1 /**********************************************************************
2 regerror.c - Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5 * Copyright (c) 2002-2007 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>
6 * All rights reserved.
7 *
8 * (C) Copyright 2015 Hewlett Packard Enterprise Development LP<BR>
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 */
31
32 #define HAVE_STDARG_PROTOTYPES
33
34 #include "regint.h"
35
36 #if 0
37 #include <stdio.h> /* for vsnprintf() */
38
39 #ifdef HAVE_STDARG_PROTOTYPES
40 #include <stdarg.h>
41 #define va_init_list(a,b) va_start(a,b)
42 #else
43 #include <varargs.h>
44 #define va_init_list(a,b) va_start(a)
45 #endif
46 #endif
47
48 extern UChar*
onig_error_code_to_format(int code)49 onig_error_code_to_format(int code)
50 {
51 char *p;
52
53 if (code >= 0) return (UChar* )0;
54
55 switch (code) {
56 case ONIG_MISMATCH:
57 p = "mismatch"; break;
58 case ONIG_NO_SUPPORT_CONFIG:
59 p = "no support in this configuration"; break;
60 case ONIGERR_MEMORY:
61 p = "fail to memory allocation"; break;
62 case ONIGERR_MATCH_STACK_LIMIT_OVER:
63 p = "match-stack limit over"; break;
64 case ONIGERR_TYPE_BUG:
65 p = "undefined type (bug)"; break;
66 case ONIGERR_PARSER_BUG:
67 p = "internal parser error (bug)"; break;
68 case ONIGERR_STACK_BUG:
69 p = "stack error (bug)"; break;
70 case ONIGERR_UNDEFINED_BYTECODE:
71 p = "undefined bytecode (bug)"; break;
72 case ONIGERR_UNEXPECTED_BYTECODE:
73 p = "unexpected bytecode (bug)"; break;
74 case ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED:
75 p = "default multibyte-encoding is not setted"; break;
76 case ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR:
77 p = "can't convert to wide-char on specified multibyte-encoding"; break;
78 case ONIGERR_INVALID_ARGUMENT:
79 p = "invalid argument"; break;
80 case ONIGERR_END_PATTERN_AT_LEFT_BRACE:
81 p = "end pattern at left brace"; break;
82 case ONIGERR_END_PATTERN_AT_LEFT_BRACKET:
83 p = "end pattern at left bracket"; break;
84 case ONIGERR_EMPTY_CHAR_CLASS:
85 p = "empty char-class"; break;
86 case ONIGERR_PREMATURE_END_OF_CHAR_CLASS:
87 p = "premature end of char-class"; break;
88 case ONIGERR_END_PATTERN_AT_ESCAPE:
89 p = "end pattern at escape"; break;
90 case ONIGERR_END_PATTERN_AT_META:
91 p = "end pattern at meta"; break;
92 case ONIGERR_END_PATTERN_AT_CONTROL:
93 p = "end pattern at control"; break;
94 case ONIGERR_META_CODE_SYNTAX:
95 p = "invalid meta-code syntax"; break;
96 case ONIGERR_CONTROL_CODE_SYNTAX:
97 p = "invalid control-code syntax"; break;
98 case ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE:
99 p = "char-class value at end of range"; break;
100 case ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE:
101 p = "char-class value at start of range"; break;
102 case ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS:
103 p = "unmatched range specifier in char-class"; break;
104 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED:
105 p = "target of repeat operator is not specified"; break;
106 case ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID:
107 p = "target of repeat operator is invalid"; break;
108 case ONIGERR_NESTED_REPEAT_OPERATOR:
109 p = "nested repeat operator"; break;
110 case ONIGERR_UNMATCHED_CLOSE_PARENTHESIS:
111 p = "unmatched close parenthesis"; break;
112 case ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS:
113 p = "end pattern with unmatched parenthesis"; break;
114 case ONIGERR_END_PATTERN_IN_GROUP:
115 p = "end pattern in group"; break;
116 case ONIGERR_UNDEFINED_GROUP_OPTION:
117 p = "undefined group option"; break;
118 case ONIGERR_INVALID_POSIX_BRACKET_TYPE:
119 p = "invalid POSIX bracket type"; break;
120 case ONIGERR_INVALID_LOOK_BEHIND_PATTERN:
121 p = "invalid pattern in look-behind"; break;
122 case ONIGERR_INVALID_REPEAT_RANGE_PATTERN:
123 p = "invalid repeat range {lower,upper}"; 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 case ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE:
133 p = "mismatch multibyte code length in char-class range"; break;
134 case ONIGERR_TOO_MANY_MULTI_BYTE_RANGES:
