[06:17] morning [06:28] morning :-) [07:29] good morning [07:30] * zyga applies updates [07:48] zyga: hey [07:49] :-) [07:49] hey, how are you doing? [07:49] zyga: caught a cold :/ [07:50] given where we live I think it is inevitable [07:50] zyga: but managed to build centos cloud image, since there's still no official ones yet [07:51] I spent most of the weekend doing homework with kids [07:52] I spent some hours on manual pages for API docs [07:52] zyga: bet is was much fun ;) [07:52] I'll try to publish my toy library next weekend [07:53] hey zyga and mborzecki [07:53] zyga: what is you library about? [07:53] hey mvo, how are you? [07:53] mvo: hey [07:53] mvo: testing! :) [07:53] mvo: libzt, robust and simple testing [07:53] for C [07:53] tested, documented, building in all major OSes [07:53] (that's the goal) [07:53] I'm maybe 60% there [07:54] zyga: zt? libzygatest? [07:55] mborzecki: over time my private libraries evolved in their naming scheme but I guess z is just z, not an acronym [07:55] t is indeed for "testing" [07:55] I ended up writing it after another private toy project needed more testing than I could do by just playing with it [07:55] and I didn't want to depend on glib [07:55] or anything that's a mess to build outside of linux [07:56] anyway [07:56] zyga: nice! I'm good, thanks, didn't sleep that well though :/ [07:56] today is snap run --explain day :) [07:56] jamie is off this week so I will focus on things that don't touch C [08:00] morning [08:00] hey pawel [08:00] o/ [08:02] gosh bugzilla is such a dinosaur [08:02] trying to find snapd bugs [08:03] pstolowski: hey [08:03] is search implemented by grepping the database? [08:03] zyga: rhbz? or suse one? [08:03] suse [08:04] I clicked "search" like two minutes ago [08:04] ... [08:04] and got 400 [08:05] https://bugzilla.novell.com/buglist.cgi?query_format=specific&order=Importance&no_redirect=1&bug_status=__all__&product=&content=snapd [08:05] hey pstolowski ! [08:05] zyga: hmm maybe they structured the project differently than rhbz has [08:06] zyga: in rhbs it's product - pick fedora or EPEL, component - pick snapd, status all open and that's it [08:06] I'll search in my mail instead [08:06] at least it's not mantis, had trouble running a trivial search in centos bug tracker [08:14] brb, tea refill [08:14] rainy morning [08:14] oh, and it's possible to install snapd on centos8, but one has to enable `cr` repository as well, sice the selinux packages are part of 8.1 that's not in base os repo yet [08:14] cr? [08:15] zyga: continuous release, iow upcoming updates [08:15] aha [08:15] oh well :) [08:25] PR snapd#7979 closed: many: drop NameAndRevision, use snap.PlaceInfo instead [08:31] PR snapd#7978 closed: data/selinux, test/main/selinux-clean: update the test to cover more scenarios [08:33] degville: hi, i've added a note about CentOS CR repo under https://forum.snapcraft.io/t/installing-snap-on-centos/10020 can you take a look? :) [08:36] heh, suprised one took the time to write this https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1859185/comments/3 [08:36] Bug #1859185: No (obvious) way to turn autorefresh off

[08:37] mborzecki: will do - thanks for letting me know. [08:44] zyga: got your new mbp yet? [09:39] pstolowski: no, it's still in netherlands [09:40] pstolowski: and only with a shipping label :/ [09:40] pstolowski: so sigh [09:40] pstolowski: I don't doubt it will arrive on time but they said 14-15th [09:40] probably waiting just to arrive "on time" [09:41] mhm [09:48] Chipaca: hi, are there open problems with #7984 (just skimmed it) or are you ok with it? if not I'm happy to discuss [09:48] PR #7984: store, overlord/snapstate, etc: SnapAction now returns a []…Result [09:49] pedronis: I'm ok with it [09:49] pedronis: it's shorter than my first attempt so that's a plus :-) [09:49] Chipaca: ok, let me know if we land in trouble further down, but let's go this way for now [09:50] pedronis: ok, next pr should be up before noon [09:50] stacked on this one i mean [09:50] understood [10:02] mborzecki: I did