[00:00] <jasoncwarner_> morning AU/NZ folks
[00:01] <TheMuso> Morning.
[00:06] <RAOF> Aha!  *That's* why nothing was showing up :)
[00:07] <RAOF> Mornin' all.
[00:24] <desrt> good "morning"
[02:53] <pikkachu> hi, what do you do when you need to change a patch between package versions?
[02:54] <pikkachu> I just overwrote the patch in debian/patches, instead of adding a new patch so that quilt pop can be applied to revert the package to the same state as previous version
[02:54] <pikkachu> I also felt the need to change the patch names, and so I did
[05:29] <pitti> Good morning
[05:48] <RAOF> What, it's pitti-morning already?  Damn you, daylight saving!
[05:54] <pitti> I actually got up comparatively late :)
[05:56] <didrocks> good morning
[05:56] <pitti> bonjour didrocks
[05:56] <didrocks> guten morgen pitti, how are you?
[05:57] <pitti> didrocks: quite fine, thanks! we had a quiet wekend; how are you?
[05:57] <didrocks> pitti: quiet week-end here as well. Everything's well :)
[05:59] <pikkachu> hi all
[05:59] <pikkachu> what do you do when a patch needs to be changed between package versions?
[06:01] <micahg> pikkachu: perhaps if you rephrase the question, you've asked this several times over the weekend and I"m not sure what you mean
[06:03] <pikkachu> micahg: can't you imagine yourself changing a patch?
[06:04] <pikkachu> micahg: imagine you maintain a local patch
[06:04] <pikkachu> micahg: for whatever reasons, you want to change the patch
[06:04] <micahg> pikkachu: the simple answer to your question is change it, but I"m guessing you're asking more
[06:04] <pikkachu> micahg: just like you change a branch, that simple
[06:05] <pikkachu> micahg: I'm simply changing the patch as needed, yes, just like we do with branches
[06:05] <micahg> and unless this is a desktop package maintained in the desktop team branches, this is OT for this channel, #ubuntu-packaging would be better
[06:05] <pikkachu> micahg: I was just confused if that's the usual way to go
[06:06] <pikkachu> micahg: for example, 01_fix_bug_xyz.patch, then you work to improve the patch and you just replace the file instead of adding 02_improvements_to_bug_xyz_fix.patch
[06:07] <micahg> pikkachu: let's move to #ubuntu-packaging unless you're working on a desktop branch for archive upload
[06:07] <pikkachu> despite in the latter case being possible to revert to prior states with quilt pop
[06:07] <pikkachu> ah sorry
[07:08] <didrocks> pitti: did you try the new compiz btw? (both in ubuntu-desktop and the unity-team/ppa)
[07:10] <pitti> didrocks: oh, I didn't; upgrading now
[07:10] <Sweetshark> G'Morning
[07:10] <pitti> hey Sweetshark
[07:15] <Sweetshark> pitti: any objections against the analysis in bug 969707 or the conclusion that we need pre-depends?
[07:15] <ubot2> Launchpad bug 969707 in libreoffice "package python-uno 1:3.4.4-0ubuntu1 failed to install/upgrade: sub-processo novo script pre-installation retornou estado de saída de erro 1" [Critical,Confirmed] https://launchpad.net/bugs/969707
[07:27] <pitti> Sweetshark: why does the preinst need to call sync_extensions in the first place? that seems very unreliable to me
[07:27] <pitti> preinsts should do as little as possible, the only real reliability you can get is in postinsts
[07:28] <chrisccoulson> good morning everyone
[07:29] <pitti> hey chrisccoulson
[07:29] <pitti> Sweetshark: but yes, in general you need to pre-depend on anything you use in a preinst, except for essential packages
[07:29] <pitti> Sweetshark: but more importantly, preinst scripts which use non-essential commands are strongly, I mean STRONGLY discouraged
[07:30] <pitti> Sweetshark: in Debian you are required to discuss adding pre-depends: publically and get consensus
[07:30] <pitti> because they really really hurt the upgrade progess
[07:30] <pitti> process
[07:30] <chrisccoulson> hi pitti, how are you?
[07:31] <pitti> chrisccoulson: very well, thanks! had a nice weekend?
[07:31] <pitti> didrocks: restarting now to get new compiz
[07:31] <chrisccoulson> pitti - yeah, it wasn't too bad thank. we cleared a lot of junk from our house and got my daughters new room painted
[07:32] <didrocks> pitti: waiting anxiously ;)
[07:32] <didrocks> hey chrisccoulson!
[07:33] <chrisccoulson> hi didrocks
[07:33] <chrisccoulson> right, now to break everyones firefox menus :)
[07:35] <pitti> didrocks: back
[07:36] <pitti> didrocks: so, when I press win+up for maximizing, I don't get the sticky launcher/help any more; the launcher still appears, though
[07:36] <pitti> didrocks: otherwise the first few seconds seem just fine :)
[07:37] <didrocks> pitti: the sticky shouldn't appear in that case in fact, but TBH, it can still appear in other cases ;)
[07:37] <didrocks> you just got lucky
[07:37] <didrocks> pitti: and I'm moving the shortcuts for ctrl + Super in a few minutes
[07:37] <didrocks> so ctrl + super + Up/Down/Right/Left
[07:39] <didrocks> jibel: as you are the master of findings bugs in a few milliseconds (even sometimes before the update is installed), can you give a try to the new compiz? (in the ubuntu-desktop ppa?) :)
[07:50] <seb128> hey
[07:51] <chrisccoulson> hi seb128
[07:51] <pitti> bonjour seb128, ca va?
[07:51] <seb128> hey pitti, chrisccoulson, how are you?
[07:51] <seb128> pitti, ca va bien merci !
[07:51] <chrisccoulson> seb128, yeah, good thanks. and you?
[07:51] <pitti> seb128: tres bien, merci!
[07:51] <seb128> good thanks ;-)
[07:51] <seb128> did you guys have a good w.e?
[07:52] <pitti> seb128: can I annoy you with udisks again? I sent a proposed patch to bug 899757, but I need you to verify it
[07:52] <ubot2> Launchpad bug 899757 in udisks "udisks-daemon assertion error: HACK: Wanting to register object at path `%s' but there is already an object there." [Medium,In progress] https://launchpad.net/bugs/899757
[07:52] <seb128> short week this week! do you guys have friday off as well?
[07:52] <pitti> seb128: it was rather quiet, but nice indeed
[07:52] <pitti> seb128: we have Friday and Monday as national holidays, and I took Thursday off
[07:52] <seb128> pitti, ok, I finish dealing with w.e backlog and give that a try
[07:52] <pitti> we'll go to Dresden again on Wednesday afternoon, so I'll work offline in the train
[07:52] <pitti> nice long weekend back home
[07:52] <seb128> pitti, ok, France as Monday, I collect Friday because I'm half german :p
[07:53] <seb128> has*
[07:53] <pitti> seb128: you should have said "I'm 'alf German"
[07:53] <seb128> ;-)
[07:53]  * pitti hugs seb128
[07:53]  * seb128 hugs pitti back
[07:53] <pitti> seb128: so, I tried for another two hours over the weekend, but scsi_debug is not good enough to reproduce this :/
[07:54] <Sweetshark> pitti: right, I saw that in the debian policy. however that change by rene for debian bug 658646 works and its a little late to look and test alternatives.
[07:54] <ubot2> Debian bug 658646 in libreoffice-pdfimport "libreoffice-pdfimport: Extension does not work at all" [Grave,Fixed] http://bugs.debian.org/658646
[07:55] <didrocks> salut seb128, bon week-end?
[07:56] <seb128> didrocks, lut, nickel, et toi ?
[07:56] <didrocks> seb128: ça va bien, même si j'ai une vrai semaine, MOI! :)
[07:56] <seb128> ;-)
[07:57] <seb128> pitti, oh, so we tie again at 306
[07:57] <pitti> indeed!
[07:57] <seb128> pitti, well done to get over the 300 mark as well ;-)
[07:57] <pitti> seb128: we need to coordinate at the end of the cycle if we are close, perhaps we can walk over the finish line together :)
[07:57] <pitti> seb128: cheers :)
[07:58] <seb128> that would be good indeed ;-)
[07:58] <pitti> if you are more than 5 ahead, you are the well deserved winner, of course
[08:02] <JamesTait> Goooooood mooooornin' intarwebz!
[08:07]  * didrocks has the feeling he will feel alone on Friday :)
[08:07] <didrocks> and maybe on Monday too if I swap it for later (for having the maximum time right now for next unity release)
[08:13] <jibel> didrocks, salut
[08:13] <jibel> didrocks, 1:0.9.7.4-0ubuntu1~ppa1  ?
[08:13] <didrocks> jibel: hello ;) oui
[08:20] <jibel> didrocks, is it enough to kill compiz to load the new version or I must logout/in ?
[08:20] <didrocks> jibel: it's enough to kill it. However, one on the interesting fix was the hang out on startup :)
[08:20] <jibel> didrocks, ok
[08:24] <jibel> didrocks, how do I disable the kb shortcuts help screen ?
[08:25] <pitti> bonjour jibel, ca va?
[08:26] <didrocks> jibel: there is nothing to disable it AFAIK
[08:26] <jibel> hey pitti , I nearly died electrocuted half an hour ago, but seems to be ok, more white hairs :)
[08:26] <didrocks> jibel: oh? what happened?
[08:26] <htorque> smspillaz: hi! about bug 940603 - compiz won't start with that plugin enabled :-(
[08:26] <ubot2> Launchpad bug 940603 in unity "white box randomly shows up at top left corner blocking applications from using stuff under it" [High,Confirmed] https://launchpad.net/bugs/940603
[08:27] <smspillaz> htorque: is it last in the plugin list ?
[08:27] <jibel> didrocks, I was playing with cables
[08:27] <htorque> smspillaz: let me check
[08:27] <smspillaz> to to ccsm, advanced settings, disable auto plugin sorting and then put it after unityshell
[08:27] <htorque> smspillaz: yes
[08:27] <smspillaz> maybe I should put that in the load rules
[08:27] <smspillaz> hmm
[08:27] <smspillaz> does it crash or somethign then ?
[08:27] <htorque> smspillaz: no, i just don't see anything but the desktop
[08:28] <ivanka> hi seb128, good morning!
[08:28] <seb128> ivanka, hey! how are you?
[08:28] <htorque> smspillaz: let me try a session restart
[08:29] <ivanka> seb128:  Very well and back in the UK. Now that I have network I just upgraded to 12.04, which is why I am here :-)
[08:29] <seb128> ivanka, oh ...  ok! tell what is wrong ;-)
[08:30] <htorque> smspillaz: yes, that did it. though i'm getting lots of flickering on the unity elements.
[08:30] <ivanka> seb128: it tells me "The package system is broken"
[08:30] <smspillaz> blargh
[08:30] <smspillaz> htorque: post your .xsession-errors ?
[08:30] <htorque> smspillaz: yeah, will try to reproduce the white window and add it to the report
[08:31] <seb128> ivanka, seems like your upgrade did it an issue? can you go on a command line and sudo apt-get -f install?
[08:31] <smspillaz> thanks
[08:34] <ivanka> seb128: that seems to have made python-uno all happy again. The package system is no longer broken. Thank you :-)
[08:34] <seb128> ivanka, you're welcome ;-)
[08:47] <htorque__> smspillaz: does this plugin has the potential to make the bug not happen?
[08:48] <smspillaz> no
[08:48] <smspillaz> why
[08:48] <htorque__> ok, then i just need to try harder. thanks.
[08:48] <smspillaz> can't see it ?
[08:48] <smspillaz> haha
[08:48] <htorque__> ha, got it
[08:50] <smspillaz> htorque__: I got a white window just now
[08:50] <smspillaz> but you know
[08:50] <smspillaz> after I disable my debugging plugin
[08:50] <smspillaz> me sad
[08:51] <smspillaz> uhhh what
[08:51] <smspillaz> hmm
[08:51] <smspillaz> I wonder why I can't select it using this thing
[08:51] <smspillaz> thats ba
[08:51] <smspillaz> *ad
[08:53] <htorque__> smspillaz: ditto :-(
[08:53] <smspillaz> I wonder if its something else then
[08:53] <smspillaz> could be a frame window
[08:53] <smspillaz> I'll change the things
[08:53] <smspillaz> to pick frame window geoemtry
[09:04] <htorque__> smspillaz: does that help anyhow? → compiz (stackdebugger) - Warn: 0x1c00007 requested invalid layer 0x7ffffff
[09:04] <htorque__> that window is the panel
[09:04] <smspillaz> FFFFFF
[09:04] <smspillaz> how on earth are they still happening !?!?
[09:04] <pitti> seb128: yay more syncs
[09:05] <seb128> pitti, ;-)
[09:06] <htorque__> smspillaz: though that might be from a different bug (bug 944701), i have no idea if that's really a dupe.
[09:06] <ubot2> Launchpad bug 944701 in unity "Sometimes a single white pixel shows in the top-left corner (dup-of: 940603)" [Undecided,Confirmed] https://launchpad.net/bugs/944701
[09:06] <ubot2> Launchpad bug 940603 in unity "white box randomly shows up at top left corner blocking applications from using stuff under it" [High,Confirmed] https://launchpad.net/bugs/940603
[09:07] <pitti> seb128: btw, new glib landed in unstable, I uploaded pygobject, gtk 3.4 will follow soon, then the rest of gnome 3.4; so we might be able to merge/sync more
[09:07] <smspillaz> whelp my connection died
[09:07] <seb128> pitti, there are quite some 3.
[09:07] <thumper> dumb skype
[09:07] <seb128> pitti, there are quite some 3.4 component landing in experimental, unstable already
[09:08] <thumper> seb128: who looks after skype?
[09:08] <seb128> thumper, microsoft
[09:08]  * thumper snorts
[09:08] <thumper> seb128: who looks after skype in ubuntu?
[09:08] <pitti> it's in the partner repository
[09:08] <seb128> thumper, microsoft? dunno it's a partner thing
[09:08] <thumper> it blows
[09:08] <pitti> last uploaders were mvo and Brian Thomasson
[09:08] <thumper> I can't get it to use pulse
[09:09] <thumper> and it tries to output to my speakers
[09:09] <thumper> makes it unusable
[09:09] <Riddell> 10:00 < debfx> Qt 4.8.1 needs testing. It's available from kubuntu-ppa/experimental
[09:09] <seb128> thumper, did that start recently?
[09:09] <pitti> thumper: right, I haven't been able to use it with my BT headset
[09:09] <thumper> seb128: a while ago, but I was still using the oneiric one
[09:09] <thumper> just changed to precies
[09:09] <thumper> but same issue
[09:09] <Riddell> anyone from unity or ubuntu one to test Qt 4.8.1?
[09:10] <seb128> Riddell, ask on #ubuntu-unity for unity2d maybe
[09:10] <pitti> Riddell: I can give unity-2d a quick test with it if that helps
[09:10] <Riddell> seb128: please go ahead, it's your flavour
[09:10] <Riddell> pitti: please
[09:10] <pitti> Error: 'kubuntu-ppa/experimental' invalid
[09:11] <seb128> Riddell, it's not my flavor, I use unity-3d :p
[09:11] <pitti> Error: 'kubuntu/experimental' invalid
[09:11] <pitti> Error: 'kubuntu/ppa' invalid
[09:11] <pitti> hmm
[09:11] <Riddell> seb128: ubuntu desktop uses unity-"2d"
[09:11] <pitti> Riddell: ^ so what's the actual PPA?
[09:11] <Riddell> pitti: sudo apt-add-repository ppa:kubuntu-ppa/experimental
[09:11] <seb128> Riddell, right, well whoever update something should be responsible to make sure the update does't break and give it some testing
[09:12] <seb128> well anyway pitti is testing it, so good ;-)
[09:12] <seb128> but it could be worth notifying the unity-2d guys
[09:12] <Riddell> yes, it would
[09:12] <pitti> Riddell: argh, silly me; forgot the ppa: prefix, thanks
[09:16] <seb128> htorque, smspillaz: there was an interesting comment on a duplicate of the white rectangle bug which claims it's triggered by chromium 18 notifications
[09:16] <seb128> "it has to do with Chrom{e,ium} 18.x and notifications.  The
[09:16] <seb128> white box appears when I have a gmail notifications specifically."
[09:16] <seb128> downgrading to chromium 17 and it doesn't happen
[09:16] <seb128> htorque: do you use chromium?
[09:19] <htorque__> seb128: yes
[09:19] <htorque__> the trigger seems to be flash content for me, though.
[09:22] <chrisccoulson> stop using chromium? sounds like a good fix to me!
[09:23] <seb128> chrisccoulson, ;-)
[09:23] <seb128> chrisccoulson, btw how are the hud issues with the new dbusmenu and firefox uploads?
[09:25] <chrisccoulson> seb128, yeah, it seems much better. i fixed quite a few things on the firefox side last week too :)
[09:25] <chrisccoulson> (which is what i've just uploaded)
[09:26] <seb128> chrisccoulson, right, I read the changelog, well done!
[09:26] <didrocks> 3 kernel panic in a morning, not happy morning :/
[09:27] <seb128> didrocks, stop doing ios! ;-)
[09:27] <didrocks> seb128: can do if I stop working :p
[09:32] <seb128> pitti, udisks patch is not good
[09:33] <pitti> seb128: what do you see?
