Jason Carver

> rlp/codec.py:72(encode_raw) > trie/validation.py:11(validate_is_bytes) > rlp/sedes/serializable.py:300(make_immutable) I think we can improve performance on all of these with `assert` statements. That way, default usages (and our tests) will work slower &...

I guess this should really be moved to py-evm/pyrlp/py-trie -- seems like we shouldn't remove the type checks in eth-hash until they are clearly a bottleneck.

I hope that having libraries break when you turn off assertions would be very rare. I would certainly consider that broken. Is that something you've run into before? @davesque

> We should be able to get roughly the same level of confidence with type checking via mypy. Which comes at the cost of *requiring* that any module that uses...

`bytes(PrivateKey)` would be especially great. I would kind of expect `bytes(PublicKey)` to be supported, by symmetry, but it also seems like something that won't be used much. `bytes(signature)` is unusable...

These are two different backends (mock and py-evm), they shouldn't import from each other. It's possible that it could make sense to have it import from a third place, which...

I think the most straightforward thing to do (and the ultimate plan) would be to support configuration via genesis json. That option wasn't available before, and now all it needs...

Drop py3.5 support (maybe even 3.6?)

Recovering a signature is a feature that we already test on lots of random inputs, even on invalid signatures. See https://github.com/ethereum/eth-keys/blob/ce2aade354c77531babaa16669529082a7835f39/tests/backends/test_native_backend_against_coincurve.py So it's going to be tricky for us to...

eth-tester beta 3? You're a good long ways behind... Can you upgrade to beta 39?