[00:07] PR snapcraft#1783 closed: tests: run codespell as part of static tests [00:08] In my snapcraft.yaml, I have a part which depends on a library and header files I build in a previous part. How do I reference those files in the build? Passing "-I/usr/include" as a CFLAG doesn't automagically resolve, so what path should I pass instead? [00:10] http://termbin.com/icbu is the work in progress [00:12] mcphail: you need to point to their location using $SNAPCRAFT_STAGE [00:12] diddledan: aah. cool. I'll experiment with that. Cheers [00:13] so if they're installing into `/usr/include` then you'd use `-I$SNAPCRAFT_STAGE/usr/include` - but you don't seem to be telling the autotools build where you want the files installed [00:13] which usually means autotools will use `/usr/local` pathjs [00:13] paths* [00:14] yeah - i'll need to work out how to pass a --prefix as well [00:16] Ta. I'll get back to work! :) [00:18] you can pass arguments to ./configure command with `configflags:` in your snapcraft.yaml [00:18] Ta. I think it sets a blank --prefix by deafult, doesn't it? [00:19] `configflags` expects a yaml list, so either new lines with `- ` indicating each item, or `[--flag1, --flag2, ...]` [00:19] difficult to do multiline pastes in IRC :-p [00:20] :) - that's brilliant. Cheers diddledan [00:22] here's an example of the two types of list https://www.irccloud.com/pastebin/8HsmUyMD/snapcraft.yaml [00:23] ta [00:23] I had a hard time grokking yaml when I first came across it because the rules weren't clear to me, but I'm learning [00:25] yay - builds! [00:25] \o/ [00:25] That's super. Thanks for your help [00:28] no probs :-) [02:45] hey anyone have a second to help me with ubuntu core for raspberry pi? [04:04] PR snapcraft#1785 opened: Add type hinting to file_utils.py === kennyloggins is now known as bujeremy [05:35] I'm very annoyed [05:37] I think I'm going to have to go back to vendoring snap-mgmt in Fedora Dist-Git if it breaks like this again [05:38] I'm incredibly disappointed that my feedback was not solicited as the author of snap-mgmt before merging the PR [05:38] whatever, good night everyone [05:53] would bytemark be an SSL or an email-provider with LibreSSL ? [06:15] morning === cpaelzer_ is now known as cpaelzer === cpaelzer is now known as cpaelzer_ === cpaelzer_ is now known as cpaelzer === cpaelzer is now known as cpaelzer_ === cpaelzer_ is now known as cpaelzer === jamesh_ is now known as jamesh [07:13] good morning! [07:17] hey [07:22] hey zyga-ubuntu and mborzecki [07:30] mvo: hey :) [07:32] mvo: how was last friday? [07:32] mvo: any fires? [07:33] zyga-ubuntu: all pretty quiet, no worries [07:33] zyga-ubuntu: quick question, Son_Goku raised an issue https://github.com/snapcore/snapd/pull/4316#discussion_r154561886 and I don't really have enough info to act on it, I see that this is an extra profile that's set up for snap-confine in the core snap, any idea where's the code that installs it? [07:33] PR #4316: cmd/snap-mgmt: introduce snap-mgmt tool [07:37] mborzecki: the interfaces/apparmor backend does this [07:37] mborzecki: setupSnapConfineReexec [07:38] looking [07:41] zyga-ubuntu: so the 'extra' profile is always '${snap-mount-dir}/core//usr/lib/snapd/snap-confine'? [07:41] mborzecki: yes [07:43] zyga-ubuntu: hmm, do you understand what Son_Goku was talking about here then? https://github.com/snapcore/snapd/pull/4316#discussion_r154561886 [07:43] PR #4316: cmd/snap-mgmt: introduce snap-mgmt tool [07:43] I think he was just confused about this aspect [07:44] base snaps, even once we have them, can have any layout but this particular one is specific to the ubuntu-based core snap [07:44] that's my understading as well, just wanted to make sure i didn't miss anything [07:44] as long as snap-confine is in the core snap this code is correct [07:48] zyga-ubuntu: another quick question, any idea why it's sleeping here? https://github.com/snapcore/snapd/blob/master/packaging/ubuntu-16.04/snapd.postrm#L7 i assumed that it's a woraround for some weirdness in older systemd, but then tracked it back to https://github.com/snapcore/snapd/commit/67902f1ef14dd29ff809fabfcb9f7b2cbcd5fb7f and there does not seem to be any reason given for adding this [07:51] mborzecki: looking [07:52] mborzecki: not 100% sure but isn't it the case that "stop" can return before the thing is really stopped? [07:52] * zyga-ubuntu observes fresh snow falling outside [07:55] yeah, this would happen if you hit a timeout in systemctl stop [07:59] hmm the opensuse package has not been updated for quite some time, and the tumbleweed build fails https://build.opensuse.org/package/show/system:snappy/snapd [08:01] ah [08:01] mborzecki: those failures are from new golang [08:01] mborzecki: they are fixed in master, we should just find some time to update the package [08:05] iirc there was some magic that you could do with obs to have it look/update package automatically [08:08] mborzecki: yes but I'm afraid packaging needs real hand care each time [08:08] mborzecki: and we need to test it before unleashing it onto our users [08:10] mborzecki: I'd like to see mvo's face if we had automatic releases to xenial, that just get made without testing :) [08:11] well, ideally all the testing you need should be part of the CI right? [08:11] mborzecki: just like we do in ubuntu [08:11] mborzecki: I think we're not ready for that, there's a class of tests disabled for opensuse [08:12] mborzecki: and nobody on the team is using it 24/7 [08:12] zyga-ubuntu: my face is nothing compared to the users faces ;) [08:16] it might make us find regressions in no time though :P [08:17] crowsouring QA [08:18] s/crowsouring/crowdsourcing/ [08:18] mborzecki: s/crowdsourcing/crowdangering/ [08:29] PR snapd#4344 opened: cmd/snap-mgmt: fixes [08:30] PR snapd#4345 opened: packaging/opensuse-42.2: package and use snap-mgmt [08:34] zyga-ubuntu: 4325 looks ready, if you could re-review we have one less PR :) [08:39] mvo: looking [08:40] o/ [08:41] done [08:41] hey kalikiana [08:42] PR snapd#4325 closed: Adding test for netlink-connector interface [09:07] PR snapd#4346 opened: Make snapd.autoimport configurable [09:07] mvo, ^^^ [09:15] ogra_: thanks, looking [09:15] ogra_: heh, thats a small patch [09:15] yeah, thanks to having a function like that :) [09:16] ogra_: yeah, I was thinking that too: "looks like this was done right" [09:17] we should look at the other bits too one day, i'm sure there is more that could be generalized in such a way [09:17] ogra_: how urgent is it? it looks trivial enough so that we call pull it into .30 even [09:17] not super urgent [09:17] ok, thanks [09:18] (like ... could wait for a 30.1 if one happens anyway or even for .31) [09:18] 2.30.1.1 [09:18] ogra_: ok, I tagged it 2.30, there will be at least one more 2.30~rc2 [09:18] ogra_: so we can pull it in there [09:18] cool [09:45] 21:08:33 I'm trying to add i18n support for snapcraft [09:45] 21:18:48 marked strings, added build_i18n to setup.py, but can't get to make it using .mo files [09:49] PR snapd#4347 opened: tests: increase warn-timeout to 15m [10:02] PR snapd#4348 opened: tests: increase workers for ubuntu-{14,16}.04-64 by one [10:13] hmm [10:13] so [10:13] on core the / inside the core snap doesn't match /snap/core/current [10:13] s/inside the core snap/inside the executiong environment/ [10:14] mborzecki: I'm looking at 4340 right now, expect a review in some minutes [10:15] mvo: thanks [10:21] hmmmm [10:23] zyga-ubuntu: that's surely expected? [10:23] * zyga-ubuntu thinks about http://pastebin.ubuntu.com/26111465/ [10:23] kalikiana: no, it's not [10:23] kalikiana: I think this is a bug [10:24] zyga-ubuntu: what do you see there that's wrong? [10:24] I'd expect mounts to only show up on / [10:25] kalikiana: well, for once, / should really be the core snap (outside the execution environment) [10:25] kalikiana: then looking at that list I find it ... too long [10:26] anyway, I just need to focus and analyze that [10:28] * kalikiana needs coffee [10:33] * ikey needs an army of testers -_- [10:33] mvo: thanks for suggestions, it looks quite nice [10:34] mborzecki: great, happy that you like them [10:34] 27 downloads and 8 of them are me lol [10:34] mborzecki: one thing I forgot, I wonder if we should also ensure that "daemon" is != "" when before/after is used [10:35] ikey: for what snap? [10:35] the lsi snap [10:36] once snapd 2.30 drops ima go enlist an army of testers :p [10:37] hey ikey [10:37] rawr [10:37] ikey: heh, sounds great. I am a bit scared of installing it, once I have steam I will not stop playing it ;) [10:37] mvo: i think after/before is applied regardless of what's in [Service], eg you can have a oneshot service that starts after some target [10:37] mvo, well just install broken games and you'll be saved :p [10:37] haha [10:38] mvo: make sure to play the games you like the least [10:38] ;-) [10:38] mvo: otoh, we cannot really express such relation in snap yaml atm [10:38] mborzecki: yeah, sorry, I was not precise. an