[05:06] morning [05:48] PR snapd#7530 opened: tests: move "centos-7" to unstable systems (2.42) [06:06] mvo: morning [06:06] hey mborzecki [06:16] PR snapd#7530 closed: tests: move "centos-7" to unstable systems (2.42) [06:17] PR snapd#7525 closed: tests: disable {contacts,calendar}-service tests on Arch Linux (2.42) [06:18] PR snapd#7524 closed: tests: add unit test for gadget defaults with a multiline string [06:19] PR snapd#7523 closed: sandbox/selinux: move SELinux related bits from 'release' to 'sandbox/selinux' [06:21] mvo: can you take a look at https://github.com/snapcore/snapd/pull/7507 ? [06:21] PR #7507: gadget: do not fail the update when old gadget snap is missing bare content [06:21] mborzecki: sure [06:22] good morning everyone [06:22] some weather ... [06:24] mvo: thanks [06:24] zyga: hey [06:55] hm can't launch anything under mulitpass [06:55] launch failed: Failed to set iptables rule for table filter: iptables v1.6.1: unknown option "--dport" [06:59] --beta works though === pstolowski|afk is now known as pstolowski [07:01] morning [07:02] good morning [07:02] Saviq: filed https://github.com/CanonicalLtd/multipass/issues/1100 let me know if you need more info [07:03] pstolowski: hey [07:12] hey pstolowski [07:18] PR snapd#7507 closed: gadget: do not fail the update when old gadget snap is missing bare content [07:18] yay :) [07:19] mvo: can you cherry-pick that to 2.42? the commit hash is 31d33fa4e [07:22] mborzecki: thanks, done [07:33] PR snapd#7531 opened: cmd/system-shutdown: include correct prototype for die [07:34] mvo: there's one more thing to cherry pick [07:34] it landed when you were away, one sec [07:34] mvo: https://github.com/snapcore/snapd/pull/7526 [07:34] PR #7526: cmd/snap-confine: allow digits in hook names [07:37] mvo: and perhaps https://github.com/snapcore/snapd/pull/7531 as well [07:37] PR #7531: cmd/system-shutdown: include correct prototype for die [07:37] mvo: so that I can drop all patches from 2.42 in suse [07:38] zyga: thanks, looking [07:38] I'll check why our builds don't catch that though, must be -Wsomething we're missing [07:38] zyga: 7526 is in [07:41] Thank you, ondra will be very happy this made it in [07:52] mborzecki: ping about https://github.com/snapcore/snapd/pull/7484 [07:52] PR #7484: osutil: generalize SyncDir with FileState interface [07:52] mvo: not sure if you want to pull that into 2.42 [07:52] it's the memory optimization for icons [07:53] it needs reviews [07:55] zyga: no [07:56] zyga: or *maybe* but it needs to be reviewed and obviously fixing something :) [07:56] yeah, just asking because you indicated you might be interested in that at the sprint when it came up [07:56] but I think it's okay to get this in 2.43 [07:57] we can use store to prevent people from sending huge icons [07:59] #7277 needs 2nd review [07:59] PR #7277: overlord/snapstate: fix undo on firstboot seeding [08:38] * Chipaca has successfully bred some kind of a throat bug over the weekend and feels this was a poor idea and would like to cancel his subscription to it [08:45] re [08:46] Chipaca: so you have my copy? :D [08:46] get well man [08:53] zyga: eh, just my throat catching a lot and the general sense of fighting something bringing me down a bit, nothing that should stop me (might slow me down some) [08:56] Chipaca: hi, sorry about that [08:57] Chipaca: #7402 can be merged ? [08:57] PR #7402: daemon, client, cmd/snap: include architecture in 'snap version' [08:59] pedronis: heading into a meeting, but … probably? [09:36] PR snapd#7516 closed: daemon: allow /v2/assertions/{assertType} to query store [09:37] PR snapd#7402 closed: daemon, client, cmd/snap: include architecture in 'snap version' [09:41] PR snapd#7531 closed: cmd/system-shutdown: include correct prototype for die [09:43] heavy rain here, hope no funny stuff with power lines toda [09:45] "hope no funny stuff with power lines toda--" KOOOOMzapWHUMP [09:54] * zyga -> quick tea [10:23] zyga: re 7484 - I wonder if it would make sense to push a PR that just adds the osutil.FileState interface and does all the mechanical stuff (no real code changes) [10:23] mvo: I can do that [10:24] note that FileState became an interface from a struct [10:24] so it would involve a lot of changes already [10:24] not sure if that's what you had on your mind [10:24] zyga: yeah, thats what I mean [10:24] let me try [10:25] zyga: it might be simpler to review one PR with just this mechanical change FileState->interface previous FileState->MemoryThing no other changes [10:25] zyga: i.e. no adding of FileReference or anything [10:25] zyga: then