/[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 75 by nigel, Sat Feb 24 21:40:37 2007 UTC revision 521 by ph10, Sat May 22 19:03:35 2010 UTC
# Line 2  Line 2 
2  *      Perl-Compatible Regular Expressions       *  *      Perl-Compatible Regular Expressions       *
3  *************************************************/  *************************************************/
4    
5  /*  /* PCRE is a library of functions to support regular expressions whose syntax
6  This is a library of functions to support regular expressions whose syntax  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. See  
 the file Tech.Notes for some information on the internals.  
   
 This module is a wrapper that provides a POSIX API to the underlying PCRE  
 functions.  
   
 Written by: Philip Hazel <ph10@cam.ac.uk>  
7    
8             Copyright (c) 1997-2004 University of Cambridge                         Written by Philip Hazel
9               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  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
40  #include "internal.h"  
41  #include "pcreposix.h"  /* This module is a wrapper that provides a POSIX API to the underlying PCRE
42  #include "stdlib.h"  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"
61    #include "pcreposix.h"
62    
 /* Corresponding tables of PCRE error messages and POSIX error codes. */  
63    
64  static const char *const estring[] = {  /* Table to translate PCRE compile time error codes into POSIX error codes. */
   ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  ERR10,  
   ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20,  
   ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR29, ERR29, ERR30,  
   ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, ERR40,  
   ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };  
65    
66  static const int eint[] = {  static const int eint[] = {
67    REG_EESCAPE, /* "\\ at end of pattern" */    0,           /* no error */
68    REG_EESCAPE, /* "\\c at end of pattern" */    REG_EESCAPE, /* \ at end of pattern */
69    REG_EESCAPE, /* "unrecognized character follows \\" */    REG_EESCAPE, /* \c at end of pattern */
70    REG_BADBR,   /* "numbers out of order in {} quantifier" */    REG_EESCAPE, /* unrecognized character follows \ */
71    REG_BADBR,   /* "number too big in {} quantifier" */    REG_BADBR,   /* numbers out of order in {} quantifier */
72    REG_EBRACK,  /* "missing terminating ] for character class" */    /* 5 */
73    REG_ECTYPE,  /* "invalid escape sequence in character class" */    REG_BADBR,   /* number too big in {} quantifier */
74    REG_ERANGE,  /* "range out of order in character class" */    REG_EBRACK,  /* missing terminating ] for character class */
75    REG_BADRPT,  /* "nothing to repeat" */    REG_ECTYPE,  /* invalid escape sequence in character class */
76    REG_BADRPT,  /* "operand of unlimited repeat could match the empty string" */    REG_ERANGE,  /* range out of order in character class */
77    REG_ASSERT,  /* "internal error: unexpected repeat" */    REG_BADRPT,  /* nothing to repeat */
78    REG_BADPAT,  /* "unrecognized character after (?" */    /* 10 */
79    REG_BADPAT,  /* "POSIX named classes are supported only within a class" */    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */
80    REG_EPAREN,  /* "missing )" */    REG_ASSERT,  /* internal error: unexpected repeat */
81    REG_ESUBREG, /* "reference to non-existent subpattern" */    REG_BADPAT,  /* unrecognized character after (? */
82    REG_INVARG,  /* "erroffset passed as NULL" */    REG_BADPAT,  /* POSIX named classes are supported only within a class */
83    REG_INVARG,  /* "unknown option bit(s) set" */    REG_EPAREN,  /* missing ) */
84    REG_EPAREN,  /* "missing ) after comment" */    /* 15 */
85    REG_ESIZE,   /* "parentheses nested too deeply" */    REG_ESUBREG, /* reference to non-existent subpattern */
86    REG_ESIZE,   /* "regular expression too large" */    REG_INVARG,  /* erroffset passed as NULL */
87    REG_ESPACE,  /* "failed to get memory" */    REG_INVARG,  /* unknown option bit(s) set */
88    REG_EPAREN,  /* "unmatched brackets" */    REG_EPAREN,  /* missing ) after comment */
89    REG_ASSERT,  /* "internal error: code overflow" */    REG_ESIZE,   /* parentheses nested too deeply */
90    REG_BADPAT,  /* "unrecognized character after (?<" */    /* 20 */
91    REG_BADPAT,  /* "lookbehind assertion is not fixed length" */    REG_ESIZE,   /* regular expression too large */
92    REG_BADPAT,  /* "malformed number after (?(" */    REG_ESPACE,  /* failed to get memory */
