/* *INDENT-OFF* */ /* keep in sync with glibc */
|
/* *INDENT-OFF* */ /* keep in sync with glibc */
|
/* Extended regular expression matching and search library,
|
/* Extended regular expression matching and search library,
|
version 0.12.
|
version 0.12.
|
(Implements POSIX draft P1003.2/D11.2, except for some of the
|
(Implements POSIX draft P1003.2/D11.2, except for some of the
|
internationalization features.)
|
internationalization features.)
|
Copyright (C) 1993, 94, 95, 96, 97, 98 Free Software Foundation, Inc.
|
Copyright (C) 1993, 94, 95, 96, 97, 98 Free Software Foundation, Inc.
|
|
|
NOTE: The canonical source of this file is maintained with the
|
NOTE: The canonical source of this file is maintained with the
|
GNU C Library. Bugs can be reported to bug-glibc@gnu.org.
|
GNU C Library. Bugs can be reported to bug-glibc@gnu.org.
|
|
|
This program is free software; you can redistribute it and/or modify it
|
This program is free software; you can redistribute it and/or modify it
|
under the terms of the GNU General Public License as published by the
|
under the terms of the GNU General Public License as published by the
|
Free Software Foundation; either version 2, or (at your option) any
|
Free Software Foundation; either version 2, or (at your option) any
|
later version.
|
later version.
|
|
|
This program is distributed in the hope that it will be useful,
|
This program is distributed in the hope that it will be useful,
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
GNU General Public License for more details.
|
GNU General Public License for more details.
|
|
|
You should have received a copy of the GNU General Public License
|
You should have received a copy of the GNU General Public License
|
along with this program; if not, write to the Free Software Foundation,
|
along with this program; if not, write to the Free Software Foundation,
|
Inc., 59 Temple Place - Suite 330,
|
Inc., 59 Temple Place - Suite 330,
|
Boston, MA 02111-1307, USA. */
|
Boston, MA 02111-1307, USA. */
|
|
|
/* AIX requires this to be the first thing in the file. */
|
/* AIX requires this to be the first thing in the file. */
|
#if defined _AIX && !defined REGEX_MALLOC
|
#if defined _AIX && !defined REGEX_MALLOC
|
#pragma alloca
|
#pragma alloca
|
#endif
|
#endif
|
|
|
#undef _GNU_SOURCE
|
#undef _GNU_SOURCE
|
#define _GNU_SOURCE
|
#define _GNU_SOURCE
|
|
|
#ifdef HAVE_CONFIG_H
|
#ifdef HAVE_CONFIG_H
|
# include <config.h>
|
# include <config.h>
|
#endif
|
#endif
|
|
|
#ifndef PARAMS
|
#ifndef PARAMS
|
# if defined __GNUC__ || (defined __STDC__ && __STDC__)
|
# if defined __GNUC__ || (defined __STDC__ && __STDC__)
|
# define PARAMS(args) args
|
# define PARAMS(args) args
|
# else
|
# else
|
# define PARAMS(args) ()
|
# define PARAMS(args) ()
|
# endif /* GCC. */
|
# endif /* GCC. */
|
#endif /* Not PARAMS. */
|
#endif /* Not PARAMS. */
|
|
|
#if defined STDC_HEADERS && !defined emacs
|
#if defined STDC_HEADERS && !defined emacs
|
# include <stddef.h>
|
# include <stddef.h>
|
#else
|
#else
|
/* We need this for `gnu-regex.h', and perhaps for the Emacs include files. */
|
/* We need this for `gnu-regex.h', and perhaps for the Emacs include files. */
|
# include <sys/types.h>
|
# include <sys/types.h>
|
#endif
|
#endif
|
|
|
/* For platform which support the ISO C amendement 1 functionality we
|
/* For platform which support the ISO C amendement 1 functionality we
|
support user defined character classes. */
|
support user defined character classes. */
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
|
/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
|
/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
|
# include <wchar.h>
|
# include <wchar.h>
|
# include <wctype.h>
|
# include <wctype.h>
|
#endif
|
#endif
|
|
|
/* This is for other GNU distributions with internationalized messages. */
|
/* This is for other GNU distributions with internationalized messages. */
|
/* CYGNUS LOCAL: ../intl will handle this for us */
|
/* CYGNUS LOCAL: ../intl will handle this for us */
|
#ifdef ENABLE_NLS
|
#ifdef ENABLE_NLS
|
# include <libintl.h>
|
# include <libintl.h>
|
#else
|
#else
|
# define gettext(msgid) (msgid)
|
# define gettext(msgid) (msgid)
|
#endif
|
#endif
|
|
|
#ifndef gettext_noop
|
#ifndef gettext_noop
|
/* This define is so xgettext can find the internationalizable
|
/* This define is so xgettext can find the internationalizable
|
strings. */
|
strings. */
|
# define gettext_noop(String) String
|
# define gettext_noop(String) String
|
#endif
|
#endif
|
|
|
/* The `emacs' switch turns on certain matching commands
|
/* The `emacs' switch turns on certain matching commands
|
that make sense only in Emacs. */
|
that make sense only in Emacs. */
|
#ifdef emacs
|
#ifdef emacs
|
|
|
# include "lisp.h"
|
# include "lisp.h"
|
# include "buffer.h"
|
# include "buffer.h"
|
# include "syntax.h"
|
# include "syntax.h"
|
|
|
#else /* not emacs */
|
#else /* not emacs */
|
|
|
/* If we are not linking with Emacs proper,
|
/* If we are not linking with Emacs proper,
|
we can't use the relocating allocator
|
we can't use the relocating allocator
|
even if config.h says that we can. */
|
even if config.h says that we can. */
|
# undef REL_ALLOC
|
# undef REL_ALLOC
|
|
|
# if defined STDC_HEADERS || defined _LIBC
|
# if defined STDC_HEADERS || defined _LIBC
|
# include <stdlib.h>
|
# include <stdlib.h>
|
# else
|
# else
|
char *malloc ();
|
char *malloc ();
|
char *realloc ();
|
char *realloc ();
|
# endif
|
# endif
|
|
|
/* When used in Emacs's lib-src, we need to get bzero and bcopy somehow.
|
/* When used in Emacs's lib-src, we need to get bzero and bcopy somehow.
|
If nothing else has been done, use the method below. */
|
If nothing else has been done, use the method below. */
|
# ifdef INHIBIT_STRING_HEADER
|
# ifdef INHIBIT_STRING_HEADER
|
# if !(defined HAVE_BZERO && defined HAVE_BCOPY)
|
# if !(defined HAVE_BZERO && defined HAVE_BCOPY)
|
# if !defined bzero && !defined bcopy
|
# if !defined bzero && !defined bcopy
|
# undef INHIBIT_STRING_HEADER
|
# undef INHIBIT_STRING_HEADER
|
# endif
|
# endif
|
# endif
|
# endif
|
# endif
|
# endif
|
|
|
/* This is the normal way of making sure we have a bcopy and a bzero.
|
/* This is the normal way of making sure we have a bcopy and a bzero.
|
This is used in most programs--a few other programs avoid this
|
This is used in most programs--a few other programs avoid this
|
by defining INHIBIT_STRING_HEADER. */
|
by defining INHIBIT_STRING_HEADER. */
|
# ifndef INHIBIT_STRING_HEADER
|
# ifndef INHIBIT_STRING_HEADER
|
# if defined HAVE_STRING_H || defined STDC_HEADERS || defined _LIBC
|
# if defined HAVE_STRING_H || defined STDC_HEADERS || defined _LIBC
|
# include <string.h>
|
# include <string.h>
|
# ifndef bzero
|
# ifndef bzero
|
# ifndef _LIBC
|
# ifndef _LIBC
|
# define bzero(s, n) (memset (s, '\0', n), (s))
|
# define bzero(s, n) (memset (s, '\0', n), (s))
|
# else
|
# else
|
# define bzero(s, n) __bzero (s, n)
|
# define bzero(s, n) __bzero (s, n)
|
# endif
|
# endif
|
# endif
|
# endif
|
# else
|
# else
|
# include <strings.h>
|
# include <strings.h>
|
# ifndef memcmp
|
# ifndef memcmp
|
# define memcmp(s1, s2, n) bcmp (s1, s2, n)
|
# define memcmp(s1, s2, n) bcmp (s1, s2, n)
|
# endif
|
# endif
|
# ifndef memcpy
|
# ifndef memcpy
|
# define memcpy(d, s, n) (bcopy (s, d, n), (d))
|
# define memcpy(d, s, n) (bcopy (s, d, n), (d))
|
# endif
|
# endif
|
# endif
|
# endif
|
# endif
|
# endif
|
|
|
/* Define the syntax stuff for \<, \>, etc. */
|
/* Define the syntax stuff for \<, \>, etc. */
|
|
|
/* This must be nonzero for the wordchar and notwordchar pattern
|
/* This must be nonzero for the wordchar and notwordchar pattern
|
commands in re_match_2. */
|
commands in re_match_2. */
|
# ifndef Sword
|
# ifndef Sword
|
# define Sword 1
|
# define Sword 1
|
# endif
|
# endif
|
|
|
# ifdef SWITCH_ENUM_BUG
|
# ifdef SWITCH_ENUM_BUG
|
# define SWITCH_ENUM_CAST(x) ((int)(x))
|
# define SWITCH_ENUM_CAST(x) ((int)(x))
|
# else
|
# else
|
# define SWITCH_ENUM_CAST(x) (x)
|
# define SWITCH_ENUM_CAST(x) (x)
|
# endif
|
# endif
|
|
|
/* How many characters in the character set. */
|
/* How many characters in the character set. */
|
# define CHAR_SET_SIZE 256
|
# define CHAR_SET_SIZE 256
|
|
|
/* GDB LOCAL: define _REGEX_RE_COMP to get BSD style re_comp and re_exec */
|
/* GDB LOCAL: define _REGEX_RE_COMP to get BSD style re_comp and re_exec */
|
#ifndef _REGEX_RE_COMP
|
#ifndef _REGEX_RE_COMP
|
#define _REGEX_RE_COMP
|
#define _REGEX_RE_COMP
|
#endif
|
#endif
|
|
|
# ifdef SYNTAX_TABLE
|
# ifdef SYNTAX_TABLE
|
|
|
extern char *re_syntax_table;
|
extern char *re_syntax_table;
|
|
|
# else /* not SYNTAX_TABLE */
|
# else /* not SYNTAX_TABLE */
|
|
|
static char re_syntax_table[CHAR_SET_SIZE];
|
static char re_syntax_table[CHAR_SET_SIZE];
|
|
|
static void
|
static void
|
init_syntax_once ()
|
init_syntax_once ()
|
{
|
{
|
register int c;
|
register int c;
|
static int done = 0;
|
static int done = 0;
|
|
|
if (done)
|
if (done)
|
return;
|
return;
|
|
|
bzero (re_syntax_table, sizeof re_syntax_table);
|
bzero (re_syntax_table, sizeof re_syntax_table);
|
|
|
for (c = 'a'; c <= 'z'; c++)
|
for (c = 'a'; c <= 'z'; c++)
|
re_syntax_table[c] = Sword;
|
re_syntax_table[c] = Sword;
|
|
|
for (c = 'A'; c <= 'Z'; c++)
|
for (c = 'A'; c <= 'Z'; c++)
|
re_syntax_table[c] = Sword;
|
re_syntax_table[c] = Sword;
|
|
|
for (c = '0'; c <= '9'; c++)
|
for (c = '0'; c <= '9'; c++)
|
re_syntax_table[c] = Sword;
|
re_syntax_table[c] = Sword;
|
|
|
re_syntax_table['_'] = Sword;
|
re_syntax_table['_'] = Sword;
|
|
|
done = 1;
|
done = 1;
|
}
|
}
|
|
|
# endif /* not SYNTAX_TABLE */
|
# endif /* not SYNTAX_TABLE */
|
|
|
# define SYNTAX(c) re_syntax_table[c]
|
# define SYNTAX(c) re_syntax_table[c]
|
|
|
#endif /* not emacs */
|
#endif /* not emacs */
|
|
|
/* Get the interface, including the syntax bits. */
|
/* Get the interface, including the syntax bits. */
|
/* CYGNUS LOCAL: call it gnu-regex.h, not regex.h, to avoid name conflicts */
|
/* CYGNUS LOCAL: call it gnu-regex.h, not regex.h, to avoid name conflicts */
|
#include "gnu-regex.h"
|
#include "gnu-regex.h"
|
|
|
/* isalpha etc. are used for the character classes. */
|
/* isalpha etc. are used for the character classes. */
|
#include <ctype.h>
|
#include <ctype.h>
|
|
|
/* Jim Meyering writes:
|
/* Jim Meyering writes:
|
|
|
"... Some ctype macros are valid only for character codes that
|
"... Some ctype macros are valid only for character codes that
|
isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when
|
isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when
|
using /bin/cc or gcc but without giving an ansi option). So, all
|
using /bin/cc or gcc but without giving an ansi option). So, all
|
ctype uses should be through macros like ISPRINT... If
|
ctype uses should be through macros like ISPRINT... If
|
STDC_HEADERS is defined, then autoconf has verified that the ctype
|
STDC_HEADERS is defined, then autoconf has verified that the ctype
|
macros don't need to be guarded with references to isascii. ...
|
macros don't need to be guarded with references to isascii. ...
|
Defining isascii to 1 should let any compiler worth its salt
|
Defining isascii to 1 should let any compiler worth its salt
|
eliminate the && through constant folding."
|
eliminate the && through constant folding."
|
Solaris defines some of these symbols so we must undefine them first. */
|
Solaris defines some of these symbols so we must undefine them first. */
|
|
|
#undef ISASCII
|
#undef ISASCII
|
#if defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
|
#if defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
|
# define ISASCII(c) 1
|
# define ISASCII(c) 1
|
#else
|
#else
|
# define ISASCII(c) isascii(c)
|
# define ISASCII(c) isascii(c)
|
#endif
|
#endif
|
|
|
#ifdef isblank
|
#ifdef isblank
|
# define ISBLANK(c) (ISASCII (c) && isblank (c))
|
# define ISBLANK(c) (ISASCII (c) && isblank (c))
|
#else
|
#else
|
# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
|
# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
|
#endif
|
#endif
|
#ifdef isgraph
|
#ifdef isgraph
|
# define ISGRAPH(c) (ISASCII (c) && isgraph (c))
|
# define ISGRAPH(c) (ISASCII (c) && isgraph (c))
|
#else
|
#else
|
# define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
|
# define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
|
#endif
|
#endif
|
|
|
#undef ISPRINT
|
#undef ISPRINT
|
#define ISPRINT(c) (ISASCII (c) && isprint (c))
|
#define ISPRINT(c) (ISASCII (c) && isprint (c))
|
#define ISDIGIT(c) (ISASCII (c) && isdigit (c))
|
#define ISDIGIT(c) (ISASCII (c) && isdigit (c))
|
#define ISALNUM(c) (ISASCII (c) && isalnum (c))
|
#define ISALNUM(c) (ISASCII (c) && isalnum (c))
|
#define ISALPHA(c) (ISASCII (c) && isalpha (c))
|
#define ISALPHA(c) (ISASCII (c) && isalpha (c))
|
#define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
|
#define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
|
#define ISLOWER(c) (ISASCII (c) && islower (c))
|
#define ISLOWER(c) (ISASCII (c) && islower (c))
|
#define ISPUNCT(c) (ISASCII (c) && ispunct (c))
|
#define ISPUNCT(c) (ISASCII (c) && ispunct (c))
|
#define ISSPACE(c) (ISASCII (c) && isspace (c))
|
#define ISSPACE(c) (ISASCII (c) && isspace (c))
|
#define ISUPPER(c) (ISASCII (c) && isupper (c))
|
#define ISUPPER(c) (ISASCII (c) && isupper (c))
|
#define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
|
#define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
|
|
|
#ifndef NULL
|
#ifndef NULL
|
# define NULL (void *)0
|
# define NULL (void *)0
|
#endif
|
#endif
|
|
|
/* We remove any previous definition of `SIGN_EXTEND_CHAR',
|
/* We remove any previous definition of `SIGN_EXTEND_CHAR',
|
since ours (we hope) works properly with all combinations of
|
since ours (we hope) works properly with all combinations of
|
machines, compilers, `char' and `unsigned char' argument types.
|
machines, compilers, `char' and `unsigned char' argument types.
|
(Per Bothner suggested the basic approach.) */
|
(Per Bothner suggested the basic approach.) */
|
#undef SIGN_EXTEND_CHAR
|
#undef SIGN_EXTEND_CHAR
|
#if __STDC__
|
#if __STDC__
|
# define SIGN_EXTEND_CHAR(c) ((signed char) (c))
|
# define SIGN_EXTEND_CHAR(c) ((signed char) (c))
|
#else /* not __STDC__ */
|
#else /* not __STDC__ */
|
/* As in Harbison and Steele. */
|
/* As in Harbison and Steele. */
|
# define SIGN_EXTEND_CHAR(c) ((((unsigned char) (c)) ^ 128) - 128)
|
# define SIGN_EXTEND_CHAR(c) ((((unsigned char) (c)) ^ 128) - 128)
|
#endif
|
#endif
|
|
|
/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we
|
/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we
|
use `alloca' instead of `malloc'. This is because using malloc in
|
use `alloca' instead of `malloc'. This is because using malloc in
|
re_search* or re_match* could cause memory leaks when C-g is used in
|
re_search* or re_match* could cause memory leaks when C-g is used in
|
Emacs; also, malloc is slower and causes storage fragmentation. On
|
Emacs; also, malloc is slower and causes storage fragmentation. On
|
the other hand, malloc is more portable, and easier to debug.
|
the other hand, malloc is more portable, and easier to debug.
|
|
|
Because we sometimes use alloca, some routines have to be macros,
|
Because we sometimes use alloca, some routines have to be macros,
|
not functions -- `alloca'-allocated space disappears at the end of the
|
not functions -- `alloca'-allocated space disappears at the end of the
|
function it is called in. */
|
function it is called in. */
|
|
|
#ifdef REGEX_MALLOC
|
#ifdef REGEX_MALLOC
|
|
|
# define REGEX_ALLOCATE malloc
|
# define REGEX_ALLOCATE malloc
|
# define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize)
|
# define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize)
|
# define REGEX_FREE free
|
# define REGEX_FREE free
|
|
|
#else /* not REGEX_MALLOC */
|
#else /* not REGEX_MALLOC */
|
|
|
/* Emacs already defines alloca, sometimes. */
|
/* Emacs already defines alloca, sometimes. */
|
# ifndef alloca
|
# ifndef alloca
|
|
|
/* Make alloca work the best possible way. */
|
/* Make alloca work the best possible way. */
|
# ifdef __GNUC__
|
# ifdef __GNUC__
|
# define alloca __builtin_alloca
|
# define alloca __builtin_alloca
|
# else /* not __GNUC__ */
|
# else /* not __GNUC__ */
|
# if HAVE_ALLOCA_H
|
# if HAVE_ALLOCA_H
|
# include <alloca.h>
|
# include <alloca.h>
|
# endif /* HAVE_ALLOCA_H */
|
# endif /* HAVE_ALLOCA_H */
|
# endif /* not __GNUC__ */
|
# endif /* not __GNUC__ */
|
|
|
# endif /* not alloca */
|
# endif /* not alloca */
|
|
|
# define REGEX_ALLOCATE alloca
|
# define REGEX_ALLOCATE alloca
|
|
|
/* Assumes a `char *destination' variable. */
|
/* Assumes a `char *destination' variable. */
|
# define REGEX_REALLOCATE(source, osize, nsize) \
|
# define REGEX_REALLOCATE(source, osize, nsize) \
|
(destination = (char *) alloca (nsize), \
|
(destination = (char *) alloca (nsize), \
|
memcpy (destination, source, osize))
|
memcpy (destination, source, osize))
|
|
|
/* No need to do anything to free, after alloca. */
|
/* No need to do anything to free, after alloca. */
|
# define REGEX_FREE(arg) ((void)0) /* Do nothing! But inhibit gcc warning. */
|
# define REGEX_FREE(arg) ((void)0) /* Do nothing! But inhibit gcc warning. */
|
|
|
#endif /* not REGEX_MALLOC */
|
#endif /* not REGEX_MALLOC */
|
|
|
/* Define how to allocate the failure stack. */
|
/* Define how to allocate the failure stack. */
|
|
|
#if defined REL_ALLOC && defined REGEX_MALLOC
|
#if defined REL_ALLOC && defined REGEX_MALLOC
|
|
|
# define REGEX_ALLOCATE_STACK(size) \
|
# define REGEX_ALLOCATE_STACK(size) \
|
r_alloc (&failure_stack_ptr, (size))
|
r_alloc (&failure_stack_ptr, (size))
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
|
r_re_alloc (&failure_stack_ptr, (nsize))
|
r_re_alloc (&failure_stack_ptr, (nsize))
|
# define REGEX_FREE_STACK(ptr) \
|
# define REGEX_FREE_STACK(ptr) \
|
r_alloc_free (&failure_stack_ptr)
|
r_alloc_free (&failure_stack_ptr)
|
|
|
#else /* not using relocating allocator */
|
#else /* not using relocating allocator */
|
|
|
# ifdef REGEX_MALLOC
|
# ifdef REGEX_MALLOC
|
|
|
# define REGEX_ALLOCATE_STACK malloc
|
# define REGEX_ALLOCATE_STACK malloc
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) realloc (source, nsize)
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) realloc (source, nsize)
|
# define REGEX_FREE_STACK free
|
# define REGEX_FREE_STACK free
|
|
|
# else /* not REGEX_MALLOC */
|
# else /* not REGEX_MALLOC */
|
|
|
# define REGEX_ALLOCATE_STACK alloca
|
# define REGEX_ALLOCATE_STACK alloca
|
|
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
|
# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
|
REGEX_REALLOCATE (source, osize, nsize)
|
REGEX_REALLOCATE (source, osize, nsize)
|
/* No need to explicitly free anything. */
|
/* No need to explicitly free anything. */
|
# define REGEX_FREE_STACK(arg)
|
# define REGEX_FREE_STACK(arg)
|
|
|
# endif /* not REGEX_MALLOC */
|
# endif /* not REGEX_MALLOC */
|
#endif /* not using relocating allocator */
|
#endif /* not using relocating allocator */
|
|
|
|
|
/* True if `size1' is non-NULL and PTR is pointing anywhere inside
|
/* True if `size1' is non-NULL and PTR is pointing anywhere inside
|
`string1' or just past its end. This works if PTR is NULL, which is
|
`string1' or just past its end. This works if PTR is NULL, which is
|
a good thing. */
|
a good thing. */
|
#define FIRST_STRING_P(ptr) \
|
#define FIRST_STRING_P(ptr) \
|
(size1 && string1 <= (ptr) && (ptr) <= string1 + size1)
|
(size1 && string1 <= (ptr) && (ptr) <= string1 + size1)
|
|
|
/* (Re)Allocate N items of type T using malloc, or fail. */
|
/* (Re)Allocate N items of type T using malloc, or fail. */
|
#define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t)))
|
#define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t)))
|
#define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t)))
|
#define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t)))
|
#define RETALLOC_IF(addr, n, t) \
|
#define RETALLOC_IF(addr, n, t) \
|
if (addr) RETALLOC((addr), (n), t); else (addr) = TALLOC ((n), t)
|
if (addr) RETALLOC((addr), (n), t); else (addr) = TALLOC ((n), t)
|
#define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t)))
|
#define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t)))
|
|
|
#define BYTEWIDTH 8 /* In bits. */
|
#define BYTEWIDTH 8 /* In bits. */
|
|
|
#define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
|
#define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
|
|
|
#undef MAX
|
#undef MAX
|
#undef MIN
|
#undef MIN
|
#define MAX(a, b) ((a) > (b) ? (a) : (b))
|
#define MAX(a, b) ((a) > (b) ? (a) : (b))
|
#define MIN(a, b) ((a) < (b) ? (a) : (b))
|
#define MIN(a, b) ((a) < (b) ? (a) : (b))
|
|
|
typedef char boolean;
|
typedef char boolean;
|
#define false 0
|
#define false 0
|
#define true 1
|
#define true 1
|
|
|
static int re_match_2_internal PARAMS ((struct re_pattern_buffer *bufp,
|
static int re_match_2_internal PARAMS ((struct re_pattern_buffer *bufp,
|
const char *string1, int size1,
|
const char *string1, int size1,
|
const char *string2, int size2,
|
const char *string2, int size2,
|
int pos,
|
int pos,
|
struct re_registers *regs,
|
struct re_registers *regs,
|
int stop));
|
int stop));
|
|
|
/* These are the command codes that appear in compiled regular
|
/* These are the command codes that appear in compiled regular
|
expressions. Some opcodes are followed by argument bytes. A
|
expressions. Some opcodes are followed by argument bytes. A
|
command code can specify any interpretation whatsoever for its
|
command code can specify any interpretation whatsoever for its
|
arguments. Zero bytes may appear in the compiled regular expression. */
|
arguments. Zero bytes may appear in the compiled regular expression. */
|
|
|
typedef enum
|
typedef enum
|
{
|
{
|
no_op = 0,
|
no_op = 0,
|
|
|
/* Succeed right away--no more backtracking. */
|
/* Succeed right away--no more backtracking. */
|
succeed,
|
succeed,
|
|
|
/* Followed by one byte giving n, then by n literal bytes. */
|
/* Followed by one byte giving n, then by n literal bytes. */
|
exactn,
|
exactn,
|
|
|
/* Matches any (more or less) character. */
|
/* Matches any (more or less) character. */
|
anychar,
|
anychar,
|
|
|
/* Matches any one char belonging to specified set. First
|
/* Matches any one char belonging to specified set. First
|
following byte is number of bitmap bytes. Then come bytes
|
following byte is number of bitmap bytes. Then come bytes
|
for a bitmap saying which chars are in. Bits in each byte
|
for a bitmap saying which chars are in. Bits in each byte
|
are ordered low-bit-first. A character is in the set if its
|
are ordered low-bit-first. A character is in the set if its
|
bit is 1. A character too large to have a bit in the map is
|
bit is 1. A character too large to have a bit in the map is
|
automatically not in the set. */
|
automatically not in the set. */
|
charset,
|
charset,
|
|
|
/* Same parameters as charset, but match any character that is
|
/* Same parameters as charset, but match any character that is
|
not one of those specified. */
|
not one of those specified. */
|
charset_not,
|
charset_not,
|
|
|
/* Start remembering the text that is matched, for storing in a
|
/* Start remembering the text that is matched, for storing in a
|
register. Followed by one byte with the register number, in
|
register. Followed by one byte with the register number, in
|
the range 0 to one less than the pattern buffer's re_nsub
|
the range 0 to one less than the pattern buffer's re_nsub
|
field. Then followed by one byte with the number of groups
|
field. Then followed by one byte with the number of groups
|
inner to this one. (This last has to be part of the
|
inner to this one. (This last has to be part of the
|
start_memory only because we need it in the on_failure_jump
|
start_memory only because we need it in the on_failure_jump
|
of re_match_2.) */
|
of re_match_2.) */
|
start_memory,
|
start_memory,
|
|
|
/* Stop remembering the text that is matched and store it in a
|
/* Stop remembering the text that is matched and store it in a
|
memory register. Followed by one byte with the register
|
memory register. Followed by one byte with the register
|
number, in the range 0 to one less than `re_nsub' in the
|
number, in the range 0 to one less than `re_nsub' in the
|
pattern buffer, and one byte with the number of inner groups,
|
pattern buffer, and one byte with the number of inner groups,
|
just like `start_memory'. (We need the number of inner
|
just like `start_memory'. (We need the number of inner
|
groups here because we don't have any easy way of finding the
|
groups here because we don't have any easy way of finding the
|
corresponding start_memory when we're at a stop_memory.) */
|
corresponding start_memory when we're at a stop_memory.) */
|
stop_memory,
|
stop_memory,
|
|
|
/* Match a duplicate of something remembered. Followed by one
|
/* Match a duplicate of something remembered. Followed by one
|
byte containing the register number. */
|
byte containing the register number. */
|
duplicate,
|
duplicate,
|
|
|
/* Fail unless at beginning of line. */
|
/* Fail unless at beginning of line. */
|
begline,
|
begline,
|
|
|
/* Fail unless at end of line. */
|
/* Fail unless at end of line. */
|
endline,
|
endline,
|
|
|
/* Succeeds if at beginning of buffer (if emacs) or at beginning
|
/* Succeeds if at beginning of buffer (if emacs) or at beginning
|
of string to be matched (if not). */
|
of string to be matched (if not). */
|
begbuf,
|
begbuf,
|
|
|
/* Analogously, for end of buffer/string. */
|
/* Analogously, for end of buffer/string. */
|
endbuf,
|
endbuf,
|
|
|
/* Followed by two byte relative address to which to jump. */
|
/* Followed by two byte relative address to which to jump. */
|
jump,
|
jump,
|
|
|
/* Same as jump, but marks the end of an alternative. */
|
/* Same as jump, but marks the end of an alternative. */
|
jump_past_alt,
|
jump_past_alt,
|
|
|
/* Followed by two-byte relative address of place to resume at
|
/* Followed by two-byte relative address of place to resume at
|
in case of failure. */
|
in case of failure. */
|
on_failure_jump,
|
on_failure_jump,
|
|
|
/* Like on_failure_jump, but pushes a placeholder instead of the
|
/* Like on_failure_jump, but pushes a placeholder instead of the
|
current string position when executed. */
|
current string position when executed. */
|
on_failure_keep_string_jump,
|
on_failure_keep_string_jump,
|
|
|
/* Throw away latest failure point and then jump to following
|
/* Throw away latest failure point and then jump to following
|
two-byte relative address. */
|
two-byte relative address. */
|
pop_failure_jump,
|
pop_failure_jump,
|
|
|
/* Change to pop_failure_jump if know won't have to backtrack to
|
/* Change to pop_failure_jump if know won't have to backtrack to
|
match; otherwise change to jump. This is used to jump
|
match; otherwise change to jump. This is used to jump
|
back to the beginning of a repeat. If what follows this jump
|
back to the beginning of a repeat. If what follows this jump
|
clearly won't match what the repeat does, such that we can be
|
clearly won't match what the repeat does, such that we can be
|
sure that there is no use backtracking out of repetitions
|
sure that there is no use backtracking out of repetitions
|
already matched, then we change it to a pop_failure_jump.
|
already matched, then we change it to a pop_failure_jump.
|
Followed by two-byte address. */
|
Followed by two-byte address. */
|
maybe_pop_jump,
|
maybe_pop_jump,
|
|
|
/* Jump to following two-byte address, and push a dummy failure
|
/* Jump to following two-byte address, and push a dummy failure
|
point. This failure point will be thrown away if an attempt
|
point. This failure point will be thrown away if an attempt
|
is made to use it for a failure. A `+' construct makes this
|
is made to use it for a failure. A `+' construct makes this
|
before the first repeat. Also used as an intermediary kind
|
before the first repeat. Also used as an intermediary kind
|
of jump when compiling an alternative. */
|
of jump when compiling an alternative. */
|
dummy_failure_jump,
|
dummy_failure_jump,
|
|
|
/* Push a dummy failure point and continue. Used at the end of
|
/* Push a dummy failure point and continue. Used at the end of
|
alternatives. */
|
alternatives. */
|
push_dummy_failure,
|
push_dummy_failure,
|
|
|
/* Followed by two-byte relative address and two-byte number n.
|
/* Followed by two-byte relative address and two-byte number n.
|
After matching N times, jump to the address upon failure. */
|
After matching N times, jump to the address upon failure. */
|
succeed_n,
|
succeed_n,
|
|
|
/* Followed by two-byte relative address, and two-byte number n.
|
/* Followed by two-byte relative address, and two-byte number n.
|
Jump to the address N times, then fail. */
|
Jump to the address N times, then fail. */
|
jump_n,
|
jump_n,
|
|
|
/* Set the following two-byte relative address to the
|
/* Set the following two-byte relative address to the
|
subsequent two-byte number. The address *includes* the two
|
subsequent two-byte number. The address *includes* the two
|
bytes of number. */
|
bytes of number. */
|
set_number_at,
|
set_number_at,
|
|
|
wordchar, /* Matches any word-constituent character. */
|
wordchar, /* Matches any word-constituent character. */
|
notwordchar, /* Matches any char that is not a word-constituent. */
|
notwordchar, /* Matches any char that is not a word-constituent. */
|
|
|
wordbeg, /* Succeeds if at word beginning. */
|
wordbeg, /* Succeeds if at word beginning. */
|
wordend, /* Succeeds if at word end. */
|
wordend, /* Succeeds if at word end. */
|
|
|
wordbound, /* Succeeds if at a word boundary. */
|
wordbound, /* Succeeds if at a word boundary. */
|
notwordbound /* Succeeds if not at a word boundary. */
|
notwordbound /* Succeeds if not at a word boundary. */
|
|
|
#ifdef emacs
|
#ifdef emacs
|
,before_dot, /* Succeeds if before point. */
|
,before_dot, /* Succeeds if before point. */
|
at_dot, /* Succeeds if at point. */
|
at_dot, /* Succeeds if at point. */
|
after_dot, /* Succeeds if after point. */
|
after_dot, /* Succeeds if after point. */
|
|
|
/* Matches any character whose syntax is specified. Followed by
|
/* Matches any character whose syntax is specified. Followed by
|
a byte which contains a syntax code, e.g., Sword. */
|
a byte which contains a syntax code, e.g., Sword. */
|
syntaxspec,
|
syntaxspec,
|
|
|
/* Matches any character whose syntax is not that specified. */
|
/* Matches any character whose syntax is not that specified. */
|
notsyntaxspec
|
notsyntaxspec
|
#endif /* emacs */
|
#endif /* emacs */
|
} re_opcode_t;
|
} re_opcode_t;
|
|
|
/* Common operations on the compiled pattern. */
|
/* Common operations on the compiled pattern. */
|
|
|
/* Store NUMBER in two contiguous bytes starting at DESTINATION. */
|
/* Store NUMBER in two contiguous bytes starting at DESTINATION. */
|
|
|
#define STORE_NUMBER(destination, number) \
|
#define STORE_NUMBER(destination, number) \
|
do { \
|
do { \
|
(destination)[0] = (number) & 0377; \
|
(destination)[0] = (number) & 0377; \
|
(destination)[1] = (number) >> 8; \
|
(destination)[1] = (number) >> 8; \
|
} while (0)
|
} while (0)
|
|
|
/* Same as STORE_NUMBER, except increment DESTINATION to
|
/* Same as STORE_NUMBER, except increment DESTINATION to
|
the byte after where the number is stored. Therefore, DESTINATION
|
the byte after where the number is stored. Therefore, DESTINATION
|
must be an lvalue. */
|
must be an lvalue. */
|
|
|
#define STORE_NUMBER_AND_INCR(destination, number) \
|
#define STORE_NUMBER_AND_INCR(destination, number) \
|
do { \
|
do { \
|
STORE_NUMBER (destination, number); \
|
STORE_NUMBER (destination, number); \
|
(destination) += 2; \
|
(destination) += 2; \
|
} while (0)
|
} while (0)
|
|
|
/* Put into DESTINATION a number stored in two contiguous bytes starting
|
/* Put into DESTINATION a number stored in two contiguous bytes starting
|
at SOURCE. */
|
at SOURCE. */
|
|
|
#define EXTRACT_NUMBER(destination, source) \
|
#define EXTRACT_NUMBER(destination, source) \
|
do { \
|
do { \
|
(destination) = *(source) & 0377; \
|
(destination) = *(source) & 0377; \
|
(destination) += SIGN_EXTEND_CHAR (*((source) + 1)) << 8; \
|
(destination) += SIGN_EXTEND_CHAR (*((source) + 1)) << 8; \
|
} while (0)
|
} while (0)
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
static void extract_number _RE_ARGS ((int *dest, unsigned char *source));
|
static void extract_number _RE_ARGS ((int *dest, unsigned char *source));
|
static void
|
static void
|
extract_number (dest, source)
|
extract_number (dest, source)
|
int *dest;
|
int *dest;
|
unsigned char *source;
|
unsigned char *source;
|
{
|
{
|
int temp = SIGN_EXTEND_CHAR (*(source + 1));
|
int temp = SIGN_EXTEND_CHAR (*(source + 1));
|
*dest = *source & 0377;
|
*dest = *source & 0377;
|
*dest += temp << 8;
|
*dest += temp << 8;
|
}
|
}
|
|
|
# ifndef EXTRACT_MACROS /* To debug the macros. */
|
# ifndef EXTRACT_MACROS /* To debug the macros. */
|
# undef EXTRACT_NUMBER
|
# undef EXTRACT_NUMBER
|
# define EXTRACT_NUMBER(dest, src) extract_number (&dest, src)
|
# define EXTRACT_NUMBER(dest, src) extract_number (&dest, src)
|
# endif /* not EXTRACT_MACROS */
|
# endif /* not EXTRACT_MACROS */
|
|
|
#endif /* DEBUG */
|
#endif /* DEBUG */
|
|
|
/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number.
|
/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number.
