git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5524 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5523 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5522 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5521 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5520 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5519 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5518 2fb0f335-1f38-0410-981e-8018bf24f1b0

- add an episode description argument to the startNextEpisode() method
- remove the maxAllowedTimeStep() stuff (the problem should take care about
  this)
- move the "timestep done" message after the time integration has been
  done, i.e. call the serialize method is now after a time integration
  is considered to be complete
- serialize and deserialize the current episode length (makes it
  easier to restart if the first episode is already over)

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5483 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5463 2fb0f335-1f38-0410-981e-8018bf24f1b0

now the problems do not need to specify 'using ParentType::neumann'
anymore. this is pretty much a temporary API, the real solution would
be that the methods get a context plus an entity index and an entity
codimension. the context stores the current solution and volume
variables, as well as the current element and its element
geometry. for other discretizations it may store other things...

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5462 2fb0f335-1f38-0410-981e-8018bf24f1b0

to cap the time step size a problem can overload the maxTimeStepSize() method.

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5461 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5460 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5459 2fb0f335-1f38-0410-981e-8018bf24f1b0

avoids a division by zero in the output of the box assembler

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5458 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5457 2fb0f335-1f38-0410-981e-8018bf24f1b0

using hints may dramatically speed up the expensive computations of
the volume variables if the calculations are expensive. This is due to
the fact that the initial guess for non-linear computation is usually
much closer to the result when using hints. (this mainly affects
calculation of the thermodynamic equilibrium when using non-linear
fluid systems.)

hints must be explicitly enabled by setting the "EnableHints" property
to true.

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5456 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5455 2fb0f335-1f38-0410-981e-8018bf24f1b0

add a base class for simple fluid systems which do not require complicated mutable parameter objects

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5454 2fb0f335-1f38-0410-981e-8018bf24f1b0

to cap the time step size a problem can overload the maxTimeStepSize() method.

Conflicts:

	dumux/boxmodels/common/boxproblem.hh

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5439 2fb0f335-1f38-0410-981e-8018bf24f1b0

   - output is again completely controlled by the problem
     (shouldWriteOutput())
   - default shouldWriteOutput() functions of base problems for
     decoupled models now make sure that an output is writen at t=t_end
   - output interval for decoupled models can now be set by the problem function
     setOutputInterval(int interval)



git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5430 2fb0f335-1f38-0410-981e-8018bf24f1b0

as primary variables, boundaryvariables are consistent with
fluxvariables


git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5427 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5413 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5412 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5411 2fb0f335-1f38-0410-981e-8018bf24f1b0

writer instead of reference to pointer to writer



git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5382 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5381 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5380 2fb0f335-1f38-0410-981e-8018bf24f1b0

adapted Andreas' last commit to decoupled models: writer is now pointer, hence setname() is usable name() it is not overwritten by the specific problem.

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5378 2fb0f335-1f38-0410-981e-8018bf24f1b0

this is just a single step of the newton method...

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5374 2fb0f335-1f38-0410-981e-8018bf24f1b0

this allows to first construct the problem, then set a name and then
run it which allows to avoid quite a few ugly hacks...

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5373 2fb0f335-1f38-0410-981e-8018bf24f1b0

timestep can now be restricted by an absolute value (maxAllowedTimeStepSize) for the whole simulation run

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5365 2fb0f335-1f38-0410-981e-8018bf24f1b0

added calculateVelocity() to fluxvariables (Kmvp, KmvpNormal) and
densityAtIp
localresidual uses now massfractions and the velocities calculated in
the fluxvariables


git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5345 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5344 2fb0f335-1f38-0410-981e-8018bf24f1b0

this is probably a temporary measure: if the new thermodynamic
framwork is adopted, then they will replace the 2p2c(ni) models in the
stable repository, if not these models will be deleted.



git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5340 2fb0f335-1f38-0410-981e-8018bf24f1b0

also, restore the 2p2c(ni) models to their previous state.in order to
compile the new_2p2c(ni) tests, you have to set symbolic links to the
devel repository for the following the direcories

dumux/material/fluidstates
dumux/material/new_fluidsystems
dumux/material/eos
dumux/material/constraintsolvers

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5335 2fb0f335-1f38-0410-981e-8018bf24f1b0

2p2c box model: make it shut up on primary variable switches, make the isothermal model work again after some API changes

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5334 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5333 2fb0f335-1f38-0410-981e-8018bf24f1b0

TODO: move code to calculate composition in twophase case to
constraint solver for composition from pressure + phase presence

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5332 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5330 2fb0f335-1f38-0410-981e-8018bf24f1b0

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@5329 2fb0f335-1f38-0410-981e-8018bf24f1b0