Problem

The equal() helper in src/pythonnative/equality.py defines how hook dependencies and state snapshots signal a change. Its contract deliberately treats identity as equal, accepts a boolean comparison result, and treats exceptions or nonboolean comparison results as unequal without coercing them to bool.

src/pythonnative/hooks.py uses it for dependency comparison and state updates. The test suite has Element equality tests, but no direct tests for this helper's unusual-comparison behavior. Replacing it with a naive bool(left == right) would break array-like application values.

Requested change

Add tests/test_equality.py with focused tests of the existing behavior. Use tiny custom Python classes to model unusual comparisons so contributors don't need NumPy or another new dependency.

Acceptance criteria

• Equal and unequal ordinary scalar/container values return actual booleans.
• The same object returns True without invoking its __eq__, even if that method would raise.
• Distinct objects whose comparison raises return False.
• A nonboolean comparison result returns False without calling the result's __bool__; make that method raise to prove it isn't called.
• A truthy nonboolean result such as integer 1 isn't treated as boolean True.
• Include a case documenting that reusing the same mutable object doesn't signal a change after an in-place mutation.

This task adds regression coverage for the existing snapshot contract; it doesn't change equality rules or hook APIs.

Verification

uv run pytest tests/test_equality.py tests/test_hooks.py -q
./scripts/check.sh

Suggested commit message

Following CONTRIBUTING.md:

test(equality): cover identity and nonboolean comparison results

Before starting

Please follow the issue-claiming process: check for an assignee or linked pull request, comment to claim the issue, and wait for a maintainer to assign it before opening a pull request. Contributor setup is in CONTRIBUTING.md.