On Mon August 17 2009 1:46:03 pm Eamonn Dwyer wrote: > Hi > I may be missin git but I can't find documentation around the use of a > CachedOutputStreamCallback so I would like some clarification on its use. > From debugging, it looks like it would typically be used to as a mechanism > to do some last bits of work just before an invocation is made on the > service. Is that a valid understanding? That seems to be the case for SOAP > over HTTP (at least) but doesn't seem to be the case for SOAP/CORBA. I'm > wondering if its a bug in the CORBA implementation or just my > misunderstanding of it.
Basically, it allows the stream to really know "at the very last possible moment" that it can modify things like headers and such. That's what http does with it. The one in CORBA is interesting and I didn't even know it did this. Basically, there are a couple interceptors and such that assume an OutputStream exists. For CORBA, the OutputStream would never be written to. It's mostly just there to satisfy those couple of interceptors as well as get the "close" event that would trigger the request/response. Ideally, this would get completely eliminated. The interceptors that require and OutputStream would be updated to be smarter and the "close" trigger could also trigger off the XMLOutputStream.close() which the CORBA version of the XMLOutputStream could implement properly. -- Daniel Kulp [email protected] http://www.dankulp.com/blog
