Pull in new L1 blocks from client state in block assembly

[delbonis]

When producing a new block we need to include new L1 blocks that we have that aren't present in the chainstate's maturation queue. We should take any blocks ahead of the current chainstate in our client state and include them in the block body to extend the maturation queue (and possibly trigger maturation of deeply buried blocks). This implementation should be fairly modularized so we can compute at a glance based on the chainstate and client state which blocks will be included and which blocks will be matured as a result (so we can prepare prover resources, etc.). This may also include extending the client state to explicitly track these local blocks.

The most important part is that this logic should check for and handle reorgs, ie when we have blocks locally that would reorg immature blocks in the chainstate. This is the harder part, since it requires having some very basic PoW checking in chainstate computing the accumulated work differential.

Don't concern ourselves too much with the contents of the blocks, like the DA proofs and deposit state update txs. That can all fit into this system when we work on those components.

[john-light]

Should have this merged by end of current sprint