But that's an extension of SerializationContextImpl, not a new implementation of SerializationContext. So this would still work fine if the interface/impl split went away.
I also agree there are other ways to do this, and in fact that's something I was going to mention re: your patches at some point soon. :) --Glen > -----Original Message----- > From: Thomas Sandholm [mailto:[EMAIL PROTECTED]] > Sent: Thursday, May 09, 2002 3:02 PM > To: [EMAIL PROTECTED]; 'Axis-Dev (E-mail)' > Subject: Re: SerializationContext/DeserializationContext cleanup? > > > One example is the AttributeSerializationContextImpl patch > that I added to > do correct serialization of simple types into XML Schema attributes > (previously only toString() was called -which didn't work for > Calendar and > Qname types for example). I'm sure there are other ways to > implement this > feature like factoring out the start, and end element calls from > serialize() for simple type serializers. But it is something > to be aware of > if you are changing the model. > /Thomas > At 02:51 PM 5/9/2002 -0400, Glen Daniels wrote: > > >Does anyone (Rich?) have any real reasons why we have > SerializationContext > >and SerializationContextImpl? I know the JAX-RPC > SerializationContext is > >an interface we have to implement, but it seems to me we > could just as > >easily (and much more conveniently) have > >axis.encoding.SerializationContext be a class instead of > introducing yet > >another interface. > > > >Same holds for DeserializationContext/DeserializationContextImpl. > > > >Do you think these are useful + important points of > extensibility? Are > >there really going to be other implementations? > > > >If we could coalesce these guys, that would make my life a > happier place. > > > >--Glen > > Thomas Sandholm <[EMAIL PROTECTED]> > The Globus Project(tm) <http://www.globus.org> > Ph: 630-252-1682, Fax: 630-252-1997 > Argonne National Laboratory >
