kanata: fix hash mismatch on case-insensitive filesystems

[jian-lin]

Description of changes

related #308089

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

---

Add a 👍 reaction to pull requests you find important.

[jian-lin]

CC @toonn for review since this is suggested by you.
CC @thefloweringash for review since the duplicated files are found by you. BTW, could you share the way you use to find that?

[SuperSandro2000]

Could you also send an upstream PR to resolve this?

[jian-lin]

Which upstream do you mean?

Presumably, the issue is in cargo. I just linked two cargo issues in #308089 (comment). cargo maintainers think they have not come to a conclusion on if it should be fixed and what the right constraints are. So I will not fix this in cargo. Fixing this in Nixpkgs eases the burden of bumping to new version a lot.

[toonn]

LGTM

Does upstream really intend for there to be two differently capitalized READMEs though, seems quirky?

[jian-lin]

Does upstream really intend for there to be two differently capitalized READMEs though

Looks like so.

---

Merge conflicts are resolved. I have verified that the hash is the same on both linux and darwin. Let's merge this.

[toonn]

Where is README.md coming from actually? The native-windows-gui repo seems to only have readme.mds, though 1.0.13, the version on crates.io, is not a tag in the repo.

I also checked Cargo.tomls, one of the issues mentioned problems when those contained a different case for README files but none of them seem to mention any.

[jian-lin]

Does upstream really intend for there to be two differently capitalized READMEs
Where is README.md coming from actually? The native-windows-gui repo seems to only have readme.mds, though 1.0.13, the version on crates.io, is not a tag in the repo.

Oh, so by "upstream" you mean native-windows-gui? I thought you meat cargo by "upstream" because the real issue is in cargo.

My understanding is that native-windows-gui's readme file is named readme.md, but cargo wants a README.md, so cargo generates a README.md from readme.md when publishing a crate. That is exactly rust-lang/cargo#14020. You can verify this following this command on a case-sensitive FS.

[jian-lin]

The subtle thing about reproducing steps in rust-lang/cargo#14020 is that you have to run cargo package on a case-insensitive FS!