=== gatox_tv is now known as gatox === gatox is now known as gatox_away [04:36] kenvandine, do you have a gdbus port for indicator-session? === dduffey is now known as dduffey_afk [04:50] robert_ancell, uh... it isn't using gdbus? [04:50] kenvandine, not in lp:indicator-session [04:51] I was about to write a patch but it seems you have a branch with some fixes like that [04:51] hummm [04:51] the communication between the service and the indicator has to be gdbus though... [04:52] kenvandine_, oh, I mean the communication with LightDM etc [04:52] ah [04:52] that old branch has been merged [04:52] so go for it :) [04:52] ok [04:52] please [04:52] :-D [04:53] everytime I come across the old bindings I groan and have to dig through the old docs [04:53] hehe [04:53] brb [05:45] jbicha: we got a freerdp 1.0.1 which is now ready for main; are you interested in pushing the update to the Debian git, or want me to do it on Monday? (I'm on vac today, need to leave soon) [05:49] pitti: I'll ping the Debian maintainer to upload the new version [05:49] and by the way, I don't think I can help myself from pronouncing your script as fat-race [05:57] jbicha: it only occurred to me a fair while after I set up the project :/ [05:57] * pitti is so bad with names [05:58] pitti: no, the name's fine, just embrace the second meaning, it should make designing a logo easier :) [05:58] heh [06:01] Good Morning [06:42] robert_ancell: can bug 897166 be fixed in precise quickly, so that the SRU can go to -updates? [06:42] Launchpad bug 897166 in lightdm "lightdm-gtk-greeter in Xubuntu, has one untranslated item" [Low,Fix committed] https://launchpad.net/bugs/897166 [06:43] pitti, does that just require updating to the latest version? [06:43] apparently the fix is in trunk [06:46] pitti, deal if you have a look at the last comment on bug 878836 and see if you agree [06:46] Launchpad bug 878836 in lightdm "Unity Greeter - Use Unity Greeter to fulfil lock screen as well as login functions" [High,In progress] https://launchpad.net/bugs/878836 === pitti_ is now known as pitti [06:48] pitti, deal if you have a look at the last comment on bug 878836 and see if you agree [06:48] Launchpad bug 878836 in lightdm "Unity Greeter - Use Unity Greeter to fulfil lock screen as well as login functions" [High,In progress] https://launchpad.net/bugs/878836 [06:48] pitti, why you always quit when I ask you a favour ;) [06:49] phew, starting the day by chasing during half an hour a kernel panic == not fun :) [06:49] robert_ancell: I just renamed [06:49] my IRC proxy timed out for some reason [06:49] robert_ancell: I'll have a look [06:49] hey didrocks [06:49] didrocks: eglibc fun? [06:50] good morning pitti :) [06:50] pitti: indeed, I reverted some packages yesterday on my system, seems I didn't revert enough of them ;) [06:50] robert_ancell: oh, I thought the "use lightdm for locking" blueprint would be for Q now? [06:50] didrocks: new eglibc should fix nvidia now [06:51] pitti: yeah, I used a livecd + chroot to see that the new version is available (and i confirm it works) [06:51] robert_ancell: we have enough bugs with VT switching that this still sounds not dependable enough? [06:52] robert_ancell: or would this avoid VT switching somehow? [06:52] pitti, the fix in the latest lightdm (just uploaded) seems to fix the switching issue for me at least [06:53] robert_ancell: well, let's just say I'm not very happy with the whole request, and I'd rather at least defer it to Q [06:54] but if you are eager to land this and have tested this extensively, I don't want to block you [07:05] pitti, did you get my 'what do you mean by "the request"?' [07:07] robert_ancell_: (no, I didn't) I mean the design team request to use lightdm for this, as long as this always requires VT switching [07:07] this sucks for multi-monitor configs [07:08] pitti, because the greeter doesn't match the same monitor layout? [07:08] (they aren't requesting lightdm for it, but they want the screen to be the same) [07:08] after entering your password, you are back in the session at instant right now; with that, it would take a second, or even several, and might involve different modes, resolutions, and output configs [07:09] robert_ancell_: yes, or you have different xrandr modes [07:18] pitti, only on certain drivers? === robert_ancell_ is now known as robert_ancell [07:18] vt switch takes about 5 secs here with my external monitor, and about 15 with the (arguably very slow) monitor of my wife [07:18] I haven't seen it to be much faster on other systems [07:18] well, these two cover intel and ati [07:18] hmm [07:18] I've never seen a switch that bad [07:18] and on nvidia, vt switching is inherently unstable [07:18] people like didrocks have eternal problems with running more than one session at once [07:18] robert_ancell: if we can get the unity greeter screen in the actual user X.org (i. e. just replacing the looks of gnome-screensaver), this would be fine, of course [07:18] but I guess not practical due to involving different users (lightdm vs. the actual user), etc.? [07:18] pitti, but that's a really hard problem [07:18] and brings a lot of risks [07:18] robert_ancell: well, even if it takes only two seconds and flickering, it's still much worse than instant and no flicker [07:18] get a better system :) [07:18] on a standard laptop with only an internal screen it should work much better [07:18] it's instantaneous for me here [07:18] but as soon as you have more than one output, it's very bad [07:18] robert_ancell: with an external monitor? [07:18] you didn't configure xrandr at all there? [07:18] it works nicely when I have the laptop undocked and in LVDS-only mode [07:18] it == vt switching [07:18] pitti, yeah, the external monitor is about 1s slower [07:18] pitti, what do you think about automatically locking displays when switching users? [07:18] I'm not sure whether xrandr mode changes are just poorly implemented, or whether there is a hardware reason why it's so slow [07:18] robert_ancell: we do already, don't we? [07:18] I think that's pretty much a must [07:18] pitti, the indicator voluntarily locks if you have screen lock enabled [07:18] if we do it in the DM automatically we don't check if the user requested it [07:18] pitti, so if you switched to your wifes account, then switched back to yours from the fast user switcher it would prompt you for a password where currently it doesnt [07:18] (if you had opted out of the screen lock) [07:18] robert_ancell: ah, I thought lightdm would kill the screensaver when you authenticate [07:18] it does [07:19] right, lightdm currently asks for my password, but not screenaver [07:19] but currently if you switch and you haven't requested a screen lock you can alt-ctrl-Fn to get back to your unlocked session [07:19] robert_ancell: oh, really? doesn't seem to work here [07:19] I disabled locking in control-center (I find it unnnerving when the screen saver kicks in) [07:19] but when I switch user, it always locks [07:20] ok, tried it, you're right [07:20] i. e. user switcher -> other account [07:20] ctrl+alt+f7 -> I'm in g-s-s [07:20] something must be picking it up automatically (CK?) [07:20] I believe it's the indicator [07:20] I think it doesn't really care for the option, it always pokes g-s-s to lock [07:20] but I haven't verified [07:20] pitti, the indicator code doesn't bother if the option is disabled [07:21] ok, what I thought [07:21] pitti, oh, I hear you're the new CK maintainer [07:21] mbiebl and I will be, yes [07:21] well, someone has to, as long as we refuse to use systemd [07:21] finally something to do in my CFT :) [07:22] copious free time? === vuntz_ is now known as vuntz [07:25] yes :) [07:25] anyway, I need to run out soon [07:25] I'm not even supposed to be here today :) [07:25] have a nice weekend everyone! [07:25] be gone! [07:25] I'll just annoy you with questions [07:26] robert_ancell: you can still ask a few, I'll do some packing and have another look at IRC [07:26] brb [07:34] pitti, robert_ancell: As long as lightdm manages to keep the same screen layout, VT switching should be ~instant, even with multiple monitors - it is here. On the free drivers, of course. No guarantees with !KMS drivers :/ [07:34] RAOF, so practically how many people aren't using the free drivers and what would the experience be for them? [07:34] hey RAOF and TheMuso and robert_ancell, how's it going tonight? [07:34] rickspencer3, hello [07:35] RAOF: it doesn't right now, though? the greeter session uses the default layout, and doesn't take the user's monitor.xml into account? [07:37] pitti: I know design's *plan* is to have it using the same layout as the last user; robert_ancell would know better if that's likely to be actually implemented. [07:38] that would certainly help a great deal [07:38] RAOF, we're aiming for the g-s-d defaults at the moment, so anyone who reconfigured their displays will have a mismatch [07:39] robert_ancell: Well, if users tick the “Install flash and other things you really want to install (hint: INSTALL THIS)” button during install they get the proprietary drivers. So, lots of people will be running the proprietary drivers. [07:39] * RAOF thinks this is a mistake for fglrx [07:40] RAOF, is there an easy 'these-drivers-are-crap' flag that we can read in gnome-screensaver and not do the vt switch in that case? === bryceh is now known as bryce [07:41] robert_ancell: Not trivially; it wouldn't be a terrible amount of effort to detect that, though. [07:43] Huh. In fact, the bgNoneRoot flag is *exactly* what you want to be checking for. Only problem is that it's not exposed outside the server at all. [07:43] RAOF, damn, you want to write an atom for that :) [07:43] ? [07:43] hey [07:43] I hear you want more extensions [07:44] Yeah. [07:44] I'm all about the extensions. [07:44] They're not painful *at all* [07:44] seb128, hey [07:44] (Hint: this is a filthy lie) [07:44] oh a robert_ancell! how are you? ;-) [07:44] hey RAOF [07:44] seb128: Yo! [07:44] so it's true, when you wake up early australians are online ;-) [07:45] hey seb128 [07:45] RAOF, sorry, that last comment came out all garbled. I'll take your other things at face value [07:45] seb128, what do you think about the lock screen VT switching. Too dangerous? [07:45] pitti, hey, wie gehts? [07:45] robert_ancell: Basically you want to check that a KMS driver is in use; there are a bunch of ways to do that. [07:45] seb128: gut, danke! [07:46] robert_ancell, if it's ready push it, ship beta with it, get feedback and see [07:46] robert_ancell, my gut feeling is that we will face driver issues and will need to revert for safety [07:46] yeah [07:46] robert_ancell: Actually, if you were to (say) query the modes from libdrm you could easily defeat pitti's problem and just use *exactly* the same modes for the lock screen. [07:47] robert_ancell, but having it as an optin for people who want it or tested oem install would be great [07:47] I've got it working in such a way that we could conditionally do it based on the drivers if RAOF can work out the right tricks [07:47] seb128, yes, exactly [07:47] RAOF, that would be very nice [07:47] RAOF, so we'd do something like X --keep-existing-modes? [07:48] We could do that. [07:48] That's been something that I've been meaning to implement for a while; it shouldn't be super-hard. [07:49] seb128, so I've been running that VT switch patch you had