test/cmd/which-formula_spec: avoid modifications outside test dir

[cho-m]

HOMEBREW_LIBRARY_PATH is the non-mocked path so setting shell_cellar based on this means modifications are done outside of test environment and instead in user's installation.

Can see this happen by adding a sleep 10 in test case and then checking cellar path.

• Before test

  ❯ /bin/ls /opt/homebrew/Cellar/foo
  ls: /opt/homebrew/Cellar/foo: No such file or directory

• During test

  ❯ /bin/ls /opt/homebrew/Cellar/foo
  1.0.0

Not sure why this isn't sandboxed. Maybe to avoid nested usage or some particular integration scenarios?

Instead, HOMEBREW_CELLAR is within our temporary test path.

---

• 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 brew benchmark results.
• 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?

---

• I did not use AI/LLM to create this PR, or I disclosed the tool/model below and reviewed its output; I did not attribute commits to AI and will answer maintainer questions and review comments myself without AI/LLM.

---

[Copilot]

Pull request overview

This PR fixes which-formula’s command spec to avoid creating/removing directories in a user’s real Homebrew installation during test runs by ensuring the test always uses the test-scoped HOMEBREW_CELLAR path.

Changes:

• Simplifies shell_cellar selection to always use HOMEBREW_CELLAR, preventing writes under a real .../Cellar when present.
• Keeps the Bash-command execution environment aligned with the test sandbox by consistently exporting HOMEBREW_CELLAR from the same temporary location.

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.

[MikeMcQuaid]

Thanks!