]> git.proxmox.com Git - mirror_edk2.git/blame - MdeModulePkg/Universal/RegularExpressionDxe/Oniguruma/regposix.c
MdeModulePkg/UfsPassThruDxe: Implement EDKII_UFS_HC_PLATFORM_PROTOCOL
[mirror_edk2.git] / MdeModulePkg / Universal / RegularExpressionDxe / Oniguruma / regposix.c
CommitLineData
14b0e578
CS
1/**********************************************************************\r
2 regposix.c - Oniguruma (regular expression library)\r
3**********************************************************************/\r
4/*-\r
b602265d 5 * Copyright (c) 2002-2018 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>\r
14b0e578
CS
6 * All rights reserved.\r
7 *\r
14b0e578
CS
8 * Redistribution and use in source and binary forms, with or without\r
9 * modification, are permitted provided that the following conditions\r
10 * are met:\r
11 * 1. Redistributions of source code must retain the above copyright\r
12 * notice, this list of conditions and the following disclaimer.\r
13 * 2. Redistributions in binary form must reproduce the above copyright\r
14 * notice, this list of conditions and the following disclaimer in the\r
15 * documentation and/or other materials provided with the distribution.\r
16 *\r
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND\r
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\r
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE\r
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE\r
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\r
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS\r
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)\r
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT\r
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY\r
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF\r
27 * SUCH DAMAGE.\r
28 */\r
29\r
30#define regex_t onig_regex_t\r
31#include "regint.h"\r
32#undef regex_t\r
33#include "onigposix.h"\r
34\r
35#define ONIG_C(reg) ((onig_regex_t* )((reg)->onig))\r
36#define PONIG_C(reg) ((onig_regex_t** )(&(reg)->onig))\r
37\r
38/* #define ENC_STRING_LEN(enc,s,len) len = strlen(s) */\r
39#define ENC_STRING_LEN(enc,s,len) do { \\r
40 if (ONIGENC_MBC_MINLEN(enc) == 1) { \\r
41 UChar* tmps = (UChar* )(s); \\r
42 while (*tmps != 0) tmps++; \\r
b602265d 43 len = (int )(tmps - (UChar* )(s));\\r
14b0e578
CS
44 } \\r
45 else { \\r
46 len = onigenc_str_bytelen_null(enc, (UChar* )s); \\r
47 } \\r
48} while(0)\r
49\r
50typedef struct {\r
51 int onig_err;\r
52 int posix_err;\r
53} O2PERR;\r
54\r
55static int\r
56onig2posix_error_code(int code)\r
57{\r
58 static const O2PERR o2p[] = {\r
59 { ONIG_MISMATCH, REG_NOMATCH },\r
60 { ONIG_NO_SUPPORT_CONFIG, REG_EONIG_INTERNAL },\r
b602265d 61 { ONIG_ABORT, REG_EONIG_INTERNAL },\r
14b0e578
CS
62 { ONIGERR_MEMORY, REG_ESPACE },\r
63 { ONIGERR_MATCH_STACK_LIMIT_OVER, REG_EONIG_INTERNAL },\r
