[07:32] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-bionic/bionic/amd64/s/snapcraft/20181108_122554_7d4c2@/log.gz [07:32] raceback (most recent call last): [07:32] File "/tmp/autopkgtest.P8IYeA/build.5xT/src/tests/integration/plugins/test_rust_plugin.py", line 107, in test_cross_compiling [07:32] self.assertThat(binary, HasArchitecture("aarch64")) [07:32] File "/usr/lib/python3/dist-packages/testtools/testcase.py", line 498, in assertThat [07:32] raise mismatch_error [07:32] testtools.matchers._impl.MismatchError: Expected 'aarch64' to be in ' x86-64' [07:32] ---------------------------------------------------------------------- [07:32] Ran 26 tests in 2831.729s [07:32] FAILED (failures=1, errors=1) [07:32] autopkgtest [11:08:32]: test integrationtests-plugins: -----------------------] [07:32] autopkgtest [11:08:33]: test integrationtests-plugins: - - - - - - - - - - results - - - - - - - - - - [07:32] integrationtests-plugins FAIL non-zero exit status 1 [07:48] zyga: ^^^ and armhf fails due to some cross compilation stuff [08:31] test [08:31] hi everybody, can somebody be of any help with the following question? [08:32] when I install Firefox via snap infrastructure, will it touch home directory of my currently installed FF ESR? [08:33] still running FF ESR 52 because of much needed adddon [08:48] mvo: so snapcraft autopkg tests still fail trying to build some obscure rust code downloaded from the net, and due to some cross compile failure [08:55] doko: good morning! this is probably best discussed with sergiusens - but let me have a quick look, maybe something trivial could be done to unblock things [08:55] hey Chipaca ! good morning [08:55] mvo: hey! how're you doing? [08:56] Chipaca: good, thank you! had a nice and refreshing weekend [08:56] Chipaca: didn't play hockey for family reasons, that made me slightly sad [08:56] Chipaca: but otoh family made me happy so a overall a net plus :) [08:56] Chipaca: and you? [08:57] mvo: :-) [08:57] mvo: I didn't play hockey either! [08:57] :-D [08:57] lol [08:57] mvo: well, at least he's highlighted [08:57] mvo: but a good weekend, yes. Planning a break early december though. [08:58] doko: I think he got highlighted last week already :/ [08:58] Chipaca: uh, I need to plan my break as well [08:58] last week was the summit though [08:58] pedronis: aha, good point [08:58] Chipaca: when? [09:01] pedronis: i've got 15 days to use-or-lose. Do you remember if it was 3, or 5 that could be carried over? [09:02] pedronis: in any case: the week from 3-7 december, for sure [09:03] doko: quite a few errors, I think I leave this best to sergio, I could disable some of the tests that look like upstream changes but other bits look like they have different causes :/ [09:03] Chipaca: 5 [09:04] mvo: ta [09:04] Chipaca: sounds like you will enjoy a long vac in dec then :) [09:04] Chipaca: good for you (not so good for us ;) [09:04] mbuahaha [09:04] * Chipaca takes all december off [09:04] (no) [09:05] fwiw 6127 needs a review, apparently this helps bringing a machine back to life that got into a bad state by disabling a snap on arm that uses gpio [09:06] mvo: it has so tests atm tough, zyga said he was working on some [09:06] pedronis: :( ok, I can look at this, zyga is off today [09:09] mvo: we also need to decide how to distribute the fix [09:10] pedronis: mvo: what's the bug? [09:11] pedronis: my understand is that putting it into the edge core is ok for now. then the user can snap refresh --edge core, snap enable affected snap and snap refresh --stable core [09:11] Chipaca: gpio interface require some order dependency across the systemd and the apparmor backend for slot and plug [09:11] seems we fix some part of that, but enable/disable of affected snaps atm is broken [09:12] ah [09:12] mvo: ok, that assumes they are happy to try edge for this [09:12] anyway we need on edge to start either way [09:12] reminds me, from the summit: we need a way to install a snap with its services disabled [09:13] pedronis: it does, we could use a hotfix branch, not sure if all the bits are in place for this though [09:13] ? [09:14] pedronis: sorry if I was unclear, a store branch in a channel [09:14] yes [09:14] those are supported [09:14] you