refactor and cleanup region errors for NLL
This is a WIP commit. It simplifies some of the code from https://github.com/rust-lang/rust/pull/51536 and extends a few more steps towards the errors that @davidtwco and I were shooting for. These are intended as a replacement for the general "unable to infer lifetime" messages -- one that is actually actionable. We're certainly not there yet, but the overall shape hopefully gets a bit clearer.
I'm thinking about trying to open up an internals thread to sketch out the overall plan and perhaps discuss how to get the wording right, which special cases to handle, etc.
r? @estebank
cc @davidtwco
Suggest not mutably borrowing a mutable reference
This PR would (hopefully) solve #45392. I deviated a bit from @estebank's instructions since the error span only included the borrowed expression (e.g. the `b` in `&mut b`). I also didn't check the mutability of the local binding, since this whole case is concerned with an immutable local.
I can see two outstanding questions:
1. `note_immutability_blame` is called in two places, but I only have one test case. I think it covers the call in `report_bckerror`, but I'm not sure how to trigger the call from `report_aliasability_violation`.
2. There is one failing test, where the local binding is `self: &mut Self`. I'm not entirely sure what the correct output should be, but I think the new message should also apply. Unfortunately, since this parameter is parsed differently, its `let_span` covers both the pattern and the type, leading to a wrong suggestion text. I'm not sure how to correctly identify this case.
This commit is concerned with the case where the user tries to mutably
borrow a mutable reference, thereby triggering an error. Instead of the
existing suggestion to make the binding mutable, the compiler will now
suggest to avoid borrowing altogether.
`PointerKind` is included in `LoanPath` and hence forms part of the
equality check; this led to having two unequal paths that both
represent `*x`, depending on whether the `*` was inserted
automatically or explicitly. Bad mojo. The `note` field, in contrast,
is intended more-or-less primarily for this purpose of adding extra
data.
In particular, I am adding an implicit injected borrow on the pattern
matches, and when we go around the loop, the compiler is reporting
that this injected borrow is conflicting with the move of the original
value when the match succeeds.
Instead of tracking the "cause" of each bit that gets added, try to
recover that by walking outlives relationships. This is currently
imprecise, since it ignores the "point" where the outlives relationship
is incurred -- but that's ok, since we're about to stop considering that
overall in a later commit. This does seem to affect one error message
negatively, I didn't dig *too* hard to find out why.
NOTE: I was careful to make each change in a manner that preserves the
existing diagnostic output (usually by ensuring that no lines were
added or removed). This means that the resulting source files are not
as nice to read as they were at the start. But we will have to review
these cases by hand anyway as follow-up work, so cleanup could
reasonably happen then (or not at all).
Extend two-phase borrows to apply to method receiver autorefs
Fixes#48598 by permitting two-phase borrows on the autorefs created when functions and methods.
- On mismatch between impl and trait method, point at the trait
signature.
- Point only at the method signature instead of the whole body on
trait/impl mismatch errors.
When refering to named lifetime conflict, point only at the method's
signature span instead of the entire method.
When the expected and found sup and sub traces are the same, avoid
redundant text.
