=== mup_ is now known as mup === zigford_ is now known as zigford [03:14] PR snapcraft#2686 closed: New app handler (command-chain and no wrappers if command allows for it) [03:24] PR snapcraft#2560 closed: meta: do not wrap commands by default === zigford_ is now known as zigford [04:58] morning [05:18] PR snapd#7457 closed: tests/main/many: increase kill-timeout to 5m [05:18] PR snapd#7458 closed: interfaces/kubernetes-support: allow systemd-run to ptrace read unconfined [05:50] mvo: morning [05:57] mborzecki: hey, good morning! [06:01] Biking home [06:09] hey zyga [06:09] zyga: nice, already doing a workout? [06:21] yes [06:21] every morning [06:21] uff :) [06:21] this and a standing desk does work wonders [06:21] and every time I complain about my back it's really when I was lazy and skipped one or the other [06:22] mvo: I'm sorry about not finishing work last night, I went to bed at 9 [06:22] last day before paris [06:22] let's make it count! [06:30] brb, new kernel (and gnome) [06:32] re [06:57] mborzecki: you look into 7133 further? should we merge as is and you build on top? or do you want to work in the PR? [06:59] mvo: left anote and merged it [06:59] PR snapd#7133 closed: overlord,daemon: adjust startup timeout via EXTEND_TIMEOUT_USEC using an estimate [06:59] mvo: systemd does not update TimeoutStartUSec even though snapd requests more time, perhaps that's how it's suposed to work anyway [07:04] - google-unstable:ubuntu-19.10-64:tests/main/interfaces-calendar-service [07:04] - google-unstable:ubuntu-19.10-64:tests/main/interfaces-contacts-service [07:04] - google-unstable:ubuntu-19.10-64:tests/unit/go [07:04] these fail on 19.10 [07:04] unit/go fails bc of gofmt === pstolowski|afk is now known as pstolowski [07:06] morning [07:09] zyga, hi, is there a workaround for that limitation about only allowing home directories under /home in snaps? [07:10] zyga, I'm trying to use "lxc" from a jenkins job, but the jenkins user has home set to /var/lib/jenkins, so it fails [07:10] ackk: only one from the user, that is, to bind-mount the actual home directory to /home/$LOGNAME/ [07:10] zyga, I see [07:11] thanks [07:12] pstolowski: hey [07:14] zyga, I also need to update the passwd though? [07:14] ackk: not really [07:14] well [07:14] it's up to you [07:14] zyga, it's not working here [07:14] all that you need to ensure that anywhere your actual $HOME points [07:14] there's a bind mount of that to /home/$LOGNAME [07:15] hmm [07:15] on second though [07:15] thought [07:15] yes, please update passwd [07:15] that's something we don't manage [07:15] I'll make a note [07:15] we have some ideas how to map that internally but that might be derailed by gecos [07:21] zyga, ok, that seems to work, but not it seems snapcraft is looking for build-packages on the host [07:21] not in the container [07:23] oh, it seems the proxy config is not forwarded to the container :/ [07:23] same issue I was hitting with multipass [07:45] PR snapd#7461 opened: run-checks, tests/main/go: allow gofmt checks to be skipped on 19.10 [07:51] zyga: FYI that pi is still going down at regular intervals, despite scheduling the updates only to be done once a month :-\ [07:51] rogpeppe: that's not good, can you give us the scheduling string please [07:51] rogpeppe: as well as "snap changes" [07:52] zyga: will do, when it gets restarted again [07:52] rogpeppe: I procured another PI and I can now replicate your setup [07:52] zyga: have you managed to replicate any of the issues i've been seeing? [07:52] rogpeppe: not yet, I needed to get the right hardware first [07:52] zyga: ok, cool [07:52] rogpeppe: and I was doing other fixes in the meantime [07:53] rogpeppe: I will put regular core-18 on the device and then load your uboot.env to see what it claims [07:53] then I plan to get the same kernel / base [07:53] and same try mode kernel and base [07:53] and see if I can reboot correctly [07:53] with serial attached and all that [07:55] zyga: i'm fascinated to find out if it all works just fine for you :) [07:56] rogpeppe: I hope it doesn't [07:57] rogpeppe: it's much harder to fix something that works for me [07:57] :) [08:08] mvo: I answered to https://github.com/snapcore/snapd/pull/7416#discussion_r324077826 probably degville or Chipaca's input on that is valuable [08:08] PR #7416: seed/seedwriter: start of Writer and internal policy16/tree16 [08:09] pedronis: I'll take a look. [08:11] mborzecki: I tweaked https://github.com/snapcore/snapd/pull/7442 based on your idea earlier [08:11] PR #7442: tests: extend mount-ns