[00:02] PR snapcraft#2291 closed: meta: put environment into runner instead of app wrapper === gurmble is now known as grumble === cpaelzer_ is now known as cpaelzer [05:12] morning [05:50] o/ [05:50] I’m feeling somewhat better. [05:51] Hello [05:55] Hey [05:56] The store is fully operational again, great! [06:01] zyga: hey [06:21] PR snapd#5805 closed: cmd/snap-update-ns: enforce trespassing checks [06:21] * zyga is happy to land that [06:25] PR snapd#5864 closed: make sure hostnamectl can be used on core too [06:38] PR snapd#5857 closed: userd: extend the list of supported XDG Desktop properties when autostarting user applications [06:39] PR snapd#5840 closed: interfaces/apparmor, interfaces/builtin: tweaks for parallel snap installs [06:51] 5571 should be an easy win (just needs a second review) [06:54] mvo: 5571 appears to be merged, did you mean another pr? [06:54] 5771 maybe? [06:54] mborzecki: yes, sorry [06:54] clearly more tea is needed :) [07:04] Morning [07:06] PR snapd#5771 closed: selftest: add test to ensure selftest.checks is up-to-date [07:06] mvo: trespassing fix is merged, shall I propose a PR to take layouts out of beta? [07:07] mvo: full disclaimer, I'm investigating a bug that is related to propagation settings [07:07] so perhaps not yet but then again, it would help a lot of people with simple layouts already [07:08] zyga: let do it [07:08] zyga: and we can wait with the merge until you finished your investigation [07:08] ok [07:08] zyga: but this way spread can run etc [07:20] nice to see the diff in #5596 so short, nearly all i had in the integration branch is now in [07:21] PR #5596: [WIP] Parallel installs integration [07:22] mborzecki: hm, this looks like the integration PR juts contains some more tests, so time to merge those too? [07:23] mvo: i'll push those separately along with couple enhancements to other tests [07:24] mborzecki: cool [07:24] but first coffee [07:35] PR snapcraft#2300 opened: [WIP] plugins: remove implicit source [07:41] PR snapd#5866 opened: tests: remove unneeded cleanup from layout tests [07:43] PR snapd#5867 opened: many: enable layouts by default [07:46] mvo: ^ done [07:51] zyga: ta [08:02] zyga: hm, curious, why switching the default instead of juts removing the option entirely? [08:02] zyga: I mean, whats the use-case, someone who feels uncomfortable with layout should still be able to disable them? [08:02] If I make my snap private, will existing users continue getting updates? [08:02] And can people still install it by snap name? [08:03] (I want a semi-private snap so people download the "official" snap instead) [08:03] zyga, mvo do you guys have any thoughts on this? https://forum.snapcraft.io/t/snaps-are-broken-in-bionic-lxd-container/7339 [08:03] PR snapd#5868 opened: tests/main/snap-env: extend to cover parallel installations of snaps [08:07] kyrofa: can you please try the switch core to "edge"? [08:08] kyrofa: we fixed this recently but its not in beta yet, hopefully by EOD though [08:09] mvo, will do, thanks [08:13] re [08:14] kyrofa: replied [08:15] hi, I'm getting the following error running snapcraft: ImportError: /snap/snapcraft/2022/lib/python3.5/site-packages/nacl/_sodium.abi3.so: undefined symbol: crypto_pwhash_argon2id_opslimit_moderate [08:16] kyrofa: or how your new neighbours might have said if they spoke broken Spanish like I do: yo ha añadido una respuesta [08:17] akk interesting, we've started getting that as well [08:17] But it doesn't seem consistent-- you? [08:17] Chipaca: heey [08:17] around? [08:18] zyga: (y) [08:18] Urgh... why can't I cancel spread... [08:18] Chipaca: I'd like to add a tweak to yams loader for snap.yaml to indicate an error if "layouts" (plural) is used over the singular correct form [08:18] Chipaca: I was thinking of a simple new entry to deserialise with the different name [08:19] Chipaca: but if you have a better