| 18 May 2026 |
hexa | https://github.com/NixOS/nixpkgs/pull/521373 | 23:28:24 |
| 19 May 2026 |
Randy Eckenrode | This fixes eval for me. I’m building the stdenv now to confirm it still bootstraps.
diff --git a/pkgs/stdenv/darwin/default.nix b/pkgs/stdenv/darwin/default.nix
index 3cf47e5a73..fe20459224 100644
--- a/pkgs/stdenv/darwin/default.nix
+++ b/pkgs/stdenv/darwin/default.nix
@@ -603,6 +603,7 @@
python3-bootstrap = super.python3.override {
self = self.python3-bootstrap;
pythonAttr = "python3-bootstrap";
| 01:07:39 |
Randy Eckenrode | * This fixes eval for me. I’m building the stdenv now to confirm it still bootstraps.
diff --git a/pkgs/stdenv/darwin/default.nix b/pkgs/stdenv/darwin/default.nix
index 3cf47e5a73..fe20459224 100644
--- a/pkgs/stdenv/darwin/default.nix
+++ b/pkgs/stdenv/darwin/default.nix
@@ -603,6 +603,7 @@
python3-bootstrap = super.python3.override {
self = self.python3-bootstrap;
pythonAttr = "python3-bootstrap";
+ zstd = null;
enableLTO = false;
};
| 01:07:51 |
hexa |  Download | 02:05:35 |
hexa | that's killing fseventsd hourly and I like it | 02:05:52 |
hexa | from the 14th fwiw | 02:06:02 |
hexa | no run-away swapping anymore | 02:06:16 |
hexa | Randy Eckenrode are you pushing a commit or should I take that patch? | 02:19:16 |
hexa | pushing to python-updates would be fine fwiw | 02:20:18 |
Randy Eckenrode | I just finished bootstrapping, building (non-bootstrap) Python 3.14, and confirming that it works. | 04:07:31 |
Randy Eckenrode | Give me a moment, and I can push to python-updates | 04:07:40 |
Randy Eckenrode | I pushed the fix. | 04:15:52 |
viraptor | Not sure who would need to approve this move. @[Randy Eckenrode] ? https://github.com/NixOS/nixpkgs/pull/520945 | 13:20:48 |
| @not-jack:matrix.org joined the room. | 14:15:47 |
@not-jack:matrix.org | anybody know if macfuse-stubs and fuse3 are similar enough to allow:
fuse3 = lowPrio (
if stdenv.hostPlatform.isDarwin then macfuse-stubs.override { isFuse3 = true; } else fuse
);
| 14:19:52 |
| whispers [& it/fae] joined the room. | 14:22:02 |
| same joined the room. | 14:47:19 |
| todo joined the room. | 15:41:35 |
| jappie joined the room. | 18:09:10 |
| 20 May 2026 |
| iamanaws joined the room. | 02:19:53 |
viraptor | In reply to @viraptor:tchncs.de Not sure who would need to approve this move. @[Randy Eckenrode] ? https://github.com/NixOS/nixpkgs/pull/520945 Or @emily ? | 09:07:22 |
toonn | I'm sure it'll get looked at, darwin-core bandwidth is just very limited. | 09:09:41 |
viraptor | Thanks, it was genuinely "who deals with base things like that". I'm happy to just wait for the group notification to do its thing. | 09:13:36 |
Austin Horstman | Something planned out beforehand? Seems a little weird to migrate from an official repo to a new fork of your own with a handful of claude commits on a single day of dev with no real usage by anyone for a stdenv package. | 13:19:07 |
Austin Horstman | * Something planned out beforehand? Seems a little weird to migrate from an official repo to a new fork of your own with a handful of claude commits on a single day of dev 4 days ago with no real usage by anyone for a stdenv package. | 13:22:14 |
| Myria joined the room. | 14:32:24 |
viraptor | In reply to @khaneliman:matrix.org Something planned out beforehand? Seems a little weird to migrate from an official repo to a new fork of your own with a handful of claude commits on a single day of dev with no real usage by anyone for a stdenv package. I posted about the process before here, but fair, new reviews could use some context. There's a tracking issue linked in the description for enabling building xcode projects using only the open software and it requires changes to lots of things. https://github.com/NixOS/nixpkgs/issues/512622 xcbuild is abandoned and archived from Facebook's side, so I took it over. If you have any issues with the actually quality of the code I'll be happy to address it. | 20:54:39 |
Austin Horstman |
xcbuild is abandoned and archived from Facebook's side
That makes more sense, then.
If you have any issues with the actually quality of the code I'll be happy to address it.
Nah, was just more concerned with the optics of the repo created/commits added/pr created to migrate on the same day just a few days ago for a foundational / widespread package.
I haven't been following it closely though so it was more of a question about what kind of problem solving / communication was done that warranted such a change.
| 20:59:23 |
| 21 May 2026 |
| rajudev joined the room. | 00:03:57 |
Gaétan Lepage | opencv is broken on darwin:
> ImportError: dlopen(/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/python3.14/site-packages/cv2/python-3.14/cv2.cpython-314-darwin.so, 0x0002): Library not loaded: @rpath/libhdf5.310.dylib
> Referenced from: <35714FA5-D465-3B35-9494-D152500A2BF3> /nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libopencv_hdf.4.13.0.dylib
> Reason: tried: '/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file), '/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file), '/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0//nix/store/glapng48by63qizjr5nxq6xw2vpnvkak-opencv-4.13.0/lib/libhdf5.310.dylib' (no such file)
| 10:29:10 |