The plastic surgery hypothesis

Abstract

Recent work on genetic-programming-based approaches to automatic program patching have relied on the insight that the content of new code can often be assembled out of fragments of code that already exist in the code base. This insight has been dubbed the plastic surgery hypothesis; successful, well-known automatic repair tools such as GenProg rest on this hypothesis, but it has never been validated. We formalize and validate the plastic surgery hypothesis and empirically measure the extent to which raw material for changes actually already exists in projects. In this paper, we mount a large-scale study of several large Java projects, and examine a history of 15,723 commits to determine the extent to which these commits are graftable, i.e., can be reconstituted from existing code, and find an encouraging degree of graftability, surprisingly independent of commit size and type of commit. For example, we find that changes are 43% graftable from the exact version of the software being changed. With a view to investigating the difficulty of finding these grafts, we study the abundance of such grafts in three possible sources: the immediately previous version, prior history, and other projects. We also examine the contiguity or chunking of these grafts, and the degree to which grafts can be found in the same file. Our results are quite promising and suggest an optimistic future for automatic program patching methods that search for raw material in already extant code in the project being patched.

Staged program repair with condition synthesis

“…The goal of the Replace part of Copy and Replace is to obtain a rich patch search space that includes variable replacement to enable copied code to operate successfully in a new naming context. The success of these modifications is consistent with previous work that shows that, without replacement, only approximately 10% of developer changes can be fully derived from existing statements without modification [23,7,16]. This fact may also provide insight into the poor performance of previous patch generation systems [21,33,28] (which only copy statements without expression-level modifications).…”

“…The rationale for the Copy and Replace modification is to exploit redundancy in the program -many successful program modifications can be constructed from code that already exists in the program [7]. The goal of the Replace part of Copy and Replace is to obtain a rich patch search space that includes variable replacement to enable copied code to operate successfully in a new naming context.…”

Mobile app and app store analysis, testing and optimisation

“…The original program is used as test oracle [3], while automated test data generation is used to assess faithfulness. Recent empirical results concerning code uniqueness [5] and the graftability of code modifications from existing code bases [1] have provided evidence that large existing systems contain a surprising amount of useful 'genetic material' for such improvement.…”

An analysis of patch plausibility and correctness for generate-and-validate patch generation systems

“…GenProg in particular is visible, widely cited, and used in subsequent research [25,9,33]. Example quotes that reflect how GenProg is widely viewed as a system that can successfully fix bugs in large C programs include: "We selected GenProg for the reason that it is almost the only state-ofthe-art automated repair tool having the ability of fixing real-world, large-scale C faulty programs" [33] and "successful, well-known automatic repair tools such as GenProg" [9].…”