feat: implement v2 weighted-majority outcome resolution

[collinsezedike]

Summary

Implements #68: weighted-majority outcome resolution for tholos-v2, the piece that turns revealed vote weight into a decided terminal_cause/final_outcome.

• After every reveal, a side locks the outcome (StrictMajorityFor/StrictMajorityAgainst) the moment its weight exceeds half of the frozen eligible total W, checked via side_weight > W - side_weight rather than division, so an odd W doesn't round the wrong way. Locking doesn't move the assertion out of Reveal: other positions can keep revealing afterward to prove entitlement for settlement, matching V2_RESOLUTION.md's state diagram (OutcomeLocked is folded into the existing Reveal phase rather than adding a new one).
• The assertion closes to Resolved once revealed_weight catches up with W, or reveal_deadline passes, whichever happens first. If neither side ever reached a majority by then, terminal_cause defaults to OptimisticTimeout and the originally asserted outcome stands, even in a lopsided-but-incomplete reveal (e.g. one side reveals 49% against with 50% of weight never showing up at all) or an exact tie at full reveal.
• New permissionless resolve_outcome entrypoint. It's needed because reveal() itself can never be the one to close things out past reveal_deadline (a call returning Err reverts all of its state changes in Soroban, so a reveal attempt made after the deadline can't leave a partial close behind), and because a dispute that drew zero third-party registrations has nobody who could ever call reveal successfully at all (the asserter/disputer's fixed positions are already marked revealed automatically when Reveal opens). resolve_outcome requires no signature: it only applies a deterministic rule to already-committed weight and elapsed time, moving no funds.

Settlement (who actually gets paid, and how much) stays out of scope here per the issue: terminal_cause is enough to know the outcome, the payout math is #69.

Test plan

• cargo fmt --check, cargo clippy --workspace --all-targets -- -D warnings, and cargo test pass locally
• cargo build -p tholos --target wasm32v1-none --release and cargo build --workspace --target wasm32v1-none --release pass locally
• mdbook build docs passes locally
• CONTRACT.md updated if the public interface changed — not applicable, tholos-v2 isn't documented there yet (tracked separately)
• scripts/testnet-smoke.sh run against testnet — not run; this only touches tholos-v2, which isn't deployed yet
• What you manually verified: added 10 new tests covering early strict-majority locking (both directions, plus an odd-total boundary case), the zero-third-party stuck-registration case, resolve_outcome's three failure modes and its idempotency, and the lopsided-but-incomplete-reveal optimistic-timeout case from V2_RESOLUTION.md. Two pre-existing reveal tests needed a second never-revealing voter added to keep exercising their original single-position scenario now that a fully-revealed dispute closes automatically.