=== hyperreal_ is now known as hyperreal === JanC is now known as Guest69049 === JanC_ is now known as JanC [04:48] morning [04:58] is there a way to build an ubuntu core image without a snapcraft account [05:05] good morning [05:07] * zyga feels more or less normal now, let's hope it will stay that way :) [05:08] zyga: hey [05:08] zyga, hey [05:22] mborzecki, you start early [05:22] cachio_: yup, always like this, leaves more day to attend kids & other things [05:22] cachio_: what time is at your place now? midnight? [05:22] mborzecki, 2 am [05:22] 2:22 [05:23] I am almost done [05:23] cachio_: wow ;) insomnia? [05:23] mborzecki, no, fixing the images [05:23] cachio_: ayy [05:23] all the builds were failing with the 16.04-32 and fedora-27 [05:24] cachio_: was that a no space left thing? [05:25] mborzecki, no, Error: Failed to synchronize cache for repo 'updates' [05:25] installing, or updating with dnf [05:26] maybe high load on fedora mirrors, people updating their systems [05:27] mborzecki, yes, it could be, I regerenated the image many times with different changes and it didin't work [05:27] didn't the same happen when 27 was released? [05:27] iirc we switched fedora to manual then [05:29] mborzecki, yes, something similar, but as yesterday I updated this iamge I thought that perhaps something new was causing this [05:31] mborzecki, I'll create a PR to set fedora as manual until it is fixed [05:31] ok [05:32] cachio_: have you looked into preparing f28 image by any chance? [05:32] mborzecki, no, should I? [05:32] mborzecki, I mean, are we going to support it? [05:33] cachio_: yeah, it'd be great [05:33] mborzecki, is it already released? [05:33] or it is comming soon? [05:34] cachio_: it's released (that's probably why the high load on the mirrors) [05:37] mborzecki, ah, ok [05:37] mborzecki, I am runnin a script to create the google image [05:37] mborzecki, not sure if it is gonna work [05:38] cachio_: no need to do it now, it's already late for you anyway [05:38] mborzecki, I jut changed few 7 for 8 [05:38] mborzecki, not big deal :) [05:38] cachio_: hah ;) [05:41] mborzecki, failed, google packages not ready for 28 [05:41] mborzecki, I'll try in few days [05:41] ok [05:49] cachio_: what does it mean "google packages not ready for 28"? [05:56] zyga, things like this package google-compute-engine-init-2.1.2-0.1488484921.el7.x86_64 requires google-compute-engine, but none of the providers can be installed [05:56] nothing provides libjson-c.so.2()(64bit) needed by google-compute-engine-oslogin-1.3.0-1.el7.x86_64 [05:56] and we cannot use fedora 28 without that package? [05:57] zyga, not in google [05:57] uh [05:57] it is required [05:57] Son_Goku: hey, do you know if f28 support in google compute engine is something that fedora is doing or is that google themselves? [05:58] that's probably something the fedora server WG would know [05:58] but I would not be shocked if it's a bit of both [05:58] but libjson-c.so.2()(64bit) is missing, which means json-c was upgraded [05:59] cachio_, where do you get gce-oslogin? [05:59] Son_Goku, so far from here https://packages.cloud.google.com/yum/repos/google-cloud-compute-el7-x86_64 [06:00] then f28 images are definitely all google [06:00] for fedora 26 and 27 [06:00] because fedora would have required these packages to be available in fedora itself [06:00] yeah, that's what I thought too [06:00] PR snapd#5152 opened: tests: move fedora 27 to manual [06:00] cachio_: could we do the same as for arch? [06:01] cachio_, it's definitely fixable though [06:01] it'd be trivial to make a compat package for it to work [06:01] mborzecki, for arch I download the image which already has that package [06:02] mborzecki, there is not any fedora image for gce afaik [06:02] cachio_: i mean build the google services form source for the time being [06:02] mborzecki, well yes [06:02] I mean, the spec files for building the gce stuff is available [06:02] so we could just _build_ them to Fedora [06:03] copr :) [06:03] Son_Goku, perfet [06:03] https://github.com/GoogleCloudPlatform/compute-image-packages/blob/master/google_compute_engine_oslogin/packaging/rpmbuild/SPECS/google-compute-engine-oslogin.spec [06:04] the packaging is actually slightly wrong :( [06:06] * Son_Goku fixes [06:12] * cachio_ EOD === cachio_ is now known as cachio [06:16] cachio, zyga, I need sleep so I'll deal with it in the morning [06:16] Son_Goku: ack, good night! [06:16] thank you both [06:26] https://www.reddit.com/r/linux/comments/8icx4w/gimp_20_flatpak_and_snap_side_by_side/ [06:44] still at 50 PRs [06:47] hello, is there a way to edit an apparmor profile and reload the profiles from a hook? [06:48] I am probably thinking it wrong.. [06:48] I should be creating an interface