| 43 |
|
|
| 44 |
|
|
| 45 |
#ifdef HAVE_CONFIG_H |
#ifdef HAVE_CONFIG_H |
| 46 |
#include <config.h> |
#include "config.h" |
| 47 |
#endif |
#endif |
| 48 |
|
|
| 49 |
#define NLBLOCK cd /* Block containing newline information */ |
#define NLBLOCK cd /* Block containing newline information */ |
| 140 |
#endif |
#endif |
| 141 |
|
|
| 142 |
|
|
| 143 |
/* Table of special "verbs" like (*PRUNE) */ |
/* Table of special "verbs" like (*PRUNE). This is a short table, so it is |
| 144 |
|
searched linearly. Put all the names into a single string, in order to reduce |
| 145 |
|
the number of relocations when a shared library is dynamically linked. */ |
| 146 |
|
|
| 147 |
typedef struct verbitem { |
typedef struct verbitem { |
|
const char *name; |
|
| 148 |
int len; |
int len; |
| 149 |
int op; |
int op; |
| 150 |
} verbitem; |
} verbitem; |
| 151 |
|
|
| 152 |
|
static const char verbnames[] = |
| 153 |
|
"ACCEPT\0" |
| 154 |
|
"COMMIT\0" |
| 155 |
|
"F\0" |
| 156 |
|
"FAIL\0" |
| 157 |
|
"PRUNE\0" |
| 158 |
|
"SKIP\0" |
| 159 |
|
"THEN"; |
| 160 |
|
|
| 161 |
static verbitem verbs[] = { |
static verbitem verbs[] = { |
| 162 |
{ "ACCEPT", 6, OP_ACCEPT }, |
{ 6, OP_ACCEPT }, |
| 163 |
{ "COMMIT", 6, OP_COMMIT }, |
{ 6, OP_COMMIT }, |
| 164 |
{ "F", 1, OP_FAIL }, |
{ 1, OP_FAIL }, |
| 165 |
{ "FAIL", 4, OP_FAIL }, |
{ 4, OP_FAIL }, |
| 166 |
{ "PRUNE", 5, OP_PRUNE }, |
{ 5, OP_PRUNE }, |
| 167 |
{ "SKIP", 4, OP_SKIP }, |
{ 4, OP_SKIP }, |
| 168 |
{ "THEN", 4, OP_THEN } |
{ 4, OP_THEN } |
| 169 |
}; |
}; |
| 170 |
|
|
| 171 |
static int verbcount = sizeof(verbs)/sizeof(verbitem); |
static int verbcount = sizeof(verbs)/sizeof(verbitem); |
| 172 |
|
|
| 173 |
|
|
| 174 |
/* Tables of names of POSIX character classes and their lengths. The list is |
/* Tables of names of POSIX character classes and their lengths. The names are |
| 175 |
terminated by a zero length entry. The first three must be alpha, lower, upper, |
now all in a single string, to reduce the number of relocations when a shared |
| 176 |
as this is assumed for handling case independence. */ |
library is dynamically loaded. The list of lengths is terminated by a zero |
| 177 |
|
length entry. The first three must be alpha, lower, upper, as this is assumed |
| 178 |
static const char *const posix_names[] = { |
for handling case independence. */ |
| 179 |
"alpha", "lower", "upper", |
|
| 180 |
"alnum", "ascii", "blank", "cntrl", "digit", "graph", |
static const char posix_names[] = |
| 181 |
"print", "punct", "space", "word", "xdigit" }; |
"alpha\0" "lower\0" "upper\0" "alnum\0" "ascii\0" "blank\0" |
| 182 |
|
"cntrl\0" "digit\0" "graph\0" "print\0" "punct\0" "space\0" |
| 183 |
|
"word\0" "xdigit"; |
| 184 |
|
|
| 185 |
static const uschar posix_name_lengths[] = { |
static const uschar posix_name_lengths[] = { |
| 186 |
5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; |
5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; |
| 219 |
/* The texts of compile-time error messages. These are "char *" because they |
/* The texts of compile-time error messages. These are "char *" because they |
| 220 |
are passed to the outside world. Do not ever re-use any error number, because |
are passed to the outside world. Do not ever re-use any error number, because |
| 221 |
they are documented. Always add a new error instead. Messages marked DEAD below |
they are documented. Always add a new error instead. Messages marked DEAD below |
| 222 |
are no longer used. */ |
