/srv/irclogs.ubuntu.com/2015/02/04/#ubuntu-desktop.txt

pittiGood morning06:40
willcookeGood morning all07:58
tsdgeosguys my indicator-time does randomly not start every say 10 or 20 times08:13
tsdgeoswhen i log in08:13
tsdgeosshitty bug report i know :/08:13
* willcooke has seen this randomly as well08:13
seb128tsdgeos, maybe check .cache/upstart log for errors?08:14
seb128what is the status of the upstart job when that happens? is the binary running?08:14
larsujournalctl!08:14
tsdgeosbinary is not running08:14
tsdgeosstatus of the job not sure since i just started it manually ^_^08:16
tsdgeoswill try next time08:16
seb128k08:16
seb128nothing in the log?08:16
tsdgeosthere's08:17
tsdgeosIndicator-Datetime-Message: indicator-datetime exiting; failed/lost bus ownership08:17
tsdgeosg_dbus_connection_real_closed: Remote peer vanished with error: Underlying GIOStream returned 0 bytes on an async read (g-io-error-quark, 0). Exiting.08:17
tsdgeosbut is not from today but yesterda08:17
tsdgeos-rw-r----- 1 tsdgeos_work tsdgeos_work 197 feb  2 18:52 indicator-datetime.log.1.gz08:17
larsutsdgeos: woah, dbus is stopped before the indicator08:20
larsuthe indicators don't handle that case at all08:20
* larsu wonders if we should assert08:20
pittibonjour mes amis, comment avez-vous et le sprint ?08:21
larsuoh cool, GDBusConnection supports killing your process when the connection disappears08:21
larsupitti: bonjour!08:21
didrocksbonjour pitti08:22
larsudidrocks: kopf hoch!08:23
* didrocks does a fake smiling then08:24
desrtdidrocks: i repeat my original suggestion :)08:24
desrt(ie: put the code in plymouth)08:24
desrtdidrocks: stand up08:25
didrockswell, the goal was to have something that would work on the long term08:25
pitti"the code"> listening to fsckd?08:30
pitti'cause, fsckd itself makes more sense in systemd itself, and it was already agreed that we want this08:31
didrockspitti: yeah, that was this that we discussed (and what I explained yesterday)08:32
seb128pitti, salut! ça va bien, et toi ?08:33
pittiseb128: je vais bien, merci ! debugging people's boot problems :)08:33
seb128pitti, my reboot/shutdown hangs until I sit on the power button, I saw you mentioned some cases where that can happen on the upstream list08:34
seb128pitti, how can I tell if I hit one of those?08:34
pittiseb128: /usr/share/doc/systemd/README.Debian describes how to enable the debug shell08:34
seb128pitti, danke08:34
pittiseb128: i. e. boot without "splash quiet", systemctl start debug-shell.service08:34
pittiseb128: then shut down, and when it hangs, Alt+F9 to get to the debug shell08:35
seb128pitti, exellent, thanks08:35
pittiseb128: then systemctl list-jobs shows you what it's waiting on08:35
pittiseb128: then you can poke further with systemctl status -l <unit>, etc.08:35
pittiseb128: thanks for doing that; please file a bug once you know what it's hanging on, I really want to track these issues08:35
seb128pitti, yw08:36
seb128pitti, that README doesn't say to boot without splash quiet08:36
seb128just to systemctl start debug-shell08:36
pittiseb128: well, it's not strictly necessary, but generally nicer to see what's going on without manual prodding08:36
seb128brb08:38
seb128pitti, ok, of course it doesn't happen when I try to reproduce, booted with the option anyway maybe later I hit it again08:49
pittiseb128: at least the issue I was discussing on the ML is a race condition; might require a few reboots indeed :/08:50
pittiseb128: you can just start the debug shell, so that you have it available once it happens08:50
seb128pitti, yeah, the README has a warning about doing that :p08:50
