[cairo] Some other bits...

[Keith Packard]

On Wed, 2005-08-17 at 00:24 -0400, Behdad Esfahbod wrote:

> cairo_fixed_t
>  _cairo_fixed_from_26_6 (uint32_t i)
>  {
> -    return i << 10;
> +    return i << (CAIRO_FIXED_FLOAT_BITS - 10);
 
This should be - 6, not - 10, I think.
> -       factor -= factor & 0xffff0000;
> +       factor &= 0xffff;
>         break;
>      case CAIRO_EXTEND_REFLECT:
>         if (factor < 0 || factor > 65536) {
>             if ((factor >> 16) % 2)
> -               factor = 65536 - (factor - (factor & 0xffff0000));
> +               factor = 65536 - (factor & 0xffff);
>             else
> -               factor -= factor & 0xffff0000;
> +               factor &= 0xffff;
>         }

Aren't all of these supposed to scale with CAIRO_FIXED_FLOAT_BITS?

-    cairo_fixed_48_16_t a_dx = a->edge.p2.x - a->edge.p1.x;
-    cairo_fixed_48_16_t a_dy = a->edge.p2.y - a->edge.p1.y;
-    cairo_fixed_48_16_t b_dx = b->edge.p2.x - b->edge.p1.x;
-    cairo_fixed_48_16_t b_dy = b->edge.p2.y - b->edge.p1.y;
+    cairo_fixed_48_16_t a_dx = (cairo_fixed_48_16_t) a->edge.p2.x -
a->edge.p1.x;
+    cairo_fixed_48_16_t a_dy = (cairo_fixed_48_16_t) a->edge.p2.y -
a->edge.p1.y;
+    cairo_fixed_48_16_t b_dx = (cairo_fixed_48_16_t) b->edge.p2.x -
b->edge.p1.x;
+    cairo_fixed_48_16_t b_dy = (cairo_fixed_48_16_t) b->edge.p2.y -
b->edge.p1.y;

Why do you believe these need to be done in 64 bits? Are you concerned
that the subtract might overflow?

-keith

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.freedesktop.org/archives/cairo/attachments/20050816/298577bd/attachment.pgp