about summary refs log tree commit diff
diff options
context:
space:
mode:
authorUlrich Drepper <drepper@redhat.com>2002-02-26 19:06:03 +0000
committerUlrich Drepper <drepper@redhat.com>2002-02-26 19:06:03 +0000
commit3b0bdc723579a7c6df2cace0115a6ca0977d73f9 (patch)
tree8b6d7f9ab35be46faadc9e778abc1ce632fe98d0
parent73f1b06797637163b8529f4c7fa4b02b90c0154c (diff)
downloadglibc-3b0bdc723579a7c6df2cace0115a6ca0977d73f9.tar.gz
glibc-3b0bdc723579a7c6df2cace0115a6ca0977d73f9.tar.xz
glibc-3b0bdc723579a7c6df2cace0115a6ca0977d73f9.zip
Update.
	* posix/Makefile (distribute): Add regcomp.c, regexec.c,
	regex_internal.c, and regex_internal.h.
	(CFLAGS-regex.c): Replace -DMBS_SUPPORT with -DRE_ENABLE_I18N.
	* posix/regex.c: Complete rewrite.
	* posix/regexec.c: New file.
	* posix/regcomp.c: New file.
	* posix/regex_internal.c: New file.
	* posix/regex_internal.h: New file.
	* posix/regex.h (RE_ICASE): New macro.
	Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
-rw-r--r--ChangeLog11
-rw-r--r--NEWS5
-rw-r--r--posix/Makefile7
-rw-r--r--posix/regcomp.c3092
-rw-r--r--posix/regex.h12
-rw-r--r--posix/regex_internal.c1095
-rw-r--r--posix/regex_internal.h542
-rw-r--r--posix/regexec.c2076
8 files changed, 6832 insertions, 8 deletions
diff --git a/ChangeLog b/ChangeLog
index b132ab7926..41b32438c0 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,16 @@
 2002-02-26  Ulrich Drepper  <drepper@redhat.com>
 
+	* posix/Makefile (distribute): Add regcomp.c, regexec.c,
+	regex_internal.c, and regex_internal.h.
+	(CFLAGS-regex.c): Replace -DMBS_SUPPORT with -DRE_ENABLE_I18N.
+	* posix/regex.c: Complete rewrite.
+	* posix/regexec.c: New file.
+	* posix/regcomp.c: New file.
+	* posix/regex_internal.c: New file.
+	* posix/regex_internal.h: New file.
+	* posix/regex.h (RE_ICASE): New macro.
+	Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
+
 	* stdio-common/vfscanf.c (_IO_vfwscanf): Always use ungetc, never
 	ungetwc.  It's a macro.
 	* libio/tst-swscanf.c (do_test): Adjust for now fixed wscanf
diff --git a/NEWS b/NEWS
index 42ca9c7493..6bd0f4b563 100644
--- a/NEWS
+++ b/NEWS
@@ -1,4 +1,4 @@
-GNU C Library NEWS -- history of user-visible changes.  2002-1-28
+GNU C Library NEWS -- history of user-visible changes.  2002-2-26
 Copyright (C) 1992-2000, 2001, 2002 Free Software Foundation, Inc.
 See the end for copying conditions.
 
@@ -25,6 +25,9 @@ Version 2.3
 
 * The malloc functions were completely rewritten by Wolfram Gloger based
   on Doug Lea's malloc-2.7.0.c.
+
+* Isamu Hasegawa contributed a completely new and POSIX conforming
+  implementation of regex.
 
 Version 2.2.5
 
diff --git a/posix/Makefile b/posix/Makefile
index 785c227ebb..008154de9a 100644
--- a/posix/Makefile
+++ b/posix/Makefile
@@ -1,4 +1,4 @@
-# Copyright (C) 1991-1999, 2000, 2001 Free Software Foundation, Inc.
+# Copyright (C) 1991-1999, 2000, 2001, 2002 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
@@ -32,7 +32,8 @@ headers	:= sys/utsname.h sys/times.h sys/wait.h sys/types.h unistd.h	      \
 distribute := confstr.h TESTS TESTS2C.sed testcases.h \
 	      PTESTS PTESTS2C.sed ptestcases.h \
 	      globtest.c globtest.sh wordexp-tst.sh annexc.c fnmatch_loop.c   \
-	      spawn_int.h tst-getconf.sh
+	      spawn_int.h tst-getconf.sh regcomp.c regexec.c regex_internal.c \
+	      regex_internal.h
 
 routines :=								      \
 	uname								      \
@@ -109,7 +110,7 @@ $(objpfx)wordexp-tst.out: wordexp-tst.sh $(objpfx)wordexp-test
 endif
 endif
 
-CFLAGS-regex.c = -Wno-strict-prototypes -DMBS_SUPPORT
+CFLAGS-regex.c = -Wno-strict-prototypes -DRE_ENABLE_I18N
 CFLAGS-getaddrinfo.c = -DRESOLVER
 tstgetopt-ARGS = -a -b -cfoobar --required foobar --optional=bazbug \
 		--none random --col --color --colour
diff --git a/posix/regcomp.c b/posix/regcomp.c
new file mode 100644
index 0000000000..12da043062
--- /dev/null
+++ b/posix/regcomp.c
@@ -0,0 +1,3092 @@
+/* Extended regular expression matching and search library.
+   Copyright (C) 2002 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
+
+   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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <assert.h>
+#include <ctype.h>
+#include <limits.h>
+#include <locale.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <wchar.h>
+#include <wctype.h>
+
+#ifdef _LIBC
+# ifndef _RE_DEFINE_LOCALE_FUNCTIONS
+#  define _RE_DEFINE_LOCALE_FUNCTIONS 1
+#   include <locale/localeinfo.h>
+#   include <locale/elem-hash.h>
+#   include <locale/coll-lookup.h>
+# endif
+#endif
+
+/* This is for other GNU distributions with internationalized messages.  */
+#if HAVE_LIBINTL_H || defined _LIBC
+# include <libintl.h>
+# ifdef _LIBC
+#  undef gettext
+#  define gettext(msgid) __dcgettext ("libc", msgid, LC_MESSAGES)
+# endif
+#else
+# define gettext(msgid) (msgid)
+#endif
+
+#ifndef gettext_noop
+/* This define is so xgettext can find the internationalizable
+   strings.  */
+# define gettext_noop(String) String
+#endif
+
+#include "regex.h"
+#include "regex_internal.h"
+
+static reg_errcode_t re_compile_internal (regex_t *preg, const char * pattern,
+                                          int length, reg_syntax_t syntax);
+static void re_compile_fastmap_iter (regex_t *bufp,
+                                     const re_dfastate_t *init_state,
+                                     char *fastmap);
+static reg_errcode_t init_dfa (re_dfa_t *dfa, int pat_len);
+static void init_word_char (re_dfa_t *dfa);
+static void free_charset (re_charset_t *cset);
+static void free_workarea_compile (regex_t *preg);
+static reg_errcode_t create_initial_state (re_dfa_t *dfa);
+static reg_errcode_t analyze (re_dfa_t *dfa);
+static reg_errcode_t analyze_tree (re_dfa_t *dfa, bin_tree_t *node);
+static void calc_first (re_dfa_t *dfa, bin_tree_t *node);
+static void calc_next (re_dfa_t *dfa, bin_tree_t *node);
+static void calc_epsdest (re_dfa_t *dfa, bin_tree_t *node);
+static int duplicate_node (re_dfa_t *dfa, int org_idx,
+                           unsigned int constraint);
+static reg_errcode_t calc_eclosure (re_dfa_t *dfa);
+static re_node_set calc_eclosure_iter (re_dfa_t *dfa, int node, int root);
+static void calc_inveclosure (re_dfa_t *dfa);
+static int fetch_number (re_string_t *input, re_token_t *token,
+                         reg_syntax_t syntax);
+static re_token_t fetch_token (re_string_t *input, reg_syntax_t syntax);
+static int peek_token (re_token_t *token, re_string_t *input,
+                        reg_syntax_t syntax);
+static int peek_token_bracket (re_token_t *token, re_string_t *input,
+                               reg_syntax_t syntax);
+static bin_tree_t *parse (re_string_t *regexp, regex_t *preg,
+                          reg_syntax_t syntax, reg_errcode_t *err);
+static bin_tree_t *parse_reg_exp (re_string_t *regexp, regex_t *preg,
+                                  re_token_t *token, reg_syntax_t syntax,
+                                  int nest, reg_errcode_t *err);
+static bin_tree_t *parse_branch (re_string_t *regexp, regex_t *preg,
+                                 re_token_t *token, reg_syntax_t syntax,
+                                 int nest, reg_errcode_t *err);
+static bin_tree_t *parse_expression (re_string_t *regexp, regex_t *preg,
+                                     re_token_t *token, reg_syntax_t syntax,
+                                     int nest, reg_errcode_t *err);
+static bin_tree_t *parse_sub_exp (re_string_t *regexp, regex_t *preg,
+                                  re_token_t *token, reg_syntax_t syntax,
+                                  int nest, reg_errcode_t *err);
+static bin_tree_t *parse_dup_op (bin_tree_t *dup_elem, re_string_t *regexp,
+                                 re_dfa_t *dfa, re_token_t *token,
+                                 reg_syntax_t syntax, reg_errcode_t *err);
+static bin_tree_t *parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa,
+                                      re_token_t *token, reg_syntax_t syntax,
+                                      reg_errcode_t *err);
+static reg_errcode_t parse_bracket_element (bracket_elem_t *elem,
+                                            re_string_t *regexp,
+                                            re_token_t *token, int token_len,
+                                            re_dfa_t *dfa,
+                                            reg_syntax_t syntax);
+static reg_errcode_t parse_bracket_symbol (bracket_elem_t *elem,
+                                          re_string_t *regexp,
+                                          re_token_t *token);
+static reg_errcode_t build_equiv_class (re_charset_t *mbcset,
+                                        re_bitset_ptr_t sbcset,
+                                        int *equiv_class_alloc,
+                                        const unsigned char *name);
+static reg_errcode_t build_charclass (re_charset_t *mbcset,
+                                      re_bitset_ptr_t sbcset,
+                                      int *char_class_alloc,
+                                      const unsigned char *name);
+static bin_tree_t *build_word_op (re_dfa_t *dfa, int not, reg_errcode_t *err);
+static void free_bin_tree (bin_tree_t *tree);
+static bin_tree_t *create_tree (bin_tree_t *left, bin_tree_t *right,
+                                re_token_type_t type, int index);
+static bin_tree_t *duplicate_tree (const bin_tree_t *src, re_dfa_t *dfa);
+
+/* This table gives an error message for each of the error codes listed
+   in regex.h.  Obviously the order here has to be same as there.
+   POSIX doesn't require that we do anything for REG_NOERROR,
+   but why not be nice?  */
+
+const char re_error_msgid[] =
+  {
+#define REG_NOERROR_IDX	0
+    gettext_noop ("Success")	/* REG_NOERROR */
+    "\0"
+#define REG_NOMATCH_IDX (REG_NOERROR_IDX + sizeof "Success")
+    gettext_noop ("No match")	/* REG_NOMATCH */
+    "\0"
+#define REG_BADPAT_IDX	(REG_NOMATCH_IDX + sizeof "No match")
+    gettext_noop ("Invalid regular expression") /* REG_BADPAT */
+    "\0"
+#define REG_ECOLLATE_IDX (REG_BADPAT_IDX + sizeof "Invalid regular expression")
+    gettext_noop ("Invalid collation character") /* REG_ECOLLATE */
+    "\0"
+#define REG_ECTYPE_IDX	(REG_ECOLLATE_IDX + sizeof "Invalid collation character")
+    gettext_noop ("Invalid character class name") /* REG_ECTYPE */
+    "\0"
+#define REG_EESCAPE_IDX	(REG_ECTYPE_IDX + sizeof "Invalid character class name")
+    gettext_noop ("Trailing backslash") /* REG_EESCAPE */
+    "\0"
+#define REG_ESUBREG_IDX	(REG_EESCAPE_IDX + sizeof "Trailing backslash")
+    gettext_noop ("Invalid back reference") /* REG_ESUBREG */
+    "\0"
+#define REG_EBRACK_IDX	(REG_ESUBREG_IDX + sizeof "Invalid back reference")
+    gettext_noop ("Unmatched [ or [^")	/* REG_EBRACK */
+    "\0"
+#define REG_EPAREN_IDX	(REG_EBRACK_IDX + sizeof "Unmatched [ or [^")
+    gettext_noop ("Unmatched ( or \\(") /* REG_EPAREN */
+    "\0"
+#define REG_EBRACE_IDX	(REG_EPAREN_IDX + sizeof "Unmatched ( or \\(")
+    gettext_noop ("Unmatched \\{") /* REG_EBRACE */
+    "\0"
+#define REG_BADBR_IDX	(REG_EBRACE_IDX + sizeof "Unmatched \\{")
+    gettext_noop ("Invalid content of \\{\\}") /* REG_BADBR */
+    "\0"
+#define REG_ERANGE_IDX	(REG_BADBR_IDX + sizeof "Invalid content of \\{\\}")
+    gettext_noop ("Invalid range end")	/* REG_ERANGE */
+    "\0"
+#define REG_ESPACE_IDX	(REG_ERANGE_IDX + sizeof "Invalid range end")
+    gettext_noop ("Memory exhausted") /* REG_ESPACE */
+    "\0"
+#define REG_BADRPT_IDX	(REG_ESPACE_IDX + sizeof "Memory exhausted")
+    gettext_noop ("Invalid preceding regular expression") /* REG_BADRPT */
+    "\0"
+#define REG_EEND_IDX	(REG_BADRPT_IDX + sizeof "Invalid preceding regular expression")
+    gettext_noop ("Premature end of regular expression") /* REG_EEND */
+    "\0"
+#define REG_ESIZE_IDX	(REG_EEND_IDX + sizeof "Premature end of regular expression")
+    gettext_noop ("Regular expression too big") /* REG_ESIZE */
+    "\0"
+#define REG_ERPAREN_IDX	(REG_ESIZE_IDX + sizeof "Regular expression too big")
+    gettext_noop ("Unmatched ) or \\)") /* REG_ERPAREN */
+  };
+
+const size_t re_error_msgid_idx[] =
+  {
+    REG_NOERROR_IDX,
+    REG_NOMATCH_IDX,
+    REG_BADPAT_IDX,
+    REG_ECOLLATE_IDX,
+    REG_ECTYPE_IDX,
+    REG_EESCAPE_IDX,
+    REG_ESUBREG_IDX,
+    REG_EBRACK_IDX,
+    REG_EPAREN_IDX,
+    REG_EBRACE_IDX,
+    REG_BADBR_IDX,
+    REG_ERANGE_IDX,
+    REG_ESPACE_IDX,
+    REG_BADRPT_IDX,
+    REG_EEND_IDX,
+    REG_ESIZE_IDX,
+    REG_ERPAREN_IDX
+  };
+
+/* Entry points for GNU code.  */
+
+/* re_compile_pattern is the GNU regular expression compiler: it
+   compiles PATTERN (of length SIZE) and puts the result in BUFP.
+   Returns 0 if the pattern was valid, otherwise an error string.
+
+   Assumes the `allocated' (and perhaps `buffer') and `translate' fields
+   are set in BUFP on entry.  */
+
+const char *
+re_compile_pattern (pattern, length, bufp)
+    const char *pattern;
+    size_t length;
+    struct re_pattern_buffer *bufp;
+{
+  reg_errcode_t ret;
+
+  /* GNU code is written to assume at least RE_NREGS registers will be set
+     (and at least one extra will be -1).  */
+  bufp->regs_allocated = REGS_UNALLOCATED;
+
+  /* And GNU code determines whether or not to get register information
+     by passing null for the REGS argument to re_match, etc., not by
+     setting no_sub.  */
+  bufp->no_sub = 0;
+
+  /* Match anchors at newline.  */
+  bufp->newline_anchor = 1;
+
+  ret = re_compile_internal (bufp, (const unsigned char *) pattern, length,
+                             re_syntax_options);
+
+  if (!ret)
+    return NULL;
+  return gettext (re_error_msgid + re_error_msgid_idx[(int) ret]);
+}
+#ifdef _LIBC
+weak_alias (__re_compile_pattern, re_compile_pattern)
+#endif
+
+/* Set by `re_set_syntax' to the current regexp syntax to recognize.  Can
+   also be assigned to arbitrarily: each pattern buffer stores its own
+   syntax, so it can be changed between regex compilations.  */
+/* This has no initializer because initialized variables in Emacs
+   become read-only after dumping.  */
+reg_syntax_t re_syntax_options;
+
+
+/* Specify the precise syntax of regexps for compilation.  This provides
+   for compatibility for various utilities which historically have
+   different, incompatible syntaxes.
+
+   The argument SYNTAX is a bit mask comprised of the various bits
+   defined in regex.h.  We return the old syntax.  */
+
+reg_syntax_t
+re_set_syntax (syntax)
+    reg_syntax_t syntax;
+{
+  reg_syntax_t ret = re_syntax_options;
+
+  re_syntax_options = syntax;
+  return ret;
+}
+#ifdef _LIBC
+weak_alias (__re_set_syntax, re_set_syntax)
+#endif
+
+int
+re_compile_fastmap (bufp)
+    struct re_pattern_buffer *bufp;
+{
+  re_dfa_t *dfa = (re_dfa_t *) bufp->buffer;
+  char *fastmap = bufp->fastmap;
+
+  memset (fastmap, '\0', sizeof (char) * SBC_MAX);
+  re_compile_fastmap_iter (bufp, dfa->init_state, fastmap);
+  if (dfa->init_state != dfa->init_state_word)
+    re_compile_fastmap_iter (bufp, dfa->init_state_word, fastmap);
+  if (dfa->init_state != dfa->init_state_nl)
+    re_compile_fastmap_iter (bufp, dfa->init_state_nl, fastmap);
+  if (dfa->init_state != dfa->init_state_begbuf)
+    re_compile_fastmap_iter (bufp, dfa->init_state_begbuf, fastmap);
+  bufp->fastmap_accurate = 1;
+  return 0;
+}
+#ifdef _LIBC
+weak_alias (__re_compile_fastmap, re_compile_fastmap)
+#endif
+
+/* Helper function for re_compile_fastmap.
+   Compile fastmap for the initial_state INIT_STATE.  */
+
+static void
+re_compile_fastmap_iter (bufp, init_state, fastmap)
+     regex_t *bufp;
+     const re_dfastate_t *init_state;
+     char *fastmap;
+{
+  re_dfa_t *dfa = (re_dfa_t *) bufp->buffer;
+  int node_cnt;
+  for (node_cnt = 0; node_cnt < init_state->nodes.nelem; ++node_cnt)
+    {
+      int node = init_state->nodes.elems[node_cnt];
+      re_token_type_t type = dfa->nodes[node].type;
+      if (type == OP_CONTEXT_NODE)
+        {
+          node = dfa->nodes[node].opr.ctx_info->entity;
+          type = dfa->nodes[node].type;
+        }
+
+      if (type == CHARACTER)
+        fastmap[dfa->nodes[node].opr.c] = 1;
+      else if (type == SIMPLE_BRACKET)
+        {
+          int i, j, ch;
+          for (i = 0, ch = 0; i < BITSET_UINTS; ++i)
+            for (j = 0; j < UINT_BITS; ++j, ++ch)
+              if (dfa->nodes[node].opr.sbcset[i] & (1 << j))
+                fastmap[ch] = 1;
+        }
+      else if (type == COMPLEX_BRACKET)
+        {
+          int i;
+          re_charset_t *cset = dfa->nodes[node].opr.mbcset;
+          if (cset->non_match || cset->ncoll_syms || cset->nequiv_classes
+              || cset->nranges || cset->nchar_classes)
+            {
+              if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0)
+                {
+                  /* In this case we want to catch the bytes which are
+                     the first byte of any collation elements.
+                     e.g. In da_DK, we want to catch 'a' since "aa"
+                          is a valid collation element, and don't catch
+                          'b' since 'b' is the only collation element
+                          which starts from 'b'.  */
+                  int j, ch;
+                  const int32_t *table = (const int32_t *)
+                    _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+                  for (i = 0, ch = 0; i < BITSET_UINTS; ++i)
+                    for (j = 0; j < UINT_BITS; ++j, ++ch)
+                      if (table[ch] < 0)
+                        fastmap[ch] = 1;
+                }
+            }
+          for (i = 0; i < cset->nmbchars; ++i)
+            {
+              unsigned char buf[256];
+              wctomb (buf, cset->mbchars[i]);
+              fastmap[buf[0]] = 1;
+            }
+        }
+      else if (type == END_OF_RE || type == COMPLEX_BRACKET
+               || type == OP_PERIOD)
+        {
+          memset (fastmap, '\1', sizeof (char) * SBC_MAX);
+          if (type == END_OF_RE)
+            bufp->can_be_null = 1;
+          return;
+        }
+    }
+}
+
+/* Entry point for POSIX code.  */
+/* regcomp takes a regular expression as a string and compiles it.
+
+   PREG is a regex_t *.  We do not expect any fields to be initialized,
+   since POSIX says we shouldn't.  Thus, we set
+
+     `buffer' to the compiled pattern;
+     `used' to the length of the compiled pattern;
+     `syntax' to RE_SYNTAX_POSIX_EXTENDED if the
+       REG_EXTENDED bit in CFLAGS is set; otherwise, to
+       RE_SYNTAX_POSIX_BASIC;
+     `newline_anchor' to REG_NEWLINE being set in CFLAGS;
+     `fastmap' to an allocated space for the fastmap;
+     `fastmap_accurate' to zero;
+     `re_nsub' to the number of subexpressions in PATTERN.
+
+   PATTERN is the address of the pattern string.
+
+   CFLAGS is a series of bits which affect compilation.
+
+     If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we
+     use POSIX basic syntax.
+
+     If REG_NEWLINE is set, then . and [^...] don't match newline.
+     Also, regexec will try a match beginning after every newline.
+
+     If REG_ICASE is set, then we considers upper- and lowercase
+     versions of letters to be equivalent when matching.
+
+     If REG_NOSUB is set, then when PREG is passed to regexec, that
+     routine will report only success or failure, and nothing about the
+     registers.
+
+   It returns 0 if it succeeds, nonzero if it doesn't.  (See regex.h for
+   the return codes and their meanings.)  */
+
+int
+regcomp (preg, pattern, cflags)
+    regex_t *preg;
+    const char *pattern;
+    int cflags;
+{
+  reg_errcode_t ret;
+  reg_syntax_t syntax = ((cflags & REG_EXTENDED) ? RE_SYNTAX_POSIX_EXTENDED
+                         : RE_SYNTAX_POSIX_BASIC);
+
+  preg->buffer = NULL;
+  preg->allocated = 0;
+  preg->used = 0;
+
+  /* Try to allocate space for the fastmap.  */
+  preg->fastmap = re_malloc (char, SBC_MAX);
+  if (preg->fastmap == NULL)
+    return REG_ESPACE;
+
+  syntax |= (cflags & REG_ICASE) ? RE_ICASE : 0;
+
+  /* If REG_NEWLINE is set, newlines are treated differently.  */
+  if (cflags & REG_NEWLINE)
+    { /* REG_NEWLINE implies neither . nor [^...] match newline.  */
+      syntax &= ~RE_DOT_NEWLINE;
+      syntax |= RE_HAT_LISTS_NOT_NEWLINE;
+      /* It also changes the matching behavior.  */
+      preg->newline_anchor = 1;
+    }
+  else
+    preg->newline_anchor = 0;
+  preg->no_sub = !!(cflags & REG_NOSUB);
+  preg->translate = NULL;
+
+  ret = re_compile_internal (preg, pattern, strlen (pattern), syntax);
+
+  /* POSIX doesn't distinguish between an unmatched open-group and an
+     unmatched close-group: both are REG_EPAREN.  */
+  if (ret == REG_ERPAREN)
+    ret = REG_EPAREN;
+
+  /* XXX Why the test for preg->fastmap != NULL?  */
+  if (ret == REG_NOERROR && preg->fastmap != NULL)
+    {
+      /* Compute the fastmap now, since regexec cannot modify the pattern
+	 buffer.  */
+      if (re_compile_fastmap (preg) == -2)
+	{
+	  /* Some error occurred while computing the fastmap, just forget
+	     about it.  */
+	  re_free (preg->fastmap);
+	  preg->fastmap = NULL;
+	}
+    }
+
+  return (int) ret;
+}
+#ifdef _LIBC
+weak_alias (__regcomp, regcomp)
+#endif
+
+/* Returns a message corresponding to an error code, ERRCODE, returned
+   from either regcomp or regexec.   We don't use PREG here.  */
+
+size_t
+regerror (errcode, preg, errbuf, errbuf_size)
+    int errcode;
+    const regex_t *preg;
+    char *errbuf;
+    size_t errbuf_size;
+{
+  const char *msg;
+  size_t msg_size;
+
+  if (errcode < 0
+      || errcode >= (int) (sizeof (re_error_msgid_idx)
+			   / sizeof (re_error_msgid_idx[0])))
+    /* Only error codes returned by the rest of the code should be passed
+       to this routine.  If we are given anything else, or if other regex
+       code generates an invalid error code, then the program has a bug.
+       Dump core so we can fix it.  */
+    abort ();
+
+  msg = gettext (re_error_msgid + re_error_msgid_idx[errcode]);
+
+  msg_size = strlen (msg) + 1; /* Includes the null.  */
+
+  if (errbuf_size != 0)
+    {
+      if (msg_size > errbuf_size)
+        {
+#if defined HAVE_MEMPCPY || defined _LIBC
+	  *((char *) __mempcpy (errbuf, msg, errbuf_size - 1)) = '\0';
+#else
+          memcpy (errbuf, msg, errbuf_size - 1);
+          errbuf[errbuf_size - 1] = 0;
+#endif
+        }
+      else
+        memcpy (errbuf, msg, msg_size);
+    }
+
+  return msg_size;
+}
+#ifdef _LIBC
+weak_alias (__regerror, regerror)
+#endif
+
+/* Free dynamically allocated space used by PREG.  */
+
+void
+regfree (preg)
+    regex_t *preg;
+{
+  int i, j;
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  if (dfa != NULL)
+    {
+      re_free (dfa->subexps);
+
+      for (i = 0; i < dfa->nodes_len; ++i)
+        {
+          re_token_t *node = dfa->nodes + i;
+          if (node->type == COMPLEX_BRACKET && node->duplicated == 0)
+            free_charset (node->opr.mbcset);
+          else if (node->type == SIMPLE_BRACKET && node->duplicated == 0)
+            re_free (node->opr.sbcset);
+          else if (node->type == OP_CONTEXT_NODE)
+            {
+              if (dfa->nodes[node->opr.ctx_info->entity].type == OP_BACK_REF)
+                {
+                  if (node->opr.ctx_info->bkref_eclosure != NULL)
+                    re_node_set_free (node->opr.ctx_info->bkref_eclosure);
+                  re_free (node->opr.ctx_info->bkref_eclosure);
+                }
+              re_free (node->opr.ctx_info);
+            }
+        }
+      re_free (dfa->firsts);
+      re_free (dfa->nexts);
+      for (i = 0; i < dfa->nodes_len; ++i)
+        {
+          if (dfa->eclosures != NULL)
+            re_node_set_free (dfa->eclosures + i);
+          if (dfa->inveclosures != NULL)
+            re_node_set_free (dfa->inveclosures + i);
+          if (dfa->edests != NULL)
+            re_node_set_free (dfa->edests + i);
+        }
+      re_free (dfa->edests);
+      re_free (dfa->eclosures);
+      re_free (dfa->inveclosures);
+      re_free (dfa->nodes);
+
+      for (i = 0; i <= dfa->state_hash_mask; ++i)
+        {
+          struct re_state_table_entry *entry = dfa->state_table + i;
+          if (entry->alloc == 0)
+            re_free (entry->entry.state);
+          else
+            {
+              for (j = 0; j < entry->num; ++j)
+                {
+                  re_dfastate_t *state = entry->entry.array[j];
+                  if (state->entrance_nodes != &state->nodes)
+                    {
+                      re_node_set_free (state->entrance_nodes);
+                      re_free (state->entrance_nodes);
+                    }
+                  re_node_set_free (&state->nodes);
+                  re_free (state->trtable);
+                  re_free (state->trtable_search);
+                  re_free (state);
+                }
+              re_free (entry->entry.array);
+            }
+        }
+      re_free (dfa->state_table);
+
+      if (dfa->word_char != NULL)
+        re_free (dfa->word_char);
+      re_free (dfa);
+    }
+  re_free (preg->fastmap);
+}
+#ifdef _LIBC
+weak_alias (__regfree, regfree)
+#endif
+
+/* Entry points compatible with 4.2 BSD regex library.  We don't define
+   them unless specifically requested.  */
+
+#if defined _REGEX_RE_COMP || defined _LIBC
+
+/* BSD has one and only one pattern buffer.  */
+static struct re_pattern_buffer re_comp_buf;
+
+char *
+# ifdef _LIBC
+/* Make these definitions weak in libc, so POSIX programs can redefine
+   these names if they don't use our functions, and still use
+   regcomp/regexec above without link errors.  */
+weak_function
+# endif
+re_comp (s)
+     const char *s;
+{
+  reg_errcode_t ret;
+
+  if (!s)
+    {
+      if (!re_comp_buf.buffer)
+	return gettext ("No previous regular expression");
+      return 0;
+    }
+
+  if (!re_comp_buf.buffer)
+    {
+      re_comp_buf.fastmap = (char *) malloc (SBC_MAX);
+      if (re_comp_buf.fastmap == NULL)
+	return (char *) gettext (re_error_msgid
+				 + re_error_msgid_idx[(int) REG_ESPACE]);
+    }
+
+  /* Since `re_exec' always passes NULL for the `regs' argument, we
+     don't need to initialize the pattern buffer fields which affect it.  */
+
+  /* Match anchors at newlines.  */
+  re_comp_buf.newline_anchor = 1;
+
+  ret = re_compile_internal (&re_comp_buf, s, strlen (s), re_syntax_options);
+
+  if (!ret)
+    return NULL;
+
+  /* Yes, we're discarding `const' here if !HAVE_LIBINTL.  */
+  return (char *) gettext (re_error_msgid + re_error_msgid_idx[(int) ret]);
+}
+#endif /* _REGEX_RE_COMP */
+
+/* Internal entry point.
