These lines:
|
HH=$(spack find --format "{version} {hash:7}" codetools | sort -rn | head -1 | awk '{print $2}' ) |
|
H1=$(spack find --format "{version} {hash:7}" llvm | sort -rn | head -1 | awk '{print $2}' ) |
|
H1=$(spack find --format "{version} {hash:7}" llvm | sort -rn | head -1 | awk '{print $2}' ) |
Have a potential future failure modes. If we have > 1 hashes with the same version number, we will get a random hash. Depending on details that might or might not be OK.
One obvious option is to just put everything in our spack view so that there is no need to load it. Need to understand other options and cost/benefit of each.