test(saya-tee): settle blocks carrying cross-chain messages

[kariy]

What

Adds a regression phase to the persistent-TEE e2e (tests/saya-tee) that settles blocks which actually carry cross-chain messages — in both directions.

The existing harness only drives plain transfer blocks, so its messages_commitment is always over an empty message set. That's why a real bug went uncaught: saya hashed L1→L2 messages with the Ethereum keccak256 formula instead of the Poseidon hash Katana commits to, so any settled block that consumes an L1→L2 message is rejected by piltover with 'tee: invalid messages' and settlement stalls.

How

After the existing plain-block loop, the test:

1. declares + deploys the contract_msg_starknet l1-handler on the appchain;
2. L1→L2: calls send_message_to_appchain on the L2 piltover core, lets the appchain messaging collector relay it into an l1-handler tx, then asserts the block settles;
3. L2→L1: emits send_message_to_l1 from the appchain, then asserts the block settles.

Supporting harness changes: enable the messaging collector on the L3 node (config.messaging), and raise the L2 node's max_event_page_size (the collector queries with chunk_size = 200; test_config capped it at 100).

Verification

• Green against a Poseidon-hashing saya-tee: both message blocks settle, test passes.
• Red against a keccak-hashing saya-tee: the L1→L2 message block fails 'tee: invalid messages', piltover stalls, the settlement assertion times out.

Note

This guards the fix in saya (dojoengine/saya#77). Until that lands, building saya-tee from saya main (pre-fix) makes this test correctly fail — so this should merge after the saya fix is available to CI.

[kariy]

Ran cargo +nightly-2025-02-20 fmt (clean) and cargo +nightly-2025-06-20 clippy -p saya-tee-e2e-test --all-targets with the repo's scripts/clippy.sh flags. The new code is lint-clean. clippy does surface 7 pre-existing lints in the harness (assertions.rs, bootstrap.rs, saya.rs, and untouched parts of main.rs/nodes.rs) — this crate isn't clippy-gated in CI (clippy.sh runs -p katana only), so they predate this PR. Left them out to keep the diff focused; happy to clean them up in a separate pass.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 68.62%. Comparing base (9bde0ae) to head (b5311c0).
⚠️ Report is 429 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #581      +/-   ##
==========================================
- Coverage   73.32%   68.62%   -4.70%     
==========================================
  Files         209      321     +112     
  Lines       23132    45352   +22220     
==========================================
+ Hits        16961    31125   +14164     
- Misses       6171    14227    +8056     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.