[05:40] <BigWhale> God Noring.
[05:40] <BigWhale> ...
[05:40] <BigWhale> Good Morning.
[06:03] <pitti> Good morning
[06:31] <RAOF> pitti: Good morning.
[06:35] <RAOF> Bah.  I don't really want to change the libxfixes soname.
[06:41] <RAOF> pitti: Ho!
[06:42] <pitti> hey RAOF, how are you?
[06:42] <pitti> my server rebooted apparently, I might have lost a few messages
[06:42] <RAOF> Trying to figure out how to not trigger a libxfixes transition :)
[06:43] <RAOF> But I need to rejig the layout of a user-visible structure.  Because it's wrong.
[06:44] <broder> RAOF: btw, i *think* i figured out that my lightdm problems were due to a combination of racing with /etc/init/hostname.conf and not having an /etc/hostname file. i don't know exactly what happened, but i added a dummy hostname file and everything was fine
[06:44] <RAOF> broder: Cool.
[06:44] <RAOF> pitti: Would it be ok to break the xfixes ABI, not bump soname, and just declare a breaks: on unity?  It's the only client of that ABI.
[06:45] <pitti> RAOF: but when that's going upstream, they most certainly want to bump the soname?
[06:45] <pitti> RAOF: i. e. what do you win by not bumping it?
[06:46] <RAOF> The patches themselves don't break ABI.
[06:46] <RAOF> Or, at least, the *final* patches don't break ABI.
[06:46] <RAOF> The current state is that it's broken, and fixing it requires breaking abi, but it's breaking a (currently) Ubuntu-only ABI.
[06:47] <pitti> RAOF: ah; that sounds alright to me then; but please, if you upload it, upload unity along, to avoid breaking the archive
[06:47]  * RAOF shall coordinate with didrocks.
[06:47] <pitti> ... precise-proposed!
[06:48] <RAOF> I can do that?
[06:48] <RAOF> If so, hells yes!
[06:48] <pitti> not yet
[06:48] <pitti> bug 930217 is standing between us and always working archives
[06:48] <ubot2`> Launchpad bug 930217 in launchpad "Make proposed pocket useful for staging uploads" [Low,Triaged] https://launchpad.net/bugs/930217
[06:49] <RAOF> Bah.  You tempter, you.
[06:49] <broder> RAOF: want a patch that fixes that bug? all you have to do is write test cases!
[06:50] <broder> (to be clear, the patch itself is dead simple)
[06:51] <RAOF> Ah.
[06:51] <RAOF> How about the test cases?  Simple?
[06:51] <broder> unknown
[06:51] <broder> http://paste.ubuntu.com/881517/ is the patch, though
[06:52] <broder> doesn't solve the whole bug - it won't cause -proposed uploads to get auto-accepted
[06:52] <broder> but it will prevent them from getting immediately rejected
[07:23] <didrocks> good morning
[07:23] <RAOF> didrocks: Good morning!  Can I coordinate a unity upload with you today (after dinner)?
[07:24] <didrocks> RAOF: sure! what about?
[07:25] <RAOF> I need to break libxfixes3 abi, and unity is the only user of that abi, so rather than introduce an Ubuntu-specific soname and do a transition on the 90-odd rdepends I'd like to just rebuild unity.
[07:25] <didrocks> RAOF: easy! yeah, let's do that after your dinner :)
[07:27] <RAOF> Awesome.
[07:28] <pitti> hey didrocks
[07:30] <didrocks> guten morgen pitti
[07:38] <pitti> meh, something changed the Alt key bindings again
[07:38] <pitti> my alt+left/right bindings are broken again in weechat
[07:38] <pitti> and alt+k doesn't work either any more to define a new one
[07:39] <didrocks> pitti: we had so much issues with alt that's not surprising (5 patches as I mentionned). I didn't notice those though
[07:40] <didrocks> I'm trying to be a full day in -2d now, do you want me to login in -3d to test?
[07:40] <didrocks> pitti: I can just tell that alt + Num worked and alt + a as well (that's what I use all the time in weechat)
[07:40] <pitti> yes, those do
[07:40] <pitti> and indeed alt+right still seems to work
[07:41] <pitti> but nto my alt+left/alt+up bindings
[07:41] <pitti> and /key (meta-k) (press key combo for binding) doesn't work either any more
[07:48] <jibel> didrocks, morning
[07:49] <didrocks> hey jibel
[07:49] <jibel> didrocks, with latest unity update, shortcut keys doesn't work if there is not a window already open on the workspace
[07:49] <didrocks> pitti: care about opening bug for those?
[07:49] <jibel> didrocks, is there an existing bug report ?
[07:49] <pitti> didrocks: can do, yes
[07:49] <didrocks> jibel: urgh, like workspace switcher?
[07:50] <didrocks> or you mean the nautilus desktop shortcuts?
[07:50] <jibel> didrocks, switcher, super key, custom shortcut, ...
[07:50] <jibel> nautilus desktop as well
[07:50] <didrocks> jibel: nautilus desktop never worked
[07:50] <didrocks> I'm surprised about the other, I didn't get that at all
[07:50] <didrocks> ok, I think I can't today be the full day in -2d, logging out, logging back in 3d
[07:51] <jibel> didrocks, for example right after login, no window open, CTRL+ALT+arrow doesn't work
[07:52] <jibel> bah, gone
[07:54] <didrocks> jibel: I definitively can switch ws after a login without any application opened here
[07:55] <didrocks> pitti: alt + right works here, not alt + left, confirmed?
[07:55] <jibel> didrocks, for example right after login, no window open, CTRL+ALT+arrow doesn't work
[07:55] <didrocks> jibel: it does work here, I just tried that :/
[07:56] <jibel> didrocks, if I open a terminal for example, it works
[07:56] <didrocks> and it's part of the tests that people did
[07:56]  * didrocks boot a second machine
[07:56]  * jibel tries on another system
[08:00] <didrocks> jibel: works on my netbook too
[08:10] <didrocks> RAOF: you really made a high score bug-wise btw with breaking the launcher reveal behavior :p
[08:19] <pitti> didrocks: right, and alt+up doesn't work either
[08:21] <didrocks> pitti: the weird part is that xev seems to get the event
[08:21] <jibel> didrocks, I'm not alone :) bug 953089
[08:21] <ubot2`> Launchpad bug 953089 in unity "Lots of regressions after update to Unity 5.6 like not working super key or alt key functionality" [Undecided,Confirmed] https://launchpad.net/bugs/953089
[08:21] <didrocks> pitti: with the modifier
[08:22] <pitti> didrocks: but in g-terminal, when I press alt+right, I get a "C"
[08:22] <pitti> I get nothing at all when I press alt+left/up
[08:23] <pitti> I thought previously this would print some garbage
[08:23] <didrocks> jibel: you mean that you don't have -2d starting as well?
[08:24] <didrocks> pitti: yeah, you're right
[08:24] <pitti> argh
[08:24] <jibel> didrocks, I didn't try 2d,broken shortcuts is my immediate problem
[08:24] <pitti> now I pressed F10 and Alt+F10; they don't do what they are supposed to do, but now they messed up unity completely
[08:24] <pitti> brb, have to restart
[08:25] <didrocks> jibel: your HUD is assigned to alt by any chance?
[08:28] <jibel> didrocks, it is, shouldn't it ? but all keybindings are broken, including <Super>
[08:28] <jibel> unless I open a window
[08:30] <didrocks> jibel: yeah, I was wondering if it can be linked to alt grabbing something
[08:30] <didrocks> do you reproduce on your second machine?
[08:30] <didrocks> I think a lot of people would have yielled before (and it would have been shown in the testing result) if it was widespread
[08:30] <jibel> didrocks, I disabled alt for the hud, and the only shortcut that works is a custom one to open a terminal
[08:31]  * jibel still waiting for his netbook to finish upgrade
[08:35] <chrisccoulson> yeah, i disabled the hud shrortcut because it breaks my alt shortcuts in vim, and now alt+tab doesn't work :(
[08:40] <didrocks> pitti: there is maybe a potential fix for alt + arrows
[08:41] <didrocks> pitti: basically a workaround in unity which has a good fix now in compiz
[08:41] <didrocks> chrisccoulson: tapping alt should work now as expected, you can bring back the HUD
[08:43] <pitti> didrocks: ah, good to hear!
[08:44] <tjaalton> hmm, if the "wall sliding duration" is 0, moving windows around gets weird
[08:44] <jibel> didrocks, another interesting effect. I've a keybinding set to CTRL+Fn to switch to workspace n
[08:45] <jibel> didrocks, if firefox has focus and I press CTRL+F4 to switch to ws4, it switches to ws4 but also closes the active firefox tab
[08:46] <didrocks> jibel: I pinged you on the upstream channel, you really should discuss it there :)
[08:46] <chrisccoulson> didrocks, i tried it yesterday, it still breaks my bindings in vim, even with the latest version :(
[08:46] <didrocks> chrisccoulson: hum, what exacty is broken? alt + some keys relveal the HUD?
[08:46] <chrisccoulson> didrocks, i map alt+arrow to move tabs in vim, and they just do nothing with the hud enabled
[08:47] <didrocks> chrisccoulson: see above ^
[08:47] <didrocks> chrisccoulson: same issue than pitti
[08:47] <chrisccoulson> ah, ok :)
[08:47] <didrocks> I'm building a fix for it, hopefully
[08:48] <chrisccoulson> thanks
[08:52] <pitti> chrisccoulson: you use alt bindings in vim? that sounds evil
[08:52] <chrisccoulson> pitti - yeah, i just use alt+l/r to switch between tabs
[08:53] <chrisccoulson> and it's always worked fine until the hud stuff landed in unity ;)
[08:53] <didrocks> pitti: btw, reminder day!
[08:53] <pitti> oh right! thanks
[08:53] <chrisccoulson> so i disabled that but then alt+tab stopped working when the hud is disabled
[08:53] <didrocks> chrisccoulson: don't speak about alt to me ever ever again :)
[08:53] <chrisccoulson> lol
[08:59] <pitti> will unity stop 0wning more and more keys for itself? :-)
[09:00]  * rickspencer3 decides to change the topic
[09:00] <rickspencer3> hey pitti, quick Python(y) question ...
[09:01] <rickspencer3> how, generally, would I go about using Python to download a file on chinstrap?
[09:01] <rickspencer3> (given that it's behind SSO?)
[09:01] <pitti> rickspencer3: ah, through http, not through ssh
[09:02] <pitti> rickspencer3: that's a good question; I tried that a while ago in apport, and it was insurmountable for me as well
[09:02] <rickspencer3> insurmountable for pitti? that doesn't sound promising :(
[09:02] <pitti> well, let's say I gave up after half an hour
[09:02] <rickspencer3> lol
[09:02] <rickspencer3> I was wondering if this SSO client package might help?
[09:02] <pitti> I suppose it actually needs to use python-ubuntu-sso-client
[09:03] <pitti> but that's not something I wanted apport to depend on
[09:03] <rickspencer3> fair enough
[09:03] <pitti> I didn't find a good way with just urllib and cookies
[09:03] <rickspencer3> yeah
[09:03] <pitti> and we got a proper launchpadlib API for this, so I didn't need it any more
[09:03] <rickspencer3> you'd probably have to open a webkit window, let the sure sign in through SSO, store the credentials, and then send them through an http request
[09:04] <rickspencer3> well, i want to automatically download some bdmurray's json reports
[09:04] <pitti> rickspencer3: so far I didn't have to use secret files
[09:04] <pitti> rickspencer3: hm, is there any reason to keep them secret?
[09:04] <rickspencer3> I may just ask him to put them somewhere not secret
[09:04] <rickspencer3> pitti, right, not that I know of
[09:04] <pitti> rickspencer3: it would be much easier to mirror them to people or macquarie or so, and just use plain http
[09:04] <rickspencer3> I just assumed it was more convenient for him to plop them on his chinstrap account
[09:05] <rickspencer3> so, I thought if I could write a line of code or two, I wouldn't have to bother him
[09:05] <pitti> they should run just as well on lillypilly, and if not, we can set up rsyncing
[09:05] <rickspencer3> ok
[09:05] <rickspencer3> I'm basically in the process of slowly rewriting bughugger to only use his json reports
[09:06] <rickspencer3> so, if the reports were accessible, my tool would be generally useful to the community, I think
[09:06] <rickspencer3> using launchpad lib was just too much code to maintain, and the perf made the tool unusable
[09:07] <rickspencer3> didrocks, btw, I'm dist-upgrading right now
[09:07] <rickspencer3> I promise I won't complain about key bindings :)
[09:08] <didrocks> rickspencer3: it's kind of desperating, 5 patches for alt, blocking for 2 weeks and you log in the morning and pouf
[09:08] <didrocks> pitti: the patch didn't work :/
[09:08]  * didrocks grumbles
[09:08] <pitti> :(
[09:08] <seb128> hey
[09:08] <rickspencer3> didrocks, oh, this is the thing to make the hud not come up so much?
