=== thumper is now known as thumper-afk [04:59] I apologize in advance for the mail spam that is about to occur [04:59] need to resubmit a bunch of stuff so that the CI bots will wake up again === thumper is now known as thumper-cooking [05:12] * duflu tries to shut eyes === thumper-cooking is now known as thumper [08:25] hey mmrazik, how are you? [08:25] morning didrocks [08:25] not so bad [08:26] mmrazik: ;) do you think it's possible to cat ap_test_debug_log.txt to the console? Would be easier to grasp some important infos in the unity/autopilot jobs? [08:26] like, I thought at first that the exit(1) was a failed installed, when it was the filtering aborting [08:26] didrocks: it would be splendid but this is something in UTAH. I've even reported a bug for it [08:26] let me find it [08:26] mmrazik: there is no way to have that as a post_trigger? [08:27] mmrazik: like we do for crash files and so on? [08:27] (not asking for real time logs, just an "after the fact") [08:27] mmrazik, bug 1075969 ? [08:27] not quite sure what you are talking about. I had an impression UTAH is very hostila towards stdout and there is no way how to print it. Not really sure if we are doing anything like that anywhere [08:27] bug 1075969 in UTAH "getting a real-time progress in jenkins would make debugging easier" [High,In progress] https://launchpad.net/bugs/1075969 [08:27] if you can give me any pointers I can give it a try [08:28] jibel: yes, looks like the one [08:29] didrocks, actually I dont think it's possible in the current state of utah [08:29] mmrazik: look at the autopilot intel job, rev 109 [08:29] jibel: ^ [08:29] in the end, we have: [08:30] + grep -E '(compiz|X|autopilot)' [08:30] _usr_bin_compiz.1000.crash [08:30] + echo 'Appears to have been a crash' [08:30] Appears to have been a crash [08:30] + exit 1 [08:30] so it seems we can show something, isn't it? [08:30] oh right... after we download it from the machine we can cat it in jenkins [08:30] urgh... [08:30] what you see on the console is the output of the process that runs on server's side [08:30] but that is _really_ stupid [08:31] agree [08:31] didrocks: but you are right we can do it [08:31] didrocks, and the test runs on a remote server [08:31] mmrazik, that doesn't improve anything, just saving a click on an artifact [08:31] jibel: mmrazik: at least, while we are waiting for UTAH to give better support to stdout, that would give us more hint when the issue is on our side [08:32] jibel: which is a lot TBH. Everybody is looking at console logs [08:32] jibel: right, but it's confusing to see exit 1 without any info [08:32] mmrazik, the really interesting feature is the bug you filed [08:32] agreed with mmrazik [08:32] I consider this one of the biggest drawbacks of utah [08:32] I was blaming UTAH for 10 minutes before having a second look :) [08:32] and I wrote that detect/abort script :p [08:34] didrocks: its there. Let me know if you don't see it in the next run [08:35] mmrazik: excellent! it's the same script for oif? [08:35] didrocks: I changed all 3 jobs -- oif, indicators and unity [08:35] mmrazik: perfect :) it's a little bit coffeeless bulletproof! Thanks a lot :) [08:35] more* [08:46] morning all [08:47] Morning! [08:47] didrocks: hi! Do you know what's up with ps-indicators-autopilot-release-testing ? I see ati red but can't get the log out of the failure from jenkins [08:48] sil2100: it's all my fault this one :) [08:49] sil2100: my filter was badly updated, I had a bug due to edubuntu-desktop :) [08:49] sil2100: fixed and rerun, this time, I hope we'll get results ;) [08:49] hey popey! [08:49] ;) [08:50] didrocks: btw.! I fixed the ibus tests finally here, now I'm just resolving some minor char encoding problems [08:50] By 'fixing' I mean 'enhanced', had to hack around the introspection limits of ibus [08:51] sil2100: excellent, this is linked to the recent change in autopilot? [08:52] didrocks: not really, I'm still using the old autopilot here it seems [08:53] ok :) [09:35] hmm [09:35] it would be really cool if we could have nightly runs of phoronix-test-suite one day [09:36] its pretty interesting to play around with [10:26] sil2100: do you mind looking at job/ps-oif-autopilot-release-testing/label=autopilot-ati/? [10:26] sil2100: one test on two failed [10:27] run 37 [10:29] smspillaz: I actually have openarena (from phoronix) running on a daily basis [10:29] on intel (HE+LE) and ati (HE+LE) with open source drivers [10:36] didrocks: checking === mmrazik is now known as mmrazik|lunch [10:45] didrocks: trying to find out why the test failed, it seems that something 'happens' when the test is ran and the dash is opened for the very first time in the session [10:47] It seems unity registered the Alt key press but the F4 keypressed went unnoticed, hmmm [10:47] sil2100: thanks, keep me posted :) [11:03] didrocks: I'll revisit it in a moment again [11:03] ok ;) [11:22] mmrazik|lunch: ah interesting === mmrazik|lunch is now known as mmrazik [11:23] I'm trying to hack pts at the moment to make it play the games in windowed mode, since that's what I'm most interested in === chaoticuk_ is now known as chaoticuk [13:01] Trevinho: bregma: I think you do see that armhf is failing to build, right? [13:03] didrocks: mh, no... looking [13:03] Trevinho: you can look at the staging ppa or the ubuntu-unity daily-build one [13:08] didrocks: ah, I see... a test that andyrock wrote is failing there... CAn we wait him to fix it or should I look at it now? [13:08] Trevinho: depends on when he will be able to fix, do you know? [13:09] Trevinho: like in the coming hours, before 00 UTC? [13:09] didrocks: he's generally available in the late afternoon [13:09] so he should yes [13:09] Trevinho: ok, in that case :) [13:12] good morning! === bambee is now known as rperier [13:16] salut