I'm asking this question in relation to this one : pre_sending Ether where the OP asked in a comment of the answer about what happens at the low-level when pre-sending ethers to a smart contract.
As smart contract addresses are computed in a deterministic way, it is possible to pre-send ethers to these ones before the deployement. The future deployed contract will then have a non zero balance.
I understand the address of the smart contract is first registered as an EOA with the "pre-sending" and then updated as a contract address (
codeHash is set to non default value). My question is about the transition between EOA and smart contract address.
From my understanding, here is the process :
Pre-send ethers to a non deployed smart contract
Let's say I am a smart contract developer and I want to pre-send ethers to my future contract.
I have finished the development of my contract and my account nonce is
n. I compute the address of the contract using
address = sha3(rlp.encode([my_account_address,n+1])). We will call it the
pre-smart contract address. Then I send 1 ether to this address and wait for the transaction to be mined.
Once my transaction is included in a block, the state changes include the following :
My account nonce is now
Nodes register the pre-smart contract address as an EOA in their world state trie with the following account sate :
nonce: 0 balance: 1000000000000000000 storageRoot: default value codeHash: default value
Now I deploy the contract. The transaction is mined and the nodes detect the address is already present in their local database. They update the account state for this address as follow :
nonce: m, the number of contract creation made by the account balance: 1000000000000000000 storageRoot: hash of the root node of the account storage contents trie where the contract data resides codeHash: hash of the EVM code of this account
If this is true then in this statement about the
codeHash in the yellow paper :
it is immutable and thus, unlike all other fields, cannot be changed after construction.
the term construction refers to the attribution of a non default value to
codeHash (and not the creation of the field).
Am I correct about all of this ? A clarification would be appreciated.