+   Compile the regular expression PATTERN, whose length is LENGTH.
+   SYNTAX indicate regular expression's syntax.  */
+
+static reg_errcode_t
+re_compile_internal (preg, pattern, length, syntax)
+     regex_t *preg;
+     const char * pattern;
+     int length;
+     reg_syntax_t syntax;
+{
+  reg_errcode_t err = REG_NOERROR;
+  re_dfa_t *dfa;
+  re_string_t regexp;
+
+  /* Initialize the pattern buffer.  */
+  preg->fastmap_accurate = 0;
+  preg->syntax = syntax;
+  preg->not_bol = preg->not_eol = 0;
+  preg->used = 0;
+  preg->re_nsub = 0;
+
+  /* Initialize the dfa.  */
+  dfa = (re_dfa_t *) preg->buffer;
+  if (preg->allocated < sizeof (re_dfa_t))
+    {
+      /* If zero allocated, but buffer is non-null, try to realloc
+	 enough space.  This loses if buffer's address is bogus, but
+	 that is the user's responsibility.  If ->buffer is NULL this
+	 is a simple allocation.  */
+      dfa = re_realloc (preg->buffer, re_dfa_t, 1);
+      if (dfa == NULL)
+	return REG_ESPACE;
+      memset (dfa, '\0', sizeof (re_dfa_t));
+      preg->allocated = sizeof (re_dfa_t);
+    }
+  preg->buffer = (unsigned char *) dfa;
+  preg->used = sizeof (re_dfa_t);
+
+  err = init_dfa (dfa, length);
+  if (err != REG_NOERROR)
+    {
+      re_free (dfa);
+      preg->buffer = NULL;
+      return err;
+    }
+
+  if (syntax & RE_ICASE)
+    err = re_string_construct_toupper (&regexp, pattern, length,
+                                       preg->translate);
+  else
+    err = re_string_construct (&regexp, pattern, length, preg->translate);
+
+  if (err != REG_NOERROR)
+    {
+      re_free (dfa);
+      preg->buffer = NULL;
+      return err;
+    }
+
+  /* Parse the regular expression, and build a structure tree.  */
+  preg->re_nsub = 0;
+  dfa->str_tree = parse (&regexp, preg, syntax, &err);
+  if (dfa->str_tree == NULL)
+    goto re_compile_internal_free_return;
+
+  /* Analyze the tree and collect information which is necessary to
+     create the dfa.  */
+  err = analyze (dfa);
+  if (err != REG_NOERROR)
+    goto re_compile_internal_free_return;
+
+  /* Then create the initial state of the dfa.  */
+  err = create_initial_state (dfa);
+  if (err != REG_NOERROR)
+    goto re_compile_internal_free_return;
+
+ re_compile_internal_free_return:
+  /* Release work areas.  */
+  free_workarea_compile (preg);
+  re_string_destruct (&regexp);
+
+  return err;
+}
+
+/* Initialize DFA.  We use the length of the regular expression PAT_LEN
+   as the initial length of some arrays.  */
+
+static reg_errcode_t
+init_dfa (dfa, pat_len)
+     re_dfa_t *dfa;
+     int pat_len;
+{
+  int table_size;
+  dfa->nodes_alloc = pat_len + 1;
+  dfa->nodes = re_malloc (re_token_t, dfa->nodes_alloc);
+
+  dfa->states_alloc = pat_len + 1;
+
+  /*  table_size = 2 ^ ceil(log pat_len) */
+  for (table_size = 1; table_size > 0; table_size <<= 1)
+    if (table_size > pat_len)
+      break;
+
+  dfa->state_table = calloc (sizeof (struct re_state_table_entry), table_size);
+  dfa->state_hash_mask = table_size - 1;
+
+  dfa->subexps_alloc = 1;
+  dfa->subexps = re_malloc (re_subexp_t, dfa->subexps_alloc);
+  dfa->word_char = NULL;
+
+  if (dfa->nodes == NULL || dfa->state_table == NULL || dfa->subexps == NULL)
+    {
+      /* We don't bother to free anything which was allocated.  Very
+	 soon the process will go down anyway.  */
+      dfa->subexps = NULL;
+      dfa->state_table = NULL;
+      dfa->nodes = NULL;
+      return REG_ESPACE;
+    }
+  return REG_NOERROR;
+}
+
+/* Initialize WORD_CHAR table, which indicate which character is
+   "word".  In this case "word" means that it is the word construction
+   character used by some operators like "\<", "\>", etc.  */
+
+static void
+init_word_char (dfa)
+     re_dfa_t *dfa;
+{
+  int i, j, ch;
+  dfa->word_char = (re_bitset_ptr_t) calloc (sizeof (bitset), 1);
+  for (i = 0, ch = 0; i < BITSET_UINTS; ++i)
+    for (j = 0; j < UINT_BITS; ++j, ++ch)
+      if (isalnum (ch) || ch == '_')
+        dfa->word_char[i] |= 1 << j;
+}
+
+/* Free the work area which are only used while compiling.  */
+
+static void
+free_workarea_compile (preg)
+     regex_t *preg;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  free_bin_tree (dfa->str_tree);
+  dfa->str_tree = NULL;
+}
+
+/* Create initial states for all contexts.  */
+
+static reg_errcode_t
+create_initial_state (dfa)
+     re_dfa_t *dfa;
+{
+  int first, i;
+  reg_errcode_t err;
+  re_node_set init_nodes;
+
+  /* Initial states have the epsilon closure of the node which is
+     the first node of the regular expression.  */
+  first = dfa->str_tree->first;
+  dfa->init_node = first;
+  err = re_node_set_init_copy (&init_nodes, dfa->eclosures + first);
+  if (err != REG_NOERROR)
+    return err;
+
+  /* The back-references which are in initial states can epsilon transit,
+     since in this case all of the subexpressions can be null.
+     Then we add epsilon closures of the nodes which are the next nodes of
+     the back-references.  */
+  if (dfa->nbackref > 0)
+    for (i = 0; i < init_nodes.nelem; ++i)
+      {
+        int node_idx = init_nodes.elems[i];
+        re_token_type_t type = dfa->nodes[node_idx].type;
+        if (type == OP_CONTEXT_NODE
+            && (dfa->nodes[dfa->nodes[node_idx].opr.ctx_info->entity].type
+                == OP_BACK_REF))
+          {
+            int prev_nelem = init_nodes.nelem;
+            re_node_set_merge (&init_nodes,
+                           dfa->nodes[node_idx].opr.ctx_info->bkref_eclosure);
+            if (prev_nelem < init_nodes.nelem)
+              i = 0;
+          }
+        else if (type == OP_BACK_REF)
+          {
+            int next_idx = dfa->nexts[node_idx];
+            if (!re_node_set_contains (&init_nodes, next_idx))
+              {
+                re_node_set_merge (&init_nodes, dfa->eclosures + next_idx);
+                i = 0;
+              }
+          }
+      }
+
+  /* It must be the first time to invoke acquire_state.  */
+  dfa->init_state = re_acquire_state_context (dfa, &init_nodes, 0);
+  if (dfa->init_state->has_constraint)
+    {
+      dfa->init_state_word = re_acquire_state_context (dfa, &init_nodes,
+                                                       CONTEXT_WORD);
+      dfa->init_state_nl = re_acquire_state_context (dfa, &init_nodes,
+                                                     CONTEXT_NEWLINE);
+      dfa->init_state_begbuf = re_acquire_state_context (dfa, &init_nodes,
+                                                         CONTEXT_NEWLINE
+                                                         | CONTEXT_BEGBUF);
+    }
+  else
+    dfa->init_state_word = dfa->init_state_nl
+      = dfa->init_state_begbuf = dfa->init_state;
+
+  if (dfa->init_state == NULL || dfa->init_state_word == NULL
+      || dfa->init_state_nl == NULL || dfa->init_state_begbuf == NULL )
+    return REG_ESPACE;
+  re_node_set_free (&init_nodes);
+  return REG_NOERROR;
+}
+
+/* Analyze the structure tree, and calculate "first", "next", "edest",
+   "eclosure", and "inveclosure".  */
+
+static reg_errcode_t
+analyze (dfa)
+     re_dfa_t *dfa;
+{
+  int i;
+  reg_errcode_t ret;
+
+  /* Allocate arrays.  */
+  dfa->firsts = re_malloc (int, dfa->nodes_alloc);
+  dfa->nexts = re_malloc (int, dfa->nodes_alloc);
+  dfa->edests = re_malloc (re_node_set, dfa->nodes_alloc);
+  dfa->eclosures = re_malloc (re_node_set, dfa->nodes_alloc);
+  dfa->inveclosures = re_malloc (re_node_set, dfa->nodes_alloc);
+  if (dfa->firsts == NULL || dfa->nexts == NULL || dfa->edests == NULL
+      || dfa->eclosures == NULL || dfa->inveclosures == NULL)
+    return REG_ESPACE;
+  /* Initialize them.  */
+  for (i = 0; i < dfa->nodes_len; ++i)
+    {
+      dfa->firsts[i] = -1;
+      dfa->nexts[i] = -1;
+      re_node_set_init_empty (dfa->edests + i);
+      re_node_set_init_empty (dfa->eclosures + i);
+      re_node_set_init_empty (dfa->inveclosures + i);
+    }
+
+  ret = analyze_tree (dfa, dfa->str_tree);
+  if (ret == REG_NOERROR)
+    {
+      ret = calc_eclosure (dfa);
+      if (ret == REG_NOERROR)
+	calc_inveclosure (dfa);
+    }
+  return ret;
+}
+
+/* Helper functions for analyze.
+   This function calculate "first", "next", and "edest" for the subtree
+   whose root is NODE.  */
+
+static reg_errcode_t
+analyze_tree (dfa, node)
+     re_dfa_t *dfa;
+     bin_tree_t *node;
+{
+  reg_errcode_t ret;
+  if (node->first == -1)
+    calc_first (dfa, node);
+  if (node->next == -1)
+    calc_next (dfa, node);
+  if (node->eclosure.nelem == 0)
+    calc_epsdest (dfa, node);
+  /* Calculate "first" etc. for the left child.  */
+  if (node->left != NULL)
+    {
+      ret = analyze_tree (dfa, node->left);
+      if (ret != REG_NOERROR)
+        return ret;
+    }
+  /* Calculate "first" etc. for the right child.  */
+  if (node->right != NULL)
+    {
+      ret = analyze_tree (dfa, node->right);
+      if (ret != REG_NOERROR)
+        return ret;
+    }
+  return REG_NOERROR;
+}
+
+/* Calculate "first" for the node NODE.  */
+static void
+calc_first (dfa, node)
+     re_dfa_t *dfa;
+     bin_tree_t *node;
+{
+  int idx, type;
+  idx = node->node_idx;
+  type = (node->type == 0) ? dfa->nodes[idx].type : node->type;
+
+  switch (type)
+    {
+#ifdef DEBUG
+    case OP_OPEN_SUBEXP:
+    case OP_CLOSE_SUBEXP:
+    case OP_OPEN_BRACKET:
+    case OP_CLOSE_BRACKET:
+    case OP_OPEN_DUP_NUM:
+    case OP_CLOSE_DUP_NUM:
+    case OP_NON_MATCH_LIST:
+    case OP_OPEN_COLL_ELEM:
+    case OP_CLOSE_COLL_ELEM:
+    case OP_OPEN_EQUIV_CLASS:
+    case OP_CLOSE_EQUIV_CLASS:
+    case OP_OPEN_CHAR_CLASS:
+    case OP_CLOSE_CHAR_CLASS:
+      /* These must not be appeared here.  */
+      assert (0);
+#endif
+    case END_OF_RE:
+    case CHARACTER:
+    case OP_PERIOD:
+    case OP_DUP_ASTERISK:
+    case OP_DUP_QUESTION:
+    case COMPLEX_BRACKET:
+    case SIMPLE_BRACKET:
+    case OP_BACK_REF:
+    case ANCHOR:
+      node->first = idx;
+      break;
+    case OP_DUP_PLUS:
+#ifdef DEBUG
+      assert (node->left != NULL);
+#endif
+      if (node->left->first == -1)
+        calc_first (dfa, node->left);
+      node->first = node->left->first;
+      break;
+    case OP_ALT:
+      node->first = idx;
+      break;
+    case SUBEXP:
+      if (node->left == NULL)
+        {
+          if (node->next == -1)
+            calc_next (dfa, node);
+          node->first = node->next;
+          break;
+        }
+      /* else fall through */
+    default:
+#ifdef DEBUG
+      assert (node->left != NULL);
+#endif
+      if (node->left->first == -1)
+        calc_first (dfa, node->left);
+      node->first = node->left->first;
+      break;
+    }
+  if (node->type == 0)
+    dfa->firsts[idx] = node->first;
+}
+
+/* Calculate "next" for the node NODE.  */
+
+static void
+calc_next (dfa, node)
+     re_dfa_t *dfa;
+     bin_tree_t *node;
+{
+  int idx, type;
+  bin_tree_t *parent = node->parent;
+  if (parent == NULL)
+    {
+      node->next = -1;
+      idx = node->node_idx;
+      if (node->type == 0)
+        dfa->nexts[idx] = node->next;
+      return;
+    }
+
+  idx = parent->node_idx;
+  type = (parent->type == 0) ? dfa->nodes[idx].type : parent->type;
+
+  switch (type)
+    {
+    case OP_DUP_ASTERISK:
+    case OP_DUP_PLUS:
+      node->next = idx;
+      break;
+    case CONCAT:
+      if (parent->left == node)
+        {
+          if (parent->right->first == -1)
+            calc_first (dfa, parent->right);
+          node->next = parent->right->first;
+          break;
+        }
+      /* else fall through */
+    default:
+      if (parent->next == -1)
+        calc_next (dfa, parent);
+      node->next = parent->next;
+      break;
+    }
+  idx = node->node_idx;
+  if (node->type == 0)
+    dfa->nexts[idx] = node->next;
+}
+
+/* Calculate "edest" for the node NODE.  */
+
+static void
+calc_epsdest (dfa, node)
+     re_dfa_t *dfa;
+     bin_tree_t *node;
+{
+  int idx;
+  idx = node->node_idx;
+  if (node->type == 0)
+    {
+      if (dfa->nodes[idx].type == OP_DUP_ASTERISK
+          || dfa->nodes[idx].type == OP_DUP_PLUS
+          || dfa->nodes[idx].type == OP_DUP_QUESTION)
+        {
+          if (node->left->first == -1)
+            calc_first (dfa, node->left);
+          if (node->next == -1)
+            calc_next (dfa, node);
+          re_node_set_init_2 (dfa->edests + idx, node->left->first,
+                              node->next);
+        }
+      else if (dfa->nodes[idx].type == OP_ALT)
+        {
+          int left, right;
+          if (node->left != NULL)
+            {
+              if (node->left->first == -1)
+                calc_first (dfa, node->left);
+              left = node->left->first;
+            }
+          else
+            {
+              if (node->next == -1)
+                calc_next (dfa, node);
+              left = node->next;
+            }
+          if (node->right != NULL)
+            {
+              if (node->right->first == -1)
+                calc_first (dfa, node->right);
+              right = node->right->first;
+            }
+          else
+            {
+              if (node->next == -1)
+                calc_next (dfa, node);
+              right = node->next;
+            }
+          re_node_set_init_2 (dfa->edests + idx, left, right);
+        }
+      else if (dfa->nodes[idx].type == ANCHOR)
+        re_node_set_init_1 (dfa->edests + idx, node->next);
+    }
+}
+
+static int
+duplicate_node (dfa, org_idx, constraint)
+     re_dfa_t *dfa;
+     int org_idx;
+     unsigned int constraint;
+{
+  re_token_t dup;
+  int dup_idx;
+
+  dup.type = OP_CONTEXT_NODE;
+  if (dfa->nodes[org_idx].type == OP_CONTEXT_NODE)
+    {
+      if (dfa->nodes[org_idx].constraint == constraint)
+        return org_idx;
+      dup.constraint = constraint |
+        dfa->nodes[org_idx].constraint;
+    }
+  else
+    dup.constraint = constraint;
+
+  /* In case that `entity' points OP_CONTEXT_NODE,
+     we correct `entity' to real entity in calc_inveclosures(). */
+  dup.opr.ctx_info = malloc (sizeof (*dup.opr.ctx_info));
+  dup.opr.ctx_info->entity = org_idx;
+  dup.opr.ctx_info->bkref_eclosure = NULL;
+  dup_idx = re_dfa_add_node (dfa, dup, 1);
+  dfa->nodes[dup_idx].duplicated = 1;
+
+  dfa->firsts[dup_idx] = dfa->firsts[org_idx];
+  dfa->nexts[dup_idx] = dfa->nexts[org_idx];
+  re_node_set_init_copy (dfa->edests + dup_idx, dfa->edests + org_idx);
+  /* Since we don't duplicate epsilon nodes, epsilon closure have
+     only itself.  */
+  re_node_set_init_1 (dfa->eclosures + dup_idx, dup_idx);
+  re_node_set_init_1 (dfa->inveclosures + dup_idx, dup_idx);
+  /* Then we must update inveclosure for this node.
+     We process them at last part of calc_eclosure(),
+     since we don't complete to calculate them here.  */
+
+  return dup_idx;
+}
+
+static void
+calc_inveclosure (dfa)
+     re_dfa_t *dfa;
+{
+  int src, idx, dest, entity;
+  for (src = 0; src < dfa->nodes_len; ++src)
+    {
+      for (idx = 0; idx < dfa->eclosures[src].nelem; ++idx)
+        {
+          dest = dfa->eclosures[src].elems[idx];
+          re_node_set_insert (dfa->inveclosures + dest, src);
+        }
+
+      entity = src;
+      while (dfa->nodes[entity].type == OP_CONTEXT_NODE)
+        {
+          entity = dfa->nodes[entity].opr.ctx_info->entity;
+          re_node_set_merge (dfa->inveclosures + src,
+                             dfa->inveclosures + entity);
+          dfa->nodes[src].opr.ctx_info->entity = entity;
+        }
+    }
+}
+
+/* Calculate "eclosure" for all the node in DFA.  */
+
+static reg_errcode_t
+calc_eclosure (dfa)
+     re_dfa_t *dfa;
+{
+  int idx, node_idx, max, incomplete = 0;
+#ifdef DEBUG
+  assert (dfa->nodes_len > 0);
+#endif
+  /* For each nodes, calculate epsilon closure.  */
+  for (node_idx = 0, max = dfa->nodes_len; ; ++node_idx)
+    {
+      re_node_set eclosure_elem;
+      if (node_idx == max)
+        {
+          if (!incomplete)
+            break;
+          incomplete = 0;
+          node_idx = 0;
+        }
+
+#ifdef DEBUG
+      assert (dfa->nodes[node_idx].type != OP_CONTEXT_NODE);
+      assert (dfa->eclosures[node_idx].nelem != -1);
+#endif
+      /* If we have already calculated, skip it.  */
+      if (dfa->eclosures[node_idx].nelem != 0)
+        continue;
+      /* Calculate epsilon closure of `node_idx'.  */
+      eclosure_elem = calc_eclosure_iter (dfa, node_idx, 1);
+
+      if (dfa->eclosures[node_idx].nelem == 0)
+        {
+          incomplete = 1;
+          re_node_set_free (&eclosure_elem);
+        }
+    }
+
+  /* for duplicated nodes.  */
+  for (idx = max; idx < dfa->nodes_len; ++idx)
+    {
+      int entity, i, constraint;
+      re_node_set *bkref_eclosure;
+      entity = dfa->nodes[idx].opr.ctx_info->entity;
+      re_node_set_merge (dfa->inveclosures + idx, dfa->inveclosures + entity);
+      if (dfa->nodes[entity].type != OP_BACK_REF)
+        continue;
+
+      /* If the node is backreference, duplicate the epsilon closure of
+         the next node. Since it may epsilon transit.  */
+      /* Note: duplicate_node() may realloc dfa->eclosures, etc.  */
+      bkref_eclosure = re_malloc (re_node_set, 1);
+      if (bkref_eclosure == NULL)
+	return REG_ESPACE;
+      re_node_set_init_empty (bkref_eclosure);
+      constraint = dfa->nodes[idx].constraint;
+      for (i = 0; i < dfa->eclosures[dfa->nexts[idx]].nelem; ++i)
+        {
+          int dest_node_idx = dfa->eclosures[dfa->nexts[idx]].elems[i];
+          if (!IS_EPSILON_NODE (dfa->nodes[dest_node_idx].type))
+            dest_node_idx = duplicate_node (dfa, dest_node_idx, constraint);
+          re_node_set_insert (bkref_eclosure, dest_node_idx);
+        }
+      dfa->nodes[idx].opr.ctx_info->bkref_eclosure = bkref_eclosure;
+    }
+
+  return REG_NOERROR;
+}
+
+/* Calculate epsilon closure of NODE.  */
+
+static re_node_set
+calc_eclosure_iter (dfa, node, root)
+     re_dfa_t *dfa;
+     int node, root;
+{
+  unsigned int constraint;
+  int i, max, incomplete = 0;
+  re_node_set eclosure;
+  re_node_set_alloc (&eclosure, 1);
+
+  /* This indicates that we are calculating this node now.
+     We reference this value to avoid infinite loop.  */
+  dfa->eclosures[node].nelem = -1;
+
+  constraint = ((dfa->nodes[node].type == ANCHOR)
+                ? dfa->nodes[node].opr.ctx_type : 0);
+
+  /* Expand each epsilon destination nodes.  */
+  if (dfa->edests[node].nelem != 0)
+    for (i = 0; i < dfa->edests[node].nelem; ++i)
+      {
+        re_node_set eclosure_elem;
+        int edest = dfa->edests[node].elems[i];
+        /* If calculating the epsilon closure of `edest' is in progress,
+           return intermediate result.  */
+        if (dfa->eclosures[edest].nelem == -1)
+          {
+            incomplete = 1;
+            continue;
+          }
+        /* If we haven't calculated the epsilon closure of `edest' yet,
+           calculate now. Otherwise use calculated epsilon closure.  */
+        if (dfa->eclosures[edest].nelem == 0)
+          eclosure_elem = calc_eclosure_iter (dfa, edest, 0);
+        else
+          eclosure_elem = dfa->eclosures[edest];
+        /* Merge the epsilon closure of `edest'.  */
+        re_node_set_merge (&eclosure, &eclosure_elem);
+        /* If the epsilon closure of `edest' is incomplete,
+           the epsilon closure of this node is also incomplete.  */
+        if (dfa->eclosures[edest].nelem == 0)
+          {
+            incomplete = 1;
+            re_node_set_free (&eclosure_elem);
+          }
+      }
+
+  /* If the current node has constraints, duplicate all non-epsilon nodes.
+     Since they must inherit the constraints.  */
+  if (constraint)
+    for (i = 0, max = eclosure.nelem; i < max; ++i)
+      {
+        int dest = eclosure.elems[i];
+        if (!IS_EPSILON_NODE (dfa->nodes[dest].type))
+          {
+            int dup_dest = duplicate_node (dfa, dest, constraint);
+            if (dest != dup_dest)
+              {
+                re_node_set_remove_at (&eclosure, i--);
+                re_node_set_insert (&eclosure, dup_dest);
+                --max;
+              }
+          }
+      }
+
+  /* Epsilon closures include itself.  */
+  re_node_set_insert (&eclosure, node);
+  if (incomplete && !root)
+    dfa->eclosures[node].nelem = 0;
+  else
+    dfa->eclosures[node] = eclosure;
+  return eclosure;
+}
+
+/* Functions for token which are used in the parser.  */
+
+/* Fetch a token from INPUT.
+   We must not use this function inside bracket expressions.  */
+
+static re_token_t
+fetch_token (input, syntax)
+     re_string_t *input;
+     reg_syntax_t syntax;
+{
+  re_token_t token;
+  int consumed_byte;
+  consumed_byte = peek_token (&token, input, syntax);
+  re_string_skip_bytes (input, consumed_byte);
+  return token;
+}
+
+/* Peek a token from INPUT, and return the length of the token.
+   We must not use this function inside bracket expressions.  */
+
+static int
+peek_token (token, input, syntax)
+     re_token_t *token;
+     re_string_t *input;
+     reg_syntax_t syntax;
+{
+  unsigned char c;
+
+  if (re_string_eoi (input))
+    {
+      token->type = END_OF_RE;
+      return 0;
+    }
+
+  c = re_string_peek_byte (input, 0);
+  token->opr.c = c;
+
+#ifdef RE_ENABLE_I18N
+  token->mb_partial = 0;
+  if (MB_CUR_MAX > 1 &&
+      !re_string_first_byte (input, re_string_cur_idx (input)))
+    {
+      token->type = CHARACTER;
+      token->mb_partial = 1;
+      return 1;
+    }
+#endif
+  if (c == '\\')
+    {
+      unsigned char c2;
+      if (re_string_cur_idx (input) + 1 >= re_string_length (input))
+        {
+          token->type = BACK_SLASH;
+          return 1;
+        }
+
+      c2 = re_string_peek_byte_case (input, 1);
+      token->opr.c = c2;
+      token->type = CHARACTER;
+      switch (c2)
+        {
+        case '|':
+          if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_NO_BK_VBAR))
+            token->type = OP_ALT;
+          break;
+        case '1': case '2': case '3': case '4': case '5':
+        case '6': case '7': case '8': case '9':
+          if (!(syntax & RE_NO_BK_REFS))
+            {
+              token->type = OP_BACK_REF;
+              token->opr.idx = c2 - '0';
+            }
+          break;
+        case '<':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = WORD_FIRST;
+            }
+          break;
+        case '>':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = WORD_LAST;
+            }
+          break;
+        case 'b':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = WORD_DELIM;
+            }
+          break;
+        case 'B':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = INSIDE_WORD;
+            }
+          break;
+        case 'w':
+          if (!(syntax & RE_NO_GNU_OPS))
+            token->type = OP_WORD;
+          break;
+        case 'W':
+          if (!(syntax & RE_NO_GNU_OPS))
+            token->type = OP_NOTWORD;
+          break;
+        case '`':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = BUF_FIRST;
+            }
+          break;
+        case '\'':
+          if (!(syntax & RE_NO_GNU_OPS))
+            {
+              token->type = ANCHOR;
+              token->opr.idx = BUF_LAST;
+            }
+          break;
+        case '(':
+          if (!(syntax & RE_NO_BK_PARENS))
+            token->type = OP_OPEN_SUBEXP;
+          break;
+        case ')':
+          if (!(syntax & RE_NO_BK_PARENS))
+            token->type = OP_CLOSE_SUBEXP;
+          break;
+        case '+':
+          if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM))
+            token->type = OP_DUP_PLUS;
+          break;
+        case '?':
+          if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM))
+            token->type = OP_DUP_QUESTION;
+          break;
+        case '{':
+          if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES)))
+            token->type = OP_OPEN_DUP_NUM;
+          break;
+        case '}':
+          if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES)))
+            token->type = OP_CLOSE_DUP_NUM;
+          break;
+        default:
+          break;
+        }
+      return 2;
+    }
+
+  token->type = CHARACTER;
+  switch (c)
+    {
+    case '\n':
+      if (syntax & RE_NEWLINE_ALT)
+        token->type = OP_ALT;
+      break;
+    case '|':
+      if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_NO_BK_VBAR))
+        token->type = OP_ALT;
+      break;
+    case '*':
+      token->type = OP_DUP_ASTERISK;
+      break;
+    case '+':
+      if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM))
+        token->type = OP_DUP_PLUS;
+      break;
+    case '?':
+      if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM))
+        token->type = OP_DUP_QUESTION;
+      break;
+    case '{':
+      if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES))
+        token->type = OP_OPEN_DUP_NUM;
+      break;
+    case '}':
+      if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES))
+        token->type = OP_CLOSE_DUP_NUM;
+      break;
+    case '(':
+      if (syntax & RE_NO_BK_PARENS)
+        token->type = OP_OPEN_SUBEXP;
+      break;
+    case ')':
+      if (syntax & RE_NO_BK_PARENS)
+        token->type = OP_CLOSE_SUBEXP;
+      break;
+    case '[':
+      token->type = OP_OPEN_BRACKET;
+      break;
+    case '.':
+      token->type = OP_PERIOD;
+      break;
+    case '^':
+      if (!(syntax & RE_CONTEXT_INDEP_ANCHORS) &&
+          re_string_cur_idx (input) != 0)
+        {
+          char prev = re_string_peek_byte (input, -1);
+          if (prev != '|' && prev != '(' &&
+              (!(syntax & RE_NEWLINE_ALT) || prev != '\n'))
+            break;
+        }
+      token->type = ANCHOR;
+      token->opr.idx = LINE_FIRST;
+      break;
+    case '$':
+      if (!(syntax & RE_CONTEXT_INDEP_ANCHORS) &&
+          re_string_cur_idx (input) + 1 != re_string_length (input))
+        {
+          re_token_t next;
+          re_string_skip_bytes (input, 1);
+          peek_token (&next, input, syntax);
+          re_string_skip_bytes (input, -1);
+          if (next.type != OP_ALT && next.type != OP_CLOSE_SUBEXP)
+            break;
+        }
+      token->type = ANCHOR;
+      token->opr.idx = LINE_LAST;
+      break;
+    default:
+      break;
+    }
+  return 1;
+}
+
+/* Peek a token from INPUT, and return the length of the token.
+   We must not use this function out of bracket expressions.  */
+
+static int
+peek_token_bracket (token, input, syntax)
+     re_token_t *token;
+     re_string_t *input;
+     reg_syntax_t syntax;
+{
+  unsigned char c;
+  if (re_string_eoi (input))
+    {
+      token->type = END_OF_RE;
+      return 0;
+    }
+  c = re_string_peek_byte (input, 0);
+  token->opr.c = c;
+
+#ifdef RE_ENABLE_I18N
+  if (MB_CUR_MAX > 1 &&
+      !re_string_first_byte (input, re_string_cur_idx (input)))
+    {
+      token->type = CHARACTER;
+      return 1;
+    }
+#endif /* RE_ENABLE_I18N */
+
+  if (c == '\\' && (syntax & RE_BACKSLASH_ESCAPE_IN_LISTS))
+    {
+      /* In this case, '\' escape a character.  */
+      unsigned char c2;
+      c2 = re_string_peek_byte (input, 1);
+      token->opr.c = c2;
+      token->type = CHARACTER;
+      return 1;
+    }
+  if (c == '[') /* '[' is a special char in a bracket exps.  */
+    {
+      unsigned char c2;
+      int token_len;
+      c2 = re_string_peek_byte (input, 1);
+      token->opr.c = c2;
+      token_len = 2;
+      switch (c2)
+        {
+        case '.':
+          token->type = OP_OPEN_COLL_ELEM;
+          break;
+        case '=':
+          token->type = OP_OPEN_EQUIV_CLASS;
+          break;
+        case ':':
+          if (syntax & RE_CHAR_CLASSES)
+            {
+              token->type = OP_OPEN_CHAR_CLASS;
+              break;
+            }
+          /* else fall through.  */
+        default:
+          token->type = CHARACTER;
+          token->opr.c = c;
+          token_len = 1;
+          break;
+        }
+      return token_len;
+    }
+  switch (c)
+    {
+    case '-':
+      token->type = OP_CHARSET_RANGE;
+      break;
+    case ']':
+      token->type = OP_CLOSE_BRACKET;
+      break;
+    case '^':
+      token->type = OP_NON_MATCH_LIST;
+      break;
+    default:
+      token->type = CHARACTER;
+    }
+  return 1;
+}
+
+/* Functions for parser.  */
+
+/* Entry point of the parser.
+   Parse the regular expression REGEXP and return the structure tree.
+   If an error is occured, ERR is set by error code, and return NULL.
+   This function build the following tree, from regular expression <reg_exp>:
+           CAT
+           / \
+          /   \
+   <reg_exp>  EOR
+
+   CAT means concatenation.
+   EOR means end of regular expression.  */
+
+static bin_tree_t *
+parse (regexp, preg, syntax, err)
+     re_string_t *regexp;
+     regex_t *preg;
+     reg_syntax_t syntax;
+     reg_errcode_t *err;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  bin_tree_t *tree, *eor, *root;
+  re_token_t current_token;
+  int new_idx;
+  current_token = fetch_token (regexp, syntax);
+  tree = parse_reg_exp (regexp, preg, &current_token, syntax, 0, err);
+  if (*err != REG_NOERROR && tree == NULL)
+    return NULL;
+  new_idx = re_dfa_add_node (dfa, current_token, 0);
+  eor = create_tree (NULL, NULL, 0, new_idx);
+  if (tree != NULL)
+    root = create_tree (tree, eor, CONCAT, 0);
+  else
+    root = eor;
+  if (new_idx == -1 || eor == NULL || root == NULL)
+    return *err = REG_ESPACE, NULL;
+  return root;
+}
+
+/* This function build the following tree, from regular expression
+   <branch1>|<branch2>:
+           ALT
+           / \
+          /   \
+   <branch1> <branch2>
+
+   ALT means alternative, which represents the operator `|'.  */
+
+static bin_tree_t *
+parse_reg_exp (regexp, preg, token, syntax, nest, err)
+     re_string_t *regexp;
+     regex_t *preg;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     int nest;
+     reg_errcode_t *err;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  bin_tree_t *tree, *branch = NULL;
+  int new_idx;
+  tree = parse_branch (regexp, preg, token, syntax, nest, err);
+  if (*err != REG_NOERROR && tree == NULL)
+    return NULL;
+
+  while (token->type == OP_ALT)
+    {
+      re_token_t alt_token = *token;
+      new_idx = re_dfa_add_node (dfa, alt_token, 0);
+      *token = fetch_token (regexp, syntax);
+      if (token->type != OP_ALT && token->type != END_OF_RE
+          && (nest == 0 || token->type != OP_CLOSE_SUBEXP))
+        {
+          branch = parse_branch (regexp, preg, token, syntax, nest, err);
+          if (*err != REG_NOERROR && branch == NULL)
+            {
+              free_bin_tree (tree);
+              return NULL;
+            }
+        }
+      tree = create_tree (tree, branch, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+    }
+  return tree;
+}
+
+/* This function build the following tree, from regular expression
+   <exp1><exp2>:
+        CAT
+        / \
+       /   \
+   <exp1> <exp2>
+
+   CAT means concatenation.  */
+
+static bin_tree_t *
+parse_branch (regexp, preg, token, syntax, nest, err)
+     re_string_t *regexp;
+     regex_t *preg;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     int nest;
+     reg_errcode_t *err;
+{
+  bin_tree_t *tree, *exp;
+  tree = parse_expression (regexp, preg, token, syntax, nest, err);
+  if (*err != REG_NOERROR && tree == NULL)
+    return NULL;
+
+  while (token->type != OP_ALT && token->type != END_OF_RE
+         && (nest == 0 || token->type != OP_CLOSE_SUBEXP))
+    {
+      exp = parse_expression (regexp, preg, token, syntax, nest, err);
+      if (*err != REG_NOERROR && exp == NULL)
+        {
+          free_bin_tree (tree);
+          return NULL;
+        }
+      if (tree != NULL && exp != NULL)
+        {
+          tree = create_tree (tree, exp, CONCAT, 0);
+          if (tree == NULL)
+            return *err = REG_ESPACE, NULL;
+        }
+      else if (tree == NULL)
+        tree = exp;
+      /* Otherwise exp == NULL, we don't need to create new tree.  */
+    }
+  return tree;
+}
+
+/* This function build the following tree, from regular expression a*:
+         *
+         |
+         a
+*/
+
+static bin_tree_t *
+parse_expression (regexp, preg, token, syntax, nest, err)
+     re_string_t *regexp;
+     regex_t *preg;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     int nest;
+     reg_errcode_t *err;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  bin_tree_t *tree;
+  int new_idx;
+  switch (token->type)
+    {
+    case CHARACTER:
+      new_idx = re_dfa_add_node (dfa, *token, 0);
+      tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+#ifdef RE_ENABLE_I18N
+      if (MB_CUR_MAX > 1)
+	{
+	  while (!re_string_eoi (regexp)
+		 && !re_string_first_byte (regexp, re_string_cur_idx (regexp)))
+	    {
+              bin_tree_t *mbc_remain;
+              *token = fetch_token (regexp, syntax);
+              new_idx = re_dfa_add_node (dfa, *token, 0);
+              mbc_remain = create_tree (NULL, NULL, 0, new_idx);
+              tree = create_tree (tree, mbc_remain, CONCAT, 0);
+              if (new_idx == -1 || mbc_remain == NULL || tree == NULL)
+                return *err = REG_ESPACE, NULL;
+            }
+	}
+#endif
+      break;
+    case OP_OPEN_SUBEXP:
+      tree = parse_sub_exp (regexp, preg, token, syntax, nest + 1, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return NULL;
+      break;
+    case OP_OPEN_BRACKET:
+      tree = parse_bracket_exp (regexp, dfa, token, syntax, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return NULL;
+      break;
+    case OP_BACK_REF:
+      if (preg->re_nsub < token->opr.idx
+          || dfa->subexps[token->opr.idx - 1].end == -1)
+        {
+          *err = REG_ESUBREG;
+          return NULL;
+        }
+      new_idx = re_dfa_add_node (dfa, *token, 0);
+      tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+      ++dfa->nbackref;
+      dfa->has_mb_node = 1;
+      break;
+    case OP_DUP_ASTERISK:
+    case OP_DUP_PLUS:
+    case OP_DUP_QUESTION:
+    case OP_OPEN_DUP_NUM:
+      if (syntax & RE_CONTEXT_INVALID_OPS)
+        return *err = REG_BADRPT, NULL;
+      else if (syntax & RE_CONTEXT_INDEP_OPS)
+        {
+          *token = fetch_token (regexp, syntax);
+          return parse_expression (regexp, preg, token, syntax, nest, err);
+        }
+      /* else fall through  */
+    case OP_CLOSE_SUBEXP:
+      if ((token->type == OP_CLOSE_SUBEXP) &&
+          !(syntax & RE_UNMATCHED_RIGHT_PAREN_ORD))
+        return *err = REG_ERPAREN, NULL;
+      /* else fall through  */
+    case OP_CLOSE_DUP_NUM:
+      /* We treat it as a normal character.  */
+
+      /* Then we can these characters as normal characters.  */
+      token->type = CHARACTER;
+      new_idx = re_dfa_add_node (dfa, *token, 0);
+      tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+      break;
+    case ANCHOR:
+      if (dfa->word_char == NULL)
+        init_word_char (dfa);
+      if (token->opr.ctx_type == WORD_DELIM)
+        {
+          bin_tree_t *tree_first, *tree_last;
+          int idx_first, idx_last;
+          token->opr.ctx_type = WORD_FIRST;
+          idx_first = re_dfa_add_node (dfa, *token, 0);
+          tree_first = create_tree (NULL, NULL, 0, idx_first);
+          token->opr.ctx_type = WORD_LAST;
+          idx_last = re_dfa_add_node (dfa, *token, 0);
+          tree_last = create_tree (NULL, NULL, 0, idx_last);
+          token->type = OP_ALT;
+          new_idx = re_dfa_add_node (dfa, *token, 0);
+          tree = create_tree (tree_first, tree_last, 0, new_idx);
+          if (idx_first == -1 || idx_last == -1 || new_idx == -1
+              || tree_first == NULL || tree_last == NULL || tree == NULL)
+            return *err = REG_ESPACE, NULL;
+        }
+      else
+        {
+          new_idx = re_dfa_add_node (dfa, *token, 0);
+          tree = create_tree (NULL, NULL, 0, new_idx);
+          if (new_idx == -1 || tree == NULL)
+            return *err = REG_ESPACE, NULL;
+        }
+      /* We must return here, since ANCHORs can't be followed
+         by repetition operators.
+         eg. RE"^*" is invalid or "<ANCHOR(^)><CHAR(*)>",
+             it must not be "<ANCHOR(^)><REPEAT(*)>".  */
+      *token = fetch_token (regexp, syntax);
+      return tree;
+    case OP_PERIOD:
+      new_idx = re_dfa_add_node (dfa, *token, 0);
+      tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+      if (MB_CUR_MAX > 1)
+        dfa->has_mb_node = 1;
+      break;
+    case OP_WORD:
+      tree = build_word_op (dfa, 0, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return NULL;
+      break;
+    case OP_NOTWORD:
+      tree = build_word_op (dfa, 1, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return NULL;
+      break;
+    case OP_ALT:
+    case END_OF_RE:
+      return NULL;
+    case BACK_SLASH:
+      *err = REG_EESCAPE;
+      return NULL;
+    default:
+      /* Must not happen?  */
+#ifdef DEBUG
+      assert (0);
+#endif
+      return NULL;
+    }
+  *token = fetch_token (regexp, syntax);
+
+  while (token->type == OP_DUP_ASTERISK || token->type == OP_DUP_PLUS
+         || token->type == OP_DUP_QUESTION || token->type == OP_OPEN_DUP_NUM)
+    {
+      tree = parse_dup_op (tree, regexp, dfa, token, syntax, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return *err = REG_ESPACE, NULL;
+    }
+
+  return tree;
+}
+
+/* This function build the following tree, from regular expression
+   (<reg_exp>):
+         SUBEXP
+            |
+        <reg_exp>
+*/
+
+static bin_tree_t *
+parse_sub_exp (regexp, preg, token, syntax, nest, err)
+     re_string_t *regexp;
+     regex_t *preg;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     int nest;
+     reg_errcode_t *err;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  bin_tree_t *tree;
+  size_t cur_nsub;
+  cur_nsub = preg->re_nsub++;
+  if (dfa->subexps_alloc < preg->re_nsub)
+    {
+      re_subexp_t *new_array;
+      dfa->subexps_alloc *= 2;
+      new_array = re_realloc (dfa->subexps, re_subexp_t, dfa->subexps_alloc);
+      if (new_array == NULL)
+	{
+	  dfa->subexps_alloc /= 2;
+	  *err = REG_ESPACE;
+	  return NULL;
+	}
+      dfa->subexps = new_array;
+    }
+  dfa->subexps[cur_nsub].start = dfa->nodes_len;
+  dfa->subexps[cur_nsub].end = -1;
+  *token = fetch_token (regexp, syntax);
+
+  /* The subexpression may be a null string.  */
+  if (token->type == OP_CLOSE_SUBEXP)
+    {
+      tree = create_tree (NULL, NULL, SUBEXP, 0);
+      if (tree == NULL)
+        return *err = REG_ESPACE, NULL;
+      dfa->subexps[cur_nsub].end = dfa->nodes_len;
+    }
+  else
+    {
+      tree = parse_reg_exp (regexp, preg, token, syntax, nest, err);
+      if (*err != REG_NOERROR && tree == NULL)
+        return NULL;
+      dfa->subexps[cur_nsub].end = dfa->nodes_len;
+      if (token->type != OP_CLOSE_SUBEXP)
+        {
+          free_bin_tree (tree);
+          *err = REG_BADPAT;
+          return NULL;
+        }
+      tree = create_tree (tree, NULL, SUBEXP, 0);
+    }
+  return tree;
+}
+
+/* This function parse repetition operators like "*", "+", "{1,3}" etc.  */
+
+static bin_tree_t *
+parse_dup_op (dup_elem, regexp, dfa, token, syntax, err)
+     bin_tree_t *dup_elem;
+     re_string_t *regexp;
+     re_dfa_t *dfa;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     reg_errcode_t *err;
+{
+  re_token_t dup_token;
+  bin_tree_t *tree = dup_elem, *work_tree;
+  int new_idx, start_idx = re_string_cur_idx (regexp);
+  re_token_t start_token = *token;
+  if (token->type == OP_OPEN_DUP_NUM)
+    {
+      int i, end, start = fetch_number (regexp, token, syntax);
+      bin_tree_t *elem;
+      if (start == -1)
+        start = 0; /* We treat "{,m}" as "{0,m}".  */
+      if (start != -2 && token->type == OP_CLOSE_DUP_NUM)
+        {
+          if (start == 0)
+            {
+              /* We treat "<re>{0}" as null string.  */
+              *token = fetch_token (regexp, syntax);
+              free_bin_tree (dup_elem);
+              return NULL;
+            }
+          end = start; /* We treat "{n}" as "{n,n}".  */
+        }
+      else if (start == -2 || token->type != CHARACTER || token->opr.c != ',')
+        /* Invalid sequence.  */
+        goto parse_dup_op_invalid_interval;
+      else
+        {
+          end = fetch_number (regexp, token, syntax);
+          if (end == -2 || token->type != OP_CLOSE_DUP_NUM)
+            /* Invalid sequence.  */
+            goto parse_dup_op_invalid_interval;
+        }
+      /* Extract "<re>{n,m}" to "<re><re>...<re><re>{0,<m-n>}".  */
+      elem = tree;
+      for (i = 0; i < start; ++i)
+        if (i != 0)
+          {
+            work_tree = duplicate_tree (elem, dfa);
+            tree = create_tree (tree, work_tree, CONCAT, 0);
+            if (work_tree == NULL || tree == NULL)
+              goto parse_dup_op_espace;
+          }
+
+      if (end == -1)
+        {
+          /* We treat "<re>{0,}" as "<re>*".  */
+          dup_token.type = OP_DUP_ASTERISK;
+          if (start > 0)
+            {
+              elem = duplicate_tree (elem, dfa);
+              new_idx = re_dfa_add_node (dfa, dup_token, 0);
+              work_tree = create_tree (elem, NULL, 0, new_idx);
+              tree = create_tree (tree, work_tree, CONCAT, 0);
+              if (elem == NULL || new_idx == -1 || work_tree == NULL
+                  || tree == NULL)
+                goto parse_dup_op_espace;
+            }
+          else
+            {
+              new_idx = re_dfa_add_node (dfa, dup_token, 0);
+              tree = create_tree (elem, NULL, 0, new_idx);
+              if (new_idx == -1 || tree == NULL)
+                goto parse_dup_op_espace;
+            }
+        }
+      else if (end - start > 0)
+        {
+          /* Then extract "<re>{0,m}" to "<re>?<re>?...<re>?".  */
+          dup_token.type = OP_DUP_QUESTION;
+          if (start > 0)
+            {
+              elem = duplicate_tree (elem, dfa);
+              new_idx = re_dfa_add_node (dfa, dup_token, 0);
+              elem = create_tree (elem, NULL, 0, new_idx);
+              tree = create_tree (tree, elem, CONCAT, 0);
+              if (elem == NULL || new_idx == -1 || tree == NULL)
+                goto parse_dup_op_espace;
+            }
+          else
+            {
+              new_idx = re_dfa_add_node (dfa, dup_token, 0);
+              tree = elem = create_tree (elem, NULL, 0, new_idx);
+              if (new_idx == -1 || tree == NULL)
+                goto parse_dup_op_espace;
+            }
+          for (i = 1; i < end - start; ++i)
+            {
+              work_tree = duplicate_tree (elem, dfa);
+              tree = create_tree (tree, work_tree, CONCAT, 0);
+              if (work_tree == NULL || tree == NULL)
+                return *err = REG_ESPACE, NULL;
+            }
+        }
+    }
+  else
+    {
+      new_idx = re_dfa_add_node (dfa, *token, 0);
+      tree = create_tree (tree, NULL, 0, new_idx);
+      if (new_idx == -1 || tree == NULL)
+        return *err = REG_ESPACE, NULL;
+    }
+  *token = fetch_token (regexp, syntax);
+  return tree;
+
+ parse_dup_op_espace:
+  free_bin_tree (tree);
+  *err = REG_ESPACE;
+  return NULL;
+
+ parse_dup_op_invalid_interval:
+  if (!(syntax & RE_INVALID_INTERVAL_ORD))
+    {
+      *err = REG_EBRACE;
+      return NULL;
+    }
+  re_string_set_index (regexp, start_idx);
+  *token = start_token;
+  token->type = CHARACTER;
+  return dup_elem;
+}
+
+/* Size of the names for collating symbol/equivalence_class/character_class.
+   I'm not sure, but maybe enough.  */
+#define BRACKET_NAME_BUF_SIZE 32
+
+static inline void *
+extend_array_for_cset (array, num, alloc, type_size)
+     void *array;
+     int num, *alloc, type_size;
+{
+  void *new_array = array;
+  if (*alloc == num)
+    {
+      if (*alloc == 0)
+        {
+          new_array = malloc (type_size);
+          *alloc = 1;
+        }
+      else
+        {
+          new_array = realloc (array, type_size * num * 2);
+          *alloc = 2 * num;
+        }
+    }
+  return new_array;
+}
+
+/* This function parse bracket expression like "[abc]", "[a-c]",
+   "[[.a-a.]]" etc.  */
+
+static bin_tree_t *
+parse_bracket_exp (regexp, dfa, token, syntax, err)
+     re_string_t *regexp;
+     re_dfa_t *dfa;
+     re_token_t *token;
+     reg_syntax_t syntax;
+     reg_errcode_t *err;
+{
+#ifdef _LIBC
+  const unsigned char *collseqmb, *collseqwc;
+  uint32_t nrules;
+  int32_t table_size;
+  const int32_t *symb_table;
+  const unsigned char *extra;
+
+  /* Local function for parse_bracket_exp.
+     Seek the collating symbol entry correspondings to NAME.
+     Return the index of the symbol in the SYMB_TABLE.  */
+
+  static inline int32_t
+  seek_collating_symbol_entry (name, name_len)
+         unsigned char *name;
+         size_t name_len;
+    {
+      int32_t hash = elem_hash (name, name_len);
+      int32_t elem = hash % table_size;
+      int32_t second = hash % (table_size - 2);
+      while (symb_table[2 * elem] != 0)
+        {
+          /* First compare the hashing value.  */
+          if (symb_table[2 * elem] == hash
+              /* Compare the length of the name.  */
+              && name_len == extra[symb_table[2 * elem + 1]]
+              /* Compare the name.  */
+              && memcmp (name, &extra[symb_table[2 * elem + 1] + 1],
+                         name_len) == 0)
+            {
+              /* Yep, this is the entry.  */
+              break;
+            }
+
+          /* Next entry.  */
+          elem += second;
+        }
+      return elem;
+    }
+
+  /* Local function for parse_bracket_exp.
+     Look up the collation sequence value of BR_ELEM.
+     Return the value if succeeded, UINT_MAX otherwise.  */
+
+  static inline unsigned int
+  lookup_collation_sequence_value (br_elem)
+         bracket_elem_t *br_elem;
+    {
+      if (br_elem->type == SB_CHAR)
+        {
+          /*
+          if (MB_CUR_MAX == 1)
+          */
+          if (nrules == 0)
+            return collseqmb[br_elem->opr.ch];
+          else
+            {
+              wint_t wc = __btowc (br_elem->opr.ch);
+              return collseq_table_lookup (collseqwc, wc);
+            }
+        }
+      else if (br_elem->type == MB_CHAR)
+        {
+          return collseq_table_lookup (collseqwc, br_elem->opr.wch);
+        }
+      else if (br_elem->type == COLL_SYM)
+        {
+          if (nrules != 0)
+            {
+              int32_t elem, idx;
+              elem = seek_collating_symbol_entry (br_elem->opr.name,
+                                                  strlen (br_elem->opr.name));
+              if (symb_table[2 * elem] != 0)
+                {
+                  /* We found the entry.  */
+                  idx = symb_table[2 * elem + 1];
+                  /* Skip the name of collating element name.  */
+                  idx += 1 + extra[idx];
+                  /* Skip the byte sequence of the collating element.  */
+                  idx += 1 + extra[idx];
+                  /* Adjust for the alignment.  */
+                  idx = (idx + 3) & ~3;
+                  /* Skip the multibyte collation sequence value.  */
+                  idx += sizeof (unsigned int);
+                  /* Skip the wide char sequence of the collating element.  */
+                  idx += sizeof (unsigned int) *
+                    (1 + *(unsigned int *) (extra + idx));
+                  /* Return the collation sequence value.  */
+                  return *(unsigned int *) (extra + idx);
+                }
+              else if (symb_table[2 * elem] == 0 &&
+                       strlen (br_elem->opr.name) == 1)
+                {
+                  /* No valid character.  Match it as a single byte
+                     character.  */
+                  return collseqmb[br_elem->opr.name[0]];
+                }
+            }
+          else if (strlen (br_elem->opr.name) == 1)
+            return collseqmb[br_elem->opr.name[0]];
+        }
+      return UINT_MAX;
+    }
+
+  /* Local function for parse_bracket_exp.
+     Build the range expression which starts from START_ELEM, and ends
+     at END_ELEM.  The result are written to MBCSET and SBCSET.
+     RANGE_ALLOC is the allocated size of mbcset->range_starts, and
+     mbcset->range_ends, is a pointer argument sinse we may
+     update it.  */
+
+  static inline reg_errcode_t
+  build_range_exp (mbcset, sbcset, range_alloc, start_elem, end_elem)
+         re_charset_t *mbcset;
+         re_bitset_ptr_t sbcset;
+         int *range_alloc;
+         bracket_elem_t *start_elem, *end_elem;
+    {
+      unsigned int ch;
+      uint32_t start_collseq;
+      uint32_t end_collseq;
+
+      /* Check the space of the arrays.  */
+      if (*range_alloc == mbcset->nranges)
+        {
+          /* There are not enough space, need realloc.  */
+          uint32_t *new_array_start;
+          uint32_t *new_array_end;
+	  int new_nranges;
+
+	  /* XXX If mbcset->range_starts and mbcset->range_ends are NULL
+	     if *range_alloc == 0 then we do not need the if.  */
+          if (*range_alloc == 0)
+            {
+              new_nranges = 1;
+              new_array_start = re_malloc (uint32_t, 1);
+              new_array_end = re_malloc (uint32_t, 1);
+            }
+          else
+            {
+              new_nranges = 2 * mbcset->nranges;
+              new_array_start = re_realloc (mbcset->range_starts, uint32_t,
+                                            new_nranges);
+              new_array_end = re_realloc (mbcset->range_ends, uint32_t,
+                                          new_nranges);
+            }
+          if (new_array_start == NULL || new_array_end == NULL)
+            return REG_ESPACE;
+
+          mbcset->range_starts = new_array_start;
+          mbcset->range_ends = new_array_end;
+	  *range_alloc = new_nranges;
+        }
+
+      if (start_elem->type == EQUIV_CLASS || start_elem->type == CHAR_CLASS
+          || end_elem->type == EQUIV_CLASS || end_elem->type == CHAR_CLASS)
+        return REG_ERANGE;
+
+      start_collseq = lookup_collation_sequence_value (start_elem);
+      end_collseq = lookup_collation_sequence_value (end_elem);
+      /* Check start/end collation sequence values.  */
+      if (start_collseq == UINT_MAX || end_collseq == UINT_MAX)
+        return REG_ECOLLATE;
+      if ((syntax & RE_NO_EMPTY_RANGES) && start_collseq > end_collseq)
+        return REG_ERANGE;
+
+      /* Got valid collation sequence values, add them as a new entry.  */
+      mbcset->range_starts[mbcset->nranges] = start_collseq;
+      mbcset->range_ends[mbcset->nranges++] = end_collseq;
+
+      /* Build the table for single byte characters.  */
+      for (ch = 0; ch <= SBC_MAX; ch++)
+        {
+          uint32_t ch_collseq;
+          /*
+          if (MB_CUR_MAX == 1)
+          */
+          if (nrules == 0)
+            ch_collseq = collseqmb[ch];
+          else
+            ch_collseq = collseq_table_lookup (collseqwc, __btowc (ch));
+          if (start_collseq <= ch_collseq && ch_collseq <= end_collseq)
+            bitset_set (sbcset, ch);
+        }
+      return REG_NOERROR;
+    }
+#endif
+
+  /* Local function for parse_bracket_exp.
+     Build the collating element which is represented by NAME.
+     The result are written to MBCSET and SBCSET.