cyphermox, ça va? [13:16] salut didrocks, ca va toi? [13:17] cyphermox: mouai mouai [13:37] sil2100: I didn't follow, did you get anything for the failing test? [14:09] didrocks: working on it on my guest session [14:29] all this red in jenkins! [14:30] mterry: we are discussing it right now, will get back to you then [14:31] mterry: do you want the short or long story? [14:32] didrocks, short I suppose :) [14:32] mterry: so basically, if I just scope to Unity: [14:32] - armhf is FTBFS [14:32] a test failing that andyrock modified [14:32] Trevinho will ask andyrock once he's online [14:32] mterry: can you ensure as well that's fixed? ^ [14:32] didrocks: I've contacted it in private, he has an exam, but he'll fix tonight... [14:33] Trevinho: thanks a lot :) [14:33] - a new issue on UTAH [14:33] or at least, povisionning the image [14:33] random failures [14:33] k [14:42] didrocks: ok, so, I tried and tried and tried, but I can't reproduce it or find a logical cause right now for the error ;/ [14:43] didrocks: the FTBFS is an alignment error, typical for ARM, but hm, maybe it might get fixed up [14:43] sil2100: ok, well, it's still part of those random tests failure [14:44] sil2100: fortunately, this time, UTAH will pass and you'll get fresh result [14:44] sil2100: it's still a little bit high, like ati shows 26 failures [14:45] didrocks: the unity release stack, yes? [14:47] sil2100: yeah, unity [15:04] sil2100, yay for ibus branch! [15:10] sil2100: the job finished, 25 failures on intel [15:15] \o/ [15:15] uh === mmrazik is now known as mmrazik|afk [16:13] Trevinho: Hi [16:34] didrocks: in the background, I'll also maybe try to fix the FTBFS for armhf, but this might take a bit longer because I'm using an ARM pbuilder for build-testing right now [16:35] didrocks: since it takes some time, I can work on the tests in the meantime ;) [16:36] sil2100: sweet! rocking ;) [17:16] hi there, what should I uninstall to disable only amazon search results? unity-lens-shopping or amazon webapp? im a bit confused it seems amazon is unity-webapp-common :s [17:23] bregma: perf test results are up [17:23] smspillaz, where? [17:23] bregma: 5x boost on unigine for some reason, 5fps on some of the other games [17:23] bregma: https://code.launchpad.net/~compiz-team/compiz/compiz.fix_1024304/+merge/147832 [17:24] I think it really depends on the type of application as to what its individual performance improvement is going to be [17:24] naturally [17:25] supertuxkart is just slow no matter what you do, unigine does a lot of postprocessing and really benefits from the freed up fill-resources on the gpu [17:25] bregma: phoronix-test-suite is pretty neat though - should probably keep it on the radar as part of the testing strategy [17:26] and not just because phoronix runs it with a great deal of prejudice [17:27] it would be very easy to hook it into compiz' framerate counters - all it does is spawn a process and watch the stdout with a regex for the fps [17:28] bregma: well, I've realized that a large part of the reaction to those benchmarks is often denial which results in lots of ad-hominem against phoronix [17:28] where attention is usually deflected away from the numbers to the journalistic quality [17:29] I'm guilty of this [17:29] its probably a natural reaction [17:29] fact is any improvement in those high-profile benchmark numbers can have a positive influence on Unity uptake [17:29] so it's a worthwhile exercise [17:29] yeah [17:30] bregma: although a warning: I modified the benchmarks slightly. those numbers are actually slower than what phoronix will report [17:30] it was only stressing the relevant codepath if I ran them windowed, which pts does not do by default [17:30] (it runs everything fullscreen, which of course, bypasses compositors) [17:31] might be a nice feature request for them though [17:31] (and of course the relevancy of the numbers comes from the fact that not everything is a fullscreen window) [17:33] smspillaz, can your change be staged (nux then unity) or does it need to go in all at once? Is is an ABI-breaking change? [17:33] bregma: it really needs to go in all at once [17:33] that's a pain [17:34] yeah I know :( [17:34] this is where direct push access comes in handy - verify that CI is green, turn off the merge bots and then just merge everything once its ready [17:36] bregma: as for ABI breaking - yes for compiz, I don't think so for nux [17:36] actually, no, it is for nux, had a change a function signature === alesage is now known as alesage|afk === alesage|afk is now known as alesage [20:41] hi whats the status/progress of widgets for Unity? Is there any code I can test or dev? === hggdh_ is now known as hggdh [21:03] osmdave: what do you mean by widgets? [21:07] thumper like adding anything to the desktop that isn't an icon. Weather, html5 apps, calculator, drag-able images and things like that. [21:09] osmdave: nothing unity specific AFAIK [21:11] thumper: ok thanks what is the part of the desktop that icons and widgets sit on? What handles that if not unity? [21:11] osmdave: that is just the desktop, and yes unity does handle that, but I'm not aware of any widget code coming any time soon [21:16] thumper: where abouts in the source is the desktop part of the desktop that deals with the icon layout etc.? [21:17] not sure exactly [22:07] thumper, hey. seems I've missed osmdave, but if he comes on asking again, you can tell him that Nautilus handles the icon rendering on the desktop, and doesn't do any desktop widgets on it's own. [22:08] robru: ok, ta [22:08] thumper, there are third-party things that can already do desktop widgets, but nothing that's installed by default or officially supported. [22:09] * thumper nods [22:09] thumper, cool. tried googling an email address for him but didn't get much ;-)