lfsml-6

Matthew Burgess ca9mbu at eos.sunderland.ac.uk
Sat Jan 18 11:32:19 PST 2003


"Matthew Burgess" <ca9mbu at eos.sunderland.ac.uk> wrote in message
news:b09i62$k9v$1 at quasar.highos.com...
> Gerard,
>
> I've finally sussed it, after trying about 5 different XML parsers,
all of
> which show the same behaviour (or don't support XInclude at all).
>
> The only reason xmllint (and others) don't perform the XInclude
instructions
> without the DTD declaration present in the child document is because
of the
> &ddash; entity references.  I don't know why the processors can't
"cache"
> the DTD from the parent document and validate the child document
against
> that if it doesn't have a DTD declaration of it's own - it might be in
the
> spec but to be honest I don't have time to wade through it.  So,
unless the
> processor can validate the child document entirely separately from the
> parent then it won't be included.
>
> Hope this helps,
>
> Matt.

And if I'd have used xsltproc instead of xmllint I would have saved
myself a whole lot of time.  Unlike xmllint which just refuses to show
the results of an XInclude (successful or otherwise) xsltproc actually
gives very useful error messages - I would have thought this should be
the other way around, with the lint tool showing errors more verbosely
than any other tool.

Here ends the biggest waste of time ever!

Matt

PS: Now I know how Ian Molton feels (on lfs-chat) - off in my own little
world, doing my own thing, and reporting back results whether or not
anyone cares.  Ah well.


-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-book' in the subject header of the message



More information about the lfs-book mailing list