Remove problematic :final from Requirement sigs

[dduugg]

• 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?

---

Partial revert of https://github.com/Homebrew/brew/pull/16640/files#diff-b75efe9741aed6830d257b35e55eb243a38c7b187774b42cf26be6deef8990db

We only enabled :final enforcement two days ago, which has caused GH Action failures, e.g. https://github.com/Homebrew/homebrew-core/actions/runs/11991937306/job/33431223236?pr=198784#step:3:322

[apainintheneck]

I can reproduce the failure locally with brew audit --formula hyperestraier --git --skip-style so reverting makes sense. It'd also be interesting to find out why this only affects some formulae. I tried a few other formulae locally and they succeeded.

[dduugg]

Thanks! Used that to discover that it also fails with :final sig on fatal?. (The remaining :final instance method here, modify_build_environment, doesn't seem to present a problem, at least with this specific formula.)

I can't immediately identify the problem, I suspect it may be a sorbet issue, it will have to wait until I have the bandwidth to dive in though.

[apainintheneck]

Here's the full backtrace for the curious: https://gist.github.com/apainintheneck/0b8154b50940b80f3114e8969cccc0e0

I can't make heads or tails of it.

[Bo98]

It'll be from old formulae: Homebrew/homebrew-core@60003ee#diff-7af2303fb6e63cb2941f96909dcf3eef5cd35a6eab39c1691e4ec06f1d6e3f7a.

Errors from basic loading/parsing of old formula must result in a FormulaUnreadableError, i.e. be in this list:

brew/Library/Homebrew/formulary.rb

Line 110 in 8d30564

rescue NameError, ArgumentError, ScriptError, MethodDeprecatedError, MacOSVersion::Error => e

Unfortunately, Sorbet does not map this to an error like they do for TypeError: https://github.com/sorbet/sorbet/blob/370510182162529f01e9f2323e37c7b27131276e/gems/sorbet-runtime/lib/types/private/methods/_methods.rb#L184.

[Bo98]

In any case: while the old syntax is not really supported, some third-party taps do use still use it (because we never really deprecated it) and we probably shouldn't make those formulae completely unreadable: https://github.com/brewsci/homebrew-science/blob/5d14d2fda3934c0eb8243c1cbdb9fb4621ec110b/Formula/lp_solve.rb#L6. Or at least not in a patch release.

[dduugg]

It's interesting that I added this test before I discovered that :final checks had to be explicitly enabled

So it seems Sorbet is aware of being run in tests, and enables :final in that context, which was counterproductive in our case…

[MikeMcQuaid]

Thanks for fix @dduugg!