[06:59] good morning [07:02] good morning desktopers [07:02] lut didrocks ;-) [08:34] seb128: just a reminder that dgadomski raised during the team meeting about his email on the desktop ML (mount options in nautilus). I'm unsure myself about the best options, but would be nice if we can answer there [08:35] didrocks, oh, right, thanks [08:35] I still feel like it's too big of a topic to be realistic that we handle it downstream === timchen1` is now known as timchen119 [08:35] but we can at least get the discussion going [08:36] agreed [08:37] didrocks, seb128: thanks guys, I'm aware that the scope is pretty wide, but any feedback is appreciated [08:58] morning desktoppers [08:59] hey willcooke, how are you? [08:59] back from London? [08:59] hey mlankhorst! Weird thing... Chromium on the N7 looks like red and blue are swapped. Did some Googling, I think it's an oddity in Chromium, few bug reports about the same thing on Android [08:59] seb128, yes! [08:59] hey willcooke [08:59] willcooke: that is probably a xmir bug :P [08:59] willcooke: does it happen with other opengl clients too? [09:00] * didrocks starts to be worried about every lightdm starting asserts failing in autopkgtests [09:00] mlankhorst, any suggestions of something to try? [09:00] didrocks, what asserts? lightdm doesn't start anymore in vivid? [09:01] seb128: from the tests, yes [09:01] didrocks, when did that start? [09:01] seb128: no idea, that's the issue with autopkgtests. We can just say last time it pass was wednesday 11 march [09:02] if only ubiquity would let me install under kwm… [09:02] kvm* [09:02] is that version blocked in proposed? [09:02] * seb128 checks [09:02] hum, no lightdm update since the 10 [09:03] yo [09:03] seb128: systemd is blocked, but I doubt the change making the tests failing is in systemd [09:03] hey Laney, wie gehts? [09:04] as even ubnutu5 is failing [09:04] hey Laney [09:04] didrocks, what's the issue with ubiquity/kvm? is that the vesa driver thing? [09:04] seb128: or a ubiquity/X11 race, couldn't get any mouse under it [09:04] had to reboot [09:04] (it's installing now) [09:05] oh ok [09:05] I'm installing a couple of week old image [09:05] that way, I will able to install, reboot, upgrade, reboot [09:05] if that passes, I'll try today's iso [09:05] let me download the current daily here [09:05] to see at least if that impacts real users [09:05] seb128: there was another bug report about ddms not starting btw [09:05] under systemd and upstart [09:05] how are those tests working? [09:06] could it be "just" xorg not working under kvm? [09:06] seb128: well, it does work outside of autopkgtests at least [09:06] hey seb128 et didrocks [09:07] didrocks, are we sure? [09:07] however, I tried an autopkgtest image, installed ubuntu-desktop on it [09:07] a few failures on glib's depends too [09:07] oh noes [09:07] seb128: well, I am under kvm, I have a xorg session here, so… yes? :) [09:07] bug #1433198 [09:07] bug 1433198 in xorg-server (Ubuntu) "15.04 daily images don't reach X using VirtualBox - VESA error: Cannot read int vect" [Undecided,Confirmed] https://launchpad.net/bugs/1433198 [09:07] didrocks, k, several people mentioned ^ this week [09:07] also desktop images failed to build [09:07] seb128: that can be a newer xorg-server issue [09:08] libqt5qml5 : Depends: libdouble-conversion1 but it is not installable [09:08] seb128: here, I'm on a week-old image [09:08] didrocks, yes, it is [09:08] * Laney looks in the direction of finland [09:08] so, I guess the update will tell us [09:08] Mirv ! [09:08] (I'll try to upgrade xorg independently) [09:08] mlankhorst, what's the status of bug 1433198? [09:08] bug 1433198 in xorg-server (Ubuntu) "15.04 daily images don't reach X using VirtualBox - VESA error: Cannot read int vect" [Undecided,Confirmed] https://launchpad.net/bugs/1433198 [09:09] https://bugs.launchpad.net/ubuntu/+source/double-conversion/+bug/1427677 [09:09] Ubuntu bug 1427677 in double-conversion (Ubuntu) "[MIR] double-conversion" [Undecided,Fix committed] [09:09] please to promote! [09:09] seb128: at least, that will explain why I can't get an autopkgtest image, where I add ubuntu-desktop on top of it