b602265d 64 { ONIGERR_RETRY_LIMIT_IN_MATCH_OVER, REG_EONIG_INTERNAL },\r
14b0e578
CS
65 { ONIGERR_TYPE_BUG, REG_EONIG_INTERNAL },\r
66 { ONIGERR_PARSER_BUG, REG_EONIG_INTERNAL },\r
67 { ONIGERR_STACK_BUG, REG_EONIG_INTERNAL },\r
68 { ONIGERR_UNDEFINED_BYTECODE, REG_EONIG_INTERNAL },\r
69 { ONIGERR_UNEXPECTED_BYTECODE, REG_EONIG_INTERNAL },\r
70 { ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED, REG_EONIG_BADARG },\r
71 { ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR, REG_EONIG_BADARG },\r
b602265d 72 { ONIGERR_FAIL_TO_INITIALIZE, REG_EONIG_INTERNAL },\r
14b0e578
CS
73 { ONIGERR_INVALID_ARGUMENT, REG_EONIG_BADARG },\r
74 { ONIGERR_END_PATTERN_AT_LEFT_BRACE, REG_EBRACE },\r
75 { ONIGERR_END_PATTERN_AT_LEFT_BRACKET, REG_EBRACK },\r
76 { ONIGERR_EMPTY_CHAR_CLASS, REG_ECTYPE },\r
77 { ONIGERR_PREMATURE_END_OF_CHAR_CLASS, REG_ECTYPE },\r
78 { ONIGERR_END_PATTERN_AT_ESCAPE, REG_EESCAPE },\r
79 { ONIGERR_END_PATTERN_AT_META, REG_EESCAPE },\r
80 { ONIGERR_END_PATTERN_AT_CONTROL, REG_EESCAPE },\r
81 { ONIGERR_META_CODE_SYNTAX, REG_BADPAT },\r
82 { ONIGERR_CONTROL_CODE_SYNTAX, REG_BADPAT },\r
83 { ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE, REG_ECTYPE },\r
84 { ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE, REG_ECTYPE },\r
85 { ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS, REG_ECTYPE },\r
86 { ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED, REG_BADRPT },\r
87 { ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID, REG_BADRPT },\r
88 { ONIGERR_NESTED_REPEAT_OPERATOR, REG_BADRPT },\r
89 { ONIGERR_UNMATCHED_CLOSE_PARENTHESIS, REG_EPAREN },\r
90 { ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS, REG_EPAREN },\r
91 { ONIGERR_END_PATTERN_IN_GROUP, REG_BADPAT },\r
92 { ONIGERR_UNDEFINED_GROUP_OPTION, REG_BADPAT },\r
93 { ONIGERR_INVALID_POSIX_BRACKET_TYPE, REG_BADPAT },\r
94 { ONIGERR_INVALID_LOOK_BEHIND_PATTERN, REG_BADPAT },\r
95 { ONIGERR_INVALID_REPEAT_RANGE_PATTERN, REG_BADPAT },\r
96 { ONIGERR_TOO_BIG_NUMBER, REG_BADPAT },\r
97 { ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE, REG_BADBR },\r
98 { ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE, REG_BADBR },\r
99 { ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS, REG_ECTYPE },\r
100 { ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE, REG_ECTYPE },\r
101 { ONIGERR_TOO_MANY_MULTI_BYTE_RANGES, REG_ECTYPE },\r
102 { ONIGERR_TOO_SHORT_MULTI_BYTE_STRING, REG_BADPAT },\r
103 { ONIGERR_TOO_BIG_BACKREF_NUMBER, REG_ESUBREG },\r
104 { ONIGERR_INVALID_BACKREF, REG_ESUBREG },\r
105 { ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED, REG_BADPAT },\r
106 { ONIGERR_TOO_BIG_WIDE_CHAR_VALUE, REG_EONIG_BADWC },\r
107 { ONIGERR_TOO_LONG_WIDE_CHAR_VALUE, REG_EONIG_BADWC },\r
108 { ONIGERR_INVALID_CODE_POINT_VALUE, REG_EONIG_BADWC },\r
109 { ONIGERR_EMPTY_GROUP_NAME, REG_BADPAT },\r
110 { ONIGERR_INVALID_GROUP_NAME, REG_BADPAT },\r
