[07:47] <dholbach> good morning
[08:07] <fgimenez> good morning
[09:08] <fgimenez> hi mvo good morning
[09:10] <mvo> hey fgimenez, good morning!
[09:12] <fgimenez> mvo, hi :) i'm reviewing the results of the latest integration suite executions and it seems that with image 270 we have a new error http://10.55.32.74:8080/job/github-snappy-integration-tests-cloud/25/console
[09:13] <fgimenez> mvo, you can search for "FAIL:" to see the details
[09:14] <fgimenez> mvo, it seems to fail while trying to write a file in $ADT_ARTIFACTS, i'll try to investigate further, are there any changes that might affect this?
[09:26] <mvo> fgimenez: the error does not ring a bell :/
[09:26] <mvo> ogra_: armhf build for rolling is back and worked, so classic mode for rpi2!
[09:27] <fgimenez> mvo, np, i'll try to get more info about it and ping you back
[09:30] <mvo> thanks
[09:44] <sergiusens> dholbach, how hard would it be to pull documentation from a 1.x branch on github for snapcraft?
[09:53] <fgimenez> mvo, it seems that ioutil.WriteFile is failing to create a file in /tmp/adt-run.pL7nNw/command1-artifacts/version, creating it in /tmp/version instead works
[09:54] <fgimenez> mvo, but now i'm getting http://paste.ubuntu.com/13619923/
[09:54] <fgimenez> mvo, "lstat /tmp/adt-run.59r10B/tree/integration-tests/data/snaps/basic-binaries: no such file or directory", but this directory exists for sure
[09:57] <fgimenez> mvo, it reminds me the issue i hit yesterday with http confinement, in that case i wassn't able to read a file from /tmp, i had to move it to /var/lib/apps/{snap}/current
[09:58] <mvo> fgimenez: interessting. is adt using some confinement that restricts tmp?
[09:58] <fgimenez> mvo, not as far as i know, it just launches the executable
[10:25] <fgimenez> mvo, it seems that the entire /tmp directory gets wiped after the first test is executed, i'll try to find out which of the steps provokes this
[10:27] <JamesTait> Good morning all; happy Wednesday, and happy International Day for the Abolition of Slavery! 😃
[10:54] <ogra_> yay, armhf builds \o/
[10:56] <ogra_> Chipaca, so i got an update on my RPi... and thought i'd check if all my services came up again .... so i either call something like "ps ax" or have to call "snappy service status" for each of them ... thats quite a task if you have multiple snaps to check against
[10:57] <Chipaca> ogra_: sudo snappy service status
[10:57] <ogra_> i was thinking ... how about making something like "snappy service status all" work
[10:57] <ogra_> oh !
[10:57] <ogra_> bah
[10:57]  * Chipaca wins
[10:58] <Chipaca> it's snappy service <operation> [package [service]]
[10:58] <ogra_> right
[10:58] <Chipaca> so start, stop, status, etc, all can operate on all
[10:58] <ogra_> oh, the others too ?
[10:58] <ogra_> wow
[10:58] <Chipaca> the only one that isn't like that is logs
[10:58]  * ogra_ <- impressed
[10:59]  * Chipaca levels up
[10:59] <ogra_> +1
[10:59] <ogra_> :)
[10:59] <Chipaca> actually logs might be like that also
[10:59] <sergiusens> Chipaca, where is the developer.ubuntu.com information pulled from for snappy?
[11:00] <Chipaca> sergiusens: dholbach knows
[11:00] <Chipaca> sergiusens: it's manual for now
[11:00] <sergiusens> Chipaca, heh; oh, good; then I don't need a 2.x branch, master can be my 2.x branch :-)
[11:01] <ogra_> hmm
[11:01] <Chipaca> sergiusens: check with dholbach, because somebody's working on automating it
[11:01] <ogra_> 2015-12-02T10:03:58.171538Z ubuntu-core-launcher unable to bind private /tmp. errmsg: No such file or directory
[11:01]  * mvo hands Chipaca a blessed +5 keyboard of awesomeness
[11:02] <mvo> ogra_: what command is triggering this?
