How does the fit function apply the robust method and user defined weights?

I want to know how exactly the fit function works. Specifically when the robust option and weights are given. If given some data y measured x with errors, yerr. We can do a robust weighted fit ...

options=fitoptions('poly2');
 set(options, 'Robust', 'on');
 set(options, 'Weights',yerr);

[best, best2]=fit(x,y, 'poly1', options);

But we could also apply the user defined weights to the data before doing the robust fitting and then do the fitting to get the same results.

   options=fitoptions('poly2');
   set(options, 'Robust', 'on');   
  [best, best2]=fit(x.*yerr,y.*yerr, 'poly1', options);

Sadly I don't, this is because I don't understand how the weights are applied when matlab does the robust fitting option.

Hope you can help