|
SOURCE must be an lvalue. */
|
SOURCE must be an lvalue. */
|
|
|
#define EXTRACT_NUMBER_AND_INCR(destination, source) \
|
#define EXTRACT_NUMBER_AND_INCR(destination, source) \
|
do { \
|
do { \
|
EXTRACT_NUMBER (destination, source); \
|
EXTRACT_NUMBER (destination, source); \
|
(source) += 2; \
|
(source) += 2; \
|
} while (0)
|
} while (0)
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
static void extract_number_and_incr _RE_ARGS ((int *destination,
|
static void extract_number_and_incr _RE_ARGS ((int *destination,
|
unsigned char **source));
|
unsigned char **source));
|
static void
|
static void
|
extract_number_and_incr (destination, source)
|
extract_number_and_incr (destination, source)
|
int *destination;
|
int *destination;
|
unsigned char **source;
|
unsigned char **source;
|
{
|
{
|
extract_number (destination, *source);
|
extract_number (destination, *source);
|
*source += 2;
|
*source += 2;
|
}
|
}
|
|
|
# ifndef EXTRACT_MACROS
|
# ifndef EXTRACT_MACROS
|
# undef EXTRACT_NUMBER_AND_INCR
|
# undef EXTRACT_NUMBER_AND_INCR
|
# define EXTRACT_NUMBER_AND_INCR(dest, src) \
|
# define EXTRACT_NUMBER_AND_INCR(dest, src) \
|
extract_number_and_incr (&dest, &src)
|
extract_number_and_incr (&dest, &src)
|
# endif /* not EXTRACT_MACROS */
|
# endif /* not EXTRACT_MACROS */
|
|
|
#endif /* DEBUG */
|
#endif /* DEBUG */
|
|
|
/* If DEBUG is defined, Regex prints many voluminous messages about what
|
/* If DEBUG is defined, Regex prints many voluminous messages about what
|
it is doing (if the variable `debug' is nonzero). If linked with the
|
it is doing (if the variable `debug' is nonzero). If linked with the
|
main program in `iregex.c', you can enter patterns and strings
|
main program in `iregex.c', you can enter patterns and strings
|
interactively. And if linked with the main program in `main.c' and
|
interactively. And if linked with the main program in `main.c' and
|
the other test files, you can run the already-written tests. */
|
the other test files, you can run the already-written tests. */
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
|
|
/* We use standard I/O for debugging. */
|
/* We use standard I/O for debugging. */
|
# include <stdio.h>
|
# include <stdio.h>
|
|
|
/* It is useful to test things that ``must'' be true when debugging. */
|
/* It is useful to test things that ``must'' be true when debugging. */
|
# include <assert.h>
|
# include <assert.h>
|
|
|
static int debug = 0;
|
static int debug = 0;
|
|
|
# define DEBUG_STATEMENT(e) e
|
# define DEBUG_STATEMENT(e) e
|
# define DEBUG_PRINT1(x) if (debug) printf (x)
|
# define DEBUG_PRINT1(x) if (debug) printf (x)
|
# define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2)
|
# define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2)
|
# define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3)
|
# define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3)
|
# define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4)
|
# define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4)
|
# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \
|
# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \
|
if (debug) print_partial_compiled_pattern (s, e)
|
if (debug) print_partial_compiled_pattern (s, e)
|
# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \
|
# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \
|
if (debug) print_double_string (w, s1, sz1, s2, sz2)
|
if (debug) print_double_string (w, s1, sz1, s2, sz2)
|
|
|
|
|
/* Print the fastmap in human-readable form. */
|
/* Print the fastmap in human-readable form. */
|
|
|
void
|
void
|
print_fastmap (fastmap)
|
print_fastmap (fastmap)
|
char *fastmap;
|
char *fastmap;
|
{
|
{
|
unsigned was_a_range = 0;
|
unsigned was_a_range = 0;
|
unsigned i = 0;
|
unsigned i = 0;
|
|
|
while (i < (1 << BYTEWIDTH))
|
while (i < (1 << BYTEWIDTH))
|
{
|
{
|
if (fastmap[i++])
|
if (fastmap[i++])
|
{
|
{
|
was_a_range = 0;
|
was_a_range = 0;
|
putchar (i - 1);
|
putchar (i - 1);
|
while (i < (1 << BYTEWIDTH) && fastmap[i])
|
while (i < (1 << BYTEWIDTH) && fastmap[i])
|
{
|
{
|
was_a_range = 1;
|
was_a_range = 1;
|
i++;
|
i++;
|
}
|
}
|
if (was_a_range)
|
if (was_a_range)
|
{
|
{
|
printf ("-");
|
printf ("-");
|
putchar (i - 1);
|
putchar (i - 1);
|
}
|
}
|
}
|
}
|
}
|
}
|
putchar ('\n');
|
putchar ('\n');
|
}
|
}
|
|
|
|
|
/* Print a compiled pattern string in human-readable form, starting at
|
/* Print a compiled pattern string in human-readable form, starting at
|
the START pointer into it and ending just before the pointer END. */
|
the START pointer into it and ending just before the pointer END. */
|
|
|
void
|
void
|
print_partial_compiled_pattern (start, end)
|
print_partial_compiled_pattern (start, end)
|
unsigned char *start;
|
unsigned char *start;
|
unsigned char *end;
|
unsigned char *end;
|
{
|
{
|
int mcnt, mcnt2;
|
int mcnt, mcnt2;
|
unsigned char *p1;
|
unsigned char *p1;
|
unsigned char *p = start;
|
unsigned char *p = start;
|
unsigned char *pend = end;
|
unsigned char *pend = end;
|
|
|
if (start == NULL)
|
if (start == NULL)
|
{
|
{
|
printf ("(null)\n");
|
printf ("(null)\n");
|
return;
|
return;
|
}
|
}
|
|
|
/* Loop over pattern commands. */
|
/* Loop over pattern commands. */
|
while (p < pend)
|
while (p < pend)
|
{
|
{
|
printf ("%d:\t", p - start);
|
printf ("%d:\t", p - start);
|
|
|
switch ((re_opcode_t) *p++)
|
switch ((re_opcode_t) *p++)
|
{
|
{
|
case no_op:
|
case no_op:
|
printf ("/no_op");
|
printf ("/no_op");
|
break;
|
break;
|
|
|
case exactn:
|
case exactn:
|
mcnt = *p++;
|
mcnt = *p++;
|
printf ("/exactn/%d", mcnt);
|
printf ("/exactn/%d", mcnt);
|
do
|
do
|
{
|
{
|
putchar ('/');
|
putchar ('/');
|
putchar (*p++);
|
putchar (*p++);
|
}
|
}
|
while (--mcnt);
|
while (--mcnt);
|
break;
|
break;
|
|
|
case start_memory:
|
case start_memory:
|
mcnt = *p++;
|
mcnt = *p++;
|
printf ("/start_memory/%d/%d", mcnt, *p++);
|
printf ("/start_memory/%d/%d", mcnt, *p++);
|
break;
|
break;
|
|
|
case stop_memory:
|
case stop_memory:
|
mcnt = *p++;
|
mcnt = *p++;
|
printf ("/stop_memory/%d/%d", mcnt, *p++);
|
printf ("/stop_memory/%d/%d", mcnt, *p++);
|
break;
|
break;
|
|
|
case duplicate:
|
case duplicate:
|
printf ("/duplicate/%d", *p++);
|
printf ("/duplicate/%d", *p++);
|
break;
|
break;
|
|
|
case anychar:
|
case anychar:
|
printf ("/anychar");
|
printf ("/anychar");
|
break;
|
break;
|
|
|
case charset:
|
case charset:
|
case charset_not:
|
case charset_not:
|
{
|
{
|
register int c, last = -100;
|
register int c, last = -100;
|
register int in_range = 0;
|
register int in_range = 0;
|
|
|
printf ("/charset [%s",
|
printf ("/charset [%s",
|
(re_opcode_t) *(p - 1) == charset_not ? "^" : "");
|
(re_opcode_t) *(p - 1) == charset_not ? "^" : "");
|
|
|
assert (p + *p < pend);
|
assert (p + *p < pend);
|
|
|
for (c = 0; c < 256; c++)
|
for (c = 0; c < 256; c++)
|
if (c / 8 < *p
|
if (c / 8 < *p
|
&& (p[1 + (c/8)] & (1 << (c % 8))))
|
&& (p[1 + (c/8)] & (1 << (c % 8))))
|
{
|
{
|
/* Are we starting a range? */
|
/* Are we starting a range? */
|
if (last + 1 == c && ! in_range)
|
if (last + 1 == c && ! in_range)
|
{
|
{
|
putchar ('-');
|
putchar ('-');
|
in_range = 1;
|
in_range = 1;
|
}
|
}
|
/* Have we broken a range? */
|
/* Have we broken a range? */
|
else if (last + 1 != c && in_range)
|
else if (last + 1 != c && in_range)
|
{
|
{
|
putchar (last);
|
putchar (last);
|
in_range = 0;
|
in_range = 0;
|
}
|
}
|
|
|
if (! in_range)
|
if (! in_range)
|
putchar (c);
|
putchar (c);
|
|
|
last = c;
|
last = c;
|
}
|
}
|
|
|
if (in_range)
|
if (in_range)
|
putchar (last);
|
putchar (last);
|
|
|
putchar (']');
|
putchar (']');
|
|
|
p += 1 + *p;
|
p += 1 + *p;
|
}
|
}
|
break;
|
break;
|
|
|
case begline:
|
case begline:
|
printf ("/begline");
|
printf ("/begline");
|
break;
|
break;
|
|
|
case endline:
|
case endline:
|
printf ("/endline");
|
printf ("/endline");
|
break;
|
break;
|
|
|
case on_failure_jump:
|
case on_failure_jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/on_failure_jump to %d", p + mcnt - start);
|
printf ("/on_failure_jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case on_failure_keep_string_jump:
|
case on_failure_keep_string_jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/on_failure_keep_string_jump to %d", p + mcnt - start);
|
printf ("/on_failure_keep_string_jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case dummy_failure_jump:
|
case dummy_failure_jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/dummy_failure_jump to %d", p + mcnt - start);
|
printf ("/dummy_failure_jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case push_dummy_failure:
|
case push_dummy_failure:
|
printf ("/push_dummy_failure");
|
printf ("/push_dummy_failure");
|
break;
|
break;
|
|
|
case maybe_pop_jump:
|
case maybe_pop_jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/maybe_pop_jump to %d", p + mcnt - start);
|
printf ("/maybe_pop_jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case pop_failure_jump:
|
case pop_failure_jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/pop_failure_jump to %d", p + mcnt - start);
|
printf ("/pop_failure_jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case jump_past_alt:
|
case jump_past_alt:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/jump_past_alt to %d", p + mcnt - start);
|
printf ("/jump_past_alt to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case jump:
|
case jump:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
printf ("/jump to %d", p + mcnt - start);
|
printf ("/jump to %d", p + mcnt - start);
|
break;
|
break;
|
|
|
case succeed_n:
|
case succeed_n:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
p1 = p + mcnt;
|
p1 = p + mcnt;
|
extract_number_and_incr (&mcnt2, &p);
|
extract_number_and_incr (&mcnt2, &p);
|
printf ("/succeed_n to %d, %d times", p1 - start, mcnt2);
|
printf ("/succeed_n to %d, %d times", p1 - start, mcnt2);
|
break;
|
break;
|
|
|
case jump_n:
|
case jump_n:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
p1 = p + mcnt;
|
p1 = p + mcnt;
|
extract_number_and_incr (&mcnt2, &p);
|
extract_number_and_incr (&mcnt2, &p);
|
printf ("/jump_n to %d, %d times", p1 - start, mcnt2);
|
printf ("/jump_n to %d, %d times", p1 - start, mcnt2);
|
break;
|
break;
|
|
|
case set_number_at:
|
case set_number_at:
|
extract_number_and_incr (&mcnt, &p);
|
extract_number_and_incr (&mcnt, &p);
|
p1 = p + mcnt;
|
p1 = p + mcnt;
|
extract_number_and_incr (&mcnt2, &p);
|
extract_number_and_incr (&mcnt2, &p);
|
printf ("/set_number_at location %d to %d", p1 - start, mcnt2);
|
printf ("/set_number_at location %d to %d", p1 - start, mcnt2);
|
break;
|
break;
|
|
|
case wordbound:
|
case wordbound:
|
printf ("/wordbound");
|
printf ("/wordbound");
|
break;
|
break;
|
|
|
case notwordbound:
|
case notwordbound:
|
printf ("/notwordbound");
|
printf ("/notwordbound");
|
break;
|
break;
|
|
|
case wordbeg:
|
case wordbeg:
|
printf ("/wordbeg");
|
printf ("/wordbeg");
|
break;
|
break;
|
|
|
case wordend:
|
case wordend:
|
printf ("/wordend");
|
printf ("/wordend");
|
|
|
# ifdef emacs
|
# ifdef emacs
|
case before_dot:
|
case before_dot:
|
printf ("/before_dot");
|
printf ("/before_dot");
|
break;
|
break;
|
|
|
case at_dot:
|
case at_dot:
|
printf ("/at_dot");
|
printf ("/at_dot");
|
break;
|
break;
|
|
|
case after_dot:
|
case after_dot:
|
printf ("/after_dot");
|
printf ("/after_dot");
|
break;
|
break;
|
|
|
case syntaxspec:
|
case syntaxspec:
|
printf ("/syntaxspec");
|
printf ("/syntaxspec");
|
mcnt = *p++;
|
mcnt = *p++;
|
printf ("/%d", mcnt);
|
printf ("/%d", mcnt);
|
break;
|
break;
|
|
|
case notsyntaxspec:
|
case notsyntaxspec:
|
printf ("/notsyntaxspec");
|
printf ("/notsyntaxspec");
|
mcnt = *p++;
|
mcnt = *p++;
|
printf ("/%d", mcnt);
|
printf ("/%d", mcnt);
|
break;
|
break;
|
# endif /* emacs */
|
# endif /* emacs */
|
|
|
case wordchar:
|
case wordchar:
|
printf ("/wordchar");
|
printf ("/wordchar");
|
break;
|
break;
|
|
|
case notwordchar:
|
case notwordchar:
|
printf ("/notwordchar");
|
printf ("/notwordchar");
|
break;
|
break;
|
|
|
case begbuf:
|
case begbuf:
|
printf ("/begbuf");
|
printf ("/begbuf");
|
break;
|
break;
|
|
|
case endbuf:
|
case endbuf:
|
printf ("/endbuf");
|
printf ("/endbuf");
|
break;
|
break;
|
|
|
default:
|
default:
|
printf ("?%d", *(p-1));
|
printf ("?%d", *(p-1));
|
}
|
}
|
|
|
putchar ('\n');
|
putchar ('\n');
|
}
|
}
|
|
|
printf ("%d:\tend of pattern.\n", p - start);
|
printf ("%d:\tend of pattern.\n", p - start);
|
}
|
}
|
|
|
|
|
void
|
void
|
print_compiled_pattern (bufp)
|
print_compiled_pattern (bufp)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
{
|
{
|
unsigned char *buffer = bufp->buffer;
|
unsigned char *buffer = bufp->buffer;
|
|
|
print_partial_compiled_pattern (buffer, buffer + bufp->used);
|
print_partial_compiled_pattern (buffer, buffer + bufp->used);
|
printf ("%ld bytes used/%ld bytes allocated.\n",
|
printf ("%ld bytes used/%ld bytes allocated.\n",
|
bufp->used, bufp->allocated);
|
bufp->used, bufp->allocated);
|
|
|
if (bufp->fastmap_accurate && bufp->fastmap)
|
if (bufp->fastmap_accurate && bufp->fastmap)
|
{
|
{
|
printf ("fastmap: ");
|
printf ("fastmap: ");
|
print_fastmap (bufp->fastmap);
|
print_fastmap (bufp->fastmap);
|
}
|
}
|
|
|
printf ("re_nsub: %d\t", bufp->re_nsub);
|
printf ("re_nsub: %d\t", bufp->re_nsub);
|
printf ("regs_alloc: %d\t", bufp->regs_allocated);
|
printf ("regs_alloc: %d\t", bufp->regs_allocated);
|
printf ("can_be_null: %d\t", bufp->can_be_null);
|
printf ("can_be_null: %d\t", bufp->can_be_null);
|
printf ("newline_anchor: %d\n", bufp->newline_anchor);
|
printf ("newline_anchor: %d\n", bufp->newline_anchor);
|
printf ("no_sub: %d\t", bufp->no_sub);
|
printf ("no_sub: %d\t", bufp->no_sub);
|
printf ("not_bol: %d\t", bufp->not_bol);
|
printf ("not_bol: %d\t", bufp->not_bol);
|
printf ("not_eol: %d\t", bufp->not_eol);
|
printf ("not_eol: %d\t", bufp->not_eol);
|
printf ("syntax: %lx\n", bufp->syntax);
|
printf ("syntax: %lx\n", bufp->syntax);
|
/* Perhaps we should print the translate table? */
|
/* Perhaps we should print the translate table? */
|
}
|
}
|
|
|
|
|
void
|
void
|
print_double_string (where, string1, size1, string2, size2)
|
print_double_string (where, string1, size1, string2, size2)
|
const char *where;
|
const char *where;
|
const char *string1;
|
const char *string1;
|
const char *string2;
|
const char *string2;
|
int size1;
|
int size1;
|
int size2;
|
int size2;
|
{
|
{
|
int this_char;
|
int this_char;
|
|
|
if (where == NULL)
|
if (where == NULL)
|
printf ("(null)");
|
printf ("(null)");
|
else
|
else
|
{
|
{
|
if (FIRST_STRING_P (where))
|
if (FIRST_STRING_P (where))
|
{
|
{
|
for (this_char = where - string1; this_char < size1; this_char++)
|
for (this_char = where - string1; this_char < size1; this_char++)
|
putchar (string1[this_char]);
|
putchar (string1[this_char]);
|
|
|
where = string2;
|
where = string2;
|
}
|
}
|
|
|
for (this_char = where - string2; this_char < size2; this_char++)
|
for (this_char = where - string2; this_char < size2; this_char++)
|
putchar (string2[this_char]);
|
putchar (string2[this_char]);
|
}
|
}
|
}
|
}
|
|
|
void
|
void
|
printchar (c)
|
printchar (c)
|
int c;
|
int c;
|
{
|
{
|
putc (c, stderr);
|
putc (c, stderr);
|
}
|
}
|
|
|
#else /* not DEBUG */
|
#else /* not DEBUG */
|
|
|
# undef assert
|
# undef assert
|
# define assert(e)
|
# define assert(e)
|
|
|
# define DEBUG_STATEMENT(e)
|
# define DEBUG_STATEMENT(e)
|
# define DEBUG_PRINT1(x)
|
# define DEBUG_PRINT1(x)
|
# define DEBUG_PRINT2(x1, x2)
|
# define DEBUG_PRINT2(x1, x2)
|
# define DEBUG_PRINT3(x1, x2, x3)
|
# define DEBUG_PRINT3(x1, x2, x3)
|
# define DEBUG_PRINT4(x1, x2, x3, x4)
|
# define DEBUG_PRINT4(x1, x2, x3, x4)
|
# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e)
|
# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e)
|
# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2)
|
# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2)
|
|
|
#endif /* not DEBUG */
|
#endif /* not DEBUG */
|
|
|
/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can
|
/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can
|
also be assigned to arbitrarily: each pattern buffer stores its own
|
also be assigned to arbitrarily: each pattern buffer stores its own
|
syntax, so it can be changed between regex compilations. */
|
syntax, so it can be changed between regex compilations. */
|
/* This has no initializer because initialized variables in Emacs
|
/* This has no initializer because initialized variables in Emacs
|
become read-only after dumping. */
|
become read-only after dumping. */
|
reg_syntax_t re_syntax_options;
|
reg_syntax_t re_syntax_options;
|
|
|
|
|
/* Specify the precise syntax of regexps for compilation. This provides
|
/* Specify the precise syntax of regexps for compilation. This provides
|
for compatibility for various utilities which historically have
|
for compatibility for various utilities which historically have
|
different, incompatible syntaxes.
|
different, incompatible syntaxes.
|
|
|
The argument SYNTAX is a bit mask comprised of the various bits
|
The argument SYNTAX is a bit mask comprised of the various bits
|
defined in gnu-regex.h. We return the old syntax. */
|
defined in gnu-regex.h. We return the old syntax. */
|
|
|
reg_syntax_t
|
reg_syntax_t
|
re_set_syntax (syntax)
|
re_set_syntax (syntax)
|
reg_syntax_t syntax;
|
reg_syntax_t syntax;
|
{
|
{
|
reg_syntax_t ret = re_syntax_options;
|
reg_syntax_t ret = re_syntax_options;
|
|
|
re_syntax_options = syntax;
|
re_syntax_options = syntax;
|
#ifdef DEBUG
|
#ifdef DEBUG
|
if (syntax & RE_DEBUG)
|
if (syntax & RE_DEBUG)
|
debug = 1;
|
debug = 1;
|
else if (debug) /* was on but now is not */
|
else if (debug) /* was on but now is not */
|
debug = 0;
|
debug = 0;
|
#endif /* DEBUG */
|
#endif /* DEBUG */
|
return ret;
|
return ret;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_set_syntax, re_set_syntax)
|
weak_alias (__re_set_syntax, re_set_syntax)
|
#endif
|
#endif
|
|
|
/* This table gives an error message for each of the error codes listed
|
/* This table gives an error message for each of the error codes listed
|
in gnu-regex.h. Obviously the order here has to be same as there.
|
in gnu-regex.h. Obviously the order here has to be same as there.
|
POSIX doesn't require that we do anything for REG_NOERROR,
|
POSIX doesn't require that we do anything for REG_NOERROR,
|
but why not be nice? */
|
but why not be nice? */
|
|
|
static const char *re_error_msgid[] =
|
static const char *re_error_msgid[] =
|
{
|
{
|
gettext_noop ("Success"), /* REG_NOERROR */
|
gettext_noop ("Success"), /* REG_NOERROR */
|
gettext_noop ("No match"), /* REG_NOMATCH */
|
gettext_noop ("No match"), /* REG_NOMATCH */
|
gettext_noop ("Invalid regular expression"), /* REG_BADPAT */
|
gettext_noop ("Invalid regular expression"), /* REG_BADPAT */
|
gettext_noop ("Invalid collation character"), /* REG_ECOLLATE */
|
gettext_noop ("Invalid collation character"), /* REG_ECOLLATE */
|
gettext_noop ("Invalid character class name"), /* REG_ECTYPE */
|
gettext_noop ("Invalid character class name"), /* REG_ECTYPE */
|
gettext_noop ("Trailing backslash"), /* REG_EESCAPE */
|
gettext_noop ("Trailing backslash"), /* REG_EESCAPE */
|
gettext_noop ("Invalid back reference"), /* REG_ESUBREG */
|
gettext_noop ("Invalid back reference"), /* REG_ESUBREG */
|
gettext_noop ("Unmatched [ or [^"), /* REG_EBRACK */
|
gettext_noop ("Unmatched [ or [^"), /* REG_EBRACK */
|
gettext_noop ("Unmatched ( or \\("), /* REG_EPAREN */
|
gettext_noop ("Unmatched ( or \\("), /* REG_EPAREN */
|
gettext_noop ("Unmatched \\{"), /* REG_EBRACE */
|
gettext_noop ("Unmatched \\{"), /* REG_EBRACE */
|
gettext_noop ("Invalid content of \\{\\}"), /* REG_BADBR */
|
gettext_noop ("Invalid content of \\{\\}"), /* REG_BADBR */
|
gettext_noop ("Invalid range end"), /* REG_ERANGE */
|
gettext_noop ("Invalid range end"), /* REG_ERANGE */
|
gettext_noop ("Memory exhausted"), /* REG_ESPACE */
|
gettext_noop ("Memory exhausted"), /* REG_ESPACE */
|
gettext_noop ("Invalid preceding regular expression"), /* REG_BADRPT */
|
gettext_noop ("Invalid preceding regular expression"), /* REG_BADRPT */
|
gettext_noop ("Premature end of regular expression"), /* REG_EEND */
|
gettext_noop ("Premature end of regular expression"), /* REG_EEND */
|
gettext_noop ("Regular expression too big"), /* REG_ESIZE */
|
gettext_noop ("Regular expression too big"), /* REG_ESIZE */
|
gettext_noop ("Unmatched ) or \\)"), /* REG_ERPAREN */
|
gettext_noop ("Unmatched ) or \\)"), /* REG_ERPAREN */
|
};
|
};
|
|
|
/* Avoiding alloca during matching, to placate r_alloc. */
|
/* Avoiding alloca during matching, to placate r_alloc. */
|
|
|
/* Define MATCH_MAY_ALLOCATE unless we need to make sure that the
|
/* Define MATCH_MAY_ALLOCATE unless we need to make sure that the
|
searching and matching functions should not call alloca. On some
|
searching and matching functions should not call alloca. On some
|
systems, alloca is implemented in terms of malloc, and if we're
|
systems, alloca is implemented in terms of malloc, and if we're
|
using the relocating allocator routines, then malloc could cause a
|
using the relocating allocator routines, then malloc could cause a
|
relocation, which might (if the strings being searched are in the
|
relocation, which might (if the strings being searched are in the
|
ralloc heap) shift the data out from underneath the regexp
|
ralloc heap) shift the data out from underneath the regexp
|
routines.
|
routines.
|
|
|
Here's another reason to avoid allocation: Emacs
|
Here's another reason to avoid allocation: Emacs
|
processes input from X in a signal handler; processing X input may
|
processes input from X in a signal handler; processing X input may
|
call malloc; if input arrives while a matching routine is calling
|
call malloc; if input arrives while a matching routine is calling
|
malloc, then we're scrod. But Emacs can't just block input while
|
malloc, then we're scrod. But Emacs can't just block input while
|
calling matching routines; then we don't notice interrupts when
|
calling matching routines; then we don't notice interrupts when
|
they come in. So, Emacs blocks input around all regexp calls
|
they come in. So, Emacs blocks input around all regexp calls
|
except the matching calls, which it leaves unprotected, in the
|
except the matching calls, which it leaves unprotected, in the
|
faith that they will not malloc. */
|
faith that they will not malloc. */
|
|
|
/* Normally, this is fine. */
|
/* Normally, this is fine. */
|
#define MATCH_MAY_ALLOCATE
|
#define MATCH_MAY_ALLOCATE
|
|
|
/* When using GNU C, we are not REALLY using the C alloca, no matter
|
/* When using GNU C, we are not REALLY using the C alloca, no matter
|
what config.h may say. So don't take precautions for it. */
|
what config.h may say. So don't take precautions for it. */
|
#ifdef __GNUC__
|
#ifdef __GNUC__
|
# undef C_ALLOCA
|
# undef C_ALLOCA
|
#endif
|
#endif
|
|
|
/* The match routines may not allocate if (1) they would do it with malloc
|
/* The match routines may not allocate if (1) they would do it with malloc
|
and (2) it's not safe for them to use malloc.
|
and (2) it's not safe for them to use malloc.
|
Note that if REL_ALLOC is defined, matching would not use malloc for the
|
Note that if REL_ALLOC is defined, matching would not use malloc for the
|
failure stack, but we would still use it for the register vectors;
|
failure stack, but we would still use it for the register vectors;
|
so REL_ALLOC should not affect this. */
|
so REL_ALLOC should not affect this. */
|
#if (defined C_ALLOCA || defined REGEX_MALLOC) && defined emacs
|
#if (defined C_ALLOCA || defined REGEX_MALLOC) && defined emacs
|
# undef MATCH_MAY_ALLOCATE
|
# undef MATCH_MAY_ALLOCATE
|
#endif
|
#endif
|
|
|
|
|
/* Failure stack declarations and macros; both re_compile_fastmap and
|
/* Failure stack declarations and macros; both re_compile_fastmap and
|
re_match_2 use a failure stack. These have to be macros because of
|
re_match_2 use a failure stack. These have to be macros because of
|
REGEX_ALLOCATE_STACK. */
|
REGEX_ALLOCATE_STACK. */
|
|
|
|
|
/* Number of failure points for which to initially allocate space
|
/* Number of failure points for which to initially allocate space
|
when matching. If this number is exceeded, we allocate more
|
when matching. If this number is exceeded, we allocate more
|
space, so it is not a hard limit. */
|
space, so it is not a hard limit. */
|
#ifndef INIT_FAILURE_ALLOC
|
#ifndef INIT_FAILURE_ALLOC
|
# define INIT_FAILURE_ALLOC 5
|
# define INIT_FAILURE_ALLOC 5
|
#endif
|
#endif
|
|
|
/* Roughly the maximum number of failure points on the stack. Would be
|
/* Roughly the maximum number of failure points on the stack. Would be
|
exactly that if always used MAX_FAILURE_ITEMS items each time we failed.
|
exactly that if always used MAX_FAILURE_ITEMS items each time we failed.
|
This is a variable only so users of regex can assign to it; we never
|
This is a variable only so users of regex can assign to it; we never
|
change it ourselves. */
|
change it ourselves. */
|
|
|
#ifdef INT_IS_16BIT
|
#ifdef INT_IS_16BIT
|
|
|
# if defined MATCH_MAY_ALLOCATE
|
# if defined MATCH_MAY_ALLOCATE
|
/* 4400 was enough to cause a crash on Alpha OSF/1,
|
/* 4400 was enough to cause a crash on Alpha OSF/1,
|
whose default stack limit is 2mb. */
|
whose default stack limit is 2mb. */
|
long int re_max_failures = 4000;
|
long int re_max_failures = 4000;
|
# else
|
# else
|
long int re_max_failures = 2000;
|
long int re_max_failures = 2000;
|
# endif
|
# endif
|
|
|
union fail_stack_elt
|
union fail_stack_elt
|
{
|
{
|
unsigned char *pointer;
|
unsigned char *pointer;
|
long int integer;
|
long int integer;
|
};
|
};
|
|
|
typedef union fail_stack_elt fail_stack_elt_t;
|
typedef union fail_stack_elt fail_stack_elt_t;
|
|
|
typedef struct
|
typedef struct
|
{
|
{
|
fail_stack_elt_t *stack;
|
fail_stack_elt_t *stack;
|
unsigned long int size;
|
unsigned long int size;
|
unsigned long int avail; /* Offset of next open position. */
|
unsigned long int avail; /* Offset of next open position. */
|
} fail_stack_type;
|
} fail_stack_type;
|
|
|
#else /* not INT_IS_16BIT */
|
#else /* not INT_IS_16BIT */
|
|
|
# if defined MATCH_MAY_ALLOCATE
|
# if defined MATCH_MAY_ALLOCATE
|
/* 4400 was enough to cause a crash on Alpha OSF/1,
|
/* 4400 was enough to cause a crash on Alpha OSF/1,
|
whose default stack limit is 2mb. */
|
whose default stack limit is 2mb. */
|
int re_max_failures = 20000;
|
int re_max_failures = 20000;
|
# else
|
# else
|
int re_max_failures = 2000;
|
int re_max_failures = 2000;
|
# endif
|
# endif
|
|
|
union fail_stack_elt
|
union fail_stack_elt
|
{
|
{
|
unsigned char *pointer;
|
unsigned char *pointer;
|
int integer;
|
int integer;
|
};
|
};
|
|
|
typedef union fail_stack_elt fail_stack_elt_t;
|
typedef union fail_stack_elt fail_stack_elt_t;
|
|
|
typedef struct
|
typedef struct
|
{
|
{
|
fail_stack_elt_t *stack;
|
fail_stack_elt_t *stack;
|
unsigned size;
|
unsigned size;
|
unsigned avail; /* Offset of next open position. */
|
unsigned avail; /* Offset of next open position. */
|
} fail_stack_type;
|
} fail_stack_type;
|
|
|
#endif /* INT_IS_16BIT */
|
#endif /* INT_IS_16BIT */
|
|
|
#define FAIL_STACK_EMPTY() (fail_stack.avail == 0)
|
#define FAIL_STACK_EMPTY() (fail_stack.avail == 0)
|
#define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0)
|
#define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0)
|
#define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size)
|
#define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size)
|
|
|
|
|
/* Define macros to initialize and free the failure stack.
|
/* Define macros to initialize and free the failure stack.
|
Do `return -2' if the alloc fails. */
|
Do `return -2' if the alloc fails. */
|
|
|
#ifdef MATCH_MAY_ALLOCATE
|
#ifdef MATCH_MAY_ALLOCATE
|
# define INIT_FAIL_STACK() \
|
# define INIT_FAIL_STACK() \
|
do { \
|
do { \
|
fail_stack.stack = (fail_stack_elt_t *) \
|
fail_stack.stack = (fail_stack_elt_t *) \
|
REGEX_ALLOCATE_STACK (INIT_FAILURE_ALLOC * sizeof (fail_stack_elt_t)); \
|
REGEX_ALLOCATE_STACK (INIT_FAILURE_ALLOC * sizeof (fail_stack_elt_t)); \
|
\
|
\
|
if (fail_stack.stack == NULL) \
|
if (fail_stack.stack == NULL) \
|
return -2; \
|
return -2; \
|
\
|
\
|
fail_stack.size = INIT_FAILURE_ALLOC; \
|
fail_stack.size = INIT_FAILURE_ALLOC; \
|
fail_stack.avail = 0; \
|
fail_stack.avail = 0; \
|
} while (0)
|
} while (0)
|
|
|
# define RESET_FAIL_STACK() REGEX_FREE_STACK (fail_stack.stack)
|
# define RESET_FAIL_STACK() REGEX_FREE_STACK (fail_stack.stack)
|
#else
|
#else
|
# define INIT_FAIL_STACK() \
|
# define INIT_FAIL_STACK() \
|
do { \
|
do { \
|
fail_stack.avail = 0; \
|
fail_stack.avail = 0; \
|
} while (0)
|
} while (0)
|
|
|
# define RESET_FAIL_STACK()
|
# define RESET_FAIL_STACK()
|
#endif
|
#endif
|
|
|
|
|
/* Double the size of FAIL_STACK, up to approximately `re_max_failures' items.
|
/* Double the size of FAIL_STACK, up to approximately `re_max_failures' items.
|
|
|
Return 1 if succeeds, and 0 if either ran out of memory
|
Return 1 if succeeds, and 0 if either ran out of memory
|
allocating space for it or it was already too large.
|
allocating space for it or it was already too large.
|
|
|
REGEX_REALLOCATE_STACK requires `destination' be declared. */
|
REGEX_REALLOCATE_STACK requires `destination' be declared. */
|
|
|
#define DOUBLE_FAIL_STACK(fail_stack) \
|
#define DOUBLE_FAIL_STACK(fail_stack) \
|
((fail_stack).size > (unsigned) (re_max_failures * MAX_FAILURE_ITEMS) \
|
((fail_stack).size > (unsigned) (re_max_failures * MAX_FAILURE_ITEMS) \
|
? 0 \
|
? 0 \
|
: ((fail_stack).stack = (fail_stack_elt_t *) \
|
: ((fail_stack).stack = (fail_stack_elt_t *) \
|
REGEX_REALLOCATE_STACK ((fail_stack).stack, \
|
REGEX_REALLOCATE_STACK ((fail_stack).stack, \
|
(fail_stack).size * sizeof (fail_stack_elt_t), \
|
(fail_stack).size * sizeof (fail_stack_elt_t), \
|
((fail_stack).size << 1) * sizeof (fail_stack_elt_t)), \
|
((fail_stack).size << 1) * sizeof (fail_stack_elt_t)), \
|
\
|
\
|
(fail_stack).stack == NULL \
|
(fail_stack).stack == NULL \
|
? 0 \
|
? 0 \
|
: ((fail_stack).size <<= 1, \
|
: ((fail_stack).size <<= 1, \
|
1)))
|
1)))
|
|
|
|
|
/* Push pointer POINTER on FAIL_STACK.
|
/* Push pointer POINTER on FAIL_STACK.
|
Return 1 if was able to do so and 0 if ran out of memory allocating
|
Return 1 if was able to do so and 0 if ran out of memory allocating
|
space to do so. */
|
space to do so. */
|
#define PUSH_PATTERN_OP(POINTER, FAIL_STACK) \
|
#define PUSH_PATTERN_OP(POINTER, FAIL_STACK) \
|
((FAIL_STACK_FULL () \
|
((FAIL_STACK_FULL () \
|
&& !DOUBLE_FAIL_STACK (FAIL_STACK)) \
|
&& !DOUBLE_FAIL_STACK (FAIL_STACK)) \
|
? 0 \
|
? 0 \
|
: ((FAIL_STACK).stack[(FAIL_STACK).avail++].pointer = POINTER, \
|
: ((FAIL_STACK).stack[(FAIL_STACK).avail++].pointer = POINTER, \
|
1))
|
1))
|
|
|
/* Push a pointer value onto the failure stack.
|
/* Push a pointer value onto the failure stack.
|
Assumes the variable `fail_stack'. Probably should only
|
Assumes the variable `fail_stack'. Probably should only
|
be called from within `PUSH_FAILURE_POINT'. */
|
be called from within `PUSH_FAILURE_POINT'. */
|
#define PUSH_FAILURE_POINTER(item) \
|
#define PUSH_FAILURE_POINTER(item) \
|
fail_stack.stack[fail_stack.avail++].pointer = (unsigned char *) (item)
|
fail_stack.stack[fail_stack.avail++].pointer = (unsigned char *) (item)
|
|
|
/* This pushes an integer-valued item onto the failure stack.
|
/* This pushes an integer-valued item onto the failure stack.
|
Assumes the variable `fail_stack'. Probably should only
|
Assumes the variable `fail_stack'. Probably should only
|
be called from within `PUSH_FAILURE_POINT'. */
|
be called from within `PUSH_FAILURE_POINT'. */
|
#define PUSH_FAILURE_INT(item) \
|
#define PUSH_FAILURE_INT(item) \
|
fail_stack.stack[fail_stack.avail++].integer = (item)
|
fail_stack.stack[fail_stack.avail++].integer = (item)
|
|
|
/* Push a fail_stack_elt_t value onto the failure stack.
|
/* Push a fail_stack_elt_t value onto the failure stack.
|
Assumes the variable `fail_stack'. Probably should only
|
Assumes the variable `fail_stack'. Probably should only
|
be called from within `PUSH_FAILURE_POINT'. */
|
be called from within `PUSH_FAILURE_POINT'. */
|
#define PUSH_FAILURE_ELT(item) \
|
#define PUSH_FAILURE_ELT(item) \
|
fail_stack.stack[fail_stack.avail++] = (item)
|
fail_stack.stack[fail_stack.avail++] = (item)
|
|
|
/* These three POP... operations complement the three PUSH... operations.
|
/* These three POP... operations complement the three PUSH... operations.
|
All assume that `fail_stack' is nonempty. */
|
All assume that `fail_stack' is nonempty. */
|
#define POP_FAILURE_POINTER() fail_stack.stack[--fail_stack.avail].pointer
|
#define POP_FAILURE_POINTER() fail_stack.stack[--fail_stack.avail].pointer
|
#define POP_FAILURE_INT() fail_stack.stack[--fail_stack.avail].integer
|
#define POP_FAILURE_INT() fail_stack.stack[--fail_stack.avail].integer
|
#define POP_FAILURE_ELT() fail_stack.stack[--fail_stack.avail]
|
#define POP_FAILURE_ELT() fail_stack.stack[--fail_stack.avail]
|
|
|
/* Used to omit pushing failure point id's when we're not debugging. */
|
/* Used to omit pushing failure point id's when we're not debugging. */
|
#ifdef DEBUG
|
#ifdef DEBUG
|
# define DEBUG_PUSH PUSH_FAILURE_INT
|
# define DEBUG_PUSH PUSH_FAILURE_INT
|
# define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_INT ()
|
# define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_INT ()
|
#else
|
#else
|
# define DEBUG_PUSH(item)
|
# define DEBUG_PUSH(item)
|
# define DEBUG_POP(item_addr)
|
# define DEBUG_POP(item_addr)
|
#endif
|
#endif
|
|
|
|
|
/* Push the information about the state we will need
|
/* Push the information about the state we will need
|
if we ever fail back to it.
|
if we ever fail back to it.
|
|
|
Requires variables fail_stack, regstart, regend, reg_info, and
|
Requires variables fail_stack, regstart, regend, reg_info, and
|
num_regs_pushed be declared. DOUBLE_FAIL_STACK requires `destination'
|
num_regs_pushed be declared. DOUBLE_FAIL_STACK requires `destination'
|
be declared.
|
be declared.
|
|
|
Does `return FAILURE_CODE' if runs out of memory. */
|
Does `return FAILURE_CODE' if runs out of memory. */
|
|
|
#define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \
|
#define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \
|
do { \
|
do { \
|
char *destination; \
|
char *destination; \
|
/* Must be int, so when we don't save any registers, the arithmetic \
|
/* Must be int, so when we don't save any registers, the arithmetic \
|
of 0 + -1 isn't done as unsigned. */ \
|
of 0 + -1 isn't done as unsigned. */ \
|
/* Can't be int, since there is not a shred of a guarantee that int \
|
/* Can't be int, since there is not a shred of a guarantee that int \
|
is wide enough to hold a value of something to which pointer can \
|
is wide enough to hold a value of something to which pointer can \
|
be assigned */ \
|
be assigned */ \
|
active_reg_t this_reg; \
|
active_reg_t this_reg; \
|
\
|
\
|
DEBUG_STATEMENT (failure_id++); \
|
DEBUG_STATEMENT (failure_id++); \
|
DEBUG_STATEMENT (nfailure_points_pushed++); \
|
DEBUG_STATEMENT (nfailure_points_pushed++); \
|
DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \
|
DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \
|
DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\
|
DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\
|
DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\
|
DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\
|
\
|
\
|
DEBUG_PRINT2 (" slots needed: %ld\n", NUM_FAILURE_ITEMS); \
|
DEBUG_PRINT2 (" slots needed: %ld\n", NUM_FAILURE_ITEMS); \
|
DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \
|
DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \
|
\
|
\
|
/* Ensure we have enough space allocated for what we will push. */ \
|
/* Ensure we have enough space allocated for what we will push. */ \
|
while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \
|
while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \
|
{ \
|
{ \
|
if (!DOUBLE_FAIL_STACK (fail_stack)) \
|
if (!DOUBLE_FAIL_STACK (fail_stack)) \
|
return failure_code; \
|
return failure_code; \
|
\
|
\
|
DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \
|
DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \
|
(fail_stack).size); \
|
(fail_stack).size); \
|
DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\
|
DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\
|
} \
|
} \
|
\
|
\
|
/* Push the info, starting with the registers. */ \
|
/* Push the info, starting with the registers. */ \
|
DEBUG_PRINT1 ("\n"); \
|
DEBUG_PRINT1 ("\n"); \
|
\
|
\
|
if (1) \
|
if (1) \
|
for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \
|
for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \
|
this_reg++) \
|
this_reg++) \
|
{ \
|
{ \
|
DEBUG_PRINT2 (" Pushing reg: %lu\n", this_reg); \
|
DEBUG_PRINT2 (" Pushing reg: %lu\n", this_reg); \
|
DEBUG_STATEMENT (num_regs_pushed++); \
|
DEBUG_STATEMENT (num_regs_pushed++); \
|
\
|
\
|
DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
|
DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
|
PUSH_FAILURE_POINTER (regstart[this_reg]); \
|
PUSH_FAILURE_POINTER (regstart[this_reg]); \
|
\
|
\
|
DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
|
DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
|
PUSH_FAILURE_POINTER (regend[this_reg]); \
|
PUSH_FAILURE_POINTER (regend[this_reg]); \
|
\
|
\
|
DEBUG_PRINT2 (" info: %p\n ", \
|
DEBUG_PRINT2 (" info: %p\n ", \
|
reg_info[this_reg].word.pointer); \
|
reg_info[this_reg].word.pointer); \
|
DEBUG_PRINT2 (" match_null=%d", \
|
DEBUG_PRINT2 (" match_null=%d", \
|
REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \
|
REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \
|
DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \
|
DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \
|
DEBUG_PRINT2 (" matched_something=%d", \
|
DEBUG_PRINT2 (" matched_something=%d", \
|
MATCHED_SOMETHING (reg_info[this_reg])); \
|
MATCHED_SOMETHING (reg_info[this_reg])); \
|
DEBUG_PRINT2 (" ever_matched=%d", \
|
DEBUG_PRINT2 (" ever_matched=%d", \
|
EVER_MATCHED_SOMETHING (reg_info[this_reg])); \
|
EVER_MATCHED_SOMETHING (reg_info[this_reg])); \
|
DEBUG_PRINT1 ("\n"); \
|
DEBUG_PRINT1 ("\n"); \
|
PUSH_FAILURE_ELT (reg_info[this_reg].word); \
|
PUSH_FAILURE_ELT (reg_info[this_reg].word); \
|
} \
|
} \
|
\
|
\
|
DEBUG_PRINT2 (" Pushing low active reg: %ld\n", lowest_active_reg);\
|
DEBUG_PRINT2 (" Pushing low active reg: %ld\n", lowest_active_reg);\
|
PUSH_FAILURE_INT (lowest_active_reg); \
|
PUSH_FAILURE_INT (lowest_active_reg); \
|
\
|
\
|
DEBUG_PRINT2 (" Pushing high active reg: %ld\n", highest_active_reg);\
|
DEBUG_PRINT2 (" Pushing high active reg: %ld\n", highest_active_reg);\
|
PUSH_FAILURE_INT (highest_active_reg); \
|
PUSH_FAILURE_INT (highest_active_reg); \
|
\
|
\
|
DEBUG_PRINT2 (" Pushing pattern %p:\n", pattern_place); \
|
DEBUG_PRINT2 (" Pushing pattern %p:\n", pattern_place); \
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \
|
PUSH_FAILURE_POINTER (pattern_place); \
|
PUSH_FAILURE_POINTER (pattern_place); \
|
\
|
\
|
DEBUG_PRINT2 (" Pushing string %p: `", string_place); \
|
DEBUG_PRINT2 (" Pushing string %p: `", string_place); \
|
DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \
|
DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \
|
size2); \
|
size2); \
|
DEBUG_PRINT1 ("'\n"); \
|
DEBUG_PRINT1 ("'\n"); \
|
PUSH_FAILURE_POINTER (string_place); \
|
PUSH_FAILURE_POINTER (string_place); \
|
\
|
\
|
DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \
|
DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \
|
DEBUG_PUSH (failure_id); \
|
DEBUG_PUSH (failure_id); \
|
} while (0)
|
} while (0)
|
|
|
/* This is the number of items that are pushed and popped on the stack
|
/* This is the number of items that are pushed and popped on the stack
|
for each register. */
|
for each register. */
|
#define NUM_REG_ITEMS 3
|
#define NUM_REG_ITEMS 3
|
|
|
/* Individual items aside from the registers. */
|
/* Individual items aside from the registers. */
|
#ifdef DEBUG
|
#ifdef DEBUG
|
# define NUM_NONREG_ITEMS 5 /* Includes failure point id. */
|
# define NUM_NONREG_ITEMS 5 /* Includes failure point id. */
|
#else
|
#else
|
# define NUM_NONREG_ITEMS 4
|
# define NUM_NONREG_ITEMS 4
|
#endif
|
#endif
|
|
|
/* We push at most this many items on the stack. */
|
/* We push at most this many items on the stack. */
|
/* We used to use (num_regs - 1), which is the number of registers
|
/* We used to use (num_regs - 1), which is the number of registers
|
this regexp will save; but that was changed to 5
|
this regexp will save; but that was changed to 5
|
to avoid stack overflow for a regexp with lots of parens. */
|
to avoid stack overflow for a regexp with lots of parens. */
|
#define MAX_FAILURE_ITEMS (5 * NUM_REG_ITEMS + NUM_NONREG_ITEMS)
|
#define MAX_FAILURE_ITEMS (5 * NUM_REG_ITEMS + NUM_NONREG_ITEMS)
|
|
|
/* We actually push this many items. */
|
/* We actually push this many items. */
|
#define NUM_FAILURE_ITEMS \
|
#define NUM_FAILURE_ITEMS \
|
(((0 \
|
(((0 \
|
? 0 : highest_active_reg - lowest_active_reg + 1) \
|
? 0 : highest_active_reg - lowest_active_reg + 1) \
|
* NUM_REG_ITEMS) \
|
* NUM_REG_ITEMS) \
|
+ NUM_NONREG_ITEMS)
|
+ NUM_NONREG_ITEMS)
|
|
|
/* How many items can still be added to the stack without overflowing it. */
|
/* How many items can still be added to the stack without overflowing it. */
|
#define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail)
|
#define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail)
|
|
|
|
|
/* Pops what PUSH_FAIL_STACK pushes.
|
/* Pops what PUSH_FAIL_STACK pushes.
|
|
|
We restore into the parameters, all of which should be lvalues:
|
We restore into the parameters, all of which should be lvalues:
|
STR -- the saved data position.
|
STR -- the saved data position.
|
PAT -- the saved pattern position.
|
PAT -- the saved pattern position.
|
LOW_REG, HIGH_REG -- the highest and lowest active registers.
|
LOW_REG, HIGH_REG -- the highest and lowest active registers.
|
REGSTART, REGEND -- arrays of string positions.
|
REGSTART, REGEND -- arrays of string positions.
|
REG_INFO -- array of information about each subexpression.
|
REG_INFO -- array of information about each subexpression.
|
|
|
Also assumes the variables `fail_stack' and (if debugging), `bufp',
|
Also assumes the variables `fail_stack' and (if debugging), `bufp',
|
`pend', `string1', `size1', `string2', and `size2'. */
|
`pend', `string1', `size1', `string2', and `size2'. */
|
|
|
#define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\
|
#define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\
|
{ \
|
{ \
|
DEBUG_STATEMENT (unsigned failure_id;) \
|
DEBUG_STATEMENT (unsigned failure_id;) \
|
active_reg_t this_reg; \
|
active_reg_t this_reg; \
|
const unsigned char *string_temp; \
|
const unsigned char *string_temp; \
|
\
|
\
|
assert (!FAIL_STACK_EMPTY ()); \
|
assert (!FAIL_STACK_EMPTY ()); \
|
\
|
\
|
/* Remove failure points and point to how many regs pushed. */ \
|
/* Remove failure points and point to how many regs pushed. */ \
|
DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \
|
DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \
|
DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \
|
DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \
|
DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \
|
DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \
|
\
|
\
|
assert (fail_stack.avail >= NUM_NONREG_ITEMS); \
|
assert (fail_stack.avail >= NUM_NONREG_ITEMS); \
|
\
|
\
|
DEBUG_POP (&failure_id); \
|
DEBUG_POP (&failure_id); \
|
DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \
|
DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \
|
\
|
\
|
/* If the saved string location is NULL, it came from an \
|
/* If the saved string location is NULL, it came from an \
|
on_failure_keep_string_jump opcode, and we want to throw away the \
|
on_failure_keep_string_jump opcode, and we want to throw away the \
|
saved NULL, thus retaining our current position in the string. */ \
|
saved NULL, thus retaining our current position in the string. */ \
|
string_temp = POP_FAILURE_POINTER (); \
|
string_temp = POP_FAILURE_POINTER (); \
|
if (string_temp != NULL) \
|
if (string_temp != NULL) \
|
str = (const char *) string_temp; \
|
str = (const char *) string_temp; \
|
\
|
\
|
DEBUG_PRINT2 (" Popping string %p: `", str); \
|
DEBUG_PRINT2 (" Popping string %p: `", str); \
|
DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \
|
DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \
|
DEBUG_PRINT1 ("'\n"); \
|
DEBUG_PRINT1 ("'\n"); \
|
\
|
\
|
pat = (unsigned char *) POP_FAILURE_POINTER (); \
|
pat = (unsigned char *) POP_FAILURE_POINTER (); \
|
DEBUG_PRINT2 (" Popping pattern %p:\n", pat); \
|
DEBUG_PRINT2 (" Popping pattern %p:\n", pat); \
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \
|
\
|
\
|
/* Restore register info. */ \
|
/* Restore register info. */ \
|
high_reg = (active_reg_t) POP_FAILURE_INT (); \
|
high_reg = (active_reg_t) POP_FAILURE_INT (); \
|
DEBUG_PRINT2 (" Popping high active reg: %ld\n", high_reg); \
|
DEBUG_PRINT2 (" Popping high active reg: %ld\n", high_reg); \
|
\
|
\
|
low_reg = (active_reg_t) POP_FAILURE_INT (); \
|
low_reg = (active_reg_t) POP_FAILURE_INT (); \
|
DEBUG_PRINT2 (" Popping low active reg: %ld\n", low_reg); \
|
DEBUG_PRINT2 (" Popping low active reg: %ld\n", low_reg); \
|
\
|
\
|
if (1) \
|
if (1) \
|
for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \
|
for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \
|
{ \
|
{ \
|
DEBUG_PRINT2 (" Popping reg: %ld\n", this_reg); \
|
DEBUG_PRINT2 (" Popping reg: %ld\n", this_reg); \
|
\
|
\
|
reg_info[this_reg].word = POP_FAILURE_ELT (); \
|
reg_info[this_reg].word = POP_FAILURE_ELT (); \
|
DEBUG_PRINT2 (" info: %p\n", \
|
DEBUG_PRINT2 (" info: %p\n", \
|
reg_info[this_reg].word.pointer); \
|
reg_info[this_reg].word.pointer); \
|
\
|
\
|
regend[this_reg] = (const char *) POP_FAILURE_POINTER (); \
|
regend[this_reg] = (const char *) POP_FAILURE_POINTER (); \
|
DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
|
DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
|
\
|
\
|
regstart[this_reg] = (const char *) POP_FAILURE_POINTER (); \
|
regstart[this_reg] = (const char *) POP_FAILURE_POINTER (); \
|
DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
|
DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
|
} \
|
} \
|
else \
|
else \
|
{ \
|
{ \
|
for (this_reg = highest_active_reg; this_reg > high_reg; this_reg--) \
|
for (this_reg = highest_active_reg; this_reg > high_reg; this_reg--) \
|
{ \
|
{ \
|
reg_info[this_reg].word.integer = 0; \
|
reg_info[this_reg].word.integer = 0; \
|
regend[this_reg] = 0; \
|
regend[this_reg] = 0; \
|
regstart[this_reg] = 0; \
|
regstart[this_reg] = 0; \
|
} \
|
} \
|
highest_active_reg = high_reg; \
|
highest_active_reg = high_reg; \
|
} \
|
} \
|
\
|
\
|
set_regs_matched_done = 0; \
|
set_regs_matched_done = 0; \
|
DEBUG_STATEMENT (nfailure_points_popped++); \
|
DEBUG_STATEMENT (nfailure_points_popped++); \
|
} /* POP_FAILURE_POINT */
|
} /* POP_FAILURE_POINT */
|
|
|
|
|
|
|
/* Structure for per-register (a.k.a. per-group) information.