135 p = "too many multibyte code ranges are specified"; break;
136 case ONIGERR_TOO_SHORT_MULTI_BYTE_STRING:
137 p = "too short multibyte code string"; break;
138 case ONIGERR_TOO_BIG_BACKREF_NUMBER:
139 p = "too big backref number"; break;
140 case ONIGERR_INVALID_BACKREF:
141 #ifdef USE_NAMED_GROUP
142 p = "invalid backref number/name"; break;
143 #else
144 p = "invalid backref number"; break;
145 #endif
146 case ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED:
147 p = "numbered backref/call is not allowed. (use name)"; break;
148 case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE:
149 p = "too big wide-char value"; break;
150 case ONIGERR_TOO_LONG_WIDE_CHAR_VALUE:
151 p = "too long wide-char value"; break;
152 case ONIGERR_INVALID_CODE_POINT_VALUE:
153 p = "invalid code point value"; break;
154 case ONIGERR_EMPTY_GROUP_NAME:
155 p = "group name is empty"; break;
156 case ONIGERR_INVALID_GROUP_NAME:
157 p = "invalid group name <%n>"; break;
158 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
159 #ifdef USE_NAMED_GROUP
160 p = "invalid char in group name <%n>"; break;
161 #else
162 p = "invalid char in group number <%n>"; break;
163 #endif
164 case ONIGERR_UNDEFINED_NAME_REFERENCE:
165 p = "undefined name <%n> reference"; break;
166 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
167 p = "undefined group <%n> reference"; break;
168 case ONIGERR_MULTIPLEX_DEFINED_NAME:
169 p = "multiplex defined name <%n>"; break;
170 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
171 p = "multiplex definition name <%n> call"; break;
172 case ONIGERR_NEVER_ENDING_RECURSION:
173 p = "never ending recursion"; break;
174 case ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY:
175 p = "group number is too big for capture history"; break;
176 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
177 p = "invalid character property name {%n}"; break;
178 case ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION:
179 p = "not supported encoding combination"; break;
180 case ONIGERR_INVALID_COMBINATION_OF_OPTIONS:
181 p = "invalid combination of options"; break;
182 case ONIGERR_OVER_THREAD_PASS_LIMIT_COUNT:
183 p = "over thread pass limit count"; break;
184
185 default:
186 p = "undefined error code"; break;
187 }
188
189 return (UChar* )p;
190 }
191
sprint_byte(char * s,unsigned int v)192 static void sprint_byte(char* s, unsigned int v)
193 {
194 sprintf_s(s, sizeof("00"), "%02x", (v & 0377));
195 }
196
sprint_byte_with_x(char * s,unsigned int v)197 static void sprint_byte_with_x(char* s, unsigned int v)
198 {
199 sprintf_s(s, sizeof("\\x00"), "\\x%02x", (v & 0377));
200 }
201
to_ascii(OnigEncoding enc,UChar * s,UChar * end,UChar buf[],int buf_size,int * is_over)202 static int to_ascii(OnigEncoding enc, UChar *s, UChar *end,
203 UChar buf[], int buf_size, int *is_over)
204 {
205 int len;
206 UChar *p;
207 OnigCodePoint code;
208
209 if (ONIGENC_MBC_MINLEN(enc) > 1) {
210 p = s;
211 len = 0;
212 while (p < end) {
213 code = ONIGENC_MBC_TO_CODE(enc, p, end);
214 if (code >= 0x80) {
215 if (code > 0xffff && len + 10 <= buf_size) {
216 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 24));
217 sprint_byte((char*)(&(buf[len+4])), (unsigned int)(code >> 16));
218 sprint_byte((char*)(&(buf[len+6])), (unsigned int)(code >> 8));
219 sprint_byte((char*)(&(buf[len+8])), (unsigned int)code);
220 len += 10;
221 }
222 else if (len + 6 <= buf_size) {
223 sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 8));
224 sprint_byte((char*)(&(buf[len+4])), (unsigned int)code);
225 len += 6;
226 }
227 else {
228 break;
229 }
230 }
231 else {
232 buf[len++] = (UChar )code;
233 }
234
235 p += enclen(enc, p);
236 if (len >= buf_size) break;
237 }
238
239 *is_over = ((p < end) ? 1 : 0);
240 }
241 else {
242 len = MIN((int)(end - s), buf_size);
243 xmemcpy(buf, s, (size_t )len);
244 *is_over = ((buf_size < (end - s)) ? 1 : 0);
245 }
246
247 return len;
248 }
249
250
251 /* for ONIG_MAX_ERROR_MESSAGE_LEN */
252 #define MAX_ERROR_PAR_LEN 30
253
254 extern int
255 #ifdef HAVE_STDARG_PROTOTYPES
onig_error_code_to_str(UChar * s,int code,...)256 onig_error_code_to_str(UChar* s, int code, ...)
