[GHC] #9426: ghci triggers 'ASSERT failed! file compiler/ghci/Linker.lhs, line 906'

[GHC]

#9426: ghci triggers 'ASSERT failed! file compiler/ghci/Linker.lhs, line 906'
-------------------------------------+-------------------------------------
              Reporter:  slyfox      |            Owner:
                  Type:  bug         |           Status:  new
              Priority:  normal      |        Milestone:
             Component:  Compiler    |          Version:  7.8.2
            Resolution:              |         Keywords:
      Operating System:              |     Architecture:  Unknown/Multiple
  Unknown/Multiple                   |       Difficulty:  Unknown
       Type of failure:              |       Blocked By:
  None/Unknown                       |  Related Tickets:
             Test Case:  ghci044     |
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------

Comment (by Simon Peyton Jones <simonpj@…>):

 In [changeset:"67a0cab6b501e2d6280b51655af66ad448b3deef/ghc"]:
 {{{
 #!CommitTicketReference repository="ghc"
 revision="67a0cab6b501e2d6280b51655af66ad448b3deef"
 Fix GHCi/GHC-API tidying and modules (Trac #9424, #9426)

 There were two related bugs here

 Trac #9426
    We must increment the ic_mod_index field of the InteractiveContext
    if we have new instances, because we maek DFunIds that should be
    distinct from previous ones.  Previously we were only incrementing
    when defining new user-visible Ids.

    The main change is in HscTypes.extendInteractiveContext, which now
    alwyas bumps the ic_mod_index.  I also added a specialised
    extendInteractiveContextWithIds for the case where we are *only*
    adding new user-visible Ids.

 Trac #9424
    In HscMain.hscDeclsWithLocations we were failing to use the
    *tidied* ClsInsts; but the un-tidied ones are LocalIds which
    causes a later ASSERT error.

    On the way I realised that, to behave consistently, the tcg_insts
    and tcg_fam_insts field of TcGblEnv should really only contain
    instances from the current GHCi command, not all the ones to date.
    That in turn meant I had to move the code for deleting replacement
    instances from addLocalInst, addLocalFamInst to
    HscTypes.extendInteractiveContext
 }}}

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9426#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler