FFI Bindings to Libraries using GMP

Isaac Dupree isaacdupree at charter.net
Tue Sep 25 09:27:39 EDT 2007

Peter Tanski wrote:
> The one problem you will find with _all_ potential 
> replacement libraries is incompatible behaviour for bitwise functions: 
> they are implemented arithmetically in GMP but logically elsewhere (when 
> they are implemented at all).

I don't fully understand this... I made sure my Haskell implementation's 
"Bits" was consistent with the existing (GMP) version, which you call 
"arithmetically"... it seemed that any other semantics would reveal 
(depend on) implementation details.  and I was able to make it 
reasonably asymptotically-efficient anyway.


More information about the Glasgow-haskell-users mailing list