2012/03/09/#ubuntu-desktop.txt

[00:11] <desrt> does anyone know how i request a retrace?
[00:11] <RAOF> It happens automatically.
[00:11] <desrt> hum
[00:11] <TheMuso> If the bug is tagged
[00:11] <desrt> oh.  this bug only came 42 minutes ago
[00:11] <desrt> that's probably why
[00:11] <TheMuso> need-$arch-retrace
[00:11] <RAOF> (Or, strictly speaking, the tags needs-amd64-retrace or needs-i386-retrace)
[00:12] <TheMuso> gah needs is correct
[00:12] <desrt> ya.  it is so tagged
[00:12] <desrt> i didn't realise the bug was just filed moments ago
[00:14]  * desrt curses the french
[00:15] <desrt> their damn cheese is so freaking good
[00:15] <desrt> brie in canada is twice as expensive and half as good
[00:58] <broder> i have a sort of embedded system which i'm trying to upgrade to oneiric, and i'm having problems with lightdm. when it runs as part of the standard boot, lightdm-session returns 1. if i look at .xsession-errors, all of the scripts are complaining that they can't open display :0
[00:58] <broder> anybody have thoughts for where i could look for what i screwed up?
[00:59] <broder> i've done plenty of terrible things to the setup on this environment, so whatever it is is probably my fault, but i'm not sure where to look
[01:02] <broder> also, it starts fine if i try again once my getty comes up
[01:08] <RAOF> That sounds like *so much fun*
[01:10] <broder> yeah, and that combined with the fact i'm using casper means i have to do a full rebuild to do any experimentation :-/
[01:11] <broder> hmm, how well tested is lightdm without accountservice?
[01:12] <RAOF> No idea, although obviously robert_ancell would be able to tell you that :)
[01:12] <RAOF> What environment do the unittests run in?
[01:13] <broder> well, there are a bunch of tests that don't have the accountservice
[01:14] <broder> (you can tell because they have names like test-language-no-accounts-service :-P)
[01:18] <robert_ancell> broder, the oneiric version is probably not as well tested as the precise version, can you pastebin the lightdm.log?
[01:21] <broder> robert_ancell: http://paste.ubuntu.com/875441/
[01:21] <broder> i don't have a greeter installed, but i'm always expecting autologin to work
[01:22] <broder> i set xserver-command=X -logverbose 0 in an attempt to get something out of the X log, which didn't seem to work
[01:22] <broder> http://paste.ubuntu.com/875442/ is the lightdm.conf
[01:24] <robert_ancell> broder, as a sanity check I take it /usr/bin/lightdm-session and /usr/bin/gnome-session exist and are executable?
[01:24] <robert_ancell> (often a cause of returning 1 on exec)
[01:25] <robert_ancell> It seems quite happy to work without Accounst Service - [+1.72s] DEBUG: org.freedesktop.Accounts does not exist, falling back to passwd file
[01:25] <broder> robert_ancell: yes. lightdm-session is definitely running and leaving this in .xsession-errors: http://paste.ubuntu.com/875444/
[01:25] <robert_ancell> ah, ** (gnome-session:1447): WARNING **: Cannot open display:
[01:25] <broder> err, i should probably note that even though it says gnome-session --session=ubuntu, i'm replacing the ubuntu session with my own session spec
[01:25] <robert_ancell> looks like a xauthority issue
[01:26] <broder> hmm...well, i do have an .Xauthority file, but it's 0-length
[01:26] <broder> but i guess that could be the result of lightdm's cleanup
[01:26] <robert_ancell> yes, it removes the entry on exit
[01:26] <robert_ancell> makes testing a pain :)
[01:27] <broder> hmm, let me see if i can add something to /etc/X11/Xsession.d to cat the xauthority file
[01:28] <robert_ancell> also check that $XAUTHORITY is set correctly
[01:28] <RAOF> Is xauthority owned by root or something?  That broke me for a while.
[01:28] <broder> no, owned by my user
[01:29] <broder> there's stuff in the .Xauthority file, and $XAUTHORITY is set
[01:29] <broder> but i think my hostname in this particular case is localhost because it's not set yet - is that a problem?
[01:29] <broder> ("localhost" showed up in the .Xauthority file)
[01:29] <robert_ancell> broder, btw you can run 'xauth list' which gives a readable form of the file
[01:29] <broder> ah, ok. i'll try that
[01:31] <broder> huh, xauth list printed nothing
[01:32] <robert_ancell> what was $XAUTHORITY set to?
[01:32] <broder> /home/mokafive/.Xauthority
[01:33] <robert_ancell> very odd.  Permissions on that file?
[01:34] <broder> 600
[01:34] <broder> oh, uh. huh. i actually don't seem to have xauth installed
[01:34] <robert_ancell> broder, can you try [LightDM] user-authority-in-system-dir=true
[01:35] <broder> do i need to have a /usr/bin/xauth if i'm not doing debugging stuff?
[01:35] <kenvandine> RAOF, i figured out the scrolling problem, which is turning out to be really nasty to work around
[01:36] <kenvandine> the gtk changes for GDK_SMOOTH_SCROLL_MASK
[01:36] <kenvandine> widgets no longer get scroll-event for GDK_SCROLL_MASK by default, so you have to explicitly add them
[01:36] <robert_ancell> broder, I don't think so
[01:36] <kenvandine> and gwibber uses a Gtk.Box with Gtk.EventBoxes on top
[01:37] <kenvandine> i have a patch for gtk to add GDK_SCROLL_MASK for GdkEventBox
[01:37] <kenvandine> but that only partly fixes it... in the eventboxes, i add more boxes and labels
[01:37] <kenvandine> which don't get the scroll-event either
[01:37] <RAOF> Sounds like fun.
[01:37] <broder> robert_ancell: no difference
[01:38] <kenvandine> so with fixing GtkEventBox, you get really hit or miss scrolling
[01:38] <kenvandine> if your cursor is on top of a label, it doesn't scroll
[01:39] <kenvandine> i really wonder why it was necessary to remove GDK_SCROLL_MASK by default
[01:39] <kenvandine> i suspect this is going to have lots of random side effects all over the place
[01:41] <broder> robert_ancell, RAOF: there's also nothing in the Xorg log, which seems suspicious: http://paste.ubuntu.com/875455
[01:41] <RAOF> Yeah, that is suspicious.
[01:42] <robert_ancell> broder, interesting though that the daemon was happy with the connection ([+3.02s] DEBUG: Connecting to XServer :0) using xcb
[01:42] <robert_ancell> broder, is $DISPLAY set correctly?
[01:42] <broder> robert_ancell: the .xsession-errors says it's trying to connect to :0
[01:43] <robert_ancell> RAOF, what's a good example program to connect to the X server and run through strace or similar to check what's actually failing on the connection?
[01:43] <broder> ah, crud. i have to run afk for the next several hours. i can try to do more testing later tonight
[01:44] <robert_ancell> ok
[01:44] <robert_ancell> 'No protocol specified' is actually an Xlib error if I remember correctly, but xlib is really bad at reporting what actually failed
[01:47] <robert_ancell> how weird, my session just exited with return value 134.  broder's problems are contagious
=== Guest16781 is now known as albrigha
[03:57] <desrt> jbicha: hi
[03:58] <jbicha> desrt: good evening!
[03:58] <desrt> jbicha: i have a 100% reliable gnome-shell crasher out of the gnome3 ppa.
[03:58] <desrt> want to check it?
[03:58] <jbicha> 100%? ok
[03:58] <desrt> open lg and click the chooser
[03:59] <desrt> then move the mouse over an open window
[03:59] <desrt> *boom*
[03:59]  * cnd cowers in fear
[04:00] <jbicha> desrt: it was more a silent boom :(
[04:00] <desrt> jbicha: i don't get it on my f17 box
[04:01] <desrt> but it's a version behind, at 3.3.5 now
[04:01] <desrt> gonna upgrade to see if it's an upstream or ubuntu issue
[04:01] <jbicha> yeah, those fedora guys take forever to package stuff, what kind of bleeding edge distro do they think they are? lol
[04:02] <desrt> it's already packaged
[04:02] <desrt> i just haven't upgraded in a while
[04:02] <desrt> it's my tablet.  i don't touch it much :p
[04:02] <jbicha> oh, I guess I need to find something else to be snarky at
[04:04] <jbicha> that crash is kinda neat though :)
[04:05] <desrt> yes.  i love when my window/compositing manager crashes.  very neat!
[04:05] <desrt> what's so interesting about it? :p
[04:05] <jbicha> the first time I did something different and it just hung with a small rectangle in the center of my screen, not sure how that happened
[04:07] <jbicha> well I don't use the lg, what's it there for? to make my desktop more like a web browser?
[04:07]  * jbicha looks around for the View Source button
[04:07] <desrt> the lg is pretty insanely cool
[04:07] <desrt> and the chooser makes it even more useful
[04:11] <jbicha> lg is far above my level
[04:16] <desrt> hum
[04:16] <desrt> looks like seb never got around to that experimental gnome-settings-daemon upload
[04:17] <jbicha> which one?
