Residual resampling using bootstrap procedure.

So I'm evaluating mutual fund performance using a four-factor model as suggested by Carhart(1997). The specification of the model is as follows: { Rpt- Rft = ap + bp (Rmt- Rft) + si SMB + hi HML + wi WML + ei } which is basically a multi-variate ols. this model assumes that the error term ei ( or residual) is normally distributed. However, daignostic analysis shows that it is not normally distributed. So to overcome this problem, we resample the residuals with replacement using the bootstrap technique. Since I am new to matlab, I would appreciate any lead on how to perform this bootstrap simulation.