refactor(homestead): avoid int() in calculate_block_difficulty

[trakshan-mishra]

Description

Removes the int() conversions from calculate_block_difficulty in the Homestead fork (src/ethereum/forks/homestead/fork.py), expressing the same computation with Uint and U256 so the ethereum-types carry the intended signedness and bitwidth. Closes #1415 for Homestead only.

The docstring and existing comments are unchanged. Variable names are descriptive (no abbreviations). The exponentiation in the difficulty bomb is kept as ** (not a bit shift). No defensive checks are added.

The time-based adjustment is restructured to branch on whether it raises or lowers difficulty, avoiding the signed "factor" (max(1 - delta//10, -99)) that could be negative. The difficulty bomb is gated on block_number >= 200_000, equivalent to the old num_bomb_periods >= 0 check but never producing a negative intermediate.

Validation performed locally:

• 101,344 differential cases matched — the old and new functions were compared over the edge-case grid (timestamp deltas, block numbers, parent difficulties around the 99-unit clamp and bomb boundaries) plus 100,000 randomized inputs; all produced identical results.
• 1,653 Homestead tests passed (25 skipped) via uv run fill --fork Homestead -m "not slow" -n auto --dist=loadgroup --skip-index --clean tests.
• Seven of eight just static checks pass (format-check, lint, spellcheck, typecheck, deadcode, lint-actions, lock-check).

One check fails: lint-spec (ethereum-spec-lint) exits 1 with a single GlacierForksHygiene diagnostic:

ethereum.forks.dao_fork - GlacierForksHygiene:
    `calculate_block_difficulty` in `.fork` has changed. Glacier forks may only differ in difficulty block.

This is because dao_fork is a glacier fork whose calculate_block_difficulty is a copy of Homestead's. The lint compares dao_fork against its predecessor (Homestead) by AST, and since only Homestead is changed here, the two now differ. The diagnostic resolves when the same refactor is applied to dao_fork in a follow-up.

Per CONTRIBUTING.md, this PR follows the recommended two-step workflow: Homestead only as a draft, with the other forks to follow after feedback. Maintainer guidance is requested before expanding to the other forks — specifically on whether to handle dao_fork together with Homestead (to keep lint-spec clean) or separately.

Related Issues or PRs

Refs #1415.

Checklist

• Ran fast static checks to avoid CI fails, see Code Standards & Verifying Changes: just static
• PR title has the form <type>(<area>): <title>, where <type> and <area> come from an appropriate C-<type>, respectively A-<area>, label. The title should match the target squash commit message.

Cute Animal Picture

[danceratopz]

Sorry, we don't have the bandwidth to review agentic-driven drive-by PRs.