[00:08] PR snapcraft#3570 closed: ROS V2 plugins: resolve conditional dependencies [01:31] whenever i invoke snap, i'm getting main.go:176: description of prepare-image's "" is lowercase in locale "pt_BR": "o directório de destino" === Ringtailed_Fox is now known as RingtailedFox === RingtailedFox is now known as ItsNeverAWasteOf === ItsNeverAWasteOf is now known as Time === Time is now known as RingtailedFox === RingtailedFox is now known as ElIsWrong === ElIsWrong is now known as RingtailedFox === RingtailedFox is now known as ZnevnaIsATroll === ZnevnaIsATroll is now known as RingtailedFox [05:34] morning === RingtailedFox is now known as Ringtailed_Fox [05:47] mborzecki, miguelpires: hi! [05:48] hi mardy :) [05:49] golang question of the day: when importing a package, will the whole package be included in the final binary, or only those parts which are actually used? [05:53] mardy: if you're thinking about lto, then go compiler isn't that smart afaik [06:24] Good morning mvo. Can you merge https://github.com/snapcore/snapd/pull/10685 please? [06:24] PR #10685: many: fix run-checks gofmt check [06:30] PR snapd#10696 closed: many: move to yaml.v3 [06:32] morning [06:33] 'morning pstolowski, mvo [06:34] good morning miguelpires, sure, happy to do that [06:35] miguelpires: it's early for you, isn't it :) ? [06:35] good morning pstolowski [06:35] and good morning mardy too :) [06:36] mborzecki: I'm not sure how the LTO idea translates to golang, but let me ask a more concrete question: I've added a method here: https://github.com/mardy/snapd/blob/access-portal/usersession/client/client.go#L297-L332 [06:37] one possibility is to have this method be called by a separate process, not snapd itself [06:38] now, it would be a bit unlucky if this process, which does not need to know anything about systemd services and other snapd features, had to carry in its binary the code for serviceControlCall() and all of its depending bits [06:39] (though, as a matter of fact, it seems that this particular package does not have many deps in systemd, so in this case it might be quite an acceptable situation) [06:40] s/systemd/snapd/ [06:40] PR snapd#10685 closed: many: fix run-checks gofmt check [06:40] PR snapd#10698 opened: wrappers: fix a bunch of duplicated service definitions in tests [06:41] miguelpires: mvo: pstolowski: hey [06:42] hey mborzecki [06:42] mardy: i suppose easiest solution is to build a minimal program and inspect what symbols show up inside, otoh if it's that minimal than runtime and net/http will take up most of the size, + that magical line/symbol lookup table that go generates, i would not expect anything smaller than 3-4MB really :P [06:43] @mvo a bit, I look like =.= [06:43] hi mborzecki [06:48] mvo: https://github.com/snapcore/snapd/pull/10634 looks mostly happy [06:48] PR #10634: many: move to go modules gradually <⛔ Blocked> [06:51] mvo: shall we dropped the blocked label then? :) [06:52] mborzecki: \o/ !!! yes [06:52] miguelpires: haha [07:17] good morning :) [07:18] hey mardy, mborzecki, mvo and miguelpires :) [07:19] hey zyga-mbp [07:25] meh, the remodel code for generating tasks is pretty hairy now [07:25] good morning zyga-mbp [07:39] Hello snapd! [07:39] So I wanted to finally release snapd 2.51.1 into -updates for all the stable series [07:39] Sadly, focal constantly FTBFS on riscv64 [07:39] https://launchpad.net/ubuntu/+source/snapd/2.51.1+20.04/+build/21701922 [07:39] I think I already retried the build like 3 times [07:46] sil2100: hi, looks like a flaky test (managers_test.go:3522: mgrsSuite.TestUpdateManyWithAutoconnect - "Settle is not converging"), i'll see if I can find something or reproduce locally in a moment [08:00] PR snapd#10699 opened: daemon: allow enabling enforce mode

[08:10] heh and that case of new kernel or base being already installed during remodel was compeltely untested, fun [08:12] it seems 3s settle timeout in this test (scaled to 18s on riscv64) may not be enough (a few lines down we've another settle but with much greater timeout - 6*45s on riscv64) [08:14] i can bump it, but 2.51.1 is only tagged (not branched, but we have 2.51 branch covering up to .7), so not sure what to patch for? [08:14] mvo ^ ? [08:16] pstolowski: in a meeting, sry, will get back to you [08:23] pstolowski: hm i see all tags up to 2.51.7? [08:24] mborzecki: could you please set "security-high" on 10661 once the feedback from alberto/ian is adressed? I would love to get it (more) ready :) [08:25] mvo: will do, wanted to finish the uc20 remodel since it's very close now [08:30] mborzecki: yes, tags are there, but can a branch target a tag? [08:33] mborzecki: totally fine, sorry for being pushy [08:36] it doesn't seem possible to target a tag with a PR (which makes sense), so i guess we would need a proper 2.51.1 branch [08:41] pstolowski: let's talk in some minutes, I need a bit more context I think [08:45] PR snapd#10700 opened: tests: use host-scaled timeout to avoid riscv64 test failure [08:46] mvo: we need this tweak ^ (here: for master), but for 2.51.1 which FTBFS on riscv64 now, see above what sil2100 reported [08:50] PR snapd#10701 