at the sprint for more than a week and it seems to solve it for me, so I've uploaded that now. Still haven't resolved the multiple switch lockup you reported [07:49] RAOF, that would be pretty awesome actually [07:49] Yeah, we should do that at startup anyway. [07:49] * RAOF even has a deferred work item for it. [07:50] RAOF, instead of mirror by default, just attempt to do "whatever happened last"? [07:50] qExactly. [07:50] RAOF, so working by monday? [07:51] So this would need to be gated on using KMS, but it shouldn't be *too* hard to not change mode. === statik is now known as 64MAAQOA7 [07:51] Hah! [07:51] robert_ancell, great [07:51] robert_ancell, I ran into the vt switch issue several times this week [07:52] (I've been using the guest session for testing stuff) [07:52] seb128, ok, good. Please look out and see if it's definitely improved for you [07:52] seb128, how are you at chasing up merges? https://code.launchpad.net/~robert-ancell/indicator-session/dont-lock-on-switch/+merge/92420 is blocking the greeter lock [07:52] robert_ancell, new lightdm,unity-greeter planned? ;-) [07:52] seb128, new lightdm is in the queue, greeter will have to wait until next week unless mterry wants to make the release (which is fine by me) [07:52] robert_ancell, will get it merged today [07:53] robert_ancell, I hinted mterry that he should step up for unity-greeter release, especially if you are not around or busy for some time [07:53] seb128, so, if you want to play there is a (disabled) patch in lp:~ubuntu-desktop/gnome-screensaver/ubuntu that changes over to doing a switch to greeter [07:53] seb128, and he ran away and blocked his ears? [07:53] robert_ancell, he said he would email you [07:54] ok [07:54] robert_ancell, he didn't do it I guess from that discussion ;-) [07:54] robert_ancell, so ok, I build indicator-session with your patch, gnome-screensaver with that patch enabled, what else? [07:54] that's it [07:54] oh ok [07:54] great [07:55] I will play with it today [07:55] and get the indicator side merged [07:55] it just calls SwitchToGreeter in LightDM. There's a new method in the next LightDM "Lock" which will hint to the greeter to go into lock mode, but we've dropped the UI changes for 12.04 anyway as there is no final design [07:56] seb128, what day is ff? [07:56] robert_ancell, thursday [07:56] ok [07:57] have you ever noticed google is awful at finding the release dates? [07:57] RAOF, (for real) is there any chance we could get that keep the same mode / driver filtering for 12.04? [07:58] robert_ancell, I rely on the firefox awesome bar for that :p [07:58] it finds https://wiki.ubuntu.com/PrecisePangolin/ReleaseSchedule just fine for me ;-) [07:59] robert_ancell, oh, we did get a new lightdm today, just noticed ;-) [08:01] robert_ancell, btw is that ok if I commit stuff to the gtk greeter directly rather than chasing you about those? ;-) [08:01] robert_ancell: I'm pretty confident I could do it by the 16th. [08:01] seb128, PLEASE DO! [08:01] robert_ancell, great ;-) [08:01] seb128, and we need someone to upload it to precise [08:01] robert_ancell: If it's a priority. [08:01] robert_ancell, well, the debian maintainer will upload once it works for him I guess [08:01] I'd really like it not to be me so they feel some ownership over it [08:01] robert_ancell, which should be the case now that you merged the config reading fix [08:02] cool [08:02] then we can sync [08:02] and let the xubuntu guys or whoever use it step up if they need patching [08:02] lubuntu have it in a ppa [08:14] hmmm, is there a reason why we don't enable the tee surface backend in cairo? [08:14] hey chrisccoulson, how are you? [08:14] seb128, yeah, good thanks. how are you? [08:15] I'm good thanks === bryceh_ is now known as bryce [08:18] chrisccoulson, debian did it it seems: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=609978 [08:18] Debian bug 609978 in libcairo2 "libcairo2: Please add support for tee surfaces in 1.10" [Normal,Fixed] [08:18] chrisccoulson, I didn't undo that in Ubuntu, cf https://launchpad.net/ubuntu/+source/cairo/1.10.2-2ubuntu1 [08:19] " * debian/libcairo2.symbols, [08:19] debian/rules: [08:19] - don't enable the tee option, it's not abi stable and not required" [08:19] chrisccoulson, I don't feel strongly about it, it just didn't seem required at the time and I don't like much maintaining stuff we don't use ;-) [08:20] especially when they are not flagged stable by upstream [08:20] yeah, we don't really need it. i just wanted to do a build of firefox with system cairo [08:20] so i can do a git bisect of cairo then without rebuilding firefox ;) [08:20] easy enough to enable [08:22] robert_ancell, I'm not convinced about your indicator-session change [08:22] seb128, yeah? [08:22] robert_ancell, it seems to increase the likeness that your screen will stay unlocked [08:22] robert_ancell, what if you use gdm? who is supposed to lock for you? [08:22] seb128, but if you try it it still locks without the change [08:22] I'm guessing ConsoleKit is doing it [08:23] with the change rather [08:23] hum ok [08:23] please confirm for me, I was suprised [08:23] robert_ancell, it's just a knee jerk reaction :p [08:23] I need to test and play with it [08:23] will do today and get an update by email [08:23] ok [08:23] I will also check with mdeslaur [08:23] he has been looking at screen locking for a while [08:24] well about making sure we don't let the screen unlocked rather [08:24] seb128, if it is risky, we can make it lock when not using LightDM (the new lightdm automatically locks for you) [08:24] robert_ancell, well it's early enough in the cycle to try this way [08:24] but I would not be surprised if it leads to some