1 | /**********************************************************************
|
---|
2 | regerror.c - Onigmo (Oniguruma-mod) (regular expression library)
|
---|
3 | **********************************************************************/
|
---|
4 | /*-
|
---|
5 | * Copyright (c) 2002-2007 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>
|
---|
6 | * Copyright (c) 2011-2016 K.Takata <kentkt AT csc DOT jp>
|
---|
7 | * All rights reserved.
|
---|
8 | *
|
---|
9 | * Redistribution and use in source and binary forms, with or without
|
---|
10 | * modification, are permitted provided that the following conditions
|
---|
11 | * are met:
|
---|
12 | * 1. Redistributions of source code must retain the above copyright
|
---|
13 | * notice, this list of conditions and the following disclaimer.
|
---|
14 | * 2. Redistributions in binary form must reproduce the above copyright
|
---|
15 | * notice, this list of conditions and the following disclaimer in the
|
---|
16 | * documentation and/or other materials provided with the distribution.
|
---|
17 | *
|
---|
18 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
---|
19 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
---|
20 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
---|
21 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
---|
22 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
---|
23 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
---|
24 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
---|
25 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
---|
26 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
---|
27 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
---|
28 | * SUCH DAMAGE.
|
---|
29 | */
|
---|
30 |
|
---|
31 | #include "regint.h"
|
---|
32 | #include <stdio.h> /* for vsnprintf() */
|
---|
33 |
|
---|
34 | #include <stdarg.h>
|
---|
35 |
|
---|
36 | extern UChar*
|
---|
37 | onig_error_code_to_format(OnigPosition code)
|
---|
38 | {
|
---|
39 | const char *p;
|
---|
40 |
|
---|
41 | if (code >= 0) return (UChar* )0;
|
---|
42 |
|
---|
43 | switch (code) {
|
---|
44 | case ONIG_MISMATCH:
|
---|
45 | p = "mismatch"; break;
|
---|
46 | case ONIG_NO_SUPPORT_CONFIG:
|
---|
47 | p = "no support in this configuration"; break;
|
---|
48 | case ONIGERR_MEMORY:
|
---|
49 | p = "failed to allocate memory"; break;
|
---|
50 | case ONIGERR_TYPE_BUG:
|
---|
51 | p = "undefined type (bug)"; break;
|
---|
52 | case ONIGERR_PARSER_BUG:
|
---|
53 | p = "internal parser error (bug)"; break;
|
---|
54 | case ONIGERR_STACK_BUG:
|
---|
55 | p = "stack error (bug)"; break;
|
---|
56 | case ONIGERR_UNDEFINED_BYTECODE:
|
---|
57 | p = "undefined bytecode (bug)"; break;
|
---|
58 | case ONIGERR_UNEXPECTED_BYTECODE:
|
---|
59 | p = "unexpected bytecode (bug)"; break;
|
---|
60 | case ONIGERR_MATCH_STACK_LIMIT_OVER:
|
---|
61 | p = "match-stack limit over"; break;
|
---|
62 | case ONIGERR_PARSE_DEPTH_LIMIT_OVER:
|
---|
63 | p = "parse depth limit over"; break;
|
---|
64 | case ONIGERR_DEFAULT_ENCODING_IS_NOT_SET:
|
---|
65 | p = "default multibyte-encoding is not set"; break;
|
---|
66 | case ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR:
|
---|
67 | p = "can't convert to wide-char on specified multibyte-encoding"; break;
|
---|
68 | case ONIGERR_INVALID_ARGUMENT:
|
---|
69 | p = "invalid argument"; break;
|
---|
70 | case ONIGERR_END_PATTERN_AT_LEFT_BRACE:
|
---|
71 | p = "end pattern at left brace"; break;
|
---|
72 | case ONIGERR_END_PATTERN_AT_LEFT_BRACKET:
|
---|
73 | p = "end pattern at left bracket"; break;
|
---|
74 | case ONIGERR_EMPTY_CHAR_CLASS:
|
---|
75 | p = "empty char-class"; break;
|
---|
76 | case ONIGERR_PREMATURE_END_OF_CHAR_CLASS:
|
---|
77 | p = "premature end of char-class"; break;
|
---|
78 | case ONIGERR_END_PATTERN_AT_ESCAPE:
|
---|
79 | p = "end pattern at escape"; break;
|
---|
80 | case ONIGERR_END_PATTERN_AT_META:
|
---|
81 | p = "end pattern at meta"; break;
|
---|
82 | case ONIGERR_END_PATTERN_AT_CONTROL:
|
---|
83 | p = "end pattern at control"; break;
|
---|
84 | case ONIGERR_META_CODE_SYNTAX:
|
---|
85 | p = "invalid meta-code syntax"; break;
|
---|
86 | case ONIGERR_CONTROL_CODE_SYNTAX:
|
---|
87 | p = "invalid control-code syntax"; break;
|
---|
88 | case ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE:
|
---|
89 | p = "char-class value at end of range"; break;
|
---|
90 | case ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE:
|
---|
91 | p = "char-class value at start of range"; break;
|
---|
92 | case ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS:
|
---|
93 | p = "unmatched range specifier in char-class"; break;
|
---|
94 | case ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED:
|
---|
95 | p = "target of repeat operator is not specified"; break;
|
---|
96 | case ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID:
|
---|
97 | p = "target of repeat operator is invalid"; break;
|
---|
98 | case ONIGERR_NESTED_REPEAT_OPERATOR:
|
---|
99 | p = "nested repeat operator"; break;
|
---|
100 | case ONIGERR_UNMATCHED_CLOSE_PARENTHESIS:
|
---|
101 | p = "unmatched close parenthesis"; break;
|
---|
102 | case ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS:
|
---|
103 | p = "end pattern with unmatched parenthesis"; break;
|
---|
104 | case ONIGERR_END_PATTERN_IN_GROUP:
|
---|
105 | p = "end pattern in group"; break;
|
---|
106 | case ONIGERR_UNDEFINED_GROUP_OPTION:
|
---|
107 | p = "undefined group option"; break;
|
---|
108 | case ONIGERR_INVALID_POSIX_BRACKET_TYPE:
|
---|
109 | p = "invalid POSIX bracket type"; break;
|
---|
110 | case ONIGERR_INVALID_LOOK_BEHIND_PATTERN:
|
---|
111 | p = "invalid pattern in look-behind"; break;
|
---|
112 | case ONIGERR_INVALID_REPEAT_RANGE_PATTERN:
|
---|
113 | p = "invalid repeat range {lower,upper}"; break;
|
---|
114 | case ONIGERR_INVALID_CONDITION_PATTERN:
|
---|
115 | p = "invalid conditional pattern"; break;
|
---|
116 | case ONIGERR_TOO_BIG_NUMBER:
|
---|
117 | p = "too big number"; break;
|
---|
118 | case ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE:
|
---|
119 | p = "too big number for repeat range"; break;
|
---|
120 | case ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE:
|
---|
121 | p = "upper is smaller than lower in repeat range"; break;
|
---|
122 | case ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS:
|
---|
123 | p = "empty range in char class"; break;
|
---|
124 | case ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE:
|
