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

[Daniel Veillard <veillard redhat com>]

On Tue, Apr 05, 2005 at 12:40:01PM -0400, Stefan Seefeld wrote:
> Daniel Veillard wrote:
> 
> >>The 'PyErr_Occurred()' function will return 1 since the (python) error
> >>callback raised an exception that was not yet dealt with.
> >>With the above hack, the example of the OP indeed generates a stack trace
> >>(a TypeError as the string in the 'raise' expression is malformed).
> >>
> >>However, things are more complex than that, since xmlTextReaderRead will
> >>potentially call the (python) callback multiple times, erasing the 
> >>exception
> >>state from former calls.
> >>Right now libxml2 seems to assume that (error) callbacks can't fail 
> >>themselfs.
> >>I'm not sure whether there is anything to do here.
> >
> >
> >  So PyErr_Occurred() check should be done when C call python at the,
> >return point, right ? 
> >I think this occurs only in a few places within python/libxml.c which is
> >not autogenerated. So hacking this might be possible.
> 
> Fine ! Though I would use some kind of polymorphism to provide this
> functionality to any language binding, not just python. What about
> setting up some global (but protected) variable (akin to an internal 
> 'errno')
> that callback wrappers may use to indicate failure ?

  I'm objecting to this because global variables then generate the question
of what happens when multiple threads raise errors, etc ... Unless there
is a clear need at the library level I would keep this a Python only 
thing.

> Make this a 'void *' normally set to 'NULL', and then abort the normal
> processing if it is != NULL. Wrapper code may then tunnel any data through 
> it.

  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.

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/