Re: [xml] using "raise" in Python error handler callbacks

[Daniel Veillard <veillard redhat com>]

On Wed, Apr 06, 2005 at 03:10:30PM -0400, Brad Clements wrote:
> On 5 Apr 2005 at 16:48, Daniel Veillard wrote:
>
> >   and what happens when 2 threads use libxml2 python bindings. Parsing of A
> > fails because parsing of B generated and error. No this is more complex
> > than that, really.
>
> Somewhat unrelated to error handler callbacks..
>
> My code uses this code when working with libxslt in Zope:
>  
>   libxml2.setEntityLoader(_Resolver) # this is a process-wide change, 
>
> and wish it were per-thread or even per-document instead of being global.

  The callback gives the parser context, usually it is sufficient to
differenciate what document is being loaded. Might be a bit trickier at
the python level but still possible.

> But anyway, with respect to the entityloader function itself. If an entityloader generates 
> a python exception, libxml2 doesn't propogate that error back through the system to 
> "cancel" the current conversion or parsing process.
>
> The last time I looked at the source, libxml2 considered an exception to be the same as 
> returning "None", which means "I can't handle this, use the default loader".
>
> When looking at handling python exceptions, it would be nice to try to figure out how to 
> propogate the exception condition back-up through the calling chain so that 
>
>     doc = libxml2.parseDoc(srcXML)
>
> or
>
>     result = style.applyStylesheet(doc,stylesheetArgs) 
>
> exposes the exception created in the EntityLoader callback..
>
> I tried looking through the source last year, but it's a bit more complicated than I had 
> first imagined, so I gave up. :-(

  clearly it's not trivial.

Daniel

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