| 24 Nov 2025 |
K900 | I mean we can just not build the Vulkan backend on x86_64 no? | 12:12:31 |
K900 | Actually we can probably yes build it | 12:12:31 |
K900 | And then just not have a driver | 12:12:31 |
niklaskorz | they only merged the MTLResidencySet PR so far but that does not require macOS 26 | 12:12:57 |
Randy Eckenrode | We want KosmicKrisp on x86_64 for Wine. | 12:14:55 |
K900 | Oof | 12:15:29 |
K900 | OK yeah | 12:15:30 |
K900 | Oof | 12:15:30 |
Randy Eckenrode | I may try to get the ARM64EC version of Wine building, but the FEX situation on Darwin is not great AFAIK. | 12:16:22 |
Randy Eckenrode | Swift comes first, which I plan to get back to doing sometime this week. | 12:16:56 |
niklaskorz | so Tom tried KosmicKrisp instead of MoltenVK for Tiny Glade, which does ray marching and all, and... it's about 4x slower than MoltenVK at the moment | 14:09:27 |
niklaskorz | might be better to not default to KosmicKrisp right on Mesa 26 release day | 14:09:50 |
niklaskorz | * might be better not to default to KosmicKrisp right on Mesa 26 release day | 14:09:56 |
niklaskorz | but that might speak in favor of making use of those Metal 4 APIs | 14:10:31 |
niklaskorz | * but that might speak in favor of making use of those Metal 4 APIs (for upstream, it's not a decision for us to make) | 14:10:42 |
Randy Eckenrode | Correct but slow is better than fast but wrong. I expect KosmicKrisp to improve performance faster than MoltenVK to improve conformance. | 14:21:34 |
Randy Eckenrode | Weβll see though. It may indeed be better to wait. | 14:21:48 |
Randy Eckenrode | * | 14:22:00 |
niklaskorz | indeed, and I don't see us shipping highly GPU intensive Vulkan applications just yet | 14:28:25 |
colemickens π³οΈβπ | So I think I'm hitting this: https://github.com/nix-darwin/nix-darwin/issues/1640
But I don't do any symlinking in my own config, I don't think, and it seems to be blowing up in home-manager font stuff. | 19:15:04 |
colemickens π³οΈβπ | I guess maybe it's an issue with home-manager, duh. I don't think I checked there yet... | 19:15:17 |
colemickens π³οΈβπ | huh, there's a fix commit in HM merged a week ago. Maybe my CI isn't working the way I expected because I should be including that -_- | 19:18:55 |
| 25 Nov 2025 |
samasaur | i would be pretty surprised if it was something in nix-darwin | 02:20:11 |
samasaur | perhaps it's a different thing you're using? buildEnv has some use cases in package definitions as well | 02:20:38 |
megmug | Download shell.nix | 10:44:01 |
megmug | Hello there, is it expected on a semi recent nixpkgs unstable commit, to have system libraries like iostream unavaillable on MacOS? I've been struggling getting a Linux project to build with a basic nix-shell that tries to use the clangStdenv | 10:44:24 |
megmug | I'm thinking I'm probably doing something wrong but I can't find a lot of info on the web | 10:45:06 |
megmug | I'm trying to compile a simple module like this, and I just can't get it to find iostream | 10:48:00 |
megmug | Download cli.cppm | 10:48:10 |
K900 | I don't think we really have C++ modules working | 10:55:30 |