+     COLL_SYM_ALLOC is the allocated size of mbcset->coll_sym, is a
+     pointer argument sinse we may update it.  */
+
+  static inline reg_errcode_t
+  build_collating_symbol (mbcset, sbcset, coll_sym_alloc, name)
+         re_charset_t *mbcset;
+         re_bitset_ptr_t sbcset;
+         int *coll_sym_alloc;
+         unsigned char *name;
+    {
+#ifdef _LIBC
+      int32_t elem, idx;
+      if (nrules != 0)
+        {
+          elem = seek_collating_symbol_entry (name, strlen (name));
+          if (symb_table[2 * elem] != 0)
+            {
+              /* We found the entry.  */
+              idx = symb_table[2 * elem + 1];
+              /* Skip the name of collating element name.  */
+              idx += 1 + extra[idx];
+            }
+          else if (symb_table[2 * elem] == 0 && strlen (name) == 1)
+            {
+              /* No valid character, treat it as a normal
+                 character.  */
+              bitset_set (sbcset, name[0]);
+              return REG_NOERROR;
+            }
+          else
+            return REG_ECOLLATE;
+
+          /* Got valid collation sequence, add it as a new entry.  */
+          /* Check the space of the arrays.  */
+          mbcset->coll_syms = extend_array_for_cset (mbcset->coll_syms,
+                                                     mbcset->ncoll_syms,
+                                                     coll_sym_alloc,
+                                                     sizeof (int32_t));
+          if (mbcset->coll_syms == NULL)
+            return REG_ESPACE;
+
+          mbcset->coll_syms[mbcset->ncoll_syms++] = idx;
+          return REG_NOERROR;
+        }
+      else
+#endif
+        {
+          if (strlen (name) != 1)
+            return REG_ECOLLATE;
+          else
+            {
+              bitset_set (sbcset, name[0]);
+              return REG_NOERROR;
+            }
+        }
+    }
+  re_token_t br_token;
+  re_bitset_ptr_t sbcset;
+  re_charset_t *mbcset;
+  bin_tree_t *work_tree;
+  int token_len, new_idx;
+  int coll_sym_alloc = 0, range_alloc = 0, mbchar_alloc = 0;
+  int equiv_class_alloc = 0, char_class_alloc = 0;
+#ifdef _LIBC
+  collseqmb = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB);
+  nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+  if (nrules)
+    {
+      /*
+      if (MB_CUR_MAX > 1)
+      */
+        collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC);
+      table_size = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_SYMB_HASH_SIZEMB);
+      symb_table = (const int32_t *) _NL_CURRENT (LC_COLLATE,
+                                                  _NL_COLLATE_SYMB_TABLEMB);
+      extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE,
+                                                   _NL_COLLATE_SYMB_EXTRAMB);
+    }
+#endif
+  sbcset = (re_bitset_ptr_t) calloc (sizeof (unsigned int), BITSET_UINTS);
+  mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1);
+  if (sbcset == NULL || mbcset == NULL)
+    {
+      *err = REG_ESPACE;
+      return NULL;
+    }
+
+  token_len = peek_token_bracket (token, regexp, syntax);
+  if (token->type == END_OF_RE)
+    {
+      re_free (sbcset);
+      free_charset (mbcset);
+      *err = REG_BADPAT;
+      return NULL;
+    }
+  if (token->type == OP_NON_MATCH_LIST)
+    {
+      int i;
+      mbcset->non_match = 1;
+      if (syntax & RE_HAT_LISTS_NOT_NEWLINE)
+        bitset_set (sbcset, '\0');
+      re_string_skip_bytes (regexp, token_len); /* Skip a token.  */
+      token_len = peek_token_bracket (token, regexp, syntax);
+      if (token->type == END_OF_RE)
+        {
+          re_free (sbcset);
+          free_charset (mbcset);
+          *err = REG_BADPAT;
+          return NULL;
+        }
+      if (MB_CUR_MAX > 1)
+        for (i = 0; i < SBC_MAX; ++i)
+          if (__btowc (i) == WEOF)
+            bitset_set (sbcset, i);
+    }
+
+  /* We treat the first ']' as a normal character.  */
+  if (token->type == OP_CLOSE_BRACKET)
+    token->type = CHARACTER;
+
+  while (1)
+    {
+      bracket_elem_t start_elem, end_elem;
+      unsigned char start_name_buf[BRACKET_NAME_BUF_SIZE];
+      unsigned char end_name_buf[BRACKET_NAME_BUF_SIZE];
+      reg_errcode_t ret;
+      int token_len2 = 0, is_range_exp = 0;
+      re_token_t token2;
+
+      start_elem.opr.name = start_name_buf;
+      ret = parse_bracket_element (&start_elem, regexp, token, token_len, dfa,
+                                   syntax);
+      if (ret != REG_NOERROR)
+        goto parse_bracket_exp_espace;
+
+      token_len = peek_token_bracket (token, regexp, syntax);
+      if (token->type == END_OF_RE)
+        {
+          re_free (sbcset);
+          free_charset (mbcset);
+          *err = REG_BADPAT;
+          return NULL;
+        }
+      if (token->type == OP_CHARSET_RANGE)
+        {
+          re_string_skip_bytes (regexp, token_len); /* Skip '-'.  */
+          token_len2 = peek_token_bracket (&token2, regexp, syntax);
+          if (token->type == END_OF_RE)
+            {
+              re_free (sbcset);
+              free_charset (mbcset);
+              *err = REG_BADPAT;
+              return NULL;
+            }
+          if (token2.type == OP_CLOSE_BRACKET)
+            {
+              /* We treat the last '-' as a normal character.  */
+              re_string_skip_bytes (regexp, -token_len);
+              token->type = CHARACTER;
+            }
+          else
+            is_range_exp = 1;
+        }
+
+      if (is_range_exp == 1)
+        {
+          end_elem.opr.name = end_name_buf;
+          ret = parse_bracket_element (&end_elem, regexp, &token2, token_len2,
+                                       dfa, syntax);
+          if (ret != REG_NOERROR)
+            goto parse_bracket_exp_espace;
+
+          token_len = peek_token_bracket (token, regexp, syntax);
+          if (token->type == END_OF_RE)
+            {
+              re_free (sbcset);
+              free_charset (mbcset);
+              *err = REG_BADPAT;
+              return NULL;
+            }
+          *err = build_range_exp (mbcset, sbcset, &range_alloc, &start_elem,
+				  &end_elem);
+          if (*err != REG_NOERROR)
+            {
+              re_free (sbcset);
+              free_charset (mbcset);
+              return NULL;
+            }
+        }
+      else
+        {
+          switch (start_elem.type)
+            {
+            case SB_CHAR:
+              bitset_set (sbcset, start_elem.opr.ch);
+              break;
+            case MB_CHAR:
+              mbcset->mbchars = extend_array_for_cset (mbcset->mbchars,
+						       mbcset->nmbchars,
+						       &mbchar_alloc,
+						       sizeof (wchar_t));
+              if (mbcset->mbchars == NULL)
+                goto parse_bracket_exp_espace;
+              mbcset->mbchars[mbcset->nmbchars++] = start_elem.opr.wch;
+              break;
+            case EQUIV_CLASS:
+              *err = build_equiv_class (mbcset, sbcset, &equiv_class_alloc,
+					start_elem.opr.name);
+              if (*err != REG_NOERROR)
+                {
+                  re_free (sbcset);
+                  free_charset (mbcset);
+                  return NULL;
+                }
+              break;
+            case COLL_SYM:
+              *err = build_collating_symbol (mbcset, sbcset, &coll_sym_alloc,
+					     start_elem.opr.name);
+              if (*err != REG_NOERROR)
+                {
+                  re_free (sbcset);
+                  free_charset (mbcset);
+                  return NULL;
+                }
+              break;
+            case CHAR_CLASS:
+              ret = build_charclass (mbcset, sbcset, &char_class_alloc,
+				     start_elem.opr.name);
+              if (ret != REG_NOERROR)
+               goto parse_bracket_exp_espace;
+              break;
+            default:
+              assert (0);
+              break;
+            }
+        }
+      if (token->type == OP_CLOSE_BRACKET)
+        break;
+    }
+
+  re_string_skip_bytes (regexp, token_len); /* Skip a token.  */
+
+  /* If it is non-matching list.  */
+  if (mbcset->non_match)
+    bitset_not (sbcset);
+
+  /* Build a tree for simple bracket.  */
+  br_token.type = SIMPLE_BRACKET;
+  br_token.opr.sbcset = sbcset;
+  new_idx = re_dfa_add_node (dfa, br_token, 0);
+  work_tree = create_tree (NULL, NULL, 0, new_idx);
+  if (new_idx == -1 || work_tree == NULL)
+    goto parse_bracket_exp_espace;
+
+  if (mbcset->nmbchars || mbcset->ncoll_syms || mbcset->nequiv_classes
+      || mbcset->nranges || (mbcset->nchar_classes && MB_CUR_MAX > 1))
+    {
+      re_token_t alt_token;
+      bin_tree_t *mbc_tree;
+      /* Build a tree for complex bracket.  */
+      br_token.type = COMPLEX_BRACKET;
+      br_token.opr.mbcset = mbcset;
+      dfa->has_mb_node = 1;
+      new_idx = re_dfa_add_node (dfa, br_token, 0);
+      mbc_tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || mbc_tree == NULL)
+        goto parse_bracket_exp_espace;
+      /* Then join them by ALT node.  */
+      alt_token.type = OP_ALT;
+      new_idx = re_dfa_add_node (dfa, alt_token, 0);
+      work_tree = create_tree (work_tree, mbc_tree, 0, new_idx);
+      if (new_idx != -1 && mbc_tree != NULL)
+        return work_tree;
+    }
+  else
+    {
+      free_charset (mbcset);
+      return work_tree;
+    }
+
+ parse_bracket_exp_espace:
+  free_charset (mbcset);
+  *err = REG_ESPACE;
+  return NULL;
+}
+
+static reg_errcode_t
+parse_bracket_element (elem, regexp, token, token_len, dfa, syntax)
+     bracket_elem_t *elem;
+     re_string_t *regexp;
+     re_token_t *token;
+     int token_len;
+     re_dfa_t *dfa;
+     reg_syntax_t syntax;
+{
+#ifdef RE_ENABLE_I18N
+  int cur_char_size;
+  cur_char_size = re_string_char_size_at (regexp, re_string_cur_idx (regexp));
+  if (cur_char_size > 1)
+    {
+      elem->type = MB_CHAR;
+      elem->opr.wch = re_string_wchar_at (regexp, re_string_cur_idx (regexp));
+      re_string_skip_bytes (regexp, cur_char_size);
+      return REG_NOERROR;
+    }
+#endif /* RE_ENABLE_I18N */
+  re_string_skip_bytes (regexp, token_len); /* Skip a token.  */
+  if (token->type == OP_OPEN_COLL_ELEM || token->type == OP_OPEN_CHAR_CLASS
+      || token->type == OP_OPEN_EQUIV_CLASS)
+    return parse_bracket_symbol (elem, regexp, token);
+  elem->type = SB_CHAR;
+  elem->opr.ch = token->opr.c;
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+parse_bracket_symbol (elem, regexp, token)
+     bracket_elem_t *elem;
+     re_string_t *regexp;
+     re_token_t *token;
+{
+  unsigned char ch, delim = token->opr.c;
+  int i = 0;
+  for (;; i++)
+    {
+#ifdef DEBUG
+      assert (i < BRACKET_NAME_BUF_SIZE);
+#endif
+      if (token->type == OP_OPEN_CHAR_CLASS)
+        ch = re_string_fetch_byte_case (regexp);
+      else
+        ch = re_string_fetch_byte (regexp);
+      if (ch == delim && re_string_peek_byte (regexp, 0) == ']')
+        break;
+      elem->opr.name[i] = ch;
+    }
+  re_string_skip_bytes (regexp, 1);
+  elem->opr.name[i] = '\0';
+  switch (token->type)
+    {
+    case OP_OPEN_COLL_ELEM:
+      elem->type = COLL_SYM;
+      break;
+    case OP_OPEN_EQUIV_CLASS:
+      elem->type = EQUIV_CLASS;
+      break;
+    case OP_OPEN_CHAR_CLASS:
+      elem->type = CHAR_CLASS;
+      break;
+    default:
+      break;
+    }
+  return REG_NOERROR;
+}
+
+  /* Helper function for parse_bracket_exp.
+     Build the equivalence class which is represented by NAME.
+     The result are written to MBCSET and SBCSET.
+     EQUIV_CLASS_ALLOC is the allocated size of mbcset->equiv_classes,
+     is a pointer argument sinse we may update it.  */
+
+static reg_errcode_t
+build_equiv_class (mbcset, sbcset, equiv_class_alloc, name)
+     re_charset_t *mbcset;
+     re_bitset_ptr_t sbcset;
+     int *equiv_class_alloc;
+     const unsigned char *name;
+{
+#ifdef _LIBC
+  uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+  if (nrules != 0)
+    {
+      const int32_t *table, *indirect;
+      const unsigned char *weights, *extra, *cp;
+      unsigned char char_buf[2];
+      int32_t idx1, idx2;
+      unsigned int ch;
+      size_t len;
+      /* This #include defines a local function!  */
+# include <locale/weight.h>
+      /* Calculate the index for equivalence class.  */
+      cp = name;
+      table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+      weights = (const unsigned char *) _NL_CURRENT (LC_COLLATE,
+					       _NL_COLLATE_WEIGHTMB);
+      extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE,
+                                                   _NL_COLLATE_EXTRAMB);
+      indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
+                                                _NL_COLLATE_INDIRECTMB);
+      idx1 = findidx (&cp);
+      if (idx1 == 0 || cp < name + strlen (name))
+        /* This isn't a valid character.  */
+        return REG_ECOLLATE;
+
+      /* Build single byte matcing table for this equivalence class.  */
+      char_buf[1] = '\0';
+      len = weights[idx1];
+      for (ch = 0; ch < SBC_MAX; ++ch)
+        {
+          char_buf[0] = ch;
+          cp = char_buf;
+          idx2 = findidx (&cp);
+/*
+          idx2 = table[ch];
+*/
+          if (idx2 == 0)
+            /* This isn't a valid character.  */
+            continue;
+          if (len == weights[idx2])
+            {
+              int cnt = 0;
+              while (cnt <= len &&
+                     weights[idx1 + 1 + cnt] == weights[idx2 + 1 + cnt])
+                ++cnt;
+
+              if (cnt > len)
+                bitset_set (sbcset, ch);
+            }
+        }
+      /* Check the space of the arrays, and extend if we need.  */
+      mbcset->equiv_classes = extend_array_for_cset (mbcset->equiv_classes,
+						     mbcset->nequiv_classes,
+						     equiv_class_alloc,
+						     sizeof (int32_t));
+      if (mbcset->equiv_classes == NULL)
+        return REG_ESPACE;
+
+      mbcset->equiv_classes[mbcset->nequiv_classes++] = idx1;
+    }
+  else
+#endif
+    {
+      if (strlen (name) != 1)
+        return REG_ECOLLATE;
+      bitset_set (sbcset, name[0]);
+    }
+  return REG_NOERROR;
+}
+
+  /* Helper function for parse_bracket_exp.
+     Build the character class which is represented by NAME.
+     The result are written to MBCSET and SBCSET.
+     CHAR_CLASS_ALLOC is the allocated size of mbcset->char_classes,
+     is a pointer argument sinse we may update it.  */
+
+static reg_errcode_t
+build_charclass (mbcset, sbcset, char_class_alloc, name)
+     re_charset_t *mbcset;
+     re_bitset_ptr_t sbcset;
+     int *char_class_alloc;
+     const unsigned char *name;
+{
+  int i;
+
+  /* Check the space of the arrays.  */
+  mbcset->char_classes = extend_array_for_cset (mbcset->char_classes,
+						mbcset->nchar_classes,
+						char_class_alloc,
+						sizeof (wctype_t));
+  if (mbcset->char_classes == NULL)
+    return REG_ESPACE;
+
+  mbcset->char_classes[mbcset->nchar_classes++] = __wctype (name);
+
+#define BUILD_CHARCLASS_LOOP(ctype_func)\
+    for (i = 0; i < SBC_MAX; ++i)	\
+      {					\
+        if (ctype_func (i))		\
+          bitset_set (sbcset, i);	\
+      }
+
+  if (strcmp (name, "alnum") == 0)
+    BUILD_CHARCLASS_LOOP (isalnum)
+  else if (strcmp (name, "cntrl") == 0)
+    BUILD_CHARCLASS_LOOP (iscntrl)
+  else if (strcmp (name, "lower") == 0)
+    BUILD_CHARCLASS_LOOP (islower)
+  else if (strcmp (name, "space") == 0)
+    BUILD_CHARCLASS_LOOP (isspace)
+  else if (strcmp (name, "alpha") == 0)
+    BUILD_CHARCLASS_LOOP (isalpha)
+  else if (strcmp (name, "digit") == 0)
+    BUILD_CHARCLASS_LOOP (isdigit)
+  else if (strcmp (name, "print") == 0)
+    BUILD_CHARCLASS_LOOP (isprint)
+  else if (strcmp (name, "upper") == 0)
+    BUILD_CHARCLASS_LOOP (isupper)
+  else if (strcmp (name, "blank") == 0)
+    BUILD_CHARCLASS_LOOP (isblank)
+  else if (strcmp (name, "graph") == 0)
+    BUILD_CHARCLASS_LOOP (isgraph)
+  else if (strcmp (name, "punct") == 0)
+    BUILD_CHARCLASS_LOOP (ispunct)
+  else if (strcmp (name, "xdigit") == 0)
+    BUILD_CHARCLASS_LOOP (isxdigit)
+  else
+    return REG_ECTYPE;
+
+  return REG_NOERROR;
+}
+
+static bin_tree_t *
+build_word_op (dfa, not, err)
+     re_dfa_t *dfa;
+     int not;
+     reg_errcode_t *err;
+{
+  re_bitset_ptr_t sbcset;
+  re_charset_t *mbcset;
+  reg_errcode_t ret;
+  re_token_t br_token;
+  bin_tree_t *tree;
+  int new_idx, alloc = 0;
+
+  sbcset = (re_bitset_ptr_t) calloc (sizeof (unsigned int), BITSET_UINTS);
+  mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1);
+  if (sbcset == NULL || mbcset == NULL)
+    {
+      *err = REG_ESPACE;
+      return NULL;
+    }
+
+  if (not)
+    {
+      int i;
+      mbcset->non_match = 1;
+      /*
+      if (syntax & RE_HAT_LISTS_NOT_NEWLINE)
+        bitset_set(cset->sbcset, '\0');
+      */
+      if (MB_CUR_MAX > 1)
+        for (i = 0; i < SBC_MAX; ++i)
+          if (__btowc (i) == WEOF)
+            bitset_set (sbcset, i);
+    }
+
+  ret = build_charclass (mbcset, sbcset, &alloc, "alpha");
+  if (ret != REG_NOERROR)
+    {
+      re_free (sbcset);
+      free_charset (mbcset);
+      *err = REG_ESPACE;
+      return NULL;
+    }
+
+  /* If it is non-matching list.  */
+  if (mbcset->non_match)
+    bitset_not (sbcset);
+
+  /* Build a tree for simple bracket.  */
+  br_token.type = SIMPLE_BRACKET;
+  br_token.opr.sbcset = sbcset;
+  new_idx = re_dfa_add_node (dfa, br_token, 0);
+  tree = create_tree (NULL, NULL, 0, new_idx);
+  if (new_idx == -1 || tree == NULL)
+    goto build_word_op_espace;
+
+  if (MB_CUR_MAX > 1)
+    {
+      re_token_t alt_token;
+      bin_tree_t *mbc_tree;
+      /* Build a tree for complex bracket.  */
+      br_token.type = COMPLEX_BRACKET;
+      br_token.opr.mbcset = mbcset;
+      dfa->has_mb_node = 1;
+      new_idx = re_dfa_add_node (dfa, br_token, 0);
+      mbc_tree = create_tree (NULL, NULL, 0, new_idx);
+      if (new_idx == -1 || mbc_tree == NULL)
+        goto build_word_op_espace;
+      /* Then join them by ALT node.  */
+      alt_token.type = OP_ALT;
+      new_idx = re_dfa_add_node (dfa, alt_token, 0);
+      tree = create_tree (tree, mbc_tree, 0, new_idx);
+      if (new_idx != -1 && mbc_tree != NULL)
+        return tree;
+    }
+  else
+    {
+      free_charset (mbcset);
+      return tree;
+    }
+ build_word_op_espace:
+  re_free (sbcset);
+  free_charset (mbcset);
+  *err = REG_ESPACE;
+  return NULL;
+}
+
+/* This is intended for the expressions like "a{1,3}".
+   Fetch a number from `input', and return the number.
+   Return -1, if the number field is empty like "{,1}".
+   Return -2, If an error is occured.  */
+
+static int
+fetch_number (input, token, syntax)
+     re_string_t *input;
+     re_token_t *token;
+     reg_syntax_t syntax;
+{
+  int num = -1;
+  unsigned char c;
+  while (1)
+    {
+      *token = fetch_token (input, syntax);
+      c = token->opr.c;
+      if (token->type == OP_CLOSE_DUP_NUM || c == ',')
+        break;
+      if (token->type != CHARACTER || c < '0' || '9' < c)
+        return -2;
+      num = (num == -1) ? c - '0' : num * 10 + c - '0';
+    }
+  if (num > RE_DUP_MAX)
+    return -2;
+  return num;
+}
+
+static void
+free_charset (re_charset_t *cset)
+{
+  re_free (cset->mbchars);
+  re_free (cset->coll_syms);
+  re_free (cset->equiv_classes);
+  re_free (cset->range_starts);
+  re_free (cset->range_ends);
+  re_free (cset->char_classes);
+  re_free (cset);
+}
+
+/* Functions for binary tree operation.  */
+
+/* Create a node of tree.
+   Note: This function automatically free left and right if malloc fails.  */
+
+static bin_tree_t *
+create_tree (left, right, type, index)
+     bin_tree_t *left;
+     bin_tree_t *right;
+     re_token_type_t type;
+     int index;
+{
+  bin_tree_t *tree;
+  tree = re_malloc (bin_tree_t, 1);
+  if (tree == NULL)
+    {
+      free_bin_tree (left);
+      free_bin_tree (right);
+      return NULL;
+    }
+  tree->parent = NULL;
+  tree->left = left;
+  tree->right = right;
+  tree->type = type;
+  tree->node_idx = index;
+  tree->first = -1;
+  tree->next = -1;
+  re_node_set_init_empty (&tree->eclosure);
+
+  if (left != NULL)
+    left->parent = tree;
+  if (right != NULL)
+    right->parent = tree;
+  return tree;
+}
+
+/* Free the sub tree pointed by TREE.  */
+
+static void
+free_bin_tree (tree)
+     bin_tree_t *tree;
+{
+  if (tree == NULL)
+    return;
+  /*re_node_set_free (&tree->eclosure);*/
+  free_bin_tree (tree->left);
+  free_bin_tree (tree->right);
+  re_free (tree);
+}
+
+/* Duplicate the node SRC, and return new node.  */
+
+static bin_tree_t *
+duplicate_tree (src, dfa)
+     const bin_tree_t *src;
+     re_dfa_t *dfa;
+{
+  bin_tree_t *left = NULL, *right = NULL, *new_tree;
+  int new_node_idx;
+  /* Since node indies must be according to Post-order of the tree,
+     we must duplicate the left at first.  */
+  if (src->left != NULL)
+    {
+      left = duplicate_tree (src->left, dfa);
+      if (left == NULL)
+        return NULL;
+    }
+
+  /* Secondaly, duplicate the right.  */
+  if (src->right != NULL)
+    {
+      right = duplicate_tree (src->right, dfa);
+      if (right == NULL)
+        {
+          free_bin_tree (left);
+          return NULL;
+        }
+    }
+
+  /* At last, duplicate itself.  */
+  if (src->type == NON_TYPE)
+    {
+      new_node_idx = re_dfa_add_node (dfa, dfa->nodes[src->node_idx], 0);
+      dfa->nodes[new_node_idx].duplicated = 1;
+      if (new_node_idx == -1)
+        {
+          free_bin_tree (left);
+          free_bin_tree (right);
+          return NULL;
+        }
+    }
+  else
+    new_node_idx = src->type;
+
+  new_tree = create_tree (left, right, src->type, new_node_idx);
+  if (new_tree == NULL)
+    {
+      free_bin_tree (left);
+      free_bin_tree (right);
+    }
+  return new_tree;
+}
diff --git a/posix/regex.h b/posix/regex.h
index 8b8bb9d5a0..d1e4b6841a 100644
--- a/posix/regex.h
+++ b/posix/regex.h
@@ -1,8 +1,8 @@
 /* Definitions for data structures and routines for the regular
-   expression library, version 0.12.
-   Copyright (C) 1985,1989-93,1995-98,2000,2001 Free Software Foundation, Inc.
-   This file is part of the GNU C Library.  Its master source is NOT part of
-   the C library, however.  The master source lives in /gd/gnu/lib.
+   expression library.
+   Copyright (C) 1985,1989-93,1995-98,2000,2001,2002
+   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
@@ -164,6 +164,10 @@ typedef unsigned long int reg_syntax_t;
    treated as 'a\{1'.  */
 #define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1)
 
+/* If this bit is set, then ignore case when matching.
+   If not set, then case is significant.  */
+#define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1)
+
 /* This global variable defines the particular regexp syntax to use (for
    some interfaces).  When a regexp is compiled, the syntax used is
    stored in the pattern buffer, so changing this does not affect
diff --git a/posix/regex_internal.c b/posix/regex_internal.c
new file mode 100644
index 0000000000..63bed420cd
--- /dev/null
+++ b/posix/regex_internal.c
@@ -0,0 +1,1095 @@
+/* Extended regular expression matching and search library.
+   Copyright (C) 2002 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
+
+   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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <assert.h>
+#include <ctype.h>
+#include <limits.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <wchar.h>
+#include <wctype.h>
+
+#ifdef _LIBC
+# ifndef _RE_DEFINE_LOCALE_FUNCTIONS
+#  define _RE_DEFINE_LOCALE_FUNCTIONS 1
+#  include <locale/localeinfo.h>
+#  include <locale/elem-hash.h>
+#  include <locale/coll-lookup.h>
+# endif
+#endif
+
+/* This is for other GNU distributions with internationalized messages.  */
+#if HAVE_LIBINTL_H || defined _LIBC
+# include <libintl.h>
+# ifdef _LIBC
+#  undef gettext
+#  define gettext(msgid) __dcgettext ("libc", msgid, LC_MESSAGES)
+# endif
+#else
+# define gettext(msgid) (msgid)
+#endif
+
+#ifndef gettext_noop
+/* This define is so xgettext can find the internationalizable
+   strings.  */
+# define gettext_noop(String) String
+#endif
+
+#include "regex.h"
+#include "regex_internal.h"
+
+static void re_string_construct_common (const unsigned char *str,
+                                        int len, re_string_t *pstr);
+#ifdef RE_ENABLE_I18N
+static reg_errcode_t build_wcs_buffer (re_string_t *pstr);
+static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr);
+#endif /* RE_ENABLE_I18N */
+static reg_errcode_t build_upper_buffer (re_string_t *pstr);
+static reg_errcode_t re_string_translate_buffer (re_string_t *pstr,
+						 RE_TRANSLATE_TYPE trans);
+static re_dfastate_t *create_newstate_common (re_dfa_t *dfa,
+                                              const re_node_set *nodes,
+                                              unsigned int hash);
+static re_dfastate_t *create_ci_newstate (re_dfa_t *dfa,
+                                          const re_node_set *nodes,
+                                          unsigned int hash);
+static re_dfastate_t *create_cd_newstate (re_dfa_t *dfa,
+                                          const re_node_set *nodes,
+                                          unsigned int context,
+                                          unsigned int hash);
+static unsigned int inline calc_state_hash (const re_node_set *nodes,
+                                            unsigned int context);
+
+/* Functions for string operation.  */
+
+/* Construct string object.  */
+static reg_errcode_t
+re_string_construct (pstr, str, len, trans)
+     re_string_t *pstr;
+     const unsigned char *str;
+     int len;
+     RE_TRANSLATE_TYPE trans;
+{
+  reg_errcode_t ret;
+  re_string_construct_common (str, len, pstr);
+#ifdef RE_ENABLE_I18N
+  if (MB_CUR_MAX >1 && pstr->len > 0)
+    {
+      ret = build_wcs_buffer (pstr);
+      if (ret != REG_NOERROR)
+        return ret;
+    }
+#endif /* RE_ENABLE_I18N  */
+  pstr->mbs_case = str;
+  if (trans != NULL)
+    {
+      ret = re_string_translate_buffer (pstr, trans);
+      if (ret != REG_NOERROR)
+        return ret;
+    }
+  return REG_NOERROR;
+}
+
+/* Construct string object. We use this function instead of
+   re_string_construct for case insensitive mode.  */
+
+static reg_errcode_t
+re_string_construct_toupper (pstr, str, len, trans)
+     re_string_t *pstr;
+     const unsigned char *str;
+     int len;
+     RE_TRANSLATE_TYPE trans;
+{
+  reg_errcode_t ret;
+  /* Set case sensitive buffer.  */
+  re_string_construct_common (str, len, pstr);
+#ifdef RE_ENABLE_I18N
+  if (MB_CUR_MAX >1)
+    {
+      if (pstr->len > 0)
+        {
+          ret = build_wcs_upper_buffer (pstr);
+          if (ret != REG_NOERROR)
+            return ret;
+        }
+    }
+  else
+#endif /* RE_ENABLE_I18N  */
+    {
+      if (pstr->len > 0)
+        {
+          ret = build_upper_buffer (pstr);
+          if (ret != REG_NOERROR)
+            return ret;
+        }
+    }
+  pstr->mbs_case = str;
+  if (trans != NULL)
+    {
+      ret = re_string_translate_buffer (pstr, trans);
+      if (ret != REG_NOERROR)
+        return ret;
+    }
+  return REG_NOERROR;
+}
+
+/* Helper functions for re_string_construct_*.  */
+static void
+re_string_construct_common (str, len, pstr)
+     const unsigned char *str;
+     int len;
+     re_string_t *pstr;
+{
+  pstr->mbs = str;
+  pstr->cur_idx = 0;
+  pstr->len = len;
+#ifdef RE_ENABLE_I18N
+  pstr->wcs = NULL;
+#endif
+  pstr->mbs_case = NULL;
+  pstr->mbs_alloc = 0;
+  pstr->mbs_case_alloc = 0;
+}
+
+#ifdef RE_ENABLE_I18N
+
+/* Build wide character buffer for `pstr'.
