validate fails on MacOS

[Matthias Kilian]

On Sat, Jan 17, 2009 at 10:46:38PM +1100, Manuel M T Chakravarty wrote:
> Thanks for writing the patch.  Nobody assumed you did that for no  
> reason.  However, a patch that fixes one and breaks another  
> architecture ultimately does not fix the problem.  From what you are  
> writing, it doesn't even seem to be clear that the patch does fix  
> NetBSD - as you haven't got a NetBSD box.

It wasn't the purpose of the patch to fix the NetBSD problem mentioned
on glasgow-haskell-users. The purpose was to help finding potential
errors in locking/unlocking as soon as they happen.

> Generally, patches that are potentially architecture dependent, should  
> be tested on multiple architectures before being committed:
> 
>   http://hackage.haskell.org/trac/ghc/wiki/TestingPatches

I'm sure Simon did test it before pushing. But aparently not on MacOS X.

Ciao,
	Kili