diff options
Diffstat (limited to 'compiler/GHC/Tc/Errors.hs')
| -rw-r--r-- | compiler/GHC/Tc/Errors.hs | 6 |
1 files changed, 1 insertions, 5 deletions
diff --git a/compiler/GHC/Tc/Errors.hs b/compiler/GHC/Tc/Errors.hs index c2c4c2c53b..71b919b4fd 100644 --- a/compiler/GHC/Tc/Errors.hs +++ b/compiler/GHC/Tc/Errors.hs @@ -2736,16 +2736,12 @@ Consider (C (Maybe (F x))), where F is a type function, and we have instances C (Maybe Int) and C (Maybe a) Since (F x) might turn into Int, this is an overlap situation, and -indeed (because of flattening) the main solver will have refrained +indeed the main solver will have refrained from solving. But by the time we get to error message generation, we've un-flattened the constraint. So we must *re*-flatten it before looking up in the instance environment, lest we only report one matching instance when in fact there are two. -Re-flattening is pretty easy, because we don't need to keep track of -evidence. We don't re-use the code in GHC.Tc.Solver.Canonical because that's in -the TcS monad, and we are in TcM here. - Note [Kind arguments in error messages] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ It can be terribly confusing to get an error message like (#9171) |