---|
125 | p = "mismatch multibyte code length in char-class range"; break;
|
---|
126 | case ONIGERR_TOO_MANY_MULTI_BYTE_RANGES:
|
---|
127 | p = "too many multibyte code ranges are specified"; break;
|
---|
128 | case ONIGERR_TOO_SHORT_MULTI_BYTE_STRING:
|
---|
129 | p = "too short multibyte code string"; break;
|
---|
130 | case ONIGERR_TOO_BIG_BACKREF_NUMBER:
|
---|
131 | p = "too big backref number"; break;
|
---|
132 | case ONIGERR_INVALID_BACKREF:
|
---|
133 | #ifdef USE_NAMED_GROUP
|
---|
134 | p = "invalid backref number/name"; break;
|
---|
135 | #else
|
---|
136 | p = "invalid backref number"; break;
|
---|
137 | #endif
|
---|
138 | case ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED:
|
---|
139 | p = "numbered backref/call is not allowed. (use name)"; break;
|
---|
140 | case ONIGERR_TOO_SHORT_DIGITS:
|
---|
141 | p = "too short digits"; break;
|
---|
142 | case ONIGERR_TOO_LONG_WIDE_CHAR_VALUE:
|
---|
143 | p = "too long wide-char value"; break;
|
---|
144 | case ONIGERR_EMPTY_GROUP_NAME:
|
---|
145 | p = "group name is empty"; break;
|
---|
146 | case ONIGERR_INVALID_GROUP_NAME:
|
---|
147 | p = "invalid group name <%n>"; break;
|
---|
148 | case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
|
---|
149 | #ifdef USE_NAMED_GROUP
|
---|
150 | p = "invalid char in group name <%n>"; break;
|
---|
151 | #else
|
---|
152 | p = "invalid char in group number <%n>"; break;
|
---|
153 | #endif
|
---|
154 | case ONIGERR_UNDEFINED_NAME_REFERENCE:
|
---|
155 | p = "undefined name <%n> reference"; break;
|
---|
156 | case ONIGERR_UNDEFINED_GROUP_REFERENCE:
|
---|
157 | p = "undefined group <%n> reference"; break;
|
---|
158 | case ONIGERR_MULTIPLEX_DEFINED_NAME:
|
---|
159 | p = "multiplex defined name <%n>"; break;
|
---|
160 | case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
|
---|
161 | p = "multiplex definition name <%n> call"; break;
|
---|
162 | case ONIGERR_NEVER_ENDING_RECURSION:
|
---|
163 | p = "never ending recursion"; break;
|
---|
164 | case ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY:
|
---|
165 | p = "group number is too big for capture history"; break;
|
---|
166 | case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
|
---|
167 | p = "invalid character property name {%n}"; break;
|
---|
168 | case ONIGERR_TOO_MANY_CAPTURE_GROUPS:
|
---|
169 | p = "too many capture groups are specified"; break;
|
---|
170 | case ONIGERR_INVALID_CODE_POINT_VALUE:
|
---|
171 | p = "invalid code point value"; break;
|
---|
172 | case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE:
|
---|
173 | p = "too big wide-char value"; break;
|
---|
174 | case ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION:
|
---|
175 | p = "not supported encoding combination"; break;
|
---|
176 | case ONIGERR_INVALID_COMBINATION_OF_OPTIONS:
|
---|
177 | p = "invalid combination of options"; break;
|
---|
178 |
|
---|
179 | default:
|
---|
180 | p = "undefined error code"; break;
|
---|
181 | }
|
---|
182 |
|
---|
183 | return (UChar* )p;
|
---|
184 | }
|
---|
185 |
|
---|
186 | static void sprint_byte(char* s, unsigned int v)
|
---|
187 | {
|
---|
188 | xsnprintf(s, 3, "%02x", (v & 0377));
|
---|
189 | }
|
---|
190 |
|
---|
191 | static void sprint_byte_with_x(char* s, unsigned int v)
|
---|
192 | {
|
---|
193 | xsnprintf(s, 5, "\\x%02x", (v & 0377));
|
---|
194 | }
|
---|
195 |
|
---|
196 | static int to_ascii(OnigEncoding enc, UChar *s, UChar *end,
|
---|
197 | UChar buf[], int buf_size, int *is_over)
|
---|
198 | {
|
---|
199 | int len;
|
---|
200 | UChar *p;
|
---|
201 | OnigCodePoint code;
|
---|
202 |
|
---|
203 | if (ONIGENC_MBC_MINLEN(enc) > 1) {
|
---|
204 | p = s;
|
---|
205 | len = 0;
|
---|
206 | while (p < end) {
|
---|
207 | code = ONIGENC_MBC_TO_CODE(enc, p, end);
|
---|
208 | if (code >= 0x80) {
|
---|
209 | if (code > 0xffff && len + 10 <= buf_size) {
|
---|
210 | sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 24));
|
---|
211 | sprint_byte((char*)(&(buf[len+4])), (unsigned int)(code >> 16));
|
---|
212 | sprint_byte((char*)(&(buf[len+6])), (unsigned int)(code >> 8));
|
---|
213 | sprint_byte((char*)(&(buf[len+8])), (unsigned int)code);
|
---|
214 | len += 10;
|
---|
215 | }
|
---|
216 | else if (len + 6 <= buf_size) {
|
---|
217 | sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 8));
|
---|
218 | sprint_byte((char*)(&(buf[len+4])), (unsigned int)code);
|
---|
219 | len += 6;
|
---|
220 | }
|
---|
221 | else {
|
---|
222 | break;
|
---|
223 | }
|
---|
224 | }
|
---|
225 | else {
|
---|
226 | buf[len++] = (UChar )code;
|
---|
227 | }
|
---|
228 |
|
---|
229 | p += enclen(enc, p, end);
|
---|
230 | if (len >= buf_size) break;
|
---|
231 | }
|
---|
232 |
|
---|
233 | *is_over = ((p < end) ? 1 : 0);
|
---|
234 | }
|
---|
235 | else {
|
---|
236 | len = (int )MIN((end - s), buf_size);
|
---|
237 | xmemcpy(buf, s, (size_t )len);
|
---|
238 | *is_over = ((buf_size < (end - s)) ? 1 : 0);
|
---|
239 | }
|
---|
240 |
|
---|
241 | return len;
|
---|
242 | }
|
---|
243 |
|
---|
244 |
|
---|
245 | /* for ONIG_MAX_ERROR_MESSAGE_LEN */
|
---|
246 | #define MAX_ERROR_PAR_LEN 30
|
---|
247 |
|
---|
248 | extern int
|
---|
249 | onig_error_code_to_str(UChar* s, OnigPosition code, ...)
|
---|
250 | {
|
---|
251 | UChar *p, *q;
|
---|
252 | OnigErrorInfo* einfo;
|
---|
253 | size_t len;
|
---|
254 | int is_over;
|
---|
255 | UChar parbuf[MAX_ERROR_PAR_LEN];
|
---|
256 | va_list vargs;
|
---|
257 |
|
---|
258 | va_start(vargs, code);
|
---|
259 |
|
---|
260 | switch (code) {
|
---|
261 | case ONIGERR_UNDEFINED_NAME_REFERENCE:
|
---|
262 | case ONIGERR_UNDEFINED_GROUP_REFERENCE:
|
---|
263 | case ONIGERR_MULTIPLEX_DEFINED_NAME:
|
---|
264 | case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
|
---|
265 | case ONIGERR_INVALID_GROUP_NAME:
|
---|
266 | case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
|
---|
267 | case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
|
---|
268 | einfo = va_arg(vargs, OnigErrorInfo*);
|
---|
269 | len = to_ascii(einfo->enc, einfo->par, einfo->par_end,
|
---|
270 | parbuf, MAX_ERROR_PAR_LEN - 3, &is_over);
|
---|
271 | q = onig_error_code_to_format(code);
|
---|
272 | p = s;
|
---|
273 | while (*q != '\0') {
|
---|
274 | if (*q == '%') {
|
---|
275 | q++;
|
---|
276 | if (*q == 'n') { /* '%n': name */
|
---|
277 | xmemcpy(p, parbuf, len);
|
---|
278 | p += len;
|
---|
279 | if (is_over != 0) {
|
---|
280 | xmemcpy(p, "...", 3);
|
---|
281 | p += 3;
|
---|
282 | }
|
---|
283 | q++;
