KibaBase

Wednesday was the most productive coding session ever for RubyTet. It was also a session devoid of Space Fighter Ace. I coded for 4 hours with agonizing slow progress. That’s mostly the fault of unit testings. I inevitably have to break some so that I can make the proper changes to the new code. This also break up the unit tests since I have to correct the assumption or update the code to reflect new usage.

For the most part, I have finished the work on loading the configuration file and read them into configuration field. I found that my translation algorithm was so robust that I didn’t even need a transformation algorithm. I might be wrong on that later, but it is my gut feeling that the algorithm will work with some modification.

The old part of the codebase and  the new part of the codebase are being seamlessly integrated. Some part of the codebase are simply legacy, especially the shape changing code, so I’ll rip them all out when the file-based configuration system is done.  It might even be possible that I don’t need all the l.rb and the i.rb, etc.

For Wednesday,  I learn a very a important coding lesson about unit testing. They are actually  a trade off. In exchange for a lot of time wasted on correcting unit tests, I built a ironcald codebase from which I can built my changes upon. It also help me seamlessly integrate the new code without changing its underlying behaviors.

A few more days to go before I am finished with the  28 hours work week. I hope to at least finish RubyTet, if nothing else.

Happy hacking.

~Kiba

Filed under:
devlog by admin