[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

[Aaron Ballman via Phabricator via cfe-commits]

aaron.ballman added a comment.

In https://reviews.llvm.org/D54258#1295158, @richardmembarth wrote:

> There are external tools (e.g. hipacc <https://hipacc-lang.org>) that generate Clang AST. This AST uses `LangAS` annotations and emits incorrect memory space specifiers for CUDA when pretty-printed.


That's good to know!

> We would need a different frontend that annotates LangAS::cuda_shared.

Do you happen to know why this behaves the way it does? e.g., is the bug that the frontend is annotating incorrectly and forces codegen to work around it, and fixing the frontend to annotate properly lets us remove some workarounds and fixes your issue?


Repository:
  rC Clang

https://reviews.llvm.org/D54258