Jean-Michel Campin

There are still few minor things to tweak such as the new test in exp. front_relax and the little sub-section in the docs, but otherwise I think this is ready...

@mjlosch Thanks fro the review. I am going to change for the English word "curl" (instead "rotational", a bit too French). And regarding the comparison P/Z, you would need to...

Regarding the `front_relax` test experiment: I am leaning towards using `GM_AdvForm` in the new secondary test `in_p` but adding also the Redi part (to do), so that both the tensor...

@mjlosch Regarding difference between P and Z run after 20 time-steps, using `data.gmredi` from `input/`: I get max diff for T ~1.5e-4, and for S ~1.4e-5, with RMS (just grid-point...

I will convert the REAME file to markdown and adjust the new test "input.in_p" in front_relax. There is one thing that I wanted to do for ocean in P-coord, is...

I did not realized earlier that many S/R are named using second convention (prefix) but belong to a file that follow the first convention (suffix).

@ifenty regarding our last point, I though that, when using pkg/obcs, the "interrior masking" was taking care of this (mask out cg2d matrix coeff and RHS and initial guess so...

A -- not so closely -- related issue: I am not sure what the resulting startTime is for some OBC fields (or OBC seaice fields) after processing startDate1/2. I can...

@mjlosch Looks like S/R CTRL_MAP_INI is also no longer called (and is now empty), so could be removed as well ?

@mjlosch I agree with leaving (4) for an other PR. We could also keep the indentation changes in ctrl_set_*pack*.F (it's just that it was a bit more tricky for me...