[09:08] <pitti> bonjour seb128
[09:08] <seb128> pitti, hey, how are you?
[09:09] <pitti> seb128: two bugs behind you!
[09:09] <pitti> and just uploaded one fix
[09:09] <seb128> pitti, ok, time for me to find stuff to fix :p
[09:09] <didrocks> rickspencer3: yeah, it's "working", but seems there is another regressions (but before coming to this state, 5 patches between unity and compiz had to been tested, failed, restested, failed…)
[09:09] <pitti> seb128: I should make you stop for a while and fix the libical FTBFS :)
[09:09] <seb128> pitti, it's Debian!
[09:09] <pitti> seb128: actually, I think it was you who requested the sync, so it falls to you anyway
[09:10] <seb128> indeed
[09:10] <seb128> grumpf, armel
[09:10] <pitti> fails in the test suite
[09:10] <pitti> and that's non-graphical, so porter box should work fine
[09:10] <seb128> pitti, I will have a look
[09:10] <seb128> right
[09:10] <pitti> merci
[09:10] <seb128> can you open a bug? :p
[09:10]  * seb128 hides
[09:11] <pitti> if you wait long enough, doko will do it for you :)
[09:11] <pitti> (open the bug, I mean, and complain bitterly)
[09:12] <didrocks> ah, on a clean session, it's better
[09:12]  * didrocks is really puzzled with all this unitcompizeries
[09:12] <seb128> didrocks, what was broken?
[09:12] <didrocks> seb128: alt + arrows on, for instance, weechat
[09:13] <seb128> what is alt-arrows supposed to do?
[09:13] <didrocks> switching between channels
[09:13] <didrocks> I personnaly use alt + number or alt + a, but pitti is using alt + arrows
[09:13] <seb128> ok, no a compiz keybinding then (i.e not in the wm actions)
[09:13] <didrocks> and chrisccoulson is doing some bad stuff in vim with it
[09:14] <seb128> the only place where I use alt-left is firefox and it works here
[09:14] <didrocks> yeah it's only happening in terminal apps
[09:14] <didrocks> xev is fine, giving back the value
[09:14] <pitti> that's broken for me as well in firefox
[09:14] <didrocks> pitti: really? it works here :/
[09:15] <seb128> wfm in firefox
[09:15]  * didrocks doesn't want to know, doesn't want to understand, it's 10am and I'm already more than tired with all the pings
[09:17]  * rickspencer3 restarts to get shiny new unity
[09:17] <jibel> didrocks, for firefox bug 953783
[09:17] <ubot2`> Launchpad bug 953783 in firefox "Alt+arrows keyboard shortcuts don't invoke Back and Forward navigation" [Undecided,New] https://launchpad.net/bugs/953783
[09:18] <didrocks> "It started to work again after some minutes. Perhaps it's related to a crash that has happened somewhere in the HUD department"
[09:18] <didrocks> wth?
[09:22] <rickspencer3> hey capitain didrocks
[09:22] <didrocks> hey hey rickspencer3
[09:22] <rickspencer3> it looks like my alt-tab bug is fixed!
[09:22] <didrocks> first good news of the day! :)
[09:23] <rickspencer3> didrocks, hmm, the update is working well for me so far
[09:23] <rickspencer3> didrocks, are there issues that I'm supposed to be having?
[09:23] <didrocks> phew
[09:23] <didrocks> rickspencer3: if you use alt + right/left
[09:23] <seb128> works fine for me as well
[09:23] <didrocks> but not 100% of the time
[09:24] <didrocks> and another issue which has been reported only on 2 machines
[09:24] <rickspencer3> didrocks, oh, my keybindings for workspace switching is still shift-windows-left/right
[09:24] <didrocks> rickspencer3: did you redefine them by hand at some point?
[09:24] <rickspencer3> didrocks, no
[09:25] <rickspencer3> but I presume they just didn't get reset at some point
[09:25] <didrocks> rickspencer3: migrating default really doesn't work well on compiz, you should unity --reset
[09:25] <didrocks> yeah
[09:25] <rickspencer3> didrocks, do you want me to try unity --reset?
[09:25] <rickspencer3> didrocks, I'll do it if you want, but, I don't much care, tbh
[09:26] <didrocks> rickspencer3: yes please, better you are on the standard :)
[09:26] <rickspencer3> ok
[09:26] <rickspencer3> doing it
[09:27] <rickspencer3> didrocks, the keybinds are now changed
[09:27] <rickspencer3> and seem to work fine
[09:27] <didrocks> great :)
[09:28] <rickspencer3> didrocks, the reveal of the autohide is now back to defaults, and seems to work WAAAY better
[09:28] <rickspencer3> didrocks, I dunno, this seems like it was a good update for me
[09:28] <didrocks> rickspencer3: yeah, I tweaked the upstream value but shhhhhh :)
[09:29] <rickspencer3> :)
[09:29] <rickspencer3> didrocks, did they update how alt-tab looks?
[09:29] <didrocks> rickspencer3: sweet! you shouldn't get false HUD reveal as well, I hope it will work well for everyone seeing the number of trials/errors it took
[09:29] <didrocks> rickspencer3: no, it's not ready yet
[09:29] <rickspencer3> ok
[09:29]  * didrocks would rather than they focus on the priority bug list though
[09:29] <didrocks> that*
[09:29] <rickspencer3> indeed
[09:29] <seb128> didrocks, I use alt a lot for "go to ws<n>", no hud false positive for a while here
[09:29] <rickspencer3> please let's not update the alt-tab imagary
[09:29] <didrocks> seb128: \o/
[09:30] <seb128> ;-)
[09:30] <rickspencer3> didrocks, seb128 the hud reveal seems to be working perfectly for me as well
[09:30] <pitti> indeed it helped a lot -- haven't seen the HUD since I rebooted yet
[09:31] <didrocks> pitti: haven't seen the HUD, like, you didn't want to see it, right? :-)
[09:31] <seb128> I triggerred it a few time by error but it's my fault
[09:31] <rickspencer3> didrocks, great job! you really rocked this update ... thanks for making it land so smoothly, though I know it wasn't a smooth road for you to get it here ;)
[09:31] <pitti> didrocks: right; it does work fine if I just press and release alt
[09:31] <seb128> I tend to put my finger on alt to switch workspaces and sometime sombody talks to me on IRC so I relax alt to reply :p
[09:31] <didrocks> rickspencer3: was quite bumpy indeed! But thanks :-)
[09:31] <didrocks> pitti: nice, one of the big blocker part of the past
[09:32] <didrocks> pitti: so, I tried to restart quite a few time, I think that alt + arrows is now fixed
[09:32]  * didrocks wants on RAOF as he wanted to rebreak ABI and so get an unity rebuild, let's slip that in
[09:32]  * pitti hugs didrocks
[09:32]  * didrocks hugs pitti back
[09:32] <didrocks> seb128: all that in less than 250ms? :)
[09:33] <seb128> didrocks, it's like my finger is over the key standing and sometime I relax a bit and hit the key by error :p
[09:34] <seb128> didrocks, like if you let you foot over one of the car pedals :p
[09:34] <didrocks> seb128: isn't that forbidden when you take your driver lessons? :p
[09:35] <seb128> it is!
[09:35] <seb128> nobody gave me keyboard navigation lessons though, maybe that's the issue ;-)
[09:35] <didrocks> you need to take again the computer driving exam!
[09:35] <didrocks> :)
[09:42] <seb128> urg, pitti went on a bugs unassignment rampage ;-)
[09:42] <pitti> seb128: I did some cleanup, yes; we have far too many assigned to the team and can't keep up
[09:43] <rickspencer3> hey didrocks the update went so well on my netbook, I'm doing my workstation now ;)
[09:43] <seb128> pitti, some of those are things that should be fixed for the lts, I'm not sure how to keep track of those out of team assignment though
[09:44] <pitti> seb128: well, feel free to assign them back; I dropped those which seemed unimportant, unrealistic, or only applied to maverick or natty
[09:44] <seb128> I guess milestoning with importance > high and hoping they don't get postponed because nobody looked at them
[09:44] <rickspencer3> oh rats, I have KDE stuff on this 'puter :/
[09:44] <didrocks> rickspencer3: excellent, but you are using one monitor isn't it? (IIRC, your netbook doesn't support dual monitor or with a really bad resolution)
[09:44] <rickspencer3> didrocks, right, my workstation has 2 though
[09:44] <seb128> pitti, I was thinking about bug #884003 for example
[09:44] <ubot2`> Launchpad bug 884003 in vino "The remote connection indicator is broken" [High,In progress] https://launchpad.net/bugs/884003
[09:44] <pitti> seb128: there's also the rls-p-tracking tag; but as I said, feel free to assign some bugs back
[09:45] <pitti> seb128: ah, fine for me
[09:45] <didrocks> rickspencer3: can you notice the compiz perf boost on it? it was quite clear on my slow laptop and netbook
[09:45] <pitti> but we actually need to assign it to someone, otherwise they just bitrot
[09:45] <rickspencer3> didrocks, multi-mon has been working really well for me since alpha1, I'll let you know how it goes
[09:45] <seb128> pitti, ok, doing that, I will try to give it to dx though :p
[09:45] <pitti> seb128: merci
[09:45] <seb128> de rien
[09:45] <rickspencer3> didrocks, I think I noticed a perf boost around a1 on my netbook
[09:45] <rickspencer3> I didn't notice one today, but I haven't used it that much yet
[09:46] <didrocks> ok, not particularly with that update? (compiz reduced by 51% it's CPU usage)
[09:46] <rickspencer3> tbh, it wasn't *that* slow
[09:46] <didrocks> its*
[09:46] <rickspencer3> didrocks, well, I'm expecting to notice it more in battery life, tbh
[09:46] <rickspencer3> which is actually more important to me
[09:47] <didrocks> right ;)  but I notice quite clearly smooth animation now on my laptop, which wasn't at all the case before (sometimes, get 4 frames for switching ws, so was seeing some "freezes" and such)
[09:50] <pitti> [ 3965.741882] hud-service[2180]: segfault at 0 ip 000000000040a640 sp 00007fffc648ca38 error 4 in hud-service[400000+11000]
[09:50] <pitti> hmm
[09:51] <pitti> I'll report it in a bit
[09:51] <seb128> pitti, was that the new version?
[09:52] <pitti> from this morning, yes
[09:52] <seb128> pitti, desrt's refactoring with gmenu support landed yesterday, I'm sure he will be happy to get bugs :p
[09:52] <seb128> iz desrt's bog!
[09:53] <rickspencer3> seb128, I guess it's too early in Cananda to troll desrt
[09:53] <seb128> rickspencer3, yeah, probably ;-)
[09:53] <seb128> rickspencer3, but I'm sure he will read feedback when he starts working :p
[09:54] <seb128> rickspencer3, in fact I was half joking, he said yesterday he was interested to have it landing to get extra feedback and fix the remaining issues
[09:59] <seb128> pitti, desrt: could be bug #953579 which seems a frequent one
[09:59] <ubot2`> Launchpad bug 953579 in indicator-appmenu "hud-service crashed with SIGSEGV in hud_string_list_get_tail()" [High,Confirmed] https://launchpad.net/bugs/953579
[10:00] <pitti> sent to LP now, will see
[10:19] <tkamppeter> pitti, what did you commit for bug 953321? Did you backport the fix from upstream?
[10:19] <ubot2`> Launchpad bug 953321 in cups "lpq and lpstat output empty lines only" [Medium,Fix committed] https://launchpad.net/bugs/953321
[10:19] <pitti> tkamppeter: yes, it works fine
[10:20] <pitti> tkamppeter: do you have anything else for cups? if not, I'd do an upload now to fix the RC bug in Debian
[10:22] <tkamppeter> pitti, do the upload, I am waiting for a patch for the broken IPP from Mike Sweet but did not get it yet. Should be no problem if we upload again tomorrow for the IPP stuff.
