about summary refs log tree commit diff
path: root/src/tools
diff options
context:
space:
mode:
authorGuillaume Gomez <guillaume1.gomez@gmail.com>2024-04-05 16:38:51 +0200
committerGitHub <noreply@github.com>2024-04-05 16:38:51 +0200
commit02ee8a8ceeff811adfb065028a48c1947d97ef91 (patch)
treee03896d1ad6a90866cf7432102cfae1b4a53e3fd /src/tools
parentf2f8d8b722fdacc6a6e02c2289e3e36571749a68 (diff)
parent55e46612c1ccceb30a7a6acf11fd485f34e393e5 (diff)
downloadrust-02ee8a8ceeff811adfb065028a48c1947d97ef91.tar.gz
rust-02ee8a8ceeff811adfb065028a48c1947d97ef91.zip
Rollup merge of #123350 - compiler-errors:async-closure-by-move, r=oli-obk
Actually use the inferred `ClosureKind` from signature inference in coroutine-closures

A follow-up to https://github.com/rust-lang/rust/pull/123349, which fixes another subtle bug: We were not taking into account the async closure kind we infer during closure signature inference.

When I pass a closure directly to an arg like `fn(x: impl async FnOnce())`, that should have the side-effect of artificially restricting the kind of the async closure to `ClosureKind::FnOnce`. We weren't doing this -- that's a quick fix; however, it uncovers a second, more subtle bug with the way that `move`, async closures, and `FnOnce` interact.

Specifically, when we have an async closure like:
```
let x = Struct;
let c = infer_as_fnonce(async move || {
  println!("{x:?}");
}
```

The outer closure captures `x` by move, but the inner coroutine still immutably borrows `x` from the outer closure. Since we've forced the closure to by `async FnOnce()`, we can't actually *do* a self borrow, since the signature of `AsyncFnOnce::call_once` doesn't have a borrowed lifetime. This means that all `async move` closures that are constrained to `FnOnce` will fail borrowck.

We can fix that by detecting this case specifically, and making the *inner* async closure `move` as well. This is always beneficial to closure analysis, since if we have an `async FnOnce()` that's `move`, there's no reason to ever borrow anything, so `move` isn't artificially restrictive.
Diffstat (limited to 'src/tools')
-rw-r--r--src/tools/miri/tests/pass/async-closure-captures.rs34
-rw-r--r--src/tools/miri/tests/pass/async-closure-captures.stdout4
2 files changed, 38 insertions, 0 deletions
diff --git a/src/tools/miri/tests/pass/async-closure-captures.rs b/src/tools/miri/tests/pass/async-closure-captures.rs
index 3e33de32efb..cac26bfe146 100644
--- a/src/tools/miri/tests/pass/async-closure-captures.rs
+++ b/src/tools/miri/tests/pass/async-closure-captures.rs
@@ -88,4 +88,38 @@ async fn async_main() {
         };
         call_once(c).await;
     }
+
+    fn force_fnonce<T>(f: impl async FnOnce() -> T) -> impl async FnOnce() -> T {
+        f
+    }
+
+    // Capture something with `move`, but infer to `AsyncFnOnce`
+    {
+        let x = Hello(6);
+        let c = force_fnonce(async move || {
+            println!("{x:?}");
+        });
+        call_once(c).await;
+
+        let x = &Hello(7);
+        let c = force_fnonce(async move || {
+            println!("{x:?}");
+        });
+        call_once(c).await;
+    }
+
+    // Capture something by-ref, but infer to `AsyncFnOnce`
+    {
+        let x = Hello(8);
+        let c = force_fnonce(async || {
+            println!("{x:?}");
+        });
+        call_once(c).await;
+
+        let x = &Hello(9);
+        let c = force_fnonce(async || {
+            println!("{x:?}");
+        });
+        call_once(c).await;
+    }
 }
diff --git a/src/tools/miri/tests/pass/async-closure-captures.stdout b/src/tools/miri/tests/pass/async-closure-captures.stdout
index a0db6d236fe..42a7999b2dc 100644
--- a/src/tools/miri/tests/pass/async-closure-captures.stdout
+++ b/src/tools/miri/tests/pass/async-closure-captures.stdout
@@ -8,3 +8,7 @@ Hello(3)
 Hello(4)
 Hello(4)
 Hello(5)
+Hello(6)
+Hello(7)
+Hello(8)
+Hello(9)