=== JoshStrobl is now known as JoshStrobl|zzz === JoshStrobl|zzz is now known as JoshStrobl [10:24] o/ === JoshStrobl is now known as JoshStrobl|Work === ShalokShalom_ is now known as ShalokShalom [12:24] hi [12:25] is here the best place to ask about build.snapcraft.io? [12:45] sSs_: hey [12:45] sSs_: I think you should ask at forum.snapcraft.io [12:45] :) [12:45] thanks! [12:59] jdstrand: hey [12:59] jdstrand: are you at the rally? [13:00] zyga-ubuntu: he is [13:00] PR snapd#3958 opened: many: add support for /home on NFS [13:02] * mwhudson waves at zyga-ubuntu [13:02] sadly [13:04] mwhudson: hey [13:04] :-( [13:04] yeah [13:05] popey: say hi to him :) [13:23] <__chip__> o/ [13:37] Bug #1681547 changed: Gnome3 on Ubuntu 17.04 doesn't find snap desktop files [13:51] zyga-ubuntu, you just need http://bit.ly/2yDzMdg [14:02] can someone point me to the seeding code? [14:04] Can you guys make a template for GNOME Builder that includes files to build a snap? [14:05] ogra_: hehe [14:05] ogra_: and someone to drive me around :) [14:06] smallfoot-: that's an interesting idea [14:06] also is there some way to make seeding extremely verbose? [14:06] zyga-ubuntu, well, thats old ... nowadays you'd have a flying drone driven by ubuntu core ;) [14:06] smallfoot-: can you please post this proposal to forum.snapcraft.io so that it doesn't get lost? [14:06] ogra_: wait, there's a core refresh [14:06] xxxxx [14:06] lol [14:06] ogra_: I see my dragonboard power usage spike whenever it reboots on core rebuild [14:07] rebuild ? you mean refresh ? [14:07] well, rebuild that triggers refresh [14:07] yes [14:08] all our default kernels boot with the performance cpufreq governor (to speed up booting) and use a script to switch to ondemand 60sec after the boot [14:08] theoretically you shoulld see such a spike on every boot ... not typically related to the refresh itself [14:08] right, but it doesn't reboot otherwise :) [14:08] it runs 24/7 [14:08] yeah [14:09] ogra_: btw, did you see my lab? [14:09] but it is just a coincidence [14:09] i saw the pics from a while ago [14:09] are there new ones ? [14:09] zyga-ubuntu, I don't have any forum account on snapcraft.io [14:09] no, nothing new. I plan to add 4 more boards but I have no time to set them up [14:09] I need a small rack to keep this tidy [14:10] well, now that you are idling at home and everyone esle is in NYC you should have some time ;) [14:13] ogra_: I'm not at home today [14:14] ogra_: but I can ssh home :) [14:14] heh [14:14] ogra_: I mainly don't have a rack [14:14] ogra_: I had some errands to do in the morning, taxes and VAT paperwork [14:14] home is where the authorized_keys is ! [14:14] ogra_: hahaha :) [14:14] yeah [14:15] i would love to give you my 2x1x1m server cabinet ... but i guess delivery would be really expensive [14:15] (trying to get rid of it since ages) [14:16] hmmm [14:16] how heavy is it? [14:16] I have space for that [14:17] but I'd rather have small desk rack as I have separate office network and no wiring to where I could keep the rack [14:17] (the big rack) [14:17] its a big metal cabinet with glass front and back doors probably somewhere between 50-100kg [14:18] nice, why do you want to let go of it? [14:18] because even 20 of the boards i work with nowadays fit on a shelf ... [14:19] it just takes space, i dont really have any big 19" server HW in use anymore [14:19] pedronis: are you looking at irc? :) [14:19] ogra_: can you send me a pic sometime (telegram is fine) [14:20] yeah, remind me next week when i'm home [14:20] thank you! [14:21] Why does 'snapcraft init' put the yaml file in snap/ instead of build-aux/snap/ ? [14:22] Now it pollutes the root, instead of be like build-aux/snap/ and build-aux/meson/ [14:22] PR snapd#3957 closed: cmd,dirs: treat "liri" the same way as "arch" [14:36] jdstrand, I assume you're at the rally? [14:42] PR snapd#3959 opened: hooks: rename refresh to after-refresh [14:58] PR snapd#3956 closed: snap-confine: bind mount /usr/lib/snapd relative to snap-confine [15:02] pstolowski, https://paste.ubuntu.com/25615089/ [15:07] zyga-ubuntu, are you here? [15:07] (on IRC, I mean) [15:07] sergiusens: is the env stuff from c#1 in LP: #1600035 done? A comment from ~1 year ago says "we are also going to allow..." without any date [15:07] Bug #1600035: no way to differentiate between build-time and run-time env vars in plugins [15:12] kyrofa: or maybe you kow (about environment: / build-environment: in the yaml). That would I think basically drop my custom wrappers and plugin altogether [15:13] nacc, environment is there, yeah, should get rid of a lot of wrappers [15:13] nacc, build-env, I don't think so [15:13] nacc, environment is a key on each app you declare [15:13] (like plugs) [15:13] There _might_ be a global one... I can't remember [15:14] kyrofa: hey [15:14] kyrofa: yes [15:14] kyrofa: oh ok, it's not here: https://snapcraft.io/docs/build-snaps/syntax [15:14] Hey zyga-ubuntu! I see the kernel-module-control interface... any chance you could give me a quick synopsis of what that does? [15:15] kyrofa: that's still better than what I have, so that's fine :) [15:15] kyrofa: yes, sure [15:15] nacc, hmm, indeed. This will help: https://stackoverflow.com/questions/42991501/snapcraft-custom-ld-library-path [15:15] kyrofa: that interface allows the snap to insert/remove kernel modules [15:16] nacc, would you mind logging a bug for the missing docs, please? Against snapcraft [15:16] kyrofa: by itself, not with snapd acting at the middle man [15:16] kyrofa: so you can modprobe essentially [15:16] zyga-ubuntu, modules contained within an app snap? [15:16] kyrofa: tip: snap interface kernel-module-contorl [15:16] PR snapd#3960 opened: travis: switch to container based test runs [15:16] kyrofa: well, any, you can wget modules or stuff [15:16] zyga-ubuntu, dude, awesome. Is that privileged? [15:16] kyrofa: it's just the permission to perform the kernel operation [15:16] kyrofa: totally [15:16] kyrofa: it requires an assertion to connect [15:16] kyrofa: canonical-livepatch uses it [15:18] kyrofa: yep will do after a mtg [15:19] * zyga-ubuntu will EOD soon [15:19] any last calls from the Ubuntu rally? [15:20] mvo: what is the motivation for https://github.com/snapcore/snapd/pull/3960/files ? [15:20] PR #3960: travis: switch to container based test runs [15:20] mvo: is it just faster [15:20] zyga-ubuntu: correct [15:20] zyga-ubuntu: also more "machines" available [15:20] zyga-ubuntu: its a bit of an experiment right now [15:20] aha, thank you for the explanation [15:20] hehe, "machines" :-) [15:21] zyga-ubuntu: your welcome, we are currently trying to find out if this is true [15:22] zyga-ubuntu: right now there is quite a bit of lack between pushing a PR and the first time travis asssigns a worker [15:22] yep [15:23] "lag" I assume [15:23] we're using the KVM based machines and those just cost them more [15:29] * zyga-ubuntu EODs [15:29] have a great rally everyone! [15:31] sergiusens: what is the 'correct "program loader"'? and how is that determined? [15:34] kyrofa: yep, I'm here. zyga asked too if you see him [15:59] PR snapd#3959 closed: hooks: rename refresh to after-refresh [16:02] ppisati, http://paste.ubuntu.com/25615430/ is the attempt to fix AlbertA and bschaefer's SD corruption ... [16:17] sergiusens: https://paste.kde.org/prox9diqs [16:21] stgraber mind parsing what that means ^... this is on arch [16:29] apol: you need the linux-userns kernel [16:29] apol: the default Arch kernel doesn't have user namespaces so only privileged containers can be used until you switch to the linux-userns kernel [16:29] apol: you can create privileged containers with -c security.privileged=true [16:31] right... [16:32] I'll try with -c security.privileged=true [16:36] sergiusens: is there black voodoo going on that allows generated wrappers (in classic snaps) to have #!