[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

Re: [xsl] Excluding Namespaces from the Result Tree


Subject: Re: [xsl] Excluding Namespaces from the Result Tree
From: "Heiko Niemann kontakt@xxxxxxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Tue, 30 Jun 2015 10:43:57 -0000

> Heiko Niemann kontakt@xxxxxxxxxxxxxxxx wrote:
>
>> using XSLT 2.0 there is the value #all for the attribute
>> exclude-result-prefixes. I try to find cases where you would NOT want to
>> use it, in other words why would you want to keep namespaces in the
>> result
>> tree that are not used there? Which leads to the question why cleaning
>> up
>> namespaces is not a default behavior.
>
> With XML ou can use qualified names not only as names for nodes but also
> in element or attribute values, for instance if you want to construct an
> XSD schema with XSLT, and in that case if you want for instance to have
>
>    <xs:element ref="pf1:foo"/>
>
> in your result then, even if the prefix 'pf1' is not used in element or
> attribute names, you want to preserve its declaration.


Thanks Martin,

now it came back to my mind. Some years ago I actually had to deal with
cleaning up namespaces but keeping the ones that were used in qualified
names as values in elements or attributes - which caused lots of
frustration as I remember. I still think that this is a confusing aspect
of XSLT/XML that there is not a real connection of parts in one context
(XSLT) that are meant to be connected in a different context (XSD schema).
So one pitfall could be using #all with exclude-result-prefixes another
one using copy-namespaces="no" within copy-of.

I do understand that for XSLT element or attribute values are just plain
values/strings although they might be qualified names. Would a
schema-aware processor solve this problem? I have never used one...

Heiko


Current Thread
Keywords