/srv/irclogs.ubuntu.com/2013/03/28/#ubuntu-unity.txt

olli_fginther, so we have a good ppa?02:22
=== racarr|lunch is now known as racarr
dufluThis makes me sad :(  https://bugs.launchpad.net/ubuntu-power-consumption/+bug/116121206:34
ubot5Launchpad bug 1161212 in unity (Ubuntu) "[regression] unityshell makes compiz wake up 190 times per second on an idle desktop" [High,New]06:34
smspillazduflu: have you looked into it ?07:24
smspillazduflu: this stuff is easy to regress when you don't have constant metrics07:24
duflusmspillaz: I don't look at such code any more07:24
dufluAnd so far have succeeded in not looking since releasing 0.9.9.007:25
smspillazduflu: I'm only get 5.9 w/s07:25
smspillaz*getting07:25
smspillazwhere is my english today07:25
smspillazduflu: its still higher than it should be (should be around 0.3)07:25
duflusmspillaz: I didn't spend long looking at the default idle rate but it went at least as low as 1307:26
dufluI was comparing it with Mir's new on-demand scheduling (which seems to work)07:27
smspillazduflu: did you check the very first result from powertop ?07:29
smspillazor after the first refresh07:29
smspillazits around 190w/s on the first check, but I suspect that's because of all the repaint-flurry caused by loading powertop07:30
smspillazafter that it returns to a stable number07:30
duflusmspillaz: No it sits at 190 for several refreshes at least07:34
dufluMaybe I needed to give it a few minutes to come back down? Doesn't sound right07:35
Saviqhey everyone, after yesterday's updates we need to update the unity tree in ../build_unity/unity09:32
Saviqthis can usually be done by ./build_unity -u09:32
Saviqbut this time: `bzr pull --overwrite -d ../build_unity/unity; ./build_unity`09:33
Saviqthat's because we rebased unity on lp:unity09:33
didrocksdednick: hey, do you mind getting all your branches merged to trunk?09:57
MacSlowSaviq, is that perhaps the reason I'm getting "Error on creating style rule:..." for my test and the shell?09:57
SaviqMacSlow, nope, that sounds like theming issues09:58
dednickdidrocks: sure09:59
didrocksthanks :)10:00
Saviqdandrader, can you https://code.launchpad.net/~saviq/unity/phablet.revert-carousel/+merge/15591910:04
dandraderSaviq, ok. Will check it now10:05
Saviqdandrader, cheers10:05
dandraderSaviq, this crash, is it easy to reproduce?10:14
Saviqdandrader, yeah, just search in the dash10:14
Saviqdandrader, click on search in top panel10:14
Saviqdandrader, type anything10:14
Saviqdandrader, in people / music / video, not home / apss10:14
Saviqapps10:14
Saviqgreyback, around?10:52
greybackSaviq: yep10:52
Saviqgreyback, can you mumble?10:52
greybacksure10:52
greyback1 sec10:52
=== dandrader is now known as dandrader|afk
davidcalledidrocks, (not sure if you have answered - network issue)11:17
didrocksdavidcalle: not sure your ping even reached me ;)11:23
didrocksbut it seems not11:23
=== dandrader|afk is now known as dandrader
dednickdidrocks: did you want me to merge with trunk (3255). there seems to be a daily release at 3254, but it's not mentioned in the changelog11:24
didrocksdednick: where do you see a daily release shipped to distro? the last one I'm seeing is rev 325211:25
dandraderSaviq, which dependency should I update to get rid of this? http://paste.ubuntu.com/5654895/11:25
Saviqdandrader, did you update unity?11:26
Saviqdandrader, see above - `bzr pull --overwrite -d ../unity_build/unity; ./build_unity`11:26
dednickdidrocks: ah. my bad. dont know how all this shipping things work. So what did you want me to merge at? 3255?11:27
didrocksdednick: 3252, but I guess this was already done :)11:28
didrocksdednick: the changelog is the authorative answer :)11:28
dednickdidrocks: yeah. it was done a few days ago. ok11:28
didrocksdednick: the merge back is done when we ship to distro11:28
didrocks(meaning, tests are running)11:31
didrocksdednick: like today, we have 40 failures on nvidia11:31
didrocksdednick: I think it's due to a service failing, (the hud most of the time) and not respawning11:31
didrocksso the tests are taking ages11:31
didrocksand all timeouts11:31
dandraderSaviq, ah, unity comes from lp:unity/phablet-mods now11:32
Saviqdandrader, it always did, only we rebased on lp:unity11:32
