[09:32] <anpok> hi
[09:32] <anpok> according to lttng we need two scheduling steps after reading the input event before it reaches the client
[09:48] <alan_g> FYA http://tgceec.tumblr.com/post/74534916370/results-of-the-grand-c-error-explosion-competition
[09:55] <alan_g> anpok: I'm not sure what your observation means nor whether it is a problem
[09:59] <anpok> i guess it means nothing,
[10:01] <anpok> i simply observed asio doing the write and the completion handler in a different thread
[12:00] <alan_g> greyback: the RPC stuff you needed is on development-branch. These tests should show how to hook it up - let me know if you need anything more. http://bazaar.launchpad.net/~mir-team/mir/development-branch/view/head:/tests/acceptance-tests/test_protobuf.cpp
[12:00] <greyback> alan_g: cool, thank you
[12:01] <alan_g> You're welcome
[17:56] <alf_> kdub: question about hwc-layerlist-improvements, could oid mga::LayerList::set_fb_target() also set the fb_target_present property, or does it make sense for fb_target_present to be true even without a set fb target?
[17:56] <kdub> alf_, fb_target_present is really 'whether to tack on a FB layer at the end' or not
[17:57] <kdub> so it won't change
[17:58] <alf_> kdub: does it make sense to have a layer without having called set_fb_target()?
[17:59] <kdub> yes, for hwc 1.0, we have layers without fb targets
[18:01] <alf_> kdub: I mean, does it make sens to have an fb_target_layer without an actual fb target set? What I am getting at is if we could remove the has_target_layer constructor argument and rely only on whether set_fb_target has been called, but I don't understand the details very well.
[18:02] <kdub> alf_, i think thats a decent improvement, it at least eliminates the funny constructor argument