[BZ #693]

Update.
2005-01-26  Jakub Jelinek  <jakub@redhat.com>

	* sysdeps/unix/sysv/linux/i386/sysdep.h
	(SYSCALL_ERROR_HANDLER_TLS_STORE): Remove unnecessary 0 imm.

	[BZ #693]
	* posix/regex_internal.h (DUMMY_CONSTRAINT): Rename to...
	(WORD_DELIM_CONSTRAINT): ...this.
	(NOT_WORD_DELIM_CONSTRAINT): Define.
	(re_context_type): Add INSIDE_NOTWORD and NOT_WORD_DELIM,
	change WORD_DELIM to use WORD_DELIM_CONSTRAINT.
	* posix/regcomp.c (peek_token): For \B create NOT_WORD_DELIM
	anchor instead of INSIDE_WORD.
	(parse_expression): Handle NOT_WORD_DELIM constraint.
	* posix/bug-regex19.c (tests): Adjust tests that relied on \B
	being inside word instead of not word delim.
	* posix/tst-rxspencer.c (mb_frob_pattern): Don't frob escaped
	characters.
	* posix/rxspencer/tests: Add some new tests.

4 files changed, 52 insertions, 21 deletions

diff --git a/posix/bug-regex19.c b/posix/bug-regex19.c
index 4000b19..3a173a6 100644
--- a/posix/bug-regex19.c
+++ b/posix/bug-regex19.c
@@ -1,5 +1,5 @@
 /* Regular expression tests.
-   Copyright (C) 2003 Free Software Foundation, Inc.
+   Copyright (C) 2003, 2005 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Jakub Jelinek <jakub@redhat.com>, 2003.
 
