feat(op): use RollupCostData instead of full encoding

[Rjected]

This is a WIP to ultimately replace the tx_envelope part of the OptimismFields with a new struct.

Currently it just introduces the struct, which basically mirrors the RollupCostData from op-geth, and uses it in a few places. The plan is to first integrate this internally, without breaking OptimismFields, and once it is properly tested, break OptimismFields to use this instead.

[DaniPopes]

non-zero

[DaniPopes]

the same functionality exists in gas costs, this should reuse that

[Rjected]

wdym / where? the counting for op data gas costs at least we want to remove, and put here instead

[DaniPopes]

revm/crates/interpreter/src/gas/calc.rs

Lines 354 to 374 in b7b92ae

	/// Initial gas that is deducted for transaction to be included.
	/// Initial gas contains initial stipend gas, gas for access list and input data.
	pub fn validate_initial_tx_gas(
	spec_id: SpecId,
	input: &[u8],
	is_create: bool,
	access_list: &[(Address, Vec<U256>)],
	) -> u64 {
	let mut initial_gas = 0;
	let zero_data_len = input.iter().filter(|v| **v == 0).count() as u64;
	let non_zero_data_len = input.len() as u64 - zero_data_len;
	// initdate stipend
	initial_gas += zero_data_len * TRANSACTION_ZERO_DATA;
	// EIP-2028: Transaction data gas cost reduction
	initial_gas += non_zero_data_len
	* if spec_id.is_enabled_in(SpecId::ISTANBUL) {
	16
	} else {
	68
	};

[Rjected]

ah i see, thx

[DaniPopes]

pretty sure all the arithmetic in this function and *fjord can be scaled down to use u64

[rakita]

@Rjected this looks nice, did you check why tests are failing?

[Rjected]

@Rjected this looks nice, did you check why tests are failing?

nope, was busy with other things, can pick this back up this week