Considering typestate verification for quantified event automata

Abstract

This paper discusses how the existing static analyses developed for typestate properties may be extended to a more expressive class of properties expressible by a specification formalism originally developed for runtime verification. The notion of typestate was introduced as a refinement of the notion of type and captures the allowed operations in certain contexts (states) as a subset of those operations allowed on the type. Typestates therefore represent per-object safety properties. There exist effective static analysis techniques for checking typestate properties and this has been an area of research since typestates were first introduced in 1986. It has already been observed that common properties monitored in runtime verification activities take the form of typestate properties. Additionally, the notion of typestate has been extended to reflect the more expressive properties seen in this area and additional static and dynamic analyses have been introduced. This paper considers a highly expressive specification language for runtime verification, quantified event automata, and discusses how these could be viewed as typestate properties and if/how the static analysis techniques could be updated accordingly. The details have not been worked out yet and are not presented, this is intended for later work.