Hi Monks, I'm using XML::DOM to parse, modify then save an XML document. I'm having a problem in that once I save a document which contained character references/entities, they are saved as multibyte characters, without beind re-encoded. XML::DOM then refuses the parse its own output. Please forgive me if I have the terminology wrong, here is an example of what happens:

Input:

<?xml version="1.0"?>
<DocumentRoot>
        <Element Attr="B&#xE4;r" />
</DocumentRoot>
[download]

Script:

use XML::DOM;
my $file = @ARGV[0];

my $parser = new XML::DOM::Parser();

my $doc;
eval {
        $doc = $parser->parsefile( $file );
};
if ($@) {
        die "parsefile() failed: $@\n";
}
$doc->printToFile($file."_out");
exit;
[download]

The output file created has the extended character (a umlaut) written un-encoded. I'm not sure if this will display properly:

<?xml version="1.0"?>
<DocumentRoot>
        <Element Attr="Bär"/>
</DocumentRoot>
[download]

If I then pass the this output back to the script, I get:

parsefile() failed:
not well-formed (invalid token) at line 3, column 17, byte 54 at /usr/
+lib/perl5/XML/Parser.pm line 187
[download]

I've had a look in the XML::DOM code, and I reckon that encodeText() would be the place to do it, however this appears to take a list of characters which should be encoded - for Unicode this would be a pretty big list :-) It's quite possible that there's something about hidden nodes which could be relevant - can anyone point me in the right direction?

I guess if all else fails, a bit of dirty regex-ery on the output could work :-P


Thanks!