=== KindTwo is now known as KindOne [04:47] morning [04:57] PR snapd#8712 opened: o/devicestate: typo fix [05:10] PR snapd#8713 opened: update avahi-daemon labelling to allow simply "avahi-daemon" [06:46] mvo: hey [06:47] mvo: how was the long weekend? [06:47] hey mborzecki - it was quite nice, thank you ! how are you? what did I miss? [06:48] hi mvo [06:50] hey jamesh ! nice to see you [06:50] mborzecki, jamesh looks like 8713 should get cherry-picked back into 2.45 (?) [06:52] mvo: yes, i think so, from what i can tell by looking at aa versions, bionic is likely the last release that had /usr/bin/avahi-daemon as the label, but i haven't checked the actual profile content [06:53] mborzecki: woah, so this went undetected for quite some time [06:53] mvo: i did not check the actual profile content, maybe it was distro patched to use the old name [06:54] mvo: the profile is shipped separately in the apparmor-profiles package, which we don't install by default. [06:55] jamesh: so without the profiles package we'd get unconfined, right? that'd explain why it went unnoticed [06:55] mborzecki: yes [06:59] good morning [07:00] hey jamesh [07:00] hey mborzecki [07:00] hey mvo :) [07:01] hey zyga [07:01] hey zyga ! good morning - how are you? [07:01] zyga: had a good weekend? [07:01] mvo: Friday was the best, yeah [07:01] hi zyga [07:02] mvo: offtopic: there was a bunch of weird snaps added to the store over weekend [07:02] I wonder if we have some sort of store response team to raise those to [07:02] or if we have some guidelines on phony/phishy snaps [07:02] zyga: some typosquatters? [07:03] no, they were even kind enough to namespace them all [07:03] one moment [07:03] morning [07:03] hey pawel [07:03] mvo: snap find mobius [07:03] looks like one large cryptoscam [07:04] good morning pstolowski [07:04] zyga: probably best to reach out to #snapstore [07:04] mvo: right, my question was mostly about that, is IRC the go-to method? [07:05] zyga: hm some forexa apps? that's what it says on their website: Мультиплатформенный терминал для торговли на рынке Forex [07:05] zyga: there is a mailing list iirc [07:05] mborzecki: dunno what those are, just looked weird [07:05] they all appeared in one day [07:05] zyga: but I would start with irc or a bugreport in LP against snapstore, if it looks serious maybe even tagged security [07:05] ok [07:05] there's even a windows app https://mtrader7.com/ru/terminals/windows [07:05] thanks! [07:06] my primary school days are coming to hount me with cyrlic [07:06] terminal dla windows? [07:06] hahah [07:07] and offtopic, to close the weekend [07:07] civ6 is free on epic games [07:07] fwiw, it'd be nice to be able to list the snaps from a given publisher in the snapstore page [07:07] if you like civ games that's the ultimate bragin [07:07] mborzecki: oh yeah [07:08] zyga: woah [07:08] mvo: it plays wonderfully (though on windows) [07:08] civ has grown quite a bit since my atari ST days [07:09] but the mood is really the same [07:15] zyga: cool, I have a windows partition somewhere :) [07:16] PR snapd#8712 closed: o/devicestate: typo fix [07:19] jamesh: hey, while I have you, do you think https://github.com/snapcore/snapd/pull/8708 is sensible? [07:19] PR #8708: tests: setup portals before starting user session [07:19] mvo: travis is often stuck on yellow on PRs btw [07:19] mvo: usually retriggering the test in travis (which is green on their side) fixes it [07:21] mborzecki: hi, did you want to vote need fixing on #8661 ? [07:21] PR #8661: configcore: add "service.console-conf.disable" config option [07:21] pedronis: hm? [07:21] ah [07:22] hey pedronis [07:22] pedronis: right, wanted to check in with mvo about this [07:22] pedronis: I didn't get to it last week but you can expect first wave of tool patches today [07:23] pedronis: I made progress on https://github.com/snapcore/snapd/pull/7825 [07:23] PR #7825: many: use transient scope for tracking apps and hooks [07:23] pedronis: it should be ready for tomorrow, perhaps we could discuss what needs to happen next [07:25] hmm, I think cachio may have left the 16.04-32 bit image unupdated [07:31] mvo: hi, shouldn't debian-sid rules rm snap-bootstrap like 14.04 ? [07:31] pedronis: I think we don't even build it there, let me doulbe check [07:31] pedronis: iirc during the build phase there is a "rm -rf cmd/snap-bootstrap" [07:32] ah, you