Sam McCall

Ah sorry, indeed I meant to say we supporte `private` today for code completion, and soon for include-cleaner.

I think you're right. export isn't really the inverse of private. If you have - foo.h: `int foo();` - umbrella1.h: `#include "foo.h" // export` - umbrella2.h: `#include "foo.h" // export`...

Ah, this makes a lot of sense, but it's hard to align with other constraints on ordering: - it should respond to quality of textual match as you type -...

This is great :-) I have a hacked up script that works-ish: https://gist.github.com/sam-mccall/1d0dedf61f45529394f4ae54345c0d25. It walks up from `root/a/b/c/d.rs` to find `root/a/b/BUILD`, and then runs `gen_rust_project //a/b:all` and returns the result...

Thanks David! > > I have a hacked up script that works-ish: https://gist.github.com/sam-mccall/1d0dedf61f45529394f4ae54345c0d25. It walks up from root/a/b/c/d.rs to find root/a/b/BUILD, and then runs gen_rust_project //a/b:all and returns the result...

@otiv-emiel-vanseveren are you currently working on this? It looks like there are a few separable pieces: - absolute file paths as discussed above (I sent #3033) - running a separate...

How much memory does clang use to parse the same file?

This came up again in #665. And again in https://github.com/ycm-core/YouCompleteMe/issues/3698 One annoyance with the `Path` idea is ownership: paths aren't cheap to copy, so I guess we need a PathRef...

Yeah, the details of those types probably don't actually work for us (e.g. C++'s paths are case-sensitive, and the method for checking equivalence does IO, reads through symlinks etc). But...

+1, this would be great. There are challenges... > the ability to offer proposals for free functions defined in files which have not yet been included would also require changes...