[XML-SIG] Problem with wddx marshaller

Thomas B. Passin tpassin at comcast.net
Tue Jun 24 21:15:48 EDT 2003


[Christian Scholz]

> But don't I have to use the encoding defined in the XML header?
> Just wondering..
>

No, not "use" - it just tells the processor what encoding to use to read the
data.  Once you have ingested it, it is up to you or the processor how to
handle the characters.  Of course, if the file says it uses one encoding but
actually uses another, you can run into trouble.

Some of the pyxml demo programs have still encoding problems, I think.  XBEL
is the one that I know about.    I hacked up some patches, but it needs more
thought.  XBEL is a really difficult case, because it has an impossible job.
It has to take browser bookmarks and turn them into xml files, but browser
bookmarks often have page titles in the wrong encoding, because the browsers
tend to just copy them from the files rather than to correct the encoding
(at least, it sure seems like that is how it works).

I combine XBEL files created from the bookmarks from several browsers, and
with my patches I can get titles with the right characters instead of As
with little hats over them and the like.  But as I say, it is a bit of a
hack.  Sometime when I have more time, maybe we can rework the XBEL files.

Cheers,

Tom P





More information about the XML-SIG mailing list