RE: DO NOT REPLY [Bug 17378] - Xalan DOM to SAX namespace "bug" alternative fix.

[Bruno Dumon <br...@outerthought.org>]

On Wed, 2003-02-26 at 08:07, Carsten Ziegeler wrote:
> Great, Bruno!
> 
> This was exactly what I was thinking about after your comments
> from monday - but you beat me :)

:-P

> 
> Now, what do you think of not creating an own streamer but
> integrating it into the usual DOMStreamer - this would be
> a little bit less performant but avoid any problems.
> Because if you stream a DOM you might not know which streamer
> you have to use, so you end up using your alternative
> in every place anyway.
> 
> What do you think?
> 

mixed feelings. How about we make it a property of the DOMStreamer,
setNamespaceCorrection(true/false) or something, and make it true by
default?

In case it is false, the normal jaxp identity transformer would be used.

BTW, these issues about namespaces made me think of an old idea: it
would be handy if there was some "CheckTransformer" that could check
that the output generated by another transformer is valid, i.e. that all
namespace-related events are correct, that the values of the
namespaceURI, localName and qName make sense, that for each startElement
there's an endElement, ...

-- 
Bruno Dumon                             http://outerthought.org/
Outerthought - Open Source, Java & XML Competence Support Center
bruno@outerthought.org

RE: DO NOT REPLY [Bug 17378] - Xalan DOM to SAX namespace"bug" alternative fix.

[Carsten Ziegeler <cz...@s-und-n.de>]

Bruno Dumon wrote:
> > 
> > Now, what do you think of not creating an own streamer but
> > integrating it into the usual DOMStreamer - this would be
> > a little bit less performant but avoid any problems.
> > Because if you stream a DOM you might not know which streamer
> > you have to use, so you end up using your alternative
> > in every place anyway.
> > 
> > What do you think?
> > 
> 
> mixed feelings. How about we make it a property of the DOMStreamer,
> setNamespaceCorrection(true/false) or something, and make it true by
> default?
> 
+1

> In case it is false, the normal jaxp identity transformer would be used.
Ok, can you please change your patch and I will apply it then asap.

> 
> BTW, these issues about namespaces made me think of an old idea: it
> would be handy if there was some "CheckTransformer" that could check
> that the output generated by another transformer is valid, i.e. that all
> namespace-related events are correct, that the values of the
> namespaceURI, localName and qName make sense, that for each startElement
> there's an endElement, ...
> 
Sounds good! Go for it :)

Thanks
Carsten