Re: XHTML validation bug (false pass)

On 06.03.00 at 10:54, David Brownell <david-b@pacbell.net> wrote:

>Gerald Oskoboiny wrote:
>> 
>>That doesn't seem good, indeed. I thought SP's XML limitations sounded
>>like obscure things that wouldn't come up much in practice, but that
>>doesn't seem to be the case.
>
>Nope.  I'd actually make that clearer in the warning that you're now
>providing -- the little footnote does make it sound like they're quite
>obscure, but it should become explicit that the consequence is that it may
>accept documents that any XML tool _must_ reject completely.

Actually, I've had several people with pretty good knowledge of XML say
that SP's limitations are truely obscure and should appear next to never in
the wild. I'm beginning to wonder if this is actually what's going on or if
there is something else involved here.

Whatever it is it needs to be dealt with, of course, but I'm not entirely
convinced it's SP's XML limitations we're butting heads with.


>>I guess I'll have to have a look at expat or something else as a
>>replacement.
>
>Expat doesn't validate.  Neither does it normally read external entities,
>though that's tweakable (as I understand).

I have "looking at Perl XML modules" on my TODO list on the assumption that
I can reimplement SP's XML support through one of those. The major sticking
point seems to be that noen of them will handle SGML correctly and so we'd
end up with SP for SGML and XML::Parser (or somesuch) for XML. 


>A longer term fix is to use a real validating XML parser.  The ones I'd
>have any reason to trust lately are all written in Java, though some C/C++
>ones are starting to crop up.

A longer term fix may actually be to replace SP with OpenSP from the
OpenJade project. AFAIK, they've more or less taken over SP and are the
only ones implementing new features. I haven't looked at it in any detail
yet, but there seems to be further developemnt going on there. Maybe they
could be talked into fixing SP's XML support.

Received on Monday, 6 March 2000 15:20:43 UTC