| 9 Aug 2025 |
 raitobezarius | yes | 19:03:32 |
| raitobezarius | you need \n | 19:03:35 |
| raitobezarius | but not a big problem if you build a tool around this | 19:03:47 |
 jade_ | maybe i need to just go fix that bug | 19:19:15 |
 Lunaphied | I don't think we even knew that you could put a full config in via environment variables | 21:00:21 |
| 10 Aug 2025 |
| jade_ | https://git.lix.systems/lix-project/lix/issues/956 how do i keep finding these bugs.... | 09:38:54 |
| 11 Aug 2025 |
 niko ⚡️ | It's a good day when you bump lix for the first time in a month and once again there are failing functional tests, I love running auto-allocate-uids on darwin | 12:33:51 |
| niko ⚡️ | https://gerrit.lix.systems/c/lix/+/3893 classic no whoami fallback | 13:01:25 |
| niko ⚡️ | Oh no this doesn't fix all the failures | 13:13:07 |
| niko ⚡️ | sigh | 13:13:08 |
| raitobezarius | It would be ideal if we could put these configurations in CI | 13:19:05 |
|  Primrose (== Léana) changed their profile picture. | 13:28:57 |
| Primrose (== Léana) changed their profile picture. | 13:41:10 |
| niko ⚡️ | In reply to @raitobezarius:matrix.org
It would be ideal if we could put these configurations in CI Is that something I could contribute as well or is the CI config somewhere else | 20:18:20 |
| raitobezarius | For Darwin, it's a bit blocked on reconfiguring things because I dont think the Sandbox can run inside the sandbox | 20:19:04 |
| niko ⚡️ | Makes sense | 20:23:52 |
| 12 Aug 2025 |
 emily | auto-allocate-uids is frankly too broken to build most things on Darwin. though I have plans for how that could be improved | 00:32:50 |
| raitobezarius | Fwiw we will be able to bump kernel version again, raboof was able to minimise the kernel issue and a kernel maintainer fixed it: https://lore.kernel.org/all/20250811-iot_iter_folio-v1-2-d9c223adf93c@codewreck.org/T/#u | 02:07:07 |
| Lunaphied | What how did we even hit that | 16:06:06 |
| raitobezarius | The local releng in a VM test exercises 9pfs and folios :-) | 16:32:46 |
| emily | fyi parse-toml-timestamps is inherently broken and also toml11 is breaking eval compat for it which is blocking the upgrade in Nixpkgs – https://github.com/NixOS/nix/pull/13741#issuecomment-3180851635 | 20:04:59 |
| emily | Sergei Zimmerman (xokdvium)'s patch restores the behaviour of the old toml11 version which seems good for stable releases but I'd suggest considering dropping the feature entirely | 20:05:30 |
| raitobezarius | Lord | 20:07:00 |
| raitobezarius | If I gather well, this is under a clear XP feature? | 20:07:46 |
| raitobezarius | So I would also lean on deletion | 20:08:25 |
 Sergei Zimmerman (xokdvium) | That'd be great | 20:08:48 |
| Sergei Zimmerman (xokdvium) | The fact that old tests didn't cover the most fucked semantics is infuriating | 20:09:25 |
| raitobezarius | Yeahh, I suppose I was too optimistic about Tom's small language being well behaving | 20:10:09 |
| raitobezarius | All these fromXYZ file format deserializers are ticking bombs :( | 20:10:34 |
| emily | yes, fromTOML by default rejects timestamps | 20:12:14 |
