[07:12] good morning desktoppers [07:20] Morning oSoMoN [07:23] hey duflu [07:39] good morning [07:47] duflu: should i965-va-driver be installed by default in bionic? [07:47] tjaalton, yes it's required... [07:47] well it's not here [07:48] just installed bionic the other day [07:48] tjaalton, but it comes from ubuntu-restricted-addons [07:48] ah [07:48] So many installs might be missing it initially [07:48] tjaalton, I was about to attempt bisecting Mesa... unless you're more practiced at at [07:48] at it [07:49] would be better to test 18.0.0-rc1 first [07:49] tjaalton, sounds optimistic :) [07:49] we'll get rcN in a week or two [07:49] in bionic [07:50] is that corruption bug hw specific? [07:50] seems fine on kbl [07:51] though the video clip probably matters too.. [07:53] yeah [07:54] reproduced the bug now [07:54] all I see is some static [07:55] I fear we might end up blaming clutter-gst-3.0 for "always having done it wrong". But at least bisecting where the problem started in Mesa will be useful [07:57] I have 18.0.0-rc2 on a ppa but it's built against libglvnd, so I'll test that combo first [07:58] broken [07:59] now this machine is tainted, going back is painful :) [07:59] so I'll grab another one [08:10] tjaalton, OK I've got my Mesa bisection groove on [08:10] Give me a while [08:47] morning gang [08:58] Morning willcooke [08:59] tjaalton, bisection complete. See the upstream bug. Looks very easy to revert [08:59] morning willcooke [08:59] getting mesa through -proposed is anything bug [08:59] but [09:00] Yeah fair enough [09:00] And morning didrocks [09:00] hey duflu [09:00] I'll see that it's fixed in 18, once we roll out glvnd [09:04] * oSoMoN does the little ifinallymanagedtobuildalibreoffice6snap dance [09:06] on to testing it actually works [09:06] tjaalton, would it be correct to assume a single patch update is too much trouble? [09:06] duflu: it's the racy/broken autopkgtests that need hand-holding [09:07] tjaalton, OK. No problem... I'll leave it with you and upstream [09:07] I'll push 18.0.0-rc3 to proposed early next week [09:11] good morning desktopers [09:11] hey seb128 [09:11] hey willcooke, how are things in the u.k today? [09:11] bit chilly, little bit of snow, but the sun is coming out [09:12] andyrock, bug #1637984 - verification done on artful and xenial [09:12] bug 1637984 in udisks2 (Ubuntu Bionic) "Disks shows all mounted snaps" [High,Fix committed] https://launchpad.net/bugs/1637984 [09:12] thanks for fixing it! [09:12] Morning seb128 [09:13] hey duflu [09:13] willcooke, thanks for testing the SRU :) [09:13] and yeah, good work andyrock [09:13] seb128, the totem bug has been bisected in Mesa. Looks easy and justifiable to undo [09:13] way to get things done properly, he worked through several iterations with upstream until they were happy, fixing tests issues not related to his changes on the way [09:17] willcooke, remind me - did we have any conversation that might have touched on this? https://trello.com/c/ptr7qnhJ [09:18] duflu, we kinda did. It needed doing but no one had time to work on it so it got pushed to the bottom of the heap [09:18] oh [09:18] duflu, good, that makes more sense than being totem itself [09:18] willcooke, OK. Was there a card/bug for it before today? [09:18] right, was just going to say, I think I misread it [09:18] duflu, willcooke, https://trello.com/c/9GI3EFh2/122-bug1686081-if-synaptics-is-installed-gnome-mouse-touchpad-settings-doesnt-work [09:18] There was one about U7 needing fixes [09:19] thanks seb128 [09:19] np [09:19] seb128, wouldn't that entail more work? We could resolve the main issue by automatically removing synaptics on upgrade [09:20] good morning seb128 [09:20] lut oSoMoN [09:20] duflu, doing what is more work? [09:21] seb128, making gnome-control-centre support synaptics would be extra unnecessary work, compared to just removing the synaptics package as part of the upgrade process [09:21] duflu, did anyone suggest doing that? [09:22] seb128, yes that's the problem discovered in today's bug. The whole panel doesn't work because he upgraded from an older series where