On Tue, Apr 17, 2007 at 07:18:10AM -0500, David Hagood wrote:
> I am having a problem with RelaxNG validation in streaming mode.
[...]
> Now, IF I tell the handler for <node> to expand the node, then all is
> well - however then I am no longer using streaming, and for my needs I
> don't think that will work.
> 
> I *think* what is happening is that the reader, upon reaching the end of
> the <first_subnode>, is deleting the node, but retaining the enclosing
> parent node. Then, when the end of <node> is hit, it calls the
> validator, which then fails the node. IF I am correct, then what needs
> to happen is that IF the reader is doing RelaxNG parsing, it needs to
> save the children nodes of any open node, OR it needs to validate on the
> fly and somehow record what it has seen.
> 
> Does this seem reasonable?

  Impossible to tell without full input and RNG. Depending on RNG it may be
compiled to a regexp or not. Then processing is completely different.
Making guesses is a very time consuming process, which I usually avoid
when I can.
  Can you reproduce with xmllint --stream --realaxng .. ... ?

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
[EMAIL PROTECTED]  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/
_______________________________________________
xml mailing list, project page  http://xmlsoft.org/
[email protected]
http://mail.gnome.org/mailman/listinfo/xml

Reply via email to