brlinCan anyone help review this one: https://github.com/ubuntu/snapcraft-desktop-helpers/pull/18402:28
gitbotubuntu issue (Pull request) 184 in snapcraft-desktop-helpers "bindtextdomain: Support locale files under share/locale" [Open]02:28
didrocksgood morning05:54
dufluMorning didrocks05:59
didrockshey duflu06:00
didrocksduflu: hum, I'm running mutter 3.32.0+git20190410-1ubuntu1 which has supposively your timestamp fix, but when I click on the dock, I still have my apps not focused when they start06:03
seb128good morning desktopers, happy friday!06:03
seb128hey duflu didrocks06:04
didrockshey seb12806:04
dufluHi seb12806:04
dufluHappy Friday06:04
didrockssome people on the french forum says that it fixes it for them, so there is maybe many causes ?06:04
* didrocks just tried if intellihide mode could have an impact, but not the case06:05
seb128GNOME has been having issue giving the focus in some cases since we switched to it (or at least ti's also an issue in bionic)06:05
seb128unsure if your issue is different06:05
didrockswell, it's quite a 100% hit here06:05
didrockswhich was never the case before 2.3106:05
seb128on any app?06:05
dufludidrocks, the bug seems to be fixed here. Remember it was Xorg sessions only06:05
didrocksduflu: I'm on Xorg06:05
didrocksah wait06:06
didrocksmaybe it was slightely updated this morning by unattended-upgrade06:06
dufluUpdate not fully installed?06:06
didrocks(I wonder why unatteded-upgrade is installed, it's not a security fix…)06:07
didrocksand so apt policy showed me I was on the latest06:07
seb128we apply dev updates with unattended-upgrade now?06:07
didrockslooks like it from my history.log :/06:07
didrocksas I wasn't expecting that, I thought apt policy -> latest meant "updated yesterday evening and running latest"06:08
didrocksso, let me alt+F206:08
dufluOr log out...06:08
didrocksduflu: all working now!06:08
dufluThough not if you want to keep your windows06:08
didrocksat least, confirming the fix :)06:08
didrocksbut really, unattended-upgrade, why?06:08
didrocksyep, unsure if this change was desired06:09
didrocksI definitively didn't tweak any config06:09
didrocksand especially on the dev cycle, I want to have control of my ugprade06:09
dufluI knew I had the fix because my test wallpaper became sharper today06:09
didrocksseb128: duflu: do you have the same unattended-upgrade experience?06:10
dufludidrocks, not sure. My dev system gets turned off every night06:10
didrocksduflu: well, same for me, it's on reboot that it updates06:10
didrocksduflu: you can look at your apt logs and see if you have unattended-upgrade lines06:10
* didrocks disables the service for now06:11
dufludidrocks, yes. It seems I am now getting unattended-upgrade06:11
didrockslet me file a bug06:12
seb128didrocks, yeah, I killed it a few time recently because it was blocking me to using apt in the morning and taking too long and I couldn't be bothered waiting to do $work06:13
didrocksseb128: do you think that should be -tracking? I have no opinion, but we may have other side-effects I guess…06:13
didrocksok, let me open the bug first06:13
duflu"Software & Updates" tells me it is set to only auto-install security updates06:13
didrocksyeah, it's supposed to only mess with those06:14
dufluHow do we know mutter wasn't a "security update"?06:14
didrockswe don't have security updates in the dev version06:14
didrocksso, it's not in the "security" pocket06:15
dufluYeah, the pocket06:15
didrocksjust the regular release one06:15
didrocksand as you can see in my pastebin those are all regular packages in the release pocket ^06:15
seb128I asked rbalint on #ubuntu-devel06:15
seb128didrocks, k, so sidestepping over, the situation is that we got confused by u-u but the focus fix works for you after all?06:48
didrocksseb128: indeed. I still think that changing the behavior of u-u in a drastic way without warning users is weird06:48
didrocksbut at least, the main point, "focus mutter issue" is fixed06:49
seb128he failed at communicating properly yeah06:59
seb128but I think on principle it's fine to have that behaviour, it gives us testing before landing changes to stable which would be difficult to get otherwise06:59
seb128and unstable series are probably mostly used by distro dev or users who like to test edge things and can deal with the behaviour (if it's documented)07:00
didrocksyeah, we need a reminder that upgrades start applying automatically07:01
dufluRAOF are you around?07:49
dufluHmm, no. Not even in the channel07:49
seb128hey willcooke, how are you?08:00
willcookehi seb128, ok, how about you?08:00
didrockshey willcooke08:00
willcookemorning didrocks08:01
seb128willcooke, good, it's friday!08:02
seb128hey Laney08:02
seb128how are you?08:02
didrockshey Laney08:02
Laneyyeah not bad08:04
dufluMorning willcooke and Laney08:05
willcookeLaney, are you all Ubiquiti'd up at home?08:42
Laneyoh sorry08:43
Laneyhey didrocks duflu ;-)08:43
Laneyand willcooke08:43
Laneywillcooke: yes08:43
LaneyI want to get another access point actually, it doesn't quite cover the whole house properly atm08:43
willcookeLaney, cool.  I think I asked you about it before, and you were a big +1.  I think the time has come for me to admit that my cobbled together collection of cheap no-name Wifi APs has had its time08:44
willcookeLet's see what May brings08:44
willcookethe month, not the witch08:44
Laneynever rely on M... yes indeed08:44
LaneyThey have two different product lines which don't particularly integrate with each other, which is a bit sad08:45
Laneyso I have https://www.ui.com/edgemax/edgerouter-poe/ and https://www.ui.com/unifi/unifi-ap-ac-pro/08:45
Laneybut they have separate software stacks08:45
willcookeedgemax is what I was looking at for the router08:46
