refactor: decouple Foundry project layout and build system

[agostbiro]

Goal

Decouple the Solidity test runner from the project layout and build system and enable support for functionality that depends on compilation artifacts in the JS runner. Decoupling the project layout and build system is not only important to remove the dependence on Foundry’s format, but also to give flexibility to iterate on Hardhat's build system.

Background

The Solidity test runner needs access to Solidity compilation artifacts for the following functionality:

• Executing tests.
  • The bytecode and JSON ABI for test suite contracts is currently passed through the TypeScript API. For large projects this can become a performance bottleneck.
• The getCode and getDeployedCode cheatcodes.
  • These were originally intended for deployment scripting which we don’t want to support, but they are also being used for mocking which we do want to support.
  • Foundry’s implementation primarily reads artifacts from memory, but it can read from disk under certain conditions.
  • getCode and getDeployedCode are currently no-op when Solidity tests are executed through from JS.
• Invariant testing targets
  • Needed to access the ABI for the contracts that will be called over the course of a given invariant test fuzzing campaign.
  • This is currently no-op when Solidity tests are executed through from JS.

Solution

This PR brings the following changes:

• Pass all the compilation artifacts through the TS API, not just the test suite contracts.
• In addition, pass the test suite artifact ids so that EDR knows which artifacts are test suites to execute.
• As discussed, drop support for the following Foundry functionality:
  • auto-linking libraries (out of scope)
  • loading artifacts from disk for getCode and getDeployedCode cheatcodes as a fallback behavior (rarely used)
  • inline configuration in Solidity source files (rarely used)

Alternatives considered

An alternative approach to achieve the same goal would be to provide a callback to EDR from JS that EDR can call to query artifacts. This approach was tested in a spike, but it was discarded, because there are several places in the Foundry codebase where the Rust code will iterate over all the artifacts which will be very slow if that involves fetching all the artifacts from JS.

In the future, once the new Hardhat build system has stabilized, we could move file loading over to Rust to address potential performance problems from passing a large amount of data through the ffi.

Notes

I tried to minimize changes to interfaces to make this PR easy to review. The MultiContractRunner is especially a candidate for future refactoring after this change.

I manually benchmarked performance and found no regression.

[changeset-bot]

⚠️ No Changeset found

Latest commit: 8ade557

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[Wodann]

Overall LGTM. Just some small suggestions and questions

[fvictorio]

I only checked the js/ts changes. LGTM.