================
@@ -987,6 +988,12 @@ void CastOperation::CheckDynamicCast() {
return;
}
+ // Similarly, dynamic_cast is not available in CUDA device code, except for
+ // dynamic_cast to void*.
+ if (Self.getLangOpts().CUDA && !DestPointee->isVoidType())
----------------
steffenlarsen wrote:
Initially I figured we should just follow the precedence set by the case above,
but with the most recent changes I tend towards agreeing with your stance.
@yxsamliu - Objections to also disallowing `dynamic_cast<void*>` for HIP/CUDA
for consistency?
https://github.com/llvm/llvm-project/pull/228469
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
