Don't talk about *the* subclass converter for A, there can be any number of them. For example, the qt, qtcanvas, qtext, qtgl, qtnetwork, qtsql and qttable all provide converters for QObject.
Ah, *now* I get it. (Did I manage to overlook the documentation on this, or is it one of those moving targets?)
It's not obvious from the SIP files -- for example, in qcanvas.sip, it looks like it's defining a subclass converter just for stuff in QCanvas and below. But in the generated file, it's clearly overlaying (? is there a better word to describe this?) the QObject converter from qobject.sip.
This aliasing would happen, though, for any converter which relies on dynamic_cast<>() (and some other kits which have their own RTTI).
The source of the type information is irrelevant. I agree that using dynamic_cast<>() would potentially mean you had overlapping converters - but that's just a bug in the converter.
Yep... I now understand the grand vision of how it all works. Thanks!
Dave
_______________________________________________ PyKDE mailing list [EMAIL PROTECTED] http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
