U++ forum: Welcome to the forum

Thu, 03 July 2025 22:33

Didier
Messages: 736
Registered: November 2008
Location: France

Contributor

Hello,

While going through Upp::Rect code searching for the right methods to use to calculate a bounding box, I stumbled on some intriguing code:
two same methods : one fully template and the same template method specialized for <double>: nothing special ... except that I don't understand why one does 'p.x + 1' and the other one doesn't.

I see no valid reason to put the '+1' here : If I use 'float' instead of 'double' i will have a +1 applied even if my values are in the [0 ,0.1] range
So I think they're may be an error here or at least with 'float'

Cor/Gtypes.h

template <class T>
void Rect_<T>::Union(Pt p) {
	if(IsNull(p)) return;
	if(IsNullInstance()) {
		right = 1 + (left = p.x);
		bottom = 1 + (top = p.y);
	}
	else
	{
		if(p.x >= right) right = p.x +1;
		else if(p.x <  left) left = p.x;
		if(p.y >= bottom) bottom = p.y +1;
		else if(p.y < top) top = p.y;
	}
}

template <>
inline void Rect_<double>::Union(Point_<double> p) {
	if(IsNull(p)) return;
	if(IsNullInstance())
	{
		left = right = p.x;
		top = bottom = p.y;
	}
	else
	{
		if(p.x < left) left = p.x;
		else if(p.x > right) right = p.x;
		if(p.y < top) top = p.y;
		else if(p.y > bottom) bottom = p.y;
	}
}

Report message to a moderator