pittiseb128: you could also "systemctl enable debug-shell" if you always want it; but be aware that anyone with access to your computer then has a root shell :)08:51
seb128yeah08:51
willcookemlankhorst, you ded?08:53
willcookemlankhorst, feeling any better today?08:53
darkxstpitti, seb128 its a security hole either way, unless you have encypted partitions08:55
pittidarkxst: these don't help -- they are usually unlocked while running :)08:55
pittithe only thing that actually helps is per-user ecryptfs for users which are not currently logged in08:56
pittibut meh -- it's a debugging tool, exactly what it says on the tin :)08:56
darkxstit does make it a little too easy to get root access to a machine though08:56
darkxstnot that its that hard to do it other ways also08:57
darkxstif you have physical access08:57
pittisure; just reboot in rescue mode08:57
pittiseb128: FYI, jibel just reminded me that it might be modemmanager09:06
pittithat occasionally hits me too, under upstart as well09:06
seb128Trevinho, larsu, there is an issue with your bamf changes, it displays a low res gedit icon for me, quite visible in alt-tab09:06
seb128pitti, how do I tell if that's it?09:06
darkxstpitti, sure, many easy  ways to get root access when you have physical access to a machine09:07
pittiseb128: list-jobs would have a job for modem-manager09:07
seb128pitti, k09:07
pittiseb128: under upstart, I'm not sure; back then I dropped splash/quiet and just saw it hang there09:07
seb128pitti, I'm under systemd, so debugging that is good enough for me09:07
larsuseb128: clearly Trevinho's fault :P09:10
tsdgeoscharles: you there?09:17
seb128tsdgeos, he's likely sleeping at this hour09:18
tsdgeosoki, i'll comment on the MR then09:19
seb128Laney, https://code.launchpad.net/~renatofilho/ubuntu/vivid/syncevolution/default-syncInterval/+merge/247768 was the mr for the syncevolution change09:19
seb128it's still unapproved and not even closed or commented on...09:20
mlankhorstwillcooke: slightly, still sick-ish10:06
willcookemlankhorst, bad luck :/10:06
willcookehope you are feeling better soon10:06
mlankhorstyep10:06
mlankhorstthanks10:06
mlankhorstbtw Xmir can't leave the ppa until 1.17 is in vivid10:07
willcookeah10:07
willcookeinteresting10:07
willcookeCan the source not be made available via that project either?10:07
mlankhorstI would prefer to use a git tree for Xmir10:08
willcookemlankhorst, can we still pull Xmir in to our projects/silos etc for inclusion in, let's say, a phone image?10:09
mlankhorstI think so10:09
willcookemlankhorst, ok - didrocks does something similar for Make10:10
willcookeLP can point to Git as an upstream project and pull it in automagically10:10
willcookelet's do that :)10:10
willcooke</handwave>10:10
mlankhorstfor an upstream project is a /debian directory needed?10:11
willcookeerm, didrocks ^ can you advise?10:11
happyaronseems not important, it'll be ignored by dpkg-source eventually10:13
Laneyyou can have this in another branch10:13
LaneyI guess pkg-xorg does something like that already10:15
Laneyseparate upstream and upstream+packaging branches10:15
mlankhorstyeah it's what happens right now10:25
mlankhorston every release we'll merge the new upstream back to the branch with packaging10:25
Laneywhat's wrong with that then?10:27
didrocksmlankhorst: no debian directory needed10:33
didrocksits all magic :)10:33
didrocksmlankhorst: I can do it for you if you create the project and gives me the creds10:33
didrocksand point to the git upstream repo :)10:34
Laneyoh for the import?10:37
mlankhorstok, I don't have a git repo currently, lets see..10:56
mlankhorstgit://people.freedesktop.org/~mlankhorst/xserver.git11:02
mlankhorstwhen it gets up in a bit11:02
didrocksmlankhorst: tell us when you registered the launchpad project11:11
