Git maintenance

clean_repos.py

import os
import subprocess
import sys


def run_git_command(repo_path, command):
    """Run a git command in the given repository path."""
    try:
        subprocess.check_call(command, cwd=repo_path)
    except subprocess.CalledProcessError as e:
        print(f"Error while running '{' '.join(command)}' in {repo_path}: {e}")
        return False
    return True


def clean_repositories(repo_dir):
    """Iterate over all repositories and run git fsck and git gc."""
    if not os.path.isdir(repo_dir):
        print(f"Error: The directory {repo_dir} does not exist.")
        return

    # Iterate over each subdirectory in the given repo directory
    for repo_name in os.listdir(repo_dir):
        repo_path = os.path.join(repo_dir, repo_name)
        if os.path.isdir(repo_path) and os.path.isdir(os.path.join(repo_path, ".git")):
            print(f"\nProcessing repository: {repo_path}")

            # Run git fsck --full
            print("Running git fsck --full...")
            if not run_git_command(repo_path, ["git", "fsck", "--full"]):
                continue  # Skip to the next repo if fsck fails

            # Run git gc --aggressive --prune=now
            print("Running git gc --aggressive --prune=now...")
            if not run_git_command(
                repo_path, ["git", "gc", "--aggressive", "--prune=now"]
            ):
                continue  # Skip to the next repo if gc fails

            # Run git commit-graph write --reachable --changed-paths
            print("Running git commit-graph write --reachable --changed-paths...")
            if not run_git_command(
                repo_path,
                ["git", "commit-graph", "write", "--reachable", "--changed-paths"],
            ):
                continue  # Skip to the next repo if commit-graph fails

            print("Repository cleaned successfully.")
        else:
            print(f"Skipping non-Git directory: {repo_path}")


def main():
    """Main function."""
    if len(sys.argv) != 2:
        print("Usage: python clean_repos.py <path-to-repositories-folder>")
        sys.exit(1)

    repo_dir = sys.argv[1]
    clean_repositories(repo_dir)


if __name__ == "__main__":
    main()

Git Maintenance

Runs the following git commands for housekeeping over a folder containing git repositories.

Verify filesystem integrity.

git fsck --full

Garbage collect - compress and optimize the repository, remove unreachable objects.

git gc --aggressive --prune=now

Write a commit graph file, which speeds up operations like git log and git blame by improving commit traversal performance.

git commit-graph write --reachable --changed-paths