synaptics was installed [09:22] duflu, the suggested fix was to make unity-control-center work with libinput so synaptic can be made to be forced removed without creating issues if you have GNOME & Unity both installed [09:23] So upgraders don't get the control panel [09:23] duflu, that's https://trello.com/c/X7gznOHG/132-libinput-support-for-unity-7 [09:23] seb128, OK so that's a completely different issue [09:23] well, it's the same [09:23] we need to remove synaptic [09:23] but doing that creates issues for unity [09:23] duflu: you still have unity (and such unity-control-center) installed on upgrade [09:23] which is why we couldn't do it [09:23] so* [09:24] didrocks, Ah yes. OK [09:27] I think we are going to need to force remove it and the unity guys are going to need to fix it [09:27] duflu, I think you wanted to keep the option to let users install synaptic instead of libinput though ... does it still stand with the recent improvements? [09:28] seb128: +1, maybe post that on the community hub on their unity topic as a head's up? [09:29] didrocks, yeah, I was thinking the same, I'm going to do that now [09:34] seb128, the reasons for reverting from libinput to synaptics are *mostly* gone. Not completely [09:36] It might also be easy to copy 'synclient' code into a patch against gnome-control-centre. [09:36] upstream is not going to want that so it would be a distro patch [09:36] and doesn't seem a good use of our resources to maintain that [09:37] we pretty much have agreement than libinput is the way to go so we should better focus on making that good enough [09:39] seb128, I think the argument against doing Unity work on it is at least as strong. Unity7 is legacy _and_ at least Unity7 users have a command line tool (synclient). If they really prefer Unity7 strongly and don't switch often then just reinstall synaptics [09:41] right, we (as our team) don't plan to do that work [09:41] the community people who picked up unity might though [09:41] up to them [09:43] seb128, what we could upstream is a change to gnome-control-centre to detect if libinput is not in use... maybe. And add a message to the control panel [09:44] Although presently it only talks gsettings or whatever. So looking for a real backend might be going too far [09:44] that would be useful [09:45] but if we remove synaptic on updates it shouldn't be a common situation [09:45] Although it does conditionally show/hide touchpad settings already. There is some detection there [09:46] right [09:46] seb128, yeah removing it on upgrade would probably create the fewest bug reports [09:46] I'm unsure if we should conflict to avoid people installing it and then getting their g-c-c that stops working [09:46] * duflu is assuming more people don't complain about libinput's performance [09:47] but at the same time we know by experience that if we do that some people are going to copy an "apt install ...synaptic" line from the internet [09:47] not read the apt text [09:47] and say yes to uninstall GNOME [09:47] or g-c-c [09:47] and then not understand why they GNOME session went away [09:47] hum, good point… [09:47] seb128, conflicts result in questions during upgrade right? (I forget, it's been so long) [09:48] should we just do that via do-release-upgrade thus? [09:48] didrocks, that would be my preference [09:48] sounds good to me [09:48] not ideal because then some people might install it for unity or because they read on a forum it solves their libinput issue [09:48] and then have g-c-c to bug [09:48] yeah… [09:48] seb128, wouldn't conflicting packages then prevent people from having both installed later? [09:48] maybe we should add a warning to the panel saying so [09:49] seb128: I guess that worthes it, shouldn't be complex [09:49] * seb128 adds to trello [09:49] I'm happy to deal with that later on (just testing a file on disk) [09:49] seb128 willcooke thanks! [09:49] can be a bug fix anyway [09:49] didrocks, thx [09:49] good morning andyrock :) how are you doing today? [09:49] and good morning all! [09:52] seb128 fine fine [09:52] I'm planning to ask upstream if they are willing to