embedaddon/pcre/pcre_compile.c - diff

Return to pcre_compile.c CVS log

Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / pcre

Diff for /embedaddon/pcre/pcre_compile.c between versions 1.1.1.2 and 1.1.1.4

version 1.1.1.2, 2012/02/21 23:50:25	version 1.1.1.4, 2013/07/22 08:25:55
Line 6	Line 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.

Written by Philip Hazel	Written by Philip Hazel
Copyright (c) 1997-2012 University of Cambridge	Copyright (c) 1997-2013 University of Cambridge

-----------------------------------------------------------------------------	-----------------------------------------------------------------------------
Redistribution and use in source and binary forms, with or without	Redistribution and use in source and binary forms, with or without
Line 53 supporting internal functions that are not used by oth	Line 53 supporting internal functions that are not used by oth
#include "pcre_internal.h"	#include "pcre_internal.h"


/* When PCRE_DEBUG is defined, we need the pcre(16)_printint() function, which	/* When PCRE_DEBUG is defined, we need the pcre(16\|32)_printint() function, which
is also used by pcretest. PCRE_DEBUG is not defined when building a production	is also used by pcretest. PCRE_DEBUG is not defined when building a production
library. We do not need to select pcre16_printint.c specially, because the	library. We do not need to select pcre16_printint.c specially, because the
COMPILE_PCREx macro will already be appropriately set. */	COMPILE_PCREx macro will already be appropriately set. */
Line 68 COMPILE_PCREx macro will already be appropriately set.	Line 68 COMPILE_PCREx macro will already be appropriately set.

/* Macro for setting individual bits in class bitmaps. */	/* Macro for setting individual bits in class bitmaps. */

#define SETBIT(a,b) a[b/8] \|= (1 << (b%8))	#define SETBIT(a,b) a[(b)/8] \|= (1 << ((b)&7))

/* Maximum length value to check against when making sure that the integer that	/* Maximum length value to check against when making sure that the integer that
holds the compiled pattern length does not overflow. We make it a bit less than	holds the compiled pattern length does not overflow. We make it a bit less than
Line 77 to check them every time. */	Line 77 to check them every time. */

#define OFLOW_MAX (INT_MAX - 20)	#define OFLOW_MAX (INT_MAX - 20)

	/* Definitions to allow mutual recursion */

	static int
	add_list_to_class(pcre_uint8 , pcre_uchar , int, compile_data ,
	const pcre_uint32 *, unsigned int);

	static BOOL
	compile_regex(int, pcre_uchar , const pcre_uchar , int *, BOOL, BOOL, int, int,
	pcre_uint32 , pcre_int32 , pcre_uint32 , pcre_int32 , branch_chain *,
	compile_data , int );



/*************************************************	/*************************************************
* Code parameters and static tables *	* Code parameters and static tables *
*************************************************/	*************************************************/
Line 110 overrun before it actually does run off the end of the	Line 122 overrun before it actually does run off the end of the

/* Private flags added to firstchar and reqchar. */	/* Private flags added to firstchar and reqchar. */

#define REQ_CASELESS 0x10000000l /* Indicates caselessness */	#define REQ_CASELESS (1 << 0) /* Indicates caselessness */
#define REQ_VARY 0x20000000l /* Reqchar followed non-literal item */	#define REQ_VARY (1 << 1) /* Reqchar followed non-literal item */
	/* Negative values for the firstchar and reqchar flags */
	#define REQ_UNSET (-2)
	#define REQ_NONE (-1)

/* Repeated character flags. */	/* Repeated character flags. */

Line 472 static const char error_texts[] =	Line 487 static const char error_texts[] =
"a numbered reference must not be zero\0"	"a numbered reference must not be zero\0"
"an argument is not allowed for (ACCEPT), (FAIL), or (*COMMIT)\0"	"an argument is not allowed for (ACCEPT), (FAIL), or (*COMMIT)\0"
/* 60 */	/* 60 */
"(*VERB) not recognized\0"	"(*VERB) not recognized or malformed\0"
"number is too big\0"	"number is too big\0"
"subpattern name expected\0"	"subpattern name expected\0"
"digit expected after (?+\0"	"digit expected after (?+\0"
Line 489 static const char error_texts[] =	Line 504 static const char error_texts[] =
"too many forward references\0"	"too many forward references\0"
"disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"	"disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"
"invalid UTF-16 string\0"	"invalid UTF-16 string\0"
	/* 75 */
	"name is too long in (MARK), (PRUNE), (SKIP), or (THEN)\0"
	"character value in \\u.... sequence is too large\0"
	"invalid UTF-32 string\0"
	"setting UTF is disabled by the application\0"
;	;

/* Table to identify digits and hex digits. This is used when compiling	/* Table to identify digits and hex digits. This is used when compiling
Line 628 static const pcre_uint8 ebcdic_chartab[] = { /* charta	Line 648 static const pcre_uint8 ebcdic_chartab[] = { /* charta
#endif	#endif


/* Definition to allow mutual recursion */

static BOOL
compile_regex(int, pcre_uchar , const pcre_uchar , int *, BOOL, BOOL, int, int,
int , int , branch_chain , compile_data , int *);



/*************************************************	/*************************************************
* Find an error text *	* Find an error text *
*************************************************/	*************************************************/
Line 655 find_error_text(int n)	Line 668 find_error_text(int n)
const char *s = error_texts;	const char *s = error_texts;
for (; n > 0; n--)	for (; n > 0; n--)
{	{
while (*s++ != 0) {};	while (*s++ != CHAR_NULL) {};
if (*s == 0) return "Error text not found (please report)";	if (*s == CHAR_NULL) return "Error text not found (please report)";
}	}
return s;	return s;
}	}
Line 739 return (*p == CHAR_RIGHT_CURLY_BRACKET);	Line 752 return (*p == CHAR_RIGHT_CURLY_BRACKET);
*************************************************/	*************************************************/

/* This function is called when a \ has been encountered. It either returns a	/* This function is called when a \ has been encountered. It either returns a
positive value for a simple escape such as \n, or a negative value which	positive value for a simple escape such as \n, or 0 for a data character
encodes one of the more complicated things such as \d. A backreference to group	which will be placed in chptr. A backreference to group n is returned as
n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When	negative n. When UTF-8 is enabled, a positive value greater than 255 may
UTF-8 is enabled, a positive value greater than 255 may be returned. On entry,	be returned in chptr.
ptr is pointing at the \. On exit, it is on the final character of the escape	On entry,ptr is pointing at the \. On exit, it is on the final character of the
sequence.	escape sequence.

Arguments:	Arguments:
ptrptr points to the pattern position pointer	ptrptr points to the pattern position pointer
	chptr points to the data character
errorcodeptr points to the errorcode variable	errorcodeptr points to the errorcode variable
bracount number of previous extracting brackets	bracount number of previous extracting brackets
options the options bits	options the options bits
isclass TRUE if inside a character class	isclass TRUE if inside a character class

Returns: zero or positive => a data character	Returns: zero => a data character
negative => a special escape sequence	positive => a special escape sequence
	negative => a back reference
on error, errorcodeptr is set	on error, errorcodeptr is set
*/	*/

static int	static int
check_escape(const pcre_uchar *ptrptr, int errorcodeptr, int bracount,	check_escape(const pcre_uchar *ptrptr, pcre_uint32 chptr, int *errorcodeptr,
int options, BOOL isclass)	int bracount, int options, BOOL isclass)
{	{
/* PCRE_UTF16 has the same value as PCRE_UTF8. */	/* PCRE_UTF16 has the same value as PCRE_UTF8. */
BOOL utf = (options & PCRE_UTF8) != 0;	BOOL utf = (options & PCRE_UTF8) != 0;
const pcre_uchar ptr = ptrptr + 1;	const pcre_uchar ptr = ptrptr + 1;
pcre_int32 c;	pcre_uint32 c;
	int escape = 0;
int i;	int i;

GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */	GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */
Line 773 ptr--; /* Set pointer back	Line 789 ptr--; /* Set pointer back

/* If backslash is at the end of the pattern, it's an error. */	/* If backslash is at the end of the pattern, it's an error. */

if (c == 0) *errorcodeptr = ERR1;	if (c == CHAR_NULL) *errorcodeptr = ERR1;

/* Non-alphanumerics are literals. For digits or letters, do an initial lookup	/* Non-alphanumerics are literals. For digits or letters, do an initial lookup
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.
Line 782 Otherwise further processing may be required. */	Line 798 Otherwise further processing may be required. */
#ifndef EBCDIC /* ASCII/UTF-8 coding */	#ifndef EBCDIC /* ASCII/UTF-8 coding */
/* Not alphanumeric */	/* Not alphanumeric */
else if (c < CHAR_0 \|\| c > CHAR_z) {}	else if (c < CHAR_0 \|\| c > CHAR_z) {}
else if ((i = escapes[c - CHAR_0]) != 0) c = i;	else if ((i = escapes[c - CHAR_0]) != 0)
	{ if (i > 0) c = (pcre_uint32)i; else escape = -i; }

#else /* EBCDIC coding */	#else /* EBCDIC coding */
/* Not alphanumeric */	/* Not alphanumeric */
else if (c < 'a' \|\| (!MAX_255(c) \|\| (ebcdic_chartab[c] & 0x0E) == 0)) {}	else if (c < CHAR_a \|\| (!MAX_255(c) \|\| (ebcdic_chartab[c] & 0x0E) == 0)) {}
else if ((i = escapes[c - 0x48]) != 0) c = i;	else if ((i = escapes[c - 0x48]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
#endif	#endif

/* Escapes that need further processing, or are illegal. */	/* Escapes that need further processing, or are illegal. */
Line 795 else if ((i = escapes[c - 0x48]) != 0) c = i;	Line 812 else if ((i = escapes[c - 0x48]) != 0) c = i;
else	else
{	{
const pcre_uchar *oldptr;	const pcre_uchar *oldptr;
BOOL braced, negated;	BOOL braced, negated, overflow;
	int s;

switch (c)	switch (c)
{	{
Line 820 else	Line 838 else
c = 0;	c = 0;
for (i = 0; i < 4; ++i)	for (i = 0; i < 4; ++i)
{	{
register int cc = *(++ptr);	register pcre_uint32 cc = *(++ptr);
#ifndef EBCDIC /* ASCII/UTF-8 coding */	#ifndef EBCDIC /* ASCII/UTF-8 coding */
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */	if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));	c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
Line 829 else	Line 847 else
c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));	c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
#endif	#endif
}	}

	#if defined COMPILE_PCRE8
	if (c > (utf ? 0x10ffffU : 0xffU))
	#elif defined COMPILE_PCRE16
	if (c > (utf ? 0x10ffffU : 0xffffU))
	#elif defined COMPILE_PCRE32
	if (utf && c > 0x10ffffU)
	#endif
	{
	*errorcodeptr = ERR76;
	}
	else if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73;
}	}
}	}
else	else
Line 855 else	Line 885 else
(3) For Oniguruma compatibility we also support \g followed by a name or a	(3) For Oniguruma compatibility we also support \g followed by a name or a
number either in angle brackets or in single quotes. However, these are	number either in angle brackets or in single quotes. However, these are
(possibly recursive) subroutine calls, _not_ backreferences. Just return	(possibly recursive) subroutine calls, _not_ backreferences. Just return
the -ESC_g code (cf \k). */	the ESC_g code (cf \k). */

case CHAR_g:	case CHAR_g:
if (isclass) break;	if (isclass) break;
if (ptr[1] == CHAR_LESS_THAN_SIGN \|\| ptr[1] == CHAR_APOSTROPHE)	if (ptr[1] == CHAR_LESS_THAN_SIGN \|\| ptr[1] == CHAR_APOSTROPHE)
{	{
c = -ESC_g;	escape = ESC_g;
break;	break;
}	}

Line 870 else	Line 900 else
if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)	if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
{	{
const pcre_uchar *p;	const pcre_uchar *p;
for (p = ptr+2; p != 0 && p != CHAR_RIGHT_CURLY_BRACKET; p++)	for (p = ptr+2; p != CHAR_NULL && p != CHAR_RIGHT_CURLY_BRACKET; p++)
if (p != CHAR_MINUS && !IS_DIGIT(p)) break;	if (p != CHAR_MINUS && !IS_DIGIT(p)) break;
if (p != 0 && p != CHAR_RIGHT_CURLY_BRACKET)	if (p != CHAR_NULL && p != CHAR_RIGHT_CURLY_BRACKET)
{	{
c = -ESC_k;	escape = ESC_k;
break;	break;
}	}
braced = TRUE;	braced = TRUE;
Line 890 else	Line 920 else
else negated = FALSE;	else negated = FALSE;

/* The integer range is limited by the machine's int representation. */	/* The integer range is limited by the machine's int representation. */
c = 0;	s = 0;
	overflow = FALSE;
while (IS_DIGIT(ptr[1]))	while (IS_DIGIT(ptr[1]))
{	{
if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */	if (s > INT_MAX / 10 - 1) /* Integer overflow */
{	{
c = -1;	overflow = TRUE;
break;	break;
}	}
c = c * 10 + *(++ptr) - CHAR_0;	s = s * 10 + (int)(*(++ptr) - CHAR_0);
}	}
if (((unsigned int)c) > INT_MAX) /* Integer overflow */	if (overflow) /* Integer overflow */
{	{
while (IS_DIGIT(ptr[1]))	while (IS_DIGIT(ptr[1]))
ptr++;	ptr++;
Line 914 else	Line 945 else
break;	break;
}	}

if (c == 0)	if (s == 0)
{	{
*errorcodeptr = ERR58;	*errorcodeptr = ERR58;
break;	break;
Line 922 else	Line 953 else

if (negated)	if (negated)
{	{
if (c > bracount)	if (s > bracount)
{	{
*errorcodeptr = ERR15;	*errorcodeptr = ERR15;
break;	break;
}	}
c = bracount - (c - 1);	s = bracount - (s - 1);
}	}

c = -(ESC_REF + c);	escape = -s;
break;	break;

/* The handling of escape sequences consisting of a string of digits	/* The handling of escape sequences consisting of a string of digits
Line 952 else	Line 983 else
{	{
oldptr = ptr;	oldptr = ptr;
/* The integer range is limited by the machine's int representation. */	/* The integer range is limited by the machine's int representation. */
c -= CHAR_0;	s = (int)(c -CHAR_0);
	overflow = FALSE;
while (IS_DIGIT(ptr[1]))	while (IS_DIGIT(ptr[1]))
{	{
if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */	if (s > INT_MAX / 10 - 1) /* Integer overflow */
{	{
c = -1;	overflow = TRUE;
break;	break;
}	}
c = c * 10 + *(++ptr) - CHAR_0;	s = s * 10 + (int)(*(++ptr) - CHAR_0);
}	}
if (((unsigned int)c) > INT_MAX) /* Integer overflow */	if (overflow) /* Integer overflow */
{	{
while (IS_DIGIT(ptr[1]))	while (IS_DIGIT(ptr[1]))
ptr++;	ptr++;
*errorcodeptr = ERR61;	*errorcodeptr = ERR61;
break;	break;
}	}
if (c < 10 \|\| c <= bracount)	if (s < 10 \|\| s <= bracount)
{	{
c = -(ESC_REF + c);	escape = -s;
break;	break;
}	}
ptr = oldptr; /* Put the pointer back and fall through */	ptr = oldptr; /* Put the pointer back and fall through */
Line 1018 else	Line 1050 else
c = 0;	c = 0;
for (i = 0; i < 2; ++i)	for (i = 0; i < 2; ++i)
{	{
register int cc = *(++ptr);	register pcre_uint32 cc = *(++ptr);
#ifndef EBCDIC /* ASCII/UTF-8 coding */	#ifndef EBCDIC /* ASCII/UTF-8 coding */
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */	if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));	c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
Line 1036 else	Line 1068 else
const pcre_uchar *pt = ptr + 2;	const pcre_uchar *pt = ptr + 2;

c = 0;	c = 0;
	overflow = FALSE;
while (MAX_255(pt) && (digitab[pt] & ctype_xdigit) != 0)	while (MAX_255(pt) && (digitab[pt] & ctype_xdigit) != 0)
{	{
register int cc = *pt++;	register pcre_uint32 cc = *pt++;
if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */	if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */

	#ifdef COMPILE_PCRE32
	if (c >= 0x10000000l) { overflow = TRUE; break; }
	#endif

#ifndef EBCDIC /* ASCII/UTF-8 coding */	#ifndef EBCDIC /* ASCII/UTF-8 coding */
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */	if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));	c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
Line 1049 else	Line 1086 else
c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));	c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
#endif	#endif

#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }	if (c > (utf ? 0x10ffffU : 0xffU)) { overflow = TRUE; break; }
#else	#elif defined COMPILE_PCRE16
#ifdef COMPILE_PCRE16	if (c > (utf ? 0x10ffffU : 0xffffU)) { overflow = TRUE; break; }
if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }	#elif defined COMPILE_PCRE32
	if (utf && c > 0x10ffffU) { overflow = TRUE; break; }
#endif	#endif
#endif
}	}

if (c < 0)	if (overflow)
{	{
while (MAX_255(pt) && (digitab[pt] & ctype_xdigit) != 0) pt++;	while (MAX_255(pt) && (digitab[pt] & ctype_xdigit) != 0) pt++;
*errorcodeptr = ERR34;	*errorcodeptr = ERR34;
Line 1080 else	Line 1117 else
c = 0;	c = 0;
while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)	while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
{	{
int cc; /* Some compilers don't like */	pcre_uint32 cc; /* Some compilers don't like */
cc = (++ptr); / ++ in initializers */	cc = (++ptr); / ++ in initializers */
#ifndef EBCDIC /* ASCII/UTF-8 coding */	#ifndef EBCDIC /* ASCII/UTF-8 coding */
if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */	if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
Line 1099 else	Line 1136 else

case CHAR_c:	case CHAR_c:
c = *(++ptr);	c = *(++ptr);
if (c == 0)	if (c == CHAR_NULL)
{	{
*errorcodeptr = ERR2;	*errorcodeptr = ERR2;
break;	break;
Line 1139 else	Line 1176 else
newline". PCRE does not support \N{name}. However, it does support	newline". PCRE does not support \N{name}. However, it does support
quantification such as \N{2,3}. */	quantification such as \N{2,3}. */

if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&	if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
!is_counted_repeat(ptr+2))	!is_counted_repeat(ptr+2))
*errorcodeptr = ERR37;	*errorcodeptr = ERR37;

/* If PCRE_UCP is set, we change the values for \d etc. */	/* If PCRE_UCP is set, we change the values for \d etc. */

if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)	if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w)
c -= (ESC_DU - ESC_D);	escape += (ESC_DU - ESC_D);

/* Set the pointer to the final character before returning. */	/* Set the pointer to the final character before returning. */

*ptrptr = ptr;	*ptrptr = ptr;
return c;	*chptr = c;
	return escape;
}	}



#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
/*************************************************	/*************************************************
* Handle \P and \p *	* Handle \P and \p *
Line 1169 escape sequence.	Line 1205 escape sequence.
Argument:	Argument:
ptrptr points to the pattern position pointer	ptrptr points to the pattern position pointer
negptr points to a boolean that is set TRUE for negation else FALSE	negptr points to a boolean that is set TRUE for negation else FALSE
dptr points to an int that is set to the detailed property value	ptypeptr points to an unsigned int that is set to the type value
	pdataptr points to an unsigned int that is set to the detailed property value
errorcodeptr points to the error code variable	errorcodeptr points to the error code variable

Returns: type value from ucp_type_table, or -1 for an invalid type	Returns: TRUE if the type value was found, or FALSE for an invalid type
*/	*/

static int	static BOOL
get_ucp(const pcre_uchar *ptrptr, BOOL negptr, int dptr, int errorcodeptr)	get_ucp(const pcre_uchar *ptrptr, BOOL negptr, unsigned int *ptypeptr,
	unsigned int pdataptr, int errorcodeptr)
{	{
int c, i, bot, top;	pcre_uchar c;
	int i, bot, top;
const pcre_uchar ptr = ptrptr;	const pcre_uchar ptr = ptrptr;
pcre_uchar name[32];	pcre_uchar name[32];

c = *(++ptr);	c = *(++ptr);
if (c == 0) goto ERROR_RETURN;	if (c == CHAR_NULL) goto ERROR_RETURN;

*negptr = FALSE;	*negptr = FALSE;

Line 1200 if (c == CHAR_LEFT_CURLY_BRACKET)	Line 1239 if (c == CHAR_LEFT_CURLY_BRACKET)
for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)	for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)
{	{
c = *(++ptr);	c = *(++ptr);
if (c == 0) goto ERROR_RETURN;	if (c == CHAR_NULL) goto ERROR_RETURN;
if (c == CHAR_RIGHT_CURLY_BRACKET) break;	if (c == CHAR_RIGHT_CURLY_BRACKET) break;
name[i] = c;	name[i] = c;
}	}
Line 1225 top = PRIV(utt_size);	Line 1264 top = PRIV(utt_size);

while (bot < top)	while (bot < top)
{	{
	int r;
i = (bot + top) >> 1;	i = (bot + top) >> 1;
c = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);	r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
if (c == 0)	if (r == 0)
{	{
*dptr = PRIV(utt)[i].value;	*ptypeptr = PRIV(utt)[i].type;
return PRIV(utt)[i].type;	*pdataptr = PRIV(utt)[i].value;
	return TRUE;
}	}
if (c > 0) bot = i + 1; else top = i;	if (r > 0) bot = i + 1; else top = i;
}	}

*errorcodeptr = ERR47;	*errorcodeptr = ERR47;
*ptrptr = ptr;	*ptrptr = ptr;
return -1;	return FALSE;

ERROR_RETURN:	ERROR_RETURN:
*errorcodeptr = ERR46;	*errorcodeptr = ERR46;
*ptrptr = ptr;	*ptrptr = ptr;
return -1;	return FALSE;
}	}
#endif	#endif

Line 1277 int max = -1;	Line 1318 int max = -1;
/* 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
an integer overflow. */	an integer overflow. */

while (IS_DIGIT(p)) min = min 10 + *p++ - CHAR_0;	while (IS_DIGIT(p)) min = min 10 + (int)(*p++ - CHAR_0);
if (min < 0 \|\| min > 65535)	if (min < 0 \|\| min > 65535)
{	{
*errorcodeptr = ERR5;	*errorcodeptr = ERR5;
Line 1292 if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else	Line 1333 if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else
if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)	if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
{	{
max = 0;	max = 0;
while(IS_DIGIT(p)) max = max 10 + *p++ - CHAR_0;	while(IS_DIGIT(p)) max = max 10 + (int)(*p++ - CHAR_0);
if (max < 0 \|\| max > 65535)	if (max < 0 \|\| max > 65535)
{	{
*errorcodeptr = ERR5;	*errorcodeptr = ERR5;
Line 1347 Arguments:	Line 1388 Arguments:
name name to seek, or NULL if seeking a numbered subpattern	name name to seek, or NULL if seeking a numbered subpattern
lorn name length, or subpattern number if name is NULL	lorn name length, or subpattern number if name is NULL
xmode TRUE if we are in /x mode	xmode TRUE if we are in /x mode
utf TRUE if we are in UTF-8 / UTF-16 mode	utf TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
count pointer to the current capturing subpattern number (updated)	count pointer to the current capturing subpattern number (updated)

