[05:13] morning [06:11] Hi [06:12] Sooooo sleepy [06:21] Hey mvo [06:22] hey zyga [06:24] zyga: if you adress the feedback from jamie in 6846, could you please force push so that I can cherry pick more easily? [06:25] PR snapd#6843 closed: tests: avoid removing snaps which are cached to speed up the prepare on boards [06:25] zyga: mvo: hey [06:28] hey mborzecki [06:32] PR snapd#6831 closed: tests: retry govendor sync [06:44] low hanging fruit: bump the delta tag please [06:45] https://github.com/snapcore/snapd/pull/6855 is a bit big [06:45] PR #6855: overlord: implement store switch remodeling [06:45] PR snapd#6846 closed: cmd/snap-confine: unshare per-user mount ns once [06:47] zyga: I cherry picked 6846 [06:47] Thank you [06:47] I'm still working on the main dish === pstolowski|afk is now known as pstolowski [07:05] morning [07:09] hey pawel [07:25] mvo: hi, the PR to do store switch remodeling is up, some TODO related to conflicts, but otherwise is all there. Re-reg is a bit more complicated but not complex, needs the dedicated context and a new task kind. [07:26] PRs are a bit big though (espcially tests) might need splitting. [07:26] mvo: hey, thanks for moving our 1-1! [07:51] pstolowski: is that time ok? I'm relatively flexible thanks to mborzecki who also moved the meeting :) [07:52] mvo: yes, it's perfect! [07:52] PR snapd#6854 closed: cmd/snap-update-ns: rename applyFstab to executeMountProfileUpdate [07:53] we are below 60 again! [07:53] PR snapd#6851 closed: cmd/snap: mangle descriptions that have indent > terminal width <⚠ Critical> [07:56] moin moin [07:58] hey Chipaca [07:59] PR snapd#6856 opened: cmd/snap-update-ns: add tests for executeMountProfileUpdate [08:02] PR snapd#6850 closed: gadget: add volume level update checks [08:02] another one lands :) [08:03] !!! [08:08] PR snapd#6857 opened: gadget: record sector size in positioned volume [08:08] and one more up /o\ [08:09] off to school, back in a bit [08:14] mborzecki: thank you [08:16] man, I needed this coffee [08:39] pedronis: you around? [08:44] Chipaca: yes [08:44] pedronis: wrt health-checks, is there anything you'd rather see up in a pr next? [08:45] Chipaca: did you push one already? [08:45] pedronis: just the draft one [08:45] pedronis: working on the tests still [08:45] Chipaca: is it too big? [08:45] pedronis: no i don't think so [08:46] I'm not sure I understand the question [08:48] pedronis: the pr that is up, that i expect to finish today while you're travelling, covers only a small portion of the feature [08:49] pedronis: so, what should i work on once that is 'done' (code complete)? [08:49] Chipaca: the PR runs the hook and let the hook set the health, right? [08:49] pedronis: yes, post-refresh [08:50] or post-install [08:50] i mean it doesn't do the pre-refresh nor the on-demand, nor the info, nor the list [08:50] I think the next bit is snap info [08:50] okie doke [08:50] now that I understand the question [08:50] :) [08:50] * Chipaca is always very clear [08:51] pedronis: and after that? (being optimistic here) [08:52] regular running of the hook? [08:52] and then the pre refresh one we need to think a bit, we can run it but we wouldn't do much with, would we? [08:53] so it well might be on-demand [08:53] I expect on-demand to be a bit complicated, to do the UX as we discussed it [08:53] because we don't have anything quite like that [08:53] atm [08:53] oh [08:54] what are you discussing? it sounds curious [08:54] health checks [08:54] or you mean the last bit in particular? [08:54] the last bit is about snap health and returning info to the user for each snap as soon as it is available [08:55] about on-demand things, I was wondering if there's some new functionality in hooks that is planned [08:55] don't think is about hooks [08:55] it's more about how we return results [08:55] from API [08:56] but it depends exactly how we implement it [08:59] Chipaca: so, info, regular, on-demand [09:00] pedronis: regular as in on-a-timer? [09:01] Chipaca: that part of the spec that says we run them (probably triggered from ensure) at given interval, [09:01] more often if unhealthy [09:01] yep [09:02] ok [09:03] there are some fun issues there to, but we can have a general sketch of that before digging [09:03] *there too [09:06] PR snapd#6858 opened: cmd/snap: unit tests for debug timings [09:13] mborzecki: not sure if intersects with your work [09:13] running snapd a while prints this a loooot [09:13] maj 10 11:00:53 yantra snapd[57785]: kernel_os.go:198: cannot get boot settings: cannot determine bootloader [09:32] pstolowski: "(and optionally the)" because code is optional! [09:32] :-p [09:33] :) [09:36] * pedronis starts early travel to Lyon [09:36] have great day and weekend! [09:36] pedronis: see you soon :) [09:36] pedronis: I'm going through geneva [09:37] pedronis: perhaps we might even share a train ride [09:37] maybe, depends on times, but will do geneva -> Lyon on Sunday [09:37] same like me [09:37] at some point [09:47] PR snapd#6751 closed: overlord/snapstate: perform hard refresh check [10:10] what was the right name for 'sideloads'? [10:10] Chipaca: none :) [10:10] snap install from a file? [10:10] unsigned? unverified? unpotatoed? [10:10] unasserted? [10:10] unasserted, thanks [10:10] note, I just made that up [10:10] but it sounds okay :) [10:11] 'twas that afair [10:11] AFAIR [10:11] also [10:20] PR snapd#6853 closed: interfaces: special-case "snapd" in sanitizeSlotReservedForOS* helper… <⚠ Critical> [10:22] second review for 6793 would be nice (maybe mborzecki because he already did a first pass :) ? [10:25] mvo: mborzecki will be back in half an hour [10:25] mvo: can I help [10:26] mvo: I ran into another bug, don't understand it yet, looks related to recent changes [10:26] mvo: snap-confine cannot open the base directory in /tmp [10:26] debugging now [10:32] mvo: https://github.com/snapcore/snapd/pull/6823 can land? [10:32] PR #6823: packaging: build empty package on powerpc [10:32] pstolowski: if it has two reviews I think it can land [10:32] zyga: oh, where do you see this bug? [10:33] mvo: when developing on opensuse [10:33] mvo: just added debugging, hold on [10:33] it's odd, I didn't see it before [10:34] mvo: 2nd review is from vorlonofportland [10:34] pstolowski: aha, indeed, then yeah, lets merge it [10:35] PR snapd#6823 closed: packaging: build empty package on powerpc [10:37] pstolowski: thank you! I also cherry-picked it into 2.39 [10:39] Chipaca: 6803 has a few conflicts [10:40] nevah [10:41] 6795 also needs a second review, should be an easy win [10:45] mvo: i've merged #6811 but we need to ping Sergio re your question about if it needs to go to 2.39 [10:45] PR #6811: tests: make create-user test support managed devices [10:45] pstolowski: indeed [10:45] pstolowski: thank you! [10:45] PR snapd#6811 closed: tests: make create-user test support managed devices [10:50] I need a break [10:50] mvo: I think master may be broken [10:50] mvo: the /tmp changes affecte stuff that is only picked up by tests running as user [10:50] meh [10:50] not fun [10:50] anyone can run edge and say if stuff is falling apart? [10:51] but first [10:51] really [10:51] break [10:59] meh [11:00] it's so odd [11:00] something must have landed [11:00] zyga: I just refreshed my bionic to edge and ran hello-world [11:00] zyga: seems to be fine [11:01] zyga: hold on, that was core not the snapd snap [11:02] perhaps it's something in my setup [11:02] * zyga explores [11:02] so I get permission denied [11:02] but it doesn't seem like apparmor [11:02] * zyga rechecks [11:03] re [11:03] zyga: on my system(s) things seem to be ok - at least with test-snapd-tools.echo [11:03] thanks for checking [11:03] mvo: what are the permissions on /tmp/snap.test-snapd-tools? [11:06] is there a potential for a race when 'snap install thing-with-home-interface', then running thing-with-home-interface and expecting :home to be connected? [11:06] Wimpress: mvo: Chipaca you guys around for catch up at half past? [11:06] Interface