[Haskell-cafe] Eta Reduction
dan.doel at gmail.com
Thu Apr 3 01:18:04 UTC 2014
On Wed, Apr 2, 2014 at 8:40 PM, Niklas Haas <haskell at nand.wakku.to> wrote:
> Oh, fair point; this particular F apparently doesn't break it but if you
> remove the Show x constraint, it does.
> Actually, is that a bug in GHC?
> > λ newtype F = F (forall x. Show x => x -> String)
> > λ F undefined `seq` ()
> > ()
> > λ undefined `seq` ()
> > *** Exception: Prelude.undefined
> I'm not sure how to interpret this output.
This is pretty weird, but here's what I think is going on:
F requires an argument of type forall x. Show x => x -> String. This
requires abstracting over a dictionary for Show x. So at the core level,
this gets expanded to something like:
\showd -> undefined
which is non-bottom.
Even if you annotate undefined with the above type, You'll probably end up
(\showd -> undefined) defaultd `seq` ()
after GHC reinstantiates the polymorphic type and then defaults, which will
be undefined. So you can only observe this by wrapping the polymorphic
expression in a newtype, to keep it abstracted.
I don't know that this qualifies as a bug, but it's definitely pretty
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Haskell-Cafe