[06:29] hello Mir people [06:30] i get 10-50% more FPS with glmark2 on OTA8 compared to OTA5 [06:30] on mx4 [06:30] good job === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === pete-woods1 is now known as pete-woods === chihchun_afk is now known as chihchun [10:25] alf_: thanks for picking up the NestedServer test issue [10:25] alan_g: np === alan_g is now known as alan_g|test === alan_g|test is now known as alan_g === greyback__ is now known as greyback === mzanetti_ is now known as mzanetti === dandrader is now known as dandrader|afk === alan_g is now known as alan_g|lunch === dandrader|afk is now known as dandrader === alan_g|lunch is now known as alan_g [14:02] alan_g, need your input https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/1517615 [14:02] Ubuntu bug 1517615 in lightdm (Ubuntu Vivid) "Need to disable "--enable-hardware-cursor=true" option in unity-system-compositor" [Undecided,Confirmed] [14:03] dandrader: looking... [14:03] alan_g, this solve https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/1515921 for ChrisTownsend [14:03] Ubuntu bug 1517615 in lightdm (Ubuntu Vivid) "duplicate for #1515921 Need to disable "--enable-hardware-cursor=true" option in unity-system-compositor" [Undecided,Confirmed] [14:03] s/solve/solved [14:04] dandrader: After chatting w/ bregma about that, the "--enable-hardware-cursor" option was added back in the day as a hack to get a cursor in the Unity 8 desktop session until Unity 8 provided its own cursor. [14:06] dandrader: Since Unity 8 now provides its own cursor, I think the option can be safely removed. [14:06] dandrader: I think enable-hardware-cursor is a USC specific "feature" [14:06] ChrisTownsend, I commented on the bug with my understanding on the issue [14:07] alan_g, ah, so it's not about a switch for using a hardware cursor instead of a software one? [14:07] // This is a workaround for u8 desktop preview in 14.04 for the lack of client cursor API. [14:07] // We need to disable the cursor for XMir but leave it on for the desktop preview. [14:07] // Luckily as it stands they run inside seperate instances of USC. ~racarr [14:08] dandrader: no, it enables a USC specific cursor [14:08] alan_g, ah, ok. great. then I think we can remove not only the parameter from the lightdm call but also its implementation in usc === chihchun is now known as chihchun_afk [14:11] Hmm. actually reading the code more carefully it is an option to enable the Mir cursor that defaults to disabling the cursor. [14:14] dandrader: ^ - so removing the parameter is likely best. [14:16] alan_g, ok [14:29] really, I'd leave it in u-s-c for now, because we may want that when prototyping kiosk applications, but remove it as default in lightdm (and make it a config option there) === chihchun_afk is now known as chihchun [14:46] alf_: any progress on lp:1517781? [14:46] alan_g: Yes, but I need to perform some more testing to ensure the fix doesn't have any side effects [14:47] Excellent! === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === chihchun is now known as chihchun_afk === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [18:47] bschaefer, do you know what causes the issue on nexus 4 where sdl2 apps show two copies of the screen side-by-side? [18:47] popey, yup! [18:47] is it an SDL/Mir issue or something which can be fixed in the game? [18:47] good! [18:47] SDL issue [18:47] awesome. [18:48] popey, fix is here: https://code.launchpad.net/~brandontschaefer/+junk/SDL2-new-mir-ABI [18:48] as well as the touch screen issue [18:48] ooooh [18:48] which issue is that? [18:48] touch wouldnt work :) [18:48] hah [18:48] okay, good one to fix :) [18:48] because i made.... up = down and down = up [18:48] yeah [18:48] lulz [18:48] I'll build that and have a play, thanks! [18:48] it's fun playing doom on an e4.5 :) [18:49] http://people.canonical.com/~alan/screenshots/device-2015-11-19-184911.png [18:49] :) [18:50] haha awesome! [18:55] bschaefer, out of interest what do you build armhf things on locally? [18:55] I'm