blob: be21d5f84c7dff8308ec920accc62ea2d8536b00 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
{-# OPTIONS_GHC -fno-warn-redundant-constraints #-}
{-# LANGUAGE MultiParamTypeClasses, FunctionalDependencies,
FlexibleInstances, UndecidableInstances #-}
-- UndecidableInstances because (L a b) is no smaller than (C a b)
-- This one shows up another rather subtle functional-dependency
-- case. The error is:
--
-- Could not deduce (C a b') from the context (C a b)
-- arising from the superclasses of an instance declaration at Foo.hs:8:0
-- Probable fix: add (C a b') to the instance declaration superclass context
-- In the instance declaration for `C (Maybe a) a'
--
-- Since L is a superclass of the (sought) constraint (C a b'), you might
-- think that we'd generate the superclasses (L a b') and (L a b), and now
-- the fundep will force b=b'. But GHC is very cautious about generating
-- superclasses when doing context reduction for instance declarations,
-- because of the danger of superclass loops.
--
-- So, today, this program fails. It's trivial to fix by adding a fundep for C
-- class (G a, L a b) => C a b | a -> b
-- Note: Sept 08: when fixing #1470, tc138 started working!
-- This test is a very strange one (fundeps, undecidable instances),
-- so I'm just marking it as "should-succeed". It's not very clear to
-- me what the "right" answer should be; when we have the type equality
-- story more worked out we might want to think about that.
module ShouldFail where
class G a
class L a b | a -> b
class (G a, L a b) => C a b
instance C a b' => G (Maybe a)
instance C a b => C (Maybe a) a
instance L (Maybe a) a
|
