1 /**********************************************************************
2 regerror.c - Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5 * Copyright (c) 2002-2019 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 #include "regint.h"
31 #include <stdio.h> /* for vsnprintf() */
32
33 #include <stdarg.h>
34
35 extern UChar*
onig_error_code_to_format(int code)36 onig_error_code_to_format(int code)
37 {
38 char *p;
39
40 switch (code) {
41 case ONIG_MISMATCH:
42 p = "mismatch"; break;
43 case ONIG_NO_SUPPORT_CONFIG:
44 p = "no support in this configuration"; break;
45 case ONIG_ABORT:
46 p = "abort"; break;
47 case ONIGERR_MEMORY:
48 p = "fail to memory allocation"; break;
49 case ONIGERR_MATCH_STACK_LIMIT_OVER:
50 p = "match-stack limit over"; break;
51 case ONIGERR_PARSE_DEPTH_LIMIT_OVER:
52 p = "parse depth limit over"; break;
53 case ONIGERR_RETRY_LIMIT_IN_MATCH_OVER:
54 p = "retry-limit-in-match over"; break;
55 case ONIGERR_TYPE_BUG:
56 p = "undefined type (bug)"; break;
57 case ONIGERR_PARSER_BUG:
58 p = "internal parser error (bug)"; break;
59 case ONIGERR_STACK_BUG:
60 p = "stack error (bug)"; break;
61 case ONIGERR_UNDEFINED_BYTECODE:
62 p = "undefined bytecode (bug)"; break;
63 case ONIGERR_UNEXPECTED_BYTECODE:
64 p = "unexpected bytecode (bug)"; break;
65 case ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED:
66 p = "default multibyte-encoding is not setted"; break;
67 case ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR:
68 p = "can't convert to wide-char on specified multibyte-encoding"; break;
69 case ONIGERR_FAIL_TO_INITIALIZE:
70 p = "fail to initialize"; break;
71 case ONIGERR_INVALID_ARGUMENT:
72 p = "invalid argument"; break;
73 case ONIGERR_END_PATTERN_AT_LEFT_BRACE:
74 p = "end pattern at left brace"; break;
75 case ONIGERR_END_PATTERN_AT_LEFT_BRACKET:
76 p = "end pattern at left bracket"; break;
77 case ONIGERR_EMPTY_CHAR_CLASS:
78 p = "empty char-class"; break;
79 case ONIGERR_PREMATURE_END_OF_CHAR_CLASS:
80 p = "premature end of char-class"; break;
81 case ONIGERR_END_PATTERN_AT_ESCAPE:
82 p = "end pattern at escape"; break;
83 case ONIGERR_END_PATTERN_AT_META:
84 p = "end pattern at meta"; break;
85 case ONIGERR_END_PATTERN_AT_CONTROL:
86 p = "end pattern at control"; break;
87 case ONIGERR_META_CODE_SYNTAX:
88 p = "invalid meta-code syntax"; break;
89 case ONIGERR_CONTROL_CODE_SYNTAX:
90 p = "invalid control-code syntax"; break;
91 case ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE:
92 p = "char-class value at end of range"; break;
93 case ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE:
94 p = "char-class value at start of range"; break;
95 case ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS:
96 p = "unmatched range specifier in char-class"; break;
97 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED:
98 p = "target of repeat operator is not specified"; break;
99 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID:
100 p = "target of repeat operator is invalid"; break;
101 case ONIGERR_NESTED_REPEAT_OPERATOR:
102 p = "nested repeat operator"; break;
103 case ONIGERR_UNMATCHED_CLOSE_PARENTHESIS:
104 p = "unmatched close parenthesis"; break;
105 case ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS:
106 p = "end pattern with unmatched parenthesis"; break;
107 case ONIGERR_END_PATTERN_IN_GROUP:
108 p = "end pattern in group"; break;
109 case ONIGERR_UNDEFINED_GROUP_OPTION:
110 p = "undefined group option"; break;
111 case ONIGERR_INVALID_POSIX_BRACKET_TYPE:
112 p = "invalid POSIX bracket type"; break;
113 case ONIGERR_INVALID_LOOK_BEHIND_PATTERN:
114 p = "invalid pattern in look-behind"; break;
115 case ONIGERR_INVALID_REPEAT_RANGE_PATTERN:
116 p = "invalid repeat range {lower,upper}"; break;
117 case ONIGERR_TOO_BIG_NUMBER:
118 p = "too big number"; break;
119 case ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE:
120 p = "too big number for repeat range"; break;
121 case ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE:
122 p = "upper is smaller than lower in repeat range"; break;
123 case ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS:
124 p = "empty range in char class"; break;
125 case ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE:
126 p = "mismatch multibyte code length in char-class range"; break;
127 case ONIGERR_TOO_MANY_MULTI_BYTE_RANGES:
128 p = "too many multibyte code ranges are specified"; break;
129 case ONIGERR_TOO_SHORT_MULTI_BYTE_STRING:
130 p = "too short multibyte code string"; break;
131 case ONIGERR_TOO_BIG_BACKREF_NUMBER:
132 p = "too big backref number"; break;
133 case ONIGERR_INVALID_BACKREF:
134 p = "invalid backref number/name"; break;
135 case ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED:
136 p = "numbered backref/call is not allowed. (use name)"; break;
137 case ONIGERR_TOO_MANY_CAPTURES:
