1GIT-READ-TREE(1) Git Manual GIT-READ-TREE(1)
2
6 git-read-tree - Reads tree information into the index
7
9 git read-tree [[-m [--trivial] [--aggressive] | --reset | --prefix=<prefix>]
10 [-u [--exclude-per-directory=<gitignore>] | -i]]
11 [--index-output=<file>] [--no-sparse-checkout]
12 (--empty | <tree-ish1> [<tree-ish2> [<tree-ish3>]])
13
16 Reads the tree information given by <tree-ish> into the index, but does
17 not actually update any of the files it "caches". (see: git-checkout-
18 index(1))
19 Optionally, it can merge a tree into the index, perform a fast-forward
21 (i.e. 2-way) merge, or a 3-way merge, with the -m flag. When used with
22 -m, the -u flag causes it to also update the files in the work tree
23 with the result of the merge.
24 Trivial merges are done by git read-tree itself. Only conflicting paths
26 will be in unmerged state when git read-tree returns.
27
29 -m
30 Perform a merge, not just a read. The command will refuse to run if
31 your index file has unmerged entries, indicating that you have not
32 finished previous merge you started.
33 --reset
35 Same as -m, except that unmerged entries are discarded instead of
36 failing.
37 -u
39 After a successful merge, update the files in the work tree with
40 the result of the merge.
41 -i
43 Usually a merge requires the index file as well as the files in the
44 working tree to be up to date with the current head commit, in
45 order not to lose local changes. This flag disables the check with
46 the working tree and is meant to be used when creating a merge of
47 trees that are not directly related to the current working tree
48 status into a temporary index file.
49 -n, --dry-run
51 Check if the command would error out, without updating the index or
52 the files in the working tree for real.
53 -v
55 Show the progress of checking files out.
56 --trivial
58 Restrict three-way merge by git read-tree to happen only if there
59 is no file-level merging required, instead of resolving merge for
60 trivial cases and leaving conflicting files unresolved in the
61 index.
62 --aggressive
64 Usually a three-way merge by git read-tree resolves the merge for
65 really trivial cases and leaves other cases unresolved in the
66 index, so that porcelains can implement different merge policies.
67 This flag makes the command resolve a few more cases internally:
68 · when one side removes a path and the other side leaves the path
70 unmodified. The resolution is to remove that path.
71 · when both sides remove a path. The resolution is to remove that
73 path.
74 · when both sides add a path identically. The resolution is to
76 add that path.
77 --prefix=<prefix>
79 Keep the current index contents, and read the contents of the named
80 tree-ish under the directory at <prefix>. The command will refuse
81 to overwrite entries that already existed in the original index
82 file.
83 --exclude-per-directory=<gitignore>
85 When running the command with -u and -m options, the merge result
86 may need to overwrite paths that are not tracked in the current
87 branch. The command usually refuses to proceed with the merge to
88 avoid losing such a path. However this safety valve sometimes gets
89 in the way. For example, it often happens that the other branch
90 added a file that used to be a generated file in your branch, and
91 the safety valve triggers when you try to switch to that branch
92 after you ran make but before running make clean to remove the
93 generated file. This option tells the command to read per-directory
94 exclude file (usually .gitignore) and allows such an untracked but
95 explicitly ignored file to be overwritten.
96 --index-output=<file>
98 Instead of writing the results out to $GIT_INDEX_FILE, write the
99 resulting index in the named file. While the command is operating,
100 the original index file is locked with the same mechanism as usual.
101 The file must allow to be rename(2)ed into from a temporary file
102 that is created next to the usual index file; typically this means
103 it needs to be on the same filesystem as the index file itself, and
104 you need write permission to the directories the index file and
105 index output file are located in.
106 --[no-]recurse-submodules
108 Using --recurse-submodules will update the content of all
109 initialized submodules according to the commit recorded in the
110 superproject by calling read-tree recursively, also setting the
111 submodules HEAD to be detached at that commit.
112 --no-sparse-checkout
114 Disable sparse checkout support even if core.sparseCheckout is
115 true.
116 --empty
118 Instead of reading tree object(s) into the index, just empty it.
119 <tree-ish#>
121 The id of the tree object(s) to be read/merged.
122
124 If -m is specified, git read-tree can perform 3 kinds of merge, a
125 single tree merge if only 1 tree is given, a fast-forward merge with 2
126 trees, or a 3-way merge if 3 or more trees are provided.
127 Single Tree Merge
129 If only 1 tree is specified, git read-tree operates as if the user did
130 not specify -m, except that if the original index has an entry for a
131 given pathname, and the contents of the path match with the tree being
132 read, the stat info from the index is used. (In other words, the
133 index’s stat()s take precedence over the merged tree’s).
134 That means that if you do a git read-tree -m <newtree> followed by a
136 git checkout-index -f -u -a, the git checkout-index only checks out the
137 stuff that really changed.
138 This is used to avoid unnecessary false hits when git diff-files is run
140 after git read-tree.
