AuroraMiddleware/glibc
1
0
Fork 0
mirror of https://sourceware.org/git/glibc.git synced 2024-11-21 12:30:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

posix: Sync gnulib regex implementation

Browse Source 
This patch syncs the regex implementation with gnulib (commit 0ee5212).
Only two changes in GLIBC regex testing are required:

  1. posix/bug-regex28.c: as previously discussed [1] the change of
     expected results on the pattern should be safe.

  2. posix/PCRE.tests: the ERE (a)|\1 is malformed (in the sense that
     the \1 doesn't mean anything) and although current GLIBC accepts
     it has undefined behavior.  This patch removes the specific test.

This sync contains some patches from thread 'Regex: Make libc regex
more usable outside GLIBC.' [2] which have been pushed upstream in
gnulib.  This patches also fixes some regex issues (BZ #23233,
BZ #21163, BZ #18986, BZ #13762) and I did not add testcases for
both #23233 and #13762 because I couldn't think a simple way to
trigger the expected failure path to trigger them.

Checked on x86_64-linux-gnu and i686-linux-gnu.

	[BZ #23233]
	[BZ #21163]
	[BZ #18986]
	[BZ #13762]
	* posix/Makefile (tests): Add bug-regex37 and bug-regex38.
	* posix/PCRE.tests: Remove invalid test.
	* posix/bug-regex28.c: Fix expected values for used syntax.
	* posix/bug-regex37.c: New file.
	* posix/bug-regex38.c: Likewise.
	* posix/regcomp.c: Sync with gnulib.
	* posix/regex.c: Likewise.
	* posix/regex.h: Likewise.
	* posix/regex_internal.c: Likewise.
	* posix/regex_internal.h: Likewise.
	* posix/regexec.c: Likewise.

[1] https://sourceware.org/ml/libc-alpha/2017-12/msg00807.html
[2] https://sourceware.org/ml/libc-alpha/2017-12/msg00237.html
This commit is contained in:
Adhemerval Zanella 2017-12-20 09:47:44 -02:00
parent b11643c21c
commit eb04c21373
12 changed files with 1610 additions and 1162 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
ChangeLog
View File

@ -1,3 +1,21 @@
2018-07-04 Adhemerval Zanella <adhemerval.zanella@linaro.org>
[BZ #23233]
[BZ #21163]
[BZ #18986]
[BZ #13762]
* posix/Makefile (tests): Add bug-regex37 and bug-regex38.
* posix/PCRE.tests: Remove invalid test.
* posix/bug-regex28.c: Fix expected values for used syntax.
* posix/bug-regex37.c: New file.
* posix/bug-regex38.c: Likewise.
* posix/regcomp.c: Sync with gnulib.
* posix/regex.c: Likewise.
* posix/regex.h: Likewise.
* posix/regex_internal.c: Likewise.
* posix/regex_internal.h: Likewise.
* posix/regexec.c: Likewise.
2018-06-26 Mike FABIAN <mfabian@redhat.com> 2018-06-26 Mike FABIAN <mfabian@redhat.com>
[BZ #23308] [BZ #23308]

3
posix/Makefile
View File

@ -95,7 +95,8 @@ tests := test-errno tstgetopt testfnm runtests runptests \
tst-posix_spawn-fd tst-posix_spawn-setsid \ tst-posix_spawn-fd tst-posix_spawn-setsid \
tst-posix_fadvise tst-posix_fadvise64 \ tst-posix_fadvise tst-posix_fadvise64 \
tst-sysconf-empty-chroot tst-glob_symlinks tst-fexecve \ tst-sysconf-empty-chroot tst-glob_symlinks tst-fexecve \
tst-glob-tilde test-ssize-max tst-spawn4 tst-glob-tilde test-ssize-max tst-spawn4 bug-regex37 \
bug-regex38
tests-internal := bug-regex5 bug-regex20 bug-regex33 \ tests-internal := bug-regex5 bug-regex20 bug-regex33 \
tst-rfc3484 tst-rfc3484-2 tst-rfc3484-3 \ tst-rfc3484 tst-rfc3484-2 tst-rfc3484-3 \
tst-glob_lstat_compat tst-spawn4-compat tst-glob_lstat_compat tst-spawn4-compat

13
posix/PCRE.tests
View File

@ -1774,19 +1774,6 @@ No match
0: abcabc 0: abcabc
1: abc 1: abc
/(a)|\1/
a
0: a
1: a
*** Failers
0: a
1: a
ab
0: a
1: a
x
No match
/abc/i /abc/i
ABC ABC
0: ABC 0: ABC

46
posix/bug-regex28.c
View File

@ -21,18 +21,22 @@
#include <stdio.h> #include <stdio.h>
#include <string.h> #include <string.h>
#include <support/test-driver.h>
#include <support/check.h>
struct tests struct tests
{ {
const char *regex; const char *regex;
const char *string; const char *string;
reg_syntax_t syntax; reg_syntax_t syntax;
int retval; int retval;
} tests[] = { };
static const struct tests tests[] = {
#define EGREP RE_SYNTAX_EGREP #define EGREP RE_SYNTAX_EGREP
#define EGREP_NL (RE_SYNTAX_EGREP | RE_DOT_NEWLINE) & ~RE_HAT_LISTS_NOT_NEWLINE #define EGREP_NL (RE_SYNTAX_EGREP | RE_DOT_NEWLINE) & ~RE_HAT_LISTS_NOT_NEWLINE
{ "a.b", "a\nb", EGREP, -1 }, { "a.b", "a\nb", EGREP, 0 },
{ "a.b", "a\nb", EGREP_NL, 0 }, { "a.b", "a\nb", EGREP_NL, 0 },
{ "a[^x]b", "a\nb", EGREP, -1 }, { "a[^x]b", "a\nb", EGREP, 0 },
{ "a[^x]b", "a\nb", EGREP_NL, 0 }, { "a[^x]b", "a\nb", EGREP_NL, 0 },
/* While \S and \W are internally handled as [^[:space:]] and [^[:alnum:]_], /* While \S and \W are internally handled as [^[:space:]] and [^[:alnum:]_],
RE_HAT_LISTS_NOT_NEWLINE did not make any difference, so ensure RE_HAT_LISTS_NOT_NEWLINE did not make any difference, so ensure
@ -42,33 +46,33 @@ struct tests
{ "a\\Wb", "a\nb", EGREP, 0 }, { "a\\Wb", "a\nb", EGREP, 0 },
{ "a\\Wb", "a\nb", EGREP_NL, 0 } { "a\\Wb", "a\nb", EGREP_NL, 0 }
}; };
static const size_t tests_size = sizeof (tests) / sizeof (tests[0]);
int static int
main (void) do_test (void)
{ {
struct re_pattern_buffer r; struct re_pattern_buffer r;
size_t i;
int ret = 0;
for (i = 0; i < sizeof (tests) / sizeof (tests[i]); ++i) for (size_t i = 0; i < tests_size; i++)
{ {
re_set_syntax (tests[i].syntax); re_set_syntax (tests[i].syntax);
memset (&r, 0, sizeof (r)); memset (&r, 0, sizeof (r));
if (re_compile_pattern (tests[i].regex, strlen (tests[i].regex), &r)) const char *re = re_compile_pattern (tests[i].regex,
{ strlen (tests[i].regex), &r);
printf ("re_compile_pattern %zd failed\n", i); TEST_VERIFY (re == NULL);
ret = 1; if (re != NULL)
continue; continue;
}
size_t len = strlen (tests[i].string); size_t len = strlen (tests[i].string);
int rv = re_search (&r, tests[i].string, len, 0, len, NULL); int rv = re_search (&r, tests[i].string, len, 0, len, NULL);
if (rv != tests[i].retval) TEST_VERIFY (rv == tests[i].retval);
{ if (test_verbose > 0)
printf ("re_search %zd unexpected value %d != %d\n", printf ("info: i=%zu rv=%d expected=%d\n", i, rv, tests[i].retval);
i, rv, tests[i].retval);
ret = 1;
}
regfree (&r); regfree (&r);
} }
return ret;
return 0;
} }
#include <support/test-driver.c>

32
posix/bug-regex37.c Normal file
View File

@ -0,0 +1,32 @@
/* Test regcomp return for invalid expression (BZ #21163).
Copyright (C) 2018 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#include <regex.h>
#include <support/check.h>
static int
do_test (void)
{
char const pattern[] = "()*)|\\1)*";
regex_t r;
TEST_VERIFY_EXIT (regcomp (&r, pattern, REG_EXTENDED) == REG_ESUBREG);
return 0;
}
#include <support/test-driver.c>

32
posix/bug-regex38.c Normal file
View File

@ -0,0 +1,32 @@
/* Diagnose invalid back-reference in the ERE '()|\1' (BZ #18986).
Copyright (C) 2018 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#include <regex.h>
#include <support/check.h>
static int
do_test (void)
{
char const pattern[] = "0|()0|\\1|0";
regex_t r;
TEST_VERIFY_EXIT (regcomp (&r, pattern, REG_EXTENDED) == REG_ESUBREG);
return 0;
}
#include <support/test-driver.c>

595
posix/regcomp.c
View File

File diff suppressed because it is too large Load Diff

21
posix/regex.c
View File

@ -15,14 +15,22 @@
You should have received a copy of the GNU Lesser General Public You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */ <https://www.gnu.org/licenses/>. */
#ifdef HAVE_CONFIG_H #ifndef _LIBC
#include "config.h" # include <config.h>
# if (__GNUC__ == 4 && 6 <= __GNUC_MINOR__) || 4 < __GNUC__
# pragma GCC diagnostic ignored "-Wsuggest-attribute=pure"
# endif
# if (__GNUC__ == 4 && 3 <= __GNUC_MINOR__) || 4 < __GNUC__
# pragma GCC diagnostic ignored "-Wold-style-definition"
# pragma GCC diagnostic ignored "-Wtype-limits"
# endif
#endif #endif
/* Make sure noone compiles this code with a C++ compiler. */ /* Make sure no one compiles this code with a C++ compiler. */
#ifdef __cplusplus #if defined __cplusplus && defined _LIBC
# error "This is C code, use a C compiler" # error "This is C code, use a C compiler"
#endif #endif
@ -56,9 +64,6 @@
#undefs RE_DUP_MAX and sets it to the right value. */ #undefs RE_DUP_MAX and sets it to the right value. */
#include <limits.h> #include <limits.h>
/* This header defines the MIN and MAX macros. */
#include <sys/param.h>
#include <regex.h> #include <regex.h>
#include "regex_internal.h" #include "regex_internal.h"

