Re: gEDA-user: PCB nanometer git tree

[karl@xxxxxxxxxxx (Karl Hammar)]

Andrew Poelstra:
> On Thu, Aug 11, 2011 at 09:30:34PM +0200, Karl Hammar wrote:
...
> > -#define LARGE_VALUE            (LONG_MAX / 2 - 1) /* maximum extent of board and elements */
> > +#define LARGE_VALUE            LONG_MAX /* (LONG_MAX / 2 - 1) maximum extent of board and elements */
...
> > Why do you want to use only half the range?
>
> So that the width of a "max width" box can be calculated as
> LARGE_VALUE - (-LARGE_VALUE) without overflow,

Hmm, wouldn't it be better to use unsigned here?
For

 typedef unsigned long Coord;

 Coord width( Coord a, Coord b ) {
   if (a > b) {
     return a - b;
   } else {
     return b - a;
   }
 }

 do work without overflow for all possible walues of a and b.

And we don't really need a signed type, do we?

> and so that two maximum widths can be averaged as (x + x) / 2.

Isn't that most easily fixed by using a wider type during addition?

 Coord mean( a, b) {
  bigger_int c = ((bigger_int) a + (bigger_int) b);
  return (Coord) c;
 }

Also, most (all?) arch. have a carry bit, which means you could
possible have a n+1 wide type during adding and div by 2. Maybe that
would involve to much hackery.

Well, not any great consern, just thinking aloud.

Regards,
/Karl Hammar

-----------------------------------------------------------------------
Aspö Data
Lilla Aspö 148
S-742 94 Östhammar
Sweden
+46 173 140 57




_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user