Re: [xml] bug 362989 -- still problem

[Daniel Veillard <veillard redhat com>]

On Wed, Nov 01, 2006 at 04:06:13PM -0800, Yong Chen (yongche) wrote:
> I'll do more debugging (including trimming down the size of b.xsd) to
> see if I can provide more info. Also did you mean if I fix the
> "non-deterministic" issue in xsd file, I'll not run into this problem?

  The non-determinism comes from the fact that all your blocks have
minOccurs = 0, which means that there is epsilon transition from any
state to any other state forward. I think the processing time is due to
the elimimation of this mass of e-transition into an even bigger mass of
transitions labelled by the element name. Solving the non-determinism
by imposing elements structure at least between some blocks would also
solve the processing cost. Solving it by renaming element and not
touching the structure won't solve it.

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/