to start xorg [09:09] lightdm logs are quite screwy though, only "starting Xorg, Xorg stopped" [09:09] with -d [09:09] seb128: just a sec I'll upload the fix.. [09:09] (nothing by default) [09:10] mlankhorst, great, thanks [09:10] got distracted by optimus [09:10] mlankhorst: do we have a ppa? I can at least try some autopkgtests here [09:10] that will answer if this is due to that quickly [09:11] [PATCH xserver] int10: Fix error check for pci_device_map_legacy [09:11] I'm confident it is.. [09:11] mlankhorst: I'm quite stuck on some autopkgtests to unblock systemd, so if you get some binaries ready, that would help :) [09:14] didrocks: sure, I'll upload in a bit [09:15] once test build finishes [09:16] hey desktopers [09:16] hey darkxst [09:17] Laney or seb128 can you land my fail whale fix? [09:17] uploaded [09:18] bug 1385572 [09:18] bug 1385572 in gnome-session (Ubuntu) "gnome-session not shutting down cleanly" [High,Triaged] https://launchpad.net/bugs/1385572 [09:18] Laney, double-conversion promoted [09:20] seb128: ! [09:21] seb128: thanks for the promotion, I mistakenly thought it'd be automatic once MIR is approved and the dependency arrives [09:22] Mirv, apparently you bugged the desktop iso build with your qt landing, I promoted double-conversion now, but it's annoying that things land in buggy state [09:22] attente_: I wonder what's the reason of 1433831? what's the difference between unity7 and unity8 desktops? [09:22] seb128: well the promotion shouldn't happen before landing either? [09:22] darkxst: ok in a bit [09:22] didrocks: uploaded to the archive, have fun :) [09:22] Mirv, no, but I wonder why we hit the issue [09:22] bbiab [09:22] Laney, are images built out of proposed? [09:22] happyaron: I saw that language-selector got uploaded, does it work? [09:22] no [09:23] proposed doesn't check components though [09:23] mlankhorst: would have been great to test beforehand but ok, let's see :) [09:23] and if not, why are things going through proposed-migration if they are not installable? [09:23] mlankhorst: also, I guess that means we need to have autopkgtests on xorg ;) [09:23] is this again train overpasses certain checks thing? [09:23] it's because CI train ignores them [09:23] in the archive it would have been depwait [09:23] but train doesn't drive proposed-migration?§ [09:23] Laney, thanks ;) [09:24] Laney, so proposed-migration doesn't check that things are actually installable? [09:24] I didn't say that [09:24] I said that the train doesn't check component mismatches [09:24] I didn't say you said that [09:24] ah ok [09:24] it's installable with universe [09:24] ok, makes sense now [09:24] correct [09:24] Laney, thanks [09:24] and the CI train lets main packages build with universe build deps [09:24] and copies with binaries [09:25] Mirv, so, yeah, it's because the ci train built without respect of the pockets and land built binaries [09:25] sounds more like a loophole, than a train [09:25] There's an option on the PPAs to respect components [09:25] Laney: it works, but gunnar didn't tell me he's uploading that... [09:25] don't know why they don't use it [09:25] happyaron: heh [09:26] is this enough for all cases? [09:26] Laney: it was enabled a year ago (we needed some fixes in launchpad first, which were done) [09:26] but I think they recreated new ppas [09:26] so probably the option wasn't enabled back [09:27] Laney, larsu, https://lists.ubuntu.com/archives/ubuntu-doc/2015-March/019404.html \o/ [09:27] Laney: almost, there are some packages (fcitx-frontend-qt5) removed unexpectedly from /target, but in general it works. tried two installs, one for en locale, another zh_CN. [09:27] Laney, can you handle the landing for larsu? you followed more than me those details I think [09:27] ok [09:28] happyaron: alright, good, gl with the fixes [09:28] probably missing a dependency to that package [09:28] guess so, haven't looked at the details [09:28] Laney: if you are going to upload gnome-session, can you also look at bug 1251281? [09:28] bug 1251281 in gnome-session (Ubuntu Vivid) "gnome-session-flashback/Unity fails to start without hardware acceleration in cloud/remote environments (Forwarded-X/XRDP/VNC/NX/X2GO/Chromoting)" [Critical,Triaged] https://launchpad.net/bugs/1251281 [09:29] seb128: \o/ [09:30] mitya57: I just went blind [09:30] mitya57: are you up to speed on it? [09:30] if so, you can do the uploading... :) [09:31] seb128: I'm skeptical about the xorg issue, I dist-upgraded an install, first with xorg, then kernel, now with everything and I'm getting to my session [09:32] Laney: shall I make an upload of im-config, everybody has fcitx as default now (current im-config behaviour) [09:32] Laney, yes it was on my todo list, of course I can do it myself [09:32] but let's see, I'll use adt-run -U option once xorg would have built in proposed [09:32] didrocks, k, I don't know then [09:32] Laney: that would mean the FFe is completed though... [09:32] mitya57: feel free [09:32] didrocks, need more details I guess [09:32] didrocks, Need me to build some virtual machines to test anything? [09:32] happyaron: what does the upload do? [09:32] do you have a diff? [09:33] Laney: make ibus default for everyone other than zh_* locales. current behaviour is: im-config prefer fcitx over ibus, but we never seed fcitx in the live cd, so ibus is default. [09:33] willcooke: I don't think you will get further more than I am right now, I tend to try with the xorg fix if adt is happy again, and we'll see (already spent some hours on this) [09:33] Laney: a minute [09:33] oki, shout if I can pitch in some help [09:34] didrocks, can you approve https://code.launchpad.net/~larsu/overlay-scrollbar/blacklist-gnome-boxes/+merge/251566 ? [09:34] willcooke: will do, thanks :) [09:34] didrocks, I think you are in the right team [09:34] I'm not [09:34] seb128: done [09:34] didrocks, thanks [09:34] thanks larsu for this patch btw :p [09:34] no more transparency on boxed! [09:34] boxes* [09:35] :) [09:36] It's still over me why we can't fix those from the theme though :-/ [09:40] Laney: https://launchpadlibrarian.net/200624463/im-config_0.27-1ubuntu8.debdiff [09:42] happyaron: ok [09:44] * Laney wonders why some i386 autopkgtests are having problems installing kde test deps [09:44] great [09:51] hey digitalalex :) How's the symmetry looking ;) [09:55] * Laney eyes adt-run [09:56] seems like it's hung unpacking libnotify4 [09:56] * didrocks eyes the archive publisher [09:56] I thought it was publishing every 10 minutes now [09:56] * Laney eyes didrocks [09:57] (last publishing was 25 minutes ago…) [09:57] * didrocks eyes Laney [09:57] :) [09:57] ogra_ also mentioned that publisher got slower again over time I think [09:57] read that on some other channel [09:57] dunno what's going on with it [09:58] ah xorg missed the publishing window… [09:59] :-/ [10:11] (at least, the xorg error is the same than in the bug) [10:13] didrocks, you get the error now? or you mean the autopkgtest one? [10:15] seb128: I used minicom to get into it while the systems are running [10:15] lightdm logs are really poor, (that's what is in the unit) and just says "lightdm failed to start" [10:15] so I looked at Xorg logs [10:16] which lightdm logs? [10:16] * didrocks didn't use minicom since 2002 [10:16] it has several logs, it's a bit confusing [10:16] seb128: the one showing up on systemctl status [10:16] so stdout, stderr + syslog forwarding [10:16] didrocks, k, those are not the logs I guess? [10:16] but even running it manually with --debug doesn't help apart from "xorg stopped" [10:17] yeah, it doesn't print anything actually [10:17] didrocks, /var/log/lightdm/lightdm.log and x-0.log probably have more details? [10:17] or X-0-greeter.log [10:17] seb128: I'm not in it anymore, lightdm.log got what --debug had (so not really useful in that case) [10:18] didrocks, the xorg errors/details are in x-0(-greeter).log [10:18] not sure about x-0.log, if it's the xorg logs, that would help I guess [10:18] right [10:18] it should [10:18] those should probably be forwarded to the journal (or rsyslog) [10:19] not sure [10:19] is syslog supposed to have the detailed logs from every service? [10:20] it would make it quite impossible to read [10:20] it's supposed to have enough logs to know what went wrong I guess [10:20] like warnings and so on [10:21] well, xorg log has the xorg log, even when there is no issue/error [10:22] but yeah, ideally real error would be fwded to the main log by some mean [10:22] indeed, and that's where the issue is [10:23] if they just print to stdout/stderr, it's not an issue anymore with the journal [10:23] but let's see, I'll poke robert [10:23] ah, finally xorg published in proposed [10:23] * didrocks tries [10:25] happyaron, Laney: I'm in doubt as regards the debian/rules changes in im-config at bug #1430893. If I understand it correctly, fcitx won't be installed by default unless Chinese is selected. There is still a Japanase remix which ships fcitx with the intention to make it the default IM framework. Considering that, wouldn't it be better to keep debian/rules as is in 15.04? [10:25] bug 1430893 in im-config (Ubuntu) "[FFe] Install Fcitx for Chinese users" [Undecided,Fix committed] https://launchpad.net/bugs/1430893 [10:26] GunnarHj: if not doing that, all live session user will have fcitx as input method [10:27] happyaron: Aha... [10:27] happyaron: Do you have some other idea how to solve the Japanese remix issue? [10:31] mlankhorst: the upgrade works and fix systemd X11-related tests! [10:32] didrocks, \o/ [10:32] mlankhorst: I really think we should have at least one autopkgtest starting lightdm and ensuring we have a xorg server running [10:32] willcooke: FYI ^ [10:32] seb128: phew! [10:32] ok, now overriding the failing autopkgtests as we discussed yesterday [10:32] didrocks, I don't understand why you didn't get the xorg not starting issue in your vm though [10:33] seb128: I don't know as well (on the whole install) [10:33] probably a driver different when starting qemu myself and under adt [10:33] (I'm using direct rendering) [10:33] could be [10:33] Laney, can we get a new desktop iso built? [10:34] Laney, with the double lib promotion and the xorg fix seems like worth starting a new build to see where we are now [10:34] seb128: xorg is still in proposed [10:34] oh, ok [10:34] is that systemd/upstart test getting fixed? [10:35] Laney: no, that was another one [10:35] the machine-id one? [10:35] not even :p [10:35] so, there is the upstart reboot issue [10:35] that we discussed yesterday [10:35] and identified the cause [10:35] then, they wanted to override the tests [10:36] but as it was one of the first test failing, I wanted that we run all the other tests still [10:36] and that's how I discovered the lightdm/xorg issue [10:37] nobody skipped the tests [10:37] Laney: Sorry about the l-s upload. Now I don't see the upload in the archive, neither in -proposed or -release. [10:38] what does 'rmadison language-selector' say? [10:38] Laney: hence the "wanted" [10:38] * didrocks feels now a little bit better on bug #1433728 should be the same [10:38] bug 1433728 in systemd (Ubuntu) "white screen after boot" [Undecided,Incomplete] https://launchpad.net/bugs/1433728 [10:38] hmm? [10:39] I guess there was another one, commenting as well [10:40] Laney: According to rmadison it's in vivid-proposed. (But not visible on the LP page.) [10:40] GunnarHj: Click on View full publishing history and you can see that it's being copied atm [10:41] Laney: Done. Looks promising. :) [10:43] Kind of surprised that this approach worked straight away. :P [11:30] happyaron: #1433831 i guess is because unity 7 runs X, so runs /etc/X11/Xsession.d/70im-config_launch. the upstart job requires IM_CONFIG_PHASE to be set by that [11:31] happyaron: and i guess the lack of initctl set-env in the job script is another bug that wasn't noticed because it already worked in unity 7 [11:44] seb128: Laney: was there some final thought about the train vs. MIR:ing? I can add the check to my Qt landing process (to ping archive admins at the right point), but it of course affects any train landings that do similar thing [11:46] Mirv: Can you find out why that flag isn't set on the PPAs? [11:46] Laney: oh, right so why the PPA:s allow building such things? [11:46] Perhaps because it's weird to promote things in the archive to satisfy builds in an external repository [11:47] blocking on that in landing PPA:s would make some sense since landing PPA is a