geoffp7
geoffp7
You may be right - but in that case, there is a bug in SDL_PointInFRect: ``` SDL_FRect fr = { 2.5f, 1.5f, 15.25f, 12.0f }; SDL_FPoint fp = { 17.5f,...
Here's another instance of the bug, in which the behaviour is obviously wrong: ``` SDL_FRect fr = { 2.5f, 1.5f, 0.25f, 12.0f }; float x1 = 0.0f, y1 = 6.0f,...
I looked at the source code. The int and float versions of the rectangle functions use the same code: the file SDL_rect_impl.h is #included twice, once to generate the Rect...
That looks like the right approach. But there is a tricky issue at the end of `SDL_ENCLOSEPOINTS`: having determined that the smallest and greatest x-coordinates of the points are `minx`...
The following code produces no output. ``` SDL_FRect result; float h = 1.0f/128.0f; for(float xmin = -1024.0f; xmin
That seems like the right general approach. But for the calculation in the floating-point case, it is not enough to just let the epsilon value be `SDL_FLT_EPSILON`, or any other...