[00:55] man, I just rebooted after accumulating a bunch of updates, and 3D performance is really poor with my intel graphics [00:55] eg, just opening the dash is really bad [00:56] interesting, .xsession-errors says: libGL error: failed to open drm device: Permission denied [00:56] libGL error: failed to load driver: i965 [00:56] I'm guessing that would explain it [00:56] permission denied? o_O [00:59] * jdstrand guesses it is the mesa update [00:59] jdstrand: you're supposed to do what everyone else does and blame apparmor first :) [01:00] hehe [01:00] I don't confine compiz or X [01:00] I probably should :P [01:00] then you fall back to blaming upstart :) [01:00] :) [01:01] http://i.qkme.me/3trz9o.jpg [01:01] hehe [01:01] hehehe [01:01] nice [01:01] perfect picture [01:02] :D === thumper-gym is now known as thumper [01:15] fyi, bug #1302270 [01:15] Launchpad bug 1302270 in xorg (Ubuntu) "Poor performance with recent update with i965: libGL error: failed to open drm device: Permission denied" [Undecided,New] https://launchpad.net/bugs/1302270 [02:36] jdstrand: your xorg.0.log is fine and from a recent boot but the xorg.0.log.old is using modesetting instead of intel and X was started again really late, guess you rebooted before filing it? all the dmesg info that might have helped was lost, might be worth attaching /var/log/kern.log that has the historical info [02:45] tjaalton: ping re: ^, arrandale and might be the mesa 10.1 bug you were talking about [02:52] Sarvatt: I did not reboot before filing. I logged in, noticed it, made a comment here, did ubuntu-bug and am still running it [02:52] * jdstrand is testing a bunch of stuff and can't really downgrade/etc atm [02:53] I don't know about the modesetting... I don't recall messing with anything like that in ages and ages [02:53] maybe I did at some point... [02:54] jdstrand: using kde? i think it might be https://bugs.freedesktop.org/show_bug.cgi?id=75723 [02:54] Freedesktop bug 75723 in Drivers/DRI/i965 "(regression since Linux 3.14?) brw_get_graphics_reset_status: Assertion `brw->hw_ctx != ((void *)0)' failed" [Major,New] [02:54] Sarvatt: unity7 [02:54] ah [02:55] I hadn't rebooted in several days at least, but had been upgrading all along [02:55] so I can't really pin point it [02:55] is any of this interesting: [02:55] [ 1.716252] fb: conflicting fb hw usage inteldrmfb vs VESA VGA - removing generic driver [02:55] [ 1.815666] fbcon: inteldrmfb (fb0) is primary device [02:55] [ 2.335498] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device [02:55] [ 2.342002] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0 [02:56] I have 965, but I can't remember if ^ is normal [02:56] yeah thats normal [02:56] I figured I'd downgrade mesa when I had a chance [02:57] your https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/1302270/+attachment/4064372/+files/XorgLogOld.txt looks bad but the current one looks fine.. [02:57] Launchpad bug 1302270 in xorg (Ubuntu) "Poor performance with recent update with i965: libGL error: failed to open drm device: Permission denied" [Undecided,New] [02:57] that's weird [02:59] I'm still in the affected session [03:06] Sarvatt: Same issue with being stuck in software rendering for me today :( Permissions are broken for both DRM and my software config too (I can't change servers any more) [03:14] duflu: [03:14] duflu: Fallout from the systemd upload? [03:17] today's changes to systemd shouldn't cause anything like that, though the upgrade may restart logind which if it can't read its state properly could cause this... in theory switching VTs may do the trick, otherwise logging out and back in should (or just reboot if you're in a hurry). [03:18] RAOF: I spy more (older duplicates). It's not specific to Intel graphics. But whatever sets the seat permissions isn't working (randomly) [03:18] Sarvatt: fyi, https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/1302270/comments/4 [03:18] Launchpad bug 1302270 in xorg (Ubuntu) "[regression] Poor performance with recent update with i965: libGL error: failed to open drm device: Permission denied" [High,Confirmed] [03:18] Also affects nouveau and my ability to change apt configs [03:18] seems duflu came upon it [03:18] * duflu goes to link up the dupes [03:18] if I add myself to the video group, it works-- so yeah, something changed with the perms [03:21] Maybe I won't link the dupes. But it seems to be a common problem with getting permissions (nouveau too) [03:24] RAOF: Yeah looking at the updates I got today, only systemd starts out as suspicious [03:25] systemd or cgroupmgr would be my high-suspicion targets. [03:25] Oh and libudev-dev [03:25] udev even [03:27] Oh, udev comes from source "systemd" [03:27] Indeed it does. [03:30] Yes! I reverted all the systemd packages from today's update and it's fixed [04:18] quick heads up, the last systemd change indeed causing this for machines without cgmanager due to extra error processing code which wasn't needed and caused an assert... [04:18] jdstrand is testing fixed packages, if those work, I'll push in the archive very soon [04:19] people affected may need to manually start logind after the update (not entirely clear if it'll do that on its own if it crashed before...) [06:52] hello [06:55] Laney, hi :) could you get a look at this branch: https://code.launchpad.net/~hikiko/unity-control-center/unity-control-center.clamp-ui-scale/+merge/214019 [06:56] it clamps the ui scale to a value at which the slider will remain visible [06:57] hiya [06:57] are things funny for anyone else on 14.04 today? [06:57] I seem to have no audio interface, just a "dummy interface" [06:57] and every now and then compiz is using 200% cpu [06:58] dholbach: https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1302264 [06:58] Launchpad bug 1302264 in systemd (Ubuntu) "systemd-logind assert failure: error.c:319: Assertion failed in nih_error_get: context_stack != NULL" [Critical,Fix released] [07:04] dholbach, I dist-upgraded a while ago and the only "strange" thing I ve noticed was some dpkg errors [07:06] hikiko, I didn't see any - which ones did you see? [07:07] dholbach, these: http://pastebin.ubuntu.com/7202139/ [07:08] hikiko, looks like it's bug 1302300 [07:08] Launchpad bug 1302300 in isc-dhcp (Ubuntu) "package isc-dhcp-client 4.2.4-7ubuntu9 failed to install/upgrade: subprocess new pre-installation script returned error exit status 2" [Undecided,Fix released] https://launchpad.net/bugs/1302300 [07:09] hikiko, is the situation fixed on your end already? if not, you might want to try another dist-upgrade; or run sudo dpkg --configure -a [07:09] no, -f didnt solve it so I guess I have to configure dpkg [07:12] dholbach, with dpkg --configure -a, and apt-get -f install it's not fixed so I suppose it's a bug [07:13] hikiko, did you try sudo apt-get update; sudo apt-get dist-upgrade as well? [07:14] yes [07:14] mvo, ^ [07:14] Preparing to unpack .../isc-dhcp-client_4.2.4-7ubuntu10_amd64.deb ... [07:14] /var/lib/dpkg/tmp.ci/preinst: 17: /var/lib/dpkg/tmp.ci/preinst: Syntax error: "fi" unexpected (expecting "then") [07:14] there's a syntax error in the installation script [07:14] mvo, do you have an idea how to recover from an issue like that if the usual tricks (dist-upgrade, dpkg --configure -a, etc.) don't work? [07:15] hikiko: you need to find a mirror that has ubuntu11 [07:16] ahhhh ok, that makes perfect sense - thanks sarnold [07:16] I changed my repositories to use ubuntu .de :) [07:17] let's see if it goes away [07:17] mvo, unping :) [07:22] mvo, can we make "apt-get changelog <...