issues [08:25] I don't like much relying on "magic" to get screen locked ;-) [08:25] robert_ancell, I'm not sure to get why it's better to have lightdm doing the locking though [08:26] I would think lightdm should stay out of deciding stuff like that for you [08:26] it's user land config mostly [08:26] well the lock screen conceptually is the gateway to your session - so once you see the lock screen you shouldn't be able to subvert it [08:27] in the case of wayland we won't need a lock as lightdm will never allow you to interact with your session until your got past the greeter [08:28] seb128, it's not magic btw - lightdm is the only part of the system that *definitely* knows you've switched [08:28] ConsoleKit knows it second hand [08:28] robert_ancell, right, but if you i.e use gdm what does the locking and when? [08:29] robert_ancell, knowing that we don't want to lock if autologin is enabled for example [08:29] seb128, I believe CK detects the VT switch and locks the session you switched away from [08:29] because those users don't want to enter a password on resume [08:29] for example [08:29] that would be a good test [08:30] I suspect that if you autologin a user that does have a password, switch to guest then switch back their session will be locked [08:31] seb128, ah, I see what you mean - if you don't sleep (i.e. don't switch) then you shouldn't see a lock screen? [08:31] robert_ancell, "don't sleep"? [08:31] I mean if you sleep [08:31] robert_ancell, well if you have a netbook for family use and autologin you probably never want to enter a password [08:31] just close the lid and reopen [08:31] so yeah, what you said [08:32] seb128, and lightdm wont lock in that case (as there hasn't been a switch) [08:32] robert_ancell, ok, great ;-) [08:32] robert_ancell, will it lock if you don't use autologin? [08:32] because we do want our laptops to lock i.e at conferences [08:32] yes, because gnome-screensaver does the lock for you [08:32] excellent [08:33] that sounds good, as said I will play with it and see how it goes for me ;-) [08:33] if you press alt-ctrl-L, timeout or sleep gnome-screensaver explicitly locks [08:34] ok, signing off. cya next week === geser_ is now known as geser [08:35] robert_ancell, have a good w.e! [08:35] see you [09:08] in fact, http://hg.mozilla.org/mozilla-central/rev/713814f07168 makes it impossible to build with system cairo now [09:11] chrisccoulson, what's the issue with cairo? [09:12] oh, I guess the issue is how mozilla patch their copy in compatible ways? ;-) [09:12] seb128_, i want to git bisect cairo for mozilla bug 716036, which should be easier with system cairo [09:12] Mozilla bug 716036 in Layout "css z-ordering renders invisible text on google search page" [Normal,Unconfirmed: ] http://bugzilla.mozilla.org/show_bug.cgi?id=716036 === seb128_ is now known as seb128 [09:12] but yeah, they have patches which change their internal cairo API [09:12] yeah, I got what you were trying to do [09:13] i think they should just remove the build option for selecting system cairo [09:13] it just took me a bit to figure why using a system lib was a pain :p [09:13] heh [09:13] i've reverted their API change now. i don't really care about the bug they fixed with it ;) [09:14] that's not going to please other distro's, anyway [09:15] i wonder if they're even aware they can no longer use system cairo ;) === mimico_ is now known as mimico === fenris_ is now known as Guest89739 === gord_ is now known as gord === Guest89739 is now known as ejat [09:46] seb128, did we just get that nice new sound settings applet from Gnome today? [09:46] rickspencer3, no [09:47] rickspencer3, we got it from conor [09:47] really? [09:47] yes [09:47] they engaged with upstream but didn't manage to move the discussion enough to get the work landing upstream this cycle [09:47] so they landed it as a new capplet for Ubuntu [09:48] I hope they managed to get that work upstream for next cycle though [09:53] yay, bug reproduced with the latest system cairo [09:53] time to do some bisecting \o/ [09:58] Yeah, it's a nice redesign. [09:58] Is anyone tracking precise-problems? [09:59] Seems to be Desktop heavy atm, http://people.canonical.com/~ubuntu-archive/testing/precise_probs.html [09:59] webkit related awesomeness [10:00] Daviey, yeah, for the third time this morning on 3 channels, webkit needs to finish building on i386 so the -common arch all binaries are available [10:00] Daviey, it built on amd64 but not on i386 yet which creates that issue [10:00] seb128: right, scrollback hasn't been helpful this morning. [10:01] Daviey, see #ubuntu-devel 45 minutes back [10:01] did I say how mute I hate webkit today? ;-) [10:01] hate hate hate [10:02] Yeah, webkit causes us (server), so much woe :) [10:02] after 3 days of trying to get it to build now buildds hate it as well and people freak out :p [10:03] Daviey, the previous build (as several of my ppa updates this week) failed I guess the buildder killing ld because it was taking too long or something [10:03] I'm guessing because somebody retried the build before I woke up [10:03] but that's what happened to ppa several times this week [10:03] *awesome* [10:04] indeed === Zdra is now known as xclaesse === infinity_ is now known as infinity === MacSlow is now known as MacSlow|afk === tkamppeter_ is now known as tkamppeter [11:30] hello all! I'm updating my precise install and this caught my attention: [11:30] The following packages will be REMOVED: [11:30] apturl nautilus-share python-webkit software-center [11:30] software-center removed? :-/ [11:30] nessita, yeah, we switched back to synaptic [11:30] j/k ;) [11:30] heh [11:33] nessita, webkit built on amd64 before i386 which is leading to issue, wait for the i386 build it will sort itself [11:33] seb128: ack, thanks! [11:33] nessita, hey btw, how are you? ;-) [11:34] seb128: very good, though the FF is occupying a lot of place y my mind [11:34] you? [11:34] nessita, I'm great thanks! [11:34] buse as well === MacSlow|afk is now known as MacSlow === artnay_ is now known as artnay === greyback is now known as greyback|lunch [13:05] i need help for fixing bluetooth in my laptop.. === MacSlow is now known as MacSlow|lunch [13:06] I can't turn ON my bluetooth, in Ubuntu 11.10 [13:09] What version of python is in Oneiric? 