feat: add contract creation execution functions

[Zer0dot]

Motivation

Currently, there isn't a way to deploy contracts directly from smart accounts. This PR addresses that by exposing two new top-level execution functions, one for create and one for create2.

Solution

Introduce two new functions create(bytes calldata initCode, uint256 value) and create2(bytes calldata initCode, bytes32 salt, uint256 value) which create a contract with initCode using their respective deployment methods.

These functions use the standard onlyAuthorized modifier.

Why not use a deployer contract?

Although practical, a deployer contract has a few limitations:

• Overrides msg.sender throughout contract construction.
• Prevents deterministic sequential create deployments (due to the nonce not being known until execution).
• Adds gas to contract creation.

Note that this doesn't mean that using a deployer contract isn't possible-- it's still entirely possible to deploy via, for example, the Create2Factory contract, or any other factory contract via standard executions.

[adamegyed]

Nice! Had a couple of questions and comments.

[jaypaik]

Awesome! A few thoughts (don't feel particularly strongly about any of them):

1. Do you think it's worth following the function signatures in https://github.com/safe-global/safe-smart-account/blob/main/contracts/libraries/CreateCall.sol for marginally easier compatibility with clients that already use Safes today?
2. CreateCall also emits events. (as does https://github.com/pcaversaccio/createx). Is this worth doing for analytics? Or do we want to save gas?
3. The examples above use a prefix, e.g., deployCreate2 or performCreate2. Is there an advantage to doing this? Perhaps makes for better grepability?

[Zer0dot]

Awesome! A few thoughts (don't feel particularly strongly about any of them):

1. Do you think it's worth following the function signatures in https://github.com/safe-global/safe-smart-account/blob/main/contracts/libraries/CreateCall.sol for marginally easier compatibility with clients that already use Safes today?
2. CreateCall also emits events. (as does https://github.com/pcaversaccio/createx). Is this worth doing for analytics? Or do we want to save gas?
3. The examples above use a prefix, e.g., deployCreate2 or performCreate2. Is there an advantage to doing this? Perhaps makes for better grepability?

I'm not sure how compatibility with Safes will work-- as this is a feature built-in to the contract, but point 3 makes me want to modify the signatures anyway!

With respect to events, my rationale is that contracts that need to be indexed can emit their own construction events without enforcing the cost. But, it would make for easier analytics, I'm open to including this as well.

[adamegyed]

Looks good!