accept an Ubuntu sso goa provider [09:53] nice [09:57] didrocks, duflu, the other way that would be robust would be to have the conflict and delete synaptic from the archive so it's not an apt away, but I don't know if some users still need synaptic for valid reason and if that's unnice to the unity people [09:58] seb128, certainly it's very useful. And the only input stack that is easily configurable on the command line. [09:58] libinput isn't there yet [09:58] seb128: yeah, also others DE might require synaptic as well? [09:59] didrocks, I think tjaalton said unity was the only remaining one [09:59] only unity is unable to configure libinput [09:59] oh, interesting [09:59] well, maybe try to convince the community unity people doing this work and remove from the archive :) [10:00] good luck with that :) [10:01] well, they decided to take the maintenance over, so, this isn't a free "there is nothing to do, easy" :p [10:02] yeah, but after trying to make it happen for ~2y(?) while unity was the thing and it didn't, now there's maybe a month or two to make the same :) [10:02] * duflu switches to chef mode [10:02] Later [10:02] of course can just drop the driver and say "live with it, or adapt" [10:03] duflu: enjoy cooking :) [10:03] duflu, enjoy your evening! [10:03] tjaalton, right, that's what I was wondering but duflu believes it's still useful to have and cover cases that libinput doesn't [10:06] right, it's not much of a maintenance burden [10:08] https://trello.com/c/ptr7qnhJ/211-bug1733032-touchpad-settings-dont-work-after-upgrading-to-1710-because-xserver-xorg-input-synaptics-is-still-installed updated for those interested [10:08] it has a summary of what was discussed here [10:09] I have a desktop that I just upgraded to bionic where it doesn't have any settings for the mouse :) [10:09] which is weird [10:09] right, we need to solve that [10:09] well, just for the left/right button, but no speed adjust [10:09] is that because synaptic is still installed after upgrade? [10:10] or another issue? [10:10] something else [10:10] :/ [10:10] synaptics isn't installed [10:10] it probably doesn't detect your touchapd as one [10:10] like it thinks it's a mouse or something [10:10] a fresh install on a laptop has it right [10:10] we had a few such reports in the past [10:10] it's a desktop [10:10] oh [10:11] the keyboard has a nipple mouse [10:11] same, maybe it doesn't detect the right device [10:11] could be [10:11] feel free to open a bug [10:11] maybe upstream as well if you can [10:11] against? [10:11] gnome-control-center [10:11] ok [10:11] thx [10:11] k, I need to step out for a bit, some errands and early lunch [10:11] bbl [10:12] Hi seb128! As a result of the util-linux upload, autopkgtest for open-iscsi failed (time out) for amd64. Maybe restart that test again? [10:12] see you later seb128 [10:12] https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#util-linux [10:15] GunnarHj, done [10:15] thx for keeping an eye on those [10:15] seb128: N.p. I was the noise maker. :) [10:22] filed bug 1748152 [10:22] bug 1748152 in gnome-control-center (Ubuntu) "Nipple mouse on a Lenovo Thinkpad USB keyboard is not detected" [Undecided,New] https://launchpad.net/bugs/1748152 [10:22] hm, changed the headline [10:27] chrisccoulson, is publishing chromium-browser 64.0.3282.140 from the stage PPA on your list? [10:28] oSoMoN, it can be added to my list with a bribe [10:28] (j/k) ;) [10:57] seb128: upstream agreed to have an UbuntuSSO provider in goa (a provider talking with snapd and not with the online server) [10:57] we need to maintain it of course [11:51] seb128: we might need an UbuntuOne icon for goa [11:52] should I ask the community to create one? [11:58] andyrock, I think we already have one [11:58] well, not for goa [11:58] Robert uses it in G-Software [11:58] not sure that's going to be ok [11:59] we need a squared one [11:59] ah right [11:59] http://www.softicons.com/social-media-icons/alike-icons-by-bokehlicia/ubuntu-one-icon [11:59] something like this [11:59] but with the proper licence [11:59] *license [12:03] andyrock, can we just cut