id summary reporter owner description type status priority milestone component version severity resolution keywords cc 1114 mismatch in array size and possibly uninitialized value in solsor.F90 avidard gm "in solsor.F90 line 140 one can find ztab(:,:) = gcr(2:nlci-1,2:nlcj-1) while ztab is (jpi,jpj). Since ztab is zeroed just before and only the max(ztab) is needed, I guess it does not really matter, but still it is not very nice. additionnaly, in both convergence test, the computation of the residual res should be put out of the IF statement since it is used anyway in the output in solver.stat " Bug closed low OCE trunk wontfix