=== nsg_ is now known as nsg === cpaelzer_ is now known as cpaelzer [05:51] morning [07:48] good morning o/ [07:53] any idea why cassandra is no longer listed in `snap find --section=database` ? [07:59] pstolowski: hey [07:59] morning! [08:01] pstolowski: what do you know about cassandra snap? [08:01] mborzecki, zero [08:01] linode:debian-9-64:tests/main/searching is failing because cassandra is no longer listed in the output of `snap find --section=database` [08:03] judging by the comments, it's amd64 only, and the test, if running on amd64, will try to look it up in the store [08:07] pstolowski: any ideas why it might be gone now? [08:09] mborzecki, oh, i see. i guess it doesn't really matter, we should just change the test to search something else [08:09] mborzecki, mongo33 looks like a good candidate [08:10] pstolowski: is it also amd64 only? [08:13] mborzecki, yes [08:51] PR snapd#4400 opened: tests/main/searching: handle changes in featured snaps list [08:53] pstolowski: can you take a look ^^ [08:53] looking [09:20] morning peeps. How's things today? [09:23] Chipaca, hey! mind taking a look at https://github.com/snapcore/snapd/pull/4400, tat should unblock tests [09:23] PR #4400: tests/main/searching: handle changes in featured snaps list [09:32] d'oh was looking a while for a "missing" property, only to notice later self wasn't the right object [09:32] * kalikiana coffee break [09:48] pstolowski: good morning [09:48] hey zyga-ubuntu ! [09:48] Chipaca: hi, master is also red with overall timeout atm [09:48] how are you doing? [09:48] * zyga-ubuntu looks at the snowstorm outside [09:48] I'm not doing my usual outdoor stuff today [09:48] decided to hop in to see how you are all doing [09:49] zyga-ubuntu, i'm good, thanks! how are your holidays? [09:49] Chipaca: so reducing to two jobs didn't help (much), start to suspect some network ops that are slow most of the times, but not always? [09:50] pstolowski: good :-) I feel much more at ease than before [09:50] pedronis: should i bump it back up then? [09:50] and catching up on loads of things I put off at home :) [09:50] zyga-ubuntu: dude, if you pop in, we're just going to tell you about things that are broken :-) [09:50] Chipaca: haha [09:50] Chipaca: what's broken? :) [09:51] zyga-ubuntu: JamieBennett [09:51] Chipaca: believe me, usually I'm outside by now but it's snowing so badly I'm not doing that now [09:51] zyga-ubuntu: we broke the boss [09:51] ?? [09:51] Chipaca: I don't know [09:51] is jamie allright? [09:51] zyga-ubuntu, hey, snowing. cmon nice sun here [09:51] Chipaca: we need to understand what is slow [09:51] zyga-ubuntu: all his snaps are stateless [09:51] koza: WAT, where are you? [09:51] zyga-ubuntu: bah [09:52] zyga-ubuntu: it looks like all his snaps can't talk to dbus [09:52] Chipaca: rm -rf ~/snap ? [09:52] zyga-ubuntu: I'm fine but yesterday was a day of finding a tonne of bugs it seems [09:52] zyga-ubuntu: hey, too bored vacationing? :) [09:52] Chipaca: apparmor denials? [09:52] zyga-ubuntu: no obvious unexpected denials [09:52] mborzecki: hehe, just stuck at home for the day [09:52] Chipaca: any missing dbus rules? [09:52] zyga-ubuntu, LDZ [09:52] (not apparmor, dbus) [09:52] zyga-ubuntu: even things like opening a url didn't work i think [09:52] zyga-ubuntu: everywhere from individual snaps to firmware issues with fwupdmgr, to USB C/Thunderbolt bugs [09:52] (and things worked here) [09:52] mmmmm [09:52] Chipaca: that was a bug with unity -> Gnome Shell [09:52] JamieBennett: did you get a priest in [09:53] JamieBennett: what happens when you run snap run --shell in hello-world and then run xdg-open http://example.org [09:53] Chipaca: on upgrade the .desktop file for Chrome when it is pinned to the dock does not contain %u so no urls were working when passed in [09:54] JamieBennett: chrome, not chromium? [09:54] zyga-ubuntu: the url thing was a Chrome issue [09:54] Chipaca: Chrome, yes [09:54] * JamieBennett should switch to Chromium [09:54] or firefox :) [09:55] * zyga-ubuntu has one more AP to update and install at home [09:55] JamieBennett: that's a really old bug: https://askubuntu.com/a/729555/711 [09:55] mborzecki: I'd love to but HO's dont work for me and all my passwords are in Chrome so it would take some time to migrate [09:55] and then the backup disks for everything at home [09:56] Chipaca: good find, I was barking up the wrong tree thinking it was a snap issue and debugging that [09:56] * zyga-ubuntu hugs everyone for using snaps and finding issues we will fix for everyone [09:56] JamieBennett: not so much a find as a memory [09:56] JamieBennett: :-) [09:57] Chipaca: I now have the pleasure of tracking down this bug - https://pastebin.canonical.com/205517/ [09:58] * Chipaca steps slowly away from the JamieBennett [09:58] lol [09:58] on the other hand [09:58] star wars was fun [09:58] go and see it [09:59] zyga-ubuntu: Booked for Christmas Eve, looking forward to it. [09:59] I'll look around for non-dubbed version, maybe without those distracting subtitles next [09:59] I hear Yoda's back, some presenter on UK TV let it slip by accident [09:59] JamieBennett: I have no comment on the matter [10:00] Well, its all over the UK news sites so hard to miss :( [10:00] * JamieBennett stops reading the news [10:02] JamieBennett: in my only comment I will say that the movie is better than I was expecting [10:02] JamieBennett: not sure if that says my expectations were low [10:03] zyga-ubuntu: was it better than the last one? [10:03] JamieBennett: yes [10:03] nice [10:03] JamieBennett: I think it ranks higher than some of IV-VI episodes [10:05] last one as in 'rouge one' or 'the force awakens' ? [10:05] mborzeck1: FA, R1 is somewhat outside [10:06] at least for my comment above === mborzeck1 is now known as mborzecki [10:09] PR snapcraft#1755 closed: tests: collect the autopkgtest results and save them in git [11:19] hello all [11:19] I'm trying to run the static tests and I'm getting an error: ./runtests.sh: line 58: codespell: command not found [11:19] Any suggestions? [11:20] gsilvapt: you'll want to `sudo apt install codespell` [11:21] thank you, kalikiana. I may have forgotten to install the dependencies required. I'll try again now [11:25] I think it should be pip instead of apt [11:25] or both should work [11:25] daniellimws, yes, both work. I thought it was strange to have only apt working === __chip__ is now known as Chipaca [11:28] pedronis, hey, is there any url to get the version of the core snap in edge, or beta? [11:28] so... why on earth is tests/main/interfacs-many taking 5 inutes? [11:28] pedronis, whithout using any header [11:29] cachio: you need to set some headers [11:29] cachio: why? [11:30] cachio: you need to at least set X-Ubuntu-Series [11:30] Chipaca, it is because I need to setup the trigger in jenkins and it just allow to add a url [11:31] Chipaca, is it possible to pass the series as part of the url? [11:31] cachio: i don't think so, no [11:31] cachio: but, give me a bit [11:31] PR snapd#4401 opened: [WIP] snapstate/ifacestate: autoconnect tasks [11:31] no, all apis we have rely on headers [11:32] * pedronis lunch [11:33] * jamesh wishes the travis-ci logs weren't quite so verbose [11:35] cachio: do you need it for both edge and beta? [11:36] Chipaca, yes, but most important is edge [11:41] cachio: https://people.canonical.com/~john/core_edge.json [11:42] Chipaca, I dont have permiossions to see that [11:42] cachio: try again [11:42] great [11:43] Chipaca, where are yo creating this? [11:43] cachio: people.c.c [11:43] cachio: crontab :-) [11:43] cachio: how often should it check? [11:44] Chipaca, every 10 minutes [11:45] Chipaca, thanks, I'll add this to the jenkins jobs [11:45] cachio: any otoher data you need in that json? [11:46] not sure yet, but I think with that info we are ok [11:47] is it for amd64, right? [11:47] because I need for arm64 and armhf [11:51] Chipaca, I'll be back for the standup [11:53] Chipaca, I need these triggrers to run the edge validation