Proposed significant breaking changes to Win32

[Elliot Robinson]

A quick reverse search [0] shows 74 packages having Win32 as a direct dependency, 3 of which are out of date. The big ones that stand out to me are `directory` (which will be up to Phil and me to fix) and `process` (which is maintained by Michael Snoyman). Both are GHC dependencies, and each has 800+ reverse dependencies, so the release timing there will need to be well managed (possibly to coincide with the next major GHC RC?).

This being said, I'm definitely in favor of better Win32 support. If it takes some extra effort to fix long-standing defects, that's the price of doing business.

[0] http://packdeps.haskellers.com/reverse/Win32

On 03/18/15, Yitzchak Gale wrote:
> There are some discussions going on in this issue:
> 
> https://github.com/haskell/win32/issues/24
> 
> with proposals for extensive breaking changes
> to the Win32 library. Well over 100 functions would
> be involved. These changes would finally fix many
> long outstanding defects in the API, some of which
> make it impossible to use important Win32 functions
> in the way originally intended.
> 
> I would generally support these proposals (after having
> seen that Neil supports them). But does anyone have
> any estimate of how much breakage it would cause on
> Hackage, or in the Windows builds of GHC?
> 
> Thanks,
> Yitz
> _______________________________________________
> Libraries mailing list
> Libraries at haskell.org
> http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries

--
Elliot Robinson
GPG Key: 9FEDE59A