The reason to do this is that it makes it calculable from inside an attestation how many epochs the attestation spans over, which is needed for proof of custody reasons. It's a relatively small change and so arguably easier to do now than to do as a patch in phase 1.
Note that this changes the meaning of latest_crosslink.epoch, from the epoch when the latest crosslink was included to the epoch that the latest crosslink was for. This affects the line:
* `state.latest_crosslinks[shard].epoch > state.validator_registry_update_epoch` for every shard number `shard` in `[(state.current_epoch_start_shard + i) % SHARD_COUNT for i in range(get_current_epoch_committee_count(state))]` (that is, for every shard in the current committees)
But this may actually make it _more_ correct, as it means that in the case where >512 shards are processed per epoch, and so a committee from the previous epoch could get finalized in the current epoch, that would no longer count toward every shard having received a "new" crosslink.
This change allows for easier maintenance of the code and the spec by
uncoupling them. Before any edit to either document resulted in having to
synchronize the other. By adding a reference to the canonical repo for the code
we avoid having to maintain a duplicate copy here.
The current spec is arguably inconsistent, in that if a set of N values gets chunked into M chunks where M is not an exact power of 2, the chunks between M and next_power_of_2(M) are filled with SSZ_CHUNK_SIZE zero bytes each, but the last chunk is not padded, and could be arbitrarily short (eg. if the values are 4 bytes and there are 257 of them, then that gets serialized into eight chunks chunks where the first four are 64 values each, the fifth is 4 bytes corresponding to the last value, and the last three chunks are SSZ_CHUNK_SIZE zero bytes). This PR fills every chunk up to exactly SSZ_CHUNK_SIZE bytes for consistency.
