6 git-read-tree - Reads tree information into the index
11 'git-read-tree' (<tree-ish> | [-m [-u|-i]] <tree-ish1> [<tree-ish2> [<tree-ish3>]])
16 Reads the tree information given by <tree-ish> into the index,
17 but does not actually *update* any of the files it "caches". (see:
20 Optionally, it can merge a tree into the index, perform a
21 fast-forward (i.e. 2-way) merge, or a 3-way merge, with the -m
22 flag. When used with -m, the -u flag causes it to also update
23 the files in the work tree with the result of the merge.
25 Trivial merges are done by "git-read-tree" itself. Only conflicting paths
26 will be in unmerged state when "git-read-tree" returns.
31 Perform a merge, not just a read.
35 Same as -m except that unmerged entries will be silently ignored.
38 After a successful merge, update the files in the work
39 tree with the result of the merge.
42 Usually a merge requires the index file as well as the
43 files in the working tree are up to date with the
44 current head commit, in order not to lose local
45 changes. This flag disables the check with the working
46 tree and is meant to be used when creating a merge of
47 trees that are not directly related to the current
48 working tree status into a temporary index file.
52 The id of the tree object(s) to be read/merged.
57 If '-m' is specified, "git-read-tree" can perform 3 kinds of
58 merge, a single tree merge if only 1 tree is given, a
59 fast-forward merge with 2 trees, or a 3-way merge if 3 trees are
65 If only 1 tree is specified, git-read-tree operates as if the user did not
66 specify '-m', except that if the original index has an entry for a
67 given pathname, and the contents of the path matches with the tree
68 being read, the stat info from the index is used. (In other words, the
69 index's stat()s take precedence over the merged tree's).
71 That means that if you do a "git-read-tree -m <newtree>" followed by a
72 "git-checkout-index -f -u -a", the "git-checkout-index" only checks out
73 the stuff that really changed.
75 This is used to avoid unnecessary false hits when "git-diff-files" is
76 run after git-read-tree.
82 Typically, this is invoked as "git-read-tree -m $H $M", where $H
83 is the head commit of the current repository, and $M is the head
84 of a foreign tree, which is simply ahead of $H (i.e. we are in a
85 fast forward situation).
87 When two trees are specified, the user is telling git-read-tree
90 1. The current index and work tree is derived from $H, but
91 the user may have local changes in them since $H;
93 2. The user wants to fast-forward to $M.
95 In this case, the "git-read-tree -m $H $M" command makes sure
96 that no local change is lost as the result of this "merge".
97 Here are the "carry forward" rules:
100 -------------------------------------------------------
101 0 nothing nothing nothing (does not happen)
102 1 nothing nothing exists use M
103 2 nothing exists nothing remove path from index
104 3 nothing exists exists use M
108 4 yes N/A N/A nothing nothing keep index
109 5 no N/A N/A nothing nothing keep index
111 6 yes N/A yes nothing exists keep index
112 7 no N/A yes nothing exists keep index
113 8 yes N/A no nothing exists fail
114 9 no N/A no nothing exists fail
116 10 yes yes N/A exists nothing remove path from index
117 11 no yes N/A exists nothing fail
118 12 yes no N/A exists nothing fail
119 13 no no N/A exists nothing fail
123 14 yes exists exists keep index
124 15 no exists exists keep index
126 clean I==H I==M (H!=M)
128 16 yes no no exists exists fail
129 17 no no no exists exists fail
130 18 yes no yes exists exists keep index
131 19 no no yes exists exists keep index
132 20 yes yes no exists exists use M
133 21 no yes no exists exists fail
135 In all "keep index" cases, the index entry stays as in the
136 original index file. If the entry were not up to date,
137 git-read-tree keeps the copy in the work tree intact when
138 operating under the -u flag.
140 When this form of git-read-tree returns successfully, you can
141 see what "local changes" you made are carried forward by running
142 "git-diff-index --cached $M". Note that this does not
143 necessarily match "git-diff-index --cached $H" would have
144 produced before such a two tree merge. This is because of cases
145 18 and 19 --- if you already had the changes in $M (e.g. maybe
146 you picked it up via e-mail in a patch form), "git-diff-index
147 --cached $H" would have told you about the change before this
148 merge, but it would not show in "git-diff-index --cached $M"
149 output after two-tree merge.
154 Each "index" entry has two bits worth of "stage" state. stage 0 is the
155 normal one, and is the only one you'd see in any kind of normal use.
157 However, when you do "git-read-tree" with three trees, the "stage"
160 This means that you can do
162 git-read-tree -m <tree1> <tree2> <tree3>
164 and you will end up with an index with all of the <tree1> entries in
165 "stage1", all of the <tree2> entries in "stage2" and all of the
166 <tree3> entries in "stage3".
168 Furthermore, "git-read-tree" has special-case logic that says: if you see
169 a file that matches in all respects in the following states, it
170 "collapses" back to "stage0":
172 - stage 2 and 3 are the same; take one or the other (it makes no
173 difference - the same work has been done on stage 2 and 3)
175 - stage 1 and stage 2 are the same and stage 3 is different; take
176 stage 3 (some work has been done on stage 3)
178 - stage 1 and stage 3 are the same and stage 2 is different take
179 stage 2 (some work has been done on stage 2)
181 The "git-write-tree" command refuses to write a nonsensical tree, and it
182 will complain about unmerged entries if it sees a single entry that is not
185 Ok, this all sounds like a collection of totally nonsensical rules,
186 but it's actually exactly what you want in order to do a fast
187 merge. The different stages represent the "result tree" (stage 0, aka
188 "merged"), the original tree (stage 1, aka "orig"), and the two trees
189 you are trying to merge (stage 2 and 3 respectively).
191 The order of stages 1, 2 and 3 (hence the order of three
192 <tree-ish> command line arguments) are significant when you
193 start a 3-way merge with an index file that is already
194 populated. Here is an outline of how the algorithm works:
196 - if a file exists in identical format in all three trees, it will
197 automatically collapse to "merged" state by git-read-tree.
199 - a file that has _any_ difference what-so-ever in the three trees
200 will stay as separate entries in the index. It's up to "porcelain
201 policy" to determine how to remove the non-0 stages, and insert a
204 - the index file saves and restores with all this information, so you
205 can merge things incrementally, but as long as it has entries in
206 stages 1/2/3 (ie "unmerged entries") you can't write the result. So
207 now the merge algorithm ends up being really simple:
209 * you walk the index in order, and ignore all entries of stage 0,
210 since they've already been done.
212 * if you find a "stage1", but no matching "stage2" or "stage3", you
213 know it's been removed from both trees (it only existed in the
214 original tree), and you remove that entry.
216 * if you find a matching "stage2" and "stage3" tree, you remove one
217 of them, and turn the other into a "stage0" entry. Remove any
218 matching "stage1" entry if it exists too. .. all the normal
221 You would normally use "git-merge-index" with supplied
222 "git-merge-one-file" to do this last step. The script
223 does not touch the files in the work tree, and the entire merge
224 happens in the index file. In other words, there is no need to
225 worry about what is in the working directory, since it is never
226 shown and never used.
228 When you start a 3-way merge with an index file that is already
229 populated, it is assumed that it represents the state of the
230 files in your work tree, and you can even have files with
231 changes unrecorded in the index file. It is further assumed
232 that this state is "derived" from the stage 2 tree. The 3-way
233 merge refuses to run if it finds an entry in the original index
234 file that does not match stage 2.
236 This is done to prevent you from losing your work-in-progress
237 changes. To illustrate, suppose you start from what has been
238 commited last to your repository:
240 $ JC=`git-rev-parse --verify "HEAD^0"`
241 $ git-checkout-index -f -u -a $JC
243 You do random edits, without running git-update-index. And then
244 you notice that the tip of your "upstream" tree has advanced
245 since you pulled from him:
247 $ git-fetch rsync://.... linus
248 $ LT=`cat .git/MERGE_HEAD`
250 Your work tree is still based on your HEAD ($JC), but you have
251 some edits since. Three-way merge makes sure that you have not
252 added or modified index entries since $JC, and if you haven't,
253 then does the right thing. So with the following sequence:
255 $ git-read-tree -m -u `git-merge-base $JC $LT` $JC $LT
256 $ git-merge-index git-merge-one-file -a
257 $ echo "Merge with Linus" | \
258 git-commit-tree `git-write-tree` -p $JC -p $LT
260 what you would commit is a pure merge between $JC and LT without
261 your work-in-progress changes, and your work tree would be
262 updated to the result of the merge.
267 gitlink:git-write-tree[1]; gitlink:git-ls-files[1]
272 Written by Linus Torvalds <torvalds@osdl.org>
276 Documentation by David Greaves, Junio C Hamano and the git-list <git@vger.kernel.org>.
280 Part of the gitlink:git[7] suite