| 13 Oct 2023 |
 infinisil | Or even automatically rebase? No that would mess with GPG signatures | 16:43:36 |
| infinisil | Merging master automatically into each PR? No that would be a super messy history | 16:43:50 |
| infinisil | Maybe it's possible to just kind of associate a new failing check run with PR's that should be rebased | 16:44:47 |
| infinisil | Let me try experimenting with that | 16:45:18 |
| infinisil | Aha! This looks pretty good https://docs.github.com/en/rest/commits/statuses?apiVersion=2022-11-28 | 16:49:02 |
| infinisil | This is such a general problem, I'm amazed it isn't solved yet..
Well, there is a solution, it's merge queues | 16:51:22 |
| infinisil | Maybe we should just enable merge queues for Nixpkgs? | 16:51:31 |
| infinisil | Okay maybe it will have to be the idea you suggested Robert Hensing (roberth): To have a CI check that ensures individual packages don't regress | 19:53:10 |
| infinisil | This won't have any of the above mess. The only disadvantage is that "rollout" of stricter checks takes a long time | 19:53:56 |
| infinisil | I guess this is like the original idea suggested in https://github.com/NixOS/nixpkgs/issues/256788#issuecomment-1737965650, but a bit more granular | 19:55:29 |
| infinisil | And at that point actually, might as well implement it without granularity initially | 19:56:28 |
| infinisil | So, I've gone full circle! | 19:56:33 |
| infinisil | Well, no there should be some granularity, because otherwise as soon as one package violates the new check, it won't be checked for all other packages either. This would mean having to fix the base branch many more times | 20:00:10 |
| infinisil | This channel is a great rubber duck lol, thanks everybody :P | 20:09:42 |
|  tgamblin changed their profile picture. | 22:25:03 |
| 17 Oct 2023 |
 Artturin | https://github.com/NixOS/nixpkgs/pull/261650 The by-name check thinks this has a merge conflict
no conflict on the merge-base or master
$ gco 9a4743b
HEAD is now at 9a4743b7e6da Merge pull request #261648 from r-ryantm/auto-update/cargo-udeps
$ git ls-remote --exit-code https://github.com/NixOS/nixpkgs.git refs/pull/261650/merge
3b9053f4b3802a4173cd2f6fa5e725999bf5a427 refs/pull/261650/merge
| 18:56:20 |
| Artturin | the message that git ls-remote outputs should be printed | 18:57:07 |
| Artturin | I'll make a pr | 19:03:14 |
| Artturin | https://github.com/NixOS/nixpkgs/pull/261693 | 19:05:59 |
| infinisil | Thanks! | 19:15:20 |
| 18 Oct 2023 |
 Rick (Mindavi) | I think adding by-name to all-packages is not intended, right? https://github.com/NixOS/nixpkgs/pull/261427 | 13:33:53 |
| infinisil | Mindavi: Nah that's fine, see https://github.com/NixOS/nixpkgs/tree/master/pkgs/by-name#changing-implicit-attribute-defaults :) | 13:34:38 |
| 20 Oct 2023 |
|  @ninjatrappeur:alternativebit.fr changed their display name from NinjaTrappeur to PicNoir (was Ninjatrappeur). | 10:33:11 |
 zmitchell | Is the binary cache ever garbage collected, or is something in the cache today guaranteed to be in the cache essentially forever given current procedures? | 14:28:30 |
 @piegames:matrix.org | No and no :) | 15:11:04 |
| Rick (Mindavi) | There's talk about it currently due to the costs of the storage | 16:07:37 |
 @quantenzitrone:matrix.org | damn firefox 82.0.3 from nixos-20.03 is still cached | 16:14:23 |
| @quantenzitrone:matrix.org | i know older stuff is cached too, but thats the first version i found that had a flake.nix | 16:14:51 |
| 21 Oct 2023 |
|  Dusk joined the room. | 04:54:03 |
| 23 Oct 2023 |
|  globin joined the room. | 09:01:02 |