building on a nexus 7, but feel I could probably benefit from maybe using a chromebook with more storage (and a screen and keyboard) [18:55] popey, on my N4 [18:55] or cross build [18:55] from pbuilder [18:55] pbuilder-dist vivid armhf build *.dsc [18:55] popey, that could work as well [18:55] good point [18:56] * bschaefer always finds arm slow to compile no matter the device :) [18:56] have to wait for the compiler to catch up [18:58] I keep running out of space [18:58] so thinking of a cheapo acer chromebook [18:58] popey, yeah thats my issue on N4 [18:58] soo i just use the pbuilder-dist for now... copy over debs [18:58] and install thought it sucks when you need to debug [18:59] when i need to debug i've to build on a device [19:39] hmm bschaefer it's failing to build here - http://paste.ubuntu.com/13351541/ [19:39] popey, its not building mir :), hmm strange what version of mir do you have? [19:41] bschaefer, which package should I look for version of? [19:41] mirclient-dev [19:42] Installed: 0.17.1+15.04.20151105.1-0ubuntu1 [19:42] hmm that should be fine... [19:42] popey, try --enable-video-mir [19:42] but it should be default... [19:43] ok [19:44] Time passes... :) [19:44] bschaefer: you can det up a symbol server and use that for remote debugging [19:44] bschaefer: saves building on the device [19:45] nice ill have to look into that [19:46] bschaefer: there's a good youtube vid from one of the valve developer conferences which explains it better than I could [19:47] o nice ill have to take a look [19:48] https://www.youtube.com/watch?v=xTmAknUbpB0 - about half way through that one [20:02] bah, build failed again [20:02] :( [20:03] http://paste.ubuntu.com/13351832/ [20:03] says mire support = no [20:03] mir* [20:04] popey, do you have libxkbcommon-dev? [20:04] yes, 0.4.3-2 [20:05] popey, you can ... try to manually edit the configure.in [20:05] under CheckMir [20:05] to force it to be enabled... as that doesnt make much sense... something is wrong [20:06] you also have libegl1-mesa-dev? [20:06] i do [20:07] :::( [20:07] its only been a week since i've tried this on vivid + overlay [20:07] and it worked... [20:07] :( [20:07] * popey dist-upgrades [20:07] dunno how old this tablet has got [20:08] probably new crack since I last updated [20:08] well ... you said mir was 0.17 [20:08] which is more then enough [20:08] (i just changed to 0.14/0.15) [20:08] * bschaefer checks again [20:11] :( [20:12] popey: can't you just cross-compile in a chroot? [20:12] probably could, yeah [20:12] will do if I can't get this working [20:12] don't understand what's going wrong [20:12] checking for Mir support... yes [20:13] Installed: 0.17.0+15.10.20151008.2+r3103~daily1~ubuntu16.04.1 [20:13] hmm you have a newer version [20:13] * bschaefer upgrades [20:14] popey: isn't it just looking for X and not finding it? [20:14] (which would be expected) [20:15] hmm [20:15] * mcphail doesn't recommend debuild for this kind of thing [20:16] popey: the failure is on line 587 [20:16] huh ... im still at 0.17.0 [20:17] you can try (and it should be there?) --disable-video-x11 [20:17] but it builds it [20:17] mcphail, theres the other issue of mir not building though === charles_ is now known as charles [20:18] bschaefer: probably just pkg-config looking in the wrong place [20:19] we need an armhf ppa for your bleeding edge sdl :) [20:19] haha... well you have a better version of mir then i do (still waiting on 0.17.1 to actually fix the color format issue [20:19] popey, o wait... that clone issue *isnt* fixed in that branch [20:19] bschaefer: what's the link to your repo again? I'll try building in a chroot [20:20] since im waiting for 0.17.1 :) [20:20] ooh! [20:20] good! [20:20] * popey ctrl-cs the build [20:20] (need to ask egl the pixel format is the issue) [20:20] ha! [20:20] vs picking a supported one [20:20] I don't want to push my silly dosbox click to the store with this problem [20:20] yeah [20:20] will get angry mail from nexus 4 owners :) [20:20] i can