[06:10] morning [06:33] good morning [06:41] zyga: hey [06:41] :-) [06:41] an easier day today [06:41] only training today and tomorrow [07:04] morning [07:20] pstolowski: hey [07:22] pstolowski: any PRs i should look at? [07:23] mvo: restarted the github jobs in https://github.com/snapcore/snapd/pull/10207 unit tests failed on gpg (again) [07:23] PR #10207: vendor: bump github.com/canonical/go-tpm2 revision [07:25] mborzecki: hey, https://github.com/snapcore/snapd/pull/10205 should be straightforward [07:25] PR #10205: cmd/snap, client: snap remove-quota command [07:25] pstolowski: ha ok, i'm already looking at it ;) [07:26] thanks! [07:32] pstolowski: +1 [07:32] ty! [07:33] pstolowski: #10208 is waiting for #10204? [07:33] Bug #10208: Needs rebuild with new libneon [07:33] PR #10208: daemon: implement REST API for quota groups (create / list / get) <⛔ Blocked> [07:33] Bug #10204: Add new launcher to desktop won't take effect after a restart [07:33] PR #10204: o/servicestate/quota_control.go: introduce basic group manipulation methods [07:34] mborzecki: yes [07:34] mborzecki: but also needs reworking a bit [07:34] ah. ok, so i'll skip it for now [07:34] mborzecki: so don't review it (but 10204 should be ready for review) [07:35] ok [07:42] pstolowski: keep me update when there is anything new to review or if I can help in any other way. I look at 10204 in the meantime [07:42] pstolowski: and good morning :) [07:44] mvo: will do, thanks [08:07] PR snapcraft#3516 closed: github: update workflows that to use pull_request_target [08:41] mvo: next up for review - https://github.com/snapcore/snapd/pull/10210 - last commit is the new code [08:41] PR #10210: cmd/snap, client: snap quotas command [08:41] PR snapd#10210 opened: cmd/snap, client: snap quotas command [09:21] PR snapd#10211 opened: o/servicestate/quota_control.go: introduce (very) basic group manipulation methods [09:34] mborzecki: hi, could you review https://github.com/snapcore/snapd/pull/10203 [09:34] PR #10203: image,c/snap: implement prepare-image --customize [09:38] pstolowski: https://github.com/snapcore/snapd/pull/10210 and https://github.com/snapcore/snapd/pull/10205 are ready for review? [09:38] PR #10210: cmd/snap, client: snap quotas command [09:38] PR #10205: cmd/snap, client: snap remove-quota command [09:38] pedronis: yes. remove is first. quotas command is just last commit on top of remove [09:39] thx === alan_g_ is now known as alan_g [09:57] PR core20#102 closed: hooks: add mDNS hostname resolution support [10:25] mvo: can you merge https://github.com/snapcore/snapd/pull/10205 (unrelated failures)? [10:25] PR #10205: cmd/snap, client: snap remove-quota command [10:37] pstolowski: sure [10:41] PR snapd#10205 closed: cmd/snap, client: snap remove-quota command [10:43] pedronis: i've rebased #10210 [10:43] PR #10210: cmd/snap, client: snap quotas command [10:43] Bug #10210: Second head doesn't work with Mobility Radeon 9600 M10 [10:52] mborzecki: when you have some spare cycles, could you please check 10203? [10:53] mvo: sure, i'll take a look now [11:04] * pstolowski lunch [11:29] pstolowski: is the quota groups tree depth limited to 2? [11:30] iow just groups & subgroups right? [11:42] mborzecki: no [11:43] you can have any depth (we should have a limit and the system probably has, but is not 2) [11:58] hi everyone, i have a quick question. in the latest ubuntu core 20 image that i prepared, no matter what i do. cloud-init never runs on boot. [11:59] i placed it as cloud-init.yaml and cloud.conf but whenever i boot into core20, i get the generic network setup and ubuntu login ncurses ui [12:06] PR snapd#10212 opened: network-setup-control: whitelist netplan generator (LP: #1926442) [12:17] mborzecki: the cli receives complete hierarchy but flattened, in case the question was triggered by snap quotas pr ;) [12:18] pstolowski: yeah, i was slightly confused by the presentation [12:20] clmsy: did it work before? depending on the grade of the model you use you might have to enable cloud-init explicitly. for security reasons some ways to drive cloud-init are not supported for the signed/secure grade of models [12:46] pedronis: do you think we should fail early if the recovery system directory already exists? [12:48] mborzecki: yes [12:49] pedronis: ok, i'll add a check to seedwriter then [12:49] degville: thanks for your input in https://github.com/snapcore/snapd/pull/10196 , I had to counterpropose something though because as you said something wasn't clear, can you look there again [12:49] PR #10196: o/devicestate,o/hookstate/ctlcmd: introduce SystemModeInfo methods and snapctl system-mode [12:49] degville: here: https://github.com/snapcore/snapd/pull/10196#discussion_r623016946 [12:59] pedronis: thank you! I'll take a look! === mup_ is now known as mup === mup_ is now known as mup [13:55] pedronis: i've updated #10181, there's a bunch of seedwriter error code paths that aren't test but afaiu, those are only triggered if operations are called in