[Phoenix] Re: Bug in source node retrieval

[Comiotto Thomas]

Hi Andi

> Hmm, well, I'm a bit wary about this. Because e.g. Atom entries are 
> mostly like this (and falling back to namespace-unaware mode is no 
> option, if I understand that mode correctly, because Atom entries can 
> contain mixed namespaces, e.g. the can contain XHTML markup inside 
> <atom:content> elements).

Should be no problem because, ehm.. ns unaware mode means ignoring 
namespaces. The idea is that only local names of elements are taken 
into account when assembling the xpath expression (*[local-name() = 
'content']/*[local-name() = 'p']/) - so mixing namespaces is no 
problem. The only problem one can run into with this approach is in 
situations where the xpath expression becomes ambiguous when ignoring 
ns information, i.e. a document that has

foo:table/foo:tr/xhtml:td/text()

xhtml:table/xhtml:tr/foo:td/text()

Never came across this, but it can't be ruled out either of course;)

--
Bests
Thomas



>
> So I better go and implement this... ;)
>
> -- 
> Kind regards,
> Andi
>