Test Co-evolution in Software Projects: A Large-Scale Empirical Study

This replication package contains the data necessary to reproduce and extend the results presented in the paper, including the identification of five test co-evolution patterns and the analysis of their impact on software quality. The package includes the curated dataset of software repositories and the raw results of the statistical tests, enabling researchers to validate our findings, explore alternative analyses, and apply the methodology to new contexts.

Notes

• Demo of the test detection approach: https://colab.research.google.com/drive/1A9LFJM0sgD9UCcHIYRcomLUJiDgeyzZG?usp=sharing

• Repository: https://github.com/CharlesMiranda/test-code-classification

In the "study" folder, you will find:

• projectinfo.csv: Raw data extracted from the GitHub API before processing. Contains information about the selected repositories.

• regular_expressions.rar: Data required for test file classification. Contains a list of regular expressions used to detect the presence of imports and function calls from popular test frameworks. We obtained these files from the work of Gonzalez et al. (2017).

• *.json: Processed data using Python scripts. Contains historical data on the total lines of production and test code.