Removing this call on close() appears to have no detrimental effects =
there must have been some reason I added it between Saxon 8.6.1 and =
8.7, so I'm not totally convinced the situation is resolved.
The other unsatisfactory thing about the resolution is that I can't see =
the error isn't causing an awful lot of my tests to fail. But it's not =
neither is the fix) so we'll just have to run with it.
Saxon 8.7.1 should be out before Easter.
> -----Original Message-----
> From: Michael Kay [mailto:mike@...
> Sent: 06 April 2006 13:39
> To: 'saxon-help@...'
> Subject: RE: [saxon] Different behaviour between 8.7b and=20
> 8.7sa when copying node in a namespace from variable
> There's clearly a bug here and I'm working on it. The problem=20
> is caused by the fact that copy-of does a close() on the=20
> output pipeline after it completes (CopyOf lines 192-194) -=20
> this is happening in SA and not in B because copy-of adds a=20
> StripValidator to the pipeline to remove any type=20
> annotations. The close() is propagating through to the tree=20
> builder, which inserts a Stopper node on the tree, and=20
> although subsequent nodes are added to the tree, they aren't=20
> seen when it comes to reading the tree.=20
> I'm now trying to work out the consequence of removing this close()...
> Michael Kay
> > -----Original Message-----
> > From: saxon-help-admin@...
> > [mailto:saxon-help-admin@...] On Behalf Of=20
> > andrew welch
> > Sent: 06 April 2006 12:14
> > To: saxon-help@...
> > Subject: Re: [saxon] Different behaviour between 8.7b and=20
> > 8.7sa when copying node in a namespace from variable
> > On 4/6/06, Trevor Nash <tcn@...> wrote:
> > > On Thu, 6 Apr 2006 10:04:46 +0100, "andrew welch"
> > > <andrew.j.welch@...> wrote:
> > >
> > > >With Saxon 8.7sa it appears to stop at the copy-of in the=20
> > <a> matching
> > > >template, but oddly still produces well-formed output:
> > >
> > > Que?? The div elements are copied by the "*" match, not=20
> the "a[..]"
> > > match. The copy-of in the <a> template only copies the variable
> > > containing the <a> element. It doesn't do anything with=20
> > the original
> > > a's siblings.
> > Que indeed. I didn't say anything to the contrary so I'm not sure
> > what your point is here.
> > > Does it work if you eliminate the superfluous 'link'=20
> variable? How
> > > about adding something (white space or a comment say) after the
> > > copy-of so it isn't the last thing in the template?
> > Yes it does work without the "superflous" variable but thats not the
> > point - its only superflous in this cut down example (I'm sure you
> > were aware of that, but I'll state the obvious anyway). And no,
> > adding comments or some content after link doesnt help (feel free to
> > try the samples)
> > > [I take it as read that this is an extract of a real=20
> > stylesheet where
> > > 'link' serves some useful purpose]
> > er, yes.
> > -------------------------------------------------------
> > This SF.Net email is sponsored by xPML, a groundbreaking=20
> > scripting language
> > that extends applications into web and mobile media. Attend=20
> > the live webcast
> > and join the prime developer group breaking into this new=20
> > coding territory!
> > =
> > _______________________________________________
> > saxon-help mailing list
> > saxon-help@...
> > https://lists.sourceforge.net/lists/listinfo/saxon-help