[00:42] <jasoncwarner_> FYI, RAOF and bryceh
[00:43] <jasoncwarner_> running noveau 3d now (since nvidia isn't working)
[00:43] <bryceh> jasoncwarner_, ok
[00:44] <jasoncwarner_> and, man, my system started running "hot". Don't know if that is related, but worth noting. Other things I've seen are laggy interfaces.
[00:45] <jasoncwarner_> not exactly scientific feedback, I know ;)
[00:45] <bryceh> :-)
[00:46] <bryceh> it's certainly conceivable that could be related to bad nouveau 3d support for the card
[00:48] <bryceh> jasoncwarner_, `ubuntu-bug xorg`.  I suspect RAOF will be interested in it.  Write up what you've seen.
[00:49] <bryceh> jasoncwarner_, might also check cpu load, just in case it's some unrelated runaway client program or something
[00:52] <jasoncwarner_> nothing going on in the CPU right now, though at one point the U1 client started to take 200% and then fell off...
[00:53] <jasoncwarner_> kenvandine: ^^
[00:53] <kenvandine> ugh
[00:53] <kenvandine> syncdaemon?
[00:53] <jasoncwarner_> yeah
[00:53] <jasoncwarner_> and I've had a TON of crashes on that, actually
[00:53] <kenvandine> haven't heard anyone else with that issue
[00:53] <kenvandine> file a bug please
[00:53] <jasoncwarner_> trying now...
[00:54] <bryceh> jasoncwarner_, oh, might need to manually install xdiagnose before reporting the X bug
[00:54] <bryceh> jasoncwarner_, sudo apt-get install xdiagnose ; ubuntu-bug xorg
[00:55] <jasoncwarner_> thanks...already installed ;)
[00:57] <jasoncwarner_> hey kenvandine, is there an easy way to get the things needed for Ubuntu One like 'ubuntu-bug ubuntu-one' or something?
[00:57] <kenvandine> ubuntu-bug ubuntuone-client
[00:57] <jasoncwarner_> FYI, I'm getting this message when I open U1 client: File Sync error. (org.freedesktop.DBus.Error.NoReply: Message did not receive a reply (timeout by message bus))
[00:57] <jasoncwarner_> thanks...filing now
[00:58] <kenvandine> wow
[00:58] <kenvandine> sounds very broken
[00:58] <bryceh> daemon not running or something?
[00:58] <kenvandine> you mean opening the control panel?
[00:58] <jasoncwarner_> yeah
[00:58] <jasoncwarner_> that message appears at top of control panel
[00:58] <kenvandine> syncdaemon must be crashing
[00:58] <kenvandine> do you have a crash file?
[00:58] <jasoncwarner_> where to look?
[00:58] <kenvandine> in /var/crash/?
[00:58] <kenvandine> for syncdaemon
[00:58] <jasoncwarner_> bryceh and RAOF, fyi #807305
[00:59] <jasoncwarner_> kenvandine: I have three
[00:59] <kenvandine> ubuntu-bug /var/crash/your.crash
[00:59] <jasoncwarner_> sorry..just one for U1
[00:59] <jasoncwarner_> others are for unity-panel-service and gnome-settings-daemon
[01:02] <jasoncwarner_> kenvandine: FYI. https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/807203 17 people effected by bug, it seems
[01:02] <ubot2> Ubuntu bug 807203 in zeitgeist "ubuntuone-syncdaemon crashed with AttributeError in __getattr__(): 'Symbol' object has no attribute 'PAGINATED_TEXT_DOCUMENT'" [High,Confirmed]
[01:02] <RAOF> jasoncwarner_: +1 to that.
[01:04] <kenvandine> ugh..
[01:05] <bryceh> jasoncwarner_, looks like you suspended during your current session; just to clarify did you notice the performance issue prior to the suspend, or did it only start after that?
[01:05] <bryceh> i.e. does it start right immediately on boot?
[01:06] <jasoncwarner_> bryceh: yeah, seems to happen at boot. I had a fussy baby and had to run upstairs and suspended while debugging baby
[01:06] <jasoncwarner_> but I did notice it before I suspended
[01:06] <bryceh> ok
[01:06] <jasoncwarner_> wow...those are some thorough logs ;)
[01:07] <bryceh> you see why we like to have them.  ;-)
[01:07] <bryceh> unfortunately, they don't seem to have captured anything of relevance to the performance issue
[01:07] <RAOF> If it starts immediately after boot then it's probably not compiz being leaky.
[01:08] <RAOF> It could concievably be a thermal issue; (sane) hardware will clock down when overheating, and that can make things slow.
[01:08] <bryceh> could there be leftover GL bits from the -nvidia install?
[01:09] <bryceh> jasoncwarner_, anything odd about the fan behavior?  like, is it not turned on when it should be?
[01:09] <jasoncwarner_> oh, yay! someone fixed the keyboard setup so I can now switch key refresh rates! Thank you rodrigo_, seb128 or whoever!
[01:09] <jasoncwarner_> bryceh: actually, fan seems fine.
[01:10] <jasoncwarner_> I might be sending you on redherring.
[01:10] <jasoncwarner_> ubuntuone-syncdaemon is actually going kinda crazy all of the sudden (just reopened system monitor)
[01:10] <jasoncwarner_> 200%
[01:10] <jasoncwarner_> 85%
[01:10] <jasoncwarner_> 150%
[01:10] <jasoncwarner_> 200%
[01:10] <bryceh> ah
[01:10] <jasoncwarner_> 0Q%
[01:10] <RAOF> Heh.
[01:11] <jasoncwarner_> it keeps going up and down.
[01:11] <bryceh> so my initial guess was right for a change :-)
[01:13] <jasoncwarner_> NM! sorry, bryceh and RAOF
[01:13] <jasoncwarner_> I always blame X
[01:13] <bryceh> heh
[01:13] <jasoncwarner_> stupid X keeps making my internets slow.
[01:13] <RAOF> As I always say: X is actually perfect.
[01:13] <jasoncwarner_> stupid X just raised my taxes
[01:13] <RAOF> Any apparent bugs can be handled with a quick and simple SEP field.
[01:13] <bryceh> you're in good company, we've been field lots and lots of non-X bugs lately ;-)
[01:14] <bryceh> guess it's either a good sign that X is more stable this time around, or a bad sign people are so suspicious of X!
[01:14] <jasoncwarner_> probably a bit of both...kind of how whenever something goes wrong in unity I start to ask didrocks about compiz ;)
[01:15] <jasoncwarner_> hey robert_ancell , btw...I let me kid play a cool video game this morning..., something called unity-greeter...seems perfect for a 3 year old!
[01:16] <robert_ancell> haha! Nice
[01:16] <bryceh> I tried to get Dutch to play with gcompris but it was not his cup of tea
[01:16] <bryceh> I need some simple app that just makes random colors and sounds when you hit different keys
[01:16] <bryceh> (and that disables all the special keys!)
[01:17] <RAOF> bryceh: Get him a multitouch screen and Chase's multitouch drawing demo!
[01:18] <jasoncwarner_> sounds like an iPad ;)
[01:18] <bryceh> RAOF, heh actually I did a while back however that computer got taken away
[01:20] <bryceh> RAOF, another guy had written that drawing program.  I was going to mod it a bit to make it into a better toy
[01:22] <TheMuso> Does anybody know if there is a tag we can mark bugs with that need translating?
[01:23]  * TheMuso asks in #ubuntu-bugs.
[04:51] <pitti> Good morning
[04:58] <cyphermox> pitti: good morning!
[04:58] <pitti> hey cyphermox, how are you?
[04:58] <cyphermox> pretty good
[04:59] <cyphermox> and you?
[04:59] <pitti> I'm great, thanks
[05:07] <cyphermox> late night upload of nm-applet, CHECK.
[05:08] <cyphermox> now I'm off to bed ;)
[05:08] <cyphermox> oh wait, are we still in soft freeze?
[05:08] <cyphermox> nm
[05:13] <pitti> cyphermox: no
[05:13] <pitti> cyphermox: it was lifted some 24 hours ago :)
[05:13] <cyphermox> yeah, I got confused by the day change now ;)
[05:25] <kenvandine> good morning pitti
[05:25] <cyphermox> hey kenvandine
[05:25] <pitti> hey kenvandine, how are you?
[05:25] <cyphermox> you're early!
[05:25] <kenvandine> and cyphermox :)
[05:25] <kenvandine> pitti, good
[05:25] <pitti> or, up late rather :)
[05:25] <cyphermox> ehehe
[05:26] <kenvandine> about to crash :)
[05:26] <cyphermox> ah, yeah
[05:26]  * kenvandine rewrote all the gwibber menus tonight... much cleaner now :)
[05:26] <kenvandine> but boring :/
[05:27] <kenvandine> good night all!
[05:27] <cyphermox> night
[05:27] <cyphermox> am leaving as well, see you later!
[05:29] <pitti> sleep well!
[06:42] <didrocks> good morning
[06:42] <pitti> bonjour didrocks
[06:43] <didrocks> guten morgen pitti! how are you?
[06:43] <pitti> quite fine, thanks! how about you?
[06:43]  * pitti in update/upload mood
[06:45] <didrocks> pitti: my flu has almost ended, which is nice for RMLL and 5h of train later today :)
[06:45] <pitti> oh, good! mine is gone as well
[06:45] <pitti> RMLL?
[06:46] <didrocks> a worldwide-french FLOSS event :) http://2011.rmll.info/
[06:46] <didrocks> I'm giving 3 conferences there (unity, how to integrate your application in unity, how to help in testing an unstable release)
[06:46] <didrocks> hum, apport keep spamming every 2s with u1 crashing there (or maybe it's update-notifier?)
[06:47] <didrocks> no, in fact u1 is really crashing. and it replaces the file in /var/crash/
[06:48] <pitti> didrocks: wow, good luck with that!
[06:48] <pitti> yes, it keeps crashing here as well :/
[06:49] <didrocks> sudo chmod -x /usr/lib/ubuntuone-client/ubuntuone-syncdaemon then :/
[06:50] <didrocks> indeed, your are in upload mood this morning ;)
[06:51] <pitti> {lib,}gnome-keyring coming up
[06:52] <didrocks> nice ;)
[06:52] <pitti> glib2.0 on amd64 seems unhappy, though :/ keeps timing out in the test suite
[06:52] <pitti> (on the builders)
[06:53] <didrocks> seif_: around?
[06:55] <didrocks> seems that zg is the guilty in u1 crashing
[07:06] <didrocks> ok, got a patch for u1
[07:09] <seif_> didrocks, can you show me the patch and the bug
[07:10] <seif_> didrocks, got it
[07:10] <seif_> the raptor 2.1 issue
[07:10] <seif_> i mean
[07:10] <didrocks> seif_: the patch is for u1 for now, on not crashing if zg is
[07:10] <seif_> raptopr2 issue
[07:10] <didrocks> seif_: right :)
[07:10] <seif_> didrocks, yeah
[07:10] <seif_> looking into it
[07:10] <didrocks> seif_: did you see my comment? https://bugs.launchpad.net/zeitgeist/+bug/807076/comments/2
[07:10] <ubot2> Ubuntu bug 807076 in zeitgeist "raptor2 not supported" [Low,In progress]
[07:11] <seif_> can i have any backlog
[07:12] <didrocks> seif_: https://launchpadlibrarian.net/74738481/Traceback.txt that one maybe?
