Transaction replay blanket ban is wrong - needed for keyless contract deployment

[SKYBITDev3]

@MicahZoltu

These both execute the constructor like a normal deployment would. No need to have them also call an initialize function (why not just put code in the constructor if you want it called on deployment?).

There can be reasons for having a separate initialize function, e.g. see Axelar’s CREATE2 factory called Constant Address Deployer which has deployAndInit “in case you need constructor arguments that are not constant across chains, as different constructor arguments result in different bytecodes”.

But my point was that your CREATE2 factory (“deterministic-deployment-proxy”), although being great work and that it’s now deployed on so many blockchains, it shouldn’t have to be the only one we can use ever. You said something similar a few years ago, i.e. users are free to create and use their own or other factory (“deployer”):

@MicahZoltu

The reason I didn’t create an EIP is because I don’t think there is significant value in standardizing it. Anyone can deploy with whatever mechanisms they want. I can use my deployer and you can use your deployer and everything works just as well as if we both used the same deployer.

Some may want to create or use some other factory that works differently. But in order to deploy the factory via keyless deployment, they need to replay a deployment transaction. With nodes increasingly forbidding transaction replay, it’s becoming almost impossible.

@MicahZoltu

Some blockchains just deploy these contracts as precompiles

There will be newer factories that people want to get onto a blockchain and start using. Hoping that a blockchain will deploy one as a precompile isn’t realistic.