Error flag instead of error message for bvp5c

During running a continuation for boundary value problems, I encounter a situation, where the problem becomes singular and bvp5c throws an error.

I would prefer to be able to catch the error, reduce the continuation stepsize and continue with the computation or possibly stop the continuation. So something like an exitflag would be fine, or the singular Jacobian could be reported in the solution stats.

Answer 1

[Alois Steindl]

The solution to my problem is quite simple: I just use try/catch, to signal the failure and continue or stop the continuation.

Nevertheless, I would prefer a return flag from bvp5c.