the u1 part from the logo and use that? What size does it need to be? [12:04] 96x96 [12:05] they I can rescale them to all the other needed sizes [12:05] having a pure squared one will feel more gnome-ish [12:05] andyrock, oki, lemme see what I can do. Asking in #design for a high res icon that I can trim [12:06] willcooke: https://www.howtogeek.com/wp-content/uploads/2017/01/ximg_588a65fed1ffc.png.pagespeed.gp+jp+jw+pj+js+rj+rp+rw+ri+cp+md.ic.pht2CIym-v.png [12:07] something that looks like these icons [12:13] andyrock, do they add the border automatically? [12:13] nope [12:14] kk [12:27] andyrock, dont suppose you've found any design guidelines for those icons? [12:28] I've made something that looks "ok" [12:28] nice :D [12:28] well I didn't had time [12:28] I had to "ok to go" from upstream today [12:29] if we don't upstream this we can't use goa [12:29] not easily at least [12:30] andyrock, good news! [12:30] andyrock, where did you discuss that? IRC? bug? [12:31] seb128: I discuessed with rishi privately [12:31] k, cool [12:33] andyrock, https://imgur.com/a/Rgyx9 (it's a screenshot because otherwise you cant see the drop shadow properly) - WDYT? (cc seb128_) [12:33] kk for me [12:33] thanks [12:34] willcooke, looks good to my non-designer eyes [12:34] :) "that'll do" [12:34] wfm [12:34] :-) [12:34] :) [12:34] seb128: willcooke https://pastebin.canonical.com/209555/ [12:34] oki, andyrock I will send you the xcf so you can fiddle if you like [12:36] andyrock, thx [12:36] i'm preparing a ppa with the patched goa [12:36] hmm, it might need more of a border, hold on [12:36] I'm just having some troubles with quilt and images [12:37] andyrock: btw, Canonical pastebin isn't public [12:37] oh maybe that was intentionally private anyway [12:37] jbicha: i didn't feel confident sharing a private conversation with asking rishi [12:37] :D [12:37] *without [12:37] ok, that's fine :) [12:38] if you're interested I can send you the logs [12:39] no need, I'm not closely following GOA stuff :) [12:39] kk [12:39] seb128: with quilt I'm getting this for images [12:39] sorry about the noise [12:39] https://www.irccloud.com/pastebin/qoTmVu04/ [12:40] andyrock, you can't include binaries in a diff ... how did you include those? [12:40] it might be possible with git [12:41] ah so I don't include them in the quilt patch but just in the debdiff [12:41] no [12:41] hmmm [12:41] andyrock, what is in debian/source/format? [12:42] is that a 3.0 (quilt)? [12:42] like on Tuesday, I experimented with cherry-picking this patch using git-buildpackage and it seemed to work. I had to add the patch filename to debian/source/include-binaries though [12:42] 3.0 (quilt) [12:42] https://git.gnome.org/browse/swell-foop/commit/?id=17791c375 [12:42] andyrock, right, then you can add the image to debian/ [12:42] and list it in debian/source/include-binaries [12:42] and conveniently enough, Ubuntu 18.04's GOA is synced with Debian which does use git… [12:42] like you add a line "debian/icon.png" to that file [12:43] jbicha: yeah I'm using salsa for this [12:43] andyrock: is this only for 18.04? [12:43] jbicha: yep [12:44] so I switch to the 3.26.2-2 branch [12:45] then "gbp pq import" [12:45] then cherry-picks [12:45] then "gbp pq export" [12:45] but I can some noise with binaries [12:45] I'll try with debian/source/include-binaries [12:46] what's the exact error you got? [12:46] no error just weird chars in the quilt patch [12:47] I cannot even pastebin them [12:47] yes, it will look funny because it's a binary diff, but I think if you add the patch filename to debian/source/include-binaries it will work [13:01] giiiiiiiiiiiiiiiiiiiiiiiiiiiiiimp [13:01] why are you so hard to us [13:01] e [13:05] would be nice to have a simple "paint" [13:17] GunnarHj, util-linux retry worked, it migrated [13:17] seb128: Great! [13:18] seb128: I've played with pkgbinarymangler. If we skip the md5sums verifications in test/run for now, it builds on bionic. Would that be an acceptable workaround for now? [13:19] GunnarHj, it would be better to fix it, that's on my list for today [13:20] seb128: Absolutely. I