I'm going to reply to this in
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=2445.
-scott
Donald Ball
<balld@websli To: <[EMAIL PROTECTED]>
ngerZ.com> cc: <[EMAIL PROTECTED]>,
<[EMAIL PROTECTED]>
Subject: Re: something funny with
namespaces and xalan2.2dev (fwd)
07/06/2001
02:42 AM
On Mon, 2 Jul 2001 [EMAIL PROTECTED] wrote:
> > I'm not 100% sure, but it looks like the current
> > SAX2DTM code expects to be passed both, and I can imagine that Cocoon
> might
> > be trying to take the shortcut...
>
> The code should work fine if passed only startPrefixMapping and
> endPrefixMapping events. I just wrote a small test for this, and
> everything seems pretty happy (though it gets confused for local-name()
if
> you pass null instead of "" for startPrefixMapping...).
>
> My suspicion is that whoever is generating SAX events within the body
> statement, i.e. the form, input, etc., is not generating the namespaceURI
> argument for startElement. The SAX2DTM will not try and resolve the
> namespace itself, as per:
(sorry for the lateness of this response)
all of the elements in the form were created using SAX by a custom
component. the component strictly creates nodes in the default namespace:
handler.startElement("","form","form",attributes);
the other elements on the page were generated using SAX by cocoon's
FileGenerator, which is ultimately using jaxp (xerces) to parse the file.
> Since a ContentHandler doesn't have a way to set the namespace property
by
> itself, SAX2DTM assumes this property is always true. (It is still in
> error in that it requires the qName argument). (This optionality on
SAX2,
> in my opinion, is really awful.)
>
> Donald, is this making any sense?
a bit. would it be helpful if i got a dump of the SAX events that are
being given to xalan for debugging? if so, let me know. for the time
being, i've simply removed the namespace from that layer. i'm actually
finding that the more i use namespaces, the more i dislike certain aspects
of working with them.
- donald
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
