[00:01] jasoncwarner_: Did you do the evtest run I suggested? [00:01] RAOF: oops...I didn't...doing it now [00:02] Make sure you note which horizontal direction corresponds to which button-press in evtest (ie: do something like move all left, then all right, and say that on the bug ☺) [00:18] jasoncwarner_: ping? === maxb_ is now known as maxb [02:10] Transmission 2.50 has been released... it's not on the website yet because we're still waiting on localizations on the Mac side, but I wanted to make sure the tarball was out before Precise's feature freeze [02:10] so: http://download-origin.transmissionbt.com/files/transmission-2.50.tar.xz [02:51] Oh wow, now unity-greeter deals with the lock screen. Now it just needs to respect my monitor config, which is what was being talked about earlier today I guess. [02:56] charles: thank you [02:56] TheMuso`: Yeah, just noticed that :) [03:09] RAOF: is it possible to sync from Debian's new queue? [03:10] jbicha: No. IIRC, it's not possible to pull packages *out* of the new queue unless you're an ftp-master, as packages in NEW are not yet confirmed to be redistributable. [03:13] ok, I guess I'll wait a bit longer... [03:58] hmm, I wasn't expecting this: http://iloveubuntu.net/redesigned-nautilus-34-probably-land-precise-pangolin-next-weeks [03:59] I wonder what the source is for that, since I don't see those changes in git [04:00] jbicha, they also wrote that google+ plugin for gwibber was likely to land in time for 12.04 [04:00] all i did was push a branch that handles authorization [04:00] haha [04:00] stalkers :) [04:00] no where near capable of doing anything useful, google doesn't even provide a useful API yet :) [04:01] that post reads like rampant speculation and wishful thinking [04:01] indeed [04:01] they are trying to get the scoop on everyone else [04:01] by guessing [04:02] oh good, GNOME UI Freeze is next week [04:02] kenvandine, uhhh has the lock screen changed dramatically recently? [04:02] yup [04:02] unity-greeter [04:03] kenvandine, it totally fucks with nvidia drivers [04:03] oh? [04:03] that isn't cool [04:03] yeah like it now takes it 30+ seconds to come out of the lock screen [04:04] * kenvandine hugs intel :) [04:04] is it possible for me to get the old lock screen back [04:04] or disable it [04:04] DBO, no idea really, last i heard that feature might not make it [04:04] and today i noticed i had it [04:05] that is all i know [04:05] can I disable teh lock screen all together? [04:05] so I dont walk away for 5 minutes and then come back and have to wait [04:05] yeah [04:05] Yeah, turn off the screen lock :) [04:05] hey RAOF [04:05] ah found it [04:06] RAOF, you experiencing what I just described at all? [04:06] RAOF, do you know anything about how its implemented or why its so different? [04:06] DBO: It's doing a VT switch [04:06] it feels like the GPU is doing a full out memory swap [04:06] DBO: So nvidia, being the mad bastard that it is, hates you. [04:07] vt switch should be fast as hell [04:07] its not [04:07] Not on nvidia [04:07] * kenvandine is glad he doesn't understand X [04:07] :) [04:07] kenvandine: If what I'm hacking on right now works out, vt switches (on intel, radeon, nouveau) should be entirely instantaneous. [04:08] why are we doing a VT switches at all? [04:08] Because we're switching back to the greeter. [04:08] why? [04:08] that *really* limits what we can do with it [04:08] that is handling the lock [04:08] that has been the plan since we decided to use lightdm [04:08] A number of reasons - (a) the login process is then the same, (b) it's trivial to switch to a different user, (c) it's more secure. [04:09] (d) it means we will never be able to transition between the log in screen and the desktop [04:09] further [04:09] it will always be slow as hell on nvidia [04:10] we need a way to use the old lock screen for nivdia users [04:10] d) we need wayland... [04:10] Which also doesn't have a way for nvidia to play, but whatever :) [04:10] right no, fucking over nvidia is not acceptable [04:14] RAOF, was nvidia screwed even if the displays were set to the same modes? [04:14] robert_ancell: Dunno [04:15] I don't think we're going to be able to enforce that constraint anyway. [04:15] k [04:16] X sets a mode during startup, and we can't change nvidia to read out the mode that was previously set. [04:17] RAOF, so does starting two nvidia X servers cause a mode switch or do they recognise they've already set the video mode [04:17] robert_ancell, mode switch [04:17] I can see it doing it [04:18] DBO, so you started an empty X server, then from that started a second one? [04:18] what? [04:18] no when I log in [04:18] I can see it mode switch [04:18] my monitor pops information about what mode its in [04:18] Do you have more than one display? [04:19] yes [04:19] DBO, yes, but that could be because the greeter isn't running the same mode as the session [04:19] I think what robert is asking about is - if you have your session, and a guest session with the same monitor configuration, does switching between them imply a modeset. [04:19] robert_ancell, I have saved my xconfig to the xorg.conf if that helps [04:19] that would do it [04:20] only works if you haven't configured your user from the system defaults [04:20] ? [04:20] well let me put it this way: [04:21] its not wholly correct to say VT switch is the whole problem [04:21] if I switch from X to VT1 [04:21] it takes about 3 seconds [04:21] and is relatively painless [04:21] switching back also takes about 3 seconds [04:21] lightdm switching however [04:21] that takes 30+ seconds [04:22] and often causes compiz to crash inside the nvidia driver [04:25] robert_ancell, ^^ [04:25] yup [04:25] its very similar to the delay nvidia incurs when switching resolutions [04:25] I would guess its probably the same [04:26] And given that it *is* switching resolutions... [04:26] Oh, it's not; you've encoded that stuff in xorg.conf. [04:27] yeah [04:27] whatever is going on though [04:27] is BAD juju [04:27] it breaks the experience [04:28] there should be testing around this :/ === TheMuso` is now known as TheMuso [04:36] RAOF, you here? [04:36] Yup. [04:36] I want to fix this nivida thing once and for all [04:36] I dont understand why we get stuck in ioctl [04:36] Oh, the slowness on modechange. [04:36] I dont understand why it only happens the first time [04:36] (except lightdm, that does it every time) [04:37] I have 1.5GB of vram [04:37] so even if that was *full* [04:37] it wouldn't take 30 seconds to completely swap that out [04:46] Sorry about that; I accidentally unplugged my router. [04:46] DBO: Yeah, 30 seconds seems like a long time. Unless it's paging to disc :) [04:47] not likely [04:47] I have 16GB of ram [04:48] :) [04:49] * RAOF shall make a coffee, upgrade an nvidia netbook, and ponder. [04:51] I have a test program that is relatively minimal [04:51] it makes it happen on nvidia [04:51] even with metacity [04:51] it just creates 40 800x800 textures with 20 FBOs [04:51] and binds them all and paints them [04:52] *eventually* the damn thing recovers [04:52] but it takes it about a minute [04:53] Good morning [04:53] pitti, are you a dbus master? [04:54] robert_ancell: I'm using the API and the CLI tool; I guess "apprentice" might be right? :-) [04:54] pitti, I'm trying to get the uid of a dbus caller. It seems a lot harder than it should be [04:54] hmm, how do I get the window menu these days? [04:55] robert_ancell: hang on, I'll look for example code; some projects I work on use this [04:56] pitti, oh good. The internet seems to have nothing. I'm looking to policykit to work it out [04:56] yes, that project would certainly need to do that [04:56] robert_ancell: There's a libdbus call - uid_for_caller, right? [04:57] I'm calling org.freedesktop.DBus.GetConnectionUnixUser - RAOF, I guess that is the same thing, and it's just not exposed in gdbus? [04:57] there's lots of decoy methods in gbus that sound like the right sort of thing but they don't seem to be any help [04:57] yes, that sounds right [04:58] robert_ancell: that's what PK does, anyway [04:58] robert_ancell: you can call this through gdbus just fine? [04:58] RAOF, do you want the test program? [04:59] pitt, yes, It's just finding any information that that is the correct method [04:59] robert_ancell: http://paste.ubuntu.com/842637/ [04:59] http://dbus.freedesktop.org/doc/api/html/group__DBusBus.html#ga24d782c710f3d82caf1b1ed582dcf474 [04:59] robert_ancell: ^^^ [04:59] RAOF, ta [04:59] DBO: Yes, please [04:59] pitti, ta also [05:00] RAOF, http://paste.ubuntu.com/842639/ [05:01] DBO: And that takes ages the first time after a modechange? [05:01] RAOF, AGES [05:02] RAOF, it screws up the GPU so bad metacity stops painting [05:02] and so does everything else [05:02] Whee! [05:02] you just get a bunch of white for a long time [05:03] pitti, and the trick is to use g_dbus_method_invocation_get_sender () I think [05:04] DBO: And I can build that with g++ $(pkg-config --libs nux), right? [05:04] RAOF, I think so yeah [05:04] RAOF, honestly there is a framebufferobject.cpp in the nux gputests/ [05:04] I just modified it a bit [05:05] to you know [05:05] make it evil [05:05] :) [05:05] Well, intel doesn't like it very much, but is perfectly happy to draw it. [05:06] I imagine its eating all of your vram [05:06] I have 16GB of vram! [05:06] no [05:06] you have 16GB of system ram [05:07] that your driver can map [05:07] Minus what the system is using, of course :) [05:07] but its freaking slow [05:07] Right. [05:07] how unhappy is it? [05:07] It slows down unity a bit. [05:07] yeah [05:07] there is no rate limiter on this [05:07] Alt tab isn't quite so responsive. [05:07] its just rendering as fast as it freaking can [05:08] which is rebinding 40 textures each frame [05:08] it eats all your GPU [05:08] The window doesn't much like being dragged, either. [05:08] when you do that it resets the textures [05:08] But it handles it at > 1/sec [05:08] I meant to disable that [05:08] what? [05:09] - [05:09] Dragging it is really slow. [05:09] *really* slow; unity draw the move at ~1fps [05:10] here [05:10] But just having it up and visible is actually barely noticable [05:10] http://paste.ubuntu.com/842647/ [05:10] that fixes the dragging slowness [05:10] Oh, maybe the alt-tab switcher is 15fps when rendered over the top of it. [05:10] thats probably due to the damage events [05:11] Sandybridge doesn't fall over on your fbo test. [05:11] nope [05:11] nvidia only [05:16] robert_ancell: ah, so you landed the lock screen stuff after all .. did you figure out how to tell lightdm to not change XRandR? [05:16] pitti, no, I think RAOF says it might not be possible [05:16] pitti, it broke everyone using nvidia blob :/ [05:16] eek [05:16] can we please revert this then? [05:17] robert_ancell: Oh, no. It's entirely possible to tell lightdm not to change XRandR. [05:17] It's just not possible to tell *nvidia* not to modeset. [05:17] That was all I was promising from the beginning - not modesetting on kms drivers. [05:17] changing VTs is bad enough and already breaks nvidia, but changing xrandr modes will make it suck for all other drivers using more than one screen [05:18] so far people on these drivers just learned the hard way to not do user switching [05:18] Yeah, we can make it work for !nvidia, !fglrx. [05:18] but now this forces it upon them [05:18] pitti, this is about more than just user switching [05:18] pitti, I have to run out for about 1-1.5 hours, but will be back after that [05:18] my lock screen now takes 30+ seconds to get out of [05:19] DBO: yes, I know; that's the xrandr part [05:19] no [05:19] I have my xorg.conf hardcoded [05:19] I'm afraid we simply don't yet have the technology to do this [05:19] (for locking screen, I mean) [05:19] robert_ancell: ok, thanks [05:19] pitti: nvidia does something weird on VT switch, like dropping all GL state or something. [05:20] pitti: You also see this after modesetting; unity takes a long time to draw anything after modesetting; sometimes so long that you can't press the “do you want to keep this mode” before it times out. [05:20] RAOF: no need to convince me :) I know how broken it is on multiple levels [05:20] (Which is clearly hilarious) [05:21] RAOF, technical question for you [05:21] RAOF: my monitor takes some 5 seconds to get back to its senses after modeswitch; my wife's display takes 30 (!) [05:21] texture A == made by generating a texture and then loading a png into it [05:21] texture B == made by generating a texture, binding it into an FBO, and painting A into it [05:21] then unbinding it [05:21] what makes these textures different? [05:22] No idea. [05:22] Sorry :) [05:24] pitti: Is it your wife's monitor, or the nvidia GPU being crazy after modeset? :) [05:25] RAOF: the monitor is excessively slow [05:25] RAOF: it's using the ati driver (the free one) [05:25] I don't have nvidia hardware in the house [05:26] hm, just dist-upgraded and rebooted to test this [05:26] tons of flickering, resetting xrandr to X.org's defaults, turns the LVDS back on, and lots of flickering when switching back [05:26] this is not a nice experience :/ [05:28] I can make it better. [05:43] robert_ancell: the indicator-session change also introduces a regression, I followed up to bug 878836 [05:43] 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 [06:03] RAOF, have you tried out the test app on nvidia? [06:03] RAOF, on my system the nvidia driver takes a full minute to change resolution and begin drawing the app again [06:03] My test system is upgrading; the last attempt to get into X failed :( [06:04] I've been trying a lot of things to hack around the slowness [06:04] unbinding the textures/using a single fbo instead of many... [06:04] nothing works [06:06] Worst case, we can throw that at nvidia via tselliot as a test case. [06:24] RAOF, you think they will take this as a test case? [06:24] RAOF, do you think its possible the driver is converting the texture formats? [06:25] good morning [06:26] bonjour didrocks [06:26] guten morgen pitti [06:26] DBO: Yes; the driver's allowed to do all sorts of things. [06:26] Oh, huh. [06:27] DBO: Does it still happen if you don't use RGBA? [06:28] RAOF, I was just