[11:02] <ogra_> it would be nice if it said *which* file or directory it was looking for ... or at the very least which was its calling process
[11:02] <ogra_> mvo, a boot
[11:02] <mvo> ogra_: a boot? you see that on boot?
[11:03] <ogra_> http://paste.ubuntu.com/13620872/
[11:03] <ogra_> it auto-rebooted after ubuntu-core upgrade
[11:03]  * ogra_ checks syslog directly
[11:04] <ogra_> ah, in syslog it shows the PID at least
[11:04] <ogra_> .... which doesnt help since the process is gone
[11:05] <Chipaca> ogra_: so which service failed?
[11:05] <ogra_> aha
[11:05] <ogra_> webdm
[11:05] <ogra_> it starts later on
[11:05] <Chipaca> ogra_: sudo journalctl -x will probably help figure out what happened
[11:06] <ogra_> http://paste.ubuntu.com/13620926/
[11:08] <Chipaca> quoth the raven: WAT
[11:09] <ogra_> Dec 02 10:03:58 localhost.localdomain audit[848]: AVC apparmor="DENIED" operation="mount" profile="/usr/bin/ubuntu-core-launcher" name="/tmp/" pid=848 comm="ubuntu-core-lau" flags="rw, bind"
[11:09] <ogra_> http://paste.ubuntu.com/13620953/
[11:09] <ogra_> the full journalctl output
[11:10] <jjohansen> ogra_: somewhat an apparmor bug, in that it shouldn't be generating a log message for that
[11:10] <ogra_> that seems to happen for two of my packages ... not webdm
[11:10] <ogra_> the failing webdm seems to be a separate thing
[11:11] <jjohansen> kern_path of the src path is being rejected, which will fail whether or not apparmor is there
[11:11] <mvo> ogra_: hm, probably a issue with the latest ubuntu-core-launcher and its apparmor rules, I did some changes there
[11:11] <jjohansen> mvo: it should fail even without apparmor
[11:12] <jjohansen> the apparmor bug is that it is logging that particular message at all, the failure is real, just not apparmor
[11:12] <mvo> aha, ok
[11:13] <ogra_> i have never seen that message when manually starting it ...
[11:13] <dholbach> sergiusens, yes, it's manual for now, but it'll be automatic soon
[11:13] <ogra_> though my system as outdated
[11:13] <dholbach> sergiusens, just let me know which branch to pull from
[11:13] <ogra_> *was
[11:14] <asac> jdstrand: no i dont, but let me think for a while and I might remember something
[11:14] <fgimenez> mvo, mmm it seems that executing any binary from a snap wipes /tmp http://paste.ubuntu.com/13621036/
[11:15] <mvo> fgimenez: ohhh, I think thats a issue with the ubuntu-core-launcher, I think its not wiping it just overlay mounting, hte same issue that ogra_ is seeing :/
[11:16] <ogra_> wow
[11:16] <ogra_> we could just sell it as a safety feature ;)
[11:18] <ogra_> hmm
[11:19] <ogra_> so while "snappy list" works now, webdm shows only installed snaps in the "snappy store" tab
[11:19] <ogra_> does it need an update ?
[11:23] <ogra_> ppisati, do you know if anything changed WRT v4l in vivid ? bug 1521710
[11:23] <ogra_> (in the kernel i mean indeed)
[11:23]  * mvo will look after lunch
[11:26] <ppisati> ogra_: no
[11:26] <ppisati> [flag@luxor ~]$ ls -la /dev/video0
[11:26] <ppisati> crw-rw----+ 1 root video 81, 0 Dec  2 09:34 /dev/video0
[11:26] <ppisati> [flag@luxor ~]$ uname -a
[11:26] <ppisati> Linux luxor 3.19.0-37-generic #42-Ubuntu SMP Fri Nov 20 18:22:05 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
[11:26] <ppisati> i don't see the kernel version there
[11:26] <ogra_> yeah, i asked for the full syslog ...
[11:26] <ppisati> this is my amd64 box + usb webcam
[11:26] <ogra_> snappy ?
[11:26] <ppisati> ogra_: nope, regular box
[11:27] <ogra_> right
[11:27] <ogra_> (amd64)ogra@aleph2:~$ uname -a
[11:27] <ogra_> Linux aleph2 3.19.0-33-generic #38-Ubuntu SMP Fri Nov 6 18:18:12 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
[11:27] <ogra_> thats my snappy ...
