[04:49] <mborzecki> morning
[04:52] <cmatsuoka> mborzecki: integration tests failed for PR #6849
[04:52] <mup> PR #6849: many: introduce a gadget helper for locating device matching given structure <Gadget update> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/6849>
[04:53] <mborzecki> cmatsuoka: hey, yes, just restarted the jobs
[04:53] <cmatsuoka> ah ok
[04:53] <mborzecki> wish we recorded timestamps before the test, and displayed only the relevant part of the journal log, otherwise it's hard to find out what is the issue when you have to scroll through screens of log output
[04:56] <mborzecki> cmatsuoka: hm mkfs.ext* seems to truncate the label on input
[04:58] <cmatsuoka> and this truncation is not on codepoint boundaries, right? or is it codepoint-aware?
[04:58] <cmatsuoka> my guess was that it would just clip on 16 narrow chars
[05:05] <zyga> Good morning
[05:05] <mborzecki> cmatsuoka: heh, so it's bit misleading, dumpe2fs and e2label seem to hide the truncated codepoint
[05:05] <zyga> Hey mborzecki, cmatsuoka :-)
[05:06] <mborzecki> cmatsuoka: but udev properties show the whole sequence: ID_FS_LABEL_ENC=abcdabcdabcdabc\xc4
[05:06] <mborzecki> zyga: hey
[05:06] <cmatsuoka> good morning zyga
[05:07] <cmatsuoka> mborzecki: and how does snapd read the label? could it panic on the string conversion or something like that?
[05:08] <mborzecki> cmatsuoka: so, our problem is reverse, the input is the value that came from gadget yaml, which is valid utf8, then we try to find something under /dev/disk/by-{part}label matching that value, we don't read the labels ourselves
[05:09] <mborzecki> cmatsuoka: in which case, there would be no match
[05:19] <zyga> cmatsuoka: I'm in the plenary room in case you're post breakfast
[05:20] <cmatsuoka> mborzecki: oh ok, we're safe then
[05:21] <mborzecki> cmatsuoka: yup, i'll ask pstolowski|afk to do antoher pass on the PR too
[05:22] <cmatsuoka> zyga: ack
[05:35] <zyga> IT WORKS!
[05:35] <zyga> mborzecki: remember ausearch
[05:35] <zyga> mborzecki: so, found another bug there
[05:35] <mborzecki> zyga: oh?
[05:35] <zyga> mborzecki: it doesn't show *anything* when apparmor denial is for propagation change
[05:35] <zyga> mborzecki: mount --make-shared -- nope, _nothing_ in dmesg, nothing in audit log
[05:36] <mborzecki> zyga: try ausearch -m AVC --debug
[05:36] <mborzecki> it'll show entries that could not be parsed too
[05:36] <zyga> mborzecki: nothing
[05:36] <mborzecki> zyga: unless, there's nothing logged :P
[05:36] <zyga> maybe deeper LSM bug
[05:36] <zyga> but man, that drove me crazy yesterday
[05:37] <zyga> mborzecki: patches coming soon, I need to adjust tests and then shove some of the patches to the earlier branch
[05:37] <mborzecki> btw, are you certain that's caused by apparmor?
[05:37] <zyga> mborzecki: yes
[05:37] <zyga> mborzecki: 100%
[05:37] <zyga> mborzecki: I'm only editing the apparmor profile
[05:37] <mborzecki> mhm
[05:37] <zyga> and now it works, found one last place that lacked the permission
[05:38] <zyga> mborzecki: haha
[05:38] <zyga> ausearch uses 100% cpu
[05:38] <zyga> I think it  failed to parse
[05:38] <zyga> anyway
[05:38] <zyga> happy this is solved now
[07:04] <pstolowski> morning
[07:04] <mborzecki> pstolowski: hey
[07:04] <mborzecki> pstolowski: can you take another look at https://github.com/snapcore/snapd/pull/6849 ? it changed a bit
[07:04] <mup> PR #6849: many: introduce a gadget helper for locating device matching given structure <Gadget update> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/6849>
[07:05] <pstolowski> k
[07:21] <mborzecki> pstolowski: thanks for the review
[07:22] <pstolowski> yw
[08:08] <mup> PR snapd#6867 opened: gadget: offset-write: fix validation, calculate absolute position <Gadget update> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/6867>
[09:47] <Chipaca> could I get reviews for #6862? the check-health pr is ready to go, but it includes this and is rather big so keeping it lean would be nice
[09:47] <mup> PR #6862: overlord/hookstate: don't run handler unless hooksup.Always <Created by chipaca> <https://github.com/snapcore/snapd/pull/6862>
[10:24] <mup> PR snapd#6868 opened: systemd: workaround systemctl show quirks on older systemd versions <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/6868>
[10:29] <Chipaca> mborzecki: nice
[10:30] <mborzecki> Chipaca: it'd be nicer if we didn't have to do that :)
[10:30] <Chipaca> totes
[10:32] <mborzecki> Chipaca: noticed we don't really have a spread test for snap services output, so the problem was not caught earlier
[10:35] <Chipaca> mborzecki: noticed
[10:54] <Chipaca> Stepping out for a walk and lunch. I'd call it a wunch, but that's the collective noun of 'banker'.
