| 19 Feb 2026 |
 emily | they can in their own tree :) | 14:45:55 |
| emily | I can look at adjusting the wording once I actually have things evaluating. | 14:46:21 |
| emily | you know you have a great patch stack going when squashing things into the first commits of it with Jujutsu is noticeably slow | 15:58:45 |
| emily | because it has to rebase the hundred commits after it | 15:58:54 |
| emily | Randy Eckenrode: do we have a plan for Wine, btw? I believe you said that x86 emulation on AArch64 didn't work on macOS upstream currently? | 16:01:06 |
 Randy Eckenrode | No plan at the moment. | 16:02:23 |
| Randy Eckenrode | It would be good to get Wine building on aarch64, so Linux users can use it with Fex, but Fex doesn’t work on Darwin AFAIK. | 16:03:01 |
| emily | seems possible we'll want to keep just enough around to keep that working, though I didn't do any work to achieve that in my branch | 16:03:06 |
| emily | FEX does ship a PE-format DLL for Wine's AArch64 emulation stuff | 16:03:29 |
| emily | is there a reason to expect that wouldn't work? | 16:03:36 |
| Randy Eckenrode | I wouldn’t bother. CrossOver will need to do something to deal with it eventually. | 16:03:37 |
| emily | Apple did signal that enough of Rosetta will remain that it's plausible CrossOver can continue using it, though I have no idea how that's actually going to work | 16:04:11 |
| Randy Eckenrode | My understanding is it doesn’t work on Darwin, but I haven’t dug into why. | 16:04:12 |
| Randy Eckenrode | For legacy games. CrossOver also needs to support non-game applications. | 16:04:43 |
| emily | I doubt they can make the system detect whether an app is a game or not :) | 16:05:21 |
| emily | especially inside Windows emulation | 16:05:28 |
| emily | https://github.com/FEX-Emu/FEX/blob/main/Data/nix/WineOnArm/shell.nix uh :D | 16:05:57 |
| emily | apparently FEX's WINE DLLs that they ship are literally built using Nix ^^ | 16:06:14 |
| Randy Eckenrode | If CrossOver needs frameworks that aren’t supported under future Rosetta 2, it wouldn’t work. Apple could also limit it to signed games with known team ids. | 16:30:08 |
| Randy Eckenrode | We don’t know because Apple isn’t saying anything about how they will limit it. | 16:30:40 |
 validitymaiden | Is it ok for nix-darwin to follow nixos-unstable or does it have to follow some variant of *-darwin? | 19:10:25 |
 samasaur | In reply to @validitymaiden:matrix.org
Is it ok for nix-darwin to follow nixos-unstable or does it have to follow some variant of *-darwin? It will work but the channel blockers are different and you will probably have a better time following a non-nixos-specific channel | 19:15:37 |
| samasaur | For unstable, the corresponding channel is nixpkgs-unstable | 19:16:07 |
| validitymaiden | I've been using nixpkgs-unstable for nix-darwin, but I'm now trying to add my this to a monolithic flake for all my systems and I'm not sure how to handle the inputs. For example: nixos on a few platforms with home-manager and nixos-wsl which follow nixos-unstable and nix-darwin on macos, also with home-manager, following nixpkgs-unstable. | 19:43:26 |
| validitymaiden | I suppose I'd have to have a bunch of different inputs and the downstream modules would need to use the one appropriate for which channel to follow. | 19:45:01 |
| samasaur | yes :( | 19:48:04 |
| validitymaiden | Well, no worse than the 2 overlays I have, I guess 😅 | 19:48:33 |
| samasaur | yeah it is ugly but in practice you can ignore it most of the time | 19:49:57 |
| validitymaiden | Yeah, I'll get there. Been making good progress with it and so far I'm very impressed tbh. Hopefully I can get proficient enough to start contributing soon. | 19:54:05 |
 Sarah Clark | I'm reviewing https://github.com/NixOS/nixpkgs/pull/486412/ to see if it builds on Darwn. It fails in the frontend build when fetchPnpmDeps returns a different hash on Darwin vs. linux. Any ideas for resolving this? | 20:12:35 |
