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       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#ifdef HAVE_STDARG_PROTOTYPES
35#include <stdarg.h>
36#define va_init_list(a,b) va_start(a,b)
37#else
38#include <varargs.h>
39#define va_init_list(a,b) va_start(a)
40#endif
41
42extern UChar*
43onig_error_code_to_format(OnigPosition code)
44{
45  const char *p;
46
47  if (code >= 0) return (UChar* )0;
48
49  switch (code) {
50  case ONIG_MISMATCH:
51    p = "mismatch"; break;
52  case ONIG_NO_SUPPORT_CONFIG:
53    p = "no support in this configuration"; break;
54  case ONIGERR_MEMORY:
55    p = "failed to allocate memory"; break;
56  case ONIGERR_MATCH_STACK_LIMIT_OVER:
57    p = "match-stack limit over"; break;
58  case ONIGERR_TYPE_BUG:
59    p = "undefined type (bug)"; break;
60  case ONIGERR_PARSER_BUG:
61    p = "internal parser error (bug)"; break;
62  case ONIGERR_STACK_BUG:
63    p = "stack error (bug)"; break;
64  case ONIGERR_UNDEFINED_BYTECODE:
65    p = "undefined bytecode (bug)"; break;
66  case ONIGERR_UNEXPECTED_BYTECODE:
67    p = "unexpected bytecode (bug)"; break;
68  case ONIGERR_DEFAULT_ENCODING_IS_NOT_SET:
69    p = "default multibyte-encoding is not set"; break;
70  case ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR:
71    p = "can't convert to wide-char on specified multibyte-encoding"; break;
72  case ONIGERR_INVALID_ARGUMENT:
73    p = "invalid argument"; break;
74  case ONIGERR_END_PATTERN_AT_LEFT_BRACE:
75    p = "end pattern at left brace"; break;
76  case ONIGERR_END_PATTERN_AT_LEFT_BRACKET:
77    p = "end pattern at left bracket"; break;
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  case ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE:
95    p = "char-class value at start of range"; break;
96  case ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS:
97    p = "unmatched range specifier in char-class"; break;
98  case ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED:
99    p = "target of repeat operator is not specified"; break;
100  case ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID:
101    p = "target of repeat operator is invalid"; break;
102  case ONIGERR_NESTED_REPEAT_OPERATOR:
103    p = "nested repeat operator"; break;
104  case ONIGERR_UNMATCHED_CLOSE_PARENTHESIS:
105    p = "unmatched close parenthesis"; break;
106  case ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS:
107    p = "end pattern with unmatched parenthesis"; break;
108  case ONIGERR_END_PATTERN_IN_GROUP:
109    p = "end pattern in group"; break;
110  case ONIGERR_UNDEFINED_GROUP_OPTION:
111    p = "undefined group option"; break;
112  case ONIGERR_INVALID_POSIX_BRACKET_TYPE:
113    p = "invalid POSIX bracket type"; break;
114  case ONIGERR_INVALID_LOOK_BEHIND_PATTERN:
115    p = "invalid pattern in look-behind"; break;
116  case ONIGERR_INVALID_REPEAT_RANGE_PATTERN:
117    p = "invalid repeat range {lower,upper}"; break;
118  case ONIGERR_INVALID_CONDITION_PATTERN:
119    p = "invalid conditional pattern"; break;
120  case ONIGERR_TOO_BIG_NUMBER:
121    p = "too big number"; break;
122  case ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE:
123    p = "too big number for repeat range"; break;
124  case ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE:
125    p = "upper is smaller than lower in repeat range"; break;
126  case ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS:
127    p = "empty range in char class"; break;
128  case ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE:
129    p = "mismatch multibyte code length in char-class range"; break;
130  case ONIGERR_TOO_MANY_MULTI_BYTE_RANGES:
131    p = "too many multibyte code ranges are specified"; break;
132  case ONIGERR_TOO_SHORT_MULTI_BYTE_STRING:
133    p = "too short multibyte code string"; break;
134  case ONIGERR_TOO_BIG_BACKREF_NUMBER:
135    p = "too big backref number"; break;
136  case ONIGERR_INVALID_BACKREF:
