Caleb Cartwright

> I remembered it not formatting last I tried using it in rustc, and sure enough it still seems to fail when I try it in playground: It's not that...

> it'd need to be a new version to add formatting support for it if we did so after stabilization. And I don't think that dance is worth it; it...

> The final step is "recommended", and that stage is reached when the feature is documented and has good support for tooling. The goal is precisely to avoid the kind...

> Can we have the body of the block get formatted still? Unfortunately no @scottmcm Though technically I'm sure we could create artificial spans to capture the parts we want...

> I'm not clear @calebcartwright on why this would be impractical I think we may be referring to two different things @nikomatsakis My sense of the thread was that we...

> there has been a https://github.com/rust-lang/rustfmt/issues/4914#issuecomment-887104615 suggestion in the dedicated rustfmt thread minor correction: That rustfmt implementation tracking issue is actually not the right place. Thoughts/perspectives/etc. on the style/formatting rules...

This is going to become significantly less of an issue now that RLS has removed the auto-publish crates from its dependency tree, and will become completely obviated if/when RLS moves...

> About rustup, it seems to only go back a few weeks to check for needed components before failing, and since RLS hasn't built in longer than that it just...

FWIW those tests can be _somewhat_ brittle because many are based on upstream values in rustc that can be changed independently. Fixing them typically involves updating the expected test fixture(s)...

Thank you for following up on this! Would you mind rebasing and ensuring this gets applied to some of the newer classes too? (e.g. merge requests)