[00:05] <stochastix> f i do,  snap find ,  is that supposed to list all of the snap available?  If I run snap find tele, I get telegram, but otherwise it is not in the long list?
[00:05] <stochastix> I am trying to just get an exhaustive list of the snaps available to look at them.
[01:06] <tsimonq2> elopio: do you use zsh? how did that idea get formed?
[01:06] <elopio> tsimonq2: I met somebody who uses zsh.
[01:06] <elopio> I have a lot of new friends today ^_^
[01:19] <tsimonq2> elopio: :D
[01:19] <tsimonq2> elopio: are you at a sprint or something? :P
[04:06] <slangasek> ogra_: https://github.com/CanonicalLtd/ubuntu-image/pull/63 landed
[04:06] <mup> PR CanonicalLtd/ubuntu-image#63: Don't build 4GB images, build images as big as the required contents <Created by vorlonofportland> <Merged by vorlonofportland> <https://github.com/CanonicalLtd/ubuntu-image/pull/63>
[05:31] <sparkin> hello my fellow crafters!!!!!!!
[05:34] <sparkin> crickets...
[06:55] <mvo> ogra_: was just thinking about the recent change to ubuntu image to make the disk only as small as needed. this does mean that if you run it directly in kvm/qemu the first snap you download causes a disk-full error?
[06:55] <dholbach> hey hey
[06:57] <mup> Bug #1623802 opened: snapd does not properly detect service/mount unit status <Snappy:New> <https://launchpad.net/bugs/1623802>
[06:59] <mvo> pitti: I'm curious why the autopkgtest test runs take so long after an upload, is it because the machines that do the testing are so busy? or is there another reason?
[06:59] <mvo> pitti: (background is my snapd upload from 12h ago that did not yet get run inside adt)
[07:03] <pitti> mvo: yeah, sorry -- look at http://autopkgtest.ubuntu.com/running
[07:03] <pitti> mvo: queues got double-Qt'ed yesterday, and the backlog lasted until this morning; s390x and armhf are still catching up
[07:05] <mvo> pitti: it says queue length is "-" for yakkety/amd64, does that mean the queue is empty?
[07:05]  * pitti orders another Binford 6100 cloud  -- we need MORE POWER
[07:05] <pitti> mvo: yes (some amd64 tests are stilll running, though
[07:06] <mvo> pitti: hm, I don't see snapd on the page but there is a new version https://launchpad.net/ubuntu/+source/snapd/2.14.2+git.1.afda4fc-0ubuntu2
[07:06] <mvo> pitti: what am I missing :) ?
[07:06] <pitti> mvo: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#snapd ?
[07:06] <pitti> mvo: you are missing the FTBFS on arm64
[07:06] <mvo> pitti: oh, meh, ok. this got fixed this morning, but I guess that is too late
[07:07] <mvo> pitti: ok, thank you!
[07:07] <mvo> pitti: so that means it will be considered soon :) ?
[07:10] <pitti> mvo: as soon as it builds, yes
[07:23] <mvo> pitti: cool, thanks. it has build ~1h ago, but I guess it takes a bit until the script catches up (or does it need to get published first?)
[07:25] <pitti> mvo: yes, britney only sees published packages
[07:25]  * mvo nods
[07:30] <zyga> quick noob question, how do I build our kernel in a way that lets me incrementally build / reinstall / reboot while tweaking some things?
[07:31] <zyga> (where what I care about is identical config and initial source as packaged kernel and incremental build after any change)
[07:46] <mup> PR snapd#1916 opened: backends: first bits of kernel-module security backend <Created by stolowski> <https://github.com/snapcore/snapd/pull/1916>
[08:00] <ogra_> mvo, yeah, we need a --size option
[08:00] <ogra_> so you can define a minimal size
[08:01] <ogra_> or allow to pre-create a target img (though thats more ugly)
[08:03] <mvo> ogra_: indeed
[08:07] <ogra_> mvo, i see steve added an image_size variable to builder.py, i guess we just need the option at the top level and hand the value over
[08:09] <mvo> ogra_: yeah, I was thinking it would be pretty easy to add
[08:10] <ogra_> as a quick hack: os.environ.get('IMAGESIZE') ... ;)
[08:10] <ogra_> wouldnt even need an option
[08:19] <ogra_> http://paste.ubuntu.com/23181304/
[08:21] <ogra_> hrm
[08:22] <ogra_> i get "Failed to mount /boot/grub." ... even with the beta channel
[08:22] <ogra_> oh
[08:23] <ogra_> FAT-fs (sda2): error, fat_get_cluster: invalid cluster chain ...
[08:25] <ogra_> hmm, weird, steve didnt touch anything in the vfat creation
[08:27] <bull> hello guys my app which is written in qt have a functionality like- when user click on open file with editor , it open the loaded file with system editor ex gedit , kwrite or whatever user having as per their configuration . after packaging my app with snap the functionality is broken, are you guys working in it ???
[08:29] <bull> am sure there is nothing wrong with my code as you can see this feature in hundreds of other applications out there. snap need to fix this :D
[08:29] <pitti> mvo: snapd tests are queued now
[08:30] <ogra_> aha ... using mtools from proposed helps :)
[08:30]  * ogra_ has a working 500MB kvm image
[08:30] <pitti> mvo: oh, so it wasn't FTBFS, just finished building around that time
[08:30] <ogra_> ogra@localhost:~$ df -h /writable
[08:30] <ogra_> Filesystem      Size  Used Avail Use% Mounted on
[08:30] <ogra_> /dev/sda3       542M  190M  313M  38% /writable
[08:31] <ogra_> bull, for now you need to ship the editor you want and tell your app to use this one ... long term i think zyga works on a sensible-editor interface or some such
[08:32] <bull> ogra_,  you mean gedit or kwrite inside snap as stage package ??
[08:32] <bull> this will fill snap with almost all kde and gnome dependencies am sure :D
[08:32] <ogra_> (i know he does something like this for browsers already, editors should just be a variant of that)
[08:32] <bull> oh
[08:33] <ogra_> bull, yes, i mean that ... just pich one with less deps then
[08:33] <ogra_> *pick
[08:33] <bull> ogra_, my app cant open url in external web browser
[08:33] <ogra_> yes, same thing
[08:34] <bull> ogra_, i have to add sensible- browser in snapcraft yaml ?
[08:34] <mvo> pitti: neat, its running
[08:34] <mvo> pitti: can't wait, I had it down to 1 test failure due to no (non-proxy) network
[08:34] <ogra_> no, you would have to add the browser you want ... there is also work on making xdg-open work if your snap runs on a classic system ... not sure where that stands
[08:35] <mvo> pitti: and in qemu its all working perfectly, so fingers crossed
[08:35] <bull> ogra_,  you mean i have to add a browser inside my snap ??
[08:35] <bull> damn
[08:35] <bull> :D
[08:35] <bull> like chrome or firefox ?
[08:35] <ogra_> hypothetically ...
[08:35] <ogra_> i'd just wait til the right interfaces are there for that one :)
[08:36] <bull> thats really funny :P
[08:36] <bull> and if my app wana talk with snapd installed on user's system ??
[08:36] <bull> is there snapd interface coming ?
[08:36] <ogra_> you can surely find a lightweight editor and seed that into your derfault config .... i wouldnt expect that for a browser
[08:36] <bull> for that too  :/
[08:37] <bull> come-on i don't want change the main aim of my feature
[08:37] <ogra_> ogra@anubis:~/datengrab/images/snappy$ snap interfaces|grep snapd
[08:37] <ogra_> :snapd-control           -
[08:37] <ogra_> that is already there
[08:38] <bull> aim of feature is to make user comfort by providing him an option to edit  file in his favorite editor not with editor of my choce
[08:38] <bull> choice
[08:38] <bull> :P
[08:38] <ogra_> yes, that was clear f4rom your first question
[08:38] <bull> okay
[08:39] <ogra_> isnt there yet ... but there are workarounds i pointed out til this landed
[08:40] <bull> ogra_,  if i will add snapd-control in my snapcraft's plugs will it make it work ?
[08:40] <ogra_> mvo, so http://paste.ubuntu.com/23181304/ seems to work fine as a quick hack, unless you want to invest time into adding an actual option
[08:41] <ogra_> bull, it allows your snap to talk to the REST api of snapd ... so if you implement the right protocol i imagine it will work ...
[08:41]  * ogra_ has no first hand experience with that API
[08:41] <bull> my app uses qprocess to call snap ,
[08:41] <liuxg> ogra_,  I have changed my network setting in my p3 to http://paste.ubuntu.com/23181339/, however, after I disconnect my cable, my wifi on the pi3 still cannot get an IP. what could be the problem for it?
[08:42] <ogra_> liuxg, sit0 isnt your wlan device :)
[08:42] <ogra_> i think it just comes up as wlan0 on the pi
[08:43] <liuxg> ogra_, what should be the wlan device? I could not find any other device there
[08:43] <ogra_> cat /proc/net/dev ?
[08:43] <liuxg> ogra_, yeah, you could be right. thanks. I will check it
[08:50] <bull> snap looking for a web browser to open a folder what ??? QGenericUnixServices::openUrl(const QUrl&): Unable to detect a web browser to launch '/home/bull/github/'
[08:54] <mup> PR snapd#1787 closed: overlord/snapstate, daemon, cmd/snap: more explicit revision support <Blocked> <Created by chipaca> <Merged by chipaca> <https://github.com/snapcore/snapd/pull/1787>
[08:54] <mup> PR snapd#1834 closed: many: mostly work to support ABA upgrades <Created by chipaca> <Merged by chipaca> <https://github.com/snapcore/snapd/pull/1834>
[08:54] <pbek> I have a problem with updating my snap in the snapstore. `snap find qownnotes` and https://uappexplorer.com/app/qownnotes.pbek tells me the latest version is 16.09.3, but my package in my https://myapps.developer.ubuntu.com account is at 16.09.7 and all releases are published in the channel `stable` as they should be.
[08:55] <pbek> All releases were published automatically by Launchpad.
[08:55] <ogra_> pbek, there is an issue with the store ... go to the last releaseed snap in the UI, click unpublish and then click publish again
[08:55] <bull> pbek,  which package you trying ?
[08:55] <ogra_> that will make it show up
[08:55] <ogra_> the fix for this should land today or tomorrow
[08:56] <pbek> ogra_: I get a `504 Gateway Time-out` when I try that :)
[08:56] <pbek> bull: qownnotes
[08:56] <ogra_> uh
[08:57] <pbek> ogra_: but it was unpublished
[08:57] <bull> pbek, does it work out-of-box as a snap ?
[08:58] <pbek> bull: what do you mean by that?
[08:58] <bull> i will check that myself by installing it
[08:59] <bull> ty
[08:59] <pbek> ogra_: thank you, 16.09.7 is now visible
[08:59] <ogra_> :)
[08:59]  * ogra_ has to do the same every morning for all ubuntu-core snaps :)
[08:59] <bull> pbek,  like most of apps dont seem to work fine with after packed in snap :D
[09:00]  * ogra_ found that most apps do work fine ... 
[09:00] <liuxg> ogra_, if I have this error "error: cannot install snap file: snap "piglow-app" has changes in progress". how can I correct this situation?
[09:00] <ogra_> really depends how complex that app is :)
[09:00] <pbek> ogra_: that explains why noone updated QOwnNotes for some time :)
[09:00] <liuxg> ogra_, I have seen this kind of errors a number of times. but sometimes it disappears.
[09:00] <ogra_> liuxg, snap changes ... find the change in progress ... snap abort $change_number
[09:01] <pbek> bull: oh, I had to do quite some work to make it happen... and there are still some issues
[09:01] <bull> ogra_,  a package management system should not worry about how developer made their app :D\
[09:02] <ogra_> every package management system does that
[09:02] <bull> they worry about what functionality developer providing ??
[09:03] <bull> i think no
[09:03] <mvo> pitti: interessting failure of adt on armhf - there is no sshd config there, is armhf driven via something else than ssh in adt?
[09:03] <ogra_> they have to make integration work
[09:03] <liuxg> ogra_, how to find the change number in the system?
[09:03] <bull> they need to install packaged stuffs to place :D
[09:03] <bull> nothing more then this
[09:03] <ogra_> liuxg, ?
[09:03] <pitti> mvo: yes, armhf and s390x run in lxc; we don't have these arches in scalingstack
[09:03] <ogra_> liuxg, does snap changes not show it ?
[09:03] <bull> snap is different case a sure :D
[09:04] <ogra_> bull, how many deb or rpm packages have you built from scratch ?
[09:04] <pitti> + sed -i s/\(PermitRootLogin\|PasswordAuthentication\)\>.*/\1 yes/ /etc/ssh/sshd_config
[09:04] <pitti> o_O
[09:04] <liuxg> ogra_, yeah, I cannot find anything about it. it seems that it is broken http://paste.ubuntu.com/23181393/
[09:04] <ogra_> liuxg, thats snap list
[09:04] <bull> ogra_, i made more then 20 debs
[09:04] <bull> from scratch
[09:04] <liuxg> ogra_, yeah, that is the list. I try to remove it, it refuses to do that.
[09:05] <bull> and they all were in ubuntu software center
[09:05] <ogra_> bull, from scratch ? into the official archive, matching all integration requirements to make then safely work ?
[09:05] <mvo> pitti: ta
[09:05] <ogra_> *them
[09:05] <bull> not in official archives
[09:05] <pbek> bull: just making debs for QOwnNotes drove me nuts :)
[09:05] <ogra_> i.e. following the debian packaging policy ...
[09:05] <liuxg> ogra_, is there anyway to find the changes, and cancel it?
[09:05] <pitti> mvo: the tests want to ssh to localhost via password? 3v1l :)
[09:05] <popey> With the git source-type, is it possible to pull a specific commit? I can use source-tag, and source-branch... but how do I get a commit?
[09:05] <popey> (in snapcraft)
[09:05] <bull> ogra_,  apps in store don't follow debian policies?
[09:06] <ogra_> liuxg, again ... snap changes ... find the change in progress ... snap abort $change_number
[09:06] <liuxg> ogra_, ok. thanks
[09:06] <ogra_> bull, they get a security check ... thats all i think
[09:07] <bull> ogra_,  snap abort change-id for a install process wont abort process until snap download the full package that is a bug
[09:07] <ogra_> thiough support for this was dropped lng ago since the review was way to time consuming
[09:07] <bull> hmm
[09:08] <bull> ogra_,  snap abort change-id for a install process wont abort process until snap download the full package that is a bug !! plz check it
[09:09] <popey> bull: we know, we know about the bug
[09:09] <ogra_> yes, you complained about this for several days, several times a day ...
