Does git pull after init remove untracked files

git

I just made a simple script and created a new repository for it on GitHub. After initializing the local repo, moving the script to the repository and pulling the GitHub repository (containing only a README.md file), the script was gone. Is this normal behavior? And is there some way to restore the file?

Here are the commands, as executed:

$ mkgithub ~/dev/cr2meta2jpg
Initialized empty Git repository in /home/username/dev/cr2meta2jpg/.git/
$ mv test.sh ~/dev/cr2meta2jpg/cr2meta2jpg.sh
$ cd ~/dev/cr2meta2jpg/
$ ls
cr2meta2jpg.sh
$ git pull
remote: Counting objects: 3, done.
remote: Compressing objects: 100% (2/2), done.
remote: Total 3 (delta 0), reused 0 (delta 0)
Unpacking objects: 100% (3/3), done.
From github.com:l0b0/cr2meta2jpg
 * [new branch]      master     -> origin/master
$ ls
README.md

Best Answer

A git pull will not overwrite local changes unless you use git add before. Even in this case, you can still recover your data.

When you issued git add before the pull:

mkgithub blub
cp test.sh blub/
cd blub/
git add test.sh
git pull

The pull will delete test.sh. It will only do this if you have no commit in the local repo. After adding a commit this won't remove test.sh (git either ignores it or reports a merge conflict)

The file is not lost. It's still in the Git repository as a dangling blob.

To recover the file:

$ git fsck
Checking object directories: 100% (256/256), done.
Checking objects: 100% (401/401), done.
dangling blob 541060d8292ff808c9715265d063f1bae220da7c
$ git show 541060d8292ff808c9715265d063f1bae220da7 > test.sh

This works as long as you did not issue git gc --prune=now afterwards.

Related Solutions

Github adding a repository as a fork from an existing clone

You can do it all from your existing repository (no need to clone the fork into a new (local) repository, create your branch, copy your commits/changes, etc.).

Get your commits ready to be published.

Refine any existing local commits (e.g. with git commit --amend and/or git rebase --interactive).

Commit any of your uncommitted changes that you want to publish (I am not sure if you meant to imply that you have some commits on your local master and some uncommitted changes, or just some uncommitted changes; incidentally, uncommitted changes are not “on a branch”, they are strictly in your working tree).

Rename your master branch to give it the name you want for your “new branch”. This is not strictly necessary (you can push from any branch to any other branch), but it will probably reduce confusion in the long run if your local branch and the branch in your GitHub fork have the same name.
```
git branch -m master my-feature
```
Fork the upstream GitHub repository
(e.g.) github.com:UpstreamOwner/repostory_name.git as
(e.g.) github.com:YourUser/repository_name.git.

This is done on the GitHub website (or a “client” that uses the GitHub APIs), there are no local Git commands involved.
In your local repository (the one that was originally cloned from the upstream GitHub repository and has your changes in its master), add your fork repository as a remote:
```
git remote add -f github github.com:YourUser/repository_name.git
```
Push your branch to your fork repository on GitHub.
```
git push github my-feature
```
Optionally, rename the remotes so that your fork is known as “origin” and the upstream as “upstream”.
```
git remote rename origin upstream
git remote rename github origin
```
One reason to rename the remotes would be because you want to be able to use git push without specifying a repository (it defaults to “origin”).

Ssh – How to avoid using ‘ssh-add ~/.ssh/id_rsa’ for every push to a github repo

below is based on another answer:

In order to not have to fill in your password even after a restart add the following to your ssh configuration file (which is commonly located at ~/.ssh/config, or on Windows at %UserProfile%/.ssh/config)

Host *
  UseKeychain yes
  AddKeysToAgent yes
  IdentityFile ~/.ssh/id_rsa

Best Answer

Related Solutions

Github adding a repository as a fork from an existing clone

Ssh – How to avoid using ‘ssh-add ~/.ssh/id_rsa’ for every push to a github repo

Related Question