Issues with Modelica.Media.IdealGases.Common.MixtureGasNasa.T_hX

[HansOlsson]

It seems that these implicit functions are problematic, see also #3399

The function:

function T_hX "Return temperature from specific enthalpy and mass fraction"
  extends Modelica.Icons.Function;
  input SpecificEnthalpy h "Specific enthalpy";
  input MassFraction[nX] X "Mass fractions of composition";
  input Boolean exclEnthForm=excludeEnthalpyOfFormation
    "If true, enthalpy of formation Hf is not included in specific enthalpy h";
  input Modelica.Media.Interfaces.Choices.ReferenceEnthalpy
                                   refChoice=referenceChoice
    "Choice of reference enthalpy";
  input SI.SpecificEnthalpy h_off=h_offset
    "User defined offset for reference enthalpy, if referenceChoice = UserDefined";
  output Temperature T "Temperature";
protected 
  MassFraction[nX] Xfull = if size(X,1) == nX then X else cat(1,X,{1-sum(X)});

  function f_nonlinear "Solve h_TX(T,X) for T with given h"
    extends Modelica.Math.Nonlinear.Interfaces.partialScalarFunction;
    input SpecificEnthalpy h "Specific enthalpy";
    input MassFraction[:] Xfull "Mass fractions of composition";
  algorithm 
    y := h_TX(T=u, X=Xfull) - h;
  end f_nonlinear;

algorithm 
  T := Modelica.Math.Nonlinear.solveOneNonlinearEquation(
    function f_nonlinear(h=h, Xfull=Xfull), 200, 6000);
  annotation(inverse(h = h_TX(T,X,exclEnthForm,refChoice,h_off)));
end T_hX;

Has the following issues:

• The inverse-annotation uses the extra arguments, exclEnthForm, refChoice, h_off, but the actual solution ignores them. I believe it is simplest to just remove them completely.
• Creating Xfull is unnecessary since the input X is declared with the size nX. The inverse also uses X directly.