testing that [06:28] RAOF, and yes [06:28] it does still do it [06:28] I also tested BGRA [06:32] pitti: Good morning, Martin. Do you have time to take a look at the MP linked to bug 928400? (It was preceded by an anything but constructive discussion on the bug report with ScottK.) :( [06:32] Launchpad bug 928400 in language-selector "systemsettings crashed with AttributeError in onSystemLanguageApply(): 'QtLanguageSelector' object has no attribute 'writeSysLangSetting'" [Undecided,Confirmed] https://launchpad.net/bugs/928400 [06:33] RAOF, the recovery time does scale perfectly with the number of textures we bind into the FBO [06:34] GunnarHj: You know, I'd be more than happy to drop language-selector-qt [06:35] pitti: But don't write that on the bug. ;-) [06:36] GunnarHj: I just did :) I think I was reasonably diplomatic [06:37] self.writeSysLanguageSetting(code) [06:37] 22- self.writeSysLangSetting(code) [06:37] GunnarHj: ah, so the first one is enough now? [06:38] GunnarHj: I guess originally the first wrote $LANGUAGE, the second $LANG [06:38] RAOF, any other ideas before I give up and go postal? [06:38] GunnarHj: but now the ls-dbus-backend does both? [06:38] DBO: I don't at the moment, no. I suggest sleeping instead of going postal; I'll keep poking. [06:40] GunnarHj: sorry that you got in between this [06:40] GunnarHj: right, just saw the code; thanks! [06:42] pitti: Sorry, went to get coffee. Is there still any clarification you need? [06:42] GunnarHj: nope, I'm good; I merged with dropping the "needs clarification" bit from the changleog [06:42] GunnarHj: thanks muchly! [06:42] uploaded [06:43] pitti: Ok, thanks! [06:46] RAOF, http://www.nvidia.com/object/linux-display-ia32-260.19.29-driver.html [06:46] check THAT out [06:46] DBO: Ah. [06:46] RAOF, do you think they regressed? [06:47] RAOF, is it possible for you to test that driver? [06:47] Yes, I can test that driver. [06:47] * DBO waits with baited breath [06:48] RAOF, it turns out my google-fu is strong today [06:48] I suggest sleeping first. It'll take a little while for me to actually install those drivers. [06:48] Or is it the morning for you? ☺ [06:48] Either way, sleeping first. It's never wrong to sleep! [06:49] Faster, laptops! FASTER! [06:50] RAOF, I'd not be able to sleep knowing that this bug might finally have a solid lead [06:57] pitti: Saw your bug comments. Please keep in mind that there are other derivatives (Lubuntu, Xubuntu) that use the gtk part of l-s, and want do keep doing so (at least for a while) when we switch to g-c-c. [06:58] GunnarHj: right, but same comment -- I won't be interested in maintaining the GUI any more [06:59] just the language_support_pkgs.py module [06:59] which we might very well split out into its own source then [06:59] then the rest can go to universe [07:00] pitti: Guess it will be up to the developers of those derivatives what to do, then. [07:02] right [07:02] well, it's not like we leave it behind completely broken [07:02] this cycle it got a lot of cleanup, which should make it a little less ugly to maintain [07:04] pitti: True. [07:05] pitti: next cycle I'd like to chat with you about alternatives to language-selector :) [07:07] pitti: I wrote a reply on the im-config/im-switch topic, with a suggestion on how to deal with it. https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2012-February/013316.html [07:07] GunnarHj: I saw, thanks! so this does change behaviour and current configuration after all? [07:07] pitti: Yes. [07:08] GunnarHj: it sounded like it would configure input support more automatically [07:08] which would be a nice simplification of the GUI as well [07:08] also wrt. control-center and other potential replacements [07:09] this is pretty much the only reason why we still use l-s [07:10] pitti: If we leave the code aside, unlike im-switch it does not care about the current locale. Good or bad? I'm not sure. [07:10] GunnarHj: oh? that sounds "bad" to my untrained ear [07:11] I thought the point was to automatically select an appropriate IM for the current locale === tkamppeter__ is now known as tkamppeter [07:11] (LC_CTYPE presumably) [07:11] pitti: That's the point with im-switch. [07:11] that's how I understood it anyway [07:11] i. e. it has a locale -> IM method map [07:12] and starts ibus, and the preferred method in those locales [07:12] and with the indicator the user can then switch between methods [07:12] pitti: Yep. [07:12] l-s just allows users to fine-tune this, but it shouldn't be necessary for most use cases [07:12] GunnarHj: ok, so at least the two of us have the same mental model here? [07:13] pitti: So it seems. [07:15] pitti: I don't know the rationale behind ignoring the language/locale in im-config, but I know that some CJK users want to be able to use ibus even if they have e.g. English as the display language. After all, display language and input method are two different things. [07:15] right [07:16] GunnarHj: should they have LC_CTYPE=zh_CN.UTF-8 and LC_MESSAGES=en_US.UTF-8 ? [07:17] I think in the new world we don't set these any more, they are derived from $LANG [07:18] pitti: No, no, currently it works just fine with im-switch. AFAICT there is no urgent need to change anything. [07:19] pitti: I suggested that we wait with converting to im-config because we don't know yet how GNOME will handle it in g-c-c, so by waiting we would avoid to expose the users twice to changes in the input method handling. [07:21] pitti: As regards which variables im-switch use, it's LC_CTYPE and if it isn't set LANG. [07:23] micahg: Out of curiosity, do have anything else bug the region module in g-c-c in mind as a replacement for language-selector? [07:29] pitti, so regarding the lock screen change - if I understand it correctly DBO says there is an instability problem with the switching as well as a performance one [07:29] (on nvidia) [07:30] You have nvidia right, are you seeing both problems? [07:30] robert_ancell: I don't have any nvidia in the house; my two laptops are intel, my wife's is using the free ati driver [07:31] robert_ancell: I'm mostly seeing the xrandr side, i. e. lots of flickering and long delays due to the monitor having to catch up [07:31] robert_ancell, yeah I get crashes inside the nvidia driver if the switch takes too long [07:31] DBO, yeah, that's the really concerning part [07:31] robert_ancell, at least they happen in libgl and not the x driver [07:32] so x stays up just the window manager dies [07:32] DBO, so it looks like a gl bug or a compiz bug? [07:32] nvidia gl bug [07:32] which we can't do anything with [07:33] yeah we have to work around those [07:33] pitti: small question, if I need to promote some binary packages (but not all), I guess I should put the source to main as well, isn't it? [07:33] didrocks: yes [07:33] robert_ancell: I left my blurb about both changes (indicator-session lock regression and the gnome-screensaver ones) in the bug [07:34] pitti: I was puzzled by the fact change-override didn't do it :) [07:34] pitti: thanks for confirming :) [07:34] didrocks: if you only change binary packages, it won't auto-change the source [07:34] pitti, ah, reading it now [07:36] robert_ancell, I have a hybrid graphic laptop, I was planning to spend some time figuring out how to use the nvidia chip today ... so let me know if you want me to do some testing [07:36] rickspencer3, pitti has used his veto powers :) [07:36] oh [07:36] nm [07:37] the xrandr part is certainly fixable on our part, as RAOF says too [07:37] but I really don't see a good way around the proprietary drivers [07:38] the crash and them not being able to handle more than one X.org at a time are both solvable in principle (but not by us) [07:38] pitti, there is blacklisting, but I'm not sure if we can reliably detect the driver from within the session [07:38] but them not using KMS, and thus having a really poor performance and flickering during VT switch is a principal problem [07:38] robert_ancell: yeah, if we can fix the randr part and blacklist non-KMS drivers, it'd be better [07:38] pitti, and the latter should hopefully solve that by matching resolutions by default [07:39] pitti: I don't see the quickly spec anymore on my status page: http://status.ubuntu.com/ubuntu-precise/u/didrocks.html, any idea? (because targeted later?) [07:39] robert_ancell: KMS == grep -q -E '(nouveau|drm)fb' /proc/fb [07:39] pitti, ok, gtg again, be back in an hour or so [07:40] that's what pm-utils uses [07:40] robert_ancell: not sure whether having /dev/dri/card0 is a good indication <- RAOF? [07:40] i. e. does nvidia have that? [07:40] (it's different from KMS, so I think checking /proc/fb is saver) [07:40] "safer" [07:41] pitti, nvidia does not make /dev/dri/card0 [07:41] pitti, do you think it's worth trying with that? Or would you prefer to revert the whole thing. Note also that all the changes are in lightdm for this to work, so it can be enabled on OEM system should they wish to patch indicator-session [07:41] I mean gnome-screensaver [07:41] DBO: ok, I remembered that much; but DRI != KMS, so it'd be only an approximation; checking /proc/fb is just as easy [07:42] robert_ancell: it still leaves the xrandr change; are you confident that this can be fixed? [07:42] robert_ancell: well, we can still revert it later, too [07:42] robert_ancell: I don't understand the indicator-session chagne, which caused the locking regression [07:42] robert_ancell: why was this necessary in the first place? [07:42] pitti, it depends on how bad you consider it - note that it is only when locking which is not normally a time critical thing [07:42] pitti, is the plan to fall back to the old lockscreen on nvidia? [07:43] DBO: yes [07:43] +1 to that good buddy [07:43] I don't think there's any other sensible option at that point on proprietary drivers [07:43] pitti, that was required because if do the lock and that asks lightdm to go to the lock screen, then do a user switch then you get two requests to the dm that spawns two VTs [07:43] didrocks: hm, curious -- https://blueprints.launchpad.net/ubuntu/+spec/desktop-p-quickly is still targetted to precise [07:43] the next version of lightdm does the locking for you so you don't have to worry about it [07:43] pitti, if we figure out the slowness issue, it might fix this for nvidia at least [07:43] RAOF, any news on that driver? :) [07:44] didrocks: ooh -- the milestone was set to "later", which doesn't really exist, I guess? [07:44] didrocks: should this be targetted to P (and mielstone -> beta-1), or moved to Q? [07:44] bbl [07:44] robert_ancell: and gdm? [07:46] pitti: ok, not sure who changed it, will fix it :) [07:47] DBO: Almost there! [07:47] RAOF, sexy [07:47] (Yes, that's slow. Sorry) [07:50] pitti, robert_ancell, RAOF a dist-upgrade this morning rather significantly horked my netbook [07:51] how in particular? [07:51] it starts in low graphics moods, then runs GMD, then won't start a session [07:51] low graphics mode [07:51] intel? [07:51] moods too, I guess, but my netbook doesn't like it when I personify it [07:51] pitti, yeah, Intel [07:52] pitti, suggestions? [07:52] rickspencer3: can you paste /var/log/Xorg.0.log and dmesg? [07:53] pitti, well, I can't log into a session, so not easily [07:53] rickspencer3: I also dist-upgraded about two hours ago and rebooted, worked hehre [07:53] so I don't have a general idea what could be wrong [07:53] robert_ancell: you can log into VT1 [07:53] pitti, yeah, I can get a VT [07:53] why would GDM run? [07:54] gdm? [07:54] pitti, yeah, GDM instead of lightdm [07:54] rickspencer3: do you have both lightdm and gdm installed? [07:55] it should work with gdm as well, of course [07:56] pitti, yes [07:56] rickspencer3: so, I think we'll need to see what dmesg/Xorg.0.log complain about [07:56] I thought gdm got removed on a previous update, I guess not, though [07:56] pitti, is this worth your time? [07:56] rickspencer3: as a first step, can you please check if "sudo apt-get -f install" complains aobut anythign? [07:56] interrupted dist-upgrade or so? [07:56] we got a new kernel and such [07:57] pitti, -f install just said there was no work to do [07:57] rickspencer3: making our manager be able to work again is worth our time, yes :) [07:57] rickspencer3: ok, so it's not that at least [07:57] pitti, well, I have a laptop too [07:57] rickspencer3: if you want to try, you could reboot, press shift to get into grub, and try with the previous kernel (-15) [07:57] if I am experiencing somehting idiosyncratic, I would want to put it off [07:57] but I am worried that it's a general issue [07:57] and see whether that makes a difference [07:58] pitti, should I try dpkg --configure maybe? [07:58] rickspencer3: no, if apt-get -f install is satisfied, packages are okay [07:58] ok [07:58] * rickspencer3 ponders [07:58] rickspencer3: so I'd first try kernel -15 again [07:58] if it works -> kernel issue, need to bug them [07:58] if it fails -> need dmesg/X.org log, some other package broke it [07:59] I wonder if I could be out of room on my hard drive [07:59] df -h [07:59] well, I'm rebooting now [07:59] that's possible; but it at least ought to mount a /tmp tmpfs in that case [07:59] oops, the -15 kernel isn't installed! [08:00] or at least not in grub [08:00] oh, you already cleaned it up? [08:00] pitti, I didn't manually [08:00] rickspencer3: it's now behind "older OS versions.." or so [08:00] not the huge long list by default any more [08:00] pitti, under Previous Linux versions I have 2 memory test options [08:01] I've been upgrading this netbook for 2 years, now [08:01] weird [08:01] rickspencer3: does uname -r say 3.2.0-16-generic ? [08:01] I have to wonder if I once configured apt to delete old kernels at some point [08:01] rickspencer3: I don't think you can [08:01] well, I don't know how, anyway [08:01] computer-janitor cleans them up nicely [08:02] pitti, so, /dev/sda1 is at 100% [08:02] let me remove some stuff and try again [08:02] aah [08:02] rickspencer3: sudo apt-get clean [08:03] that woudl make a nice test case [08:03] sudo rm -r *.wmv [08:03] defualt install, fill up / to 100%, reboot and verify that a session starts [08:03] it ought to survive that [08:03] I'm assuming I'm not going to watch all these