feat: deterministic proxies with CREATE2 and EIP1167

[skellet0r]

What I did

Extended the create_forwarder_to built-in by adding functionality to deploy minimal proxies using the CREATE2 opcode.

The following is now possible:

@external
def foo(_target: address, _salt: bytes32) -> address:
    return create_forwarder_to(_target, salt=_salt)

The deployment address of the proxy can be determined with the following:

keccak256(0xff ++ deployingAddr ++ salt ++ keccak256(bytecode))[12:]

where bytecode is equal to:

pre = "0x602D3D8160093D39F3363d3d373d3d3d363d73"
post = "0x5af43d82803e903d91602b57fd5bf3"
bytecode = pre ++ target ++ post

How I did it

Modified the built-in to accept and handle new keyword salt.

How to verify it

Check the new tests added. A set of tests verifying syntax, and a test verifying runtime execution.

Description for the changelog

Added support for deterministic deployment of minimal proxies using CREATE2

Cute Animal Picture

[codecov-commenter]

Codecov Report

Merging #2460 (fd4b30c) into master (b418e9f) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2460      +/-   ##
==========================================
- Coverage   84.67%   84.66%   -0.02%     
==========================================
  Files          94       94              
  Lines        9438     9446       +8     
  Branches     2206     2208       +2     
==========================================
+ Hits         7992     7997       +5     
- Misses        937      939       +2     
- Partials      509      510       +1     

Impacted Files	Coverage Δ
vyper/builtin_functions/functions.py	88.59% <100.00%> (-0.25%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b418e9f...fd4b30c. Read the comment docs.

[charles-cooper]

Nice work and neat PR. Is there a corresponding issue that this closes?

Also I think this can be done without the is_deterministic flag. You can just look for the presence of the salt kwarg and use that to decide whether to use CREATE2.

[skellet0r]

True, I started off without the is_deterministic flag but didn't want to exclude EMPTY_BYTES32 as the salt value. Although, how often will that ever be used as a salt ? 🤷‍♀️

And hmm, technically it's related to issue #649, but since this is embedded in create_forwarder_to, not sure it'd be eligible to close that issue.

Edit: ah actually I could just annotate that it's deterministic during the annotation phase ...

[charles-cooper]

I think you can just check in the AST whether the keyword is set at all. Not sure if you can do it in the implementation of CreateForwarderTo itself

[charles-cooper]

Yeah I think in build_LLL you can just check expr.keywords

[skellet0r]

Closes #649

[charles-cooper]

🔥 🔥

[fubuloubu]

sick!