2.6 or 2.7? [13:14] BigWhale: 2.7 on CD (and by default), 2.6 is also available in repositories [13:14] 2.6 is dropped in Precise [13:14] mitya57, thanks! That's excellent news! === ts2 is now known as tsimpson === ayan_ is now known as ayan [13:33] pitti, hi [13:35] tkamppeter, he's not working today, he took his friday off [13:42] seb128, thanks, skipping this to Monday ... === greyback|lunch is now known as greyback [14:01] hate hate hate webkit [14:02] "/usr/bin/ld: failed to set dynamic section sizes: Memory exhausted" [14:02] that's how the i386 built failed after 8 hours [14:05] seb128: fun times. fun times. at least you don't have to build it on windows :) === MacSlow|lunch is now known as MacSlow === stgraber_ is now known as stgraber === dduffey_afk is now known as dduffey === soren_ is now known as soren [14:58] didrocks, no need to be today but do you have any opinion on [14:58] https://code.launchpad.net/~andrewsomething/ubuntu/precise/compizconfig-settings-manager/disable_unity_checkbox/+merge/91375 [14:58] https://code.launchpad.net/~andrewsomething/ubuntu/precise/compizconfig-settings-manager/first_run_warning/+merge/91374 [14:58] [14:59] didrocks, andrew has upload right for ccsm but said he would welcome a review, the first mr1 stops from disabling unity from unity, the second displays a warning on first run [15:01] seb128, don't forget to look at the panel loading in g-c-c :) [15:01] kind of nice to nag seb128 for a change ;) [15:01] kenvandine, didn't I say yesterday I was unsure it was a good idea after all? ;-) joke aside I'm thinking about it but I don't like much opening a subpanel where you can't go back [15:02] it's a bit of a confusing experience [15:02] you did... but remember the corner case we are solving [15:02] upgraders that setup their chat accounts that way, empathy will send them to that panel [15:03] although i am open to other ideas [15:03] seb128: yeah, I append that for mon Monday list :) [15:03] kenvandine, right, I'm thinking maybe doing it the autostart way, conditional on a gsettings or something which would be set if you have an account configured with it [15:03] didrocks, thanks [15:03] (I saw the ML discussion this morning) [15:03] just can't today :p [15:03] kenvandine, so it would show in the control center only for those who have an account [15:09] achiang`, have you experienced any more hangs in chat? i haven't in a while [15:14] HI all, is it normal that the today's dist-upgrade would like to remove i.ex. ubuntu-desktop, sotware-center and rhythmbox-plugins ? [15:15] kenvandine, audio? [15:15] seb128, webkit [15:15] elvisd, "normal" if you are on amd64 yes, or rather known issue, webkit failed to build on i386 [15:15] seb128, yes i'm on amd64 [15:15] so the -common package from webkit is missing [15:16] kenvandine, ok, because the libc update from yesterday fixed audio issues [15:16] cool [15:16] kenvandine, right dunno if you read my [15:16] hate hate hate webkit [15:16] "/usr/bin/ld: failed to set dynamic section sizes: Memory exhausted" [15:16] that's how the i386 built failed after 8 hours [15:16] seb128, better to not upgrade? [15:16] ugh [15:16] kenvandine, the i386 builders don't have enough ram to build it [15:16] elvisd, don't let it remove that stuff :) [15:16] ld takes over 3gb [15:16] if the box is not pae it sucks [15:16] seb128, ok thank you :) [15:16] it's hitting the box limits [15:17] seb128, rock meet hard place! [15:17] kenvandine, indeed :-( [15:17] the builders should really be amd64, and just build for i386 [15:17] kenvandine, Colin sent it to a beefier buildd, let's see if that works [15:17] dobey: The choices for the start date in the credit card details window only go [15:17] up to 2011. [15:17] whoops [15:17] er [15:17] kenvandine, right... [15:17] dobey: where do ubuntuone music store bugs go now? [15:18] dobey: there's this one filed against banshee: https://bugs.launchpad.net/bugs/930117 [15:18] Launchpad bug 930117 in banshee "Credit card details" [Undecided,New] [15:18] hyperair: that should probably be filed against ubuntuone-music-store project [15:19] alright, i'll reassign it then. thanks [15:19] ok, thanks [15:21] jjardon, heyo! ping about https://code.launchpad.net/~mterry/indicator-power/show-indicator/+merge/92103 [15:22] mterry, I'm not sure jjardon is still maintaining that code [15:24] tedg, maybe you then? :) ^ [15:26] mterry, seb128, yeah I'll fix up the branches there [15:26] tedg, thanks === achiang` is now known as achiang [15:29] kenvandine: yesterday i still had hangs. i'll see about today [15:33] mterry: looks like that's bug 811769 [15:33] Launchpad bug 811769 in indicator-power "Power Menu should respect icon-policy hide setting" [High,In progress] https://launchpad.net/bugs/811769 [15:35] achiang, i haven't seen a hang in like 24hours... but could just be luck :) [15:35] nothing has changed that i know of [15:37] kenvandine: hm, ok. i haven't paid close attention, but i remember someone blogging about a new notify-osd yesterday [15:37] kenvandine: perhaps it's related to that [15:37] nope.. the chat window is it's own