/[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 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 518 by ph10, Tue May 18 15:47:01 2010 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) 1997-2007 University of Cambridge             Copyright (c) 1997-2010 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 43  functions. */ Line 43  functions. */
43    
44    
45  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
46  #include <config.h>  #include "config.h"
47  #endif  #endif
48    
49    
# Line 57  previously been set. */ Line 57  previously been set. */
57  #  define PCREPOSIX_EXP_DEFN __declspec(dllexport)  #  define PCREPOSIX_EXP_DEFN __declspec(dllexport)
58  #endif  #endif
59    
60  #include <pcre.h>  #include "pcre.h"
61  #include "pcre_internal.h"  #include "pcre_internal.h"
62  #include "pcreposix.h"  #include "pcreposix.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. */
66    
67  static const int eint[] = {  static const int eint[] = {
# Line 71  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 or name after (?( */    REG_BADPAT,  /* malformed number or name after (?( */
100    REG_BADPAT,  /* conditional group contains 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 in subpattern name (missing terminator) */    REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
119    REG_BADPAT,  /* two named subpatterns 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) */    REG_BADPAT,  /* subpattern name is too long (maximum 32 characters) */
126    REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */    REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */
127      /* 50 */
128    REG_BADPAT,  /* repeated subpattern is too long */    REG_BADPAT,  /* repeated subpattern is too long */
129    REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */    REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */
130    REG_BADPAT,  /* internal error: overran compiling workspace */    REG_BADPAT,  /* internal error: overran compiling workspace */
131    REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */    REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */
132    REG_BADPAT,  /* DEFINE group contains more than one branch */    REG_BADPAT,  /* DEFINE group contains more than one branch */
133      /* 55 */
134    REG_BADPAT,  /* repeating a DEFINE group is not allowed */    REG_BADPAT,  /* repeating a DEFINE group is not allowed */
135    REG_INVARG,  /* inconsistent NEWLINE options */    REG_INVARG,  /* inconsistent NEWLINE options */
136    REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */    REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */
137    REG_BADPAT   /* (?+ or (?- must be followed by a non-zero number */    REG_BADPAT,  /* a numbered reference must not be zero */
138      REG_BADPAT,  /* an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT) */
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      /* 65 */
146      REG_BADPAT,  /* different names for subpatterns of the same number are not allowed */
147      REG_BADPAT,  /* (*MARK) must have an argument */
148      REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UCP support */
149  };  };
150    
151  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 157  static const char *const pstring[] = { Line 178  static const char *const pstring[] = {
178  *          Translate error code to string        *  *          Translate error code to string        *
179  *************************************************/  *************************************************/
180    
181  PCREPOSIX_EXP_DEFN size_t  PCREPOSIX_EXP_DEFN size_t PCRE_CALL_CONVENTION
182  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)
183  {  {
184  const char *message, *addmessage;  const char *message, *addmessage;
# Line 192  return length + addlength; Line 213  return length + addlength;
213  *           Free store held by a regex           *  *           Free store held by a regex           *
214  *************************************************/  *************************************************/
215    
216  PCREPOSIX_EXP_DEFN void  PCREPOSIX_EXP_DEFN void PCRE_CALL_CONVENTION
217  regfree(regex_t *preg)  regfree(regex_t *preg)
218  {  {
219  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 215  Returns: 0 on success Line 236  Returns: 0 on success
236                various non-zero codes on failure                various non-zero codes on failure
237  */  */
238    
239  PCREPOSIX_EXP_DEFN int  PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
240  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
241  {  {
242  const char *errorptr;  const char *errorptr;
# Line 223  int erroffset; Line 244  int erroffset;
244  int errorcode;  int errorcode;
245  int options = 0;  int options = 0;
246    
247  if ((cflags & REG_ICASE) != 0)   options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)    options |= PCRE_CASELESS;
248  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0)  options |= PCRE_MULTILINE;
249  if ((cflags & REG_DOTALL) != 0)  options |= PCRE_DOTALL;  if ((cflags & REG_DOTALL) != 0)   options |= PCRE_DOTALL;
250  if ((cflags & REG_NOSUB) != 0)   options |= PCRE_NO_AUTO_CAPTURE;  if ((cflags & REG_NOSUB) != 0)    options |= PCRE_NO_AUTO_CAPTURE;
251  if ((cflags & REG_UTF8) != 0)    options |= PCRE_UTF8;  if ((cflags & REG_UTF8) != 0)     options |= PCRE_UTF8;
252    if ((cflags & REG_UCP) != 0)      options |= PCRE_UCP;
253    if ((cflags & REG_UNGREEDY) != 0) options |= PCRE_UNGREEDY;
254    
255  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
256    &erroffset, NULL);    &erroffset, NULL);
257  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
258    
259  if (preg->re_pcre == NULL) return eint[errorcode];  /* Safety: if the error code is too big for the translation vector (which
260    should not happen, but we all make mistakes), return REG_BADPAT. */
261    
262    if (preg->re_pcre == NULL)
263      {
264      return (errorcode < sizeof(eint)/sizeof(const int))?