[11:28] <ogra_> (we havent done a build in a while in 15.04)
[12:07] <dholbach> kyrofa, do you know if Sergio wanted to upload 0.6 to xenial as well? Right now it's only in the ppa.
[13:05] <sergiusens> dholbach, https://github.com/ubuntu-core/snapcraft/tree/1.x for 1.0 and master for 2.x
[13:07] <sergiusens> kyrofa, ok, master is no 2.x and 1.x is, well, 1.x :-)
[13:07] <kyrofa> sergiusens, ah alright-- so do we need to stop updating documentation? Or can we switch the server pull?
[13:08] <sergiusens> kyrofa, already discussed with dholbach ;-)
[13:08] <kyrofa> sergiusens, great!
[13:08] <sergiusens> dholbach, btw, can you create a calendar event for the clinic? I almost forgot about ti
[13:11] <amp_> I have a quick stupid question: If I can reach my snappy from http://webdm.local:4200, on which URL should I look for xkcd-webserver?
[13:12] <kyrofa> amp_, same url, just port 80 I believe
[13:15] <amp_> kyrofa, thanks, I see it on 'http:wedm.local/' although the actual xkcd image doesn't load. What if I have multiple webservers? How would they be sorted?
[13:16] <ogra_> then you should set a hostname via "snappy config ubuntu-core"
[13:16] <dholbach> sergiusens, yep
[13:16] <ogra_> this will change "webdm.local" to "$hostname.local"
[13:16] <kyrofa> dholbach, sorry for not answering your question about 0.6 earlier-- I didn't get on until about 15 minutes ago. Did you get that question answered?
[13:17] <dholbach> kyrofa, no worries
[13:17] <dholbach> sergiusens, do you want to upload 0.6 to xenial as well? Right now it's only in the ppa.
[13:17] <dholbach> kyrofa, it wasn't that urgent - but sergiusens was offline for a bit, so I thought I'd ask you :)
[13:19] <sergiusens> dholbach, yeah, I don't mind
[13:19] <dholbach> sergiusens, ok
[13:19] <kyrofa> dholbach, yeah I haven't been on snapcraft that long-- I'm not up to speed on the release process just yet :)
[13:19] <dholbach> sergiusens, done
[13:19] <dholbach> kyrofa, no worries - I just thought you maybe had talked about it :)
[13:20] <sergiusens> dholbach, but I do want a calendar event as I don't know when it starts and want to prep 30' before it starts ;-)
[13:21] <dholbach> sergiusens, that's already done
[13:21] <sergiusens> kyrofa, also, if you want to look at https://github.com/ubuntu-core/snapcraft/pull/134
[13:21] <kyrofa> sergiusens, funny, I just finished
[13:21] <sergiusens> dholbach, oh, great; thanks :-)
[13:32] <kyrofa> sergiusens, after I've already built a .snap from snapcraft, if I change the code in a part being built, should I just be able to run `snapcraft build` again and expect it to build the new part but not pull down all the dependencies again?
[13:35] <sergiusens> kyrofa, that's what I intend to fix with 2.x; it was just too complex in the current code base
[13:35] <sergiusens> kyrofa, also, --force seems to not do what we want (at least what I want)
[13:36] <kyrofa> sergiusens, ah, so there's no way to do that in 1.x?
[13:36] <sergiusens> kyrofa, you can --force but it will go through `pull` again
[13:37] <sergiusens> which is what I find backwards :-)
[13:37] <kyrofa> Indeed
[13:37] <kyrofa> Alright, yeah we should change that for 2.0. Makes developing plugins slow
[13:38] <kyrofa> Most of my day is spent watching ROS download :P
[13:39] <sergiusens> kyrofa, there is a trick you can play; edit parts/<part-name>/state to be 'pull' iirc and that should make it 'build'
[13:39] <kyrofa> sergiusens, ooo, I'll test that out, thanks!
[13:40] <sergiusens> kyrofa, I don't know the extact syntax, but check what it does after just running `snapcraft pull`
[13:40] <kyrofa> sergiusens, alright
[13:41] <kyrofa> sergiusens, is #135 ready for review?
