Less API calls to get the blockHeader to write a conflation.

[letypequividelespoubelles]

so, the blockhash makes 256 + (nb_of_blocks_in_the_conflation -1) call to get the header (and then the blockhash), while blobBaseFee does 2 * nb_of_blocks_in_the_conflation.
So yes, we could save nb_of_blocks_in_the_conflation -1 calls to get the header ...
We could optimize it indeen, but not a priority imho.

Originally posted by @letypequividelespoubelles in #2513 (comment)