[cairo] Cairo version now at 1.5.1

[Vladimir Vukicevic]

Behdad Esfahbod wrote:
> On Thu, 2007-06-28 at 08:25 -0400, Adrian Johnson wrote:
>> Carl Worth wrote:
>>> But I also think I was unfair to call Vlad's work "24.8 work". I
>>> believe his code is just a #define away from supporting any of the
>>> options you propose above. So we can definitely explore things easily
>>> here.
>>>
>>> That said, anybody have any interesting opinions in that discussion?
>> In the PS/PDF backends where 1 device unit is 1/72 inch there needs to
>> be enough precision to position objects to the accuracy of the printer
>> resolution.
> 
> Right.  28.4 gives us 1152 dpi precision.  That's even less than 1200
> dpi which is the norm these days.  26.6 goes up to 4608 dpi which should
> be enough for a few more years.  24.8 goes up to 18432 dpi that is
> probably more than we will ever need.

24.8 also gives us a coordinate space that's -8,388,608 .. 8,388,607, or 
somewhere around +/- 116,508 inches at 72dpi.  That, I think, should 
also be sufficient for quite a while.  There will probably be some use 
cases of people wanting to handle extremely large data sets with a high 
amount of precision, but I don't think any fixed point choice will 
handle those cases; that should be up to the application to handle at a 
level higher than cairo.

But as Carl said, as long as the fixed point type is 32 bits, it's just 
one #define to change where the point goes.  If we ever want to go more 
than 32 bits, it'd be a bit more work (but not a lot).

     - Vlad