| * RAOF would like gmock to hurry up and make NiceMock the default. | 05:30 | |
| duflu_ | RAOF: That would be troublesome for finding things you missed | 06:08 |
|---|---|---|
| * duflu_ -> lunch | 06:08 | |
| RAOF | duflu_: You can always explicitly make them naggy mocks if you want. | 06:09 |
| RAOF | duflu_: Another question is: what sort of things you missed? | 06:11 |
| === duflu_ is now known as duflu | ||
| duflu | RAOF: Naggy can be useful, yes :) | 07:07 |
| === WebbyIT is now known as rpadovani | ||
| Saviq | hey guys, do these test failures from devel mean anything to you https://launchpadlibrarian.net/170204996/buildlog_ubuntu-trusty-armhf.mir_0.1.8%2B14.04.20140321-0ubuntu1_FAILEDTOBUILD.txt.gz ? | 09:16 |
| alan_g | Saviq: it is only a guess - but is that an attempt to run the tests on a machine with no graphics hardware? | 09:20 |
| Saviq | alan_g, definitely, it's a launchpad builder | 09:21 |
| Saviq | alan_g, if those tests require graphics, they shouldn't be included in make check | 09:21 |
| Saviq | alan_g, OTOH it completed for i386 and amd64 | 09:21 |
| Saviq | alan_g, it's basically a PPA build of lp:mir/devel | 09:22 |
| alan_g | alf__: ^^ you were looking into which tests ran where recently? | 09:22 |
| Saviq | alan_g, alf__ ugh, so we lost that log right now, since the build was rekicked | 09:22 |
| Saviq | let's see if it's consistent, then, will report back in some 37 minutes | 09:23 |
| alan_g | Saviq: ack | 09:23 |
| alan_g | alf__: FWLW there was a lot of "C++ exception with description "Could not open hardware module" thrown in the test body." | 09:27 |
| alf__ | alan_g: hmm, that sounds like android hw, but didn't kgunn disable them for package building? Let me see.. | 09:29 |
| alf__ | alan_g: them = integration and acceptance tests | 09:29 |
| alan_g | alf__: "them" is what failed | 09:30 |
| alan_g | I thought so too | 09:30 |
| alf__ | alan_g: ah, they were disabled only in trunk not devel | 09:31 |
| alan_g | Saviq: ^ | 09:31 |
| alf__ | alan_g: don't we sync back from trunk to devel after release? | 09:31 |
| Saviq | oh, explanations, /me likes :) | 09:32 |
| alf__ | alan_g: well, it's clear we didn't in this case, but aren't we supposed to do so? | 09:32 |
| alan_g | alf__: we do. But this time there were some "trivial conflicts" maybe it got lost | 09:32 |
| alf__ | alan_g: Saviq: ok, let me sync again | 09:33 |
| Saviq | alf__, looks like "nothing to do" | 09:33 |
| Saviq | alf__, but indeed that change is not there | 09:33 |
| alf__ | Saviq: ok, I will propose an MP with the fix | 09:34 |
| Saviq | alf__, great, thanks | 09:34 |
| Saviq | alf__, please let me know when you get the MP, I'd like to put it in a silo | 09:43 |
| alf__ | Saviq: ok | 09:43 |
| duflu | Whee, circular make_shared dependencies eat all the memories | 09:51 |
| alan_g | duflu: stick to a DAG | 09:51 |
| duflu | alan_g: I know but I didn't see the cycle coming | 09:52 |
| alan_g | I guessed that ;) | 09:52 |
| alan_g | I presume no-one else is looking into https://bugs.launchpad.net/mir/+bug/1295231? | 09:53 |
| ubot5 | Ubuntu bug 1295231 in Mir "TestClientInput memory leak in CI" [Medium,Confirmed] | 09:53 |
| alf__ | RAOF: do we require libumockdev >= 0.6? Trunk says no, devel says yes, but the change in devel predates last sync of devel => trunk | 09:53 |
| duflu | alf__: I think that just landed for 0.1.8. It's documented in the upcoming changelog ;) | 09:54 |
| duflu | Oh, that part's not documented | 09:54 |
| alf__ | duflu: you are right, it's post 0.1.7 | 09:58 |
| alf__ | Saviq: https://code.launchpad.net/~afrantzis/mir/no-display-hw-tests-on-armhf-package-builds/+merge/212122 | 10:09 |
| Saviq | alf__, thanks! | 10:09 |
| * alan_g looks at TestClientInput and feels a strong urge to delete and rewrite | 10:26 | |