@@ -170,22 +170,22 @@ static struct test_s
   {ERE, "[^k]\\B[^k]", "kBk", 0, -1},
   {ERE, "[^C]\\B[^C]", "CCCABA", 0, 3},
   {ERE, "[^C]\\B[^C]", "CBC", 0, -1},
-  {ERE, ".(\\b|\\B).", "=~AB", 0, 1},
+  {ERE, ".(\\b|\\B).", "=~AB", 0, 0},
   {ERE, ".(\\b|\\B).", "A=C", 0, 0},
   {ERE, ".(\\b|\\B).", "ABC", 0, 0},
-  {ERE, ".(\\b|\\B).", "=~\\!", 0, -1},
-  {ERE, "[^k](\\b|\\B)[^k]", "=~AB", 0, 1},
+  {ERE, ".(\\b|\\B).", "=~\\!", 0, 0},
+  {ERE, "[^k](\\b|\\B)[^k]", "=~AB", 0, 0},
   {ERE, "[^k](\\b|\\B)[^k]", "A=C", 0, 0},
   {ERE, "[^k](\\b|\\B)[^k]", "ABC", 0, 0},
-  {ERE, "[^k](\\b|\\B)[^k]", "=~kBD", 0, 3},
-  {ERE, "[^k](\\b|\\B)[^k]", "=~\\!", 0, -1},
-  {ERE, "[^k](\\b|\\B)[^k]", "=~kB", 0, -1},
-  {ERE, "[^C](\\b|\\B)[^C]", "=~AB", 0, 1},
+  {ERE, "[^k](\\b|\\B)[^k]", "=~kBD", 0, 0},
+  {ERE, "[^k](\\b|\\B)[^k]", "=~\\!", 0, 0},
+  {ERE, "[^k](\\b|\\B)[^k]", "=~kB", 0, 0},
+  {ERE, "[^C](\\b|\\B)[^C]", "=~AB", 0, 0},
   {ERE, "[^C](\\b|\\B)[^C]", "A=C", 0, 0},
   {ERE, "[^C](\\b|\\B)[^C]", "ABC", 0, 0},
-  {ERE, "[^C](\\b|\\B)[^C]", "=~CBD", 0, 3},
-  {ERE, "[^C](\\b|\\B)[^C]", "=~\\!", 0, -1},
-  {ERE, "[^C](\\b|\\B)[^C]", "=~CB", 0, -1},
+  {ERE, "[^C](\\b|\\B)[^C]", "=~CBD", 0, 0},
+  {ERE, "[^C](\\b|\\B)[^C]", "=~\\!", 0, 0},
+  {ERE, "[^C](\\b|\\B)[^C]", "=~CB", 0, 0},
   {ERE, "\\b([A]|[!]|.B)", "A=AC", 0, 0},
   {ERE, "\\b([A]|[!]|.B)", "=AC", 0, 1},
   {ERE, "\\b([A]|[!]|.B)", "!AC", 0, 1},
diff --git a/posix/regcomp.c b/posix/regcomp.c
index a8c1f74..2ac9953 100644
--- a/posix/regcomp.c
+++ b/posix/regcomp.c
@@ -1859,7 +1859,7 @@ peek_token (token, input, syntax)
 	  if (!(syntax & RE_NO_GNU_OPS))
 	    {
 	      token->type = ANCHOR;
-	      token->opr.ctx_type = INSIDE_WORD;
+	      token->opr.ctx_type = NOT_WORD_DELIM;
 	    }
 	  break;
 	case 'w':
@@ -2349,15 +2349,25 @@ parse_expression (regexp, preg, token, syntax, nest, err)
       break;
     case ANCHOR:
       if ((token->opr.ctx_type
-	   & (WORD_DELIM | INSIDE_WORD | WORD_FIRST | WORD_LAST))
+	   & (WORD_DELIM | NOT_WORD_DELIM | WORD_FIRST | WORD_LAST))
 	  && dfa->word_ops_used == 0)
 	init_word_char (dfa);
-      if (token->opr.ctx_type == WORD_DELIM)
+      if (token->opr.ctx_type == WORD_DELIM
+          || token->opr.ctx_type == NOT_WORD_DELIM)
 	{
 	  bin_tree_t *tree_first, *tree_last;
-	  token->opr.ctx_type = WORD_FIRST;
-	  tree_first = re_dfa_add_tree_node (dfa, NULL, NULL, token);
-	  token->opr.ctx_type = WORD_LAST;
+	  if (token->opr.ctx_type == WORD_DELIM)
+	    {
+	      token->opr.ctx_type = WORD_FIRST;
+	      tree_first = re_dfa_add_tree_node (dfa, NULL, NULL, token);
+	      token->opr.ctx_type = WORD_LAST;
+            }
+          else
+            {
+	      token->opr.ctx_type = INSIDE_WORD;
+	      tree_first = re_dfa_add_tree_node (dfa, NULL, NULL, token);
+	      token->opr.ctx_type = INSIDE_NOTWORD;
+            }
 	  tree_last = re_dfa_add_tree_node (dfa, NULL, NULL, token);
 	  token->type = OP_ALT;
 	  tree = re_dfa_add_tree_node (dfa, tree_first, tree_last, token);
diff --git a/posix/rxspencer/tests b/posix/rxspencer/tests
index a724252..a8b6e4b 100644
--- a/posix/rxspencer/tests
+++ b/posix/rxspencer/tests
@@ -526,3 +526,12 @@ a((b+|((c)*)))+d	-	abcd	abcd	c,c,c,c
 (((\b))){0}	-	x	@x	-,-,-
 a(((.*)))b((\2)){0}c	-	abc	abc	@bc,@bc,@bc,-,-
 a(((.*)))b((\1)){0}c	-	axbc	axbc	x,x,x,-,-
+
+\b	&	SaT	@aT
+\b	&	aT	@aT
+a.*\b	&	abT	ab
+\b	&	STSS
+\B	&	abc	@bc
+\B	&	aSbTc
+\B	&	SaT	@SaT
+\B	&	aSTSb	@TSb
diff --git a/posix/tst-rxspencer.c b/posix/tst-rxspencer.c
index cb40421..a68bab2 100644
--- a/posix/tst-rxspencer.c
+++ b/posix/tst-rxspencer.c
@@ -1,5 +1,5 @@
 /* Regular expression tests.
-   Copyright (C) 2003 Free Software Foundation, Inc.
+   Copyright (C) 2003, 2005 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Jakub Jelinek <jakub@redhat.com>, 2003.
 
@@ -127,14 +127,15 @@ mb_frob_string (const char *str, const char *letters)
 }
 
 /* Like mb_frob_string, but don't replace anything between
-   [: and :], [. and .] or [= and =].  */
+   [: and :], [. and .] or [= and =] or characters escaped
+   with a backslash.  */
 
 static char *
 mb_frob_pattern (const char *str, const char *letters)
 {
   char *ret, *dst;
   const char *src;
-  int in_class = 0;
+  int in_class = 0, escaped = 0;
 
   if (str == NULL)
     return NULL;
@@ -144,7 +145,18 @@ mb_frob_pattern (const char *str, const char *letters)
     return NULL;
 
   for (src = str, dst = ret; *src; ++src)
-    if (!in_class && strchr (letters, *src))
+    if (*src == '\\')
+      {
+	escaped ^= 1;
+	*dst++ = *src;
+      }
+    else if (escaped)
+      {
+	escaped = 0;
+	*dst++ = *src;
+	continue;
+      }
+    else if (!in_class && strchr (letters, *src))
       dst = mb_replace (dst, *src);
     else
       {