138 p = "too many captures"; break;
139 case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE:
140 p = "too big wide-char value"; break;
141 case ONIGERR_TOO_LONG_WIDE_CHAR_VALUE:
142 p = "too long wide-char value"; break;
143 case ONIGERR_INVALID_CODE_POINT_VALUE:
144 p = "invalid code point value"; break;
145 case ONIGERR_EMPTY_GROUP_NAME:
146 p = "group name is empty"; break;
147 case ONIGERR_INVALID_GROUP_NAME:
148 p = "invalid group name <%n>"; break;
149 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
150 p = "invalid char in group name <%n>"; break;
151 case ONIGERR_UNDEFINED_NAME_REFERENCE:
152 p = "undefined name <%n> reference"; break;
153 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
154 p = "undefined group <%n> reference"; break;
155 case ONIGERR_MULTIPLEX_DEFINED_NAME:
156 p = "multiplex defined name <%n>"; break;
157 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
158 p = "multiplex definition name <%n> call"; break;
159 case ONIGERR_NEVER_ENDING_RECURSION:
160 p = "never ending recursion"; break;
161 case ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY:
162 p = "group number is too big for capture history"; break;
163 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
164 p = "invalid character property name {%n}"; break;
165 case ONIGERR_INVALID_IF_ELSE_SYNTAX:
166 p = "invalid if-else syntax"; break;
167 case ONIGERR_INVALID_ABSENT_GROUP_PATTERN:
168 p = "invalid absent group pattern"; break;
169 case ONIGERR_INVALID_ABSENT_GROUP_GENERATOR_PATTERN:
170 p = "invalid absent group generator pattern"; break;
171 case ONIGERR_INVALID_CALLOUT_PATTERN:
172 p = "invalid callout pattern"; break;
173 case ONIGERR_INVALID_CALLOUT_NAME:
174 p = "invalid callout name"; break;
175 case ONIGERR_UNDEFINED_CALLOUT_NAME:
176 p = "undefined callout name"; break;
177 case ONIGERR_INVALID_CALLOUT_BODY:
178 p = "invalid callout body"; break;
179 case ONIGERR_INVALID_CALLOUT_TAG_NAME:
180 p = "invalid callout tag name"; break;
181 case ONIGERR_INVALID_CALLOUT_ARG:
182 p = "invalid callout arg"; break;
183 case ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION:
184 p = "not supported encoding combination"; break;
185 case ONIGERR_INVALID_COMBINATION_OF_OPTIONS:
186 p = "invalid combination of options"; break;
187 case ONIGERR_LIBRARY_IS_NOT_INITIALIZED:
188 p = "library is not initialized"; break;
189
190 default:
191 p = "undefined error code"; break;
192 }
193
194 return (UChar* )p;
195 }
196
sprint_byte(char * s,unsigned int v)197 static void sprint_byte(char* s, unsigned int v)
198 {
199 xsnprintf(s, 3, "%02x", (v & 0377));
200 }
201
sprint_byte_with_x(char * s,unsigned int v)202 static void sprint_byte_with_x(char* s, unsigned int v)
203 {
204 xsnprintf(s, 5, "\\x%02x", (v & 0377));
205 }
206
to_ascii(OnigEncoding enc,UChar * s,UChar * end,UChar buf[],int buf_size,int * is_over)207 static int to_ascii(OnigEncoding enc, UChar *s, UChar *end,
208 UChar buf[], int buf_size, int *is_over)
209 {
210 int len;
211 UChar *p;
212 OnigCodePoint code;
213
214 if (!s) {
215 len = 0;
216 *is_over = 0;
217 }
218 else if (ONIGENC_MBC_MINLEN(enc) > 1) {
219 p = s;
220 len = 0;
221 while (p < end) {
222 code = ONIGENC_MBC_TO_CODE(enc, p, end);
223 if (code >= 0x80) {
224 if (code > 0xffff && len + 10 <= buf_size) {
225 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 24));
226 sprint_byte((char*)(&(buf[len+4])), (unsigned int)(code >> 16));
227 sprint_byte((char*)(&(buf[len+6])), (unsigned int)(code >> 8));
228 sprint_byte((char*)(&(buf[len+8])), (unsigned int)code);
229 len += 10;
230 }
231 else if (len + 6 <= buf_size) {
232 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 8));
233 sprint_byte((char*)(&(buf[len+4])), (unsigned int)code);
234 len += 6;
235 }
236 else {
237 break;
238 }
239 }
240 else {
241 buf[len++] = (UChar )code;
242 }
243
244 p += enclen(enc, p);
245 if (len >= buf_size) break;
246 }
247
248 *is_over = p < end;
249 }
250 else {
251 len = MIN((int )(end - s), buf_size);
252 xmemcpy(buf, s, (size_t )len);
253 *is_over = ((buf_size < (end - s)) ? 1 : 0);
254 }
255
256 return len;
257 }
258
259
260 extern int
onig_is_error_code_needs_param(int code)261 onig_is_error_code_needs_param(int code)
262 {
263 switch (code) {
264 case ONIGERR_UNDEFINED_NAME_REFERENCE:
265 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
266 case ONIGERR_MULTIPLEX_DEFINED_NAME:
267 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
268 case ONIGERR_INVALID_GROUP_NAME:
269 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
270 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
271 return 1;
272 default:
273 return 0;
274 }
275 }
276
277 /* for ONIG_MAX_ERROR_MESSAGE_LEN */
278 #define MAX_ERROR_PAR_LEN 30
279
onig_error_code_to_str(UChar * s,int code,...)280 extern int onig_error_code_to_str(UChar* s, int code, ...)
