPAT   /* ] is an invalid data character in JavaScript compatibility mode */
145  };  };
146    
147  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 131  static const char *const pstring[] = { Line 174  static const char *const pstring[] = {
174  *          Translate error code to string        *  *          Translate error code to string        *
175  *************************************************/  *************************************************/
176    
177  EXPORT size_t  PCREPOSIX_EXP_DEFN size_t PCRE_CALL_CONVENTION
178  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)
179  {  {
180  const char *message, *addmessage;  const char *message, *addmessage;
# Line 166  return length + addlength; Line 209  return length + addlength;
209  *           Free store held by a regex           *  *           Free store held by a regex           *
210  *************************************************/  *************************************************/
211    
212  EXPORT void  PCREPOSIX_EXP_DEFN void PCRE_CALL_CONVENTION
213  regfree(regex_t *preg)  regfree(regex_t *preg)
214  {  {
215  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 189  Returns:      0 on success Line 232  Returns:      0 on success
232                various non-zero codes on failure                various non-zero codes on failure
233  */  */
234    
235  EXPORT int  PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
236  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
237  {  {
238  const char *errorptr;  const char *errorptr;
# Line 197  int erroffset; Line 240  int erroffset;
240  int errorcode;  int errorcode;
241  int options = 0;  int options = 0;
242    
243  if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)    options |= PCRE_CASELESS;
244  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0)  options |= PCRE_MULTILINE;
245  if ((cflags & REG_DOTALL) != 0) options |= PCRE_DOTALL;  if ((cflags & REG_DOTALL) != 0)   options |= PCRE_DOTALL;
246    if ((cflags & REG_NOSUB) != 0)    options |= PCRE_NO_AUTO_CAPTURE;
247    if ((cflags & REG_UTF8) != 0)     options |= PCRE_UTF8;
248    if ((cflags & REG_UNGREEDY) != 0) options |= PCRE_UNGREEDY;
249    
250  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
251    &erroffset, NULL);    &erroffset, NULL);
252  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
253    
254  if (preg->re_pcre == NULL) return eint[errorcode];  /* Safety: if the error code is too big for the translation vector (which
255    should not happen, but we all make mistakes), return REG_BADPAT. */
256    
257    if (preg->re_pcre == NULL)
258      {
259      return (errorcode < sizeof(eint)/sizeof(const int))?
260        eint[errorcode] : REG_BADPAT;
261      }
262    
263  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
264  return 0;  return 0;
# Line 223  substring, so we have to get and release Line 276  substring, so we have to get and release
276  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
277  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
278  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
279  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
280    
281  EXPORT int  If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
282    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
283    the only result is yes/no/error. */
284    
285    PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
286  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
287    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
288  {  {
289  int rc;  int rc, so, eo;
290  int options = 0;  int options = 0;
291  int *ovector = NULL;  int *ovector = NULL;
292  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
293  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
294    BOOL nosub =
295      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
296    
297  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
298  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
299    if ((eflags & REG_NOTEMPTY) != 0) options |= PCRE_NOTEMPTY;
300    
301  ((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 */
302    
303  if (nmatch > 0)  /* When no string data is being returned, or no vector has been passed in which
304    to put it, ensure that nmatch is zero. Otherwise, ensure the vector for holding
305    the return data is large enough. */
306    
307    if (nosub || pmatch == NULL) nmatch = 0;
308    
309    else if (nmatch > 0)
310    {    {
311    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
312      {      {
# Line 248  if (nmatch > 0) Line 314  if (nmatch > 0)
314      }      }
315    else    else
316      {      {
317        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
318      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
319      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
320      allocated_ovector = TRUE;      allocated_ovector = TRUE;
321      }      }
322    }    }
323    
324  rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string, (int)strlen(string),  /* REG_STARTEND is a BSD extension, to allow for non-NUL-terminated strings.
325    The man page from OS X says "REG_STARTEND affects only the location of the
326    string, not how it is matched". That is why the "so" value is used to bump the
327    start location rather than being passed as a PCRE "starting offset". */
328    
329    if ((eflags & REG_STARTEND) != 0)
330      {
331      so = pmatch[0].rm_so;
332      eo = pmatch[0].rm_eo;
333      }
334    else
335      {
336      so = 0;
337      eo = strlen(string);
338      }
339    
340    rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so),
341    0, options, ovector, nmatch * 3);    0, options, ovector, nmatch * 3);
342    
343  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */
# Line 262  if (rc == 0) rc = nmatch;    /* All capt Line 345  if (rc == 0) rc = nmatch;    /* All capt
345  if (rc >= 0)  if (rc >= 0)
346    {    {
347    size_t i;    size_t i;
348    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
349      {      {
350      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
351      pmatch[i].rm_eo = ovector[i*2+1];        {
352          pmatch[i].rm_so = ovector[i*2];
353          pmatch[i].rm_eo = ovector[i*2+1];
354          }
355        if (allocated_ovector) free(ovector);
356        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
357      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
358    return 0;    return 0;
359    }    }
360    

Legend:
Removed from v.77  
changed lines
  Added in v.432

  ViewVC Help
Powered by ViewVC 1.1.5