The task of debugging software failures is generally time consuming and involves substantial manual effort. A crucial part of this task lies in the reproduction of the reported failure at the developer’s site. In this paper, we propose a novel framework that aims to address the problem of failure reproduction by employing an adaptive search-based approach in combination with a limited amount of instrumentation. In particular, we formulate the problem of reproducing failures as a search problem: reproducing a software failure can be viewed as the search for a set of inputs that lead its execution to the failing path. The search is guided by information obtained through instrumentation. Preliminary experiments on small-size programs show promising results in which the proposed approach outperforms random search.
A Search-Based Framework for Failure Reproduction
Kifetew, Fitsum Meshesha
2012-01-01
Abstract
The task of debugging software failures is generally time consuming and involves substantial manual effort. A crucial part of this task lies in the reproduction of the reported failure at the developer’s site. In this paper, we propose a novel framework that aims to address the problem of failure reproduction by employing an adaptive search-based approach in combination with a limited amount of instrumentation. In particular, we formulate the problem of reproducing failures as a search problem: reproducing a software failure can be viewed as the search for a set of inputs that lead its execution to the failing path. The search is guided by information obtained through instrumentation. Preliminary experiments on small-size programs show promising results in which the proposed approach outperforms random search.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
