/srv/irclogs.ubuntu.com/2015/11/20/#ubuntu-desktop.txt

=== mhall119 is now known as mhall119|fossetc
=== happyaro1 is now known as happyaron
=== maclin1 is now known as maclin
pittiGood morning04:06
TheMusoHey pitti, how are you?04:15
pittihey TheMuso! a bit tired still, but good, thanks! how about yourself?04:25
TheMusopitti: Not too bad thanks, glad to have air conditioning here today, 40+ degrees in Sydney, can't wait for the southerly cool change so I can open my windows again. :)04:27
pittiurgh, steaming04:27
TheMusoPretty much.04:27
pittiLaney: meh, more weird MemoryErrors and segfaults -- time to reboot this box04:33
pittiLaney: maybe the piled up ksplices eventually made the kernel's head explode04:33
darkxstpitti, how did this manage to migrate through proposed? bug 151803505:51
ubot5bug 1518035 in libjpeg9 (Ubuntu) "package libjpeg-progs 1:9a-2ubuntu1 failed to install/upgrade: trying to overwrite '/usr/share/man/man1/djpeg.1.gz', which is also in package libjpeg-turbo-progs 1.3.0-0ubuntu2" [Undecided,New] https://launchpad.net/bugs/151803505:51
pittidarkxst: we don't do upgrade tests as part of proposed-migration05:52
pittiapparently a fresh install works05:52
darkxstpitti, right that makes sense05:57
didrocksgood morning06:06
pittibonjour didrocks !06:14
hikikoMorning !06:20
didrocksgood morning pitti, hikiko!06:27
pittiLaney: dbus tests are reeeeally unhappy, the root tests keep timing out06:39
* pitti cancels the currently running ones, no point06:39
seb128good morning desktopers06:52
seb128hey didrocks, pitti06:52
seb128pitti, Laney is off today :-/06:52
pittibonjour seb128 !06:52
seb128though I would not be surprised if he reads IRC06:52
didrocksre seb12806:57
didrocksLaney: preventively: "go away"! :)06:58
pittiok, that should have cleared a lot of -proposed stuff08:04
seb128pitti, "that"?08:06
pittiseb128: retrying tests, adding hints, retrying some failed builds08:07
seb128ah, ok08:07
pittia lot of arm64 builds failed in a weird way08:07
* seb128 wonders what transitions is the solid block at the bottom of update_output08:07
pittiseb128: and bringing moah powah to my clouds, so the queues finally cleared up \o/08:07
pittiseb128: isn't that just the concatenation of everything it tries to migrate?08:08
pittithat "autohinter" thing08:08
seb128does it always does that?08:08
pittithere's certainly a bunch of valid candidates which wait for finishing lib transitions, like imobiledevice08:09
seb128libimobiledevice seems to miss a upower rebuild08:09
seb128pitti, did you have any content to include for that one or should we go for a no change?08:10
pittiah, easy; want me to, or are you on it?08:10
seb128I can do it08:10
pittiheh08:10
seb128just wanted to check if you had actually patches pending or something08:10
pittiseb128: I suppose your "upload a no-change rebuild" script is as good as everyone else's :)08:10
seb128doing it then ;-)08:10
pitticheers08:11
seb128yw!08:11
larsuhappy Friday!08:21
seb128hey larsu, happy friday!08:21
seb128wie gehts?08:21
Trevinhomorning08:23
seb128hey Trevinho!08:27
Trevinhohey seb12808:27
seb128how are you?08:28
larsuseb128: gut danke. und idr?08:28
larsu*dir08:28
didrockshappy Friday larsu, Trevinho!08:28
larsuhi Trevinho! Sup?08:28
seb128larsu, auch gut, danke :-)08:28
larsubonjour didrocks08:28
Trevinhohi larsu .... Trying to fix my broken head... :-/08:38
larsuTrevinho: oh what's up?08:40
Trevinholarsu: still the same.... I've not been able to fix myself yet. Trying to fight with nightly panic attacks.08:42
larsuTrevinho: sorry to hear. All the best!08:42
Trevinholarsu: thanks08:42
pittihey larsu!08:43
pittihello Trevinho08:43
* seb128 just spent ~1.5 hours going through the 300 or so "most recently changed" bugs on launchpad, enough triaging for today08:43
seb128(trying to do a bit of that every day since wily to keep on top of new issues reported/important ones)08:44
larsumorgen pitti! Wie gehts?08:44
Trevinhohi pitti08:44
larsuseb128: good work! Thanks08:44
pittilarsu: prima, danke!08:44
seb128:-)08:44
seb128shotwell is another unhappy program because of the glib file monitor improvements :-/08:45
seb128segfault when editing photos with an external editor08:45
* pitti hugs seb12808:45
* seb128 hugs pitti back08:46
larsuseb128: "improvements"08:49
seb128larsu, yeah :-)08:50
* didrocks will not get the card of backward compatibility in term of behaviors :p08:52
larsudidrocks: that would mean we could never fix behaviour bugs08:52
larsuthere's a balance08:52
larsubut yeah ... this situation is breaking all kinds of things08:52
didrockslarsu: well, some frameworks do that and add new calls08:53
didrocksor flags to enable the new "fixed" behavior08:53
didrockslarsu: the extreme would be to fix "this" in javascript for instance :p08:53
didrocksit's the kind of things we do, and thus, the result in crashing apps…08:53
larsuindeed08:56
larsulike I say, it's a balance :)08:56
larsuseb128: which nautilus version do you need the menubar patch for? Will we stay on 3.14?08:58
seb128larsu, 3.1808:58
larsunice :)08:58
seb128larsu, no, if we were to stay on 3.14 we wouldn't need a patch since that has a menubar :p08:58
larsunot for me :/08:58
seb128well or at least our package has one08:58
seb128but it's built on old code08:59
larsuah, restart08:59
seb128which got rewriten to use moder gaction, etc08:59
larsuweird08:59
larsuright right08:59
larsuwoah, got asked about which login manager I want to use, again09:06
larsuand I need to restart nautilus after logging in before I have correct fonts or the global menubar09:12
* larsu blames xsettings not being there from the beginning09:12
didrockspitti: do you know well how subprocess and os.environ interacts? I'm really puzzled on some weird behaviors when running tests…09:12
larsuhm firefox theme as well09:14
larsuis anyone else seeing this?09:14
seb128larsu, I don't09:20
seb128but I didn't upgrade yet today09:20
happyaronwhere is the latest snappy image?09:22
seb128happyaron, #snappy should know09:25
pittididrocks: normally the spawned process should inherit the parent's env, unless you explicitly set env=09:37
happyaronseb128: thanks, :)09:37
didrockspitti: agreed, it's not what happens here though09:37
didrockspitti: I have reduced it to few lines, and it behaves different if I have one test class or multiple…09:38
didrockspitti: if you have a minute, I have a gist to reproduce the issue: https://gist.github.com/didrocks/ae12d61138090b837c1809:56
pittididrocks: oh, are you sure that assigning the whole os.environ array works?09:57
pittididrocks: I'm not sure, but I faintly remember having run into this before too09:58
pittididrocks: untested, but can you try instead:09:58
didrockspitti: it does from the output, like the output doesn't contain http_proxy for instance09:58
pittios.environ.clear()09:58
pittios.environ.update(self.original_env)09:58
didrocksah09:58
didrockspitti: but it's still weird that a impacts b, isn't it?09:59
pittii. e. keep the os.environ dict object itself unchanged, but only change its keys09:59
didrockslet me try09:59
didrocks(I can see that SHELL is restored though)09:59
didrockspitti: yeah, it passes now :)10:00
didrocksI'm still unsure what happens under the hood though10:00
didrocksohhhhhhh, I see10:00
didrocksso, what happens is that subprocess still reference os.environ, the original one10:00
pittididrocks: os.environ is "magic" -- changing the reference to a different object apparenlty loses that magic10:00
didrocksnot the one we replaced10:00
pittiand you assign it to a normal "dumb" dict10:00
pittiright10:01
pittiyou shouldn't change the object itself (i. e. the reference to it)10:01
didrocksthat's why it still got the env variable only if a ran10:01
didrockshappened before*10:01
didrockspitti: it all makes sense now :)10:01
didrocksthanks a bunch, I was really fighting on that one (first, to reduce at that state)10:01
didrocksand so, if I only ran b, there was no issue, because I didn't touch os.environ assignement before10:02
didrocksstill weird that in this magic, when I changed b, it was changing the new os.environ that I assigned10:02
didrocksbut subprocess keeps using the old os.environ10:02
didrockseven if not called before10:02
pittiself.original_env = os.environ.copy()10:03
pittididrocks: ^ maybe the .copy() just returns an ordinary dict, not the special "magic" one that actually calls setenv()?10:03
pittibut I don't know for sure10:03
didrockspitti: that would make sense, indeed10:04
pittijust as far as "never assign os.environ", got burned and only remembered the "don't do that then"10:04
didrocksso we don't have anymore _Environ10:04
didrocksand so, it doesn't setenv()10:04
didrocksahah :)10:04
didrockspitti: rightly noted now!10:04
didrocksthanks a lot10:04
* didrocks hugs pitti10:04
pittide rien !10:05
Laneypitti: ERROR:../../test/uid-permissions.c:129:test_uae: assertion failed (dbus_message_get_type (m) == DBUS_MESSAGE_TYPE_METHOD_RETURN): (3 == 2) <- there's a related UpdateActivationEnvironment change in this release - check with smcv?10:26
Laneymight be that this should fail now10:26
Laney(3 is error)10:27
seb128hey Laney10:32
Laneyhi seb12810:37
* Laney is looking at cake recipes10:37
Laneywhat's up?10:37
seb128how did the cake go?10:37
Laneynot made yet10:37
seb128oh, not done yet :-)10:37
seb128new gtk migrated!10:37
Laneyso I saw10:38
Laneyyay cloud10:38
Laneysomeone can follow up with that dbus thing if they want or I can look on monday10:40
seb128if somebody things it's urgent I can have a look10:46
seb128otherwise I think it can wait monday10:46
Laneyhttps://bugs.freedesktop.org/show_bug.cgi?id=9303610:51