[09:09] <ogra_> hard to not notice :P
[09:09] <bull> several times a day ??
[09:09] <bull> lol
[09:09] <bull> i was just reminding :P liuxg
[09:10] <bull> popey, yeah we tested it :P
[09:10] <liuxg> bull, yeah, it takes 3 mins to remove a snap.
[09:11] <bull> liuxg, not about removing snap man , it was about to kill a install process . snap wont kill install process untill it download the whole package
[09:11]  * ogra_ sighs and goes for brakfast 
[09:11] <liuxg> liuxg, I am still waiting for the completion of the command. I am sideload a snap. it is not downloading.
[09:12] <bull> liuxg,  use snapcraft-gui
[09:12] <bull> it has package manager to manage snaps
[09:13] <liuxg> bull, snapgraft-gui is a desktop app, right? is it in the store already?
[09:13] <bull> no
[09:13] <bull> it is here https://github.com/snapcraft-gui/snapcraft-gui
[09:13] <bull> repo coming soon
[09:13] <bull> i developed it :P
[09:14] <bull> with all convenience for snap user in mind
[09:14] <bull> you can write recipes and make snaps , with clicks
[09:14] <liuxg> bull, thanks for letting me know. I will try it later on.
[09:15] <bull> okay
[09:16] <bull> ii was having lp ppa but tsimonq2  said he will make new one with better sourcebuild features so we deleted old one
[09:19] <ogra_> liuxg, so what does "snap changes" show you
[09:19] <ogra_> (gimme a pastebin)
[09:25] <bull> QSqlError("", "Driver not loaded", "Driver not loaded")
[09:25] <bull> QSqlQuery::prepare: database not open
[09:27] <bull> ogra_, can we have access to partitions  of hard disk ?? i cant see any such things in the file dialog
[09:28] <ogra_> there is the home interface
[09:28] <bull> home interface show partitions too ?
[09:28] <ogra_> no, only files in home
[09:28] <ogra_> there is no way to get such sensible info yet
[09:29] <ogra_> like partitions ...
[09:29] <bull> wow
[09:29] <bull> thats why i cant load music from mymusic partition using vlc
[09:30] <ogra_> bind mount it under your home somewhere ... i.e. into ~/Music
[09:31] <bull> ogra_,  how :(
[09:31] <ogra_> by bind mounting ?
[09:31] <bull> yea
[09:32] <bull> ogra_,  we have to do it manually before using vlc ?
[09:32] <ogra_> fi you put your music into exotic places you have to make it available in a common place first, yes
[09:33] <bull> :D
[09:33] <ogra_> i have no probs with vlc here
[09:33] <bull> exotic places :P ?
[09:34] <ogra_> (and i think 90% of users wont)
[09:34] <bull> you put music in home /music ?
[09:34] <ogra_> yes
[09:34] <bull> cool
[09:34] <bull> some people do use partitions and external hd
[09:35] <ogra_> good for them
[09:35] <mup> PR snapd#1917 opened: tests: ensure openssh-server is installed in autopkgtest <Created by mvo5> <https://github.com/snapcore/snapd/pull/1917>
[09:35] <bull> not good for them if they dont know how to load music in vlc
[09:35] <bull> people will ditch whole ubuntu for vlc :D
[09:36] <bull> if it will not fixed and snap will be default format :D
[09:36]  * ogra_ goes away for a few hourst ... i really cant take it anymore
[09:36] <bull> *some people
[09:37] <bull> ogra_, LOL
[09:37] <ogra_> not funny, really
[09:37] <bull> yeah i have seen cases
[09:38] <ogra_> if you have something constructive to tell me, fine ... if you find a bug, please file it ... but dont rave about it for hours
[09:44] <bull> hi
[09:44] <bull> tsimonq2, hi
[09:46] <bull> with button icons is it looking sexy > https://raw.githubusercontent.com/snapcraft-gui/snapcraft-gui/master/screenshots/sc1.png
[09:56] <mup> PR snapd#1918 opened: spread.yaml, tests: add adhoc-ubuntu-core spread backend <Created by fgimenez> <https://github.com/snapcore/snapd/pull/1918>
[10:16] <abeato> sergiusens, any idea why I can get this message when creating a snap? : "Copying needed target link from the system /dev/null"
[10:16] <abeato> sergiusens, it created a prime/dev/null file
[10:16] <abeato> which provoked some issues until I noticed
[10:18] <ogra_> mvo, http://paste.ubuntu.com/23181669/ works (not actually more beautiful but i get an extra 500MB when i set WIGGLEROOM=500)
[10:34] <mup> PR snapd#1919 opened: tests: make ubuntu-core tests more robust <Created by mvo5> <https://github.com/snapcore/snapd/pull/1919>
[10:34] <mvo> ogra_: ok
[10:35] <ogra_> mvo, oh, and ouch ... we build the core snap against proposed ...
[10:35] <ogra_> mvo, but -proposed isnt in the sources.list ...
[10:35] <ogra_> i currently can not install vim in classic due to that
[10:35] <ogra_> (in the classic shell i mean)
[10:36] <ogra_> mvo, should i stop using proposed during imagge build or should we add proposed to sources.list ?
[10:39] <mvo> ogra_: I think stop using proposed is best
[10:39] <ogra_> ok
[10:39] <mvo> ogra_: we can also sed it out in classic.create
[10:39] <mvo> (for now)
[10:39] <ogra_> lets see if that builds :)
[10:39] <mvo> but I think !-proposed is best
[10:39]  * ogra_ drops it and does a test build
[10:48] <mvo> pitti: hm, snapd (latest version)vanished from the running page for amd64 but it did not appear on the results page yet. is there a (cron) delay or something?
[10:51] <pitti> mvo: wow, watching it like a hawk, aren't you :) -- there's ~ 5 minutes cron delay indeed
[10:51] <pitti> mvo: still running on http://autopkgtest.ubuntu.com/running#pkg-snapd though on some arches
[10:52] <pitti> mvo: could also be testbed failure, I'll check the logs in a minute
[10:52] <mup> PR snapd#1920 opened: --lazy is only available with -l in trusty <Created by vosst> <https://github.com/snapcore/snapd/pull/1920>
[10:57] <morphis__> pitti: can we configure netplan to generate wifi settings files via NetworkManager and ethernet ones via networkd depending on some flag we put somewhere?
[10:57] <pitti> morphis__: that is the default mode indeed
[10:57] <morphis__> oh really?
[10:57] <pitti> morphis__: oh wait, actually not any more, default is networkd
[10:57] <morphis__> can we change that somehow?
[10:57] <pitti> morphis__: it used to be NM for wifi until I taught it to do wifi via networkd+wpasupplicant
[10:58] <pitti> morphis__: anyway, you can select the backed for every individual definition, yes
[10:58] <pitti> renderer: NetworkManager|networkd, see the manpage
[10:58] <morphis__> pitti: and how can we change that on Ubuntu Core, not sure what configures this globally
[10:58] <morphis__> or would that be a thing console-conf has to know about?
