[PATCH] XML validity of kcfg files

[Frans Englich]

On Wednesday 03 November 2004 00:23, David Faure wrote:
> On Wednesday 03 November 2004 01:24, Frans Englich wrote:
> > Anyway, attached is an XML Schema version for the KConfigXT format, most
> > of it is written but could use some polishing, and tighter validation can
> > also be done. If someone wants to do anything with it, or in the future
> > will do fancy stuff, it is here in the archives(I'll not push it).
>
> Thanks, this is the right solution. IMHO it should replace the DTD.

Christian had some concerns with this(and I didn't follow up because I thought 
it would be irrelevant):

On Saturday 30 October 2004 20:48, Christian Mueller wrote:
> Changing from DTDs to XSDs would mean changing the prolog of all the kcfg
> files. 

We wouldn't remove the DTD(that would be more or less equivalent to breaking 
source compatibility), but those who are interested in ensuring their kcfg 
files doesn't contain any latent bugs, could switch to use the Schema. In 
other words, it would be a gradual migration for those who cares.


Now if people want to use it, it might be worth to mention it doesn't solve 
the order restriction...

I've asked about it one the xml-schema-dev list(w3c) and that's what they told 
me, I asked why it is so but haven't received a reply(yet). The exact problem 
is of combining maxOccurs="0" with the all indicator(perhaps the Schema 
becomes ambiguous, dunno). There were work-arounds, but it involved switching 
to RELAX NG or doing a XSLT pass inbetween etc -- overkill. 

Well, I think I'll contradict what I said before: it isn't that hard to ensure 
those two/three elements are in a certain order(get over it), it's done once 
and then other significant errors can be ironed out.


Cheers,

		Frans