It is ambiguous whether this should produce several `.span_suggestions()`
calls or one `.multipart_suggestions()` call.

The check previously matched this, and suggested adding a missing
`struct`:

pub Foo(...):

It was probably intended to match this instead (semicolon instead of
colon):

pub Foo(...);

This is required in order to support translatable diagnostics.

#[derive(Subdiagnostic)] does not allow multiple subdiagnostics on one
variant, as in NonItemInItemListSub::Other.

Because the old spelling is widely used, some projects may need time to
migrate their uses to the new triple spelling. The old spelling may
eventually be removed altogether.

…flatten_doc, r=cuviper

Add example for iterator_flatten

Adds an Example to iterator_flatten
Fixes rust-lang#82687

… r=davidtwco

rustc_parse: diagnostics migration, v4

This is all the `rustc_parse` migrations I have in store right now; unfortunately life is pretty busy right now, so I won't be able to do much more in the near future.

cc rust-lang#100717

r? `@davidtwco`

…agisa

Enable atomic cas for bpf targets

It seems like LLVM now supports it.

https://reviews.llvm.org/D72184 - the PR in LLVM

…=oli-obk

Fix bad import suggestion with nested `use` tree

Fixes rust-lang#105566
Fixes rust-lang#105373

Ideally, we'd find some way to turn these into structured suggestions -- perhaps on a separate line as a different `use` statement, but I have no idea how to access the span for the whole `use` from this point in the import resolution code.

…-compare, r=oli-obk

No need to take opaques in `check_type_bounds`

`InferCtxt` already has its defining use anchor set to err

r? `@oli-obk`

…didates, r=oli-obk

Prefer non-`[type error]` candidates during selection

Fixes rust-lang#102130
Fixes rust-lang#106351

r? types

note: Alternatively we could filter out error where-clauses during param-env construction? But we still need to filter out impls with errors during `match_impl`, I think.

…erbose, r=estebank

Render missing generics suggestion verbosely

It's a bit easier to read like this, especially ones that are appending new generics onto an existing list, like ": `, T`" which render somewhat poorly inline.

Also don't suggest `dyn` as a type parameter to add, even if technically that's valid in edition 2015.

…e, r=tmandry

Accept old spelling of Fuchsia target triples

The old spelling of Fuchsia target triples was changed in rust-lang#106429 to add a proper vendor. Because the old spelling is widely used, some projects may need time to migrate their uses to the new triple spelling. The old spelling may eventually be removed altogether.

r? `@tmandry`

update test for inductive canonical cycles

the previous test always resulted in a cycle 😅 cc rust-lang/chalk#787.

I checked with rust-lang#102713 and this is the only test which fails with that PR.

r? `@jackh726`

…Gomez

rustdoc: merge common CSS for `a`