Continuing the discussion from 2019 Strategy for Rustc and the RLS, let’s talk about ways to improve compilation time in a big way (over a 1-2 year time horizon, but with incremental wins along the way).
What are the big changes that make sense? What data can we gather to validate whether they make sense? What are the dependencies between them? Can we make these changes in a way that yields gains as we go, and not all at the end?