257 #else
258 onig_error_code_to_str(s, code, va_alist)
259 UChar* s;
260 int code;
261 va_dcl
262 #endif
263 {
264 UChar *p, *q;
265 OnigErrorInfo* einfo;
266 int len, is_over;
267 UChar parbuf[MAX_ERROR_PAR_LEN];
268 va_list vargs;
269
270 va_init_list(vargs, code);
271
272 switch (code) {
273 case ONIGERR_UNDEFINED_NAME_REFERENCE:
274 case ONIGERR_UNDEFINED_GROUP_REFERENCE:
275 case ONIGERR_MULTIPLEX_DEFINED_NAME:
276 case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
277 case ONIGERR_INVALID_GROUP_NAME:
278 case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
279 case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
280 einfo = va_arg(vargs, OnigErrorInfo*);
281 len = to_ascii(einfo->enc, einfo->par, einfo->par_end,
282 parbuf, MAX_ERROR_PAR_LEN - 3, &is_over);
283 q = onig_error_code_to_format(code);
284 p = s;
285 while (*q != '\0') {
286 if (*q == '%') {
287 q++;
288 if (*q == 'n') { /* '%n': name */
289 xmemcpy(p, parbuf, len);
290 p += len;
291 if (is_over != 0) {
292 xmemcpy(p, "...", 3);
293 p += 3;
294 }
295 q++;
296 }
297 else
298 goto normal_char;
299 }
300 else {
301 normal_char:
302 *p++ = *q++;
303 }
304 }
305 *p = '\0';
306 len = (int)(p - s);
307 break;
308
309 default:
310 q = onig_error_code_to_format(code);
311 if (q == NULL) {
312 len = 0;
313 break;
314 }
315 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, q);
316 xmemcpy(s, q, len);
317 s[len] = '\0';
318 break;
319 }
320
321 va_end(vargs);
322 return len;
323 }
324
325
326 void
327 #ifdef HAVE_STDARG_PROTOTYPES
onig_snprintf_with_pattern(UChar buf[],int bufsize,OnigEncoding enc,UChar * pat,UChar * pat_end,const UChar * fmt,...)328 onig_snprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
329 UChar* pat, UChar* pat_end, const UChar *fmt, ...)
330 #else
331 onig_snprintf_with_pattern(buf, bufsize, enc, pat, pat_end, fmt, va_alist)
332 UChar buf[];
333 int bufsize;
334 OnigEncoding enc;
335 UChar* pat;
336 UChar* pat_end;
337 const UChar *fmt;
338 va_dcl
339 #endif
340 {
341 int n, need, len;
342 UChar *p, *s, *bp;
343 UChar bs[6];
344 va_list args;
345
346 va_init_list(args, fmt);
347 n = xvsnprintf((char* )buf, bufsize, (const char* )fmt, args);
348 va_end(args);
349
350 need = (int)(pat_end - pat) * 4 + 4;
351
352 if (n + need < bufsize) {
353 strcat_s((char* )buf, bufsize, ": /");
354 s = buf + onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, buf);
355
356 p = pat;
357 while (p < pat_end) {
358 if (*p == '\\') {
359 *s++ = *p++;
360 len = enclen(enc, p);
361 while (len-- > 0) *s++ = *p++;
362 }
363 else if (*p == '/') {
364 *s++ = (unsigned char )'\\';
365 *s++ = *p++;
366 }
367 else if (ONIGENC_IS_MBC_HEAD(enc, p)) {
368 len = enclen(enc, p);
369 if (ONIGENC_MBC_MINLEN(enc) == 1) {
370 while (len-- > 0) *s++ = *p++;
371 }
372 else { /* for UTF16 */
373 int blen;
374
375 while (len-- > 0) {
376 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
377 blen = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
378 bp = bs;
379 while (blen-- > 0) *s++ = *bp++;
380 }
381 }
382 }
383 else if (!ONIGENC_IS_CODE_PRINT(enc, *p) &&
384 !ONIGENC_IS_CODE_SPACE(enc, *p)) {
385 sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
386 len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
387 bp = bs;
388 while (len-- > 0) *s++ = *bp++;
389 }
390 else {
391 *s++ = *p++;
392 }
393 }
394
395 *s++ = '/';
396 *s = '\0';
397 }
398 }
399