willcookeLaney, did you choose the pro over the lite for a specific reason?08:47
LaneyI think it could achieve faster speeds or something? More channels maybe?08:49
willcookeoki, I'll do some more reading08:50
Laneywas a few years ago already08:50
seb128Laney, with the freeze nothing is autoaccepted now right? like I don't need to worry about packages being in a set or not for gstreamer right?09:21
seb128(I don't want to e.G have some universe plugin autoaccepted while the base is not)09:21
Laneyseb128: the script is off09:23
seb128Laney, k, gstreamer stack uploaded to disco and desktop ppa (https://launchpad.net/~ubuntu-desktop/+archive/ubuntu/ppa)09:28
seb128needs to build now (seems like ppa doesn't do depwait?)09:28
seb128let's give it some testing from the team in the ppa once it will be built09:28
Laneysure they do, you just get the × icon09:29
seb128ah, indeed09:29
tseliotseb128, Laney: do you know how to detect that a udev rules is running on the live installer. I would like to make it quit in this specific case, so that LP: #1824177 can be fixed09:35
ubot5`Launchpad bug 1824177 in nvidia-graphics-drivers-410 (Ubuntu) "dmesg spammed with nvidia-nvlink messages during install" [Undecided,Confirmed] https://launchpad.net/bugs/182417709:35
tseliotwillcooke: ^09:35
willcookesorry, I dont know09:38
seb128tseliot, sorry but I don't know either09:39
Laneytseliot: isn't vorlon saying that you shouldn't be doing anything until reboot at all?09:44
Laneybut no, sorry, I'm not aware of an official way to detect a live environment09:44
Laneyyou might want to ship something in casper09:44
LaneyAFAIK udev rules in /run shadow ones in /lib09:45
tseliotLaney: I don't think udev rules have that kind of logic in place. What I was thinking is to make the udev rule quit when on the live installer. So far, I've found out that update-initramfs is diverted on the live installer. Maybe I can use that09:46
Laneyright, that's a casper script09:47
tseliotso, I want to test a condition in the udev rule, and avoid doing any big changes before the release09:48
LaneyI think you probably want to make casper ship a udev rule in /run/udev/rules.d/ that overrides the packaged one09:50
Laney"files in /run take precedence over files with the same name in /lib. This can be used to override a system-supplied rules file with a local file if needed"09:50
tseliotLaney: do you mean a rule with the same name?09:50
tseliotthat's even easier. I can try it here now09:51
Laneyslightly hard bit will be testing the casper change09:51
tseliotLaney: just an empty udev rules with the same seems to do the trick, and I don't get spammed any more09:57
willcookewoot thanks tseliot Laney09:57
Laneytseliot: so that's a casper-bottom script fwiw09:58
Laneylike the one to disable snap refresh09:58
tseliotLaney: ok, then I'll have a look at the sources, and come up with a patch09:58
Laneyto test it you can do break=top from an iso boot and then hack it into place in /scripts09:59
Laneybut it's very bare bones, be warned09:59
tseliotLaney: ok, I'll try that. Thanks10:03
Laney(well, you could probably re-pack the ISO or something too...)10:04
brlinCan anyone review this PR?: https://github.com/ubuntu/snapcraft-desktop-helpers/pull/18510:15
gitbotubuntu issue (Pull request) 185 in snapcraft-desktop-helpers "gnome-platform: Drop unnecessary build-packages" [Open]10:15
tseliotLaney: I wrote the script manually (using tee), and made the file executable (chmod a+x), after booting with break=top, and then I run the "exit" command. /run/udev/rules.d/ wasn't created. Is there any way to check what happened? Maybe a log from casper?11:04
tseliotI don't see anything useful in casper.log11:05
tseliotthe script was in scripts/casper-bottom/11:05
seb128brlin, try asking kenvandine when he gets online11:07
willcookenew ups battery arrived, going off line for a few mins while I change it over.  I can't remember exactly what's connected and what's not, but I assume my modem is.11:18
Laneytseliot: weird, I just tried the same and it worked for me -- did you add it to ORDER and copy all of the boilerplate stuff that the other scripts have at the top?11:22
Laneycasper.log isn't very helpful tho, agreet11:23
brlinseb128: Sure, thanks!11:28
Laney(not sure what ORDER does tbh)11:29
willcookewell, that went better than I expected.11:35
willcookeHardly anything caught on fire11:35
willcookeso it seems upowerd is holding on to one of my upses, and I want nut to manage it.  Any ideas how I can stop upower doing that, or just stop upowerd altogether?11:54
willcookeheh, I just killed it and started nut before it could restart. Problem solved11:55
willcookealso, nut is *the* hardest thing to work with in the world.11:56
seb128gstreamer stack update in https://launchpad.net/~ubuntu-desktop/+archive/ubuntu/ppa if others want to test it11:57
seb128(seems to work fine here)11:57
seb128disco update11:57
* willcooke updates11:57
seb128(tested totem/rb/cheese here)11:59
seb128brb, changing location11:59
andyrockseb128: is there a but about the g-i-s/snap thing?12:15
seb128andyrock, hey, good timing I was about to ping you :)12:16
seb128bug #182418812:16
ubot5`bug 1824188 in gnome-initial-setup (Ubuntu) "Software tab is empty on clean 19.04 install" [High,Confirmed] https://launchpad.net/bugs/182418812:16
seb128andyrock, ^12:16
andyrockI took a look at the code, it's going to take a while to fix it12:16
andyrockalso because reproducing the bug it's not easy12:17
willcookecould you kill snapd, and run g-i-s without it running, then start once g-i-s is running?12:18
seb128andyrock, attach gdb to the snapd service to block it from responding? ;)12:20
willcooke(side note:  a problem I was having with a specific bluetooth mouse not reconnecting automatically is seemingly fixed :) )12:21
seb128clobrano, hey, is that gnome-initial-setup bug about using the incorrect icon still valid? the description pointed out to the welcome-app yaru icon but that has been removed12:21
seb128willcooke, good! was that the gnome-bluetooth update?12:22
seb128willcooke, also did you test the gstreamer updates? I dropped from IRC for a bit so might have missed that12:22
clobranohey seb128! Was the bug opened on launchpad? I can't find it on github12:23
willcookeseb128, just testing now.  Cheese works, downloading b.b.bunny now...12:24
willcookerb radio works12:24
seb128clobrano, yes, bug 180352112:24
ubot5`bug 1803521 in gnome-initial-setup (Ubuntu) "gnome-initial-setup uses wrong icon" [Undecided,New] https://launchpad.net/bugs/180352112:24
clobranolet me check12:24
seb128clobrano, thx12:25
willcookeseb128, @ bluetooth - I think so.  I found an upstream bug which said it was fixed, so sounds right12:26
seb128willcooke, good :)12:26
willcookeSaviq, do you still have your Logitech M557 mouse?  If so, I think that disco has the fix now for reconnecting problems12:27
Saviqwillcooke: yeah I've been working with duflu to verify the fix12:27
tseliotLaney: unless I missed something obvious: http://paste.ubuntu.com/p/vMRpJdRccV/12:29
willcookeSaviq, woot12:29
seb128Saviq, what fix is it?12:29
tseliotLaney: and no, I didn't add it to ORDER. I'll check that12:29
seb128Saviq, I mean are you refering to a specific commit we can SRU?12:29
seb128(bug number?)12:29
willcookeI think its adding the device ids to a file so that it auto tries with pin 000012:29
willcookeI'll find it12:30
SaviqI'll get you guys the big number when I'm back in Ubuntu12:30
willcookeseb128, looks like this one: https://bugs.launchpad.net/ubuntu/+source/bluez/+bug/124698112:30
seb128willcooke, Saviq, https://salsa.debian.org/gnome-team/gnome-bluetooth/commit/ba99a3ce then?12:30
ubot5`Ubuntu bug 1246981 in bluez (Ubuntu) "Bluetooth mouse fails to re-connect after sleep." [Medium,Confirmed]12:30
SaviqThe latest gnome-bluetooth update I believe12:31
willcookeseb128, yeah that looks like it12:31
seb128well the diff doesn't make sense to me though12:31
willcookehm, nothing about this mouse in there12:32
willcookeone sec12:32
seb128willcooke, no, it's fine, I misread it12:34
seb128it's the12:34
seb128 12:34
seb128<!-- Mice don't usually need pincodes -->12:34
seb128<device type="mouse" pin="NULL"/>12:34
seb128bit from the diff12:34
willcookeah, I see12:35
seb128I'm going to SRU that now12:35
willcookewoo, thanks seb12812:35
andyrockseb128: using gdb it's not going to work because gnome-initial-setup will just block12:37
andyrockbecause it's using sync operation12:37
seb128andyrock, ah, it's a sync call without timeout?12:37
andyrockI'll try to make it async12:38
seb128andyrock, or maybe ask #snappy if they can help you putting snapd in the state which makes it not respond12:38
Saviqseb128, willcooke: my bug is https://bugs.launchpad.net/ubuntu/+source/bluez/+bug/177706212:38
ubot5`Ubuntu bug 1777062 in gnome-bluetooth (Ubuntu Cosmic) "Logitech M535 mouse only pairs successfully after 4-5 tries" [Undecided,New]12:38
clobranoseb128: yes, since the Yaru icon is not there anymore, that specific behavior is invalid. However the "problem" should still be there12:39
seb128clobrano, what's the icon you expect to be used between the two?12:39
willcookeSaviq, thx12:40
seb128Saviq, willcooke, that bug description doesn't make sense to me. gnome-bluetooth only handle the pairing from the settings, not happens later12:40
seb128I mean if it pairs it in a way that is not persistent/trusted in bluez that explain it has to be paired again12:40
willcookeseb128, Totem working with an h264 working.  Promoted me to install the codec, did that from the ppa, restart totem, loaded the video, moar rabbits12:40
seb128but if you pair using bluetoothctl then gnome-bluetooth is out o fthe picture12:40
seb128willcooke, great, thx for testing!12:41
willcookeseb128, is the pin-code-database shared between bluetoothctl and gnome-bluetooth perhaps?12:41
seb128willcooke, gnome-bluetooth used to have overrides, the change is basically to drop that to use the bluez one12:42
seb128bluez doesn't use anything from gnome-bluetooth no12:42
seb128it's a lower level component, doesn't require GNOME nor event a graphical session12:43
seb128willcooke, the gnome-bluetooth change should fix the case of "pairing done through GNOME doesn't persist"12:43
seb128which is probably most reports12:43
seb128the "pairing with bluetoothctl has problem" from Saviq sounds like it's different though12:44
seb128maybe that was a kernel issue or such also fixed in disco12:44
willcookeI bet duflu will know12:45
seb128I will SRU the gnome-bluetooth fixes since they make sense12:46
seb128that's a good first step12:46
clobranoseb128: yes, the bug was opened because the unexpected icon was used, and this issue is gone12:46
clobranoseb128: but I thought that the problem is due to a difference in how the .desktop files where written12:46
seb128clobrano, right, if we standardize we need to pick one icon or the other, I just don't know which one makes more sense ... do you have an opinion on that?12:47
clobranoseb128: I have no strong opinion, I can only say that "preferences-system" doesn't seem specific enough if the real target is the ubuntu-logo12:48
seb128clobrano, thx12:48
clobranoseb128: np :)12:48
clobranoomg, just noticed... s/where/were ^12:49
kenvandinebrlin: i'm on it13:11
brlinkenvandine: <313:14
willcookemorning kenvandine13:20
kenvandinehey willcooke13:23
andyrockseb128: if we want to get rid of most of the gedit crashes it would be nice to have this in disco: https://gitlab.gnome.org/GNOME/gedit/merge_requests/3213:26
gitbotGNOME issue (Merge request) 32 in gedit "build: Reintroduce enable-gvfs-metadata option" [Merged]13:26