dandraderSaviq,  hmm, it seems I was using bzr+ssh://bazaar.launchpad.net/~anjali-team/anjali/unity.trunk/11:33
Saviqdandrader, ah, that was before11:33
dednickdidrocks: hm. ok. looks like the latest exp ppa ones are pretty good though.11:33
Saviqdandrader, drop the whole of ../unity then11:33
Saviqdandrader, to be sure, and ./build_unity -s11:33
didrocksdednick: yeah, I'm talking about unity trunk ones11:33
dednickdidrocks: sure11:33
didrocksdednick: it's really "random crash of hud (especially on nvidia) -> everything fails"11:34
=== tvoss is now known as tvoss|lunch
MacSlowSaviq, any idea... although I did "bzr pull --overwrite -d ../unity_build/unity; ./build_unity" I'm still getting this with current lp:unity/phablet http://pastebin.ubuntu.com/565491711:38
SaviqMacSlow, ./build, not straight cmake11:38
SaviqMacSlow, and make sure you ./unity_build11:38
SaviqMacSlow, that error means your UnityCore build failed11:39
SaviqMacSlow, just drop ../unity_build11:39
MacSlowSaviq, ok... trying again11:39
SaviqMacSlow, and go `./build_unity -s; ./build_unity` from a clean slate11:39
MacSlowSaviq, also took a first look at the notification-interface test... only found one minor thing11:42
SaviqMacSlow, why do we need to test for that hint?11:46
MacSlowSaviq, just a suggestion... it is implicitly covered by the Type.Confirmation... still it has to be passed by clients, which intend to trigger a confirmation/synchronous notification11:48
SaviqMacSlow, yes, an implementation detail that the shell does not need to know about11:48
MacSlowSaviq, ok11:48
SaviqMacSlow, imagine we move away from libnotify at some point11:48
SaviqMacSlow, we'd have to artificially set that hing11:49
Saviqhint11:49
MacSlowSaviq, true... safer to hide it then11:49
SaviqMacSlow, about that, what does the tint hint do?11:50
MacSlowSaviq, a way to indicate that the positive answer for a snap-decision is colored (Design wants this)11:51
SaviqMacSlow, yeah, that's fine, just wanted to know what it does11:51
MacSlowSaviq, is meant to be reserved for "system-level" snap-decision... regular apps should not use that hint11:51
SaviqMacSlow, k11:52
=== MacSlow is now known as MacSlow|lunch
=== _salem is now known as salem_
=== alan_g is now known as alan_g|lunch
=== MacSlow|lunch is now known as MacSlow
seb128cyphermox, hey, do you think you could trigger a raring landing of the indicators stack? there is an indicator-sync fix in there I would like to see in raring before the freeze tonight13:29
seb128(that's the 3rd most reported issue on e.u.c/raring)13:29
kgunnpaulliu: just checking in, saw you had a feature branch based on unity/phablet, you able to build ok now?13:48
paulliukgunn: no..that doesn't builld13:49
didrocksmhr3: \o/13:49
paulliukgunn: I'm still checking.13:49
kgunnpaulliu: pastebin your build probs if you need to...lots of folks can help13:51
paulliukgunn: http://paste.ubuntu.com/5654866/13:51
paulliukgunn: I've tried bzr pull --overwrite -d ../unity_build/unity; ./build_unity13:52
paulliukgunn: still not working.13:52
kgunnpaulliu: yeah...i had the same problem13:52
kgunnpaulliu: Saviq told me the same :) but the bzr pull didn't work for me either13:52
paulliukgunn: how did you solve your problem?13:52
kgunnpaulliu: i ended up rm -rf my unity-build dir13:53
kgunnpaulliu: then re-run ./build -s13:53
kgunnpaulliu: and that seemed to work13:53
paulliukgunn: ok, let me try it again. I thought I remove unity_build few hours ago.13:54
mterrydidrocks, slangasek was asking yesterday why we don't plan to just publish the touch stack directly to raring.  Apparently we have a standing FFe13:54
didrocksmterry:13:55
didrocksmterry: I'll talk to him, but reasons are:13:55
didrocks- we depend on hud 2.0 which will break existing desktop infra13:55
kgunnpaulliu: yeah, try again...let me know how you get on13:55
didrocks- I was busy with 100 scopes/in dash payment as I told him13:55
paulliukgunn: ok13:55
didrocks- libhybris and such needs to build on all arch13:55
mterrydidrocks, don't get your last comment, but OK.  I do understand hud2 issue13:57
mterrydidrocks, (Good EST morning btw!  :)13:57
didrockshey mterry ;)13:58
