[PATCH] XML validity of kcfg files

[Frans Englich]

On Saturday 20 November 2004 01:15, Benjamin Meyer wrote:
> On Wednesday 03 November 2004 1:19 am, Frans Englich wrote:
> > On Wednesday 03 November 2004 02:22, Cornelius Schumacher wrote:
> > > On Wednesday 03 November 2004 01:24, Frans Englich wrote:
<snip>

> So what is the proper way to validate the xml files?

These lines:

      xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
      xsi:schemaLocation="http://www.kde.org/standards/kcfg/1.0
                          http://www.kde.org/standards/kcfg/1.0/kcfg.xsd"

says "the prefix xsi refers to the XML application XMLSchema-instance" by 
identifying it with the 'http://www.w3.org/2001/XMLSchema-instance' 
namespace, and then it binds a WXS(W3C XML Schema) to that namespace with 
schemaLocation, such that the validator can figure out the schema for a 
certain namespace. It's possible to specify multiple schemas by putting pairs 
of namespaces+WXS in schemaLocation. The validator may choose to ignore what 
schemaLocation says, and fetch the Schema from some other place.

But. libxml2 does not yet support XMLSchema-instance. In other words, it can't 
figure out the Schema from the document instance, so it must be explicitly 
specified from the outside. Typically, like this:

xmllint --noout --schema http://www.kde.org/standards/kcfg/1.0/kcfg.xsd 
foo.kcfg

(And the kcfg file must have a proper declaration, as the docs in kcfg.xsd 
explains, and HEAD:kdepim/korganizer/korganizer.kcfg exemplifies.)

There's a bug report for Schema Instance support in xmllint, and there's 
preliminary support for it, so I wouldn't be surprised if it turned up in the 
next release.

That you get a complain about missing DTD is probably because you passed  
--valid(and that's for XML validness as opposed to Schema validness). 


As for how good xmllint is as a validator, to reflect on Cornelis' comment; 
xmllint can certainly be improved, but I don't know if there's any other 
which are particularly better(not that WXS validators exactly grows on 
threes). It have bad usability: the error messages don't identify 
what's wrong, but I've filed bug reports like a monkey for that, so it will 
be fixed(and in some cases is fixed). I've also found small quirks in its 
compliance, but there BRs for that(or fixed). When those small annoyances are 
ironed out, I don't know what to ask for, especially considering libxml2's 
speed.	

(I learn when I try to retell.)


Cheers,

		Frans