Ab=ᗺa

Abstract

This paper introduces a new type of automated testing oracle, called the execution equivalence (EE) invariants. These invariants can be mined from application logs that capture both application events and application states. The EEinvariants express an equivalence relation on the sequences of application events in terms of equality of respective initial and final states, which these sequences leave in the logs during the run-time. We claim that even equivalences up to a length of four events already provide useful testing oracle. We extended our tool LOPI (LOg-based Pattern Inferencer) with the algorithm for mining EE-invariants, and evaluated the effectiveness of these invariants on a case-study -the web application Flex Store. The evaluation is carried out based on two parameters: the false positive rate and the fault finding capability. Moreover, we compared the strength of LOPI's execution equivalences with Daikon's data invariants. This comparison has shown that Daikon was slightly more effective than LOPI in testing Flex Store. However, we have found a suitable confidence level for LOPI which allows to outperform Daikon.

Bridging the Gap Between GUI Functional System Testing and Functional System Logic Testing

“…In addition to Testar, Elyasov and Vos extend a tool called Log-based pattern interface (LOPI) to present a new type of testing oracle. The oracle uses the information in log files that software generate to obtain data which help make a decision whether two sequences of events are equivalent [90]. Esparcia-Alcázar integrated genetic algorithms to improve the events selection process and traversal procedure used in Testar [91].…”

Representation learning for algorithms implementation search using pseudo code

“…Importantly, getting effective representation learning for mappings between the algorithms and code can help identify semantic clones that implement the same algorithmic steps. Such semantic clones are known to be very valuable for detecting bugs [33,18], generating test oracles and performing differential testings [5,9], fixing and improving programs [31], designing APIs and optimizing code [8], and providing data and downstream tasks for evaluating deep learning-based source code modeling tools [41,34,37,36,14] are mainly dependent on getting better representation learning. The mapping from pseudo code to source code can also provide insights on how to use pseudo code to synthesize the programs [21].…”