Box Cox transformation in gretl?

I was trying to transform some data using the boxcox criteria, but i couldn't.
I know the command is
boxplot(series, d)
being d the lamdba parameter.
But the issue here is gretl is not maximizing the MLE(maximum-likelihood estimation) when i run the boxcox command.
I know other softwares do this automatically.
I tried to run mle on gretl, but i get lost.
Does Anyone know how to maximize the chance of obtaining the value x with the following transformation of the x-value?
x*= (x^lambda-1)/lambda if x>0 
x*=log(x) if x=0

Thanks for your comments.