Barry Smith wrote:

A couple of weeks ago these examples ran fine and did not crash. Now they crash. Perhaps it was not a change to fieldsplit that caused the recent crashes, but it is a recent change to something.

Indeed, but I am still responsible. When I changed MatSetBlockSize to not error when called (with compatible bs) on BAIJ, I forgot to actually change the block size for the MPI* matrices, just for a->A and a->B. Can we run TESTEXAMPLES_DATAFILESPATH in the nightlies? Note that PETSC_MAT_LIB does not depend on PETSC_DM_LIB. This is violated in mhyp.c, all over impls/adic (but I never got ADIC working), and in tests/ex96.c (which can be trivially moved to dm/da/examples/tests). This breaks anything that tries to link against PETSC_MAT_LIB (but not PETSC_KSP_LIB which does depend on PETSC_DM_LIB). Jed