| 6 |
and semantics are as close as possible to those of the Perl 5 language. |
and semantics are as close as possible to those of the Perl 5 language. |
| 7 |
|
|
| 8 |
Written by Philip Hazel |
Written by Philip Hazel |
| 9 |
Copyright (c) 1997-2010 University of Cambridge |
Copyright (c) 1997-2011 University of Cambridge |
| 10 |
|
|
| 11 |
----------------------------------------------------------------------------- |
----------------------------------------------------------------------------- |
| 12 |
Redistribution and use in source and binary forms, with or without |
Redistribution and use in source and binary forms, with or without |
| 261 |
cbit_xdigit,-1, 0 /* xdigit */ |
cbit_xdigit,-1, 0 /* xdigit */ |
| 262 |
}; |
}; |
| 263 |
|
|
| 264 |
/* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class |
/* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class |
| 265 |
substitutes must be in the order of the names, defined above, and there are |
substitutes must be in the order of the names, defined above, and there are |
| 266 |
both positive and negative cases. NULL means no substitute. */ |
both positive and negative cases. NULL means no substitute. */ |
| 267 |
|
|
| 268 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 272 |
(uschar *)"\\P{Xsp}", /* \S */ /* NOTE: Xsp is Perl space */ |
(uschar *)"\\P{Xsp}", /* \S */ /* NOTE: Xsp is Perl space */ |
| 273 |
(uschar *)"\\p{Xsp}", /* \s */ |
(uschar *)"\\p{Xsp}", /* \s */ |
| 274 |
(uschar *)"\\P{Xwd}", /* \W */ |
(uschar *)"\\P{Xwd}", /* \W */ |
| 275 |
(uschar *)"\\p{Xwd}" /* \w */ |
(uschar *)"\\p{Xwd}" /* \w */ |
| 276 |
}; |
}; |
| 277 |
|
|
| 278 |
static const uschar *posix_substitutes[] = { |
static const uschar *posix_substitutes[] = { |
| 279 |
(uschar *)"\\p{L}", /* alpha */ |
(uschar *)"\\p{L}", /* alpha */ |
| 280 |
(uschar *)"\\p{Ll}", /* lower */ |
(uschar *)"\\p{Ll}", /* lower */ |
| 281 |
(uschar *)"\\p{Lu}", /* upper */ |
(uschar *)"\\p{Lu}", /* upper */ |
| 282 |
(uschar *)"\\p{Xan}", /* alnum */ |
(uschar *)"\\p{Xan}", /* alnum */ |
| 283 |
NULL, /* ascii */ |
NULL, /* ascii */ |
| 284 |
(uschar *)"\\h", /* blank */ |
(uschar *)"\\h", /* blank */ |
| 285 |
NULL, /* cntrl */ |
NULL, /* cntrl */ |
| 289 |
NULL, /* punct */ |
NULL, /* punct */ |
| 290 |
(uschar *)"\\p{Xps}", /* space */ /* NOTE: Xps is POSIX space */ |
(uschar *)"\\p{Xps}", /* space */ /* NOTE: Xps is POSIX space */ |
| 291 |
(uschar *)"\\p{Xwd}", /* word */ |
(uschar *)"\\p{Xwd}", /* word */ |
| 292 |
NULL, /* xdigit */ |
NULL, /* xdigit */ |
| 293 |
/* Negated cases */ |
/* Negated cases */ |
| 294 |
(uschar *)"\\P{L}", /* ^alpha */ |
(uschar *)"\\P{L}", /* ^alpha */ |
| 295 |
(uschar *)"\\P{Ll}", /* ^lower */ |
(uschar *)"\\P{Ll}", /* ^lower */ |
| 296 |
(uschar *)"\\P{Lu}", /* ^upper */ |
(uschar *)"\\P{Lu}", /* ^upper */ |
| 297 |
(uschar *)"\\P{Xan}", /* ^alnum */ |
(uschar *)"\\P{Xan}", /* ^alnum */ |
| 298 |
NULL, /* ^ascii */ |
NULL, /* ^ascii */ |
| 299 |
(uschar *)"\\H", /* ^blank */ |
(uschar *)"\\H", /* ^blank */ |
| 300 |
NULL, /* ^cntrl */ |
NULL, /* ^cntrl */ |
| 304 |
NULL, /* ^punct */ |
NULL, /* ^punct */ |
| 305 |
(uschar *)"\\P{Xps}", /* ^space */ /* NOTE: Xps is POSIX space */ |
(uschar *)"\\P{Xps}", /* ^space */ /* NOTE: Xps is POSIX space */ |
| 306 |
(uschar *)"\\P{Xwd}", /* ^word */ |
(uschar *)"\\P{Xwd}", /* ^word */ |
| 307 |
NULL /* ^xdigit */ |
NULL /* ^xdigit */ |
| 308 |
}; |
}; |
| 309 |
#define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *)) |
#define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *)) |
| 310 |
#endif |
#endif |
| 311 |
|
|
| 312 |
#define STRING(a) # a |
#define STRING(a) # a |
| 313 |
#define XSTRING(s) STRING(s) |
#define XSTRING(s) STRING(s) |
| 407 |
/* 65 */ |
/* 65 */ |
| 408 |
"different names for subpatterns of the same number are not allowed\0" |
"different names for subpatterns of the same number are not allowed\0" |
| 409 |
"(*MARK) must have an argument\0" |
"(*MARK) must have an argument\0" |
| 410 |
"this version of PCRE is not compiled with PCRE_UCP support\0" |
"this version of PCRE is not compiled with PCRE_UCP support\0" |
| 411 |
|
"\\c must be followed by an ASCII character\0" |
| 412 |
; |
; |
| 413 |
|
|
| 414 |
/* Table to identify digits and hex digits. This is used when compiling |
/* Table to identify digits and hex digits. This is used when compiling |
| 842 |
break; |
break; |
| 843 |
|
|
| 844 |
/* For \c, a following letter is upper-cased; then the 0x40 bit is flipped. |
/* For \c, a following letter is upper-cased; then the 0x40 bit is flipped. |
| 845 |
This coding is ASCII-specific, but then the whole concept of \cx is |
An error is given if the byte following \c is not an ASCII character. This |
| 846 |
|
coding is ASCII-specific, but then the whole concept of \cx is |
| 847 |
ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ |
ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ |
| 848 |
|
|
| 849 |
case CHAR_c: |
case CHAR_c: |
| 853 |
*errorcodeptr = ERR2; |
*errorcodeptr = ERR2; |
| 854 |
break; |
break; |
| 855 |
} |
} |
| 856 |
|
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
| 857 |
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
if (c > 127) /* Excludes all non-ASCII in either mode */ |
| 858 |
|
{ |
| 859 |
|
*errorcodeptr = ERR68; |
| 860 |
|
break; |
| 861 |
|
} |
| 862 |
if (c >= CHAR_a && c <= CHAR_z) c -= 32; |
if (c >= CHAR_a && c <= CHAR_z) c -= 32; |
| 863 |
c ^= 0x40; |
c ^= 0x40; |
| 864 |
#else /* EBCDIC coding */ |
#else /* EBCDIC coding */ |
| 865 |
if (c >= CHAR_a && c <= CHAR_z) c += 64; |
if (c >= CHAR_a && c <= CHAR_z) c += 64; |
| 866 |
c ^= 0xC0; |
c ^= 0xC0; |
| 867 |
#endif |
#endif |
| 1105 |
start at a parenthesis. It scans along a pattern's text looking for capturing |
start at a parenthesis. It scans along a pattern's text looking for capturing |
| 1106 |
subpatterns, and counting them. If it finds a named pattern that matches the |
subpatterns, and counting them. If it finds a named pattern that matches the |
| 1107 |
name it is given, it returns its number. Alternatively, if the name is NULL, it |
name it is given, it returns its number. Alternatively, if the name is NULL, it |
| 1108 |
returns when it reaches a given numbered subpattern. We know that if (?P< is |
returns when it reaches a given numbered subpattern. Recursion is used to keep |
| 1109 |
encountered, the name will be terminated by '>' because that is checked in the |
track of subpatterns that reset the capturing group numbers - the (?| feature. |
| 1110 |
first pass. Recursion is used to keep track of subpatterns that reset the |
|
| 1111 |
capturing group numbers - the (?| feature. |
This function was originally called only from the second pass, in which we know |
| 1112 |
|
that if (?< or (?' or (?P< is encountered, the name will be correctly |
| 1113 |
|
terminated because that is checked in the first pass. There is now one call to |
| 1114 |
|
this function in the first pass, to check for a recursive back reference by |
| 1115 |
|
name (so that we can make the whole group atomic). In this case, we need check |
| 1116 |
|
only up to the current position in the pattern, and that is still OK because |
| 1117 |
|
and previous occurrences will have been checked. To make this work, the test |
| 1118 |
|
for "end of pattern" is a check against cd->end_pattern in the main loop, |
| 1119 |
|
instead of looking for a binary zero. This means that the special first-pass |
| 1120 |
|
call can adjust cd->end_pattern temporarily. (Checks for binary zero while |
| 1121 |
|
processing items within the loop are OK, because afterwards the main loop will |
| 1122 |
|
terminate.) |
| 1123 |
|
|
| 1124 |
Arguments: |
Arguments: |
| 1125 |
ptrptr address of the current character pointer (updated) |