[10:30] <chrisccoulson> sigh
[10:30] <chrisccoulson> i can't think of a more unproductive use of my time than refreshing a monster sized CSS patch, again :(
[10:30] <pitti> chrisccoulson: what do you patch?
[10:31] <chrisccoulson> pitti - the monochrome icons in thunderbird
[10:31] <didrocks> RAOF: around for breaking an ABI?
[10:31] <pitti> chrisccoulson: and I guess it's not an option to drop them until they get accepted upstream?
[10:31] <chrisccoulson> pitti - http://bazaar.launchpad.net/~mozillateam/thunderbird/thunderbird-trunk.head/view/head:/debian/patches/theme-refresh-messenger-toolbar-icons.patch, and that's one of 3 ;)
[10:31] <chrisccoulson> yeah, hopefully they will land upstream soon
[10:32] <chrisccoulson> a lot of the theme changes already did land
[10:32] <pitti> that looks pretty ridiculous
[10:34] <chrisccoulson> pitti - yeah, it is quite ridiculous
[10:34] <chrisccoulson> and some people think that they're going to be doing this for firefox too ;)
[10:34] <pitti> chrisccoulson: is that request coming from design team?
[10:35] <pitti> it's the kind of change that we usually reject in the sponsoring process
[10:35] <pitti> and say "go upstream first"
[10:37] <seb128> didn't we work with andreas for the tb theme thing?
[10:38] <seb128> how come that didn't go upstream?
[10:39] <chrisccoulson> seb128, most of it has landed upstream now, it's just the icon changes are still in the review process
[10:39] <chrisccoulson> i should ping andreasn about that really :)
[10:39] <seb128> chrisccoulson, would our tb look different from the rest of the desktop without that patch?
[10:39] <seb128> I sort of like the monochrome icons :p
[10:40] <chrisccoulson> seb128, it just wouldn't have the monochrome icons
[10:40] <chrisccoulson> yeah, i like those too ;)
[10:40] <pitti> still, couldn't we just wait until they get changed upstream, instead of spending hours each time to update that patch?
[10:40] <seb128> or do it once at the end of the cycle
[10:40] <pitti> well, for tbird/ffox there is no real "end" of the cycle
[10:41] <seb128> pitti, well, why upstream would care about our branding? it seems like the sort of things it's hard to block on upstream for
[10:41] <pitti> seb128: chrisccoulson said it's almost accepted
[10:41] <pitti> seb128: and if it's not, it's all the more reason not to patch it
[10:41] <seb128> right, but it has probably been taking time because it's not a priority for them
[10:41] <pitti> I mean, 7000 lines of patch, to be re-done from scratch every month?
[10:41] <seb128> pitti, I disagree with the "let our OS looks like crap because we don't want to distro patch for branding and upstream doesn't care about our branding"
[10:42] <seb128> but yeah, for sure it's an issue
[10:42] <pitti> well, 'like crap' is certainly an overstatement
[10:43] <seb128> in fact having tb looking decently integrated to Ubuntu was one of the condition to take it by default in oneiric
[10:43] <pitti> but if that stays a distro patch, then we need a less intrusive form than http://bazaar.launchpad.net/~mozillateam/thunderbird/thunderbird-trunk.head/view/head:/debian/patches/theme-refresh-messenger-toolbar-icons.patch times three
[10:43] <seb128> but yeah, somewhat the price to pay to not use an application in our default toolkit
[10:43] <seb128> pitti, right, agreed on that
[10:44] <pitti> seb128: nice syncs
[10:45] <seb128> pitti, mbiel is launched on GNOME 3.4 it seems ;-)
[10:55] <Sweetshark> chrisccoulson: heya
[10:55] <Sweetshark> chrisccoulson: where do I find the npapi-sdk pkg-config package these days?
[10:57] <RAOF> didrocks: Around now.
[10:58] <chrisccoulson> Sweetshark, it's in firefox-dev
[11:09] <tkamppeter> pitti, hi
[11:11] <pitti> hey tkamppeter
[11:13] <tkamppeter> pitti, I have the following problem: We use Ghostscript for pdftops (PostScript printers) in Precise. Ghostscript's PostScript output prints perfectly on HP and displays perfectly with Ghostscript, but it crashes Kyocera and Brother printers. Oneiric used Poppler with inferior color management. What should we do?
[11:18] <pitti> hm, no off-hand idea
[11:18] <pitti> it seems we switched between gs and poppler several times already
[11:20] <tkamppeter> pitti, yes, afdter FF when there are relatively many people testing, the current method disqualifies and I switch to the other method.
[11:22] <tkamppeter> pitti, perhaps I have to put a by-manufacture or by-PPD-keyword switch into the filter to keep as many printers as possible well color-managed but also as many printers as possible at least working.
[11:22] <pitti> tkamppeter: I thought back then we switched from gs to poppler because gs gave too many problems
[11:23] <pitti> tkamppeter: and you can't work around that by telling gs to emit PS level 1 or 2 only?
[11:25] <tkamppeter> pitti, also a way to try. I will work out an appropriate recipe for the Kyocera and Brother users now to try.
[11:46]  * ritz wondering, why are our patches so ubuntu specific , wrt global menus ( liberal use of ubuntu_ in variable naming ) 
[11:48] <tkamppeter> pitti, new recipes for all the ideas from you and the GS devs are out in bug 951627 and bug 950713.
[11:48] <ubot2`> Launchpad bug 951627 in cups "all Kyocera network printers stopped working with cups 1.5.2" [Undecided,Incomplete] https://launchpad.net/bugs/951627
[11:48] <ubot2`> Launchpad bug 950713 in cups "Brother printer fails to print anything but a blank page" [Undecided,Incomplete] https://launchpad.net/bugs/950713
[12:14] <chrisccoulson> nice, just freed 28GB by emptying my trash
[12:14] <seb128> chrisccoulson, waouh, you use nautilus a lot! ;-)
[12:15] <mdeslaur> chrisccoulson: FYI, you don't have to download all the pr0n locally, you can just watch it inside your firefox window
[12:16] <chrisccoulson> lol ;)
[12:53] <didrocks> pitti: seems that unity-3d merged some multimonitor changed without the feature freeze excpection approved. There was no bug linked to it, but it seems that bug #950136 is the one
[12:53] <ubot2`> Launchpad bug 950136 in unity-2d "multimonitor, launcher: Provide an option to display either a single launcher or a launcher on each display in a multi-monitor environment" [High,In progress] https://launchpad.net/bugs/950136
[12:53] <didrocks> pitti: it seems that there is some g-c-c changes as well, it would have been nice if design involved platform a little bit more with it
[12:56] <pitti> re
[12:57] <pitti> tkamppeter: ah, thanks
[12:57] <pitti> didrocks: hm, I vaguely remember reading an UIFE about this before, but perhaps it was very similar
[12:57] <pitti> didrocks: or it was by mail
[12:57] <pitti> it seems ok to me
[13:00] <didrocks> pitti: ok, can you acked on the bug? The g-c-c changes are still not completly defined
[13:00] <didrocks> (and of course, -3d and -2d will again use different keys…)
[13:00] <pitti> erk, why?
[13:00] <didrocks> so a lot of fun
[13:00] <pitti> gsettings vs. gconf?
[13:00] <didrocks> right
[13:01] <didrocks> and despite asking them beforehand, there is no way they use gconf…
[13:01] <didrocks> so all the fun about concialiting the keys as I did for the other changes, listing them to not segfault if 2d is not installed…
[13:01] <didrocks> also not sure how we can set the primary monitor and detect the nvidia case
[13:02] <pitti> acked
[13:06] <seb128> pitti, did you have any chance to test the user dialog stuff on amd64?
[13:06] <pitti> I didn't yet, sorry
[13:06] <pitti> I'm swamped in mail backlog, it's a crazy time
[13:06] <seb128> pitti, yeah, I noticed you were busy on bugs ;-)
[13:07] <pitti> I'm down from 50 to 30 in my ubuntu folder
[13:07] <seb128> pitti, you just made me less busy, I had that gucharmap on my list for today ;-)
[13:07] <pitti> bug 952612? feel free to grab if you want
[13:07] <ubot2`> Launchpad bug 952612 in gucharmap "Package @GCONFPKGS@ was not found in the pkg-config search path." [Medium,In progress] https://launchpad.net/bugs/952612
[13:08] <seb128> pitti, no, that's fine, you grabbed it, it's yours :p
[13:08] <pitti> I wanted to take the opportunity to merge with Debain
[13:08] <pitti> and sync
[13:08] <seb128> I just read titles daily and "flag" stuff that seem like they need to be fixed
[13:08] <seb128> pitti, it's all yours ;-)
[13:08] <pitti> ok
[13:08] <seb128> I'm not running out of bugs don't worry
[13:08] <seb128> though I'm busy valgrinding around today
[13:10] <pitti> I'm close to a fix for bug 868400, just waiting on Bastian's second review
[13:10] <ubot2`> Launchpad bug 868400 in gnome-settings-daemon "Synaptics touchpad stops working - two syndaemon instances running" [High,In progress] https://launchpad.net/bugs/868400
[13:19] <cyphermox> pitti: desktop team PPA reasonable for the testing of the NM updates?
[13:19] <pitti> cyphermox: yes, it is; it's the staging area for stuff that's meant to go into precise
[13:19] <cyphermox> if we're still using it for that yea
[13:19] <cyphermox> alright :)
[13:20] <cyphermox> well, I'll upload shortly
[13:29] <desrt> seb128, sigh.  really stupid issue.
[13:29] <desrt> i don't understand why i didn't see it already.
[13:29] <seb128> desrt, hey
[13:30] <seb128> desrt, you got quite some segfault, I assigned at least 3 to you
[13:30] <seb128> desrt, do you get bug emails nowadays or should I mention them there? ;-)
[13:30] <seb128> desrt, hey btw, good morning :p
[13:30] <desrt> i get them
[13:30] <desrt> good morning :)
[13:30] <desrt> i already know what one of the issues is before looking at the code :p
[13:31] <seb128> desrt, btw archive testing is good, showed some issues in a night ;-)
[13:32] <desrt> okay.  first issue fixed.
[13:34] <desrt> https://bugs.launchpad.net/ubuntu/+source/indicator-appmenu/+bug/953620 is more interesting
[13:34] <ubot2`> Launchpad bug 953620 in indicator-appmenu "hud-service crashed with SIGSEGV in g_str_equal()" [Medium,New]
[13:35] <desrt> looks like some bad interaction with bamf
[13:36] <Trevinho> desrt: always about bamf there's this problem: https://launchpad.net/bugs/926208
[13:36] <ubot2`> Launchpad bug 926208 in bamf "bamfdaemon crashed with SIGABRT in g_assertion_message()" [Critical,Confirmed]
[13:36] <Trevinho> It seems like that when an object is removed from bus too quickly, then we get that problem
[13:36] <Trevinho> looking at the glib sources, the crashing function is basically an idle that is never removed when it should (i.e. when the vtable is removed), but that has an assert that causes the crash
[13:36] <desrt> sorry.  xchat suddenly crashed
[13:37] <Trevinho> desrt: did you read my lines?
[13:37] <desrt> i'll add a guard to ensure that this function is not giving NULL
[13:37] <desrt> that should fix it
[13:37] <desrt> Trevinho, just about the 'object removed too quickly' situation
[13:37] <Trevinho> ok
[13:38] <seb128> desrt, can you bzr push your first fix so I can look at it? ;-)
[13:38] <desrt> so it's possible that bamf_view_get_name() for a window could return NULL if the window is currently disappearing from the bus?
[13:38] <Trevinho> desrt: it shouldn't....
[13:38] <desrt> but it does?
[13:38] <Trevinho> at least, the name is generally saved on initialization
[13:39] <Trevinho> let me check
[13:39] <desrt> what if the window exists only very briefly?  could there be a race?
[13:40] <Trevinho> desrt: well, that's possible...
