/[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 220 by ph10, Thu Aug 16 13:29:39 2007 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) 1997-2007 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    #ifdef HAVE_CONFIG_H
46    #include <config.h>
47    #endif
48    
49    
50  /* Corresponding tables of PCRE error messages and POSIX error codes. */  /* 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  static const char *const estring[] = {  #if defined(_WIN32) && !defined(PCRE_STATIC)
56    ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  ERR10,  #  define PCREPOSIX_EXP_DECL extern __declspec(dllexport)
57    ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20,  #  define PCREPOSIX_EXP_DEFN __declspec(dllexport)
58    ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR29, ERR29, ERR30,  #endif
59    ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, ERR40,  
60    ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };  #include <pcre.h>
61    #include "pcre_internal.h"
62    #include "pcreposix.h"
63    
64    
65    
66    /* Table to translate PCRE compile time error codes into POSIX error codes. */
67    
68  static const int eint[] = {  static const int eint[] = {
69    REG_EESCAPE, /* "\\ at end of pattern" */    0,           /* no error */
70    REG_EESCAPE, /* "\\c at end of pattern" */    REG_EESCAPE, /* \ at end of pattern */
71    REG_EESCAPE, /* "unrecognized character follows \\" */    REG_EESCAPE, /* \c at end of pattern */
72    REG_BADBR,   /* "numbers out of order in {} quantifier" */    REG_EESCAPE, /* unrecognized character follows \ */
73    REG_BADBR,   /* "number too big in {} quantifier" */    REG_BADBR,   /* numbers out of order in {} quantifier */
74    REG_EBRACK,  /* "missing terminating ] for character class" */    REG_BADBR,   /* number too big in {} quantifier */
75    REG_ECTYPE,  /* "invalid escape sequence in character class" */    REG_EBRACK,  /* missing terminating ] for character class */
76    REG_ERANGE,  /* "range out of order in character class" */    REG_ECTYPE,  /* invalid escape sequence in character class */
77    REG_BADRPT,  /* "nothing to repeat" */    REG_ERANGE,  /* range out of order in character class */
78    REG_BADRPT,  /* "operand of unlimited repeat could match the empty string" */    REG_BADRPT,  /* nothing to repeat */
79    REG_ASSERT,  /* "internal error: unexpected repeat" */    REG_BADRPT,  /* operand of unlimited repeat could match the empty string */
80    REG_BADPAT,  /* "unrecognized character after (?" */    REG_ASSERT,  /* internal error: unexpected repeat */
81    REG_BADPAT,  /* "POSIX named classes are supported only within a class" */    REG_BADPAT,  /* unrecognized character after (? */
82    REG_EPAREN,  /* "missing )" */    REG_BADPAT,  /* POSIX named classes are supported only within a class */
83    REG_ESUBREG, /* "reference to non-existent subpattern" */    REG_EPAREN,  /* missing ) */
84    REG_INVARG,  /* "erroffset passed as NULL" */    REG_ESUBREG, /* reference to non-existent subpattern */
85    REG_INVARG,  /* "unknown option bit(s) set" */    REG_INVARG,  /* erroffset passed as NULL */
86    REG_EPAREN,  /* "missing ) after comment" */    REG_INVARG,  /* unknown option bit(s) set */
87    REG_ESIZE,   /* "parentheses nested too deeply" */    REG_EPAREN,  /* missing ) after comment */
88    REG_ESIZE,   /* "regular expression too large" */    REG_ESIZE,   /* parentheses nested too deeply */
89    REG_ESPACE,  /* "failed to get memory" */    REG_ESIZE,   /* regular expression too large */
90    REG_EPAREN,  /* "unmatched brackets" */    REG_ESPACE,  /* failed to get memory */
91    REG_ASSERT,  /* "internal error: code overflow" */    REG_EPAREN,  /* unmatched brackets */
92    REG_BADPAT,  /* "unrecognized character after (?<" */    REG_ASSERT,  /* internal error: code overflow */
93    REG_BADPAT,  /* "lookbehind assertion is not fixed length" */    REG_BADPAT,  /* unrecognized character after (?< */
94    REG_BADPAT,  /* "malformed number after (?(" */    REG_BADPAT,  /* lookbehind assertion is not fixed length */
95    REG_BADPAT,  /* "conditional group containe more than two branches" */    REG_BADPAT,  /* malformed number or name after (?( */
96    REG_BADPAT,  /* "assertion expected after (?(" */    REG_BADPAT,  /* conditional group contains more than two branches */
97    REG_BADPAT,  /* "(?R or (?digits must be followed by )" */    REG_BADPAT,  /* assertion expected after (?( */
98    REG_ECTYPE,  /* "unknown POSIX class name" */    REG_BADPAT,  /* (?R or (?[+-]digits must be followed by ) */
