Yuji Kanagawa

Results 88 comments of Yuji Kanagawa

> Should the Debug::fmt be what is actually used as a fallback for no "display attribute" Agreed, I modified to do so.

Now I noticed it's a breaking change. Like https://github.com/rust-lang-nursery/failure/blob/master/failure_derive/tests/no_derive_display.rs , now we can write manual implementation if a struct/enum derives `Fail` and it has no `fail` attribute. So `Display` implementation...

Hmm... So, should we set threshold to display eldoc? (e.g. If cursor doesn't move 0.5 sec then call racer)

> I think my main problem is that I am using winapi. Just trying to look for one of the things I imported it started looking through tons and tons...

@apiraino > is the RUST_SRC_PATH env var needed at all or is it better to not set it? Now racer can work without `RUST_SRC_PATH`, but I think you have no...

I often use `-r` and `--no-install` simultaneously to speed up tests. Also, my point is to keep the behavior of `session.install` and `session.run("pip-sync")` consistent when `--no-install` or `--install-only` is given.

@FollowTheProcess Like this snippet: ```python if not session.no_install or session.install_only: session.run("pip-sync", requirements) ```

Thanks, I just added some tests to make coverage happy

> This looks like the use case for session.run_always, or am I missing something? That function won't run under --no-install. Ah, you're right, thank you. I feel like it'd be...

I'm sorry but I don't have clear motivation now