[07:12] <cdbs-android> didrocks: Problems after problems, looks like I won't be able to work on the spec till next week, my 2-month old xps died again
[07:12] <seif_> didrocks, i am fixing this
[07:13] <didrocks> seif_: you rock! :)
[07:13] <didrocks> cdbs-android: argh, it's not your hardware time it seems :/
[07:13] <didrocks> seif_: take your time, I'm uploading the u1 fixes, and zg is still working for file/application search
[07:14] <cdbs-android> cyphermox: I won't be able to fix liboauth, next week probably. My laptop died
[07:15] <jasoncwarner_> hey didrocks...able to help me debug something?
[07:16] <jasoncwarner_> when I try to run nautilus, nothing happens...so when I run 'nautilus' from the command line, I get an error: Could not register the application: Timeout was reached
[07:16] <jasoncwarner_> not sure where to look for more information!
[07:17] <didrocks> jasoncwarner_: sounds like a gnome-session message
[07:17] <didrocks> jasoncwarner_: one sec
[07:17] <jasoncwarner_> didrocks: no worries...
[07:18] <didrocks> jasoncwarner_: finishing the u1 crash first, get back to you then :)
[07:19] <jasoncwarner_> didrocks: oh, please! that thing was wreaking havoc on my system all day!
[07:22] <chrisccoulson> good morning everyone
[07:22] <didrocks> hum, u1 never did any rebuilt with gcc 4.6? Fixing the errors…
[07:22] <didrocks> hey chrisccoulson
[07:25] <chrisccoulson> hey didrocks, how are you?
[07:26] <didrocks> chrisccoulson: racing against crashes ;) otherwise good, almost no more flu! :) and you?
[07:27] <chrisccoulson> didrocks, yeah, good thanks. just trying to finish the last few WI's on https://blueprints.launchpad.net/ubuntu/+spec/desktop-o-mozilla-rapid-release-maintenance
[07:28] <chrisccoulson> hmmm, does anyone have any idea what is going on here? https://launchpadlibrarian.net/74754356/buildlog_ubuntu-oneiric-i386.dehydra_0.9.hg20110312-0ubuntu2_FAILEDTOBUILD.txt.gz
[07:28] <chrisccoulson> i have no problem installing the packages locally or in a chroot :/
[07:30] <didrocks> chrisccoulson: maybe temporary inconsistency? did you try a give back?
[07:31] <chrisccoulson> didrocks, i don't normally like doing that without understanding what the original problem was :)
[07:33] <geser> chrisccoulson: have you tried installing all build-dependencies or only the packages from the error?
[07:33] <chrisccoulson> geser, all of them (and i can build it locally too)
[07:35] <pitti> hey chrisccoulson, good morning
[07:35] <pitti> hey jasoncwarner_
[07:35] <chrisccoulson> hey pitti, how are you?
[07:35] <jasoncwarner_> hey pitti, morning!
[07:36] <pitti> quite fine, thanks! back in Dresden
[07:38] <didrocks> dobey: you will have https://code.launchpad.net/~didrocks/ubuntuone-client/dont-suffer-zg-crash/+merge/67285 and https://code.launchpad.net/~didrocks/ubuntuone-client/use_result_var/+merge/67287 to review. It's already as distro patches in the ubuntuone-client package to fix the current oneiric crash
[07:39] <mvo> hey pitti! would you mind if I cherry pick the forward_search fix you did yesterday and upload it? I think update-manager is ready mostly now
[07:39] <geser> chrisccoulson: have you tried grabbing the buildd chroot and trying it there?
[07:39] <pitti> mvo: wohoo!
[07:39] <mvo> plus gconf->gsettings etc
[07:40] <geser> chrisccoulson: else try giving it back and see if it was a transient error or not
[07:40] <mvo> one thing I will need is get_xid() at some point, that is currently not introspectable, but its not critical at this point
[07:40] <pitti> mvo: it's in testing, so I'll commit/upload to Debian and sync
[07:40] <mvo> pitti: awsome, thanks
[07:41] <pitti> mvo: right, I didn't find a good solution for this one either so far
[07:41] <mvo> pitti: I will put in a version dependency and wait until its there
[07:41] <pitti> mvo: I'm almost done with gnome-keyring, doing pygobject next
[07:41] <mvo> is there something fundamental blocking to make get_xid() available?
[07:41] <didrocks> jasoncwarner_: is it exactly the message you get? It's not "Unable to register" rather?
[07:41] <jasoncwarner_> didrocks: yeah, copy and pasted from terminal
[07:42] <jasoncwarner_> nothing else was there
[07:42] <didrocks> jasoncwarner_: ok, maybe it's gtkapplication then, let me check
[07:42] <didrocks> not gnome-session, not nautilus anyway :)
[07:42] <pitti> mvo: I don't remember right now
[07:42] <pitti> mvo: might be because gdk_x11_window_get_xid() returns a Window datatype?
[07:42] <mvo> ok, no worries, not urgent
[07:42] <mvo> but its just a int, no?
[07:42] <pitti> but anyway, it should work in theory
[07:43]  * mvo nods
[07:43] <pitti> I suppose it stumbles over
[07:43] <pitti>           <type name="xlib.Window" c:type="Window"/>
[07:44] <pitti> and can't resolve this any further
[07:44] <mvo> yeah, that sounds very plausible
[07:44] <pitti> but maybe that can be annotated away
[07:44] <mvo> what will be the version of python-gobject with the fix? so that I don't forget to update my control file
[07:44] <chrisccoulson> geser, oh, i recreated it in pbuilder now
[07:45] <pitti> mvo: 2.28.6-2
[07:45] <mvo> ta
[07:45] <chrisccoulson> libgcc1: Breaks: gcc-4.5 (< 4.5.3-2) but 4.5.3-1ubuntu2 is to be installed.
[07:45] <chrisccoulson> the last gcc-4.6 upload added a breaks which makes gcc-4.5 uninstallable
[07:47] <didrocks> jasoncwarner_: hum, no gio, not gtk :/ do you have it at each nautilus start?
[07:48] <jasoncwarner_> yeah...it is really behaving weirdly...
[07:48] <jasoncwarner_> actually...let me reboot, see if I am having a weird update problem
[07:48] <jasoncwarner_> rebooting...
[07:50] <didrocks> jasoncwarner_: you should maybe install the ubuntuone-client update first ;)
[07:50] <didrocks> or do not let it start
[08:00] <bigon> could someone have a look at clutter-1.0? it needs a merge (or at least a rebuild) for the rename on gir1.2-json-glib-1.0 to gir1.2-json-1.0
[08:01] <chrisccoulson> huh, "gcc -print-file-name=plugin" doesn't work in gcc-4.6 :/
[08:04] <pitti> mvo: hm, seems something in glib/gio changed underneath, I have a test suite failure/hang now; need to investigate this first
[08:04] <mvo> ok, no worries
[08:06] <alex3f> hi mvo!
[08:06] <alex3f> I finally pushed version abstraction (as discussed in an old email) in my backend-refactor
[08:06] <alex3f> please have a look at it
[08:07] <alex3f> i'm also addressing the origin abstraction now
[08:07] <seif> didrocks, back
[08:08] <seif> didrocks, what other outstanding issues are there
[08:09] <seif> i can look into them
[08:09] <seif> (unity wise)
[08:09] <didrocks> seif: did the zg datahub not starting zg is fixed?
[08:09] <seif>  we have the solution
[08:09] <seif> we need mhr3 or kamstrup to work on it though
[08:10] <seif> its a flushing mechanism
[08:10] <seif> which means zeitgeist does not start until the first thing is requested
[08:10] <seif> if any1 is pushing on startup we wont care
[08:10] <seif> we will put it in some C process that will be flushed upon startup
[08:10] <seif> didrocks, can you report a bug on that please
[08:10] <seif> :)
[08:11] <didrocks> seif: sure doing
[08:11] <seif> so we can get more peopls attentnion
[08:11] <seif> didrocks, i am thinking of putting a 100$ bounty on that
[08:11] <seif> and requirments hsould be
[08:11] <seif> that
[08:11] <seif> no unity not zeitgeist person is to hack on it
[08:11] <seif> this way we can get new developers
[08:11] <seif> :)
[08:11] <didrocks> yeah, seems a smart way of getting new contributors for easy tasks :)
[08:12] <seif> exactly
[08:12] <seif> ahhhh njpatel is here
[08:12] <didrocks> that, and the indexing are the two big ones AFAIK
[08:12] <didrocks> but for the indexing, we need John
[08:12] <seif> didrocks, what kind of indexing
[08:12] <didrocks> for some design input
[08:12] <seif> ah that one
[08:12] <njpatel> What what?
[08:12] <seif> didrocks, well we have all the tools ready
[08:12] <didrocks> seif: well, I call it indexing, even if it's not :)
[08:12] <didrocks> seif: right, we need to see how to integrate it at best
[08:12] <seif> its just a design
[08:12] <seif> exactlx
[08:12] <seif> the script is there
[08:13] <seif> now we need to know how to run the scriopt or when
[08:13] <didrocks> exactly
[08:13] <seif> didrocks, we reduced memory consumption if you have not noticed with 0.8.1
[08:13] <seif> njpatel, did u have your coffe yet oh brown one :P
[08:13] <didrocks> seif: yeah, I checked after the update, and indeed, nice work!
[08:13] <seif> fts got updated too
[08:14] <seif> but needs to be released
[08:14] <njpatel> seif, coffee? Tea my friend :)
[08:15] <didrocks> seif: ok, no hurry though for it
[08:15] <njpatel> seif, also, I logged in an ZG crashed....fix your crap man! ;P
[08:15] <seif> njpatel, give me your log please
[08:15] <seif> its in .cache/zeitgeist
[08:15] <seif> :)
[08:15] <seif> unless you logged porn then please remember the title of the movie and delete those lines :P
[08:15] <njpatel> seif, you just want to see what, er, "special" files I have
[08:15] <didrocks> seif: you already had a bug btw: https://bugs.launchpad.net/zeitgeist-datahub/+bug/772265
[08:15] <ubot2> Ubuntu bug 772265 in zeitgeist-datahub "The datahub starts zeitgeist-daemon on startup" [Undecided,New]
[08:16] <njpatel> seif, I think apport caught it so i might just be able to provide a useful bug
[08:16] <seb128> hey didrocks pitti
[08:16] <seb128> hey everybody
[08:16]  * njpatel plays with apport
[08:16] <didrocks> salut seb128!
[08:16] <pitti> bonjour seb128, ca va?
[08:16] <njpatel> hey didrocks
[08:16] <didrocks> hey njpatel
[08:16] <seb128> thanks to whoever newer indicators
[08:16] <njpatel> hey pitti
[08:16] <seif> hi pitti
[08:16] <pitti> seif: de rien :)
[08:16] <seb128> pitti, ca va bien, et toi ?
[08:17] <seif> didrocks, if i fix the bug now you buy me 10 beers ok
[08:17] <seif> ?
[08:17] <pitti> hallo seif, wie gehts?
[08:17] <seif> pitti, ganz gut
[08:17] <pitti> seif: I'm great, thanks! in upload mood :)
[08:17] <mvo> hey seif!