[13:40] <desrt> okay.  i'd better protect my code, then
[13:40] <Trevinho> as the name is saved only the first time that is used, so if in the mean time the bus is already down, the value is null
[13:41] <desrt> ya.  that's what i figured
[13:41]  * desrt is not a huge fan of how bamf works like that
[13:41] <Trevinho> desrt: there are some values that are saved on initialization, but not the name apparentyl
[13:41] <desrt> okay.  then i'm convinced that that explains this bug
[13:41] <desrt> seb128, pushing 2 fixes
[13:41] <seb128> desrt, \o/
[13:42] <seb128> desrt, btw Trevinho had what he thinks is a glib bug for you :p
[13:42] <desrt> he's probably right :)
[13:42] <seb128> Trevinho, you should repeat the line desrt didn't read when his xchat closed ;-)
[13:43] <Trevinho> 14:36:35 <Trevinho> It seems like that when an object is removed from bus too quickly, then we get that problem
[13:43] <Trevinho> 14:36:48 <Trevinho> looking at the glib sources, the crashing function is basically an idle that is never removed when it should (i.e. when the vtable is removed), but that has an assert that causes the crash
[13:43] <Trevinho> bug is https://launchpad.net/bugs/926208
[13:43] <ubot2`> Launchpad bug 926208 in bamf "bamfdaemon crashed with SIGABRT in g_assertion_message()" [Critical,Confirmed]
[13:44] <Trevinho> desrt: we could also figure to fix it on bamf, adding an idle there, but I wonder why you don't avoid your idle to run if the vtable has been shut down
[13:44] <Trevinho> anyway call_in_idle_cb seems to cause other issues too
[13:45] <Trevinho> I mean, in other packages there are crashes caused by that call_in_idle_cb
[13:45] <desrt> Trevinho, please file this in bugzilla.gnome.org
[13:46] <desrt> by a quick look at the code, i more or less agree
[13:46] <Trevinho> sure
[13:46] <desrt> the idle dispatch can be queued on the main context while the code that is unregistering the object is running
[13:46] <Trevinho> yes, that's what I figured...
[13:46] <desrt> it's actually a tricky bug
[13:46] <desrt> because normally the worker thread checks for the existence of the object (so that it knows which main context to dispatch to)
[13:47] <Trevinho> desrt: I'm using gdbus-codegen on bamf side, so I've not so much low-level control...
[13:47] <desrt> so it's the one that sends the "no such object!" error
[13:47] <desrt> in this case, though, it saw that there was an object and did the dispatch
[13:47] <desrt> but no object will be there to receive it
[13:47] <Trevinho> mh, races!
[13:47] <desrt> this is like the exact opposite of a bug that mikkel and i were talking about with davidz for a long long time
[13:48] <desrt> ie: what should be done about messages that arrive while we're *registering* objects in the main context :)
[13:48] <desrt> which is actually a far far more common case
[13:48] <desrt> ie: service activation
[13:49] <desrt> i wonder if this is the same bug  https://bugs.launchpad.net/ubuntu/+source/indicator-appmenu/+bug/953562   :)
[13:49] <ubot2`> Launchpad bug 953562 in indicator-appmenu "hud-service crashed with SIGSEGV in menuproxy_build_cb()" [High,Confirmed]
[13:50] <desrt> looks more like a dbusmenu bug, actually
[13:53] <seb128> desrt, you are welcome to fix bugs in there as well :p
[13:53] <desrt> i'll propose a fix, i guess
[13:54] <seb128> ;-)
[13:55] <desrt> oh.  interesting.
[13:55] <desrt> there is already some code here that attempts to deal with that
[13:55] <desrt> i wonder why it is not working
[13:56] <desrt> btw: what's with all this 0xaaaaaaaa business i see everywhere?  do you have some debug options in the memory allocator during the unstable cycle?
[13:56] <desrt> it's making my life a lot easier :)
[13:56] <pitti> desrt: I indeed see this in quite a few stack traces
[13:56] <seb128> desrt, no idea but I noticed as well
[13:57] <pitti> I suspect g_free() or similar does this
[13:57] <pitti> to make it more obvious when you try to access freed memory
[13:57]  * desrt likes it :D
[13:57] <pitti> too bad that R, G, and H aren't hex digits as well :)
[13:57] <desrt> RAGHH!
[13:58]  * desrt raises an eyebrow
[13:58] <pitti> 0xDEADFAFF
[13:58] <desrt> r would be good: 0x1FREED11
[13:59] <Sweetshark> 0xCAFEBABE is still a nifty one
[13:59] <pitti> hah, good one
[14:00] <Trevinho> desrt: here you are https://bugzilla.gnome.org/show_bug.cgi?id=671988
[14:00] <ubot2`> Gnome bug 671988 in gio "Quickly registering / unregistering objects on bus causes crash" [Major,Unconfirmed]
[14:02] <hallyn> since yesterday's update, unity-3d doesn't give me an indicator panel.
[14:03] <Trevinho> hallyn: no indicators at all? :o
[14:06] <desrt> seb128, the strip() bug is extremely annoying
[14:06] <seb128> desrt, why?
[14:06] <seb128> hallyn, can you pastebin a dpkg -l | grep unity
[14:06] <desrt> seb128, it's really a rather core feature of how the search algorithm works (and i disagree with the algorithm that is being used)
[14:07] <desrt> seb128, we can deal with some of the fallout by stripping yes... but there's a bigger issue here
[14:07] <desrt> i need to talk to ted about that
[14:07] <seb128> desrt, ok, I'm not sure to understand why any algo would drop things that match
[14:07] <desrt> seb128, it doesn't drop them.  they get lost in the noise
[14:08] <seb128> desrt, like "zoom " should obviously match results which have "zoom" in them?
[14:08] <hallyn> Trevinho: well actually also no panel.  can't even start a term with ctrl-alt-t
[14:08] <hallyn> one sec
[14:08] <seb128> desrt, well a full word matching should be a high score?
[14:08] <desrt> the algorithm works like so: it splits the incoming string into tokens and matches each token against the available items in the menu, marking distance for each token
[14:08] <hallyn> seb128: http://paste.ubuntu.com/881858/
[14:08] <desrt> teach token is counted equivalently
[14:08] <desrt> so when you search "file " you have tokens "file" and ""
[14:08] <desrt> "file" matches things... good
[14:08] <desrt> but "" doesn't match anything
[14:08] <desrt> which introduces a lot of noise
[14:09] <seb128> well a full matching should be scored enough to keep you out of noise?
[14:09] <desrt> ie: the distance from "" to anything is high -- which means no result has close distance
[14:09] <desrt> so even the good results get lost in the noise
[14:09] <desrt> i don't think so
[14:09] <seb128> hum
[14:09] <seb128> ok
[14:09] <desrt> for example, "" will match far better with a shorter string
[14:09] <seb128> well "" shouldn't be tried to be matched at all
[14:09] <desrt> i agree
[14:09] <desrt> but here's where it gets annoying:
[14:09] <desrt> "file x"
[14:10] <desrt> same issue there -- only one character less severe
[14:10] <desrt> that will match things with short strings starting with 'x' quite strongly
[14:10] <desrt> possibly to the detriment of the 'file' part
[14:10] <seb128> desrt, I see what you mean
[14:10] <seb128> desrt, to me it seems like a full word matching should score much higher
[14:11] <desrt> i think the search algorithm we use is a bit bogus
[14:11] <desrt> i'd be happier with a simple subset matching
[14:11] <desrt> ie: ignore all this fancy fuzzy-string business
[14:11] <seb128> desrt, anyway something to sort with ted
[14:11] <seb128> desrt, imho what we have now, not scoring "" and scoring a full word match very high should be fine
[14:12] <seb128> for most cases at least
[14:12] <desrt> the algorithm is not based on 'high scores'
[14:12] <desrt> it's based on 'low distances'
[14:12] <desrt> and a full word match can only have a distance of 0
[14:12] <desrt> not -10 :)
[14:13] <seb128> desrt, well somehow you managed to "trick" the system indicators score
[14:13] <desrt> (although maybe we should tweak that...)
[14:13] <desrt> seb128, right.  that's a penalty that gets added
[14:13] <desrt> whatever their distance is, it adds 50%
[14:13] <desrt> so 10 becomes 15
[14:13] <seb128> desrt, well anyway I don't think we will sort the algo here today between us
[14:14] <seb128> desrt, so let's move on to the next bug ;-)
[14:14] <desrt> i agree
[14:14] <desrt> so the menuproxy_build_cb() affects 17 people, eh?
[14:14] <desrt> crazy....
[14:14] <desrt> is this only showing up after the update yesterday?
[14:14] <desrt> or what that the first time you pushed the new code to the archive?
[14:15] <seb128> desrt, first bug was 15 hours ago
[14:15] <seb128> desrt, so we get a dup by hour basically
[14:15] <desrt> seb128, did you ever push my hud code to precise before then?
[14:15] <seb128> desrt, no
[14:15] <desrt> okay.  good
[14:15] <seb128> desrt, it was in the ppa before
[14:15] <desrt> because i didn't think anything i did in the past few days could cause this
[14:16] <seb128> desrt, ted strongly hinted he would be unhappy to land a version without the highlight and usage db so I decided to wait for those
[14:16] <seb128> desrt, he said it was my call but still, I figured waiting for those was a good midway compromise
[14:17] <desrt> ya.  i don't disagree
[14:17] <desrt> regressions are ugly
[14:18] <desrt> launchpad needs some sort of intelligence for trying to find out what all the duplicated reports have in common :)
[14:18] <desrt> seb128, so the other issue is the DoS bug
[14:18] <desrt> i've known about that for some time, actually
[14:18] <desrt> there will be two fixes there: 1) algorithmic improvements
[14:19] <desrt> but even with those, i think it would be reasonable to 2) limit the search string length to some preset maximum?
[14:20] <seb128> desrt, well, mhr3 said that typing "i want to rotate this image by ninety degrees" with gimp selected was enough to bring your hud version down
[14:20] <seb128> desrt, he also wrote "it is a bug, the methods normally finish in a couple hundred milliseconds, but if you do this, each search takes *seconds* longer and longer"
[14:21] <mhr3> right
[14:21] <mhr3> desrt, https://bugs.launchpad.net/indicator-appmenu/+bug/953910/+attachment/2862038/+files/Screenshot%20from%202012-03-13%2011%3A48%3A14.png
[14:21] <ubot2`> Launchpad bug 953910 in indicator-appmenu "HUD: Very high CPU usage" [Undecided,New]
[14:21] <seb128> desrt, are you done with the segfaults?
[14:22] <mhr3> (although that wasn't the latest version)
[14:22] <seb128> desrt, https://bugs.launchpad.net/ubuntu/+source/indicator-appmenu/+bug/953771 seems quite duppy as well
[14:22] <ubot2`> Launchpad bug 953771 in indicator-appmenu "hud-service crashed with SIGSEGV in g_hash_table_lookup_node()" [Medium,New]
[14:24] <desrt> suspicious!!
[14:24] <desrt> not so suspicious!!
[14:28]  * Sweetshark grumbles about nvidia optimus still being broken on precise. That was the one thing that worked well with natty.
[14:30] <Sweetshark> (well, just plainly running on nvidia worked on natty.)
[14:31] <dupondje> Sweetshark: you should use bumblebee :) then it 'works' :)
[14:31] <dupondje> at least you can shutdown 1 gfx card
[14:34] <Sweetshark> dupondje: http://geek.co.il/wp/2012/02/19/nvidia-optimus-on-ubuntu-12-04 <- this doesnt sound as if that stuff is too stable currently either.
[14:35] <desrt> seb128, okay.  pushed another fix.
[14:35] <desrt> any other crashers you see? :)
[14:36] <dupondje> Sweetshark: its indeed not a finished solution. I currently use it just to disable 1 gfx card, cause else I have to high powerdrain on my laptop.
[14:36] <seb128> desrt, bug #953620 and bug #953517 are the same bug?
[14:36] <ubot2`> Launchpad bug 953620 in indicator-appmenu "hud-service crashed with SIGSEGV in g_str_equal()" [Medium,Confirmed] https://launchpad.net/bugs/953620
[14:36] <ubot2`> Launchpad bug 953517 in indicator-appmenu "hud-service crashed with SIGSEGV in g_str_equal()" [High,Confirmed] https://launchpad.net/bugs/953517
[14:36] <desrt> seb128, ya.  those are both the bamf issue (now worked-around)
[14:37] <desrt> Trevinho, you should probably add some code to bamf to be more careful in that case...
