/srv/irclogs.ubuntu.com/2012/11/08/#ubuntu-unity.txt

dmj_nova1Does anyone know exactly what widgets are supported in app indicators?00:22
thumperdmj_nova1: charles may be able to help there01:12
thumperdmj_nova1: but past his EOD I think01:12
thumperdmj_nova1: depending on your TZ, lars could help when he gets up (europe time)01:12
tsdgeosis there any way other than logginout+login to clear a program that obviously is not running from the launcher?09:17
dmj_nova1larsu: Do you know exactly what widgets are supported in app indicators?09:19
larsudmj_nova1, no, sorry, I don't09:19
Mirvtsdgeos: possibly alt+f2 unity and/or killing/restarting bamfdaemon09:30
tsdgeosMirv: yeah restarting unity worked, now let me see if i can reproduce how to get that ghost in there09:32
tsdgeoswhich obviously i can't :-/09:34
=== dbarth__ is now known as dbarth
=== mike_ is now known as Guest69661
=== mike is now known as Guest50749
tsdgeosMirv: what about this simple thingie? https://code.launchpad.net/~aacid/unity/initialize_using_nofilters_background/+merge/13229811:04
Mirvtsdgeos: looks good11:09
Mirvit's then read from style11:10
=== _salem is now known as salem_
=== mmrazik is now known as mmrazik|lunch
=== mmrazik|lunch is now known as mmrazik
=== MacSlow is now known as MacSlow|lunch
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== mpt_ is now known as mpt
ppdhi, quick question: Is the current shortcut behaviour of unity desired? I can't define shortcuts as a single keypress. I always need a modifier too. So is this by design or just a bug?13:28
=== MacSlow|lunch is now known as MacSlow
mterryfginther, hello!  So about the PPA upload issues: you were right about unity-lens-applications having out-of-date packaging, and I have a merge out there for that.  But compiz looks up to date to me14:00
fginthermterry, thanks for checking on compiz. I wasn't sure about that one, it just looked 'different'14:02
mterrydidrocks, what's the protocol for reviews?  How many approvals before the branch can be marked as approved?  (And does that happen automatically by a bot or manually?)14:09
didrocksmterry: one sec, in a hangout, be back then :)14:10
didrocksmterry: oh just one14:31
didrocksmterry: someone need to change the big switch though14:31
mterrydidrocks, OK.  I'll mark a couple build modernization branches then14:32
didrocksmterry: sure :)14:32
mterrydidrocks, and once the big switch is pulled, the bot takes over, right?14:33
didrocksmterry: I didn't push them for all as I wasn't sure that after the first merge, everything was all right on jenkins side14:33
didrocksmterry: yep :)14:33
didrocksmterry: btw, reproposed: https://code.launchpad.net/~didrocks/dee/add_python_3/+merge/13349114:33
didrockshey bregma14:36
bregmahey didrocks14:36
didrocksbregma: did you see bug #132978?14:36
ubot5Launchpad bug 132978 in Open Library "timestamp is missing the tdb.log" [Critical,Fix released] https://launchpad.net/bugs/13297814:36
didrockshum14:36
didrocksbug #1076129 rather14:36
ubot5Launchpad bug 1076129 in Nux "nux autolanding builds encounter make check-headless failures on i386 builds" [Undecided,New] https://launchpad.net/bugs/107612914:36
bregmasaw it, looks like some bad assumptions somewhere14:37
=== dandrader is now known as dandrader|lunch
didrocksbregma: do you have someone working on it?14:37
didrockseither fixing the test or the code :)14:37
didrocksmterry: looking at libunity again, one sec14:38
bregmamy team is studiously avoiding touching nux except where necessary to fix our bugs14:38
bregmaso we don't tread on other people's toes14:38
bregmaso no, I don;t have anoyone looking at that issue14:39
mterrydidrocks, bregma: I looked at that, it's a floating point rounding error14:39
mterryNot sure how to fix yet14:39
didrocksbregma: didn't we assign someone to nux in your team? is anyone looking at it?14:40
didrocksmterry: at STR_EQUAL? comparaison?14:40
didrocksbregma: I'm concerned that nobody in the 13.04 team isn't actively looking at nux14:40
bregmadidrocks, nobody is on nux in my team because of INFORMATION REDACTED14:40
