=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [07:53] Hi, [07:54] Just want to let you know that I have publish a few games, native game based on sdl2 [07:55] so there are directly using mirclient, through sdl2, and they can be a way to test mir [07:55] (indeed, it seemed one app have made mir crashed, when disconnecting) [07:56] so if you want to try, here's the link : https://uappexplorer.com/app/com.ubuntu.developer.1bsyl.cartes [08:11] slvn_, hehe i have almost all of them installed :)) [08:11] slvn_, negative space is really nice [08:12] thanks, great you enjoyed ! [08:13] slvn_, was it hard to port your games to ubuntu phone? i you also have android versions [08:13] usually people gives up quickly with negative space :) [08:13] they run out of ink ))) [08:14] not really hard to port .. actually there were developed first under ubuntu and android [08:14] there was a few issue with mir ... for orientation .. and couples of bugs.. [08:15] and, since I use sdl2, I can re-use all the code for all platforms [08:18] gatobaubau, I was told then when quitting my app, sometime mir crashed ... maybe you could double check ... because I have no ubuntu-deice :D [08:20] slvn_, i never had any problems with your apps, i'll check to logs later :P just in case [08:48] slvn_: hm mir as in the unity8 nested server? or unity-system-compostior? or really a shutdown crash of the application? [08:49] I just copy paste, since it's short ... [08:49] qtmir.mir: SessionListener::stopping - this= SessionListener(0xb194a134) session= 0x5f935c [08:49] terminate called after throwing an instance of 'boost::exception_detail::clone_impl >' [08:49] what(): Failure sending input event : Unknown channel provided [08:49] () [08:50] to be confirmed ... [08:51] ah [08:52] this is actually possible.. and if not handled by the input dispatcher the server would be affected [08:52] *plausible [08:53] hm since qtmir has its own we need to fix that there too [09:00] .. don't really understand this :) ... but, maybe one should double check if this really happens with an official version ... my app were a use-case to reproduce this .. [09:02] well i guess you need to hit the right point in time to trigger the problem.. looking at the code.. it is caused by the fact that the input communication is done in a separate thread and keeps track of active surfaces.. [09:02] and in the case described the component knows that the surface is already gone, before it gets the request to dispatch the event.. [09:04] hmm so I either change the way the input channels of surfaces are tracked and kept alive a little longer or we fix the caller to handle the case.. [09:05] https://bugs.launchpad.net/mir/+bug/1521529 [09:05] Ubuntu bug 1521529 in Mir "handle failures during input dispatch when client disconnet" [Undecided,New] [09:06] one of things that could be reworked.. the abstraction around InputChannel is still kind of odd becaue we use the android input framework for that.. === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [12:15] sometimes I see line of code that makes me wonder... [12:15] if (write(p.write_fd(), "a", 1)) {} [12:25] write returns non-zero on error and success, doesn't it? [12:25] * mcphail doesn't have "man" installed on this machine [12:36] Yes, but is this really the clearest way to ignore the return code. === alan_g is now known as alan_g|lunch [13:00] Those curly brackets were actually empty? Probably meant to check for success but couldn't understand why write() was still returning true on failure and deleted the block. === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === alan_g|lunch is now known as alan_g === gatobaubau is now known as gato_aw === dandrader is now known as dandrader|afk [16:10] /join #ubuntu-touch === francisco is now known as Guest98423 [17:07] is there a daily iso for testing unity8+MIR? [17:08] the desktop-next daily-live is gone, but that is what most wiki pages link to [17:12] k1l: I don't think so (we don't create/publish ISOs etc around here, just churn the code). There is a package: unity8-desktop-session-mir - does that help? [17:13] i just wanted to give it a test run to see what the state is, while not touching my actual system. [17:14] (just tested a fedora rawhide iso which should have wayland as default now but that didnt boot to the login/desktop so i was curious what the MIR/unity is doing on my thinkpad hardware) [17:17] I guess you could boot a vanilla ISO, and install unity8-desktop-session-mir before switching sessions. [17:17] will try that, thanks === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|EOD === shiznix is now known as Guest4941