yxsamliu wrote: Since this changes HIP device code to accept zero-sized arrays, can we keep a positive SemaHIP test for the new behavior instead of only deleting the old negative test? It would be useful to cover a normal device local array, a templated array, `__host__ __device__`, and `__shared__ T x[0]` so this does not regress later.
https://github.com/llvm/llvm-project/pull/195716 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
