Building is an integral part of the software development process. However, little is known about the compiler errors that occur in this process. In this paper, we present an empirical study of 26.6 million builds produced during a period of nine months by thousands of developers. We describe the workflow through which those builds are generated, and we analyze failure frequency, compiler error types, and resolution efforts to fix those compiler errors. The results provide insights on how a large organization build process works, and pinpoints errors for which further developer support would be most effective.

One of the biggest challenges in concolic testing, an automatic test generation technique, is its huge search space. Concolic testing generates next inputs by selecting branches from previous execution paths. However, a large number of candidate branches makes a simple exhaustive search infeasible, which often leads to poor test coverage. Several search strategies have been proposed to explore high-priority branches only. Each strategy applies different criteria to the branch selection process but most do not consider context, how we got to the branch, in the selection process.In this paper, we introduce a context-guided search (CGS) strategy. CGS looks at preceding branches in execution paths and selects a branch in a new context for the next input. We evaluate CGS with two publicly available concolic testing tools, CREST and CarFast, on six C subjects and six Java subjects. The experimental results show that CGS achieves the highest coverage of all twelve subjects and reaches a target coverage with a much smaller number of iterations on most subjects than other strategies.

No abstract