intermittent type checker AWOL bug?
Simon Peyton-Jones
simonpj at microsoft.com
Fri Jun 27 12:36:22 EDT 2008
Interesting. It doesn't ring a bell, and I don't want to build a special compiler from Feb to check, so I'm just going to immortalise your code in the test suite, so if it ever happens again we'll know about it.
Thanks for fwding it.
Simon
| -----Original Message-----
| From: cvs-ghc-bounces at haskell.org [mailto:cvs-ghc-bounces at haskell.org] On Behalf Of Claus Reinke
| Sent: 18 June 2008 18:52
| To: cvs-ghc at haskell.org
| Subject: intermittent type checker AWOL bug?
|
| have any type checker loop bugs been fixed recently?
| I have a piece of code for which:
|
| GHCi 6.6.1 reports an error
| GHCi, version 6.9.20080217 loops
| GHCi, version 6.9.20080514 reports an error
|
| and I just want to make sure that this is a known and fixed
| bug, not something that has come out of hiding too briefly
| to be fixed properly.
|
| Claus
|
| --
| this erroneous code fragment sends `GHCi, version 6.9.20080217` into a loop:
| {{{
| {-# OPTIONS_GHC -fglasgow-exts #-}
| data HsDoc id
| = DocEmpty
| | DocParagraph (HsDoc id)
|
| gfoldl' :: (forall a b . c (a -> b) -> a -> c b) -> (forall g . g -> c g) -> a -> c a
| gfoldl' k z hsDoc = case hsDoc of
| DocEmpty -> z DocEmpty
| (DocParagraph hsDoc) -> z DocParagraph `k` hsDoc
| }}}
| I haven't waited more than a couple of minutes, because GHCi 6.6.1 immediately finds the bug:
| {{{
| C:/Documents and Settings/cr3/Desktop/Typo.hs:10:2:
| Couldn't match expected type `a' (a rigid variable)
| against inferred type `HsDoc id'
| `a' is bound by the type signature for `gfoldl''
| at C:/Documents and Settings/cr3/Desktop/Typo.hs:8:77
| In the pattern: DocEmpty
| In a case alternative: DocEmpty -> z DocEmpty
| In the expression:
| case hsDoc of
| DocEmpty -> z DocEmpty
| (DocParagraph hsDoc) -> (z DocParagraph) `k` hsDoc
| Failed, modules loaded: none.
| }}}
| I was about to create a ticket for this when I recalled having a
| newer ghc to check with and, surprise, GHCi, version 6.9.20080514
| does not suffer from this issue.
|
|
| _______________________________________________
| Cvs-ghc mailing list
| Cvs-ghc at haskell.org
| http://www.haskell.org/mailman/listinfo/cvs-ghc
More information about the Cvs-ghc
mailing list