[06:45] <pitti> Bonjour tout le monde
[06:46] <duflu> tout le monde: bonjour ;)
[07:06] <larsu> good morning everyone
[07:10] <Mirv> mlankhorst: thanks, LTS-S working great here :) (with one workaround I needed related to the understandable upstream changes)
[09:01] <Laney> morning
[09:03] <Laney> happyaron: ok, doing
[09:07] <seb128> good morning desktopers!
[09:07] <seb128> Laney, hey
[09:07] <Laney> hey seb128
[09:15] <mlankhorst> Mirv: well if the same workaround would have been needed for saucy, great. :P
[09:15] <mlankhorst> oh grr, I need to figure out a bug on that one still
[09:17] <happyaron> Laney: thanks, :)
[09:17] <Laney> fonts-droid will need promoting first though
[09:18] <Laney> https://bugs.launchpad.net/ubuntu/+source/fonts-android/+bug/1249132
[09:18] <ubot2> Launchpad bug 1249132 in fonts-android (Ubuntu) "[MIR] fonts-android" [Undecided,Fix committed]
[09:18] <happyaron> mterry said he did that, and just waiting it to being seeded. but I have no idea where we are.
[09:18] <Laney> seb128: could you do the honours?
[09:20] <seb128> Laney, sure
[09:20] <Laney> merci
[09:20] <Mirv> mlankhorst: yeah, it would have, it's upstream change and I just finished a blog post on it too http://losca.blogspot.fi/2013/11/workaround-for-setting-full-rgb-when.html
[09:21] <mlankhorst> ah :p
[09:22] <seb128> Laney, done
[09:22] <Laney> \o/
[10:03]  * didrocks feels guilty
[10:04] <didrocks> seb128: do you know who is active on indicator-messages?
[10:04] <seb128> didrocks, what did you do again?
[10:04] <didrocks> I'm going to ask for something!
[10:04] <didrocks> ;)
[10:04] <seb128> lol
[10:04] <seb128> larsu, HIDE
[10:04] <didrocks> roh, you made the connection too quick :p
[10:04] <seb128> didrocks, is that the "timestamp and content is sometime wrong on the phone"?
[10:04] <didrocks> basically we have that one: bug #1253810
[10:04] <ubot2> Launchpad bug 1253810 in indicator-messages (Ubuntu) "Messages in Incoming not always display the correct date and content" [High,Confirmed] https://launchpad.net/bugs/1253810
[10:04] <didrocks> yep
[10:05] <seb128> I had a feeling that would come our way soon :p
[10:05] <seb128> did that start with https://launchpad.net/ubuntu/+source/indicator-messages/13.10.1+14.04.20131125-0ubuntu1 ?
[10:05] <seb128> no
[10:05] <seb128> report is from 21
[10:05] <didrocks> seb128: some people mentionned image 27 at least, so yeah, a little bit older
[10:05] <didrocks> not sure if it was a hidden bug which is now easier to report
[10:06] <seb128> weird, the indicator didn't change before 10-29 and 11-25 though
[10:06] <didrocks> yep, can be as well something else in the pipe
[10:06] <didrocks> rsalveti told that he checked the ofono side though
[10:07] <didrocks> (and messaging app is fine)
[10:07] <didrocks> so not a service side
[10:07]  * didrocks hopes for a trivial bug
[10:08] <seb128> it would be more trivial to debug with a reproducer
[10:08] <seb128> do you have an easy way to trigger it?
[10:08] <didrocks> agreed, I have no more information than the bug report
[10:08] <didrocks> it was mentionned in the meeting, without the regression tag (as my comment and tag addition)
[10:09] <didrocks> seems people were upset (what ogra_ told) about it, but it wasn't raised :)
[10:10] <ogra_> seb128, i still see the 12h/24h issue (once the panel clock loads my clock goes to 12h ...), it smells a bit like these are related
[10:10] <seb128> ogra_, that sounds like a bug in the sdk/some of the qml component
[10:11] <seb128> didrocks, ^
[10:11] <larsu> ya, I can't think of a change in the indicator itself that might trigger that
[10:11] <ogra_> well to me it looks like the clock backend fires up before ~/.pam-environment is read ... but i dont know how that would be technically possible
[10:11] <didrocks> rsalveti: received 10 sms, got the issue 4 times
[10:11] <didrocks> from the irlogs
[10:11] <larsu> whoever reproduces it next, please ping me before interacting with the mesage
[10:11] <larsu> *message
[10:11] <seb128> rsalveti, ^ when you wake up
[10:12] <seb128> we need help debugging from somebody who gets the issue
[10:12] <ogra_> or Ursinha ^^^
[10:12] <didrocks> ogra_: reading the backlog, people don't seem to have been as upset or doing buzz around it as you told doing the meeting :)
[10:12] <didrocks> seb128: http://irclogs.ubuntu.com/2013/11/26/%23ubuntu-ci-eng.html#t20:45 FYI
[10:13] <seb128> didrocks, thanks
[10:13] <seb128> seems like Ursinha and rsalveti have it easily
[10:13] <ogra_> didrocks, did i say that ? i only said ricardo complained about another regression
[10:13] <seb128> so they can probably help us debug when they get up
[10:14] <didrocks> ogra_: well, you clearly used the word "upset" in the hangout, so I found it quite strong and thought people raised alarms :)
[10:14] <ogra_> (probably complained was the wrong word, sorry)
[10:14] <didrocks> ogra_: no worry, just be careful with words, people can imagine other things :)
[10:22] <Kaleo> seb128, can install the indicator-datetime?
[10:22] <Kaleo> +I
[10:23] <seb128> Kaleo, the ido fix is in, so you should be fine yes
[10:24] <Kaleo> seb128, thanks!
[10:24] <seb128> yw
[10:24] <seb128> Kaleo, just for info, we have a less frequent issue that is pending landing (https://code.launchpad.net/~larsu/ido/remove-idle-source/+merge/196841)
[10:25] <seb128> but I've seen that one only once
[10:26] <Kaleo> ok
[10:26] <Kaleo> so if I see my cpu 100%
[10:26] <Kaleo> I know why :)
[10:27] <ogra_> yeah, you clearly need more CPU then :)
[10:28] <seb128> brb
[10:35] <Laney> xnox: just uploaded mythtv, want to wrangle the transitional package or whatever it is for samba?
[10:37] <xnox> Laney: yeap. was poking it already.
[10:37] <Laney> neat
[10:37] <Laney> then it'll just be tbird
[10:37] <Laney> \o/
[10:37] <xnox> Laney: thanks a lot for your help.
[10:38] <Laney> np
[10:38] <Laney> looking forward to seeing it in
[10:38] <xnox> well, it will have to go through new...
[10:38] <xnox> cause imho samba*4 dummy transitional packages should be in src:samba, and src:samba4 dropped.
[10:38] <seb128> larsu, overlay scrollbars on gtk2 are happy again with your new commit ;-)
[10:38] <Laney> I'm sure you are well connected enough to make that happen :P
[10:39] <xnox> Laney: LOL =)
[10:50] <Laney> oops
[10:50] <Laney> just rm -rf * in my home directory
[10:50] <Laney> luckily caught it before it went very far ...
[10:55] <seb128> urg
[10:56] <seb128> I tend to try to avoid that command, I usually go down a directory and rm subdir/...
[10:59] <Laney> yeah...
[10:59] <Laney> I thought I was in temp/
[11:00] <Laney> oh well, most of the stuff it deleted is synced anyway
[11:11] <sil2100> seb128: hello! :D
[11:11] <seb128> sil2100, hey, how are you?
[11:11] <sil2100> seb128: fine, thanks :) And you?
[11:11] <seb128> I'm good thanks
[11:12] <didrocks> seb128: runnnnnnnn :)
[11:12] <seb128> didrocks, yeah, I was thinking about that ;-)
[11:14] <didrocks> seb128: I think you saw that, but indicator-messages is out of the loop, seems Unity8 :)
[11:16] <seb128> didrocks, no, I didn't, but good to read
[11:16] <seb128> larsu, ^
[11:18] <sil2100> seb128: I'm not sure if this was preNEWed already or not, since we were doing a packaging review like a few months ago, but could you take a look at lp:unity-voice ?
[11:18] <seb128> didrocks, was it again Saviq making bugs and trying to blame it on us?
[11:18] <seb128> sil2100, sure
[11:18] <sil2100> seb128: there's one copyright mismatch that I'll fix (one of the autopilot files has an invalid LGPL header), but besides that it looks fine
[11:18] <sil2100> seb128: thank you!
[11:18] <Saviq> seb128, *making* them? it's not a regression :P
[11:19] <Saviq> seb128, and I haven't seen steps to reproduce yet :P
[11:19] <didrocks> seb128: completely all their fault! :)
[11:20] <seb128> chrisccoulson, hey, so turned out that the addon porting was non trivial?  or did you get lazy and just didn't want to do it? ;-)
[11:21] <seb128> didrocks, btw tb got uploaded, can we get an ido landing for the other fix?
