[05:52] <oSoMoN> good morning desktoppers!
[06:01] <jibel> morning
[06:11] <oSoMoN> bonjour jibel
[06:11] <oSoMoN> hey jamesh
[06:11] <jibel> salut oSoMoN, ça va?
[06:11] <oSoMoN> jibel, bien et toi?
[06:11] <jibel> oSoMoN, très bien, avec les enfants en vacances tout est calme :)
[06:12] <oSoMoN> hé hé, veinard :)
[06:19] <jamesh> hi oSoMoN
[06:21] <didrocks> good morning
[07:11] <andyrock> good morning!
[07:13] <didrocks> hey andyrock
[07:35] <ricotz> hey desktopers
[07:36] <ricotz> oSoMoN, hey
[07:37] <willcooke> morning all
[07:37] <willcooke> didrocks, I installed a new machine from the daily ISO yesterday, and both "Ubuntu" and "Ubuntu Wayland" are running Wayland :)  Is that a bug?
[07:38] <willcooke> I'm seeing a strange keyboard-auto-repeat problem, which I'd like to test on X
[07:38] <duflu> willcooke, I suspect gdm is using Wayland/Xwayland in the bacdkground... ??
[07:39] <duflu> I noticed the Xorg login still had (X?)wayland running too
[07:42] <didrocks> willcooke: oh, that's indeed weird, nothing changed from this, but I only tested on a vm when selecting the wayland session starts wayland and ubuntu starts xorg
[07:42] <didrocks> willcooke: we saw that as well on iso testing
[07:42] <didrocks> XDG_SESSION_TYPE is showing up wayland for your ubuntu session thus?
[07:44] <willcooke> didrocks, confirmed, it is
[07:44] <didrocks> could be gdm doing things "smartish", but surprised by the mecanism as I don't reproduce it on vms
[07:44] <didrocks> willcooke: and you confirm XDG_SESSION_DESKTOP=ubuntu ?
[07:44] <didrocks> (it didn't select ubuntu-wayland for whatever reasons?)
[07:45] <willcooke> didrocks, ah ha  - it's ubuntu-wayland, but I specifically selected the not-wayland session
[07:45] <didrocks> could be a gdm bug :p
[07:45] <didrocks> like, "are you sure? you can do wayland" :p
[07:45] <didrocks> so, I guess if you logout
[07:45] <didrocks> and see current session selected, is it ubuntu or ubuntu wayland?
[07:46] <willcooke> it's ubuntu
[07:46] <didrocks> waow
[07:46] <didrocks> so, sounds like a gdm bug
[07:46] <didrocks> when logged in
[07:47] <didrocks> ~/.dmrc has ubuntu or ubuntu-wayland?
[07:47] <willcooke> that doesnt exist
[07:48] <duflu> Both scenarios will have "Xwayland" running. Seems the first instance is used by gdm itself.
[07:48] <didrocks> yeah, I think gdm tries wayland in its own session if possible (for the gdm user)
[07:48] <didrocks> but doesn't impact which session is then selected
[07:49] <didrocks> I wonder where gdm store which session is saved as selected by the user
[07:49] <flexiondotorg> Morning didrocks willcooke duflu ricotz
[07:49] <willcooke> morning flexiondotorg
[07:49] <didrocks> not .dmrc, I don't see anything in /var
[07:49] <didrocks> hey flexiondotorg
[07:49] <duflu> Morning flexiondotorg
[07:52] <jibel> is 2 fingers scrolling broken just for me?
[07:52] <didrocks> gdm doesn't log it
[07:52] <duflu> jibel, I found two-finger right click is broken. You can re-enable it in Gnome Tweak Tool
[07:53] <duflu> Maybe scrolling too?
[07:53] <didrocks> willcooke: I'm doing xenial -> yakkety (in progress) -> zesty -> artful upgrades today, I'll let you know if I have the same behavior
[07:53] <didrocks> maybe the others already in artful can confirm the bug
[07:53] <didrocks> (can't in a vm)
[07:55] <jibel> didrocks, i've X here on hw on 2 machines
[07:55] <didrocks> ok, so doesn't seem widespreadd, thanks jibel
[07:55] <didrocks> willcooke: ^
[07:55] <didrocks> would be interesting to know why your gdm is login you in the wayland session…
[07:57] <didrocks> AccountsService maybe?
[07:58] <didrocks> willcooke: mind pasting /var/lib/AccountsService/users/<your_user> ?
[07:59] <willcooke> didrocks, http://paste.ubuntu.com/25066927/
[08:00] <didrocks> XSession=ubuntu-wayland
[08:00] <didrocks> so, it confirms gdm selects ubuntu-wayland for you whatever you set
[08:00] <didrocks> for $reasons
[08:00] <willcooke> can I just edit that?
[08:00] <didrocks> well, you should do it when gdm is stopped and not logged in into any session
[08:00] <didrocks> but gdm will probably rewrite it anyway
[08:01] <willcooke> kk, lemme try
[08:01] <willcooke> oh
[08:01] <didrocks> maybe try to select the wayland session
[08:01] <didrocks> log in
[08:01] <didrocks> logout
[08:01] <didrocks> select the ubuntu session
[08:01] <didrocks> and log in ofc :p
[08:01] <willcooke> didrocks, tried all that
[08:01] <jibel> willcooke, is it with gdm you rebuilt to change the theme or from the archive?
[08:01] <willcooke> (related note - are we going to move vtys to 7 & 8?)
