Now that the translation is finished, it was time for some cleanup/restructuring.
The most noticeable changes are:
there is only one shared library now (libaquacrop.so)
the executables are now just called aquacrop (instead of PluginBareV70)
FORTRAN_EXE=1 is now the default (instead 0)
Edit: the full testsuite now passes for both DEBUG=0 and DEBUG=1, both in the Singularity image
and with foss-2018a. This was done with the (now-default) FORTRAN_EXE=1 option.
For the FORTRAN_EXE=0 option, the same combinations were tested, but only for the Harvest test
(this is OK since the Fortran-Pascal coupling is now reduced to a trivial StartTheProgram() call).
Forget my last comment. Of course, Europe test still missing. Saw also you merged the getters and setters already into this branch. Will use this for the LIS coupling now.
As you can see in 673522f, there was still one failure for the Harvest test,
but only for DEBUG=0 and with the Singularity image. That one is also solved now
(though the real cause rather seems to be a bug in the v10.2.1 gfortran compiler).
Now that the translation is finished, it was time for some cleanup/restructuring.
The most noticeable changes are:
libaquacrop.so
)aquacrop
(instead ofPluginBareV70
)FORTRAN_EXE=1
is now the default (instead0
)Edit: the full testsuite now passes for both
DEBUG=0
andDEBUG=1
, both in the Singularity image and withfoss-2018a
. This was done with the (now-default)FORTRAN_EXE=1
option. For theFORTRAN_EXE=0
option, the same combinations were tested, but only for the Harvest test (this is OK since the Fortran-Pascal coupling is now reduced to a trivialStartTheProgram()
call).