A Foundation of Escape Analysis

Escape analysis of object-oriented languages determines, for every program point, the creation points of the objects reachable from the variables. This information is useful to stack allocate dynamically created objects and to reduce the overhead of synchronisation in Java-like languages. We formalise the escape property E as a property of concrete states. We define the optimal abstract operations induced by E for a framework of analysis known as watchpoint semantics. The implementation of E inside that framework is a formally correct abstract semantics (analyser) for escape analysis. We claim that E is the basis for more refined and precise domains for escape analysis