]>
Commit | Line | Data |
---|---|---|
1 | MERGE STRATEGIES | |
2 | ---------------- | |
3 | ||
4 | The merge mechanism (`git merge` and `git pull` commands) allows the | |
5 | backend 'merge strategies' to be chosen with `-s` option. Some strategies | |
6 | can also take their own options, which can be passed by giving `-X<option>` | |
7 | arguments to `git merge` and/or `git pull`. | |
8 | ||
9 | resolve:: | |
10 | This can only resolve two heads (i.e. the current branch | |
11 | and another branch you pulled from) using a 3-way merge | |
12 | algorithm. It tries to carefully detect criss-cross | |
13 | merge ambiguities and is considered generally safe and | |
14 | fast. | |
15 | ||
16 | recursive:: | |
17 | This can only resolve two heads using a 3-way merge | |
18 | algorithm. When there is more than one common | |
19 | ancestor that can be used for 3-way merge, it creates a | |
20 | merged tree of the common ancestors and uses that as | |
21 | the reference tree for the 3-way merge. This has been | |
22 | reported to result in fewer merge conflicts without | |
23 | causing mismerges by tests done on actual merge commits | |
24 | taken from Linux 2.6 kernel development history. | |
25 | Additionally this can detect and handle merges involving | |
26 | renames, but currently cannot make use of detected | |
27 | copies. This is the default merge strategy when pulling | |
28 | or merging one branch. | |
29 | + | |
30 | The 'recursive' strategy can take the following options: | |
31 | ||
32 | ours;; | |
33 | This option forces conflicting hunks to be auto-resolved cleanly by | |
34 | favoring 'our' version. Changes from the other tree that do not | |
35 | conflict with our side are reflected in the merge result. | |
36 | For a binary file, the entire contents are taken from our side. | |
37 | + | |
38 | This should not be confused with the 'ours' merge strategy, which does not | |
39 | even look at what the other tree contains at all. It discards everything | |
40 | the other tree did, declaring 'our' history contains all that happened in it. | |
41 | ||
42 | theirs;; | |
43 | This is the opposite of 'ours'; note that, unlike 'ours', there is | |
44 | no 'theirs' merge strategy to confuse this merge option with. | |
45 | ||
46 | patience;; | |
47 | With this option, 'merge-recursive' spends a little extra time | |
48 | to avoid mismerges that sometimes occur due to unimportant | |
49 | matching lines (e.g., braces from distinct functions). Use | |
50 | this when the branches to be merged have diverged wildly. | |
51 | See also linkgit:git-diff[1] `--patience`. | |
52 | ||
53 | diff-algorithm=[patience|minimal|histogram|myers];; | |
54 | Tells 'merge-recursive' to use a different diff algorithm, which | |
55 | can help avoid mismerges that occur due to unimportant matching | |
56 | lines (such as braces from distinct functions). See also | |
57 | linkgit:git-diff[1] `--diff-algorithm`. | |
58 | ||
59 | ignore-space-change;; | |
60 | ignore-all-space;; | |
61 | ignore-space-at-eol;; | |
62 | ignore-cr-at-eol;; | |
63 | Treats lines with the indicated type of whitespace change as | |
64 | unchanged for the sake of a three-way merge. Whitespace | |
65 | changes mixed with other changes to a line are not ignored. | |
66 | See also linkgit:git-diff[1] `-b`, `-w`, | |
67 | `--ignore-space-at-eol`, and `--ignore-cr-at-eol`. | |
68 | + | |
69 | * If 'their' version only introduces whitespace changes to a line, | |
70 | 'our' version is used; | |
71 | * If 'our' version introduces whitespace changes but 'their' | |
72 | version includes a substantial change, 'their' version is used; | |
73 | * Otherwise, the merge proceeds in the usual way. | |
74 | ||
75 | renormalize;; | |
76 | This runs a virtual check-out and check-in of all three stages | |
77 | of a file when resolving a three-way merge. This option is | |
78 | meant to be used when merging branches with different clean | |
79 | filters or end-of-line normalization rules. See "Merging | |
80 | branches with differing checkin/checkout attributes" in | |
81 | linkgit:gitattributes[5] for details. | |
82 | ||
83 | no-renormalize;; | |
84 | Disables the `renormalize` option. This overrides the | |
85 | `merge.renormalize` configuration variable. | |
86 | ||
87 | no-renames;; | |
88 | Turn off rename detection. This overrides the `merge.renames` | |
89 | configuration variable. | |
90 | See also linkgit:git-diff[1] `--no-renames`. | |
91 | ||
92 | find-renames[=<n>];; | |
93 | Turn on rename detection, optionally setting the similarity | |
94 | threshold. This is the default. This overrides the | |
95 | 'merge.renames' configuration variable. | |
96 | See also linkgit:git-diff[1] `--find-renames`. | |
97 | ||
98 | rename-threshold=<n>;; | |
99 | Deprecated synonym for `find-renames=<n>`. | |
100 | ||
101 | subtree[=<path>];; | |
102 | This option is a more advanced form of 'subtree' strategy, where | |
103 | the strategy makes a guess on how two trees must be shifted to | |
104 | match with each other when merging. Instead, the specified path | |
105 | is prefixed (or stripped from the beginning) to make the shape of | |
106 | two trees to match. | |
107 | ||
108 | octopus:: | |
109 | This resolves cases with more than two heads, but refuses to do | |
110 | a complex merge that needs manual resolution. It is | |
111 | primarily meant to be used for bundling topic branch | |
112 | heads together. This is the default merge strategy when | |
113 | pulling or merging more than one branch. | |
114 | ||
115 | ours:: | |
116 | This resolves any number of heads, but the resulting tree of the | |
117 | merge is always that of the current branch head, effectively | |
118 | ignoring all changes from all other branches. It is meant to | |
119 | be used to supersede old development history of side | |
120 | branches. Note that this is different from the -Xours option to | |
121 | the 'recursive' merge strategy. | |
122 | ||
123 | subtree:: | |
124 | This is a modified recursive strategy. When merging trees A and | |
125 | B, if B corresponds to a subtree of A, B is first adjusted to | |
126 | match the tree structure of A, instead of reading the trees at | |
127 | the same level. This adjustment is also done to the common | |
128 | ancestor tree. | |
129 | ||
130 | With the strategies that use 3-way merge (including the default, 'recursive'), | |
131 | if a change is made on both branches, but later reverted on one of the | |
132 | branches, that change will be present in the merged result; some people find | |
133 | this behavior confusing. It occurs because only the heads and the merge base | |
134 | are considered when performing a merge, not the individual commits. The merge | |
135 | algorithm therefore considers the reverted change as no change at all, and | |
136 | substitutes the changed version instead. |