idea I'm all ears [08:19] kyrofa: you can, it just takes a bit [08:19] kyrofa: it delays because it shuts down the remotes, which can take a while to respond [08:20] kyrofa: you can also ctrl-backslash, and then use the .spread* to clean up [08:21] zyga: I wonder if there's a way to throw an error on anything unparsed [08:21] Chipaca: mmm, that may be future-proof nasty [08:21] I'd rather not [08:21] Chipaca, ah, ctrl-backslash will do it! [08:23] kyrofa: usually you can then tell spread itself to clean up using e.g. spread -reuse-pid= -discard [08:23] kyrofa: spread itself will also tell you what the right argument for -reuse-pid is in its log, if it's still in your scrollback [08:23] zyga: good point [08:24] Chipaca, just --reuse --discard seems to work [08:25] Chipaca: but we could _warn_ about it [08:25] (wink wink) [08:25] zyga: if only we had something like that [08:25] right [08:25] * Chipaca gets back to writing stuff on the forum [08:26] Chipaca: https://github.com/snapcore/snapd/pull/5869 [08:26] PR #5869: snap: detect layouts vs layout in snap.yaml [08:27] PR snapd#5869 opened: snap: detect layouts vs layout in snap.yaml [08:27] zyga: what about: TypeLouts errorCatcher [08:27] zyga: and then, type errorCatcher struct{} [08:27] errorCatcher? [08:27] ah [08:27] zyga: and then, func (errorCatcher) UnmarshalYAML(...) [08:27] and fail on deserialise? [08:27] :D [08:27] yeah [08:27] ++ [08:27] I'll do that, great idea [08:28] but... [08:28] how to tweak the error message then? [08:28] hmmmm [08:28] no < > to specialise [08:28] still [08:28] I could use a generic type [08:28] zyga: make it a layoutsErrorCatcher for now, we can worry about making it generic later [08:29] and subtype it [08:29] yes [08:29] +1 [08:32] Chipaca: actually, we can use the struct tags for that [08:32] zyga: we can? [08:32] zyga: i thought that'd be visible one step up [08:33] Chipaca: TypoLayouts errorCatcher `yaml:"layouts,omitempty",hint:"yada yada"` [08:33] something like that? [08:33] zyga: does the errorCatcher deserialiser "see" that hint? [08:33] not sure yte [08:33] looking at this part now [08:34] #5833 needs a 2nd review [08:34] PR #5833: asserts,interfaces/policy: add support for on-store/on-brand/on-model plug/slot rule constraints [08:41] If errorCatcher is implementing yaml.Marshaler, I doubt it would have access to the tag [08:42] yeah, that's what I see so far [08:42] I'm still looking for _a_ way [08:42] it'd have a pointer to the struct member passed as the receiver, but no way of knowing that it is a member of the struct [08:42] making the struct itself implement yaml.Marshaler on the other hand might do the trick [08:43] but then we're back to square one with explicit checking per-field like I do nwo [08:43] *now [08:44] it _feels_ like schema [08:44] PR snapd#5813 closed: image: warn on missing default-providers [08:44] so maybe the simple fix I did now is OK before we have schema support [08:44] niemeyer: mvo: how much do we still care about old local revisions? I'm talking about local revisions from before we went negative for local [08:45] that is, revisions above 100000 [08:45] we have tests for them, but I suspect we don't really support them any more outside of those tests [08:50] PR snapd#5870 opened: tests/main/parallel-install-services: add spread test for snaps with services [08:59] Chipaca: My only concern is people using these after installing an old image [08:59] Chipaca: I abandon this idea now, [09:00] niemeyer: ack === abeato_ is now known as abeato [09:19] * zyga -> coffee and snack [09:26] mvo: what should I tag warnings as? 2.36? [09:29] Chipaca: yes [09:29] Chipaca: what is missing there? [09:29] Chipaca: I kind of hope to do 2.36 later today [09:30] mvo: https://forum.snapcraft.io/t/warnings/7599 [09:30] mvo: that :-) [09:30] degville: https://forum.snapcraft.io/t/warnings/7599 [09:31] degville: let me