Yes I would say that is a bug. It looks like old integer code is being
used for floating point, resulting in it rounding down multiple times.
Fixed in svn.
Peter Li wrote:
> Hi,all
> I was using fltk2.0 r5864 on Windows.
> I found a bug about scale transform error!
> Can you discuss it?
>
> For example, next code:
>
> readimage(data , RGB, Rectangle(x, y, 200, 24));
> push_matrix();
> scale(1, 2);
> drawimage(data, RGB, Rectangle(x, y+50, 200, 24));
> pop_matrix();
>
> Now, x = 20; y = 0;
> When I trace drawimage to void Image::draw(const fltk::Rectangle& from,
> const fltk::Rectangle& to) ,
> I watch the Rectangle to, is: to {x_=20 y_=50 w_=200 ...} const
> fltk::Rectangle &. this is my expectant.
> Then run to "fltk::Rectangle R; fltk::transform(to,R);", I hope
> Rectangle R is {20, 100, 200, 48}.
> But I run into void fltk::transform(const Rectangle& from, Rectangle&
> to),
> float w = rintf(sqrtf(d1x*d1x+d2x*d2x)); // w is 200, it is ok.
> x = floorf(x - (w+1)/2); // x is 19, why? it
> is wrong.
>
> float h = rintf(sqrtf(d1y*d1y+d2y*d2y)); // h is 48, it is
> ok.
> y = floorf(y - (h+1)/2); // y is 99, it's
> wrong.
> So I get the R, R{x_=19 y_=99 w_=200 ...} fltk::Rectangle, it's wrong.
>
> Is this a bug?
>
>
_______________________________________________
fltk-dev mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk-dev
