function whose arguments contains a function with possibly a variable number of outputs

I am writing a class which calls matrix and vector matlab functions adding some stuff to obtain an approximation of the number of significant bits.

For example with A=hilb(10) whose condition number is about 1e13 if I compute det(A) A beeing an object of my class I obtain 2.1640e-53 the number of displayed digits is automatically limited to the expected correct ones. To achive this I overload matlab function as shown below:

%%%%
methods
        function C=mldivide(A,B), C=rfpa.OP(@mldivide,A,B);end
        function d=det(A), d=rfpa.OP(@det,A);end
        function B=inv(A), B=rfpa.OP(@inv,A);end
        function c=cond(A), c=rfpa.OP(@cond,A);end
    ......

%%%

OP is a static method :

function xr=OP(fun,varargin)
            x1=fun(varargin{:});
            xr=zeros(size(x1),'rfpa');
            .....

This works fine but I have not found how to do, when function has a variable number of outputs as QR SVD for example.

Is it possible to implement something like

function varargout=OP(fun,varargin)
    [varargout{1:nargout}]=fun(varargin{:});
    .....

My current solution is to have specific wrapper for QR SVD etc... which is no intellectually satisfactory.

Any help will be welcome.

Alain

Answer 1

[Steven Lord]

See this documentation page, specifically the section on How to Use Comma-Separated Lists.

[Alain Barraud]

OK, According to the Function Return Values example it is necessary to add nargout as a input parameter within my function OP. Now it works as expected.

Thanks a lot

Alain