Compiling LIBSVM for Matlab R2016a - matlab

Question

I want to use LiBSVM with Matlab R2016a under Windows 7 64 Bit with the MinGW64 compiler and the precompiled mex files didn't work for me.
When I tried to run the standard make file that comes with LibSVM the following error occurs:
*Building with 'MinGW64 Compiler (C)'.
Error using mex
gcc: error: -fexceptions: No such file or directory
*
The makefile looks like this:
mex CFLAGS="\$CFLAGS -std=c99" -largeArrayDims libsvmread.c
mex CFLAGS="\$CFLAGS -std=c99" -largeArrayDims libsvmwrite.c
mex CFLAGS="\$CFLAGS -std=c99" -I.. -largeArrayDims svmtrain.c ../svm.cpp svm_model_matlab.c
mex CFLAGS="\$CFLAGS -std=c99" -I.. -largeArrayDims svmpredict.c ../svm.cpp svm_model_matlab.c
In another answer to a stackoverflow question I saw that somebody used this code for compiling:
mex -O -largeArrayDims libsvmread.c
mex -O -largeArrayDims libsvmwrite.c
mex -O -I.. -largeArrayDims svmtrain.c ../svm.cpp svm_model_matlab.c
mex -O -I.. -largeArrayDims svmpredict.c ../svm.cpp svm_model_matlab.c
This also works for me. What is the difference between the two ways of compiling LibSVM and are there any disadvantages of using the second option?
edit:
this is the so question I mentioned: http://ch.mathworks.com/matlabcentral/newsreader/view_thread/297195
after doing a bit more research I understood that -std=c99 selects a c language version which is not fully implemented for gcc under windows and -O optimizes the object code. In http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html#f213 the developers mention that they actually did not use any c99 languge features.
So compiling the library should have no disadvantages, right?

Answer 1

The "error: -fexceptions: No such file or directory." May also be solved by changing CFLAGS to COMPFLAGS.
Credit https://github.com/cjlin1/libsvm/issues/55