dev: change type of branch node subnodes

[enitrat]

What was wrong?

The type for the BranchNode's subnodes was List[rlp.Extended]. While this is not inherently wrong, a more suitable type would be a tuple with 16 rlp.Extended elements, each representing a BranchNode of the trie, as there's no intent to mutate the subnodes after instantiation.

Related to Issue #

How was it fixed?

Changed to Annotated[Tuple[rlp.Extended, ...], 16]. Also open to just repeating 16 times the rlp.Extended argument, but using an annotation sounds more convenient rather than repeating. The annotation conveys at type hint that the tuple is of size 16.

Cute Animal Picture

Our beloved moodeng friend

[SamWilsn]

Does the annotation get enforced? If not, I'd prefer the tuple with 16 elements, probably as a proper type or type alias.

[enitrat]

The Annotation is not enforced. It's only a "hint" for the reader that it expects 16 elements. We could create a proper type that

Perhaps something like:

BranchSubnodes = Tuple[rlp.Extended, rlp.Extended, rlp.Extended, rlp.Extended,
                      rlp.Extended, rlp.Extended, rlp.Extended, rlp.Extended,
                      rlp.Extended, rlp.Extended, rlp.Extended, rlp.Extended,
                      rlp.Extended, rlp.Extended, rlp.Extended, rlp.Extended]


@slotted_freezable
@dataclass
class BranchNode:
    """Branch node in the Merkle Trie"""

    subnodes: BranchSubnodes
    value: rlp.Extended

[SamWilsn]

Kinda ugly, but I don't see a shorter way of making the type annotation meaningful to mypy. Let's go with the sixteen rlp.Extended type alias 🤣

[enitrat]

done :)

[enitrat]

Hi @SamWilsn, any chance you can help me understand why CI is failing ?