
2015/05
=======

- ggcm_mhd_iono: fix calculation of theta in current map 

  This corrects (I believe) the mapping of the currents to the
  ionosphere, but we don't know what difference it makes.  It's now
  inconsistent with Fortran, so this should be resolved eventually.

- ggcm_mhd_iono: fix current map shifted by 2

  The C code was definitely wrong before this -- but it's not clear
  whether the Fortran version was (and still is) subject to the same
  issue.

- ggcm_mhd_step: too many time steppers have their own implementation
  of a debug_dump option -- this should be handled more generically

- amr: vlct B fluxes now not destroyed by flux correction
    
  Basically, the vlct scheme needs B fluxes up to one ghost point outside,
  and the flux correction ended up zeroing those, which is not good.
    
  But this current solution is kinda hacky and changes behavior for
  stencil_file / hardcodes BX/Y/Z (>= 5)

- ggcm_mhd_iono: zero_potential

  Would be a good opportunity to provide a much simpler inner boundary w/o
  mapping / commmunication with ionosphere


old
===

- fix #res1border hardcoded

- there are still some ggcm_mhd_iono changes in "ganymede-before-lws"

- mirdip i.c. cleanup/consolidation

- get rid of default_box()

- better handling of b.c.s

- step c3 still needs mrc_fld_max() etc.

- eta0i normalization

- use mrc_fld_spatial_dims()