[flang-commits] [flang] [flang][cuda] Resolve generic tie‑breaks using per‑arg CUDA distance (PR #181038)

[Valentin Clement バレンタイン クレメン via flang-commits]

https://github.com/clementval approved this pull request.

LGTM. Maybe just add a comment that explain why we do it this way. 

https://github.com/llvm/llvm-project/pull/181038