[PATCH] Implement CWG496: Is a volatile-qualified type really a POD?

[Richard Smith]

Looks great, thanks!


================
Comment at: test/CXX/drs/dr4xx.cpp:1187
@@ -1186,3 +1186,3 @@
 
-namespace dr496 { // dr496: no
+namespace dr496 { // dr496: yes
   struct A { int n; };
----------------
`dr496: 3.7`, please.

================
Comment at: test/SemaCXX/type-traits.cpp:1863-1865
@@ -1859,2 +1862,5 @@
   { int arr[T(__is_trivially_copyable(DerivesHasRef))]; }
+  { int arr[T(__is_trivially_copyable(NonTrivialDefault))]; }
+  { int arr[T(__is_trivially_copyable(NonTrivialDefault[]))]; }
+  { int arr[T(__is_trivially_copyable(NonTrivialDefault[3]))]; }
 
----------------
Please commit this test update and the bugfix for the array case of `QualType::isTriviallyCopyableType` separately from the rest of the patch.

http://reviews.llvm.org/D7060

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/