Git

https://git-scm.com/

Notes

  • When you run the git init command, Git creates a .git directory in the directory from where the command was launched. Everything Git needs to work is in there.

  • When you run the git add command, Git writes files to the staging area of the repository. When you run git commit, Git creates a commit object (with an ID), and it applies references (= labels) to that commit (HEAD and the branch you are on).

  • When you run the git branch command, Git just created a new reference with the name of the newly created branch, and applies it to the HEAD commit.

  • When you run the git checkout command, Git switches the branch it's going to advance when you commit.

  • An ID in Git is a checksum of the content of the commit, the author, the date, the log message, and the previous commit ID. Every ID is unique, so every commit is unique (as changing even slightly just one of the variables above will generate a completely different checksum).

  • When you run the git merge <branch> command, two situations are possible:

    • Git recognises the current branch is an ancestor of the branch to be merged: in this case, Git will perform the merge by fast-forwarding, i.e. it will just update references;

    • Git recognises the current branch is not an ancestor of the branch to be merged: in this case, Git will create a new commit (with two parents), and will move the master and the HEAD references (but not the branch one).

  • Thanks to the fact that every ID is unique, every commit is unique, commits do not change, and IDs depend on previous commits, Git is able to retrieve the full history of a commit just given its ID. This is very useful when you run the git push command, as it will send to origin only the minimum set of commits to make the two repositories the same.

  • When you run git tag, Git will create a new reference and attach it to a commit. The main difference of a tag is that it never moves after being created.

  • The reflog is a summary of the latest commits Git has interacted with. It can be useful in rescue situations when some of the commits have been done from a detached HEAD situation, meaning there was no branch associated with them.

Useful commands

Add a new remote repository

git remote add <remote_name> <remote_url>

Create a fixup commit

git commit --fixup <hash>

Rename current branch

git branch -m <new_name>

Resources

Alternatives

Articles

Books

Images

Tools

Videos

Websites

PreviousUnreal EngineNextGoals

Last updated