test to handle mimics [08:12] the diff is short and pretty comprehensible [08:49] degville: thanks [08:49] pedronis: np. [08:55] pedronis: yeah, sorry for the bikesheed there [08:56] pedronis: aha, sorry, I missed part of the information. its snaps coming from the new "options" file, I missed that piece. let me medidate a bit but with that my suggestions are indeed not good [08:57] mvo: not from a file, but is snaps coming from ubuntu-image --snaps [08:57] or anything external from the model really [08:57] pedronis: yeah, I get it now [08:57] given that is quite generic and ubuntu-image/prepare-image are just the first use case [08:58] pedronis: sorry, that was not clear to me, let me comment in the PR but with that my comment does not make much sense anymore [08:58] pedronis: +1 [08:58] mvo: to be fair as I said the doc comment is not super clear [08:58] and I should fix it [08:58] also OptionsSnap might be less confusable with optional snap [08:59] pedronis: yeah, I like that [08:59] pedronis: and +1 on a comment fix, maybe with a small example like "E.g. OptionsSnaps may come via ubuntu-image --snaps ..." [09:00] yea, as I said I might leave a todo for the actual rename to do later in the chain [09:00] because otherwise I fear conflicts [09:00] all over the place [09:00] in it [09:00] how do I run a clean build with snapcraft --use-lxd ? [09:04] ackk: I'm not sure, sorry, perhaps one of the others knows [09:10] pedronis: +1 [09:16] hey [09:16] a quick question [09:16] does anyone know of "git diff" like tool [09:16] that shows differences inside one line [09:17] like what github can do [09:17] emacs :) [09:19] any command I can try to see it [09:19] zyga, perhaps git diff --word-diff does what you want? https://makandracards.com/makandra/6585-git-diff-changes-in-a-long-line-efficiently [09:19] I'm not after a two semester curriculum ;) [09:19] oh, that's probably exactly that, thank you so much [09:19] np, TIL too [09:24] * Chipaca goes for coffee [09:27] PR snapd#7462 opened: asserts: introduce explicit support for grade for Core 20 models [09:29] PR snapd#7463 opened: tests/main/interfaces-{calendar,contacts}-service: disable on 19.10 [09:30] * Chipaca *really* goes for coffee [09:42] augh [09:48] PR snapd#7460 closed: interfaces/kubernetes-support: allow systemd-run to ptrace read unconfined - 2.41 [09:51] Chipaca: hey, wdyt about https://bugs.launchpad.net/snapd/+bug/1843725 ? should we simply de-duplicate args or error out right away with "nonsense argument given"? [09:51] Bug #1843725: snap install with duplicated snap names gives confusing errors [09:59] pstolowski: we need to make electrified keyboards mandatory [09:59] Chipaca: :) [10:02] pstolowski: more serious answer involves better checking for this case in both snapstate and cmd/snap [10:03] pstolowski: -1 to de-duplication [10:03] mvo: applied feedback: https://github.com/snapcore/snapd/pull/7416/commits/fea6a6c7f694b9489f6655c1af66209f76ba576b [10:03] PR #7416: seed/seedwriter: start of Writer and internal policy16/tree16 [10:27] zyga: about in-line diff tool [10:27] zyga: it's in git itself, but not built [10:28] zyga: $ git config --get core.pager [10:28] /home/john/bin/diff-highlight | less [10:28] zyga: diff-highlight is /usr/share/doc/git/contrib/diff-highlight [10:28] thanks, I'll check that out [10:29] zyga: OTOH, snap install icdiff, git icdiff [10:29] it will come in handy for some of the changes I'm doing :) [10:29] zyga: side-by-side diff with in-line highlighting [10:31] zyga: something's buggy with the pager detection in icdiff so you want to set icdiff.pager explicitly, I need to dig into that this weekend [10:31] but it's otherwise good :-) [10:31] I'll check it out but I think you're way ahead of me :) [10:37] Chipaca: thanks for feedback on the bug === ricab is now known as ricab|bbl [10:51] the diff on https://github.com/snapcore/snapd/pull/7436/files _finally_ makes some sense [10:51] PR #7436: many: make per-snap mount namespace MS_SHARED [10:51] I need to double check that it's good [10:51] but I feel this branch is close [10:52] mborzecki: yeah, as I thought [10:52] mborzecki: I think we still got sharing wrong, even with MS_SHARED [10:52] the per-user stuff is not a real slave [10:52] it's still bi-directional [10:52] I need to check that [10:52] it seems wrong [10:53] and indicates more tests needed [11:06] mborzecki: I commented out the MS_SLAVED | MS_REC change for the per-user mount namespace [11:06] I wonder if it doesn't work [11:06] I want to see the effect it has [11:06] pedronis: wrt your comment about gadgets having bases, does that mean a gadget without an explicit base needs core? [11:06] Chipaca: yes [11:06] pedronis: was there a special value of base that meant "no base needed"? [11:07] none [11:07] PR snapd#6839 closed: devicestate: allow remodel to different kernels [11:07] but we implemented it but is not used [11:07] zyga: can you take another look at https://github.com/snapcore/snapd/pull/7451 ? [11:07] PR #7451: sandbox/cgroup: introduce cgroup wrappers package [11:07] yes, sorry [11:07] looking now [11:07] pedronis: the string "none"? [11:08] the string "none" yes [11:08] pedronis: all these changes to canremove, should they go into the same refactor pr? [11:08] Chipaca: as you prefer [11:09] PR snapd#7464 opened: snapstate: add missing tests for checkGadgetOrKernel [11:12] mborzecki: only small remark about capitalizing sentences and stuff like that [11:13] zyga: ack, thx [11:14] mborzecki: sent now [11:14] https://github.com/snapcore/snapd/pull/7451#pullrequestreview-287952596 [11:14] PR #7451: sandbox/cgroup: introduce cgroup wrappers package [11:15] I'd like to land https://github.com/snapcore/snapd/pull/7442 -- it will show up more behavior of the sharing bug in the next PR [11:15] PR #7442: tests: extend mount-ns test to handle mimics [11:16] PR snapd#7465 opened: snap-confine: fix return value checks for udev functions [11:17] oh, interesting mvo [11:17] I had a look at this code recently, let me double check if this bug is there too [11:17] thank you! [11:19] daaaaaaaaaaaaaaaaaaaaaaaaaaaaamn [11:19] aww crap :) [11:20] zyga: ...? [11:20] Chipaca: I removed one line from snap-confine [11:20] and ... no change to tests [11:20] it's a dead line [11:20] but one we hoped not to be dead [11:20] I'll check out what's going on next [11:28] PR snapd#7466 opened: cmd/snap: special handling of snap set refresh.hold [11:28] test rolling now [11:29] wonder what do you guys thing about 7466 [11:30] yeah, another bug [11:30] maaaaan [11:30] tests [11:30] mborzecki: +1; 'hold' feels like a duration anyway [11:30] thank you [11:30] zyga: https://www.youtube.com/watch?v=dpWmlRNfLck but with 'tests' [11:33] mvo: so, we have another bug, not sure how we never noticed before [11:34] zyga: where? [11:34] https://www.irccloud.com/pastebin/2UL7crS0/ [11:34] this has absolutely no impact [11:34] hah [11:35] I was digging at something that showed up in the MS_SHARED test [11:35] it was showing something that made no sense [11:35] so I kept digging [11:35] and now this [11:35] zyga: mayeb the tests are too limited to catch whetehr this has any impact or not [11:35] no [11:35] they 100% would catch this [11:35] zyga: could it be a 1404ism? [11:35] or something like that? [11:35] 1404, no [11:35] it's really not working [11:36] I'll write a test [11:36] eh, ok [11:36] sucky [11:36] i'm going to go make lunch and let you have your fun [11:36] mmmmmmmmmmmmm lunch [11:36] but yeah, little by little test show crap [11:36] * Chipaca needs food [11:36] oh, I'm hungry now [11:36] don't tell me about food please [11:36] * Chipaca was about to, but stops [11:37] sergiusens, hey [11:37] zyga: I'll tell you that i'm hungry because apparently running 30k makes me hungry for more than 12 hours [11:37] sergiusens, about https://travis-ci.org/snapcore/snapcraft/jobs/584308760#L175 error [11:37] zyga: or maybe it's just my excuse :-) [11:37] Chipaca: congratulations, that's nice! [11:37] * Chipaca goes [11:37] hello! [11:37] sergiusens, this seems to be missing the key to connect to gce [11:38] sergiusens, is this a new travis job? [11:38] yes and no [11:39] sergiusens, what's the new part? [11:39] cachio: https://travis-ci.org/snapcore/snapcraft/builds/584308757/config I moved the unit tests into spread https://github.com/snapcore/snapcraft/pull/2712/commits/b871fd4777d690a3b51c858084e03d27184f33e8 [11:39] PR snapcraft#2712: snap: migrate to core18 [11:40] cachio: in the second link [11:41] cachio: hmm, the only difference is that I am using the snap for spread compared to just downloading it for the spread task that works [11:41] so, you didn't change the project [11:41] or the google key? [11:41] right [11:42] # SPREAD_GOOGLE_KEY [11:43] cachio: do you use the spread snap for your jobs? [11:43] sergiusens, I really don't know how spread snap could affect [11:43] we dont we it [11:43] we download the bin [11:44] hah, that might explain it [11:44] is the spread snap known to be working? [11:44] sergiusens: I don't know if it is maintained [11:45] would be good