* regex.c (is_in_list): should work weill with UTF-8.

* regex.c (re_match_exec): ditto.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@3363 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

ChangeLog

@@ -23,6 +23,12 @@ Sun Jan 19 23:08:18 2003  Akinori MUSHA  <knu@iDaemons.org>
 	  quotes should not be regarded as meta character.  This bug or
 	  maybe feature was inherited from Perl's shellwords.pl.
 
+Sun Jan 19 14:01:12 2003  UENO Katsuhiro  <unnie@blue.sky.or.jp>
+
+	* regex.c (is_in_list): should work weill with UTF-8.
+
+	* regex.c (re_match_exec): ditto.
+
 Sat Jan 18 14:53:49 2003  Nobuyoshi Nakada  <nobu.nokada@softhome.net>
 
 	* bignum.c (rb_cstr_to_inum): should not erase all 0s, but

hash.c

@@ -101,10 +101,7 @@ rb_any_hash(a)
       default:
 	DEFER_INTS;
 	hval = rb_funcall(a, id_hash, 0);
-	if (FIXNUM_P(hval)) {
-	    hval %= 536870923;
-	}
-	else {
+	if (!FIXNUM_P(hval)) {
 	    hval = rb_funcall(hval, '%', 1, INT2FIX(536870923));
 	}
 	ENABLE_INTS;

lib/cgi.rb

@@ -942,21 +942,24 @@ convert string charset, and set language to "ja".
     private :initialize_query
 
     def [](key)
-      value = @params[key][0]
-      def value.to_ary
-        [self.dup]
-      end
-      def value.[](key)
-        $stderr.puts <<END
+      value, = @params[key]
+      unless value.nil?
+	value = value.dup
+	def value.to_ary
+	  [self]
+	end
+	def value.[](key)
+	  $stderr.puts <<END
 CAUTION! cgi['key'] == cgi.params['key'][0] If want Array, use cgi.params['key']
 END
-        self
-      end
-      def value.first
-        $stderr.puts <<END
+	  self
+	end
+	def value.first
+	  $stderr.puts <<END
 CAUTION! cgi['key'] == cgi.params['key'][0] If want Array, use cgi.params['key']
 END
-        self
+	  self
+	end
       end
       value
     end

regex.c

@@ -698,7 +698,18 @@ set_list_bits(c1, c2, b)
 }
 
 static int
-is_in_list(c, b)
+is_in_list_sbc(c, b)
+    unsigned long c;
+    const unsigned char *b;
+{
+  unsigned short size;
+
+  size = *b++;
+  return ((int)c / BYTEWIDTH < (int)size && b[c / BYTEWIDTH] & 1 << c % BYTEWIDTH);
+}
+  
+static int
+is_in_list_mbc(c, b)
     unsigned long c;
     const unsigned char *b;
 {
@@ -706,9 +717,6 @@ is_in_list(c, b)
   unsigned short i, j;
 
   size = *b++;
-  if ((int)c / BYTEWIDTH < (int)size && b[c / BYTEWIDTH] & 1 << c % BYTEWIDTH) {
-    return 1;
-  }
   b += size + 2;
   size = EXTRACT_UNSIGNED(&b[-2]);
   if (size == 0) return 0;
@@ -727,6 +735,14 @@ is_in_list(c, b)
   return 0;
 }
 
+static int
+is_in_list(c, b)
+    unsigned long c;
+    const unsigned char *b;
+{
+  return is_in_list_sbc(c, b) || is_in_list_mbc(c, b);
+}
+
 static void
 print_partial_compiled_pattern(start, end)
     unsigned char *start;
@@ -3815,19 +3831,25 @@ re_match_exec(bufp, string_arg, size, pos, beg, regs)
 	  int cc, c;
 
 	  PREFETCH;
-	  cc = c = (unsigned char)*d++;
+	  c = (unsigned char)*d++;
 	  if (ismbchar(c)) {
 	    if (d + mbclen(c) - 1 <= dend) {
+	      cc = c;
 	      MBC2WC(c, d);
+	      not = is_in_list_mbc(c, p);
+	      if (!not) {
+		part = not = is_in_list_sbc(cc, p);
+	      }
+	    } else {
+	      not = is_in_list_sbc(c, p);
 	    }
 	  }
-	  else if (TRANSLATE_P())
-	    cc = c = (unsigned char)translate[c];
-
-	  not = is_in_list(c, p);
-	  if (!not && cc != c) {
-	      part = not = is_in_list(cc, p);
+	  else {
+	    if (TRANSLATE_P())
+	      c = (unsigned char)translate[c];
+	    not = is_in_list_sbc(c, p);
 	  }
+
 	  if (*(p - 1) == (unsigned char)charset_not) {
 	    not = !not;
 	  }

sample/test.rb

@@ -375,6 +375,7 @@ r([1,2,[]]){next *[1,2]}
 r([nil,nil,[]]){next *[*[]]}
 r([1,nil,[]]){next *[*[1]]}
 r([1,2,[]]){next *[*[1,2]]}
+
 test_check "condition"
 
 $x = '0';