111 { ONIGERR_INVALID_CHAR_IN_GROUP_NAME, REG_BADPAT },\r
112 { ONIGERR_UNDEFINED_NAME_REFERENCE, REG_BADPAT },\r
113 { ONIGERR_UNDEFINED_GROUP_REFERENCE, REG_BADPAT },\r
114 { ONIGERR_MULTIPLEX_DEFINED_NAME, REG_BADPAT },\r
115 { ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL, REG_BADPAT },\r
116 { ONIGERR_NEVER_ENDING_RECURSION, REG_BADPAT },\r
117 { ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY, REG_BADPAT },\r
118 { ONIGERR_INVALID_CHAR_PROPERTY_NAME, REG_BADPAT },\r
b602265d
DG
119 { ONIGERR_INVALID_IF_ELSE_SYNTAX, REG_BADPAT },\r
120 { ONIGERR_INVALID_ABSENT_GROUP_PATTERN, REG_BADPAT },\r
121 { ONIGERR_INVALID_ABSENT_GROUP_GENERATOR_PATTERN, REG_BADPAT },\r
122 { ONIGERR_INVALID_CALLOUT_PATTERN, REG_BADPAT },\r
123 { ONIGERR_INVALID_CALLOUT_NAME, REG_BADPAT },\r
124 { ONIGERR_UNDEFINED_CALLOUT_NAME, REG_BADPAT },\r
125 { ONIGERR_INVALID_CALLOUT_BODY, REG_BADPAT },\r
126 { ONIGERR_INVALID_CALLOUT_TAG_NAME, REG_BADPAT },\r
127 { ONIGERR_INVALID_CALLOUT_ARG, REG_BADPAT },\r
14b0e578 128 { ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION, REG_EONIG_BADARG },\r
b602265d 129 { ONIGERR_LIBRARY_IS_NOT_INITIALIZED, REG_EONIG_INTERNAL }\r
14b0e578
CS
130 };\r
131\r
132 int i;\r
133\r
134 if (code >= 0) return 0;\r
135\r
136 for (i = 0; i < (int )(sizeof(o2p) / sizeof(o2p[0])); i++) {\r
137 if (code == o2p[i].onig_err)\r
138 return o2p[i].posix_err;\r
139 }\r
140\r
141 return REG_EONIG_INTERNAL; /* but, unknown error code */\r
142}\r
143\r
144extern int\r
145regcomp(regex_t* reg, const char* pattern, int posix_options)\r
146{\r
147 int r, len;\r
148 OnigSyntaxType* syntax = OnigDefaultSyntax;\r
149 OnigOptionType options;\r
150\r
151 if ((posix_options & REG_EXTENDED) == 0)\r
152 syntax = ONIG_SYNTAX_POSIX_BASIC;\r
153\r
154 options = syntax->options;\r
155 if ((posix_options & REG_ICASE) != 0)\r
156 ONIG_OPTION_ON(options, ONIG_OPTION_IGNORECASE);\r
157 if ((posix_options & REG_NEWLINE) != 0) {\r
158 ONIG_OPTION_ON( options, ONIG_OPTION_NEGATE_SINGLELINE);\r
159 ONIG_OPTION_OFF(options, ONIG_OPTION_SINGLELINE);\r
160 }\r
161\r
162 reg->comp_options = posix_options;\r
163\r
164 ENC_STRING_LEN(OnigEncDefaultCharEncoding, pattern, len);\r
165 r = onig_new(PONIG_C(reg), (UChar* )pattern, (UChar* )(pattern + len),\r
166 options, OnigEncDefaultCharEncoding, syntax,\r
167 (OnigErrorInfo* )NULL);\r
168 if (r != ONIG_NORMAL) {\r
169 return onig2posix_error_code(r);\r
170 }\r
171\r
172 reg->re_nsub = ONIG_C(reg)->num_mem;\r
173 return 0;\r
174}\r
175\r
176extern int\r
177regexec(regex_t* reg, const char* str, size_t nmatch,\r
178 regmatch_t pmatch[], int posix_options)\r
179{\r
180 int r, i, len;\r
181 UChar* end;\r
182 regmatch_t* pm;\r
183 OnigOptionType options;\r
184\r
185 options = ONIG_OPTION_POSIX_REGION;\r
186 if ((posix_options & REG_NOTBOL) != 0) options |= ONIG_OPTION_NOTBOL;\r
187 if ((posix_options & REG_NOTEOL) != 0) options |= ONIG_OPTION_NOTEOL;\r