[08:17] <didrocks> njpatel: already reported :)
[08:17] <seif> viel arbeit
[08:17] <didrocks> seif: 10! I'm sure I can trade with something else :)
[08:17] <seif> pitti, wie gehts ihnen
[08:17] <seif> :=)
[08:17] <njpatel> didrocks, awesome :)
[08:17] <seb128> pitti, thanks for the updates as well ;-)
[08:17] <seif> didrocks, your french you occupied egypt you owe me beer already
[08:17] <pitti> seif: prima, danke!
[08:17] <seif> neil can go live off british guilt
[08:18] <seif> :P
[08:18] <didrocks> seif: ahah, I think all europe owe a lot of beers then in various country ;)
[08:18] <seif> lol
[08:18] <seb128> desrt, glib hates the builder
[08:18] <seif> guys i am having a gvaraint issue here
[08:18] <seif> gnome-tweak-tools is not working
[08:18] <seb128> how not working?
[08:19] <seb128> pitti, did you retry the glib build earlier today?
[08:20] <pitti> seb128: yes
[08:20] <seb128> ok, trying again for the 4th time
[08:20] <pitti> seb128: it was a bit unclear what happened there
[08:20] <pitti> seb128: seems to hang eternally in a test case
[08:20] <seb128> it worked on the second try on i386
[08:20] <seb128> pitti, right, it's random, it hangs on different testcases :-(
[08:20] <pitti> ah, it finally timed out again
[08:21] <seb128> desrt and ricotz tried to debug it a bit during the rally by adding some printfs but it worked on that upload
[08:21] <seb128> "funny" but it worked without issue on armel and powerpc on the first build
[08:21] <seb128> usually that's the archs which work not so great :p
[08:21] <pitti> seb128: they are slow enough!
[08:25] <seb128> hum
[08:25] <seb128> session-service.c:592:11: error: format '%li' expects argument of type 'long int', but argument 4 has type 'gint64' [-Werror=format]
[08:26] <pitti> seb128: were you meaning to retry the build again?
[08:26] <pitti> want me to trigger it?
[08:26] <mvo> does anyone know whats up with gnome-tweak-tool? seems to be crashing snce a recent upgrade
[08:27] <seb128> mvo, stacktrace? but "don't use it"
[08:27] <mvo> "NotImplementedError: structure type 'GVariant' is not supported yet" <- error
[08:27] <mvo> why should I not use it?
[08:27] <seb128> pitti, ups, just did, thanks
[08:28] <seb128> mvo, because it's buggy :p
[08:28] <mvo> haha, I noticed that ;)
[08:28] <seb128> mvo, I blame pitti's gobject-introspection update
[08:29] <seb128> I doubt glib broke GVariant, must but a gir issue
[08:29] <pitti> mvo: that's the very thing I'm debugging right now
[08:29] <pitti> it breaks pygobject's test suite all over
[08:30] <pitti> (which holds me from uploading pygobject)
[08:30] <pitti> mvo: so, I'm on it
[08:30] <seb128> mvo, just curious but what settings do you tweak from gtt?
[08:30] <seb128> mvo, just to know if there are some we should bring on the default gcc in some way
[08:32] <mvo> seb128: I want to see what monospace font I have, it looks unfamiliar and check my theme, I use the light theme, but the indicators are really hard to see currently
[08:32] <mvo> seb128: so it looks like some mixup in the theme
[08:32] <seb128> they are blurry right?
[08:32] <seb128> the text at least is for me
[08:32] <seb128> mvo, dconf-editor is your friend :p
[08:33] <mvo> meh
[08:34] <pitti> ok, downgrading gir1.2-glib-2.0 to 0.10.8-1ubuntu1 fixes it
[08:34] <pitti> well, it makes it slightly better, anyway
[08:36] <seb128> didrocks, oh about gnome-session I assigned you the bug because I though you had a spec to re-enable the option for those who want to use it from the command line
[08:36] <seb128> didrocks, i.e letting the ui part off, ignored session with timestamps older than oneiric
[08:36] <didrocks> seb128: yeah, but I updated the spec to tell that finally it wasn't still working
[08:36] <seb128> ignoring
[08:37] <seb128> ok, I thought we were ok letting people shot themself in the foot from the command line
[08:37] <didrocks> seb128: and I think we don't want to spent more than a day for something which is half-broken anyway
[08:37] <seb128> since some clearly want to do it
[08:37] <didrocks> seb128: well, it's really show in the foot if they try to change from session
[08:37] <seb128> no, I would just have dropped the patch, but let the tab hidden
[08:37] <didrocks> required components are still register and conflicts
[08:38] <seb128> right but some are clearly asking for it
[08:38] <seb128> usually geeks that use custom sessions and don't switch
[08:38] <didrocks> so we will get bugs about "changed my session but unity still starts"
[08:38] <didrocks> or whatever
[08:38] <seb128> well, if those used the command line to store a session they are not basic users
[08:38] <didrocks> took me a lot of time previous cycles about all people getting this
[08:38] <seb128> but ok, I might have a look later in the cycle
[08:39] <didrocks> right, but that doesn't mean they will know what's the cause
[08:39] <seb128> right, but we had the option in the ui by then
[08:39] <didrocks> indeed
[08:39] <seb128> we had a tab with "click there to store your session"
[08:39] <seb128> which lusers clicked
[08:39] <didrocks> but still, we will ship a feature half-broken, just not visible
[08:39] <didrocks> not sure we want to do that
[08:39] <didrocks> you convinced me that this option was not useful :)
[08:39] <seb128> I've read enough comments on the internet of technical users being annoyed enough by it being dropped that they consider switching distro that I would like it back in some way
[08:40] <seb128> like a "--will-break-my-session" command line option
[08:40] <seb128> yeah, and you convinced me we should try to be nice to our tech users ;-)
[08:40] <didrocks> not sure that shipping something broken is what we want, but ok, can look at that :)
[08:41] <didrocks> oh, you reconsidered finally? ;)
[08:41] <seb128> ;-)
[08:41] <didrocks> because you were quite clear about this, I'm surprized you changed your opinion for that particular one :p
[08:41] <seb128> well anyway I was still on the impression that you said you would turn it on for oneiric but only from a command line at UDS
[08:41] <didrocks> but ok, in that case, I can revisit
[08:41] <seb128> that's why I assigned you the bug
[08:41] <seb128> didrocks, only idiots don't revisit their positions sometimes ;-)
[08:41] <didrocks> yeah, I changed the blueprint and talk here IIRC about it
[08:45] <pitti> seb128: seems to hang again :/
[08:52] <seb128> pitti, :-(
[08:52] <didrocks> seb128: so, thinking about it, I think that people rather want to "start applications reliably at each boot" (they won't spaw the command line each time they want to register the session). So basically, the application list at startup without system apps can do it
[08:52] <seb128> pitti, I'm pondering turning off the testsuit and do a new upload
[08:53] <didrocks> seb128: we probably need to get compiz and this ui let you "start app on ws <number>"
[08:53] <seb128> didrocks, right
[08:53] <didrocks> at least, we will have a more generic feature which will work with every apps
[08:53] <seb128> we have a busy cycle though so let's keep thinking a bit about it and see next cycle
[08:53] <didrocks> seb128: yeah, my tomboy won a new note :-)
[08:54] <seb128> ;-)
[09:07] <didrocks> chrisccoulson: can I bother you with a thunderbird issue? ;)
[09:08] <chrisccoulson> didrocks, sure. but i'm pretty busy already ;)
[09:08] <didrocks> chrisccoulson: I get the impression that some folders are only refreshed if I click on them.
[09:09] <chrisccoulson> didrocks, i think that might be the case, visually
[09:09] <didrocks> chrisccoulson: like, I have my folder with ubuntu-devel ML content, my thunderbird was opened since this morning, no "new number "showing
[09:09] <didrocks> oh, already known then?
[09:09] <chrisccoulson> oh, you mean the actual number is not refreshed?
[09:09] <chrisccoulson> that sounds like a bug
[09:09] <didrocks> chrisccoulson: well, there is no number at all
[09:09] <didrocks> but when I click on it
[09:10] <didrocks> I get, like this time "5", with the 5 emails
[09:10] <chrisccoulson> what type of account is this?
[09:10] <didrocks> but the icon shows that number only when I click on the folder
[09:10] <didrocks> imap
[09:10] <didrocks> and it's not the case for every folders in that account
[09:10] <didrocks> but seems to always be on the same ones
[09:11] <chrisccoulson> do you get entries in the messaging indicator for those folders before you click on them?
[09:12] <didrocks> chrisccoulson: hum, I didn't try to notice, I'll for next time
[09:12] <chrisccoulson> yeah, it would be useful
[09:12] <chrisccoulson> that will tell us if it is just a visual problem with the folder pane or not
[09:13] <didrocks> chrisccoulson: ok, that would be the first step then :)
[09:13] <didrocks> I'll tell you
[09:37] <desrt> seb128: i hate the builders.
[09:37] <seb128> desrt, they hate you back
[09:37] <desrt> actually, it seems that they hate ricotz :)
[09:37] <seb128> desrt, https://launchpad.net/ubuntu/+source/glib2.0/2.29.10-0ubuntu1/+build/2613839
[09:38] <seb128> desrt, well now they hate pitti
[09:38] <seb128> 2.29.10 doesn't want to build on amd64
[09:38] <seb128> it built on i386 on the second retry though so I'm happy :p
[09:38] <desrt> glib is having a rough ride lately
[09:39] <desrt> i should probably pay more attention to actively attempting to maintain it
[09:39] <seb128> well I didn't notice particular issue out of that testsuit hang
[09:40] <seb128> seems to be hanging on gdbus tests, not the same one each time though
[09:40] <desrt> we had some internal issues lately
[09:40] <desrt> it has nothing to do with gdbus, i think
[09:40] <desrt> the real failure is that we don't unit-test inter-thread dispatches of gmainloop
[09:40] <desrt> and gdbus does that
[09:40] <desrt> so it's the first test that sees the problem
[09:41] <desrt> i think it's very likely related to the eventfd work
[09:41] <desrt> colin probably made some weird mistake somewhere
[09:41] <desrt> and introduced a race
[09:42] <desrt> seb128: an interesting possibility would be to try to use a vendor patch to revert his changes and see if it improves the situation
[09:42] <desrt> seb128: if i'm right then it would be a relatively 'cheap' solution for you since you could do it right now, and quite easily
[09:43] <desrt> and if it works out nicely then it points a finger directly at the source of the problem, which is useful feedback for us
[09:43] <rodrigo_> is ubuntu going to switch to systemd anytime soon?
[09:43] <desrt> rodrigo_: rumour has it that it will happen after the LTS release
[09:43] <seb128> rodrigo_, no
[09:43] <seb128> not "any time soon"
[09:43] <pitti> rodrigo_: right now it seems the foundations team wants to stick with upstart
[09:43] <seb128> like not before a year
[09:43] <seb128> if we switch
[09:44] <rodrigo_> ok, then not sure what to do with the system-wide setting stuff in g-c-c
[09:44] <rodrigo_> I guess we could implement the DBus interfaces of systemd in ubuntu-system-service
[09:44] <pitti> rodrigo_: can't it use ubuntu-system-service?
[09:44] <seb128> did GNOME accept the systemd depends?
[09:44] <pitti> right, I thought that was the idea
[09:44] <desrt> seb128: i don't think that we did.
[09:44] <seb128> the discussion stopped on the list without consensus
[09:45] <desrt> seb128: for the record, i added a sort of systemd dependency to dconf recently, in git
[09:45] <rodrigo_> pitti, yes, but for that we'd need to create a distro patch, I am trying to make it work for both ubuntu and upstream
[09:45] <seb128> desrt, the xdg run dir thing?
[09:45] <pitti> rodrigo_: we could change ubuntu-system-service to have the same dbus names/API
[09:45] <desrt> seb128: yes.  exactly.
[09:45] <pitti> rodrigo_: nothing else uses it, after all
[09:45] <rodrigo_> seb128, the consensus was to define dbus interfaces that anyone could implement
[09:45] <desrt> it's a standard....
[09:45] <seb128> desrt, that should be easy to get on !systemd as well
[09:45] <rodrigo_> pitti, yes, that's what I'm thinking
[09:45] <desrt> seb128: indeed.
[09:46] <desrt> seb128: and i use it through glib which, itself, has a reasonable fallback
[09:46] <desrt> (~/.cache/)
[09:46] <seb128> desrt, great
[09:46] <pitti> mvo, seb128: gobject-introspection hit over the head, and uploaded; I now know what's going on, but the real solution will still take some time; but this hack should do
[09:46] <desrt> but dconf will break on NFS systems now if systemd is not used
[09:46] <seb128> pitti, ok
[09:46] <desrt> since ~/.cache/ is a particularly bad spot in case of NFS
[09:47] <desrt> someone like pitti should come up with a solution to this problem :)
[09:47] <seb128> desrt, could you open an upstart bug about xdg_run_dir? or do you want me to do it?
[09:47] <pitti> moving to systemd? *cough*
[09:47] <desrt> seb128: you should do it, i guess
[09:47] <rodrigo_> another problem is the packagekit thing, as we don't want in g-c-c upstream code to deal with different package systems, all that should be lower in the stack
[09:47] <seb128> desrt, ok, will do and then I will ping foundation guys saying we need it this cycle if possible
[09:47] <rodrigo_> so is it possible to have packagekit on the CD?
[09:47] <seb128> rodrigo_, no
[09:47] <pitti> rodrigo_: no
[09:47] <desrt> annoyingly, this is all rather related to the question of dbus lifecycle
[09:48] <seb128> rodrigo_, we have sessioninstaller, it works with i.e file-roller
[09:48] <rodrigo_> then distro patch then :(
[09:48] <pitti> rodrigo_: PK doesn't work for us as a general package installer; we have aptdaemon
[09:48] <rodrigo_> sessioninstaller?
[09:48] <desrt> seb128: link them the xdg spec.  should help to convince them.
[09:48] <seb128> desrt, right
[09:48] <pitti> rodrigo_: sessioninstaller is the user session component of it, yes
[09:48] <seb128> rodrigo_, apt-cache show sessioninstaller
[09:48] <rodrigo_> pitti, right, just thinking about the API, which is more and more used in gnome
[09:48] <pitti> rodrigo_: sessioninstaller even implements PK's API in terms of aptdaemon
[09:49] <pitti> oh, where did he go
[09:49] <seb128> rodrigo_, don't run away! ;-)
[09:49] <pitti> rodrigo_: wb
[09:49] <pitti> rodrigo_: sessioninstaller even implements PK's API in terms of aptdaemon
[09:49] <pitti> rodrigo_: and we install it by default
[09:49] <rodrigo_> sorry, xchat crashed
[09:49] <pitti> rodrigo_: so ideally it would "just work"
[09:49] <rodrigo_> pitti, ah, ok that's good
[09:50] <desrt> seb128: will you try the eventfd revert vendor patch today?
[09:50] <desrt> seb128: because i intend to roll a glib release of my own today with a different set of vendor patches....
[09:50] <seb128> desrt, yes, I just finish what I do, then it's next
[09:50] <desrt> okay.  awesome
[09:50] <desrt> i will wait for you, then :)
[09:50] <seb128> k ;-)
[09:50] <rodrigo_> pitti, ok, so apps can safely use packagekit api on a default ubuntu install?
[09:50] <seb128> rodrigo_, that should work, test it though
[09:50] <pitti> rodrigo_: only the session bus
[09:51] <seb128> rodrigo_, but it works for i.e file-roller
[09:51] <pitti> rodrigo_: sessioninstaller doesn't implement the system bus PK API
[09:51] <rodrigo_> I had to install packagekit for the info panel in g-c-c to detect upgrades
[09:51] <rodrigo_> pitti, ah, it's got the same API but on the session bus?
[09:51] <pitti> rodrigo_: no
[09:51] <pitti> rodrigo_: it's got the same session bus API as PK
[09:51] <chrisccoulson> b'ah, my internet connection has gone down for the second day in a row
[09:52] <rodrigo_> ah
[09:52] <pitti> rodrigo_: but it doesn't have PK's system bus API; that's aptdaemon, and has a different API due to how apt works
[09:53] <pitti> rodrigo_: but the session bus API doesn't do stuff like "updates available", only querying and installing packages
[09:53] <pitti> rodrigo_: for "updates available" it should use aptdaemon
[09:54] <rodrigo_> pitti, right, but then we need a distro-only patch :(
[09:55] <glatzor> pitti, rodrigo_ aptdaemon doesn't provide a query/search interface in contrast to packagekit. by design it only takes care of the actions which modify the system and require higher privileges
[09:55] <pitti> rodrigo_: yes
[09:55] <pitti> rodrigo_: packaging is inherentl distro specific :/
[09:55] <pitti> rodrigo_: there were some discussions in the past to make PK work properly for apt, but it was found to be impractical with the current design
[09:56] <rodrigo_> pitti, well, since packagekit system bus is being used more and more, we'll have to carry quite big patches
[09:56] <glatzor> pitti, rodrigo_ so there isn't any "updates available" method in aptdaemon
[09:56] <rodrigo_> yeah
[09:56] <glatzor> rodrigo_, "normal" applications should not use the system bus directly but make use of the session bus
[09:57] <glatzor> rodrigo_, the system bus interface is designed for package managers and system update managers
[09:57] <rodrigo_> ok
[09:57] <pitti> rodrigo_: might be easier to disable the package updates part in g-c-c? we alreayd have update-notifier/update-manager for that anyway, with notifications
[09:58] <rodrigo_> pitti, yeah, it would just not work by default
[09:58] <rodrigo_> unless packagekit is installed
[09:58] <pitti> rodrigo_: the button/label could be hidden if PK is not available? that should even be upstreamable
[09:58] <rodrigo_> ok, but if upstream already uses the session bus interface, that would solve the problem for other cases, like installing languages
[09:59] <pitti> for installing languages, sessioninstaller should work fine
[09:59] <rodrigo_> pitti, yeah, the button is already hidden, only shows up if it gets an answer from packagekit
[09:59] <pitti> rodrigo_: ah, sounds perfect then -- we don't need to change anything then?
[09:59] <rodrigo_> pitti, not there, no
[10:00] <rodrigo_> and for installing languages, if the session bus interface is enough, it should be ok also
[10:02] <didrocks> ok, time to take the train. I'll try to hop on IRC next week if the internet connexion is good enough, otherwise, I should be reachable by emails
[10:03] <glatzor> rodrigo_, perhaps you need to a new resource to InstallResource method of the session API
[10:03] <rodrigo_> didrocks, going on vacation?
[10:03] <didrocks> rodrigo_: no, at RMLL (french floss event. at the ubuntu booth and doing some talks)
[10:03] <glatzor> rodrigo_, InstallResource("language", "German") to be really distro neutral on the g-c-c side
[10:03] <rodrigo_> glatzor, yeah, I was hoping InstallProvideFiles would be enough
[10:03] <rodrigo_> glatzor, yeah, sounds good
[10:04] <didrocks> see you guys :)
[10:05] <glatzor> rodrigo_, The InstallProvideFiles doesn't yet work for Debian/Ubuntu properly. It would require to add support for remote apt-file
[10:11] <glatzor> rodrigo_, the aptcc backend of packagekit also doesn't support the SearchFiles method with uninstalled files
[10:11] <rodrigo_> glatzor, ok, I'm discussing with upstream your idea of InstallResource
[10:11] <glatzor> rodrigo_, so InstallProvideFiles won't work too with packagekit on Debian/Ubuntu
[10:12] <seb128> desrt, do you have a diff handy of what a revert of the eventfd change would be? ;-) I guess it includes several commits
[10:12] <desrt> seb128: i guess so too.  i'll prepare one for you.
[10:12] <seb128> desrt, thanks
[10:13] <desrt> seb128: step 1: revert initial patch
[10:13] <desrt> step 2: notice conflicts
[10:13] <desrt> step 3: resolve.  repeat.
[10:13] <desrt> :)
[10:13] <seb128> desrt, you said you will prepare one I decide to stay on that :p
[10:14] <desrt> step 1: ask desrt
[10:14] <seb128> \o/
[10:14] <desrt> :)
[10:14] <mvo> rodrigo_: what is InstallFile used for or how? for the language support? we have InstallFile for bnaries (where we can us ethe command-not-found data)
[10:14] <rodrigo_> mvo, not being used yet, was just thinking how it could be done upstream that would work on ubuntu
[10:15] <rodrigo_> mvo, sending a message now to the PK list to see if the InstallResource idea works for everyone
[10:15] <rodrigo_> and then we'll have to provide that method in sessioninstaller
[10:15] <mvo> aha, ok. I'm keen on learning how InstallFile is used, we can not make it universally work without the need t odownload the *huge* Contents-$arch file, but for specific cases like binaries we can
[10:15] <mvo> great, thanks rodrigo_
[10:15] <mvo> thanks thanks glatzor
[10:15] <rodrigo_> :)
[10:15] <glatzor> hey mvo!
[10:16] <glatzor> mvo, sorry for my huge disappear in the last weeks
[10:16] <mvo> glatzor: I got a request about a "UpdatesAvailable" method over dbus, could we (ab)use the sessioninstaller bus for this?
[10:16] <mvo> glatzor: no worries
[10:16] <mvo> glatzor: real-life is also important :)
[10:17] <mvo> rodrigo_: hm, if we had UpdatesAvailable on the sessionbus that could be used by gnome sa well, right? I assume its currently using the systembus only because there is no such method on the session bus?
[10:17] <glatzor> mvo, you want it to be distro neutral?
[10:17] <mvo> glatzor: ideally
[10:17] <mvo> glatzor: the plan is to have it in unity in a indicator
[10:18] <glatzor> packagekit was written originally to have an universal update application on all distros
[10:18] <mvo> glatzor: and if gnome is also using something like this (from what I read earlier from rodrigo_) it sounds like there is a API there :)
[10:18] <mvo> right, I'm thinking of a more lightweight way that can be used by e.g. a panel indicator
[10:19] <mvo> enough so that it can display a icon if updates are available but not the full thing with all details
[10:19] <rodrigo_> yes, not sure why some operations are on the session and others on the system bus
[10:20] <glatzor> rodrigo_, because the session bus provides a complete ui for handling package installations
[10:20] <rodrigo_> ah
[10:20] <glatzor> you just call InstallPackage xterm and all the download, progress and error handling will be done for you
[10:21] <glatzor> rodrigo_, you can even modify the to be shown dialogs and error messages. but the later one isn't supported by sessioninstaller
[10:22] <desrt> seb128: think i have a patch.  doing a local test build first.
