Found the bug. Turned out to have nothing to do with the validation code. It
was simply a bug in the code that builds the in-memory representation of the
pattern.
James
> Thank you.
>
> I'm working on this now. I'd hoped this would be same bug as
> James Clark
> pointed out but it isn't.
>
> James
>
> > -----Original Message-----
> > From: Murata Makoto [mailto:]
> > Sent: Thursday, March 29, 2001 8:17 AM
> > To: ''
> > Subject: Re: PyTREX feature complete
> >
> >
> > James Tauber wrote:
> >
> > > I've just release PyTREX 0.7.0 at
> > >
> > > http://sourceforge.net/projects/pytrex/
> > >
> > > PyTREX now implements the entire TREX specification. There
> > are a couple of
> > > known bugs and no anonymous types are supported but
> > everything in the spec
> > > is implemented.
> >
> > My bug report is available at:
> >
> > http://sourceforge.net/tracker/index.php?func=detail&aid=41213
> > 0&group_id=21330&atid=371829
> >
> > Content models of TREX handle both elements and attributes,
> > thus representing
> > interdepencies between attributes and elements very well.
> > However, implementation
> > is not as easy as it looks like. Although I am sure that
> > James Clark's implementation
> > is correct, I would like to see correct implementations based
> > on different algorithms.
> >
> >
> > Cheers,
> >
> > Makoto
> >
>
