Re: [boost] Library Proposal: Boost.Geom 2D and 3D Geometric Primitives

From: Sérgio Vale e Pace (svpace.forum_at_[hidden])
Date: 2005-06-21 08:25:18

On 6/21/05, Anis Benyelloul <benyelloul_at_[hidden]> wrote:
> > Sérgio Vale e Pace wrote:
> 
> > graphical libraries usually are incompatible and sometimes even
> > "unusual" for a reason, performance, when you´re processing lots and
> > lots of points per second, the simple act of calling a non inline
> > function is too much overhead.
> 
> 
> I totally agree with you. But, who said that my code was calling
> non-inline functions ?
Sorry I didn´t mean your code does that, just that even <insert simple
action here> could be bad
> 
> > Don´t take me wrong I also believe that
> > a "unique" implementation for points would be great, but such
> > implementation is not trivial.
> 
> 
> Sorry, I actually don't beleive so. I do not intend to create just one
> more implementation for points and rectangles, I want to give points
> and rectangles of *others* the same inteface. Yes, after having
> specialized some traits class you'll be able to write :
> geom::point<QPoint> pt; // use QPoint as an implementation
> geom::box<SDL_Rect> bx; // use SDL Rect as an implementation
> // Now mix !
> bx+=pt; // !!
> // And still be able to extract the QPoint out of pt;
> QDrawPoint(pt.impl());
> SDL_UpdateRect(&(bx.impl()); // And the SDL_Rect out of bx, at no cost !
> > I´m not sure about the internals of
> > your implementation,
> 
> 
> Then you should check the 'Boost Sandbox File vault' for a file named
> geom.zip and make your own opinion. Reading through the documentation
> should give a better idea of the general "spirit" of the library.

Ahh, now I get it... I will take a look at your code.

> -- Anis Benyelloul
Sérgio