=== the-mentor7 is now known as the-mentor === oerheks1 is now known as oerheks [07:04] good morning [07:14] morning [07:14] chilly morning, -25C [08:07] morning [08:15] mvo: pstolowski: hey [08:15] good morning mborzecki and pstolowski [08:16] pstolowski: was it equally cold in your area in the morning? [08:16] mborzecki: -18C atm [08:16] is that cold enough? ;) [08:20] pstolowski: haha ;) nice, was -25C in the morning, couldn't open my car [08:21] oh wow [08:22] fwtw i thinking it's going to be >0 again this week [08:22] *i think [08:23] PR snapd#9777 closed: gadget: add gadget.ResolveContentPaths() [08:24] pstolowski: yeah, back to normal :) [08:24] mvo: do you have a log with failure of https://github.com/snapcore/snapd/pull/9844/commits/a0bda4cce16da09fed44d70c1eaa5e56516024fa ? [08:24] PR #9844: cmd: make string/error code more robust against errno leaking [08:24] mborzecki: yes, need to dig but should be able to find it [08:25] mborzecki: https://buildd.debian.org/status/fetch.php?pkg=snapd&arch=i386&ver=2.48.2-2&stamp=1610734191&raw=0 [08:26] mborzecki: does what I write in my PR make sense? [08:26] mvo: thanks, let me see [08:26] mborzecki: pretty sure it's test related stuff leaking but it highlights our code is buggy, a strquote will never have errno set so we need to clear it [08:27] mborzecki: but yeah, thanks for looking into this, I'm not sure what exactly in the tests is causing this leaking [08:27] mborzecki: also a bit hard to reproduce, I had to setup a i386 sid vm [08:27] mborzecki: but there it failed without sbuild irrc, just by running latest gcc/i386 [08:27] mvo: it built on amd64 though? [08:28] mborzecki: correct [08:29] mborzecki: here is the full build status https://buildd.debian.org/status/package.php?p=snapd&suite=sid [08:29] mborzecki: it also has issues on mips it seems [08:29] mvo: and riscv go does not support pie [08:29] mborzecki: but afaict we don't have to worry about mips* [08:30] yeah, unlikely one is going to run snapd on their mips router :P [08:32] mborzecki: +1 [08:56] mvo: hm, so i'm trying debian i386 in podman which i upgraded to sid [09:03] mborzecki: and no luck with the error? [09:07] heh, network is down in the container after updating to sid ;) [09:23] PR snapd#9845 opened: osutil: update go-udev package [09:24] woot, cla-check fails with PermissionError: [Errno 13] Permission denied: '/home/ubuntu/.launchpadlib' [09:25] i wonder if it's going to like subtree of other repository nowadays... it wasn't an issue 2 years ago [09:35] mborzecki, pstolowski, degville I accidently deleted the standup today, feel free to still have it from tomorrow on everything in the calendar is normal [09:35] mvo: ack [09:37] mvo: i can't reproduce the build failure in a container [09:37] mvo: all unit tests are passing [09:37] oh w8, but silly me tried the master branch [09:39] unit tests are happy on 2.48 too, what gives? [10:16] heh, what's not to hate about implicit errno [10:18] zyga: hi, you like C, do you have any ideas about improving the errno mess in https://github.com/snapcore/snapd/pull/9844 ? [10:18] PR #9844: cmd: make string/error code more robust against errno leaking [10:19] mvo: were you able to reproduce it in a vm too? [10:22] mborzecki: I think I was, it was friday evening, it's a bit of a blur but I can reboot this vm [10:23] mborzecki: trying it now, should have results in a few min [10:24] mborzecki: worst case is that it's sbuild (which definitely does a lot of strange things) [10:28] mborzecki: yes, i386 qemu based vm is enough for me to reproduce [10:28] mborzecki: do you need anything? gcc ver? kernel ver? [10:28] mborzecki: I don't think all sid updates are applied I could try with that [10:29] mborzecki, I'll get back to you in a moment [10:30] mborzecki, interesting [10:31] mborzecki, I have some ideas, I did run into this earlier but I didn't fix there [10:32] mvo: so the only difference is the kernel then, the continer is using my host's kernel, while you have the kernel from sid [10:33] mborzecki, my feeling around the _right_ solution is to drift away from global errno [10:33] mborzecki, deprecate die() that uses implicit errno [10:34] mborzecki, add die_just_because and die_because_errno [10:34] mborzecki, then those no longer have the problem that we are papering over [10:34] mborzecki: I was using 4.19, let me try a newer one [10:35] mvo: sid on 4.19? [10:35] mborzecki, then we go one by one and convert those to the new API [10:35] mborzecki: ha! fun, yes, with 5.10 I don't get the error [10:35] mborzecki, this is also nicely in line with the trend to avoid using global errno and start using errno _values_ but not the variable [10:36] zyga: yeah, thought about just adding a parameter to die(), but maybe a new die_() would be better [10:36] mborzecki, I think it's better to be explicit, especially that in all the cases I can think of, it will be clear that we want to die_with_errno() and pass the errno [10:36] or die just because but ignore errno, even if there is some non-zero value there [10:36] how does that sound to you? [10:38] sound lines something we should have done a while ago :) maybe now's the time [10:40] mborzecki, I think it was always strangled by the feeling that touching C is bad and reviews never materialized [10:40] mborzecki, if that is not true anymore I'd love to see some progress on this [10:40] mborzecki, and I can help with review or coding if you want [10:41] CC mvo ^ [11:19] PR snapd#9846 opened: [RFC] systemd, many: no reload when enabling or disabling services [11:32] silly gofmt [11:39] pedronis: hi, i'm unclear about max-format=.. argument for v2/assertions/validation-sets/...; it's not mentioned in the spec