Understanding Exceptional Behavior and Exception Flow in Implicit Invocation Systems by Mehdi Bagherzadeh, Hridesh Rajan and Ali Darvish

Exceptions and their implicit control flow complicate modular understanding (reasoning) of both exceptional behavior and exception flow of a system. Such reasoning tasks about exceptions are more complicated in implicit invocation (II) systems, where a subject module can invoke other observer modules without knowing about them and especially behavior and flow of their exceptions. In this work, we motivate and discuss design of an II language model called Ptolemyχ which enables modular reasoning about exceptional behavior and exception flow of event announcement and handling. To enable modular reasoning, first Ptolemyχ’s exception aware specification expressions and boundary exceptions limit the set of exceptions subjects and observers may throw, both checked and unchecked. Second, behavior of such exceptions is specified using exceptional postconditions, adapted to II settings. And third, greybox-based specifications along with exception aware specification expressions expose information about places, in the observers’ implementations, where exceptions are thrown or handled. This work illustrates Ptolemyχ’s reasoning techniques about exceptional behavior and exception flow and evaluates utility of the exception flow reasoning by using it to understand a set of bug patterns for aspect-oriented (AO) programs. It also presents Ptolemyχ’s operational semantics and sound static semantics.