If a minimizer returns no error information (gradient-free method, aborted fit, or a parameter at a bound), we pass this further as 0.0 error value. This masks the actual issue since 0.0 can be an actual error.
It's better to make it None or something.
If a minimizer returns no error information (gradient-free method, aborted fit, or a parameter at a bound), we pass this further as
0.0error value. This masks the actual issue since 0.0 can be an actual error.It's better to make it
Noneor something.