[00:00] <robert_ancell> jbicha, oh, and just prompts you after you get into your desktop "updates are available" and they can then install automatically
[00:00] <robert_ancell> install immediately I should have said
[00:00] <jbicha> I'd like to propose that we add one more checkbox to actually install those updates since I think that's what a lot of people expect
[00:00] <robert_ancell> Yeah, I expected that...
[00:00] <jbicha> or just change the wording and do it
[00:01] <robert_ancell> In this particular case, the bug is probably not a problem (it's only a crash on shutdown)
[00:01] <jbicha> the counter-argument is that apparently it can make the install take a rather long time once a release has been out for several months
[00:01] <robert_ancell> I don't know of any other package that gets updated on install, so it's probably a new feature that would need adding.
[00:02] <robert_ancell> jbicha, "snap it" is probably the answer :)
[00:02] <jbicha> new feature: which means it wouldn't happen until 18.04.1 anyway! :|
[00:03] <robert_ancell> indeed
[00:04] <robert_ancell> I think the .1 is release is where you'll get the majority of users, so it's not a big issue to wait.
[00:05] <jbicha> robert_ancell: btw, we got approval for the chose fix, not sure how to handle that smoothly with translations
[00:05] <jbicha> https://github.com/CanonicalLtd/desktop-design/issues/102
[00:05] <ubot5-ng`> CanonicalLtd bug 102 in desktop-design "first-login: "location you choose" or "location you chose"" (comments: 2) [Closed] - Assigned to matthewpaulthomas (Matthew Thomas)
[00:05] <jbicha> also https://github.com/CanonicalLtd/desktop-design/issues/101
[00:05] <ubot5-ng`> CanonicalLtd bug 101 in desktop-design "First login: What's new: wording suggestions" (comments: 3) [Open] - Assigned to matthewpaulthomas (Matthew Thomas)
[00:06] <sarnold> ubot pls
[06:41] <seb128> good morning desktopers
[06:51] <duflu> Morning seb128
[06:51] <seb128> hey duflu, how are you?
[06:51] <duflu> seb128, going OK... you?
[06:51] <duflu> In Berlin?
[06:52] <seb128> I'm good thanks, yes in Berlin
[06:52] <duflu> ICE FTW
[06:53] <seb128> indeed!
[06:54] <seb128> travel was nice
[06:59] <seb128> duflu, https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/1768012 ... do you think it could be a kernel issue? it's Sergio who has the issue and the touch screen seem to work fine mostly, the bug seems rather in g-s touch events handling
[07:03] <duflu> seb128, I will try some other touchscreens this week. But it is slightly suspicious to be running a custom "surface" kernel
[07:03] <duflu> Guess the manufactureer
[07:03] <duflu> -e
[07:04] <willcooke> seb128, duflu re: https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1751252
[07:04] <willcooke> So do we thinkthat's a glib bug?
[07:05] <duflu> willcooke, No idea. The two crashes are seconds apart
[07:05] <duflu> But always in order
[07:06] <duflu> willcooke, actually in person you can see in advance it is going to crash. The installer splash is plain black
[07:07] <duflu> It's not that we never tested bionic on the affected machines, only that the ubiquity regression occurred more recently
[07:07] <duflu> Because it used to work
[07:07] <duflu> Or glib regression
[07:08] <seb128> duflu, can you reproduce on that machine? can you see if ubiquity hits an XError for you like on https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1767720 ?
[07:10] <duflu> seb128, I was going to try and fix the zoom issue first because I at least know that code (and the language). But yes I will check the ubiquity crash again now
[07:10] <seb128> duflu, thanks
[07:28] <duflu> seb128, yes. Bug 1751252 is also BadAccess with the same X error. It's highly unlikely they are different bugs
[07:29] <seb128> duflu, good, do you think you are able to the bt with gtk/libx11 dbgsym installed?
[07:29] <duflu> seb128, not sure. For starters X errors are always delayed. Will have to force it to synchronous mode and retry.
[07:30] <seb128> duflu, right, see my comment on that bug
[07:30] <seb128> duflu, but basically need to exportGDK_SYNCHRONIZE=1
[07:31] <seb128> then gdb --args /usr/bin/python3 /usr/lib/ubiquity/bin/ubiquity gtk_ui
[07:31] <duflu> Interesting this is always on specific machines...