on cm3 and dragonboard [11:53] cachio: i can easily change it to be per arch as well [11:54] cachio: different urls though; i'll tell you where when it's done [11:54] Chipaca, :) [11:54] Chipaca, great, thansks [11:55] Chipaca, I'll be back for daily [11:55] cachio: i'll probably be late for the daily [11:55] * cachio afk [11:59] cachio: https://people.canonical.com/~john/core/ [12:01] Chipaca: I was trying to figure about Snappy and Ubuntu Core, I wasn't sure if it was a single project, two Canonical's projects working togheter with Core relying on Snappy, or a third party project that's just being used by Core [12:01] I got curious about "containerized" apps in a desktop system, so Ubuntu Core came out [12:04] PR snapd#4397 closed: tests/main/lxd: temporarily switch to manual [12:05] anddam: ubuntu core is a thing from before snappy [12:05] anddam: snappy ubuntu core is ubuntu core based on snappy [12:05] anddam: in snappy there is a special snap that is called 'core', jsut to keep people on their toes [12:06] (in seriousness, the snap called 'core' is or was when we started a lot like what used to be ubuntu core) [12:06] anddam: but note snaps aren't containerised as i understand that terminology; they're confined [12:08] mborzeck1: you mentioned a mvo script that analysed spread output, do you have that somewhere? [12:08] Chipaca: https://gist.github.com/mvo5/06a206e991b0ae5a53606821dded4bdf [12:09] Chipaca: ah thanks for the info. What's the difference between being containerised and confined there? [12:09] mborzeck1: i was about to dig into interfaces-many's suspicious 5 minute test runs, but thught i should check all of them first [12:09] and, i need to run to the school -- will bbl [12:10] also does snappy aim to be a full system package manager, like can it be seen replacing apt on an ubuntu system at some point in future/ [12:10] anddam: snaps get a restricted view of the actual system (possibly with a different filesystem) [12:10] s,/$,? [12:10] * Chipaca -> afk [12:10] k [12:15] * pstolowski lunch [12:34] Bug #1738197 opened: Daemons do not have an /run/user/* dir created [12:43] hi, I'm getting this error when trying to install a snap with "try": http://paste.ubuntu.com/26182870/ am I doing something wrong? [13:02] pstolowski: standup [13:02] mborzeck1, coming [13:02] cachio: standup === mborzeck1 is now known as mborzecki [13:03] ackk: that's strange, let me test something locally [13:03] Chipaca, I think I found the issue, it seems I can't install the snap anymore in an LXD unless it's privileged [13:04] Chipaca: fyi, I helped him with opening urls, it was a local issue. it's sorted [13:04] (because of a systemd issue) [13:04] ackk: :-( [13:05] Chipaca: ah, I see you know that already :) [13:06] * jdstrand goes back to reading backscroll and commenting on things that no longer need discussion [13:06] * Chipaca hugs jdstrand [13:09] Chipaca: the syscall denial in https://pastebin.canonical.com/205517/ is for listen. that is not in the dbus interface (interesting) [13:09] JamieBennett: hey, would you mind trying something? [13:09] * jdstrand hugs Chipaca [13:11] oh, the new one better than force awakens? wow. /me tries to temper expectations [13:12] (I like force awakens a lot) [13:12] Hello! [13:12] hey niemeyer :) [13:21] * kalikiana going to lunch in ~10 [13:21] niemeyer: o/ what's up [13:24] niemeyer: actually, would you have time for a quick chat? can be later today. there's something I wanted to ask you [13:25] (about architectures) [13:26] hey jdstrand, sure [13:26] JamieBennett: hey. let me try to summarize the problem as I understand it [13:27] JamieBennett: the corebird snap isn't able to communicate on dbus and dies/stops working/doesn't work correctly [13:27] JamieBennett: is that accurate? [13:28] jdstrand: Actually after more debugging I don't think it is a dbus issue [13:28] I used dbus-monitor to see that the settings in Corebird were actually being set [13:28] JamieBennett: I saw a seccomp denial for 