[14:37] <seb128> desrt, ok, so remains the dbusmenu issue and the DoS issue
[14:37] <desrt> ie: return "" or something, instead of NULL
[14:37] <desrt> seb128, yup
[14:37] <desrt> seb128, any other than those, though?
[14:37] <seb128> desrt, not on my list
[14:37] <desrt> nice
[14:37] <desrt> not bad for a complete rewrite :p
[14:37] <seb128> ;-)
[14:38]  * desrt was vaguely dreading waking up this morning
[14:38] <seb128> lol
[14:38] <seb128> desrt, you still have enough for the day I think
[14:39] <seb128> desrt, especially if you count Trevinho's glib issue ;-)
[14:39] <desrt> pish.  i'm only on my first coffee so far :)
[14:39] <Trevinho> eheh
[14:39] <desrt> get another release out and let's find out about all the new issues i've caused :D
[14:39] <desrt> actually -- i guess i should really try to get to the bottom of this dbusmenu one
[14:39] <Trevinho> desrt: what you mean about adding some code to bamf?
[14:39] <desrt> it looks pretty serious
[14:39] <seb128> desrt, that would be good, it collects dups
[14:39] <desrt> Trevinho, you have a known non-bug case where that function can return NULL
[14:40] <Trevinho> I've not looked at the issues, btw
[14:40] <desrt> Trevinho, you should either document it as such ("this function can return NULL in certain races, so you better check that") or fix it
[14:40] <Trevinho> desrt: and that's not specified?
[14:40] <Trevinho> mh, ok
[14:40] <desrt> well, funny you should say that
[14:40] <desrt> because the developer docs are still completely empty :p
[14:40] <Trevinho> One the things in my TODO is to add a nice documentation to bamf
[14:40] <Trevinho> yes, I know
[14:41] <Trevinho> I didn't touch these things yet... I would love to get more introspection as well
[14:43] <desrt> seb128, btw... dunno if you know anything about this
[14:43] <desrt> but i installed gdm the other day to test something
[14:43] <desrt> set it as my default login manager
[14:43] <desrt> then i removed it
[14:43] <desrt> and now lightdm is completely messed up
[14:43] <desrt> i've tried dpkg-reconfigure, uninstalling, reinstalling,... nothing
[14:43] <desrt> the symptoms are that (a) it doesn't start on bootup anymore and (b) once i do start it up (manually) it fails the first login
[14:44] <desrt> ie: i type my password and it accepts it and the screen flickers and i'm back at the lightdm greeter again -- but the second time it works
[14:45] <Trevinho> oh... that looks bad
[14:45] <desrt> it's clearly my fault, at least a little
[14:45] <desrt> but the system should probably not be so fragile that installing/removing a package can do that
[14:45] <Trevinho> well, yes... but it could happen that an user try to change that
[14:45] <Trevinho> of course...
[14:46] <seb128> desrt, sudo edit /etc/X11/default-display-manager
[14:46] <seb128> desrt, change lightdm to /usr/bin/lightdm there
[14:46] <seb128> desrt, it's https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/864174
[14:46] <ubot2`> Launchpad bug 864174 in lightdm "boot hangs waiting for lightdm after purging gdm (wrong default-display-manager)" [High,Won't fix]
[14:46] <desrt> known issue, i guess :)
[14:46] <Trevinho> :)
[14:46] <desrt> thanks
[14:46] <desrt> wontfix, eh?
[14:47] <desrt> btw: why such old gdm?
[14:49] <mterry> kenvandine, two unity-greeter reviews for ya, if you can get a chance.  Should get an email from LP for both
[14:49] <seb128> desrt, because nobody cared enough to update, jbicha and ricotz tried and hit issues with the custom dconf gsettings hackery
[14:50] <desrt> ah...
[14:50] <desrt> i've talked to ray about that.  it's evil stuff.
[14:50] <seb128> desrt, the one which got refactored 3 times since to work
[14:50] <kenvandine> mterry, just saw them :)
[14:50] <desrt> the entire concept of blacklisting parts of the session from starting seems deeply wrong
[14:50] <desrt> he has a plan to change that, but no help for now
[14:51] <desrt> seb128, thanks for the explanation
[14:52] <seb128> desrt, I think it's just that nobody wanted to invest enough energy to get that working
[14:52] <seb128> desrt, I mean that's why the package update got stalled
[14:52] <seb128> i.e downstream manpower,motivation issue :-(
[14:52] <seb128> yw
[14:52] <desrt> ya.  totally understandable
[14:52] <desrt> if it was a simple update, that's one thing
[14:52] <desrt> but it doesn't make sense to put so much effort for something that people won't install anyway
[14:58] <jbicha> pitti: what do you think of the clutter/gnome-shell FFe's?
[14:59] <pitti> jbicha: I haven't looked at them this week yet, need to catch up
[14:59] <jbicha> ok, no problem
[15:01] <jbicha> seb128: good morning, I merged this with Debian https://code.launchpad.net/~jbicha/+junk/gnome-system-log-3-3/
[15:01] <jbicha> I believe we can just sync baobab as we don't have an LPI patch there
[15:01] <seb128> jbicha, do we have any interest to sync those this cycle?
[15:01] <seb128> it seems like most of those source didn't get any change during the cycle
[15:02] <seb128> well I'm fine either way, I just ignored them since they are unmaintained upstream
[15:02] <jbicha> seb128: both baobab & gnome-system-log (at least in its next tarball) got new Mallard help, but yeah, they're nearly unchanged from 3.2
[15:03] <seb128> jbicha, works for me
[15:03] <seb128> pitti, ^ what do you think about syncing those? it shouldn't be a big change, mbiebl just packaged the new standalone versions which have improved documentation
[15:04] <pitti> sounds fine to me
[15:05] <seb128> jbicha, ok, I will sync them and update gnome-utils in a bit
[15:05] <seb128> well, sync baobad, review your system log
[15:05] <seb128> bab
[15:10] <dupondje> aha do we get a gnome-shell update ? :)
[15:13] <jbicha> dupondje: maybe
[15:14] <dupondje> so my battery indicator is finally correct ^^
[15:15] <seb128> desrt, ok, those 3 fixes backported and uploaded to precise
[15:15] <seb128> on that I'm out for some exercice, be back in ~1 hour for the meeting if we have one today ;-)
[15:16] <jbicha> dupondje: yes, I'd rather not try to backport several fixes if we end up not needing to
[15:17] <dupondje> just blow in a new version ;)
[15:20] <dupondje> jbicha: you got my vote ;) not that it matters but ok :D
[15:23] <asac> unity disappeared :/... window manager still running
[15:23] <asac> dont see a unity binary anymore
[15:29] <Sarvatt> asac: ya let a dist-upgrade remove it?
[15:32] <asac> Sarvatt: yeah
[15:32] <asac> seems so
[15:32] <asac> restartin X now
[15:32] <asac> cannot work in this state
[15:36] <asac> back on unity2d
[15:36] <asac> until proper unity comes back
[15:37] <ogra_> how did you get rid of it ?
[15:37]  * ogra_ has unity running fine on his newly set up x86 box 
[15:37] <Sarvatt> there was a libxfixes transition that needed a unity rebuild, was a period of time dist-upgrade would remove unity
[15:38] <ogra_> well, thats why apt or u-m tell you about it and ask a Y/N question ;)
[15:38] <Sarvatt> yep :)
[15:38] <asac> they told me not strong enough
[15:39] <ogra_> admit that you just didnt clean your glasses before upgrading !
[15:39] <asac> didnt clean my screen, yes
[15:39] <ogra_> heh
[15:41] <Saviq> didrocks, automerge-unity stuck?
[15:42] <Saviq> or... wait... maybe it's the time diff
[15:42] <Saviq> last build took 4 mins and now it's 13 mins and counting
[15:43] <Saviq> we'll wait :/
[15:51] <didrocks> Saviq: yeah, it's depending if they are other merges in progress
[15:52] <didrocks> Saviq: I hope you did notice that unity-common is know installed for unity-2d :)
[16:23] <cyphermox> fyi, I uploaded network-manager and network-manager-applet to ppa:ubuntu-desktop/ppa; help with testing this and making sure wired, wifi, mobile broadband et al. still work properly would be much appreciated :)
[16:24] <cyphermox> so if you run into explosions and cries of agony, etc; ping me :)
[16:30] <didrocks> cyphermox: well, if it explodes, we would have no network, so no crie? :)
[16:30] <ogra_> you could put your mobile # into the changelog so people can send SMS
[16:30] <glatzor> evening mvo
[16:33] <mvo> hey glatzor!
[16:33] <mvo> glatzor: thanks for your info on #950791, I'm happy to fix that now that I know what approach to take
[16:36] <cyphermox> didrocks: you know my dark dark secret :)
[16:37] <didrocks> cyphermox: that's so easy! :-)
[16:37] <Sweetshark> dupondje: bumblebee doesnt work for me. but I can switch that stuff in bios anywhy. with natty it worked when I switched to pure discrete gfx (no optimus) with the nvidia driver. but somewhere in oneiric/precise that was broken
[16:37] <Sweetshark> dupondje: now it just boots into a blank screen.
[16:38] <cyphermox> didrocks: realistically though, not likely to explode. I've already been running it for two days :)
[16:38] <didrocks> cyphermox: well, computer science is full of surprises
[16:39] <cyphermox> didrocks: I limit surprises by careful testing and blind faith :)
[16:42] <cyphermox> yay, the daily builds appear to be unbroken now
[17:09] <glatzor> mvo, I already have fixed it
[17:09] <glatzor> mvo, i was on a train ride
[17:31] <stgraber> mterry: ping
[17:34] <mterry> stgraber, yo
[17:34] <stgraber> mterry: bug 954255
[17:34] <ubot2`> Launchpad bug 954255 in unity-greeter "unity-greeter crashed with SIGSEGV in g_closure_invoke()" [Undecided,New] https://launchpad.net/bugs/954255
[17:34] <stgraber> mterry: it's technically a duplicate of bug 943037
[17:34] <ubot2`> Launchpad bug 943037 in casper "Log out/in not working in live session" [High,Triaged] https://launchpad.net/bugs/943037
[17:34] <stgraber> though I reported a new one to have the apport data in it
[17:35] <stgraber> mterry: it's been affecting all our live media for a while apparently, whenever you logout, lightdm gets into an infinite loop trying to spawn unity-greeter and unity-greeter crashing
[17:35] <mterry> stgraber, ah.  OK.  will look at it
[17:36] <stgraber> mterry: it could be that unity-greeter simply doesn't work on a live media as we usually don't see it (as we use autologin)
[17:47] <desrt> does anyone know what the status of TRIM support in ubuntu is?
[17:53] <seb128> jbicha, still there?
[17:55] <desrt> so this is a pretty big fail.  out of the box, ubuntu doesn't have TRIM enabled
[17:55] <desrt> which means that if you run ubuntu on an SSD your computer will get slower and slower and slower the more you use it
[17:56] <pitti> what is necessary to enable this?
[17:56] <desrt> the 'discard' mount option to ext4
[17:56] <desrt> which can also be set in the default mount options using tune2fs (or presumably at the time the filesystem is created)
[17:57] <pitti> bug 867794 has some discussion
[17:57] <ubot2`> Launchpad bug 867794 in ubiquity "Ubuntu 11.10 does not enter 'discard' filesystem option on ext4 install on SSD drive" [Wishlist,Confirmed] https://launchpad.net/bugs/867794
[18:00] <pitti> I think I looked at that a few months ago, and it didn't seem to be unanimous
[18:00] <desrt> ya.  i just asked in the kernel channel
[18:01] <desrt> fstrim sounds awesome.  i was wondering if something like this exists or if i would need to force the trim of my filesystem by filling it with a large file and deleting it :)
[18:02] <desrt> /home: 60428242944 bytes were trimmed
[18:02] <desrt> christ...
[18:02] <desrt> 60 gigs...
