Re: [Gestalt-users] Bug in namespace-uri-for-prefix() ?

[Colin A. <col...@ho...>]

>From: Florent Georges <dar...@ya...>

>   I don't know Eiffel, but if I understand correctly, the

You understand perfectly.

>item type passed to XM_XPATH_ORPHAN.make is not correct.
>It is taken in XM_XSLT_TEXT_CONSTRUCTOR from
>item_type.primitive_type, where item_type is set in
>XM_XSLT_COMPILED_NAMESPACE:
>
>     item_type: XM_XPATH_ITEM_TYPE is
>       do
>         Result := any_item
>       end
>
>   I wonder if the primitive_type of any_type is correct
>here.  Doesn't have it to be Namespace_node ?

Yes. As you will have seen, I had a comment in the code querying that.
In fact, its an accident of history.
Originally, I had full support for namespace nodes.
But when I realized they were deprecated and optional, I removed them - 
overzealously.
finally, the spec. was clarified so that support for the namespace axis is 
required for XPath 1.0 compatibility mode.

After some agonizing, I reluctantly decided to retain support for XPath 1.0 
compatibility, and so I restored namespace axis support, and set about 
adding back stuff like Namespace_node - just sufficient to get it to compile 
again.

This was an area I overlooked, seeing as I had never fully implemented it, 
there were no regression tests to fail.

>   But once again I don't know Eiffel, I just tried to read
>the sources and figure what happen.  Maybe that will help
>you, maybe not.

I had already fixed it (although there turned out to be a lot more than just 
that to fix).

Since you find it so easy to understand Eiffel (it is designed for maximum 
clarity), why not take a look at 
http:?se.inf.ethz.ch?teaching?ws2006?0001?english_index.html (substitute 
slashes for ?)?

_________________________________________________________________
Think you're a film buff? Play the Movie Mogul quiz and win fantastic 
prizes!  http://www.msnmoviemogul.com