Auflistung nach Schlagwort "Conditional Model Checking"
1 - 2 von 2
Treffer pro Seite
Sortieroptionen
- KonferenzbeitragCombining Verifiers in Conditional Model Checking via Reducers(Software Engineering and Software Management 2019, 2019) Beyer, Dirk; Jakobs, Marie-Christine; Lemberger, Thomas; Wehrheim, HeikeSoftware verification received lots of attention in the past two decades. Nonetheless, it remains an extremely difficult problem. Some verification tasks cannot be solved automatically by any of today’s verifiers. To still verify such tasks, one can combine the strengths of different verifiers. A promising approach to create combinations is conditional model checking (CMC). In CMC, the first verifier outputs a condition that describes the parts of the program state space that it successfully verified, and the next verifier uses that condition to steer its exploration towards the unverified state space. Despite the benefits of CMC, only few verifiers can handle conditions. To overcome this problem, we propose an automatic plug-and-play extension for verifiers. Instead of modifying verifiers, we suggest to add a preprocessor: the reducer. The reducer takes the condition and the original program and computes a residual program that encodes the unverified state space in program code. We developed one such reducer and use it to integrate existing verifiers and test-case generators into the CMC process. Our experiments show that we can solve many additional verification tasks with this reducer-based construction.
- KonferenzbeitragCooperative Test-Case Generation with Verifiers(Software Engineering 2020, 2020) Beyer, Dirk; Jakobs, Marie-ChristineSoftware testing is widely applied in software quality assurance. Often, test suites fulfilling a certain coverage measure must be constructed. Manually constructing them is laborious. However, numerous automatic test-generation approaches exist. Due to various strengths and weaknesses of individual approaches, hybrid approaches, which combine different approaches, construct test suites that achieve higher coverage values than test suites generated by individual approaches. We propose the hybrid test-generation approach CoVeriTest. CoVeriTest is flexible, cooperative, and based on verification technology. It iteratively executes a sequence of verifiers that may exchange analysis information between each other and output a test case whenever they reach a test goal. The verifiers, their individual time limits, and which analysis information is exchanged between them is configurable. We experimented with different CoVeriTest configurations. The best configuration participated in the 1st International Competition on Software Testing (Test-Comp’19) and won the third place. This proves the value of our CoVeriTest approach.