Abstract

We systematically explore the design space of constraint systems for interprocedural analysis both of imperative and logic languages. Our framework is based on a small-step operational semantics where both the concrete and the abstract operational semantics are formalized by means of (input-free) pushdown automata. We consider the analysis problem of derivability and present constraint systems for a corresponding relational analysis with forward as well as backward (intraprocedural) accumulation. We abstract these constraint systems to obtain constraint systems for corresponding functional analyses. We clarify the relative precision of forward versus backward accumulation in presence of more or less complex intraprocedural control-flow graphs.

Citations