this renames the 'test' directory to 'tests' and 'test/implicit' to
'tests/models'. the latter change reflects the fact that in eWoms all
models are implicit since the IMPET models have been removed.
now, we only compare pressure, saturation, mass/mole fractions and
velocity fields and we only do a qualitative comparison (i.e., much
higher tolerances). Hopefully this should avoid issues with sightly
different reference solutions in the future.
Only public member names have been treated so far. Local variables
will be done shortly, classes maybe not before the next release
cycle. Everything old could be marked deprecated.
Reviewed by Benjamin and Christoph.
Dumux-Svn-Revison: 10761
Ported-By: Andreas Lauser <andreas.lauser@iws.uni-stuttgart.de>
- special status of saturation. In particular, it deals with the
capitalized "S"-indices:
S0Idx -> s0Idx
SnIdx -> snIdx
SwIdx -> swIdx
SOrX1Idx -> sOrX1Idx
SOrX2Idx -> sOrX2Idx
SnOrXidx -> snOrXIdx
SwOrXIdx -> swOrXIdx
Reviewed by Christoph.
Dumux-Svn-Revison: 10756
Ported-By: Andreas Lauser <andreas.lauser@iws.uni-stuttgart.de>
- time steps for the pressure model can be choosen larger than for the
transport model
- locally small time steps of the std::explicit scheme are less
restrictive
- reviewed by Bernd
Dumux-Svn-Revison: 10700
Ported-By: Andreas Lauser <andreas.lauser@iws.uni-stuttgart.de>
- it now can generate and install the API documentation ('make install-html')
- the information of the dune.module file is now used much more extensively
- presence of the valgrind headers is tested
- tests which cannot be run for a configuration will now appear as
'skipped' in CDash
- adding tests does not require modifying the toplevel CMakeList.txt anymore
- introduction of the 'EwomsAddTest' cmake macro which removes quite a bit of the
copy-and-paste from the CMakeLists.txt.
- various small loose ends have been fixed
if one of the files matches (fuzzyly, obviously), the test counts as
passed if none matches it counts as failed. this lets us add reference
results for hosts that need the same number of time steps, but for
whatever reason produce a different result. also, the number of time
steps required to produce the result do not matter anymore, as long as
the data is the same (in the fuzzy sense).
this is somewhat similar to what is currently done in the 'implicit'
branch of dumux, but it takes things further:
- ewoms/boxmodels/common -> ewoms/disc/vcfv
(vcfv == vertex centered finite volumes)
- ewoms/boxmodels/vtk -> ewoms/vtk
(possibly this will move to ewoms/io/vtk, TBD)
- ewoms/boxmodels -> ewoms/models
- test/boxmodels -> test/implicit
- the files for the common VCFV code have been renamed from "box*" to
"vcfv\1" and the classes they contain have been renamed accordingly.
The rationale is to reflect the fact that multiple discretizations may
be chosen for any physical model. (Once these discretizations are
implemented, obviously.) All tests still compile for the autotools and
cmake build systems
instead of typing 'make test/common/propertysystem/test_propertysystem'
in the toplevel build directory, one can now just type
'make test_propertysystem'
eWoms hereby declares full independence. Humor aside, the main
technical advantage of this is, that it is now possible to easily
install both, Dumux and eWoms on a system using a package management
system without bad tricks.
this field is only written by the semi-implicit models (the
fully-implicit ones write "filterVelocity_$PHASENAME") and caused
problems for "test_1p". Since I don't care about the semi-implicit
models too much, let's just increase the tolerance.
- don't use references to entity pointers in the semi-implicit models
anymore. this caused the element pointer to go out of scope and made
it dangle. It was just pure luck that it (mostly) worked anyway.
- fix the non-isothermal navier-stokes model
- reduce the default value for the final time of the water-air problem
to 5000 seconds. this should avoid time-out problems on slower
machines (nalle, I'm looking at you)
- reduce the final time for the reservoir problem to 100 seconds to avoid
numeric errors from accumulating too much.
- don't complain about saturation differences less than .1% anymore
- make the fix for the end time detection work for negative end times
- reduce the total time the stokes problem is run to 10 seconds. this
is probably not a real fix, but the navier-stokes model does not
work nicely yet anyway...
- do not compare floating point values with the '==' operator in the
immiscible flash constraint solver.
- run one simulation in parallel if MPI is available
- test the parameter passing infrastructure
- reduce end time of the navier stokes problem to 1e-3 to make it pass
on the first try
- add CUSTOM_CXX_FLAGS to CMakeLists.txt to allow passing compiler
flags from the command line to make clang shut up about the dune
issues it encounters.
