[flang-commits] [flang] [llvm] [flang-rt] Enable more runtime functions for the GPU target (PR #183649)

[Joseph Huber via flang-commits]

================
@@ -16,13 +16,13 @@
 
 // Avoid <mutex> if possible to avoid introduction of C++ runtime
 // library dependence.
-#ifndef _WIN32
+#if !defined(_WIN32)
----------------
jhuber6 wrote:

Oh yeah, I had another check here until I just inherited the pseudo lock form the CUDA build.

https://github.com/llvm/llvm-project/pull/183649