Bugfix#210 intermittent fails #244
Conversation
Codecov Report
@@ Coverage Diff @@
## master #244 +/- ##
==========================================
- Coverage 61.43% 61.41% -0.03%
==========================================
Files 19 19
Lines 3558 3558
==========================================
- Hits 2186 2185 -1
- Misses 1372 1373 +1
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #244 +/- ##
==========================================
+ Coverage 61.43% 61.46% +0.02%
==========================================
Files 19 19
Lines 3558 3558
==========================================
+ Hits 2186 2187 +1
+ Misses 1372 1371 -1
Continue to review full report at Codecov.
|
nip5
added
the
Bugfix
There was a problem hiding this comment.
This is not fixing #210. It still can fail, e.g.,
soilTemp_rng.state=fe4d618741a96818, soilTemp_rng.inc=11
test/test_SW_Flow_lib_temp.cc:589: Failure
printed by TEST(SWFlowTempTest, MainSoilTemperatureFunction_LyrMAX)
...
RandSeed(0,&soilTemp_rng);
#include <inttypes.h>
swprintf("state=%" PRIx64 ", inc=%" PRIx64 "\n", soilTemp_rng.state, soilTemp_rng.inc);
...
- keep soil bulk density fixed -- instead of setting it by a random number - keep SWC(wilting point), SWC(field capacity), SWC(saturation) fixed and physically well defined (see code comments) -- instead of setting field capacity by a random number - only set SWC by a uniform random number and keep it between field capacity and wilting point -> goal: keep setup simple: this is to test the soil temperature function and not a random number function - new script `many_test_runs.sh` to repeatedly run unit tests -> this commit has not failed in 2,500 test runs: close #210 ``` N=2500 ./many_test_runs.sh ```
This fixes the problem but is a hack so while it fixes this so our CIs pass, further investigation is necessary for a more permanent fix later on. This just changes the random number lower bound for
bDensity2[i]from.1to.129. It still generates the value ofbDensity2[i]randomly.