[01:58] PR snapd#9872 opened: overlord/servicestate: expose dbus activators of a service === philroche_ is now known as philroche [06:59] morning [07:27] mvo: hey [07:29] good morning mborzecki ! [07:30] school run [07:34] PR snapd#9869 closed: cmd: make string/error code more robust against errno leaking (2.49) [07:39] PR snapd#9861 closed: gadget: run resolveContentPaths() when laying out a volume [07:39] PR snapd#9873 opened: gadget: add new ResolvedContent and populate from LayoutVolume() [07:57] re [08:00] mvo: something is off in https://github.com/snapcore/snapd/pull/9870 [08:00] PR #9870: vendor: update go-tpm2/secboot to latest version [08:02] mborzecki: yeah, it's strange, no idea yet (but haven't digged, jumped into dtb) [08:02] mvo: i can take a look once i get my morning coffee [08:08] yeah, unit tests don't build even [08:10] morning [08:19] mvo: looks like a problem could be in the secboot package itself: https://paste.ubuntu.com/p/XC86YRCRFD/ [08:19] pstolowski: hey [08:26] mborzecki: makes sense, seems like chris needs to update some code? [08:26] good morning pstolowski [08:33] mvo: yes, do you recall what is the fix in go-tpm2 we want to pull in? [08:36] mborzecki: iirc there was a fix to be less strict in the tpm2 checks or differently strict that fixes uc20 install on davids test PC [08:37] hmm maybe there's a PR still up [08:38] mvo: this PR most likely: https://github.com/snapcore/secboot/pull/108 [08:38] PR secboot#108: Make AddEFISecureBootPolicyProfile less strict [08:39] mborzecki: yeah, that's what I remember. maybe we just wait for chris to come online? [08:43] mvo: ok, i think i've found a set of revisions that works [08:44] mborzecki: hm, ok [08:44] mvo: i'll push it to your branch and leave links to the relevant commits [08:44] PR snapd#9874 opened: gadget: remove resolvedSource from VolumeContent [08:50] mvo: https://github.com/snapcore/snapd/pull/9870#issuecomment-768898417 [08:50] PR #9870: vendor: update go-tpm2/secboot to latest version [10:05] PR snapd#9875 opened: gadget: use ResolvedSource in MountedFilesystemWriter [10:57] mvo: it'd be nice to include https://github.com/snapcore/snapd/pull/9744 in 2.49 if we can get security team to look at it (cc amurray), wdyt? [10:57] PR #9744: OpenGL interface: Support more Tegra libs [10:57] (added a milestone) [10:58] mborzecki: +1 [10:59] mborzecki: I gess we could try to ping alex directly in the PR [11:22] mvo: https://github.com/snapcore/snapd/pull/9865 can be merged and cherry picked, the failure on 16.04-32 is unrelated [11:22] PR #9865: interfaces/browser-support: allow sched_setaffinity with browser-sandbox: true [11:51] mborzecki: cool, will do [11:52] mborzecki: merged and cherry-picked [11:52] mvo: thank you! [11:55] PR snapd#9865 closed: interfaces/browser-support: allow sched_setaffinity with browser-sandbox: true [11:57] meh, it's snowing again [13:05] how can I remove all snapshots? `snap forget *` does not work. [13:11] PR snapcraft#3431 closed: godeps spread test: use latest/stable go snap [13:21] oxek: little for loop over `snap saved | awk {..}` output? [13:56] PR snapcraft#3432 closed: spread tests: remove legacy plugin tests [15:13] errands time === alan_g_ is now known as alan_g [15:46] PR snapd#9855 closed: [RFC] gadget: enable sector-sizes in gadget.yaml; refactor SectorSize handling <â›” Blocked> [15:51] PR snapd#9876 opened: Add vcu iface [15:54] cachio: I released core 2.49~rc1 to beta now too, so this should also be ready for validation [17:16] mvo: any reason not to merge #9856? I see there's discussion about the signature but doesn't seem a blocker and could be changed later [17:16] PR #9856: gadget,overlord: pass kernelRoot to install.Run() [17:16] yeah the signature I don't think is a blocker on that pr [17:21] pedronis: no reason, I just did not see the second +1 yet, had a very busy day [17:22] * cachio lunch === Ringtailed-Fox is now known as RingtailedFox [17:22] hi, is the network-manager interface pretty much the only way for a snap to receive network manager events like connectivity change? [17:28] bandali: there's both