[11:04] <mborzecki> meh, gofmt 1.10 vs gofmt 1.11 is super annoying
[11:13]  * pstolowski lunch
[11:23] <cachio> degville, hey
[11:23] <degville> cachio: hello!
[11:32] <mup> PR snapd#6849 closed: many: introduce a gadget helper for locating device matching given structure <Gadget update> <Created by bboozzoo> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/6849>
[12:11] <mborzecki> off to school, back before standup
[12:51] <diddledan> @jdstrand I appreciate I took forever and a day to try your suggestions, but we have a successfully working makemkv with some fiddling with the apparmor and udev rules: https://forum.snapcraft.io/t/openat-returning-eperm-for-dev-sg0/10069/15
[13:00] <zyga> Hey hey
[13:01] <zyga> mborzecki: found one more bug in the same area
[13:01] <zyga> The synthetic mount change retention logic is incorrect
[13:16] <mborzecki> cachio: Fedora-Cloud-Base-30-1.2.x86_64.qcow2 seems to behave just fine, just updated the kernel, was at 5.0.9-301.fc30.x86_64, rebooted and 5.0.13-300.fc30.x86_64 is running now
[13:17] <cachio> mborzecki, weird
[13:17] <cachio> I am using http://fedora.c3sl.ufpr.br/linux/releases/30/Cloud/x86_64/images/Fedora-Cloud-Base-30-1.2.x86_64.raw.xz
[13:18] <cachio> it is the same
[13:18] <mborzecki> cachio: shuld be the same
[13:18] <cachio> yes
[13:18] <mborzecki> cachio: all i'm doing is sudo dnf upgrade -y && sudo reboot
[13:18] <cachio> I'll need to review if something installed during the image creation is causing this
[13:19] <mborzecki> pstolowski: idk if you noticed, but tests/main/auto-refresh-retry is failing more often now, usually the last line is:
[13:19] <mborzecki> May 14 11:30:49 arch nsenter[2289]: retry.go:113: DEBUG: Retrying because of temporary net error (DNS): &net.DNSError{Err:"Temporary failure in name resolution", Name:"api.snapcraft.io", Server:"", IsTimeout:false, IsTemporary:false}
[13:32] <pstolowski> mborzecki: hmm i saw it once
[13:34] <mborzecki> cachio: i'm looking at fedora, it's using some bootloader specification now, https://systemd.io/BOOT_LOADER_SPECIFICATION
[13:35] <cachio> mborzecki, I updated the scripts to create the image
[13:35] <cachio> mborzecki, I am running now
[13:39] <Chipaca> mborzecki: nice mess, the volumes thing
[13:43] <mborzecki> Chipaca: heh ;) maybe the spec is user friendly at least
[13:44] <mborzecki> for sure it's not parser friendly
[13:44] <Chipaca> mborzecki: the yaml is very easy to follow
[13:44] <Chipaca> mborzecki: but even the names of the things in there seem to be designed to conflict with other homonymous things
[13:45] <Chipaca> anyhoo
[13:45]  * Chipaca makes tea
[14:45] <jdstrand> diddledan: ack, thanks! :) I'm sprinting this week and will give it some thought. I've added it to my list for profile additions for 2.40. I too am not sure if it should be a separate interface or not
[14:46] <diddledan> cool, many thanks :-)
[14:46] <jdstrand> thank you :)
[14:58] <Chipaca> brb rebooting
[15:19] <cachio> mborzecki, it is fixed the issue with the kernel
[15:19] <cachio> on fedora
[15:19] <mborzecki> cachio: what was the fix?
[15:20] <cachio> Initially I was installing kernel-modules with the base image
[15:20] <cachio> now I am installing kernel modules and kernel core during the update
[15:20] <cachio> then reboot
[15:21] <cachio> basically
[15:21] <cachio> I was installing kernel-modules then doing upgrade then instlaling kernel-core then during upgrade
[15:21] <cachio> in 2 scripts
[15:21] <mborzecki> ah
[15:22] <mborzecki> cachio: so once you push the new image, we can restart the PR with fedora 30
[15:22] <cachio> now I upgrade, then install the kernel-modules and kernel core and then reboot
[15:23] <cachio> mborzecki, done
[15:24] <cachio> thanks for the support
[15:36]  * cachio lunch
[17:19] <mup> PR snapd#6865 closed: cmd/snap: allow option descriptions to start with the command <Created by chipaca> <Merged by chipaca> <https://github.com/snapcore/snapd/pull/6865>