[01:38] PR snapcraft#3493 opened: cli: rename experimental login environment variable [06:01] morning [07:03] morning [07:04] good morning pstolowski [07:08] good morning [07:12] mvo: zyga: pstolowski morning guys [07:12] :-) [07:12] I wonder what the ZFS investigation turns out [07:16] mvo: did you ping colin-king about zfs-linux or should i? [07:17] good morning mborzecki and zyga ! [07:17] pstolowski: go ahead, I did not see him on irc yesterday so I did not ping him [07:17] hej mvo [07:17] mvo: yeah me neither, will try todaty [07:17] *today [07:17] I will fire up my machine, just need to find a power cable [07:18] pstolowski: thanks! [07:31] ok, my zfs box is up === Ps1-Jack is now known as Psi-Jack [08:54] xnox: hi, fyi https://github.com/snapcore/snapd/pull/10113 [08:54] PR #10113: systemd: wait for zfs mounts [08:56] mvo: re yesterday's point about implicit dependency of snaps units on fs-local.target, i don't see it with systemctl list-dependencies [09:00] oh well, the gpg unit test error keeps delivering... [09:00] didn't Ian look at it? [09:58] PR snapd#10116 opened: gadget: be more flexible with kernel content resolving [11:00] re [11:03] PR snapd#10096 closed: gadget: add support for kernel command line provided by the gadget [11:13] if /tmp/.X11-unix isn't a directory snap-confine moans: https://www.irccloud.com/pastebin/LVMtW7q0/ [11:14] or maybe that's snapd..? someone moans. [11:14] (looking at X11 support for WSL2's native gui capabilities) [11:14] I've updated LP#1922262 [11:19] * ogra hands diddledan a pair of earplugs against the moaning ... [11:19] * diddledan inserts one into each nostrill [11:20] * ogra tikles diddledan to see where they shoot at [11:20] *tickles [11:20] hahaha [11:44] PR snapcraft#3493 closed: cli: rename experimental login environment variable [12:30] pstolowski: hi, I looked at https://github.com/snapcore/snapd/pull/10053 again [12:30] PR #10053: o/snapstate: helper for getting snaps affected by refresh, define new hook [12:33] pedronis: thanks! [12:42] thank you [13:11] diddledan that's expected sadly [13:11] what is it for you? [13:13] zyga: the .X11-unix dir on WSL2 is at /mnt/wslg/.X11-unix - they symlink that to /tmp/.X11-unix (it's all in the bug report) [13:13] oh [13:13] interesting [13:13] oh well [13:13] then it's hard to make that work [13:14] I tried a bind mount, but the snapd bind mount from there results in an empty directory inside the snap world [13:14] i.e. snapd /tmp/.X11-unix bind-> host /tmp/.X11-unix bind-> /mnt/wslg/.X11-unix [13:15] seems that a bind mount of a path that is already bind-mounted from elsewhere doesn't propagate that bind mount chain [13:15] it binds to the underlying directory, not the bind mounted view [13:49] diddledan propagation is a separate topic [13:49] diddledan if you remove the symlink and bind mount it instead: mount --bind /mnt/wslg/.X11-unix /tmp/.X11-unix [13:49] and _then_ start a snap app [13:49] that should work [13:53] mvo: the PR you mentioned in the SU that you want reviews for is #10116 right ? [13:53] Bug #10116: qgis: new changes from Debian require merging [13:53] PR #10116: gadget: be more flexible with kernel content resolving [13:54] zyga: no, because it it bind mounted again, which is why I'm talking about propagation [13:54] diddledan, it's not bind mounted on the host [13:54] I was suggesting to bind mount it on the host [13:55] then once /tmp/.X11-unix is a real directory, not a symlink [13:55] that's precisely what I'm telling you won't (does not - I tried it) work [13:55] then that is explicitly (not propagated) bind mounted on snap startup [13:55] what happened when you tried? [13:55] inside the snap the folder is empty [13:57] can you provide /proc/mountinfo from the host [13:57] and then from snap run --shell [13:59] there is no /proc/mountinfo [13:59] there's a /proc/mounts [14:13] diddledan sorry, I meant /proc/self/mountinfo [14:13] it's a per-process view [14:15] ergh. now it's not failing the same way [14:16] it's actually able to see X now [14:16] I hate heisenbugs [14:17] I wonder how I got it to break before? oh well, you were right, a bind mount seems to fix it [14:19] you had a bind mount after the mount namespace was constructed [14:19] so it was not propagating into it [14:20] ijohnson correct, thanks! [14:22] got it, will take a look [14:59] PR snapcraft#3494 opened: tests: use candidate for autopkgtests [15:15] * cachio lunch [15:16] ijohnson: I approved https://github.com/snapcore/snapd/pull/10061 you can chat with Alex to see if there's any blocker [15:16] PR #10061: interfaces/pwm: add PWM interface [15:16] pedronis: ack sounds good [15:20] and it's snowing again, 3rd time today [15:51] mborzecki yeah [15:51] same here :) [15:51] huge snowflakes [16:14] PR snapd#9171 closed: [RFC] config: "virtual" configuration for timezone <⛔ Blocked> [18:13] * cachio afk [18:28] Issue classic-snap#34 opened: Classic Dimension for Ubuntu Core 20 [19:03] Issue classic-snap#34 closed: Classic Dimension for Ubuntu Core 20 [19:45] PR snapcraft#3495 opened: conda v2 plugin: support for more architectures [23:31] PR snapd#10117 opened: cmd/snap-bootstrap/initramfs-mounts: write active boot-flags to /run [23:36] PR snapd#10118 opened: boot/flags.go: add RunModeRootfs()