are right, I'm reading the rm wrong [07:33] pedronis: no worries, it would be nice to be more uniform but the dh-golang tool does not respect buildtags accross the board. I filed a debian bugreport a while ago (with patch) but that got no traction [07:37] mvo: unrelated, do you remember why we have a specific ubuntu-17.04 packaging symlink ? [07:46] pedronis: historic reasons, we can kill all of that, actually let me do this [07:53] PR snapd#8714 opened: packaging: remove obsolete 16.10,17.04 symlinks [07:56] brb [08:02] pstolowski: mvo: hi, I tried to re-run #8697, there's a lot of weird errors, I don't understand if we broke something on master or if it's related to packaging changes in odd ways [08:02] PR #8697: packaging: build cmd/snap and cmd/snap-bootstrap with nomanagers tag [08:07] pedronis: thanks, i will keep an eye on it; i pushed some changes last friday but only run them on a subset of systems & only with the new spread test, so it is possible it breaks something [08:08] pstolowski: we really need it unblock your config PRs [08:09] pedronis: in a meeting but I can check later [08:18] pstolowski: are you working on the feedback to #8304 ? [08:18] PR #8304: usersession/userd: add zoommtg url support [08:18] sorry [08:18] pstolowski: I meant, #8704 [08:18] PR #8704: cmd/snap-preseed: improve mountpoint checks of the preseeded chroot (1/3) [08:19] pedronis: yes, i'm going to update it today [08:20] ok, thx [08:43] small breakfast and back to work [08:56] re [09:05] mvo: can you please force merge https://github.com/snapcore/snapd/pull/8708 [09:05] PR #8708: tests: setup portals before starting user session [09:05] not sure how to unstuck travis there [09:12] zyga: sure [09:12] thanks! [09:13] PR snapd#8708 closed: tests: setup portals before starting user session [09:14] mvo: some questions/comments in #8351 [09:14] PR #8351: config: apply vitality-hint immediately when the config changes [09:15] pedronis: cool, thank you [09:51] pstolowski: I reviewed #8709 [09:51] PR #8709: cmd/snap-preseed, systemd: fix handling of fuse.squashfuse when preseeding (2/3) [09:51] pedronis: thank you [10:11] PR snapd#8715 opened: tests: port interfaces-network-status-classic to session-tool [10:13] mborzecki: thanks for the review of the Pool stuff [10:26] brb [10:26] small coffee break [10:26] hmmm [10:26] 2020-05-25 10:24:19 Cannot allocate google-unstable:opensuse-15.1-64: cannot allocate new Google server for opensuse-15.1-64: invalid value for field 'resource.shieldedInstanceConfig': ''. Shielded VM Config can only be set when using a UEFI-compatible disk [10:28] mvo: ^ FYI, this is the outdated image used after deploying the new spread binary [10:28] we should discuss spread at the standup [10:35] zyga: ok [10:37] PR snapd#8714 closed: packaging: remove obsolete 16.10,17.04 symlinks === Eleventh_Doctor is now known as Eighth_Doctor [10:51] PR snapd#8716 opened: o/devicestate: refactor current system handling [10:51] pedronis: ^^ [10:53] zyga: I reworked 8508 and it uses outputs now, much nicer indeed [10:53] wow :) [10:53] zyga: please have another look, I think this is good to go now [10:53] looking [10:53] it was just a suggestion, I didn't know you'd go all the way :) [10:55] zyga: it is much nicer this way [11:00] mvo: https://github.com/snapcore/snapd/pull/8508#pullrequestreview-417632557 [11:00] PR #8508: github: run all spread systems in a single go with cached results [11:01] zyga: thanks! yeah, I think the root-owned one is fixed, I will do a followup one this one is in [11:01] * mvo didn't want to cram too much into this tiny PR [11:03] mborzecki: are you (still) working on not accepting mode not to be set in modeenv? [11:04] pedronis: yes, i'm back at that branch [11:04] mborzecki: thx [11:04] mvo: thank you! [11:10] pedronis: small step towards what we discussed https://github.com/snapcore/snapd/pull/8717 [11:10] PR #8717: test: session-tool cli tweaks [11:10] PR snapd#8717 opened: test: session-tool cli tweaks [11:12] I' afraid my dog just told me I should go for a walk :D [11:12] brb [11:37] PR snapd#8718 opened: boot, many: require mode in modeenv [11:40] pedronis: fyi, my packaging change fails because of panic: user: Current not implemented on linux/amd64 [11:40] pstolowski: ? [11:42] pedronis: something isn't included in the build and snap cli panics in