Optimize Core

[b-studios]

This PR builds on the work by @fritzladwig and supersedes #206.

It performs (in this order)

1. Deadcode elimination
2. Inlines uniquely used blocks (functions and objects).
3. Normalizes a bit (return(run(e)) = e, and so on)

It does not implement a static argument transformation. We need to reimplement that later at some point. Ideally on a CPS representation. We should also look into the standard shrinking techniques (https://www.cs.princeton.edu/~appel/papers/shrink.pdf) at that point.

It turned out that the simplifications were "too good" and loops became fully direct style (test: raytracer.effekt). In order to work around this issue in the JS backend, I introduced a new microphase that patches up recursive (non-guarded) function calls by thunking them under a trivial val _ = return (); recursive(). In the JS implementation this will cause a frame to be pushed on the monadic trampoline, which effectively makes recursive calls stack safe.

[b-studios]

Thanks @fritzladwig for your work again! It might look like I re-implemented everything, but that was much easier after you having implemented it already once.

[b-studios]

@CodiumAI-Agent /review

[CodiumAI-Agent]

PR Analysis

• 🎯 Main theme: Optimizing the Core of the Effekt language
• 📌 Type of PR: Enhancement
• ✨ Focused PR: True
• 🔒 Security concerns: No security concerns found

PR Feedback

• General suggestions: The PR is well-structured and the changes are focused on optimizing the core of the Effekt language. However, it would be beneficial to add tests to ensure that the optimizations do not introduce any regressions.

• 🤖 Code feedback:

  • relevant file: effekt/shared/src/main/scala/effekt/core/Optimizer.scala
    suggestion: Consider adding more comments to the code to explain the logic behind the optimizations. This would make the code more maintainable and easier for others to understand. [medium]
    relevant line: object Optimizer extends Phase[CoreTransformed, CoreTransformed] {

  • relevant file: effekt/shared/src/main/scala/effekt/core/Optimizer.scala
    suggestion: It seems that the 'shouldInline' and 'shouldKeep' methods are checking for the same conditions but returning opposite results. Consider refactoring these methods to remove redundancy. [medium]
    relevant line: def shouldInline(id: Id)(using ctx: InlineContext): Boolean =

  • relevant file: effekt/shared/src/main/scala/effekt/core/MakeStackSafe.scala
    suggestion: The 'thunk' method is not very descriptive. Consider renaming it to something more meaningful or adding a comment to explain its purpose. [medium]
    relevant line: def thunk(s: Stmt): Stmt = Stmt.Val(TmpValue(), Stmt.Return(Literal((), core.Type.TUnit)), s)

How to use

Tag me in a comment '@CodiumAI-Agent' and add one of the following commands:
/review [-i]: Request a review of your Pull Request. For an incremental review, which only considers changes since the last review, include the '-i' option.
/describe: Modify the PR title and description based on the contents of the PR.
/improve: Suggest improvements to the code in the PR.
/ask <QUESTION>: Pose a question about the PR.

To edit any configuration parameter from 'configuration.toml', add --config_path=new_value
For example: /review --pr_reviewer.extra_instructions="focus on the file: ..."
To list the possible configuration parameters, use the /config command.