[11:21] <chrisccoulson> seb128, the latter for now. i'll have a proper look at it when i have a bit more time ;)
[11:21] <seb128> chrisccoulson, k
[11:21] <Laney> ooh, getting close
[11:22] <didrocks> seb128: sure, Mirv: mind taking care of that? ^
[11:22] <Laney> turns out my rm accident wiped out a load of config files
[11:22] <Laney> which then synced them to the cloud and deleted them from my desktop too
[11:22]  * Laney coughs
[11:23] <didrocks> Laney: syncing to the cloud is not a backup! :)
[11:23] <didrocks> but you had deja-dup right? ;)
[11:23] <Laney> I did have not-too-old copies
[11:23] <Laney> of most of them, not every single one
[11:23] <Laney> but it's alright, got the important ones back
[11:27] <seb128> Laney, :-(
[11:28] <Laney> $ mv configs zzzconfigs ;-)
[11:30] <larsu> seb128: ah, cool
[11:30] <larsu> did you ping dednick about it?
[11:31] <seb128> larsu, no, but I guess they are aware of the issue/on it since didrocks said it's a bug on their side
[11:32] <larsu> ah okay
[11:33]  * larsu goes back to his pizza then
[11:37] <Mirv> didrocks: so, ido/trusty + unity7 AP (what indicators stack check job would do if it'd work at the moment) for it?
[11:38] <didrocks> Mirv: I don't think ido is used on touch
[11:38] <didrocks> seb128: right? ^
[11:38] <didrocks> Mirv: so yeah, only a rebuild with the indicators stack check tests
[11:39] <seb128> larsu, hum, pizza, me wants!
[11:39] <seb128> didrocks, Mirv: no, it's not, it's a set of GTK widgets
[11:39] <didrocks> POA
[11:39] <didrocks> Pizza On Air
[11:39] <seb128> didrocks, is that like the virtual beer at vUDS?
[11:39] <Mirv> didrocks: unfortunately the check part doesn't work because of evolution migration (or at least didn't on Monday), so I'll run them manually if needed
[11:39]  * seb128 wants real pizza and beers!
[11:41] <larsu> seb128: beer for lunch? Are you German? :P
[11:41]  * Mirv building ido in daily-build
[11:41] <didrocks> seb128: yeah, but more solid :p
[11:42] <didrocks> Mirv: ok ;)
[11:44] <seb128> larsu, well, some things just match ;-)
[11:45] <seb128> (the other option is soda+pizza, but that would makes me feel american, I prefer being german ;-)
[11:46] <larsu> great, now I feel bad for having water with my pizza
[11:46] <Mirv> I'm noticing btw now the same python problem I had on device on desktop :S "... module incorrectly imported from [path]. Expected [the same path]. . Is this module globally installed?"
[11:46] <seb128> larsu, what, are you french?!
[11:46]  * seb128 hugs larsu
[11:46] <larsu> haha
[11:47] <didrocks> Mirv: what have you done? ;)
[11:49] <Mirv> didrocks: I have zero idea :) filed bug #1255505
[11:49] <ubot2> Launchpad bug 1255505 in autopilot (Ubuntu) "autopilot crashed with ImportError in _find_tests(): 'test_switcher' module incorrectly imported from '/usr/lib/python2.7/dist-packages/unity/tests'. Expected '/usr/lib/python2.7/dist-packages/unity/tests'. Is this module globally installed?" [Undecided,New] https://launchpad.net/bugs/1255505
[11:49] <Mirv> because apport kindly offered it
[11:50] <didrocks> ;)
[11:53] <sil2100> hm, on the desktop, what component exactly does control the desktop wallpaper?
[11:54] <didrocks> sil2100: nautilus displays it
[12:09] <didrocks> waow, Microsoft starts to be agressive (maybe feeling that this christmas a lot of people will buy a chromebook): https://www.youtube.com/watch?v=-y2mqoDjQXI&noredirect=1
[12:10] <didrocks> (funny though that they post the video on … youtube :p)
[12:12] <Mirv> bingtube would reach so much more people and all
[12:13] <didrocks> heh, sure! :)
[12:14] <larsu> didrocks: that's hilarious
[12:14] <didrocks> larsu: I find it funny as well, if you take that with a big grain of salt :)
[12:14] <larsu> I have the feeling that whole campaign is worse for microsoft than it is for google
[12:15] <larsu> "scroogled", seriously?
[12:15] <didrocks> larsu: knowing that for being on the chromebook, all chrome apps need to work offline by default for instance, so their main argument are totally FUD :)
[12:15] <didrocks> larsu: yeah, the "scroogled" term is awesome, I recognize that :)
[12:16] <didrocks> "see this logo, it means it's not a real computer"
[12:16] <didrocks> :)
[12:16] <larsu> ha, yeah
[12:16]  * ogra_ googles bingtube
[12:16] <larsu> also, I don't think people buying those think that there's windows on them
[12:16] <ogra_> (for the irony)
[12:16] <larsu> I think they're buying them because there's _no_ windows on them
[12:16] <didrocks> right, I agree with you
[12:17] <didrocks> ogra_: argh, it seems it's not an url safe for work :)
[12:17] <ogra_> lol
[12:18] <ogra_> didrocks, did you know that our new sensor handling service is calles usensord ... try that one ;)
[12:18] <ogra_> oh, it changed !
[12:18] <didrocks> ogra_: my whole first page are only launchpad and google plus results
[12:18] <ogra_> yeah, great
[12:18] <didrocks> from a guy, "Michael… something?" ;)
[12:18] <ogra_> it used to find a lot chinese portn
[12:18] <didrocks> interesting…
[12:19] <ogra_> "ucensored"
[12:19] <didrocks> oh right
[12:19] <didrocks> ah
[12:26] <seb128> larsu, oh, fcitx preferences UI has the same issue than update-manager with 3.10 (e.g a list reduced to one line rather than taking the space it should)
[12:28] <larsu> fcitx?
[12:28] <seb128> larsu, the new ibus :p
[12:29] <seb128> larsu, you clearly need to learn more about input, you should take over attente's job for a cycle, just trade or something ;-)
[12:30]  * larsu runs
[12:30] <seb128> didrocks, I like how the video get 3 times more thumbs down
[12:30] <larsu> seb128: hm, works fine for me. You mean the window that pops up when you click on indicator/configure?
[12:30] <seb128> larsu, no, fcitx is another input framework, which is not ibus
[12:31] <seb128> larsu, that's what Kylin is using and they would like us to switch to it
[12:31] <larsu> right, I just installed that
[12:31] <larsu> and it gives me an app indicator
[12:31] <seb128> right
[12:31] <seb128> click the configure entry
[12:31] <larsu> yep
[12:31] <didrocks> seb128: yeah, bad commercials are punished, I have faith again!
[12:31] <seb128> the input tab (the first one)
[12:31] <larsu> weirdness, it works for me. Are you on 3.10.3?
[12:32] <seb128> larsu, well, for me the first tab has its language list being one line at the top
[12:32] <seb128> then lot of empty space
[12:32] <seb128> then one line at the bottom
[12:33] <larsu> seb128: http://i.imgur.com/EGuJeSB.png
[12:33] <seb128> larsu, let me try again, I'm on 3.8.8 atm, was testing the new version before uploading and testing that scrollbars work on 3.8
[12:33] <larsu> ah right
[12:33] <seb128> larsu, well, 3.8 works, I was having the issue yesterday on 3.10.4
[12:33] <larsu> I still have the issue in update-manager
[12:34] <seb128> larsu, oh, sorry
[12:34] <seb128> larsu, click the "+" at the bottom
[12:34] <mdeslaur> seb128: hey, if you do decide to stay on ibus...can we at least kill the stupid notification on first login?
[12:34] <seb128> that's the dialog which has the issue
[12:35] <larsu> seb128: ah, got the same. Thanks
[12:35] <seb128> mdeslaur, the one saying that your key migrated? I wish I was able to get that one, people mentioned it but it just doesn't happen here
[12:35] <seb128> larsu, yw
[12:35] <Mirv> didrocks: FYI it was python 2.7 update that breaks autopilot for me. downgrading to 2.7.5-8ubuntu4 made it possible for me to run the tests (and sooin release ido)
[12:35] <mdeslaur> seb128: that's odd, I see it in all my vms, and I see it on the laptop I just installed yesterday
[12:36] <seb128> mdeslaur, I don't have fresh installs, but I do use guest sessions a lot and it doesn't happen there
[12:36] <seb128> mdeslaur, I guess I need to try a fresh install in a vm, that's annoying as a testcase though
[12:37] <didrocks> Mirv: maybe you should ping doko about it
[12:37] <mdeslaur> seb128: how weird, the guest session in my vm doesn't show it
[12:37] <seb128> mdeslaur, see! ;-)
[12:38] <Mirv> didrocks: ok
[12:38] <mdeslaur> seb128: I think my vm just got scared of you and stopped showing it
[12:39] <seb128> mdeslaur, good, so I fixed it, you own me a beer, I can deal with that ;-)
[12:39] <mdeslaur> lol
[12:42] <Mirv> seb128: ido uploaded
[12:42] <seb128> Mirv, thanks!
