fatxpool: view revalidation shall revalidate futures transactions

[michalkucharczyk]

Future transactions shall also be revalidated during view revalidation process. Otherwise some of them could stuck there forever.

[skunert]

TODO from discussion: Verify if this is still relevant. Quite old and note recently witnessed.

[michalkucharczyk]

To add more context:

The future transaction that is in the txpool for quite long time, could become invalid (e.g. due to runtime upgrade, or some other transactions executed).
Future transactions are not revalidated in the view - so theoretically such transaction could stay in the pool for very long time.

It will be dropped when the pool hits future limits or when it be revalidated in the mempool (which happens on finalization), but till then it probably will sit in the pool.

By validating future transaction in the view we could avoid this situation.

After that being written it is rather niche issue.

[michalkucharczyk]

invalid future transaction can also be promoted to ready. So theoretically it may happen that invalid transaction will be provided to the block builder.

This transaction will be revalidated later in view-revalidation, but we could avoid this situation by validating also futures.