* string.c (count_utf8_lead_bytes_with_ulong): fix shift size.

  [ruby-dev:33993]

* string.c (str_utf8_nth) fix wrong counting.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@15700 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

4 files changed, 28 insertions, 14 deletions

diff --git a/ChangeLog b/ChangeLog
index 00aad39c73..e289a15b78 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+Thu Mar  6 04:32:06 2008  NARUSE, Yui  <naruse@ruby-lang.org>
+
+	* string.c (count_utf8_lead_bytes_with_ulong): fix shift size.
+	  [ruby-dev:33993]
+
+	* string.c (str_utf8_nth) fix wrong counting.
+
 Thu Mar  6 00:34:00 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* sprintf.c (rb_str_format): size_t returned from strlen() can be
@@ -7,7 +14,7 @@ Thu Mar  6 00:31:39 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* struct.c (make_struct): preserve encoding of struct name.
 
-Wed Mar 05 22:49:20 2008  NARUSE, Yui  <naruse@ruby-lang.org>
+Wed Mar  5 22:49:20 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* string.c (is_utf8_lead_byte, count_utf8_lead_bytes_with_ulong):
 	  defined for UTF-8 optimization.
@@ -22,7 +29,7 @@ Wed Mar  5 17:53:01 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 	* file.c (rb_file_flock): returns false on EAGAIN if non-blocking.
 	  [ruby-core:15795]
 
-Web Mar  5 17:43:43 2008  Martin Duerst  <duerst@it.aoyama.ac.jp>
+Wed Mar  5 17:43:43 2008  Martin Duerst  <duerst@it.aoyama.ac.jp>
 
 	* transcode.c (transcode_loop): Adjusted detection of invalid
 	  (ill-formed) UTF-8 sequences. Fixing potential security issue, see
@@ -85,7 +92,7 @@ Tue Mar  4 16:29:06 2008  Yukihiro Matsumoto  <matz@ruby-lang.org>
 
 	* parse.y (parser_yylex): disallow non digits '0o' expression.
 
-Tue Mar 04 14:35:12 2008  NARUSE, Yui  <naruse@ruby-lang.org>
+Tue Mar  4 14:35:12 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* io.c (open_key_args): use rb_io_open_with_args instead of rb_f_open.
 	  [ruby-core:15763]
@@ -99,7 +106,7 @@ Tue Mar  4 10:21:03 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 	* gc.c (add_heap): use binary search to find the place to insert the
 	  new heap slot.  [ruby-dev:33983]
 
-Tue Mar 04 05:30:31 2008  NARUSE, Yui  <naruse@ruby-lang.org>
+Tue Mar  4 05:30:31 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* io.c (open_key_args): use rb_io_open instead of rb_f_open.
 	  [ruby-core:15746]
@@ -189,7 +196,7 @@ Sat Mar  1 12:15:42 2008  Yukihiro Matsumoto  <matz@ruby-lang.org>
 	* thread.c (remove_event_hook): should not access freed memory.
 	  [ruby-dev:31820]
 
-Sat Mar 01 10:31:19 2008  NARUSE, Yui  <naruse@ruby-lang.org>
+Sat Mar  1 10:31:19 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* io.c (read_all, rb_io_getline_fast): encoding is io_input_encoding.
 
diff --git a/string.c b/string.c
index 79288ceeb2..89a8b465f2 100644
--- a/string.c
+++ b/string.c
@@ -763,7 +763,7 @@ count_utf8_lead_bytes_with_ulong(const unsigned long *s)
     unsigned long d = *s;
     d |= ~(d>>1);
     d >>= 6;
-    d &= NONASCII_MASK >> 3;
+    d &= NONASCII_MASK >> 7;
     d += (d>>8);
     d += (d>>16);
 #if NONASCII_MASK == 0x8080808080808080UL
@@ -1177,11 +1177,10 @@ str_utf8_nth(const char *p, const char *e, int nth)
 	    if (is_utf8_lead_byte(*p)) nth--;
 	    p++;
 	}
-	while (s < t) {
+	do {
 	    nth -= count_utf8_lead_bytes_with_ulong(s);
-	    if (nth < sizeof(long)) break;
 	    s++;
-	}
+	} while (s < t && sizeof(long) <= nth);
 	p = (char *)s;
     }
     if (0 < nth) {
diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb
index 069554a127..7a9e24f4a1 100644
--- a/test/ruby/test_m17n.rb
+++ b/test/ruby/test_m17n.rb
@@ -810,6 +810,17 @@ class TestM17N < Test::Unit::TestCase
     assert_equal(false, str[0..-1].ascii_only?)
   end
 
+  def test_utf8str_aref
+    s = "abcdefghijklmnopqrstuvwxyz\u{3042 3044 3046 3048 304A}"
+    assert_equal("a", s[0])
+    assert_equal("h", s[7])
+    assert_equal("i", s[8])
+    assert_equal("j", s[9])
+    assert_equal("\u{3044}", s[27])
+    assert_equal("\u{3046}", s[28])
+    assert_equal("\u{3048}", s[29])
+  end
+
   def test_str_aref_len
     assert_equal(a("\xa1"), a("\xc2\xa1\xc2\xa2\xc2\xa3")[1, 1])
     assert_equal(a("\xa1\xc2"), a("\xc2\xa1\xc2\xa2\xc2\xa3")[1, 2])
diff --git a/test/ruby/test_regexp.rb b/test/ruby/test_regexp.rb
index 29783d4153..17fe9fa45f 100644
--- a/test/ruby/test_regexp.rb
+++ b/test/ruby/test_regexp.rb
@@ -123,6 +123,8 @@ class TestRegexp < Test::Unit::TestCase
     r = /./
     m = r.match("a")
     assert_equal(r, m.regexp)
+    re = /foo/
+    assert_equal(re, re.match("foo").regexp)
   end
 
   def test_source
@@ -188,11 +190,6 @@ class TestRegexp < Test::Unit::TestCase
     assert_equal(/foo/, m.dup.regexp)
   end
 
-  def test_match_regexp
-    re = /foo/
-    assert_equal(re, re.match("foo").regexp)
-  end
-
   def test_match_size
     m = /(.)(.)(\d+)(\d)/.match("THX1138.")
     assert_equal(5, m.size)