[12:46] <didrocks> Mirv: argh, I have another script broken by this python upload as well
[12:52] <mdeslaur> seb128: what does ibus-ui-gtk3 do?
[12:53] <seb128> mdeslaur, I think that's what gives you the "predictions"
[12:53] <seb128> mdeslaur, e.g http://hanboyang.files.wordpress.com/2010/01/ibus.jpeg
[12:54] <seb128> not sure though
[12:54] <seb128> happyaron probably knows better
[13:04] <rsalveti> larsu: what do you need for bug 1253810
[13:04] <ubot2> Launchpad bug 1253810 in unity8 (Ubuntu) "Messages in Incoming not always display the correct date and content" [High,Confirmed] https://launchpad.net/bugs/1253810
[13:05] <happyaron> mdeslaur: basically the small windows that displays the candidate list.
[13:06] <larsu> rsalveti: apparently somebody already figured out that it is an issue in unity8 and not the indicator. I'm afraid I can't help much anymore...
[13:08] <seb128> larsu, well, it seems you are involved in the unity8 branch linked to this bug as well ;-)
[13:08] <seb128> but yeah
[13:08] <seb128> rsalveti, check with Saviq I guess, maybe you can try the unity8 fix in there
[13:08] <rsalveti> hm, cool
[13:09] <larsu> hm, are you guys sure that's the issue?
[13:09] <seb128> larsu, I've no idea, just relaying what Saviq and didrocks said
[13:09] <seb128> Saviq, ^?
[13:09]  * ogra_ still thinks it is related to the werid clock behavior that shows when the datetime indicator starts 
[13:10] <ogra_> ... which reminds me ... to filnally file a bug about that :P
[13:10] <larsu> ogra_: that's unlikely
[13:10] <larsu> but please file a bug
[13:10] <larsu> :)
[13:10] <larsu> rsalveti: do you have one of those messages in the menu right now
[13:10] <larsu> ?
[13:10] <Ursinha> seb128, ogra_, if you're talking about the indicator bug that I filed, it's easily reproducible
[13:11] <Saviq> larsu, no I'm not sure that's the issue, but then I'm not sure unity8 should even care - as in why couldn't the indicator give up preformatted strings?
[13:11] <rsalveti> larsu: nops, but can try to reproduce
[13:11] <ogra_> Ursinha, the one that restets my clocks on the system ?
[13:11] <Ursinha> I'm seeing those since r27, and it's like 4 out of 10 msgs
[13:11] <ogra_> (from 24h to 12h)
[13:11] <seb128> Ursinha, yeah, if you have it, larsu might want to have some details
[13:11] <Saviq> larsu, it's what we're doing for i18n - strings come pre-translated, I'm not sure why date formatting should be different
[13:12] <Ursinha> ogra_, no, I haven't noticed that... but I can pay more attention to see if I find it too
[13:12] <ogra_> Ursinha, does brazil use 24h time format ?
[13:12] <ogra_> i think it is closely tied to your locale
[13:13] <Ursinha> ogra_, afaik there's not really a standard here
[13:13] <larsu> Saviq: so that we don't have to send a dbus message every minute
[13:13] <Ursinha> I never use brazilian portuguese or brazilian locale for anything
[13:13] <larsu> or every second for people who like seeing seconds
[13:13] <ogra_> Ursinha, well, the locale setting sets the display format for the time
[13:14] <Saviq> larsu, maybe, yeah
[13:15] <larsu> Saviq: tbh, that's not an issue with the current design of the messaging menu (it shows absolute times). It used to be relative times, like on the desktop. And the time indicator has the issue anyway
[13:15] <Ursinha> btw I'm having a transient (it seems) issue that I click on icons but applications don't open
[13:15] <Ursinha> Saviq, is that your stuff?
[13:15] <larsu> whoever has this issue right now, please execute `gdbus call --session --dest com.canonical.indicator.messages  --object-path /com/canonical/indicator/messages/phone  --method org.gtk.Menus.Start [0,1]`
[13:16] <larsu> and tell me when the message should have arrived :)
[13:16] <Ursinha> once I'm able to open applications I'll test that for you
[13:16] <larsu> thanks
[13:16] <ogra_> bug 1255530
[13:16] <ubot2> Launchpad bug 1255530 in indicator-datetime (Ubuntu) "system clock gets forcefully set to 12h (AM/PM) as soon as the panel clock starts" [Medium,New] https://launchpad.net/bugs/1255530
[13:16] <Ursinha> I don't want to reboot and miss the bug that I seem to have here right now
[13:17] <Saviq> Ursinha, depends, can you see interesting things in unity8.log?
[13:17] <larsu> ogra_: cool bug. What does `gsettings get com.canonical.indicator.datetime time-format?` give you?
[13:18] <ogra_> (process:4391): dconf-CRITICAL **: unable to create directory '/root/.cache/dconf': Read-only file system.  dconf will not work properly.
[13:18] <ogra_> 'locale-default'
[13:18] <rsalveti> running as root?
[13:18] <larsu> ogra_: no as root please..
[13:18] <ogra_> (would probably be nice to tell dconf to not try to use /root/ :) )
[13:19] <larsu> dconf should find out if its running as root and bail?
[13:19] <Ursinha> Saviq, nothing is added to unity8 log when I try to open applications
[13:19] <ogra_> no, it should just not try to create a user db for the root user
[13:19] <Ursinha> but I tried to open them via shell and:
[13:19] <ogra_> 'locale-default' is the output btw
[13:20] <Ursinha> phablet@ubuntu-phablet:~$ upstart-app-launch webbrowser-app
[13:20] <Ursinha> ** (process:11378): WARNING **: Unable to connect to the Upstart Session: Empty address ''
[13:20] <ogra_> Ursinha, how did you become phabletz
[13:20] <ogra_> (dont use su !)
[13:20] <Ursinha> ogra_, how should I do that? :)
[13:20] <popey> sudo -u phablet -i
[13:20] <larsu> ogra_: okay that sounds good. I'll ping charles about this bug when he's up.
[13:20] <ogra_> sudo -u phablet -i
[13:20] <popey> \o/
[13:20] <ogra_> popey, wins
[13:20] <Saviq> ssh phablet@ ;P
[13:20] <ogra_> :)
[13:20] <Ursinha> thanks :) but same error
[13:21] <Saviq> sounds like the upstart session got b0rked
[13:21] <popey> you need to specify the APP_ID bit don't you?
[13:21] <Saviq> popey, webbrowser-app is no click
[13:21] <Ursinha> used to work before
[13:21] <popey> ah
[13:21] <ogra_> phablet@ubuntu-phablet:~$ upstart-app-launch webbrowser-app
[13:21] <ogra_> phablet@ubuntu-phablet:~$
[13:21] <ogra_> works fine for me
[13:21] <rsalveti> me 2
[13:21] <Saviq> ogra_, Ursinha yeah, your upstart session got borked somehow
[13:21] <popey> me 4
[13:21] <Ursinha> ogra_, I'm not saying the bug is permanent, I'm saying it's happening right now
[13:21] <Saviq> me 8
[13:21] <seb128> me 128
[13:21] <Ursinha> lol
[13:21] <popey> hah
[13:21] <rsalveti> :P
[13:22] <seb128> ;-)
[13:22] <Ursinha> seb128 won
[13:22] <seb128> \o/
[13:22] <Saviq> seb128, you can't count :P
[13:22] <Ursinha> but I'm reporting a transient issue, people
[13:22] <Saviq> lots!
[13:22] <seb128> Saviq, who cares, I won :p
[13:22] <popey> Ursinha: can you start webbrowser from the icon?
[13:22] <Saviq> popey, no, that was the first thing she tried
[13:22] <Ursinha> popey, no, as I said, I can't open
[13:22] <popey> ah
[13:22] <ogra_> Ursinha, well, i never had it not working before ... did you restart your session but did not re-login the user ?
[13:22] <Ursinha> thanks Saviq :)
[13:23] <Ursinha> ogra_, I did nothing, that stopped working and I came here to ask directions to debug that
[13:23] <Ursinha> because I didn't want to dismiss the bug
[13:23] <ogra_> well if your unity session crashed, the session env vars will be updated in ~/.bashrc
[13:23] <ogra_> if you dont log in again the vars are wrong
[13:24] <ogra_> there ios sadly no sane way around this without making unity kick you out of the shell when it restarts or some such
[13:24] <rsalveti> ogra_: hm, thought unity8 would use the same env if restarted
[13:25] <Ursinha> ogra_, what should I look for in the logs?