[04:35] <desrt> jbicha: the one that lets you replace the value of arbitrary xsettings using a gsettings key
[04:37] <desrt> jbicha: so f17 with the same upstream gnome-shell version as the gnome3 ppa => no crash
[04:41] <jbicha> desrt: we don't really patch gnome-shell, gjs, mutter so I don't know
[04:41] <stgraber> pitti: would be great if you could have a look at bug 950087
[04:41] <ubot2`> Launchpad bug 950087 in libxklavier "python2.7 crashed with SIGSEGV in g_object_newv() when using gi binding for xklavier (gir1.2-xkl-1.0)" [High,New] https://launchpad.net/bugs/950087
=== m_conley is now known as m_conley_away
[05:45] <BigWhale> Good Morning.
[05:46] <pitti> Good morning
[05:46] <pitti> stgraber: yep, will do today
[05:55] <pitti> dobey: still here?
[05:57] <pitti> dobey: ah, found the rb MP; seems rb-ubuntuone was uploaded prematurely, it's not installable right now
[06:08] <pitti> dobey: ok, one failing-to-build upgrade of bzr-builddeb lalter, it now actually merges your branch
[06:08] <pitti> pre-applied patches, would you please just die!
[06:28] <pitti> Sweetshark: seems LibO stopped building some packages on armhf? http://people.canonical.com/~ubuntu-archive/testing/precise_probs.html
[06:28] <didrocks> good morning
[06:29] <didrocks> guten morgen pitti
[06:29] <pitti> hey didrocks
[06:43] <pitti> stgraber: responded in the bug, needs to be fixed in ubiquity
[06:46] <didrocks> ok, so latest compiz fixes regressed somehow the tap detection… again :(
[06:46] <didrocks> endless story
[06:47] <didrocks> and latest nux fix regressed again ^o
[06:47]  * didrocks *sigh*, I don't smell an unity release today
[06:49]  * didrocks takes all his changes in a nux branch which is left around from some weeks already and nobody reviewed
[06:52] <RAOF> didrocks: But at least <super><space> brings up Do again :)
[06:57] <RAOF> Ah, stupid git.  Why can't you apply a stash to my unclean tree.
[06:59] <didrocks> RAOF: well, for how long, I heard about that key to bring up the hud :)
[06:59] <didrocks> RAOF: so, just a piece of advice, enjoy for now! :)
[07:00] <RAOF> Unity's going to claim *my* keybinding?  NEVER!
[07:01] <didrocks> :)
[07:18] <RAOF> didrocks: I take it that MIRs for gtest et al aren't likely to be done before the weekend?
[07:19] <didrocks> RAOF: well, with all this unity noise, not sure. gtest is doable I guess, but I have opened questions
[07:19] <didrocks> the other xorg-gtests need mterry to be around
[07:19] <didrocks> as I NEWed it in universe
[07:19] <didrocks> RAOF: for gtest, did you see my uploads?
[07:19] <RAOF> I saw one, not the other.
[07:20] <didrocks> (that take us some time as making the test build failing, and so the merging rejected for unity)
[07:20] <didrocks> RAOF: well, tere were two, but the idea is that we need to include a makefile to be able to rebuild the upstream source
[07:20] <didrocks> RAOF: I did it quickly for cmake, but I'm sure there is a better way to do it
[07:20] <RAOF> Ah, cmake file.  I didn't notice that, because I was following the google FAQ for how to build it.
[07:20] <didrocks> and same for Makefile.am?
[07:21] <didrocks> ah, there is a FAQ?
[07:21] <RAOF> We could certainly distribute a m4 snippet.
[07:21] <didrocks> yeah
[07:21] <didrocks> and something for cmake (maybe a README in the package?)
[07:21] <RAOF> Yeah; the FAQ basically says “Add gtest-all.cc to your makefile”
[07:21] <RAOF> README.Debian points at the upstream faq.
[07:22] <didrocks> but how do you detect the location of gtest-all.cc?
[07:22] <didrocks> hardcoded?
[07:22] <RAOF> Indeed.
[07:22] <RAOF> It wouldn't be a terrible idea to add an m4 macro + a pkg-config file (so we can do nice things like check the version).
[07:22] <didrocks> RAOF: agreed
[07:22] <didrocks> RAOF: volonteering? :)
[07:22] <RAOF> Ok, I guess.
[07:22] <RAOF> :P
[07:23] <didrocks> RAOF: you can remove the CMakeFile then I guess if there is the pkg-config one
[07:23] <RAOF> Well, you won't be able to PKG_CHECK_MODULES, so it might still be useful to have a CMakeFile.
[07:25] <didrocks> hum
[07:25] <didrocks> indeed
[07:25] <didrocks> but putting the cmake file in the cmake directory
[07:26] <didrocks> (so generate it with the prefix path)
[07:35] <RAOF> Anyway, I'll upload the fixed Xserver without the tests enabled now, and we can enable the tests once gtest is all done.
[07:37] <didrocks> RAOF: sounds fine
[07:57] <mvo> pitti: thanks for your aptdaemon review from the other day! do you mind if I do a upload today of the current bzr?
[08:04] <didrocks> each time I have a Xorg crash (every 1.5 days), I loose 15 minutes :/
[08:05] <RAOF> This server will probably help.
[08:07] <didrocks> RAOF: you think so? I'm experiencing this kind of crash since Oneiric (the one I talked about at UDS and at the rally)
[08:07] <didrocks> it's the "when there is too many writings on disk"
[08:07] <RAOF> Ah, right.  No, this won't help that.
[08:07] <didrocks> loosing builds are not fun :(
[08:08] <RAOF> Or, if you'd like it, I can say that it'll totally help.
[08:09] <didrocks> RAOF: just to make me happy for the week-end? And then, totally desperated again? ;)
[08:11] <RAOF> Indeed!
[08:29] <pitti> mvo: aptdaemon upload> no objection
[08:30] <mvo> ta, uploaded
[08:39] <didrocks> and a 5th attempt to try to fix alt tap detection on compiz, let's see how it goes :)
[08:45] <pitti> didrocks: got used to run large builds in screen?
[08:46] <didrocks> pitti: yeah, and replying/retweaking patches as well as we are really diverging from trunk
[08:46] <didrocks> and ABI break each times, so rebuilding everything
[08:47] <didrocks> a lot of fun :)
[08:49] <didrocks> pitti: but I still bless the ccache tricks :)
[08:49] <pitti> yeah, that should help quite a bit
[08:50] <pitti> it's even more effective for C++ (since g++ is just sloooow), but still helps for C
[08:50] <didrocks> yeah, I notice it quite clearly on compiz/nux/unity :)
[08:50] <didrocks> and as it can work with pbuilder too, it's awesome!
[08:53]  * pitti raises fist about http://www.horaoficial.cl/cambio.htm and races to update tzdata everywhere
[08:53] <pitti> that's a change that becomes effective 11 days after the announcement
[08:55] <Sweetshark> pitti: ahh indeed, some fly-by changes merged from rene. will fix
[08:55] <pitti> Sweetshark: good morning
[08:55] <pitti> Sweetshark: how's your cold today?
[08:55] <Sweetshark> pitti: better
[08:55] <Sweetshark> well, its worse. I am better.
[08:55] <pitti> heh, good to hear
[08:56] <Sweetshark> didrocks: /me builds in a pbuilder with ccache running from a jenkins ...
[08:56] <didrocks> Sweetshark: ^5s :)
[08:57] <Sweetshark> didrocks: well, I get that build down to ~2h on my notebook
[08:57] <Sweetshark> :/
[08:58] <seb128> hey
[08:59] <pitti> bonjour seb128
[09:00] <seb128> hey pitti, how are you?
[09:00]  * pitti shakes fist at seb128
[09:00] <seb128> happy friday!
[09:00] <pitti> seb128: I'm great, thanks!
[09:00] <didrocks> salut seb128
[09:00] <pitti> seb128: but busy with tzdata updates, release reports, etc.; need more time to fix 3 bugs!
[09:00] <seb128> pitti, ahah, I'm ahead again? ;-)
[09:00] <pitti> yeah, you got almost 10 yesterday
[09:00] <seb128> now I just need to find some unity bugs so didrocks keeps stalling a bit longer!
[09:00] <seb128> didrocks, lut
[09:01] <didrocks> it's really time to release unity, those guys are catching up
[09:01] <didrocks> not good :)
[09:01] <didrocks> seb128: well, don't worry, people find regressions before you :)
[09:01] <didrocks> (or rather *I* find regressions :p)
[09:01] <seb128> didrocks, yet another one?
[09:01] <pitti> seb128: the more bugs you find, the more didrocks will get on his fixed list
[09:02] <pitti> I for one yearn for a new unity
[09:02] <didrocks> seb128: yeah
[09:02] <didrocks> seb128: tapping issue back
[09:02] <didrocks> seb128: fix building…
[09:02] <pitti> the launcher AND hud constantly opening without me wanting them is a bit unnerving
[09:02] <didrocks> seb128: needing to try as well the nux jay's fix, but ABI break without a branch proposal for unity rebuildd
[09:02] <didrocks> so doing
[09:02] <pitti> and what's worse, HUD opens on top of fullscreen applications
[09:03] <seb128> that's not wanted?
