1 /**********************************************************************
2   regposix.c - Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5  * Copyright (c) 2002-2021  K.Kosako
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 
34 #include "onigposix.h"
35 
36 #undef regex_t
37 #undef regmatch_t
38 #undef regoff_t
39 #undef regcomp
40 #undef regexec
41 #undef regfree
42 #undef regerror
43 #undef reg_set_encoding
44 #undef reg_name_to_group_numbers
45 #undef reg_foreach_name
46 #undef reg_number_of_names
47 
48 #define ONIG_C(reg)    ((onig_regex_t* )((reg)->onig))
49 #define PONIG_C(reg)   ((onig_regex_t** )(&(reg)->onig))
50 
51 /* #define ENC_STRING_LEN(enc,s,len)    len = strlen(s) */
52 #define ENC_STRING_LEN(enc,s,len) do { \
53   if (ONIGENC_MBC_MINLEN(enc) == 1) { \
54     UChar* tmps = (UChar* )(s); \
55     while (*tmps != 0) tmps++; \
56     len = (int )(tmps - (UChar* )(s));\
57   } \
58   else { \
59     len = onigenc_str_bytelen_null(enc, (UChar* )s); \
60   } \
61 } while(0)
62 
63 typedef struct {
64   int onig_err;
65   int posix_err;
66 } O2PERR;
67 
68 static int
onig2posix_error_code(int code)69 onig2posix_error_code(int code)
70 {
71   static const O2PERR o2p[] = {
72     { ONIG_MISMATCH,                                      REG_NOMATCH },
73     { ONIG_NO_SUPPORT_CONFIG,                             REG_EONIG_INTERNAL },
74     { ONIG_ABORT,                                         REG_EONIG_INTERNAL },
75     { ONIGERR_MEMORY,                                     REG_ESPACE  },
76     { ONIGERR_MATCH_STACK_LIMIT_OVER,                     REG_EONIG_INTERNAL },
77     { ONIGERR_RETRY_LIMIT_IN_MATCH_OVER,                  REG_EONIG_INTERNAL },
78     { ONIGERR_RETRY_LIMIT_IN_SEARCH_OVER,                 REG_EONIG_INTERNAL },
79     { ONIGERR_SUBEXP_CALL_LIMIT_IN_SEARCH_OVER,           REG_EONIG_INTERNAL },
80     { ONIGERR_TYPE_BUG,                                   REG_EONIG_INTERNAL },
81     { ONIGERR_PARSER_BUG,                                 REG_EONIG_INTERNAL },
82     { ONIGERR_STACK_BUG,                                  REG_EONIG_INTERNAL },
83     { ONIGERR_UNDEFINED_BYTECODE,                         REG_EONIG_INTERNAL },
84     { ONIGERR_UNEXPECTED_BYTECODE,                        REG_EONIG_INTERNAL },
85     { ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED,             REG_EONIG_BADARG },
86     { ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR, REG_EONIG_BADARG },
87     { ONIGERR_FAIL_TO_INITIALIZE,                         REG_EONIG_INTERNAL },
88     { ONIGERR_INVALID_ARGUMENT,                           REG_EONIG_BADARG },
89     { ONIGERR_END_PATTERN_AT_LEFT_BRACE,                  REG_EBRACE  },
90     { ONIGERR_END_PATTERN_AT_LEFT_BRACKET,                REG_EBRACK  },
91     { ONIGERR_EMPTY_CHAR_CLASS,                           REG_ECTYPE  },
92     { ONIGERR_PREMATURE_END_OF_CHAR_CLASS,                REG_ECTYPE  },
93     { ONIGERR_END_PATTERN_AT_ESCAPE,                      REG_EESCAPE },
94     { ONIGERR_END_PATTERN_AT_META,                        REG_EESCAPE },
95     { ONIGERR_END_PATTERN_AT_CONTROL,                     REG_EESCAPE },
96     { ONIGERR_META_CODE_SYNTAX,                           REG_BADPAT  },
97     { ONIGERR_CONTROL_CODE_SYNTAX,                        REG_BADPAT  },
98     { ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE,           REG_ECTYPE  },
99     { ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE,         REG_ECTYPE  },
100     { ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS,    REG_ECTYPE  },
101     { ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED,    REG_BADRPT  },
102     { ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID,          REG_BADRPT  },
103     { ONIGERR_NESTED_REPEAT_OPERATOR,                     REG_BADRPT  },
104     { ONIGERR_UNMATCHED_CLOSE_PARENTHESIS,                REG_EPAREN  },
105     { ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS,     REG_EPAREN  },
106     { ONIGERR_END_PATTERN_IN_GROUP,                       REG_BADPAT  },
107     { ONIGERR_UNDEFINED_GROUP_OPTION,                     REG_BADPAT  },
108     { ONIGERR_INVALID_POSIX_BRACKET_TYPE,                 REG_BADPAT  },
109     { ONIGERR_INVALID_LOOK_BEHIND_PATTERN,                REG_BADPAT  },
110     { ONIGERR_INVALID_REPEAT_RANGE_PATTERN,               REG_BADPAT  },
111     { ONIGERR_TOO_BIG_NUMBER,                             REG_BADPAT  },
112     { ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE,            REG_BADBR   },
113     { ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE,   REG_BADBR   },
114     { ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS,                  REG_ECTYPE  },
115     { ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE,        REG_ECTYPE  },
116     { ONIGERR_TOO_MANY_MULTI_BYTE_RANGES,                 REG_ECTYPE  },
117     { ONIGERR_TOO_SHORT_MULTI_BYTE_STRING,                REG_BADPAT  },
118     { ONIGERR_TOO_BIG_BACKREF_NUMBER,                     REG_ESUBREG },
119     { ONIGERR_INVALID_BACKREF,                            REG_ESUBREG },
120     { ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED,       REG_BADPAT  },
121     { ONIGERR_TOO_BIG_WIDE_CHAR_VALUE,                    REG_EONIG_BADWC },
122     { ONIGERR_TOO_LONG_WIDE_CHAR_VALUE,                   REG_EONIG_BADWC },
123     { ONIGERR_UNDEFINED_OPERATOR,                         REG_BADPAT  },
124     { ONIGERR_INVALID_CODE_POINT_VALUE,                   REG_EONIG_BADWC },
125     { ONIGERR_EMPTY_GROUP_NAME,                           REG_BADPAT },
126     { ONIGERR_INVALID_GROUP_NAME,                         REG_BADPAT },
127     { ONIGERR_INVALID_CHAR_IN_GROUP_NAME,                 REG_BADPAT },
128     { ONIGERR_UNDEFINED_NAME_REFERENCE,                   REG_BADPAT },
129     { ONIGERR_UNDEFINED_GROUP_REFERENCE,                  REG_BADPAT },
130     { ONIGERR_MULTIPLEX_DEFINED_NAME,                     REG_BADPAT },
131     { ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL,             REG_BADPAT },
132     { ONIGERR_NEVER_ENDING_RECURSION,                     REG_BADPAT },
133     { ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY,      REG_BADPAT },
134     { ONIGERR_INVALID_CHAR_PROPERTY_NAME,                 REG_BADPAT },
135     { ONIGERR_INVALID_IF_ELSE_SYNTAX,                     REG_BADPAT },
136     { ONIGERR_INVALID_ABSENT_GROUP_PATTERN,               REG_BADPAT },
137     { ONIGERR_INVALID_ABSENT_GROUP_GENERATOR_PATTERN,     REG_BADPAT },
138     { ONIGERR_INVALID_CALLOUT_PATTERN,                    REG_BADPAT },
139     { ONIGERR_INVALID_CALLOUT_NAME,                       REG_BADPAT },
140     { ONIGERR_UNDEFINED_CALLOUT_NAME,                     REG_BADPAT },
141     { ONIGERR_INVALID_CALLOUT_BODY,                       REG_BADPAT },
142     { ONIGERR_INVALID_CALLOUT_TAG_NAME,                   REG_BADPAT },
143     { ONIGERR_INVALID_CALLOUT_ARG,                        REG_BADPAT },
144     { ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION,         REG_EONIG_BADARG },
145     { ONIGERR_VERY_INEFFICIENT_PATTERN,                   REG_BADPAT },
146     { ONIGERR_LIBRARY_IS_NOT_INITIALIZED,                 REG_EONIG_INTERNAL }
147   };
148 
149   int i;
150 
151   if (code >= 0) return 0;
152 
153   for (i = 0; i < (int )(sizeof(o2p) / sizeof(o2p[0])); i++) {
154     if (code == o2p[i].onig_err)
155       return o2p[i].posix_err;
156   }
157 
158   return REG_EONIG_INTERNAL;  /* but, unknown error code */
159 }
160 
161 extern int