ptrptr address of the current character pointer (updated) |
| 1127 |
name name to seek, or NULL if seeking a numbered subpattern |
name name to seek, or NULL if seeking a numbered subpattern |
| 1128 |
lorn name length, or subpattern number if name is NULL |
lorn name length, or subpattern number if name is NULL |
| 1129 |
xmode TRUE if we are in /x mode |
xmode TRUE if we are in /x mode |
| 1130 |
|
utf8 TRUE if we are in UTF-8 mode |
| 1131 |
count pointer to the current capturing subpattern number (updated) |
count pointer to the current capturing subpattern number (updated) |
| 1132 |
|
|
| 1133 |
Returns: the number of the named subpattern, or -1 if not found |
Returns: the number of the named subpattern, or -1 if not found |
| 1135 |
|
|
| 1136 |
static int |
static int |
| 1137 |
find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn, |
find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn, |
| 1138 |
BOOL xmode, int *count) |
BOOL xmode, BOOL utf8, int *count) |
| 1139 |
{ |
{ |
| 1140 |
uschar *ptr = *ptrptr; |
uschar *ptr = *ptrptr; |
| 1141 |
int start_count = *count; |
int start_count = *count; |
| 1147 |
|
|
| 1148 |
if (ptr[0] == CHAR_LEFT_PARENTHESIS) |
if (ptr[0] == CHAR_LEFT_PARENTHESIS) |
| 1149 |
{ |
{ |
| 1150 |
if (ptr[1] == CHAR_QUESTION_MARK && |
/* Handle specials such as (*SKIP) or (*UTF8) etc. */ |
| 1151 |
ptr[2] == CHAR_VERTICAL_LINE) |
|
| 1152 |
|
if (ptr[1] == CHAR_ASTERISK) ptr += 2; |
| 1153 |
|
|
| 1154 |
|
/* Handle a normal, unnamed capturing parenthesis. */ |
| 1155 |
|
|
| 1156 |
|
else if (ptr[1] != CHAR_QUESTION_MARK) |
| 1157 |
|
{ |
| 1158 |
|
*count += 1; |
| 1159 |
|
if (name == NULL && *count == lorn) return *count; |
| 1160 |
|
ptr++; |
| 1161 |
|
} |
| 1162 |
|
|
| 1163 |
|
/* All cases now have (? at the start. Remember when we are in a group |
| 1164 |
|
where the parenthesis numbers are duplicated. */ |
| 1165 |
|
|
| 1166 |
|
else if (ptr[2] == CHAR_VERTICAL_LINE) |
| 1167 |
{ |
{ |
| 1168 |
ptr += 3; |
ptr += 3; |
| 1169 |
dup_parens = TRUE; |
dup_parens = TRUE; |
| 1170 |
} |
} |
| 1171 |
|
|
| 1172 |
/* Handle a normal, unnamed capturing parenthesis */ |
/* Handle comments; all characters are allowed until a ket is reached. */ |
| 1173 |
|
|
| 1174 |
else if (ptr[1] != CHAR_QUESTION_MARK && ptr[1] != CHAR_ASTERISK) |
else if (ptr[2] == CHAR_NUMBER_SIGN) |
| 1175 |
{ |
{ |
| 1176 |
*count += 1; |
for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break; |
| 1177 |
if (name == NULL && *count == lorn) return *count; |
goto FAIL_EXIT; |
|
ptr++; |
|
| 1178 |
} |
} |
| 1179 |
|
|
| 1180 |
/* Handle a condition. If it is an assertion, just carry on so that it |
/* Handle a condition. If it is an assertion, just carry on so that it |
| 1181 |
is processed as normal. If not, skip to the closing parenthesis of the |
is processed as normal. If not, skip to the closing parenthesis of the |
| 1182 |
condition (there can't be any nested parens. */ |
condition (there can't be any nested parens). */ |
| 1183 |
|
|
| 1184 |
else if (ptr[2] == CHAR_LEFT_PARENTHESIS) |
else if (ptr[2] == CHAR_LEFT_PARENTHESIS) |
| 1185 |
{ |
{ |
| 1191 |
} |
} |
| 1192 |
} |
} |
| 1193 |
|
|
| 1194 |
/* We have either (? or (* and not a condition */ |
/* Start with (? but not a condition. */ |
| 1195 |
|
|
| 1196 |
else |
else |
| 1197 |
{ |
{ |
| 1220 |
} |
} |
| 1221 |
|
|
| 1222 |
/* Past any initial parenthesis handling, scan for parentheses or vertical |
/* Past any initial parenthesis handling, scan for parentheses or vertical |
| 1223 |
bars. */ |
bars. Stop if we get to cd->end_pattern. Note that this is important for the |
| 1224 |
|
first-pass call when this value is temporarily adjusted to stop at the current |
| 1225 |
|
position. So DO NOT change this to a test for binary zero. */ |
| 1226 |
|
|
| 1227 |
for (; *ptr != 0; ptr++) |
for (; ptr < cd->end_pattern; ptr++) |
| 1228 |
{ |
{ |
| 1229 |
/* Skip over backslashed characters and also entire \Q...\E */ |
/* Skip over backslashed characters and also entire \Q...\E */ |
| 1230 |
|
|
| 1298 |
|
|
| 1299 |
if (xmode && *ptr == CHAR_NUMBER_SIGN) |
if (xmode && *ptr == CHAR_NUMBER_SIGN) |
| 1300 |
{ |
{ |
| 1301 |
while (*(++ptr) != 0 && *ptr != CHAR_NL) {}; |
ptr++; |
| 1302 |
|
while (*ptr != 0) |
| 1303 |
|
{ |
| 1304 |
|
if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } |
| 1305 |
|
ptr++; |
| 1306 |
|
#ifdef SUPPORT_UTF8 |
| 1307 |
|
if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++; |
| 1308 |
|
#endif |
| 1309 |
|
} |
| 1310 |
if (*ptr == 0) goto FAIL_EXIT; |
if (*ptr == 0) goto FAIL_EXIT; |
| 1311 |
continue; |
continue; |
| 1312 |
} |
} |
| 1315 |
|
|
| 1316 |
if (*ptr == CHAR_LEFT_PARENTHESIS) |
if (*ptr == CHAR_LEFT_PARENTHESIS) |
| 1317 |
{ |
{ |
| 1318 |
int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, count); |
int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count); |
| 1319 |
if (rc > 0) return rc; |
if (rc > 0) return rc; |
| 1320 |
if (*ptr == 0) goto FAIL_EXIT; |
if (*ptr == 0) goto FAIL_EXIT; |
| 1321 |
} |
} |
| 1323 |
else if (*ptr == CHAR_RIGHT_PARENTHESIS) |
else if (*ptr == CHAR_RIGHT_PARENTHESIS) |
| 1324 |
{ |
{ |
| 1325 |
if (dup_parens && *count < hwm_count) *count = hwm_count; |
if (dup_parens && *count < hwm_count) *count = hwm_count; |
| 1326 |
*ptrptr = ptr; |
goto FAIL_EXIT; |
|
return -1; |
|
| 1327 |
} |
} |
| 1328 |
|
|
| 1329 |
else if (*ptr == CHAR_VERTICAL_LINE && dup_parens) |
else if (*ptr == CHAR_VERTICAL_LINE && dup_parens) |
| 1361 |
name name to seek, or NULL if seeking a numbered subpattern |
name name to seek, or NULL if seeking a numbered subpattern |
| 1362 |
lorn name length, or subpattern number if name is NULL |
lorn name length, or subpattern number if name is NULL |
| 1363 |
xmode TRUE if we are in /x mode |
xmode TRUE if we are in /x mode |
| 1364 |
|
utf8 TRUE if we are in UTF-8 mode |
| 1365 |
|
|
| 1366 |
Returns: the number of the found subpattern, or -1 if not found |
Returns: the number of the found subpattern, or -1 if not found |
| 1367 |
*/ |
*/ |
| 1368 |
|
|
| 1369 |
static int |
static int |
| 1370 |
find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode) |
find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode, |
| 1371 |
|
BOOL utf8) |
| 1372 |
{ |
{ |
| 1373 |
uschar *ptr = (uschar *)cd->start_pattern; |
uschar *ptr = (uschar *)cd->start_pattern; |
| 1374 |
int count = 0; |
int count = 0; |
| 1381 |
|
|
| 1382 |
for (;;) |
for (;;) |
| 1383 |
{ |
{ |
| 1384 |
rc = find_parens_sub(&ptr, cd, name, lorn, xmode, &count); |
rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count); |
| 1385 |
if (rc > 0 || *ptr++ == 0) break; |
if (rc > 0 || *ptr++ == 0) break; |
| 1386 |
} |
} |
| 1387 |
|
|
| 1397 |
|
|
| 1398 |
/* This is called by several functions that scan a compiled expression looking |
/* This is called by several functions that scan a compiled expression looking |
| 1399 |
for a fixed first character, or an anchoring op code etc. It skips over things |
for a fixed first character, or an anchoring op code etc. It skips over things |
| 1400 |
that do not influence this. For some calls, a change of option is important. |
that do not influence this. For some calls, it makes sense to skip negative |
| 1401 |
For some calls, it makes sense to skip negative forward and all backward |
forward and all backward assertions, and also the \b assertion; for others it |
| 1402 |
assertions, and also the \b assertion; for others it does not. |
does not. |
| 1403 |
|
|
| 1404 |
Arguments: |
Arguments: |
| 1405 |
code pointer to the start of the group |
code pointer to the start of the group |
| 1419 |
{ |
{ |
| 1420 |
switch ((int)*code) |
switch ((int)*code) |
| 1421 |
{ |
{ |
|
case OP_OPT: |
|
|
if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit)) |
|
|
*options = (int)code[1]; |
