| 8 May 2026 |
 emily | (I guess this is still kicking the userdbd can down the road though) | 11:13:48 |
| emily | @vcunat:matrix.org do you think you could ping me a few days in advance of the first post-branch-off staging cycle? I would like to coordinate around x86_64-darwin drop.
I think we should skip building x86_64-darwin starting on the first staging cycle and land the "no longer supported" error with it, because it will be bad UX if things stop being cached before the error (or if trying to override the error "seemingly just works" because we haven't had a world rebuild yet - and then goes kaput on the next staging cycle).
the basic support drop PRs are ~ready but there are some other important PRs on top I need to clean up and push out (e.g. update scripts that will break if x86_64-darwin dies).
| 11:18:25 |
| emily | (I'm not sure how close we are to that cycle because what's a release schedule 😅) | 11:18:55 |
| emily | relatedly, maybe <half the Darwin will help make this more feasible... | 11:19:32 |
 leona | and hopefully at some time in the future more builders 😅😉 | 11:20:11 |
| emily | something something queue runner? | 11:20:23 |
| emily | a pony? | 11:20:26 |
| emily | feel like it wasn't that long ago cycles took half the time | 11:20:48 |
| emily | (zstd when??) | 11:21:04 |
 Alyssa Ross | I might like a gnutls rebuild FYI | 11:23:21 |
| Alyssa Ross | I guess it's a bit late... that's probably fine — I'll just send it to staging | 11:52:14 |
 winston | checking this out atm, finally got a full gnome build on staging-next :)
if the PR works its still much preferable to reverting that gdm-user commit I linked | 12:00:42 |
| winston | gnome-shell doesn't build right now: https://github.com/NixOS/nixpkgs/pull/518038 | 12:01:14 |
| emily | Redacted or Malformed Event | 12:03:35 |
| emily | even GNOME is just spawning argvs rather than going through DBus/systemd units? 😅 | 12:04:03 |
| emily | uh, are all my messages sending twice or is Element X just being weird | 12:04:17 |
| winston | deleted the first, then sent a second one | 12:04:52 |
 jappie | I saw one message, then it got deleted / removed?, and then a second one appeared ~30s later | 12:04:53 |
| winston | stuff like unzip apparently, yea, but this change is removing one such subprocess spawn | 12:05:43 |
| emily | yeah it appeared twice for me, I deleted one, it deleted both, sent it again, appeared twice again, restarted the app, went back to normal. Matrix! | 12:07:08 |
| emily | do we need to do something to ensure Tecla gets hooked up in the session so that the spawn works? | 12:08:15 |
| emily | i.e. is there a chance that this was load-bearing to get Tecla "installed" at all? | 12:08:28 |
| winston | lemme see | 12:08:41 |
| emily | since ofc we will need org.gnome.Tecla.desktop to be around globally now | 12:09:01 |
| emily | (what even is Tecla) | 12:09:04 |
| winston | "keyboard layout viewer" | 12:10:13 |
| emily | the gnome thing of having internal codenames that aren't exposed in the UI is confusing :) | 12:14:16 |
| emily | they think they're Google! | 12:14:36 |
| emily | anyway happy to merge if you can confirm it's still hooked up enough to spawn | 12:15:06 |
| emily | (can merge the GDM thing too if you can check that it works, vouch for the code changes, get the commit messages fixed, and pinky promise to look into userdbd for 26.11) | 12:22:50 |
