#1781: Type equality class leads to non-termination
----------------------------------------+-----------------------------------
Reporter: chak | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler (Type checker) | Version: 6.9
Severity: normal | Resolution:
Keywords: | Difficulty: Unknown
Os: Unknown | Testcase:
Architecture: Unknown |
----------------------------------------+-----------------------------------
Changes (by chak):
* severity: blocker => normal
Comment:
Replying to [comment:2 guest]:
> I set the milestone to 6.8.1 and the severity to blocker, and I hope for
your understanding. The reason is that at the moment it seems to be
impossible to implement a HList-like equality test. At some point you
need a type equality constraint, either {{{t1 ~ t2}}} or {{{E t1 t2}}}
(with {{{E}}} being defined as above). But neither of them works! The
latter doesn’t work because of this very bug, the former doesn’t work
probably because of bug #1754. So if GHC 6.8.1 is released without this
bug being fixed I’ll probably have to tell my users (of Grapefruit) that
they need something like GHC 6.8.0.20070916. :-(
You may have misunderstood the problem exposed by this bug - maybe my
explanation just wasn't clear. The problem is '''not''' with the
equality class `E` as such. It's with this particular use in the
definition of `plus`, where the signature type contains a type variable
that is constrained to be a function by the class context. This is
something that has '''never''' worked in GHC, so I doubt that it breaks
any existing code.
Having said that, maybe you found some other problems with `E`. Then,
please document them in a separate bug (if you haven't yet). In any case,
this particular bug is surely no blocker, in fact it is fairly obscure and
I'd doubt anybody is going to run into it.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/1781#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
