From: KIRK, BENJAMIN (JSCEG) (NASA) <benjamin.kirk1@na...>  20050321 18:09:35

In the course of some transient AMR stuff for the E.coli problem I found that at some adaptive steps the mesh does not actually change. In such a case it is obviously wasteful to project vectors (identity operator) and resolve the nonlinear problem (which converges in one iteration since it is the same problem you just converged. :) MeshRefinement::refine_and_coarsen_elements() and friends now return true if the mesh actually changes, false otherwise. This can be used to decide whether or not things need updating. So, in ex10, for instance, mesh_refinement.refine_and_coarsen_elements(); equation_systems.reinit(); could become if (mesh_refinement.refine_and_coarsen_elements()) equation_systems.reinit();  Benjamin S. Kirk 