Revert "wayland: mark as broken on darwin"#214828
Conversation
This reverts commit d9986a5. That commit turned a package that built for macOS into a package that was marked broken for both (also breaking the wayland-scanner attribute, which is important for cross-compiling to Linux), and complicated the mesa-demos expression, with no explanation given and only four hours of opportunity for review.
|
Strong objection. The compiled Wayland package contains only stub files: $ find result/
result/
result//share
result//share/wayland
result//share/wayland/wayland.dtd
result//share/wayland/wayland.xml
result//share/wayland/wayland-scanner.mkwhich means that almost all packages previously marked as broken transitively by Wayland are now throwing an error on Hydra complaining about So I'm going to post another PR in a few hours to reject this one. You are welcome to disagree with me in that PR. Besides, this PR left three hours for review, and that's from 11pm to 2am for me. |
This PR was independently reviewed and merged, not an unreviewed self merge. |
I admit it's better than what I did, anyway. |
Yes. @wegank: please write a proper commit message next time (important for any non-trivial changes) and leave more time for reviews. Thanks :)
Not sure if that's true. AFAIK Wayland shouldn't be required on Darwin so those packages might only depend on it due to Nixpkgs dependency inaccuracies. I'm neither using Darwin nor am I interested in Darwin though so I have no idea / didn't look into it.
IMO that's fine for reverting problematic changes / significant regressions (they can always be re-landed / fixed later if necessary). I'm not sure how significant this regression was but I'd probably merged this PR as soon as CI passed. Anyway, no hard feelings but as maintainer of the Wayland package I just wanted to share my opinion on this matter as well. |
Description of changes
This reverts commit d9986a5.
That commit turned a package that built for macOS into a package that was marked broken for both (also breaking the wayland-scanner attribute, which is important for cross-compiling to Linux), and complicated the mesa-demos expression, with no explanation given and only four hours of opportunity for review.
Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)nixos/doc/manual/md-to-db.shto update generated release notes