Returns: the number of the named subpattern, or -1 if not found	Returns: the number of the named subpattern, or -1 if not found
Line 1369 if (ptr[0] == CHAR_LEFT_PARENTHESIS)	Line 1410 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
{	{
/* Handle specials such as (SKIP) or (UTF8) etc. */	/* Handle specials such as (SKIP) or (UTF8) etc. */

if (ptr[1] == CHAR_ASTERISK) ptr += 2;	if (ptr[1] == CHAR_ASTERISK)
	{
	ptr += 2;
	while (ptr < cd->end_pattern && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
	}

/* Handle a normal, unnamed capturing parenthesis. */	/* Handle a normal, unnamed capturing parenthesis. */

Line 1393 if (ptr[0] == CHAR_LEFT_PARENTHESIS)	Line 1438 if (ptr[0] == CHAR_LEFT_PARENTHESIS)

else if (ptr[2] == CHAR_NUMBER_SIGN)	else if (ptr[2] == CHAR_NUMBER_SIGN)
{	{
for (ptr += 3; ptr != 0; ptr++) if (ptr == CHAR_RIGHT_PARENTHESIS) break;	for (ptr += 3; *ptr != CHAR_NULL; ptr++)
	if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
goto FAIL_EXIT;	goto FAIL_EXIT;
}	}

Line 1406 if (ptr[0] == CHAR_LEFT_PARENTHESIS)	Line 1452 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
ptr += 2;	ptr += 2;
if (ptr[1] != CHAR_QUESTION_MARK)	if (ptr[1] != CHAR_QUESTION_MARK)
{	{
while (ptr != 0 && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;	while (ptr != CHAR_NULL && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
if (*ptr != 0) ptr++;	if (*ptr != CHAR_NULL) ptr++;
}	}
}	}

Line 1423 if (ptr[0] == CHAR_LEFT_PARENTHESIS)	Line 1469 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&	if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
ptr[1] != CHAR_EQUALS_SIGN) \|\| *ptr == CHAR_APOSTROPHE)	ptr[1] != CHAR_EQUALS_SIGN) \|\| *ptr == CHAR_APOSTROPHE)
{	{
int term;	pcre_uchar term;
const pcre_uchar *thisname;	const pcre_uchar *thisname;
*count += 1;	*count += 1;
if (name == NULL && count == lorn) return count;	if (name == NULL && count == lorn) return count;
Line 1431 if (ptr[0] == CHAR_LEFT_PARENTHESIS)	Line 1477 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;	if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
thisname = ptr;	thisname = ptr;
while (*ptr != term) ptr++;	while (*ptr != term) ptr++;
if (name != NULL && lorn == ptr - thisname &&	if (name != NULL && lorn == (int)(ptr - thisname) &&
STRNCMP_UC_UC(name, thisname, lorn) == 0)	STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
return *count;	return *count;
term++;	term++;
}	}
Line 1450 for (; ptr < cd->end_pattern; ptr++)	Line 1496 for (; ptr < cd->end_pattern; ptr++)

if (*ptr == CHAR_BACKSLASH)	if (*ptr == CHAR_BACKSLASH)
{	{
if (*(++ptr) == 0) goto FAIL_EXIT;	if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
if (*ptr == CHAR_Q) for (;;)	if (*ptr == CHAR_Q) for (;;)
{	{
while ((++ptr) != 0 && ptr != CHAR_BACKSLASH) {};	while ((++ptr) != CHAR_NULL && ptr != CHAR_BACKSLASH) {};
if (*ptr == 0) goto FAIL_EXIT;	if (*ptr == CHAR_NULL) goto FAIL_EXIT;
if (*(++ptr) == CHAR_E) break;	if (*(++ptr) == CHAR_E) break;
}	}
continue;	continue;
Line 1498 for (; ptr < cd->end_pattern; ptr++)	Line 1544 for (; ptr < cd->end_pattern; ptr++)

while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)	while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
{	{
if (*ptr == 0) return -1;	if (*ptr == CHAR_NULL) return -1;
if (*ptr == CHAR_BACKSLASH)	if (*ptr == CHAR_BACKSLASH)
{	{
if (*(++ptr) == 0) goto FAIL_EXIT;	if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
if (*ptr == CHAR_Q) for (;;)	if (*ptr == CHAR_Q) for (;;)
{	{
while ((++ptr) != 0 && ptr != CHAR_BACKSLASH) {};	while ((++ptr) != CHAR_NULL && ptr != CHAR_BACKSLASH) {};
if (*ptr == 0) goto FAIL_EXIT;	if (*ptr == CHAR_NULL) goto FAIL_EXIT;
if (*(++ptr) == CHAR_E) break;	if (*(++ptr) == CHAR_E) break;
}	}
continue;	continue;
Line 1519 for (; ptr < cd->end_pattern; ptr++)	Line 1565 for (; ptr < cd->end_pattern; ptr++)
if (xmode && *ptr == CHAR_NUMBER_SIGN)	if (xmode && *ptr == CHAR_NUMBER_SIGN)
{	{
ptr++;	ptr++;
while (*ptr != 0)	while (*ptr != CHAR_NULL)
{	{
if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }	if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
ptr++;	ptr++;
Line 1527 for (; ptr < cd->end_pattern; ptr++)	Line 1573 for (; ptr < cd->end_pattern; ptr++)
if (utf) FORWARDCHAR(ptr);	if (utf) FORWARDCHAR(ptr);
#endif	#endif
}	}
if (*ptr == 0) goto FAIL_EXIT;	if (*ptr == CHAR_NULL) goto FAIL_EXIT;
continue;	continue;
}	}

Line 1537 for (; ptr < cd->end_pattern; ptr++)	Line 1583 for (; ptr < cd->end_pattern; ptr++)
{	{
int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);	int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
if (rc > 0) return rc;	if (rc > 0) return rc;
if (*ptr == 0) goto FAIL_EXIT;	if (*ptr == CHAR_NULL) goto FAIL_EXIT;
}	}

else if (*ptr == CHAR_RIGHT_PARENTHESIS)	else if (*ptr == CHAR_RIGHT_PARENTHESIS)
Line 1581 Arguments:	Line 1627 Arguments:
name name to seek, or NULL if seeking a numbered subpattern	name name to seek, or NULL if seeking a numbered subpattern
lorn name length, or subpattern number if name is NULL	lorn name length, or subpattern number if name is NULL
xmode TRUE if we are in /x mode	xmode TRUE if we are in /x mode
utf TRUE if we are in UTF-8 / UTF-16 mode	utf TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode

Returns: the number of the found subpattern, or -1 if not found	Returns: the number of the found subpattern, or -1 if not found
*/	*/
Line 1602 matching closing parens. That is why we have to have a	Line 1648 matching closing parens. That is why we have to have a
for (;;)	for (;;)
{	{
rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);	rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
if (rc > 0 \|\| *ptr++ == 0) break;	if (rc > 0 \|\| *ptr++ == CHAR_NULL) break;
}	}

return rc;	return rc;
Line 1684 and doing the check at the end; a flag specifies which	Line 1730 and doing the check at the end; a flag specifies which

Arguments:	Arguments:
code points to the start of the pattern (the bracket)	code points to the start of the pattern (the bracket)
utf TRUE in UTF-8 / UTF-16 mode	utf TRUE in UTF-8 / UTF-16 / UTF-32 mode
atend TRUE if called when the pattern is complete	atend TRUE if called when the pattern is complete
cd the "compile data" structure	cd the "compile data" structure

Line 1710 for (;;)	Line 1756 for (;;)
{	{
int d;	int d;
pcre_uchar ce, cs;	pcre_uchar ce, cs;
register int op = *cc;	register pcre_uchar op = *cc;

switch (op)	switch (op)
{	{
Line 1830 for (;;)	Line 1876 for (;;)
case OP_EXACTI:	case OP_EXACTI:
case OP_NOTEXACT:	case OP_NOTEXACT:
case OP_NOTEXACTI:	case OP_NOTEXACTI:
branchlength += GET2(cc,1);	branchlength += (int)GET2(cc,1);
cc += 2 + IMM2_SIZE;	cc += 2 + IMM2_SIZE;
#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);	if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
Line 1839 for (;;)	Line 1885 for (;;)

case OP_TYPEEXACT:	case OP_TYPEEXACT:
branchlength += GET2(cc,1);	branchlength += GET2(cc,1);
if (cc[1 + IMM2_SIZE] == OP_PROP \|\| cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2;	if (cc[1 + IMM2_SIZE] == OP_PROP \|\| cc[1 + IMM2_SIZE] == OP_NOTPROP)
	cc += 2;
cc += 1 + IMM2_SIZE + 1;	cc += 1 + IMM2_SIZE + 1;
break;	break;

Line 1874 for (;;)	Line 1921 for (;;)

/* Check a class for variable quantification */	/* Check a class for variable quantification */

#if defined SUPPORT_UTF \|\| defined COMPILE_PCRE16
case OP_XCLASS:
cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
/* Fall through */
#endif

case OP_CLASS:	case OP_CLASS:
case OP_NCLASS:	case OP_NCLASS:
	#if defined SUPPORT_UTF \|\| defined COMPILE_PCRE16 \|\| defined COMPILE_PCRE32
	case OP_XCLASS:
	/* The original code caused an unsigned overflow in 64 bit systems,
	so now we use a conditional statement. */
	if (op == OP_XCLASS)
	cc += GET(cc, 1);
	else
	cc += PRIV(OP_lengths)[OP_CLASS];
	#else
cc += PRIV(OP_lengths)[OP_CLASS];	cc += PRIV(OP_lengths)[OP_CLASS];
	#endif

switch (*cc)	switch (*cc)
{	{
Line 1897 for (;;)	Line 1948 for (;;)
case OP_CRRANGE:	case OP_CRRANGE:
case OP_CRMINRANGE:	case OP_CRMINRANGE:
if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;	if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
branchlength += GET2(cc,1);	branchlength += (int)GET2(cc,1);
cc += 1 + 2 * IMM2_SIZE;	cc += 1 + 2 * IMM2_SIZE;
break;	break;

Line 2013 length.	Line 2064 length.

Arguments:	Arguments:
code points to start of expression	code points to start of expression
utf TRUE in UTF-8 / UTF-16 mode	utf TRUE in UTF-8 / UTF-16 / UTF-32 mode
number the required bracket number or negative to find a lookbehind	number the required bracket number or negative to find a lookbehind

Returns: pointer to the opcode for the bracket, or NULL if not found	Returns: pointer to the opcode for the bracket, or NULL if not found
Line 2024 PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, i	Line 2075 PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, i
{	{
for (;;)	for (;;)
{	{
register int c = *code;	register pcre_uchar c = *code;

if (c == OP_END) return NULL;	if (c == OP_END) return NULL;

Line 2047 for (;;)	Line 2098 for (;;)
else if (c == OP_CBRA \|\| c == OP_SCBRA \|\|	else if (c == OP_CBRA \|\| c == OP_SCBRA \|\|
c == OP_CBRAPOS \|\| c == OP_SCBRAPOS)	c == OP_CBRAPOS \|\| c == OP_SCBRAPOS)
{	{
int n = GET2(code, 1+LINK_SIZE);	int n = (int)GET2(code, 1+LINK_SIZE);
if (n == number) return (pcre_uchar *)code;	if (n == number) return (pcre_uchar *)code;
code += PRIV(OP_lengths)[c];	code += PRIV(OP_lengths)[c];
}	}
Line 2077 for (;;)	Line 2128 for (;;)
case OP_TYPEMINUPTO:	case OP_TYPEMINUPTO:
case OP_TYPEEXACT:	case OP_TYPEEXACT:
case OP_TYPEPOSUPTO:	case OP_TYPEPOSUPTO:
if (code[1 + IMM2_SIZE] == OP_PROP	if (code[1 + IMM2_SIZE] == OP_PROP \|\| code[1 + IMM2_SIZE] == OP_NOTPROP)
\|\| code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;	code += 2;
break;	break;

case OP_MARK:	case OP_MARK:
case OP_PRUNE_ARG:	case OP_PRUNE_ARG:
case OP_SKIP_ARG:	case OP_SKIP_ARG:
code += code[1];
break;

case OP_THEN_ARG:	case OP_THEN_ARG:
code += code[1];	code += code[1];
break;	break;
Line 2100 for (;;)	Line 2148 for (;;)
a multi-byte character. The length in the table is a minimum, so we have to	a multi-byte character. The length in the table is a minimum, so we have to
arrange to skip the extra bytes. */	arrange to skip the extra bytes. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf) switch(c)	if (utf) switch(c)
{	{
case OP_CHAR:	case OP_CHAR:
Line 2152 instance of OP_RECURSE.	Line 2200 instance of OP_RECURSE.

Arguments:	Arguments:
code points to start of expression	code points to start of expression
utf TRUE in UTF-8 / UTF-16 mode	utf TRUE in UTF-8 / UTF-16 / UTF-32 mode

Returns: pointer to the opcode for OP_RECURSE, or NULL if not found	Returns: pointer to the opcode for OP_RECURSE, or NULL if not found
*/	*/
Line 2162 find_recurse(const pcre_uchar *code, BOOL utf)	Line 2210 find_recurse(const pcre_uchar *code, BOOL utf)
{	{
for (;;)	for (;;)
{	{
register int c = *code;	register pcre_uchar c = *code;
if (c == OP_END) return NULL;	if (c == OP_END) return NULL;
if (c == OP_RECURSE) return code;	if (c == OP_RECURSE) return code;

Line 2197 for (;;)	Line 2245 for (;;)
case OP_TYPEUPTO:	case OP_TYPEUPTO:
case OP_TYPEMINUPTO:	case OP_TYPEMINUPTO:
case OP_TYPEEXACT:	case OP_TYPEEXACT:
if (code[1 + IMM2_SIZE] == OP_PROP	if (code[1 + IMM2_SIZE] == OP_PROP \|\| code[1 + IMM2_SIZE] == OP_NOTPROP)
\|\| code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;	code += 2;
break;	break;

case OP_MARK:	case OP_MARK:
case OP_PRUNE_ARG:	case OP_PRUNE_ARG:
case OP_SKIP_ARG:	case OP_SKIP_ARG:
code += code[1];
break;

case OP_THEN_ARG:	case OP_THEN_ARG:
code += code[1];	code += code[1];
break;	break;
Line 2220 for (;;)	Line 2265 for (;;)
by a multi-byte character. The length in the table is a minimum, so we have	by a multi-byte character. The length in the table is a minimum, so we have
to arrange to skip the extra bytes. */	to arrange to skip the extra bytes. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf) switch(c)	if (utf) switch(c)
{	{
case OP_CHAR:	case OP_CHAR:
case OP_CHARI:	case OP_CHARI:
	case OP_NOT:
	case OP_NOTI:
case OP_EXACT:	case OP_EXACT:
case OP_EXACTI:	case OP_EXACTI:
	case OP_NOTEXACT:
	case OP_NOTEXACTI:
case OP_UPTO:	case OP_UPTO:
case OP_UPTOI:	case OP_UPTOI:
	case OP_NOTUPTO:
	case OP_NOTUPTOI:
case OP_MINUPTO:	case OP_MINUPTO:
case OP_MINUPTOI:	case OP_MINUPTOI:
	case OP_NOTMINUPTO:
	case OP_NOTMINUPTOI:
case OP_POSUPTO:	case OP_POSUPTO:
case OP_POSUPTOI:	case OP_POSUPTOI:
	case OP_NOTPOSUPTO:
	case OP_NOTPOSUPTOI:
case OP_STAR:	case OP_STAR:
case OP_STARI:	case OP_STARI:
	case OP_NOTSTAR:
	case OP_NOTSTARI:
case OP_MINSTAR:	case OP_MINSTAR:
case OP_MINSTARI:	case OP_MINSTARI:
	case OP_NOTMINSTAR:
	case OP_NOTMINSTARI:
case OP_POSSTAR:	case OP_POSSTAR:
case OP_POSSTARI:	case OP_POSSTARI:
	case OP_NOTPOSSTAR:
	case OP_NOTPOSSTARI:
case OP_PLUS:	case OP_PLUS:
case OP_PLUSI:	case OP_PLUSI:
	case OP_NOTPLUS:
	case OP_NOTPLUSI:
case OP_MINPLUS:	case OP_MINPLUS:
case OP_MINPLUSI:	case OP_MINPLUSI:
	case OP_NOTMINPLUS:
	case OP_NOTMINPLUSI:
case OP_POSPLUS:	case OP_POSPLUS:
case OP_POSPLUSI:	case OP_POSPLUSI:
	case OP_NOTPOSPLUS:
	case OP_NOTPOSPLUSI:
case OP_QUERY:	case OP_QUERY:
case OP_QUERYI:	case OP_QUERYI:
	case OP_NOTQUERY:
	case OP_NOTQUERYI:
case OP_MINQUERY:	case OP_MINQUERY:
case OP_MINQUERYI:	case OP_MINQUERYI:
	case OP_NOTMINQUERY:
	case OP_NOTMINQUERYI:
case OP_POSQUERY:	case OP_POSQUERY:
case OP_POSQUERYI:	case OP_POSQUERYI:
	case OP_NOTPOSQUERY:
	case OP_NOTPOSQUERYI:
if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);	if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);
break;	break;
}	}
Line 2278 bracket whose current branch will already have been sc	Line 2351 bracket whose current branch will already have been sc
Arguments:	Arguments:
code points to start of search	code points to start of search
endcode points to where to stop	endcode points to where to stop
utf TRUE if in UTF-8 / UTF-16 mode	utf TRUE if in UTF-8 / UTF-16 / UTF-32 mode
cd contains pointers to tables etc.	cd contains pointers to tables etc.

Returns: TRUE if what is matched could be empty	Returns: TRUE if what is matched could be empty
Line 2288 static BOOL	Line 2361 static BOOL
could_be_empty_branch(const pcre_uchar code, const pcre_uchar endcode,	could_be_empty_branch(const pcre_uchar code, const pcre_uchar endcode,
BOOL utf, compile_data *cd)	BOOL utf, compile_data *cd)
{	{
register int c;	register pcre_uchar c;
for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);	for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
code < endcode;	code < endcode;
code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))	code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
Line 2322 for (code = first_significant_code(code + PRIV(OP_leng	Line 2395 for (code = first_significant_code(code + PRIV(OP_leng
/* Test for forward reference */	/* Test for forward reference */

for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)	for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;	if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;

/* Not a forward reference, test for completed backward reference */	/* Not a forward reference, test for completed backward reference */

Line 2495 for (code = first_significant_code(code + PRIV(OP_leng	Line 2568 for (code = first_significant_code(code + PRIV(OP_leng
case OP_TYPEUPTO:	case OP_TYPEUPTO:
case OP_TYPEMINUPTO:	case OP_TYPEMINUPTO:
case OP_TYPEPOSUPTO:	case OP_TYPEPOSUPTO:
if (code[1 + IMM2_SIZE] == OP_PROP	if (code[1 + IMM2_SIZE] == OP_PROP \|\| code[1 + IMM2_SIZE] == OP_NOTPROP)
\|\| code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;	code += 2;
break;	break;

/* End of branch */	/* End of branch */
Line 2511 for (code = first_significant_code(code + PRIV(OP_leng	Line 2584 for (code = first_significant_code(code + PRIV(OP_leng
/* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,	/* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
MINUPTO, and POSUPTO may be followed by a multibyte character */	MINUPTO, and POSUPTO may be followed by a multibyte character */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
case OP_STAR:	case OP_STAR:
case OP_STARI:	case OP_STARI:
case OP_MINSTAR:	case OP_MINSTAR:
Line 2543 for (code = first_significant_code(code + PRIV(OP_leng	Line 2616 for (code = first_significant_code(code + PRIV(OP_leng
case OP_MARK:	case OP_MARK:
case OP_PRUNE_ARG:	case OP_PRUNE_ARG:
case OP_SKIP_ARG:	case OP_SKIP_ARG:
code += code[1];
break;

case OP_THEN_ARG:	case OP_THEN_ARG:
code += code[1];	code += code[1];
break;	break;
Line 2577 Arguments:	Line 2647 Arguments:
code points to start of the recursion	code points to start of the recursion
endcode points to where to stop (current RECURSE item)	endcode points to where to stop (current RECURSE item)
bcptr points to the chain of current (unclosed) branch starts	bcptr points to the chain of current (unclosed) branch starts
utf TRUE if in UTF-8 / UTF-16 mode	utf TRUE if in UTF-8 / UTF-16 / UTF-32 mode
cd pointers to tables etc	cd pointers to tables etc

Returns: TRUE if what is matched could be empty	Returns: TRUE if what is matched could be empty
Line 2643 Returns: TRUE or FALSE	Line 2713 Returns: TRUE or FALSE
static BOOL	static BOOL
check_posix_syntax(const pcre_uchar ptr, const pcre_uchar *endptr)	check_posix_syntax(const pcre_uchar ptr, const pcre_uchar *endptr)
{	{
int terminator; /* Don't combine these lines; the Solaris cc */	pcre_uchar terminator; /* Don't combine these lines; the Solaris cc */
terminator = (++ptr); / compiler warns about "non-constant" initializer. */	terminator = (++ptr); / compiler warns about "non-constant" initializer. */
for (++ptr; *ptr != 0; ptr++)	for (++ptr; *ptr != CHAR_NULL; ptr++)
{	{
if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)	if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
ptr++;	ptr++;
Line 2692 register int yield = 0;	Line 2762 register int yield = 0;
while (posix_name_lengths[yield] != 0)	while (posix_name_lengths[yield] != 0)
{	{
if (len == posix_name_lengths[yield] &&	if (len == posix_name_lengths[yield] &&
STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;	STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
pn += posix_name_lengths[yield] + 1;	pn += posix_name_lengths[yield] + 1;
yield++;	yield++;
}	}
Line 2724 value in the reference (which is a group number).	Line 2794 value in the reference (which is a group number).
Arguments:	Arguments:
group points to the start of the group	group points to the start of the group
adjust the amount by which the group is to be moved	adjust the amount by which the group is to be moved
utf TRUE in UTF-8 / UTF-16 mode	utf TRUE in UTF-8 / UTF-16 / UTF-32 mode
cd contains pointers to tables etc.	cd contains pointers to tables etc.
save_hwm the hwm forward reference pointer at the start of the group	save_hwm the hwm forward reference pointer at the start of the group

Line 2747 while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) !=	Line 2817 while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) !=

for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)	for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
{	{
offset = GET(hc, 0);	offset = (int)GET(hc, 0);
if (cd->start_code + offset == ptr + 1)	if (cd->start_code + offset == ptr + 1)
{	{
PUT(hc, 0, offset + adjust);	PUT(hc, 0, offset + adjust);
Line 2760 while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) !=	Line 2830 while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) !=

if (hc >= cd->hwm)	if (hc >= cd->hwm)
{	{
offset = GET(ptr, 1);	offset = (int)GET(ptr, 1);
if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);	if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
}	}

Line 2828 PUT(previous_callout, 2 + LINK_SIZE, length);	Line 2898 PUT(previous_callout, 2 + LINK_SIZE, length);
*************************************************/	*************************************************/

/* This function is passed the start and end of a class range, in UTF-8 mode	/* This function is passed the start and end of a class range, in UTF-8 mode
with UCP support. It searches up the characters, looking for internal ranges of	with UCP support. It searches up the characters, looking for ranges of
characters in the "other" case. Each call returns the next one, updating the	characters in the "other" case. Each call returns the next one, updating the
start address.	start address. A character with multiple other cases is returned on its own
	with a special return value.