+   If the byte sequence of the string are:
+     <mb1>(0), <mb1>(1), <mb2>(0), <mb2>(1), <sb3>
+   Then wide character buffer will be:
+     <wc1>   , WEOF    , <wc2>   , WEOF    , <wc3>
+   We use WEOF for padding, they indicate that the position isn't
+   a first byte of a multibyte character.  */
+
+static reg_errcode_t
+build_wcs_buffer (pstr)
+     re_string_t *pstr;
+{
+  mbstate_t state, prev_st;
+  wchar_t wc;
+  int char_idx, char_len, mbclen;
+
+  pstr->wcs = re_malloc (wchar_t, pstr->len + 1);
+  if (pstr->wcs == NULL)
+    return REG_ESPACE;
+
+  memset (&state, '\0', sizeof (mbstate_t));
+  char_len = pstr->len;
+  for (char_idx = 0; char_idx < char_len ;)
+    {
+      int next_idx, remain_len = char_len - char_idx;
+      prev_st = state;
+      mbclen = mbrtowc (&wc, pstr->mbs + char_idx, remain_len, &state);
+      if (mbclen == (size_t) -2 || mbclen == (size_t) -1 || mbclen == 0)
+        /* We treat these cases as a singlebyte character.  */
+        {
+          mbclen = 1;
+          wc = (wchar_t) pstr->mbs[char_idx++];
+          state = prev_st;
+        }
+      /* Write wide character and padding.  */
+      pstr->wcs[char_idx++] = wc;
+      for (next_idx = char_idx + mbclen - 1; char_idx < next_idx ;)
+        pstr->wcs[char_idx++] = WEOF;
+    }
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+build_wcs_upper_buffer (pstr)
+     re_string_t *pstr;
+{
+  mbstate_t state, prev_st;
+  wchar_t wc;
+  unsigned char *mbs_upper;
+  int char_idx, char_len, mbclen;
+
+  pstr->wcs = re_malloc (wchar_t, pstr->len + 1);
+  mbs_upper = re_malloc (unsigned char, pstr->len + 1);
+  if (pstr->wcs == NULL || mbs_upper == NULL)
+    {
+      pstr->wcs = NULL;
+      return REG_ESPACE;
+    }
+
+  memset (&state, '\0', sizeof (mbstate_t));
+  char_len = pstr->len;
+  for (char_idx = 0 ; char_idx < char_len ; char_idx += mbclen)
+    {
+      int byte_idx, remain_len = char_len - char_idx;
+      prev_st = state;
+      mbclen = mbrtowc (&wc, pstr->mbs + char_idx, remain_len, &state);
+      if (mbclen == 1)
+        {
+          pstr->wcs[char_idx] = wc;
+          if (islower (pstr->mbs[char_idx]))
+            mbs_upper[char_idx] = toupper (pstr->mbs[char_idx]);
+          else
+            mbs_upper[char_idx] = pstr->mbs[char_idx];
+        }
+      else if (mbclen == (size_t) -2 || mbclen == (size_t) -1 || mbclen == 0)
+        /* We treat these cases as a singlebyte character.  */
+        {
+          mbclen = 1;
+          pstr->wcs[char_idx] = (wchar_t) pstr->mbs[char_idx];
+          mbs_upper[char_idx] = pstr->mbs[char_idx];
+          state = prev_st;
+        }
+      else /* mbclen > 1 */
+        {
+          pstr->wcs[char_idx] = wc;
+          if (iswlower (wc))
+            wcrtomb (mbs_upper + char_idx, towupper (wc), &prev_st);
+          else
+            memcpy (mbs_upper + char_idx, pstr->mbs + char_idx, mbclen);
+          for (byte_idx = 1 ; byte_idx < mbclen ; byte_idx++)
+            pstr->wcs[char_idx + byte_idx] = WEOF;
+        }
+    }
+  pstr->mbs = mbs_upper;
+  pstr->mbs_alloc = 1;
+  return REG_NOERROR;
+}
+#endif /* RE_ENABLE_I18N  */
+
+static reg_errcode_t
+build_upper_buffer (pstr)
+     re_string_t *pstr;
+{
+  unsigned char *mbs_upper;
+  int char_idx, char_len;
+
+  mbs_upper = re_malloc (unsigned char, pstr->len + 1);
+  if (mbs_upper == NULL)
+    return REG_ESPACE;
+
+  char_len = pstr->len;
+  for (char_idx = 0 ; char_idx < char_len ; char_idx ++)
+    {
+      if (islower (pstr->mbs[char_idx]))
+        mbs_upper[char_idx] = toupper (pstr->mbs[char_idx]);
+      else
+        mbs_upper[char_idx] = pstr->mbs[char_idx];
+    }
+  pstr->mbs = mbs_upper;
+  pstr->mbs_alloc = 1;
+  return REG_NOERROR;
+}
+
+/* Apply TRANS to the buffer in PSTR.  We assume that wide char buffer
+   is already constructed if MB_CUR_MAX > 1.  */
+
+static reg_errcode_t
+re_string_translate_buffer (pstr, trans)
+     re_string_t *pstr;
+     RE_TRANSLATE_TYPE trans;
+{
+  int buf_idx;
+  unsigned char *transed_buf, *transed_case_buf;
+#ifdef DEBUG
+  assert (trans != NULL);
+#endif
+  if (pstr->mbs_alloc)
+    {
+      transed_buf = (unsigned char *) pstr->mbs;
+      transed_case_buf = re_malloc (unsigned char, pstr->len + 1);
+      if (transed_case_buf == NULL)
+        return REG_ESPACE;
+      pstr->mbs_case_alloc = 1;
+    }
+  else
+    {
+      transed_buf = re_malloc (unsigned char, pstr->len + 1);
+      if (transed_buf == NULL)
+        return REG_ESPACE;
+      transed_case_buf = NULL;
+      pstr->mbs_alloc = 1;
+    }
+  for (buf_idx = 0 ; buf_idx < pstr->len ; buf_idx++)
+    {
+#ifdef RE_ENABLE_I18N
+      if (MB_CUR_MAX > 1 && !re_string_is_single_byte_char (pstr, buf_idx))
+        transed_buf[buf_idx] = pstr->mbs[buf_idx];
+      else
+#endif
+        transed_buf[buf_idx] = trans[pstr->mbs[buf_idx]];
+      if (transed_case_buf)
+        {
+#ifdef RE_ENABLE_I18N
+         if (MB_CUR_MAX > 1 && !re_string_is_single_byte_char (pstr, buf_idx))
+            transed_case_buf[buf_idx] = pstr->mbs_case[buf_idx];
+          else
+#endif
+            transed_case_buf[buf_idx] = trans[pstr->mbs_case[buf_idx]];
+        }
+    }
+  if (pstr->mbs_case_alloc == 1)
+    {
+      pstr->mbs = transed_buf;
+      pstr->mbs_case = transed_case_buf;
+    }
+  else
+    {
+      pstr->mbs = transed_buf;
+      pstr->mbs_case = transed_buf;
+    }
+  return REG_NOERROR;
+}
+
+static void
+re_string_destruct (pstr)
+     re_string_t *pstr;
+{
+#ifdef RE_ENABLE_I18N
+  re_free (pstr->wcs);
+#endif /* RE_ENABLE_I18N  */
+  if (pstr->mbs_alloc)
+    re_free ((void *) pstr->mbs);
+  if (pstr->mbs_case_alloc)
+    re_free ((void *) pstr->mbs_case);
+}
+
+/* Return the context at IDX in INPUT.  */
+static unsigned int
+re_string_context_at (input, idx, eflags, newline_anchor)
+     const re_string_t *input;
+     int idx, eflags, newline_anchor;
+{
+  int c;
+  if (idx < 0 || idx == input->len)
+    {
+      unsigned int context = 0;
+      if (idx < 0)
+        context = CONTEXT_BEGBUF;
+      else if (idx == input->len)
+        context = CONTEXT_ENDBUF;
+
+      if ((idx < 0 && !(eflags & REG_NOTBOL))
+          || (idx == input->len && !(eflags & REG_NOTEOL)))
+        return CONTEXT_NEWLINE | context;
+      else
+        return context;
+    }
+  c = re_string_byte_at (input, idx);
+  if (IS_WORD_CHAR (c))
+    return CONTEXT_WORD;
+  return (newline_anchor && IS_NEWLINE (c)) ? CONTEXT_NEWLINE : 0;
+}
+
+/* Functions for set operation.  */
+
+static reg_errcode_t
+re_node_set_alloc (set, size)
+     re_node_set *set;
+     int size;
+{
+  set->alloc = size;
+  set->nelem = 0;
+  set->elems = re_malloc (int, size);
+  if (set->elems == NULL)
+    return REG_ESPACE;
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_init_1 (set, elem)
+     re_node_set *set;
+     int elem;
+{
+  set->alloc = 1;
+  set->nelem = 1;
+  set->elems = re_malloc (int, 1);
+  if (set->elems == NULL)
+    return REG_ESPACE;
+  set->elems[0] = elem;
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_init_2 (set, elem1, elem2)
+     re_node_set *set;
+     int elem1, elem2;
+{
+  set->alloc = 2;
+  set->elems = re_malloc (int, 2);
+  if (set->elems == NULL)
+    return REG_ESPACE;
+  if (elem1 == elem2)
+    {
+      set->nelem = 1;
+      set->elems[0] = elem1;
+    }
+  else
+    {
+      set->nelem = 2;
+      if (elem1 < elem2)
+        {
+          set->elems[0] = elem1;
+          set->elems[1] = elem2;
+        }
+      else
+        {
+          set->elems[0] = elem2;
+          set->elems[1] = elem1;
+        }
+    }
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_init_copy (dest, src)
+     re_node_set *dest;
+     const re_node_set *src;
+{
+  dest->nelem = src->nelem;
+  if (src->nelem > 0)
+    {
+      dest->alloc = dest->nelem;
+      dest->elems = re_malloc (int, dest->alloc);
+      if (dest->elems == NULL)
+        return REG_ESPACE;
+      memcpy (dest->elems, src->elems, src->nelem * sizeof (int));
+    }
+  else
+    re_node_set_init_empty (dest);
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_intersect (dest, src1, src2)
+     re_node_set *dest;
+     const re_node_set *src1, *src2;
+{
+  int i1, i2, id;
+  if (src1->nelem > 0 && src2->nelem > 0)
+    {
+      if (src1->nelem + src2->nelem > dest->alloc)
+        {
+          int *new_array;
+          if (dest->alloc == 0)
+            new_array = re_malloc (int, src1->nelem + src2->nelem);
+          else
+            new_array = re_realloc (dest->elems, int,
+                                    src1->nelem + src2->nelem);
+          dest->alloc = src1->nelem + src2->nelem;
+          if (new_array == NULL)
+            return REG_ESPACE;
+          dest->elems = new_array;
+        }
+    }
+  else
+    {
+      dest->nelem = 0;
+      return REG_NOERROR;
+    }
+
+  for (i1 = i2 = id = 0 ; i1 < src1->nelem && i2 < src2->nelem ;)
+    {
+      if (src1->elems[i1] > src2->elems[i2])
+        {
+          ++i2;
+          continue;
+        }
+      if (src1->elems[i1] == src2->elems[i2])
+        dest->elems[id++] = src2->elems[i2++];
+      ++i1;
+    }
+  dest->nelem = id;
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_add_intersect (dest, src1, src2)
+     re_node_set *dest;
+     const re_node_set *src1, *src2;
+{
+  int i1, i2, id;
+  if (src1->nelem > 0 && src2->nelem > 0)
+    {
+      if (src1->nelem + src2->nelem + dest->nelem > dest->alloc)
+        {
+          int *new_array;
+          if (dest->alloc == 0)
+            new_array = re_malloc (int, src1->nelem + src2->nelem);
+          else
+            new_array = re_realloc (dest->elems, int,
+                                    src1->nelem + src2->nelem + dest->nelem);
+          dest->alloc = src1->nelem + src2->nelem + dest->nelem;
+          if (new_array == NULL)
+            return REG_ESPACE;
+          dest->elems = new_array;
+        }
+    }
+  else
+    return REG_NOERROR;
+
+  for (i1 = i2 = id = 0 ; i1 < src1->nelem && i2 < src2->nelem ;)
+    {
+      if (src1->elems[i1] > src2->elems[i2])
+        {
+          ++i2;
+          continue;
+        }
+      if (src1->elems[i1] == src2->elems[i2])
+        {
+          while (id < dest->nelem && dest->elems[id] < src2->elems[i2])
+            ++id;
+          if (id < dest->nelem && dest->elems[id] == src2->elems[i2])
+            ++id;
+          else
+            {
+              memmove (dest->elems + id + 1, dest->elems + id,
+                       sizeof (int) * (dest->nelem - id));
+              dest->elems[id++] = src2->elems[i2++];
+              ++dest->nelem;
+            }
+        }
+      ++i1;
+    }
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_init_union (dest, src1, src2)
+     re_node_set *dest;
+     const re_node_set *src1, *src2;
+{
+  int i1, i2, id;
+  if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0)
+    {
+      dest->alloc = src1->nelem + src2->nelem;
+      dest->elems = re_malloc (int, dest->alloc);
+      if (dest->elems == NULL)
+        return REG_ESPACE;
+    }
+  else
+    {
+      if (src1 != NULL && src1->nelem > 0)
+        return re_node_set_init_copy (dest, src1);
+      else if (src2 != NULL && src2->nelem > 0)
+        return re_node_set_init_copy (dest, src2);
+      else
+        re_node_set_init_empty (dest);
+      return REG_NOERROR;
+    }
+  for (i1 = i2 = id = 0 ; i1 < src1->nelem && i2 < src2->nelem ;)
+    {
+      if (src1->elems[i1] > src2->elems[i2])
+        {
+          dest->elems[id++] = src2->elems[i2++];
+          continue;
+        }
+      if (src1->elems[i1] == src2->elems[i2])
+        ++i2;
+      dest->elems[id++] = src1->elems[i1++];
+    }
+  if (i1 < src1->nelem)
+    {
+      memcpy (dest->elems + id, src1->elems + i1,
+             (src1->nelem - i1) * sizeof (int));
+      id += src1->nelem - i1;
+    }
+  else if (i2 < src2->nelem)
+    {
+      memcpy (dest->elems + id, src2->elems + i2,
+             (src2->nelem - i2) * sizeof (int));
+      id += src2->nelem - i2;
+    }
+  dest->nelem = id;
+  return REG_NOERROR;
+}
+
+static reg_errcode_t
+re_node_set_merge (dest, src)
+     re_node_set *dest;
+     const re_node_set *src;
+{
+  int si, di;
+  if (src == NULL || src->nelem == 0)
+    return REG_NOERROR;
+  else if (dest == NULL)
+    {
+      dest = re_malloc (re_node_set, 1);
+      return re_node_set_init_copy (dest, src);
+    }
+  if (dest->alloc < src->nelem + dest->nelem)
+    {
+      dest->alloc = 2 * (src->nelem + dest->alloc);
+      dest->elems = re_realloc (dest->elems, int, dest->alloc);
+    }
+
+  for (si = 0, di = 0 ; si < src->nelem && di < dest->nelem ;)
+    {
+      int cp_from, ncp, mid, right, src_elem = src->elems[si];
+      /* Binary search the spot we will add the new element.  */
+      right = dest->nelem;
+      while (di < right)
+        {
+          mid = (di + right) / 2;
+          if (dest->elems[mid] < src_elem)
+            di = mid + 1;
+          else
+            right = mid;
+        }
+      if (di >= dest->nelem)
+        break;
+
+      if (dest->elems[di] == src_elem)
+        {
+          /* Skip since, DEST already has the element.  */
+          ++di;
+          ++si;
+          continue;
+        }
+
+      /* Skip the src elements which are less than dest->elems[di].  */
+      cp_from = si;
+      while (si < src->nelem && src->elems[si] < dest->elems[di])
+        ++si;
+      /* Copy these src elements.  */
+      ncp = si - cp_from;
+      memmove (dest->elems + di + ncp, dest->elems + di,
+               sizeof (int) * (dest->nelem - di));
+      memcpy (dest->elems + di, src->elems + cp_from,
+              sizeof (int) * ncp);
+      /* Update counters.  */
+      di += ncp;
+      dest->nelem += ncp;
+    }
+
+  /* Copy remaining src elements.  */
+  if (si < src->nelem)
+    {
+      memcpy (dest->elems + di, src->elems + si,
+              sizeof (int) * (src->nelem - si));
+      dest->nelem += src->nelem - si;
+    }
+  return REG_NOERROR;
+}
+
+/* Insert the new element ELEM to the re_node_set* SET.
+   return 0 if SET already has ELEM,
+   return -1 if an error is occured, return 1 otherwise.  */
+
+static int
+re_node_set_insert (set, elem)
+     re_node_set *set;
+     int elem;
+{
+  int idx, right, mid;
+  /* In case of the set is empty.  */
+  if (set->elems == NULL || set->alloc == 0)
+    {
+      if (re_node_set_init_1 (set, elem) == REG_NOERROR)
+        return 1;
+      else
+        return -1;
+    }
+
+  /* Binary search the spot we will add the new element.  */
+  idx = 0;
+  right = set->nelem;
+  while (idx < right)
+    {
+      mid = (idx + right) / 2;
+      if (set->elems[mid] < elem)
+        idx = mid + 1;
+      else
+        right = mid;
+    }
+
+  /* Realloc if we need.  */
+  if (set->alloc < set->nelem + 1)
+    {
+      int *new_array;
+      set->alloc = set->alloc * 2;
+      new_array = re_malloc (int, set->alloc);
+      if (new_array == NULL)
+        return -1;
+      /* Copy the elements they are followed by the new element.  */
+      if (idx > 0)
+        memcpy (new_array, set->elems, sizeof (int) * (idx));
+      /* Copy the elements which follows the new element.  */
+      if (set->nelem - idx > 0)
+        memcpy (new_array + idx + 1, set->elems + idx,
+		sizeof (int) * (set->nelem - idx));
+      set->elems = new_array;
+    }
+  else
+    {
+      /* Move the elements which follows the new element.  */
+      if (set->nelem - idx > 0)
+        memmove (set->elems + idx + 1, set->elems + idx,
+                 sizeof (int) * (set->nelem - idx));
+    }
+  /* Insert the new element.  */
+  set->elems[idx] = elem;
+  ++set->nelem;
+  return 1;
+}
+
+/* Compare two node sets SET1 and SET2.
+   return 1 if SET1 and SET2 are equivalent, retrun 0 otherwise.  */
+
+static int
+re_node_set_compare (set1, set2)
+     const re_node_set *set1, *set2;
+{
+  int i;
+  if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem)
+    return 0;
+  for (i = 0 ; i < set1->nelem ; i++)
+    if (set1->elems[i] != set2->elems[i])
+      return 0;
+  return 1;
+}
+
+/* Return 1 if SET contains the element ELEM, return 0 otherwise.  */
+
+static int
+re_node_set_contains (set, elem)
+     const re_node_set *set;
+     int elem;
+{
+  int idx, right, mid;
+  if (set->nelem <= 0)
+    return 0;
+
+  /* Binary search the element.  */
+  idx = 0;
+  right = set->nelem - 1;
+  while (idx < right)
+    {
+      mid = (idx + right) / 2;
+      if (set->elems[mid] < elem)
+        idx = mid + 1;
+      else
+        right = mid;
+    }
+  return set->elems[idx] == elem;
+}
+
+static void
+re_node_set_remove_at (set, idx)
+     re_node_set *set;
+     int idx;
+{
+  if (idx < 0 || idx >= set->nelem)
+    return;
+  if (idx < set->nelem - 1)
+    memmove (set->elems + idx, set->elems + idx + 1,
+             sizeof (int) * (set->nelem - idx - 1));
+  --set->nelem;
+}
+
+
+/* Add the token TOKEN to dfa->nodes, and return the index of the token.
+   Or return -1, if an error will be occured.  */
+
+static int
+re_dfa_add_node (dfa, token, mode)
+     re_dfa_t *dfa;
+     re_token_t token;
+     int mode;
+{
+  if (dfa->nodes_len >= dfa->nodes_alloc)
+    {
+      re_token_t *new_array;
+      dfa->nodes_alloc *= 2;
+      new_array = re_realloc (dfa->nodes, re_token_t, dfa->nodes_alloc);
+      if (new_array == NULL)
+        return -1;
+      else
+        dfa->nodes = new_array;
+      if (mode)
+        {
+          int *new_firsts, *new_nexts;
+          re_node_set *new_edests, *new_eclosures, *new_inveclosures;
+
+          new_firsts = re_realloc (dfa->firsts, int, dfa->nodes_alloc);
+          new_nexts = re_realloc (dfa->nexts, int, dfa->nodes_alloc);
+          new_edests = re_realloc (dfa->edests, re_node_set, dfa->nodes_alloc);
+          new_eclosures = re_realloc (dfa->eclosures, re_node_set,
+                                      dfa->nodes_alloc);
+          new_inveclosures = re_realloc (dfa->inveclosures, re_node_set,
+                                         dfa->nodes_alloc);
+          if (new_firsts == NULL || new_nexts == NULL || new_edests == NULL
+              || new_eclosures == NULL || new_inveclosures == NULL)
+            return -1;
+          dfa->firsts = new_firsts;
+          dfa->nexts = new_nexts;
+          dfa->edests = new_edests;
+          dfa->eclosures = new_eclosures;
+          dfa->inveclosures = new_inveclosures;
+        }
+    }
+  dfa->nodes[dfa->nodes_len] = token;
+  dfa->nodes[dfa->nodes_len].duplicated = 0;
+  return dfa->nodes_len++;
+}
+
+static unsigned int inline
+calc_state_hash (nodes, context)
+     const re_node_set *nodes;
+     unsigned int context;
+{
+  unsigned int hash = nodes->nelem + context;
+  int i;
+  for (i = 0 ; i < nodes->nelem ; i++)
+    hash += nodes->elems[i];
+  return hash;
+}
+
+/* Search for the state whose node_set is equivalent to NODES.
+   Return the pointer to the state, if we found it in the DFA.
+   Otherwise create the new one and return it.  */
+
+static re_dfastate_t *
+re_acquire_state (dfa, nodes)
+     re_dfa_t *dfa;
+     const re_node_set *nodes;
+{
+  unsigned int hash;
+  struct re_state_table_entry *spot;
+  int i;
+  if (nodes->nelem == 0)
+    return NULL;
+  hash = calc_state_hash (nodes, 0);
+  spot = dfa->state_table + (hash & dfa->state_hash_mask);
+
+  if (spot->alloc == 0)
+    {
+      /* Currently there are only one state in this spot.  */
+      if (spot->entry.state != NULL && hash == spot->entry.state->hash
+          && re_node_set_compare (&spot->entry.state->nodes, nodes))
+        return spot->entry.state;
+    }
+  else
+    for (i = 0 ; i < spot->num ; i++)
+      {
+        re_dfastate_t *state = spot->entry.array[i];
+        if (hash != state->hash)
+          continue;
+        if (re_node_set_compare (&state->nodes, nodes))
+          return state;
+      }
+
+  /* There are no appropriate state in the dfa, create the new one.  */
+  return create_ci_newstate (dfa, nodes, hash);
+}
+
+/* Search for the state whose node_set is equivalent to NODES and
+   whose context is equivalent to CONTEXT.
+   Return the pointer to the state, if we found it in the DFA.
+   Otherwise create the new one and return it.  */
+
+static re_dfastate_t *
+re_acquire_state_context (dfa, nodes, context)
+     re_dfa_t *dfa;
+     const re_node_set *nodes;
+     unsigned int context;
+{
+  unsigned int hash;
+  struct re_state_table_entry *spot;
+  int i;
+  if (nodes->nelem == 0)
+    return NULL;
+  hash = calc_state_hash (nodes, context);
+  spot = dfa->state_table + (hash & dfa->state_hash_mask);
+
+  if (spot->alloc == 0)
+    {
+      /* Currently there are only one state in this spot.  */
+      if (spot->entry.state != NULL && hash == spot->entry.state->hash
+          && re_node_set_compare (&spot->entry.state->nodes, nodes)
+          && spot->entry.state->context == context)
+        return spot->entry.state;
+    }
+  else
+    for (i = 0 ; i < spot->num ; i++)
+      {
+        re_dfastate_t *state = spot->entry.array[i];
+        if (hash != state->hash)
+          continue;
+        if (re_node_set_compare (state->entrance_nodes, nodes)
+            && state->context == context)
+          return state;
+      }
+  /* There are no appropriate state in `dfa', create the new one.  */
+  return create_cd_newstate (dfa, nodes, context, hash);
+}
+
+static re_dfastate_t *
+create_newstate_common (dfa, nodes, hash)
+     re_dfa_t *dfa;
+     const re_node_set *nodes;
+     unsigned int hash;
+{
+  re_dfastate_t *newstate;
+  newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1);
+  re_node_set_init_copy (&newstate->nodes, nodes);
+  newstate->trtable = NULL;
+  newstate->trtable_search = NULL;
+  newstate->hash = hash;
+  return newstate;
+}
+
+static void
+register_state (dfa, newstate, hash)
+     re_dfa_t *dfa;
+     re_dfastate_t *newstate;
+     unsigned int hash;
+{
+  struct re_state_table_entry *spot;
+  spot = dfa->state_table + (hash & dfa->state_hash_mask);
+
+  if (spot->alloc <= spot->num)
+    {
+      re_dfastate_t **new_array;
+
+      /* XXX Is spot->entry.array == NULL if spot->alloc == 0?  If yes
+	 the if can go away and only realloc is needed.  */
+      if (spot->alloc == 0)
+        {
+          spot->alloc = 4;
+          new_array = re_malloc (re_dfastate_t *, spot->alloc);
+	  if (new_array == NULL)
+	    /* XXX return value */
+	    return;
+          new_array[0] = spot->entry.state;
+        }
+      else
+        {
+          spot->alloc = 2 * spot->num;
+          new_array = re_realloc (spot->entry.array, re_dfastate_t *,
+                                  spot->alloc);
+        }
+      spot->entry.array = new_array;
+    }
+  spot->entry.array[spot->num++] = newstate;
+}
+
+static re_dfastate_t *
+create_ci_newstate (dfa, nodes, hash)
+     re_dfa_t *dfa;
+     const re_node_set *nodes;
+     unsigned int hash;
+{
+  int i;
+  re_dfastate_t *newstate;
+  newstate = create_newstate_common (dfa, nodes, hash);
+  newstate->entrance_nodes = &newstate->nodes;
+
+  for (i = 0 ; i < nodes->nelem ; i++)
+    {
+      re_token_t *node = dfa->nodes + nodes->elems[i];
+      re_token_type_t type = node->type;
+      if (type == CHARACTER)
+        continue;
+
+      /* If the state has the halt node, the state is a halt state.  */
+      else if (type == END_OF_RE)
+        newstate->halt = 1;
+      else if (type == COMPLEX_BRACKET
+               || (type == OP_PERIOD && MB_CUR_MAX > 1))
+        newstate->accept_mb = 1;
+      else if (type == OP_BACK_REF)
+        newstate->has_backref = 1;
+      else if (type == ANCHOR || OP_CONTEXT_NODE)
+        {
+          newstate->has_constraint = 1;
+          if (type == OP_CONTEXT_NODE
+              && dfa->nodes[node->opr.ctx_info->entity].type == END_OF_RE)
+            newstate->halt = 1;
+        }
+    }
+
+  register_state (dfa, newstate, hash);
+  return newstate;
+}
+
+static re_dfastate_t *
+create_cd_newstate (dfa, nodes, context, hash)
+     re_dfa_t *dfa;
+     const re_node_set *nodes;
+     unsigned int context, hash;
+{
+  int i, nctx_nodes = 0;
+  re_dfastate_t *newstate;
+
+  newstate = create_newstate_common (dfa, nodes, hash);
+  newstate->context = context;
+  newstate->entrance_nodes = &newstate->nodes;
+
+  for (i = 0 ; i < nodes->nelem ; i++)
+    {
+      unsigned int constraint = 0;
+      re_token_t *node = dfa->nodes + nodes->elems[i];
+      re_token_type_t type = node->type;
+      if (type == CHARACTER)
+        continue;
+
+      /* If the state has the halt node, the state is a halt state.  */
+      else if (type == END_OF_RE)
+        newstate->halt = 1;
+      else if (type == COMPLEX_BRACKET
+               || (type == OP_PERIOD && MB_CUR_MAX > 1))
+        newstate->accept_mb = 1;
+      else if (type == OP_BACK_REF)
+        newstate->has_backref = 1;
+      else if (type == ANCHOR)
+        constraint = node->opr.ctx_type;
+      else if (type == OP_CONTEXT_NODE)
+        {
+          re_token_type_t ctype = dfa->nodes[node->opr.ctx_info->entity].type;
+          constraint = node->constraint;
+          if (ctype == END_OF_RE)
+            newstate->halt = 1;
+          else if (ctype == OP_BACK_REF)
+            newstate->has_backref = 1;
+          else if (ctype == COMPLEX_BRACKET
+                   || (type == OP_PERIOD && MB_CUR_MAX > 1))
+            newstate->accept_mb = 1;
+        }
+
+      if (constraint)
+        {
+          if (newstate->entrance_nodes == &newstate->nodes)
+            {
+              newstate->entrance_nodes = re_malloc (re_node_set, 1);
+	      if (newstate->entrance_nodes == NULL)
+		/* XXX Return which value?  */
+		return NULL;
+              re_node_set_init_copy (newstate->entrance_nodes, nodes);
+              nctx_nodes = 0;
+              newstate->has_constraint = 1;
+            }
+
+          if (NOT_SATISFY_PREV_CONSTRAINT (constraint,context))
+            {
+              re_node_set_remove_at (&newstate->nodes, i - nctx_nodes);
+              ++nctx_nodes;
+            }
+        }
+    }
+  register_state (dfa, newstate, hash);
+  return newstate;
+}
diff --git a/posix/regex_internal.h b/posix/regex_internal.h
new file mode 100644
index 0000000000..35f9f4a868
--- /dev/null
+++ b/posix/regex_internal.h
@@ -0,0 +1,542 @@
+/* Extended regular expression matching and search library.