are no longer used. This used to be a table of strings, but in order to reduce |
| 223 |
|
the number of relocations needed when a shared library is loaded dynamically, |
| 224 |
static const char *error_texts[] = { |
it is now one long string. We cannot use a table of offsets, because the |
| 225 |
"no error", |
lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we |
| 226 |
"\\ at end of pattern", |
simply count through to the one we want - this isn't a performance issue |
| 227 |
"\\c at end of pattern", |
because these strings are used only when there is a compilation error. */ |
| 228 |
"unrecognized character follows \\", |
|
| 229 |
"numbers out of order in {} quantifier", |
static const char error_texts[] = |
| 230 |
|
"no error\0" |
| 231 |
|
"\\ at end of pattern\0" |
| 232 |
|
"\\c at end of pattern\0" |
| 233 |
|
"unrecognized character follows \\\0" |
| 234 |
|
"numbers out of order in {} quantifier\0" |
| 235 |
/* 5 */ |
/* 5 */ |
| 236 |
"number too big in {} quantifier", |
"number too big in {} quantifier\0" |
| 237 |
"missing terminating ] for character class", |
"missing terminating ] for character class\0" |
| 238 |
"invalid escape sequence in character class", |
"invalid escape sequence in character class\0" |
| 239 |
"range out of order in character class", |
"range out of order in character class\0" |
| 240 |
"nothing to repeat", |
"nothing to repeat\0" |
| 241 |
/* 10 */ |
/* 10 */ |
| 242 |
"operand of unlimited repeat could match the empty string", /** DEAD **/ |
"operand of unlimited repeat could match the empty string\0" /** DEAD **/ |
| 243 |
"internal error: unexpected repeat", |
"internal error: unexpected repeat\0" |
| 244 |
"unrecognized character after (?", |
"unrecognized character after (? or (?-\0" |
| 245 |
"POSIX named classes are supported only within a class", |
"POSIX named classes are supported only within a class\0" |
| 246 |
"missing )", |
"missing )\0" |
| 247 |
/* 15 */ |
/* 15 */ |
| 248 |
"reference to non-existent subpattern", |
"reference to non-existent subpattern\0" |
| 249 |
"erroffset passed as NULL", |
"erroffset passed as NULL\0" |
| 250 |
"unknown option bit(s) set", |
"unknown option bit(s) set\0" |
| 251 |
"missing ) after comment", |
"missing ) after comment\0" |
| 252 |
"parentheses nested too deeply", /** DEAD **/ |
"parentheses nested too deeply\0" /** DEAD **/ |
| 253 |
/* 20 */ |
/* 20 */ |
| 254 |
"regular expression is too large", |
"regular expression is too large\0" |
| 255 |
"failed to get memory", |
"failed to get memory\0" |
| 256 |
"unmatched parentheses", |
"unmatched parentheses\0" |
| 257 |
"internal error: code overflow", |
"internal error: code overflow\0" |
| 258 |
"unrecognized character after (?<", |
"unrecognized character after (?<\0" |
| 259 |
/* 25 */ |
/* 25 */ |
| 260 |
"lookbehind assertion is not fixed length", |
"lookbehind assertion is not fixed length\0" |
| 261 |
"malformed number or name after (?(", |
"malformed number or name after (?(\0" |
| 262 |
"conditional group contains more than two branches", |
"conditional group contains more than two branches\0" |
| 263 |
"assertion expected after (?(", |
"assertion expected after (?(\0" |
| 264 |
"(?R or (?[+-]digits must be followed by )", |
"(?R or (?[+-]digits must be followed by )\0" |
| 265 |
/* 30 */ |
/* 30 */ |
| 266 |
"unknown POSIX class name", |
"unknown POSIX class name\0" |
| 267 |
"POSIX collating elements are not supported", |