old episodes of Futurama and southpark again [08:03] that's why we mount a tmpfs on /tmp [08:04] at least we used to, this might be broken now [08:04] pitti, I agree that my netbook should be able to survive all these old episodes [08:04] /run is a /tmpfs now [08:04] but, it seems like this is a bit of a corner case [08:05] micahg: not by default [08:05] micahg: sorry, yes; /run is [08:05] I meant /tmp [08:05] * DBO gives RAOF the big puppy dog eyes [08:05] if it can't create X or D-BUS sockets, your system won't be very happy [08:05] pitti, making room fixed it [08:05] sorry to take up so much of your time [08:05] also .. [08:05] rickspencer3: np [08:05] phew [08:05] I was having flashbacks to the bad old days [08:14] moin! [08:14] hey Sweetshark, good morning [08:14] rickspencer3: are you still having trouble with LO? [08:14] good morning Sweetshark [08:14] I assume so, I haven't tried today [08:14] hold on [08:16] Sweetshark, so, yes, I am still having trouble [08:16] when I run it, the splash screen appears, and then when it disappears, nothing [08:16] when I run libreoffice from a terminal, there is no output [08:16] just the splashscreen coming and going [08:17] rickspencer3, does the process terminate? [08:17] DBO yes [08:17] what does "strace /usr/lib/libreoffice/program/soffice.bin" do? [08:18] Sweetshark, it segfaults [08:18] pitti, this behaviour only occurs under lightdm [08:18] robert_ancell: ah, I see === MacSlow is now known as MacSlow|afk [08:19] rickspencer3: strace itself? or the soffice.bin? [08:19] Sweetshark, not sure [08:19] the error dialog says strace [08:20] rickspencer3: interesting, I like it if debugging tools break themselves ... [08:21] pitti, actually it appears that indicator-session only switches under lightdm now [08:21] rickspencer3: was there any output before the segfault? if so, could you pastebin the last dying breath of it? [08:22] Sweetshark, yes, hold on [08:23] Sweetshark, http://pastebin.ubuntu.com/842759/ [08:29] rickspencer3: seems like it cant find libicui18n.so.48, hmmm [08:30] Sweetshark, ok [08:30] so, how can I help? [08:30] siimilar to bug 917153 ? [08:31] Launchpad bug 917153 in libreoffice "failed to upgrade from oneiric to precise: /usr/lib/libreoffice/program/unopkg.bin: error while loading shared libraries: libicule.so.48: cannot open shared object file: No such file or directory" [High,Triaged] https://launchpad.net/bugs/917153 [08:32] woot, no more webkit-gtk2 dependency for gwibber! [08:32] finally all merged [08:33] pitti: yes, I find that suspicious too. But I thought that to be an issue only with update order, after completing the install all libs should be there anyway .... [08:39] rickspencer3: does 'sudo apt-get install --reinstall libreoffice-core' and then 'sudo apt-get install --reinstall libreoffice-common' help? [08:39] rickspencer3: and do you know when the problem first appeared? There was no recent LO update in the last days. [08:44] Sweetshark, it started yesterday [08:44] and it worked on Monday [08:44] so, something when I updated yesterday morning, I guess [08:44] And this wasn't an oneiric->precise upgrade between Monday and yesterday, I assume? [08:45] rickspencer3: libicu48 is installed and configured? [08:45] (dpkg -l libicu48) [08:45] infinity, it was a dist-upgrade, but precise to precise [08:46] Hrm. === MacSlow|afk is now known as MacSlow [08:47] At first glance, it sure looks like the bug that pitti brought up, but I don't see how that could happen on a recent precise->precise upgrade. [08:47] meh [08:47] pitti, should I log a bug from my "can't start when the disk is full" thing from this morning? [08:47] Given that the last time I synced ICU was 3 weeks ago. [08:47] again wrong unpack order? [08:48] I thought that was the problem during dist-upgrade, although it should not really hit that hard on precise->precise [08:48] pitti: Sure, unpack order issues are a sound theory, but the problem on dist-upgrade was the icu44->icu48 SOVER bump. [08:48] rickspencer3: you migth as well; I'm sure that there are a thousand (unindentifyalbe) dupes around, but a clean one might be better [08:48] pitti: rick would have already had icu48 installd, so... [08:49] pitti, which package should I log it against? [08:50] rickspencer3: hm, I'm not sure about the root cause yet, so perhaps start with an "ubuntu" only task [08:50] rickspencer3: it's probably in more than one [08:50] Sweetshark, hey, so, those reinstalls didn't work [08:50] rickspencer3: what was the exact problem, X.org failed to start? [08:51] rickspencer3: I think we lost the automatic tmpfs /tmp creation [08:51] erm, what does Super+R sho [08:51] rickspencer3: so please start with "mountall" [08:51] pitti, ack [08:51] ... what should Super+R do? it does something weird now for me [08:52] something like panning virtual desktop with zoom [08:52] eek [08:52] and I can't disable this mode, mouse pointer is also not displayed [08:52] how do I get out of that again? [08:52] TheMuso: ^ do you know? [08:52] rickspencer3: Regarding your libreoffice troubles, does dpkg list "libicu48" as installed? [08:52] pitti, phew [08:52] ah [08:53] pitti, press super and scroll down on mouse/trackpad [08:53] everything is so big [08:53] help [08:53] rye: thanks [08:53] phew [08:53] ok, so that's pretty cool [08:53] so super+r is indeed confusing [08:53] quite unexpected, sorry about panicking [08:54] pitti, I got an apport error! [08:54] * rickspencer3 resists the yak shaving excursion [08:55] rickspencer3: which? [08:56] infinity, according to apt-cache, it is installed === fenris is now known as Guest77257 [08:57] hey [08:58] bonjour seb128 [08:58] hey pitti, how are you? [08:58] rickspencer3: Hrm. Then it's vaguely curious that libreoffice can't find it. :/ [08:58] seb128: not too bad; desperately trying to keep teh archive in one piece :) [08:58] seb128: and arguing about the gnome-screensaver changes [08:59] pitti, oh, sorry you were not around when I uploaded so not kept in the loop, it's rather a "let's see how it goes and get some feedback", I don't think we will stick to it for this cycle [08:59] seb128: I left my ramblings in the bug, and discussed possible options with robert_ancell [09:00] ok [09:00] good morning everyone [09:00] mdeslaur had good feedback yesterday [09:00] hey chrisccoulson [09:00] hi pitti, how are you? [09:00] seb128: in other news, g-s-d works like a charm here [09:00] I didn't think much about the issue due to the fact that your switch away from your user seat so i.e disable sound [09:00] chrisccoulson: quite fine, thanks! [09:00] or don't respect the user settings [09:00] seb128: ah, haven't even considered those [09:00] pitti, great, I will upload that to precise in a bit, jbicha filled the libwacom mir and mterry acked it [09:01] seb128: it totally breaks nvidia users, and it's a really bad experience with multiple monitors [09:01] salut seb128 [09:01] pitti, multimonitor? because vt switch sucks? [09:01] didrocks, lut [09:01] chrisccoulson, hey, how are you? [09:02] pitti, https://bugs.launchpad.net/ubuntu/+source/mountall/+bug/932598 [09:02] Launchpad bug 932598 in mountall "Full hard drive causes xorg and desktop to fail to load" [Undecided,New] [09:02] like I dope, I logged it from a different computer :/ [09:02] seb128, yeah, not too bad thanks. stressing about mozilla bug 694594 though ;) [09:02] rickspencer3: thanks, I'll queue that [09:02] how are you? [09:02] Mozilla bug 694594 in JavaScript Engine "Crashes with gcc 4.4.3" [Critical,New: ] http://bugzilla.mozilla.org/show_bug.cgi?id=694594 [09:02] rickspencer3, pitti: isn't that issue years old? [09:03] seb128: it got much better when we had the fallback in sysvinit [09:03] seb128: it used to mount a tmpfs on /tmp if you had < 1 MB free on / [09:04] but I think we'd be better off with always having that [09:04] pitti, rickspencer3: bug #35217 [09:04] Launchpad bug 35217 in gdm "Unable to login when disk space is exhausted" [Wishlist,In progress] https://launchpad.net/bugs/35217 [09:04] it's a 5 digits bug! [09:04] lol [09:04] seb128: so there's two parts [09:04] pitti, I'm pretty sure it was discussed on another bug recently but I think I cleared the emails [09:04] and we're nearly at 7 digits! [09:05] seb128: if you have /home on a full partition, the session fails to start [09:05] but if /tmp is full, then not even lightdm/X.org start up [09:05] as D-BUS etc. can't even create their sockets there, etc. [09:05] pitti, oh, right [09:05] but still, we can join the bugs indeed [09:05] pitti, I don't think we ever really solved the xorg,session part [09:05] thanks for the pointer [09:06] right [09:06] good morning seb128 [09:06] i know it must be time for bed if seb128 is here :) [09:06] kenvandine, hey, what are you doing still there? ;-) [09:06] kenvandine, indeed :p [09:06] merging branches :) [09:07] jibel: how much effort would it be to add a test case which fills up the disk with a dummy file, then reboots, and checks that X and lightdm come up properly? [09:07] DBO: Why is the nvidia installer give such totally useless error messages?! [09:07] jibel: this case kept breaking several times in the past [09:08] pitti, rickspencer3: ah, found it [09:08] bug #873368 [09:08] Launchpad bug 873368 in gnome-session "Session fails to start when root system full" [High,Triaged] https://launchpad.net/bugs/873368 [09:09] pitti, ^ that's the one I read comment on recently [09:09] seb128: that sounds like a proper dupe of 35217 indeed [09:09] it got assigned to our team [09:16] is anyone else using the new unity from the PPA? [09:16] yeah [09:16] pitti, doing a fresh desktop install, reboot, check that essential services are running, fill the disk, reboot, check again. That's it ? [09:16] didrocks, do you notice an annoying flicker when switching workspaces now? [09:17] chrisccoulson: in the preview, you mean? [09:17] perhaps that's a compiz thing actually. did compiz change recently as well? [09:17] didrocks, no, this is when using the keyboard shortcuts [09:17] you don't read the team report? :p [09:17] heh [09:17] i do, honest ;) === Guest77257 is now known as ejat [09:18] chrisccoulson: one sec, I have a video [09:19] chrisccoulson: bug #930122 [09:19] Launchpad bug 930122 in compiz "some window preview glitches on workspace switcher and expo" [High,Confirmed] https://launchpad.net/bugs/930122 [09:20] didrocks, yeah, that looks the same as what i'm seeing :) [09:20] thanks! [09:20] chrisccoulson: can you confirm as well? [09:20] pitti, I think we can create a test case for this in the current iso testing framework. not very difficult but need to find the time to do it. [09:20] chrisccoulson: it's compiz [09:20] pitti, is there any level of emergency ? [09:20] (spotted before the upload, but didn't worth blocking on it) [09:20] didrocks, yeah, i can confirm that [09:25] meh. still cant reproduce the icu bug. [09:26] jibel: no, it's not really something new; it just came up in the discussion again [09:27] jibel: it doesn't need to be an entirely new install; a post-instal test of an existing one is sufficient [09:27] pitti: good morning [09:27] jibel: so that this takes a minute, not 20 [09:27] hey desrt [09:27] pitti: i think introspection in libxklavier is broken [09:27] desrt: how so? [09:27] GISCAN Xkl-1.0.gir [09:27] gives [09:28] /home/desrt/jhbuild/checkout/libxklavier-5.2/libxklavier/xkl_config_registry.h:24:36: fatal error: libxklavier/xkl_engine.h: No such file or directory [09:28] the reason seems to be because Xkl_1_0_gir_CFLAGS = $(INCLUDES) $(X_CFLAGS) $(XML_CFLAGS) $(GLIB_CFLAGS) $(XINPUT_CFLAGS) [09:28] my checkout is still from January 16 [09:28] let me check [09:28] but meanwhile the real AM_CFLAGS contain -I. -I$(top_srcdir) [09:28] oh. i'm looking at the 5.2 release, btw [09:28] it may well have been fixed since then. silly me. [09:29] my checkout is before 5.2 even [09:29] I'll compare my checkout with current master [09:29] ok, still works in my checkout (c8addc9) [09:30] oh. indeed [09:30] http://cgit.freedesktop.org/libxklavier/commit/?id=57785c952444e4cb9d13e43d4f9cc9713d66958b [09:30] made after the 5.2 release [09:30] hmm; master also works [09:30] yes. you're right. [09:30] aah [09:30] right, I noticed that on the next day [09:30] i don't suppose you're interested in doing a point release :) [09:30] but in the meantime 5.2 was already released [09:30] so I cherry-picked this into our packages [09:30] (i guess i ask svu for that) [09:31] desrt: yes, I pointed that out to him [09:31] okay. sorry for the noise. [09:31] jhbuilding is a lot like packaging ;) [09:31] no problem at all [09:31] desrt: let's say, if it's done well in jhbuild, packaging becomes much easier :) [09:32] but I should update our packages to 5.2 [09:32] not many changes, but it's cleaner [09:32] pitti: i like to think i've been actively making seb's life easier lately by catching ubuntu issues in gnome before he tries to package them :p [09:33] desrt, hey ;-) [09:33] seb128: good morning :) [09:34] desrt, how are you? are you in europe yet or just on european tz adjustment? ;-) [09:34] working on the adjustment :) [09:34] seb128, there's a new lightdm/u-g about to be uploaded - did you want to do any smoke-testing? [09:34] * desrt will be in europe tomorrow [09:35] robert_ancell, hey, I can do if you want [09:35] robert_ancell, everything is in the packaging vcs? [09:35] seb128, yup [09:35] robert_ancell, ok, I'm on it [09:35] robert_ancell: latest u-g theme changes look really nice, btw [09:35] the seethrough dropdown is a nice touch [09:35] desrt, that's all mterry and cimi's work [09:36] seb128, if you've got multi-monitor then the greeter may look a bit weird - it does multi-monitor now, but the default is to mirror and on my laptop it trims the sides [09:36] robert_ancell: hah! [09:36] (default being the g-s-d default) [09:36] i was just typing a question to ask you how to change that :) [09:36] robert_ancell, I'm using a laptop docked but with lid closed, i.e just the external monitor [09:37] robert_ancell, that usually doesn't stop the closed screen for some reason and put me in minor [09:37] because it looks quite bad [09:37] which is annoying because I get the laptop resolution on the external screen [09:37] the resolution ends up being sort of the worst-choice of all connected monitors [09:37] seb128: it's not even the laptop's resolution [09:37] it's the highest resolution supported by both devices [09:37] which for me is something extremely obnoxious like 1024x768 [09:37] desrt, I think it is, at least it's in mirror mode, both screen should have the same resolution, I figured it would pick the lower one of the two [09:38] seb128, right, so if I start with my monitor plugged in I get that, but if you disconnect and reconnect the monitor it has the correct resolution [09:38] seb128: not all monitors support all lower resolutions [09:38] desrt, right, I see what you mean [09:38] so if your laptop has an odd panel size, chances are the external monitor does not directly support it [09:39] seb128, so not sure what we're going to do here, the changes are nice but if we can't get g-s-d/x to play nice they might have to be reverted [09:39] i think this is what the "Note: may limit resolution options" string in gnome-control-center is hinting at (which at first i thought only meant what you thought it meant and i was like "well... duh... obviously") [09:39] robert_ancell, I'm pleased to see that you used merge-upstream this time :p [09:40] seb128, I remembered :) [09:40] robert_ancell, I was not even sure you read my email [09:40] robert_ancell: do you have an idea of what you'd ideally like to happen? [09:40] desrt, side by side - always [09:40] seems right [09:40] how do you select the default though? [09:40] desrt, heuristics I guess [09:41] * desrt suggests a good one: largest panel [09:41] you will not plug in an external monitor and ignore your laptop screen if the laptop screen is larger... [09:41] desrt, or the one closest to the hardware, i.e. laptop display overrides external monitor [09:41] robert_ancell: i don't like that option because the lid may be closed [09:41] desrt, then you only have one screen [09:42] is that how it works these days? [09:42] robert_ancell, should I test lightdm first and then the greeter or both together? [09:42] lid closed -> xrandr 'disappears' the monitor for you? [09:42] seb128, might as well both [09:42] desrt, it should [09:43] that's pretty clever logic [09:43] (no idea if it does) [09:43] well, i've seen behaviour that suggest that there is at least some hinting going on in this direction [09:43] robert_ancell, ok, installed, brb [09:43] maybe g-s-d is doing something clever there [09:44] desrt, oh, I noticed that u-g is using the slow method of doing images (like we had in g-g) [09:44] is there an easy way to confirm the current method is the slow one? [09:44] burn it! [09:44] desrt, the main part is being done in a thread and it didn't like me doing a trivial replacement [09:44] for gnome-games i put a for loop for 1000 iterations around the 'draw' function [09:45] that was enough to show a rather visible difference between the code as it was and my improvements [09:45] desrt, also, is there a reason why there isn't a Gdk.cairo_make_native_surface (w, h)? [09:45] robert_ancell: oh. yes. don't do that. [09:45] robert_ancell: you'd also need a GdkDisplay there (or maybe even a Gdk.Screen) [09:46] or just grab the default [09:47] gdk tries not to introduce APIs that assume the default gdkdisplay these days [09:47] in order to be safe for multiple GdkDisplay in one process [09:47] (which i think is pretty bogus, but whatever) [09:47] actually, i guess with our new support for multiple display backends this makes more sense than it used to [09:47] but i still don't know of anyone who is actually using this support for anything [09:48] robert_ancell: anyway.. it makes a lot of sense [09:48] cairo is threadsafe itself [09:48] robert_ancell, is buggy [09:48] and an image surface is fully detached from the X server -- it's just in local process memory [09:48] X on the other hand is _not_ threadsafe [09:48] and the cairo 'similar' surface for an X window is going to be doing drawing via X [09:48] seb128, in what way? [09:48] so you can't write to the similar surface in a thread at the same time that you do other things to X in the main thread [09:49] robert_ancell, I can't start a guest session, it just switches me to the greeter [09:49] robert_ancell, (it always did that when a greeter is running) [09:49] did you make it so there is always a greeter running? [09:49] desrt, yeah, I was moving all that stuff outside of the thread and just doing the GdkPixbuf scaling in the thread [09:50] robert_ancell: use cairo to do the scaling on the GPU [09:50] robert_ancell, usually the greeter would close when I logged in or typed my password on the greeter to log back in, and then I could start a guest session from indicator-session [09:50] robert_ancell: at the very worst you will get a fallback that's no worse than GdkPixbuf [09:50] robert_ancell, now picking guest in the session menu (from my session) send me to the greeter [09:50] robert_ancell: and maybe you get something a lot better... [09:50] seb128, confirmed [09:52] robert_ancell, otherwise login on ecryptfs and normal, switch user, keyboard layout, etc work fine [09:52] (although you may lose the concurrency benefits, of course) [09:52] robert_ancell, small 'glitch' in lightdm as a greeter, if I log with 2 users and switch between those from the session indicator it sends me to the other session and I get gnome-screensaver, not the lightdm greeter [09:52] always having a greeter running will again break on nvidia & friends [09:52] (if that's intended) [09:53] pitti: it is intended to break nvidia & friends? ;) [09:53] pitti, I think you made that point yes ;-) [09:53] does unity have an overview like 'spread' view? [09:53] seb128: different context here, though; it's independent from the gnome-screensaver changes [09:53] robert_ancell, otherwise multimonitor -> it doesn't behave differently from before, I get my external monitor in 1440 and not 1920, I guess it goes to mirror still by default for me [09:54] seb128, yeah, you'll need to unplug/replug it [09:54] desrt: hm, we used to with super+w [09:54] desrt, super-w [09:54] desrt, sit on super for a second [09:54] desrt: but now that only shows the apps on the current ws [09:54] desrt, you will get a keys overview [09:54] super+w works, thanks [09:54] seb128: super+w is not documented there [09:55] pitti, it is for me [09:55] seb128, ok, I don't have the energy to chase down the problem tonight, I'll look into it tomorrow. Can you revert the gnome-screensaver patch / indicator-session change? I don't think I'm going to have the time to get it all working [09:55] seb128: interesting [09:55] seb128: odd -- translation issue? [09:55] pitti, first in the bottom right section [09:55] oh, there, right [09:55] seb128: same here [09:55] hum. vi is calling unity a liar [09:55] robert_ancell, ok [09:55] it's telling me that it's quite sure i have this file open in another terminal [09:56] * desrt thinks it's wrong, after all [09:56] robert_ancell, do you want to upload the new lightdm,unity-greeter anyway? [09:56] robert_ancell, I would say the guest session stuff is minitor [09:56] minor [09:56] robert_ancell, you can start it from the greeter [09:56] robert_ancell: I have locally reverted the screensaver one; can upload that if you want to [09:56] though if the greeter keeps running that's an issue for nvidia as pitti pointed [09:56] hmm [09:56] desrt 12637 0.0 0.0 53908 5076 pts/2 S+ 04:19 0:00 vi gnome-suites-core-deps-3.4.modules [09:56] pitti, please do [09:56] pitti, I know your just waiting to do it - go for it! [09:56] does super+W also show minimised windows? [09:57] robert_ancell: heh [09:57] seb128, nah, it's not pressing - just let mterry know the new ones are available if he wants to do any testing [09:57] robert_ancell: good night, sleep well! [09:57] robert_ancell: thanks for staying up late to discuss this [09:57] pitti, can you revert the indicator-session patch as well if you think it's wrong? be aware that upstream acked it and merged it to their trunk so we will need to discuss it there if we want to revert that [09:57] seb128: I was told it's incomplete [09:57] pitti, dx is keen to not enforce screensaver from the session menu [09:57] seb128: the next lightdm will do the locking [09:58] so I'm fine with keeping this for now [09:58] pitti, well as you pointed on the bug that's an issue for other dms [09:58] ok [09:58] seb128: robert told me that scrensaver will only do that if lightdm is running [09:58] robert_ancell, 'night [09:58] pitti, ok, good [09:58] seb128: I guess he meant "indicator" [09:58] seb128, though I just grepped the indicator code and it appears they only support lightdm (i.e. no reference to gdmflexiserver) [09:58] pitti, so let's wait for that, you upload gnome-screensaver? [09:58] or that [09:59] seb128: yes, will do [09:59] thanks [09:59] I'm fine wiht keeping indicator-session for now [09:59] robert_ancell, ok, we can sort the indicator stuff later [09:59] ok, bye all [10:00] how does one regain access to windows that they have minimised? [10:00] desrt, they click on them in the launcher [10:00] that seems to be lauching new windows for me [10:00] it shouldn't [10:00] ohhhh [10:00] it means you have a bamf bug and it doesn't see the current instance [10:00] fascinating [10:01] super+w -is- showing my minimised windows [10:01] left click is focus the current one [10:01] they're just invisible [10:01] middle click is start a new one [10:01] like, it does the spread, and leaves a space where the window should be and clicking into the space brings it back [10:02] found it :) [10:02] desrt, works for me, but I've seen issues with that in compiz before [10:03] seb128: it was also not possible to get to it using alt+tab :/ [10:03] desrt, that seems a bamf bug [10:03] desrt, did you restart it or did it segfault during your session? [10:03] i don't recall either of those things happening, no [10:04] desrt, unity has some issues when bamf restart, it doesn't restore its state always correctly [10:04] that's a known bug [10:04] it will be looked at after ff [10:04] cool [10:04] i kill bamf a lot while working but i did a cold boot when i got home and haven't done any work on it since then [10:04] so something to keep in mind in any case if i see this bug pop up later while working [10:05] thanks :) [10:05] right [10:05] yw [10:05] seb128: done, and bug 878836 updated accordingly [10:05] 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 [10:05] pitti, danke! [10:05] seb128: any new package versions etc. which we need to get in today/by FF? [10:05] I'm doing some stable+1 stuff to hold the archive in one piece, but I have some time in between to help out if necessary [10:06] pitti, there is some red on versions but I don't think any is ffish [10:06] like they are point releases, i.e bug fix updates [10:11] Sweetshark, so, thoughts on getting LibreOffice working again? [10:14] * rickspencer3 reinstalls everything, restarts session [10:24] rickspencer3: Im a bit out of ideas for now. I just updated my precise VM from ~Desktop-Sprint and had no issues. [10:34] pitti, were the retracers broken and catching up or did precise just turned to have segfault issues yesterday? [10:34] seb128: the former, I restarted it some hours ago [10:35] ok, good, I was starting being a bit concerned, I've no time before ff to go through the bug emails but I can see those pilling up there ;-) [10:54] pitti, is it possible to add xsession-errors to apport bugs reported for the lenses? [10:55] mhr3: it should be attached by default [10:55] mhr3: well, right now we do that for programs which link with either libgtk or libX11 [10:56] mhr3: if lenses don't do that, we could either add a new 'trigger' library to apport [10:56] pitti, yea, lenses don't do that [10:56] pitti, mhr3: we don't add .xsession-errors, we do a grep on known warning markers [10:56] or you manually call apport.hookutils.xsession_errors(pattern) to the hook [10:56] well, yes [10:56] pattern = re.compile('^(\(.*:\d+\): \w+-(WARNING|CRITICAL|ERROR))|(Error: .*No Symbols named)') [10:56] that's our current filter [10:56] pitti, mhr3: that's on my list of things to "get fix" [10:56] to avoid exposing arbitrary data there [10:56] "get fixed" [10:57] pitti, I think we should not worry about arbitrary datas [10:57] we loose lot of useful informations with the current way [10:57] in practice you don't have often real datas in there [10:57] what bug-buddy was doing is attaching like the n most recent lines to GNOME bugs, I don't think it has been an issue [10:58] well, we introduced this check because we had bugs with that [10:59] any package hook can fish out other things it's interested in, of course [10:59] well for one unity is using custom log handlers, so it should include that as well [10:59] like using just "WARN" [11:00] pitti, ok, I will open a bug about including Xerrors in some way there [11:00] pitti, we miss the actual infos for half the g-s-d issues which are in BadMath xerrors [11:00] but that errors is not collected [11:01] so we don't get the numbers associated with those [11:01] if we can grep X errors, that sounds fine [11:04] dear all, libproxy is ready for reviewing if you have time git://git.debian.org/git/users/laney/libproxy.git [11:04] i still do not know what to do about the modules [11:05] pitti, https://bugs.launchpad.net/ubuntu/+source/apport/+bug/932660 [11:05] Launchpad bug 932660 in apport "XsessionErrors should include X errors" [Undecided,New] [11:07] Laney, hey, how are you? [11:07] yeah, good thanks :-) [11:07] you? [11:07] pitti, ^ libproxy seems something we want in for ff if you still have some spare cycle [11:07] Laney, I'm great thanks, a bit in a rush for ff [11:07] heh [11:07] pitti, our version doesn't support GNOME3 so we need that update [11:08] Laney: you want me to do a Debian upload, or merge with Ubuntu, or both? [11:08] pitti: I think there's a bit of a Joss situation with this package atm, so … [11:09] I mean, I'm fine with committing changes to the Debian svn [11:09] well Joss hates it but he agreed the update was needed since the current version is broken [11:09] I'll try and negotiate with pkg-gnome; seeking reviews and probably an Ubuntu upload atm [11:09] just not sure what this git branch is [11:09] seb128: oh, why doesn't he like it? [11:09] there's some stuff I am not entirely confident about, like the multiarching or the naming of the module packages [11:09] or how users are supposed to get the modules [11:11] pitti, dunno, he justs hate the design or the code or something [11:11] thinks it should be better done in glib-networking [11:11] pitti, he's wanting to replace it with something else or write something new instead, he considers it just broken by design and something we should get ride of if we can [11:12] seb128: ah, ok; so nothing in particular about the new version [11:12] I didn't bother trying to figure why exactly he hates it, I just know he does ;-) [11:12] pitti, no, he said he agrees we need the update since the current one is broken with GNOME3 [11:12] but he would prefer to just get ride of it if we could :p [11:13] actually probably the modules don't need to depend on the shared lib === MacSlow is now known as MacSlow|lunch [12:18] pitti: A transition issue: As long as bug 926207 is not resolved, some of the people who install Ubuntu will have LC_MESSAGES, LC_CTYPE and LC_COLLATE set in /etc/default/locale. Therefore I think we should comment the 'dpkg --compare-versions' condition in http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/precise/accountsservice/precise/view/head:/debian/accountsservice.postinst for now. It can be uncommented later on, b [12:18] Launchpad bug 926207 in ubiquity "Set formats related LC_* variables when applicable instead of LC_MESSAGES, LC_CTYPE and LC_COLLATE" [Undecided,New] https://launchpad.net/bugs/926207 [12:18] ut with a later accountsservice version than 0.6.15-2ubuntu3. Any thoughts? [12:18] GunnarHj: that doesn't help much, though [12:18] as it's unlikely that there is a package update after a precise installation [12:18] pitti: Why not? [12:18] accountsservice is not the kind of package that gets many post-release updates [12:19] pitti: Then what would you suggest? [12:19] didrocks, not sure if you've seen this one https://bugs.launchpad.net/ubuntu/+source/unity/+bug/932520 [12:19] Launchpad bug 932520 in unity "Maximzed windows on start up don't show full window" [Undecided,New] [12:20] GunnarHj: well, thaht bug needs to get fixed :) [12:20] pgraner: this is with latest unity trunk? [12:20] pgraner: or new to the compiz update yesterday? [12:20] didrocks, no yesterday's archive [12:20] pitti: Yes. Of course, my assumption was that it has to be done before the 12.04 release. [12:20] pgraner: ok, seems a compiz regression then, adding to the list, thanks [12:21] didrocks, started on after monday's updates and I've been able to repro tho it does take some time [12:21] GunnarHj: after that we can do another accountsservice upload and bump the comparison, to fix it for people who installed alpha-2 etc. [12:21] pgraner: are you sure they are maximized not just "big enough"? [12:21] pitti: Yes, that's exactly what I meant. [12:21] (it's hard to tell from the screenshot) [12:21] but yeyah, seems anyway that something is cut [12:22] didrocks, yea look at the screen shot, the buttons are showing the right state and when you click the max/min button it goes to a window [12:22] GunnarHj: ah, ok; misunderstood you then [12:22] didrocks, to note once you min then max the window all is working well [12:22] pitti: ok, I'll add about the button showing in the right state and add to the regression ist [12:22] list* [12:22] oupss, I meant pgraner ^ [12:23] didrocks, ack, let me know if I can do anything else, I won't update that box for awhile yet [12:23] pgraner: smspillaz FYI, adding that to the list to look after ^ [12:29] pitti, good job on the apport-request-retrace stuff, we got to try it on one of the dx bugs yesterday and it worked great [12:29] ah, nice! [12:29] this list is growing larger and larger :( [12:30] pitti, it's the first time I saw it in action, works as it should, kudos for that ;-) [12:32] (this one is probably another silly race condition that I don't have time for again :( ) === Ursinha_ is now known as Guest37786 === Guest37786 is now known as Ursula [12:44] Ran 47 tests in 453.606s [12:45] ok, time to make an additional one that won't pass :) [13:05] mterry, hey , I resubmitted that branch for merging using canberra exclusively === smspillaz is now known as sm|break === greyback is now known as greyback|lunch [13:37] Laney: looking at your git now [13:37] brill [13:37] don't forget to git-dch [13:37] Laney: but this confuses me -- it has 0.3.1-3 [13:37] ^ [13:37] :-) [13:37] but Deiban already has 0.3.1-4 [13:38] I haven't imported -4 yet [13:38] but it's not massively important (I have reworked the packaging anyway) [13:39] Laney: so, what shoudl I do with this git branch? [13:40] make sure to branch upstream and pristine-tar [13:40] ronoc, ok, will look [13:40] git-dch --auto, git commit debian/changelog -m snapshot === MacSlow|lunch is now known as MacSlow [13:40] git-buildpackage [13:40] seb128, morning. who changed the power dialog to add a dropdown for the battery indicator? I had a branch ready to go for that [13:40] TA [13:40] mterry, hey [13:40] oh ta [13:41] Laney: oh, this way around (I'm using to committing the debian/changelog together with the changes, and debcommit) [13:41] check the multiarching is OK; see if you agree with me that the plugins shouldn't depend on the shlib; see if you have an idea about how users should get the plugins [13:41] caps lock ... [13:41] mterry, wasn't that part of charles's work? i.e the stuff you both worked on? [13:41] Laney: plugins? [13:41] ah, yeah, that's another workflow, but I don't do it that way because it makes cherry-pickink and other vcs manipulations easier [13:41] Laney: (NB that I have NFC about what this pacakge does) [13:41] pitti: yeah, like the gsettings plugin or the kconfig one [13:41] it lets apps query the system for the proxy to use [13:41] seb128, I didn't know he had a g-c-c side branch too; just knew about the indicator side [13:42] and has plugins for various data sources [13:42] mterry, sorry about that :-( [13:42] so e.g. it won't query gsettings if you don't have libproxy1-plugin-gsettings installed [13:42] mterry, the patch I applied was from charles, I though you both sorted that out the other day [13:42] mterry, I hope you didn't spend too much time on it [13:42] and I don't really know what the best way for users to have that installed is [13:43] have g-c-c or something recommend it? [13:44] seb128, no, just the indicator side. Seeing a screenshot of the panel now, I'm inclined to still use my patch, as mine also makes the panel look more like the google doc design in other ways besides just the dropdown [13:44] mterry, yes please do, I just tried to clean the queue yesterday and get stuff landing before ff [13:44] mterry, I don't like much how his version looks like, alignment is weird, but I didn't have time to argue over it [13:45] seb128, yar, of course. This was just a miscommunication [13:45] Laney: how did you test this? [13:45] mterry, and it would be fair that you each land one of your commits ;-) [13:45] heh [13:45] install the tools package and run 'proxy http://some.url' [13:45] Laney: the only rdepends are glib-networking and glib-networking-services [13:45] apt-cache rdepends libproxy0 [13:45] Laney: ah, and if I set a proxy in control-center, it should pick that up? [13:45] if you have the gsettings plugin installed, yeah [13:46] http_proxy=http://foo proxy http://www.ubuntu.com is a simple test that it's working at all [13:46] Laney: I meant for testing that plugins still work with the multi-archification; I guess $http_proxy is built in [13:46] ah, yeah [13:47] Laney: but I guess you already did that [13:47] set _MM_DEBUG=1 _PX_DEBUG=1 [13:47] and you'll get debugging output [13:49] Laney: so I'll take git diff debian/0.3.1-3 .. -- debian, review that, and put it into svn [13:53] If you like... I was going to tread a more careful political course ;-) [13:54] Laney: why did you drop debian/control.in? [13:55] part of switching to dh [13:55] possibly I shouldn't have done that [13:55] Laney: so you intended this to be an Ubuntu-only update? [13:55] (for the time being) [13:56] (wrt. "political course..') [13:56] For now I think that would be easier; I am planning on speaking to pochu as maintainer about what to do [13:58] Laney: multi-archification was already done in -4ubuntu3, and it looks like -4ubuntu2 and -4ubuntu1 are obsolete [13:58] right? [13:58] * Laney looks [13:59] 4ubuntu1 because there are now separate plugisn [13:59] yeah, that looks accurate [13:59] sorry I haven't been on top of merging in changes since I forked it [14:01] possibly in Debian it would be easier to target exp at first [14:01] and for debian we'll probably need to revert o cdbs [14:02] likely unless I take it over [14:02] I don't want to commit this to svn as it is, as it woudl be the only pkg-gnome package which doesn't use control.in and cdbs [14:02] yes [14:02] I mean at least not without an ack from mbiebl, pochu, or joss [14:03] I don't think the changes would be too difficult to make that so [14:03] it was mainly the cli/python2 sequences that made me do it this way [14:03] anyway, the diff looks ok to me (for Ubuntu), now I just need to create a real changelog [14:04] cool [14:04] I'd just remove all except the most important entries [14:06] Laney: http://paste.ubuntu.com/843046/ [14:06] how does that sound? [14:06] the shlibs:Recommends thing was only transient [14:06] actually, the diff doesn't have any rpath [14:06] * pitti drops that as well [14:06] and that [14:06] and I'd like Ken's change to be credited if you can [14:07] http://paste.ubuntu.com/843049/ [14:08] wfm [14:08] ok, I'll do a test build and upload then [14:08] Laney: thanks! [14:08] we can think about adding recommends or whatever later I guess [14:09] thanks for your review! [14:09] Laney: I think it's right for plugins to be in multi-arch dirs [14:09] they can't be M-A: foreign [14:10] yeah [14:11] $ http_proxy=http://127.0.0.1 proxy http://piware.de [14:11] direct:// [14:11] hmmm [14:12] Laney: ^ that doesn't look right? [14:12] hmmm [14:12] set _MM_DEBUG=1 _PX_DEBUG=1 [14:12] also tried with other hosts [14:12] maybe it prefers gsettings? [14:12] laney@raleigh> http_proxy=http://127.0.0.1 proxy http://piware.de ~/dev/debian/packaging/pkg-gnome/libproxy [14:13] http://127.0.0.1 [14:13] preload : /usr/lib/x86_64-linux-gnu/libproxy/0.4.7/modules/config_gnome3.so [14:13] Using config: 22gnome_config_extension [14:13] Using ignore: localhost,127.0.0.0/8 [14:13] Config is: direct:// [14:13] direct:// [14:13] ah, perhaps with that plugin it _only_ considers the gsettings bits [14:13] yeah [14:13] not sure if that is desirable [14:13] right, that's it [14:16] gnome-control-center network crashes, can't test this ATM [14:16] seb128: ^ for you as well? [14:16] (as soon as I click on "proxy") [14:16] might be fallout from g--s-d [14:17] * Laney needs to thesis for a bit, ttyl [14:17] gsettings set org.gnome.system.proxy mode auto [14:17] Laney: now it queries WPAD [14:17] seems to work fine [14:17] :-) [14:17] with "manual", too [14:18] pitti, wfm [14:18] pitti, stacktrace? [14:19] Laney: uploaded, thanks! [14:19] sweet, cheers again for your help [14:19] seb128: http://paste.ubuntu.com/843063/ [14:20] pitti, hum, you have a11y on? [14:20] by accident perhaps [14:20] pitti, I will need to check on that, it's off there, but it's not a gsd side effect, it's an a11y issue [14:20] migth have enabled it in the past to test something [14:20] org.gnome.settings-daemon.plugins.a11y-keyboard active true [14:20] seb128: ^ this one? [14:21] org.gnome.settings-daemon.plugins.a11y-settings active true [14:21] pitti, if that is active a11y will be one yes [14:21] * pitti resets the whole two a11y trees === greyback|lunch is now known as greyback [14:22] restarting session [14:22] pitti, the gst plugin one is fine [14:22] pitti, check on org.gnome.desktop.a11y.keyboard enable [14:25] desrt: ping [14:25] seb128: hm, still crashes with everything reset [14:25] anyway, meeting, will check again later [14:26] seb128: also crashes in a guest session [14:26] pitti, hum, wtf, will debug tomorrow after ff rush if that works for you [14:27] pitti, but it works here so it's not an everybody having it issue [14:27] Laney, kenvandine: ah, bummer -- libmozjs185-dev is in universe, so libproxy is depwait [14:27] argh [14:27] :( [14:27] was it intended to go to main? [14:28] I guess not [14:28] we might need to disable that plugin from libproxy [14:29] kenvandine: pinged you because you apparently contributed the patch to build against it [14:29] just a rename [14:29] it had depended on the unverioned one [14:29] unversioned [14:30] just means that people have to use libjscore to interpret PAC files [14:32] kenvandine: I see libmozjs-dev was in main in hardy, but I don't konw whether chrisccoulson is happy about it being in main [14:32] Laney, pitti: until now we built with webkit and not gjs in ubunt [14:33] * kenvandine would rather not use libmozjs [14:36] agateau: hey [14:36] * desrt wonders: is it common practice to have patches directly in the .diff.gz anymore? [14:36] is this a result of packaging out of git or something? [14:36] does anyone here make any changes to the expo compiz plugin? should the active desktop be getting colored or highlighted somehow? [14:36] * desrt is used to seeing debian/patches/ [14:37] (bc it isn't) [14:37] desrt: some time ago you mentionned a linked attribute for GtkButtonBox or something similar, which would be useful to implement the gnome-control-center breadcrumbs [14:37] desrt: in the .diff.gz? [14:37] desrt: I can't find it :/ [14:38] desrt: usually in the dpkg source files, the patches end up in the .diff.gz or the debian.tar.gz, depending on which source format is used [14:39] agateau: let me bang you up some demo code [14:39] desrt: thanks! [14:40] agateau: http://fpaste.org/c4GB/ [14:41] desrt, well, depends, typically you will get patches in the diff.gz for dx packages [14:41] seb128: i see it here for libxml2 [14:41] desrt, because we bzr merge -c in the packaging vcs to backport upstream fixes [14:41] desrt, which is handled by the vcs without needing a patch system [14:41] so I'd like to get gnome-boxes into Precise, don't know if it'll happen by feature freeze though, prerequisites are bug 931783 & bug 931803 [14:41] Launchpad bug 931783 in ubuntu "Please package libosinfo 0.1.0" [Wishlist,New] https://launchpad.net/bugs/931783 [14:41] Launchpad bug 931803 in spice-gtk "[needs-packaging] Please package spice-gtk 0.9" [Wishlist,New] https://launchpad.net/bugs/931803 [14:41] desrt, I think libxml is "old school debian" [14:42] seb128: seems to be [14:42] debian packaging is so broken :) [14:42] jbicha, I think we should delay that to next week and ffe [14:42] jbicha, it's universe should be easier to get, we still rush for stuff in main this week [14:42] (or at least, debian packaging as it was apparently originally intended to be used) [14:42] desrt: I assume this needs style support to work? I don't see any difference with Ambiance [14:42] * agateau tries Adwaita [14:42] agateau: yes. probably. [14:43] seb128: yeah by next week, spice-gtk should just be a sync from Debian, maybe libosinfo too [14:43] if ambiance doesn't support it it should start supporting it soon [14:43] because gnome is starting to use it [14:43] jbicha, did spice-gtk made it to debian already? [14:43] seb128: the new version is waiting in the new quee [14:43] *queue [14:43] jbicha, ok, we can sync when it goes through, debian new is not slow nowadays [14:44] desrt: nicer with Adwaita, it turns into segmented buttons [14:45] agateau: yup [14:45] desrt: I like your solution, but I would need to create my own "class" and thus it won't work well with any theme which does not explicitly support this [14:45] mmm [14:45] agateau: themes will need to support this [14:46] agateau: i think the correct course of action here is to poke cimi for getting support in ambiance [14:46] and then proceed as if you know you have it [14:46] desrt: ok, assuming this is not a problem, is it possible for the theme to make the buttons overlap? [14:46] no. probably not. [14:47] gtk won't let the theme get involved in size allocation to that extent [14:47] and in general it's not possible to have overlapping widgets in gtk [14:47] * agateau is messing with GtkFixed right now, don't tell desrt [14:47] you want some array effect like [ / > home > agateau ] ? [14:47] *arrow [14:47] yes [14:47] ya. that will be quite difficult. [14:47] or rather design team wants [14:48] well [14:48] depending on what is supposed to happen when you hover/click [14:48] you could very obviously (and easily) "cheat" [14:49] but that probably would only work for when you're not interacting with it [14:49] is there a video somewhere of the intended behaviour? [14:49] you mean put the whole arrow as part of one of the button? [14:49] i was thinking half-and-half [14:49] but ya... that's the basic idea [14:50] iirc software-center widget used to just put the whole arrow in the button on the left [14:50] it would be extremely easy for you to do that, obviously [14:50] and honestly, i'd leave it to the theme to implement it [14:50] and this way it can be done on the theme side [14:51] like put "linked" and "breadcrumb" attributes on it [14:51] and then let the theme do what it pleases [14:51] the css-based styling gives enough power to do this, i'm pretty sure [14:51] if it doesn't understand "breadcrumb" then at least it will be linked [14:51] that sounds easier to do than what I was about to try [14:52] I am going to check with Cimi then [14:52] thanks! [14:52] the power of css! [14:52] ;) [14:52] :) [14:52] on the actual 'code' size, you will end up with an extreme trivial GtkButtonBox subclass, i guess [14:52] *extremely [14:52] yes, I already have that [14:53] so if Cimi agrees, I am basically done :) [14:53] you should join #gtk+ on gimpnet [14:53] i'd be happier if you talked to Company or cosimoc about this [14:54] they probably have some thoughts on the topic and they both know the area better than i do [14:54] I would rather not get people yell at me for butchering upstream (cf. comments didrock got on his blog post about display settings) [14:55] i think the yelling only comes when there is no attempt to collaborate [14:55] asking for opinions first is a good way to avoid the yelling [14:55] maybe they already have some (better?) plans here and you are just duplicating efforts [14:56] mterry, hey, can you do me another favor? https://bugs.launchpad.net/ubuntu/+source/indicator-printers/+bug/932759 [14:56] Launchpad bug 932759 in indicator-printers "[MIR] indicator-printers" [Undecided,New] [14:56] mterry, it's a new indicator should be easy to review [14:57] seb128, nothing like FF to bring MIRs out the of the woodwork :) [14:57] desrt: that is true [14:57] mterry, ;-) [14:57] desrt: I assume you are on this chan, in case things get nasty? [14:58] mterry, well that one is a combo, it was upload, get it NEWed (thanks didrocks), now need a mir and promotion [14:58] agateau: yes. of course. [14:58] mterry: just my conditional note on the build: the priority needs to be extra and the description be fixed [14:59] desrt: ok, let's go [14:59] * desrt goes to have a shower [14:59] =) [14:59] (you asked if i was on the channel... not if i would be active) ;) [15:04] agateau: http://diylol.com/meme-generator/socially-awkward-penguin/memes/says-hi-in-gtk-gets-kickbanned [15:05] desrt: so much for trying to be polite [15:07] Laney, kenvandine: so, want me to try and build without mozjs, or are you on it already? [15:10] kenvandine: or could we build against firefox-dev somehow? [15:11] pitti, i don't think so, but haven't looked at it in quite a while [15:14] hi guys, I've got a small enhancement for the nautilus-terminal project, remapping the F4 toggle button to F7, thus not toggling the terminal when closing overlaying apps with alt+f4 [15:21] maybe someone here has upload rights to the nautilus-terminal project? :) [15:23] pitti - build what against firefox-dev? i bet you don't really want to do that ;) [15:23] chrisccoulson: I'm just disabling the mozjs plugin from libproxy for now [15:23] pitti - oh, you couldn't build that against firefox-dev anyway [15:24] well, you might be able to do it by linking against a 31MB libxul.so, but there isn't a separate libmozjs in firefox [15:24] it's all statically linked ;) [15:24] agateau: was that as bad as you thought? :p [15:25] desrt: nope :) [15:25] desrt: I expect some snarky comments when I pasted the mockup url, but that didn't come [15:26] agateau: i can give them to you now, if you like :) [15:26] "ew. dark theme? you weirdo." [15:26] desrt: ah thanks, that feels better [15:27] todays' updates broke my custom navigation shortcuts. I filed bug 932807 [15:27] Launchpad bug 932807 in unity "applications grab custom navigation shortcuts" [High,New] https://launchpad.net/bugs/932807 [15:27] and also now g-c-c is crashing [15:27] :( bad day for an update [15:28] jibel: be thankful you didn't update yesterday :) [15:59] bah [16:00] mterry, I'm hiding in shame, the german and french let go through without noticing the missing translations and you the american spoted it [16:00] ;-) [16:00] * kenvandine rofl [16:00] kenvandine, there is nothing fun there, NOTHING :p [16:01] i am used to being the one missing that stuff :) [16:01] mterry, thanks for the review and nice catch ;-) [16:04] has anyone experienced the following: http://paste.ubuntu.com/843187/ with gnome-control-center, to be precise when trying to access the network settings [16:10] sil2100: I saw that you removed some of the _descritpion in the test, which makes the result difficulty readable. Not your fault, I should have done the review better, but just a warning of being careful :) [16:10] mandel, can you get a backtrace using gdb? [16:12] seb128, sure, let me take a look [16:14] seb128, :) [16:14] sil2100: is there any progress on the compiz side btw? [16:18] seb128, should I bee looking for something? [16:19] mandel, no, just pastebin the backtrace [16:19] ack [16:20] didrocks: wait, what description ;)? [16:20] Ah, from checkbox? [16:20] sil2100: yeah [16:21] sil2100: it screwed the result collecter though, I just pushed a fixed version. The issue is that checkbox is not forgiving on the format :/ [16:21] good night everyone! [16:21] 'night pitti [16:21] didrocks: oh my, sorry about that! [16:21] hmm [16:21] have a good night pitti [16:21] pitti: goodnight [16:21] sil2100: no worry, just to tell you we have to be really careful with it :) [16:22] Ouch! Now I see what you mean [16:22] How is it possible? I didn't even remember removing those ;) [16:23] Those probably got cut out by my misuse of vim and 'd' [16:25] seb128, I hope this is all the info you need from gdb: http://paste.ubuntu.com/843221/ [16:25] mandel, report a bug using ubuntu-bug if it's detecting the segfault, the stacktrace is not that useful it needs debug symbols or retracing [16:27] seb128, ubuntu-bug does not detect it, do you know how can get the debug symbols? [16:27] mandel, are you sure it doesn't? [16:27] mandel, ls /var/crash [16:28] mandel, check /var/log/apport.log [16:28] you might want to rm stuff in that dir and try again [16:28] * mandel looks [16:30] seb128, apport complains with the following 'Could not determine the package or source package name.' [16:30] seb128, let me see the report of the crash [16:31] mandel, do you have an apt source containing the binary which is hitting the bug? [16:31] what is the command in the report? [16:31] what dpkg -S on it says? [16:32] seb128, the command in the report is: ProcCmdline: gnome-control-center [16:33] mandel, ok, dunno, to debug with pitti tomorrow [16:33] seb128, and dpkg -S gives the following: http://paste.ubuntu.com/843228/ [16:33] didrocks: about that compiz keybinding bug - I'm in the middle of investigating right now [16:34] sil2100: ok, thanks :) [16:34] mandel, sorry I've no time for that today, ff tonight, come back tomorrow [16:34] Sam told me to check the most recent source version of compiz-core, and it's still there [16:34] sil2100: the hot list for compiz is https://bugs.launchpad.net/compiz-core/+bugs?field.tag=compiz-ff-precise [16:34] seb128, ok, I'll ping him about it tom, he is in germany atm, right? [16:34] seb128, no problem :) [16:34] mandel, yes [16:34] So I'm looking into it, and it seems that once a keybinding is modified, it stops working [16:34] weird that sam asked you that, we are already on tip of compiz-core :/ [16:34] mandel, but pitti hit a bug in the panel earlier so I suspect there is an issue and we will get reports about it [16:34] seb128, I'll just change the proxy settings via gconf-editor [16:34] But when it's not touched, it works, so hm, will keep you informed [16:34] sil2100: ok, thanks :) [16:34] didrocks: thanks for the list ;) [16:35] mandel, oh, it's the proxy, yes, pitti reported the issue earlier [16:35] didrocks: hey [16:35] seb128, do you know the bug number so that i can listen to it? [16:35] didrocks: did you know about https://bugs.launchpad.net/ubuntu/+source/bamf/+bug/924471 ? [16:35] Launchpad bug 924471 in bamf "Developer doc is empty in devhelp" [Undecided,Confirmed] [16:35] mandel, it's probably bug #932644 [16:35] Launchpad bug 932644 in gnome-control-center "gnome-control-center crashed with SIGSEGV in g_type_check_instance_cast()" [Medium,Confirmed] https://launchpad.net/bugs/932644 [16:36] mandel, pitti's stacktrace was similar [16:36] seb128, superb, thx for the help! [16:36] mandel, yw [16:36] desrt: oh interesting [16:37] desrt: trevino assigned it to himself, I think that the doc generation is maybe broken, I'll have a look [16:37] awesome. thanks :) [16:40] kenvandine: hey, around? [16:41] pitti: remmina got promoted to main, but there is a security issue in it: LP: #760381 [16:42] pitti: oh, wait a sec, it may be fixed already [16:43] * mdeslaur got confused by date of bug [16:44] hey didrocks [16:44] kenvandine: small question on the video lens [16:44] pitti: yeah, false alert, please ignore me [16:44] kenvandine: you have two sources, is that on purpose? can't that be merged in one? (maybe easier for the maintainance) [16:45] davidcalle, ^^^ [16:45] didrocks, i didn't really question that [16:45] kenvandine: also, no COPYING file in unity-lens-video :/ [16:45] i think the remote scope might get updated more often [16:45] there should be... [16:46] he merged my branch adding it :) [16:46] maybe it didn't make it into the tarball [16:46] kenvandine: I just apt-get source the one you sponsored :p [16:46] kenvandine, the COPYING was in only one of the sources [16:46] :) [16:46] not in the manifest [16:46] whoops :) [16:47] davidcalle, can you add that to MANIFEST.in ? [16:47] kenvandine, done ;) [16:47] debian/source/format should be (quilt) or version 1.0 [16:47] not native [16:47] oh... i didn't bzr add it either :) [16:47] it's not a native package :) [16:47] didrocks, i removed that... grr [16:47] maybe i uploaded the wrong source :) [16:48] kenvandine: no worry :) [16:48] didrocks, reject it and i'll re-upload :) [16:48] didrocks, is that in both of them? [16:48] yeah, I'm finishing looking at it! [16:48] kenvandine: just in one for now [16:49] pitti, hi [16:50] davidcalle: kenvandine: it's weird that po/en_US.po isn't turned in a mo file with a local build [16:50] oh [16:50] davidcalle: kenvandine: not import for ubuntu with the langpack but something should be broken in setup.py [16:50] the problem is the debian dir is in the orig tarball! [16:51] ah, also [16:51] the lens refers to /usr/share/unity/5/lens-nav-video.svg [16:51] which isn't installed by the package [16:51] that should be it for the first one, rejecting === rodrigo__ is now known as rodrigo_ [16:54] davidcalle, can you create tarballs without the debian directory in them? [16:54] kenvandine, sure. [16:58] so same thing on scope-video-remote [16:58] no COPYING [16:58] same debhelper hint, davidcalle, should be 8 [16:59] same remark on the (native) for source format [16:59] also, just to be nitpicky: ${shlibs:Depends}, is not necessary :) [16:59] didrocks: can you check this #930091 - is this blocked by FF? [16:59] https://bugs.launchpad.net/activity-log-manager/+bug/930091 [16:59] Launchpad bug 930091 in