作者: Milos Gligoric , Rupak Majumdar , Rohan Sharma , Lamyaa Eloussi , Darko Marinov
DOI: 10.1007/978-3-319-08867-9_19
关键词:
摘要: Regression test selection analyzes incremental changes to a codebase and chooses run only those tests whose behavior may be affected by the latest in code. By focusing on small subset of all tests, testing process runs faster can more tightly integrated into development process. Existing techniques for regression consider two versions code at time, effectively assuming where occur linear sequence. Modern processes that use distributed version-control systems are complex. Software version histories generally modeled as directed graphs; addition occurring linearly, multiple related other commands, e.g., branch, merge, rebase, cherry-pick, revert, etc. This paper describes test-selection technique software developed using modern systems. modeling different branch or merge commands directly our technique, it computes safe sets substantially smaller than applying previous linearization history. We evaluate several large open-source projects. The results encouraging: obtained an average 10.89× reduction number over existing while still selecting differ.