The change was done long ago.. I don't remember the reason [and it wasn't documented in the commit message :(] https://bitbucket.org/petsc/petsc/commits/aa4d2986619717de6f4da384f5fcafb392... Perhaps we should revert it. I see a bunch of warnings with petsc fortran includes/examples. Will fix and push.. Satish On Sun, 10 Apr 2016, Lisandro Dalcin wrote:

Why is this warning silenced? Why should we treat Fortran different than C/C++ regarding unused variables? Maybe some old gfortran confused local variables with dummy arguments?