335
posix/regex.h
View File

@ -15,7 +15,7 @@
You should have received a copy of the GNU Lesser General Public You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */ <https://www.gnu.org/licenses/>. */
#ifndef _REGEX_H #ifndef _REGEX_H
#define _REGEX_H 1 #define _REGEX_H 1
@ -27,6 +27,36 @@
extern "C" { extern "C" {
#endif #endif
/* Define __USE_GNU to declare GNU extensions that violate the
POSIX name space rules. */
#ifdef _GNU_SOURCE
# define __USE_GNU 1
#endif
#ifdef _REGEX_LARGE_OFFSETS
/* Use types and values that are wide enough to represent signed and
unsigned byte offsets in memory. This currently works only when
the regex code is used outside of the GNU C library; it is not yet
supported within glibc itself, and glibc users should not define
_REGEX_LARGE_OFFSETS. */
/* The type of object sizes. */
typedef size_t __re_size_t;
/* The type of object sizes, in places where the traditional code
uses unsigned long int. */
typedef size_t __re_long_size_t;
#else
/* The traditional GNU regex implementation mishandles strings longer
than INT_MAX. */
typedef unsigned int __re_size_t;
typedef unsigned long int __re_long_size_t;
#endif
/* The following two types have to be signed and unsigned integer type /* The following two types have to be signed and unsigned integer type
wide enough to hold a value of a pointer. For most ANSI compilers wide enough to hold a value of a pointer. For most ANSI compilers
ptrdiff_t and size_t should be likely OK. Still size of these two ptrdiff_t and size_t should be likely OK. Still size of these two
@ -108,9 +138,9 @@ typedef unsigned long int reg_syntax_t;
If not set, newline is literal. */ If not set, newline is literal. */
# define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
/* If this bit is set, then `{...}' defines an interval, and \{ and \} /* If this bit is set, then '{...}' defines an interval, and \{ and \}
are literals. are literals.
If not set, then `\{...\}' defines an interval. */ If not set, then '\{...\}' defines an interval. */
# define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
/* If this bit is set, (...) defines a group, and \( and \) are literals. /* If this bit is set, (...) defines a group, and \( and \) are literals.
@ -165,8 +195,8 @@ typedef unsigned long int reg_syntax_t;
whether ^ should be special. */ whether ^ should be special. */
# define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1)
/* If this bit is set, then \{ cannot be first in an bre or /* If this bit is set, then \{ cannot be first in a regex or
immediately after an alternation or begin-group operator. */ immediately after an alternation, open-group or \} operator. */
# define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1)
/* If this bit is set, then no_sub will be set to 1 during /* If this bit is set, then no_sub will be set to 1 during
@ -185,9 +215,9 @@ extern reg_syntax_t re_syntax_options;
(The [[[ comments delimit what gets put into the Texinfo file, so (The [[[ comments delimit what gets put into the Texinfo file, so
don't delete them!) */ don't delete them!) */
/* [[[begin syntaxes]]] */ /* [[[begin syntaxes]]] */
#define RE_SYNTAX_EMACS 0 # define RE_SYNTAX_EMACS 0
#define RE_SYNTAX_AWK \ # define RE_SYNTAX_AWK \
(RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
| RE_NO_BK_PARENS | RE_NO_BK_REFS \ | RE_NO_BK_PARENS | RE_NO_BK_REFS \
| RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
@ -195,52 +225,49 @@ extern reg_syntax_t re_syntax_options;
| RE_CHAR_CLASSES \ | RE_CHAR_CLASSES \
| RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
#define RE_SYNTAX_GNU_AWK \ # define RE_SYNTAX_GNU_AWK \
((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
| RE_INVALID_INTERVAL_ORD) \ | RE_INVALID_INTERVAL_ORD) \
& ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \ & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \
| RE_CONTEXT_INVALID_OPS )) | RE_CONTEXT_INVALID_OPS ))
#define RE_SYNTAX_POSIX_AWK \ # define RE_SYNTAX_POSIX_AWK \
(RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
| RE_INTERVALS | RE_NO_GNU_OPS \ | RE_INTERVALS | RE_NO_GNU_OPS \
| RE_INVALID_INTERVAL_ORD) | RE_INVALID_INTERVAL_ORD)
#define RE_SYNTAX_GREP \ # define RE_SYNTAX_GREP \
(RE_BK_PLUS_QM | RE_CHAR_CLASSES \ ((RE_SYNTAX_POSIX_BASIC | RE_NEWLINE_ALT) \
| RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ & ~(RE_CONTEXT_INVALID_DUP | RE_DOT_NOT_NULL))
| RE_NEWLINE_ALT)
#define RE_SYNTAX_EGREP \ # define RE_SYNTAX_EGREP \
(RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ ((RE_SYNTAX_POSIX_EXTENDED | RE_INVALID_INTERVAL_ORD | RE_NEWLINE_ALT) \
| RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ & ~(RE_CONTEXT_INVALID_OPS | RE_DOT_NOT_NULL))
| RE_NEWLINE_ALT | RE_NO_BK_PARENS \
| RE_NO_BK_VBAR)
#define RE_SYNTAX_POSIX_EGREP \ /* POSIX grep -E behavior is no longer incompatible with GNU. */
(RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ # define RE_SYNTAX_POSIX_EGREP \
| RE_INVALID_INTERVAL_ORD) RE_SYNTAX_EGREP
/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC # define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC # define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
/* Syntax bits common to both basic and extended POSIX regex syntax. */ /* Syntax bits common to both basic and extended POSIX regex syntax. */
#define _RE_SYNTAX_POSIX_COMMON \ # define _RE_SYNTAX_POSIX_COMMON \
(RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
| RE_INTERVALS | RE_NO_EMPTY_RANGES) | RE_INTERVALS | RE_NO_EMPTY_RANGES)
#define RE_SYNTAX_POSIX_BASIC \ # define RE_SYNTAX_POSIX_BASIC \
(_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP)
/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
isn't minimal, since other operators, such as \`, aren't disabled. */ isn't minimal, since other operators, such as \`, aren't disabled. */
#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ # define RE_SYNTAX_POSIX_MINIMAL_BASIC \
(_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
#define RE_SYNTAX_POSIX_EXTENDED \ # define RE_SYNTAX_POSIX_EXTENDED \
(_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
| RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
| RE_NO_BK_PARENS | RE_NO_BK_VBAR \ | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
@ -248,25 +275,35 @@ extern reg_syntax_t re_syntax_options;
/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is
removed and RE_NO_BK_REFS is added. */ removed and RE_NO_BK_REFS is added. */
#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ # define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
(_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
| RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
| RE_NO_BK_PARENS | RE_NO_BK_REFS \ | RE_NO_BK_PARENS | RE_NO_BK_REFS \
| RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
/* [[[end syntaxes]]] */ /* [[[end syntaxes]]] */
/* Maximum number of duplicates an interval can allow. Some systems /* Maximum number of duplicates an interval can allow. POSIX-conforming
(erroneously) define this in other header files, but we want our systems might define this in <limits.h>, but we want our
value, so remove any previous define. */ value, so remove any previous define. */
# ifdef _REGEX_INCLUDE_LIMITS_H
# include <limits.h>
# endif
# ifdef RE_DUP_MAX # ifdef RE_DUP_MAX
# undef RE_DUP_MAX # undef RE_DUP_MAX
# endif # endif
/* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */
/* RE_DUP_MAX is 2**15 - 1 because an earlier implementation stored
the counter as a 2-byte signed integer. This is no longer true, so
RE_DUP_MAX could be increased to (INT_MAX / 10 - 1), or to
((SIZE_MAX - 9) / 10) if _REGEX_LARGE_OFFSETS is defined.
However, there would be a huge performance problem if someone
actually used a pattern like a\{214748363\}, so RE_DUP_MAX retains
its historical value. */
# define RE_DUP_MAX (0x7fff) # define RE_DUP_MAX (0x7fff)
#endif #endif
/* POSIX `cflags' bits (i.e., information for `regcomp'). */ /* POSIX 'cflags' bits (i.e., information for 'regcomp'). */
/* If this bit is set, then use extended regular expression syntax. /* If this bit is set, then use extended regular expression syntax.
If not set, then use basic regular expression syntax. */ If not set, then use basic regular expression syntax. */
@ -274,19 +311,19 @@ extern reg_syntax_t re_syntax_options;
/* If this bit is set, then ignore case when matching. /* If this bit is set, then ignore case when matching.
If not set, then case is significant. */ If not set, then case is significant. */
#define REG_ICASE (REG_EXTENDED << 1) #define REG_ICASE (1 << 1)
/* If this bit is set, then anchors do not match at newline /* If this bit is set, then anchors do not match at newline
characters in the string. characters in the string.
If not set, then anchors do match at newlines. */ If not set, then anchors do match at newlines. */
#define REG_NEWLINE (REG_ICASE << 1) #define REG_NEWLINE (1 << 2)
/* If this bit is set, then report only success or fail in regexec. /* If this bit is set, then report only success or fail in regexec.
If not set, then returns differ between not matching and errors. */ If not set, then returns differ between not matching and errors. */
#define REG_NOSUB (REG_NEWLINE << 1) #define REG_NOSUB (1 << 3)
/* POSIX `eflags' bits (i.e., information for regexec). */ /* POSIX 'eflags' bits (i.e., information for regexec). */
/* If this bit is set, then the beginning-of-line operator doesn't match /* If this bit is set, then the beginning-of-line operator doesn't match
the beginning of the string (presumably because it's not the the beginning of the string (presumably because it's not the
@ -304,41 +341,60 @@ extern reg_syntax_t re_syntax_options;
/* If any error codes are removed, changed, or added, update the /* If any error codes are removed, changed, or added, update the
`re_error_msg' table in regex.c. */ '__re_error_msgid' table in regcomp.c. */
typedef enum typedef enum
{ {
#if defined _XOPEN_SOURCE || defined __USE_XOPEN2K _REG_ENOSYS = -1, /* This will never happen for this implementation. */
REG_ENOSYS = -1, /* This will never happen for this implementation. */ _REG_NOERROR = 0, /* Success. */
#endif _REG_NOMATCH, /* Didn't find a match (for regexec). */
REG_NOERROR = 0, /* Success. */
REG_NOMATCH, /* Didn't find a match (for regexec). */
/* POSIX regcomp return error codes. (In the order listed in the /* POSIX regcomp return error codes. (In the order listed in the
standard.) */ standard.) */
REG_BADPAT, /* Invalid pattern. */ _REG_BADPAT, /* Invalid pattern. */
REG_ECOLLATE, /* Invalid collating element. */ _REG_ECOLLATE, /* Invalid collating element. */
REG_ECTYPE, /* Invalid character class name. */ _REG_ECTYPE, /* Invalid character class name. */
REG_EESCAPE, /* Trailing backslash. */ _REG_EESCAPE, /* Trailing backslash. */
REG_ESUBREG, /* Invalid back reference. */ _REG_ESUBREG, /* Invalid back reference. */
REG_EBRACK, /* Unmatched left bracket. */ _REG_EBRACK, /* Unmatched left bracket. */
REG_EPAREN, /* Parenthesis imbalance. */ _REG_EPAREN, /* Parenthesis imbalance. */
REG_EBRACE, /* Unmatched \{. */ _REG_EBRACE, /* Unmatched \{. */
REG_BADBR, /* Invalid contents of \{\}. */ _REG_BADBR, /* Invalid contents of \{\}. */
REG_ERANGE, /* Invalid range end. */ _REG_ERANGE, /* Invalid range end. */
REG_ESPACE, /* Ran out of memory. */ _REG_ESPACE, /* Ran out of memory. */
REG_BADRPT, /* No preceding re for repetition op. */ _REG_BADRPT, /* No preceding re for repetition op. */
/* Error codes we've added. */ /* Error codes we've added. */
REG_EEND, /* Premature end. */ _REG_EEND, /* Premature end. */
REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ _REG_ESIZE, /* Too large (e.g., repeat count too large). */
REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
} reg_errcode_t; } reg_errcode_t;
#if defined _XOPEN_SOURCE || defined __USE_XOPEN2K
# define REG_ENOSYS _REG_ENOSYS
#endif
#define REG_NOERROR _REG_NOERROR
#define REG_NOMATCH _REG_NOMATCH
#define REG_BADPAT _REG_BADPAT
#define REG_ECOLLATE _REG_ECOLLATE
#define REG_ECTYPE _REG_ECTYPE
#define REG_EESCAPE _REG_EESCAPE
#define REG_ESUBREG _REG_ESUBREG
#define REG_EBRACK _REG_EBRACK
#define REG_EPAREN _REG_EPAREN
#define REG_EBRACE _REG_EBRACE
#define REG_BADBR _REG_BADBR
#define REG_ERANGE _REG_ERANGE
#define REG_ESPACE _REG_ESPACE
#define REG_BADRPT _REG_BADRPT
#define REG_EEND _REG_EEND
#define REG_ESIZE _REG_ESIZE
#define REG_ERPAREN _REG_ERPAREN
/* This data structure represents a compiled pattern. Before calling /* This data structure represents a compiled pattern. Before calling
the pattern compiler, the fields `buffer', `allocated', `fastmap', the pattern compiler, the fields 'buffer', 'allocated', 'fastmap',
and `translate' can be set. After the pattern has been compiled, and 'translate' can be set. After the pattern has been compiled,
the fields `re_nsub', `not_bol' and `not_eol' are available. All the fields 're_nsub', 'not_bol' and 'not_eol' are available. All
other fields are private to the regex routines. */ other fields are private to the regex routines. */
#ifndef RE_TRANSLATE_TYPE #ifndef RE_TRANSLATE_TYPE
@ -356,16 +412,15 @@ typedef enum
struct re_pattern_buffer struct re_pattern_buffer
{ {
/* Space that holds the compiled pattern. It is declared as /* Space that holds the compiled pattern. The type
`unsigned char *' because its elements are sometimes used as 'struct re_dfa_t' is private and is not declared here. */
array indexes. */ struct re_dfa_t *__REPB_PREFIX(buffer);
unsigned char *__REPB_PREFIX(buffer);
/* Number of bytes to which `buffer' points. */ /* Number of bytes to which 'buffer' points. */
unsigned long int __REPB_PREFIX(allocated); __re_long_size_t __REPB_PREFIX(allocated);
/* Number of bytes actually used in `buffer'. */ /* Number of bytes actually used in 'buffer'. */
unsigned long int __REPB_PREFIX(used); __re_long_size_t __REPB_PREFIX(used);
/* Syntax setting with which the pattern was compiled. */ /* Syntax setting with which the pattern was compiled. */
reg_syntax_t __REPB_PREFIX(syntax); reg_syntax_t __REPB_PREFIX(syntax);
@ -385,13 +440,13 @@ struct re_pattern_buffer
size_t re_nsub; size_t re_nsub;
/* Zero if this pattern cannot match the empty string, one else. /* Zero if this pattern cannot match the empty string, one else.
Well, in truth it's used only in `re_search_2', to see whether or Well, in truth it's used only in 're_search_2', to see whether or
not we should use the fastmap, so we don't set this absolutely not we should use the fastmap, so we don't set this absolutely
perfectly; see `re_compile_fastmap' (the `duplicate' case). */ perfectly; see 're_compile_fastmap' (the "duplicate" case). */
unsigned __REPB_PREFIX(can_be_null) : 1; unsigned __REPB_PREFIX(can_be_null) : 1;
/* If REGS_UNALLOCATED, allocate space in the `regs' structure /* If REGS_UNALLOCATED, allocate space in the 'regs' structure
for `max (RE_NREGS, re_nsub + 1)' groups. for 'max (RE_NREGS, re_nsub + 1)' groups.
If REGS_REALLOCATE, reallocate space if necessary. If REGS_REALLOCATE, reallocate space if necessary.
If REGS_FIXED, use what's there. */ If REGS_FIXED, use what's there. */
#ifdef __USE_GNU #ifdef __USE_GNU
@ -401,11 +456,11 @@ struct re_pattern_buffer
#endif #endif
unsigned __REPB_PREFIX(regs_allocated) : 2; unsigned __REPB_PREFIX(regs_allocated) : 2;
/* Set to zero when `regex_compile' compiles a pattern; set to one /* Set to zero when 're_compile_pattern' compiles a pattern; set to
by `re_compile_fastmap' if it updates the fastmap. */ one by 're_compile_fastmap' if it updates the fastmap. */
unsigned __REPB_PREFIX(fastmap_accurate) : 1; unsigned __REPB_PREFIX(fastmap_accurate) : 1;
/* If set, `re_match_2' does not return information about /* If set, 're_match_2' does not return information about
subexpressions. */ subexpressions. */
unsigned __REPB_PREFIX(no_sub) : 1; unsigned __REPB_PREFIX(no_sub) : 1;
@ -423,7 +478,17 @@ struct re_pattern_buffer
typedef struct re_pattern_buffer regex_t; typedef struct re_pattern_buffer regex_t;
/* Type for byte offsets within the string. POSIX mandates this. */ /* Type for byte offsets within the string. POSIX mandates this. */
#ifdef _REGEX_LARGE_OFFSETS
/* POSIX 1003.1-2008 requires that regoff_t be at least as wide as
ptrdiff_t and ssize_t. We don't know of any hosts where ptrdiff_t
is wider than ssize_t, so ssize_t is safe. ptrdiff_t is not
visible here, so use ssize_t. */
typedef ssize_t regoff_t;
#else
/* The traditional GNU regex implementation mishandles strings longer
than INT_MAX. */
typedef int regoff_t; typedef int regoff_t;
#endif
#ifdef __USE_GNU #ifdef __USE_GNU
@ -431,15 +496,15 @@ typedef int regoff_t;
regex.texinfo for a full description of what registers match. */ regex.texinfo for a full description of what registers match. */
struct re_registers struct re_registers
{ {
unsigned num_regs; __re_size_t num_regs;
regoff_t *start; regoff_t *start;
regoff_t *end; regoff_t *end;
}; };
/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, /* If 'regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
`re_match_2' returns information about at least this many registers 're_match_2' returns information about at least this many registers
the first time a `regs' structure is passed. */ the first time a 'regs' structure is passed. */
# ifndef RE_NREGS # ifndef RE_NREGS
# define RE_NREGS 30 # define RE_NREGS 30
# endif # endif
@ -447,7 +512,7 @@ struct re_registers
/* POSIX specification for registers. Aside from the different names than /* POSIX specification for registers. Aside from the different names than
`re_registers', POSIX uses an array of structures, instead of a 're_registers', POSIX uses an array of structures, instead of a
structure of arrays. */ structure of arrays. */
typedef struct typedef struct
{ {
@ -459,17 +524,17 @@ typedef struct
#ifdef __USE_GNU #ifdef __USE_GNU
/* Sets the current default syntax to SYNTAX, and return the old syntax. /* Sets the current default syntax to SYNTAX, and return the old syntax.
You can also simply assign to the `re_syntax_options' variable. */ You can also simply assign to the 're_syntax_options' variable. */
extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax);
/* Compile the regular expression PATTERN, with length LENGTH /* Compile the regular expression PATTERN, with length LENGTH
and syntax given by the global `re_syntax_options', into the buffer and syntax given by the global 're_syntax_options', into the buffer
BUFFER. Return NULL if successful, and an error string if not. BUFFER. Return NULL if successful, and an error string if not.
To free the allocated storage, you must call `regfree' on BUFFER. To free the allocated storage, you must call 'regfree' on BUFFER.
Note that the translate table must either have been initialised by Note that the translate table must either have been initialized by
`regcomp', with a malloc'ed value, or set to NULL before calling 'regcomp', with a malloc'ed value, or set to NULL before calling
`regfree'. */ 'regfree'. */
extern const char *re_compile_pattern (const char *__pattern, size_t __length, extern const char *re_compile_pattern (const char *__pattern, size_t __length,
struct re_pattern_buffer *__buffer); struct re_pattern_buffer *__buffer);
@ -485,47 +550,52 @@ extern int re_compile_fastmap (struct re_pattern_buffer *__buffer);
characters. Return the starting position of the match, -1 for no characters. Return the starting position of the match, -1 for no
match, or -2 for an internal error. Also return register match, or -2 for an internal error. Also return register
information in REGS (if REGS and BUFFER->no_sub are nonzero). */ information in REGS (if REGS and BUFFER->no_sub are nonzero). */
extern int re_search (struct re_pattern_buffer *__buffer, const char *__string, extern regoff_t re_search (struct re_pattern_buffer *__buffer,
int __length, int __start, int __range, const char *__String, regoff_t __length,
struct re_registers *__regs); regoff_t __start, regoff_t __range,
struct re_registers *__regs);
/* Like `re_search', but search in the concatenation of STRING1 and /* Like 're_search', but search in the concatenation of STRING1 and
STRING2. Also, stop searching at index START + STOP. */ STRING2. Also, stop searching at index START + STOP. */
extern int re_search_2 (struct re_pattern_buffer *__buffer, extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer,
const char *__string1, int __length1, const char *__string1, regoff_t __length1,
const char *__string2, int __length2, int __start, const char *__string2, regoff_t __length2,
int __range, struct re_registers *__regs, int __stop); regoff_t __start, regoff_t __range,
struct re_registers *__regs,
regoff_t __stop);
/* Like `re_search', but return how many characters in STRING the regexp /* Like 're_search', but return how many characters in STRING the regexp
in BUFFER matched, starting at position START. */ in BUFFER matched, starting at position START. */
extern int re_match (struct re_pattern_buffer *__buffer, const char *__string, extern regoff_t re_match (struct re_pattern_buffer *__buffer,
int __length, int __start, struct re_registers *__regs); const char *__String, regoff_t __length,
regoff_t __start, struct re_registers *__regs);
/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ /* Relates to 're_match' as 're_search_2' relates to 're_search'. */
extern int re_match_2 (struct re_pattern_buffer *__buffer, extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer,
const char *__string1, int __length1, const char *__string1, regoff_t __length1,
const char *__string2, int __length2, int __start, const char *__string2, regoff_t __length2,
struct re_registers *__regs, int __stop); regoff_t __start, struct re_registers *__regs,
regoff_t __stop);
/* 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 BUFFER and REGS will use this memory ENDS. Subsequent matches using BUFFER and REGS will use this memory
for recording register information. STARTS and ENDS must be for recording register information. STARTS and ENDS must be
allocated with malloc, and must each be at least `NUM_REGS * sizeof allocated with malloc, and must each be at least 'NUM_REGS * sizeof
(regoff_t)' bytes long. (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 BUFFER will allocate its own register data, without
freeing the old data. */ freeing the old data. */
extern void re_set_registers (struct re_pattern_buffer *__buffer, extern void re_set_registers (struct re_pattern_buffer *__buffer,
struct re_registers *__regs, struct re_registers *__regs,
unsigned int __num_regs, __re_size_t __num_regs,
regoff_t *__starts, regoff_t *__ends); regoff_t *__starts, regoff_t *__ends);
#endif /* Use GNU */ #endif /* Use GNU */
@ -537,39 +607,46 @@ extern int re_exec (const char *);
# endif # endif
#endif #endif
/* GCC 2.95 and later have "__restrict"; C99 compilers have /* For plain 'restrict', use glibc's __restrict if defined.
"restrict", and "configure" may have defined "restrict". */ Otherwise, GCC 2.95 and later have "__restrict"; C99 compilers have
#ifndef __restrict "restrict", and "configure" may have defined "restrict".
# if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) Other compilers use __restrict, __restrict__, and _Restrict, and
# if defined restrict || 199901L <= __STDC_VERSION__ 'configure' might #define 'restrict' to those words, so pick a
# define __restrict restrict different name. */
# else #ifndef _Restrict_
# define __restrict # if defined __restrict || 2 < __GNUC__ + (95 <= __GNUC_MINOR__)
# endif # define _Restrict_ __restrict
# elif 199901L <= __STDC_VERSION__ || defined restrict
# define _Restrict_ restrict
# else
# define _Restrict_
# endif # endif
#endif #endif
/* gcc 3.1 and up support the [restrict] syntax. */ /* For [restrict], use glibc's __restrict_arr if available.
#ifndef __restrict_arr Otherwise, GCC 3.1 (not in C++ mode) and C99 support [restrict]. */
# if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) \ #ifndef _Restrict_arr_
&& !defined __GNUG__ # ifdef __restrict_arr
# define __restrict_arr __restrict # define _Restrict_arr_ __restrict_arr
# elif ((199901L <= __STDC_VERSION__ || 3 < __GNUC__ + (1 <= __GNUC_MINOR__)) \
&& !defined __GNUG__)
# define _Restrict_arr_ _Restrict_
# else # else
# define __restrict_arr # define _Restrict_arr_
# endif # endif
#endif #endif
/* POSIX compatibility. */ /* POSIX compatibility. */
extern int regcomp (regex_t *__restrict __preg, extern int regcomp (regex_t *_Restrict_ __preg,
const char *__restrict __pattern, const char *_Restrict_ __pattern,
int __cflags); int __cflags);
extern int regexec (const regex_t *__restrict __preg, extern int regexec (const regex_t *_Restrict_ __preg,
const char *__restrict __string, size_t __nmatch, const char *_Restrict_ __String, size_t __nmatch,
regmatch_t __pmatch[__restrict_arr], regmatch_t __pmatch[_Restrict_arr_],
int __eflags); int __eflags);
extern size_t regerror (int __errcode, const regex_t *__restrict __preg, extern size_t regerror (int __errcode, const regex_t *_Restrict_ __preg,
char *__restrict __errbuf, size_t __errbuf_size); char *_Restrict_ __errbuf, size_t __errbuf_size);
extern void regfree (regex_t *__preg); extern void regfree (regex_t *__preg);

