Taylor McDonnell

Here's something to think about related to this issue: ```julia using ForwardDiff using LinearAlgebra # start with zero valued vector of dual numbers v = zeros(ForwardDiff.Dual{Nothing, Float64, 1}, 3); #...

I created a new issue related to my comment as it is somewhat tangential to this issue. I'll post my workaround there.

Is it possible we could use [Preferences.jl](https://github.com/JuliaPackaging/Preferences.jl) to enable/disable NaN-safe mode now? See https://github.com/JuliaLang/Pkg.jl/issues/458. I think that would effectively solve this long-standing issue.

One possible workaround is to replace the zero-testing in the NaN-safe function definitions with approximate zero-testing as follows. ```julia if NANSAFE_MODE_ENABLED isapproxzero(x) = isapprox(x, zero(x), atol=eps(promote_type(eltype(x), Float64)), norm=(v)->norm(v, Inf)) @inline...

Just implemented this for fmincon with a combination of MATLAB.jl and mexjulia. See https://github.com/byuflowlab/Fmincon.jl. I had to revive mexjulia (update to Julia 1.0) to get it to work, (which I...

Theoretically AD would generate the same code with the right overloads, but tape based methods don't work for long computations and my understanding is that pervasive Zygote doesn't quite work...

It seems the problem is limited to the automated tests. I'm able to build and run on Julia 1.8 locally.

What MATLAB release are you using? One of this packages dependencies, [MATLAB.jl](https://github.com/JuliaInterop/MATLAB.jl) doesn't work with 2022a and 2022b. I have some automated tests that passed a year ago but fail...

The latest versions of MATLAB (2022a and 2022b) and Julia (v1.8) don't appear to work with this package. The automatic tests currently work with MATLAB 2021b and Julia 1.6. Perhaps...

I can confirm that the master branch (i.e. `dev Mex`) works on Linux with MATLAB 2021b and Julia 1.8. I also tested to make sure Fmincon.jl still works, though I...