Static methods not recognized (initially) in R2015a

Below is a copy/paste of a sequence of commands that I executed at the command line in R2015a and its output. Here GeoCalCho and TrajClass are both value classes that I have defined and GeoCalCho.split() is a static method of the GeoCalCho class. It is supposed to take a TrajClass object, th, as input and return two TrajClass objects, thu and thl, as output.

    >> clear GeoCalCho

    >> [thu,thl]=GeoCalCho.split(th)
    Undefined variable "GeoCalCho" or class "GeoCalCho.split".

    >> GeoCalCho  %call class constructor
    Error using GeoCalCho (line 31)
    Not enough input arguments.

    >> [thu,thl]=GeoCalCho.split(th)

    thu = 

      TrajClass with properties:

            data: [2x8x558 double]
        coordsys: 'ij'

    thl = 

      TrajClass with properties:

            data: [2x8x558 double]
        coordsys: 'ij'

So why doesn't MATLAB recognize GeoCalCho as a static method, initially? Clearly it is not a path problem. When I called the GeoCalCho class constructor, MATLAB finds it successfully. Moreover, MATLAB is thereafter able to find GeoCalCho.split(), although I haven't done anything to the path.

This is just an example. I have seen the same kind of behavior with other classes, though regrettably, I haven't found a way to reproduce it reliably. Has anyone else noticed this?

[Matt J]

I find that REHASH also works as a way of making MATLAB aware of the static method, but I don't see why that would be necessary

    >> clear GeoCalCho
    >> [thu,thl]=GeoCalCho.split(th)
    Undefined variable "GeoCalCho" or class "GeoCalCho.split".

    >> rehash
    >> [thu,thl]=GeoCalCho.split(th)

    thu = 

      TrajClass with properties:

            data: [2x8x558 double]
        coordsys: 'ij'

    thl = 

      TrajClass with properties:

            data: [2x8x558 double]
        coordsys: 'ij'

[Jan Siegmund]

Thank you, rehash saved me for a different problem, where a static method was recognized as a variable.