[00:00] robert_ancell: who's been dealing with the packaging in lp:mir? [00:01] thomi, a mix of people, but I can help out with that [00:03] robert_ancell: I'm trying to add the mir_stress binary to a new mir-stress package, but for some reason it's not adding the binary to the package [00:03] I wonder if you have a second to look? [00:03] thomi, sure [00:03] lp:~thomir/mir/add-mir-stress [00:11] thomi, hmm, it looks ok to me [00:11] you mean, you think it should work, or it actually does work for you? [00:12] Greetings... [00:12] Hey why doesn't Gnome clock in raring know _any_ NZ timezone/locations? :) [00:12] thomi, I haven't compiled it but will try now [00:12] duflu, works for me.. [00:13] duflu: me too, in saucy too :-/ [00:14] duflu, time to upgrade to saucy anyway :) [00:14] It also reckons _every_ town in West Aust is called Perth :) [00:14] there's other towns in WA? [00:14] duflu: well, that's accurate, isn't it? [00:14] zing! [00:15] * thomi high-fives robert_ancell [00:15] ka-pow! [00:15] we is so witty [00:15] heh [00:15] Don't mention the S word yet. I only just got my server to raring [00:16] duflu, phone is switching, so we are switching [00:16] Fair enough. But not my stable server :) [00:16] I'll put saucy on a couple of others [00:19] TBH, the biggest issues I had with development releases in the past was the email client. Thankfully that particular problem has been solved for me now :) [00:20] I've been running saucy for ages now, no serious issues yet === duflu is now known as duflu|breakfast === duflu|breakfast is now known as duflu [02:00] thomi, debuild failed: /home/bob/bzr/mir/add-mir-stress/src/server/options/program_option.cpp:98:13: error: ‘cerr’ is not a member of ‘std’ [02:00] std::cerr << "ERROR in " << filename << ": " << error.what() << std::endl; [02:01] continuing after adding [02:01] * duflu reconfigures modem [02:26] thomi, hmm, now it fails because examples/demo_client_accelerated.cpp gives: [02:26] /home/bob/bzr/mir/add-mir-stress/examples/demo_client_accelerated.cpp: In function ‘int main(int, char**)’: [02:26] /home/bob/bzr/mir/add-mir-stress/examples/demo_client_accelerated.cpp:82:26: error: variable ‘rc’ set but not used [-Werror=unused-but-set-variable] [02:26] int major, minor, n, rc ; [02:26] this is because it's being compiled with -DNDEBUG and the asserts don't do anything [02:53] robert_ancell: huh? I don't get that at all [02:53] thomi, this is just a bzr branch and a debuild [02:53] robert_ancell: hmmm, I use 'bzr bd' [02:54] robert_ancell: can you try that? AFAIK that's the way the distro team are recommending people build PS upstream projects [02:54] thomi, ok === smspilla1 is now known as smspillaz [03:11] thomi, same issue as before [03:11] (i.e std::cerr, is compiling with -DNDEBUG) [03:11] that's odd, it builds fine for me [03:11] I mean, it fails to sign the packages, but that's expected [05:59] hikiko, RAOF, meeting [06:00] thomi, ^ [06:00] coming [06:00] kdub_, ^ [06:25] hikiko, ping [08:19] alf_: do you want to review fix-input-registrar-locking before I top-approve? [10:10] hi duflu - welcome back [10:10] alan_g: Thanks. [10:10] I was about to log off ;) [10:10] OK bye. Catch you tomorrow. ;) === alan_g is now known as alan_g|afk === alan_g|afk is now known as alan_g === alan_g is now known as alan_g|lunch === pete-woods1 is now known as pete-woods-lunch === alan_g|lunch is now known as alan_g === pete-woods-lunch is now known as pete-woods === jono is now known as Guest19882 [13:54] hi, question: what's the gbm_cursor? [13:56] hikiko: it's the hardware cursor [13:59] question: what do the ints in native_handle represent? [13:59] alan_g, we set a bo as drm cursor and we fill it using gbm? [14:01] hikiko: sounds about right [14:01] I mean cursor == a gbm bo rendered using drm (overlayed by gpu etc)? [14:01] yes [14:01] ok :) [14:01] thank you alan_g ! [14:05] alan_g: which native_handle is that? [14:05] android cutils [14:23] alan_g: I think the information there supposed to be somewhat opaque and driver dependent [14:24] alf_: I knew that. ;) === Guest19882 is now known as jono [14:25] The thing is when I try passing out a framebuffer there's a difference that affects the client behaviour. [14:25] alan_g: ok, so I guess the answer is "Who knows" (Kevin will probably have mor info) :) === jono is now known as Guest85454 [14:25] alf_: yep [14:26] alan_g: what happens? [14:27] alf_: a "normal" buffer gets posted back to the server, a frame buffer doesn't [14:27] the second int (index 1) seems to be key [14:27] "normal" buffers have 8, frame buffers 3 [14:28] If I overwrite this with 8 the client attempts to post the buffer but then things (predictably) go wrong [14:28] So I'm wondering what this int is about... [15:13] status: Started working on providing facilities for session/surface snapshots, in particular, implementing a BufferSwapper::compositor_peek() method [15:22] Morning === racarr_ is now known as racarr [15:25] Afternoon [15:33] greyback: Hangout? I dont have so much realy today (platform-API continues) [15:33] making progress now though === Cimi_ is now known as Cimi === lool- is now known as lool === mmrazik is now known as mmrazik|afk [16:24] Going to work on client-focus-notifications for a