[git] The following untracked working tree files would be overwritten by merge, but I don't care

On my branch I had some files in .gitignore

On a different branch those files are not.

I want to merge the different branch into mine, and I don't care if those files are no longer ignored or not.

Unfortunately I get this:

The following untracked working tree files would be overwritten by merge

How would I modify my pull command to overwrite those files, without me having to find, move or delete those files myself?

The problem is when we have incoming changes that will merge untracked file, git complains. These commands helped me:

git clean -dxf
git pull origin master

Remove all untracked files:

git clean  -d  -fx .

Caution: this will delete IDE files and any useful files as long as you donot track the files. Use this command with care

Update - a better version

This tool (https://github.com/mklepaczewski/git-clean-before-merge) will:

• delete untracked files that are identical to their git pull equivalents,
• revert changes to modified files who's modified version is identical to their git pull equivalents,
• report modified/untracked files that differ from their git pull version,
• the tool has the --pretend option that will not modify any files.

Old version

How this answer differ from other answers?

The method presented here removes only files that would be overwritten by merge. If you have other untracked (possibly ignored) files in the directory this method won't remove them.

The solution

This snippet will extract all untracked files that would be overwritten by git pull and delete them.

git pull 2>&1|grep -E '^\s'|cut -f2-|xargs -I {} rm -rf "{}"

and then just do:

git pull

This is not git porcelain command so always double check what it would do with:

git pull 2>&1|grep -E '^\s'|cut -f2-|xargs -I {} echo "{}"

Explanation - because one liners are scary:

Here's a breakdown of what it does:

1. git pull 2>&1 - capture git pull output and redirect it all to stdout so we can easily capture it with grep.
2. grep -E '^\s - the intent is to capture the list of the untracked files that would be overwritten by git pull. The filenames have a bunch of whitespace characters in front of them so we utilize it to get them.
3. cut -f2- - remove whitespace from the beginning of each line captured in 2.
4. xargs -I {} rm -rf "{}" - us xargs to iterate over all files, save their name in "{}" and call rm for each of them. We use -rf to force delete and remove untracked directories.

It would be great to replace steps 1-3 with porcelain command, but I'm not aware of any equivalent.

In my case when I had this problem. I had a local file that I had renamed on the remote.

When trying to git pull Git told me the new filename was not tracked -- which it was on the remote although it didn't yet exist on local.

Because there was no instance of it locally I couldn't do git pull until I did git rm on the old filename (which wasn't obvious at first because of my stupid idea of renaming it).

Neither clean/reset/hard checkout/rebase worked for me.

So I just removed files that git complained about*

rm /path/to/files/that/git/complained/about

*I checked if this files can be removed by checking out a brand new repo in a separate folder (files were not there)

You can try that command

git clean -df

In addition to the accepted answer you can of course remove the files if they are no longer needed by specifying the file:

git clean -f '/path/to/file/'

Remember to run it with the -n flag first if you would like to see which files git clean will remove. Note that these files will be deleted. In my case I didn't care about them anyway, so that was a better solution for me.

One way to do this is by stashing you local changes and pulling from the remote repo. In this way, you will not lose your local files as the files will go to the stash.

git add -A
git stash
git pull

You can check your local stashed files using this command - git stash list

If you have the files written under .gitignore, remove the files and run git pull again. That helped me out.

If this is a one-time operation, you could just remove all untracked files from the working directory before doing the pull. Read How to remove local (untracked) files from the current Git working tree? for information on how to remove all untracked files.

Be sure to not accidentally remove untracked file that you still need ;)

For those who don't know, git ignores uppercase/lowercase name differences in files and folders. This turns out to be a nightmare when you rename them to the exact same name with a different case.

I encountered this issue when I renamed a folder from "Petstore" to "petstore" (uppercase to lowercase). I had edited my .git/config file to stop ignoring case, made changes, squashed my commits, and stashed my changes to move to a different branch. I could not apply my stashed changes to this other branch.

The fix that I found that worked was to temporarily edit my .git/config file to temporarily ignore case again. This caused git stash apply to succeed. Then, I changed ignoreCase back to false. I then added everything except for the new files in the petstore folder which git oddly claimed were deleted, for whatever reason. I committed my changes, then ran git reset --hard HEAD to get rid of those untracked new files. My commit appeared exactly as expected: the files in the folder were renamed.

I hope that this helps you avoid my same nightmare.

The only commands that worked for me were:

git fetch --all
git reset --hard origin/{{your branch name}}

A replacement for git merge that will overwrite untracked files

The comments below use 'FOI' for the 'files of interest', the files that

• exist in the donor branch,
• do not exist in the receiving branch,
• and are blocking the merge because they are present and untracked in your working directory.

git checkout -f donor-branch   # replace FOI with tracked `donor` versions
git checkout receiving-branch  # FOI are not in `receiving`, so they disapppear
git merge donor-branch  # now the merge works

A replacement for git pull that will overwrite untracked files

pull = fetch + merge, so we do git fetch followed by the git checkout -f, git checkout, git merge trick above.

git fetch origin  # fetch remote commits
git checkout -f origin/mybranch  # replace FOI with tracked upstream versions
git checkout mybranch  # FOI are not in mybranch, so they disapppear
git merge origin/mybranch  # Now the merge works. fetch + merge completes the pull.

Detailed explanation

git merge -f does not exist, but git checkout -f does.

We will use git checkout -f + git checkout to remove the Files Of Interest (see above), and then your merge can proceed normally.

Step 1. This step forcibly replaces untracked FOI with tracked versions of the donor branch (it also checks out the donor branch, and updates the rest of the working dir).

git checkout -f donor-branch

Step 2. This step removes the FOI because they they are tracked in our current (donor) branch, and absent in the receiving-branch we switch to.

git checkout receiving-branch

Step 3. Now that the FOI are absent, merging in the donor branch will not overwrite any untracked files, so we get no errors.

git merge donor-branch

You can try command to clear the untracked files from the local

Git 2.11 and newer versions:

git clean  -d  -f .

Older versions of Git:

git clean  -d  -f ""

Where -d can be replaced with the following:

• -x ignored files are also removed as well as files unknown to Git.

• -d remove untracked directories in addition to untracked files.

• -f is required to force it to run.

Here is the link that can be helpful as well.

If you consider using the -f flag you might first run it as a dry-run. Just that you know upfront what kind of interesting situation you will end up next ;-P

-n 
--dry-run 
    Don’t actually remove anything, just show what would be done.