Abstracting Event-Driven Systems with Lifestate Rules

TL;DR

This paper introduces lifestate rules as a unified approach to abstract and understand complex event-driven object protocols, specifically applied to Android framework types, aiding developers in managing intricate callback and API call sequences.

Contribution

It presents a novel model for event-driven systems and a specification mining method to derive lifestate rules from Android framework behavior.

Findings

Successfully derived rules characterizing Android framework behavior

Lifestate rules unify lifecycle and typestate constraints

Enhanced understanding of event-driven object protocols

Abstract

We present lifestate rules--an approach for abstracting event-driven object protocols. Developing applications against event-driven software frameworks is notoriously difficult. One reason why is that to create functioning applications, developers must know about and understand the complex protocols that abstract the internal behavior of the framework. Such protocols intertwine the proper registering of callbacks to receive control from the framework with appropriate application programming interface (API) calls to delegate back to it. Lifestate rules unify lifecycle and typestate constraints in one common specification language. Our primary contribution is a model of event-driven systems from which lifestate rules can be derived. We then apply specification mining techniques to learn lifestate specifications for Android framework types. In the end, our implementation is able to find several rules that characterize actual behavior of the Android framework.