[certi-dev] Java Binding

[Eric Noulard]

Hi Andrej,

As you may have noticed I finally reached the point were we can
totally rely on CERTI_Message.msg specification for RTIA<-->Federate messages.

Could you tell me if you did work on the Java generator backend?

We may work together on this in order to makes it easier your work.
There is now many difference on the message spec' you generated
and the message spec I did check-in, mostly because I did prune
some messages that were conveying too much information.

We currently have one discrepancy regarding the Attributes/Parameters
sequences in the CERTI messages. I did use sequence and not directly
the RTI provided interface "AttributeHandleSet", "AttributeHandleValuePairSet"
etc...

I did wrote some conversion functions from the RTI interface to CERTI
internal interface. I did this mostly because since we plan to support 1516
we would have to do conversion anyway because 1516 defines other types
for the very same items. So if we won't to support a multi-standard RTIA/RTIG
we better separate the RTI specific types from CERTI internal types.

Try to evaluate if this generates a "big" change for you and if you do
not have time to go for it I may try to help you with that.

Whatever is your answer I did tag in order to clearly identify the compatibility
breakage in the CERTI CVS tree.

I will now work on some pending patches and go back to enhance
the CERTI message generator in order to at least modularize the language
generator backend in such a way that we can write a single file for each
language generator backend.


-- 
Erk
Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org