Redundant specialise pragmas

[Duncan Coutts]

On Thu, 2009-10-08 at 13:11 +0000, Simon Peyton-Jones wrote:
> Duncan, Don
> 
> GHC will soon start warning about redundant SPECIALISE pragmas.  A few
> weeks ago, I mentioned that bytestring has two such examples.  Could
> you remove them from the upstream bytestring, please?  Otherwise we
> won't be able to validate.

Done.

Note that ghc HEAD is currently using a tarball of the release and not a
time-lagged mirror of the upstream bytestring darcs repo.

Ian will need to hasten the switch over from the tarball back to the
repo before this warning becomes fatal. I think he was hoping to do that
slowly so that people using the old bytestring repo (remember we had an
accidental ghc fork) have longer to switch away from the old repo.

Another option is to not make warnings fatal in the libs during the
early part of the development cycle.

Duncan