Make LibXML transcoding-passing regex stricter and more readable.

1 files changed, 10 insertions, 5 deletions

diff --git a/httpd/cgi-bin/check b/httpd/cgi-bin/check
index 0cf6ca6..91adf57 100755
--- a/httpd/cgi-bin/check
+++ b/httpd/cgi-bin/check
@@ -14,7 +14,7 @@
 # This source code is available under the license at:
 #     http://www.w3.org/Consortium/Legal/copyright-software
 #
-# $Id: check,v 1.751 2009-12-12 20:06:36 ville Exp $
+# $Id: check,v 1.752 2009-12-14 20:44:49 ville Exp $
 
 #
 # We need Perl 5.8.0+.
@@ -197,7 +197,7 @@ EOF
 
     #
     # Strings
-    $VERSION = q$Revision: 1.751 $;
+    $VERSION = q$Revision: 1.752 $;
     $VERSION =~ s/Revision: ([\d\.]+) /$1/;
 
     # Read friendly error message file
@@ -619,9 +619,14 @@ if (&is_xml($File)) {
         # the XML parser will check the value of encoding attribute in XML
         # declaration so we have to amend it to reflect transcoding.
         # see Bug 4867
-        $xml_string =~ s/(<\?xml.*)
-  (encoding[\x20|\x09|\x0D|\x0A]*=[\x20|\x09|\x0D|\x0A]*(?:"[A-Za-z][a-zA-Z0-9_-]+"|'[A-Za-z][a-zA-Z0-9_-]+'))
-  (.*\?>)/$1encoding="utf-8"$3/sx;
+        $xml_string =~ s/
+               (^<\?xml\b[^>]*[\x20\x09\x0D\x0A])
+               (encoding[\x20\x09\x0D\x0A]*=[\x20\x09\x0D\x0A]*
+                   (?:(["'])[A-Za-z][a-zA-Z0-9_-]+\3)
+               )
+               ([^>].*\?>)
+           /$1encoding="UTF-8"$4/sx;
+
         eval { $xmlparser->parse_string($xml_string); };
         $xml_string = undef;
         my $xml_parse_errors_line = undef;