maybe a floating point issue?

[czylabsonasa]

> And you said you made some random tests - and results are the 
> same?
> even for 7 digits precision?
>
> It's weird

Yes, i tested the C++ and D code with a thousand (literally) of 
random instances and got no difference, even comparing the 
outputs as strings.
(For compiling g++ -O2 and gdc-12 -O2 were used in accordance 
with the kattis site). On my system (lmde/faye) only gdc-12 is 
available, and i don't want to "mess" it with a different 
compiler.