[05:20] tsimonq2, yes, you should be able to use it to authenticate yourself when performing git operations over HTTPS [07:42] tsimonq2: Yeah, I suppose in principle it doesn't _have_ to be tied to the username since tokens are unique, but it seems clearer that way and why not [15:25] Hello! I am trying to add my SSH public key to my Launchpad profile, but it keeps saying "invalid key". My public key is sk-ecdsa-sha2-nistp256@openssh.com, and the private key is on a Yubikey (so it's not easy for me to re-generate it). What could be the issue with my SSH key? [15:26] The Launchpad says "Only SSH v2 keys are supported", but I am not sure exactly what it means... is the key format I have unsupported? [15:40] We don't support sk-* keys yet - they'll require upstream work in Twisted first [15:43] I thought I'd mentioned this in https://www.chiark.greenend.org.uk/~cjwatson/blog/lp-new-ssh-features.html but I seem to have forgotten. But that should give a sense of the amount of work involved [15:45] (And maybe a pointer to where somebody who wanted to contribute this sort of thing could look) [18:36] Hey, has anyone else noticed that something seems to be seriously awry with the Nextcloud arm builds on Soyuz? It looks like some of them have been running for a couple days already... [18:47] * RogueScholar slowly looks the nicklist up and down and feels a bit chagrined, "Dang, I didn't realize that it's like 80% FOSS heavy hitters around here. Please allow me to clarify that I had no intentions of sharpshooting anyone just now. I had assumed this was more of a user-centric space but now it's clear that probably everyone here is better informed than me and probably a half dozen or more are sitting within view of a screen showing the [18:48] ...live resource monitoring for the whole deal." 😯 [19:09] RogueScholar, there were some issues with the builder this week that we have resolved today. So I'd recommend cancelling stuck builds and retrying them. [19:09] *builders [19:10] guruprasad: Thanks for the insight. None of mine are currently building, I'm afraid, just holding position in the queue. [19:20] I'm retrying this one. [19:20] And a dozen or so others. [19:24] Issues specifically: we tried to upgrade builder VMs to jammy, but it turned out that LXD 5.0 (which is the default in jammy cloud images) has some subtle issues that cause some builds to eventually hang after producing some quantity of output. === Saviq is now known as Saviq_ === Saviq_ is now known as Saviq [20:24] cjwatson: Ugh, as much as I love the sound of the builders finally moving off of Xenial, that sounds like a ugly one to try and troubleshoot. You know this week I've noticed that some of my build recipes (and I can't find a common thread yet as to what they have in common) have started erroring out over the version string, they won't build if there's a package revision. Is it possible that's all part of the same set of issues? [20:26] RogueScholar: moving off focal, not xenial [20:26] RogueScholar: But maybe? Link? [20:28] (builders moved off xenial in 2020 or so) [20:31] At a guess, that sounds more likely to be an independent problem possibly also prompted by switching the base system to jammy; if that's the case it ought to have gone away again now, since we've rolled back to focal. But I'd like to check. [20:31] Oh right, my bad. Well sassafras, I think I might have rewritten all the versions on the problematic recipes and the logs of the failures got pushed out of the queue, but I'm pretty sure this one was one of them [20:34] That one has had seven builds ever of which all appear to have been successful, unless I'm missing something [20:34] Oh, maybe https://launchpad.net/~appimagelauncher-team/+archive/ubuntu/daily/+build/26457484 [20:35] Failures without a build log like that are normally either your build killed the builder (unlikely for something like dh-python) or infrastructure issues [20:35] So I think that one is probably ignorable given that you've had more recent successful builds [20:39] Yeah, it must have been one of the other ones. Sorry I can't illustrate it better, but it was pretty clear that it was an issue with the version string parsing. I know the builds always get run in native package mode after quilt runs the patch series, but before this week it had always let the package revisions through. [20:41] Spoke too soon, just found one! [20:42] RogueScholar: Thanks, probably a behaviour change in dpkg-dev in jammy that we need to take account of. It should have gone away for now as per what I said above, but could you file a bug on https://bugs.launchpad.net/launchpad-buildd/+filebug please as a reminder for us to look into that? [20:43] With that link, of courase [20:43] *course [20:43] You got it, boss; can do. [20:44] That code in dpkg doesn't seem to have changed at all recently, but I only looked superficially [22:27] cjwatson: hmm I'm afraid it's not clear to me what would need to be done in Twisted Conch in order to support sk-* keys. First, I am not a cryptographer. Does that rule me out for that? I can write Python code though. Then... does it look like a good "first contribution", or does it look like something pretty involved? [23:58] 99% sure the build farm is stuck.