Git Add

git add does one thing: it copies the current content of a file into the staging area, the list of changes your next commit will record. Nothing is committed yet, and running it again after further edits picks up the newer version rather than adding a second copy.

That is the part worth internalising. git add snapshots a file, it does not watch it. If you edit a file after staging it, the staged version is now out of date and you need to git add it again.

Add file to staging

for single file

git add <file>

for all files

git add .

or

git add -A

for all files in a directory

git add <directory>

Staging part of a file

By default git add takes the whole file. If you have changed several things in one file and only want some of them in this commit, stage it interactively and choose hunk by hunk.

git add -p src/app.ts

Git walks through each change and asks whether to stage it, so you end up with one commit per logical change rather than one commit containing everything you happened to be working on.

Unstaging

To take a change back out of the staging area while keeping it in your working directory:

git restore --staged src/app.ts

What not to add

Build output, dependency directories and credentials do not belong in a commit. Tell Git to leave them alone with a .gitignore rather than remembering to skip them — see .gitignore. A secret that reaches a commit stays in the history even after you delete it in a later one. Adds a file or directory to the staging area. This command prepares the changes for the next commit. It adds the specified file or directory to the index.