know if you can't edit that and want to [09:31] (i know you want to :-p) [09:31] Chipaca: how do you feel about adding a warning that warnings are added to everyone refreshing from old snapd? :) [09:31] (and then we remove it in the next release) [09:32] zyga: before answering I need to know how many other bad ideas you've had today [09:32] Chipaca: ... I need to re-think the idea to get up in the morning [09:32] zyga: I liked at the sprint where you'd announce "this is bad idea #N++" [09:33] :D [09:33] * zyga is now being honest [09:33] zyga: my main problem with your idea here is that it'd be the _only_ warning the user'd see [09:33] oh [09:33] in that case it'd be futile [09:33] zyga: because warnings are there, but nothing uses them as yet [09:33] that's 2.37 material [09:33] any low hanging fruit? [09:34] in that case I'd postpone the idea to 2.37 [09:34] zyga: egrep -ir '(todo|xxx).*warn' [09:34] zyga: and, my secondary problem with your idea is that if we need to do that, then we failed to make them obvious and discoverable and subtle and non-intrusive [09:35] warning: your snapd has updated to x.yz, featuring new functionality and bug fixes [09:35] but that's not a warning [09:36] zyga: and it'll train the user to ignore them [09:36] mmm, perhaps but that would be a nice touch [09:36] so, yeah, -1 [09:36] "snapd is getting better while you are to looking" [09:37] please rate snapd [09:37] mmm, those are way more annoying [09:37] PR snapd#5871 opened: snapstate: only report errors if there is an actual error [09:37] because they are like beggars asking for a coin, here it's more of a "what's new" [09:38] anyway lots of our users will not see warnings unless gnome-software learns about them [09:38] yeah, that's true [09:38] it'd be more welcome if any app could do it [09:38] well, back to propagation [09:38] the app can do it inside themselves [09:39] yeah but if all apps could do it in a standardised way through gnome-software it would be nice [09:39] maybe, maybe not [09:39] everything already wants to speak to me [09:40] pedronis: that's because you're famous [09:42] PR snapd#5872 opened: tests/main/parallel-install-local: rename from *-sideload, extend to run snaps [09:43] today Alice in wonderland would have to agree to a privacy policy, EULA and limited warranty before falling down the rabbit hole [09:50] she wouldn't read them thoroughly, though, or at all, and therefore be violated without her knowing [09:52] PR snapd#5866 closed: tests: remove unneeded cleanup from layout tests [09:53] thanks! [10:09] PR snapd#5873 opened: tests/main/parallel-install-store: run installed snap [10:13] is there any particular reason why core18 edge hasnt seen an update for about 3 weeks now? [10:14] joc: is there a problem with what's there now? [10:14] was hoping for at least a build with ping back as per https://forum.snapcraft.io/t/core18-included-binaries/7191/14 [10:17] joc: yes [10:18] joc: it's because when it is stable we cannot break it [10:18] so that decision was postponed [10:18] joc: I _believe_ this is now the decision of the foundation team though [10:18] mvo: ^ [10:18] zyga: makes sense for stable channel but was looking at edge [10:19] oh, I missed that, sorry [10:19] in that case I don't know [10:19] maybe the transfer to foundations is the reason for the hiatus [10:31] Bug #1771793 changed: Interface for hardware temperature monitoring [10:31] mup is slow, I changed that bug 4 minutes ago! :-p [10:31] Bug #4: Importing finished po doesn't change progressbar