[13:41] <sergiusens> kyrofa, yeah; it might fail the tests since doctopt might not be there for trusty and I would need to backport and rerun; but code review can start
[13:43] <kyrofa> sergiusens, oh that's not quite as long as I anticipated
[13:43] <sergiusens> kyrofa, I split it up into many smaller reviews
[13:43] <kyrofa> sergiusens, you're a good man
[13:44] <sergiusens> kyrofa, the only problem this brings in is that the integration tests need to be redone in the last one
[13:45] <kyrofa> sergiusens, hmm. So the tests will be broken between these? Perhaps we should merge (and review) into a temporary branch so we don't have broken commits?
[13:45] <sergiusens> kyrofa, sounds good
[13:45]  * sergiusens creates a `new-cli` branch
[13:46] <sergiusens> kyrofa, unit tests all work though ;-) but point taken
[13:47] <kyrofa> sergiusens, good to know, and I don't feel super strongly about it, just some git OCD you know
[13:48] <sergiusens> kyrofa, ok, so 136 it is
[13:53] <mvo> fgimenez, ogra_: I can reproduce the /tmp overmount issue, looking into it
[13:53] <ogra_> thanks !
[13:58] <dholbach> sergiusens, I'm not quite sure what you had planned for the clinic, how about this? "Sergio Schvezov will give us an overview of what's new in Snapcraft and how to take advantage of its new features. Of course we will also have time to answer questions." - or is there anything missing you'd rather talk about?
[14:01] <sergiusens> dholbach, sure, I will show the new plugins and such as well
[14:02] <dholbach> awesome
[14:05] <dholbach> sergiusens, shall we direct people here or to #ubuntu-on-air?
[14:05] <dholbach> here should be better, right - some might just stay here
[14:08] <sergiusens> dholbach, yeah, here is fine
[14:08] <sergiusens> kyrofa, you can join the snappy/snapcraft clinic if you want
[14:08] <sergiusens> dholbach, also, why is only snapcraft doing clinics? ;-)
[14:09] <dholbach> sergiusens, I pinged mvo and niemeyer about doing a snappy one in between UOS and this one today and I never got a reply :-(
[14:09] <dholbach> maybe I should have brought this up in the standup
[14:10] <dholbach> I know how many emails we all get
[14:10] <kyrofa> sergiusens, I definitely plan on watching, at least. I'm not sure I know enough to have anything to add to the discussion itself just yet, unless you want me there
[14:14] <sergiusens> kyrofa, ah, I can introduce you as the ros goto person ;-)
[14:14] <sergiusens> then there is no escape! har har
[14:14] <kyrofa> sergiusens, hahaha
[14:14] <kyrofa> sergiusens, alright I'll be there. If for nothing else than moral support ;)
[14:30] <ogra_> hmm, is there a way for me to find out my own IP from a snap ?
[14:30] <sergiusens> ogra_, you can with go's net pkg
[14:31] <ogra_> hmm, and without any extra SW ?
[14:31] <sergiusens> ogra_, no software means no way of doing it ;-)
[14:31] <longsleep> ogra_: if your snap has network admin it can call all the tools
[14:31] <ogra_> (i only want to put it into the server name field of a config file so i can tell apart different machines with the same snap)
[14:32] <ogra_> hmm, but does network-admin pass the store automatically ?
[14:32] <longsleep> no idea
[14:34] <sergiusens> ogra_, no, it doesn't
[14:34] <ogra_> yeah, thats what i thought
[14:34] <kyrofa> sergiusens, sorry I'm late, grabbing headphones
[14:35] <ogra_> well, how does go do it ? i assume it needs to read some proc node
[14:36] <longsleep> ogra_: https://golang.org/src/net/interface_linux.go
[14:36] <longsleep> syscalls are your friend
[14:36] <ogra_> yeah
[14:51] <sergiusens> dholbach, hey, can you invite kyrofa to the clinic as well?
[14:51] <dholbach> sure
[14:55] <dholbach> one
[14:55] <ogra_> dholbach, do we need a meeting today (/me just noticed the calendar)
[14:55] <dholbach> ogra_, not for me...