[18:02] <pitti> I did notice that IO sucks on my SSD these days
[18:02] <desrt> maybe my SSD starts being fast again :)
[18:02] <pitti> but I had "discard" enabled for quite some time, and it didn't seem to help really
[18:02] <desrt> well, 'discard' concerns me, honestly
[18:03] <desrt> it doesn't make a whole lot of sense because it's non-atomic
[18:03] <desrt> i'm a filesystem... someone erases a file
[18:03] <pitti> fstrim: /: FITRIM ioctl failed: Die Operation wird nicht unterstützt
[18:03] <pitti> hm, too bad
[18:03] <desrt> i have to do (simplifying for the sake of argument) two things:
[18:03] <desrt> 1) remove the link from the directory
[18:03] <desrt> 2) trim the contents of the space taken by the file
[18:03] <desrt> clearly i should do #1 before #2
[18:03] <desrt> but then that means if i get shutdown and i did #1 and not #2 then there is untrimmed space...
[18:04] <desrt> pitti, root?
[18:04] <seb128> desrt, btw new stacktrace on this dbusmenu menuproxy_build_cb() bug, https://launchpadlibrarian.net/96654548/Stacktrace.txt
[18:04] <pitti> desrt: I did
[18:04] <pitti> desrt: it says "operation not supported"
[18:05] <desrt> i figured that :)
[18:05] <desrt> pitti, ext4?
[18:05] <pitti> yep
[18:05] <pitti> /dev/sda2 on / type ext4 (rw,errors=remount-ro,discard)
[18:05] <desrt> maybe your drive doesn't have TRIM support
[18:05] <seb128> desrt, it's different from the old ones (I asked a new stacktrace with the current versions to the retracers)
[18:05] <desrt> is it intel or an off brand?
[18:05] <pitti> desrt: oh - it seems I still have the "discard" mount option, perhaps that interferes
[18:05] <desrt> pitti, i have it too -- *shrug*
[18:06] <pitti> desrt: SAMSUNG_MMCRE28G8MXP-0VBL1
[18:06] <pitti> it's in a ThinkPad X201, so shouldn't be the worst one
[18:06] <desrt> i'm afraid it is
[18:06] <pitti> desrt: but yes, I guess it's missing the support for it
[18:06] <desrt> when i was shopping for my x201 i called lenovo to ask about the drive that you get when you pick the 'SSD' option
[18:06] <desrt> and did some looking at benchmarks
[18:06] <desrt> it's quite bad...
[18:06] <desrt> i ended up buying the smallest possible traditional drive with the laptop and throwing it out and putting in an SSD i bought myself
[18:07] <desrt> i think lenovo cleaned up their acts a bit since then, but the x201 was quite bad....
[18:07] <pitti> well, 250 MB/s isn't exactly bad
[18:07] <seb128> desrt, I did the same for my dell
[18:08] <seb128> not to mention that the ssd option at dell was more expensive than buying and intel ssd myself
[18:08] <desrt> ya... buying bigger HD/memory with laptops is a waste of money
[18:08] <desrt> always better to buy&install it yourself
[18:10] <desrt> SSDs are particularly weird... they're such a new and complex technology that depending on the brand and controller there can easily be 2-3x performance difference between them for particular workloads
[18:10] <desrt> which is not really something you see anywhere else anymore
[18:11] <jbicha> seb128: half here
[18:11] <didrocks> ok, time for some exercice, see you tomorrow guys
[18:11] <desrt> didrocks, enjoy :)
[18:11] <didrocks> desrt: (not sure enjoy is the term, but thanks nevertheless) ;)
[18:12] <seb128> desrt, btw did you see my note about the new stacktrace on https://launchpadlibrarian.net/96654548/Stacktrace.txt ?
[18:12] <desrt> yes.  looking at it now.
[18:12] <desrt> it's similar, but different
[18:13] <desrt> i have an idea for how i might be able to reproduce this locally.
[18:13] <desrt> give me a moment
[18:13] <desrt> in theory writing an app that spawns many windows and closes them quickly may trigger this
[18:13] <seb128> desrt, I was just pointing it because the retracers seem to think they are the same bugs but the stacktraces look a bit different
[18:14] <seb128> desrt, I wanted to make sure you saw there was a new stacktrace matching the new version
[18:14] <seb128> desrt, no hurry ;-)
[18:14] <desrt> seb128, so the old hud kept dbusmenu clients around forever
[18:14] <desrt> the new code unrefs them when the window is destroyed
[18:14] <desrt> at the same time the window is being destroyed there may be outstanding things going on with the client
[18:14] <desrt> and i think that's what's causing this new class of bugs
[18:15] <desrt> so i'm going to write a program that just creates and destroys a whole lot of windows in an attempt to lose this race
[18:15] <seb128> jbicha, so we can't push to lp:~ubuntu-desktop/component/ubuntu if "component" is not registered to launchpad, do you want to register gnome-system-log or should we use lp:ubuntu/gnome-system-log? It's a small source I'm fine using the lp:ubuntu location
[18:15] <seb128> desrt, ok
[18:16] <desrt> i'm quite sure this is a dbusmenu bug caused by the way that i'm using it
[18:16] <desrt> but not 100% sure :)
[18:20] <jbicha> seb128: oh, I can register gnome-system-log tonight, as far as branches go, I prefer a packaging-only branch but whatever you like
[18:20] <seb128> jbicha, I prefer those as well, I'm uploading without pushing, I will push tomorrow once the product exists
[18:23] <desrt> sigh
[18:23] <desrt> my torture-test program crashed unity
[18:23] <desrt> and failed to crash the hud...
[18:23] <seb128> :-(
[18:24] <seb128> desrt, unity or gnome-session,X?
[18:24] <desrt> could have been X
[18:24] <seb128> desrt, your IRC closed, unity going down should not end the session
[18:24] <seb128> usually compiz respawns
[18:24] <desrt> i had to logout
[18:24] <desrt> sorry.. not crashed
[18:24] <desrt> froze
[18:24]  * desrt went to the console and did kill -9 -1
[18:24] <seb128> desrt, go to a vt next time and type "unity"
[18:25] <seb128> that will do a compiz --replace with DISPLAY set for you basically
[18:25] <desrt> f i slow it down it's better :)
[18:25] <desrt> 1 new window per 1ms is too fast :p
[18:26] <desrt> so... many many bamf bugs here
[18:26] <desrt> but no dbusmenu crashes yet
[18:33] <desrt> seb128, is the gtk menu hiding broken for you?
[18:34] <desrt> in gtk3 apps...
[18:35] <seb128> desrt, no
[18:35] <seb128> desrt, you use the ppa g-s-d with the gnome-shell xsettings?
[18:35] <desrt> ah.  could be.
[18:36] <desrt> let me purge that
[18:36] <desrt>  *** 3.3.91-0ubuntu3+appmenu1 0
[18:36] <desrt>         500 http://ppa.launchpad.net/gnome3-team/gnome3/ubuntu/ precise/main amd64 Packages
[18:36] <desrt> yup :)
[18:37] <desrt> another thing: do you notice that 'killall' is not working properly anymore?
[18:37] <desrt> like "killall gnome-settings-daemon" does not work
[18:39] <desrt> sigh.  now compiz is spinning causing xorg to use 100% CPU.  better than being completely pegged i guess
[18:39]  * desrt needs to stop using xchat
[18:46] <kenvandine> mterry, is this known?
[18:46] <kenvandine> [+0.00s] WARNING: unity-greeter.vala:129: Failed to load state from /home/ken/.cache/unity-greeter/state: Key file contains line '@�' which is not a key-value pair, group, or comment
[18:46] <kenvandine> then later on i see lots of these
[18:46] <kenvandine> [+0.23s] DEBUG: unity-greeter.vala:487: Failed to write state: Error writing to file: Bad address
[18:46] <mterry> kenvandine, in --test-mode, yeah, it's common
[18:46] <kenvandine> mterry, just wondering if i should delete that file
[18:47] <kenvandine> or if it would be useful for a bug report
[18:47] <kenvandine> ok
[18:47] <mterry> kenvandine, eh, don't worry about it I don't think
[18:47] <pitti> good night everyone
[18:47] <kenvandine> good night pitti
[18:47] <kenvandine> mterry, ok, thx
[18:47] <kenvandine> mterry, almost done with the reviews
[18:47] <mterry> pitti, night!
[18:48] <seb128> desrt, yeah, I got annoyed by it too but I didn't report it, a foundation team bog
[18:48] <kenvandine> mterry, only comment from me so far is you removed some really nice descriptive comments
[18:48] <seb128> 'night pitti
[18:48] <kenvandine> the ones that described how the animations get calculated
[18:49] <mterry> kenvandine, are they still needed?
[18:49] <mterry> kenvandine, I thought I removed the code that went with them
[18:49] <kenvandine> i am sure they changed
[18:49] <kenvandine> perhaps :)
[18:49] <Darxus> seb128: So what would it take to get you to --enable-wayland-backend in gtk for precise?
[18:50] <seb128> Darxus, I think there is no way I want to turn on experimental backends a month after ff in a lts
[18:50] <seb128> Darxus, its easy enough to have a ppa with that option on for precise
[18:50] <kenvandine> mterry, but the new code doesn't include nice explanations
[18:50] <desrt> heh.  my torture test is causing all kinds of fun unity bugs
[18:50] <mterry> kenvandine, fair.  I can add some before commit.  I got so excited about deleting code  :)
[18:50] <kenvandine> maybe it just isn't needed anymore
[18:51] <kenvandine> indeed
[18:51] <seb128> desrt, what sort of bugs?
[18:51] <desrt> there seems to be some bad race involved in maintaining the list of open applications in the dash
[18:51] <kenvandine> it is great to delete
[18:51] <Darxus> seb128: I suppose that's reasonable, thanks.
[18:51] <kenvandine> mterry, but the old code set the code comment bar pretty high :)
[18:51] <kenvandine> i actually enjoyed reading the removed code :-p
[18:52] <seb128> Darxus, yw
[18:52] <desrt> http://imgur.com/yGu91
[18:52] <mterry> kenvandine, hah  :)
[18:52] <desrt> ^ a program that opens and closes windows quickly causes that to happen, even after it has completely exited
[18:53] <seb128> urg
[18:53] <desrt> and now compiz is spinning again
[18:53] <desrt> sigh
[18:54] <desrt> 'killall compiz' (and allow it to restart) solves both issues
[18:54]  * desrt switches back to gnome-shell :p
[18:56] <desrt> huh.  seems that unity-panel-service has some kinda of race in the appmenu registrar that it doesn't notice that the window has been closed if you close it too quickly :X
[18:57] <seb128> desrt, the count for this bug is at 26 users affected now btw
[18:57] <seb128> it's solid on the 1 by hour ;-)
[18:57] <seb128> slightly over that now
[18:57] <desrt> seb128: i'm having trouble coming up with a testcase that reproduces it without also triggering bugs in other places :p
[18:58] <seb128> I noticed ;-)
[18:59] <desrt> the appmenu registrar actually has some odd behaviour...
[18:59] <desrt> it seems to be on some sort of timeout
[19:02]  * desrt gets fed up and goes back to the other approach
[19:04] <desrt> seb128: crash is happening on x86_64, x86 only, what?
[19:05]  * desrt notes at least x86_64 from this one backtrace
[19:05] <seb128> desrt, i386 and amd64
[19:05] <seb128> probably others but most of ours users are on those
[19:06] <desrt> wait a minute.......
[19:06]  * desrt gets a hunch
[19:13]  * mvo hugs glatzor
[19:19] <mterry> kenvandine, thanks for the reviews!
[19:19] <kenvandine> mterry, anytime
[19:19] <desrt> seb128: found the bug
[19:19] <desrt> seb128: and it's not one of those easy sort of things to fix....
[19:20] <seb128> desrt, oh? what's the issue?
[19:20] <desrt> seb128: essentially, dbusmenu is assuming that if it calls g_cancellable_cancel() on a cancellable connected to an async request then the async request will ALWAYS return an "operation was cancelled" error
[19:20] <desrt> GIO makes no such guarantee
[19:20] <desrt> if the operation was already completed and ready for dispatch at the time you call cancel() you may get the actual result instead of the cancellation error
[19:21] <seb128> oh
[19:23] <desrt> so short story: definitely dbusmenu bug (for making incorrect assumptions about underlying library behaviour)
[19:24] <desrt> longer story: maybe the incorrect assumption was reasonable to expect, actually, and we should consider fixing the library
[19:24] <desrt> asking davidz's opinion
[19:36] <desrt> seb128: we decide we're going to try to fix the problem in GIO
[19:36] <desrt> i'm going to whip up a patch in a hurry
[19:37] <seb128> desrt, yeah, I just read the #gtk+ backlog when you highlighted me there ;-)
[19:50] <micahg> should we provide transitional packages from telepathy-butterfly to telepathy-haze?