[10:58] <ogra_> but please dont make NM a hard requirement
[10:58] <pitti> you can specify the renderer globally, by device type, or by individual definition
[10:59] <pitti> and you can put it into a separate yaml snippet
[10:59] <ogra_> if it isnt installed on an image it netpplan needs to fall back to using wpa-supplicant
[10:59] <pitti> it normally detects the availability
[10:59] <ogra_> ok
[10:59] <morphis__> ogra_: no :-)
[10:59] <ogra_> :P
[10:59] <pitti> but if you say "use NM" and NM is not installed, those definitions wil just fail
[10:59] <morphis__> pitti: I see
[10:59] <ogra_> right ...
[11:00] <morphis__> pitti: where do I specify the renderer globally?
[11:00] <pitti> morphis__: at the global yaml level, right under the top-level "network:" (again, see manpage)
[11:01] <ogra_> pitti, btw, my network setup on the pi3 now goes really mad ... i'm not yet sure if thats netplan or console-conf that is at fault ... console-conf offers me both interfaces (wired and wireless) but i always end up with only wired usable
[11:02] <ogra_> it works just fine on the dragonboard where i only have a wlan interface ... but having both makes something go mad
[11:02] <morphis__> pitti: thanks
[11:03] <mwhudson> ogra_: there is https://bugs.launchpad.net/ubuntu/+source/subiquity/+bug/1623756 in this area
[11:03] <mup> Bug #1623756: console-conf generates invalid config when wifi is not configured <subiquity (Ubuntu):In Progress> <https://launchpad.net/bugs/1623756>
[11:03] <ogra_> mwhudson, oh, thanks !
[11:03] <ogra_> yeah, i wanted to file that one but collect some more data ... just got distracted from it
[11:04] <mwhudson> just sent a PR with a presumed fix, really need to get the repo made public now...
[11:04] <ogra_> mwhudson, though mine is slightly different ... i actually configure it with ssid and it even obtains an IP
[11:04] <mwhudson> ogra_: ah ok
[11:05] <ogra_> it still tries to use wired though ... evven if i say "dont configure"
[11:05]  * ogra_ will collect more data first ... perhaps it is the same cause, who knows
[11:05] <mwhudson> ah hm
[11:05] <mwhudson> yeah, that's certainly possible
[11:06] <mwhudson> although if the wired interface is up, you remove it from the config entirely and run netplan apply, i don't know if the interface would be taken down
[11:06] <mwhudson> pitti: would know
[11:07] <ogra_> well, if there is no cable, it should not be tried at all
[11:07] <pitti> no, apply doesn't currently down interfaces which are up
[11:07] <mwhudson> oh
[11:07] <ogra_> not sure if netplan takes mii state into account here
[11:07] <pitti> seemed too dangerous to me
[11:07] <pitti> if we want/need that, bug report please
[11:08] <pitti> ogra_: you mean up, but no link beat?
[11:08] <mwhudson> ogra_: how can you tell that it's being tried? sorry it's late here and i'm probably being dense
[11:08] <pitti> I guess we can safely twiddle those too
[11:09] <pitti> "netplan --debug apply" might be useful
[11:09] <bull> Snapcraft GUI version 2.0 released with everything working :P download .deb here https://github.com/snapcraft-gui/snapcraft-gui/releases/tag/2.0
[11:09] <ogra_> pitti, i mean when i boot i get offered to configure eth0 in console-conf even if there is no cable plugged in ... it shouldnt offer that i think
[11:09] <pitti> hmm
[11:09] <pitti> ogra_: there are routers with broken link beat
[11:09] <pitti> or wifi devices have no link until you connect to an AP
[11:10] <ogra_> hmm, yeah
[11:10] <pitti> so I'm not sure if that's so obvious
[11:10] <pitti> it coudl certainly tell you that
[11:10] <pitti> en1 (disconnected)
[11:10] <ogra_> yeah, i guess something like that in the UI would already help
[11:10] <mup> PR snapd#1921 opened: Replace systemd-run with on-the-fly generation of units <Created by vosst> <https://github.com/snapcore/snapd/pull/1921>
[11:15] <mwhudson> ogra_: sounds good, bug pls
[11:16] <ogra_> yay ...
[11:16]  * ogra_ can install vim again
[11:20] <mup> Bug #1623897 opened: console-conf should show if wired interfaces are connected <Snappy:New> <subiquity (Ubuntu):New> <https://launchpad.net/bugs/1623897>
[11:20] <ogra_> mwhudson, ^^
[11:24] <mwhudson> ogra_: thanks
[11:24]  * mwhudson zzz
[11:25] <ogra_> sleep well
[11:25] <mup> PR snapd#1922 opened: Use apt in accordance with functionality available on trusty <Created by vosst> <https://github.com/snapcore/snapd/pull/1922>
[11:33] <liuxg> is snapweb currently working already?
[11:36] <mup> PR snapd#1923 opened: Replace realpath with readlink -f for trusty support <Created by vosst> <https://github.com/snapcore/snapd/pull/1923>
[11:47] <bull> liuxg, checkout https://github.com/snapcraft-gui/snapcraft-gui/releases/tag/2.0
[11:48] <liuxg> bull, it looks very nice. I will promote it in my place to the community :)
[11:48] <bull> ty
[11:49] <bull> liuxg, please create issue if you want any changes , and finds any bug
[11:49] <liuxg> bull sure, I will do that. it is actually my strong point :)
[11:50] <liuxg> bull, in fact, I had thought of doing this some time ago :) it is really a nice work.
[11:50] <bull> thanks :) i hope you have found the project page already
[11:50] <bull> hmm
[12:03] <mup> PR snapd#1917 closed: tests: ensure openssh-server is installed in autopkgtest <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1917>
[12:03] <mup> PR snapd#1919 closed: tests: make ubuntu-core tests more robust <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1919>
[12:04] <mup> PR snapd#1915 closed: tests: add http_proxy to /etc/environment in the autopkgtest environment <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1915>
[12:04] <mup> PR snapd#1916 closed: backends: first bits of kernel-module security backend <Created by stolowski> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1916>
[12:08] <mup> PR snapd#1924 opened: tests: disable prepare-image-grub test in autopkgtest <Created by mvo5> <https://github.com/snapcore/snapd/pull/1924>
[12:10] <mvo> mwhudson: do you happen to know what is the deal with govendor(ing) and gccgo? it seems to be not supported in 16.04 or am I just using it wrong?
[12:11] <liuxg> may I know how to stop auto-update in the background? previously I could use config to disable it.
[12:11] <pitti> mvo: you have a few i386 and amd64 logs now
[12:12] <mvo> mwhudson: aha, I see you commented already, thank you!
[12:13] <mvo> pitti: ha! close
[12:13] <mvo> pitti: still one failure
[12:13] <mup> PR snapd#1925 opened: Adjust regex to account for changes in stat output <Created by vosst> <https://github.com/snapcore/snapd/pull/1925>
[12:14] <mup> PR snapcraft#803 closed: Do not run the bootstrap directory as a script (autotools plugin) <Created by evandandrea> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/803>
[12:14] <mup> PR snapd#1924 closed: tests: disable prepare-image-grub test in autopkgtest <Created by mvo5> <Closed by mvo5> <https://github.com/snapcore/snapd/pull/1924>
[12:15] <mvo> pitti: hm, i386 is strange it says ppa:snappy-dev/image can not be added (it does not exist). this worked on the amd64 run
[12:16] <sergiusens> abeato oh, neat! ldd for one of your elf files has that.
