What does .deployed() do?

[fvictorio]

Users tend to do this:

const contract = Contract.deploy()
await contract.deployed()

but that doesn't seem to quite work on testnets (it does on local networks). My answer usually is to tell them to use await contract.deployTransaction.wait(1) instead, but I'm not quite sure about why, and what's the difference. I think one would assume that they are equivalent, but as far as I know they aren't.

So, what is the difference? What does .deployed() do exactly? And what is the correct way to use them?

[ricmoo]

The .deployTransaction only exists on transactions which we deployed by a ContractFactory during this execution.

The .deployed() method will use the .deployTransaction if it is present (i.e. was deployed by a ContractFactrory during this execution), otherwise it will use provider.getCode to ensure there is a contract deployed at this address, which means it can be used in another process or otherwise orthogonally to a ContractFactory. This does not verify the bytecode is the correct contract (as it doesn't even know what the correct bytecode is).

So, the .deployed() will make sure there is a contract deployed at contract.address.

Does that make sense?

[fvictorio]

So what you are saying is that foo.deployed() resolves when there is code under foo.address, right? (If that's the case, I think it would be great if this was added to the docs.)

That makes sense, though I do wonder why some users have had problems that went away by replacing await foo.deployed() with await foo.deployTransaction.wait(). I'll try to dig deeper if it happens again and let you know if I find something weird going on.

[AlexTheLion123]

Ric, Is it necessary or recommended to use .deployed() after creating every contract instance? Since it queries the blockchain, it could be overkill?