the wrong order [13:55] Bug #10181: Usability: Cursor notifications doesn't work [13:55] PR #10181: overlord/devicestate: add helper for creating recovery systems at runtime [13:55] mborzecki: thx [13:56] degville: I pushed my changes there, for expedience, but it means you double check if I applied your other suggestions properly [13:56] *you can [13:56] pedronis: yes, absolutely - thanks for letting me know. [13:58] pedronis: just looked and it all looks good! thank you! [13:58] thx [14:09] re [14:21] pedronis: in https://github.com/snapcore/snapd/pull/10211#discussion_r622952750, to be clear, you mean if we try to update the memory limit for a sub-group to be greater than the limit allowed by the parent group ? [14:21] PR #10211: o/servicestate/quota_control.go: introduce (very) basic group manipulation methods [14:21] ijohnson: yes [14:22] ok, thanks for clarifying [14:33] ijohnson: I tried to answer some of your questions there [14:34] ok, thanks that makes more sense now [14:35] ijohnson: sorry, two of my comments were more statement than feedback asking for a change, though they might mean that we need a comment nearby [14:37] hi all [14:41] In case someone has time to look at a few (ok, quite a few) questions regarding first-boot / seeding UX, I just opened https://forum.snapcraft.io/t/first-boot-device-initialization-ux/24152 – thankful for helpful comments [14:43] hi dot-tobias we are on a tight deadline at the moment, so it will be a few days before anyone can devote a lot of attention (well myself at least) to the post, but I have got it in my backlog to look at when I have more time [14:44] dot-tobias: what ijohnson said, but a short thing: detect from within a snap if the seeding process is still running?, yes we are adding something like that atm [14:44] mvo, pedronis i've updated and unblocked #10208 (it's going to fail now due to missing servicestate methods) [14:44] Bug #10208: Needs rebuild with new libneon [14:44] PR #10208: daemon: implement REST API for quota groups (create / list / get) [14:45] nice thanks pstolowski [14:46] ijohnson: btw i'll work on a spread test for cli [14:46] awesome thank you! [14:46] ijohnson, pedronis: Thanks for the quick feedback, and I hope you don't need to pull too many night shifts for this deadline 😊 Great to hear that seeding detection is already underway! [14:47] bad mup for highlighting noise [14:57] pstolowski: \o/ thank you [15:32] PR snapd#10213 opened: tests: basic spread test for snap quota commands [16:14] pedronis: the PR is updated now with more tests and checks and the things you mentioned [16:15] ijohnson: thanks, I'll look at it in a bit [16:20] ijohnson: which one, 10211 ? [16:21] pstolowski: #10211 [16:21] Bug #10211: nautilus crash when writing on smbfs [16:21] PR #10211: o/servicestate/quota_control.go: introduce (very) basic group manipulation methods [16:21] thx [16:21] pstolowski: let me know if you can update your pr or if it's getting too late, I can help and pickup things if needed [16:21] pstolowski: but hopefully very straightforward :) [16:21] mvo: will do it [16:22] thanks [16:25] done [16:47] mvo, ijohnson i though we wanted to leave UpdateQuota out for the first iteration? [16:47] *thought [16:48] pstolowski: we will [16:48] pstolowski: it will be a followup, just ignore it for now [16:48] my understanding was it will be left out of the daemon [16:48] ok, got confused reviewing the PR [16:48] mvo: oh should I have removed servicestate.UpdateQuota entirely? [16:48] * ijohnson is very confused again [16:53] ijohnson: no I think it's fine [16:53] ijohnson: unless I'm confused [16:53] ijohnson: I just wanted to unexport it [16:53] ahhhhhhhhhhhhhhhhhh [16:54] ijohnson: but we can do that in a followup, I think/hope this in the current form is uncontroverial [16:54] mvo: you mean UpdateQuotas note the s [16:54] ijohnson: oh, sorry, yes [16:54] ijohnson: I think [16:54] I thought you were talking about UpdateQuota (note the no s) [16:55] ok, I think we are in alignment, UpdateQuota (which will be patchQuotas in the future) will be in the PR as-is, and we will un-export that function in the followup [16:55] ijohnson: sorry, my bad [16:55] ijohnson: PatchQuotaState was hte one I had in mind [16:55] ijohnson: yeah :) looks like we are aligned [16:55] the full thing, servicestate.UpdateQuota(st *state.State, name string, updateOpts QuotaGroupUpdate) will stay as-is in the PR [16:56] ijohnson: yes, I think so [16:56] great yeah we're good then [18:02] PR snapd#10204 closed: o/servicestate/quota_control.go: introduce basic group manipulation methods <⛔ Blocked> [18:03] PR snapd#10211 closed: o/servicestate/quota_control.go: introduce (very) basic group manipulation methods [18:38] PR snapd#10214 opened: features,servicestate: add experimental.quota-control flag === JanC is now known as Guest66672 === JanC_ is now known as JanC [20:53] PR snapd#10196 closed: o/devicestate,o/hookstate/ctlcmd: introduce SystemModeInfo methods and snapctl system-mode