seb128Laney, ^ do you have slots to help reviewing the few pending fixes from andyrock for gedit today?13:26
andyrockseb128, Laney: the other fixes are not strictly necessary if we get this in debian/disco13:27
seb128andyrock, why not? they don't fix other problems?13:27
seb128well you mean they are less important?13:28
andyrockthey do, but if we enable gvfs metadata the crash is not going to happen13:28
andyrockthey're necessary if and only if gvfs metadata is disabled13:29
andyrockto summarize:13:31
andyrock1. https://gitlab.gnome.org/GNOME/gedit/merge_requests/34 is necessary both if gvfs is enabled and if it is disabled (but we already distro patched it)13:32
gitbotGNOME issue (Merge request) 34 in gedit "open-document-selector: Properly remove idle" [Opened]13:32
andyrock2.  https://gitlab.gnome.org/GNOME/gedit/merge_requests/35 it is necessary only if gvfs is disabled13:33
gitbotGNOME issue (Merge request) 35 in gedit "metadata-manager: Remove singleton" [Opened]13:33
seb128andyrock, oh ok13:33
andyrock3. tab: Cancel loading operation when tab is disposed this does not fix a crash but I saw it while debugging so I fixed it13:33
andyrocknow the commit that just got merged re-enabled gvfs by default13:33
seb128jbicha, ^ or maye you can review since you have been sort of gedit maintainer in recent cycles :-)13:34
andyrockseb128: jbicha said that I can merge them if I want but I can't because I'm not a GNOME developer13:34
seb128andyrock, I try to get it later today if Laney can't but I'm currently on a packed todolist for the day :/13:35
andyrockand now it's too late for me to ask 💔13:35
andyrockmy suggestion for Laney is to review/merge https://gitlab.gnome.org/GNOME/gedit/merge_requests/3413:35
gitbotGNOME issue (Merge request) 34 in gedit "open-document-selector: Properly remove idle" [Opened]13:35
andyrockit should be safe enough13:35
seb128willcooke, Saviq, https://launchpadlibrarian.net/418989286/gnome-bluetooth_3.28.0-2ubuntu0.2_source.changes (used another bug for the SRU which looked like the issue and was less old/busy, those reports which collect problems over years often lead to difficult SRU validation because some of the people subscribed have another problem and will claim your SRU doesn't work)13:36
seb128andyrock, that's the wrong one right?13:36
seb128that one we already backported/SRUed13:36
seb128you mean !3213:36
seb128ah, you mean upstream13:37
andyrockI know that it has been backported but it still nice to have it upstream13:37
seb128you are correct :)13:37
willcookeseb128, @ bluetooth: will confirm on this machine13:37
andyrockseb128:  also I can distro-patch the commit to enable gvfs by default after I take care of g-i-s13:37
seb128andyrock, k, let's see what L_aney says when he's back from lunch13:38
seb128andyrock, thx13:38
willcookeW: An error occurred during the signature verification. The repository is not updated and the previous index files will be used. GPG error: http://dl.google.com/linux/chrome/deb stable Release: The following signatures were invalid: EXPKEYSIG 1397BC53640DB551 Google Inc. (Linux Packages Signing Authority) <linux-packages-keymaster@google.com>13:40
willcookeW: Failed to fetch http://dl.google.com/linux/chrome/deb/dists/stable/Release.gpg  The following signatures were invalid: EXPKEYSIG 1397BC53640DB551 Google Inc. (Linux Packages Signing Authority) <linux-packages-keymaster@google.com>13:40
willcookeW: Some index files failed to download. They have been ignored, or old ones used instead.13:40
seb128well done google :p13:41
kenvandinebrlin: thanks, that's a nice improvement.  Merged!14:00
brlinkenvandine: Thanks!  Here's another one: https://github.com/ubuntu/snapcraft-desktop-helpers/pull/18414:01
gitbotubuntu issue (Pull request) 184 in snapcraft-desktop-helpers "bindtextdomain: Support locale files under share/locale" [Open]14:01
kenvandinebrlin: saw it, that only helps snaps with --prefix=/14:06
kenvandinecan't imagine there are a ton of those14:06
brlinkenvandine: Snapcraft autotools plugin by default built with empty prefix14:06
kenvandinealthough it time cost should be negligible14:07
kenvandineempty usually defaults to usr or usr/local14:07
brlinkenvandine: That's not what I encountered though :-/14:08
brlinIt does specify `--prefix` while configuring , just set it as null: https://github.com/snapcore/snapcraft/blob/77c17b26573e474783e9535bf66ce3fe358b71f1/snapcraft/plugins/autotools.py#L12114:09
kenvandinebrlin: indeed14:10
kenvandinebrlin: merged14:11
brlinkenvandine: Thanks a bunch!14:11
kenvandinebrlin: thank you!14:11
Laneyseb128: yeah I saw the ping about gedit somewhere else14:18
seb128Laney, it was trello :)14:19
Laneyso I had the tab open already, but thx for the re-ping14:19
Laneyseems SRUable to me14:19
seb128Laney, but that other bug andyrock was pinging about seems worth trying to get into disco proper so maybe one for today if we can squeeze14:19
seb128Laney, thx for catching that tracker buggy upload, trying too much this week and not taking time to do some of them properly :/14:44
seb128Laney, I was going to upload to Debian but didn't want to wait to wait for launchpad to pick the uploads/sync and miss friday to get it reviewed14:45
Laneynp, good when reviews work14:45
seb128Laney, but feel free to take over that :)14:45
Laneywell you can do a ~fakesync upload14:45
Laneylike for gstreamer14:45
Laneybut also things can be SRUed at this point, maybe that makes for fewer mistakes with less rushing14:45
seb128right, it's just that jbicha commited 2.2.1 things to the vcs14:45
seb128and I didn't want spend time sorting out what to do there at the time14:46
Laneyyep, I'm making a branch from the 2.1.8-1 tag14:46
Laneygit checkout -b debian/buster debian/2.1.8-1 and update debian/gbp.conf14:46