a quick pass about the changes in #7972 [10:02] PR #7972: overlord/snapstate, wrappers: undo of snapd on core [10:03] pedronis: cool, thanks [10:03] pedronis: hey, do you have a moment to talk quickly about some aspect of snap disconnect --forget? [10:06] pstolowski: can do in ~20 mins ? [10:06] pedronis: sure [10:24] brb [10:27] pstolowski: I'm available now [10:28] pedronis: ok, coming to standup ho [11:04] PR snapd#7985 opened: tests: use snap remove with --purge flag in most of the spread tests [11:05] ^ pretty mechanical [11:07] looks like core18 is failing right now in master, has anyone already looked? if not, I'm doing so now [11:07] haven't looked, can look if you are busy [11:09] Chipaca: ty! [11:10] pstolowski: i was about to ask about just setting the option, but this way it's explicit so +1 [11:12] zyga: thanks, should be fine [11:12] Chipaca: yeah i was thinking about some magic, but preferred explicit [11:17] pstolowski: nice PR, thanks for this [11:17] mvo: i should have thought about it earlier! [11:19] pstolowski: do we have (at least) one test left that run snap remove without purge? I assume we do, just want to double check :) [11:20] mvo: yes, a few tests that test remove itself, snapshots test, and selinux tests (i mentioned this in the commit and PR desc) [11:20] pstolowski: \o/ [11:45] mvo, hey [11:46] mvo, about beta release [11:46] are you planning a 2.43.1 ? [11:46] or it needs to be discussed? [11:52] cachio: I'm not aware of 2.43 issues right now, so hopefully we don't need 2.43.1. or am I missing something? [11:52] mvo, I fwd an email on friday [11:55] PR snapd#7985 closed: tests: use snap remove --purge flag in most of the spread tests [11:56] cachio: oh, missed that, yes, this sounds very much like we need 2.43.1 [11:56] cachio: let me read the full bug [11:56] mvo, nice, thanks [11:57] cachio: thank you! [12:08] pedronis: https://github.com/snapcore/snapd/pull/7986/commits/479e46211c3ec18c95b6634d1b33fb54acc5ff4c [12:08] PR snapd#7986 opened: store, o/snapstate: send default-tracks header, use RedirectChannel [12:08] PR #7986: store, o/snapstate: send default-tracks header, use RedirectChannel [12:08] * Chipaca will start abbreviating overlord to 😎 [12:09] alas there is no panopticon emoji [12:09] emojus? [12:09] emoja? [12:09] * Chipaca hides [12:10] PR snapd#7987 opened: snap: make `snap version` output host without extra whitespace <⚠ Critical> [12:11] mvo: is the host-and-virt-in-version only just getting released? [12:12] Chipaca: yes, part of 2.43 which is in beta right now, never made it beyond beta afaict [12:15] mvo: use U+2620 instead of / :-p [12:15] Chipaca: heh [12:16] mvo: did you mean to target it to master? [12:17] Chipaca: mvo: we need to discuss that bit [12:17] Chipaca: yes, I will backport it from there once we have a agreement [12:17] pedronis: sure, should we do it right after the standup? [12:18] mvo: yes [12:18] ok === ricab is now known as ricab|lunch === msalvatore_ is now known as msalvatore === ricab|lunch is now known as ricab [14:32] PR snapd#7987 closed: snap: make `snap version` output host without extra whitespace <⚠ Critical> [14:32] re [14:40] PR snapcraft#2868 opened: cli: implement progressive releases [14:42] PR snapd#7988 opened: snap: remove "host" output from `snap version` <⚠ Critical> [14:49] PR snapcraft#2869 opened: build providers: use multipass from stable [14:52] mborzecki: wrt 596,523 hours, it's weird because nowhere do we format numbers with a comma like that [14:52] mborzecki: that's not snapd :) [14:54] mborzecki: in particular snapd's progress bar would've printed "68y" if it had to render that number [14:54] Chipaca: libc may do that [14:54] Chipaca: , is a common 1k separator [14:54] which, sure, weird, but that's why i'm asking where that's coming from [14:54] zyga: yes i know it may [14:54] zyga: but we don't use libc to format how long is left for a download in snapd [14:55] so that's why i'm asking :) [14:56] Chipaca: hmm right [14:57] there's not much info there, wouldn't do anything unless they give