[09:34] <seb128> pitti, http://pastebin.ubuntu.com/911256/
[09:35] <seb128> pitti, wait, maybe I screwed the packaging
[09:35] <seb128> debuild binary didn't do what I wanted
[09:35] <seb128> i.e seems it didn't do the make install again
[09:36] <rickspencer3> where am I supposed to find the new privacy settings?
[09:36] <seb128> rickspencer3, system settings -> privacy
[09:36] <rickspencer3> seb128, hmm, I guess it didn't get installed
[09:36] <seb128> rickspencer3, it's on the bottom line
[09:37] <seb128> rickspencer3, install activity-log-manager-control-center
[09:37] <rickspencer3> seb, is it not part of the default?
[09:37] <pitti> seb128: ok, waiting for your next round then
[09:37] <seb128> rickspencer3, it is
[09:37] <rickspencer3> I would have thought I would have gotten it in a dist-upgrade
[09:38] <seb128> pitti, I've something weird, I cped the binary and retried, now udisks exit without error
[09:38] <pitti> seb128: btw, no reason to cp; you can just sudo src/udisks-daemon -r
[09:38] <seb128> rickspencer3, maybe you used "upgrade" rather than "dist-upgrade" on some days?
[09:38] <seb128> pitti, ok, well, now I get udisks to exit without error :-(
[09:39] <pitti> seb128: what's the last three or so lines of output?
[09:40] <seb128> pitti, http://pastebin.ubuntu.com/911261/
[09:40] <seb128> pitti, ok, let me do a clean get a new source, reapply the patch round
[09:41] <seb128> rickspencer3, using "upgrade" avoid having things uninstalled but make you miss new recommends as well
[09:41] <rickspencer3> seb128, right, but i never upgrade, I always dist-upgrade
[09:41] <rickspencer3> anyway, it's installed now
[09:42] <seb128> rickspencer3, well maybe you did catch dist-upgrade at a time the new binary was not available
[09:42] <rickspencer3> seb128, right, it's probably just one of those things
[09:42] <seb128> rickspencer3, since that's only a recommends apt is happy to ignore those when that happens
[09:42] <rickspencer3> I'm not worried
[09:44] <seb128> pitti, oh
[09:45] <seb128> pitti, ok, got it
[09:45] <seb128> pitti,
[09:45] <seb128>      g_warning ("Ignoring device with nonexisting native path: %s", device->priv->native_path);
[09:45] <seb128> (gdb) p device
[09:45] <seb128> $1 = (Device *) 0x0
[09:45] <seb128> pitti, no cookie for you
[09:45] <seb128> -> SEGV
[09:48] <seb128> pitti, out of that new bug the patch works fine, i.e I dropped the %s from the g_warning and got the buggy case -> udisks keeps running
[10:01] <pitti> seb128: erk, indeed; should be device->priv->native_path -> native_path
[10:01] <pitti> seb128: did you test with this/
[10:01] <pitti> seb128: as I said, I have no way of actually triggering this condition :/
[10:01] <seb128> pitti, I tested dropping the %s as said
[10:02] <pitti> or that
[10:02] <seb128> pitti, that works, udisks2 seems solid as well
[10:02] <pitti> -      g_warning ("Ignoring device with nonexisting native path: %s", device->priv->native_path);
[10:02] <pitti> +      g_warning ("Ignoring device with nonexisting native path: %s", native_path);
[10:02] <pitti> seb128: ^
[10:02] <seb128> pitti, ok, I had just dropped the argument for testing, works fine
[10:02] <pitti> seb128: ok, many thanks for testing! good to hear that udisks2 is unaffected
[10:02] <seb128> pitti, yw
[10:02] <pitti> seb128: udisks1 still has a lot of hacks and workarounds for earlier kernels
[10:02] <pitti> this is one of them
[10:03] <seb128> ok
[10:03]  * pitti sends upstream and will upload this week
[10:16] <didrocks> seb128: thanks for the libindicator revert :)
[10:17] <seb128> didrocks, yw
[10:17] <didrocks> seb128: was spamming us with crashes :)
[10:17] <seb128> g_variant is hard, people keep screwing up using it ;-)
[10:17] <thumper> C++ is so much easier...
[10:17]  * thumper trolls
[10:23] <glatzor> mvo, hello. Have you time to look at the bug fix merge proposals for session-installer and aptdaemon?
[10:28] <mvo> glatzor: yes, sorry that it took so long
[10:32] <mvo> seb128: do you have a idea why there is no underline in the label in this example https://bugs.launchpad.net/ubuntu/+source/software-center/+bug/955030/+attachment/2988350/+files/lala.py
[10:32] <ubot2> Launchpad bug 955030 in software-center ""Show N technical items" doesnt look clickable" [High,Triaged]
[10:39] <seb128> mvo, hey
[10:41] <mvo> seb128: this is a bit puzzling, almost like there is some special behavior triggered for a bottom hbox that is expand=False
[10:42] <mvo> seb128: this makes the underline appear:
[10:42] <mvo> -box.pack_start(hbox, False, True, 5)
[10:42] <mvo> +box.pack_start(hbox, True, True, 5)
[10:46] <seb128> mvo, puzzling
[10:48] <mvo> seb128: yeah, works fine in 11.10 fwiw
[10:48] <mvo> seb128: I played around with various stuff, putting the bottom hbox in a alignment etc but can't make it display the underscore, funny enough bold color etc all work
[10:51] <seb128> mvo, did you try in C to see if that's python specific?
[10:52] <seb128> mvo, I tried ld preloading some gtk version back, i.e 3.3.10, still the same bug...
[10:52] <seb128> mvo, do you know when that issue started?
[10:52] <mvo> seb128: unfortunately not, I can start working on a workaround for this though
[10:53] <seb128> mvo, I'm about to go for lunch but I will keep looking for a bit and try the C equivalent then
[10:55] <mvo> seb128: ok, lets talk in a bit
[11:28] <didrocks> seb128: so, if we want the HUD key being editable, we need to touch GTK_TYPE_CELL_RENDERER_ACCEL I guess in Gtk
[11:28] <didrocks> seb128: not sure we want that though
[11:29] <seb128> didrocks, you mean editable in the ui?
[11:29] <didrocks> seb128: right, for accepting key modifiers like "Alt"
[11:29] <didrocks> (without any keybinding
[11:29]  * didrocks looks at other masks
[11:29] <seb128> didrocks, I don't like the sound of it ;-)
[11:30] <didrocks> seb128: I'm not surprised :p me neither for the record
[11:30] <seb128> didrocks, can we bind a special ui in the appearance behaviour tab rather?
[11:30] <didrocks> JohnLea: ^
[11:31]  * didrocks looks at Gtk meanwhile
[11:31] <didrocks> seb128: in addition to the Gtk change, there is some changes needed for g-c-c to not reject Alt afterwards anyway :)
[11:32] <seb128> didrocks, the HUD on Alt is madness it's about time they change it to alt-space or something ;-)
[11:32] <seb128> JohnLea, ^
[11:32]  * didrocks *sighs* and agrees
[11:33] <seb128> didrocks, joke aside we are not going to patch gtk to allow a "alt" tapping in the renderer, it's a crazy usecase and not something to change in the toolkit, using alt is special enough that they should come with a custom ui like a 2 choices selection,
[11:33] <seb128> * use alt
[11:33] <seb128> * use a custom keybinding
[11:34] <seb128> then the custom keybindings would let you enter a normal keybinding use the gtk widget
[11:34] <didrocks> seb128: agreed, I can imagine tons of side-effects in various apps
[11:34] <didrocks> hum
[11:34] <didrocks> not quite sure how that can be shown, but yeah
[11:34] <didrocks> let me look if another value can work though
[11:36] <seb128> didrocks, well it can't in the keybindings section, that would need to go to the behaviour appearance tab imho
[11:37] <didrocks> I had some hope in "accel-key" mode that the filtering wasn't here
[11:37] <didrocks> but it's not the case
[11:37] <didrocks> for accel-mode, the Gtk doc is quite clear:
[11:37] <didrocks> "Determines if the edited accelerators are GTK+ accelerators. If they are, consumed modifiers are suppressed, only accelerators accepted by GTK+ are allowed, and the accelerators are rendered in the same way as they are in menus."
