Re: enabling utf8 encoding for sax parer XML::SAX::Expat

eyal edri schrieb:

> i'm getting an error while parsing an xml with non-ascii chars: "not
> well formed xml at line..... "

The document content doesn't match the specified encoding. The specified
encoding is the one in the XML declaration, like this:

   <?xml version="1.0" encoding="Windows-1252"?>

Where there is no XML declaration, the encoding must be either UTF-8 or
UTF-16, and when UTF-16, this must be indicated by a BOM (byte-order
mark).

I guess your parser is correct, and your document is broken.

>    1. i don't have control of the src xml, so i can't make them give
>    it to me properly.

If it's broken, it's broken. Should be easy to fix at the source.

Failing that, you could preprocess the file using some non-XML
technology, like Perl, and hack the missing XML declaration in,
which may or may not be sufficient to repair the document.

>    - how can i tell the parser to open the xml in a certain encoding
>    (despite the sender fucked up and didn't add the magic numbers) ?

As explained above: fix the file.

>    - should i use a different sax parser?

You can try others, but I think Expat is okay.

> and should i always use :encoding(UTF-8) when opening/writing files?

When using open() to open a file encoded in UTF-8, yes; else no.
Check the following, recorded on a Latin1 terminal:

milu@colinux:~/ds > cat Kaese.pl
use strict;
use warnings;
my $filename = shift or die 'Datei!';
open my $fh, '<encoding(utf-8)', $filename;
print <$fh>;
close $fh;
milu@colinux:~/ds > cat Kaese.iso88591.txt
Käse
milu@colinux:~/ds > cat Kaese.utf8.txt
Käse
milu@colinux:~/ds > perl Kaese.pl Kaese.utf8.txt
Käse
milu@colinux:~/ds > perl Kaese.pl Kaese.iso88591.txt
utf8 "\xE4" does not map to Unicode at Kaese.pl line 5.
K\xE4se

Michael Ludwig
_______________________________________________
Perl-XML mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs