If you run a multi-stage workflow with status checks in branch protection rules, you might found that skipped jobs are treated as "passed". That is, if you have a workflow with a chain of dependent jobs A -> B -> C. When B fails and C is skipped, it still counts as passed when checking job C.1 Thus, auto-merging may merge broken code.
Job matrix creates multiple job runs that are based on the combinations of the variables. Sometimes we want a dynamic number of matrix jobs, which requires a JSON array as an output. Here we use json and glob modules in Python to generate that JSON list.12
The Python package mattduck/gh2md exports Github repository issues and pull requests to a single, readable markdown file.
Sources:
HEAD: the current state of the repo.
Clone a git repo from a remote repository:
git clone <url>
Checking out a specific branch:
git clone <url> -b <branchname>
If there are submodule(s) in the Git repository, you might want to clone them as well using the --recursive option.
git clone <url> --recursive
See also: SSH login to Git services like GitHub and GitLab.
git status # The current state of the repository.
git add <file> # Add a new or edited file to the staging area. i.e. telling git to track this file
git add -A # Track all files at once
git commit -m "Commit message" # Commit staged (added) file
git commit -am "Commit message" # Commit modified files without having to run git add beforehand
git revert <SHA> # Make a counter commit to undo the changes. The tracked files will go back to the <SHA> commit.
git fetch # Download objects and refs from another repository without really pull in the changes
git merge # After git fetch, merge the changes done in the remote to the local repo
git push <remote> <branch-name> # Push commits in to remote
git push --set-upstream <remote> <name-of-your-branch> # Setup remote url before push
git pull <remote> # Pull changes from the remote
To temporarily store untracked files.
git stash -u # Store current work with untracked files
git stash pop # Bring stashed work back to the working directory
git branch <branch_name> # Create a new branch
git branch -a # List all branches
git branch -d <branch_name> # Delete a branch
git checkout <branch_name> # checkout an existing branch
git checkout -b <branch_name> # Create a new branch and checkout it
git switch <branch_name> # Switch to a specified branch. If the branch name does not exist, create one.
git merge <branch_name> # Merge the branch into the current branch
Orphan branches are unrelated to others in history. For example, gh-pages branch dedicated to GitHub pages.
git branch --orphan <branchname> # Create a orphan branch
Frequently used commands for Git submodules.
TO add the reference to another git project as a submodule:
git submodule add $url $path
git submodule update --init --recursive
Alternatively, you can use GUI tools like or GitHub desktop. They download and initiate submodules automatically.
Add you will see the file .gitmodules with information about the submodule(s). For instance,
[submodule "themes/DoIt"]
path = themes/DoIt
url = https://github.com/HEIGE-PCloud/DoIt.git
With -b $branch option
git submodule add -b $branch $url $path
Or set-branch -b $branch if you already have added a submodule
git submodule set-branch -b $branch $path
From a stackOverflow post and Git docs
git submodule update --remote --merge
For automated updates by bots, see automatic dependency update.
From Git docs
# Remove submodule from config
git submodule deinit $path
# Delete submodule tracking data
git rm <submodule path> && git commit
# Complete removal
rm -rf $GIT_DIR/modules/$name/
ssh-keygen -t ed25519 -C "your_email@example.com"
The SSH agent will ask you to enter a location to save the keys. e.g. /home/user/.ssh/id_ed25519. Passphrase is optional.
Then there will be two SSH key files:
~/.ssh/id_ed25519 is the private key. Protect it at all costs.~/.ssh/id_ed25519.pub is the public key.Using different keys for GitHub and GitLab access is more secure. However, the same pair of keys is used for this demonstration purposes.
Edit ~/.ssh/config
mkdir -p ~/.ssh
touch ~/.ssh/config
chmod 600 ~/.ssh/config
nano ~/.ssh/config
Add the following content to set the private key as the IdentityFile.
Host GitHub
HostName github.com
IdentityFile ~/.ssh/id_ed25519
Host GitLab
HostName gitlab.com
IdentityFile ~/.ssh/id_ed25519
According to 📖 Github docs, add the SSH key here
~/.ssh/id_ed25519.pub to the key field.Add SSH key green button. If prompted, confirm your GitHub password.Add the SSH key here
~/.ssh/id_ed25519.pub to the key field.Add key button.GitHub:
ssh -vT git@github.com
Accept its fingerprint if prompted. If you see "Hi user! You've successfully authenticated, but GitHub does not provide shell access" that means login is successful.
GitLab:
ssh -vT git@gitlab.com
Accept its fingerprint if prompted.
Jupyter notebooks without multimedia outputs are more friendly to source control since git is not good at comparing binary data (e.g., plots, pictures, videos) in jupyter notebooks. And they tend to bloat the size of git repositories.
Use pathlib to read and tomllib to parse the file. (requires Python >= 3.11)
Git filter-repo is a filter-branch replacement for rewriting history written in a single-file python script.
To wipe large binary files entirely:
git filter-repo --strip-blobs-bigger-than 100M
Bonus: Remove sensitive content
git filter-repo --use-base-name --path id_dsa --path id_rsa --invert-paths
git filter-repo --replace-text passwords.txt
commit are discarded.commit)git reset --hard <SHA> # Reset git history to a specific commit
git reset HEAD~ # Reset state to the previous commit (~)