patch applied (ghc): Avoid hanging on to old unfoldings; fixes #4367 (compiler space regression)

[Ian Lynagh]

Hi Simon,

On Thu, Oct 21, 2010 at 07:57:28AM +0000, Simon Peyton-Jones wrote:
> 
> Don't you think it'd be better to put that `seq` into 'mkUnfolding'
> rather than into the calls to mkUnfolding?

Hmm, well, it's at the call site that we are creating (isBottomingId
id), and trying to avoid holding on to (id), so having the seq where it
is is clearer to me. We could do it in mkUnfolding itself, but there
we'd just be forcing (is_bottoming) without it being obvious why forcing
it is beneficial.

I don't feel strongly, though; what do you think?


Thanks
Ian