diff options
| author | Stephan T. Lavavej <stl@exchange.microsoft.com> | 2017-01-17 22:24:45 +0000 |
|---|---|---|
| committer | Stephan T. Lavavej <stl@exchange.microsoft.com> | 2017-01-17 22:24:45 +0000 |
| commit | c8fc92cc09a98fde9087b3270259e4444251a5a2 (patch) | |
| tree | 9f1185e322033621e32a9b69ead9f66e98ef72b4 /test/std/utilities | |
| parent | 37b2be9c5810d3f43146f7d20522cac2db32e7a0 (diff) | |
| download | external_libcxx-c8fc92cc09a98fde9087b3270259e4444251a5a2.tar.gz external_libcxx-c8fc92cc09a98fde9087b3270259e4444251a5a2.tar.bz2 external_libcxx-c8fc92cc09a98fde9087b3270259e4444251a5a2.zip | |
[libcxx] [test] Don't ask whether Incomplete& can be assigned to.
This is the subject of an active NB comment. Regardless of what the Working
Paper currently says, asking this question is morally wrong, because the
answer can change when the type is completed. C1XX now detects such
precondition violations and complains about them; perhaps Clang should too.
Fixes D28591.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@292281 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/std/utilities')
| -rw-r--r-- | test/std/utilities/meta/meta.unary/meta.unary.prop/is_assignable.pass.cpp | 1 |
1 files changed, 0 insertions, 1 deletions
diff --git a/test/std/utilities/meta/meta.unary/meta.unary.prop/is_assignable.pass.cpp b/test/std/utilities/meta/meta.unary/meta.unary.prop/is_assignable.pass.cpp index f4736e713..3955d4bc3 100644 --- a/test/std/utilities/meta/meta.unary/meta.unary.prop/is_assignable.pass.cpp +++ b/test/std/utilities/meta/meta.unary/meta.unary.prop/is_assignable.pass.cpp @@ -82,5 +82,4 @@ int main() // pointer to incomplete template type test_is_assignable<X<D>*&, X<D>*> (); - test_is_not_assignable<Incomplete&, Incomplete const&>(); } |