this hopefully reduces confusion about what which parameter means. the
Domain* parameters specify the physical size of the simulated domain,
the Cells* specify how many cells the grid has into the direction of
each axis.
the immiscible(ni) model is a model that is independent on the number
of phases and assumes immiscibility. this change is along the lines of
the previous commit for the compositional primary variable switching
model.
Tensor -> DimMatrix
refElem -> refElement
gv -> gridView
remove "AtIP" suffix
*Index -> *Idx for variables
remove ${MODEL_NAME}Index property
consistent argument names for compute(Storage|Flux|Source) in the local residuals
also remove the "MassUpwindWeight" and "UpwindWeight" properties as
they were not really useful.
most includes are now sorted by abstraction levels, with the most
abstract header first. also some system headers like <string> are now
explicitly included in every file where some of their classes are
used.
the switches are the introduction of "smart primary variables", rate
vectors, and changing the capillary pressure API for the multi-phase
models to the one previously used by the MpNc model.
- this renames historyIdx to timeIdx (I think it expresses what it is meant for better)
- rename localIdx to "spaceIdx" in the problems/spatial parameters
- add a "timeIdx" argument to the problem's methods
tested on GCC 4.3 and GCC 4.6. Due to an incompatibility between those
two compilers (GCC 4.3 expects that there's no semicolon behind
static_assert(), GCC 4.6 requires one), it only compiles on GCC 4.6.
- the constructor of VtkMultiWriter now expects a grid view as first
argument. unfortunately the old API cannot be kept around with this
change.
- as a follow up beginWrite() does not need the grid view as an
argument anymore
- if the grid has been changed during two time steps, the new method
gridChanged() must be called before beginWrite() is called
as described in my e-mail from wednesday. Everything in the stable
repository compiles.
also, some compile fixes for the Mp-Nc model where made to some fluid
systems and the 1p2c index structure got an offset template argument
like the other models.
