in reply to Re^3: Cygwin perl and (internal) numeric representations.
in thread Cygwin perl and (internal) numeric representations.

As to your question about using 64-bit integers on 32-bit machine, well that allows to exceed 2**32

Yes, but I can use integers up to 53 bits (before losing precision) on perls that haven't been compiled with use64bitint ... and on this Cygwin perl 5.8.7 I still only get integers up to 53 bits (before losing precision). So, in this instance, the use64bitint has achieved nothing.

My first reading of that thread you linked to leads me to believe that use64bitint should provide 64 bits on perl 5.8 (but is guaranteed only to provide 53 bits on perl 5.6). But, since the Cygwin perl is 5.8.7, use64bitint should be providing the full 64 bits ..... I'll have to go back and go through that Google thread carefully.

Thanks for steering me back onto the track.

Cheers,
Rob
  • Comment on Re^4: Cygwin perl and (internal) numeric representations.