Andreas Lauser
2f8118e5a6
validate the results produced by the parallel tests
...
this implied a few changes to the "runtest.sh" script and requires
reference solutions...
2014-03-06 14:37:32 +01:00
Andreas Lauser
d9c4d1058e
revert the output redirection stuff of ebdcba686ea10c5da5dd4f4da98f19ef388c2af4
...
it seems like shellcheck complains about this but once the order of
'>' and '2>&1' is reversed, bash does not like it anymore. since
making bash happy is way more important than shellcheck, let's revert
this part of the change...
2013-11-26 12:20:26 +01:00
Andreas Lauser
79bfc102cc
runtest.sh: fix shellcheck warninings
...
at least the ones which were no false positives
2013-11-25 15:04:16 +01:00
Andreas Lauser
537bf6a1bd
parameter system: fix handling of --help/-h flag
2013-09-24 18:58:47 +02:00
Andreas Lauser
8276063c68
test driver: test -> tests
2013-09-24 18:33:28 +02:00
Andreas Lauser
2b4f07e18a
make the tests work
...
i.e., update the CMakeLists* and tweak runtest.sh
2013-08-28 17:28:26 +02:00
Andreas Lauser
ec369207bc
test runner script: be more verbose if something goes wrong
2013-02-07 19:44:15 +01:00
Andreas Lauser
bff2b8c72d
add a third reference solution for the tutorial_coupled test
2013-02-07 19:44:15 +01:00
Andreas Lauser
75cbe3f132
test runner: more verbose output
2013-02-06 15:48:03 +01:00
Andreas Lauser
4a2b4ceb3b
write the newton convergence for the restart test.
...
this slightly increases test coverage, and raises an alarm if
something goes horribly wrong in the code writing the convergence.
2013-01-30 22:46:24 +01:00
Andreas Lauser
973e5349f0
test runner: print some infos to ease debugging before comparing the simulation results
2013-01-30 15:39:10 +01:00
Andreas Lauser
fb0f3902ee
update randomization once more
...
It turns out that using the output of base64 for filenames is not a
too good idea.
2013-01-25 11:31:37 +01:00
Andreas Lauser
896beaad90
(hopefully) better random number generation to avoid test failures
2013-01-24 23:28:57 +01:00
Andreas Lauser
03af1e1ecb
fix testrunner
...
*grr* stupid mistakes!
2013-01-13 16:09:15 +01:00
Andreas Lauser
0a5b63f70e
test runner: allow an arbitrary number of reference results
...
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).
2013-01-11 18:16:29 +01:00
Andreas Lauser
28a0fb487c
test driver script: make it more robust
...
it does not get confused by the result files of previous runs
anymore...
2012-11-28 20:30:30 +01:00
Andreas Lauser
38d7b97157
CMake: Teach it good manners when it comes to caching
...
also improve testing by adding a restart test and making sure that
invalid parameter names are not accepted.
2012-10-29 18:48:10 +01:00
Andreas Lauser
93049a223d
fix tests for test_diffusion, test_navierstokes
2012-10-23 16:55:44 +02:00
Andreas Lauser
63d554ff08
make the test runner script also work for semi-implicit simulations
2012-10-23 13:06:20 +02:00
Andreas Lauser
9f13b80118
improved unit testing
...
- 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.
2012-10-23 11:55:26 +02:00
Andreas Lauser
c601a65931
runtest.sh: make it work in parallel
2012-10-22 22:49:35 +02:00
Andreas Lauser
95ae6949d8
fix the CMake build
...
also, fix and simplify the CTest system
2012-10-19 19:39:53 +02:00
Andreas Lauser
12cffbb074
update reference solutions
...
also rename runTest.sh to runtest.sh and improve printed messages of
fuzzycompare.py slightly
2012-07-12 21:25:17 +02:00
