Opened 7 years ago
Closed 5 years ago
#2544 closed bug (fixed)
Improve "Can't unify" error messages from type functions
Reported by: | simonpj | Owned by: | |
---|---|---|---|
Priority: | low | Milestone: | 7.0.2 |
Component: | Compiler (Type checker) | Version: | 6.8.3 |
Keywords: | Cc: | jwlato@… | |
Operating System: | Unknown/Multiple | Architecture: | Unknown/Multiple |
Type of failure: | None/Unknown | Test Case: | indexed-types/should_fail/T2544 |
Blocked By: | Blocking: | ||
Related Tickets: | Differential Rev(s): | ||
Wiki Page: |
Description
Consider this example from a Haskell Cafe thread:
data (:|:) a b = Inl a | Inr b class Ix i where type IxMap i :: * -> * empty :: IxMap i [Int] data BiApp a b c = BiApp (a c) (b c) instance (Ix l, Ix r) => Ix (l :|: r) where type IxMap (l :|: r) = BiApp (IxMap l) (IxMap r) empty = BiApp empty empty
This elicits the following confusing error message:
Couldn't match expected type `IxMap l' against inferred type `IxMap i' Expected type: IxMap (l :|: r) [Int] Inferred type: BiApp (IxMap i) (IxMap i1) [Int] In the expression: BiApp empty empty In the definition of `empty': empty = BiApp empty empty
As Alexander Dunlap responds, the error message is correct, but I can't help feeling that we should try harder to give a better error message. Something like "Since IxMap is a type function, knowing that IxMap l = IxMap i does not require that l = i". Or something.
This ticket is just to make sure we don't forget. The thread is here http://www.haskell.org/pipermail/haskell-cafe/2008-August/046371.html
Simon
Change History (7)
comment:1 Changed 7 years ago by simonmar
- Architecture changed from Unknown to Unknown/Multiple
comment:2 Changed 7 years ago by simonmar
- Operating System changed from Unknown to Unknown/Multiple
comment:3 Changed 5 years ago by igloo
- Milestone changed from 6.12 branch to 6.12.3
comment:4 Changed 5 years ago by igloo
- Milestone changed from 6.12.3 to 6.14.1
- Priority changed from normal to low
comment:5 Changed 5 years ago by igloo
- Milestone changed from 7.0.1 to 7.0.2
comment:6 Changed 5 years ago by jwlato
- Cc jwlato@… added
- Type of failure set to None/Unknown
comment:7 Changed 5 years ago by simonpj
- Resolution set to fixed
- Status changed from new to closed
- Test Case set to indexed-types/should_fail/T2544
It's even improved a bit more in HEAD
T2544.hs:15:12: Could not deduce (IxMap l ~ IxMap i0) from the context (Ix l, Ix r) bound by the instance declaration at T2544.hs:13:10-37 NB: `IxMap' is a type function, and may not be injective Expected type: IxMap (l :|: r) [Int] Actual type: BiApp (IxMap i0) (IxMap r) [Int] In the return type of a call of `BiApp' In the expression: BiApp empty empty In an equation for `empty': empty = BiApp empty empty T2544.hs:15:24: Could not deduce (IxMap i1 ~ IxMap r) from the context (Ix l, Ix r) bound by the instance declaration at T2544.hs:13:10-37 NB: `IxMap' is a type function, and may not be injective Expected type: IxMap r [Int] Actual type: IxMap i1 [Int] In the second argument of `BiApp', namely `empty' In the expression: BiApp empty empty In an equation for `empty': empty = BiApp empty empty
So let's close. I'll add a test for completeness.
Thanks for bringing it up.
Simon
With ghc-7.0.1, the error message is now:
Can this be closed? These error messages are much more useful (at least to my eye).