281 {
282 UChar *p, *q;
283 OnigErrorInfo* einfo;
284 int len, is_over;
285 UChar parbuf[MAX_ERROR_PAR_LEN];
286 va_list vargs;
287
288 va_start(vargs, code);
289
290 switch (code) {
291 case ONIGERR_UNDEFINED_NAME_REFERENCE:
292 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
293 case ONIGERR_MULTIPLEX_DEFINED_NAME:
294 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
295 case ONIGERR_INVALID_GROUP_NAME:
296 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
297 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
298 einfo = va_arg(vargs, OnigErrorInfo*);
299 len = to_ascii(einfo->enc, einfo->par, einfo->par_end,
300 parbuf, MAX_ERROR_PAR_LEN - 3, &is_over);
301 q = onig_error_code_to_format(code);
302 p = s;
303 while (*q != '\0') {
304 if (*q == '%') {
305 q++;
306 if (*q == 'n') { /* '%n': name */
307 xmemcpy(p, parbuf, len);
308 p += len;
309 if (is_over != 0) {
310 xmemcpy(p, "...", 3);
311 p += 3;
312 }
313 q++;
314 }
315 else
316 goto normal_char;
317 }
318 else {
319 normal_char:
320 *p++ = *q++;
321 }
322 }
323 *p = '\0';
324 len = (int )(p - s);
325 break;
326
327 default:
328 q = onig_error_code_to_format(code);
329 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, q);
330 xmemcpy(s, q, len);
331 s[len] = '\0';
332 break;
333 }
334
335 va_end(vargs);
336 return len;
337 }
338
339
onig_snprintf_with_pattern(UChar buf[],int bufsize,OnigEncoding enc,UChar * pat,UChar * pat_end,const UChar * fmt,...)340 void onig_snprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
341 UChar* pat, UChar* pat_end, const UChar *fmt, ...)
342 {
343 int n, need, len;
344 UChar *p, *s, *bp;
345 UChar bs[6];
346 va_list args;
347
348 va_start(args, fmt);
349 n = xvsnprintf((char* )buf, bufsize, (const char* )fmt, args);
350 va_end(args);
351
352 need = (int )(pat_end - pat) * 4 + 4;
353
354 if (n + need < bufsize) {
355 xstrcat((char* )buf, ": /", bufsize);
356 s = buf + onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, buf);
357
358 p = pat;
359 while (p < pat_end) {
360 if (ONIGENC_IS_MBC_HEAD(enc, p)) {
361 len = enclen(enc, p);
362 if (ONIGENC_MBC_MINLEN(enc) == 1) {
363 while (len-- > 0) *s++ = *p++;
364 }
365 else { /* for UTF16/32 */
366 int blen;
367
368 while (len-- > 0) {
369 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
370 blen = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
371 bp = bs;
372 while (blen-- > 0) *s++ = *bp++;
373 }
374 }
375 }
376 else if (*p == '\\') {
377 *s++ = *p++;
378 len = enclen(enc, p);
379 while (len-- > 0) *s++ = *p++;
380 }
381 else if (*p == '/') {
382 *s++ = (unsigned char )'\\';
383 *s++ = *p++;
384 }
385 else if (!ONIGENC_IS_CODE_PRINT(enc, *p) &&
386 !ONIGENC_IS_CODE_SPACE(enc, *p)) {
387 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
388 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
389 bp = bs;
390 while (len-- > 0) *s++ = *bp++;
391 }
392 else {
393 *s++ = *p++;
394 }
395 }
396
397 *s++ = '/';
398 *s = '\0';
399 }
400 }
401
