deepseq: instance NFData (a -> b)

[Henning Thielemann]

On Sat, 7 May 2016, Edward Kmett wrote:

> Let's turn this around, is there anybody actually in favor of keeping
> the instance?

Looking in the git-logs I found git-commit 
3b5c957ce7bba7b63b4483a43c6762c3f5d8ee28 by Herbert Valerio Riedel saying:

"Merge `deepseq-generics` into `deepseq` ..."

and it points to the proposal:
    http://permalink.gmane.org/gmane.comp.lang.haskell.libraries/23031


So, I guess Herbert must be in favor of the instance because he added it. 
:-)

And this commit confirms that the instance NFData (a->b) was a side 
product of Generics support.