I VERY much doubt we'll do much to it unless it manifests in 6.10, so don't 
burn cycles on a test case.

As a matter of interest, though, if it's cheap to try, does -dcore-lint show up 
anything?

Simon

| -----Original Message-----
| From: [email protected] [mailto:[email protected]] On Behalf Of David
| Roundy
| Sent: 15 April 2009 17:42
| To: Simon Peyton-Jones
| Cc: [email protected]
| Subject: Re: error: Iface type variable out of scope:
|
| I've just tried doing the same compile with ghc 6.10, and it doesn't
| run into the same issue.  Would you still be interested in a test
| case, even though it seems to already be fixed?
|
| David
|
| P.S. This was definitely in a clean tree.  I thought it sounded like
| an incorrect .hi file, so I cleaned it again and again (thinking maybe
| I hadn't removed the right .hi file).
|
| On Wed, Apr 15, 2009 at 9:22 AM, Simon Peyton-Jones
| <[email protected]> wrote:
| > Usually this kind of thing indicates a stale .hi file.
| >
| > If it's stale, but GHC should know, then it's a bug.  But sometimes it
| happens because file modification times are messed up, or you changed version
| of GHC or something.
| >
| > If it happens in a clean tree, it's definitely a bug.
| >
| > It's not a "known bug" so far as I know, so a reproducible test case would
| be fantastic.
| >
| > Oh: use -dcore-lint. That often nails the error much earlier
| >
| > Simon
| >
| > | -----Original Message-----
| > | From: [email protected] [mailto:[email protected]] On
| Behalf Of
| > | David Roundy
| > | Sent: 15 April 2009 15:02
| > | To: [email protected]
| > | Subject: error: Iface type variable out of scope:
| > |
| > | Hello ghc folks,
| > |
| > | I'm getting the following error compiling darcs:
| > |
| > | [ghc] src/Darcs/Repository/HashedRepo.hs
| > | src/Darcs/Patch/Set.hi
| > | Declaration for a:
| > |   Iface type variable out of scope:  x10s
| > | Cannot continue after interface file error
| > |
| > | This is with the ghc 6.8.2 that is included in ubuntu Hardy Heron.
| > | The code is a bit large, and I haven't yet committed the changes that
| > | trigger this (since it won't compile), but I thought I'd just check
| > | whether this is a known bug before providing a better report.  A
| > | googling revealed a bug with the same error message (also revealed by
| > | darcs developers) that was in ghc 6.6, but doesn't seem to be related.
| > |  In that case, the workaround was to avoid using the GADT syntax.
| > |
| > | Any suggestions?
| > | --
| > | David Roundy
| > |
| > | _______________________________________________
| > | Cvs-ghc mailing list
| > | [email protected]
| > | http://www.haskell.org/mailman/listinfo/cvs-ghc
| >
| >
|
|
|
| --
| David Roundy

_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to