On Wed, May 14, 2008 at 5:46 PM, Jeffrey Hutzelman <[EMAIL PROTECTED]> wrote: > --On Wednesday, May 14, 2008 05:36:42 PM -0400 Matt Benjamin > <[EMAIL PROTECTED]> wrote: > > > > Create AFSExtendedCallbackV2, which aggregates an AFSExtendedCallbackV1, > > and then adds new members. Create ExtendedCallbackV2 taking > > AFSExtendedCallbackV2 as argument. > > > > Oh, you don't need to do that. You _can_ gracefully extend the union, as > long as you are willing to accept that calls made using a discriminator not > known to the server you are calling will abort with RXGEN_SS_UNMARSHALL. > This is the moral equivalent of a call to an RPC not supported by the server > aborting with RXGEN_OPCODE, and is perfectly acceptable, given that you > shouldn't be sending notifications with unknown types anyway. >
one presumes new operation types would be the cause of new callback types, and there'd be a capability bit anyway. _______________________________________________ AFS3-standardization mailing list [email protected] http://michigan-openafs-lists.central.org/mailman/listinfo/afs3-standardization