| === psivaa_ is now known as psivaa | ||
| alan_g | alf__: happy for me to top-approve https://code.launchpad.net/~alan-griffiths/mir/once-Surface-interface-to-rule-them-all/+merge/212012? | 11:57 |
| alf__ | alan_g: hmm, I thought I had review that (or something similarly named :)) | 11:58 |
| alf__ | alan_g: looking | 11:58 |
| alf__ | alan_g: @//resolve ambiguous member function names, what's ambiguous about them? | 12:12 |
| alan_g | They exist in multiple base classes | 12:12 |
| alan_g | (Which makes code calling them ambiguous) | 12:13 |
| alf__ | alan_g: but they are pure virtual in all of them, so shouldn't this just work? | 12:13 |
| alan_g | I know what you mean, but no | 12:14 |
| * alan_g tries to come up with code where it matters which name is resolved... | 12:16 | |
| alan_g | auto eg1 = &Surface::name; // eg1 is Base1::*() | 12:19 |
| alan_g | auto eg2 = &Surface::name; // eg2 is Base2::*() | 12:19 |
| alan_g | eg1 & eg2 have different types | 12:19 |
| alan_g | Even though they call the same member function | 12:20 |
| alan_g | There's no rule in the standard to disambiguate Base1::name and Base2::name | 12:21 |
| anpok_ | hm a using of either Base1::name or Base2::name? | 12:22 |
| alan_g | that would work too | 12:24 |
| alf__ | alan_g: I don't get what the first two line you pasted show | 12:26 |
| alan_g | The two alternative resolutions of Surface::name | 12:27 |
| alan_g | It isn't real code - as both lines are ambiguous | 12:28 |
| alf__ | alan_g: ok, got it, thanks | 12:33 |
| anpok_ | mterry: hi | 12:34 |
| anpok_ | I was reviewing and now testing the u-s-c branch.. | 12:34 |
| anpok_ | one question came up .. in which cases should the spinner show up? | 12:35 |
| mterry | anpok_, hi | 12:35 |
| mterry | anpok_, well, where you testing with other split branches or just the usc branch? | 12:36 |
| anpok_ | oh - only usc - and I only see it on startup .. i need a unity8 branch too? | 12:42 |
| mterry | anpok_, so the branch is OK in isolation. But you won't see all the scenarios unless you also install some other branches... Let me give you links, and I can explain when you'll see others | 12:49 |
| mterry | anpok_, https://code.launchpad.net/~mterry/unity8/split/+merge/210664 and https://code.launchpad.net/~mterry/ubuntu-touch-session/split/+merge/211549 | 12:50 |
| mterry | anpok_, the times you'll see it are 3-fold: | 12:50 |
| mterry | anpok_, boot, waiting for greeter to come up; boot, after greeter comes up but if the session behind the greeter is slow to appear and you swipe greeter away very fast, you'll see spinner until user session finishes appearing; and when you lock device, if you turn it back on very quickly before greeter has had a chance to finish appearing, you'll see it | 12:52 |
| mterry | anpok_, that second instance Design actually wants me to get rid of -- I'll make a small adjustment to that branch so that the greeter doesn't appear until the user session is also ready | 12:52 |
| === alan_g is now known as alan_g|lunch | ||
| === alan_g|lunch is now known as alan_g | ||
| === psivaa is now known as psivaa-afk | ||
| === psivaa-afk is now known as psivaa | ||
| === dandrader is now known as dandrader|lunch | ||
| === dandrader|lunch is now known as dandrader | ||
| === dandrader is now known as dandrader|afk | ||
| mterry | racarr_, hello! | 20:49 |
| mterry | racarr_, we've got a testing PPA for the split greeter stuff going (which includes your remove-ensure-display-powered merge) | 20:50 |
| mterry | racarr_, but I'm still seeing the screen turn on after locking | 20:50 |
| mterry | racarr_, I was wondering if you had time to re-investigate that. Might be easier now that there is a PPA with all the code for easier testing | 20:50 |
| mterry | racarr_, https://launchpad.net/~ci-train-ppa-service/+archive/landing-004 | 20:51 |
