Hi, I came along a small bug in the numerical evaluation of alternating Euler sums of depth 1, when in the input no lists are used:

evalf(zeta({5},{1})); 1.0369277551433699263 evalf(zeta(5,1)); 1.0369277551433699263 evalf(zeta({5},{-1})); -0.972119770446909306 evalf(zeta(5,-1)); zeta(5,-1)

The attached patch fixes the last case, so zeta(5,-1) is now also evaluated numerically. Best wishes, Stefan