>" act on source package names as well? :) [07:24] +1 :) [07:25] dholbach: let me look into that [07:25] hello [07:25] hikiko: success I hope? :) [07:25] hikiko, better? [07:26] noooo [07:26] :( [07:26] well [07:26] dpkg [07:26] decided to remove isc-dhcp-common :D [07:26] so... [07:26] I cant run dhclient eth0 [07:27] and I have no clue how to install the dhclient :p could you help me? [07:27] hikiko: check /var/cache/apt/archives/ to find an old package.. [07:27] :D [07:27] well [07:28] the archived package [07:28] has the syntax error [07:28] so it cant be installed [07:28] you may have e.g. ubuntu7 or something? [07:28] no I ll download one [07:29] but wont there be any dependency issues? [07:30] mmm what if I download the daily build and change my sources.list to use the cd-rom to install the latest package? [07:30] ooof that sounds like a lot of work :) [07:30] yes :/ [07:30] is there any easiest way to do it? [07:32] hikiko: I'd use ip or ifconfig to set an ip at random, use ip or route to add a route to your gateway, then use apt-get install isc-dhcp-whatever=specific.version.number [07:32] yes!! static ip !! [07:32] hikiko: bed tie for me, good luck :) [07:33] thanks a lot sarnold [07:33] have a good night [07:33] thanks :) [07:36] morning [07:51] just for the record I have internet with my static ip but the isc-dhcp-common cannot be installed... [08:01] good morning desktopers [08:01] hey hikiko [08:01] hikiko, what's the issue/error when you try to install it? [08:02] line 17 in a script [08:02] just a sec [08:02] to switch desktop [08:02] and paste it [08:02] hey seb128 [08:02] pitti, hey, how are you? how is IoM? [08:03] seb128: small, green, cold, and wet :) [08:03] back :) [08:03] haha [08:03] pitti, how cold? [08:03] sorry there's a new error now: http://pastebin.ubuntu.com/7202265/ [08:03] seb128: about 8 to 10 degrees.. [08:03] brrr [08:04] hikiko, do you use an outdated mirror? [08:04] greetings [08:04] happy friday, how did that happen so soon? [08:05] hikiko, current version is ubuntu11, it should fix the shell script error you had earlier [08:05] Laney, happy friday! [08:05] which mirror do you use in your sources.list? [08:05] I used .de [08:06] no mirror, archive.ubuntu.com [08:06] ok I will change it then [08:10] hikiko: the idea seems reasonable [08:10] guys i dont know what was in todays update.. but 14.04 is barely usable for me untill i kill deja-dup-monitor .. compiz is also CPU hungry [08:10] Not sure if it's arbitrary to use the current window size though [08:11] Laney, seb128 [08:11] Maybe it could instead say 'a 640x480' (or whatever) window must fit on the screen [08:11] hello [08:11] hey! [08:11] also can you have the same issue with the gtk values or not? [08:11] I have a branch for review [08:12] Laney, seb128 there was a problem with the mpt's design for u-c-c: that when a toplevel win is larger than the screen size, the window manager places it to a predefined position and ignores the program's requests [08:12] so I ve done a new branch where we just clamp the values [08:12] https://code.launchpad.net/~hikiko/unity-control-center/unity-control-center.clamp-ui-scale [08:12] yeah that's what I'm looking at [08:13] this branch has another "issue" but bregma said it's acceptable [08:13] the issue is that [08:13] the window size depends on 2 parameters [08:13] 1- an integer scale factor [08:14] 2- the size of the widgets it contains that depend on their font size and number of characters etc and it's not predictable [08:15] It's why a dialog would be better, but I get that we are where we are [08:15] so, since in order to clamp the slider we take the window size into account [08:15] for different scale factors in gsettings [08:15] the user might see the slider clamped to a different value [08:16] eg: for scale factor 1, clamp=1.88 for scale factor 1.5 clamp = 1.75 [08:16] what do you call "clamp"? you limit the max value to what would put the dialog offscreen? [08:16] yes [08:16] not the dialog, the u-c-c window [08:16] there's no dialog [08:16] that's what I called dialog, sorry [08:17] but limiting the values range seems fine to me [08:17] so, the window never exceeds the screen boundaries [08:17] before we landed the slider I was already suggesting to limit to 2 [08:18] but the user might see different clamp values [08:18] because in practice nobody is going to want to go over 2 [08:18] well now it's limited to 1.75-1.88 in normal screens [08:18] because in 2 it goes offline [08:18] :s/offline/offscreen :p [08:18] sorry [08:18] :) [08:19] that approach seems reasonable to me [08:19] great :) [08:19] I just wanted to make sure [08:19] that there's no problem! [08:19] it's a bit hackish, but that should do [08:19] Laney, what do you think? [08:20] I was suggesting to use some reference size instead of the size of the current window [08:20] but the idea is fine [08:20] that would work for me as well [08:20] seems even easier [08:20] what do you mean by reference size? [08:20] don't get the geometry, just a fixed value [08:21] like "you can't make a 600px window go oversize" [08:21] or whatever constant [08:21] that's what I am doing [08:21] aaa [08:21] no [08:21] well [08:21] I might not explain it well [08:22] you do [08:22] instead of limiting the win size to [08:22] I limit it to [08:22] in both cases there will be an unpredictable maximum value [08:22] But you perform your calculation based on the height of the window [08:23] right? [08:23] no [08:23] what I do [08:23] is that: [08:23] + gtk_window_get_size (win, &win_width, &win_height); [08:23] I check that the window height [08:23] doesnt exceed [08:23] the screen height [08:23] because height <= width [08:23] and I want the window [08:23] (not necessarily, what about pivot setups?) [08:24] to be all visible [08:24] I used to choose [08:24] MIN (width, height) [08:24] but that was resulting [08:24] hikiko, can you please not use enter as punctuation? it makes the text really hard to read [08:25] ok sorry [08:26] We understand what you're doing, my suggestion is to instead say "int win_height = 600;" [08:26] But I don't think it's a big deal [08:26] I see :) [08:26] sorry [08:27] I thought you mean [08:27] win_height < 600 [08:27] It means you use a fixed size and say "A window of height 600px must always fit on the screen" instead of "This panel window must always fit on the screen" which you do at the minute [08:27] Laney, I'll fix this right now! I just misunderstood it! [08:28] you are right :) [08:28] that seems a bit less "random" [08:28] Well, see if it works - it'll make the code simpler [08:28] yes, it's a very good idea :) [08:28] seb128: can you test something for me? bdmurray said it didn't work for him but it seems to here... [08:29] Laney, I was just about to ask you to test something for me as well :p [08:29] let's trade testing! [08:29] woo! [08:29] Laney, https://launchpad.net/~ci-train-ppa-service/+archive/landing-017/+packages [08:29] Laney, the url-dispatcher workaround from ted [08:29] ok, first try to reproduce this please https://bugs.launchpad.net/ubuntu/+source/whoopsie-preferences/+bug/1261096 [08:29] Launchpad bug 1261096 in whoopsie-preferences (Ubuntu) "whoopsie-preferences respawns whoopsie in an infinite loop" [High,Confirmed] [08:29] it's "start a-l-m → diagnostics, watch top for whoopsie-preferences/init spinning" [08:30] while I put my possible fix into a branch [08:33] Laney, ok, init/dbus toping [08:33] dbus-monitor says [08:34] string "ARG0=/com/ubuntu/Upstart/jobs/startpar_2dbridge/whoopsie_2d_2dstopped" [08:34] Laney, there's a define: #define MINIMUM_HEIGHT 530 because according to gnome wiki: In order to fit on Netbooks screens all panels should be 675x530px [08:34] maybe I should just use this [08:36] seb128: lp:~laney/whoopsie-preferences/g-spawn-sync [08:38] man that guy loves doing things in upstart jobs [08:38] lol [08:38] I'd have expected some defensive measures in the code too [08:39] With this