process now [15:38] kenvandine: ok. i'll just continue to monitor [15:38] good morning, world\n [15:38] achiang, thx [15:39] jbicha, is that icon-policy key still a thing? No schema installs it and I can't find GNOME code that references it [15:39] Ah, comment 16 confirms [15:59] DBO: ....daily bamf poke [16:23] hmmm, i'm struggling to work out how http://cgit.freedesktop.org/cairo/commit/?id=baaf312e047a9bea6f54e63cd6534c2ed7448523 could cause https://bugzilla.mozilla.org/show_bug.cgi?id=716036 only on some displays :/ [16:23] Mozilla bug 716036 in Layout "css z-ordering renders invisible text on google search page" [Normal,Unconfirmed: ] [16:54] i am really surprised gwibber is the only thing that depends on gir1.2-indicate-0.6 [16:55] kenvandine: why? [16:55] anybody knows what happens when you have two versioned typelibs installed at the same time? [16:56] the newest one gets used unless you explicitly request the older one first [16:56] that doesn't sound so bad... [16:56] assuming the new one doesn't totally break the api, it's probably fine [16:56] just trying to make the upgrade path not suck [16:57] is there a new indicate gir version? [16:57] libindicate5 (= ${binary:Version}), [16:57] dobey, working on it [16:57] gir1.2-indicate-0.7 and 0.6 have that [16:57] what is it? 0.7? [16:57] yes [16:57] so upgrading libindicate5 to a new versions forces gir1.2-indicate-0.6 to get removed [16:58] gwibber-service has a depends on gir1.2-indicate-0.6 [16:58] i guess you need to update gwibber also [16:58] awkward to change that without a soname bump [16:58] yeah, that is a bit odd [16:58] yeah... just thinking there should be a better way to handle it [16:58] like the gir could have a Provides [16:58] make the -0.7 breaks/replaces the 0.6? [16:58] that packages can depend on [16:59] it does that automatically by having the explicit binary:Version depends [16:59] when the lib updates it forces the old one to be removed [16:59] right, but the new gir doesn't get automatically installed; i guess replaces doesn't fix that? [17:00] good point [17:00] so it removes the old and anything that depends on it [17:00] right [17:00] it would be nice if it had a Provides, like gir-indicate [17:00] so both versions could provide that, and gwibber could depend on it [17:01] fwiw, we will Recommends: it next week in ubuntuone-client [17:01] seb128, ^^^ [17:01] what do you think? [17:01] kenvandine, do you I have to read screens of backlog? ;-) [17:01] no [17:01] virtual packages for girs would be nice [17:01] having to tweak the recommends every time unity has a new release is annoying [17:02] gir1.2-indicate-0.6 -> gir1.2-indicate-0.7 without libindicate5 getting a soname bump [17:02] packages that depend on gir1.2-indicate-0.6 need to be changed [17:02] kenvandine, hum [17:02] it would be cool if gir1.2-indicate-0.* had a Provides [17:02] that packages could dep on [17:02] to provide what? [17:02] a virtual package [17:02] gir-indicate [17:03] or something [17:03] if the abi number changed that's probably a good think that you need to transition [17:03] i think all gir typelib packages should do that [17:03] gir1.2-indicate, gir1.2-unity [17:03] seb128, but the versioned deps on libindicate5 force the gir to get transitioned for you [17:03] but without changing packages that depend on the old one, those will get removed [17:04] gir1.2-indicate-0.* have this: [17:04] libindicate5 (= ${binary:Version}), [17:04] kenvandine, we should maybe change the depends to be >= ${binary... [17:04] kenvandine: patch it to bump the soname :P [17:04] so libindicate5 upgrade removes the old one [17:04] kenvandine, do a rebuild of the current version with that [17:04] kenvandine, wait for it to be published and then upload the new one [17:04] but i think it would be a good idea to make the old version get removed [17:05] i guess that should be fine [17:05] until there is a soname bump [17:05] don't you love gir? :( [17:05] the soname change doesn't break anything [17:05] you would keep the old gir and old soname [17:05] right [17:05] imho just change the = to >= [17:06] that's the easiest way [17:06] we can discuss a better way later [17:06] ok [17:06] this is probably kind of a corner case, usually these changes come with soname bumps [17:06] right [17:06] it's weird that the gir abi changed when the lib abi didn't [17:07] it didn't really... it is just to keep the versioning similar [17:07] i think [17:07] indicate-0.7 for pkgconfig [17:07] those versioning are screwed up imho [17:07] they don't serve any purpose [17:08] agreed [17:08] well anyway we will not "fix" that today ;-) [17:08] yup [17:08] "_ [17:08] :) [17:09] seb128: the API changed, not the ABI; perhaps it has new API, but all the old API is still the same, and thus the ABI is fine [17:09] the -N.M is API not ABI, version [17:09] or at least, that's how it's supposed to be [18:16] webkit failed on i386 due to symbols changing which is troubling [18:28] micahg, they didn't change as said on #ubuntu-devel, it's cross arch mangling and cpp on a new symbol [18:28] which I didn't see because I didn't build it on 2 arches there [18:28] got enough difficulties getting it to build on one :p [18:28] heh, ok, that's good then [18:29] * micahg was worried about ABI breaks [19:44] seb128, mind if i upload your unreleased changes in g-c-c? [19:44] kenvandine, stealing my bugs!!! [19:45] hehe :) [19:45] refreshing ronoc's patch [19:45] kenvandine, you got lucky, I just walked back from dinner, feel free to upload thanks ;-) [19:45] kenvandine, you can assume that if I commited something without warning in the changelog that it's fine to upload btw [19:45] seb128, libindicate is in binNEW and