Plug Slot Notes [11:06] home documentation-builder:home - - [11:06] trying to work out how that happened, in a fresh LXD container, immediately after 'snap install documentation-builder' [11:07] sparkiegeek: odd, is core or snapd installed? [11:08] zyga: drwx------ 3 root root 4,0K Mai 10 13:00 /tmp/snap.test-snapd-tools [11:08] zyga: good question, this is repeatedly happening in aforementioned container - in a test run, so can add debugging but loop is longer than I'd like [11:08] mvo: thanks [11:08] popey_, Wimpress, Chipaca I'm around but I would not mind skipping given that we meet in lyon next week anyway [11:08] mvo: for me it is root.users! [11:08] mvo: whaaat? [11:09] * zyga inspects [11:09] mvo: can you remove that [11:09] and run it again [11:09] and see if that fails? [11:10] mvo: I think I know what's wrong [11:10] mvo: look at https://github.com/snapcore/snapd/pull/6714/files [11:10] PR #6714: cmd/snap-confine: reject crafted /tmp/snap.$SNAP_NAME [11:10] mvo: you don't see it on debian [11:10] see that code that changes groups? [11:11] zyga: yes [11:11] I think we need that now [11:11] what's the permission of /usr/lib/snapd/snap-confine? [11:11] mvo: ok, that's cool [11:11] zyga: -rwsr-sr-x 1 root root 99K Mär 15 20:54 /usr/lib/snapd/snap-confine [11:11] mvo: also in snapd.snap [11:12] PR snapd#6857 closed: gadget: record sector size in positioned volume [11:12] * zyga is confused [11:12] mvo: thank you [11:13] popey_: cool, declined it now [11:13] zyga: appears to be neither core nor snapd, but our good friend core18 [11:13] sparkiegeek: that's a known bug [11:13] * zyga looks for the bug URL [11:13] bug 1819318 [11:13] ? [11:14] https://bugs.launchpad.net/snapd/+bug/1819318 [11:14] yep [11:14] Bug #1819318: no interfaces when installing only core18 [11:14] zyga: ok, and workaround is to ... manually install core too? [11:14] yes [11:15] sparkiegeek: yes, this will be fixed in 2.39 (which is in candidate already and should go to stable early next week) [11:15] sparkiegeek: sorry for the trouble [11:16] mvo: np. Due to a separate SNAFU that took me out all of yesterday, glad to finally be able to get debugging on this and look under the curtain, happy it's a known issue [11:24] PR snapd#6795 closed: cmd,sandbox: tweak seccomp version info handling [11:27] is there a 'grep-friendly' way of determining if I'm hitting ^ that bug. i.e. snap connected documentation-builder | grep -v ':home' || snap install core ? [11:27] sparkiegeek: just snap list is sufficient [11:27] sparkiegeek: then snap connections will show you there are no slots [11:28] zyga: right, just trying to build a snippet in a Makefile to workaround the bug if I need to, but don't install core if the interface is connected [11:54] mborzecki: hey, do you want to repeat our experiment? [11:54] mborzecki: I'm stuck, perhaps you will just see what's obviously broken [11:55] https://meet.google.com/rha-xhdp-vic?authuser=0 [11:56] mvo: or you perhaps? [11:56] anyone? [12:00] zyga: in 5? i'm finishing reading through pstolowski's pr [12:00] sure [12:08] PR snapcraft#2500 closed: cli: add remote build [12:08] \o/ [12:10] Chipaca, pstolowski: saving data of multipass, on removal, is loooong [12:11] Chipaca: curious, how do we remove snapshots? [12:11] zyga: snap forget [12:11] zyga: uhmmm.. it can be disabled [12:11] can I forget all snapshots of a given snap? [12:11] zyga: yes [12:11] cachio: hey, any chance for a fedora-30 image please? :) [12:12] zyga: not easily tho :-| [12:12] zyga: maybe we should add that, although it'd be tricky [12:12] Chipaca: I noticed [12:12] given that the things are called NNN_$SNAP_NAME_* [12:13] is that enough of a hint? [12:13] Chipaca: can I just rm them manually? [12:13] zyga: snap saved thesnap | | xargs snap forget {} thesnap \;, or something [12:13] or is there state [12:13] :/ [12:13] zyga: there is no state [12:13] ah [12:13] off they go :) [12:13] that was part of the design of snapshots [12:13] thanks! [12:13] though for usability [12:13] killing snapshots of a snap [12:13] very likely useful [12:14] zyga: there is state about auto-snapshots, but it shouldn't break anything if it tries to remove them and they're gone [12:14] zyga: file bugs if it breaks and eats your thesis [12:15] who knew [12:15] sudo has --background and --command-timeout options [12:15] should we add a --purge to snap remove to mean "don't auto-snapshot"? [12:15] Chipaca: yeah [12:15] feels like something I'd do in tests [12:15] our tests don't have much data but removing multipass hit my fans running [12:15] pstolowski: wtyt? [12:15] 2GBs [12:16] wdyt* [12:18] Chipaca: +1 to 'snap remove --purge', i like that. gives more control than all-or-nothing flag we have now [12:19] we should either put up a pr, or write something in the forum, before we forget [12:19] me, i'm going to have lunch [12:19] * Chipaca prioritises like a baws === ricab is now known as ricab|lunch [12:25] mborzecki: PR #6849 failed some integration tests [12:25] PR #6849: many: introduce a gadget helper for locating device matching given structure [12:33] mborzecki: it's != its ;) [12:33] https://github.com/snapcore/snapd/pull/6849/files#diff-f57f6320470091f8bb514620cc2fbbddR35 [12:33] PR #6849: many: introduce a gadget helper for locating device matching given structure [12:38] mvo: mystery solved [12:41] Saviq: duh, missed that, thanks for spotting [12:44] pstolowski: May 10 12:01:02 may101137-895818 nsenter[32024]: retry.go:120: DEBUG: Retrying because of temporary net error: &net.OpError{Op:"dial", Net:"tcp", Source:net.Addr(nil), Addr:net.Addr(nil), Err:(*net.DNSError)(0xc00006c5c0)} [12:45] we're logging the error as #+v ? [12:45] %+v [12:47] mborzecki: %#v, yes [12:48] this is Debug though [12:51] pstolowski: yup, so probably fine [12:56] mborzecki: yeah, an it helps understand error conditions better if we need to peel them :} [12:57] Saviq, sure, I'll try today [12:58] zyga: what is the mystery? [13:00] cachio: hey, can you take a look if that PR that we landed today needs to go to 2.39? https://github.com/snapcore/snapd/pull/6811 [13:00] PR #6811: tests: make create-user test support managed devices [13:03] joining [13:04] PR snapd#6793 closed: cmd/snap: support for --ensure argument for snap debug timings [13:39] Chipaca: so, i'll create forum topic re --purge and do impl sometime next week [13:39] pstolowski: <3 [13:44] k [13:45] degville: btw, we need to update documentation with automatic snapshots; is it ok if i send you a short description of it, and you flesh it out/reword and put in a proper context of existing doc? === ricab|lunch is now known as ricab [13:49] pstolowski: absolutely, yes, of course. [13:54] Chipaca: I couldn't help it and de-conflicted 6803 [13:54] Chipaca: hope you don't mind [13:54] mvo: HULK SMASH [13:55] mvo: sorry, i meant "thank you" [13:55] mvo: typo [13:55] :-þ [13:57] haha [13:57] * mvo hugs Chipaca [14:17] PR core#38 closed: Add another pi-config option [14:17] PR core#83 closed: move most of the ubuntu-core config deb into the snap snap build [14:17] PR core#104 closed: snapcraft.yaml: use remote fc-cache-builder [14:18] PR core#38 opened: Add another pi-config option [14:18] PR core#83 opened: move most of the ubuntu-core config deb into the snap snap build [14:18] PR core#104 opened: snapcraft.yaml: use remote fc-cache-builder [14:43] mborzecki, zyga if you could take another look to #6541 should be great, thanks [14:43] PR #6541: tests: change how xdg-desktop-portal is prepared and restored for desktop-portal-* tests [14:43] cachio: ok [14:43] mborzecki, thanks [15:01] Chipaca: https://forum.snapcraft.io/t/automatic-snapshots-snap-remove-purge-proposal/11294 [15:01] * Chipaca reads [15:02] pstolowski: thank you [16:15] cmatsuoka: almost forgot - safe travels! [16:15] thanks! :) === pstolowski is now known as pstolowski|afk [16:31] pstolowski|afk, hey [16:32] cachio: hey [16:32] I have few issues with core [16:32] for example sudo usermod -a -G dialout user1 [16:32] does not work [16:33]