Skip to content

Conversation

@Frontear
Copy link
Owner

@Frontear Frontear commented Dec 2, 2025

Hello there, another bi-weekly staging cycle has started up. The last cycle was 2025-11-10 -> 2025-11-30.

Breaking Changes

  • feat!: update to NixOS 26.05 (7f95eb9)

New Features

  • feat: significantly reduce the amount of runs done in nix-benchmark (cfec198)
  • feat: add Polychromatic to control my Razer devices (c6228fc)
  • feat: bump flake.lock (ff7cfd7)

Fixes

  • fix: update select flake inputs (205d947)
  • fix: update flake.lock, migrate to new DMS package, update nix-benchmark (49b8a89)
  • fix: improve Nix configuration for nix-benchmark (9fa0a1d)
  • fix: persist OpenRazer and Polychromatic paths (6f47bd8)
  • fix: disable mouse acceleration in Niri (6252162)
  • fix: drop the TODO for OpenRazer detection (7bc3048)

DankMaterialShell's author decided to be dumb and destroyed their git commit history, and apparently also broke the way their app and flake works too.. fun. I've gone ahead and fixed it, though this entire situation has made me want to quit using this shell entirely.

On other, more pleasant news, we are now on the latest NixOS release. The most noticable change should be the update to GNOME 49. I don't visually notice much but I did notice an airplane mode button, cute.

Lix 2.94 also released and removed the `no-url-literals` experimental feature, since it has now become a default. This is yet again another breaking change from upstream, since having this in your experimental features list will error out. Fair enough I suppose, though it does become a little annoying when one setup doesn't translate to the other cleanly. Overall a good change though, so I suppose there's that to be happy about.
I updated `nixpkgs`, `home-manager`, and `stylix`, mainly to stop the version mismatch warning.
DankMaterialShell changed their flake in a backwards incompatible way (again). This time, they've pinned quickshell as their own input, and also changed their package to `dms-shell`, which comprises both the Quickshell configuration as well as the CLI application (formerly known as `dankMaterialShell` and `dmsCli` respectively).
Instead of attaching both possible options into one, I dynamically generate the options based on which Nix variant is being used.
I feel like 20 runs is mildly excessive. It takes up a lot of CI time and from local testing, doesn't seem to really improve the accuracy of the final time measurement. At the very least, the accuracy is negligible compared to the runtime cost, so I'm dropping it by many factors. Warmup runs are going from 5 -> 4, and full runs are going from 20 -> 8.
This also adds the OpenRazer daemon, as that is the backend that Polychromatic communicates with.
The most important of these is the Polychromatic cache, and the OpenRazer config. I'm unsure if I really care much about the Polychromatic config, but for now I'm keeping it.
Doesn't make much sense, since peripherals can be swapped at any time.
@Frontear Frontear marked this pull request as ready for review December 17, 2025 03:30
@Frontear Frontear merged commit ff7cfd7 into main Dec 17, 2025
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants