Compiler team proactive agenda


#1

This thread is intended to be a permanent thread where we can discuss the agenda for compiler team proactive meetings. See this older thread for details on the proactive concept.

Meeting time:

  • Thursdays at 4pm Boston time
  • #rustc on IRC

Agenda:

  • 2017-06-15: landing async-await (notes)
  • 2017-06-01: mentorship and accessibility (proposal, thread)

#2

Agenda proposal: Discussing stable clippy via the method of making clippy a rustup component that is built with rustc.

The current form of the proposal is in https://github.com/rust-lang/rust/issues/40921#issuecomment-299073597


#3

OK so we didn’t really set any agenda for this week! I propose that it’d be great to talk about mentorship and accessibility of the compiler source. I’ve been thinking about this a lot and I think there are a lot of “small hurdles” that contributors find when they come to rustc, and I’d like to discuss what we can do to solve those.

Some links and thoughts:

  • issue 40902: more forge docs for mentoring
  • rustdocs for the compiler crates are not available
  • most of our types don’t have documentation
    • proposal: sort by which are used most frequently, open up a “checklist issue” with the goal of documenting all structs, starting with those used most frequently
  • no design documentations clarifying the overall way that the compiler works
    • what is AST? what is HIR? what is ty::Ty<'tcx>? how do these things relate?
    • what are the 'tcx and 'gcx lifetimes?
      • (I think we should consider changing those names, too.)

Basically I’d like to decide on what are most urgent problems and try to draw up a plan to address.


#4

cc @mgattozzi, with whom I have spoken about this – you interested or able to attend meeting today at 4pm Boston time (#rustc on IRC)?


#5

@nikomatsakis I can jump on while I’m doing some work at work today!


#6

I haven’t seen a lot of suggestions for our pro-active agenda this week. I propose that we talk about @Zoxc’s async-await branch and what we think it might take to land it (obviously in a highly experimental form). I will read into the implementation in the meantime and suggest others try to do the same.


#7

There was also a relatively large amount of discussion a few weeks ago on IRC in #rust-lang which may be worth reviewing.