[09:20] PR snapd#7954 opened: changelog: fix typos [09:39] Hi, is there any newer docs to setup an snapstore, like this: https://ubuntu.com/blog/howto-host-your-own-snap-store ? [09:43] sdhd-sascha: not that I am aware of [09:44] Chipaca: How can i build my own ubuntu-core and add my own kernel and gadget. [09:44] In the store it needs a review and i have to wait [09:44] sdhd-sascha: I'd expect ubuntu-image to let you use a local (unasserted) snap [09:45] Chipaca: thank you :-) yesterday i already cloned ubuntu-image, because i'm looking for an error message :-) [09:46] sudo ubuntu-image -c beta -O rk3318-test rk3318.model [09:46] Warning: for backwards compatibility, `ubuntu-image` falls back to `ubuntu-image snap` if no subcommand is given [09:46] error: model with series "18" != "16" unsupported [09:46] COMMAND FAILED: snap prepare-image --channel=beta rk3318.model /tmp/tmpdh568a0l/unpack [09:46] sergiusens: a little Christmas present for you: https://forum.snapcraft.io/t/call-for-testing-github-action-for-snapcrtaft/14930 [09:47] hello jamesh, thanks! [09:47] jamesh: +1 [09:47] sdhd-sascha: series 18?? [09:47] sdhd-sascha: that's correct, series is fixed to 16, there is nothing like series 18 [09:47] even if you use core18 [09:47] as your base [09:47] ? [09:48] we thought we would have a series 16, 18, 20 etc [09:48] but in the end it's not what happened [09:48] https://www.irccloud.com/pastebin/RwdFllwF/ [09:48] it might change at some point in the future if something changes is an very incompatible [09:48] way [09:48] sdhd-sascha: different values for series imply incompatible platforms. You can mix snaps built for core and core18 on a single system [09:48] but hasn't so far [09:49] so series: 16 is fixed everywhere atm [09:49] i copied the linux raspi [09:49] sdhd-sascha: change series:18 to series:16 in that model and it'll be better [09:50] sdhd-sascha: there is no series 18 [09:50] sdhd-sascha: there never has been, and never will be [09:50] What is "series" for ? [09:50] sdhd-sascha: resetting the universe [09:50] :-D [09:50] sdhd-sascha: if we find a fundamental problem with the way we do snaps, and have to restart from scratch in an incompatible way [09:51] sdhd-sascha: the idea was that at each lts we could have completely disjoin sets of snaps, but we didn't proceed that way atm. it's just a format version number at this point [09:51] whose change would represent a non-backward compatible change [09:51] to a lots of things [09:51] i understand - thank you :-) [09:53] Is there a way, to store my passphrase with "snap sign" ? Currently i only can sign if i call "register-key" again [10:33] sdhd-sascha: I'm not sure I understood your question [10:35] Chipaca: "snap sign" says: `error: cannot sign assertion: cannot sign using GPG: /usr/bin/gpg --personal-digest-preferences SHA512 --default-key 0x.... --detach-sign failed: exit status 2 ("gpg: signing failed: No such file or directory\ngpg: signing failed: No such file or directory\n") ` [10:35] you dont have the key ;) [10:35] Then i call "snap register-key" again. Then it works for a moment. [10:36] "register-key" ask me for my passphrase of the key [10:37] sdhd-sascha: there is no register-key [10:37] thats a snapcraft option... [10:38] Oh, sorry "snapcraft register-key" from https://docs.ubuntu.com/core/en/guides/build-device/image-building#image-building [10:40] In this documentation on step 2 is "snap sign", which only works after "snapcraft register-key" [10:41] Did i need some "gpg-agent" or something? [10:43] does "snapcraft list-keys" list your key ? [10:43] ogra: yes, "default" key [10:44] and you are calling register-key with "default" as option ? [10:45] https://www.irccloud.com/pastebin/a7OgLbOb/ [10:46] And going to 2FA site https://login.ubuntu.com/device-list [10:46] i get a 404 page [10:47] well, and you get a 500 error in that paste [10:47] yes [10:48] the first time it works. [10:48] but i need to call "register-key", that "snap sign" works for a moment [10:49] https://www.irccloud.com/pastebin/TOfRf5RN/ [10:49] and snapcraft:3.8 [11:02] hi, I have a (go-based) snap which has been failing for a while on BSI, but I can't reproduce the issue locally. https://build.snapcraft.io/user/albertodonato/h2static/785325 is an example of a failing build [11:02] I tried both with and without go-importpath, they both work locally but both fail over there [11:06] ackk: is this a local path in the log? `go get -t -d ./github.com/albertodonato/h2static/...