Use applyChainTick before transaction validation

[edsko]

The mempool must be able to validate transactions which have not yet been included in a block. At the moment, it validates these transactions against the current ledger state, in the assumption that they will be included in the "next" block. The set of transactions in the mempool is then updated every time a block comes in, and reevaluated against the new current ledger state.

This is not correct. The transactions in the mempool will be part of the body of a block, but a block consists of a header and a body, and the full validation of a block insists of first processing its header and only then processing the body. This is important, because processing the header may change the state of the ledger: the update system might be updated, scheduled delegations might be applied, etc., and such changes should take affect before we validate any transactions.

We therefore need to make sure, consensus side, that the transactions in the mempool are validated against the current ledger state to which we have applied the "chain tick" function for the header of the block in which they will end up. Of course, we don't have that header yet; it turns out, all we need is a slot, but we don't even have that slot number yet. We will therefore have to make some assumptions in the mempool: when we are validating transactions without producing a block (that is, when adding transactions into the mempool), we could assume the slot number is the next available one; when we are actually producing a block, then at that point we do have the correct slot number and we should use it.

This will also require some changes to the Byron wrapper. At the moment, the chain tick function only updates the update proposal state; it does not change either cvsLastSlot or the delegation state; for the validation to be correct, it should change both. Technically this requires changes to the spec (see IntersectMBO/cardano-ledger#1046), but since updates to the delegation state are idempotent, it's okay to (at least for the time being) do this consensus side (it will happen twice when we apply a block: once in the tick function and then again when ledger processes the block body, but that's not a big concern).

[edsko]

Unfortunately applyChainTick was only recently removed from the API (#1226); we will have to bring it back.

[edsko]

#1293 reintroduces the chain tick function and extends the Byron chain tick function.