have a patch ready if you give up. ;) [13:27] GunnarHj, can you share it in any case so I see how you worked around it? [13:28] seb128: Sure, then I'll submit it on the bug report. [13:29] thanks [13:33] seb128: https://bugs.launchpad.net/ubuntu/+source/pkgbinarymangler/+bug/1688994/comments/10 [13:33] Ubuntu bug 1688994 in pkgbinarymangler (Ubuntu) "dh-translations: cannot build POT without intltool" [High,In progress] [13:42] jbicha: for debian/source/include-binaries to work, the new icons have to be in debian/* ? [13:44] no, the patch itself is a "binary", so just add the patch file name (debian/patches/0001_make-it-more-awesome.patch) to debian/source/include-binaries [13:46] ahh [13:46] thanks [14:21] you know how it's super-click instead of alt-click to drag a window around by the middle, is that a Wayland thing or a gnome-shell thing? [14:22] willcooke: it's a GNOME default. You can change it back to Alt in GNOME Tweaks > Windows > Window Action Key [14:24] thanks jbicha [14:25] I saw talk of it the other day in here, did we decide to change it to alt-click? [14:25] Cos I /think/ we should. It's been that way for ever, and I only discovered it because I read it in here [14:26] hm, since my upgrade to bionic my mouse is unbearable fast. is there something to tweak this? [14:27] willcooke: but why are you familiar with it? is it more important to maintain compatibility with how things used to work in GNOME 2 or do things more logically for now? [14:28] willcooke, is that https://community.ubuntu.com/t/ubuntu-18-04-move-window-without-tilebar-by-press-and-hold-alt-click-inside-window/3801/4 ? [14:29] jbicha, fair question, but I think even Motif did it that way (alt-click), so I do wonder if we should stick with alt [14:29] willcooke, the issues from understand are that 1- other actions use super so it would be consitent, 2- alt is used by some apps and claiming it for shell actions creates conflicts and make some actions impossible to do in those apps [14:29] willcooke, e.g https://blender.stackexchange.com/questions/24473/alt-click-is-not-selecting-edge-loops-in-linux [14:29] ah, yes that's where I saw it [14:30] kk, then I think we need to socialise it a bit more. I can do that. [14:30] history/habits has values [14:30] but sometime it's worth forcing users to adapt their habits [14:30] right [14:30] and the nice shortcuts overlay will help :) [14:31] hey kenvandine! [14:31] right [14:31] but that's not for this cycle [14:31] yeah [14:31] this is yet another case for why we need that overlay [14:36] andyrock, this time you can move the disks card to done :) [15:36] https://forum.snapcraft.io/t/call-for-testing-libreoffice-6-0-0/3917 [15:38] and https://plus.google.com/+OlivierTilloy/posts/ADH83TytCab [15:38] oSoMoN, woooot [15:41] oSoMoN, woot [15:42] I wish I'd had it ready *before* FOSDEM, but better late than never… [15:42] * ricotz feels pressured now ;) [15:55] seb128: I think I'll upload adwaita-icon-theme 3.27.90 to bionic now https://git.gnome.org/browse/adwaita-icon-theme/tree/NEWS [15:55] oSoMoN, yeah, it's not late after :) [16:06] jbicha, sounds fine to me [17:24] nice work oSoMoN [17:37] didrocks, re: dock. Say I have two terminal windows open and I switch between them via the dock. The most recently used one is always at the top right? Personally I find that very confusing, sometimes the terminal that I want is the top one, and sometimes its the bottom one. Would it be possible to fix it (probably as an option? :( ) so that the windows are always the same place in the stack? [17:37] jibel: could you forward your nm-config-connectivity change to Debian? I mentioned it to mbiebl earlier and he didn't seem to understand why it was needed [17:42] oh nice, Khurshid is working on getting GOA working in unity-control-center (see the Community Hub) [17:57] willcooke: hum, I need to look at the code, I find the current (stack order) making more sense to me [17:57] willcooke: but it doesn't prevent to open a bug upstream [17:58] * didrocks finally started to get some reviews on G-S, but on styling and naming, so doing the changes (but