[03:11] PR snapd#6632 opened: tests: restore sbuild test [05:30] Good morning [06:10] morning [06:37] o/ [06:37] hey mborzecki :) [06:37] zyga: hey [06:38] love when the PRs are green [06:39] like the smell of fresh coffee in the morning [06:39] I restarted a few PRs last night [06:50] all but one of mine are green :) [06:51] https://pact.kadena.io/example/Accounts [06:52] neat stuff [07:15] btw. today is the first day of spring for those in the northern part of the globe [07:15] must explain my back pain :/ [07:17] zyga: what's with the lispy thing you linked? [07:17] it's a formal verification engine that works over the web [07:17] splice an algorithm [07:17] add constraints [07:17] and it formally proves it (or provides a violating input set) [07:18] https://twitter.com/_wjmartino_/status/1105496676754751494 <- example [07:28] * zyga goes to the office [07:28] ttys (talk to you soon :) === pstolowski|afk is now known as pstolowski [08:04] heyas [08:05] pstolowski: hey [08:06] coffee time [08:13] hey pawel :) [08:16] * zyga will work on some debugging today [08:40] mvo: morning [08:45] hey mborzecki [08:45] mborzecki: did I miss anything ? [08:45] mvo: nope [08:47] hmm, google:ubuntu-18.04-64:tests/main/desktop-portal-filechooser strikes again? [08:47] booo [08:47] ok [08:48] PR snapd#6632 closed: tests: restore sbuild test === pachulo_ is now known as pachulo [09:25] * zyga looks at https://github.com/snapcore/snapd/pull/6485 [09:25] PR #6485: interfaces/seccomp: regenerate changed profiles only === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [09:52] mborzecki: reviewed [09:58] zyga: thanks [10:06] zyga: mborzecki: I added a comment there [10:07] zyga: also shouldn't the system-key mechanism in part prevent trying to use old .bin ? [10:12] pedronis: isnt't that in snap? [10:12] mborzecki: ? [10:12] it's the wait in snap-confine that zyga mentioned [10:12] so there's this mechanism, one in snap-confine that waits for the profile to appear, and then we can fix the PR and use EnsureFileState or similar [10:13] pedronis: that's the wait for compiled bpf profile [10:13] mborzecki: yes, ok, then we agree, and yes, some form of atomic rename, EnsureFIleState would be better [10:13] for that case [10:13] iirc there was another mechanism in snap (?) that checked system-key [10:14] mborzecki: there's a mechanism that also wait if there's a system-key mismatch I think [10:14] in snap run [10:14] yes [10:14] likely [10:15] ok, will push an update, got to figure out the gadget structure overlap first [10:15] there's also some weird assumption about offset-write that isn't really clear [10:15] mborzecki: thx [10:22] mborzecki: I think ubuntu-image also has a readme or some design doc or something that also explains some of this (not sure if that is helpful for your specific case though) [10:24] mvo: i think i got it, idk maybe i'm bit too much used to anaconda/mic/wic way of specifying paritions/offsets and have some prior assumptions that don't hold up with u-i and gadget.yaml [10:25] * mvo nods [10:49] mborzecki, pedronis: I added some comments there as well [10:49] I think the behavior of snap-run and snap-seccomp is important context for the reader there [10:50] watchdog timeout in https://api.travis-ci.org/v3/job/509326704/log.txt [10:50] very interesting [10:51] Chipaca: looks like a deadlock? [10:52] snapd thinks it's waiting for reboot, system has no inkling of it [10:54] but then, because again we don't store the 'needs reboot' flag, snapd rolls back the refresh [10:58] Chipaca: we should fix that, but if we are waiting for a reboot and there's no reboot something else is also going on [10:58] pedronis: yeah, the reboot happens later due to the revert [10:59] what kicks the watchdog timer? [10:59] is that from ensure loops? [10:59] or from a dedicated goroutine? [10:59] dunno, things are weird in that log [11:00] zyga: we have a dedicated goroutine [11:00] in main [11:00] so that would have to die [11:00] that's odd [11:00] dedicated goroutine in cmd/snapd/main.go [11:04] well [11:04] not too odd [11:04] we stop poking the watchdog if the daemon is Dying() [11:04] ahh [11:05] and the first thing daemon.Stop does is tomb.Kill [11:07] mvo: 2.38 is on releases [11:07] is that the day we package? [11:08] zyga: we can wait for the beta validation but it should be fine [11:08] ok [11:08] tomorrow then [11:08] thank you :) [11:08] mborzecki: I'll do debian, fedora and suse [11:08] zyga: yeah - or later today, more automation makes it quicker [11:09] heh, someone already flagged snapd out of date in aur [11:50] mvo, errm [11:50] I think your releasing automation is broken [11:50] this is the second release I've looked at with a corrupted changelog entry [11:50] https://github.com/snapcore/snapd/commit/7d3222250d98ff1baf8ad4e7df283b40a35d960c [11:51] Son_Goku: corrupted how? [11:51] it's missing the parent level "- New upstream release " in the changelog entry [11:51] Son_Goku: uh, sorry :/ [11:52] it's alright [11:52] I've been fixing it locally and couldn't get ahold of you to let you know, so that's my bad :( [11:52] Son_Goku: thanks for letting me know now :) [11:52] and sorry again [11:52] I'm looking at the fedora 2.38 update now [11:53] Son_Goku: thank you! [11:54] did we ever deal with the global MongoDB removal thing? [11:54] PR