| 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) 1997-2008 University of Cambridge |
Copyright (c) 1997-2009 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 |
| 97 |
on. Zero means further processing is needed (for things like \x), or the escape |
on. Zero means further processing is needed (for things like \x), or the escape |
| 98 |
is invalid. */ |
is invalid. */ |
| 99 |
|
|
| 100 |
#ifndef EBCDIC /* This is the "normal" table for ASCII systems */ |
#ifndef EBCDIC |
| 101 |
|
|
| 102 |
|
/* This is the "normal" table for ASCII systems or for EBCDIC systems running |
| 103 |
|
in UTF-8 mode. */ |
| 104 |
|
|
| 105 |
static const short int escapes[] = { |
static const short int escapes[] = { |
| 106 |
0, 0, 0, 0, 0, 0, 0, 0, /* 0 - 7 */ |
0, 0, |
| 107 |
0, 0, ':', ';', '<', '=', '>', '?', /* 8 - ? */ |
0, 0, |
| 108 |
'@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E, 0, -ESC_G, /* @ - G */ |
0, 0, |
| 109 |
-ESC_H, 0, 0, -ESC_K, 0, 0, 0, 0, /* H - O */ |
0, 0, |
| 110 |
-ESC_P, -ESC_Q, -ESC_R, -ESC_S, 0, 0, -ESC_V, -ESC_W, /* P - W */ |
0, 0, |
| 111 |
-ESC_X, 0, -ESC_Z, '[', '\\', ']', '^', '_', /* X - _ */ |
CHAR_COLON, CHAR_SEMICOLON, |
| 112 |
'`', 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, /* ` - g */ |
CHAR_LESS_THAN_SIGN, CHAR_EQUALS_SIGN, |
| 113 |
-ESC_h, 0, 0, -ESC_k, 0, 0, ESC_n, 0, /* h - o */ |
CHAR_GREATER_THAN_SIGN, CHAR_QUESTION_MARK, |
| 114 |
-ESC_p, 0, ESC_r, -ESC_s, ESC_tee, 0, -ESC_v, -ESC_w, /* p - w */ |
CHAR_COMMERCIAL_AT, -ESC_A, |
| 115 |
0, 0, -ESC_z /* x - z */ |
-ESC_B, -ESC_C, |
| 116 |
|
-ESC_D, -ESC_E, |
| 117 |
|
0, -ESC_G, |
| 118 |
|
-ESC_H, 0, |
| 119 |
|
0, -ESC_K, |
| 120 |
|
0, 0, |
| 121 |
|
0, 0, |
| 122 |
|
-ESC_P, -ESC_Q, |
| 123 |
|
-ESC_R, -ESC_S, |
| 124 |
|
0, 0, |
| 125 |
|
-ESC_V, -ESC_W, |
| 126 |
|
-ESC_X, 0, |
| 127 |
|
-ESC_Z, CHAR_LEFT_SQUARE_BRACKET, |
| 128 |
|
CHAR_BACKSLASH, CHAR_RIGHT_SQUARE_BRACKET, |
| 129 |
|
CHAR_CIRCUMFLEX_ACCENT, CHAR_UNDERSCORE, |
| 130 |
|
CHAR_GRAVE_ACCENT, 7, |
| 131 |
|
-ESC_b, 0, |
| 132 |
|
-ESC_d, ESC_e, |
| 133 |
|
ESC_f, 0, |
| 134 |
|
-ESC_h, 0, |
| 135 |
|
0, -ESC_k, |
| 136 |
|
0, 0, |
| 137 |
|
ESC_n, 0, |
| 138 |
|
-ESC_p, 0, |
| 139 |
|
ESC_r, -ESC_s, |
| 140 |
|
ESC_tee, 0, |
| 141 |
|
-ESC_v, -ESC_w, |
| 142 |
|
0, 0, |
| 143 |
|
-ESC_z |
| 144 |
}; |
}; |
| 145 |
|
|
| 146 |
#else /* This is the "abnormal" table for EBCDIC systems */ |
#else |
| 147 |
|
|
| 148 |
|
/* This is the "abnormal" table for EBCDIC systems without UTF-8 support. */ |
| 149 |
|
|
| 150 |
static const short int escapes[] = { |
static const short int escapes[] = { |
| 151 |
/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', |
/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', |
| 152 |
/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, |
/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, |
| 177 |
|
|
| 178 |
/* Table of special "verbs" like (*PRUNE). This is a short table, so it is |
/* Table of special "verbs" like (*PRUNE). This is a short table, so it is |
| 179 |
searched linearly. Put all the names into a single string, in order to reduce |
searched linearly. Put all the names into a single string, in order to reduce |
| 180 |
the number of relocations when a shared library is dynamically linked. */ |
the number of relocations when a shared library is dynamically linked. The |
| 181 |
|
string is built from string macros so that it works in UTF-8 mode on EBCDIC |
| 182 |
|
platforms. */ |
| 183 |
|
|
| 184 |
typedef struct verbitem { |
typedef struct verbitem { |
| 185 |
int len; |
int len; |
| 187 |
} verbitem; |
} verbitem; |
| 188 |
|
|
| 189 |
static const char verbnames[] = |
static const char verbnames[] = |
| 190 |
"ACCEPT\0" |
STRING_ACCEPT0 |
| 191 |
"COMMIT\0" |
STRING_COMMIT0 |
| 192 |
"F\0" |
STRING_F0 |
| 193 |
"FAIL\0" |
STRING_FAIL0 |
| 194 |
"PRUNE\0" |
STRING_PRUNE0 |
| 195 |
"SKIP\0" |
STRING_SKIP0 |
| 196 |
"THEN"; |
STRING_THEN; |
| 197 |
|
|
| 198 |
static const verbitem verbs[] = { |
static const verbitem verbs[] = { |
| 199 |
{ 6, OP_ACCEPT }, |
{ 6, OP_ACCEPT }, |
| 215 |
for handling case independence. */ |
for handling case independence. */ |
| 216 |
|
|
| 217 |
static const char posix_names[] = |
static const char posix_names[] = |
| 218 |
"alpha\0" "lower\0" "upper\0" "alnum\0" "ascii\0" "blank\0" |
STRING_alpha0 STRING_lower0 STRING_upper0 STRING_alnum0 |
| 219 |
"cntrl\0" "digit\0" "graph\0" "print\0" "punct\0" "space\0" |
STRING_ascii0 STRING_blank0 STRING_cntrl0 STRING_digit0 |
| 220 |
"word\0" "xdigit"; |
STRING_graph0 STRING_print0 STRING_punct0 STRING_space0 |
| 221 |
|
STRING_word0 STRING_xdigit; |
| 222 |
|
|
| 223 |
static const uschar posix_name_lengths[] = { |
static const uschar posix_name_lengths[] = { |
| 224 |
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 }; |
| 360 |
|
|
| 361 |
Then we can use ctype_digit and ctype_xdigit in the code. */ |
Then we can use ctype_digit and ctype_xdigit in the code. */ |
| 362 |
|
|
| 363 |
#ifndef EBCDIC /* This is the "normal" case, for ASCII systems */ |
#ifndef EBCDIC |
| 364 |
|
|
| 365 |
|
/* This is the "normal" case, for ASCII systems, and EBCDIC systems running in |
| 366 |
|
UTF-8 mode. */ |
| 367 |
|
|
| 368 |
static const unsigned char digitab[] = |
static const unsigned char digitab[] = |
| 369 |
{ |
{ |
| 370 |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ |
| 400 |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ |
| 401 |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ |
| 402 |
|
|
| 403 |
#else /* This is the "abnormal" case, for EBCDIC systems */ |
#else |
| 404 |
|
|
| 405 |
|
/* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */ |
| 406 |
|
|
| 407 |
static const unsigned char digitab[] = |
static const unsigned char digitab[] = |
| 408 |
{ |
{ |
| 409 |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ |
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ |
| 500 |
find_error_text(int n) |
find_error_text(int n) |
| 501 |
{ |
{ |
| 502 |
const char *s = error_texts; |
const char *s = error_texts; |
| 503 |
for (; n > 0; n--) while (*s++ != 0); |
for (; n > 0; n--) while (*s++ != 0) {}; |
| 504 |
return s; |
return s; |
| 505 |
} |
} |
| 506 |
|
|
| 548 |
in a table. A non-zero result is something that can be returned immediately. |
in a table. A non-zero result is something that can be returned immediately. |
| 549 |
Otherwise further processing may be required. */ |
Otherwise further processing may be required. */ |
| 550 |
|
|
| 551 |
#ifndef EBCDIC /* ASCII coding */ |
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
| 552 |
else if (c < '0' || c > 'z') {} /* Not alphanumeric */ |
else if (c < CHAR_0 || c > CHAR_z) {} /* Not alphanumeric */ |
| 553 |
else if ((i = escapes[c - '0']) != 0) c = i; |
else if ((i = escapes[c - CHAR_0]) != 0) c = i; |
| 554 |
|
|
| 555 |
#else /* EBCDIC coding */ |
#else /* EBCDIC coding */ |
| 556 |
else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphanumeric */ |
else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphanumeric */ |
| 569 |
/* A number of Perl escapes are not handled by PCRE. We give an explicit |
/* A number of Perl escapes are not handled by PCRE. We give an explicit |
| 570 |
error. */ |
error. */ |
| 571 |
|
|
| 572 |
case 'l': |
case CHAR_l: |
| 573 |
case 'L': |
case CHAR_L: |
| 574 |
case 'N': |
case CHAR_N: |
| 575 |
case 'u': |
case CHAR_u: |
| 576 |
case 'U': |
case CHAR_U: |
| 577 |
*errorcodeptr = ERR37; |
*errorcodeptr = ERR37; |
| 578 |
break; |
break; |
| 579 |
|
|
| 593 |
(possibly recursive) subroutine calls, _not_ backreferences. Just return |
(possibly recursive) subroutine calls, _not_ backreferences. Just return |
| 594 |
the -ESC_g code (cf \k). */ |
the -ESC_g code (cf \k). */ |
| 595 |
|
|
| 596 |
case 'g': |
case CHAR_g: |
| 597 |
if (ptr[1] == '<' || ptr[1] == '\'') |
if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE) |
| 598 |
{ |
{ |
| 599 |
c = -ESC_g; |
c = -ESC_g; |
| 600 |
break; |
break; |
| 602 |
|
|
| 603 |
/* Handle the Perl-compatible cases */ |
/* Handle the Perl-compatible cases */ |
| 604 |
|
|
| 605 |
if (ptr[1] == '{') |
if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) |
| 606 |
{ |
{ |
| 607 |
const uschar *p; |
const uschar *p; |
| 608 |
for (p = ptr+2; *p != 0 && *p != '}'; p++) |
for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++) |
| 609 |
if (*p != '-' && (digitab[*p] & ctype_digit) == 0) break; |
if (*p != CHAR_MINUS && (digitab[*p] & ctype_digit) == 0) break; |
| 610 |
if (*p != 0 && *p != '}') |
if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET) |
| 611 |
{ |
{ |
| 612 |
c = -ESC_k; |
c = -ESC_k; |
| 613 |
break; |
break; |
| 617 |
} |
} |
| 618 |
else braced = FALSE; |
else braced = FALSE; |
| 619 |
|
|
| 620 |
if (ptr[1] == '-') |
if (ptr[1] == CHAR_MINUS) |
| 621 |
{ |
{ |
| 622 |
negated = TRUE; |
negated = TRUE; |
| 623 |
ptr++; |
ptr++; |
| 626 |
|
|
| 627 |
c = 0; |
c = 0; |
| 628 |
while ((digitab[ptr[1]] & ctype_digit) != 0) |
while ((digitab[ptr[1]] & ctype_digit) != 0) |
| 629 |
c = c * 10 + *(++ptr) - '0'; |
c = c * 10 + *(++ptr) - CHAR_0; |
| 630 |
|
|
| 631 |
if (c < 0) /* Integer overflow */ |
if (c < 0) /* Integer overflow */ |
| 632 |
{ |
{ |
| 634 |
break; |
break; |
| 635 |
} |
} |
| 636 |
|
|
| 637 |
if (braced && *(++ptr) != '}') |
if (braced && *(++ptr) != CHAR_RIGHT_CURLY_BRACKET) |
| 638 |
{ |
{ |
| 639 |
*errorcodeptr = ERR57; |
*errorcodeptr = ERR57; |
| 640 |
break; |
break; |
| 671 |