[11:39] <didrocks> anyway, it doesn't seem reasonable a week from the finale freeze
[11:39] <didrocks> rickspencer3: FYI, as you were asking me this morning ^
[11:47] <seb128> didrocks, rickspencer3, JohnLea: it's crazy that months after ff we are still looking at adding options, we should move the hud to alt-space and get some unity bugs fixed indeed, we already spent way too much time on trying to fix that alt tapping
[11:47] <seb128> indeed->instead
[11:48] <didrocks> well, I fully agree with this sentence and Tim did as well (tapping alt for getting the HUD costed a lot and it's still not fully fixed…)
[11:49] <didrocks> I think as well any game using alt will be screwed as well (try reset ccsm with pressing "alt")
[11:51] <tjaalton> didrocks: is it a known bug that fullscreen apps (games) aren't really fullscreen, the unity panel/launcher is always visible? easy to repro by running phoronix-test-suite with some test, possibly a regression in the current unity release
[11:52] <didrocks> tjaalton: hum, I didn't notice a bug about it yet, maybe there is one, please poke upstream on their channel directly ;)
[11:52] <didrocks> tjaalton: jay and duflu in particular
[11:52] <tjaalton> didrocks: which one was that?
[11:52] <tjaalton> the channel
[11:53] <didrocks> tjaalton: #ubuntu-unity
[11:54] <tjaalton> didrocks: ah, of course.. thanks
[11:54] <didrocks> yw ;)
[12:02] <asac> UPDATE: i am on latest unity proper, i didnt see any .crash file for a day, by bluetooth headset needs less pairing (e.g. auto re-connects after loosing connectionm and also after suspend/resume) => i believe you guys are on right track. rock on!
[12:03] <asac> seb128: hud on alt-space? that currently pops up the window menu for me :)
[12:03] <seb128> asac, hey, good to read such comments, thanks ;-)
[12:03] <asac> whats the hud?
[12:03] <seb128> asac, control menus from the keyboard, what you get in unity if you tap alt
[12:04] <asac> ah
[12:04] <asac> that one stick around now if you do something more common, right?
[12:04] <asac> i remember this getting in my way  a bit
[12:04] <asac> dont know which use case that was though
[12:04] <didrocks> seb128: translated to "tête haute" in French if you didn't notice :)
[12:04] <seb128> asac, it used to popup on alt based keybinding, like alt-f2 or ws changes
[12:04] <seb128> didrocks, ;-)
[12:04] <asac> btw, i still have the problem that i bought a new cherry keyboard and it has NO WINDOWS KEY
[12:05] <seb128> didrocks, I didn't, weird translation :p
[12:05] <asac> what feature am i missing now?
[12:05] <asac> seb128: oh yeah ... i cannot maximize windows with alt-f10 anymore
[12:05] <asac> that was one
[12:05] <seb128> asac, most of the nice stuff, like opening the dash or keyboard shortcuts to use the launcher
[12:05] <didrocks> seb128: I just noted it because of the shortcut keybinding: "Raccourci pour faire apparaitre l'affichage tête haute" :)
[12:05] <seb128> i.e super 1 launching,focussing your first launcher icon
[12:05] <didrocks> the first who can understand wins a price
[12:05] <asac> hmm
[12:06] <didrocks> noticed*
[12:06] <asac> oh if you tab alt i also start seeing the menu
[12:06] <didrocks> asac: maximizing is Super + Up arrow
[12:06] <didrocks> asac: tomorrow, it will normally be Control + Super + Up arrow
[12:06] <asac> didrocks: heh. ok. so i dont have that feature anymore
[12:07] <asac> why do we need to take this alt- stuff away again?
[12:07] <didrocks> asac: tapping alt for triggering an action isn't really supported anywhere
[12:08] <didrocks> asac: so you end with the HUD showing in front of a lot of application when you wanted to do alt + <something>
[12:08] <didrocks> there are a lot of tricks and workarounds in compiz for that, but it's not fully working and there are still some edge cases
[12:08] <asac> aha
[12:09] <asac> if i press on the ubutu button in the launcher, is that the HUD?
[12:09] <seb128> no, that's the dash
[12:10] <seb128> if you tap alt it's the hud
[12:10] <asac> "Type  your Command" ?
[12:10] <seb128> yes
[12:10] <asac> ah the "smart command console"
[12:10] <seb128> that's basically doing matching on the menus of the focussed app
[12:10] <seb128> the app and the indicators
[12:11] <asac> wow
[12:11] <asac> indeed :)
[12:11] <seb128> well at the moment, it will be improved over cycles to be smarter than using only menus, but for this cycle the "commands" source are the menus
[12:11] <asac> yeah
[12:12] <asac> is there any less type intensive HUD future? like something innovative that would allow you to reach all avaialable/focussed action through mouse navigation?
[12:12] <asac> or is this a general keyboard centric UI?
[12:12] <seb128> asac, keyboard for now, voice over time
[12:12] <seb128> asac, not sure about other input methods
[12:13] <asac> right
[12:13] <asac> so a freeform interaction thing
[12:13] <seb128> asac, but I assume you rather want gestures for mouse and you don't need the hud ui to use gestures
[12:13] <asac> lets see if an how i can work that into my daily use
[12:13] <asac> not sure about that
[12:13] <asac> i rarely hit menus anyway
[12:13] <asac> i guess the case with most mouse usage is really selecting text
[12:13] <seb128> asac, well it's handy that it controls indicators
[12:13] <didrocks> asac: people reports good thing in application with complex menus, like drawing software, gimp, inkscape…
[12:14] <asac> seb128: how often do you hit indicators?
[12:14] <asac> mine are usually sitting happily in the bar :)
[12:14] <asac> but yeah
[12:14] <asac> i often try to find the sound thing
[12:14] <asac> to poke around
[12:14] <asac> but that is mostly bug induced imo
[12:15] <seb128> asac, the messaging one, if you talk to me I can alt- asac - enter and type my reply to your without using the mouse for example...
[12:15] <seb128> well that's for im
[12:15] <asac> oh... ok. i think my messaging isnt properly integrated there
[12:15] <asac> i see wifi connecting is also potentially nice
[12:15] <seb128> on IRC I would type desktop or #ubu or something
[12:16] <seb128> asac, emails as well, you receive a message in your linaro box in tb you can alt-lin-enter
[12:18] <asac> seb128: yep. i dont use mail anymore
[12:18] <asac> that was declared ancient technology :)
[12:18] <asac> in my unit
[12:19] <seb128> asac, how do you guys communicate nowadays? google hangouts?
[12:19] <didrocks> asac: you sent your orders by g+? ;)
[12:19] <asac> yeah hangouts are slowly taking over
[12:19] <asac> mumble is still kicking though and IRC is the river that continues to flow
[12:19] <asac> didrocks: good idea ... at least they would read it there :)
[12:19] <didrocks> ;)
[12:20] <asac> i actually try to move back to super-ancient... hence tried to chop some cycles to fix/improve mutt for gmail http://code.google.com/p/mutt-gmail/
[12:21] <asac> later i noticed how much spaghetti code mutt is so I believe this will take a couple of more weeks to change the user experience paradigm to match that
[12:21] <asac> of gmail
[12:22] <asac> hehe
[12:22] <asac> ok have to work hard now
[12:22] <asac> sorry :)
[12:25] <seb128> asac, have "fun" ;-)
[12:27]  * desrt yawn
[12:27] <desrt> happy monday
[12:28] <seb128> desrt, hey, how are you? had a good w.e?
[12:28] <desrt> seb128: pretty good.  hud-service is faster now :)
[12:28] <seb128> desrt, ;-)
[12:29] <seb128> desrt, we got some users for get hud-service using 1G ram, there might be a leak somewhere, is that on your list of issues to watch?
[12:29] <seb128> users who*
[12:30] <desrt> seb128: i pushed a relatively large leakfix as part of my performance work
[12:30] <desrt> when i added the icon support, each HudItem got a copy of the icon
[12:30] <seb128> desrt, ok, when do you hope to land that work?
[12:30] <desrt> and didn't free it
[12:30] <seb128> utch
[12:30] <desrt> so you would leak one icon-name-sized string per-menuitem-per-window that ever existed
[12:30] <desrt> seb128: i'd appreciate a test build right now, actually
[12:31] <desrt> out of my hud-performance branch
[12:31] <desrt> i now have a performance testcase in there that we can use to make informed optimisations if more are needed
[12:32] <desrt> currently runs in 0:33.  i don't know how long the old code takes to run it -- i left it running when i went out and it was still on when i came back
[12:32] <mvo> seb128: any ideas about my test-case for the underscore stuff?
[12:32] <seb128> mvo, no :-( maybe ask on #gtk+?
[12:33] <seb128> mvo, it's still on my list but I don't see anything obviously wrong ... let me finish what I'm doing and test with the C api
[12:33] <mvo> ok, I can convert to C too if that helps
[12:34] <seb128> mvo, I can do it that's ok
[12:35] <seb128> desrt, I guess I want to backport r200 to fix the leak?