[12:16] <sergiusens> abeato care to fgure out which one?
[12:16] <abeato> sergiusens, lol
[12:16] <mup> PR snapd#1926 opened: tests: add https_proxy into environment as well <Created by mvo5> <https://github.com/snapcore/snapd/pull/1926>
[12:17] <abeato> sergiusens, yes, let me look. Interestingly that happens only when building for ARM, but not for x86
[12:18] <sergiusens> abeato if you can, find the offending lib and give me the ldd output for it
[12:18] <abeato> ack
[12:18] <abeato> sergiusens, anohter question: what does
[12:18] <abeato> Issues while validating snapcraft.yaml: The '0' property does not match the required schema: ['usr/bin/Xorg'] is not of type 'string'
[12:18] <abeato> mean?
[12:19] <pitti> mvo: hm, that add-apt-repo call works in a i386 chroot
[12:23] <sergiusens> abeato it means, remove the brackets :-)
[12:23] <sergiusens> abeato is this `organize`?
[12:24] <abeato> sergiusens, no brackets
[12:26] <abeato> sergiusens, kind of, it works now after removing this line: "# organize: ??
[12:26] <abeato> "# organize: ??"
[12:26] <abeato> which I added to remind me to investigate something
[12:27] <mup> PR snapd#1927 opened: Account for different PWD handling on trusty <Created by vosst> <https://github.com/snapcore/snapd/pull/1927>
[12:28] <sergiusens> abeato let me get you an example
[12:28] <abeato> sergiusens, a parsing error??
[12:28] <sergiusens> abeato https://github.com/snapcore/snapcraft/blob/master/demos/mosquitto/snapcraft.yaml#L28
[12:29] <sergiusens> abeato you are not matching the schema;; we use jsonschema that by default gives us those horrible errors and no easy way to compose our own
[12:29] <abeato> sergiusens, hmm... but why does it complain for a line that is a comment?
[12:32] <sergiusens> abeato oh, I see; give me the full yaml
[12:33] <sergiusens> abeato the lines are off by one btw
[12:40] <bull> sergiusens, can https://github.com/snapcraft-gui/snapcraft-gui/ be part of snapcraft source-code ?? please read this issue here https://github.com/snapcraft-gui/snapcraft-gui/issues/4
[12:41] <abeato> sergiusens, off by one? some whitespace issue?
[12:56] <abeato> sergiusens, I have found no references to something that migth pull /dev/null in the binaries... what kind of thing do you expect me to see?
[13:19] <abeato> sergiusens, and why can this happen:
[13:19] <abeato> Failed to clean step 'prime': Missing necessary state. This won't work until a complete clean has occurred.
[13:19] <abeato> ??
[13:20] <liuxg> ogra_, does Pi3 has a built-in bluetooth in it? thanks. I just try to make use of it, however, I cannot find it.
[13:26] <liuxg> morphis__,  is bluetooth currently supported on pi3?
[13:28] <ogra_> liuxg, pi3 has builtin BT ... but i dont know if we have all firmware and drivers yet
[13:29] <ogra_> theoretically it should be the wlan firmware since i think it is the same device
[13:29] <ogra_> practically i dont know though
[13:29] <liuxg> ogra_, ok. thanks for your reply. I basically duplicate the same step as 410c, but it does not work. I have tried to use bluetoothctrl, but it seems that it is not working.
[13:29] <bull> my bt dont work on ubuntu 16.04 :(  it was working on 14.04
[13:30] <ogra_> liuxg, i'm not really sure if the kernel has all bits and pieces yet for the pi3 img
[13:30] <liuxg> ogra_, yes, I see the wlan there http://paste.ubuntu.com/23182127/
[13:30] <ogra_> yeah, that definiteely works
[13:31] <liuxg> ogra_, currently, the bluetooth works on 410c board.
[13:31] <abeato> ogra, an image on beta channel for rpi3 should work?
[13:31] <ogra_> yup, i noticed that you got it working yesterday
[13:31] <ogra_> abeato, yes
[13:31] <abeato> ogra_, ack
[13:32] <ogra_> we'll release new ones later today or tomorrow on cdimage
[13:32] <ogra_> http://people.canonical.com/~ogra/snappy/all-snaps/ has an interim one i just rolled for someone else
[13:32] <ogra_> in case you need something to dd right now
[13:32] <abeato> ogra_, thanks, will try to build one with u-d-f nonetheless
[13:33] <ogra_> u-d-f is unsupported now
[13:33] <ogra_> that will most likely not produce anything usable anymore
[13:33] <abeato> what do we use now?
[13:33] <ogra_> ubuntu-image ... and you need signed assertions
[13:33] <ogra_> just grab mine from the above url
[13:34] <abeato> yeah, that will be easier...
[13:35] <bull> ogra_,  http://people.canonical.com/~ogra/2CAF65BCd01.jpg :D
[13:43]  * zyga has a theory about what might be broken with ns sharing
[13:46] <bull> bye friends :*
[13:47] <morphis__> ogra_: why is HDMI output disabled on the official pi2 image?
[13:47] <ogra_> morphis__, ?
[13:47] <ogra_> works here
[13:47] <morphis__> ogra_: I had to change the config file on the boot partition to get it
[13:47] <morphis__> hdmi_safe=1
[13:48] <ogra_> weird, i tested console-conf and login onn various monitors here
[13:48] <ogra_> (well, two monitors and a TV)
[13:49] <ogra_> hdmi_safe should only be needed if your monitor has issues with the spec i heard
[13:50] <ogra_> "hdmi_safe Use "safe mode" settings to try to boot with maximum hdmi compatibility. This is the same as the combination of: hdmi_force_hotplug=1, hdmi_ignore_edid=0xa5000080, config_hdmi_boost=4, hdmi_group=2, hdmi_mode=4, disable_overscan=0, overscan_left=24, overscan_right=24, overscan_top=24, overscan_bottom=24 "
[13:50] <ogra_> haha
[13:50] <ogra_> "hdmi_ignore_edid Enables the ignoring of EDID/display data if your display is a crappy Chinese one "
[13:51] <bull> ogra_,  bye
[13:51] <bull> :Zzz
[13:51] <ogra_> ciao
[13:52] <ogra_> morphis__, so whats your monitor ... ?
[13:54] <morphis__> ogra_: hm, it was working well before from what I remember
[13:54] <ogra_> before what ?
[13:55] <ogra_> i dot thinnk we have changed anything in config.txt since 6 months
[13:55] <morphis__> ok
[13:55] <morphis__> then its a local problem here
[13:55] <ogra_> IIRC liuxg had probs as well and needed to use another HDMI port on his monitor, there it then jjust worked
[13:57] <ogra_> but perhaps we should actually default to hdmi_safe to get the widest compatibility ... i'm just a bit reluctant to add display options at all since these are dev boards ... mainly used for embedded development
[14:10]  * zyga -> coffee and food
[14:15] <zyga> jdstrand: ping me when you are around please
[14:16] <mectors> I am trying to upgrade a node-red snap to U16.04. The snap installs without problems. When I execute the snap/bin/nodered.red everything works fine but no .service is created. I am trying devmode. Don't see an error in syslog. What should I do?
[14:18] <zyga> mectors: look at the apps declaration, does it define this is a deamon?