profound changes incoming I guess) [18:15] didrocks, ack, thanks [18:19] oSoMoN: libreoffice is now blocking migrations. will you upload 6 before the weekend, or could you fix the current build? [18:23] night all [18:38] oSoMoN: never mind, upload building with the internal liborcus [18:43] doko, ack, that's good news because I have other urgent work before EOW [18:57] oSoMoN: crap, now fails with another error: [18:57] In file included from /usr/include/glm/gtx/norm.hpp:18:0, [18:57] from /<>/vcl/inc/opengl/VertexUtils.hxx:16, [18:57] from /<>/vcl/opengl/gdiimpl.cxx:39: [18:57] /usr/include/glm/gtx/quaternion.hpp:23:3: error: #error "GLM: GLM_GTX_quaternion is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it." [18:57] # error "GLM: GLM_GTX_quaternion is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it." [18:57] ^~~~~ [18:57] In file included from /<>/vcl/inc/opengl/VertexUtils.hxx:16:0, [18:57] from /<>/vcl/opengl/gdiimpl.cxx:39: [18:57] /usr/include/glm/gtx/norm.hpp:21:3: error: #error "GLM: GLM_GTX_norm is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it." [18:57] # error "GLM: GLM_GTX_norm is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it." [18:57] ^~~~~ [18:57] there is a reason why you should keep packages buildable ... [19:00] doko, that's already fixed in the 6.0 branch [19:00] https://git.launchpad.net/~libreoffice/ubuntu/+source/libreoffice/commit/?id=25c4af24b858087b3bd375ce8fad847b1affd484 [19:02] oSoMoN: I assume it's not yet ready for upload? [19:02] doko, "keep packages buildable", that's a joke right? the package are kept buildable, it's just that when things keep changing and create issues it takes time to keep up [19:02] obviously LibreOffice built fine a week ago [19:03] no, a week ago, it wasn't building [19:03] if there something buggy there is to upload a new version without handling fixing rdepends [19:03] https://launchpad.net/ubuntu/+source/libreoffice/1:5.4.4-0ubuntu2 [19:03] seb128: no, it's not a joke. it's real life [19:03] new version of glm [19:03] I didn't upload [19:03] right, neither did o_SoMoN [19:04] so do we have magic dwarfs and elfs fixing build failures? [19:04] your complain is that whoever synced glm didn't handle the transition [19:05] jbicha: ^^^ [19:05] doko, stop that, this glm update was done a week ago [19:05] doko, libreoffice is already fixed in the vcs [19:05] so it has been actively handled [19:05] seb128: it doesn't help if it's not uploaded. you know the current transition mess [19:05] it's just that updates sometime take more than a day [19:05] sure, a day would be fine ... [19:06] right, well talk to whoever starts those stack of transitions [19:06] not to those who are victim of the mess [19:07] transitons could be better prepared upfront [19:07] seb128: YOU ARE WRONG. most of of the unrelated ones are imported from debian. you are responsible for these as I am [19:08] doko, well maybe we should delete some stuff from proposed [19:08] and do transitions one by one [19:08] stopping the autoimporter for a while [19:08] there are way to deal better with those [19:08] liborcus and libixion are tied to the LibreOffice version, if we're not ready for LO 6 yet, then we don't need those new versions yet either [19:08] seb128: go ahead and propose that, and DRIVE that. in Oct/Nov I did that. [19:09] doko, I've other things to work on and the situation doesn't bother me [19:09] but I sympathize with you trying to sort that out [19:09] sure, until everything gets fucked up [19:10] we could flush proposed [19:10] and start reuploading in wanted order [19:10] oSoMoN: could doko upload https://git.launchpad.net/~libreoffice/ubuntu/+source/libreoffice/?h=ubuntu-bionic-6.0 now? [19:10] if really needed [19:10] shrug [19:10] I'm uploading a build with internal glm for now [19:10] jbicha, could be delete you glm update from proposed? [19:10] be->we [19:11] no, reverting would be worse I assume [19:11] why was that transition even started if the proposed situation is that complex? [19:11] the way out is usually not to pile more [19:11] glm wasn't supposed to be a "transition" [19:11]