"POSIX collating elements are not supported\0" |
| 268 |
"this version of PCRE is not compiled with PCRE_UTF8 support", |
"this version of PCRE is not compiled with PCRE_UTF8 support\0" |
| 269 |
"spare error", /** DEAD **/ |
"spare error\0" /** DEAD **/ |
| 270 |
"character value in \\x{...} sequence is too large", |
"character value in \\x{...} sequence is too large\0" |
| 271 |
/* 35 */ |
/* 35 */ |
| 272 |
"invalid condition (?(0)", |
"invalid condition (?(0)\0" |
| 273 |
"\\C not allowed in lookbehind assertion", |
"\\C not allowed in lookbehind assertion\0" |
| 274 |
"PCRE does not support \\L, \\l, \\N, \\U, or \\u", |
"PCRE does not support \\L, \\l, \\N, \\U, or \\u\0" |
| 275 |
"number after (?C is > 255", |
"number after (?C is > 255\0" |
| 276 |
"closing ) for (?C expected", |
"closing ) for (?C expected\0" |
| 277 |
/* 40 */ |
/* 40 */ |
| 278 |
"recursive call could loop indefinitely", |
"recursive call could loop indefinitely\0" |
| 279 |
"unrecognized character after (?P", |
"unrecognized character after (?P\0" |
| 280 |
"syntax error in subpattern name (missing terminator)", |
"syntax error in subpattern name (missing terminator)\0" |
| 281 |
"two named subpatterns have the same name", |
"two named subpatterns have the same name\0" |
| 282 |
"invalid UTF-8 string", |
"invalid UTF-8 string\0" |
| 283 |
/* 45 */ |
/* 45 */ |
| 284 |
"support for \\P, \\p, and \\X has not been compiled", |
"support for \\P, \\p, and \\X has not been compiled\0" |
| 285 |
"malformed \\P or \\p sequence", |
"malformed \\P or \\p sequence\0" |
| 286 |
"unknown property name after \\P or \\p", |
"unknown property name after \\P or \\p\0" |
| 287 |
"subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)", |
"subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0" |
| 288 |
"too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")", |
"too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0" |
| 289 |
/* 50 */ |
/* 50 */ |
| 290 |
"repeated subpattern is too long", /** DEAD **/ |
"repeated subpattern is too long\0" /** DEAD **/ |
| 291 |
"octal value is greater than \\377 (not in UTF-8 mode)", |
"octal value is greater than \\377 (not in UTF-8 mode)\0" |
| 292 |
"internal error: overran compiling workspace", |
"internal error: overran compiling workspace\0" |
| 293 |
"internal error: previously-checked referenced subpattern not found", |
"internal error: previously-checked referenced subpattern not found\0" |
| 294 |
"DEFINE group contains more than one branch", |
"DEFINE group contains more than one branch\0" |
| 295 |
/* 55 */ |
/* 55 */ |
| 296 |
"repeating a DEFINE group is not allowed", |
"repeating a DEFINE group is not allowed\0" |
| 297 |
"inconsistent NEWLINE options", |
"inconsistent NEWLINE options\0" |
| 298 |
"\\g is not followed by a braced name or an optionally braced non-zero number", |
"\\g is not followed by a braced name or an optionally braced non-zero number\0" |
| 299 |
"(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number", |
"(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number\0" |
| 300 |
"(*VERB) with an argument is not supported", |
"(*VERB) with an argument is not supported\0" |
| 301 |
/* 60 */ |
/* 60 */ |
| 302 |
"(*VERB) not recognized", |
"(*VERB) not recognized\0" |
| 303 |
"number is too big" |
"number is too big\0" |
| 304 |
}; |
"subpattern name expected\0" |
| 305 |
|
"digit expected after (?+"; |
| 306 |
|
|
| 307 |
|
|
| 308 |
