bonobo-activation-server does not tolerate crashing factories

[ERDI Gergo <cactus cactus rulez org>]

If you request an object and its factory is started by
bonobo-activation-server, the server does not later check if the factory
is still working. So if it crashes and you request another object of the
same type, instead of starting the factory again, the request simply
hangs.  To get things back to normal again, you have to kill the
activation server and remove 
/tmp/orbit-$USER/reg.IDL:Bonobo_ActivationContext:1.0*.

The problem with this is that this obviously sucks. In fact, I wonder if
the activation server doesn't get a CORBA::COMM_FAILURE exception when
calling Bonobo::GenericFactory::createObject on the factory, which could
be caught and a new factory be started instead.

-- 
   .--= ULLA! =---------------------.   `We are not here to give users what
   \     http://cactus.rulez.org     \   they want'  -- RMS, at GUADEC 2001
    `---= cactus cactus rulez org =---'
Nem vagyok sznob. Kérdezz meg bárkit. Úgy értem, bárkit, aki számít.