or something, right? [06:51] mborzecki: we should be able to improve the file pickers once portals lands [06:52] jamesh: looking forward to it ;) [06:53] that's assuming Gimp uses the FileChooserNative APIs [06:53] which is not guaranteed === pstolowski|afk is now known as pstolowski [07:08] morning [07:09] pstolowski: hey [07:12] afk for ~2h, need to go to the car shop, start the claims process [08:12] *yawn* [08:13] hey john [08:13] I was thinking [08:13] about that seeding issue [08:13] it would be interesting if snapd could model being online or offline [08:13] because then we could be smarter about many decisions [08:13] e.g.: refreshing snap foo will make us offline temporarily [08:13] or stopping it will make us offline [08:14] or we are offline/online and a system package (not snap) is managing that [08:15] zyga: i'm wary of adding dimensions to our state [08:16] zyga: also, http://www.linusakesson.net/programming/pipelogic/index.php [08:16] ogra_: you might enjoy it too ^ :-) [08:24] woah [09:24] zyga: I think https://github.com/snapcore/snapd/pull/5115 should be complete now. Everything is passing except the fedora 27 issue (which in theory is down to load caused by the fedora 28 release) [09:24] PR #5115: interfaces: add xdg-document-portal support to desktop interface [09:24] ack [09:24] * zyga looks [09:25] zyga: I've also put together some instructions on testing the current state of the code at https://forum.snapcraft.io/t/snapd-support-for-xdg-desktop-portal/161/9?u=jamesh [09:26] one thing that may be controversial is getting "snap run" to try and launch various D-Bus session services before transferring control to snap-confine [09:26] we can't very well rely on service activation for the mount point. [09:27] jamesh: what normally starts the document portal? [09:27] is it run by the session [09:27] or is that run by snap/flatpak as needed [09:28] zyga: it doesn't auto-start with the session on Ubuntu. I suspect flatpak just activates it via dbus. Let me check [09:29] thanks [09:31] jamesh: if snap run would need to start the document portal, how would it know [09:31] zyga: yep. It's relying on D-Bus service activation, doing a synchronous method call: https://github.com/flatpak/flatpak/blob/master/common/flatpak-run.c#L1682 [09:31] jamesh: the desktop interface in one of the plugs [09:31] jamesh: even if so, we don't know if it is connected [09:32] looking at that code the nice thing is that it doesn't seem to be specific to an app [09:32] so we could even start it from snapd userd === Guest68883 is now known as devil__ === devil__ is now known as devil_ [09:33] zyga: is that any better? Isn't userd a dbus activated service too, and not generally run until you use xdg-open? [09:34] userd has an auto-start feature now [09:34] though it still feels racy [09:34] my point is that I don't know what would be the condition that would make 'snap run' start the portal [09:34] re [09:35] I think at some point we will want a long running session service, and niemeyer seemed to think userd shouldn't do that [09:35] and if we had that? [09:35] I'm still trying to piece together how it is expected to work in the end [09:35] (i.e. he wanted userd to be able to restart at any point) [09:36] if "snap run" is going to ensure userd is running, then having userd try to activate xdg-document-portal should be fine [09:37] how would that work in a headless system without sessions [09:38] if there is no session bus, then there's no portals [09:38] and we'd continue without it [09:39] The safe scenarios are (1) document-portal starts before confined app, and (2) document-portal is never started while confined app is running [09:39] unsafe is "document-portal started while confined app is running" [09:40] For flatpak, this last scenario is safe because the entire /run/user/$uid tree is private [09:48] ck [09:48] ack [09:50] when i try to execute snapcraft prime or snapcraft pull or snapcraft build , we are getting the error message in Linux mint 18.1 64 bit "Native builds aren't supported on Linux Mint. You can however use 'snapcraft cleanbuild' with a container". [09:52] newbee: that's right [09:54] @zyga : also getting error for snapcraft cleanbuild "The container you are starting doesn't have any network attached to it." what it means.. [09:54] newbee: it means the container doesn't have a network interface (probably) [09:54] perhaps kalikiana can help you, I'm not an expert on snapcraft [09:56] something wrong with local lxc/lxd installation? [09:57] eg. on arch you have to tweak the configuration shipped with distro package [09:57] otherwise the containers that are started have no network interfaces [10:00] @mborzecki / @zyga : please tell us, can i run snapcraft in linux