Describe the bug
It appears that the configuration summary does not sanitize the cmake build options. For example, if HDF5 is built with -DHDF5_ENABLE_PARALLEL=42, the configuration summary shown by h5pcc will be Parallel HDF5: 42. Since cmake accepts so many values as true (e.g, ON, Y, TRUE, YES, 1, 42, 3.14, etc.), this becomes an issue. For example, this spack issue (spack/spack#35546) is caused by spack using ON to enable cmake build options whereas CMake's FindHDF5 module checking yes for Parallel HDF5 support.
Expected behavior
Maybe HDF5 should always write YES or ON or yes (because that's what CMake checks) for all cmake true values.
Platform (please complete the following information)
- HDF5 version (if building from a maintenance branch, please include the commit hash)
- OS and version
- Compiler and version
- Build system (e.g. CMake, Autotools) and version
CMake
- Any configure options you specified
- MPI library and version (parallel HDF5)
Describe the bug
It appears that the configuration summary does not sanitize the cmake build options. For example, if HDF5 is built with
-DHDF5_ENABLE_PARALLEL=42, the configuration summary shown byh5pccwill beParallel HDF5: 42. Since cmake accepts so many values as true (e.g,ON,Y,TRUE,YES,1,42,3.14, etc.), this becomes an issue. For example, this spack issue (spack/spack#35546) is caused by spack usingONto enable cmake build options whereas CMake's FindHDF5 module checkingyesfor Parallel HDF5 support.Expected behavior
Maybe HDF5 should always write
YESorONoryes(because that's what CMake checks) for all cmake true values.Platform (please complete the following information)
CMake