New SoT/QP DSL
I had to delete and completely redo the original qp-dsl. The reason is quite simple: the original language and its concepts wouldn't work well with the common constriant representation. Furthermore, the way I did it was a nightmare for code/text generation. This is going to be cleaner, easier to generate, and above all compliant to CCR.
- The most interesting one is rethinking the priority concept. No longer priority is a property of the task. Instead it is an atribute given by the SoT. This concept, however, doesn't easily scale up to Herman's notion of weighted priorities. Although it is still possible.
- Add feedback gains to tasks (are they in all tasks?)
- What about the situations where I have only one task? (regarding priority scheme)