Redeploying a contract using bytecode on the blockchain?

Question

It seems redeploying a contract using bytecode on the blockchain is not possible (even assuming the ABI is known). The reason for this assumption is that the compiled code includes instructions for initialization including any constructor actions, however, the return data on contract initialization written to the block does not include constructor instructions. If it is possible to reconstruct and redeploy with information already extant on the chain, I would love to see how this is possible.

[In a side note: recent versions of the compiler provide "runtime bytecode" which allows one to check the validity of source code. This code should match contract code deployed to the block.]

Answer 1

You can do that.

Everything on the blockchain is deterministic and immutable. Therefore, a contract cannot exist on the blockchain without a corresponding transaction that deployed it. This transaction will include the constructor arguments and constructor code. This transaction is the first contract transaction shown on Etherscan.io, labeled "contract creation".

You would be able to redeploy an exact replica by sending a raw transaction with the exact same data payload. Context is important in many constructors, e.g. owner = msg.sender. Even if you don't know what all the code does, you would be the msg.sender in this case and you might get desirable effects, e.g. you own the new contract. You could even manipulate the constructor arguments to create a similar deployed contract with different constructor arguments.

It won't be especially easy to decode the payload without source code but it's a matter of degree of difficulty. A tool like this can help you decipher the bytecode. https://ethervm.io/decompile.

Hope it helps.