ubot5Freedesktop bug 93036 in core "/uid-permissions/uae/root test fails with 1.10.4" [Normal,New]10:51
* Laney goes to buy more ingredients now10:51
Laneylater!10:51
seb128Laney, have fun!10:51
seb128oh, good, libimobiledevice transition went through10:52
Sweet5harkhmmm, no will.cooke in here today?10:52
seb128he said he would swap the morning10:52
seb128he should be there after lunch10:52
seb128speaking of lunch I start being hungry ;-)10:52
* didrocks does as well, going out for lunch, see you later guys :)10:53
seb128didrocks, enjoy!10:53
didrocksthanks, you too seb12810:53
pittiLaney: FYI, memory errors seem to be real; committed https://git.launchpad.net/~ubuntu-release/+git/autopkgtest-cloud/commit/?id=17776035 and will re-deploy once the queue settled down11:25
desrthello desktop-types.  i hope you are having a good morning.12:26
seb128hey desrt, happy friday!12:40
desrtindeed!12:40
=== hikiko is now known as hikiko|ln
=== alan_g is now known as alan_g|lunch
=== dpm is now known as dpm-afk
=== hikiko|ln is now known as hikiko
didrocksphew, all tests passed, finally! :p13:56
pittididrocks: \o/13:56
didrockspitti: thanks again! And with the large tests refactoring, I divided the testing time by two!13:57
didrocks(~50 minutes vs 1h40)13:57
pittiwow13:57
=== alan_g|lunch is now known as alan_g
* didrocks celebrated with pushing 3 new tests14:14
=== dpm-afk is now known as dpm
didrockstkamppeter: hey, why is hplip depending on pexpect? It seems just to be for testing, right? So build-dep, but not runtime dep?14:51
didrockshum, seems that hplip has a pexpect embeeded version…14:52
didrocksand the dep was added in 2009 with the idea of removing the embeeded version14:53
didrockswhich wasn't done in 5 years :p14:53
didrocks6*14:53
didrocksok, they are using pexpect in utils.py for handling some subprocess and giving password and such to it (urgh)14:54
tkamppeterdidrocks, this is not done by me but by the Debian folks, perhaps a Debian bug should get reported.14:56
didrockstkamppeter: mind opening one? are you maintaining this stack as well? :)14:56
didrockstkamppeter: sounds like this kind of things shouldn't even be in main14:56
didrocksmaybe I should ask for some security team review, but passing password to subprocess sounds like something I would have kicked the MIR out14:57
didrocksseb128: mind creating a https://launchpad.net/~ubuntu-desktop/+archive/ubuntu/ubuntu-make-builddeps ppa?15:02
didrocksI want to ensure that testing/build-deps are not installed on user's machine if they add the ubuntu make ppa15:02
didrocks(like… pexpect that we install by default :/)15:03
seb128didrocks, done15:03
didrocksthanks!15:03
seb128yw!15:05
didrocksSweet5hark: it seems that willcooke is here. I think you were looking for him ^15:18
didrocksbtw, good morning willcooke :-)15:18
willcookemorning15:18
Sweet5harkheya! :)15:31
Sweet5harkwillcooke: was just wondering about the lack of feedback on the 2oitg thing ... but I assume, if this is "morning" you are sprinting somewhere in the US?15:32
willcookeSweet5hark, just swapped some hours this week15:32
willcookeSweet5hark, seen you mail, but not digested it yet15:33
attentehow does someone grep the archive?16:29
didrocksattente: what do you mean, grepping through all the sources ?16:31
attentedidrocks: yeah. i think Laney had a web interface for it, but the url doesn't work any more16:32
didrocksattente: oh, I don't know about a web interface. archive admins can grep directly in the source file system, but that's it16:32
didrocksattente: if you only look for file names, though, you can do it throuh a web interface16:33
attentedidrocks: could you do a couple greps for me? i'm looking for QT_IM_MODULE and maliitphablet16:34
didrocksattente: quite late to do that (I'm leaving in ~10 minutes), could do on Monday for sure16:34
didrocksor if seb128 is still there for some times, he might (or wait for Laney if he thinks there is a web interface)16:34
attenteah, ok, sure. have a good w.e!16:34
didrocksthanks (not left yet, but soon) ;)16:35
seb128I don't know of an handy way no16:35
seb128jdstrand had a script I think16:35
seb128Laney had a https://codesearch.debian.net/ equivalent up somewhere16:35
seb128but unsure if that's still a thing16:35
attenteyeah, i think it used to be http://ubuntu-codesearch.surgut.co.uk/, but it hasn't been online in a while16:36
* didrocks goes, see you on Monday guys!16:54
willcookesee ya didrocks16:54
didrockshave a good week-end willcooke :)16:54
=== popey_ is now known as popey
seb128have a good w.e!17:09
willcookesee ya seb12817:11
seb128see you on monday ;-)17:11
Trevinhoguys, have nice weekend !17:19
willcookesee ya Trevinho17:20
Trevinhowillcooke: see u17:20
=== alan_g is now known as alan_g|EOD
willcookeg'night19:13

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