Arguments:	Arguments:
cptr points to starting character value; updated	cptr points to starting character value; updated
Line 2838 Arguments:	Line 2909 Arguments:
ocptr where to put start of othercase range	ocptr where to put start of othercase range
odptr where to put end of othercase range	odptr where to put end of othercase range

Yield: TRUE when range returned; FALSE when no more	Yield: -1 when no more
	0 when a range is returned
	>0 the CASESET offset for char with multiple other cases
	in this case, ocptr contains the original
*/	*/

static BOOL	static int
get_othercase_range(unsigned int cptr, unsigned int d, unsigned int ocptr,	get_othercase_range(pcre_uint32 cptr, pcre_uint32 d, pcre_uint32 ocptr,
unsigned int *odptr)	pcre_uint32 *odptr)
{	{
unsigned int c, othercase, next;	pcre_uint32 c, othercase, next;
	unsigned int co;

	/* Find the first character that has an other case. If it has multiple other
	cases, return its case offset value. */

for (c = *cptr; c <= d; c++)	for (c = *cptr; c <= d; c++)
{ if ((othercase = UCD_OTHERCASE(c)) != c) break; }	{
	if ((co = UCD_CASESET(c)) != 0)
	{
	ocptr = c++; / Character that has the set */
	cptr = c; / Rest of input range */
	return (int)co;
	}
	if ((othercase = UCD_OTHERCASE(c)) != c) break;
	}

if (c > d) return FALSE;	if (c > d) return -1; /* Reached end of range */

*ocptr = othercase;	*ocptr = othercase;
next = othercase + 1;	next = othercase + 1;
Line 2861 for (++c; c <= d; c++)	Line 2947 for (++c; c <= d; c++)
next++;	next++;
}	}

*odptr = next - 1;	odptr = next - 1; / End of othercase range */
*cptr = c;	cptr = c; / Rest of input range */
	return 0;
return TRUE;
}	}


Line 2886 Returns: TRUE if auto-possessifying is OK	Line 2971 Returns: TRUE if auto-possessifying is OK
*/	*/

static BOOL	static BOOL
check_char_prop(int c, int ptype, int pdata, BOOL negated)	check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
{	{
	#ifdef SUPPORT_UCP
	const pcre_uint32 *p;
	#endif

const ucd_record *prop = GET_UCD(c);	const ucd_record *prop = GET_UCD(c);

switch(ptype)	switch(ptype)
{	{
case PT_LAMP:	case PT_LAMP:
Line 2926 switch(ptype)	Line 3016 switch(ptype)
return (PRIV(ucp_gentype)[prop->chartype] == ucp_L \|\|	return (PRIV(ucp_gentype)[prop->chartype] == ucp_L \|\|
PRIV(ucp_gentype)[prop->chartype] == ucp_N \|\|	PRIV(ucp_gentype)[prop->chartype] == ucp_N \|\|
c == CHAR_UNDERSCORE) == negated;	c == CHAR_UNDERSCORE) == negated;

	#ifdef SUPPORT_UCP
	case PT_CLIST:
	p = PRIV(ucd_caseless_sets) + prop->caseset;
	for (;;)
	{
	if (c < *p) return !negated;
	if (c == *p++) return negated;
	}
	break; /* Control never reaches here */
	#endif
}	}

return FALSE;	return FALSE;
}	}
#endif /* SUPPORT_UCP */	#endif /* SUPPORT_UCP */
Line 2943 sense to automatically possessify the repeated item.	Line 3045 sense to automatically possessify the repeated item.

Arguments:	Arguments:
previous pointer to the repeated opcode	previous pointer to the repeated opcode
utf TRUE in UTF-8 / UTF-16 mode	utf TRUE in UTF-8 / UTF-16 / UTF-32 mode
ptr next character in pattern	ptr next character in pattern
options options bits	options options bits
cd contains pointers to tables etc.	cd contains pointers to tables etc.
Line 2955 static BOOL	Line 3057 static BOOL
check_auto_possessive(const pcre_uchar *previous, BOOL utf,	check_auto_possessive(const pcre_uchar *previous, BOOL utf,
const pcre_uchar ptr, int options, compile_data cd)	const pcre_uchar ptr, int options, compile_data cd)
{	{
pcre_int32 c, next;	pcre_uint32 c = NOTACHAR;
int op_code = *previous++;	pcre_uint32 next;
	int escape;
	pcre_uchar op_code = *previous++;

/* Skip whitespace and comments in extended mode */	/* Skip whitespace and comments in extended mode */

Line 2968 if ((options & PCRE_EXTENDED) != 0)	Line 3072 if ((options & PCRE_EXTENDED) != 0)
if (*ptr == CHAR_NUMBER_SIGN)	if (*ptr == CHAR_NUMBER_SIGN)
{	{
ptr++;	ptr++;
while (*ptr != 0)	while (*ptr != CHAR_NULL)
{	{
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }	if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
ptr++;	ptr++;
Line 2987 value is a character, a negative value is an escape va	Line 3091 value is a character, a negative value is an escape va
if (*ptr == CHAR_BACKSLASH)	if (*ptr == CHAR_BACKSLASH)
{	{
int temperrorcode = 0;	int temperrorcode = 0;
next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);	escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options,
	FALSE);
if (temperrorcode != 0) return FALSE;	if (temperrorcode != 0) return FALSE;
ptr++; /* Point after the escape sequence */	ptr++; /* Point after the escape sequence */
}	}
else if (!MAX_255(ptr) \|\| (cd->ctypes[ptr] & ctype_meta) == 0)	else if (!MAX_255(ptr) \|\| (cd->ctypes[ptr] & ctype_meta) == 0)
{	{
	escape = 0;
#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
if (utf) { GETCHARINC(next, ptr); } else	if (utf) { GETCHARINC(next, ptr); } else
#endif	#endif
Line 3010 if ((options & PCRE_EXTENDED) != 0)	Line 3116 if ((options & PCRE_EXTENDED) != 0)
if (*ptr == CHAR_NUMBER_SIGN)	if (*ptr == CHAR_NUMBER_SIGN)
{	{
ptr++;	ptr++;
while (*ptr != 0)	while (*ptr != CHAR_NULL)
{	{
if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }	if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
ptr++;	ptr++;
Line 3029 if (ptr == CHAR_ASTERISK \|\| ptr == CHAR_QUESTION_MAR	Line 3135 if (ptr == CHAR_ASTERISK \|\| ptr == CHAR_QUESTION_MAR
STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)	STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
return FALSE;	return FALSE;

/* Now compare the next item with the previous opcode. First, handle cases when	/* If the previous item is a character, get its value. */
the next item is a character. */

if (next >= 0) switch(op_code)	if (op_code == OP_CHAR \|\| op_code == OP_CHARI \|\|
	op_code == OP_NOT \|\| op_code == OP_NOTI)
{	{
case OP_CHAR:
#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
GETCHARTEST(c, previous);	GETCHARTEST(c, previous);
#else	#else
c = *previous;	c = *previous;
#endif	#endif
return c != next;	}

/* For CHARI (caseless character) we must check the other case. If we have	/* Now compare the next item with the previous opcode. First, handle cases when
Unicode property support, we can use it to test the other case of	the next item is a character. */
high-valued characters. */

case OP_CHARI:	if (escape == 0)
#ifdef SUPPORT_UTF	{
GETCHARTEST(c, previous);	/* For a caseless UTF match, the next character may have more than one other
#else	case, which maps to the special PT_CLIST property. Check this first. */
c = *previous;
	#ifdef SUPPORT_UCP
	if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
	{
	unsigned int ocs = UCD_CASESET(next);
	if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
	}
#endif	#endif
if (c == next) return FALSE;
#ifdef SUPPORT_UTF	switch(op_code)
if (utf)
{	{
unsigned int othercase;	case OP_CHAR:
if (next < 128) othercase = cd->fcc[next]; else	return c != next;

	/* For CHARI (caseless character) we must check the other case. If we have
	Unicode property support, we can use it to test the other case of
	high-valued characters. We know that next can have only one other case,
	because multi-other-case characters are dealt with above. */

	case OP_CHARI:
	if (c == next) return FALSE;
	#ifdef SUPPORT_UTF
	if (utf)
	{
	pcre_uint32 othercase;
	if (next < 128) othercase = cd->fcc[next]; else
#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
othercase = UCD_OTHERCASE((unsigned int)next);	othercase = UCD_OTHERCASE(next);
#else	#else
othercase = NOTACHAR;	othercase = NOTACHAR;
#endif	#endif
return (unsigned int)c != othercase;	return c != othercase;
}	}
else	else
#endif /* SUPPORT_UTF */	#endif /* SUPPORT_UTF */
return (c != TABLE_GET((unsigned int)next, cd->fcc, next)); /* Non-UTF-8 mode */	return (c != TABLE_GET(next, cd->fcc, next)); /* Not UTF */

/* For OP_NOT and OP_NOTI, the data is always a single-byte character. These	case OP_NOT:
opcodes are not used for multi-byte characters, because they are coded using	return c == next;
an XCLASS instead. */

case OP_NOT:	case OP_NOTI:
return (c = *previous) == next;	if (c == next) return TRUE;

case OP_NOTI:
if ((c = *previous) == next) return TRUE;
#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
if (utf)	if (utf)
{	{
unsigned int othercase;	pcre_uint32 othercase;
if (next < 128) othercase = cd->fcc[next]; else	if (next < 128) othercase = cd->fcc[next]; else
#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
othercase = UCD_OTHERCASE(next);	othercase = UCD_OTHERCASE(next);
#else	#else
othercase = NOTACHAR;	othercase = NOTACHAR;
#endif	#endif
return (unsigned int)c == othercase;	return c == othercase;
}	}
else	else
#endif /* SUPPORT_UTF */	#endif /* SUPPORT_UTF */
return (c == (int)(TABLE_GET((unsigned int)next, cd->fcc, next))); /* Non-UTF-8 mode */	return (c == TABLE_GET(next, cd->fcc, next)); /* Not UTF */

/* Note that OP_DIGIT etc. are generated only when PCRE_UCP is not set.	/* Note that OP_DIGIT etc. are generated only when PCRE_UCP is not set.
When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */	When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */

case OP_DIGIT:	case OP_DIGIT:
return next > 127 \|\| (cd->ctypes[next] & ctype_digit) == 0;	return next > 255 \|\| (cd->ctypes[next] & ctype_digit) == 0;

case OP_NOT_DIGIT:	case OP_NOT_DIGIT:
return next <= 127 && (cd->ctypes[next] & ctype_digit) != 0;	return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;

case OP_WHITESPACE:	case OP_WHITESPACE:
return next > 127 \|\| (cd->ctypes[next] & ctype_space) == 0;	return next > 255 \|\| (cd->ctypes[next] & ctype_space) == 0;

case OP_NOT_WHITESPACE:	case OP_NOT_WHITESPACE:
return next <= 127 && (cd->ctypes[next] & ctype_space) != 0;	return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;

case OP_WORDCHAR:	case OP_WORDCHAR:
return next > 127 \|\| (cd->ctypes[next] & ctype_word) == 0;	return next > 255 \|\| (cd->ctypes[next] & ctype_word) == 0;

case OP_NOT_WORDCHAR:	case OP_NOT_WORDCHAR:
return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;	return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;

case OP_HSPACE:	case OP_HSPACE:
case OP_NOT_HSPACE:	case OP_NOT_HSPACE:
switch(next)	switch(next)
{	{
case 0x09:	HSPACE_CASES:
case 0x20:	return op_code == OP_NOT_HSPACE;
case 0xa0:
case 0x1680:
case 0x180e:
case 0x2000:
case 0x2001:
case 0x2002:
case 0x2003:
case 0x2004:
case 0x2005:
case 0x2006:
case 0x2007:
case 0x2008:
case 0x2009:
case 0x200A:
case 0x202f:
case 0x205f:
case 0x3000:
return op_code == OP_NOT_HSPACE;
default:
return op_code != OP_NOT_HSPACE;
}

case OP_ANYNL:	default:
case OP_VSPACE:	return op_code != OP_NOT_HSPACE;
case OP_NOT_VSPACE:	}
switch(next)
{
case 0x0a:
case 0x0b:
case 0x0c:
case 0x0d:
case 0x85:
case 0x2028:
case 0x2029:
return op_code == OP_NOT_VSPACE;
default:
return op_code != OP_NOT_VSPACE;
}

	case OP_ANYNL:
	case OP_VSPACE:
	case OP_NOT_VSPACE:
	switch(next)
	{
	VSPACE_CASES:
	return op_code == OP_NOT_VSPACE;

	default:
	return op_code != OP_NOT_VSPACE;
	}

#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
case OP_PROP:	case OP_PROP:
return check_char_prop(next, previous[0], previous[1], FALSE);	return check_char_prop(next, previous[0], previous[1], FALSE);

case OP_NOTPROP:	case OP_NOTPROP:
return check_char_prop(next, previous[0], previous[1], TRUE);	return check_char_prop(next, previous[0], previous[1], TRUE);
#endif	#endif

default:	default:
return FALSE;	return FALSE;
	}
}	}


/* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP	/* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are	is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
generated only when PCRE_UCP is not set, that is, when only ASCII	generated only when PCRE_UCP is not set, that is, when only ASCII
Line 3183 switch(op_code)	Line 3279 switch(op_code)
{	{
case OP_CHAR:	case OP_CHAR:
case OP_CHARI:	case OP_CHARI:
#ifdef SUPPORT_UTF	switch(escape)
GETCHARTEST(c, previous);
#else
c = *previous;
#endif
switch(-next)
{	{
case ESC_d:	case ESC_d:
return c > 127 \|\| (cd->ctypes[c] & ctype_digit) == 0;	return c > 255 \|\| (cd->ctypes[c] & ctype_digit) == 0;

case ESC_D:	case ESC_D:
return c <= 127 && (cd->ctypes[c] & ctype_digit) != 0;	return c <= 255 && (cd->ctypes[c] & ctype_digit) != 0;

case ESC_s:	case ESC_s:
return c > 127 \|\| (cd->ctypes[c] & ctype_space) == 0;	return c > 255 \|\| (cd->ctypes[c] & ctype_space) == 0;

case ESC_S:	case ESC_S:
return c <= 127 && (cd->ctypes[c] & ctype_space) != 0;	return c <= 255 && (cd->ctypes[c] & ctype_space) != 0;

case ESC_w:	case ESC_w:
return c > 127 \|\| (cd->ctypes[c] & ctype_word) == 0;	return c > 255 \|\| (cd->ctypes[c] & ctype_word) == 0;

case ESC_W:	case ESC_W:
return c <= 127 && (cd->ctypes[c] & ctype_word) != 0;	return c <= 255 && (cd->ctypes[c] & ctype_word) != 0;

case ESC_h:	case ESC_h:
case ESC_H:	case ESC_H:
switch(c)	switch(c)
{	{
case 0x09:	HSPACE_CASES:
case 0x20:	return escape != ESC_h;
case 0xa0:
case 0x1680:
case 0x180e:
case 0x2000:
case 0x2001:
case 0x2002:
case 0x2003:
case 0x2004:
case 0x2005:
case 0x2006:
case 0x2007:
case 0x2008:
case 0x2009:
case 0x200A:
case 0x202f:
case 0x205f:
case 0x3000:
return -next != ESC_h;
default:	default:
return -next == ESC_h;	return escape == ESC_h;
}	}

case ESC_v:	case ESC_v:
case ESC_V:	case ESC_V:
switch(c)	switch(c)
{	{
case 0x0a:	VSPACE_CASES:
case 0x0b:	return escape != ESC_v;
case 0x0c:
case 0x0d:
case 0x85:
case 0x2028:
case 0x2029:
return -next != ESC_v;
default:	default:
return -next == ESC_v;	return escape == ESC_v;
}	}

/* When PCRE_UCP is set, these values get generated for \d etc. Find	/* When PCRE_UCP is set, these values get generated for \d etc. Find
their substitutions and process them. The result will always be either	their substitutions and process them. The result will always be either
-ESC_p or -ESC_P. Then fall through to process those values. */	ESC_p or ESC_P. Then fall through to process those values. */

#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
case ESC_du:	case ESC_du:
Line 3265 switch(op_code)	Line 3334 switch(op_code)
case ESC_SU:	case ESC_SU:
{	{
int temperrorcode = 0;	int temperrorcode = 0;
ptr = substitutes[-next - ESC_DU];	ptr = substitutes[escape - ESC_DU];
next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);	escape = check_escape(&ptr, &next, &temperrorcode, 0, options, FALSE);
if (temperrorcode != 0) return FALSE;	if (temperrorcode != 0) return FALSE;
ptr++; /* For compatibility */	ptr++; /* For compatibility */
}	}
Line 3275 switch(op_code)	Line 3344 switch(op_code)
case ESC_p:	case ESC_p:
case ESC_P:	case ESC_P:
{	{
int ptype, pdata, errorcodeptr;	unsigned int ptype = 0, pdata = 0;
	int errorcodeptr;
BOOL negated;	BOOL negated;

ptr--; /* Make ptr point at the p or P */	ptr--; /* Make ptr point at the p or P */
ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);	if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
if (ptype < 0) return FALSE;	return FALSE;
ptr++; /* Point past the final curly ket */	ptr++; /* Point past the final curly ket */

/* If the property item is optional, we have to give up. (When generated	/* If the property item is optional, we have to give up. (When generated
Line 3293 switch(op_code)	Line 3363 switch(op_code)

/* Do the property check. */	/* Do the property check. */

return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);	return check_char_prop(c, ptype, pdata, (escape == ESC_P) != negated);
}	}
#endif	#endif

Line 3308 switch(op_code)	Line 3378 switch(op_code)
these op-codes are never generated.) */	these op-codes are never generated.) */

case OP_DIGIT:	case OP_DIGIT:
return next == -ESC_D \|\| next == -ESC_s \|\| next == -ESC_W \|\|	return escape == ESC_D \|\| escape == ESC_s \|\| escape == ESC_W \|\|
next == -ESC_h \|\| next == -ESC_v \|\| next == -ESC_R;	escape == ESC_h \|\| escape == ESC_v \|\| escape == ESC_R;

case OP_NOT_DIGIT:	case OP_NOT_DIGIT:
return next == -ESC_d;	return escape == ESC_d;

case OP_WHITESPACE:	case OP_WHITESPACE:
return next == -ESC_S \|\| next == -ESC_d \|\| next == -ESC_w \|\| next == -ESC_R;	return escape == ESC_S \|\| escape == ESC_d \|\| escape == ESC_w;

case OP_NOT_WHITESPACE:	case OP_NOT_WHITESPACE:
return next == -ESC_s \|\| next == -ESC_h \|\| next == -ESC_v;	return escape == ESC_s \|\| escape == ESC_h \|\| escape == ESC_v \|\| escape == ESC_R;

case OP_HSPACE:	case OP_HSPACE:
return next == -ESC_S \|\| next == -ESC_H \|\| next == -ESC_d \|\|	return escape == ESC_S \|\| escape == ESC_H \|\| escape == ESC_d \|\|
next == -ESC_w \|\| next == -ESC_v \|\| next == -ESC_R;	escape == ESC_w \|\| escape == ESC_v \|\| escape == ESC_R;

case OP_NOT_HSPACE:	case OP_NOT_HSPACE:
return next == -ESC_h;	return escape == ESC_h;

/* Can't have \S in here because VT matches \S (Perl anomaly) */	/* Can't have \S in here because VT matches \S (Perl anomaly) */
case OP_ANYNL:	case OP_ANYNL:
case OP_VSPACE:	case OP_VSPACE:
return next == -ESC_V \|\| next == -ESC_d \|\| next == -ESC_w;	return escape == ESC_V \|\| escape == ESC_d \|\| escape == ESC_w;

case OP_NOT_VSPACE:	case OP_NOT_VSPACE:
return next == -ESC_v \|\| next == -ESC_R;	return escape == ESC_v \|\| escape == ESC_R;

case OP_WORDCHAR:	case OP_WORDCHAR:
return next == -ESC_W \|\| next == -ESC_s \|\| next == -ESC_h \|\|	return escape == ESC_W \|\| escape == ESC_s \|\| escape == ESC_h \|\|
next == -ESC_v \|\| next == -ESC_R;	escape == ESC_v \|\| escape == ESC_R;

case OP_NOT_WORDCHAR:	case OP_NOT_WORDCHAR:
return next == -ESC_w \|\| next == -ESC_d;	return escape == ESC_w \|\| escape == ESC_d;

default:	default:
return FALSE;	return FALSE;
Line 3352 switch(op_code)	Line 3422 switch(op_code)


/*************************************************	/*************************************************
	* Add a character or range to a class *
	*************************************************/

	/* This function packages up the logic of adding a character or range of
	characters to a class. The character values in the arguments will be within the
	valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is
	mutually recursive with the function immediately below.

	Arguments:
	classbits the bit map for characters < 256
	uchardptr points to the pointer for extra data
	options the options word
	cd contains pointers to tables etc.
	start start of range character
	end end of range character

	Returns: the number of < 256 characters added
	the pointer to extra data is updated
	*/

	static int
	add_to_class(pcre_uint8 classbits, pcre_uchar *uchardptr, int options,
	compile_data *cd, pcre_uint32 start, pcre_uint32 end)
	{
	pcre_uint32 c;
	int n8 = 0;

	/* If caseless matching is required, scan the range and process alternate
	cases. In Unicode, there are 8-bit characters that have alternate cases that
	are greater than 255 and vice-versa. Sometimes we can just extend the original
	range. */

	if ((options & PCRE_CASELESS) != 0)
	{
	#ifdef SUPPORT_UCP
	if ((options & PCRE_UTF8) != 0)
	{
	int rc;
	pcre_uint32 oc, od;

	options &= ~PCRE_CASELESS; /* Remove for recursive calls */
	c = start;

	while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
	{
	/* Handle a single character that has more than one other case. */

	if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
	PRIV(ucd_caseless_sets) + rc, oc);

	/* Do nothing if the other case range is within the original range. */

	else if (oc >= start && od <= end) continue;

	/* Extend the original range if there is overlap, noting that if oc < c, we
	can't have od > end because a subrange is always shorter than the basic
	range. Otherwise, use a recursive call to add the additional range. */

	else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
	else if (od > end && oc <= end + 1) end = od; /* Extend upwards */
	else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
	}
	}
	else
	#endif /* SUPPORT_UCP */

	/* Not UTF-mode, or no UCP */

	for (c = start; c <= end && c < 256; c++)
	{
	SETBIT(classbits, cd->fcc[c]);
	n8++;
	}
	}

	/* Now handle the original range. Adjust the final value according to the bit
	length - this means that the same lists of (e.g.) horizontal spaces can be used
	in all cases. */

	#if defined COMPILE_PCRE8
	#ifdef SUPPORT_UTF
	if ((options & PCRE_UTF8) == 0)
	#endif
	if (end > 0xff) end = 0xff;

	#elif defined COMPILE_PCRE16
	#ifdef SUPPORT_UTF
	if ((options & PCRE_UTF16) == 0)
	#endif
	if (end > 0xffff) end = 0xffff;

	#endif /* COMPILE_PCRE[8\|16] */

	/* If all characters are less than 256, use the bit map. Otherwise use extra
	data. */

	if (end < 0x100)
	{
	for (c = start; c <= end; c++)
	{
	n8++;
	SETBIT(classbits, c);
	}
	}

	else
	{
	pcre_uchar uchardata = uchardptr;

	#ifdef SUPPORT_UTF
	if ((options & PCRE_UTF8) != 0) /* All UTFs use the same flag bit */
	{
	if (start < end)
	{
	*uchardata++ = XCL_RANGE;
	uchardata += PRIV(ord2utf)(start, uchardata);
	uchardata += PRIV(ord2utf)(end, uchardata);
	}
	else if (start == end)
	{
	*uchardata++ = XCL_SINGLE;
	uchardata += PRIV(ord2utf)(start, uchardata);
	}
	}
	else
	#endif /* SUPPORT_UTF */

	/* Without UTF support, character values are constrained by the bit length,
	and can only be > 256 for 16-bit and 32-bit libraries. */

	#ifdef COMPILE_PCRE8
	{}
	#else
	if (start < end)
	{
	*uchardata++ = XCL_RANGE;
	*uchardata++ = start;
	*uchardata++ = end;
	}
	else if (start == end)
	{
	*uchardata++ = XCL_SINGLE;
	*uchardata++ = start;
	}
	#endif

	uchardptr = uchardata; / Updata extra data pointer */
	}

	return n8; /* Number of 8-bit characters */
	}




	/*************************************************
	* Add a list of characters to a class *
	*************************************************/

	/* This function is used for adding a list of case-equivalent characters to a
	class, and also for adding a list of horizontal or vertical whitespace. If the
	list is in order (which it should be), ranges of characters are detected and
	handled appropriately. This function is mutually recursive with the function
	above.

	Arguments:
	classbits the bit map for characters < 256
	uchardptr points to the pointer for extra data
	options the options word
	cd contains pointers to tables etc.
	p points to row of 32-bit values, terminated by NOTACHAR
	except character to omit; this is used when adding lists of
	case-equivalent characters to avoid including the one we
	already know about

	Returns: the number of < 256 characters added
	the pointer to extra data is updated
	*/

	static int
	add_list_to_class(pcre_uint8 classbits, pcre_uchar *uchardptr, int options,
	compile_data cd, const pcre_uint32 p, unsigned int except)
	{
	int n8 = 0;
	while (p[0] < NOTACHAR)
	{
	int n = 0;
	if (p[0] != except)
	{
	while(p[n+1] == p[0] + n + 1) n++;
	n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
	}
	p += n + 1;
	}
	return n8;
	}



	/*************************************************
	* Add characters not in a list to a class *
	*************************************************/

	/* This function is used for adding the complement of a list of horizontal or
	vertical whitespace to a class. The list must be in order.

	Arguments:
	classbits the bit map for characters < 256
	uchardptr points to the pointer for extra data
	options the options word
	cd contains pointers to tables etc.
	p points to row of 32-bit values, terminated by NOTACHAR

	Returns: the number of < 256 characters added
	the pointer to extra data is updated
	*/

	static int
	add_not_list_to_class(pcre_uint8 classbits, pcre_uchar *uchardptr,
	int options, compile_data cd, const pcre_uint32 p)
	{
	BOOL utf = (options & PCRE_UTF8) != 0;
	int n8 = 0;
	if (p[0] > 0)
	n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
	while (p[0] < NOTACHAR)
	{
	while (p[1] == p[0] + 1) p++;
	n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
	(p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
	p++;
	}
	return n8;
	}



	/*************************************************
* Compile one branch *	* Compile one branch *
*************************************************/	*************************************************/

Line 3366 Arguments:	Line 3674 Arguments:
codeptr points to the pointer to the current code point	codeptr points to the pointer to the current code point
ptrptr points to the current pattern pointer	ptrptr points to the current pattern pointer
errorcodeptr points to error code variable	errorcodeptr points to error code variable
firstcharptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)	firstcharptr place to put the first required character
reqcharptr set to the last literal character required, else < 0	firstcharflagsptr place to put the first character flags, or a negative number
	reqcharptr place to put the last required character
	reqcharflagsptr place to put the last required character flags, or a negative number
bcptr points to current branch chain	bcptr points to current branch chain
cond_depth conditional nesting depth	cond_depth conditional nesting depth
cd contains pointers to tables etc.	cd contains pointers to tables etc.
Line 3380 Returns: TRUE on success	Line 3690 Returns: TRUE on success

static BOOL	static BOOL
compile_branch(int optionsptr, pcre_uchar *codeptr,	compile_branch(int optionsptr, pcre_uchar *codeptr,
const pcre_uchar *ptrptr, int errorcodeptr, pcre_int32 *firstcharptr,	const pcre_uchar *ptrptr, int errorcodeptr,
pcre_int32 reqcharptr, branch_chain bcptr, int cond_depth,	pcre_uint32 firstcharptr, pcre_int32 firstcharflagsptr,
	pcre_uint32 reqcharptr, pcre_int32 reqcharflagsptr,
	branch_chain *bcptr, int cond_depth,
compile_data cd, int lengthptr)	compile_data cd, int lengthptr)
{	{
int repeat_type, op_type;	int repeat_type, op_type;
int repeat_min = 0, repeat_max = 0; /* To please picky compilers */	int repeat_min = 0, repeat_max = 0; /* To please picky compilers */
int bravalue = 0;	int bravalue = 0;
int greedy_default, greedy_non_default;	int greedy_default, greedy_non_default;
pcre_int32 firstchar, reqchar;	pcre_uint32 firstchar, reqchar;
pcre_int32 zeroreqchar, zerofirstchar;	pcre_int32 firstcharflags, reqcharflags;
	pcre_uint32 zeroreqchar, zerofirstchar;
	pcre_int32 zeroreqcharflags, zerofirstcharflags;
pcre_int32 req_caseopt, reqvary, tempreqvary;	pcre_int32 req_caseopt, reqvary, tempreqvary;
int options = optionsptr; / May change dynamically */	int options = optionsptr; / May change dynamically */
int after_manual_callout = 0;	int after_manual_callout = 0;
int length_prevgroup = 0;	int length_prevgroup = 0;
register int c;	register pcre_uint32 c;
	int escape;
register pcre_uchar code = codeptr;	register pcre_uchar code = codeptr;
pcre_uchar *last_code = code;	pcre_uchar *last_code = code;
pcre_uchar *orig_code = code;	pcre_uchar *orig_code = code;
Line 3414 must not do this for other options (e.g. PCRE_EXTENDED	Line 3729 must not do this for other options (e.g. PCRE_EXTENDED
dynamically as we process the pattern. */	dynamically as we process the pattern. */

#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
/* PCRE_UTF16 has the same value as PCRE_UTF8. */	/* PCRE_UTF[16\|32] have the same value as PCRE_UTF8. */
BOOL utf = (options & PCRE_UTF8) != 0;	BOOL utf = (options & PCRE_UTF8) != 0;
	#ifndef COMPILE_PCRE32
pcre_uchar utf_chars[6];	pcre_uchar utf_chars[6];
	#endif
#else	#else
BOOL utf = FALSE;	BOOL utf = FALSE;
#endif	#endif

/* Helper variables for OP_XCLASS opcode (for characters > 255). */	/* Helper variables for OP_XCLASS opcode (for characters > 255). We define
	class_uchardata always so that it can be passed to add_to_class() always,
	though it will not be used in non-UTF 8-bit cases. This avoids having to supply
	alternative calls for the different cases. */

	pcre_uchar *class_uchardata;
#if defined SUPPORT_UTF \|\| !defined COMPILE_PCRE8	#if defined SUPPORT_UTF \|\| !defined COMPILE_PCRE8
BOOL xclass;	BOOL xclass;
pcre_uchar *class_uchardata;
pcre_uchar *class_uchardata_base;	pcre_uchar *class_uchardata_base;
#endif	#endif

Line 3448 to take the zero repeat into account. This is implemen	Line 3768 to take the zero repeat into account. This is implemen
zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual	zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
item types that can be repeated set these backoff variables appropriately. */	item types that can be repeated set these backoff variables appropriately. */

firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;	firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
	firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;

/* The variable req_caseopt contains either the REQ_CASELESS value	/* The variable req_caseopt contains either the REQ_CASELESS value
or zero, according to the current setting of the caseless flag. The	or zero, according to the current setting of the caseless flag. The
Line 3469 for (;; ptr++)	Line 3790 for (;; ptr++)
BOOL is_recurse;	BOOL is_recurse;
BOOL reset_bracount;	BOOL reset_bracount;
int class_has_8bitchar;	int class_has_8bitchar;
int class_single_char;	int class_one_char;
int newoptions;	int newoptions;
int recno;	int recno;
int refsign;	int refsign;
int skipbytes;	int skipbytes;
int subreqchar;	pcre_uint32 subreqchar, subfirstchar;
int subfirstchar;	pcre_int32 subreqcharflags, subfirstcharflags;
int terminator;	int terminator;
int mclength;	unsigned int mclength;
int tempbracount;	unsigned int tempbracount;
	pcre_uint32 ec;
pcre_uchar mcbuffer[8];	pcre_uchar mcbuffer[8];

/* Get next character in the pattern */	/* Get next character in the pattern */
Line 3488 for (;; ptr++)	Line 3810 for (;; ptr++)
/* If we are at the end of a nested substitution, revert to the outer level	/* If we are at the end of a nested substitution, revert to the outer level
string. Nesting only happens one level deep. */	string. Nesting only happens one level deep. */

if (c == 0 && nestptr != NULL)	if (c == CHAR_NULL && nestptr != NULL)
{	{
ptr = nestptr;	ptr = nestptr;
nestptr = NULL;	nestptr = NULL;
Line 3563 for (;; ptr++)	Line 3885 for (;; ptr++)

/* If in \Q...\E, check for the end; if not, we have a literal */	/* If in \Q...\E, check for the end; if not, we have a literal */

if (inescq && c != 0)	if (inescq && c != CHAR_NULL)
{	{
if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)	if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
{	{
Line 3611 for (;; ptr++)	Line 3933 for (;; ptr++)
if (c == CHAR_NUMBER_SIGN)	if (c == CHAR_NUMBER_SIGN)
{	{
ptr++;	ptr++;
while (*ptr != 0)	while (*ptr != CHAR_NULL)
{	{
if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }	if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
ptr++;	ptr++;
Line 3619 for (;; ptr++)	Line 3941 for (;; ptr++)
if (utf) FORWARDCHAR(ptr);	if (utf) FORWARDCHAR(ptr);
#endif	#endif
}	}
if (*ptr != 0) continue;	if (*ptr != CHAR_NULL) continue;

/* Else fall through to handle end of string */	/* Else fall through to handle end of string */
c = 0;	c = 0;
Line 3641 for (;; ptr++)	Line 3963 for (;; ptr++)
case CHAR_VERTICAL_LINE: /* or \| or ) */	case CHAR_VERTICAL_LINE: /* or \| or ) */
case CHAR_RIGHT_PARENTHESIS:	case CHAR_RIGHT_PARENTHESIS:
*firstcharptr = firstchar;	*firstcharptr = firstchar;
	*firstcharflagsptr = firstcharflags;
*reqcharptr = reqchar;	*reqcharptr = reqchar;
	*reqcharflagsptr = reqcharflags;
*codeptr = code;	*codeptr = code;
*ptrptr = ptr;	*ptrptr = ptr;
if (lengthptr != NULL)	if (lengthptr != NULL)
Line 3665 for (;; ptr++)	Line 3989 for (;; ptr++)
previous = NULL;	previous = NULL;
if ((options & PCRE_MULTILINE) != 0)	if ((options & PCRE_MULTILINE) != 0)
{	{
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
*code++ = OP_CIRCM;	*code++ = OP_CIRCM;
}	}
else *code++ = OP_CIRC;	else *code++ = OP_CIRC;
Line 3680 for (;; ptr++)	Line 4004 for (;; ptr++)
repeats. The value of reqchar doesn't change either. */	repeats. The value of reqchar doesn't change either. */

case CHAR_DOT:	case CHAR_DOT:
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;
previous = code;	previous = code;
*code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;	*code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
break;	break;
Line 3756 for (;; ptr++)	Line 4082 for (;; ptr++)
(cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)	(cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
{	{
*code++ = negate_class? OP_ALLANY : OP_FAIL;	*code++ = negate_class? OP_ALLANY : OP_FAIL;
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
break;	break;
}	}

Line 3767 for (;; ptr++)	Line 4094 for (;; ptr++)

should_flip_negation = FALSE;	should_flip_negation = FALSE;

/* For optimization purposes, we track some properties of the class.	/* For optimization purposes, we track some properties of the class:
class_has_8bitchar will be non-zero, if the class contains at least one	class_has_8bitchar will be non-zero if the class contains at least one <
< 256 character. class_single_char will be 1 if the class contains only	256 character; class_one_char will be 1 if the class contains just one
a single character. */	character. */

class_has_8bitchar = 0;	class_has_8bitchar = 0;
class_single_char = 0;	class_one_char = 0;

/* Initialize the 32-char bit map to all zeros. We build the map in a	/* Initialize the 32-char bit map to all zeros. We build the map in a
temporary bit of memory, in case the class contains only 1 character (less	temporary bit of memory, in case the class contains fewer than two
than 256), because in that case the compiled code doesn't use the bit map.	8-bit characters because in that case the compiled code doesn't use the bit
*/	map. */

memset(classbits, 0, 32 * sizeof(pcre_uint8));	memset(classbits, 0, 32 * sizeof(pcre_uint8));

#if defined SUPPORT_UTF \|\| !defined COMPILE_PCRE8	#if defined SUPPORT_UTF \|\| !defined COMPILE_PCRE8
xclass = FALSE; /* No chars >= 256 */	xclass = FALSE;
class_uchardata = code + LINK_SIZE + 2; /* For UTF-8 items */	class_uchardata = code + LINK_SIZE + 2; /* For XCLASS items */
class_uchardata_base = class_uchardata; /* For resetting in pass 1 */	class_uchardata_base = class_uchardata; /* Save the start */
#endif	#endif

/* Process characters until ] is reached. By writing this as a "do" it	/* Process characters until ] is reached. By writing this as a "do" it
means that an initial ] is taken as a data character. At the start of the	means that an initial ] is taken as a data character. At the start of the
loop, c contains the first byte of the character. */	loop, c contains the first byte of the character. */

if (c != 0) do	if (c != CHAR_NULL) do
{	{
const pcre_uchar *oldptr;	const pcre_uchar *oldptr;

Line 3807 for (;; ptr++)	Line 4134 for (;; ptr++)
/* In the pre-compile phase, accumulate the length of any extra	/* In the pre-compile phase, accumulate the length of any extra
data and reset the pointer. This is so that very large classes that	data and reset the pointer. This is so that very large classes that
contain a zillion > 255 characters no longer overwrite the work space	contain a zillion > 255 characters no longer overwrite the work space
(which is on the stack). */	(which is on the stack). We have to remember that there was XCLASS data,
	however. */

if (lengthptr != NULL)	if (lengthptr != NULL && class_uchardata > class_uchardata_base)
{	{
	xclass = TRUE;
*lengthptr += class_uchardata - class_uchardata_base;	*lengthptr += class_uchardata - class_uchardata_base;
class_uchardata = class_uchardata_base;	class_uchardata = class_uchardata_base;
}	}
Line 3912 for (;; ptr++)	Line 4241 for (;; ptr++)
for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];	for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];
}	}

/* Not see if we need to remove any special characters. An option	/* Now see if we need to remove any special characters. An option
value of 1 removes vertical space and 2 removes underscore. */	value of 1 removes vertical space and 2 removes underscore. */

if (tabopt < 0) tabopt = -tabopt;	if (tabopt < 0) tabopt = -tabopt;
Line 3928 for (;; ptr++)	Line 4257 for (;; ptr++)
for (c = 0; c < 32; c++) classbits[c] \|= pbits[c];	for (c = 0; c < 32; c++) classbits[c] \|= pbits[c];

ptr = tempptr + 1;	ptr = tempptr + 1;
/* Every class contains at least one < 256 characters. */	/* Every class contains at least one < 256 character. */
class_has_8bitchar = 1;	class_has_8bitchar = 1;
/* Every class contains at least two characters. */	/* Every class contains at least two characters. */
class_single_char = 2;	class_one_char = 2;
continue; /* End of POSIX syntax handling */	continue; /* End of POSIX syntax handling */
}	}

Line 3939 for (;; ptr++)	Line 4268 for (;; ptr++)
of the specials, which just set a flag. The sequence \b is a special	of the specials, which just set a flag. The sequence \b is a special
case. Inside a class (and only there) it is treated as backspace. We	case. Inside a class (and only there) it is treated as backspace. We
assume that other escapes have more than one character in them, so	assume that other escapes have more than one character in them, so
speculatively set both class_has_8bitchar and class_single_char bigger	speculatively set both class_has_8bitchar and class_one_char bigger
than one. Unrecognized escapes fall through and are either treated	than one. Unrecognized escapes fall through and are either treated
as literal characters (by default), or are faulted if	as literal characters (by default), or are faulted if
PCRE_EXTRA is set. */	PCRE_EXTRA is set. */

if (c == CHAR_BACKSLASH)	if (c == CHAR_BACKSLASH)
{	{
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);	escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options,
	TRUE);
if (*errorcodeptr != 0) goto FAILED;	if (*errorcodeptr != 0) goto FAILED;
	if (escape == 0) c = ec;
if (-c == ESC_b) c = CHAR_BS; /* \b is backspace in a class */	else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
else if (-c == ESC_N) /* \N is not supported in a class */	else if (escape == ESC_N) /* \N is not supported in a class */
{	{
*errorcodeptr = ERR71;	*errorcodeptr = ERR71;
goto FAILED;	goto FAILED;
}	}
else if (-c == ESC_Q) /* Handle start of quoted string */	else if (escape == ESC_Q) /* Handle start of quoted string */
{	{
if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)	if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
{	{
Line 3964 for (;; ptr++)	Line 4294 for (;; ptr++)
else inescq = TRUE;	else inescq = TRUE;
continue;	continue;
}	}
else if (-c == ESC_E) continue; /* Ignore orphan \E */	else if (escape == ESC_E) continue; /* Ignore orphan \E */

if (c < 0)	else
{	{
register const pcre_uint8 *cbits = cd->cbits;	register const pcre_uint8 *cbits = cd->cbits;
/* Every class contains at least two < 256 characters. */	/* Every class contains at least two < 256 characters. */
class_has_8bitchar++;	class_has_8bitchar++;
/* Every class contains at least two characters. */	/* Every class contains at least two characters. */
class_single_char += 2;	class_one_char += 2;

switch (-c)	switch (escape)
{	{
#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
case ESC_du: /* These are the values given for \d etc */	case ESC_du: /* These are the values given for \d etc */
Line 3984 for (;; ptr++)	Line 4314 for (;; ptr++)
case ESC_su: /* of the default ASCII testing. */	case ESC_su: /* of the default ASCII testing. */
case ESC_SU:	case ESC_SU:
nestptr = ptr;	nestptr = ptr;
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */	ptr = substitutes[escape - ESC_DU] - 1; /* Just before substitute */
class_has_8bitchar--; /* Undo! */	class_has_8bitchar--; /* Undo! */
continue;	continue;
#endif	#endif
Line 4008 for (;; ptr++)	Line 4338 for (;; ptr++)

/* Perl 5.004 onwards omits VT from \s, but we must preserve it	/* Perl 5.004 onwards omits VT from \s, but we must preserve it
if it was previously set by something earlier in the character	if it was previously set by something earlier in the character
class. */	class. Luckily, the value of CHAR_VT is 0x0b in both ASCII and
	EBCDIC, so we lazily just adjust the appropriate bit. */

case ESC_s:	case ESC_s:
classbits[0] \|= cbits[cbit_space];	classbits[0] \|= cbits[cbit_space];
Line 4022 for (;; ptr++)	Line 4353 for (;; ptr++)
classbits[1] \|= 0x08; /* Perl 5.004 onwards omits VT from \s */	classbits[1] \|= 0x08; /* Perl 5.004 onwards omits VT from \s */
continue;	continue;

