| 6 Jan 2026 |
 spewdins | uh well, the upstream mr from gitlab is what the darwin.patch is generated from
https://gitlab.freedesktop.org/wayland/wayland/-/merge_requests/481/diffs | 06:50:18 |
| spewdins | so, it will be 1:1 if the mr is merged right now, to upstream wayland in its current state | 06:50:47 |
| spewdins | any other changes from mr481 could be generated again to replace darwin.patch with the updated mr changes from mr481.
and, if it does become merged, we could always just remove the patch anyway | 06:52:22 |
| spewdins | I guess, in this case, I'm fixing previous bitrot | 07:03:04 |
 raitobezarius | Is there macOS folks who could take a look at https://github.com/NixOS/nixpkgs/pull/476848 ? | 10:09:53 |
 Alyssa Ross | No it won't, it'll be 1:1 to an unmerged upstream PR | 16:05:29 |
| Alyssa Ross | The current situation is not really sustainable. Wayland is going to keep breaking on macOS. Somebody who wants it needs to actually work with upstream to figure out how to get that MR merged. | 16:06:01 |
 emily | yeah, macOS is only going to have a healthy Wayland ecosystem if libwayland upstream actually builds | 17:52:10 |
| emily | if it's important to you then it's better to be the maintainer of Darwin support in libwayland than the maintainer of Darwin support in libwayland in Nixpkgs | 17:52:32 |
| emily | even if we merged the patch, it'll inevitably break on updates, and people aren't going to want to block Linux on that | 17:52:55 |
| spewdins | Makes sense. I understand the reasoning there | 23:58:59 |
| 7 Jan 2026 |
 Randy Eckenrode | Does by-name require the pname to match the folder name? | 01:00:19 |
| Randy Eckenrode | A lot of Swift packages are named swift-foo. I’d like to put them in fo/foo because otherwise sw ends up very full. | 01:00:42 |
| Randy Eckenrode | Just answered my own question. The folder name becomes the package name in nixpkgs. | 01:02:31 |
| Randy Eckenrode | * | 01:02:42 |
| Randy Eckenrode | It appears that Swift SDKs are just a sysroot named foo.sdk. | 03:23:55 |
| emily | "New attribute names should be the same as the value in pname." | 04:56:18 |
| emily | (https://github.com/NixOS/nixpkgs/tree/master/pkgs) | 04:56:23 |
 vcunat |
there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course.
| 08:19:46 |
| vcunat | *
"Should": there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course.
| 08:20:04 |
| vcunat | I don't know. I'm just dropping the tests to unblock:
https://github.com/NixOS/nixpkgs/commit/1551a606a703 | 08:51:32 |
 kdn | FYI: I have recreated the VM (launchctl bootout ... && rm -r /shared/dir) and it seems to work just fine now | 12:11:30 |
 eveeifyeve | I can't believe the apple-sdk didn't have a license. | 13:49:38 |
| eveeifyeve | * I can't believe the apple-sdk didn't have a license I just made a pr that adds it. https://github.com/NixOS/nixpkgs/pull/477742 | 13:50:42 |
| emily | we only use API stubs and headers from the SDK | 14:01:43 |
| eveeifyeve | Would that still count tho? | 14:02:01 |
| emily | which are either uncopyrightable or fair use depending on jurisdiction (Oracle v. Google etc.) | 14:02:13 |
| emily | we strip out all non-free binary code | 14:02:19 |
| eveeifyeve | * Would that still count tho? Because it's under copies or redistribution. | 14:02:25 |
| eveeifyeve | I would say still add the licence, because it still matters. | 14:03:21 |