/bin/sh in them? [16:37] is it possible to mount a rw slot (write: $SNAP_DATA) from a ro plug (target: $SNAP/foo)? with core 16-2.27.6, i get permission denied: http://paste.ubuntu.com/25615637/ === ShalokShalom_ is now known as ShalokShalom [17:47] nacc check snapcraft.internal.meta, but we don't generate wrappers for classic [17:47] ... sure you do :) [17:48] http://paste.ubuntu.com/25615968/ [17:48] I don't konw *why* you do [17:48] http://paste.ubuntu.com/25615970/ [17:51] nacc to allow you to specify `command: some-command -w $SNAP_USER_DATA` [17:51] nacc so, environment variables [17:52] sergiusens: right, so you *do* make wrappers :) [17:52] nacc yeah, we do; that dropped my mind [17:52] but... [17:52] sergiusens: yep, i imagine it's some voodoo so that /bin/sh isn't used from the host FS [17:53] nacc your thoughts on https://github.com/snapcore/snapcraft/pull/1420 would be appreciated [17:53] PR snapcraft#1420: add new "no-wrapper" property to apps [17:53] sergiusens: i'll take a look :) [17:54] nacc for classic the voodoo is incomplete though, we would need to change it to `#!/snap/core/current/bin/sh` [17:54] sergiusens: i know little of the internals, but can give feedback [17:54] sergiusens: yep, that's what i'm havig to do in my scripts, or they segfault :) [17:55] nacc if you want to create a PR with that fix to use the correctly shebanged sh in the wrapper I would appreciate that ;-) [17:55] if not I'll take care of it [17:55] sergiusens: heh, i'll put it on my backlog :) [17:56] nacc if you can't get to it, just tell me and I'll take care of it [17:57] sergiusens: it's probably not going to bubble to the top of my queue anytime soon, so it's probably better for you to do [18:11] mvo, when you have a minute https://github.com/snapcore/spread-cron/pull/46 [18:11] PR spread-cron#46: Add configure files as part of the sync [18:11] cachio: sure, coming in a minute [18:15] ogra_: http://people.canonical.com/~ppisati/linux-image-4.4.0-1075-raspi2_4.4.0-1075.83~ograsdquirks_armhf.deb [18:15] * ogra_ hugs ppisati ... [18:26] PR snapd#3961 opened: git: make the .gitingore file a bit more targeted [18:27] nessita: hi! I just noticed that https://dashboard.snapcraft.io/dev/snaps/6058/rev/661/ is 'stuck'. the automated review passed, but it didn't get approved and there are a bunch of revisions after it pending review [18:28] nessita: (and it's showing up in https://dashboard.snapcraft.io/dev/snaps/reviewer/ubuntu/) [18:29] jdstrand: hm, similar to what just happened with hugo, which I just unblocked [18:29] jdstrand: ah! 6058 *is* hugo :) [18:31] jdstrand: 661 is now ready to release? so is 662... [18:31] jdstrand: revision 620 was pending resolution, so all subsequent uploads got held. I unblocked 620 and the queue is being processed... [18:33] almost there, there are 714 revisions, it's up to 702 and working [18:38] elopio1, jdstrand : all outstanding hugo uploads are reviewed, passed and ready to release [18:48] pstolowski, i need the output of: grep . /sys/class/mmc_host/mmc0/mmc0:*/* 2>/dev/null [18:49] ogra_, http://pastebin.ubuntu.com/25616270/ [18:49] perfect ! [19:00] I was trying the tutorial found at snapcraft.io and followed it exactly but keep getting this error when running the snapcraft command: "Issues while validating snapcraft.yaml: found character '\t' that cannot start any token on line 14 of snap/snapcraft.yaml". I cannot find the "\t" character anywhere in the .yaml file. Anyone else run into this issue? [19:01] navyguns: it's a tab character [19:02] Thanks. I'll check. [19:03] navyguns: my guess would be mixing tabs and spaces, possibly? [19:07] Yes, that was it. Now i'm getting a "mapping here on line 2" which contains Version: '2.10" [19:12] Well, got rid of those errors. Executing snapcraft now produces a bunch of python3 code errors and then quits. Running on Mint 18.2 Cinnamon. [19:21] thank you. I've notified them in the bug. And thanks jdstrand for looking at it too. [19:23] navyguns: 'code errors'? Can