Do you have any idea how to solve the following error?

[yxchng]

>> mex compute_dense_overlap.cc CXXFLAGS='$CXXFLAGS -fopenmp' ...
LDFLAGS='$LDFLAGS -fopenmp' CXXOPTIMFLAGS='-O3 -DNDEBUG' 
Building with 'g++'.
Error using mex
/home/yxchng/tiny/utils/compute_dense_overlap.cc: In function ‘void mexFunction(int, mxArray**,
int, const mxArray**)’:
/home/yxchng/tiny/utils/compute_dense_overlap.cc:70:77: error: cannot convert ‘const int*’ to
‘const size_t* {aka const long unsigned int*}’ for argument ‘2’ to ‘mxArray*
mxCreateNumericArray(size_t, const size_t*, mxClassID, mxComplexity)’
   mxArray *mx_overlap = mxCreateNumericArray(4, dims, mxDOUBLE_CLASS, mxREAL);
                                                                             ^
/home/yxchng/tiny/utils/compute_dense_overlap.cc:75:38: warning: narrowing conversion of
‘(((double)((int)vsy + -1) * ((double)zmy)) + 1.0e+0)’ from ‘double’ to ‘const int’ inside { }
[-Wnarrowing]
   const int tmp_dims[] = {(vsy-1)*zmy+1, (vsx-1)*zmx+1, nt};
                                      ^
/home/yxchng/tiny/utils/compute_dense_overlap.cc:75:53: warning: narrowing conversion of
‘(((double)((int)vsx + -1) * ((double)zmx)) + 1.0e+0)’ from ‘double’ to ‘const int’ inside { }
[-Wnarrowing]
   const int tmp_dims[] = {(vsy-1)*zmy+1, (vsx-1)*zmx+1, nt};
                                                     ^
/home/yxchng/tiny/utils/compute_dense_overlap.cc:76:85: error: cannot convert ‘const int*’ to
‘const size_t* {aka const long unsigned int*}’ for argument ‘2’ to ‘mxArray*
mxCreateNumericArray(size_t, const size_t*, mxClassID, mxComplexity)’
   mxArray *mx_tmp_overlap = mxCreateNumericArray(3, tmp_dims, mxDOUBLE_CLASS, mxREAL);
                                                                                  ^

[yxchng]

Solved. You are using 32bit system I think, so I need to compile mex with "mex -DMX_COMPACT_32"

[peiyunh]

I'm not sure why you had to compile with that option, but I'm glad you solved it. My system is 64-bit. What's the version of MATLAB that you are using?

[myungsub]

I had the same error, but couldn't solve it by adding the 32-bit option.
Instead, solved it by changing to the cross-platform type mwSize in the mex file compute_dense_overlap.cc. (Followed this)
On line 67:
const int dims[] = {vsy, vsx, nt, ng}; --> const mwSize dims[] = {vsy, vsx, nt, ng};
Same on line 75 with tmp_dims[].

I'm using MATLAB 2017a on Ubuntu 16.04. Hope it helps someone else who use this code!

[varunagrawal]

@peiyunh can you please add this to the README?
I encountered the same issue and @myungsub's solution worked.
I'm using MATLAB 2017b 64-bit and had this issue as well.

[peiyunh]

@varunagrawal Yes! Thanks for reminding me to do so.

@myungsub Thanks for pointing out the solution. I've updated the source code.