What do git checkouts really mean?

As you noted, HEAD is a label noting where you are in the commit tree. It moves with you when you move from one commit to another. git checkout <commit> is the basic mechanism for moving around in the commit tree, moving your focus (HEAD) to the specified commit.

The commit can be specified by any of a number of ways, commit hash, branch name, tag name, the relative syntax (HEAD^, HEAD~1, etc.) and so on. It is often useful to consider a checkout to be changing branches, and there are some options that work from that perspective, but they all reference commits.

To checkout a commit has some side affects other than moving HEAD around.

• The working directory is updated to the state of the checked out commit.
• if a branch name is specified, checkout makes that branch active. The active branch will move along with any new commits that are added.
  • with the -b option a new branch will be created based on the current commit and then made active.
  • with the --track option the checked out branch can be made aware of a remote branch
  • with the --orphan option a new branch is created (like with -b) but will not be based on any existing commit.

There are a few more options, which you can read about in the git checkout man-page, all of which revolve around moving from one commit to another -- just varying in what effect that move has in addition to moving HEAD.

"To check out" means that you take any given commit from the repository and re-create the state of the associated file and directory tree in the working directory.

When you check out a commit that is not a branch head (e.g. git checkout HEAD~2), you are on a so-called detached head. You can create commits here, but once you switch to a different branch, those commits will not be recoverable by a branch name and might even get removed by the garbage collector after some time.

Let me explain some use cases of checkout with file, folder and branches so that it might be helpful in understanding.

Let say we have folder named dev and index.html also Everything is tracked and working directory is clean.

If I accidentally change file name index.html and I want to undo that i will simply use git checkout index.html it will recover that file state from repository currently selected branch.

Now if I made some change in dev folder and want to recover that. I can use git checkout dev but what if there is already branch named dev instead of checking out that folder it will pull down that branch. To avoid that I would rather do git checkout -- dev.

Now here bare double dash stands for current branch and asking git for folder dev from currently selected branch.

Similarly If I do git checkout alpha dev it will pull down dev folder from alpha branch.

This answer is for your first question 'git checkout really mean'.