mterrydidrocks, the rounding isn't happening in the test code, but rather in (for at least one of the failures) AnimateValue.14:40
didrocksmterry: ah…14:41
bregmabugs in nux need to go to Tim's team right now14:42
mterrydidrocks, so it's not as simple as switching EQ to FLOAT_EQ or whatever.  And it happens in template-based code so it's hard to specifically say "round from float to int"14:42
didrocksmterry: ok, it was my bet14:43
bregmasounds like the animation code has some serious flaws, 'cos floating point results can be very different on arm vs. intel, too14:44
didrocksmterry: removed the override and pushed, that's what happen when you try different parameters to filter and nothing worked :p14:46
didrocksmterry: thanks14:50
mterryfginther, https://code.launchpad.net/~mterry/dee/modern-build/+merge/133344 seems like a recipe failure14:54
fginthermterry, you are correct. It should be fixed now and re-approved.14:56
didrocksmterry: so failure on libunity here14:56
mterrydidrocks, :(14:56
didrocksmterry: oh, I build with -j…14:57
didrockslet me look at the exec first14:57
didrocks(shouldn't impact tests though)14:57
didrocksls -l ./test/python/bug-1062331.py14:58
didrocks-rw-rw-r-- 1 pbuilder pbuilder 263 Nov  8 14:40 ./test/python/bug-1062331.py14:58
didrocksmterry: ^14:58
didrocksso definitively not executable in my pbuilder14:58
mterryhm14:58
mterrybecause I think it tries to exec them directly14:58
mterryyeah it does14:58
didrocksyeTESTS = bug-1062331.py extras.py14:59
didrocksso yeah, that's what should happen14:59
mterrydidrocks, so if they aren't executable, that's your problem.  Are they executable outside the pbuilder?14:59
didrockshum, they are14:59
didrocksin the branch I took from you15:00
didrocksbut not when I bzr bd --builder pdebuild15:00
didrockshow do you use pbuilder? you don't have that issue?15:00
didrocksmterry: (and python3 was already as build-dep, I think you saw that changing back the status :p)15:02
mterrydidrocks, yeah I think I commented too15:03
mterrydidrocks, sorry, brain fair15:03
mterrybrain fart even15:03
didrocksmterry: still relying on email and launchpad is slower on that :)15:03
didrocksmterry: no worry! you did some amazing job with all those MR! :)15:03
mterrydidrocks, thanks.  Not done yet though!  :)15:04
didrocksa lot done :)15:04
mterrydidrocks, about libunity, I use my pbuilder-scripts, which call pbuilder directly without bzr bd15:04
didrocksmterry: ah, maybe that's the difference, let me try15:04
didrocksah, doesn't work so well with split mode though :p15:05
didrocksmterry: I'm afraid the same will happen on the merger then15:05
didrocksmterry: same with plain pdebuild for me, do you know of anything that can strip those? I looked again at my .devscript and .pbuilderrc but don't see anything15:13
mterrydidrocks, no.  it's in the upstream tarball which normally (like, via 'make dist' style) keeps them15:14
mterrydidrocks, I mean, "in the upstream source"15:14
mterrydidrocks, wait, I use pdebuild too15:14
didrocksyeah, I wonder why I'm seeing something different here…15:14
didrocksmterry: you pbuilder are created with the buildd environment, right?15:15
mterryyes.  but that shouldn't affect this I shouldn't think15:15
didrocksI don't think so as well, just trying to find the eventual difference :)15:16
=== dandrader|lunch is now known as dandrader
didrocksgrrr, when you resubmit a MR, it ignores the commit message…15:21
didrocksmterry: seems seb128 has the same issue with pdebuild. Do you have time to look at why? (I can look, but surely not today, maybe tomorrow morning)15:29
seb128just for the record I didn't use pdebuild15:30
seb128but I logged into pbuilder and did a debuild manually there (like to do that because it doesn't wipe my env when the build ends)15:30
didrocksseb128: brz bd --builder pdebuild isn't the command I gave you?15:30
mterrydidrocks, I can try...  I have to do sponsoring today and I'm out tomorrow, so might not get to it, but no biggie.  Just the libunity merge is held up by it15:30
didrocksah ok :)15:30
seb128didrocks, well, I assumed you wanted a build in pbuilder rather than the exact command ;-)15:30