/* Table to identify digits and hex digits. This is used when compiling |
/* Table to identify digits and hex digits. This is used when compiling |
| 438 |
|
|
| 439 |
|
|
| 440 |
/************************************************* |
/************************************************* |
| 441 |
|
* Find an error text * |
| 442 |
|
*************************************************/ |
| 443 |
|
|
| 444 |
|
/* The error texts are now all in one long string, to save on relocations. As |
| 445 |
|
some of the text is of unknown length, we can't use a table of offsets. |
| 446 |
|
Instead, just count through the strings. This is not a performance issue |
| 447 |
|
because it happens only when there has been a compilation error. |
| 448 |
|
|
| 449 |
|
Argument: the error number |
| 450 |
|
Returns: pointer to the error string |
| 451 |
|
*/ |
| 452 |
|
|
| 453 |
|
static const char * |
| 454 |
|
find_error_text(int n) |
| 455 |
|
{ |
| 456 |
|
const char *s = error_texts; |
| 457 |
|
for (; n > 0; n--) while (*s++ != 0); |
| 458 |
|
return s; |
| 459 |
|
} |
| 460 |
|
|
| 461 |
|
|
| 462 |
|
/************************************************* |
| 463 |
* Handle escapes * |
* Handle escapes * |
| 464 |
*************************************************/ |
*************************************************/ |
| 465 |
|
|
| 816 |
while (bot < top) |
while (bot < top) |
| 817 |
{ |
{ |
| 818 |
i = (bot + top) >> 1; |
i = (bot + top) >> 1; |
| 819 |
c = strcmp(name, _pcre_utt[i].name); |
c = strcmp(name, _pcre_utt_names + _pcre_utt[i].name_offset); |
| 820 |
if (c == 0) |
if (c == 0) |
| 821 |
{ |
{ |
| 822 |
*dptr = _pcre_utt[i].value; |
*dptr = _pcre_utt[i].value; |
| 1773 |
static int |
static int |
| 1774 |
check_posix_name(const uschar *ptr, int len) |
check_posix_name(const uschar *ptr, int len) |
| 1775 |
{ |
{ |
| 1776 |
|
const char *pn = posix_names; |
| 1777 |
register int yield = 0; |
register int yield = 0; |
| 1778 |
while (posix_name_lengths[yield] != 0) |
while (posix_name_lengths[yield] != 0) |
| 1779 |
{ |
{ |
| 1780 |
if (len == posix_name_lengths[yield] && |
if (len == posix_name_lengths[yield] && |
| 1781 |
strncmp((const char *)ptr, posix_names[yield], len) == 0) return yield; |
strncmp((const char *)ptr, pn, len) == 0) return yield; |
| 1782 |
|
pn += posix_name_lengths[yield] + 1; |
| 1783 |
yield++; |
yield++; |
| 1784 |
} |
} |
| 1785 |
return -1; |
return -1; |
| 2385 |
for (;; ptr++) |
for (;; ptr++) |
| 2386 |
{ |
{ |
| 2387 |
BOOL negate_class; |
BOOL negate_class; |
| 2388 |
|
BOOL should_flip_negation; |
| 2389 |
BOOL possessive_quantifier; |
BOOL possessive_quantifier; |
| 2390 |
BOOL is_quantifier; |
BOOL is_quantifier; |
| 2391 |
BOOL is_recurse; |
BOOL is_recurse; |
| 2634 |
else break; |
else break; |
| 2635 |
} |
} |
| 2636 |
|
|
| 2637 |
|
/* If a class contains a negative special such as \S, we need to flip the |
| 2638 |
|
negation flag at the end, so that support for characters > 255 works |
| 2639 |
|
correctly (they are all included in the class). */ |
| 2640 |
|
|
| 2641 |
|
should_flip_negation = FALSE; |
| 2642 |
|
|
| 2643 |
/* Keep a count of chars with values < 256 so that we can optimize the case |
/* Keep a count of chars with values < 256 so that we can optimize the case |
| 2644 |
of just a single character (as long as it's < 256). However, For higher |
of just a single character (as long as it's < 256). However, For higher |
| 2645 |
valued UTF-8 characters, we don't yet do any optimization. */ |
