CODEWITHSUNDEEP
gitlogging
explorer
explorer

Reputation: 89

Is it possible to carry history of a file from a git branch to another (using command line)

I've two branches in one repo : branch_1 and branch_2.

Scenario 1

somebody added a file in branch_1 let's say first_file.txt. To get this file in branch_2 - I'll simply raise a pull request and once the PR is merged I'll have the file in branch_2.

Scenario 2

somebody added a file in branch_1 let's say first_file.txt. Now to get this file, I do below on terminal : 

(on branch branch_1) $ git pull  

(on branch branch_1) $ git checkout branch_2

(on branch branch_2) $ git checkout branch_1 first_file.txt

and then performs git add,commit and push.

In both above scenarios I get the file in my branch (branch_2). but the issue is when I see git log (or history on git portal) they are completely different.

For Scenario 1 : it shows that file came from somebody.

For Scenario 2 : it shows that file is checked-in by me.

Is there a solution to copy files from one branch to another using command line, keeping history safe?

Thanks

Upvotes: 2

Views: 867

Answers (1)

torek
torek

Reputation: 488183

The only history in a Git repository is the set of commits in the repository.

Each commit contains a complete snapshot of every file—or rather, of every file that is in that commit. That is, suppose we start with a simple repository with one commit containing one file:

A   <-- master (HEAD)

(A stands in for some raw hash ID). Commit A contains a README.md file, and nothing else.

Now you make a new commit in which you add a new file, such as main.py:

A--B   <-- master (HEAD)

New commit B contains two files: README.md—which is the same as the copy in commit A—and main.py.

The history in this repository is that commit B leads back to commit A. That's the only history there is.

In your case, you have a more complex starting setup. Simplified, it probably looks about like this:

          I--J   <-- branch_1
         /
...--G--H
         \
          K   <-- branch_2

Commit H is the first commit where the two branches join, in terms of history: history from J backwards goes J, I, H, G, and so on. History from K goes K, H, G, and so on.

In commits up through I, and your commits up through K, there is no file named first_file.txt, but in commit J, someone else added this new file. Therefore commit J has all the files that I has, plus this one new file. Viewing the history from J backwards, the difference between I and J is that the file is added.

If you now make a new commit L that comes after K that has the file in it, the difference between commit K and your new commit L is that this file is added. So the history as viewed from L backwards shows the file added.

That's simply how Git works. There is no file history, there is only commit history. You can ask Git to produce a subset of commit history that traces which commits had activity in specific files—but since the commit history will be that the file newly appears in L as compared to K, that is what you will see.

Upvotes: 1

Related Questions