[04:17] Bug #1625279 changed: cmd_run.go:179: WARNING: cannot create user data directory: cannot create "/home/$USER/snap/$SNAP/$VERSION": mkdir /home/$USER/snap/$SNAP: permission-denied [06:38] who knows how to do the os.snap? [06:53] morning [06:54] starting late again today [07:04] mvo: morning [07:06] hey mborzecki - good morning! [07:09] mvo: i played with /dev/random patchset over the weekend, had mixed success rate [07:11] mborzecki: oh, why? is gpg too smart? or does it not really help? [07:11] one thing that bugs me is that if gpg-agent starts reading /dev/random (the real one) and that out of entropy, simple SIGTERM to gpg-agent has no effect, I actually had to SIGKILL it [07:13] and the highest success rate was when i juggled /dev/random in task's prepare/restore rather than on the project level [07:18] mborzecki: woah, that is interessting, I would not have expected this [07:38] PR snapd#4385 closed: interfaces/desktop,unity7: allow status/activate/lock of screensavers for 2.30 [07:39] PR snapd#4384 closed: interfaces/desktop,unity7: allow status/activate/lock of screensavers [07:40] PR snapd#4378 closed: tests: do not disable refresh timer on external backend [07:40] PR snapd#4379 closed: client: send all snap related bool json fields [07:41] PR snapd#4375 closed: interfaces: interfaces: also add an app/hook-specific udev RUN rule for hotplugging for 2.30 [07:41] PR snapd#4376 closed: tests: fix external backend for tests that need DEBUG output [07:52] good morning! [07:52] hey kalikiana, good morning [08:04] mvo: hey, i've just switched over to golang 1.9 being the default in bionic [08:04] mvo: snapd seems to build fine with it, just wanted you to know [08:10] mwhudson: great, thank you [08:11] mwhudson: any discussion on having a golang-1.9 in xenial-updates (just like we have 1.6 for trusty) ? [08:11] mvo: no [08:11] mvo: the latest juju in xenial-updates is a multi-orig bundling go 1.8.3 ... [08:12] mwhudson: right, I guess we could do the same [08:13] mvo: i /think/ the security team prefer this because it means that we don't end up with golang 1.8 and then 1.9 and then 1.10 all arriving in main and being there until EOL [08:13] mvo: but you should check that with them :) [08:20] mwhudson: right, it makes (some) sense [08:30] good morning! I have a question. I've not find any info about it. and also i ask the question on forum.snapcraft.io, but nobody reply it. [08:30] https://forum.snapcraft.io/t/how-to-make-the-os-snap/3127 [08:30] the question is Ubuntu core has three snaps at least : a Core snap, a Kernel snap, a Gadget snap . And now i have a ubuntu minimal system. how to snapcraft the system to a core snap? anyone can help me? thanks very much. [08:31] PR core-build#24 closed: ubuntu-core-rootfs: mount os/kernel snaps RO [08:33] thanks very much [08:36] hey haisheng-from-ch - there is now way today now to replace the core snap. however we are working on a new feature called "base" snaps that will allow you to have your own core (it will be called differently) from your own set of binaries. [08:38] PR snapd#4374 closed: interfaces: interfaces: also add an app/hook-specific udev RUN rule for hotplugging [08:44] mvo, mup.thanks. But i don't understand it very well. === davidcalle is now known as davidcalle_bluef [09:41] mwhudson: FYI I updated snapcraft#pull/1771 now based on the review conversation [09:41] snapcraft#1771 even [09:41] PR snapcraft#1771: lxd: suppress traceback when lxc launch / init fails [09:42] * kalikiana going for a coffee break [10:32] PR snapd#4386 opened: release: 2.30.rc3 [10:39] cachio: hey, good morning. today is release to stable day. lets notify the store and aim for a release after the standup, ok? [10:39] mvo, yes [10:40] mvo, good afternon [10:41] cachio: thank you! not quite afternoon yet, shortly before lunchtime :) [10:41] cachio: feels like afternoon already though ;) (kidding) [10:41] mvo, >( [10:41] :) [10:42] any PRs needign a second review? [10:43] pstolowski: 4221 looks good to me now, do you want to have a final look at it and then merge ? [10:43] mvo, rc3 is ready for beta validation? [10:43] cachio: correct [10:43] mvo, great [10:44] mborzecki: not sure, if I see one I send it your way. 4221 might be one (given that zyga is out) [10:44] mvo, yes, looking [10:44] mborzecki: if pstolowski is already looking I think thats enough though :) [10:44] pstolowski: thank you [10:44] ok [10:45] mborzecki: 4354 has conflicts now [10:45] ok, let me merge master there real quick [10:46] mborzecki: ta [10:47] mvo: done [10:47] i'm surprised there's no command line switch or a config that you can pass to gpg gen-key to tell it to use some particular device [10:48] although there's a switch when calling gpg --get-random, you can specify the quality of the random source, anything below 2 seems to pick /dev/urandom [11:04] mvo, +1 [11:12] pstolowski: thank you [11:13] PR snapd#4221 closed: tests: new test