	/* The rest apply in both UCP and non-UCP cases. */

case ESC_h:	case ESC_h:
SETBIT(classbits, 0x09); /* VT */	(void)add_list_to_class(classbits, &class_uchardata, options, cd,
SETBIT(classbits, 0x20); /* SPACE */	PRIV(hspace_list), NOTACHAR);
SETBIT(classbits, 0xa0); /* NSBP */
#ifndef COMPILE_PCRE8
xclass = TRUE;
*class_uchardata++ = XCL_SINGLE;
*class_uchardata++ = 0x1680;
*class_uchardata++ = XCL_SINGLE;
*class_uchardata++ = 0x180e;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x2000;
*class_uchardata++ = 0x200a;
*class_uchardata++ = XCL_SINGLE;
*class_uchardata++ = 0x202f;
*class_uchardata++ = XCL_SINGLE;
*class_uchardata++ = 0x205f;
*class_uchardata++ = XCL_SINGLE;
*class_uchardata++ = 0x3000;
#elif defined SUPPORT_UTF
if (utf)
{
xclass = TRUE;
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata);
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x200a, class_uchardata);
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata);
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata);
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata);
}
#endif
continue;	continue;

case ESC_H:	case ESC_H:
for (c = 0; c < 32; c++)	(void)add_not_list_to_class(classbits, &class_uchardata, options,
{	cd, PRIV(hspace_list));
int x = 0xff;
switch (c)
{
case 0x09/8: x ^= 1 << (0x09%8); break;
case 0x20/8: x ^= 1 << (0x20%8); break;
case 0xa0/8: x ^= 1 << (0xa0%8); break;
default: break;
}
classbits[c] \|= x;
}
#ifndef COMPILE_PCRE8
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x0100;
*class_uchardata++ = 0x167f;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x1681;
*class_uchardata++ = 0x180d;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x180f;
*class_uchardata++ = 0x1fff;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x200b;
*class_uchardata++ = 0x202e;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x2030;
*class_uchardata++ = 0x205e;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x2060;
*class_uchardata++ = 0x2fff;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x3001;
#ifdef SUPPORT_UTF
if (utf)
class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
else
#endif
*class_uchardata++ = 0xffff;
#elif defined SUPPORT_UTF
if (utf)
{
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x200b, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
}
#endif
continue;	continue;

case ESC_v:	case ESC_v:
SETBIT(classbits, 0x0a); /* LF */	(void)add_list_to_class(classbits, &class_uchardata, options, cd,
SETBIT(classbits, 0x0b); /* VT */	PRIV(vspace_list), NOTACHAR);
SETBIT(classbits, 0x0c); /* FF */
SETBIT(classbits, 0x0d); /* CR */
SETBIT(classbits, 0x85); /* NEL */
#ifndef COMPILE_PCRE8
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x2028;
*class_uchardata++ = 0x2029;
#elif defined SUPPORT_UTF
if (utf)
{
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);
}
#endif
continue;	continue;

case ESC_V:	case ESC_V:
for (c = 0; c < 32; c++)	(void)add_not_list_to_class(classbits, &class_uchardata, options,
{	cd, PRIV(vspace_list));
int x = 0xff;
switch (c)
{
case 0x0a/8: x ^= 1 << (0x0a%8);
x ^= 1 << (0x0b%8);
x ^= 1 << (0x0c%8);
x ^= 1 << (0x0d%8);
break;
case 0x85/8: x ^= 1 << (0x85%8); break;
default: break;
}
classbits[c] \|= x;
}

#ifndef COMPILE_PCRE8
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x0100;
*class_uchardata++ = 0x2027;
*class_uchardata++ = XCL_RANGE;
*class_uchardata++ = 0x202a;
#ifdef SUPPORT_UTF
if (utf)
class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
else
#endif
*class_uchardata++ = 0xffff;
#elif defined SUPPORT_UTF
if (utf)
{
xclass = TRUE;
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata);
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(0x202a, class_uchardata);
class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
}
#endif
continue;	continue;

#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
Line 4203 for (;; ptr++)	Line 4380 for (;; ptr++)
case ESC_P:	case ESC_P:
{	{
BOOL negated;	BOOL negated;
int pdata;	unsigned int ptype = 0, pdata = 0;
int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);	if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
if (ptype < 0) goto FAILED;	goto FAILED;
xclass = TRUE;	*class_uchardata++ = ((escape == ESC_p) != negated)?
*class_uchardata++ = ((-c == ESC_p) != negated)?
XCL_PROP : XCL_NOTPROP;	XCL_PROP : XCL_NOTPROP;
*class_uchardata++ = ptype;	*class_uchardata++ = ptype;
*class_uchardata++ = pdata;	*class_uchardata++ = pdata;
Line 4226 for (;; ptr++)	Line 4402 for (;; ptr++)
goto FAILED;	goto FAILED;
}	}
class_has_8bitchar--; /* Undo the speculative increase. */	class_has_8bitchar--; /* Undo the speculative increase. */
class_single_char -= 2; /* Undo the speculative increase. */	class_one_char -= 2; /* Undo the speculative increase. */
c = ptr; / Get the final character and fall through */	c = ptr; / Get the final character and fall through */
break;	break;
}	}
}	}

/* Fall through if we have a single character (c >= 0). This may be	/* Fall through if the escape just defined a single character (c >= 0).
greater than 256. */	This may be greater than 256. */

	escape = 0;

} /* End of backslash handling */	} /* End of backslash handling */

/* A single character may be followed by '-' to form a range. However,	/* A character may be followed by '-' to form a range. However, Perl does
Perl does not permit ']' to be the end of the range. A '-' character	not permit ']' to be the end of the range. A '-' character at the end is
at the end is treated as a literal. Perl ignores orphaned \E sequences	treated as a literal. Perl ignores orphaned \E sequences entirely. The
entirely. The code for handling \Q and \E is messy. */	code for handling \Q and \E is messy. */

CHECK_RANGE:	CHECK_RANGE:
while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)	while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
Line 4248 for (;; ptr++)	Line 4426 for (;; ptr++)
inescq = FALSE;	inescq = FALSE;
ptr += 2;	ptr += 2;
}	}

oldptr = ptr;	oldptr = ptr;

/* Remember \r or \n */	/* Remember if \r or \n were explicitly used */

if (c == CHAR_CR \|\| c == CHAR_NL) cd->external_flags \|= PCRE_HASCRORLF;	if (c == CHAR_CR \|\| c == CHAR_NL) cd->external_flags \|= PCRE_HASCRORLF;

Line 4259 for (;; ptr++)	Line 4436 for (;; ptr++)

if (!inescq && ptr[1] == CHAR_MINUS)	if (!inescq && ptr[1] == CHAR_MINUS)
{	{
int d;	pcre_uint32 d;
ptr += 2;	ptr += 2;
while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;	while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;

Line 4275 for (;; ptr++)	Line 4452 for (;; ptr++)
break;	break;
}	}

if (ptr == 0 \|\| (!inescq && ptr == CHAR_RIGHT_SQUARE_BRACKET))	/* Minus (hyphen) at the end of a class is treated as a literal, so put
	back the pointer and jump to handle the character that preceded it. */

	if (ptr == CHAR_NULL \|\| (!inescq && ptr == CHAR_RIGHT_SQUARE_BRACKET))
{	{
ptr = oldptr;	ptr = oldptr;
goto LONE_SINGLE_CHARACTER;	goto CLASS_SINGLE_CHARACTER;
}	}

	/* Otherwise, we have a potential range; pick up the next character */

#ifdef SUPPORT_UTF	#ifdef SUPPORT_UTF
if (utf)	if (utf)
{ /* Braces are required because the */	{ /* Braces are required because the */
Line 4296 for (;; ptr++)	Line 4478 for (;; ptr++)

if (!inescq && d == CHAR_BACKSLASH)	if (!inescq && d == CHAR_BACKSLASH)
{	{
d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);	int descape;
	descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE);
if (*errorcodeptr != 0) goto FAILED;	if (*errorcodeptr != 0) goto FAILED;

/* \b is backspace; any other special means the '-' was literal */	/* \b is backspace; any other special means the '-' was literal. */

if (d < 0)	if (descape != 0)
{	{
if (d == -ESC_b) d = CHAR_BS; else	if (descape == ESC_b) d = CHAR_BS; else
{	{
ptr = oldptr;	ptr = oldptr;
goto LONE_SINGLE_CHARACTER; /* A few lines below */	goto CLASS_SINGLE_CHARACTER; /* A few lines below */
}	}
}	}
}	}

/* Check that the two values are in the correct order. Optimize	/* Check that the two values are in the correct order. Optimize
one-character ranges */	one-character ranges. */

if (d < c)	if (d < c)
{	{
*errorcodeptr = ERR8;	*errorcodeptr = ERR8;
goto FAILED;	goto FAILED;
}	}
	if (d == c) goto CLASS_SINGLE_CHARACTER; /* A few lines below */

if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */	/* We have found a character range, so single character optimizations
	cannot be done anymore. Any value greater than 1 indicates that there
	is more than one character. */

/* Remember \r or \n */	class_one_char = 2;

	/* Remember an explicit \r or \n, and add the range to the class. */

if (d == CHAR_CR \|\| d == CHAR_NL) cd->external_flags \|= PCRE_HASCRORLF;	if (d == CHAR_CR \|\| d == CHAR_NL) cd->external_flags \|= PCRE_HASCRORLF;

/* Since we found a character range, single character optimizations	class_has_8bitchar +=
cannot be done anymore. */	add_to_class(classbits, &class_uchardata, options, cd, c, d);
class_single_char = 2;

/* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless	continue; /* Go get the next char in the class */
matching, we have to use an XCLASS with extra data items. Caseless	}
matching for characters > 127 is available only if UCP support is
available. */

#if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)	/* Handle a single character - we can get here for a normal non-escape
if ((d > 255) \|\| (utf && ((options & PCRE_CASELESS) != 0 && d > 127)))	char, or after \ that introduces a single character or for an apparent
#elif defined SUPPORT_UTF	range that isn't. Only the value 1 matters for class_one_char, so don't
if (utf && (d > 255 \|\| ((options & PCRE_CASELESS) != 0 && d > 127)))	increase it if it is already 2 or more ... just in case there's a class
#elif !(defined COMPILE_PCRE8)	with a zillion characters in it. */
if (d > 255)
#endif
#if defined SUPPORT_UTF \|\| !(defined COMPILE_PCRE8)
{
xclass = TRUE;

/* With UCP support, we can find the other case equivalents of	CLASS_SINGLE_CHARACTER:
the relevant characters. There may be several ranges. Optimize how	if (class_one_char < 2) class_one_char++;
they fit with the basic range. */

#ifdef SUPPORT_UCP	/* If class_one_char is 1, we have the first single character in the
#ifndef COMPILE_PCRE8	class, and there have been no prior ranges, or XCLASS items generated by
if (utf && (options & PCRE_CASELESS) != 0)	escapes. If this is the final character in the class, we can optimize by
#else	turning the item into a 1-character OP_CHAR[I] if it's positive, or
if ((options & PCRE_CASELESS) != 0)	OP_NOT[I] if it's negative. In the positive case, it can cause firstchar
#endif	to be set. Otherwise, there can be no first char if this item is first,
{	whatever repeat count may follow. In the case of reqchar, save the
unsigned int occ, ocd;	previous value for reinstating. */
unsigned int cc = c;
unsigned int origd = d;
while (get_othercase_range(&cc, origd, &occ, &ocd))
{
if (occ >= (unsigned int)c &&
ocd <= (unsigned int)d)
continue; /* Skip embedded ranges */

if (occ < (unsigned int)c &&	if (class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
ocd >= (unsigned int)c - 1) /* Extend the basic range */	{
{ /* if there is overlap, */	ptr++;
c = occ; /* noting that if occ < c */	zeroreqchar = reqchar;
continue; /* we can't have ocd > d */	zeroreqcharflags = reqcharflags;
} /* because a subrange is */
if (ocd > (unsigned int)d &&
occ <= (unsigned int)d + 1) /* always shorter than */
{ /* the basic range. */
d = ocd;
continue;
}

if (occ == ocd)	if (negate_class)
{	{
*class_uchardata++ = XCL_SINGLE;	#ifdef SUPPORT_UCP
}	int d;
else
{
*class_uchardata++ = XCL_RANGE;
class_uchardata += PRIV(ord2utf)(occ, class_uchardata);
}
class_uchardata += PRIV(ord2utf)(ocd, class_uchardata);
}
}
#endif /* SUPPORT_UCP */

/* Now record the original range, possibly modified for UCP caseless
overlapping ranges. */

*class_uchardata++ = XCL_RANGE;
#ifdef SUPPORT_UTF
#ifndef COMPILE_PCRE8
if (utf)
{
class_uchardata += PRIV(ord2utf)(c, class_uchardata);
class_uchardata += PRIV(ord2utf)(d, class_uchardata);
}
else
{
*class_uchardata++ = c;
*class_uchardata++ = d;
}
#else
class_uchardata += PRIV(ord2utf)(c, class_uchardata);
class_uchardata += PRIV(ord2utf)(d, class_uchardata);
#endif	#endif
#else /* SUPPORT_UTF */	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
*class_uchardata++ = c;	zerofirstchar = firstchar;
*class_uchardata++ = d;	zerofirstcharflags = firstcharflags;
#endif /* SUPPORT_UTF */

/* With UCP support, we are done. Without UCP support, there is no	/* For caseless UTF-8 mode when UCP support is available, check
caseless matching for UTF characters > 127; we can use the bit map	whether this character has more than one other case. If so, generate
for the smaller ones. As for 16 bit characters without UTF, we	a special OP_NOTPROP item instead of OP_NOTI. */
can still use */

#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
#ifndef COMPILE_PCRE8	if (utf && (options & PCRE_CASELESS) != 0 &&
if (utf)	(d = UCD_CASESET(c)) != 0)
#endif
continue; /* With next character in the class */
#endif /* SUPPORT_UCP */

#if defined SUPPORT_UTF && !defined(SUPPORT_UCP) && !(defined COMPILE_PCRE8)
if (utf)
{	{
if ((options & PCRE_CASELESS) == 0 \|\| c > 127) continue;	*code++ = OP_NOTPROP;
/* Adjust upper limit and fall through to set up the map */	*code++ = PT_CLIST;
d = 127;	*code++ = d;
}	}
else	else
{	#endif
if (c > 255) continue;	/* Char has only one other case, or UCP not available */
/* Adjust upper limit and fall through to set up the map */
d = 255;
}
#elif defined SUPPORT_UTF && !defined(SUPPORT_UCP)
if ((options & PCRE_CASELESS) == 0 \|\| c > 127) continue;
/* Adjust upper limit and fall through to set up the map */
d = 127;
#else
if (c > 255) continue;
/* Adjust upper limit and fall through to set up the map */
d = 255;
#endif /* SUPPORT_UTF && !SUPPORT_UCP && !COMPILE_PCRE8 */
}
#endif /* SUPPORT_UTF \|\| !COMPILE_PCRE8 */

/* We use the bit map for 8 bit mode, or when the characters fall
partially or entirely to [0-255] ([0-127] for UCP) ranges. */

class_has_8bitchar = 1;

/* We can save a bit of time by skipping this in the pre-compile. */

if (lengthptr == NULL) for (; c <= d; c++)
{
classbits[c/8] \|= (1 << (c&7));
if ((options & PCRE_CASELESS) != 0)
{	{
int uc = cd->fcc[c]; /* flip case */	*code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
classbits[uc/8] \|= (1 << (uc&7));	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
	if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
	code += PRIV(ord2utf)(c, code);
	else
	#endif
	*code++ = c;
}	}
}

continue; /* Go get the next char in the class */	/* We are finished with this character class */
}

/* Handle a lone single character - we can get here for a normal	goto END_CLASS;
non-escape char, or after \ that introduces a single character or for an
apparent range that isn't. */

LONE_SINGLE_CHARACTER:

/* Only the value of 1 matters for class_single_char. */
if (class_single_char < 2) class_single_char++;

/* If class_charcount is 1, we saw precisely one character. As long as
there were no negated characters >= 128 and there was no use of \p or \P,
in other words, no use of any XCLASS features, we can optimize.

In UTF-8 mode, we can optimize the negative case only if there were no
characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
operate on single-bytes characters only. This is an historical hangover.
Maybe one day we can tidy these opcodes to handle multi-byte characters.

The optimization throws away the bit map. We turn the item into a
1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.
Note that OP_NOT[I] does not support multibyte characters. In the positive
case, it can cause firstchar to be set. Otherwise, there can be no first
char if this item is first, whatever repeat count may follow. In the case
of reqchar, save the previous value for reinstating. */

#ifdef SUPPORT_UTF
if (class_single_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET
&& (!utf \|\| !negate_class \|\| c < (MAX_VALUE_FOR_SINGLE_CHAR + 1)))
#else
if (class_single_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
#endif
{
ptr++;
zeroreqchar = reqchar;

/* The OP_NOT[I] opcodes work on single characters only. */

if (negate_class)
{
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
zerofirstchar = firstchar;
*code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
*code++ = c;
goto NOT_CHAR;
}	}

/* For a single, positive character, get the value into mcbuffer, and	/* For a single, positive character, get the value into mcbuffer, and
then we can handle this with the normal one-character code. */	then we can handle this with the normal one-character code. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)	if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
mclength = PRIV(ord2utf)(c, mcbuffer);	mclength = PRIV(ord2utf)(c, mcbuffer);
else	else
Line 4536 for (;; ptr++)	Line 4599 for (;; ptr++)
goto ONE_CHAR;	goto ONE_CHAR;
} /* End of 1-char optimization */	} /* End of 1-char optimization */

/* Handle a character that cannot go in the bit map. */	/* There is more than one character in the class, or an XCLASS item
	has been generated. Add this character to the class. */

#if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)	class_has_8bitchar +=
if ((c > 255) \|\| (utf && ((options & PCRE_CASELESS) != 0 && c > 127)))	add_to_class(classbits, &class_uchardata, options, cd, c, c);
#elif defined SUPPORT_UTF
if (utf && (c > 255 \|\| ((options & PCRE_CASELESS) != 0 && c > 127)))
#elif !(defined COMPILE_PCRE8)
if (c > 255)
#endif

#if defined SUPPORT_UTF \|\| !(defined COMPILE_PCRE8)
{
xclass = TRUE;
*class_uchardata++ = XCL_SINGLE;
#ifdef SUPPORT_UTF
#ifndef COMPILE_PCRE8
/* In non 8 bit mode, we can get here even if we are not in UTF mode. */
if (!utf)
*class_uchardata++ = c;
else
#endif
class_uchardata += PRIV(ord2utf)(c, class_uchardata);
#else /* SUPPORT_UTF */
*class_uchardata++ = c;
#endif /* SUPPORT_UTF */

#ifdef SUPPORT_UCP
#ifdef COMPILE_PCRE8
if ((options & PCRE_CASELESS) != 0)
#else
/* In non 8 bit mode, we can get here even if we are not in UTF mode. */
if (utf && (options & PCRE_CASELESS) != 0)
#endif
{
unsigned int othercase;
if ((int)(othercase = UCD_OTHERCASE(c)) != c)
{
*class_uchardata++ = XCL_SINGLE;
class_uchardata += PRIV(ord2utf)(othercase, class_uchardata);
}
}
#endif /* SUPPORT_UCP */

}
else
#endif /* SUPPORT_UTF \|\| COMPILE_PCRE16 */

/* Handle a single-byte character */
{
class_has_8bitchar = 1;
classbits[c/8] \|= (1 << (c&7));
if ((options & PCRE_CASELESS) != 0)
{
c = cd->fcc[c]; /* flip case */
classbits[c/8] \|= (1 << (c&7));
}
}
}	}

/* Loop until ']' reached. This "while" is the end of the "do" far above.	/* Loop until ']' reached. This "while" is the end of the "do" far above.
If we are at the end of an internal nested string, revert to the outer	If we are at the end of an internal nested string, revert to the outer
string. */	string. */

while (((c = *(++ptr)) != 0 \|\|	while (((c = *(++ptr)) != CHAR_NULL \|\|
(nestptr != NULL &&	(nestptr != NULL &&
(ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&	(ptr = nestptr, nestptr = NULL, c = *(++ptr)) != CHAR_NULL)) &&
(c != CHAR_RIGHT_SQUARE_BRACKET \|\| inescq));	(c != CHAR_RIGHT_SQUARE_BRACKET \|\| inescq));

/* Check for missing terminating ']' */	/* Check for missing terminating ']' */

if (c == 0)	if (c == CHAR_NULL)
{	{
*errorcodeptr = ERR6;	*errorcodeptr = ERR6;
goto FAILED;	goto FAILED;
}	}

	/* We will need an XCLASS if data has been placed in class_uchardata. In
	the second phase this is a sufficient test. However, in the pre-compile
	phase, class_uchardata gets emptied to prevent workspace overflow, so it
	only if the very last character in the class needs XCLASS will it contain
	anything at this point. For this reason, xclass gets set TRUE above when
	uchar_classdata is emptied, and that's why this code is the way it is here
	instead of just doing a test on class_uchardata below. */

	#if defined SUPPORT_UTF \|\| !defined COMPILE_PCRE8
	if (class_uchardata > class_uchardata_base) xclass = TRUE;
	#endif

/* If this is the first thing in the branch, there can be no first char	/* If this is the first thing in the branch, there can be no first char
setting, whatever the repeat count. Any reqchar setting must remain	setting, whatever the repeat count. Any reqchar setting must remain
unchanged after any kind of repeat. */	unchanged after any kind of repeat. */

if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;

/* If there are characters with values > 255, we have to compile an	/* If there are characters with values > 255, we have to compile an
extended class, with its own opcode, unless there was a negated special	extended class, with its own opcode, unless there was a negated special
Line 4674 for (;; ptr++)	Line 4699 for (;; ptr++)
memcpy(code, classbits, 32);	memcpy(code, classbits, 32);
}	}
code += 32 / sizeof(pcre_uchar);	code += 32 / sizeof(pcre_uchar);
NOT_CHAR:
	END_CLASS:
break;	break;


Line 4712 for (;; ptr++)	Line 4738 for (;; ptr++)
if (repeat_min == 0)	if (repeat_min == 0)
{	{
firstchar = zerofirstchar; /* Adjust for zero repeat */	firstchar = zerofirstchar; /* Adjust for zero repeat */
	firstcharflags = zerofirstcharflags;
reqchar = zeroreqchar; /* Ditto */	reqchar = zeroreqchar; /* Ditto */
	reqcharflags = zeroreqcharflags;
}	}

/* Remember whether this is a variable length repeat */	/* Remember whether this is a variable length repeat */
Line 4775 for (;; ptr++)	Line 4803 for (;; ptr++)

/* Now handle repetition for the different types of item. */	/* Now handle repetition for the different types of item. */

