Re: gnomecanvas question

[Joe Van Dyk <joevandyk gmail com>]

I'm not familar at all with the gnomecanvas implementation, but I
highly doubt it stores information about every pixel in memory.


On Mon, 14 Mar 2005 17:31:59 -0600, Greg Breland
<gbreland mozillanews org> wrote:
> I'm amazed its even possible to create a canvas that is 1million x
> 1million.  The shear amount of memory needed to do this is staggering.
> Lets say it a 16bit canvas, that would be (1000000*1000000*2)/1024/1024
> = 1,907,349MB.
> 
> Your going to have to manage scrolling yourself and only make the canvas
> slightly larger than the current viewport size.
> 
> On Mon, 2005-03-14 at 16:24, Joe Van Dyk wrote:
> > Say I have a 2D space that's potentially very large.  I need to be
> > able to display various player icons on this world (they rotate and
> > move around in this space) and draw a grid on it.
> >
> > Players shouldn't really be limited to a certain area of movement,
> > they should be able to move pretty darned far.  So if I initialize a
> > gnomecanvas that's 1000000x1000000, they could probably move off of
> > that and I'd have to resize the viewable area.
> >
> > And if I initialize a huge world up front and draw grids all over it,
> > the application takes a long time to start.
> >
> > So, what I really need to do is make the world big enough to contain
> > all the players, even when the players move around (the world should
> > expand as the players move away from the center).
> >
> > Ideas on how to properly implement something like this?  Another
> > problem I've run in to is that when the world / viewable area is very
> > large, the scroll bars don't provide very good movement (think
> > scrolling in a million page document or something).
> >
> > Thanks,
> > Joe
> > _______________________________________________
> > gtk-list mailing list
> > gtk-list gnome org
> > http://mail.gnome.org/mailman/listinfo/gtk-list
> 
>