[10:23] <glatzor> rodrigo_, furthermore the system bus is quite complex. it is transaction based and requires to "perform a transaction dance" - but this is nowadays covered by the pk_tasks methods
[10:24] <seb128> desrt, thanks
[10:25] <rodrigo_> glatzor, ok, so the system bus is for package managers and the session one for desktop apps
[10:25] <glatzor> rodrigo_, right
[10:28] <desrt> seb128: http://fpaste.org/98Fa/ should do the trick
[10:28] <seb128> desrt, thanks
[10:28] <pitti> rodrigo_: I see that the g-c-c task of bug 787694 is fixed; what needs to happen in jockey, scp, and friends now to make them appear again? (the other tasks)
[10:28] <ubot2> Launchpad bug 787694 in gnome-control-center "GNOME 3 System Settings: Provide stub icons for launching Ubuntu specific capplets" [High,Fix released] https://launchpad.net/bugs/787694
[10:29] <rodrigo_> pitti, working on them
[10:29] <pitti> rodrigo_: apparently having a simple "Settings" Category in the .desktop isn't sufficient any more?
[10:29] <rodrigo_> pitti, no, we need to provide a .so implementing the ExternalPanel class
[10:29] <pitti> rodrigo_: do we need g-c-c patches to display the ones we want, or can this be done in the packages themselves?
[10:29] <mvo> so … gtk3 update-manager is hitting the archive soon, please let me know if everything explodes
[10:29] <desrt> man.  glib testcases are getting heavy these days
[10:30] <desrt> either that or this laptop is very slow :)
[10:30] <rodrigo_> pitti, in the packages themselves, or in g-c-c, since some of those packages use python's distutils and I'm not sure how to integrate a C build there
[10:30] <seb128> mvo, \o/
[10:30] <pitti> desrt: I wouldn't call my laptop "slow", and it's still taking ages
[10:30] <rodrigo_> pitti, but then we'll have to depend on those packages, so I guess it's better in each package
[10:30] <pitti> rodrigo_: right, that would have been my next question
[10:30] <chrisccoulson> pitti, https://blueprints.launchpad.net/ubuntu/+spec/desktop-o-mozilla-rapid-release-maintenance is so close now! i might be asking you to do some package removals later ;)
[10:30] <desrt> this spawn-sync test is deadly
[10:30] <pitti> chrisccoulson: yay
[10:30] <seb128> desrt, it takes longer to run the testsuite than to do the 3 builds the package is doing
[10:31] <seb128> desrt, I tend to disable the testsuite locally when I've to do rebuilds because of .symbols changes or something ;-)
[10:31] <pitti> mvo: https://launchpad.net/ubuntu/+source/pygobject/2.28.6-2 FYI
[10:31] <desrt> seb128: at the same time it's clearly a good idea that we have the tests to catch problems like the one we have now, for example
[10:31] <pitti> rodrigo_: i. e. we need to change all these packages to be arch:any and add a .c file for integrating into g-c-c?
[10:32] <seb128> desrt, oh, don't get me wrong, I'm really in favor of the testsuite ;-) it just takes ages!
[10:32] <pitti> rodrigo_: as I suppose all these .c files would be pretty much identical, would it be possible to have a generic one which takes a .desktop file and then reads the name, icon, and Exec= from that?
[10:32] <desrt> i'm starting to wonder if this test has livelocked
[10:32] <mvo> pitti: \o/
[10:32] <desrt> it's spinning at 100% CPU for 3 minutes now
[10:34] <desrt> hm.  this is a new test.  i wonder if it has (or has found) a bug.
[10:36] <pitti> hm, the amd64 build of gnome-keyring is also stuck in the test suite
[10:36] <pitti> might be the same glib bug?
[10:36] <mpt> mvo, glatzor, good morning. Random question for you: What is the use case for URLs of the form "apt:package-name?minver=1.0"?
[10:38] <seb128> desrt, bah, I'm having stupid browser issues with pastebin, any change you can email me the patch?
[10:39] <pitti> oh?
[10:39] <pitti> seb128: http://fpaste.org/98Fa/raw/ ?
[10:39] <seb128> pitti, oh, stupid me...
[10:39] <seb128> pitti, thanks ;-)
[10:40]  * pitti hugs seb128
[10:40]  * pitti hugs desrt, too
[10:40]  * seb128 hugs pitti
[10:42] <seb128> pitti, you wouldn't want to apply that to current glib and upload to a ppa by any chance to see if that workaround the hang issue? ;-)
[10:42] <seb128> pitti, I will do it in a bit otherwise, I want to clean the indicators updates and unity rebuild first
[10:42] <seb128> they need to transition in once we are half broken atm
[10:42] <pitti> seb128: yep, can do
[10:42] <seb128> pitti, feel free to say no, will do glib in a bit if nobody else does
[10:43] <pitti> seb128: can do it in parallel with my release team update, that's fine
[10:43] <seb128> danke
[10:43] <pitti> I'm also eager to get keyring, etc. built
[10:45] <desrt> y'all could give me upload rights
[10:45]  * desrt applies a 'vendor patch' to unity to replace it with gnome-shell
[10:46] <pitti> and then see how long it takes people to tell the difference? :-)
[10:46] <desrt> pitti: ;)
[10:47] <pitti> seb128: uploaded to my PPA; knocking on wood
[10:47] <seb128> pitti, danke
[10:47]  * desrt elects to read some manpages and look at the code in question for obvious errors
[10:52]  * desrt finds at least one more bug
[10:52] <desrt> seb128: what is the test case that was failing?
[10:53] <seb128> desrt, I don't see it's one testcase
[10:53] <desrt> there's a bug in the GCancellable code
[10:53] <seb128> desrt, I've seen hangs on 3 different gdbus tests
[10:53] <desrt> same as the bug i found in the mainloop
[10:53] <desrt> checking for ENOSYS but not EINVAL
[10:53] <desrt> i'll file a new upstream bug about that one
[10:53] <seb128> thanks
[10:54] <desrt> would be nice to know if the tests that it hangs on involve using GCancellable or not
[10:56] <seb128> desrt, current build: https://launchpad.net/ubuntu/+source/glib2.0/2.29.10-0ubuntu1/+build/2613839
[10:56] <seb128> desrt, (it's the stock 2.29.10 archive one, not the patched one)
[10:57] <seb128> it's sitting there
[10:57] <desrt> seb128: do you have superpowers to get your builds pushed to the front of the line or something? :)
[10:57] <seb128> pitti does
[10:57] <desrt> nice
[10:57] <desrt> so i assume this means that the test case that follows 'signals' is hanging
[10:58] <desrt> that's "filter"
[10:58] <desrt> which, iirc, was one of the ones causing trouble before
[10:58] <seb128> I think that's one it did hit yesterday as well
[10:59] <pitti> bah, not my day -- the pygobject build hangs as well
[10:59] <seb128> :-(
[10:59]  * desrt fears the prospect of bisecting a month or two of history using builders to find a non-deterministic bug
[10:59] <desrt> this is some kind of nightmare
[11:00] <desrt> has anyone ever seen the hang outside of a builder?
[11:00] <seb128> no
[11:00] <seb128> but I did only one build
[11:00] <pitti> I didn't
[11:00] <desrt> and you're absolutely sure my patch is applied for this build?
[11:00] <seb128> desrt, which one?
[11:00] <desrt> the one to remove eventfd
[11:00] <seb128>  desrt, (it's the stock 2.29.10 archive one, not the patched one)
[11:01] <desrt> oh
[11:01] <desrt> okay.  please try with my patch, then :)
[11:01] <pitti> desrt: https://launchpad.net/~pitti/+archive/ppa/+build/2615620 and https://launchpad.net/~pitti/+archive/ppa/+build/2615619 have your patch applied
[11:01] <pitti> desrt: but they aren't at the test suite yet, I just uploaded it a couple of minutes ago
[11:01] <seb128> desrt, the patched version didn't reach testsuit yet
[11:01] <desrt> okay.  great.
[11:01] <desrt> we'll wait for those, then
[11:01] <desrt> i thought the one that seb showed me was with my patch
[11:01] <seb128> desrt, but you asked what case were causing issues
[11:01] <seb128> that's why I showed you the log
[11:01] <desrt> ahhhhh.  thanks :)
[11:01] <desrt> okay
[11:01] <desrt> because if my patch fails, then we have some serious trouble
[11:01] <desrt> i don't even know where to start
[11:04] <pitti> it seems we kill buildds by the dozens today :/ I asked infinity/lamont to kill these
[11:04] <pitti> (glib, g-keyring, pygobject)
[11:04] <seb128> well at least they will stop those over 3 hours
[11:04] <seb128> but yeah, still
[11:05] <desrt> did we have a post-eventfd build of glib pass the tests (by chance?) and get into the archive?
[11:05] <desrt> if so, it could be responsible for all of the failures
[11:08] <mvo> mpt: I guess if you blog about foo and 1.0 is not good, but 2.0 is awsome to ensure that people really get 2.0 and will not be disappointed about 1.0 (if 1.0 is the only version availalble for them)
[11:09] <chrisccoulson> i'm not the only one hanging buildd's now then? ;)
[11:12] <mpt> mvo, fair enough, thanks
[11:13] <desrt> builders are slow :)
[11:13] <desrt> s/slow/busy/, probably
[11:14] <pitti> oh, https://launchpad.net/~pitti/+archive/ppa/+build/2615620 is at the test suite now
[11:15] <desrt> ya.  i see that.
[11:15]  * desrt waits for it to get to the good part
[11:15] <desrt> this is like watching a hockey game :)
[11:16] <ogra_> now you know how exciting the work on arm is :)
[11:17] <desrt> ogra_: that's more like watching a hockey game in slow motion
[11:17] <pitti> "and now the test case of the blue team is going to attack the dbus, aims, and ... PAAAAAAAAAAAASS!"
[11:17] <ogra_> *g*
[11:18] <desrt> actually
[11:18] <desrt> PASS: atomic-test
[11:18] <desrt> PASS: bit-test
[11:18] <desrt> PASS: cxx-test
[11:18] <desrt> PASS: child-test
[11:18] <desrt> PASS: completion-test
[11:18] <desrt> PASS: dirname-test
[11:18] <desrt> this part sounds like the announcer
[11:18] <desrt> "atomic passes to bit.  bit passes to cxx.  cxx to childs..."
[11:20] <desrt> filter test passed...
[11:21] <desrt> all of the gdbus and gapplication tests are passed
[11:21] <desrt> gio/tests/ is done
[11:21] <desrt> so either we just got very lucky or the patch fixes the problem (indicating that the problem is, indeed, eventfd)
[11:26] <mpt> mvo, if you have a section like Multiverse disabled, could USC still easily tell whether a given package is present in that section? Or would you need to enable it first?
[11:27] <pitti> I need to disappear for a few hours, see you back on the release meeting
[11:27] <pitti> seb128: ^
[11:27] <desrt> pitti: thanks for the builds
[11:27] <pitti> seb128: lamont is killing the hanging builds
[11:28] <pitti> seb128: can you check my PPA build, and if it works, upload the ubuntu bzr?
[11:29] <desrt> amd64 just passed the testsuite too
[11:32] <pitti> yay
[11:36] <mvo> mpt: for application it knows that currently because we have the app-instlal-data locally. in the future when the metadata moves to the server it may not be possilbe anymore
[11:36] <mpt> ok, thanks mvo
[11:36] <mvo> yw
[11:37] <desrt> seb128: i'd do the 'official' upload with my patch for now
[11:37] <desrt> seb128: i'll take the issue up with colin when he's awake
[11:40] <seb128> pitti, ok
[11:40] <seb128> desrt, thanks
[12:00] <cyphermox> good morning!
[12:34] <chrisccoulson> hi Laney, do you have any ideas about bug 798941? i think your mono knowledge is probably much greater than mine ;)
[12:34] <ubot2> Launchpad bug 798941 in moon "moon FTBFS in oneiric with libav 0.7" [Medium,Triaged] https://launchpad.net/bugs/798941
[12:34] <chrisccoulson> (ignore the title btw, there is a patch in the bug for that specific issue already)
[12:36] <Sweetshark> pitti: so how can I restart https://launchpad.net/~libreoffice/+archive/ppa/+sourcepub/1807234/+listing-archive-extra now that it has been aborted?
[12:36] <Sweetshark> pitti: ah, found it!
[13:01] <chrisccoulson> sigh, wth has happened to my connection
[13:12] <kenvandine> seb128, thx for helping with those merges
[13:12] <kenvandine> are you doing datetime already?
[13:12] <kenvandine> if not i can do it now
[13:12] <seb128> hey kenvandine
[13:12] <seb128> kenvandine, no, I'm on unity
[13:12] <kenvandine> ok
[13:12] <seb128> if you can do datetime that would be nice ;-)
[13:12] <seb128> yw btw ;-)
[13:13] <kenvandine> :)
[13:14] <kenvandine> awesome, i see didrocks fixed that nasty u1 crasher
[13:14] <kenvandine> woot
[13:14] <kenvandine> this morning is looking up already :)
[13:15] <mvo> eh, silly question (what is the best channel to ask such questions?). but in the gtk3 world a "get_allocation() call return GdkRectangle" which no longer provides width, height nor methods to access them. how do I get the allocation height in e.g. size-allocated callbacks?
[13:16] <seb128> mvo, you can try #gnome-hackers or #gtk+ on irc.gnome.org I guess
[13:16] <kenvandine> mvo, i think i have an example... one sec
[13:19] <kenvandine> mvo, the GdkRectangle has width and height properties
[13:20] <mvo> aha, let me try that
[13:25] <mvo> kenvandine: do you have a example snippet? I can not get it to work
[13:26] <kenvandine> only in vala
[13:26] <kenvandine> look at lp:gwibber
[13:26] <kenvandine> in client/tab-bar.vala
[13:27] <m_conley> rodrigo_: ping
[13:29] <dobey> rodrigo_: ping
[13:31] <kiwinote> mvo: I use on_size_allocate(self, widget, allocation): print widget.get_allocation().width which seems to work, I haven't yet worked out how to use the allocation directly yet
[13:35] <mvo> kiwinote: oh, with gtk3? hmmm, makes me wonder what I'm doing wrong here. do you hvae example code?
[13:35] <kiwinote> mvo: let me quickly put a sample together
[13:36] <seb128> Laney, there?
[13:38] <kiwinote> mvo: http://paste.ubuntu.com/640136/
[13:39] <mvo> kiwinote: oh, you use the widget.get_allocation() instead of the allocation one? clever :)
[13:39] <kiwinote> mvo: yeah, the allocation one seems to be broken :/
[13:40] <rodrigo_> m_conley, dobey: pong
[13:40] <mvo> thanks kiwinote! that helped me and the coding can continue \o/
[13:41] <kiwinote> ;)
[13:42] <dobey> rodrigo_: hey, how is it that evo-couchdb is compiling in oneiric? i was looking at the create_u1_addressbook patch last night, and it's exactly the same as we have in the nightlies packages, but it always fails to apply
[13:43] <rodrigo_> dobey, hmm, it shouldn't build, unless it's using an older release
[13:43] <rodrigo_> dobey, but thanks for the reminder, I'll rebase the patch now
[13:43] <dobey> rodrigo_: so in oneiric proper, it is FTBFS right now?
[13:44] <rodrigo_> hmm, no, haven't uploaded the newer version iirc
[13:44]  * rodrigo_ checks
[13:44] <dobey> rodrigo_: lp says 0.5.90 is built
[13:44] <dobey> rodrigo_: or was the big change that breaks the patch, after that release?
[13:46] <Laney> chrisccoulson: we're going to remove moon, don't worry about that
[13:46] <Laney> ping directhex about it
[13:46] <Laney> seb128: yeah a bit
[13:47] <seb128> Laney, that's ok, I figured it out
[13:47] <seb128> Laney, got some build failure due to mono
[13:47] <Laney> ah ok, yeah that's the transition
[13:47] <Laney> if it's stuff not fixed in debian then please send fixes to us too
[13:47] <seb128> but it seems libunit-cil has some binaries in universe
[13:48] <seb128> pitti, ^ do you have any clue about that? is that an error?
[13:48] <Laney> it got a new abi
[13:48] <Laney> should have been newed to main probably
[13:48] <rodrigo_> dobey, yes, was after that change
[13:49] <rodrigo_> dobey, not released yet
[13:49] <seb128> tedg, kenvandine: but I commited a one liner to indicator-session, there was a g_debug leftover which was breaking the build, robert_ancell asked for it to be cleaned in the merge request but that was overlooked it seemed
[13:49] <seb128> Laney, right
[13:49] <dobey> rodrigo_: oh, ok :(
[13:51] <rodrigo_> dobey, will do a release and rebase the patch
[13:53] <rodrigo_> hmm, firefox is not in debian? apt-cache doesn't find anything that looks like firefox package
[13:53] <kenvandine> seb128, weird it only broke 386 though, but i didn't look at the patch
[13:53] <seb128> kenvandine, it didn't like %li with a guint64
[13:54] <m_conley> rodrigo_: hey - I was wondering if you had a minute to talk about EDS
[13:54] <seb128> in the g_debug
[13:54] <seb128> kenvandine, it's using -Werror
[13:54] <kenvandine> ah
[13:54] <rodrigo_> m_conley, yes
[13:54] <kenvandine> seb128, thx for fixing that
[13:54] <seb128> np
[13:54] <seb128> just mentioning it because I commited to trunk without review
[13:54] <seb128> but it was just dropping the g_debug so should be alright ;-)
[13:54] <kenvandine> :)
[13:55] <kenvandine> good call :)
[13:55] <tedg> seb128, Oh, I thought it was useful for a while. Didn't realize it'd break 386.
[13:55] <m_conley> rodrigo_: so I want to monitor changes to EContacts within EBooks.  Most of the examples I've seen (and the way I believe Evo does it) is by using an EBookView query, and attaching signal handlers for when contacts are added/removed/changed in that view.
[13:55] <tedg> seb128, People should upgrade their kernels ;-)
[13:56] <seb128> tedg, what does it have to do with the kernel?
[13:56] <seb128> btw didn't we say we wouldn't use Werror in tarball?
[13:56] <m_conley> rodrigo_: however, one problem is that I believe those signals are coming in via "random" threads, and that attempts to trigger callbacks in the JS main thread, which freaks Gecko's thread-safety mechanisms out.
[13:56] <tedg> seb128, I'm guessing most people have processors that are amd64 capable, just choosing a system that is not.
[13:56] <seb128> njpatel, unity is failing to build as well :p
[13:57] <seb128> tedg, right, I use i386 there on my i5 ;-)
[13:57] <m_conley> rodrigo_: so I'm considering doing polling instead, but I've been looking for examples of e_book_get_changes_async, and cannot seem to find any
[13:57] <chrisccoulson> Laney, oh, how come moon is being removed?
[13:57] <rodrigo_> m_conley, the get_changes call is going away
[13:57] <m_conley> rodrigo_: ah.  That's very good to know.
[13:58] <m_conley> rodrigo_: how would you suggest monitoring for changes then?
[13:58] <rodrigo_> m_conley, I guess the signals are sent in each operation's thread, although not sure that's the case really
[13:58] <Laney> chrisccoulson: general firefox breakage and being out of date afaik
[13:58] <Laney> it didn't work with the new somethingorother
[13:58] <rodrigo_> m_conley, if on a thread, you should install an idle/timeout, etc, and make any not-thread safe call on the handler for that timeout/signal/whatever
[13:59] <seb128> njpatel, complaining about a variable set and not used and Werror (I'm turning Werror off in the package for this upload)
[13:59] <chrisccoulson> Laney, my change avoids the firefox breakage issues
[13:59] <rodrigo_> m_conley, but are you sure the signals are on threads? afaik all signals are sent on the main thread
[13:59] <Laney> chrisccoulson: oh cool
[13:59] <Laney> does the plugin work?
[13:59] <njpatel> seb128, thanks, will fix in trunk when i get some time
[13:59] <chrisccoulson> Laney, yeah, it seems to on natty
[14:00] <chrisccoulson> (which has the same change now)
[14:00] <m_conley> rodrigo_: hrm.  Well, it's kind of hard to tell.  All I know is that listening to signals from EBookView was causing me to fail thread-safety assertions.
[14:00] <Laney> oneiric might be a different matter, given mono 2.10
[14:00] <m_conley> rodrigo_: I must admit, Gecko's thread-safety model is very new to me.  :)
[14:01] <m_conley> rodrigo_: but thanks for telling me about get_changes going away.  So, just to reiterate, you'd suggest for me to listen for signals from EBookView?
[14:01] <rodrigo_> m_conley, can you show some code to see if I can see something?
[14:01] <rodrigo_> m_conley, yes, that's the correct way
[14:01] <m_conley> rodrigo_: Sure, let me whip something up, and I'll show you.  I'll let you know.  Thanks for your help!
[14:01] <chrisccoulson> Laney, i spoke to shana on IRC a few weeks ago, and she said we could just turn off the firefox bridge and provide the curl bridge instead, and it should carry on working
[14:01] <chrisccoulson> (which is what i've done)
[14:04] <Laney> excellent
[14:04] <Laney> you should pass the news on to directhex
[14:06] <kenvandine> seb128, i uploaded a new indicator-power that will build against the current libindicator
[14:06] <kenvandine> should be able to get the goodness soon :)
[14:06] <seb128> kenvandine, ok, what is missing in the stack?
[14:06] <seb128> kenvandine, why are the breaks on the old libindicator soname require again btw?
[14:06] <kenvandine> indicator-me
[14:06] <kenvandine> which i am doing now
[14:06] <kenvandine> to make sure everything that depends on it can be updated at the same time
[14:07] <seb128> kenvandine, i've dropped those locally to be able to do the updates without uninstalling half my system
[14:07] <seb128> kenvandine, so it's not a case of "will break", it's a case of "will only load those rebuilt"
[14:07] <seb128> right?
[14:07] <kenvandine> yeah
[14:11] <kenvandine> oh, indicator-messages is still missing
[14:12] <tedg> kenvandine, ?
[14:13] <kenvandine> tedg, did you do a release for indicator-messages?
[14:13] <tedg> kenvandine, Hmm, sees I did all the work but didn't do a release.  Sorry, my bad.
[14:13] <kenvandine> i didn't get a merge proposal
[14:13] <kenvandine> hehe :)
[14:13] <kenvandine> tedg,  i am doing an indicator-me release now too
[14:30] <kenvandine> seb128, indicator-power is in binNEW now
[14:30]  * kenvandine files MIR bug
[14:30] <seb128> kenvandine, ok, looking in a bit once I'm done sorting unity issues
[14:31] <kenvandine> no rush
[14:32] <kenvandine> its weird, now that i have the power indicator, it feels like oneiric is kind of complete again :)
[14:33] <kenvandine> s/again/finally
[14:34] <desrt> rickspencer3: ping
[14:41] <rickspencer3> hi desrt
[14:42] <rickspencer3> wow, been a while since I've been pinged in the desktop channel!