` [11:06] sdhd-sascha, yeah that's based on go-importpath [11:06] sdhd-sascha, https://github.com/albertodonato/h2static/blob/master/snap/snapcraft.yaml is the config [11:07] ackk: i try it here on my machine and will report, what happens here [11:07] sdhd-sascha, I assume something has changes either with snapcraft or with the building env on launchpad, as I also had to add gcc as build-packages, which wasn't needed before [11:08] sdhd-sascha, thanks [11:10] ackk: it works here, too [11:11] but i build it with multipass + vm [11:13] I will test it now with: `snapcraft build --use-lxd --debug` [11:13] sdhd-sascha, yeah I did --use-lxd, worked for me [11:18] ackk: you are right about the env or something. --use-lxd runs without problem here [11:18] I wonder if it's running the right "go" binary on LP [11:20] actually, it doesn't seem it's running "go mod download", which looking at the plugin source code shoud do when go.mod is presente [11:20] *present [11:22] if i call snapctl from within an app, it does properly set the value for a key but doesnt run the configure hook at all ... is that a known issue ? [11:24] ogra I think I've encountered that issue before, ISTR it was intentional (probably to avoid infinite loops?) [11:26] sdhd-sascha, oddly, I get that error if I build with --destructive-mode [11:27] ackk, well, sadly thats not helpful if you need to force-restart an app to make it pick up the new configuration ... :) [11:28] (even worse if the configure hook actually re-writes a config file ... then the app config and snap values get completely out of sync) [11:29] pedronis, ^^^ is that true (not running hooks for snapctl calls when they come from inside an app) ? [11:29] ogra, yeah I agree it should call the hook, maybe it could either detect when it's being run inside a hook and not call it just in that case, or have a flag to do so [11:29] ogra: yes, that's by design, there has been discussions to support snapctl set --configure to get the other behavior [11:29] ah, thanks ... [11:29] very confusing behaviour :) [11:30] sdhd-sascha, oh, it seems the stable snapcraft doesn't handle go.mod [11:31] ackk: oh [11:31] i can't use snapcraft:3.9 because of https://github.com/snapcore/snapcraft/pull/2840 [11:31] PR snapcraft#2840: plugs: plugs can have no element [11:33] i use snapcraft:3.8 [11:33] stable [11:36] hey mvo [11:37] are we supposed to work today or did I get my dates messed up? [11:37] ackk: i need to ask again. I didn't understand the error cause. What's wrong and in which case? [11:38] zyga: isn't there a public holiday in .pl today? [11:39] yeah, I think so [11:39] ah, but not in .de? [11:39] sdhd-sascha, I get the same error as the build on LP if I run "snapcraft --destructive-mode" [11:41] zyga: not in .de [11:41] ah, that explains everything, I was worried seeing the release that I should be in the office for hours :) [11:41] zyga: well, my understanding is that you guys have a public holiday but you should know better than me :) [11:44] mvo: https://mobile.twitter.com/qikipedia/status/1213731861660413952 :) [11:47] pedronis: reviewed #7935, i think i'll hold off going on down the stack for a bit [11:47] PR #7935: boot,overlord: introduce internal abstraction bootState and use it for InUse/GetCurrentBoot [11:47] Chipaca: haha - good one [11:49] sdhd-sascha, ah, I think I got it. it seems that the stable snapcraft doesn't handle go modules, but since it's using go1.13 by default now, go tried to pull them. if I remove go.mod it builds. [11:49] ah, ok [11:50] so it probably broke for me when go1.13 snap was pushed to stable [12:00] is there a way to extend the environment for pull/build steps? it doesn't seem that exporting a var in override-* works [12:03] ackk: do you need to compute the value? [12:03] otherwise the environment block should work [12:03] Chipaca, no, it's afixed one. basically a workaround for the issue above ^ would be to set GO111MODULE=off for pull [12:04] Chipaca, can I use "environment" in a part? [12:04] ackk: build-environment IIRC [12:04] ah, TIL [12:06] Chipaca, out of curiosity, why is that a list of dicts rather than a dict like the app environment? [12:07] ackk: hmm? [12:08] ackk: I don't know [12:08] ackk: question for sergiusens maybe [12:14] disabling modules via build-environment worked, thanks sdhd-sascha and Chipaca [12:34] Chipaca: thx for the review [12:35] pedronis: i change-requested it because of the error, if i goofed just holler :) [12:48] Chipaca: no, I think you are right, but need to do some other things before going back