valued UTF-8 characters, we don't yet do any optimization. */ |
| 2712 |
if (*ptr == '^') |
if (*ptr == '^') |
| 2713 |
{ |
{ |
| 2714 |
local_negate = TRUE; |
local_negate = TRUE; |
| 2715 |
|
should_flip_negation = TRUE; /* Note negative special */ |
| 2716 |
ptr++; |
ptr++; |
| 2717 |
} |
} |
| 2718 |
|
|
| 2815 |
continue; |
continue; |
| 2816 |
|
|
| 2817 |
case ESC_D: |
case ESC_D: |
| 2818 |
|
should_flip_negation = TRUE; |
| 2819 |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; |
| 2820 |
continue; |
continue; |
| 2821 |
|
|
| 2824 |
continue; |
continue; |
| 2825 |
|
|
| 2826 |
case ESC_W: |
case ESC_W: |
| 2827 |
|
should_flip_negation = TRUE; |
| 2828 |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; |
| 2829 |
continue; |
continue; |
| 2830 |
|
|
| 2834 |
continue; |
continue; |
| 2835 |
|
|
| 2836 |
case ESC_S: |
case ESC_S: |
| 2837 |
|
should_flip_negation = TRUE; |
| 2838 |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; |
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; |
| 2839 |
classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */ |
classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */ |
| 2840 |
continue; |
continue; |
| 3340 |
zeroreqbyte = reqbyte; |
zeroreqbyte = reqbyte; |
| 3341 |
|
|
| 3342 |
/* If there are characters with values > 255, we have to compile an |
/* If there are characters with values > 255, we have to compile an |
| 3343 |
extended class, with its own opcode. If there are no characters < 256, |
extended class, with its own opcode, unless there was a negated special |
| 3344 |
we can omit the bitmap in the actual compiled code. */ |
such as \S in the class, because in that case all characters > 255 are in |
| 3345 |
|
the class, so any that were explicitly given as well can be ignored. If |
| 3346 |
|
(when there are explicit characters > 255 that must be listed) there are no |
| 3347 |
|
characters < 256, we can omit the bitmap in the actual compiled code. */ |
| 3348 |
|
|
| 3349 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 3350 |
if (class_utf8) |
if (class_utf8 && !should_flip_negation) |
| 3351 |
{ |
{ |
| 3352 |
*class_utf8data++ = XCL_END; /* Marks the end of extra data */ |
*class_utf8data++ = XCL_END; /* Marks the end of extra data */ |
| 3353 |
*code++ = OP_XCLASS; |
*code++ = OP_XCLASS; |
| 3373 |
} |
} |
| 3374 |
#endif |
#endif |
| 3375 |
|
|
| 3376 |
/* If there are no characters > 255, negate the 32-byte map if necessary, |
/* If there are no characters > 255, set the opcode to OP_CLASS or |
| 3377 |
and copy it into the code vector. If this is the first thing in the branch, |
OP_NCLASS, depending on whether the whole class was negated and whether |
| 3378 |
there can be no first char setting, whatever the repeat count. Any reqbyte |
there were negative specials such as \S in the class. Then copy the 32-byte |
| 3379 |
setting must remain unchanged after any kind of repeat. */ |
map into the code vector, negating it if necessary. */ |
| 3380 |
|
|
| 3381 |
|
*code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS; |
| 3382 |
if (negate_class) |
if (negate_class) |
| 3383 |
{ |
{ |
|
*code++ = OP_NCLASS; |
|
| 3384 |
if (lengthptr == NULL) /* Save time in the pre-compile phase */ |
if (lengthptr == NULL) /* Save time in the pre-compile phase */ |
| 3385 |
for (c = 0; c < 32; c++) code[c] = ~classbits[c]; |
for (c = 0; c < 32; c++) code[c] = ~classbits[c]; |
| 3386 |
} |
} |
| 3387 |
else |
else |
| 3388 |
