Re: Object activation for GNOME.

[Phil Dawes <philipd parallax co uk>]

Geoff Brimhall wrote:
> 
> Though this might be opening up a sore, what was decided on the issue
of being able to work directly with a v-table for an object, which is
something that COM gives you if you instantiate an object in-process ?
> 
> The last I had seen it was decided against this since COBRA doesn't
allow such direct access.
> 
> Thanks for any enlightenment.
> 
> Geoff Brimhall
> 

At the time it was decided (rightly I think) to delay these sort of
architectural decisions until after we had a basic working orb.

The corba spec doesn't prohibit 'collation' per-se, it just means you
can't do other things like forwarding, interceptors and some poa states.
I submitted a proposal which used a run time 'veto' system between the
client, server and orb to decide whether object collation should be
allowed. Basically the user sets the appropriate flags whilst generating
idl stubs to enable collation stubs to be produced (in addition to
marshalling stubs), and the mediating orb must be configured at start-up
time to allow collation.

Cheers,

Phil.

-- 
_______________________________________________________________________
 Phil Dawes                               |   My opinions are my own
 WWW:    err.. temporarily non-existant   |   and nothing to do with
 Email:  philipd@parallax.co.uk           |      my employer.