[Haskell-cafe] Strange problem with inference

[Nikolay Amiantov]

An interesting fix, indeed! I think this might be a bug in GHC -- these
should (if I understand correctly) be equivalent signatures. I'll try to
find a simpler example too before reporting this, it might shed some
light on things (maybe even enough to show this isn't a bug at all ^_^).

On 09/11/2015 05:17 AM, Sumit Sahrawat, Maths & Computing, IIT (BHU) wrote:
> One possible fix (tested on GHC-7.10.1 with lens-4.12.3):
> 
> test2 :: (Test a, t ~ TestT a) => t -> a
> test2 = view (from myiso)
> 
> This might have something to do with type families not being injective,
> but I'm not completely sure.
> 
> I also agree that it might be possible to trigger this without lens,
> will try to find an example and post if I succeed.
-- 
Nikolay.