How to tell which version is ours and which is theirs?

See the following example:

$ git branch
      develop
    * master
$ git merge develop
$ git log
commit 5aa63defd7d552544348deaad88a22d212c43038 (HEAD -> master)
Merge: 011eeae bdd631b
Author: Symbolk <symbolk@163.com>
Date:   Sat Jul 6 17:12:25 2019 +0800

    Merge branch 'develop'

The current branch master is called ours, on which the merge commit will be submitted. The branch develop in the git-merge command is called theirs, which will not be affected by the merging.

• Windows /Linux/macOS
• Python 3.7
• Git 2.18.0
• PyCharm

1. Open the clone repository as a project with PyCharm;
2. Under the root directory of the cloned repository, run the following command in the terminal:

pip install -r requirements.txt 

Collect Java files involved in merge scenarios that contain merge conflict(s) from the whole commit history.

A Git repository with the name of the target branch (usually master).

Edit the main.py to set the necessary variables, then run it:

if __name__ == "__main__":
    repo_name = "cassandra"
    # get the default branch
    branch_name = "trunk"
    # if the repo is not present in the repo_dir, it will be cloned, but better to clone in advance
    git_url = "https://github.com/apache/cassandra"
    repo_dir = os.path.join(home, "coding/data/repos", repo_name)
    result_dir = os.path.join(home, "coding/data/merges", repo_name)

    # Usage1: Collect Java files involved in merge scenarios that contain merge conflict(s) from the whole commit history
    statistic_path = result_dir + "/statistics.csv"
    git_service = GitService(repo_name, git_url, repo_dir, branch_name, result_dir)
    git_service.collect_from_repo(statistic_path)

During the mining process, the brief summary of each merge scenario with merge conflict(s) is printed in the Run console of PyCharm:

Cloning into 'D:\github\rep\javaparser'...
POST git-upload-pack (gzip 7425 to 3775 bytes)
remote: Enumerating objects: 47, done.        
remote: Counting objects: 100% (47/47), done.        
remote: Compressing objects: 100% (21/21), done.        
remote: Total 92708 (delta 10), reused 43 (delta 10), pack-reused 92661        
Receiving objects: 100% (92708/92708), 21.11 MiB | 830.00 KiB/s, done.
Resolving deltas: 100% (49120/49120), done.
Checking out files: 100% (2028/2028), done.
Ready to process repo: javaparser at branch: master
Commit: e6063bb10d6d41cb2b258540bb47edbd18b4646b, #Unmerged_blobs: 4, #Conflict java files: 2, #Conflict blocks: 2
Commit: 0258e273bfd2dca550a27d3204cf22227a41e772, #Unmerged_blobs: 5, #Conflict java files: 4, #Conflict blocks: 4
Commit: 25c4bbf796034c987e7517d4d3c596026a0142e3, #Unmerged_blobs: 16, #Conflict java files: 2, #Conflict blocks: 3
...

Collected data will be saved in the result_dir, which contains:

1. Sub-folders named with merge commit ids, each of them contains conflicting Java files in that merge scenario.

   

   

2. A csv file that provides a statistical summary of each merge scenario, which consist of 4 commit ids (merge commit, HEAD of ours/theirs branch, base commit), the number of conflicting Java files and their paths, and the number of conflict blocks.

   

   In the column #conflict blocks, the numbers denotes the number of conflict blocks in every conflicting Java file. For example, in the first row, there are 5 conflicting Java files, the first file javaparser-core-serialization/src/main/java/com/github/javaparser/serialization/JavaParserJsonSerializer.java has 1 conflict block inside it.

Collect Java files involved in merge scenarios that contain refactoring-related merge conflict(s) from the csv file generated by https://github.com/Symbolk/RefConfMiner.git (python scripts to analyze MySql data generated by https://github.com/ualberta-smr/RefactoringsInMergeCommits).

1. A Git repository with the name of the target branch (usually the main branch, like master).
2. The csv file that records refactoring-related merge commit ids, generated from the tool RefactoringsInMergeCommits (https://github.com/ualberta-smr/RefactoringsInMergeCommits).

Edit the main.py to set the necessary variables, then run it:

if __name__ == "__main__":
    repo_name = "cassandra"
    # get the default branch
    branch_name = "trunk"
    # if the repo is not present in the repo_dir, it will be cloned, but better to clone in advance
    git_url = "https://github.com/apache/cassandra"
    repo_dir = os.path.join(home, "coding/data/repos", repo_name)
    result_dir = os.path.join(home, "coding/data/merges", repo_name)

    # Usage2: Collect Java files involved in merge scenarios that contain refactoring-related merge conflict(s) 
    # from the csv file generated by https://github.com/Symbolk/RefConfMiner.git
    result_dir = os.path.join(home, "coding/data/ref_conflicts", repo_name)
    # csv_file = "merge_scenarios_involved_refactorings_test.csv"
    csv_file = os.path.join(home, "coding/data/merge_scenarios_involved_refactorings", repo_name + ".csv")
    git_service = GitService(repo_name, repo_dir, branch_name, result_dir)
    git_service.collect_from_csv(csv_file)

The output is basically same with that in Usage 1.