137#ifdef USE_NAMED_GROUP
138    p = "invalid backref number/name"; break;
139#else
140    p = "invalid backref number"; break;
141#endif
142  case ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED:
143    p = "numbered backref/call is not allowed. (use name)"; break;
144  case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE:
145    p = "too big wide-char value"; break;
146  case ONIGERR_TOO_SHORT_DIGITS:
147    p = "too short digits"; break;
148  case ONIGERR_TOO_LONG_WIDE_CHAR_VALUE:
149    p = "too long wide-char value"; break;
150  case ONIGERR_INVALID_CODE_POINT_VALUE:
151    p = "invalid code point value"; break;
152  case ONIGERR_EMPTY_GROUP_NAME:
153    p = "group name is empty"; break;
154  case ONIGERR_INVALID_GROUP_NAME:
155    p = "invalid group name <%n>"; break;
156  case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
157#ifdef USE_NAMED_GROUP
158    p = "invalid char in group name <%n>"; break;
159#else
160    p = "invalid char in group number <%n>"; break;
161#endif
162  case ONIGERR_UNDEFINED_NAME_REFERENCE:
163    p = "undefined name <%n> reference"; break;
164  case ONIGERR_UNDEFINED_GROUP_REFERENCE:
165    p = "undefined group <%n> reference"; break;
166  case ONIGERR_MULTIPLEX_DEFINED_NAME:
167    p = "multiplex defined name <%n>"; break;
168  case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
169    p = "multiplex definition name <%n> call"; break;
170  case ONIGERR_NEVER_ENDING_RECURSION:
171    p = "never ending recursion"; break;
172  case ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY:
173    p = "group number is too big for capture history"; break;
174  case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
175    p = "invalid character property name {%n}"; break;
176  case ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION:
177    p = "not supported encoding combination"; break;
178  case ONIGERR_INVALID_COMBINATION_OF_OPTIONS:
179    p = "invalid combination of options"; break;
180  case ONIGERR_OVER_THREAD_PASS_LIMIT_COUNT:
181    p = "over thread pass limit count"; break;
182
183  default:
184    p = "undefined error code"; break;
185  }
186
187  return (UChar* )p;
188}
189
190static void sprint_byte(char* s, unsigned int v)
191{
192  sprintf(s, "%02x", (v & 0377));
193}
194
195static void sprint_byte_with_x(char* s, unsigned int v)
196{
197  sprintf(s, "\\x%02x", (v & 0377));
198}
199
200static int to_ascii(OnigEncoding enc, UChar *s, UChar *end,
201		    UChar buf[], int buf_size, int *is_over)
202{
203  int len;
204  UChar *p;
205  OnigCodePoint code;
206
207  if (ONIGENC_MBC_MINLEN(enc) > 1) {
208    p = s;
209    len = 0;
210    while (p < end) {
211      code = ONIGENC_MBC_TO_CODE(enc, p, end);
212      if (code >= 0x80) {
213	if (code > 0xffff && len + 10 <= buf_size) {
214	  sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 24));
215	  sprint_byte((char*)(&(buf[len+4])),      (unsigned int)(code >> 16));
216	  sprint_byte((char*)(&(buf[len+6])),      (unsigned int)(code >>  8));
217	  sprint_byte((char*)(&(buf[len+8])),      (unsigned int)code);
218	  len += 10;
219	}
220	else if (len + 6 <= buf_size) {
221	  sprint_byte_with_x((char*)(&(buf[len])), (unsigned int)(code >> 8));
222	  sprint_byte((char*)(&(buf[len+4])),      (unsigned int)code);
223	  len += 6;
224	}
225	else {
226	  break;
227	}
228      }
229      else {
230	buf[len++] = (UChar )code;
231      }
232
233      p += enclen(enc, p, end);
234      if (len >= buf_size) break;
235    }
236
237    *is_over = ((p < end) ? 1 : 0);
238  }
239  else {
240    len = (int )MIN((end - s), buf_size);
241    xmemcpy(buf, s, (size_t )len);
242    *is_over = ((buf_size < (end - s)) ? 1 : 0);
243  }
244
245  return len;
246}
247
248
249/* for ONIG_MAX_ERROR_MESSAGE_LEN */
250#define MAX_ERROR_PAR_LEN   30
251
252extern int
253#ifdef HAVE_STDARG_PROTOTYPES
254onig_error_code_to_str(UChar* s, OnigPosition code, ...)