[12:36] <seb128> desrt, I will do that and get a build from your hud-performance locally and then to the desktop team ppa if it works fine for me
[12:36] <desrt> sure
[12:36] <desrt> i tweaked the algorithms a lot too
[12:36] <desrt> (ie: not just faster -- they will be returning different results now)
[12:37] <desrt> so local french-testing is highly appreicated :)
[12:37] <seb128> desrt, I think #system does tarballs on wednesday so 2 days ppa testing and precise on wednesday seems good
[12:37] <desrt> great
[12:37] <desrt> the biggest change is that search terms are now order-sensitive
[12:37] <seb128> desrt, yeah, I will give it some of my preferred testcases including stuff like préférences with and without accents ;-)
[12:37] <desrt> so "preferences edit" is not a great way to find "Edit > Preferences"
[12:38] <desrt> in trade for that, though, we now have penalising of missing end terms (but not start terms)
[12:38] <seb128> works for me
[12:38] <desrt> so "save" matches equally "File > Save" and "Some > Weird > Menu > Save"
[12:38] <seb128> well I mean I'm fine with the ordering
[12:38] <desrt> but less good for "File > Save As"
[12:39] <desrt> we have a bug about that..
[12:39] <didrocks> hey desrt ;)
[12:39] <desrt> didrocks: g'morning
[12:41] <didrocks> desrt: ah, I had a question for you!
[12:41] <desrt> time for a nap!
[12:41] <didrocks> :p
[12:41] <didrocks> desrt: imagine that we have a funny story, like tapping alt would do something on your system
[12:41] <didrocks> now, imagine that you want the user to configure that shortcut
[12:41] <didrocks> like, on g-c-c
[12:41] <desrt> you'd wire up alt as a hotkey?
[12:41] <desrt> that's pretty fucking stupid imho
[12:42] <desrt> but i assume it's only a silly example, so please continue
[12:42] <didrocks> yeah, let's keep on the theorical side :p
[12:42] <didrocks> then, this one would appear to be using accel-mode for it's GtkCellRenderer and so discare assigning "alt" to anything
[12:43] <didrocks> as in this case, it wouldn't send the "accel_edited" callback
[12:43] <didrocks> and even would filter the alt key in the cell render to show "disabled"
[12:43] <didrocks> so, in this hypothetical world
[12:43] <didrocks> would you think of a hack to avoid any more pain and still achieve this?
[12:44] <didrocks> (oh, and assume than someone else, rightly so, doesn't want us to patch Gtk :p)
[12:44] <desrt> i'm confused.
[12:44] <seb128> you can assume than the someone else is french
[12:44] <desrt> alt is used by a system shortcut
[12:44] <desrt> and that's causing problems with the toolkit?
[12:44] <didrocks> of course, we are at the beginning of a cycle and we have the full time to consider any options :)
[12:45] <didrocks> seb128: ;)
[12:45] <didrocks> desrt: basically, the g-c-c shortcut dialog is using a gtk treeview to assign shortcuts
[12:45] <seb128> desrt, the toolkit is what provides the gcc widget used to change shortcuts
[12:45] <didrocks> this is where you can't set "Alt" to it
[12:45] <desrt> oh.
[12:45] <desrt> that's very bad.
[12:45] <seb128> desrt, but the toolkit guys didn't consider that you might want to use "alt" as a shortcut
[12:46] <desrt> sounds like the way in which alt is being grabbed is broken
[12:46] <didrocks> The "accel-mode" property
[12:46] <didrocks> "Determines if the edited accelerators are GTK+ accelerators. If they are, consumed modifiers are suppressed, only accelerators accepted by GTK+ are allowed, and the accelerators are rendered in the same way as they are in menus."
[12:46] <didrocks> from devhelp
[12:46] <desrt> what happens for tapping other modifiers?
[12:46] <desrt> ctrl, say?
[12:47] <desrt> so interesting...
[12:47] <didrocks> desrt: same, you don't have the edit signal
[12:47] <didrocks> desrt: but Ctrl+alt does
[12:48] <desrt> the windows key comes to mind as the obvious example of a modifier that could possibly cause an action to occur on another desktop environment...
[12:48] <desrt> and it gets ignored as well
[12:48] <didrocks> right
[12:48] <desrt> and if you look in the settings, you see that alt+f1 is assigned for this purpose
[12:48] <didrocks> then, you have a filtering g-c-c side as well
[12:48] <didrocks> but this one would be easy to bypass
[12:48] <desrt> i assume the changing of the accel occurs by taking a keyboard grab
[12:49] <desrt> so that no existing system-level grabs have an effect
[12:49] <desrt> ie: toolkit is entirely in charge here
[12:49] <didrocks> desrt: how would Ctrl + Alt takes a keyboard grab then?
[12:49] <desrt> for me it doesn't
[12:50] <didrocks> ok, you're right
[12:50] <didrocks> I should have touched another key while trying
[12:50] <desrt> so it's complicated
[12:50] <desrt> because those are modifiers
[12:50] <desrt> they are not supposed to be perceived as input
[12:50] <didrocks> indeed
[12:51] <desrt> (windows key aside), the idea of tying one to an accel is a bit... odd
[12:51] <didrocks> desrt: remember, it's still hypothetical :p
[12:51] <desrt> (and i only put the windows key aside because the real oddity there is that the windows key is considered a modifier)
[12:52] <desrt> so i'm sure the code explicitly ignores modifiers right now so that when you go to press alt+x, it doesn't assume that you wanted 'alt'
[12:53] <desrt> and therefore i assume that what you're proposing is to have the key sequence <press alt><release alt> to come up with alt as the accel
[12:53] <didrocks> yeah, and detecting tapping would be a little bit hackish, risky, and won't make seb128 happy? ;)
[12:53] <didrocks> exactly :p
[12:53] <seb128> <- not happy
[12:53] <seb128> ;-)
[12:53] <didrocks> see!
[12:53] <didrocks> ;)
 <- never happy
[12:54] <seb128> desrt, !!!
[12:54] <didrocks> I'm not sure why g-c-c has a filter on its side as well for modifiers only shortcut btw, seems a double checking (but easy to workaround) :)
[12:55] <desrt> the other concern: this only makes sense at all if the toolkit is able to see 'alt' as an accel
[12:55] <desrt> and i have some doubts about that
[12:55] <desrt> i've heard that treating the process of tapping alt to be a shortcut is quite... hard
[12:56] <didrocks> desrt: just wanted to know your feeling about it
[12:56] <desrt> my feeling are that you are in a world of pain :)
[12:56] <didrocks> we pinged JohnLea about the technical limitation, let's see his answer
[12:56] <didrocks> desrt: may I quote you? ;-)
[12:57] <desrt> sure
[12:57] <didrocks> so not so-easy-hack-without-any-regression on top of your head for that matter?
[12:58] <desrt> i don't think it's a case of 'filtering' as you suggest
[12:58] <desrt> i think it's rather a case of the toolkit treating modifier keys as modifier keys
[12:58] <desrt> pressing alt on its own is not supposed to have an effect...
[12:58] <didrocks> yeah, so the wanted event isn't even triggered on pressing alt and we would have to hack around it in a very bad way
[12:58]  * desrt is looking at the code now
[12:58] <desrt> well
[12:59] <desrt> i'm sure that there's a keyboard grab setup and a function that gets *all* keystrokes at that time
[12:59] <desrt> ignoring modifiers
[13:01] <desrt> maybe not...
[13:02] <desrt> looks like that function is not even told of modifiers
[13:02] <didrocks> oh?
[13:02] <desrt> it takes the first keypress it sees at all...
[13:03] <didrocks> the callback though have the GdkModifierType
[13:03] <didrocks> so this is factored in a higher level?
[13:03] <desrt> ya. but gtk tells it what that is
[13:03] <desrt> indeed
[13:03] <didrocks> "fun" :)
[13:03] <desrt> it would seem that if you press alt+s then you see only one keypress
[13:03] <desrt> which is "s" with modifier mask set for alt
[13:04] <desrt> this is really cool
[13:06] <desrt> so gdk reports the event as a normal keypress
[13:06]  * desrt continues to dig
[13:07] <bcurtiswx> good morning
[13:07] <didrocks> desrt: "the event", you mean, an "alt tapping"?
[13:07] <didrocks> hey bcurtiswx
[13:07] <didrocks> well, an alt press
[13:07] <desrt> i mean an alt press
[13:08] <didrocks> ok
[13:09] <desrt> doing some tests...
[13:09] <desrt> this may not be _that_ hard....
[13:09] <desrt> it may have to be opt-in
[13:10] <desrt> like with GTK_CELL_RENDER_ACCEL_MODE_BARE_MODIFIERS or something
[13:11] <bcurtiswx> is kenvandine out today?
[13:12] <didrocks> desrt: opt in would be great, yeah. People will stil be able to assign "alt" to some other stuff that Gtk wouldn't be able to handle, or maybe I can prevent that g-c-c side
[13:12] <didrocks> like "only enable alt only on this key"
[13:12] <desrt> didrocks: doesn't work like that
[13:13] <desrt> keep in mind that this is a cellrenderer and only one copy of the cell renderer exists :)
[13:13] <desrt> although, strictly speaking, i guess you could probably put a key in the treemodel to enable/disable this behaviour and bind that to a property...