295
posix/regex_internal.c
View File

@ -15,19 +15,29 @@
You should have received a copy of the GNU Lesser General Public You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */ <https://www.gnu.org/licenses/>. */
static void re_string_construct_common (const char *str, int len, static void re_string_construct_common (const char *str, Idx len,
re_string_t *pstr, re_string_t *pstr,
RE_TRANSLATE_TYPE trans, int icase, RE_TRANSLATE_TYPE trans, bool icase,
const re_dfa_t *dfa); const re_dfa_t *dfa);
static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa, static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa,
const re_node_set *nodes, const re_node_set *nodes,
unsigned int hash); re_hashval_t hash);
static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa, static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa,
const re_node_set *nodes, const re_node_set *nodes,
unsigned int context, unsigned int context,
unsigned int hash); re_hashval_t hash);
static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
Idx new_buf_len);
#ifdef RE_ENABLE_I18N
static void build_wcs_buffer (re_string_t *pstr);
static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr);
#endif /* RE_ENABLE_I18N */
static void build_upper_buffer (re_string_t *pstr);
static void re_string_translate_buffer (re_string_t *pstr);
static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
int eflags) __attribute__ ((pure));
/* Functions for string operation. */ /* Functions for string operation. */
@ -36,11 +46,11 @@ static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa,
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_string_allocate (re_string_t *pstr, const char *str, int len, int init_len, re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx init_len,
RE_TRANSLATE_TYPE trans, int icase, const re_dfa_t *dfa) RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa)
{ {
reg_errcode_t ret; reg_errcode_t ret;
int init_buf_len; Idx init_buf_len;
/* Ensure at least one character fits into the buffers. */ /* Ensure at least one character fits into the buffers. */
if (init_len < dfa->mb_cur_max) if (init_len < dfa->mb_cur_max)
@ -64,8 +74,8 @@ re_string_allocate (re_string_t *pstr, const char *str, int len, int init_len,
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_string_construct (re_string_t *pstr, const char *str, int len, re_string_construct (re_string_t *pstr, const char *str, Idx len,
RE_TRANSLATE_TYPE trans, int icase, const re_dfa_t *dfa) RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa)
{ {
reg_errcode_t ret; reg_errcode_t ret;
memset (pstr, '\0', sizeof (re_string_t)); memset (pstr, '\0', sizeof (re_string_t));
@ -127,7 +137,7 @@ re_string_construct (re_string_t *pstr, const char *str, int len,
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_string_realloc_buffers (re_string_t *pstr, int new_buf_len) re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len)
{ {
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
if (pstr->mb_cur_max > 1) if (pstr->mb_cur_max > 1)
@ -135,8 +145,8 @@ re_string_realloc_buffers (re_string_t *pstr, int new_buf_len)
wint_t *new_wcs; wint_t *new_wcs;
/* Avoid overflow in realloc. */ /* Avoid overflow in realloc. */
const size_t max_object_size = MAX (sizeof (wint_t), sizeof (int)); const size_t max_object_size = MAX (sizeof (wint_t), sizeof (Idx));
if (BE (SIZE_MAX / max_object_size < new_buf_len, 0)) if (BE (MIN (IDX_MAX, SIZE_MAX / max_object_size) < new_buf_len, 0))
return REG_ESPACE; return REG_ESPACE;
new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len); new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len);
@ -145,7 +155,7 @@ re_string_realloc_buffers (re_string_t *pstr, int new_buf_len)
pstr->wcs = new_wcs; pstr->wcs = new_wcs;
if (pstr->offsets != NULL) if (pstr->offsets != NULL)
{ {
int *new_offsets = re_realloc (pstr->offsets, int, new_buf_len); Idx *new_offsets = re_realloc (pstr->offsets, Idx, new_buf_len);
if (BE (new_offsets == NULL, 0)) if (BE (new_offsets == NULL, 0))
return REG_ESPACE; return REG_ESPACE;
pstr->offsets = new_offsets; pstr->offsets = new_offsets;
@ -166,15 +176,15 @@ re_string_realloc_buffers (re_string_t *pstr, int new_buf_len)
static void static void
re_string_construct_common (const char *str, int len, re_string_t *pstr, re_string_construct_common (const char *str, Idx len, re_string_t *pstr,
RE_TRANSLATE_TYPE trans, int icase, RE_TRANSLATE_TYPE trans, bool icase,
const re_dfa_t *dfa) const re_dfa_t *dfa)
{ {
pstr->raw_mbs = (const unsigned char *) str; pstr->raw_mbs = (const unsigned char *) str;
pstr->len = len; pstr->len = len;
pstr->raw_len = len; pstr->raw_len = len;
pstr->trans = trans; pstr->trans = trans;
pstr->icase = icase ? 1 : 0; pstr->icase = icase;
pstr->mbs_allocated = (trans != NULL || icase); pstr->mbs_allocated = (trans != NULL || icase);
pstr->mb_cur_max = dfa->mb_cur_max; pstr->mb_cur_max = dfa->mb_cur_max;
pstr->is_utf8 = dfa->is_utf8; pstr->is_utf8 = dfa->is_utf8;
@ -206,7 +216,7 @@ build_wcs_buffer (re_string_t *pstr)
unsigned char buf[64]; unsigned char buf[64];
#endif #endif
mbstate_t prev_st; mbstate_t prev_st;
int byte_idx, end_idx, remain_len; Idx byte_idx, end_idx, remain_len;
size_t mbclen; size_t mbclen;
/* Build the buffers from pstr->valid_len to either pstr->len or /* Build the buffers from pstr->valid_len to either pstr->len or
@ -269,7 +279,7 @@ __attribute_warn_unused_result__
build_wcs_upper_buffer (re_string_t *pstr) build_wcs_upper_buffer (re_string_t *pstr)
{ {
mbstate_t prev_st; mbstate_t prev_st;
int src_idx, byte_idx, end_idx, remain_len; Idx src_idx, byte_idx, end_idx, remain_len;
size_t mbclen; size_t mbclen;
#ifdef _LIBC #ifdef _LIBC
char buf[MB_LEN_MAX]; char buf[MB_LEN_MAX];
@ -307,14 +317,13 @@ build_wcs_upper_buffer (re_string_t *pstr)
mbclen = __mbrtowc (&wc, mbclen = __mbrtowc (&wc,
((const char *) pstr->raw_mbs + pstr->raw_mbs_idx ((const char *) pstr->raw_mbs + pstr->raw_mbs_idx
+ byte_idx), remain_len, &pstr->cur_state); + byte_idx), remain_len, &pstr->cur_state);
if (BE (mbclen + 2 > 2, 1)) if (BE (mbclen < (size_t) -2, 1))
{ {
wchar_t wcu = wc; wchar_t wcu = __towupper (wc);
if (__iswlower (wc)) if (wcu != wc)
{ {
size_t mbcdlen; size_t mbcdlen;
wcu = __towupper (wc);
mbcdlen = __wcrtomb (buf, wcu, &prev_st); mbcdlen = __wcrtomb (buf, wcu, &prev_st);
if (BE (mbclen == mbcdlen, 1)) if (BE (mbclen == mbcdlen, 1))
memcpy (pstr->mbs + byte_idx, buf, mbclen); memcpy (pstr->mbs + byte_idx, buf, mbclen);
@ -377,14 +386,13 @@ build_wcs_upper_buffer (re_string_t *pstr)
else else
p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx; p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx;
mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state); mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state);
if (BE (mbclen + 2 > 2, 1)) if (BE (mbclen < (size_t) -2, 1))
{ {
wchar_t wcu = wc; wchar_t wcu = __towupper (wc);
if (__iswlower (wc)) if (wcu != wc)
{ {
size_t mbcdlen; size_t mbcdlen;
wcu = __towupper (wc);
mbcdlen = __wcrtomb ((char *) buf, wcu, &prev_st); mbcdlen = __wcrtomb ((char *) buf, wcu, &prev_st);
if (BE (mbclen == mbcdlen, 1)) if (BE (mbclen == mbcdlen, 1))
memcpy (pstr->mbs + byte_idx, buf, mbclen); memcpy (pstr->mbs + byte_idx, buf, mbclen);
@ -400,7 +408,7 @@ build_wcs_upper_buffer (re_string_t *pstr)
if (pstr->offsets == NULL) if (pstr->offsets == NULL)
{ {
pstr->offsets = re_malloc (int, pstr->bufs_len); pstr->offsets = re_malloc (Idx, pstr->bufs_len);
if (pstr->offsets == NULL) if (pstr->offsets == NULL)
return REG_ESPACE; return REG_ESPACE;
@ -483,11 +491,11 @@ build_wcs_upper_buffer (re_string_t *pstr)
/* Skip characters until the index becomes greater than NEW_RAW_IDX. /* Skip characters until the index becomes greater than NEW_RAW_IDX.
Return the index. */ Return the index. */
static int static Idx
re_string_skip_chars (re_string_t *pstr, int new_raw_idx, wint_t *last_wc) re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc)
{ {
mbstate_t prev_st; mbstate_t prev_st;
int rawbuf_idx; Idx rawbuf_idx;
size_t mbclen; size_t mbclen;
wint_t wc = WEOF; wint_t wc = WEOF;
@ -496,11 +504,11 @@ re_string_skip_chars (re_string_t *pstr, int new_raw_idx, wint_t *last_wc)
rawbuf_idx < new_raw_idx;) rawbuf_idx < new_raw_idx;)
{ {
wchar_t wc2; wchar_t wc2;
int remain_len = pstr->raw_len - rawbuf_idx; Idx remain_len = pstr->raw_len - rawbuf_idx;
prev_st = pstr->cur_state; prev_st = pstr->cur_state;
mbclen = __mbrtowc (&wc2, (const char *) pstr->raw_mbs + rawbuf_idx, mbclen = __mbrtowc (&wc2, (const char *) pstr->raw_mbs + rawbuf_idx,
remain_len, &pstr->cur_state); remain_len, &pstr->cur_state);
if (BE ((ssize_t) mbclen <= 0, 0)) if (BE (mbclen == (size_t) -2 || mbclen == (size_t) -1 || mbclen == 0, 0))
{ {
/* We treat these cases as a single byte character. */ /* We treat these cases as a single byte character. */
if (mbclen == 0 || remain_len == 0) if (mbclen == 0 || remain_len == 0)
@ -511,7 +519,7 @@ re_string_skip_chars (re_string_t *pstr, int new_raw_idx, wint_t *last_wc)
pstr->cur_state = prev_st; pstr->cur_state = prev_st;
} }
else else
wc = (wint_t) wc2; wc = wc2;
/* Then proceed the next character. */ /* Then proceed the next character. */
rawbuf_idx += mbclen; rawbuf_idx += mbclen;
} }
@ -526,7 +534,7 @@ re_string_skip_chars (re_string_t *pstr, int new_raw_idx, wint_t *last_wc)
static void static void
build_upper_buffer (re_string_t *pstr) build_upper_buffer (re_string_t *pstr)
{ {
int char_idx, end_idx; Idx char_idx, end_idx;
end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;
for (char_idx = pstr->valid_len; char_idx < end_idx; ++char_idx) for (char_idx = pstr->valid_len; char_idx < end_idx; ++char_idx)
@ -534,10 +542,7 @@ build_upper_buffer (re_string_t *pstr)
int ch = pstr->raw_mbs[pstr->raw_mbs_idx + char_idx]; int ch = pstr->raw_mbs[pstr->raw_mbs_idx + char_idx];
if (BE (pstr->trans != NULL, 0)) if (BE (pstr->trans != NULL, 0))
ch = pstr->trans[ch]; ch = pstr->trans[ch];
if (islower (ch)) pstr->mbs[char_idx] = toupper (ch);
pstr->mbs[char_idx] = toupper (ch);
else
pstr->mbs[char_idx] = ch;
} }
pstr->valid_len = char_idx; pstr->valid_len = char_idx;
pstr->valid_raw_len = char_idx; pstr->valid_raw_len = char_idx;
@ -548,7 +553,7 @@ build_upper_buffer (re_string_t *pstr)
static void static void
re_string_translate_buffer (re_string_t *pstr) re_string_translate_buffer (re_string_t *pstr)
{ {
int buf_idx, end_idx; Idx buf_idx, end_idx;
end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;
for (buf_idx = pstr->valid_len; buf_idx < end_idx; ++buf_idx) for (buf_idx = pstr->valid_len; buf_idx < end_idx; ++buf_idx)
@ -567,10 +572,13 @@ re_string_translate_buffer (re_string_t *pstr)
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_string_reconstruct (re_string_t *pstr, int idx, int eflags) re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags)
{ {
int offset = idx - pstr->raw_mbs_idx; Idx offset;
if (BE (offset < 0, 0))
if (BE (pstr->raw_mbs_idx <= idx, 0))
offset = idx - pstr->raw_mbs_idx;
else
{ {
/* Reset buffer. */ /* Reset buffer. */
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