value is greater than 377, the least significant 8 bits are taken. Inside a |
value is greater than 377, the least significant 8 bits are taken. Inside a |
| 672 |
character class, \ followed by a digit is always an octal number. */ |
character class, \ followed by a digit is always an octal number. */ |
| 673 |
|
|
| 674 |
case '1': case '2': case '3': case '4': case '5': |
case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: case CHAR_5: |
| 675 |
case '6': case '7': case '8': case '9': |
case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: |
| 676 |
|
|
| 677 |
if (!isclass) |
if (!isclass) |
| 678 |
{ |
{ |
| 679 |
oldptr = ptr; |
oldptr = ptr; |
| 680 |
c -= '0'; |
c -= CHAR_0; |
| 681 |
while ((digitab[ptr[1]] & ctype_digit) != 0) |
while ((digitab[ptr[1]] & ctype_digit) != 0) |
| 682 |
c = c * 10 + *(++ptr) - '0'; |
c = c * 10 + *(++ptr) - CHAR_0; |
| 683 |
if (c < 0) /* Integer overflow */ |
if (c < 0) /* Integer overflow */ |
| 684 |
{ |
{ |
| 685 |
*errorcodeptr = ERR61; |
*errorcodeptr = ERR61; |
| 697 |
generates a binary zero byte and treats the digit as a following literal. |
generates a binary zero byte and treats the digit as a following literal. |
| 698 |
Thus we have to pull back the pointer by one. */ |
Thus we have to pull back the pointer by one. */ |
| 699 |
|
|
| 700 |
if ((c = *ptr) >= '8') |
if ((c = *ptr) >= CHAR_8) |
| 701 |
{ |
{ |
| 702 |
ptr--; |
ptr--; |
| 703 |
c = 0; |
c = 0; |
| 710 |
to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more |
to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more |
| 711 |
than 3 octal digits. */ |
than 3 octal digits. */ |
| 712 |
|
|
| 713 |
case '0': |
case CHAR_0: |
| 714 |
c -= '0'; |
c -= CHAR_0; |
| 715 |
while(i++ < 2 && ptr[1] >= '0' && ptr[1] <= '7') |
while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7) |
| 716 |
c = c * 8 + *(++ptr) - '0'; |
c = c * 8 + *(++ptr) - CHAR_0; |
| 717 |
if (!utf8 && c > 255) *errorcodeptr = ERR51; |
if (!utf8 && c > 255) *errorcodeptr = ERR51; |
| 718 |
break; |
break; |
| 719 |
|
|
| 721 |
than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is |
than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is |
| 722 |
treated as a data character. */ |
treated as a data character. */ |
| 723 |
|
|
| 724 |
case 'x': |
case CHAR_x: |
| 725 |
if (ptr[1] == '{') |
if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) |
| 726 |
{ |
{ |
| 727 |
const uschar *pt = ptr + 2; |
const uschar *pt = ptr + 2; |
| 728 |
int count = 0; |
int count = 0; |
| 731 |
while ((digitab[*pt] & ctype_xdigit) != 0) |
while ((digitab[*pt] & ctype_xdigit) != 0) |
| 732 |
{ |
{ |
| 733 |
register int cc = *pt++; |
register int cc = *pt++; |
| 734 |
if (c == 0 && cc == '0') continue; /* Leading zeroes */ |
if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */ |
| 735 |
count++; |
count++; |
| 736 |
|
|
| 737 |
#ifndef EBCDIC /* ASCII coding */ |
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
| 738 |
if (cc >= 'a') cc -= 32; /* Convert to upper case */ |
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ |
| 739 |
c = (c << 4) + cc - ((cc < 'A')? '0' : ('A' - 10)); |
c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); |
| 740 |
#else /* EBCDIC coding */ |
#else /* EBCDIC coding */ |
| 741 |
if (cc >= 'a' && cc <= 'z') cc += 64; /* Convert to upper case */ |
if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ |
| 742 |
c = (c << 4) + cc - ((cc >= '0')? '0' : ('A' - 10)); |
c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); |
| 743 |
#endif |
#endif |
| 744 |
} |
} |
| 745 |
|
|
| 746 |
if (*pt == '}') |
if (*pt == CHAR_RIGHT_CURLY_BRACKET) |
| 747 |
{ |
{ |
| 748 |
if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34; |
if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34; |
| 749 |
ptr = pt; |
ptr = pt; |
| 759 |
c = 0; |
c = 0; |
| 760 |
while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0) |
while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0) |
| 761 |
{ |
{ |
| 762 |
int cc; /* Some compilers don't like ++ */ |
int cc; /* Some compilers don't like */ |
| 763 |
cc = *(++ptr); /* in initializers */ |
cc = *(++ptr); /* ++ in initializers */ |
| 764 |
#ifndef EBCDIC /* ASCII coding */ |
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
| 765 |
if (cc >= 'a') cc -= 32; /* Convert to upper case */ |
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ |
| 766 |
c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10)); |
c = c * 16 + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); |
| 767 |
#else /* EBCDIC coding */ |
#else /* EBCDIC coding */ |
| 768 |
if (cc <= 'z') cc += 64; /* Convert to upper case */ |
if (cc <= CHAR_z) cc += 64; /* Convert to upper case */ |
| 769 |
c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10)); |
c = c * 16 + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); |
| 770 |
#endif |
#endif |
| 771 |
} |
} |
| 772 |
break; |
break; |
| 775 |
This coding is ASCII-specific, but then the whole concept of \cx is |
This coding is ASCII-specific, but then the whole concept of \cx is |
| 776 |
ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ |
ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ |
| 777 |
|
|
| 778 |
case 'c': |
case CHAR_c: |
| 779 |
c = *(++ptr); |
c = *(++ptr); |
| 780 |
if (c == 0) |
if (c == 0) |
| 781 |
{ |
{ |
| 783 |
break; |
break; |
| 784 |
} |
} |
| 785 |
|
|
| 786 |
#ifndef EBCDIC /* ASCII coding */ |
#ifndef EBCDIC /* ASCII/UTF-8 coding */ |
| 787 |
if (c >= 'a' && c <= 'z') c -= 32; |
if (c >= CHAR_a && c <= CHAR_z) c -= 32; |
| 788 |
c ^= 0x40; |
c ^= 0x40; |
| 789 |
#else /* EBCDIC coding */ |
#else /* EBCDIC coding */ |
| 790 |
if (c >= 'a' && c <= 'z') c += 64; |
if (c >= CHAR_a && c <= CHAR_z) c += 64; |
| 791 |
c ^= 0xC0; |
c ^= 0xC0; |
| 792 |
#endif |
#endif |
| 793 |
break; |
break; |
| 849 |
/* \P or \p can be followed by a name in {}, optionally preceded by ^ for |
/* \P or \p can be followed by a name in {}, optionally preceded by ^ for |
| 850 |
negation. */ |
negation. */ |
| 851 |
|
|
| 852 |
if (c == '{') |
if (c == CHAR_LEFT_CURLY_BRACKET) |
| 853 |
{ |
{ |
| 854 |
if (ptr[1] == '^') |
if (ptr[1] == CHAR_CIRCUMFLEX_ACCENT) |
| 855 |
{ |
{ |
| 856 |
*negptr = TRUE; |
*negptr = TRUE; |
| 857 |
ptr++; |
ptr++; |
| 860 |
{ |
{ |
| 861 |
c = *(++ptr); |
c = *(++ptr); |
| 862 |
if (c == 0) goto ERROR_RETURN; |
if (c == 0) goto ERROR_RETURN; |
| 863 |
if (c == '}') break; |
if (c == CHAR_RIGHT_CURLY_BRACKET) break; |
| 864 |
name[i] = c; |
name[i] = c; |
| 865 |
} |
} |
| 866 |
if (c !='}') goto ERROR_RETURN; |
if (c != CHAR_RIGHT_CURLY_BRACKET) goto ERROR_RETURN; |
| 867 |
name[i] = 0; |
name[i] = 0; |
| 868 |
} |
} |
| 869 |
|
|
| 928 |
{ |
{ |
| 929 |
if ((digitab[*p++] & ctype_digit) == 0) return FALSE; |
if ((digitab[*p++] & ctype_digit) == 0) return FALSE; |
| 930 |
while ((digitab[*p] & ctype_digit) != 0) p++; |
while ((digitab[*p] & ctype_digit) != 0) p++; |
| 931 |
if (*p == '}') return TRUE; |
if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; |
| 932 |
|
|
| 933 |
if (*p++ != ',') return FALSE; |
if (*p++ != CHAR_COMMA) return FALSE; |
| 934 |
if (*p == '}') return TRUE; |
if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; |
| 935 |
|
|
| 936 |
if ((digitab[*p++] & ctype_digit) == 0) return FALSE; |
if ((digitab[*p++] & ctype_digit) == 0) return FALSE; |
| 937 |
while ((digitab[*p] & ctype_digit) != 0) p++; |
while ((digitab[*p] & ctype_digit) != 0) p++; |
| 938 |
|
|
| 939 |
return (*p == '}'); |
return (*p == CHAR_RIGHT_CURLY_BRACKET); |
| 940 |
} |
} |
| 941 |
|
|
| 942 |
|
|
| 969 |
/* Read the minimum value and do a paranoid check: a negative value indicates |
/* Read the minimum value and do a paranoid check: a negative value indicates |
| 970 |
an integer overflow. */ |
an integer overflow. */ |
| 971 |
|
|
| 972 |
while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - '0'; |
while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - CHAR_0; |
| 973 |
if (min < 0 || min > 65535) |
if (min < 0 || min > 65535) |
| 974 |
{ |
{ |
| 975 |
*errorcodeptr = ERR5; |
*errorcodeptr = ERR5; |
| 979 |
/* Read the maximum value if there is one, and again do a paranoid on its size. |
/* Read the maximum value if there is one, and again do a paranoid on its size. |
| 980 |
Also, max must not be less than min. */ |
Also, max must not be less than min. */ |
| 981 |
|
|
| 982 |
if (*p == '}') max = min; else |
if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else |
| 983 |
{ |
{ |
| 984 |
if (*(++p) != '}') |
if (*(++p) != CHAR_RIGHT_CURLY_BRACKET) |
| 985 |
{ |
{ |
| 986 |
max = 0; |
max = 0; |
| 987 |
while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - '0'; |
while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - CHAR_0; |
| 988 |
if (max < 0 || max > 65535) |
if (max < 0 || max > 65535) |
| 989 |
{ |
{ |
| 990 |
*errorcodeptr = ERR5; |
*errorcodeptr = ERR5; |
| 1042 |
|
|
| 1043 |
/* Skip over backslashed characters and also entire \Q...\E */ |
/* Skip over backslashed characters and also entire \Q...\E */ |
| 1044 |
|
|
| 1045 |
if (*ptr == '\\') |
if (*ptr == CHAR_BACKSLASH) |
| 1046 |
{ |
{ |
| 1047 |
if (*(++ptr) == 0) return -1; |
if (*(++ptr) == 0) return -1; |
| 1048 |
if (*ptr == 'Q') for (;;) |
if (*ptr == CHAR_Q) for (;;) |
| 1049 |
{ |
{ |
| 1050 |
while (*(++ptr) != 0 && *ptr != '\\'); |
while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {}; |
| 1051 |
if (*ptr == 0) return -1; |
if (*ptr == 0) return -1; |
| 1052 |
if (*(++ptr) == 'E') break; |