to check revision history [11:46] sergiusens, we don't use it, and we don't maintain this [11:46] sergiusens, so, not sure [11:46] sergiusens, niemeyer could kno [11:46] w [11:49] mborzecki: and I know why that is now too [11:49] sergiusens: I do use the snap locally [11:50] it's old though apparently [11:51] pedronis: and it reads your google key just fine when exported into the environment? [11:51] yes [11:51] I just used it yesterday [11:53] ok, I will continue checking to see if it is something else, but this is the only diference between a job that works and one that does not on travis [11:56] PR snapd#7467 opened: seed/seedwriter: resolve channels using channel.Resolve* for snaps [11:57] PR snapd#7468 opened: seed/seedwriter: support local snaps [11:59] PR snapd#7469 opened: seed/seedwriter,snap/naming: support classic models [12:07] pedronis: that's a long chain of PRs :-( [12:07] PR snapd#7423 closed: overlord/snapstate: config revision code cleanup and extra tests [12:09] Chipaca: it's not even the end of it, but I'll try not to push more up today [12:09] heh [12:10] Chipaca: it's not visible atm but except the first one that mvo skimmed they are all reasonably sized [12:10] that's why they are many [12:10] pedronis: cool [12:10] pedronis: i'm reviewing 7416 fwiw [12:11] Chipaca: indeed that's the start and the biggest one [12:11] thank you [12:11] Chipaca: I pushed some more in case people want to see where things head to review the start [12:11] btw [12:11] in case [12:11] as well [12:12] very confusing that asserts' core snaps are type "core" [12:12] (well, that's not the confusing bit, but ykwim) [12:14] Chipaca: we could probably do something in types.go [12:14] anyway there is exactly two snaps with type "os" [12:14] :) [12:14] (not counting strange test ones) [12:18] pedronis: wouldn't "AdditionalSnap" have been clearer for OptionsSnap? [12:18] Chipaca: they are not additional [12:18] they can also tweak model ones [12:19] you can see --snap core=edge [12:19] core is not additional there [12:19] s/see/say/ [12:24] pedronis: unrelatedly, you know how to write graphviz graphs, yes? [12:24] yes [12:24] pedronis: graph-easy produces ASCII output from 'em [12:24] pedronis: as in sudo apt install libgraph-easy-perl [12:24] didn't know that [12:25] then 'graph-easy foo.dot' → ascii graph [12:25] PR snapd#7408 closed: tests: fix interfaces-timeserver-control on 19.10 [12:27] PR snapd#7126 closed: tests: part3 making tests work on ubuntu-core-18 [12:33] pedronis: looking at that graph, is the line on the left that goes from Start to SnapsToDownload supposed to be there? The 'no local snaps' label seems to be about the one immediately to its right that goes from LocalSnaps to SnapsToDownload === ricab|bbl is now known as ricab [12:33] and the leftmost one just mirrors the rightmost one === pedronis_ is now known as pedronis [12:40] ok my lunch smells ready now [12:40] * Chipaca grabs a bowl [12:41] Chipaca: right side jumps: if you didn't call SetOptionsSnaps, you can jump straight from Start to SnapsToDownload, left side jumps: if you called it but none of them where local you can also do that, if you don't have local snaps you can also skip InfoDerived even if you called LocalSnaps [12:42] the right side one is a bit redundant, if you have no option snaps then of course you don't local either [12:47] PR snapd#7470 opened: DRAFT: core20 snap install [12:59] PR snapd#7463 closed: tests/main/interfaces-{calendar,contacts}-service: disable on 19.10 [13:04] PR snapd#7471 opened: tests: fix newline and wrong test name pointed out in previous PRs === genii_ is now known as genii [13:29] ijohnson: hey, I know it's not fair because I don't review much lately but https://github.com/snapcore/snapd/pull/7442 is +60,-0 and I've addressed your comments now [13:29] PR #7442: tests: extend mount-ns test to handle mimics [13:29] zyga: sure happy to take a look [14:01] Hey guys can someone guide me through an install of Snappy to install an app for Kali Linux , problem "home is not in /home/user its just in /user" [14:02] I know snappy doesd not function so, but I found a few work around guides but want perhaps more guidance before i just blindly punch in commands [14:10] Chipaca: updated #7411 [14:10] ijohnson: ack [14:10] PR #7411: cmd/model: output tweaks, add'l tests [14:10] ijohnson: i'll get to a savepoint in #7416 and look at that [14:10] PR #7416: seed/seedwriter: start of Writer and internal policy16/tree16 [14:10] PRs have savepoints, right? [14:10] sure, no rush [14:10]