+   Copyright (C) 2002 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
+
+   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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#ifndef _REGEX_INTERNAL_H
+#define _REGEX_INTERNAL_H 1
+
+/* Number of bits in a byte.  */
+#define BYTE_BITS 8
+/* Number of single byte character.  */
+#define SBC_MAX 256
+
+#define COLL_ELEM_LEN_MAX 8
+
+/* The character which represents newline.  */
+#define NEWLINE_CHAR '\n'
+
+/* Rename to standard API for using out of glibc.  */
+#ifndef _LIBC
+# define __wctype wctype
+# define __iswctype iswctype
+# define __btowc btowc
+# define __mempcpy memcpy
+#endif
+
+extern const char re_error_msgid[];
+extern const size_t re_error_msgid_idx[];
+
+/* Number of bits in an unsinged int.  */
+#define UINT_BITS (sizeof (unsigned int) * BYTE_BITS)
+/* Number of unsigned int in an bit_set.  */
+#define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS)
+typedef unsigned int bitset[BITSET_UINTS];
+typedef unsigned int *re_bitset_ptr_t;
+
+#define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS)
+#define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS))
+#define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS))
+#define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS)
+#define bitset_set_all(set) \
+  memset (set, 255, sizeof (unsigned int) * BITSET_UINTS)
+#define bitset_copy(dest,src) \
+  memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS)
+static inline void bitset_not (bitset set);
+static inline void bitset_merge (bitset dest, const bitset src);
+static inline void bitset_not_merge (bitset dest, const bitset src);
+
+#define PREV_WORD_CONSTRAINT 0x0001
+#define PREV_NOTWORD_CONSTRAINT 0x0002
+#define NEXT_WORD_CONSTRAINT 0x0004
+#define NEXT_NOTWORD_CONSTRAINT 0x0008
+#define PREV_NEWLINE_CONSTRAINT 0x0010
+#define NEXT_NEWLINE_CONSTRAINT 0x0020
+#define PREV_BEGBUF_CONSTRAINT 0x0040
+#define NEXT_ENDBUF_CONSTRAINT 0x0080
+#define DUMMY_CONSTRAINT 0x0100
+
+typedef enum
+{
+  INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
+  WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
+  WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
+  LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
+  LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
+  BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
+  BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
+  WORD_DELIM = DUMMY_CONSTRAINT
+} re_context_type;
+
+typedef struct
+{
+  int alloc;
+  int nelem;
+  int *elems;
+} re_node_set;
+
+typedef enum
+{
+  NON_TYPE = 0,
+
+  /* Token type, these are used only by token.  */
+  OP_OPEN_SUBEXP,
+  OP_CLOSE_SUBEXP,
+  OP_OPEN_BRACKET,
+  OP_CLOSE_BRACKET,
+  OP_CHARSET_RANGE,
+  OP_OPEN_DUP_NUM,
+  OP_CLOSE_DUP_NUM,
+  OP_NON_MATCH_LIST,
+  OP_OPEN_COLL_ELEM,
+  OP_CLOSE_COLL_ELEM,
+  OP_OPEN_EQUIV_CLASS,
+  OP_CLOSE_EQUIV_CLASS,
+  OP_OPEN_CHAR_CLASS,
+  OP_CLOSE_CHAR_CLASS,
+  OP_WORD,
+  OP_NOTWORD,
+  BACK_SLASH,
+
+  /* Tree type, these are used only by tree. */
+  CONCAT,
+  ALT,
+  SUBEXP,
+  SIMPLE_BRACKET,
+  COMPLEX_BRACKET,
+
+  /* Node type, These are used by token, node, tree.  */
+  OP_PERIOD,
+  CHARACTER,
+  END_OF_RE,
+  OP_ALT,
+  OP_DUP_ASTERISK,
+  OP_DUP_PLUS,
+  OP_DUP_QUESTION,
+  OP_BACK_REF,
+  ANCHOR,
+  OP_CONTEXT_NODE,
+
+  /* Dummy marker.  */
+  END_OF_RE_TOKEN_T
+} re_token_type_t;
+
+typedef struct
+{
+  /* If this character set is the non-matching list.  */
+  unsigned int non_match : 1;
+
+  /* Multibyte characters.  */
+  wchar_t *mbchars;
+  int nmbchars;
+
+  /* Collating symbols.  */
+#ifdef _LIBC
+  int32_t *coll_syms;
+#endif
+  int ncoll_syms;
+
+  /* Equivalence classes. */
+#ifdef _LIBC
+  int32_t *equiv_classes;
+#endif
+  int nequiv_classes;
+
+  /* Range expressions. */
+#ifdef _LIBC
+  uint32_t *range_starts;
+  uint32_t *range_ends;
+#endif
+  int nranges;
+
+  /* Character classes. */
+  wctype_t *char_classes;
+  int nchar_classes;
+} re_charset_t;
+
+typedef struct
+{
+  re_token_type_t type;
+  union
+  {
+    unsigned char c;		/* for CHARACTER */
+    re_bitset_ptr_t sbcset;	/* for SIMPLE_BRACKET */
+    re_charset_t *mbcset;	/* for COMPLEX_BRACKET */
+    int idx;			/* for BACK_REF */
+    re_context_type ctx_type;	/* for ANCHOR */
+    struct
+    {
+      int entity;		/* for OP_CONTEXT_NODE, index of the entity */
+      re_node_set *bkref_eclosure;
+    } *ctx_info;
+  } opr;
+  unsigned int constraint : 10;	/* context constraint */
+  unsigned int duplicated : 1;
+#ifdef RE_ENABLE_I18N
+  unsigned int mb_partial : 1;
+#endif
+} re_token_t;
+
+#define IS_EPSILON_NODE(type) \
+  ((type) == OP_ALT || (type) == OP_DUP_ASTERISK || (type) == OP_DUP_PLUS || \
+   (type) == OP_DUP_QUESTION || (type) == ANCHOR)
+
+#define ACCEPT_MB_NODE(type) \
+  ((type) == COMPLEX_BRACKET || (type) == OP_PERIOD)
+
+struct re_string_t
+{
+  /* Store the multibyte string.  In case of "case insensitive mode" like
+     REG_ICASE, upper cases of the string are stored.  */
+  const unsigned char *mbs;
+  /* Store the case sensitive multibyte string.  In case of
+     "case insensitive mode", the original string are stored,
+     otherwise MBS_CASE points the same address that MBS points.  */
+  const unsigned char *mbs_case;
+  int cur_idx;
+  int len;
+#ifdef RE_ENABLE_I18N
+  /* Store the wide character string which is corresponding to MBS.  */
+  wchar_t *wcs;
+#endif
+  /* 1 if mbs is allocated by regex library.  */
+  unsigned int mbs_alloc : 1;
+  /* 1 if mbs_case is allocated by regex library.  */
+  unsigned int mbs_case_alloc : 1;
+};
+typedef struct re_string_t re_string_t;
+
+static reg_errcode_t re_string_construct (re_string_t *pstr,
+					  const unsigned char *str, int len,
+					  RE_TRANSLATE_TYPE trans);
+static reg_errcode_t re_string_construct_toupper (re_string_t *pstr,
+						  const unsigned char *str,
+						  int len,
+						  RE_TRANSLATE_TYPE trans);
+static void re_string_destruct (re_string_t *pstr);
+#ifdef RE_ENABLE_I18N
+static int re_string_elem_size_at (const re_string_t *pstr, int idx);
+static inline int re_string_char_size_at (const re_string_t *pstr, int idx);
+static inline wint_t re_string_wchar_at (const re_string_t *pstr, int idx);
+#endif /* RE_ENABLE_I18N */
+static unsigned int re_string_context_at (const re_string_t *input, int idx,
+					  int eflags, int newline_anchor);
+#define re_string_peek_byte(pstr, offset) \
+  ((pstr)->mbs[(pstr)->cur_idx + offset])
+#define re_string_peek_byte_case(pstr, offset) \
+  ((pstr)->mbs_case[(pstr)->cur_idx + offset])
+#define re_string_fetch_byte(pstr) \
+  ((pstr)->mbs[(pstr)->cur_idx++])
+#define re_string_fetch_byte_case(pstr) \
+  ((pstr)->mbs_case[(pstr)->cur_idx++])
+#define re_string_first_byte(pstr, idx) \
+  ((idx) == (pstr)->len || (pstr)->wcs[idx] != WEOF)
+#define re_string_is_single_byte_char(pstr, idx) \
+  ((pstr)->wcs[idx] != WEOF && ((pstr)->len == (idx) \
+                                || (pstr)->wcs[(idx) + 1] != WEOF))
+#define re_string_eoi(pstr) ((pstr)->len == (pstr)->cur_idx)
+#define re_string_cur_idx(pstr) ((pstr)->cur_idx)
+#define re_string_get_buffer(pstr) ((pstr)->mbs)
+#define re_string_length(pstr) ((pstr)->len)
+#define re_string_byte_at(pstr,idx) \
+  ((pstr)->mbs[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_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
+#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
+#define re_free(p) free (p)
+
+struct bin_tree_t
+{
+  struct bin_tree_t *parent;
+  struct bin_tree_t *left;
+  struct bin_tree_t *right;
+
+  /* `node_idx' is the index in dfa->nodes, if `type' == 0.
+     Otherwise `type' indicate the type of this node.  */
+  re_token_type_t type;
+  int node_idx;
+
+  int first;
+  int next;
+  re_node_set eclosure;
+};
+typedef struct bin_tree_t bin_tree_t;
+
+struct re_backref_cache_entry
+{
+  int node;
+  int from;
+  int to;
+  int flag;
+};
+
+typedef struct
+{
+  int eflags;
+  int match_first;
+  int match_last;
+  int state_log_top;
+  /* Back reference cache.  */
+  int nbkref_ents;
+  int abkref_ents;
+  struct re_backref_cache_entry *bkref_ents;
+  int max_bkref_len;
+} re_match_context_t;
+
+
+#define CONTEXT_WORD 1
+#define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
+#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
+#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
+
+#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
+#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
+#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
+#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
+#define IS_ORDINARY_CONTEXT(c) ((c) == 0)
+
+#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
+#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
+
+#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
+ ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
+  || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
+  || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
+  || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
+
+#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
+ ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
+  || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
+  || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
+  || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
+
+struct re_dfastate_t
+{
+  unsigned int hash;
+  re_node_set nodes;
+  re_node_set *entrance_nodes;
+  struct re_dfastate_t **trtable;
+  struct re_dfastate_t **trtable_search;
+  /* If this state is a special state.
+     A state is a special state if the state is the halt state, or
+     a anchor.  */
+  unsigned int context : 2;
+  unsigned int halt : 1;
+  /* If this state can accept `multi byte'.
+     Note that we refer to multibyte characters, and multi character
+     collating elements as `multi byte'.  */
+  unsigned int accept_mb : 1;
+  /* If this state has backreference node(s).  */
+  unsigned int has_backref : 1;
+  unsigned int has_constraint : 1;
+};
+typedef struct re_dfastate_t re_dfastate_t;
+
+typedef struct
+{
+  /* start <= node < end  */
+  int start;
+  int end;
+} re_subexp_t;
+
+struct re_state_table_entry
+{
+  int num;
+  int alloc;
+  union
+  {
+    re_dfastate_t *state;
+    re_dfastate_t **array;
+  } entry;
+};
+
+struct re_dfa_t
+{
+  re_bitset_ptr_t word_char;
+
+  /* number of subexpressions `re_nsub' is in regex_t.  */
+  int subexps_alloc;
+  re_subexp_t *subexps;
+
+  re_token_t *nodes;
+  int nodes_alloc;
+  int nodes_len;
+  bin_tree_t *str_tree;
+  int *firsts;
+  int *nexts;
+  re_node_set *edests;
+  re_node_set *eclosures;
+  re_node_set *inveclosures;
+  struct re_state_table_entry *state_table;
+  unsigned int state_hash_mask;
+  re_dfastate_t *init_state;
+  re_dfastate_t *init_state_word;
+  re_dfastate_t *init_state_nl;
+  re_dfastate_t *init_state_begbuf;
+  int states_alloc;
+  int init_node;
+  int nbackref; /* The number of backreference in this dfa.  */
+  /* If this dfa has "multibyte node", which is a backreference or
+     a node which can accept multibyte character or multi character
+     collating element.  */
+  unsigned int has_mb_node : 1;
+};
+typedef struct re_dfa_t re_dfa_t;
+
+static reg_errcode_t re_node_set_alloc (re_node_set *set, int size);
+static reg_errcode_t re_node_set_init_1 (re_node_set *set, int elem);
+static reg_errcode_t re_node_set_init_2 (re_node_set *set, int elem1,
+                                         int elem2);
+#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
+static reg_errcode_t re_node_set_init_copy (re_node_set *dest,
+                                            const re_node_set *src);
+static reg_errcode_t re_node_set_intersect (re_node_set *dest,
+                                            const re_node_set *src1,
+                                            const re_node_set *src2);
+static reg_errcode_t re_node_set_add_intersect (re_node_set *dest,
+                                                const re_node_set *src1,
+                                                const re_node_set *src2);
+static reg_errcode_t re_node_set_init_union (re_node_set *dest,
+                                             const re_node_set *src1,
+                                             const re_node_set *src2);
+static reg_errcode_t re_node_set_merge (re_node_set *dest,
+                                        const re_node_set *src);
+static int re_node_set_insert (re_node_set *set, int elem);
+static int re_node_set_compare (const re_node_set *set1,
+                                const re_node_set *set2);
+static int re_node_set_contains (const re_node_set *set, int elem);
+static void re_node_set_remove_at (re_node_set *set, int idx);
+#define re_node_set_empty(p) ((p)->nelem = 0)
+#define re_node_set_free(set) re_free ((set)->elems)
+static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token, int mode);
+static re_dfastate_t *re_acquire_state (re_dfa_t *dfa,
+                                        const re_node_set *nodes);
+static re_dfastate_t *re_acquire_state_context (re_dfa_t *dfa,
+                                                const re_node_set *nodes,
+                                                unsigned int context);
+
+
+typedef enum
+{
+  SB_CHAR,
+  MB_CHAR,
+  EQUIV_CLASS,
+  COLL_SYM,
+  CHAR_CLASS
+} bracket_elem_type;
+
+typedef struct
+{
+  bracket_elem_type type;
+  union
+  {
+    unsigned char ch;
+    unsigned char *name;
+    wchar_t wch;
+  } opr;
+} bracket_elem_t;
+
+
+/* Inline functions for bitset operation.  */
+static inline void
+bitset_not (set)
+     bitset set;
+{
+  int bitset_i;
+  for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
+    set[bitset_i] = ~set[bitset_i];
+}
+
+static inline void
+bitset_merge (dest, src)
+     bitset dest;
+     const bitset src;
+{
+  int bitset_i;
+  for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
+    dest[bitset_i] |= src[bitset_i];
+}
+
+static inline void
+bitset_not_merge (dest, src)
+     bitset dest;
+     const bitset src;
+{
+  int i;
+  for (i = 0; i < BITSET_UINTS; ++i)
+    dest[i] |= ~src[i];
+}
+
+#ifdef RE_ENABLE_I18N
+/* Inline functions for re_string.  */
+static inline int
+re_string_char_size_at (pstr, idx)
+     const re_string_t *pstr;
+     int idx;
+{
+  int byte_idx;
+  if (MB_CUR_MAX == 1)
+    return 1;
+  for (byte_idx = 1; idx + byte_idx < pstr->len; ++byte_idx)
+    if (pstr->wcs[idx + byte_idx] != WEOF)
+      break;
+  return byte_idx;
+}
+
+static inline wint_t
+re_string_wchar_at (pstr, idx)
+     const re_string_t *pstr;
+     int idx;
+{
+  if (MB_CUR_MAX == 1)
+    return (wint_t) pstr->mbs[idx];
+  return (wint_t) pstr->wcs[idx];
+}
+
+static int
+re_string_elem_size_at (pstr, idx)
+     const re_string_t *pstr;
+     int idx;
+{
+#ifdef _LIBC
+  const unsigned char *p;
+  const char *extra;
+  const int32_t *table, *indirect;
+  int32_t tmp;
+# include <locale/weight.h>
+  uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+
+  if (nrules != 0)
+    {
+      table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+      extra = (const char *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
+      indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
+						_NL_COLLATE_INDIRECTMB);
+      p = pstr->mbs + idx;
+      tmp = findidx (&p);
+      return p - (const unsigned char *) pstr->mbs - idx;
+    }
+  else
+#endif /* _LIBC */
+    return 1;
+}
+#endif /* RE_ENABLE_I18N */
+
+#endif /*  _REGEX_INTERNAL_H */
diff --git a/posix/regexec.c b/posix/regexec.c
new file mode 100644
index 0000000000..cf8f304b48
--- /dev/null
+++ b/posix/regexec.c
@@ -0,0 +1,2076 @@
+/* Extended regular expression matching and search library.
+   Copyright (C) 2002 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
+
+   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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <assert.h>
+#include <ctype.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <wchar.h>
+#include <wctype.h>
+
+#ifdef _LIBC
+# ifndef _RE_DEFINE_LOCALE_FUNCTIONS
+#  define _RE_DEFINE_LOCALE_FUNCTIONS 1
+#  include <locale/localeinfo.h>
+#  include <locale/elem-hash.h>
+#  include <locale/coll-lookup.h>
+# endif
+#endif
+
+#include "regex.h"
+#include "regex_internal.h"
+
+static void match_ctx_init (re_match_context_t *cache, int eflags, int n);
+static void match_ctx_free (re_match_context_t *cache);
+static void match_ctx_add_entry (re_match_context_t *cache, int node, int from,
+                                 int to);
+static int re_search_internal (const regex_t *preg, const char *string,
+                               int length, int start, int range, size_t nmatch,
+                               regmatch_t pmatch[], int eflags);
+static inline re_dfastate_t *acquire_init_state_context (const regex_t *preg,
+                                const re_string_t *input, int idx, int eflags);
+static int check_matching (const regex_t *preg, re_string_t *input,
+                           re_match_context_t *mctx, re_dfastate_t **state_log,
+                           int start_idx, int fl_search, int fl_longest_match);
+static int check_halt_node_context (const re_dfa_t *dfa, int node,
+                                    unsigned int context);
+static int check_halt_state_context (const regex_t *preg,
+                                     const re_dfastate_t *state,
+                                     const re_string_t *input, int idx,
+                                     int eflags);
+static int proceed_next_node (const regex_t *preg,
+                              re_dfastate_t **state_log,
+                              const re_match_context_t *mctx,
+                              const re_string_t *input,
+                              int *pidx, int node, re_node_set *eps_via_nodes);
+static void set_regs (const regex_t *preg, re_dfastate_t **state_log,
+                      const re_match_context_t *mctx, const re_string_t *input,
+                      size_t nmatch, regmatch_t *pmatch, int last);
+static int sift_states_iter_mb (const regex_t *preg, re_dfastate_t **state_log,
+                                const re_match_context_t *mctx,
+                                const re_string_t *input, int node_idx,
+                                int str_idx, int max_str_idx);
+static int sift_states_iter_bkref (const re_dfa_t *dfa,
+                                   re_dfastate_t **state_log,
+                                   struct re_backref_cache_entry *mctx_entry,
+                                   int node_idx, int idx, int match_first,
+                                   int match_last);
+static void sift_states_backward (const regex_t *preg,
+                                  re_dfastate_t **state_log,
+                                  const re_match_context_t *mctx,
+                                  const re_string_t *input, int last_node);
+static void add_epsilon_backreference (const re_dfa_t *dfa,
+                                       const re_match_context_t *mctx,
+                                       const re_node_set *plog, int idx,
+                                       re_node_set *state_buf);
+static re_dfastate_t *transit_state (const regex_t *preg, re_dfastate_t *state,
+                                     re_string_t *input, int fl_search,
+                                     re_dfastate_t **state_log,
+                                     re_match_context_t *mctx);
+static re_dfastate_t *transit_state_sb (const regex_t *preg,
+                                        re_dfastate_t *pstate,
+                                        re_string_t *input, int fl_search,
+                                        re_match_context_t *mctx);
+static void transit_state_mb (const regex_t *preg, re_dfastate_t *pstate,
+                              const re_string_t *input,
+                              re_dfastate_t **state_log,
+                              re_match_context_t *mctx);
+static void transit_state_bkref (const regex_t *preg, re_dfastate_t *pstate,
+                                 const re_string_t *input,
+                                 re_dfastate_t **state_log,
+                                 re_match_context_t *mctx);
+static void transit_state_bkref_loop (const regex_t *preg,
+                                      const re_string_t *input,
+                                      re_node_set *nodes,
+                                      re_dfastate_t **work_state_log,
+                                      re_dfastate_t **state_log,
+                                      re_match_context_t *mctx);
+static re_dfastate_t **build_trtable (const regex_t *dfa,
+                                      const re_dfastate_t *state,
+                                      int fl_search);
+static int check_node_accept_bytes (const regex_t *preg, int node_idx,
+                                    const re_string_t *input, int idx);
+static unsigned int find_collation_sequence_value (const unsigned char *mbs,
+                                                   size_t name_len);
+static int group_nodes_into_DFAstates (const regex_t *dfa,
+                                       const re_dfastate_t *state,
+                                       re_node_set *states_node,
+                                       bitset *states_ch);
+static int check_node_accept (const regex_t *preg, const re_token_t *node,
+                              const re_string_t *input, int idx, int eflags);
+
+/* Entry point for POSIX code.  */
+
+/* regexec searches for a given pattern, specified by PREG, in the
+   string STRING.
+
+   If NMATCH is zero or REG_NOSUB was set in the cflags argument to
+   `regcomp', we ignore PMATCH.  Otherwise, we assume PMATCH has at
+   least NMATCH elements, and we set them to the offsets of the
+   corresponding matched substrings.
+
+   EFLAGS specifies `execution flags' which affect matching: if
+   REG_NOTBOL is set, then ^ does not match at the beginning of the
+   string; if REG_NOTEOL is set, then $ does not match at the end.
+
+   We return 0 if we find a match and REG_NOMATCH if not.  */
+
+int
+regexec (preg, string, nmatch, pmatch, eflags)
+    const regex_t *preg;
+    const char *string;
+    size_t nmatch;
+    regmatch_t pmatch[];
+    int eflags;
+{
+  int length = strlen (string);
+  if (preg->no_sub)
+    return re_search_internal (preg, string, length, 0, length, 0,
+                               NULL, eflags);
+  else
+    return re_search_internal (preg, string, length, 0, length, nmatch,
+                               pmatch, eflags);
+}
+#ifdef _LIBC
+weak_alias (__regexec, regexec)
+#endif
+
+/* Entry points for GNU code.  */
+
+/* re_match is like re_match_2 except it takes only a single string.  */
+
+int
+re_match (buffer, string, length, start, regs)
+    struct re_pattern_buffer *buffer;
+    const char *string;
+    int length, start;
+    struct re_registers *regs;
+{
+  int i, nregs, result, rval, eflags = 0;
+  regmatch_t *pmatch;
+
+  eflags |= (buffer->not_bol) ? REG_NOTBOL : 0;
+  eflags |= (buffer->not_eol) ? REG_NOTEOL : 0;
+
+  /* We need at least 1 register.  */
+  nregs = ((regs == NULL) ? 1
+           : ((regs->num_regs > buffer->re_nsub) ? buffer->re_nsub + 1
+              : regs->num_regs + 1));
+  pmatch = re_malloc (regmatch_t, nregs);
+  if (pmatch == NULL)
+    return -2;
+  result = re_search_internal (buffer, string, length, start, 0,
+                               nregs, pmatch, eflags);
+
+  /* If caller wants register contents data back, do it.  */
+  if (regs && !buffer->no_sub)
+    {
+      /* Have the register data arrays been allocated?  */
+      if (buffer->regs_allocated == REGS_UNALLOCATED)
+        { /* No.  So allocate them with malloc.  We need one
+             extra element beyond `num_regs' for the `-1' marker
+             GNU code uses.  */
+          regs->num_regs = ((RE_NREGS > buffer->re_nsub + 1) ? RE_NREGS
+                            : buffer->re_nsub + 1);
+          regs->start = re_malloc (regoff_t, regs->num_regs);
+          regs->end = re_malloc (regoff_t, regs->num_regs);
+          if (regs->start == NULL || regs->end == NULL)
+            {
+              re_free (pmatch);
+              return -2;
+            }
+          buffer->regs_allocated = REGS_REALLOCATE;
+        }
+      else if (buffer->regs_allocated == REGS_REALLOCATE)
+        { /* Yes.  If we need more elements than were already
+             allocated, reallocate them.  If we need fewer, just
+             leave it alone.  */
+          if (regs->num_regs < buffer->re_nsub + 1)
+            {
+              regs->num_regs = buffer->re_nsub + 1;
+              regs->start = re_realloc (regs->start, regoff_t, regs->num_regs);
+              regs->end = re_realloc (regs->end, regoff_t, regs->num_regs);
+              if (regs->start == NULL || regs->end == NULL)
+                {
+                  re_free (pmatch);
+                  return -2;
+                }
+            }
+        }
+      else
+        {
+          /* These braces fend off a "empty body in an else-statement"
+             warning under GCC when assert expands to nothing.  */
+          assert (buffer->regs_allocated == REGS_FIXED);
+        }
+    }
+
+  /* Restore registers.  */
+  if (regs != NULL)
+    {
+      for (i = 0; i <= nregs; ++i)
+        {
+          regs->start[i] = pmatch[i].rm_so;
+          regs->end[i] = pmatch[i].rm_eo;
+        }
+      for ( ; i < regs->num_regs; ++i)
+        {
+          regs->start[i] = -1;
+          regs->end[i] = -1;
+        }
+    }
+  /* Return value is -1 if not match, the length of mathing otherwise.  */
+  rval = (result) ? -1 : pmatch[0].rm_eo - pmatch[0].rm_so;
+  re_free (pmatch);
+  return rval;
+}
+#ifdef _LIBC
+weak_alias (__re_match, re_match)
+#endif
+
+/* re_match_2 matches the compiled pattern in BUFP against the
+   the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1
+   and SIZE2, respectively).  We start matching at POS, and stop
+   matching at STOP.
+
+   If REGS is non-null and the `no_sub' field of BUFP is nonzero, we
+   store offsets for the substring each group matched in REGS.  See the
+   documentation for exactly how many groups we fill.
+
+   We return -1 if no match, -2 if an internal error.
+   Otherwise, we return the length of the matched substring.  */
+
+int
+re_match_2 (buffer, string1, length1, string2, length2, start, regs, stop)
+     struct re_pattern_buffer *buffer;
+     const char *string1, *string2;
+     int length1, length2, start, stop;
+     struct re_registers *regs;
+{
+  int len, ret;
+  char *str = re_malloc (char, length1 + length2);
+  if (str == NULL)
+    return -2;
+  memcpy (str, string1, length1);
+  memcpy (str + length1, string2, length2);
+  len = (length1 + length2 < stop) ? length1 + length2 : stop;
+  ret = re_match (buffer, str, len, start, regs);
+  re_free (str);
+  return ret;
+}
+#ifdef _LIBC
+weak_alias (__re_match_2, re_match_2)
+#endif
+
+/* Like re_search_2, below, but only one string is specified, and
+   doesn't let you say where to stop matching.  */
+
+int
+re_search (bufp, string, size, startpos, range, regs)
+     struct re_pattern_buffer *bufp;
+     const char *string;
+     int size, startpos, range;
+     struct re_registers *regs;
+{
+  int i, nregs, result, real_range, rval, eflags = 0;
+  regmatch_t *pmatch;
+
+  eflags |= (bufp->not_bol) ? REG_NOTBOL : 0;
+  eflags |= (bufp->not_eol) ? REG_NOTEOL : 0;
+
+  /* Check for out-of-range.  */
+  if (startpos < 0 || startpos > size)
+    return -1;
+
+  /* We need at least 1 register.  */
+  nregs = ((regs == NULL) ? 1
+           : ((regs->num_regs > bufp->re_nsub) ? bufp->re_nsub + 1
+              : regs->num_regs + 1));
+  pmatch = re_malloc (regmatch_t, nregs);
+
+  /* Correct range if we need.  */
+  real_range = ((startpos + range > size) ? size - startpos
+                : ((startpos + range < 0) ? -startpos : range));
+
+  /* Compile fastmap if we haven't yet.  */
+  if (bufp->fastmap != NULL && !bufp->fastmap_accurate)
+    re_compile_fastmap (bufp);
+
+  result = re_search_internal (bufp, string, size, startpos, real_range,
+                               nregs, pmatch, eflags);
+
+  /* If caller wants register contents data back, do it.  */
+  if (regs && !bufp->no_sub)
+    {
+      /* Have the register data arrays been allocated?  */
+      if (bufp->regs_allocated == REGS_UNALLOCATED)
+        { /* No.  So allocate them with malloc.  We need one
+             extra element beyond `num_regs' for the `-1' marker
+             GNU code uses.  */
+          regs->num_regs = ((RE_NREGS > bufp->re_nsub + 1) ? RE_NREGS
+                            : bufp->re_nsub + 1);
+          regs->start = re_malloc (regoff_t, regs->num_regs);
+          regs->end = re_malloc (regoff_t, regs->num_regs);
+          if (regs->start == NULL || regs->end == NULL)
+            {
+              re_free (pmatch);
+              return -2;
+            }
+          bufp->regs_allocated = REGS_REALLOCATE;
+        }
+      else if (bufp->regs_allocated == REGS_REALLOCATE)
+        { /* Yes.  If we need more elements than were already
+             allocated, reallocate them.  If we need fewer, just
+             leave it alone.  */
+          if (regs->num_regs < bufp->re_nsub + 1)
+            {
+              regs->num_regs = bufp->re_nsub + 1;
+              regs->start = re_realloc (regs->start, regoff_t, regs->num_regs);
+              regs->end = re_realloc (regs->end, regoff_t, regs->num_regs);
+              if (regs->start == NULL || regs->end == NULL)
+                {
+                  re_free (pmatch);
+                  return -2;
+                }
+            }
+        }
+      else
+        {
+          /* These braces fend off a "empty body in an else-statement"
+             warning under GCC when assert expands to nothing.  */
+          assert (bufp->regs_allocated == REGS_FIXED);
+        }
+    }
+
+  /* Restore registers.  */
+  if (regs != NULL)
+    {
+      for (i = 0; i <= bufp->re_nsub; ++i)
+        {
+          regs->start[i] = pmatch[i].rm_so;
+          regs->end[i] = pmatch[i].rm_eo;
+        }
+      for ( ; i < regs->num_regs; ++i)
+        {
+          regs->start[i] = -1;
+          regs->end[i] = -1;
+        }
+    }
+  /* Return value is -1 if not match, the position where the mathing starts
+     otherwise.  */
+  rval = (result) ? -1 : pmatch[0].rm_so;
+  re_free (pmatch);
+  return rval;
+}
+#ifdef _LIBC
+weak_alias (__re_search, re_search)
+#endif
+
+/* Using the compiled pattern in BUFP, first tries to match the virtual
+   concatenation of STRING1 and STRING2, starting first at index
+   STARTPOS, then at STARTPOS + 1, and so on.
+
+   STRING1 and STRING2 have length SIZE1 and SIZE2, respectively.
+
+   RANGE is how far to scan while trying to match.  RANGE = 0 means try
+   only at STARTPOS; in general, the last start tried is STARTPOS +
+   RANGE.
