Reasoning and Improving on Software Resilience against Unanticipated Exceptions

In software, there are the errors anticipated at specification and design time, those encountered at development and testing time, and those that happen in production mode yet never anticipated. In this paper, we aim at reasoning on the ability of software to correctly handle unanticipated exceptions. We propose an algorithm, called short-circuit testing, which injects exceptions during test suite execution so as to simulate unanticipated errors. This algorithm collects data that is used as input for verifying two formal exception contracts that capture two resilience properties. Our evaluation on 9 test suites, with 78% line coverage in average, analyzes 241 executed catch blocks, shows that 101 of them expose resilience properties and that 84 can be transformed to be more resilient.