From c71ede368ce71888d8fa13708bd2cb4df9e840a3 Mon Sep 17 00:00:00 2001
From: Ben Kimock <kimockb@gmail.com>
Date: Thu, 29 Aug 2024 15:39:05 -0400
Subject: [PATCH] Add a test for trait solver overflow in MIR inliner cycle
detection
---
tests/mir-opt/inline/type_overflow.rs | 32 +++++++++++++++++++++++++++
1 file changed, 32 insertions(+)
create mode 100644 tests/mir-opt/inline/type_overflow.rs
diff --git a/tests/mir-opt/inline/type_overflow.rs b/tests/mir-opt/inline/type_overflow.rs
new file mode 100644
index 0000000000000..bfd9e71b9e73f
--- /dev/null
+++ b/tests/mir-opt/inline/type_overflow.rs
@@ -0,0 +1,32 @@
+// This is a regression test for one of the problems in #128887; it checks that the
+// strategy in #129714 avoids trait solver overflows in this specific case.
+
+// skip-filecheck
+//@ compile-flags: -Zinline-mir
+
+pub trait Foo {
+ type Associated;
+ type Chain: Foo<Associated = Self::Associated>;
+}
+
+trait FooExt {
+ fn do_ext() {}
+}
+impl<T: Foo<Associated = f64>> FooExt for T {}
+
+#[allow(unconditional_recursion)]
+fn recurse<T: Foo<Associated = f64>>() {
+ T::do_ext();
+ recurse::<T::Chain>();
+}
+
+macro_rules! emit {
+ ($($m:ident)*) => {$(
+ pub fn $m<T: Foo<Associated = f64>>() {
+ recurse::<T>();
+ }
+ )*}
+}
+
+// Increase the chance of triggering the bug
+emit!(m00 m01 m02 m03 m04 m05 m06 m07 m08 m09 m10 m11 m12 m13 m14 m15 m16 m17 m18 m19);