rsa.compareRefRDM2candRDMs bug with case 'randomisation' when nan in models

[iancharest]

case 'randomisation'
        fprintf('\nPerforming condition-label randomisation test of RDM relatedness (fixed effects).\n');
        nRandomisations = userOptions.nRandomisations;
        for rdmI = 1:numel(candRDMs)
            % do the randomisation test, also keep the randomistion correltions
            % in a separte matrix
            rdms(rdmI,:) = vectorizeRDM(meanCandRDMs(:,:,rdmI));
        end
        [n,n]=size(meanRefRDM);
        exhaustPermutations = false;
        if n < 8
            allPermutations = exhaustivePermutations(n);
            nRandomisations = size(allPermutations, 1);
            exhaustPermutations = true;
            warning('(!) Comparing RDMs with fewer than 8 conditions (per conditions set) will produce unrealiable results!\n  + I''ll partially compensate by using exhaustive instead of random permutations...');
        end%if

Performing condition-label randomisation test of RDM relatedness (fixed effects).
Undefined function or variable 'exhaustivePermutations'.

Error in rsa.compareRefRDM2candRDMs (line 622)
allPermutations = exhaustivePermutations(n);

The bug comes in the randomisation and seems to be driven by the fact that one of the model rdms had nans in it.

Should the possibility of having nans in the model RDMs be added to the toolbox?