afwu / git-dumper

A tool to dump a git repository from a website

git-dumper

A tool to dump a git repository from a website.

Install

This can be installed easily with pip:

pip install git-dumper

Usage

usage: git-dumper [options] URL DIR

Dump a git repository from a website.

positional arguments:
  URL                   url
  DIR                   output directory

optional arguments:
  -h, --help            show this help message and exit
  --proxy PROXY         use the specified proxy
  -j JOBS, --jobs JOBS  number of simultaneous requests
  -r RETRY, --retry RETRY
                        number of request attempts before giving up
  -t TIMEOUT, --timeout TIMEOUT
                        maximum time in seconds before giving up
  -u USER_AGENT, --user-agent USER_AGENT
                        user-agent to use for requests
  -H HEADER, --header HEADER
                        additional http headers, e.g `NAME=VALUE`

Example

git-dumper http://website.com/.git ~/website

Disclaimer

Use this software at your own risk!

You should know that if the repository you are downloading is controlled by an attacker, this could lead to remote code execution on your machine.

Build from source

Simply install the dependencies with pip:

pip install -r requirements.txt

Then, simply use:

./git_dumper.py http://website.com/.git ~/website

How does it work?

The tool will first check if directory listing is available. If it is, then it will just recursively download the .git directory (what you would do with wget).

If directory listing is not available, it will use several methods to find as many files as possible. Step by step, git-dumper will:

• Fetch all common files (.gitignore, .git/HEAD, .git/index, etc.);
• Find as many refs as possible (such as refs/heads/master, refs/remotes/origin/HEAD, etc.) by analyzing .git/HEAD, .git/logs/HEAD, .git/config, .git/packed-refs and so on;
• Find as many objects (sha1) as possible by analyzing .git/packed-refs, .git/index, .git/refs/* and .git/logs/*;
• Fetch all objects recursively, analyzing each commits to find their parents;
• Run git checkout . to recover the current working tree