[14:42]  * rickspencer3 feels needed again :)
[14:44] <seb128> rickspencer3, salut
[14:44] <chrisccoulson> hi rickspencer3!
[14:44] <kenvandine> hey rickspencer3!
[14:44] <kenvandine> tedg, how's indicator-messages coming?
[14:44] <kenvandine> tedg, just making sure you didn't forget, again :-D\
[14:44] <rickspencer3> hi all
[14:44] <rickspencer3> bon jour mes amis!
[14:46] <rodrigo_> is gksu on the CD and installed by default?
[14:46] <rodrigo_> hi rickspencer3
[14:46] <seb128> rodrigo_, it has been but there is an effort to drop it this cycle
[14:46] <seb128> rodrigo_, why?
[14:47] <rodrigo_> seb128, for the software panel in g-c-c, seems software-properties-gtk can only be run as root
[14:47] <seb128> right
[14:47] <rodrigo_> so, sudo s-p-gtk won't work if it asks for a password
[14:47] <seb128> gksu.. is what the .desktop does right?
[14:48] <rodrigo_> yes: gksu --desktop /usr/share/applications/software-properties-gtk.desktop /usr/bin/software-properties-gtk
[14:49] <seb128> that's basically what we should do until software-properties it teached to use modern techs
[14:49] <rodrigo_> yeah
[14:53] <seb128> kenvandine, there is still indicator-appmenu as well
[14:56] <kenvandine> seb128, oh yeah... tedg ^^
[14:56]  * kenvandine makes whip cracking noises
[15:01] <pitti> seb128: seems the patch worked?
[15:01] <seb128> pitti, or we got lucky...
[15:01] <seb128> could e-addressbook-factory stop segfaulting in ldap code when I send emails?
[15:02] <seb128> (using evo)
[15:02] <pitti> seb128: -cil> probaly just an AA error
[15:02] <seb128> pitti, ok, can you have a look, I'm finishing sorting unity, indicators there
[15:02] <seb128> we should have most of stack uploaded
[15:02] <seb128> they just need to build now
[15:03] <kenvandine> anyone have suggestions on what to do when you loose all input in lightdm?
[15:03] <kenvandine> just upgraded my test box and no mouse or keyboard input
[15:04] <kenvandine> can't even switch VTs
[15:04] <kenvandine> however if i ssh in and use chvt, keyboard input works on the console
[15:06] <seb128> kenvandine, urg, weird, boot in failsafe?
[15:06] <kenvandine> doesn't work either
[15:06] <kenvandine> just in lightdm
[15:06] <kenvandine> which means i can't login to unity
[15:06] <kenvandine> i guess i could try startx :)
[15:06] <kenvandine> actually, maybe lightdm has logs :)
[15:07] <m_conley> rodrigo_: hey - have a second to look at my EBookView code?
[15:08] <pitti> Laney, seb128: is that libnunit-cil, i. e. nunit source?
[15:08] <rodrigo_> not sure where to put my code for this -> http://people.gnome.org/~rodrigo/ubuntu-control-center.png as it's C code, and all of s-p, u1 and jockey are python only modules
[15:08] <rodrigo_> m_conley, sure
[15:08] <seb128> pitti, it seems so from the build log
[15:08] <pitti> Laney, seb128: anyway, promoted to main
[15:09] <seb128> pitti, https://launchpadlibrarian.net/74804374/buildlog_ubuntu-oneiric-i386.libappindicator_0.3.90-0ubuntu1_FAILEDTOBUILD.txt.gz
[15:09] <seb128> pitti, danke
[15:11] <pitti> seb128: want me to upload glib?
[15:12] <seb128> desrt, ^ do you recommand uploading the eventfd revert to the archive?
[15:12] <seb128> desrt, or was that meant to be for testing only?
[15:12] <desrt> seb128: it was for testing, but i recommend uploading it
[15:12] <desrt> since it clearly fixes the problem
[15:12] <desrt> and since we clearly cannot proceed without the fix
[15:14] <pitti> it also fixes some annoying bugs like the nautilus crashes
[15:15] <seb128> pitti, the new glib does
[15:15] <seb128> pitti, we but could let the code it and disable the testsuite as an another workaround to get it built
[15:16] <pitti> seb128: but that wouldn't fix the actual problem in eventfd?
[15:18] <seb128> pitti, the actual problem seems to happen only on old kernels
[15:18] <seb128> pitti, it wouldn't happen on oneiric
[15:19] <seb128> pitti, it's in the fallbacking
[15:19] <seb128> pitti, eventfd should work fine on the natty or oneiric kernels
[15:19] <pitti> ah, so that's why it only happens on the buildds?
[15:19] <seb128> yes
[15:20] <pitti> ok, so then I'd be fine with disabling tests temporarily
[15:28] <seb128> pitti, I'm fine either way, feel free to upload what you uploaded to your ppa
[15:28] <seb128> we don't need the eventfd thing for anything and we know it built fine
[15:28] <seb128> it will give confirmation that it builds fine on the official builders as well without it
[15:28] <pitti> ok
[15:28] <pitti> as I have little time right now, I'll upload bzr (eventfd revert)
[15:29] <pitti> we can reconsider the next time we upload
[15:29] <seb128> thanks
[15:29] <pitti> (/me @ release team meeting, in the context of a university graduate celebrations party..)
[15:30] <seb128> pitti, you should have asked me to cover for you
[15:30] <seb128> pitti, I can still do if you want
[15:30] <seb128> though I've not prepared
[15:30] <pitti> it's fine; we already covered the desktop part
[15:30] <seb128> ok
[15:31] <pitti> I asked skaet to go early today
[15:31] <pitti> but I don't want to start a new glib build right now
[15:31] <seb128> pitti, I can handle glib if you want
[15:31] <seb128> but we should upload what you put in your ppa anyway
[15:33] <pitti> seb128: just committed the debcommit -r; could you just bzr bd -S and upload?
[15:33] <seb128> pitti, sure
[15:34] <pitti> seb128: cheers
[15:34] <pitti> (little bandwidth here, so that's appreciated)
[15:34] <seb128> kenvandine, did you fix your lightdm issue?
[15:34] <seb128> jibel pointed several bugs about that
[15:34] <seb128> ie bug #807291
[15:34] <ubot2> Launchpad bug 807291 in lightdm "Mouse and keyboard unresponsive until re-plugged in" [Undecided,New] https://launchpad.net/bugs/807291
[15:35] <pitti> good bye everyone then, have a nice weekend!
[15:36] <seb128> pitti, bye
[15:38] <seb128> kenvandine, is your issue on real box or in a vm?
[15:38] <kenvandine> seb128, real box
[15:39] <seb128> hum k
[15:39] <seb128> it started this night around midnight utc
[15:39] <seb128> from bug reports
[15:39] <kenvandine> looking at the X logs, it isn't finding input devices
[15:39] <seb128> but I don't see any obvious candidate for issue in the uploads
[15:39]  * kenvandine tries to replugin
[15:39] <kenvandine> startx fails too
[15:39] <kenvandine> not just lightdm
[15:39] <seb128> jibel: ^
[15:40] <kenvandine> i am guessing udev
[15:40] <kenvandine> maybe udev gets the plugin event and configures the device
[15:41] <kenvandine> confirmed, re-plugging in the mouse and keyboard worked
[15:41] <kenvandine> but doesn't work at boot
[15:42] <kenvandine> would be hard to narrow which update broke it though
[15:42] <kenvandine> i hadn't updated it since before the sprint
[15:42] <kenvandine> downloaded like 400M of debs :)
[15:42] <davmor2> seb128, kenvandine: saw that earlier on my netbook trying to resolve an issue with U1 bit more awkward trying to re-plugin the keyboard and mouse there though :(  but I was doing a fresh alpha2 install to check when the u1 issue appeared
[15:42] <davmor2> since then it has been okay
[15:42] <jibel>  xorg-server 2:1.10.2.902-1ubuntu2 is a candidate
[15:42] <kenvandine> oh interesting
[15:43] <kenvandine> i did a logout
[15:43] <kenvandine> and lost input
[15:43] <seb128> jibel: no it's not
[15:43] <seb128> jibel: the bugs you pointed have been filed around 2am and xorg-server uploaded at 6am
[15:44] <seb128> so 0utc published means it was uploaded before 10utc-ish yesterday
[15:44] <seb128> 10pm
[15:44] <jibel> right and user is running xserver-xorg 1:7.6+7ubuntu1
[15:45] <seb128> hum, I can't stop any obvious candidate on -changes
[16:16] <rodrigo_> is launchpad down?
[16:17] <rodrigo_> ah, no, it's not
[16:41] <rodrigo_> pitti, you are the maintainer of jockey, right?
[16:42] <rodrigo_> pitti, if so, could you please review https://code.launchpad.net/~rodrigo-moya/jockey/fix-787694 ?
[16:56] <seb128> kenvandine, did you try the rebuilt indicators in some way?
[16:59] <kenvandine> seb128, yes
[16:59] <kenvandine> indicator-application seems to be broken
[16:59] <kenvandine> looking into that right now
[16:59] <seb128> right, just went to the same conclusion
[16:59] <seb128> it's taking the unity service down
[16:59] <kenvandine> :)
[16:59] <seb128> thanks
[17:00] <kenvandine> seb128, figure anything out?
[17:00] <seb128> no
[17:00] <kenvandine> i had removed the indicators and been adding them back one at a time
[17:00] <kenvandine> indicator-application was the last i grabbed
[17:00] <seb128> well indicator-application failed to build
[17:01] <seb128> so no hurry
[17:01] <seb128> due to the libappindicator build issue
[17:01] <kenvandine> i need to leave in an hour :)
[17:01] <seb128> but libappindicator has built since so when somebody fix i-a and upload it we should be fine
[17:01] <seb128> I can do the packaging
[17:01] <seb128> get ted to fix the bug ;-)
[17:01] <kenvandine> tedg, ^^
[17:01] <seb128> I'm out for half an hour but will be back then
[17:01] <Sweetshark> pitti: http://ftp-master.debian.org/new/libreoffice_1:3.4.1-1.html is there and I have a completed 3.4.1 build, however packaged with one huge *.orig tarball for now (instead of the multiple ones debian uses). Should I wait for debian waving the release through the queue to use their tarballs or should I do a ppa upload with our own *.org tarball now?
[17:01] <kenvandine> seb128, ok
[17:02] <kenvandine> i'll be leaving in an hour for the day
[17:02] <kenvandine> sort of, back on later tonight
[17:02] <kenvandine> hopefully you'll be sleeping by then :)
[17:02] <seb128> ok
[17:02] <tedg> seb128, What's the build issue?
[17:02] <seb128> have a nice weekend then if we don't cross each other
[17:03] <seb128> tedg, build issue was a distro problem that's fixed
[17:03] <seb128> tedg, but indicator-application segfaults on start
[17:03] <seb128> kenvandine can get you details I guess
[17:03] <seb128> need to run but I will be back in half an hour
[17:03] <kenvandine> later seb128!
