Accelerating Search-Based Program Repair

PRF: a framework for building automatic program repair prototypes for JVM-based languages

“…Our measurements indicate that patch validation takes up 77.7% of total repair time, on average. Similar observations are also reported in previous research by Mehne et al [15]. As we shall discuss in §4, there are a number of methods to reduce patch validation time, but in this work we employ a novel approach: (1) each patch is validated in a separate process; (2) only the tests covering patched location are executed;…”

“…Le Goues et al [11] highlight the high cost of patch validation in test based G&V APR. Mehne et al [15] report that patch validation can take between 40% to 92% of total repair time and propose to prune the patches needed to be tested as well as test case selection to reduce this cost. A recent line of research [6,9], proposes to use the HotSwap trick offered by the JVM to validate the patches on-the-fly, without restarting the JVM.…”

Enhancing Genetic Improvement of Software with Regression Test Selection

“…During GI optimisation, software undergoes transformations in order to improve a set of properties, either functional or non-functional [1]. The most common type of functional improvement is APR [13], [14], [25], in which a faulty program is modified until the failing test suite passes. In non-functional GI however, the goal is to improve the software's memory usage [8], [9], execution time [5]- [7], energy consumption [10]- [12], and other non-functional properties, whilst maintaining the functional properties of the software, measured with the use of the program's test suite, i.e., test cases should pass after the program transformation.…”

“…This process may go on for hundreds or thousands of iterations, each of which imposing the cost of executing the test suite against the candidate solution. The efficiency of GI becomes a problem when the software is accompanied by costly test suites, which has been pointed out in APR work [13], [14].…”

“…As far as we are aware, there is only one related work in the literature. Mehne et al [14] used an ad-hoc RTS (called "Test-Case Pruning") and a localisation technique with GenProg [2], an APR tool for C programs. Their technique only applies patches in specific places, selects test cases that cover the patched functions, and then executes only these test cases when the patch needs to be validated.…”