style: fix on_macos/on_linux resource block checks

[Rylan12]

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same change?
• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your changes? Here's an example.
• Have you successfully run brew style with your changes locally?
• Have you successfully run brew typecheck with your changes locally?
• Have you successfully run brew tests with your changes locally?
• Have you successfully run brew man locally and committed any changes?

---

This PR fixes a brew style bug for the on_macos and on_linux checks with resource blocks.

Previously, the following would cause an error that broke brew style (this example is from Homebrew/homebrew-core#67535):

resource "cargobootstrap" do
  on_macos do
    if Hardware::CPU.arch == :arm64
      url "https://static.rust-lang.org/dist/2020-12-23/cargo-beta-aarch64-apple-darwin.tar.gz"
      sha256 "efbc0e72533d4ca7def9a985feef4b3e43d24f1f6792815bdba9125af1f8ecdf"
    else
      url "https://static.rust-lang.org/dist/2020-11-19/cargo-1.48.0-x86_64-apple-darwin.tar.gz"
      sha256 "ce00d796cf5a9ac8d88d9df94c408e5d7ccd3541932a829eae833cc8e57efb15"
    end
  end
end

Now, this will trigger the style issue and give the following message:

only a url and a sha256 (in the right order) are allowed in a `on_macos` block within a resource block.

---

Edit: this is now allowed. The check will still verify that only url and sha256 or url, version, and sha256 methods are used in on_macos/on_linux blocks in resource blocks (even if they are nested in if statements)

[BrewTestBot]

Review period ended.

[dtrodrigues]

The crystal formula has a version section within its on_macos block, not just url and sha256:

Library/Taps/homebrew/homebrew-core/Formula/crystal.rb:62:3: C: FormulaAudit/ComponentsOrder: only a url and a sha256 (in the right order) are allowed in a on_macos block within a resource block.
  resource "boot" do ...
  ^^^^^^^^^^^^^^^^^^

5437 files inspected, 1 offense detected, 1 offense auto-correctable

[dtrodrigues]

Changes look good once CI passes on all of homebrew/core.

[Rylan12]

Why do resource blocks need a version spec? Isn't it just based on the URL?

If it's needed, I think we should allow both:

resource "..." do
  on_macos do
    url "..."
    sha256 "..."
  end
end

and

resource "..." do
  on_macos do
    url "..."
    version "..."
    sha256 "..."
  end
end

[SeekingMeaning]

I wrote some of this code so my review might be biased 😝

[Rylan12]

I wrote some of this code so my review might be biased 😝

Well, you can approve my code and I'll approve your code 😉 ✅

I'll let this sit for a tad longer in case some other maintainers have opinions. Since it's blocking a core PR, though, I'll merge it later today.

[SeekingMeaning]

One thing I noticed was that a lot of the tests in the tests/rubocops directory don't have very good descriptions for it and context blocks. (I don't blame you @Rylan12 in this particular PR since the surrounding descriptions here aren't very... good)

For example, if the test surrounding this comment fails, here's what brew tests spits out:

Failures:

  1) RuboCop::Cop::FormulaAudit::ComponentsOrder resource the content of the on_linux block is wrong in a resource block

In contrast, here's what is outputted when this test in test/utils/curl_spec fails:

Failures:

  1) Utils::Curl curl_args returns --disable as the first argument when HOMEBREW_CURLRC is not set

This is an issue that probably should be addressed in one or more future PRs (by willing volunteers, of course)

[Rylan12]

Yeah, I've certainly noticed this as well. I went ahead and fixed up just this section. Later, I can start going through and fixing up other sections as well.