[cairo] big point handling

[Vladimir Vukicevic]

Behdad Esfahbod wrote:
> On Tue, 2007-06-12 at 12:55 -0400, Dan Amelang wrote:
>> Could be. I bet what you'd really like is for there to be no error at
>> all. If we move to 24.8 fixed as planned, we avoid the whole overflow
>> issue.
> 
> Not really.  It would just make it overflow at a further point.

Yep.  Though if we're going down the mitigation route, the move to 24.8 
work should really be a move to a compile-time selection of precision 
and type size -- would be nice to be able to use 64-bit types if 
necessary (though that would mean needing to use 128-bit intermediate 
types in a few places, I think).

     - Vlad