[08:02] <didrocks> willcooke: we didn't came to any decision (even didn't discuss it AFAIK)
[08:02] <willcooke> jibel, this is a plain gdm (with with some hacks for https://bugs.launchpad.net/ubuntu/+source/gdm3/+bug/1703415
[08:03] <didrocks> willcooke: I'm quite puzzled if it's not a stale cache and such, will be easier if I can reproduce after my today's 3 dist-upgrade :p
[08:03] <didrocks> (hoping I can get the same behavior…)
[08:03] <didrocks> Exec=gnome-session --session=ubuntu
[08:03] <didrocks> TryExec=gnome-shell
[08:03] <didrocks> this is the ubuntu session ^
[08:04] <didrocks> Exec=gnome-session --session=ubuntu
[08:04] <didrocks> TryExec=gnome-shell
[08:04] <didrocks> this is the wayland session ^
[08:04] <didrocks> so, they both have the same conditions (TryExec)
[08:04] <didrocks> can't be a fallback due to a failing tryexec
[08:04] <willcooke> didrocks, if I stop gdm3 I still have my session running on vty2 - how do I stop that too?  Just log out?
[08:04] <didrocks> willcooke: yes, log out
[08:05] <didrocks> you should stop account service daemon as well
[08:05] <didrocks> sudo systemctl stop accounts-daemon.service
[08:05] <Laney> bah
[08:06] <didrocks> hey Laney
[08:06] <willcooke> morning Laney
[08:06] <seb128> good morning desktopers
[08:06] <didrocks> re seb128
[08:06] <Laney> i'm on tethering
[08:06] <seb128> hey Laney willcooke flexiondotorg
[08:06] <Laney> normal interwebs is down
[08:06] <Laney> hi didrocks willcooke seb128
[08:07] <willcooke> morning seb128
[08:07] <willcooke> doh
[08:07] <willcooke> I used the ctrl-f2 -> r test
[08:07] <willcooke> and it worked
[08:08] <willcooke> should have used a different test there I think
[08:08] <willcooke> oh, actually, it was fine
[08:08] <willcooke> oki, didrocks I am now running Xorg
[08:08] <willcooke> :)
[08:08] <willcooke> on vty 7
[08:09] <didrocks> willcooke: hum, that's why I asked you for the env variable :p
[08:09] <didrocks> willcooke: still, something to keep a look at
[08:09] <didrocks> will tell you after my 3 dist-ugprades if I get that
[08:10] <willcooke> didrocks, cool, thanks!
[08:10] <willcooke> ok, next topic:
[08:10] <willcooke> https://bugs.launchpad.net/ubuntu/+source/gdm3/+bug/1703415
[08:11] <willcooke> j_bicha shared the debian way of doing it with me:
[08:11] <willcooke> https://paste.debian.net/975780/
[08:11] <willcooke> They are using postinst to copy the file, and mine was using rules.
[08:11] <willcooke> Line 25 of that diff
[08:12] <willcooke> If that dir. doesn't exist, then it wont work afaict.
[08:13] <willcooke> j_bicha suggested I talk to the debian dev who worked on it, but I haven't done yet.  Before I do, what do you think about that patch / concept - it means breaking BT audio in GDM for a11y users - but then unbreaking it for all users in the session itself
[08:18] <didrocks> I would follow and trust mbiebl's decision. Sounds like the breakage in GDM (which doesn't seem to be conditional for a11y users only, but everyone) to have it in the session sounds like a global win
[08:18] <didrocks> but that's just IMHO
[08:23] <seb128> didrocks, willcooke, loginctl show-session <id> also gives the type, that can be handy and easier to get that the env if you have a non working session but can get to a vt
[08:23] <seb128> just as a fyi
[08:34] <jbicha> good morning
[08:34] <seb128> good "morning" jbicha
[08:34] <jbicha> didrocks: I tried apt source ubuntu-meta then ran ./update but it reports "No changes found"
[08:36] <didrocks> jbicha: interesting, I'll have a look later if you want
[08:37] <jbicha> thanks
[08:38] <jbicha> actually, maybe there aren't any changes
[08:38] <didrocks> could be
[08:38] <jbicha> the nenw stuff did show up on http://people.canonical.com/~ubuntu-archive/component-mismatches.html
[08:38] <jbicha> *new
[08:59] <jbicha> and I guess we need the new lang pkgs promoted to main for them to end up on the Ubuntu iso
[09:18] <didrocks> jbicha:
[09:18] <didrocks> ibus-table-quick-classic
[09:18] <didrocks> ?
[09:20] <Laney> no WAIT, it's back
[09:21] <didrocks> ?
[09:23] <jibel> didrocks, do you have a vm with today's build and a french kbd layout?
[09:24] <jibel> didrocks, when I press altgr it does an 'enter'
[09:24] <didrocks> jibel: no, but I can maybe start it (on yakkety -> zesty dist-upgrade though)
[09:24] <didrocks> hum
[09:24] <didrocks> gnome-boxes is crashing
[09:24] <didrocks> probably due to the dist-upgrade
[09:26] <didrocks> jbicha: so, please, enlight me on ibus-mozc and ibus-table-quick-classic as we are still using fcitx
[09:26] <jbicha> didrocks: ibus-table-quick-classic is the final line in language-support's data/pkg_depends
[09:27] <didrocks> jbicha: maybe we didn't update it when moving to fcitx
[09:27] <didrocks> so it should be the other way around
[09:27] <didrocks> (until we have a decision on fcitx vs ibus)
[09:28] <jibel> didrocks, and up does print screen :/
[09:28] <jbicha> im:zh-hant:gnome-shell:ibus-table-quick-classic
[09:29] <jibel> what default session do we want for automatic login on a fresh installation?
