/[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 418 by ph10, Fri Apr 24 09:17:18 2009 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.  
7    
8  Written by: Philip Hazel <ph10@cam.ac.uk>                         Written by Philip Hazel
9               Copyright (c) 1997-2009 University of Cambridge
            Copyright (c) 1997-2004 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    
 #include "internal.h"  
 #include "pcreposix.h"  
 #include "stdlib.h"  
40    
41    /* This module is a wrapper that provides a POSIX API to the underlying PCRE
42    functions. */
43    
44    
45  /* Corresponding tables of PCRE error messages and POSIX error codes. */  #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49  static const char *const estring[] = {  
50    ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  ERR10,  /* Ensure that the PCREPOSIX_EXP_xxx macros are set appropriately for
51    ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20,  compiling these functions. This must come before including pcreposix.h, where
52    ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR29, ERR29, ERR30,  they are set for an application (using these functions) if they have not
53    ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, ERR40,  previously been set. */
54    ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };  
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"
62    #include "pcreposix.h"
63    
64    
65    /* Table to translate PCRE compile time error codes into POSIX error codes. */
66    
67  static const int eint[] = {  static const int eint[] = {
68    REG_EESCAPE, /* "\\ at end of pattern" */    0,           /* no error */
69    REG_EESCAPE, /* "\\c at end of pattern" */    REG_EESCAPE, /* \ at end of pattern */
70    REG_EESCAPE, /* "unrecognized character follows \\" */    REG_EESCAPE, /* \c at end of pattern */
71    REG_BADBR,   /* "numbers out of order in {} quantifier" */    REG_EESCAPE, /* unrecognized character follows \ */
72    REG_BADBR,   /* "number too big in {} quantifier" */    REG_BADBR,   /* numbers out of order in {} quantifier */
73    REG_EBRACK,  /* "missing terminating ] for character class" */    /* 5 */
74    REG_ECTYPE,  /* "invalid escape sequence in character class" */    REG_BADBR,   /* number too big in {} quantifier */
75    REG_ERANGE,  /* "range out of order in character class" */    REG_EBRACK,  /* missing terminating ] for character class */
76    REG_BADRPT,  /* "nothing to repeat" */    REG_ECTYPE,  /* invalid escape sequence in character class */
77    REG_BADRPT,  /* "operand of unlimited repeat could match the empty string" */    REG_ERANGE,  /* range out of order in character class */
78    REG_ASSERT,  /* "internal error: unexpected repeat" */    REG_BADRPT,  /* nothing to repeat */
79    REG_BADPAT,  /* "unrecognized character after (?" */    /* 10 */
80    REG_BADPAT,  /* "POSIX named classes are supported only within a class" */    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */
81    REG_EPAREN,  /* "missing )" */    REG_ASSERT,  /* internal error: unexpected repeat */
82    REG_ESUBREG, /* "reference to non-existent subpattern" */    REG_BADPAT,  /* unrecognized character after (? */
83    REG_INVARG,  /* "erroffset passed as NULL" */    REG_BADPAT,  /* POSIX named classes are supported only within a class */
84    REG_INVARG,  /* "unknown option bit(s) set" */    REG_EPAREN,  /* missing ) */
85    REG_EPAREN,  /* "missing ) after comment" */    /* 15 */
86    REG_ESIZE,   /* "parentheses nested too deeply" */    REG_ESUBREG, /* reference to non-existent subpattern */
87    REG_ESIZE,   /* "regular expression too large" */    REG_INVARG,  /* erroffset passed as NULL */
88    REG_ESPACE,  /* "failed to get memory" */    REG_INVARG,  /* unknown option bit(s) set */
89    REG_EPAREN,  /* "unmatched brackets" */    REG_EPAREN,  /* missing ) after comment */
90    REG_ASSERT,  /* "internal error: code overflow" */    REG_ESIZE,   /* parentheses nested too deeply */
91    REG_BADPAT,  /* "unrecognized character after (?<" */    /* 20 */
92    REG_BADPAT,  /* "lookbehind assertion is not fixed length" */    REG_ESIZE,   /* regular expression too large */
93    REG_BADPAT,  /* "malformed number after (?(" */    REG_ESPACE,  /* failed to get memory */
94    REG_BADPAT,  /* "conditional group containe more than two branches" */    REG_EPAREN,  /* unmatched parentheses */
95    REG_BADPAT,  /* "assertion expected after (?(" */    REG_ASSERT,  /* internal error: code overflow */
96    REG_BADPAT,  /* "(?R or (?digits must be followed by )" */    REG_BADPAT,  /* unrecognized character after (?< */
97    REG_ECTYPE,  /* "unknown POSIX class name" */    /* 25 */
98    REG_BADPAT,  /* "POSIX collating elements are not supported" */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
99    REG_INVARG,  /* "this version of PCRE is not compiled with PCRE_UTF8 support" */    REG_BADPAT,  /* malformed number or name after (?( */
100    REG_BADPAT,  /* "spare error" */    REG_BADPAT,  /* conditional group contains more than two branches */
101    REG_BADPAT,  /* "character value in \x{...} sequence is too large" */    REG_BADPAT,  /* assertion expected after (?( */
102    REG_BADPAT,  /* "invalid condition (?(0)" */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
103    REG_BADPAT,  /* "\\C not allowed in lookbehind assertion" */    /* 30 */
104    REG_EESCAPE, /* "PCRE does not support \\L, \\l, \\N, \\U, or \\u" */    REG_ECTYPE,  /* unknown POSIX class name */
105    REG_BADPAT,  /* "number after (?C is > 255" */    REG_BADPAT,  /* POSIX collating elements are not supported */
106    REG_BADPAT,  /* "closing ) for (?C expected" */    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */
107    REG_BADPAT,  /* "recursive call could loop indefinitely" */    REG_BADPAT,  /* spare error */
108    REG_BADPAT,  /* "unrecognized character after (?P" */    REG_BADPAT,  /* character value in \x{...} sequence is too large */
109    REG_BADPAT,  /* "syntax error after (?P" */    /* 35 */
110    REG_BADPAT,  /* "two named groups have the same name" */    REG_BADPAT,  /* invalid condition (?(0) */
111    REG_BADPAT,  /* "invalid UTF-8 string" */    REG_BADPAT,  /* \C not allowed in lookbehind assertion */
112    REG_BADPAT,  /* "support for \\P, \\p, and \\X has not been compiled" */    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
113    REG_BADPAT,  /* "malformed \\P or \\p sequence" */    REG_BADPAT,  /* number after (?C is > 255 */
114    REG_BADPAT   /* "unknown property name after \\P or \\p" */    REG_BADPAT,  /* closing ) for (?C expected */
115      /* 40 */
116      REG_BADPAT,  /* recursive call could loop indefinitely */
117      REG_BADPAT,  /* unrecognized character after (?P */
118      REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
119      REG_BADPAT,  /* two named subpatterns have the same name */
120      REG_BADPAT,  /* invalid UTF-8 string */
121      /* 45 */
122      REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
123      REG_BADPAT,  /* malformed \P or \p sequence */
124      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 135  static const char *const pstring[] = { Line 171  static const char *const pstring[] = {
171    
172    
173  /*************************************************  /*************************************************
 *          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;  
 }  
   
