nixos/calibre-web: customize data directory

[aymanbagabas]

This commit adds the ability to customize the data directory where Calibre-Web stores its data. It uses systemd tmpfiles to create the directory with the correct permissions.

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.

[FliegendeWurst]

Seems reasonable

[zehauser]

Wasn't this a breaking change?

[FliegendeWurst]

If you changed the value of the option, yes.

[zehauser]

I did indeed set the option. 🙂 This change caused calibre-web to crash-loop with a confusing error.

Do we not normally try to avoid needlessly breaking changes in NixOS modules? I'm not super familiar with NixOS community standards yet, though I have been using NixOS for a year or two now and regularly updating (I'm on nixos-unstable) and have not seen any other cases like this yet. I have seen options get deprecated or renamed with a clear eval-time warning or error message.

This breakage wasn't an edge case — we just changed the meaning of the option unconditionally, so that anyone who set it would be broken (at runtime). Wouldn't it have been better to deprecate and create a new option? Or even to try to do something clever like try to recognize an absolute path (the new semantics) vs a relative path (the old semantics)?

I'm not trying to criticize anybody here — I'd genuinely like to learn.

[FliegendeWurst]

You are right, this should have included an assertion that the provided value is an absolute path, and a changelog entry.
Check out #375539 for a different approach, where I restore the old behaviour whilst also allowing a fully customized data directory.