Opened 7 years ago
Last modified 4 years ago
#11739 closed Bugs
operator / for point2 is hard coded to operate on type double — at Initial Version
Reported by: | Owned by: | Hailin Jin | |
---|---|---|---|
Milestone: | To Be Determined | Component: | gil USE GITHUB |
Version: | Boost 1.59.0 | Severity: | Problem |
Keywords: | Cc: | mateusz@… |
Description
When using boost::gil::point2<float>, applying a division operation results in a point2<double>. This is undesirable.
The fix appears to be easy, change this code
template <typename T> BOOST_FORCEINLINE point2<double> operator/(const point2<T>& p, double t) { return t==0 ? point2<double>(0,0):point2<double>(p.x/t,p.y/t); }
to operate on T instead of double. But, that seems overly simple so I am trying to figure out why this code was written this way in the first place.
A possibly safer alternative is to cast the result back to point2<T>, ie;
template <typename T> BOOST_FORCEINLINE point2<double> operator/(const point2<T>& p, double t) { return t==0 ? point2<T>(0,0):point2<T>(T(p.x/t),T(p.y/t)); }
This change makes the / operator act the same as the /= operator, which is also hard coded to double, but because it's a member, doesn't yield a new type, so it's probably fine (other than the fact that I asked for floats and am getting double operations).
While we're at it, I also find it weird that this code is protecting against division by zero, but that's a different problem.