/* If previous was a character match, abolish the item and generate a	/* If previous was a character or negated character match, abolish the item
repeat item instead. If a char item has a minumum of more than one, ensure	and generate a repeat item instead. If a char item has a minimum of more
that it is set in reqchar - it might not be if a sequence such as x{3} is	than one, ensure that it is set in reqchar - it might not be if a sequence
the first thing in a branch because the x will have gone into firstchar	such as x{3} is the first thing in a branch because the x will have gone
instead. */	into firstchar instead. */

if (previous == OP_CHAR \|\| previous == OP_CHARI)	if (previous == OP_CHAR \|\| previous == OP_CHARI
	\|\| previous == OP_NOT \|\| previous == OP_NOTI)
{	{
op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;	switch (*previous)
	{
	default: /* Make compiler happy. */
	case OP_CHAR: op_type = OP_STAR - OP_STAR; break;
	case OP_CHARI: op_type = OP_STARI - OP_STAR; break;
	case OP_NOT: op_type = OP_NOTSTAR - OP_STAR; break;
	case OP_NOTI: op_type = OP_NOTSTARI - OP_STAR; break;
	}

/* Deal with UTF characters that take up more than one character. It's	/* Deal with UTF characters that take up more than one character. It's
easier to write this out separately than try to macrify it. Use c to	easier to write this out separately than try to macrify it. Use c to
hold the length of the character in bytes, plus UTF_LENGTH to flag that	hold the length of the character in bytes, plus UTF_LENGTH to flag that
it's a length rather than a small character. */	it's a length rather than a small character. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && NOT_FIRSTCHAR(code[-1]))	if (utf && NOT_FIRSTCHAR(code[-1]))
{	{
pcre_uchar *lastchar = code - 1;	pcre_uchar *lastchar = code - 1;
Line 4806 for (;; ptr++)	Line 4842 for (;; ptr++)
with UTF disabled, or for a single character UTF character. */	with UTF disabled, or for a single character UTF character. */
{	{
c = code[-1];	c = code[-1];
if (repeat_min > 1) reqchar = c \| req_caseopt \| cd->req_varyopt;	if (*previous <= OP_CHARI && repeat_min > 1)
	{
	reqchar = c;
	reqcharflags = req_caseopt \| cd->req_varyopt;
	}
}	}

/* If the repetition is unlimited, it pays to see if the next thing on	/* If the repetition is unlimited, it pays to see if the next thing on
Line 4825 for (;; ptr++)	Line 4865 for (;; ptr++)
goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */	goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */
}	}

/* If previous was a single negated character ([^a] or similar), we use
one of the special opcodes, replacing it. The code is shared with single-
character repeats by setting opt_type to add a suitable offset into
repeat_type. We can also test for auto-possessification. OP_NOT and OP_NOTI
are currently used only for single-byte chars. */

else if (previous == OP_NOT \|\| previous == OP_NOTI)
{
op_type = ((*previous == OP_NOT)? OP_NOTSTAR : OP_NOTSTARI) - OP_STAR;
c = previous[1];
if (!possessive_quantifier &&
repeat_max < 0 &&
check_auto_possessive(previous, utf, ptr + 1, options, cd))
{
repeat_type = 0; /* Force greedy */
possessive_quantifier = TRUE;
}
goto OUTPUT_SINGLE_REPEAT;
}

/* If previous was a character type match (\d or similar), abolish it and	/* If previous was a character type match (\d or similar), abolish it and
create a suitable repeat item. The code is shared with single-character	create a suitable repeat item. The code is shared with single-character
repeats by setting op_type to add a suitable offset into repeat_type. Note	repeats by setting op_type to add a suitable offset into repeat_type. Note
Line 4883 for (;; ptr++)	Line 4903 for (;; ptr++)

if (repeat_max == 0) goto END_REPEAT;	if (repeat_max == 0) goto END_REPEAT;

/--------------------------------------------------------------------/
/* This code is obsolete from release 8.00; the restriction was finally
removed: */

/* All real repeats make it impossible to handle partial matching (maybe
one day we will be able to remove this restriction). */

/* if (repeat_max != 1) cd->external_flags \|= PCRE_NOPARTIAL; */
/--------------------------------------------------------------------/

/* Combine the op_type with the repeat_type */	/* Combine the op_type with the repeat_type */

repeat_type += op_type;	repeat_type += op_type;
Line 4945 for (;; ptr++)	Line 4955 for (;; ptr++)

if (repeat_max < 0)	if (repeat_max < 0)
{	{
#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && (c & UTF_LENGTH) != 0)	if (utf && (c & UTF_LENGTH) != 0)
{	{
memcpy(code, utf_chars, IN_UCHARS(c & 7));	memcpy(code, utf_chars, IN_UCHARS(c & 7));
Line 4970 for (;; ptr++)	Line 4980 for (;; ptr++)

else if (repeat_max != repeat_min)	else if (repeat_max != repeat_min)
{	{
#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && (c & UTF_LENGTH) != 0)	if (utf && (c & UTF_LENGTH) != 0)
{	{
memcpy(code, utf_chars, IN_UCHARS(c & 7));	memcpy(code, utf_chars, IN_UCHARS(c & 7));
Line 5000 for (;; ptr++)	Line 5010 for (;; ptr++)

/* The character or character type itself comes last in all cases. */	/* The character or character type itself comes last in all cases. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && (c & UTF_LENGTH) != 0)	if (utf && (c & UTF_LENGTH) != 0)
{	{
memcpy(code, utf_chars, IN_UCHARS(c & 7));	memcpy(code, utf_chars, IN_UCHARS(c & 7));
Line 5039 for (;; ptr++)	Line 5049 for (;; ptr++)
goto END_REPEAT;	goto END_REPEAT;
}	}

/--------------------------------------------------------------------/
/* This code is obsolete from release 8.00; the restriction was finally
removed: */

/* All real repeats make it impossible to handle partial matching (maybe
one day we will be able to remove this restriction). */

/* if (repeat_max != 1) cd->external_flags \|= PCRE_NOPARTIAL; */
/--------------------------------------------------------------------/

if (repeat_min == 0 && repeat_max == -1)	if (repeat_min == 0 && repeat_max == -1)
*code++ = OP_CRSTAR + repeat_type;	*code++ = OP_CRSTAR + repeat_type;
else if (repeat_min == 1 && repeat_max == -1)	else if (repeat_min == 1 && repeat_max == -1)
Line 5208 for (;; ptr++)	Line 5208 for (;; ptr++)

else	else
{	{
if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;	if (groupsetfirstchar && reqcharflags < 0)
	{
	reqchar = firstchar;
	reqcharflags = firstcharflags;
	}

for (i = 1; i < repeat_min; i++)	for (i = 1; i < repeat_min; i++)
{	{
Line 5583 for (;; ptr++)	Line 5587 for (;; ptr++)
if (*ptr == CHAR_COLON)	if (*ptr == CHAR_COLON)
{	{
arg = ++ptr;	arg = ++ptr;
while (ptr != 0 && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;	while (ptr != CHAR_NULL && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
arglen = (int)(ptr - arg);	arglen = (int)(ptr - arg);
	if ((unsigned int)arglen > MAX_MARK)
	{
	*errorcodeptr = ERR75;
	goto FAILED;
	}
}	}

if (*ptr != CHAR_RIGHT_PARENTHESIS)	if (*ptr != CHAR_RIGHT_PARENTHESIS)
Line 5600 for (;; ptr++)	Line 5609 for (;; ptr++)
if (namelen == verbs[i].len &&	if (namelen == verbs[i].len &&
STRNCMP_UC_C8(name, vn, namelen) == 0)	STRNCMP_UC_C8(name, vn, namelen) == 0)
{	{
	int setverb;

/* Check for open captures before ACCEPT and convert it to	/* Check for open captures before ACCEPT and convert it to
ASSERT_ACCEPT if in an assertion. */	ASSERT_ACCEPT if in an assertion. */

Line 5617 for (;; ptr++)	Line 5628 for (;; ptr++)
*code++ = OP_CLOSE;	*code++ = OP_CLOSE;
PUT2INC(code, 0, oc->number);	PUT2INC(code, 0, oc->number);
}	}
*code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;	setverb = *code++ =
	(cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;

/* Do not set firstchar after ACCEPT /	/* Do not set firstchar after ACCEPT /
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
}	}

/* Handle other cases with/without an argument */	/* Handle other cases with/without an argument */
Line 5632 for (;; ptr++)	Line 5644 for (;; ptr++)
*errorcodeptr = ERR66;	*errorcodeptr = ERR66;
goto FAILED;	goto FAILED;
}	}
*code = verbs[i].op;	setverb = *code++ = verbs[i].op;
if (*code++ == OP_THEN) cd->external_flags \|= PCRE_HASTHEN;
}	}

else	else
Line 5643 for (;; ptr++)	Line 5654 for (;; ptr++)
*errorcodeptr = ERR59;	*errorcodeptr = ERR59;
goto FAILED;	goto FAILED;
}	}
*code = verbs[i].op_arg;	setverb = *code++ = verbs[i].op_arg;
if (*code++ == OP_THEN_ARG) cd->external_flags \|= PCRE_HASTHEN;
*code++ = arglen;	*code++ = arglen;
memcpy(code, arg, IN_UCHARS(arglen));	memcpy(code, arg, IN_UCHARS(arglen));
code += arglen;	code += arglen;
*code++ = 0;	*code++ = 0;
}	}

	switch (setverb)
	{
	case OP_THEN:
	case OP_THEN_ARG:
	cd->external_flags \|= PCRE_HASTHEN;
	break;

	case OP_PRUNE:
	case OP_PRUNE_ARG:
	case OP_SKIP:
	case OP_SKIP_ARG:
	cd->had_pruneorskip = TRUE;
	break;
	}

break; /* Found verb, exit loop */	break; /* Found verb, exit loop */
}	}

Line 5676 for (;; ptr++)	Line 5701 for (;; ptr++)
{	{
case CHAR_NUMBER_SIGN: /* Comment; skip to ket */	case CHAR_NUMBER_SIGN: /* Comment; skip to ket */
ptr++;	ptr++;
while (ptr != 0 && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;	while (ptr != CHAR_NULL && ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
if (*ptr == 0)	if (*ptr == CHAR_NULL)
{	{
*errorcodeptr = ERR18;	*errorcodeptr = ERR18;
goto FAILED;	goto FAILED;
Line 5700 for (;; ptr++)	Line 5725 for (;; ptr++)
/* ------------------------------------------------------------ */	/* ------------------------------------------------------------ */
case CHAR_LEFT_PARENTHESIS:	case CHAR_LEFT_PARENTHESIS:
bravalue = OP_COND; /* Conditional group */	bravalue = OP_COND; /* Conditional group */
	tempptr = ptr;

/* A condition can be an assertion, a number (referring to a numbered	/* A condition can be an assertion, a number (referring to a numbered
group), a name (referring to a named group), or 'R', referring to	group), a name (referring to a named group), or 'R', referring to
Line 5712 for (;; ptr++)	Line 5738 for (;; ptr++)
be the recursive thing or the name 'R' (and similarly for 'R' followed	be the recursive thing or the name 'R' (and similarly for 'R' followed
by digits), and (b) a number could be a name that consists of digits.	by digits), and (b) a number could be a name that consists of digits.
In both cases, we look for a name first; if not found, we try the other	In both cases, we look for a name first; if not found, we try the other
cases. */	cases.

	For compatibility with auto-callouts, we allow a callout to be
	specified before a condition that is an assertion. First, check for the
	syntax of a callout; if found, adjust the temporary pointer that is
	used to check for an assertion condition. That's all that is needed! */

	if (ptr[1] == CHAR_QUESTION_MARK && ptr[2] == CHAR_C)
	{
	for (i = 3;; i++) if (!IS_DIGIT(ptr[i])) break;
	if (ptr[i] == CHAR_RIGHT_PARENTHESIS)
	tempptr += i + 1;
	}

/* For conditions that are assertions, check the syntax, and then exit	/* For conditions that are assertions, check the syntax, and then exit
the switch. This will take control down to where bracketed groups,	the switch. This will take control down to where bracketed groups,
including assertions, are processed. */	including assertions, are processed. */

if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN \|\|	if (tempptr[1] == CHAR_QUESTION_MARK &&
ptr[2] == CHAR_EXCLAMATION_MARK \|\| ptr[2] == CHAR_LESS_THAN_SIGN))	(tempptr[2] == CHAR_EQUALS_SIGN \|\|
	tempptr[2] == CHAR_EXCLAMATION_MARK \|\|
	tempptr[2] == CHAR_LESS_THAN_SIGN))
break;	break;

/* Most other conditions use OP_CREF (a couple change to OP_RREF	/* Most other conditions use OP_CREF (a couple change to OP_RREF
Line 5753 for (;; ptr++)	Line 5793 for (;; ptr++)
}	}
else	else
{	{
terminator = 0;	terminator = CHAR_NULL;
if (ptr[1] == CHAR_MINUS \|\| ptr[1] == CHAR_PLUS) refsign = *(++ptr);	if (ptr[1] == CHAR_MINUS \|\| ptr[1] == CHAR_PLUS) refsign = *(++ptr);
}	}

Line 5773 for (;; ptr++)	Line 5813 for (;; ptr++)
while (MAX_255(ptr) && (cd->ctypes[ptr] & ctype_word) != 0)	while (MAX_255(ptr) && (cd->ctypes[ptr] & ctype_word) != 0)
{	{
if (recno >= 0)	if (recno >= 0)
recno = (IS_DIGIT(ptr))? recno 10 + *ptr - CHAR_0 : -1;	recno = (IS_DIGIT(ptr))? recno 10 + (int)(*ptr - CHAR_0) : -1;
ptr++;	ptr++;
}	}
namelen = (int)(ptr - name);	namelen = (int)(ptr - name);

if ((terminator > 0 && *ptr++ != terminator) \|\|	if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) \|\|
*ptr++ != CHAR_RIGHT_PARENTHESIS)	*ptr++ != CHAR_RIGHT_PARENTHESIS)
{	{
ptr--; /* Error offset */	ptr--; /* Error offset */
Line 5843 for (;; ptr++)	Line 5883 for (;; ptr++)
code[1+LINK_SIZE]++;	code[1+LINK_SIZE]++;
}	}

/* If terminator == 0 it means that the name followed directly after	/* If terminator == CHAR_NULL it means that the name followed directly
the opening parenthesis [e.g. (?(abc)...] and in this case there are	after the opening parenthesis [e.g. (?(abc)...] and in this case there
some further alternatives to try. For the cases where terminator != 0	are some further alternatives to try. For the cases where terminator !=
[things like (?(<name>... or (?('name')... or (?(R&name)... ] we have	0 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
now checked all the possibilities, so give an error. */	now checked all the possibilities, so give an error. */

else if (terminator != 0)	else if (terminator != CHAR_NULL)
{	{
*errorcodeptr = ERR15;	*errorcodeptr = ERR15;
goto FAILED;	goto FAILED;
Line 6018 for (;; ptr++)	Line 6058 for (;; ptr++)

if (lengthptr != NULL)	if (lengthptr != NULL)
{	{
if (*ptr != terminator)	if (*ptr != (pcre_uchar)terminator)
{	{
*errorcodeptr = ERR42;	*errorcodeptr = ERR42;
goto FAILED;	goto FAILED;
Line 6160 for (;; ptr++)	Line 6200 for (;; ptr++)
*errorcodeptr = ERR62;	*errorcodeptr = ERR62;
goto FAILED;	goto FAILED;
}	}
if (*ptr != terminator)	if (*ptr != (pcre_uchar)terminator)
{	{
*errorcodeptr = ERR42;	*errorcodeptr = ERR42;
goto FAILED;	goto FAILED;
Line 6266 for (;; ptr++)	Line 6306 for (;; ptr++)
while(IS_DIGIT(*ptr))	while(IS_DIGIT(*ptr))
recno = recno * 10 + *ptr++ - CHAR_0;	recno = recno * 10 + *ptr++ - CHAR_0;

if (*ptr != terminator)	if (*ptr != (pcre_uchar)terminator)
{	{
*errorcodeptr = ERR29;	*errorcodeptr = ERR29;
goto FAILED;	goto FAILED;
Line 6370 for (;; ptr++)	Line 6410 for (;; ptr++)

/* Can't determine a first byte now */	/* Can't determine a first byte now */

if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
continue;	continue;


Line 6504 for (;; ptr++)	Line 6544 for (;; ptr++)
cond_depth +	cond_depth +
((bravalue == OP_COND)?1:0), /* Depth of condition subpatterns */	((bravalue == OP_COND)?1:0), /* Depth of condition subpatterns */
&subfirstchar, /* For possible first char */	&subfirstchar, /* For possible first char */
	&subfirstcharflags,
&subreqchar, /* For possible last char */	&subreqchar, /* For possible last char */
	&subreqcharflags,
bcptr, /* Current branch chain */	bcptr, /* Current branch chain */
cd, /* Tables block */	cd, /* Tables block */
(lengthptr == NULL)? NULL : /* Actual compile phase */	(lengthptr == NULL)? NULL : /* Actual compile phase */
Line 6565 for (;; ptr++)	Line 6607 for (;; ptr++)
*errorcodeptr = ERR27;	*errorcodeptr = ERR27;
goto FAILED;	goto FAILED;
}	}
if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;	if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
}	}
}	}

Line 6614 for (;; ptr++)	Line 6656 for (;; ptr++)
back off. */	back off. */

zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;
zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
groupsetfirstchar = FALSE;	groupsetfirstchar = FALSE;

if (bravalue >= OP_ONCE)	if (bravalue >= OP_ONCE)
Line 6625 for (;; ptr++)	Line 6669 for (;; ptr++)
no firstchar, set "none" for the whole branch. In both cases, a zero	no firstchar, set "none" for the whole branch. In both cases, a zero
repeat forces firstchar to "none". */	repeat forces firstchar to "none". */

if (firstchar == REQ_UNSET)	if (firstcharflags == REQ_UNSET)
{	{
if (subfirstchar >= 0)	if (subfirstcharflags >= 0)
{	{
firstchar = subfirstchar;	firstchar = subfirstchar;
	firstcharflags = subfirstcharflags;
groupsetfirstchar = TRUE;	groupsetfirstchar = TRUE;
}	}
else firstchar = REQ_NONE;	else firstcharflags = REQ_NONE;
zerofirstchar = REQ_NONE;	zerofirstcharflags = REQ_NONE;
}	}

/* If firstchar was previously set, convert the subpattern's firstchar	/* If firstchar was previously set, convert the subpattern's firstchar
into reqchar if there wasn't one, using the vary flag that was in	into reqchar if there wasn't one, using the vary flag that was in
existence beforehand. */	existence beforehand. */

else if (subfirstchar >= 0 && subreqchar < 0)	else if (subfirstcharflags >= 0 && subreqcharflags < 0)
subreqchar = subfirstchar \| tempreqvary;	{
	subreqchar = subfirstchar;
	subreqcharflags = subfirstcharflags \| tempreqvary;
	}

/* If the subpattern set a required byte (or set a first byte that isn't	/* If the subpattern set a required byte (or set a first byte that isn't
really the first byte - see above), set it. */	really the first byte - see above), set it. */

if (subreqchar >= 0) reqchar = subreqchar;	if (subreqcharflags >= 0)
	{
	reqchar = subreqchar;
	reqcharflags = subreqcharflags;
	}
}	}

/* For a forward assertion, we take the reqchar, if set. This can be	/* For a forward assertion, we take the reqchar, if set. This can be
Line 6657 for (;; ptr++)	Line 6709 for (;; ptr++)
of a firstchar. This is overcome by a scan at the end if there's no	of a firstchar. This is overcome by a scan at the end if there's no
firstchar, looking for an asserted first char. */	firstchar, looking for an asserted first char. */

else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;	else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
	{
	reqchar = subreqchar;
	reqcharflags = subreqcharflags;
	}
break; /* End of processing '(' */	break; /* End of processing '(' */


Line 6665 for (;; ptr++)	Line 6721 for (;; ptr++)
/* Handle metasequences introduced by \. For ones like \d, the ESC_ values	/* Handle metasequences introduced by \. For ones like \d, the ESC_ values
are arranged to be the negation of the corresponding OP_values in the	are arranged to be the negation of the corresponding OP_values in the
default case when PCRE_UCP is not set. For the back references, the values	default case when PCRE_UCP is not set. For the back references, the values
are ESC_REF plus the reference number. Only back references and those types	are negative the reference number. Only back references and those types
that consume a character may be repeated. We can test for values between	that consume a character may be repeated. We can test for values between
ESC_b and ESC_Z for the latter; this may have to change if any new ones are	ESC_b and ESC_Z for the latter; this may have to change if any new ones are
ever created. */	ever created. */

case CHAR_BACKSLASH:	case CHAR_BACKSLASH:
tempptr = ptr;	tempptr = ptr;
c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);	escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
if (*errorcodeptr != 0) goto FAILED;	if (*errorcodeptr != 0) goto FAILED;

if (c < 0)	if (escape == 0) /* The escape coded a single character */
	c = ec;
	else
{	{
if (-c == ESC_Q) /* Handle start of quoted string */	if (escape == ESC_Q) /* Handle start of quoted string */
{	{
if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)	if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
ptr += 2; /* avoid empty string */	ptr += 2; /* avoid empty string */
Line 6685 for (;; ptr++)	Line 6743 for (;; ptr++)
continue;	continue;
}	}

if (-c == ESC_E) continue; /* Perl ignores an orphan \E */	if (escape == ESC_E) continue; /* Perl ignores an orphan \E */

/* For metasequences that actually match a character, we disable the	/* For metasequences that actually match a character, we disable the
setting of a first character if it hasn't already been set. */	setting of a first character if it hasn't already been set. */

if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)	if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
firstchar = REQ_NONE;	firstcharflags = REQ_NONE;

/* Set values to reset to if this is followed by a zero repeat. */	/* Set values to reset to if this is followed by a zero repeat. */

zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;

/* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'	/* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
is a subroutine call by number (Oniguruma syntax). In fact, the value	is a subroutine call by number (Oniguruma syntax). In fact, the value
-ESC_g is returned only for these cases. So we don't need to check for <	ESC_g is returned only for these cases. So we don't need to check for <
or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is	or ' if the value is ESC_g. For the Perl syntax \g{n} the value is
-ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as	-n, and for the Perl syntax \g{name} the result is ESC_k (as
that is a synonym for a named back reference). */	that is a synonym for a named back reference). */

if (-c == ESC_g)	if (escape == ESC_g)
{	{
const pcre_uchar *p;	const pcre_uchar *p;
save_hwm = cd->hwm; /* Normally this is set when '(' is read */	save_hwm = cd->hwm; /* Normally this is set when '(' is read */
Line 6725 for (;; ptr++)	Line 6785 for (;; ptr++)
if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)	if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
{	{
BOOL is_a_number = TRUE;	BOOL is_a_number = TRUE;
for (p = ptr + 1; p != 0 && p != terminator; p++)	for (p = ptr + 1; p != CHAR_NULL && p != (pcre_uchar)terminator; p++)
{	{
if (!MAX_255(*p)) { is_a_number = FALSE; break; }	if (!MAX_255(*p)) { is_a_number = FALSE; break; }
if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;	if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
if ((cd->ctypes[*p] & ctype_word) == 0) break;	if ((cd->ctypes[*p] & ctype_word) == 0) break;
}	}
if (*p != terminator)	if (*p != (pcre_uchar)terminator)
{	{
*errorcodeptr = ERR57;	*errorcodeptr = ERR57;
break;	break;
Line 6749 for (;; ptr++)	Line 6809 for (;; ptr++)

p = ptr + 2;	p = ptr + 2;
while (IS_DIGIT(*p)) p++;	while (IS_DIGIT(*p)) p++;
if (*p != terminator)	if (*p != (pcre_uchar)terminator)
{	{
*errorcodeptr = ERR57;	*errorcodeptr = ERR57;
break;	break;
Line 6761 for (;; ptr++)	Line 6821 for (;; ptr++)
/* \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).
We also support \k{name} (.NET syntax). */	We also support \k{name} (.NET syntax). */

if (-c == ESC_k)	if (escape == ESC_k)
{	{
if ((ptr[1] != CHAR_LESS_THAN_SIGN &&	if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))	ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
Line 6780 for (;; ptr++)	Line 6840 for (;; ptr++)
not set to cope with cases like (?=(\w+))\1: which would otherwise set	not set to cope with cases like (?=(\w+))\1: which would otherwise set
':' later. */	':' later. */

if (-c >= ESC_REF)	if (escape < 0)
{	{
open_capitem *oc;	open_capitem *oc;
recno = -c - ESC_REF;	recno = -escape;

HANDLE_REFERENCE: /* Come here from named backref handling */	HANDLE_REFERENCE: /* Come here from named backref handling */
if (firstchar == REQ_UNSET) firstchar = REQ_NONE;	if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
previous = code;	previous = code;
*code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;	*code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
PUT2INC(code, 0, recno);	PUT2INC(code, 0, recno);
Line 6810 for (;; ptr++)	Line 6870 for (;; ptr++)
/* So are Unicode property matches, if supported. */	/* So are Unicode property matches, if supported. */

#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
else if (-c == ESC_P \|\| -c == ESC_p)	else if (escape == ESC_P \|\| escape == ESC_p)
{	{
BOOL negated;	BOOL negated;
int pdata;	unsigned int ptype = 0, pdata = 0;
int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);	if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
if (ptype < 0) goto FAILED;	goto FAILED;
previous = code;	previous = code;
*code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;	*code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
*code++ = ptype;	*code++ = ptype;
*code++ = pdata;	*code++ = pdata;
}	}
Line 6826 for (;; ptr++)	Line 6886 for (;; ptr++)
/* If Unicode properties are not supported, \X, \P, and \p are not	/* If Unicode properties are not supported, \X, \P, and \p are not
allowed. */	allowed. */

else if (-c == ESC_X \|\| -c == ESC_P \|\| -c == ESC_p)	else if (escape == ESC_X \|\| escape == ESC_P \|\| escape == ESC_p)
{	{
*errorcodeptr = ERR45;	*errorcodeptr = ERR45;
goto FAILED;	goto FAILED;
Line 6836 for (;; ptr++)	Line 6896 for (;; ptr++)
/* For the rest (including \X when Unicode properties are supported), we	/* For the rest (including \X when Unicode properties are supported), we
can obtain the OP value by negating the escape value in the default	can obtain the OP value by negating the escape value in the default
situation when PCRE_UCP is not set. When it is set, we substitute	situation when PCRE_UCP is not set. When it is set, we substitute
Unicode property tests. */	Unicode property tests. Note that \b and \B do a one-character
	lookbehind, and \A also behaves as if it does. */

else	else
{	{
	if ((escape == ESC_b \|\| escape == ESC_B \|\| escape == ESC_A) &&
	cd->max_lookbehind == 0)
	cd->max_lookbehind = 1;
#ifdef SUPPORT_UCP	#ifdef SUPPORT_UCP
if (-c >= ESC_DU && -c <= ESC_wu)	if (escape >= ESC_DU && escape <= ESC_wu)
{	{
nestptr = ptr + 1; /* Where to resume */	nestptr = ptr + 1; /* Where to resume */
ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */	ptr = substitutes[escape - ESC_DU] - 1; /* Just before substitute */
}	}
else	else
#endif	#endif
Line 6852 for (;; ptr++)	Line 6916 for (;; ptr++)
so that it works in DFA mode and in lookbehinds. */	so that it works in DFA mode and in lookbehinds. */

{	{
previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;	previous = (escape > ESC_b && escape < ESC_Z)? code : NULL;
*code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;	*code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape;
}	}
}	}
continue;	continue;
Line 6863 for (;; ptr++)	Line 6927 for (;; ptr++)
a value > 127. We set its representation in the length/buffer, and then	a value > 127. We set its representation in the length/buffer, and then
handle it as a data character. */	handle it as a data character. */

#ifdef SUPPORT_UTF	#if defined SUPPORT_UTF && !defined COMPILE_PCRE32
if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)	if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
mclength = PRIV(ord2utf)(c, mcbuffer);	mclength = PRIV(ord2utf)(c, mcbuffer);
else	else
Line 6896 for (;; ptr++)	Line 6960 for (;; ptr++)

ONE_CHAR:	ONE_CHAR:
previous = code;	previous = code;

	/* For caseless UTF-8 mode when UCP support is available, check whether
	this character has more than one other case. If so, generate a special
	OP_PROP item instead of OP_CHARI. */

	#ifdef SUPPORT_UCP
	if (utf && (options & PCRE_CASELESS) != 0)
	{
	GETCHAR(c, mcbuffer);
	if ((c = UCD_CASESET(c)) != 0)
	{
	*code++ = OP_PROP;
	*code++ = PT_CLIST;
	*code++ = c;
	if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
	break;
	}
	}
	#endif

	/* Caseful matches, or not one of the multicase characters. */

*code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;	*code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];	for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];

Line 6909 for (;; ptr++)	Line 6995 for (;; ptr++)
Otherwise, leave the firstchar value alone, and don't change it on a zero	Otherwise, leave the firstchar value alone, and don't change it on a zero
repeat. */	repeat. */

if (firstchar == REQ_UNSET)	if (firstcharflags == REQ_UNSET)
{	{
zerofirstchar = REQ_NONE;	zerofirstcharflags = REQ_NONE;
zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;

/* If the character is more than one byte long, we can set firstchar	/* If the character is more than one byte long, we can set firstchar
only if it is not to be matched caselessly. */	only if it is not to be matched caselessly. */
Line 6920 for (;; ptr++)	Line 7007 for (;; ptr++)
if (mclength == 1 \|\| req_caseopt == 0)	if (mclength == 1 \|\| req_caseopt == 0)
{	{
firstchar = mcbuffer[0] \| req_caseopt;	firstchar = mcbuffer[0] \| req_caseopt;
if (mclength != 1) reqchar = code[-1] \| cd->req_varyopt;	firstchar = mcbuffer[0];
	firstcharflags = req_caseopt;

	if (mclength != 1)
	{
	reqchar = code[-1];
	reqcharflags = cd->req_varyopt;
	}
}	}
else firstchar = reqchar = REQ_NONE;	else firstcharflags = reqcharflags = REQ_NONE;
}	}

/* firstchar was previously set; we can set reqchar only if the length is	/* firstchar was previously set; we can set reqchar only if the length is
Line 6931 for (;; ptr++)	Line 7025 for (;; ptr++)
else	else
{	{
zerofirstchar = firstchar;	zerofirstchar = firstchar;
	zerofirstcharflags = firstcharflags;
zeroreqchar = reqchar;	zeroreqchar = reqchar;
	zeroreqcharflags = reqcharflags;
if (mclength == 1 \|\| req_caseopt == 0)	if (mclength == 1 \|\| req_caseopt == 0)
reqchar = code[-1] \| req_caseopt \| cd->req_varyopt;	{
	reqchar = code[-1];
	reqcharflags = req_caseopt \| cd->req_varyopt;
	}
}	}

break; /* End of literal character handling */	break; /* End of literal character handling */
Line 6952 return FALSE;	Line 7051 return FALSE;




/*************************************************	/*************************************************
* Compile sequence of alternatives *	* Compile sequence of alternatives *
*************************************************/	*************************************************/
Line 6973 Arguments:	Line 7071 Arguments:
reset_bracount TRUE to reset the count for each branch	reset_bracount TRUE to reset the count for each branch
skipbytes skip this many bytes at start (for brackets and OP_COND)	skipbytes skip this many bytes at start (for brackets and OP_COND)
cond_depth depth of nesting for conditional subpatterns	cond_depth depth of nesting for conditional subpatterns
firstcharptr place to put the first required character, or a negative number	firstcharptr place to put the first required character
reqcharptr place to put the last required character, or a negative number	firstcharflagsptr place to put the first character flags, or a negative number
	reqcharptr place to put the last required character
	reqcharflagsptr place to put the last required character flags, or a negative number
bcptr pointer to the chain of currently open branches	bcptr pointer to the chain of currently open branches
cd points to the data block with tables pointers etc.	cd points to the data block with tables pointers etc.
lengthptr NULL during the real compile phase	lengthptr NULL during the real compile phase
Line 6986 Returns: TRUE on success	Line 7086 Returns: TRUE on success
static BOOL	static BOOL
compile_regex(int options, pcre_uchar codeptr, const pcre_uchar ptrptr,	compile_regex(int options, pcre_uchar codeptr, const pcre_uchar ptrptr,
int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,	int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
int cond_depth, pcre_int32 firstcharptr, pcre_int32 reqcharptr,	int cond_depth,
	pcre_uint32 firstcharptr, pcre_int32 firstcharflagsptr,
	pcre_uint32 reqcharptr, pcre_int32 reqcharflagsptr,
branch_chain bcptr, compile_data cd, int *lengthptr)	branch_chain bcptr, compile_data cd, int *lengthptr)
{	{
const pcre_uchar ptr = ptrptr;	const pcre_uchar ptr = ptrptr;
Line 6996 pcre_uchar *start_bracket = code;	Line 7098 pcre_uchar *start_bracket = code;
pcre_uchar *reverse_count = NULL;	pcre_uchar *reverse_count = NULL;
open_capitem capitem;	open_capitem capitem;
int capnumber = 0;	int capnumber = 0;
pcre_int32 firstchar, reqchar;	pcre_uint32 firstchar, reqchar;
pcre_int32 branchfirstchar, branchreqchar;	pcre_int32 firstcharflags, reqcharflags;
	pcre_uint32 branchfirstchar, branchreqchar;
	pcre_int32 branchfirstcharflags, branchreqcharflags;
int length;	int length;
int orig_bracount;	unsigned int orig_bracount;
int max_bracount;	unsigned int max_bracount;
branch_chain bc;	branch_chain bc;

bc.outer = bcptr;	bc.outer = bcptr;
bc.current_branch = code;	bc.current_branch = code;

firstchar = reqchar = REQ_UNSET;	firstchar = reqchar = 0;
	firstcharflags = reqcharflags = REQ_UNSET;

/* Accumulate the length for use in the pre-compile phase. Start with the	/* Accumulate the length for use in the pre-compile phase. Start with the
length of the BRA and KET and any extra bytes that are required at the	length of the BRA and KET and any extra bytes that are required at the
Line 7066 for (;;)	Line 7171 for (;;)
into the length. */	into the length. */

if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,	if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
&branchreqchar, &bc, cond_depth, cd,	&branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
(lengthptr == NULL)? NULL : &length))	cond_depth, cd, (lengthptr == NULL)? NULL : &length))
{	{
*ptrptr = ptr;	*ptrptr = ptr;
return FALSE;	return FALSE;
Line 7088 for (;;)	Line 7193 for (;;)
if (*last_branch != OP_ALT)	if (*last_branch != OP_ALT)
{	{
firstchar = branchfirstchar;	firstchar = branchfirstchar;
	firstcharflags = branchfirstcharflags;
reqchar = branchreqchar;	reqchar = branchreqchar;
	reqcharflags = branchreqcharflags;
}	}

/* If this is not the first branch, the first char and reqchar have to	/* If this is not the first branch, the first char and reqchar have to
Line 7102 for (;;)	Line 7209 for (;;)
we have to abandon the firstchar for the regex, but if there was	we have to abandon the firstchar for the regex, but if there was
previously no reqchar, it takes on the value of the old firstchar. */	previously no reqchar, it takes on the value of the old firstchar. */

if (firstchar >= 0 && firstchar != branchfirstchar)	if (firstcharflags >= 0 &&
	(firstcharflags != branchfirstcharflags \|\| firstchar != branchfirstchar))
{	{
if (reqchar < 0) reqchar = firstchar;	if (reqcharflags < 0)
firstchar = REQ_NONE;	{
	reqchar = firstchar;
	reqcharflags = firstcharflags;
	}
	firstcharflags = REQ_NONE;
}	}

/* If we (now or from before) have no firstchar, a firstchar from the	/* If we (now or from before) have no firstchar, a firstchar from the
branch becomes a reqchar if there isn't a branch reqchar. */	branch becomes a reqchar if there isn't a branch reqchar. */

if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)	if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
branchreqchar = branchfirstchar;	{
	branchreqchar = branchfirstchar;
	branchreqcharflags = branchfirstcharflags;
	}

/* Now ensure that the reqchars match */	/* Now ensure that the reqchars match */

if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))	if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) \|\|
reqchar = REQ_NONE;	reqchar != branchreqchar)
else reqchar \|= branchreqchar; /* To "or" REQ_VARY */	reqcharflags = REQ_NONE;
	else
	{
	reqchar = branchreqchar;
	reqcharflags \|= branchreqcharflags; /* To "or" REQ_VARY */
	}
}	}

/* If lookbehind, check that this branch matches a fixed-length string, and	/* If lookbehind, check that this branch matches a fixed-length string, and
Line 7147 for (;;)	Line 7267 for (;;)
*ptrptr = ptr;	*ptrptr = ptr;
return FALSE;	return FALSE;
}	}
else { PUT(reverse_count, 0, fixed_length); }	else
	{
	if (fixed_length > cd->max_lookbehind)
	cd->max_lookbehind = fixed_length;
	PUT(reverse_count, 0, fixed_length);
	}
}	}
}	}

Line 7209 for (;;)	Line 7334 for (;;)
*codeptr = code;	*codeptr = code;
*ptrptr = ptr;	*ptrptr = ptr;
*firstcharptr = firstchar;	*firstcharptr = firstchar;
	*firstcharflagsptr = firstcharflags;
*reqcharptr = reqchar;	*reqcharptr = reqchar;
	*reqcharflagsptr = reqcharflags;
if (lengthptr != NULL)	if (lengthptr != NULL)
{	{
if (OFLOW_MAX - *lengthptr < length)	if (OFLOW_MAX - *lengthptr < length)
Line 7279 and the highest back reference was greater than or equ	Line 7406 and the highest back reference was greater than or equ
However, by keeping a bitmap of the first 31 back references, we can catch some	However, by keeping a bitmap of the first 31 back references, we can catch some
of the more common cases more precisely.	of the more common cases more precisely.

	... A second exception is when the .* appears inside an atomic group, because
	this prevents the number of characters it matches from being adjusted.

Arguments:	Arguments:
code points to start of expression (the bracket)	code points to start of expression (the bracket)
bracket_map a bitmap of which brackets we are inside while testing; this	bracket_map a bitmap of which brackets we are inside while testing; this
handles up to substring 31; after that we just have to take	handles up to substring 31; after that we just have to take
the less precise approach	the less precise approach
backref_map the back reference bitmap	cd points to the compile data block
	atomcount atomic group level

Returns: TRUE or FALSE	Returns: TRUE or FALSE
*/	*/

static BOOL	static BOOL
is_anchored(register const pcre_uchar *code, unsigned int bracket_map,	is_anchored(register const pcre_uchar *code, unsigned int bracket_map,
unsigned int backref_map)	compile_data *cd, int atomcount)
{	{
do {	do {
const pcre_uchar *scode = first_significant_code(	const pcre_uchar *scode = first_significant_code(
Line 7303 do {	Line 7434 do {
if (op == OP_BRA \|\| op == OP_BRAPOS \|\|	if (op == OP_BRA \|\| op == OP_BRAPOS \|\|
op == OP_SBRA \|\| op == OP_SBRAPOS)	op == OP_SBRA \|\| op == OP_SBRAPOS)
{	{
if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;	if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
}	}

/* Capturing brackets */	/* Capturing brackets */
Line 7313 do {	Line 7444 do {
{	{
int n = GET2(scode, 1+LINK_SIZE);	int n = GET2(scode, 1+LINK_SIZE);
int new_map = bracket_map \| ((n < 32)? (1 << n) : 1);	int new_map = bracket_map \| ((n < 32)? (1 << n) : 1);
if (!is_anchored(scode, new_map, backref_map)) return FALSE;	if (!is_anchored(scode, new_map, cd, atomcount)) return FALSE;
}	}

/* Other brackets */	/* Positive forward assertions and conditions */

else if (op == OP_ASSERT \|\| op == OP_ONCE \|\| op == OP_ONCE_NC \|\|	else if (op == OP_ASSERT \|\| op == OP_COND)
op == OP_COND)
{	{
if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;	if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
}	}

	/* Atomic groups */

	else if (op == OP_ONCE \|\| op == OP_ONCE_NC)
	{
	if (!is_anchored(scode, bracket_map, cd, atomcount + 1))
	return FALSE;
	}

/* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and	/* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and
it isn't in brackets that are or may be referenced. */	it isn't in brackets that are or may be referenced or inside an atomic
	group. */

else if ((op == OP_TYPESTAR \|\| op == OP_TYPEMINSTAR \|\|	else if ((op == OP_TYPESTAR \|\| op == OP_TYPEMINSTAR \|\|
op == OP_TYPEPOSSTAR))	op == OP_TYPEPOSSTAR))
{	{
if (scode[1] != OP_ALLANY \|\| (bracket_map & backref_map) != 0)	if (scode[1] != OP_ALLANY \|\| (bracket_map & cd->backref_map) != 0 \|\|
	atomcount > 0 \|\| cd->had_pruneorskip)
return FALSE;	return FALSE;
}	}

/* Check for explicit anchoring */	/* Check for explicit anchoring */

else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;	else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;

code += GET(code, 1);	code += GET(code, 1);
}	}
while (code == OP_ALT); / Loop for each alternative */	while (code == OP_ALT); / Loop for each alternative */
Line 7354 return TRUE;	Line 7495 return TRUE;
matching and for non-DOTALL patterns that start with .* (which must start at	matching and for non-DOTALL patterns that start with .* (which must start at
the beginning or after \n). As in the case of is_anchored() (see above), we	the beginning or after \n). As in the case of is_anchored() (see above), we
have to take account of back references to capturing brackets that contain .*	have to take account of back references to capturing brackets that contain .*
because in that case we can't make the assumption.	because in that case we can't make the assumption. Also, the appearance of .*
	inside atomic brackets or in a pattern that contains PRUNE or SKIP does not
	count, because once again the assumption no longer holds.

Arguments:	Arguments:
code points to start of expression (the bracket)	code points to start of expression (the bracket)
bracket_map a bitmap of which brackets we are inside while testing; this	bracket_map a bitmap of which brackets we are inside while testing; this
handles up to substring 31; after that we just have to take	handles up to substring 31; after that we just have to take
the less precise approach	the less precise approach
backref_map the back reference bitmap	cd points to the compile data
	atomcount atomic group level

Returns: TRUE or FALSE	Returns: TRUE or FALSE
*/	*/

static BOOL	static BOOL
is_startline(const pcre_uchar *code, unsigned int bracket_map,	is_startline(const pcre_uchar *code, unsigned int bracket_map,
unsigned int backref_map)	compile_data *cd, int atomcount)
{	{
do {	do {
const pcre_uchar *scode = first_significant_code(	const pcre_uchar *scode = first_significant_code(
Line 7394 do {	Line 7538 do {
return FALSE;	return FALSE;

default: /* Assertion */	default: /* Assertion */
if (!is_startline(scode, bracket_map, backref_map)) return FALSE;	if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
do scode += GET(scode, 1); while (*scode == OP_ALT);	do scode += GET(scode, 1); while (*scode == OP_ALT);
scode += 1 + LINK_SIZE;	scode += 1 + LINK_SIZE;
break;	break;
Line 7408 do {	Line 7552 do {
if (op == OP_BRA \|\| op == OP_BRAPOS \|\|	if (op == OP_BRA \|\| op == OP_BRAPOS \|\|
op == OP_SBRA \|\| op == OP_SBRAPOS)	op == OP_SBRA \|\| op == OP_SBRAPOS)
{	{
if (!is_startline(scode, bracket_map, backref_map)) return FALSE;	if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
}	}

/* Capturing brackets */	/* Capturing brackets */
Line 7418 do {	Line 7562 do {
{	{
int n = GET2(scode, 1+LINK_SIZE);	int n = GET2(scode, 1+LINK_SIZE);
int new_map = bracket_map \| ((n < 32)? (1 << n) : 1);	int new_map = bracket_map \| ((n < 32)? (1 << n) : 1);
if (!is_startline(scode, new_map, backref_map)) return FALSE;	if (!is_startline(scode, new_map, cd, atomcount)) return FALSE;
}	}

/* Other brackets */	/* Positive forward assertions */

else if (op == OP_ASSERT \|\| op == OP_ONCE \|\| op == OP_ONCE_NC)	else if (op == OP_ASSERT)
{	{
if (!is_startline(scode, bracket_map, backref_map)) return FALSE;	if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
}	}

/* .* means "start at start or after \n" if it isn't in brackets that	/* Atomic brackets */
may be referenced. */

	else if (op == OP_ONCE \|\| op == OP_ONCE_NC)
	{
	if (!is_startline(scode, bracket_map, cd, atomcount + 1)) return FALSE;
	}

	/* .* means "start at start or after \n" if it isn't in atomic brackets or
	brackets that may be referenced, as long as the pattern does not contain
	PRUNE or SKIP, because these break the feature. Consider, for example,
	/.?a(PRUNE)b/ with the subject "aab", which matches "ab", i.e. not at the
	start of a line. */

else if (op == OP_TYPESTAR \|\| op == OP_TYPEMINSTAR \|\| op == OP_TYPEPOSSTAR)	else if (op == OP_TYPESTAR \|\| op == OP_TYPEMINSTAR \|\| op == OP_TYPEPOSSTAR)
{	{
if (scode[1] != OP_ANY \|\| (bracket_map & backref_map) != 0) return FALSE;	if (scode[1] != OP_ANY \|\| (bracket_map & cd->backref_map) != 0 \|\|
	atomcount > 0 \|\| cd->had_pruneorskip)
	return FALSE;
}	}

/* Check for explicit circumflex */	/* Check for explicit circumflex; anything else gives a FALSE result. Note
	in particular that this includes atomic brackets OP_ONCE and OP_ONCE_NC
	because the number of characters matched by .* cannot be adjusted inside
	them. */

else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;	else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;

Line 7464 we return that char, otherwise -1.	Line 7623 we return that char, otherwise -1.