if (*(++ptr) == CHAR_E) break; |
| 1053 |
} |
} |
| 1054 |
continue; |
continue; |
| 1055 |
} |
} |
| 1057 |
/* Skip over character classes; this logic must be similar to the way they |
/* Skip over character classes; this logic must be similar to the way they |
| 1058 |
are handled for real. If the first character is '^', skip it. Also, if the |
are handled for real. If the first character is '^', skip it. Also, if the |
| 1059 |
first few characters (either before or after ^) are \Q\E or \E we skip them |
first few characters (either before or after ^) are \Q\E or \E we skip them |
| 1060 |
too. This makes for compatibility with Perl. */ |
too. This makes for compatibility with Perl. Note the use of STR macros to |
| 1061 |
|
encode "Q\\E" so that it works in UTF-8 on EBCDIC platforms. */ |
| 1062 |
|
|
| 1063 |
if (*ptr == '[') |
if (*ptr == CHAR_LEFT_SQUARE_BRACKET) |
| 1064 |
{ |
{ |
| 1065 |
BOOL negate_class = FALSE; |
BOOL negate_class = FALSE; |
| 1066 |
for (;;) |
for (;;) |
| 1067 |
{ |
{ |
| 1068 |
int c = *(++ptr); |
int c = *(++ptr); |
| 1069 |
if (c == '\\') |
if (c == CHAR_BACKSLASH) |
| 1070 |
{ |
{ |
| 1071 |
if (ptr[1] == 'E') ptr++; |
if (ptr[1] == CHAR_E) |
| 1072 |
else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3; |
ptr++; |
| 1073 |
else break; |
else if (strncmp((const char *)ptr+1, |
| 1074 |
|
STR_Q STR_BACKSLASH STR_E, 3) == 0) |
| 1075 |
|
ptr += 3; |
| 1076 |
|
else |
| 1077 |
|
break; |
| 1078 |
} |
} |
| 1079 |
else if (!negate_class && c == '^') |
else if (!negate_class && c == CHAR_CIRCUMFLEX_ACCENT) |
| 1080 |
negate_class = TRUE; |
negate_class = TRUE; |
| 1081 |
else break; |
else break; |
| 1082 |
} |
} |
| 1084 |
/* If the next character is ']', it is a data character that must be |
/* If the next character is ']', it is a data character that must be |
| 1085 |
skipped, except in JavaScript compatibility mode. */ |
skipped, except in JavaScript compatibility mode. */ |
| 1086 |
|
|
| 1087 |
if (ptr[1] == ']' && (cd->external_options & PCRE_JAVASCRIPT_COMPAT) == 0) |
if (ptr[1] == CHAR_RIGHT_SQUARE_BRACKET && |
| 1088 |
|
(cd->external_options & PCRE_JAVASCRIPT_COMPAT) == 0) |
| 1089 |
ptr++; |
ptr++; |
| 1090 |
|
|
| 1091 |
while (*(++ptr) != ']') |
while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET) |
| 1092 |
{ |
{ |
| 1093 |
if (*ptr == 0) return -1; |
if (*ptr == 0) return -1; |
| 1094 |
if (*ptr == '\\') |
if (*ptr == CHAR_BACKSLASH) |
| 1095 |
{ |
{ |
| 1096 |
if (*(++ptr) == 0) return -1; |
if (*(++ptr) == 0) return -1; |
| 1097 |
if (*ptr == 'Q') for (;;) |
if (*ptr == CHAR_Q) for (;;) |
| 1098 |
{ |
{ |
| 1099 |
while (*(++ptr) != 0 && *ptr != '\\'); |
while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {}; |
| 1100 |
if (*ptr == 0) return -1; |
if (*ptr == 0) return -1; |
| 1101 |
if (*(++ptr) == 'E') break; |
if (*(++ptr) == CHAR_E) break; |
| 1102 |
} |
} |
| 1103 |
continue; |
continue; |
| 1104 |
} |
} |
| 1108 |
|
|
| 1109 |
/* Skip comments in /x mode */ |
/* Skip comments in /x mode */ |
| 1110 |
|
|
| 1111 |
if (xmode && *ptr == '#') |
if (xmode && *ptr == CHAR_NUMBER_SIGN) |
| 1112 |
{ |
{ |
| 1113 |
while (*(++ptr) != 0 && *ptr != '\n'); |
while (*(++ptr) != 0 && *ptr != CHAR_NL) {}; |
| 1114 |
if (*ptr == 0) return -1; |
if (*ptr == 0) return -1; |
| 1115 |
continue; |
continue; |
| 1116 |
} |
} |
| 1117 |
|
|
| 1118 |
/* An opening parens must now be a real metacharacter */ |
/* An opening parens must now be a real metacharacter */ |
| 1119 |
|
|
| 1120 |
if (*ptr != '(') continue; |
if (*ptr != CHAR_LEFT_PARENTHESIS) continue; |
| 1121 |
if (ptr[1] != '?' && ptr[1] != '*') |
if (ptr[1] != CHAR_QUESTION_MARK && ptr[1] != CHAR_ASTERISK) |
| 1122 |
{ |
{ |
| 1123 |
count++; |
count++; |
| 1124 |
if (name == NULL && count == lorn) return count; |
if (name == NULL && count == lorn) return count; |
| 1126 |
} |
} |
| 1127 |
|
|
| 1128 |
ptr += 2; |
ptr += 2; |
| 1129 |
if (*ptr == 'P') ptr++; /* Allow optional P */ |
if (*ptr == CHAR_P) ptr++; /* Allow optional P */ |
| 1130 |
|
|
| 1131 |
/* We have to disambiguate (?<! and (?<= from (?<name> */ |
/* We have to disambiguate (?<! and (?<= from (?<name> */ |
| 1132 |
|
|
| 1133 |
if ((*ptr != '<' || ptr[1] == '!' || ptr[1] == '=') && |
if ((*ptr != CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_EXCLAMATION_MARK || |
| 1134 |
*ptr != '\'') |
ptr[1] == CHAR_EQUALS_SIGN) && *ptr != CHAR_APOSTROPHE) |
| 1135 |
continue; |
continue; |
| 1136 |
|
|
| 1137 |
count++; |
count++; |
| 1138 |
|
|
| 1139 |
if (name == NULL && count == lorn) return count; |
if (name == NULL && count == lorn) return count; |
| 1140 |
term = *ptr++; |
term = *ptr++; |
| 1141 |
if (term == '<') term = '>'; |
if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN; |
| 1142 |
thisname = ptr; |
thisname = ptr; |
| 1143 |
while (*ptr != term) ptr++; |
while (*ptr != term) ptr++; |
| 1144 |
if (name != NULL && lorn == ptr - thisname && |
if (name != NULL && lorn == ptr - thisname && |
| 1501 |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
| 1502 |
break; |
break; |
| 1503 |
} |
} |
| 1504 |
|
#else |
| 1505 |
|
(void)(utf8); /* Keep compiler happy by referencing function argument */ |
| 1506 |
#endif |
#endif |
| 1507 |
} |
} |
| 1508 |
} |
} |
| 1596 |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; |
| 1597 |
break; |
break; |
| 1598 |
} |
} |
| 1599 |
|
#else |
| 1600 |
|
(void)(utf8); /* Keep compiler happy by referencing function argument */ |
| 1601 |
#endif |
#endif |
| 1602 |
} |
} |
| 1603 |
} |
} |
| 1664 |
BOOL empty_branch; |
BOOL empty_branch; |
| 1665 |
if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ |
if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ |
| 1666 |
|
|
| 1667 |
/* Scan a closed bracket */ |
/* If a conditional group has only one branch, there is a second, implied, |
| 1668 |
|
empty branch, so just skip over the conditional, because it could be empty. |
| 1669 |
|
Otherwise, scan the individual branches of the group. */ |
| 1670 |
|
|
| 1671 |
empty_branch = FALSE; |
if (c == OP_COND && code[GET(code, 1)] != OP_ALT) |
|
do |
|
|
{ |
|
|
if (!empty_branch && could_be_empty_branch(code, endcode, utf8)) |
|
|
empty_branch = TRUE; |
|
| 1672 |
code += GET(code, 1); |
code += GET(code, 1); |
| 1673 |
|
else |
| 1674 |
|
{ |
| 1675 |
|
empty_branch = FALSE; |
| 1676 |
|
do |
| 1677 |
|
{ |
| 1678 |
|
if (!empty_branch && could_be_empty_branch(code, endcode, utf8)) |
| 1679 |
|
empty_branch = TRUE; |
| 1680 |
|
code += GET(code, 1); |
| 1681 |
|
} |
| 1682 |
|
while (*code == OP_ALT); |
| 1683 |
|
if (!empty_branch) return FALSE; /* All branches are non-empty */ |
| 1684 |
} |
} |
| 1685 |
while (*code == OP_ALT); |
|
|
if (!empty_branch) return FALSE; /* All branches are non-empty */ |
|
| 1686 |
c = *code; |
c = *code; |
| 1687 |
continue; |
continue; |
| 1688 |
} |
} |
| 1884 |
terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ |
terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ |
| 1885 |
for (++ptr; *ptr != 0; ptr++) |
for (++ptr; *ptr != 0; ptr++) |
| 1886 |
{ |
{ |
| 1887 |
if (*ptr == '\\' && ptr[1] == ']') ptr++; else |
if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) ptr++; else |
| 1888 |
{ |
{ |
| 1889 |
if (*ptr == ']') return FALSE; |
if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE; |
| 1890 |
if (*ptr == terminator && ptr[1] == ']') |
if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) |
| 1891 |
{ |
{ |
| 1892 |
*endptr = ptr; |
*endptr = ptr; |
| 1893 |
return TRUE; |
return TRUE; |
| 2078 |
unsigned int c, othercase, next; |
unsigned int c, othercase, next; |
| 2079 |
|
|
| 2080 |
for (c = *cptr; c <= d; c++) |
for (c = *cptr; c <= d; c++) |
| 2081 |
{ if ((othercase = _pcre_ucp_othercase(c)) != NOTACHAR) break; } |
{ if ((othercase = UCD_OTHERCASE(c)) != c) break; } |
| 2082 |
|
|
| 2083 |
if (c > d) return FALSE; |
if (c > d) return FALSE; |
| 2084 |
|
|
| 2087 |
|
|
| 2088 |
for (++c; c <= d; c++) |
for (++c; c <= d; c++) |
| 2089 |
{ |
{ |
| 2090 |
if (_pcre_ucp_othercase(c) != next) break; |
if (UCD_OTHERCASE(c) != next) break; |
| 2091 |
next++; |
next++; |
| 2092 |
} |
} |
| 2093 |
|
|
| 2133 |
for (;;) |
for (;;) |
| 2134 |
{ |
{ |
| 2135 |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
| 2136 |
if (*ptr == '#') |
if (*ptr == CHAR_NUMBER_SIGN) |
| 2137 |
{ |
{ |
| 2138 |
while (*(++ptr) != 0) |
while (*(++ptr) != 0) |
| 2139 |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
| 2145 |
/* If the next item is one that we can handle, get its value. A non-negative |
/* If the next item is one that we can handle, get its value. A non-negative |
| 2146 |
value is a character, a negative value is an escape value. */ |
value is a character, a negative value is an escape value. */ |
| 2147 |
|
|
| 2148 |
if (*ptr == '\\') |
if (*ptr == CHAR_BACKSLASH) |
| 2149 |
{ |
{ |
| 2150 |
int temperrorcode = 0; |
int temperrorcode = 0; |
| 2151 |
next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE); |
next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE); |
| 2170 |
for (;;) |
for (;;) |
| 2171 |
{ |
{ |
| 2172 |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; |
| 2173 |
if (*ptr == '#') |
if (*ptr == CHAR_NUMBER_SIGN) |
| 2174 |
{ |
{ |
| 2175 |
while (*(++ptr) != 0) |
while (*(++ptr) != 0) |
| 2176 |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } |
| 2181 |
|
|
| 2182 |
/* If the next thing is itself optional, we have to give up. */ |
/* If the next thing is itself optional, we have to give up. */ |
| 2183 |
