ethclient: add AuthList field in toCallArg

[colinlyguo]

This PR adds the AuthList field in toCallArg to support eth_call and eth_estimateGas of SetCodeTx.

Some unit tests are added in this PR (containing uncovered cases e.g., ErrSetCodeTxCreate).

A typo is also fixed.

[fjl]

We have been discussing this. At the moment, it doesn't feel very useful since you need a signed authorization in order to invoke the call. It's not super convenient. And it's not really necessary to override an account with code via EIP-7702 in eth_call because it already supports arbitrary state overrides in Geth. And the new method eth_simulateV1 supports overrides in all clients.

[colinlyguo]

We have been discussing this. At the moment, it doesn't feel very useful since you need a signed authorization in order to invoke the call. It's not super convenient. And it's not really necessary to override an account with code via EIP-7702 in eth_call because it already supports arbitrary state overrides in Geth. And the new method eth_simulateV1 supports overrides in all clients.

aha. I see and thanks. IMHO the main intention for this PR is not to overwrite accounts but to return an accurate gas estimation. e.g. bundlers may need to estimate gas (if the bundle transactions have some delegation submissions. I learned this from here. Though they might be written in other languages, too). In the meantime, it doesn't make it worse to support AuthList here, haven't thought of any side effects.

btw, do you think the added unit tests are useful here? might tweak some if duplicated.

[s1na]

This is a mind-twisting test but I think I get it now :)

Authorization: account0 to itself
Execute: recipient is account0, it will fetch the delegate which is again itself (recursive). So it will return the raw delegation code (prefix+account0) and try to execute it which will fail.

Is that the intention?