Avoiding an int8 underflow (24d1a7e208b90436b01e743c974ef45f7444e1af).
In certain Europe test cases (e.g. True-50-12) this leads to (small) changes in the
output, so we'll need to switch to a new set of reference outputs (currently inside
a aquacrop_tests_PR300 folder on stg_00024).
This will also be fixed in the Pascal version, because the same bug is present there.
Fixing the FORTRAN_EXE=0 build (7c0522d8fd63ee436b78115715b9a65467dfe243),
which got broken after InitializeRun was split up into InitializeRunPart1/Part2.
This branch passes all tests (with aquacrop_tests_PR300 reference results), performed
for FORTRAN_EXE=1 with both DEBUG=0 and DEBUG=1 and with foss-2018a and with
the Singularity image. The same tests also passed with FORTRAN_EXE=0 but only
considering ~50 Europe cases.
Two (unrelated) bug fixes in this PR:
Avoiding an int8 underflow (24d1a7e208b90436b01e743c974ef45f7444e1af). In certain Europe test cases (e.g. True-50-12) this leads to (small) changes in the output, so we'll need to switch to a new set of reference outputs (currently inside a
aquacrop_tests_PR300folder onstg_00024).This will also be fixed in the Pascal version, because the same bug is present there.
Fixing the
FORTRAN_EXE=0build (7c0522d8fd63ee436b78115715b9a65467dfe243), which got broken afterInitializeRunwas split up intoInitializeRunPart1/Part2.This branch passes all tests (with
aquacrop_tests_PR300reference results), performed for FORTRAN_EXE=1 with both DEBUG=0 and DEBUG=1 and with foss-2018a and with the Singularity image. The same tests also passed with FORTRAN_EXE=0 but only considering ~50 Europe cases.