|
|
| 2184 |
if (*ptr == '*' || *ptr == '?' || strncmp((char *)ptr, "{0,", 3) == 0) |
if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK || |
| 2185 |
return FALSE; |
strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0) |
| 2186 |
|
return FALSE; |
| 2187 |
|
|
| 2188 |
/* Now compare the next item with the previous opcode. If the previous is a |
/* Now compare the next item with the previous opcode. If the previous is a |
| 2189 |
positive single character match, "item" either contains the character or, if |
positive single character match, "item" either contains the character or, if |
| 2198 |
case OP_CHAR: |
case OP_CHAR: |
| 2199 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
| 2200 |
if (utf8 && item > 127) { GETCHAR(item, utf8_char); } |
if (utf8 && item > 127) { GETCHAR(item, utf8_char); } |
| 2201 |
|
#else |
| 2202 |
|
(void)(utf8_char); /* Keep compiler happy by referencing function argument */ |
| 2203 |
#endif |
#endif |
| 2204 |
return item != next; |
return item != next; |
| 2205 |
|
|
| 2218 |
unsigned int othercase; |
unsigned int othercase; |
| 2219 |
if (next < 128) othercase = cd->fcc[next]; else |
if (next < 128) othercase = cd->fcc[next]; else |
| 2220 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 2221 |
othercase = _pcre_ucp_othercase((unsigned int)next); |
othercase = UCD_OTHERCASE((unsigned int)next); |
| 2222 |
#else |
#else |
| 2223 |
othercase = NOTACHAR; |
othercase = NOTACHAR; |
| 2224 |
#endif |
#endif |
| 2239 |
unsigned int othercase; |
unsigned int othercase; |
| 2240 |
if (next < 128) othercase = cd->fcc[next]; else |
if (next < 128) othercase = cd->fcc[next]; else |
| 2241 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
| 2242 |
othercase = _pcre_ucp_othercase(next); |
othercase = UCD_OTHERCASE(next); |
| 2243 |
#else |
#else |
| 2244 |
othercase = NOTACHAR; |
othercase = NOTACHAR; |
| 2245 |
#endif |
#endif |
| 2621 |
|
|
| 2622 |
if (inescq && c != 0) |
if (inescq && c != 0) |
| 2623 |
{ |
{ |
| 2624 |
if (c == '\\' && ptr[1] == 'E') |
if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) |
| 2625 |
{ |
{ |
| 2626 |
inescq = FALSE; |
inescq = FALSE; |
| 2627 |
ptr++; |
ptr++; |
| 2647 |
/* Fill in length of a previous callout, except when the next thing is |
/* Fill in length of a previous callout, except when the next thing is |
| 2648 |
a quantifier. */ |
a quantifier. */ |
| 2649 |
|
|
| 2650 |
is_quantifier = c == '*' || c == '+' || c == '?' || |
is_quantifier = |
| 2651 |
(c == '{' && is_counted_repeat(ptr+1)); |
c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK || |
| 2652 |
|
(c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1)); |
| 2653 |
|
|
| 2654 |
if (!is_quantifier && previous_callout != NULL && |
if (!is_quantifier && previous_callout != NULL && |
| 2655 |
after_manual_callout-- <= 0) |
after_manual_callout-- <= 0) |
| 2664 |
if ((options & PCRE_EXTENDED) != 0) |
if ((options & PCRE_EXTENDED) != 0) |
| 2665 |
{ |
{ |
| 2666 |
if ((cd->ctypes[c] & ctype_space) != 0) continue; |
if ((cd->ctypes[c] & ctype_space) != 0) continue; |
| 2667 |
if (c == '#') |
if (c == CHAR_NUMBER_SIGN) |
| 2668 |
{ |
{ |
| 2669 |
while (*(++ptr) != 0) |
while (*(++ptr) != 0) |
| 2670 |
{ |
{ |
| 2689 |
{ |
{ |
| 2690 |
/* ===================================================================*/ |
/* ===================================================================*/ |
| 2691 |
case 0: /* The branch terminates at string end */ |
case 0: /* The branch terminates at string end */ |
| 2692 |
case '|': /* or | or ) */ |
case CHAR_VERTICAL_LINE: /* or | or ) */ |
| 2693 |
case ')': |
case CHAR_RIGHT_PARENTHESIS: |
| 2694 |
*firstbyteptr = firstbyte; |
*firstbyteptr = firstbyte; |
| 2695 |
*reqbyteptr = reqbyte; |
*reqbyteptr = reqbyte; |
| 2696 |
*codeptr = code; |
*codeptr = code; |
| 2712 |
/* Handle single-character metacharacters. In multiline mode, ^ disables |
/* Handle single-character metacharacters. In multiline mode, ^ disables |
| 2713 |
the setting of any following char as a first character. */ |
the setting of any following char as a first character. */ |
| 2714 |
|
|
| 2715 |
case '^': |
case CHAR_CIRCUMFLEX_ACCENT: |
| 2716 |
if ((options & PCRE_MULTILINE) != 0) |
if ((options & PCRE_MULTILINE) != 0) |
| 2717 |
{ |
{ |
| 2718 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 2721 |
*code++ = OP_CIRC; |
*code++ = OP_CIRC; |
| 2722 |
break; |
break; |
| 2723 |
|
|
| 2724 |
case '$': |
case CHAR_DOLLAR_SIGN: |
| 2725 |
previous = NULL; |
previous = NULL; |
| 2726 |
*code++ = OP_DOLL; |
*code++ = OP_DOLL; |
| 2727 |
break; |
break; |
| 2729 |
/* 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 |
| 2730 |
repeats. The value of reqbyte doesn't change either. */ |
repeats. The value of reqbyte doesn't change either. */ |
| 2731 |
|
|
| 2732 |
case '.': |
case CHAR_DOT: |
| 2733 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 2734 |
zerofirstbyte = firstbyte; |
zerofirstbyte = firstbyte; |
| 2735 |
zeroreqbyte = reqbyte; |
zeroreqbyte = reqbyte; |
| 2753 |
In JavaScript compatibility mode, an isolated ']' causes an error. In |
In JavaScript compatibility mode, an isolated ']' causes an error. In |
| 2754 |
default (Perl) mode, it is treated as a data character. */ |
default (Perl) mode, it is treated as a data character. */ |
| 2755 |
|
|
| 2756 |
case ']': |
case CHAR_RIGHT_SQUARE_BRACKET: |
| 2757 |
if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) |
if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) |
| 2758 |
{ |
{ |
| 2759 |
*errorcodeptr = ERR64; |
*errorcodeptr = ERR64; |
| 2761 |
} |
} |
| 2762 |
goto NORMAL_CHAR; |
goto NORMAL_CHAR; |
| 2763 |
|
|
| 2764 |
case '[': |
case CHAR_LEFT_SQUARE_BRACKET: |
| 2765 |
previous = code; |
previous = code; |
| 2766 |
|
|
| 2767 |
/* PCRE supports POSIX class stuff inside a class. Perl gives an error if |
/* PCRE supports POSIX class stuff inside a class. Perl gives an error if |
| 2768 |
they are encountered at the top level, so we'll do that too. */ |
they are encountered at the top level, so we'll do that too. */ |
| 2769 |
|
|
| 2770 |
if ((ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && |
if ((ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || |
| 2771 |
|
ptr[1] == CHAR_EQUALS_SIGN) && |
| 2772 |
check_posix_syntax(ptr, &tempptr)) |
check_posix_syntax(ptr, &tempptr)) |
| 2773 |
{ |
{ |
| 2774 |
*errorcodeptr = (ptr[1] == ':')? ERR13 : ERR31; |
*errorcodeptr = (ptr[1] == CHAR_COLON)? ERR13 : ERR31; |
| 2775 |
goto FAILED; |
goto FAILED; |
| 2776 |
} |
} |
| 2777 |
|
|
| 2783 |
for (;;) |
for (;;) |
| 2784 |
{ |
{ |
| 2785 |
c = *(++ptr); |
c = *(++ptr); |
| 2786 |
if (c == '\\') |
if (c == CHAR_BACKSLASH) |
| 2787 |
{ |
{ |
| 2788 |
if (ptr[1] == 'E') ptr++; |
if (ptr[1] == CHAR_E) |
| 2789 |
else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3; |
ptr++; |
| 2790 |
else break; |
else if (strncmp((const char *)ptr+1, |
| 2791 |
|
STR_Q STR_BACKSLASH STR_E, 3) == 0) |
| 2792 |
|
ptr += 3; |
| 2793 |
|
else |
| 2794 |
|
break; |
| 2795 |
} |
} |
| 2796 |
else if (!negate_class && c == '^') |
else if (!negate_class && c == CHAR_CIRCUMFLEX_ACCENT) |
| 2797 |
negate_class = TRUE; |
negate_class = TRUE; |
| 2798 |
else break; |
else break; |
| 2799 |
} |
} |
| 2803 |
that. In JS mode, [] must always fail, so generate OP_FAIL, whereas |
that. In JS mode, [] must always fail, so generate OP_FAIL, whereas |
| 2804 |
[^] must match any character, so generate OP_ALLANY. */ |
[^] must match any character, so generate OP_ALLANY. */ |
| 2805 |
|
|
| 2806 |
if (c ==']' && (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) |
if (c == CHAR_RIGHT_SQUARE_BRACKET && |
| 2807 |
|
(cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) |
| 2808 |
{ |
{ |
| 2809 |
*code++ = negate_class? OP_ALLANY : OP_FAIL; |
*code++ = negate_class? OP_ALLANY : OP_FAIL; |
| 2810 |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; |
| 2869 |
|
|
| 2870 |
if (inescq) |
if (inescq) |
| 2871 |
{ |
{ |
| 2872 |
if (c == '\\' && ptr[1] == 'E') /* If we are at \E */ |
if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) /* If we are at \E */ |
| 2873 |
{ |
{ |
| 2874 |
inescq = FALSE; /* Reset literal state */ |
inescq = FALSE; /* Reset literal state */ |
| 2875 |
ptr++; /* Skip the 'E' */ |
ptr++; /* Skip the 'E' */ |
| 2884 |
[.ch.] and [=ch=] ("collating elements") and fault them, as Perl |
[.ch.] and [=ch=] ("collating elements") and fault them, as Perl |
| 2885 |
5.6 and 5.8 do. */ |
5.6 and 5.8 do. */ |
| 2886 |
|
|
| 2887 |
if (c == '[' && |
if (c == CHAR_LEFT_SQUARE_BRACKET && |
| 2888 |
(ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && |
(ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || |
| 2889 |
check_posix_syntax(ptr, &tempptr)) |
ptr[1] == CHAR_EQUALS_SIGN) && check_posix_syntax(ptr, &tempptr)) |
| 2890 |
{ |
{ |
| 2891 |
BOOL local_negate = FALSE; |
BOOL local_negate = FALSE; |
| 2892 |
int posix_class, taboffset, tabopt; |
int posix_class, taboffset, tabopt; |
| 2893 |
register const uschar *cbits = cd->cbits; |
register const uschar *cbits = cd->cbits; |
| 2894 |
uschar pbits[32]; |
uschar pbits[32]; |
| 2895 |
|
|
| 2896 |
if (ptr[1] != ':') |
if (ptr[1] != CHAR_COLON) |
| 2897 |
{ |
{ |
| 2898 |
*errorcodeptr = ERR31; |
*errorcodeptr = ERR31; |
| 2899 |
goto FAILED; |
goto FAILED; |
| 2900 |
} |
} |
| 2901 |
|
|
| 2902 |
ptr += 2; |
ptr += 2; |
| 2903 |
if (*ptr == '^') |
if (*ptr == CHAR_CIRCUMFLEX_ACCENT) |
| 2904 |
{ |
{ |
| 2905 |
local_negate = TRUE; |