@ -599,7 +607,7 @@ re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
if (BE (pstr->offsets_needed, 0)) if (BE (pstr->offsets_needed, 0))
{ {
int low = 0, high = pstr->valid_len, mid; Idx low = 0, high = pstr->valid_len, mid;
do do
{ {
mid = (high + low) / 2; mid = (high + low) / 2;
@ -683,7 +691,7 @@ re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
{ {
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
/* No, skip all characters until IDX. */ /* No, skip all characters until IDX. */
int prev_valid_len = pstr->valid_len; Idx prev_valid_len = pstr->valid_len;
if (BE (pstr->offsets_needed, 0)) if (BE (pstr->offsets_needed, 0))
{ {
@ -696,7 +704,7 @@ re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
if (pstr->mb_cur_max > 1) if (pstr->mb_cur_max > 1)
{ {
int wcs_idx; Idx wcs_idx;
wint_t wc = WEOF; wint_t wc = WEOF;
if (pstr->is_utf8) if (pstr->is_utf8)
@ -726,7 +734,7 @@ re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
{ {
mbstate_t cur_state; mbstate_t cur_state;
wchar_t wc2; wchar_t wc2;
int mlen = raw + pstr->len - p; Idx mlen = raw + pstr->len - p;
unsigned char buf[6]; unsigned char buf[6];
size_t mbclen; size_t mbclen;
@ -826,10 +834,11 @@ re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
} }
static unsigned char static unsigned char
__attribute ((pure)) __attribute__ ((pure))
re_string_peek_byte_case (const re_string_t *pstr, int idx) re_string_peek_byte_case (const re_string_t *pstr, Idx idx)
{ {
int ch, off; int ch;
Idx off;
/* Handle the common (easiest) cases first. */ /* Handle the common (easiest) cases first. */
if (BE (!pstr->mbs_allocated, 1)) if (BE (!pstr->mbs_allocated, 1))
@ -870,7 +879,8 @@ re_string_fetch_byte_case (re_string_t *pstr)
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
if (pstr->offsets_needed) if (pstr->offsets_needed)
{ {
int off, ch; Idx off;
int ch;
/* For tr_TR.UTF-8 [[:islower:]] there is /* For tr_TR.UTF-8 [[:islower:]] there is
[[: CAPITAL LETTER I WITH DOT lower:]] in mbs. Skip [[: CAPITAL LETTER I WITH DOT lower:]] in mbs. Skip
@ -911,7 +921,7 @@ re_string_destruct (re_string_t *pstr)
/* Return the context at IDX in INPUT. */ /* Return the context at IDX in INPUT. */
static unsigned int static unsigned int
re_string_context_at (const re_string_t *input, int idx, int eflags) re_string_context_at (const re_string_t *input, Idx idx, int eflags)
{ {
int c; int c;
if (BE (idx < 0, 0)) if (BE (idx < 0, 0))
@ -925,7 +935,7 @@ re_string_context_at (const re_string_t *input, int idx, int eflags)
if (input->mb_cur_max > 1) if (input->mb_cur_max > 1)
{ {
wint_t wc; wint_t wc;
int wc_idx = idx; Idx wc_idx = idx;
while(input->wcs[wc_idx] == WEOF) while(input->wcs[wc_idx] == WEOF)
{ {
#if defined DEBUG && DEBUG #if defined DEBUG && DEBUG
@ -956,23 +966,23 @@ re_string_context_at (const re_string_t *input, int idx, int eflags)
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_alloc (re_node_set *set, int size) re_node_set_alloc (re_node_set *set, Idx size)
{ {
set->alloc = size; set->alloc = size;
set->nelem = 0; set->nelem = 0;
set->elems = re_malloc (int, size); set->elems = re_malloc (Idx, size);
if (BE (set->elems == NULL, 0)) if (BE (set->elems == NULL, 0) && (MALLOC_0_IS_NONNULL || size != 0))
return REG_ESPACE; return REG_ESPACE;
return REG_NOERROR; return REG_NOERROR;
} }
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_init_1 (re_node_set *set, int elem) re_node_set_init_1 (re_node_set *set, Idx elem)
{ {
set->alloc = 1; set->alloc = 1;
set->nelem = 1; set->nelem = 1;
set->elems = re_malloc (int, 1); set->elems = re_malloc (Idx, 1);
if (BE (set->elems == NULL, 0)) if (BE (set->elems == NULL, 0))
{ {
set->alloc = set->nelem = 0; set->alloc = set->nelem = 0;
@ -984,10 +994,10 @@ re_node_set_init_1 (re_node_set *set, int elem)
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_init_2 (re_node_set *set, int elem1, int elem2) re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2)
{ {
set->alloc = 2; set->alloc = 2;
set->elems = re_malloc (int, 2); set->elems = re_malloc (Idx, 2);
if (BE (set->elems == NULL, 0)) if (BE (set->elems == NULL, 0))
return REG_ESPACE; return REG_ESPACE;
if (elem1 == elem2) if (elem1 == elem2)
@ -1020,13 +1030,13 @@ re_node_set_init_copy (re_node_set *dest, const re_node_set *src)
if (src->nelem > 0) if (src->nelem > 0)
{ {
dest->alloc = dest->nelem; dest->alloc = dest->nelem;
dest->elems = re_malloc (int, dest->alloc); dest->elems = re_malloc (Idx, dest->alloc);
if (BE (dest->elems == NULL, 0)) if (BE (dest->elems == NULL, 0))
{ {
dest->alloc = dest->nelem = 0; dest->alloc = dest->nelem = 0;
return REG_ESPACE; return REG_ESPACE;
} }
memcpy (dest->elems, src->elems, src->nelem * sizeof (int)); memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx));
} }
else else
re_node_set_init_empty (dest); re_node_set_init_empty (dest);
@ -1042,7 +1052,7 @@ __attribute_warn_unused_result__
re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1,
const re_node_set *src2) const re_node_set *src2)
{ {
int i1, i2, is, id, delta, sbase; Idx i1, i2, is, id, delta, sbase;
if (src1->nelem == 0 || src2->nelem == 0) if (src1->nelem == 0 || src2->nelem == 0)
return REG_NOERROR; return REG_NOERROR;
@ -1050,8 +1060,8 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1,
conservative estimate. */ conservative estimate. */
if (src1->nelem + src2->nelem + dest->nelem > dest->alloc) if (src1->nelem + src2->nelem + dest->nelem > dest->alloc)
{ {
int new_alloc = src1->nelem + src2->nelem + dest->alloc; Idx new_alloc = src1->nelem + src2->nelem + dest->alloc;
int *new_elems = re_realloc (dest->elems, int, new_alloc); Idx *new_elems = re_realloc (dest->elems, Idx, new_alloc);
if (BE (new_elems == NULL, 0)) if (BE (new_elems == NULL, 0))
return REG_ESPACE; return REG_ESPACE;
dest->elems = new_elems; dest->elems = new_elems;
@ -1073,7 +1083,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1,
--id; --id;
if (id < 0 || dest->elems[id] != src1->elems[i1]) if (id < 0 || dest->elems[id] != src1->elems[i1])
dest->elems[--sbase] = src1->elems[i1]; dest->elems[--sbase] = src1->elems[i1];
if (--i1 < 0 || --i2 < 0) if (--i1 < 0 || --i2 < 0)
break; break;
@ -1120,7 +1130,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1,
} }
/* Copy remaining SRC elements. */ /* Copy remaining SRC elements. */
memcpy (dest->elems, dest->elems + sbase, delta * sizeof (int)); memcpy (dest->elems, dest->elems + sbase, delta * sizeof (Idx));
return REG_NOERROR; return REG_NOERROR;
} }
@ -1133,11 +1143,11 @@ __attribute_warn_unused_result__
re_node_set_init_union (re_node_set *dest, const re_node_set *src1, re_node_set_init_union (re_node_set *dest, const re_node_set *src1,
const re_node_set *src2) const re_node_set *src2)
{ {
int i1, i2, id; Idx i1, i2, id;
if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0) if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0)
{ {
dest->alloc = src1->nelem + src2->nelem; dest->alloc = src1->nelem + src2->nelem;
dest->elems = re_malloc (int, dest->alloc); dest->elems = re_malloc (Idx, dest->alloc);
if (BE (dest->elems == NULL, 0)) if (BE (dest->elems == NULL, 0))
return REG_ESPACE; return REG_ESPACE;
} }
@ -1165,13 +1175,13 @@ re_node_set_init_union (re_node_set *dest, const re_node_set *src1,
if (i1 < src1->nelem) if (i1 < src1->nelem)
{ {
memcpy (dest->elems + id, src1->elems + i1, memcpy (dest->elems + id, src1->elems + i1,
(src1->nelem - i1) * sizeof (int)); (src1->nelem - i1) * sizeof (Idx));
id += src1->nelem - i1; id += src1->nelem - i1;
} }
else if (i2 < src2->nelem) else if (i2 < src2->nelem)
{ {
memcpy (dest->elems + id, src2->elems + i2, memcpy (dest->elems + id, src2->elems + i2,
(src2->nelem - i2) * sizeof (int)); (src2->nelem - i2) * sizeof (Idx));
id += src2->nelem - i2; id += src2->nelem - i2;
} }
dest->nelem = id; dest->nelem = id;
@ -1185,13 +1195,13 @@ static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_merge (re_node_set *dest, const re_node_set *src) re_node_set_merge (re_node_set *dest, const re_node_set *src)
{ {
int is, id, sbase, delta; Idx is, id, sbase, delta;
if (src == NULL || src->nelem == 0) if (src == NULL || src->nelem == 0)
return REG_NOERROR; return REG_NOERROR;
if (dest->alloc < 2 * src->nelem + dest->nelem) if (dest->alloc < 2 * src->nelem + dest->nelem)
{ {
int new_alloc = 2 * (src->nelem + dest->alloc); Idx new_alloc = 2 * (src->nelem + dest->alloc);
int *new_buffer = re_realloc (dest->elems, int, new_alloc); Idx *new_buffer = re_realloc (dest->elems, Idx, new_alloc);
if (BE (new_buffer == NULL, 0)) if (BE (new_buffer == NULL, 0))
return REG_ESPACE; return REG_ESPACE;
dest->elems = new_buffer; dest->elems = new_buffer;
@ -1201,7 +1211,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src)
if (BE (dest->nelem == 0, 0)) if (BE (dest->nelem == 0, 0))
{ {
dest->nelem = src->nelem; dest->nelem = src->nelem;
memcpy (dest->elems, src->elems, src->nelem * sizeof (int)); memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx));
return REG_NOERROR; return REG_NOERROR;
} }
@ -1222,7 +1232,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src)
{ {
/* If DEST is exhausted, the remaining items of SRC must be unique. */ /* If DEST is exhausted, the remaining items of SRC must be unique. */
sbase -= is + 1; sbase -= is + 1;
memcpy (dest->elems + sbase, src->elems, (is + 1) * sizeof (int)); memcpy (dest->elems + sbase, src->elems, (is + 1) * sizeof (Idx));
} }
id = dest->nelem - 1; id = dest->nelem - 1;
@ -1251,7 +1261,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src)
{ {
/* Copy remaining SRC elements. */ /* Copy remaining SRC elements. */
memcpy (dest->elems, dest->elems + sbase, memcpy (dest->elems, dest->elems + sbase,
delta * sizeof (int)); delta * sizeof (Idx));
break; break;
} }
} }
@ -1262,38 +1272,33 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src)
/* Insert the new element ELEM to the re_node_set* SET. /* Insert the new element ELEM to the re_node_set* SET.
SET should not already have ELEM. SET should not already have ELEM.
return -1 if an error is occured, return 1 otherwise. */ Return true if successful. */
static int static bool
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_insert (re_node_set *set, int elem) re_node_set_insert (re_node_set *set, Idx elem)
{ {
int idx; Idx idx;
/* In case the set is empty. */ /* In case the set is empty. */
if (set->alloc == 0) if (set->alloc == 0)
{ return BE (re_node_set_init_1 (set, elem) == REG_NOERROR, 1);
if (BE (re_node_set_init_1 (set, elem) == REG_NOERROR, 1))
return 1;
else
return -1;
}
if (BE (set->nelem, 0) == 0) if (BE (set->nelem, 0) == 0)
{ {
/* We already guaranteed above that set->alloc != 0. */ /* We already guaranteed above that set->alloc != 0. */
set->elems[0] = elem; set->elems[0] = elem;
++set->nelem; ++set->nelem;
return 1; return true;
} }
/* Realloc if we need. */ /* Realloc if we need. */
if (set->alloc == set->nelem) if (set->alloc == set->nelem)
{ {
int *new_elems; Idx *new_elems;
set->alloc = set->alloc * 2; set->alloc = set->alloc * 2;
new_elems = re_realloc (set->elems, int, set->alloc); new_elems = re_realloc (set->elems, Idx, set->alloc);