mterrydidrocks, regarding the touch stack, any issues with starting to enable some apps?   gallery app should be good to go13:58
=== alan_g|lunch is now known as alan_g
didrocksmterry: you are adding it to head/, right?13:58
mterrydidrocks, no, let me add it to configs13:59
mterrydidrocks, so raring is fully split off by now?14:00
mterrydidrocks, should I add a new stack?  touch-apps?14:00
didrocksmterry: not fully split yet14:00
mterryI recall, we were considering "touch-platform", "touch-apps", and one more?14:00
didrocksmterry: jibel is fixing a bug14:00
didrocksmterry: we didn't decide on the stack, I was waiting for sergio to get a list before his holidays14:01
didrockswhat I didn't get btw14:01
cyphermoxseb128: yeah, let's just make sure it's ok with didrocks14:01
cyphermoxdidrocks: indicator-sync landing ^ ??14:02
mterrydidrocks, I remember in the hangout, they said they split the packages along team lines.  They had a team for platform, one for apps, and one for something else.  That's where I got my idea I guess14:02
didrockscyphermox: good for me14:02
didrocksmterry: yeah, there is an existing platform one, let's use that one14:02
cyphermoxok, I was just making sure with the branch changes and all14:03
didrocksmterry: and add an apps one14:03
mterrysure14:03
didrocksmterry: we'll figure out later moving stuff14:03
mterrydidrocks, OK, pushed a bare new apps.cfg.  I have it disabled via to_transition right now, but if you like it, we could try enabling it14:11
mhr3didrocks, ?14:11
didrocksmterry: you think it would build?14:12
didrocksmhr3: on the "let's have the scope shutting down" :)14:12
mhr3oh14:12
mhr3i don't like it much though :P14:12
didrocks:)14:13
mhr3i'd rather have all python scopes in a single process and see how that works14:13
mhr3something to try over the weekend i guess :)14:13
Saviqpaulliu, let me know if you have any issues14:13
paulliuSaviq: ok.14:14
paulliuSaviq: BTW, just pushed a new commit to the merge request.14:14
Saviqpaulliu, cheers14:14
mhr3didrocks, is there any way to make bzr bd use parallel build?14:15
mterrydidrocks, gallery-app?  it should.  It needs stuff not in the PPA to run...  But it should build fine14:16
=== tvoss|lunch is now known as tvoss
tvosskgunn, ping14:17
didrocksmhr3: --parallel in debian/rules (the dh line)14:18
didrocksmterry: ok14:18
mhr3didrocks, any envvar for that? not sure i'd want it permanent14:18
mterrymhr3, if the package can handle parallel build, why not always?14:19
mhr3hmm14:19
mhr3good point14:19
didrocks:)14:19
mhr3didrocks, and since we have time to clean up now, how can i remove/hide all unity_internal_* symbols from the so?14:21
mhr3i remember there being a ldflag for it14:22
mhr3or was that a libtool thing?14:23
didrocksmhr3: it was a libtool14:23
didrocksthing14:23
didrocksthat we used in dee IIRC14:23
mhr3oh? /me checks14:23
kgunnMacSlow: hey, could you rebase your notification branch to latest trunk?....the build is giving me fits since unity update is needed14:33
jibeldidrocks, https://code.launchpad.net/~jibel/cupstream2distro-config/skip_downstream_task_if_no_project/+merge/15597514:33
didrocksjibel: thanks!14:33
MacSlowkgunn, I did that already and pushed 2 hours ago... rev 531 did you get that?14:34
kgunnMacSlow: :) my bad....i was trying last night...will pull14:34
MacSlowIs the script run_on_device still meant to be the way to get the qml-phone-shell from the dev-machine cross-compiled, put on the phone and executed?14:46
MacSlowof course running "./run_on_device -s" first to get it all setup14:46
=== tvoss is now known as tvoss|otp
greybackMacSlow: yes14:50
greybackMacSlow: where are things going wrong on you? Does the setup succeed?14:51
MacSlowgreyback, wait... jsut reflashed the phone... so it's missing rsync again14:52
greybackMacSlow: my usual procedure after a reflash is "phablet-network-setup -i, run_on_device -s"14:53
kgunnMacSlow: phablet-network-setup14:53
Saviqpaulliu, come here btw14:54
Saviqpaulliu, http://pastebin.ubuntu.com/5655340/14:54
Saviqpaulliu, we need a fix ApplicationManager.startProcess, though, to take the stage hint into account14:55
paulliuSaviq: but now we don't have stage hint in daemon?14:55