[09:29] <jibel> ubuntu on X ?
[09:29] <jbicha> so it was only installed if a person had gnome-shell installed
[09:29] <didrocks> jbicha: yeah, because you didn't move to fcitx I guess
[09:29] <didrocks> jbicha: ubuntu on X, correct
[09:30] <didrocks> oupss jibel ^
[09:30] <didrocks> jbicha: so, I guess those should be either reverted or we need to have the fcitx/ibus discussion quickly
[09:38] <jbicha> I'm always uncomfortable making those input method decisions since I don't use them
[09:38] <didrocks> same
[09:38] <didrocks> that's why I would say, let's stick with what we have until it's really discuss
[09:38] <didrocks> discussed*
[09:38] <didrocks> and make a card to align language-selector with the seed once the decision is made?
[09:40] <willcooke> humm, getting a strange abort when trying to apt install stuff.
[09:40] <willcooke> Google says its because apt cant read from std in
[09:40] <willcooke> wayland issue?
[09:41] <willcooke> using "-y" works
[09:43] <didrocks> jbicha: ok, most (apart from ibus stuff) are done. I'll wait for a publisher cycle before looking at them again (I may have missed some, the list was long, some were binary only, other sources only, we are not protected against a typo ;))
[09:43] <jbicha> and I dropped those 2 from ubuntu.artful for now
[09:43] <didrocks> excellent!
[09:44] <jibel> didrocks, the waird input behaviour with VMs seems like another wayland bug
[09:44] <jibel> weird
[09:44] <didrocks> jbicha: mind making a trello card or adding to the fcitx one?
[09:44] <didrocks> jibel: "nice" :/ which VMs are you using?
[09:44] <didrocks> gnome-boxes worked fine for me until recently (but I hope I can blame the current crash on dist-upgrading in progress)
[09:45] <jibel> didrocks, qemu
[09:46] <didrocks> with kvm accell, I guess?
[09:46] <jibel> didrocks, yes
[09:46] <didrocks> regressed today?
[09:47] <didrocks> or just transient bug/you didn't spot it because of non wayland?
[09:47] <jbicha> note added to fcitx card
[09:47] <didrocks> thanks!
[09:48] <didrocks> jbicha: do you want me to ping you to refresh ubuntu-meta once all promotions/demotions are visible?
[09:49] <didrocks> (I can do it otherwise)
[09:52] <jibel> didrocks, not sure, I usually don't use wayland, it was for verifying something I saw on another machine
[09:54] <jibel> didrocks, re 'default session on autologin' the session is Ubuntu but with wayland not X
[09:54] <jibel> sounds like the same bug than willcooke discussed earlier
[09:54] <didrocks> interesting
[09:54] <didrocks> sounds like this happens only from the gdm switch
[09:55] <didrocks> but not your hw, correct?
[09:55] <didrocks> ah, autologin
[09:55] <didrocks> did willcooke mentioned autologin as well?
[09:55] <willcooke> no
[09:55] <willcooke> this problem was on a machine I installed yesterday from yesterday's daily iso
[09:56] <jibel> didrocks, no it was not autologin but a fresh installation
[09:56] <didrocks> but jibel didn't get it on real hw, I bet this is something related with gdm
[09:56] <didrocks> if the issue was:
[09:56] <didrocks> - default selected session is ubuntu-wayland
[09:57] <didrocks> - then, switch in gdm to "ubuntu" and it's ubuntu on X
[09:57] <didrocks> then yeah, the issue is that without any config, gdm selects ubuntu-wayland (as we saw on the iso)
[09:57] <didrocks> but here, willcooke was selecting the "ubuntu" session and ended up with ubuntu-wayland
[09:58] <willcooke> just spotted another, unrelated, issue.  On Wayland the icons down the left of the dash are a lot smaller than on X
[09:59] <willcooke> oh
[09:59] <willcooke> suddenly they're not
[09:59] <willcooke> meh, ignore
[09:59] <didrocks> this gives very low confidence… :p
[10:00] <jbicha> I think we don't need an ubuntu-meta update
[10:00] <willcooke> durr.  I r dumb.  They shrink when you have more things open.  Ignore me.
[10:01] <didrocks> few
[10:01] <didrocks> phew*
[10:02] <didrocks> so yeah, we would need to set DefaultSession in ubiquity anyway, as the gdm code seems to favor the wayland session if available by default
[10:02] <didrocks> on first log in
[10:03] <didrocks> (that doesn't explain willcooke's "I select ubuntu and it's ubuntu-wayland")
[10:03] <didrocks> but contrary to lightdm which uses alphabetical order, gdm seems to favor wayland sessions
[10:04] <didrocks> seems like that should have been tested in the gdm transition :p
[10:04] <didrocks> Laney: I can handle this later this week, apart if kenvandine wants to finish his transition's work with this, as you both prefer :)
[10:04] <Laney> what's the problem?
[10:04] <didrocks> basically, if you don't set a default session, gdm will favor the wayland ones
[10:04] <didrocks> (lightdm was going alphabetically)
[10:05] <didrocks> so you need to explicitely set your default session in ubiquity
[10:05] <Laney> isn't that how it is supposed to work?