if (BE (new_elems == NULL, 0)) if (BE (new_elems == NULL, 0))
return -1; return false;
set->elems = new_elems; set->elems = new_elems;
} }
@ -1314,56 +1319,56 @@ re_node_set_insert (re_node_set *set, int elem)
/* Insert the new element. */ /* Insert the new element. */
set->elems[idx] = elem; set->elems[idx] = elem;
++set->nelem; ++set->nelem;
return 1; return true;
} }
/* Insert the new element ELEM to the re_node_set* SET. /* Insert the new element ELEM to the re_node_set* SET.
SET should not already have any element greater than or equal to ELEM. SET should not already have any element greater than or equal to ELEM.
Return -1 if an error is occured, return 1 otherwise. */ Return true if successful. */
static int static bool
__attribute_warn_unused_result__ __attribute_warn_unused_result__
re_node_set_insert_last (re_node_set *set, int elem) re_node_set_insert_last (re_node_set *set, Idx elem)
{ {
/* Realloc if we need. */ /* Realloc if we need. */
if (set->alloc == set->nelem) if (set->alloc == set->nelem)
{ {
int *new_elems; Idx *new_elems;
set->alloc = (set->alloc + 1) * 2; set->alloc = (set->alloc + 1) * 2;
new_elems = re_realloc (set->elems, int, set->alloc); new_elems = re_realloc (set->elems, Idx, set->alloc);
if (BE (new_elems == NULL, 0)) if (BE (new_elems == NULL, 0))
return -1; return false;
set->elems = new_elems; set->elems = new_elems;
} }
/* Insert the new element. */ /* Insert the new element. */
set->elems[set->nelem++] = elem; set->elems[set->nelem++] = elem;
return 1; return true;
} }
/* Compare two node sets SET1 and SET2. /* Compare two node sets SET1 and SET2.
return 1 if SET1 and SET2 are equivalent, return 0 otherwise. */ Return true if SET1 and SET2 are equivalent. */
static int static bool
__attribute ((pure)) __attribute__ ((pure))
re_node_set_compare (const re_node_set *set1, const re_node_set *set2) re_node_set_compare (const re_node_set *set1, const re_node_set *set2)
{ {
int i; Idx i;
if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem) if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem)
return 0; return false;
for (i = set1->nelem ; --i >= 0 ; ) for (i = set1->nelem ; --i >= 0 ; )
if (set1->elems[i] != set2->elems[i]) if (set1->elems[i] != set2->elems[i])
return 0; return false;
return 1; return true;
} }
/* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */ /* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */
static int static Idx
__attribute ((pure)) __attribute__ ((pure))
re_node_set_contains (const re_node_set *set, int elem) re_node_set_contains (const re_node_set *set, Idx elem)
{ {
unsigned int idx, right, mid; __re_size_t idx, right, mid;
if (set->nelem <= 0) if (set->nelem <= 0)
return 0; return 0;
@ -1382,7 +1387,7 @@ re_node_set_contains (const re_node_set *set, int elem)
} }
static void static void
re_node_set_remove_at (re_node_set *set, int idx) re_node_set_remove_at (re_node_set *set, Idx idx)
{ {
if (idx < 0 || idx >= set->nelem) if (idx < 0 || idx >= set->nelem)
return; return;
@ -1393,37 +1398,42 @@ re_node_set_remove_at (re_node_set *set, int idx)
/* Add the token TOKEN to dfa->nodes, and return the index of the token. /* Add the token TOKEN to dfa->nodes, and return the index of the token.
Or return -1, if an error will be occured. */ Or return -1 if an error occurred. */
static int static Idx
re_dfa_add_node (re_dfa_t *dfa, re_token_t token) re_dfa_add_node (re_dfa_t *dfa, re_token_t token)
{ {
int type = token.type;
if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0)) if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0))
{ {
size_t new_nodes_alloc = dfa->nodes_alloc * 2; size_t new_nodes_alloc = dfa->nodes_alloc * 2;
int *new_nexts, *new_indices; Idx *new_nexts, *new_indices;
re_node_set *new_edests, *new_eclosures; re_node_set *new_edests, *new_eclosures;
re_token_t *new_nodes; re_token_t *new_nodes;
/* Avoid overflows in realloc. */ /* Avoid overflows in realloc. */
const size_t max_object_size = MAX (sizeof (re_token_t), const size_t max_object_size = MAX (sizeof (re_token_t),
MAX (sizeof (re_node_set), MAX (sizeof (re_node_set),
sizeof (int))); sizeof (Idx)));
if (BE (SIZE_MAX / max_object_size < new_nodes_alloc, 0)) if (BE (MIN (IDX_MAX, SIZE_MAX / max_object_size) < new_nodes_alloc, 0))
return -1; return -1;
new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc); new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc);
if (BE (new_nodes == NULL, 0)) if (BE (new_nodes == NULL, 0))
return -1; return -1;
dfa->nodes = new_nodes; dfa->nodes = new_nodes;
new_nexts = re_realloc (dfa->nexts, int, new_nodes_alloc); new_nexts = re_realloc (dfa->nexts, Idx, new_nodes_alloc);
new_indices = re_realloc (dfa->org_indices, int, new_nodes_alloc); new_indices = re_realloc (dfa->org_indices, Idx, new_nodes_alloc);
new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc); new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc);
new_eclosures = re_realloc (dfa->eclosures, re_node_set, new_nodes_alloc); new_eclosures = re_realloc (dfa->eclosures, re_node_set, new_nodes_alloc);
if (BE (new_nexts == NULL || new_indices == NULL if (BE (new_nexts == NULL || new_indices == NULL
|| new_edests == NULL || new_eclosures == NULL, 0)) || new_edests == NULL || new_eclosures == NULL, 0))
return -1; {
re_free (new_nexts);
re_free (new_indices);
re_free (new_edests);
re_free (new_eclosures);
return -1;
}
dfa->nexts = new_nexts; dfa->nexts = new_nexts;
dfa->org_indices = new_indices; dfa->org_indices = new_indices;
dfa->edests = new_edests; dfa->edests = new_edests;
@ -1434,7 +1444,8 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token)
dfa->nodes[dfa->nodes_len].constraint = 0; dfa->nodes[dfa->nodes_len].constraint = 0;
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
dfa->nodes[dfa->nodes_len].accept_mb = dfa->nodes[dfa->nodes_len].accept_mb =
(type == OP_PERIOD && dfa->mb_cur_max > 1) || type == COMPLEX_BRACKET; ((token.type == OP_PERIOD && dfa->mb_cur_max > 1)
|| token.type == COMPLEX_BRACKET);
#endif #endif
dfa->nexts[dfa->nodes_len] = -1; dfa->nexts[dfa->nodes_len] = -1;
re_node_set_init_empty (dfa->edests + dfa->nodes_len); re_node_set_init_empty (dfa->edests + dfa->nodes_len);
@ -1442,11 +1453,11 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token)
return dfa->nodes_len++; return dfa->nodes_len++;
} }
static inline unsigned int static re_hashval_t
calc_state_hash (const re_node_set *nodes, unsigned int context) calc_state_hash (const re_node_set *nodes, unsigned int context)
{ {
unsigned int hash = nodes->nelem + context; re_hashval_t hash = nodes->nelem + context;
int i; Idx i;
for (i = 0 ; i < nodes->nelem ; i++) for (i = 0 ; i < nodes->nelem ; i++)
hash += nodes->elems[i]; hash += nodes->elems[i];
return hash; return hash;
@ -1466,10 +1477,14 @@ __attribute_warn_unused_result__
re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa,
const re_node_set *nodes) const re_node_set *nodes)
{ {
unsigned int hash; re_hashval_t hash;
re_dfastate_t *new_state; re_dfastate_t *new_state;
struct re_state_table_entry *spot; struct re_state_table_entry *spot;
int i; Idx i;
#if defined GCC_LINT || defined lint
/* Suppress bogus uninitialized-variable warnings. */
*err = REG_NOERROR;
#endif
if (BE (nodes->nelem == 0, 0)) if (BE (nodes->nelem == 0, 0))
{ {
*err = REG_NOERROR; *err = REG_NOERROR;
@ -1510,10 +1525,14 @@ __attribute_warn_unused_result__
re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa, re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa,
const re_node_set *nodes, unsigned int context) const re_node_set *nodes, unsigned int context)
{ {
unsigned int hash; re_hashval_t hash;
re_dfastate_t *new_state; re_dfastate_t *new_state;
struct re_state_table_entry *spot; struct re_state_table_entry *spot;
int i; Idx i;
#if defined GCC_LINT || defined lint
/* Suppress bogus uninitialized-variable warnings. */
*err = REG_NOERROR;
#endif
if (nodes->nelem == 0) if (nodes->nelem == 0)
{ {
*err = REG_NOERROR; *err = REG_NOERROR;
@ -1530,7 +1549,7 @@ re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa,
&& re_node_set_compare (state->entrance_nodes, nodes)) && re_node_set_compare (state->entrance_nodes, nodes))
return state; return state;
} }
/* There are no appropriate state in `dfa', create the new one. */ /* There are no appropriate state in 'dfa', create the new one. */
new_state = create_cd_newstate (dfa, nodes, context, hash); new_state = create_cd_newstate (dfa, nodes, context, hash);
if (BE (new_state == NULL, 0)) if (BE (new_state == NULL, 0))
*err = REG_ESPACE; *err = REG_ESPACE;
@ -1545,11 +1564,11 @@ re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa,
static reg_errcode_t static reg_errcode_t
__attribute_warn_unused_result__ __attribute_warn_unused_result__
register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, register_state (const re_dfa_t *dfa, re_dfastate_t *newstate,
unsigned int hash) re_hashval_t hash)
{ {
struct re_state_table_entry *spot; struct re_state_table_entry *spot;
reg_errcode_t err; reg_errcode_t err;
int i; Idx i;
newstate->hash = hash; newstate->hash = hash;
err = re_node_set_alloc (&newstate->non_eps_nodes, newstate->nodes.nelem); err = re_node_set_alloc (&newstate->non_eps_nodes, newstate->nodes.nelem);
@ -1557,16 +1576,16 @@ register_state (const re_dfa_t *dfa, re_dfastate_t *newstate,
return REG_ESPACE; return REG_ESPACE;
for (i = 0; i < newstate->nodes.nelem; i++) for (i = 0; i < newstate->nodes.nelem; i++)
{ {
int elem = newstate->nodes.elems[i]; Idx elem = newstate->nodes.elems[i];
if (!IS_EPSILON_NODE (dfa->nodes[elem].type)) if (!IS_EPSILON_NODE (dfa->nodes[elem].type))
if (re_node_set_insert_last (&newstate->non_eps_nodes, elem) < 0) if (! re_node_set_insert_last (&newstate->non_eps_nodes, elem))
return REG_ESPACE; return REG_ESPACE;
} }
spot = dfa->state_table + (hash & dfa->state_hash_mask); spot = dfa->state_table + (hash & dfa->state_hash_mask);
if (BE (spot->alloc <= spot->num, 0)) if (BE (spot->alloc <= spot->num, 0))
{ {
int new_alloc = 2 * spot->num + 2; Idx new_alloc = 2 * spot->num + 2;
re_dfastate_t **new_array = re_realloc (spot->array, re_dfastate_t *, re_dfastate_t **new_array = re_realloc (spot->array, re_dfastate_t *,
new_alloc); new_alloc);
if (BE (new_array == NULL, 0)) if (BE (new_array == NULL, 0))
@ -1600,9 +1619,9 @@ free_state (re_dfastate_t *state)
static re_dfastate_t * static re_dfastate_t *
__attribute_warn_unused_result__ __attribute_warn_unused_result__
create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes,
unsigned int hash) re_hashval_t hash)
{ {
int i; Idx i;
reg_errcode_t err; reg_errcode_t err;
re_dfastate_t *newstate; re_dfastate_t *newstate;
@ -1650,9 +1669,9 @@ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes,
static re_dfastate_t * static re_dfastate_t *
__attribute_warn_unused_result__ __attribute_warn_unused_result__
create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes,
unsigned int context, unsigned int hash) unsigned int context, re_hashval_t hash)
{ {
int i, nctx_nodes = 0; Idx i, nctx_nodes = 0;
reg_errcode_t err; reg_errcode_t err;
re_dfastate_t *newstate; re_dfastate_t *newstate;

