=== jamesh_ is now known as jamesh [07:02] morning [07:02] zyga: hey [07:02] hey :) [07:02] https://listed.zygoon.pl/20686/raspberry-pi-4b-and-4k-display-at-60hz battle from last evening [07:17] zyga: btw. this is interesting: https://gist.github.com/dmp1ce/8a321608fbcf4c3ea61fce134e1f0b0d there's apparmor enabled, but /proc/self/attr/apparmor/current does not exist [07:17] mborzecki how about /proc/self/attr/current [07:17] the apparmor sub-directory is relatively recent addition [07:18] the user has that, maybe it's the lts kernel they are booting, suggested using the non lts one to see [07:18] fwiw, this code is supposed to guess the right location: https://gitlab.com/apparmor/apparmor/-/blob/master/libraries/libapparmor/src/kernel.c#L235 [07:19] zyga: it eventually gets called from aa_getcon() [07:19] right [07:20] what arch? do you have uname? [07:20] maybe wonky kernel from some cloud [07:20] or arm gizmo kernel [07:20] zyga: they said it's an lts server, so probably 5.4.82-1 [07:21] hm maybe i should ask them to build and run a snippet like this: https://gist.github.com/dmp1ce/8a321608fbcf4c3ea61fce134e1f0b0d [07:21] hmm [07:26] * zyga changes hosts [07:38] PR snapcraft#3391 opened: plugins v2: port the qmake plugin [07:44] mborzecki, pi4 desktop is quite nice [07:44] such a long way from "it barely works" to "4K 60Hz" [08:03] morning [08:03] good morning pawel [08:03] I managed to sort out my alt-desktop for now :) [08:22] hello mvo [08:22] good morning zyga! how are you doing? [08:22] hey :-) [08:22] I'm working on a pi desktop today [08:23] feeling good, trying to build a pipeline for debian packaging now [08:26] hey mvo, zyga! [08:30] zyga: nice! [08:30] pstolowski: good morning [08:36] pstolowski: mvo: hey [08:37] mvo: we should land it when the spread jobs finishes: https://github.com/snapcore/snapd/pull/9785 [08:37] PR #9785: tests/main/fake-netplan-apply: disable test on xenial for now <⚠ Critical> [08:51] mborzecki: ineed [08:52] mborzecki: we need this for 2.48 too, right? [08:53] a review for 9758 would be really good [08:57] mvo: yes, i think so, unless we can get #9784 (or both?) [08:57] PR #9784: interfaces/builtin/network-setup-{control,observe}: allow using netplan directly [08:59] I don't understand the last change in #9784 fwiw [08:59] PR #9784: interfaces/builtin/network-setup-{control,observe}: allow using netplan directly [08:59] also it really needs security reviews [09:05] PR snapd#9785 closed: tests/main/fake-netplan-apply: disable test on xenial for now

<⚠ Critical> [09:10] PR snapd#9786 opened: tests: backport netplan workarounds from #9785 (2.48) [09:31] mvo: I re-reviewed 9783, it's now failing on classic/hotplug :/ [09:35] Sergio mentioned it, i'll take a look at this test [09:37] pedronis: thanks for this! [09:56] mvo: we need this now? https://github.com/snapcore/snapd/pull/9783 [09:56] PR #9783: gadget: use "sealed-keys" to determine what method to use for reseal

[09:58] mborzecki: yeah, this pr will unblock master [10:03] mborzecki: I reviewed 9781, it conflicts with some of my incoming PRs but not too badly fwiw [10:04] yes, we should really try to land 9783 as quickly as possible, it also unblocks my PRs [10:07] mvo: in theory we could just force land it, and then have a follow with the small changes suggested by Ian and me? [10:08] mvo: is it blocked by something else? [10:08] pedronis: that sounds good, no [10:08] pedronis: not blocked [10:08] pedronis: I can squash merge it now [10:09] pedronis: thanks, we can land it before of after, wahtever is easier to do [10:09] whichever [10:13] mborzecki: I tried after [10:13] fwiw [10:13] pedronis: merged and cherry-picked, I will do the tweaks in a small followup [10:13] mvo: thanks [10:13] I will merge master in my gadget PR [10:14] +1 [10:14] I have follow ups but that one needs to land first [10:15] PR snapd#9783 closed: gadget: use "sealed-keys" to determine what method to use for reseal

[10:20] yay [10:22] pedronis: so let's merge #9775, then #9781 [10:22] PR #9775: gadget,o/devicestate,tests: drop EffectiveFilesystemLabel and instead set the implicit labels when loading the yaml [10:22] PR #9781: gadget/quantity: introduce Offset, start using it for offset related fields in the gadget [10:22] in the meantime i'll look at updating #9771 [10:22] PR #9771: boot: boot config update & reseal [10:25] mborzecki: should I do this? [10:28] mborzecki: I will force-merge 9781 [10:30] PR snapd#9781 closed: gadget/quantity: introduce Offset, start using it for offset related fields in the gadget [10:40] mvo: that's not the order mborzecki suggested though [10:40] PR snapd#9758 closed: secboot: add new LockSealedKeys() that uses either TPM/fde-reveal-key [10:40] PR snapd#9787 opened: boot: tweak resealing with fde-setup