local_negate = TRUE; |
| 2906 |
should_flip_negation = TRUE; /* Note negative special */ |
should_flip_negation = TRUE; /* Note negative special */ |
| 2973 |
to 'or' into the one we are building. We assume they have more than one |
to 'or' into the one we are building. We assume they have more than one |
| 2974 |
character in them, so set class_charcount bigger than one. */ |
character in them, so set class_charcount bigger than one. */ |
| 2975 |
|
|
| 2976 |
if (c == '\\') |
if (c == CHAR_BACKSLASH) |
| 2977 |
{ |
{ |
| 2978 |
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); |
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); |
| 2979 |
if (*errorcodeptr != 0) goto FAILED; |
if (*errorcodeptr != 0) goto FAILED; |
| 2980 |
|
|
| 2981 |
if (-c == ESC_b) c = '\b'; /* \b is backspace in a class */ |
if (-c == ESC_b) c = CHAR_BS; /* \b is backspace in a class */ |
| 2982 |
else if (-c == ESC_X) c = 'X'; /* \X is literal X in a class */ |
else if (-c == ESC_X) c = CHAR_X; /* \X is literal X in a class */ |
| 2983 |
else if (-c == ESC_R) c = 'R'; /* \R is literal R in a class */ |
else if (-c == ESC_R) c = CHAR_R; /* \R is literal R in a class */ |
| 2984 |
else if (-c == ESC_Q) /* Handle start of quoted string */ |
else if (-c == ESC_Q) /* Handle start of quoted string */ |
| 2985 |
{ |
{ |
| 2986 |
if (ptr[1] == '\\' && ptr[2] == 'E') |
if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) |
| 2987 |
{ |
{ |
| 2988 |
ptr += 2; /* avoid empty string */ |
ptr += 2; /* avoid empty string */ |
| 2989 |
} |
} |
| 3209 |
entirely. The code for handling \Q and \E is messy. */ |
entirely. The code for handling \Q and \E is messy. */ |
| 3210 |
|
|
| 3211 |
CHECK_RANGE: |
CHECK_RANGE: |
| 3212 |
while (ptr[1] == '\\' && ptr[2] == 'E') |
while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) |
| 3213 |
{ |
{ |
| 3214 |
inescq = FALSE; |
inescq = FALSE; |
| 3215 |
ptr += 2; |
ptr += 2; |
| 3219 |
|
|
| 3220 |
/* Remember \r or \n */ |
/* Remember \r or \n */ |
| 3221 |
|
|
| 3222 |
if (c == '\r' || c == '\n') cd->external_flags |= PCRE_HASCRORLF; |
if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; |
| 3223 |
|
|
| 3224 |
/* Check for range */ |
/* Check for range */ |
| 3225 |
|
|
| 3226 |
if (!inescq && ptr[1] == '-') |
if (!inescq && ptr[1] == CHAR_MINUS) |
| 3227 |
{ |
{ |
| 3228 |
int d; |
int d; |
| 3229 |
ptr += 2; |
ptr += 2; |
| 3230 |
while (*ptr == '\\' && ptr[1] == 'E') ptr += 2; |
while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2; |
| 3231 |
|
|
| 3232 |
/* If we hit \Q (not followed by \E) at this point, go into escaped |
/* If we hit \Q (not followed by \E) at this point, go into escaped |
| 3233 |
mode. */ |
mode. */ |
| 3234 |
|
|
| 3235 |
while (*ptr == '\\' && ptr[1] == 'Q') |
while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_Q) |
| 3236 |
{ |
{ |
| 3237 |
ptr += 2; |
ptr += 2; |
| 3238 |
if (*ptr == '\\' && ptr[1] == 'E') { ptr += 2; continue; } |
if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) |
| 3239 |
|
{ ptr += 2; continue; } |
| 3240 |
inescq = TRUE; |
inescq = TRUE; |
| 3241 |
break; |
break; |
| 3242 |
} |
} |
| 3243 |
|
|
| 3244 |
if (*ptr == 0 || (!inescq && *ptr == ']')) |
if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET)) |
| 3245 |
{ |
{ |
| 3246 |
ptr = oldptr; |
ptr = oldptr; |
| 3247 |
goto LONE_SINGLE_CHARACTER; |
goto LONE_SINGLE_CHARACTER; |
| 3260 |
not any of the other escapes. Perl 5.6 treats a hyphen as a literal |
not any of the other escapes. Perl 5.6 treats a hyphen as a literal |
| 3261 |
in such circumstances. */ |
in such circumstances. */ |
| 3262 |
|
|
| 3263 |
if (!inescq && d == '\\') |
if (!inescq && d == CHAR_BACKSLASH) |
| 3264 |
{ |
{ |
| 3265 |
d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); |
d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); |
| 3266 |
if (*errorcodeptr != 0) goto FAILED; |
if (*errorcodeptr != 0) goto FAILED; |
| 3270 |
|
|
| 3271 |
if (d < 0) |
if (d < 0) |
| 3272 |
{ |
{ |
| 3273 |
if (d == -ESC_b) d = '\b'; |
if (d == -ESC_b) d = CHAR_BS; |
| 3274 |
else if (d == -ESC_X) d = 'X'; |
else if (d == -ESC_X) d = CHAR_X; |
| 3275 |
else if (d == -ESC_R) d = 'R'; else |
else if (d == -ESC_R) d = CHAR_R; else |
| 3276 |
{ |
{ |
| 3277 |
ptr = oldptr; |
ptr = oldptr; |
| 3278 |
goto LONE_SINGLE_CHARACTER; /* A few lines below */ |
goto LONE_SINGLE_CHARACTER; /* A few lines below */ |
| 3293 |
|
|
| 3294 |
/* Remember \r or \n */ |
/* Remember \r or \n */ |
| 3295 |
|
|
| 3296 |
if (d == '\r' || d == '\n') cd->external_flags |= PCRE_HASCRORLF; |
if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; |
| 3297 |
|
|
| 3298 |
/* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless |
/* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless |
| 3299 |
matching, we have to use an XCLASS with extra data items. Caseless |
matching, we have to use an XCLASS with extra data items. Caseless |
| 3413 |
if ((options & PCRE_CASELESS) != 0) |
if ((options & PCRE_CASELESS) != 0) |
| 3414 |
{ |
{ |
| 3415 |
unsigned int othercase; |
unsigned int othercase; |
| 3416 |
if ((othercase = _pcre_ucp_othercase(c)) != NOTACHAR) |
if ((othercase = UCD_OTHERCASE(c)) != c) |
| 3417 |
{ |
{ |
| 3418 |
*class_utf8data++ = XCL_SINGLE; |
*class_utf8data++ = XCL_SINGLE; |
| 3419 |
class_utf8data += _pcre_ord2utf8(othercase, class_utf8data); |
class_utf8data += _pcre_ord2utf8(othercase, class_utf8data); |
| 3440 |
|
|
| 3441 |
/* Loop until ']' reached. This "while" is the end of the "do" above. */ |
/* Loop until ']' reached. This "while" is the end of the "do" above. */ |
| 3442 |
|
|
| 3443 |
while ((c = *(++ptr)) != 0 && (c != ']' || inescq)); |
while ((c = *(++ptr)) != 0 && (c != CHAR_RIGHT_SQUARE_BRACKET || inescq)); |
| 3444 |
|
|
| 3445 |
if (c == 0) /* Missing terminating ']' */ |
if (c == 0) /* Missing terminating ']' */ |
| 3446 |
{ |
{ |
| 3585 |
/* Various kinds of repeat; '{' is not necessarily a quantifier, but this |
/* Various kinds of repeat; '{' is not necessarily a quantifier, but this |
| 3586 |
has been tested above. */ |
has been tested above. */ |
| 3587 |
|
|
| 3588 |
case '{': |
case CHAR_LEFT_CURLY_BRACKET: |
| 3589 |
if (!is_quantifier) goto NORMAL_CHAR; |
if (!is_quantifier) goto NORMAL_CHAR; |
| 3590 |
ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr); |
ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr); |
| 3591 |
if (*errorcodeptr != 0) goto FAILED; |
if (*errorcodeptr != 0) goto FAILED; |
| 3592 |
goto REPEAT; |
goto REPEAT; |
| 3593 |
|
|
| 3594 |
case '*': |
case CHAR_ASTERISK: |
| 3595 |
repeat_min = 0; |
repeat_min = 0; |
| 3596 |
repeat_max = -1; |
repeat_max = -1; |
| 3597 |
goto REPEAT; |
goto REPEAT; |
| 3598 |
|
|
| 3599 |
case '+': |
case CHAR_PLUS: |
| 3600 |
repeat_min = 1; |
repeat_min = 1; |
| 3601 |
repeat_max = -1; |
repeat_max = -1; |
| 3602 |
goto REPEAT; |
goto REPEAT; |
| 3603 |
|
|
| 3604 |
case '?': |
case CHAR_QUESTION_MARK: |
| 3605 |
repeat_min = 0; |
repeat_min = 0; |
| 3606 |
repeat_max = 1; |
repeat_max = 1; |
| 3607 |
|
|
| 3636 |
but if PCRE_UNGREEDY is set, it works the other way round. We change the |
but if PCRE_UNGREEDY is set, it works the other way round. We change the |
| 3637 |
repeat type to the non-default. */ |
repeat type to the non-default. */ |
| 3638 |
|
|
| 3639 |
if (ptr[1] == '+') |
if (ptr[1] == CHAR_PLUS) |
| 3640 |
{ |
{ |
| 3641 |
repeat_type = 0; /* Force greedy */ |
repeat_type = 0; /* Force greedy */ |
| 3642 |
possessive_quantifier = TRUE; |
possessive_quantifier = TRUE; |
| 3643 |
ptr++; |
ptr++; |
| 3644 |
} |
} |
| 3645 |
else if (ptr[1] == '?') |
else if (ptr[1] == CHAR_QUESTION_MARK) |
| 3646 |
{ |
{ |
| 3647 |
repeat_type = greedy_non_default; |
repeat_type = greedy_non_default; |
| 3648 |
ptr++; |
ptr++; |
| 4275 |
lookbehind or option setting or condition or all the other extended |
lookbehind or option setting or condition or all the other extended |
| 4276 |
parenthesis forms. */ |
parenthesis forms. */ |
| 4277 |
|
|
| 4278 |
case '(': |
case CHAR_LEFT_PARENTHESIS: |
| 4279 |
newoptions = options; |
newoptions = options; |
| 4280 |
skipbytes = 0; |
skipbytes = 0; |
| 4281 |
bravalue = OP_CBRA; |
bravalue = OP_CBRA; |
| 4284 |
|
|
| 4285 |
/* First deal with various "verbs" that can be introduced by '*'. */ |
/* First deal with various "verbs" that can be introduced by '*'. */ |
| 4286 |
|
|
| 4287 |
if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0) |
if (*(++ptr) == CHAR_ASTERISK && (cd->ctypes[ptr[1]] & ctype_letter) != 0) |
| 4288 |
{ |
{ |
| 4289 |
int i, namelen; |
int i, namelen; |
| 4290 |
const char *vn = verbnames; |
const char *vn = verbnames; |
| 4291 |
const uschar *name = ++ptr; |
const uschar *name = ++ptr; |
| 4292 |
previous = NULL; |
previous = NULL; |
| 4293 |
while ((cd->ctypes[*++ptr] & ctype_letter) != 0); |
while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {}; |
| 4294 |
if (*ptr == ':') |
if (*ptr == CHAR_COLON) |
| 4295 |
{ |
{ |
| 4296 |
*errorcodeptr = ERR59; /* Not supported */ |
*errorcodeptr = ERR59; /* Not supported */ |
| 4297 |
goto FAILED; |
goto FAILED; |
| 4298 |
} |
} |
| 4299 |
if (*ptr != ')') |
if (*ptr != CHAR_RIGHT_PARENTHESIS) |
| 4300 |
{ |
{ |
| 4301 |
*errorcodeptr = ERR60; |
*errorcodeptr = ERR60; |
| 4302 |
goto FAILED; |
goto FAILED; |
| 4321 |
/* Deal with the extended parentheses; all are introduced by '?', and the |
/* Deal with the extended parentheses; all are introduced by '?', and the |
| 4322 |
appearance of any of them means that this is not a capturing group. */ |
appearance of any of them means that this is not a capturing group. */ |
| 4323 |
|
|
| 4324 |
else if (*ptr == '?') |
else if (*ptr == CHAR_QUESTION_MARK) |
| 4325 |
{ |
{ |
| 4326 |
int i, set, unset, namelen; |
int i, set, unset, namelen; |
| 4327 |
int *optset; |
int *optset; |
| 4330 |
|
|
| 4331 |
switch (*(++ptr)) |
switch (*(++ptr)) |
| 4332 |
{ |
{ |
| 4333 |
case '#': /* Comment; skip to ket */ |
case CHAR_NUMBER_SIGN: /* Comment; skip to ket */ |
| 4334 |
ptr++; |
ptr++; |
| 4335 |
while (*ptr != 0 && *ptr != ')') ptr++; |
while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; |
| 4336 |
if (*ptr == 0) |
if (*ptr == 0) |
| 4337 |
{ |
{ |
| 4338 |
*errorcodeptr = ERR18; |
*errorcodeptr = ERR18; |
| 4342 |
|
|
| 4343 |
|
|
| 4344 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4345 |
case '|': /* Reset capture count for each branch */ |
case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */ |
| 4346 |
reset_bracount = TRUE; |
reset_bracount = TRUE; |
| 4347 |
/* Fall through */ |
/* Fall through */ |
| 4348 |
|
|
| 4349 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4350 |
case ':': /* Non-capturing bracket */ |
case CHAR_COLON: /* Non-capturing bracket */ |
| 4351 |
bravalue = OP_BRA; |
bravalue = OP_BRA; |
| 4352 |
ptr++; |
ptr++; |
| 4353 |
break; |
break; |
| 4354 |
|
|
| 4355 |
|
|
| 4356 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4357 |
case '(': |
case CHAR_LEFT_PARENTHESIS: |
| 4358 |
bravalue = OP_COND; /* Conditional group */ |
bravalue = OP_COND; /* Conditional group */ |
| 4359 |
|
|
| 4360 |
/* A condition can be an assertion, a number (referring to a numbered |
/* A condition can be an assertion, a number (referring to a numbered |
| 4374 |
the switch. This will take control down to where bracketed groups, |
the switch. This will take control down to where bracketed groups, |
| 4375 |
including assertions, are processed. */ |
including assertions, are processed. */ |
| 4376 |
|
|
| 4377 |
if (ptr[1] == '?' && (ptr[2] == '=' || ptr[2] == '!' || ptr[2] == '<')) |
if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN || |
| 4378 |
|
ptr[2] == CHAR_EXCLAMATION_MARK || ptr[2] == CHAR_LESS_THAN_SIGN)) |
| 4379 |
break; |
break; |
| 4380 |
|
|
| 4381 |
/* Most other conditions use OP_CREF (a couple change to OP_RREF |
/* Most other conditions use OP_CREF (a couple change to OP_RREF |
| 4387 |
|
|
| 4388 |
/* Check for a test for recursion in a named group. */ |
/* Check for a test for recursion in a named group. */ |
| 4389 |
|
|
| 4390 |
if (ptr[1] == 'R' && ptr[2] == '&') |
if (ptr[1] == CHAR_R && ptr[2] == CHAR_AMPERSAND) |
| 4391 |
{ |
{ |
| 4392 |
terminator = -1; |
terminator = -1; |
| 4393 |
ptr += 2; |
ptr += 2; |
| 4397 |
/* Check for a test for a named group's having been set, using the Perl |
/* Check for a test for a named group's having been set, using the Perl |
| 4398 |
syntax (?(<name>) or (?('name') */ |
syntax (?(<name>) or (?('name') */ |
| 4399 |
|
|
| 4400 |
else if (ptr[1] == '<') |
else if (ptr[1] == CHAR_LESS_THAN_SIGN) |
| 4401 |
{ |
{ |
| 4402 |
terminator = '>'; |
terminator = CHAR_GREATER_THAN_SIGN; |
| 4403 |
ptr++; |
ptr++; |
| 4404 |
} |
} |
| 4405 |
else if (ptr[1] == '\'') |
else if (ptr[1] == CHAR_APOSTROPHE) |
| 4406 |
{ |
{ |
| 4407 |
terminator = '\''; |
terminator = CHAR_APOSTROPHE; |
| 4408 |
ptr++; |
ptr++; |
| 4409 |
} |
} |
| 4410 |
else |
else |
| 4411 |
{ |
{ |
| 4412 |
terminator = 0; |
terminator = 0; |
| 4413 |
if (ptr[1] == '-' || ptr[1] == '+') refsign = *(++ptr); |
if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr); |
| 4414 |
} |
} |
| 4415 |
|
|
| 4416 |
/* We now expect to read a name; any thing else is an error */ |
/* We now expect to read a name; any thing else is an error */ |
| 4430 |
{ |
{ |
| 4431 |
if (recno >= 0) |
if (recno >= 0) |
| 4432 |
recno = ((digitab[*ptr] & ctype_digit) != 0)? |
recno = ((digitab[*ptr] & ctype_digit) != 0)? |
| 4433 |
recno * 10 + *ptr - '0' : -1; |
recno * 10 + *ptr - CHAR_0 : -1; |
| 4434 |
ptr++; |
ptr++; |
| 4435 |
} |
} |
| 4436 |
namelen = ptr - name; |
namelen = ptr - name; |
| 4437 |
|
|
| 4438 |
if ((terminator > 0 && *ptr++ != terminator) || *ptr++ != ')') |
if ((terminator > 0 && *ptr++ != terminator) || |
| 4439 |
|
*ptr++ != CHAR_RIGHT_PARENTHESIS) |
| 4440 |
{ |
{ |
| 4441 |
ptr--; /* Error offset */ |
ptr--; /* Error offset */ |
| 4442 |
*errorcodeptr = ERR26; |
*errorcodeptr = ERR26; |
| 4458 |
*errorcodeptr = ERR58; |
*errorcodeptr = ERR58; |
| 4459 |
goto FAILED; |
goto FAILED; |
| 4460 |
} |
} |
| 4461 |
recno = (refsign == '-')? |
recno = (refsign == CHAR_MINUS)? |
| 4462 |
cd->bracount - recno + 1 : recno +cd->bracount; |
cd->bracount - recno + 1 : recno +cd->bracount; |
| 4463 |
if (recno <= 0 || recno > cd->final_bracount) |
if (recno <= 0 || recno > cd->final_bracount) |
| 4464 |
{ |
{ |
| 4510 |
/* Check for (?(R) for recursion. Allow digits after R to specify a |
/* Check for (?(R) for recursion. Allow digits after R to specify a |
| 4511 |
specific group number. */ |
specific group number. */ |
| 4512 |
|
|
| 4513 |
else if (*name == 'R') |
else if (*name == CHAR_R) |
| 4514 |
{ |
{ |
| 4515 |
recno = 0; |
recno = 0; |
| 4516 |
for (i = 1; i < namelen; i++) |
for (i = 1; i < namelen; i++) |
| 4520 |
*errorcodeptr = ERR15; |
*errorcodeptr = ERR15; |
| 4521 |
goto FAILED; |
goto FAILED; |
| 4522 |
} |
} |
| 4523 |
recno = recno * 10 + name[i] - '0'; |
recno = recno * 10 + name[i] - CHAR_0; |
| 4524 |
} |
} |
| 4525 |
if (recno == 0) recno = RREF_ANY; |
if (recno == 0) recno = RREF_ANY; |
| 4526 |
code[1+LINK_SIZE] = OP_RREF; /* Change test type */ |
code[1+LINK_SIZE] = OP_RREF; /* Change test type */ |
| 4530 |
/* Similarly, check for the (?(DEFINE) "condition", which is always |
/* Similarly, check for the (?(DEFINE) "condition", which is always |
| 4531 |
false. */ |
false. */ |
| 4532 |
|
|
| 4533 |
else if (namelen == 6 && strncmp((char *)name, "DEFINE", 6) == 0) |
else if (namelen == 6 && strncmp((char *)name, STRING_DEFINE, 6) == 0) |
| 4534 |
{ |
{ |
| 4535 |
code[1+LINK_SIZE] = OP_DEF; |
code[1+LINK_SIZE] = OP_DEF; |
| 4536 |
skipbytes = 1; |
skipbytes = 1; |
| 4555 |
|
|
| 4556 |
|
|
| 4557 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4558 |
case '=': /* Positive lookahead */ |
case CHAR_EQUALS_SIGN: /* Positive lookahead */ |
| 4559 |
bravalue = OP_ASSERT; |
bravalue = OP_ASSERT; |
| 4560 |
ptr++; |
ptr++; |
| 4561 |
break; |
break; |
| 4562 |
|
|
| 4563 |
|
|
| 4564 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4565 |
case '!': /* Negative lookahead */ |
case CHAR_EXCLAMATION_MARK: /* Negative lookahead */ |
| 4566 |
ptr++; |
ptr++; |
| 4567 |
if (*ptr == ')') /* Optimize (?!) */ |
if (*ptr == CHAR_RIGHT_PARENTHESIS) /* Optimize (?!) */ |
| 4568 |
{ |
{ |
| 4569 |
*code++ = OP_FAIL; |
*code++ = OP_FAIL; |
| 4570 |
previous = NULL; |
previous = NULL; |
| 4575 |
|
|
| 4576 |
|
|
| 4577 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4578 |
case '<': /* Lookbehind or named define */ |
case CHAR_LESS_THAN_SIGN: /* Lookbehind or named define */ |
| 4579 |
switch (ptr[1]) |
switch (ptr[1]) |
| 4580 |
{ |
{ |
| 4581 |
case '=': /* Positive lookbehind */ |
case CHAR_EQUALS_SIGN: /* Positive lookbehind */ |
| 4582 |
bravalue = OP_ASSERTBACK; |
bravalue = OP_ASSERTBACK; |
| 4583 |
ptr += 2; |
ptr += 2; |
| 4584 |
break; |
break; |
| 4585 |
|
|
| 4586 |
case '!': /* Negative lookbehind */ |
case CHAR_EXCLAMATION_MARK: /* Negative lookbehind */ |
| 4587 |
bravalue = OP_ASSERTBACK_NOT; |
bravalue = OP_ASSERTBACK_NOT; |
| 4588 |
ptr += 2; |
ptr += 2; |
| 4589 |
break; |
break; |
| 4598 |
|
|
| 4599 |
|
|
| 4600 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4601 |
case '>': /* One-time brackets */ |
case CHAR_GREATER_THAN_SIGN: /* One-time brackets */ |
| 4602 |
bravalue = OP_ONCE; |
bravalue = OP_ONCE; |
| 4603 |
ptr++; |
ptr++; |
| 4604 |
break; |
break; |
| 4605 |
|
|
| 4606 |
|
|
| 4607 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4608 |
case 'C': /* Callout - may be followed by digits; */ |
case CHAR_C: /* Callout - may be followed by digits; */ |
| 4609 |
previous_callout = code; /* Save for later completion */ |
previous_callout = code; /* Save for later completion */ |
| 4610 |
after_manual_callout = 1; /* Skip one item before completing */ |
after_manual_callout = 1; /* Skip one item before completing */ |
| 4611 |
*code++ = OP_CALLOUT; |
*code++ = OP_CALLOUT; |
| 4612 |
{ |
{ |
| 4613 |
int n = 0; |
int n = 0; |
| 4614 |
while ((digitab[*(++ptr)] & ctype_digit) != 0) |
while ((digitab[*(++ptr)] & ctype_digit) != 0) |
| 4615 |
n = n * 10 + *ptr - '0'; |
n = n * 10 + *ptr - CHAR_0; |
| 4616 |
if (*ptr != ')') |
if (*ptr != CHAR_RIGHT_PARENTHESIS) |
| 4617 |
{ |
{ |
| 4618 |
*errorcodeptr = ERR39; |
*errorcodeptr = ERR39; |
| 4619 |
goto FAILED; |
goto FAILED; |
| 4633 |
|
|
| 4634 |
|
|
| 4635 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4636 |
case 'P': /* Python-style named subpattern handling */ |
case CHAR_P: /* Python-style named subpattern handling */ |
| 4637 |
if (*(++ptr) == '=' || *ptr == '>') /* Reference or recursion */ |
if (*(++ptr) == CHAR_EQUALS_SIGN || |
| 4638 |
|
*ptr == CHAR_GREATER_THAN_SIGN) /* Reference or recursion */ |
| 4639 |
{ |
{ |
| 4640 |
is_recurse = *ptr == '>'; |
is_recurse = *ptr == CHAR_GREATER_THAN_SIGN; |
| 4641 |
terminator = ')'; |
terminator = CHAR_RIGHT_PARENTHESIS; |
| 4642 |
goto NAMED_REF_OR_RECURSE; |
goto NAMED_REF_OR_RECURSE; |
| 4643 |
} |
} |
| 4644 |