while to change pace from platorm-api [16:49] ghashashasfa over and over this test hangs when logging is disabled [16:49] over and over (100 iterations) it starts passing as soon as I enable logging [16:56] racarr_: which test is that? [17:07] alf_: My new test in client-focus-notifications === IReboot is now known as 36DAARST9 === Stskeepz is now known as Stskeeps [17:23] Finally got a failing run out of valgrind + logs...narrowed it down to the client or the [17:23] test itself [17:29] fascinating, the event makes it all the way to MirSurace::handle_event but sometimes [17:29] handle_event_callback is NULL [17:29] but that should all be set up beore things even create and how can it race [17:31] and [17:31] mir_wait_for(mir_connection_create_surface(connection, &request_params, create_surface_callback, this)); [17:31] mir_surface_set_event_handler(surface, &event_delegate); [17:31] there it is [17:31] ...*kick self* [17:32] lol [17:32] I thought for sure it was the socket code [17:32] not handling this sending an event before the surface is fully created well [17:32] and went through all the message serialization and parsing :( [17:34] does anyone remember [17:34] why event_delegate was removed from the constructor? [17:37] It's possible to fix now by setting the event_delegate rom the surface_created callback [17:37] as completion of that blocks reading of more messages === mmrazik|afk is now known as mmrazik === yofel_ is now known as yofel [18:18] Lunch for racarr === racarr_ is now known as racarr|lunch === racarr|lunch is now known as racarr === Stskeepz is now known as Stskeeps === Stskeepz is now known as stskeeps === stskeeps is now known as Stskeeps [19:24] * kdub_ has waded into the mesa weeds [19:25] its strange that we put our egl native display type in the egl native surface type definition... [19:30] namespace ubuntu { namespace application { namespace mir { [19:30] feels pretty good :) [19:31] kdub_: ? [19:31] refresh my memory -.- [19:33] racarr, isn't it more fun to just do ubuntu_application_api_mir_client_api_query_for_targetable_surfaces_that_are_not_focused? ;-) [19:33] kdub_: How do you know my password === yofel_ is now known as yofel === francisco is now known as Guest71674 [20:11] kdub_: Do you happen to know anything about building platform-API? I want to build phone packages in pbuilder but it says it needs hybris-dev which is a virtual package [20:11] and uh...I don't really know what that means :p [20:11] racarr, you probably have to add the ppa with hybris in it to the pbuilder setup [20:12] oh I didn't know it was still only in a ppa :) [20:19] lol now aptitude [20:19] segfaults [20:19] what [21:40] robert_ancell, quick q [21:41] robert_ancell, in a full converged world where we have support for legacy X apps... [21:41] where I have a touch-only device (e.g. tablet) and a legacy X app, that is not touch enabled [21:42] where would we need to inject touch support for that app? [21:42] I would assume still in X & the resp. toolkit [21:49] olli, yeah, in the toolkit afaik [21:49] they would work as well as running Ubuntu on the nexus 7 at the moment I suspect [21:50] i.e. mir would take input and pass it on down to x [21:50] basically 1:1 [21:50] none of the OIF translations would happen at the Mir level [21:50] is that close? [21:51] I think that's correct [21:52] I'm not sure if it's been completely designed yet, might be worth asking tvoss if he had any thoughts on it [22:18] olli: That sounds correct to me. [22:23] OIF does a lot of math to conclude which touch action might have been triggered, might be interesting to see if mir couldn't just pass the result onto X [22:23] rather than the series of events [22:27] robert_ancell: I wonder if you have time for a MP? src/mir-stress.cpp src/client.cpp src/threading.cpp src/results.cpp [22:27] otherwise I'll wait till the Aussies get online [22:27] thomi, can have a look [22:27] thanks [22:27] thomi, did you fix the packaging problem? [22:28] robert_ancell: yep :) [22:28] robert_ancell: turns out I was being dumb [22:28] and had forgotten to install the binary [22:28] olli: Maybe. Someone would have to do a close look at how they map up [22:28] certainly step one is sending the "raw" events through [22:29] racarr, ack [22:29] will bug tvoss next week [22:29] not sure who approved his vacation [22:29] Let's just hope he doesn't relax [22:29] ;) [22:29] maybe I should call him [22:30] "hey, I just had this great idea about somthing I have no clue of" [22:30] his wife will kill me [22:30] "How is your vacation?" "Good, yester..." Yeah, great so about X11..." [22:30] you should wait 2-3h [22:30] its way to early to call him :P [22:31] racarr, when you write "Yeah" I can't help but think of Office Space [22:31] yeah...., so [22:32] haha [22:32] yeah, so... that jenkins template you wrote the other day.... [22:33] * thomi hides at the mention of the "j" word [22:33] thomi, did you not get my memo? [22:33] ... [22:34] http://www.youtube.com/watch?v=Fy3rjQGc6lA === Cheery_ is now known as Cheery [23:51] whee, can build platform api-2 with hybris graphics/sensors, mir server/hybris sensors, mirclient/hybris sensors, and mirserver/mirclient desktop [23:51] the story of how one shared library became 5