#4981: inconsistent class requirements with TypeFamilies and FlexibleContexts
---------------------------------+------------------------------------------
Reporter: ganesh | Owner:
Type: bug | Status: new
Priority: normal | Component: Compiler (Type checker)
Version: 7.0.1 | Keywords:
Testcase: | Blockedby:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: GHC rejects valid program
---------------------------------+------------------------------------------
Changes (by dimitris):
* component: Compiler => Compiler (Type checker)
Comment:
Right, I know what's going on. The problem we are solving is:
{{{
[Given] Conflict (OnPrim p),
+ its superclass: [Given] (PatchInspect (PrimOf (FL (OnPrim p))))
[Given] PrimOf (OnPrim p) ~ WithName (PrimOf p)
[Wanted] Conflict (FL (OnPrim p))
}}}
Now, we apply the instance to get new work:
{{{
[Wanted] Conflict (OnPrim p)
[Wanted] PatchInspect (PrimOf (FL (OnPrim p)))
}}}
The second guy is the "silent parameter" wanted. Now, the Conflict wanted
can be
readily discharged so let's not worry about him. However:
{{{
[Wanted] PatchInspect (PrimOf (FL (OnPrim p))) ~~>
[Wanted] PatchInspect (PrimOf (OnPrim p))
}}}
Now, this ''could'' be readily discharged by the given but instead what
happens
is that it gets rewritten with the given equality to:
{{{
[Wanted] PatchInspect (WithName (PrimOf p))
}}}
Then the instance triggers, and we get:
{{{
[Wanted] PatchInspect (PrimOf p)
}}}
which is the error we see.
So indeed this program is in muddy territory where givens (superclasses of
givens, actually!) overlap with top-level instances.
I know why GHC is not picking the given up: it has to do with the fact
that we have
not saturated all possible equalities before we look for instances, but
luckily this
is something Simon and I are planning to fix pretty soon. The other thing
that we should also do is remove these silent superclass parameters --
they used to exist
for reasons related to recursive dictionaries but they are not necessary
any more
the way we have dealt with recursive dictionaries at the end.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4981#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