[13:14] <didrocks> desrt: I meant, filtering that back on the callback
[13:14] <didrocks> and then eventually rejecting the change
[13:14] <desrt> oh.  interesting.
[13:15] <desrt> quite possible, i suppose
[13:15] <didrocks> yeah, g-c-c structure is well done for that, there are already some cases
[13:15] <didrocks> like you can't bind "a" shorcut :)
[13:15] <desrt> so i feel quite dumb now
[13:15] <desrt> i see what you were trying to tell me about the 'filter'
[13:15] <desrt> the if statement right at the top :p
[13:15] <didrocks> ;)
[13:16] <desrt> okay.  this is doable
[13:16] <desrt> but it won't be pretty
[13:16] <desrt> how about multi-modifier taps, btw?
[13:16] <desrt> is ctrl+alt a shortcut?
[13:17] <desrt> also.. what do i do in the case that i see this sequence:
[13:17] <desrt> <ctrl down> <alt down> <alt up> <shift down> <shift up> <ctrl up>
[13:17] <desrt> or even
[13:17] <didrocks> desrt: TBH, this is only only for one key… so I would assume let's be crazy (like, not consider edge case) on setting that one.  I think people who wants to redefine alt to something else would use something like alt + Space or less crazy
[13:17] <desrt> <alt down> <ctrl down> <ctrl up> <'x' down> <'x' up> <alt up>
[13:17] <desrt> so only single-modifier tapping supported
[13:18] <didrocks> I wouls assume, just <modifier down><modifier up>
[13:18] <didrocks> yeah
[13:18] <desrt> okay
[13:24] <didrocks> patching gtk just for one use case isn't nice though, do you think it's possible to put as much possible this divergence in the code back in g-c-c (like getting the widget here)?
[13:24] <didrocks> that's something that seb128 won't oppose I guess ;)
[13:25] <seb128> desrt, I blame you for suggesting adding another distro patch to gtk we will never get ride of :p
[13:25] <desrt> seb128: UPSTREAM!
[13:25] <desrt> seb128: you may only apply this patch as a cherry-pick from git
[13:26] <desrt> otherwise blame lies with only you :)
[13:26] <didrocks> desrt: you are willing to push that upstream directly? Awesome :)
[13:26] <desrt> why not?
[13:26] <desrt> we could use it to set the windows key for the shell :)
[13:26] <seb128> desrt, ok ;-)
[13:27] <smspillaz> desrt: very very random question
[13:27] <smspillaz> desrt: do you know if gtk releases grabs
[13:27] <smspillaz> if the grabbed window intentionally loses focus
[13:27] <smspillaz> eg
[13:27] <smspillaz> pops up a modal dialog
[13:27] <smspillaz> or does it wait on some condition from the window manager
[13:27] <didrocks> desrt: you're right, there is this why-is-this-a-modifier super annoying key :)
[13:28] <desrt> smspillaz: no idea
[13:28] <smspillaz> okay
[13:28]  * desrt is not a toolkit guy :)
[13:28] <smspillaz> time for me to go digging into gtk
[13:28] <smspillaz> desrt: but I see your name in the copyright :P !
[13:28] <smspillaz> or do you have some
[13:28] <smspillaz> doppelganger
[13:28] <desrt> doppelganger theory sounds good to me
[13:28] <desrt> gotta run!
[13:30] <kenvandine> BigWhale, kazam sponsored
[13:31] <BigWhale> kenvandine, +1 beer
[13:31] <BigWhale> :))
[13:33] <desrt> didrocks: you have a fun new problem
[13:33] <desrt> didrocks: my chnages are working perfectly
[13:34] <desrt> didrocks: but of course you know about how Alt-L and Alt-R are really two different keys...
[13:34] <didrocks> desrt: indeed :)
[13:35] <desrt> got a plan? :)
[13:35] <seb128> desrt, why is that an issue? it means you can map the hud to r-alt which is good
[13:35] <seb128> that means less conflicts?
[13:35] <didrocks> desrt: I don't get the issue? the keycode are differents?
[13:35] <desrt> i dunno.  is that how it works?
[13:35] <seb128> desrt, well current only lalt works by default
[13:36] <desrt> oh.  interesting.
[13:36] <didrocks> it's not the same key in the french keyboard and it doesn't modify the keys the same way
[13:36] <seb128> but I would like to remap to ralt now that you mention it
[13:36] <desrt> so for alt maybe that makes sense -- particularly for non-US keyboards
[13:36] <didrocks> so, left alt seems ok to me :)
[13:36] <desrt> but same happens for ctrl and shift
[13:36] <didrocks> desrt: I guess it makes sense that if you want to assign a physical key for something, you just assign one
[13:36] <desrt> it's quite rare to think that the left shift key is different from the right one
[13:37] <desrt> very well *shrug*
[13:37] <didrocks> not sure, just how I would understand it without trying
[13:37] <desrt> btw: i managed to make multiple-modifier chords word
[13:37] <desrt> *work
[13:38] <desrt> so you can do ctrl down alt down alt up ctrl up
[13:38] <desrt> and then you get Ctrl + Alt L
[13:38] <didrocks> desrt: oh, in addition to that? :) awesome!
[13:38] <desrt> just adding the conditional support now...
[13:38] <didrocks> not sure if some people are mad enough for that, but nice ;)
[13:42] <jbicha> didrocks: howdy, when's Unity 5.10 scheduled for?
[13:42] <didrocks> jbicha: hey ;) will be next week, it will hopefully be the finale release :)
[13:43] <didrocks> this week is all about the preparation and last bug fixes for it
[13:43] <jbicha> ah, I was hoping for a quick fix for bug 964325
[13:43] <ubot2> Launchpad bug 964325 in unity-lens-applications "Regression: Installed apps not sorted alphabetically" [High,Fix committed] https://launchpad.net/bugs/964325
[13:43] <didrocks> jbicha: is that impact you a lot? doesn't seem to make a difference for screenshots?
[13:44] <didrocks> jbicha: if we start backporting every commit, we will never end ;) (and some commits in trunk regressed some test cases already)
[13:45] <jbicha> I wouldn't say it's a big impact, but we do use it for the big screenshot at https://help.ubuntu.com/11.10/ubuntu-help/unity-introduction.html
[13:46] <jbicha> it's not a big deal at this point
[13:46] <desrt> seb128: i didn't mark that patch as fixing that leak because i'm not sure that it does
[13:46] <seb128> desrt, yeah, I will reopen I think, launchpad lack a "fix part of lp: #nnn" which would reference without closing
[13:47] <desrt> well
[13:47] <desrt> more like maybe-fixes :)
[13:47] <seb128> desrt, I like having the cross referencing
[13:47] <desrt> i did a valgrind and it's the only thing that popped out
[13:47] <seb128> desrt, so I listed anyway
[13:47] <desrt> doesn't mean there won't be others
[13:47] <desrt> seems a bit odd to have a few leaking icon names take up 1gig
[13:47] <seb128> desrt, ok, I will comment asking "if you still see a leak please reopen"
[13:47] <desrt> sounds good
[13:48]  * desrt thinks a bit
[13:48] <seb128> desrt, is that a few? or is one for each menu item?
[13:48] <desrt> assume icon name plus malloc overhead is like 100 bytes
[13:48] <desrt> assume each window has like 200 menu items, tops
[13:48] <didrocks> jbicha: do not hesitate to tell me if you have big changes that you need to have backported
[13:48] <didrocks> jbicha: you can still use the unity-team/staging ppa for screenshots
[13:48] <desrt> so like 20k leaked per window open/close
[13:48] <desrt> ah... one thing, though -- network manager is changing its menus all the time
[13:48] <didrocks> jbicha: stability not garanteed, but good enough on a spare machine to take those (it's always latest upstream trunks)
[13:48] <desrt> when new networks come and go
[13:48] <desrt> and signal strength is changing
[13:48] <desrt> so that would be more memory leaked every 10 seconds
[13:49] <desrt> that could add up....
[13:49] <jbicha> didrocks: ok I'll try the PPA, thanks!
[13:49] <didrocks> jbicha: yw ;)
[13:49] <seb128> desrt, one of the guy had a 4 days uptime, so yeah, every 10 seconds on 4 days ...
[13:49] <desrt> it's in the realm of possibility
[13:50] <desrt> i'd have loved to see the bug get to a point where after a few people reporting the same we notice "this bug only seems to happen on laptops...."
[13:50] <desrt> would be fun :)
[13:50] <seb128> desrt, sort of "wouldn't print on tuesday" bug? ;-)
[13:51] <desrt> didrocks: patching control-centre now to use my new API (to test if it works)
[13:52] <didrocks> desrt: ok, in that case, you need to look at the keyval_is_forbidden() function (just return False if you want for now) to see if it works :)
[13:52] <desrt> is that the one that prevents me from binding the lowercase 'x'?