mean building it on our side? [09:14] I think we have not used one for us yet [09:15] pedronis: yeah, we could take a stable core and do a one-off build for them, its a bit of work though (mostly because we need this for arm) [09:15] so not sure its really worth it === cpaelzer_ is now known as cpaelzer [09:15] as I said we need to ask [09:15] ok [09:18] Chipaca: I updated 6039 with your suggested error message (I think its a good choice) [09:22] mvo: over the weekend I was reminded of #1669000 [09:22] Bug #1669000: classic snap can't use confinement override [09:24] Chipaca: is jacknorriswebserver a classic snap? [09:25] Chipaca: we need to adjust our error message here, correct? [09:25] Chipaca: or is there more to it? [09:25] mvo: ignore the last comment, i think they're off piste [09:25] Chipaca: ok [09:26] mvo: it's very close to 6039, but from the other side [09:26] mvo: yes we probably should check and return a better error, like with --classic, but i'm not sure how current the problem is [09:26] i haven't tried [09:26] i just saw it, thought "oh that's similar to 6039" and went on with my weekend [09:27] Chipaca: heh, ok [09:27] Chipaca: heh [09:27] anyway doesn't sound super urgent [09:28] correct [09:28] talking about urgent, i'm off to epochs land [09:41] Chipaca: let me know how that goes, we should chat about how to do the further requests bit [09:41] soon [09:43] pedronis: yep [09:46] hey [09:46] I'm off, just wanted to say a few words [09:46] I ran into some interesting bugs while reinstalling my devices [09:47] https://bugs.launchpad.net/snapd/+bug/1802773 [09:47] Bug #1802773: Cannot refresh from ubuntu-core 2.15.2 on raspberry pi "ubuntu core 16" image - snap-declaration for "snapweb": parsing assertion headers: expected 4 chars nesting prefix after multiline introduction "plugs:": " network:" [09:47] our stable images cannot update [09:51] with that I'm off to fix the printer [09:51] ttyl [09:52] zyga: that's a store problem I think [09:52] yes [09:59] pedronis: the action context does need to carry epochs always, right? [10:00] Chipaca: correct [10:00] afaiu [10:00] pedronis: a lot of the test changes are going to remain ¯\_(ツ)_/¯ [10:01] oh well [10:01] we still want the change [10:01] saner api use [10:08] * Chipaca ~> afk [10:12] * Chipaca ~> really afk === chihchun_afk is now known as chihchun [11:29] Chipaca, hi, FWIW I filed https://bugs.launchpad.net/snapd/+bug/1802721 about that issue I'm having with bash completion [11:29] Bug #1802721: bash-completion not working on core18-base classic snap === chihchun is now known as chihchun_afk [12:09] Chipaca: were snapshots done in 2.36 or will in 2.37 ? [12:20] mvo, hey [12:21] hey cachio [12:21] mvo, about beta validation [12:22] there were some issues on dragonboard [12:22] cachio: oh, tell me more please [12:22] mvo, first I saw this error [12:22] on 2 executions [12:23] same error [12:23] cachio: do you have a link to the error text? [12:24] https://paste.ubuntu.com/p/jZbsWd7x36/ [12:24] next execution 13 tests failed [12:24] mvo, I am gonna debug nor [12:25] now [12:25] mvo, I go slow with the db because mine is broken so I am using a remote one [12:25] mvo, now I'll try to use mine with external display and usb keyboard [12:26] cachio: hm, Cannot access MTD device /boot/uboot/uboot.env: No such file or directory look scary [12:26] mvo, the rest of the devies are ok [12:26] mvo, yes [12:26] cachio: please check if /boot is mounted once you have a debug session [12:26] I check back after lunch [12:27] mvo, I could reproduce that in 2 different devices [12:28] mvo, sure [12:28] I'll continur with this [13:03] sergiusens: o/ [13:03] sergiusens: does snapcraft support use of the before/after stanzas for daemon application startup ordering within a snap yet? [13:03] or have I got the wrong end if the stick on that feature being implemented [13:09] cachio: mvo: what were the errors during the weekend about snapd-vendor-daily? [13:16] pedronis, checking [13:17] pedronis, I dont see any error on snapd-vendor-sync [13:17] at least the spread cron branches worked [13:18] and all of them are in green [13:18] cachio: I'm talking about the recipes on LP [13:18] [recipe build #1985875] of ~snappy-dev snapd-vendor-daily-cosmic in cosmic: Failed to build [13:18] etc [13:19] pedronis, ahh, ok, the recipe failed to build [13:20] I'll check that [13:20] I am gonna add a new spread cron branch to check the result of that recipe [13:20] currently we are just triggering but not checking the resutls [13:25] pedronis, seems to be a infrastructure issue [13:26] failed doing "git fetch" [13:26] on xenial and cosmic as well [13:36] ackk: yes that issue is true. For now, install core to get tab completion even if your snap is core18 [13:36] ackk: #6126 addresses it [13:36] PR #6126: dirs, wrappers, overlord/snapstate: make completion + bases work [13:37] Chipaca, as noted in that LP bug, even with core it doesn't work [13:38] ackk: with core, it works [13:38] Chipaca, not in my case, see the bug [13:38] sigh [13:38] ackk: the bug just says "it doesn't work" [13:39] ackk: https://forum.snapcraft.io/t/debugging-tab-completion/4198 [13:39] Chipaca, well, no also says I followed the forum thread, and everything except step 4 works [13:39] Chipaca, including the last step, if I load completion in the snap manually [13:39] ackk: ah sorry [13:40] ackk: I'd misread and misremembered it and just scanned it and jumped to conclusions and now I feel terrible and I'm sorry [13:40] give me a sec to recover and i'll walk through it with you in a moment [13:40] Chipaca, heh, np [13:41] ackk: for step 4, where's the -x output? [13:41] Chipaca, you want the full output or just the output of compelte -P? [13:42] Chipaca, https://paste.ubuntu.com/p/wdfXH7yHcf/ [13:43] PR snapd#6128 opened: overlord/ifacestate: setup security backends phased by backends first [13:43] ackk: right [13:43] ackk: so there's something missing [13:43] ackk: in that pastebin, see how the returned string points to a "cannot" [13:43] Chipaca, yeah but that's in the script [13:44] I mean that's not the output of complete, it's the script tha compares with "cannot" [13:44] ackk: so there _probably_ is a "cannot find completion specification for yadda.yadda" [13:44] hmm [13:44] * Chipaca looks [13:45] ackk: you're right [13:45] Chipaca, the problem I think is that complete -P only returns "default" and empty lines (as mentioned in the bug ) [13:45] Chipaca, but if I run it in the --shell it works [13:45] (hence I was suspecting that --command=complete does something different [13:49] ackk: so [13:49] ackk: if you run snap run --command=complete sshoot 9 9 7 1 ' ' 'sshoot ' sshoot '' [13:50] ackk: what do you get? [13:51] ackk: because I get a clear indication of where the problem is [13:51] ackk: so I was wondering what you saw :-) [13:57] Chipaca, https://paste.ubuntu.com/p/tntr3dNtSW/ [13:57] Chipaca, with multiple blank lines [13:58] gah pastebin is very slow here [13:58] not sure why [13:59] ackk: https://dpaste.de/jr9h [13:59] Chipaca, are you running on bionic? [13:59] ackk: no [14:00] Chipaca, that's another bug I filed, that mvo is fixing, classic bionic snaps don't work on xenial [14:00] ah this is the bad symlink thing? [14:00] yeah [14:00] mvo: is that fixed on core18 edge? [14:00] * Chipaca checks [14:00] i'm on edge [14:00] so probably no :-) [14:01] Chipaca: standup? :) === chihchun_afk is now known as chihchun [14:01] gah, standup o'clock [14:01] pedronis: omw [14:21] ackk: so the next step would be to do the 'snap run --shell', and run etelpmoc.sh [14:22] ackk: which takes a lot of arguments, but it's got a comment explaining them [14:22] ackk: that's what 'snap run --command=complete' thing does, once it's 'inside' [14:23] Chipaca, /snap/core/current/usr/lib/snapd/etelpmoc.sh is not executable, is that right? [14:23] does snap run run it with "bash" ? [14:25] ackk: yes [14:25] Chipaca, https://paste.ubuntu.com/p/PYy443Xw9S/ [14:26] Chipaca, note that if I actually source the sshoot-completion script inside a --shell, it works [14:26] Chipaca, why is it not found? [14:26] ackk: because I don't think the arguments are right [14:26] uhm [14:26] ackk: dunno [14:26] ackk: i'm technically in a meeting [14:26] I passed the same as the --command=complete, plus the completion script name at the beginning [14:26] ackk: half a neuron on this [14:26] Chipaca, ok sorry [14:26] :-) [14:27] _die "USAGE: $0