No abstract

Syntax Trees (ASTs) are widely used beyond compilers in many tools that measure and improve code quality, such as code analysis, bug detection, mining code metrics, refactoring. With the advent of fast software evolution and multistage releases, the temporal analysis of an AST history is becoming useful to understand and maintain code.However, jointly analyzing thousands versions of ASTs independently faces scalability issues, mostly combinatorial, both in terms of memory and CPU usage. In this paper, we propose a novel type of AST, called HyperAST , that enables efficient temporal code analysis on a given software history by: 1/ leveraging code redundancy through space (between code elements) and time (between versions); 2/ reusing intermediate computation results. We show how the HyperAST can be built incrementally on a set of commits to capture all multiple ASTs at once in an optimized way. We evaluated the HyperAST on a curated list of large software projects. Compared to Spoon, a state-of-the-art technique, we observed that the HyperAST outperforms it with an order-of-magnitude difference from ×6 up to ×8076 in CPU construction time and from ×12 up to ×1159 in memory footprint. While the HyperAST requires up to 2 h 22 min and 7.2 GB for the biggest project, Spoon requires up to 93 h and 31 min and 2.2 TB. The gains in construction time varied from 83.4 % to 99.99 % and the gains in memory footprint varied from 91.8 % to 99.9 %. We further compared the task of finding references of declarations with the HyperAST and Spoon. We observed on average 90 % precision and 97 % recall without a significant difference in search time.

Version Control Systems are key elements of modern software development. They provide the history of software systems, serialized as lists of commits. Practitioners may rely on this history to understand and study the evolutions of software systems, including the co-evolution amongst strongly coupled development artifacts such as production code and their tests. However, a precise identification of code and test coevolutions requires practitioners to manually untangle spaghetti of evolutions. In this paper, we propose an automated approach for detecting co-evolutions between code and test, independently of the commit history. The approach creates a sound knowledge base of code and test co-evolutions that practitioners can use for various purposes in their projects. We conducted an empirical study on a curated set of 45 open-source systems having Git histories. Our approach exhibits a precision of 100 % and an underestimated recall of 37.5 % in detecting the code and test coevolutions. Our approach also spotted different kinds of code and test co-evolutions, including some of those researchers manually identified in previous work.