[01:58] PR snapd#7201 closed: HACKING.md: GitHub not Github [05:25] morning === pstolowski|afk is now known as pstolowski [07:22] mornings [07:22] pstolowski: hey [08:02] PR snapd#7203 opened: i18n, vendor, packaging: drop github.com/ojii/gettext.go, use github.com/snapcore/go-gettext <⛔ Blocked> [08:04] pstolowski: can you take a look at https://github.com/snapcore/snapd/pull/7186 ? [08:04] PR #7186: gadget: ensure filesystem labels are unique [08:14] PR snapd#7204 opened: cmd/snap: use showDone helper with 'snap switch' [08:15] Chipaca: good morning! something for you sir ^ [08:15] mborzecki: will do [08:16] pstolowski: nice. Will look in a bit. [08:20] pstolowski: thanks [08:20] Chipaca: morning [08:20] mborzecki: mo'in [08:22] PR snapd#7175 closed: gadget: effective structure role fallback, extra tests [09:03] is mvo the only one with write access to https://github.com/snapcore/go-gettext ? [09:04] Chipaca: pstolowski: can you take a quick look at https://github.com/snapcore/go-gettext/pull/2 ? [09:04] PR go-gettext#2: Fix unit tests on hosts where /tmp is on a separate device [09:07] anyways, wonder why this didn't come up earlier [09:13] k [10:01] mborzecki: we weren't running the unit tests of it i guess? [10:02] Chipaca: or /tmp was on the same device as the source dir :/ [10:02] mborzecki: or they set TEMPDIR [10:02] or is it TMPDIR [10:03] Chipaca: or that ;) [10:13] mborzecki: can you take a look at #7204? [10:13] PR #7204: cmd/snap: use showDone helper with 'snap switch' [10:13] pstolowski: sure [10:45] mborzecki: thanks === alan_g is now known as alan_g_ [10:55] hm weird, perhaps soemthing wrong with our spread images, this minimal task.yaml https://paste.ubuntu.com/p/B4ZBQc5pnT/ REBOOT fails on 16.04 but works on 18.04 [11:06] sorry, I asked this before, but my PC may have fallen asleep by the time I got a response. Anyone think of a reason why snapd 2.40 complains unless I add libncurses to the apparmor profile? [11:14] zigford: complains how? [11:20] 1 sec, it was about a week ago that I tested. I'll remove my entry and repro [11:21] Chipaca: do you have any idea what could be wrong here? https://paste.ubuntu.com/p/988jjnTcw2/ [11:22] "/bin/sh: error while loading shared libraries: libncurses.so.6: cannot open shared object file: No such file or directory [11:22] child exited with status 127" [11:22] "apparmor="DENIED" operation="open" profile="/usr/lib64/snapd/snap-confine" name="/lib64/libncurses.so.6.1" pid=13806 comm="snap-device-hel" requested_mask="r" denied_mask="r" fsuid=0 ouid=0 [11:22] " [11:22] mborzecki: what am i looking at? [11:22] zigford: what distro is this? [11:23] Sorry, that's after running snap run packagename [11:23] Gentoo [11:23] Chipaca: a spread restore script, executed on 16.04, returning 1, same script run on 18.04 returns 213 (as expected) [11:23] If I add an entry to apparmor.d snap-confine profile, the snaps run fine. [11:24] mborzecki: why doesn't it have a shebang? [11:24] Just wondering if something had changed in snapd that I need to account for in my snapd ebuild [11:25] Chipaca: running bash script.sh gives the same result, 1 on 16.04, 213 on 18.04 [11:26] zigford: there is a libncurses entry in the apparmor profile of snap-confine, at least on master [11:26] mborzecki: hmmmm [11:26] mborzecki: does shellcheck say anything? [11:26] Chipaca: nothing useful [11:27] Chipaca: is it libncurses or libncursesw? the makefile that generated the profile has libncursesw not libncurses [11:27] mborzecki: w [11:28] although why it even has that is unclear to me [11:28] er [11:28] zigford: libncursesw [11:29] zigford: actually, it