you pastebin the command and output? [19:24] quit [19:26] https://pastebin.com/WprAeU7Q [19:28] navyguns: it would appear that mint's snapcraft doesn't work? [19:29] navyguns: dunno, that error implies no subclass is being used [19:31] Thanks for looking. I'm going to have to deal with it later. Work calls. [19:36] heyo jdstrand: i can't figure out a hook that i can call to setup a content plug subdirectory early enough to beat the interface auto-connect. does such a thing exist? (install, prepare-plug-foo don't seem to work). when you get some time, can you advise on https://bugs.launchpad.net/snapd/+bug/1719370 [19:36] Bug #1719370: snaps with content plug fail to install with auto-connect [19:38] i have a couple workarounds: 1) install the plug snap first, then the slot snap, or 2) work out something using the home interface. ideally though, i'd like to keep these bits in a $SNAP_x directory vs every users' $HOME. [20:18] PR snapd#3961 closed: git: make the .gitingore file a bit more targeted [20:31] PR snapd#3962 opened: tests: Increase SNAPD_CONFIGURE_HOOK_TIMEOUT to 3 minutes to install real snaps [20:34] hi [20:34] to use https://build.snapcraft.io [20:34] the snapcraft.yaml needs to be in the root or snap directory only ? [20:35] it can't look other directory ? [20:36] gouchi: where else would you like it to live? snap sources often have toplevel snap/ [20:37] some software may organize their source differently with sub directory [20:37] for instance pkg/deb, pkg/rpm , pkg/snap ... [20:38] not necessarily in snap/ [20:45] it will be great if you can specify custom path to snapcraft.yaml for the building process [20:48] +1 gouchi.. fwiw, i'm writing snaps for apache bigtop which keeps their packaging bits in ./pkgs/src/[deb|rpm]. my workaround for now is to maintain intermediate repos for use by the snap builders (1 snap per repo, each with ./snap/snapcraft.yaml). i plan to manually sync those with upstream periodically, but if the builders supported a user-specified snapcraft.yaml location, that'd be awesome! [20:51] gouchi: it depends, i guess [20:51] as kwmonroe said, it's trivial to create just a basically empty repo with snap/ inn it [20:51] that fetches the 'upstream' source via git [20:51] (in the yaml) [20:52] yes [20:53] kwmonroe: why do you need to sync yours with upstream? [20:53] kwmonroe: that is, why isn't the upstream a part in your yaml [20:55] nacc: the upstream is a part of my yaml (https://github.com/juju-solutions/snap-hadoop/blob/master/snap/snapcraft.yaml#L100) just like you said -- fetching upstream in one of my parts. but since bigtop houses rpm and deb control files, it'd be nice to stick ./snap up there too. [20:55] to clarify, the nice part would be if a snap builder (snapcraft or lp) would let me specify the snapcraft.yaml location in the repo when i register it to build. [20:56] kwmonroe: oh i see, so they could do some testing? and for the latter, so you don't need the intermediary [20:56] yup [20:56] (and for the former, to match their other stuff) [20:56] got it [20:59] fwiw nacc, projects like bigtop house 30ish other projects (https://github.com/apache/bigtop/tree/master/bigtop-packages/src/deb), so if i wanted to have a snap for each of those, i'd be maintaining 30is intermediate repos. [21:00] kwmonroe: true enough [21:11] PR snapcraft#1570 opened: Don't fail over empty or faulty lines [21:14] PR snapcraft#1571 opened: Make sure we don't try to include an empty set [21:14] kyrofa: https://github.com/snapcore/snapcraft/pull/1571 https://github.com/snapcore/snapcraft/pull/1570 [21:14] PR snapcraft#1571: Make sure we don't try to include an empty set [21:14] PR snapcraft#1570: Don't fail over empty or faulty lines [21:48] PR snapd#3963 opened: cmd/snap-confine: add support for per-user mounts [21:54] flexiondotorg `export SNAPCRAFT_CONTAINER_BUILDS=1` [21:56] sergiusens: thanks [22:05] PR snapcraft#1572 opened: catkin plugin: allow ROS_MASTER_URI change