ok, I wanne say that double is not a precise number in computer.
there must be many unusual mistakes if you use double...
so, a better way is read them as strings,and then translate them to a long long number,which is 100 times as big as the input number.
so,now you can solve this without the trouble of precise....

p.s. you can also times the double with 100, but I think you will get a TLE....

sorry for my english....