[13:25] <ogra_> rsalveti, well, lightdm sets new vars
[13:25] <ogra_> Ursinha, try a second adb session and see if it works from there
[13:25] <ogra_> if it does, then it is this ... well, ... i dont like to call it bug ... its a design flaw
[13:26] <ogra_> rsalveti, it cant, there is a new dbus address and a new upstart session ... the env vars get other values
[13:26] <Ursinha> no man, I have like five different shells and none work
[13:26] <mdeslaur> seb128, happyaron: how about something like this? http://paste.ubuntu.com/6484307/
[13:26] <ogra_> and there is no way to tell bash to reaload it from outside the shell session
[13:26] <ogra_> *reload
[13:27] <seb128> mdeslaur, where is prev_version coming from?
[13:27] <rsalveti> Ursinha: in theory the upstart session address is located at the UPSTART_SESSION env var
[13:28] <mdeslaur> seb128: gsettings I believe
[13:28] <seb128> mdeslaur, where is that not happening in a guest session?
[13:28] <mdeslaur> seb128: it is happening for me in a guest session on real hardware
[13:28] <seb128> not here
[13:28] <mdeslaur> seb128: I think the vm screen isn't coming up fast enough to see the actual notification
[13:28] <seb128> seems a bit random :/
[13:28] <ogra_> Ursinha, rsalveti http://paste.ubuntu.com/6484317/ are the code snippüets in question
[13:29] <ogra_> -ü
[13:29] <Ursinha> rsalveti, UPSTART_SESSION is empty
[13:29] <seb128> mdeslaur, in any case the patch looks fine to me
[13:29] <Ursinha> how did that happen?
[13:29] <mdeslaur> seb128: ok, let me send it upstream, file a bug, etc. thanks
[13:29] <seb128> mdeslaur, thank you!
[13:29] <rsalveti> ogra_: right
[13:30] <rsalveti> let me try killing my unity8 to see
[13:30] <ogra_> Ursinha, even if you enter a new adb session and use "sudo -u phablet -i"
[13:30] <Ursinha> ogra_, yes sir
[13:30] <ogra_> can you check if /etc/profile.d/ still has the dbus and upstart snippets ?
[13:31] <Ursinha> sure
[13:31] <rsalveti> hahah, my screen is now flashing after killing unity8
[13:31] <ogra_> they usually get processed by sudo-ing
[13:31] <Ursinha> lol
[13:31] <ogra_> rsalveti, how did you "kill" it ?
[13:31] <rsalveti> kill -9
[13:31] <ogra_> ugh
[13:31] <Ursinha> ogra_, upstart part is there
[13:31] <ogra_> use upstart
[13:32] <rsalveti> ogra_: well, I'm simulating a crash
[13:32] <rsalveti> I don't want to nicely kill it
[13:32] <ogra_> it wont clean up after itself if you just kill it
[13:32] <rsalveti> :-)
[13:32] <ogra_> Ursinha, dbus not ?
[13:32] <rsalveti> I want to see if unity8 would recover after a crash
[13:32]  * ogra_ definitely has it here on r31
[13:32] <Ursinha> ogra_, but if it dies it won't clean up after itself, so it's a bug
[13:32] <rsalveti> we should probably have a test case for that
[13:32] <Ursinha> ogra_, yes, both are here
[13:32] <ogra_> Ursinha, well, upstart should clean up
[13:33] <Ursinha> ogra_, but if it dies... as rsalveti said, it should be tested
[13:33] <rsalveti> Ursinha: what do you have with /sbin/initctl list-sessions
[13:33] <Ursinha> rsalveti, that's what I was looking right now
[13:33] <ogra_> rsalveti, nothing if UPSTART_SESSION is empty :)
[13:33] <rsalveti> well, just confirming :-)
[13:33] <ogra_> initctl needs that var
[13:33] <rsalveti> maybe a random output lol
[13:33] <Ursinha> yeah, nothing
[13:34] <Ursinha> you never know hahaha
[13:34] <rsalveti> so it could be that the session died and unity8 restarted without a valid session var
[13:34] <ogra_> no
[13:34] <ogra_> then nothing would start
[13:34] <rsalveti> but not sure if unity8 would open without a valid session
[13:35] <rsalveti> yeah
[13:35] <ogra_> unity itself might, but nothing else
[13:35] <rsalveti> guess the session just died then
[13:35] <Ursinha> the session died, wtf
[13:35] <ogra_> right
[13:35] <ogra_> and properly restarted
[13:35] <Ursinha> okay, so this is a bug as it it should recover properly
[13:35] <Ursinha> right?
[13:35] <ogra_> it obviously did
[13:35] <rsalveti> hm, rebooted and now I don't have the indicator clock
[13:35] <ogra_> else you wouldnt have a UI
[13:35] <ogra_> rsalveti, yeah, longstanding bug
[13:36] <rsalveti> argh
[13:36] <ogra_> it disappears randomly
[13:36] <ogra_> i was told that also happens on desktop
[13:36] <ogra_> (havent seen it here though)
[13:36] <rsalveti> oh, hm
[13:37] <Ursinha> ogra_, obviously recovered properly? if it did I should have my apps opening here haha
[13:37] <ogra_> Ursinha, do the apps open if you tap them ?
[13:37] <Ursinha> haha, man, no
[13:37] <Ursinha> that was the first thing I said :)
[13:37] <ogra_> oh
[13:37] <rsalveti> nops, there's no session
[13:37] <ogra_> you didnt say that
 btw I'm having a transient (it seems) issue that I click on icons but applications don't open
[13:38] <ogra_> yeah, i see it now
[13:38] <rsalveti> Saviq: ogra_: should unity8 be able to recover itself if I kill it with 'kill -9'?
[13:39] <ogra_> xnox, any idea how to debug such a case ?
[13:39] <Saviq> rsalveti, upstart will restart it, yes
[13:39] <ogra_> should
[13:39] <ogra_> :P
[13:39] <rsalveti> right, so wonder if mir is now causing issues here
[13:39] <rsalveti> my screen is flashing again
[13:39] <rsalveti> tested with mako
[13:39]  * ogra_ doubts it is mir
[13:39] <rsalveti> the new unity8 is running it seems
[13:39] <rsalveti> yeah, probably
[13:40] <rsalveti> I/ServiceManager(  704): service 'display.qservice' died
[13:40] <rsalveti> causing the qcom component to die
[13:40] <ogra_> i guess thats fallout
[13:40] <rsalveti> so specific to mako
[13:40] <rsalveti> logcat is going crazy here
[13:41] <xnox> (disappearning indicators are/will be solved soon, each indicator will be managed by session init / respawned)
[13:41] <seb128> ogra_, we fixed the indicator lock issues in saucy a month ago, it would be interesting to see if it still happen with the trusty-proposed version (could be a different bug)
[13:41] <ogra_> so doing the same kill -9 here i see that it wants to start but approt is keeping the system to busy
[13:41] <rsalveti> ogra_: when did we get a mir update? want to test with an older image
[13:41] <seb128> ogra_, yeah for a world where devel serie is slower than SRUs btw :p
[13:41] <ogra_> (load of 5 after a few seconds)
[13:41] <ogra_> seb128, lock or clock ?
[13:42] <xnox> ogra_: not sure what debug advise i can give you for unity8/mir.
[13:42] <rsalveti> xnox: how to debug when the user session died?
[13:42] <ogra_> xnox, for upstart user sessions
[13:42] <seb128> ogra_, lock (they were not showing in the panel because the service is hitting a glib deadlock on init)
[13:42] <ogra_> seb128, ah
[13:42] <ogra_> seb128, i'm running image r31 and there should be nothing in proposed anymore
[13:43] <ogra_> rsalveti, so clearly apport is heavilly hogging my mako here
[13:43] <seb128> ogra_, indicator-datetime is, it's part of the e-d-s/samba/libav transitions
[13:43] <ogra_> load went above 6 now
[13:43] <xnox> rsalveti: by user session - you mean session init or logind session? cause it's quite valid to be thrown out, but session init still running. Check ps tree and see which dangling processes there are and if they are paranted to any session inits or not.
[13:43] <ogra_> seb128, ah
[13:43] <xnox> session init will try to respawn.
[13:43] <ogra_> so it is held back, k
[13:43] <seb128> yeah
[13:44] <seb128> should be cleared soon hopefully
[13:44] <ogra_> k
[13:44] <xnox> rsalveti: if there are any session inits left, one can get/guess the session init token and investigate running commands. etc. If it really is dead, increase debugging to "--debug" on the session-init, reproduce, check logs why that happened.
[13:45] <xnox> most likely something is killing / terminating it.
[13:46] <xnox> i'm not sure how the session-init is started on the phone, but e.g. mir/unity8 crash shouldn't stop the user-session. Instead they should respawn / restarted into an existing one.