|
|
code += 2; |
|
|
break; |
|
|
|
|
| 1422 |
case OP_ASSERT_NOT: |
case OP_ASSERT_NOT: |
| 1423 |
case OP_ASSERTBACK: |
case OP_ASSERTBACK: |
| 1424 |
case OP_ASSERTBACK_NOT: |
case OP_ASSERTBACK_NOT: |
| 1555 |
case OP_RREF: |
case OP_RREF: |
| 1556 |
case OP_NRREF: |
case OP_NRREF: |
| 1557 |
case OP_DEF: |
case OP_DEF: |
|
case OP_OPT: |
|
| 1558 |
case OP_CALLOUT: |
case OP_CALLOUT: |
| 1559 |
case OP_SOD: |
case OP_SOD: |
| 1560 |
case OP_SOM: |
case OP_SOM: |
| 1562 |
case OP_EOD: |
case OP_EOD: |
| 1563 |
case OP_EODN: |
case OP_EODN: |
| 1564 |
case OP_CIRC: |
case OP_CIRC: |
| 1565 |
|
case OP_CIRCM: |
| 1566 |
case OP_DOLL: |
case OP_DOLL: |
| 1567 |
|
case OP_DOLLM: |
| 1568 |
case OP_NOT_WORD_BOUNDARY: |
case OP_NOT_WORD_BOUNDARY: |
| 1569 |
case OP_WORD_BOUNDARY: |
case OP_WORD_BOUNDARY: |
| 1570 |
cc += _pcre_OP_lengths[*cc]; |
cc += _pcre_OP_lengths[*cc]; |
| 1573 |
/* Handle literal characters */ |
/* Handle literal characters */ |
| 1574 |
|
|
| 1575 |
case OP_CHAR: |
case OP_CHAR: |
| 1576 |
case OP_CHARNC: |
case OP_CHARI: |
| 1577 |
case OP_NOT: |
case OP_NOT: |
| 1578 |
|
case OP_NOTI: |
| 1579 |
branchlength++; |
branchlength++; |
| 1580 |
cc += 2; |
cc += 2; |
| 1581 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 1750 |
case OP_MARK: |
case OP_MARK: |
| 1751 |
case OP_PRUNE_ARG: |
case OP_PRUNE_ARG: |
| 1752 |
case OP_SKIP_ARG: |
case OP_SKIP_ARG: |
|
case OP_THEN_ARG: |
|
| 1753 |
code += code[1]; |
code += code[1]; |
| 1754 |
break; |
break; |
| 1755 |
|
|
| 1756 |
|
case OP_THEN_ARG: |
| 1757 |
|
code += code[1+LINK_SIZE]; |
| 1758 |
|
break; |
| 1759 |
} |
} |
| 1760 |
|
|
| 1761 |
/* Add in the fixed length from the table */ |
/* Add in the fixed length from the table */ |
| 1770 |
if (utf8) switch(c) |
if (utf8) switch(c) |
| 1771 |
{ |
{ |
| 1772 |
case OP_CHAR: |
case OP_CHAR: |
| 1773 |
case OP_CHARNC: |
case OP_CHARI: |
| 1774 |
case OP_EXACT: |
case OP_EXACT: |
| 1775 |
|
case OP_EXACTI: |
| 1776 |
case OP_UPTO: |
case OP_UPTO: |
| 1777 |
|
case OP_UPTOI: |
| 1778 |
case OP_MINUPTO: |
case OP_MINUPTO: |
| 1779 |
|
case OP_MINUPTOI: |
| 1780 |
case OP_POSUPTO: |
case OP_POSUPTO: |
| 1781 |
|
case OP_POSUPTOI: |
| 1782 |
case OP_STAR: |
case OP_STAR: |
| 1783 |
|
case OP_STARI: |
| 1784 |
case OP_MINSTAR: |
case OP_MINSTAR: |
| 1785 |
|
case OP_MINSTARI: |
| 1786 |
case OP_POSSTAR: |
case OP_POSSTAR: |
| 1787 |
|
case OP_POSSTARI: |
| 1788 |
case OP_PLUS: |
case OP_PLUS: |
| 1789 |
|
case OP_PLUSI: |
| 1790 |
case OP_MINPLUS: |
case OP_MINPLUS: |
| 1791 |
|
case OP_MINPLUSI: |
| 1792 |
case OP_POSPLUS: |
case OP_POSPLUS: |
| 1793 |
|
case OP_POSPLUSI: |
| 1794 |
case OP_QUERY: |
case OP_QUERY: |
| 1795 |
|
case OP_QUERYI: |
| 1796 |
case OP_MINQUERY: |
case OP_MINQUERY: |
| 1797 |
|
case OP_MINQUERYI: |
| 1798 |
case OP_POSQUERY: |
case OP_POSQUERY: |
| 1799 |
|
case OP_POSQUERYI: |
| 1800 |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
| 1801 |
break; |
break; |
| 1802 |
} |
} |
| 1869 |
case OP_MARK: |
case OP_MARK: |
| 1870 |
case OP_PRUNE_ARG: |
case OP_PRUNE_ARG: |
| 1871 |
case OP_SKIP_ARG: |
case OP_SKIP_ARG: |
|
case OP_THEN_ARG: |
|
| 1872 |
code += code[1]; |
code += code[1]; |
| 1873 |
break; |
break; |
| 1874 |
|
|
| 1875 |
|
case OP_THEN_ARG: |
| 1876 |
|
code += code[1+LINK_SIZE]; |
| 1877 |
|
break; |
| 1878 |
} |
} |
| 1879 |
|
|
| 1880 |
/* Add in the fixed length from the table */ |
/* Add in the fixed length from the table */ |
| 1889 |
if (utf8) switch(c) |
if (utf8) switch(c) |
| 1890 |
{ |
{ |
| 1891 |
case OP_CHAR: |
case OP_CHAR: |
| 1892 |
case OP_CHARNC: |
case OP_CHARI: |
| 1893 |
case OP_EXACT: |
case OP_EXACT: |
| 1894 |
|
case OP_EXACTI: |
| 1895 |
case OP_UPTO: |
case OP_UPTO: |
| 1896 |
|
case OP_UPTOI: |
| 1897 |
case OP_MINUPTO: |
case OP_MINUPTO: |
| 1898 |
|
case OP_MINUPTOI: |
| 1899 |
case OP_POSUPTO: |
case OP_POSUPTO: |
| 1900 |
|
case OP_POSUPTOI: |
| 1901 |
case OP_STAR: |
case OP_STAR: |
| 1902 |
|
case OP_STARI: |
| 1903 |
case OP_MINSTAR: |
case OP_MINSTAR: |
| 1904 |
|
case OP_MINSTARI: |
| 1905 |
case OP_POSSTAR: |
case OP_POSSTAR: |
| 1906 |
|
case OP_POSSTARI: |
| 1907 |
case OP_PLUS: |
case OP_PLUS: |
| 1908 |
|
case OP_PLUSI: |
| 1909 |
case OP_MINPLUS: |
case OP_MINPLUS: |
| 1910 |
|
case OP_MINPLUSI: |
| 1911 |
case OP_POSPLUS: |
case OP_POSPLUS: |
| 1912 |
|
case OP_POSPLUSI: |
| 1913 |
case OP_QUERY: |
case OP_QUERY: |
| 1914 |
|
case OP_QUERYI: |
| 1915 |
case OP_MINQUERY: |
case OP_MINQUERY: |
| 1916 |
|
case OP_MINQUERYI: |
| 1917 |
case OP_POSQUERY: |
case OP_POSQUERY: |
| 1918 |
|
case OP_POSQUERYI: |
| 1919 |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
| 1920 |
break; |
break; |
| 1921 |
} |
} |
| 2093 |
case OP_ALLANY: |
case OP_ALLANY: |
| 2094 |
case OP_ANYBYTE: |
case OP_ANYBYTE: |
| 2095 |
case OP_CHAR: |
case OP_CHAR: |
| 2096 |
case OP_CHARNC: |
case OP_CHARI: |
| 2097 |
case OP_NOT: |
case OP_NOT: |
| 2098 |
|
case OP_NOTI: |
| 2099 |
case OP_PLUS: |
case OP_PLUS: |
| 2100 |
case OP_MINPLUS: |
case OP_MINPLUS: |
| 2101 |
case OP_POSPLUS: |
case OP_POSPLUS: |
| 2143 |
|
|
| 2144 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2145 |
case OP_STAR: |
case OP_STAR: |
| 2146 |
|
case OP_STARI: |
| 2147 |
case OP_MINSTAR: |
case OP_MINSTAR: |
| 2148 |
|
case OP_MINSTARI: |
| 2149 |
case OP_POSSTAR: |
case OP_POSSTAR: |
| 2150 |
|
case OP_POSSTARI: |
| 2151 |
case OP_QUERY: |
case OP_QUERY: |
| 2152 |
|
case OP_QUERYI: |
| 2153 |
case OP_MINQUERY: |
case OP_MINQUERY: |
| 2154 |
|
case OP_MINQUERYI: |
| 2155 |
case OP_POSQUERY: |
case OP_POSQUERY: |
| 2156 |
|
case OP_POSQUERYI: |
| 2157 |
if (utf8 && code[1] >= 0xc0) code += _pcre_utf8_table4[code[1] & 0x3f]; |
if (utf8 && code[1] >= 0xc0) code += _pcre_utf8_table4[code[1] & 0x3f]; |
| 2158 |
break; |
break; |
| 2159 |
|
|
| 2160 |
case OP_UPTO: |
case OP_UPTO: |
| 2161 |
|
case OP_UPTOI: |
| 2162 |
case OP_MINUPTO: |
case OP_MINUPTO: |
| 2163 |
|
case OP_MINUPTOI: |
| 2164 |
case OP_POSUPTO: |
case OP_POSUPTO: |
| 2165 |
|
case OP_POSUPTOI: |
| 2166 |
if (utf8 && code[3] >= 0xc0) code += _pcre_utf8_table4[code[3] & 0x3f]; |
if (utf8 && code[3] >= 0xc0) code += _pcre_utf8_table4[code[3] & 0x3f]; |
| 2167 |
break; |
break; |
| 2168 |
#endif |
#endif |
| 2173 |
case OP_MARK: |
case OP_MARK: |
| 2174 |
case OP_PRUNE_ARG: |
case OP_PRUNE_ARG: |
| 2175 |
case OP_SKIP_ARG: |
case OP_SKIP_ARG: |
|
case OP_THEN_ARG: |
|
| 2176 |
code += code[1]; |
code += code[1]; |
| 2177 |
break; |
break; |
| 2178 |
|
|
| 2179 |
|
case OP_THEN_ARG: |
| 2180 |
|
code += code[1+LINK_SIZE]; |
| 2181 |
|
break; |
| 2182 |
|
|
| 2183 |
/* None of the remaining opcodes are required to match a character. */ |
/* None of the remaining opcodes are required to match a character. */ |
| 2184 |
|
|
| 2185 |
default: |
default: |
| 2491 |
ptype the property type |
ptype the property type |
| 2492 |
pdata the data for the type |
pdata the data for the type |
| 2493 |
negated TRUE if it's a negated property (\P or \p{^) |
negated TRUE if it's a negated property (\P or \p{^) |
| 2494 |
|
|
| 2495 |
Returns: TRUE if auto-possessifying is OK |
Returns: TRUE if auto-possessifying is OK |
| 2496 |
*/ |
*/ |
| 2497 |
|
|
| 2498 |
static BOOL |
static BOOL |
| 2499 |
check_char_prop(int c, int ptype, int pdata, BOOL negated) |
check_char_prop(int c, int ptype, int pdata, BOOL negated) |
| 2537 |
_pcre_ucp_gentype[prop->chartype] == ucp_N || |
_pcre_ucp_gentype[prop->chartype] == ucp_N || |
| 2538 |
c == CHAR_UNDERSCORE) == negated; |
c == CHAR_UNDERSCORE) == negated; |
| 2539 |
} |
} |
| 2540 |
return FALSE; |
return FALSE; |
| 2541 |
} |
} |
| 2542 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
| 2543 |
|
|
| 2562 |
*/ |
*/ |
| 2563 |
|
|
| 2564 |
static BOOL |
static BOOL |
| 2565 |
check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr, |
