@nikomatsakis: @vlad and I were discussing the problem of "impls inside functions leak outside". Do you think it makes sense to submit some RFC on the topic, just to start things moving? I think the delta which we want to apply (modulo backwards compat) is:
#[macro_export]for local macros
It seems like we don't have capacity to work on this right now, but perhaps it makes sense to submit an RFC just to keep the problem documented?
It would be enough though to deprecate / lint against this with a message saying "don't do this, or it won't work with IDEs", right?
#[allow(leaking_impls)] on the function would cause IDE to look inside
Like, the act of silencing the lint makes it non-problematic, which is a fun property :)
unless you silence it in the whole crate ;)
I have been thinking about this a bit as well. I don't think I'd submit an RFC -- but it'd be nice to have some place to track it as a "future idea".
What that place would be?
There is no good place right now I don't think
But I dont' like using open RFCs for it, I know tha t:)
I guess in theory issues on rust-lang/rfcs is the official place
(I have wanted however to create something i the lang-team repository)
but I didn't yet :)
FWIW, I feel like my and Vlad 's heads provide enough redundancy here, so no additional storage is required. Logistically, it maybe makes sense to thing about this once chalk is fully in production?