[flang-commits] [flang] [flang][cuda] Split realloc for host-accessible CUDA Fortran allocatable assigns (PR #224195)

[via flang-commits]

https://github.com/jeanPerier commented:

Please harden derived type handling inside allocateAndInitNewStorage, LGTM otherwise.

https://github.com/llvm/llvm-project/pull/224195