[09:03] <pitti> nothing worse than having a HUD window on top of TIE fighter, which makes you miss that A-wing you were hunting down
[09:03] <seb128> I would assume it needs to be able to keyboard control those?
[09:03] <didrocks> pitti: launcher and hud opening -> that's the tapping detection
[09:03] <pitti> seb128: nah, I'm just trolling
[09:03] <seb128> lol
[09:03] <didrocks> you know, the thing that got 5 fixes :p
[09:03] <pitti> seb128: but HUD opening on Ctrl+Alt or alt+someother key or you looking the wrong way is bad
[09:04] <seb128> indeed
[09:04] <pitti> I keep getting it when switching workspaces or terminal tabs
[09:04] <didrocks> well, that's should be fixed when we can release
[09:05] <didrocks> that's one of the two things we are fighting for a week
[09:05] <seb128> \o/ slangasek fixed that gconf bug
[09:06] <didrocks> ok, found one more bug, but shoudln't be a new one
[09:09] <mitya57> hey, can anybody please triage & raise importance of bug 885730?
[09:09] <ubot2`> Launchpad bug 885730 in gnome-settings-daemon "keyboard layout indicator reacts incorrectly" [Undecided,Confirmed] https://launchpad.net/bugs/885730
[09:09] <mitya57> from comments: "I survived Unity migration, but not this "tiny" bug"
[09:09] <mitya57> :)
[09:09] <seb128> hey mitya57
[09:09] <mitya57> hey
[09:09] <seb128> mitya57, right, we have trolls on random bugs
[09:09] <mitya57> considering that one, I'm personally annoyed by it
[09:10] <seb128> charles from dx said he would look at it IIRC
[09:10] <mitya57> ok, thanks
[09:19] <mvo> seb128: hey, aptdaemon is uploaded, so the restart required bit from my branch should work now (just fyi)
[09:19] <seb128> mvo, I noticed the upload, I will try that in a bit, thanks a lot!
[09:19] <mvo> anyone where with a nouveau driver? I would love to get the output of /usr/share/pyshared/debtagshw/opengl.py from precise
[09:19] <mvo> seb128: yw
[09:19] <seb128> mvo, is there a way to emulate an upgrade that needs a restart?
[09:19] <mvo> seb128: do you happen to use the nouveau driver ;)
[09:19] <seb128> mvo, would downgrading and reupgrading a deb which set the flag be enough?
[09:19] <seb128> mvo, no, intel only for me for years
[09:20] <micahg> mvo: I've got one
[09:20] <seb128> well one ati on my old desktop
[09:20] <mvo> seb128: you could reinstall a kernel, you can also probably just run "sudo touch /var/run/reboot-required"
[09:21] <mvo> micahg: could you run the above python command for me please so that I can get the render/vendor/version that nouveau is using? that would rock :)
[09:21] <micahg> mvo: http://paste.ubuntu.com/875795/
[09:21] <seb128> mvo, thanks
[09:21] <mvo> micahg: thanks, much appreciated!
[09:21] <micahg> mvo: yw
[09:22] <mvo> the final missing bit in the video driver detection :)
[09:37] <seb128> pitti, do you want me to look at this gvfs bug?
[09:37] <pitti> seb128: you mean bug 899858 ?
[09:37] <ubot2`> Launchpad bug 899858 in oem-priority/precise "regression in gvfs to connect/browse using obex" [High,In progress] https://launchpad.net/bugs/899858
[09:37] <seb128> pitti, yes
[09:37] <pitti> if you have some time, that'd be appreciated
[09:38] <seb128> pitti, ok, adding to my list, I've some idea about it
[09:38] <pitti> seb128: unfortunately android devices don't seem to do obex file transfer, so it's not easy to test here
[09:38] <pitti> my old Nokia could, but I gave that away
[09:38] <seb128> I can test with my samsung
[09:38] <pitti> well, perhaps there's an app for it
[09:38] <pitti> seb128: oh, that's not android?
[09:39] <seb128> pitti, no, I've a semi smart bada phone
[09:57] <chrisccoulson> good morning everyone
[09:58] <didrocks> hey chrisccoulson
[09:58] <chrisccoulson> hi didrocks
[09:58]  * micahg glares at the rhythmbox-mozilla package and wonders if it's NPAPI and why it recommends epiphany :)
[09:59] <pitti> dobey: ^
[10:01] <pitti> Recommends: firefox | epiphany-browser | www-browser
[10:01] <pitti> micahg: ^ seems alright?
[10:01] <micahg> pitti: description says xulrunner based browsers which epiphany has not been in Ubuntu since karmic
[10:02] <micahg> so either it's NPAPI and should work in all browsers or it needs to be dropped
[10:02] <chrisccoulson> it's fine, the descriptions is just out of date. it's not that difficult to check ;)
[10:03] <chrisccoulson> if it's still in the archive, then it's ok
[10:03] <micahg> chrisccoulson: this is a new package
[10:05] <micahg> pitti: looks Mozilla specific, do we need this?
[10:05] <pitti> micahg: we can certainly drop the epiphany recommends
[10:05] <micahg> pitti: no, I meant the plugin
[10:05] <chrisccoulson> it isn't mozilla specific at all
[10:05] <chrisccoulson> it's the same plugin that we've had in the archive for years, just in a separate package now
[10:05] <pitti> micahg: I'm not sure; deferring to dobey
[10:06] <pitti> but I think it just got split out
[10:06] <chrisccoulson> we don't need to do anything with it, it's fine
[10:06]  * micahg defers to chrisccoulson and will file a bug for the description
[10:06] <chrisccoulson> how did you come to the conclusion that it's mozilla specific?
[10:07] <seb128> pitti, dobey: btw why adding all those new binaries to rb?
[10:07] <ricotz> dobey, hi :), where is this rhythmbox tarball coming from? cant find it anywhere and there is no 2.95.5 tag
[10:07] <seb128> did we really need to ship magnitude etc in their own binaries?
[10:07] <ricotz> seb128, hi, that would be my next question ;)
[10:07] <micahg> chrisccoulson: grepping for mozilla in the source, but now I realize that I was just looking at an alias and it's just referencing NPAPI stuff :-/
[10:08] <xclaesse> seb128, do you know if someone is working to package latest empathy (+farstream and tp-farstream) in ubuntu precise?
[10:08] <pitti> seb128: I think it was to make it easier to remove them
[10:08] <seb128> dobey, pitti: I hate adding such divergence over Debian
[10:08] <pitti> seb128: yes, I mentioned it in the FFE
[10:08] <seb128> xclaesse, kenvandine is, it's in the ubuntu-desktop ppa
[10:08] <pitti> I sponsored it this morning because it made the desktop uninstallable
[10:08] <xclaesse> seb128, good, thanks !
[10:08] <xclaesse> a ppa, exactly what I needed :D
[10:09] <micahg> pitti: you mean dobey made the desktop uninstallable by uploading a package depending on his creation before it was uploaded
[10:09] <seb128> pitti, ok, I'm all for getting the new version, I'm just unsure those packaging changes are good idea
[10:09] <pitti> it should ceratainly be done in Debian as well, or we'll revert them
[10:11] <seb128> pitti, ok, great, thanks
[10:11] <pitti> well, not "great", but let's see what debian says to the split
[10:12] <seb128> pitti, well great to know that people are not pushing to keep the divergence over Debian
[10:12] <seb128> though I would have avoided the packaging change at this point
[10:12] <seb128> not a big deal though
[10:12] <seb128> we can merge back,sort that later
[10:15] <pitti> chrisccoulson: hm, latest firefox forgot how to drag&drop tabs?
[10:16] <pitti> i. e. change the sorting?
[10:17] <chrisccoulson> pitti - that seems to work here. have you got an addon that is breaking it?
[10:17] <pitti> it was working until yesterday at least
[10:18] <pitti> chrisccoulson: adblock plus and easy youtube downloader; neither seem relevant?
[10:18] <pitti> and menubar, testpilot, ubufox of course
[10:18] <chrisccoulson> yeah, they should be
[10:19] <chrisccoulson> **shouldn't be
[10:19]  * chrisccoulson needs more coffee
=== webmaster is now known as davidcalle
[10:28] <pitti> chrisccoulson: hm, a restart fixed it; seems the automatic restart broke a tad
[10:29] <pitti> chrisccoulson: I also noticed that in that auto-restarted firefox it did a poor job of remembering the value of input fields, let's hope that's gone now as well :)
[10:29] <pitti> so perhaps the auto-restart just didn't  reset itself properly
[10:36] <micahg> pitti: I just noticed that the I got an unable to start web browser error, but the browser opened fine, any debug info I can get you before I close it?
[10:37] <pitti> micahg: was this a root/system user report or a crash from your user?
[10:37] <micahg> idk, it was unity-2d
[10:37] <pitti> micahg: "unable to start web browser" -> terminal or dialog?
[10:37] <micahg> dialog
[10:37] <pitti> micahg: ok, your user then
[10:38] <pitti> micahg: apport-cli, -gtk, or -kde?
[10:38] <micahg> apport-gtk
[10:38] <pitti> ok, that means that xdg-open exited with nonzero
[10:39] <pitti> micahg: if you run xdg-open <the URL apport opened> in a terminal, does it exit with 0?
