There's no evidence of a bug in reading/printing, because the only way I can currently get a number big enough to trigger the supposed bug is by multiplying, which doesn't work.