{ |
{ |
|
*code++ = OP_CLASS; |
|
| 3389 |
memcpy(code, classbits, 32); |
memcpy(code, classbits, 32); |
| 3390 |
} |
} |
| 3391 |
code += 32; |
code += 32; |
| 4079 |
if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0) |
if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0) |
| 4080 |
{ |
{ |
| 4081 |
int i, namelen; |
int i, namelen; |
| 4082 |
|
const char *vn = verbnames; |
| 4083 |
const uschar *name = ++ptr; |
const uschar *name = ++ptr; |
| 4084 |
previous = NULL; |
previous = NULL; |
| 4085 |
while ((cd->ctypes[*++ptr] & ctype_letter) != 0); |
while ((cd->ctypes[*++ptr] & ctype_letter) != 0); |
| 4097 |
for (i = 0; i < verbcount; i++) |
for (i = 0; i < verbcount; i++) |
| 4098 |
{ |
{ |
| 4099 |
if (namelen == verbs[i].len && |
if (namelen == verbs[i].len && |
| 4100 |
strncmp((char *)name, verbs[i].name, namelen) == 0) |
strncmp((char *)name, vn, namelen) == 0) |
| 4101 |
{ |
{ |
| 4102 |
*code = verbs[i].op; |
*code = verbs[i].op; |
| 4103 |
if (*code++ == OP_ACCEPT) cd->had_accept = TRUE; |
if (*code++ == OP_ACCEPT) cd->had_accept = TRUE; |
| 4104 |
break; |
break; |
| 4105 |
} |
} |
| 4106 |
|
vn += verbs[i].len + 1; |
| 4107 |
} |
} |
| 4108 |
if (i < verbcount) continue; |
if (i < verbcount) continue; |
| 4109 |
*errorcodeptr = ERR60; |
*errorcodeptr = ERR60; |
| 4248 |
*errorcodeptr = ERR58; |
*errorcodeptr = ERR58; |
| 4249 |
goto FAILED; |
goto FAILED; |
| 4250 |
} |
} |
| 4251 |
if (refsign == '-') |
recno = (refsign == '-')? |
| 4252 |
|
cd->bracount - recno + 1 : recno +cd->bracount; |
| 4253 |
|
if (recno <= 0 || recno > cd->final_bracount) |
| 4254 |
{ |
{ |
| 4255 |
recno = cd->bracount - recno + 1; |
*errorcodeptr = ERR15; |
| 4256 |
if (recno <= 0) |
goto FAILED; |
|
{ |
|
|
*errorcodeptr = ERR15; |
|
|
goto FAILED; |
|
|
} |
|
| 4257 |
} |
} |
|
else recno += cd->bracount; |
|
| 4258 |
PUT2(code, 2+LINK_SIZE, recno); |
PUT2(code, 2+LINK_SIZE, recno); |
| 4259 |
break; |
break; |
| 4260 |
} |
} |
| 4326 |
skipbytes = 1; |
skipbytes = 1; |
| 4327 |
} |
} |
| 4328 |
|
|
| 4329 |
/* Check for the "name" actually being a subpattern number. */ |
/* Check for the "name" actually being a subpattern number. We are |
| 4330 |
|
in the second pass here, so final_bracount is set. */ |
| 4331 |
|
|
| 4332 |
else if (recno > 0) |
else if (recno > 0 && recno <= cd->final_bracount) |
| 4333 |
{ |
{ |
| 4334 |
PUT2(code, 2+LINK_SIZE, recno); |
PUT2(code, 2+LINK_SIZE, recno); |
| 4335 |
} |
} |
| 4523 |
|
|
| 4524 |
/* We come here from the Python syntax above that handles both |
/* We come here from the Python syntax above that handles both |
| 4525 |
references (?P=name) and recursion (?P>name), as well as falling |
references (?P=name) and recursion (?P>name), as well as falling |
| 4526 |
through from the Perl recursion syntax (?&name). */ |
through from the Perl recursion syntax (?&name). We also come here from |
| 4527 |
|
the Perl \k<name> or \k'name' back reference syntax and the \k{name} |
| 4528 |
|
.NET syntax. */ |
| 4529 |
|
|
| 4530 |
NAMED_REF_OR_RECURSE: |
NAMED_REF_OR_RECURSE: |
| 4531 |
name = ++ptr; |
name = ++ptr; |
| 4537 |
|
|
| 4538 |
if (lengthptr != NULL) |
if (lengthptr != NULL) |
| 4539 |
{ |
{ |
| 4540 |
|
if (namelen == 0) |
| 4541 |
|
{ |
| 4542 |
|
*errorcodeptr = ERR62; |
| 4543 |
|
goto FAILED; |
| 4544 |
|
} |
| 4545 |
if (*ptr != terminator) |
if (*ptr != terminator) |
| 4546 |
{ |
{ |
| 4547 |
*errorcodeptr = ERR42; |
*errorcodeptr = ERR42; |
| 4555 |
