Block forge should check for pending future blocks in ChainDB

[nfrisby]

The ChainDB only processes blocks it set aside as future blocks the next time a block is added. From the report:

In a way, calls to addBlock are how the chain database sees time advance.

Suppose we receive a block B that is slightly in the future, then idle long enough for B to no longer be in the future, and then forge a new block. Since we didn't receive any blocks after idling, we will not forge on top of B.

Since leadership opportunities are precious, we wouldn't want to squander one like that: forging on top of B would have made a more competitive block.

[dnadales]

@nfrisby, is this still relevant given the fact that we're working on this patch

[amesgen]

I think we can let this be closed by #525 👍, as this scenario will not arise anymore.

[nfrisby]

Closing. Yeah, PR 525 should prevent this.