mlankhorsthttps://launchpad.net/xmir already exists?11:13
mlankhorstcould we use that?11:14
didrocksmlankhorst: yeah, ask duflu to fix the ownership though11:16
didrocksmlankhorst: or kgunn11:16
mlankhorstwho do you want to have ownership?11:22
mlankhorstI can adjust it to anything it seems11:23
dufludidrocks, mlankhorst: Might be better to add Maarten to mir-team or create an xmir team11:24
mlankhorstduflu: https://launchpad.net/~mir-team/+members ;-)11:24
didrocksmlankhorst: ok, so, if you create a new branch, there is an option to tell "hosted in a git repo"11:25
didrocksand then, just point to your git ^11:25
dufluOK, I need to fix nothing.11:25
dufluLunch11:25
didrocksthanks duflu11:25
mlankhorstadded11:26
didrocksmlankhorst: should appear in ~15min11:28
didrockskeep me posted11:28
mlankhorstoke11:53
=== Cimi_ is now known as Cimi
=== MacSlow is now known as MacSlow|lunch
=== alan_g is now known as alan_g|lunch
seb128shrug13:48
seb128kenvandine, u-s-s tests seems unreliable again?13:48
seb128got a few different ones hitting CI errors :-/13:48
kenvandineseb128, :-(13:59
tedgseb128, kenvandine, do you guys know of a package overriding ctest? I was trying to do it and I can't seem to get the directory right. Looking for something to crib off of.14:03
kenvandinetedg, i haven't seen one14:03
=== alan_g|lunch is now known as alan_g
seb128tedg, no, I don't14:12
didrockslarsu: http://paste.ubuntu.com/10054163/14:13
tedg:-(14:14
tedgSomeone had to have wanted verbose test logs, somewhere! :-)14:14
tedgSo I think I need to be in the object directory.14:18
tedgHow do I find that?14:18
tedgIn my case it's: /home/ted/Development/indicator-sound/build-area/indicator-sound-12.10.2+15.04.20150129.1/obj-x86_64-linux-gnu14:19
xnoxtedg: dh_auto_make -> will put you in the cmake's build directory as far as I can tell.14:25
xnoxtedg: one can derive it from DEB_HOST_GNU_TYPE it's14:27
xnox"obj-" . dpkg_architecture_value("DEB_HOST_GNU_TYPE");14:27
xnoxinclude /usr/share/dpkg/default.mk14:27
xnox"obj-"$(DEB_HOST_GNU_TYPE)14:27
xnox"14:27
tedgCool, let me try that. I think I found an environment variable that might pass things to CTest as well.14:28
tedgYes, so this works to pass an arg to CTest:14:29
tedg+override_dh_auto_test:14:29
tedg+ARGS=-V dh_auto_test14:29
tedgWe can talk about horrible variable naming later :-)14:29
tedg(wasn't me)14:29
=== MacSlow|lunch is now known as MacSlow
tedgThanks xnox14:34
xnoxtedg: ewh14:34
xnoxtedg: dh_auto_test -- -V=114:34
xnoxoverride_dh_auto_test:14:34
xnox     dh_auto_test -- -V=114:35
xnoxis aught to work14:35
ogra_but it doesnt use the beautiful "ARGS"14:35
ogra_:P14:35
tedgxnox, I think that passes the -V to make, which doesn't work.14:35
tedgxnox, It has to pierce through the makefile and into ctest.14:35
* tedg heard you liked wrappers do he wrapped your make script in a dh_ helper so that it can wrap the test tool which wraps your tests14:36
xnoxtedg: i would have thought:14:37
xnoxdh_auto_configure -- -DCTEST_OUTPUT_ON_FAILURE=TRUE14:37
xnoxwould work14:37
xnoxor14:37
tedgI don't want the output on failure, I want it on progress.14:37
xnoxCTEST_OUTPUT_ON_FAILURE=TRUE dh_auto_test14:37
xnoxoh, horum.14:37
* xnox hates that cmake is non-free14:37
tedgThe problem is the test is hanging on Jenkins.14:37
xnoxdocumentation that is.14:37
xnoxhehe =) Jenkins, got to love ya.14:37
tedgYeah :-/14:38
ochosixnox: hey there! i know it's very low priority and all, but also extremely low hanging fruit, mind to take a 1min peek at my 3-line ubiquity MR? https://code.launchpad.net/~ochosi/ubiquity/xubuntu-panel-bg/+merge/24700114:38
tedgAnd since Jenkins just kills the whole thing instead of killing the leaf processes and letting make clean up, I don't get any logs.14:38
