yann@1365: Original patch from: gentoo/src/patchsets/glibc/2.9/0052_all_glibc-2.9-regex-BZ9697.patch
yann@1365: 
yann@1365: -= BEGIN original header =-
yann@1365: http://sourceware.org/ml/libc-alpha/2009-01/msg00005.html
yann@1365: 
yann@1365: From ea8ca0dfcbf2721bcf2c08ce3c01d5764b827572 Mon Sep 17 00:00:00 2001
yann@1365: From: Ulrich Drepper <drepper@redhat.com>
yann@1365: Date: Thu, 8 Jan 2009 00:42:28 +0000
yann@1365: Subject: [PATCH] (re_compile_fastmap_iter): Rewrite COMPLEX_BRACKET handling.
yann@1365: 
yann@1365: -= END original header =-
yann@1365: 
yann@1365: diff -durN glibc-2_9.orig/posix/regcomp.c glibc-2_9/posix/regcomp.c
yann@1365: --- glibc-2_9.orig/posix/regcomp.c	2008-05-15 05:07:21.000000000 +0200
yann@1365: +++ glibc-2_9/posix/regcomp.c	2009-02-02 22:00:41.000000000 +0100
yann@1365: @@ -350,47 +350,67 @@
yann@1365:  #ifdef RE_ENABLE_I18N
yann@1365:        else if (type == COMPLEX_BRACKET)
yann@1365:  	{
yann@1365: -	  int i;
yann@1365:  	  re_charset_t *cset = dfa->nodes[node].opr.mbcset;
yann@1365: -	  if (cset->non_match || cset->ncoll_syms || cset->nequiv_classes
yann@1365: -	      || cset->nranges || cset->nchar_classes)
yann@1365: -	    {
yann@1365: +	  int i;
yann@1365: +
yann@1365:  # ifdef _LIBC
yann@1365: -	      if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0)
yann@1365: +	  /* See if we have to try all bytes which start multiple collation
yann@1365: +	     elements.
yann@1365: +	     e.g. In da_DK, we want to catch 'a' since "aa" is a valid
yann@1365: +		  collation element, and don't catch 'b' since 'b' is
yann@1365: +		  the only collation element which starts from 'b' (and
yann@1365: +		  it is caught by SIMPLE_BRACKET).  */
yann@1365: +	      if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0
yann@1365: +		  && (cset->ncoll_syms || cset->nranges))
yann@1365:  		{
yann@1365: -		  /* In this case we want to catch the bytes which are
yann@1365: -		     the first byte of any collation elements.
yann@1365: -		     e.g. In da_DK, we want to catch 'a' since "aa"
yann@1365: -			  is a valid collation element, and don't catch
yann@1365: -			  'b' since 'b' is the only collation element
yann@1365: -			  which starts from 'b'.  */
yann@1365:  		  const int32_t *table = (const int32_t *)
yann@1365:  		    _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
yann@1365:  		  for (i = 0; i < SBC_MAX; ++i)
yann@1365:  		    if (table[i] < 0)
yann@1365:  		      re_set_fastmap (fastmap, icase, i);
yann@1365:  		}
yann@1365: -# else
yann@1365: -	      if (dfa->mb_cur_max > 1)
yann@1365: -		for (i = 0; i < SBC_MAX; ++i)
yann@1365: -		  if (__btowc (i) == WEOF)
yann@1365: -		    re_set_fastmap (fastmap, icase, i);
yann@1365: -# endif /* not _LIBC */
yann@1365: -	    }
yann@1365: -	  for (i = 0; i < cset->nmbchars; ++i)
yann@1365: +# endif /* _LIBC */
yann@1365: +
yann@1365: +	  /* See if we have to start the match at all multibyte characters,
yann@1365: +	     i.e. where we would not find an invalid sequence.  This only
yann@1365: +	     applies to multibyte character sets; for single byte character
yann@1365: +	     sets, the SIMPLE_BRACKET again suffices.  */
yann@1365: +	  if (dfa->mb_cur_max > 1
yann@1365: +	      && (cset->nchar_classes || cset->non_match
yann@1365: +# ifdef _LIBC
yann@1365: +		  || cset->nequiv_classes
yann@1365: +# endif /* _LIBC */
yann@1365: +		 ))
yann@1365:  	    {
yann@1365: -	      char buf[256];
yann@1365: -	      mbstate_t state;
yann@1365: -	      memset (&state, '\0', sizeof (state));
yann@1365: -	      if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
yann@1365: -		re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
yann@1365: -	      if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
yann@1365: +	      unsigned char c = 0;
yann@1365: +	      do
yann@1365:  		{
yann@1365: -		  if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
yann@1365: -		      != (size_t) -1)
yann@1365: -		    re_set_fastmap (fastmap, 0, *(unsigned char *) buf);
yann@1365: +		  mbstate_t mbs;
yann@1365: +		  memset (&mbs, 0, sizeof (mbs));
yann@1365: +		  if (__mbrtowc (NULL, (char *) &c, 1, &mbs) == (size_t) -2)
yann@1365: +		    re_set_fastmap (fastmap, false, (int) c);
yann@1365:  		}
yann@1365: +	      while (++c != 0);
yann@1365:  	    }
yann@1365: +
yann@1365: +	  else
yann@1365: +	    {
yann@1365: +	      /* ... Else catch all bytes which can start the mbchars.  */
yann@1365: +	      for (i = 0; i < cset->nmbchars; ++i)
yann@1365: +		{
yann@1365: +		  char buf[256];
yann@1365: +		  mbstate_t state;
yann@1365: +		  memset (&state, '\0', sizeof (state));
yann@1365: +		  if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1)
yann@1365: +		    re_set_fastmap (fastmap, icase, *(unsigned char *) buf);
yann@1365: +		  if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1)
yann@1365: +		    {
yann@1365: +		      if (__wcrtomb (buf, towlower (cset->mbchars[i]), &state)
yann@1365: +			  != (size_t) -1)
yann@1365: +			re_set_fastmap (fastmap, false, *(unsigned char *) buf);
yann@1365: +		    }
yann@1365: + 		}
yann@1365: + 	    }
yann@1365:  	}
yann@1365:  #endif /* RE_ENABLE_I18N */
yann@1365:        else if (type == OP_PERIOD