Compile error regression from GHC 7.10 to 7.11

Consider the following program,

{-# LANGUAGE OverloadedStrings #-}
-- {-# LANGUAGE FlexibleContexts #-}

chunksOf :: Int -> String -> [String]
chunksOf n = go
  where
    -- go :: String -> [String]
    go "" = []
    go s@(_:_) = a : go b
      where
        (a,b) = splitAt n s

when compiled with GHC 7.8.4:

GHCi, version 7.8.4: http://www.haskell.org/ghc/  :? for help
Loading package ghc-prim ... linking ... done.
Loading package integer-gmp ... linking ... done.
Loading package base ... linking ... done.
[1 of 1] Compiling Main             ( chunksof.hs, interpreted )

chunksof.hs:8:5: Warning:
    Pattern match(es) are overlapped
    In an equation for ‘go’: go s@(_ : _) = ...
Ok, modules loaded: Main.
λ:2> 

when compiled with GHC 7.10:

GHCi, version 7.10.1.20150630: http://www.haskell.org/ghc/  :? for help
[1 of 1] Compiling Main             ( chunksof.hs, interpreted )

chunksof.hs:8:5:
    Non type-variable argument in the constraint: Data.String.IsString [t]
    (Use FlexibleContexts to permit this)
    When checking that ‘go’ has the inferred type
      go :: forall t. (Eq t, Data.String.IsString [t]) => [t] -> [[t]]
    In an equation for ‘chunksOf’:
        chunksOf n
          = go
          where
              go "" = []
              go s@(_ : _)
                = a : go b
                where
                    (a, b) = splitAt n s
Failed, modules loaded: none.

NB: FlexibleContexts is rightly suggested!

However, when compiled with GHC HEAD:

GHCi, version 7.11.20150630: http://www.haskell.org/ghc/  :? for help
[1 of 1] Compiling Main             ( chunksof.hs, interpreted )

chunksof.hs:8:8: error:
    Could not deduce (IsString [t]) arising from the literal ‘""’
    from the context: Eq t bound by the inferred type of go :: Eq t => [t] -> [[t]] at chunksof.hs:(8,5)-(11,27)
    In the pattern: ""
    In an equation for ‘go’: go "" = []
    In an equation for ‘chunksOf’:
        chunksOf n
          = go
          where
              go "" = []
              go s@(_ : _)
                = a : go b
                where
                    (a, b) = splitAt n s
Failed, modules loaded: none.

Trac metadata

Trac field	Value
Version	7.11
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Compiler
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system
Architecture

changed weight to 5

added Tbug Trac import labels

Harump. This was deliberate:

commit b83160d07e626bee685f329a9a73e90a4a6074ae
Author: Simon Peyton Jones <simonpj@microsoft.com>
Date:   Thu Apr 30 09:18:49 2015 +0100

    Tidy up treatment of FlexibleContexts
    
    Previously (Trac #10351) we could get
    
        Non type-variable argument in the constraint: C [t]
        (Use FlexibleContexts to permit this)
        When checking that `f' has the inferred type
          f :: forall t. C [t] => t -> ()
    
    which is a bit stupid: we have *inferred* a type that we
    immediately *reject*.  This patch arranges that that the
    generalisation mechanism (TcSimplify.decideQuantification)
    doesn't pick a predicate that will be rejected by the
    subsequent validity check.

Moreover #10351 (closed) comment 3 makes essentially the same point as you do here.

I'll just reverse the change, which is easy to do. Two people have complained about the new behaviour and I don't think anyone complained about the old!

Simon

Replying to [ticket:10608#comment:102069 simonpj]:

I'll just reverse the change, which is easy to do. Two people have complained about the new behaviour and I don't think anyone complained about the old!

Please don't count me yet as complaining... but rather as wondering :-)

The new error message hint

Could not deduce (IsString [t]) arising from the literal ‘""’

it provides some additional information the previous message didn't. The only downside of the new message is that it doesn't also give the hint of what an explicit type-sig would look like or that FlexibleContexts may let GHC continue

closed

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedFixed

mentioned in commit b5aabfbd

changed milestone to %8.0.1

Milestone renamed

added Pnormal label

mentioned in issue #20686

mentioned in merge request !10123 (merged)