you can still get it created with the wrong permissions, they'll just get fixed up later [08:41] do you want to block the landing on that? [08:41] lemme see what happens now [08:43] Laney, I don't understand your patch [08:44] https://developer.gnome.org/glib/stable/glib-Spawning-Processes.html#g-spawn-sync [08:44] I guessed that the call was going wrong because those flags weren't set [08:45] * seb128 kicks google [08:45] Laney, first result gave me http://www.gtk.org/api/2.6/glib/glib-Spawning-Processes.html#g-spawn-sync [08:45] which doesn't have that note [08:46] also the note doesn't explain what happens when you don't use those flags [08:46] it feels weird it's looping on the command [08:46] maybe we know a guy who knows [08:46] desrt, ! [08:46] ;-) [08:51] Laney, that branch doesn't fix it [08:51] hmm ok, I didn't manage to get it again with that [08:52] Laney, fixed and it works very well in my monitor :) [08:52] hikiko: great [08:52] CrazyLemon, your deja-dup issue might be https://bugs.launchpad.net/ubuntu/+source/deja-dup/+bug/1302416 [08:53] Launchpad bug 1302416 in deja-dup (Ubuntu) "deja-dup monitor was taking 6GB of memory" [Undecided,New] [08:54] seb128 it might.. but there was a wide number of issues which are now gone - thankfully. I changed the server from si.archive.ubuntu.com to archive.ubuntu.com and that one had all the latest bugfixed which solved every issue. [08:54] good [08:55] is there any way i can close the bug report on LP? i opened a u-c-c bug repored and i can find the 'close' button [08:55] cant* [08:56] change the status to invalid [08:56] the yellow round icon next to "new" in the status column [08:57] i found it..thanks :) [09:00] Laney, things go crazy when whoopsie-preferences exit [09:01] yeah I got it to happen again in the end [09:01] it's easy to trigger, just kill whoopsie-preference while the panel is open [09:03] url-dispatcher seems okay [09:03] great [09:03] I didn't manage to get wrong permissions yet anyway [09:03] I still need to restart my session [09:04] to see if that fix my permissions [09:06] it did for me [09:08] Laney, the whoopsie issue is that whoopsie_preferences_on_set_report_crashes() keeps being called in loop [09:08] g_signal_connect (interface, "handle-set-report-crashes", [09:08] G_CALLBACK (whoopsie_preferences_on_set_report_crashes), [09:08] NULL); [09:10] hum, no [09:11] hiya [09:12] I don't know what just happened, but my keyboard settings changed to english on their own [09:12] do we have a bug report for that somewhere? anything I can do to debug? [09:12] maybe the ibus bug? [09:12] dholbach, hey, we don't know either ;-) [09:12] there's one we see in xubuntu where it always falls back to en_US [09:13] Laney, ok, I'm going to let you debug the whoopsie thing, it's weird, it seems like whoopsie-preference keeps respawning, I guess a bug in a-l-m's code [09:13] dholbach, did you use another session than Unity? [09:13] seb128, in the settings dialog it says that I'm still on German [09:13] seb128, not recently [09:13] I hate keyboards [09:14] :( === psivaa_ is now known as psivaa [09:14] dholbach: https://bugs.launchpad.net/ubuntu/trusty/+source/ibus/+bug/1284635 [09:14] Launchpad bug 1284635 in ibus (Ubuntu Trusty) "Keyboard layout changes after login" [High,Triaged] [09:14] possibly unrelated, but who knows [09:15] attente, not saying that because of you! [09:15] knome, ochosi: you guys are for sure nagging about that issue :p [09:15] dholbach, bug #1292412 as well [09:15] Launchpad bug 1292412 in unity-settings-daemon (Ubuntu) "Keyboard layout is always English(!)" [Medium,Confirmed] https://launchpad.net/bugs/1292412 [09:15] dholbach, can you change back using the indicator? [09:15] seb128: in all fairness, it's the first time *i* mentioned it [09:15] ochosi, knome: I didn't use another session and it was German up until a few minutes ago - it just