   
   
 /*************************************************  
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 191  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 214  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;
239  int erroffset;  int erroffset;
240    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;
246    if ((cflags & REG_NOSUB) != 0)   options |= PCRE_NO_AUTO_CAPTURE;
247    if ((cflags & REG_UTF8) != 0)    options |= PCRE_UTF8;
248    
249  preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset, NULL);  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
250      &erroffset, NULL);
251  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
252    
253  if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);  /* Safety: if the error code is too big for the translation vector (which
254    should not happen, but we all make mistakes), return REG_BADPAT. */
255    
256    if (preg->re_pcre == NULL)
257      {
258      return (errorcode < sizeof(eint)/sizeof(const int))?
259        eint[errorcode] : REG_BADPAT;
260      }
261    
262  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
263  return 0;  return 0;
# Line 245  substring, so we have to get and release Line 275  substring, so we have to get and release
275  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
276  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
277  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
278  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
279    
280  EXPORT int  If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
281    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
282    the only result is yes/no/error. */
283    
284    PCREPOSIX_EXP_DEFN int PCRE_CALL_CONVENTION
285  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
286    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
287  {  {
288  int rc;  int rc, so, eo;
289  int options = 0;  int options = 0;
290  int *ovector = NULL;  int *ovector = NULL;
291  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
292  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
293    BOOL nosub =
294      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
295    
296  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
297  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
298    if ((eflags & REG_NOTEMPTY) != 0) options |= PCRE_NOTEMPTY;
299    
300  ((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 */
301    
302  if (nmatch > 0)  /* When no string data is being returned, ensure that nmatch is zero.
303    Otherwise, ensure the vector for holding the return data is large enough. */
304    
305    if (nosub) nmatch = 0;
306    
307    else if (nmatch > 0)
308    {    {
309    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
310      {      {
# Line 270  if (nmatch > 0) Line 312  if (nmatch > 0)
312      }      }
313    else    else
314      {      {
315        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
316      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
317      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
318      allocated_ovector = TRUE;      allocated_ovector = TRUE;
319      }      }
320    }    }
321    
322  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.
323    The man page from OS X says "REG_STARTEND affects only the location of the
324    string, not how it is matched". That is why the "so" value is used to bump the
325    start location rather than being passed as a PCRE "starting offset". */
326    
327    if ((eflags & REG_STARTEND) != 0)
328      {
329      so = pmatch[0].rm_so;
330      eo = pmatch[0].rm_eo;
331      }
332    else
333      {
334      so = 0;
335      eo = strlen(string);
336      }
337    
338    rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so),
339    0, options, ovector, nmatch * 3);    0, options, ovector, nmatch * 3);
340    
341  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */  if (rc == 0) rc = nmatch;    /* All captured slots were filled in */
# Line 284  if (rc == 0) rc = nmatch;    /* All capt Line 343  if (rc == 0) rc = nmatch;    /* All capt
343  if (rc >= 0)  if (rc >= 0)
344    {    {
345    size_t i;    size_t i;
346    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
347      {      {
348      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
349      pmatch[i].rm_eo = ovector[i*2+1];        {
350          pmatch[i].rm_so = ovector[i*2];
351          pmatch[i].rm_eo = ovector[i*2+1];
352          }
353        if (allocated_ovector) free(ovector);
354        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
355      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
356    return 0;    return 0;
357    }    }
358    

Legend:
Removed from v.75  
changed lines
  Added in v.418

  ViewVC Help
Powered by ViewVC 1.1.5