sort of proposed-proposed anyway [11:47] but then again indeed it's weird to prepromote [11:48] best would be if proposed could detect it even with copy-binaried packages [11:48] anyhow, I'll add the check to my own process doc [11:56] (added to https://docs.google.com/a/canonical.com/spreadsheets/d/1xXhsnTwzvm5p6gDbUrDpRPSJV9tX4SHCptC9yUYCMmE/edit#gid=1330430878 for 5.5) [12:00] charles: do I need to merge indicator-power myself, or is that handled automatically? [12:01] oh, I can't merge it [12:02] tedg: ^ ? === MacSlow is now known as MacSlow|lunch [12:44] GunnarHj: any problem with uploading a new version of the default wallpaper? [12:44] xnox: want to check it? [12:45] Laney: yo, where is it? same bug report? [12:45] ya [12:46] Laney: No. We haven't changed the screenshots in the desktop guide yet, so just go ahead. [12:46] ty [12:46] damn [12:47] i got a ticket for Election Debates with Pink News, Evan Davis and all major parties and my question got selected. [12:47] but i can't go - as I have a meeting and a volleyball match =( [12:47] /o\ [12:47] miss them! [12:50] does lightdm do any sort of scaling on the greeter? [12:50] on vivid [12:51] I've got a system with a highres panel which starts with the native resolution but the display goes blank for a second and comes back with a lower looking res, or scaled. X log doesn't show resolution changing === alan_g is now known as alan_g|afk [12:53] Laney: i approve, both pngs look good with little artifacts and better symmetry. [12:53] Laney: the greeter logo doesn't align, but i ponder if the greeter is at fault here, rather than the wallpaper. [12:55] tjaalton, since yesterday u-s-d sets the scale factor on hidpi systems yes [12:56] xnox: ok, thx [12:56] the logo doesn't look equally spaced between rows to me [12:56] probably because it is heavier at the bottom [12:56] (independent of background image) [12:57] seb128: ok.. [12:57] tjaalton, is that looking wrong. [12:58] well it should happen earlier, it looks a bit ugly now [12:58] the greeter starts first with the native resolution [12:58] out of the transition [12:58] unscaled [12:58] does the result looks fine? [12:58] result is ok-ish to me [12:58] ok-ish only? [12:59] haven't compared to another one with same dimensions but 1080p [12:59] can't keep them both powered on :) [12:59] Laney: i do think it might be a greeter problem. Cause we overlay the grid pattern at an offset from the screen edges, But we should still, imho find the centre and do calculations from there [13:00] to the edges, sans padding. [13:00] * xnox will play with vala later. === MacSlow|lunch is now known as MacSlow [13:01] mdeslaur, Oh, I put it in a silo, but I just realized I forgot about it.... [13:02] Silos need push notifications and to be a little more annoying ☺ [13:06] tedg: ah, thanks! [13:08] willcooke: where are red and blue swapped in Cr on N7? Images? Text? Backgrounds? All? I can imagine something specific to Cr for the fractional pixels in drawing text. [13:10] qengho, hey! Just checking, I think it's everything in Cr [13:14] qengho, text is, images are too [13:14] mlankhorst said he thought it was likely an Xmir thing [13:16] definitely :P [13:16] are you running with a compositor or without? [13:17] without I think [13:19] mlankhorst: Wrong visual-class or something? I'll peek in the X layer details so I can answer questions you might have. [13:19] qengho: probably just xmir-dri2.c bugging, does it show when you run Xmir with -egl too? [13:22] * willcooke tries [13:22] mlankhorst, same [13:23] interesting, I thought it would have been a bug related to dri2.. [13:24] willcooke: what's the smallest testcase? [13:24] eg do i need icewm or anything? [13:32] mlankhorst, don't need icewm in order to see the issue, so I'd say not. [13:32] smallest testcase would be, install it, run it, witness the rainbow [13:32] (but icewm might make things a bit easier for you, if you've got a mouse) [13:32] ok I'll take a look :P [13:32] thanks mlankhorst [13:33] Cr is surprisingly fast on this N7 btw qengho [13:34] I compiled-in support for the GOFAST register. [13:34] Ha ha [13:35] I'm not surprised, willcooke. The N7 is pretty