On Fri, 16 Sep 2005, Brian Lindahl wrote:
The segfault was due to a really obscure bug that I really don't have time to investigate - it involved wrapping initialization of the orb, and the mimic object in one try block and the "work_pending, perform work" in another try block. I didn't bother to investigate further as a work around was just to wrap the entire thing in one try block.
It looks like the problem was in your code. Have you reviewed your _var type variables usage?
Anyhow, it turns out my scheme won't work because in order for the ServerRequestInterceptor::receive_request to be called, you have to parse the arguments in the DynamicImplementation::invoke method, first.
True, you definitelly need to use IR in case you're going to proxify using DSI/DII.
As such, I've had to abandon this solution for a hard-coded non-generic solution (the other option being implementing TII or a GIOP server that goes low level to forward requests to different objects - both of which would involve too much time).
And do you really need TII? Isn't siple callback-based AMI enough for your application? As I said, this shouldn't be that hard to implement... or you can even use DII and poll for replies periodicaly (kind of what AMI does)
On a side note, apparently 'ForwardRequest' exceptions are not caught during invocation, but are only caught during Interceptor, ServantManager and ServantLocator upcalls.
True, I've just checked the spec and it seems MICO is right here. Cheers, Karel -- Karel Gardas [EMAIL PROTECTED] ObjectSecurity Ltd. http://www.objectsecurity.com --- Need experienced, fast, reliable technical MICO support? ---> http://www.objectsecurity.com/commsup.html <--- --- _______________________________________________ Mico-devel mailing list [email protected] http://www.mico.org/mailman/listinfo/mico-devel