Saviqpaulliu, yes, the ApplicationManager needs to read it and launch the app correctly14:56
paulliuSaviq: ok.. Yes, since we have desktop file path in daemon, Qt should be able to parse the .desktop files.14:56
Saviqpaulliu, yeah, we just pass the desktop file name to the ApplicationManager and it should do its thing14:57
paulliuok.14:57
didrocksmterry: did you look at why unity can't publish btw?15:01
mterrydidrocks, guh, just more nvidia test failures.  Went up by 3015:02
didrocksmterry: yeah, I think it's the hud failing again15:02
didrocksmterry: so maybe try to restart it and get integration tests running?15:02
didrocksmterry: I'm blocked on knowing if I should redeploy head or not yet for unity :)15:02
mterryk15:02
mterryI can restart no problem15:03
mterrydidrocks, done; I wish these builds were more reliable15:03
didrocksmterry: I wish as well :)15:04
MacSlowgreyback, kgunn: working now again... mock notifications show up on the phone15:07
greybackcool15:07
kgunnMacSlow: cool!...now break it :)15:07
kgunnby adding more code that is :)15:07
MacSlowkgunn, nooooo :)15:07
MacSlowkgunn, sure15:07
kgunnMacSlow: reminds me of my favorite bug ever....working on audio for a phone, a real user submitted a bug15:08
kgunnMacSlow: saying after they submerged their phone in water15:08
kgunnMacSlow: it didn't make any sound :)15:08
MacSlowkgunn, btw... I'm off tomorrow and Monday... forgot to mention that in standup15:09
kgunnMacSlow: enjoy15:09
MacSlowkgunn, you honestly got that filed in your bug-db?15:09
kgunnMacSlow: it was a classic :)15:09
MacSlowkgunn, as long as the phone wasn't meant to be water-proof, how can people expect that to work15:10
kgunnMacSlow: i told them i fixed in sw & to please retest :)15:10
MacSlowkgunn, these days phone-hw is cheap anyway ;)15:10
kgunnSaviq: been talking to kalikiana about dash blueprint work item "merge DeeVariantText into dee-qt"15:13
kgunnbut looks like it out of date in lp, branch superceded15:14
kgunnis there something to be done? ....wonder if someone on strehl's team would need to own this?15:14
Saviqkgunn, yeah, the superseding branch needs review/fixing/merging https://code.launchpad.net/~unity-team/dee-qt/deevarianttext-and-tests/+merge/15353015:15
Saviqkgunn, the tests there are most useful15:17
Saviqkgunn, DeeVariantText... it depends whether GVariant string representation will be used anywhere passed to the shell15:17
Saviqkgunn, but all in all, yeah, it's the API team that needs to own it15:17
kgunnSaviq: thanks, i'll ping tstrehl15:18
Saviqkgunn, the DeeVariantText does not even belong with dee either15:19
Saviqkgunn, we just slapped it on because we needed it, and quick15:19
kgunnSaviq: where's the best "conceptual" home?15:19
Saviqkgunn, I don't think there is, yet, it would have to be some kind of qt<>gio compatibility layer15:20
=== alan_g is now known as alan_g|tea
kgunndidrocks: you joining ? https://plus.google.com/hangouts/_/47a4dfa53a9b99896e07a788f1bdf66c07bafd0115:33
didrockskgunn: sorry, 3 more minutes?15:38
kgunndidrocks: :) sure...we'll just talk about you until you're here15:39
Saviqpaulliu, if you're still around (go to sleep)15:40
paulliuSaviq: hmm..15:41
Saviqpaulliu, we're in testing mode, so just look in tests/{unittests,qmluitests}15:41
paulliuSaviq: ok15:41
=== alan_g|tea is now known as alan_g
Saviqpaulliu, so just find a component (start with the Components dir)15:44
Saviqpaulliu, that doesn't have tests15:45
paulliuSaviq: Yeah. I can take that one.15:47
paulliuSaviq: got it.15:48
kgunndidrocks: you want to do tomorrow? i have something in 10 min15:50
didrockskgunn: just joining now15:50
didrocksTrevinho: my obfsucated perl code!!! how dare you? :)16:12
didrocksTrevinho: well, it's still in perl, so still obfuscated I guess :p16:13
Trevinhodidrocks: eheh16:13
Trevinhodidrocks: I wrote it in-line, for what it worth... Then I thought it could be easier to maintain it in that form, don't you think? :)16:13
didrocksTrevinho: but but… I felt like a powa-user :-)16:14
didrocksTrevinho: more seriously, yeah, way much better, thanks!16:14
Trevinhodidrocks: you are16:14
didrocksTrevinho: I hope you did the change on a Friday16:14
didrocksTrevinho: that's my rule: always touching perl on a Friday :)16:14
