[commit: ghc] master: Reify data family instances correctly. (e319d6d)

[Joachim Breitner]

Hi Richard,

Travis is complaining about your commit:

Actual stderr output differs from expected:
--- ./th/T9692.stderr	2014-10-21 13:22:53.783212762 +0000
+++ ./th/T9692.comp.stderr	2014-10-21 13:52:36.314083513 +0000
@@ -1,2 +0,0 @@
-data family T9692.F (a_0 :: k_1) (b_2 :: k_3) :: *
-data instance T9692.F GHC.Types.Int x_4 = T9692.FInt x_4
*** unexpected failure for T9692(normal)
https://s3.amazonaws.com/archive.travis-ci.org/jobs/38604151/log.txt

It maybe be due to a missing flushing of stdout; I added that line and
will let you know if that was not it.

Greetings,
Joachim

Greetings,
Joachim

Am Dienstag, den 21.10.2014, 13:21 +0000 schrieb git at git.haskell.org:
> Repository : ssh://git@git.haskell.org/ghc
> 
> On branch  : master
> Link       : http://ghc.haskell.org/trac/ghc/changeset/e319d6d2704edc2696f47409f85f4d4ce58a6cc4/ghc
> 
> >---------------------------------------------------------------
> 
> commit e319d6d2704edc2696f47409f85f4d4ce58a6cc4
> Author: Richard Eisenberg <eir at cis.upenn.edu>
> Date:   Mon Oct 20 15:36:57 2014 -0400
> 
>     Reify data family instances correctly.
>     
>     Summary:
>     Fix #9692.
>     
>     The reifier didn't account for the possibility that data/newtype
>     instances are sometimes eta-reduced. It now eta-expands as necessary.
>     
>     Test Plan: th/T9692
>     
>     Reviewers: simonpj, austin
>     
>     Subscribers: thomie, carter, ezyang, simonmar
>     
>     Differential Revision: https://phabricator.haskell.org/D355
> 
> 
> >---------------------------------------------------------------
> 
> e319d6d2704edc2696f47409f85f4d4ce58a6cc4
>  compiler/typecheck/TcSplice.lhs | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/compiler/typecheck/TcSplice.lhs b/compiler/typecheck/TcSplice.lhs
> index bb6af8c..e952a27 100644
> --- a/compiler/typecheck/TcSplice.lhs
> +++ b/compiler/typecheck/TcSplice.lhs
> @@ -1338,8 +1338,16 @@ reifyFamilyInstance (FamInst { fi_flavor = flavor
>        DataFamilyInst rep_tc ->
>          do { let tvs = tyConTyVars rep_tc
>                   fam' = reifyName fam
> +
> +                   -- eta-expand lhs types, because sometimes data/newtype
> +                   -- instances are eta-reduced; See Trac #9692
> +                   -- See Note [Eta reduction for data family axioms]
> +                   -- in TcInstDcls
> +                 (_rep_tc, rep_tc_args) = splitTyConApp rhs
> +                 etad_tyvars            = dropList rep_tc_args tvs
> +                 eta_expanded_lhs = lhs `chkAppend` mkTyVarTys etad_tyvars
>             ; cons <- mapM (reifyDataCon (mkTyVarTys tvs)) (tyConDataCons rep_tc)
> -           ; th_tys <- reifyTypes lhs
> +           ; th_tys <- reifyTypes (filter (not . isKind) eta_expanded_lhs)
>             ; return (if isNewTyCon rep_tc
>                       then TH.NewtypeInstD [] fam' th_tys (head cons) []
>                       else TH.DataInstD    [] fam' th_tys cons        []) }
> 
> _______________________________________________
> ghc-commits mailing list
> ghc-commits at haskell.org
> http://www.haskell.org/mailman/listinfo/ghc-commits
> 

-- 
Joachim “nomeata” Breitner
  mail at joachim-breitner.de • http://www.joachim-breitner.de/
  Jabber: nomeata at joachim-breitner.de  • GPG-Key: 0xF0FBF51F
  Debian Developer: nomeata at debian.org

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20141022/67fa11a8/attachment.sig>