How do I create a commit without a parent in Git?

Question

I have a small project I was working on without any kind of version control software, and I ended up with three different versions of the same script, where each one does something slightly different from the others. I initialized a git repo with one of the versions, and I'd really like to commit each of the other two without parents. That way, I can merge the differences as though each of the three versions of the script were the tips of different branches.

I'm not even sure whether this is necessarily a sane or reasonable way to go about solving this problem... Any advice on how to do things this way, or advice on a better way to handle this situation would be much appreciated.

Why don't you create two branches and add the two other versions to them and then work on the branches? — manojlds, Apr 16 '11 at 22:16
See: [how to insert a commit as the first, shifting all the others?](http://stackoverflow.com/questions/645450/git-how-to-insert-a-commit-as-the-first-shifting-all-the-others) — Josh Lee, Apr 25 '11 at 23:00

score 172 · Accepted Answer · edited May 14 '12 at 08:21

From my comment:

Why don't you create two branches and add the two other versions to them and then work on the branches?

Which I think would agree with ctcherry's answer as well.

But if you really do want to create a completely different "branch" that is unrelated to master, then you can do it with git checkout --orphan. The manual for git checkout describes to work of the option --orphan:

Create a new orphan branch, named <new_branch>, started from <start_point> and switch to it. The first commit made on this new branch will have no parents and it will be the root of a new history totally disconnected from all the other branches and commits.

You might use it like this:

git checkout --orphan version2
git rm -rf .
<add version2 script>
git add your files
git commit -m 'Initial commit for version 2'

If your Git is older than 1.7.2 (when git checkout --orphan was introduced), then you can use the alternate command sequence from “Creating New Empty Branches” in the Git Community Book:

Ocasionally, you may want to keep branches in your repository that do not share an ancestor with your normal code. Some examples of this might be generated documentation or something along those lines. If you want to create a new branch head that does not use your current codebase as a parent, you can create an empty branch like this:
git symbolic-ref HEAD refs/heads/newbranch 
rm .git/index 
git clean -fdx 
<do work> 
git add your files 
git commit -m 'Initial commit'

+1 for mentioning `git checkout --orphan` - even if my git 1.7.1 does not support this yet. But it looks like [Ubuntu 11.04 will contain git 1.7.4](http://packages.ubuntu.com/natty/git), so I only have to wait for more two weeks. — Paŭlo Ebermann, Apr 16 '11 at 23:12
+1 for explaining why the OP might want to rethink their strategy, but then answering the question anyway, instead of saying "Your approach is wrong" and leaving it at that. — Tyler, Apr 17 '11 at 01:28
After performing `git checkout`, you can also use `git reset --hard` to remove the staged files. — Thai, Oct 24 '14 at 08:28

score 1 · Answer 2 · answered Apr 16 '11 at 22:16

1

You can create a branch for each version of the script, and commit them to those individual branches. Then merge them one at a time back into master. That should have the effect you are looking for.

answered Apr 16 '11 at 22:16

Chris Cherry

28,118
6
68
71

score 1 · Answer 3 · answered Apr 25 '11 at 21:50

Why not start with an 'Empty' repo (e.g. a single readme of your problem statement above). Then you can branch the three times and add the different script versions into each of the branches, and commit them separately.

You can then branch/edit/commit/merge and try out all you options in a well managed way.

How do I create a commit without a parent in Git?

3 Answers3

Linked

Related