[10:05] <jibel> didrocks, on autologin the session selected is ubuntu but of type wayland (output of loginctl) is it what you call ubuntu-wayland?
[10:05] <didrocks> yes, once we +1 on having wayland by default
[10:05] <didrocks> which isn't what we decided yet
[10:05] <didrocks> jibel: yes
[10:06] <jibel> ok
[10:06] <Laney> well that's how it is in GDM by default
[10:06] <Laney> so the change is actually a decision to go back to X
[10:06] <didrocks> Laney: right, hence the fact we should set "ubuntu" as default session
[10:06] <didrocks> as right now, we are unsure about X vs wayland
[10:07] <didrocks> and we spotted quite some issues with wayland anyway
[10:07] <didrocks> (which forced us to set X unconditionnally on live)
[10:08] <didrocks> jibel: you meant "ubuntu on wayland" is the selected label, correct?
[10:16] <didrocks> urgh, was dicsonnected
[10:16] <didrocks> jibel: do you have your testing box handy? like can you do something before first login?
[10:17] <didrocks> edit /etc/gdm3/custom.conf and uncomment WaylandEnable=false
[10:17] <didrocks> I want to know:
[10:17] <didrocks> 1. which default session is then selected
[10:17] <didrocks> 2. if ubuntu (using wayland) is available on your login screen
[10:17] <didrocks> 3. if you select ubuntu (using wayland), if wayland is really selected
[10:18] <didrocks> (I wonder if that switch doesn't disable autodetection as well, as it worked on the live)
[10:19] <Laney> oh right
[10:19] <Laney> so it prefers /usr/share/wayland-sessions/...
[10:19] <didrocks> yep
[10:20] <didrocks> we discovered that on the live
[10:20] <didrocks> ubiquity is crashing under wayland (and apport)
[10:20] <didrocks> I added some hooks in casper to force non wayland default session
[10:22] <Laney> didrocks: https://git.gnome.org/browse/gdm/tree/daemon/gdm-session.c#n363 <- append instead of prepend?
[10:23] <didrocks> Laney: yeah, if we can do that with a settings though and not distro-patching, let's try it
[10:24] <didrocks> I think the "ignore_wayland" may be the options I talked about above ^
[10:24] <Laney> doesn't that hide all the sessions though?
[10:24] <didrocks> Laney: doesn't seem from a quick test I made
[10:24] <didrocks> just want jibel to confirm
[10:25] <Laney> https://git.gnome.org/browse/gdm/tree/daemon/gdm-session.c#n361
[10:25] <didrocks> it just append them
[10:25] <didrocks> (it's just a runtime settings, not a #ifdef compilation option)
[10:26] <Laney> I know, that's line 361
[10:26] <Laney> 362 is the compile time thing
[10:26] <didrocks> however, as I'm dist(upgrading, I just tried to stop gdm, change the setting and start it
[10:26] <didrocks> so I can't confirm I'm sure at 100%
[10:26] <didrocks> let's wait for jean-baptiste's test for now
[10:27] <Laney> although this is set in only this case
[10:27] <Laney>         g_object_set (G_OBJECT (session), "ignore-wayland", !greeter_is_wayland, NULL);
[10:27] <didrocks> yeah, seems so
[10:27]  * didrocks really needs to take a lunch break, bbl
[10:27] <Laney> hf!
[10:28] <didrocks> at worst, indeed, quick distro patch
[10:30] <duflu> willcooke, I think I hit your Bluetooth bug. That's new with gdm right?
[10:31] <duflu> Nevermind
[10:31]  * duflu prepares dinner
[10:31] <willcooke> duflu, yeah
[10:32] <duflu> Alright. That's a good boundary to call it a day
[10:32] <duflu> or night
[11:21] <jibel> didrocks, trying now
[11:55] <jibel> didrocks, I've name ubuntu, type x11. Trying on a fresh install now
[11:55] <didrocks> jibel: do you have access to the ubuntu on wayland session?
[11:57] <jibel> didrocks, no I trashed it. I'll redo it after current install if you want
[11:58] <didrocks> yes please :)
[11:58] <didrocks> that will confirm we can turn that switch on if the session is available
[12:03] <jibel> yay another GS crash ...
[12:11] <jibel> didrocks, with WaylandEnabled set to false I get an x11 session
[12:11] <jibel> and autologin enabled
[12:11] <jibel> redoing an install without this change
[12:12] <didrocks> jibel: but do you see the wayland session available in gdm?
[12:12] <jibel> didrocks, with autologin I didn't see gdm :) let me logout
[12:14] <jibel> didrocks, weird, in gdm I don't see the cog to select the session
[12:15] <didrocks> I guess that means only one session is available
[12:15] <jibel> I'll disable autologin and reboot just to check
[12:17] <jibel> didrocks, same after a reboot, there is no option to select a session
[12:18] <jibel> and autologin disabled
[12:18] <jibel> didrocks, do you still need to reproduce previous scenario (default install and ubuntu-wayland) ?
[12:19] <didrocks> jibel: no, that was what I wanted to know, thanks for looking!
[12:19] <didrocks> we'll need to patch gdm to prefer non wayland to wayland thus
[12:19] <didrocks> will do that tomorrow
[12:20] <jibel> didrocks, do you have a bug # already otherwise I'll file one
[12:20] <didrocks> jibel: I don't, please do it and assign it to me
[12:23] <willcooke> bigon, hi!  jbicha pointed me your way to talk about the gdm/pulseaudio issues with it locking A2DP.  I got this (reverse) diff: https://paste.debian.net/975780/ and I can see that you're copying the file in postinst.  Line 25 in the diff copies the file, but I can't see where that directory is created.  I have a similar diff for Ubuntu which uses rules to copy the file.
[12:24] <willcooke> so a couple of questions:  any idea when/if that will land in debian & is that a bug regarding the creation of the /var/lib/gdm3/.config/pulse dir?
[12:26] <jibel> didrocks, bug 1703601
[12:26] <didrocks> jibel: reset it to gdm please
[12:29] <jibel> done
[12:29] <didrocks> thx
[13:56] <andyrock> mpt hey hey, any progress on the design?
[14:00] <jbicha> seb128: we could cherry-pick this if you wanted it now: https://git.gnome.org/browse/evolution/commit/?id=736f774d
[14:23] <seb128> jbicha, that would be good
[14:30] <seb128> jbicha, thanks
[14:33] <jbicha> seb128: done for evolution and evolution-data-server, do you need to wait for it to build to change the LP setting?
[14:37] <willcooke> jbicha, changed the MP for gdm pulse audio to be the same as the Debian one: https://code.launchpad.net/~willcooke/gdm/ubuntu/+merge/327165
[14:38] <willcooke> Interested to know what you think about merging it.  Particularly re: a11y users
[14:38] <Laney> that mkdir in the postinst is wrong, it should refer to the real filesystem
[14:38] <willcooke> gah
[14:38] <willcooke> copy & paste error
[14:39] <jbicha> willcooke: oops, you should contact mbiebl not big_on; he's in #debian-gnome on irc.oftc.net
[14:40] <Laney> or #ubuntu-devel
[14:41] <willcooke> Laney, fixed the mkdir
[14:41] <Laney> nice one
[14:46] <seb128> jbicha, yes, I need for the template to be in the import queue to accept it and set the name/target
[14:53] <kenvandine> Laney, https://asciinema.org/a/kt3NIQ0IJigsvM3CY6VCE9N5s
[14:54] <Laney> hey kenvandine
[14:54]  * Laney gets popcorn
[14:54] <ricotz> may I draw some attention to https://bugs.launchpad.net/ubuntu/+source/xmlsec1/+bug/1703556 :)
[14:55] <kenvandine> Laney, i still need to make sure it does prompt when we want to :)
[14:55]  * kenvandine prepares test case for that :)
[14:55] <Laney> nice
[14:55] <kenvandine> i was also able to drop the ubuntu-session depends now too
[14:55] <kenvandine> which is nice
[14:55] <Laney> ya
[14:55] <Laney> and the changes there too I guess
[14:55] <kenvandine> yes
[14:56] <kenvandine> working on that still
[14:56] <Laney> grrrrrreat
[14:57] <bigon> willcooke: well that might break accessibility in GDM isn't it?
[14:57] <bigon> if people have bluetooth headsets
[14:57] <bigon> (or speaker)
[14:57] <bigon> that's why it was not uploaded
[15:02] <bigon> There is a bug opened against pa IIRC
[15:03] <willcooke> bigon, yeah, no movement on that pa bug for a while.  I'm trying to decide if it'd be "ok" for 17.10 and then we can help fix it in time for 18.04.
[15:30] <seb128> ricotz, hey, https://launchpad.net/ubuntu/+source/xmlsec1/1.2.24-3
[15:32] <ricotz> seb128, thanks!
[15:32] <seb128> ricotz, yw!
[15:32] <seb128> willcooke, meeting?
[15:33] <seb128> or did you ask me to handle it afterall and I missed the memo?
[15:33] <seb128> it's the right day/time?
[15:33] <willcooke> yeah, would you mind seb128
[15:34] <willcooke> I've got this other call
[15:34] <seb128> no problem
[15:34] <jbicha> o/
[15:34] <willcooke> thanks
[15:34] <didrocks> o/
[15:34] <seb128> #startmeeting Desktop team 2017-07-11
[15:34] <meetingology> Meeting started Tue Jul 11 15:34:51 2017 UTC.  The chair is seb128. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
[15:34] <meetingology> Available commands: action commands idea info link nick
[15:35] <oSoMoN> 🐵/
[15:35] <heber> o/
[15:35] <seb128> Roll call: andyrock, dgadomski, didrocks, duflu (out), jbicha, jamesh (out), jibel, heber, kenvandine, laney, oSoMoN, seb128, tkamppeter, trevinho, robert_ancell (out)
[15:35] <andyrock> o/
[15:35] <seb128> oSoMoN, lol :-)
[15:35] <kenvandine> o/
[15:35] <Trevinho> :)
[15:35] <seb128> ok, seems like most people are around, let's get started
[15:35] <seb128> hope everybody is doing fine!