app can be any sort of daemon or a standalone binary. so we might want to check to ensure it is really != a standalone applicaton [10:40] * ikey goes off in search of copious amounts of caffeine [10:40] mvo: ok, what you suggest is to raise an error when !app.IsService() ? [10:40] an IV drip will do.. [10:40] ikey: try geen tea [10:40] oh you mean monster? ok [10:41] >_> [10:41] mborzecki: yeah, I think so [10:41] ok [10:41] mborzecki: or do you think thats inappropriate? [10:41] zyga-ubuntu, ikey: green tea++ [10:42] i think it's ok, once we'll probably learn more about use cases once this is introduced into the wild [10:42] mborzecki: *nod* [10:49] mvo: btw, https://gist.github.com/mvo5/6e0ac963aacce9d0c78f1c7cf829f9c4#file-validate-go-L15 this will exhaust stack calling a.String() in a loop :) [10:50] mborzecki: oh, silly me [10:51] mborzecki: indeed, tests ftw :) [10:52] aha, I think I found the bug now [10:53] I think I need that coffee and then I can fix it :) [10:54] PR snapd#4349 opened: Make snapd.autoimport configurable [10:55] so i dont know how much use this is to anyone but i ported the remainder of the 4.13 apparmor stuff into the solus 4.14 kernel last night [10:55] and its working fine for snapd [11:00] ikey: thank you for doing that! [11:00] ikey: is there a lot of delta left? (vanilla vs apparmor) [11:00] 19 files changed, 1882 insertions(+), 28 deletions(-) [11:00] https://dev.solus-project.com/source/linux-current/browse/master/files/security/0001-apparmor-Merge-items-still-missing-from-upstream.patch [11:01] interesting, thank you [11:01] majority of it is because of af_unix/dbus [11:01] well lets generalise that as "socket" [11:01] because net.c is a big component of it too [11:02] PR snapd#4346 closed: Make snapd.autoimport configurable [11:04] honestly i think some of that stuff could probably belong to securityfs and lsm APIs by now [11:04] reduce duplication, same as for the secureexec bits [11:08] mvo: I added Gustavo to ogra PR, not sure we want to surface in the config the implementation detail that this is a separate service [11:09] I'm also a b it worried that people will not understand the implications of turning it off [11:10] it isnt like it is publically exposed anywhere or so [11:11] pedronis: ok [11:12] mvo: alsos why does it make boot slow? do we wait on it? [11:12] PR snapd#4345 closed: packaging/opensuse-42.2: package and use snap-mgmt [11:14] pedronis, the particular customer has a bunch of other disk devices on the HW (that are used for other stuff but some contain vfat, are attached via the usb bus and are slow in responding) [11:14] pedronis: we don't wait for it afaict [11:15] (and they will never use the feature) [11:15] an option to list devices it should use might make more sense [11:15] review for 4338 looks like an easy win [11:15] I don't know [11:15] either way I suspect Gustavo might have opinions how this is spelled [11:16] well, if we cant use the services function it will become a lot more complex i suspect [11:16] but yeah, understood [11:16] mborzecki: does 4308 needs a master merge? [11:18] PR snapd#4304 closed: debian: demote gnupg from dependencies to recommends [11:20] mvo: i'd like to have #4344 resolved first [11:20] PR #4344: cmd/snap-mgmt: fixes [11:20] mvo: shall i add blocked tag 4308? [11:21] mborzecki: sounds good [11:21] mborzecki: and a note ideally [11:21] mborzecki: why its blocked [11:22] Son_Goku: thanks for looking at 4344 [11:23] * Son_Goku mumbles something about reviewed... [11:27] PR snapcraft#1784 closed: cli: show helpful message for 'snapcraft list-keys' with no keys [11:30] PR snapd#4350 opened: debian: make "gnupg" a recommends [11:33] PR snapd#4344 closed: cmd/snap-mgmt: fixes [11:40] mvo: do you have a second? [11:45] hey mpt, can you please take a look at the UX side of snapcraft#1780 ? I've added some comments you can agree or disagree on and pasted some output for it to be easier to review [11:45] PR snapcraft#1780: cli: add export-login command [11:48] mvo, I'll be 5 minutes late today [11:49] mvo, still 2 weeks of kinesiology missing [11:51] PR snapd#4341 closed: tests: new test to validate location control interface [11:56] PR snapd#4351 opened: tests: new test to validate location control interface [11:57] Hi! [11:57] How can I find what those changes are "error: cannot refresh "core": snap "core" has changes in progress"? [11:57] I get the same when trying to update my own snap [11:57] And it does not seem to really have any changes [12:02] looking [12:32] ikey: hey, cool. fyi, you might be interested in https://gitlab.com/apparmor/apparmor-kernel. I don't know how jjohansen is going to have separate branches for different kernels, but you might want to talk to him about that [12:33] jdstrand: hey [12:33] jdstrand: I'm still on that stale core issue, it misbehaves on core systems; I'm looking at it [12:34] zyga-ubuntu: is the misbehavior the needing of the additional rule? [12:34] '/'? [12:34] jdstrand: no, it's a bug where the namespace is always stale [12:34] oh, ok [12:34] jdstrand: trying to check if this is bad logic or bad test setup [12:34] zyga-ubuntu: and hi :) [12:36] zyga-ubuntu: fyi, mvo said that this is not a 2.30 item, but I have a number of 2.30 items to work on this week, so I plan to focus on those and be responsive to your layouts work as have time and hopefully swing back into it next week [12:36] jdstrand: sure [12:37] jdstrand: I think we are close, I need one review on https://github.com/snapcore/snapd/pull/4315 -- I plan to remove the panic but other than that I'd like your eyes whenver you can [12:37] PR #4315: cmd/snap-update-ns: add execWritableMimic [12:38] jdstrand: I'll focus on bugfixes and finishing this this week [12:44] zyga-ubuntu: cool. 4315 is already on the list for background this and planned focus next week [12:44] jdstrand: thank you! good hunting :) [12:47] mvo: hey [12:47] mvo: do you have a moment? [12:53] PR snapd#4324 closed: cmd/snap-confine: discard stale mount namespaces [12:58] zyga-ubuntu: sure [12:58] mvo: hey, let's talk in the standup [12:59] cachio: 5min late> no problme [12:59] zyga-ubuntu: ok [13:09] PR snapcraft#1786 opened: lxd: always install squashfuse [13:19] PR snapd#4078 closed: tests: new test to check interfaces after reboot the system [13:21] PR snapcraft#1787 opened: tests: move refresh to unit folder [13:22] mvo: https://travis-ci.org/snapcore/snapd/builds/311190657?utm_source=github_status&utm_medium=notification this is a build related to arch CI that timed out [13:23] sergiusens: Hey! I'd like to highlight snapcraft#1787 - the refresh test isn't currently run at all and at least one PR erroneously passed because of it, although master seems fine [13:24] PR snapcraft#1787: tests: move refresh to unit folder [13:24] Also elopio, when you get online, you may want to have a look. I'm wondering how this happened, and if maybe we can have some checks to catch "lost" test cases ^^ [13:28] elopio: Hello, are you back from your holidays ? [13:31] mborzecki: thank you! [13:43] * pstolowski heading to the dentist, bb in ~1,5h [13:48] mvo: btw in 2.30 on edge there's no nice transition with tasks from configure-snapd to back to configure, afaict the refresh/task stays stuck [13:49] PR snapd#4347 closed: tests: increase warn-timeout to 15m [13:51] PR snapd#4348 closed: tests: increase workers for ubuntu-{14,16}.04-64 by one [13:55] niemeyer, sorry, I lost internet connection [13:55] cachio: It's okay, we're done.. if issue is entroy we agreed to create /dev/random with the same device as /dev/urandom [13:56] cachio: We really don't care about having proper entropy for tests [13:56] niemeyer, sure [13:56] cachio: This should solve the issue for all cases [13:56] niemeyer, perfect [13:58] niemeyer, is it going to be part of that PR? [14:08] sergiusens: when you have some time, can you have a look at snapcraft#1746 ? to see what's the right way of dealing with the mypy error [14:08] PR snapcraft#1746: cli: add version command [14:09] kalikiana ok, but have you noticed you have an assigned milestone task https://github.com/snapcore/snapcraft/milestone/12 ? [14:10] PR snapcraft#1785 closed: Add type hinting to file_utils.py [14:10] I'll get something to eat [14:15] sergiusens: Yeah, I'm just going through the queue atm [14:19] kalikiana well that issue is due for tomorrow, so I suggest you give it priority [14:19] kyrofa this is something for you https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/armhf/s/snapcraft/20171204_024823_8b570@/log.gz (blocking 2.35 from entering xenial) [14:20] Aye === cachio_ is now known as cachio [14:31] zyga-ubuntu, when you have some time could you take a look to #4218 [14:31] PR #4218: tests: adding tests for time*-control interfaces [14:31] Good morning! [14:31] cachio: sure [14:31] sergiusens: do you think you or your team might come up with a simple dotnet app I can use for docs? [14:31] morning elopio [14:32] popey I can, I have been playing with it [14:32] ooh [14:32] I tried building something I randomly found - openra, and it all fall apart looking for some .net components [14:32] so looking for a simpler example [14:33] elopio also note you have tasks due tomorrow https://github.com/snapcore/snapcraft/milestone/12 ... ideally merged today to catch issues for tomorrow [14:33] On it... [14:33] popey I have a simple forum parser; before New York I had been automating our reports [14:34] I can polish that up a bit and share [14:34] or write something similar to wethr [14:34] * sergiusens commutes back home [14:37] sergiusens: something small and simple would be fine [14:42] mborzecki: I added some more code for the log analyzing and ran it through your timed out PR: http://paste.ubuntu.com/26112553/ - so it looks like we were short on machines when this PR ran and core had just 2 instead of 4 workers allocated (cc niemeyer). lets keep an eye on PRs that time out and see if we find a pattern [14:42] mvo: thanks for investigating this [14:42] cachio, zyga-ubuntu, pedronis if you ran accross prs that time out, please let me know, I want to get to the bottom of those :) [14:42] mborzecki: sure, its a bit annoying so we need to fix it [14:42] mvo: We'll indeed need to increase a bit the number of machines to accommodate the extra distributions being tested [14:43] mvo, https://travis-ci.org/snapcore/snapd/builds/311266636?utm_source=github_status&utm_medium=notification [14:43] thanks cachio [14:43] niemeyer: thanks [14:43] mvo, this is being caused by Cannot allocate linode:debian-9-64: no powered off servers in Linode account exceed halt-timeout [14:43] mvo: Getting to the bottom of it sounds quite correct here :P [14:43] niemeyer: lol [14:43] mvo, I think most of the issues that we see is because lack of linode servers [14:43] mvo, we are in the limit [14:44] cachio, mborzecki: either way works for me.. either mborzecki tries to fix his PR, and then we generalize it.. or mborzecki already generalizes upfront.. whatever sounds easier [14:44] cachio: yeah, looks like your PR was particularly "bad": http://paste.ubuntu.com/26112570/ [14:45] mvo, do you have other travis logs to take a look? [14:45] cachio: I was looking at the one from mborzecki for arch but that was about it [14:46] hmm set up a bot, monitor travis results for jobs started on snapd repo, send email to mvo's inbox :) [14:46] cachio: I wrote two scripts, one looks for allocation times (which are around 1min-2min) for spread machines, except when it can't find any [14:46] cachio: and the other that looks at the individual tests to see how long those take, to see if we waste time somewhere [14:47] mborzecki: heh :) [14:47] mvo, I saw many problems to allocate linode servers the last weeks [14:47] cachio: but it looks like the alloc script is silly, spread already tells us when it cannot allocate things, I was wondering if sometimes linode is slow providing machines but that seems to be not the case [14:47] cachio: yeah, exactly [14:56] PR snapd#4218 closed: tests: adding tests for time*-control interfaces [15:01] sergiusens, done [15:02] mpt thanks! [15:03] PR snapd#4352 opened: tests: increase amount of workers for ubuntu-16.04-64 by one [15:25] mvo: did you see my question about 2.30 ? [15:38] sergiusens, yikes, this is xenial armhf? That means packages have different contents on different archs, that's sad [15:42] Although usr/share/doc/python-minimal/changelog.Debian.gz really shouldn't conflict unless a new version was released [15:43] That seems odd [15:44] sergiusens, that failure is consistent? [16:04] kyrofa you can check all the runs here: http://autopkgtest.ubuntu.com/packages/s/snapcraft/xenial/armhf [16:04] kyrofa look for triggers "snapcraft/2.35" [16:05] kyrofa would appreciate a follow up on snapcraft#1781 [16:05] PR snapcraft#1781: many: set rpath for elf files for classic [16:06] pedronis: sorry, I have not seen your question [16:06] pedronis: here? [16:09] mvo: yes, what should we do/tell people that had 2.30 edge with configure-snapd, our change/fix doesn't unstuck them afaict (I'm in that situation) [16:13] pedronis: snap abort chg-id should get them out of misery [16:13] pedronis: if too many people are affected we may need to do something automatically [16:13] mvo: it doesn't , the next try will still make a change with configure-snapd [16:14] and then we don't understand it anymore, so we get stuck again [16:17] pedronis: hm, ok. thats annoying, so we need a dummy configure-snapd task for edge/2.30 ? [16:17]