A version control system, such as Git, requires a way to integrate changes from different developers or branches. Given a merge sce- nario, a merge tool either outputs a clean integration of the changes or outputs a conflict for manual resolution. A clean integration is correct if it preserves intended program behavior, and is incorrect otherwise (e.g., if it causes a build failure or test failure). Manual resolution consumes valuable developer time, and correcting a defect introduced by an incorrect merge is even more costly.

New merge tools have been proposed, but they have not yet been evaluated against one another. This paper does so, providing essential context for researchers and practitioners. Our novel experimental methodology combines running test suites, examining merges on deleted branches, and accounting for the cost of incorrect merges.

Based on these evaluations, we created a family of merge tools that outperform all previous tools under most assumptions. They correctly handle the most common merge conflicts produced by Git.