some use cases (login?). i'm unclear, investigating [11:55] re [11:55] pstolowski: re [11:56] pstolowski: when I looked at godbus code there was some interesting difference between how godbus is built [11:56] pstolowski: and the code handled variants with cgo, without cgo, on various systems [11:56] pstolowski: perhaps we need somiething similar [11:57] zyga: it was caused by passing CGO_ENABLED=0 [11:59] pstolowski: https://github.com/godbus/dbus/blob/master/homedir.go https://github.com/godbus/dbus/blob/master/homedir_dynamic.go and https://github.com/godbus/dbus/blob/master/homedir_static.go [12:00] linux sound magic, after unplugging headphones there was no sound for like 30 seconds [12:00] anyway [12:00] back to work [12:01] PR snapd#8715 closed: tests: port interfaces-network-status-classic to session-tool [12:03] PR snapd#8719 opened: tests: remove dbus.sh [12:51] mvo zyga so if I just go to 8508 and try to restart one of the spread jobs there that passed it should be really quick and effectively a no-op right ? [12:51] ijohnson: looking [12:51] ah [12:51] * ijohnson is excited to try this [12:51] yes, I think so :) [12:52] if not it'd be quite funny [12:52] let's try? [12:52] * ijohnson excitedly clicks buttons [12:52] ijohnson: correct [12:53] ijohnson: I guess there are a bunch of people starring at this now :) [12:53] haha [12:55] well now it seems the jobs are stuck waiting in the queue because multiple people tried it at once [12:55] aka "insert coin" [12:56] ijohnson: not as exciting but super simple https://github.com/snapcore/snapd/pull/8719 [12:56] PR #8719: tests: remove dbus.sh [12:57] Nice I'll take a look after SU [12:57] thanks :) [12:57] it's just a rm -f [12:59] indeed it is! In that case here's your +1 [13:00] standup :) [13:28] mborzecki: https://github.com/snapcore/snapd/pull/8720/files <- any ideas on better "gimme all processes" command? [13:28] PR #8720: spread.yaml: add ps aux to debug section [13:28] PR snapd#8720 opened: spread.yaml: add ps aux to debug section [13:29] at this rate we will have beta and then some more releases till stable [13:42] zyga: gamma? [13:46] mborzecki: like adom version [13:46] hey folks any idea who's a good person to poke about snap-store (the gui client)? kenvandine maybe? (but he's off today) [13:46] roadmr: no idea, but probably ken indeed [13:48] thanks zyga :) [13:52] PR snapd#8721 opened: devicestate, sysconfig: revert support for cloud.cfg.d/ in the gadget [13:52] zyga: do you know if there is something funny going on with our gh action runners? I see that 8508 takes forever to be re-run, any idea? [14:01] mvo: looking [14:02] lots of busy runners [14:02] mvo: queue size ~10 [14:02] mvo: it should clear in 30 minutes [14:03] looks like we got a number of tests re triggered or started [14:03] oddly one stuck at -discard for 2 hours 44 minutes [14:04] and one 16.04-64 running for over 3 hours 40 minutes [14:04] but most of the queue looks happy [14:04] ah, sorry, those are zombies [14:04] perhaps actions runnner bug / not collecting zombies [14:05] note that https://github.com/snapcore/snapd/actions?query=is%3Ain_progress gives pretty good visibility [14:05] remember that each really consumes a number of runners at once === bdx9 is now known as bdx [14:29] ijohnson: 8508 did successfully run the cached results, I triggered another run to see if restarting twice works too (I see no reason why it would not but I'm paranoid) [14:29] ijohnson: but yeah, looks like this is ready [14:29] * mvo hugs mborzecki for this [14:32] yeah, it will be the best of both worlds [14:32] yay [14:33] mvo: great news! [14:33] thanks mborzecki and mvo this will be super amazing [14:34] but we fixed the random issues in the tests :P [14:37] mborzecki: haha [14:37] mborzecki: only some [14:37] mborzecki: invariant-tool is amazing because a simple few lines shows you what's left TODO [14:39] hmm why on earth unsquashfs would read /var/run/systemd/userdb which obviously has systemd_userdbd_runtime_t context [14:39] unless it's some another pam/nss/systemd thing [14:40] I bet it's pam/nss/whatnot [14:41] heh, man 8 nss-systemd, This module preferably utilizes systemd-userdbd.service(8) for resolving users and groups, but also works without the service running. [14:41] squashfs is not getting that many commits these days, I don't think code for that