Revert "Return a shared type instance from type constructors when runtime checking is disabled"

[dduugg]

Summary

Reverts #23078 ("Return a shared type instance from type constructors when runtime checking is disabled").

Changes

Reverts the 33-line change to Library/Homebrew/standalone/sorbet.rb in full, restoring the original type constructor behaviour when $HOMEBREW_SORBET_RUNTIME is unset.

---

• Have you followed our Contributing guidelines?
• Have you checked for other open Pull Requests for the same change?
• Have you explained what your changes do? Performance claims (e.g. "this is faster") must include Hyperfine benchmarks.
• Have you explained why you'd like these changes included, not just what they do?
• For bug fixes, have you given step-by-step brew commands to reproduce the bug?
• Have you written new tests (excluding integration tests)? Here's an example.
• Have you successfully run brew lgtm (style, typechecking and tests) locally?

---

• AI was used to generate or assist with generating this PR.

[Copilot]

Pull request overview

This pull request reverts the prior optimization in standalone/sorbet.rb that returned a shared type instance from Sorbet type constructors when $HOMEBREW_SORBET_RUNTIME is unset, restoring the original runtime-disabled behavior for type construction.

Changes:

• Removes the no-op overrides for Sorbet type constructors (e.g., T.nilable, T.any, and typed collection indexers) in the runtime-disabled branch.
• Stops prepending the removed constructor-override module into T’s singleton class, so type constructors behave as they did before PR #23078.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[dduugg]

Closing in favor of #23093, which fixes the root cause (splitting the shared no-op type instance so only nilable returns a type containing NilClass) instead of reverting #23078 outright — keeps its measured performance win while restoring correct default: application on deserialization.

[dduugg]

The additional failure of https://github.com/Homebrew/brew/actions/runs/29296519551/job/86971139791 leads me to think that #23078 was a bit too creative with sorbet-runtime overrides (although the tapioca code should probably use the unpatched sorbet-runtime path, it still illustrates the dragons lurking with the approach). I will more thoughtfully revisit this approach another time.