Would something like:
#include "numpy/arrayobject.h"
// for compatibility with Numpy version <= 1.6
#if NPY_FEATURE_VERSION < 0x0007
#define NPY_ARRAY_FARRAY NPY_FARRAY
// other defines for deprecated stuff
// ...
#endif
Be robust enough ?
2014-02-28 14:31 GMT+00:00 G
Hi everyone,
I have got code for some python wrappers of a scientific library which
needs to support both Numpy 1.6 and later versions.
The build of the wrapper (using swig) stopped working because of the
deprecated API introduced in v1.7. The error only concerns the renaming of
some macros from