[10:39] <micahg> pitti: what's the special shell call to check the return value of the last command?
[10:40] <pitti> micahg: echo $?
[10:40] <micahg> yep, 0
[10:40] <pitti> hmm
[10:40] <pitti> so, xdg-open is evil
[10:40] <pitti> micahg: at this point I can't think of further debug questions, I'm afraid
[10:40] <pitti> FYI, I have $? in my $PS1, it's quite handy
[10:41] <micahg> pitti: is it worth a report since i don't think I can reproduce?
[10:41] <pitti> micahg: checking existing bugs..
[10:41] <micahg> apport?  I didn't see any
[10:41] <pitti> right, doesn't seem to be reported
[10:42] <pitti> they might be against xdg-utils of course (which is where it actually belongs)
[10:42] <Sweetshark> pitti: I can has another bug pattern for bug 925049?
[10:42] <ubot2`> Launchpad bug 925049 in libreoffice "soffice.bin crashed with SIGSEGV in QCoreApplication::notifyInternal()" [Undecided,Fix committed] https://launchpad.net/bugs/925049
[10:42] <Sweetshark> pitti: please
[10:42] <pitti> Sweetshark: lazy you! :-)
[10:42] <pitti> Sweetshark: yes, can do
[10:43] <Sweetshark> pitti: no, lazy reporters not checking for obvious dupes
[10:44] <pitti> (I meant for not writing it yourself, but nevermind)
[10:45] <pitti> Sweetshark: committed
[10:47] <Sweetshark> pitti: I didnt even try, fearing some hardcore launchpad permissions and group membership minefield on the way ...
[10:47] <pitti> Sweetshark: ~ubuntu-bugcontrol
[10:48] <Sweetshark> pitti: im a member, so that can be all of it!
[10:48] <Sweetshark> s/can/cant/
[10:48] <Sweetshark> ;)
[10:48] <pitti> that's the group who can commit to lp:~ubuntu-bugcontrol/apport/ubuntu-bugpatterns
[10:48] <pitti> anyway, not a biggie
[10:49] <pitti> Sweetshark: I just wonder why the client-side duplication detection doesn't pick that up
[10:49] <micahg> pitti: I don't see an xdg-utils bug
[10:49] <pitti> http://cdimage.ubuntu.com/daily-live/20120309/
[10:49] <pitti> \o/ not oversized any more
[10:51]  * Sweetshark hopes to get to <700 open bugs in LibreOffice before release.
[10:54] <micahg> Sweetshark: wow, your not in such bad shape, we have >1k more in just the firefox source
[10:54] <micahg> s/your/you're/
[10:56] <Sweetshark> micahg: Most of those are harmless, but the bigger the number of bugs, the easier it is to let a critical one fly by under the radar.
[11:00] <seb128> mvo, new aptdaemon is buggy
[11:00] <seb128>   File "/usr/lib/python2.7/dist-packages/aptdaemon/pkcompat.py", line 2157, in _emit_require_restart
[11:00] <seb128>     trans.pktrans.RequireRestart(pk_enums.RESTART_SYSTEM, "")
[11:00] <seb128> AttributeError: 'Transaction' object has no attribute 'pktrans'
[11:00] <seb128> mvo, when installing an update which should require a restart
[11:01] <mvo> seb128: meh, give me a minute
[11:01] <seb128> mvo, the indicator-session works
[11:01] <seb128> mvo, it turned red
[11:02] <mvo> seb128: very nice
[11:09] <Sweetshark> eh, how do I remove a 'security issue' flag from a bug?
[11:09] <mvo> seb128: I need to get lunch now, sorry, I will look afterwards - do you have more backtrace? I'm not 100% certain what is causing this from my initial look
[11:10] <seb128> mvo, http://pastebin.ubuntu.com/875890/
[11:11] <seb128> mvo, enjoy lunch
[11:30] <mitya57> Something that shouldn't normally happen: I get yelp crash every time I try to search — what can be the reason? (see bug 947188)
[11:30] <ubot2`> Launchpad bug 947188 in yelp "yelp assert failure: *** glibc detected *** yelp: double free or corruption (fasttop): 0xacb71ea0 ***" [Medium,New] https://launchpad.net/bugs/947188
[11:42] <seb128> mitya57, bug?
[11:42] <mitya57> bug 947188
[11:42] <ubot2`> Launchpad bug 947188 in yelp "yelp assert failure: *** glibc detected *** yelp: double free or corruption (fasttop): 0xacb71ea0 ***" [Medium,New] https://launchpad.net/bugs/947188
[11:43] <seb128> mitya57, right, I saw that, you ask "what can be the reason" -> a bug
[11:44] <seb128> that's the reason
[11:44] <seb128> what's the reason to segfaults and issues?
[11:44] <seb128> that's a weird question to ask...
[11:44] <mitya57> what's the reason for it working for everyone except me? <- that's what I mean
[11:47] <chrisccoulson> you should try running it in valgrind
=== greyback is now known as greyback|lunch
=== MacSlow is now known as MacSlow|lunch
[12:08] <mitya57> chrisccoulson: http://paste.ubuntu.com/875954/
[12:08] <mitya57> (sorry for the delay, it hanged after start and failed to load the home page for 30 mins)
[13:02] <xclaesse> is it known that killall does not seems to work correctly on precise anymore?
[13:03] <xclaesse> killall telepathy-gabble says telepathy-gabble: no process found
[13:03] <xclaesse> even though it auto-completed it with tab
[13:03] <xclaesse> and of course, ps shows it, and kill its pid works
[13:08] <Sweetshark> xclaesse: file a bug, I guess. If you need a workaround try pkill (killall is evil anyway for breaking unix tradition)
[13:11] <pitti> chrisccoulson: I'd appreciate if you could take a look at bug 950097, I left a question there for you
[13:11] <ubot2`> Launchpad bug 950097 in ubuntu-defaults-builder "offline start page doesn't work" [Undecided,New] https://launchpad.net/bugs/950097
=== MacSlow|lunch is now known as MacSlow
[13:31] <chrisccoulson> pitti, answered now
[13:37] <pitti> chrisccoulson: ah, so I should remove that functionality from defaults-builder?
[13:38] <chrisccoulson> pitti - yeah, i think that would be best
=== greyback|lunch is now known as greyback
[14:02] <dobey> micahg, pitti: i used the same description/etc that totem-mozilla uses, for rhythmbox-mozilla. if you think it's wrong to say "xulrunner" there, we should change both i guess.
[14:02] <mdeslaur> seb128: I got a UIFe for the new lock setting
[14:02] <seb128> mdeslaur, \o/
[14:03] <mdeslaur> seb128: do I just commit to the three trees, or do you want me to upload too?
[14:04] <seb128> mdeslaur, I can handle both if you want, i.e from your debdiff
[14:04] <seb128> mdeslaur, I need to upload gsettings-desktop-schemas and gnome-settings-daemon today anyway
[14:04] <seb128> mdeslaur, and I can sneak g-c-c work if there is an upload for that
[14:05] <dobey> seb128, pitti: intend to push binnews to debian as well, yes. but rhythmbox-ubuntuone needed new crasher fixes
[14:05] <mdeslaur> seb128: wait a sec, I've bumped the gsettings-desktop-schemas depends in gnome-settings-daemon and g-c-c since I posted the debdiffs
[14:05] <seb128> dobey, it's just surprising that you went for new binaries changes
[14:06] <seb128> mdeslaur, that's fine, I can fix the depends
[14:06] <dobey> seb128: sorry. i listed them in the ffe bug report i filed the other day
[14:06] <mdeslaur> seb128: ok, then the debdiffs + the bumped depends
[14:06] <mdeslaur> seb128: thanks!
[14:06] <seb128> dobey, well it's just that I don't see the point to divert from Debian on that
[14:06] <seb128> mdeslaur, thanks
[14:07] <seb128> mdeslaur, I might rename the key
[14:07] <mdeslaur> seb128: sure!
[14:07] <seb128> mdeslaur, i.e ubuntu-lock-on-suspend
[14:07] <seb128> mdeslaur, I like to have non upstream stuff ubuntu- prefixed, avoid confusion for everybody
[14:07] <mdeslaur> that sounds good
[14:07] <seb128> great ;-)
[14:09] <desrt> chrisccoulson: hello
[14:10] <desrt> chrisccoulson: i've set the 3 different settings for "send fucking plain text mail" in thunderbird and it's still trying to send html
[14:10] <dobey> seb128: well, external plug-ins having to Depends: rhythmbox-plugins, to work, is a bit broken. and it's nice to be able to remove someo of the plug-ins. anyway, i will work to push the changes into debian as well. was just a bit rushed to get it into ubuntu as there are crash fixes rhythmbox-ubuntuone needs :-/
[14:10] <desrt> is there a 4th and 5th setting somewhere?
[14:10] <dobey> desrt: don't stress so much. the hair on your neck is turning grey.
[14:11] <seb128> dobey, ok, thanks
[14:11] <seb128> hey desrt
[14:11] <kenvandine> good morning folks
[14:11] <kenvandine> seb128, farstream sync is in NEW
[14:12] <dobey> pitti, seb128, kenvandine: care to look at https://code.launchpad.net/~dobey/ubuntu/precise/twisted/fix-935756/+merge/96617 ? :)
[14:12] <seb128> hey kenvandine
[14:13] <kenvandine> dobey, sure
[14:13] <dobey> thanks
[14:13] <seb128> kenvandine, no it isn't ;-)
[14:13] <kenvandine> seb128, old joke, i'm onto you!