[14:19] <mectors> No was a command. Will change to daemon
[14:19] <qengho> mectors: You're not making that a service. You're making it an app that you can run.
[14:20] <qengho> mectors: "daemon: OneOfTheDaemonTypes" underneath "command:".
[14:20] <mectors> thanks
[14:20] <zyga> jdstrand: http://pastebin.ubuntu.com/23182286/
[14:21] <zyga> jdstrand: aka WTF?
[14:21] <zyga> jdstrand: NOTE: ip in the core snap is not the one in classic
[14:21] <zyga> jdstrand: not sure if this matters in what you did earlier
[14:21] <zyga> jdstrand: I was also thinking if the way I pivot_root is buggy, I tweaked that locally but it has no observable effect
[14:22] <zyga> jdstrand: what bugs me is that I *can* open ns/net but ip cannot
[14:22] <mectors> fixed it. many thanks
[14:22] <zyga> (ip the command)
[14:28] <mup> PR snapd#1928 opened: hookstate,daemon: don't mock HookRunner, mock command <Created by kyrofa> <https://github.com/snapcore/snapd/pull/1928>
[14:29] <jdstrand> zyga: it is very puzzling. I mean, it worked at one point!! I downgraded kernels (even to 4.3-- pre-trusty), used a fresh install of xenial desktop without upgrading, built an image using udf with os, kernel and gadget snaps from June, etc, etc
[14:30] <jdstrand> zyga: I even saw on the old xenial desktop that a new core snap was downloaded when installing hello-world so I replaced current symlink to use the old one. I think I have a few more things to play with there
[14:31] <zyga> jdstrand: maybe the core snap updated?
[14:31] <zyga> jdstrand: and old VMs may still have reeexec anbled
[14:31] <zyga> jdstrand: as a quick exercise, I can run unshare --net=foo
[14:32] <zyga> jdstrand: I can open /proc/self/ns/net from python
[14:32] <zyga> jdstrand: I'm utterly lost as to what this might be
[14:33] <zyga> jdstrand: quick question about pivot_root, does this change make sense: http://paste.ubuntu.com/23182323/
[14:33] <zyga> jdstrand: I read what pivot_root does again and I suspect that either docs are wrong or pivot_root . . should not be used
[14:33] <zyga> (this disabled nvidia binding because it's done in a way that needs the hostfs thing to be in place but I can move that to always happen after pivot root)
[14:37] <jdstrand> zyga: I think tyhicks may be able to better answer that, or possibly an lxd dev
[14:38] <jdstrand> (the pivot_root question)
[14:38] <jdstrand> I can do it, but it'll take me a while to say for sure
[14:39] <jdstrand> zyga: re the core snap updated-- it definitely did, which is why I tried to undo that. It's possible I didn't do that right and have a few more things to try there
[14:39] <zyga> jdstrand: don't worry about pivot root; I'll keep digging
[14:39] <tyhicks> zyga: what are you trying to achieve there?
[14:41] <jdstrand> zyga: regarding the permission denied-- it is seriously puzzling
[14:41] <zyga> tyhicks: use pivot_root correctly
[14:41] <zyga> tyhicks: perhaps avoid one more bind mount
[14:41] <zyga> tyhicks: ignore pivot_root unless you think it affects the big ns issue
[14:41] <zyga> tyhicks: I don't know if you are in the loop
[14:42] <zyga> tyhicks: if you want we can sync quickly
[14:42] <jdstrand> I need to look at another PR and will circle back
[14:42] <zyga> ok
[14:42] <tyhicks> I'm probably not up on what you've been up to the last day or two
[14:43] <tyhicks> the lxd devs will definitely have more real world experience w/ pivot_root than I do
[14:44] <zyga> tyhicks: long story short: the kernel behaves in super odd ways wrt net namespace; with snap-confine built from pull request https://github.com/snapcore/snap-confine/pull/145 "ip netns" gets permission denied to open /proc/$pid/ns/net (as shown by strace on http://pastebin.ubuntu.com/23182286/) but nothing apparently prevents this if you just try to open it directy (see the pastebin)
[14:44] <mup> PR snap-confine#145: Enable snap-confine namespace sharing <Created by zyga> <https://github.com/snapcore/snap-confine/pull/145>
[14:44] <zyga> this is all with devmode
[14:46] <zyga> tyhicks: and in that pastebin I also don't understand why this fails:
[14:46] <zyga> fstat(4, 0x7fff7fe39650)                = -1 EACCES (Permission denied)
[14:47] <zyga> 4 is stderr after dup()
[14:48] <zyga> might that be some kernel memory corruption?
[14:50] <tyhicks> I don't know much about network namespaces but that strace sure does look odd
[14:50] <tyhicks> let me think about it a little more
[15:00] <mup> PR snapd#1929 opened: Return proper icon URL for non-installed snaps <Created by jamiedbennett> <https://github.com/snapcore/snapd/pull/1929>
[15:06] <tyhicks> zyga: I don't have any good ideas about what's going on there
[15:06] <tyhicks> zyga: I'd have to spend quite a bit of time reading kernel code
[15:07] <zyga> tyhicks: ok, thanks
[15:24] <jdstrand> zyga: ok, I've circled back
[15:25] <jdstrand> I'll keep plugging away at trying to see when this once worked and try other things that I think of along the way
[15:27] <jdstrand> zyga: that said ^, there is perhaps something else to consider. the fact that this isn't working is odd to be sure, but this test (ip netns ...) working within a snap may not be as important if we think about the long term goal: a network-namespace-support interface that allows snaps to add network namespaces to the system and for other snaps to use
[15:28] <zyga> jdstrand: hmm
[15:28] <zyga> jdstrand: for that we'd have to solve the bi-directional mounting in general
[15:28] <jdstrand> zyga: in other words, since we know that network-namespace-support is the desired goal, if we can figure out how to achieve that without fixing this oddity, maybe that's ok
[15:28] <zyga> jdstrand: or do you plan snapd to create the namepaces centrally?
[15:29] <jdstrand> that's up for discussion
[15:29] <zyga> jdstrand: ok, so what shall we do
[15:29] <jdstrand> it seems clear that people will want to use 'ip netns' to setup namespaces. lots of existing code does that
[15:29] <jdstrand> our 'ip' command could be modified to talk to snapd
[15:30] <zyga> jdstrand: I'm inclined to merge this, add more spread tests
[15:30] <zyga> jdstrand: for both working and OMG/WTF broken cases
[15:30] <zyga> jdstrand: and decide if we want to enable this for release or not
[15:40] <mup> PR snapd#1924 opened: tests: disable prepare-image-grub test in autopkgtest <Created by mvo5> <https://github.com/snapcore/snapd/pull/1924>
[16:00] <mectors> just uploaded nodered.mectors. Says it is published but can't see it in the uApp Explorer or install it via snap install.
[16:01] <kyrofa> mectors, did you publish it in a specific channel?
[16:03] <mectors> no just did snapcraft push
[16:04] <kyrofa> mectors, are you familiar with channels?