activity-log-manager "offer deletion of existing events for newly-added blacklists" [Undecided,Confirmed] [17:00] anyone here who can help me on a sync problem? [17:00] m4n1sh: sorry, missing context, can you illustrate with an example? [17:01] kenvandine: davidcalle: reject the scope as well [17:01] didrocks, no shlibs for both? [17:01] davidcalle: yep [17:01] Ok and debhelper (>= 8) [17:01] do not forget debian/compat to bump to 8 [17:01] didrocks: example. you blacklisted gedit. so a dialog box will come up and ask "do you want to delete all the previously logged events for this application?" [17:01] Yep [17:02] I want to sync the "pnm2ppa" package with Debian, once Debian's package needs all our needs, and second, to get rid of a broken upstream source tarball. [17:03] m4n1sh: it will need a FFe, yeah [17:06] probably then I will avoid it. [17:06] didrocks: any build system changes are not part? right? like the --with-ccpanel [17:07] and it doesnt probe itself and has to be passed explicitly [17:09] m4n1sh: nothing too fanzy, but avoid changing build system [17:10] okay [17:31] seb128: fwiw, I'm glad for mterry's g-c-c patch to go in instead of mine, it's all good [17:32] kenvandine, i have a little patch here in works for gwibber lens :) [17:32] charles, great, sorry that you guys ended duplicating work [17:32] he didn't see the ticket so we didn't know each other was working on it. life goes on [17:32] charles, we need to be a bit better in track cross team specs next cycle [17:32] well, it /was/ tracked ;) [17:33] yeah, too much tracking :p [17:33] doubled ;-) [17:33] anyway I agree with you wrt the alignment in my patch & am glad you can swap another one in before freeze [17:33] hm, when exactly is the freeze? [17:34] I needed to talk to mterry about something else, do you know what tz he's in? [17:34] sil2100, in 5 hours [17:34] charles, thanks [17:34] seb128: wrt to blocklist hosting - what about putting it somewhere under *.debian.net (with a redirect to debian mainainer's server)? They have the same issue with squeeze version of transmission so we could maybe use the same place - and it's going to be much easier to get a hosting there ;) [17:34] charles, east coast u.s [17:34] Holy shit! [17:34] charles, he should be around [17:34] kklimonda, would work for me ;-) [17:35] but of course the most urgent question is, Is Transmission 2.50 uploaded ;) [17:35] charles, yes, jbicha did it before I woke up today [17:35] ;-) [17:35] yay! [17:36] \o/ [17:39] mhr3, patches welcome, what does it do? [17:40] kenvandine, fixes the search for multiple words etc [17:47] ok [17:47] I got gnome-control-center 3.3.5 working with all our patches I think [17:47] I will upload to the ppa in a bit [17:57] seb128: awesome! [17:57] hey jbicha [17:58] jbicha, I had to revert some your changes, we can't drop the library that would break stuff that we integrate there [17:59] dropping the library is an upstream change, the headers still work [17:59] I still have the privacy, additional drivers panels etc in my System Settings 3.3.5 for instance [18:05] jbicha, that doesn't make sense [18:05] jbicha, you probably didn't uninstall libgnome-control-center1 3.3.2 [18:05] jbicha, i.e probably still have the old binary on disk [18:06] seb128: no, it doesn't need it, the library is part of the main binary now [18:06] * didrocks waves good evening [18:07] jbicha, I'm pretty surprised, i.e deja-dup depends on libgnome-control-center1 [18:07] jbicha, can you rm /usr/lib/libgnome-control-center.so.1.0.0 and confirm it still works? [18:09] well it might resolve the symbols in the shell binary but that seems something I'm not sure I want to rely on [18:09] seb128: yes it still works after deleting that file [18:09] jbicha, well anyway I restored the lib for that version, I will play with dropping if after ff when I've time to look at details [18:10] jbicha, and you can rebuild the rdepends without the lib? [18:10] jbicha, I'm surprised, only have the .h shouldn't be enough, you need the lib to resolve the symbols are build time [18:10] jbicha, I doubt stuff will link against the g-c-c binary [18:10] it's not a lib [18:10] seb128: yes, I tested rebuilding indicator-datetime for instance a month ago [18:11] jbicha, with /usr/lib/libgnome-control-center.so.1.0.0 deleted? [18:11] jbicha, upstream would have dropped the lib earlier if it was not needed, they kept it to be able to build gnome-bluetooth [18:11] jbicha, well anyway, I'm just really suprised if the lib was not needed at all, I will check, thanks for the infos, it's interesting ;-) [18:12] seb128: I built with my sbuild so it was a clean environment, and it doesn't need the library to run [18:16] jbicha, ok, I don't understand why ;-) [18:16] jbicha, it should need to resolve the symbols at build time and they are in a binary in /usr/bin, that's not somewhere the linker look for symbols [18:19] jbicha, ok, I don't understand why ;-) [18:19] jbicha, it should need to resolve the symbols at build time and they are in a binary in /usr/bin, that's not somewhere the linker look for symbols [18:19] jbicha, I'm interested to understand why it works but I will look after ff ;-) [18:19] jbicha, I'm also curious why bastien didn't drop the lib earlier if it was not needed [18:29] is sshd not working for anyone else on precise? [18:32] dobey, works for me [18:35] seb128: hrmm, it seems to be working now, after i restarted it a few times :-/ [18:48] desrt: ping [18:49] hi [18:49] TheMuso, I reverted your gtk uploaded, it turned g-c-c to segfault land and several people need to land and test g-c-c changes and I've no time to debug gtk [18:50] seb128: new upstream release causing problems? [18:50] TheMuso, please talk to me and get review before doing another gtk upload, we can't create too much installability issues around freeze time, and I would prefer getting a new tarball that playing backporting 15 commits [18:50] desrt, no, backporting of a stack of git commits [18:50] ah. ya. [18:50] desrt, https://i92984147.restricted.launchpadlibrarian.net/92984147/Stacktrace.txt?token=e21bc64c12937ea01a6eddd00ee120c1 [18:50] is the stacktrace [18:51] desrt, https://launchpadlibrarian.net/92923817/gtk%2B3.0_3.3.14-0ubuntu1_3.3.14-0ubuntu2.diff.gz is the upload [18:51] achiang: how can i help? [18:51] it seems broken as well, not sure what TheMuso did [18:51] + get_effective_focus_column@Base 3.3.14-0ubuntu2 [18:51] that doesn't seem an upstream symbol, or that's really buggy [18:52] in fact it's a commit from Company [18:52] not sure where the commit stack is commit from though [18:53] pitti, ^ your g-c-c segfaulting when trying to set a proxy was due to that [18:53] we got like a dozen bugs similar today [18:54] seb128: something to do with a11y. looks to be a bug upstream as well, actually [18:55] desrt, right, I just don't have time to debug it before ff and it's create lot of issues, I reverted for today, I guess TheMuso will follow up upstream [18:55] TheMuso, ^ can you do that, the bugs it created are listed in the changelog [18:55] seb128: it'll get fixed next time someone types 'make check' :) [18:56] desrt, ;-) [18:56] typical symbol leak mistake [18:57] desrt, btw g-c-c 3.3.5 is in the ubuntu-desktop ppa, I got it before ff! (not built yet though) [18:57] seb128: you're a brave man :) [18:58] ;-) [18:58] that being done -> dinner [18:59] bbiab [18:59] hum [18:59] i forgot to eat today, it seems [18:59] * desrt should have some lunch :) [19:05] hi, i've posted a bug report requesting a merge for scribus (from sid), and i've also attached a debdiff to take care of the merge to it. could someone upload it? [19:06] https://bugs.launchpad.net/ubuntu/+source/scribus/+bug/932962 [19:06] Launchpad bug 932962 in scribus "Please merge scribus 1.4.0.dfsg+r17300-1 (main) from Debian unstable (main)" [Undecided,New] [19:06] ockham: did you see bug 930639 ? [19:06] Launchpad bug 930639 in scribus "Please merge scribus 1.4.0.dfsg+r17287-1 (main) from Debian sid (main)" [Undecided,Confirmed] https://launchpad.net/bugs/930639 [19:07] jbicha: no, unfortunately :-( [19:08] jbicha: but that debian version is also a bit outdated since they're now at r17300 [19:11] so, anyone? === Amaranthus is now known as Amaranth [19:59] tedg, charles: Anyone want to review an indicator-messages branch for me? Doesn't need a release, I can distro patch, but I'd like another set of eyeballs before releasing unto the world. [19:59] https://code.launchpad.net/~mterry/indicator-messages/tell-accounts-services/+merge/93290 [20:01] mterry: larsu would be an excellent choice for that, he just got done doing patches for indicator-messages [20:01] I saw him in #systems just a minute ago but don't see him in here [20:01] charles, to be fair it's 9pm in europe [20:02] charles, so it seems it would be better to have u.s guys to look at it rather than asking lars to keep working [20:02] hm, you're right [20:03] ok then, I'll grab it [20:03] Do we have tomorrow to push things in too? Or just tonight? I can never remember [20:03] this is my first time through, so I don't know :) [20:03] mterry, nobody can, I asked earlier, you have still 23h57 min [20:03] mterry, they keep changing it :p [20:03] nice [20:03] mterry, it's 21utc tomorrow [20:03] Don't think I need it, but still good for disasters [20:04] mterry, do you want me to find a reason for you to need it? ;-) [20:04] seb128, I need a head start for UI Freeze! [20:05] hum [20:05] mterry, ok, if that's for bringing the bling I'm happy with that ;-) [20:06] actually, 24hr and 55 min [20:07] mterry: one superficial but useful trick I learned recently, you can replace all those "if (foo != NULL) { g_object_unref(foo); foo = NULL; }" blocks with "g_clear_object(&foo)" which is a little like g_clear_error() but for GObjects [20:07] charles, ooh, fancy [20:07] * mterry hates writing async gobject code in C [20:16] hmm, why do we have a poppler-0.18.3-0 source package, when sid is still at 0.16.7-3 but it is producing a libpopper-private-dev package for later migration, that we do not yet create. should I file a bug? [20:18] Sweetshark, no [20:18] dunno what debian is doing but that's broken [20:18] mterry: in "g_variant_get(foo, "(&o)", &bar)", the '&' before 'o' tells gvariant to return its own internal pointer so that we can skip an unncessary g_strdup()/g_free() sequence, is that right? [20:18] we updated and ported all the rdepends [20:20] seb128: "dd a libpoppler-private-dev package: it will contain the private poppler core headers, but at the moment it is empty to allow sources to migrate their (build-)dependencies from libpoppler-dev to it. [20:21] from the sid poppler changelog at http://packages.debian.org/changelogs/pool/main/p/poppler/poppler_0.16.7-3/changelog [20:21] Sweetshark, seems like the debian poppler maintainer is on crack [20:21] chrisccoulson: around? [20:22] chrisccoulson, if you are *hide* [20:22] ;-) [20:22] desrt, sort of ;) [20:23] seb128: please communicate that verbatim to _rene_ [20:23] * Sweetshark doesnt dare to. [20:23] lol [20:24] let's look at that tomorrow [20:26] charles, yah [20:27] charles, pointer lasts until the variant dies, obvi [20:30] *nod* [21:05] chrisccoulson, kenvandine, mterry, cyphermox, jbicha: the new gnome-control-center should be in the ppa, if you could test it that would be welcome, I want to upload a bit later if it's working ok [21:05] just make sure you got the new gtk or yesterday's one, the update today broke stuff [21:05] i.e don't blame g-c-c for the buggy gtk update :p [21:05] okie dokie [21:05] * dobey wonders who to beg to get twisted 12.0 into debian/ubuntu before tomorrow [21:06] you can still try with the buggy gtk but if like proxy segfault that's not my fault [21:06] dobey, good luck with that [21:06] dobey, i should have gwibber uploaded without the webkit-gtk2 depends today :) [21:06] if that helps at all... i suspect not [21:06] kenvandine: yay [21:07] well, it doesn't help with twisted :) [21:08] seb128: will test in a second. what's the good gtk version? [21:08] guess i can just propose a twisted package that patches in the bits we need/want, for now [21:08] cyphermox, 3.3.14-0ubuntu2 is the bad one [21:08] gotcha [21:08] cyphermox, so either 0ubuntu1 or 0ubuntu3 is good [21:14] seb128: any noteworthy changes? [21:14] nevermind me, I should just read changelog [21:15] cyphermox, the changelog is not very informative (yet), just see if stuff work fine for you [21:15] ok [21:15] cyphermox, you can easily see that you got the new version by looking at the top left icon grid which replaced the "all settings" label [21:16] cyphermox, wacom is the biggest change if you have one of those devices I think (I don't have one to test) [21:17] cyphermox, otherwise that's mostly lot of small visual improvements and bug fixes at different places, hard to summarize or give a big change [21:17] if things seem to work fine that's good enough info for me ;-) [21:17] I assume the accesibility graphical tab has larger labels on purpose? [21:18] yes [21:18] seems to work just fine to me [21:18] cyphermox, thanks for testing! [21:18] I wish it didn't shift so much as modules are opened [21:18] yeah, that's a bit disturbing [21:19] but I guess normal users don't change panels so much [21:19] oh well, in general it works, the rest is cosmetics [21:19] right [21:19] they most open the shell and click on one or directly access it by the dash or indicator [21:19] sure [21:19] now that reminds me I should fix the bluetooth pin dialog if possible [21:19] it shows up under most windows :( [21:23] mterry, do you still have your gimp update locally and forgot to push or did you overlook the vcs? [21:23] mterry, don't bother if that's the second, I can commit it [21:24] mterry, just noticed that the version I was about to upload is already in the archive, the vcs just didn't get it [21:31] seb128: Sure, apologies that it caused issues... I wouldn't have backported if orca upstream didn't explicitly ask me to backport them so that they wouldn't have been hounded by users on precise running orca, but I guess in the long term they can just suffer hangs as it is still a devel release after all. [21:32] TheMuso, sorry for getting angry before, today