Feature request: Optimism support about eth-gas-reporter

[andrewshvv]

First calculation:

[smartcontracts]

FYI, we just updated the Optimism docs to give more information about how to accurately estimate the current L1 data cost for a given transaction: https://community.optimism.io/docs/developers/build/transaction-fees/#estimating-the-l1-data-fee

[smartcontracts]

When computing the total cost for a transaction on Optimism you need to account for two things, the L2 "execution" fee and the L1 "data" fee.

The L2 execution fee can be computed with the standard formula of provider.estimateGas(tx) * provider.getGasPrice().

The L1 data fee can be most easily estimated by making a query to OVM_GasPriceOracle.getL1Fee(serialize(tx)).

[andrewshvv]

(in progress, will update with time)

Optimism research:

• It is possible to estimate the L1 gas cost if we have any trouble getting the proper values required for cost calculations.
• From this article we can check how Optimism estimates the L1 gas cost in OVM_GasPriceOracle getL1Fee function.
• Found a way to see L1 fee by injecting L1 context into ether.js provider.

eth-gas-reporter research:

• Place where gas is accumulated from "it" block, needs to be populated with additional L1 gas.
• Async RPC provider in hardhat reporter which substitute the sync calls made by the core reporter
• deasync, turns async function to sync

Open questions:

• Is estimation is sufficient or do we need to get the actual gas cost?
• What is required to get the actual L1 gas cost on Optimism?
• How exactly injectL2Context function works in Optimism? Do we need to replicate it in sync rpc?

[cgewecke]

Yes, that looks right to me.

As a practical matter this tool is usually run with HardhatEVM or ganache-cli (both ethereumjs-vm based) It would nice to know if gasUsedL2 on an optimism vm is reasonably similar to what is used on an L1 ethereum vm.

Do you know if that's the case?

[cgewecke]

Cross-linking relevant issue at hardhat-gas-reporter

cgewecke/hardhat-gas-reporter#78

Will think a little bit about how this should be done....if we take the simplest approach and don't worry about representing the different layer costs in the table, most of the code changes would be in this file where the utilities to fetch and calculate gasPrices are:

https://github.com/cgewecke/eth-gas-reporter/blob/master/lib/utils.js

[andrewshvv]

Cool, glad to hear that!

It would nice to know if gasUsedL2 on an optimism vm is reasonably similar to what is used on an L1 ethereum VM.
Do you know if that's the case?

By reading this it seems so https://research.paradigm.xyz/optimism
But I will take a deeper dive in a couple of days.

Cross-linking relevant issue at hardhat-gas-reporter

Yeah, hardhat-gas-reporter is what I am actually using.

https://github.com/cgewecke/eth-gas-reporter/blob/master/lib/utils.js

Alright, I will take a look in utils.js, optimism rpc node documentation, ethereumjs-vm docs.

[math77]

this was implemented?

[andrewshvv]

@math77 nah, sorry, never got it working properly
I ended up rewriting too much, and got stuck introducing new bugs
It is open for anyone to take

[cgewecke]

Hi! eth-gas-reporter is being deprecated in favor of hardhat-gas-reporter.

In the latest version at Hardhat, Optimism Ecotone l1GasUsed and total cost are available by setting the L2 option to "optimism". The reporter synthesizes these costs locally so there's no performance hit for collecting this data while running your tests.

https://github.com/cgewecke/hardhat-gas-reporter/releases/tag/v2.0.0