/[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 345 by ph10, Mon Apr 28 15:10:02 2008 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) 1997-2008 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    
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  /* Corresponding tables of PCRE error messages and POSIX error codes. */  #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  static const char *const estring[] = {  
65    ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  ERR10,  /* Table to translate PCRE compile time error codes into POSIX error codes. */
   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 };  
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" */    REG_BADBR,   /* number too big in {} quantifier */
74    REG_ECTYPE,  /* "invalid escape sequence in character class" */    REG_EBRACK,  /* missing terminating ] for character class */
75    REG_ERANGE,  /* "range out of order in character class" */    REG_ECTYPE,  /* invalid escape sequence in character class */
76    REG_BADRPT,  /* "nothing to repeat" */    REG_ERANGE,  /* range out of order in character class */
77    REG_BADRPT,  /* "operand of unlimited repeat could match the empty string" */    REG_BADRPT,  /* nothing to repeat */
78    REG_ASSERT,  /* "internal error: unexpected repeat" */    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */
79    REG_BADPAT,  /* "unrecognized character after (?" */    REG_ASSERT,  /* internal error: unexpected repeat */
80    REG_BADPAT,  /* "POSIX named classes are supported only within a class" */    REG_BADPAT,  /* unrecognized character after (? */
81    REG_EPAREN,  /* "missing )" */    REG_BADPAT,  /* POSIX named classes are supported only within a class */
82    REG_ESUBREG, /* "reference to non-existent subpattern" */    REG_EPAREN,  /* missing ) */
83    REG_INVARG,  /* "erroffset passed as NULL" */    REG_ESUBREG, /* reference to non-existent subpattern */
84    REG_INVARG,  /* "unknown option bit(s) set" */    REG_INVARG,  /* erroffset passed as NULL */
85    REG_EPAREN,  /* "missing ) after comment" */    REG_INVARG,  /* unknown option bit(s) set */
86    REG_ESIZE,   /* "parentheses nested too deeply" */    REG_EPAREN,  /* missing ) after comment */
87    REG_ESIZE,   /* "regular expression too large" */    REG_ESIZE,   /* parentheses nested too deeply */
88    REG_ESPACE,  /* "failed to get memory" */    REG_ESIZE,   /* regular expression too large */
89    REG_EPAREN,  /* "unmatched brackets" */    REG_ESPACE,  /* failed to get memory */
90    REG_ASSERT,  /* "internal error: code overflow" */    REG_EPAREN,  /* unmatched brackets */
91    REG_BADPAT,  /* "unrecognized character after (?<" */    REG_ASSERT,  /* internal error: code overflow */
92    REG_BADPAT,  /* "lookbehind assertion is not fixed length" */    REG_BADPAT,  /* unrecognized character after (?< */
93    REG_BADPAT,  /* "malformed number after (?(" */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
94    REG_BADPAT,  /* "conditional group containe more than two branches" */    REG_BADPAT,  /* malformed number or name after (?( */
95    REG_BADPAT,  /* "assertion expected after (?(" */    REG_BADPAT,  /* conditional group contains more than two branches */
96    REG_BADPAT,  /* "(?R or (?digits must be followed by )" */    REG_BADPAT,  /* assertion expected after (?( */
97    REG_ECTYPE,  /* "unknown POSIX class name" */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
98    REG_BADPAT,  /* "POSIX collating elements are not supported" */    REG_ECTYPE,  /* unknown POSIX class name */
99    REG_INVARG,  /* "this version of PCRE is not compiled with PCRE_UTF8 support" */    REG_BADPAT,  /* POSIX collating elements are not supported */
100    REG_BADPAT,  /* "spare error" */    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */
101    REG_BADPAT,  /* "character value in \x{...} sequence is too large" */    REG_BADPAT,  /* spare error */
102    REG_BADPAT,  /* "invalid condition (?(0)" */    REG_BADPAT,  /* character value in \x{...} sequence is too large */
103    REG_BADPAT,  /* "\\C not allowed in lookbehind assertion" */    REG_BADPAT,  /* invalid condition (?(0) */
104    REG_EESCAPE, /* "PCRE does not support \\L, \\l, \\N, \\U, or \\u" */    REG_BADPAT,  /* \C not allowed in lookbehind assertion */
105    REG_BADPAT,  /* "number after (?C is > 255" */    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
106    REG_BADPAT,  /* "closing ) for (?C expected" */    REG_BADPAT,  /* number after (?C is > 255 */
107    REG_BADPAT,  /* "recursive call could loop indefinitely" */    REG_BADPAT,  /* closing ) for (?C expected */
108    REG_BADPAT,  /* "unrecognized character after (?P" */    REG_BADPAT,  /* recursive call could loop indefinitely */
109    REG_BADPAT,  /* "syntax error after (?P" */    REG_BADPAT,  /* unrecognized character after (?P */
110    REG_BADPAT,  /* "two named groups have the same name" */    REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
111    REG_BADPAT,  /* "invalid UTF-8 string" */    REG_BADPAT,  /* two named subpatterns have the same name */
112    REG_BADPAT,  /* "support for \\P, \\p, and \\X has not been compiled" */    REG_BADPAT,  /* invalid UTF-8 string */
113    REG_BADPAT,  /* "malformed \\P or \\p sequence" */    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
114    REG_BADPAT   /* "unknown property name after \\P or \\p" */    REG_BADPAT,  /* malformed \P or \p sequence */
115      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      REG_BADPAT,  /* number is too big */
128      REG_BADPAT,  /* subpattern name expected */
129      REG_BADPAT,  /* digit expected after (?+ */
130      REG_BADPAT   /* ] is an invalid data character in JavaScript compatibility mode */
131  };  };
132    
133  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 135  static const char *const pstring[] = { Line 157  static const char *const pstring[] = {
157    
158    
159  /*************************************************  /*************************************************
 *          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;  
 }  
   