99    REG_BADPAT,  /* "POSIX collating elements are not supported" */    REG_ECTYPE,  /* unknown POSIX class name */
100    REG_INVARG,  /* "this version of PCRE is not compiled with PCRE_UTF8 support" */    REG_BADPAT,  /* POSIX collating elements are not supported */
101    REG_BADPAT,  /* "spare error" */    REG_INVARG,  /* this version of PCRE is not compiled with PCRE_UTF8 support */
102    REG_BADPAT,  /* "character value in \x{...} sequence is too large" */    REG_BADPAT,  /* spare error */
103    REG_BADPAT,  /* "invalid condition (?(0)" */    REG_BADPAT,  /* character value in \x{...} sequence is too large */
104    REG_BADPAT,  /* "\\C not allowed in lookbehind assertion" */    REG_BADPAT,  /* invalid condition (?(0) */
105    REG_EESCAPE, /* "PCRE does not support \\L, \\l, \\N, \\U, or \\u" */    REG_BADPAT,  /* \C not allowed in lookbehind assertion */
106    REG_BADPAT,  /* "number after (?C is > 255" */    REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
107    REG_BADPAT,  /* "closing ) for (?C expected" */    REG_BADPAT,  /* number after (?C is > 255 */
108    REG_BADPAT,  /* "recursive call could loop indefinitely" */    REG_BADPAT,  /* closing ) for (?C expected */
109    REG_BADPAT,  /* "unrecognized character after (?P" */    REG_BADPAT,  /* recursive call could loop indefinitely */
110    REG_BADPAT,  /* "syntax error after (?P" */    REG_BADPAT,  /* unrecognized character after (?P */
111    REG_BADPAT,  /* "two named groups have the same name" */    REG_BADPAT,  /* syntax error in subpattern name (missing terminator) */
112    REG_BADPAT,  /* "invalid UTF-8 string" */    REG_BADPAT,  /* two named subpatterns have the same name */
113    REG_BADPAT,  /* "support for \\P, \\p, and \\X has not been compiled" */    REG_BADPAT,  /* invalid UTF-8 string */
114    REG_BADPAT,  /* "malformed \\P or \\p sequence" */    REG_BADPAT,  /* support for \P, \p, and \X has not been compiled */
115    REG_BADPAT   /* "unknown property name after \\P or \\p" */    REG_BADPAT,  /* malformed \P or \p sequence */
116      REG_BADPAT,  /* unknown property name after \P or \p */
117      REG_BADPAT,  /* subpattern name is too long (maximum 32 characters) */
118      REG_BADPAT,  /* too many named subpatterns (maximum 10,000) */
119      REG_BADPAT,  /* repeated subpattern is too long */
120      REG_BADPAT,  /* octal value is greater than \377 (not in UTF-8 mode) */
121      REG_BADPAT,  /* internal error: overran compiling workspace */
122      REG_BADPAT,  /* internal error: previously-checked referenced subpattern not found */
123      REG_BADPAT,  /* DEFINE group contains more than one branch */
124      REG_BADPAT,  /* repeating a DEFINE group is not allowed */
125      REG_INVARG,  /* inconsistent NEWLINE options */
126      REG_BADPAT,  /* \g is not followed followed by an (optionally braced) non-zero number */
127      REG_BADPAT,  /* (?+ or (?- must be followed by a non-zero number */
128      REG_BADPAT   /* number is too big */
129  };  };
130    
131  /* Table of texts corresponding to POSIX error codes */  /* Table of texts corresponding to POSIX error codes */
# Line 135  static const char *const pstring[] = { Line 155  static const char *const pstring[] = {
155    
156    
157  /*************************************************  /*************************************************
 *          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;  
 }  
   
   
   
 /*************************************************  
158  *          Translate error code to string        *  *          Translate error code to string        *
159  *************************************************/  *************************************************/
160    
161  EXPORT size_t  PCREPOSIX_EXP_DEFN size_t
162  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)
163  {  {
164  const char *message, *addmessage;  const char *message, *addmessage;
# Line 191  return length + addlength; Line 193  return length + addlength;
193  *           Free store held by a regex           *  *           Free store held by a regex           *
194  *************************************************/  *************************************************/
195    
196  EXPORT void  PCREPOSIX_EXP_DEFN void
197  regfree(regex_t *preg)  regfree(regex_t *preg)
198  {  {
199  (pcre_free)(preg->re_pcre);  (pcre_free)(preg->re_pcre);
# Line 214  Returns: 0 on success Line 216  Returns: 0 on success
216                various non-zero codes on failure                various non-zero codes on failure
217  */  */