[15:36] <seb128> #topic andyrock
[15:36] <seb128> andyrock, hey
[15:36] <andyrock> #1 Review SRU
[15:36] <andyrock> #2 Test SRU
[15:36] <andyrock> #3 Discovered, reported and fixed [BUG] If the background file does not exists, the lockscreen draws a transparent background
[15:36] <andyrock> #4 Preparing the demo for livepatch
[15:37] <andyrock> #5 europython
[15:37] <andyrock> # eow
[15:37] <seb128> SRU of? BUG? :p
[15:37] <andyrock> Blocked on other livepatch works waiting for design
[15:37] <andyrock> unity7
[15:37] <andyrock> sorry
[15:37] <seb128> haha
[15:37] <seb128> no worry
[15:38] <andyrock> ;D
[15:38] <seb128> thanks andyrock, hope design (is that mpt?) can unblock you this week
[15:38] <andyrock> yep
[15:38] <seb128> k
[15:38] <seb128> thanks
[15:38] <seb128> next is Ken since he might have to go away from the computer at some point
[15:38] <seb128> #topic kenvandine
[15:38] <seb128> kenvandine, hey
[15:39] <kenvandine> * First pass lightdm->gdm3 transition, found some issues which I'm testing fixes for now.
[15:39] <kenvandine> * Replaced lightdm with gdm3 in the desktop seed
[15:39] <kenvandine> * Worked on snapping more gnome apps
[15:39] <kenvandine> * Preparing demo using gnome-3-24 platform snap
[15:39]  * kenvandine drops the mic
[15:39] <seb128> thanks kenvandine
[15:39] <seb128> #topic dgadomski
[15:39] <seb128> dgadomski, hey
[15:39] <dgadomski> hey
[15:39] <dgadomski> * working on bug #1700827
[15:39] <dgadomski> eof
[15:40] <seb128> thanks dgadomski
[15:40] <seb128> #topic didrocks
[15:40] <seb128> didrocks, hey
[15:40] <didrocks> * Clean a big part of the seeds. Did some analyses with seb on some of them. Lot of demotions (more than 70 packages). indicator-, touch-*, unity-*.
[15:40] <didrocks> * Still some to ongoing investigation to remove other parts of the stack (like zeitgeist, ubuntu-app-launcher, qt…). The biggest one pulling packages now is fcitx. We need to discuss about that one. We still have some, like dee, libunity, indicator-application (due to its libs), indicator-messages (due to its libs).
[15:40] <didrocks> * Split into separate packages for optional dependencies, like deja-dup and unity-control-center panel to be able to keep the functionality (pulled now in only in the unity session) the unity, but without having the dependencies in the main iso in our gnome-shell session.
[15:40] <didrocks> * Did some ubiquity debugging (was wrongly accused! :p) which in fact revealed that the gdm transition was defaulting to wayland, and ubiquity is crashing under it. We wanted to force X for now anyway, so patched casper for this.
[15:40] <didrocks> * language discussion, change the seeds for -frami german dictionary variant to align with language-selector. Reviewed MP.
[15:40] <didrocks> * Main promotion and test lightm -> gdm transitions + debug some postinst issue.
[15:40] <didrocks> * Look at existing gtk golang bindings. Found one which is still actively developed, quite feature complete on gtk, glib, gio, updated up to gtk 3.24
[15:40] <didrocks> * Discussed keeping unity7 session somewhat functional and impacts on code, migrations…
[15:40] <didrocks> * discussion on snapcraft forum on interface UI
[15:40] <didrocks> * AA: newed some packages
[15:40] <didrocks> * . *
[15:42] <seb128> thanks didrocks, busy week :-)
[15:42] <didrocks> yep ;)
[15:42] <seb128> #topic duflu
[15:42] <seb128> * Video acceleration:
[15:42] <seb128>   - Still working breadth-first on the VA-API failures in Wayland sessions: https://bugs.launchpad.net/bugs/1698287
[15:42] <seb128>     . Formally proposed upstream fix for totem/gstreamer-vaapi - https://bugzilla.gnome.org/show_bug.cgi?id=783169
[15:42] <seb128>     . Investigated gst-play-1.0 failure on Gnome Shell, but not Weston.
[15:42] <seb128>       . Narrowed down to Mutter missing a feature. Stopped there for now.
[15:42] <seb128>     . Investigated graphics corruption in totem (gstreamer-clutter plugin)
[15:42] <seb128>       . Narrowed down to clutter, confirming the same problem with gst-launch-1.0 chains.
[15:42] <seb128>       . Will fix this one soon as a priority.
[15:42] <seb128>   - Also lost a day and part of the weekend investigating graphics corruption in mpv+vaapi: https://github.com/mpv-player/mpv/issues/4555
[15:42] <seb128>   - Fun fact: Modern Intel Atom chips (Cherry Trail, Apollo Lake) and cheap notebook chips (Braswell) can play 4K H.265 without breaking a sweat. Even on a 2-watt CPU. Unfortunately they usually come with low quality screens and never HDMI 2.0...
[15:43] <seb128> * PulseAudio:
[15:43] <seb128>   - Set up launchpad git for pulseaudio maintenance properly: https://git.launchpad.net/~ubuntu-audio-dev/pulseaudio
[15:43] <seb128>   - Work in progress on an update for artful: https://git.launchpad.net/~ubuntu-audio-dev/pulseaudio/diff/debian/changelog?h=ubuntu
[15:43] <seb128> * Daily bug maintenance across gnome-shell, bluez, pulseaudio and mir.
[15:43] <seb128>   - Noticed hundreds of pulseaudio bugs are finally expiring this week, which is nice. Should total about 1000 closed soon.
