From d50cee3fa401246061c106248b749be03e108298 Mon Sep 17 00:00:00 2001 From: Pierre-Marie Pédrot Date: Mon, 12 Feb 2018 16:58:52 +0100 Subject: Adding a test for variance subtyping in the module system. --- test-suite/modules/cumpoly.v | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) create mode 100644 test-suite/modules/cumpoly.v (limited to 'test-suite/modules') diff --git a/test-suite/modules/cumpoly.v b/test-suite/modules/cumpoly.v new file mode 100644 index 000000000..654b86cb4 --- /dev/null +++ b/test-suite/modules/cumpoly.v @@ -0,0 +1,19 @@ +Set Universe Polymorphism. + +(** Check that variance subtyping is respected. The signature T is asking for + invariance, while M provide an irrelevant implementation, which is deemed + legit. + + There is currently no way to go the other way around, so it's not possible + to generate a counter-example that should fail with the wrong subtyping. +*) + +Module Type T. +Parameter t@{i|Set <= i} : Type@{i}. +Cumulative Inductive I@{i|Set <= i} : Type@{i} := C : t@{i} -> I. +End T. + +Module M : T. +Definition t@{i|Set <= i} : Type@{i} := nat. +Cumulative Inductive I@{i|Set <= i} : Type@{i} := C : t@{i} -> I. +End M. -- cgit v1.2.3