Really fix regular expression to avoid overcapture

Previously there was a bogus .* in the regular expression that matched
newlines which is bad and would cause RDoc to find the wrong comment
(which was fixed by another .* removed in the previous commit).

Now we remove these things properly.

Fixes #327 (for real).

Author: drbrain

lib/rdoc/parser/c.rb

@@ -746,10 +746,10 @@ def find_class_comment class_name, class_mod
                          (?:<\s+[:,\w]+)?\n((?>.*?\*/))%xm then
       comment = "/*\n#{$1}"
     elsif @content =~ %r%((?>/\*.*?\*/\s+))
-                         ([\w\.\s]+\s* = \s+)?rb_define_(class|module).*?"(#{class_name})"%xm then
+                         ([\w\.\s]+\s* = \s+)?rb_define_(class|module)[\t (]*?"(#{class_name})"%xm then
       comment = $1
     elsif @content =~ %r%((?>/\*.*?\*/\s+))
-                         ([\w\.\s]+\s* = \s+)?rb_define_(class|module)_under.*?"(#{class_name.split('::').last})"%xm then
+                         ([\w\. \t]+ = \s+)?rb_define_(class|module)_under[\t (]*?"(#{class_name.split('::').last})"%xm then
       comment = $1
     else
       comment = ''