onig_posix_regcomp(onig_posix_regex_t * reg,const char * pattern,int posix_options)162 onig_posix_regcomp(onig_posix_regex_t* reg, const char* pattern, int posix_options)
163 {
164   int r, len;
165   OnigSyntaxType* syntax = OnigDefaultSyntax;
166   OnigOptionType options;
167 
168   reg->onig = (void* )0;
169 
170   if ((posix_options & REG_EXTENDED) == 0)
171     syntax = ONIG_SYNTAX_POSIX_BASIC;
172 
173   options = syntax->options;
174   if ((posix_options & REG_ICASE)   != 0)
175     ONIG_OPTION_ON(options, ONIG_OPTION_IGNORECASE);
176   if ((posix_options & REG_NEWLINE) != 0) {
177     ONIG_OPTION_ON( options, ONIG_OPTION_NEGATE_SINGLELINE);
178     ONIG_OPTION_OFF(options, ONIG_OPTION_SINGLELINE);
179   }
180 
181   reg->comp_options = posix_options;
182 
183   ENC_STRING_LEN(OnigEncDefaultCharEncoding, pattern, len);
184   r = onig_new(PONIG_C(reg), (UChar* )pattern, (UChar* )(pattern + len),
185                options, OnigEncDefaultCharEncoding, syntax,
186                (OnigErrorInfo* )NULL);
187   if (r != ONIG_NORMAL) {
188     return onig2posix_error_code(r);
189   }
190 
191   reg->re_nsub = ONIG_C(reg)->num_mem;
192   return 0;
193 }
194 
195 extern int
onig_posix_regexec(onig_posix_regex_t * reg,const char * str,size_t nmatch,onig_posix_regmatch_t pmatch[],int posix_options)196 onig_posix_regexec(onig_posix_regex_t* reg, const char* str, size_t nmatch,
197                    onig_posix_regmatch_t pmatch[], int posix_options)
198 {
199   int r, i, len;
200   UChar* end;
201   onig_posix_regmatch_t* pm;
202   OnigOptionType options;
203 
204   options = ONIG_OPTION_POSIX_REGION;
205   if ((posix_options & REG_NOTBOL) != 0) options |= ONIG_OPTION_NOTBOL;
206   if ((posix_options & REG_NOTEOL) != 0) options |= ONIG_OPTION_NOTEOL;
207 
208   if (nmatch == 0 || (reg->comp_options & REG_NOSUB) != 0) {
209     pm = (onig_posix_regmatch_t* )NULL;
210     nmatch = 0;
211   }
212   else if ((int )nmatch < ONIG_C(reg)->num_mem + 1) {
213     pm = (onig_posix_regmatch_t* )xmalloc(sizeof(onig_posix_regmatch_t)
214                                * (ONIG_C(reg)->num_mem + 1));
215     if (pm == NULL)
216       return REG_ESPACE;
217   }
218   else {
219     pm = pmatch;
220   }
221 
222   ENC_STRING_LEN(ONIG_C(reg)->enc, str, len);
223   end = (UChar* )(str + len);
224   r = onig_search(ONIG_C(reg), (UChar* )str, end, (UChar* )str, end,
225                   (OnigRegion* )pm, options);
226 
227   if (r >= 0) {
228     r = 0; /* Match */
229     if (pm != pmatch && pm != NULL) {
230       xmemcpy(pmatch, pm, sizeof(onig_posix_regmatch_t) * nmatch);
231     }
232   }
233   else if (r == ONIG_MISMATCH) {
234     r = REG_NOMATCH;
235     for (i = 0; i < (int )nmatch; i++)
236       pmatch[i].rm_so = pmatch[i].rm_eo = ONIG_REGION_NOTPOS;
237   }
238   else {
239     r = onig2posix_error_code(r);
240   }
241 
242   if (pm != pmatch && pm != NULL)
243     xfree(pm);
244 
245 #if 0
246   if (reg->re_nsub > nmatch - 1)
247     reg->re_nsub = (nmatch <= 1 ? 0 : nmatch - 1);
248 #endif
249 
250   return r;
251 }
252 
253 extern void
onig_posix_regfree(onig_posix_regex_t * reg)254 onig_posix_regfree(onig_posix_regex_t* reg)
255 {
256   onig_free(ONIG_C(reg));
257   reg->onig = (void* )0;
258 }
259 
260 
261 extern void
onig_posix_reg_set_encoding(int mb_code)262 onig_posix_reg_set_encoding(int mb_code)
263 {
264   OnigEncoding enc;
265 
266   switch (mb_code) {
267   case REG_POSIX_ENCODING_ASCII:
268     enc = ONIG_ENCODING_ASCII;
269     break;
270   case REG_POSIX_ENCODING_EUC_JP:
271     enc = ONIG_ENCODING_EUC_JP;
272     break;
273   case REG_POSIX_ENCODING_SJIS:
274     enc = ONIG_ENCODING_SJIS;
275     break;
276   case REG_POSIX_ENCODING_UTF8:
277     enc = ONIG_ENCODING_UTF8;
278     break;
279   case REG_POSIX_ENCODING_UTF16_BE:
280     enc = ONIG_ENCODING_UTF16_BE;
281     break;
