[02:32] PR snapcraft#3263 closed: colcon v2 plugin: honour http(s) proxy for all build commands [06:23] o/ [06:23] mvo: I saw your message yesterday [06:23] oh boy I open my laptop and Lucy wakes up [06:24] zyga-x240: good morning [06:24] :) [06:24] zyga-x240: no worries, either way is fine [06:25] until she's fully awake, changed and fed I will not be fully available [06:25] just a few more days till next week [06:26] zyga-x240: no worries [06:32] can someone please approve 9214? it's just changelog updaes [06:40] PR snapd#9215 closed: mkversion.sh: do not use git describe --dirty [06:45] PR snapd#9215 opened: mkversion.sh: do not use git describe --dirty [07:04] morniing [07:05] good morning pstolowski [07:18] re [07:18] mvo looking [07:19] mvo merged [07:20] PR snapd#9214 closed: release: 2.46 [07:23] zyga: \o/ [07:24] mvo in a way it's good we'll need 2.47 [07:24] we should release it in a few weeks at most [07:24] and get back to small releases [07:24] that are fast [07:25] PR snapd#9219 opened: run-checks: only check files in git for misspelling [07:27] mvo reviewed with a question [07:28] zyga: \o/ [07:28] mvo I'm available now, shall I pick up snapshot export PR? [07:28] zyga: if you want, I did not manage to and have meetings soon :/ [07:28] ok [07:28] * mvo hugs zyga [07:28] gladly [07:28] on it [07:28] it's a new topic [07:28] that feels good sometimes [07:28] btw, offtopic, my lxd install stopped working [07:28] it says it needs manual network configuration [07:29] as it cannot find a free subnet or something [07:29] I have a 10.x.y.z/8 netmask [07:29] so I'm quite surprised [07:31] mvo shellcheck complained about the spell changing [07:34] zyga: yeah, I am a muppet [07:34] zyga: I force pushed a fix, I don't have $GOPATH/bin in my $PATH [07:34] k [07:34] zyga: so for local testing I did that (I don't like the idea of something like $GOPATH/bin in my PATH, feels too dangerous) [07:34] anyway [07:34] silly me [07:34] it's okay [07:35] zyga: I replied, I can add comments about where the ignore comes from, maybe a good idea [07:36] mvo replied too [07:50] PR snapd#9205 closed: boot/initramfs_test.go: reset boot vars on the bootloader for each iteration [09:08] mvo I've updated https://github.com/snapcore/snapd/pull/8982 and left two unresolved comments: [09:08] PR #8982: snapshots: export of snapshots [09:08] mvo one is about the help message of export-snapshot, I need to ask Graham for help here [09:08] it can also happen in a follow up I guess but it might be easy enough to do todat [09:08] *today [09:09] the second one is about error handling in https://github.com/snapcore/snapd/pull/8982#discussion_r476372693 [09:09] and I think I need pstolowski help on that [09:09] can you look and express your opinion please pawel? [09:09] other code seems to handle errors from requests in a different way [09:09] and can use the body to unpack error details [09:09] we should either be consistent about it or leave a comment why we are not doing this [09:10] mvo I left the tar improvements out for a new PR [09:10] and adjusted everything else [09:10] so it means someone other than me should do a review [09:10] zyga: in a meeting, but *thanks* [09:11] zyga: sure [09:15] I can look about the error handling as well but in a bit, I suspect the difference might relate that the happy response is not json here [09:15] but I suspect the error ones should be [09:19] yeah, I think so too [09:19] thanks! [09:20] I'll return to my other tasks and try to finish the tar optimize after PT in the evening [09:26] zyga: it looks like e.g. maintenance status on error may be interesting to have [10:03] zyga: just looked over your changes, looks fine [10:03] zyga: thank you! [10:05] mvo cheers :) [10:05] pstolowski: I wonder if the client error handling should be done in a followup, I'm not sure we need more here tbh but even if it seems the exiting code is not wrong, just that it could be improved(?) [10:05] pstolowski: I'm in favor of merging (if green :) [10:05] mvo: yes, absolutely, a followup is gine [10:05] *fine [10:05] mvo I'm okay with that as well [10:26] hmm services PR failed and the failure looks real on 14.04 [10:41] PR snapd#9220 opened: o/snapstate: disk space check with single snap install <⛔ Blocked> [10:55] pstolowski how did it fail? [10:56] zyga: state of services not as expected [10:56] mmm [10:56] oddly only on 14.04. i'm going to investigate in a bit [10:57] what was the unexpected status? [11:06] zyga: enabled/inactive instead of disabled/inactive [11:06] PR snapd#9221 opened: tests: disk space awareness spread test <⛔ Blocked> [11:16] PR snapd#8982 closed: snapshots: export of snapshots [11:21] PR snapd#9199 closed: snapstate: installSize helper that calculates total size of snaps and their prerequisites [11:31] pstolowski: should I re-review the service PR or wait? [11:51] PR snapd#9222 opened: osutil: add a package doc comment (via doc.go) [12:04] pedronis: please wait, i'll let you know [12:04] pstolowski: ok, thx [12:05] pstolowski: when you have a moment, can you merge master into #9211 ? [12:05] PR #9211: o/snapstate: disk space check with InstallMany <⛔ Blocked> [12:07] yes [12:11] mvo: #9222 is small/simple and touches something you did [12:11] PR #9222: osutil: add a package doc comment (via doc.go) === benfrancis7 is now known as benfrancis [12:58] pedronis: thank you [13:01] PR snapd#9206 closed: boot: complain about reused asset name during initial install [13:01] PR snapd#9222 closed: osutil: add a package doc comment (via doc.go) [13:09] hey-o [13:09] i've recently found that i need to restart snapd.socket manually every so often === benfrancis6 is now known as benfrancis [14:06] pstolowski: 9211 has some unhappy unit test but looks great otherwise [14:08] mvo: ah, i forgot to rename after mergin master, thanks [14:08] zyga, hey [14:08] about the error related to the failober tests [14:09] re [14:09] of rexternal backend [14:09] Chipaca, oh what's the state of the socket when you do? [14:09] cachio, yeah? [14:09] do you have in mind anython to fix that? [14:09] zyga: a fair question. i'll look next time it happens. [14:09] cachio sorry, the failure we talked about before or some new one? [14:09] Chipaca I did observe something weird related to pulse socket like that [14:09] oh, sorry [14:09] not to the pulse socket [14:09] zyga, the one we talked before [14:09] to the snapd socket in fact [14:10] it was related to snapd-userd socket [14:10] when we changed the socket definition [14:10] (just rewrite it) [14:10] and not reload systemd-userd [14:10] zyga, which failed because external backend does "sudo su to run the spread script" [14:10] cachio right [14:10] cachio that's fundamentally broken sadly :/ [14:11] cachio we have to change that [14:11] it should, at minimum, do sudo -l [14:11] zyga, so change spread [14:11] er, wait [14:11] not -l [14:11] su -l, for sudo that's .... [14:12] PR snapd#9217 closed: run-checks: add typos from auto-tools when using `make hack` [14:12] cachio sudo -i (or sudo --login for clarity) [14:12] not tested [14:12] but that should be much better if it works as I expect [14:12] if that's convenient, change spread and see if this fixes it [14:12] zyga, ok, I'll make the change [14:12] ok [14:12] zyga, thanks for the suggestion [14:13] Chipaca in any case, thanks for the heads up [14:13] I'm running snapd without rebooting on metal to observe weird things as well but I have been unable to spend the time to dig into the things I saw [14:13] (insert commas for best sense) [14:16] zyga: mvo: so for #9216, what should I do? the changelog typo fix is not on release/2.46 so I'm not sure what to force push ? [14:16] PR #9216: vendor.json: update mysterious secboot SHA again <⚠ Critical> [14:17] PR snapd#9215 closed: mkversion.sh: do not use git describe --dirty <⚠ Critical> [14:17] ijohnson I was unclear, sorry, I expected to find only the hash change, not the typo fixes [14:17] zyga: sure but then no tests run against the PR if I don't fix the typos here [14:17] are they necessary in this branch? [14:17] oh [14:17] meh [14:17] that sucks, can we just merge the fixes separately? [14:17] i.e. open any PR against release/2.46 and tests don't run [14:17] it's not a biggie [14:18] I'll leave to mvo to decide what to do, I'll leave the PR as-is for now [14:18] ijohnson due to the typos? [14:18] ok [14:18] zyga: yes [14:18] I trust mvo's merge button [14:18] the unit / static tests fail on the typos and don't do spread runs [14:18] PR snapcraft#3264 closed: colcon v2 plugin: don't strip environment for stage-runtime-dependencies [14:18] normally I'd say it's fine but this sha thing is mysterious enough to make me want to see spread results of building with this vendor.json change [14:20] ijohnson I would just merge the typos separately [14:20] and then rebase this branch [14:20] so we can see a single thing to track, not that with other bits [14:22] zyga: sure but I would rather wait til mvo says what he prefers [14:22] ok [14:23] PR snapcraft#3265 opened: colcon v2 plugin: honour http(s) proxy for stage-runtime-dependencies [14:29] mvo: pstolowski: there is no daemon support yet for ErrInsuffiecientSpace ? also that name is not following style conventions [14:30] pedronis: correct, no daemon yet [14:35] zyga, so currently it is doing "sudo -i /bin/bash /tmp/tmp.zwRt33mi04" [14:35] to execute the script [14:36] pstolowski: I left some comments [14:36] cachio currently as in with some changes or vanilla? [14:37] ijohnson, zyga sorry, meetings will look at the backlog [14:37] zyga, vanilla [14:37] cachio ok, perhaps it needs some more oomph [14:37] pedronis: ty [14:38] can you try changing that to su -l [14:38] zyga, sure [14:40] mvo: I opened https://github.com/snapcore/snapd/pull/9223 to fix the --dirty handling in mkversion.sh [14:40] PR #9223: mkversion.sh: simple hack to include dirty in version if the tree is dirty [14:41] hopefully that is the best of all possible worlds (but all possible worlds are still terrible) [14:41] mvo: also please advice on next steps for #9216 [14:41] PR #9216: vendor.json: update mysterious secboot SHA again <⚠ Critical> [14:42] PR snapd#9223 opened: mkversion.sh: simple hack to include dirty in version if the tree is dirty [15:07] PR snapd#9216 closed: vendor.json: update mysterious secboot SHA again <⚠ Critical> [15:07] pstolowski: mvo: do we plan to have checks also for refresh ? [15:08] PR snapcraft#3266 opened: cli: add --enable-experimental-extensions option for expand-extensions [15:09] pstolowski, pedronis: I think that would be good [15:09] pedronis: yes i think so [15:15] pstolowski: mvo: given that we remove old revisions only after installing the new ones, it's probably not too different form install [15:15] *from [15:15] pedronis: yeah [15:16] pedronis: one interessting point for really low disk space system is that we may need to provide a mode in the future to change this behavior to "remove-before-install" [15:19] ok, sounds something to think about when we get there [15:21] it means no rollback unless you keep two old revisions [15:22] which would take more space [15:23] mvo: if you really want just one revision there's no undo [15:28] * cachio lunch [15:29] * mvo nods [15:42] PR snapd#9224 opened: Added MIRKey to U2F devices list in interfaces/builtin [15:47] hmm there is something wonky with 'systemctl enable --root=/ on 14.04 ...' [15:48] mvo: interesting idea [15:49] pstolowski: all later [15:49] sure, ack [15:49] pstolowski: just an idea because we were talking about really low disspace systems [15:51] systemctl enable --root=/ on 14.04 creates symlinks sunch as "snap.disabled-svcs-kept.svc.service -> ///etc/systemd/system/snap.disabled-svcs-kept.svc.service" and things seem to go wrong from there [15:52] pstolowski: what is using enable --root ? a test? [15:53] pedronis: Enable(..) in our systemd package, and it's now used in my services PR. no problems except for 14.04 [15:54] so it's used for real, not in test [15:55] pstolowski: is that related to preseeding ? [15:55] pedronis: no [15:55] or it was always like that? [15:56] pedronis: as far as systemd package is concerned , --root is 1 year old [15:56] it's related to user services it seems [15:56] pedronis: but it's now used for snap start.. in my PR. i need to compare with master [15:57] pedronis: likely, the change was from James [15:57] not sure [15:57] James didn't change that [15:58] at least not in the obvious PR [15:59] * mvo hockey, available