`setUp()` function proof not reinitializing when the tests change about evm-semantics

[ehildenb]

We should not be copying the code cell from the output of the setUp function, just the storage, origStorage, balance, and nonce. That way, it won't break in this way. Requires modification here:

evm-semantics/kevm-pyk/src/kontrol/foundry.py

Line 1338 in 75583c1

if init_state:

[nwatson22]

We should not be copying the code cell from the output of the setUp function, just the storage, origStorage, balance, and nonce. That way, it won't break in this way. Requires modification here:

evm-semantics/kevm-pyk/src/kontrol/foundry.py

Line 1338 in 75583c1

if init_state:

@ehildenb Are you sure that is the preferable behavior (just copying storage, origStorage, balance, and nonce? I am thinking about the case where setUp() creates a contract. In that case wouldn't we want the account for that contract to be copied with its code?

I tried to address the issue in #2064 now by, after copying the accounts cell, overwriting the code cell for the account Foundry.address_TEST_CONTRACT() to make sure it gets the newer version of the code.

[ehildenb]

Yeah, if the setup function creates another contract, that should be copied as well. But the code from the test contract shouldn't be included I think, because that code can change even if the setup or storage doesn't.