Re: Asynchronous procedure call with ORBit.

[Mark Wigzell <markwigzell home com>]

With regards to COMM_FAILURE being thrown:
Shouldn't the ORB distinguish between a failure to reach object (TRANSIENT)
and session falure (COMM_FAILURE)?
If you have your IOR in a file, and the server is down, and you run a client
from it you get COMM_FAILURE. 
BUT: what if it was a temporary inability to reach the  server?(TRANSIENT).
Afterall, the IOR file is still there, so no clean up has been
done yet, maybe the server is busy?.
... Mark

> Message: 2
> Date: Mon, 19 Nov 2001 17:22:44 -0500 (EST)
> From: Michael Meeks <michael@trna.ximian.com>
> To: orbit-list <orbit-list@gnome.org>
> Subject: Re: Asynchronous procedure call with ORBit.
> Reply-To: orbit-list@gnome.org
> 
> 
> On 18 Nov 2001, Diego González wrote:
> > > oneway void notify (...)  :-)
> > 
> > if you use oneway calls nobody can ensure that the remote object in
> > which you are executing that method has died, and if you execute the
> > method and the object has died you also receive no exeption telling    
> > you that.
>   
>         Not quite the case in fact. It is true, that you are never sure
> that the method has fully completed successfully. However - if the
> connection has died, any subsequent oneway invocations on that object will
> result in a comm failure exception being thrown.
> 
>         Regards,
>    
>                 Michael.
> 
> -- 
>  mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot