The main question is does this even make sense? I suspect that we need to cherry-pick the optimizations that help and opt out of the harmful ones (extra bundle size, wasted CPU cycles, extra memory), using some sort of test app. Because we are almost exclusively leaf components, maybe this could be mostly harmful, not having much in the render tree that is pruned.
[ ] Lint the codebase with eslint-plugin-react-compiler.
[ ] Run all the tests on the output of the compiler. I believe we already do this with the Babel optimizations that we have, so won't be too hard. It's important to guarantee behavior.
[ ] Publish the source with the output of the compiler.
Summary
Per https://youtu.be/qd5yk2gxbtg?t=1061&si=ZueAd7hJ3rvYQBod we might want to explore adopting the react compiler.
The main question is does this even make sense? I suspect that we need to cherry-pick the optimizations that help and opt out of the harmful ones (extra bundle size, wasted CPU cycles, extra memory), using some sort of test app. Because we are almost exclusively leaf components, maybe this could be mostly harmful, not having much in the render tree that is pruned.
Another blocker is the overall community interest in this, which seems low: https://npm-stat.com/charts.html?package=babel-plugin-react-compiler&package=eslint-plugin-react-compiler for now.
The action plan would involve:
Examples
Motivation
Performance.
Search keywords: -