Bugfix: fix a bug where invalid data would be read through the Fortran API
The new Fortran API uses pointer to ensure correct semantics. We have to test if this is correctly supported by compilers.
Also fixes a bug in the example where ii would not be initialized to 0, leading to a bug in Release mode
Fixes #7 (closed)
@croussel can you check this works on Poincare & Curie with the Intel compiler please