more info [15:00] Chipaca: otoh, i should probably open a PR to tweak quantity formatting in the progress bar, sub second values are the funny ones i mentioned [15:00] mborzecki: :) [15:00] mborzecki: there are probably bugs (tm) [15:00] s/bugs/features/ [15:01] mborzecki: rather, we have yet to adequately sample the probability cloud of bugs [15:02] certainly bug(x⃗) > 0 ∀ x⃗ [15:02] 🧐 [15:05] mborzecki: what kind of weirdness, btw? [15:06] Chipaca: when the download is close to finishing sometimes the left time reported is in us/ns range [15:07] Chipaca: as in right before it's finished [15:07] mborzecki: ah, yes [15:07] mborzecki: I've wondered what it goes on to do after finishing :) [15:07] pedronis: mvo: is it safe to say that all the tests that are uc20 specific are given by function name `Test*20*` ? Using a similar regex I found these tests: https://pastebin.ubuntu.com/p/NHq7sB68dj/ [15:07] do y'all think that's exhaustive or are there other packages/test suites that I need to look at as well ? [15:08] ijohnson: yes, it's reasonable, but some one have the 20 in the suite [15:08] not each test [15:08] ah good point [15:09] pedronis: so that adds these 3: https://pastebin.ubuntu.com/p/FNdtRsttBc/ [15:09] is that probably all of the uc20 specific tests ? [15:10] ijohnson: seems right to me [15:10] ack, thanks [15:11] (I'm in a meeting) [15:24] mvo, this is the model I am usign https://github.com/sergiocazzolato/snapd/blob/tests-enable-nested-on-core20/tests/lib/assertions/nested-20-amd64.model.json [15:24] brb [15:24] cachio: thank you! I think I have an idea what might need tweaking. I am in a meeting right now but let me try to help [15:25] mvo, great, thanks [15:26] cachio: I think this needs "grade: dangerous" (or something) and a snaps: subsection, here is an example https://github.com/snapcore/snapd/blob/master/tests/lib/assertions/ubuntu-core-20-amd64.model [15:26] mvo, nice, thanks [15:34] mvo: pedronis: cmatsuoka: thoughts on having modeenv.Write() also return the filepath where the modeenv was written out? this will make mocking easier in tests [15:34] ijohnson: still in meeting [15:34] oh ok [15:41] grr jq in 14.04 doesn't like dashes in object keys [15:51] Chipaca: use jq snap [15:56] yeah [15:59] ijohnson: sounds fine to me [15:59] mvo: ack, I will file a quick PR for that this morning then [15:59] thanks [16:01] ijohnson: I have another meeting, but I'm a bit unclear why you need that? it writes thing in a predictable place [16:01] no? [16:02] pedronis: yes it's predictable, but in all these tests I have to call modeenv.Write(), then immediately do a filepath.Join(dirs.Root /*or c.tmpdir*/,dirs.SnapModeenvFile) to remove it after the tests [16:02] pedronis: so it would be nice to just have `f, err := m.Write(); defer os.Remove(f)` [16:03] * cachio lunch [16:09] ijohnson: ok, I would prefer not to do it, because usually we don't have that style [16:10] ijohnson: you can have a test helper if needed, you can use AddCleanup in it [16:10] pedronis: ok [16:10] ijohnson: not a lot of the tests you listed have a devicemgr around [16:11] ijohnson: are you adding write modeenv to a lof of places? [16:11] I fear I'm missing something here [16:11] pedronis: well it's probably fine to just have the same modeenv in most tests, but other tests are not consistent about the name of the base/kernel for uc20 tests iiuc [16:12] it's fine I guess to have a helper [16:15] ijohnson: ok, notice that we might need a higher level helper anyway at some point [16:15] * zyga-laptop had dinner and need a moment to rest [16:15] code review time! [16:15] yes perhaps [16:15] zyga's laptop had dinner? 😋 [16:15] roadmr sadly the laptop recharges while I'm not looking ;) [16:15] ijohnson: we bootloadertest.MockBootloader but it doesn't fit quite the fact that Core20 has boot state in more than one location [16:16] *we have [16:16] that one has things like SetBootBase and SetBootKernel etc [16:16] but there's mismatch with Core20 [16:19] pedronis: yes I added a number of functions to the MockBootloader for the ExtractedRunKernelImage interface [16:29] hello, i have a lot of time for your kind reply, i get this error when installing youtube music cannot self-bind mount /run/snapd/ns: Cannot allocate memory [16:29] 5.5.0-1-MANJARO === heather is now known as hellsworth [16:33] mvo, hey [16:34] i will check if there's a new kernel beta soon [16:34] thank you [16:34] mvo, using snaps as you did in the other model [16:34] I get error: cannot decode model assertion "/home/gopath/src/github.com/snapcore/snapd/tests/lib/assertions/nested-20-amd64.model": assertion model: type of snap "snapd" must be one of app|base|gadget|kernel|core [16:34] Zylop: hello [16:34] mvo, which type we should use for snapd? [16:34] Zylop: that's interesting, I've never seen an error like that before [16:36] zyga: Zylop: I seem to remember seeing that error on the forum recently [16:36] Zylop: perhaps it's some kind of new hardening or bug [16:36] or just downgrade to 5.3 [16:36] its a bug in 5.4 and 5.5 [16:36] kernel [16:36] patched in 5.5rc5 [16:36] Zylop: zyga: see https://forum.snapcraft.io/t/on-ubuntu-18-04-3-with-5-4-5-5-kernels-snaps-are-not-launching/14662/18 [16:36] witch atm its not officially launched so im downgrading to 5.3 for while [16:36] i founf this [16:36] ah ok [16:36] ty [16:37] i see in arch forum [16:37] * zyga l boks [16:37] thanks ijohnson [16:37] PR snapcraft#2869 closed: build providers: use multipass from stable [16:37] PR snapcraft#2870 opened: snap: set PYLXD_WARNINGS to inhibit unknown LXD attribute warnings [16:47] PR snapd#7989 opened: devicestate: do not allow remodel between core20 models [17:02] cachio: that means the snapd is too old [17:03] we allow type: snapd now [17:03] pedronis, ah, ok, that makes sense [17:03] thanks!! === alan_g is now known as alan_g_ [17:18] working [17:18] thanks [17:18] :-) [17:19] PR snapcraft#2871 opened: snaps: account for forwarded (effective) installation channels [17:24] pedronis: oh also I was going to ask you if we expect classic devices with the new uc20 models to have modeenv? currently my HasModeenv method just checks if the model is not ModelGradeUnset === Spacehog is now known as Zylop [17:35] is the forum down again ? [17:36] seems like it [17:37] it should really be added to status.snapcraft.io [17:37] +1 for that [17:37] pedronis, which version of snapd is needed? [17:37] snapcraft.io/docs should be on status.snapcraft.io too [17:38] i just tried with the 2.42.1 and did't work [17:46] pedronis, and core 2.42.8 [17:46] 2.42.5 [17:47] pedronis, with 2.43 works [18:05] PR snapcraft#2861 closed: meta: remove Application's `prepend_command_chain` [18:22] ijohnson: we don't let you set classic and grade at the same time atm, we will need to think through what a core20 model means for classic [18:25] cachio: yes, it was in the 2.43 dev cycle [18:25] pedronis: ack that's what I thought [18:25] pedronis, yes, the image is created now [18:25] pedronis, thanks for the help+ [18:38] noice [19:08] PR snapcraft#2866 closed: spread tests: limit adapter test to amd64 [19:08] PR snapcraft#2870 closed: snap: set PYLXD_WARNINGS to inhibit unknown LXD attribute warnings [19:15] zyga: could you have a quick look at https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1859160 to see if it's something you recognize? [19:15] Bug #1859160: Snapd fails to delete obsolete dev-loops, thus causing delay at lockscreen [19:23] PR snapcraft#2858 closed: add support for system-usernames [19:29] PR snapcraft#2872 opened: Release/3.9.5/cherry picks [20:23] PR snapcraft#2867 closed: elf: remove return parameters for ElfFile's _extract() [20:40] PR snapd#7990 opened: many: misc tweaks [20:48] PR snapd#7947 closed: boot,bootloader: support new UC20 style kernel extraction <⛔ Blocked> [20:56] PR snapd#7991 opened: boot: add HasModeenv to Device [21:11] PR snapcraft#2872 closed: Release/3.9.5/cherry picks [21:53] PR snapcraft#2868 closed: cli: implement progressive releases [22:13] PR snapd#7992 opened: bootloader: add ExtractedRunKernelImageBootloader interface, implement in grub