RelaxNG support on nALFS?

James Robertson jwrober at linuxfromscratch.org
Tue Nov 30 13:42:00 PST 2004


Matthew Burgess wrote:
> James Robertson wrote:
> 
>> Matthew Burgess wrote:
>>
>>> Well, depends on your definition of "easy" I suppose :)  Bear in mind 
>>> that (I think) we'd like a syntax that those not proficient with the 
>>> various XML-related specs/technologies can easily adapt for their own 
>>> uses.  Sure, given enough example profiles, customising xpointer 
>>> syntax may be entirely suitable.  I was thinking of something like a 
>>> top-level <alfsCollection> tag or something, that would then allow 
>>> one or more <alfs> children underneath it.  Like your change, it 
>>> would only require a change to the top-level LFS.xml profile.
>>
>>
>>
>> That would be a DTD change?  Is there a way to get nALFS to rad the 
>> profile in and "figure out" what method was used to then allow 
>> validation?
> 
> 
> I was assuming DTD/schema changes would be permissible for nALFS2, hence 
> the suggestion.  If that's not possible then neither is validation IMO. 
>  An XML document is either entirely valid (proved via validation), or 
> you hope for the best by not doing any validation and accepting anything 
> that is thrown at you.  Although there is some value in implementing 
> something like "Validate everything except LFS.xml" I doubt it's 
> expected behaviour from a validating parser.
> 
> Regards,
> 
> Matt.

Oh sure, DTD changes are definitely in for the new tool.  We just need 
to come up with suggested changes and get them in BZ.  See what I am 
doing here:

http://wiki.linuxfromscratch.org/index.php?pagename=nALFS2Develpment

James



More information about the alfs-discuss mailing list