93    REG_BADPAT,  /* "conditional group containe more than two branches" */    REG_EPAREN,  /* unmatched parentheses */
94    REG_BADPAT,  /* "assertion expected after (?(" */    REG_ASSERT,  /* internal error: code overflow */
95    REG_BADPAT,  /* "(?R or (?digits must be followed by )" */    REG_BADPAT,  /* unrecognized character after (?< */
96    REG_ECTYPE,  /* "unknown POSIX class name" */    /* 25 */
97    REG_BADPAT,  /* "POSIX collating elements are not supported" */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
98    REG_INVARG,  /* "this version of PCRE is not compiled with PCRE_UTF8 support" */    REG_BADPAT,  /* malformed number or name after (?( */
99    REG_BADPAT,  /* "spare error" */    REG_BADPAT,  /* conditional group contains more than two branches */
100    REG_BADPAT,  /* "character value in \x{...} sequence is too large" */    REG_BADPAT,  /* assertion expected after (?( */
101    REG_BADPAT,  /* "invalid condition (?(0)" */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
102    REG_BADPAT,  /* "\\C not allowed in lookbehind assertion" */    /* 30 */
103    REG_EESCAPE, /* "PCRE does not support \\L, \\l, \\N, \\U, or \\u" */    REG_ECTYPE,  /* unknown POSIX class name */
104    REG_BADPAT,  /* "number after (?C is > 255" */    REG_BADPAT,  /* POSIX collating elements are not supported */
105    REG_BADPAT,  /* "closing ) for (?C expected" */    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */
106    REG_BADPAT,  /* "recursive call could loop indefinitely" */    REG_BADPAT,  /* spare error */
107    REG_BADPAT,  /* "unrecognized character after (?P" */    REG_BADPAT,  /* character value in \x{...} sequence is too large */
108    REG_BADPAT,  /* "syntax error after (?P" */    /* 35 */
109    REG_BADPAT,  /* "two named groups have the same name" */    REG_BADPAT,  /* invalid condition (?(0) */
110    REG_BADPAT,  /* "invalid UTF-8 string" */    REG_BADPAT,  /* \C not allowed in lookbehind assertion */
111    REG_BADPAT,  /* "support for \\P, \\p, and \\X has not been compiled" */    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
112    REG_BADPAT,  /* "malformed \\P or \\p sequence" */    REG_BADPAT,  /* number after (?C is > 255 */
113    REG_BADPAT   /* "unknown property name after \\P or \\p" */    REG_BADPAT,  /* closing ) for (?C expected */
114      /* 40 */
115      REG_BADPAT,  /* recursive call could loop indefinitely */
116      REG_BADPAT,  /* unrecognized character after (?P */
117      REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
118      REG_BADPAT,  /* two named subpatterns have the same name */
119      REG_BADPAT,  /* invalid UTF-8 string */
120      /* 45 */
121      REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
122      REG_BADPAT,  /* malformed \P or \p sequence */
123      REG_BADPAT,  /* unknown property name after \P or \p */
124      REG_BADPAT,  /* subpattern name is too long (maximum 32 characters) */
125      REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */
126      /* 50 */
127      REG_BADPAT,  /* repeated subpattern is too long */
128      REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */
129      REG_BADPAT,  /* internal error: overran compiling workspace */
130      REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */
131      REG_BADPAT,  /* DEFINE group contains more than one branch */
132      /* 55 */
133      REG_BADPAT,  /* repeating a DEFINE group is not allowed */
134      REG_INVARG,  /* inconsistent NEWLINE options */
135      REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */
136      REG_BADPAT,  /* a numbered reference must not be zero */
137      REG_BADPAT,  /* an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT) */
138      /* 60 */
139      REG_BADPAT,  /* (*VERB) not recognized */
140      REG_BADPAT,  /* number is too big */
141      REG_BADPAT,  /* subpattern name expected */
142      REG_BADPAT,  /* digit expected after (?+ */
143      REG_BADPAT,  /* ] is an invalid data character in JavaScript compatibility mode */
144      /* 65 */
145      REG_BADPAT,  /* different names for subpatterns of the same number are not allowed */
146      REG_BADPAT,  /* (*MARK) must have an argument */
147      REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UCP support */
148  };  };
149    
150  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 135  static const char *const pstring[] = { Line 174  static const char *const pstring[] = {
174    
175    
176  /*************************************************  /*************************************************
 *          Translate PCRE text code to int       *  
 *************************************************/  
   