   
   
 /*************************************************  
160  *          Translate error code to string        *  *          Translate error code to string        *
161  *************************************************/  *************************************************/
162    
163  EXPORT size_t  PCREPOSIX_EXP_DEFN size_t
164  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)
165  {  {
166  const char *message, *addmessage;  const char *message, *addmessage;
# Line 191  return length + addlength; Line 195  return length + addlength;
195  *           Free store held by a regex           *  *           Free store held by a regex           *
196  *************************************************/  *************************************************/
197    
198  EXPORT void  PCREPOSIX_EXP_DEFN void
199  regfree(regex_t *preg)  regfree(regex_t *preg)
200  {  {
201  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 214  Returns:      0 on success Line 218  Returns:      0 on success
218                various non-zero codes on failure                various non-zero codes on failure
219  */  */
220    
221  EXPORT int  PCREPOSIX_EXP_DEFN int
222  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
223  {  {
224  const char *errorptr;  const char *errorptr;
225  int erroffset;  int erroffset;
226    int errorcode;
227  int options = 0;  int options = 0;
228    
229  if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)   options |= PCRE_CASELESS;
230  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;
231    if ((cflags & REG_DOTALL) != 0)  options |= PCRE_DOTALL;
232    if ((cflags & REG_NOSUB) != 0)   options |= PCRE_NO_AUTO_CAPTURE;
233    if ((cflags & REG_UTF8) != 0)    options |= PCRE_UTF8;
234    
235  preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset, NULL);  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
236      &erroffset, NULL);
237  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
238    
239  if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);  if (preg->re_pcre == NULL) return eint[errorcode];
240    
241  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
242  return 0;  return 0;
# Line 245  substring, so we have to get and release Line 254  substring, so we have to get and release
254  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
255  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
256  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
257  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
258    
259    If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
260    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
261    the only result is yes/no/error. */
262    
263  EXPORT int  PCREPOSIX_EXP_DEFN int
264  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
265    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
266  {  {
267  int rc;  int rc, so, eo;
268  int options = 0;  int options = 0;
269  int *ovector = NULL;  int *ovector = NULL;
270  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
271  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
272    BOOL nosub =
273      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
274    
275  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
276  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
277    
278  ((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 */
279    
280  if (nmatch > 0)  /* When no string data is being returned, ensure that nmatch is zero.
281    Otherwise, ensure the vector for holding the return data is large enough. */
282    
283    if (nosub) nmatch = 0;
284    
285    else if (nmatch > 0)
286    {    {
287    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
288      {      {
# Line 270  if (nmatch > 0) Line 290  if (nmatch > 0)
290      }      }
291    else    else
292      {      {
293        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
294      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
295      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
296      allocated_ovector = TRUE;      allocated_ovector = TRUE;
297      }      }
298    }    }
299    
300  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.
301    The man page from OS X says "REG_STARTEND affects only the location of the
302    string, not how it is matched". That is why the "so" value is used to bump the
303    start location rather than being passed as a PCRE "starting offset". */
304    
305    if ((eflags & REG_STARTEND) != 0)
306      {
307      so = pmatch[0].rm_so;
308      eo = pmatch[0].rm_eo;
309      }
310    else
311      {
312      so = 0;
313      eo = strlen(string);
314      }
315    
316    rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so),
317    0, options, ovector, nmatch * 3);    0, options, ovector, nmatch * 3);
318    
319  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 321  if (rc == 0) rc = nmatch;    /* All capt
321  if (rc >= 0)  if (rc >= 0)
322    {    {
323    size_t i;    size_t i;
324    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
325      {      {
326      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
327      pmatch[i].rm_eo = ovector[i*2+1];        {
328          pmatch[i].rm_so = ovector[i*2];
329          pmatch[i].rm_eo = ovector[i*2+1];
330          }
331        if (allocated_ovector) free(ovector);
332        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
333      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
334    return 0;    return 0;
335    }    }
336    

Legend:
Removed from v.75  
changed lines
  Added in v.345

  ViewVC Help
Powered by ViewVC 1.1.5