Bugfix, minor: properly alternate VPM/Buhlmann in deco loop The idea of the counter in this commit is to compute the opposite deco algorithm in 1:10 calls. So, when Buhlmann is selected, also compute VPM with 1:10 sample rate (and vise versa). The idea is good, but this never worked as the counter was local on the stack, so was 0 for every call here. So, when switching between deco algorithm, it took some refresh cycles as the computations where not in sync. Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>