3-fold cross validation for svm

Hello, I'm trying to execute a rid search 3 fold cross validation for svm. I am using an svm classifier from this package:

http://asi.insa-rouen.fr/enseignants/~arakoto/toolbox/index.html

and the svm classifiers from exclass1.m.

I am relying on matlab's crossval package, but am having trouble with the function component and calling the right arguments.

Here is my attempt:

%Gaussian

C = [.00000001, .0000001, .000001, .00001, .0001, .001, .01, .1, 1, 10, 100, 1000]
G = [.00000001, .0000001, .000001, .00001, .0001, .001, .01, .1, .3, .5, 1]
grid = zeros(numel(C), numel(G));
[Xapp,yapp,xtest,ytest]=datasets('gaussian',n,0,sigma);
X =[Xapp,yapp,xtest,ytest]

for c = 1:numel(C)
    for g = 1:numel(G)

          vals = crossval(fun(Xapp, xtest,yapp, C(c), G(g)), X, 'kfold', 3);
          grid(c,s) = mean(vals);
      end
  end

[cbest, sbest] = find(grid == max(grid(:)));

C_final = C(cbest);
S_final = S(sbest);

function svm = fun(Xapp, xtest, yapp, C, G)

      [xsup,w,w0,pos,tps,alpha] = svmclass(Xapp,yapp,C,G,'Gaussian',1,1); 

      ypred = svmval(xtest,xsup,w,w0,'Gaussian',1,[ones(length(xtest),1)]);
  end

The error I get is here:

Output argument "svm" (and maybe others) not assigned during call to "Problem_Set_2>fun".

Error in Problem_Set_2 (line 14) vals = crossval(fun(Xapp, xtest,yapp, C(c), G(g)), X, 'kfold', 3);

Any suggestions?