Arguments:	Arguments:
code points to start of expression (the bracket)	code points to start of expression (the bracket)
	flags points to the first char flags, or to REQ_NONE
inassert TRUE if in an assertion	inassert TRUE if in an assertion

Returns: -1 or the fixed first char	Returns: the fixed first char, or 0 with REQ_NONE in flags
*/	*/

static int	static pcre_uint32
find_firstassertedchar(const pcre_uchar *code, BOOL inassert)	find_firstassertedchar(const pcre_uchar code, pcre_int32 flags,
	BOOL inassert)
{	{
register int c = -1;	register pcre_uint32 c = 0;
	int cflags = REQ_NONE;

	*flags = REQ_NONE;
do {	do {
int d;	pcre_uint32 d;
	int dflags;
int xl = (code == OP_CBRA \|\| code == OP_SCBRA \|\|	int xl = (code == OP_CBRA \|\| code == OP_SCBRA \|\|
code == OP_CBRAPOS \|\| code == OP_SCBRAPOS)? IMM2_SIZE:0;	code == OP_CBRAPOS \|\| code == OP_SCBRAPOS)? IMM2_SIZE:0;
const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,	const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
TRUE);	TRUE);
register int op = *scode;	register pcre_uchar op = *scode;

switch(op)	switch(op)
{	{
default:	default:
return -1;	return 0;

case OP_BRA:	case OP_BRA:
case OP_BRAPOS:	case OP_BRAPOS:
Line 7496 do {	Line 7661 do {
case OP_ONCE:	case OP_ONCE:
case OP_ONCE_NC:	case OP_ONCE_NC:
case OP_COND:	case OP_COND:
if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)	d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
return -1;	if (dflags < 0)
if (c < 0) c = d; else if (c != d) return -1;	return 0;
	if (cflags < 0) { c = d; cflags = dflags; } else if (c != d \|\| cflags != dflags) return 0;
break;	break;

case OP_EXACT:	case OP_EXACT:
Line 7509 do {	Line 7675 do {
case OP_PLUS:	case OP_PLUS:
case OP_MINPLUS:	case OP_MINPLUS:
case OP_POSPLUS:	case OP_POSPLUS:
if (!inassert) return -1;	if (!inassert) return 0;
if (c < 0) c = scode[1];	if (cflags < 0) { c = scode[1]; cflags = 0; }
else if (c != scode[1]) return -1;	else if (c != scode[1]) return 0;
break;	break;

case OP_EXACTI:	case OP_EXACTI:
Line 7522 do {	Line 7688 do {
case OP_PLUSI:	case OP_PLUSI:
case OP_MINPLUSI:	case OP_MINPLUSI:
case OP_POSPLUSI:	case OP_POSPLUSI:
if (!inassert) return -1;	if (!inassert) return 0;
if (c < 0) c = scode[1] \| REQ_CASELESS;	if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
else if (c != scode[1]) return -1;	else if (c != scode[1]) return 0;
break;	break;
}	}

code += GET(code, 1);	code += GET(code, 1);
}	}
while (*code == OP_ALT);	while (*code == OP_ALT);

	*flags = cflags;
return c;	return c;
}	}

Line 7558 Returns: pointer to compiled data block, or NUL	Line 7726 Returns: pointer to compiled data block, or NUL
with errorptr and erroroffset set	with errorptr and erroroffset set
*/	*/

#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION	PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
pcre_compile(const char pattern, int options, const char *errorptr,	pcre_compile(const char pattern, int options, const char *errorptr,
int erroroffset, const unsigned char tables)	int erroroffset, const unsigned char tables)
#else	#elif defined COMPILE_PCRE16
PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION	PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,	pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
int erroroffset, const unsigned char tables)	int erroroffset, const unsigned char tables)
	#elif defined COMPILE_PCRE32
	PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
	pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr,
	int erroroffset, const unsigned char tables)
#endif	#endif
{	{
#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);	return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
#else	#elif defined COMPILE_PCRE16
return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);	return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
	#elif defined COMPILE_PCRE32
	return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
#endif	#endif
}	}


#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION	PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
pcre_compile2(const char pattern, int options, int errorcodeptr,	pcre_compile2(const char pattern, int options, int errorcodeptr,
const char *errorptr, int erroroffset, const unsigned char *tables)	const char *errorptr, int erroroffset, const unsigned char *tables)
#else	#elif defined COMPILE_PCRE16
PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION	PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,	pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
const char *errorptr, int erroroffset, const unsigned char *tables)	const char *errorptr, int erroroffset, const unsigned char *tables)
	#elif defined COMPILE_PCRE32
	PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
	pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr,
	const char *errorptr, int erroroffset, const unsigned char *tables)
#endif	#endif
{	{
REAL_PCRE *re;	REAL_PCRE *re;
int length = 1; /* For final END opcode */	int length = 1; /* For final END opcode */
pcre_int32 firstchar, reqchar;	pcre_int32 firstcharflags, reqcharflags;
	pcre_uint32 firstchar, reqchar;
	pcre_uint32 limit_match = PCRE_UINT32_MAX;
	pcre_uint32 limit_recursion = PCRE_UINT32_MAX;
int newline;	int newline;
int errorcode = 0;	int errorcode = 0;
int skipatstart = 0;	int skipatstart = 0;
BOOL utf;	BOOL utf;
	BOOL never_utf = FALSE;
size_t size;	size_t size;
pcre_uchar *code;	pcre_uchar *code;
const pcre_uchar *codestart;	const pcre_uchar *codestart;
Line 7652 if ((options & ~PUBLIC_COMPILE_OPTIONS) != 0)	Line 7834 if ((options & ~PUBLIC_COMPILE_OPTIONS) != 0)
goto PCRE_EARLY_ERROR_RETURN;	goto PCRE_EARLY_ERROR_RETURN;
}	}

	/* If PCRE_NEVER_UTF is set, remember it. */

	if ((options & PCRE_NEVER_UTF) != 0) never_utf = TRUE;

/* 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
the offset for later. */	the offset for later. */

	cd->external_flags = 0; /* Initialize here for LIMIT_MATCH/RECURSION */

while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&	while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&
ptr[skipatstart+1] == CHAR_ASTERISK)	ptr[skipatstart+1] == CHAR_ASTERISK)
{	{
int newnl = 0;	int newnl = 0;
int newbsr = 0;	int newbsr = 0;

	/* For completeness and backward compatibility, (*UTFn) is supported in the
	relevant libraries, but (*UTF) is generic and always supported. Note that
	PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */

#ifdef COMPILE_PCRE8	#ifdef COMPILE_PCRE8
if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)	if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0)
{ skipatstart += 7; options \|= PCRE_UTF8; continue; }	{ skipatstart += 7; options \|= PCRE_UTF8; continue; }
#endif	#endif
#ifdef COMPILE_PCRE16	#ifdef COMPILE_PCRE16
if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)	if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0)
{ skipatstart += 8; options \|= PCRE_UTF16; continue; }	{ skipatstart += 8; options \|= PCRE_UTF16; continue; }
#endif	#endif
	#ifdef COMPILE_PCRE32
	if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0)
	{ skipatstart += 8; options \|= PCRE_UTF32; continue; }
	#endif

	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0)
	{ skipatstart += 6; options \|= PCRE_UTF8; continue; }
else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
{ skipatstart += 6; options \|= PCRE_UCP; continue; }	{ skipatstart += 6; options \|= PCRE_UCP; continue; }
else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
{ skipatstart += 15; options \|= PCRE_NO_START_OPTIMIZE; continue; }	{ skipatstart += 15; options \|= PCRE_NO_START_OPTIMIZE; continue; }

	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_MATCH_EQ, 12) == 0)
	{
	pcre_uint32 c = 0;
	int p = skipatstart + 14;
	while (isdigit(ptr[p]))
	{
	if (c > PCRE_UINT32_MAX / 10 - 1) break; /* Integer overflow */
	c = c*10 + ptr[p++] - CHAR_0;
	}
	if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break;
	if (c < limit_match)
	{
	limit_match = c;
	cd->external_flags \|= PCRE_MLSET;
	}
	skipatstart = p;
	continue;
	}

	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_RECURSION_EQ, 16) == 0)
	{
	pcre_uint32 c = 0;
	int p = skipatstart + 18;
	while (isdigit(ptr[p]))
	{
	if (c > PCRE_UINT32_MAX / 10 - 1) break; /* Integer overflow check */
	c = c*10 + ptr[p++] - CHAR_0;
	}
	if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break;
	if (c < limit_recursion)
	{
	limit_recursion = c;
	cd->external_flags \|= PCRE_RLSET;
	}
	skipatstart = p;
	continue;
	}

if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0)	if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0)
{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; }	{ skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3) == 0)	else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3) == 0)
Line 7697 while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&	Line 7934 while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&
else break;	else break;
}	}

/* PCRE_UTF16 has the same value as PCRE_UTF8. */	/* PCRE_UTF(16\|32) have the same value as PCRE_UTF8. */
utf = (options & PCRE_UTF8) != 0;	utf = (options & PCRE_UTF8) != 0;
	if (utf && never_utf)
	{
	errorcode = ERR78;
	goto PCRE_EARLY_ERROR_RETURN2;
	}

/* Can't support UTF unless PCRE has been compiled to include the code. The	/* Can't support UTF unless PCRE has been compiled to include the code. The
return of an error code from PRIV(valid_utf)() is a new feature, introduced in	return of an error code from PRIV(valid_utf)() is a new feature, introduced in
Line 7709 not used here. */	Line 7951 not used here. */
if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&	if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
(errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)	(errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
{	{
#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
errorcode = ERR44;	errorcode = ERR44;
#else	#elif defined COMPILE_PCRE16
errorcode = ERR74;	errorcode = ERR74;
	#elif defined COMPILE_PCRE32
	errorcode = ERR77;
#endif	#endif
goto PCRE_EARLY_ERROR_RETURN2;	goto PCRE_EARLY_ERROR_RETURN2;
}	}
Line 7817 cd->start_pattern = (const pcre_uchar *)pattern;	Line 8061 cd->start_pattern = (const pcre_uchar *)pattern;
cd->end_pattern = (const pcre_uchar )(pattern + STRLEN_UC((const pcre_uchar )pattern));	cd->end_pattern = (const pcre_uchar )(pattern + STRLEN_UC((const pcre_uchar )pattern));
cd->req_varyopt = 0;	cd->req_varyopt = 0;
cd->assert_depth = 0;	cd->assert_depth = 0;
	cd->max_lookbehind = 0;
cd->external_options = options;	cd->external_options = options;
cd->external_flags = 0;
cd->open_caps = NULL;	cd->open_caps = NULL;

/* Now do the pre-compile. On error, errorcode will be set non-zero, so we	/* Now do the pre-compile. On error, errorcode will be set non-zero, so we
Line 7831 ptr += skipatstart;	Line 8075 ptr += skipatstart;
code = cworkspace;	code = cworkspace;
*code = OP_BRA;	*code = OP_BRA;
(void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,	(void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);	FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
	cd, &length);
if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;	if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;

DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,	DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
Line 7867 re->magic_number = MAGIC_NUMBER;	Line 8112 re->magic_number = MAGIC_NUMBER;
re->size = (int)size;	re->size = (int)size;
re->options = cd->external_options;	re->options = cd->external_options;
re->flags = cd->external_flags;	re->flags = cd->external_flags;
re->dummy1 = 0;	re->limit_match = limit_match;
	re->limit_recursion = limit_recursion;
re->first_char = 0;	re->first_char = 0;
re->req_char = 0;	re->req_char = 0;
re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar);	re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar);
Line 7876 re->name_count = cd->names_found;	Line 8122 re->name_count = cd->names_found;
re->ref_count = 0;	re->ref_count = 0;
re->tables = (tables == PRIV(default_tables))? NULL : tables;	re->tables = (tables == PRIV(default_tables))? NULL : tables;
re->nullpad = NULL;	re->nullpad = NULL;
	#ifdef COMPILE_PCRE32
	re->dummy = 0;
	#else
	re->dummy1 = re->dummy2 = re->dummy3 = 0;
	#endif

/* The starting points of the name/number translation table and of the code are	/* The starting points of the name/number translation table and of the code are
passed around in the compile data block. The start/end pattern and initial	passed around in the compile data block. The start/end pattern and initial
Line 7887 field; this time it's used for remembering forward ref	Line 8138 field; this time it's used for remembering forward ref
cd->final_bracount = cd->bracount; /* Save for checking forward references */	cd->final_bracount = cd->bracount; /* Save for checking forward references */
cd->assert_depth = 0;	cd->assert_depth = 0;
cd->bracount = 0;	cd->bracount = 0;
	cd->max_lookbehind = 0;
cd->names_found = 0;	cd->names_found = 0;
cd->name_table = (pcre_uchar *)re + re->name_table_offset;	cd->name_table = (pcre_uchar *)re + re->name_table_offset;
codestart = cd->name_table + re->name_entry_size * re->name_count;	codestart = cd->name_table + re->name_entry_size * re->name_count;
Line 7894 cd->start_code = codestart;	Line 8146 cd->start_code = codestart;
cd->hwm = (pcre_uchar *)(cd->start_workspace);	cd->hwm = (pcre_uchar *)(cd->start_workspace);
cd->req_varyopt = 0;	cd->req_varyopt = 0;
cd->had_accept = FALSE;	cd->had_accept = FALSE;
	cd->had_pruneorskip = FALSE;
cd->check_lookbehind = FALSE;	cd->check_lookbehind = FALSE;
cd->open_caps = NULL;	cd->open_caps = NULL;

Line 7905 ptr = (const pcre_uchar *)pattern + skipatstart;	Line 8158 ptr = (const pcre_uchar *)pattern + skipatstart;
code = (pcre_uchar *)codestart;	code = (pcre_uchar *)codestart;
*code = OP_BRA;	*code = OP_BRA;
(void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,	(void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
&firstchar, &reqchar, NULL, cd, NULL);	&firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
re->top_bracket = cd->bracount;	re->top_bracket = cd->bracount;
re->top_backref = cd->top_backref;	re->top_backref = cd->top_backref;
	re->max_lookbehind = cd->max_lookbehind;
re->flags = cd->external_flags \| PCRE_MODE;	re->flags = cd->external_flags \| PCRE_MODE;

if (cd->had_accept) reqchar = REQ_NONE; /* Must disable after (ACCEPT) /	if (cd->had_accept)
	{
	reqchar = 0; /* Must disable after (ACCEPT) /
	reqcharflags = REQ_NONE;
	}

/* If not reached end of pattern on success, there's an excess bracket. */	/* If not reached end of pattern on success, there's an excess bracket. */

if (errorcode == 0 && *ptr != 0) errorcode = ERR22;	if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22;

/* Fill in the terminating state and check for disastrous overflow, but	/* Fill in the terminating state and check for disastrous overflow, but
if debugging, leave the test till after things are printed out. */	if debugging, leave the test till after things are printed out. */
Line 7925 if debugging, leave the test till after things are pri	Line 8183 if debugging, leave the test till after things are pri
if (code - codestart > length) errorcode = ERR23;	if (code - codestart > length) errorcode = ERR23;
#endif	#endif

	#ifdef SUPPORT_VALGRIND
	/* If the estimated length exceeds the really used length, mark the extra
	allocated memory as unaddressable, so that any out-of-bound reads can be
	detected. */
	VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
	#endif

/* Fill in any forward references that are required. There may be repeated	/* Fill in any forward references that are required. There may be repeated
references; optimize for them, as searching a large regex takes time. */	references; optimize for them, as searching a large regex takes time. */

Line 7960 if (errorcode == 0 && re->top_backref > re->top_bracke	Line 8225 if (errorcode == 0 && re->top_backref > re->top_bracke

/* If there were any lookbehind assertions that contained OP_RECURSE	/* If there were any lookbehind assertions that contained OP_RECURSE
(recursions or subroutine calls), a flag is set for them to be checked here,	(recursions or subroutine calls), a flag is set for them to be checked here,
because they may contain forward references. Actual recursions can't be fixed	because they may contain forward references. Actual recursions cannot be fixed
length, but subroutine calls can. It is done like this so that those without	length, but subroutine calls can. It is done like this so that those without
OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The	OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The
exceptional ones forgo this. We scan the pattern to check that they are fixed	exceptional ones forgo this. We scan the pattern to check that they are fixed
Line 7995 if (cd->check_lookbehind)	Line 8260 if (cd->check_lookbehind)
(fixed_length == -4)? ERR70 : ERR25;	(fixed_length == -4)? ERR70 : ERR25;
break;	break;
}	}
	if (fixed_length > cd->max_lookbehind) cd->max_lookbehind = fixed_length;
PUT(cc, 1, fixed_length);	PUT(cc, 1, fixed_length);
}	}
cc += 1 + LINK_SIZE;	cc += 1 + LINK_SIZE;
Line 8015 if (errorcode != 0)	Line 8281 if (errorcode != 0)
}	}

/* If the anchored option was not passed, set the flag if we can determine that	/* If the anchored option was not passed, set the flag if we can determine that
the pattern is anchored by virtue of ^ characters or \A or anything else (such	the pattern is anchored by virtue of ^ characters or \A or anything else, such
as starting with .* when DOTALL is set).	as starting with non-atomic .* when DOTALL is set and there are no occurrences
	of PRUNE or SKIP.

Otherwise, if we know what the first byte has to be, save it, because that	Otherwise, if we know what the first byte has to be, save it, because that
speeds up unanchored matches no end. If not, see if we can set the	speeds up unanchored matches no end. If not, see if we can set the
PCRE_STARTLINE flag. This is helpful for multiline matches when all branches	PCRE_STARTLINE flag. This is helpful for multiline matches when all branches
start with ^. and also when all branches start with .* for non-DOTALL matches.	start with ^. and also when all branches start with non-atomic .* for
*/	non-DOTALL matches when PRUNE and SKIP are not present. /

if ((re->options & PCRE_ANCHORED) == 0)	if ((re->options & PCRE_ANCHORED) == 0)
{	{
if (is_anchored(codestart, 0, cd->backref_map))	if (is_anchored(codestart, 0, cd, 0)) re->options \|= PCRE_ANCHORED;
re->options \|= PCRE_ANCHORED;
else	else
{	{
if (firstchar < 0)	if (firstcharflags < 0)
firstchar = find_firstassertedchar(codestart, FALSE);	firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
if (firstchar >= 0) /* Remove caseless flag for non-caseable chars */	if (firstcharflags >= 0) /* Remove caseless flag for non-caseable chars */
{	{
#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
re->first_char = firstchar & 0xff;	re->first_char = firstchar & 0xff;
#else	#elif defined COMPILE_PCRE16
#ifdef COMPILE_PCRE16
re->first_char = firstchar & 0xffff;	re->first_char = firstchar & 0xffff;
	#elif defined COMPILE_PCRE32
	re->first_char = firstchar;
#endif	#endif
#endif	if ((firstcharflags & REQ_CASELESS) != 0)
if ((firstchar & REQ_CASELESS) != 0)
{	{
#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)	#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
/* We ignore non-ASCII first chars in 8 bit mode. */	/* We ignore non-ASCII first chars in 8 bit mode. */
Line 8064 if ((re->options & PCRE_ANCHORED) == 0)	Line 8330 if ((re->options & PCRE_ANCHORED) == 0)

re->flags \|= PCRE_FIRSTSET;	re->flags \|= PCRE_FIRSTSET;
}	}
else if (is_startline(codestart, 0, cd->backref_map))
re->flags \|= PCRE_STARTLINE;	else if (is_startline(codestart, 0, cd, 0)) re->flags \|= PCRE_STARTLINE;
}	}
}	}

Line 8073 if ((re->options & PCRE_ANCHORED) == 0)	Line 8339 if ((re->options & PCRE_ANCHORED) == 0)
variable length item in the regex. Remove the caseless flag for non-caseable	variable length item in the regex. Remove the caseless flag for non-caseable
bytes. */	bytes. */

if (reqchar >= 0 &&	if (reqcharflags >= 0 &&
((re->options & PCRE_ANCHORED) == 0 \|\| (reqchar & REQ_VARY) != 0))	((re->options & PCRE_ANCHORED) == 0 \|\| (reqcharflags & REQ_VARY) != 0))
{	{
#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
re->req_char = reqchar & 0xff;	re->req_char = reqchar & 0xff;
#else	#elif defined COMPILE_PCRE16
#ifdef COMPILE_PCRE16
re->req_char = reqchar & 0xffff;	re->req_char = reqchar & 0xffff;
	#elif defined COMPILE_PCRE32
	re->req_char = reqchar;
#endif	#endif
#endif	if ((reqcharflags & REQ_CASELESS) != 0)
if ((reqchar & REQ_CASELESS) != 0)
{	{
#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)	#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
/* We ignore non-ASCII first chars in 8 bit mode. */	/* We ignore non-ASCII first chars in 8 bit mode. */
Line 8133 if ((re->flags & PCRE_REQCHSET) != 0)	Line 8399 if ((re->flags & PCRE_REQCHSET) != 0)
else printf("Req char = \\x%02x%s\n", ch, caseless);	else printf("Req char = \\x%02x%s\n", ch, caseless);
}	}

#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
pcre_printint((pcre *)re, stdout, TRUE);	pcre_printint((pcre *)re, stdout, TRUE);
#else	#elif defined COMPILE_PCRE16
pcre16_printint((pcre *)re, stdout, TRUE);	pcre16_printint((pcre *)re, stdout, TRUE);
	#elif defined COMPILE_PCRE32
	pcre32_printint((pcre *)re, stdout, TRUE);
#endif	#endif

/* This check is done here in the debugging case so that the code that	/* This check is done here in the debugging case so that the code that
Line 8152 if (code - codestart > length)	Line 8420 if (code - codestart > length)
}	}
#endif /* PCRE_DEBUG */	#endif /* PCRE_DEBUG */

#ifdef COMPILE_PCRE8	#if defined COMPILE_PCRE8
return (pcre *)re;	return (pcre *)re;
#else	#elif defined COMPILE_PCRE16
return (pcre16 *)re;	return (pcre16 *)re;
	#elif defined COMPILE_PCRE32
	return (pcre32 *)re;
#endif	#endif
}	}

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>

Removed from v.1.1.1.2
changed lines
	Added in v.1.1.1.4