check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr, |
| 2566 |
int options, compile_data *cd) |
int options, compile_data *cd) |
| 2567 |
{ |
{ |
| 2568 |
int c, next; |
int c, next; |
| 2577 |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
| 2578 |
if (*ptr == CHAR_NUMBER_SIGN) |
if (*ptr == CHAR_NUMBER_SIGN) |
| 2579 |
{ |
{ |
| 2580 |
while (*(++ptr) != 0) |
ptr++; |
| 2581 |
|
while (*ptr != 0) |
| 2582 |
|
{ |
| 2583 |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
| 2584 |
|
ptr++; |
| 2585 |
|
#ifdef SUPPORT_UTF8 |
| 2586 |
|
if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++; |
| 2587 |
|
#endif |
| 2588 |
|
} |
| 2589 |
} |
} |
| 2590 |
else break; |
else break; |
| 2591 |
} |
} |
| 2621 |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
| 2622 |
if (*ptr == CHAR_NUMBER_SIGN) |
if (*ptr == CHAR_NUMBER_SIGN) |
| 2623 |
{ |
{ |
| 2624 |
while (*(++ptr) != 0) |
ptr++; |
| 2625 |
|
while (*ptr != 0) |
| 2626 |
|
{ |
| 2627 |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
| 2628 |
|
ptr++; |
| 2629 |
|
#ifdef SUPPORT_UTF8 |
| 2630 |
|
if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++; |
| 2631 |
|
#endif |
| 2632 |
|
} |
| 2633 |
} |
} |
| 2634 |
else break; |
else break; |
| 2635 |
} |
} |
| 2647 |
if (next >= 0) switch(op_code) |
if (next >= 0) switch(op_code) |
| 2648 |
{ |
{ |
| 2649 |
case OP_CHAR: |
case OP_CHAR: |
| 2650 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2651 |
GETCHARTEST(c, previous); |
GETCHARTEST(c, previous); |
| 2652 |
#else |
#else |
| 2653 |
c = *previous; |
c = *previous; |
| 2654 |
#endif |
#endif |
| 2655 |
return c != next; |
return c != next; |
| 2656 |
|
|
| 2657 |
/* For CHARNC (caseless character) we must check the other case. If we have |
/* For CHARI (caseless character) we must check the other case. If we have |
| 2658 |
Unicode property support, we can use it to test the other case of |
Unicode property support, we can use it to test the other case of |
| 2659 |
high-valued characters. */ |
high-valued characters. */ |
| 2660 |
|
|
| 2661 |
case OP_CHARNC: |
case OP_CHARI: |
| 2662 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2663 |
GETCHARTEST(c, previous); |
GETCHARTEST(c, previous); |
| 2664 |
#else |
#else |
| 2665 |
c = *previous; |
c = *previous; |
| 2666 |
#endif |
#endif |
| 2667 |
if (c == next) return FALSE; |
if (c == next) return FALSE; |
| 2668 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2669 |
if (utf8) |
if (utf8) |
| 2681 |
#endif /* SUPPORT_UTF8 */ |
#endif /* SUPPORT_UTF8 */ |
| 2682 |
return (c != cd->fcc[next]); /* Non-UTF-8 mode */ |
return (c != cd->fcc[next]); /* Non-UTF-8 mode */ |
| 2683 |
|
|
| 2684 |
/* For OP_NOT, its data is always a single-byte character. */ |
/* For OP_NOT and OP_NOTI, the data is always a single-byte character. These |
| 2685 |
|
opcodes are not used for multi-byte characters, because they are coded using |
| 2686 |
|
an XCLASS instead. */ |
| 2687 |
|
|
| 2688 |
case OP_NOT: |
case OP_NOT: |
| 2689 |
|
return (c = *previous) == next; |
| 2690 |
|
|
| 2691 |
|
case OP_NOTI: |
| 2692 |
if ((c = *previous) == next) return TRUE; |
if ((c = *previous) == next) return TRUE; |
|
if ((options & PCRE_CASELESS) == 0) return FALSE; |
|
| 2693 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2694 |
if (utf8) |
if (utf8) |
| 2695 |
{ |
{ |
| 2705 |
else |
else |
| 2706 |
#endif /* SUPPORT_UTF8 */ |
#endif /* SUPPORT_UTF8 */ |
| 2707 |
return (c == cd->fcc[next]); /* Non-UTF-8 mode */ |
return (c == cd->fcc[next]); /* Non-UTF-8 mode */ |
| 2708 |
|
|
| 2709 |
/* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set. |
/* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set. |
| 2710 |
When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */ |
When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */ |
| 2711 |
|
|
| 2712 |
case OP_DIGIT: |
case OP_DIGIT: |
| 2713 |
return next > 127 || (cd->ctypes[next] & ctype_digit) == 0; |
return next > 127 || (cd->ctypes[next] & ctype_digit) == 0; |
| 2775 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 2776 |
case OP_PROP: |
case OP_PROP: |
| 2777 |
return check_char_prop(next, previous[0], previous[1], FALSE); |
return check_char_prop(next, previous[0], previous[1], FALSE); |
| 2778 |
|
|
| 2779 |
case OP_NOTPROP: |
case OP_NOTPROP: |
| 2780 |
return check_char_prop(next, previous[0], previous[1], TRUE); |
return check_char_prop(next, previous[0], previous[1], TRUE); |
| 2781 |
#endif |
#endif |
| 2785 |
} |
} |
| 2786 |
|
|
| 2787 |
|
|
| 2788 |
/* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP |
/* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP |
| 2789 |
is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are |
is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are |
| 2790 |
generated only when PCRE_UCP is *not* set, that is, when only ASCII |
generated only when PCRE_UCP is *not* set, that is, when only ASCII |
| 2791 |
characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are |
characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are |
| 2792 |
replaced by OP_PROP codes when PCRE_UCP is set. */ |
replaced by OP_PROP codes when PCRE_UCP is set. */ |
| 2793 |
|
|
| 2794 |
switch(op_code) |
switch(op_code) |
| 2795 |
{ |
{ |
| 2796 |
case OP_CHAR: |
case OP_CHAR: |
| 2797 |
case OP_CHARNC: |
case OP_CHARI: |
| 2798 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2799 |
GETCHARTEST(c, previous); |
GETCHARTEST(c, previous); |
| 2800 |
#else |
#else |
| 2801 |
c = *previous; |
c = *previous; |
| 2802 |
#endif |
#endif |
| 2803 |
switch(-next) |
switch(-next) |
| 2804 |
{ |
{ |
| 2805 |
case ESC_d: |
case ESC_d: |
| 2863 |
default: |
default: |
| 2864 |
return -next == ESC_v; |
return -next == ESC_v; |
| 2865 |
} |
} |
| 2866 |
|
|
| 2867 |
/* When PCRE_UCP is set, these values get generated for \d etc. Find |
/* When PCRE_UCP is set, these values get generated for \d etc. Find |
| 2868 |
their substitutions and process them. The result will always be either |
their substitutions and process them. The result will always be either |
| 2869 |
-ESC_p or -ESC_P. Then fall through to process those values. */ |
-ESC_p or -ESC_P. Then fall through to process those values. */ |
| 2870 |
|
|
| 2871 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 2872 |
case ESC_du: |
case ESC_du: |
| 2873 |
case ESC_DU: |
case ESC_DU: |
| 2882 |
if (temperrorcode != 0) return FALSE; |
if (temperrorcode != 0) return FALSE; |
| 2883 |
ptr++; /* For compatibility */ |
ptr++; /* For compatibility */ |
| 2884 |
} |
} |
| 2885 |
/* Fall through */ |
/* Fall through */ |
| 2886 |
|
|
| 2887 |
case ESC_p: |
case ESC_p: |
| 2888 |
case ESC_P: |
case ESC_P: |
| 2889 |
{ |
{ |
| 2890 |
int ptype, pdata, errorcodeptr; |
int ptype, pdata, errorcodeptr; |
| 2891 |
BOOL negated; |
BOOL negated; |
| 2892 |
|
|
| 2893 |
ptr--; /* Make ptr point at the p or P */ |
ptr--; /* Make ptr point at the p or P */ |
| 2894 |
ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr); |
ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr); |
| 2895 |
if (ptype < 0) return FALSE; |
if (ptype < 0) return FALSE; |
| 2896 |
ptr++; /* Point past the final curly ket */ |
ptr++; /* Point past the final curly ket */ |
| 2897 |
|
|
| 2898 |
/* If the property item is optional, we have to give up. (When generated |
/* If the property item is optional, we have to give up. (When generated |
| 2899 |
from \d etc by PCRE_UCP, this test will have been applied much earlier, |
from \d etc by PCRE_UCP, this test will have been applied much earlier, |
| 2900 |
to the original \d etc. At this point, ptr will point to a zero byte. */ |
to the original \d etc. At this point, ptr will point to a zero byte. */ |
| 2901 |
|
|
| 2902 |
if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK || |
if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK || |
| 2903 |
strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0) |
strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0) |
| 2904 |
return FALSE; |
return FALSE; |
| 2905 |
|
|
| 2906 |
/* Do the property check. */ |
/* Do the property check. */ |
| 2907 |
|
|
| 2908 |
return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated); |
return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated); |
| 2909 |
} |
} |
| 2910 |
#endif |
#endif |
| 2911 |
|
|
| 2912 |
default: |
default: |
| 2913 |
return FALSE; |
return FALSE; |
| 2914 |
} |
} |
| 2915 |
|
|
| 2916 |
/* In principle, support for Unicode properties should be integrated here as |
/* In principle, support for Unicode properties should be integrated here as |
| 2917 |
well. It means re-organizing the above code so as to get hold of the property |
well. It means re-organizing the above code so as to get hold of the property |
| 2918 |
values before switching on the op-code. However, I wonder how many patterns |
values before switching on the op-code. However, I wonder how many patterns |
| 2919 |
combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set, |
combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set, |
| 2920 |
these op-codes are never generated.) */ |
these op-codes are never generated.) */ |
| 2921 |
|
|
| 2922 |
case OP_DIGIT: |
case OP_DIGIT: |
| 2923 |
return next == -ESC_D || next == -ESC_s || next == -ESC_W || |
return next == -ESC_D || next == -ESC_s || next == -ESC_W || |
| 2933 |
return next == -ESC_s || next == -ESC_h || next == -ESC_v; |
return next == -ESC_s || next == -ESC_h || next == -ESC_v; |
| 2934 |
|
|
| 2935 |
case OP_HSPACE: |
case OP_HSPACE: |
| 2936 |
return next == -ESC_S || next == -ESC_H || next == -ESC_d || |
return next == -ESC_S || next == -ESC_H || next == -ESC_d || |
| 2937 |
next == -ESC_w || next == -ESC_v || next == -ESC_R; |
next == -ESC_w || next == -ESC_v || next == -ESC_R; |
| 2938 |
|
|
| 2939 |
case OP_NOT_HSPACE: |
case OP_NOT_HSPACE: |
| 2940 |
return next == -ESC_h; |
return next == -ESC_h; |
| 2941 |
|
|
| 2942 |
/* Can't have \S in here because VT matches \S (Perl anomaly) */ |
/* Can't have \S in here because VT matches \S (Perl anomaly) */ |
| 2943 |
case OP_ANYNL: |
case OP_ANYNL: |
| 2944 |
case OP_VSPACE: |
case OP_VSPACE: |
| 2945 |
return next == -ESC_V || next == -ESC_d || next == -ESC_w; |
return next == -ESC_V || next == -ESC_d || next == -ESC_w; |
| 2946 |
|
|
| 2948 |
return next == -ESC_v || next == -ESC_R; |
return next == -ESC_v || next == -ESC_R; |
| 2949 |
|
|
| 2950 |
case OP_WORDCHAR: |
case OP_WORDCHAR: |
| 2951 |
return next == -ESC_W || next == -ESC_s || next == -ESC_h || |
return next == -ESC_W || next == -ESC_s || next == -ESC_h || |
| 2952 |
next == -ESC_v || next == -ESC_R; |
next == -ESC_v || next == -ESC_R; |
| 2953 |
|
|
| 2954 |
case OP_NOT_WORDCHAR: |
case OP_NOT_WORDCHAR: |
| 3084 |
|
|
| 3085 |
c = *ptr; |
c = *ptr; |
| 3086 |
|
|
| 3087 |
/* If we are at the end of a nested substitution, revert to the outer level |
/* If we are at the end of a nested substitution, revert to the outer level |
| 3088 |
string. Nesting only happens one level deep. */ |
string. Nesting only happens one level deep. */ |
| 3089 |
|
|
| 3090 |
if (c == 0 && nestptr != NULL) |
if (c == 0 && nestptr != NULL) |
| 3206 |
if ((cd->ctypes[c] & ctype_space) != 0) continue; |
if ((cd->ctypes[c] & ctype_space) != 0) continue; |
| 3207 |
if (c == CHAR_NUMBER_SIGN) |
if (c == CHAR_NUMBER_SIGN) |
| 3208 |
{ |
{ |
| 3209 |
while (*(++ptr) != 0) |
ptr++; |
| 3210 |
|
while (*ptr != 0) |
| 3211 |
{ |
{ |
| 3212 |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } |
| 3213 |
|
ptr++; |
| 3214 |
|
#ifdef SUPPORT_UTF8 |
| 3215 |
|
if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++; |
| 3216 |
|
#endif |
| 3217 |
} |
} |
| 3218 |
if (*ptr != 0) continue; |
if (*ptr != 0) continue; |
| 3219 |
|
|
| 3258 |
the setting of any following char as a first character. */ |
the setting of any following char as a first character. */ |
| 3259 |
|
|
| 3260 |
case CHAR_CIRCUMFLEX_ACCENT: |
case CHAR_CIRCUMFLEX_ACCENT: |
| 3261 |
|
previous = NULL; |
| 3262 |
if ((options & PCRE_MULTILINE) != 0) |
if ((options & PCRE_MULTILINE) != 0) |
| 3263 |
{ |
{ |
| 3264 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 3265 |
|
*code++ = OP_CIRCM; |
| 3266 |
} |
} |
| 3267 |
previous = NULL; |
else *code++ = OP_CIRC; |
|
*code++ = OP_CIRC; |
|
| 3268 |
break; |
break; |
| 3269 |
|
|
| 3270 |
case CHAR_DOLLAR_SIGN: |
case CHAR_DOLLAR_SIGN: |
| 3271 |
previous = NULL; |
previous = NULL; |
| 3272 |
*code++ = OP_DOLL; |
*code++ = ((options & PCRE_MULTILINE) != 0)? OP_DOLLM : OP_DOLL; |
| 3273 |
break; |
break; |
| 3274 |
|
|
| 3275 |
/* There can never be a first char if '.' is first, whatever happens about |
/* There can never be a first char if '.' is first, whatever happens about |
| 3397 |
{ /* Braces are required because the */ |
{ /* Braces are required because the */ |
| 3398 |
GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ |
GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ |
| 3399 |
} |
} |
| 3400 |
|
|
| 3401 |
/* In the pre-compile phase, accumulate the length of any UTF-8 extra |
/* In the pre-compile phase, accumulate the length of any UTF-8 extra |
| 3402 |
data and reset the pointer. This is so that very large classes that |
data and reset the pointer. This is so that very large classes that |
| 3403 |
contain a zillion UTF-8 characters no longer overwrite the work space |
contain a zillion UTF-8 characters no longer overwrite the work space |
| 3466 |
|
|
| 3467 |
if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) |
if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) |
| 3468 |
posix_class = 0; |
posix_class = 0; |
| 3469 |
|
|
| 3470 |
/* When PCRE_UCP is set, some of the POSIX classes are converted to |
/* When PCRE_UCP is set, some of the POSIX classes are converted to |
| 3471 |
different escape sequences that use Unicode properties. */ |
different escape sequences that use Unicode properties. */ |
| 3472 |
|
|
| 3473 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 3474 |
if ((options & PCRE_UCP) != 0) |
if ((options & PCRE_UCP) != 0) |
| 3475 |
{ |
{ |
| 3476 |
int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0); |
int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0); |
| 3477 |
if (posix_substitutes[pc] != NULL) |
if (posix_substitutes[pc] != NULL) |
| 3478 |
{ |
{ |
| 3479 |
nestptr = tempptr + 1; |
nestptr = tempptr + 1; |
| 3480 |
ptr = posix_substitutes[pc] - 1; |
ptr = posix_substitutes[pc] - 1; |
| 3481 |
continue; |
continue; |
| 3482 |
} |
} |
| 3483 |
} |
} |
| 3484 |
#endif |
#endif |
| 3485 |
/* In the non-UCP case, we build the bit map for the POSIX class in a |
/* In the non-UCP case, we build the bit map for the POSIX class in a |
| 3486 |
chunk of local store because we may be adding and subtracting from it, |
chunk of local store because we may be adding and subtracting from it, |
| 3487 |
and we don't want to subtract bits that may be in the main map already. |
and we don't want to subtract bits that may be in the main map already. |
| 3568 |
case ESC_SU: |
case ESC_SU: |
| 3569 |
nestptr = ptr; |
nestptr = ptr; |
| 3570 |
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ |
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ |
| 3571 |
class_charcount -= 2; /* Undo! */ |
class_charcount -= 2; /* Undo! */ |
| 3572 |
continue; |
continue; |
| 3573 |
#endif |
#endif |
| 3574 |
case ESC_d: |
case ESC_d: |
| 3589 |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; |
| 3590 |
continue; |
continue; |
| 3591 |
|
|
| 3592 |
|
/* Perl 5.004 onwards omits VT from \s, but we must preserve it |
| 3593 |
|