|
---|
284 | }
|
---|
285 | else
|
---|
286 | goto normal_char;
|
---|
287 | }
|
---|
288 | else {
|
---|
289 | normal_char:
|
---|
290 | *p++ = *q++;
|
---|
291 | }
|
---|
292 | }
|
---|
293 | *p = '\0';
|
---|
294 | len = p - s;
|
---|
295 | break;
|
---|
296 |
|
---|
297 | default:
|
---|
298 | q = onig_error_code_to_format(code);
|
---|
299 | if (q) {
|
---|
300 | len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, q);
|
---|
301 | xmemcpy(s, q, len);
|
---|
302 | } else {
|
---|
303 | len = 0;
|
---|
304 | }
|
---|
305 | s[len] = '\0';
|
---|
306 | break;
|
---|
307 | }
|
---|
308 |
|
---|
309 | va_end(vargs);
|
---|
310 | return (int )len;
|
---|
311 | }
|
---|
312 |
|
---|
313 | void
|
---|
314 | onig_vsnprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
|
---|
315 | UChar* pat, UChar* pat_end, const UChar *fmt, va_list args)
|
---|
316 | {
|
---|
317 | size_t need;
|
---|
318 | int n, len;
|
---|
319 | UChar *p, *s, *bp;
|
---|
320 | UChar bs[6];
|
---|
321 |
|
---|
322 | n = xvsnprintf((char* )buf, bufsize, (const char* )fmt, args);
|
---|
323 |
|
---|
324 | need = (pat_end - pat) * 4 + 4;
|
---|
325 |
|
---|
326 | if (n + need < (size_t )bufsize) {
|
---|
327 | static const char sep[] = ": /";
|
---|
328 | memcpy((char* )buf + n, sep, sizeof(sep));
|
---|
329 | s = buf + onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, buf);
|
---|
330 |
|
---|
331 | p = pat;
|
---|
332 | while (p < pat_end) {
|
---|
333 | if (ONIGENC_IS_MBC_HEAD(enc, p, pat_end)) {
|
---|
334 | len = enclen(enc, p, pat_end);
|
---|
335 | if (ONIGENC_MBC_MINLEN(enc) == 1) {
|
---|
336 | while (len-- > 0) *s++ = *p++;
|
---|
337 | }
|
---|
338 | else { /* for UTF16/32 */
|
---|
339 | int blen;
|
---|
340 |
|
---|
341 | while (len-- > 0) {
|
---|
342 | sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
|
---|
343 | blen = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
|
---|
344 | bp = bs;
|
---|
345 | while (blen-- > 0) *s++ = *bp++;
|
---|
346 | }
|
---|
347 | }
|
---|
348 | }
|
---|
349 | else if (*p == '\\') {
|
---|
350 | *s++ = *p++;
|
---|
351 | len = enclen(enc, p, pat_end);
|
---|
352 | while (len-- > 0) *s++ = *p++;
|
---|
353 | }
|
---|
354 | else if (*p == '/') {
|
---|
355 | *s++ = (unsigned char )'\\';
|
---|
356 | *s++ = *p++;
|
---|
357 | }
|
---|
358 | else if (!ONIGENC_IS_CODE_PRINT(enc, *p) &&
|
---|
359 | !ONIGENC_IS_CODE_SPACE(enc, *p)) {
|
---|
360 | sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
|
---|
361 | len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
|
---|
362 | bp = bs;
|
---|
363 | while (len-- > 0) *s++ = *bp++;
|
---|
364 | }
|
---|
365 | else {
|
---|
366 | *s++ = *p++;
|
---|
367 | }
|
---|
368 | }
|
---|
369 |
|
---|
370 | *s++ = '/';
|
---|
371 | *s = '\0';
|
---|
372 | }
|
---|
373 | }
|
---|
374 |
|
---|
375 | #if 0 /* unused */
|
---|
376 | void
|
---|
377 | onig_snprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
|
---|
378 | UChar* pat, UChar* pat_end, const UChar *fmt, ...)
|
---|
379 | {
|
---|
380 | va_list args;
|
---|
381 | va_start(args, fmt);
|
---|
382 | onig_vsnprintf_with_pattern(buf, bufsize, enc,
|
---|
383 | pat, pat_end, fmt, args);
|
---|
384 | va_end(args);
|
---|
385 | }
|
---|
386 | #endif
|
---|