[14:55] <ogra_> then lets skip :)
[14:59] <fgimenez> Chipaca, one quick question about the http client
[15:00] <fgimenez> Chipaca, in one of the corner case tests i'm getting http://paste.ubuntu.com/13624161/, is that ok?
[15:02] <asac> lool: can you make the tomcat webapp use upstream maven rather than from archvie?
[15:02] <asac> that would be cool
[15:03] <asac> i mean the maven plugin
[15:03] <asac> not urgent, but i think that would save me from apt getting half of the world during snapcraft build?
[15:04] <lool> asac: I am not sure it's good; we're getting java from Ubuntu anyway and that's the biggest thing
[15:04] <asac> lool: hmm ... what about getting java not from ubuntu?
[15:04] <lool> and typically developers iterating want to be able to run maven locally, and it's not clear how we'd install it for them
[15:04] <asac> i think java from ubuntu is almost always what folks dont want
[15:05] <lool> if you see what I mean
[15:05] <asac> not sure... arent those packages installed in a chroot?
[15:05] <asac> we could jsut put maven and java into such chroot instead of apt-get
[15:05] <lool> asac: while I would consider this upstream java not from Ubuntu a nice thing to have, it's not something straightforward; it requires maintenance and I dont think this is available straightaway
[15:06] <lool> consider finding an openjdk for ARM tarball with runtime deps, it's not that easy with some kind of security / QA commitmenet
[15:06] <asac> sergiusens: same prob
[15:06] <asac> http://paste.ubuntu.com/13624280/
[15:06] <asac> even with working link
[15:06] <lool> while we know the one from Ubuntu is up-to-date, tested, maintained; even if we generate an embedded copy, the embedded copy is up-to-date
[15:07] <asac> lool: i think what folks want they want oracle java
[15:07] <lool> asac: yes, and that's even harder to automate
[15:07] <sergiusens> asac, I plan to move to using maven from upstream
[15:07] <asac> sergiusens: what about java?
[15:07] <asac> oracle
[15:07] <asac> i think trick is the license accept
[15:07] <asac> on instlal
[15:08] <asac> but i think with Chipaca's feature thats avail too
[15:08] <sergiusens> asac, oracle java is easy; I have a plan for the license
[15:08] <asac> please!!!!!
[15:08] <lool> I don't think Oracle is keen on people automating the download + license acceptance of their stuff
[15:08] <asac> i need it
[15:08] <lool> unless you spawn a browser and pickup the tarball from its Downloads
[15:08] <asac> the license accept is on console for oracle
[15:08] <asac> afaik
[15:09] <asac> and i its supposely fine to remember
[15:09] <sergiusens> asac, I think lool is mentioning the fact that the download link is hidden until the license is accepted on the webpage
[15:09] <lool> right, and we specifically pulled off the oracle java installer from Ubuntu IIRC
[15:10] <asac> i dont think thats the case
[15:10] <lool> which was this .deb with a license prompt d/ling it from oracle
[15:10] <asac> hmm. it is
[15:10] <asac> guess they changed that from sun approach
[15:12] <lool> this is how I see it's being done:
[15:12] <lool> https://github.com/dockerfile/java/blob/master/oracle-java8/Dockerfile
[15:12] <asac> nice they even prevent you from wget that link that you got on browser
[15:12] <lool> installer is from PPA, you set the debconf option not to prompt
[15:12] <lool> the installer connects to the site, saves the cookie, prompts for license, downloads with cookie
[15:12] <asac> right
[15:12] <asac> thats awful
[15:12] <lool> and I don't think what this PPA does is legal
[15:13] <asac> well, the user using it is doing something not legal
[15:13] <asac> not sure if making such code avail is illegal in itself
[15:13] <asac> but then the code facilitates
[15:13] <asac> :)
[15:13] <asac> a crime
[15:13] <asac> theft of key IP
[15:14] <lool> you're kind of an accomplice by suggesting we should do this!  :-)
[15:14] <asac> so guess the plugin could support the ability to point to a local tar
[15:14] <asac> lool: i dont think we should
[15:14] <ogra_> lool, at least we have someone to blame then ...
[15:14] <lool> asac: yes, local tar is what I had suggested in the past
[15:14] <asac> we should just make it so that we hav a plugin that allows user to point to a URL that can be wget without cookie
[15:14] <ogra_> is that the main job of managers anyway ? :)
[15:14] <asac> local on disk
[15:14] <asac> or on local desktop
[15:15] <asac> and then have the openjdk option
[15:15] <asac> preferably not from archive
[15:15] <ogra_> why not ?
[15:15] <asac> but if thats archive its probably also fine
[15:15] <ogra_> yeah, i guess for openjdk the archive should be fine
[15:15] <asac> because in archive you dont have control over what exactly you get
[15:15] <asac> which veresion etc.
[15:16] <asac> sure its fine for most, but then you are not in charge about what you get
[15:16] <ogra_> but you can at least rely on the fact that it is well maintained ... and you can look up the version easily
[15:16] <asac> and put in your sofrtware
[15:16] <asac> doesnt matter
[15:16] <asac> devs should have choice to upgrade when they want... if they can auto get latest that is great
[15:16] <ogra_> but we should laso provide comfortable defaults :)
[15:16] <asac> thats the whole point of bundling dependencies
[15:17] <asac> thats fine, but we have no way to stay on the old version on respins
[15:17] <ogra_> as a dev i perhaps only care about my jar
[15:17] <ogra_> i dont want to have to care for what java i get etc to run it
[15:17] <asac> so you might wake up witha  firedrill in your code, want to respin and then BOOM get the heap from doko with many bugs and changes you didnt want during that night too
[15:17] <asac> so your night getse long and your firedrill delayed.
[15:17] <asac> you always want both
[15:17] <ogra_> well, you indeed dont want to get doko's experimental crap :)
[15:18] <ogra_> but his well tested packages from the archive
[15:18] <sergiusens> asac, there is no easy way to get openjdk from NOT the archive
[15:18] <asac> if you cant have both than you want explicit control
[15:18] <asac> sergiusens: lool found a place i think
[15:18] <asac> nice tarball to unpack
[15:18] <asac> he didnt like it because he is idealisitic :P
[15:18] <ogra_> binary ?
[15:18] <asac> yeah
[15:18] <lool> it was like a random guy
[15:18] <lool> probably building things on his fridge
[15:18] <lool> over IPv6
[15:19] <ogra_> yay, thats surely better than dokos packages
[15:19] <asac> right. so thing is that in the past linux distros refused to really consume such binaries and offer that conveniently, but maybe we can talk tot hem
[15:19] <ogra_> :P
[15:19] <lool> and embedding a SPAM bot in openjdk
[15:19] <asac> and tell them we would love to see official binaries
[15:19] <lool> yes!
[15:19] <asac> so if you look around python also has no official binaries i think
[15:19] <asac> while those that target all platofrms usually have agood binaries
[15:19] <asac> like golang etc.
[15:19] <sergiusens> lool, asac as the snapcraft team I don't mind; I was planning on maintaining the go binaries for armhf as well
[15:20] <asac> sergiusens: the prob is that openjdk is afaik super nasty to build
[15:20] <lool> it is
[15:20] <asac> more like building the toolchain
[15:20] <sergiusens> asac, python is the worst thing to snap
[15:20] <asac> imo deserves a redo
[15:20] <asac> cant be that folks do something that cant be built with just make make install in 2015
[15:20] <sergiusens> asac, no one cares for openjdk in any case, only distros
[15:20] <asac> right
[15:20] <lool> let's invest 2-3 months into making openjdk easier to build to provide static binaries, then another 2-3 months on Python and then we're good
[15:20] <lool> :-P
[15:21] <ogra_> thats only half a year !
[15:21] <asac> but maybe having great binaries for windows and mac would boost how much folks care
[15:21] <asac> fact ist hat most folks  deploy on linux
[15:21] <asac> but dev on windows :P
[15:21] <asac> so if you dont have openjdk for windows conveniently its like the dead before the start
[15:21] <ogra_> C:\snappy.exe install myjavatool-0.1.snap
[15:22] <nikna> what's the new ?
[15:22] <rospppse> ?
[15:25] <elopio> fgimenez: are you ok if we skip our meeting today?
[15:26] <asac> lool: sergiusens: ok -ranting, can we make the current maven/java plugin have a flag that makes it consume the JVM from a path that user has to configure; otherwise it takes the archive one?
[15:26] <fgimenez> elopio, sure, np
[15:26] <asac> sergiusens: or what is the plan?
[15:27] <asac> in any case think nothing speaks against using maven from upstream always?
[15:28] <asac> (well, besides not having a current link on apache mirros for course)
[15:29] <asac> sergiusens: going back... http://paste.ubuntu.com/13624629/ is really broken
[16:06] <dholbach> if you have questions for the snappy clinic, make sure you prefix them with QUESTION: so we can more easily pick them up
[16:08] <dholbach> (currently on on http://ubuntuonair.com)
[16:16] <dholbach> any questions for Kyle and Sergio so far?
[16:24] <shout-user> hello
[16:41] <dholbach> kyrofa, elopio, sergiusens: great work guys! thanks a lot!
[17:01] <nigel380> byeee
[17:02] <kyrofa> Hey, maybe that netsplit is finally over
[17:51] <dvdbot> Is Ubuntu Snappy only meant for web applications or are you able to show stuff on the screen the rpi for example is connected to? Like a website or something
[18:01] <ogra_> dvdbot, there is no Mir framework yet (but planned)
[18:02] <dvdbot> Ah okay, I'm starting with a Digital Signage project and came across Ubuntu Snappy and thought maybe that's it. Gonna have to look further
[18:02] <dvdbot> Anyone got any suggestions?
[18:03] <dasjoe> Just the normal Ubuntu Core, not snappy?
[18:06] <dvdbot> I meant the Ubuntu Core, the thing is I need a UI or something to show stuff on the screen the RPI is connected to, tried the Windows IoT already but it can't handle video's yet
[18:16] <kyrofa> sergiusens, that pull state hack has already saved me probably two hours
[18:21] <sergiusens> kyrofa, awesome
[18:41] <sergiusens> Chipaca, how do we purge snaps? seems `snappy purge` seems to not do anything
[18:41] <sergiusens> too  many seems seems bad
[18:41] <sergiusens> :-P
[19:10] <jerryG> in snapcraft.yaml, how do i specify the version # of dependencies?
[20:03] <plars> elopio_: I think I already know the answer to this, but just checking... the tests you have in snappy are pretty heavily dependent on the framework there, correct? That is, I couldn't compile them, take a bunch of binaries, and just run them as individual tests locally on a snappy instance?
[20:37] <Chipaca> sergiusens: what do you mean, how do you purge them?
[20:37] <Chipaca> sergiusens: tell me what you want 'purge' to do :-)
[20:38] <sergiusens> Chipaca, to really purge, snappy list --verbose keeps listing inactive snaps
[20:38] <Chipaca> sergiusens: purge is about removing the data
[20:39] <sergiusens> Chipaca, ah
[20:39] <Chipaca> sergiusens: sudo snappy remove app=version will do what you want
[20:39] <sergiusens> pcoca, ^
[20:39] <Chipaca> although you lose the ability to rollback when doing that :)
[20:40] <sergiusens> Chipaca, no worries, no more disk space
[22:09] <balloons> afternoon all. I'm curious who might be the best person for help with snapcraft atm?
[22:10] <balloons> sergiusens perhaps?
[22:11]  * balloons puts on his recruitment hat
[22:11] <balloons> I'd like to ask someone to volunteer to be a mentor for some snapcraft tasks for GCI
[22:13] <sergiusens> balloons, I would gladly say yes, but I have offered too much of my free time already; I can give it some thought on the 14th
[22:13] <sergiusens> if it is too late I propose elopio_
[22:14] <balloons> sergiusens, if it's ok with you I'll add the tasks and add you as mentor. I am happy to stand in as well ofc. It would just be useful to have someone closer to the core to tap for help
[22:15] <balloons> sergiusens, also, do you have a list of packages that would make good targets for snaps? I know dholbach had a list at some point, but I don't have it atm
[22:15] <sergiusens> balloons, some on that list are really hard, well at least not trivial
[22:15] <balloons> I'd like to get as many as I can, as I think growing snaps would be quite useful
[22:16] <balloons> sergiusens, ahh, yea, we're looking for things that in general would be a bit easier
[22:16] <sergiusens> balloons, snaps of the parts wiki with useful parts
[22:16] <balloons> if you don't have such a list, I'll use my own thoughts
[22:17] <sergiusens> balloons, not really, no; but this is where elopio_ can be of help too ;-)
[22:18] <sergiusens> balloons, in general, easy ones are nodejs or go based projects
[22:18] <balloons> ok. I was also thinking of simple command line tools like pastebinit
[22:19] <balloons> so something like your snap of shout irc (which is nodejs) was simple then eh?