if it was previously set by something earlier in the character |
| 3594 |
|
class. */ |
| 3595 |
|
|
| 3596 |
case ESC_s: |
case ESC_s: |
| 3597 |
for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; |
classbits[0] |= cbits[cbit_space]; |
| 3598 |
classbits[1] &= ~0x08; /* Perl 5.004 onwards omits VT from \s */ |
classbits[1] |= cbits[cbit_space+1] & ~0x08; |
| 3599 |
|
for (c = 2; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; |
| 3600 |
continue; |
continue; |
| 3601 |
|
|
| 3602 |
case ESC_S: |
case ESC_S: |
| 4015 |
|
|
| 4016 |
In UTF-8 mode, we can optimize the negative case only if there were no |
In UTF-8 mode, we can optimize the negative case only if there were no |
| 4017 |
characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR |
characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR |
| 4018 |
operate on single-bytes only. This is an historical hangover. Maybe one day |
operate on single-bytes characters only. This is an historical hangover. |
| 4019 |
we can tidy these opcodes to handle multi-byte characters. |
Maybe one day we can tidy these opcodes to handle multi-byte characters. |
| 4020 |
|
|
| 4021 |
The optimization throws away the bit map. We turn the item into a |
The optimization throws away the bit map. We turn the item into a |
| 4022 |
1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note |
1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative. |
| 4023 |
that OP_NOT does not support multibyte characters. In the positive case, it |
Note that OP_NOT[I] does not support multibyte characters. In the positive |
| 4024 |
can cause firstbyte to be set. Otherwise, there can be no first char if |
case, it can cause firstbyte to be set. Otherwise, there can be no first |
| 4025 |
this item is first, whatever repeat count may follow. In the case of |
char if this item is first, whatever repeat count may follow. In the case |
| 4026 |
reqbyte, save the previous value for reinstating. */ |
of reqbyte, save the previous value for reinstating. */ |
| 4027 |
|
|
| 4028 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 4029 |
if (class_charcount == 1 && !class_utf8 && |
if (class_charcount == 1 && !class_utf8 && |
| 4030 |
(!utf8 || !negate_class || class_lastchar < 128)) |
(!utf8 || !negate_class || class_lastchar < 128)) |
| 4034 |
{ |
{ |
| 4035 |
zeroreqbyte = reqbyte; |
zeroreqbyte = reqbyte; |
| 4036 |
|
|
| 4037 |
/* The OP_NOT opcode works on one-byte characters only. */ |
/* The OP_NOT[I] opcodes work on one-byte characters only. */ |
| 4038 |
|
|
| 4039 |
if (negate_class) |
if (negate_class) |
| 4040 |
{ |
{ |
| 4041 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 4042 |
zerofirstbyte = firstbyte; |
zerofirstbyte = firstbyte; |
| 4043 |
*code++ = OP_NOT; |
*code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT; |
| 4044 |
*code++ = class_lastchar; |
*code++ = class_lastchar; |
| 4045 |
break; |
break; |
| 4046 |
} |
} |
| 4104 |
} |
} |
| 4105 |
#endif |
#endif |
| 4106 |
|
|
| 4107 |
/* If there are no characters > 255, or they are all to be included or |
/* If there are no characters > 255, or they are all to be included or |
| 4108 |
excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the |
excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the |
| 4109 |
whole class was negated and whether there were negative specials such as \S |
whole class was negated and whether there were negative specials such as \S |
| 4110 |
(non-UCP) in the class. Then copy the 32-byte map into the code vector, |
(non-UCP) in the class. Then copy the 32-byte map into the code vector, |
| 4198 |
the first thing in a branch because the x will have gone into firstbyte |
the first thing in a branch because the x will have gone into firstbyte |
| 4199 |
instead. */ |
instead. */ |
| 4200 |
|
|
| 4201 |
if (*previous == OP_CHAR || *previous == OP_CHARNC) |
if (*previous == OP_CHAR || *previous == OP_CHARI) |
| 4202 |
{ |
{ |
| 4203 |
|
op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR; |
| 4204 |
|
|
| 4205 |
/* Deal with UTF-8 characters that take up more than one byte. It's |
/* Deal with UTF-8 characters that take up more than one byte. It's |
| 4206 |
easier to write this out separately than try to macrify it. Use c to |
easier to write this out separately than try to macrify it. Use c to |
| 4207 |
hold the length of the character in bytes, plus 0x80 to flag that it's a |
hold the length of the character in bytes, plus 0x80 to flag that it's a |
| 4246 |
/* If previous was a single negated character ([^a] or similar), we use |
/* If previous was a single negated character ([^a] or similar), we use |
| 4247 |
one of the special opcodes, replacing it. The code is shared with single- |
one of the special opcodes, replacing it. The code is shared with single- |
| 4248 |
character repeats by setting opt_type to add a suitable offset into |
character repeats by setting opt_type to add a suitable offset into |
| 4249 |
repeat_type. We can also test for auto-possessification. OP_NOT is |
repeat_type. We can also test for auto-possessification. OP_NOT and OP_NOTI |
| 4250 |
currently used only for single-byte chars. */ |
are currently used only for single-byte chars. */ |
| 4251 |
|
|
| 4252 |
else if (*previous == OP_NOT) |
else if (*previous == OP_NOT || *previous == OP_NOTI) |
| 4253 |
{ |
{ |
| 4254 |
op_type = OP_NOTSTAR - OP_STAR; /* Use "not" opcodes */ |
op_type = ((*previous == OP_NOT)? OP_NOTSTAR : OP_NOTSTARI) - OP_STAR; |
| 4255 |
c = previous[1]; |
c = previous[1]; |
| 4256 |
if (!possessive_quantifier && |
if (!possessive_quantifier && |
| 4257 |
repeat_max < 0 && |
repeat_max < 0 && |
| 4448 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 4449 |
*previous == OP_XCLASS || |
*previous == OP_XCLASS || |
| 4450 |
#endif |
#endif |
| 4451 |
*previous == OP_REF) |
*previous == OP_REF || |
| 4452 |
|
*previous == OP_REFI) |
| 4453 |
{ |
{ |
| 4454 |
if (repeat_max == 0) |
if (repeat_max == 0) |
| 4455 |
{ |
{ |
| 4503 |
|
|
| 4504 |
/* If the maximum repeat count is unlimited, find the end of the bracket |
/* If the maximum repeat count is unlimited, find the end of the bracket |
| 4505 |
by scanning through from the start, and compute the offset back to it |
by scanning through from the start, and compute the offset back to it |
| 4506 |
from the current code pointer. There may be an OP_OPT setting following |
from the current code pointer. */ |
|
the final KET, so we can't find the end just by going back from the code |
|
|
pointer. */ |
|
| 4507 |
|
|
| 4508 |
if (repeat_max == -1) |
if (repeat_max == -1) |
| 4509 |
{ |
{ |
| 4803 |
case OP_QUERY: *tempcode = OP_POSQUERY; break; |
case OP_QUERY: *tempcode = OP_POSQUERY; break; |
| 4804 |
case OP_UPTO: *tempcode = OP_POSUPTO; break; |
case OP_UPTO: *tempcode = OP_POSUPTO; break; |
| 4805 |
|
|
| 4806 |
case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break; |
case OP_STARI: *tempcode = OP_POSSTARI; break; |
| 4807 |
case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break; |
case OP_PLUSI: *tempcode = OP_POSPLUSI; break; |
| 4808 |
case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break; |
case OP_QUERYI: *tempcode = OP_POSQUERYI; break; |
| 4809 |
case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break; |
case OP_UPTOI: *tempcode = OP_POSUPTOI; break; |
| 4810 |
|
|
| 4811 |
case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break; |
case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break; |
| 4812 |
case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break; |
case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break; |
| 4813 |
case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break; |
case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break; |
| 4814 |
