feat: Add a wrapper for matter-labs-eip1962 for EIP196

[kevaundray]

This PR does not change make any breaking changes as the default library is still sustrate-bn.

The main difference now is that the substrate-bn library is optional (this follows the same strategy that the repository uses when there are two libraries for a given functionality).

This means that substrate-bn is now explciitly specified as in a crates default features in order to enable it. Moreover, config macros were added to various parts of the code in order to pass cargo checks and tests that run with --no-default-features.

[kevaundray]

Based off of #2264

[kevaundray]

Currently to run tests against it, we do:

cargo run --features revm/matter-labs-eip1962 -p revme statetest tests/develop/state_tests

I don't think optional dependencies are being tested against in CI, only cargo check -- basing this off of looking into ci.yml and observing kzg-rs under the check-no-std job

[codspeed-hq]

CodSpeed Performance Report

Merging #2266 will not alter performance

_{Comparing kevaundray:kw/add-matter-labs (d56d192) with main (b7a4e0a)}

Summary

✅ 8 untouched benchmarks

[kevaundray]

On my laptop I have:

Before

Crypto Precompile benchmarks/precompile bench | ecrecover precompile
                        time:   [28.906 µs 29.387 µs 29.990 µs]
                        change: [+1.2033% +2.1702% +3.3946%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 15 outliers among 100 measurements (15.00%)
  3 (3.00%) low mild
  6 (6.00%) high mild
  6 (6.00%) high severe
Crypto Precompile benchmarks/precompile bench | bn128 add precompile
                        time:   [2.4155 µs 2.4253 µs 2.4394 µs]
                        change: [-0.1218% +0.4023% +0.9439%] (p = 0.12 > 0.05)
                        No change in performance detected.
Found 7 outliers among 100 measurements (7.00%)
  5 (5.00%) high mild
  2 (2.00%) high severe
Crypto Precompile benchmarks/precompile bench | ecpairing precompile
                        time:   [2.5919 ms 2.5974 ms 2.6031 ms]
                        change: [-0.3921% -0.0444% +0.2900%] (p = 0.80 > 0.05)
                        No change in performance detected.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

After

Crypto Precompile benchmarks/precompile bench | bn128 add precompile
                        time:   [1.5845 µs 1.5889 µs 1.5940 µs]
                        change: [-35.061% -34.745% -34.436%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) high mild
  2 (2.00%) high severe
Benchmarking Crypto Precompile benchmarks/precompile bench | ecpairing precompile: Warming up for 3.0000 s
Warning: Unable to complete 100 samples in 5.0s. You may wish to increase target time to 9.4s, enable flat sampling, or reduce sample count to 50.
Crypto Precompile benchmarks/precompile bench | ecpairing precompile
                        time:   [1.7464 ms 1.7546 ms 1.7657 ms]
                        change: [-32.660% -32.324% -31.960%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 15 outliers among 100 measurements (15.00%)
  4 (4.00%) low mild
  7 (7.00%) high mild
  4 (4.00%) high severe

To run before: cargo bench
To run after: cargo bench --features matter-labs-eip1962

Both commands were ran in the precompile directory

[kevaundray]

The config flags get complicated because the optimism crate re-uses it in granite and there are tests/checks which run with --no-default-features, so tests and functions need to feature gated so that no-default-features tests/checks pass

[rakita]

Superseded by #2305