15:52
I have an idea for the compiler tester. I can have WinterCoat run the code that is output by the compiler. Then this way I will know if it works correctly. It can also support various register counts and instruction sizes too potentially. That way it can test a bunch of things on the same piece of code. This would be useful and it makes WinterCoat more useful.
18:08
The LCDUI code is definitely old and needs to be updated for how the new code works, with the better services now.
18:13
For the kernel, I am going to split the server code to another class and have that instead. That way for each created task, services and such can be initialized accordingly.
18:38
I think what could be better for the kernel instead of abstract interfaces would be a native kernel infrastructure kind of thing where the kernel and such refers to those things for task creation and other things as such.
19:38
This definitely should be better, along with having stuff such as a primitive process and thread set.
19:43
If I remove the ability for the Java SE environment to fork new processes, the design of the kernel could be a bit simpler. Although it could be done. For now I will intend to write it for real environments.
20:51
I need a class to easily extract system call arguments to the required types and casted for ease of use.