xref: /PHP-5.5/ext/mbstring/oniguruma/regposix.c (revision fe92d64a)
1 /**********************************************************************
2   regposix.c - Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5  * Copyright (c) 2002-2008  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
6  * All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 #define regex_t   onig_regex_t
31 #include "regint.h"
32 #undef regex_t
33 #include "onigposix.h"
34 
35 #define ONIG_C(reg)    ((onig_regex_t* )((reg)->onig))
36 #define PONIG_C(reg)   ((onig_regex_t** )(&(reg)->onig))
37 
38 /* #define ENC_STRING_LEN(enc,s,len)    len = strlen(s) */
39 #define ENC_STRING_LEN(enc,s,len) do { \
40   if (ONIGENC_MBC_MINLEN(enc) == 1) { \
41     UChar* tmps = (UChar* )(s); \
42     while (*tmps != 0) tmps++; \
43     len = tmps - (UChar* )(s); \
44   } \
45   else { \
46     len = onigenc_str_bytelen_null(enc, (UChar* )s); \
47   } \
48 } while(0)
49 
50 typedef struct {
51   int onig_err;
52   int posix_err;
53 } O2PERR;
54 
55 static int
onig2posix_error_code(int code)56 onig2posix_error_code(int code)
57 {
58   static const O2PERR o2p[] = {
59     { ONIG_MISMATCH,                                      REG_NOMATCH },
60     { ONIG_NO_SUPPORT_CONFIG,                             REG_EONIG_INTERNAL },
61     { ONIGERR_MEMORY,                                     REG_ESPACE  },
62     { ONIGERR_MATCH_STACK_LIMIT_OVER,                     REG_EONIG_INTERNAL },
63     { ONIGERR_TYPE_BUG,                                   REG_EONIG_INTERNAL },
64     { ONIGERR_PARSER_BUG,                                 REG_EONIG_INTERNAL },
65     { ONIGERR_STACK_BUG,                                  REG_EONIG_INTERNAL },
66     { ONIGERR_UNDEFINED_BYTECODE,                         REG_EONIG_INTERNAL },
67     { ONIGERR_UNEXPECTED_BYTECODE,                        REG_EONIG_INTERNAL },
68     { ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED,             REG_EONIG_BADARG },
69     { ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR, REG_EONIG_BADARG },
70     { ONIGERR_INVALID_ARGUMENT,                           REG_EONIG_BADARG },
71     { ONIGERR_END_PATTERN_AT_LEFT_BRACE,                  REG_EBRACE  },
72     { ONIGERR_END_PATTERN_AT_LEFT_BRACKET,                REG_EBRACK  },
73     { ONIGERR_EMPTY_CHAR_CLASS,                           REG_ECTYPE  },
74     { ONIGERR_PREMATURE_END_OF_CHAR_CLASS,                REG_ECTYPE  },
75     { ONIGERR_END_PATTERN_AT_ESCAPE,                      REG_EESCAPE },
76     { ONIGERR_END_PATTERN_AT_META,                        REG_EESCAPE },
77     { ONIGERR_END_PATTERN_AT_CONTROL,                     REG_EESCAPE },
78     { ONIGERR_META_CODE_SYNTAX,                           REG_BADPAT  },
79     { ONIGERR_CONTROL_CODE_SYNTAX,                        REG_BADPAT  },
80     { ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE,           REG_ECTYPE  },
81     { ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE,         REG_ECTYPE  },
82     { ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS,    REG_ECTYPE  },
83     { ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED,    REG_BADRPT  },
84     { ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID,          REG_BADRPT  },
85     { ONIGERR_NESTED_REPEAT_OPERATOR,                     REG_BADRPT  },
86     { ONIGERR_UNMATCHED_CLOSE_PARENTHESIS,                REG_EPAREN  },
87     { ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS,     REG_EPAREN  },
88     { ONIGERR_END_PATTERN_IN_GROUP,                       REG_BADPAT  },
89     { ONIGERR_UNDEFINED_GROUP_OPTION,                     REG_BADPAT  },
90     { ONIGERR_INVALID_POSIX_BRACKET_TYPE,                 REG_BADPAT  },
91     { ONIGERR_INVALID_LOOK_BEHIND_PATTERN,                REG_BADPAT  },
