Re: hang in gconfd

[Michael Meeks <michael ximian com>]

Well,

	Everyone likes to moan about how terrible ref counting is for
lifecycle problems - but it seems that this is somewhat analagous.

	If you want to be sure the data arrives, you get to block.

	If you don't care - then just drop it on the floor.

	If you have a remote application that is blocking indefinately -
what do you do ? what can you do ? time it out ? but then it sits there
forever just blocking, kill it ? but perhaps it was doing some interesting
calculation etc. etc. Ultimately life is non-obvious it seems [ unless
I've missed the obvious ].

On Mon, 10 Dec 2001, Alex Larsson wrote:
> On Mon, 10 Dec 2001, Michael Meeks wrote:
> > Hi Havoc,
> >
> > > I do agree the "oneway method, exit()" sequence breaking is bad as
> > > well. atexit() doesn't un-bad it
> >
> > 	Quite - I don't like atexit; I like shutdown methods :-)
>
> Having the shutdown method block, possibly for an unlimited time sounds
> bad too.

	But at least you can not do it if you don't want to; and you can
pass args like 'dont_block':

void
CORBA_ORB_shutdown (CORBA_ORB            orb,
		    const CORBA_boolean  wait_for_completion,
		    CORBA_Environment   *ev)

	Regards,

		Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot