I just noticed that compiling nfft with Octave support fails in fastsummex in the branch develop and release/3.4.0 with the code from #39 / #47. GCC error message:
fastsummex.c:267:27: error: 'INT_MAX' undeclared (first use in this function)
if (mxGetM(prhs[2]) > INT_MAX)
The INT_MAX macro seems to be available via the MATLAB MEX interface but not by the Octave MEX interface.
The INT_MAX macro is defined for glibc in limits.h and this is included by MATLAB tmwtypes.h included by matrix.h included by mex.h.
I just noticed that compiling nfft with Octave support fails in fastsummex in the branch develop and release/3.4.0 with the code from #39 / #47. GCC error message:
The
INT_MAX
macro seems to be available via the MATLAB MEX interface but not by the Octave MEX interface. TheINT_MAX
macro is defined for glibc inlimits.h
and this is included by MATLABtmwtypes.h
included bymatrix.h
included bymex.h
.