On Wed, Dec 10, 2014 at 5:10 PM, Andrew Stitcher <[email protected]> wrote:
> An off the cuff thought - worth what you paid for it - > > On Wed, 2014-12-10 at 11:23 -0500, Ken Giusti wrote: > > ... > > Furthermore, I think the event API would benefit from a way to 'opt-in' > to specific events. For example, for qpidd we would not want to receive > PN_TRANSPORT nor PN_LINK_LOCAL_FLOW events. > > > > As far as I can see it should be hardly more expensive to do this > filtering in the listening code than in the proton code itself. So I'm a > little suspicious that the large number of ignored events is causing the > problem. If they are causing a lot of app work to happen then perhaps > they can be filtered out closer to reception. > +1 > I think it would be difficult to resolve filtering the events out in > proton, with allowing multiple independent and ignorant of each other > layers that can coexist. > We could probably get something to work, i.e. essentially build a tiny little pub/sub model for events so we only produce them if someone is listening for them, but there's no guarantee that would be faster than just ignoring the events you don't care about. --Rafael