[07:36] <duflu> seb128, also worth noting a bunch of manual bug reports to ubquitiy the past few days report "installer crashed" but we seem to not attach sufficient logs
[07:37] <seb128> right
[07:37] <seb128> it's always tricky
[07:38] <duflu> Just the journal would help
[07:38] <seb128> i/o errors on usb stick are also common
[07:39] <seb128> like some of the snaps reports on e.u.c which are having a few hundred reports a week are squashfs read errors, which are likely corrupted image/read error from the device being used
[07:39] <seb128> sometimes segfaults are also a side result of a corrupted image/faulty key
[07:43] <duflu> seb128, I thought of that. The same key works on other machines even after failing on the unlucky one
[07:44] <seb128> well if some people can reproduce on some configs there is probably something in the machine configuration that confuses ubiquity, e.g a real bug
[07:44] <seb128> we just don't have enough info atm to be able to figure out what exactly that is
[07:44] <seb128> maybe a sync/debug bt is going to tell us more
[07:54] <seb128> duflu, hum, "log_domain=0x7f5a7365304e <_gdk_wayland_display_utf8_to_string_target+270>", why are we getting wayland mentioned there?
[07:58] <duflu> Not sure. More comments in a minute
[08:09] <duflu> seb128, OK, I have documented a workaround in the bug. Install complete
[08:10] <duflu> No let's see if I can get a synchronous python stack too
[08:10] <willcooke_> ahhhh, nicely deduced duflu
[08:13] <seb128> duflu, good finding!
[08:14] <seb128> there is webkit/cairo calls in that bt, I wonder if something in the slideshow is trying to display something too big for the video memory or something
[08:15] <duflu> Or too small
[08:16] <duflu> Still, if anyone knows how to debug ubiquity I think we now have a test case
[08:16] <duflu> (set Scale to 200% before installing)
[08:17] <seb128> well, it sounds like the issue there is webkitgtk
[08:17] <seb128> or something the installer is doing with it
[08:17] <seb128> duflu, well done figuring that out in any case
[08:23] <seb128> duflu, do you know if there is a way to set the scaling factor from an environment variable or such?
[08:24] <duflu> seb128, not sure. Probably better to rely on the config
[08:24] <seb128> k
[08:24] <duflu> seb128, GDK_SCALE={1,2}
[08:46] <Laney> helo
[08:48] <seb128> hey Laney, how is sunny u.k today?
[08:48] <Laney> hey seb128
[08:49] <Laney> it's quite nice today
[08:49] <Laney> except we're in the main office now
[08:49] <Laney> how's the land of currywurst?
[09:06] <seb128> nice and sunny
[09:06] <seb128> but we are also lock inside
[09:07] <seb128> at least not in a dungeon, pretty nice view on the outside
[09:07] <seb128> bah, bug #1758120
[09:07] <seb128> having a /usr/share/pixmaps file in your deb isn't going to fly
[10:03] <ahayzen> Hey, i had an issue where i login with user A, then logout and try to login with user B - it results in a purple display with the mouse (switching tty back to 1 then doing the login again works). I wanted to check if this story is covered by bug 1765261 and bug 1766137 or wht
[10:03] <ahayzen> bah
[10:03] <ahayzen> *or whether i should report another bug
[10:04] <ahayzen> as there were a few different user stories in those bugs
[10:04] <seb128> Laney, do you know how I can set a DNS server from a rescue.target env?
[10:06] <seb128> ahayzen, would be good to report a new bug upstream, they can dup if needed, but even if it's a dup it might help raise awareness
[10:06] <Laney> seb128: resolv.conf doesn't work?
[10:06] <seb128> Laney, can't edit in /etc
[10:07] <Laney> ????
[10:07] <ahayzen> seb128, ok thanks, is https://gitlab.gnome.org/GNOME/gnome-shell/issues "upstream" is this case? or is there another place for gdm issues ?
[10:07] <Laney> that's a symlink to /run/something if you're using resolved
[10:07] <seb128> Laney, saving the file fails with a "no such file or directory" for some reason?
[10:08] <seb128> dnsmasq  --server fails as well due to missing /etc/resolv.conf
[10:08] <Laney> is resolved actually running?
[10:08] <Laney> systemctl start network-online.target usually gets me online
[10:09] <seb128> it was not
[10:09] <seb128> that did it
[10:09] <seb128> thx
[10:09] <duflu> seb128, I've verified the workaround works on two machines and attached a new stack with debug info. But I think I am about to give up for the day
[10:10] <seb128> duflu, thx, it's good enough, we have a way to reproduce now so others can pick up
[13:17] <lech1111[m]> Hi, anyone have idea how network-manager (with vpnc and network-manager-vpnc, and network-manager-vpnc-gnome) can connect to vpn using sms token?
[13:37] <jbicha> seb128: kenvandine: meeting cancelled today?
[13:37] <kenvandine> jbicha, oh sorry, yeah
[13:38] <kenvandine> jbicha, seems half the team has national holidays today :)
[13:48] <Laney> (also they've been slacking)
[13:49]  * seb128 crack the whip toward Laney
[14:08] <kenvandine> seb128, that reminded me of rick's whip cracking sound he used to make :)
[14:08] <seb128> lol
[14:08] <kenvandine> oh the memories :)