Fix CLA max_sharpe with equal expected returns

[anmol0705]

Description

Fixes #738.

This fixes a TypeError in CLA.max_sharpe() when all expected returns are equal.

In this edge case, _compute_lambda() can return None for lam, but _solve() later compares lam against floats. This PR applies the existing CLA._infnone() helper consistently in the lambda comparison path, allowing the algorithm to fall through deterministically to the minimum-volatility solution.

Changes

• Adds a regression test using the sample from [BUG] CLA crashes with TypeError and produces non-deterministic results when expected returns are equal #738.
• Verifies max_sharpe() no longer crashes for equal expected returns.
• Verifies the result matches min_volatility() for that edge case.
• Keeps the CLA logic change narrow and localized.

Validation

• uv run --extra dev pytest tests/test_cla.py
• uv run --with ruff ruff check pypfopt/cla.py tests/test_cla.py
• git diff --check