Re-computing Coverage Information to Assist Regression Testing Pavan Kumar Chittimalli

This paper presents a technique that leverages an exist-ing regression test-selection algorithm to compute accurate, updated coverage data on a version of the software, Pi+1, without rerunning any test cases that do not execute the changes from the previous version of the software, Pi, to Pi+1. Users of our technique can avoid the expense of re-running the entire test suite on Pi+1 or the inaccuracy pro-duced by previous approaches that estimate coverage data for Pi+1 or reuse outdated coverage data from Pi. This pa-per also presents a tool, RECOVER, that implements our technique, along with a set of empirical studies. The studies show the inaccuracies that can exist when an application— regression-test selection—uses estimated and outdated cov-erage data. The studies also show that the overhead in-curred by our technique is negligible.