While there is a concept of "default branch name" in git, I do not see any mention of "default branch" in git documentation or Pro Git book.

However, I do see mentions of "default branch" in git hosting platforms such as GitHub and GitLab.

So, is there a concept of "default branch" in vanilla git? Or is it only a concept used by git hosting platforms?

CodePudding user response：

In git's model, there are no special "server" or "central" repositories. When you clone a repository, then git will by default check out whatever branch is checked out on that other repository. More specifically, it will copy the .git/HEAD file from the other repository into the new clone.

At the most basic, a git server (including hosting platforms such as Github and GitLab) is just what git calls a "bare repository", which contains all the files which would normally be in the .git directory, but no working copy. See the manual for git clone --bare and "Git on the Server" in the Pro Git book.

That means the repository still has a file called HEAD which contains a reference to the "current" branch. When you perform a clone from the server, that is the reference copied to your new .git/HEAD. As such, it acts as the "default branch" for new clones, but in git's terminology it is actually the "current" branch.

Management UIs commonly also use this as a convenient way to set the default branch shown in the file browser, the default target for pull/merge requests, etc.

CodePudding user response：

The closest thing seems to be the init.defaultBranch config entry.

It allows to set which name will be used at git init for the first commit/branch. (see also here)

The default value for this parameter is master.