seb128thx for the tip :)14:46
seb128yeah, I know for SRU, I've difficulties to be in peace with that though :(14:46
seb128we always end up getting reports in the error tracker and launchpad until end of times from users who don't apply updates for some reasons14:47
Laneyyeah like we get people who insist on running devel-proposed14:47
seb128but that's probably my issue, I should accept that such are software users14:47
* Laney has made peace with these things14:47
seb128one day you need to teach me how :)14:47
seb128anyway, I do try to land some of the crashers fix in the release for those reasons14:48
seb128but I'm not pushing so fine if they sit in the queue or in proposed until SRU time14:48
Laneyokey, maybe make sure they have a bug report attached then14:48
Laneynot saying they don't, haven't checked for that14:48
seb128yeah, that one had14:49
seb128g-c-c doesn't for the translation fix I added14:49
seb128worth thing it get rejected and it's a bit more work14:49
seb128I try to that for the next ones14:49
seb128should I reupload the gstreamer stack for a SRU reference?14:49
seb128probably for -release14:49
Laneyprobably going to accept those ones in a minute14:49
seb128Laney, thx14:51
seb128hey Trevinho, how are you?14:52
Trevinhoseb128: hey, all good!14:53
seb128happy friday btw :)14:53
seb128the work day only starts for you, but any fancy plan for the w.e ;)14:54
andyrockseb128: do you want me to add an error message in g-i-s in case it fails to get the featured snaps ?14:55
andyrockI've added a retry behaviour14:55
andyrockit's going to try to every 250 ms for a maximum of 5s14:55
andyrockbut what if it fails? showing an error will required a UIFe14:56
seb128andyrock, I think new string is fine in that case14:59
seb128willcooke, Laney, ^ opinion? (about bug #1824188)15:00
ubot5`bug 1824188 in gnome-initial-setup (Ubuntu) "Software tab is empty on clean 19.04 install" [High,Confirmed] https://launchpad.net/bugs/182418815:00
willcookeFor the string, how about... "Unable to retrieve applications at this time.  Please open Ubuntu Software to see what's available."15:01
seb128willcooke, we do want a string/UI change right? even if it's untranslated at this point?15:01
seb128andyrock, also chipaca was saying yesterday that it takes 90s for snapd to be ready in his kvm15:02
seb128willcooke, did you do any timing testing at the end?15:02
Trevinhoseb128: happy Friday you too :-), well... some more sightseeing I expect15:02
willcookeI did, and it was no where near 90 seconds.  Let me fire up the vm15:02
seb128andyrock, so maybe 1s for 30s or something would be better15:02
willcookeoh, except I will need to reinstall15:02
seb128willcooke, no hurry, but I think we still want to push a bit further than 5s retry total15:03
kenvandinei noticed the same problem in hyperv yesterday15:03
kenvandineempty page15:03
willcookeIt certainly feels like something has got worse there on the snapd side15:03
andyrockwho's going to wait 90s for that page? XD15:04
Laneyis that a spinner while waiting for the list of snaps?15:04
willcookeLaney, correct15:05
seb128andyrock, well, I think that's total from boot that it takes 90s15:05
andyrockLaney: I'm going to add a spinner too15:06
willcookeandyrock, seb128 - if we don't have apps to show, we should remove the "You can use software to install apps like these" too15:06
seb128andyrock, so time your session start, the wizard load, you go through previous steps15:06
andyrockseb128: kk15:06
seb128and 90s is on slow vm15:06
seb128hopefully it's not that slow for everyone :p15:06
kenvandineeven 30 seconds is way to long15:06
seb128I think the snapd team should care/own those sort of issues15:07
kenvandineimo it should be able to return results without it needing to seed15:07
seb128instead of bouncing back to us as our problem every time15:07
seb128that's a bit irritating15:07
kenvandinei can see not being able to install15:07
kenvandineit would queue15:07
Laneydone is insensitive while the spinner is showing or what?15:07
kenvandinebut get the list from the store shouldn't be blocked by seeding15:07
LaneyI don't really like showing a weird error15:07
kenvandinemvo: ^^ thoughts?15:07
Laneyjust let them quit it15:08
willcookeThe other option imo, is to not launch g-i-s until snapd is ready.  But that could be a long wait and pop up seemingly randomly15:08
Laneykeep trying forever as long as the page is open15:08
Laneybut let people get out of it whenever they want15:08
seb128+1 for that15:08
willcookeah, that sounds good Laney15:09
willcookeseb128, what are you +1 ing?15:09
seb128andyrock, ^15:09
kenvandineand we should bug the snapd team about making it handle queries while it's seeding15:09
seb128willcooke, spin while the page is on screen, let user close anyway if they don't want to wait15:09
seb128also it's no new string and less UI change15:09
Laneyotoh an eternal spinner could be confusing15:09
Laneybut probably less bad than an error?15:09
seb128hopefully it's never eternel15:09
willcookeI think then, we should hide the "you can use software like...." as well, until its ready15:10
willcookeseb128, it could be for ever if the machine is off line15:10
seb128it should take a 1 minute or so max15:10
willcookeso 90 seconds max seems sensible15:10
seb128what do we do today for those?15:10
willcookewell, 9515:10
kenvandineif we're offline we shouldn't show the page at all15:10
willcookeah, kk15:10
kenvandineand i think snapd can tell you if it's offline15:11
mvokenvandine: we looked at this (pawel to be precise) and its the fc-cache generation that takes the time - if we could generate caches for v6/v7 as part of the install we cut a significant amount of time (>20s in our tests)15:11
kenvandinehopefully it doesn't report as offline when it's seeding15:11
