/srv/irclogs.ubuntu.com/2020/06/25/#ubuntu-desktop.txt

=== guiverc2 is now known as guiverc
seb128goood morning desktopers!06:08
luna_morning06:32
dufluHi seb128 and luna_ 06:35
Laneyguten morgen!08:02
jibelcpaelzer_, is there anything in latest libvirtd in groovy that could affect networking? I'm trying to figure out why suddenly I've no network inside the guest.08:02
jibelthe guest is a focal live session with daily from yesterday08:03
dufluHi Laney08:06
Laneymoin duflu 08:15
seb128hey duflu, Laney, jibel, how are you in a summer day?08:30
dufluseb128, it's winter but sunny. How are you?08:31
seb128duflu, I'm alright, hopefully having a 'normal' work day today and I can get some work done, though it's going to be over 30°C and melting in front of the keyboard sucks a bit08:33
seb128we have extra heat until tomorrow and then rain over the weekend :/08:33
dufluTo be fair, a month into winter and we're only now below 20 degrees so it's OK08:33
jibelseb128, going well.enjoying the beach on the evening08:35
seb128jibel, :-)08:41
duflujibel is winning June08:42
Laneymoin seb128 08:45
Laneybit sad that all the nice days are during work time and then the weekend sucks08:45
seb128indeed :/08:46
* Laney does the 'rm fontconfig cache' trick to get emojis back in Chromium08:48
Laneywhy did I switch to Chromium like a month ago? I can't even remember any more, might go back to the nice Firefox deb :>08:48
Laneydear chromium, please exit when I exit you08:52
seb128Laney, if I understand that bug correctly it's snapd screwing the font cache which breaks deb apps as well (but maybe firefox doesn't rely on fontconfig for emoji?)08:52
seb128unless you do purge snapd along chromium which would fix the problem on your system08:53
LaneyI never saw broken colour emoji anywhere else08:54
Laneybut actually purging the cache didn't fix it in chromium either08:54
Laney🤷08:54
Laney(^- that one is coloured for me)08:54
seb128callmepk, hey, you mentioned you had wslu SRUs waiting for sponsoring, do you have merge requests for packages for those?08:54
LaneyFirefox works08:55
seb128Laney, I was commented based on https://bugs.launchpad.net/ubuntu/+source/chromium-browser/+bug/1858636/comments/31 also the noto-color-emoji is missing from the fontcache08:55
ubot5Ubuntu bug 1858636 in snapd (Ubuntu) "snapd generates incomplete fontconfig caches, result in emoji rendering issue in chromium" [High,Confirmed]08:55
seb128so I don't understand why it doesn't impact e.g gnome-characters08:55
Laneymaybe I have a different problem then08:55
seb128speaking of that bug08:56
seb128jamesh, ^ the snapd team bounced back to us, could you maybe help and reply to some of the questions for Ian on the bug?08:56
* jamesh looks08:57
seb128jamesh, start reading at comment 3708:59
seb128before it's noise of users and triager trying to understand what's going on and how to trigger the issue09:00
jameshseb128: I suspect the problem is down to freetype versions: I think snapd is running bionic's version of fc-cache statically linked against xenial's freetype09:15
seb128:/09:16
seb128why are they not using versions from the same serie?09:16
seb128also any easy way to confirm that? would be nice if snaps logged the fc-cache output09:17
jameshit's trying to build binaries that will run against the core snap's libraries09:17
jameshI think they assumed the cache differences were only down to fontconfig versions09:17
jameshI've got a general idea about how we can handle this better with private fontconfig caches for snaps, but haven't built the prototype yet09:18
seb128k09:18
jameshthis would be roughly one cache for each gnome platform snap we publish09:19
seb128jamesh, thanks, that's useful input ... can you comment that on the bug?09:19
jameshI'm writing the comment now09:19
seb128great09:19
seb128thx again!09:19
seb128Trevinho, did you notice that the bionic SRU for bug #1793496 failed to build?09:19
ubot5bug 1793496 in Mutter "scaling changes when closing/re-opening the lid" [Undecided,New] https://launchpad.net/bugs/179349609:19
seb128not only that bug, it's just the one reference, but that's the update to a git snapshot you did09:20
jibelLaney, is there a way to simulate an rst drive in a vm?09:32
jibelI want to test the rst code of the installer09:32
seb128jibel, ubiquity/plugins/ubi-prepare.py:        return os.environ.get('SHOW_RST_UI', '0') == '1'09:34
seb128try setting that env maybe?09:34
seb128well, depends if you just want to see the page or actually test the detection of the disk I guess09:35
jibelsounds like what I need. I'll have a look. However I want an RST drive and another that is not to verify that you can install on the second drive even if the first is bitlocker'd09:36
jibelotherwise I'll play with the code a bit09:37
Laneyjibel: not that I know of, no09:41
LaneyHack the detection code to return true for one of the drives or something09:42
jibelyeah, i'll do that09:43
jibelnot really sru'ish but well...09:43
Laneyah for testing instructions?09:46
Laneymaybe having an env var to do that would be good09:46
LaneyUBIQUITY_PRETEND_DRIVE_IS_RST=...09:47
ricotzgood morning desktopers!09:59
ricotzseb128, hi :), could you process https://launchpad.net/ubuntu/+source/nodejs/12.18.0~dfsg-3ubuntu1 which is in NEW10:00
seb128ricotz, hey ricotz, k10:10
ricotzseb128, thank you10:12
seb128yw!10:12
seb128ricotz, bug #1885006 is something for you or for Olivier when he's back?10:47
ubot5bug 1885006 in firefox (Ubuntu) "Search Service broken on Ubuntu Nightly builds" [Undecided,Confirmed] https://launchpad.net/bugs/188500610:47
ricotzseb128, thanks, this looks like an issue for both of us which would hit firefox 79 beta10:52
seb128ricotz, I will drop an email to Olivier so he saws it once he's back10:53
ricotzseb128, looks like a simple variable rename of upstream, so an easy fix11:08
seb128ricotz, great!11:08
seb128ricotz, do you have an easy way to flag that to Olivier or should I better drop him an email as a FYI?11:09
ricotzI can leave a comment in the bug report point to the fix11:10
seb128ricotz, thx11:11
seb128Laney, did you see the discussion on #debian-gnome about our geoclue delta the other day? do you have any opinion on that?11:32
seb128I'm pondering just dropping the delta and syncing back11:32
Laneyseb128: I did but it seemed Simon handled it with you so I didn't think about it much11:32
Laneylet me see what the delta is11:33
seb128Laney, it's making the lib depends on the service11:33
Laneythis is something from me from a few years ago11:34
Laneydid bigon say anything about adding it?11:35
LaneyI think it still makes sense, but it could maybe be a recommends as most apps which use geoclue can work ok without it11:36
Laneythat amazon launcher probably should have been made to not crash11:36
Laneyah I forgot bigon is in here :>11:36
seb128Laney, we discussed on #debian-gnome if you have the backlog of 2 days ago (iirc) but summary is that him and Simon think it's better to let the apps that need the service recommends/depends on it11:40
seb128Simon's argument was that most GNOME components that links again the lib don't really need the service11:40
seb128e.g calendar, initial setup, settings-daemon11:41
Laneythe way I read that conversation they both mentioned recommends11:41
seb128from the lib to the service?11:42
Laneyyes11:42
seb128 <smcv>seb128: in cases like libdbus-1-3 Recommends dbus, we still get occasional complaints about "installing $app pulled in $service, this is GNOME bloat"11:42
seb128so yeah, that discussion has argument for ang against the recommends 11:42
Laneybut libdbus still does have that recommends so ...11:43
seb128it feels like we failed to have a conclusion at the end, I will ask bigon again11:43
Laneythat was a mention of what sometimes happens11:43
seb128well just did I guess :p11:43
seb128right11:43
seb128Laney, thx, I try another round to see if the recommends could be added in Debian11:45
Laneysilly non team maintained packages!11:46
LaneyI would have done something about it in 2017 otherwise :p11:46
seb128:-)11:46
bigonfor https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=834176 I think I will close it as wontfix, there is no reason that the -dev pkg needs the daemon11:53
ubot5Debian bug 834176 in libgeoclue-2-dev "libgeoclue-2-dev should depend on package geoclue-2.0" [Normal,Open]11:53