[13:46] <rsalveti> xnox: in Ursinha's case /sbin/initctl list-sessions shows nothing
[13:46] <rsalveti> seems the upstart user session died, and she can't open any other app anymore
[13:47] <ogra_> xnox, the session runs but UPSTART_SESSION is unset
[13:47] <ogra_> xnox, so apps dont start (since they use upstart-app-launch)
[13:47] <ogra_> xnox, the "unity8 not respawning" issue is a differet one
[13:48] <ogra_> rsalveti, iirfc i heard someone talk about the mir socket ... might be that it is still open when unity tries to start anew
[13:48] <xnox> rsalveti: Ursinha: can you get session-init & unity8 processes environment variables from /proc/*/environ ? the intersting ones are UPSTART_SESSION & DBUS_SESSION_ADDRESS
[13:48] <xnox> and see if they are inconsistent.
[13:49] <ogra_> terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::system::system_error> >'
[13:49] <ogra_>   what():  bind: Address already in use
[13:49] <xnox> (and also check the dbus address that is stored somewhere under ~/.cache ~/.config)
[13:49] <ogra_> aha, and there we go
[13:49] <Ursinha> xnox,UPSTART_SESSION is empty
[13:49] <xnox> =(
[13:49] <ogra_> rsalveti, tail -f /home/phablet/.cache/upstart/unity8.log
[13:49] <ogra_> rsalveti, watch it try toi respawn every second
[13:49] <Ursinha> xnox, yep, that was the problem that we noticed because
[13:50] <Ursinha> phablet@ubuntu-phablet:~$ upstart-app-launch webbrowser-app
[13:50] <Ursinha> ** (process:12659): WARNING **: Unable to connect to the Upstart Session: Empty address ''
[13:50] <ogra_> rsalveti, its the mir socket, i have seen that error msg before
[13:50] <Ursinha> xnox, DBUS_SESSION_BUS_ADDRESS  points to a tmp folder
[13:50] <Ursinha> that doesn't exist
[13:50] <Ursinha> oops
[13:51] <ogra_> rsalveti,  rm /var/run/user/32011/mir_socket
[13:51] <ogra_> and see it start
[13:51] <ogra_> Saviq, should we add a post-stop script to the upstart job to remove that file ?
[13:51] <xnox> Ursinha: ls /run/user/1000/upstart/sessions/
[13:51] <xnox> Ursinha: anything there?
[13:51] <ogra_> xnox, not 1000
[13:51] <ogra_> :)
[13:52] <Ursinha> xnox, nope, no such file or directory
[13:52] <ogra_> 32011
[13:52] <xnox> Ursinha: well ok, $ ls $XDG_RUNTIME_DIR/upstart/sessions
[13:52] <Ursinha> there's one file
[13:53] <ogra_> should be the pid
[13:53] <Ursinha> and in it there's something that should be the content of UPSTART_SESSION, I presume
[13:53] <rsalveti> ogra_: oh, ok
[13:53] <rsalveti> it might be the same issue in here
[13:53] <rsalveti> let me check
[13:53] <ogra_> phablet@ubuntu-phablet:~$ env |grep UPSTART
[13:53] <ogra_> UPSTART_SESSION=unix:abstract=/com/ubuntu/upstart-session/32011/1826
[13:53] <ogra_> Ursinha, ^^^
[13:53] <xnox> Ursinha: correct, you should be able to do: UPSTART_SESSION=unix:abstract:..... initctl set-env --global UPSTART_SESSION=unix:abstract:.....
[13:54] <rsalveti> ogra_: yeah, same issue
[13:54] <xnox> Ursinha: this should re-export UPSTART_SESSION to all user-session process and things should start working again.
[13:54] <rsalveti> ogra_: that's so annoying
[13:54] <ogra_> rsalveti, right, i think a post-stop script would help
[13:54] <xnox> Ursinha: but it is very confusing how session-init managed to loose it!
[13:54] <Ursinha> xnox, okay, but why UPSTART_SESSION variable is empty? that's a bug, right?
[13:54] <ogra_> yeah
[13:54] <Saviq> ogra_, or pre-start, for that matter
[13:54] <ogra_> Saviq, yeah, whatever fits best :)
[13:54] <rsalveti> ogra_: actually, I got a different one
[13:54] <xnox> Ursinha: yes, we should never loose that variable no-matter what. Ditto DBUS_SESSION_ADDRESS.
[13:54] <rsalveti> terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >'
[13:54] <rsalveti>   what():  Could not unblank display
[13:55] <ogra_> rsalveti, different what ?
[13:55] <Saviq> rsalveti, press power
[13:55] <ogra_> oh
[13:55] <xnox> Ursinha: is there a way I can reproduce this?
[13:55] <Saviq> rsalveti, bug #1235000
[13:55] <ubot2> Launchpad bug 1236525 in unity-mir "duplicate for #1235000 unity8 killed/crash then restart can result in mir unable "could not unblank display"" [Critical,Fix released] https://launchpad.net/bugs/1236525
[13:55] <rsalveti> Saviq: then it moves to   what():  bind: Address already in use
[13:55] <Saviq> rsalveti, yeah, and now remove the socket ;)
[13:56] <Saviq> ogra_, will a pre-start script be run on restart / respawn?
[13:56] <Ursinha> xnox, it was the first time I had it, the  last thing I did yesterday was to open the check for new updates screen and let it (really slowly) download the changes
[13:56] <ogra_> sure
[13:56] <Ursinha> this morning that app was frozen and then nothing else worked
[13:56] <ogra_> it will be run before the exec ...
[13:56] <xnox> Saviq: ogra_: rsalveti: why is mir using file socket, instead of abstract one? using abstract socket would avoid stale file sockets.
[13:56]  * ogra_ points xnox to #ubuntu-mir 
[13:56] <Saviq> xnox, apport-bug mir please :)
[13:56] <ogra_> :)
[13:57] <ogra_> xnox, probably the same reason why mir needs /dev/input/* to be 777 ;)
[13:57] <rsalveti> Saviq: why is this fixed released?
[13:57] <ogra_> (it worked when someone tried it ... so it stayed that way)
[13:58] <Saviq> rsalveti, good question
[13:58] <ogra_> rsalveti, apparently Mir takes weeks to land in the images
[13:58] <ogra_> it is probably fixed in some branch
[13:58] <rsalveti> oh, there's no bugtask for the package :-(
[13:58] <rsalveti> only tracking it in the upstream project
[13:58] <Saviq> rsalveti, https://bugs.launchpad.net/mir/+bug/1236525/comments/6
[13:58] <ubot2> Launchpad bug 1236525 in unity-mir "unity8 killed/crash then restart can result in mir unable "could not unblank display"" [Critical,Fix released]
[13:59] <ogra_> and another one that wont be on anyones radar
[13:59] <Ursinha> rsalveti, ogra_, are you guys following the "filing bugs against ubuntu packages" thread? Saviq pointed that not everyone can triage ubuntu tasks, and that might be what caused people filing bugs only upstream
[14:00] <Saviq> Ursinha, well, not only that, but just that we didn't know the importance of the difference
[14:00] <ogra_> Ursinha, yes, i saw that, funnily i tried that with some random packages and i actually get the bugmail
[14:01] <ogra_> i dont really get why it doesnt work for Saviq ...
[14:01] <Ursinha> ogra_, you need to be a bugsquad member (or something) iirc
[14:01] <Saviq> ogra_, actually Mir has a clear distinction between upstream and distro bugs and they're looking at both
[14:01] <Saviq> Ursinha, yeah, which kind of sucks
[14:01] <ogra_> Saviq, well, it just means adding a few teams to the bugsquad team ... if thats really the case
[14:02] <Saviq> ogra_, yeah, but that's not ideal is it... it's not like those people/teams *are* actually part of the bugsquad team
[14:02] <Saviq> ogra_, granularity is lost
[14:02] <seb128> Saviq, you are asking for double list/work?
[14:03] <ogra_> usually when a package enters the archive such bits should be set up by the person adding it to the archive
[14:03] <Saviq> seb128, I'm not *asking*
[14:03] <Saviq> seb128, I'm pointing out what doesn't work for us with just distro bugs
[14:03] <ogra_> seems that step wasnt done for any pof the touch packages
[14:03] <Saviq> seb128, I'm fine with defaulting to distro bugs
[14:04] <seb128> Saviq, well, it's a cost/benefit thing
[14:04] <seb128> we should fix the issues for sure
[14:04] <xnox> Ursinha: Saviq: the barrier of entry is to read wiki page of "statuses & priorities" for Ubuntu Packages, special casing slight different statuses for SRUs & then people can be mass added to bugsquad.
[14:04] <seb128> well, I guess it's up to each team/project
[14:04] <seb128> I'm glad we consolidated on one bug list for ubuntu-system-settings
[14:05] <seb128> having to manage upstream/downstream lists is really annoying
[14:05] <rsalveti> ogra_: maybe it's easier to become per package uploader, which would make them ubuntu-dev, and would also be part of bugsquad?
