Gitalong is a tool for Git repositories that seek to prevent conflicts between files when working with a team. It uses hooks and a store to communicate local changes across all clones of a given remote. In turns this information can be leveraged by integrations to prevent modifying files that are already changed elsewhere.
Tip
Setting up Python and Git can be intimidating on Windows. You can make your life easier by installing Scoop and running scoop install python git in a Windows command prompt.
pip install gitalong
# Creating a dummy project repository and its clone in current working directory.
git init --bare project.git
git clone project.git project
# Creating a repository that Gitalong will use to store and share local changes.
# You would normally host this somewhere like GitHub so your entire team has access to it.
git init --bare store.git
# Setting up Gitalong in your project repository.
# This will clone the store repository in an ignored `.gitalong` folder.
# It will also start tracking a `.gitalong.json` configuration file.
gitalong -C project setup store.git --modify-permissions --tracked-extensions .jpg,.png --track-uncommitted --update-gitignore --update-hooks
# Creating some files.
touch project/untracked.txt
touch project/uncommitted.png
touch project/local.png
touch project/current.jpg
touch project/remote.jpg
# Spreading them across branches.
git -C project add current.jpg
git -C project commit -m "Add current.jpg"
git -C project add remote.jpg
git -C project commit -m "Add remote.jpg"
git -C project push
git -C project reset --hard HEAD^
git -C project add local.png
git -C project commit -m "Add local.png"
# Updating tracked commits with current local changes.
# Because we specified `track_uncommitted`. Uncommitted changes will be stored as sha-less commit.
# Update permissions of all files based on track commits.
# Because `modify_permssions` was passed this will update all permissions of tracked files.
# Permission updates currently comes at high performance cost and is not recommended.
gitalong -C project update
# Checking the status for the files we created.
# Each status will show <spread> <filename> <commit> <local-branches> <remote-branches> <host> <author>.
# Spread flags represent where the commit live.
# It will be displayed in the following order:
# <mine-uncommitted><mine-active-branch><mine-other-branch><remote-matching-branch><remote-other-branch><other-other-branch><other-matching-branch><other-uncomitted>
# A `+` sign means is true, while a `-` sign means false or unknown.
gitalong -C project untracked.txt status uncommited.png local.png current.jpg remote.jpgYou can find a usage example in example.py.
As mentioned earlier, Gitalong needs an accessible place to store and share local changes with all clones of the managed repository. Multiple options are offered here.
A Git repository can be used for this purpose. The advantage of this solution is that you won't need more infrastructure and security mechanisms than what is needed to access your project's repository. That said, pulling and pushing the data that way is pretty slow. This method is used in the usage examples above.
JSONBin.io is a simple JSON hosting service. You will get faster operations with this option but it may come at a cost depending on your usage. See how to set this up below.
gitalong -C project setup https://api.jsonbin.io/v3/b/<BIN_ID> --store-header X-Access-Key=<ACCESS_KEY> ...
repository = Repository.setup(
store_url="https://api.jsonbin.io/v3/b/<BIN_ID>",
store_headers={"X-Access-Key": "<ACCESS_KEY>"},
...Worth noting that <ACCESS_KEY> can be an environment variable such as $ACCESS_KEY.
Setup a Python virtual environment and run the following command.
python -m venv .venv
source .venv/bin/activate
pip install -e ".[ci]"pytest - -cov - report = html - -cov = gitalong - -profile - svgsphinx-build ./docs/source ./docs/buildpython setup.py sdist bdist_wheeltwine upload --username __token__ --verbose dist/*