used to have both i think? [11:29] ec4f3c07746 added both [11:30] Chipaca: curious. Guess the best bet might be to dig through the commits and see if/when/why it changed. I know Gentoo isn't supported, but I want to do my best to keep it running [11:33] er [11:33] mborzecki: the subshell [11:36] * Chipaca pokes more [11:38] Chipaca: strace is a bit confusing too https://paste.ubuntu.com/p/VrXKfhQhkf/ [11:38] Chipaca: notice how on 16.04 3359 exits with 213, but then 3358 exits with 1 :?? [11:42] mborzecki: so [11:42] mborzecki: on 16.04 [11:42] ( foo ) [11:42] actually, let me double check that [11:44] yes [11:44] so [11:44] mborzecki: on 16.04, given: boo() { exit 213; } [11:44] mborzecki: and 'set -e' [11:44] mborzecki: then ( ( boo ); ( echo hi ) ) [11:45] exits with 1 [11:45] mborzecki: replace the outer subshell-causing ()s with {}s and it exits with 213 [11:46] mborzecki: a workaround [11:46] mborzecki: ( boo ) || exit $? [11:47] so, two workarounds [11:48] either replace the outer subshell with a group, or, explicitly re-raise the exit value [11:48] Chipaca: yet, we still need to get the fix through spread :/ [11:48] mborzecki: when did this start breaking? [11:48] what changed? [11:50] Chipaca: noticed this in https://github.com/snapcore/snapd/pull/7198 [11:51] PR #7198: tests: reboot the node when restoring after a test involving lxd [11:56] mborzecki: so I imagine the issue is that the restore is run in a further subshell, when compared to the test itself [11:56] mborzecki: but I don't know, care to check that? [11:57] the other REBOOT-using tests use it in the body afaik [11:59] anyway I should make lunch [11:59] and then, apparently, i need to go to the bank [11:59] mborzecki: can you lead the standup if I don't make it back in time? [12:01] Chipaca: sure [12:01] Chipaca: commented out restore-each in my spread.yaml, https://paste.ubuntu.com/p/6MkKSf9sqy/ [12:01] mborzecki: yes [12:02] mborzecki: I'd call it a bug in bash :) [12:02] Chipaca: yup, trying to find a bug report or a changelog entry [12:02] PR snapd#7204 closed: cmd/snap: use showDone helper with 'snap switch' [12:02] Chipaca: since it's fixed in newer version anyway :) [12:05] Good morning [12:05] I’m sorry for not doing more last night [12:06] I sent some reviews but didn’t have time to work on blockers [12:06] mborzecki, Chipaca what does reboot expand to? [12:07] Is it the function or the executable [12:07] Reboot returns 130 AFAIR [12:07] Did you check what spread does to detect a reboot request? [12:08] zyga: awe already know that set -e ( () () ) behaves differently in bash 4.3 and 4.4 [12:09] zyga: the bottom snippet here shows how the script looks like in the minimal test case that fails https://paste.ubuntu.com/p/6MkKSf9sqy/ [12:09] Bash, the bane of development :-( [12:09] Thank you for digging [12:09] I read that [12:10] Did you see my question about command vs function? [12:10] zyga: I don't think that makes a difference [12:10] zyga: it's complicated, REBOOT is a function which echos and exists with special exit code (213), the spread separately runs `reboot` command [12:11] zyga: you can change the 'REBOOT' for a direct 'exit 213' and it still misbehaves [12:11] mborzecki: REBOOT can also be tests/lib/bin/REBOOT which is what zyga is asking about i think [12:11] mborzecki: we also unset the function and do the equivalent in a reboot command [12:12] I wonder if that matters [12:12] Yeah [12:12] but, as i say, it's not the function, it's the double subshell followed by a subshell [12:12] or as mborzecki said [12:12] anyway why am i not making lunch [12:12] Chipaca: i know, though i'm