I'm having some trouble getting Transformiix to parse a document that uses namespaces. The attached test case outputs "[]" instead of "[foo]" See http://groups.google.com/groups?hl=en&th=b162e03e2cdfa928&seekm=3BF3A3EF.5DF9450D%40numerik.uni-kiel.de&frame=off for netnews discussion.

bug 102293 could fix this. We need to check a few landing gears for this.

we can't really fix this one fully until we resolve namespace during parsing, which require the new contexts. We could however work around this and make it work for the most common cases by setting mXPathParseContext to xslDocument.documentElement at the top of ::findTemplate. It's a hack and will not fully work, but it should cover the most common cases and i think that that was in effect what we used to do. We should at least get this in better shape for the next milestone, but until then it might IMHO be better to concentrate our efforts on the RightThing.

Since the plans for the contexts has become somewhat bigger i doubt that we'll be able to get them in before 0.9.7. So this hack-around will make us do about the same as we used to do, which should take care of most cases. If we still aim the contexts for 0.9.7 then just ignore the patch

Comment on attachment 60326 [details] [diff] [review] workaround hack OK, contexts won't make it to 0.9.7. Suggest a text to be put up on the project page, so we can describe the horkage in detail. Add a XXX wallpaper to the comment, this is not a solution. Other than that, r=me

added the comment in my local tree. I suggest we put something like "Transformiix namespace handling currently has many problems, watch bug 96410 and bug 113611 for progress"

Comment on attachment 60326 [details] [diff] [review] workaround hack sr=jst

checked in, we'll close this once namespacehandling actually works. Pike, do you want this bug now?

Fixed.

we didn't verify for a long time. I really checked, so VERIFIED.