andyrockwait so should we show an error or not?15:11
andyrockiirc I can hide a page in g-i-s if there is no connection15:11
andyrocklet me check15:11
kenvandinemvo: my real question was can we let snapd respond to queries while it's seeding?15:11
Laneythe other option is don't show it at all if it's not ready15:12
kenvandineso let s get the list of snaps in ubuntu-firstrun category15:12
kenvandines/let s/let us/15:12
mvokenvandine: queries like store searches?15:12
Laneyandyrock: we're saying just keep trying forever as long as they are sitting on the page I think15:12
Laneybut make "Done" sensitive if you're doing that15:12
seb128Laney, we did fear that in the current state of things it would mean not showing it for most users because snapd got slow in disco15:12
kenvandinemvo: we just need a list of snaps from a category15:12
andyrockLaney: kk15:12
mvokenvandine: let me quickly check with the team15:12
kenvandineeven if we can't install them yet15:12
kenvandineright now we can't get that list until seeding is done15:13
kenvandineit blocks15:13
willcookehm, if we get the list, and someone clicks on it they get taken to Software.  What happens then?  Will Software error, or will it just wait?15:13
seb128brb, changing location15:13
kenvandinewillcooke: it should queue15:13
willcookeI'd like to test that15:14
willcookeI think you're right, but ya know15:14
willcookeahoy Chipaca15:14
mvokenvandine: Chipaca says "yes we can"15:14
mvokenvandine: it seems the problem is that the connect is too early? sorry I was not following all the details but Chipaca  I think was15:14
Chipacathe problem isn't querying while seeding (although "list of installed snaps" when it's in the middle of installing will be racy)15:14
mvokenvandine: and for fc-cache we have pstolowski  who knows the details15:14
Chipacathe problem is not retrying on connect/read error15:15
Chipacabecause snapd will go away in the middle of seeding, that's normal and expected15:15
Chipacait takes less than a second to come back (~200ms on a loaded machine)15:16
willcookeso a few retries might just be enough then15:16
Chipacanote that there's a difference between a low-level error like 'the socket went away' or sth, and an error response -- but i don't know how snapd-glib exposes that15:17
Chipacawhat you're wanting to retry is the former15:17
andyrockactually snapd-glib has code to retry on connect but no on read15:18
willcookeyeah we're getting  Connection reset by peer on read15:19
Chipacaalso, snapd won't go away in the _middle_ of a request15:19
Chipaca(or it shouldn't! :-) bugs might exist)15:19
Chipacawe go away gracefully15:19
kenvandineok, so maybe this is something we can improve in snapd-glib?15:20
Chipacai need to repeat that getting the list of installed snaps is dubious when snapd is still seeding, so maybe that logic needs tweaking15:20
Chipacabut sections, featured snaps, etc, as long as the network is up should just work15:21
kenvandineChipaca: we don't need the list of installed for this use case15:21
kenvandinejust the list for a single category15:21
Chipacakenvandine: i mean, you're getting it15:22
Chipacait's the first thing you do15:22
andyrockkenvandine: snapd-glib can be smarter but it's easier to fix it in g-i-s15:23
andyrockat least at this point of the ccle15:23
kenvandineandyrock: great :)15:23
Chipacakenvandine: line 826 of debian/patches/0001-Add-Ubuntu-mode-with-special-pages.patch15:24
Chipacainstalled_snaps = snapd_client_list_sync (client, NULL, &error);15:24
kenvandinei wonder why15:24
kenvandineandyrock: ^^15:24
Chipacathat's why i say that logic needs tweaking :-)15:24
kenvandinethat seems pointless15:24
Chipacakenvandine: /* Skip if already installed */15:25
kenvandineunless it's to filter out15:25
Chipacakenvandine: also /* Ignore common snaps that have default .debs */15:25
kenvandinewe should just trust that section is properly currated15:26
kenvandineandyrock: what do you think?15:26
Chipacayou _could_ look at what's going to get seeded and skip those15:26
kenvandineubuntu-firstrun should only container what we want displayed in g-i-s15:27
Chipacabecause otherwise in 2 years, are you going to remember that the snap you seeded in 19.10 is now featured15:27
Laneyplease leave bigger refactorings for after final freeze15:27
seb128(back online)15:27
seb128willcooke, on my inspiron on a fresh install from today it took like a minute for the wizard to open15:27
ChipacaLaney: it's C, the binary is tiny, so that's small right?15:27
* Chipaca hides15:27
seb128willcooke, still the software page was empty :/15:28
Laneyand we all know C developers are the smartest, so actually the smallest chance of any bugs appearing :>15:28
Chipacakenvandine: because the code carries on even if that snapd_client_list_sync call fails, it should be fine to just comment out that line15:28
Chipacakenvandine: (although I have no idea if the code actually _works_ if that call fails, as opposed to freaking out over it being nil)15:29
Chipacaalso it's starting to feel like i'm teaching my grandmother something about eggs15:29
Chipacaso i'm going to step away from this now15:29
kenvandineChipaca: thanks!15:30
Chipacakenvandine: you know where to find me15:30
GunnarHjHi seb128, Laney! Me again with another g-c-c MP. Not perfect, but a step in the right direction.15:31
GunnarHjTo late?15:31
willcookeseb128, I've seen it be slow to open on a few installs, but could never work out why15:31
pstolowskiwillcooke: have you discussed anything re fc-cache issue?15:31
willcookepstolowski, erm, dont /think/ so15:32
LaneyGunnarHj: might be better to keep working on it and SRU this?15:32
pstolowskiwillcooke: ah, i see, i was told it was brought up here already15:33
pstolowskiwillcooke: ok, let me explain15:33
kenvandinepstolowski: i don't think the fc-cache bit is really our issue15:33
