6 git-merge - Join two or more development histories together
12 'git-merge' [-n] [--summary] [--no-commit] [--squash] [-s <strategy>]...
13 [-m <msg>] <remote> <remote>...
17 This is the top-level interface to the merge machinery
18 which drives multiple merge strategy scripts.
23 include::merge-options.txt[]
26 The commit message to be used for the merge commit (in case
27 it is created). The `git-fmt-merge-msg` script can be used
28 to give a good default for automated `git-merge` invocations.
31 Our branch head commit. This has to be `HEAD`, so new
32 syntax does not require it
35 Other branch head merged into our branch. You need at
36 least one <remote>. Specifying more than one <remote>
37 obviously means you are trying an Octopus.
39 include::merge-strategies.txt[]
42 If you tried a merge which resulted in a complex conflicts and
43 would want to start over, you can recover with
50 Whether to include summaries of merged commits in newly
51 created merge commit. False by default.
54 Controls the amount of output shown by the recursive merge
55 strategy. Level 0 outputs nothing except a final error
56 message if conflicts were detected. Level 1 outputs only
57 conflicts, 2 outputs conflicts and file changes. Level 5 and
58 above outputs debugging information. The default is level 2.
59 Can be overriden by 'GIT_MERGE_VERBOSITY' environment variable.
65 A merge is always between the current `HEAD` and one or more
66 remote branch heads, and the index file must exactly match the
67 tree of `HEAD` commit (i.e. the contents of the last commit) when
68 it happens. In other words, `git-diff --cached HEAD` must
72 This is a bit of lie. In certain special cases, your index are
73 allowed to be different from the tree of `HEAD` commit. The most
74 notable case is when your `HEAD` commit is already ahead of what
75 is being merged, in which case your index can have arbitrary
76 difference from your `HEAD` commit. Otherwise, your index entries
77 are allowed have differences from your `HEAD` commit that match
78 the result of trivial merge (e.g. you received the same patch
79 from external source to produce the same result as what you are
80 merging). For example, if a path did not exist in the common
81 ancestor and your head commit but exists in the tree you are
82 merging into your repository, and if you already happen to have
83 that path exactly in your index, the merge does not have to
86 Otherwise, merge will refuse to do any harm to your repository
87 (that is, it may fetch the objects from remote, and it may even
88 update the local branch used to keep track of the remote branch
89 with `git pull remote rbranch:lbranch`, but your working tree,
90 `.git/HEAD` pointer and index file are left intact).
92 You may have local modifications in the working tree files. In
93 other words, `git-diff` is allowed to report changes.
94 However, the merge uses your working tree as the working area,
95 and in order to prevent the merge operation from losing such
96 changes, it makes sure that they do not interfere with the
97 merge. Those complex tables in read-tree documentation define
98 what it means for a path to "interfere with the merge". And if
99 your local modifications interfere with the merge, again, it
100 stops before touching anything.
102 So in the above two "failed merge" case, you do not have to
103 worry about loss of data --- you simply were not ready to do
104 a merge, so no merge happened at all. You may want to finish
105 whatever you were in the middle of doing, and retry the same
106 pull after you are done and ready.
108 When things cleanly merge, these things happen:
110 1. the results are updated both in the index file and in your
112 2. index file is written out as a tree,
113 3. the tree gets committed, and
114 4. the `HEAD` pointer gets advanced.
116 Because of 2., we require that the original state of the index
117 file to match exactly the current `HEAD` commit; otherwise we
118 will write out your local changes already registered in your
119 index file along with the merge result, which is not good.
120 Because 1. involves only the paths different between your
121 branch and the remote branch you are pulling from during the
122 merge (which is typically a fraction of the whole tree), you can
123 have local modifications in your working tree as long as they do
124 not overlap with what the merge updates.
126 When there are conflicts, these things happen:
128 1. `HEAD` stays the same.
130 2. Cleanly merged paths are updated both in the index file and
131 in your working tree.
133 3. For conflicting paths, the index file records up to three
134 versions; stage1 stores the version from the common ancestor,
135 stage2 from `HEAD`, and stage3 from the remote branch (you
136 can inspect the stages with `git-ls-files -u`). The working
137 tree files have the result of "merge" program; i.e. 3-way
138 merge result with familiar conflict markers `<<< === >>>`.
140 4. No other changes are done. In particular, the local
141 modifications you had before you started merge will stay the
142 same and the index entries for them stay as they were,
143 i.e. matching `HEAD`.
145 After seeing a conflict, you can do two things:
147 * Decide not to merge. The only clean-up you need are to reset
148 the index file to the `HEAD` commit to reverse 2. and to clean
149 up working tree changes made by 2. and 3.; `git-reset` can
152 * Resolve the conflicts. `git-diff` would report only the
153 conflicting paths because of the above 2. and 3.. Edit the
154 working tree files into a desirable shape, `git-add` or `git-rm`
155 them, to make the index file contain what the merge result
156 should be, and run `git-commit` to commit the result.
161 gitlink:git-fmt-merge-msg[1], gitlink:git-pull[1]
166 Written by Junio C Hamano <junkio@cox.net>
171 Documentation by Junio C Hamano and the git-list <git@vger.kernel.org>.
175 Part of the gitlink:git[7] suite