[22:20] <sergiusens> balloons, it took me an hour to write the plugin with 100% unit test code coverage, create the shout snap and publish to the store
[22:20] <balloons> sergiusens, that sounds good. We're looking at 3-5 hour tasks for students
[22:20] <balloons> right on the money. I'll pick some cool stuff I know of :-)
[22:21] <sergiusens> balloons, right, at most, any nodejs source would need to be mangled a bit to use the writable parts of the filesystem instead of `./`
[22:21] <sergiusens> not the case for shout
[22:22] <balloons> sergiusens, so do I have a hesitant yes from you for later in the contest? It runs through Janurary, so we can hold this stuff until a bit later
[22:23] <sergiusens> balloons, sure, just take into account that that is summer time here and I might take a couple of more days off ;-)
[22:23] <balloons> :-) I don't want you to not enjoy summer!
[22:23] <sergiusens> I'm also going to be at ubuconn
[22:23] <balloons> right, me too
[22:57] <elopio> plars: if you pass the build tag excludereboots, then they are standalone. Maybe you will have problems finding the errors in the output, because it won't generate the subunit this way.
[22:59] <elopio> balloons: sergiusens: I can also give a hesitant yes.
[22:59] <balloons> elopio, :-) I'll add you
[22:59] <balloons> elopio, see my pm about your donations request. We should chat about it tomorrow
[22:59] <elopio> balloons: pm as in email?
[23:24] <sergiusens> elopio, another one for you to enjoy ;-)
[23:24] <sergiusens> https://github.com/ubuntu-core/snapcraft/pull/141
[23:29] <plars> elopio: I'm not sure what you are referring to
[23:32] <sergiusens> balloons, I can't seem to accept your invitation; I get logged out
[23:33] <elopio> plars: the framework is needed for handling reboots and merging results.
[23:33] <elopio> if you don't run the reboots, then the binary with the tests will work nicely if run from inside the testbed.
[23:36] <plars> elopio: I should elaborate on what I'm looking to do...
[23:36] <plars> elopio: I'm looking into whether the tests can be used by something else, such as checkbox, so that the cert team can get results produced by checkbox for running the individual tests in your suite, without having to port them all over individually
[23:37] <plars> elopio: of course that means that we would need to somehow build, then pull them out - because checkbox does not run remotely, it runs *on* the device under test
[23:38] <elopio> plars: I talked a little with sylvain about that. Personally, I would just first run the autopkgtest suite and then the checkbox suite.
[23:38] <plars> elopio: ah, good. we are probably looking at the same thing :)
[23:38] <elopio> but if you get checkbox to handle the reboot flags, then you could even run the reboots from inside the testbed.
[23:39] <elopio> plars: read the integration-tests/README. We explain in there the reboot details.
[23:39] <balloons> sergiusens, ohh.. is there another google account I should invite?
[23:39] <balloons> I invited your canonical one I believe
[23:39] <balloons> feel free to pm
[23:40] <plars> elopio: I'll take a look, but would you be interested in possibly having a sync up with us later in the week, or maybe even next week? I'd like to talk about this, but also other plans you guys have around testing snappy
[23:40] <elopio> plars: of course. Just send me the invite.
[23:42] <plars> elopio: great! I'll mention it to sylvain and others at tomorrow morning's standup and see who else might want to attend. thanks, and have a good evening :)
[23:43] <sergiusens> balloons, no worries, remind me to click on the link tomorrow just in case ;-)
[23:43] <balloons> ack