case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break; |
case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break; |
| 4815 |
|
|
| 4816 |
|
case OP_NOTSTARI: *tempcode = OP_NOTPOSSTARI; break; |
| 4817 |
|
case OP_NOTPLUSI: *tempcode = OP_NOTPOSPLUSI; break; |
| 4818 |
|
case OP_NOTQUERYI: *tempcode = OP_NOTPOSQUERYI; break; |
| 4819 |
|
case OP_NOTUPTOI: *tempcode = OP_NOTPOSUPTOI; break; |
| 4820 |
|
|
| 4821 |
|
case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break; |
| 4822 |
|
case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break; |
| 4823 |
|
case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break; |
| 4824 |
|
case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break; |
| 4825 |
|
|
| 4826 |
/* Because we are moving code along, we must ensure that any |
/* Because we are moving code along, we must ensure that any |
| 4827 |
pending recursive references are updated. */ |
pending recursive references are updated. */ |
| 4828 |
|
|
| 4881 |
arg = ++ptr; |
arg = ++ptr; |
| 4882 |
while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0 |
while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0 |
| 4883 |
|| *ptr == '_') ptr++; |
|| *ptr == '_') ptr++; |
| 4884 |
arglen = ptr - arg; |
arglen = (int)(ptr - arg); |
| 4885 |
} |
} |
| 4886 |
|
|
| 4887 |
if (*ptr != CHAR_RIGHT_PARENTHESIS) |
if (*ptr != CHAR_RIGHT_PARENTHESIS) |
| 4919 |
*errorcodeptr = ERR66; |
*errorcodeptr = ERR66; |
| 4920 |
goto FAILED; |
goto FAILED; |
| 4921 |
} |
} |
| 4922 |
*code++ = verbs[i].op; |
*code = verbs[i].op; |
| 4923 |
|
if (*code++ == OP_THEN) |
| 4924 |
|
{ |
| 4925 |
|
PUT(code, 0, code - bcptr->current_branch - 1); |
| 4926 |
|
code += LINK_SIZE; |
| 4927 |
|
} |
| 4928 |
} |
} |
| 4929 |
|
|
| 4930 |
else |
else |
| 4934 |
*errorcodeptr = ERR59; |
*errorcodeptr = ERR59; |
| 4935 |
goto FAILED; |
goto FAILED; |
| 4936 |
} |
} |
| 4937 |
*code++ = verbs[i].op_arg; |
*code = verbs[i].op_arg; |
| 4938 |
|
if (*code++ == OP_THEN_ARG) |
| 4939 |
|
{ |
| 4940 |
|
PUT(code, 0, code - bcptr->current_branch - 1); |
| 4941 |
|
code += LINK_SIZE; |
| 4942 |
|
} |
| 4943 |
*code++ = arglen; |
*code++ = arglen; |
| 4944 |
memcpy(code, arg, arglen); |
memcpy(code, arg, arglen); |
| 4945 |
code += arglen; |
code += arglen; |
| 5133 |
/* Search the pattern for a forward reference */ |
/* Search the pattern for a forward reference */ |
| 5134 |
|
|
| 5135 |
else if ((i = find_parens(cd, name, namelen, |
else if ((i = find_parens(cd, name, namelen, |
| 5136 |
(options & PCRE_EXTENDED) != 0)) > 0) |
(options & PCRE_EXTENDED) != 0, utf8)) > 0) |
| 5137 |
{ |
{ |
| 5138 |
PUT2(code, 2+LINK_SIZE, i); |
PUT2(code, 2+LINK_SIZE, i); |
| 5139 |
code[1+LINK_SIZE]++; |
code[1+LINK_SIZE]++; |
| 5434 |
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; |
| 5435 |
namelen = (int)(ptr - name); |
namelen = (int)(ptr - name); |
| 5436 |
|
|
| 5437 |
/* In the pre-compile phase, do a syntax check and set a dummy |
/* In the pre-compile phase, do a syntax check. We used to just set |
| 5438 |
reference number. */ |
a dummy reference number, because it was not used in the first pass. |
| 5439 |
|
However, with the change of recursive back references to be atomic, |
| 5440 |
|
we have to look for the number so that this state can be identified, as |
| 5441 |
|
otherwise the incorrect length is computed. If it's not a backwards |
| 5442 |
|
reference, the dummy number will do. */ |
| 5443 |
|
|
| 5444 |
if (lengthptr != NULL) |
if (lengthptr != NULL) |
| 5445 |
{ |
{ |
| 5446 |
|
const uschar *temp; |
| 5447 |
|
|
| 5448 |
if (namelen == 0) |
if (namelen == 0) |
| 5449 |
{ |
{ |
| 5450 |
*errorcodeptr = ERR62; |
*errorcodeptr = ERR62; |
| 5460 |
*errorcodeptr = ERR48; |
*errorcodeptr = ERR48; |
| 5461 |
goto FAILED; |
goto FAILED; |
| 5462 |
} |
} |
| 5463 |
recno = 0; |
|
| 5464 |
|
/* The name table does not exist in the first pass, so we cannot |
| 5465 |
|
do a simple search as in the code below. Instead, we have to scan the |
| 5466 |
|
pattern to find the number. It is important that we scan it only as |
| 5467 |
|
far as we have got because the syntax of named subpatterns has not |
| 5468 |
|
been checked for the rest of the pattern, and find_parens() assumes |
| 5469 |
|
correct syntax. In any case, it's a waste of resources to scan |
| 5470 |
|
further. We stop the scan at the current point by temporarily |
| 5471 |
|
adjusting the value of cd->endpattern. */ |
| 5472 |
|
|
| 5473 |
|
temp = cd->end_pattern; |
| 5474 |
|
cd->end_pattern = ptr; |
| 5475 |
|
recno = find_parens(cd, name, namelen, |
| 5476 |
|
(options & PCRE_EXTENDED) != 0, utf8); |
| 5477 |
|
cd->end_pattern = temp; |
| 5478 |
|
if (recno < 0) recno = 0; /* Forward ref; set dummy number */ |
| 5479 |
} |
} |
| 5480 |
|
|
| 5481 |
/* In the real compile, seek the name in the table. We check the name |
/* In the real compile, seek the name in the table. We check the name |
| 5500 |
} |
} |
| 5501 |
else if ((recno = /* Forward back reference */ |
else if ((recno = /* Forward back reference */ |
| 5502 |
find_parens(cd, name, namelen, |
find_parens(cd, name, namelen, |
| 5503 |
(options & PCRE_EXTENDED) != 0)) <= 0) |
(options & PCRE_EXTENDED) != 0, utf8)) <= 0) |
| 5504 |
{ |
{ |
| 5505 |
*errorcodeptr = ERR15; |
*errorcodeptr = ERR15; |
| 5506 |
goto FAILED; |
goto FAILED; |
| 5611 |
if (called == NULL) |
if (called == NULL) |
| 5612 |
{ |
{ |
| 5613 |
if (find_parens(cd, NULL, recno, |
if (find_parens(cd, NULL, recno, |
| 5614 |
(options & PCRE_EXTENDED) != 0) < 0) |
(options & PCRE_EXTENDED) != 0, utf8) < 0) |
| 5615 |
{ |
{ |
| 5616 |
*errorcodeptr = ERR15; |
*errorcodeptr = ERR15; |
| 5617 |
goto FAILED; |
goto FAILED; |
| 5727 |
} |
} |
| 5728 |
else |
else |
| 5729 |
{ |
{ |
|
if ((options & PCRE_IMS) != (newoptions & PCRE_IMS)) |
|
|
{ |
|
|
*code++ = OP_OPT; |
|
|
*code++ = newoptions & PCRE_IMS; |
|
|
} |
|
| 5730 |
greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); |
greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); |
| 5731 |
greedy_non_default = greedy_default ^ 1; |
greedy_non_default = greedy_default ^ 1; |
| 5732 |
req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; |
req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; |
| 5945 |
|
|
| 5946 |
/* ===================================================================*/ |
/* ===================================================================*/ |
| 5947 |
/* Handle metasequences introduced by \. For ones like \d, the ESC_ values |
/* Handle metasequences introduced by \. For ones like \d, the ESC_ values |
| 5948 |
are arranged to be the negation of the corresponding OP_values in the |
are arranged to be the negation of the corresponding OP_values in the |
| 5949 |
default case when PCRE_UCP is not set. For the back references, the values |
default case when PCRE_UCP is not set. For the back references, the values |
| 5950 |
are ESC_REF plus the reference number. Only back references and those types |
are ESC_REF plus the reference number. Only back references and those types |
| 5951 |
that consume a character may be repeated. We can test for values between |
that consume a character may be repeated. We can test for values between |
| 6064 |
HANDLE_REFERENCE: /* Come here from named backref handling */ |
HANDLE_REFERENCE: /* Come here from named backref handling */ |
| 6065 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 6066 |
previous = code; |
previous = code; |
| 6067 |
*code++ = OP_REF; |
*code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF; |
| 6068 |
PUT2INC(code, 0, recno); |
PUT2INC(code, 0, recno); |
| 6069 |
cd->backref_map |= (recno < 32)? (1 << recno) : 1; |
cd->backref_map |= (recno < 32)? (1 << recno) : 1; |
| 6070 |
if (recno > cd->top_backref) cd->top_backref = recno; |
if (recno > cd->top_backref) cd->top_backref = recno; |
