[GHC] #14559: When overlapping instances lead to a compile error, show import chains for the instances

[GHC]

#14559: When overlapping instances lead to a compile error, show import chains for
the instances
-------------------------------------+-------------------------------------
           Reporter:  Artyom.Kazak   |             Owner:  (none)
               Type:  feature        |            Status:  new
  request                            |
           Priority:  normal         |         Milestone:
          Component:  Compiler       |           Version:  8.2.1
           Keywords:                 |  Operating System:  Unknown/Multiple
       Architecture:                 |   Type of failure:  None/Unknown
  Unknown/Multiple                   |
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):                 |         Wiki Page:
-------------------------------------+-------------------------------------
 Here's an example from work:

 {{{
     • Overlapping instances for ToJSON BackupPhrase
         arising from a use of 'aesonRoundtripProp'
       Matching instances:
         instance ToJSON BackupPhrase
           -- Defined in 'Cardano.Wallet.Orphans.Aeson'
         instance ToJSON BackupPhrase
           -- Defined in 'Pos.Wallet.Aeson.ClientTypes'
 }}}

 Currently resolving such issues requires looking carefully at all imports
 and trying to guess which ones bring the instance into scope. It'd be
 quite helpful if the error was explicit about the import chain:

 {{{
     • Overlapping instances for ToJSON BackupPhrase
         arising from a use of 'aesonRoundtripProp'
       Matching instances:
         instance ToJSON BackupPhrase
           -- Defined in 'Cardano.Wallet.Orphans.Aeson',
           --   imported by 'Cardano.Wallet.Orphans'
         instance ToJSON BackupPhrase
           -- Defined in 'Pos.Wallet.Aeson.ClientTypes',
           --   imported by 'Pos.Wallet.Web.Mode',
           --   and then by 'Cardano.Wallet.API.V1.Migration'
 }}}

 Thoughts?

 I'm worried about two things here:

   1. Might these chains get too long?

   2. Is it sufficiently obvious that “imported by `A` and then by `B`”
 means that you have imported `B` and `B` imported `A`, and not vice-versa?

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14559>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler