Report 6 - Final
After my last report, I changed my plans. I wasn't able to continue with mexlib because my mentor went in holidays. I started with GIWS, because I was stucked in some last problems in mexlib.
During these weeks, I learned about Python C/API and how to generate code like GIWS do. It looks easier than JNI code, but I did just basic features in a experimental way. I sent some patches with it to Sylvestre LEDRU.
Me and GSoC
I really enjoyed GSoC and Scilab project. In three months I was able to go deep in Scilab code to be able to implement mexlib. It was not easy, but was an awesome challenge. I learned a lot about C++, git and Scilab. I never worked with C++ or Scilab, and knew just a bit of git before GSoC.
I had some trouble with most complicated functions of mexlib, but even my mentor didn't know how to implement some functions. I'm working on YaSp and some things were done during GSoC, like dynamic link, which enabled unit tests to mexlib. Before I can test anything, I wrote a lot of code, but part of it I wrote without know if I'm doing it right and completely rewrote after unit test it. It made mexlib project harder than it should be.