Shuhei Kadowaki

These JuliaFormatter versions aren't compatible with CSTParser-related code of Atom.jl. So we won't merge this unless we update that code as well. Anyone who can take that job is very...

xref: https://github.com/JunoLab/atom-julia-client/pull/641 xref: https://github.com/JunoLab/atom-ink/pull/236

Here are some edge cases I've noticed but still I'm not sure about how to handle. ### local refactoring If we have the code like: ```julia function outer() val =...

I agree the logic inside [`renamerefactor`](https://github.com/JunoLab/Atom.jl/blob/86b3b2c795cda7a3accd99b7b72476e3bf36040a/src/refactor.jl#L18-L77) would look messy, and please feel free to ask me if you find anything weird -- I probably miss some edge cases. https://github.com/JunoLab/Atom.jl/blob/86b3b2c795cda7a3accd99b7b72476e3bf36040a/src/refactor.jl#L18-L77

#208, #209, #210 need to be merged in order to finish up this PR (this only thing remain is to add tests)

> We'll also need to think about proper UIs for all of this, especially the global refactors. I'd love to reuse `find-and-replace`s pane, but that doesn't seem to be possible....

Honestly speaking, now repl.jl is kinda beyond my understanding, so I don't feel I can give a good code review on this. Well, I tested this branch for a while...

@nanosoldier `runbenchmarks("inference", vs=":master")`

Indeed, this could be a great chance to introduce something similar to LLVM's optimization manager, which would centralize the management of optimization passes. I'll start thinking about the design.

> There's a few different usage modes: > > * The external absint sets `def` to a Method. In this case, the set of > `edges` is completely partitioned between...