|
/* Structure for per-register (a.k.a. per-group) information.
|
Other register information, such as the
|
Other register information, such as the
|
starting and ending positions (which are addresses), and the list of
|
starting and ending positions (which are addresses), and the list of
|
inner groups (which is a bits list) are maintained in separate
|
inner groups (which is a bits list) are maintained in separate
|
variables.
|
variables.
|
|
|
We are making a (strictly speaking) nonportable assumption here: that
|
We are making a (strictly speaking) nonportable assumption here: that
|
the compiler will pack our bit fields into something that fits into
|
the compiler will pack our bit fields into something that fits into
|
the type of `word', i.e., is something that fits into one item on the
|
the type of `word', i.e., is something that fits into one item on the
|
failure stack. */
|
failure stack. */
|
|
|
|
|
/* Declarations and macros for re_match_2. */
|
/* Declarations and macros for re_match_2. */
|
|
|
typedef union
|
typedef union
|
{
|
{
|
fail_stack_elt_t word;
|
fail_stack_elt_t word;
|
struct
|
struct
|
{
|
{
|
/* This field is one if this group can match the empty string,
|
/* This field is one if this group can match the empty string,
|
zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */
|
zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */
|
#define MATCH_NULL_UNSET_VALUE 3
|
#define MATCH_NULL_UNSET_VALUE 3
|
unsigned match_null_string_p : 2;
|
unsigned match_null_string_p : 2;
|
unsigned is_active : 1;
|
unsigned is_active : 1;
|
unsigned matched_something : 1;
|
unsigned matched_something : 1;
|
unsigned ever_matched_something : 1;
|
unsigned ever_matched_something : 1;
|
} bits;
|
} bits;
|
} register_info_type;
|
} register_info_type;
|
|
|
#define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p)
|
#define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p)
|
#define IS_ACTIVE(R) ((R).bits.is_active)
|
#define IS_ACTIVE(R) ((R).bits.is_active)
|
#define MATCHED_SOMETHING(R) ((R).bits.matched_something)
|
#define MATCHED_SOMETHING(R) ((R).bits.matched_something)
|
#define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something)
|
#define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something)
|
|
|
|
|
/* Call this when have matched a real character; it sets `matched' flags
|
/* Call this when have matched a real character; it sets `matched' flags
|
for the subexpressions which we are currently inside. Also records
|
for the subexpressions which we are currently inside. Also records
|
that those subexprs have matched. */
|
that those subexprs have matched. */
|
#define SET_REGS_MATCHED() \
|
#define SET_REGS_MATCHED() \
|
do \
|
do \
|
{ \
|
{ \
|
if (!set_regs_matched_done) \
|
if (!set_regs_matched_done) \
|
{ \
|
{ \
|
active_reg_t r; \
|
active_reg_t r; \
|
set_regs_matched_done = 1; \
|
set_regs_matched_done = 1; \
|
for (r = lowest_active_reg; r <= highest_active_reg; r++) \
|
for (r = lowest_active_reg; r <= highest_active_reg; r++) \
|
{ \
|
{ \
|
MATCHED_SOMETHING (reg_info[r]) \
|
MATCHED_SOMETHING (reg_info[r]) \
|
= EVER_MATCHED_SOMETHING (reg_info[r]) \
|
= EVER_MATCHED_SOMETHING (reg_info[r]) \
|
= 1; \
|
= 1; \
|
} \
|
} \
|
} \
|
} \
|
} \
|
} \
|
while (0)
|
while (0)
|
|
|
/* Registers are set to a sentinel when they haven't yet matched. */
|
/* Registers are set to a sentinel when they haven't yet matched. */
|
static char reg_unset_dummy;
|
static char reg_unset_dummy;
|
#define REG_UNSET_VALUE (®_unset_dummy)
|
#define REG_UNSET_VALUE (®_unset_dummy)
|
#define REG_UNSET(e) ((e) == REG_UNSET_VALUE)
|
#define REG_UNSET(e) ((e) == REG_UNSET_VALUE)
|
|
|
/* Subroutine declarations and macros for regex_compile. */
|
/* Subroutine declarations and macros for regex_compile. */
|
|
|
static reg_errcode_t regex_compile _RE_ARGS ((const char *pattern, size_t size,
|
static reg_errcode_t regex_compile _RE_ARGS ((const char *pattern, size_t size,
|
reg_syntax_t syntax,
|
reg_syntax_t syntax,
|
struct re_pattern_buffer *bufp));
|
struct re_pattern_buffer *bufp));
|
static void store_op1 _RE_ARGS ((re_opcode_t op, unsigned char *loc, int arg));
|
static void store_op1 _RE_ARGS ((re_opcode_t op, unsigned char *loc, int arg));
|
static void store_op2 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
static void store_op2 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
int arg1, int arg2));
|
int arg1, int arg2));
|
static void insert_op1 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
static void insert_op1 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
int arg, unsigned char *end));
|
int arg, unsigned char *end));
|
static void insert_op2 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
static void insert_op2 _RE_ARGS ((re_opcode_t op, unsigned char *loc,
|
int arg1, int arg2, unsigned char *end));
|
int arg1, int arg2, unsigned char *end));
|
static boolean at_begline_loc_p _RE_ARGS ((const char *pattern, const char *p,
|
static boolean at_begline_loc_p _RE_ARGS ((const char *pattern, const char *p,
|
reg_syntax_t syntax));
|
reg_syntax_t syntax));
|
static boolean at_endline_loc_p _RE_ARGS ((const char *p, const char *pend,
|
static boolean at_endline_loc_p _RE_ARGS ((const char *p, const char *pend,
|
reg_syntax_t syntax));
|
reg_syntax_t syntax));
|
static reg_errcode_t compile_range _RE_ARGS ((const char **p_ptr,
|
static reg_errcode_t compile_range _RE_ARGS ((const char **p_ptr,
|
const char *pend,
|
const char *pend,
|
char *translate,
|
char *translate,
|
reg_syntax_t syntax,
|
reg_syntax_t syntax,
|
unsigned char *b));
|
unsigned char *b));
|
|
|
/* Fetch the next character in the uncompiled pattern---translating it
|
/* Fetch the next character in the uncompiled pattern---translating it
|
if necessary. Also cast from a signed character in the constant
|
if necessary. Also cast from a signed character in the constant
|
string passed to us by the user to an unsigned char that we can use
|
string passed to us by the user to an unsigned char that we can use
|
as an array index (in, e.g., `translate'). */
|
as an array index (in, e.g., `translate'). */
|
#ifndef PATFETCH
|
#ifndef PATFETCH
|
# define PATFETCH(c) \
|
# define PATFETCH(c) \
|
do {if (p == pend) return REG_EEND; \
|
do {if (p == pend) return REG_EEND; \
|
c = (unsigned char) *p++; \
|
c = (unsigned char) *p++; \
|
if (translate) c = (unsigned char) translate[c]; \
|
if (translate) c = (unsigned char) translate[c]; \
|
} while (0)
|
} while (0)
|
#endif
|
#endif
|
|
|
/* Fetch the next character in the uncompiled pattern, with no
|
/* Fetch the next character in the uncompiled pattern, with no
|
translation. */
|
translation. */
|
#define PATFETCH_RAW(c) \
|
#define PATFETCH_RAW(c) \
|
do {if (p == pend) return REG_EEND; \
|
do {if (p == pend) return REG_EEND; \
|
c = (unsigned char) *p++; \
|
c = (unsigned char) *p++; \
|
} while (0)
|
} while (0)
|
|
|
/* Go backwards one character in the pattern. */
|
/* Go backwards one character in the pattern. */
|
#define PATUNFETCH p--
|
#define PATUNFETCH p--
|
|
|
|
|
/* If `translate' is non-null, return translate[D], else just D. We
|
/* If `translate' is non-null, return translate[D], else just D. We
|
cast the subscript to translate because some data is declared as
|
cast the subscript to translate because some data is declared as
|
`char *', to avoid warnings when a string constant is passed. But
|
`char *', to avoid warnings when a string constant is passed. But
|
when we use a character as a subscript we must make it unsigned. */
|
when we use a character as a subscript we must make it unsigned. */
|
#ifndef TRANSLATE
|
#ifndef TRANSLATE
|
# define TRANSLATE(d) \
|
# define TRANSLATE(d) \
|
(translate ? (char) translate[(unsigned char) (d)] : (d))
|
(translate ? (char) translate[(unsigned char) (d)] : (d))
|
#endif
|
#endif
|
|
|
|
|
/* Macros for outputting the compiled pattern into `buffer'. */
|
/* Macros for outputting the compiled pattern into `buffer'. */
|
|
|
/* If the buffer isn't allocated when it comes in, use this. */
|
/* If the buffer isn't allocated when it comes in, use this. */
|
#define INIT_BUF_SIZE 32
|
#define INIT_BUF_SIZE 32
|
|
|
/* Make sure we have at least N more bytes of space in buffer. */
|
/* Make sure we have at least N more bytes of space in buffer. */
|
#define GET_BUFFER_SPACE(n) \
|
#define GET_BUFFER_SPACE(n) \
|
while ((unsigned long) (b - bufp->buffer + (n)) > bufp->allocated) \
|
while ((unsigned long) (b - bufp->buffer + (n)) > bufp->allocated) \
|
EXTEND_BUFFER ()
|
EXTEND_BUFFER ()
|
|
|
/* Make sure we have one more byte of buffer space and then add C to it. */
|
/* Make sure we have one more byte of buffer space and then add C to it. */
|
#define BUF_PUSH(c) \
|
#define BUF_PUSH(c) \
|
do { \
|
do { \
|
GET_BUFFER_SPACE (1); \
|
GET_BUFFER_SPACE (1); \
|
*b++ = (unsigned char) (c); \
|
*b++ = (unsigned char) (c); \
|
} while (0)
|
} while (0)
|
|
|
|
|
/* Ensure we have two more bytes of buffer space and then append C1 and C2. */
|
/* Ensure we have two more bytes of buffer space and then append C1 and C2. */
|
#define BUF_PUSH_2(c1, c2) \
|
#define BUF_PUSH_2(c1, c2) \
|
do { \
|
do { \
|
GET_BUFFER_SPACE (2); \
|
GET_BUFFER_SPACE (2); \
|
*b++ = (unsigned char) (c1); \
|
*b++ = (unsigned char) (c1); \
|
*b++ = (unsigned char) (c2); \
|
*b++ = (unsigned char) (c2); \
|
} while (0)
|
} while (0)
|
|
|
|
|
/* As with BUF_PUSH_2, except for three bytes. */
|
/* As with BUF_PUSH_2, except for three bytes. */
|
#define BUF_PUSH_3(c1, c2, c3) \
|
#define BUF_PUSH_3(c1, c2, c3) \
|
do { \
|
do { \
|
GET_BUFFER_SPACE (3); \
|
GET_BUFFER_SPACE (3); \
|
*b++ = (unsigned char) (c1); \
|
*b++ = (unsigned char) (c1); \
|
*b++ = (unsigned char) (c2); \
|
*b++ = (unsigned char) (c2); \
|
*b++ = (unsigned char) (c3); \
|
*b++ = (unsigned char) (c3); \
|
} while (0)
|
} while (0)
|
|
|
|
|
/* Store a jump with opcode OP at LOC to location TO. We store a
|
/* Store a jump with opcode OP at LOC to location TO. We store a
|
relative address offset by the three bytes the jump itself occupies. */
|
relative address offset by the three bytes the jump itself occupies. */
|
#define STORE_JUMP(op, loc, to) \
|
#define STORE_JUMP(op, loc, to) \
|
store_op1 (op, loc, (int) ((to) - (loc) - 3))
|
store_op1 (op, loc, (int) ((to) - (loc) - 3))
|
|
|
/* Likewise, for a two-argument jump. */
|
/* Likewise, for a two-argument jump. */
|
#define STORE_JUMP2(op, loc, to, arg) \
|
#define STORE_JUMP2(op, loc, to, arg) \
|
store_op2 (op, loc, (int) ((to) - (loc) - 3), arg)
|
store_op2 (op, loc, (int) ((to) - (loc) - 3), arg)
|
|
|
/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */
|
/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */
|
#define INSERT_JUMP(op, loc, to) \
|
#define INSERT_JUMP(op, loc, to) \
|
insert_op1 (op, loc, (int) ((to) - (loc) - 3), b)
|
insert_op1 (op, loc, (int) ((to) - (loc) - 3), b)
|
|
|
/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */
|
/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */
|
#define INSERT_JUMP2(op, loc, to, arg) \
|
#define INSERT_JUMP2(op, loc, to, arg) \
|
insert_op2 (op, loc, (int) ((to) - (loc) - 3), arg, b)
|
insert_op2 (op, loc, (int) ((to) - (loc) - 3), arg, b)
|
|
|
|
|
/* This is not an arbitrary limit: the arguments which represent offsets
|
/* This is not an arbitrary limit: the arguments which represent offsets
|
into the pattern are two bytes long. So if 2^16 bytes turns out to
|
into the pattern are two bytes long. So if 2^16 bytes turns out to
|
be too small, many things would have to change. */
|
be too small, many things would have to change. */
|
/* Any other compiler which, like MSC, has allocation limit below 2^16
|
/* Any other compiler which, like MSC, has allocation limit below 2^16
|
bytes will have to use approach similar to what was done below for
|
bytes will have to use approach similar to what was done below for
|
MSC and drop MAX_BUF_SIZE a bit. Otherwise you may end up
|
MSC and drop MAX_BUF_SIZE a bit. Otherwise you may end up
|
reallocating to 0 bytes. Such thing is not going to work too well.
|
reallocating to 0 bytes. Such thing is not going to work too well.
|
You have been warned!! */
|
You have been warned!! */
|
#if defined _MSC_VER && !defined WIN32
|
#if defined _MSC_VER && !defined WIN32
|
/* Microsoft C 16-bit versions limit malloc to approx 65512 bytes.
|
/* Microsoft C 16-bit versions limit malloc to approx 65512 bytes.
|
The REALLOC define eliminates a flurry of conversion warnings,
|
The REALLOC define eliminates a flurry of conversion warnings,
|
but is not required. */
|
but is not required. */
|
# define MAX_BUF_SIZE 65500L
|
# define MAX_BUF_SIZE 65500L
|
# define REALLOC(p,s) realloc ((p), (size_t) (s))
|
# define REALLOC(p,s) realloc ((p), (size_t) (s))
|
#else
|
#else
|
# define MAX_BUF_SIZE (1L << 16)
|
# define MAX_BUF_SIZE (1L << 16)
|
# define REALLOC(p,s) realloc ((p), (s))
|
# define REALLOC(p,s) realloc ((p), (s))
|
#endif
|
#endif
|
|
|
/* Extend the buffer by twice its current size via realloc and
|
/* Extend the buffer by twice its current size via realloc and
|
reset the pointers that pointed into the old block to point to the
|
reset the pointers that pointed into the old block to point to the
|
correct places in the new one. If extending the buffer results in it
|
correct places in the new one. If extending the buffer results in it
|
being larger than MAX_BUF_SIZE, then flag memory exhausted. */
|
being larger than MAX_BUF_SIZE, then flag memory exhausted. */
|
#define EXTEND_BUFFER() \
|
#define EXTEND_BUFFER() \
|
do { \
|
do { \
|
unsigned char *old_buffer = bufp->buffer; \
|
unsigned char *old_buffer = bufp->buffer; \
|
if (bufp->allocated == MAX_BUF_SIZE) \
|
if (bufp->allocated == MAX_BUF_SIZE) \
|
return REG_ESIZE; \
|
return REG_ESIZE; \
|
bufp->allocated <<= 1; \
|
bufp->allocated <<= 1; \
|
if (bufp->allocated > MAX_BUF_SIZE) \
|
if (bufp->allocated > MAX_BUF_SIZE) \
|
bufp->allocated = MAX_BUF_SIZE; \
|
bufp->allocated = MAX_BUF_SIZE; \
|
bufp->buffer = (unsigned char *) REALLOC (bufp->buffer, bufp->allocated);\
|
bufp->buffer = (unsigned char *) REALLOC (bufp->buffer, bufp->allocated);\
|
if (bufp->buffer == NULL) \
|
if (bufp->buffer == NULL) \
|
return REG_ESPACE; \
|
return REG_ESPACE; \
|
/* If the buffer moved, move all the pointers into it. */ \
|
/* If the buffer moved, move all the pointers into it. */ \
|
if (old_buffer != bufp->buffer) \
|
if (old_buffer != bufp->buffer) \
|
{ \
|
{ \
|
b = (b - old_buffer) + bufp->buffer; \
|
b = (b - old_buffer) + bufp->buffer; \
|
begalt = (begalt - old_buffer) + bufp->buffer; \
|
begalt = (begalt - old_buffer) + bufp->buffer; \
|
if (fixup_alt_jump) \
|
if (fixup_alt_jump) \
|
fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\
|
fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\
|
if (laststart) \
|
if (laststart) \
|
laststart = (laststart - old_buffer) + bufp->buffer; \
|
laststart = (laststart - old_buffer) + bufp->buffer; \
|
if (pending_exact) \
|
if (pending_exact) \
|
pending_exact = (pending_exact - old_buffer) + bufp->buffer; \
|
pending_exact = (pending_exact - old_buffer) + bufp->buffer; \
|
} \
|
} \
|
} while (0)
|
} while (0)
|
|
|
|
|
/* Since we have one byte reserved for the register number argument to
|
/* Since we have one byte reserved for the register number argument to
|
{start,stop}_memory, the maximum number of groups we can report
|
{start,stop}_memory, the maximum number of groups we can report
|
things about is what fits in that byte. */
|
things about is what fits in that byte. */
|
#define MAX_REGNUM 255
|
#define MAX_REGNUM 255
|
|
|
/* But patterns can have more than `MAX_REGNUM' registers. We just
|
/* But patterns can have more than `MAX_REGNUM' registers. We just
|
ignore the excess. */
|
ignore the excess. */
|
typedef unsigned regnum_t;
|
typedef unsigned regnum_t;
|
|
|
|
|
/* Macros for the compile stack. */
|
/* Macros for the compile stack. */
|
|
|
/* Since offsets can go either forwards or backwards, this type needs to
|
/* Since offsets can go either forwards or backwards, this type needs to
|
be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */
|
be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */
|
/* int may be not enough when sizeof(int) == 2. */
|
/* int may be not enough when sizeof(int) == 2. */
|
typedef long pattern_offset_t;
|
typedef long pattern_offset_t;
|
|
|
typedef struct
|
typedef struct
|
{
|
{
|
pattern_offset_t begalt_offset;
|
pattern_offset_t begalt_offset;
|
pattern_offset_t fixup_alt_jump;
|
pattern_offset_t fixup_alt_jump;
|
pattern_offset_t inner_group_offset;
|
pattern_offset_t inner_group_offset;
|
pattern_offset_t laststart_offset;
|
pattern_offset_t laststart_offset;
|
regnum_t regnum;
|
regnum_t regnum;
|
} compile_stack_elt_t;
|
} compile_stack_elt_t;
|
|
|
|
|
typedef struct
|
typedef struct
|
{
|
{
|
compile_stack_elt_t *stack;
|
compile_stack_elt_t *stack;
|
unsigned size;
|
unsigned size;
|
unsigned avail; /* Offset of next open position. */
|
unsigned avail; /* Offset of next open position. */
|
} compile_stack_type;
|
} compile_stack_type;
|
|
|
|
|
#define INIT_COMPILE_STACK_SIZE 32
|
#define INIT_COMPILE_STACK_SIZE 32
|
|
|
#define COMPILE_STACK_EMPTY (compile_stack.avail == 0)
|
#define COMPILE_STACK_EMPTY (compile_stack.avail == 0)
|
#define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size)
|
#define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size)
|
|
|
/* The next available element. */
|
/* The next available element. */
|
#define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail])
|
#define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail])
|
|
|
|
|
/* Set the bit for character C in a list. */
|
/* Set the bit for character C in a list. */
|
#define SET_LIST_BIT(c) \
|
#define SET_LIST_BIT(c) \
|
(b[((unsigned char) (c)) / BYTEWIDTH] \
|
(b[((unsigned char) (c)) / BYTEWIDTH] \
|
|= 1 << (((unsigned char) c) % BYTEWIDTH))
|
|= 1 << (((unsigned char) c) % BYTEWIDTH))
|
|
|
|
|
/* Get the next unsigned number in the uncompiled pattern. */
|
/* Get the next unsigned number in the uncompiled pattern. */
|
#define GET_UNSIGNED_NUMBER(num) \
|
#define GET_UNSIGNED_NUMBER(num) \
|
{ if (p != pend) \
|
{ if (p != pend) \
|
{ \
|
{ \
|
PATFETCH (c); \
|
PATFETCH (c); \
|
while (ISDIGIT (c)) \
|
while (ISDIGIT (c)) \
|
{ \
|
{ \
|
if (num < 0) \
|
if (num < 0) \
|
num = 0; \
|
num = 0; \
|
num = num * 10 + c - '0'; \
|
num = num * 10 + c - '0'; \
|
if (p == pend) \
|
if (p == pend) \
|
break; \
|
break; \
|
PATFETCH (c); \
|
PATFETCH (c); \
|
} \
|
} \
|
} \
|
} \
|
}
|
}
|
|
|
/* Use this only if they have btowc(), since wctype() is used below
|
/* Use this only if they have btowc(), since wctype() is used below
|
together with btowc(). btowc() is defined in the 1994 Amendment 1
|
together with btowc(). btowc() is defined in the 1994 Amendment 1
|
to ISO C and may not be present on systems where we have wchar.h
|
to ISO C and may not be present on systems where we have wchar.h
|
and wctype.h. */
|
and wctype.h. */
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H && defined HAVE_BTOWC)
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H && defined HAVE_BTOWC)
|
/* The GNU C library provides support for user-defined character classes
|
/* The GNU C library provides support for user-defined character classes
|
and the functions from ISO C amendement 1. */
|
and the functions from ISO C amendement 1. */
|
# ifdef CHARCLASS_NAME_MAX
|
# ifdef CHARCLASS_NAME_MAX
|
# define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
|
# define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
|
# else
|
# else
|
/* This shouldn't happen but some implementation might still have this
|
/* This shouldn't happen but some implementation might still have this
|
problem. Use a reasonable default value. */
|
problem. Use a reasonable default value. */
|
# define CHAR_CLASS_MAX_LENGTH 256
|
# define CHAR_CLASS_MAX_LENGTH 256
|
# endif
|
# endif
|
|
|
# ifdef _LIBC
|
# ifdef _LIBC
|
# define IS_CHAR_CLASS(string) __wctype (string)
|
# define IS_CHAR_CLASS(string) __wctype (string)
|
# else
|
# else
|
# define IS_CHAR_CLASS(string) wctype (string)
|
# define IS_CHAR_CLASS(string) wctype (string)
|
# endif
|
# endif
|
#else
|
#else
|
# define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
|
# define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
|
|
|
# define IS_CHAR_CLASS(string) \
|
# define IS_CHAR_CLASS(string) \
|
(STREQ (string, "alpha") || STREQ (string, "upper") \
|
(STREQ (string, "alpha") || STREQ (string, "upper") \
|
|| STREQ (string, "lower") || STREQ (string, "digit") \
|
|| STREQ (string, "lower") || STREQ (string, "digit") \
|
|| STREQ (string, "alnum") || STREQ (string, "xdigit") \
|
|| STREQ (string, "alnum") || STREQ (string, "xdigit") \
|
|| STREQ (string, "space") || STREQ (string, "print") \
|
|| STREQ (string, "space") || STREQ (string, "print") \
|
|| STREQ (string, "punct") || STREQ (string, "graph") \
|
|| STREQ (string, "punct") || STREQ (string, "graph") \
|
|| STREQ (string, "cntrl") || STREQ (string, "blank"))
|
|| STREQ (string, "cntrl") || STREQ (string, "blank"))
|
#endif
|
#endif
|
|
|
#ifndef MATCH_MAY_ALLOCATE
|
#ifndef MATCH_MAY_ALLOCATE
|
|
|
/* If we cannot allocate large objects within re_match_2_internal,
|
/* If we cannot allocate large objects within re_match_2_internal,
|
we make the fail stack and register vectors global.
|
we make the fail stack and register vectors global.
|
The fail stack, we grow to the maximum size when a regexp
|
The fail stack, we grow to the maximum size when a regexp
|
is compiled.
|
is compiled.
|
The register vectors, we adjust in size each time we
|
The register vectors, we adjust in size each time we
|
compile a regexp, according to the number of registers it needs. */
|
compile a regexp, according to the number of registers it needs. */
|
|
|
static fail_stack_type fail_stack;
|
static fail_stack_type fail_stack;
|
|
|
/* Size with which the following vectors are currently allocated.
|
/* Size with which the following vectors are currently allocated.
|
That is so we can make them bigger as needed,
|
That is so we can make them bigger as needed,
|
but never make them smaller. */
|
but never make them smaller. */
|
static int regs_allocated_size;
|
static int regs_allocated_size;
|
|
|
static const char ** regstart, ** regend;
|
static const char ** regstart, ** regend;
|
static const char ** old_regstart, ** old_regend;
|
static const char ** old_regstart, ** old_regend;
|
static const char **best_regstart, **best_regend;
|
static const char **best_regstart, **best_regend;
|
static register_info_type *reg_info;
|
static register_info_type *reg_info;
|
static const char **reg_dummy;
|
static const char **reg_dummy;
|
static register_info_type *reg_info_dummy;
|
static register_info_type *reg_info_dummy;
|
|
|
/* Make the register vectors big enough for NUM_REGS registers,
|
/* Make the register vectors big enough for NUM_REGS registers,
|
but don't make them smaller. */
|
but don't make them smaller. */
|
|
|
static
|
static
|
regex_grow_registers (num_regs)
|
regex_grow_registers (num_regs)
|
int num_regs;
|
int num_regs;
|
{
|
{
|
if (num_regs > regs_allocated_size)
|
if (num_regs > regs_allocated_size)
|
{
|
{
|
RETALLOC_IF (regstart, num_regs, const char *);
|
RETALLOC_IF (regstart, num_regs, const char *);
|
RETALLOC_IF (regend, num_regs, const char *);
|
RETALLOC_IF (regend, num_regs, const char *);
|
RETALLOC_IF (old_regstart, num_regs, const char *);
|
RETALLOC_IF (old_regstart, num_regs, const char *);
|
RETALLOC_IF (old_regend, num_regs, const char *);
|
RETALLOC_IF (old_regend, num_regs, const char *);
|
RETALLOC_IF (best_regstart, num_regs, const char *);
|
RETALLOC_IF (best_regstart, num_regs, const char *);
|
RETALLOC_IF (best_regend, num_regs, const char *);
|
RETALLOC_IF (best_regend, num_regs, const char *);
|
RETALLOC_IF (reg_info, num_regs, register_info_type);
|
RETALLOC_IF (reg_info, num_regs, register_info_type);
|
RETALLOC_IF (reg_dummy, num_regs, const char *);
|
RETALLOC_IF (reg_dummy, num_regs, const char *);
|
RETALLOC_IF (reg_info_dummy, num_regs, register_info_type);
|
RETALLOC_IF (reg_info_dummy, num_regs, register_info_type);
|
|
|
regs_allocated_size = num_regs;
|
regs_allocated_size = num_regs;
|
}
|
}
|
}
|
}
|
|
|
#endif /* not MATCH_MAY_ALLOCATE */
|
#endif /* not MATCH_MAY_ALLOCATE */
|
|
|
static boolean group_in_compile_stack _RE_ARGS ((compile_stack_type
|
static boolean group_in_compile_stack _RE_ARGS ((compile_stack_type
|
compile_stack,
|
compile_stack,
|
regnum_t regnum));
|
regnum_t regnum));
|
|
|
/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX.
|
/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX.
|
Returns one of error codes defined in `gnu-regex.h', or zero for success.
|
Returns one of error codes defined in `gnu-regex.h', or zero for success.
|
|
|
Assumes the `allocated' (and perhaps `buffer') and `translate'
|
Assumes the `allocated' (and perhaps `buffer') and `translate'
|
fields are set in BUFP on entry.
|
fields are set in BUFP on entry.
|
|
|
If it succeeds, results are put in BUFP (if it returns an error, the
|
If it succeeds, results are put in BUFP (if it returns an error, the
|
contents of BUFP are undefined):
|
contents of BUFP are undefined):
|
`buffer' is the compiled pattern;
|
`buffer' is the compiled pattern;
|
`syntax' is set to SYNTAX;
|
`syntax' is set to SYNTAX;
|
`used' is set to the length of the compiled pattern;
|
`used' is set to the length of the compiled pattern;
|
`fastmap_accurate' is zero;
|
`fastmap_accurate' is zero;
|
`re_nsub' is the number of subexpressions in PATTERN;
|
`re_nsub' is the number of subexpressions in PATTERN;
|
`not_bol' and `not_eol' are zero;
|
`not_bol' and `not_eol' are zero;
|
|
|
The `fastmap' and `newline_anchor' fields are neither
|
The `fastmap' and `newline_anchor' fields are neither
|
examined nor set. */
|
examined nor set. */
|
|
|
/* Return, freeing storage we allocated. */
|
/* Return, freeing storage we allocated. */
|
#define FREE_STACK_RETURN(value) \
|
#define FREE_STACK_RETURN(value) \
|
return (free (compile_stack.stack), value)
|
return (free (compile_stack.stack), value)
|
|
|
static reg_errcode_t
|
static reg_errcode_t
|
regex_compile (pattern, size, syntax, bufp)
|
regex_compile (pattern, size, syntax, bufp)
|
const char *pattern;
|
const char *pattern;
|
size_t size;
|
size_t size;
|
reg_syntax_t syntax;
|
reg_syntax_t syntax;
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
{
|
{
|
/* We fetch characters from PATTERN here. Even though PATTERN is
|
/* We fetch characters from PATTERN here. Even though PATTERN is
|
`char *' (i.e., signed), we declare these variables as unsigned, so
|
`char *' (i.e., signed), we declare these variables as unsigned, so
|
they can be reliably used as array indices. */
|
they can be reliably used as array indices. */
|
register unsigned char c, c1;
|
register unsigned char c, c1;
|
|
|
/* A random temporary spot in PATTERN. */
|
/* A random temporary spot in PATTERN. */
|
const char *p1;
|
const char *p1;
|
|
|
/* Points to the end of the buffer, where we should append. */
|
/* Points to the end of the buffer, where we should append. */
|
register unsigned char *b;
|
register unsigned char *b;
|
|
|
/* Keeps track of unclosed groups. */
|
/* Keeps track of unclosed groups. */
|
compile_stack_type compile_stack;
|
compile_stack_type compile_stack;
|
|
|
/* Points to the current (ending) position in the pattern. */
|
/* Points to the current (ending) position in the pattern. */
|
const char *p = pattern;
|
const char *p = pattern;
|
const char *pend = pattern + size;
|
const char *pend = pattern + size;
|
|
|
/* How to translate the characters in the pattern. */
|
/* How to translate the characters in the pattern. */
|
RE_TRANSLATE_TYPE translate = bufp->translate;
|
RE_TRANSLATE_TYPE translate = bufp->translate;
|
|
|
/* Address of the count-byte of the most recently inserted `exactn'
|
/* Address of the count-byte of the most recently inserted `exactn'
|
command. This makes it possible to tell if a new exact-match
|
command. This makes it possible to tell if a new exact-match
|
character can be added to that command or if the character requires
|
character can be added to that command or if the character requires
|
a new `exactn' command. */
|
a new `exactn' command. */
|
unsigned char *pending_exact = 0;
|
unsigned char *pending_exact = 0;
|
|
|
/* Address of start of the most recently finished expression.
|
/* Address of start of the most recently finished expression.
|
This tells, e.g., postfix * where to find the start of its
|
This tells, e.g., postfix * where to find the start of its
|
operand. Reset at the beginning of groups and alternatives. */
|
operand. Reset at the beginning of groups and alternatives. */
|
unsigned char *laststart = 0;
|
unsigned char *laststart = 0;
|
|
|
/* Address of beginning of regexp, or inside of last group. */
|
/* Address of beginning of regexp, or inside of last group. */
|
unsigned char *begalt;
|
unsigned char *begalt;
|
|
|
/* Place in the uncompiled pattern (i.e., the {) to
|
/* Place in the uncompiled pattern (i.e., the {) to
|
which to go back if the interval is invalid. */
|
which to go back if the interval is invalid. */
|
const char *beg_interval;
|
const char *beg_interval;
|
|
|
/* Address of the place where a forward jump should go to the end of
|
/* Address of the place where a forward jump should go to the end of
|
the containing expression. Each alternative of an `or' -- except the
|
the containing expression. Each alternative of an `or' -- except the
|
last -- ends with a forward jump of this sort. */
|
last -- ends with a forward jump of this sort. */
|
unsigned char *fixup_alt_jump = 0;
|
unsigned char *fixup_alt_jump = 0;
|
|
|
/* Counts open-groups as they are encountered. Remembered for the
|
/* Counts open-groups as they are encountered. Remembered for the
|
matching close-group on the compile stack, so the same register
|
matching close-group on the compile stack, so the same register
|
number is put in the stop_memory as the start_memory. */
|
number is put in the stop_memory as the start_memory. */
|
regnum_t regnum = 0;
|
regnum_t regnum = 0;
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
DEBUG_PRINT1 ("\nCompiling pattern: ");
|
DEBUG_PRINT1 ("\nCompiling pattern: ");
|
if (debug)
|
if (debug)
|
{
|
{
|
unsigned debug_count;
|
unsigned debug_count;
|
|
|
for (debug_count = 0; debug_count < size; debug_count++)
|
for (debug_count = 0; debug_count < size; debug_count++)
|
putchar (pattern[debug_count]);
|
putchar (pattern[debug_count]);
|
putchar ('\n');
|
putchar ('\n');
|
}
|
}
|
#endif /* DEBUG */
|
#endif /* DEBUG */
|
|
|
/* Initialize the compile stack. */
|
/* Initialize the compile stack. */
|
compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t);
|
compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t);
|
if (compile_stack.stack == NULL)
|
if (compile_stack.stack == NULL)
|
return REG_ESPACE;
|
return REG_ESPACE;
|
|
|
compile_stack.size = INIT_COMPILE_STACK_SIZE;
|
compile_stack.size = INIT_COMPILE_STACK_SIZE;
|
compile_stack.avail = 0;
|
compile_stack.avail = 0;
|
|
|
/* Initialize the pattern buffer. */
|
/* Initialize the pattern buffer. */
|
bufp->syntax = syntax;
|
bufp->syntax = syntax;
|
bufp->fastmap_accurate = 0;
|
bufp->fastmap_accurate = 0;
|
bufp->not_bol = bufp->not_eol = 0;
|
bufp->not_bol = bufp->not_eol = 0;
|
|
|
/* Set `used' to zero, so that if we return an error, the pattern
|
/* Set `used' to zero, so that if we return an error, the pattern
|
printer (for debugging) will think there's no pattern. We reset it
|
printer (for debugging) will think there's no pattern. We reset it
|
at the end. */
|
at the end. */
|
bufp->used = 0;
|
bufp->used = 0;
|
|
|
/* Always count groups, whether or not bufp->no_sub is set. */
|
/* Always count groups, whether or not bufp->no_sub is set. */
|
bufp->re_nsub = 0;
|
bufp->re_nsub = 0;
|
|
|
#if !defined emacs && !defined SYNTAX_TABLE
|
#if !defined emacs && !defined SYNTAX_TABLE
|
/* Initialize the syntax table. */
|
/* Initialize the syntax table. */
|
init_syntax_once ();
|
init_syntax_once ();
|
#endif
|
#endif
|
|
|
if (bufp->allocated == 0)
|
if (bufp->allocated == 0)
|
{
|
{
|
if (bufp->buffer)
|
if (bufp->buffer)
|
{ /* If zero allocated, but buffer is non-null, try to realloc
|
{ /* If zero allocated, but buffer is non-null, try to realloc
|
enough space. This loses if buffer's address is bogus, but
|
enough space. This loses if buffer's address is bogus, but
|
that is the user's responsibility. */
|
that is the user's responsibility. */
|
RETALLOC (bufp->buffer, INIT_BUF_SIZE, unsigned char);
|
RETALLOC (bufp->buffer, INIT_BUF_SIZE, unsigned char);
|
}
|
}
|
else
|
else
|
{ /* Caller did not allocate a buffer. Do it for them. */
|
{ /* Caller did not allocate a buffer. Do it for them. */
|
bufp->buffer = TALLOC (INIT_BUF_SIZE, unsigned char);
|
bufp->buffer = TALLOC (INIT_BUF_SIZE, unsigned char);
|
}
|
}
|
if (!bufp->buffer) FREE_STACK_RETURN (REG_ESPACE);
|
if (!bufp->buffer) FREE_STACK_RETURN (REG_ESPACE);
|
|
|
bufp->allocated = INIT_BUF_SIZE;
|
bufp->allocated = INIT_BUF_SIZE;
|
}
|
}
|
|
|
begalt = b = bufp->buffer;
|
begalt = b = bufp->buffer;
|
|
|
/* Loop through the uncompiled pattern until we're at the end. */
|
/* Loop through the uncompiled pattern until we're at the end. */
|
while (p != pend)
|
while (p != pend)
|
{
|
{
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
switch (c)
|
switch (c)
|
{
|
{
|
case '^':
|
case '^':
|
{
|
{
|
if ( /* If at start of pattern, it's an operator. */
|
if ( /* If at start of pattern, it's an operator. */
|
p == pattern + 1
|
p == pattern + 1
|
/* If context independent, it's an operator. */
|
/* If context independent, it's an operator. */
|
|| syntax & RE_CONTEXT_INDEP_ANCHORS
|
|| syntax & RE_CONTEXT_INDEP_ANCHORS
|
/* Otherwise, depends on what's come before. */
|
/* Otherwise, depends on what's come before. */
|
|| at_begline_loc_p (pattern, p, syntax))
|
|| at_begline_loc_p (pattern, p, syntax))
|
BUF_PUSH (begline);
|
BUF_PUSH (begline);
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
}
|
}
|
break;
|
break;
|
|
|
|
|
case '$':
|
case '$':
|
{
|
{
|
if ( /* If at end of pattern, it's an operator. */
|
if ( /* If at end of pattern, it's an operator. */
|
p == pend
|
p == pend
|
/* If context independent, it's an operator. */
|
/* If context independent, it's an operator. */
|
|| syntax & RE_CONTEXT_INDEP_ANCHORS
|
|| syntax & RE_CONTEXT_INDEP_ANCHORS
|
/* Otherwise, depends on what's next. */
|
/* Otherwise, depends on what's next. */
|
|| at_endline_loc_p (p, pend, syntax))
|
|| at_endline_loc_p (p, pend, syntax))
|
BUF_PUSH (endline);
|
BUF_PUSH (endline);
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
}
|
}
|
break;
|
break;
|
|
|
|
|
case '+':
|
case '+':
|
case '?':
|
case '?':
|
if ((syntax & RE_BK_PLUS_QM)
|
if ((syntax & RE_BK_PLUS_QM)
|
|| (syntax & RE_LIMITED_OPS))
|
|| (syntax & RE_LIMITED_OPS))
|
goto normal_char;
|
goto normal_char;
|
handle_plus:
|
handle_plus:
|
case '*':
|
case '*':
|
/* If there is no previous pattern... */
|
/* If there is no previous pattern... */
|
if (!laststart)
|
if (!laststart)
|
{
|
{
|
if (syntax & RE_CONTEXT_INVALID_OPS)
|
if (syntax & RE_CONTEXT_INVALID_OPS)
|
FREE_STACK_RETURN (REG_BADRPT);
|
FREE_STACK_RETURN (REG_BADRPT);
|
else if (!(syntax & RE_CONTEXT_INDEP_OPS))
|
else if (!(syntax & RE_CONTEXT_INDEP_OPS))
|
goto normal_char;
|
goto normal_char;
|
}
|
}
|
|
|
{
|
{
|
/* Are we optimizing this jump? */
|
/* Are we optimizing this jump? */
|
boolean keep_string_p = false;
|
boolean keep_string_p = false;
|
|
|
/* 1 means zero (many) matches is allowed. */
|
/* 1 means zero (many) matches is allowed. */
|
char zero_times_ok = 0, many_times_ok = 0;
|
char zero_times_ok = 0, many_times_ok = 0;
|
|
|
/* If there is a sequence of repetition chars, collapse it
|
/* If there is a sequence of repetition chars, collapse it
|
down to just one (the right one). We can't combine
|
down to just one (the right one). We can't combine
|
interval operators with these because of, e.g., `a{2}*',
|
interval operators with these because of, e.g., `a{2}*',
|
which should only match an even number of `a's. */
|
which should only match an even number of `a's. */
|
|
|
for (;;)
|
for (;;)
|
{
|
{
|
zero_times_ok |= c != '+';
|
zero_times_ok |= c != '+';
|
many_times_ok |= c != '?';
|
many_times_ok |= c != '?';
|
|
|
if (p == pend)
|
if (p == pend)
|
break;
|
break;
|
|
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
if (c == '*'
|
if (c == '*'
|
|| (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?')))
|
|| (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?')))
|
;
|
;
|
|
|
else if (syntax & RE_BK_PLUS_QM && c == '\\')
|
else if (syntax & RE_BK_PLUS_QM && c == '\\')
|
{
|
{
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
|
|
PATFETCH (c1);
|
PATFETCH (c1);
|
if (!(c1 == '+' || c1 == '?'))
|
if (!(c1 == '+' || c1 == '?'))
|
{
|
{
|
PATUNFETCH;
|
PATUNFETCH;
|
PATUNFETCH;
|
PATUNFETCH;
|
break;
|
break;
|
}
|
}
|
|
|
c = c1;
|
c = c1;
|
}
|
}
|
else
|
else
|
{
|
{
|
PATUNFETCH;
|
PATUNFETCH;
|
break;
|
break;
|
}
|
}
|
|
|
/* If we get here, we found another repeat character. */
|
/* If we get here, we found another repeat character. */
|
}
|
}
|
|
|
/* Star, etc. applied to an empty pattern is equivalent
|
/* Star, etc. applied to an empty pattern is equivalent
|
to an empty pattern. */
|
to an empty pattern. */
|
if (!laststart)
|
if (!laststart)
|
break;
|
break;
|
|
|
/* Now we know whether or not zero matches is allowed
|
/* Now we know whether or not zero matches is allowed
|
and also whether or not two or more matches is allowed. */
|
and also whether or not two or more matches is allowed. */
|
if (many_times_ok)
|
if (many_times_ok)
|
{ /* More than one repetition is allowed, so put in at the
|
{ /* More than one repetition is allowed, so put in at the
|
end a backward relative jump from `b' to before the next
|
end a backward relative jump from `b' to before the next
|
jump we're going to put in below (which jumps from
|
jump we're going to put in below (which jumps from
|
laststart to after this jump).