xnoxochosi: you want cyphermox14:39
cyphermoxah?14:39
xnoxcyphermox: ^^^ from ochosi14:39
cyphermoxyes yes14:39
ochosiah14:40
xnoxah ah14:40
* xnox joins the ah game14:40
ochosixnox: sorry, i was under the impression you were *the* ubiquity guy :)14:40
ochosiotherwise i wouldn't have done the ad hominem ping14:40
ochosithanks for fwding me14:40
xnoxochosi: i am just a minion.14:40
ochosiand hi cyphermox :)14:40
cyphermoxhowdy :)14:40
xnoxochosi: cyphermox is a more responsive minion =)14:40
ochosixnox: you mean one of those little yellow ones? :]14:40
ogra_ubiquity makes you crazy if you work in it to long ... so we try to rotate the devs before we need to send them to asylum14:41
ochosihaha14:41
xnoxochosi: BA-NA-NA14:41
cyphermoxI see myself more like an umpa-lumpa, but minion works ;)14:41
ochosiheh, well given that its codebase drives you crazy, ubiquity works well enough :)14:41
davmor2cyphermox: I don't believe it there is no oompaloompa vs minion videos :D14:55
cyphermoxomg.14:55
andyrockseb128, you here?14:58
seb128andyrock, hey14:58
andyrockso i'm working on the large text bug again (was busy in the last few days)14:58
seb128cyphermox, hey, how is the segfault issue going?14:58
andyrockknow unity correctly updates the value of text-scaling-factor (the gnome setting)14:59
andyrockbut the applications need to be refocused14:59
seb128andyrock, that's a bug in GTK I think14:59
andyrockto get the large text15:00
seb128Laney, larsu ^15:00
andyrockalso the gtk widgets in 14.10 scales nicely15:00
andyrock*scale15:00
=== SuperLag_ is now known as SuperLag
andyrockin 15.04 they fail to scale and text does not fit15:01
andyrockanother no-unity bug15:01
seb128larsu, Laney ^15:06
seb128andyrock, what widgets?15:06
larsuLaney: ^15:16
LaneyWHAT15:17
desrtseb128: ping15:17
seb128desrt, PONG15:17
didrocksthose guys…15:17
desrtseb128: so we're playing tennis now, or what?15:17
didrocksseb128: Laney: larsu: desrt: those guys!15:17
seb128oh yeah15:17
seb128double with a refereee = 515:17
desrtsize of a shift in hockey, excluding the goalie15:18
desrtalso five!15:18
desrtand omg15:18
desrtseb128 + didrocks + Laney + larsu + desrt ...15:18
desrtalso five15:18
larsu+ willcooke is also 515:19
desrtwillcooke can be the goalie15:19
larsuballboy15:19
desrtit would suck to have a nick that didn't have 5 letters in it15:20
desrtsure glad i don't have to live that life15:20
andyrockseb128, http://imagebin.ca/v/1qVP9Bni8IMD15:20
andyrockafter a while they get the correct dimensions15:21
andyrockthe text just need a refocus15:21
andyrockthe widgets (in the pictures) need some time more15:21
andyrockseb128, and it happens on nautilus, u-c-c and others15:22
seb128andyrock, yeah, looks like a gtk issue, please fix the unity side still15:23
andyrockok, i can look inside the gtk side as well but I need some help to start15:24
seb128Laney, larsu ^15:24
larsuLaney: ^15:25
LaneyWHAT15:26
larsuWAT15:26
Laneythanks it's in our brain space and will look soon15:26
cyphermoxseb128: still in my PPA, with the other fix too15:26
=== tyhicks` is now known as tyhicks
andyrockLaney, regarding the issue with "large text"15:27
andyrockLaney, https://bugs.launchpad.net/unity/+bug/140821215:27
ubot5Launchpad bug 1408212 in unity (Ubuntu) "Turn on/off the Large Text in Universal Access, it doesn't work immediately" [High,In progress]15:27
andyrockLaney, a fix for Unity is in progress15:28
andyrockbut likely there is some issue on gtk as well15:28
andyrockseb128, I proposed a fix right now15:32
andyrockseb128, I'm working on an utility class to avoid this issue in the feature but for the moment is better to fix it asap15:32