to those Prs [13:02] mvo: Chipaca: what should we do with this bug, it's not clear to me which bits are snapd and which are lubuntu? https://bugs.launchpad.net/snapd/+bug/1858303 [13:02] Bug #1858303: Snaps don't work unless pre-installed or user logs out and logs back in [13:14] pedronis: hmhm, interessting, looking [13:20] PR snapd#7955 opened: tests: remove "test-snapd-tools" in smoke/sandbox on restore [13:22] Bug #1555140 changed: unit test apiSuite.TestGetOpInfoIntegration fails when building the deb in jenkins [13:23] PR snapd#7956 opened: packaging: ship var/lib/snapd/desktop/applications in the pkg [13:23] pedronis: that was figured out in the forum [13:23] mvo: ^ [13:23] ah, mvo spotted that already [13:23] anyway it's https://forum.snapcraft.io/t/snapd-required-dependencies-to-start/14904 [13:25] Bug #1611068 changed: 401 when trying to install any snap due to invalid credentials in ~/.snap/auth.json

[13:25] Chipaca: yeah, 7956 should address it [13:25] mvo: ah, nice [13:28] Bug #1665756 changed: environment variable setting issue === ricab is now known as ricab|lunch [13:57] PR snapd#7948 closed: spread: drop copr repo with F30 build dependencies === ijohnson|EOY is now known as ijohnson [14:23] degville: turing tumble is aweseme (only ever seen it in video -- i assume it's as good as it looks?) [14:23] Chipaca: yes! it's really good! :) [14:24] degville: woah, that looks fun! [14:24] it's brilliant actually seeing and trying to physically work out how to solve the problems. [14:50] ogra: i think i found the problem with "snap sign". The passphrase wasn't shown because of the stdin-pipe. [14:51] ah, ouch [14:51] Now i want to refactor the "snap sign" [14:52] But can i break backward-compatibility or not [14:52] (i always use a dedictaed "build" or "image" key for my models that goes witjout passphrase) [14:53] (snapcraft allows to give them names at creation time) [14:53] ogra: normally, i too ... but this time ... [14:54] i'm looking for a "snap delete-key" command, yesterday. Or how can registered keys be deleted ? [14:54] i dont think they can easily [14:54] Ok, no delete [14:55] (ask in the forum, i might be wrong ... i know they couldnt a year ago) [14:55] Then how about to make "snap sign -k default " === ricab|lunch is now known as ricab [15:07] * cachio lunch [15:08] PR snapcraft#2844 closed: Add punctuation rule for comments [15:16] * Chipaca wonders at what time he should stop with the coffee [15:16] grr the snapcraft forum seems even slower this year [15:20] ijohnson, just append "?slow=false" to the url ! [15:21] ... and happy new year :) [15:21] can I just do ?speed=plaid instead ? :-) [15:21] happy new year ogra! [15:23] PR snapcraft#2837 closed: remote-build: gpg-signing and usability fixes === pedronis_ is now known as pedronis [15:36] Chipaca: I updated #7935 [15:36] pedronis: tks [15:36] PR #7935: boot,overlord: introduce internal abstraction bootState and use it for InUse/GetCurrentBoot [15:37] pedronis: I presume the dropping of the NoState error is because it's umpossible :) [15:37] unpossible* [15:37] Chipaca: yes, it should not happen, if we have kernel, we should have a model [15:37] and we check whether we have a kernel just before [15:38] yep [15:38] +1, thank you [15:47] PR snapd#7957 opened: snap-bootstrap: mount the correct snapd snap to /run/mnt/snapd [15:56] PR core20#18 opened: static: try using /run/mnt/snapd first in run-snapd-from-snap [16:26] mvo, hey, having troubles to login into the uc20 instnace as root [16:26] using your branch [16:27] did you change anything today? [16:28] Chipaca: I reviewed your channel PRs [16:28] thank you [16:33] cachio: I don't think I did - stange [16:33] pedronis: thank you! [16:34] pedronis: you're right about InstallPath, i'll fix that in a bit [16:34] need to pop some state first :) [16:36] also, need to reboot for a new kernel, and go to the shops and stuff. So AFK for a bit. [16:39] mvo, np, I'll continue researching [16:40] cachio: thank you [16:40] cachio: s/stange/strange/ :) [16:40] cachio: there were some core20/gadget changes over the break, maybe something there broke it, but AFAICT the spread test itself is still working on the uc20 tests [16:45] mvo, with your branch didnt work today, the system reboots and it seems to be ok but I cant connect using root [16:46] I was using a branch with my chnges and it was manually adding a access for the users [16:46] and restarting ssh [16:47] I just restarted the travis job to cehck it is wotking after the break [16:50] cachio: oh, ok [16:51] cachio: interessting, so something did change :/ please keep me update! [16:51] updated [16:56] mvo, sure [17:34] PR snapd#7958 opened: snap-bootstrap: refactor partition creation [17:38] PR snapcraft#2849 closed: rust plugin: split RUSTUP_HOME and CARGO_HOME [17:56] PR snapcraft#2833 closed: catkin: remove rospack workaround [18:20] Hi there. Upon installing chromium-browser - which in my current version of Ubuntu (Kubuntu 19.10) subsequently installs the snapd version - the cursor theme is different to the one I currently use (Breeze theme) whilst it hovers over Chromium's contents. I have gtk-common-themes installed via snapd. Is there a way to force the snap to use the appropriate cursor theme? Thanks. [18:41] J_C: you might try asking in #ubuntu-desktop or on the forums, I think most folks here who would be able to help you with that are offline due to TZ's === heather is now known as hellsworth [19:11] ijohnson: Heh, I was actually suggested to ask in here from #ubuntu. No worries. [19:11] J_C: ah ok, well anybody else in #snappy is free to speak up but I myself don't know :-) [19:12] ijohnson: It's nothing too urgent anyway, just a small vidual quirk and nothing that impacts functionality at the very least. :) [20:23] PR snapd#7956 closed: packaging: ship var/lib/snapd/desktop/applications in the pkg [20:49] J_C: that will happen for any themes not included in the gtk-common-themes snap [20:52] Chipaca: I was under the impression that both Breeze and Breeze Dark were in gtk-common-themes. [20:53] J_C: cursor themes are separate though [20:53] and I don't see anything under Breeze/cursors [20:54] That's a shame. At least I know why the cursor isn't showing properly now. Thanks. [20:55] PR snapd#7959 opened: tests: fix classic-ubuntu-core-transition-two-cores after refactor of MATCH -v [20:55] J_C: what this says to me is that the cursor you're seeing is not of the theme but of some kde default [20:55] J_C: if you figure out where it's from, maybe see about getting it included? [20:56] if it is the default kde one, we should probably include it :) [20:58] kenvandine: does that ^ sound sane to you? (about where the cursor is coming from) [20:59] hello [20:59] Chipaca: Sure, I'll have a look into it :) Should I submit a ticket to the gtk-common-themes repo or elsewhere? Would you like a screenshot as to what the current cursor looks like? [20:59] is there a way to give local filesystem permission to snap package? [21:00] thiras: not in general no. What are you trying to do in particular? [21:00] specifically dbeaver package. it needs to access native mysql client installed on my system through apt [21:00] J_C: i wouldn't know what to do with the screenshot other than agree with you, and i already believe you :) [21:01] thiras: that sounds wrong [21:01] you cannot use it's internal drivers to dump a database. it needs native mysql client to do that [21:01] https://github.com/dbeaver/dbeaver/issues/1361 [21:02] obviously package maintainer seems be missed that detail. so the current package is totally isolated [21:02] to be* [21:03] thiras: there isn't a way of doing that, no [21:03] too bad [21:04] should i open a bug report to package maintainer? [21:04] i guess some of the snap packages can have access to filesystem (like vscode) [21:04] classic snaps can, yes [21:05] but i don't think it's a bug, this seems more like working as intended [21:05] thiras: what is the snap name by the way? [21:05] Chipaca, dbeaver-ce [21:06] maintainer is the dbeaver company itself [21:06] then how modern snaps handles if they need to access to the filesystem? [21:07] thiras: 'access to the filesystem' is one thing, what you're wanting is being able to run arbitrary things [21:07] thiras: the short answer is that you don't, you can't, that's not how it's supposed to work [21:08] anyway, EOD for me [21:08] so i assume that snap package should contain native client within? [21:08] 👋 [21:08] thanks for help [21:08] thiras: exactly, yes [21:09] bye! [21:29] PR snapd#7960 opened: tests: use unbuffered python output for daemons, misc formatting [22:20] Where is the place to ask question about "ubuntu-image" ? `error: cannot use kernel "linux-generic-allwinner" published by "QfOqF7d2M1Pk2O0SbEKqTdB9Ry2aI0BP" for model by "go2ddtVdG0SlCsLDKX7cd9IWiZoEbSym"` [22:21] Oh, sorry. This is a "snap sign" error [22:24] (revert last line) [22:32] PR snapd#7961 opened: cmd: sign: add filename param [23:06] PR snapcraft#2852 opened: wstool: don't rely on host git