[Haskell-cafe] naturally, length :: a -> Int

[Carter Schonwald]

I think we’d also wanna add compiler suport for ints and words and signal
on over/underflow using cpu machinery .  Most other approachs have pretty
fun changes in memory rep

On Tue, Mar 2, 2021 at 11:10 AM Sylvain Henry <sylvain at haskus.fr> wrote:

> On 01/03/2021 17:45, Johannes Waldmann wrote:
> > Dear Cafe -
> >
> >
> > let's fight Int.
>
> I'm in too.
>
> One first step would be to use `Word#` instead of `Int#` in ghc primops
> where it makes sense. This could be done as it should be transparent to
> most users.
>
> The natural next step would be to modify functions and classes in `base`
> to use appropriate types too. I don't think it's going to happen as it
> would require modifying the Haskell report and breaking almost every
> existing code/books/etc.
>
> So an alternative path would be to extract every wired-in things from
> `base` and to put them into a `ghc-base` package that `base`,
> `foundation` and other alternative preludes would depend on. This way we
> could write codes that don't depend on `base` and the legacy stuff it
> contains at all (even transitively).
>
> Could you open a ticket on GHC's bug tracker about this?
>
> Cheers,
> Sylvain
>
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> Only members subscribed via the mailman list are allowed to post.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20210303/1ab136b0/attachment.html>