Must every git commit be an ancestor of a branch?

Given this git tree:

A <--- B <--- C (HEAD, master)

is it possible to put the tree into this state:

  • Can the most recent common ancestor in git be ambiguous?
  • With Git, how can I show the tree of commits that all share a common ancestor?
  • A <--- B (HEAD, master) <--- C

    (There are no branches or tags other than master.) In other words, can a commit (C) exist without either being a commit pointed to by a branch (or tag) or the ancestor of a commit pointed to by a branch (or tag)?

    (This question is purely theoretical to help me understand better how git branches works.)

  • Putting uncommitted changes at Master to a new branch by Git
  • Creating branch inside a branch in git
  • Differences Between Fixup and Removing a Commit in Git Rebase
  • Git checkout 'virtual' branch
  • How can I retrieve the local changed files which I wrongly reset in git
  • Git branch vs commit
  • 2 Solutions collect form web for “Must every git commit be an ancestor of a branch?”


    git checkout -b tmp
    touch whatever.txt
    git add -A
    git commit -m "committed!"
    git checkout master
    git branch -D tmp

    The commit whose predecessor is the current HEAD of master is now still available, but no name is pointing to it. You can still get to it e.g. with the reflog commands.

    Another possibility, given that C already exists:

    git checkout -b tmp
    git checkout master
    git reset --hard HEAD-1
    git branch -D tmp

    ps, cannot find the tilde key on this kb…

    Commit C can exist temporarily (as a dangling commit); once a GC occurs it will be reclaimed.


    git fsck --unreachable

    to see dangling commits and blobs. Note that because git has a notion of reflog what you’ve described won’t create a dangling commit right away, C will stay in the reflog for a while. git maintains the reflog just in case you’ve made a mistake and need to recover something.

    Git Baby is a git and github fan, let's start git clone.