New function Gauss_test_powerAnalysis for
computation of power or sufficient sample size for the one-sided Gauss
test. The power (respectively sufficient sample size) are computed in an
exact (non-asymptotic) way based on bounds on the Edgeworth expansion.
This power holds non-asymptotically and uniformly over all alternatives
with a given effect size (under some regularity conditions: bounded
kurtosis, tail constraint kappa).
Adding a verbose option to Bound_EE1 to
indicate the values of the intermediary terms that are summed to produce
the final bound. This can be useful to understand which terms has the
largest contribution to the bound.
Removing pipes in
Lower_incomplete_gamma_for_negative_x to be compatible with
R (< 4.1.0).