141 Two Tree Merge
143 Typically, this is invoked as git read-tree -m $H $M, where $H is the
144 head commit of the current repository, and $M is the head of a foreign
145 tree, which is simply ahead of $H (i.e. we are in a fast-forward
146 situation).
147 When two trees are specified, the user is telling git read-tree the
149 following:
150 1. The current index and work tree is derived from $H, but the user
152 may have local changes in them since $H.
153 2. The user wants to fast-forward to $M.
155 In this case, the git read-tree -m $H $M command makes sure that no
157 local change is lost as the result of this "merge". Here are the "carry
158 forward" rules, where "I" denotes the index, "clean" means that index
159 and work tree coincide, and "exists"/"nothing" refer to the presence of
160 a path in the specified commit:
161 I H M Result
163 -------------------------------------------------------
164 0 nothing nothing nothing (does not happen)
165 1 nothing nothing exists use M
166 2 nothing exists nothing remove path from index
167 3 nothing exists exists, use M if "initial checkout",
168 H == M keep index otherwise
169 exists, fail
170 H != M
171 clean I==H I==M
173 ------------------
174 4 yes N/A N/A nothing nothing keep index
175 5 no N/A N/A nothing nothing keep index
176 6 yes N/A yes nothing exists keep index
178 7 no N/A yes nothing exists keep index
179 8 yes N/A no nothing exists fail
180 9 no N/A no nothing exists fail
181 10 yes yes N/A exists nothing remove path from index
183 11 no yes N/A exists nothing fail
184 12 yes no N/A exists nothing fail
185 13 no no N/A exists nothing fail
186 clean (H==M)
188 ------
189 14 yes exists exists keep index
190 15 no exists exists keep index
191 clean I==H I==M (H!=M)
193 ------------------
194 16 yes no no exists exists fail
195 17 no no no exists exists fail
196 18 yes no yes exists exists keep index
197 19 no no yes exists exists keep index
198 20 yes yes no exists exists use M
199 21 no yes no exists exists fail
200 In all "keep index" cases, the index entry stays as in the original
202 index file. If the entry is not up to date, git read-tree keeps the
203 copy in the work tree intact when operating under the -u flag.
204 When this form of git read-tree returns successfully, you can see which
206 of the "local changes" that you made were carried forward by running
207 git diff-index --cached $M. Note that this does not necessarily match
208 what git diff-index --cached $H would have produced before such a two
209 tree merge. This is because of cases 18 and 19 --- if you already had
210 the changes in $M (e.g. maybe you picked it up via e-mail in a patch
211 form), git diff-index --cached $H would have told you about the change
212 before this merge, but it would not show in git diff-index --cached $M
213 output after the two-tree merge.
214 Case 3 is slightly tricky and needs explanation. The result from this
216 rule logically should be to remove the path if the user staged the
217 removal of the path and then switching to a new branch. That however
218 will prevent the initial checkout from happening, so the rule is
219 modified to use M (new tree) only when the content of the index is
220 empty. Otherwise the removal of the path is kept as long as $H and $M
221 are the same.
222 3-Way Merge
224 Each "index" entry has two bits worth of "stage" state. stage 0 is the
225 normal one, and is the only one you’d see in any kind of normal use.
226 However, when you do git read-tree with three trees, the "stage" starts
228 out at 1.
229 This means that you can do
231 $ git read-tree -m <tree1> <tree2> <tree3>
233 and you will end up with an index with all of the <tree1> entries in
236 "stage1", all of the <tree2> entries in "stage2" and all of the <tree3>
237 entries in "stage3". When performing a merge of another branch into the
238 current branch, we use the common ancestor tree as <tree1>, the current
239 branch head as <tree2>, and the other branch head as <tree3>.
240 Furthermore, git read-tree has special-case logic that says: if you see
242 a file that matches in all respects in the following states, it
243 "collapses" back to "stage0":
244 · stage 2 and 3 are the same; take one or the other (it makes no
246 difference - the same work has been done on our branch in stage 2
247 and their branch in stage 3)
248 · stage 1 and stage 2 are the same and stage 3 is different; take
250 stage 3 (our branch in stage 2 did not do anything since the
251 ancestor in stage 1 while their branch in stage 3 worked on it)
252 · stage 1 and stage 3 are the same and stage 2 is different take
254 stage 2 (we did something while they did nothing)
255 The git write-tree command refuses to write a nonsensical tree, and it
257 will complain about unmerged entries if it sees a single entry that is
258 not stage 0.
259 OK, this all sounds like a collection of totally nonsensical rules, but
261 it’s actually exactly what you want in order to do a fast merge. The
262 different stages represent the "result tree" (stage 0, aka "merged"),
263 the original tree (stage 1, aka "orig"), and the two trees you are
264 trying to merge (stage 2 and 3 respectively).
265 The order of stages 1, 2 and 3 (hence the order of three <tree-ish>
267 command-line arguments) are significant when you start a 3-way merge
268 with an index file that is already populated. Here is an outline of how
269 the algorithm works:
270 · if a file exists in identical format in all three trees, it will
272 automatically collapse to "merged" state by git read-tree.
