[01:03] Bug #1906327 opened: aws-iot-greengrass snap *was temporarily gone [01:12] Bug #1906327 changed: aws-iot-greengrass snap *was temporarily gone [01:15] Bug #1906327 opened: aws-iot-greengrass snap *was temporarily gone [07:05] morning [07:15] brb, reboot for a new kernel [07:17] re [07:18] PR snapd#9421 closed: interfaces/utf: Add Ledger to U2F devices [07:30] mvo: morning [07:30] good morning mborzecki [07:57] mornings [08:01] good morning pstolowski [08:18] pstolowski: hey [08:23] PR snapd#9722 closed: daemon: split apps and logs endpoints to api_apps.go and tests [08:49] mvo: I commented on #9718 [08:49] PR #9718: secboot,devicestate: add scaffoling for "fde-reveal-key" support

[08:50] mvo: mborzecki: hi, #9713 needs 2nd reviews/re-reviews, my fix seems to have worked [08:50] PR #9713: tests: sign new nested-18|20* models to allow for generic serials [08:51] mvo: mborzecki: thanks for reviewing my daemon cleanups [08:52] pedronis: hi, i think you can force merge https://github.com/snapcore/snapd/pull/9723 when the spread jobs finish, i've restarted them but one of the tests still failed on amazon due to 408 from the store [08:52] PR #9723: daemon: move some more api tests from daemon to daemon_test [08:53] pedronis: and https://github.com/snapcore/snapd/pull/9719 is trivial and could use your review [08:53] PR #9719: boot: set kernel command line in modeenv during install [09:03] mborzecki: it's in my queue [09:10] pedronis: thanks [09:27] mvo: I re-reviewed #9715 , I made a non-blocking suggestion but maybe is too complicated for now [09:27] PR #9715: hookstate: add new HookManager.EphemeralRunHook()

[09:36] mborzecki: I asked a question in #9719 [09:36] PR #9719: boot: set kernel command line in modeenv during install [09:59] good morning guys [10:00] got out of most of the planning mess and I will be back to normal on-irc working on foss-y bits tomorrow [10:00] I will resume snapd patches as well [10:01] the lat few days were sort of an internal planning sprint where I'm just always in meetings and planning [10:04] mvo spread has no missing deps in Debian anymore [10:05] coffee done and back to planning :) [10:05] zyga: woah! did you update it to use only pkgs that are already packaged? [10:05] mvo yeah, I just need to give you the package [10:05] zyga: let me know when you need sponsoring [10:05] given that I run part of my own planning now [10:05] I added a task group for handling debian packaging [10:05] so I will definitely have time for shared work [10:06] it's just not today, time to write down all decisions and fill in various comms systems for others to know [10:06] good luck and see you soon :) [10:15] pedronis: thanks, let me take a look [10:15] zyga: sure, no rush! see you [10:22] mborzecki: what go version we have on on amazon linux? [10:23] pedronis: let me double check but iirc it was 1.11+ [10:23] 1.15 [10:23] sorry, we print nowadays [10:23] pedronis: yup that's quite likely [10:27] pedronis: centos 7 should have 1.15 too [10:28] PR snapd#9723 closed: daemon: move some more api tests from daemon to daemon_test [10:43] pedronis: i've updated https://github.com/snapcore/snapd/pull/9719 [10:43] PR #9719: boot: set kernel command line in modeenv during install [11:45] mborzecki: +1 [11:49] thanks pedronis ! [11:49] 9715 needs a second review, maybe pstolowski can have a look, hopefully not too long/difficult [11:50] mvo: sure, will do [11:50] thanks a lot! [11:55] pedronis: thanks! [11:57] mvo: I commented on the cleanup: https://github.com/snapcore/snapd/pull/9715#discussion_r533353541 [11:57] PR #9715: hookstate: add new HookManager.EphemeralRunHook()

[11:57] looks like apt hooks and snap advise is failing [11:58] (the tests i mean) [11:58] pedronis: excellent, I poke on this next [11:58] heh HTTP/1.0 429 Too Many Requests [11:59] when updating snap catalog [12:16] mvo: can you explain what does it mean to run hooks in ephemeral mode? maybe i'm confused by what epehemeral meant for apps & snapctl. what is the problem this is addressing? [12:18] pstolowski: we want to run a hook from snapd without a change or task [12:18] synchronously [12:19] pedronis: ah, i see, makes sense, thanks [12:39] PR snapd#9724 opened: boot: observe successful command line update, provide a default [14:50] PR snapd#9725 opened: bootloader: remove installableBootloader interface and methods [15:13] Bug #1906327 changed: aws-iot-greengrass snap *was temporarily gone [15:20] mborzecki: do you know if we have unit tests for snap-confine? if so where are they? [15:21] ijohnson: we have some, they're in cmd/snap-confine/*-test.c files [15:21] mborzecki: I commented on #9724, some questions there [15:21] PR #9724: boot: observe successful command line update, provide a default [15:21] mborzecki: ah thanks I think snap-confine-invocation-test.c is what I was looking for [15:21] ijohnson: though it's not per se the whole of snap-confine functionality covered by the test, rather some smaller individual bits [15:21] pedronis: thanks [15:22] ijohnson: you asked about unit tests, there's also spread tests for it in case, under spread-tests [15:23] yeah I saw the spread tests, just wondering if we also had unit test coverage for snap-confine [15:23] w8, do we run those tests? [15:23] haaaaa [15:24] 🤦 [15:24] mborzecki: uhhhh no we don't seem to run them in github actions that is [15:25] perhaps we still run them in nightly travis runs? cachio do you know if we run the spread tests in ./cmd/snap-confine/spread-tests anywhere? [15:30] mmh, I don't see anything mentioning them [15:39] Before snap-confine was merged with snapd it used to have its own spread [15:39] tests [15:39] that's from commit e3d6f6aae6bb48e585ff3de1a25277b680e736b4 [15:39] https://github.com/bboozzoo/snapd/commit/e3d6f6aae6bb48e585ff3de1a25277b680e736b4 [15:41] * cachio lunch [15:42] mborzecki: I updated #9720, could you take a look? thanks [15:42] PR #9720: many: rename disks.FindMatching... to FindMatching...WithFsLabel and err type [15:47] ijohnson: +1, little copy paste typo :P [15:49] thanks mborzecki ! [16:25] PR snapd#9726 opened: cmd/snap: suppress a case of spurious stdout logging from tests [16:47] is it possible to persistently set some env variable for specific snap by user other than editing desktop file? [16:50] PR snapd#9727 opened: seed: cleanup/drop some no longer valid TODOS, clarify some other points [16:51] vidal72[m]: not atm, it would need some discussion because it kind of sidesteps/overlap with the configuration support for snaps [17:10] PR snapd#9728 opened: Disable user creation on system-user auto-import [17:16] Hey my /var/lib/snapd/hostfs/ is empty, is that expected ? Also how can I check what my snap could "see" in that directory ? [17:19] om26er: /var/lib/snapd/hostfs will be empty if you are running outside of snap confinement [17:20] While trying to run a strict snap on Nvidia Xavier NX, my snap won't start, but if I copy my libGL* files to /var/lib/snapd/lib/gl then it starts fine. [17:20] om26er: that directory has "/" bind mounted to it when snaps run [17:20] om26er: so whatever is in "/" is what will show up in /var/lib/snapd/hostfs [17:21] hm, interesting. So that means the opengl interface might need some update(s) to support Yocto based environments [17:21] ? [17:22] well you mentioned /var/lib/snapd/lib/gl and /var/lib/snapd/hostfs [17:22] /var/lib/snapd/hostfs should not need any changes for yocto since that's just a simple bind mount [17:23] /var/lib/snapd/lib/gl will most likely require changes to snap-confine to do extra mounts depending on where the gl files are on the host on yocto, yes [17:23] is it possible that my snapd is not able to see /var/lib/snapd/hostfs, due to a bug or something ? [17:23] what do you mean that your snapd is not able to see /var/lib/snapd/hostfs ? [17:24] I mean't is it possible that due to a bug in snapd/confine, my snaps are not able to "access" stuff on the rootfs ? [17:25] om26er: what do you see when you run `snap run --shell . -c "ls -lah /var/lib/snapd/hostfs"` ? [17:25] root@jetson-xavier-nx-devkit:~# snap run --shell kivy-sample -c "ls -lah /var/lib/snapd/hostfs" [17:25] ls: cannot open directory '/var/lib/snapd/hostfs': Permission denied [17:28] also: on that system libGLESv2.so.2 is under /usr/lib/libGLESv2.so.2 which isn't the case on Ubuntu at least. [17:29] also when I run `strace`, it's not able to find this directory as well https://github.com/snapcore/snapd/blob/master/interfaces/builtin/opengl.go#L56 -- it clearly exists on the host [17:29] s/directory/file [17:31] ijohnson btw, that "permission denied" also happens on my desktop, so looks like its "expected" ? [17:32] om26er: oh yeah you will need to run a devmode snap, can you reinstall the snap in devmode [17:32] and then try that same command again [17:33] yeah, now it prints the directory structure [17:34] well in devmode, my snap starts anyways, I have been working on enable confinement though [17:35] PR snapd#9729 opened: bootloader/lkenv: specify backup file as arg to NewEnv(), use "" as path+"bak" [17:35] actually I take that back, it does't work in devmode as well unless I put the libGL* files under `/var/lib/snapd/lib/gl/` [17:41] btw, these are the minimal set of .so files that I copied from /usr/lib to /var/lib/snapd/lib/gl/ for the Kivy to work from a snap environment https://gist.github.com/om26er/b4aaf190f5ce3629067bd18b5ff502f2 [17:45] PR snapd#9715 closed: hookstate: add new HookManager.EphemeralRunHook()

[18:11] hello. i'm installing snapcraft in a ubuntu 20.04 docker container (RUN apt install snapcraft...), and it seems it has issues connecting to snap store while running a debian preinstall script. It tries asking me on a TTY what to do, but there is no terminal. How can install it without user interaction? [18:11] https://code.videolan.org/-/snippets/1301/raw is the log output [18:31] thresh: you shouldn't be using snapcraft as a deb package, you should be using snapcraft as a snap, there's examples of how to do this from inside a docker container on snapcraft.io/docs === ijohnson is now known as ijohnson|lunch [18:46] thanks ijohnson|lunch [18:46] why does https://snapcraft.io/docs/build-on-docker mention FROM ubuntu:bionic as builder, but later on the Dockerfile it's using FROM ubuntu:xenial ? [18:47] X-Ubuntu-Series: 16 is also suspicious wrt core18... === pedronis__ is now known as pedronis [18:54] "X-Ubuntu-Series: 16" is normal even for core18 [18:54] (yes, this is a bit confusing. It's more of an API version nowadays) [19:11] PR snapd#9730 opened: hookstate: refactor around EphemeralRunHook [19:26] PR snapd#9731 opened: daemon: split out snapctl support and snap configuration support to their own api_*.go files [19:46] PR snapd#9726 closed: cmd/snap: suppress a case of spurious stdout logging from tests [19:51] cjwatson, ah yeah, it's still a bit weird that it doesnt fetch core18, while says so... I've found https://github.com/snapcore/snapcraft/blob/master/docker/stable.Dockerfile which helped be to figure it out [19:58] thanks a lot and have a good day :) [21:30] * cachio afk