Duplicated lines of code after a merge

Recovery steps

1. 1

   See exactly what the merge changed

   Right after the merge, ORIG_HEAD is your branch before it. Otherwise compare with the merge commit's first parent (HEAD^1).

   Safe

   $ git diff ORIG_HEAD HEAD -- path/to/file
   # later on:
   $ git diff HEAD^1 HEAD -- path/to/file

2. 2

   Find who added each copy

   blame on the duplicated lines shows the commit of each copy: you'll know which branch brought which one.

   Safe

   $ git blame -L 10,60 path/to/file

3. 3

   Look for forgotten conflict markers

   A resolution done too fast often leaves both sides AND the markers.

   Safe

   $ git grep -n -e "<<<<<<<" -e ">>>>>>>"

4. 4

   Fix it by hand (works pushed or not)

   Remove the wrong copy, build, test, then commit the fix. The safest option in every case.

   Safe

   $ git add path/to/file
   $ git commit -m "fix: remove code duplicated by the merge"

5. 5

   Not pushed: redo the merge properly

   Cancel the merge, then merge again with a conflict style that shows the original version: much easier to choose.

   Destructive

   $ git reset --hard ORIG_HEAD
   $ git config --global merge.conflictStyle zdiff3
   $ git merge feature/my-branch

6. 6

   Already pushed and badly broken: revert the merge

   Cancels the whole merge without rewriting history. Merging the branch again later will need a revert of this revert.

   Caution

   $ git revert -m 1 <merge-commit-hash>
   $ git push

These procedures are provided for information only. Read each command and adapt it to your situation before running it. Disclaimer

Other disasters