Note in the tests how the code to compute partials of matrix and gradients is soooo much simpler to read! It's the same code as for normal VectorN =)
I moved some functions around and it seems that the diff got confused. I've added a screenshot of the diff. Mainly, the big difference is that the function which needs to be derived has a signature which directly returns the evaluation of the function and its matrix of partials.
Note in the tests how the code to compute partials of matrix and gradients is soooo much simpler to read! It's the same code as for normal VectorN =)
I moved some functions around and it seems that the diff got confused. I've added a screenshot of the diff. Mainly, the big difference is that the function which needs to be derived has a signature which directly returns the evaluation of the function and its matrix of partials.
Closes #18