kenvandineso you can disregard15:33
GunnarHjLaney: May be. I'd say that if you do yet another upload for some other reason, it would be good to include it. Otherwise wait.15:34
seb128GunnarHj, what does it fix compared to the previous change?15:34
pstolowskikenvandine, willcooke: it's about the installer and populating fontconfig cache on install, before reboot15:35
GunnarHjseb128: I displays correct labels, for instance "Serbian" and "Serbian - latin" instead of "Serbian" and "Serbian".15:35
seb128GunnarHj, ah, we can probably do that in a SRU, I'm sure we are going to do some for g-c-c15:35
Laneybut they don't work properly right, so it's still going to be a bit weird/confusing15:36
GunnarHjLaney: True.15:36
seb128k, so SRU sounds better, sort it out properly and then upload15:40
willcookepstolowski, on install of the snap?15:41
seb128willcooke, btw on that install on the inspiron the world map was really streched horizontally and the dots such misplaced (like Paris was in the north sea), did you see that problem?15:41
willcookeseb128, negative.  Let me download todays ISO15:41
seb128could have a one off, I'm going to try again15:42
andyrockkenvandine: the installed snaps is to filter them out15:43
andyrocknot sure about the firefox thing15:43
andyrockI didn't write that page :D15:43
pstolowskiwillcooke: snapd generates fc-cache if it doesn't exist, this currently happens after seeding on the first run after installation and takes around 20s. we could probably save a lot of that if the installer simply populated font cache of the installed system15:44
pstolowskiwillcooke: i mean simply copied /var/cache/fontconfig from live session15:44
seb128pstolowski, would that be enough? shaving 20s of 90s is still making likely that snapd is not ready on time for the UI15:45
willcookeandyrock, kenvandine  - I sort of remember something about that.  It was when people upgraded that we wouldnt pimp stuff they already had, so that other stuff had a chance to fill the gaps15:46
seb128willcooke, andyrock, kenvandine, well, not only upgrade, no point recommending things we pre-install as deb/they already have installed out of the box15:46
willcookeseb128, @fc-cache I guess this is orthogonal to the g-i-s issue15:47
willcookeseb128, @ preinstall - +115:47
seb128the fc-cache thing would be an optimization15:47
pstolowskiseb128: what is the 90s you're talking about?15:47
seb128but I don't think it's rls topic at this point15:47
willcookeso it seems like a nice time saver anyway, but too late now for 19.04 of course :)15:47
willcookepstolowski, Chipaca saw 90s from booting his disco vm to snapd being "done"15:48
willcookedesktop started at around 30s from what I remember15:48
seb128pstolowski, https://irclogs.ubuntu.com/2019/04/11/%23snappy.html#t12:4615:48
willcookepstolowski, Given that it's too late for this cycle, could you log a bug with the suggestion?15:49
pstolowskiwillcooke: will do, thanks15:50
pstolowskithanks seb128, i'll check this out15:50
andyrockseb128: kenvandine we should just remove firefox from the ubuntu-firstrun section15:51
kenvandinei'd be ok with that15:56
willcookeI dont understand why it's even there15:56
willcookeit shouldnt be in the curated list15:56
andyrockseb128: willcooke https://usercontent.irccloud-cdn.com/file/oyoTc3Dn/image.png16:05
andyrocknot showing "You can use...." will make the spinner weird16:05
willcookehow weird?  Just that you don't know what it's doing?16:06
willcookegot ya16:06
andyrockin theory the spinner will never be visible16:06
andyrockbecause what's going to happen is the following:16:07
andyrock1. g-i-s ask snaps the list of snaps to show16:07
andyrock2. snapd restarts in the middle of the request16:07
andyrock3. g-i-s receives a failure16:08
andyrock4. g-i-s try to connect to snapd again16:08
andyrock5. not it works16:08
willcookes/not/now ?16:08
andyrockyeah now :D16:09
andyrock1-5 should be pretty fast16:09
andyrockbut a spinner still makes sense in case it's not fast16:09
willcookeShould we add "Loading..." or something as well, for cases where its not fast?16:10
Laneywe're trying to avoid a new string16:12
willcookefair enough16:12
andyrockthe problem really needs to be fixed in snapd-glib because (I guess) snapd dies after he wrote/send a request to snapd16:12
andyrockthere is some logic to re-send the request if write fails because snapd died but there is no logic to retry if snapd died before/while reading16:13
Laneysounds worthy of a bug report / trello card16:13
willcookeandyrock, would you mind opening a bug against snapd-glib so we can get Robert to take a look16:13
willcookewhat L_aney  said16:13
andyrockyeah I'll do that later16:13
andyrocklet me try to finish the branch asap16:13
willcookethanks andyrock, nice work16:14
seb128Trevinho, https://bileto.ubuntu.com/#/ticket/3698 ... the scope upload revert a change which was probably missing from the vcs but is in the ubuntu archive16:17
seb128-  * Add "X-Ubuntu-Use-Langpack: yes" to debian/control now that this package is16:17
seb128-    in Universe (LP: #1760435).16:17
ubot5`Launchpad bug 1760435 in unity-scope-manpages (Ubuntu) "Use Ubuntu language packs for various Unity packages" [Undecided,Fix released] https://launchpad.net/bugs/176043516:17
seb128-X-Ubuntu-Use-Langpack: yes16:17
seb128GunnarHj, did you see bug #1823722 btw?16:20
ubot5`bug 1823722 in mutter (Ubuntu) "keyboard shortcuts are displayed not translated" [Low,Confirmed] https://launchpad.net/bugs/182372216:20
seb128bug #182449716:24
ubot5`bug 1824497 in nautilus (Ubuntu) "unable to copy small files with nautilus on cifs mount" [Low,New] https://launchpad.net/bugs/182449716:24
seb128ups, sorry, wrong tab16:24
=== pstolowski is now known as pstolowski|afk
willcookeseb128, did you see the big map in UEFI or legacy mode?16:33
