Guido van Rossum a écrit : > This is way above my head. :-) > > The only requirement *I* would like to see is that for floats that > exactly represent ints (or longs for that matter) the result ought of > x%y ought to have the same value as the same operation on the > corresponding ints (except if the result can't be represented exactly > as a float -- I don't know what's best then).
Which is exactly the case at hand. As a proposed change, I would say that when the implementation of x%y returns y, then we should return 0 instead, or maybe the biggest possible number that is still < y. That way we would have a result that respects the "x%y < y" rule and still be as good as possible with the reduced float precision. And besides, mathematics say that "0 = y = 2y = 3y [y]" and so there is no problem for returning 0 instead of y :) -- http://mail.python.org/mailman/listinfo/python-list