recno = 0; |
recno = 0; |
| 4556 |
} |
} |
| 4557 |
|
|
| 4558 |
/* In the real compile, seek the name in the table */ |
/* In the real compile, seek the name in the table. We check the name |
| 4559 |
|
first, and then check that we have reached the end of the name in the |
| 4560 |
|
table. That way, if the name that is longer than any in the table, |
| 4561 |
|
the comparison will fail without reading beyond the table entry. */ |
| 4562 |
|
|
| 4563 |
else |
else |
| 4564 |
{ |
{ |
| 4565 |
slot = cd->name_table; |
slot = cd->name_table; |
| 4566 |
for (i = 0; i < cd->names_found; i++) |
for (i = 0; i < cd->names_found; i++) |
| 4567 |
{ |
{ |
| 4568 |
if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break; |
if (strncmp((char *)name, (char *)slot+2, namelen) == 0 && |
| 4569 |
|
slot[2+namelen] == 0) |
| 4570 |
|
break; |
| 4571 |
slot += cd->name_entry_size; |
slot += cd->name_entry_size; |
| 4572 |
} |
} |
| 4573 |
|
|
| 4604 |
{ |
{ |
| 4605 |
const uschar *called; |
const uschar *called; |
| 4606 |
|
|
| 4607 |
if ((refsign = *ptr) == '+') ptr++; |
if ((refsign = *ptr) == '+') |
| 4608 |
|
{ |
| 4609 |
|
ptr++; |
| 4610 |
|
if ((digitab[*ptr] & ctype_digit) == 0) |
| 4611 |
|
{ |
| 4612 |
|
*errorcodeptr = ERR63; |
| 4613 |
|
goto FAILED; |
| 4614 |
|
} |
| 4615 |
|
} |
| 4616 |
else if (refsign == '-') |
else if (refsign == '-') |
| 4617 |
{ |
{ |
| 4618 |
if ((digitab[ptr[1]] & ctype_digit) == 0) |
if ((digitab[ptr[1]] & ctype_digit) == 0) |
| 5939 |
no longer needed, so hopefully this workspace will never overflow, though there |
no longer needed, so hopefully this workspace will never overflow, though there |
| 5940 |
is a test for its doing so. */ |
is a test for its doing so. */ |
| 5941 |
|
|
| 5942 |
cd->bracount = 0; |
cd->bracount = cd->final_bracount = 0; |
| 5943 |
cd->names_found = 0; |
cd->names_found = 0; |
| 5944 |
cd->name_entry_size = 0; |
cd->name_entry_size = 0; |
| 5945 |
cd->name_table = NULL; |
cd->name_table = NULL; |
| 6016 |
field; this time it's used for remembering forward references to subpatterns. |
field; this time it's used for remembering forward references to subpatterns. |
| 6017 |
*/ |
*/ |
| 6018 |
|
|
| 6019 |
|
cd->final_bracount = cd->bracount; /* Save for checking forward references */ |
| 6020 |
cd->bracount = 0; |
cd->bracount = 0; |
| 6021 |
cd->names_found = 0; |
cd->names_found = 0; |
| 6022 |
cd->name_table = (uschar *)re + re->name_table_offset; |
cd->name_table = (uschar *)re + re->name_table_offset; |
| 6081 |
PCRE_EARLY_ERROR_RETURN: |
PCRE_EARLY_ERROR_RETURN: |
| 6082 |
*erroroffset = ptr - (const uschar *)pattern; |
*erroroffset = ptr - (const uschar *)pattern; |
| 6083 |
PCRE_EARLY_ERROR_RETURN2: |
PCRE_EARLY_ERROR_RETURN2: |
| 6084 |
*errorptr = error_texts[errorcode]; |
*errorptr = find_error_text(errorcode); |
| 6085 |
if (errorcodeptr != NULL) *errorcodeptr = errorcode; |
if (errorcodeptr != NULL) *errorcodeptr = errorcode; |
| 6086 |
return NULL; |
return NULL; |
| 6087 |
} |
} |
| 6166 |
if (code - codestart > length) |
if (code - codestart > length) |
| 6167 |
{ |
{ |
| 6168 |
(pcre_free)(re); |
(pcre_free)(re); |
| 6169 |
*errorptr = error_texts[ERR23]; |
*errorptr = find_error_text(ERR23); |
| 6170 |
*erroroffset = ptr - (uschar *)pattern; |
*erroroffset = ptr - (uschar *)pattern; |
| 6171 |
if (errorcodeptr != NULL) *errorcodeptr = ERR23; |
if (errorcodeptr != NULL) *errorcodeptr = ERR23; |
| 6172 |
return NULL; |
return NULL; |