[05:52] Good morning [06:05] morning [06:58] PR snapd#6611 closed: interfaces/builtin: add add exec "/" to docker-support (for 2.38) [07:08] Hey mborzecki [07:08] zyga: hey [07:09] https://github.com/systemd/systemd/issues/12018 [07:10] hey zyga and mborzecki [07:10] Hey mvo :-) [07:11] zyga: interesting bug report [07:15] Systemd ran into the same parsing issue [07:16] In that bug report it would really clobber a tmp-foo.mount unit [07:34] back in the office [07:35] * zyga focuses on https://github.com/snapcore/snapd/pull/6575 [07:35] PR #6575: cmd/snap-confine: pass sc_invocation instead of numerous args around [07:35] mvo: when you have a break from backporting systemd patches, could you take a look at https://github.com/snapcore/snapd/pull/6609 ? [07:35] PR #6609: snap/gadget: introduce volume update info [07:38] mborzecki: did you see my question about case sensitivity? [07:40] zyga: yeah, we could probably warn when there's a volume which is case insensitive and more than one 'version' of file is listed in the path [07:47] mborzecki: I think we should not warn, we should use case-insensitive matching [07:47] do the right thing [07:48] mborzecki: and it should be invalid to specifcy case-clashing entries === pstolowski|afk is now known as pstolowski [08:05] morning [08:08] hey pstolowski - good morning [08:28] pstolowski: hey [09:14] kjackal, zyga: I stopped using OVH over this "we'll call it ubuntu but use a random kernel we found down a ginnel" nonsense [09:15] damn, that error with removing of snap data directory is annoying [09:17] mborzecki: and i can't reproduce it on its own :-( [09:17] Chipaca: since it's so well randomly reproducible, i'm actually thinging about adding a list of files in snap data dir in the error message [09:18] mborzecki: sgtm [09:18] mborzecki: find -ls in a debug section? [09:18] hm or that [09:19] let me open a PR [09:22] PR snapd#6615 opened: spread: some debug info to catch /var/snap/.. not being empty [09:26] Chipaca: it's not really reproducing locally either, 40+ loop iterations of install/remove of 2 snaps and still nothing [09:27] mborzecki: what turns on SPREAD_DEBUG_EACH? [09:27] -debug-each ? [09:27] ah, found it [09:28] SPREAD_DEBUG_EACH: "$(HOST: echo ${SPREAD_DEBUG_EACH:-1})" [09:28] you can't turn it on, but you can turn it off :-) [09:29] heh [09:53] Chipaca: you can turn it off [09:53] I use it all the time [09:53] SPREAD_DEBUG_EACH=0 [09:54] hello sir :) [09:54] zyga: that's what i said innit? [09:54] ah [09:54] I read that backwards, sorry [09:55] it's good I got my coffe :) [09:56] Chipaca: can you please look at the commit message in https://github.com/snapcore/snapd/pull/6614 [09:56] PR #6614: cmd/snap-confine: use fixed private tmp directory [09:56] Chipaca: if this lands, I'd like to start removing per-snap /tmp data [09:56] Chipaca: when we remove a snap [09:56] Chipaca: just FYI [09:57] zyga: i reviewed the whole thing, friday [09:58] spent some time remembering the semantics of mkdir and mkdirat to make sure it was ok wrt symlink attack [09:58] zyga: but i then got distracted and didn't +1 or anything [09:59] zyga: but [10:00] zyga: what happens if somebody mkdir's the expected directory, and then mkdir's tmp inside it to not be 01777? [10:00] Chipaca: we fix it on startup === cpaelzer__ is now known as cpaelzer [10:01] ah, there [10:01] * Chipaca found it [10:03] Chipaca: we don't fix it if you break it after taht [10:03] *that [10:03] but that was true before [10:03] and it wasn't hidden, just obfuscated [10:03] if you are root you can break it [10:03] but this is also true now [10:03] zyga: right, root can do w/e [10:04] zyga: but i didn't