| 6123 |
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ |
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ |
| 6124 |
} |
} |
| 6125 |
else |
else |
| 6126 |
#endif |
#endif |
| 6127 |
{ |
{ |
| 6128 |
previous = (-c > ESC_b && -c < ESC_Z)? code : NULL; |
previous = (-c > ESC_b && -c < ESC_Z)? code : NULL; |
| 6129 |
*code++ = -c; |
*code++ = -c; |
| 6130 |
} |
} |
| 6131 |
} |
} |
| 6132 |
continue; |
continue; |
| 6133 |
} |
} |
| 6172 |
|
|
| 6173 |
ONE_CHAR: |
ONE_CHAR: |
| 6174 |
previous = code; |
previous = code; |
| 6175 |
*code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR; |
*code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR; |
| 6176 |
for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; |
for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; |
| 6177 |
|
|
| 6178 |
/* Remember if \r or \n were seen */ |
/* Remember if \r or \n were seen */ |
| 6236 |
/* On entry, ptr is pointing past the bracket character, but on return it |
/* On entry, ptr is pointing past the bracket character, but on return it |
| 6237 |
points to the closing bracket, or vertical bar, or end of string. The code |
points to the closing bracket, or vertical bar, or end of string. The code |
| 6238 |
variable is pointing at the byte into which the BRA operator has been stored. |
variable is pointing at the byte into which the BRA operator has been stored. |
|
If the ims options are changed at the start (for a (?ims: group) or during any |
|
|
branch, we need to insert an OP_OPT item at the start of every following branch |
|
|
to ensure they get set correctly at run time, and also pass the new options |
|
|
into every subsequent branch compile. |
|
|
|
|
| 6239 |
This function is used during the pre-compile phase when we are trying to find |
This function is used during the pre-compile phase when we are trying to find |
| 6240 |
out the amount of memory needed, as well as during the real compile phase. The |
out the amount of memory needed, as well as during the real compile phase. The |
| 6241 |
value of lengthptr distinguishes the two phases. |
value of lengthptr distinguishes the two phases. |
| 6327 |
|
|
| 6328 |
if (reset_bracount) cd->bracount = orig_bracount; |
if (reset_bracount) cd->bracount = orig_bracount; |
| 6329 |
|
|
|
/* Handle a change of ims options at the start of the branch */ |
|
|
|
|
|
if ((options & PCRE_IMS) != oldims) |
|
|
{ |
|
|
*code++ = OP_OPT; |
|
|
*code++ = options & PCRE_IMS; |
|
|
length += 2; |
|
|
} |
|
|
|
|
| 6330 |
/* Set up dummy OP_REVERSE if lookbehind assertion */ |
/* Set up dummy OP_REVERSE if lookbehind assertion */ |
| 6331 |
|
|
| 6332 |
if (lookbehind) |
if (lookbehind) |
| 6483 |
cd->open_caps = cd->open_caps->next; |
cd->open_caps = cd->open_caps->next; |
| 6484 |
} |
} |
| 6485 |
|
|
|
/* Reset options if needed. */ |
|
|
|
|
|
if ((options & PCRE_IMS) != oldims && *ptr == CHAR_RIGHT_PARENTHESIS) |
|
|
{ |
|
|
*code++ = OP_OPT; |
|
|
*code++ = oldims; |
|
|
length += 2; |
|
|
} |
|
|
|
|
| 6486 |
/* Retain the highest bracket number, in case resetting was used. */ |
/* Retain the highest bracket number, in case resetting was used. */ |
| 6487 |
|
|
| 6488 |
cd->bracount = max_bracount; |
cd->bracount = max_bracount; |
| 6542 |
/* Try to find out if this is an anchored regular expression. Consider each |
/* Try to find out if this is an anchored regular expression. Consider each |
| 6543 |
alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket |
alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket |
| 6544 |
all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then |
all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then |
| 6545 |
it's anchored. However, if this is a multiline pattern, then only OP_SOD |
it's anchored. However, if this is a multiline pattern, then only OP_SOD will |
| 6546 |
counts, since OP_CIRC can match in the middle. |
be found, because ^ generates OP_CIRCM in that mode. |
| 6547 |
|
|
| 6548 |
We can also consider a regex to be anchored if OP_SOM starts all its branches. |
We can also consider a regex to be anchored if OP_SOM starts all its branches. |
| 6549 |
This is the code for \G, which means "match at start of match position, taking |
This is the code for \G, which means "match at start of match position, taking |
| 6617 |
|
|
| 6618 |
/* Check for explicit anchoring */ |
/* Check for explicit anchoring */ |
| 6619 |
|
|
| 6620 |
else if (op != OP_SOD && op != OP_SOM && |
else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE; |
|
((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC)) |
|
|
return FALSE; |
|
| 6621 |
code += GET(code, 1); |
code += GET(code, 1); |
| 6622 |
} |
} |
| 6623 |
while (*code == OP_ALT); /* Loop for each alternative */ |
while (*code == OP_ALT); /* Loop for each alternative */ |
| 6717 |
|
|
| 6718 |
/* Check for explicit circumflex */ |
/* Check for explicit circumflex */ |
| 6719 |
|
|
| 6720 |
else if (op != OP_CIRC) return FALSE; |
else if (op != OP_CIRC && op != OP_CIRCM) return FALSE; |
| 6721 |
|
|
| 6722 |
/* Move on to the next alternative */ |
/* Move on to the next alternative */ |
| 6723 |
|
|
| 6778 |
scode += 2; |
scode += 2; |
| 6779 |
|
|
| 6780 |
case OP_CHAR: |
case OP_CHAR: |
| 6781 |
case OP_CHARNC: |
case OP_CHARI: |
| 6782 |
case OP_PLUS: |
case OP_PLUS: |
| 6783 |
case OP_MINPLUS: |
case OP_MINPLUS: |
| 6784 |
case OP_POSPLUS: |
case OP_POSPLUS: |
| 6911 |
{ skipatstart += 7; options |= PCRE_UTF8; continue; } |
{ skipatstart += 7; options |= PCRE_UTF8; continue; } |
| 6912 |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0) |
| 6913 |
{ skipatstart += 6; options |= PCRE_UCP; continue; } |
{ skipatstart += 6; options |= PCRE_UCP; continue; } |
| 6914 |
|
else if (strncmp((char *)(ptr+skipatstart+2), STRING_NO_START_OPT_RIGHTPAR, 13) == 0) |
| 6915 |
|
{ skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; } |
| 6916 |
|
|
| 6917 |
if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0) |
if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0) |
| 6918 |
{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; } |
{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; } |
| 6936 |
options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr; |
options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr; |
| 6937 |
else break; |
else break; |
| 6938 |
} |
} |
| 6939 |
|
|
| 6940 |
utf8 = (options & PCRE_UTF8) != 0; |
utf8 = (options & PCRE_UTF8) != 0; |
| 6941 |
|
|
| 6942 |
/* Can't support UTF8 unless PCRE has been compiled to include the code. */ |
/* Can't support UTF8 unless PCRE has been compiled to include the code. The |
| 6943 |
|
return of an error code from _pcre_valid_utf8() is a new feature, introduced in |
| 6944 |
|
release 8.13. The only use we make of it here is to adjust the offset value to |
| 6945 |
|
the end of the string for a short string error, for compatibility with previous |
| 6946 |
|
versions. */ |
| 6947 |
|
|
| 6948 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 6949 |
if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 && |
if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 && |
| 6950 |
(*erroroffset = _pcre_valid_utf8((USPTR)pattern, -1)) >= 0) |
(*erroroffset = _pcre_valid_utf8((USPTR)pattern, -1, &errorcode)) >= 0) |
| 6951 |
{ |
{ |
| 6952 |
errorcode = ERR44; |
errorcode = ERR44; |
| 6953 |
goto PCRE_EARLY_ERROR_RETURN2; |
goto PCRE_EARLY_ERROR_RETURN2; |
| 6966 |
if ((options & PCRE_UCP) != 0) |
if ((options & PCRE_UCP) != 0) |
| 6967 |
{ |
{ |
| 6968 |
errorcode = ERR67; |
errorcode = ERR67; |
| 6969 |
goto PCRE_EARLY_ERROR_RETURN; |
goto PCRE_EARLY_ERROR_RETURN; |
| 6970 |
} |
} |
| 6971 |
#endif |
#endif |
| 6972 |
|
|
| 6973 |
/* Check validity of \R options. */ |
/* Check validity of \R options. */ |