'listen' [13:28] kalikiana: Today and tomorrow will be somewhat tight as we're having a two-days meeting, but I will be around next week if we don't find a time slot [13:28] JamieBennett: so this would've caused some grief [13:28] jdstrand: Corebird does not reflect the new settings in the UI [13:29] jdstrand: happy to continue debugging though to see what the real issue is [13:29] JamieBennett: ok, the listen denial probably isn't the cause of that, but it is a bug in the dbus interface regardless [13:29] ok [13:30] JamieBennett: but, if you wouldn't mind added this to /var/lib/snapd/seccomp/bpf/snap.corebird.corebird.src: [13:30] niemeyer: Alright. I'd say if there's time today that'd nice grand, but it's not an emergency so we can do it next week [13:30] listen [13:30] accept [13:30] accept4 [13:30] JamieBennett: then running /snap/core/current/usr/lib/snapd/snap-seccomp compile /var/lib/snapd/seccomp/bpf/snap.corebird.corebird.src /var/lib/snapd/seccomp/bpf/snap.corebird.corebird.bin [13:31] JamieBennett: and then trying again, that would be great [13:31] jdstrand: OK [13:31] i'm off to the 'christmas thing' at my son's kidergarten, bbl [13:31] I'm already preparing a PR for this regardless if it fixes your config issue [13:32] mborzecki: have fun! [13:32] * kalikiana going for lunch now [13:37] Chipaca, do you have the files for armhf and arm64? [13:37] cachio: https://people.canonical.com/~john/core/ [13:37] cachio: I got the time wrong.. I need to head over to the meeting now, but will investigate this issue over the day and ping you when I have something [13:38] Chipaca, thanks a lot [13:45] niemeyer, sure, thanks [13:46] niemeyer, this is a good example https://travis-ci.org/snapcore/snapd/builds/316056668 [13:53] so, interfaces-many legitimately takes 5 minuntes, because it does a 65 connects at an average of just over 4 seconds per connect [13:53] why does a connect take 4 seconds? [13:57] python3 -c $'import yaml;y=yaml.load(open("task.yaml"));\nfor i in ("execute","prepare","restore"):\n if i in y: open(i+".sh", "w").write(y[i])' [13:57] oops [14:07] elopio kalikiana can you enlighten me on why these cleanbuild/container builds fail https://travis-ci.org/snapcore/snapcraft/jobs/316397431 [14:37] Chipaca, could you please dhance the version in the file https://people.canonical.com/~john/core/arm64_edge.json [14:37] I need to test if the trigger is workping properly [14:39] re [14:42] sergiusens: Looking [14:43] kalikiana if you find the reason, fix it up in a follow up PR to that one; I've skipped and logged a bug for that for now [14:43] running the actual tests is more important the having things green running against the wrong things [14:51] Chipaca, there? [15:19] Issue snapcraft#1767 closed: stage-packages does not respect architectures [15:19] PR snapcraft#1788 closed: repo: error for packages with broken dependencies [15:22] PR snapcraft#1447 closed: add support for the "contact" field in snapcraft [15:22] PR snapcraft#1579 closed: Make it possible to use the cmake ninja generator [15:25] PR snapcraft#1726 closed: schema: sources should not have defaults [15:36] Bug #1738222 opened: FAIL: main_test.go:769: snapSeccompSuite.TestCompatArchWorks [15:47] PR snapd#4400 closed: tests/main/searching: handle changes in featured snaps list [15:52] yay, at least one fire is put out [15:53] guys, if you have PRs open and they fail in tests/main/searching, then you pull the changes from master [15:59] PR snapcraft#1807 opened: tests: run test_cleanbuild in LXD on Travis [15:59] cachio: sorry was afk. can i still help? [15:59] (was at physio) [15:59] Chipaca, it is ok, I already validated the triggers [16:00] cachio: ok [16:00] cachio: now, what was the thing about images? [16:00] cachio: I can put .png files on p.c.c, but I don't think that's what you meant :-p [16:04] Chipaca, seems to be working well [16:04] Chipaca, tomorrow I'll check it again to see if the whole thing is working [16:04] Chipaca, hehe, no it is ok [16:04] Chipaca, I