| 7 Oct 2025 |
 Vladimír Čunát | I don't think this is a breaking change. | 08:09:26 |
| Vladimír Čunát | Looks even backportable to 25.05 to me. (though there probably isn't any need to, as we don't pass the option) | 08:10:06 |
| Vladimír Čunát | Though is it OK to let this be unfixed on 25.11? | 08:11:44 |
| Vladimír Čunát | I suppose we could always deal with it once someone runs into it. | 08:11:58 |
 K900 | I don't think it's unfixed | 08:12:10 |
| K900 | It's just that we build our bootstrap GCCs extra weird | 08:12:22 |
 dramforever | this is not going to affect any user if it builds fine | 08:13:03 |
| Vladimír Čunát | And apart from this gcc patch, there's consensus on everything else, starting staging-next now-ish from the current staging? | 08:13:42 |
| K900 | I think we should be fine | 08:14:24 |
| K900 | Unless anyone has anything else | 08:14:44 |
| Vladimír Čunát | I'm a bit unclear about when we do staging-next-25.05, as we should at some point. What will be the best moment. | 08:16:18 |
| Vladimír Čunát | (I don't recall anything else) | 08:17:18 |
| Vladimír Čunát | Maybe there will be a suitable point before 25.11 mass-rebuild-diverges from master. | 08:18:30 |
 Marie | gcc 15 cross fix? 👉️ 👈️ | 08:19:08 |
| Vladimír Čunát | Because once we have three different mass-rebuild branches, it becomes tough. | 08:19:08 |
| K900 | Do we need it in 25.11? | 08:19:30 |
 leona | I would personally say after the next staging-next would be fine | 08:19:32 |
| K900 | I'd rather not touch GCC at this point | 08:19:53 |
| leona | as then we probably have enough time to do another staiging-next before ZHF (or at least near that) | 08:20:02 |
| Marie | not really to be fair | 08:20:05 |
| leona | ah, maybe i miscalculated, it would be fine anyway i think | 08:20:47 |
| leona | (not fully awake yet....) | 08:20:55 |
| Vladimír Čunát | For these big changes it's like... if we want them in 25.11, better now. So it's good to decide early. | 08:20:59 |
| Vladimír Čunát | Oh, on Hydra I should first deal with blockers that prevented the previous staging-next from reaching nixpkgs-unstable. But after that I count on new staging-next eval. | 08:39:35 |
| Vladimír Čunát | at least once postgres deadlocks let it pass through. | 08:49:58 |
 sterni | K900, Alyssa Ross: yes we use ld.gold for Haskell. It’s technically not required anymore, but we tried ld.bfd in the summer and it just takes a million years to link anything | 09:14:10 |
| sterni | using lld would probably be worth a try in order to get rid of gold, but it’s more complicated to wire up and I’m not sure how linking C++ libs would fare on linux | 09:14:52 |
| Vladimír Čunát | 🤔 coreutils update?
https://github.com/NixOS/nixpkgs/pull/445644 | 09:47:31 |
| Vladimír Čunát | So that this cycle isn't too boring 😉 | 09:47:49 |
| Vladimír Čunát | Based on the announcement I'd say it's unlikely to break stuff, though. | 09:49:27 |