has been busy with ff rush and that fall as an issue in the middle blocking testing etc [21:33] seb128: I understand, I'll get upstream to talk to you first if such a situation comes up again. [21:33] TheMuso, well, gtk upstream roll a tarball every 2.5 weeks or so atm so I think we can wait a few extra days and get something tested [21:33] seb128: and you had every right to be annoyed. [21:33] TheMuso, I was mostly surprised at the stack of commits you backported, usually we backport some small fixes or wait for the next version [21:34] TheMuso, anyway I didn't have time to investigate but could you check if trunk has the same issue as your backports? [21:34] Yep. [21:34] seb128, dobey gwibber uploaded, no more depends on webkit-gtk2 :) [21:34] seb128: Once i've read up on it a bit more, I will do so yes. [21:34] TheMuso, basically pitti said that opening the network capplet and selecting proxy was enough to segfault it for him [21:34] kenvandine, \o/ [21:35] Ok. [21:35] TheMuso, do you get that issue on your version? I didn't update today so I didn't have a chance to try if that was doing the same here [21:38] seb128: Yup, crash, will test with git master. [21:38] TheMuso, ok, good, that gives you start point ;-) [21:38] TheMuso, thanks for looking to it [21:39] No problem, and apologies. I feel terrible. [21:39] But you live and learn. [21:39] TheMuso, if you get another version you want to upload please bunch stuff in one patch and put it in the vcs, I will review,test it tomorrow [21:39] kenvandine: yay [21:39] TheMuso, no need to feel terrible, it can happen to anyone, it's not like session was not starting so it was easy to overlook even with testing [21:40] Yep. [21:40] TheMuso, I just found it because it traced back what changed which could explain lot of segfaults in gtk coming from retracers today [21:40] Yep. [21:43] seb128: your new System Settings works here [21:45] jbicha, thanks for testing! [22:07] mterry, hey [22:12] seb128: So upstrea forgot to tell me that they had to rebuild gcc for things to work properly again, same with nautilus, so I dare say when we get the new tarball we will have to do the same thing. I'll check locally to be 100% sure. [22:12] oh, a robert_ancell [22:13] TheMuso, that seems like an abi incompatible change [22:13] desrt, ^ is that "normal" that a gtk upload requires rebuild of gtk users to not segfault? [22:13] seb128: Agreed. WIll check with git master anyway. [22:13] * robert_ancell hides [22:13] seb128: no. [22:14] seb128: but certainly not 'unacceptable' during an unstable series [22:14] that said, i don't know of anything right now that would require that [22:14] desrt, ok, let's see what that is about if,when it's confirmed [22:14] that seems weird to be [22:20] seb128, we are picking up the latest shotwell right? I'm packaging gexiv2 atm [22:20] robert_ancell, yes [22:20] robert_ancell, thanks for doing that update [22:20] robert_ancell, btw I'm updating g-c-c to 3.3 as we speak [22:20] robert_ancell, well rather I did the update today, I'm uploading [22:21] seb128, btw I'm going to try and do a mad crazy sprint today and get gnome-screensaver to look like u-g so we have a solution for 12.04... [22:21] robert_ancell, \o/ [22:21] it's not going to be pretty code wise though :) [22:21] robert_ancell, technically it's uif work, i.e you still have a week [22:21] robert_ancell, but better to land something and improve then ;-) [22:21] really? ok then [22:22] robert_ancell, no, forget that [22:22] I have it on public record :P [22:22] robert_ancell, (never remove the motivation to get stuff done :p) [22:22] damn [22:26] seems strange to have a week between FF and UIF, isn't that usually longer? [22:27] kenvandine, it was 2 weeks last cycle [22:27] i added something that was very ugly in gwibber specifically to motivate njpatel to make it prettier :) [22:27] i hope he does it :) [22:27] one month in natty [22:27] so yeah [22:27] it's shrinking :p [22:28] ok [22:28] seb128, i poked around in g-c-c [22:28] g-c-c 3.3.5 upload to precise, fear ;-) [22:28] seems ok [22:28] good [22:28] kenvandine, I would have sweared if you just told me a minute after upload that it was broken :p [22:29] hehe [22:31] if it's broken I blame it on robert_ancell [22:31] he did all those gdbus ports! [22:31] seb128, what's broken? [22:32] i need to get away from the computer for a couple hours, 19 hours yesterday has wiped me out [22:32] davidcalle, i'll re-upload the lens and scope when i come back, just get me new tarballs [22:32] robert_ancell, nothing, I just updated g-c-c to the new serie, the ratio bug fixes,crazy changes was way in favor of the bug fixes [22:33] robert_ancell, I'm covering in case the update has issues ;-) [22:33] robert_ancell, but it seems to work fine so far [22:33] kenvandine, ok :) [22:34] robert_ancell, I reverted some stuff like the keybindings switch to gsettings though, but it seems it should be easy to get that done dynamically at runtime, if jbicha or ricotz look at doing that we might get the new gnome-shell after all for this cycle [22:34] seb128, sorry I missed that - are we going with g-c-c 3.4? [22:35] robert_ancell, yes [22:35] robert_ancell, that's what I was saying, I just uploaded 3.3.5 [22:35] seb128, what drugs did desrt give you? [22:35] lol [22:35] but yay \o/ [22:35] hey robert_ancell and RAOF see this bug.... do you think it is in greeter/lightdm or x? https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/932611 I originally thought compiz, but seems not so much [22:35] Launchpad bug 932611 in xorg-server "visual corruption on startup (before desktop)" [High,Confirmed] [22:35] robert_ancell, I spent a week looking what fixes I wanted to backport, and I was like over 35 commits and I figured it would be easier to update and back out stuff I don't want [22:36] robert_ancell, well "spent a week", look to it a bit and spent some time thinking about it, it didn't take me a week ;-) [22:36] jasoncwarner_, since after the multi-monitor u-g upload? [22:36] seb128, yaeh [22:36] jasoncwarner_: yeah, I think so...I saw it first time 2 days ago [22:37] robert_ancell, rather ^^ [22:37] hey jasoncwarner_ [22:37] robert_ancell: and I thought it was compiz [22:37] the multimonitor is not 2 days old I think [22:37] seb128: ! you should be sleeping, I'm pretty sure [22:37] jasoncwarner_: That's basically the long standing “there's nothing drawn on the screen between unity-greeter and g-s-d setting the wallpaper” bug, isn't it?+ [22:37] RAOF: not sure, but I do know that I had not had the issue before, it just started [22:37] jasoncwarner_, hang on, I haven't uploaded the latest u-g. [22:37] RAOF: makes book look pretty, uh, bad ;) lots of visual corruptions etc [22:38] RAOF, what happens if the resolution changes between u-g and the desktop? I'm leaving the root window there so if it stays the same it should be seamless (it was for me) [22:38] jasoncwarner_, RAOF, robert_ancell: just as a piece of info, when I updated to the new g-s-d (which was in the ppa for 2 days), my external monitor changed from showing the lightdm wallpaper tiled (login is 1440 because mirror, desktop 1920) to see nothing [22:38] but if it changes, no idea what the driver could do [22:38] during login [22:39] robert_ancell: Right. There's no sensible thing to do; except maybe blank. [22:39] like before I could see the lightdm wallpaper in 1440 and tiled around [22:39] seb128, does that happen with the u-g you tried last night? [22:39] since the g-s-d update I see a blank screen [22:39] robert_ancell, yeah, I go from login to nothing to desktop [22:39] which I find better than the tiled background [22:39] Well, that's an improvement. [22:39] none is perfect though [22:39] I think blanking is better than the visual corruptions or tiling for sure [22:40] antyhing that makes the user think something borked their desktop == bad ;) [22:40] jasoncwarner_, want to get your hands dirty and compile the new u-g and try it? [22:40] Well, what we really *should* be doing is fading to black, doing the modeset ensuring we end up at black, and then fading in to the desktop. [22:41] jasoncwarner_, if you want to give it a try, wget https://launchpad.net/ubuntu/+source/gnome-settings-daemon/3.2.2-0ubuntu17/+build/3210683/+files/gnome-settings-daemon_3.2.2-0ubuntu17_amd64.deb and dpkg -i it [22:41] RAOF, or being optimistic and having the resolutions line up to be seamless [22:41] robert_ancell: sure, can do...but probably not for a couple hours. otp with tim about compiz, unity and hud and then I have to do some x and gnome recruiting [22:41] I do not volunteer to make that happen this cycle :) [22:41] jasoncwarner_, just to see if that's gsd which changed behaviour for you [22:41] jasoncwarner_, (the deb assuming you are on amd64) [22:41] seb128: that will work for me [22:41] jasoncwarner_, sleeping> I would if we wouldn't have ff tomorrow :p [22:41] seb128: coffee? ;) [22:42] lol [22:42] maybe not, I might not be in bed yet but I do plan to go soon ;-) [22:42] I just learnt earlier that ff was tomorrow 21utc [22:42] I though it was 0utc [22:42] that gives me an extra day [22:45] jasoncwarner_, RAOF, robert_ancell: see lp bug #850885 [22:45] Launchpad bug 850885 in gnome-control-center "(oneiric) greeter screen displayed incorrectly with external monitor during login" [High,Triaged] https://launchpad.net/bugs/850885 [22:46] bryceh: thanks... [22:46] RAOF, do you know of anyone doing nested X servers? I was just thinking we could get the multi-session composition benefits of Wayland using a nested X servers and I can't think of any technical reason why it wouldn't work (suspect based on Xephyr that there isn't the support for this being fully functional though) [22:48] I don't know of anyone doing nested X servers; the problem would be forwarding on all the protocol stuff to the base server. Currently you can bring up a nested server, as long as you don't want acceleration, 3D, randr, etc. [22:48] bryceh: mine is slightly different, no tiling...mine is all jumbled and broken (I'll use my phone to take a pic) [22:49] jasoncwarner_: What driver? [22:49] RAOF: nvidia [22:49] GHreau [22:49] AONEuaohenudaoenstuha [22:49] ! [22:49] RAOF: no, not that one [22:49] ;) [22:50] RAOF: was that the cat? [22:50] This is driver dependent; what you're getting is a framebuffer consisting of unitinialised memory. Which just happens to contain a bunch of stuff from previous sessions. [22:50] No, that was me being annoyed with the nvidia driver. [22:51] Also the tiny, tiny netbook keyboard that I'm using to try and verify that an older nvidia driver is less sucky at FBOs for DBO. [22:51] seb128, have you seen the crazy strobe effect when importing in shotwell? [22:51] robert_ancell, I didn't try to import recently [22:51] RAOF!!! [22:51] hows it going babe? [22:51] jasoncwarner_, if you know which day the change occurred, your /var/log/dpkg.log can be used to narrow down what package(s) might be suspect [22:52] robert_ancell, could be https://bugs.launchpad.net/light-themes/+bug/931630 [22:52] Launchpad bug 931630 in light-themes "don't set GtkDrawingArea to have transparent background" [Undecided,Triaged] [22:52] seb128, I just did an import of ~/Desktop on a blank database and it seems to flash the default gray background between each photo being imported [22:52] jasoncwarner_, then you'd downgrade each package one by one until the problem resolves, and then we know what to blame [22:52] bryceh: ok...I'll take a look [22:52] seb128, yeah, there's a lot of rendering stuff going crazy from light themes it seems [22:52] robert_ancell, if you read the emails from adam he mentioned that is causing some rendering issues, Cimi will look at it after ff he said [22:52] jasoncwarner_, you may still have the older debs for things in your apt cache (/var/cache/apt), or can download them off launchpad if not [22:53] RAOF: that sounds exactly like what I'm getting, btw...what you described above [22:53] seb128: oh that was fast, I was going to copy g-c-c to the gnome3 ppa but I guess I don't need to now :) [22:53] robert_ancell, just drop the http://bazaar.launchpad.net/~ubuntu-art-pkg/light-themes/trunk/revision/184#Ambiance/gtk-3.0/gtk-widgets.css GtkDrawingArea 3 lines and see if that solves it? [22:53] jbicha, yeah, ff is on us soon ;-) [22:53] jbicha, it seemed to work fine so I went for it [22:54] jasoncwarner_, try gsd first in case [22:54] yeah, it works [22:54] jasoncwarner_, I'm sure it changed the lightdm to desktop background transition for me [22:54] seb128, that fixes it [22:54] robert_ancell, cool [22:54] robert_ancell, I will make sure Cimi look at it [22:55] kenvandine, bug #931630, can you maybe revert the GtkDrawingArea change when you get a free slot? [22:55] seb128, is there a bug for it? [22:55] Launchpad bug 931630 in light-themes "don't set GtkDrawingArea to have transparent background" [Undecided,Triaged] https://launchpad.net/bugs/931630 [22:55] robert_ancell, ^ [22:55] robert_ancell, or feel free to do it if you want [22:55] seb128, I'll link it to shotwell so no-one goes looking for the problem [22:55] we can sort if with Cimi later [22:55] robert_ancell, thanks [23:00] is Cimi breaking more stuff? ;) [23:07] jasoncwarner_: I'm not sure what we can do about this for nvidia, except perhaps raise a bug with them. Possibly there's something we can do in X. [23:08] RAOF: my question would be why I'm getting it now, but never got it before? [23:09] RAOF: I still have to try the g-s-d that seb128 mentioned, but, yeah, why now? [23:10] robert_ancell, pong [23:10] mterry, oh hey, I was just going to say there's a small bug in switching to greeter that is stopping the lightdm release, but otherwise is good to go [23:12] robert_ancell, sure, k [23:16] seb128 RAOF just tried the latest g-s-d and still have it, fyi. [23:20] jasoncwarner_: Does the engineer role you posted require any fluency in a specific language? [23:20] I didn't notice a mention of any [23:20] bkerensa, the X position? [23:21] bryceh: I would have to look [23:22] bkerensa, the X position would require fluency in C, and python/bash skills would be desirable [23:22] it was the other one :P [23:22] not the X [23:22] The GNOME engineer