mint.. also give a clue to configure the network interface to lxc [10:01] newbee: I don't know how to setup lxc in linux mint [10:01] you can run snapcraft (I suspect) but I don't know how to set it up for you [10:33] Chipaca, mosfets !!! now it just needs to play music ! [10:33] (lovely article indeed) [10:34] PR snapd#5148 closed: boot: clear "snap_mode" when needed (2.32) [10:43] PR snapd#5150 closed: snapd.core-fixup.sh: add workaround for corrupted uboot.env (2.32) [10:46] hello there [10:46] would anyone accept a bribe to get https://github.com/snapcore/snapcraft/pull/2119 merged? [10:46] PR snapcraft#2119: repo: automatically prune unneeded stage-packages [10:48] thresh: I hear kyrofa likes craft beer [10:50] I doubt he has powers to do self-review and merges :-) [10:51] thresh: sergiusens might be needing a new aikido outfit, a slimmer one\ [10:52] Chipaca, thresh: so on that one, mvo was going to review this week, but apparently some issues showed up and he got put on the hook for them [10:53] these resolver ones can have tricky side effects not viewable at plain sight and his input is really welcomed on it [10:53] we can ship one of these nice costumes: https://pbs.twimg.com/media/DVMNuAoX0AU7gKS.jpg:large [10:54] thresh: mvo would rock hockey in one of those [10:54] sergiusens, ah, good to know. It kinda looked abandoned since no reviewers set etc, so I thought to nudge here. Thanks! [10:54] thresh: (mvo spent most of his week hunting down a bug in the interplay of uboot and linux's FAT implementations [10:54] ) [10:55] the horrors [10:55] IKR [10:56] sergiusens: yeah, this week was slightly bad, just go ahead with the PR it looks good and I can do a proper indepth PR next week when things are a bit more calm [10:57] mvo: why are you here at all! :-) [10:57] mvo: get out of here [10:57] mvo: shoo! [10:58] sergiusens: because this update is not released yet that we promised for tihs week [10:58] Chipaca: yeah, I know :/ [10:58] oh, :-( [10:59] mvo: what's missing? [11:00] Chipaca: one test is still running, once that is in I will do .7 and run away [11:00] Chipaca: i.e. 2.32.7, sru to bionic (that is what we promised) and push to beta for screenly to test [11:05] thresh: btw, I am a happy user of the vlc snap, everyday user here :-) [11:07] sweet :-) [11:08] PR snapd#5149 closed: many: add wait command and seeded target (2.32) [11:51] * cachio afk === pstolowski is now known as pstolowski|lunch [11:57] mmm, good idea [11:57] mvo: ppc failed on tests [11:57] https://launchpadlibrarian.net/369787126/buildlog_ubuntu-xenial-powerpc.snapd_2.32.7_BUILDING.txt.gz [11:58] FAIL: devicestate_test.go:478: TestFullDeviceRegistrationHappyClassicFallback.pN66_github_com_snapcore_snapd_overlord_devicestate_test.deviceMgrSuite [12:14] zyga: hm, I think this is just flaky on gccgo [12:14] ack [12:14] is it an issue for the release? [12:15] zyga: this one is not (if its transient) but https://github.com/snapcore/snapd/pull/5147/commits/f73fa6f48687aeae69a26c95951b72fcc9404e03 is [12:15] PR #5147: snapd.core-fixup.sh: add workaround for corrupted uboot.env [12:16] zyga: 5147 has a proper test now (yay) but it shows an issue with the fixup (meh) so I think I need .8 :/ [12:16] :-( [12:16] well [12:16] zyga: yeah [12:16] * zyga hugs mvo [12:16] zyga: could be worse, at least we have a real test and real fix this way :) [12:18] zyga: a re-review of that pr would be great, tests are running locally (and in spread) right now [12:19] looking [12:22] zyga: fwiw, the binary file with the partition image is 200kb which I think is acceptable given the importance of the bug [12:25] mvo: done [12:26] PR snapd#5153 opened: snapd.core-fixup.sh: add workaround for corrupted uboot.env (2.32) [12:26] zyga: \o/ thank you [12:26] mvo: i was looking into writing something to generate that 'semi-corrupted' vfat in python [12:27] Mornings [12:27] pstolowski|lunch: Replied on #5120.. btw, the PR summary isn't following the usual pattern [12:27] PR #5120: interfaces: interface hooks for refresh [12:27] mborzecki: yeah, I think that would be good, in the meantime I took the original corrupted image, cleaned it, zeroed all unused blocks and this means its compressed small enough to put into the tree === pstolowski|lunch is now known as pstolowski [12:27] pstolowski|lunch: I've slightly adjusted it, but still needs further tweaking to reflect what's in the PR [12:27] niemeyer: thanks, looking [12:28] mvo: as for the partition image, yeah, no doubt it is useful [12:38] niemeyer: updated the summary. re the methods with identicals args, do you want me to change to a single method? [12:40] I need to go to the school. Will miss the standup today.... ttfn [12:46] bumped arch package to 2.32.7 [12:47] mborzecki: .8 is