refactor(benchmark): rename zkevm tests to benchmark

[spencer-tb]

🗒️ Description

Renames tests/zkevm to tests/benchmark. Just to get the process started.

Additionally renames the releases & markers to benchmark.

🔗 Related Issues

This change will affect all open zkevm/benchmark test PRs.

✅ Checklist

• All: Set appropriate labels for the changes.
• All: Considered squashing commits to improve commit history.
• All: Added an entry to CHANGELOG.md.
• All: Considered updating the online docs in the ./docs/ directory.

[marioevz]

Looks good to me, I think we should go ahead and do the rename.

The fixture issue has an easy fix IMO: #1778 (comment)

We should also modify https://github.com/ethereum/execution-spec-tests/blob/main/.github/configs/feature.yaml in this same PR and rename all zkevm references to benchmark.

[danceratopz]

Thanks @spencer-tb. Would you mind taking care of the release naming (zkevm -> benchmark in .github/configs/evm.yaml and .github/configs/evm.yaml if appropriate?

[spencer-tb]

I updated the docs in this commit to assume a [email protected] release in the future.

I think after we merge this we finish the following:

• Include all test formats in benchmark releases #1778 in a follow up PR.
• Update all past/current PR titles & labels from zkevm to benchmark. I'm happy to do this.
• Finish the nethermind test ports: tracker(nethermind): Porting Nethermind tests #1734 cc @LouisTsai-Csie

Then we can spin up a [email protected] release, and use the new nomenclature moving forward! Wdy(guys)t?

[marioevz]

LGTM, just two small comments. Thanks!

[marioevz]

I feel like there's no way the Evmone Coverage Report / evmone-coverage-diff (pull_request) is going to complete successfully due to the amount of cpu intensive tests that are being touched.

The only way the coverage would complete is if it could intelligently determine that the fixtures are exactly the same and have just been renamed, and then skip those, but that's not yet a capability of the coverage script sadly.

If it fails IMO we should still merge.

[spencer-tb]

I feel like there's no way the Evmone Coverage Report / evmone-coverage-diff (pull_request) is going to complete successfully due to the amount of cpu intensive tests that are being touched.

The only way the coverage would complete is if it could intelligently determine that the fixtures are exactly the same and have just been renamed, and then skip those, but that's not yet a capability of the coverage script sadly.

If it fails IMO we should still merge.

We should consider adding the hasher logic to the coverage script for this!

(holds urge to YOLO merge :P)