Investigate constraint count increase caused by enabling poseidon gate for the Tube circuit
See: https://github.com/AztecProtocol/aztec-packages/pull/7494
also for the ultra recursive verifier. I observed a ~46k gate difference before and after poseidon2 was added, when it should be way lower.
We should confirm that hashing in the tube does not result in unexpected numbers of gates.
TL;DR when we actually enabled the Poseidon2 gate in recursive verifiers arithmetised as ultra circuits (my big Georgia arc) we were a bit surprised by the increase in constraint count. It's not specifically the tube circuit, just ensuring the number of gates produced by Poseidon in ultra circuit is not higher than expected. Especially given that, at the time, there were still some things left slightly unconstrained in bb::field_conversions - we are missing a bunch of validate_on_curve checks