[Follow-up from "Small refactorizations in TimeLoop"] Interaction between setEquationSystem() based on ProcessData and nl_solver::setEquationSystem.
The following discussion from !3960 (merged) should be addressed:
-
@chleh started a discussion: For me it looks like this should be implemented in the nonlinear solver. The NL solver needs to know everything, anyway:
nl_solver->setEquationSystem(eq_sys_, conv_crit);
Who is responsible for that NL solver lib design?
😠 ;-)This is, of course, more work an probably should be done in a future MR.