want a random user to mkdir and then have fun deleting other user's stuff [10:04] but that's fine [10:04] yeah, that's prevented now [10:04] I have one more aspect of that to fix but this is an unrelated bug related to /tmp [10:05] Chipaca: https://api.travis-ci.org/v3/job/507758121/log.txt [10:06] wait what [10:06] * zyga looks at the same one [10:06] that is very much unexpected [10:06] I expected a random directory [10:07] not the current symlink! [10:07] wtf [10:07] yup [10:07] - Remove security profile for snap "test-snapd-tools_instance" (6) (cannot find installed snap "test-snapd-tools_instance" at revision 6: missing file /snap/test-snapd-tools_instance/6/meta/snap.yaml) [10:07] mborzecki: ^ some fuckery going on [10:08] Chipaca: I love our test suite [10:08] perfect job security [10:10] waaaaait [10:10] mborzecki: quick check with you [10:10] - Remove data for snap "test-snapd-tools_instance" (5) (failed to remove snap "test-snapd-tools" base directory: remove /var/snap/test-snapd-tools: directory not empty) [10:10] mborzecki: why is test-snapd-tools_instance removal removing /var/snap/test-snapd-tools ? [10:13] Chipaca: ooh [10:13] feels like incorrect use of snap name vs instance name? [10:14] zyga: or incorrect determination of hasOtherInstances [10:14] Chipaca: seen it happen without any extra _instance entries too [10:22] mborzecki: what does hasOtherInstances() return if you have two instances installed, and remove them both in the same change? [10:22] * Chipaca ponders [10:35] * zyga is in a ethernet cable bonanza [10:35] I can finally fix my messy office wiring [10:37] zyga: is this the part of bonanza where everything is on fire and you ride away into the sunset? [10:38] Chipaca: that is after I plug them :) [10:38] given the time it is probably rather the part where the cook comes out and calls for lunch [11:02] Chipaca: so you're sayng a scenario where both snaps are removed, and at the time hasOtherInstances is run, it returns true for both, but then the directory is left behind? [11:03] * zyga small break to install new network cables [11:03] cables? :P [11:03] mborzecki: wifi runs on cables :)" [11:03] you don't want to see how it looks like now [11:14] 1/3 there, [11:14] now the longest one [11:15] mborzecki: I'm mainly using ethernet because NAS over wifi is not as fast [11:15] mborzecki: my NAS would be noticeably faster if I got the 10Gbit network update now [11:15] mborzecki: wifi here is also very crowded and noisy [11:23] pstolowski: https://www.apple.com/us/shop/ <- roll your 8ball what the updates are [11:23] zyga: interesting [11:24] pstolowski: that's apple talk for "new products today" [11:25] uhm [11:36] pedronis: any concerns if I chery-pick 6598 into 2.38? I think we talked about it and it was fine but I may misremember (the "set a header when auto-refreshing" change from john [11:36] mvo: it's fine [11:37] mborzecki: or conversely, where it returns false to both [11:38] mborzecki, can we get all the selinux stuff to land for 2.38? [11:38] Son_Goku: 2.39 [11:38] arrrgh [11:38] you're killing me here :P [11:38] Son_Goku: sorry :) [11:39] I was hoping to get this in with F30 devel so that the majority of people would be doing reboot upgrades no matter what [11:51] mvo: possible spam edit on https://bugs.launchpad.net/arduino-debug/+bug/1791691 [11:51] Bug #1791691: PATH broken in systemd units [11:58] * Chipaca disappears in a cloud of lunch [11:59] Yum [12:02] PR # closed: snapd#5644, snapd#5822, snapd#5915, snapd#6108, snapd#6177, snapd#6238, snapd#6258, snapd#6325, snapd#6327, snapd#6329, snapd#6341, snapd#6347, snapd#6360, snapd#6367, snapd#6404, snapd#6410, snapd#6418, snapd#6436, snapd#6485, snapd#6491, snapd#6502, snapd#6541, snapd#6559, [12:02] snapd#6564, snapd#6575, snapd#6577, snapd#6583, snapd#6588, snapd#6591, snapd#6594, snapd#6595, snapd#6597, snapd#6599, snapd#6602, snapd#6603, snapd#6604, snapd#6605, snapd#6606, snapd#6607, snapd#6608, snapd#6609, snapd#6610, snapd#6612, snapd#6613, snapd#6614, snapd#6615 [12:03] PR # opened: snapd#5644, snapd#5822, snapd#5915, snapd#6108, snapd#6177, snapd#6238, snapd#6258, snapd#6325, snapd#6327, snapd#6329, snapd#6341, snapd#6347, snapd#6360, snapd#6367, snapd#6404, snapd#6410, snapd#6418, snapd#6436, snapd#6485, snapd#6491, snapd#6502, snapd#6541, snapd#6559, [12:03] snapd#6564, snapd#6575, snapd#6577, snapd#6583, snapd#6588, snapd#6591, snapd#6594, snapd#6595, snapd#6597, snapd#6599, snapd#6602, snapd#6603, snapd#6604, snapd#6605, snapd#6606, snapd#6607, snapd#6608, snapd#6609, snapd#6610, snapd#6612, snapd#6613, snapd#6614, snapd#6615 [12:03] * zyga breaks [12:03] wow @mup [12:10] finishing up with an update to the yocto layer [12:11] hm things were a bit behind [12:12] mborzecki: "a bit" [12:12] mborzecki: any chance you can use snapd.mK/ [12:14] zyga: maybe, although the main problem was handling of go in poky, which is fine to 'just' go packages, but little funky when there's go and something else [12:14] mmm [12:26] re [13:01] Chipaca: restarted the travis job in #6615, want to get another sample [13:01] PR #6615: spread: some debug info to catch /var/snap/.. not being empty <⛔ Blocked> [13:02] off to pick up the kids === ricab is now known as ricab|lunch [13:48] brb, daughter is hungry [14:01] mborzecki, Chipaca we might be a bit late, meeting overrunning [14:39] jdstrand: some low hanging fruit: https://github.com/snapcore/snapd/pull/6607 [14:39] PR #6607: cmd: typedef mountinfo structures [14:39] zyga: yes it is on my list [14:39] thanks! [14:39] it's boring, there are some more interesting ones in the pipe :) === epod is now known as luk3yx === ricab|lunch is now known as ricab [15:55] PR snapd#6616 opened: [RFC] boot: add flag file "force-kernel-extraction" [15:57] pstolowski: I did a pass over the timings PR [16:10] pedronis: k, thanks [16:21] mvo, pedronis, Chipaca I think my fix for error handling of non-nummeric snapshot IDs in snap snapshots commands broke 'snap saved' (2.37.4 is ok) :( [16:22] pstolowski: ok, so we need this for 2.38? [16:23] mvo: when was 2.38 branched away? [16:24] my change landed the week after the sprint. let me check [16:25] pstolowski: ok [16:25] pstolowski: about 2 weeks ago it was branched [16:26] mvo: ok, confirmed, 2.38 requires a fix [16:28] pstolowski: ta [16:28] pstolowski: I will wait for 2.38-final then [16:28] pstolowski: until that is available [16:33] cachio: hey, could you redo the rawhide image? it's cut to version 30 now so GPG is failing https://travis-ci.org/MirServer/mir/jobs/507765669#L1529 [16:33] Saviq, hi, sure [16:33] thanks :) [16:33] I'll ping you when it is ready [16:33] alan_g: FYI ↑ [16:33] Saviq, np [17:13] PR snapd#6617 opened: cmd/snap: fix regression of snap saved command <⚠ Critical> [17:13] mvo: ^ this fix for master === pstolowski is now known as pstolowski|afk [17:20] niemeyer, hey, https://github.com/snapcore/spread/pull/75 readu [17:20] PR spread#75: Make spread tests for spread project run on google backend [17:20] ready [17:31] PR snapcraft#2504 opened: cli: validate plugin schema before provider [17:40] PR snapcraft#2503 closed: cli: disable raven if not running from package [17:43] re [18:12] Saviq, the image is ready [18:12] Saviq please tell me if it works for you [19:47] * cachio afk [23:53] PR snapcraft#2505 opened: Test store