changed [09:15] dholbach, is that a one time thing? [09:15] seb128, yes, first time [09:16] ochosi, yeah, that's true, I included you in the xubuntu group who keeps mentioning it ;-) [09:16] seb128, hum, let me see if I can re-enable the indicator somehow [09:16] dholbach, is unity-settings-daemon running? [09:16] seb128: and i only mentioned it cause it seemed ibus related and also fell back to en_US. so i think that's not too bad ;) [09:16] seb128, yes [09:16] dholbach, since when? [09:16] e.g did it segfault or something? [09:17] seb128: hi :) [09:17] desrt, hey [09:17] what am i supposed to know? :) [09:17] seb128, no crash file [09:17] many things, but not the thing that we thought was the thing [09:17] which turns out not to be the thing [09:17] so the thing is still an unknown thing [09:18] desrt, what happens if you use g_spawn_sync () with standard_output and standard_error being NULL without using G_SPAWN_STDOUT_TO_DEV_NULL and G_SPAWN_STDERR_TO_DEV_NULL [09:18] but the thing is still an interesting thing, but no longer the thing [09:18] they go to the stdout/err of the parent process [09:18] seb128, hey, it's our best bet... :) [09:18] dholbach, since when is the process running (ps should tell you) [09:19] desrt, well anyway, that's probably not our issue, there is a command that seems to be ran in loop, Laney spotted that error and though it could be creating the bug, but that's not it [09:20] seb128, which ps flag do I have to use for that_ [09:20] ? [09:20] dholbach, ps ux | grep unity-settings-daemon [09:20] seb128, since I rebooted the machine [09:21] k [09:21] 09:34 [09:21] but the issue started after that? [09:21] attente, do we have a gsettings key or something supposed to contain the current config? [09:22] seb128, yes, just 5m before I pinged you guys in here [09:22] dholbach, do you remember what you did just before it happened? [09:23] brb, need to restart my sesison [09:23] seb128: like under /org/gnome/desktop/input-sources? [09:24] dholbach: are you using an ibus input method? [09:25] attente, no [09:26] dholbach: if you add a different keyboard layout and switch to that, does it still persist as en_US? [09:26] attente, http://paste.ubuntu.com/7202445/ [09:26] ok, let me trz [09:26] try [09:27] attente, no, doesn't seem to have an effect [09:27] dholbach: it just happened randomly during the session or at start up? [09:28] seems like a similar description to bug #1292412 [09:28] Launchpad bug 1292412 in unity-settings-daemon (Ubuntu) "Keyboard layout is always English(!)" [Medium,Confirmed] https://launchpad.net/bugs/1292412 [09:29] attente, it happened during the session [09:29] attente, the "current" key should be 0 not 1 right? [09:29] in his dconf's pastbin [09:29] attente, during the session [09:29] what changes that index? [09:29] seb128: yeah, but what flipped it [09:30] does any log help? [09:30] unity-settings-daemon, indicator-keyboard [09:30] attente, that's a good question... is that source handled by u-s-d only, or is ibus touching it? [09:30] I think I had indicator-keyboard disabled for a long time [09:30] dholbach, those are in .cache/upstart [09:30] i don't think ibus is touching it [09:30] can't remember having seen it in months [09:31] it's not supposed to show if there is only one layout (I think) [09:31] it shows if there is one, but doesn't if there are none [09:31] unable to create file '/home/daniel/.cache/dconf/user': [09:32] dholbach, can you pastebin the full logs? [09:33] attente, seb128: no indicator-keyboard log [09:34] http://paste.ubuntu.com/7202465/ [09:35] shrug [09:35] .cache/dconf was owned by root.root [09:35] wth? [09:35] I have never touched it [09:35] I'm quite certain of that [09:35] what was the timestamp? [09:36] Dec 23 [09:36] well, dconf should use XDG_RUNTIME_DIR [09:36] dholbach, echo $XDG_RUNTIME_DIR [09:36] I installed the machine on Dec 22 [09:36]