[02:31] PR snapcraft#3369 closed: sources: enable 7z, bzr, hg, svn, zip for non-linux [06:37] morning [07:34] good morning [07:35] zyga: hey [08:03] quick errand, back in 30 [08:20] morning [08:44] re [08:50] hey [09:27] anyone looking at fixing tests broken by cgorup v2 on sid? [09:28] pstolowski: mvo: trvial comment fix https://github.com/snapcore/snapd/pull/9826 can you take a look? [09:28] PR #9826: interfaces/apparmor: drive by comment fix [09:28] sure [09:29] PR snapd#9826 opened: interfaces/apparmor: drive by comment fix [09:34] thanks! [09:40] heh, found some apparmor related deadcode too [09:44] mborzecki: sure thing, in a meeting right now but will do when this is finished [09:50] pstolowski: another trivial one: https://github.com/snapcore/snapd/pull/9827 [09:50] PR #9827: interfaces/builtin: more drive by fixes, import ordering, removing dead code [09:50] ackk: hi, https://bugs.launchpad.net/snapd/+bug/1901212 is yours right? [09:50] Bug #1901212: snaps break after upgrade, snapd claims meta/snap.yaml is missing [09:51] pstolowski, yes [09:52] ackk: i guess it's safe to assume unclean/abrupt shutdowns or fs corruption can be ruled out right? [09:52] pstolowski, correct. I just found them broken at some point. I don't have any broken right now, though [09:53] ackk: may i ask you to send me your /var/lib/snapd/state.json next time you see this problem (you may want to remove your store macaroon from it before sending)? [09:54] pstolowski, sure, can you comment on the bug ? [09:54] PR snapd#9827 opened: interfaces/builtin: more drive by fixes, import ordering, removing dead code [09:56] ackk: thanks. done. [09:56] ps [09:56] pstolowski, ty [10:59] PR snapd#9826 closed: interfaces/apparmor: drive by comment fix [11:14] PR snapd#9821 closed: tests: skip interfaces-openvswitch spread test on debian sid [11:14] PR snapd#9824 closed: interfaces/greengrass-support: back-port interface changes to 2.48 [11:19] PR snapd#9827 closed: interfaces/builtin: more drive by fixes, import ordering, removing dead code [11:19] PR snapd#9828 opened: sandbox/cgroup, tess/main: fix unit tests on v2 system, disable broken tests on sid [11:59] pedronis: hi, do you think you'll be able to take a look at https://github.com/snapcore/snapd/pull/9771 this week? [11:59] PR #9771: boot: boot config update & reseal [11:59] mborzecki: yes, either today or tomorrow [11:59] pedronis: great, thank you [13:10] mvo: shall I push the change with gzip compression of the install log to #9545 ? [13:10] PR #9545: devicestate: keep log from install-mode on installed system [13:20] PR snapd#9829 opened: tests: cleanup the run-checks script [13:33] mborzecki: yeah, please do [13:33] mborzecki: I can do to but busy with $stuff :/ [13:33] mvo: ok, spread test is running now, i'll push when it's done [13:34] mborzecki: \o/ [13:42] mvo: and pushed [13:43] nice [14:27] did anythig change WRT hook handling in UC20 ?`i end up with weird apparmor denials and a non-functioning hook [14:28] cachio: i tried to run 21.04 image with your PR (also tried with the specific image you found was ok) but spread was failing (on prepare i think) [14:31] https://paste.ubuntu.com/p/V2NHdz9tFd/ if anyone has an idea why the configure hook would nowadas hard-require CAP_NET_ADMIN without reason, please speak up :) [14:33] (and no i is neither date, not md5sum or tr that trigger the denial, the code was changed several times and still triggers the denial) [14:42] ogra: can you reproduce only on pi, or also on amd64 ? [14:44] i have not tried on amd64 yet [14:47] (that snap does not run on non-pi HW and i have not really used any x86 UC20 yet) [14:47] it smells like something else beyond the hook is executed by snapd or so ... [14:48] well the denial has snap-exec as the comm [14:48] so it's not your hook, it's something in snapd [14:48] right [14:49] the snap has the desktop plug definied (but indeed not connected on UC20) ... i wonder if it is related to the recent fontconfig hacks [15:02] * ogra moves the hook from confgure to install ... perhaps that helps [15:18] ... doesnt ... moving it to a launch wrapper then ... that will work in any case [15:34] mvo: no sync today in case you were not following preseed channel [15:41] pstolowski: yeah, just got the mail, thank you! [15:42] pedronis: just fyi, I updated 9819 and moved the test to a different snap [15:42] pedronis: eh, moved the test to a different file :) [15:42] thx [15:42] I'm going back to reviewing now [15:53] ijohnson, so it seems snapctl also causes this denial (i have a wrapper calling "snapctl is-connected", disconnecting network-control for the app causes the same denial on app restart) [15:53] yeah not surprised that's what I would have expected [16:28] pstolowski,mvo: #9513 needs a 2nd review from sombody that didn't look at/touch it yet [16:28] PR #9513: snapshotstate: detect duplicated snapshot imports [16:29] yes [16:32] maybe ijohnson ? [16:35] sure I'll have a look [16:37] thanks! [16:50] pstolowski: I asked some questions in #9820 [16:50] PR #9820: o/snapshotstate: handle conflicts between snapshot forget, export and import [16:52] cachio, hey [16:52] cachio, I was pretty busy with office work today but I'll send some more fixes for spread in my late evening [16:53] pedronis: thanks. i'm tweaking this right now, found an issue causing spread test failures [17:13] zyga, hey, nice [17:13] I have time for that today [17:13] I can review that for sure [17:17] PR snapcraft#3411 opened: repo: install apt build packages marked for installation [17:26] * cachio afk === ijohnson is now known as ijohnson|lunch [18:57] PR snapcraft#3411 closed: repo: only install build packages marked for installation [20:27] PR snapd#9822 closed: tests: new action-add-labels workflow to autotically tag a PR with "Run Nested" <â›” Blocked> === ijohnson|lunch is now known as ijohnson [22:38] cachio, https://github.com/snapcore/spread/pull/112 [22:38] PR spread#112: Bump base version of go to 1.13 [22:39] cachio, trivial bump, just to see if it passes [22:39] it worked for me locally (for go get) and then failed on the go vet issues [22:41] cachio, if you agree we should prepare a single branch that has 1.10 -> 1.13 bump, and two go vet fixes and propose that instead, it should pass ci [23:28] PR snapcraft#3412 opened: grammar: ensure all dictionary primitives are captured