+
+   In REGS, return the indices of the virtual concatenation of STRING1
+   and STRING2 that matched the entire BUFP->buffer and its contained
+   subexpressions.
+
+   Do not consider matching one past the index STOP in the virtual
+   concatenation of STRING1 and STRING2.
+
+   We return either the position in the strings at which the match was
+   found, -1 if no match, or -2 if error.  */
+
+int
+re_search_2 (bufp, string1, length1, string2, length2, start, range, regs,
+             stop)
+    struct re_pattern_buffer *bufp;
+    const char *string1, *string2;
+    int length1, length2, start, range, stop;
+    struct re_registers *regs;
+{
+  int len, ret;
+  char *str = re_malloc (char, length1 + length2);
+  memcpy (str, string1, length1);
+  memcpy (str + length1, string2, length2);
+  len = (length1 + length2 < stop) ? length1 + length2 : stop;
+  ret = re_search (bufp, str, len, start, range, regs);
+  re_free (str);
+  return ret;
+}
+#ifdef _LIBC
+weak_alias (__re_search_2, re_search_2)
+#endif
+
+/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
+   ENDS.  Subsequent matches using PATTERN_BUFFER and REGS will use
+   this memory for recording register information.  STARTS and ENDS
+   must be allocated using the malloc library routine, and must each
+   be at least NUM_REGS * sizeof (regoff_t) bytes long.
+
+   If NUM_REGS == 0, then subsequent matches should allocate their own
+   register data.
+
+   Unless this function is called, the first search or match using
+   PATTERN_BUFFER will allocate its own register data, without
+   freeing the old data.  */
+
+void
+re_set_registers (bufp, regs, num_regs, starts, ends)
+    struct re_pattern_buffer *bufp;
+    struct re_registers *regs;
+    unsigned num_regs;
+    regoff_t *starts, *ends;
+{
+  if (num_regs)
+    {
+      bufp->regs_allocated = REGS_REALLOCATE;
+      regs->num_regs = num_regs;
+      regs->start = starts;
+      regs->end = ends;
+    }
+  else
+    {
+      bufp->regs_allocated = REGS_UNALLOCATED;
+      regs->num_regs = 0;
+      regs->start = regs->end = (regoff_t *) 0;
+    }
+}
+#ifdef _LIBC
+weak_alias (__re_set_registers, re_set_registers)
+#endif
+
+/* Entry points compatible with 4.2 BSD regex library.  We don't define
+   them unless specifically requested.  */
+
+#if defined _REGEX_RE_COMP || defined _LIBC
+int
+# ifdef _LIBC
+weak_function
+# endif
+re_exec (s)
+     const char *s;
+{
+  return 0 == regexec (&re_comp_buf, s, 0, NULL, 0);
+}
+#endif /* _REGEX_RE_COMP */
+
+static re_node_set empty_set;
+
+/* Internal entry point.  */
+
+/* Searches for a compiled pattern PREG in the string STRING, whose
+   length is LENGTH.  NMATCH, PMATCH, and EFLAGS have the same
+   mingings with regexec.  START, and RANGE have the same meanings
+   with re_search.
+   Return 0 if we find a match and REG_NOMATCH if not.
+   Note: We assume front end functions already check ranges.
+   (START + RANGE >= 0 && START + RANGE <= LENGTH)  */
+
+static int
+re_search_internal (preg, string, length, start, range, nmatch, pmatch, eflags)
+    const regex_t *preg;
+    const char *string;
+    int length, start, range, eflags;
+    size_t nmatch;
+    regmatch_t pmatch[];
+{
+  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
+  re_string_t input;
+  re_dfastate_t **state_log;
+  int fl_longest_match, match_first, match_last = -1;
+  re_match_context_t mctx;
+  char *fastmap = ((preg->fastmap != NULL && preg->fastmap_accurate)
+                   ? preg->fastmap : NULL);
+
+  /* Check if the DFA haven't been compiled.  */
+  if (preg->used == 0 || dfa->init_state == NULL
+      || dfa->init_state_word == NULL || dfa->init_state_nl == NULL
+      || dfa->init_state_begbuf == NULL)
+    return 1;
+
+  re_node_set_init_empty (&empty_set);
+
+  /* We must check the longest matching, if nmatch > 0.  */
+  fl_longest_match = (nmatch != 0);
+
+  /* We will log all the DFA states through which the dfa pass,
+     if nmatch > 1, or this dfa has "multibyte node", which is a
+     back-reference or a node which can accept multibyte character or
+     multi character collating element.  */
+  if (nmatch > 1 || dfa->has_mb_node)
+    state_log = re_malloc (re_dfastate_t *, length + 1);
+  else
+    state_log = NULL;
+
+  if (preg->syntax & RE_ICASE)
+    re_string_construct_toupper (&input, string, length, preg->translate);
+  else
+    re_string_construct (&input, string, length, preg->translate);
+
+  match_ctx_init (&mctx, eflags, dfa->nbackref * 2);
+
+#ifdef DEBUG
+  /* We assume front-end functions already check them.  */
+  assert (start + range >= 0 && start + range <= length);
+#endif
+
+  /* Check incrementally whether of not the input string match.  */
+  for (match_first = start; ;)
+    {
+      if ((match_first < length
+           && (fastmap == NULL
+               || fastmap[re_string_byte_at (&input, match_first)]))
+          || preg->can_be_null)
+        {
+#ifdef RE_ENABLE_I18N
+          if (MB_CUR_MAX == 1 || re_string_first_byte (&input, match_first))
+#endif
+            {
+              /* We assume that the matching starts from `match_first'.  */
+              re_string_set_index (&input, match_first);
+              mctx.match_first = mctx.state_log_top = match_first;
+              mctx.nbkref_ents = mctx.max_bkref_len = 0;
+              match_last = check_matching (preg, &input, &mctx, state_log,
+                                           match_first, 0, fl_longest_match);
+              if (match_last != -1)
+                break;
+            }
+        }
+      /* Update counter.  */
+      if (range < 0)
+        {
+          --match_first;
+          if (match_first < start + range)
+            break;
+        }
+      else
+        {
+          ++match_first;
+          if (match_first > start + range)
+            break;
+        }
+    }
+
+  /* Set pmatch[] if we need.  */
+  if (match_last != -1 && nmatch > 0)
+    {
+      int reg_idx;
+
+      /* Initialize registers.  */
+      for (reg_idx = 0; reg_idx < nmatch; ++reg_idx)
+        pmatch[reg_idx].rm_so = pmatch[reg_idx].rm_eo = -1;
+
+      /* Set the points where matching start/end.  */
+      pmatch[0].rm_so = mctx.match_first;
+      mctx.match_last = pmatch[0].rm_eo = match_last;
+
+      if (!preg->no_sub && nmatch > 1)
+        {
+          /* We need the ranges of all the subexpressions.  */
+          int halt_node;
+          re_dfastate_t *pstate = state_log[match_last];
+#ifdef DEBUG
+          assert (state_log != NULL);
+#endif
+          halt_node = check_halt_state_context (preg, pstate, &input,
+                                                match_last, eflags);
+          sift_states_backward (preg, state_log, &mctx, &input, halt_node);
+          set_regs (preg, state_log, &mctx, &input, nmatch, pmatch, halt_node);
+        }
+    }
+
+  re_free (state_log);
+  if (dfa->nbackref)
+    match_ctx_free (&mctx);
+  re_string_destruct (&input);
+  return match_last == -1;
+}
+
+/* Acquire an initial state.
+   We must select appropriate initial state depending on the context,
+   since initial states may have constraints like "\<", "^", etc..  */
+
+static inline re_dfastate_t *
+acquire_init_state_context (preg, input, idx, eflags)
+    const regex_t *preg;
+    const re_string_t *input;
+    int idx, eflags;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+
+  if (dfa->init_state->has_constraint)
+    {
+      unsigned int context;
+      context =  re_string_context_at (input, idx - 1, eflags,
+                                       preg->newline_anchor);
+      if (IS_WORD_CONTEXT (context))
+        return dfa->init_state_word;
+      else if (IS_ORDINARY_CONTEXT (context))
+        return dfa->init_state;
+      else if (IS_BEGBUF_CONTEXT (context) && IS_NEWLINE_CONTEXT (context))
+        return dfa->init_state_begbuf;
+      else if (IS_NEWLINE_CONTEXT (context))
+        return dfa->init_state_nl;
+      else if (IS_BEGBUF_CONTEXT (context))
+        /* It is relatively rare case, then calculate on demand.  */
+        return re_acquire_state_context (dfa, dfa->init_state->entrance_nodes,
+                                         context);
+      else
+        /* Must not happen?  */
+        return dfa->init_state;
+    }
+  else
+    return dfa->init_state;
+}
+
+/* Check whether the regular expression match input string INPUT or not,
+   and return the index where the matching end, or return -1 if not match.
+   FL_SEARCH means we must search where the matching starts,
+   FL_LONGEST_MATCH means we want the POSIX longest matching.  */
+
+static int
+check_matching (preg, input, mctx, state_log, start_idx, fl_search,
+                fl_longest_match)
+    const regex_t *preg;
+    re_string_t *input;
+    re_match_context_t *mctx;
+    re_dfastate_t **state_log;
+    int start_idx, fl_search, fl_longest_match;
+{
+  int match = 0, match_last = -1;
+  re_dfastate_t *cur_state;
+
+  cur_state = acquire_init_state_context (preg, input, start_idx,
+                                          mctx->eflags);
+  if (state_log != NULL)
+    state_log[start_idx] = cur_state;
+  /* If the RE accepts NULL string.  */
+  if (cur_state->halt)
+    {
+      if (!cur_state->has_constraint
+          || check_halt_state_context (preg, cur_state, input, start_idx,
+                                       mctx->eflags))
+        {
+          if (!fl_longest_match)
+            return start_idx;
+          else
+            {
+              match_last = start_idx;
+              match = 1;
+            }
+        }
+    }
+
+  while (!re_string_eoi (input))
+    {
+      cur_state = transit_state (preg, cur_state, input, fl_search && !match,
+                                 state_log, mctx);
+      if (cur_state == NULL) /* Reached at the invalid state.  */
+        {
+          int cur_str_idx = re_string_cur_idx (input);
+          if (fl_search && !match)
+            {
+              /* Restart from initial state, since we are searching
+                 the point from where matching start.  */
+#ifdef RE_ENABLE_I18N
+              if (MB_CUR_MAX == 1 || re_string_first_byte (input, cur_str_idx))
+#endif /* RE_ENABLE_I18N */
+                cur_state = acquire_init_state_context (preg, input,
+                                                        cur_str_idx,
+                                                        mctx->eflags);
+              if (state_log != NULL)
+                state_log[cur_str_idx] = cur_state;
+            }
+          else if (!fl_longest_match && match)
+            break;
+          else /* (fl_longest_match && match) || (!fl_search && !match)  */
+            {
+              if (state_log == NULL)
+                break;
+              else
+                {
+                  int max = mctx->state_log_top;
+                  for (; cur_str_idx <= max; ++cur_str_idx)
+                    if (state_log[cur_str_idx] != NULL)
+                      break;
+                  if (cur_str_idx > max)
+                    break;
+                }
+            }
+        }
+
+      if (cur_state != NULL && cur_state->halt)
+        {
+          /* Reached at a halt state.
+             Check the halt state can satisfy the current context.  */
+          if (!cur_state->has_constraint
+              || check_halt_state_context (preg, cur_state, input,
+                                           re_string_cur_idx (input),
+                                           mctx->eflags))
+            {
+              /* We found an appropriate halt state.  */
+              match_last = re_string_cur_idx (input);
+              match = 1;
+              if (!fl_longest_match)
+                break;
+            }
+        }
+   }
+  return match_last;
+}
+
+/* Check NODE match the current context.  */
+
+static int check_halt_node_context (dfa, node, context)
+    const re_dfa_t *dfa;
+    int node;
+    unsigned int context;
+{
+  int entity;
+  re_token_type_t type = dfa->nodes[node].type;
+  if (type == END_OF_RE)
+    return 1;
+  if (type != OP_CONTEXT_NODE)
+    return 0;
+  entity = dfa->nodes[node].opr.ctx_info->entity;
+  if (dfa->nodes[entity].type != END_OF_RE
+      || NOT_SATISFY_NEXT_CONSTRAINT (dfa->nodes[node].constraint, context))
+    return 0;
+  return 1;
+}
+
+/* Check the halt state STATE match the current context.
+   Return 0 if not match, if the node, STATE has, is a halt node and
+   match the context, return the node.  */
+
+static int
+check_halt_state_context (preg, state, input, idx, eflags)
+    const regex_t *preg;
+    const re_dfastate_t *state;
+    const re_string_t *input;
+    int idx, eflags;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int i;
+  unsigned int context;
+#ifdef DEBUG
+  assert (state->halt);
+#endif
+  context = re_string_context_at (input, idx, eflags, preg->newline_anchor);
+  for (i = 0; i < state->nodes.nelem; ++i)
+    if (check_halt_node_context (dfa, state->nodes.elems[i], context))
+      return state->nodes.elems[i];
+  return 0;
+}
+
+/* Compute the next node to which "NFA" transit from NODE.
+   Return the destination node, and update EPS_VIA_NODES.
+   ("NFA" is a NFA corresponding to the DFA.  */
+
+static int
+proceed_next_node (preg, state_log, mctx, input, pidx, node, eps_via_nodes)
+    const regex_t *preg;
+    re_dfastate_t **state_log;
+    const re_match_context_t *mctx;
+    const re_string_t *input;
+    int *pidx, node;
+    re_node_set *eps_via_nodes;
+{
+  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
+  int i, dest_node = -1;
+  if (IS_EPSILON_NODE (dfa->nodes[node].type))
+    {
+      re_node_set_insert (eps_via_nodes, node);
+      for (i = 0; i < state_log[*pidx]->nodes.nelem; ++i)
+        {
+          int candidate = state_log[*pidx]->nodes.elems[i];
+          if (!re_node_set_contains (dfa->edests + node, candidate)
+              && !(dfa->nodes[candidate].type == OP_CONTEXT_NODE
+                   && re_node_set_contains (dfa->edests + node,
+                            dfa->nodes[candidate].opr.ctx_info->entity)))
+            continue;
+          dest_node = candidate;
+          /* In order to avoid infinite loop like "(a*)*".  */
+          if (!re_node_set_contains (eps_via_nodes, dest_node))
+            break;
+        }
+#ifdef DEBUG
+      assert (dest_node != -1);
+#endif
+      return dest_node;
+    }
+  else
+    {
+      int naccepted = 0, entity = node;
+      re_token_type_t type = dfa->nodes[node].type;
+      if (type == OP_CONTEXT_NODE)
+        {
+          entity = dfa->nodes[node].opr.ctx_info->entity;
+          type = dfa->nodes[entity].type;
+        }
+
+      if (ACCEPT_MB_NODE (type))
+        naccepted = check_node_accept_bytes (preg, entity, input, *pidx);
+      else if (type == OP_BACK_REF)
+        {
+          for (i = 0; i < mctx->nbkref_ents; ++i)
+            {
+              if (mctx->bkref_ents[i].node == node
+                  && mctx->bkref_ents[i].from == *pidx)
+                naccepted = mctx->bkref_ents[i].to - *pidx;
+            }
+          if (naccepted == 0)
+            {
+              re_node_set_insert (eps_via_nodes, node);
+              dest_node = dfa->nexts[node];
+              if (re_node_set_contains (&state_log[*pidx]->nodes, dest_node))
+                return dest_node;
+              for (i = 0; i < state_log[*pidx]->nodes.nelem; ++i)
+                {
+                  dest_node = state_log[*pidx]->nodes.elems[i];
+                  if ((dfa->nodes[dest_node].type == OP_CONTEXT_NODE
+                       && (dfa->nexts[node]
+                           == dfa->nodes[dest_node].opr.ctx_info->entity)))
+                    return dest_node;
+                }
+            }
+        }
+
+      if (naccepted != 0
+          || check_node_accept (preg, dfa->nodes + node, input, *pidx,
+                                mctx->eflags))
+        {
+          dest_node = dfa->nexts[node];
+          *pidx = (naccepted == 0) ? *pidx + 1 : *pidx + naccepted;
+#ifdef DEBUG
+          assert (state_log[*pidx] != NULL);
+#endif
+          re_node_set_empty (eps_via_nodes);
+          return dest_node;
+        }
+    }
+  /* Must not reach here.  */
+#ifdef DEBUG
+  assert (0);
+#endif
+  return 0;
+}
+
+/* Set the positions where the subexpressions are starts/ends to registers
+   PMATCH.
+   Note: We assume that pmatch[0] is already set, and
+   pmatch[i].rm_so == pmatch[i].rm_eo == -1 (i > 1).  */
+
+static void
+set_regs (preg, state_log, mctx, input, nmatch, pmatch, last_node)
+    const regex_t *preg;
+    re_dfastate_t **state_log;
+    const re_match_context_t *mctx;
+    const re_string_t *input;
+    size_t nmatch;
+    regmatch_t *pmatch;
+    int last_node;
+{
+  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
+  int idx, cur_node, node_entity, real_nmatch;
+  re_node_set eps_via_nodes;
+  int i;
+#ifdef DEBUG
+  assert (nmatch > 1);
+  assert (state_log != NULL);
+#endif
+  cur_node = dfa->init_node;
+  real_nmatch = (nmatch <= preg->re_nsub) ? nmatch : preg->re_nsub + 1;
+  re_node_set_init_empty (&eps_via_nodes);
+  for (idx = pmatch[0].rm_so; idx <= pmatch[0].rm_eo ;)
+    {
+      node_entity = ((dfa->nodes[cur_node].type == OP_CONTEXT_NODE)
+                     ? dfa->nodes[cur_node].opr.ctx_info->entity : cur_node);
+      for (i = 1; i < real_nmatch; ++i)
+        {
+          if (dfa->subexps[i - 1].start == dfa->subexps[i - 1].end)
+            {
+              /* In case of the null subexpression like '()'.  */
+              if (dfa->subexps[i - 1].start == node_entity)
+                {
+                  pmatch[i].rm_so = idx;
+                  pmatch[i].rm_eo = idx;
+                }
+            }
+          else if (dfa->subexps[i - 1].start <= node_entity
+                   && node_entity < dfa->subexps[i - 1].end)
+            {
+              if (pmatch[i].rm_so == -1 || pmatch[i].rm_eo != -1)
+                /* We are at the first node of this sub expression.  */
+                {
+                  pmatch[i].rm_so = idx;
+                  pmatch[i].rm_eo = -1;
+                }
+            }
+          else
+            {
+              if (pmatch[i].rm_so != -1 && pmatch[i].rm_eo == -1)
+                /* We are at the last node of this sub expression.  */
+                pmatch[i].rm_eo = idx;
+            }
+        }
+      if (idx == pmatch[0].rm_eo && cur_node == last_node)
+        break;
+
+      /* Proceed to next node.  */
+      cur_node = proceed_next_node (preg, state_log, mctx, input, &idx,
+                                    cur_node, &eps_via_nodes);
+    }
+  re_node_set_free (&eps_via_nodes);
+  return;
+}
+
+#define NUMBER_OF_STATE 1
+
+/* This function checks the STATE_LOG from the MCTX->match_last
+   to MCTX->match_first and sift the nodes in each states according to
+   the following rules.  Updated state_log will be wrote to STATE_LOG.
+
+   Rules: We throw away the Node `a' in the STATE_LOG[STR_IDX] if...
+     1. When STR_IDX == MATCH_LAST(the last index in the state_log):
+        If `a' isn't the LAST_NODE and `a' can't epsilon transit to
+        the LAST_NODE, we throw away the node `a'.
+     2. When MATCH_FIRST <= STR_IDX < MATCH_LAST and `a' accepts
+        string `s' and transit to `b':
+        i. If 'b' isn't in the STATE_LOG[STR_IDX+strlen('s')], we throw
+           away the node `a'.
+        ii. If 'b' is in the STATE_LOG[STR_IDX+strlen('s')] but 'b' is
+            throwed away, we throw away the node `a'.
+     3. When 0 <= STR_IDX < n and 'a' epsilon transit to 'b':
+        i. If 'b' isn't in the STATE_LOG[STR_IDX], we throw away the
+           node `a'.
+        ii. If 'b' is in the STATE_LOG[STR_IDX] but 'b' is throwed away,
+            we throw away the node `a'.  */
+
+#define STATE_NODE_CONTAINS(state,node) \
+  ((state) != NULL && re_node_set_contains (&(state)->nodes, node))
+
+static void
+sift_states_backward (preg, state_log, mctx, input, last_node)
+    const regex_t *preg;
+    re_dfastate_t **state_log;
+    const re_match_context_t *mctx;
+    const re_string_t *input;
+    int last_node;
+{
+  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
+  re_node_set state_buf;
+  int str_idx = mctx->match_last;
+  re_node_set *plog;	/* Points the state_log[str_idx]->nodes  */
+
+#ifdef DEBUG
+  assert (state_log != NULL && state_log[str_idx] != NULL);
+#endif
+  re_node_set_alloc (&state_buf, NUMBER_OF_STATE);
+  plog = &state_log[str_idx]->nodes;
+
+  /* Build sifted state_log[str_idx].  It has the nodes which can epsilon
+     transit to the last_node and the last_node itself.  */
+  re_node_set_intersect (&state_buf, plog, dfa->inveclosures + last_node);
+
+  if (state_log[str_idx] != NULL && state_log[str_idx]->has_backref)
+    add_epsilon_backreference (dfa, mctx, plog, str_idx, &state_buf);
+
+  /* Update state log.  */
+  state_log[str_idx] = re_acquire_state (dfa, &state_buf);
+
+  /* Then check each states in the state_log.  */
+  while (str_idx > mctx->match_first)
+    {
+      int i, j;
+      /* Update counters.  */
+      re_node_set_empty (&state_buf);
+      --str_idx;
+      plog = ((state_log[str_idx] == NULL) ? &empty_set
+              : &state_log[str_idx]->nodes);
+
+      /* Then build the next sifted state.
+         We build the next sifted state on `state_buf', and update
+         `state_log[str_idx]' with `state_buf'.
+         Note:
+         `state_buf' is the sifted state from `state_log[str_idx + 1]'.
+         `plog' points the node_set of the old `state_log[str_idx]'.  */
+      for (i = 0; i < plog->nelem; i++)
+        {
+          int prev_node = plog->elems[i];
+          int entity = prev_node;
+          int naccepted = 0;
+          re_token_type_t type = dfa->nodes[prev_node].type;
+          if (type == OP_CONTEXT_NODE)
+            {
+              entity = dfa->nodes[prev_node].opr.ctx_info->entity;
+              type = dfa->nodes[entity].type;
+            }
+
+          /* If the node may accept `multi byte'.  */
+          if (ACCEPT_MB_NODE (type))
+            naccepted = sift_states_iter_mb (preg, state_log, mctx, input,
+                                             entity, str_idx,
+                                             mctx->match_last);
+
+          /* If the node is a back reference.  */
+          else if (type == OP_BACK_REF)
+            for (j = 0; j < mctx->nbkref_ents; ++j)
+              {
+                naccepted = sift_states_iter_bkref (dfa, state_log,
+                                                    mctx->bkref_ents + j,
+                                                    prev_node, str_idx,
+                                                    mctx->match_first,
+                                                    mctx->match_last);
+                if (naccepted)
+                  break;
+              }
+
+          if (!naccepted
+              && check_node_accept (preg, dfa->nodes + prev_node, input,
+                                    str_idx, mctx->eflags)
+              && STATE_NODE_CONTAINS (state_log[str_idx + 1],
+                                      dfa->nexts[prev_node]))
+            naccepted = 1;
+
+          if (naccepted == 0)
+            continue;
+
+          /* `prev_node' may point the entity of the OP_CONTEXT_NODE,
+             then we use plog->elems[i] instead.  */
+          re_node_set_add_intersect (&state_buf, plog,
+                                     dfa->inveclosures + prev_node);
+        }
+      if (state_log[str_idx] != NULL && state_log[str_idx]->has_backref)
+        add_epsilon_backreference (dfa, mctx, plog, str_idx, &state_buf);
+
+      /* Update state_log.  */
+      state_log[str_idx] = re_acquire_state (dfa, &state_buf);
+    }
+
+  re_node_set_free (&state_buf);
+}
+
+/* Helper functions.  */
+
+static inline void
+clean_state_log_if_need (state_log, mctx, next_state_log_idx)
+    re_dfastate_t **state_log;
+    re_match_context_t *mctx;
+    int next_state_log_idx;
+{
+  int top = mctx->state_log_top;
+  if (top < next_state_log_idx)
+    {
+      memset (state_log + top + 1, '\0',
+              sizeof (re_dfastate_t *) * (next_state_log_idx - top));
+      mctx->state_log_top = next_state_log_idx;
+    }
+}
+
+static int
+sift_states_iter_mb (preg, state_log, mctx, input, node_idx, str_idx,
+                     max_str_idx)
+    const regex_t *preg;
+    re_dfastate_t **state_log;
+    const re_match_context_t *mctx;
+    const re_string_t *input;
+    int node_idx, str_idx, max_str_idx;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int naccepted;
+  /* Check the node can accept `multi byte'.  */
+  naccepted = check_node_accept_bytes (preg, node_idx, input, str_idx);
+  if (naccepted > 0 && str_idx + naccepted <= max_str_idx &&
+      !STATE_NODE_CONTAINS (state_log[str_idx + naccepted],
+                            dfa->nexts[node_idx]))
+    /* The node can't accept the `multi byte', or the
+       destination was already throwed away, then the node
+       could't accept the current input `multi byte'.   */
+    naccepted = 0;
+  /* Otherwise, it is sure that the node could accept
+     `naccepted' bytes input.  */
+  return naccepted;
+}
+
+static int
+sift_states_iter_bkref (dfa, state_log, mctx_entry, node_idx, idx, match_first,
+                        match_last)
+    const re_dfa_t *dfa;
+    re_dfastate_t **state_log;
+    struct re_backref_cache_entry *mctx_entry;
+    int node_idx, idx, match_first, match_last;
+{
+  int naccepted = 0;
+  int from_idx, to_idx;
+  from_idx = mctx_entry->from;
+  to_idx = mctx_entry->to;
+  if (mctx_entry->node == node_idx
+      && from_idx == idx && to_idx <= match_last
+      && STATE_NODE_CONTAINS (state_log[to_idx], dfa->nexts[node_idx]))
+    naccepted = to_idx - from_idx;
+  return naccepted;
+}
+
+static void
+add_epsilon_backreference (dfa, mctx, plog, idx, state_buf)
+    const re_dfa_t *dfa;
+    const re_match_context_t *mctx;
+    const re_node_set *plog;
+    int idx;
+    re_node_set *state_buf;
+{
+  int i, j;
+  for (i = 0; i < plog->nelem; ++i)
+    {
+      int node_idx = plog->elems[i];
+      re_token_type_t type = dfa->nodes[node_idx].type;
+      if (type == OP_CONTEXT_NODE)
+        type = dfa->nodes[dfa->nodes[node_idx].opr.ctx_info->entity].type;
+
+      if (type == OP_BACK_REF &&
+          !re_node_set_contains (state_buf, node_idx))
+        {
+          for (j = 0; j < mctx->nbkref_ents; ++j)
+            {
+              struct re_backref_cache_entry *entry;
+              entry = mctx->bkref_ents + j;
+              if (entry->from == entry->to && entry->from == idx)
+                break;
+            }
+          if (j < mctx->nbkref_ents || idx == mctx->match_first)
+            {
+              re_node_set_add_intersect (state_buf, plog,
+                                         dfa->inveclosures + node_idx);
+              i = 0;
+            }
+        }
+    }
+}
+
+/* Functions for state transition.  */
+
+/* Return the next state to which the current state STATE will transit by
+   accepting the current input byte, and update STATE_LOG if necessary.