seb128didrocks, so I sticked to my usual workflow ;-)15:30
didrocksmterry: no worry, just drop me an email if you can't do it today, so that I remind to do that tomorrow :)15:31
didrocksseb128: well, I think that's close enough to pdebuild anyway15:31
seb128yeah15:31
* didrocks goes back to the automation/jenkins/launchpad party15:32
mterrydidrocks, well no...  pbuilder won't magically destroy the executable bit, but when pdebuild bundles up the source, it might very well15:32
mterrydidrocks, seb128: so those two very well might give different results15:32
mterry(I mean, logging into pbuilder won't destroy the executable bit)15:32
seb128# ./test-vala15:33
seb128...15:33
mterryunless you only have +x for your user/group and not root?15:33
seb128Error spawning command line `dbus-launch --autolaunch=59e939bd4642e27890eb8f84509bcf30 --binary-syntax --close-stderr': Child process exited with code 1 (g-spawn-exit-error-quark, 1)15:33
seb128unity-launcher.c: line 125015:33
seb128didrocks, mterry: running tests under dbus-test-runner seems to work15:35
seb128you should do that15:35
mterryoh...  so your error wasn't related to the executable bit?15:36
seb128mterry, no, mine it's "Unable to connect to session bus: Could not connect: Connection refused"15:36
mterryThree people, three different results.  :)15:36
seb128mterry, indicator-sync already use dbus-test-runner if you want an example15:37
seb128override_dh_auto_test:15:37
seb128        dbus-test-runner -m 300 -t dh_auto_test15:37
seb128basically15:37
mterrysure...  I don't see that error so I didn't think it was needed...15:37
seb128mterry, do you have a session bus in your pbuilder?15:37
mterryseb128, I don't know why I would15:38
seb128mterry, it's weird that the tests work without one for you15:38
mterryseb128, agreed (I didn't notice they needed dbus so I didn't ask questions).  I'll look at the branch again later15:39
seb128didrocks, mterry: vala tests success for me with15:40
seb128override_dh_auto_test:15:40
seb128dbus-test-runner -t dh_auto_test15:40
seb128but then the python ones fail15:40
seb128/bin/bash: line 9: ./bug-1062331.py: Permission denied15:40
seb128task-0: FAIL: bug-1062331.py15:40
dmj_nova1charles: Do you know what widgets can be put in app indicators?15:40
seb128same for extras.py15:40
mterryseb128, aha!  that was the permissions issue that we were looking into15:40
mterryseb128, it seems that the executable bit is lost for some reason15:41
seb128you shouldn't rely on it, just call python test.py...?15:41
seb128mterry, didrocks: well "the executable bit is lost for some reason" is an obvious reason15:43
seb128mterry, didrocks: you have those files in the diff.gz and diff doesn't handle permissions15:43
seb128dpkg-source: warning: executable mode 0775 of 'test/python/bug-1062331.py' will not be represented in diff15:43
mterryseb128, why are they in diff.gz?  they should be in orig.tar.gz15:43
mterryseb128, does split mode put everything in a diff?15:43
seb128mterry, because15:44
seb128$ tar tzf libunity_6.10.0.orig.tar.gz | grep test | grep python15:44
seb128$15:44
seb128mterry, where do you got your tarball?15:44
seb128the archive one doesn't have python tests15:44
mterryseb128, I just run pdebuild15:44
mterryseb128, on a bzr checkout15:45
seb128did you get*15:45
seb128mterry, I guess it's building the pristine tarball from the 6.10.0 tag15:45
seb128or getting it from the archive15:45
seb128mterry, in any case 6.10.0 doesn't contain those files, that's why they are in the diff15:45
mterryseb128, I bet it's recreating it from scratch, because I don't have the 6.10.0 orig tarball hanging around15:45
mterryseb128, OK.  So executable thing is a non-issue then.  didrocks ^15:46
mterryseb128, which doesn't explain why you got dbus issues while didrocks and I did not15:46
seb128mterry, it should be just fine once you do a release with those in the tar15:46
mterryseb128, yeah, and for the moment, we're concerned about automated tests from trunk, where it also won't be an issue15:46
seb128mterry,15:48
seb128Program received signal SIGTRAP, Trace/breakpoint trap.15:48
