[03:33] PR snapd#8943 opened: wrappers: generate D-Bus service activation files [05:23] morning [06:10] hmm ubuntu-core-20 prepare sometimes fails right at the start, before we transtion from 20.04 host to actual uc20 https://paste.ubuntu.com/p/XrZQbS6mFf/ [06:21] mvo: hey [06:29] mborzecki: good morning [06:39] PR snapd#8912 closed: tests: Remove unity test from nightly test suite [06:40] mvo: there was an interesting failur in prepare of uc20: https://paste.ubuntu.com/p/XrZQbS6mFf/ [06:40] mvo: happens right at the beginning when we purge snapd on the host [06:41] surprisingly, haven't seen it in regular 20.04 tests [06:48] mborzecki: looking [06:49] mborzecki: so purge did not actually work and we have leftover data? [06:49] mborzecki: my theory is that we need to stop snapd earlier in postrm [06:49] mvo: that's your pr right? [06:49] mborzecki: still does not explain everything, but something like 8883 [06:49] mborzecki: I was thinking about it [06:49] mborzecki: maybe we need to do the cleanup in two phases [06:50] mborzecki: or N phases, cleanup, stop snapd, cleanup again if anything is left [06:50] mborzecki: to ensure we catch stuff in flight - just a theory right now [06:50] mhm [06:51] mborzecki: would be cool to get the state.json content [06:51] mborzecki: maybe that's the next debug output and/or snapd journal output, then we can see if e.g. snapd is restarting whle we purge it [06:51] also, debug section is apparently not executed when prepare fails [06:53] mborzecki: oh no [06:53] mborzecki: meh, ok [06:59] PR snapd#8944 opened: tests/lib/prepare-restore: collect debug info when prepare purge fails [06:59] mvo: ^^ ok, let's see if anythign comes up [07:01] mborzecki: \o/ thank you so much [07:02] morning [07:02] mborzecki: the state.json maybe a bit unwieldy but pushing it through python3 -m json.tool will just make it incredible long :/ and jq seems overkill (plus we may important info) so let's do it [07:03] (do it this way) [07:03] hey pstolowski good morning! [07:03] pstolowski: hey [07:03] mvo: hmm it should be failry empty at this point, it's a clean system [07:03] (or at least should be) [07:04] mvo: i don't want to interrupt this run, so let's wait to see what happens and i can push the python bit next [07:04] mborzecki: aha, indeed, it will just have the seeding for lxd/core18 - good point [07:05] mborzecki: yeah, no python bits needed I think, was mostly thinking alound/wanted to talk it over. I think we can paste locally and run through jq or similar [07:49] hello [07:49] sorry for being offline, I was working on the other laptop and I haven't set IRC up there yet [07:49] updated refresh backend branch and fixed two failing tests [07:50] how are things? [07:53] hey zyga [07:53] hey mvo, welcome back [07:56] zyga: hey [08:04] mvo: haha https://github.com/snapcore/snapd/pull/8944 passed on uc20 [08:04] PR #8944: tests/lib/prepare-restore: collect debug info when prepare purge fails [08:04] PR snapd#8925 closed: bootloader: allow managed bootloader to update its boot config [08:08] mborzecki: of course it did ;) [08:08] mborzecki: fun with races [08:08] mborzecki: I guess the rule is "if there is a race you always loose - even when you win!" [08:08] mborzecki: let's just merge it and we will get good debug in one of the subsequent failures, wdyt? [08:09] yup [08:09] mvo: mborzecki: hi, #8817 needs 2nd reviews [08:09] PR #8817: tests: new test to validate refresh and revert of kernel and gadget on uc20 [08:10] zyga: pstolowski: can you take a look at https://github.com/snapcore/snapd/pull/8944 ? [08:10] PR #8944: tests/lib/prepare-restore: collect debug info when prepare purge fails [08:10] sure [08:11] yep [08:11] mborzecki merge it! [08:12] i made a syggestion... to late [08:12] *too [08:12] pstolowski: hah, do we have a wrapper for -m json.tool? [08:14] hmm nope, also very few test use json.tool; no tests use json_pp [08:14] PR snapd#8944 closed: tests/lib/prepare-restore: collect debug info when prepare purge fails [08:18] * zyga pushed to https://github.com/snapcore/snapd/pull/7825 - only ~ 600 lines change now [08:18] PR #7825: many: use transient scope for tracking apps and hooks [08:20] pstolowski I added one more comment to https://github.com/snapcore/snapd/pull/8939/files#r447500153 [08:20] PR #8939: [RFC] snap-confine: don't die if a device from sysfs path cannot be found by udev [08:22] mborzecki I made one more, super small this time, branch carved out of the big backend branch: https://github.com/snapcore/snapd/pull/8938 [08:22] PR #8938: sandbox/cgroup: extend SnapNameFromPid with tracking cgroup data [08:26] pedronis: reviewing something else right now but will put it on my list [08:34] PR snapd#8817 closed: tests: new test to validate refresh and revert of kernel and gadget on uc20 [08:39] PR snapd#8923 closed: wrappers: helper for enabling services (8/9) [08:42] need to run an errand, i'll probably work from some cafe or sth for a while, bbl [08:51] brb [09:19] PR snapd#8940 closed: tests: fix "restart.service" [09:59] PR snapd#8945 opened: tests: fix leaked dbus-daemon in selinux-clean [10:21] and re, at least for some time [10:30] PR snapd#8946 opened: client: move all error kinds into errors.go and add doc strings [10:34] pedronis: added a comment under https://github.com/snapcore/snapd/pull/8930#discussion_r446897486 [10:34] PR #8930: many: managed boot config during run mode setup [10:35] it should probably be both flags tbh (if we ever have a script that's not aware of kernel.efi