On 10 Oct 2014, at 23:20, Lester Caine <[email protected]> wrote:
> Is this the right approach to implement BIGINT?
> I don't see the use of GMP to implement something as simple as native 64
> bit numbers on 64 bit platforms as the right base.
Um, we already have this since the 64-bit patch.
> All we are missing is
> correctly handling two word data on a 32bit platform. The wrapping and
> everything still applies, but only on a 32 bit platform ... no need for
> the complication of GMP. The bit I'm looking at here IS using BIGINT as
> array keys without the problems of them changing to long strings and I
> don't see how GMP fixes that?
What you want is 64-bit data handling. This is arbitrary-bit data handling. It’s not a “wrong
approach”.
--
Andrea Faulds
http://ajf.me/