[10:31] haha [10:32] silly bot [10:32] when the machines rebel mup will remember you said that ;) [10:33] I, for one, welcome our shiny metal overlords [10:34] https://memegenerator.net/img/instances/27465850/kiss-my-shiny-metal-ass.jpg [10:35] https://imgflip.com/i/2iu79t [10:41] nice! [10:46] haha [10:46] ah bender and morpheus [10:48] cachio has set up centos images, 2018-09-27 12:47:06 Sending project content to google:centos-7-64 (sep271046-416264)... [10:50] joc: ping is back but the core18 snap is held up in manual review because of this, sil2100 is working on fixing this AIUI [10:51] hi, question about the nodejs plugin: shouldn't it run "yarn install" (with node-package-manager: yarn) automatically during the build phase? [10:52] ackk: I don't think yarn support works right now [10:53] diddledan, looking at https://github.com/snapcore/snapcraft/blob/master/snapcraft/plugins/nodejs.py#L178 it seems it does [10:53] diddledan, I mean by setting node-package-manager to yarn I do get yarn installed, but it doesn't really install stuff from the yarn.lock file automatically [10:53] yes, that's what I'm referring to. it's broken. [10:54] ah, I see [10:54] diddledan, fwiw https://paste.ubuntu.com/p/tYJPk9bt6K/ worked for me [10:57] diddledan, akk can one of you please log a bug about that? [10:58] diddledan, maybe you have a bit more context for the bug ^ ? [10:59] I haven't really dug into it. I meant to, but then forgot :-) [11:04] https://bugs.launchpad.net/snapcraft/+bug/1794727 [11:04] Bug #1794727: node plugin doesn't run yarn install during build [11:07] snapcraft edge channel broken? beta works. edge failure: https://paste.ubuntu.com/p/GW4gWH4nrC/ [11:07] kyrofa, sergiusens: ^ [11:09] edge 2022; beta 1871 [11:09] rbasak: kyrofa is working on it [11:10] rbasak, thank you for letting us know, very nice to have folks testing edge :) [11:11] All over it [11:11] Great, thanks :) [11:12] Chipaca: warnings aren't pushed by anything yet, are they? [11:17] * zyga needs to take the dog for a walk [11:19] zyga: going dogging, eh? ;-p [11:20] I really should take myself for a walk occasionally ;-D [11:20] <-- lazy [11:26] mborzecki: correct; sudo http snapd:///v2/debug action=add-warning message="hello world" [11:28] mborzecki: hi! fyi, was looking at pt 5594 and noticed something. see https://github.com/snapcore/snapd/pull/5594/files#r220886654 [11:28] PR #5594: snap/snapenv: add snap instance specific variables [11:28] jdstrand: looking [11:30] # bind mount *not* used here (see 'parallel installs', above) [11:30] jdstrand: this was fixed later on in https://github.com/snapcore/snapd/pull/5735 which came as a result of discussion in #5713 (parallel installs mount mappings) [11:30] owner @{HOME}/snap/@{SNAP_INSTANCE_NAME}/ r, [11:30] PR #5735: snap/snapenv: drop some instance specific variables, use instance-specific ones for user locations [11:30] PR #5713: many: mount namespace mapping for parallel installs of snaps

[11:30] owner @{HOME}/snap/@{SNAP_INSTANCE_NAME}/** mrkix, [11:31] jdstrand: so $HOME is instance specific path now, same as $SNAP_USER_DATA [11:32] ok, I see that in that pr. thanks! [11:33] jdstrand: thank you for being so thorough [11:33] mborzecki: it sounds like there isn't anything left for me with this feature? [11:34] jdstrand: not at the moment, i'll ping you if i end up doing any tweaks to the interfaces [11:34] mborzecki: ok, thanks! I bet you're pretty happy to see this coming to a close :) [11:34] jdstrand: and thanks for the reviews again :) [11:34] np :) [11:35] Chipaca: aargh, need http unix sockets extension [11:35] mborzecki: snap install http dude [11:36] publisher: John Lenton (chipaca), nice! [11:36] :-) [11:36] mborzecki: that's why it understands snapd:/// at all :-) [11:37] mborzecki: also how it's able to use snapd-control :D [11:39] Chipaca: yay, works now [11:39] pedronis, is it normal that i dont get a snap id here ? https://paste.ubuntu.com/p/ZNcjS2P53M/ [11:39] Chipaca: any plan on where do we start springking warnings first? [11:39] mborzecki: also relevant, there's a unshow-all-warnings debug action [11:40] (i'm trying to pre-connect some interfaces from gadget.yaml but if the system doesnt see an id i can not really do that) [11:40] mborzecki: egrep -ir '(todo|xxx).*warn' [11:49] mvo: how're you feeling about dotfiles for 2.36? [11:50] diddledan: adopting a dog helped with my step count average :) [11:50] :-) [11:52] hmm [11:52] Process '/usr/bin/unshare -m /usr/bin/snap auto-import --mount=/dev/sda1' failed with exit code 1 [11:52] is there any chance to find out *why* it actually failed ? [11:52] i can manually mount the device (and it properly carries an auto-import.assertion file) [11:53] journalctl sadly does only print that one line ... and snap changes doesnt mention it either [11:54] (it is a vfat formatted usb stick) [11:55] ogra: dmesg might have more info [11:55] doesnt [11:55] ogra: that command is run via an udev rule [11:55] manually running the program helps a bit (tells me the user already exists, which is indeed true because i manually created it after the import failed) [11:55] ogra: snapd-autoimport.rules (or 66-snapd-autoimport.rules) [11:55] Chright, might help to connect stderr to stdout here === Chipaca is now known as Chright [11:56] Chipaca, ^^ [11:56] haha [11:57] * Chright ~> lunch-making [11:58] off to pick up the kids === Chright is now known as Chipaca [12:02] niemeyer: do you want to review #5856, or should I merge it? (it has two +1s already) [12:02] PR #5856: cmd/snap: tame the help zoo [12:03] * Chipaca ~> really lunch [12:07] is there a valid way to find the name of the model from a hook without using snapd-control ? [12:09] Chipaca: If it follows what we discussed, +1 [12:10] PR snapd#5874 opened: Adding missing permission to create /run/wpa_supplicant directory [12:20] PR snapd#5875 opened: Avahi interface update [12:24] niemeyer: there's a change, with a new Admin section that has version, warnings, okay, and get, set and wait (meaning the old Config section went away) [12:25] Chipaca: Hmm [12:25] Chipaca: Sounds a bit awkward at first sight [12:25] Chipaca: The whole of "snap foo" is admin [12:25] ogra: looks like is locally installed [12:25] Chipaca: and get/set/wait seem unrelated to the rest [12:25] in which case I think that's the current behavior [12:26] pedronis, indeed it is, i'm trying to develop as bunch of default connections for the interfaces [12:26] (trying to test them before uploading indeed) [12:26] ah, thats pretty unfortunate [12:26] ogra: we don't print a snap id for local snaps (even if there's a corresponding store one) [12:26] i thought it would still pull the id from the store [12:27] that output is actually a bit confusing [12:27] well, it results in the connections to not happening [12:27] yes, connections are only for store snaps [12:28] given they are snap id based [12:28] yeah, thats a bit bad for developing ... [12:28] would be nice if we could use some kind of fake id or some such ... i imaging other developers also would like to test their changes before pushing them to the store [12:28] *imagine [12:29] niemeyer: thing is, what we discussed didn't include warnings, so those need adding somewhere; i tried to iterate on this at the sprint but we were never not busy with other stuff :-) [12:29] niemeyer: so I thought the second best thing would be to iterate in a PR [12:30] Chipaca: Sure, and that's what we're doing here too I guess.. [12:30] ogra: you can always push first as private snap [12:30] niemeyer: dunno, I'm having lunch, not iterating [12:30] ubuntu-image gets annoying but possible [12:31] pedronis, sure, i can also push just to edge and leave it there but preferably i'd like to test locally first [12:31] Chipaca: Well, you're iterating on lunch.. agile style? :) [12:31] anyway, i'll get around ... but it would be great to improve in that area [12:32] niemeyer: not enough beer for this to be XP, so maybe it's scrum? [12:41] Chipaca: dotfiles> I want it in but we can pull it into ~pre2 or ~rc1. it also need a proper name [12:44] I think dotfiles is fine [12:46]