 /* PCRE compile-time errors are given as strings defined as macros. We can just  
 look them up in a table to turn them into POSIX-style error codes. */  
   
 static int  
 pcre_posix_error_code(const char *s)  
 {  
 size_t i;  
 for (i = 0; i < sizeof(estring)/sizeof(char *); i++)  
   if (strcmp(s, estring[i]) == 0) return eint[i];  
 return REG_ASSERT;  
 }  
   
   
   
 /*************************************************  
177  *          Translate error code to string        *  *          Translate error code to string        *
178  *************************************************/  *************************************************/
179    
180  EXPORT size_t  PCREPOSIX_EXP_DEFN size_t PCRE_CALL_CONVENTION
181  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)
182  {  {
183  const char *message, *addmessage;  const char *message, *addmessage;
# Line 191  return length + addlength; Line 212  return length + addlength;
212  *           Free store held by a regex           *  *           Free store held by a regex           *
213  *************************************************/  *************************************************/
214    
215  EXPORT void  PCREPOSIX_EXP_DEFN void PCRE_CALL_CONVENTION
216  regfree(regex_t *preg)  regfree(regex_t *preg)
217  {  {
218  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 214  Returns:      0 on success Line 235  Returns:      0 on success
235                various non-zero codes on failure                various non-zero codes on failure
236  */  */
237    
238  EXPORT int  PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
239  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
240  {  {
241  const char *errorptr;  const char *errorptr;
242  int erroffset;  int erroffset;
243    int errorcode;
244  int options = 0;  int options = 0;
245    
246  if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)    options |= PCRE_CASELESS;
247  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0)  options |= PCRE_MULTILINE;
248    if ((cflags & REG_DOTALL) != 0)   options |= PCRE_DOTALL;
249    if ((cflags & REG_NOSUB) != 0)    options |= PCRE_NO_AUTO_CAPTURE;
250    if ((cflags & REG_UTF8) != 0)     options |= PCRE_UTF8;
251    if ((cflags & REG_UCP) != 0)      options |= PCRE_UCP;
252    if ((cflags & REG_UNGREEDY) != 0) options |= PCRE_UNGREEDY;
253    
254  preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset, NULL);  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
255      &erroffset, NULL);
256  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
257    
258  if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);  /* Safety: if the error code is too big for the translation vector (which
259    should not happen, but we all make mistakes), return REG_BADPAT. */
260    
261    if (preg->re_pcre == NULL)
262      {
263      return (errorcode < sizeof(eint)/sizeof(const int))?
264        eint[errorcode] : REG_BADPAT;
265      }
266    
267  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
268  return 0;  return 0;
# Line 245  substring, so we have to get and release Line 280  substring, so we have to get and release
280  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
281  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
282  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
283  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
284    
285  EXPORT int  If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
286    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
287    the only result is yes/no/error. */
288    
289    PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
290  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
291    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
292  {  {
293  int rc;  int rc, so, eo;
294  int options = 0;  int options = 0;
295  int *ovector = NULL;  int *ovector = NULL;
296  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
297  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
298    BOOL nosub =
299      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
300    
301  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
302  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
303    if ((eflags & REG_NOTEMPTY) != 0) options |= PCRE_NOTEMPTY;
304    
