Michael Goulet

See https://github.com/rust-lang/rust/pull/100046#issuecomment-1214446815 This lifetime bound `S: 'a` should've been required the whole time, but the compiler was not smart enough and had unsoundness :sweat_smile:

Both use `:` in Rust, but the colon is always parsed as outlives in formality. This was confusing when @jackh726 wrote his projection PR, since it looked like trait goals...

Normalize type outlives obligations in NLL for new solver

10

Normalize the type outlives assumptions and obligations in MIR borrowck. This should fix any of the lazy-norm-related MIR borrowck problems. Also some cleanups from last PR: 1. Normalize obligations in...

Rework support for async closures; allow them to return futures that borrow from the closure's captures

32

This PR implements a new lowering for async closures via `TyKind::CoroutineClosure` which handles the curious relationship between the closure and the coroutine that it returns. I wrote up a bunch...

For a rigid projection, recursively look at the self type's item bounds to fix the `associated_type_bounds` feature

3

Given a deeply nested rigid projection like `::Assoc3`, this PR adjusts both trait solvers to look at the item bounds for all of `Assoc3`, `Assoc2`, and `Assoc1` in order to...

Don't allow try-builds to be rolled up

1

Pending try builds have their checkboxes unchecked, even though the code suggests they should be disabled: https://github.com/rust-lang/homu/blob/c8f51b8e261541c13747e50097cd8899aae9537b/homu/html/queue.html#L180-L184  I can see someone accidentally rolling up a try build because of...

Format trailing where clauses in type aliases

13

Self-explanatory. cc rust-lang/rust#114901 This implementation is probably super busted. I have no idea what most of these options in rustfmt mean, lol.

Move stability attribute section to std dev guide?

3

https://rustc-dev-guide.rust-lang.org/stability.html probably could live in https://std-dev-guide.rust-lang.org/, since it's not that relevant to the compiler. Just a thought.

Use fulfillment in method probe, not evaluation

7

This PR reworks method probing to use fulfillment instead of a `for`-loop of `evaluate_predicate` calls, and moves normalization from method candidate assembly into the `consider_probe`, where it's applied to *all*...

During `predicates_of`, we install outlives obligations which link the bivariant lifetimes of an opaque to their invariant (captured) copies. During NLL, we don't enforce well-formedness of an opaque that we...