willcookewere you in the live session when you saw it16:33
willcookeseb128, tried both options - it looks "ok" to me I think.  Doesn't look wrong at least, but maybe a little stretched out, nothing that would cause me to think "oooh, that's broken"16:40
GunnarHjseb128: I can take a look at bug #1823722 this weekend.16:42
ubot5`bug 1823722 in mutter (Ubuntu) "keyboard shortcuts are displayed not translated" [Low,Confirmed] https://launchpad.net/bugs/182372216:42
* willcooke -> dinner. bbiab16:51
Trevinhoseb128: ah... sorry I missed the mention. Mh, yeah... wasn't in the vcs. I need to add that and should we do another upload I suppose?16:52
Laneydidn't get to gedit, sorry, might get some time later tho17:11
willcookeLooks like we're suspending by default now.  I have a feeling we turned that off by default before.  Anyone remember?17:19
willcookeby which I mean, if you leave the machine alone it's suspending17:19
willcookerather than just turning off the screen17:19
willcookeMaybe I just turned that off before17:20
willcookeah, this is when on battery only17:20
andyrockseb128: https://code.launchpad.net/~azzar1/ubuntu/+source/gnome-initial-setup/+git/gnome-initial-setup/+merge/36596617:24
andyrockhopefully it's enough17:24
seb128willcooke, it was on "install only", uefi17:54
seb128willcooke, I will try again on monday, no worry17:54
seb128Trevinho, I guess yes...17:54
seb128willcooke, right, I wonder if that's a requirement from the new energy standards also to do that (suspend on idle on battery)17:55
seb128well it does make sense on battery imho17:55
seb128andyrock, great, thx17:55
Trevinhoseb128: ok, there are not many translated strings, but indeed better17:55
Trevinhoi'll prepare it17:55
seb128Trevinho, well, better and also I don't think the r-t is going to let you revert a change without a solid reason at this point and "it was not commited to the vcs" isn't one17:56
willcookeseb128, @ battery - yeah, I think it makes sense, once I realised it was only on battery.  Easy to change anyway17:58
Trevinhoseb128: better to use a .1 or jut bumb the ubuntu number?17:58
willcookeseb128, secure boot on or off?17:59
willcookeah ha!  Also "safe graphics mode" appears :)18:00
seb128willcooke, secure boot on18:10
seb128Trevinho, either is fine18:10
willcookeseb128, oki, thx.  Does it look ok to you from the photo?18:10
seb128I'm doing another install to see if it was a one time thing18:11
willcookeoki.  Probably the weekend now though? :)18:11
willcooke(I'm hiding from the kids atm)18:12
seb128yeah, that and upload gis and then done18:12
seb128we already had dinner18:12
seb128he's going to bed now so I can be back a bit at the computer :)18:12
willcookeI'm trying to pair an Amazon Fire TV stick with an Echo dot18:13
willcookeand doing some install testing at the same time18:13
GunnarHjseb128: Think I fixed the mutter translation issue. (MP)18:22
seb128GunnarHj, looking, oh nice, it's easier that I though :)18:26
seb128Trevinho, Laney, https://code.launchpad.net/~gunnarhj/ubuntu/+source/mutter/+git/mutter/+merge/365971 looks like, we should probably have it in for the next upload/SRU18:26
GunnarHjseb128: I noticed that a full language pack update has just been accomplished. But maybe another delta update can be made before release?18:28
Trevinhoseb128: ok18:28
seb128GunnarHj, I doubt it's important enough to be worth doing that before release18:28
seb128it's only keybinding in settings18:28
GunnarHjseb128: Ok.18:28
seb128it's not a LTS cycle, we accept a little less polish18:28
seb128but thanks for carring :)18:28
Trevinhohopefully 3.32.1 is also coming soon18:28
seb128g-s upstream is like you Marco18:29
seb128liking to do things when it suits them :)18:29
Trevinholate? xD18:29
seb128schedule? what? who care, enjoy life18:29
Trevinhoah, well... You know it's maintained by someone living in Spain for some time now, so....18:30
seb128worth than the italians right? ;)18:30
Trevinhoquite likely18:30
Trevinhocan't complain this cycle though, so far I was in the defined times :P18:32
seb128on the edge but in indeed!18:33
Laneyseb128: If you want to add mutter delta, I'm not the person to talk to. Don't see why that should be delta, but whatever.18:34
* Laney goes away again18:35
seb128Laney, I don't think it should be a delta?18:35
seb128never said that?18:35
seb128oh, you mean it should go on salsa/debian, yeah I'm not arguing against that ... GunnarHj, do you think you can mp to debian too?18:36
seb128Laney, I meant that the delta seemed to make sense, I didn't even think about whether it was targetting debian or Ubuntu, just thinking about the diff by itself18:36
seb128sorry if that looked like me being careless or something18:37
seb128anyway, g-i-s uploaded, time to call it a day (might still look at backlog before going to bed in case there is a problem with that upload or something else)18:38
seb128andyrock, ^18:38
seb128enjoy the w.e desktopers!18:38
GunnarHjseb128: Sure, I can MP to Debian if you like. They probably don't need it since they don't run dh_translations, but OTOH it doesn't hurt them.18:39
seb128GunnarHj, right, and L_aney is right that we shouldn't add Delta to Ubuntu if we don't need to (I didn't plan to for the record, I was mostly thinking about Disco/SRU at this point which is probably not going to be a sync, but going forward it should go to Debian and be synced)18:40
GunnarHjseb128: Got it.18:41
seb128thx again GunnarHj18:41
seb128have a good w.e :)18:41
GunnarHjseb128: Same to you.18:41
andyrockseb_128 thanks!18:57
willcookesee you seb12819:00
willcookeswitching off now too.  In London on Monday, might be late online19:01
willcookeWill be on email etcs19:01
willcookenight all19:01

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