[14:05] <Saviq> seb128, +1
[14:05] <xnox> Ursinha: Saviq: all people that land code into Ubuntu should roughly know what the bug conventions for Ubuntu Packages are.
[14:05] <rsalveti> or just core-dev that is part of bugsquad? don't remember
[14:05] <xnox> well, bugsquad is open to non-developers & non-ubuntu-members =)
[14:05] <ogra_> rsalveti, it would indeed be desired that they become per-package-uploader ... but there is a process ... it involves interest ...
[14:06] <Saviq> xnox, sure, but if someone's not really in bugsquad, 'cause he's not doing what bugsquad does, at least not for projects other than their own
[14:06] <ogra_> (and investing the time to go through it)
[14:06] <xnox> bugsquad is meant to be very lightweight - read a doc, show comment what statuses you'd set on some sample, and off you go caring about a package that you want to care about.
[14:06] <Saviq> ogra_, q: where can I see the reason for "Job failed to start" ?
[14:06] <ogra_> upstart job ?
[14:07] <Saviq> ogra_, ah there ;)
[14:07] <Saviq> right
[14:07] <xnox> Saviq: there is blanket exception for upstreams. All upstream developers can be auto-added to bugsquad, if they promise to set/manager their upstreams' packages bugs only.
[14:07] <ogra_>  /home/phablet/.cache/upstart/$jobid.log
[14:07] <Saviq> ogra_, hmm no, not going there
[14:07] <Saviq> ogra_, since the job never started (pre-start script failed, I presume)
[14:07] <ogra_> is it a system job ?
[14:08] <Saviq> ogra_, session
[14:08] <ogra_> hmm
[14:08] <ogra_> then you need to add debugging to your job script i guess
[14:08] <xnox> Saviq: if any command exists non-zero and doesn't print anything the job will fail to start and there won't be anything in the log.
[14:08] <ogra_> right
[14:08] <ogra_> add some echos to the pre-start script
[14:09] <xnox> Saviq: pre-start/script/exec are all run under "set -e"
[14:09] <Saviq> xnox, ogra_ah ok
[14:12] <xnox> Ursinha: can you try as phablet user "$ initctl --user --no-wait start re-exec" and check if that looses UPSTART_SESSION variables everywhere?
[14:12] <mitya57> Hi, what is blocking e-d-s migration? Quite a few packages are stuck in proposed because of that.
[14:13] <xnox> mitya57: very little left, samba transitional packages and thunderbird.
[14:13] <xnox> mitya57: it's actually entangled together with libav9 (~100) packages transition, osg transition, and a few others.
[14:13] <xnox> mitya57: it will migrate soon.
[14:15] <mitya57> xnox: thanks
[14:21] <Ursinha> xnox, sure
[14:22] <Ursinha> sorry, was getting the daily dose of coffee
[14:22] <xnox> =))))) coffee is good!
[14:22] <rsalveti> ogra_: did you open a bug for the mit socket issue?
[14:22] <rsalveti> *mir
[14:22] <ogra_> nope
[14:22] <Ursinha> xnox, oh yeah :)
[14:22] <ogra_> rsalveti, there must be one, this has occured so often in the past
[14:22] <Ursinha> xnox, initctl: invalid option: --no-wait
[14:23] <rsalveti> ogra_: right
[14:23] <xnox> Ursinha: weird, move "--no-wait" after "start" ?
[14:24] <xnox> or try without --no-wait
[14:24] <Ursinha> xnox, initctl: Unable to connect to Upstart: Empty address ''
[14:24] <xnox> Ursinha: ah, so you didn't reset the UPSTART_SESSION from the file? or did it now loose it again.
[14:25] <Ursinha> xnox, no, I haven't reset that so I didn't lose the "buggy" condition :)
[14:25] <Ursinha> but I can do that, yes
[14:26] <xnox> Ursinha: well, I know how to artificially cause buggy condition - initctl unset-env --global UPSTART_SESSION
[14:27] <xnox> i'm trying to figure out how to reproduce it. My thought is that re-exec maybe loosing it, thus the attempt to re-exec session init and see if that goes from good -> buggy.
[14:27] <xnox> (under actual operating conditions, not artificially induced)
[14:27] <Ursinha> xnox, right
[14:28] <Ursinha> xnox, I tried to set the current env to that and I got this: initctl: Unable to connect to Upstart: Failed to connect to socket /com/ubuntu/upstart-session/32011/1625: Connection refused
[14:28] <Ursinha> s/env/UPSTART_SESSION/
[14:30] <xnox> hm.
[14:30] <xnox> $  export `cat $XDG_RUNTIME_DIR/upstart/sessions/*.session`
[14:31] <xnox> $ initctl --user set-env -g UPSTART_SESSION=$UPSTART_SESSION
[14:31] <xnox> $ initctl list-env
[14:31] <xnox> $ initctl list-env | grep UPSTART_SESSION
[14:32] <xnox> Ursinha: ^ does above sequence recovers ? and everything starts to know about UPSTART_SESSION?
[14:32] <Ursinha> xnox, the initctl command gives me the Connection refused error
[14:32] <xnox> Ursinha: all executed as the phablet user?
[14:33] <Ursinha> it seems the upstart session information isn't valid somehow?
[14:33] <Ursinha> xnox, yes
[14:33] <xnox> Ursinha: does the pid of the session filename matches the pid of session upstart?
[14:34] <xnox> to be honest it's weird that we encode PID, given that we can re-exec.
[14:35] <Ursinha> xnox, what's the name of the process I should look? there are a couple upstart-something when I ps (none with that PID)
[14:36] <ogra_> ps ax|grep "init --user"
[14:36] <ogra_> try that
[14:36] <xnox> Ursinha: pidof init
[14:36] <Ursinha> oh, init
[14:36] <ogra_> or that :)
[14:36] <Ursinha> ogra_, your command returns only grep itself :)
[14:37] <ogra_> phablet@ubuntu-phablet:~$ ps ax|grep "init --user"
[14:37] <ogra_>  1826 ?        S      0:17 init --user
[14:37] <ogra_>  9051 pts/27   S+     0:00 grep --color=auto init --user
[14:37] <Ursinha> xnox, 730 and 1
[14:37] <ogra_> returns the pid for me
[14:37] <xnox> Ursinha: and the session file name is 730.session? in ls $XDG_RUNTIME_DIR/upstart/sessions/*
[14:38] <xnox> Ursinha: also looks like your session started a lot earlier than ogra_'s =)
[14:38] <Ursinha> xnox, haha no: /run/user/32011/upstart/sessions/1625.session
[14:38] <ogra_> yeah, thats a strangely low PID
[14:38] <xnox> ogra_: unless Ursinha's looped ;-)
[14:38] <Ursinha> ogra_, I don't reboot my mako often, I use it as my regular phone so no need to
[14:38] <xnox> and re-execed.
[14:39] <xnox> Ursinha: ah, what's the upstart package version?
[14:39] <ogra_> well, i get 681 here when i run it as root
[14:39] <xnox>  ... we did fix a couple user-session bugs in latest upload.
[14:39] <ogra_> (alongside with 1)
[14:39] <Ursinha> hmm
[14:39] <Ursinha> xnox, 1.11-0ubuntu1
[14:40] <xnox> which is the latest one =(
[14:40] <ogra_> upstart-app-lauch was uploaded recently
[14:40] <ogra_> but i dont see how that would have any impact on the session
[14:41] <ogra_> nor do i understand why the UPSTART_SESSION var isnt set
[14:41] <ogra_> since that has to be processed on a new login
[14:41] <ogra_> so at least in the second adb session you should have it set properly
[14:41] <Ursinha> ogra_, yeah, I opened many of them and they have all the same behavior
[14:42] <ogra_> Ursinha, does the unity UI itself recat ... can you scroll etc ?
[14:42] <ogra_> *react
[14:42] <Ursinha> ogra_, I can use it normally, it seems, can't only open apps
[14:42] <ogra_> k
[14:42] <xnox> Ursinha: let's see if can find the active one: does $ lsof -U | grep init
[14:43] <xnox> Ursinha: returns anything like @/com/ubuntu/upstart-session/*/730
[14:43] <xnox> ?
[14:43] <Ursinha> xnox, is that supposed to be in the read-only image? if not I don't have it
[14:43] <ogra_> no lsof
[14:43] <Ursinha> lsof
[14:43] <xnox> darn.
[14:43] <Ursinha> yeah :/
[14:43] <ogra_> (no fuser either in case you wanted to fall back to that)
[14:44] <xnox> .... quickly cross-compile statically linked armhf lsof and run that..... (got to love debugging read-only images)
[14:44] <Ursinha> lol
[14:44] <xnox> hm, i wonder how we can guess it. well. it should be.
[14:44] <Ursinha> ogra_, can't we have that now that debugging is important, and remove it later?