everything that needs a rebuild has been uploaded [19:46] seb128, if you weren't online i would have just uploaded it :) [19:46] kenvandine, no it's not ;-) [19:46] oh, not yet? [19:46] oh, not all built yet [19:46] kenvandine, not anymore you mean? ;-) [19:47] oh [19:47] kenvandine, newed indicator-applet as well [19:47] you're sneaky [19:47] guys i did a new session indicator release during the week aswell [19:47] kenvandine, I stole the joke from pitti to be honest ;-) [19:47] hehe [19:47] ronoc, you know it is friday already right? :) [19:47] oh, i think i uploaded that already [19:47] kenvandine, yep just waiting on something else here [19:48] kenvandine, today ? [19:48] i'm still on 0.3.90 [19:48] ah [19:48] https://launchpad.net/indicator-session/+milestone/0.3.91 [19:48] 2 releases this week :) [19:48] kenvandine, you don't look at versions enough ;-) [19:48] i'll do .91 in a few [19:49] no panic [19:49] it fixes that apt menuitem [19:49] kenvandine, that's a 1 commit release, let's see what could go wrong in one commit :p [19:49] seb128, i've had plenty of backload without looking at the versions page :) [19:49] seb128, lots :)) [19:49] ronoc, I bet ;-) [19:49] i have special skills come into play [19:49] how many memory leaks can you introduce in one commit [19:50] ronoc, not sure I've enough finger to count that far :p [19:51] seb128, don't know i have lost count :-)) [19:51] ronoc, 1 commit... but not "small" [19:51] ronoc, btw it's 8pm on a friday for you, what are you still don't there? [19:51] 2 files changed, 185 insertions(+), 397 deletions(-) [19:51] kenvandine, yep focused work [19:52] seb128, waiting on friend coming from the airport [19:52] no panic [19:52] i started a little late [19:52] finished now anyway with ubuntu work [19:52] ronoc, btw robert_ancell said he is going to port the rest of indicator-session to gdbus [19:53] kenvandine, brilliant !, tedg did you hear that robert_ancell has volunteered to maintain i-session [19:53] lol [19:53] haha ;) [19:53] :) [19:53] he just needs it for the greeter [19:54] it's like a hot potato that project [19:54] ronoc, congrat btw, by the same principle you just won gnome-control-center [19:54] I love it when Australia is on weekend and we can volunteer them for things! [19:54] I still we still win ;-) [19:54] ah shit [19:54] * tedg still wins even then ;-) [19:54] just the sound panel seb128 [19:55] ronoc, no no, it's a by source game [19:55] i don't know about anything else :) [19:55] you got the whole package [19:55] the beast that is [19:55] it takes about 10 minutes to branch [19:55] well [19:55] ronoc, so what should i focus on when i test the package? [19:55] lucky you, you don't have to build webkit [19:55] true [19:55] I spent half my week waiting on builds from it [19:55] and Design doesn't think we have enough patches in control center yet :) [19:56] kenvandine, do you have a usb headset ? [19:56] it fails to build on the i386 builders because they don't have enough ram [19:56] nope [19:56] kenvandine, hmmm [19:56] but i meant indicator-session anyway [19:56] :) [19:56] oh [19:56] ronoc, like you wait 8 hours to see that ld got killed because it was using over 3gb ram [19:56] i had one... but my son ripped the wire out of it [19:56] destructive little guy [19:56] seb128, wow [19:56] 3 headsets he destroyed... and he is only 3.5 years old [19:57] kenvandine, not afraid to explore how things work ! [19:57] ronoc, I learnt that most buildds don't use pae and don't like you using over 3gb :-( [19:57] kenvandine, you need to hide your stuff better! [19:57] kenvandine, i.e. rip it apart and see from there [19:57] exactly [19:57] high shelves [19:58] :) [19:58] jbicha, btw I looked at gsd today, I think it would be easy enough to support both gsettings and gconf keybind according to the session in use [19:58] jbicha, I need to look at the g-c-c side still though [19:58] ronoc, gnome-control-center uploaded [19:58] i remember i smashed open the first LED watch to come into our house - it was my older brothers - stole from him and smashed upon the screen and took at the LED panel just to see what 'he was doing in there' [19:58] jbicha, we could maybe hack up a patch that would allow to update gnome-shell [19:59] kenvandine, on i-session the apt menuitem should now be working correctly [19:59] "correctly" [19:59] it should report correctly when updates are available and when you are up to date [19:59] like is there a way for me to trick it into thinking there are updates from packagekit? [19:59] kenvandine, no [20:00] kenvandine, you will have to wait for an update :) [20:00] ok, i can stick something in my local repo :) [20:00] kenvandine, anything would be better than how it currently is [20:00] kenvandine: add a ppa and refresh? [20:00] ronoc, the only new build dep shoudl be libpackagekit-glib2-dev [20:00] right [20:00] kenvandine, yep [20:00] dobey, i have a local directory for an apt repo [20:01] well aren't you cool :) [20:01] * kenvandine wishes he could install that without packagekit :) [20:01] kenvandine, oh yeah don't pull packagekit [20:01] dobey, it is a must for testing apt upgrades for transitions [20:01] * kenvandine blames tedg [20:01] ronoc, i won't [20:01] i'll just trust pbuilder [20:01] cool [20:01] seb128: cool, big project for next week then [20:01] i just wish pbuilder would let me say "use this PPA to satisfy dependencies for this build" [20:02] jbicha, let's see what I can get rolling before feature freeze ;-) [20:02] seb128: and it sounds like you've won g-c-c and g-s-d :) [20:02] jbicha, I'm done with other stuff on my list so next week is for g-s-d and g-c-c [20:02] kenvandine, ? === s9iper1_ is now known as s9iper1 [20:02] jbicha, hum, is that really "winning"? ;-) [20:02] kenvandine, pk is all ronoc [20:02] i just worry about someone else running "apt-get build-dep indicator-session" [20:02] seb128: it is for the rest of us! [20:02] hehe [20:02] kenvandine, is that not then a mvo issue [20:03] tedg, i was blaming you for me having to test apt upgrades on package transitions [20:03] kenvandine, which they will [20:03] I don't always test my package transitions, but when I do, I do it on kenvandine's machine. [20:03] kenvandine, I blame ted for not knowing what "stability" means [20:03] seb128: do you want to use the existing desktop branches for g-c-c/g-s-d or should we use separate branches for now? [20:04] jbicha, for what? what I will look at next week or..? [20:04] jbicha, if you want to update to the unstable serie please create a new vcs [20:04] jbicha, I'm not sure yet if I want to backport patches and tweak or upgrade and back out patches or ... [20:05] jbicha, I need to think a bit over it and look a g-c-c on monday [20:05] seb128, haha :) [20:05] tedg, libindicate and all the rebuilds are uploaded [20:05] jbicha, but either way it should be doable to adapt the keybinding patch to do gconf or gsettings according to the session [20:05] seb128: we could use the gnome3 branches for that then [20:05] first time ever pidgin-libnotify didn't need me to fix an unrelated issue :) [20:05] jbicha, one other option is to do what kenvandine and ronoc did [20:06] kenvandine, Cool! Thanks! [20:06] branching and merging g-c-c is a pain because of quilt [20:06] i.e have 2 different .so, one being the old gconf one, one being the new gsettings one, and load the right one for the session [20:06] yup [20:06] ronoc, nice work on the sound settings btw :) [20:06] chrisccoulson, thanks Chris ! [20:06] getting there now hopefully [20:07] ronoc, do you plan to change the default tab btw? I've never understood why the Sound Effects tab is the one which is active when you open it :) [20:07] chrisccoulson, oh yes forgot about that - output first [20:07] excellent, that's good :) [20:07] ronoc, +1 on that! [20:08] cool, we'll do it next week [20:08] sweet [20:08] kenvandine, when do you guys want to freeze the repo [20:08] wednedayish [20:08] or do you freeze for FF [20:09] i can't remember [20:10] chrisccoulson, ronoc, kenvandine: I fixed that first tab bug in precise sound capplet [20:10] so I blame it on ronoc from dropping that patch his the merge :p [20:10] output was first before the update [20:10] seb128, oh sorry about that [20:10] heh [20:10] seb128, i seem to remember that being that way before [20:11] seb128, doh! [20:12] ronoc, your (incorrect) order was this way until a few weeks ago when I backported the git patch to make "output" first [20:12] ronoc, no worry it's a detail ;-) [20:15] ronoc, ok... not seeing that there are updates available [20:16] kenvandine, it takes 60 seconds after it starts up [20:16] pitti, didn't want it checking straight away [20:16] ah [20:16] effects startup time [20:16] makes sense [20:17] what about after i do an apt-get update? [20:17] seb128, cool [20:17] kenvandine, i don't think command line apt-get brings up the packagekit dbus api [20:17] right [20:17] so indicator-session will bring that up [20:18] but what sends the signals that there are updates? [20:18] or does the indicator query for updates? [20:18] it will check with the package kit glib 1 minute after it starts [20:18] ok, still nothing [20:18] kenvandine, are there updates available [20:18] ? [20:19] 3 minutes and i have 4 updates availalbe [20:19] hmmm [20:19] according to apt [20:19] it's ubuntu, there are always updates! [20:19] update-manager did fire [20:19] weird it worked fine here [20:19] packagekitd is running [20:20] why did you install that? [20:20] the goal is to use aptdaemon [20:20] dunno === fenris_ is now known as Guest28407 [20:20] --no-install-recommends ;-) [20:20] I wonder if pk and aptdaemon conflict in some way [20:21] oh [20:21] libpackagekit-glib2-14 depends on gnome-packagekit [20:21] and packagekit [20:21] ronoc, said it recommends it [20:21] ? [20:21] recommends get installed [20:21] right [20:21] --no-install-recommends [20:21] yep [20:22] kenvandine, that's what ronoc meant earlier that we should talk to mvo about fixing the recommends issue [20:22] but I guess that's for next week [20:22] seb128, i think they do [20:22] so maybe hold on this upload until monday [20:22] seb128, kenvandine i spoke with mvo during the week about this [20:22] definately [20:22] pk will cause problems if running [20:22] i thought it was just for the build depends :) [20:23] sounds sensible [20:23] * ronoc mutters control center under his breadth [20:23] did i just say that === Guest28407 is now known as ejat [20:24] oh new treats from the mastering house [20:25] i really need to start mastering my own stuff [20:25] they get me every time - i just need to buy a manley compressor [20:28] okay guys i'm out - good w/end all! [21:43] micahg: ping. webkit is broken on amd64 only it seems? [21:44] dobey: new breakage or just not being able to install stuff (arch skew) [21:44] micahg: looks like 1.7.5 is not installable on amd64 (and isn't available on i386) [21:44] new builds are running now, should be ready in a few hours [21:44] i386 was broke for ~ 1 day [21:44] ah ok [21:45] seb128: fixed it about 3 hours ago [21:45] seb128: sorry, didn't mean to ping [21:45] dobey, that's what I was complaining about earlier [21:45] micahg, no worry, I close IRC when I don't want to be pinged ;-) [21:46] ok [21:46] I'm still around though I will call it a day in a bit [21:46] seb128: lucky you :) [22:08] 'night === dduffey is now known as dduffey_afk