it becomes a mechanical review and the actual implementation change is smaller and more obvious [10:25] mhm [10:26] man, what a downpour! [10:26] (but of course if e.g. pedronis thinks 7418 is fine as is and should not be split that is fine for me too) [10:36] mvo: no strong opinion, is not very big either way [10:36] * mvo nods [10:49] duh, dates suuuck [10:56] mborzecki: you're reminding me of this book that looked into systems that needed something like three different times referring to everything so as to answer questions such as "what price would the sales person have had on their laptop in the field at the time they gave the presentation?" [11:03] Bug #1628289 changed: snapd should depend on squashfuse (for use in containers)

[11:07] ok, pushed a fix to https://github.com/snapcore/snapd/pull/7443 think i got all the edge cases right [11:07] PR #7443: timeutil: fix schedules with ambiguous nth weekday spans [11:09] Bug #1548225 changed: Snapcraft/snappy should detect incompatible version mismatches [11:09] Bug #1636891 changed: Unable to log in using email with + character [11:12] Bug #1636242 changed: snap fails to install a package if it doesn't start a daemon successfully [11:13] Chipaca: do the current retries handle this better: https://bugs.launchpad.net/snappy/+bug/1643893 ? [11:15] Bug #1637153 changed: plugging in a network cable when wifi config failed causes a traceback when restarting console-conf [11:15] Bug #1644546 changed: [console-conf] Network screen continues to account setup with failing configuration [11:15] back from call, returning to branches [11:19] pedronis: dunno [11:27] mborzecki: could you update/answer here: https://bugs.launchpad.net/snappy/+bug/1839709 [11:27] Bug #1839709: does not work under cgroup v2 [11:27] pedronis: sure [11:27] thx [11:28] btw. on this note, people are starting to notice cgroups v2 in fedora since docker broke :P [11:28] mborzecki: I saw that [11:28] mborzecki: note there's a fallback plan still, to push it for another release [11:28] but I'm curious [11:29] since it seems like fedora might want to push to the docker replacement stack (podman and friends) instead [11:29] which would be perfectly in line with what they are trying to achieve (ship brand new software early) [11:30] yeah, still it kind of assumes that you have the time to play with it and 1st, debug why docker doesn't work anymore, 2nd actually switch to podman [11:33] mborzecki: I'm suprised docker is not a transitional package with a loads of wrapper scripts [11:33] zyga: well podman should be able to consume the same arguments as you normally pass to docker [11:34] and it does not require a daemon [11:34] mborzecki: right my point is that it feels like user-machine friction [11:34] "podman" != "docker" [11:34] so people get docker and bam [11:34] zyga: right, and docker inc. adds to that by having docker and moby :P [11:35] zyga: where moby may as well come from distro repos and behave a bit differently than docker from docker repos [11:36] part of FOSS complexity I guess [11:49] we should get podman into a snap [11:53] mvo: we got an answer to your request here: https://bugs.launchpad.net/snapd/+bug/1843077 [11:53] Bug #1843077: Data corruption when UC18 is not powered down cleanly [12:01] zyga: you got one +1 to 7484, splitting it now seems a bit a waste of time [12:02] pedronis: I'm happy NOT to if mvo agrees :) [12:02] I can do more useful stuff then [12:03] zyga: I don't think mvo had a strong opinion on this :) if you add the tests and get a +1 from mborzecki it can land I think [12:03] cool, I'll do that then [12:08] pedronis: zyga: not splitting is ok for me [12:09] mborzecki: +1 [12:10] mborzecki, hey, here there is a log when centos cannot start activate a service [12:10] and gives a timeout [12:10] https://paste.ubuntu.com/p/pNYYqPKJM7/ [12:11] mborzecki, then you see [12:11] + systemctl daemon-reload [12:11] Failed to execute operation: Connection timed out [12:11] in the log [12:12] cachio: yup, i filed a bug in centos with this last week, there's a fixed systemd build in the fasttrack repo, but we have to wait for it to end up in updates [12:12] mborzecki: https://github.com/snapcore/snapd/pull/7484/commits/9481c3cba1a8a4a62c76df78fb7cb7bece6ee89d [12:12] PR #7484: osutil: generalize SyncDir with FileState interface [12:13] mborzecki, nice [12:13] thanks for the update [12:13] cachio: https://bugs.centos.org/view.php?id=16441 if you want the details [12:14] zyga: yeah, thats fine. was mostly wondering, especially since you suggested it for 2.42 [12:14] zyga: so not splitting is fine [12:16] break for more tea [12:16] brb [12:34] PR snapd#7532 opened: Allow ovs interface access to bridge sockets [12:51] * Chipaca stops editing mborzecki's standup thing and goes to get something to nom [12:56] PR snapd#7533 opened: client: add support to use the new "download" API === ricab is now known as ricab|lunch [13:09] PR snapd#7534 opened: sandbox/seccomp: move the remaining sandbox bits to a corresponding sandbox package [13:09] zyga: ^^ [13:10] sure [13:12] mborzecki: done [13:12] mborzecki: can you review /writable PR please :) [13:12] zyga: hmm quite sure i posted a review [13:12] oh [13:13] silly github stale stuff [13:13] thanks! [13:14] mborzecki: replied [13:14] thanks [13:29] jdstrand: hey [13:29] jdstrand: are you back this week? [13:29] zyga: today and tomorrow, yes [13:29] (hey) [13:30] hey :) [13:30] I did an update over one branch and I'd like your re-review, just one today [13:30] https://github.com/snapcore/snapd/pull/7421 [13:30] it's also shorter and easier [13:30] PR #7421: cmd/snap-confine: unmount /writable from snap view [13:31] jdstrand: the whole complexity got reduced to one function with two if statements [13:31] https://github.com/snapcore/snapd/pull/7421/files#diff-af477950316a096b57d91c74478bc4d2R521 [13:32] zyga: ok [13:35] thanks, I'm working on more :) [13:45] * zyga -> walking the dog in the rain [13:47] Chipaca: so this one https://github.com/snapcore/snapd/pull/7166#discussion_r329554850 is actually quite interesting, bc the way i read it, timeout.C serves as the upper timit (time based though) for the retry mechanism (and that one is a workaround for snapd daemon restarts) [13:48] PR #7166: client: add doTimeout to http.Client{Timeout} [13:49] Chipaca: but since it doesn't affec the actual client.raw() call, it will have no effect when client.raw() gets stuck [14:12] hey folks, could I get reviews for #7437? ping mborzecki :-) [14:12] PR #7437: wrappers/services.go: add disabled svc list arg to AddSnapServices [14:15] PR snapd#7484 closed: osutil: generalize SyncDir with FileState interface [14:45] mborzecki: fair enough; that one wasn't the blocker :-) [15:00] * zyga EODs for now, will return to garden one branch later tonight after homework and rest [15:00] o/ [15:03] I'm going to slither off early as well, got my monthly physio check-up thing and then maybe a run if I'm up to it (doubt it because bleh) [15:04] PR snapd#7534 closed: sandbox/seccomp: move the remaining sandbox bits to a corresponding sandbox package [15:06] * cachio lunch [15:12] mvo: pushed commits, still to check the test errors [15:13] mvo: i'll be back after lunch [15:16] cmatsuoka: thank you [15:28] Using xdg-open within a snap: should I be using snap-xdg-open, or just xdg-open? [15:54] sil2100, fyi: https://github.com/CanonicalLtd/ubuntu-image/pull/175 [15:54] PR CanonicalLtd/ubuntu-image#175: Little kernel bootloader support === pstolowski is now known as pstolowski|afk [16:17] mvo got a minute to talk about uc18 spread test setup? I'm still having difficulties building an image with hacks in it by following parts of the spread prepare suite, and I'm not sure how much of the spread stuff I should be copying... [16:17] we could also talk tomorrow if you're EOD or about to EOD [16:22] abeato: see it, thanks! Will review, but it looks solid from a first look [16:22] ijohnson: pretty much eod, so unless its something specfic tomorrow is better. any particular thing that you are looking for? [16:23] well how to make it work haha [16:23] let's just talk tomorrow then [16:23] abeato: could we get the gadget.yaml spec updated to include the new bootloader? https://forum.snapcraft.io/t/the-gadget-snap/696 [16:27] sil2100, good point, will give that a go. or maybe ondra prefers to do that himself ^^ [16:27] abeato: anyway, thanks for the PR! I'll make sure it's reviewed and (hopefully) merged tomorrow [16:27] sil2100, awesome, thanks! [16:27] o/ [17:04] cachio: is there an easy way to increase the size of the disk space provided to a qemu machine when setup with spread? [17:04] I'm trying to run the ubuntu-core-18-64 tests (well a modified one) and it runs out of space before finishing preparing [17:07] hmm actually I think if I just increase the empty space in the image that spread uses it should automatically expand to use that I think [17:18] ijohnson, which image are you using? [17:18] I am trying to use the qemu backend [17:18] cachio: ^ [17:20] ijohnson, it is weird [17:21] ijohnson, did you change the prapare phose of the test suite? [17:21] no, not yet, just trying to see how uc18 in the tests