imagine, unless they like that sort of thing [20:21] (though input would be all wonky) [20:21] popey: tell them it is virtual reality [20:21] hah, need to buy a proprietary lens and cardboard from me [20:21] it's a shame we can't detect device in the store, and say "No to you!" for people with duff devices [20:22] bschaefer, is there no workaround for this? [20:22] in the app [20:23] neverball had this issue, and I think they worked around it in the app.. [20:23] popey, whats a duff device? [20:23] nexus 4, nexus 7 [20:23] (devices where it doesn't render correctly) [20:23] but as far as sdl2 cares ... its just weather its supports the rendering library or not [20:24] hell, armhf thinks glx works because the headers exist :) [20:24] soo theres nothing to fancy for checking rendering atm [20:24] sdl2 armhf* thinks glx works because opengl installs those headers [20:30] wonder if I can tell dosbox to request some other mode from sdl [20:30] is there some other mode I could use maybe [20:30] lower colour depth maybe [20:30] * popey is clutching at straws now [20:34] https://github.com/pseuudonym404/neverball-touch/commit/1f50b5f20d4eaa0d024379ca2363a4f2bae2f631 that's what the neverball person did [20:39] try it [20:42] * popey rebuilds dosbox [21:07] didn't fix it :( [21:08] popey, only way is to recompile with [21:08] umm [21:09] SDL_mirwindow.c:111: surfaceparm.pixel_format = FindValidPixelFormat(mir_data); [21:09] to [21:09] SDL_mirwindow.c:111: surfaceparm.pixel_format = mir_pixel_format_rgb_565; [21:09] popey, but this will cause issues on other devices (maybae?) [21:10] hmmm [21:11] * popey looks for a cheeky ppa to build it in [21:11] ppas dont usually build for armf :( [21:11] armhf* [21:11] I have one that does, on bare metal [21:12] o nice [21:21] https://launchpad.net/~canonical-community/+archive/ubuntu/ppa/+build/8320572 [21:21] lets see [21:24] bschaefer: any chance you can persuade them to release sdl2.0.4 soon? We've been on 2.0.3 for ages and we're never going to get the mir fixes in the repos if we don't get a new version [21:25] them being Sam? [21:25] Is it all down to Sam? [21:26] Looking at some of the 2.0.4 grubles on mailing lists I thought it was more complex than just going through Sam [21:26] IIRC deb is still on 2.0.2 [21:26] mcphail, and i cant push upstream until the LTS is released [21:27] since upstream follows the LTS [21:27] bschaefer: aah. OK [21:27] mcphail, worth taking a snapshot of 2.0.4 and putting in this ppa for fun and lulz? [21:27] the plan is to get all the fixes merged into archive [21:27] from a ubuntu patch [21:27] but need to wait for 0.17.1 ... [21:28] popey: my old sdl build was a "2.0.4" snapshot, but, of course, won't have bschaefer's fixes [21:28] right [21:28] bschaefer: presumably that isn't in OTA8? [21:29] OTA8? [21:29] whats on the phone today [21:29] bschaefer: current release on the phone [21:29] yeah its not [21:29] trying to get the patch done but got side tracked on something i have to finish for mir [21:29] before 0.18 :) [21:29] :) [21:35] hm, failed to build in the same way [21:35] https://launchpadlibrarian.net/227045641/buildlog_ubuntu-vivid-armhf.libsdl2_2.0.3~ppa1~vivid1_BUILDING.txt.gz [21:35] bah [21:36] * popey gives up for tonight [21:37] mcphail, is it worth (for fun) taking your 2.0.4 snapshot and integrating bschaefer's patches? [21:37] or not worth it until 0.17.1 of mir? [21:38] it doesnt have any video drivers :( [21:38] not sure what causing the failure at the end, but even if it worked, nothing would work :) [21:38] popey: don't hink it is worth it, tbh. I'm sure bschaefer's branch will be better in every way [21:38] yeah [21:39] I'll just play with stuff on my bq for now and not upload anything till it's working [21:39] Shame [21:39] appreciate the help chaps, this is one of the more fun parts of hacking [21:39] haha [21:39] * bschaefer enjoys the programming bits more [21:39] A device isn't viable until it has run doom [21:41] haha === Guest63074 is now known as UukGoblin