273 · a file that has any difference what-so-ever in the three trees will
275 stay as separate entries in the index. It’s up to "porcelain
276 policy" to determine how to remove the non-0 stages, and insert a
277 merged version.
278 · the index file saves and restores with all this information, so you
280 can merge things incrementally, but as long as it has entries in
281 stages 1/2/3 (i.e., "unmerged entries") you can’t write the result.
282 So now the merge algorithm ends up being really simple:
283 · you walk the index in order, and ignore all entries of stage 0,
285 since they’ve already been done.
286 · if you find a "stage1", but no matching "stage2" or "stage3",
288 you know it’s been removed from both trees (it only existed in
289 the original tree), and you remove that entry.
290 · if you find a matching "stage2" and "stage3" tree, you remove
292 one of them, and turn the other into a "stage0" entry. Remove
293 any matching "stage1" entry if it exists too. .. all the normal
294 trivial rules ..
295 You would normally use git merge-index with supplied git merge-one-file
297 to do this last step. The script updates the files in the working tree
298 as it merges each path and at the end of a successful merge.
299 When you start a 3-way merge with an index file that is already
301 populated, it is assumed that it represents the state of the files in
302 your work tree, and you can even have files with changes unrecorded in
303 the index file. It is further assumed that this state is "derived" from
304 the stage 2 tree. The 3-way merge refuses to run if it finds an entry
305 in the original index file that does not match stage 2.
306 This is done to prevent you from losing your work-in-progress changes,
308 and mixing your random changes in an unrelated merge commit. To
309 illustrate, suppose you start from what has been committed last to your
310 repository:
311 $ JC=`git rev-parse --verify "HEAD^0"`
313 $ git checkout-index -f -u -a $JC
314 You do random edits, without running git update-index. And then you
317 notice that the tip of your "upstream" tree has advanced since you
318 pulled from him:
319 $ git fetch git://.... linus
321 $ LT=`git rev-parse FETCH_HEAD`
322 Your work tree is still based on your HEAD ($JC), but you have some
325 edits since. Three-way merge makes sure that you have not added or
326 modified index entries since $JC, and if you haven’t, then does the
327 right thing. So with the following sequence:
328 $ git read-tree -m -u `git merge-base $JC $LT` $JC $LT
330 $ git merge-index git-merge-one-file -a
331 $ echo "Merge with Linus" | \
332 git commit-tree `git write-tree` -p $JC -p $LT
333 what you would commit is a pure merge between $JC and $LT without your
336 work-in-progress changes, and your work tree would be updated to the
337 result of the merge.
338 However, if you have local changes in the working tree that would be
340 overwritten by this merge, git read-tree will refuse to run to prevent
341 your changes from being lost.
342 In other words, there is no need to worry about what exists only in the
344 working tree. When you have local changes in a part of the project that
345 is not involved in the merge, your changes do not interfere with the
346 merge, and are kept intact. When they do interfere, the merge does not
347 even start (git read-tree complains loudly and fails without modifying
348 anything). In such a case, you can simply continue doing what you were
349 in the middle of doing, and when your working tree is ready (i.e. you
350 have finished your work-in-progress), attempt the merge again.
351
353 "Sparse checkout" allows populating the working directory sparsely. It
354 uses the skip-worktree bit (see git-update-index(1)) to tell Git
355 whether a file in the working directory is worth looking at.
356 git read-tree and other merge-based commands (git merge, git
358 checkout...) can help maintaining the skip-worktree bitmap and working
359 directory update. $GIT_DIR/info/sparse-checkout is used to define the
360 skip-worktree reference bitmap. When git read-tree needs to update the
361 working directory, it resets the skip-worktree bit in the index based
362 on this file, which uses the same syntax as .gitignore files. If an
363 entry matches a pattern in this file, skip-worktree will not be set on
364 that entry. Otherwise, skip-worktree will be set.
365 Then it compares the new skip-worktree value with the previous one. If
367 skip-worktree turns from set to unset, it will add the corresponding
368 file back. If it turns from unset to set, that file will be removed.
369 While $GIT_DIR/info/sparse-checkout is usually used to specify what
371 files are in, you can also specify what files are not in, using negate
372 patterns. For example, to remove the file unwanted:
373 /*
375 !unwanted
376 Another tricky thing is fully repopulating the working directory when
379 you no longer want sparse checkout. You cannot just disable "sparse
380 checkout" because skip-worktree bits are still in the index and your
381 working directory is still sparsely populated. You should re-populate
382 the working directory with the $GIT_DIR/info/sparse-checkout file
383 content as follows:
384 /*
386 Then you can disable sparse checkout. Sparse checkout support in git
389 read-tree and similar commands is disabled by default. You need to turn
390 core.sparseCheckout on in order to have sparse checkout support.
391
393 git-write-tree(1); git-ls-files(1); gitignore(5)
394
396 Part of the git(1) suite
397Git 2.20.1 12/15/2018 GIT-READ-TREE(1)