What's the difference between git rebase and merge --no-ff

12,866

Solution 1

It depends how you want your history to look.

A really simple example:

git merge --no-ff

*   33ad16c (HEAD, master) Merge branch 'feature-1'
|\  
| * e61c6a6 Various bug fixes
| * e15a356 Started feature
|/  
* 38b8ada Initial commit

git rebase

* e61c6a6 (HEAD, master) Various bug fixes
* e15a356 Started feature
* 38b8ada Initial commit

These two histories reflect the exact same commits but using the --no-ff, the first history makes it really clear how 2 of the commits where, in fact, part of the same set of work. This allows you to know exactly what work went into what feature. This makes it easier to roll back changes or just give yourself context when navigating the history.

The rebase history does not show this distinction and it makes it impossible to see any grouping of work. It makes it look like everyone is committing to master.

The empty merge commit you get with --no-ff also gives a helpful place to hang a commit message, whereas a plain commit gives you no easy place to say "This commit merges in the work for feature-x", which again is helpful for history

Solution 2

Generally if you merge some feature branch you want to use --no-ff to enforce the creation of an explicit merge commit.

If you are just doing your work on some branch you generally want to avoid merge commits that might arise from a git pull. Therfore you use git rebase or git pull --rebase to not create a merge commit and get a clean history.

Share:
12,866

Related videos on Youtube

James Lin
Author by

James Lin

Passionate about writing software to solve real life problems.

Updated on August 08, 2022

Comments