282   case REG_POSIX_ENCODING_UTF16_LE:
283     enc = ONIG_ENCODING_UTF16_LE;
284     break;
285 
286   default:
287     return ;
288     break;
289   }
290 
291   onig_initialize(&enc, 1);
292 
293   onigenc_set_default_encoding(enc);
294 }
295 
296 extern int
onig_posix_reg_name_to_group_numbers(onig_posix_regex_t * reg,const unsigned char * name,const unsigned char * name_end,int ** nums)297 onig_posix_reg_name_to_group_numbers(onig_posix_regex_t* reg,
298   const unsigned char* name, const unsigned char* name_end, int** nums)
299 {
300   return onig_name_to_group_numbers(ONIG_C(reg), name, name_end, nums);
301 }
302 
303 typedef struct {
304   int (*func)(const unsigned char*, const unsigned char*,int,int*,onig_posix_regex_t*,void*);
305   onig_posix_regex_t* reg;
306   void* arg;
307 } i_wrap;
308 
309 static int
i_wrapper(const UChar * name,const UChar * name_end,int ng,int * gs,onig_regex_t * reg ARG_UNUSED,void * arg)310 i_wrapper(const UChar* name, const UChar* name_end, int ng, int* gs,
311           onig_regex_t* reg ARG_UNUSED, void* arg)
312 {
313   i_wrap* warg = (i_wrap* )arg;
314 
315   return (*warg->func)(name, name_end, ng, gs, warg->reg, warg->arg);
316 }
317 
318 extern int
onig_posix_reg_foreach_name(onig_posix_regex_t * reg,int (* func)(const unsigned char *,const unsigned char *,int,int *,onig_posix_regex_t *,void *),void * arg)319 onig_posix_reg_foreach_name(onig_posix_regex_t* reg,
320  int (*func)(const unsigned char*, const unsigned char*,int,int*,onig_posix_regex_t*,void*),
321  void* arg)
322 {
323   i_wrap warg;
324 
325   warg.func = func;
326   warg.reg  = reg;
327   warg.arg  = arg;
328 
329   return onig_foreach_name(ONIG_C(reg), i_wrapper, &warg);
330 }
331 
332 extern int
onig_posix_reg_number_of_names(onig_posix_regex_t * reg)333 onig_posix_reg_number_of_names(onig_posix_regex_t* reg)
334 {
335   return onig_number_of_names(ONIG_C(reg));
336 }
337 
338 
339 #ifdef USE_BINARY_COMPATIBLE_POSIX_API
340 
341 extern int
regcomp(onig_posix_regex_t * reg,const char * pattern,int posix_options)342 regcomp(onig_posix_regex_t* reg, const char* pattern, int posix_options)
343 {
344   return onig_posix_regcomp(reg, pattern, posix_options);
345 }
346 
347 extern int
regexec(onig_posix_regex_t * reg,const char * str,size_t nmatch,onig_posix_regmatch_t pmatch[],int posix_options)348 regexec(onig_posix_regex_t* reg, const char* str, size_t nmatch,
349         onig_posix_regmatch_t pmatch[], int posix_options)
350 {
351   return onig_posix_regexec(reg, str, nmatch, pmatch, posix_options);
352 }
353 
354 extern void
regfree(onig_posix_regex_t * reg)355 regfree(onig_posix_regex_t* reg)
356 {
357   onig_posix_regfree(reg);
358 }
359 
360 extern void
reg_set_encoding(int mb_code)361 reg_set_encoding(int mb_code)
362 {
363   onig_posix_reg_set_encoding(mb_code);
364 }
365 
366 extern int
reg_name_to_group_numbers(onig_posix_regex_t * reg,const unsigned char * name,const unsigned char * name_end,int ** nums)367 reg_name_to_group_numbers(onig_posix_regex_t* reg,
368   const unsigned char* name, const unsigned char* name_end, int** nums)
369 {
370   return onig_posix_reg_name_to_group_numbers(reg, name, name_end, nums);
371 }
372 
373 extern int
reg_foreach_name(onig_posix_regex_t * reg,int (* func)(const unsigned char *,const unsigned char *,int,int *,onig_posix_regex_t *,void *),void * arg)374 reg_foreach_name(onig_posix_regex_t* reg,
375   int (*func)(const unsigned char*, const unsigned char*,int,int*,onig_posix_regex_t*,void*),
376   void* arg)
377 {
378   return onig_posix_reg_foreach_name(reg, func, arg);
379 }
380 
381 extern int
reg_number_of_names(onig_posix_regex_t * reg)382 reg_number_of_names(onig_posix_regex_t* reg)
383 {
384   return onig_posix_reg_number_of_names(reg);
385 }
386 
387 #endif /* USE_BINARY_COMPATIBLE_POSIX_API */
388