Fix block production race condition in integration test

[mrmr1993]

This PR tweaks the behaviour of the block producer to check that the best tip isn't at the slot that its currently trying to produce at.

This race condition shouldn't happen in practice on a full network, but arises in some of the snarkless integration tests in #10005 due to the change in timings of the new poseidon hash function for kimchi.

Checklist:

• Modified the current draft of release notes with details on what is completed or incomplete within this project
• Document code purpose, how to use it
  • Mention expected invariants, implicit constraints
• Tests were added for the new behavior
  • Document test purpose, significance of failures
  • Test names should reflect their purpose
• All tests pass (CI will check this if you didn't)
• Serialized types are in stable-versioned modules
• Does this close issues? List them