On Fri, 2009-05-29 at 06:17 -0700, Simon Peyton Jones wrote:
> Fri May 29 06:11:37 PDT 2009  [email protected]
>   * Implement -XMonoLocalBinds: a radical new flag
>   Ignore-this: b52744bdde2e8ea52a9b6d4374a3e049

>   The new flag -XMonoLocalBinds tells GHC not to generalise nested
>   bindings in let or where clauses, unless there is a type signature,
>   in which case we use it.  

If eventually you intend this to be for general consumption then
remember to register the extension in Language.Haskell.Extension. The
key test is whether packages uploaded to hackage should be able to
specify that they use the extension.

There's nothing wrong with experimental unregistered extensions. For
example -XPArr is still not registered (and when it's ready we're
assuming it'll be renamed to ParallelArrays). However unregistered
extensions cannot be used in distributed packages.

Registering involves making a patch to Language.Haskell.Extension
including a brief description (and example?) and a link to further
description (eg a paper or ghc user guide). The patch should be sent to
the libraries list so people can complain about clashing/duplicate/bad
names.

Yes, this is a new hard line policy! :-) Ian was the first to be hit
with it when he added a late extension for 6.10. Niklas Broberg has
recently been doing the same for his XML and pattern extensions. I'm
also hoping to get someone to properly document / cross-reference the
existing list of extensions.

Thanks

Duncan

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

Reply via email to