Hilding Elmqvist

Thank you for pointing this out. This shows an error in a feature to remove singularities for certain models. In this case you can turn off this feature: simulate(PartialTest, 1,...

Toivo and I have started to work on that. We will continue next week. So I suggest you just wait a couple of weeks.

Yes. You are right.

Thank you for highlighting this. It is a general property of the Julia parser. `julia> e = [1 + 2] 1-element Vector{Int64}: 3 julia> e = [1 +2] 1×2 Matrix{Int64}:...

It is also of interest to study the Modelica [specification](https://modelica.org/documents/ModelicaSpec34.pdf) The section: 12.7 Declaring Derivatives of Functions (page 160) describes how one can provide derivative functions manually. It shows several...

The current Modia implementation regarding differentation of non-standard functions uses a naming convention. The tests for [symbolic transformation](https://github.com/ModiaSim/Modia.jl/blob/master/test/symbolic/DAEquations/testSymbolicTransform.jl) illustrates the current convention: ``` der = showDifferentiate(:(y = f(x, 5, z)))...

Excellent: Some test cases: ``` straightFlanksCamJacobian(0.0)*[1; 0; 0; 0] == [0.0, 2.5] straightFlanksCamJacobian(pi/2)*[1; 0; 0; 0] == [-1.0, 6.123233995736766e-17] straightFlanksCamJacobian(pi)*[1; 0; 0; 0] == [-1.2246467991473532e-16, -1.0] straightFlanksCamJacobian(3*pi/2)*[1; 0; 0; 0]...

It seems the user function needs to be wrapped in order to have a function with a single vector input for ForwardDiff. What would be the best way of automatically...

There is an issue how to make the lookup of the derivative function. Presently a new Symbol is created for the derivative function: ` f_der = Symbol(string(op, "_der"))` and a...

The solution was to use the getfield() function. So now, using Modia master, derivative functions are found and called. But a wrapper needs to be made maually using, for example,...