EVM: non-empty "empty" account bugfixes

[jochem-brouwer]

Thanks to @kchojn who found this bug. It was introduced by #2054

The PR fixes 4 bugs in our lib.

This is the situation:

EOA Account A has balance B, nonce 0, and (of course) no code and no storage.
Account A now sends a Tx to contract C and it forwards ALL balance upfront to pay for gas. (So, at this point, account A has 0 balance as it starts running the contract).

Now JS has 3 bugs in these situations:
Calling EXTCODEHASH on account A will put 0 on the stack, not the empty code hash, because we report that account is empty.
CALL will charge callNewAccount gas when we forward nonzero amount from the contract back into the account
SELFDESTRUCT will do the same if we selfdestruct nonzero value to account A.

[codecov]

Codecov Report

Merging #2383 (1cffeab) into master (a85081e) will decrease coverage by 5.19%.
The diff coverage is 91.66%.

Additional details and impacted files

Flag	Coverage Δ
block	?
blockchain	?
client	?
common	?
devp2p	91.57% <ø> (?)
ethash	?
evm	79.50% <91.66%> (?)
rlp	∅ <ø> (∅)
statemanager	?
trie	?
tx	?
util	?
vm	?

Flags with carried forward coverage won't be shown. Click here to find out more.

[kchojn]

Awesome, I am glad that you helped to solve the problem so quickly! it looks like there are no more problems with it, gj

[acolytec3]

LGTM

[gitpoap-bot]

Congrats, your important contribution to this open-source project has earned you a GitPOAP!

GitPOAP: 2022 EthereumJS Contributor:

Head to gitpoap.io & connect your GitHub account to mint!

Learn more about GitPOAPs here.

[qbzzt]

Did you run https://github.com/ethereum/tests/blob/develop/src/GeneralStateTestsFiller/stExtCodeHash/extcodehashEmptyFiller.yml? It should have detected this bug because after EOA A sends its balance it should have a non-zero nonce.

Or was the problem that the balance sending tx didn't raise the nonce?