Ruby 4.1.0dev (2026-05-17 revision ba43e07c12093273e46630c55e93c7491dbbd372)
regerror.c (ba43e07c12093273e46630c55e93c7491dbbd372)
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) 2011-2019 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
36extern UChar*
37onig_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#if 0
67 case ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR:
68 p = "can't convert to wide-char on specified multibyte-encoding"; break;
69#endif
70 case ONIGERR_INVALID_ARGUMENT:
71 p = "invalid argument"; break;
72 case ONIGERR_END_PATTERN_AT_LEFT_BRACE:
73 p = "end pattern at left brace"; break;
74#if 0
75 case ONIGERR_END_PATTERN_AT_LEFT_BRACKET:
76 p = "end pattern at left bracket"; break;
77#endif
78 case ONIGERR_EMPTY_CHAR_CLASS:
79 p = "empty char-class"; break;
80 case ONIGERR_PREMATURE_END_OF_CHAR_CLASS:
81 p = "premature end of char-class"; break;
82 case ONIGERR_END_PATTERN_AT_ESCAPE:
83 p = "end pattern at escape"; break;
84 case ONIGERR_END_PATTERN_AT_META:
85 p = "end pattern at meta"; break;
86 case ONIGERR_END_PATTERN_AT_CONTROL:
87 p = "end pattern at control"; break;
88 case ONIGERR_META_CODE_SYNTAX:
89 p = "invalid meta-code syntax"; break;
90 case ONIGERR_CONTROL_CODE_SYNTAX:
91 p = "invalid control-code syntax"; break;
92 case ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE:
93 p = "char-class value at end of range"; break;
94#if 0
95 case ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE:
96 p = "char-class value at start of range"; break;
97#endif
98 case ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS:
99 p = "unmatched range specifier in char-class"; break;
100 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED:
101 p = "target of repeat operator is not specified"; break;
102 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID:
103 p = "target of repeat operator is invalid"; break;
104#if 0
105 case ONIGERR_NESTED_REPEAT_OPERATOR:
106 p = "nested repeat operator"; break;
107#endif
108 case ONIGERR_UNMATCHED_CLOSE_PARENTHESIS:
109 p = "unmatched close parenthesis"; break;
110 case ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS:
111 p = "end pattern with unmatched parenthesis"; break;
112 case ONIGERR_END_PATTERN_IN_GROUP:
113 p = "end pattern in group"; break;
114 case ONIGERR_UNDEFINED_GROUP_OPTION:
115 p = "undefined group option"; break;
116 case ONIGERR_INVALID_POSIX_BRACKET_TYPE:
117 p = "invalid POSIX bracket type"; break;
118 case ONIGERR_INVALID_LOOK_BEHIND_PATTERN:
119 p = "invalid pattern in look-behind"; break;
120 case ONIGERR_INVALID_REPEAT_RANGE_PATTERN:
121 p = "invalid repeat range {lower,upper}"; break;
122 case ONIGERR_INVALID_CONDITION_PATTERN:
123 p = "invalid conditional pattern"; break;
124 case ONIGERR_TOO_BIG_NUMBER:
125 p = "too big number"; break;
126 case ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE:
127 p = "too big number for repeat range"; break;
128 case ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE:
129 p = "upper is smaller than lower in repeat range"; break;
130 case ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS:
131 p = "empty range in char class"; break;
132#if 0
133 case ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE:
134 p = "mismatch multibyte code length in char-class range"; break;
135#endif
136 case ONIGERR_TOO_MANY_MULTI_BYTE_RANGES:
137 p = "too many multibyte code ranges are specified"; break;
138 case ONIGERR_TOO_SHORT_MULTI_BYTE_STRING:
139 p = "too short multibyte code string"; break;
140#if 0
141 case ONIGERR_TOO_BIG_BACKREF_NUMBER:
142 p = "too big backref number"; break;
143#endif
144 case ONIGERR_INVALID_BACKREF:
145#ifdef USE_NAMED_GROUP
146 p = "invalid backref number/name"; break;
147#else
148 p = "invalid backref number"; break;
149#endif
