From: Daniel Veillard (Daniel.Veillard@w3.org)
Date: Tue Aug 22 2000 - 16:22:37 EDT
On Tue, Aug 22, 2000 at 10:59:39AM -0400, Aaron M. Ucko wrote:
> Daniel Veillard <Daniel.Veillard@w3.org> writes:
>
> > My regression tests of libxml have had & cases since the very
> > beginning. Like test/ent3
> >
> > <EXAMPLE prop1="a&b" prop2="&xml;">
> >
> > If libxml was choking on them I think I would have detected it
> > immediately, before doing the packaging.
> > Do you have the testcase exhibiting the problem ?
>
> Compare the output of
>
> xmllint --valid test/ent3
>
> to that of
>
> sed -e 's/&/\&/g' test/ent3 > ent3b
> xmllint --valid ent3b
>
> > Are you using the SAX interface or the DOM one ?
>
> I'm using the DOM-style interface, but with validation enabled.
Woah, this is a real bug, impressive that it never showed before,
seems that all the document i used to test the validations, none
uses & in attribute values !!!
I fixed the problem (your suggested fix is invalid it doesn't
deliver the proper attribute value at the end and expose a number
of other problems). I enclose the patch, I will commit it to the CVS
bases in the next hour.
thanks a lot for catching this, and apologies to the person
who reported it previously it seems I didn't look at the problem.
Daniel
-- Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes | Today's Bookmarks : Tel : +33 476 615 257 | 655, avenue de l'Europe | Linux XML libxml WWW Fax : +33 476 615 207 | 38330 Montbonnot FRANCE | Gnome rpm2html rpmfind http://www.w3.org/People/all#veillard%40w3.org | RPM badminton Kaffe ---- Message from the list xml@xmlsoft.org Archived at : http://xmlsoft.org/messages/ to unsubscribe: echo "unsubscribe xml" | mail majordomo@xmlsoft.org
This archive was generated by hypermail 2b29 : Tue Aug 22 2000 - 13:43:11 EDT