at some point) [10:47] mborzecki: answered, still we should add the TODO [10:48] ok [11:06] re === msalvatore_ is now known as msalvatore [12:00] PR snapd#8947 opened: many: update managd boot config when refreshing snapd [12:03] zyga: two independent reports for a namespace error when refreshing gimp: https://bugs.launchpad.net/snapd/+bug/1884849 [12:03] Bug #1884849: Error message trying to refresh or uninstalling a snap ( gimp ) [12:16] Bug #1873452 changed: 'snap install' | "INFO Waiting for restart..." [12:22] Bug #1835342 changed: disconnect fails if plug is gone from snap meta [12:30] PR snapd#8899 closed: tests: add servicestate.Control tests (7/9) [12:37] Bug #1835342 opened: disconnect fails if plug is gone from snap meta [12:43] Bug #1835342 changed: disconnect fails if plug is gone from snap meta [14:06] ijohnson: there's some little tweaks in https://github.com/snapcore/snapd/pull/8918 and it could land, i can push a patch there if you haven't started with the fixes yet [14:06] PR #8918: many: make nested spread tests more reliable [14:08] mborzecki: no unfortunately I haven't had time to apply those yet, so feel free to do so, thanks! [14:09] ijohnson: cool, will do [14:29] zyga, hey, about the rename of the functions on the core-config helper, do you have something in mind? [14:29] zyga, I though prepare_model -> prepare_core_model [14:30] zyga, does it make sense? [14:36] PR snapd#8948 opened: cmd/snap-update-ns: detect unmounted mounts [14:49] zyga, I already updated the names,. perhaps this is better [14:50] https://github.com/snapcore/snapd/pull/8918 needs a 2nd review [14:50] PR #8918: many: make nested spread tests more reliable [15:22] * cachio lunch [17:16] PR snapd#8945 closed: tests: fix leaked dbus-daemon in selinux-clean [17:39] hey recently I had an issue where after a dist upgrade all my snaps lost there connections... e.g. plugs and slots all disconnected [17:40] so I ran into some interesting things while poking around... 1. I can't find a snap command to list the default connections from meta/snap.yaml [17:42] 2. I was thinking that somewhere in connections it might list the default/installed state but I can't seem to see from the output where that might be indicated [17:43] it just connects them at install time ... so snap connections will just show the default ones as connected [17:43] 3. is there a command that can restore a snap to it's default connections, might be helpful for users that want to reset a snap if they've manually connected or disconnect things [17:43] ogra: so defaults as connected currently right? [17:43] anything you connected later will be tagged as "manual" [17:43] (in the most right column of the connections output) [17:43] it's interesting, my output shows - for notes on all of them with all of them disconnected [17:44] naturally the app crashes [17:44] ogra@pi4:~$ snap connections fabrica [17:44] Interface Plug Slot Notes [17:44] lxd fabrica:lxd lxd:lxd manual [17:44] mount-observe fabrica:mount-observe :mount-observe manual [17:44] network fabrica:network :network - [17:44] network-bind fabrica:network-bind :network-bind - [17:44] ssh-keys fabrica:ssh-keys :ssh-keys manual [17:44] system-observe fabrica:system-observe :system-observe manual [17:44] bah ... [17:45] so here (even though the formatting is totally screwed) you can see that network and network-bind were connected by default [17:45] in my case if the default output lists a plug with no slot... is it listed because it was defaulted? [17:45] Interface Plug Slot Notes [17:45] alsa google-play-music-desktop-player:alsa - - [17:46] audio-playback google-play-music-desktop-player:audio-playback - - [17:46] ... [17:46] desktop google-play-music-desktop-player:desktop :desktop manual [17:46] hmm, audio-playback should definitely autoconnect ... as should desktop [17:46] I manually started reconnecting but I thought it an odd thing to have to do [17:46] it is [17:46] the obvious thing here is to just re-install [17:47] the snap [17:47] either open a topic in the forum (see channel topic) or open a bug on launchpad [17:47] but I thought it an interesting thing to have happened [17:47] this should definitely not happen [17:47] ogra, thanks just wanted a sanity check [17:47] FSVO "interesting" ... 🙂 [17:47] I thought maybe I was missing something [17:48] nah, thats surely a bug [17:48] interesting thing here is I've not idea how to reproduce the issue as this happened after a dist upgrade of popos [17:48] who the hell knows what happened during that process [17:48] LOL [17:48] "popos" ... /me grins [17:49] (popo is a nickly term for "butt" in german ... ) [17:50] haha [17:50] I'll have to avoid that with my sister in law [17:50] our family likes to use mushy for things that are squishy and she gets a kick out of that too [17:51] haha [18:21] PR snapd#8949 opened: tests: new fs-tool which replaces the files.sh helper [20:22] PR snapd#8950 opened: tests: new str-tool which replaces the strings.sh helper [20:27] PR snapd#8924 closed: gadget, bootloader: preserve managed boot assets during gadget updates [20:27] PR snapd#8951 opened: gadget/install: move udev trigger to gadget/install === pedronis_ is now known as pedronis [20:45] PR snapcraft#3192 closed: Flutter [21:05] PR snapcraft#3193 opened: extensions: plug the opengl interface for GNOME [23:18] PR snapd#8952 opened: tests: enable tests on uc20 which now work with the real model assertion [23:58] PR snapd#8953 opened: tests: enable system-snap-refresh test on uc20