Remove dependence on ethash (pyethash)

[fselmo]

What was wrong?

The move toward prioritization of the execution layer & EVM logic.

• The goal for py-evm was once to provide the Trinity client with an EVM accessible via Python. This is not the case anymore as Trinity was sunset. With different consensus configurations, py-evm still serves well today for EVM logic testing. With the move to PoS (Proof of Stake) after the merge, there is a drive to remove the dependency on the ethash library. There are a few implementations of the Ethash algorithm in python and the goal of internalizing this logic into python-only code if desired is accessible without too much dev time being taken up. Using ethereum.org and the execution-specs as references, implement an in-house ethash algorithm that works well enough, even if it is significantly slower.

• Added motivation: Python 3.10+ support in ethash leads to unknown segfaults. Maintaining the C bindings file in ethash for little benefit becomes an added chore to debug, find someone to review + approve, and update. That repository does not seem to be very actively maintained anymore and larger projects like go-ethereum have already dropped mining and ethash dependence altogether. Internalizing the ethash algorithm gives ease of control for supporting new python versions.

How was it fixed?

• ethash implemented within py-evm as python code and POW is un-prioritized as a consensus choice within the library.

Todo:

• Clean up commit history
• Add or update documentation related to these changes
• Add entry to the release notes

Cute Animal Picture

[wolovim]

not at all convinced i know the full implications, but can definitely appreciate the dep removal and better modern python support 👏 for our purposes, i don't see the algo performance being a blocker.