TY - GEN
T1 - ArbitCheck
T2 - 7th IEEE International Conference on Software Testing, Verification and Validation Workshops, ICSTW 2014
AU - Yatoh, Kohsuke
AU - Sakamoto, Kazunori
AU - Ishikawa, Fuyuki
AU - Honiden, Shinichi
PY - 2014/1/1
Y1 - 2014/1/1
N2 - Lightweight property-based testing tools are becoming popular these days. With property-based testing, developers can test properties of the system under test against large varieties of randomly generated inputs without writing test cases. Despite the advantages of property-based testing, current property-based testing tools have a major drawback: they require developers to write generator functions for user-defined types. This is because it is difficult for a tool to infer the possible values for the type. However, user-defined generators sometimes fail to find faults by only producing overly limited varieties of values. In this paper, we present a new property-based testing tool, called ArbitCheck, which automates object generation by adapting the feedback-directed random test generation technique. With the help of feedback-directed random test generation, ArbitCheck exhaustively generates possible values of user-defined types and tests properties with them, so that it can reveal faults that are hard to find with either manually written tests or existing property-based testing tools.
AB - Lightweight property-based testing tools are becoming popular these days. With property-based testing, developers can test properties of the system under test against large varieties of randomly generated inputs without writing test cases. Despite the advantages of property-based testing, current property-based testing tools have a major drawback: they require developers to write generator functions for user-defined types. This is because it is difficult for a tool to infer the possible values for the type. However, user-defined generators sometimes fail to find faults by only producing overly limited varieties of values. In this paper, we present a new property-based testing tool, called ArbitCheck, which automates object generation by adapting the feedback-directed random test generation technique. With the help of feedback-directed random test generation, ArbitCheck exhaustively generates possible values of user-defined types and tests properties with them, so that it can reveal faults that are hard to find with either manually written tests or existing property-based testing tools.
KW - Feedback-directed random test generation
KW - Java
KW - Object-oriented
KW - Property-based testing
KW - QuickCheck
KW - Random testing
KW - Randoop
KW - Unit testing
UR - http://www.scopus.com/inward/record.url?scp=84903624779&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84903624779&partnerID=8YFLogxK
U2 - 10.1109/ICSTW.2014.68
DO - 10.1109/ICSTW.2014.68
M3 - Conference contribution
AN - SCOPUS:84903624779
SN - 9780769551944
T3 - Proceedings - IEEE 7th International Conference on Software Testing, Verification and Validation Workshops, ICSTW 2014
SP - 405
EP - 412
BT - Proceedings - IEEE 7th International Conference on Software Testing, Verification and Validation Workshops, ICSTW 2014
PB - IEEE Computer Society
Y2 - 31 March 2014 through 4 April 2014
ER -