| 16 Oct 2025 |
 ivy | and currently packages expose them as outputs but simply cannot use buildEnv on those specific paths due to the parsing being based on spaces | 15:20:05 |
 matthewcroughan | Can we merge https://github.com/NixOS/nixpkgs/pull/446139 ? | 15:45:33 |
| matthewcroughan | I've been testing it extensively in nixos-musl | 15:45:41 |
| matthewcroughan | https://github.com/MatthewCroughan/nixos-musl/blob/refactor/musl.nix#L105-L119 | 15:46:02 |
| matthewcroughan | ghpzin: | 15:46:37 |
| matthewcroughan | Like frankly, there are only a few PRs left to make Nixpkgs build with musl and work with vmtests, etc for 25.11 | 15:48:50 |
| matthewcroughan | Are fixes for musl considered unmergeable during the feature freeze jopejoe1 ? | 15:50:36 |
| matthewcroughan | IMO they're just fixes, especially if conditional, so most can target master | 15:51:01 |
 K900 | You need to lay off the gas | 15:51:34 |
| K900 | There's a very small group of people interested in musl fixes | 15:51:47 |
| matthewcroughan | Okay, I have my answer then, keep copying things into my tracking flake | 15:52:04 |
| K900 | And release managers aren't those people | 15:52:07 |
| matthewcroughan | I'm not looking to go faster, I'm just looking to know that this is the limit | 15:52:11 |
| K900 | The limit is, frankly, Alyssa and Yureka's bandwidth | 15:52:39 |
| K900 | Unless you find more people with musl experience to review this stuff | 15:52:49 |
| matthewcroughan | My tracking flake can help them make decisions, and that's my current goal with it | 15:53:01 |
| matthewcroughan | since there is no CI for it atm, I'm hoping to fill that in at least a little bit | 15:53:14 |
| matthewcroughan | I just think it would be cool if a default pkgs.nixos {} built with musl for 25.11 | 15:54:35 |
| matthewcroughan | And it's really very close | 15:54:38 |
| matthewcroughan | The onetbb PR has been up for 3 weeks though, and that's not musl, so I'm going to nixpkgs-review that and merge it | 15:56:37 |
 ghpzin | Afaik there is basically no timeline for gcc15 update (other than after branchoff), so from that side there is no hurry.
If it somehow also fixes musl issues, then it is another thing (did not test and did not know about it). | 16:04:38 |
| ghpzin | * Afaik there is basically no timeline for gcc15 update (other than after branchoff), so from that side there is no hurry.
If it somehow also fixes musl issues, then it is another thing (did not test and did not know about it myself). | 16:05:16 |
| ghpzin | * Afaik there is basically no timeline for gcc15 update (other than it being after branchoff), so from that side there is no hurry.
If it somehow also fixes musl issues, then it is another thing, I did not test and did not know about it myself. | 16:06:21 |
| matthewcroughan | It does happen to fix a musl issue yeah | 16:07:25 |
| matthewcroughan | all of the gcc15 fixes seem to fix musl issues | 16:07:32 |
| matthewcroughan | That's what I'm including in my nixpkgs-review now | 16:08:45 |
| ghpzin | That is probably just lucky coincidences, considering PR for gcc15 update breaks pkgsMusl.stdenv itself. | 16:12:24 |
| matthewcroughan | Ah yes, okay | 16:14:26 |
| matthewcroughan | Thanks for pointing that out | 16:14:36 |
| matthewcroughan | I have not tested that PR | 16:14:45 |