265        eint[errorcode] : REG_BADPAT;
266      }
267    
268  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
269  return 0;  return 0;
# Line 257  If REG_NOSUB was specified at compile ti Line 287  If REG_NOSUB was specified at compile ti
287  be set. When this is the case, the nmatch and pmatch arguments are ignored, and  be set. When this is the case, the nmatch and pmatch arguments are ignored, and
288  the only result is yes/no/error. */  the only result is yes/no/error. */
289    
290  PCREPOSIX_EXP_DEFN int  PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
291  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
292    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
293  {  {
294  int rc;  int rc, so, eo;
295  int options = 0;  int options = 0;
296  int *ovector = NULL;  int *ovector = NULL;
297  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
# Line 271  BOOL nosub = Line 301  BOOL nosub =
301    
302  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
303  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
304    if ((eflags & REG_NOTEMPTY) != 0) options |= PCRE_NOTEMPTY;
305    
306  ((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 */
307    
308  /* When no string data is being returned, ensure that nmatch is zero.  /* When no string data is being returned, or no vector has been passed in which
309  Otherwise, ensure the vector for holding the return data is large enough. */  to put it, ensure that nmatch is zero. Otherwise, ensure the vector for holding
310    the return data is large enough. */
311    
312  if (nosub) nmatch = 0;  if (nosub || pmatch == NULL) nmatch = 0;
313    
314  else if (nmatch > 0)  else if (nmatch > 0)
315    {    {
# Line 294  else if (nmatch > 0) Line 326  else if (nmatch > 0)
326      }      }
327    }    }
328    
329  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.
330    The man page from OS X says "REG_STARTEND affects only the location of the
331    string, not how it is matched". That is why the "so" value is used to bump the
332    start location rather than being passed as a PCRE "starting offset". */
333    
334    if ((eflags & REG_STARTEND) != 0)
335      {
336      so = pmatch[0].rm_so;
337      eo = pmatch[0].rm_eo;
338      }
339    else
340      {
341      so = 0;
342      eo = strlen(string);
343      }
344    
345    rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so),
346    0, options, ovector, nmatch * 3);    0, options, ovector, nmatch * 3);
347    
348  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */
349    
350    /* Successful match */
351    
352  if (rc >= 0)  if (rc >= 0)
353    {    {
354    size_t i;    size_t i;
# Line 315  if (rc >= 0) Line 365  if (rc >= 0)
365    return 0;    return 0;
366    }    }
367    
368  else  /* Unsuccessful match */
369    
370    if (allocated_ovector) free(ovector);
371    switch(rc)
372    {    {
373    if (allocated_ovector) free(ovector);  /* ========================================================================== */
374    switch(rc)    /* These cases are never obeyed. This is a fudge that causes a compile-time
375      {    error if the vector eint, which is indexed by compile-time error number, is
376      case PCRE_ERROR_NOMATCH: return REG_NOMATCH;    not the correct length. It seems to be the only way to do such a check at
377      case PCRE_ERROR_NULL: return REG_INVARG;    compile time, as the sizeof() operator does not work in the C preprocessor.
378      case PCRE_ERROR_BADOPTION: return REG_INVARG;    As all the PCRE_ERROR_xxx values are negative, we can use 0 and 1. */
379      case PCRE_ERROR_BADMAGIC: return REG_INVARG;  
380      case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;    case 0:
381      case PCRE_ERROR_NOMEMORY: return REG_ESPACE;    case (sizeof(eint)/sizeof(int) == ERRCOUNT):
382      case PCRE_ERROR_MATCHLIMIT: return REG_ESPACE;    return REG_ASSERT;
383      case PCRE_ERROR_BADUTF8: return REG_INVARG;  /* ========================================================================== */
384      case PCRE_ERROR_BADUTF8_OFFSET: return REG_INVARG;  
385      default: return REG_ASSERT;    case PCRE_ERROR_NOMATCH: return REG_NOMATCH;
386      }    case PCRE_ERROR_NULL: return REG_INVARG;
387      case PCRE_ERROR_BADOPTION: return REG_INVARG;
388      case PCRE_ERROR_BADMAGIC: return REG_INVARG;
389      case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;
390      case PCRE_ERROR_NOMEMORY: return REG_ESPACE;
391      case PCRE_ERROR_MATCHLIMIT: return REG_ESPACE;
392      case PCRE_ERROR_BADUTF8: return REG_INVARG;
393      case PCRE_ERROR_BADUTF8_OFFSET: return REG_INVARG;
394      default: return REG_ASSERT;
395    }    }
396  }  }
397    

Legend:
Removed from v.200  
changed lines
  Added in v.518

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12