nixos: remove allowSubstitutes = false from core modules

[lovesegfault]

Description of changes

These are used almost anywhere, and make building only uncached parts of
a NixOS system extremely annoying. They all already have preferLocal = true set, which should be enough to prevent them from being built
remotely, where the largest perf hit would be seen.

Alternatively, we could introduce a top-level config option to NixOS like
substituteTrivialDerivations to let users toggle this behavior.

IMO, these derivations produce tiny files, and fetching them from a substituter
is probably not significantly worse than building them locally.

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• 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/)
• 23.05 Release Notes (or backporting 22.11 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.

[lovesegfault]

Another idea would be a config to nixpkgs like alwaysAllowSubstitutes where, if it's true, it just flips the value for any derivations which would set it.

[lovesegfault]

I made another PR with the config-based approach here: #221048

[ElvishJerricco]

I don't understand what's annoying about not using substitutes for these? Surely, spending the milliseconds it takes to create a trivial file is preferable to reaching out over the network, especially considering how unreliable networking can be. Especially considering debugging network connectivity problems is likely to involve tweaking systemd units directly or indirectly, so you wouldn't want these attempting to reach out to a substituter.

[lovesegfault]

I don't understand what's annoying about not using substitutes for these? Surely, spending the milliseconds it takes to create a trivial file is preferable to reaching out over the network, especially considering how unreliable networking can be. Especially considering debugging network connectivity problems is likely to involve tweaking systemd units directly or indirectly, so you wouldn't want these attempting to reach out to a substituter.

It means your CI is always having to fetch everything it needs to build these, instead of them being properly cached. It's extremely annoying because it balloons the build time of seemingly minor system changes by causing hundreds of paths to be fetched just to build a text file that was already built and cached previously.