is "bootstrapped" from classic 16.04 [17:22] ijohnson, so, you want to run uc18 on qemu right? [17:22] yes, but I want to eventually modify the image that boots in order to add stuff to the root filesystem [17:23] but for now, just getting one that boots with spread is sufficient for me [17:23] and how did you create that image? [17:23] ijohnson, the uc18 one [17:23] I followed HACKING.md :-) [17:23] https://github.com/snapcore/snapd/blob/master/HACKING.md#building-spread-vm-images [17:24] I created the ubuntu-16.04-64.img with `adt-buildvm-ubuntu-cloud` [17:24] ijohnson, ah, and then you want to transform it to core18, right? [17:25] yes [17:25] ijohnson, did you try to generate the image directly by using ubuntu-image? [17:25] I tried just modifying the official amd64 image by adding stuff to the rootfs but it didn't take, it seems to be ignored [17:25] and the using external system? [17:25] I have not tried that yet [17:26] s/official amd64 image/official amd64 UC18 image/ [17:27] so, you can generate the image by running [17:28] sudo ubuntu-image "$image_option" "$snaps" \ [17:28] -c "$CHANNEL" \ [17:28] -O "$output" \ [17:28] "./models/${platform}-${VERSION}.model" [17:28] with [17:28] image_option="--image-size 3G" [17:28] snaps= --snap [17:29] if you need to add any new snap [17:29] how would I do that for the qemu image? [17:29] then you run [17:29] sudo kvm -snapshot -smp 2 -m 2000 -net nic,model=virtio -net user,hostfwd=tcp::8022-:22 -nographic -serial mon:stdio [17:30] and you can connect to it [17:30] setting the [17:30] SPREAD_EXTERNAL_ADDRESS=localhost:8022 [17:30] and using the external backend4 [17:30] backend [17:30] the qemu backend in fact is doing something similar to that [17:31] the qemu backend is doing this: https://github.com/snapcore/spread/blob/master/spread/qemu.go#L117 [17:32] ijohnson, I didnt try it, beu you could save the image generated with ubuntu-image into "$HOME/.spread/qemu/" [17:32] and then try to use the qemu backend [17:33] yes I've been able to boot the image with qemu, I guess the problem I have is 2 problems [17:33] , first is that if I modify my UC18 image directly, the changes don't seem to work properly (I think this is because of the special UC18 boot process which is different from UC16 where my modifications do work) [17:33] ijohnson, but it is not gonna work because ubuntu core needs to run console-conf [17:33] and the second problem is that if I try to use spread to boot the default classic xenial image and let spread "prepare" the xenial image into a UC18, then preparing fails because the default image doesn't have enough space [17:34] cachio: I successfully disabled console-conf in the image, that much still works from my days hacking on UC16 images :-) [17:35] I need to break for lunch but will look at this again in a little bit, cachio do you know if the adt-build-vm-ubuntu-cloud created images are in raw format or a different format? [17:35] ijohnson, no idea [17:36] ok, np [17:36] ijohnson, but you should be able to resize the image [17:36] hmm actually there is a qemu resize image command isn't there [17:36] ijohnson, for google images we run resize2fs [17:37] to resize them when it is needed [17:37] ack, thanks I'll look into that after lunch [17:37] thanks cachio! [17:37] * ijohnson lunches [17:37] ijohnson, sure [17:43] PR snapd#7535 opened: tests: disable {contacts,calendar}-service tests on debian-sid [17:45] review for -^ would be great [17:45] simple and will unbreak master [17:48] PR snapcraft#2734 opened: SnapcraftError refactoring [17:52] mvo, do you know why it is failing in debian [17:52] same reason than arch? [17:52] cachio: looks like it [17:53] cachio: https://metadata.ftp-master.debian.org/changelogs//main/e/evolution-data-server/evolution-data-server_3.34.0-3_changelog [17:53] cachio: looks like the updated version hit unstable today [17:54] mvo, today we updated debian sid [17:59] mvo: reviewed [17:59] ta! [19:18] PR snapd#7536 opened: gadget: accept system-seed role and ubuntu-data label [19:57] PR snapd#7535 closed: tests: disable {contacts,calendar}-service tests on debian-sid [21:42] roadmr: hey, not urgent at all, can you pull 20190930-2141UTC. it is the override for spread for Chipaca [21:42] jdstrand: sure, anything for Chipaca 👍 [21:43] :) [21:43] I'll aim to deploy that tomorrow but no promises - it'll be in the queue anyway, worst case a couple of days [22:22] PR snapcraft#2735 opened: extensions: add gsettings plug to gnome-3-28 extension [22:24] PR snapd#7537 opened: tests: fix ubuntu-core-device-reg test for arm devices on core18