[14:45] <ogra_> hmm
[14:46] <xnox> well it should be UPSTART_SESSION=unix:abstract=/com/ubuntu/upstart-session/$uid/$pid, so try something like: UPSTART_SESSION=unix:abstract=/com/ubuntu/upstart-session/$uid/730 initctl --user list-env
[14:46] <ogra_> doesnt have any extra deps, i suppose we could add it, yeah
[14:46] <xnox> Ursinha: ^
[14:46] <xnox> Ursinha: and 730 process is owned by phablet user right?
[14:46] <Ursinha> xnox, no, root
[14:46] <Ursinha> is that right?
[14:47] <ogra_> xnox, thats androids init
[14:47] <ogra_> Ursinha, ps ax|grep 730
[14:47] <ogra_> Ursinha, does it have a / in front ?
[14:47] <xnox> Ursinha: these are not the inits we are looking for
[14:47] <Ursinha> ogra_, I did that, I know how to check a process ownership haha
[14:47] <ogra_> Ursinha, but is there a / ?
[14:47] <ogra_> then it is androids init
[14:47] <Ursinha> pid 1 is /sbin/init, pid 730 is /init
[14:48] <ogra_> right
[14:48] <Ursinha> both root's
[14:48] <Saviq> ogra_, https://code.launchpad.net/~saviq/unity8/drop-stale-socket/+merge/196917
[14:48] <Ursinha> and that's about it
[14:48] <ogra_> there needs to be a third one
[14:48] <xnox> Ursinha: so where did the 1625 go =) upstart logs should have a disconnect from session-init then.
[14:48] <Ursinha> xnox, good question hehe
[14:48] <ogra_> Ursinha, that means there is no upstart session at all ... now how did you get unity to ever run ...
[14:49] <xnox> Ursinha: are there any /var/crash/* files?
[14:49] <xnox> (unless whoopsie puts them elsewhere on the phone)
[14:49] <ogra_> nope, same place
[14:49] <Ursinha> there are a bunch
[14:49] <xnox> Ursinha: anything initish?
[14:49] <xnox> Ursinha: you may want to fetch all of them.
[14:50] <Ursinha> there's a unity8 one, one upstart-app-launch, one sbin-system-image-dbus
[14:50] <Ursinha> others are webbrowser, system-settings and other apps
[14:51] <ogra_> no upstart ?
[14:51] <ogra_> hmm
[14:51] <Ursinha> let me fetch them all
[14:52] <ogra_> did you by chance run out of disk space
[14:52] <Ursinha> ogra_, doesn't seem so, only with 99% is /dev/loop1 (/lib/modules)
[14:52] <xnox> ogra_: where is init --user spawned on the phone? and is the output redirected? e.g. on desktops init's stdout/err ends up in ~/.xsessions-errors
[14:53] <xnox> ogra_: do we have equivalent on the phone?
[14:53] <ogra_> xnox, lightdm spawns it iirc
[14:53] <xnox> ogra_: and redirects stderr to /dev/null?
[14:54] <ogra_> not sure where lightdm redirects ...
[14:54] <xnox> yeah, i want to know where the output is redirected.
[14:54] <ogra_> /var/log/lightdm/lightdm.log should have some info
[14:54] <xnox> ogra_: well lightdm spawns a session init for itself (lightdm-session) or does it do (the session session-init) on the phone as well?
[14:54] <Ursinha> whoa, unity8 crash has 55Mb
[14:55] <ogra_> xnox, not in autologin mode
[14:55] <xnox> ah, right.
[14:55] <ogra_> [+0.26s] DEBUG: Registering session with bus path /org/freedesktop/DisplayManager/Session0
[14:55] <ogra_> [+0.26s] DEBUG: Session pid=1772: Running command /usr/sbin/lightdm-session ubuntu-touch-session
[14:55] <ogra_> [+0.27s] DEBUG: Session pid=1772: Logging to .xsession-errors
[14:55] <ogra_> aha
[14:55] <ogra_> so yeah, some bits are logged there
[14:56] <xnox> Ursinha: cat ~/.xsession-errors please =)
[14:56] <Ursinha> yay, there's a segfault
[14:56] <xnox> \o/
[14:56] <ogra_> great
[14:56] <xnox> Ursinha: pastebin that to me.
[14:56] <Ursinha> \o/
[14:56] <xnox> on the other hand, init just segfaulted /o\
[14:56] <Ursinha> xnox, http://pastebin.ubuntu.com/6484724/
[14:56] <Ursinha> lol

[14:57] <xnox> Ursinha: since it core-dumped there should have been a whoopsie .crash file.... ?!
[14:57] <Ursinha> so it seems my phone is now kind of a zombie?
[14:58] <Ursinha> xnox, http://paste.ubuntu.com/6484729/
[14:58] <ogra_> xnox, well, it doesnt say it was init that segfaulted ... "something" segfaulted
[14:58] <Ursinha> the list of crash files
[14:58] <ogra_> heh, apport crashed :P
[14:59] <Ursinha> crashed while processing a unity8 crash?
[15:00] <Ursinha> oh well
[15:00] <xnox> Ursinha: interesting why apps were getting killed with KILL / TERM / ABRT
[15:00] <ogra_> because the session died
[15:01] <xnox> Ursinha: are there any processes left as phablet user running? and what's their environ. We want something pid larger 1625 but very close to it?
[15:01] <Ursinha> let me see
[15:02] <xnox> hm, .xsession-errors ways that process 1625 asserted.
[15:02] <xnox> s/ways/says/
[15:02] <Ursinha> xnox, http://pastebin.ubuntu.com/6484736/
[15:02] <Ursinha> all others have much higher pid
[15:04] <xnox> Ursinha: How odd.... session-init should start dbus, yet it looks like it was started ahead of us by pid 1618
[15:05] <Ursinha> xnox, it seems nothing makes much sense in this scenario.. how did it get there?
[15:06] <xnox> Ursinha: can you cat /proc/1672/environ and check if has DBUS_SESSION_BUS_ADDRESS set & UPSTART_SESSION
[15:07] <xnox> Ursinha: if not, all of them are self-spawned zombies. Cause all of the indicators, for example, are dbus activated and prefer to do upstart activation, but silently fallback to direct exec.
[15:07] <xnox> actually.
[15:07] <xnox> Ursinha: i bet all of these are dbus-activated, unity8 including.
[15:07] <happyaron> seb128: RHEL7 is using ibus 1.5.3 at the moment
[15:07] <Ursinha> UPSTART_SESSION=unix:abstract=/com/ubuntu/upstart-session/32011/1625 and DBUS_SESSION_BUS_ADDRESS=unix:abstract=/tmp/dbus-Qrg3Aw75Zy,guid=dee63f588c7e05843689419552949b55
[15:07] <xnox> Ursinha: or not, so dconf was spawned by now dead session-init.
[15:09] <Ursinha> xnox, I assume I should file a bug about it... where should I start, it seems everything is odd
[15:09] <xnox> with pid1 atleast kernel panics, but with session-init it all just keeps on running.
[15:10] <Ursinha> right
[15:10] <xnox> Ursinha: i'm downloading all the pastebins and filing a bug report. I'll paste a link once i'm done.
[15:10] <Ursinha> xnox, thank you :)
[15:10] <xnox> Ursinha: hence, i don't think post-segfault upstart code-path has ever been possible to exercise =)
[15:11] <Ursinha> hehe
[15:11] <Ursinha> let me know if you need more info, I'll wait for you to finish so I can restart my phone and use it :)
[15:11] <xnox> imho system init should notice, disconnect, no-reconnect, stale session file & re-create one.
[15:11] <xnox> Ursinha: yes, asap. Thanks a lot for this!
[15:11] <ogra_> xnox, well, the user session is in the hands of lightdm
[15:12] <Ursinha> xnox, no problem. thanks for helping debugging it :)
[15:12] <ogra_> should system init still pick that up even if i.e. lightdm wouldnt inform it
[15:13] <seb128> happyaron, I saw your comment on the google doc, thanks for checking
[15:13] <happyaron> seb128: so do you still want to merge 1.5.4 (not counting the GTK input purpose problem)?
[15:14] <seb128> happyaron, is there big differences between those versions?
[15:15] <happyaron> seb128: not too many commits' diff, https://github.com/ibus/ibus/commits/master
[15:16] <seb128> happyaron, ok, seems like a stable update
[15:16] <happyaron> but some of them could be big diff in a single commit, like "Implement IME state per window. "
[15:16] <xnox> Ursinha: one more thing. I wonder if upstart session-state was serialised anywhere. cause one could start session init & pass fd to it with json state information. No need to reboot at all =)
[15:17] <happyaron> seb128: seems there's not the concept of stable update at ibus upstream, every point release is implementing new stuff plus bug fixes. and downstream maintain a list of patches in their distro, including Fedora.
[15:18] <seb128> happyaron, is that commit fixing the state issue we were discussing?
