I am using an API that uses integers where the calculations I am doing require floating point precision. The API provides a templat'd data structure and I am converting between the two template types to satisfy the API. Is this the reasonable thing to do for the conversion?
Code:template <typename T, typename U> void cut( core::coords<T> & t, core::coords<U> & u) { for all t.coords = (T) u.coords; }
Last edited by Tonto; 03-18-2008 at 10:32 PM.
You've provided no real information about the "API" (other than a hint that it's some set of templated functions and classes) so no answer is possible.
As a rough rule, if the "API" is documented as being limited to integral types, there is usually a good reason to assume it will work incorrectly with floating point types. You will need to confirm or deny that suspicion through suitable reading of documentation and/or testing.
Yea the type system won't allow anything but integers. I just require floating point in my calculations that can withstand the truncation to integers. I'm just wondering if there's any peculiar thing like operator () for automatic conversions between types for this sort of deal.
