Sharing recovery fix

[sseefried]

The NodeCounts data structure has changed substantially. It is no longer a list of
node/count pairs with the invariant that all child nodes appear after parent nodes. It is
now a list of "dependency groups" each dependency group is a directed acyclic graph (DAG)
that captures the dependencies that an AccSharing node can have on VarSharing nodes.

The function 'filterCompleted' now checks and filters out nodes on a per-dependency group
basis.

This commit also introduces a new trace function in D.A.A.Debug and adds tests for sharing
recovery.

There are also some new dependencies added to accelerate.cabal because we are using
unordered containers.

[sseefried]

I realise you've done a lot of changes. This would actually be best pulled into master after commit c16f19d.

I'm loathe to combine this patch with your partially complete recovery of scalar expressions.

[sseefried]

Just got your last comment. Yes, I realise this conflicts with your last chunk of code. I've mainly sent it through just so you can validate it. If it passes muster then I will integrate with your stuff, or you can do the same.

[mchakravarty]

Thanks again for the fix and examples that highlighted an important problem in the original implementation. I actually managed to find a significantly simpler fix. Instead of tracking entire dependency groups, it is sufficient to track the height of shared trees, and then, to ensure that the entries in NodeCounts are ordered by height (higher ones first). When scanning for completed subtrees (to generate let bindings), only pick entries from the front of the list.

This ensures a well-formed result as subtrees are necessarily less high than the parent trees that contain them.

[mchakravarty]

(This fix is currently only on my "sharing" branch.)