[13:53] <didrocks> desrt: no, that's accel_edited_callback ()
[13:53] <didrocks> look for /* Check for unmodified keys */
[13:53] <didrocks> (the check for keyval_is_forbidden () is part of this if)
[13:54] <didrocks> as well as other check for standalone key
[13:55] <desrt> seems to be working without additional changes
[13:55] <didrocks> oh? keyval_is_forbidden () has some forbidden_keyvals checks which GDK_KEY_Mode_switch
[13:55] <didrocks> uses*
[13:55] <didrocks> I would have assume this one would cause issues
[14:00] <desrt> didrocks: https://bugzilla.gnome.org/show_bug.cgi?id=673360
[14:00] <ubot2> Gnome bug 673360 in general "accel cell renderer: support modifier 'tapping'" [Normal,Unconfirmed]
[14:01] <desrt> seems to work good in gtk and also working well with a small modification in the control centre (ie: changing the 'mode')
[14:01] <desrt> up to you to figure out how to implement it in unity :)
[14:02]  * desrt makes coffee
[14:03] <didrocks> desrt: the patch looks nice and not intrusive. Also I like the new feature. Thinking about it, I guess I even have an hypothetical use case for it! :-)
[14:07] <seb128> cyphermox, hey
[14:07] <cyphermox> hey
[14:08] <seb128> cyphermox, how are you? had a good w.e?
[14:08] <cyphermox> yeah pretty good
[14:08] <cyphermox> and you?
[14:09] <seb128> cyphermox, I'm good thanks
[14:10] <seb128> cyphermox, we just had an archive rebuild test, http://people.ubuntuwire.org/~wgrant/rebuild-ftbfs-test/test-rebuild-20120328-precise.html
[14:10] <cyphermox> cool, will look into it :)
[14:10] <seb128> cyphermox, evolution-exchange and evolution-indicator show are ftbfsing (not finding some eds libs when linking it seems)
[14:10] <seb128> cyphermox, do you think you could look at those 2?
[14:10] <cyphermox> story of the same think as is happening every release, probably
[14:10] <seb128> cyphermox, oh, evolution is on there as well, libpst
[14:10] <cyphermox> yes, should not take much time
[14:10] <seb128> cyphermox, thanks
[14:11] <cyphermox> I'll tackle all the ones I can figure out
[14:11] <cyphermox> also -couchdb, -mapi, -rss, urgh
[14:12] <mdeslaur> cyphermox: thanks for fixing NM adhoc WPA
[14:12] <cyphermox> ah, why is heartbeat and this apache module under ubuntu-desktop?
[14:12] <cyphermox> mdeslaur: "fixing" :)
[14:12] <mdeslaur> cyphermox: well, hiding it :)
[14:17] <chrisccoulson> hmmm, so, i made firefox crash!
[14:26] <tkamppeter> pitti, hi
[14:27] <pitti> hello tkamppeter
[14:28] <tkamppeter> pitti, I have subscribed you to several "CUPS does not restart after update" bugs, on bug 968956 there is already an answer, seems to be AppArmor.
[14:28] <ubot2> Launchpad bug 968956 in cups "package cups 1.5.2-8bzr2 failed to install/upgrade: Unterprozess installiertes post-installation-Skript gab den Fehlerwert 1 zurück" [Undecided,Incomplete] https://launchpad.net/bugs/968956
[14:28] <pitti> tkamppeter: I noticed, tahnks
[14:29] <pitti> tkamppeter: reassigned to apparmor
[14:56] <Sweetshark> pitti: _rene_ only wants conflicts, not pre-depends ...
[14:57] <Sweetshark> pitti: I guess that will work too, although it is not really "correct" ....
[14:57] <pitti> Sweetshark: conflicts to what?
[14:58] <pitti> you need to pre-depend on the package that provides the command your preinst calls
[15:00] <Sweetshark> pitti: conflicts to libreoffice-common/-core <3.5
[15:01] <pgraner> didrocks, are you hearing of unity segfaults after todays updates?
[15:02] <didrocks> pgraner: panel service or unity itself?
[15:02] <didrocks> pgraner: there has been no unity update for a week
[15:02] <Sweetshark> pitti: well, the sync_extensions call uses unopkg.bin (from -core), but is does nothing if that is not there, so with a conflicts:-line we should either get a working 3.5 unopkg.bin or a missing one resulting in a noop
[15:03] <pitti> Sweetshark: does the preinst DTRT if sync_extensions does not exist?
[15:07] <pgraner> didrocks, unity --reset gives me: Segmentation fault (core dumped)
[15:08] <pgraner> didrocks, just got the crash its uploading to LP
[15:08] <Sweetshark> DTRT?
[15:08] <didrocks> pgraner: so you meant, it's a crash on --reset, right?
[15:08] <pitti> Sweetshark: "does the right thing"
[15:08] <pitti> Sweetshark: i. e. succeeds
[15:08] <didrocks> pgraner: keep me posted with the bug #
[15:08] <pgraner> didrocks, yep, after I updated I rebooted and I got a desktop with no panel, launcher etc...
[15:09] <pgraner> didrocks, so ctl-alt-f1 unity --reset give a segfault
[15:10] <bcurtiswx> kenvandine, on first login to my desktop I loaded empathy and used the messaging menu to switch my status to away. All accounts switched except for AIM (telepathy-haze). Once I used empathy to switch to away all accounts switched and upon using the messaging menu after that all accounts switched to the correct status
[15:10] <pitti> Sweetshark: http://anonscm.debian.org/gitweb/?p=pkg-openoffice/libreoffice.git;a=commitdiff;h=f34451f9ef4a7508904704c9abd5c1767b247055 does not have any || true after the sync_extensions
[15:11] <Sweetshark> pitti: but http://anonscm.debian.org/gitweb/?p=pkg-openoffice/libreoffice.git;a=blob;f=shell-lib-extensions.sh;h=21dafe32323e261d0943bcb99cb2429b9aab0602;hb=ubuntu-precise-3.5 check for unopkg.bin being there.
[15:11] <Sweetshark> s/check/checks
[15:11] <bcurtiswx> kenvandine, i searched for bugs in haze or messaging menu or empathy and found none.
[15:11] <didrocks> pgraner: give me the bug # once retraced
[15:11] <pgraner> didrocks, looks like it nux https://bugs.launchpad.net/ubuntu/+source/nux/+bug/971617
[15:11] <ubot2> pgraner: Error: <Bugtracker.plugin.Launchpad instance at 0xa91202c> bug 971617 not found
[15:12] <pgraner> didrocks, not retraced yet
[15:12] <pitti> Sweetshark: but that function/script itself is in l-common
[15:12] <didrocks> pgraner: we got one last week on --reset, but it's suppossed to be fixed, one sec
[15:12] <pitti> Sweetshark: so what if l-common is too old, or not installed?
[15:12] <didrocks> pgraner: ok, I keep the tab open until it's retraced and the bot subscribed me
[15:12] <pitti> Sweetshark: in general, it's so much easier to do such complex stuff in postinst only
[15:13] <pgraner> didrocks, looks like its dumping in the nvidia binary driver
[15:13] <didrocks> pgraner: ah, tseliot ^ maybe?
[15:14] <pgraner> didrocks, possibly
[15:14] <pgraner> didrocks, I subscribed you so you can see the stack trace
[15:15] <Sweetshark> pitti: right.
[15:15] <kenvandine> bcurtiswx, is that consistently reproducable?
[15:16] <kenvandine> like always aim that fails the first time?
[15:16] <didrocks> pgraner: hum, I would be surprised that the crash is because of an invalid unity::ui::IconRenderer::RenderElement reference (as this code is called a lot and didn't change for the past months). So maybe nvidia race? Let's see once it will be retraced
[15:16] <pgraner> didrocks, ack, I'll be here :)
[15:16] <didrocks> :)
[15:17] <bcurtiswx> kenvandine, i will test brb
[15:17] <Sweetshark> pitti: I think the sync_extensions in preinst in superfluous as we do trigger that (via trigger in core) anyway ....
[15:18] <bcurtiswx> kenvandine, using menu Chat->quit then starting empathy again can reproduce the issue
[15:19] <bcurtiswx> kenvandine, my dist is from an upgrade from 11.10 so IDK if there are any residual configs
[15:22] <Sweetshark> pitti: also: shell-lib-extension.sh is not used from an installed package, but added to each maintainer script (see debian/rules:1863), so it is not used from -common, but in fact part of the maintscript itself.
[15:23] <pitti> Sweetshark: oh, it gets put into the .preinst during build time?
[15:23] <Sweetshark> pitti: thats what I assumed, I will recheck.
