[NO MRG] Test CI

[jakirkham]

Opening to test CI

[jakirkham]

Seeing this error in various forms on CI

/__w/pynvjitlink/pynvjitlink/test_binary_generation /__w/pynvjitlink/pynvjitlink
GPU CC: 70
Alternative CC: 80
# Compile all test objects
nvcc -O3 -rdc true -gencode arch=compute_70,code=sm_70 --cubin -o ../pynvjitlink/tests/undefined_extern.cubin undefined_extern.cu
<command-line>: fatal error: cuda_runtime.h: No such file or directory
compilation terminated.
make: *** [Makefile:41: all] Error 1

Looks like we are picking up cuda-nvcc version 12.4 from the nvidia channel

cuda-nvcc                 12.4.131                      0    nvidia

This was before important packaging fixes were ported over. Hence why cuda-cudart is not being pulled in by cuda-nvcc

[bdice]

I'm exploring strict channel priority as a potential solution in #109.

[bdice]

#109 resolved the CI issue, so this can be closed.

[jakirkham]

Thanks Bradley! 🙏