Isaac Dupree wrote:
would implementing that make it harder or easier for ghci to support
unboxed tuples?
Slightly easier, in that construction (of unboxed tuples with lifted
components only) could be done by calling the curried construction
functions in PrimopWrappers.
Deconstruction could be done in a similar way; that is we need a family of
functions like
liftUbx2 :: (State# -> (# a, b #)) -> (a, b)
the State# argument is just a hack to lift the unboxed tuple. So you could
then compile
case f x of
(# a, b #) -> ...
into
let f' = \s -> f x in
case liftUbx2 f' of
(a, b) -> ...
but this only works for unboxed tuples with lifted components. You could
also generate some similar wrappers for particular instances of unboxed
tuples with unlifted components. Doesn't sound particularly attractive,
though.
Cheers,
Simon
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