218    
219  EXPORT int  PCREPOSIX_EXP_DEFN int
220  regcomp(regex_t *preg, const char *pattern, int cflags)  regcomp(regex_t *preg, const char *pattern, int cflags)
221  {  {
222  const char *errorptr;  const char *errorptr;
223  int erroffset;  int erroffset;
224    int errorcode;
225  int options = 0;  int options = 0;
226    
227  if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS;  if ((cflags & REG_ICASE) != 0)   options |= PCRE_CASELESS;
228  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;  if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE;
229    if ((cflags & REG_DOTALL) != 0)  options |= PCRE_DOTALL;
230    if ((cflags & REG_NOSUB) != 0)   options |= PCRE_NO_AUTO_CAPTURE;
231    if ((cflags & REG_UTF8) != 0)    options |= PCRE_UTF8;
232    
233  preg->re_pcre = pcre_compile(pattern, options, &errorptr, &erroffset, NULL);  preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr,
234      &erroffset, NULL);
235  preg->re_erroffset = erroffset;  preg->re_erroffset = erroffset;
236    
237  if (preg->re_pcre == NULL) return pcre_posix_error_code(errorptr);  if (preg->re_pcre == NULL) return eint[errorcode];
238    
239  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);  preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL);
240  return 0;  return 0;
# Line 245  substring, so we have to get and release Line 252  substring, so we have to get and release
252  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
253  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
254  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
255  in a macro that can be changed at configure time. */  in a macro that can be changed at configure time.
256    
257    If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will
258    be set. When this is the case, the nmatch and pmatch arguments are ignored, and
259    the only result is yes/no/error. */
260    
261  EXPORT int  PCREPOSIX_EXP_DEFN int
262  regexec(const regex_t *preg, const char *string, size_t nmatch,  regexec(const regex_t *preg, const char *string, size_t nmatch,
263    regmatch_t pmatch[], int eflags)    regmatch_t pmatch[], int eflags)
264  {  {
# Line 256  int options = 0; Line 267  int options = 0;
267  int *ovector = NULL;  int *ovector = NULL;
268  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];  int small_ovector[POSIX_MALLOC_THRESHOLD * 3];
269  BOOL allocated_ovector = FALSE;  BOOL allocated_ovector = FALSE;
270    BOOL nosub =
271      (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0;
272    
273  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;  if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL;
274  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;  if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL;
275    
276  ((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 */
277    
278  if (nmatch > 0)  /* When no string data is being returned, ensure that nmatch is zero.
279    Otherwise, ensure the vector for holding the return data is large enough. */
280    
281    if (nosub) nmatch = 0;
282    
283    else if (nmatch > 0)
284    {    {
285    if (nmatch <= POSIX_MALLOC_THRESHOLD)    if (nmatch <= POSIX_MALLOC_THRESHOLD)
286      {      {
# Line 270  if (nmatch > 0) Line 288  if (nmatch > 0)
288      }      }
289    else    else
290      {      {
291        if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE;
292      ovector = (int *)malloc(sizeof(int) * nmatch * 3);      ovector = (int *)malloc(sizeof(int) * nmatch * 3);
293      if (ovector == NULL) return REG_ESPACE;      if (ovector == NULL) return REG_ESPACE;
294      allocated_ovector = TRUE;      allocated_ovector = TRUE;
# Line 284  if (rc == 0) rc = nmatch; /* All capt Line 303  if (rc == 0) rc = nmatch; /* All capt
303  if (rc >= 0)  if (rc >= 0)
304    {    {
305    size_t i;    size_t i;
306    for (i = 0; i < (size_t)rc; i++)    if (!nosub)
307      {      {
308      pmatch[i].rm_so = ovector[i*2];      for (i = 0; i < (size_t)rc; i++)
309      pmatch[i].rm_eo = ovector[i*2+1];        {
310          pmatch[i].rm_so = ovector[i*2];
311          pmatch[i].rm_eo = ovector[i*2+1];
312          }
313        if (allocated_ovector) free(ovector);
314        for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;
315      }      }
   if (allocated_ovector) free(ovector);  
   for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1;  
316    return 0;    return 0;
317    }    }
318    

Legend:
Removed from v.75  
changed lines
  Added in v.220

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12