92     { ONIGERR_INVALID_REPEAT_RANGE_PATTERN,               REG_BADPAT  },
93     { ONIGERR_TOO_BIG_NUMBER,                             REG_BADPAT  },
94     { ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE,            REG_BADBR   },
95     { ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE,   REG_BADBR   },
96     { ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS,                  REG_ECTYPE  },
97     { ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE,        REG_ECTYPE  },
98     { ONIGERR_TOO_MANY_MULTI_BYTE_RANGES,                 REG_ECTYPE  },
99     { ONIGERR_TOO_SHORT_MULTI_BYTE_STRING,                REG_BADPAT  },
100     { ONIGERR_TOO_BIG_BACKREF_NUMBER,                     REG_ESUBREG },
101     { ONIGERR_INVALID_BACKREF,                            REG_ESUBREG },
102     { ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED,       REG_BADPAT  },
103     { ONIGERR_TOO_BIG_WIDE_CHAR_VALUE,                    REG_EONIG_BADWC },
104     { ONIGERR_TOO_LONG_WIDE_CHAR_VALUE,                   REG_EONIG_BADWC },
105     { ONIGERR_INVALID_CODE_POINT_VALUE,                   REG_EONIG_BADWC },
106     { ONIGERR_EMPTY_GROUP_NAME,                           REG_BADPAT },
107     { ONIGERR_INVALID_GROUP_NAME,                         REG_BADPAT },
108     { ONIGERR_INVALID_CHAR_IN_GROUP_NAME,                 REG_BADPAT },
109     { ONIGERR_UNDEFINED_NAME_REFERENCE,                   REG_BADPAT },
110     { ONIGERR_UNDEFINED_GROUP_REFERENCE,                  REG_BADPAT },
111     { ONIGERR_MULTIPLEX_DEFINED_NAME,                     REG_BADPAT },
112     { ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL,             REG_BADPAT },
113     { ONIGERR_NEVER_ENDING_RECURSION,                     REG_BADPAT },
114     { ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY,      REG_BADPAT },
115     { ONIGERR_INVALID_CHAR_PROPERTY_NAME,                 REG_BADPAT },
116     { ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION,         REG_EONIG_BADARG },
117     { ONIGERR_OVER_THREAD_PASS_LIMIT_COUNT,               REG_EONIG_THREAD }
118 
119   };
120 
121   int i;
122 
123   if (code >= 0) return 0;
124 
125   for (i = 0; i < (int )(sizeof(o2p) / sizeof(o2p[0])); i++) {
126     if (code == o2p[i].onig_err)
127       return o2p[i].posix_err;
128   }
129 
130   return REG_EONIG_INTERNAL;  /* but, unknown error code */
131 }
132 
133 extern int
regcomp(regex_t * reg,const char * pattern,int posix_options)134 regcomp(regex_t* reg, const char* pattern, int posix_options)
135 {
136   int r, len;
137   OnigSyntaxType* syntax = OnigDefaultSyntax;
138   OnigOptionType options;
139 
140   if ((posix_options & REG_EXTENDED) == 0)
141     syntax = ONIG_SYNTAX_POSIX_BASIC;
142 
143   options = syntax->options;
144   if ((posix_options & REG_ICASE)   != 0)
145     ONIG_OPTION_ON(options, ONIG_OPTION_IGNORECASE);
146   if ((posix_options & REG_NEWLINE) != 0) {
147     ONIG_OPTION_ON( options, ONIG_OPTION_NEGATE_SINGLELINE);
148     ONIG_OPTION_OFF(options, ONIG_OPTION_SINGLELINE);
149   }
150 
151   reg->comp_options = posix_options;
152 
153   ENC_STRING_LEN(OnigEncDefaultCharEncoding, pattern, len);
154   r = onig_new(PONIG_C(reg), (UChar* )pattern, (UChar* )(pattern + len),
155 	       options, OnigEncDefaultCharEncoding, syntax,
156 	       (OnigErrorInfo* )NULL);
157   if (r != ONIG_NORMAL) {
158     return onig2posix_error_code(r);
159   }
160 
161   reg->re_nsub = ONIG_C(reg)->num_mem;
162   return 0;
163 }
164 
165 extern int
regexec(regex_t * reg,const char * str,size_t nmatch,regmatch_t pmatch[],int posix_options)166 regexec(regex_t* reg, const char* str, size_t nmatch,
167 	regmatch_t pmatch[], int posix_options)
168 {
169   int r, i, len;
170   UChar* end;
171   regmatch_t* pm;
172   OnigOptionType options;
173 
