With --no-commit perform the merge but pretend the merge failed and do
not autocommit, to give the user a chance to inspect and further tweak
the merge result before
committing.
For anyone who just needs to micromanage a merge, jump to the Hybrid section below.
For anyone who is wondering about the difference between @True's answer using git difftool and the other answers that use git merge, see Git mergetool vs difftool.
Here are more details:
Difftool
If you have git configured to use a modern diff.tool such as kdiff3, meld, or vimdiff, you'll be able to manually merge using that diff tool, and the command line can be simple:
git difftool other_branch
...this will let you do a two-way manual merge between your current branch and other_branch (described as $LOCAL and $REMOTE in man git-config).
Mergetool
The "correct" way the other answers discuss would be to instead configure git to use e.g. kdiff3 or vimdiff as your merge.tool, and use:
...this command can do an N-way manual merge between $BASE, $LOCAL, and $REMOTE, into $MERGED. See https://stackoverflow.com/a/2235841/1264797 for one example of how to configure git. You many not need to configure the mergetool.*.cmd entry at all if you use one of the tools git already knows about. (Meld can only show three panes, so if you use meld with the default settings, you'll not see $BASE.)
Difftool vs mergetool
Someone might jump in to correct me, but the main differences between the above difftool and mergetool techniques seem to be:
mergetool can do an N-way merge if configured accordingly
mergetool adds other_branch as a parent on the new commit so history works correctly
difftool lets you manually see and select each and every changed line, but you lose the above two benefits of mergetool
Hybrid -- best of both worlds
A method that combines merging and diffing would look something like this:
...this does the N-way merge to resolve conflicts and make history work right, and then shows you the complete set of diffs so you can review and tweak before you commit.
The merge mechanism (git merge and git pull commands) allows the backend 'merge strategies' to be chosen with -s option. Some strategies can also take their own options, which can be passed by giving -X arguments to git merge and/or git pull.
ours
This resolves any number of heads, but the resulting tree of the merge
is always that of the current branch head, effectively ignoring all
changes from all other branches. It is meant to be used to supersede
old development history of side branches. Note that this is different
from the -Xours option to the 'recursive' merge strategy.
What Bitbucket see later however is a mystery to me, it recognise the commit as merge but fail to actually merge the branch (does not solve a pull-request) - probably Bitbucket guru could help on this issue, I can not even give you any logs/error messages since I do not have that visibility - git/TortoiseGit does not complain at all though.
I found the other answers unsatisfactory and became frustrated searching for an answer. A solution to this question I finally found here: https://stackoverflow.com/a/11593308/1351182
If you run these commands, you will create a new commit which essentially takes the latest commit of branchToMergeFrom and allows you to apply a patch on top of it, which I think is like an additional commit on top.
You will then be prompted (file-by-file if you didn't specify file) on exactly which 'hunks' you want to merge. In this way it walks you through each part of what would have been the automatic merge process and instead asks for manual arbitration on which bits and pieces you want to accept from the mergefrom branch. Here's an example of what it looked like in my project:
@@ -249,7 +251,8 @@ def draw_everything():
draw_bg()
draw_balls(ax)
- plt.show(block=False)
+ if show:
+ plt.show(block=False)
def advance(ms, accel_fun, collision_matrix_fun):
global balls
(3/6) Apply this hunk to index and worktree [y,n,q,a,d,K,j,J,g,/,e,?]?
After typing y and <Enter>, I was presented with the next hunk, (4/6) for that file. This prompt at the bottom lets you simply accept the merge 'hunk' with y, reject it with n, or even go in and edit it manually. Here are the options:
y - apply this hunk to index and worktree
n - do not apply this hunk to index and worktree
q - quit; do not apply this hunk or any of the remaining ones
a - apply this hunk and all later hunks in the file
d - do not apply this hunk or any of the later hunks in the file
g - select a hunk to go to
/ - search for a hunk matching the given regex
j - leave this hunk undecided, see next undecided hunk
J - leave this hunk undecided, see next hunk
K - leave this hunk undecided, see previous hunk
s - split the current hunk into smaller hunks
e - manually edit the current hunk
? - print help
I wanted to go in and manually edit one hunk, as I didn't want to accept or reject the merge exactly as it was posed. So I chose e and was given a file to edit. I was pleased when I noticed there were even instructions at the bottom on how to edit the hunk properly. You can even split hunks into smaller ones with the s option as above.
I would recommend this process if what you want is manual merging where you still leverage the automatic process as much as possible. The difference is that you get to oversee every merge 'hunk' and edit them as you please. I hope this helps future readers.
After this process, you probably want to run git checkout branchToMergeTo && git merge branchToMergeFrom in order to formally merge the history of branchToMergeFrom into branchToMergeTo.