+ALTERNATIVE 3: ignore the reboxing problem. The trouble is that
+the conservative reboxing story prevents many useful functions from being
+specialised. Example:
+ foo :: Maybe Int -> Int -> Int
+ foo (Just m) 0 = 0
+ foo x@(Just m) n = foo x (n-m)
+Here the use of 'x' will clearly not require boxing in the specialised function.
+
+The strictness analyser has the same problem, in fact. Example:
+ f p@(a,b) = ...
+If we pass just 'a' and 'b' to the worker, it might need to rebox the
+pair to create (a,b). A more sophisticated analysis might figure out
+precisely the cases in which this could happen, but the strictness
+analyser does no such analysis; it just passes 'a' and 'b', and hopes
+for the best.
+
+So my current choice is to make SpecConstr similarly aggressive, and
+ignore the bad potential of reboxing.
+