[14:13] <kenvandine> :-D
[14:13] <seb128> kenvandine, hehe
[14:14] <seb128> kenvandine, I'm going to upload a gtk in a bit, how did it go for the eventbox scrolling stuff?
[14:14] <kenvandine> dobey, i would rather have someone more familiar with twisted, it scares me
[14:15] <kenvandine> maybe ping doko?
[14:15] <seb128> dobey, same than kenvandine
[14:15] <dobey> ok
[14:15] <kenvandine> seb128, sort of... having to explicitly set GDK_SCROLL_MASK on widgets seems like it is going to have tons of side effects
[14:15] <kenvandine> adding it to the eventbox sort of helped gwibber
[14:16] <kenvandine> but all the widgets in the eventbox blocks the scroll events
[14:16] <seb128> kenvandine, can you move to discuss it on the #gtk+ channel?
[14:16] <kenvandine> so everywhere there is a label it gets ignored
[14:16] <kenvandine> yeah, i about to do that
[14:16] <seb128> kenvandine, they will probably unblock you quickly compared at having to poke and figure it by yourself
[14:16] <dobey> awesome
[14:16] <dobey> it's like scrolling a web page and being like "oh hi flash!"
[14:20] <ricotz> dobey, hi
[14:21] <dobey> ricotz: hi
[14:21] <kenvandine> seb128, i also suspect we might have more xi2 problems, if i add GDK_SCROLL_MASK and GDK_SMOOTH_SCROLL_MASK all i get are GDK_SMOOTH_SCROLL events
[14:21] <ricotz> dobey, where is this rhythmbox tarball coming from? and there is no 2.95.5 tag
[14:21] <kenvandine> which afaik i shouldn't, since i don't have the right kind of device
[14:21] <dobey> ricotz: http://people.gnome.org/~jmatthew/rhythmbox-2.95.5.tar.xz
[14:22] <seb128> kenvandine, yeah, it's for sure having some issues, I hope they get sorted soon
[14:22] <dobey> ricotz: it's a "prerelease" he made for me, and not an official tagged release. yeah i know :-/
[14:22] <kenvandine> not sure if that would be in gdk or X though
[14:22] <ricotz> dobey, hmm, i see, do you know which git hash it is based on?
[14:23] <stgraber> pitti: thanks, I must admit I simply tried to find equivalents for our current code and as Engine existed I simply used it ;)
[14:24] <stgraber> pitti: will do the changes in ubiquity now and see if that works as expected
[14:24] <ricotz> dobey, nm, the changelog says it
[14:24] <dobey> ricotz: ok
[14:25] <dpm> hi pitti, it seems we've still got the language pack version 20120221 in precise, but I see that the cron job is enabled. Do you know why langpacks are not being released after the beta freeze?
[14:26] <Beret> does anyone know if there's going to be a way to disable the HUD in precise?
[14:26] <Beret> (CCSM doesn't count)
[14:27] <ricotz> dobey, the mpris patch isnt included in this version
[14:27] <dobey> Beret: dconf-editor
[14:27] <dobey> ricotz: yes it is
[14:28] <ricotz> dobey, no, it isnt
[14:28] <ricotz> could you check again?
[14:29] <dobey> ricotz: MPRIS_BUS_NAME_PREFIX ".rhythmbox",
[14:29] <dobey> ricotz: that's what's in the code in the tarball
[14:30] <ricotz> dobey, i mean this one http://git.gnome.org/browse/rhythmbox/commit/?id=778d3317900337dc7177d0a80ee052f966094485
[14:32] <dobey> ricotz: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/precise/rhythmbox/precise/revision/201 doesn't include that. the changes in the patch added in that revision are included in this tarball.
[14:32] <ricotz> dobey, ah there where two mpris fixes
[14:32] <dobey> ricotz: so i guess that patch was missing this change too then?
[14:33] <ricotz> dobey, yeah sorry, i think so
[14:34] <dobey> ricotz: i think moch will make a 2.96 release in the next few days. if you want to add that patch or wait for tarball, either way is fine by me. :)
[14:35] <ricotz> dobey, shouldnt be that important, just thought this was the patch that was added at last
[14:36] <dobey> ricotz: ok
[14:50] <stgraber> pitti: any idea why the .name property of a ConfigItem contains that much garbage http://paste.ubuntu.com/876095/?
[14:51] <stgraber> pitti: http://paste.ubuntu.com/876097/ is my current diff for converting from xklavier to gir Xkl
[15:00] <kenvandine> seb128, farstream is in binNEW
[15:01]  * kenvandine waits for seb128's response :)
[15:01] <seb128> kenvandine: is that aimed to main? I didn't follow much
[15:01] <kenvandine> yes
[15:01] <kenvandine> pitti said no MIR
[15:01] <seb128> kenvandine: I'm sure you didn't wait for this response :p
[15:01] <kenvandine> since it is a rename
[15:01] <seb128> kenvandine: ok, good
[15:02] <pitti> yay, my network is back
[15:02] <pitti> stgraber: ah, great
[15:02] <kenvandine> pitti, having network is good, mkay'
[15:02] <seb128> kenvandine: I will wait for the amd64 binaries
[15:02] <kenvandine> :)
[15:02] <pitti> dpm: I only re-enabled the cronjob yesterday or so
[15:02] <kenvandine> seb128, ok
[15:02] <pitti> dpm: and I saw langpacks on the buildds today
[15:02] <seb128> kenvandine: should be there in less than half an hour
[15:02] <seb128> kenvandine: is it blocking you?
[15:02] <kenvandine> seb128, thx
[15:02] <kenvandine> well, i was going to wait for it to upload all the things that need it
[15:03] <kenvandine> i hate having to hit retry a bunch of time for things
[15:03] <seb128> kenvandine: things will dep wait, don't bother
[15:03] <kenvandine> yeah, but they will queue to build and then dep wait then queue to build again
[15:03] <kenvandine> i guess the builders aren't busy today
[15:05] <pitti> stgraber: it's indeed a bit of a nasty API, see the wrappers I had to do in the test-gi.py script
[15:05] <pitti> stgraber: it's delivered as an array, not a staring
[15:05] <pitti> string
[15:05] <kenvandine> seb128, so with the smooth scrolling landing in gtk, vala apps that depend on GDK_SCROLL events will be broken until vala updates to the latest gtk
[15:05] <seb128> kenvandine: is that the same will all bindings?
[15:05] <kenvandine> i still think we need to add that back to the eventbox, that is a common use case for an eventbox
[15:05] <kenvandine> i assume
[15:06] <seb128> :-(
[15:06] <seb128> what changed?
[15:06] <kenvandine> well, gir is easier
[15:06] <kenvandine> the gtk gir should already have it right?
[15:06] <seb128> gtk stopped emiting scroll events in favor of smooth scrolling ones?
[15:06] <kenvandine> yes
[15:06] <kenvandine> if you have a xi2 server
[15:07] <kenvandine> you get the smooth scroll event
[15:07] <pitti> stgraber:
[15:07] <pitti>     i = s.find(b'\x00')
[15:07] <pitti>     return s[:i].decode()
[15:07] <kenvandine> and the vapi doesn't know about it
[15:07] <seb128> kenvandine: still it's an abi change late in the cycle, I start pondering reverting smooth scrolling for precise
[15:07] <pitti> stgraber: that's my item_str(s) method which converts a zero-terminated int8 array into a proper str
[15:07] <stgraber> pitti: oh, fun ;) ok will update to that then...
[15:07] <kenvandine> seb128, i do fear how many unknown side effects there are
[15:08] <pitti> stgraber: accessing bare structs with static data is fun :/
[15:08] <kenvandine> i could manually go add the SCROLL_MASK to every place i create an event box in gwibber and if we don't patch gtk to add them to the eventbox
[15:08] <kenvandine> gwibber will be fine
[15:09] <kenvandine> but i think the right thing to do is add both scroll and smooth scroll to the eventbox, it is common reason to use an eventbox
[15:09] <kenvandine> but if it has GDK_SMOOTH_SCROLL_MASK on it, then that is what gets emitted
[15:09] <stgraber> pitti: wow, the test actually pass now! thanks
[15:09] <kenvandine> instead of GDK_SCROLL
[15:09] <pitti> stgraber: \o/
[15:09] <kenvandine> so if apps depend on that, they will break
[15:09] <kenvandine> or in the vala case, the vapi doesn't know about them yet
[15:10] <dobey> kenvandine: gwibbre has an eventbox with a box with a bunch of eventboxes?
[15:19] <chrisccoulson> desrt, sorry, was on the bike
[15:19] <chrisccoulson> there aren't any more settings that i'm aware of :/
[15:19] <kenvandine> dobey, yes... because we do our own scrolling
[15:19] <chrisccoulson> desrt, you're not using it under unity, are you?
[15:20] <dpm> pitti, ah, ok, thanks!
