Document using liveness to infer auto traits for async blocks #1088
Conversation
|
@rustbot label +I-nominate +T-lang I'd appreciate some feedback. I'm also open to doing a video or chat meeting if that's helpful. |
|
Error: Label I-nominate can only be set by Rust team members Please let |
|
Niko mentioned this in today's T-lang meeting and I raised that it'd be good to document the current state for this inference before adding changes, so that there's a good diff to look at. |
|
@Mark-Simulacrum - That's a good idea, I'll put up a PR documenting the current rules and then rebase this one on top of that so that the differences are clearer. |
eholk
force-pushed
the
async-liveness-auto-traits
branch
from
f0ddca7
to
652008e
Compare
We are in the process of changing this (rust-lang/#69663), but it would be good to document the existing rules before changing them. This should also help explain the compilation errors people are getting in the meantime.
eholk
force-pushed
the
async-liveness-auto-traits
branch
from
652008e
to
e810f63
Compare
|
I just pushed #1093 which documents the current behavior. This PR now includes that one as well, so once we merge that one it will be easier to see what is changing. |
This documents the proposed liveness-based rules for inferring auto traits for async block. Currently we use a scope-based approach, so this also serves as a description of we would like to change.
eholk
force-pushed
the
async-liveness-auto-traits
branch
from
e810f63
to
c91c126
Compare
In order to fix some of the issues around overly conservative generator captures (rust-lang/rust#69663), I am working on changing the analysis to be based on liveness of values rather than their scope. This is an update to the reference that describes how I propose things to work in the future, and I'd appreciate some feedback on it. Obviously we shouldn't merge this until the rustc changes are ready too, but I at least wanted to notify the lang team and get some feedback.