hooks [10:41] mvo: now 9775 has conflicts :/ [10:41] so it will need a new test run [10:42] :/ [10:44] pedronis: oh no! I'm sorry! [10:45] sorry, should not have interefreed [11:00] pedronis: i can help you and fix conflicts in 9775 [11:00] mborzecki: I'm fixing them but we need anyway to wait the current test run [11:00] right [11:01] if we hope to avoid two nested runs [11:01] heh is it fair to duplicate the questions from #snapcraft here ? [11:01] and don't get banned ? [11:06] paul424: you won't get banned, but forum.snapcraft.io is the best place for questions with most visibility and chances for answers [11:07] Heh witam jakie miasto ? :D [11:08] mogę dobrze napisać po polsku w czym rzecz :D [11:08] oki no more polish [11:09] paul424: sorry, please use our forums [11:35] the nested hotplug test failure seems to be caused by a change in qemu, the usb serial port doesn't show up on the bus [11:48] oSoMoN, do we not update the chromium snap installer version alongside a new stable release ? (chromium-browser is at 85...) [11:49] if we dont, we should probably not try to match the chromium version string in the deb package [11:49] (there is a user asking about it in #ubuntu) [11:49] ogra, we don't, unless there are other changes to the packaging [11:50] I agree it's misleading, but it's also harmless [11:50] right ... so it should probably just use a native debian version instead of suggesting to the users they get an old version [11:50] because the user will always get the latest stable update through the snap [11:50] indeed [11:51] ogra, do you have a suggestion for a version number that would not be misleading? [11:52] 1 [11:52] 🙂 [11:52] or 1-ubuntu1 if you want [11:53] just not the chromium version string [11:58] ogra, that's pretty much what we did for the maas package in focal, which installs/transitions to the snap [11:58] yeah [11:58] (plus epoch) [12:12] we already have an epoch, so it would need to be bumped to ensure a greater version number than the current 1:85.0.4183.83-0ubuntu2 [12:12] "2:1-0ubuntu1" [12:13] yeah, something like that I guess [12:17] ackk, so for maas, you changed the version number to 0.x and added an epoch in focal, and removed the package altogether in groovy onwards? [12:31] oSoMoN, correct [12:47] * cachio afk [13:09] mvo: https://github.com/snapcore/snapd/pull/9786/ mostly passed, there's a failure in tests/core20/basic but unrelated to the PR [13:09] PR #9786: tests: backport netplan workarounds from #9785 (2.48) [13:10] the failure is intersting though, become-operational does not complete [13:16] PR snapd#9775 closed: gadget,o/devicestate,tests: drop EffectiveFilesystemLabel and instead set the implicit labels when loading the yaml [13:17] mborzecki: I'm getting internal errors from shellcheck :/ [13:17] pedronis: got log? which test? [13:17] pedronis: or do you mean actual shellcheck is failing? [13:18] I'm getting this [13:18] running unit tests in wrappers [13:19] mborzecki: https://paste.ubuntu.com/p/QXBgGQxf6j/ [13:21] mborzecki: it's https://github.com/snapcore/snapd/pull/9779 perhaps ? [13:21] PR #9779: testutil: make DBusTest use a custom bus configuration file [13:22] I got it only now after merging master [13:23] ah, mess [13:23] hmm [13:23] I have shellcheck installed as a snap [13:24] but now we can't run snaps from inside our tests I suppose [13:24] it's snap run trying to talk to systemd and failing [13:25] heh, as in real system apparently [13:26] so why dont i see it too? [13:26] do you have shellcheck as a snap? [13:27] I'm not even sure why that code gets involved [13:27] hm i have, maybe it isn't used [13:28] pedronis: dbustest.go sets DBUS_SESSION_BUS_ADDRESS [13:29] in env [13:29] that's probably why snap run connects to the session bus (although is the one for tests) and attempts to chat with systemd user session manager [13:32] so, yes, if I got back before that was merged things don't explode [13:32] for me [13:35] bit of a mess [13:37] pedronis: https://paste.ubuntu.com/p/33RHztGMJd/ ? [13:38] i can propose taht [13:38] I suppose, but there should be something easier [13:38] no, let me try something else [13:38] first [13:38] with the diff the tests work for me if they involve shellcheck from a snap [13:39] (actually had to remove the shellcheck package, somehow it was picked first in the go workspace i have) [13:39] mborzecki: I have a shorter change [13:40] oh, nice, can you propose it? [13:40] yes [13:43] mborzecki: https://github.com/snapcore/snapd/pull/9788 [13:43] PR #9788: testutil: use the original environment when calling shellcheck [13:46] PR snapd#9788 opened: testutil: use the original environment when calling shellcheck [13:46] PR snapd#9789 opened: many: separate consistency/content validation into gadget.Validate|Content [13:51] PR snapd#9790 opened: gadget: move BuildPartitionList to install and make it unexported [13:54] pedronis: thanks, lgtm and works locally here too [14:44] pedronis: mvo: the cloud-init pr is ready to merge, I reviewed all the spread failures they are all unrelated or already fixed in master, see https://github.com/snapcore/snapd/pull/9770#issuecomment-743231366 [14:44] PR #9770: sysconfig/cloudinit.go: add "manual_cache_clean: true" to cloud-init restrict file [14:45] after a day of work on aarch64 desktop the only thing that I find annoying is ports.ubuntu.com vs archive.ubuntu.com, it's surprisingly smooth otherwise [14:59] hey zyga [14:59] hey :) [14:59] I noticed you mentioned in your blog post the whole switching between bootloaders on your rpi [14:59] did you try dave's new tool for doing just that ? [14:59] ijohnson, no [14:59] I think it's like pibootctl or something [14:59] I just dsit-upgraded, I wasn't aware [14:59] ah [14:59] I know about pibootctl [15:00] I think here the problem was that 20.04 desktop and 20.10 desktop have totally different boot setup [15:00] it's pretty snazzy I used it for my pi that I upgraded from focal to groovy and eliminated uboot that way [15:00] and regular package updates just don't handle the transition [15:00] zyga: see https://waldorf.waveform.org.uk/2020/introducing-pibootctl.html [15:00] right, he shared that with me [15:00] you can even backup/restore changes to config.txt and such [15:00] my point was for simple guide, to avoid debugging, just reinstall [15:01] anyways I thought it was a cool tool that worked really well when I used it [15:01] but yes agreed reinstall is much simpler [15:01] if you copy the right dbts and config yeah, it will wrok [15:01] *work [15:01] are you using your groovy pi? [15:01] yeah [15:01] well I use it headlessly [15:02] PR snapd#9791 opened: devicestate: implement checkFDEFeatures() [15:03] I don't use desktop just server [15:03] ijohnson, I was mainly after a working desktop with 60Hz in 4K [15:03] now it works but I had my doubts along the way :) [15:04] ijohnson: merged and cherry-picked [15:04] thanks [15:07] PR snapd#9770 closed: sysconfig/cloudinit.go: add "manual_cache_clean: true" to cloud-init restrict file [15:25] need to wrap it up, enjoy your holidays [15:49] reviews for 9769 and 9791 would be amazing, it's the last fde-hook bits [15:52] * cachio lunch [15:58] ok, i think i know what's wrong with hotplug tests... it's shockingly simple [15:58] and embarassing [16:04] pstolowski, what is ti? [16:05] zyga: apt install -y "linux-modules-extra-$(uname -r)"' || true [16:05] || true :D [16:06] lol [16:06] it seems to be linux-modules-extra-...-generic now [16:06] kernel vs headers mismatch? [16:06] ohhh [16:06] man [16:06] nice :) [16:06] and a multitude of other variants [16:06] I'll gladly review that :) [16:06] the above is still fine on old releases [16:07] needs to support both i think, testing now... nested is terribly sloooow [16:24] wow look at that a green pr [16:24] a thing of beauty that is [16:24] nice work pstolowski ! [16:27] PR snapd#9788 closed: testutil: use the original environment when calling shellcheck [16:46] hmm there is more to it but i'm confused [16:49] apart from package name change, it doesn't provide ftdi_sio and usbserial modules for 5.8.0-31-generic kernel; in fact no other package has them [16:49] that's 20.10 [16:50] pstolowski, I see this on my pi [16:50] linux-modules-5.8.0-1006-raspi: /lib/modules/5.8.0-1006-raspi/kernel/drivers/usb/serial/usbserial.ko [16:51] zyga: i'm talking about -31- specifically. it see it for older versions too [16:53] zyga: i mean this: https://pastebin.ubuntu.com/p/WnmkdMmZ7y/ [16:54] yeah, that's weird for sure [16:54] no idea [16:58] hmm no fix for nested today i guess :/ [17:06] cachio: ping [17:20] cachio: the reason nested/classic/hostplug is failing is we are missing ftdi_sio and usbserial modules; we try to install modules-extra in the test (and ignore errors if it's missing), but this package was renamed with -31- kernel. but oddly, these modules are no longer provided by any package for that kernel. i'm stopping investigating this for today; if you have some time or any ideas then feel free to look at it; if not then [17:20] i'll continue on Monday. in any case, happy holidays! [17:20] * pstolowski eow === ijohnson is now known as ijohnson|lunch [18:23] PR snapd#9792 opened: tests: enable ubuntu 21.04 for spread tests === ijohnson|lunch is now known as ijohnson [22:29] PR snapd#9737 closed: tests: use os.query tool instead of comparing the system var [22:29] PR snapd#9793 opened: tests: add os query commands for subsystems and architectures === King_InuYasha is now known as Conan_Kudo === Conan_Kudo is now known as King_InuYasha