150 case ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED:
151 p = "numbered backref/call is not allowed. (use name)"; break;
152 case ONIGERR_TOO_SHORT_DIGITS:
153 p = "too short digits"; break;
154 case ONIGERR_TOO_LONG_WIDE_CHAR_VALUE:
155 p = "too long wide-char value"; break;
156 case ONIGERR_EMPTY_GROUP_NAME:
157 p = "group name is empty"; break;
158 case ONIGERR_INVALID_GROUP_NAME:
159 p = "invalid group name <%n>"; break;
160 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
161#ifdef USE_NAMED_GROUP
162 p = "invalid char in group name <%n>"; break;
163#else
164 p = "invalid char in group number <%n>"; break;
165#endif
166 case ONIGERR_UNDEFINED_NAME_REFERENCE:
167 p = "undefined name <%n> reference"; break;
168 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
169 p = "undefined group <%n> reference"; break;
170 case ONIGERR_MULTIPLEX_DEFINED_NAME:
171 p = "multiplex defined name <%n>"; break;
172 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
173 p = "multiplex definition name <%n> call"; break;
174 case ONIGERR_NEVER_ENDING_RECURSION:
175 p = "never ending recursion"; break;
176#ifdef USE_CAPTURE_HISTORY
177 case ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY:
178 p = "group number is too big for capture history"; break;
179#endif
180 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
181 p = "invalid character property name {%n}"; break;
182 case ONIGERR_TOO_MANY_RANGE_REPEAT:
183 p = "too many range repeat"; break;
184 case ONIGERR_TOO_MANY_NULL_CHECK:
185 p = "too many null check"; break;
186 case ONIGERR_TOO_MANY_CAPTURE_GROUPS:
187 p = "too many capture groups are specified"; break;
188 case ONIGERR_INVALID_CODE_POINT_VALUE:
189 p = "invalid code point value"; break;
190 case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE:
191 p = "too big wide-char value"; break;
192 case ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION:
193 p = "not supported encoding combination"; break;
194 case ONIGERR_INVALID_COMBINATION_OF_OPTIONS:
195 p = "invalid combination of options"; break;
196
197 default:
198 p = "undefined error code"; break;
199 }
200
201 return (UChar* )p;
202}
203
204static void sprint_byte(char* s, unsigned int v)
205{
206 xsnprintf(s, 3, "%02x", (v & 0377));
207}
208
209static void sprint_byte_with_x(char* s, unsigned int v)
210{
211 xsnprintf(s, 5, "\\x%02x", (v & 0377));
212}
213
214static int to_ascii(OnigEncoding enc, UChar *s, UChar *end,
215 UChar buf[], int buf_size, int *is_over)
216{
217 int len;
218 UChar *p;
219 OnigCodePoint code;
220
221 if (ONIGENC_MBC_MINLEN(enc) > 1) {
222 p = s;
223 len = 0;
224 while (p < end) {
225 code = ONIGENC_MBC_TO_CODE(enc, p, end);
226 if (code >= 0x80) {
227 if (code > 0xffff && len + 10 <= buf_size) {
228 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 24));
229 sprint_byte((char*)(&(buf[len+4])), (unsigned int)(code >> 16));
230 sprint_byte((char*)(&(buf[len+6])), (unsigned int)(code >> 8));
231 sprint_byte((char*)(&(buf[len+8])), (unsigned int)code);
232 len += 10;
233 }
234 else if (len + 6 <= buf_size) {
235 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 8));
236 sprint_byte((char*)(&(buf[len+4])), (unsigned int)code);
237 len += 6;
238 }
239 else {
240 break;
241 }
242 }
243 else {
244 buf[len++] = (UChar )code;
245 }
246
247 p += enclen(enc, p, end);
248 if (len >= buf_size) break;
249 }
250
251 *is_over = ((p < end) ? 1 : 0);
252 }
253 else {
254 len = (int )MIN((end - s), buf_size);
255 xmemcpy(buf, s, (size_t )len);
256 *is_over = ((buf_size < (end - s)) ? 1 : 0);
257 }
258
259 return len;
260}
261
262
263/* < ONIG_MAX_ERROR_MESSAGE_LEN - max length of messages with %n */
264#define MAX_ERROR_PAR_LEN 50
265
266extern int
267onig_error_code_to_str(UChar* s, OnigPosition code, ...)
268{
269 UChar *p, *q;
270 OnigErrorInfo* einfo;
271 size_t len;
272 int is_over;
273 UChar parbuf[MAX_ERROR_PAR_LEN];
274 va_list vargs;
275
276 va_start(vargs, code);
277
278 switch (code) {
279 case ONIGERR_UNDEFINED_NAME_REFERENCE:
280 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
281 case ONIGERR_MULTIPLEX_DEFINED_NAME:
282 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
283 case ONIGERR_INVALID_GROUP_NAME:
284 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
285 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
286 einfo = va_arg(vargs, OnigErrorInfo*);
287 len = to_ascii(einfo->enc, einfo->par, einfo->par_end,
288 parbuf, MAX_ERROR_PAR_LEN - 3, &is_over);
289 q = onig_error_code_to_format(code);
290 p = s;
291 while (*q != '\0') {
292 if (*q == '%') {
293 q++;
294 if (*q == 'n') { /* '%n': name */
295 xmemcpy(p, parbuf, len);
296 p += len;
297 if (is_over != 0) {
298 xmemcpy(p, "...", 3);
299 p += 3;
300 }
301 q++;
302 }
303 else
304 goto normal_char;
305 }
306 else {
307 normal_char:
308 *p++ = *q++;
309 }
310 }
311 *p = '\0';
312 len = p - s;
313 break;
314
315 default:
316 q = onig_error_code_to_format(code);
317 if (q) {
318 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, q);
319 xmemcpy(s, q, len);
320 }
321 else {
322 len = 0;
323 }
324 s[len] = '\0';
325 break;
326 }
327
328 va_end(vargs);
329 return (int )len;
330}
331
332void
333onig_vsnprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
334 UChar* pat, UChar* pat_end, const char *fmt, va_list args)
335{
336 size_t need;
337 int n, len;
338 UChar *p, *s, *bp;
339 UChar bs[6];
340
341 n = xvsnprintf((char* )buf, bufsize, (const char* )fmt, args);
342
343 need = (pat_end - pat) * 4 + 4;
344
345 if (n + need < (size_t )bufsize) {
346 static const char sep[] = ": /";
347 memcpy((char* )buf + n, sep, sizeof(sep));
348 s = buf + onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, buf);
349
350 p = pat;
351 while (p < pat_end) {
352 if (ONIGENC_IS_MBC_HEAD(enc, p, pat_end)) {
353 len = enclen(enc, p, pat_end);
354 if (ONIGENC_MBC_MINLEN(enc) == 1) {
355 while (len-- > 0) *s++ = *p++;
356 }
357 else { /* for UTF16/32 */
358 int blen;
359
360 while (len-- > 0) {
361 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
362 blen = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
363 bp = bs;
364 while (blen-- > 0) *s++ = *bp++;
365 }
366 }
367 }
368 else if (*p == '\\') {
369 *s++ = *p++;
370 len = enclen(enc, p, pat_end);
371 while (len-- > 0) *s++ = *p++;
372 }
373 else if (*p == '/') {
374 *s++ = (unsigned char )'\\';
375 *s++ = *p++;
376 }
377 else if (!ONIGENC_IS_CODE_PRINT(enc, *p) &&
378 (!ONIGENC_IS_CODE_SPACE(enc, *p) ||
379 ONIGENC_IS_CODE_CNTRL(enc, *p))) {
380 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
381 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
382 bp = bs;
383 while (len-- > 0) *s++ = *bp++;
384 }
385 else {
386 *s++ = *p++;
387 }
388 }
389
390 *s++ = '/';
391 *s = '\0';
392 }
393}
394
395#if 0 /* unused */
396void
397onig_snprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
398 UChar* pat, UChar* pat_end, const char *fmt, ...)
399{
400 va_list args;
401 va_start(args, fmt);
402 onig_vsnprintf_with_pattern(buf, bufsize, enc,
403 pat, pat_end, fmt, args);
404 va_end(args);
405}
406#endif
int len
Length of the buffer.
Definition io.h:8
Defines old _.