need to build the image that I use for cm3 on beta validation [16:04] so, when there is a new core snap in beta (armhf) I need to generate a new image to test the cm3 [16:04] Chipaca, but, first I prefer to make the edge validation work properly and then go to beta [16:04] cachio: ok, fair enough [16:05] cachio: I don't know if p.c.c has any sort of a SLA though; if this is going to block work if it goes down, we should look into moving it somwehre more solid (my main aim was to unblock you today) [16:05] Chipaca, ok, in that case I can setup a service on prodstack [16:06] Chipaca, it was the original idea [16:06] cachio: right, that's a better solution long term [16:06] Chipaca, agree [16:07] Chipaca, thanks for the json files :) [16:07] cachio: i can share the super high tech software that's behind them if you want :-p [16:07] Chipaca, hehehe, yes please [16:08] I am imaging the for [16:08] cachio: https://people.canonical.com/~john/core/get-core [16:08] kyrofa: would you have time for a chat in a bit? about garmmar stuff [16:09] Chipaca, awesome [16:15] kyrofa: elopio: sergiusens Also heads-up I'll be off tomorrow and back on Monday [16:20] PR snapcraft#1803 closed: ci: correctly run from snap [16:22] kalikiana, sure, let me know [16:32] kyrofa: in say 5 minutes? just got out of another meeting and getting a coffee [16:33] Who helps pack the fedora rpm, I'm trying to update suse and I'm just trying to understand the differences and fix some of the issues [16:35] hurricanehrndz, that's Son_Goku [16:35] kalikiana, sure [16:35] oh yeah that's right [16:35] hurricanehrndz: hi [16:36] what's up? [16:36] Son_Goku: I'm just trying to understand some of the linker flags used in the spec file for fedora [16:37] Son_Goku: I have suse build working, but I'm not ending up with static binaries for snap-exec and snap-update-ns, which I understand is required for base snaps [16:37] yeah [16:37] it's a tricky mess [16:37] Son_Goku: Yup, I gather that much [16:37] I had been trying to do the suse one myself, but the golang toolchain is screwy :( [16:38] Son_Goku: Well, I'm almost there https://build.opensuse.org/package/show/home:hurricanehernandez:branches:system:snappy/snapd [16:38] Son_Goku: I'm just updating it. [16:39] Son_Goku: the go macros are especially troublesome and I see how you handle it on fedora, but I'm confused about the linker flags you use in your gobuild macro [16:42] Could you please walk me through them, specifically the '-B' flag, which I thought was for static and dynamic linking but the spec file uses a random hex or is the hex the builduuid [16:42] sure, [16:42] kyrofa: I'm in the weekly now [16:45] hurricanehrndz, also, `gpg2` is not correct [16:46] Son_Goku: Thanks. I started with the spec that was on github within snapcore, go figure [16:46] Son_Goku: ah.. I see you are going through my spec [16:46] it has to be gpg1 [16:47] unfortunately, upstream snapd has not fixed their stuff to use gpg2 instead of gpg1 [16:48] I don't know why you're pulling in the 32-bit libs, though [16:48] apparmor needs to be disabled on %suse_version < 1500 [16:49] hurricanehrndz: I guess you are going to only use vendored tarball? [16:50] kalikiana, sorry, hopping in now [16:50] hurricanehrndz: also... bundling appears to still be outright banned in openSUSE: https://en.opensuse.org/openSUSE:Packaging_guidelines#Bundling_of_multiple_projects [16:50] Son_Goku: Son_Goku: yes [16:51] No worries, I used the time to update some notes from today [16:51] Tedious but useful :-P [16:53] Son_Goku: Hmm Yeah, I will see about unbundling later, I'm using zyga's work as a staring point [16:56] hurricanehrndz: I'd suggest starting from my packaging and start from scratch [16:56] I've made some effort to make the spec relatively portable [16:57] the main thing that'll need to change is dropping the mandatory requires for snapd-selinux on snapd, since openSUSE supports either SELinux or AppArmor [16:58] Son_Goku: Okay, I'm guess this is yours: https://github.com/snapcore/snapd/blob/master/packaging/fedora/snapd.spec [16:58] https://src.fedoraproject.org/rpms/snapd/blob/master/f/snapd.spec [16:58] but yes [16:58] the canonical source tree is always in src.fedoraproject.org [16:58] I don't release from snapcore/snapd [16:59] Son_Goku:Okay I will start there [16:59] Can you explain to me the ldflags on line 52 [17:00] sure [17:00] in Fedora, we put a great deal of effort to make Go generate useful debug symbols [17:00] the ldflags there enable that [17:01] iirc, SUSE doesn't quite care as much about making useful debuginfo as Fedora does [17:01] Cool, what's '-B' flag do? [17:02] I tried reading the linker's man page but could find info on the hex following the flag [17:03] it's a weird go thing [17:04] it basically forces a build-id to be injected into the binaries [17:04] because rpm uses that for debuginfo splitting and packaging [17:05] Son_Goku: Thought so... had a feeling but couldn't find documentation [17:05] should have looked up go linker [17:05] ha ha [17:05] https://fedoraproject.org/wiki/PackagingDrafts/Go#Debuginfo [17:05] Chipaca: fyi, I'm looking into the corebird issue from JamieBennett. we may hand off to the desktop team if I can't reproduce [17:06] Son_Goku: Thank you, love find documentation... [17:06] Chipaca: it seems unrelated to snapd [17:06] Son_Goku: Awesome, you've been a great help, thanks everything you have done so far [17:06] np [17:06] Chipaca: not 100% sure yet since I can't reproduce, but it doesn't look like it [17:07] Son_Goku: One more question, the vendor json specifies the exact versions of the supporting golibraries [17:07] I ignore them [17:07] aside from a few, they're mostly irrelevant [17:08] the latest stuff that matches those library references work [17:08] Son_Goku: Cool [17:08] jdstrand: ack [17:08] jdstrand: thank you! [17:15] sergiusens: FYI I couldn't find a way to reproduce the failing cleanbuild test locally, so I pushed snapcraft##1807 to expose the failure in Travis (which eats the errors by the looks of it) and will investigate this more on Monday - unless elopio beats me to figuring it out [17:15] PR #1807: interfaces/builtin: allow /dev/vhci on bluetooth-control [17:16] * kalikiana hugs mup very affectionately [17:16] ^^ snapcraft#1807 [17:16] PR snapcraft#1807: tests: run test_cleanbuild in LXD on Travis [17:17] * kalikiana wrapping up for today/ this week [17:21] snappy-m-o: let me introduce you to daniellimws. [17:21] Command ":" / ": let" not found. [17:22] no manners... [17:22] haha [17:27] * kalikiana hugs snappy-m-o [17:33] kyrofa: Is SNAPCRAFT_SETUP_CORE still necessary for building classic snaps inside docker containers? [17:33] flexiondotorg, if snapd can't run, it's necessary [17:33] Thanks [17:34] flexiondotorg, without some jiggery, docker falls into that category [17:34] Sure thing! [17:39] flexiondotorg: kyrofa it's becoming automatic as of snapcraft#1801 which will be in the next release [17:39] PR snapcraft#1801: lifecycle: detect docker to auto setup core [18:04] * Chipaca EODish [18:50] Hi guys, does anyone know how can I replicate the GET requests to the snapcraft API ? e.g. sudo snap info [18:50] Is there any public API documentation? [18:52] flexiondotorg: can I pm a sec? [19:02] hurricanehrndz: everything going well with suse packaging of snapd? [19:03] Pharaoh_Atem: Well I just started tackling, I got all the info I need.. just time is required... [19:03] cool [19:05] Pharaoh_Atem: are you on suse? [19:11] hurricanehrndz: I'm on Fedora, openSUSE, and Mageia :) [19:13] Pharaoh_Atem: Cool, I'll send you a msg once it's ready for testing [19:24] flexiondotorg: tis ok, I got ahold of popey [19:32] brunosfer: try running snapd with SNAPD_DEBUG=1 [19:33] brunosfer: example curl based on what's in the logs: https://paste.ubuntu.com/26184762/ [19:42] mborzecki: I stopped the service snapd, but now I can't find the path to the snapd binary file... [22:47] Bug #1738295 opened: snap auto-refresh re-installs removed snaps