[15:20] <dobey> kenvandine: why not just shove it in a scrolledwindow?
[15:20] <kenvandine> oh i wish i could
[15:20] <kenvandine> i will evently do something like that
[15:20] <kenvandine> eventually
[15:21] <dobey> if you can't, your code is broken :)
[15:21] <kenvandine> right now it is a several boxes, side by side in a group
[15:21] <kenvandine> and each box scrolls
[15:22] <dobey> meh
[15:22] <desrt> chrisccoulson: no.
[15:22] <kenvandine> and they also only contain as many tiles as you can see
[15:23] <kenvandine> njpatel was getting really fancy with that stuff
[15:23] <kenvandine> it just ended up being rather complicated
[15:24] <chrisccoulson> desrt, are you replying to a mail?
[15:25] <desrt> yes
[15:25] <desrt> the mail in question is in plain text
[15:25] <desrt> restarting thunderbird fixed it
[15:25] <desrt> i really only did require 3 settings
[15:25] <chrisccoulson> lol
[15:25] <chrisccoulson> bugger :(
[15:26] <desrt> thunderbird seems to be built on the assumption that you'd only ever want to send text email because something is wrong
[15:26] <desrt> ie: it lets you list particular people to avoid sending html mail to
[15:27] <chrisccoulson> the reason i asked if you were using it under unity before, is because the radio buttons in the Format submenu seem to be broken
[15:27] <chrisccoulson> ie, when i select "Plain text" here, i get 2 items selected ;)
[15:31] <dobey> desrt: gsettings/dconf-daemon talk over dbus, right?
[15:31] <desrt> gsettings uses the dconfgsettings backend which talks to the dconf-service over dbus, yes
[15:32] <desrt> it's possible for gsettings to use non-dconf backends, though
[15:32] <desrt> and the 'talking' only happens for writes
[15:32] <dobey> oh. reads happen directly via open() on the db?
[15:32] <desrt> yes
[15:32] <dobey> bummer
[15:32] <desrt> that's why it's fast
[15:32] <desrt> it's also why i'm on a campaign to remove all writes on login
[15:33] <desrt> because then the service doesn't need to be running at all
[15:33] <dobey> is there any way to make it not do that?
[15:33] <desrt> not do what?  reads?
[15:33] <dobey> not read settings, that is.
[15:33] <desrt> don't call g_settings_get()?
[15:33] <dobey> yes, so user settings don't screw up tests
[15:33] <desrt> oh
[15:33] <desrt> GSETTINGS_BACKEND=memory
[15:33] <dobey> ah
[15:34] <desrt> that will do all gsettings lookups against an (initially empty) GHashTable in the local process
[15:34] <dobey> desrt: excellent. thanks!
[15:36] <jbicha> what does this mean, re: the clutter 1.9 ffe ? http://lists.clutter-project.org/pipermail/clutter-devel-list/2012-March/000358.html
=== JanC_ is now known as JanC
[15:54] <seb128> cyphermox, hey
[15:55] <cyphermox> hey seb128
[15:56] <cyphermox> what's up
[15:57] <seb128> cyphermox, how are you?
[15:57] <seb128> cyphermox, I noticed there is a new bluez,obex, want to have a look next week if that's updates we should do?
[15:57] <cyphermox> pretty good, an you
[15:57] <seb128> cyphermox, the bluez.org description suggest it's mostly bug fixes
[15:57] <cyphermox> yes, I'll merge from debian for bluez
[15:57] <seb128> cyphermox, I'm good thanks ;-)
[15:58] <cyphermox> obex I hadn't seen but there's one or two things I want to look at today in bluetooth, once I'm done with the update of NM.
[15:58] <cyphermox> seb128: btw, are you in the release team?
[15:58] <seb128> cyphermox, no, I'm not
[15:58] <cyphermox> ah
[15:59] <cyphermox> I'm preparing an update of NM now to another snapshot, and 0.9.4 is supposed to land around March 20.
[16:00] <cyphermox> pitti:  ^^ I want to get a release tarball in precise rather than snapshots if possible
[16:00] <pitti> cyphermox: indeed, appreciated
[16:00] <cyphermox> read that "around march 20" as "before March 20"
[16:01] <cyphermox> pitti: there's a FFe bug open for that: https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/950313
[16:01] <ubot2`> Launchpad bug 950313 in network-manager "[FFe] Update NetworkManager to a recent snapshot in prevision for the 0.9.4 release." [Undecided,New]
[16:01] <pitti> cyphermox: does it actually have new features?
[16:02] <desrt> hum
[16:02] <desrt> my system is running like a pig
[16:02] <cyphermox> kind of. connectivity checking, vlan and bonding weren't completely in by feature freeze
[16:02] <desrt> like, when i open a new firefox window, it takes about 3 seconds for the window to first draw itself
[16:02] <cyphermox> at the sprint I tried to merge in connectivity checking but it didn't work well enough
[16:03] <pitti> cyphermox: ah, saw the commit list, very detailled; thanks
[16:03] <desrt> anyone else having the same?
[16:03] <pitti> my computer freezes to a crawl whenever its busy with large files
[16:03] <pitti> heavey IO really sucks
[16:03] <desrt> pitti: i also noticed that compiling was quite a lot slower than i am used to...
[16:03] <cyphermox> pitti: hopefully that's detailed enough for the release team to make an informed decision. I'm not overly worried, and I just finished building a pacakge to give it a good test run
[16:03] <pitti> and this is "fast" class SSD
[16:03] <chrisccoulson> mine freezes to a crawl every time i try to compile anything now
[16:04] <desrt> new kernel?
[16:04] <chrisccoulson> i can literally watch the frames when i try to switch workspaces
[16:04] <chrisccoulson> the workspace switching animation takes in excess of 5 seconds when i'm compiling things now
[16:04] <pitti> desrt: not for me; linux is like that for years
[16:04] <chrisccoulson> that only started recently
[16:04] <desrt> ya.  for me it's definitely gotten quite a lot worse recently
[16:04] <desrt> like, i only noticed for the first time today
[16:05] <ogra_> switch to a 2.2 kernel ... lots faster !
[16:08] <pitti> ogra_: I'm sure udev and upstart will like that a lot!
[16:08] <ogra_> haha
[16:09] <desrt> extra points for systemd :)
[16:10] <ogra_> shudder
[16:14] <BigWhale> is network manager installed on ubuntu-server?
[16:15] <pitti> no
[16:15] <desrt> MacSlow: hey
[16:16] <BigWhale> pitti, I thought so. Thanks.
[16:16] <desrt> MacSlow: i want to try to tackle this unity average-bg-color thing today
[16:16] <MacSlow> desrt, still busy with design-stuff here
[16:16] <desrt> MacSlow: can you point me at a definitive list of all the places that it is used?  i know that unity sets it, the dash uses it and also notify-osd
[16:16] <desrt> MacSlow: i'll do the work
[16:16] <desrt> MacSlow: i just need to know where to do the work :)
[16:17] <MacSlow> desrt, so it's only used for two things... to tint the dash and picked up in notify-osd (although there's still a bug somewhere with picking up the color so it turns out all wrong)
[16:18] <desrt> MacSlow: i plan to turn it into a property on the root window
[16:20] <MacSlow> desrt, iirc it's used in class BGHash (unity side) and in notify-osd picked up in class Defaults (defaults.c)
[16:20] <desrt> MacSlow: i guess unity is capable of communicating with itself for the purpose of the dash?
[16:20] <MacSlow> desrt, sure
[16:20]  * desrt sees only the setting being set from unity
[16:21] <desrt> MacSlow: so one writer: unity and one reader: notify-osd
[16:21] <MacSlow> desrt, correct
[16:21] <desrt> okay.  should be easy.
[16:21] <desrt> particularly since notify-osd is transient
[16:21] <desrt> it means i don't have to watch for changes -- only query it just before showing each bubble
[16:21] <desrt> i assume you don't care that the bubble dynamically changes background colour if a background fade is in progress at the time the bubble is showing :)
[16:22] <MacSlow> desrt, no it won't do that... it'll pick the avg. color currently stored locally when creating a bubble
[16:23] <MacSlow> if created and the avg. color changes only the next bubbles will pick that one up
[16:23] <desrt> cool
[16:23] <desrt> i'm not even going to store it locally, though
[16:23] <desrt> just query the server every time
[16:27] <desrt> interesting.  _XROOTPMAP_ID(PIXMAP): pixmap id # 0x1000001
=== zyga is now known as zyga-afk
[16:41] <mdeslaur> seb128: I've attached a new debdiff that changes the switch to a checkbox, much nicer
[16:42] <seb128> mdeslaur, ok, good
[16:42] <seb128> I've almost done with g-s-d
[16:42] <seb128> soon to gcc ;-)
[16:42] <mdeslaur> cool :)
[16:49] <pitti> good night everyone! have a nice weekend
[16:49] <pitti> seb128: and stop upping your bug counter, impossible to keep up :)
[16:50] <seb128> pitti, 'night ;-)
[16:50] <seb128> have a good w.e
=== zyga-afk is now known as zyga
[17:44] <chrisccoulson> right, i've got a valgrind capable build of thunderbird. now time to see who is leaking memory here :)
[18:14] <desrt> chrisccoulson: EVERYONE
[18:14] <chrisccoulson> heh
[18:14] <chrisccoulson> it's a pain having to rebuild it just so i can run it in valgrind
[18:17] <seb128> mdeslaur, there?