446
posix/regex_internal.h
View File

@ -15,7 +15,7 @@
You should have received a copy of the GNU Lesser General Public You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */ <https://www.gnu.org/licenses/>. */
#ifndef _REGEX_INTERNAL_H #ifndef _REGEX_INTERNAL_H
#define _REGEX_INTERNAL_H 1 #define _REGEX_INTERNAL_H 1
@ -26,35 +26,78 @@
#include <stdlib.h> #include <stdlib.h>
#include <string.h> #include <string.h>
#if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC #include <langinfo.h>
# include <langinfo.h> #include <locale.h>
#endif #include <wchar.h>
#if defined HAVE_LOCALE_H || defined _LIBC #include <wctype.h>
# include <locale.h> #include <stdbool.h>
#endif #include <stdint.h>
#if defined HAVE_WCHAR_H || defined _LIBC
# include <wchar.h> /* Properties of integers. Although Gnulib has intprops.h, glibc does
#endif /* HAVE_WCHAR_H || _LIBC */ without for now. */
#if defined HAVE_WCTYPE_H || defined _LIBC #ifndef _LIBC
# include <wctype.h> # include "intprops.h"
#endif /* HAVE_WCTYPE_H || _LIBC */
#if defined HAVE_STDBOOL_H || defined _LIBC
# include <stdbool.h>
#endif /* HAVE_STDBOOL_H || _LIBC */
#if defined HAVE_STDINT_H || defined _LIBC
# include <stdint.h>
#endif /* HAVE_STDINT_H || _LIBC */
#if defined _LIBC
# include <libc-lock.h>
#else #else
# define __libc_lock_define(CLASS,NAME) /* True if the real type T is signed. */
# define __libc_lock_init(NAME) do { } while (0) # define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
# define __libc_lock_lock(NAME) do { } while (0)
# define __libc_lock_unlock(NAME) do { } while (0) /* True if adding the nonnegative Idx values A and B would overflow.
If false, set *R to A + B. A, B, and R may be evaluated more than
once, or zero times. Although this is not a full implementation of
Gnulib INT_ADD_WRAPV, it is good enough for glibc regex code.
FIXME: This implementation is a fragile stopgap, and this file would
be simpler and more robust if intprops.h were migrated into glibc. */
# define INT_ADD_WRAPV(a, b, r) \
(IDX_MAX - (a) < (b) ? true : (*(r) = (a) + (b), false))
#endif
#ifdef _LIBC
# include <libc-lock.h>
# define lock_define(name) __libc_lock_define (, name)
# define lock_init(lock) (__libc_lock_init (lock), 0)
# define lock_fini(lock) ((void) 0)
# define lock_lock(lock) __libc_lock_lock (lock)
# define lock_unlock(lock) __libc_lock_unlock (lock)
#elif defined GNULIB_LOCK && !defined USE_UNLOCKED_IO
# include "glthread/lock.h"
/* Use gl_lock_define if empty macro arguments are known to work.
Otherwise, fall back on less-portable substitutes. */
# if ((defined __GNUC__ && !defined __STRICT_ANSI__) \
|| (defined __STDC_VERSION__ && 199901L <= __STDC_VERSION__))
# define lock_define(name) gl_lock_define (, name)
# elif USE_POSIX_THREADS
# define lock_define(name) pthread_mutex_t name;
# elif USE_PTH_THREADS
# define lock_define(name) pth_mutex_t name;
# elif USE_SOLARIS_THREADS
# define lock_define(name) mutex_t name;
# elif USE_WINDOWS_THREADS
# define lock_define(name) gl_lock_t name;
# else
# define lock_define(name)
# endif
# define lock_init(lock) glthread_lock_init (&(lock))
# define lock_fini(lock) glthread_lock_destroy (&(lock))
# define lock_lock(lock) glthread_lock_lock (&(lock))
# define lock_unlock(lock) glthread_lock_unlock (&(lock))
#elif defined GNULIB_PTHREAD && !defined USE_UNLOCKED_IO
# include <pthread.h>
# define lock_define(name) pthread_mutex_t name;
# define lock_init(lock) pthread_mutex_init (&(lock), 0)
# define lock_fini(lock) pthread_mutex_destroy (&(lock))
# define lock_lock(lock) pthread_mutex_lock (&(lock))
# define lock_unlock(lock) pthread_mutex_unlock (&(lock))
#else
# define lock_define(name)
# define lock_init(lock) 0
# define lock_fini(lock) ((void) 0)
/* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
# define lock_lock(lock) ((void) dfa)
# define lock_unlock(lock) ((void) 0)
#endif #endif
/* In case that the system doesn't have isblank(). */ /* In case that the system doesn't have isblank(). */
#if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
# define isblank(ch) ((ch) == ' ' || (ch) == '\t') # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
#endif #endif
@ -75,6 +118,7 @@
__dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES) __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
# endif # endif
#else #else
# undef gettext
# define gettext(msgid) (msgid) # define gettext(msgid) (msgid)
#endif #endif
@ -84,23 +128,17 @@
# define gettext_noop(String) String # define gettext_noop(String) String
#endif #endif
/* For loser systems without the definition. */
#ifndef SIZE_MAX
# define SIZE_MAX ((size_t) -1)
#endif
#if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC
# define RE_ENABLE_I18N # define RE_ENABLE_I18N
#endif #endif
#if __GNUC__ >= 3 #define BE(expr, val) __builtin_expect (expr, val)
# define BE(expr, val) __builtin_expect (expr, val)
#else
# define BE(expr, val) (expr)
#endif
/* Number of single byte character. */ /* Number of ASCII characters. */
#define SBC_MAX 256 #define ASCII_CHARS 0x80
/* Number of single byte characters. */
#define SBC_MAX (UCHAR_MAX + 1)
#define COLL_ELEM_LEN_MAX 8 #define COLL_ELEM_LEN_MAX 8
@ -110,11 +148,15 @@
/* Rename to standard API for using out of glibc. */ /* Rename to standard API for using out of glibc. */
#ifndef _LIBC #ifndef _LIBC
# undef __wctype
# undef __iswctype
# define __wctype wctype # define __wctype wctype
# define __iswalnum iswalnum
# define __iswctype iswctype # define __iswctype iswctype
# define __towlower towlower
# define __towupper towupper
# define __btowc btowc # define __btowc btowc
# define __mbrtowc mbrtowc # define __mbrtowc mbrtowc
# define __mempcpy mempcpy
# define __wcrtomb wcrtomb # define __wcrtomb wcrtomb
# define __regfree regfree # define __regfree regfree
# define attribute_hidden # define attribute_hidden
@ -124,32 +166,70 @@
# define __attribute__(arg) # define __attribute__(arg)
#endif #endif
extern const char __re_error_msgid[] attribute_hidden; #ifndef SSIZE_MAX
extern const size_t __re_error_msgid_idx[] attribute_hidden; # define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
#endif
/* The type of indexes into strings. This is signed, not size_t,
since the API requires indexes to fit in regoff_t anyway, and using
signed integers makes the code a bit smaller and presumably faster.
The traditional GNU regex implementation uses int for indexes.
The POSIX-compatible implementation uses a possibly-wider type.
The name 'Idx' is three letters to minimize the hassle of
reindenting a lot of regex code that formerly used 'int'. */
typedef regoff_t Idx;
#ifdef _REGEX_LARGE_OFFSETS
# define IDX_MAX SSIZE_MAX
#else
# define IDX_MAX INT_MAX
#endif
/* A hash value, suitable for computing hash tables. */
typedef __re_size_t re_hashval_t;
/* An integer used to represent a set of bits. It must be unsigned, /* An integer used to represent a set of bits. It must be unsigned,
and must be at least as wide as unsigned int. */ and must be at least as wide as unsigned int. */
typedef unsigned long int bitset_word_t; typedef unsigned long int bitset_word_t;
/* All bits set in a bitset_word_t. */ /* All bits set in a bitset_word_t. */
#define BITSET_WORD_MAX ULONG_MAX #define BITSET_WORD_MAX ULONG_MAX
/* Number of bits in a bitset_word_t. */
#define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT) /* Number of bits in a bitset_word_t. For portability to hosts with
/* Number of bitset_word_t in a bit_set. */ padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
#define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS) instead, deduce it directly from BITSET_WORD_MAX. Avoid
greater-than-32-bit integers and unconditional shifts by more than
31 bits, as they're not portable. */
#if BITSET_WORD_MAX == 0xffffffffUL
# define BITSET_WORD_BITS 32
#elif BITSET_WORD_MAX >> 31 >> 4 == 1
# define BITSET_WORD_BITS 36
#elif BITSET_WORD_MAX >> 31 >> 16 == 1
# define BITSET_WORD_BITS 48
#elif BITSET_WORD_MAX >> 31 >> 28 == 1
# define BITSET_WORD_BITS 60
#elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
# define BITSET_WORD_BITS 64
#elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
# define BITSET_WORD_BITS 72
#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
# define BITSET_WORD_BITS 128
#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
# define BITSET_WORD_BITS 256
#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
# define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
# if BITSET_WORD_BITS <= SBC_MAX
# error "Invalid SBC_MAX"
# endif
#else
# error "Add case for new bitset_word_t size"
#endif
/* Number of bitset_word_t values in a bitset_t. */
#define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
typedef bitset_word_t bitset_t[BITSET_WORDS]; typedef bitset_word_t bitset_t[BITSET_WORDS];
typedef bitset_word_t *re_bitset_ptr_t; typedef bitset_word_t *re_bitset_ptr_t;
typedef const bitset_word_t *re_const_bitset_ptr_t; typedef const bitset_word_t *re_const_bitset_ptr_t;
#define bitset_set(set,i) \
(set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
#define bitset_clear(set,i) \
(set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
#define bitset_contain(set,i) \
(set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
#define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
#define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
#define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
#define PREV_WORD_CONSTRAINT 0x0001 #define PREV_WORD_CONSTRAINT 0x0001
#define PREV_NOTWORD_CONSTRAINT 0x0002 #define PREV_NOTWORD_CONSTRAINT 0x0002
#define NEXT_WORD_CONSTRAINT 0x0004 #define NEXT_WORD_CONSTRAINT 0x0004
@ -177,9 +257,9 @@ typedef enum
typedef struct typedef struct
{ {
int alloc; Idx alloc;
int nelem; Idx nelem;
int *elems; Idx *elems;
} re_node_set; } re_node_set;
typedef enum typedef enum
@ -265,19 +345,19 @@ typedef struct
unsigned int non_match : 1; unsigned int non_match : 1;
/* # of multibyte characters. */ /* # of multibyte characters. */
int nmbchars; Idx nmbchars;
/* # of collating symbols. */ /* # of collating symbols. */
int ncoll_syms; Idx ncoll_syms;
/* # of equivalence classes. */ /* # of equivalence classes. */
int nequiv_classes; Idx nequiv_classes;
/* # of range expressions. */ /* # of range expressions. */
int nranges; Idx nranges;
/* # of character classes. */ /* # of character classes. */
int nchar_classes; Idx nchar_classes;
} re_charset_t; } re_charset_t;
#endif /* RE_ENABLE_I18N */ #endif /* RE_ENABLE_I18N */
@ -290,10 +370,10 @@ typedef struct
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
re_charset_t *mbcset; /* for COMPLEX_BRACKET */ re_charset_t *mbcset; /* for COMPLEX_BRACKET */
#endif /* RE_ENABLE_I18N */ #endif /* RE_ENABLE_I18N */
int idx; /* for BACK_REF */ Idx idx; /* for BACK_REF */
re_context_type ctx_type; /* for ANCHOR */ re_context_type ctx_type; /* for ANCHOR */
} opr; } opr;
#if __GNUC__ >= 2 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
re_token_type_t type : 8; re_token_type_t type : 8;
#else #else
re_token_type_t type; re_token_type_t type;
@ -324,30 +404,30 @@ struct re_string_t
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
/* Store the wide character string which is corresponding to MBS. */ /* Store the wide character string which is corresponding to MBS. */
wint_t *wcs; wint_t *wcs;
int *offsets; Idx *offsets;
mbstate_t cur_state; mbstate_t cur_state;
#endif #endif
/* Index in RAW_MBS. Each character mbs[i] corresponds to /* Index in RAW_MBS. Each character mbs[i] corresponds to
raw_mbs[raw_mbs_idx + i]. */ raw_mbs[raw_mbs_idx + i]. */
int raw_mbs_idx; Idx raw_mbs_idx;
/* The length of the valid characters in the buffers. */ /* The length of the valid characters in the buffers. */
int valid_len; Idx valid_len;
/* The corresponding number of bytes in raw_mbs array. */ /* The corresponding number of bytes in raw_mbs array. */
int valid_raw_len; Idx valid_raw_len;
/* The length of the buffers MBS and WCS. */ /* The length of the buffers MBS and WCS. */
int bufs_len; Idx bufs_len;
/* The index in MBS, which is updated by re_string_fetch_byte. */ /* The index in MBS, which is updated by re_string_fetch_byte. */
int cur_idx; Idx cur_idx;
/* length of RAW_MBS array. */ /* length of RAW_MBS array. */
int raw_len; Idx raw_len;
/* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
int len; Idx len;
/* End of the buffer may be shorter than its length in the cases such /* End of the buffer may be shorter than its length in the cases such
as re_match_2, re_search_2. Then, we use STOP for end of the buffer as re_match_2, re_search_2. Then, we use STOP for end of the buffer
instead of LEN. */ instead of LEN. */
int raw_stop; Idx raw_stop;
/* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
