Steven G. Johnson

Do you have a simple test case that illustrates this?

See also stevengj/nlopt#28 (a vaguely similar issue with SLSQP, which I'm also having difficulty reproducing).

If the options are incorrect it would throw a different exception. The BFGS failures are a bit harder to diagnose (in part because the Fortran BFGS code I am calling...

The philosophy of NLopt has always been that this sort of thing is best implemented by just adding a line or two to your objective function. (Either to print things...

I'm still struggling to see what is to be gained by supporting this. How is something like: ``` trace_vals!(opt, true) .... vals = get_trace_vals(opt) ``` so much easier than just...

@mlubin, but in that case shouldn't tracing be added to the JuMP interface?

I'm still confused; if you are using the JuMP interface, how would you hook into an NLopt-specific method for printouts? Shouldn't there be a solver-independent JuMP interface to request "informational...

Since JuMP generates the objective function for you, so you can't insert tracing if you want to, it seems like this should be a JuMP issue. It seems like you'd...

Something along these lines seems reasonable to me.

Rather than `min_objective!(opt, f, autodiff=:forward)`, I would suggest something like `min_objective!(opt, autodiff(f))`, where `autodiff(f)` takes a function `f(x)` and returns an `(x, grad) -> ...` function of the form expected...