Test 🧪: add accessList gas cost tests

[pLabarta]

What does it do?

This PR adds test for 3 cases related to the accessList transaction field introduced in EIP-2930 and implemented through the evm pallet in order to confirm that the field does not present an attack vector for Moonbeam networks. In Moonbeam tho, behavior is somewhat different from the expected for the first 4 addresses and storage keys added to the access list of a transaction; costing less than expected.

Cases are the following:

• Transactions with big access lists should not be gossiped if they exceed the block gas limit
• After the 4th one, additional addresses should cost 2400 gas (exact value declared in the EIP)
• After the 4th one, additional storage keys should cost 1900 gas (exact value declared in the EIP)

Is there something left for follow-up PRs?

While working on these tests it was detected that transactions using the accessList field cost more gas that those without it, even when reusing the same storage and addresses multiple times. This means the feature does not work as expected as it is meant to allow users to save some gas by pre-warming the storage they will access.

[github-actions]

WASM runtime size check:

Compared to target branch

Moonbase runtime: 2288 KB (no changes) ✅

Moonbeam runtime: 2256 KB (no changes) ✅

Moonriver runtime: 2256 KB (no changes) ✅

Compared to latest release (runtime-3401)

Moonbase runtime: 2288 KB (+256 KB compared to latest release) ⚠️

Moonbeam runtime: 2256 KB (+244 KB compared to latest release) ⚠️

Moonriver runtime: 2256 KB (+244 KB compared to latest release) ⚠️

[github-actions]

Coverage Report

@@                    Coverage Diff                     @@
##           master   pablo/access-list-test      +/-   ##
==========================================================
- Coverage   74.58%                   74.47%   -0.11%     
  Files         377                      377              
- Lines       96042                    95720     -322     
==========================================================
- Hits        71632                    71283     -349     
+ Misses      24410                    24437      +27     

Files Changed	Coverage
/primitives/account/src/lib.rs	62.50% (-2.34%)	🔽
/runtime/moonbase/src/weights/pallet_treasury.rs	0.00% (-13.24%)	🔽
/runtime/moonbase/tests/integration_test.rs	99.39% (-0.03%)	🔽
/runtime/moonbeam/src/lib.rs	44.41% (-0.02%)	🔽
/runtime/moonbeam/src/weights/pallet_treasury.rs	0.00% (-13.24%)	🔽
/runtime/moonbeam/tests/integration_test.rs	99.39% (-0.02%)	🔽
/runtime/moonriver/src/lib.rs	44.63% (-0.02%)	🔽
/runtime/moonriver/src/weights/pallet_treasury.rs	0.00% (-13.24%)	🔽
/runtime/moonriver/tests/integration_test.rs	99.37% (-0.02%)	🔽

Coverage generated Tue Jan 21 17:12:28 UTC 2025

[TarekkMA]

LGTM, just few nits

[RomarQ]

I think this should be a mapping, accessing the same value multiple times is not useful for checking the AccessList usage. The value is considered warm after the first read, and subsequent reads should be cheaper.

[pLabarta]

I was mostly for testing if warm storage was cheaper after X subsequent reads when using an access list or not (they did not prove to be cheaper than a TX without the access list field).

For the current tests, I don't this it is necessary to have a mapping unless you think we should add some other scenario.