305  ((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 */
306    
307  if (nmatch > 0)  /* When no string data is being returned, or no vector has been passed in which
308    to put it, ensure that nmatch is zero. Otherwise, ensure the vector for holding
309    the return data is large enough. */
310    
311    if (nosub || pmatch == NULL) nmatch = 0;
312    
313    else if (nmatch > 0)
314    {    {
315    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
316      {      {
# Line 270  if (nmatch > 0) Line 318  if (nmatch > 0)
318      }      }
319    else    else
320      {      {
321        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
322      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
323      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
324      allocated_ovector = TRUE;      allocated_ovector = TRUE;
325      }      }
326    }    }
327    
328  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.
329    The man page from OS X says "REG_STARTEND affects only the location of the
330    string, not how it is matched". That is why the "so" value is used to bump the
331    start location rather than being passed as a PCRE "starting offset". */
332    
333    if ((eflags & REG_STARTEND) != 0)
334      {
335      so = pmatch[0].rm_so;
336      eo = pmatch[0].rm_eo;
337      }
338    else
339      {
340      so = 0;
341      eo = strlen(string);
342      }
343    
344    rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so),
345    0, options, ovector, nmatch * 3);    0, options, ovector, nmatch * 3);
346    
347  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */
348    
349    /* Successful match */
350    
351  if (rc >= 0)  if (rc >= 0)
352    {    {
353    size_t i;    size_t i;
354    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
355      {      {
356      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
357      pmatch[i].rm_eo = ovector[i*2+1];        {
358          pmatch[i].rm_so = ovector[i*2];
359          pmatch[i].rm_eo = ovector[i*2+1];
360          }
361        if (allocated_ovector) free(ovector);
362        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
363      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
364    return 0;    return 0;
365    }    }
366    
367  else  /* Unsuccessful match */
368    
369    if (allocated_ovector) free(ovector);
370    switch(rc)
371    {    {
372    if (allocated_ovector) free(ovector);  /* ========================================================================== */
373    switch(rc)    /* These cases are never obeyed. This is a fudge that causes a compile-time
374      {    error if the vector eint, which is indexed by compile-time error number, is
375      case PCRE_ERROR_NOMATCH: return REG_NOMATCH;    not the correct length. It seems to be the only way to do such a check at
376      case PCRE_ERROR_NULL: return REG_INVARG;    compile time, as the sizeof() operator does not work in the C preprocessor.
377      case PCRE_ERROR_BADOPTION: return REG_INVARG;    As all the PCRE_ERROR_xxx values are negative, we can use 0 and 1. */
378      case PCRE_ERROR_BADMAGIC: return REG_INVARG;  
379      case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;    case 0:
380      case PCRE_ERROR_NOMEMORY: return REG_ESPACE;    case (sizeof(eint)/sizeof(int) == ERRCOUNT):
381      case PCRE_ERROR_MATCHLIMIT: return REG_ESPACE;    return REG_ASSERT;
382      case PCRE_ERROR_BADUTF8: return REG_INVARG;  /* ========================================================================== */
383      case PCRE_ERROR_BADUTF8_OFFSET: return REG_INVARG;  
384      default: return REG_ASSERT;    case PCRE_ERROR_NOMATCH: return REG_NOMATCH;
385      }    case PCRE_ERROR_NULL: return REG_INVARG;
386      case PCRE_ERROR_BADOPTION: return REG_INVARG;
387      case PCRE_ERROR_BADMAGIC: return REG_INVARG;
388      case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT;
389      case PCRE_ERROR_NOMEMORY: return REG_ESPACE;
390      case PCRE_ERROR_MATCHLIMIT: return REG_ESPACE;
391      case PCRE_ERROR_BADUTF8: return REG_INVARG;
392      case PCRE_ERROR_BADUTF8_OFFSET: return REG_INVARG;
393      default: return REG_ASSERT;
394    }    }
395  }  }
396    

Legend:
Removed from v.75  
changed lines
  Added in v.521

  ViewVC Help
Powered by ViewVC 1.1.5