[17:03] <seb128> kenvandine, don't bother with the packaging I will deal with it
[17:03] <kenvandine> have a great weekend
[17:03] <tedg> backtrace or it didn't happen
[17:03] <tedg> ;-)
[17:03] <seb128> thanks, you too
[17:07] <kenvandine> tedg, http://paste.ubuntu.com/640246
[17:09] <kenvandine> tedg, gvariant type problem
[17:10] <kenvandine> perhaps
[17:10] <kenvandine> (indicator-loader3:26752): GLib-CRITICAL **: the GVariant format string `(a(sisossss))' has a type of `(a(sisossss))' but the given value has a type of `(a(sisosssss))'
[17:30] <kenvandine> tedg, did you see the backtrace?
[17:30] <desrt> i love GVariant error messages sometimes
[17:30] <desrt> that one is particularly hilarious
[17:30]  * Sweetshark dances a bit. libreoffice-3.4.1-1ubuntu1 build finished!
[17:30] <desrt> particularly if you attempt to read it out loud
[17:30] <tedg> kenvandine, Just now, sorry grabbing a bit of lunch.
[17:31]  * kenvandine wonders why desrt tried to read it out loud
[17:31] <kenvandine> tedg, no worries
[17:31] <kenvandine> just making sure it didn't get missed before i take off for the day :)
[17:31] <kenvandine> tedg, don't rely on merge proposal mail requests to give seb128 a fix... i still never got the last couple you sent me
[17:31] <kenvandine> LP must be angry :)
[17:32] <dobey> kenvandine: it seems to be having issues generating diffs today
[17:32] <kenvandine> yeah, i think that is blocking the mail
[17:33] <dobey> yep
[17:33] <dobey> looks like someone is finally looking at it though
[17:34] <Sweetshark> desrt: at least you can finish reading that in linear time. Try that with a average stl-template-plus-UNO-error in c++
[17:38] <seb128> re
[17:38] <seb128> desrt, btw same glib built in oneiric
[17:39] <seb128> desrt, so it's 2 built which went fine on all arch
[17:39] <seb128> desrt, seems confirming that the issue is somewhere in the eventfd patch
[17:40] <desrt> seb128: i'm quite happy
[17:40] <desrt> seb128: because if it wasn't there, ... i think my head would explode
[17:41] <desrt> seb128: question for you: do you gtk-doc when you build glib, or do you rely on the html from the tarball?
[17:41] <desrt> i didn't see it showing up in the builds....
[17:41] <seb128> desrt, we rely on the html from the tarball for GNOME components usually
[17:41] <desrt> ugh
[17:41] <seb128> there is no reason a local build should generate different html from the same source
[17:42] <seb128> so why wasting cpu cycles to rebuild those?
[17:42] <desrt> except when excessively vendorpatching =)
[17:42] <seb128> it's just changing the configure flags in rules
[17:42] <desrt> and depends, i guess
[17:42] <seb128> that and adding the build-depends
[17:42] <desrt> yup.  thanks.
[17:45] <seb128> desrt, the other way is to make dist tarballs from your git rather than distro patching the diff ;-)
[17:45] <seb128> I didn't think about that when you asked yesterday
[17:47] <desrt> seb128: that's specifically what i was asking about yesterday
[17:47] <desrt> and i decided against it
[17:48] <desrt> since it would require rebasing and typing 'make dist' and going through the whole bother for every update
[17:48] <desrt> if i do it with a patch then i can just update the package and hold on to the same patch
[17:48] <tedg> seb128, Do you want to just distro patch this?  lp:~ted/indicator-application/more-s-is-better/
[17:48] <desrt> (ie: let you do most of the work)
[17:48] <seb128> tedg, yes
[17:48] <seb128> kenvandine, tedg: will test and backport it, thanks
[17:48] <seb128> tedg, bonus point if you get it in trunk
[17:49] <seb128> easier to backport ;-)
[17:49] <tedg> seb128, It doesn't matter really?  no?  I mean doesn't Bazaar handle the details either way?
[17:49]  * desrt goes into a docs-writing coma
[17:49] <seb128> tedg, since it's a vcs probably not you are right
[17:49] <desrt> GCC needs an optimiser pass to deduce the semantics of my code and write documentation about it
[17:49] <seb128> tedg, if you don't change the code before merging ;-)
[17:50] <seb128> but dinner first
[17:50] <seb128> then backport
[17:50] <seb128> be back in a bit
[17:50] <seb128> kenvandine, feel free to test and upload if you have time otherwise I will do it after dinner
[18:05] <kenvandine> seb128, i gotta run now
[18:05] <kenvandine> seb128, so please do it when you can
[18:05] <kenvandine> thx!
[18:06] <kenvandine> have a great weekend!
[18:06]  * kenvandine runs
[18:32] <Sweetshark> And libreoffice-l10n-3.3.3-1ubuntu1 finished!
[18:32]  * Sweetshark dances again.
[18:51] <seb128> tedg, still there?
[18:52]  * tedg is wondering if this is a trick question
[18:52] <seb128> tedg, the fix of yours is not enough
[18:52] <seb128> unity-panel-service still goes down
[18:53] <tedg> seb128, bt?
[18:56] <tedg> seb128, Did you see the dbus-test-runner test on the merge request?  Does that work for you?
[18:57] <seb128> tedg, http://pastebin.ubuntu.com/640299/
[18:57] <seb128> tedg, let me try
[18:57] <seb128> tedg, indicator-loader stopped segfaulting but unity still does
[19:00] <seb128> tedg, v
[19:00] <seb128> tedg, http://pastebin.ubuntu.com/640302/
[19:01] <seb128> tedg, that's the indicator-loader debug output
[19:09] <seb128> tedg, is that useful info? do you need extra ones?
[19:09] <tedg> seb128, Trying to figure out why mine doesn't do that.  It did before.
[19:10] <tedg> seb128, But with the patch it  does not.
[19:10] <seb128> tedg, hum, maybe I screwed something
[19:10] <seb128> well the loader was segfaulting before
[19:10] <seb128> so I think I got the update loaded right
[19:11] <seb128> since now it only displays those warning and have an empty indicator
[19:11] <seb128> unity segfaults though
[19:13] <seb128> tedg, the dbus command line "works", i.e doesn't segfault, but the indicator it displays is empty
[19:13] <tedg> seb128, Yeah, do you have a build of libappindicator handy?
[19:13] <seb128> yes
[19:13] <tedg> seb128, dbus-test-runner -t /usr/lib/libindicator/indicator-loader3 -p .libs/libapplication.so -t ./indicator-application-service -t ../../../libappindicator/trunk/example/simple-client
[19:13] <tedg> seb128, That puts something in the indicator
[19:16] <seb128> tedg, k
[19:16] <seb128> tedg, so you don't get any issue locally? not warning using the loader?
[19:17] <tedg> seb128, no
[19:17] <seb128> how I can help you to debug?
[19:18] <tedg> seb128, Well, I'm hoping having the simple client will cause a segfault perhaps..
[19:20] <seb128> tedg, ok, in fact my build tree was cleaned after build so I'm doing a new build, almost done
[19:20] <tedg> Hmm, I thought unity-panel-service had a command line...
[19:20] <seb128> /usr/lib/unity/unity-panel-service
[19:20] <seb128> ?
[19:22] <tedg> seb128, I mean a way to specify modules to load on the command line.
[19:23] <seb128> tedg, you mean?
[19:23] <seb128> tedg, it does load all the indicators?
[19:25] <tedg> seb128, Yeah it does
[19:25] <tedg> seb128, I thought there was a way to say "load this one"
[19:27] <seb128> bah, simple-client is not built
[19:30] <seb128> tedg, ok, your dbus-test-runner command works
[19:30] <seb128> it displays an envelop icon correctly and the menu works
[19:30] <tedg> seb128, ? So it worked with the client but not without?
[19:31] <seb128> tedg, well as said before the loader displays warning but runs
[19:31] <seb128> unity segfaults still
[19:31] <seb128> tedg, (indicator-loader3:1014): GLib-CRITICAL **: the GVariant format string `(a(sisosssss))' has a type of `(a(sisosssss))' but the given value has a type of `(a(sisossss))'
[19:31] <seb128> is still displayed
[19:32] <seb128> but the loader works
[19:32] <tedg> seb128, So literally if you take off the last "-t" it starts having those warnings?
[19:32] <tedg> Oh, I'm so confused.
[19:32] <seb128> I'm confused myself
[19:34] <seb128> tedg, ok, let's start again
[19:34] <seb128> $ dbus-test-runner -t /usr/lib/libindicator/indicator-loader3 -p .libs/libapplication.so
[19:35] <seb128> -> leads to a indicator-application-service signal 5
[19:35] <tedg> Yeah, the no watchers case.
[19:36] <tedg> Wait, I mean the unable to start service
[19:37] <seb128> $ dbus-test-runner -t /usr/lib/libindicator/indicator-loader3 -p .libs/libapplication.so -t ./indicator-application-service -t ../../../../../../libappindicator/libappindicator-0.3.90/example/.libs/simple-client
[19:37] <seb128> works
[19:38] <seb128>  
[19:38] <seb128> $ /usr/lib/libindicator/indicator-loader3 .libs/libapplication.so
[19:38] <seb128> ** (indicator-loader3:21994): DEBUG: Looking at Module: .libs/libapplication.so
[19:38] <seb128> ** (indicator-loader3:21994): DEBUG: Loading Module: .libs/libapplication.so
[19:38] <seb128> (indicator-loader3:21994): Indicator-Application-DEBUG: Connected to Application Indicator Service.
[19:38] <seb128> (indicator-loader3:21994): Indicator-Application-DEBUG: Request current apps
[19:38] <seb128> (indicator-loader3:21994): GLib-CRITICAL **: the GVariant format string `(a(sisosssss))' has a type of `(a(sisosssss))' but the given value has a type of `(a(sisossss))'
[19:38] <seb128>  
[19:38] <seb128> tedg, ^ does that makes sense?
[19:38] <seb128> bah
[19:38] <seb128> tedg, sorry it works
[19:39] <seb128> stupid indicators!
[19:39] <seb128> ;-)
[19:39] <seb128> tedg, seems like there was a system indicator-application-service still running
[19:39] <seb128> once stopped the new run started working
[19:40] <tedg> seb128, Woot!
[19:40] <seb128> sorry for wasting your time on it
[19:40] <tedg> Probably should have bumped the version there so they wouldn' connect.
[19:40] <tedg> Hmm.
[19:40] <chrisccoulson> my inbox is full of pedro this evening!
[19:40] <chrisccoulson> pedro_, did you get replaced by a bot? ;)
[19:40] <seb128> nothing in the log indicated it was connecting to a running one
[19:40] <seb128> \o/ unity working
[19:40] <seb128> tedg, thanks!
[19:41] <seb128> chrisccoulson, hum, my bug box counts says 181
[19:41] <seb128> it was empty before dinner
[19:41] <seb128> pedro_, !!!
[19:41] <chrisccoulson> seb128, yeah, i've got a lot too
[19:41] <seb128> those mexicans...
[19:41] <seb128> pedro_, ;-)
[19:41] <pedro_> lol!
[19:41] <chrisccoulson> lol
[19:42]  * micahg thought pedro was peruvian
[19:42] <pedro_> chrisccoulson, just delete those emails i'm checking old bugs without comments
[19:42] <pedro_> micahg, oh Chilean actually
[19:42] <micahg> oh right
[19:42] <pedro_> micahg, and seb128 is from Germany
[19:42] <chrisccoulson> lol
[19:45] <seb128> brb trying indicator fix