seb1280xb7ca53fd in g_logv () from /lib/i386-linux-gnu/libglib-2.0.so.015:48
seb128(gdb) bt15:48
seb128#0  0xb7ca53fd in g_logv () from /lib/i386-linux-gnu/libglib-2.0.so.015:48
seb128#1  0xb7ca5583 in g_log () from /lib/i386-linux-gnu/libglib-2.0.so.015:48
seb128#2  0xb7f7f0f3 in unity_launcher_entry_constructor (type=134653544,15:48
seb128    n_construct_properties=1, construct_properties=0x806f9b0)15:48
seb128    at unity-launcher.c:125015:48
mterryseb128, that's from the dbus failure?15:49
seb128yes15:49
seb128can you run "dbus-monitor --session" from your pbuilder?15:49
didrocksseb128: urgh, I didn't see that even if I modify the versioning to 6.12 to discare that case, it did bring 6.12 from an old test release I've done locally so was in the diff.gz again… :/15:49
didrocksso I thought about it, but was trapped by a local tarball and didn't notice it was still in the diff :/15:50
seb128didrocks, no worry, glad we sorted it out ;-)15:50
mterryseb128, no, because dbus-monitor isn't installed by default (will install now and check).  Though I do notice that DISPLAY=:015:50
didrocksyeah, thanks for noticing it :)15:50
didrocksmterry: so just bump to 6.11 the changelog and we'll be fine for this one15:50
seb128mterry, same here... well dunno why dbus is working for you, it shouldn't in a pbuilder15:50
seb128mterry, but anyway, if the tests work no worry, if they don't just add dbus-test-runner15:51
mterryseb128, well, dbus-x11 is a build-dep, so it should be pulled in and started if needed?15:51
seb128unity-launcher.c: line 1250: unexpected error: Error spawning command line `dbus-launch --autolaunch=59e939bd4642e27890eb8f84509bcf30 --binary-syntax --close-stderr': Child process exited with code 1 (g-spawn-exit-error-quark, 1)15:52
seb128 15:52
mterrydidrocks, why bump the changelog?  Won't the auto-builders always use a fresh tarball?15:52
seb128not sure why the autospawning doesn't work15:52
didrocksmterry: no, it will reconstruct the tarball for already published version15:52
seb128mterry, though15:53
seb128# dbus-launch --autolaunch=59e939bd4642e27890eb8f84509bcf30 --binary-syntax --close-stderr nano15:53
mterrydidrocks, how will we handle auto-built versioning in future?  We were going to have automatic changelog entries and such right?15:53
seb128No protocol specified15:53
seb128Autolaunch error: X11 initialization failed.15:53
seb128#15:53
seb128mterry, here15:53
didrocksmterry: yeah, basically the versionning will be:15:54
didrocks<major_upstream>dailydd.mm.yy(.1…)15:54
didrocksso always a new version15:54
mterrydidrocks, branch updated, try now15:54
seb128mterry, "xvfb-run dbus-launch ./test-vala" works15:55
seb128well anyway, letting you guys continue on that, if you are interested by extra debug infos let me know15:55
mterryseb128, I will look at it later, I really should do my sponsoring shift or fear the wrath of dholbach15:55
seb128hum, I should do some sponsoring as well, I missed my shift during UDS and that would make Daniel happy15:57
didrocksmterry: built! approving :)16:02
mterrydidrocks, yay!  I hope the dbus thing is just a weird seb-specific issue16:02
didrocksmterry: I guess we'll see soon :)16:03
seb128didrocks, mterry: in any case it's easy to workaroud with dbus-test-runner16:04
mterryyar16:04
conscioususermpt: le ping16:08
=== gatox is now known as gatox_lunch
mterryfginther, is https://code.launchpad.net/~mterry/libunity/enable-tests/+merge/133350 jenkins failure just a recipe not being updated?16:27
fginthermterry, yes. I fixed the job and re-approved the MP16:30
mterrythanks!16:30
tsdgeos_Mirv: there's also the 6.0 branch if we are interested https://code.launchpad.net/~aacid/unity/initialize_using_nofilters_background_6.0/+merge/13229716:42
mterryfginther, again with that libunity update, though a different error17:04
fginthermterry, yep, I saw it also. I patched the scripts to resolve the issue. Hopefully it will go through now17:05
fgintherI'll continue to watch it17:05
mterrythanks!17:06
=== gatox_lunch is now known as gatox
=== dandrader is now known as dandrader|afk
fginthermterry, one more time...17:27
mterry:)17:27
=== dandrader|afk is now known as dandrader