[16:04] <mectors> kyrofa pushing it to stable now
[16:04] <kyrofa> mectors, ah very good, okay
[16:05] <kyrofa> mectors, so yeah, you can snapcraft push to specific channels, or snapcraft release after the fact, or do it all in from the myapps site
[16:05] <kyrofa> s/all in/all/
[16:05] <kyrofa> But snapcraft push by default doesn't put it in any channels
[16:07] <mectors> kyrofa I updated the channel to stable on myapps but nothing. it says it passed all 40 checks there
[16:08] <mectors> kyrofa sorry my mistake was trying nodered.mectors but nodered works
[16:08] <kyrofa> mectors, good deal!
[16:09] <kyrofa> ev, I've been noticing numerous misunderstandings stemming from the fact that the store shows <snapname>.<publisher> in the header. Is there a reason for leaving it that way?
[16:10] <ev> kyrofa: no, there isn’t and as I’ve come across that myself I’ve considered it part of the task to switch from “short namespaces” to SSO usernames, but they’re orthogonal.
[16:16] <shuduo> matteo: hi, so you can see embedded wifi interface with your image/kernel, right? do you see it in console-conf phase?
[16:30] <zyga> jdstrand: it is quite easy to crash the kernel with various mount calls
[16:30]  * zyga hugs VMs
[16:40] <sergiusens> mectors next time, you can do `snapcraft release <snap-name> <revno-from-store> <channel>`
[16:41] <matteo> shuduo: hi
[16:41] <sergiusens> kyrofa hey, you don't push to channels, you release to channels ;-)
[16:41] <matteo> shuduo: you mean the web interface? yes
[16:42] <shuduo> matteo: do you use pi2-kernel?
[16:44] <matteo> let me check
[16:44] <matteo> pi2-kernel   4.4.0-1021-3  14   canonical  -
[16:45] <zyga> matteo: hey :)
[16:45] <mup> PR snapd#1929 closed: Return proper icon URL for non-installed snaps <Created by jamiedbennett> <Merged by chipaca> <https://github.com/snapcore/snapd/pull/1929>
[16:45] <matteo> hi zyga
[16:45] <sergiusens> wow cjwatson, thanks for taking that macaroon task!
[16:47] <cjwatson> sergiusens: np, it happened to hit me and I already have code in LP that does much the same thing
[16:48] <shuduo> ogra_: hi, i'm following the doc "Ubuntu Core Image building" and generated an image for pi3. Now I see some issues like:
[16:49] <shuduo> 1, it booting up time is very long if I plug ethernet cable. if no ethernet cable, it can boot up to console-conf soon. Is it a known issue?
[16:50] <ogra_> nope, definitely not known
[16:50] <shuduo> 2, I can't see integrated wireless interface, is it a known issue? I use pi2-kernel.
[16:50] <matteo> shuduo: do you want my ubuntu core image and openwrt snap?
[16:50] <ogra_> yes, the wifi interface is only seen in very recent console-conf versions ... in the edge channel
[16:51] <ogra_> but it doesnt not work right yet
[16:51] <matteo> I'm using ethernet to access internet, and the wifi as AP
[16:51] <ogra_> shuduo, http://people.canonical.com/~ogra/snappy/all-snaps/u-image-pi3.img.xz
[16:51] <shuduo> matteo: yes, pls.
[16:51] <shuduo> ogra_: okay, let me try your image too.
[16:52] <matteo> shuduo: I'm pushing everything here: https://people.canonical.com/~teknoraver/
[16:52] <ogra_> shuduo, we'll release an image like that this week
[16:53] <matteo> zyga: why this test fails, in your opinion= http://162.213.35.179:8080/job/github-snapcraft-autopkgtest-yakkety-snaps-cloud/5/console
[16:53] <shuduo> ogra_: good to know. we want to demo openwrt snap with pi3 at end of Sept. it would be great if later image do everything well. :)
[16:54] <ogra_> well, there are still issues with wlan vs eth on the pi3 ... but these should surely be fixed before end of the month :)
[16:54] <zyga> matteo: looking
[16:55] <zyga> ah, vpn
[17:03] <matteo> ping mvo
[17:07] <mvo> matteo: pong (with some latency, but I will read/reply eventually)
[17:08] <matteo> I have an unit test which fails in an odd way
[17:08] <matteo> maybe it's not related to my commit
[17:08] <matteo> so you may want to have a look
[17:08] <matteo> mvo: http://162.213.35.179:8080/job/github-snapcraft-autopkgtest-yakkety-snaps-cloud/5/console
[17:10] <mvo> matteo: fix in git should actually be fixed in yakkety already
[17:10] <matteo> so if I trigger a new test it will pass?
[17:23] <kyrofa> sergiusens, I mean the push command supports releasing as well, no?
[17:24] <sergiusens> kyrofa yes it does, `snapcraft push <snap-file> --release <channel>`
[17:24] <sergiusens> kyrofa that's release on push, not push to a channel though ;-)
[17:24] <kyrofa> sergiusens, meh :P
[18:04] <sergiusens> stgraber any way to figure out what is going on here? http://paste.ubuntu.com/23183087/
[18:06] <stgraber> sergiusens: hmm, well, LXD forks tar and tar got sigkilled, not sure why
[18:06] <stgraber> sergiusens: out of memory maybe?
[18:06] <stgraber> sergiusens: anything in dmesg?
[18:07] <sergiusens> stgraber yeah, out of mem; strange
[18:08] <sergiusens> I am on a fresh boot
[18:08] <stgraber> oh, actually, that's unsquashfs not tar for that image I guess
[18:09] <sergiusens> stgraber yup
[18:09] <sergiusens> [  598.420356] Out of memory: Kill process 8056 (unsquashfs) score 506 or sacrifice child [  598.420664] Killed process 8056 (unsquashfs) total-vm:645224kB, anon-rss:260812kB, file-rss:0kB
[18:09] <stgraber> how much memory do you have on that system?
[18:10] <stgraber> unsquashfs using 260MB of RAM while not light doesn't seem entirely unreasonable to me
[18:11] <sergiusens> stgraber ah, I have 512
[18:12] <stgraber> how many CPUs in /proc/cpuinfo?
[18:13] <sergiusens> stgraber just the one http://paste.ubuntu.com/23183119/ its a DO droplet
[18:13] <stgraber> weird, not sure why unsquashfs would use more memory than the size of the whole image
[18:14] <mup> PR snapd#1928 closed: hookstate,daemon: don't mock HookRunner, mock command <Created by kyrofa> <Merged by kyrofa> <https://github.com/snapcore/snapd/pull/1928>
[18:18] <stgraber> sergiusens: I'm testing some unsquashfs options to work on low memory systems
[18:19] <stgraber> sergiusens: running on a 64MB system now with 48 cores, so kinda worst case scenario :)
[18:25] <sergiusens> stgraber yeah, this used to work before, I guess before the move to squash, more so, I have 3 containers here running fine
[18:27] <stgraber> sergiusens: hmm, so I managed to launch a container using a squashfs image just fine on that crazy 64MB system...
[18:27] <stgraber> sergiusens: unsquashfs memory use was showing at 40MB or so
[18:28]  * stgraber pokes things harder
[18:36] <mup> PR snapd#1914 closed: docs: add a little documentation on hooks <Created by kyrofa> <Merged by kyrofa> <https://github.com/snapcore/snapd/pull/1914>
[18:39] <mup> PR snapd#1888 closed: interfaces: allow special casing for auto-connect until we have assertions <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1888>
[18:40] <mup> PR snapd#1924 closed: tests: disable prepare-image-grub test in autopkgtest <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/1924>
[18:44] <kgunn> kyrofa: hey there, so i'm seeing something really interesting...
[18:44] <kyrofa> kgunn, yeah what's up?
[18:45] <kgunn> we've got a snap we're building that's pretty big admittedly...unity8...and it pulls in qtmir
[18:45] <kgunn> so the qtmir pkgs are there in the snap on amd64
[18:45] <kgunn> however, on arm64
[18:45] <kgunn> they are missing....
[18:45] <kgunn> this is true for a snap that was built with lp builders and then one built in a cross build chroot locallly
[18:46] <kgunn> and...to boot...the one built locally
[18:46] <kgunn> the files are in stage and prime!
[18:46] <kgunn> but not in the snap
[18:46] <kgunn> ....oh, and i just added qtmir pkgs specifically to part as stage (using nil plugin)
[18:46] <kgunn> and they still didn't make the snap
[18:46] <kgunn> any ideas?
[18:46] <kyrofa> Hmm
[18:46] <kgunn> AlbertA: ^
[18:47] <kyrofa> If files are in prime, they should 100% definitely be in the snap
[18:47] <kyrofa> Because making the snap is literally running mksquashfs on the prime dir
[18:47] <kyrofa> kgunn, so something is weird here
[18:47] <kgunn> kyrofa: don't suppose you know if you can verify what's in the prime from launchpad ?
[18:48] <kyrofa> kgunn, unsquashfs the resulting snap. I can't think of a single way the prime dir could be different
[18:48] <kgunn> AlbertA: wanna shoot me link for arm64 version one more time from the chroot that definitely had the qtmir stuff in prime?
[18:48] <AlbertA> kgunn: yeah it's
[18:48] <kgunn> just to make sure
[18:49] <AlbertA> https://private-fileshare.canonical.com/~alberto/unity8-session_0_arm64.snap
[19:02] <kyrofa> kgunn, when you say qtmir, are you talking about qtdeclarative5-qtmir-plugin? Or qtmir-desktop?
[19:04] <kyrofa> kgunn, doesn't matter-- both of them are in there
[19:04] <kgunn> right
[19:04] <kyrofa> kgunn, so am I missing something?
[19:05] <kgunn> kyrofa: one sec....got distracted lemme install and check
[19:13] <kgunn> kyrofa: so upon install of that same snap on my dragonboard i get
[19:13] <kgunn> ubuntu@localhost:/snap/unity8-session/current$ sudo find -iname qtdeclarative5-qtmir-plugin
[19:13] <kgunn> ./usr/share/doc/qtdeclarative5-qtmir-plugin
[19:13] <kgunn> ubuntu@localhost:/snap/unity8-session/current$
[19:13] <kyrofa> kgunn, that makes sense, no?
[19:14] <kyrofa> kgunn, find -iname libunityscreensplugin.so should find the lib contained within that package as well
[19:14] <kgunn> kyrofa: sorry i meant this one
[19:14] <kgunn> ubuntu@localhost:/snap/unity8-session/current$ sudo find libqpa-mirserver*
[19:14] <kgunn> find: ‘libqpa-mirserver*’: No such file or directory
[19:15] <kgunn> ubuntu@localhost:/snap/unity8-session/current$
[19:15] <kgunn> kyrofa: ah...yeah, i see it now
[19:16] <kgunn> ubuntu@localhost:/snap/unity8-session/current$ sudo find -iname libqpa*
[19:16] <kgunn> ./usr/lib/aarch64-linux-gnu/qt5/plugins/platforms/libqpa-mirserver.so
[19:16] <kgunn> ./usr/lib/aarch64-linux-gnu/qt5/plugins/platforms/libqpa-ubuntumirclient.so
[19:16] <kgunn> the damn "minus" sign i forgot find get's angry
[19:17] <kyrofa> Hahaha
[19:18] <kyrofa> So does that all look okay then?
[19:19] <kgunn> kyrofa: yes....now i probably made a stink over nothing
[19:20] <kgunn> i bet the lp built one is fine
[19:20] <kgunn> kyrofa: fwiw, we did really have one at one point where it wasn't included....so it got us all amped
[19:20] <kgunn> i think it was a fluke
[19:20] <kgunn> ...and we verified the pkgs weren't there in the sstage/prime area...
[19:20] <kyrofa> kgunn, heh, well if you run into it again, let me know!
[19:21] <kgunn> yeah, see something once like that, makes you all paranoid...
[20:03] <stgraber> sergiusens: https://github.com/lxc/lxd/pull/2383
[20:03] <mup> PR lxc/lxd#2383: Tweak squashfs for low-memory systems <Created by stgraber> <https://github.com/lxc/lxd/pull/2383>
[20:03] <stgraber> sergiusens: with that change in place, I could get LXD to mostly work on a system with just 64MB of RAM
[20:03] <stgraber> sergiusens: mostly in the sense that container creation worked fine, but as soon as the container started, the OOM killer started killing everything :)
[20:04] <stgraber> sergiusens: though that was when running a full Ubuntu in the container, something like busybox would have run without triggering OOM kills I expect
[20:12] <sergiusens> stgraber great, in the meantime I upped the mem for my droplet :-)
[20:13] <stgraber> sergiusens: ok, you could have put a simple wrapper around unsquashfs in /usr/local/bin setting the same options I set in that patch
[20:14] <sergiusens> stgraber I will try your branch later tonight and get back to you to see how it went
[20:14] <stgraber> sergiusens: ok. it should be fine since I got it to run on a system with just 64MB :)
[20:18] <mup> PR snapd#1930 opened: many: support snapctl -h <Created by kyrofa> <https://github.com/snapcore/snapd/pull/1930>
[20:41] <chiluk> hey guys I'm still trying to learn snappy, and I'm trying ot snap up the file command, because I thought it would be simple
[20:42] <chiluk> this is what I have so far http://paste.ubuntu.com/23183657/
[20:42] <chiluk> but whenever I go to run file after installing my snap I get /etc/magic, 0: Warning: using regular magic file `/usr/share/misc/magic'
[20:42] <chiluk> file: could not find any valid magic files!
[21:03] <chiluk> I tried to explicitly copy the configs per https://github.com/owncloud/ubuntu-snap/wiki/Building-your-first-snap and got http://paste.ubuntu.com/23183708/   but I get the same results.
[21:04] <chiluk> what am I missing here?
[21:04] <chiluk> I guess I'll send mail to the mailing list.
[21:20] <mup> PR snapcraft#804 opened: Adjust login to conform to UX specification <Created by cjwatson> <https://github.com/snapcore/snapcraft/pull/804>
[21:59] <mwhudson> ogra_: https://github.com/CanonicalLtd/subiquity is public now fwiw
[22:01] <jdstrand> morphis: fyi, I approved udisks2 but you need to press the publish button
[22:37] <ahoneybun> was there a change in snapd that you can install a snap that is not signed?
[22:37] <ahoneybun> *can't
[22:37] <lpotter> ahoneybun: yes there was. add --force-dangerous
[22:37] <ahoneybun> thanks lpotter
[22:38] <ahoneybun> tried to install a inkscape snap from LP and it failed because of a missing signature
[22:41] <ahoneybun> is it "snap install --force-dangerous?
[22:43] <lpotter> yes
[23:05] <sergiusens> stgraber that did the trick btw
[23:43] <stgraber> sergiusens: cool