PROJECT TITLE :

Global-Aware Recommendations for Repairing Violations in Exception Handling - 2017

ABSTRACT:

Empirical evidence suggests exception handling is not reliably implemented. Most faults in exception handling are related to global exceptions violating the intended exception handling style. However, repairing these violations is a cumbersome and error-prone task. It needs knowing the supposed design and understanding how the supply code violates it. It also requires changing the supply code to form it compliant with the supposed style. However changing the exception handling code may be a tough task, since changes in exception handling requires changing different elements of a program. Currently, there's still no answer to assist the repair of this kind of violations. To bridge this gap, we tend to present RAVEN, a heuristic strategy alert to the worldwide context of exceptions that produces recommendations of how violations in exception handling could be repaired. This strategy takes advantage of express specifications of the supposed style, though their availability isn't obligatory. Our results revealed RAVEN provides recommendations in a position to repair violations in 69p.c of the cases when policy specifications are not on the market and in 97p.c of the cases when specifications are accessible. Thus, development teams might benefit from RAVEN, even when exception handling style choices don't seem to be documented in their comes.