didrocks(evening for bonus points)16:15
Trevinhodidrocks: ops, I was one day ahead :)16:15
didrocks:p16:15
didrocksTrevinho: more seriouly, the perl part looks good, I'll let someone else testing the rest :)16:15
Trevinhodidrocks: ok, thanks16:16
* Trevinho hates the change in DesktopAppInfo that needs a valid exec to return a valid GAppINfo..16:17
Trevinhoit makes testing so bad :/16:17
didrockscyphermox: in case you didn't see: http://10.97.0.1:8080/view/cu2d/view/Raring/view/Indicators/ failure on the -check16:17
cyphermoxdah16:17
didrocksTrevinho: are we impacted a lot?16:17
didrocksTrevinho: I had to fix some tests in libunity for that16:18
Trevinhodidrocks: well, also in BAMF16:18
didrocksTrevinho: it's in the 100scopes16:18
didrocksbranch16:18
Trevinhodidrocks: yes, I've seen it16:18
didrocksTrevinho: I'm using /bin/true as you saw then16:18
Trevinhodidrocks: however in bamf changing the exec to /bin/true changes the world16:18
didrocksdon't think about a better way…16:18
didrocksyeah16:18
didrocksseb128 was proposing to eventually revert it ^16:18
didrocksif we have too many issues16:18
Trevinhofor real world it's a good thing, but I don't like the fact that the app-info should be null... they could have been added just a getter16:19
didrocksyep16:20
didrocksTrevinho: do you think we'll get badly impacted, like what's your feeling?16:20
didrocksshould we revert this glib change16:20
didrocksor just fix whatever we need to fix16:21
Trevinhodidrocks: well, I was trying to fix the tests... And I can, but the actual result is different16:21
Trevinhodidrocks: I mean, in BAMF we have that a .desktop file has less or more priority based on the fact that its name matches the exec basename, and this is broken by that on tests16:22
didrocksTrevinho: ok, do you think it won't take too much of your time to fix that this release?16:22
Trevinhodidrocks: I'm not sure how I can fix it :)16:23
Trevinhodidrocks: let me check something16:23
Trevinhodidrocks: ok, I've faked it adding a local program and changing the PATH during the test16:29
didrocksTrevinho: sounds good to me16:30
=== alan_g is now known as alan_g|afk
didrockskenvandine: robru: mterry: cyphermox: so, FYI, apart from Unity, I redeployed everything. The "raring" view is now the one with daily releases.16:49
kenvandinedidrocks, cool16:50
didrockscyphermox: the indicator has the new branches for the 13.04 branch enabled, I just disabled daily as I want for head to run a full time one first16:50
=== alan_g|afk is now known as alan_g
didrockskenvandine: robru: cyphermox: so, think about looking at the "raring" view now16:50
didrocksand when upstream want to diverge trunk from a feature branch16:51
didrockshhttps://wiki.ubuntu.com/DailyRelease/MovingNewRelease#Diverging_.22trunk.22_and_a_.22maintenance.22_branch16:51
cyphermoxdidrocks: cool16:52
* cyphermox finishing up review of sil2100's branches16:52
cyphermoxdidrocks: symbols file for C++ ?16:53
didrockscyphermox: no, we can't have those :(16:54
didrockscyphermox: the symbols are different between archs16:54
cyphermoxright, that's what I thought16:54
cyphermoxjust drop them16:54
didrocksyep16:57
cyphermoxsil2100: did you see my comments?17:04
cyphermoxI'd like to see this land like, real soon17:04
cyphermoxesp. given that we only have 4 more hours17:04
mterrydidrocks, anything I can help with for 100scopes?  You know, I can create that MIR bug for ya, take some paperwork off your back17:14
didrocksmterry: ok, need to get you up to date with latest news17:14
mterrydidrocks, oh boy news17:14
cyphermoxsil2100: what PPA were you using to test the -qt stuff?17:22
mhall119didrocks: seb128: I remember last physical UDS there was discussion around automated performance testing of Ubuntu core17:39
mhall119do you have a system for that now?17:39
seb128no17:40
seb128I don't think much work went into that (yet)17:40
mhall119:/17:40
mhall119ok17:40
seb128mhall119, try talking the the qa team maybe?17:40
mhall119ok, thanks17:41
sil2100cyphermox: ppa:sil2100/qt mostly17:48
=== alan_g is now known as alan_g|wine
=== salem_ is now known as _salem
tigrangandy__ == andyrock?22:02

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!