Open robertgj opened 4 years ago
Indeed, I have to admit that the Octave interface is practically not maintained anymore; however, the MATLAB interface should work under Octave, it is sufficient to call the configure script with:
./configure MEX="mkoctfile --mex"
I am inclined to remove the Octave files in future releases, and use the mex file only (to avoid duplicating the effort). Could you try if that works for you?
On Wed, 02 Sep 2020 03:54:30 -0700 Leonardo Robol notifications@github.com wrote:
Indeed, I have to admit that the Octave interface is practically not maintained anymore; however, the MATLAB interface should work under Octave, it is sufficient to call the configure script with: ./configure MEX=mkoctfile --mex I am inclined to remove the Octave files in future releases, and use the mex file only (to avoid duplicating the effort). Could you try if that works for you?
Hi,
There seems to be a conflict between mps_boolean and the standard library boolean type:
make[3]: Entering directory '/home/robj/TMP/MPSolve-master/examples/matlab'
Best regards,
Rob Jenssen
There is a compile time option to avoid these defines, since they are in the mex headers as well. For MATLAB it works, and for Octave it used to work; apparently it doesn't anymore. I'll investigate this as soon as I find some time.
Firstly, thank you for the mpsolve package.
Building the mpsolve-3.1.7 zipfile downloaded from github.com:
Testing:
Unfortunately, the version of MPSolve available at https://numpi.dm.unipi.it/_media/software/mpsolve/mpsolve-3.2.1.tar.gz fails to compile under octave-5.2.0 because octave_support.h, mps_polyeig.m and DESCRIPTION are missing from examples/octave. However, at the command line:
Compiling with debugging symbols:
and running valgrind: