summaryrefslogtreecommitdiffstats
path: root/src/test/ui/traits/reservation-impl/non-lattice-ok.rs
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-17 12:02:58 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-17 12:02:58 +0000
commit698f8c2f01ea549d77d7dc3338a12e04c11057b9 (patch)
tree173a775858bd501c378080a10dca74132f05bc50 /src/test/ui/traits/reservation-impl/non-lattice-ok.rs
parentInitial commit. (diff)
downloadrustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.tar.xz
rustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.zip
Adding upstream version 1.64.0+dfsg1.upstream/1.64.0+dfsg1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/test/ui/traits/reservation-impl/non-lattice-ok.rs')
-rw-r--r--src/test/ui/traits/reservation-impl/non-lattice-ok.rs59
1 files changed, 59 insertions, 0 deletions
diff --git a/src/test/ui/traits/reservation-impl/non-lattice-ok.rs b/src/test/ui/traits/reservation-impl/non-lattice-ok.rs
new file mode 100644
index 000000000..a71051243
--- /dev/null
+++ b/src/test/ui/traits/reservation-impl/non-lattice-ok.rs
@@ -0,0 +1,59 @@
+// build-pass
+
+// Check that a reservation impl does not force other impls to follow
+// a lattice discipline.
+
+// Why did we ever want to do this?
+//
+// We want to eventually add an `impl<T> From<!> for T` impl. That impl conflicts
+// with existing impls - at least the `impl<T> From<T> for T` impl. There are
+// 2 ways we thought of for dealing with that conflict:
+//
+// 1. Using specialization and doing some handling for the
+// overlap. The current thought is to require ["intersection
+// impls"][ii], specialization", which means providing an
+// (higher-priority) impl for the intersection of every 2 conflicting
+// impls that determines what happens in the intersection case. That's
+// the first thing we thought about - see e.g.
+// https://github.com/rust-lang/rust/issues/57012#issuecomment-452150775
+//
+// 2. The other way is to notice that `impl From<!> for T` is basically a
+// marker trait since its only method is uninhabited, and allow for "marker
+// trait overlap", where the conflict "doesn't matter" because it can't
+// actually cause any ambiguity.
+//
+// Now it turned out lattice specialization doesn't work it, because an
+// `impl<T> From<T> for Smaht<T>` would require an `impl From<!> for Smaht<!>`,
+// breaking backwards-compatibility in a fairly painful way. So if we want to
+// go with a known approach, we should go with a "marker trait overlap"-style
+// approach.
+//
+// [ii]: https://smallcultfollowing.com/babysteps/blog/2016/09/24/intersection-impls/
+
+#![feature(rustc_attrs, never_type)]
+
+trait MyTrait {}
+
+impl MyTrait for ! {}
+
+trait MyFrom<T> {
+ fn my_from(x: T) -> Self;
+}
+
+// Given the "normal" impls for From
+#[rustc_reservation_impl="this impl is reserved"]
+impl<T> MyFrom<!> for T {
+ fn my_from(x: !) -> Self { match x {} }
+}
+
+impl<T> MyFrom<T> for T {
+ fn my_from(x: T) -> Self { x }
+}
+
+// ... we *do* want to allow this common pattern, of `From<!> for MySmaht<T>`
+struct MySmaht<T>(T);
+impl<T> MyFrom<T> for MySmaht<T> {
+ fn my_from(x: T) -> Self { MySmaht(x) }
+}
+
+fn main() {}