opened: many: remove unused parameters [08:51] pstolowski: I see, I can handle this, no worries, we will do a 2.51.1+20.04ubuntu1 or something equally nice outside for the git repo by just applying the patch [08:52] ok, great [09:29] stgraber: I'm getting this error when running `snapcraft --use-lxd`: An error occurred with the instance when trying to launch with 'LXD': Failed creating instance from image: Failed to run: zfs clone lxd/images/e076b3174006b985a6298feb03d74f565be43d1f3df21b530a31c8372f6c27be@readonly lxd/containers/snapcraft-snapd: cannot create 'lxd/containers/snapcraft-snapd': dataset already exists. [09:53] removing (with --purge) the lxd snap and reinstalling it helped [10:23] pstolowski: thanks! [10:24] In the meantime I'll try rebuilding the existing package again, maybe the flacky test will pass this time - since I'd like to unblock snapd releases ASAP [10:29] sil2100: i hope my fix solves it. the 3s timeout we had was very aggressive for riscv64, in other places it was 270s (!) with a comment saying that riscv64 builder is 5x slower than armhf [10:44] hmm gcc -fanalyzer does not like g_assert_nonull() and concludes it will not stop the program execution [10:49] and clang-analyzer isn't coping with the cleanup attribute too well? eh [11:04] pstolowski: wow, 3s? Yeah, that sounds like a bad idea for riscv64 machines that are generally quite slow [11:11] PR snapd#10702 opened: interfaces: introduce snap-refresh-control interface [11:18] woah, i didn't know we have "Missing high-level test for interface 'snap-refresh-control'. Please add to: ..." test check [11:19] nice touch [11:49] PR snapcraft#3577 opened: cli: add experimental --offline option (CRAFT-480) [12:12] what do you think of this way of mocking systemd? https://github.com/mardy/snapd/commit/0a524128abd3021f98e6ed1a66d1e1ecd26f0ea0 [12:26] PR snapd#10703 opened: tests/main/security-device-cgroups-strict-enforced: demonstrate device cgroup being enforced [12:42] feedback very welcome: https://github.com/snapcore/snapd/pull/10704 [12:42] PR #10704: systemd: add fake type for easier mocking of systemd functions [12:46] PR snapd#10704 opened: systemd: add fake type for easier mocking of systemd functions [13:09] PR 10634 now needs reviews, it seems to be ready finally [13:09] PR #10634: many: move to go modules gradually [13:18] hi [13:19] i am trying to setup up can adapter mcp2518fd on spi6 [13:19] so i wrote an overlay for it [13:19] but i cant find what the issue is in this [13:19] * Device tree overlay for mcp251x/mcp2518fd on spi0.0 [13:19] */ [13:19] compatible = "brcm,bcm2835", "brcm,bcm2836", "brcm,bcm2708", "brcm,bcm2709"; [13:19] /* disable spi-dev for spi3.0 */ [13:19] fragment@0{ [13:19] target = <&spidev3>; [13:19] __overlay__ { [13:19] status = "disabled"; [13:19] }; [13:19] }; [13:19] /* the interrupt pin of the can-controller */ [13:19] fragment@2 { [13:19] target = <&gpio>; [13:19] __overlay__ { [13:19] can_int_pins: can_int_pins { [13:19] brcm,pins = <6>; [13:19] brcm,function = <0>; /* input */ [13:19] }; [13:19] spi6_pins: spi6_pins { [13:19] brcm,pins = <19 20 21>; [13:20] target = <&spi6>; [13:20] __overlay__ { [13:20] /* needed to avoid dtc warning */ [13:20] #address-cells = <1>; [13:20] #size-cells = <0>; [13:20] cs-gpios = <&gpio 18 1>; [13:20] status = "okay"; [13:20] can0: can@0 { [13:20] reg = <0>; [13:20] pinctrl-names = "default"; [13:20] pinctrl-0 = <&spi6_pins &spi6_cs_pins>; [13:20] compatible = "microchip,mcp2518fd"; [13:20] spi-max-frequency = <20000000>; [13:20] interrupt-parent = <&gpio>; [13:20] interrupts = <6 8>; /* IRQ_TYPE_LEVEL_LOW */ [13:20] clocks = <&mcp2518fd_osc>; [13:20] }; [13:20] }; [13:22] later i want to set it up for ucore too, so first i am trying it on ubuntu server [13:56] PR snapd#10698 closed: wrappers: fix a bunch of duplicated service definitions in tests [13:56] PR snapd#10700 closed: tests: use host-scaled timeout to avoid riscv64 test failure [13:56] ares1925: this is a very detailed and low level question that you should probably ask elsewhere (i'm not sure how is it related to snapd and snaps but if you think it is, you may try to ask on the forum along with some more context). also, please don't paste huge snippets like this on irc, use pastebin.ubuntu.com and post links instead [14:01] ohk sorry [14:20] hm so the spread test is happy on a remodel branch commit just before i merged master :/ [14:20] and master passed too [14:21] PR snapd#8926 closed: interfaces: add microstack-support interface <:birthday:> [14:51] PR snapd#10705 opened: tests: add minimal smoke test for microstack [14:55] * mvo switches network, bbiab === itu is now known as Guest2168 [16:16] PR snapd#10524 closed: packaging: add libfuse3-dev build dependency === Ringtailed_Fox is now known as God === God is now known as RingtailedFox [18:02] PR snapd#10706 opened: tests: clean up test tools [18:25] PR snapcraft#3566 closed: ROS 2 plugins v2 misc fixes [19:35] PR snapcraft#3577 closed: cli: add experimental --offline option (CRAFT-480) === Guest2168 is now known as itu === itu is now known as Guest6671 === Guest6671 is now known as itu