[18:17] <mdeslaur> seb128: yes
[18:18] <seb128> mdeslaur, is there any reason why you packed a checkbox and a label in a box?
[18:18] <mdeslaur> seb128: what's up?
[18:18] <seb128> mdeslaur, rather than just setting the label on the GtkCheckButton
[18:18] <seb128> mdeslaur, just checking before "fixing" it
[18:18] <seb128> mdeslaur, the way you did it you can't click on the label to check,uncheck the box
[18:18] <mdeslaur> seb128: d'oh, no...it's just the result of me converting it from the switch
[18:19] <seb128> mdeslaur, ok, I'm changing it then, no worry,
[18:19] <seb128> mdeslaur, thanks
[18:19] <mdeslaur> seb128: sorry about that
[18:19] <mdeslaur> thanks
[18:19] <seb128> no worry, thanks for the good work ;-)
[18:19] <desrt> gord: ping
[18:23] <didrocks> have a good week-end everyone!
[18:23] <desrt> didrocks: you too. bye :)
[18:23] <didrocks> desrt: thanks ;)
[18:23] <mterry> didrocks, bye!
[18:23]  * mterry hides again
[18:29] <seb128> oh, a mterry!
[18:29] <seb128> mterry, how are you?
[18:30] <seb128> mterry, still busy trying to catch up with design requests? ;-)
[18:30] <mterry> seb128, yeah a bit.  Trying to figure out performance issues in unity-greeter, which design didn't like
[18:30] <seb128> mterry, do you also get that 1 second "hang" in the select user animation when switching to the greeter?
[18:32] <mterry> seb128, not sure.  explain?
[18:35] <seb128> mterry, not sure how I trigger it exactly now, sometime when switching users I'm send to the greeter, and when the greeter starts there is a "sliding effect" to select my users which is not at the top of the list
[18:35] <seb128> mterry, that "slide to the user to select" often is jerky
[18:36] <mterry> hm
[18:36] <seb128> like it stops for some 0.1s and resume
[18:36] <mterry> i guess that's the general thing.  stopping jerkiness
[18:36] <mterry> Not sure the cause yet
[18:36] <seb128> ok
[18:36] <mterry> seb128, do you have users with long names?
[18:36] <seb128> well I was curious rather than complaining
[18:36] <mterry> such that they fade off?
[18:36] <seb128> no
[18:36] <mterry> bummer, i fixed a slow down with those
[18:36] <seb128> well animation is smooth once the greeter is loaded
[18:36] <seb128> I though it was because it was io busy or something
[18:37] <mterry> maybe it's just cpu bound or something on startup...
[18:37] <seb128> right
[18:37] <mterry> but I've seen jerkiness in other situations
[18:40] <seb128> mvo, bug #950791
[18:40] <ubot2`> Launchpad bug 950791 in aptdaemon "<type 'exceptions.AttributeError'>: 'Transaction' object has no attribute 'pktrans'" [Undecided,New] https://launchpad.net/bugs/950791
[18:40] <seb128> mvo, that's the bug I mentioned this morning with the new aptdaemon
=== jalcine_ is now known as jalcine
[19:31] <gatox> hi, does anyone know where i can report a bug about the hud?
[19:31] <desrt> seb128: seeing #gnome-hackers?
[19:34] <dobey> gatox: unity
[19:35] <gatox> dobey, ok, i was asking because maybe was something specific for that
[19:35] <gatox> dobey, thanks
[19:35] <dobey> gatox: dx team will file it against unity for tracking anyway
[19:46] <chrisccoulson> uh-oh, "Solihull Police (@SolihullPolice) is now following you on Twitter"
[19:46] <chrisccoulson> that's not what i want to see!
[19:46] <chrisccoulson> ;)
[19:51] <dobey> chrisccoulson: haha
[19:55] <seb128> desrt, specifics? it's friday night and I don't feel like reading 5 pages of scrollback to figure what you pinged about ;-)
[19:58] <desrt> seb128: so how unity works now is something like this:
[19:58] <desrt> gnome-control-center modifies the background image settings, writes to gsettings
[19:58] <desrt> g-s-d picks those up, loads the image, writes to a pixmap in X and sets a property on the root window to point to the pixmap
[19:58] <desrt> unity sees this change, loads the image out of the X server, does an average calculation on it, sets the average-bg-color in GSettings
[19:58] <desrt> notify-osd picks it up
[19:58] <seb128> what has g-s-d to do with that?
[19:59] <desrt> the g-s-d background plugin is what is responsible for actually setting the background
[19:59] <seb128> no
[19:59] <seb128> nautilus is
[19:59] <seb128> but I guess same difference
[19:59] <desrt> very well
[19:59] <desrt> in either case, they're using libgnome-desktop "gnome-bg"
[19:59] <seb128> it's probably g-s-d in GNOME because nautilus doesn't handle the background
[19:59] <seb128> sorry, please keep going ;-)
[20:00] <desrt> i'm doing a patch now to libgnome-desktop to set the average colour in a new property on the X root window
[20:00] <seb128> great
[20:00] <desrt> and notify-osd can just read it straight from there
[20:00] <desrt> so it will work even without unity
[20:00] <seb128> \o/
[20:00] <desrt> looks like it will be accepted too
[20:00] <seb128> is there a standard definition on what the average color is?
[20:00] <desrt> Company and halflife are already arguing over ways to make the algorithm 'more intelligent' than just using the average :)
[20:00] <seb128> like something everybody agrees on?
[20:00] <desrt> well
[20:00] <desrt> there is some existing code there, and i use that
[20:01] <desrt> it's the same code the answers the gnome_bg_is_dark() question
[20:01] <desrt> basically, just adds up all the pixels
[20:01] <desrt> and divides by the number of pixels
[20:01] <desrt> it sounds like we may get fancy here, though
[20:01] <desrt> a great example is if you have green grass and a blue sky
[20:02] <desrt> in that case it sounds like Company wants to say that there are two representative colours
[20:02] <desrt> taking the average blue and the average green, separately
[20:02] <seb128> right
[20:02] <seb128> it opens the way to complex logic though
[20:02] <seb128> then you need to consider what are the regions and where your ui is placed
[20:02] <desrt> ya.  i was thinking about that.
[20:03] <desrt> i think we should not worry about it
[20:03] <desrt> this approach is already better than the existing one
[20:03] <seb128> right, we can still improve over time if needed
[20:03] <seb128> but the medium value should be a good start
[20:03] <chrisccoulson> ouch: https://twitter.com/#!/paulrouget/status/178198646188294144
[20:03] <seb128> desrt, thanks for working on that ;-)
[20:03] <desrt> if needed, we could do an average for each 'quadrant' or so
[20:03] <chrisccoulson> G+, Y U USE SO MUCH MEMORY????
[20:03] <chrisccoulson> :)
[20:04] <seb128> chrisccoulson, stop twitter, it's bad for you! ;-)
[20:04] <micahg> Firefox trunk with load on demand tabs seems to be sipping memory now :)
[20:04] <chrisccoulson> lol
[20:04] <chrisccoulson> seb128, i get all of my information from twitter now. even the news!
[20:04] <chrisccoulson> in fact, i actually do find things out on twitter before i see them anywhere else now
[20:04] <chrisccoulson> that's bad ;)
[20:05] <seb128> chrisccoulson, seeing the noise ration you must be spending a lot of time to get the infos you need ;-à
[20:05] <seb128> ;-)
[20:05] <chrisccoulson> lol
[20:05] <chrisccoulson> i don't post that much tbh
[20:05] <seb128> well posting is fine :p
[20:05] <seb128> it's dealing with the flood of infos which is not :p
[20:05] <seb128> chrisccoulson, are you one of those who have notify-osd always on screen with twitter infos? ;-)
[20:06] <seb128> like constant activities all day long? ;-)
[20:06] <chrisccoulson> seb128, no, i don't have gwibber set up at all
[20:06] <chrisccoulson> i find the desktop notifications really annoying ;)
[20:06] <chrisccoulson> for twitter, anyway
[20:06] <seb128> chrisccoulson, well it's a way to get the tweets on screen before going to a browser
[20:06] <seb128> chrisccoulson, but I guess a browser is your desktop anyway :p
[20:06] <chrisccoulson> having twitter open in my browser is much less distracting :)
[20:07] <seb128> chrisccoulson, hum, I need to bug something to get you to do desktop work again ;-)
[20:07] <seb128> you need to get you of your browser
[20:07] <chrisccoulson> lol
[20:07] <seb128> that would be good for you :p
[20:08] <seb128> get *out*
[20:08] <chrisccoulson> you just need to wait for DX to break something ;)
[20:08] <dobey> chrisccoulson: so give it 5 minutes?
