zhcxww / icse2017 Goto Github PK

This repository contains the materials of our paper "How do Developers Fix Cross-project Correlated Bugs? 
A case study on the GitHub scientific Python ecosystem", including:

1. script
* the Python script used to collect bugs in GitHub repository.

2. method & data
* Manual Analysis: the detailed description of our manual inspection
* Survey Analysis: the detailed description of the online survey and a summary of the survey responses
* cross-project bugs: the 271 pairs of investigated cross-project correlated bugs
* Statistical Comparision: the results of statistical tests for comparing the difficulty of fixing cross-project bugs and other bugs in terms of fixing time, number of participants, and number of comments
* numpy, scipy, astropy, ipython, scikit-learn, matplotlib, pandas: the bug data of numpy, scipy, astropy, ipython, scikit-learn, matplotlib, pandas collected by the script "getBugs.py"

3. figures
* Dependency graph (Figure 3): Figure 3 in the paper which shows part of the inter-dependencies among projects in the scientific Python ecosystem
* Distribution of hidden duration (Figure 4): Figure 4 in RQ1 which shows the distribution of hidden duration
A version-dependent workaround (Figure 6): Figure 6 in RQ3 which shows an example of a version-dependent workaround
* The identities of reporters and restorers (Figure 7): Figure 7 in RQ3 which shows the identities of reporters and restorers of the upstream bugs in the dowsntream projects