64826 - wildcard not working for attribute axis

Status: VERIFIED FIXED

(Core :: XSLT, defect)

Description

[Axel Hecht]

attribute::* doesn't match, because WildCardExpr::matches only returns
Element nodes. I will add a optional argument to matches, defaulting to
Node::ELEMENT_NODE, to match nodetype against.
I will call nodeExpr->matches(attr, context, cs, Node::ATTRIBUTE_NODE) then in
LocationStep.cpp:128.
So much, that I remember what to do tomorrow.

Axel

Comment 1

[Peter Van der Beken [:peterv]]

Axel, still working on this?

Comment 2

[Axel Hecht]

This one is worse than I thought it would be.
Basically because I can't just add another parameter to ::matches, it's all
over the code.
What we need to fix this is really a matter of finding the axis from within
WildcardExpr::matches. Keith, that's really above my head, please help.

Axel

Comment 3

[Keith Visco]

I'll do it this weekend. 

Comment 4

[Keith Visco]

Reassinging to myself.

Comment 5

[Keith Visco]

Axel,

I just spend some time looking into this. I am just going to kill off
WildCardExpr. I am putting the proper logic in place for AttributeExpr and 
ElementExpr.

--Keith

Comment 6

[Keith Visco]

I've noticed the namespace related methods for ProcessorState are being used for 
both the Result tree as well as the stylesheet tree....this is not good. I am in 
the process of trying to clean this up. I need to do this, before I can check in 
my fix for this bug, since it depends on having the proper namespace.

Comment 7

[Keith Visco]

I checked in my patches. This should now be fixed. I am awaiting verification 
from Axel. I added a quick test case in test.xsl under the axis identifiers 
section for this.

--Keith

Comment 8

[Axel Hecht]

Not fixed yet, we don't get the document order right here.
How do we do document order for attributes?

This is too hefty for now, maybe later today, this week, or even this life.

Axel

Comment 9

[Keith Visco]

For attributes I use the document order for the element in which the attribute
was declared and I append a ".0"  All attributes for an element have the same 
document order.  

I still think this bug is fixed. If there is a document order problem then it's 
a different bug.

--Keith

Comment 10

[chris hofmann]

not updates for a while on this bug. 
unsetting the target milestone.
maybe it should just be marked fixed.

Comment 11

[Keith Visco]

Axel, is this still an issue?

Comment 12

[Axel Hecht]

resolving. bugs have evolved alot since this one was filed.
remaining issues are in bug 92106

Axel

Comment 13

[Axel Hecht]

bitching buttons, verfication spam