Re: [xml] [Fwd: [xml-libxml] dangling doc setfaults]

[Daniel Veillard <veillard redhat com>]

On Fri, May 09, 2003 at 12:19:59PM -0400, Bruce Miller wrote:
> Daniel Veillard wrote:
> > On Fri, May 09, 2003 at 10:26:35AM -0400, Bruce Miller wrote:
> [...]
> > > But I wanted to repost this here, in case a similar bug is lurking
> > > in xmlParseBalancedChunkMemory that Daniel might recognize, or whether
> > > it suggests a way that XML::LibXML should be dealing with it.
> >
> >   I think some things have been fixed there too since the early 2.5.x
> > releases, but without a reproductible test case at the libxml2/xmllint
> > level it might be difficult for me to really see and check.
>
> I'm running 2.5.7, now, but I probably should bite the bullet & run the CVS;
> I've put that off, since there's so many dependences on libxml2's rpm
> in RH9. (But I guess that just reflects the usefulness of it, doesn't it! :>)

   cvs update, make rpm and rpm -U --force with the newly built ones.

> As for test cases: I'll do what I can. I certainly appreciate your preference
> for a finite C case. But since it seems (?) to be pointer related, it
> takes all sorts of permutations & head-standing to get it to happen reliable
> in perl (and the permutations are easier to generate in perl).  Of course,
> if i were to translate it to C (if I could!), it wouldn't fail the same way!
> At any rate, probably the best I can do will be something like "Don't we
> need a <foo> here?" rather than an actual, useful patch.... Apologies in advance!!

  Well python tests would be okay too... but understood it's not trivial.

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/