Basic support for create and create2

[g-r-a-n-t]

What was wrong?

There was no support for contract creation.

How was it fixed?

• Built out support for type attribute calls and added create functions to the contract type.
• The analyzer now records contract types that are created and the compiler adds their Yul objects to the contract's runtime.
• Added a compiler flag for Yul optimization.

To-Do

• Update Spec if applicable

• Add entry to the release notes (may forgo for trivial changes)

• Clean up commit history

[codecov-io]

Codecov Report

Merging #239 (7c166a0) into master (584f5ac) will decrease coverage by 0.13%.
The diff coverage is 88.33%.

@@            Coverage Diff             @@
##           master     #239      +/-   ##
==========================================
- Coverage   94.08%   93.94%   -0.14%     
==========================================
  Files          54       54              
  Lines        3701     3768      +67     
==========================================
+ Hits         3482     3540      +58     
- Misses        219      228       +9     

Impacted Files	Coverage Δ
compiler/src/yul/mappers/expressions.rs	97.16% <0.00%> (ø)
compiler/src/yul/mappers/module.rs	78.57% <75.00%> (+3.57%)	⬆️
analyzer/src/traversal/expressions.rs	92.95% <83.05%> (-1.93%)	⬇️
analyzer/src/namespace/scopes.rs	95.93% <85.71%> (-0.87%)	⬇️
analyzer/src/lib.rs	91.93% <100.00%> (+0.06%)	⬆️
analyzer/src/traversal/declarations.rs	96.77% <100.00%> (ø)
compiler/src/evm/mod.rs	95.83% <100.00%> (+0.83%)	⬆️
compiler/src/lib.rs	88.88% <100.00%> (ø)
compiler/src/yul/constructor.rs	100.00% <100.00%> (ø)
compiler/src/yul/mod.rs	100.00% <100.00%> (ø)
... and 5 more

---

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 584f5ac...7c166a0. Read the comment docs.

[cburgdorf]

Should we consider giving contracts a special attribute to obtain the address such so that the above could be written as:

pub def create2_foo() -> address:
        # `0` is the value being sent and `52` is the address salt
        foo: Foo = Foo.create2(0, 52)
        return foo.__address__

With the cast, a bug such as the following would be possible that would result in returning the wrong address.

pub def create2_foo() -> address:
        # `0` is the value being sent and `52` is the address salt
        foo: Foo = Foo.create2(0, 52)
        bar: Bar = Bar.create2(0, 11)
        return address(bar)

In the example above, we mean to return the address of the newly created foo but we accidentally return the address of bar instead. Sure, one can also accidentially return bar.__address__ but it feels to me that having the address accessible as a property removes one layer of error potential. That said, I really dislike the __address__ syntax but wouldn't know a good way to get around the potential for naming conflicts. At least not until we have something like traits that we could bring into scope conditionally to expose special functions/attributes.

[g-r-a-n-t]

Yeah, I wasn't sure how to deal with this.

We can discuss this more in #179 and if we come up with a better idea, it can be changed. For now, I'm inclined to just merge this as is.

[cburgdorf]

Yeah, should be good for now. 👍

[cburgdorf]

Looks good! Left some comments inline.