[15:18]  * seb128 reads http://code.google.com/p/ibus/issues/detail?id=1568
[15:18] <happyaron> no
[15:18] <Ursinha> xnox, so what's the black magic I should apply here? :)
[15:19] <seb128> happyaron, are you sure? ;-)
[15:19] <happyaron> seb128: so far it still doesn't work for me, so I think it's not complete at least.
[15:21] <seb128> happyaron, seems like ibus upstream is wanting to get that resolved though, that's good
[15:21] <xnox> Ursinha: i'm not sure =) something like $ find -name "upstart.state*" across home directory & XDG_RUNTIME_DIR?
[15:22] <xnox> but i highly daubt it will turn anything up.
[15:22] <happyaron> seb128: yeah good, restoring old codes...
[15:23] <Ursinha> xnox, .cache/upstart has a lot of logs, and I could find a dbus-session file that contains DBUS_SESSION_BUS_ADDRESS
[15:23] <xnox> yeah, that's normal.
[15:23] <xnox> no files with "state" in their name?
[15:23] <xnox> and json inside them
[15:24] <Ursinha> xnox, nope
[15:24] <Ursinha> neither at ~ or XDG_RUNTIME_DIR
[15:25] <xnox> Ursinha: most of the crash files are probably caused by not being able to connect to session init. So i guess it's reboot time.
[15:25] <xnox> =(
[15:25] <Ursinha> :(
[15:25] <Ursinha> okay
[15:26] <xnox> Ursinha: https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/1255569
[15:26] <ubot2> Launchpad bug 1255569 in upstart (Ubuntu) "session-init-less session-init session" [Medium,Confirmed]
[15:27] <ogra_> lovely title
[15:28] <Ursinha> thanks xnox
[15:29] <Ursinha> updating to r33
[15:58] <rsalveti> larsu: got the issue
[15:58] <rsalveti> larsu: http://paste.ubuntu.com/6485009/
[15:59] <rsalveti> larsu: the first message from 27800 is wrong in here
[16:04] <rsalveti> larsu: http://paste.ubuntu.com/6485021/ and http://people.canonical.com/~rsalveti/indicator.png
[16:06] <rsalveti> also updated bug 1253810
[16:06] <ubot2> Launchpad bug 1253810 in unity8 (Ubuntu) "Messages in Incoming not always display the correct date and content" [High,Confirmed] https://launchpad.net/bugs/1253810
[16:37] <Laney> bah, we really should parallelise the png optimisation step
[16:38] <ogra_> we should stick it into dput :P
[16:39] <ogra_> (or debuild)
[17:06] <attente> seb128, do you know anything about the packaging in this bug? https://bugs.launchpad.net/unity-gtk-module/+bug/1250213
[17:06] <ubot2> Launchpad bug 1250213 in Unity GTK+ module "dependency error" [Undecided,New]
[17:35] <robru> seb128, around? can you preNEW lp:unity-voice?
[17:47] <kenvandine> seb128, do you know if there is a way to set the opacity on the header?
[17:47] <kenvandine> or i guess change the style programatically?
[17:47] <kenvandine> or anyone ^^
[17:53] <seb128> attente, hey, if the content is not arch dependent, I think it should just be marked "Multi-Arch: foreign" (see https://wiki.debian.org/Multiarch/Implementation)
[17:53] <seb128> robru, sil2100 already asked me about that, I said I would try to have a look but I had a busy day, maybe tomorrow if didrocks doesn't beat me to it
[17:54] <seb128> kenvandine, no idea sorry, what are you trying to do? sounds like against visual consistency to do that
[17:54] <sil2100> seb128: thanks :)
[17:54] <seb128> sil2100, yw ;-)
[17:54] <kenvandine> seb128, the background panel
[17:55] <kenvandine> the spec calls for a preview page that makes the header and toolbar semi-transparent
[17:55] <kenvandine> so you can see what the wallpaper would look like if you set it
[17:56] <seb128> kenvandine, seems like something that could be useful in other places and could be built into the toolkit...
[17:56] <seb128> kenvandine, but no, dunno how to do it otherwise, sorry
[17:57] <kenvandine> Kaleo, ideas? ^^
[17:58] <attente> seb128, thanks, i'll try it
[18:10] <kenvandine> seb128, oh.. i found a hack that works :)
[18:10] <kenvandine>         pageStack.currentPage.header.opacity = 0.5;
[18:11] <seb128> kenvandine, nice
[18:11] <kenvandine> sort of
[18:11] <kenvandine> you just have to make sure when you pop to reset it :)
[18:11] <seb128> kenvandine, I hope it doesn't stop working at some point because you use non documented stuff and they rework Page or something
[18:11] <kenvandine> indeed
[18:11] <seb128> we should have a test for that sort of stuff
[18:12] <seb128> so if it does stop working we know
[18:12] <kenvandine> works for the toolbar as well
[18:15]  * didrocks waves good evening
[18:18] <robru> pitti, still around at this late hour? how goes the python3 transition for autopilot? I'm seeing a weird friends_app AP failure that's not my fault, wondering if you broke it ;-)
[18:32] <Kaleo> kenvandine, sorry, let me read
[18:33] <Kaleo> "seb128: kenvandine, no idea sorry, what are you trying to do? sounds like against visual consistency to do that"
[18:33] <Kaleo> he is right
[18:33] <kenvandine> i have a hack that  works, but not elegant
[18:33] <Kaleo> kenvandine, who is the designer? did he talk with the others?
[18:33] <kenvandine> Kaleo, it's for previewing what the wallpaper will look like
[18:33] <kenvandine> mpt
[18:33] <Kaleo> kenvandine, specifically katie and jouni
[18:34] <Kaleo> kenvandine, you mean the app to be see through?
[18:34] <seb128> kenvandine, imho we should just hide the header/go fullscreen if we want to use the space for content
[18:34] <kenvandine> yeah
[18:34] <kenvandine> hiding the header probably makes the most sense
[18:34] <Kaleo> kenvandine, yes to seb128? or to me?
[18:34] <Kaleo> :)
[18:34] <Kaleo> ok
[18:35] <kenvandine> although, showing it does hint the user that they are previewing
[18:35] <seb128> kenvandine, let's forget about the hack and talk to the designers about an alternative solution
[18:36] <kenvandine> yeah
[18:36] <Kaleo> seb128, kenvandine, if the agreed solution is still the same we can see to put it in the toolkit
[18:37] <kenvandine> cool
[18:37] <seb128> Kaleo, thanks
[19:38] <attente> sil2100, is there a reason for this line in u-g-m's debian/control? http://bazaar.launchpad.net/~indicator-applet-developers/unity-gtk-module/trunk.14.04/view/head:/debian/control#L10
[19:38] <attente> it seems to issue a warning that it's not a known field
[19:43] <seb128> attente, that seems buggy placed, it should be on the binaries, not on the source
[19:43] <seb128> well, binaries that provide a shared lib
[19:43] <seb128> cf https://wiki.debian.org/Multiarch/Implementation
[19:44] <seb128> if there is no shared lib, just drop the line
[19:44] <attente> seb128, thanks
[19:44] <seb128> yw!
[19:44] <attente> seb128, when they say shared libs, do they also mean the -dev packages?
[19:45] <seb128> no, just libraries (I think it's for shlibs/depends resolution)
[19:46] <attente> ok, cool
[19:47] <attente> so.. uh.. how do i know if this all works in the end?
[19:47] <attente> the result seems indistinguishable from what trunk produces on my machine...
[19:48] <seb128> I guess you use amd64
[19:49] <seb128> can you install the :i386 variant of one of the binaries without making apt unhappy?
[19:49] <seb128> reading the description it seems that didn't work before because the -common was not multiarched correctly
[19:50] <seb128> I guess you can build a package on amd64 using the same package version than the archive
[19:50] <seb128> then try to apt-get install the binary which is in the description
[19:50] <seb128> (you might need to add an i386 source to your sources.list if you never did that)
[19:53] <attente> seb128, the default repositories don't already provide i386 packages?
[19:55] <seb128> attente, I don't know, I'm on i386, try to sudo apt-get install gedit:i386 or something and see if that works
[19:57] <attente> seb128, i get a lot of unmet dependencies errors in that case
[19:57] <seb128> attente, dpkg --print-architecture
[19:58] <seb128> attente, you might need to sudo dpkg --add-architecture i386
[19:58] <attente> yeah, forgot to apt-get update
[19:59] <seb128> I don't know, gedit might have depends that are not properly mutilarched
[19:59] <seb128> try maybe something easier like file-roller or eog
[19:59] <seb128> sorry, that's as much as I know on the topic
[20:00] <sil2100> attente: indeed, as seb128 said ^
[20:01] <attente> hmm.. maybe i need a reboot
[20:10] <seb128> attente, reboot? that's not windows here ;-) you might need some sources.list hacks
[20:10] <seb128> e.g https://wiki.debian.org/Multiarch/HOWTO has "deb [arch=amd64,i386]..."
[20:11] <seb128> ok, need to go, good luck with that
[20:11] <seb128> bye