|
laststart to after this jump).
|
|
|
But if we are at the `*' in the exact sequence `.*\n',
|
But if we are at the `*' in the exact sequence `.*\n',
|
insert an unconditional jump backwards to the .,
|
insert an unconditional jump backwards to the .,
|
instead of the beginning of the loop. This way we only
|
instead of the beginning of the loop. This way we only
|
push a failure point once, instead of every time
|
push a failure point once, instead of every time
|
through the loop. */
|
through the loop. */
|
assert (p - 1 > pattern);
|
assert (p - 1 > pattern);
|
|
|
/* Allocate the space for the jump. */
|
/* Allocate the space for the jump. */
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
|
|
/* We know we are not at the first character of the pattern,
|
/* We know we are not at the first character of the pattern,
|
because laststart was nonzero. And we've already
|
because laststart was nonzero. And we've already
|
incremented `p', by the way, to be the character after
|
incremented `p', by the way, to be the character after
|
the `*'. Do we have to do something analogous here
|
the `*'. Do we have to do something analogous here
|
for null bytes, because of RE_DOT_NOT_NULL? */
|
for null bytes, because of RE_DOT_NOT_NULL? */
|
if (TRANSLATE (*(p - 2)) == TRANSLATE ('.')
|
if (TRANSLATE (*(p - 2)) == TRANSLATE ('.')
|
&& zero_times_ok
|
&& zero_times_ok
|
&& p < pend && TRANSLATE (*p) == TRANSLATE ('\n')
|
&& p < pend && TRANSLATE (*p) == TRANSLATE ('\n')
|
&& !(syntax & RE_DOT_NEWLINE))
|
&& !(syntax & RE_DOT_NEWLINE))
|
{ /* We have .*\n. */
|
{ /* We have .*\n. */
|
STORE_JUMP (jump, b, laststart);
|
STORE_JUMP (jump, b, laststart);
|
keep_string_p = true;
|
keep_string_p = true;
|
}
|
}
|
else
|
else
|
/* Anything else. */
|
/* Anything else. */
|
STORE_JUMP (maybe_pop_jump, b, laststart - 3);
|
STORE_JUMP (maybe_pop_jump, b, laststart - 3);
|
|
|
/* We've added more stuff to the buffer. */
|
/* We've added more stuff to the buffer. */
|
b += 3;
|
b += 3;
|
}
|
}
|
|
|
/* On failure, jump from laststart to b + 3, which will be the
|
/* On failure, jump from laststart to b + 3, which will be the
|
end of the buffer after this jump is inserted. */
|
end of the buffer after this jump is inserted. */
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump
|
INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump
|
: on_failure_jump,
|
: on_failure_jump,
|
laststart, b + 3);
|
laststart, b + 3);
|
pending_exact = 0;
|
pending_exact = 0;
|
b += 3;
|
b += 3;
|
|
|
if (!zero_times_ok)
|
if (!zero_times_ok)
|
{
|
{
|
/* At least one repetition is required, so insert a
|
/* At least one repetition is required, so insert a
|
`dummy_failure_jump' before the initial
|
`dummy_failure_jump' before the initial
|
`on_failure_jump' instruction of the loop. This
|
`on_failure_jump' instruction of the loop. This
|
effects a skip over that instruction the first time
|
effects a skip over that instruction the first time
|
we hit that loop. */
|
we hit that loop. */
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6);
|
INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6);
|
b += 3;
|
b += 3;
|
}
|
}
|
}
|
}
|
break;
|
break;
|
|
|
|
|
case '.':
|
case '.':
|
laststart = b;
|
laststart = b;
|
BUF_PUSH (anychar);
|
BUF_PUSH (anychar);
|
break;
|
break;
|
|
|
|
|
case '[':
|
case '[':
|
{
|
{
|
boolean had_char_class = false;
|
boolean had_char_class = false;
|
|
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
|
|
/* Ensure that we have enough space to push a charset: the
|
/* Ensure that we have enough space to push a charset: the
|
opcode, the length count, and the bitset; 34 bytes in all. */
|
opcode, the length count, and the bitset; 34 bytes in all. */
|
GET_BUFFER_SPACE (34);
|
GET_BUFFER_SPACE (34);
|
|
|
laststart = b;
|
laststart = b;
|
|
|
/* We test `*p == '^' twice, instead of using an if
|
/* We test `*p == '^' twice, instead of using an if
|
statement, so we only need one BUF_PUSH. */
|
statement, so we only need one BUF_PUSH. */
|
BUF_PUSH (*p == '^' ? charset_not : charset);
|
BUF_PUSH (*p == '^' ? charset_not : charset);
|
if (*p == '^')
|
if (*p == '^')
|
p++;
|
p++;
|
|
|
/* Remember the first position in the bracket expression. */
|
/* Remember the first position in the bracket expression. */
|
p1 = p;
|
p1 = p;
|
|
|
/* Push the number of bytes in the bitmap. */
|
/* Push the number of bytes in the bitmap. */
|
BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH);
|
BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH);
|
|
|
/* Clear the whole map. */
|
/* Clear the whole map. */
|
bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH);
|
bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH);
|
|
|
/* charset_not matches newline according to a syntax bit. */
|
/* charset_not matches newline according to a syntax bit. */
|
if ((re_opcode_t) b[-2] == charset_not
|
if ((re_opcode_t) b[-2] == charset_not
|
&& (syntax & RE_HAT_LISTS_NOT_NEWLINE))
|
&& (syntax & RE_HAT_LISTS_NOT_NEWLINE))
|
SET_LIST_BIT ('\n');
|
SET_LIST_BIT ('\n');
|
|
|
/* Read in characters and ranges, setting map bits. */
|
/* Read in characters and ranges, setting map bits. */
|
for (;;)
|
for (;;)
|
{
|
{
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
|
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
/* \ might escape characters inside [...] and [^...]. */
|
/* \ might escape characters inside [...] and [^...]. */
|
if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\')
|
if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\')
|
{
|
{
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
|
|
PATFETCH (c1);
|
PATFETCH (c1);
|
SET_LIST_BIT (c1);
|
SET_LIST_BIT (c1);
|
continue;
|
continue;
|
}
|
}
|
|
|
/* Could be the end of the bracket expression. If it's
|
/* Could be the end of the bracket expression. If it's
|
not (i.e., when the bracket expression is `[]' so
|
not (i.e., when the bracket expression is `[]' so
|
far), the ']' character bit gets set way below. */
|
far), the ']' character bit gets set way below. */
|
if (c == ']' && p != p1 + 1)
|
if (c == ']' && p != p1 + 1)
|
break;
|
break;
|
|
|
/* Look ahead to see if it's a range when the last thing
|
/* Look ahead to see if it's a range when the last thing
|
was a character class. */
|
was a character class. */
|
if (had_char_class && c == '-' && *p != ']')
|
if (had_char_class && c == '-' && *p != ']')
|
FREE_STACK_RETURN (REG_ERANGE);
|
FREE_STACK_RETURN (REG_ERANGE);
|
|
|
/* Look ahead to see if it's a range when the last thing
|
/* Look ahead to see if it's a range when the last thing
|
was a character: if this is a hyphen not at the
|
was a character: if this is a hyphen not at the
|
beginning or the end of a list, then it's the range
|
beginning or the end of a list, then it's the range
|
operator. */
|
operator. */
|
if (c == '-'
|
if (c == '-'
|
&& !(p - 2 >= pattern && p[-2] == '[')
|
&& !(p - 2 >= pattern && p[-2] == '[')
|
&& !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^')
|
&& !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^')
|
&& *p != ']')
|
&& *p != ']')
|
{
|
{
|
reg_errcode_t ret
|
reg_errcode_t ret
|
= compile_range (&p, pend, translate, syntax, b);
|
= compile_range (&p, pend, translate, syntax, b);
|
if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
|
if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
|
}
|
}
|
|
|
else if (p[0] == '-' && p[1] != ']')
|
else if (p[0] == '-' && p[1] != ']')
|
{ /* This handles ranges made up of characters only. */
|
{ /* This handles ranges made up of characters only. */
|
reg_errcode_t ret;
|
reg_errcode_t ret;
|
|
|
/* Move past the `-'. */
|
/* Move past the `-'. */
|
PATFETCH (c1);
|
PATFETCH (c1);
|
|
|
ret = compile_range (&p, pend, translate, syntax, b);
|
ret = compile_range (&p, pend, translate, syntax, b);
|
if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
|
if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
|
}
|
}
|
|
|
/* See if we're at the beginning of a possible character
|
/* See if we're at the beginning of a possible character
|
class. */
|
class. */
|
|
|
else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':')
|
else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':')
|
{ /* Leave room for the null. */
|
{ /* Leave room for the null. */
|
char str[CHAR_CLASS_MAX_LENGTH + 1];
|
char str[CHAR_CLASS_MAX_LENGTH + 1];
|
|
|
PATFETCH (c);
|
PATFETCH (c);
|
c1 = 0;
|
c1 = 0;
|
|
|
/* If pattern is `[[:'. */
|
/* If pattern is `[[:'. */
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
|
|
for (;;)
|
for (;;)
|
{
|
{
|
PATFETCH (c);
|
PATFETCH (c);
|
if ((c == ':' && *p == ']') || p == pend
|
if ((c == ':' && *p == ']') || p == pend
|
|| c1 == CHAR_CLASS_MAX_LENGTH)
|
|| c1 == CHAR_CLASS_MAX_LENGTH)
|
break;
|
break;
|
str[c1++] = c;
|
str[c1++] = c;
|
}
|
}
|
str[c1] = '\0';
|
str[c1] = '\0';
|
|
|
/* If isn't a word bracketed by `[:' and `:]':
|
/* If isn't a word bracketed by `[:' and `:]':
|
undo the ending character, the letters, and leave
|
undo the ending character, the letters, and leave
|
the leading `:' and `[' (but set bits for them). */
|
the leading `:' and `[' (but set bits for them). */
|
if (c == ':' && *p == ']')
|
if (c == ':' && *p == ']')
|
{
|
{
|
/* CYGNUS LOCAL: Skip this code if we don't have btowc(). btowc() is */
|
/* CYGNUS LOCAL: Skip this code if we don't have btowc(). btowc() is */
|
/* defined in the 1994 Amendment 1 to ISO C and may not be present on */
|
/* defined in the 1994 Amendment 1 to ISO C and may not be present on */
|
/* systems where we have wchar.h and wctype.h. */
|
/* systems where we have wchar.h and wctype.h. */
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H && defined HAVE_BTOWC)
|
#if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H && defined HAVE_BTOWC)
|
boolean is_lower = STREQ (str, "lower");
|
boolean is_lower = STREQ (str, "lower");
|
boolean is_upper = STREQ (str, "upper");
|
boolean is_upper = STREQ (str, "upper");
|
wctype_t wt;
|
wctype_t wt;
|
int ch;
|
int ch;
|
|
|
wt = IS_CHAR_CLASS (str);
|
wt = IS_CHAR_CLASS (str);
|
if (wt == 0)
|
if (wt == 0)
|
FREE_STACK_RETURN (REG_ECTYPE);
|
FREE_STACK_RETURN (REG_ECTYPE);
|
|
|
/* Throw away the ] at the end of the character
|
/* Throw away the ] at the end of the character
|
class. */
|
class. */
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
|
|
for (ch = 0; ch < 1 << BYTEWIDTH; ++ch)
|
for (ch = 0; ch < 1 << BYTEWIDTH; ++ch)
|
{
|
{
|
# ifdef _LIBC
|
# ifdef _LIBC
|
if (__iswctype (__btowc (ch), wt))
|
if (__iswctype (__btowc (ch), wt))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
#else
|
#else
|
if (iswctype (btowc (ch), wt))
|
if (iswctype (btowc (ch), wt))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
#endif
|
#endif
|
|
|
if (translate && (is_upper || is_lower)
|
if (translate && (is_upper || is_lower)
|
&& (ISUPPER (ch) || ISLOWER (ch)))
|
&& (ISUPPER (ch) || ISLOWER (ch)))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
}
|
}
|
|
|
had_char_class = true;
|
had_char_class = true;
|
#else
|
#else
|
int ch;
|
int ch;
|
boolean is_alnum = STREQ (str, "alnum");
|
boolean is_alnum = STREQ (str, "alnum");
|
boolean is_alpha = STREQ (str, "alpha");
|
boolean is_alpha = STREQ (str, "alpha");
|
boolean is_blank = STREQ (str, "blank");
|
boolean is_blank = STREQ (str, "blank");
|
boolean is_cntrl = STREQ (str, "cntrl");
|
boolean is_cntrl = STREQ (str, "cntrl");
|
boolean is_digit = STREQ (str, "digit");
|
boolean is_digit = STREQ (str, "digit");
|
boolean is_graph = STREQ (str, "graph");
|
boolean is_graph = STREQ (str, "graph");
|
boolean is_lower = STREQ (str, "lower");
|
boolean is_lower = STREQ (str, "lower");
|
boolean is_print = STREQ (str, "print");
|
boolean is_print = STREQ (str, "print");
|
boolean is_punct = STREQ (str, "punct");
|
boolean is_punct = STREQ (str, "punct");
|
boolean is_space = STREQ (str, "space");
|
boolean is_space = STREQ (str, "space");
|
boolean is_upper = STREQ (str, "upper");
|
boolean is_upper = STREQ (str, "upper");
|
boolean is_xdigit = STREQ (str, "xdigit");
|
boolean is_xdigit = STREQ (str, "xdigit");
|
|
|
if (!IS_CHAR_CLASS (str))
|
if (!IS_CHAR_CLASS (str))
|
FREE_STACK_RETURN (REG_ECTYPE);
|
FREE_STACK_RETURN (REG_ECTYPE);
|
|
|
/* Throw away the ] at the end of the character
|
/* Throw away the ] at the end of the character
|
class. */
|
class. */
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
|
|
|
for (ch = 0; ch < 1 << BYTEWIDTH; ch++)
|
for (ch = 0; ch < 1 << BYTEWIDTH; ch++)
|
{
|
{
|
/* This was split into 3 if's to
|
/* This was split into 3 if's to
|
avoid an arbitrary limit in some compiler. */
|
avoid an arbitrary limit in some compiler. */
|
if ( (is_alnum && ISALNUM (ch))
|
if ( (is_alnum && ISALNUM (ch))
|
|| (is_alpha && ISALPHA (ch))
|
|| (is_alpha && ISALPHA (ch))
|
|| (is_blank && ISBLANK (ch))
|
|| (is_blank && ISBLANK (ch))
|
|| (is_cntrl && ISCNTRL (ch)))
|
|| (is_cntrl && ISCNTRL (ch)))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
if ( (is_digit && ISDIGIT (ch))
|
if ( (is_digit && ISDIGIT (ch))
|
|| (is_graph && ISGRAPH (ch))
|
|| (is_graph && ISGRAPH (ch))
|
|| (is_lower && ISLOWER (ch))
|
|| (is_lower && ISLOWER (ch))
|
|| (is_print && ISPRINT (ch)))
|
|| (is_print && ISPRINT (ch)))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
if ( (is_punct && ISPUNCT (ch))
|
if ( (is_punct && ISPUNCT (ch))
|
|| (is_space && ISSPACE (ch))
|
|| (is_space && ISSPACE (ch))
|
|| (is_upper && ISUPPER (ch))
|
|| (is_upper && ISUPPER (ch))
|
|| (is_xdigit && ISXDIGIT (ch)))
|
|| (is_xdigit && ISXDIGIT (ch)))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
if ( translate && (is_upper || is_lower)
|
if ( translate && (is_upper || is_lower)
|
&& (ISUPPER (ch) || ISLOWER (ch)))
|
&& (ISUPPER (ch) || ISLOWER (ch)))
|
SET_LIST_BIT (ch);
|
SET_LIST_BIT (ch);
|
}
|
}
|
had_char_class = true;
|
had_char_class = true;
|
#endif /* libc || wctype.h */
|
#endif /* libc || wctype.h */
|
}
|
}
|
else
|
else
|
{
|
{
|
c1++;
|
c1++;
|
while (c1--)
|
while (c1--)
|
PATUNFETCH;
|
PATUNFETCH;
|
SET_LIST_BIT ('[');
|
SET_LIST_BIT ('[');
|
SET_LIST_BIT (':');
|
SET_LIST_BIT (':');
|
had_char_class = false;
|
had_char_class = false;
|
}
|
}
|
}
|
}
|
else
|
else
|
{
|
{
|
had_char_class = false;
|
had_char_class = false;
|
SET_LIST_BIT (c);
|
SET_LIST_BIT (c);
|
}
|
}
|
}
|
}
|
|
|
/* Discard any (non)matching list bytes that are all 0 at the
|
/* Discard any (non)matching list bytes that are all 0 at the
|
end of the map. Decrease the map-length byte too. */
|
end of the map. Decrease the map-length byte too. */
|
while ((int) b[-1] > 0 && b[b[-1] - 1] == 0)
|
while ((int) b[-1] > 0 && b[b[-1] - 1] == 0)
|
b[-1]--;
|
b[-1]--;
|
b += b[-1];
|
b += b[-1];
|
}
|
}
|
break;
|
break;
|
|
|
|
|
case '(':
|
case '(':
|
if (syntax & RE_NO_BK_PARENS)
|
if (syntax & RE_NO_BK_PARENS)
|
goto handle_open;
|
goto handle_open;
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
|
|
|
|
case ')':
|
case ')':
|
if (syntax & RE_NO_BK_PARENS)
|
if (syntax & RE_NO_BK_PARENS)
|
goto handle_close;
|
goto handle_close;
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
|
|
|
|
case '\n':
|
case '\n':
|
if (syntax & RE_NEWLINE_ALT)
|
if (syntax & RE_NEWLINE_ALT)
|
goto handle_alt;
|
goto handle_alt;
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
|
|
|
|
case '|':
|
case '|':
|
if (syntax & RE_NO_BK_VBAR)
|
if (syntax & RE_NO_BK_VBAR)
|
goto handle_alt;
|
goto handle_alt;
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
|
|
|
|
case '{':
|
case '{':
|
if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES)
|
if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES)
|
goto handle_interval;
|
goto handle_interval;
|
else
|
else
|
goto normal_char;
|
goto normal_char;
|
|
|
|
|
case '\\':
|
case '\\':
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
|
|
|
/* Do not translate the character after the \, so that we can
|
/* Do not translate the character after the \, so that we can
|
distinguish, e.g., \B from \b, even if we normally would
|
distinguish, e.g., \B from \b, even if we normally would
|
translate, e.g., B to b. */
|
translate, e.g., B to b. */
|
PATFETCH_RAW (c);
|
PATFETCH_RAW (c);
|
|
|
switch (c)
|
switch (c)
|
{
|
{
|
case '(':
|
case '(':
|
if (syntax & RE_NO_BK_PARENS)
|
if (syntax & RE_NO_BK_PARENS)
|
goto normal_backslash;
|
goto normal_backslash;
|
|
|
handle_open:
|
handle_open:
|
bufp->re_nsub++;
|
bufp->re_nsub++;
|
regnum++;
|
regnum++;
|
|
|
if (COMPILE_STACK_FULL)
|
if (COMPILE_STACK_FULL)
|
{
|
{
|
RETALLOC (compile_stack.stack, compile_stack.size << 1,
|
RETALLOC (compile_stack.stack, compile_stack.size << 1,
|
compile_stack_elt_t);
|
compile_stack_elt_t);
|
if (compile_stack.stack == NULL) return REG_ESPACE;
|
if (compile_stack.stack == NULL) return REG_ESPACE;
|
|
|
compile_stack.size <<= 1;
|
compile_stack.size <<= 1;
|
}
|
}
|
|
|
/* These are the values to restore when we hit end of this
|
/* These are the values to restore when we hit end of this
|
group. They are all relative offsets, so that if the
|
group. They are all relative offsets, so that if the
|
whole pattern moves because of realloc, they will still
|
whole pattern moves because of realloc, they will still
|
be valid. */
|
be valid. */
|
COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer;
|
COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer;
|
COMPILE_STACK_TOP.fixup_alt_jump
|
COMPILE_STACK_TOP.fixup_alt_jump
|
= fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0;
|
= fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0;
|
COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer;
|
COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer;
|
COMPILE_STACK_TOP.regnum = regnum;
|
COMPILE_STACK_TOP.regnum = regnum;
|
|
|
/* We will eventually replace the 0 with the number of
|
/* We will eventually replace the 0 with the number of
|
groups inner to this one. But do not push a
|
groups inner to this one. But do not push a
|
start_memory for groups beyond the last one we can
|
start_memory for groups beyond the last one we can
|
represent in the compiled pattern. */
|
represent in the compiled pattern. */
|
if (regnum <= MAX_REGNUM)
|
if (regnum <= MAX_REGNUM)
|
{
|
{
|
COMPILE_STACK_TOP.inner_group_offset = b - bufp->buffer + 2;
|
COMPILE_STACK_TOP.inner_group_offset = b - bufp->buffer + 2;
|
BUF_PUSH_3 (start_memory, regnum, 0);
|
BUF_PUSH_3 (start_memory, regnum, 0);
|
}
|
}
|
|
|
compile_stack.avail++;
|
compile_stack.avail++;
|
|
|
fixup_alt_jump = 0;
|
fixup_alt_jump = 0;
|
laststart = 0;
|
laststart = 0;
|
begalt = b;
|
begalt = b;
|
/* If we've reached MAX_REGNUM groups, then this open
|
/* If we've reached MAX_REGNUM groups, then this open
|
won't actually generate any code, so we'll have to
|
won't actually generate any code, so we'll have to
|
clear pending_exact explicitly. */
|
clear pending_exact explicitly. */
|
pending_exact = 0;
|
pending_exact = 0;
|
break;
|
break;
|
|
|
|
|
case ')':
|
case ')':
|
if (syntax & RE_NO_BK_PARENS) goto normal_backslash;
|
if (syntax & RE_NO_BK_PARENS) goto normal_backslash;
|
|
|
if (COMPILE_STACK_EMPTY)
|
if (COMPILE_STACK_EMPTY)
|
{
|
{
|
if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
|
if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
|
goto normal_backslash;
|
goto normal_backslash;
|
else
|
else
|
FREE_STACK_RETURN (REG_ERPAREN);
|
FREE_STACK_RETURN (REG_ERPAREN);
|
}
|
}
|
|
|
handle_close:
|
handle_close:
|
if (fixup_alt_jump)
|
if (fixup_alt_jump)
|
{ /* Push a dummy failure point at the end of the
|
{ /* Push a dummy failure point at the end of the
|
alternative for a possible future
|
alternative for a possible future
|
`pop_failure_jump' to pop. See comments at
|
`pop_failure_jump' to pop. See comments at
|
`push_dummy_failure' in `re_match_2'. */
|
`push_dummy_failure' in `re_match_2'. */
|
BUF_PUSH (push_dummy_failure);
|
BUF_PUSH (push_dummy_failure);
|
|
|
/* We allocated space for this jump when we assigned
|
/* We allocated space for this jump when we assigned
|
to `fixup_alt_jump', in the `handle_alt' case below. */
|
to `fixup_alt_jump', in the `handle_alt' case below. */
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1);
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1);
|
}
|
}
|
|
|
/* See similar code for backslashed left paren above. */
|
/* See similar code for backslashed left paren above. */
|
if (COMPILE_STACK_EMPTY)
|
if (COMPILE_STACK_EMPTY)
|
{
|
{
|
if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
|
if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
|
goto normal_char;
|
goto normal_char;
|
else
|
else
|
FREE_STACK_RETURN (REG_ERPAREN);
|
FREE_STACK_RETURN (REG_ERPAREN);
|
}
|
}
|
|
|
/* Since we just checked for an empty stack above, this
|
/* Since we just checked for an empty stack above, this
|
``can't happen''. */
|
``can't happen''. */
|
assert (compile_stack.avail != 0);
|
assert (compile_stack.avail != 0);
|
{
|
{
|
/* We don't just want to restore into `regnum', because
|
/* We don't just want to restore into `regnum', because
|
later groups should continue to be numbered higher,
|
later groups should continue to be numbered higher,
|
as in `(ab)c(de)' -- the second group is #2. */
|
as in `(ab)c(de)' -- the second group is #2. */
|
regnum_t this_group_regnum;
|
regnum_t this_group_regnum;
|
|
|
compile_stack.avail--;
|
compile_stack.avail--;
|
begalt = bufp->buffer + COMPILE_STACK_TOP.begalt_offset;
|
begalt = bufp->buffer + COMPILE_STACK_TOP.begalt_offset;
|
fixup_alt_jump
|
fixup_alt_jump
|
= COMPILE_STACK_TOP.fixup_alt_jump
|
= COMPILE_STACK_TOP.fixup_alt_jump
|
? bufp->buffer + COMPILE_STACK_TOP.fixup_alt_jump - 1
|
? bufp->buffer + COMPILE_STACK_TOP.fixup_alt_jump - 1
|
: 0;
|
: 0;
|
laststart = bufp->buffer + COMPILE_STACK_TOP.laststart_offset;
|
laststart = bufp->buffer + COMPILE_STACK_TOP.laststart_offset;
|
this_group_regnum = COMPILE_STACK_TOP.regnum;
|
this_group_regnum = COMPILE_STACK_TOP.regnum;
|
/* If we've reached MAX_REGNUM groups, then this open
|
/* If we've reached MAX_REGNUM groups, then this open
|
won't actually generate any code, so we'll have to
|
won't actually generate any code, so we'll have to
|
clear pending_exact explicitly. */
|
clear pending_exact explicitly. */
|
pending_exact = 0;
|
pending_exact = 0;
|
|
|
/* We're at the end of the group, so now we know how many
|
/* We're at the end of the group, so now we know how many
|
groups were inside this one. */
|
groups were inside this one. */
|
if (this_group_regnum <= MAX_REGNUM)
|
if (this_group_regnum <= MAX_REGNUM)
|
{
|
{
|
unsigned char *inner_group_loc
|
unsigned char *inner_group_loc
|
= bufp->buffer + COMPILE_STACK_TOP.inner_group_offset;
|
= bufp->buffer + COMPILE_STACK_TOP.inner_group_offset;
|
|
|
*inner_group_loc = regnum - this_group_regnum;
|
*inner_group_loc = regnum - this_group_regnum;
|
BUF_PUSH_3 (stop_memory, this_group_regnum,
|
BUF_PUSH_3 (stop_memory, this_group_regnum,
|
regnum - this_group_regnum);
|
regnum - this_group_regnum);
|
}
|
}
|
}
|
}
|
break;
|
break;
|
|
|
|
|
case '|': /* `\|'. */
|
case '|': /* `\|'. */
|
if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR)
|
if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR)
|
goto normal_backslash;
|
goto normal_backslash;
|
handle_alt:
|
handle_alt:
|
if (syntax & RE_LIMITED_OPS)
|
if (syntax & RE_LIMITED_OPS)
|
goto normal_char;
|
goto normal_char;
|
|
|
/* Insert before the previous alternative a jump which
|
/* Insert before the previous alternative a jump which
|
jumps to this alternative if the former fails. */
|
jumps to this alternative if the former fails. */
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
INSERT_JUMP (on_failure_jump, begalt, b + 6);
|
INSERT_JUMP (on_failure_jump, begalt, b + 6);
|
pending_exact = 0;
|
pending_exact = 0;
|
b += 3;
|
b += 3;
|
|
|
/* The alternative before this one has a jump after it
|
/* The alternative before this one has a jump after it
|
which gets executed if it gets matched. Adjust that
|
which gets executed if it gets matched. Adjust that
|
jump so it will jump to this alternative's analogous
|
jump so it will jump to this alternative's analogous
|
jump (put in below, which in turn will jump to the next
|
jump (put in below, which in turn will jump to the next
|
(if any) alternative's such jump, etc.). The last such
|
(if any) alternative's such jump, etc.). The last such
|
jump jumps to the correct final destination. A picture:
|
jump jumps to the correct final destination. A picture:
|
_____ _____
|
_____ _____
|
| | | |
|
| | | |
|
| v | v
|
| v | v
|
a | b | c
|
a | b | c
|
|
|
If we are at `b', then fixup_alt_jump right now points to a
|
If we are at `b', then fixup_alt_jump right now points to a
|
three-byte space after `a'. We'll put in the jump, set
|
three-byte space after `a'. We'll put in the jump, set
|
fixup_alt_jump to right after `b', and leave behind three
|
fixup_alt_jump to right after `b', and leave behind three
|
bytes which we'll fill in when we get to after `c'. */
|
bytes which we'll fill in when we get to after `c'. */
|
|
|
if (fixup_alt_jump)
|
if (fixup_alt_jump)
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
|
|
|
/* Mark and leave space for a jump after this alternative,
|
/* Mark and leave space for a jump after this alternative,
|
to be filled in later either by next alternative or
|
to be filled in later either by next alternative or
|
when know we're at the end of a series of alternatives. */
|
when know we're at the end of a series of alternatives. */
|
fixup_alt_jump = b;
|
fixup_alt_jump = b;
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
b += 3;
|
b += 3;
|
|
|
laststart = 0;
|
laststart = 0;
|
begalt = b;
|
begalt = b;
|
break;
|
break;
|
|
|
|
|
case '{':
|
case '{':
|
/* If \{ is a literal. */
|
/* If \{ is a literal. */
|
if (!(syntax & RE_INTERVALS)
|
if (!(syntax & RE_INTERVALS)
|
/* If we're at `\{' and it's not the open-interval
|
/* If we're at `\{' and it's not the open-interval
|
operator. */
|
operator. */
|
|| ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES))
|
|| ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES))
|
|| (p - 2 == pattern && p == pend))
|
|| (p - 2 == pattern && p == pend))
|
goto normal_backslash;
|
goto normal_backslash;
|
|
|
handle_interval:
|
handle_interval:
|
{
|
{
|
/* If got here, then the syntax allows intervals. */
|
/* If got here, then the syntax allows intervals. */
|
|
|
/* At least (most) this many matches must be made. */
|
/* At least (most) this many matches must be made. */
|
int lower_bound = -1, upper_bound = -1;
|
int lower_bound = -1, upper_bound = -1;
|
|
|
beg_interval = p - 1;
|
beg_interval = p - 1;
|
|
|
if (p == pend)
|
if (p == pend)
|
{
|
{
|
if (syntax & RE_NO_BK_BRACES)
|
if (syntax & RE_NO_BK_BRACES)
|
goto unfetch_interval;
|
goto unfetch_interval;
|
else
|
else
|
FREE_STACK_RETURN (REG_EBRACE);
|
FREE_STACK_RETURN (REG_EBRACE);
|
}
|
}
|
|
|
GET_UNSIGNED_NUMBER (lower_bound);
|
GET_UNSIGNED_NUMBER (lower_bound);
|
|
|
if (c == ',')
|
if (c == ',')
|
{
|
{
|
GET_UNSIGNED_NUMBER (upper_bound);
|
GET_UNSIGNED_NUMBER (upper_bound);
|
if (upper_bound < 0) upper_bound = RE_DUP_MAX;
|
if (upper_bound < 0) upper_bound = RE_DUP_MAX;
|
}
|
}
|
else
|
else
|
/* Interval such as `{1}' => match exactly once. */
|
/* Interval such as `{1}' => match exactly once. */
|
upper_bound = lower_bound;
|
upper_bound = lower_bound;
|
|
|
if (lower_bound < 0 || upper_bound > RE_DUP_MAX
|
if (lower_bound < 0 || upper_bound > RE_DUP_MAX
|
|| lower_bound > upper_bound)
|
|| lower_bound > upper_bound)
|
{
|
{
|
if (syntax & RE_NO_BK_BRACES)
|
if (syntax & RE_NO_BK_BRACES)
|
goto unfetch_interval;
|
goto unfetch_interval;
|
else
|
else
|
FREE_STACK_RETURN (REG_BADBR);
|
FREE_STACK_RETURN (REG_BADBR);
|
}
|
}
|
|
|
if (!(syntax & RE_NO_BK_BRACES))
|
if (!(syntax & RE_NO_BK_BRACES))
|
{
|
{
|
if (c != '\\') FREE_STACK_RETURN (REG_EBRACE);
|
if (c != '\\') FREE_STACK_RETURN (REG_EBRACE);
|
|
|
PATFETCH (c);
|
PATFETCH (c);
|
}
|
}
|
|
|
if (c != '}')
|
if (c != '}')
|
{
|
{
|
if (syntax & RE_NO_BK_BRACES)
|
if (syntax & RE_NO_BK_BRACES)
|
goto unfetch_interval;
|
goto unfetch_interval;
|
else
|
else
|
FREE_STACK_RETURN (REG_BADBR);
|
FREE_STACK_RETURN (REG_BADBR);
|
}
|
}
|
|
|
/* We just parsed a valid interval. */
|
/* We just parsed a valid interval. */
|
|
|
/* If it's invalid to have no preceding re. */
|
/* If it's invalid to have no preceding re. */
|
if (!laststart)
|
if (!laststart)
|
{
|
{
|
if (syntax & RE_CONTEXT_INVALID_OPS)
|
if (syntax & RE_CONTEXT_INVALID_OPS)
|
FREE_STACK_RETURN (REG_BADRPT);
|
FREE_STACK_RETURN (REG_BADRPT);
|
else if (syntax & RE_CONTEXT_INDEP_OPS)
|
else if (syntax & RE_CONTEXT_INDEP_OPS)
|
laststart = b;
|
laststart = b;
|
else
|
else
|
goto unfetch_interval;
|
goto unfetch_interval;
|
}
|
}
|
|
|
/* If the upper bound is zero, don't want to succeed at
|
/* If the upper bound is zero, don't want to succeed at
|
all; jump from `laststart' to `b + 3', which will be
|
all; jump from `laststart' to `b + 3', which will be
|
the end of the buffer after we insert the jump. */
|
the end of the buffer after we insert the jump. */
|
if (upper_bound == 0)
|
if (upper_bound == 0)
|
{
|
{
|
GET_BUFFER_SPACE (3);
|
GET_BUFFER_SPACE (3);
|
INSERT_JUMP (jump, laststart, b + 3);
|
INSERT_JUMP (jump, laststart, b + 3);
|
b += 3;
|
b += 3;
|
}
|
}
|
|
|
/* Otherwise, we have a nontrivial interval. When
|
/* Otherwise, we have a nontrivial interval. When
|
we're all done, the pattern will look like:
|
we're all done, the pattern will look like:
|
set_number_at <jump count> <upper bound>
|
set_number_at <jump count> <upper bound>
|
set_number_at <succeed_n count> <lower bound>
|
set_number_at <succeed_n count> <lower bound>
|
succeed_n <after jump addr> <succeed_n count>
|
succeed_n <after jump addr> <succeed_n count>
|
<body of loop>
|
<body of loop>
|
jump_n <succeed_n addr> <jump count>
|
jump_n <succeed_n addr> <jump count>
|
(The upper bound and `jump_n' are omitted if
|
(The upper bound and `jump_n' are omitted if
|
`upper_bound' is 1, though.) */
|
`upper_bound' is 1, though.) */
|
else
|
else
|
{ /* If the upper bound is > 1, we need to insert
|
{ /* If the upper bound is > 1, we need to insert
|
more at the end of the loop. */
|
more at the end of the loop. */
|
unsigned nbytes = 10 + (upper_bound > 1) * 10;
|
unsigned nbytes = 10 + (upper_bound > 1) * 10;
|
|
|
GET_BUFFER_SPACE (nbytes);
|
GET_BUFFER_SPACE (nbytes);
|
|
|
/* Initialize lower bound of the `succeed_n', even
|
/* Initialize lower bound of the `succeed_n', even
|
though it will be set during matching by its
|
though it will be set during matching by its
|
attendant `set_number_at' (inserted next),
|
attendant `set_number_at' (inserted next),
|
because `re_compile_fastmap' needs to know.
|
because `re_compile_fastmap' needs to know.
|
Jump to the `jump_n' we might insert below. */
|
Jump to the `jump_n' we might insert below. */
|
INSERT_JUMP2 (succeed_n, laststart,
|
INSERT_JUMP2 (succeed_n, laststart,
|
b + 5 + (upper_bound > 1) * 5,
|
b + 5 + (upper_bound > 1) * 5,
|
lower_bound);
|
lower_bound);
|
b += 5;
|
b += 5;
|
|
|
/* Code to initialize the lower bound. Insert
|
/* Code to initialize the lower bound. Insert
|
before the `succeed_n'. The `5' is the last two
|
before the `succeed_n'. The `5' is the last two
|
bytes of this `set_number_at', plus 3 bytes of
|
bytes of this `set_number_at', plus 3 bytes of
|
the following `succeed_n'. */
|
the following `succeed_n'. */
|
insert_op2 (set_number_at, laststart, 5, lower_bound, b);
|
insert_op2 (set_number_at, laststart, 5, lower_bound, b);
|
b += 5;
|
b += 5;
|
|
|
if (upper_bound > 1)
|
if (upper_bound > 1)
|
{ /* More than one repetition is allowed, so
|
{ /* More than one repetition is allowed, so
|
append a backward jump to the `succeed_n'
|
append a backward jump to the `succeed_n'
|
that starts this interval.
|
that starts this interval.
|
|
|
When we've reached this during matching,
|
When we've reached this during matching,
|
we'll have matched the interval once, so
|
we'll have matched the interval once, so
|
jump back only `upper_bound - 1' times. */
|
jump back only `upper_bound - 1' times. */
|
STORE_JUMP2 (jump_n, b, laststart + 5,
|
STORE_JUMP2 (jump_n, b, laststart + 5,
|
upper_bound - 1);
|
upper_bound - 1);
|
b += 5;
|
b += 5;
|
|
|
/* The location we want to set is the second
|
/* The location we want to set is the second
|
parameter of the `jump_n'; that is `b-2' as
|
parameter of the `jump_n'; that is `b-2' as
|
an absolute address. `laststart' will be
|
an absolute address. `laststart' will be
|
the `set_number_at' we're about to insert;
|
the `set_number_at' we're about to insert;
|
`laststart+3' the number to set, the source
|
`laststart+3' the number to set, the source
|
for the relative address. But we are
|
for the relative address. But we are
|
inserting into the middle of the pattern --
|
inserting into the middle of the pattern --
|
so everything is getting moved up by 5.
|
so everything is getting moved up by 5.
|
Conclusion: (b - 2) - (laststart + 3) + 5,
|
Conclusion: (b - 2) - (laststart + 3) + 5,
|
i.e., b - laststart.
|
i.e., b - laststart.
|
|
|
We insert this at the beginning of the loop
|
We insert this at the beginning of the loop
|
so that if we fail during matching, we'll
|
so that if we fail during matching, we'll
|
reinitialize the bounds. */
|
reinitialize the bounds. */
|
insert_op2 (set_number_at, laststart, b - laststart,
|
insert_op2 (set_number_at, laststart, b - laststart,
|
upper_bound - 1, b);
|
upper_bound - 1, b);
|
b += 5;
|
b += 5;
|
}
|
}
|
}
|
}
|
pending_exact = 0;
|
pending_exact = 0;
|
beg_interval = NULL;
|
beg_interval = NULL;
|
}
|
}
|
break;
|
break;
|
|
|
unfetch_interval:
|
unfetch_interval:
|
/* If an invalid interval, match the characters as literals. */
|
/* If an invalid interval, match the characters as literals. */
|
assert (beg_interval);
|
assert (beg_interval);
|
p = beg_interval;
|
p = beg_interval;
|
beg_interval = NULL;
|
beg_interval = NULL;
|
|
|
/* normal_char and normal_backslash need `c'. */
|
/* normal_char and normal_backslash need `c'. */
|
PATFETCH (c);
|
PATFETCH (c);
|
|
|
if (!(syntax & RE_NO_BK_BRACES))
|
if (!(syntax & RE_NO_BK_BRACES))
|
{
|
{
|
if (p > pattern && p[-1] == '\\')
|
if (p > pattern && p[-1] == '\\')
|
goto normal_backslash;
|
goto normal_backslash;
|
}
|
}
|
goto normal_char;
|
goto normal_char;
|
|
|
#ifdef emacs
|
#ifdef emacs
|
/* There is no way to specify the before_dot and after_dot
|
/* There is no way to specify the before_dot and after_dot
|
operators. rms says this is ok. --karl */
|
operators. rms says this is ok. --karl */
|
case '=':
|
case '=':
|
BUF_PUSH (at_dot);
|
BUF_PUSH (at_dot);
|
break;
|
break;
|
|
|
case 's':
|
case 's':
|
laststart = b;
|
laststart = b;
|
PATFETCH (c);
|
PATFETCH (c);
|
BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]);
|
BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]);
|
break;
|
break;
|
|
|
case 'S':
|
case 'S':
|
laststart = b;
|
laststart = b;
|
PATFETCH (c);
|
PATFETCH (c);
|
BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]);
|
BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]);
|
break;
|
break;
|
#endif /* emacs */
|
#endif /* emacs */
|
|
|
|
|
case 'w':
|
case 'w':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
laststart = b;
|
laststart = b;
|
BUF_PUSH (wordchar);
|
BUF_PUSH (wordchar);
|
break;
|
break;
|
|
|
|
|
case 'W':
|
case 'W':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
laststart = b;
|
laststart = b;
|
BUF_PUSH (notwordchar);
|
BUF_PUSH (notwordchar);
|
break;
|
break;
|
|
|
|
|
case '<':
|
case '<':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (wordbeg);
|
BUF_PUSH (wordbeg);
|
break;
|
break;
|
|
|
case '>':
|
case '>':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (wordend);
|
BUF_PUSH (wordend);
|
break;
|
break;
|
|
|
case 'b':
|
case 'b':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (wordbound);
|
BUF_PUSH (wordbound);
|
break;
|
break;
|
|
|
case 'B':
|
case 'B':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (notwordbound);
|
BUF_PUSH (notwordbound);
|
break;
|
break;
|
|
|
case '`':
|
case '`':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (begbuf);
|
BUF_PUSH (begbuf);
|
break;
|
break;
|
|
|
case '\'':
|
case '\'':
|
if (syntax & RE_NO_GNU_OPS)
|
if (syntax & RE_NO_GNU_OPS)
|
goto normal_char;
|
goto normal_char;
|
BUF_PUSH (endbuf);
|
BUF_PUSH (endbuf);
|
break;
|
break;
|
|
|
case '1': case '2': case '3': case '4': case '5':
|
case '1': case '2': case '3': case '4': case '5':
|
case '6': case '7': case '8': case '9':
|
case '6': case '7': case '8': case '9':
|
if (syntax & RE_NO_BK_REFS)
|
if (syntax & RE_NO_BK_REFS)
|
goto normal_char;
|
goto normal_char;
|
|
|
c1 = c - '0';
|
c1 = c - '0';
|
|
|
if (c1 > regnum)
|
if (c1 > regnum)
|
FREE_STACK_RETURN (REG_ESUBREG);
|
FREE_STACK_RETURN (REG_ESUBREG);
|
|
|
/* Can't back reference to a subexpression if inside of it. */
|
/* Can't back reference to a subexpression if inside of it. */
|
if (group_in_compile_stack (compile_stack, (regnum_t) c1))
|
if (group_in_compile_stack (compile_stack, (regnum_t) c1))
|
goto normal_char;
|
goto normal_char;
|
|
|
laststart = b;
|
laststart = b;
|
BUF_PUSH_2 (duplicate, c1);
|
BUF_PUSH_2 (duplicate, c1);
|
break;
|
break;
|
|
|
|
|
case '+':
|
case '+':
|
case '?':
|
case '?':
|
if (syntax & RE_BK_PLUS_QM)
|
if (syntax & RE_BK_PLUS_QM)
|
goto handle_plus;
|
goto handle_plus;
|
else
|
else
|
goto normal_backslash;
|
goto normal_backslash;
|
|
|
default:
|
default:
|
normal_backslash:
|
normal_backslash:
|
/* You might think it would be useful for \ to mean
|
/* You might think it would be useful for \ to mean
|
not to translate; but if we don't translate it
|
not to translate; but if we don't translate it
|
it will never match anything. */
|
it will never match anything. */
|
c = TRANSLATE (c);
|
c = TRANSLATE (c);
|
goto normal_char;
|
goto normal_char;
|
}
|
}
|
break;
|
break;
|
|
|
|
|
default:
|
default:
|
/* Expects the character in `c'. */
|
/* Expects the character in `c'. */
|
normal_char:
|
normal_char:
|
/* If no exactn currently being built. */
|
/* If no exactn currently being built. */
|
if (!pending_exact
|
if (!pending_exact
|
|
|
/* If last exactn not at current position. */
|
/* If last exactn not at current position. */
|
|| pending_exact + *pending_exact + 1 != b
|
|| pending_exact + *pending_exact + 1 != b
|
|
|
/* We have only one byte following the exactn for the count. */
|
/* We have only one byte following the exactn for the count. */
|
|| *pending_exact == (1 << BYTEWIDTH) - 1
|
|| *pending_exact == (1 << BYTEWIDTH) - 1
|
|
|
/* If followed by a repetition operator. */
|
/* If followed by a repetition operator. */
|
|| *p == '*' || *p == '^'
|
|| *p == '*' || *p == '^'
|
|| ((syntax & RE_BK_PLUS_QM)
|
|| ((syntax & RE_BK_PLUS_QM)
|
? *p == '\\' && (p[1] == '+' || p[1] == '?')
|
? *p == '\\' && (p[1] == '+' || p[1] == '?')
|
: (*p == '+' || *p == '?'))
|
: (*p == '+' || *p == '?'))
|
|| ((syntax & RE_INTERVALS)
|
|| ((syntax & RE_INTERVALS)
|
&& ((syntax & RE_NO_BK_BRACES)
|
&& ((syntax & RE_NO_BK_BRACES)
|
? *p == '{'
|
? *p == '{'
|
: (p[0] == '\\' && p[1] == '{'))))
|
: (p[0] == '\\' && p[1] == '{'))))
|
{
|
{
|
/* Start building a new exactn. */
|
/* Start building a new exactn. */
|
|
|
laststart = b;
|
laststart = b;
|
|
|
BUF_PUSH_2 (exactn, 0);
|
BUF_PUSH_2 (exactn, 0);
|
pending_exact = b - 1;
|
pending_exact = b - 1;
|
}
|
}
|
|
|
BUF_PUSH (c);
|
BUF_PUSH (c);
|
(*pending_exact)++;
|
(*pending_exact)++;
|
break;
|
break;
|
} /* switch (c) */
|
} /* switch (c) */
|
} /* while p != pend */
|
} /* while p != pend */
|
|
|
|
|
/* Through the pattern now. */
|
/* Through the pattern now. */
|
|
|
if (fixup_alt_jump)
|
if (fixup_alt_jump)
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
|
STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
|
|
|
if (!COMPILE_STACK_EMPTY)
|
if (!COMPILE_STACK_EMPTY)
|
FREE_STACK_RETURN (REG_EPAREN);
|
FREE_STACK_RETURN (REG_EPAREN);
|
|
|
/* If we don't want backtracking, force success
|
/* If we don't want backtracking, force success
|
the first time we reach the end of the compiled pattern. */
|
the first time we reach the end of the compiled pattern. */
|
if (syntax & RE_NO_POSIX_BACKTRACKING)
|
if (syntax & RE_NO_POSIX_BACKTRACKING)
|
BUF_PUSH (succeed);
|
BUF_PUSH (succeed);
|
|
|
free (compile_stack.stack);
|
free (compile_stack.stack);
|
|
|
/* We have succeeded; set the length of the buffer. */
|
/* We have succeeded; set the length of the buffer. */
|
bufp->used = b - bufp->buffer;
|
bufp->used = b - bufp->buffer;
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
if (debug)
|
if (debug)
|
{
|
{
|
DEBUG_PRINT1 ("\nCompiled pattern: \n");
|
DEBUG_PRINT1 ("\nCompiled pattern: \n");
|
print_compiled_pattern (bufp);
|
print_compiled_pattern (bufp);
|
}
|
}
|
#endif /* DEBUG */
|
#endif /* DEBUG */
|
|
|
#ifndef MATCH_MAY_ALLOCATE
|
#ifndef MATCH_MAY_ALLOCATE
|
/* Initialize the failure stack to the largest possible stack. This
|
/* Initialize the failure stack to the largest possible stack. This
|
isn't necessary unless we're trying to avoid calling alloca in
|
isn't necessary unless we're trying to avoid calling alloca in
|
the search and match routines. */
|
the search and match routines. */
|
{
|
{
|
int num_regs = bufp->re_nsub + 1;
|
int num_regs = bufp->re_nsub + 1;
|
|
|
/* Since DOUBLE_FAIL_STACK refuses to double only if the current size
|
/* Since DOUBLE_FAIL_STACK refuses to double only if the current size
|
is strictly greater than re_max_failures, the largest possible stack
|
is strictly greater than re_max_failures, the largest possible stack
|
is 2 * re_max_failures failure points. */
|
is 2 * re_max_failures failure points. */
|
if (fail_stack.size < (2 * re_max_failures * MAX_FAILURE_ITEMS))
|
if (fail_stack.size < (2 * re_max_failures * MAX_FAILURE_ITEMS))
|
{
|
{
|
fail_stack.size = (2 * re_max_failures * MAX_FAILURE_ITEMS);
|
fail_stack.size = (2 * re_max_failures * MAX_FAILURE_ITEMS);
|
|
|
# ifdef emacs
|
# ifdef emacs
|
if (! fail_stack.stack)
|
if (! fail_stack.stack)
|
fail_stack.stack
|
fail_stack.stack
|
= (fail_stack_elt_t *) xmalloc (fail_stack.size
|
= (fail_stack_elt_t *) xmalloc (fail_stack.size
|
* sizeof (fail_stack_elt_t));
|
* sizeof (fail_stack_elt_t));
|
else
|
else
|
fail_stack.stack
|
fail_stack.stack
|
= (fail_stack_elt_t *) xrealloc (fail_stack.stack,
|
= (fail_stack_elt_t *) xrealloc (fail_stack.stack,
|
(fail_stack.size
|
(fail_stack.size
|
* sizeof (fail_stack_elt_t)));
|
* sizeof (fail_stack_elt_t)));
|
# else /* not emacs */
|
# else /* not emacs */
|
if (! fail_stack.stack)
|
if (! fail_stack.stack)
|
fail_stack.stack
|
fail_stack.stack
|
= (fail_stack_elt_t *) malloc (fail_stack.size
|
= (fail_stack_elt_t *) malloc (fail_stack.size
|
* sizeof (fail_stack_elt_t));
|
* sizeof (fail_stack_elt_t));
|
else
|
else
|
fail_stack.stack
|
fail_stack.stack
|
= (fail_stack_elt_t *) realloc (fail_stack.stack,
|
= (fail_stack_elt_t *) realloc (fail_stack.stack,
|
(fail_stack.size
|
(fail_stack.size
|
* sizeof (fail_stack_elt_t)));
|
* sizeof (fail_stack_elt_t)));
|
# endif /* not emacs */
|
# endif /* not emacs */
|
}
|
}
|
|
|
regex_grow_registers (num_regs);
|
regex_grow_registers (num_regs);
|
}
|
}
|
#endif /* not MATCH_MAY_ALLOCATE */
|
#endif /* not MATCH_MAY_ALLOCATE */
|
|
|
return REG_NOERROR;
|
return REG_NOERROR;
|
} /* regex_compile */
|
} /* regex_compile */
|
|
|
/* Subroutines for `regex_compile'. */
|
/* Subroutines for `regex_compile'. */
|
|
|
/* Store OP at LOC followed by two-byte integer parameter ARG. */
|
/* Store OP at LOC followed by two-byte integer parameter ARG. */
|
|
|
static void
|
static void
|
store_op1 (op, loc, arg)
|
store_op1 (op, loc, arg)
|
re_opcode_t op;
|
re_opcode_t op;
|
unsigned char *loc;
|
unsigned char *loc;
|
int arg;
|
int arg;
|
{
|
{
|
*loc = (unsigned char) op;
|
*loc = (unsigned char) op;
|
STORE_NUMBER (loc + 1, arg);
|
STORE_NUMBER (loc + 1, arg);
|
}
|
}
|
|
|
|
|
/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */
|
/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */
|
|
|
static void
|
static void
|
store_op2 (op, loc, arg1, arg2)
|
store_op2 (op, loc, arg1, arg2)
|
re_opcode_t op;
|
re_opcode_t op;
|
unsigned char *loc;
|
unsigned char *loc;
|
int arg1, arg2;
|
int arg1, arg2;
|
{
|
{
|
*loc = (unsigned char) op;
|
*loc = (unsigned char) op;
|
STORE_NUMBER (loc + 1, arg1);
|
STORE_NUMBER (loc + 1, arg1);
|
STORE_NUMBER (loc + 3, arg2);
|
STORE_NUMBER (loc + 3, arg2);
|
}
|
}
|
|
|
|
|
/* Copy the bytes from LOC to END to open up three bytes of space at LOC
|
/* Copy the bytes from LOC to END to open up three bytes of space at LOC
|
for OP followed by two-byte integer parameter ARG. */
|
for OP followed by two-byte integer parameter ARG. */
|
|
|
static void
|
static void
|
insert_op1 (op, loc, arg, end)
|
insert_op1 (op, loc, arg, end)
|
re_opcode_t op;
|
re_opcode_t op;
|
unsigned char *loc;
|
unsigned char *loc;
|
int arg;
|
int arg;
|
unsigned char *end;
|
unsigned char *end;
|
{
|
{
|
register unsigned char *pfrom = end;
|
register unsigned char *pfrom = end;
|
register unsigned char *pto = end + 3;
|
register unsigned char *pto = end + 3;
|
|
|
while (pfrom != loc)
|
while (pfrom != loc)
|
*--pto = *--pfrom;
|
*--pto = *--pfrom;
|
|
|
store_op1 (op, loc, arg);
|
store_op1 (op, loc, arg);
|
}
|
}
|
|
|
|
|
/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */
|
/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */
|
|
|
static void
|
static void
|
insert_op2 (op, loc, arg1, arg2, end)
|
insert_op2 (op, loc, arg1, arg2, end)
|
re_opcode_t op;
|
re_opcode_t op;
|
unsigned char *loc;
|
unsigned char *loc;
|
int arg1, arg2;
|
int arg1, arg2;
|
unsigned char *end;
|
unsigned char *end;
|
{
|
{
|
register unsigned char *pfrom = end;
|
register unsigned char *pfrom = end;
|
register unsigned char *pto = end + 5;
|
register unsigned char *pto = end + 5;
|
|
|
while (pfrom != loc)
|
while (pfrom != loc)
|
*--pto = *--pfrom;
|
*--pto = *--pfrom;
|
|
|
store_op2 (op, loc, arg1, arg2);
|
store_op2 (op, loc, arg1, arg2);
|
}
|
}
|
|
|
|
|
/* P points to just after a ^ in PATTERN. Return true if that ^ comes
|
/* P points to just after a ^ in PATTERN. Return true if that ^ comes
|
after an alternative or a begin-subexpression. We assume there is at
|
after an alternative or a begin-subexpression. We assume there is at
|
least one character before the ^. */
|
least one character before the ^. */
|
|
|
static boolean
|
static boolean
|
at_begline_loc_p (pattern, p, syntax)
|
at_begline_loc_p (pattern, p, syntax)
|
const char *pattern, *p;
|
const char *pattern, *p;
|
reg_syntax_t syntax;
|
reg_syntax_t syntax;
|
{
|
{
|
const char *prev = p - 2;
|
const char *prev = p - 2;
|
boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\';
|
boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\';
|
|
|
return
|
return
|
/* After a subexpression? */
|
/* After a subexpression? */
|
(*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash))
|
(*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash))
|
/* After an alternative? */
|
/* After an alternative? */
|
|| (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash));
|
|| (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash));
|
}
|
}
|
|
|
|
|
/* The dual of at_begline_loc_p. This one is for $. We assume there is
|
/* The dual of at_begline_loc_p. This one is for $. We assume there is
|
at least one character after the $, i.e., `P < PEND'. */
|
at least one character after the $, i.e., `P < PEND'. */
|
|
|
static boolean
|
static boolean
|
at_endline_loc_p (p, pend, syntax)
|
at_endline_loc_p (p, pend, syntax)
|
const char *p, *pend;
|
const char *p, *pend;
|
reg_syntax_t syntax;
|
reg_syntax_t syntax;
|
{
|
{
|
const char *next = p;
|
const char *next = p;
|
boolean next_backslash = *next == '\\';
|
boolean next_backslash = *next == '\\';
|
const char *next_next = p + 1 < pend ? p + 1 : 0;
|
const char *next_next = p + 1 < pend ? p + 1 : 0;
|
|
|
return
|
return
|
/* Before a subexpression? */
|
/* Before a subexpression? */
|
(syntax & RE_NO_BK_PARENS ? *next == ')'
|
(syntax & RE_NO_BK_PARENS ? *next == ')'
|
: next_backslash && next_next && *next_next == ')')
|
: next_backslash && next_next && *next_next == ')')
|
/* Before an alternative? */
|
/* Before an alternative? */
|
|| (syntax & RE_NO_BK_VBAR ? *next == '|'
|
|| (syntax & RE_NO_BK_VBAR ? *next == '|'
|
: next_backslash && next_next && *next_next == '|');
|
: next_backslash && next_next && *next_next == '|');
|
}
|
}
|
|
|
|
|
/* Returns true if REGNUM is in one of COMPILE_STACK's elements and
|
/* Returns true if REGNUM is in one of COMPILE_STACK's elements and
|
false if it's not. */
|
false if it's not. */
|
|
|
static boolean
|
static boolean
|
group_in_compile_stack (compile_stack, regnum)
|
group_in_compile_stack (compile_stack, regnum)
|
compile_stack_type compile_stack;
|
compile_stack_type compile_stack;
|
regnum_t regnum;
|
regnum_t regnum;
|
{
|
{
|
int this_element;
|
int this_element;
|
|
|
for (this_element = compile_stack.avail - 1;
|
for (this_element = compile_stack.avail - 1;
|
this_element >= 0;
|
this_element >= 0;
|
this_element--)
|
this_element--)
|
if (compile_stack.stack[this_element].regnum == regnum)
|
if (compile_stack.stack[this_element].regnum == regnum)
|
return true;
|
return true;
|
|
|
return false;
|
return false;
|
}
|
}
|
|
|
|
|
/* Read the ending character of a range (in a bracket expression) from the
|
/* Read the ending character of a range (in a bracket expression) from the
|
uncompiled pattern *P_PTR (which ends at PEND). We assume the
|
uncompiled pattern *P_PTR (which ends at PEND). We assume the
|
starting character is in `P[-2]'. (`P[-1]' is the character `-'.)
|
starting character is in `P[-2]'. (`P[-1]' is the character `-'.)
|
Then we set the translation of all bits between the starting and
|
Then we set the translation of all bits between the starting and
|
ending characters (inclusive) in the compiled pattern B.
|
ending characters (inclusive) in the compiled pattern B.
|
|
|
Return an error code.
|
Return an error code.
|
|
|
We use these short variable names so we can use the same macros as
|
We use these short variable names so we can use the same macros as
|
`regex_compile' itself. */
|
`regex_compile' itself. */
|
|
|
static reg_errcode_t
|
static reg_errcode_t
|
compile_range (p_ptr, pend, translate, syntax, b)
|
compile_range (p_ptr, pend, translate, syntax, b)
|
const char **p_ptr, *pend;
|
const char **p_ptr, *pend;
|
RE_TRANSLATE_TYPE translate;
|
RE_TRANSLATE_TYPE translate;
|
reg_syntax_t syntax;
|
reg_syntax_t syntax;
|
unsigned char *b;
|
unsigned char *b;
|
{
|
{
|
unsigned this_char;
|
unsigned this_char;
|
|
|
const char *p = *p_ptr;
|
const char *p = *p_ptr;
|
unsigned int range_start, range_end;
|
unsigned int range_start, range_end;
|
|
|
if (p == pend)
|
if (p == pend)
|
return REG_ERANGE;
|
return REG_ERANGE;
|
|
|
/* Even though the pattern is a signed `char *', we need to fetch
|
/* Even though the pattern is a signed `char *', we need to fetch
|
with unsigned char *'s; if the high bit of the pattern character
|
with unsigned char *'s; if the high bit of the pattern character
|
is set, the range endpoints will be negative if we fetch using a
|
is set, the range endpoints will be negative if we fetch using a
|
signed char *.
|
signed char *.
|
|
|
We also want to fetch the endpoints without translating them; the
|
We also want to fetch the endpoints without translating them; the
|
appropriate translation is done in the bit-setting loop below. */
|
appropriate translation is done in the bit-setting loop below. */
|
/* The SVR4 compiler on the 3B2 had trouble with unsigned const char *. */
|
/* The SVR4 compiler on the 3B2 had trouble with unsigned const char *. */
|
range_start = ((const unsigned char *) p)[-2];
|
range_start = ((const unsigned char *) p)[-2];
|
range_end = ((const unsigned char *) p)[0];
|
range_end = ((const unsigned char *) p)[0];
|
|
|
/* Have to increment the pointer into the pattern string, so the
|
/* Have to increment the pointer into the pattern string, so the
|
caller isn't still at the ending character. */
|
caller isn't still at the ending character. */
|
(*p_ptr)++;
|
(*p_ptr)++;
|
|
|
/* If the start is after the end, the range is empty. */
|
/* If the start is after the end, the range is empty. */
|
if (range_start > range_end)
|
if (range_start > range_end)
|
return syntax & RE_NO_EMPTY_RANGES ? REG_ERANGE : REG_NOERROR;
|
return syntax & RE_NO_EMPTY_RANGES ? REG_ERANGE : REG_NOERROR;
|
|
|
/* Here we see why `this_char' has to be larger than an `unsigned
|
/* Here we see why `this_char' has to be larger than an `unsigned
|
char' -- the range is inclusive, so if `range_end' == 0xff
|
char' -- the range is inclusive, so if `range_end' == 0xff
|
(assuming 8-bit characters), we would otherwise go into an infinite
|
(assuming 8-bit characters), we would otherwise go into an infinite
|
loop, since all characters <= 0xff. */
|
loop, since all characters <= 0xff. */
|
for (this_char = range_start; this_char <= range_end; this_char++)
|
for (this_char = range_start; this_char <= range_end; this_char++)
|
{
|
{
|
SET_LIST_BIT (TRANSLATE (this_char));
|
SET_LIST_BIT (TRANSLATE (this_char));
|
}
|
}
|
|
|
return REG_NOERROR;
|
return REG_NOERROR;
|
}
|
}
|
|
|
/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in
|
/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in
|
BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible
|
BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible
|
characters can start a string that matches the pattern. This fastmap
|
characters can start a string that matches the pattern. This fastmap
|
is used by re_search to skip quickly over impossible starting points.
|
is used by re_search to skip quickly over impossible starting points.
|
|
|
The caller must supply the address of a (1 << BYTEWIDTH)-byte data
|
The caller must supply the address of a (1 << BYTEWIDTH)-byte data
|
area as BUFP->fastmap.
|
area as BUFP->fastmap.
|
|
|
We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in
|
We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in
|
the pattern buffer.
|
the pattern buffer.
|
|
|
Returns 0 if we succeed, -2 if an internal error. */
|
Returns 0 if we succeed, -2 if an internal error. */
|
|
|
int
|
int
|
re_compile_fastmap (bufp)
|
re_compile_fastmap (bufp)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
{
|
{
|
int j, k;
|
int j, k;
|
#ifdef MATCH_MAY_ALLOCATE
|
#ifdef MATCH_MAY_ALLOCATE
|
fail_stack_type fail_stack;
|
fail_stack_type fail_stack;
|
#endif
|
#endif
|
#ifndef REGEX_MALLOC
|
#ifndef REGEX_MALLOC
|
char *destination;
|
char *destination;
|
#endif
|
#endif
|
|
|
register char *fastmap = bufp->fastmap;
|
register char *fastmap = bufp->fastmap;
|
unsigned char *pattern = bufp->buffer;
|
unsigned char *pattern = bufp->buffer;
|
unsigned char *p = pattern;
|
unsigned char *p = pattern;
|
register unsigned char *pend = pattern + bufp->used;
|
register unsigned char *pend = pattern + bufp->used;
|
|
|
#ifdef REL_ALLOC
|
#ifdef REL_ALLOC
|
/* This holds the pointer to the failure stack, when
|
/* This holds the pointer to the failure stack, when
|
it is allocated relocatably. */
|
it is allocated relocatably. */
|
fail_stack_elt_t *failure_stack_ptr;
|
fail_stack_elt_t *failure_stack_ptr;
|
#endif
|
#endif
|
|
|
/* Assume that each path through the pattern can be null until
|
/* Assume that each path through the pattern can be null until
|
proven otherwise. We set this false at the bottom of switch
|
proven otherwise. We set this false at the bottom of switch
|
statement, to which we get only if a particular path doesn't
|
statement, to which we get only if a particular path doesn't
|
match the empty string. */
|
match the empty string. */
|
boolean path_can_be_null = true;
|
boolean path_can_be_null = true;
|
|
|
/* We aren't doing a `succeed_n' to begin with. */
|
/* We aren't doing a `succeed_n' to begin with. */
|
boolean succeed_n_p = false;
|
boolean succeed_n_p = false;
|
|
|
assert (fastmap != NULL && p != NULL);
|
assert (fastmap != NULL && p != NULL);
|
|
|
INIT_FAIL_STACK ();
|
INIT_FAIL_STACK ();
|
bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */
|
bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */
|
bufp->fastmap_accurate = 1; /* It will be when we're done. */
|
bufp->fastmap_accurate = 1; /* It will be when we're done. */
|
bufp->can_be_null = 0;
|
bufp->can_be_null = 0;
|
|
|
while (1)
|
while (1)
|
{
|
{
|
if (p == pend || *p == succeed)
|
if (p == pend || *p == succeed)
|
{
|
{
|
/* We have reached the (effective) end of pattern. */
|
/* We have reached the (effective) end of pattern. */
|
if (!FAIL_STACK_EMPTY ())
|
if (!FAIL_STACK_EMPTY ())
|
{
|
{
|
bufp->can_be_null |= path_can_be_null;
|
bufp->can_be_null |= path_can_be_null;
|
|
|
/* Reset for next path. */
|
/* Reset for next path. */
|
path_can_be_null = true;
|
path_can_be_null = true;
|
|
|
p = fail_stack.stack[--fail_stack.avail].pointer;
|
p = fail_stack.stack[--fail_stack.avail].pointer;
|
|
|
continue;
|
continue;
|
}
|
}
|
else
|
else
|
break;
|
break;
|
}
|
}
|
|
|
/* We should never be about to go beyond the end of the pattern. */
|
/* We should never be about to go beyond the end of the pattern. */
|
assert (p < pend);
|
assert (p < pend);
|
|
|
switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
|
switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
|
{
|
{
|
|
|
/* I guess the idea here is to simply not bother with a fastmap
|
/* I guess the idea here is to simply not bother with a fastmap
|
if a backreference is used, since it's too hard to figure out
|
if a backreference is used, since it's too hard to figure out
|
the fastmap for the corresponding group. Setting
|
the fastmap for the corresponding group. Setting
|
`can_be_null' stops `re_search_2' from using the fastmap, so
|
`can_be_null' stops `re_search_2' from using the fastmap, so
|
that is all we do. */
|
that is all we do. */
|
case duplicate:
|
case duplicate:
|
bufp->can_be_null = 1;
|
bufp->can_be_null = 1;
|
goto done;
|
goto done;
|
|
|
|
|
/* Following are the cases which match a character. These end
|
/* Following are the cases which match a character. These end
|
with `break'. */
|
with `break'. */
|
|
|
case exactn:
|
case exactn:
|
fastmap[p[1]] = 1;
|
fastmap[p[1]] = 1;
|
break;
|
break;
|
|
|
|
|
case charset:
|
case charset:
|
for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
|
for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
|
if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))
|
if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
case charset_not:
|
case charset_not:
|
/* Chars beyond end of map must be allowed. */
|
/* Chars beyond end of map must be allowed. */
|
for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++)
|
for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
|
|
for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
|
for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
|
if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))))
|
if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))))
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
case wordchar:
|
case wordchar:
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
if (SYNTAX (j) == Sword)
|
if (SYNTAX (j) == Sword)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
case notwordchar:
|
case notwordchar:
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
if (SYNTAX (j) != Sword)
|
if (SYNTAX (j) != Sword)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
case anychar:
|
case anychar:
|
{
|
{
|
int fastmap_newline = fastmap['\n'];
|
int fastmap_newline = fastmap['\n'];
|
|
|
/* `.' matches anything ... */
|
/* `.' matches anything ... */
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
|
|
/* ... except perhaps newline. */
|
/* ... except perhaps newline. */
|
if (!(bufp->syntax & RE_DOT_NEWLINE))
|
if (!(bufp->syntax & RE_DOT_NEWLINE))
|
fastmap['\n'] = fastmap_newline;
|
fastmap['\n'] = fastmap_newline;
|
|
|
/* Return if we have already set `can_be_null'; if we have,
|
/* Return if we have already set `can_be_null'; if we have,
|
then the fastmap is irrelevant. Something's wrong here. */
|
then the fastmap is irrelevant. Something's wrong here. */
|
else if (bufp->can_be_null)
|
else if (bufp->can_be_null)
|
goto done;
|
goto done;
|
|
|
/* Otherwise, have to check alternative paths. */
|
/* Otherwise, have to check alternative paths. */
|
break;
|
break;
|
}
|
}
|
|
|
#ifdef emacs
|
#ifdef emacs
|
case syntaxspec:
|
case syntaxspec:
|
k = *p++;
|
k = *p++;
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
if (SYNTAX (j) == (enum syntaxcode) k)
|
if (SYNTAX (j) == (enum syntaxcode) k)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
case notsyntaxspec:
|
case notsyntaxspec:
|
k = *p++;
|
k = *p++;
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
for (j = 0; j < (1 << BYTEWIDTH); j++)
|
if (SYNTAX (j) != (enum syntaxcode) k)
|
if (SYNTAX (j) != (enum syntaxcode) k)
|
fastmap[j] = 1;
|
fastmap[j] = 1;
|
break;
|
break;
|
|
|
|
|
/* All cases after this match the empty string. These end with
|
/* All cases after this match the empty string. These end with
|
`continue'. */
|
`continue'. */
|
|
|
|
|
case before_dot:
|
case before_dot:
|
case at_dot:
|
case at_dot:
|
case after_dot:
|
case after_dot:
|
continue;
|
continue;
|
#endif /* emacs */
|
#endif /* emacs */
|
|
|
|
|
case no_op:
|
case no_op:
|
case begline:
|
case begline:
|
case endline:
|
case endline:
|
case begbuf:
|
case begbuf:
|
case endbuf:
|
case endbuf:
|
case wordbound:
|
case wordbound:
|
case notwordbound:
|
case notwordbound:
|
case wordbeg:
|
case wordbeg:
|
case wordend:
|
case wordend:
|
case push_dummy_failure:
|
case push_dummy_failure:
|
continue;
|
continue;
|
|
|
|
|
case jump_n:
|
case jump_n:
|
case pop_failure_jump:
|
case pop_failure_jump:
|
case maybe_pop_jump:
|
case maybe_pop_jump:
|
case jump:
|
case jump:
|
case jump_past_alt:
|
case jump_past_alt:
|
case dummy_failure_jump:
|
case dummy_failure_jump:
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
p += j;
|
p += j;
|
if (j > 0)
|
if (j > 0)
|
continue;
|
continue;
|
|
|
/* Jump backward implies we just went through the body of a
|
/* Jump backward implies we just went through the body of a
|
loop and matched nothing. Opcode jumped to should be
|
loop and matched nothing. Opcode jumped to should be
|
`on_failure_jump' or `succeed_n'. Just treat it like an
|
`on_failure_jump' or `succeed_n'. Just treat it like an
|
ordinary jump. For a * loop, it has pushed its failure
|
ordinary jump. For a * loop, it has pushed its failure
|
point already; if so, discard that as redundant. */
|
point already; if so, discard that as redundant. */
|
if ((re_opcode_t) *p != on_failure_jump
|
if ((re_opcode_t) *p != on_failure_jump
|
&& (re_opcode_t) *p != succeed_n)
|
&& (re_opcode_t) *p != succeed_n)
|
continue;
|
continue;
|
|
|
p++;
|
p++;
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
p += j;
|
p += j;
|
|
|
/* If what's on the stack is where we are now, pop it. */
|
/* If what's on the stack is where we are now, pop it. */
|
if (!FAIL_STACK_EMPTY ()
|
if (!FAIL_STACK_EMPTY ()
|
&& fail_stack.stack[fail_stack.avail - 1].pointer == p)
|
&& fail_stack.stack[fail_stack.avail - 1].pointer == p)
|
fail_stack.avail--;
|
fail_stack.avail--;
|
|
|
continue;
|
continue;
|
|
|
|
|
case on_failure_jump:
|
case on_failure_jump:
|
case on_failure_keep_string_jump:
|
case on_failure_keep_string_jump:
|
handle_on_failure_jump:
|
handle_on_failure_jump:
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
EXTRACT_NUMBER_AND_INCR (j, p);
|
|
|
/* For some patterns, e.g., `(a?)?', `p+j' here points to the
|
/* For some patterns, e.g., `(a?)?', `p+j' here points to the
|
end of the pattern. We don't want to push such a point,
|
end of the pattern. We don't want to push such a point,
|
since when we restore it above, entering the switch will
|
since when we restore it above, entering the switch will
|
increment `p' past the end of the pattern. We don't need
|
increment `p' past the end of the pattern. We don't need
|
to push such a point since we obviously won't find any more
|
to push such a point since we obviously won't find any more
|
fastmap entries beyond `pend'. Such a pattern can match
|
fastmap entries beyond `pend'. Such a pattern can match
|
the null string, though. */
|
the null string, though. */
|
if (p + j < pend)
|
if (p + j < pend)
|
{
|
{
|
if (!PUSH_PATTERN_OP (p + j, fail_stack))
|
if (!PUSH_PATTERN_OP (p + j, fail_stack))
|
{
|
{
|
RESET_FAIL_STACK ();
|
RESET_FAIL_STACK ();
|
return -2;
|
return -2;
|
}
|
}
|
}
|
}
|
else
|
else
|
bufp->can_be_null = 1;
|
bufp->can_be_null = 1;
|
|
|
if (succeed_n_p)
|
if (succeed_n_p)
|
{
|
{
|
EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */
|
EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */
|
succeed_n_p = false;
|
succeed_n_p = false;
|
}
|
}
|
|
|
continue;
|
continue;
|
|
|
|
|
case succeed_n:
|
case succeed_n:
|
/* Get to the number of times to succeed. */
|
/* Get to the number of times to succeed. */
|
p += 2;
|
p += 2;
|
|
|
/* Increment p past the n for when k != 0. */
|
/* Increment p past the n for when k != 0. */
|
EXTRACT_NUMBER_AND_INCR (k, p);
|
EXTRACT_NUMBER_AND_INCR (k, p);
|
if (k == 0)
|
if (k == 0)
|
{
|
{
|
p -= 4;
|
p -= 4;
|
succeed_n_p = true; /* Spaghetti code alert. */
|
succeed_n_p = true; /* Spaghetti code alert. */
|
goto handle_on_failure_jump;
|
goto handle_on_failure_jump;
|
}
|
}
|
continue;
|
continue;
|
|
|
|
|
case set_number_at:
|
case set_number_at:
|
p += 4;
|
p += 4;
|
continue;
|
continue;
|
|
|
|
|
case start_memory:
|
case start_memory:
|
case stop_memory:
|
case stop_memory:
|
p += 2;
|
p += 2;
|
continue;
|
continue;
|
|
|
|
|
default:
|
default:
|
abort (); /* We have listed all the cases. */
|
abort (); /* We have listed all the cases. */
|
} /* switch *p++ */
|
} /* switch *p++ */
|
|
|
/* Getting here means we have found the possible starting
|
/* Getting here means we have found the possible starting
|
characters for one path of the pattern -- and that the empty
|
characters for one path of the pattern -- and that the empty
|
string does not match. We need not follow this path further.
|
string does not match. We need not follow this path further.
|
Instead, look at the next alternative (remembered on the
|
Instead, look at the next alternative (remembered on the
|
stack), or quit if no more. The test at the top of the loop
|
stack), or quit if no more. The test at the top of the loop
|
does these things. */
|
does these things. */
|
path_can_be_null = false;
|
path_can_be_null = false;
|
p = pend;
|
p = pend;
|
} /* while p */
|
} /* while p */
|
|
|
/* Set `can_be_null' for the last path (also the first path, if the
|
/* Set `can_be_null' for the last path (also the first path, if the
|
pattern is empty). */
|
pattern is empty). */
|
bufp->can_be_null |= path_can_be_null;
|
bufp->can_be_null |= path_can_be_null;
|
|
|
done:
|
done:
|
RESET_FAIL_STACK ();
|
RESET_FAIL_STACK ();
|
return 0;
|
return 0;
|
} /* re_compile_fastmap */
|
} /* re_compile_fastmap */
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_compile_fastmap, re_compile_fastmap)
|
weak_alias (__re_compile_fastmap, re_compile_fastmap)
|
#endif
|
#endif
|
|
|
/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
|
/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
|
ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use
|
ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use
|
this memory for recording register information. STARTS and ENDS
|
this memory for recording register information. STARTS and ENDS
|
must be allocated using the malloc library routine, and must each
|
must be allocated using the malloc library routine, and must each
|
be at least NUM_REGS * sizeof (regoff_t) bytes long.
|
be at least NUM_REGS * sizeof (regoff_t) bytes long.
|
|
|
If NUM_REGS == 0, then subsequent matches should allocate their own
|
If NUM_REGS == 0, then subsequent matches should allocate their own
|
register data.
|
register data.
|
|
|
Unless this function is called, the first search or match using
|
Unless this function is called, the first search or match using
|
PATTERN_BUFFER will allocate its own register data, without
|
PATTERN_BUFFER will allocate its own register data, without
|
freeing the old data. */
|
freeing the old data. */
|
|
|
void
|
void
|
re_set_registers (bufp, regs, num_regs, starts, ends)
|
re_set_registers (bufp, regs, num_regs, starts, ends)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
unsigned num_regs;
|
unsigned num_regs;
|
regoff_t *starts, *ends;
|
regoff_t *starts, *ends;
|
{
|
{
|
if (num_regs)
|
if (num_regs)
|
{
|
{
|
bufp->regs_allocated = REGS_REALLOCATE;
|
bufp->regs_allocated = REGS_REALLOCATE;
|
regs->num_regs = num_regs;
|
regs->num_regs = num_regs;
|
regs->start = starts;
|
regs->start = starts;
|
regs->end = ends;
|
regs->end = ends;
|
}
|
}
|
else
|
else
|
{
|
{
|
bufp->regs_allocated = REGS_UNALLOCATED;
|
bufp->regs_allocated = REGS_UNALLOCATED;
|
regs->num_regs = 0;
|
regs->num_regs = 0;
|
regs->start = regs->end = (regoff_t *) 0;
|
regs->start = regs->end = (regoff_t *) 0;
|
}
|
}
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_set_registers, re_set_registers)
|
weak_alias (__re_set_registers, re_set_registers)
|
#endif
|
#endif
|
|
|
/* Searching routines. */
|
/* Searching routines. */
|
|
|
/* Like re_search_2, below, but only one string is specified, and
|
/* Like re_search_2, below, but only one string is specified, and
|
doesn't let you say where to stop matching. */
|
doesn't let you say where to stop matching. */
|
|
|
int
|
int
|
re_search (bufp, string, size, startpos, range, regs)
|
re_search (bufp, string, size, startpos, range, regs)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
const char *string;
|
const char *string;
|
int size, startpos, range;
|
int size, startpos, range;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
{
|
{
|
return re_search_2 (bufp, NULL, 0, string, size, startpos, range,
|
return re_search_2 (bufp, NULL, 0, string, size, startpos, range,
|
regs, size);
|
regs, size);
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_search, re_search)
|
weak_alias (__re_search, re_search)
|
#endif
|
#endif
|
|
|
|
|
/* Using the compiled pattern in BUFP->buffer, first tries to match the
|
/* Using the compiled pattern in BUFP->buffer, first tries to match the
|
virtual concatenation of STRING1 and STRING2, starting first at index
|
virtual concatenation of STRING1 and STRING2, starting first at index
|
STARTPOS, then at STARTPOS + 1, and so on.
|
STARTPOS, then at STARTPOS + 1, and so on.
|
|
|
STRING1 and STRING2 have length SIZE1 and SIZE2, respectively.
|
STRING1 and STRING2 have length SIZE1 and SIZE2, respectively.
|
|
|
RANGE is how far to scan while trying to match. RANGE = 0 means try
|
RANGE is how far to scan while trying to match. RANGE = 0 means try
|
only at STARTPOS; in general, the last start tried is STARTPOS +
|
only at STARTPOS; in general, the last start tried is STARTPOS +
|
RANGE.
|
RANGE.
|
|
|
In REGS, return the indices of the virtual concatenation of STRING1
|
In REGS, return the indices of the virtual concatenation of STRING1
|
and STRING2 that matched the entire BUFP->buffer and its contained
|
and STRING2 that matched the entire BUFP->buffer and its contained
|
subexpressions.
|
subexpressions.
|
|
|
Do not consider matching one past the index STOP in the virtual
|
Do not consider matching one past the index STOP in the virtual
|
concatenation of STRING1 and STRING2.
|
concatenation of STRING1 and STRING2.
|
|
|
We return either the position in the strings at which the match was
|
We return either the position in the strings at which the match was
|
found, -1 if no match, or -2 if error (such as failure
|
found, -1 if no match, or -2 if error (such as failure
|
stack overflow). */
|
stack overflow). */
|
|
|
int
|
int
|
re_search_2 (bufp, string1, size1, string2, size2, startpos, range, regs, stop)
|
re_search_2 (bufp, string1, size1, string2, size2, startpos, range, regs, stop)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
const char *string1, *string2;
|
const char *string1, *string2;
|
int size1, size2;
|
int size1, size2;
|
int startpos;
|
int startpos;
|
int range;
|
int range;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
int stop;
|
int stop;
|
{
|
{
|
int val;
|
int val;
|
register char *fastmap = bufp->fastmap;
|
register char *fastmap = bufp->fastmap;
|
register RE_TRANSLATE_TYPE translate = bufp->translate;
|
register RE_TRANSLATE_TYPE translate = bufp->translate;
|
int total_size = size1 + size2;
|
int total_size = size1 + size2;
|
int endpos = startpos + range;
|
int endpos = startpos + range;
|
|
|
/* Check for out-of-range STARTPOS. */
|
/* Check for out-of-range STARTPOS. */
|
if (startpos < 0 || startpos > total_size)
|
if (startpos < 0 || startpos > total_size)
|
return -1;
|
return -1;
|
|
|
/* Fix up RANGE if it might eventually take us outside
|
/* Fix up RANGE if it might eventually take us outside
|
the virtual concatenation of STRING1 and STRING2.
|
the virtual concatenation of STRING1 and STRING2.
|
Make sure we won't move STARTPOS below 0 or above TOTAL_SIZE. */
|
Make sure we won't move STARTPOS below 0 or above TOTAL_SIZE. */
|
if (endpos < 0)
|
if (endpos < 0)
|
range = 0 - startpos;
|
range = 0 - startpos;
|
else if (endpos > total_size)
|
else if (endpos > total_size)
|
range = total_size - startpos;
|
range = total_size - startpos;
|
|
|
/* If the search isn't to be a backwards one, don't waste time in a
|
/* If the search isn't to be a backwards one, don't waste time in a
|
search for a pattern that must be anchored. */
|
search for a pattern that must be anchored. */
|
if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == begbuf && range > 0)
|
if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == begbuf && range > 0)
|
{
|
{
|
if (startpos > 0)
|
if (startpos > 0)
|
return -1;
|
return -1;
|
else
|
else
|
range = 1;
|
range = 1;
|
}
|
}
|
|
|
#ifdef emacs
|
#ifdef emacs
|
/* In a forward search for something that starts with \=.
|
/* In a forward search for something that starts with \=.