would have made it (but I could be wrong of course) [14:42] and we call unsquashfs -l during install to check the sanity of a snap [14:43] it's still the same user tho [14:43] ah ok, unsquashfs may be calling getpwent() or somesuch to translate uid -> name [14:44] mvo: mborzecki: I commented quickly on #8661 [14:44] PR #8661: configcore: add "service.console-conf.disable" config option [14:44] that hits nss, and eventually may reach nss-systemd which probably reads the db directly for simple queries [14:44] mborzecki: must be nss [14:46] https://github.com/snapcore/snapd/pull/8717 is simple and needs a 2nd review [14:46] PR #8717: test: session-tool cli tweaks [14:49] gentle rain outside, what a lovely sound [14:51] pedronis: thanks [14:53] mborzecki: tbc, the main comment was about no restarting things if nothing has changed [14:54] the other one is cosmetics [14:55] zyga, mborzecki: hey all [14:55] hey [14:55] how are you? [14:56] ehh, as well as I could be [14:56] day one billion of the quarantine [14:56] I talk to the napkins now :) [14:56] Eighth_Doctor: really? I must go out more [14:56] haha it's been less than 3 months :P [14:57] Schools here closed on March 12th or so [14:57] Datto closed on March 13, but I stopped one day early [14:58] PR snapd#8568 closed: asserts: rest of the Pool API [15:06] zyga: #8717 has 2 +1's and is almost entirely green [15:06] thanks, merging! [15:06] PR #8717: test: session-tool cli tweaks [15:07] PR snapd#8717 closed: test: session-tool cli tweaks [15:07] afk for some time [15:10] PR snapd#8722 opened: tests: check that host settings like hostname are settable on core [15:10] cmatsuoka: I pushed a very simple spread test, could you please check that this is essentially what you did with hostname when you tested this manually? [15:11] * zyga reboots for updates [15:12] PR snapd#8508 closed: github: run all spread systems in a single go with cached results [15:14] PR snapd#8723 opened: github: remove workaround for bug 133 in actions/cache [15:28] mvo: I made also this comment: https://github.com/snapcore/snapd/pull/8661#discussion_r429969902 [15:28] PR #8661: configcore: add "service.console-conf.disable" config option [15:28] mvo: checking [15:28] mvo: btw uc18 is also slow, but it doesn't have the 14s delay before the ttyS0 initialization [15:29] pstolowski: mborzecki: thanks for the Pool reviews, I landed the bits, now there's the main branch using it, but probably can wait a little bit before reviews, I want to discuss something first [15:29] pedronis: is it updated with master? [15:30] pstolowski: it is, but as I said can wait a little bit [15:30] pedronis: ack [15:36] ijohnson: #8718 is probably something you can help reviewing [15:36] PR #8718: boot, many: require mode in modeenv [15:36] pedronis: sure will take a look today [15:37] thx [15:49] PR snapd#8720 closed: spread.yaml: add ps aux to debug section [15:50] PR snapd#8704 closed: cmd/snap-preseed: improve mountpoint checks of the preseeded chroot (1/3) [15:50] PR snapd#8724 opened: interfaces/block_devices: add NVMe subsystem devices, support multipath paths <⛔ Blocked> [15:51] pstolowski: I merged 8704 to unblock you, still one suggestion inline that might be worth looking at in one of the followups [15:52] mvo: thanks! right, will do [15:52] thanks mvo! [15:55] mborzecki: I reviewed #8716 [15:55] PR #8716: o/devicestate: refactor current system handling [15:58] mvo: did you see my comment about missing newlines? you pushed many times to that PR but not a fix for that [16:00] pedronis: aha, sorry, missed it [16:43] I'm so confused right now [16:44] somehow a change in github.com/snapcore/snapd/snap dir is making github.com/snapcore/snapd/wrappers/ fail by calling systemctl is-system-running seemingly out of thin air??? [16:52] ijohnson: afaik only snap run is using is-system-running [16:52] pedronis: yes I figured it out [16:52] we mock systemctl, then we run shellcheck [16:52] on my system shellcheck is a snap [16:53] and it just so happened that I had a system key mismatch, so when `shellcheck` as a snap runs, it tries to check if the system is running using the mocked systemctl [16:53] very very confusing bug [16:53] interesting combination [16:53] we should probably separate mocking of path vs running shellcheck [16:55] my fix is just to set PATH _after_ we call shellcheck [16:56]