so probably not applicable there? or is it something implicitly supported by all assertion endpoints, and should be passed too? [11:43] pstolowski, IIRC max-format was for the case where an old device wakes up and talks to the store and is not confused by assertion format it cannot parse [11:43] pstolowski, so it was meant to be for all assertions [11:44] but I'm not sure if that's true [11:44] PR snapd#9847 opened: cgroup-support.c: fix link to CGROUP DELEGATION [11:46] ijohnson, hey [11:46] hey zyga [12:09] PR snapd#9848 opened: overlord/devicestate, sysconfig: do nothing when cloud-init is not present === pedronis_ is now known as pedronis [12:14] PR snapd#9838 closed: asserts: sort by revision with Sort interface [12:15] pstolowski: it should be supported everywhere, double check with suligap [12:15] pedronis: sure, thanks [12:23] ijohnson: thanks for the review https://github.com/snapcore/snapd/pull/9848 [12:23] PR #9848: overlord/devicestate, sysconfig: do nothing when cloud-init is not present [12:23] mborzecki: you're welcome :-) [12:45] ijohnson: is this more less what you had in mind? https://paste.ubuntu.com/p/dvGMDwT832/ [12:45] mborzecki: yes thank you [12:45] cool, thanks for taking a look [13:15] cmatsuoka: hey do you still have that spread-runner worker online? it doesn't seem to be running cla check properly, anymore: https://pastebin.ubuntu.com/p/pngjxpRWbz/ [14:34] ijohnson: i retried #9845 5 times and it fails every time on cla check, do you think i should keep trying some more? [14:34] PR #9845: osutil: update go-udev package [14:34] pstolowski: I tried like 3 times and it worked, but check the very start of the raw log from the github runner to see which runner got allocated [14:34] ijohnson: right, good idea [14:35] cause afaik it's random as to which runners get allocated [14:35] ijohnson: Machine name: 'claudio-spread-1'.... [14:35] yeah that's the broken one right now [14:36] unfortunately it seems it's also the one that really wants to run over my PR right now ;) [14:37] haha it's very attached [14:40] ok, it picked a different one finally.. but failed checking one of the contributors of go-udev upstream :( [14:45] PR snapd#9849 opened: tests/main/snap-network-errors: disable dns-caching on centos-7 [14:45] :-( [14:49] not sure what to do about that... it doesn't make sense to chase radom people of 3rd party project about signing cla for us, obviously [14:49] i could --squash, but that sucks [14:51] pstolowski, ijohnson sorry a bit distracted but should I remove a runner (claudios?) [14:51] mvo: yeah if you have perms to do so, that would make cla-checks less likely to fail until cmatsuoka can fix the runner [14:51] +1 [14:57] forced remove it [14:58] ty [14:59] is there a py-lib to use snap from within python ? haven't found one yet, just looking at the wrong spots? [15:00] dariball, use in which sense? [15:00] dariball, drive snap operations or manage a snap system? [15:01] I would like to check if there is a new version of a snap, refresh it, restart, stop [15:10] dariball, snapd is doing that work internally [15:11] dariball, there are REST apis for that as well, those can be scripted [15:11] but they only complement the internal refresh check [15:11] dariball, there are Glib bindings that I know of [15:11] dariball, I don't know if there are python bindings yet [15:13] oh, rest api sounds very good, just found it via unix-socket, perfect! thank you this is what I was looking for ... [15:18] dariball: https://snapcraft.io/docs/snapd-api may help [15:20] jap exactly what I was looking for, nice [15:45] PR snapd#9815 closed: {,sec}boot: pass "key-name" to the FDE hooks [16:49] mvo: hey, would it make sense for me to be a collaborator on the snappy-debug snap? I was dropped from snappy-dev so can't commit to it any more. I can do MPs and hand off to emitorino. honestly, it would probably be best if it was converted to git and moved to github under the snapcore umbrella (perhaps with a rename to snap-debug). that all requires various bits of work. I'd be happy to continue to [16:49] contribute to the project though [16:50] mvo: let me know how you'd like me to do that [16:54] jdstrand: in a meeting but I want to absolutely add you as collaborator, can you please /msg me the email I should use for this? [16:55] mvo: so there is both collaborating on the snap and committing to the tree [17:00] PR snapcraft#3420 opened: plainbox spread tests: set tasks to manual [17:02] mvo: in other news, it looks like I am still part of the snapcore and ubuntu-core organizations in github. I think I'll leave that as is for now unless you want me to change it [17:05] PR snapd#9850 opened: store: method for fetching validation set assertion

[17:06] pedronis_: ^ [17:06] pedronis_: some dillemas here re primary key and potential reuse (after tweaks) of asserts.HeadersFromPrimaryKey [17:07] and it's based on the PR that switches to v2 api [17:09] https://twitter.com/zygoon/status/1351215032475004942/photo/1 :) [17:09] niemeyer, ^ [17:09] hey jdstrand [17:10] * zyga hugs jdstrand for being here [17:11] hey zyga :) [17:21] mvo, is cachio off / sprinting this week? [17:45] zyga: cachio is moving to a new house this week, I think he will be back on like wed or thurs [17:47] ijohnson, oh, that's quite a big change [17:47] cool, thanks for letting me know [17:47] np [17:47] btw nice logo for spread, I like it [17:52] ijohnson, :) [17:53] I like it because it's so logo-like it's just lines spreaing outwards === ijohnson is now known as ijohnsonlunch === ijohnsonlunch is now known as ijohnson|lunch === ijohnson|lunch is now known as ijohnson === tianon is now known as tianon-- === tianon- is now known as tianon