Closed thompsonnoahe closed 1 year ago
@thompsonnoahe I'm looking into this now. Looks like CUDA reserves __noinline__
for it's own purposes, causing this issue: https://stackoverflow.com/questions/70301375/noinline-macro-conflict-between-glib-and-cuda/
Alright I've pushed a fix for this. I'm not easily able to test this myself currently so if you still have further issues feel free to re-open it and I'll have a deeper look into it. Thanks for the report!
Environment
toml++ version and/or commit hash: v3.3.0
Compiler: nvcc V10.2.300
C++ standard mode: 17
Target arch: aarch64
Library configuration overrides: None
Relevant compilation flags: --compiler-options -std=c++17
Describe the bug
During compilation of
toml++
with nvcc, NVIDIA's CUDA Compiler, the compiler errors out with the following status:Additional information
NVIDIA CUDA Runtime v10.2.