git 分支和 git checkout -b 有什么区别？

Question

I'm very new to git/gitlab and want to confirm the difference between the 2 commands in the first screen shot. What is the purpose of having 2 commands which both create new branches. Why is 'git branch' even needed?

Is it purely there in case i want to create a new branch but not switch to that branch? So i very minor time save instead of doing 'git checkout -b new_branch' and then 'git checkout previous_branch'.

Also why does it say to reference the current HEAD? I assume this is related to the 'remotes/origin/HEAD -> origin/master' in the second screen shot?

I tried 'git branch test_branch_name' which created the new branch without needed to reference the current HEAD, so why does it mention referencing the current HEAD?

Probably not necessary, but here's a link for where the first screen shot comes from: https://about.gitlab.com/images/press/git-cheat-sheet.pdf

Answer 1

To fill in a few more details:

... Why is 'git branch' even needed?

The git branch command has many options, such as:

git branch --set-upstream-to

or:

git branch --list

or:

git branch -d

which do things that neither git checkout nor the new (Git 2.23 or later) git switch will do.

The things that git branch --list does can be done with git for-each-ref . The thing that git branch --set-upstream-to does can be done with git config . If you'd like an analogy, Git commands are tools . You can cut wood with a bandsaw, a hacksaw, a coping saw, a drill, or a chisel. If you'd like to make dovetails, most of these tools could do the job—but it will go a lot faster and better and easier with a router and a dovetailing jig. If your goal is to put in a lock and doorknob set, consider a hole saw and the drill.

(The git checkout command actually has too many tools built into it—in my opinion anyway. Apparently, finally, the official Git maintenance folks are coming around to this same opinion, as two of the main jobs that git checkout can perform are now provided by two separate front-end commands: git switch and git restore . Of course, in typical Git fashion, git restore can now do a few things that git checkout can't.)

Git is in some ways mainly a graph-manipulating suite of tools that happen to drag source code along with them. The graph you manipulate is the commit graph , which is made up of nodes (graph vertices) that are commits, connected by one-way edge links (arcs) that are the parent hash IDs stored in each commit-node. See Think Like (a) Git .

Answer 2

git checkout -b [name] is short for git branch [name] && git checkout [name] . The recent versions of Git also support git switch to checkout a branch.

git branch [name] is equivalent to git branch [name] HEAD . git branch [name] is a special case of git branch [name] [start-point] . When start-point is omitted, it defaults to HEAD .

For more details, see git branch .