On Wed, 20 Mar 2002, R J Scheuerle Jr wrote: > Costin, > > Could you please send the test, wsdl and/or Request/Reply message.
The code is quite complex and dependent of the application. I'll try to extract a test case. The problem is with objects that are received after the href was used. It works fine for the first level of indirection, but the second level doesn't work. It used to work fine, I just upgraded to beta ( and then the 'real' beta) and it stoped working. I spent the night trying to understand how deserialization works. Maybe I'm not familiar enough with SAX... BTW, can someone explain at least the relation between DeserializationContextImpl.idMap and DCI.LocalIDResolver.idMap ? And how is the fixup supposed to work its way up in the object tree ? Costin > > I thing the problem is related with DeserializerContextImpl.fixups > > By adding some loggings, it seems getFixupDeserializer is called a lot, > but I never saw any calls to setFixupDeserializer. It seems the only > place that could set it is registerElementById, but on all calls > to the method fixups==null, so it never gets called. > > If the fixup deserializer plays any role in the process - I have no > idea. The code is far too spaghetti to understand what's going on, > I'm just hoping adding printlns() will show something. > > This is pretty serious problem IMHO, it affects the objects that are > sent after the href is used. > > > Costin > > > On Tue, 19 Mar 2002 [EMAIL PROTECTED] wrote: > > > > > Is anyone able to deserialize complex objects ( using id/href ) > > with the current axis ? It used to work, but neither beta or > > the nightly seems to be able to deal with that - I get null > > for all the referenced objects. > > > > I'm trying to debug it, but if anyone has any hint it would > > save me a lot of time. I'm adding more debug statements > > in BodyBuilder/DererializationContext/MessageElement/DeserializerImpl - > it > > seems most of the action happens there. > > > > Costin > > > > > > > > > > > > > > >
