Stabilize RFC3324 dyn upcasting coercion by Urgau · Pull Request #118133 · rust-...

r? @b-naber

(rustbot has picked a reviewer for you, use r? to override)

The Miri subtree was changed

cc @rust-lang/miri

Should we keep the deref_into_dyn_supertrait lint around, maybe changing its wording? These impls probably don't do what the user expects them to, given that they can't be used implicitly. I feel like pointing out "hey, this impl is now implicit via a coercion, consider removing it" is preferable to completely removing the lint

Sure, if those impls don't do what the user expects we should tell them.

What about this diagnostic change?

- warning: `dyn Foo<'_>` implements `Deref` with supertrait `Bar<'_>` as target
+ warning: `dyn Foo<'_>` implements `Deref` with supertrait `Bar<'_>` as target is shadowed by coercion
   --> $DIR/migrate-lint-deny-regions.rs:8:1
    |
 LL | impl<'a> Deref for dyn Foo<'a> {
    | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 ...
 LL |     type Target = dyn Bar<'a>;
    |     -------------------------- target type is set here
    |
-   = warning: this will change its meaning in a future release!
-   = note: for more information, see issue #89460 <https://github.com/rust-lang/rust/issues/89460>
+   = help: consider removing this implementation in favour for the implicit coercion

maybe something like

- warning: `dyn Foo<'_>` implements `Deref` with supertrait `Bar<'_>` as target
+ warning: this `Deref` implementation is covered by the implicit super-trait coercion
   --> $DIR/migrate-lint-deny-regions.rs:8:1
    |
 LL | impl<'a> Deref for dyn Foo<'a> {
    | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 ...
 LL |     type Target = dyn Bar<'a>;
    |     -------------------------- target type is set here
    |
-   = warning: this will change its meaning in a future release!
-   = note: for more information, see issue #89460 <https://github.com/rust-lang/rust/issues/89460>
+   = help: consider removing this implementation or replacing it with a method instead

something like this maybe? 🤔

Done. I've restored the lint, removed the future incompatibility mark and changed the diagnostic output as discussed.

@bors delegate+

✌️ @Urgau, you can now approve this pull request!

If @WaffleLapkin told you to "r=me" after making some further change, please make that change, then do @bors r=@WaffleLapkin

Also, could you open a reference PR editing this page?

@crlf0710 already has one opened rust-lang/reference#1259 (from his previous attempt as stabilizing this feature).

📌 Commit 4c2d6de has been approved by WaffleLapkin

It is now in the queue for this repository.

⌛ Testing commit 4c2d6de with merge 6d2b84b...

A job failed! Check out the build log: (web) (plain)

☀️ Test successful - checks-actions
Approved by: WaffleLapkin
Pushing 6d2b84b to master...

@rustbot label +relnotes

Finished benchmarking commit (6d2b84b): comparison URL.

Overall result: ❌ regressions - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

Max RSS (memory usage)

Results

Cycles

Results

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 675.017s -> 675.132s (0.02%)
Artifact size: 313.79 MiB -> 313.79 MiB (0.00%)