[15:28]  * didrocks needs to go for the co-owner meeting for his building, ttyl guys! (will probably either reconnect, either work offline)
[15:28] <didrocks> (not sure about the co-owner translation, but it's the closer I can think of :p)
[15:28] <chrisccoulson> heh, pitti - http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread/thread/93b7f392176df62a#. it seems they started to notice that their downloads are getting bigger ;)
[15:29] <pitti> chrisccoulson: hah, at last :)
[15:48] <pitti> good night everyone!
[15:49] <seb128> 'night pitti
[15:49] <pitti> seb128: btw, any luck with bug 899858?
[15:50] <ubot2> Launchpad bug 899858 in oem-priority/precise "regression in gvfs to connect/browse using obex" [High,In progress] https://launchpad.net/bugs/899858
[15:50] <seb128> pitti, it's the remaining item on my todo for today ;-)
[15:50] <pitti> hah
[15:50] <pitti> I'll be back for TB meeting tonight
[15:50] <seb128> pitti, I just got finished with the other stuff I was doing
[15:50] <seb128> pitti, see you later!
[15:50] <pitti> enjoy the evening!
[15:51] <zyga> I'd like to report a bug on thunderbird theming, I'm not 100% sure how to report it, also, I'd like someone to confirm they see this on their end and that it is not too monitor specific
[15:51] <zyga> the bug is: unselected item background color is very hard to differentiate from background (white)
[15:52] <zyga> on my monitor it is virtually the same, there is a very subtle tint but I can barely see it
[15:53] <zyga> the problem is also magnified by message listing that (at least for me) cycles even/odd colors for each row
[15:57] <zyga> anyone? ^^
[15:58] <seb128> zyga, try pinging Cimi on #ubuntu-unity, I mentioned it to him already, he said that's a tb bug they do custom theming stuff
[15:58] <zyga> seb128, thanks
[15:58] <seb128> not sure know if he said he would fix it, he said he was not interested to look at the tb code but that he could maybe workaround it in the theme
[15:59] <zyga> seb128, I'm sure the bug is in the theme, installing an alternate theme allows me to "fix" this
[15:59] <seb128> or the other theme you use has a workaround for tb
[16:26] <G__81> hi jono
[16:31] <jono> hey G__81
[16:32] <G__81> hey jono i had sent my feedback to Daniel after going through the getting started with Ubuntu Devel. I am not sure whether he received it or whether he had a chance to look @ it. I hope i have sent it to the right email id :o
[16:41] <jono> G__81, which email address did you send it to?
[16:42] <G__81> jono: the one mentioned in the blog which talks about this ubuntu development
[16:46] <jono> G__81, then probably just hasnt replied to it yet
[16:46] <ritz> heya, do we have a lp bzr tree for qt ?
[16:46] <ritz> https://code.launchpad.net/qt lists a few different tree
[16:46] <ritz> s
[16:46] <G__81> jono: oh ok just wanted to know whether he received it
[16:46] <jono> G__81, np
[16:46] <jono> you should ask him, I am not sure which email he has or has not recieved
[16:48] <G__81> yeah i tried to see if i could catch him but i dont find him here :)
[16:50] <jono> G__81, he is in #ubuntu-community-team in European hours
[16:52] <jbicha> desktoppers, what do you think of bug 969235 ?
[16:52] <ubot2> Launchpad bug 969235 in metacity "Keyboard shortcuts - Update some window management shortcuts to use "Ctrl + Super" instead of just "Super" " [Undecided,Confirmed] https://launchpad.net/bugs/969235
[17:14] <chrisccoulson> lol @ https://twitter.com/#!/bwinton/status/186863971746316289
[17:14] <mvo> seb128: any luck with my label underline issue?
[17:14] <mvo> (sorry for nagging)
[17:16] <seb128> mvo, the C testcase has the same issue so for once it's not the bindings ;-)
[17:16] <mvo> heh :) and it should be reasonable clear for a bugreport
[17:16] <seb128> mvo, indeed
[17:16] <mvo> ok, that puts my mind at ease for now and I ahve dinner
[17:17] <seb128> mvo, ld preloading gtk from oneiric has the same issue
[17:17] <seb128> mvo, so I wonder if it could be pango or something
[17:17] <seb128> mvo, still investigating
[17:17] <seb128> mvo, I will let you know tomorrow
[17:17] <seb128> mvo, enjoy dinner
[17:18] <mvo> ta! could be pango too, running it in my oneiric VM worked
[17:18] <mvo> I think at least :)
[17:20] <seb128> mvo, no luck downgrading pango, a bit puzzling...
[18:25] <davmor2> seb128: do you happen to have a camera?
[18:28] <davmor2> seb128: if so could you try plugging it in to your precise box, for me I get the "What do you want to do box" it correctly gives me shotwell as the only app option but fails to open it
[18:33] <davmor2> seb128: nevermind missing dep
[18:50] <dobey> bcurtiswx: around?
[18:51] <bcurtiswx> dobey, yes
[18:51] <dobey> bcurtiswx: can you test a fix for bug #969262 ?
[18:51] <ubot2> Launchpad bug 969262 in libubuntuone/trunk "Error while getting credentials: The '/org/freedesktop/secrets/collection/login/24' object does not exist" [High,Fix committed] https://launchpad.net/bugs/969262
[18:52] <bcurtiswx> dobey, sure
[18:53] <dobey> bcurtiswx: if you install the libubuntuoneui-3.0-1 package from the ubuntuone nightlies ppa (ppa:ubuntuone/nightlies), quit rhythmbox, and then search for a song in the music lens for purchase, and click on it, it should open correctly to the page instead of giving that weird error, now
[19:00] <bcurtiswx> dobey, im trying to reproduce the issue so I can confirm it's squashed, but now everytime i try to load rhythmbox with ubuntuone using the lens it keeps asking me to signup for ubuntu one although i already have an account :-\
[19:00] <bcurtiswx> so once i can get that bug reproduced I will install that build
[19:01] <dobey> bcurtiswx: quit rhythmbox and run u2sdtool -d && u1sdtool -c in a terminal
[19:01] <dobey> but i think that is another incantation of the same problem :(
[19:02] <dobey> which means i keep chasing symptoms because i can't find the actual problem. yay memory corruption from race conditions :(
[19:02] <bcurtiswx> dobey, i'm assuming you meant u1sdtool for both ?
[19:02] <dobey> oh yes
[19:03] <dobey> there is no ubuntu two :)
[19:04] <bcurtiswx> i figured out how to reproduce the bug, lol but the tool didn't help it
[19:05] <dobey> ah, no, the u1sdtool bit was just to help you get back logged in
[19:05] <dobey> since the music store widget deleted your token :(
[19:05] <bcurtiswx> ah i c
[19:06] <dobey> which is what the original error was about. the token got deleted from the keyring while it was being used, so the client (music store widget) had a reference to a keyring object on dbus that no longer existed :(
[19:06] <dobey> but this definitely seems to be memory corruption induced by io bound operations :(
[19:11] <bcurtiswx> dobey, ok i've installed 3.1+r149-31~precise1 and i'm going to try to reproduce
[19:12] <bcurtiswx> dobey, still error :(
[19:13] <dobey> sigh :(
[19:16] <dobey> bcurtiswx: thanks, i'll try to see if i can't find something else that will cause it.
[19:16] <bcurtiswx> I open rhythmbox by itself (no lens), click the ubuntu music store in there, let it load, close rhythmbox.  Using the music lens type in 'adele' and click on her album '21' and it will cause error
[19:18] <dobey> right
[19:18] <bcurtiswx> if i keep reproducing the error, the number at the end increases
[19:18] <dobey> right
[19:18] <dobey> because the id for the entry in the keyring has changed
[19:19] <dobey> or it keeps trying to get the next wrong one, or something
[19:19] <dobey> i'm not sure, but that's just a symptom and not the problem itself
[19:19] <dobey> albeit that error message is kind of useless
[19:20] <bcurtiswx> dobey, OK. Good luck, and I'm more than happy to test whenever you think of another potential fix. I am heading home though and will be available tomorrow starting at 9AM ET
[19:20] <dobey> and it's a very io-bound problem. if i increase startup io load for rbox enough to get it on my workstation, then fix it, it still breaks on slower machines it seems
[19:21] <dobey> ok, thanks bcurtiswx
[19:21] <bcurtiswx> see everyone tomorrow
[20:25] <chrisccoulson> can anyone else trigger bug 971800?
[20:25] <ubot2> Launchpad bug 971800 in firefox "Firefox crashes while trying to delete a bookmark" [Undecided,Incomplete] https://launchpad.net/bugs/971800
[23:01] <Sweetsha1k> oh, mayhem in oakland and its not even UDS.
[23:02] <jbicha> Sweetsha1k: what's with the IRC nick?