good, aside from the terrible "disk". [13:36] :D [13:38] bah [13:38] Laney, thanks for the theme landing, there was another approved branch from larsu fixing issues with the new gtk, would have been nice to land it with it, but I guess for next landing now... [13:44] seb128: is there a way to disable scaling the greeter? sometimes it seems to take a long time to get the greeter up, and I don't want to blame the i915 driver for nothing [13:44] tjaalton, no [13:44] ok [13:44] tjaalton, move the u-s-d xsettings plugin away [13:45] indicator menus are misplaced [13:45] tjaalton, the change was https://code.launchpad.net/~seb128/unity-settings-daemon/non-unity-sessions-need-scaling/+merge/250153 [13:45] we might want to exclude the greeter [13:45] the main goal was to get ubiquity/oem-setup scaling [13:46] ok thanks, I'll try that [13:46] tjaalton, can you open an unity-greeter bug about the menu thing? [13:46] sure [13:48] thanks [13:50] bug 1434094 [13:50] bug 1434094 in unity-greeter (Ubuntu) "indicator menus are misplaced when the display scaler is used" [Undecided,New] https://launchpad.net/bugs/1434094 [13:51] I had the old package cached, can try that on the other machine [13:56] tjaalton, thanks, I'm going to try on my non-hdpi laptop in a bit by forcing the value from u-s-d [14:03] ok it's not the scaler that's confusing my skl system but loading the background [14:03] or whatever the greeter does [14:03] oh well, bisect it is then [14:47] Trevinho is going insane [14:47] STOP MAKING MY PHONE BEEP [14:48] desrt: yeah... let's go back to IRC:.... way better :D [14:48] desrt: mute it :P [14:48] desrt: it's your baby... [14:48] * didrocks is happy to have mute is a while ago [14:48] desrt: your going to hate it... :D [14:48] desrt: but.... in few days it will calm down again [14:48] Trevinho: my baby is all grown up [14:48] Trevinho: ya.. no doubt [14:48] this happened in brussels as well :) [14:48] desrt: someone had to do this dirty job at some point [14:49] i'm sort of glad there is a hard limit at 200 :p [14:49] lol [14:52] updating my xmir setup, lightdm fails to log in probably because i haven't updated since systemd switch :P [15:06] seb128: quick question that you might be able to help with, On vivid desktop where are audio cd's mounted there is nothing in media, but if I plug in a usb device that shows in media [15:07] davmor2, audio cds are not data, they are not really mounted in the fs meaning, what would you want to see in /media exactly? [15:10] seb128: I just expected to see a folder for it in media :) [15:10] davmor2, yeah ... but no ;-) [15:10] NOTABUG [15:11] we could create a folder but it would be empty since there is no file on the disk [15:11] seb128: I didn't think it was a bug I just figured it lived somewhere else for audio cd's :) but now it appears they don't mount as such, learn something new everyday :) [15:11] oh ok [15:11] well, it's a gvfs mount on the desktop [15:12] it's a bit like if you mount a smb share [15:12] seb128: ah okay thanks [15:12] those don't have entries in the local fs [15:12] seb128: that makes more sense now ta :) [15:12] yw! [15:27] willcooke: it shows the correct colors here? [15:27] ?!?! [15:28] I'll risk an upgrade and see if that helps [15:28] xserver-xorg-core and xmir in archive are newer, need to downgrade afterwards if you do [15:29] apt-get install xserver-xorg-core=2:1.17.1-0ubuntu2+sa8 xserver-xorg-core-dbg=2:1.17.1-0ubuntu2+sa8 xserver-xorg-xmir=2:1.17.1-0ubuntu2+sa8 [15:29] something like that [15:29] thx [15:36] darn - ran out of space upgrading. [15:37] :/ [15:37] I don't see it happening on the desktop, could you give the spam Xmir gives on startup? [15:40] oops - device rebooted -> brick [15:40] I'll have to reflash [15:40] \o/ [15:41] oke, bbiab then :P [15:41] :D [15:41] can't wait to test if I fixed my little tegra or not [15:41] tested all day on radeon + intel, almost the same thing :p [15:41] just different hw [15:42] bb [15:42] didrocks: https://bugzilla.freedesktop.org/show_bug.cgi?id=85572btw [15:42] Freedesktop bug 85572 in core "Call /bin/dbus-uuidgen in systemd unit file" [Enhancement,Needinfo] [15:42] with a space [15:42]