174   options = ONIG_OPTION_POSIX_REGION;
175   if ((posix_options & REG_NOTBOL) != 0) options |= ONIG_OPTION_NOTBOL;
176   if ((posix_options & REG_NOTEOL) != 0) options |= ONIG_OPTION_NOTEOL;
177 
178   if (nmatch == 0 || (reg->comp_options & REG_NOSUB) != 0) {
179     pm = (regmatch_t* )NULL;
180     nmatch = 0;
181   }
182   else if ((int )nmatch < ONIG_C(reg)->num_mem + 1) {
183     pm = (regmatch_t* )xmalloc(sizeof(regmatch_t)
184                                * (ONIG_C(reg)->num_mem + 1));
185     if (pm == NULL)
186       return REG_ESPACE;
187   }
188   else {
189     pm = pmatch;
190   }
191 
192   ENC_STRING_LEN(ONIG_C(reg)->enc, str, len);
193   end = (UChar* )(str + len);
194   r = onig_search(ONIG_C(reg), (UChar* )str, end, (UChar* )str, end,
195 		  (OnigRegion* )pm, options);
196 
197   if (r >= 0) {
198     r = 0; /* Match */
199     if (pm != pmatch && pm != NULL) {
200       xmemcpy(pmatch, pm, sizeof(regmatch_t) * nmatch);
201     }
202   }
203   else if (r == ONIG_MISMATCH) {
204     r = REG_NOMATCH;
205     for (i = 0; i < (int )nmatch; i++)
206       pmatch[i].rm_so = pmatch[i].rm_eo = ONIG_REGION_NOTPOS;
207   }
208   else {
209     r = onig2posix_error_code(r);
210   }
211 
212   if (pm != pmatch && pm != NULL)
213     xfree(pm);
214 
215 #if 0
216   if (reg->re_nsub > nmatch - 1)
217     reg->re_nsub = (nmatch <= 1 ? 0 : nmatch - 1);
218 #endif
219 
220   return r;
221 }
222 
223 extern void
regfree(regex_t * reg)224 regfree(regex_t* reg)
225 {
226   onig_free(ONIG_C(reg));
227 }
228 
229 
230 extern void
reg_set_encoding(int mb_code)231 reg_set_encoding(int mb_code)
232 {
233   OnigEncoding enc;
234 
235   switch (mb_code) {
236   case REG_POSIX_ENCODING_ASCII:
237     enc = ONIG_ENCODING_ASCII;
238     break;
239   case REG_POSIX_ENCODING_EUC_JP:
240     enc = ONIG_ENCODING_EUC_JP;
241     break;
242   case REG_POSIX_ENCODING_SJIS:
243     enc = ONIG_ENCODING_SJIS;
244     break;
245   case REG_POSIX_ENCODING_UTF8:
246     enc = ONIG_ENCODING_UTF8;
247     break;
248   case REG_POSIX_ENCODING_UTF16_BE:
249     enc = ONIG_ENCODING_UTF16_BE;
250     break;
251   case REG_POSIX_ENCODING_UTF16_LE:
252     enc = ONIG_ENCODING_UTF16_LE;
253     break;
254 
255   default:
256     return ;
257     break;
258   }
259 
260   onigenc_set_default_encoding(enc);
261 }
262 
263 extern int
reg_name_to_group_numbers(regex_t * reg,const unsigned char * name,const unsigned char * name_end,int ** nums)264 reg_name_to_group_numbers(regex_t* reg,
265   const unsigned char* name, const unsigned char* name_end, int** nums)
266 {
267   return onig_name_to_group_numbers(ONIG_C(reg), name, name_end, nums);
268 }
269 
270 typedef struct {
271   int (*func)(const unsigned char*, const unsigned char*,int,int*,regex_t*,void*);
272   regex_t* reg;
273   void* arg;
274 } i_wrap;
275 
276 static int
i_wrapper(const UChar * name,const UChar * name_end,int ng,int * gs,onig_regex_t * reg ARG_UNUSED,void * arg)277 i_wrapper(const UChar* name, const UChar* name_end, int ng, int* gs,
278 	  onig_regex_t* reg ARG_UNUSED, void* arg)
279 {
280   i_wrap* warg = (i_wrap* )arg;
281 
282   return (*warg->func)(name, name_end, ng, gs, warg->reg, warg->arg);
283 }
284 
285 extern int
reg_foreach_name(regex_t * reg,int (* func)(const unsigned char *,const unsigned char *,int,int *,regex_t *,void *),void * arg)286 reg_foreach_name(regex_t* reg,
287  int (*func)(const unsigned char*, const unsigned char*,int,int*,regex_t*,void*),
288  void* arg)
289 {
290   i_wrap warg;
291 
292   warg.func = func;
293   warg.reg  = reg;
294   warg.arg  = arg;
295 
296   return onig_foreach_name(ONIG_C(reg), i_wrapper, &warg);
297 }
298 
299 extern int
reg_number_of_names(regex_t * reg)300 reg_number_of_names(regex_t* reg)
301 {
302   return onig_number_of_names(ONIG_C(reg));
303 }
304