Fix mms

[hugary1995]

A few things:

1. You simply forgot a minus sign on the body force. See the updated overleaf document.
2. I simplified your kernels.
3. I modified the MMS input file to use full Dirichlet BCs -- we don't lose generality by doing so.
4. Your magic preconditioner didn't work for me, and so I changed it to lu.
5. This particular manufactured solution isn't interesting, and you won't see convergence. The error will stay at machine precision. You need to use a higher order function to examine convergence. I usually use some exponential or sinusoidal functions. I don't think your manufactured solution have to satisfy the incompressibility constraint, so that should make things easier.

[hugary1995]

I don't think your manufactured solution have to satisfy the incompressibility constraint, so that should make things easier.

By this I mean you can add a body force to the pressure equation.