didrockswillcooke: bregma: https://bugs.launchpad.net/unity/+bug/136407015:32
ubot5Launchpad bug 1364070 in unity (Ubuntu) "utopic and trusty unity (7.3.1+14.10.20140811-0ubuntu1) launcher addition through gsettings isn't picked up" [Medium,Confirmed]15:32
andyrockdidrocks, last time it was working fine15:33
seb128attente_,  gsettings set org.gnome.desktop.interface text-scaling-factor15:33
andyrockdidrocks, likely you where using new gsettings15:34
andyrock*were15:34
didrocksandyrock: did you look at the bug report?15:35
didrocksandyrock: it was reporting way before the gsettings change15:35
didrocksreported*15:35
didrocksandyrock: and it's 80% of the time15:35
andyrockdidrocks, yep i know15:35
didrocks(see description)15:35
andyrockbut I can't reproduce 100%15:35
didrocksandyrock: so, it's not due to the new gsettings15:35
didrocksyeah, 80%15:35
didrocksgetting bug reports everyday15:35
didrocksseb128: you got it as well, isn't it?15:35
seb128I saw it yeah15:36
andyrockdidrocks, how do you create the desktop file?15:41
didrocksandyrock: in ~/.local/share/applications15:41
didrocksjust writing to it through python15:41
didrocksandyrock: want an example?15:42
didrocks(it shows up in the dash, and I can add it afterward)15:42
didrocksandyrock: again, it's working like 20% of time15:42
didrocksso, maybe a timestamp issue?15:43
didrocks(gsettings has the right values with gsettings get)15:43
andyrockdidrocks, k I'll try to reproduce it again15:45
didrocksthanks!15:45
andyrockdidrocks, maybe bamf's fault as well15:46
bregmawillcooke, another thing we need to have a strategy for is applications that make explicit X11 calls despite using a toolkit15:48
willcookebregma, is that not convered by Xmir?15:49
bregmaonly if it's using XMir15:49
willcookeAhh, so, e.g. Gtk apps which also call X1115:49
bregmayeah, like gksudo15:50
willcookeerk15:50
didrocksandyrock: oh, good one15:50
bregmaor QtCreator15:50
didrocksandyrock: when it happens, a logout/login "fixes" it, so can be bamf,yeah15:50
willcookeSounds like another "exceptions list" thing15:50
Trevinhoandyrock: I think it's not bamf, but mostly related to the icons priority15:50
TrevinhoI had a change that was resetting it after a reorder, so maybe that's enough but it had other downsides15:51
bregmawillcooke, we're thinking a small fake Xlib might solve some problems (setting WM atoms is a common reason to call X11 directly)15:51
andyrockTrevinho, unity asks bamf is a desktop file is a good one right?15:51
andyrock*if15:51
Trevinhono15:51
willcookebregma, oh - much nicer :)15:52
andyrockmmm so it's not the problem15:52
larsusomeone hug desrt16:38
desrtlarsu: ?16:41
larsudesrt: you looked like you needed one when you were fighting with wp16:41
desrtah.  ya.16:41
desrtevil wp.16:42
andyrockdidrocks, ok i can reproduce the issue16:44
Laneywee16:47
Laney46,000th launchpad email16:47
Laneysince Date: Wed, 04 Aug 2010 19:29:21 -000016:47
desrtLaney: good job.  have a cookie.16:50
LaneyI'll take one at 50,000 instead16:51
desrt:)16:51
desrtlarsu seems happy16:51
didrocksandyrock: sweet!16:51
larsudesrt: he is!16:52
=== alan_g is now known as alan_g|EOD
=== mterry_ is now known as mterry
cyphermoxochosi: your ubiquity change looks fine, but you should know that ubiquity-panel-bg.png doesn't currently seem to be installed from shimmer-themes.20:04
ochosicyphermox: hm, weird, it's in git and i thought there was a release+upload already. i'll check again. good thing is, even if it fails, i.e. that file isn't there, ubiquity won't fail/break23:16
ochosicyphermox: thanks for reviewing/approving and letting me know!23:16
andyrockdesrt, ping23:42

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