via tg for ugent cases [16:01] pedronis: right, he just touched the line but didn't change semantics. anyway, something in my PR to fix, i need to check what did we do in master, most likely --root wasn't and shouldn't be passed in this case [16:02] afaict even on master we pass it everywhere [16:02] ah, no [16:02] sorry, I'm confused is the problem enable or start? [16:04] pedronis: start --enable our our side, which translates to separate enable followed by start [16:06] pstolowski: I see, anyway --root was there since that code started using systemctl in 2016 [16:07] pstolowski: 5106ed3bcb193bf [16:07] PR snapd#9225 opened: many: cloud-init cleanups from previous PR's [16:11] pedronis: yes, but afaict it wasn't used with this exec-command tasks where systemctl was called directly, that's the thing [16:11] anyway i need to eod, appointment [16:12] PR snapd#9226 opened: cmd/snap-bootstrap/initramfs-mounts: compute string outside of loop [16:17] ijohnson: does #9225 need to run with nested tests enabled? [16:17] PR #9225: many: cloud-init cleanups from previous PR's [16:18] zyga, no luck with sudo -l [16:18] pedronis: ah yes good point, let me re-open with that label [16:21] ijohnson: I reviewed the cleanup PRs [16:21] thanks [16:27] PR snapd#9219 closed: run-checks: only check files in git for misspelling [16:29] cachio: can you add to your nested test TODO to fix the regexp here for the ubuntu-16.04-64:tests/nested/core/extra-snaps-assertions test? https://pastebin.ubuntu.com/p/8y5pYY5PQw/ [16:30] pedronis: the last change you requested re cloud-init was in nested.sh from the private reviews, but I pushed it to https://github.com/snapcore/snapd/pull/9208/commits/9d1a8d64f0eb1242ce912875fc9e9f977fe72fe3 since that branch needed a restart anyways and is already nested.sh related [16:30] PR #9208: tests/nested/core20/kernel-failover: add test for failed refresh of uc20 kernel [16:30] pedronis: you originally requested it here: https://github.com/anonymouse64/snapd-private/pull/8/files/bbfe9f9fc266d9cc43fe2c45c44620f4ad6ff2d0#r451357700 [16:30] ijohnson, to fix? [16:30] let me check [16:31] cachio: see the pastebin, the regexp for the core snap is wrong, it doesn't expect to see a pre version there [16:32] PR snapd#9227 opened: snap: add stat to the random access file return interface [16:33] ijohnson, ahhh, the pre + git [16:33] yes [16:33] in my pr [16:33] I am not checkin hte version when the rev ix x1 [16:34] because we test that as part of the listing test in main suite [16:34] so what I did is to simplify what we test in case we build core [16:34] ijohnson, does it make sense for you? [16:35] cachio: sure if it doesn't make sense to check that you can just remove that check that's fine [16:35] just wanted to make sure you knew about the failure [16:38] ijohnson, nice, so that problem is covered in #9098 [16:38] PR #9098: tests: new organization for nested tests [16:39] great [16:51] cachio, can you check if if su -l creates a logind session? [16:51] zyga, sure [16:51] * zyga feels so so [16:52] I think I may be getting the same thing my wife has [16:52] zyga, go to rest [16:52] zyga: oh no [16:53] cachio alternatively try this [16:53] cachio run without sudo [16:53] cachio use sudo to create a systemd unit, maybe via systemd-run or manually [16:53] have the unit run the shell script [16:54] zyga, run the script without sudo? [16:54] and start the unit similarly to how session-tool does it, with runuser -l [16:55] cachio start the script with runuser -l, as the desired user (root) but inside a systemd system unit [16:55] cachio it's all a bit convoluted but that's one scenario that I know should behave correctly [16:55] cachio you can then try to reduce this [16:56] zyga, ok, I'll try that [17:06] cmatsuoka: what's the keyboard shortcut to open a debug shell for a uc20 vm again? I seem to remember it was ctrl+f8 or something but that doesn't do anything :-/ [18:04] ijohnson, cmatsuoka: please review simple bug fix https://github.com/snapcore/snapd/pull/9228 [18:04] PR #9228: interfaces/systemd: compare dereferenced Service [18:04] customer bug [18:04] and my shame [18:04] as I wrote this [18:04]