[15:43] <seb128> #topic jbicha
[15:43] <seb128> jbicha, hey
[15:43] <jbicha> • Sponsored GunnarHJ's ubuntu-meta l10n update
[15:43] <jbicha> • Fixed a few broken po files upstream discovered by first Artful language pack
[15:44] <jbicha> • Synced onboard from Debian (completing a year-long project to sync the packaging)
[15:44] <jbicha> • Cherry-picked fix for Qt file chooser crashing in Wayland, zesty SRU in progress LP:# 1702741
[15:44] <jbicha> LP: #1702741
[15:44] <jbicha> • Helped verify a few old SRUs (wine1.6, unity-control-center)
[15:44] <jbicha> • Packaged libdazzle (pending sponsorship into Debian new queue), will be needed for gnome-builder 3.26 https://bugs.debian.org/867852
[15:44] <jbicha> • Jumpstarted libunique3 removal process in Debian
[15:44] <jbicha> https://bugs.debian.org/cgi-bin/pkgreport.cgi?users=pkg-gnome-maintainers@lists.alioth.debian.org;tag=libunique3-removal
[15:45] <jbicha> eof
[15:45] <seb128> thanks jbicha
[15:45] <seb128> #topic jamesh
[15:46] <seb128> gnome-software: prepared upstream versions of the patches I'd worked
[15:46] <seb128> on.  Multi-app snaps patch has been accepted.
[15:46] <seb128> snapd: continued conversation about use of polkit in snapd.  Last word
[15:46] <seb128> from Gustavo was to create a pull request so there would be something
[15:46] <seb128> concrete to discuss.  I'm working on a minimal one that basically just
[15:46] <seb128> covers what snapd-login-service does.
[15:46] <seb128> network-manager: still need to do testing of captive portal UI.
[15:46] <seb128> #topic jibel / heber
[15:46] <seb128> jibel, heber, hey
[15:46] <heber> Hey guys! QA updates:
[15:46] <heber> * Verification of upgrades to artful
[15:46] <heber> * Tested unity7 on artful
[15:46] <heber> * Reported several issues with desktop images (no images, the live session on wayland, installer)
[15:46] <heber> * Exploratory testing of unity7 on artful
[15:46] <heber> * Exploratory testing of the default session with wayland
[15:46] <heber> * Manual review of upgrades from xenial to zesty
[15:46] <heber> * Manual review of upgrades to artful
[15:46] <heber> * Update auto-upgrade-tests to persist the resulting image in order to be able to run post upgrade tests
[15:46] <heber> * Update CI to run post upgrade tests after upgrade jobs
[15:46] <heber> * Address review feedback on MPs for running post-install tests.
[15:46] <heber> EOF
[15:47] <seb128> thanks heber
[15:47] <seb128> #topic Laney
[15:47] <Laney> • Short week, was off last week
[15:47] <Laney> • Catching up on emails and stuff
[15:47] <Laney> • Small autopkgtest-web debug/fix to the trigger validation regex so that some packages can be retried
[15:47] <Laney> • Upload PackageKit and appstream-glib to artful with some extra patches, prep for GS 3.25
[15:47] <seb128> Laney, hey, hope you had a good week off!
[15:47] <Laney> ∘ appstream-glib fix to properly get apps marked as unremovable
[15:47] <Laney> ∘ patch (upstream) to gnome-software to not list installed packages when doing PackageKit WhatProvides queries, works around some false positives
[15:47] <Laney> ∘ small patch in ubuntu branch to turn off updates via PackageKit since we do those using update-manager atm
[15:47] <Laney> • Some advice/testing/review on GDM migration stuff & comments on GDM/wayland/X11 session selection
[15:47] <Laney> 😏
[15:47] <Laney> oh & hi!
[15:47] <seb128> :-)
[15:48] <Laney> back to the 0s and 1s
[15:48] <seb128> thanks Laney, welcome back in the binary world
[15:48] <seb128> #topic oSoMoN
[15:48] <seb128> oSoMoN, hey
[15:48] <oSoMoN> hey there
[15:48] <Laney> #ubuntu-desktop is the best bits
[15:48] <oSoMoN> • a chromium crasher affecting 59.0.3071.109 on trusty/xenial/yakkety was widely reported last week (bug #1702407), worked right away to identify a fix, had a number of affected users confirm the fix through a PPA, and finally pushed the update out today (thanks Chris!)
[15:48] <oSoMoN> • chromium beta: updated HTML5 conformance autopkgtest to not fail early, instead display discrepancies and fail only at the end of the test
[15:48] <oSoMoN> • experimenting with is_official_build=true && is_component_build=false for chromium, with the goal of producing builds that are more similar to chrome (branding not affected, there are different flags for that)
[15:48] <oSoMoN> • got little but positive feedback about the libreoffice 5.3.4 snap in the beta channel, published to the stable channel today
[15:49] <oSoMoN> • working on temporarily disabling jvm-related unit tests in libreoffice x86 build to allow the migration to move forward, given that the issue is in the kernel and that the current version is also affected (bug #1699772)
[15:49] <oSoMoN> • will resume work on packaging libreoffice 5.4.0 RC1 now that cppunit 1.14 has been synced from debian
[15:49] <oSoMoN> • off on holidays for two full weeks starting this Thursday evening, anything urgent I should look into speak now or forever hold your peace
[15:49] <oSoMoN> ⛱
[15:49] <ricotz> oSoMoN, https://launchpad.net/~ricotz/+archive/ubuntu/red/+packages?field.name_filter=&field.status_filter=published&field.series_filter=artful
[15:49] <oSoMoN> ricotz, yup, seen that, thanks!
[15:50] <seb128> thanks oSoMoN
[15:51] <seb128> #topic seb128
[15:51] <seb128> • reviews (u-c-c) and sponsoring (synced xmlsect1)
[15:51] <seb128> • tested daily iso and helped debugged a bit issues with wayland session being default
[15:51] <seb128> • talked with Didier about the iso cleanups he was doing
[15:51] <seb128> • discussed language packages tweaks and impact on the iso
[15:51] <seb128> • changed g-c-c to build the -faces binary instead of u-c-c so that one can be moved to universe
[15:51] <seb128> • talked about the u7 session status in artful, started a wikipage listing known issues
[15:51] <seb128> • debugged gnome-shell translations missing from langpack in some locales (due to formatting issues in the corresponding .po leading to launchpad not being to import them)
[15:51] <seb128> • verified unity-control-center nopassword login yakkety SRU

[15:51] <Laney> yakkety (H)
[15:52] <oSoMoN> yakkety is EOL at the end of the month
[15:52] <ricotz> oSoMoN, going to push the branch when amd64 succeeds
[15:52] <seb128> SRU team was insisting that the yakkety SRU got verified to unblock the xenial one
[15:52] <seb128> was easier to verify it that argue with them more on IRC
[15:52] <seb128> let's move on to the next one ;-)
[15:53] <seb128> #topic tkamppeter
[15:53] <seb128> - cups-filters: Modified cups-browsed to improve the cooperation with CUPS' mechanism of auto-creating temporary print queue for DNS-SD-discovered printers. Naming of cups-browsed-created queues is now the same as CUPS' naming for temporary queues but can be configured to the formerly used schemes for clustering. Temporary queues are recognized and replaced by cups-browsed queues.
[15:53] <seb128> - GNOME printer setup tool: Reported 9 bugs on shortcomings of GNOME's tool against system-config-printer.
[15:53] <seb128> - Google Summer of Code 2017: Student project coordination and mentoring. Especially helped the student working on PCLm printing support.
[15:53] <seb128> - Bugs.
[15:53] <seb128> good work Till doing that review of the g-c-c print panel!
[15:53] <seb128> #topic Trevinho
[15:53] <seb128> Trevinho, hey
[15:53] <Trevinho> · Preparing SRU for unity and compiz X with improved lowgfx
[15:53] <Trevinho>  - Blocked by some issues I've noticed, and fixing them
[15:53] <Trevinho> · Some work to move unity settings daemon to use different schemas
[15:53] <Trevinho> . Doing multiple tests in X
[15:53] <Trevinho>  /EOF
[15:54] <seb128> thanks Trevinho
[15:54] <seb128> #topic robert_ancell
[15:54] <seb128> To: Will Cooke <will.cooke@canonical.com>
[15:54] <seb128> - Wrote Snappy sprint blog post
[15:54] <seb128> - Worked on snapd license patch
[15:54] <seb128> - Ongoing discussions on forum.snapcraft.io about various desktop related snap issues
[15:54] <seb128> - SRU fix to stop installed snaps showing up twice (bug 1700994)
[15:54] <seb128> - Prod snapd-glib SRU to try and speed up progress and confirm this fixes wrongly translated Polkit prompts
[15:54] <seb128> - Work on GNOME Software 3.20.5 SRU
[15:55] <seb128> - Release snapd-glib 1.14 and 1.15
[15:55] <seb128> - Update GNOME Software to support new title field and system confinement status
[15:55] <seb128> - Investigating gnome-software errors.ubuntu.com crashes
[15:55] <seb128> ups, sorry for the first line, copied one too much
[15:55] <seb128> #topic aob
[15:56] <seb128> is there anything else you want to discuss this week?
[15:57] <seb128> seems not?
[15:57] <seb128> so let's wrap, thanks everyone!
[15:57] <seb128> #endmeeting
[15:57] <meetingology> Meeting ended Tue Jul 11 15:57:24 2017 UTC.
[15:57] <meetingology> Minutes:        http://ubottu.com/meetingology/logs/ubuntu-desktop/2017/ubuntu-desktop.2017-07-11-15.34.moin.txt
[15:57] <oSoMoN> cheers
[15:57] <didrocks> thx
[18:43] <willcooke> anyone else getting timeouts uploading to LP?
[18:46] <willcooke> and of course, as soon as I say that, after having had 6 time outs, it works
[18:47] <cyphermox> willcooke: I did have some timeouts, but not uploading
[18:47] <cyphermox> loading a page that should have been very easy to load
[18:47] <willcooke> kk, thx cyphermox
[19:09] <willcooke> night all
[19:47] <xclaesse> are notifications in Unity clickable?
[19:48] <xclaesse> can the app know when user clicked them?
[20:53] <fossfreedom> jbicha - just reinstalled with today's Ubuntu Budgie daily.  Strangely the GNOME session is the default and I see GNOME Wayland as well.  Any ideas why GNOME Shell is now installed in UB?
[21:13] <jbicha> fossfreedom: please file a bug for that…against mutter I guess since mutter recommends gnome-session
[21:20] <fossfreedom> ok - cheers
[21:27] <fossfreedom> bug 1703685 filed
[21:34] <jbicha> fossfreedom: it's actually an upstream bug (Debian) https://launchpad.net/ubuntu/+source/gnome-session/3.24.1-0ubuntu7
[21:39] <fossfreedom> hmm -  that build has failed at the beginning of july.