coming :( [12:47] mborzecki: I will update opensuse after lunch/standup [13:02] I'm also going to be a bit late as I have a conflicting meeting today [13:02] (in theory, low on attendance atm) [13:04] Hi [13:04] AnaValencia: hey [13:22] niemeyer: we all went (including mvo) so I think that's it for today [13:22] niemeyer: apart from .8 release nothing major to report [13:23] niemeyer, zyga: yeah, just waiting for tests on this one [13:23] zyga, mvo: Cool, sorry for being late.. should have rescheduled it [13:37] * zyga needs to run an errand for CE, [13:50] * ogra_ grins about the forum ... [13:51] jdstrand, you need to implement an "open-giant-security-hole" interface that bind-mounts /usr/lib/flashplugin-installer into browser snaps ;) [14:01] PR snapd#5153 closed: snapd.core-fixup.sh: add workaround for corrupted uboot.env (2.32) [14:08] https://forum.snapcraft.io/t/feature-proposal-source-submodules-option/5372 [14:14] ey jdstrand, you are spot on the kill signal recieve issue of docker containers. I couldn't find a way to provide my own apparmor profile for each container the dockerd spawns, however I could add a single line in the docker-default apparmor profile and reload the profiles. This does not feel right though. Do you think the only option we have is to build our own dockerd and ship it? [14:14] *hey [14:18] PR snapd#5147 closed: snapd.core-fixup.sh: add workaround for corrupted uboot.env [14:19] https://forum.snapcraft.io/t/bug-docs-should-specify-the-default-behavior-when-stage-key-is-absent/5384 [14:19] PR snapcraft#2130 opened: tests: remove obsolete env var [14:20] PR snapd#5154 opened: releases: merge 2.32.8 back into master [14:22] PR snapcraft#2129 closed: Swap FROM in Dockerfiles from 'xenial' to 'bionic' [14:37] another store error on travis.. cannot get nonce from store: store server returned status 418 [14:49] PR snapcraft#2131 opened: No user site for snapcraft [14:53] cachio: could you please sru test 2.32.8 for bionic only? it contians an important customer fix [14:54] cachio: and please ping apw on monday and ask for the release of .5 into -updates so that we can get .8 into proposed on all !bionic series :) thank you! [14:59] zyga, snap run --shell doesn't seem to take stdin. Is there any way I can use it in a script? [15:04] Wait I lied, shell script fail [15:04] cachio: also 2.32.8 is in the beta channel now ready for validation [15:04] re [15:05] * zyga delivered edge gateway to a colleague, needs water and will be hacking soon [15:06] zyga, ubiquity? Did you switch to something else? [15:06] kyrofa: hmm? :) [15:06] kyrofa: I meant that I returned a borrowed dell edge gatweay [15:06] kyrofa: to a colleague that lives nearby [15:06] Ahh [15:07] Too much similar-sounding network hardware :P [15:07] it's so hot today I could drink a river [15:07] * ogra_ de-routes the rhine to warsaw to zyga's house [15:08] * kyrofa is impressed with ogra_'s `route` foo [15:08] route add ... blah [15:08] :) [15:29] * zyga breaks for some time [15:33] currently have this for manipulating FAT entries: https://gist.github.com/bboozzoo/e68254507eef4673dd8c6f9b82f65d92 [15:36] PR snapd#4588 closed: Snapshots! [15:38] niemeyer: would you like me to tweak these Reconnect/Autoconnect functions in #5120 or can I merge? [15:38] PR #5120: interfaces: interface hooks for refresh [15:38] Nice :-) === pstolowski is now known as pstolowski|afk [16:20] pstolowski|afk: Feel free to merge either way as this is a trivial point and easy to fix afterwards, but you do have two functions there which have exactly the same parameters and do exactly the same thing, apparently for no reason.. [16:20] pstolowski|afk: No pressing need, but this is really a single function [16:37] niemeyer: yes I get your point. the reason though is that I couldn't come up with a good single name. but i'm ok to change this in a followup [16:39] PR snapd#5120 closed: interfaces: interface hooks for refresh === alan_g is now known as alan_g|EOW [19:06] zyga, cachio: https://copr.fedorainfracloud.org/coprs/ngompa/gce-oslogin/build/752832/ [19:06] zyga, since I am waiting tor my flight boarding, I figured I'd just take care of this ;) [19:07] it covers F27, F28, and Rawhide (targeting F29) [19:14] Thank you :-) [19:53] zyga, cachio: built successfully: https://copr.fedorainfracloud.org/coprs/ngompa/gce-oslogin/build/752841/ [20:04] PR snapcraft#1769 closed: lxd: add an --image argument to cleanbuild [20:04] PR snapcraft#1969 closed: Add a "--profile" parameter to cleanbuild [21:01] PR snapcraft#2132 opened: errors: generic exception for common.run[_output] [21:31] PR snapcraft#2131 closed: No user site for snapcraft [23:44] Issue snapcraft#2133 opened: gradle plugin treats spaces as delimiter in gradle-options