using a minimal task sample https://paste.ubuntu.com/p/B4ZBQc5pnT/ outside of your spread setup ;) [12:12] zyga: ^^ [12:13] Ack [12:13] zyga: with 'set -e', this: ( ( exit 213 ); ( echo hi ) ) [12:13] zyga: exits with 1 [12:13] ¯\_(ツ)_/¯ [12:22] duh, bash git log is most useless :/ [12:25] (set -e ; ( ( exit 213 ) || exit $?; ( echo hi ) || exit $? ) || exit $? ); echo $? [12:25] omg /o\ [12:54] PR snapcraft#2651 opened: Extension confinement [12:57] PR snapcraft#2648 closed: remote-build: include project name in build-id [12:57] PR snapcraft#2650 closed: store: send snapcraft-started-at in push requests [13:01] cachio: standup? [13:37] cachio: I talked to field and they said there's a ufw snap - would that work for you? if not, then iptables is included in network-manager, but there's no app exposed for it so you'd need to do `snap run --shell ...` [13:38] ijohnson, nice, thanks, I'll try ufw snap [13:38] +1 [13:39] and see if it works on the tests [13:39] ijohnson, thanks!! [13:39] if you have problems with it, blame ogra he recommended it to me so it's all his fault [13:40] ijohnson, hehhe [13:40] ijohnson, I'll do [13:40] thanks [13:44] ijohnson, cachio i'll pass that blame on to the owner of the snap .... jdstrand_ ;) [13:48] what do we do with changes to spread? shall I open a PR? [13:49] pstolowski, i'm just wondering, what happens to my hotplugged usb-serial device if i unplug and re-plug ... will the interface an app uses stay connected ? [13:53] ogra: when you unplug the device, the interface gets disconnected and slot disappears from the system; when you plug it back, the connection is restored [13:53] prefect, thanks ! [14:09] funny: https://www.reddit.com/r/ProgrammerHumor/comments/btkri6/i_call_for_death_penalty/?ocid=AID740620_FACEBOOK_oo_spl100000841935315 [14:27] pstolowski: would you like to join the meeting about the channel vs risk thing? [14:27] pstolowski: wrt snapd-glib and the semantics change [14:27] Chipaca: sure [14:27] pstolowski: ok, i'll add you to the invite [14:27] it starts in ~3 :) [14:28] Chipaca: ok, np [14:43] man, i love short meetings of busy but smiling people [14:43] haha [14:44] Chipaca: thanks for overseeing this change & making Robert aware of it [14:45] "overseeing" is a big label :-) [14:45] * Chipaca adds "overseer" to his business cards [14:46] lol [14:47] * ogra senses the next holliwood movie "The Overseer - featuring J. Lenton as the overseer" [14:47] * Chipaca upgrades his cash-raking rake [14:51] Chipaca, pstolowski I just noticed the current G-S code doesn't use the full channel names, but I was intending to change that (to match snapcraft.io). Either way this still shouldn't be an issue. [14:52] robert_ancell: 👍 [14:52] robert_ancell: good, ty [14:53] * cachio lunch === pstolowski is now known as pstolowski|afk [16:28] PR snapcraft#2652 opened: meta: transparently support command-chain [19:59] PR snapd#7205 opened: many: introduce security profile failsafe flag [21:10] * cachio EOD [21:34] kenvandine popey: fyi my lame attempt to get drawing snapped yesterday evening - https://github.com/alexmurray/drawing-snap [21:56] amurray: I have one that's pretty close. Needs newer things, which are in the gnome build snap [21:56] amurray: which is still a WIP === JamieBennett_ is now known as JamieBennett === cprov_ is now known as cprov === marosg_ is now known as marosg === kalikiana_ is now known as kalikiana === tedg_ is now known as tedg === jamespage_ is now known as jamespage === kenvandine_ is now known as kenvandine === zyga_ is now known as zyga === benoitc_ is now known as benoitc === joedborg_ is now known as joedborg