bigonfor https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=864305, I guess a patch that conditionally add the Depends/Recommends for ubuntu is fine as a 1st step11:54
ubot5Debian bug 864305 in libgeoclue-2-0 "libgeoclue-2-0 doesn't depend on geoclue-2.0" [Normal,Open]11:54
bigonwas the m-a co-instability evalatuated?11:54
bigonTBH I'm way more annoyed by https://gitlab.freedesktop.org/geoclue/geoclue/-/issues/111 not being fixed11:54
Laneygeoclue-2.0 should be foreign probably11:55
ricotzis someone able to do a simple runtime test of a firefox build on s390x?12:19
ricotzhttps://launchpad.net/~ubuntu-mozilla-daily/+archive/ubuntu/ppa/+build/19506686/+files/firefox-trunk_79.0~a1~hg20200625r537255-0ubuntu0.20.04.1~umd3_s390x.deb12:21
seb128bigon, right, that upstream issue is annoying :/ also I got no traction when I asked on our side about the BTS bug12:28
bigonZeeshan said he will fix it "soon" but he's not opposed if somebody is doing it https://gitlab.freedesktop.org/geoclue/geoclue/-/merge_requests/70#note_52893512:30
seb128take over :)12:30
seb128ricotz, it starts fine, compared to the archive version that segfaults12:33
ricotzseb128, that is quite something then :)12:34
seb128:-)12:36
jibelxnox, I'm verifying bug 1847898. I created a VM in legacy mode with 3 disks and selected to install on vdc. grub-install runs on the /dev/vdc. 12:57
ubot5bug 1847898 in ubiquity (Ubuntu Focal) "System doesn't boot after installation - Legacy mode / 2 disks" [High,Fix committed] https://launchpad.net/bugs/184789812:57
jibelBut after installation the machine doesn't find a bootable device.12:57
jibelxnox, my test case is correct? 12:57
jibelis there any information I should dig for?12:57
jibelah actually wait 12:57
jibelhm, to make it boot I've to set disk3 as the first bootable device12:58
jibelnot sure it's specific to VMs12:59
jibel^ WDYT?12:59
xnoxjibel:  what is the drive order set in like libvirt / virsh / qemu command line?14:05
xnoxjibel:  or like can you ask to bring up the boot menu, and select the third drive, it then should boot fine.14:05
xnoxjibel:  multi-disk selection in bios is a bit unpredictable.14:06
xnoxjibel:  unlike UEFI where  we create ubuntu entry, and order it to be first, even though it's disk3 install.14:06
jibelxnox, yes, it boots fine when I changed the drive order. It's more a pb with libvirt/qemu which didn't try all the available drives14:06
xnoxjibel:  it does, if one passes more options to it.14:07
xnoxjibel:  i use virt-manager, and i can tick which drives to try, and in which order. I think there are qemu cmdline options / xml for that too14:07
xnoxjibel:  <boot order="1"/> inside the <disk></disk in xml14:08
xnoxand let me check qemu cmdline options14:08
xnoxjibel:  ,bootindex=2 on the -device .... stanza14:09
xnoxthen it will try them all, in the defined order.14:09
xnoxjibel:  but grub-install matches the selected drive means we did the right thing!14:09
jibelindeed14:12
jibel5.6 didn't improve my sound issues. I keep losing hdmi output and all that remains are dummy entries14:24
ricotzseb128, please don't forget about https://launchpad.net/ubuntu/+source/nodejs/12.18.0~dfsg-3ubuntu114:53
ricotz(it is a required build-dep for firefox 78+)14:54
seb128ricotz, I know, Olivier asked me to get it uploaded before his holidays, I did the nagging and LocustusOfBorg uploaded14:56
ricotzseb128, ok, I thought you are on ArchiveAdmin duty today14:57
seb128ricotz, I'm not really doing archive days but I'm happy to process pings between other things14:58
seb128ricotz, binaries accepted now14:59
ricotzI see, thanks!14:59
seb128yw!14:59
ricotzmaybe tweak this a bit - https://wiki.ubuntu.com/ArchiveAdministration#Archive_days14:59
seb128ricotz, it has a warning 'Warning /!\ This is currently being re-assessed to a more task oriented approach, rather than regular admin days. '15:24
seb128but yeah15:24
=== ijohnson is now known as ijohnson|lunch
Laneysee you monday!17:17
luna_Rolling Rhino :D17:28
=== ijohnson|lunch is now known as ijohnson
callmepkseb128 I create a MR for wslu to fix the build issue, tested on 1804 and the build should work now22:37

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!