Flaky: quil-engine storage attestation test aborts during shutdown

[blacks1ne]

quil-engine::e2e_consensus worker_active_storage_attestation intermittently reports test result: ok, then exits with SIGABRT and pthread lock: Invalid argument.

This now reproduces on upstream itself, without the earlier PR queue, and in isolated runs. Twenty same-tree repetitions produced eight aborts and twelve passes. The earlier conclusion that it required concurrent workspace load does not hold. Investigate shutdown and native resource lifetimes; no root cause or product impact is established. Keep the storage-attestation regression enabled.

Reproduce after resolving the execution test dependency and lockfile blockers:

cargo nextest run --workspace --no-fail-fast --config-file .nextest.toml --features quil-hypergraph/test-utils -E 'test(=worker_active_storage_attestation)'

Tests: one full-workspace abort, followed by 8/20 isolated aborts.

Tested upstream: 58de7bd0a5dc5f57e63d1500f0d479e1c05d305b

[blacks1ne]

Resolved by #666, merged into v2.1.0.25 at a564a345f8bc2aa6c35381eafe9d717bc79b59c9. The storage-attestation harness now requests cooperative engine shutdown and joins its engine and event-drain tasks before test exit. Same-executable A/B reproduced 4/20 aborts with the old teardown and none in 40 joined runs. The final fix passed 40 isolated repetitions, all 21 e2e_consensus tests, and the affected handle unit test before merge; its patch is preserved in the upstream squash. Closing this test-shutdown issue; the exact failing native mutex was not identified.