Re: [liblo] Callback handler and bundles

On Wed, 2010-02-24 at 17:49 -0500, Stephen Sinclair wrote:
> On Fri, Feb 19, 2010 at 4:55 PM, David Robillard <da...@dr...> wrote:
> > On Thu, 2010-02-18 at 13:21 -0500, Stephen Sinclair wrote:
> >> On Tue, Feb 16, 2010 at 6:41 PM, David Robillard <da...@dr...> wrote:
> >> > P.S. It would be also possible for apps that use lo_server_recv and
> >> > friends directly to know about bundles if it only dispatched one message
> >> > at a time and lo_server_events_pending worked as advertised, but the
> >> > callback is more useful and generic anyway
> >>
> >> By the way, I don't actually this is an appropriate use of
> >> events_pending(), but in what way is it broken?
> >
> > I was expecting each individual message of a bundle to be dispatched
> > independently, thus events_pending() could be used to check if you're in
> > a bundle.  Unfortunately the entire bundle is dispatched at once so this
> > is not the case.
> >
> > A proper callback is better anyway, so whatever
> 
> Right.  Technically they aren't "pending" until liblo sees them, and
> since it goes through the bundle one at a time, it hasn't seen the
> rest of the messages in the bundle yet so they aren't yet pending..
> 
> I wonder if this should be changed/fixed?
> 
> Perhaps not, because I think events_pending() should be used mostly to
> check if there are _future_ messages, not more _now_ messages.

It's probably best to leave it as-is, it's not the right way to check
for more _now_ messages anyway.

Attached is a revised bundle handlers patch which uses separate
callbacks and passes the timetag to the start handler.

Cheers,

-dr