255#else
256onig_error_code_to_str(s, code, va_alist)
257  UChar* s;
258  OnigPosition code;
259  va_dcl
260#endif
261{
262  UChar *p, *q;
263  OnigErrorInfo* einfo;
264  size_t len;
265  int is_over;
266  UChar parbuf[MAX_ERROR_PAR_LEN];
267  va_list vargs;
268
269  va_init_list(vargs, code);
270
271  switch (code) {
272  case ONIGERR_UNDEFINED_NAME_REFERENCE:
273  case ONIGERR_UNDEFINED_GROUP_REFERENCE:
274  case ONIGERR_MULTIPLEX_DEFINED_NAME:
275  case ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL:
276  case ONIGERR_INVALID_GROUP_NAME:
277  case ONIGERR_INVALID_CHAR_IN_GROUP_NAME:
278  case ONIGERR_INVALID_CHAR_PROPERTY_NAME:
279    einfo = va_arg(vargs, OnigErrorInfo*);
280    len = to_ascii(einfo->enc, einfo->par, einfo->par_end,
281		   parbuf, MAX_ERROR_PAR_LEN - 3, &is_over);
282    q = onig_error_code_to_format(code);
283    p = s;
284    while (*q != '\0') {
285      if (*q == '%') {
286	q++;
287	if (*q == 'n') { /* '%n': name */
288	  xmemcpy(p, parbuf, len);
289	  p += len;
290	  if (is_over != 0) {
291	    xmemcpy(p, "...", 3);
292	    p += 3;
293	  }
294	  q++;
295	}
296	else
297	  goto normal_char;
298      }
299      else {
300      normal_char:
301	*p++ = *q++;
302      }
303    }
304    *p = '\0';
305    len = p - s;
306    break;
307
308  default:
309    q = onig_error_code_to_format(code);
310    len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, q);
311    xmemcpy(s, q, len);
312    s[len] = '\0';
313    break;
314  }
315
316  va_end(vargs);
317  return (int )len;
318}
319
320void
321onig_vsnprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
322                           UChar* pat, UChar* pat_end, const UChar *fmt, va_list args)
323{
324  size_t need;
325  int n, len;
326  UChar *p, *s, *bp;
327  UChar bs[6];
328
329  n = xvsnprintf((char* )buf, bufsize, (const char* )fmt, args);
330
331  need = (pat_end - pat) * 4 + 4;
332
333  if (n + need < (size_t )bufsize) {
334    strcat((char* )buf, ": /");
335    s = buf + onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, buf);
336
337    p = pat;
338    while (p < pat_end) {
339      if (*p == '\\') {
340	*s++ = *p++;
341	len = enclen(enc, p, pat_end);
342	while (len-- > 0) *s++ = *p++;
343      }
344      else if (*p == '/') {
345	*s++ = (unsigned char )'\\';
346	*s++ = *p++;
347      }
348      else if (ONIGENC_IS_MBC_HEAD(enc, p, pat_end)) {
349        len = enclen(enc, p, pat_end);
350        if (ONIGENC_MBC_MINLEN(enc) == 1) {
351          while (len-- > 0) *s++ = *p++;
352        }
353        else { /* for UTF16 */
354          int blen;
355
356          while (len-- > 0) {
357            sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
358            blen = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
359            bp = bs;
360            while (blen-- > 0) *s++ = *bp++;
361          }
362        }
363      }
364      else if (!ONIGENC_IS_CODE_PRINT(enc, *p) &&
365	       !ONIGENC_IS_CODE_SPACE(enc, *p)) {
366	sprint_byte_with_x((char* )bs, (unsigned int )(*p++));
367	len = onigenc_str_bytelen_null(ONIG_ENCODING_ASCII, bs);
368        bp = bs;
369	while (len-- > 0) *s++ = *bp++;
370      }
371      else {
372	*s++ = *p++;
373      }
374    }
375
376    *s++ = '/';
377    *s   = '\0';
378  }
379}
380
381void
382#ifdef HAVE_STDARG_PROTOTYPES
383onig_snprintf_with_pattern(UChar buf[], int bufsize, OnigEncoding enc,
384                           UChar* pat, UChar* pat_end, const UChar *fmt, ...)
385#else
386onig_snprintf_with_pattern(buf, bufsize, enc, pat, pat_end, fmt, va_alist)
387    UChar buf[];
388    int bufsize;
389    OnigEncoding enc;
390    UChar* pat;
391    UChar* pat_end;
392    const UChar *fmt;
393    va_dcl
394#endif
395{
396  va_list args;
397  va_init_list(args, fmt);
398  onig_vsnprintf_with_pattern(buf, bufsize, enc,
399	  pat, pat_end, fmt, args);
400  va_end(args);
401}
402
403