DDC/ClosureTyping

From HaskellWiki

1 Region sharing

Consider the following function:

Without closure information this function would have the following type:

Remember that the at the front of the type is supposed to indicate that the return value is freshly allocated. This is certainly true if we apply both arguments: In there are different regions annotation on each of the constructors, which means they do not alias, and its safe to treat one as and the other as . But what happens if we partially apply ? The standard type system will re-generalize the type for the new binding and we're left with: We've now got a function which returns the same every time we call it, but the type says it's supposed to be fresh! The problem here is that was free in our original definition of so is shared between applications of it.

2 Closure typing for region sharing

Closure typing is used to track the sharing of regions between function calls like this one.

The type inferred for is actually: The annotation tells us that this function has a object free in its closure, and that the object contains region . The syntax gives a name to this term, and is pronounced "x of region zero".

Similarly to variables in the expression language, the exact names given to closure terms is not significant to the type system. However, to make inferred types easier to read, DDC names the closure terms in a function's type after the free variables that gave rise to those terms. The user is free to use whatever names they wish in type signatures. Two closure terms with different names but identical right hand sides are taken as being equivalent.

If we use this new type and apply the first argument we have:

This type says that is a function that takes a unit value and returns an , but the is free in its closure so is shared by all calls to it. The type system does not generalize regions which are free in these closures.