int stop; Idx stop;
/* The context of mbs[0]. We store the context independently, since /* The context of mbs[0]. We store the context independently, since
the context of mbs[0] may be different from raw_mbs[0], which is the context of mbs[0] may be different from raw_mbs[0], which is
@ -357,7 +437,7 @@ struct re_string_t
RE_TRANSLATE_TYPE trans; RE_TRANSLATE_TYPE trans;
/* Copy of re_dfa_t's word_char. */ /* Copy of re_dfa_t's word_char. */
re_const_bitset_ptr_t word_char; re_const_bitset_ptr_t word_char;
/* 1 if REG_ICASE. */ /* true if REG_ICASE. */
unsigned char icase; unsigned char icase;
unsigned char is_utf8; unsigned char is_utf8;
unsigned char map_notascii; unsigned char map_notascii;
@ -373,18 +453,10 @@ typedef struct re_string_t re_string_t;
struct re_dfa_t; struct re_dfa_t;
typedef struct re_dfa_t re_dfa_t; typedef struct re_dfa_t re_dfa_t;
#if IS_IN (libc) #ifndef _LIBC
static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, # define IS_IN(libc) false
int new_buf_len);
# ifdef RE_ENABLE_I18N
static void build_wcs_buffer (re_string_t *pstr);
static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr);
# endif /* RE_ENABLE_I18N */
static void build_upper_buffer (re_string_t *pstr);
static void re_string_translate_buffer (re_string_t *pstr);
static unsigned int re_string_context_at (const re_string_t *input, int idx,
int eflags) __attribute__ ((pure));
#endif #endif
#define re_string_peek_byte(pstr, offset) \ #define re_string_peek_byte(pstr, offset) \
((pstr)->mbs[(pstr)->cur_idx + offset]) ((pstr)->mbs[(pstr)->cur_idx + offset])
#define re_string_fetch_byte(pstr) \ #define re_string_fetch_byte(pstr) \
@ -402,7 +474,9 @@ static unsigned int re_string_context_at (const re_string_t *input, int idx,
#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
#include <alloca.h> #if defined _LIBC || HAVE_ALLOCA
# include <alloca.h>
#endif
#ifndef _LIBC #ifndef _LIBC
# if HAVE_ALLOCA # if HAVE_ALLOCA
@ -414,9 +488,24 @@ static unsigned int re_string_context_at (const re_string_t *input, int idx,
# else # else
/* alloca is implemented with malloc, so just use malloc. */ /* alloca is implemented with malloc, so just use malloc. */
# define __libc_use_alloca(n) 0 # define __libc_use_alloca(n) 0
# undef alloca
# define alloca(n) malloc (n)
# endif # endif
#endif #endif
#ifdef _LIBC
# define MALLOC_0_IS_NONNULL 1
#elif !defined MALLOC_0_IS_NONNULL
# define MALLOC_0_IS_NONNULL 0
#endif
#ifndef MAX
# define MAX(a,b) ((a) < (b) ? (b) : (a))
#endif
#ifndef MIN
# define MIN(a,b) ((a) < (b) ? (a) : (b))
#endif
#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
#define re_free(p) free (p) #define re_free(p) free (p)
@ -431,9 +520,9 @@ struct bin_tree_t
re_token_t token; re_token_t token;
/* `node_idx' is the index in dfa->nodes, if `type' == 0. /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
Otherwise `type' indicate the type of this node. */ Otherwise 'type' indicate the type of this node. */
int node_idx; Idx node_idx;
}; };
typedef struct bin_tree_t bin_tree_t; typedef struct bin_tree_t bin_tree_t;
@ -477,7 +566,7 @@ typedef struct bin_tree_storage_t bin_tree_storage_t;
struct re_dfastate_t struct re_dfastate_t
{ {
unsigned int hash; re_hashval_t hash;
re_node_set nodes; re_node_set nodes;
re_node_set non_eps_nodes; re_node_set non_eps_nodes;
re_node_set inveclosure; re_node_set inveclosure;
@ -485,9 +574,9 @@ struct re_dfastate_t
struct re_dfastate_t **trtable, **word_trtable; struct re_dfastate_t **trtable, **word_trtable;
unsigned int context : 4; unsigned int context : 4;
unsigned int halt : 1; unsigned int halt : 1;
/* If this state can accept `multi byte'. /* If this state can accept "multi byte".
Note that we refer to multibyte characters, and multi character Note that we refer to multibyte characters, and multi character
collating elements as `multi byte'. */ collating elements as "multi byte". */
unsigned int accept_mb : 1; unsigned int accept_mb : 1;
/* If this state has backreference node(s). */ /* If this state has backreference node(s). */
unsigned int has_backref : 1; unsigned int has_backref : 1;
@ -497,8 +586,8 @@ typedef struct re_dfastate_t re_dfastate_t;
struct re_state_table_entry struct re_state_table_entry
{ {
int num; Idx num;
int alloc; Idx alloc;
re_dfastate_t **array; re_dfastate_t **array;
}; };
@ -506,8 +595,8 @@ struct re_state_table_entry
typedef struct typedef struct
{ {
int next_idx; Idx next_idx;
int alloc; Idx alloc;
re_dfastate_t **array; re_dfastate_t **array;
} state_array_t; } state_array_t;
@ -515,8 +604,8 @@ typedef struct
typedef struct typedef struct
{ {
int node; Idx node;
int str_idx; /* The position NODE match at. */ Idx str_idx; /* The position NODE match at. */
state_array_t path; state_array_t path;
} re_sub_match_last_t; } re_sub_match_last_t;
@ -526,20 +615,20 @@ typedef struct
typedef struct typedef struct
{ {
int str_idx; Idx str_idx;
int node; Idx node;
state_array_t *path; state_array_t *path;
int alasts; /* Allocation size of LASTS. */ Idx alasts; /* Allocation size of LASTS. */
int nlasts; /* The number of LASTS. */ Idx nlasts; /* The number of LASTS. */
re_sub_match_last_t **lasts; re_sub_match_last_t **lasts;
} re_sub_match_top_t; } re_sub_match_top_t;
struct re_backref_cache_entry struct re_backref_cache_entry
{ {
int node; Idx node;
int str_idx; Idx str_idx;
int subexp_from; Idx subexp_from;
int subexp_to; Idx subexp_to;
char more; char more;
char unused; char unused;
unsigned short int eps_reachable_subexps_map; unsigned short int eps_reachable_subexps_map;
@ -557,18 +646,18 @@ typedef struct
/* EFLAGS of the argument of regexec. */ /* EFLAGS of the argument of regexec. */
int eflags; int eflags;
/* Where the matching ends. */ /* Where the matching ends. */
int match_last; Idx match_last;
int last_node; Idx last_node;
/* The state log used by the matcher. */ /* The state log used by the matcher. */
re_dfastate_t **state_log; re_dfastate_t **state_log;
int state_log_top; Idx state_log_top;
/* Back reference cache. */ /* Back reference cache. */
int nbkref_ents; Idx nbkref_ents;
int abkref_ents; Idx abkref_ents;
struct re_backref_cache_entry *bkref_ents; struct re_backref_cache_entry *bkref_ents;
int max_mb_elem_len; int max_mb_elem_len;
int nsub_tops; Idx nsub_tops;
int asub_tops; Idx asub_tops;
re_sub_match_top_t **sub_tops; re_sub_match_top_t **sub_tops;
} re_match_context_t; } re_match_context_t;
@ -576,23 +665,23 @@ typedef struct
{ {
re_dfastate_t **sifted_states; re_dfastate_t **sifted_states;
re_dfastate_t **limited_states; re_dfastate_t **limited_states;
int last_node; Idx last_node;
int last_str_idx; Idx last_str_idx;
re_node_set limits; re_node_set limits;
} re_sift_context_t; } re_sift_context_t;
struct re_fail_stack_ent_t struct re_fail_stack_ent_t
{ {
int idx; Idx idx;
int node; Idx node;
regmatch_t *regs; regmatch_t *regs;
re_node_set eps_via_nodes; re_node_set eps_via_nodes;
}; };
struct re_fail_stack_t struct re_fail_stack_t
{ {
int num; Idx num;
int alloc; Idx alloc;
struct re_fail_stack_ent_t *stack; struct re_fail_stack_ent_t *stack;
}; };
@ -601,8 +690,8 @@ struct re_dfa_t
re_token_t *nodes; re_token_t *nodes;
size_t nodes_alloc; size_t nodes_alloc;
size_t nodes_len; size_t nodes_len;
int *nexts; Idx *nexts;
int *org_indices; Idx *org_indices;
re_node_set *edests; re_node_set *edests;
re_node_set *eclosures; re_node_set *eclosures;
re_node_set *inveclosures; re_node_set *inveclosures;
@ -616,10 +705,10 @@ struct re_dfa_t
re_bitset_ptr_t sb_char; re_bitset_ptr_t sb_char;
int str_tree_storage_idx; int str_tree_storage_idx;
/* number of subexpressions `re_nsub' is in regex_t. */ /* number of subexpressions 're_nsub' is in regex_t. */
unsigned int state_hash_mask; re_hashval_t state_hash_mask;
int init_node; Idx init_node;
int nbackref; /* The number of backreference in this dfa. */ Idx nbackref; /* The number of backreference in this dfa. */
/* Bitmap expressing which backreference is used. */ /* Bitmap expressing which backreference is used. */
bitset_word_t used_bkref_map; bitset_word_t used_bkref_map;
@ -636,11 +725,11 @@ struct re_dfa_t
int mb_cur_max; int mb_cur_max;
bitset_t word_char; bitset_t word_char;
reg_syntax_t syntax; reg_syntax_t syntax;
int *subexp_map; Idx *subexp_map;
#ifdef DEBUG #ifdef DEBUG
char* re_str; char* re_str;
#endif #endif
__libc_lock_define (, lock) lock_define (lock)
}; };
#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
@ -671,16 +760,60 @@ typedef struct
} bracket_elem_t; } bracket_elem_t;
/* Inline functions for bitset operation. */ /* Functions for bitset_t operation. */
static void __attribute__ ((unused))
static inline void
bitset_set (bitset_t set, Idx i)
{
set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
}
static inline void
bitset_clear (bitset_t set, Idx i)
{
set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
}
static inline bool
bitset_contain (const bitset_t set, Idx i)
{
return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
}
static inline void
bitset_empty (bitset_t set)
{
memset (set, '\0', sizeof (bitset_t));
}
static inline void
bitset_set_all (bitset_t set)
{
memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
if (SBC_MAX % BITSET_WORD_BITS != 0)
set[BITSET_WORDS - 1] =
((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
}
static inline void
bitset_copy (bitset_t dest, const bitset_t src)
{
memcpy (dest, src, sizeof (bitset_t));
}
static inline void
bitset_not (bitset_t set) bitset_not (bitset_t set)
{ {
int bitset_i; int bitset_i;
for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
set[bitset_i] = ~set[bitset_i]; set[bitset_i] = ~set[bitset_i];
if (SBC_MAX % BITSET_WORD_BITS != 0)
set[BITSET_WORDS - 1] =
((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
& ~set[BITSET_WORDS - 1]);
} }
static void __attribute__ ((unused)) static inline void
bitset_merge (bitset_t dest, const bitset_t src) bitset_merge (bitset_t dest, const bitset_t src)
{ {
int bitset_i; int bitset_i;
@ -688,7 +821,7 @@ bitset_merge (bitset_t dest, const bitset_t src)
dest[bitset_i] |= src[bitset_i]; dest[bitset_i] |= src[bitset_i];
} }
static void __attribute__ ((unused)) static inline void
bitset_mask (bitset_t dest, const bitset_t src) bitset_mask (bitset_t dest, const bitset_t src)
{ {
int bitset_i; int bitset_i;
@ -697,10 +830,10 @@ bitset_mask (bitset_t dest, const bitset_t src)
} }
#ifdef RE_ENABLE_I18N #ifdef RE_ENABLE_I18N
/* Inline functions for re_string. */ /* Functions for re_string. */
static int static int
__attribute__ ((pure, unused)) __attribute__ ((pure, unused))
re_string_char_size_at (const re_string_t *pstr, int idx) re_string_char_size_at (const re_string_t *pstr, Idx idx)
{ {
int byte_idx; int byte_idx;
if (pstr->mb_cur_max == 1) if (pstr->mb_cur_max == 1)
@ -713,23 +846,22 @@ re_string_char_size_at (const re_string_t *pstr, int idx)
static wint_t static wint_t
__attribute__ ((pure, unused)) __attribute__ ((pure, unused))
re_string_wchar_at (const re_string_t *pstr, int idx) re_string_wchar_at (const re_string_t *pstr, Idx idx)
{ {
if (pstr->mb_cur_max == 1) if (pstr->mb_cur_max == 1)
return (wint_t) pstr->mbs[idx]; return (wint_t) pstr->mbs[idx];
return (wint_t) pstr->wcs[idx]; return (wint_t) pstr->wcs[idx];
} }
# if IS_IN (libc) # ifdef _LIBC
# ifdef _LIBC # include <locale/weight.h>
# include <locale/weight.h> # endif
# endif
static int static int
__attribute__ ((pure, unused)) __attribute__ ((pure, unused))
re_string_elem_size_at (const re_string_t *pstr, int idx) re_string_elem_size_at (const re_string_t *pstr, Idx idx)
{ {
# ifdef _LIBC # ifdef _LIBC
const unsigned char *p, *extra; const unsigned char *p, *extra;
const int32_t *table, *indirect; const int32_t *table, *indirect;
uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
@ -746,10 +878,34 @@ re_string_elem_size_at (const re_string_t *pstr, int idx)
return p - pstr->mbs - idx; return p - pstr->mbs - idx;
} }
else else
# endif /* _LIBC */ # endif /* _LIBC */
return 1; return 1;
} }
# endif
#endif /* RE_ENABLE_I18N */ #endif /* RE_ENABLE_I18N */
#ifndef __GNUC_PREREQ
# if defined __GNUC__ && defined __GNUC_MINOR__
# define __GNUC_PREREQ(maj, min) \
((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
# else
# define __GNUC_PREREQ(maj, min) 0
# endif
#endif
#if __GNUC_PREREQ (3,4)
# undef __attribute_warn_unused_result__
# define __attribute_warn_unused_result__ \
__attribute__ ((__warn_unused_result__))
#else
# define __attribute_warn_unused_result__ /* empty */
#endif
#ifndef FALLTHROUGH
# if __GNUC__ < 7
# define FALLTHROUGH ((void) 0)
# else
# define FALLTHROUGH __attribute__ ((__fallthrough__))
# endif
#endif
#endif /* _REGEX_INTERNAL_H */ #endif /* _REGEX_INTERNAL_H */

936
posix/regexec.c
View File

File diff suppressed because it is too large Load Diff