+   If STATE can accept a multibyte char/collating element/back reference
+   update the destination of STATE_LOG.  */
+
+static re_dfastate_t *
+transit_state (preg, state, input, fl_search, state_log, mctx)
+    const regex_t *preg;
+    re_dfastate_t *state, **state_log;
+    re_string_t *input;
+    int fl_search;
+    re_match_context_t *mctx;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  re_dfastate_t **trtable, *next_state;
+  unsigned char ch;
+
+  if (state == NULL)
+    {
+      next_state = state;
+      re_string_skip_bytes (input, 1);
+    }
+  else
+    {
+      /* If the current state can accept multibyte.  */
+      if (state->accept_mb)
+        transit_state_mb (preg, state, input, state_log, mctx);
+
+      /* Then decide the next state with the single byte.  */
+      if (1)
+        {
+          /* Use transition table  */
+          ch = re_string_fetch_byte (input);
+          trtable = fl_search ? state->trtable_search : state->trtable;
+          if (trtable == NULL)
+            {
+              trtable = build_trtable (preg, state, fl_search);
+              if (fl_search)
+                state->trtable_search = trtable;
+              else
+                state->trtable = trtable;
+            }
+          next_state = trtable[ch];
+        }
+      else
+        {
+          /* don't use transition table  */
+          next_state = transit_state_sb (preg, state, input, fl_search, mctx);
+        }
+    }
+
+  /* Update the state_log if we need.  */
+  if (state_log != NULL)
+    {
+      int cur_idx = re_string_cur_idx (input);
+      if (cur_idx > mctx->state_log_top)
+        {
+          state_log[cur_idx] = next_state;
+          mctx->state_log_top = cur_idx;
+        }
+      else if (state_log[cur_idx] == 0)
+        {
+          state_log[cur_idx] = next_state;
+        }
+      else
+        {
+          re_dfastate_t *pstate;
+          unsigned int context;
+          re_node_set next_nodes, *log_nodes, *table_nodes = NULL;
+          /* If (state_log[cur_idx] != 0), it implies that cur_idx is
+             the destination of a multibyte char/collating element/
+             back reference.  Then the next state is the union set of
+             these destinations and the results of the transition table.  */
+          pstate = state_log[cur_idx];
+          log_nodes = pstate->entrance_nodes;
+          if (next_state != NULL)
+            {
+              table_nodes = next_state->entrance_nodes;
+              re_node_set_init_union (&next_nodes, table_nodes, log_nodes);
+            }
+          else
+            next_nodes = *log_nodes;
+          /* Note: We already add the nodes of the initial state,
+                   then we don't need to add them here.  */
+
+          context = re_string_context_at (input, re_string_cur_idx (input) - 1,
+                                          mctx->eflags, preg->newline_anchor);
+          next_state = state_log[cur_idx]
+              = re_acquire_state_context (dfa, &next_nodes, context);
+          if (table_nodes != NULL)
+            re_node_set_free (&next_nodes);
+        }
+      /* If the next state has back references.  */
+      if (next_state != NULL && next_state->has_backref)
+        {
+          transit_state_bkref (preg, next_state, input, state_log, mctx);
+          next_state = state_log[cur_idx];
+        }
+    }
+  return next_state;
+}
+
+/* Helper functions for transit_state.  */
+
+/* Return the next state to which the current state STATE will transit by
+   accepting the current input byte.  */
+
+static re_dfastate_t *
+transit_state_sb (preg, state, input, fl_search, mctx)
+    const regex_t *preg;
+    re_dfastate_t *state;
+    re_string_t *input;
+    int fl_search;
+    re_match_context_t *mctx;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  re_node_set next_nodes;
+  re_dfastate_t *next_state;
+  int node_cnt, cur_str_idx = re_string_cur_idx (input);
+  unsigned int context;
+
+  re_node_set_alloc (&next_nodes, state->nodes.nelem + 1);
+  for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt)
+    {
+      int cur_node = state->nodes.elems[node_cnt];
+      if (check_node_accept (preg, dfa->nodes + cur_node, input,
+                             cur_str_idx, mctx->eflags))
+        re_node_set_merge (&next_nodes,
+                           dfa->eclosures + dfa->nexts[cur_node]);
+    }
+  if (fl_search)
+    {
+#ifdef RE_ENABLE_I18N
+      int not_initial = 0;
+      if (MB_CUR_MAX > 1)
+        for (node_cnt = 0; node_cnt < next_nodes.nelem; ++node_cnt)
+          if (dfa->nodes[next_nodes.elems[node_cnt]].type == CHARACTER)
+            {
+              not_initial = dfa->nodes[next_nodes.elems[node_cnt]].mb_partial;
+              break;
+            }
+      if (!not_initial)
+#endif
+        re_node_set_merge (&next_nodes, dfa->init_state->entrance_nodes);
+    }
+  context = re_string_context_at (input, cur_str_idx, mctx->eflags,
+                                  preg->newline_anchor);
+  next_state = re_acquire_state_context (dfa, &next_nodes, context);
+  re_node_set_free (&next_nodes);
+  re_string_skip_bytes (input, 1);
+  return next_state;
+}
+
+static void
+transit_state_mb (preg, pstate, input, state_log, mctx)
+    const regex_t *preg;
+    re_dfastate_t *pstate, **state_log;
+    const re_string_t *input;
+    re_match_context_t *mctx;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int i;
+
+  for (i = 0; i < pstate->nodes.nelem; ++i)
+    {
+      re_node_set dest_nodes, *new_nodes;
+      int cur_node_idx = pstate->nodes.elems[i];
+      int naccepted = 0, dest_idx;
+      unsigned int context;
+      re_dfastate_t *dest_state;
+
+      if (dfa->nodes[cur_node_idx].type == OP_CONTEXT_NODE)
+        {
+          context = re_string_context_at (input, re_string_cur_idx (input),
+                                          mctx->eflags, preg->newline_anchor);
+          if (NOT_SATISFY_NEXT_CONSTRAINT (dfa->nodes[cur_node_idx].constraint,
+                                        context))
+            continue;
+          cur_node_idx = dfa->nodes[cur_node_idx].opr.ctx_info->entity;
+        }
+
+      /* How many bytes the node can accepts?  */
+      if (ACCEPT_MB_NODE (dfa->nodes[cur_node_idx].type))
+        naccepted = check_node_accept_bytes (preg, cur_node_idx, input,
+                                             re_string_cur_idx (input));
+      if (naccepted == 0)
+        continue;
+
+      /* The node can accepts `naccepted' bytes.  */
+      dest_idx = re_string_cur_idx (input) + naccepted;
+      clean_state_log_if_need (state_log, mctx, dest_idx);
+#ifdef DEBUG
+      assert (dfa->nexts[cur_node_idx] != -1);
+#endif
+      /* `cur_node_idx' may point the entity of the OP_CONTEXT_NODE,
+         then we use pstate->nodes.elems[i] instead.  */
+      new_nodes = dfa->eclosures + dfa->nexts[pstate->nodes.elems[i]];
+
+      dest_state = state_log[dest_idx];
+      if (dest_state == NULL)
+        dest_nodes = *new_nodes;
+      else
+        re_node_set_init_union (&dest_nodes, dest_state->entrance_nodes,
+                                new_nodes);
+      context = re_string_context_at (input, dest_idx - 1, mctx->eflags,
+                                      preg->newline_anchor);
+      state_log[dest_idx] = re_acquire_state_context (dfa, &dest_nodes, context);
+      if (dest_state != NULL)
+        re_node_set_free (&dest_nodes);
+    }
+}
+
+static void
+transit_state_bkref (preg, pstate, input, state_log, mctx)
+    const regex_t *preg;
+    re_dfastate_t *pstate, **state_log;
+    const re_string_t *input;
+    re_match_context_t *mctx;
+{
+  re_dfastate_t **work_state_log;
+
+#ifdef DEBUG
+  assert (mctx->match_first != -1);
+#endif
+  work_state_log = re_malloc (re_dfastate_t *, re_string_cur_idx (input) + 1);
+
+  transit_state_bkref_loop (preg, input, &pstate->nodes, work_state_log,
+                            state_log, mctx);
+
+  re_free (work_state_log);
+}
+
+/* Caller must allocate `work_state_log'.  */
+
+static void
+transit_state_bkref_loop (preg, input, nodes, work_state_log, state_log, mctx)
+    const regex_t *preg;
+    const re_string_t *input;
+    re_node_set *nodes;
+    re_dfastate_t **work_state_log, **state_log;
+    re_match_context_t *mctx;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int i, j;
+  regmatch_t *cur_regs = re_malloc (regmatch_t, preg->re_nsub + 1);
+  int cur_str_idx = re_string_cur_idx (input);
+
+  for (i = 0; i < nodes->nelem; ++i)
+    {
+      int dest_str_idx, subexp_idx, prev_nelem, subexp_len;
+      int node_idx = nodes->elems[i];
+      unsigned int context;
+      re_token_t *node = dfa->nodes + node_idx;
+      re_dfastate_t *dest_state;
+      re_node_set *new_dest_nodes;
+
+      /* Check whether `node' is a backreference or not.  */
+      if (node->type == OP_BACK_REF)
+        subexp_idx = node->opr.idx;
+      else if (node->type == OP_CONTEXT_NODE &&
+               dfa->nodes[node->opr.ctx_info->entity].type == OP_BACK_REF)
+        {
+          context = re_string_context_at (input, cur_str_idx, mctx->eflags,
+                                          preg->newline_anchor);
+          if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context))
+            continue;
+          subexp_idx = dfa->nodes[node->opr.ctx_info->entity].opr.idx;
+        }
+      else
+        continue;
+
+      /* `node' is a backreference.
+         At first, set registers to check the backreference. */
+      cur_regs[0].rm_so = mctx->match_first;
+      cur_regs[0].rm_eo = cur_str_idx;
+      memcpy (work_state_log + mctx->match_first,
+              state_log + mctx->match_first,
+              sizeof (re_dfastate_t *)
+	      * (cur_str_idx - mctx->match_first + 1));
+      mctx->match_last = cur_str_idx;
+      sift_states_backward (preg, work_state_log, mctx, input, node_idx);
+      if (!STATE_NODE_CONTAINS (work_state_log[mctx->match_first],
+                                dfa->init_node))
+        continue;
+      for (j = 1; j <= preg->re_nsub; ++j)
+        cur_regs[j].rm_so = cur_regs[j].rm_eo = -1;
+      set_regs (preg, work_state_log, mctx, input,
+                subexp_idx + 1, cur_regs, node_idx);
+
+      /* Then check that the backreference can match the input string.  */
+      subexp_len = cur_regs[subexp_idx].rm_eo - cur_regs[subexp_idx].rm_so;
+      if (subexp_len < 0
+          || (strncmp ((re_string_get_buffer (input)
+                        + cur_regs[subexp_idx].rm_so),
+                       re_string_get_buffer (input) + cur_str_idx, subexp_len)
+              != 0))
+        continue;
+
+      /* Successfully matched, add a new cache entry.  */
+      dest_str_idx = cur_str_idx + subexp_len;
+      match_ctx_add_entry (mctx, node_idx, cur_str_idx, dest_str_idx);
+      clean_state_log_if_need (state_log, mctx, dest_str_idx);
+
+      /* And add the epsilon closures (which is `new_dest_nodes') of
+         the backreference to appropriate state_log.  */
+#ifdef DEBUG
+      assert (dfa->nexts[node_idx] != -1);
+#endif
+      if (node->type == OP_CONTEXT_NODE && subexp_len == 0)
+        new_dest_nodes = dfa->nodes[node_idx].opr.ctx_info->bkref_eclosure;
+      else
+        new_dest_nodes = dfa->eclosures + dfa->nexts[node_idx];
+      context = (IS_WORD_CHAR (re_string_byte_at (input, dest_str_idx - 1))
+                 ? CONTEXT_WORD : 0);
+      dest_state = state_log[dest_str_idx];
+
+      prev_nelem = ((state_log[cur_str_idx] == NULL) ? 0
+                    : state_log[cur_str_idx]->nodes.nelem);
+      /* Add `new_dest_node' to state_log.  */
+      if (dest_state == NULL)
+        state_log[dest_str_idx] = re_acquire_state_context (dfa,
+                                                            new_dest_nodes,
+                                                            context);
+      else
+        {
+          re_node_set dest_nodes;
+          re_node_set_init_union (&dest_nodes, dest_state->entrance_nodes,
+                                  new_dest_nodes);
+          state_log[dest_str_idx] = re_acquire_state_context (dfa, &dest_nodes,
+                                                              context);
+          re_node_set_free (&dest_nodes);
+        }
+
+      /* We need to check recursively if the backreference can epsilon
+         transit.  */
+      if (subexp_len == 0 && state_log[cur_str_idx]->nodes.nelem > prev_nelem)
+        transit_state_bkref_loop (preg, input, new_dest_nodes, work_state_log,
+                                  state_log, mctx);
+    }
+  re_free (cur_regs);
+}
+
+/* Build transition table for the state.  */
+
+static re_dfastate_t **
+build_trtable (preg, state, fl_search)
+    const regex_t *preg;
+    const re_dfastate_t *state;
+    int fl_search;
+{
+  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int i, j, k, ch;
+  int ndests; /* Number of the destination states from `state'.  */
+  re_dfastate_t **trtable, **dest_states, **dest_states_word, **dest_states_nl;
+  re_node_set follows, *dests_node;
+  bitset *dests_ch;
+  bitset acceptable;
+
+  /* We build DFA states which corresponds to the destination nodes
+     from `state'.  `dests_node[i]' represents the nodes which i-th
+     destination state contains, and `dests_ch[i]' represents the
+     characters which i-th destination state accepts.  */
+  dests_node = re_malloc (re_node_set, SBC_MAX);
+  dests_ch = re_malloc (bitset, SBC_MAX);
+
+  /* Initialize transiton table.  */
+  trtable = (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX);
+
+  /* At first, group all nodes belonging to `state' into several
+     destinations.  */
+  ndests = group_nodes_into_DFAstates (preg, state, dests_node, dests_ch);
+  if (ndests == 0)
+    {
+      re_free (dests_node);
+      re_free (dests_ch);
+      return trtable;
+    }
+
+  dest_states = re_malloc (re_dfastate_t *, ndests);
+  dest_states_word = re_malloc (re_dfastate_t *, ndests);
+  dest_states_nl = re_malloc (re_dfastate_t *, ndests);
+  bitset_empty (acceptable);
+
+  re_node_set_alloc (&follows, ndests + 1);
+  /* Then build the states for all destinations.  */
+  for (i = 0; i < ndests; ++i)
+    {
+      int next_node;
+      re_node_set_empty (&follows);
+      /* Merge the follows of this destination states.  */
+      for (j = 0; j < dests_node[i].nelem; ++j)
+        {
+          next_node = dfa->nexts[dests_node[i].elems[j]];
+          if (next_node != -1)
+            {
+              re_node_set_merge (&follows, dfa->eclosures + next_node);
+            }
+        }
+      /* If search flag is set, merge the initial state.  */
+      if (fl_search)
+        {
+#ifdef RE_ENABLE_I18N
+          int not_initial = 0;
+          for (j = 0; j < follows.nelem; ++j)
+            if (dfa->nodes[follows.elems[j]].type == CHARACTER)
+              {
+                not_initial = dfa->nodes[follows.elems[j]].mb_partial;
+                break;
+              }
+          if (!not_initial)
+#endif
+            re_node_set_merge (&follows, dfa->init_state->entrance_nodes);
+        }
+      dest_states[i] = re_acquire_state_context (dfa, &follows, 0);
+      /* If the new state has context constraint,
+         build appropriate states for these contexts.  */
+      if (dest_states[i]->has_constraint)
+        {
+          dest_states_word[i] = re_acquire_state_context (dfa, &follows,
+                                                          CONTEXT_WORD);
+          dest_states_nl[i] = re_acquire_state_context (dfa, &follows,
+                                                        CONTEXT_NEWLINE);
+        }
+      else
+        {
+          dest_states_word[i] = dest_states[i];
+          dest_states_nl[i] = dest_states[i];
+        }
+      bitset_merge (acceptable, dests_ch[i]);
+    }
+
+  /* Update the transition table.  */
+  for (i = 0, ch = 0; i < BITSET_UINTS; ++i)
+    for (j = 0; j < UINT_BITS; ++j, ++ch)
+      if ((acceptable[i] >> j) & 1)
+        {
+          if (IS_WORD_CHAR (ch))
+            {
+              for (k = 0; k < ndests; ++k)
+                if ((dests_ch[k][i] >> j) & 1)
+                  trtable[ch] = dest_states_word[k];
+            }
+          else /* not WORD_CHAR */
+            {
+              for (k = 0; k < ndests; ++k)
+                if ((dests_ch[k][i] >> j) & 1)
+                  trtable[ch] = dest_states[k];
+            }
+        }
+  /* new line */
+  for (k = 0; k < ndests; ++k)
+    if (bitset_contain (acceptable, NEWLINE_CHAR))
+      trtable[NEWLINE_CHAR] = dest_states_nl[k];
+
+  re_free (dest_states_nl);
+  re_free (dest_states_word);
+  re_free (dest_states);
+
+  re_node_set_free (&follows);
+  for (i = 0; i < ndests; ++i)
+    re_node_set_free (dests_node + i);
+
+  re_free (dests_ch);
+  re_free (dests_node);
+
+  return trtable;
+}
+
+/* Group all nodes belonging to STATE into several destinations.
+   Then for all destinations, set the nodes belonging to the destination
+   to DESTS_NODE[i] and set the characters accepted by the destination
+   to DEST_CH[i].  This function return the number of destinations.  */
+
+static int
+group_nodes_into_DFAstates (preg, state, dests_node, dests_ch)
+    const regex_t *preg;
+    const re_dfastate_t *state;
+    re_node_set *dests_node;
+    bitset *dests_ch;
+{
+  const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  int i, j, k;
+  int ndests; /* Number of the destinations from `state'.  */
+  bitset accepts; /* Characters a node can accept.  */
+  const re_node_set *cur_nodes = &state->nodes;
+  bitset_empty (accepts);
+  ndests = 0;
+
+  /* For all the nodes belonging to `state',  */
+  for (i = 0; i < cur_nodes->nelem; ++i)
+    {
+      unsigned int constraint = 0;
+      re_token_t *node = &dfa->nodes[cur_nodes->elems[i]];
+      re_token_type_t type = node->type;
+
+      if (type == OP_CONTEXT_NODE)
+        {
+          constraint = node->constraint;
+          node = dfa->nodes + node->opr.ctx_info->entity;
+          type = node->type;
+        }
+
+      /* Enumerate all single byte character this node can accept.  */
+      if (type == CHARACTER)
+        bitset_set (accepts, node->opr.c);
+      else if (type == SIMPLE_BRACKET)
+        {
+          bitset_merge (accepts, node->opr.sbcset);
+        }
+      else if (type == OP_PERIOD)
+        {
+          bitset_set_all (accepts);
+          if (!(preg->syntax & RE_DOT_NEWLINE))
+            bitset_clear (accepts, '\n');
+          if (preg->syntax & RE_DOT_NOT_NULL)
+            bitset_clear (accepts, '\0');
+        }
+      else
+        continue;
+
+      /* Check the `accepts' and sift the characters which are not
+         match it the context.  */
+      if (constraint)
+        {
+          if (constraint & NEXT_WORD_CONSTRAINT)
+            for (j = 0; j < BITSET_UINTS; ++j)
+              accepts[j] &= dfa->word_char[j];
+          else if (constraint & NEXT_NOTWORD_CONSTRAINT)
+            for (j = 0; j < BITSET_UINTS; ++j)
+              accepts[j] &= ~dfa->word_char[j];
+          else if (constraint & NEXT_NEWLINE_CONSTRAINT)
+            {
+              int accepts_newline = bitset_contain (accepts, NEWLINE_CHAR);
+              bitset_empty (accepts);
+              if (accepts_newline)
+                bitset_set (accepts, NEWLINE_CHAR);
+              else
+                continue;
+            }
+        }
+
+      /* Then divide `accepts' into DFA states, or create a new
+         state.  */
+      for (j = 0; j < ndests; ++j)
+        {
+          bitset intersec; /* Intersection sets, see below.  */
+          bitset remains;
+          /* Flags, see below.  */
+          int has_intersec, not_subset, not_consumed;
+
+          /* Optimization, skip if this state doesn't accept the character.  */
+          if (type == CHARACTER && !bitset_contain (dests_ch[j], node->opr.c))
+            continue;
+
+          /* Enumerate the intersection set of this state and `accepts'.  */
+          has_intersec = 0;
+          for (k = 0; k < BITSET_UINTS; ++k)
+            has_intersec |= intersec[k] = accepts[k] & dests_ch[j][k];
+          /* And skip if the intersection set is empty.  */
+          if (!has_intersec)
+            continue;
+
+          /* Then check if this state is a subset of `accepts'.  */
+          not_subset = not_consumed = 0;
+          for (k = 0; k < BITSET_UINTS; ++k)
+            {
+              not_subset |= remains[k] = ~accepts[k] & dests_ch[j][k];
+              not_consumed |= accepts[k] = accepts[k] & ~dests_ch[j][k];
+            }
+
+          /* If this state isn't a subset of `accepts', create a
+             new group state, which has the `remains'. */
+          if (not_subset)
+            {
+              bitset_copy (dests_ch[ndests], remains);
+              bitset_copy (dests_ch[j], intersec);
+              re_node_set_init_copy (dests_node + ndests, &dests_node[j]);
+              ++ndests;
+            }
+
+          /* Put the position in the current group. */
+          re_node_set_insert (&dests_node[j], cur_nodes->elems[i]);
+
+          /* If all characters are consumed, go to next node. */
+          if (!not_consumed)
+            break;
+        }
+      /* Some characters remain, create a new group. */
+      if (j == ndests)
+        {
+          bitset_copy (dests_ch[ndests], accepts);
+          re_node_set_init_1 (dests_node + ndests, cur_nodes->elems[i]);
+          ++ndests;
+          bitset_empty (accepts);
+        }
+    }
+  return ndests;
+}
+
+/* Check how many bytes the node `dfa->nodes[node_idx]' accepts.  */
+
+static int
+check_node_accept_bytes (preg, node_idx, input, str_idx)
+    const regex_t *preg;
+    int node_idx, str_idx;
+    const re_string_t *input;
+{
+  const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  const re_token_t *node = dfa->nodes + node_idx;
+  int elem_len = re_string_elem_size_at (input, str_idx);
+  int char_len = re_string_char_size_at (input, str_idx);
+  int i, j;
+#ifdef _LIBC
+  uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+#endif /* _LIBC */
+  if (elem_len <= 1 && char_len <= 1)
+    return 0;
+  if (node->type == OP_PERIOD)
+    {
+      if ((!(preg->syntax & RE_DOT_NEWLINE) &&
+           re_string_byte_at (input, str_idx) == '\n') ||
+          ((preg->syntax & RE_DOT_NOT_NULL) &&
+           re_string_byte_at (input, str_idx) == '\0'))
+        return 0;
+      return char_len;
+    }
+  else if (node->type == COMPLEX_BRACKET)
+    {
+      const re_charset_t *cset = node->opr.mbcset;
+      const unsigned char *pin = re_string_get_buffer (input) + str_idx;
+#ifdef _LIBC
+      if (nrules != 0)
+        {
+          int match_len = 0;
+          unsigned int in_collseq = 0;
+          const int32_t *table, *indirect;
+          const unsigned char *weights, *extra, *collseqwc;
+          int32_t idx;
+          wchar_t wc = 0;
+          /* This #include defines a local function!  */
+# include <locale/weight.h>
+
+          /* match with collating_symbol?  */
+          if (cset->ncoll_syms)
+            extra = (const unsigned char *)
+              _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);
+          for (i = 0; i < cset->ncoll_syms; ++i)
+            {
+              const unsigned char *coll_sym = extra + cset->coll_syms[i];
+              /* Compare the length of input collating element and
+                 the length of current collating element.  */
+              if (*coll_sym != elem_len)
+                continue;
+              /* Compare each bytes.  */
+              for (j = 0; j < *coll_sym; j++)
+                if (pin[j] != coll_sym[1 + j])
+                  break;
+              if (j == *coll_sym)
+                {
+                  /* Match if every bytes is equal.  */
+                  match_len = j;
+                  goto check_node_accept_bytes_match;
+                }
+            }
+
+          if (cset->nranges || cset->nchar_classes || cset->nmbchars)
+            wc = re_string_wchar_at (input, str_idx);
+
+          if (cset->nranges)
+            {
+              if (elem_len <= char_len)
+                {
+                  collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC);
+                  in_collseq = collseq_table_lookup (collseqwc, wc);
+                }
+              else
+                in_collseq = find_collation_sequence_value (pin, elem_len);
+            }
+          /* match with range expression?  */
+          for (i = 0; i < cset->nranges; ++i)
+            if (cset->range_starts[i] <= in_collseq
+                && in_collseq <= cset->range_ends[i])
+              {
+                match_len = elem_len;
+                goto check_node_accept_bytes_match;
+              }
+
+          /* match with equivalence_class?  */
+          if (cset->nequiv_classes)
+            {
+              const unsigned char *cp = pin;
+              table = (const int32_t *)
+                _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+              weights = (const unsigned char *)
+                _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
+              extra = (const unsigned char *)
+                _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
+              indirect = (const int32_t *)
+                _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
+              idx = findidx (&cp);
+              if (idx > 0)
+                for (i = 0; i < cset->nequiv_classes; ++i)
+                  {
+                    int32_t equiv_class_idx = cset->equiv_classes[i];
+                    size_t weight_len = weights[idx];
+                    if (weight_len == weights[equiv_class_idx])
+                      {
+                        int cnt = 0;
+                        while (cnt <= weight_len
+                               && (weights[equiv_class_idx + 1 + cnt]
+                                   == weights[idx + 1 + cnt]))
+                          ++cnt;
+                        if (cnt > weight_len)
+                          {
+                            match_len = elem_len;
+                            goto check_node_accept_bytes_match;
+                          }
+                      }
+                  }
+            }
+
+          /* match with multibyte character?  */
+          for (i = 0; i < cset->nmbchars; ++i)
+            if (wc == cset->mbchars[i])
+              {
+                match_len = char_len;
+                goto check_node_accept_bytes_match;
+              }
+
+          /* match with character_class?  */
+          for (i = 0; i < cset->nchar_classes; ++i)
+            {
+              wctype_t wt = cset->char_classes[i];
+              if (__iswctype (wc, wt))
+                {
+                  match_len = char_len;
+                  goto check_node_accept_bytes_match;
+                }
+            }
+
+        check_node_accept_bytes_match:
+          if (!cset->non_match)
+            return match_len;
+          else
+            {
+              if (match_len > 0)
+                return 0;
+              else
+                return re_string_elem_size_at (input, str_idx);
+            }
+        }
+#endif
+    }
+  return 0;
+}
+
+#ifdef _LIBC
+static unsigned int
+find_collation_sequence_value (mbs, mbs_len)
+    const unsigned char *mbs;
+    size_t mbs_len;
+{
+  uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+  if (nrules == 0)
+    {
+      if (mbs_len == 1)
+        {
+          /* No valid character.  Match it as a single byte character.  */
+          const unsigned char *collseq = (const unsigned char *)
+            _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB);
+          return collseq[mbs[0]];
+        }
+      return UINT_MAX;
+    }
+  else
+    {
+      int32_t idx;
+      const unsigned char *extra = (const unsigned char *)
+        _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);
+
+      for (idx = 0; ;)
+        {
+          int mbs_cnt, found = 0;
+          int32_t elem_mbs_len;
+          /* Skip the name of collating element name.  */
+          idx = idx + extra[idx] + 1;
+          elem_mbs_len = extra[idx++];
+          if (mbs_len == elem_mbs_len)
+            {
+              for (mbs_cnt = 0; mbs_cnt < elem_mbs_len; ++mbs_cnt)
+                if (extra[idx + mbs_cnt] != mbs[mbs_cnt])
+                  break;
+              if (mbs_cnt == elem_mbs_len)
+                /* Found the entry.  */
+                found = 1;
+            }
+          /* Skip the byte sequence of the collating element.  */
+          idx += elem_mbs_len;
+          /* Adjust for the alignment.  */
+          idx = (idx + 3) & ~3;
+          /* Skip the collation sequence value.  */
+          idx += sizeof (uint32_t);
+          /* Skip the wide char sequence of the collating element.  */
+          idx = idx + sizeof (uint32_t) * (extra[idx] + 1);
+          /* If we found the entry, return the sequence value.  */
+          if (found)
+            return *(uint32_t *) (extra + idx);
+          /* Skip the collation sequence value.  */
+          idx += sizeof (uint32_t);
+        }
+    }
+}
+#endif
+
+/* Check whether the node accepts the byte which is IDX-th
+   byte of the INPUT.  */
+
+static int
+check_node_accept (preg, node, input, idx, eflags)
+    const regex_t *preg;
+    const re_token_t *node;
+    const re_string_t *input;
+    int idx, eflags;
+{
+  const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
+  const re_token_t *cur_node;
+  unsigned char ch;
+  if (node->type == OP_CONTEXT_NODE)
+    {
+      /* The node has constraints.  Check whether the current context
+         satisfies the constraints.  */
+      unsigned int context = re_string_context_at (input, idx, eflags,
+                                                   preg->newline_anchor);
+      if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context))
+        return 0;
+      cur_node = dfa->nodes + node->opr.ctx_info->entity;
+    }
+  else
+    cur_node = node;
+
+  ch = re_string_byte_at (input, idx);
+  if (cur_node->type == CHARACTER)
+    return cur_node->opr.c == ch;
+  else if (cur_node->type == SIMPLE_BRACKET)
+    return bitset_contain (cur_node->opr.sbcset, ch);
+  else if (cur_node->type == OP_PERIOD)
+    return !((ch == '\n' && !(preg->syntax & RE_DOT_NEWLINE))
+             || (ch == '\0' && (preg->syntax & RE_DOT_NOT_NULL)));
+  else
+    return 0;
+}
+
+/* Functions for matching context.  */
+
+static void
+match_ctx_init (mctx, eflags, n)
+    re_match_context_t *mctx;
+    int eflags;
+    int n;
+{
+  mctx->eflags = eflags;
+  mctx->match_first = mctx->match_last = -1;
+  if (n > 0)
+    mctx->bkref_ents = re_malloc (struct re_backref_cache_entry, n);
+  else
+    mctx->bkref_ents = NULL;
+  mctx->nbkref_ents = 0;
+  mctx->abkref_ents = n;
+  mctx->max_bkref_len = 0;
+}
+
+static void
+match_ctx_free (mctx)
+    re_match_context_t *mctx;
+{
+  re_free (mctx->bkref_ents);
+}
+
+/* Add a new backreference entry to the cache.  */
+
+static void
+match_ctx_add_entry (mctx, node, from, to)
+    re_match_context_t *mctx;
+    int node, from, to;
+{
+  if (mctx->nbkref_ents >= mctx->abkref_ents)
+    {
+      mctx->bkref_ents = re_realloc (mctx->bkref_ents,
+                                     struct re_backref_cache_entry,
+                                     mctx->abkref_ents * 2);
+      memset (mctx->bkref_ents + mctx->nbkref_ents, '\0',
+             sizeof (struct re_backref_cache_entry) * mctx->abkref_ents);
+      mctx->abkref_ents *= 2;
+    }
+  mctx->bkref_ents[mctx->nbkref_ents].node = node;
+  mctx->bkref_ents[mctx->nbkref_ents].from = from;
+  mctx->bkref_ents[mctx->nbkref_ents++].to = to;
+  if (mctx->max_bkref_len < to - from)
+    mctx->max_bkref_len = to - from;
+}