| racarr_ | mterry: Ok I will investigate | 20:53 |
| * mterry buys a beer for racarr_, keeps it cold in fridge for ya | 20:54 | |
| mterry | tsk.. this beer will go bad before I see you next | 20:54 |
| * mterry opens beer | 20:54 | |
| dansuf | Hi, I am porting ubuntu touch to my phone and I've got graphics-related issues. Here's a log generated by logcat when running lightdm: http://pastebin.com/ag6iSVkW . Also, in dmesg I've got "mdp4_overlay_play: pmem Error" and in unity-system-compositor.log: "terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >' | 20:56 |
| dansuf | what(): error during hwc set() | 20:56 |
| dansuf | " | 20:56 |
| dansuf | Any ideas what could be the cause of these errors? | 21:00 |
| === dandrader is now known as dandrader|afk | ||
| racarr_ | mterry: Cheers lol | 21:12 |
| racarr_ | dansuf: Hmm I mean abstractly it looks like some of the hardware compositor usage through the android HAL...isnt working | 21:13 |
| racarr_ | the different devices have different quirks. kdub or AlbertA would know more. | 21:13 |
| kdub | dansuf, yeah, something is going wrong with hwc and the drivers | 21:15 |
| racarr_ | We used to have a fallback mode...presumably still do, that might work for you | 21:15 |
| racarr_ | dansuf: in mir/src/platorm/graphics/android/output_buider.cpp I think you could force_backup_display(true) in the constructor | 21:15 |
| dansuf | so it may be something connected to device files | 21:15 |
| dansuf | racarr_, I will have to build mir? | 21:16 |
| kdub | fallback is discontinued in most up to date qcom drivers | 21:16 |
| kdub | which device? | 21:17 |
| racarr_ | Oh. :( maybe we dont have a fallback | 21:17 |
| dansuf | aa far as i know its legacy version if this does change anything | 21:17 |
| dansuf | i mean legacy qcom | 21:17 |
| dansuf | i don't have big knowledge about all this stuff | 21:17 |
| dansuf | it's sony ericson live with walkman, codename coconut | 21:18 |
| AlbertA | it does look like dev permissions from the log | 21:18 |
| AlbertA | or maybe your kernel driver does not instantiate those devices | 21:19 |
| dansuf_ | i applied udev permissions from android files | 21:19 |
| kdub | yeah, maybe run as root and see if that helps | 21:19 |
| dansuf_ | i run it in root shell | 21:20 |
| kdub | also, those drivers have a hwc.mdpcomp.enable (or similar), maybe try turning that off | 21:20 |
| === dandrader|afk is now known as dandrader | ||
| dansuf_ | kdub, I set that to false in build.prop, it still fails but the logcat is slightly extended http://pastebin.com/Au3Paj1x | 21:45 |
| kdub | i'd still guess there's some permissions or chroot mix up somewhere | 21:48 |
| kdub | do the hybris test programs work? | 21:48 |
| dansuf_ | kdub, I didn't run them | 21:49 |
| dansuf_ | Where can I find them? | 21:49 |
| kdub | /usr/bin/test_glesv2 from libhybris-test | 21:50 |
| dansuf_ | kdub, wow, it works | 21:51 |
| dansuf_ | I mean, the tests | 21:51 |
| kdub | dansuf, good sign | 21:51 |
| dansuf_ | Although, egl seems to crash | 21:52 |
| dansuf_ | I'll try maybe to build mir with fallback | 21:56 |
| dansuf_ | So, I tried to build mir for ubuntu touch and It just throws error with no reason "[ 7%] Built target android-input | 22:43 |
| dansuf_ | make: *** [all] Error 2 | 22:43 |
| dansuf_ | " | 22:43 |
| bregma | racarr_, just so you know, I pushed your u-s-c and the lightdm tweaks to https://launchpad.net/~unity8-desktop-session-team/+archive/custom and installed them, I now have a working mouse and trackpad in the Unity8 desktop session | 22:57 |
| bregma | just. like. that. | 22:58 |
| bregma | I look forward to those hittong the archive soon for winder consumption | 22:58 |
| bregma | *wider | 22:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!