fginthermterry, libunty MP failed again, but this time it's a unit test failure, not a jenkins infrastructure issue17:58
mterryfginther, yay?17:59
mterryfginther, I'll look into it in a bit17:59
fginther:-P one step closer17:59
conscioususermpt, you online?18:30
fgintherHello, unity-team/staging is broken. Can someone help me figure out why?20:53
fgintherlog is here: https://launchpadlibrarian.net/122383545/buildlog_ubuntu-quantal-amd64.unity_6.10.0bzr2884pkg0quantal0_FAILEDTOBUILD.txt.gz20:53
fgintherI think some CMakeLists.txt changes in rev 2884 might be the cause20:54
=== salem_ is now known as _salem
fgintherandyrock, bschaefer, charles, thumper, anybody ^^ ?21:07
thumpermorning fginther21:07
fginthergood afternoon, sorry for the blast-o-gram21:08
fgintherI think the unity broken-ness could be impacting all builds21:09
* thumper is looking at bzr log21:09
fginther2884 is the last merger21:10
thumperfginther: is libgtest-dev listed as a build dep?21:13
fgintherthumper, no21:13
thumperfginther: it looks like what changed was something that was looked for with a test include now has a default value21:13
thumperfginther: so the search after doesn't find gtest and falls back to the default21:14
thumperthen  it tries to include a src dir which doesn't exist21:14
thumperhence the boom21:14
thumperfginther: can you add libgtest-dev and google-mock as build deps?21:14
fgintherthumper, to lp:unity?21:15
thumperyeah21:15
fginthersure21:16
thumperfginther: alternatively you could comment out those 5 lines added to the base CMakeLists.txt21:16
fgintherthumper, that sounds like a better alternative for now21:18
thumperfginther: that should get things working again21:18
thumperand we can add the build depends after21:18
andyrockfginther, hey21:26
fgintherandyrock, hello21:27
fgintherthumper helped me out21:27
andyrockok sorry i was afk21:27
* bschaefer just got back from lunch21:30
fgintherthumper, can you take a look please: https://code.launchpad.net/~fginther/unity/unity-revert-gtest/+merge/13357622:03
bschaeferfginther, I can approve it if you need22:11
fgintherbschaefer, thanks22:12
bschaeferfginther, that is fixing the issue?22:12
fgintherI consider it a workaround. a better fix is needed (I'm creating a bug now)22:12
fgintherbut yes, the build proceeds past the point it was failing22:13
bschaeferfginther, hmm alright, ill confirm everything compiles and Ill merge it22:13
fgintherbschaefer, thanks!22:14
bschaeferthough nux has those dependencies22:14
bschaefershouldn't unity have those as well?22:14
bschaeferfginther, ^22:15
fgintherbschaefer, are you sure nux has libgtest-dev and google-mock as build deps? I don't see it in lp:nux22:16
bschaeferhmm I know I use to have to manually install those22:16
bschaeferand a fresh install otherwise nux wont compile its tests22:16
bschaeferand I know nux uses google=mock for its tests...22:16
fgintherbschaefer, that makes sense. most packages I'm familar with don't have dependencies on libs that are purely fur executing tests22:17
bschaeferoo that would explain that then22:17
thumperfginther: approved22:17
TheMusoWhat is the best list to send mail to if I want to discuss something about unity technically?22:17
TheMusoAs in the technical workings of unity22:17
thumperfginther: nux should have them as build deps22:17
thumperfginther: they are needed for running the tests in make check22:17
bschaeferthumper, I was about to!22:17
thumperbschaefer: bet you to it22:17
bschaeferthumper, also how was your flight?22:17
thumperlong and uneventful22:17
bschaefersounds like mine22:18
bschaeferbut half has long22:18
thumper:)22:18
bschaeferhalf as*22:18
thumperfginther: the nux branch you reapproved failed landing again22:23
* fginther looks at nux branches...22:31
fgintherthumper, nux is failing due to https://bugs.launchpad.net/nux/+bug/1076129. I'm working with Jay to reproduce outside of jenkins22:35
ubot5Ubuntu bug 1076129 in Nux "nux autolanding builds encounter make check-headless failures on i386 builds" [Critical,Triaged]22:35
thumperfginther: ok, ta22:35

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