interop: Remove checkInterop from Block Building Path

[axelKingsley]

Through recent transaction handling designs, we have decided that putting a blocking RPC call against every interop transaction during block building is inappropriate. For the following reasons:

• It costs excessive time during building
• Transactions are filtered at mempool ingress already
  • The only way a transaction would go from pass->fail in this time is if a remote reorg occurred on the referenced chain.
• Remote reorgs can still occur after this check, and even after this block is published, so we don't gain 100% accuracy by having this check here.

Design
FMA

We may want to include the recurring (timer) checks here in this same PR, but I wanted to kick off the work with this change, which is purely removal.

[protolambda]

This needs to be scrutinized very closely, due to reorg behavior (note: the local chain, not the remote chain) breaking pre-validation assumptions.

[axelKingsley]

This needs to be scrutinized very closely, due to reorg behavior (note: the local chain, not the remote chain) breaking pre-validation assumptions.

Ah right, the behavior in which all transactions from a reorged block get dumped back into the mempool. I would hope that mempool ingress would catch those, but I remember you and I talking about how we need to confirm this.