[20:08] <chrisccoulson> :)
[20:08] <chrisccoulson> i didn't want to be the person who said that
[20:11] <kenvandine> chrisccoulson, you know you can change what notifications you get... like none at all or just for mentions and private
[20:11] <kenvandine> which is the default now
[20:11] <chrisccoulson> kenvandine, ah, i might try that. i haven't used gwibber for quite a long time now
[20:11] <chrisccoulson> i should try it again :)
[20:11] <kenvandine> you should :)
[20:12] <dobey> just don't try to scroll
[20:12] <dobey> or well, you can try i guess
[20:12] <dobey> just won't go anywhere :)
[20:12] <kenvandine> dobey, that should be fixed with the gtk upload seb128 is doing
[20:13] <seb128> kenvandine, is doing -> did 3 hours ago
[20:13] <chrisccoulson> oh, i don't even have gwibber installed. i wonder when that disappeared
[20:13] <kenvandine> seb128, great :)
[20:16] <desrt> seb128: xsettings override branch just merged
[20:17] <seb128> desrt, waouh
[20:17] <seb128> desrt, I will look to it on monday for sure ;-)
[20:17] <desrt> ya.  it's not super-high priority
[20:18]  * desrt is more curious to see what with happen with the representative colours patch
[20:19] <seb128> desrt, btw I was close to consider adding gconf usage again today ;-)
[20:19] <desrt> why?
[20:20] <seb128> desrt, because of the gsettings schemas stuff making it so difficult to add an option key for a "gnome_me_harder" use
[20:20] <seb128> you always have to consider where to add the schemas to make sure it's there first
[20:20] <desrt> seb128: why don't we have a global schema for gnome-me-harder?
[20:20] <desrt> you could patch it into gsettings-desktop-schemas
[20:20] <seb128> or to add safegards around it
[20:20] <desrt> which the entire world already depends on
[20:20] <seb128> desrt, I did the gsettings-desktop-schemas thing at the end
[20:20] <desrt> great
[20:20] <desrt> btw: you can use dconf
[20:21] <desrt> if you don't want to write a schema
[20:21] <seb128> is there a C api?
[20:21] <desrt> yes
[20:21] <seb128> ok, I should look at that ;-)
[20:21] <desrt> don't you package this stuff? :)
[20:21] <seb128> I do, but I never saw anyone using the dconf C api in a software
[20:21] <desrt> vuntz uses it in gnome-panel
[20:21] <seb128> so I was not sure how supported it was :p
[20:21] <desrt> robert_ancell uses it in dconf-editor
[20:21] <desrt> uhm... that's all, i think :)
[20:22] <seb128> hehe
[20:22] <seb128> don't get me started on dconf-editor :p
[20:22] <seb128> I stopped used it
[20:22] <desrt> the great thing about dconf is that the #1 thing that people complain about is dconf-editor
[20:22] <desrt> and i can say "robert's fault!"
[20:22] <seb128> lol
[20:22] <seb128> well he blames it on you :p
[20:22] <seb128> you made it so it's very hard to write an editor
[20:23] <desrt> yes and no...
[20:23] <desrt> the general design makes it difficult
[20:23] <jalcine> lol
[20:23] <desrt> which is caused by gsetting's aim to be something more than just a replacement for gconf
[20:23] <desrt> ie: you use it unconditionally from your software and it acts as a frontend to the registry on windows, property lists on the mac and dconf on linux
[20:24] <desrt> so there's this split.... do we want an editor for gsettings (that would also work on windows...)
[20:24] <desrt> or one just for dconf?
[20:24] <desrt> robert does a reasonable job of standing on both sides of that line....
[20:24] <desrt> but of course there are complications caused by that
[20:25] <seb128> desrt, my main complain with it is that it lies about key values
[20:25] <seb128> it ignores overrides
[20:25] <desrt> ya
[20:25] <seb128> which screwed me a few time in debugging
[20:25] <seb128> I stopped trusting it
[20:25] <desrt> it wouldn't be that hard for him to fix that.....
[20:25] <desrt> alas
[20:26] <seb128> it's like if you can trust a viewer to show the right value it's useless
[20:26] <stgraber> gsettings list-recursively does a reasonable job for debugging (I also gave up on dconf-editor a while ago)
[20:26] <seb128> can't
[20:26] <desrt> seb128: do you have any idea how notify-osd works?
[20:26] <johan> seb128: seems broadway/html5 isn't enabled in the latest gtk-3 packaged in precise, is there anything I can do to make that change go in?
[20:26] <desrt> stgraber: dconf dump as well
[20:26] <seb128> johan, hey, you already pinged me about that on #gnome-hackers ;-)
[20:27] <johan> seb128: yes, but that was before you uploaded the latest version :D
[20:27] <seb128> johan, it's not enabled upstream and mclasen nacked a fedora request to enable is saying it's experimental and shouldn't be enabled
[20:27] <seb128> johan, if it's too experimental for fedora it makes me wonder if I really want to enable it :p
[20:27] <johan> seb128: :(
[20:27] <desrt> ubuntu: always copying what fedora does :p
[20:27] <johan> to get it tested of course
[20:28] <johan> you do need to set an environment variable to enable it, so it's not for the light hearted
[20:28] <desrt> seb128: you could add yet another build variant of gtk :)
[20:28] <seb128> well, lts are not where you get stuff tested, it's where you get stuff solid ;-)
[20:28] <seb128> johan, I don't know enough about it, I wanted to check what it means for depends and build-depends
[20:28] <seb128> but maybe desrt knows?
[20:29] <seb128> desrt, does enabling a backend like that adds run time depends?
[20:29]  * desrt knows nothing
[20:29] <desrt> seb128: yes.  i'd imagine so.
[20:29] <desrt> seb128: the backends are not loaded as separate modules, if that's what you mean
[20:31] <seb128> desrt, yeah, so enabling them at build-time has an impact, it's not just adding a .so somewhere you can opt in
=== zyga is now known as zyga-weekend
[20:31] <desrt> seb128: looks like a pretty minimal impact in terms of extra deps...
[20:31] <desrt> libz, basically
[20:31] <desrt> so mostly just increasing the code size, i guess
[20:32] <desrt> 6000 extra lines of C
[20:33] <desrt> it's my opinion that there's no obvious harm to turning it on...
[20:33] <desrt> that doesn't mean that there's no harm, of course :)
[20:34] <seb128> desrt, yeah, I will check with mclasen another day
[20:34] <seb128> johan, my gut feeling is that a ppa for that is good enough
[20:34] <seb128> it's clearly not something import as it in the lts
[20:34] <seb128> like no normal user need it
[20:34] <johan> seb128: yeah, I guess you're right
[20:36] <desrt> ugh
[20:36] <desrt> _get_unity_schema()
[20:36] <desrt> leaks...
[20:44] <dobey> GLib-GObject-WARNING **: /build/buildd/glib2.0-2.31.20/./gobject/gsignal.c:2572: instance `0x91f1200' has no handler with id `3170'
[20:44] <dobey> hrmm; i am getting a *LOT* of those getting printed for various things
[20:45] <cyphermox> seb128: I'm not as certain that bluez 4.99 looks like bugfix :/
[20:45] <seb128> cyphermox, ok, I didn't look, I just saw the summary and wanted us to have a look
[20:45] <cyphermox> I see a bunch of things that make me worry a bit about how much fun this risks being to re-test everything and behavior changes
[20:45] <cyphermox> ok
[20:45] <seb128> cyphermox, if you think we shouldn't update that's fine
[20:46] <cyphermox> though I noticed a few patches we might want to cherry-pick :)
[20:46] <seb128> dobey, update your libdbusmenu to 0.5.93-0ubuntu2
[20:49] <Arlo> Hi everybody, long time listener, first time caller. I work for a VoIP startup that recently released a voip/chat/video chat tool for ubuntu 32/64. I am wondering what the process would be for us to get it into the ubuntu repo?
[20:49] <desrt> Arlo: is it free software?
[20:50] <Arlo> yes. http://www.icall.com/get
[20:50] <desrt> did you read http://developer.ubuntu.com/publish/ ?
[20:51] <Arlo> apparently not, that answers all of my questions. thank you.
[20:51] <desrt> :)
=== jalcine is now known as JackyAlcine
=== JackyAlcine is now known as jalcine
[21:18] <desrt> seb128: got a working notify-osd patch for the new X property
[21:18] <seb128> desrt, nice!
[21:21] <dobey> why is the "File System" icon in my nautilus a paper icon, instead of a hard disk?
[21:27] <seb128> dobey, because you use a broken icon theme?
[21:27] <dobey> no
[21:27] <dobey> it's correct in the gtk file chooser
[21:28] <dobey> it's only broken in nautilus
[21:28] <dobey> and my windows partition has the right icon
[21:30] <dobey> and the larger icon onder computer:/// is also correct
[21:32] <seb128> dobey, ok, wfm
[21:33] <seb128> so dunno
[21:33] <dobey> hrmm
[21:34] <seb128> desrt, you probably need a configure.ac requirement update on gnome-desktop for _GNOME_BACKGROUND_REPRESENTATIVE_COLORS
[21:34] <seb128> desrt, or at least mention how it's defined and by what version in the merge request
[21:34] <desrt> seb128: the fallback case is just the same for if the unity schema was missing
[21:35] <dobey> i wonder how long it's been broken. i don't use nautilus much
[21:40] <seb128> ok, time to call it a week
[21:40] <seb128> good w.e everybody
[21:41] <dobey> cheers seb128