[
Lists Home |
Date Index |
Thread Index
]
David Carlisle wrote:
> <x>
> <x select="{data:,x}:y[.='{data:,x}:y']"/>
> <x select="{data:,y}:y[.='{data:,y}:y']"/>
> </x>
>
> Here I have effectively destroyed the prefixes, so my result could only be
>
>You also effectively destroyed the meaning. See I said it wasn't so easy
>you need a full Xpath parser.
>The point of my example was that the x:y in [.='x:y'] is a _string
>literal_ and not a Qname, it shoul dbe no more subject to namespace
>processing than if it had said 'colour: red'.
>
>
>
I was about to say "oops", but what if the content of y is specified by
an XML Schema to be a QName?
But you're right in that I did not take care and replaced inside the
string without switching on my mental XPath parser. my bad.
But I still have the feeling that thinking it for each special case (the
set of triple-x documents above needs some careful rethinking) can yield
sane documents.
> I am not sure if anybody really auto-generates XSLT with clashing
> namespace declaration.
>
>Oh sure it's _very_ common (also hand written. Most common re-usimg
>the default namespace. People often prefer to use xhtaml and mathml say
>both unprefixed and pay teh price of re-declaring mathml every time you
>go in to mathematics.
>
>
>
Oh but that one does not give any problems at all: sanitization just
assigns a prefix to it. Clashing would be, explicitly using the same
prefix for different namespaces (and thus, the need to handle scopes).
And if MathML and XHTML contain QNames, it is better to not touch them
(one can learn from one's mistakes:)
Hell, anything that does not use Qnames in content would work right
away, and that would be pretty much everything if there was not this
incomprehensible urge to use XML for descriptions of XML. There would be
no need for incredibly specialized corner cases and maybe be people
would actually start to use XML on a wide scale.
Maybe a big red sticker "use sanitization when there are no QNames in
content" is quite enough, and one has to give up on those.
Thanks for the example anyway, I shall be more careful next time.
cheers,
Burak
--
Burak Emir
http://lamp.epfl.ch/~buraq
|
