This question already has an answer here:
I eagerly ducked into code mode and modified some files, but neglected to branch from master first. The mods aren't so extensive that I can't redo them, but what's a good way of taking my (so far, uncommitted) changes in master and migrating them to a new branch, leaving master untouched in the end?
If not yet committed anywhere (git status
shows a bunch of stuff modified, it's OK if it's "git add"-ed too):
$ git checkout -b newbranch
Despite the name checkout
this usage (with -b
) does not check anything out. The -b
flag says "create a new branch", so git creates the branch-name and makes it correspond to the current HEAD
commit. Then it makes HEAD
point to the new branch, and stops there.
Your next commit is therefore on newbranch
, which has as its parent commit, the commit you were on when you started modifying files. So assuming you were on master
, and you had these commits:
A - B - C <-- HEAD=master
the checkout -b
makes this read:
A - B - C <-- master, HEAD=newbranch
and a later commit adds a new commit D
:
A - B - C <-- master
\
D <-- newbranch
git branch -M master my-branch
and then
git fetch origin refs/heads/master:refs/heads/master
or
git branch master my-branch (or another ref)
git stash
git stash branch <branchname>
来源:https://stackoverflow.com/questions/19612439/forgot-to-branch-in-git-need-to-move-changes-from-master