Error when running matlab setup file for bbob-biobj-matlab-smsemoa example

[paulduf]

When I run the "setup.m" file in the example folder I have compilation errors (with g++):

For online documentation, see https://www.mathworks.com/support
For product information, visit www.mathworks.com.

compiling Hypervolume.cpp...Building with 'g++'.
Error using mex
/home/pol/workspace/1+1_mo_cma_es/hv.cpp: In function ‘void mexFunction(int, mxArray**, int, const mxArray**)’:
/home/pol/workspace/1+1_mo_cma_es/hv.cpp:31:32: error: cannot convert ‘const mwSize* {aka const long unsigned int*}’ to ‘const int*’ in assignment
dims = mxGetDimensions(prhs[0]);
^
/home/pol/workspace/1+1_mo_cma_es/hv.cpp:37:34: error: cannot convert ‘const mwSize* {aka const long unsigned int*}’ to ‘const int*’ in assignment
dims_r = mxGetDimensions(prhs[1]);
^

Error in setup (line 8)
mex -I. hv.cpp Hypervolume.cpp

I use Debian 9 and g++ (Debian 6.3.0-18+deb9u1) 6.3.0 20170516.

[ttusar]

I can duplicate the problem

[nikohansen]

I am slightly lost which step of the installation we are talking about. Is it that

python do.py run-matlab

calls setup.m which then fails?

[brockho]

The problem seems to lie in the hypervolume code that is borrowed from the Shark library and has nothing to do directly with COCO.

[paulduf]

In the do.py file this refers to the run-matlab-sms method. but when I run this, the program is stuck after prnting:

RUN matlab -nodisplay -nosplash -r setup, exit in code-experiments/examples/bbob-biobj-matlab-smsemoa

The error I show comes from a direct call matlab -nodisplay -nosplash -r setup

[ttusar]

The SMS-EMOA example fails. I already have a fix.

[paulduf]

The problem seems to lie in the hypervolume code that is borrowed from the Shark library and has nothing to do directly with COCO.

I agree but then it should be either solved or removed from the examples ?

[brockho]

If @ttusar has a fix, that's the best way to proceed. We had never tested the Matlab code under a non-Windows non-MacOS operating system (and therefore also never claimed that it should work :-), see here). But, of course, many thanks for catching the error.

[ttusar]

I have just created a pull request with the fix (#1854). When the tests pass, we need to merge it into the development branch.

[ttusar]

Merged