#4485: Unexplained change in type-inference + OverlappingInstances in
---------------------------------+------------------------------------------
Reporter: JeremyShaw | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 7.0.1 RC1
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: GHC rejects valid program
---------------------------------+------------------------------------------
Comment(by JeremyShaw):
Regarding, XMLGenT Identity (), I just meant that:
{{{
instance EmbedAsChild (IdentityT IO) FooBar where
asChild b = asChild $ ((genElement "foo") :: XMLGenT Identity ())
}}}
produces a type error. So clearly genElement does not actually have that
type. But I make no informed claims that the GHC type checker should be
able to figure that out.
Your explanation makes sense (and matches what is in the documentation).
The fact that things worked under GHC 6.12, made me wonder if there was
some other secret voodoo that I did not know about, or if 6.12 was just
conveniently broken.
I am a little bit frightened of enabling IncohorentInstances. If they
could be enabled just in the original module where the class is enabled,
that might not be so bad. But if I have to enabled them everywhere I used
the class, that is not so thrilling. In my experience, enabling
IncohorentInstances is almost never the right thing to do. Since I use
other classes and instances, it would also allow those to be incoherent
when I would really prefer a compiler error.
It would be somewhat nice if IncoherentInstances could be enabled on a
per-class basis due to its 'unsafe' nature.
thanks!
- jeremy
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4485#comment:5>
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