|
don't keep searching past point. */
|
don't keep searching past point. */
|
if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == at_dot && range > 0)
|
if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == at_dot && range > 0)
|
{
|
{
|
range = PT - startpos;
|
range = PT - startpos;
|
if (range <= 0)
|
if (range <= 0)
|
return -1;
|
return -1;
|
}
|
}
|
#endif /* emacs */
|
#endif /* emacs */
|
|
|
/* Update the fastmap now if not correct already. */
|
/* Update the fastmap now if not correct already. */
|
if (fastmap && !bufp->fastmap_accurate)
|
if (fastmap && !bufp->fastmap_accurate)
|
if (re_compile_fastmap (bufp) == -2)
|
if (re_compile_fastmap (bufp) == -2)
|
return -2;
|
return -2;
|
|
|
/* Loop through the string, looking for a place to start matching. */
|
/* Loop through the string, looking for a place to start matching. */
|
for (;;)
|
for (;;)
|
{
|
{
|
/* If a fastmap is supplied, skip quickly over characters that
|
/* If a fastmap is supplied, skip quickly over characters that
|
cannot be the start of a match. If the pattern can match the
|
cannot be the start of a match. If the pattern can match the
|
null string, however, we don't need to skip characters; we want
|
null string, however, we don't need to skip characters; we want
|
the first null string. */
|
the first null string. */
|
if (fastmap && startpos < total_size && !bufp->can_be_null)
|
if (fastmap && startpos < total_size && !bufp->can_be_null)
|
{
|
{
|
if (range > 0) /* Searching forwards. */
|
if (range > 0) /* Searching forwards. */
|
{
|
{
|
register const char *d;
|
register const char *d;
|
register int lim = 0;
|
register int lim = 0;
|
int irange = range;
|
int irange = range;
|
|
|
if (startpos < size1 && startpos + range >= size1)
|
if (startpos < size1 && startpos + range >= size1)
|
lim = range - (size1 - startpos);
|
lim = range - (size1 - startpos);
|
|
|
d = (startpos >= size1 ? string2 - size1 : string1) + startpos;
|
d = (startpos >= size1 ? string2 - size1 : string1) + startpos;
|
|
|
/* Written out as an if-else to avoid testing `translate'
|
/* Written out as an if-else to avoid testing `translate'
|
inside the loop. */
|
inside the loop. */
|
if (translate)
|
if (translate)
|
while (range > lim
|
while (range > lim
|
&& !fastmap[(unsigned char)
|
&& !fastmap[(unsigned char)
|
translate[(unsigned char) *d++]])
|
translate[(unsigned char) *d++]])
|
range--;
|
range--;
|
else
|
else
|
while (range > lim && !fastmap[(unsigned char) *d++])
|
while (range > lim && !fastmap[(unsigned char) *d++])
|
range--;
|
range--;
|
|
|
startpos += irange - range;
|
startpos += irange - range;
|
}
|
}
|
else /* Searching backwards. */
|
else /* Searching backwards. */
|
{
|
{
|
register char c = (size1 == 0 || startpos >= size1
|
register char c = (size1 == 0 || startpos >= size1
|
? string2[startpos - size1]
|
? string2[startpos - size1]
|
: string1[startpos]);
|
: string1[startpos]);
|
|
|
if (!fastmap[(unsigned char) TRANSLATE (c)])
|
if (!fastmap[(unsigned char) TRANSLATE (c)])
|
goto advance;
|
goto advance;
|
}
|
}
|
}
|
}
|
|
|
/* If can't match the null string, and that's all we have left, fail. */
|
/* If can't match the null string, and that's all we have left, fail. */
|
if (range >= 0 && startpos == total_size && fastmap
|
if (range >= 0 && startpos == total_size && fastmap
|
&& !bufp->can_be_null)
|
&& !bufp->can_be_null)
|
return -1;
|
return -1;
|
|
|
val = re_match_2_internal (bufp, string1, size1, string2, size2,
|
val = re_match_2_internal (bufp, string1, size1, string2, size2,
|
startpos, regs, stop);
|
startpos, regs, stop);
|
#ifndef REGEX_MALLOC
|
#ifndef REGEX_MALLOC
|
# ifdef C_ALLOCA
|
# ifdef C_ALLOCA
|
alloca (0);
|
alloca (0);
|
# endif
|
# endif
|
#endif
|
#endif
|
|
|
if (val >= 0)
|
if (val >= 0)
|
return startpos;
|
return startpos;
|
|
|
if (val == -2)
|
if (val == -2)
|
return -2;
|
return -2;
|
|
|
advance:
|
advance:
|
if (!range)
|
if (!range)
|
break;
|
break;
|
else if (range > 0)
|
else if (range > 0)
|
{
|
{
|
range--;
|
range--;
|
startpos++;
|
startpos++;
|
}
|
}
|
else
|
else
|
{
|
{
|
range++;
|
range++;
|
startpos--;
|
startpos--;
|
}
|
}
|
}
|
}
|
return -1;
|
return -1;
|
} /* re_search_2 */
|
} /* re_search_2 */
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_search_2, re_search_2)
|
weak_alias (__re_search_2, re_search_2)
|
#endif
|
#endif
|
|
|
/* This converts PTR, a pointer into one of the search strings `string1'
|
/* This converts PTR, a pointer into one of the search strings `string1'
|
and `string2' into an offset from the beginning of that string. */
|
and `string2' into an offset from the beginning of that string. */
|
#define POINTER_TO_OFFSET(ptr) \
|
#define POINTER_TO_OFFSET(ptr) \
|
(FIRST_STRING_P (ptr) \
|
(FIRST_STRING_P (ptr) \
|
? ((regoff_t) ((ptr) - string1)) \
|
? ((regoff_t) ((ptr) - string1)) \
|
: ((regoff_t) ((ptr) - string2 + size1)))
|
: ((regoff_t) ((ptr) - string2 + size1)))
|
|
|
/* Macros for dealing with the split strings in re_match_2. */
|
/* Macros for dealing with the split strings in re_match_2. */
|
|
|
#define MATCHING_IN_FIRST_STRING (dend == end_match_1)
|
#define MATCHING_IN_FIRST_STRING (dend == end_match_1)
|
|
|
/* Call before fetching a character with *d. This switches over to
|
/* Call before fetching a character with *d. This switches over to
|
string2 if necessary. */
|
string2 if necessary. */
|
#define PREFETCH() \
|
#define PREFETCH() \
|
while (d == dend) \
|
while (d == dend) \
|
{ \
|
{ \
|
/* End of string2 => fail. */ \
|
/* End of string2 => fail. */ \
|
if (dend == end_match_2) \
|
if (dend == end_match_2) \
|
goto fail; \
|
goto fail; \
|
/* End of string1 => advance to string2. */ \
|
/* End of string1 => advance to string2. */ \
|
d = string2; \
|
d = string2; \
|
dend = end_match_2; \
|
dend = end_match_2; \
|
}
|
}
|
|
|
|
|
/* Test if at very beginning or at very end of the virtual concatenation
|
/* Test if at very beginning or at very end of the virtual concatenation
|
of `string1' and `string2'. If only one string, it's `string2'. */
|
of `string1' and `string2'. If only one string, it's `string2'. */
|
#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2)
|
#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2)
|
#define AT_STRINGS_END(d) ((d) == end2)
|
#define AT_STRINGS_END(d) ((d) == end2)
|
|
|
|
|
/* Test if D points to a character which is word-constituent. We have
|
/* Test if D points to a character which is word-constituent. We have
|
two special cases to check for: if past the end of string1, look at
|
two special cases to check for: if past the end of string1, look at
|
the first character in string2; and if before the beginning of
|
the first character in string2; and if before the beginning of
|
string2, look at the last character in string1. */
|
string2, look at the last character in string1. */
|
#define WORDCHAR_P(d) \
|
#define WORDCHAR_P(d) \
|
(SYNTAX ((d) == end1 ? *string2 \
|
(SYNTAX ((d) == end1 ? *string2 \
|
: (d) == string2 - 1 ? *(end1 - 1) : *(d)) \
|
: (d) == string2 - 1 ? *(end1 - 1) : *(d)) \
|
== Sword)
|
== Sword)
|
|
|
/* Disabled due to a compiler bug -- see comment at case wordbound */
|
/* Disabled due to a compiler bug -- see comment at case wordbound */
|
#if 0
|
#if 0
|
/* Test if the character before D and the one at D differ with respect
|
/* Test if the character before D and the one at D differ with respect
|
to being word-constituent. */
|
to being word-constituent. */
|
#define AT_WORD_BOUNDARY(d) \
|
#define AT_WORD_BOUNDARY(d) \
|
(AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \
|
(AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \
|
|| WORDCHAR_P (d - 1) != WORDCHAR_P (d))
|
|| WORDCHAR_P (d - 1) != WORDCHAR_P (d))
|
#endif
|
#endif
|
|
|
/* Free everything we malloc. */
|
/* Free everything we malloc. */
|
#ifdef MATCH_MAY_ALLOCATE
|
#ifdef MATCH_MAY_ALLOCATE
|
# define FREE_VAR(var) if (var) REGEX_FREE (var); var = NULL
|
# define FREE_VAR(var) if (var) REGEX_FREE (var); var = NULL
|
# define FREE_VARIABLES() \
|
# define FREE_VARIABLES() \
|
do { \
|
do { \
|
REGEX_FREE_STACK (fail_stack.stack); \
|
REGEX_FREE_STACK (fail_stack.stack); \
|
FREE_VAR (regstart); \
|
FREE_VAR (regstart); \
|
FREE_VAR (regend); \
|
FREE_VAR (regend); \
|
FREE_VAR (old_regstart); \
|
FREE_VAR (old_regstart); \
|
FREE_VAR (old_regend); \
|
FREE_VAR (old_regend); \
|
FREE_VAR (best_regstart); \
|
FREE_VAR (best_regstart); \
|
FREE_VAR (best_regend); \
|
FREE_VAR (best_regend); \
|
FREE_VAR (reg_info); \
|
FREE_VAR (reg_info); \
|
FREE_VAR (reg_dummy); \
|
FREE_VAR (reg_dummy); \
|
FREE_VAR (reg_info_dummy); \
|
FREE_VAR (reg_info_dummy); \
|
} while (0)
|
} while (0)
|
#else
|
#else
|
# define FREE_VARIABLES() ((void)0) /* Do nothing! But inhibit gcc warning. */
|
# define FREE_VARIABLES() ((void)0) /* Do nothing! But inhibit gcc warning. */
|
#endif /* not MATCH_MAY_ALLOCATE */
|
#endif /* not MATCH_MAY_ALLOCATE */
|
|
|
/* These values must meet several constraints. They must not be valid
|
/* These values must meet several constraints. They must not be valid
|
register values; since we have a limit of 255 registers (because
|
register values; since we have a limit of 255 registers (because
|
we use only one byte in the pattern for the register number), we can
|
we use only one byte in the pattern for the register number), we can
|
use numbers larger than 255. They must differ by 1, because of
|
use numbers larger than 255. They must differ by 1, because of
|
NUM_FAILURE_ITEMS above. And the value for the lowest register must
|
NUM_FAILURE_ITEMS above. And the value for the lowest register must
|
be larger than the value for the highest register, so we do not try
|
be larger than the value for the highest register, so we do not try
|
to actually save any registers when none are active. */
|
to actually save any registers when none are active. */
|
#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH)
|
#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH)
|
#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1)
|
#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1)
|
|
|
/* Matching routines. */
|
/* Matching routines. */
|
|
|
#ifndef emacs /* Emacs never uses this. */
|
#ifndef emacs /* Emacs never uses this. */
|
/* re_match is like re_match_2 except it takes only a single string. */
|
/* re_match is like re_match_2 except it takes only a single string. */
|
|
|
int
|
int
|
re_match (bufp, string, size, pos, regs)
|
re_match (bufp, string, size, pos, regs)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
const char *string;
|
const char *string;
|
int size, pos;
|
int size, pos;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
{
|
{
|
int result = re_match_2_internal (bufp, NULL, 0, string, size,
|
int result = re_match_2_internal (bufp, NULL, 0, string, size,
|
pos, regs, size);
|
pos, regs, size);
|
# ifndef REGEX_MALLOC
|
# ifndef REGEX_MALLOC
|
# ifdef C_ALLOCA
|
# ifdef C_ALLOCA
|
alloca (0);
|
alloca (0);
|
# endif
|
# endif
|
# endif
|
# endif
|
return result;
|
return result;
|
}
|
}
|
# ifdef _LIBC
|
# ifdef _LIBC
|
weak_alias (__re_match, re_match)
|
weak_alias (__re_match, re_match)
|
# endif
|
# endif
|
#endif /* not emacs */
|
#endif /* not emacs */
|
|
|
static boolean group_match_null_string_p _RE_ARGS ((unsigned char **p,
|
static boolean group_match_null_string_p _RE_ARGS ((unsigned char **p,
|
unsigned char *end,
|
unsigned char *end,
|
register_info_type *reg_info));
|
register_info_type *reg_info));
|
static boolean alt_match_null_string_p _RE_ARGS ((unsigned char *p,
|
static boolean alt_match_null_string_p _RE_ARGS ((unsigned char *p,
|
unsigned char *end,
|
unsigned char *end,
|
register_info_type *reg_info));
|
register_info_type *reg_info));
|
static boolean common_op_match_null_string_p _RE_ARGS ((unsigned char **p,
|
static boolean common_op_match_null_string_p _RE_ARGS ((unsigned char **p,
|
unsigned char *end,
|
unsigned char *end,
|
register_info_type *reg_info));
|
register_info_type *reg_info));
|
static int bcmp_translate _RE_ARGS ((const char *s1, const char *s2,
|
static int bcmp_translate _RE_ARGS ((const char *s1, const char *s2,
|
int len, char *translate));
|
int len, char *translate));
|
|
|
/* re_match_2 matches the compiled pattern in BUFP against the
|
/* re_match_2 matches the compiled pattern in BUFP against the
|
the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1
|
the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1
|
and SIZE2, respectively). We start matching at POS, and stop
|
and SIZE2, respectively). We start matching at POS, and stop
|
matching at STOP.
|
matching at STOP.
|
|
|
If REGS is non-null and the `no_sub' field of BUFP is nonzero, we
|
If REGS is non-null and the `no_sub' field of BUFP is nonzero, we
|
store offsets for the substring each group matched in REGS. See the
|
store offsets for the substring each group matched in REGS. See the
|
documentation for exactly how many groups we fill.
|
documentation for exactly how many groups we fill.
|
|
|
We return -1 if no match, -2 if an internal error (such as the
|
We return -1 if no match, -2 if an internal error (such as the
|
failure stack overflowing). Otherwise, we return the length of the
|
failure stack overflowing). Otherwise, we return the length of the
|
matched substring. */
|
matched substring. */
|
|
|
int
|
int
|
re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop)
|
re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
const char *string1, *string2;
|
const char *string1, *string2;
|
int size1, size2;
|
int size1, size2;
|
int pos;
|
int pos;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
int stop;
|
int stop;
|
{
|
{
|
int result = re_match_2_internal (bufp, string1, size1, string2, size2,
|
int result = re_match_2_internal (bufp, string1, size1, string2, size2,
|
pos, regs, stop);
|
pos, regs, stop);
|
#ifndef REGEX_MALLOC
|
#ifndef REGEX_MALLOC
|
# ifdef C_ALLOCA
|
# ifdef C_ALLOCA
|
alloca (0);
|
alloca (0);
|
# endif
|
# endif
|
#endif
|
#endif
|
return result;
|
return result;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_match_2, re_match_2)
|
weak_alias (__re_match_2, re_match_2)
|
#endif
|
#endif
|
|
|
/* This is a separate function so that we can force an alloca cleanup
|
/* This is a separate function so that we can force an alloca cleanup
|
afterwards. */
|
afterwards. */
|
static int
|
static int
|
re_match_2_internal (bufp, string1, size1, string2, size2, pos, regs, stop)
|
re_match_2_internal (bufp, string1, size1, string2, size2, pos, regs, stop)
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
const char *string1, *string2;
|
const char *string1, *string2;
|
int size1, size2;
|
int size1, size2;
|
int pos;
|
int pos;
|
struct re_registers *regs;
|
struct re_registers *regs;
|
int stop;
|
int stop;
|
{
|
{
|
/* General temporaries. */
|
/* General temporaries. */
|
int mcnt;
|
int mcnt;
|
unsigned char *p1;
|
unsigned char *p1;
|
|
|
/* Just past the end of the corresponding string. */
|
/* Just past the end of the corresponding string. */
|
const char *end1, *end2;
|
const char *end1, *end2;
|
|
|
/* Pointers into string1 and string2, just past the last characters in
|
/* Pointers into string1 and string2, just past the last characters in
|
each to consider matching. */
|
each to consider matching. */
|
const char *end_match_1, *end_match_2;
|
const char *end_match_1, *end_match_2;
|
|
|
/* Where we are in the data, and the end of the current string. */
|
/* Where we are in the data, and the end of the current string. */
|
const char *d, *dend;
|
const char *d, *dend;
|
|
|
/* Where we are in the pattern, and the end of the pattern. */
|
/* Where we are in the pattern, and the end of the pattern. */
|
unsigned char *p = bufp->buffer;
|
unsigned char *p = bufp->buffer;
|
register unsigned char *pend = p + bufp->used;
|
register unsigned char *pend = p + bufp->used;
|
|
|
/* Mark the opcode just after a start_memory, so we can test for an
|
/* Mark the opcode just after a start_memory, so we can test for an
|
empty subpattern when we get to the stop_memory. */
|
empty subpattern when we get to the stop_memory. */
|
unsigned char *just_past_start_mem = 0;
|
unsigned char *just_past_start_mem = 0;
|
|
|
/* We use this to map every character in the string. */
|
/* We use this to map every character in the string. */
|
RE_TRANSLATE_TYPE translate = bufp->translate;
|
RE_TRANSLATE_TYPE translate = bufp->translate;
|
|
|
/* Failure point stack. Each place that can handle a failure further
|
/* Failure point stack. Each place that can handle a failure further
|
down the line pushes a failure point on this stack. It consists of
|
down the line pushes a failure point on this stack. It consists of
|
restart, regend, and reg_info for all registers corresponding to
|
restart, regend, and reg_info for all registers corresponding to
|
the subexpressions we're currently inside, plus the number of such
|
the subexpressions we're currently inside, plus the number of such
|
registers, and, finally, two char *'s. The first char * is where
|
registers, and, finally, two char *'s. The first char * is where
|
to resume scanning the pattern; the second one is where to resume
|
to resume scanning the pattern; the second one is where to resume
|
scanning the strings. If the latter is zero, the failure point is
|
scanning the strings. If the latter is zero, the failure point is
|
a ``dummy''; if a failure happens and the failure point is a dummy,
|
a ``dummy''; if a failure happens and the failure point is a dummy,
|
it gets discarded and the next next one is tried. */
|
it gets discarded and the next next one is tried. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
|
fail_stack_type fail_stack;
|
fail_stack_type fail_stack;
|
#endif
|
#endif
|
#ifdef DEBUG
|
#ifdef DEBUG
|
static unsigned failure_id = 0;
|
static unsigned failure_id = 0;
|
unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0;
|
unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0;
|
#endif
|
#endif
|
|
|
#ifdef REL_ALLOC
|
#ifdef REL_ALLOC
|
/* This holds the pointer to the failure stack, when
|
/* This holds the pointer to the failure stack, when
|
it is allocated relocatably. */
|
it is allocated relocatably. */
|
fail_stack_elt_t *failure_stack_ptr;
|
fail_stack_elt_t *failure_stack_ptr;
|
#endif
|
#endif
|
|
|
/* We fill all the registers internally, independent of what we
|
/* We fill all the registers internally, independent of what we
|
return, for use in backreferences. The number here includes
|
return, for use in backreferences. The number here includes
|
an element for register zero. */
|
an element for register zero. */
|
size_t num_regs = bufp->re_nsub + 1;
|
size_t num_regs = bufp->re_nsub + 1;
|
|
|
/* The currently active registers. */
|
/* The currently active registers. */
|
active_reg_t lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
active_reg_t lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
active_reg_t highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
active_reg_t highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
|
|
/* Information on the contents of registers. These are pointers into
|
/* Information on the contents of registers. These are pointers into
|
the input strings; they record just what was matched (on this
|
the input strings; they record just what was matched (on this
|
attempt) by a subexpression part of the pattern, that is, the
|
attempt) by a subexpression part of the pattern, that is, the
|
regnum-th regstart pointer points to where in the pattern we began
|
regnum-th regstart pointer points to where in the pattern we began
|
matching and the regnum-th regend points to right after where we
|
matching and the regnum-th regend points to right after where we
|
stopped matching the regnum-th subexpression. (The zeroth register
|
stopped matching the regnum-th subexpression. (The zeroth register
|
keeps track of what the whole pattern matches.) */
|
keeps track of what the whole pattern matches.) */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
const char **regstart, **regend;
|
const char **regstart, **regend;
|
#endif
|
#endif
|
|
|
/* If a group that's operated upon by a repetition operator fails to
|
/* If a group that's operated upon by a repetition operator fails to
|
match anything, then the register for its start will need to be
|
match anything, then the register for its start will need to be
|
restored because it will have been set to wherever in the string we
|
restored because it will have been set to wherever in the string we
|
are when we last see its open-group operator. Similarly for a
|
are when we last see its open-group operator. Similarly for a
|
register's end. */
|
register's end. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
const char **old_regstart, **old_regend;
|
const char **old_regstart, **old_regend;
|
#endif
|
#endif
|
|
|
/* The is_active field of reg_info helps us keep track of which (possibly
|
/* The is_active field of reg_info helps us keep track of which (possibly
|
nested) subexpressions we are currently in. The matched_something
|
nested) subexpressions we are currently in. The matched_something
|
field of reg_info[reg_num] helps us tell whether or not we have
|
field of reg_info[reg_num] helps us tell whether or not we have
|
matched any of the pattern so far this time through the reg_num-th
|
matched any of the pattern so far this time through the reg_num-th
|
subexpression. These two fields get reset each time through any
|
subexpression. These two fields get reset each time through any
|
loop their register is in. */
|
loop their register is in. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
|
register_info_type *reg_info;
|
register_info_type *reg_info;
|
#endif
|
#endif
|
|
|
/* The following record the register info as found in the above
|
/* The following record the register info as found in the above
|
variables when we find a match better than any we've seen before.
|
variables when we find a match better than any we've seen before.
|
This happens as we backtrack through the failure points, which in
|
This happens as we backtrack through the failure points, which in
|
turn happens only if we have not yet matched the entire string. */
|
turn happens only if we have not yet matched the entire string. */
|
unsigned best_regs_set = false;
|
unsigned best_regs_set = false;
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
const char **best_regstart, **best_regend;
|
const char **best_regstart, **best_regend;
|
#endif
|
#endif
|
|
|
/* Logically, this is `best_regend[0]'. But we don't want to have to
|
/* Logically, this is `best_regend[0]'. But we don't want to have to
|
allocate space for that if we're not allocating space for anything
|
allocate space for that if we're not allocating space for anything
|
else (see below). Also, we never need info about register 0 for
|
else (see below). Also, we never need info about register 0 for
|
any of the other register vectors, and it seems rather a kludge to
|
any of the other register vectors, and it seems rather a kludge to
|
treat `best_regend' differently than the rest. So we keep track of
|
treat `best_regend' differently than the rest. So we keep track of
|
the end of the best match so far in a separate variable. We
|
the end of the best match so far in a separate variable. We
|
initialize this to NULL so that when we backtrack the first time
|
initialize this to NULL so that when we backtrack the first time
|
and need to test it, it's not garbage. */
|
and need to test it, it's not garbage. */
|
const char *match_end = NULL;
|
const char *match_end = NULL;
|
|
|
/* This helps SET_REGS_MATCHED avoid doing redundant work. */
|
/* This helps SET_REGS_MATCHED avoid doing redundant work. */
|
int set_regs_matched_done = 0;
|
int set_regs_matched_done = 0;
|
|
|
/* Used when we pop values we don't care about. */
|
/* Used when we pop values we don't care about. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
|
const char **reg_dummy;
|
const char **reg_dummy;
|
register_info_type *reg_info_dummy;
|
register_info_type *reg_info_dummy;
|
#endif
|
#endif
|
|
|
#ifdef DEBUG
|
#ifdef DEBUG
|
/* Counts the total number of registers pushed. */
|
/* Counts the total number of registers pushed. */
|
unsigned num_regs_pushed = 0;
|
unsigned num_regs_pushed = 0;
|
#endif
|
#endif
|
|
|
DEBUG_PRINT1 ("\n\nEntering re_match_2.\n");
|
DEBUG_PRINT1 ("\n\nEntering re_match_2.\n");
|
|
|
INIT_FAIL_STACK ();
|
INIT_FAIL_STACK ();
|
|
|
#ifdef MATCH_MAY_ALLOCATE
|
#ifdef MATCH_MAY_ALLOCATE
|
/* Do not bother to initialize all the register variables if there are
|
/* Do not bother to initialize all the register variables if there are
|
no groups in the pattern, as it takes a fair amount of time. If
|
no groups in the pattern, as it takes a fair amount of time. If
|
there are groups, we include space for register 0 (the whole
|
there are groups, we include space for register 0 (the whole
|
pattern), even though we never use it, since it simplifies the
|
pattern), even though we never use it, since it simplifies the
|
array indexing. We should fix this. */
|
array indexing. We should fix this. */
|
if (bufp->re_nsub)
|
if (bufp->re_nsub)
|
{
|
{
|
regstart = REGEX_TALLOC (num_regs, const char *);
|
regstart = REGEX_TALLOC (num_regs, const char *);
|
regend = REGEX_TALLOC (num_regs, const char *);
|
regend = REGEX_TALLOC (num_regs, const char *);
|
old_regstart = REGEX_TALLOC (num_regs, const char *);
|
old_regstart = REGEX_TALLOC (num_regs, const char *);
|
old_regend = REGEX_TALLOC (num_regs, const char *);
|
old_regend = REGEX_TALLOC (num_regs, const char *);
|
best_regstart = REGEX_TALLOC (num_regs, const char *);
|
best_regstart = REGEX_TALLOC (num_regs, const char *);
|
best_regend = REGEX_TALLOC (num_regs, const char *);
|
best_regend = REGEX_TALLOC (num_regs, const char *);
|
reg_info = REGEX_TALLOC (num_regs, register_info_type);
|
reg_info = REGEX_TALLOC (num_regs, register_info_type);
|
reg_dummy = REGEX_TALLOC (num_regs, const char *);
|
reg_dummy = REGEX_TALLOC (num_regs, const char *);
|
reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type);
|
reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type);
|
|
|
if (!(regstart && regend && old_regstart && old_regend && reg_info
|
if (!(regstart && regend && old_regstart && old_regend && reg_info
|
&& best_regstart && best_regend && reg_dummy && reg_info_dummy))
|
&& best_regstart && best_regend && reg_dummy && reg_info_dummy))
|
{
|
{
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
return -2;
|
return -2;
|
}
|
}
|
}
|
}
|
else
|
else
|
{
|
{
|
/* We must initialize all our variables to NULL, so that
|
/* We must initialize all our variables to NULL, so that
|
`FREE_VARIABLES' doesn't try to free them. */
|
`FREE_VARIABLES' doesn't try to free them. */
|
regstart = regend = old_regstart = old_regend = best_regstart
|
regstart = regend = old_regstart = old_regend = best_regstart
|
= best_regend = reg_dummy = NULL;
|
= best_regend = reg_dummy = NULL;
|
reg_info = reg_info_dummy = (register_info_type *) NULL;
|
reg_info = reg_info_dummy = (register_info_type *) NULL;
|
}
|
}
|
#endif /* MATCH_MAY_ALLOCATE */
|
#endif /* MATCH_MAY_ALLOCATE */
|
|
|
/* The starting position is bogus. */
|
/* The starting position is bogus. */
|
if (pos < 0 || pos > size1 + size2)
|
if (pos < 0 || pos > size1 + size2)
|
{
|
{
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
return -1;
|
return -1;
|
}
|
}
|
|
|
/* Initialize subexpression text positions to -1 to mark ones that no
|
/* Initialize subexpression text positions to -1 to mark ones that no
|
start_memory/stop_memory has been seen for. Also initialize the
|
start_memory/stop_memory has been seen for. Also initialize the
|
register information struct. */
|
register information struct. */
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
{
|
{
|
regstart[mcnt] = regend[mcnt]
|
regstart[mcnt] = regend[mcnt]
|
= old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE;
|
= old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE;
|
|
|
REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE;
|
REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE;
|
IS_ACTIVE (reg_info[mcnt]) = 0;
|
IS_ACTIVE (reg_info[mcnt]) = 0;
|
MATCHED_SOMETHING (reg_info[mcnt]) = 0;
|
MATCHED_SOMETHING (reg_info[mcnt]) = 0;
|
EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0;
|
EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0;
|
}
|
}
|
|
|
/* We move `string1' into `string2' if the latter's empty -- but not if
|
/* We move `string1' into `string2' if the latter's empty -- but not if
|
`string1' is null. */
|
`string1' is null. */
|
if (size2 == 0 && string1 != NULL)
|
if (size2 == 0 && string1 != NULL)
|
{
|
{
|
string2 = string1;
|
string2 = string1;
|
size2 = size1;
|
size2 = size1;
|
string1 = 0;
|
string1 = 0;
|
size1 = 0;
|
size1 = 0;
|
}
|
}
|
end1 = string1 + size1;
|
end1 = string1 + size1;
|
end2 = string2 + size2;
|
end2 = string2 + size2;
|
|
|
/* Compute where to stop matching, within the two strings. */
|
/* Compute where to stop matching, within the two strings. */
|
if (stop <= size1)
|
if (stop <= size1)
|
{
|
{
|
end_match_1 = string1 + stop;
|
end_match_1 = string1 + stop;
|
end_match_2 = string2;
|
end_match_2 = string2;
|
}
|
}
|
else
|
else
|
{
|
{
|
end_match_1 = end1;
|
end_match_1 = end1;
|
end_match_2 = string2 + stop - size1;
|
end_match_2 = string2 + stop - size1;
|
}
|
}
|
|
|
/* `p' scans through the pattern as `d' scans through the data.
|
/* `p' scans through the pattern as `d' scans through the data.
|
`dend' is the end of the input string that `d' points within. `d'
|
`dend' is the end of the input string that `d' points within. `d'
|
is advanced into the following input string whenever necessary, but
|
is advanced into the following input string whenever necessary, but
|
this happens before fetching; therefore, at the beginning of the
|
this happens before fetching; therefore, at the beginning of the
|
loop, `d' can be pointing at the end of a string, but it cannot
|
loop, `d' can be pointing at the end of a string, but it cannot
|
equal `string2'. */
|
equal `string2'. */
|
if (size1 > 0 && pos <= size1)
|
if (size1 > 0 && pos <= size1)
|
{
|
{
|
d = string1 + pos;
|
d = string1 + pos;
|
dend = end_match_1;
|
dend = end_match_1;
|
}
|
}
|
else
|
else
|
{
|
{
|
d = string2 + pos - size1;
|
d = string2 + pos - size1;
|
dend = end_match_2;
|
dend = end_match_2;
|
}
|
}
|
|
|
DEBUG_PRINT1 ("The compiled pattern is:\n");
|
DEBUG_PRINT1 ("The compiled pattern is:\n");
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend);
|
DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend);
|
DEBUG_PRINT1 ("The string to match is: `");
|
DEBUG_PRINT1 ("The string to match is: `");
|
DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2);
|
DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2);
|
DEBUG_PRINT1 ("'\n");
|
DEBUG_PRINT1 ("'\n");
|
|
|
/* This loops over pattern commands. It exits by returning from the
|
/* This loops over pattern commands. It exits by returning from the
|
function if the match is complete, or it drops through if the match
|
function if the match is complete, or it drops through if the match
|
fails at this starting point in the input data. */
|
fails at this starting point in the input data. */
|
for (;;)
|
for (;;)
|
{
|
{
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT2 ("\n%p: ", p);
|
DEBUG_PRINT2 ("\n%p: ", p);
|
#else
|
#else
|
DEBUG_PRINT2 ("\n0x%x: ", p);
|
DEBUG_PRINT2 ("\n0x%x: ", p);
|
#endif
|
#endif
|
|
|
if (p == pend)
|
if (p == pend)
|
{ /* End of pattern means we might have succeeded. */
|
{ /* End of pattern means we might have succeeded. */
|
DEBUG_PRINT1 ("end of pattern ... ");
|
DEBUG_PRINT1 ("end of pattern ... ");
|
|
|
/* If we haven't matched the entire string, and we want the
|
/* If we haven't matched the entire string, and we want the
|
longest match, try backtracking. */
|
longest match, try backtracking. */
|
if (d != end_match_2)
|
if (d != end_match_2)
|
{
|
{
|
/* 1 if this match ends in the same string (string1 or string2)
|
/* 1 if this match ends in the same string (string1 or string2)
|
as the best previous match. */
|
as the best previous match. */
|
boolean same_str_p = (FIRST_STRING_P (match_end)
|
boolean same_str_p = (FIRST_STRING_P (match_end)
|
== MATCHING_IN_FIRST_STRING);
|
== MATCHING_IN_FIRST_STRING);
|
/* 1 if this match is the best seen so far. */
|
/* 1 if this match is the best seen so far. */
|
boolean best_match_p;
|
boolean best_match_p;
|
|
|
/* AIX compiler got confused when this was combined
|
/* AIX compiler got confused when this was combined
|
with the previous declaration. */
|
with the previous declaration. */
|
if (same_str_p)
|
if (same_str_p)
|
best_match_p = d > match_end;
|
best_match_p = d > match_end;
|
else
|
else
|
best_match_p = !MATCHING_IN_FIRST_STRING;
|
best_match_p = !MATCHING_IN_FIRST_STRING;
|
|
|
DEBUG_PRINT1 ("backtracking.\n");
|
DEBUG_PRINT1 ("backtracking.\n");
|
|
|
if (!FAIL_STACK_EMPTY ())
|
if (!FAIL_STACK_EMPTY ())
|
{ /* More failure points to try. */
|
{ /* More failure points to try. */
|
|
|
/* If exceeds best match so far, save it. */
|
/* If exceeds best match so far, save it. */
|
if (!best_regs_set || best_match_p)
|
if (!best_regs_set || best_match_p)
|
{
|
{
|
best_regs_set = true;
|
best_regs_set = true;
|
match_end = d;
|
match_end = d;
|
|
|
DEBUG_PRINT1 ("\nSAVING match as best so far.\n");
|
DEBUG_PRINT1 ("\nSAVING match as best so far.\n");
|
|
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
{
|
{
|
best_regstart[mcnt] = regstart[mcnt];
|
best_regstart[mcnt] = regstart[mcnt];
|
best_regend[mcnt] = regend[mcnt];
|
best_regend[mcnt] = regend[mcnt];
|
}
|
}
|
}
|
}
|
goto fail;
|
goto fail;
|
}
|
}
|
|
|
/* If no failure points, don't restore garbage. And if
|
/* If no failure points, don't restore garbage. And if
|
last match is real best match, don't restore second
|
last match is real best match, don't restore second
|
best one. */
|
best one. */
|
else if (best_regs_set && !best_match_p)
|
else if (best_regs_set && !best_match_p)
|
{
|
{
|
restore_best_regs:
|
restore_best_regs:
|
/* Restore best match. It may happen that `dend ==
|
/* Restore best match. It may happen that `dend ==
|
end_match_1' while the restored d is in string2.
|
end_match_1' while the restored d is in string2.
|
For example, the pattern `x.*y.*z' against the
|
For example, the pattern `x.*y.*z' against the
|
strings `x-' and `y-z-', if the two strings are
|
strings `x-' and `y-z-', if the two strings are
|
not consecutive in memory. */
|
not consecutive in memory. */
|
DEBUG_PRINT1 ("Restoring best registers.\n");
|
DEBUG_PRINT1 ("Restoring best registers.\n");
|
|
|
d = match_end;
|
d = match_end;
|
dend = ((d >= string1 && d <= end1)
|
dend = ((d >= string1 && d <= end1)
|
? end_match_1 : end_match_2);
|
? end_match_1 : end_match_2);
|
|
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
|
{
|
{
|
regstart[mcnt] = best_regstart[mcnt];
|
regstart[mcnt] = best_regstart[mcnt];
|
regend[mcnt] = best_regend[mcnt];
|
regend[mcnt] = best_regend[mcnt];
|
}
|
}
|
}
|
}
|
} /* d != end_match_2 */
|
} /* d != end_match_2 */
|
|
|
succeed_label:
|
succeed_label:
|
DEBUG_PRINT1 ("Accepting match.\n");
|
DEBUG_PRINT1 ("Accepting match.\n");
|
|
|
/* If caller wants register contents data back, do it. */
|
/* If caller wants register contents data back, do it. */
|
if (regs && !bufp->no_sub)
|
if (regs && !bufp->no_sub)
|
{
|
{
|
/* Have the register data arrays been allocated? */
|
/* Have the register data arrays been allocated? */
|
if (bufp->regs_allocated == REGS_UNALLOCATED)
|
if (bufp->regs_allocated == REGS_UNALLOCATED)
|
{ /* No. So allocate them with malloc. We need one
|
{ /* No. So allocate them with malloc. We need one
|
extra element beyond `num_regs' for the `-1' marker
|
extra element beyond `num_regs' for the `-1' marker
|
GNU code uses. */
|
GNU code uses. */
|
regs->num_regs = MAX (RE_NREGS, num_regs + 1);
|
regs->num_regs = MAX (RE_NREGS, num_regs + 1);
|
regs->start = TALLOC (regs->num_regs, regoff_t);
|
regs->start = TALLOC (regs->num_regs, regoff_t);
|
regs->end = TALLOC (regs->num_regs, regoff_t);
|
regs->end = TALLOC (regs->num_regs, regoff_t);
|
if (regs->start == NULL || regs->end == NULL)
|
if (regs->start == NULL || regs->end == NULL)
|
{
|
{
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
return -2;
|
return -2;
|
}
|
}
|
bufp->regs_allocated = REGS_REALLOCATE;
|
bufp->regs_allocated = REGS_REALLOCATE;
|
}
|
}
|
else if (bufp->regs_allocated == REGS_REALLOCATE)
|
else if (bufp->regs_allocated == REGS_REALLOCATE)
|
{ /* Yes. If we need more elements than were already
|
{ /* Yes. If we need more elements than were already
|
allocated, reallocate them. If we need fewer, just
|
allocated, reallocate them. If we need fewer, just
|
leave it alone. */
|
leave it alone. */
|
if (regs->num_regs < num_regs + 1)
|
if (regs->num_regs < num_regs + 1)
|
{
|
{
|
regs->num_regs = num_regs + 1;
|
regs->num_regs = num_regs + 1;
|
RETALLOC (regs->start, regs->num_regs, regoff_t);
|
RETALLOC (regs->start, regs->num_regs, regoff_t);
|
RETALLOC (regs->end, regs->num_regs, regoff_t);
|
RETALLOC (regs->end, regs->num_regs, regoff_t);
|
if (regs->start == NULL || regs->end == NULL)
|
if (regs->start == NULL || regs->end == NULL)
|
{
|
{
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
return -2;
|
return -2;
|
}
|
}
|
}
|
}
|
}
|
}
|
else
|
else
|
{
|
{
|
/* These braces fend off a "empty body in an else-statement"
|
/* These braces fend off a "empty body in an else-statement"
|
warning under GCC when assert expands to nothing. */
|
warning under GCC when assert expands to nothing. */
|
assert (bufp->regs_allocated == REGS_FIXED);
|
assert (bufp->regs_allocated == REGS_FIXED);
|
}
|
}
|
|
|
/* Convert the pointer data in `regstart' and `regend' to
|
/* Convert the pointer data in `regstart' and `regend' to
|
indices. Register zero has to be set differently,
|
indices. Register zero has to be set differently,
|
since we haven't kept track of any info for it. */
|
since we haven't kept track of any info for it. */
|
if (regs->num_regs > 0)
|
if (regs->num_regs > 0)
|
{
|
{
|
regs->start[0] = pos;
|
regs->start[0] = pos;
|
regs->end[0] = (MATCHING_IN_FIRST_STRING
|
regs->end[0] = (MATCHING_IN_FIRST_STRING
|
? ((regoff_t) (d - string1))
|
? ((regoff_t) (d - string1))
|
: ((regoff_t) (d - string2 + size1)));
|
: ((regoff_t) (d - string2 + size1)));
|
}
|
}
|
|
|
/* Go through the first `min (num_regs, regs->num_regs)'
|
/* Go through the first `min (num_regs, regs->num_regs)'
|
registers, since that is all we initialized. */
|
registers, since that is all we initialized. */
|
for (mcnt = 1; (unsigned) mcnt < MIN (num_regs, regs->num_regs);
|
for (mcnt = 1; (unsigned) mcnt < MIN (num_regs, regs->num_regs);
|
mcnt++)
|
mcnt++)
|
{
|
{
|
if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt]))
|
if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt]))
|
regs->start[mcnt] = regs->end[mcnt] = -1;
|
regs->start[mcnt] = regs->end[mcnt] = -1;
|
else
|
else
|
{
|
{
|
regs->start[mcnt]
|
regs->start[mcnt]
|
= (regoff_t) POINTER_TO_OFFSET (regstart[mcnt]);
|
= (regoff_t) POINTER_TO_OFFSET (regstart[mcnt]);
|
regs->end[mcnt]
|
regs->end[mcnt]
|
= (regoff_t) POINTER_TO_OFFSET (regend[mcnt]);
|
= (regoff_t) POINTER_TO_OFFSET (regend[mcnt]);
|
}
|
}
|
}
|
}
|
|
|
/* If the regs structure we return has more elements than
|
/* If the regs structure we return has more elements than
|
were in the pattern, set the extra elements to -1. If
|
were in the pattern, set the extra elements to -1. If
|
we (re)allocated the registers, this is the case,
|
we (re)allocated the registers, this is the case,
|
because we always allocate enough to have at least one
|
because we always allocate enough to have at least one
|
-1 at the end. */
|
-1 at the end. */
|
for (mcnt = num_regs; (unsigned) mcnt < regs->num_regs; mcnt++)
|
for (mcnt = num_regs; (unsigned) mcnt < regs->num_regs; mcnt++)
|
regs->start[mcnt] = regs->end[mcnt] = -1;
|
regs->start[mcnt] = regs->end[mcnt] = -1;
|
} /* regs && !bufp->no_sub */
|
} /* regs && !bufp->no_sub */
|
|
|
DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n",
|
DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n",
|
nfailure_points_pushed, nfailure_points_popped,
|
nfailure_points_pushed, nfailure_points_popped,
|
nfailure_points_pushed - nfailure_points_popped);
|
nfailure_points_pushed - nfailure_points_popped);
|
DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed);
|
DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed);
|
|
|
mcnt = d - pos - (MATCHING_IN_FIRST_STRING
|
mcnt = d - pos - (MATCHING_IN_FIRST_STRING
|
? string1
|
? string1
|
: string2 - size1);
|
: string2 - size1);
|
|
|
DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt);
|
DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt);
|
|
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
return mcnt;
|
return mcnt;
|
}
|
}
|
|
|
/* Otherwise match next pattern command. */
|
/* Otherwise match next pattern command. */
|
switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
|
switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
|
{
|
{
|
/* Ignore these. Used to ignore the n of succeed_n's which
|
/* Ignore these. Used to ignore the n of succeed_n's which
|
currently have n == 0. */
|
currently have n == 0. */
|
case no_op:
|
case no_op:
|
DEBUG_PRINT1 ("EXECUTING no_op.\n");
|
DEBUG_PRINT1 ("EXECUTING no_op.\n");
|
break;
|
break;
|
|
|
case succeed:
|
case succeed:
|
DEBUG_PRINT1 ("EXECUTING succeed.\n");
|
DEBUG_PRINT1 ("EXECUTING succeed.\n");
|
goto succeed_label;
|
goto succeed_label;
|
|
|
/* Match the next n pattern characters exactly. The following
|
/* Match the next n pattern characters exactly. The following
|
byte in the pattern defines n, and the n bytes after that
|
byte in the pattern defines n, and the n bytes after that
|
are the characters to match. */
|
are the characters to match. */
|
case exactn:
|
case exactn:
|
mcnt = *p++;
|
mcnt = *p++;
|
DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt);
|
|
|
/* This is written out as an if-else so we don't waste time
|
/* This is written out as an if-else so we don't waste time
|
testing `translate' inside the loop. */
|
testing `translate' inside the loop. */
|
if (translate)
|
if (translate)
|
{
|
{
|
do
|
do
|
{
|
{
|
PREFETCH ();
|
PREFETCH ();
|
if ((unsigned char) translate[(unsigned char) *d++]
|
if ((unsigned char) translate[(unsigned char) *d++]
|
!= (unsigned char) *p++)
|
!= (unsigned char) *p++)
|
goto fail;
|
goto fail;
|
}
|
}
|
while (--mcnt);
|
while (--mcnt);
|
}
|
}
|
else
|
else
|
{
|
{
|
do
|
do
|
{
|
{
|
PREFETCH ();
|
PREFETCH ();
|
if (*d++ != (char) *p++) goto fail;
|
if (*d++ != (char) *p++) goto fail;
|
}
|
}
|
while (--mcnt);
|
while (--mcnt);
|
}
|
}
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
break;
|
break;
|
|
|
|
|
/* Match any character except possibly a newline or a null. */
|
/* Match any character except possibly a newline or a null. */
|
case anychar:
|
case anychar:
|
DEBUG_PRINT1 ("EXECUTING anychar.\n");
|
DEBUG_PRINT1 ("EXECUTING anychar.\n");
|
|
|
PREFETCH ();
|
PREFETCH ();
|
|
|
if ((!(bufp->syntax & RE_DOT_NEWLINE) && TRANSLATE (*d) == '\n')
|
if ((!(bufp->syntax & RE_DOT_NEWLINE) && TRANSLATE (*d) == '\n')
|
|| (bufp->syntax & RE_DOT_NOT_NULL && TRANSLATE (*d) == '\000'))
|
|| (bufp->syntax & RE_DOT_NOT_NULL && TRANSLATE (*d) == '\000'))
|
goto fail;
|
goto fail;
|
|
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
DEBUG_PRINT2 (" Matched `%d'.\n", *d);
|
DEBUG_PRINT2 (" Matched `%d'.\n", *d);
|
d++;
|
d++;
|
break;
|
break;
|
|
|
|
|
case charset:
|
case charset:
|
case charset_not:
|
case charset_not:
|
{
|
{
|
register unsigned char c;
|
register unsigned char c;
|
boolean not = (re_opcode_t) *(p - 1) == charset_not;
|
boolean not = (re_opcode_t) *(p - 1) == charset_not;
|
|
|
DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : "");
|
DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : "");
|
|
|
PREFETCH ();
|
PREFETCH ();
|
c = TRANSLATE (*d); /* The character to match. */
|
c = TRANSLATE (*d); /* The character to match. */
|
|
|
/* Cast to `unsigned' instead of `unsigned char' in case the
|
/* Cast to `unsigned' instead of `unsigned char' in case the
|
bit list is a full 32 bytes long. */
|
bit list is a full 32 bytes long. */
|
if (c < (unsigned) (*p * BYTEWIDTH)
|
if (c < (unsigned) (*p * BYTEWIDTH)
|
&& p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
|
&& p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
|
not = !not;
|
not = !not;
|
|
|
p += 1 + *p;
|
p += 1 + *p;
|
|
|
if (!not) goto fail;
|
if (!not) goto fail;
|
|
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
d++;
|
d++;
|
break;
|
break;
|
}
|
}
|
|
|
|
|
/* The beginning of a group is represented by start_memory.
|
/* The beginning of a group is represented by start_memory.
|
The arguments are the register number in the next byte, and the
|
The arguments are the register number in the next byte, and the
|
number of groups inner to this one in the next. The text
|
number of groups inner to this one in the next. The text
|
matched within the group is recorded (in the internal
|
matched within the group is recorded (in the internal
|
registers data structure) under the register number. */
|
registers data structure) under the register number. */
|
case start_memory:
|
case start_memory:
|
DEBUG_PRINT3 ("EXECUTING start_memory %d (%d):\n", *p, p[1]);
|
DEBUG_PRINT3 ("EXECUTING start_memory %d (%d):\n", *p, p[1]);
|
|
|
/* Find out if this group can match the empty string. */
|
/* Find out if this group can match the empty string. */
|
p1 = p; /* To send to group_match_null_string_p. */
|
p1 = p; /* To send to group_match_null_string_p. */
|
|
|
if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE)
|
if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE)
|
REG_MATCH_NULL_STRING_P (reg_info[*p])
|
REG_MATCH_NULL_STRING_P (reg_info[*p])
|
= group_match_null_string_p (&p1, pend, reg_info);
|
= group_match_null_string_p (&p1, pend, reg_info);
|
|
|
/* Save the position in the string where we were the last time
|
/* Save the position in the string where we were the last time
|
we were at this open-group operator in case the group is
|
we were at this open-group operator in case the group is
|
operated upon by a repetition operator, e.g., with `(a*)*b'
|
operated upon by a repetition operator, e.g., with `(a*)*b'
|
against `ab'; then we want to ignore where we are now in
|
against `ab'; then we want to ignore where we are now in
|
the string in case this attempt to match fails. */
|
the string in case this attempt to match fails. */
|
old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
|
old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
|
? REG_UNSET (regstart[*p]) ? d : regstart[*p]
|
? REG_UNSET (regstart[*p]) ? d : regstart[*p]
|
: regstart[*p];
|
: regstart[*p];
|
DEBUG_PRINT2 (" old_regstart: %d\n",
|
DEBUG_PRINT2 (" old_regstart: %d\n",
|
POINTER_TO_OFFSET (old_regstart[*p]));
|
POINTER_TO_OFFSET (old_regstart[*p]));
|
|
|
regstart[*p] = d;
|
regstart[*p] = d;
|
DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p]));
|
DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p]));
|
|
|
IS_ACTIVE (reg_info[*p]) = 1;
|
IS_ACTIVE (reg_info[*p]) = 1;
|
MATCHED_SOMETHING (reg_info[*p]) = 0;
|
MATCHED_SOMETHING (reg_info[*p]) = 0;
|
|
|
/* Clear this whenever we change the register activity status. */
|
/* Clear this whenever we change the register activity status. */
|
set_regs_matched_done = 0;
|
set_regs_matched_done = 0;
|
|
|
/* This is the new highest active register. */
|
/* This is the new highest active register. */
|
highest_active_reg = *p;
|
highest_active_reg = *p;
|
|
|
/* If nothing was active before, this is the new lowest active
|
/* If nothing was active before, this is the new lowest active
|
register. */
|
register. */
|
if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
|
if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
|
lowest_active_reg = *p;
|
lowest_active_reg = *p;
|
|
|
/* Move past the register number and inner group count. */
|
/* Move past the register number and inner group count. */
|
p += 2;
|
p += 2;
|
just_past_start_mem = p;
|
just_past_start_mem = p;
|
|
|
break;
|
break;
|
|
|
|
|
/* The stop_memory opcode represents the end of a group. Its
|
/* The stop_memory opcode represents the end of a group. Its
|
arguments are the same as start_memory's: the register
|
arguments are the same as start_memory's: the register
|
number, and the number of inner groups. */
|
number, and the number of inner groups. */
|
case stop_memory:
|
case stop_memory:
|
DEBUG_PRINT3 ("EXECUTING stop_memory %d (%d):\n", *p, p[1]);
|
DEBUG_PRINT3 ("EXECUTING stop_memory %d (%d):\n", *p, p[1]);
|
|
|
/* We need to save the string position the last time we were at
|
/* We need to save the string position the last time we were at
|
this close-group operator in case the group is operated
|
this close-group operator in case the group is operated
|
upon by a repetition operator, e.g., with `((a*)*(b*)*)*'
|
upon by a repetition operator, e.g., with `((a*)*(b*)*)*'
|
against `aba'; then we want to ignore where we are now in
|
against `aba'; then we want to ignore where we are now in
|
the string in case this attempt to match fails. */
|
the string in case this attempt to match fails. */
|
old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
|
old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
|
? REG_UNSET (regend[*p]) ? d : regend[*p]
|
? REG_UNSET (regend[*p]) ? d : regend[*p]
|
: regend[*p];
|
: regend[*p];
|
DEBUG_PRINT2 (" old_regend: %d\n",
|
DEBUG_PRINT2 (" old_regend: %d\n",
|
POINTER_TO_OFFSET (old_regend[*p]));
|
POINTER_TO_OFFSET (old_regend[*p]));
|
|
|
regend[*p] = d;
|
regend[*p] = d;
|
DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p]));
|
DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p]));
|
|
|
/* This register isn't active anymore. */
|
/* This register isn't active anymore. */
|
IS_ACTIVE (reg_info[*p]) = 0;
|
IS_ACTIVE (reg_info[*p]) = 0;
|
|
|
/* Clear this whenever we change the register activity status. */
|
/* Clear this whenever we change the register activity status. */
|
set_regs_matched_done = 0;
|
set_regs_matched_done = 0;
|
|
|
/* If this was the only register active, nothing is active
|
/* If this was the only register active, nothing is active
|
anymore. */
|
anymore. */
|
if (lowest_active_reg == highest_active_reg)
|
if (lowest_active_reg == highest_active_reg)
|
{
|
{
|
lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
}
|
}
|
else
|
else
|
{ /* We must scan for the new highest active register, since
|
{ /* We must scan for the new highest active register, since
|
it isn't necessarily one less than now: consider
|
it isn't necessarily one less than now: consider
|
(a(b)c(d(e)f)g). When group 3 ends, after the f), the
|
(a(b)c(d(e)f)g). When group 3 ends, after the f), the
|
new highest active register is 1. */
|
new highest active register is 1. */
|
unsigned char r = *p - 1;
|
unsigned char r = *p - 1;
|
while (r > 0 && !IS_ACTIVE (reg_info[r]))
|
while (r > 0 && !IS_ACTIVE (reg_info[r]))
|
r--;
|
r--;
|
|
|
/* If we end up at register zero, that means that we saved
|
/* If we end up at register zero, that means that we saved
|
the registers as the result of an `on_failure_jump', not
|
the registers as the result of an `on_failure_jump', not
|
a `start_memory', and we jumped to past the innermost
|
a `start_memory', and we jumped to past the innermost
|
`stop_memory'. For example, in ((.)*) we save
|
`stop_memory'. For example, in ((.)*) we save
|
registers 1 and 2 as a result of the *, but when we pop
|
registers 1 and 2 as a result of the *, but when we pop
|
back to the second ), we are at the stop_memory 1.
|
back to the second ), we are at the stop_memory 1.
|
Thus, nothing is active. */
|
Thus, nothing is active. */
|
if (r == 0)
|
if (r == 0)
|
{
|
{
|
lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
lowest_active_reg = NO_LOWEST_ACTIVE_REG;
|
highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
highest_active_reg = NO_HIGHEST_ACTIVE_REG;
|
}
|
}
|
else
|
else
|
highest_active_reg = r;
|
highest_active_reg = r;
|
}
|
}
|
|
|
/* If just failed to match something this time around with a
|
/* If just failed to match something this time around with a
|
group that's operated on by a repetition operator, try to
|
group that's operated on by a repetition operator, try to
|
force exit from the ``loop'', and restore the register
|
force exit from the ``loop'', and restore the register
|
information for this group that we had before trying this
|
information for this group that we had before trying this
|
last match. */
|
last match. */
|
if ((!MATCHED_SOMETHING (reg_info[*p])
|
if ((!MATCHED_SOMETHING (reg_info[*p])
|
|| just_past_start_mem == p - 1)
|
|| just_past_start_mem == p - 1)
|
&& (p + 2) < pend)
|
&& (p + 2) < pend)
|
{
|
{
|
boolean is_a_jump_n = false;
|
boolean is_a_jump_n = false;
|
|
|
p1 = p + 2;
|
p1 = p + 2;
|
mcnt = 0;
|
mcnt = 0;
|
switch ((re_opcode_t) *p1++)
|
switch ((re_opcode_t) *p1++)
|
{
|
{
|
case jump_n:
|
case jump_n:
|
is_a_jump_n = true;
|
is_a_jump_n = true;
|
case pop_failure_jump:
|
case pop_failure_jump:
|
case maybe_pop_jump:
|
case maybe_pop_jump:
|
case jump:
|
case jump:
|
case dummy_failure_jump:
|
case dummy_failure_jump:
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
if (is_a_jump_n)
|
if (is_a_jump_n)
|
p1 += 2;
|
p1 += 2;
|
break;
|
break;
|
|
|
default:
|
default:
|
/* do nothing */ ;
|
/* do nothing */ ;
|
}
|
}
|
p1 += mcnt;
|
p1 += mcnt;
|
|
|
/* If the next operation is a jump backwards in the pattern
|
/* If the next operation is a jump backwards in the pattern
|
to an on_failure_jump right before the start_memory
|
to an on_failure_jump right before the start_memory
|
corresponding to this stop_memory, exit from the loop
|
corresponding to this stop_memory, exit from the loop
|
by forcing a failure after pushing on the stack the
|
by forcing a failure after pushing on the stack the
|
on_failure_jump's jump in the pattern, and d. */
|
on_failure_jump's jump in the pattern, and d. */
|
if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump
|
if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump
|
&& (re_opcode_t) p1[3] == start_memory && p1[4] == *p)
|
&& (re_opcode_t) p1[3] == start_memory && p1[4] == *p)
|
{
|
{
|
/* If this group ever matched anything, then restore
|
/* If this group ever matched anything, then restore
|
what its registers were before trying this last
|
what its registers were before trying this last
|
failed match, e.g., with `(a*)*b' against `ab' for
|
failed match, e.g., with `(a*)*b' against `ab' for
|
regstart[1], and, e.g., with `((a*)*(b*)*)*'
|
regstart[1], and, e.g., with `((a*)*(b*)*)*'
|
against `aba' for regend[3].
|
against `aba' for regend[3].
|
|
|
Also restore the registers for inner groups for,
|
Also restore the registers for inner groups for,
|
e.g., `((a*)(b*))*' against `aba' (register 3 would
|
e.g., `((a*)(b*))*' against `aba' (register 3 would
|
otherwise get trashed). */
|
otherwise get trashed). */
|
|
|
if (EVER_MATCHED_SOMETHING (reg_info[*p]))
|
if (EVER_MATCHED_SOMETHING (reg_info[*p]))
|
{
|
{
|
unsigned r;
|
unsigned r;
|
|
|
EVER_MATCHED_SOMETHING (reg_info[*p]) = 0;
|
EVER_MATCHED_SOMETHING (reg_info[*p]) = 0;
|
|
|
/* Restore this and inner groups' (if any) registers. */
|
/* Restore this and inner groups' (if any) registers. */
|
for (r = *p; r < (unsigned) *p + (unsigned) *(p + 1);
|
for (r = *p; r < (unsigned) *p + (unsigned) *(p + 1);
|
r++)
|
r++)
|
{
|
{
|
regstart[r] = old_regstart[r];
|
regstart[r] = old_regstart[r];
|
|
|
/* xx why this test? */
|
/* xx why this test? */
|
if (old_regend[r] >= regstart[r])
|
if (old_regend[r] >= regstart[r])
|
regend[r] = old_regend[r];
|
regend[r] = old_regend[r];
|
}
|
}
|
}
|
}
|
p1++;
|
p1++;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
PUSH_FAILURE_POINT (p1 + mcnt, d, -2);
|
PUSH_FAILURE_POINT (p1 + mcnt, d, -2);
|
|
|
goto fail;
|
goto fail;
|
}
|
}
|
}
|
}
|
|
|
/* Move past the register number and the inner group count. */
|
/* Move past the register number and the inner group count. */
|
p += 2;
|
p += 2;
|
break;
|
break;
|
|
|
|
|
/* \<digit> has been turned into a `duplicate' command which is
|
/* \<digit> has been turned into a `duplicate' command which is
|
followed by the numeric value of <digit> as the register number. */
|
followed by the numeric value of <digit> as the register number. */
|
case duplicate:
|
case duplicate:
|
{
|
{
|
register const char *d2, *dend2;
|
register const char *d2, *dend2;
|
int regno = *p++; /* Get which register to match against. */
|
int regno = *p++; /* Get which register to match against. */
|
DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno);
|
DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno);
|
|
|
/* Can't back reference a group which we've never matched. */
|
/* Can't back reference a group which we've never matched. */
|
if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno]))
|
if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno]))
|
goto fail;
|
goto fail;
|
|
|
/* Where in input to try to start matching. */
|
/* Where in input to try to start matching. */
|
d2 = regstart[regno];
|
d2 = regstart[regno];
|
|
|
/* Where to stop matching; if both the place to start and
|
/* Where to stop matching; if both the place to start and
|
the place to stop matching are in the same string, then
|
the place to stop matching are in the same string, then
|
set to the place to stop, otherwise, for now have to use
|
set to the place to stop, otherwise, for now have to use
|
the end of the first string. */
|
the end of the first string. */
|
|
|
dend2 = ((FIRST_STRING_P (regstart[regno])
|
dend2 = ((FIRST_STRING_P (regstart[regno])
|
== FIRST_STRING_P (regend[regno]))
|
== FIRST_STRING_P (regend[regno]))
|
? regend[regno] : end_match_1);
|
? regend[regno] : end_match_1);
|
for (;;)
|
for (;;)
|
{
|
{
|
/* If necessary, advance to next segment in register
|
/* If necessary, advance to next segment in register
|
contents. */
|
contents. */
|
while (d2 == dend2)
|
while (d2 == dend2)
|
{
|
{
|
if (dend2 == end_match_2) break;
|
if (dend2 == end_match_2) break;
|
if (dend2 == regend[regno]) break;
|
if (dend2 == regend[regno]) break;
|
|
|
/* End of string1 => advance to string2. */
|
/* End of string1 => advance to string2. */
|
d2 = string2;
|
d2 = string2;
|
dend2 = regend[regno];
|
dend2 = regend[regno];
|
}
|
}
|
/* At end of register contents => success */
|
/* At end of register contents => success */
|
if (d2 == dend2) break;
|
if (d2 == dend2) break;
|
|
|
/* If necessary, advance to next segment in data. */
|
/* If necessary, advance to next segment in data. */
|
PREFETCH ();
|
PREFETCH ();
|
|
|
/* How many characters left in this segment to match. */
|
/* How many characters left in this segment to match. */
|
mcnt = dend - d;
|
mcnt = dend - d;
|
|
|
/* Want how many consecutive characters we can match in
|
/* Want how many consecutive characters we can match in
|
one shot, so, if necessary, adjust the count. */
|
one shot, so, if necessary, adjust the count. */
|
if (mcnt > dend2 - d2)
|
if (mcnt > dend2 - d2)
|
mcnt = dend2 - d2;
|
mcnt = dend2 - d2;
|
|
|
/* Compare that many; failure if mismatch, else move
|
/* Compare that many; failure if mismatch, else move
|
past them. */
|
past them. */
|
if (translate
|
if (translate
|
? bcmp_translate (d, d2, mcnt, translate)
|
? bcmp_translate (d, d2, mcnt, translate)
|
: memcmp (d, d2, mcnt))
|
: memcmp (d, d2, mcnt))
|
goto fail;
|
goto fail;
|
d += mcnt, d2 += mcnt;
|
d += mcnt, d2 += mcnt;
|
|
|
/* Do this because we've match some characters. */
|
/* Do this because we've match some characters. */
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
}
|
}
|
}
|
}
|
break;
|
break;
|
|
|
|
|
/* begline matches the empty string at the beginning of the string
|
/* begline matches the empty string at the beginning of the string
|
(unless `not_bol' is set in `bufp'), and, if
|
(unless `not_bol' is set in `bufp'), and, if
|
`newline_anchor' is set, after newlines. */
|
`newline_anchor' is set, after newlines. */
|
case begline:
|
case begline:
|
DEBUG_PRINT1 ("EXECUTING begline.\n");
|
DEBUG_PRINT1 ("EXECUTING begline.\n");
|
|
|
if (AT_STRINGS_BEG (d))
|
if (AT_STRINGS_BEG (d))
|
{
|
{
|
if (!bufp->not_bol) break;
|
if (!bufp->not_bol) break;
|
}
|
}
|
else if (d[-1] == '\n' && bufp->newline_anchor)
|
else if (d[-1] == '\n' && bufp->newline_anchor)
|
{
|
{
|
break;
|
break;
|
}
|
}
|
/* In all other cases, we fail. */
|
/* In all other cases, we fail. */
|
goto fail;
|
goto fail;
|
|
|
|
|
/* endline is the dual of begline. */
|
/* endline is the dual of begline. */
|
case endline:
|
case endline:
|
DEBUG_PRINT1 ("EXECUTING endline.\n");
|
DEBUG_PRINT1 ("EXECUTING endline.\n");
|
|
|
if (AT_STRINGS_END (d))
|
if (AT_STRINGS_END (d))
|
{
|
{
|
if (!bufp->not_eol) break;
|
if (!bufp->not_eol) break;
|
}
|
}
|
|
|
/* We have to ``prefetch'' the next character. */
|
/* We have to ``prefetch'' the next character. */
|
else if ((d == end1 ? *string2 : *d) == '\n'
|
else if ((d == end1 ? *string2 : *d) == '\n'
|
&& bufp->newline_anchor)
|
&& bufp->newline_anchor)
|
{
|
{
|
break;
|
break;
|
}
|
}
|
goto fail;
|
goto fail;
|
|
|
|
|
/* Match at the very beginning of the data. */
|
/* Match at the very beginning of the data. */
|
case begbuf:
|
case begbuf:
|
DEBUG_PRINT1 ("EXECUTING begbuf.\n");
|
DEBUG_PRINT1 ("EXECUTING begbuf.\n");
|
if (AT_STRINGS_BEG (d))
|
if (AT_STRINGS_BEG (d))
|
break;
|
break;
|
goto fail;
|
goto fail;
|
|
|
|
|
/* Match at the very end of the data. */
|
/* Match at the very end of the data. */
|
case endbuf:
|
case endbuf:
|
DEBUG_PRINT1 ("EXECUTING endbuf.\n");
|
DEBUG_PRINT1 ("EXECUTING endbuf.\n");
|
if (AT_STRINGS_END (d))
|
if (AT_STRINGS_END (d))
|
break;
|
break;
|
goto fail;
|
goto fail;
|
|
|
|
|
/* on_failure_keep_string_jump is used to optimize `.*\n'. It
|
/* on_failure_keep_string_jump is used to optimize `.*\n'. It
|
pushes NULL as the value for the string on the stack. Then
|
pushes NULL as the value for the string on the stack. Then
|
`pop_failure_point' will keep the current value for the
|
`pop_failure_point' will keep the current value for the
|
string, instead of restoring it. To see why, consider
|
string, instead of restoring it. To see why, consider
|
matching `foo\nbar' against `.*\n'. The .* matches the foo;
|
matching `foo\nbar' against `.*\n'. The .* matches the foo;
|
then the . fails against the \n. But the next thing we want
|
then the . fails against the \n. But the next thing we want
|
to do is match the \n against the \n; if we restored the
|
to do is match the \n against the \n; if we restored the
|
string value, we would be back at the foo.
|
string value, we would be back at the foo.
|
|
|
Because this is used only in specific cases, we don't need to
|
Because this is used only in specific cases, we don't need to
|
check all the things that `on_failure_jump' does, to make
|
check all the things that `on_failure_jump' does, to make
|
sure the right things get saved on the stack. Hence we don't
|
sure the right things get saved on the stack. Hence we don't
|
share its code. The only reason to push anything on the
|
share its code. The only reason to push anything on the
|
stack at all is that otherwise we would have to change
|
stack at all is that otherwise we would have to change
|
`anychar's code to do something besides goto fail in this
|
`anychar's code to do something besides goto fail in this
|
case; that seems worse than this. */
|
case; that seems worse than this. */
|
case on_failure_keep_string_jump:
|
case on_failure_keep_string_jump:
|
DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump");
|
DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump");
|
|
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT3 (" %d (to %p):\n", mcnt, p + mcnt);
|
DEBUG_PRINT3 (" %d (to %p):\n", mcnt, p + mcnt);
|
#else
|
#else
|
DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt);
|
DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt);
|
#endif
|
#endif
|
|
|
PUSH_FAILURE_POINT (p + mcnt, NULL, -2);
|
PUSH_FAILURE_POINT (p + mcnt, NULL, -2);
|
break;
|
break;
|
|
|
|
|
/* Uses of on_failure_jump:
|
/* Uses of on_failure_jump:
|
|
|
Each alternative starts with an on_failure_jump that points
|
Each alternative starts with an on_failure_jump that points
|
to the beginning of the next alternative. Each alternative
|
to the beginning of the next alternative. Each alternative
|
except the last ends with a jump that in effect jumps past
|
except the last ends with a jump that in effect jumps past
|
the rest of the alternatives. (They really jump to the
|
the rest of the alternatives. (They really jump to the
|
ending jump of the following alternative, because tensioning
|
ending jump of the following alternative, because tensioning
|
these jumps is a hassle.)
|
these jumps is a hassle.)
|
|
|
Repeats start with an on_failure_jump that points past both
|
Repeats start with an on_failure_jump that points past both
|
the repetition text and either the following jump or
|
the repetition text and either the following jump or
|
pop_failure_jump back to this on_failure_jump. */
|
pop_failure_jump back to this on_failure_jump. */
|
case on_failure_jump:
|
case on_failure_jump:
|
on_failure:
|
on_failure:
|
DEBUG_PRINT1 ("EXECUTING on_failure_jump");
|
DEBUG_PRINT1 ("EXECUTING on_failure_jump");
|
|
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT3 (" %d (to %p)", mcnt, p + mcnt);
|
DEBUG_PRINT3 (" %d (to %p)", mcnt, p + mcnt);
|
#else
|
#else
|
DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt);
|
DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt);
|
#endif
|
#endif
|
|
|
/* If this on_failure_jump comes right before a group (i.e.,
|
/* If this on_failure_jump comes right before a group (i.e.,
|
the original * applied to a group), save the information
|
the original * applied to a group), save the information
|
for that group and all inner ones, so that if we fail back
|
for that group and all inner ones, so that if we fail back
|
to this point, the group's information will be correct.
|
to this point, the group's information will be correct.
|
For example, in \(a*\)*\1, we need the preceding group,
|
For example, in \(a*\)*\1, we need the preceding group,
|
and in \(zz\(a*\)b*\)\2, we need the inner group. */
|
and in \(zz\(a*\)b*\)\2, we need the inner group. */
|
|
|
/* We can't use `p' to check ahead because we push
|
/* We can't use `p' to check ahead because we push
|
a failure point to `p + mcnt' after we do this. */
|
a failure point to `p + mcnt' after we do this. */
|
p1 = p;
|
p1 = p;
|
|
|
/* We need to skip no_op's before we look for the
|
/* We need to skip no_op's before we look for the
|
start_memory in case this on_failure_jump is happening as
|
start_memory in case this on_failure_jump is happening as
|
the result of a completed succeed_n, as in \(a\)\{1,3\}b\1
|
the result of a completed succeed_n, as in \(a\)\{1,3\}b\1
|
against aba. */
|
against aba. */
|
while (p1 < pend && (re_opcode_t) *p1 == no_op)
|
while (p1 < pend && (re_opcode_t) *p1 == no_op)
|
p1++;
|
p1++;
|
|
|
if (p1 < pend && (re_opcode_t) *p1 == start_memory)
|
if (p1 < pend && (re_opcode_t) *p1 == start_memory)
|
{
|
{
|
/* We have a new highest active register now. This will
|
/* We have a new highest active register now. This will
|
get reset at the start_memory we are about to get to,
|
get reset at the start_memory we are about to get to,
|
but we will have saved all the registers relevant to
|
but we will have saved all the registers relevant to
|
this repetition op, as described above. */
|
this repetition op, as described above. */
|
highest_active_reg = *(p1 + 1) + *(p1 + 2);
|
highest_active_reg = *(p1 + 1) + *(p1 + 2);
|
if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
|
if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
|
lowest_active_reg = *(p1 + 1);
|
lowest_active_reg = *(p1 + 1);
|
}
|
}
|
|
|
DEBUG_PRINT1 (":\n");
|
DEBUG_PRINT1 (":\n");
|
PUSH_FAILURE_POINT (p + mcnt, d, -2);
|
PUSH_FAILURE_POINT (p + mcnt, d, -2);
|
break;
|
break;
|
|
|
|
|
/* A smart repeat ends with `maybe_pop_jump'.
|
/* A smart repeat ends with `maybe_pop_jump'.
|
We change it to either `pop_failure_jump' or `jump'. */
|
We change it to either `pop_failure_jump' or `jump'. */
|
case maybe_pop_jump:
|
case maybe_pop_jump:
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt);
|
{
|
{
|
register unsigned char *p2 = p;
|
register unsigned char *p2 = p;
|
|
|
/* Compare the beginning of the repeat with what in the
|
/* Compare the beginning of the repeat with what in the
|
pattern follows its end. If we can establish that there
|
pattern follows its end. If we can establish that there
|
is nothing that they would both match, i.e., that we
|
is nothing that they would both match, i.e., that we
|
would have to backtrack because of (as in, e.g., `a*a')
|
would have to backtrack because of (as in, e.g., `a*a')
|
then we can change to pop_failure_jump, because we'll
|
then we can change to pop_failure_jump, because we'll
|
never have to backtrack.
|
never have to backtrack.
|
|
|
This is not true in the case of alternatives: in
|
This is not true in the case of alternatives: in
|
`(a|ab)*' we do need to backtrack to the `ab' alternative
|
`(a|ab)*' we do need to backtrack to the `ab' alternative
|
(e.g., if the string was `ab'). But instead of trying to
|
(e.g., if the string was `ab'). But instead of trying to
|
detect that here, the alternative has put on a dummy
|
detect that here, the alternative has put on a dummy
|
failure point which is what we will end up popping. */
|
failure point which is what we will end up popping. */
|
|
|
/* Skip over open/close-group commands.
|
/* Skip over open/close-group commands.
|
If what follows this loop is a ...+ construct,
|
If what follows this loop is a ...+ construct,
|
look at what begins its body, since we will have to
|
look at what begins its body, since we will have to
|
match at least one of that. */
|
match at least one of that. */
|
while (1)
|
while (1)
|
{
|
{
|
if (p2 + 2 < pend
|
if (p2 + 2 < pend
|
&& ((re_opcode_t) *p2 == stop_memory
|
&& ((re_opcode_t) *p2 == stop_memory
|
|| (re_opcode_t) *p2 == start_memory))
|
|| (re_opcode_t) *p2 == start_memory))
|
p2 += 3;
|
p2 += 3;
|
else if (p2 + 6 < pend
|
else if (p2 + 6 < pend
|
&& (re_opcode_t) *p2 == dummy_failure_jump)
|
&& (re_opcode_t) *p2 == dummy_failure_jump)
|
p2 += 6;
|
p2 += 6;
|
else
|
else
|
break;
|
break;
|
}
|
}
|
|
|
p1 = p + mcnt;
|
p1 = p + mcnt;
|
/* p1[0] ... p1[2] are the `on_failure_jump' corresponding
|
/* p1[0] ... p1[2] are the `on_failure_jump' corresponding
|
to the `maybe_finalize_jump' of this case. Examine what
|
to the `maybe_finalize_jump' of this case. Examine what
|
follows. */
|
follows. */
|
|
|
/* If we're at the end of the pattern, we can change. */
|
/* If we're at the end of the pattern, we can change. */
|
if (p2 == pend)
|
if (p2 == pend)
|
{
|
{
|
/* Consider what happens when matching ":\(.*\)"
|
/* Consider what happens when matching ":\(.*\)"
|
against ":/". I don't really understand this code
|
against ":/". I don't really understand this code
|
yet. */
|
yet. */
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT1
|
DEBUG_PRINT1
|
(" End of pattern: change to `pop_failure_jump'.\n");
|
(" End of pattern: change to `pop_failure_jump'.\n");
|
}
|
}
|
|
|
else if ((re_opcode_t) *p2 == exactn
|
else if ((re_opcode_t) *p2 == exactn
|
|| (bufp->newline_anchor && (re_opcode_t) *p2 == endline))
|
|| (bufp->newline_anchor && (re_opcode_t) *p2 == endline))
|
{
|
{
|
register unsigned char c
|
register unsigned char c
|
= *p2 == (unsigned char) endline ? '\n' : p2[2];
|
= *p2 == (unsigned char) endline ? '\n' : p2[2];
|
|
|
if ((re_opcode_t) p1[3] == exactn && p1[5] != c)
|
if ((re_opcode_t) p1[3] == exactn && p1[5] != c)
|
{
|
{
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n",
|
DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n",
|
c, p1[5]);
|
c, p1[5]);
|
}
|
}
|
|
|
else if ((re_opcode_t) p1[3] == charset
|
else if ((re_opcode_t) p1[3] == charset
|
|| (re_opcode_t) p1[3] == charset_not)
|
|| (re_opcode_t) p1[3] == charset_not)
|
{
|
{
|
int not = (re_opcode_t) p1[3] == charset_not;
|
int not = (re_opcode_t) p1[3] == charset_not;
|
|
|
if (c < (unsigned char) (p1[4] * BYTEWIDTH)
|
if (c < (unsigned char) (p1[4] * BYTEWIDTH)
|
&& p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
|
&& p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
|
not = !not;
|
not = !not;
|
|
|
/* `not' is equal to 1 if c would match, which means
|
/* `not' is equal to 1 if c would match, which means
|
that we can't change to pop_failure_jump. */
|
that we can't change to pop_failure_jump. */
|
if (!not)
|
if (!not)
|
{
|
{
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
}
|
}
|
}
|
}
|
}
|
}
|
else if ((re_opcode_t) *p2 == charset)
|
else if ((re_opcode_t) *p2 == charset)
|
{
|
{
|
#ifdef DEBUG
|
#ifdef DEBUG
|
register unsigned char c
|
register unsigned char c
|
= *p2 == (unsigned char) endline ? '\n' : p2[2];
|
= *p2 == (unsigned char) endline ? '\n' : p2[2];
|
#endif
|
#endif
|
|
|
#if 0
|
#if 0
|
if ((re_opcode_t) p1[3] == exactn
|
if ((re_opcode_t) p1[3] == exactn
|
&& ! ((int) p2[1] * BYTEWIDTH > (int) p1[5]
|
&& ! ((int) p2[1] * BYTEWIDTH > (int) p1[5]
|
&& (p2[2 + p1[5] / BYTEWIDTH]
|
&& (p2[2 + p1[5] / BYTEWIDTH]
|
& (1 << (p1[5] % BYTEWIDTH)))))
|
& (1 << (p1[5] % BYTEWIDTH)))))
|
#else
|
#else
|
if ((re_opcode_t) p1[3] == exactn
|
if ((re_opcode_t) p1[3] == exactn
|
&& ! ((int) p2[1] * BYTEWIDTH > (int) p1[4]
|
&& ! ((int) p2[1] * BYTEWIDTH > (int) p1[4]
|
&& (p2[2 + p1[4] / BYTEWIDTH]
|
&& (p2[2 + p1[4] / BYTEWIDTH]
|
& (1 << (p1[4] % BYTEWIDTH)))))
|
& (1 << (p1[4] % BYTEWIDTH)))))
|
#endif
|
#endif
|
{
|
{
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n",
|
DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n",
|
c, p1[5]);
|
c, p1[5]);
|
}
|
}
|
|
|
else if ((re_opcode_t) p1[3] == charset_not)
|
else if ((re_opcode_t) p1[3] == charset_not)
|
{
|
{
|
int idx;
|
int idx;
|
/* We win if the charset_not inside the loop
|
/* We win if the charset_not inside the loop
|
lists every character listed in the charset after. */
|
lists every character listed in the charset after. */
|
for (idx = 0; idx < (int) p2[1]; idx++)
|
for (idx = 0; idx < (int) p2[1]; idx++)
|
if (! (p2[2 + idx] == 0
|
if (! (p2[2 + idx] == 0
|
|| (idx < (int) p1[4]
|
|| (idx < (int) p1[4]
|
&& ((p2[2 + idx] & ~ p1[5 + idx]) == 0))))
|
&& ((p2[2 + idx] & ~ p1[5 + idx]) == 0))))
|
break;
|
break;
|
|
|
if (idx == p2[1])
|
if (idx == p2[1])
|
{
|
{
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
}
|
}
|
}
|
}
|
else if ((re_opcode_t) p1[3] == charset)
|
else if ((re_opcode_t) p1[3] == charset)
|
{
|
{
|
int idx;
|
int idx;
|
/* We win if the charset inside the loop
|
/* We win if the charset inside the loop
|
has no overlap with the one after the loop. */
|
has no overlap with the one after the loop. */
|
for (idx = 0;
|
for (idx = 0;
|
idx < (int) p2[1] && idx < (int) p1[4];
|
idx < (int) p2[1] && idx < (int) p1[4];
|
idx++)
|
idx++)
|
if ((p2[2 + idx] & p1[5 + idx]) != 0)
|
if ((p2[2 + idx] & p1[5 + idx]) != 0)
|
break;
|
break;
|
|
|
if (idx == p2[1] || idx == p1[4])
|
if (idx == p2[1] || idx == p1[4])
|
{
|
{
|
p[-3] = (unsigned char) pop_failure_jump;
|
p[-3] = (unsigned char) pop_failure_jump;
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
|
}
|
}
|
}
|
}
|
}
|
}
|
}
|
}
|
p -= 2; /* Point at relative address again. */
|
p -= 2; /* Point at relative address again. */
|
if ((re_opcode_t) p[-1] != pop_failure_jump)
|
if ((re_opcode_t) p[-1] != pop_failure_jump)
|
{
|
{
|
p[-1] = (unsigned char) jump;
|
p[-1] = (unsigned char) jump;
|
DEBUG_PRINT1 (" Match => jump.\n");
|
DEBUG_PRINT1 (" Match => jump.\n");
|
goto unconditional_jump;
|
goto unconditional_jump;
|
}
|
}
|
/* Note fall through. */
|
/* Note fall through. */
|
|
|
|
|
/* The end of a simple repeat has a pop_failure_jump back to
|
/* The end of a simple repeat has a pop_failure_jump back to
|
its matching on_failure_jump, where the latter will push a
|
its matching on_failure_jump, where the latter will push a
|
failure point. The pop_failure_jump takes off failure
|
failure point. The pop_failure_jump takes off failure
|
points put on by this pop_failure_jump's matching
|
points put on by this pop_failure_jump's matching
|
on_failure_jump; we got through the pattern to here from the
|
on_failure_jump; we got through the pattern to here from the
|
matching on_failure_jump, so didn't fail. */
|
matching on_failure_jump, so didn't fail. */
|
case pop_failure_jump:
|
case pop_failure_jump:
|
{
|
{
|
/* We need to pass separate storage for the lowest and
|
/* We need to pass separate storage for the lowest and
|
highest registers, even though we don't care about the
|
highest registers, even though we don't care about the
|
actual values. Otherwise, we will restore only one
|
actual values. Otherwise, we will restore only one
|
register from the stack, since lowest will == highest in
|
register from the stack, since lowest will == highest in
|
`pop_failure_point'. */
|
`pop_failure_point'. */
|
active_reg_t dummy_low_reg, dummy_high_reg;
|
active_reg_t dummy_low_reg, dummy_high_reg;
|
unsigned char *pdummy;
|
unsigned char *pdummy;
|
const char *sdummy;
|
const char *sdummy;
|
|
|
DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n");
|
DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n");
|
POP_FAILURE_POINT (sdummy, pdummy,
|
POP_FAILURE_POINT (sdummy, pdummy,
|
dummy_low_reg, dummy_high_reg,
|
dummy_low_reg, dummy_high_reg,
|
reg_dummy, reg_dummy, reg_info_dummy);
|
reg_dummy, reg_dummy, reg_info_dummy);
|
}
|
}
|
/* Note fall through. */
|
/* Note fall through. */
|
|
|
unconditional_jump:
|
unconditional_jump:
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT2 ("\n%p: ", p);
|
DEBUG_PRINT2 ("\n%p: ", p);
|
#else
|
#else
|
DEBUG_PRINT2 ("\n0x%x: ", p);
|
DEBUG_PRINT2 ("\n0x%x: ", p);
|
#endif
|
#endif
|
/* Note fall through. */
|
/* Note fall through. */
|
|
|
/* Unconditionally jump (without popping any failure points). */
|
/* Unconditionally jump (without popping any failure points). */
|
case jump:
|
case jump:
|
EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */
|
EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */
|
DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt);
|
DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt);
|
p += mcnt; /* Do the jump. */
|
p += mcnt; /* Do the jump. */
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT2 ("(to %p).\n", p);
|
DEBUG_PRINT2 ("(to %p).\n", p);
|
#else
|
#else
|
DEBUG_PRINT2 ("(to 0x%x).\n", p);
|
DEBUG_PRINT2 ("(to 0x%x).\n", p);
|
#endif
|
#endif
|
break;
|
break;
|
|
|
|
|
/* We need this opcode so we can detect where alternatives end
|
/* We need this opcode so we can detect where alternatives end
|
in `group_match_null_string_p' et al. */
|
in `group_match_null_string_p' et al. */
|
case jump_past_alt:
|
case jump_past_alt:
|
DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n");
|
DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n");
|
goto unconditional_jump;
|
goto unconditional_jump;
|
|
|
|
|
/* Normally, the on_failure_jump pushes a failure point, which
|
/* Normally, the on_failure_jump pushes a failure point, which
|
then gets popped at pop_failure_jump. We will end up at
|
then gets popped at pop_failure_jump. We will end up at
|
pop_failure_jump, also, and with a pattern of, say, `a+', we
|
pop_failure_jump, also, and with a pattern of, say, `a+', we
|
are skipping over the on_failure_jump, so we have to push
|
are skipping over the on_failure_jump, so we have to push
|
something meaningless for pop_failure_jump to pop. */
|
something meaningless for pop_failure_jump to pop. */
|
case dummy_failure_jump:
|
case dummy_failure_jump:
|
DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n");
|
DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n");
|
/* It doesn't matter what we push for the string here. What
|
/* It doesn't matter what we push for the string here. What
|
the code at `fail' tests is the value for the pattern. */
|
the code at `fail' tests is the value for the pattern. */
|
PUSH_FAILURE_POINT (NULL, NULL, -2);
|
PUSH_FAILURE_POINT (NULL, NULL, -2);
|
goto unconditional_jump;
|
goto unconditional_jump;
|
|
|
|
|
/* At the end of an alternative, we need to push a dummy failure
|
/* At the end of an alternative, we need to push a dummy failure
|
point in case we are followed by a `pop_failure_jump', because
|
point in case we are followed by a `pop_failure_jump', because
|
we don't want the failure point for the alternative to be
|
we don't want the failure point for the alternative to be
|
popped. For example, matching `(a|ab)*' against `aab'
|
popped. For example, matching `(a|ab)*' against `aab'
|
requires that we match the `ab' alternative. */
|
requires that we match the `ab' alternative. */
|
case push_dummy_failure:
|
case push_dummy_failure:
|
DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n");
|
DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n");
|
/* See comments just above at `dummy_failure_jump' about the
|
/* See comments just above at `dummy_failure_jump' about the
|
two zeroes. */
|
two zeroes. */
|
PUSH_FAILURE_POINT (NULL, NULL, -2);
|
PUSH_FAILURE_POINT (NULL, NULL, -2);
|
break;
|
break;
|
|
|
/* Have to succeed matching what follows at least n times.
|
/* Have to succeed matching what follows at least n times.
|
After that, handle like `on_failure_jump'. */
|
After that, handle like `on_failure_jump'. */
|
case succeed_n:
|
case succeed_n:
|
EXTRACT_NUMBER (mcnt, p + 2);
|
EXTRACT_NUMBER (mcnt, p + 2);
|
DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt);
|
|
|
assert (mcnt >= 0);
|
assert (mcnt >= 0);
|
/* Originally, this is how many times we HAVE to succeed. */
|
/* Originally, this is how many times we HAVE to succeed. */
|
if (mcnt > 0)
|
if (mcnt > 0)
|
{
|
{
|
mcnt--;
|
mcnt--;
|
p += 2;
|
p += 2;
|
STORE_NUMBER_AND_INCR (p, mcnt);
|
STORE_NUMBER_AND_INCR (p, mcnt);
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p - 2, mcnt);
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p - 2, mcnt);
|
#else
|
#else
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p - 2, mcnt);
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p - 2, mcnt);
|
#endif
|
#endif
|
}
|
}
|
else if (mcnt == 0)
|
else if (mcnt == 0)
|
{
|
{
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT2 (" Setting two bytes from %p to no_op.\n", p+2);
|
DEBUG_PRINT2 (" Setting two bytes from %p to no_op.\n", p+2);
|
#else
|
#else
|
DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n", p+2);
|
DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n", p+2);
|
#endif
|
#endif
|
p[2] = (unsigned char) no_op;
|
p[2] = (unsigned char) no_op;
|
p[3] = (unsigned char) no_op;
|
p[3] = (unsigned char) no_op;
|
goto on_failure;
|
goto on_failure;
|
}
|
}
|
break;
|
break;
|
|
|
case jump_n:
|
case jump_n:
|
EXTRACT_NUMBER (mcnt, p + 2);
|
EXTRACT_NUMBER (mcnt, p + 2);
|
DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt);
|
|
|
/* Originally, this is how many times we CAN jump. */
|
/* Originally, this is how many times we CAN jump. */
|
if (mcnt)
|
if (mcnt)
|
{
|
{
|
mcnt--;
|
mcnt--;
|
STORE_NUMBER (p + 2, mcnt);
|
STORE_NUMBER (p + 2, mcnt);
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p + 2, mcnt);
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p + 2, mcnt);
|
#else
|
#else
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p + 2, mcnt);
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p + 2, mcnt);
|
#endif
|
#endif
|
goto unconditional_jump;
|
goto unconditional_jump;
|
}
|
}
|
/* If don't have to jump any more, skip over the rest of command. */
|
/* If don't have to jump any more, skip over the rest of command. */
|
else
|
else
|
p += 4;
|
p += 4;
|
break;
|
break;
|
|
|
case set_number_at:
|
case set_number_at:
|
{
|
{
|
DEBUG_PRINT1 ("EXECUTING set_number_at.\n");
|
DEBUG_PRINT1 ("EXECUTING set_number_at.\n");
|
|
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
p1 = p + mcnt;
|
p1 = p + mcnt;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p);
|
#ifdef _LIBC
|
#ifdef _LIBC
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p1, mcnt);
|
DEBUG_PRINT3 (" Setting %p to %d.\n", p1, mcnt);
|
#else
|
#else
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt);
|
DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt);
|
#endif
|
#endif
|
STORE_NUMBER (p1, mcnt);
|
STORE_NUMBER (p1, mcnt);
|
break;
|
break;
|
}
|
}
|
|
|
#if 0
|
#if 0
|
/* The DEC Alpha C compiler 3.x generates incorrect code for the
|
/* The DEC Alpha C compiler 3.x generates incorrect code for the
|
test WORDCHAR_P (d - 1) != WORDCHAR_P (d) in the expansion of
|
test WORDCHAR_P (d - 1) != WORDCHAR_P (d) in the expansion of
|
AT_WORD_BOUNDARY, so this code is disabled. Expanding the
|
AT_WORD_BOUNDARY, so this code is disabled. Expanding the
|
macro and introducing temporary variables works around the bug. */
|
macro and introducing temporary variables works around the bug. */
|
|
|
case wordbound:
|
case wordbound:
|
DEBUG_PRINT1 ("EXECUTING wordbound.\n");
|
DEBUG_PRINT1 ("EXECUTING wordbound.\n");
|
if (AT_WORD_BOUNDARY (d))
|
if (AT_WORD_BOUNDARY (d))
|
break;
|
break;
|
goto fail;
|
goto fail;
|
|
|
case notwordbound:
|
case notwordbound:
|
DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
|
DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
|
if (AT_WORD_BOUNDARY (d))
|
if (AT_WORD_BOUNDARY (d))
|
goto fail;
|
goto fail;
|
break;
|
break;
|
#else
|
#else
|
case wordbound:
|
case wordbound:
|
{
|
{
|
boolean prevchar, thischar;
|
boolean prevchar, thischar;
|
|
|
DEBUG_PRINT1 ("EXECUTING wordbound.\n");
|
DEBUG_PRINT1 ("EXECUTING wordbound.\n");
|
if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
|
if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
|
break;
|
break;
|
|
|
prevchar = WORDCHAR_P (d - 1);
|
prevchar = WORDCHAR_P (d - 1);
|
thischar = WORDCHAR_P (d);
|
thischar = WORDCHAR_P (d);
|
if (prevchar != thischar)
|
if (prevchar != thischar)
|
break;
|
break;
|
goto fail;
|
goto fail;
|
}
|
}
|
|
|
case notwordbound:
|
case notwordbound:
|
{
|
{
|
boolean prevchar, thischar;
|
boolean prevchar, thischar;
|
|
|
DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
|
DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
|
if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
|
if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
|
goto fail;
|
goto fail;
|
|
|
prevchar = WORDCHAR_P (d - 1);
|
prevchar = WORDCHAR_P (d - 1);
|
thischar = WORDCHAR_P (d);
|
thischar = WORDCHAR_P (d);
|
if (prevchar != thischar)
|
if (prevchar != thischar)
|
goto fail;
|
goto fail;
|
break;
|
break;
|
}
|
}
|
#endif
|
#endif
|
|
|
case wordbeg:
|
case wordbeg:
|
DEBUG_PRINT1 ("EXECUTING wordbeg.\n");
|
DEBUG_PRINT1 ("EXECUTING wordbeg.\n");
|
if (WORDCHAR_P (d) && (AT_STRINGS_BEG (d) || !WORDCHAR_P (d - 1)))
|
if (WORDCHAR_P (d) && (AT_STRINGS_BEG (d) || !WORDCHAR_P (d - 1)))
|
break;
|
break;
|
goto fail;
|
goto fail;
|
|
|
case wordend:
|
case wordend:
|
DEBUG_PRINT1 ("EXECUTING wordend.\n");
|
DEBUG_PRINT1 ("EXECUTING wordend.\n");
|
if (!AT_STRINGS_BEG (d) && WORDCHAR_P (d - 1)
|
if (!AT_STRINGS_BEG (d) && WORDCHAR_P (d - 1)
|
&& (!WORDCHAR_P (d) || AT_STRINGS_END (d)))
|
&& (!WORDCHAR_P (d) || AT_STRINGS_END (d)))
|
break;
|
break;
|
goto fail;
|
goto fail;
|
|
|
#ifdef emacs
|
#ifdef emacs
|
case before_dot:
|
case before_dot:
|
DEBUG_PRINT1 ("EXECUTING before_dot.\n");
|
DEBUG_PRINT1 ("EXECUTING before_dot.\n");
|
if (PTR_CHAR_POS ((unsigned char *) d) >= point)
|
if (PTR_CHAR_POS ((unsigned char *) d) >= point)
|
goto fail;
|
goto fail;
|
break;
|
break;
|
|
|
case at_dot:
|
case at_dot:
|
DEBUG_PRINT1 ("EXECUTING at_dot.\n");
|
DEBUG_PRINT1 ("EXECUTING at_dot.\n");
|
if (PTR_CHAR_POS ((unsigned char *) d) != point)
|
if (PTR_CHAR_POS ((unsigned char *) d) != point)
|
goto fail;
|
goto fail;
|
break;
|
break;
|
|
|
case after_dot:
|
case after_dot:
|
DEBUG_PRINT1 ("EXECUTING after_dot.\n");
|
DEBUG_PRINT1 ("EXECUTING after_dot.\n");
|
if (PTR_CHAR_POS ((unsigned char *) d) <= point)
|
if (PTR_CHAR_POS ((unsigned char *) d) <= point)
|
goto fail;
|
goto fail;
|
break;
|
break;
|
|
|
case syntaxspec:
|
case syntaxspec:
|
DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt);
|
mcnt = *p++;
|
mcnt = *p++;
|
goto matchsyntax;
|
goto matchsyntax;
|
|
|
case wordchar:
|
case wordchar:
|
DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n");
|
DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n");
|
mcnt = (int) Sword;
|
mcnt = (int) Sword;
|
matchsyntax:
|
matchsyntax:
|
PREFETCH ();
|
PREFETCH ();
|
/* Can't use *d++ here; SYNTAX may be an unsafe macro. */
|
/* Can't use *d++ here; SYNTAX may be an unsafe macro. */
|
d++;
|
d++;
|
if (SYNTAX (d[-1]) != (enum syntaxcode) mcnt)
|
if (SYNTAX (d[-1]) != (enum syntaxcode) mcnt)
|
goto fail;
|
goto fail;
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
break;
|
break;
|
|
|
case notsyntaxspec:
|
case notsyntaxspec:
|
DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt);
|
DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt);
|
mcnt = *p++;
|
mcnt = *p++;
|
goto matchnotsyntax;
|
goto matchnotsyntax;
|
|
|
case notwordchar:
|
case notwordchar:
|
DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n");
|
DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n");
|
mcnt = (int) Sword;
|
mcnt = (int) Sword;
|
matchnotsyntax:
|
matchnotsyntax:
|
PREFETCH ();
|
PREFETCH ();
|
/* Can't use *d++ here; SYNTAX may be an unsafe macro. */
|
/* Can't use *d++ here; SYNTAX may be an unsafe macro. */
|
d++;
|
d++;
|
if (SYNTAX (d[-1]) == (enum syntaxcode) mcnt)
|
if (SYNTAX (d[-1]) == (enum syntaxcode) mcnt)
|
goto fail;
|
goto fail;
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
break;
|
break;
|
|
|
#else /* not emacs */
|
#else /* not emacs */
|
case wordchar:
|
case wordchar:
|
DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n");
|
DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n");
|
PREFETCH ();
|
PREFETCH ();
|
if (!WORDCHAR_P (d))
|
if (!WORDCHAR_P (d))
|
goto fail;
|
goto fail;
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
d++;
|
d++;
|
break;
|
break;
|
|
|
case notwordchar:
|
case notwordchar:
|
DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n");
|
DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n");
|
PREFETCH ();
|
PREFETCH ();
|
if (WORDCHAR_P (d))
|
if (WORDCHAR_P (d))
|
goto fail;
|
goto fail;
|
SET_REGS_MATCHED ();
|
SET_REGS_MATCHED ();
|
d++;
|
d++;
|
break;
|
break;
|
#endif /* not emacs */
|
#endif /* not emacs */
|
|
|
default:
|
default:
|
abort ();
|
abort ();
|
}
|
}
|
continue; /* Successfully executed one pattern command; keep going. */
|
continue; /* Successfully executed one pattern command; keep going. */
|
|
|
|
|
/* We goto here if a matching operation fails. */
|
/* We goto here if a matching operation fails. */
|
fail:
|
fail:
|
if (!FAIL_STACK_EMPTY ())
|
if (!FAIL_STACK_EMPTY ())
|
{ /* A restart point is known. Restore to that state. */
|
{ /* A restart point is known. Restore to that state. */
|
DEBUG_PRINT1 ("\nFAIL:\n");
|
DEBUG_PRINT1 ("\nFAIL:\n");
|
POP_FAILURE_POINT (d, p,
|
POP_FAILURE_POINT (d, p,
|
lowest_active_reg, highest_active_reg,
|
lowest_active_reg, highest_active_reg,
|
regstart, regend, reg_info);
|
regstart, regend, reg_info);
|
|
|
/* If this failure point is a dummy, try the next one. */
|
/* If this failure point is a dummy, try the next one. */
|
if (!p)
|
if (!p)
|
goto fail;
|
goto fail;
|
|
|
/* If we failed to the end of the pattern, don't examine *p. */
|
/* If we failed to the end of the pattern, don't examine *p. */
|
assert (p <= pend);
|
assert (p <= pend);
|
if (p < pend)
|
if (p < pend)
|
{
|
{
|
boolean is_a_jump_n = false;
|
boolean is_a_jump_n = false;
|
|
|
/* If failed to a backwards jump that's part of a repetition
|
/* If failed to a backwards jump that's part of a repetition
|
loop, need to pop this failure point and use the next one. */
|
loop, need to pop this failure point and use the next one. */
|
switch ((re_opcode_t) *p)
|
switch ((re_opcode_t) *p)
|
{
|
{
|
case jump_n:
|
case jump_n:
|
is_a_jump_n = true;
|
is_a_jump_n = true;
|
case maybe_pop_jump:
|
case maybe_pop_jump:
|
case pop_failure_jump:
|
case pop_failure_jump:
|
case jump:
|
case jump:
|
p1 = p + 1;
|
p1 = p + 1;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
p1 += mcnt;
|
p1 += mcnt;
|
|
|
if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n)
|
if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n)
|
|| (!is_a_jump_n
|
|| (!is_a_jump_n
|
&& (re_opcode_t) *p1 == on_failure_jump))
|
&& (re_opcode_t) *p1 == on_failure_jump))
|
goto fail;
|
goto fail;
|
break;
|
break;
|
default:
|
default:
|
/* do nothing */ ;
|
/* do nothing */ ;
|
}
|
}
|
}
|
}
|
|
|
if (d >= string1 && d <= end1)
|
if (d >= string1 && d <= end1)
|
dend = end_match_1;
|
dend = end_match_1;
|
}
|
}
|
else
|
else
|
break; /* Matching at this starting point really fails. */
|
break; /* Matching at this starting point really fails. */
|
} /* for (;;) */
|
} /* for (;;) */
|
|
|
if (best_regs_set)
|
if (best_regs_set)
|
goto restore_best_regs;
|
goto restore_best_regs;
|
|
|
FREE_VARIABLES ();
|
FREE_VARIABLES ();
|
|
|
return -1; /* Failure to match. */
|
return -1; /* Failure to match. */
|
} /* re_match_2 */
|
} /* re_match_2 */
|
|
|
/* Subroutine definitions for re_match_2. */
|
/* Subroutine definitions for re_match_2. */
|
|
|
|
|
/* We are passed P pointing to a register number after a start_memory.
|
/* We are passed P pointing to a register number after a start_memory.
|
|
|
Return true if the pattern up to the corresponding stop_memory can
|
Return true if the pattern up to the corresponding stop_memory can
|
match the empty string, and false otherwise.
|
match the empty string, and false otherwise.
|
|
|
If we find the matching stop_memory, sets P to point to one past its number.
|
If we find the matching stop_memory, sets P to point to one past its number.
|
Otherwise, sets P to an undefined byte less than or equal to END.
|
Otherwise, sets P to an undefined byte less than or equal to END.
|
|
|
We don't handle duplicates properly (yet). */
|
We don't handle duplicates properly (yet). */
|
|
|
static boolean
|
static boolean
|
group_match_null_string_p (p, end, reg_info)
|
group_match_null_string_p (p, end, reg_info)
|
unsigned char **p, *end;
|
unsigned char **p, *end;
|
register_info_type *reg_info;
|
register_info_type *reg_info;
|
{
|
{
|
int mcnt;
|
int mcnt;
|
/* Point to after the args to the start_memory. */
|
/* Point to after the args to the start_memory. */
|
unsigned char *p1 = *p + 2;
|
unsigned char *p1 = *p + 2;
|
|
|
while (p1 < end)
|
while (p1 < end)
|
{
|
{
|
/* Skip over opcodes that can match nothing, and return true or
|
/* Skip over opcodes that can match nothing, and return true or
|
false, as appropriate, when we get to one that can't, or to the
|
false, as appropriate, when we get to one that can't, or to the
|
matching stop_memory. */
|
matching stop_memory. */
|
|
|
switch ((re_opcode_t) *p1)
|
switch ((re_opcode_t) *p1)
|
{
|
{
|
/* Could be either a loop or a series of alternatives. */
|
/* Could be either a loop or a series of alternatives. */
|
case on_failure_jump:
|
case on_failure_jump:
|
p1++;
|
p1++;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
|
|
/* If the next operation is not a jump backwards in the
|
/* If the next operation is not a jump backwards in the
|
pattern. */
|
pattern. */
|
|
|
if (mcnt >= 0)
|
if (mcnt >= 0)
|
{
|
{
|
/* Go through the on_failure_jumps of the alternatives,
|
/* Go through the on_failure_jumps of the alternatives,
|
seeing if any of the alternatives cannot match nothing.
|
seeing if any of the alternatives cannot match nothing.
|
The last alternative starts with only a jump,
|
The last alternative starts with only a jump,
|
whereas the rest start with on_failure_jump and end
|
whereas the rest start with on_failure_jump and end
|
with a jump, e.g., here is the pattern for `a|b|c':
|
with a jump, e.g., here is the pattern for `a|b|c':
|
|
|
/on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6
|
/on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6
|
/on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3
|
/on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3
|
/exactn/1/c
|
/exactn/1/c
|
|
|
So, we have to first go through the first (n-1)
|
So, we have to first go through the first (n-1)
|
alternatives and then deal with the last one separately. */
|
alternatives and then deal with the last one separately. */
|
|
|
|
|
/* Deal with the first (n-1) alternatives, which start
|
/* Deal with the first (n-1) alternatives, which start
|
with an on_failure_jump (see above) that jumps to right
|
with an on_failure_jump (see above) that jumps to right
|
past a jump_past_alt. */
|
past a jump_past_alt. */
|
|
|
while ((re_opcode_t) p1[mcnt-3] == jump_past_alt)
|
while ((re_opcode_t) p1[mcnt-3] == jump_past_alt)
|
{
|
{
|
/* `mcnt' holds how many bytes long the alternative
|
/* `mcnt' holds how many bytes long the alternative
|
is, including the ending `jump_past_alt' and
|
is, including the ending `jump_past_alt' and
|
its number. */
|
its number. */
|
|
|
if (!alt_match_null_string_p (p1, p1 + mcnt - 3,
|
if (!alt_match_null_string_p (p1, p1 + mcnt - 3,
|
reg_info))
|
reg_info))
|
return false;
|
return false;
|
|
|
/* Move to right after this alternative, including the
|
/* Move to right after this alternative, including the
|
jump_past_alt. */
|
jump_past_alt. */
|
p1 += mcnt;
|
p1 += mcnt;
|
|
|
/* Break if it's the beginning of an n-th alternative
|
/* Break if it's the beginning of an n-th alternative
|
that doesn't begin with an on_failure_jump. */
|
that doesn't begin with an on_failure_jump. */
|
if ((re_opcode_t) *p1 != on_failure_jump)
|
if ((re_opcode_t) *p1 != on_failure_jump)
|
break;
|
break;
|
|
|
/* Still have to check that it's not an n-th
|
/* Still have to check that it's not an n-th
|
alternative that starts with an on_failure_jump. */
|
alternative that starts with an on_failure_jump. */
|
p1++;
|
p1++;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
if ((re_opcode_t) p1[mcnt-3] != jump_past_alt)
|
if ((re_opcode_t) p1[mcnt-3] != jump_past_alt)
|
{
|
{
|
/* Get to the beginning of the n-th alternative. */
|
/* Get to the beginning of the n-th alternative. */
|
p1 -= 3;
|
p1 -= 3;
|
break;
|
break;
|
}
|
}
|
}
|
}
|
|
|
/* Deal with the last alternative: go back and get number
|
/* Deal with the last alternative: go back and get number
|
of the `jump_past_alt' just before it. `mcnt' contains
|
of the `jump_past_alt' just before it. `mcnt' contains
|
the length of the alternative. */
|
the length of the alternative. */
|
EXTRACT_NUMBER (mcnt, p1 - 2);
|
EXTRACT_NUMBER (mcnt, p1 - 2);
|
|
|
if (!alt_match_null_string_p (p1, p1 + mcnt, reg_info))
|
if (!alt_match_null_string_p (p1, p1 + mcnt, reg_info))
|
return false;
|
return false;
|
|
|
p1 += mcnt; /* Get past the n-th alternative. */
|
p1 += mcnt; /* Get past the n-th alternative. */
|
} /* if mcnt > 0 */
|
} /* if mcnt > 0 */
|
break;
|
break;
|
|
|
|
|
case stop_memory:
|
case stop_memory:
|
assert (p1[1] == **p);
|
assert (p1[1] == **p);
|
*p = p1 + 2;
|
*p = p1 + 2;
|
return true;
|
return true;
|
|
|
|
|
default:
|
default:
|
if (!common_op_match_null_string_p (&p1, end, reg_info))
|
if (!common_op_match_null_string_p (&p1, end, reg_info))
|
return false;
|
return false;
|
}
|
}
|
} /* while p1 < end */
|
} /* while p1 < end */
|
|
|
return false;
|
return false;
|
} /* group_match_null_string_p */
|
} /* group_match_null_string_p */
|
|
|
|
|
/* Similar to group_match_null_string_p, but doesn't deal with alternatives:
|
/* Similar to group_match_null_string_p, but doesn't deal with alternatives:
|
It expects P to be the first byte of a single alternative and END one
|
It expects P to be the first byte of a single alternative and END one
|
byte past the last. The alternative can contain groups. */
|
byte past the last. The alternative can contain groups. */
|
|
|
static boolean
|
static boolean
|
alt_match_null_string_p (p, end, reg_info)
|
alt_match_null_string_p (p, end, reg_info)
|
unsigned char *p, *end;
|
unsigned char *p, *end;
|
register_info_type *reg_info;
|
register_info_type *reg_info;
|
{
|
{
|
int mcnt;
|
int mcnt;
|
unsigned char *p1 = p;
|
unsigned char *p1 = p;
|
|
|
while (p1 < end)
|
while (p1 < end)
|
{
|
{
|
/* Skip over opcodes that can match nothing, and break when we get
|
/* Skip over opcodes that can match nothing, and break when we get
|
to one that can't. */
|
to one that can't. */
|
|
|
switch ((re_opcode_t) *p1)
|
switch ((re_opcode_t) *p1)
|
{
|
{
|
/* It's a loop. */
|
/* It's a loop. */
|
case on_failure_jump:
|
case on_failure_jump:
|
p1++;
|
p1++;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
p1 += mcnt;
|
p1 += mcnt;
|
break;
|
break;
|
|
|
default:
|
default:
|
if (!common_op_match_null_string_p (&p1, end, reg_info))
|
if (!common_op_match_null_string_p (&p1, end, reg_info))
|
return false;
|
return false;
|
}
|
}
|
} /* while p1 < end */
|
} /* while p1 < end */
|
|
|
return true;
|
return true;
|
} /* alt_match_null_string_p */
|
} /* alt_match_null_string_p */
|
|
|
|
|
/* Deals with the ops common to group_match_null_string_p and
|
/* Deals with the ops common to group_match_null_string_p and
|
alt_match_null_string_p.
|
alt_match_null_string_p.
|
|
|
Sets P to one after the op and its arguments, if any. */
|
Sets P to one after the op and its arguments, if any. */
|
|
|
static boolean
|
static boolean
|
common_op_match_null_string_p (p, end, reg_info)
|
common_op_match_null_string_p (p, end, reg_info)
|
unsigned char **p, *end;
|
unsigned char **p, *end;
|
register_info_type *reg_info;
|
register_info_type *reg_info;
|
{
|
{
|
int mcnt;
|
int mcnt;
|
boolean ret;
|
boolean ret;
|
int reg_no;
|
int reg_no;
|
unsigned char *p1 = *p;
|
unsigned char *p1 = *p;
|
|
|
switch ((re_opcode_t) *p1++)
|
switch ((re_opcode_t) *p1++)
|
{
|
{
|
case no_op:
|
case no_op:
|
case begline:
|
case begline:
|
case endline:
|
case endline:
|
case begbuf:
|
case begbuf:
|
case endbuf:
|
case endbuf:
|
case wordbeg:
|
case wordbeg:
|
case wordend:
|
case wordend:
|
case wordbound:
|
case wordbound:
|
case notwordbound:
|
case notwordbound:
|
#ifdef emacs
|
#ifdef emacs
|
case before_dot:
|
case before_dot:
|
case at_dot:
|
case at_dot:
|
case after_dot:
|
case after_dot:
|
#endif
|
#endif
|
break;
|
break;
|
|
|
case start_memory:
|
case start_memory:
|
reg_no = *p1;
|
reg_no = *p1;
|
assert (reg_no > 0 && reg_no <= MAX_REGNUM);
|
assert (reg_no > 0 && reg_no <= MAX_REGNUM);
|
ret = group_match_null_string_p (&p1, end, reg_info);
|
ret = group_match_null_string_p (&p1, end, reg_info);
|
|
|
/* Have to set this here in case we're checking a group which
|
/* Have to set this here in case we're checking a group which
|
contains a group and a back reference to it. */
|
contains a group and a back reference to it. */
|
|
|
if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE)
|
if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE)
|
REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret;
|
REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret;
|
|
|
if (!ret)
|
if (!ret)
|
return false;
|
return false;
|
break;
|
break;
|
|
|
/* If this is an optimized succeed_n for zero times, make the jump. */
|
/* If this is an optimized succeed_n for zero times, make the jump. */
|
case jump:
|
case jump:
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
if (mcnt >= 0)
|
if (mcnt >= 0)
|
p1 += mcnt;
|
p1 += mcnt;
|
else
|
else
|
return false;
|
return false;
|
break;
|
break;
|
|
|
case succeed_n:
|
case succeed_n:
|
/* Get to the number of times to succeed. */
|
/* Get to the number of times to succeed. */
|
p1 += 2;
|
p1 += 2;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
|
|
if (mcnt == 0)
|
if (mcnt == 0)
|
{
|
{
|
p1 -= 4;
|
p1 -= 4;
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
EXTRACT_NUMBER_AND_INCR (mcnt, p1);
|
p1 += mcnt;
|
p1 += mcnt;
|
}
|
}
|
else
|
else
|
return false;
|
return false;
|
break;
|
break;
|
|
|
case duplicate:
|
case duplicate:
|
if (!REG_MATCH_NULL_STRING_P (reg_info[*p1]))
|
if (!REG_MATCH_NULL_STRING_P (reg_info[*p1]))
|
return false;
|
return false;
|
break;
|
break;
|
|
|
case set_number_at:
|
case set_number_at:
|
p1 += 4;
|
p1 += 4;
|
|
|
default:
|
default:
|
/* All other opcodes mean we cannot match the empty string. */
|
/* All other opcodes mean we cannot match the empty string. */
|
return false;
|
return false;
|
}
|
}
|
|
|
*p = p1;
|
*p = p1;
|
return true;
|
return true;
|
} /* common_op_match_null_string_p */
|
} /* common_op_match_null_string_p */
|
|
|
|
|
/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN
|
/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN
|
bytes; nonzero otherwise. */
|
bytes; nonzero otherwise. */
|
|
|
static int
|
static int
|
bcmp_translate (s1, s2, len, translate)
|
bcmp_translate (s1, s2, len, translate)
|
const char *s1, *s2;
|
const char *s1, *s2;
|
register int len;
|
register int len;
|
RE_TRANSLATE_TYPE translate;
|
RE_TRANSLATE_TYPE translate;
|
{
|
{
|
register const unsigned char *p1 = (const unsigned char *) s1;
|
register const unsigned char *p1 = (const unsigned char *) s1;
|
register const unsigned char *p2 = (const unsigned char *) s2;
|
register const unsigned char *p2 = (const unsigned char *) s2;
|
while (len)
|
while (len)
|
{
|
{
|
if (translate[*p1++] != translate[*p2++]) return 1;
|
if (translate[*p1++] != translate[*p2++]) return 1;
|
len--;
|
len--;
|
}
|
}
|
return 0;
|
return 0;
|
}
|
}
|
|
|
/* Entry points for GNU code. */
|
/* Entry points for GNU code. */
|
|
|
/* re_compile_pattern is the GNU regular expression compiler: it
|
/* re_compile_pattern is the GNU regular expression compiler: it
|
compiles PATTERN (of length SIZE) and puts the result in BUFP.
|
compiles PATTERN (of length SIZE) and puts the result in BUFP.
|
Returns 0 if the pattern was valid, otherwise an error string.
|
Returns 0 if the pattern was valid, otherwise an error string.
|
|
|
Assumes the `allocated' (and perhaps `buffer') and `translate' fields
|
Assumes the `allocated' (and perhaps `buffer') and `translate' fields
|
are set in BUFP on entry.
|
are set in BUFP on entry.
|
|
|
We call regex_compile to do the actual compilation. */
|
We call regex_compile to do the actual compilation. */
|
|
|
const char *
|
const char *
|
re_compile_pattern (pattern, length, bufp)
|
re_compile_pattern (pattern, length, bufp)
|
const char *pattern;
|
const char *pattern;
|
size_t length;
|
size_t length;
|
struct re_pattern_buffer *bufp;
|
struct re_pattern_buffer *bufp;
|
{
|
{
|
reg_errcode_t ret;
|
reg_errcode_t ret;
|
|
|
/* GNU code is written to assume at least RE_NREGS registers will be set
|
/* GNU code is written to assume at least RE_NREGS registers will be set
|
(and at least one extra will be -1). */
|
(and at least one extra will be -1). */
|
bufp->regs_allocated = REGS_UNALLOCATED;
|
bufp->regs_allocated = REGS_UNALLOCATED;
|
|
|
/* And GNU code determines whether or not to get register information
|
/* And GNU code determines whether or not to get register information
|
by passing null for the REGS argument to re_match, etc., not by
|
by passing null for the REGS argument to re_match, etc., not by
|
setting no_sub. */
|
setting no_sub. */
|
bufp->no_sub = 0;
|
bufp->no_sub = 0;
|
|
|
/* Match anchors at newline. */
|
/* Match anchors at newline. */
|
bufp->newline_anchor = 1;
|
bufp->newline_anchor = 1;
|
|
|
ret = regex_compile (pattern, length, re_syntax_options, bufp);
|
ret = regex_compile (pattern, length, re_syntax_options, bufp);
|
|
|
if (!ret)
|
if (!ret)
|
return NULL;
|
return NULL;
|
return gettext (re_error_msgid[(int) ret]);
|
return gettext (re_error_msgid[(int) ret]);
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__re_compile_pattern, re_compile_pattern)
|
weak_alias (__re_compile_pattern, re_compile_pattern)
|
#endif
|
#endif
|
|
|
/* Entry points compatible with 4.2 BSD regex library. We don't define
|
/* Entry points compatible with 4.2 BSD regex library. We don't define
|
them unless specifically requested. */
|
them unless specifically requested. */
|
|
|
#if defined _REGEX_RE_COMP || defined _LIBC
|
#if defined _REGEX_RE_COMP || defined _LIBC
|
|
|
/* BSD has one and only one pattern buffer. */
|
/* BSD has one and only one pattern buffer. */
|
static struct re_pattern_buffer re_comp_buf;
|
static struct re_pattern_buffer re_comp_buf;
|
|
|
char *
|
char *
|
#ifdef _LIBC
|
#ifdef _LIBC
|
/* Make these definitions weak in libc, so POSIX programs can redefine
|
/* Make these definitions weak in libc, so POSIX programs can redefine
|
these names if they don't use our functions, and still use
|
these names if they don't use our functions, and still use
|
regcomp/regexec below without link errors. */
|
regcomp/regexec below without link errors. */
|
weak_function
|
weak_function
|
#endif
|
#endif
|
re_comp (s)
|
re_comp (s)
|
const char *s;
|
const char *s;
|
{
|
{
|
reg_errcode_t ret;
|
reg_errcode_t ret;
|
|
|
if (!s)
|
if (!s)
|
{
|
{
|
if (!re_comp_buf.buffer)
|
if (!re_comp_buf.buffer)
|
return gettext ("No previous regular expression");
|
return gettext ("No previous regular expression");
|
return 0;
|
return 0;
|
}
|
}
|
|
|
if (!re_comp_buf.buffer)
|
if (!re_comp_buf.buffer)
|
{
|
{
|
re_comp_buf.buffer = (unsigned char *) malloc (200);
|
re_comp_buf.buffer = (unsigned char *) malloc (200);
|
if (re_comp_buf.buffer == NULL)
|
if (re_comp_buf.buffer == NULL)
|
return (char *) gettext (re_error_msgid[(int) REG_ESPACE]);
|
return (char *) gettext (re_error_msgid[(int) REG_ESPACE]);
|
re_comp_buf.allocated = 200;
|
re_comp_buf.allocated = 200;
|
|
|
re_comp_buf.fastmap = (char *) malloc (1 << BYTEWIDTH);
|
re_comp_buf.fastmap = (char *) malloc (1 << BYTEWIDTH);
|
if (re_comp_buf.fastmap == NULL)
|
if (re_comp_buf.fastmap == NULL)
|
return (char *) gettext (re_error_msgid[(int) REG_ESPACE]);
|
return (char *) gettext (re_error_msgid[(int) REG_ESPACE]);
|
}
|
}
|
|
|
/* Since `re_exec' always passes NULL for the `regs' argument, we
|
/* Since `re_exec' always passes NULL for the `regs' argument, we
|
don't need to initialize the pattern buffer fields which affect it. */
|
don't need to initialize the pattern buffer fields which affect it. */
|
|
|
/* Match anchors at newlines. */
|
/* Match anchors at newlines. */
|
re_comp_buf.newline_anchor = 1;
|
re_comp_buf.newline_anchor = 1;
|
|
|
ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf);
|
ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf);
|
|
|
if (!ret)
|
if (!ret)
|
return NULL;
|
return NULL;
|
|
|
/* Yes, we're discarding `const' here if !HAVE_LIBINTL. */
|
/* Yes, we're discarding `const' here if !HAVE_LIBINTL. */
|
return (char *) gettext (re_error_msgid[(int) ret]);
|
return (char *) gettext (re_error_msgid[(int) ret]);
|
}
|
}
|
|
|
|
|
int
|
int
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_function
|
weak_function
|
#endif
|
#endif
|
re_exec (s)
|
re_exec (s)
|
const char *s;
|
const char *s;
|
{
|
{
|
const int len = strlen (s);
|
const int len = strlen (s);
|
return
|
return
|
0 <= re_search (&re_comp_buf, s, len, 0, len, (struct re_registers *) 0);
|
0 <= re_search (&re_comp_buf, s, len, 0, len, (struct re_registers *) 0);
|
}
|
}
|
|
|
#endif /* _REGEX_RE_COMP */
|
#endif /* _REGEX_RE_COMP */
|
|
|
/* POSIX.2 functions. Don't define these for Emacs. */
|
/* POSIX.2 functions. Don't define these for Emacs. */
|
|
|
#ifndef emacs
|
#ifndef emacs
|
|
|
/* regcomp takes a regular expression as a string and compiles it.
|
/* regcomp takes a regular expression as a string and compiles it.
|
|
|
PREG is a regex_t *. We do not expect any fields to be initialized,
|
PREG is a regex_t *. We do not expect any fields to be initialized,
|
since POSIX says we shouldn't. Thus, we set
|
since POSIX says we shouldn't. Thus, we set
|
|
|
`buffer' to the compiled pattern;
|
`buffer' to the compiled pattern;
|
`used' to the length of the compiled pattern;
|
`used' to the length of the compiled pattern;
|
`syntax' to RE_SYNTAX_POSIX_EXTENDED if the
|
`syntax' to RE_SYNTAX_POSIX_EXTENDED if the
|
REG_EXTENDED bit in CFLAGS is set; otherwise, to
|
REG_EXTENDED bit in CFLAGS is set; otherwise, to
|
RE_SYNTAX_POSIX_BASIC;
|
RE_SYNTAX_POSIX_BASIC;
|
`newline_anchor' to REG_NEWLINE being set in CFLAGS;
|
`newline_anchor' to REG_NEWLINE being set in CFLAGS;
|
`fastmap' and `fastmap_accurate' to zero;
|
`fastmap' and `fastmap_accurate' to zero;
|
`re_nsub' to the number of subexpressions in PATTERN.
|
`re_nsub' to the number of subexpressions in PATTERN.
|
|
|
PATTERN is the address of the pattern string.
|
PATTERN is the address of the pattern string.
|
|
|
CFLAGS is a series of bits which affect compilation.
|
CFLAGS is a series of bits which affect compilation.
|
|
|
If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we
|
If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we
|
use POSIX basic syntax.
|
use POSIX basic syntax.
|
|
|
If REG_NEWLINE is set, then . and [^...] don't match newline.
|
If REG_NEWLINE is set, then . and [^...] don't match newline.
|
Also, regexec will try a match beginning after every newline.
|
Also, regexec will try a match beginning after every newline.
|
|
|
If REG_ICASE is set, then we considers upper- and lowercase
|
If REG_ICASE is set, then we considers upper- and lowercase
|
versions of letters to be equivalent when matching.
|
versions of letters to be equivalent when matching.
|
|
|
If REG_NOSUB is set, then when PREG is passed to regexec, that
|
If REG_NOSUB is set, then when PREG is passed to regexec, that
|
routine will report only success or failure, and nothing about the
|
routine will report only success or failure, and nothing about the
|
registers.
|
registers.
|
|
|
It returns 0 if it succeeds, nonzero if it doesn't. (See gnu-regex.h for
|
It returns 0 if it succeeds, nonzero if it doesn't. (See gnu-regex.h for
|
the return codes and their meanings.) */
|
the return codes and their meanings.) */
|
|
|
int
|
int
|
regcomp (preg, pattern, cflags)
|
regcomp (preg, pattern, cflags)
|
regex_t *preg;
|
regex_t *preg;
|
const char *pattern;
|
const char *pattern;
|
int cflags;
|
int cflags;
|
{
|
{
|
reg_errcode_t ret;
|
reg_errcode_t ret;
|
reg_syntax_t syntax
|
reg_syntax_t syntax
|
= (cflags & REG_EXTENDED) ?
|
= (cflags & REG_EXTENDED) ?
|
RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC;
|
RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC;
|
|
|
/* regex_compile will allocate the space for the compiled pattern. */
|
/* regex_compile will allocate the space for the compiled pattern. */
|
preg->buffer = 0;
|
preg->buffer = 0;
|
preg->allocated = 0;
|
preg->allocated = 0;
|
preg->used = 0;
|
preg->used = 0;
|
|
|
/* Don't bother to use a fastmap when searching. This simplifies the
|
/* Don't bother to use a fastmap when searching. This simplifies the
|
REG_NEWLINE case: if we used a fastmap, we'd have to put all the
|
REG_NEWLINE case: if we used a fastmap, we'd have to put all the
|
characters after newlines into the fastmap. This way, we just try
|
characters after newlines into the fastmap. This way, we just try
|
every character. */
|
every character. */
|
preg->fastmap = 0;
|
preg->fastmap = 0;
|
|
|
if (cflags & REG_ICASE)
|
if (cflags & REG_ICASE)
|
{
|
{
|
unsigned i;
|
unsigned i;
|
|
|
preg->translate
|
preg->translate
|
= (RE_TRANSLATE_TYPE) malloc (CHAR_SET_SIZE
|
= (RE_TRANSLATE_TYPE) malloc (CHAR_SET_SIZE
|
* sizeof (*(RE_TRANSLATE_TYPE)0));
|
* sizeof (*(RE_TRANSLATE_TYPE)0));
|
if (preg->translate == NULL)
|
if (preg->translate == NULL)
|
return (int) REG_ESPACE;
|
return (int) REG_ESPACE;
|
|
|
/* Map uppercase characters to corresponding lowercase ones. */
|
/* Map uppercase characters to corresponding lowercase ones. */
|
for (i = 0; i < CHAR_SET_SIZE; i++)
|
for (i = 0; i < CHAR_SET_SIZE; i++)
|
preg->translate[i] = ISUPPER (i) ? tolower (i) : i;
|
preg->translate[i] = ISUPPER (i) ? tolower (i) : i;
|
}
|
}
|
else
|
else
|
preg->translate = NULL;
|
preg->translate = NULL;
|
|
|
/* If REG_NEWLINE is set, newlines are treated differently. */
|
/* If REG_NEWLINE is set, newlines are treated differently. */
|
if (cflags & REG_NEWLINE)
|
if (cflags & REG_NEWLINE)
|
{ /* REG_NEWLINE implies neither . nor [^...] match newline. */
|
{ /* REG_NEWLINE implies neither . nor [^...] match newline. */
|
syntax &= ~RE_DOT_NEWLINE;
|
syntax &= ~RE_DOT_NEWLINE;
|
syntax |= RE_HAT_LISTS_NOT_NEWLINE;
|
syntax |= RE_HAT_LISTS_NOT_NEWLINE;
|
/* It also changes the matching behavior. */
|
/* It also changes the matching behavior. */
|
preg->newline_anchor = 1;
|
preg->newline_anchor = 1;
|
}
|
}
|
else
|
else
|
preg->newline_anchor = 0;
|
preg->newline_anchor = 0;
|
|
|
preg->no_sub = !!(cflags & REG_NOSUB);
|
preg->no_sub = !!(cflags & REG_NOSUB);
|
|
|
/* POSIX says a null character in the pattern terminates it, so we
|
/* POSIX says a null character in the pattern terminates it, so we
|
can use strlen here in compiling the pattern. */
|
can use strlen here in compiling the pattern. */
|
ret = regex_compile (pattern, strlen (pattern), syntax, preg);
|
ret = regex_compile (pattern, strlen (pattern), syntax, preg);
|
|
|
/* POSIX doesn't distinguish between an unmatched open-group and an
|
/* POSIX doesn't distinguish between an unmatched open-group and an
|
unmatched close-group: both are REG_EPAREN. */
|
unmatched close-group: both are REG_EPAREN. */
|
if (ret == REG_ERPAREN) ret = REG_EPAREN;
|
if (ret == REG_ERPAREN) ret = REG_EPAREN;
|
|
|
return (int) ret;
|
return (int) ret;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__regcomp, regcomp)
|
weak_alias (__regcomp, regcomp)
|
#endif
|
#endif
|
|
|
|
|
/* regexec searches for a given pattern, specified by PREG, in the
|
/* regexec searches for a given pattern, specified by PREG, in the
|
string STRING.
|
string STRING.
|
|
|
If NMATCH is zero or REG_NOSUB was set in the cflags argument to
|
If NMATCH is zero or REG_NOSUB was set in the cflags argument to
|
`regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at
|
`regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at
|
least NMATCH elements, and we set them to the offsets of the
|
least NMATCH elements, and we set them to the offsets of the
|
corresponding matched substrings.
|
corresponding matched substrings.
|
|
|
EFLAGS specifies `execution flags' which affect matching: if
|
EFLAGS specifies `execution flags' which affect matching: if
|
REG_NOTBOL is set, then ^ does not match at the beginning of the
|
REG_NOTBOL is set, then ^ does not match at the beginning of the
|
string; if REG_NOTEOL is set, then $ does not match at the end.
|
string; if REG_NOTEOL is set, then $ does not match at the end.
|
|
|
We return 0 if we find a match and REG_NOMATCH if not. */
|
We return 0 if we find a match and REG_NOMATCH if not. */
|
|
|
int
|
int
|
regexec (preg, string, nmatch, pmatch, eflags)
|
regexec (preg, string, nmatch, pmatch, eflags)
|
const regex_t *preg;
|
const regex_t *preg;
|
const char *string;
|
const char *string;
|
size_t nmatch;
|
size_t nmatch;
|
regmatch_t pmatch[];
|
regmatch_t pmatch[];
|
int eflags;
|
int eflags;
|
{
|
{
|
int ret;
|
int ret;
|
struct re_registers regs;
|
struct re_registers regs;
|
regex_t private_preg;
|
regex_t private_preg;
|
int len = strlen (string);
|
int len = strlen (string);
|
boolean want_reg_info = !preg->no_sub && nmatch > 0;
|
boolean want_reg_info = !preg->no_sub && nmatch > 0;
|
|
|
private_preg = *preg;
|
private_preg = *preg;
|
|
|
private_preg.not_bol = !!(eflags & REG_NOTBOL);
|
private_preg.not_bol = !!(eflags & REG_NOTBOL);
|
private_preg.not_eol = !!(eflags & REG_NOTEOL);
|
private_preg.not_eol = !!(eflags & REG_NOTEOL);
|
|
|
/* The user has told us exactly how many registers to return
|
/* The user has told us exactly how many registers to return
|
information about, via `nmatch'. We have to pass that on to the
|
information about, via `nmatch'. We have to pass that on to the
|
matching routines. */
|
matching routines. */
|
private_preg.regs_allocated = REGS_FIXED;
|
private_preg.regs_allocated = REGS_FIXED;
|
|
|
if (want_reg_info)
|
if (want_reg_info)
|
{
|
{
|
regs.num_regs = nmatch;
|
regs.num_regs = nmatch;
|
regs.start = TALLOC (nmatch, regoff_t);
|
regs.start = TALLOC (nmatch, regoff_t);
|
regs.end = TALLOC (nmatch, regoff_t);
|
regs.end = TALLOC (nmatch, regoff_t);
|
if (regs.start == NULL || regs.end == NULL)
|
if (regs.start == NULL || regs.end == NULL)
|
return (int) REG_NOMATCH;
|
return (int) REG_NOMATCH;
|
}
|
}
|
|
|
/* Perform the searching operation. */
|
/* Perform the searching operation. */
|
ret = re_search (&private_preg, string, len,
|
ret = re_search (&private_preg, string, len,
|
/* start: */ 0, /* range: */ len,
|
/* start: */ 0, /* range: */ len,
|
want_reg_info ? ®s : (struct re_registers *) 0);
|
want_reg_info ? ®s : (struct re_registers *) 0);
|
|
|
/* Copy the register information to the POSIX structure. */
|
/* Copy the register information to the POSIX structure. */
|
if (want_reg_info)
|
if (want_reg_info)
|
{
|
{
|
if (ret >= 0)
|
if (ret >= 0)
|
{
|
{
|
unsigned r;
|
unsigned r;
|
|
|
for (r = 0; r < nmatch; r++)
|
for (r = 0; r < nmatch; r++)
|
{
|
{
|
pmatch[r].rm_so = regs.start[r];
|
pmatch[r].rm_so = regs.start[r];
|
pmatch[r].rm_eo = regs.end[r];
|
pmatch[r].rm_eo = regs.end[r];
|
}
|
}
|
}
|
}
|
|
|
/* If we needed the temporary register info, free the space now. */
|
/* If we needed the temporary register info, free the space now. */
|
free (regs.start);
|
free (regs.start);
|
free (regs.end);
|
free (regs.end);
|
}
|
}
|
|
|
/* We want zero return to mean success, unlike `re_search'. */
|
/* We want zero return to mean success, unlike `re_search'. */
|
return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH;
|
return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__regexec, regexec)
|
weak_alias (__regexec, regexec)
|
#endif
|
#endif
|
|
|
|
|
/* Returns a message corresponding to an error code, ERRCODE, returned
|
/* Returns a message corresponding to an error code, ERRCODE, returned
|
from either regcomp or regexec. We don't use PREG here. */
|
from either regcomp or regexec. We don't use PREG here. */
|
|
|
size_t
|
size_t
|
regerror (errcode, preg, errbuf, errbuf_size)
|
regerror (errcode, preg, errbuf, errbuf_size)
|
int errcode;
|
int errcode;
|
const regex_t *preg;
|
const regex_t *preg;
|
char *errbuf;
|
char *errbuf;
|
size_t errbuf_size;
|
size_t errbuf_size;
|
{
|
{
|
const char *msg;
|
const char *msg;
|
size_t msg_size;
|
size_t msg_size;
|
|
|
if (errcode < 0
|
if (errcode < 0
|
|| errcode >= (int) (sizeof (re_error_msgid)
|
|| errcode >= (int) (sizeof (re_error_msgid)
|
/ sizeof (re_error_msgid[0])))
|
/ sizeof (re_error_msgid[0])))
|
/* Only error codes returned by the rest of the code should be passed
|
/* Only error codes returned by the rest of the code should be passed
|
to this routine. If we are given anything else, or if other regex
|
to this routine. If we are given anything else, or if other regex
|
code generates an invalid error code, then the program has a bug.
|
code generates an invalid error code, then the program has a bug.
|
Dump core so we can fix it. */
|
Dump core so we can fix it. */
|
abort ();
|
abort ();
|
|
|
msg = gettext (re_error_msgid[errcode]);
|
msg = gettext (re_error_msgid[errcode]);
|
|
|
msg_size = strlen (msg) + 1; /* Includes the null. */
|
msg_size = strlen (msg) + 1; /* Includes the null. */
|
|
|
if (errbuf_size != 0)
|
if (errbuf_size != 0)
|
{
|
{
|
if (msg_size > errbuf_size)
|
if (msg_size > errbuf_size)
|
{
|
{
|
#if defined HAVE_MEMPCPY || defined _LIBC
|
#if defined HAVE_MEMPCPY || defined _LIBC
|
*((char *) __mempcpy (errbuf, msg, errbuf_size - 1)) = '\0';
|
*((char *) __mempcpy (errbuf, msg, errbuf_size - 1)) = '\0';
|
#else
|
#else
|
memcpy (errbuf, msg, errbuf_size - 1);
|
memcpy (errbuf, msg, errbuf_size - 1);
|
errbuf[errbuf_size - 1] = 0;
|
errbuf[errbuf_size - 1] = 0;
|
#endif
|
#endif
|
}
|
}
|
else
|
else
|
memcpy (errbuf, msg, msg_size);
|
memcpy (errbuf, msg, msg_size);
|
}
|
}
|
|
|
return msg_size;
|
return msg_size;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__regerror, regerror)
|
weak_alias (__regerror, regerror)
|
#endif
|
#endif
|
|
|
|
|
/* Free dynamically allocated space used by PREG. */
|
/* Free dynamically allocated space used by PREG. */
|
|
|
void
|
void
|
regfree (preg)
|
regfree (preg)
|
regex_t *preg;
|
regex_t *preg;
|
{
|
{
|
if (preg->buffer != NULL)
|
if (preg->buffer != NULL)
|
free (preg->buffer);
|
free (preg->buffer);
|
preg->buffer = NULL;
|
preg->buffer = NULL;
|
|
|
preg->allocated = 0;
|
preg->allocated = 0;
|
preg->used = 0;
|
preg->used = 0;
|
|
|
if (preg->fastmap != NULL)
|
if (preg->fastmap != NULL)
|
free (preg->fastmap);
|
free (preg->fastmap);
|
preg->fastmap = NULL;
|
preg->fastmap = NULL;
|
preg->fastmap_accurate = 0;
|
preg->fastmap_accurate = 0;
|
|
|
if (preg->translate != NULL)
|
if (preg->translate != NULL)
|
free (preg->translate);
|
free (preg->translate);
|
preg->translate = NULL;
|
preg->translate = NULL;
|
}
|
}
|
#ifdef _LIBC
|
#ifdef _LIBC
|
weak_alias (__regfree, regfree)
|
weak_alias (__regfree, regfree)
|
#endif
|
#endif
|
|
|
#endif /* not emacs */
|
#endif /* not emacs */
|
|
|
