DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4336>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4336

Xalan 2.2.D11 adds a strange Attribute





------- Additional Comments From [EMAIL PROTECTED]  2001-11-12 06:16 -------
> the problem is that this namespace declaration did not appear in the 
> input document.

It didn't appear explicitly. It's implicitly declared by the Namespaces In XML 
spec. The XPath data model requires that namespace declarations, including this 
implicit one, be visible as Namespace Nodes, so we have to synthesize a 
Namespace Node to handle it. (We failed to do so in previous versions of Xalan, 
but this was correctly reported as a bug.)

I understand the issue re canonical XML... but our output is not currently 
guaranteed to be in Canoncial Form, and a canonicalizer should already know how 
to discard this re-declaration of the default.

>it looks like a default attribute which is not explicitly defined in 
>the input but only in the DTD. 

Good analogy. This effectively _is_ a default attribute -- defined not by the 
DTD, but the XML langauge and XML Namespaces -- and you should probably be 
processing it similarly. Since the XPath data model doesn't distinguish between 
default and specified attribute values, you shouldn't be too suprised that 
passing the data through Xalan loses that distinction.


There are a few subtleties in our Namespace Node implementation where we don't 
completely match the spec (partly because our best guess is that the spec is 
going to change). But leaving this node out was a non-subtle error, and really 
did need to be fixed.

... I think we're working as specified. It may not be what you'd prefer -- I 
happen to dislike the whole concept of namespace nodes, personally -- but I 
don't think dropping this node from the DTM is an available option.

Reply via email to