network-manager plug and the network-manager-observe plug, I'm not sure if there's a specific access you are looking for [17:31] ijohnson, i see. i'm not entirely sure myself, but i'm told we need to be able to detect and act upon wifi/ethernet/vpn changes [17:52] bandali: well I would try first with network-manager-observe and see if that works for your snap === ijohnson is now known as ijohnson|lunch [18:07] right; i'll try with that (again); but i seem to recall it not being enough [19:27] ijohnson|lunch, hi [19:27] I see the unit tests run much slower in xenial than in bionic [19:27] do you see any contrain to run those tests in bionic? [19:32] hey guys [19:32] what a long day [19:32] how are things? [19:33] zyga, hey [19:33] hey :-) [19:33] almost on vacations :) [19:33] still need to work on the log analyzer [19:33] thanks for the review but I was fixing other stuff and couldnt address that yet [19:35] oh, nice, what are you going to do on holidays? [19:36] that's fine, I'm stretched think anyway, I cannot review things quickly myself [19:36] feel free to ping me as a reminder [19:37] PR snapd#9877 opened: tests: run unit tests in Focal instead of Xenial [19:38] cachio do you know why there's such a big difference in speed? [19:39] but I'm not sure this is a good idea [19:39] I'll comment on the PR [19:40] done [20:00] back === ijohnson|lunch is now known as ijohnson [20:08] hey cachio [20:09] cachio: I think it's probably fine to run the unit tests in focal instead of xenial, if it's faster seems fine to me [20:09] hey ijohnson [20:10] hey zyga [20:12] zyga: I commented on that pr, I disagree with your request for changes, I think the pr is fine [20:13] ijohnson hmm [20:14] ijohnson do I understand incorrectly that the snapd build is now happening on focal? [20:14] zyga: the github action we use to build the snap uses LXD to build the snap [20:14] ohhh [20:14] I missed that [20:14] I will revise my review [20:14] thanks! [20:14] thank you for pointing that out [20:15] done [20:20] zyga, no idea [20:21] but in github machines the time is similar [20:21] the difference is in gce [20:23] still trying to see if there is any way to run this checks in parallel at least because is taking 15/6 minutes [20:52] cachio: this morning (or maybe yesterday I don't remember) you mentioned something about uc20-recovery task failing now? it should be passing now everywhere the kernel has been updated with everything for a while now [21:32] PR snapd#9878 opened: tests/core/uc20-recovery: move recover mode helpers to generic testslib script [22:05] ijohnson, hi, sorry for the delay [22:05] no worries [22:05] uc20 recovery test is failing [22:06] it failed on pi3 pi4 and amd64 [22:06] let me check the errors [22:08] ijohnson, this is pi3 https://paste.ubuntu.com/p/9CwtrYwWrq/ [22:08] I remember the problem was related to an ip change [22:08] right? [22:09] cachio: let me have a look [22:09] cachio: also I just noticed that our github actions spread runners are almost out of space in canonistack can you take a look before you eod ? https://pastebin.ubuntu.com/p/9g6RfyFMz8/ [22:10] cachio: as to those logs, it could be related to a different pr, which kernel + snapd snaps are you trying with [22:10] ijohnson, yes, I need to add a new script to clean up the jobs [22:11] cachio: ack [22:11] because the runner logs are huge [22:12] ijohnson, this is amd64 https://paste.ubuntu.com/p/K6FQzvy62Y/ [22:12] pretty the same [22:13] cachio: hmm yeah it does appear to be the same, when this happens can you check the IP address of the external device manually? if it is different then we have a bug/regression somehow [22:13] cachio: I don't have time this afternoon, but I can try to run the external tests tomorrow to see if I can reproduce [22:13] for that I need to manually reserve the device and run the test [22:13] I can do it [22:53] PR snapd#9879 opened: osutil/many: make all test pkgs osutil_test instead of "osutil" [23:03] PR snapd#9880 opened: tests/lib/fakestore: support repair assertions too === niemeyer_ is now known as niemeyer === boxrick_ is now known as boxrick