for interface network status [11:15] np [12:35] mvo: project prepare is run once before all the tests that execute on given worker right? [12:45] that's weird, https://paste.ubuntu.com/26163192/ it's like the code that switches /dev/random did not execute at all, a restore was run before the task started or the machine was rebooted [12:54] mvo: hey, I think I can get rid of the write on ~/.gnupg/random_seed with '--no-random-seed-file'. We would just need to document that [12:54] mvo: shall I test and submit a PR? [12:57] mvo: (talking about with the new gpg-keys interface) [12:58] mvo: you know what, I'll submit it and then people can discuss [12:59] mborzecki: hm, strange. is that your arch PR? I ca nhave a look [12:59] jdstrand: +1 [12:59] it's the fake random thing [12:59] mborzecki: ta [13:00] #4354 [13:00] PR #4354: tests/lib: introduce helpers for setting up /dev/random using /dev/urandom in project prepare [13:03] mborzecki: standup? [13:03] joining, chrome got stuck [13:26] * kalikiana going to leave for lunch in a few minutes [13:28] PR snapd#4387 opened: interfaces/gpg-keys: force use of '--no-random-seed-file' via explicit deny [13:28] mvo: https://github.com/snapcore/snapd/pull/4387. this addresses Gustavo's remaining concern and the more I think about it, the more I think it is the correct choice (see my full description) [13:28] PR #4387: interfaces/gpg-keys: force use of '--no-random-seed-file' via explicit deny [13:29] mvo: if you agree, I can submit a PR for 2.30 too [13:31] jdstrand: looking [13:40] I’m trying to run SDP server on bluetooth using my snap with sdptool, however I get the error “Failed to connect to SDP server on FF:FF:FF:00:00:00: No such file or directory” do you know how can I solve this problem in ubuntu snappy core? [13:45] jdstrand: I replied in the PR, thanks for looking into this [13:46] mvo: I did too, which is why I mentioned both sides of it :) [13:47] mvo: pushed a trivial fix to your #4334 PR [13:47] jdstrand: heh, I leave that for gustavo to do the final decision [13:47] PR #4334: tests: ensure snap-confine apparmor profile is parsable [13:47] * jdstrand nods [13:47] mborzecki: nice, thank you [13:47] mborzecki: heh, thanks, just read the diff [13:52] I’m trying to run SDP server on bluetooth using my snap with sdptool, however I get the error “Failed to connect to SDP server on FF:FF:FF:00:00:00: No such file or directory” do you know how can I solve this problem in ubuntu snappy core? [14:09] mvo: I found a fun bug, if auto refresh refreshes exactly 2 snaps, we get no summary :) [14:09] because C and Go switch don't work the same [14:27] re [14:47] pedronis: ha! fun [15:09] mvo: can you do a second pass on #4354? [15:09] PR #4354: tests/lib: introduce helpers for setting up /dev/random using /dev/urandom in project prepare [15:11] I know you will all hate me, but I'm freezing. 14C, this is not nice. [15:12] I need to run to the beach, can't feel my hands :( [15:13] mborzecki: sure [15:14] * genii throws a long-sleeve shirt on to go out and have a smoke in the -6C Toronto weather [15:14] mborzecki: woah, its *green* :) [15:14] genii: I cry for you. [15:14] heh [15:16] mvo: yeah, I know :) [15:19] mvo: thanks [15:19] PR snapd#4354 closed: tests/lib: introduce helpers for setting up /dev/random using /dev/urandom in project prepare [15:36] elopio: hey [15:36] when you're around, can I ask you something about codespell results? [15:44] kalikiana: of course, shoot. But that was done by a gci student while I was on holidays, so I might have ignorant answers :) [15:47] elopio: ah, right [15:47] PR snapd#4388 opened: overlord/snapstate: fix auto-refresh summary for 2 snaps [15:48] mvo: ^ , I didn't find any preexisting tests for that code [15:49] elopio: so my question might be a request for enhancement of sorts. Was wondering what to expect when running it locally. I'm getting a ton of "WARNING: Binary file" and it seems to search .git files as well and complaining about errors from other branches. [15:50] pedronis: looking, thank you [15:50] elopio: So my main issue is that it's very noisy.... although I've managed to identify the real complaints from noise so far [15:50] pedronis: I can look into the test, a welcome change from wrestling with the default-provider tests :) [15:56] mborzecki: you had a gocheck that does pretty-print for struct diffs, right? where can I find that? [15:57] mvo: https://github.com/bboozzoo/check branch bboozzoo/pretty-equals-not-equals [15:57] mborzecki: ta [15:57] you will need to do go get the dependencies [16:00] mborzecki: I think you should submit that to check (if you haven't already). iirc gustavo asked to change the difflib to the other pretty print thing (kr/pretty iirc) [16:01] i'll change the pretty printer before submitting [16:01] i also started working on subtests a bit [16:02] mborzecki: neat [16:02] mborzecki: makes my life already easier, thanks for this! [16:02] yw [16:03] PR snapd#4389 opened: overlord/snapstate: override Snapstate.UserID in refresh if the installing user is gone [16:04] mvo: ^ this deals with the other issue I mentioned in the standup, not the only option to handle it though [16:05] pedronis: looking [16:13] ok, i'm done for today, need to get back to preparing something for the meetup [16:13] enjoy the rest of the day guys [16:37] o/ [16:37] how are things [16:37] sergiusens: [16:37] oops [16:37] sergiusens: the phrase SNAPCRAFT_SETUP_CORE=1 appears in search results in bug reports and PRs, but I see no mention in documentation. What does it do? [16:37] (also, should I file a bug about missing documentation)? [16:51] popey, if set, snapcraft will download and unpack the core snap in order to build classic snaps [16:52] yeah, found that in one of the many bug reports, but couldn't find any official docs about it [16:52] popey, if memory serves correctly, it was added so classic snaps could be build on the LP builders back when they were only chroots that couldn't run snaps [16:52] It was a workaround easter egg, not really a feature [16:52] seems other people are using it, like microsoft in vscode and others [16:52] haha [16:52] Huh [16:53] Typically it's expected that one has the core snap installed (or installs it) in order to build classic snaps [16:53] That wasn't an option on LP before they started using lxd [16:53] These are building on docker [16:53] Ahhhh, same problem then [16:59] * kalikiana waves at kyrofa o/ [17:00] Hey there kalikiana [17:00] * kalikiana about to wrap up for the day [17:00] I've got some reviews heading your way [17:00] * Son_Goku hates compiler bugs... [17:01] kyrofa: sure. feel free to drop me an email and I'll check them in the morning [17:22] * kalikiana out [17:40] popey no, we do not want to be documenting that, we want to detect we are in a docker container instead [17:47] kyrofa elopio why is lxd from edge installed for our tests? [17:47] that could explain some of the randomness seen [17:47] Good question, no idea [17:47] also, I changed the approach in snapcraft#1798 to be quirk based instead of plugin driven [17:47] PR snapcraft#1798: go plugin: strip sections that patchelf does not handle [18:09] sergiusens: there was a weird combination of lxd stable and snappy stable that didn't work. We can try to get back to having both on stable. [18:10] elopio I think, that that should work these days, well I hope at least, and if not, let's log bugs [18:10] kalikiana: sorry, I missed the thing about codespell. You can run it with ./runtests.sh static. That's not noisy for me. [19:56] elopio I don't think we are running from the snapcraft snap anymore -> https://github.com/snapcore/snapcraft/pull/1798 [19:56] PR snapcraft#1798: elf: strip the .note.go.buildid to make room for patching elf [20:25] kalikiana: thanks [20:26] kalikiana: at some point it's just easier for the reviewer to make the change that beat the concept through the dense skull of the submitter :) === alan_g is now known as alan_g|EOD [21:49] PR snapcraft#1771 closed: lxd: suppress traceback when lxc launch / init fails [21:50] kyrofa popey snapcraft#1801 [21:50] PR snapcraft#1801: ci: detect docker to auto setup core [21:52] PR snapcraft#1801 opened: lifecycle: detect docker to auto setup core === ondra_ is now known as ondra === StoneTable is now known as aisrael === Kamilion|ZNC is now known as Kamilion [22:19] elopio, you simply must buy one of these kits [22:29] elopio, hey [22:29] elopio, I see some snapcraft errors running sru validation for snapd 2.29.4.2 [22:29] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/amd64/s/snapcraft/20171201_043445_970ee@/log.gz [22:29] elopio, any idea what could be causing that? [22:32] cachio, I see at least one yarn timeout [22:32] Yeah, looks like a few timeouts [22:34] kyrofa, ok, thanks, I'll run it again tomorrow in that case [22:35] kyrofa, it is weird because it also failed in the other systems [22:35] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-zesty/zesty/i386/s/snapcraft/20171201_020201_0a045@/log.gz [22:35] this is for 32 bits [22:36] snaps_tests.CommandError: Error running command 'snapcraft snap' in '/tmp/tmpg4dwm24g/webchat'. Expected output 'Snapped .*\\.snap' not found. [22:36] same error [22:36] cachio, yeah, same "unexpected error occurred: "https://registry.yarnpkg.com/express: ETIMEDOUT"" [22:36] Sounds like yarn generally hiccuped for a few minutes [22:36] Or the adt infra had network issues [22:37] kyrofa, yes [22:37] makes sense [22:37] I'll run it again [22:38] kyrofa, thanks [22:38] Any time :) [22:38] I know how hard those are to parse [22:39] :) [23:32] Hello, I have one specific task that I would like to run on a single processor core of my machine, and have nothing else interrupt that core. This seems to be impossible on full desktop Ubuntu, but is it possible on Ubuntu Core? === ubott2 is now known as ubottu