=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [07:08] PR snapd#2807 opened: snap: add new `snap switch` command [08:17] PR snapd#3607 opened: cmd: fix re-exec bug when starting from snapd 2.21 [08:28] PR snapd#3608 opened: cmd: rework reexec detection [08:35] zyga-ubuntu: the crash: http://pastebin.ubuntu.com/25131529/ [08:46] Saviq: thank you [09:35] PR snapcraft#1413 closed: core: minimal windows support === jospoortvliet_ is now known as jospoortvliet [10:21] PR snapd#3609 opened: Introduce the kvm interface [10:35] PR snapd#3610 opened: snap: do not always quote the snap info summary [11:06] * Chipaca in a warren of undoes [11:08] PR snapcraft#1417 opened: Move cross-compiling from kernel to kbuild plugin [11:36] Chipaca: hey, can you have a look at a small pr for cmd.go please? [11:36] https://github.com/snapcore/snapd/pull/3608 [11:36] PR snapd#3608: cmd: rework reexec detection === chihchun is now known as chihchun_afk [11:42] zyga-ubuntu: sure [11:43] zyga-ubuntu: how will this work in places where it's not /snap? [11:49] https://lwn.net/Articles/698073/ [11:49] Chipaca: it's not yet, we're aware [11:49] Chipaca: but that's for future [11:52] PR snapd#3607 closed: cmd: fix re-exec bug when starting from snapd 2.21 [12:00] pedronis: how're you feeling today, vis-a-vis being asked for reviews? [12:04] Chipaca: I'm ok [12:09] pedronis: when you have a slot, I'd appreciate a look at snapd#3600 [12:09] PR snapd#3600: many: expose service status in 'snap info' [12:11] meanwhile I'm fixing a bug in snapstate's copydata thing [12:18] lip 20 14:17:55 fyke kernel: audit: type=1400 audit(1500553075.963:253): apparmor="DENIED" operation="capable" profile="/snap/core/2445/usr/lib/snapd/snap-confine" pid=10030 comm="snap-confine" capability=4 capname="fsetid" [13:12] PR snapd#3611 opened: overlord/snapstate/backend: some copydata improvements [13:22] Chipaca: done [13:29] pedronis: thanks! [13:36] pedronis: I'll address your comments right now (as i just finished the copydata pr) [13:36] "just" == "just now before reading your comments and making me some tea" [13:42] Bug #1705486 opened: SPI not working on Raspberry Pi 2 with ubuntu core [13:55] zyga-ubuntu: github.com/rogpeppe/govers [13:58] thanks! [13:59] Chipaca: to clarify, I thing you were trying to solve that client and deamon API are the same, so why two struct, but now you are positing client == daemon == systemd, which I'm not sure about [14:02] fair 'nuf [14:02] thing is [14:02] bah, forward-thinking to the actual services work [14:02] not sure this distinction will continue to be as clear, but ok [14:03] which distinction? [14:03] about systemd? [14:03] the package [14:03] bah [14:03] i'm probably hung up on details of the past implementation [14:03] PR snapd#3606 closed: cmd: fix broken double re-executiong from snapd 2.21 [14:03] PR snapd#3608 closed: cmd: rework reexec detection [14:04] pedronis: you're right [14:04] there's no reason this needs to be like this [14:04] * Chipaca fixes [14:05] I understand that having two structs is more annoying than 1, but reading client in the signatures of systemd is really strange, a bit unclear if the relationship will hold [14:05] but I agree we don't want/need 3 structs for this [14:05] niemeyer: we don't have to switch, upstream just merged the fix :) [14:13] PR snapd#3612 opened: vendor: update go-flags to address crash in "snap debug" [14:19] Chipaca: FYI https://forum.snapcraft.io/t/snap-remove-doesnt-remove-data-from-root-snap-snap-name/1387 [14:19] maybe true, maybe red herring [14:19] maybe red dwarf === chihchun_afk is now known as chihchun [14:28] PR snapd#3525 closed: interfaces: add password-manager-service implicit classic interface (LP: #1653769) [14:30] zyga-ubuntu: Dude. [14:30] yes? [14:31] zyga-ubuntu, I have updated PR 3604 to make seccomp work for opensuse as you requested [14:31] zyga-ubuntu, could you take a look when you have some time? [14:31] yes, looking [14:33] tx [14:34] cachio: commented, === chihchun is now known as chihchun_afk [14:36] zyga-ubuntu, i see this error [14:36] https://paste.ubuntu.com/25133113/ [14:37] when LDFLAGS contain a static link to seccomp [14:38] zyga-ubuntu, any idea how to deal with it? [14:42] PR snapd#3548 closed: interfaces: Add /run/uuid/request to openvswitch [14:51] cachio: looking [14:51] cachio: are we installing a package that ships libseccomp.a? [14:51] zyga-ubuntu: wrt "dude", commented in the forum [14:51] Chipaca: aha, thank you [14:52] :D [14:52] thanks! [14:52] so pawel reported it too :D [14:52] it must have been him then [14:52] * zyga-ubuntu has rusty memory [15:05] zyga-ubuntu, not sure but in snap-seccomp/main.go we are linking to it [15:10] in opensuse libseccomp.a is not included in any package that we install [15:11] Hey jdstrand, you wanted to ping you about that classic snap I'm playing with [15:11] zyga-ubuntu, but we need to link statically to libseccomp [15:12] I'm not awake yet and thus can't type [15:12] zyga-ubuntu, at least there is a test for that [15:13] jdstrand, it contains libraries for building its clients, and a tool for running several clients in a specific configuration based on a text file [15:13] It looks for those clients in the PATH [15:14] And even if the PATH is accessible in confinement, snap-confine strips it off [15:14] So if the snap is strictly confined, it can't find anything to orchestrate in e.g. the home dir [15:14] Thus classic [15:16] zyga-ubuntu, why we need to link statically to libseccomp in ubuntu? [15:18] cachio: because of reexec [15:19] pedronis, to make sure both are using the same one? [15:19] yes (though we don't do that with libudev because there the libudev<->kernel part is the more fragile one, but is not ideal) [15:19] ok [15:20] pedronis, thanks [15:36] zyga-ubuntu: not only it was reported, but we discussed not to touch it until we have snapshots, at least that was the conclusion in London [15:45] jdstrand, moos-kyrofa is now in the review queue [16:33] pedronis: fixed snapd#3600 as per your review; thanks! [16:33] PR snapd#3600: many: expose service status in 'snap info' [16:40] PR snapd#3596 closed: tests: disable snapd-notify for the external backend [16:43] PR snapd#3503 closed: tests: add browser-support interface test [16:54] Chipaca: generally +1 with a comment about expected in the parser [16:58] pedronis: gotcha. Yeah, it's a little convoluted as is [16:59] pedronis: by having expected be an array, you mean then doing strutil.ListContains to check for unexpected stuff? [16:59] bah, no real need, the switch handles that [17:00] hmmm [20:17] PR snapd#3600 closed: many: expose service status in 'snap info'