else if (*ptr != '<') /* Test for Python-style definition */ |
else if (*ptr != CHAR_LESS_THAN_SIGN) /* Test for Python-style defn */ |
| 4645 |
{ |
{ |
| 4646 |
*errorcodeptr = ERR41; |
*errorcodeptr = ERR41; |
| 4647 |
goto FAILED; |
goto FAILED; |
| 4651 |
|
|
| 4652 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4653 |
DEFINE_NAME: /* Come here from (?< handling */ |
DEFINE_NAME: /* Come here from (?< handling */ |
| 4654 |
case '\'': |
case CHAR_APOSTROPHE: |
| 4655 |
{ |
{ |
| 4656 |
terminator = (*ptr == '<')? '>' : '\''; |
terminator = (*ptr == CHAR_LESS_THAN_SIGN)? |
| 4657 |
|
CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; |
| 4658 |
name = ++ptr; |
name = ++ptr; |
| 4659 |
|
|
| 4660 |
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; |
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; |
| 4728 |
|
|
| 4729 |
|
|
| 4730 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4731 |
case '&': /* Perl recursion/subroutine syntax */ |
case CHAR_AMPERSAND: /* Perl recursion/subroutine syntax */ |
| 4732 |
terminator = ')'; |
terminator = CHAR_RIGHT_PARENTHESIS; |
| 4733 |
is_recurse = TRUE; |
is_recurse = TRUE; |
| 4734 |
/* Fall through */ |
/* Fall through */ |
| 4735 |
|
|
| 4804 |
|
|
| 4805 |
|
|
| 4806 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4807 |
case 'R': /* Recursion */ |
case CHAR_R: /* Recursion */ |
| 4808 |
ptr++; /* Same as (?0) */ |
ptr++; /* Same as (?0) */ |
| 4809 |
/* Fall through */ |
/* Fall through */ |
| 4810 |
|
|
| 4811 |
|
|
| 4812 |
/* ------------------------------------------------------------ */ |
/* ------------------------------------------------------------ */ |
| 4813 |
case '-': case '+': |
case CHAR_MINUS: case CHAR_PLUS: /* Recursion or subroutine */ |
| 4814 |
case '0': case '1': case '2': case '3': case '4': /* Recursion or */ |
case CHAR_0: case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: |
| 4815 |
case '5': case '6': case '7': case '8': case '9': /* subroutine */ |
case CHAR_5: case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: |
| 4816 |
{ |
{ |
| 4817 |
const uschar *called; |
const uschar *called; |
| 4818 |
terminator = ')'; |
terminator = CHAR_RIGHT_PARENTHESIS; |
| 4819 |
|
|
| 4820 |
/* Come here from the \g<...> and \g'...' code (Oniguruma |
/* Come here from the \g<...> and \g'...' code (Oniguruma |
| 4821 |
compatibility). However, the syntax has been checked to ensure that |
compatibility). However, the syntax has been checked to ensure that |
| 4825 |
|
|
| 4826 |
HANDLE_NUMERICAL_RECURSION: |
HANDLE_NUMERICAL_RECURSION: |
| 4827 |
|
|
| 4828 |
if ((refsign = *ptr) == '+') |
if ((refsign = *ptr) == CHAR_PLUS) |
| 4829 |
{ |
{ |
| 4830 |
ptr++; |
ptr++; |
| 4831 |
if ((digitab[*ptr] & ctype_digit) == 0) |
if ((digitab[*ptr] & ctype_digit) == 0) |
| 4834 |
goto FAILED; |
goto FAILED; |
| 4835 |
} |
} |
| 4836 |
} |
} |
| 4837 |
else if (refsign == '-') |
else if (refsign == CHAR_MINUS) |
| 4838 |
{ |
{ |
| 4839 |
if ((digitab[ptr[1]] & ctype_digit) == 0) |
if ((digitab[ptr[1]] & ctype_digit) == 0) |
| 4840 |
goto OTHER_CHAR_AFTER_QUERY; |
goto OTHER_CHAR_AFTER_QUERY; |
| 4843 |
|
|
| 4844 |
recno = 0; |
recno = 0; |
| 4845 |
while((digitab[*ptr] & ctype_digit) != 0) |
while((digitab[*ptr] & ctype_digit) != 0) |
| 4846 |
recno = recno * 10 + *ptr++ - '0'; |
recno = recno * 10 + *ptr++ - CHAR_0; |
| 4847 |
|
|
| 4848 |
if (*ptr != terminator) |
if (*ptr != terminator) |
| 4849 |
{ |
{ |
| 4851 |
goto FAILED; |
goto FAILED; |
| 4852 |
} |
} |
| 4853 |
|
|
| 4854 |
if (refsign == '-') |
if (refsign == CHAR_MINUS) |
| 4855 |
{ |
{ |
| 4856 |
if (recno == 0) |
if (recno == 0) |
| 4857 |
{ |
{ |
| 4865 |
goto FAILED; |
goto FAILED; |
| 4866 |
} |
} |
| 4867 |
} |
} |
| 4868 |
else if (refsign == '+') |
else if (refsign == CHAR_PLUS) |
| 4869 |
{ |
{ |
| 4870 |
if (recno == 0) |
if (recno == 0) |
| 4871 |
{ |
{ |
| 4951 |
set = unset = 0; |
set = unset = 0; |
| 4952 |
optset = &set; |
optset = &set; |
| 4953 |
|
|
| 4954 |
while (*ptr != ')' && *ptr != ':') |
while (*ptr != CHAR_RIGHT_PARENTHESIS && *ptr != CHAR_COLON) |
| 4955 |
{ |
{ |
| 4956 |
switch (*ptr++) |
switch (*ptr++) |
| 4957 |
{ |
{ |
| 4958 |
case '-': optset = &unset; break; |
case CHAR_MINUS: optset = &unset; break; |
| 4959 |
|
|
| 4960 |
case 'J': /* Record that it changed in the external options */ |
case CHAR_J: /* Record that it changed in the external options */ |
| 4961 |
*optset |= PCRE_DUPNAMES; |
*optset |= PCRE_DUPNAMES; |
| 4962 |
cd->external_flags |= PCRE_JCHANGED; |
cd->external_flags |= PCRE_JCHANGED; |
| 4963 |
break; |
break; |
| 4964 |
|
|
| 4965 |
case 'i': *optset |= PCRE_CASELESS; break; |
case CHAR_i: *optset |= PCRE_CASELESS; break; |
| 4966 |
case 'm': *optset |= PCRE_MULTILINE; break; |
case CHAR_m: *optset |= PCRE_MULTILINE; break; |
| 4967 |
case 's': *optset |= PCRE_DOTALL; break; |
case CHAR_s: *optset |= PCRE_DOTALL; break; |
| 4968 |
case 'x': *optset |= PCRE_EXTENDED; break; |
case CHAR_x: *optset |= PCRE_EXTENDED; break; |
| 4969 |
case 'U': *optset |= PCRE_UNGREEDY; break; |
case CHAR_U: *optset |= PCRE_UNGREEDY; break; |
| 4970 |
case 'X': *optset |= PCRE_EXTRA; break; |
case CHAR_X: *optset |= PCRE_EXTRA; break; |
| 4971 |
|
|
| 4972 |
default: *errorcodeptr = ERR12; |
default: *errorcodeptr = ERR12; |
| 4973 |
ptr--; /* Correct the offset */ |
ptr--; /* Correct the offset */ |
| 4998 |
both phases. |
both phases. |
| 4999 |
|
|
| 5000 |
If we are not at the pattern start, compile code to change the ims |
If we are not at the pattern start, compile code to change the ims |
| 5001 |
options if this setting actually changes any of them. We also pass the |
options if this setting actually changes any of them, and reset the |
| 5002 |
new setting back so that it can be put at the start of any following |
greedy defaults and the case value for firstbyte and reqbyte. */ |
|
branches, and when this group ends (if we are in a group), a resetting |
|
|
item can be compiled. */ |
|
| 5003 |
|
|
| 5004 |
if (*ptr == ')') |
if (*ptr == CHAR_RIGHT_PARENTHESIS) |
| 5005 |
{ |
{ |
| 5006 |
if (code == cd->start_code + 1 + LINK_SIZE && |
if (code == cd->start_code + 1 + LINK_SIZE && |
| 5007 |
(lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE)) |
(lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE)) |
| 5008 |
{ |
{ |
| 5009 |
cd->external_options = newoptions; |
cd->external_options = newoptions; |
|
options = newoptions; |
|
| 5010 |
} |
} |
| 5011 |
else |
else |
| 5012 |
{ |
{ |
| 5015 |
*code++ = OP_OPT; |
*code++ = OP_OPT; |
| 5016 |
*code++ = newoptions & PCRE_IMS; |
*code++ = newoptions & PCRE_IMS; |
| 5017 |
} |
} |
|
|
|
|
/* Change options at this level, and pass them back for use |
|
|
in subsequent branches. Reset the greedy defaults and the case |
|
|
value for firstbyte and reqbyte. */ |
|
|
|
|
|
*optionsptr = options = newoptions; |
|
| 5018 |
greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); |
greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); |
| 5019 |
greedy_non_default = greedy_default ^ 1; |
greedy_non_default = greedy_default ^ 1; |
| 5020 |
req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; |
req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; |
| 5021 |
} |
} |
| 5022 |
|
|
| 5023 |
|
/* Change options at this level, and pass them back for use |
| 5024 |
|
in subsequent branches. When not at the start of the pattern, this |
| 5025 |
|
information is also necessary so that a resetting item can be |
| 5026 |
|
compiled at the end of a group (if we are in a group). */ |
| 5027 |
|
|
| 5028 |
|
*optionsptr = options = newoptions; |
| 5029 |
previous = NULL; /* This item can't be repeated */ |
previous = NULL; /* This item can't be repeated */ |
| 5030 |
continue; /* It is complete */ |
continue; /* It is complete */ |
| 5031 |
} |
} |
| 5141 |
|
|
| 5142 |
/* Error if hit end of pattern */ |
/* Error if hit end of pattern */ |
| 5143 |
|
|
| 5144 |
if (*ptr != ')') |
if (*ptr != CHAR_RIGHT_PARENTHESIS) |
| 5145 |
{ |
{ |
| 5146 |
*errorcodeptr = ERR14; |
*errorcodeptr = ERR14; |
| 5147 |
goto FAILED; |
goto FAILED; |
| 5239 |
We can test for values between ESC_b and ESC_Z for the latter; this may |
We can test for values between ESC_b and ESC_Z for the latter; this may |
| 5240 |
have to change if any new ones are ever created. */ |
have to change if any new ones are ever created. */ |
| 5241 |
|
|
| 5242 |
case '\\': |
case CHAR_BACKSLASH: |
| 5243 |
tempptr = ptr; |
tempptr = ptr; |
| 5244 |
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE); |
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE); |
| 5245 |
if (*errorcodeptr != 0) goto FAILED; |
if (*errorcodeptr != 0) goto FAILED; |
| 5248 |
{ |
{ |
| 5249 |
if (-c == ESC_Q) /* Handle start of quoted string */ |
if (-c == ESC_Q) /* Handle start of quoted string */ |
| 5250 |
{ |
{ |
| 5251 |
if (ptr[1] == '\\' && ptr[2] == 'E') ptr += 2; /* avoid empty string */ |
if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) |
| 5252 |
else inescq = TRUE; |
ptr += 2; /* avoid empty string */ |
| 5253 |
|
else inescq = TRUE; |
| 5254 |
continue; |
continue; |
| 5255 |
} |
} |
| 5256 |
|
|
| 5278 |
{ |
{ |
| 5279 |
const uschar *p; |
const uschar *p; |
| 5280 |
save_hwm = cd->hwm; /* Normally this is set when '(' is read */ |
save_hwm = cd->hwm; /* Normally this is set when '(' is read */ |
| 5281 |
terminator = (*(++ptr) == '<')? '>' : '\''; |
terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? |
| 5282 |
|
CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; |
| 5283 |
|
|
| 5284 |
/* These two statements stop the compiler for warning about possibly |
/* These two statements stop the compiler for warning about possibly |
| 5285 |
unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In |
unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In |
| 5291 |
|
|
| 5292 |
/* Test for a name */ |
/* Test for a name */ |
| 5293 |
|
|
| 5294 |
if (ptr[1] != '+' && ptr[1] != '-') |
if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS) |
| 5295 |
{ |
{ |
| 5296 |
BOOL isnumber = TRUE; |
BOOL isnumber = TRUE; |
| 5297 |
for (p = ptr + 1; *p != 0 && *p != terminator; p++) |
for (p = ptr + 1; *p != 0 && *p != terminator; p++) |
| 5329 |
/* \k<name> or \k'name' is a back reference by name (Perl syntax). |
/* \k<name> or \k'name' is a back reference by name (Perl syntax). |
| 5330 |
We also support \k{name} (.NET syntax) */ |
We also support \k{name} (.NET syntax) */ |
| 5331 |
|
|
| 5332 |
if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\'' || ptr[1] == '{')) |
if (-c == ESC_k && (ptr[1] == CHAR_LESS_THAN_SIGN || |
| 5333 |
|
ptr[1] == CHAR_APOSTROPHE || ptr[1] == CHAR_LEFT_CURLY_BRACKET)) |
| 5334 |
{ |
{ |
| 5335 |
is_recurse = FALSE; |
is_recurse = FALSE; |
| 5336 |
terminator = (*(++ptr) == '<')? '>' : (*ptr == '\'')? '\'' : '}'; |
terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? |
| 5337 |
|
CHAR_GREATER_THAN_SIGN : (*ptr == CHAR_APOSTROPHE)? |
| 5338 |
|
CHAR_APOSTROPHE : CHAR_RIGHT_CURLY_BRACKET; |
| 5339 |
goto NAMED_REF_OR_RECURSE; |
goto NAMED_REF_OR_RECURSE; |
| 5340 |
} |
} |
| 5341 |
|
|
| 5438 |
|
|
| 5439 |
/* Remember if \r or \n were seen */ |
/* Remember if \r or \n were seen */ |
| 5440 |
|
|
| 5441 |
if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n') |
if (mcbuffer[0] == CHAR_CR || mcbuffer[0] == CHAR_NL) |
| 5442 |
cd->external_flags |= PCRE_HASCRORLF; |
cd->external_flags |= PCRE_HASCRORLF; |
| 5443 |
|
|
| 5444 |
/* Set the first and required bytes appropriately. If no previous first |
/* Set the first and required bytes appropriately. If no previous first |
| 5683 |
compile a resetting op-code following, except at the very end of the pattern. |
compile a resetting op-code following, except at the very end of the pattern. |
| 5684 |
Return leaving the pointer at the terminating char. */ |
Return leaving the pointer at the terminating char. */ |
| 5685 |
|
|
| 5686 |
if (*ptr != '|') |
if (*ptr != CHAR_VERTICAL_LINE) |
| 5687 |
{ |
{ |
| 5688 |
if (lengthptr == NULL) |
if (lengthptr == NULL) |
| 5689 |
{ |
{ |
| 5706 |
|
|
| 5707 |
/* Resetting option if needed */ |
/* Resetting option if needed */ |
| 5708 |
|
|
| 5709 |
if ((options & PCRE_IMS) != oldims && *ptr == ')') |
if ((options & PCRE_IMS) != oldims && *ptr == CHAR_RIGHT_PARENTHESIS) |
| 5710 |
{ |
{ |
| 5711 |
*code++ = OP_OPT; |
*code++ = OP_OPT; |
| 5712 |
*code++ = oldims; |
*code++ = oldims; |
| 5888 |
NULL, 0, FALSE); |
NULL, 0, FALSE); |
| 5889 |
register int op = *scode; |
register int op = *scode; |
| 5890 |
|
|
| 5891 |
|
/* If we are at the start of a conditional assertion group, *both* the |
| 5892 |
|
conditional assertion *and* what follows the condition must satisfy the test |
| 5893 |
|
for start of line. Other kinds of condition fail. Note that there may be an |
| 5894 |
|
auto-callout at the start of a condition. */ |
| 5895 |
|
|
| 5896 |
|
if (op == OP_COND) |
| 5897 |
|
{ |
| 5898 |
|
scode += 1 + LINK_SIZE; |
| 5899 |
|
if (*scode == OP_CALLOUT) scode += _pcre_OP_lengths[OP_CALLOUT]; |
| 5900 |
|
switch (*scode) |
| 5901 |
|
{ |
| 5902 |
|
case OP_CREF: |
| 5903 |
|
case OP_RREF: |
| 5904 |
|
case OP_DEF: |
| 5905 |
|
return FALSE; |
| 5906 |
|
|
| 5907 |
|
default: /* Assertion */ |
| 5908 |
|
if (!is_startline(scode, bracket_map, backref_map)) return FALSE; |
| 5909 |
|
do scode += GET(scode, 1); while (*scode == OP_ALT); |
| 5910 |
|
scode += 1 + LINK_SIZE; |
| 5911 |
|
break; |
| 5912 |
|
} |
| 5913 |
|
scode = first_significant_code(scode, NULL, 0, FALSE); |
| 5914 |
|
op = *scode; |
| 5915 |
|
} |
| 5916 |
|
|
| 5917 |
/* Non-capturing brackets */ |
/* Non-capturing brackets */ |
| 5918 |
|
|
| 5919 |
if (op == OP_BRA) |
if (op == OP_BRA) |
| 5932 |
|
|
| 5933 |
/* Other brackets */ |
/* Other brackets */ |
| 5934 |
|
|
| 5935 |
else if (op == OP_ASSERT || op == OP_ONCE || op == OP_COND) |
else if (op == OP_ASSERT || op == OP_ONCE) |
| 5936 |
{ if (!is_startline(scode, bracket_map, backref_map)) return FALSE; } |
{ |
| 5937 |
|
if (!is_startline(scode, bracket_map, backref_map)) return FALSE; |
| 5938 |
|
} |
| 5939 |
|
|
| 5940 |
/* .* means "start at start or after \n" if it isn't in brackets that |
/* .* means "start at start or after \n" if it isn't in brackets that |
| 5941 |
may be referenced. */ |
may be referenced. */ |
| 6052 |
with errorptr and erroroffset set |
with errorptr and erroroffset set |
| 6053 |
*/ |
*/ |
| 6054 |
|
|
| 6055 |
PCRE_EXP_DEFN pcre * |
PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION |
| 6056 |
pcre_compile(const char *pattern, int options, const char **errorptr, |
pcre_compile(const char *pattern, int options, const char **errorptr, |
| 6057 |
int *erroroffset, const unsigned char *tables) |
int *erroroffset, const unsigned char *tables) |
| 6058 |
{ |
{ |
| 6060 |
} |
} |
| 6061 |
|
|
| 6062 |
|
|
| 6063 |
PCRE_EXP_DEFN pcre * |
PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION |
| 6064 |
pcre_compile2(const char *pattern, int options, int *errorcodeptr, |
pcre_compile2(const char *pattern, int options, int *errorcodeptr, |
| 6065 |
const char **errorptr, int *erroroffset, const unsigned char *tables) |
const char **errorptr, int *erroroffset, const unsigned char *tables) |
| 6066 |
{ |
{ |
| 6132 |
} |
} |
| 6133 |
#endif |
#endif |
| 6134 |
|
|
| 6135 |
if ((options & ~PUBLIC_OPTIONS) != 0) |
if ((options & ~PUBLIC_COMPILE_OPTIONS) != 0) |
| 6136 |
{ |
{ |
| 6137 |
errorcode = ERR17; |
errorcode = ERR17; |
| 6138 |
goto PCRE_EARLY_ERROR_RETURN; |
goto PCRE_EARLY_ERROR_RETURN; |
| 6149 |
/* Check for global one-time settings at the start of the pattern, and remember |
/* Check for global one-time settings at the start of the pattern, and remember |
| 6150 |
the offset for later. */ |
the offset for later. */ |
| 6151 |
|
|
| 6152 |
while (ptr[skipatstart] == '(' && ptr[skipatstart+1] == '*') |
while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS && |
| 6153 |
|
ptr[skipatstart+1] == CHAR_ASTERISK) |
| 6154 |
{ |
{ |
| 6155 |
int newnl = 0; |
int newnl = 0; |
| 6156 |
int newbsr = 0; |
int newbsr = 0; |
| 6157 |
|
|
| 6158 |
if (strncmp((char *)(ptr+skipatstart+2), "CR)", 3) == 0) |
if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0) |
| 6159 |
{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; } |
{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; } |
| 6160 |
else if (strncmp((char *)(ptr+skipatstart+2), "LF)", 3) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_LF_RIGHTPAR, 3) == 0) |
| 6161 |
{ skipatstart += 5; newnl = PCRE_NEWLINE_LF; } |
{ skipatstart += 5; newnl = PCRE_NEWLINE_LF; } |
| 6162 |
else if (strncmp((char *)(ptr+skipatstart+2), "CRLF)", 5) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_CRLF_RIGHTPAR, 5) == 0) |
| 6163 |
{ skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; } |
{ skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; } |
| 6164 |
else if (strncmp((char *)(ptr+skipatstart+2), "ANY)", 4) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_ANY_RIGHTPAR, 4) == 0) |
| 6165 |
{ skipatstart += 6; newnl = PCRE_NEWLINE_ANY; } |
{ skipatstart += 6; newnl = PCRE_NEWLINE_ANY; } |
| 6166 |
else if (strncmp((char *)(ptr+skipatstart+2), "ANYCRLF)", 8) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_ANYCRLF_RIGHTPAR, 8) == 0) |
| 6167 |
{ skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; } |
{ skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; } |
| 6168 |
|
|
| 6169 |
else if (strncmp((char *)(ptr+skipatstart+2), "BSR_ANYCRLF)", 12) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_BSR_ANYCRLF_RIGHTPAR, 12) == 0) |
| 6170 |
{ skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; } |
{ skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; } |
| 6171 |
else if (strncmp((char *)(ptr+skipatstart+2), "BSR_UNICODE)", 12) == 0) |
else if (strncmp((char *)(ptr+skipatstart+2), STRING_BSR_UNICODE_RIGHTPAR, 12) == 0) |
| 6172 |
{ skipatstart += 14; newbsr = PCRE_BSR_UNICODE; } |
{ skipatstart += 14; newbsr = PCRE_BSR_UNICODE; } |
| 6173 |
|
|
| 6174 |
if (newnl != 0) |
if (newnl != 0) |
| 6196 |
switch (options & PCRE_NEWLINE_BITS) |
switch (options & PCRE_NEWLINE_BITS) |
| 6197 |
{ |
{ |
| 6198 |
case 0: newline = NEWLINE; break; /* Build-time default */ |
case 0: newline = NEWLINE; break; /* Build-time default */ |
| 6199 |
case PCRE_NEWLINE_CR: newline = '\r'; break; |
case PCRE_NEWLINE_CR: newline = CHAR_CR; break; |
| 6200 |
case PCRE_NEWLINE_LF: newline = '\n'; break; |
case PCRE_NEWLINE_LF: newline = CHAR_NL; break; |
| 6201 |
case PCRE_NEWLINE_CR+ |
case PCRE_NEWLINE_CR+ |
| 6202 |
PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; |
PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; |
| 6203 |
case PCRE_NEWLINE_ANY: newline = -1; break; |
case PCRE_NEWLINE_ANY: newline = -1; break; |
| 6204 |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
| 6205 |
default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; |
default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; |