snapd#6612 closed: interfaces/builtin: add dev/pts/ptmx access to docker_support [11:54] mvo or am I still going to need this? https://src.fedoraproject.org/rpms/snapd/blob/master/f/1001-errtracker-neuter-error-tracker.patch [11:55] Son_Goku: why is that needed? [11:56] the mgo golang package requires mongodb itself, and mongodb was removed [11:56] Chipaca: https://src.fedoraproject.org/rpms/snapd/c/e923663082f5ce01052fe6299bbb1e0b6780bfc5?branch=master [11:56] ah, i didn't remember the first half of that [11:56] that's a bit silly [11:56] :-/ [11:56] maybe we should vendor bson itself, directly [11:56] * Son_Goku shrugs [11:57] MongoDB has been removed, effective Fedora 30 [11:57] that makes me happy, for historical reasons :-) [11:58] Son_Goku: how grumpy would fedora be if we vendored bson? [11:58] vendored as in pulled it into our tree directly, not via fancy golang vendoring [11:58] actually, it looks like my problem might be solved anyway [11:58] oh no [11:58] :-) [11:58] PR snapd#6633 opened: release: merge 2.38 changelogs back into master [11:58] it looks like they might have mutated the driver sources so that mongodb isn't pulled in as part of using the sources to build a thing [11:59] sweet [11:59] I'll need to test to see if it chokes on itself during the build or not [11:59] if not, then I can drop that patch [12:05] Son_Goku: thank you for looking at the 2.38 update, I was planning on doing that tomorrow but please go ahead :) [12:05] well, I saw the bug report filed by anitya [12:05] (aka release-monitoring.org) [12:07] Son_Goku: I saw it too [12:08] PR snapd#6634 opened: snap: add validation of gadget.yaml [12:11] the title of the forum topic: https://forum.snapcraft.io/t/layout-magic-collection/10501 [12:11] :D [12:11] (I love that people use layouts to do things they could not do before) [12:15] hoooh [12:15] we have some SELinux stuff in 2.38 [12:16] Son_Goku: some [12:16] none of it is enabled in the spec for 2.38 [12:16] so I guess I should leave it be then === ricab is now known as ricab|lunch [13:36] PR snapd#6619 closed: partition,bootloader: rename 'partition' package to 'bootloader' [13:56] PR snapd#6635 opened: client, daemon, store: search by common-id [13:57] cprov: ^ fwiw [14:02] zyga: u standup today? [14:03] oh [14:03] sorry [14:04] I muted all notifications === ricab|lunch is now known as ricab [14:51] PR snapcraft#2507 opened: build providers: improve handling in snap logic [15:18] pstolowski: what's the status of #6491? is it close to landing? [15:18] PR #6491: interfaces: hotplug nested vm test, updated serial-port interface [15:24] pedronis: zyga promised to take another look [15:42] zyga: added a comment here https://github.com/snapcore/snapd/pull/6485#discussion_r267821309 and in the code [15:42] PR #6485: interfaces/seccomp: regenerate changed profiles only [15:44] * cachio lunch === chihchun is now known as chihchun_afk [16:03] PR snapcraft#2508 opened: Build provider errors [16:06] zyga: 6575 can be landed [16:11] jdstrand, hey, are you sure that the behaviour re: empty cgroup is not a bug? See my last comment: https://github.com/snapcore/snapd/pull/6623#discussion_r267811338 [16:11] PR #6623: interfaces/builtin/opengl: allow access to Tegra X1 [16:19] Chipaca: could you look at #6591 when you have a bit of time [16:19] PR #6591: overlord/ifacemgr: basic measurements [16:26] pedronis: ye [16:27] thx [16:32] zyga: I did another pass on some of your PRs [16:32] pedronis: thank you! [16:42] 2.38 is in disco! [16:51] PR snapd#6609 closed: snap/gadget: introduce volume update info [16:55] PR snapd#6636 opened: cmd: prevent umask from breaking snap-run chain [17:02] pedronis: any objections to merge https://github.com/snapcore/snapd/pull/6575 [17:02] PR #6575: cmd/snap-confine: pass sc_invocation instead of numerous args around [17:02] zyga: no, I actually mentioned in the channel a while ago [17:03] great, sorry I must have missed that [17:03] mvo: I left an after the fact comment about 6609 [17:03] PR snapd#6575 closed: cmd/snap-confine: pass sc_invocation instead of numerous args around [17:05] pedronis: thanks, looking [17:07] pedronis: uh, thank you for the comment in 6609 - looks like I was too trigger happy [17:08] mvo: not the end of the world [17:15] pedronis: I updated https://github.com/snapcore/snapd/pull/6597 [17:15] PR #6597: cmd/snap-update-ns: refactor of profile application (1/N) [17:15] I hope this part is ready now [17:16] zyga: I'll look in the morning at this point [17:22] PR core18#63 closed: [RFC] snapcraft.yaml: add current date to core18 rev [17:25] PR core18#121 opened: Make the version number date-based [17:42] pedronis: sure, thank you [17:42] * zyga EOEs [17:43] have a nice evening everyone [17:43] * zyga goes to study [17:44] jdstrand: hello, can you please enqueue https://github.com/snapcore/snapd/pull/6502 -- it is the next feature work I need your review for [17:44] PR #6502: dirs,overlord/snapstate: add Soft and Hard refresh checks [17:53] zyga: it is already enqueued [18:01] EOD from me [18:13] PR snapcraft#2507 closed: build providers: improve handling in snap logic [18:13] PR snapcraft#2508 closed: Build provider errors [18:28] PR snapcraft#2509 opened: build providers: initial support for LXD [20:40] PR snapcraft#2510 opened: Release changelog for 3.3