188\r
189 if (nmatch == 0 || (reg->comp_options & REG_NOSUB) != 0) {\r
190 pm = (regmatch_t* )NULL;\r
191 nmatch = 0;\r
192 }\r
193 else if ((int )nmatch < ONIG_C(reg)->num_mem + 1) {\r
194 pm = (regmatch_t* )xmalloc(sizeof(regmatch_t)\r
195 * (ONIG_C(reg)->num_mem + 1));\r
196 if (pm == NULL)\r
197 return REG_ESPACE;\r
198 }\r
199 else {\r
200 pm = pmatch;\r
201 }\r
202\r
203 ENC_STRING_LEN(ONIG_C(reg)->enc, str, len);\r
204 end = (UChar* )(str + len);\r
205 r = onig_search(ONIG_C(reg), (UChar* )str, end, (UChar* )str, end,\r
206 (OnigRegion* )pm, options);\r
207\r
208 if (r >= 0) {\r
209 r = 0; /* Match */\r
210 if (pm != pmatch && pm != NULL) {\r
211 xmemcpy(pmatch, pm, sizeof(regmatch_t) * nmatch);\r
212 }\r
213 }\r
214 else if (r == ONIG_MISMATCH) {\r
215 r = REG_NOMATCH;\r
216 for (i = 0; i < (int )nmatch; i++)\r
217 pmatch[i].rm_so = pmatch[i].rm_eo = ONIG_REGION_NOTPOS;\r
218 }\r
219 else {\r
220 r = onig2posix_error_code(r);\r
221 }\r
222\r
223 if (pm != pmatch && pm != NULL)\r
224 xfree(pm);\r
225\r
226#if 0\r
227 if (reg->re_nsub > nmatch - 1)\r
228 reg->re_nsub = (nmatch <= 1 ? 0 : nmatch - 1);\r
229#endif\r
230\r
231 return r;\r
232}\r
233\r
234extern void\r
235regfree(regex_t* reg)\r
236{\r
237 onig_free(ONIG_C(reg));\r
238}\r
239\r
240\r
241extern void\r
242reg_set_encoding(int mb_code)\r
243{\r
244 OnigEncoding enc;\r
245\r
246 switch (mb_code) {\r
247 case REG_POSIX_ENCODING_ASCII:\r
248 enc = ONIG_ENCODING_ASCII;\r
249 break;\r
14b0e578
CS
250 case REG_POSIX_ENCODING_UTF16_LE:\r
251 enc = ONIG_ENCODING_UTF16_LE;\r
252 break;\r
14b0e578
CS
253 }\r
254\r
b602265d
DG
255 onig_initialize(&enc, 1);\r
256\r
14b0e578
CS
257 onigenc_set_default_encoding(enc);\r
258}\r
259\r
260extern int\r
261reg_name_to_group_numbers(regex_t* reg,\r
262 const unsigned char* name, const unsigned char* name_end, int** nums)\r
263{\r
264 return onig_name_to_group_numbers(ONIG_C(reg), name, name_end, nums);\r
265}\r
266\r
267typedef struct {\r
268 int (*func)(const unsigned char*, const unsigned char*,int,int*,regex_t*,void*);\r
269 regex_t* reg;\r
270 void* arg;\r
271} i_wrap;\r
272\r
273static int\r
274i_wrapper(const UChar* name, const UChar* name_end, int ng, int* gs,\r
275 onig_regex_t* reg ARG_UNUSED, void* arg)\r
276{\r
277 i_wrap* warg = (i_wrap* )arg;\r
278\r
279 return (*warg->func)(name, name_end, ng, gs, warg->reg, warg->arg);\r
280}\r
281\r
282extern int\r
283reg_foreach_name(regex_t* reg,\r
284 int (*func)(const unsigned char*, const unsigned char*,int,int*,regex_t*,void*),\r
285 void* arg)\r
286{\r
287 i_wrap warg;\r
288\r
289 warg.func = func;\r
290 warg.reg = reg;\r
291 warg.arg = arg;\r
292\r
293 return onig_foreach_name(ONIG_C(reg), i_wrapper, &warg);\r
294}\r
295\r
296extern int\r
297reg_number_of_names(regex_t* reg)\r
298{\r
299 return onig_number_of_names(ONIG_C(reg));\r
300}\r