[19:51] <seb128> no
[19:51] <seb128> haze use pidgin to do any protocol
[19:52] <seb128> it's not an equivalent of butterfly
[19:52] <micahg> right, but butterfly was removed
[19:52] <seb128> so?
[19:52] <seb128> what issue do you try to solve? ;-)
[19:52] <micahg> aptitude broken upgrades :), dist-upgrade seems to work fine
[19:53] <seb128> what is aptitude unhappy about?
[19:53] <seb128> telepathy-butterfly is not something any "user" will have installed
[19:53] <seb128> it will have been pulled out as an implementation detail by some im
[19:54] <micahg> the farsight -> farstream change WRT dependencies, dist-upgrade removes telepathy-butterfly (probably because it's not in the archive anymore) and that breaks the bad dependency chain
[19:54] <seb128> so imho it should be solved by fixed the recommends,depends of those ims
[19:55] <micahg> hmm, the issue seems to have been telepathy-core used to depend on it, so I'm not sure why it's not removed on upgrade, I guess I have to keep digging, thanks
[19:56] <seb128> yw
[19:58] <desrt> seb128: you gonna be around much longer?
[19:58]  * desrt will have some glib patches to test soon
[19:59] <seb128> desrt, "test"
[19:59] <seb128> ?
[19:59] <seb128> desrt, I'm not able to trigger the bug on willing so I'm not sure what to test
[19:59] <desrt> that's a good point
[19:59] <desrt> test by uploading to the archive, then? :)
[19:59] <seb128> desrt, I'm about to go for dinner, I can run a build and do testing later or tomorrow morning
[19:59] <seb128> desrt, don't hurry, I will not do a glib upload before going to bed, I prefer to stay around for those, so tomorrow morning
[20:00] <desrt> okay
[20:00] <seb128> desrt, it's a bit of apport noise but not the end of the world, can wait tomorrow
[20:00] <desrt> you seemed to have a sense of urgency because of the level of incoming bugs
[20:00] <seb128> desrt, though I would prefer if you get a patch today our day so I can pick it up in the morning
[20:00] <desrt> seb128: i have the GSimpleAsyncResult part done already
[20:01] <desrt> going to open a bug about that and push it
[20:01] <desrt> then i'll fix GDBus to use the new API
[20:01] <seb128> desrt, well, "urgency", I didn't want to slip by a week or to be in beta2
[20:01] <seb128> desrt, but tomorrow is fine
[20:01] <desrt> seb128: oh... pfft
[20:01] <seb128> desrt, thanks, efforts are appreciated ;-)
[20:01] <desrt> this was a tough bug, but it wouldn't have taken an entire week to find :p
[20:01] <desrt> have some more faith :p
[20:02] <seb128> desrt, I didn't doubt of that, I just wanted to make sure it was not landing at the bottom of a list ;-)
[20:02] <desrt> "your new code is crashing" lands at the top of my list :)
[20:02] <seb128> good!
[20:02] <seb128> that said I'm off for dinner
[20:02] <seb128> bbiab
[21:04] <dobey> chrisccoulson: hrmm. spellchecker.dictionary was set to en_AU for me, and specified as "user set" though i'm pretty sure i have never set it as such. because why would i want to spell things like that
[21:09] <bryceh> dobey, interesting; same thing happened to me a few times a couple weeks ago
[21:10] <dobey> also, i wonder how to havae it support multiple languages
[21:11] <dobey> have even.
[21:11] <bryceh> Feb 22 15:20:09 <chrisccoulson>	bryceh, yes, i guess it's fallout from this change: https://developer.mozilla.org/en/HTML/Controlling_spell_checking_in_HTML_forms#Controlling_the_spellchecker_language
[21:12] <bryceh> dobey, ^^
[21:13] <dobey> ah, fail
[21:20] <dobey> i can't see how doing that would ever be "the right thing to do"
[21:20] <Sarvatt> is there any bug regarding the apparmor boot speed regression discussed in here a few weeks ago? http://reports.qa.ubuntu.com/reports/boot-speed/acer-veriton-02/index.html looks nasty
[21:22] <micahg> Sarvatt: bug 949891
[21:22] <ubot2`> Launchpad bug 949891 in apparmor "apparmor caching is not working which has severely regressed boot time" [High,Fix released] https://launchpad.net/bugs/949891
[21:22] <Sarvatt> micahg: thanks a ton, i couldn't find it
[21:23] <Sarvatt> darn, fix released, it did reduce the time a good chunk but its still hugely regressed
[21:32] <Sarvatt> https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/954469
[21:32] <ubot2`> Launchpad bug 954469 in apparmor "apparmor changes caused boot speed regression" [Undecided,New]
[21:34] <mdeslaur> Sarvatt: hrm, that's not normal..it should have fixed that
[22:30] <desrt> seb128: sup
[22:31] <seb128> desrt, sort of around, not working but laptop on the table next to TV ;-)
[22:31] <seb128> desrt, i.e I will reply to you but not build or test anything else tonight :p
[22:32] <desrt> seb128: just wanted you to know that the two patches on https://bugzilla.gnome.org/show_bug.cgi?id=672013 should be enough to fix the last of the crashes you were seeing
[22:32] <ubot2`> Gnome bug 672013 in gio "GSimpleAsyncResult: support reliable cancellation" [Normal,New]
[22:33] <seb128> desrt, no dbusmenu fix needed then? it just make the assumption it's doing true?
[22:33] <desrt> yup
[22:33] <seb128> excellent
[22:33] <desrt> alex thought that it's a generally useful guarantee to make
[22:33] <desrt> and we managed to convince david too
[22:33] <desrt> so i did the patch
[22:33] <seb128> desrt, do you feel confident in the patches enough to get them distro patched without review?
[22:34] <desrt> i ran the glib testcase on them with no issues
[22:34] <desrt> let me jhbuild and login to them
[22:34] <seb128> desrt, I think I will do a test build tomorrow morning in any case
[22:34] <seb128> desrt, I was just wondering if I should wait for a review before uploading
[22:35] <desrt> seb128: tell you what
[22:35] <seb128> desrt, do you plan to try to get davidz or walters to review it today?
[22:35] <desrt> i'm going to build and login to them tonight
[22:35] <seb128> or alex tomorrow?
[22:35] <desrt> if there's anything bad happening i'll say so on the bug
[22:35] <seb128> desrt, I'm fine waiting for you to show up tomorrow
[22:35] <desrt> otherwise, i'd assume they're fine
[22:35] <seb128> desrt, I will build them and get them reading in the morning and wait for you to upload
[22:35] <seb128> ok?
[22:35] <seb128> ready
[22:35] <seb128> so I can test them etc for a few hours
[22:35] <desrt> my only worry is that there is some weird application out there that unknowingly relies on the race being won in the other direction
[22:35] <desrt> which seems... unlikely
[22:36] <desrt> sure
[22:36] <desrt> what is beta2?
[22:36] <desrt> *when
[22:36] <seb128> desrt, the freeze is in thursday after next
[22:37] <seb128> i.e 9 days
[22:37] <desrt> so no need to rush
[22:38] <seb128> right
[22:39] <seb128> desrt, I will not say that hud bugs are not the end of the world again or you will try to quote me on it! ;-)
[22:39] <desrt> s/quote/misquote/ :)
[22:39] <desrt> and it's not a hud bug... it's a dbusmenu bug :)
[22:42] <seb128> desrt, same difference, who cares about dbusme^W forget that ;-)
[22:45] <desrt> hah.  i wish :)
[22:45] <desrt> btw: did you get to the bottom of that gee thing?
[22:45] <desrt> i dont' see anything more about it in xsession errors
[22:45] <desrt> i think it just needed a new release to be built?
[22:51] <seb128> desrt, funny that you ask about that, mhr3 pinged me about it today since it was still happening but gee didn't build with the current vala yet so I triggered a rebuild today, I will tell you tomorrow
[22:51] <desrt> color-plugin-WARNING **: failed to reset xrandr-AU Optronics gamma tables: gamma size is zero
[22:51] <desrt> are you seeing this one?
[22:52] <desrt> or this one: gnome-session[12518]: WARNING: Could not launch application 'gdu-notification-daemon.desktop': Unable to start application: Failed to execute child process "/usr/lib/gnome-disk-utility/gdu-notification-daemon" (No such file or directory)
[22:52] <desrt> that's just about all that's left in my xsession-errors
[22:52] <desrt> other than some stupid debug cruft from gnome-shell:
[22:52] <desrt> ** Message: applet now removed from the notification area
[22:52] <desrt> ** Message: using fallback from indicator to GtkStatusIcon
[22:52] <desrt> ** Message: applet now embedded in the notification area
[22:53] <seb128> desrt, no way
[22:53] <seb128> desrt, not seing the colord one no
[22:54] <desrt> i wonder if it's related to some hardware i have
[22:54] <desrt> they make TFTs so i guess it's the panel in my laptop or my external monitors
[22:55] <RAOF> “xrandr” suggests so, yes.
[22:55] <desrt> ya.  that might have been a hint as well :p
[22:56] <seb128> hey RAOF, how are you?
[22:56] <RAOF> seb128: Good evening.  Pretty good.
[22:57] <jdstrand> Sarvatt: fyi, we tracked done the apparmor boot speed regression
[22:57] <jdstrand> s/done/down/
[22:57] <Sarvatt> jdstrand: AWESOME!
[22:57] <jdstrand> it'll be fixed in the next upload which should be this week
[22:58] <jdstrand> hopefully sooner than later in the week
[23:04] <desrt> RAOF: did you fix all my X bugs yet? :)
[23:04] <RAOF> desrt: Yes.
[23:04] <desrt> hum!  none of them are fixed!!
[23:05] <desrt> i still scroll backwards and my barriers in gnome-shell are not working
[23:05] <RAOF> Your barriers should be working.
[23:05] <desrt> as of when?
[23:05] <RAOF> When did you last restart your xserver?
[23:05] <desrt> earlier today thanks to compiz :)
[23:06] <RAOF> Hm.  Should be fixed in 2:1.11.4-0ubuntu6
[23:06] <desrt> and it looks like no X updates since then
[23:06] <RAOF> As of last Friday.
[23:06] <desrt> ya... i'd definitely have installed that
[23:07] <desrt> odd.  my versions are called like 1:7.6 or so
[23:07] <desrt> ah
[23:07] <desrt>   Installed: 2:1.11.4-0ubuntu6
[23:07] <RAOF> xserver-xorg-core is the package you're after.
[23:07] <RAOF> ...
[23:07] <RAOF> All my tests pass!
[23:07] <RAOF> So gnome-shell is working!  WORKING, I SAY!
[23:08] <desrt> gonna dist-upgrade once more and logout/in just to make absolutely sure
[23:09] <desrt> ya.. not working
[23:09] <desrt> the barrier is there... as long as i approach it at an ultra-low speed
[23:10] <desrt> ie: almost useless
[23:12] <htorque> hi all! i need some help determining the right package for filing a bug report:
[23:12] <htorque> when i set a background from "Colors & Gradients", the 'picture-uri' key for the background is set to null,
[23:12] <htorque> which isn't allowed (i get some critical warnings in ~/.xsession-errors). don't know what package to pick for that.
[23:13] <seb128> htorque, isn't it allowed or is something unhappy about it having that value?
[23:14] <seb128> htorque, gnome-control-center is what set the value if you use it to change the background
[23:14] <htorque> seb128: ~/.xsession-errors says "assertion `value != NULL' failed" a couple of times
[23:14] <seb128> htorque, then whatever is generating the warning seems buggy
[23:14] <seb128> could be nautilus
[23:14] <seb128> or gnome-settings-daemon
[23:14] <desrt> people use assert way too much
[23:14] <seb128> or unity
[23:14] <htorque> ha!
[23:14] <desrt> we need to start complaining to those people to stop it
[23:15] <bryceh> desrt, yeah it crashes my program every time it executes, it sucks.
[23:17] <htorque> seb128: i'll pick g-c-c as i also see "GLib-GIO-CRITICAL **: g_settings_set_value: ..."
[23:17] <htorque> seb128: thanks! :-)
[23:17]  * desrt raises an eyebrow
[23:17] <desrt> htorque: anything immediately before that?
[23:17] <desrt> like perhaps something about invalid utf8?
[23:18] <desrt> calling g_settings_set_value() directly is slightly rare
[23:18] <htorque> it just complains about the empty string: http://paste.ubuntu.com/882500/
[23:18] <desrt> that's not the empty string :)
[23:19] <htorque> oops, right, null.
[23:19] <desrt> ya.. so that's clear as day :)
[23:19] <desrt> someone calls g_settings_set_string (something, "picture-uri", NULL);
[23:19] <bryceh> htorque, btw I've a ppa that purports to fix gpu lockups while playing certain games.  https://launchpad.net/~bryce/+archive/lp899159
[23:19] <seb128> we should fix stuff to set their domain as well
[23:19] <seb128> so we know what displays warnings
[23:19] <bryceh> htorque, probably will take some hours to build mesa tho
[23:19] <desrt> we should fix the default log handler to put the process name and/or pid
[23:20] <seb128> desrt, there is,was a discussion about that on ubuntu-devel list
[23:20] <htorque> bryceh: is this the same as from the xorg-edgers ppa?
[23:20] <rye> hello, I am currently reinstalling precise on acer aspire one and noticed that ubiquity slideshow is flickering quite heavy when the images are switched
[23:20] <rye> I suppose the installer is ubiquity :)
[23:21] <desrt> seb128: why not upstream in bugzilla?
[23:21] <seb128> desrt, systemd-log? ;-)
[23:21] <seb128> desrt, or you might glib log handler?
[23:21] <desrt> glib log handler
[23:21] <htorque> seb128: sorry to interrupt - last question: file on launchpad or upstream?
[23:21] <desrt> so you dont' have to guess what's in .xsession-errors
[23:21] <seb128> desrt, the ubuntu-devel discussion was about fixing syslog to do that for you
[23:22] <seb128> but yeah, that's a different issue from the .xsession-errors one
[23:22] <seb128> htorque, launchpad please, bonus point if you get a stacktrace
[23:22] <rye> the question was is this known (checked in launchpad - does not seem to be filed)?
[23:22] <desrt> seb128: i'll raise the issue upstream
[23:23] <rye> this reminds me of the flicking the shotwell was doing due to doublebuffer being disabled
[23:23] <rye> flickering
[23:23] <desrt> xsession-errors tracking is a whole lot less fun when you don't know the pid
[23:23] <seb128> desrt, thanks
[23:23] <seb128> desrt, indeed
[23:23] <seb128> desrt, like those pygobject warnings from yesterday
[23:23] <desrt> ya...
[23:23] <seb128> desrt, it took us a while to have a testcase which was "run python and import gtk"
[23:24] <seb128> rye, the installer is ubiquity
[23:24] <seb128> rye, people on #gtk+ mentioned today that webkit disable doublebuffer
[23:25] <seb128> webkit is used for the slideshow I think
[23:25] <desrt> seb128: 'import gio' was sufficient, as it turns out
[23:25] <seb128> desrt, right, well "use a python app" was the clue ;-)
[23:25] <desrt> ya
[23:25] <desrt> only took me a few minutes to track it down after i had that hint
[23:25] <desrt> i had no idea otherwise
[23:25] <seb128> desrt, I can open a bug tomorrow on glib if you want
[23:25] <htorque> seb128 desrt: when setting it from g-c-c, i _do_ see something about invalid utf-8: http://paste.ubuntu.com/882511/ (the paste from above was just session start)
[23:25] <desrt> seb128: i already did: https://bugzilla.gnome.org/show_bug.cgi?id=672026
[23:25] <ubot2`> Gnome bug 672026 in general "default log output should include pid and/or prgname" [Normal,New]
[23:25] <desrt> gonna do a patch now
[23:25] <seb128> desrt, thanks
[23:26] <bryceh> htorque, yes the patch should be included in xorg-edgers as well
[23:26] <desrt> at the very least you can apply it locally to help you with your hunting
[23:26] <seb128> right
[23:26] <seb128> htorque, blame desrt
[23:26] <seb128> it's clearly a a g_variant bug
[23:26] <desrt> g_variant is the world's leading cause of invalid utf8
[23:26] <seb128> desrt, ;-)
[23:27] <desrt> funny thing is that gvariant was totally not designed with unicode in mind
[23:27] <desrt> the first version accepted arbitrary strings until we realised it was incompatible with dbus
[23:27] <desrt> we made an unstable change there...
[23:27] <desrt> nobody noticed :)
[23:27] <desrt> retroactively, gvariant became a very very good unicode citizen
[23:27] <seb128> htorque, ok, it's an Ubuntu bug
[23:27] <seb128> desrt, ;-)
[23:28] <desrt> s/unstable change/incompatible change/
[23:28] <seb128> it's a mterry bug, but he's not around so I will not troll him further about it
[23:28] <seb128> I didn't check the code, it's too late and I'm about to go
[23:28] <htorque> bryceh: in that case i probably already tested it (i opened bug 950300 as i couldn't get the other one hang to show again, so i guess *that* one is fixed).
[23:28] <ubot2`> Launchpad bug 950300 in mesa "[snb-gt2] GPU lockup render.IPEHR: 0x7a000002" [Undecided,New] https://launchpad.net/bugs/950300
[23:28] <desrt> seb is a good guy
[23:28] <desrt> only trolls to people's faces
[23:28] <desrt> never behind their backs :)
[23:28] <seb128> but I'm 95% sure it's the "set the background in accountsservice so lightdm can use it" patch
[23:29] <seb128> desrt, ;-)
[23:29] <desrt> seb128: i hate that patch....
[23:29] <seb128> desrt, you hate all Ubuntu patches
[23:29] <desrt> seb128: since it totally and completely fails to work if your home directory is secure
[23:29] <seb128> desrt, that's why you keep rewriting them in an upstream friendly way ;-)
[23:29] <desrt> seb128: this has nothing to do with upstream... it just doesn't work :p
[23:29] <seb128> desrt, it works as designed
[23:30] <desrt> it should copy the background image to somewhere in /var
[23:30] <htorque> maybe there are two bugs? i'm not sure but i think i've seen this before lightdm got that user background feature.
[23:30] <seb128> desrt, and even working like that some people have a go about exposing their private infos on a public screen
[23:30] <desrt> your background image is private? :)
[23:30] <seb128> desrt, oh yes
[23:31]  * desrt is suspicious
[23:31] <seb128> desrt, or some people think so at least
[23:31] <desrt> there are about two usecases i can think of:
[23:31] <jbicha> desrt: you don't have a private background image? ;)
[23:31] <seb128> desrt, some of them had a "I have a sexy picture of gf as my background but I don't anyone to look at it"
[23:31] <desrt> computer in my house in a room that i never let people go into: maybe i have a naked picture as the background image
[23:31] <desrt> laptop that i take outside with me: ... uh.. probably not so much
[23:31] <desrt> unless i install ubuntu-calendar, of course...
[23:32] <desrt> jbicha: i only have one computer
[23:32] <seb128> good times ;-)
[23:32] <desrt> i dock my laptop...
[23:32] <desrt> since it leaves the house with me and i use it on trains and aeroplanes and so on... no :)
[23:33] <desrt> so here's the funny thing...
[23:33] <desrt> my background image is the blue stripes
[23:34] <desrt> which is at the system level somehow... i don't know if it's on the CD or with gnome-shell installed or what...
[23:34] <desrt> i don't see it in lightdm
[23:34] <seb128> desrt, are you sure you don't have the slideshow?
[23:34] <desrt> i do have the slideshow
[23:34] <seb128> ok, lightdm doesn't support slideshows
[23:34] <desrt> i would expect it to select one of the images
[23:35] <seb128> right, "feature request"
[23:35] <seb128> or "known issue"
[23:35] <desrt> hum
[23:35] <desrt> was i supposed to have real work tonight? :)
[23:35] <desrt> oh right.  the log thing
[23:35] <desrt> this is way more important
[23:35] <seb128> hehe
[23:35]  * desrt figures out how slideshows work, anyway
[23:36] <desrt> Background=/usr/share/themes/Adwaita/backgrounds/adwaita-timed.xml
[23:36] <desrt> account service is working as expected.  good.
[23:36] <desrt> just have to teach lightdm what that means, i guess :)
[23:36] <RAOF> Bonus points for making it switch more frequently than on the desktop; every 30 sec, maybe.
[23:37] <desrt> interesting idea
[23:37] <rye> bug #918617
[23:37] <ubot2`> Launchpad bug 918617 in unity-greeter "Greeter does not support dynamic wallpapers" [Low,Triaged] https://launchpad.net/bugs/918617
[23:37] <bryceh> htorque, thanks
[23:37] <desrt> rye: thanks
[23:38] <seb128> desrt, if you want to read arguing over the private stuff there are comment on https://bugs.launchpad.net/unity-greeter/+bug/844081
[23:38] <desrt> RAOF: so i think the most-correct thing to do is to display the correct image for the current time
[23:38] <ubot2`> Launchpad bug 844081 in accountsservice "Unity Greeter - Background of the Unity Greeter should reflect the background chosen by the user that is currently selected" [Wishlist,Confirmed]
[23:38] <rye> By the way - current precise, freeshly installed - two syndaemons running
[23:38] <seb128> rye, did g-s-d segfault or got restarted?
[23:38] <desrt> cnd: hey.  rye was just talking about you :)
[23:38] <rye> seb128: no, first one is from lightdm, second one is from my user
[23:38] <seb128> desrt, no, he was talking about bastien :p
[23:39] <cnd> hi rye, who are you? :)
[23:39] <desrt> seb128: all bugs are ubuntu bugs :)
[23:39] <desrt> (one way or the other)
[23:39] <seb128> rye, ok, "known issue" but less known that the gsd restart one, it's a bit weird since unity-greeter doesn't load the mouse gsd plugin so it shouldn't run syncdaemon
[23:40] <seb128> desrt, well that one is an upstream issue, pitti is fixing it, he had some ping pong with bastien and others about how to fix correctly
[23:40] <rye> seb128: this iso is from mar 13 with full reinstall
[23:40] <desrt> cnd: disregard, i guess :p
[23:40] <seb128> rye, it doesn't help much saying why you have those running when most people don't ;-)
[23:41] <seb128> desrt, it's basically "g-s-d spans syncdaemon to block pad while typing", but since g-s-d is keen to segfault and restart it happens often that people get several syncdaemon running and fighting
[23:41] <desrt> this bug is really easy to fix...
[23:42] <seb128> desrt, since g-s-d segfaulting and respawning of course neither stop the first spawned nor check if one is running before starting a new one
[23:42] <desrt> unity-greeter should just be using libgnome-desktop to do its job for it
[23:42] <seb128> desrt, great, you might be able to fix that and the glib log stuff then ;-)
[23:42] <desrt> okay
[23:42] <seb128> that said I'm off to bed
[23:42] <desrt> not *really* easy
[23:42] <seb128> 'night everybody
[23:43] <desrt> but maybe only one night work :)
[23:43] <rye> seb128: there was no g-s-d- failures and syndaemon is running by lightdm, but, since the init is now parent I can't find out who was it originally
[23:43] <desrt> g'night
[23:43] <seb128> rye, well it should be run by lightdm
[23:43] <seb128> rye, lightdm has nothing to do with it, g-s-d has but the unity-greeter disable that plugin
[23:43] <seb128> ups
[23:43] <seb128> "shouldn't
[23:43] <seb128> but anyway that's for another day
[23:44] <seb128> if you figure why your config start it that would be useful info ;-)
[23:44] <rye> seb128: sure, will track this if it starts
[23:44] <seb128> thanks
[23:49] <rye> actually i am working on Ubuntu One stuff, but i want to make sure that Precise works well on all my machines, therefore I am so noisy here
[23:51] <jbicha> hmm, I should have tried to get someone to build cogl on ARM first
[23:55] <htorque> desrt: bug 954590 (sorry, took a while)
[23:55] <ubot2`> Launchpad bug 954590 in gnome-settings-daemon "GLib-CRITICAL warning when setting a solid color as background" [Undecided,New] https://launchpad.net/bugs/954590
[23:58] <desrt> hum!
[23:58] <desrt> so this is a bit harder after all
[23:58] <desrt> the current code is threaded and GnomeBG is not threadsafe