Bug#686926: liboctave-dev: creates broken .oct files when the OpenMPI flavor of HDF5 (libhdf5-openmpi-dev) is installed

[Bradley M. Froehle]

> More practically, and more importantly, Octave users who are compiling
> extensions from within Octave are yet another step removed from both the
> compiler and the HDF5 library. They may not even be aware of HDF5 but
> the compilation and linking of their oct-file will be affected.
> Mkoctfile simply compiles extensions using the same set of flags and
> libraries that Octave itself was built with. And those extensions should
> work without needing to recompile if the HDF5 library is switched out.
> Compiling them with CXX=mpicxx has the opposite effect.

Yes. I agree completely.

Fundamentally I blame #598227 (http://bugs.debian.org/598227) which introduced the -I/usr/include/mpi flag to the mkoctfile build script.  This was done in the name of pragmatism, but it really opens up a can of worms (as you can see).  If that flag weren't there your experience would instead have been a "mpi.h" not found bug at compile time.  A quick search on Google would hopefully have pointed you towards adding CC=mpicc CXX=mpicxx to the mkoctfile invocation and there wouldn't have been future issues.

The plan, originally, was to have the octave headers depend on libhdf5-serial-dev (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=598227#29).  That would certainly resolve the issue, but it seems a bit overly restrictive.

If only there was a way for the different variants of HDF5 to be simultaneously installable…