checkout.txt: note about losing staged changes with --merge
[git/git.git] / Documentation / git-checkout.txt
1 git-checkout(1)
2 ===============
3
4 NAME
5 ----
6 git-checkout - Switch branches or restore working tree files
7
8 SYNOPSIS
9 --------
10 [verse]
11 'git checkout' [-q] [-f] [-m] [<branch>]
12 'git checkout' [-q] [-f] [-m] --detach [<branch>]
13 'git checkout' [-q] [-f] [-m] [--detach] <commit>
14 'git checkout' [-q] [-f] [-m] [[-b|-B|--orphan] <new_branch>] [<start_point>]
15 'git checkout' [-f|--ours|--theirs|-m|--conflict=<style>] [<tree-ish>] [--] <paths>...
16 'git checkout' [<tree-ish>] [--] <pathspec>...
17 'git checkout' (-p|--patch) [<tree-ish>] [--] [<paths>...]
18
19 DESCRIPTION
20 -----------
21 Updates files in the working tree to match the version in the index
22 or the specified tree. If no paths are given, 'git checkout' will
23 also update `HEAD` to set the specified branch as the current
24 branch.
25
26 'git checkout' <branch>::
27 To prepare for working on <branch>, switch to it by updating
28 the index and the files in the working tree, and by pointing
29 HEAD at the branch. Local modifications to the files in the
30 working tree are kept, so that they can be committed to the
31 <branch>.
32 +
33 If <branch> is not found but there does exist a tracking branch in
34 exactly one remote (call it <remote>) with a matching name, treat as
35 equivalent to
36 +
37 ------------
38 $ git checkout -b <branch> --track <remote>/<branch>
39 ------------
40 +
41 If the branch exists in multiple remotes and one of them is named by
42 the `checkout.defaultRemote` configuration variable, we'll use that
43 one for the purposes of disambiguation, even if the `<branch>` isn't
44 unique across all remotes. Set it to
45 e.g. `checkout.defaultRemote=origin` to always checkout remote
46 branches from there if `<branch>` is ambiguous but exists on the
47 'origin' remote. See also `checkout.defaultRemote` in
48 linkgit:git-config[1].
49 +
50 You could omit <branch>, in which case the command degenerates to
51 "check out the current branch", which is a glorified no-op with
52 rather expensive side-effects to show only the tracking information,
53 if exists, for the current branch.
54
55 'git checkout' -b|-B <new_branch> [<start point>]::
56
57 Specifying `-b` causes a new branch to be created as if
58 linkgit:git-branch[1] were called and then checked out. In
59 this case you can use the `--track` or `--no-track` options,
60 which will be passed to 'git branch'. As a convenience,
61 `--track` without `-b` implies branch creation; see the
62 description of `--track` below.
63 +
64 If `-B` is given, <new_branch> is created if it doesn't exist; otherwise, it
65 is reset. This is the transactional equivalent of
66 +
67 ------------
68 $ git branch -f <branch> [<start point>]
69 $ git checkout <branch>
70 ------------
71 +
72 that is to say, the branch is not reset/created unless "git checkout" is
73 successful.
74
75 'git checkout' --detach [<branch>]::
76 'git checkout' [--detach] <commit>::
77
78 Prepare to work on top of <commit>, by detaching HEAD at it
79 (see "DETACHED HEAD" section), and updating the index and the
80 files in the working tree. Local modifications to the files
81 in the working tree are kept, so that the resulting working
82 tree will be the state recorded in the commit plus the local
83 modifications.
84 +
85 When the <commit> argument is a branch name, the `--detach` option can
86 be used to detach HEAD at the tip of the branch (`git checkout
87 <branch>` would check out that branch without detaching HEAD).
88 +
89 Omitting <branch> detaches HEAD at the tip of the current branch.
90
91 'git checkout' [<tree-ish>] [--] <pathspec>...::
92
93 Overwrite paths in the working tree by replacing with the
94 contents in the index or in the <tree-ish> (most often a
95 commit). When a <tree-ish> is given, the paths that
96 match the <pathspec> are updated both in the index and in
97 the working tree.
98 +
99 The index may contain unmerged entries because of a previous failed merge.
100 By default, if you try to check out such an entry from the index, the
101 checkout operation will fail and nothing will be checked out.
102 Using `-f` will ignore these unmerged entries. The contents from a
103 specific side of the merge can be checked out of the index by
104 using `--ours` or `--theirs`. With `-m`, changes made to the working tree
105 file can be discarded to re-create the original conflicted merge result.
106
107 'git checkout' (-p|--patch) [<tree-ish>] [--] [<pathspec>...]::
108 This is similar to the "check out paths to the working tree
109 from either the index or from a tree-ish" mode described
110 above, but lets you use the interactive interface to show
111 the "diff" output and choose which hunks to use in the
112 result. See below for the description of `--patch` option.
113
114
115 OPTIONS
116 -------
117 -q::
118 --quiet::
119 Quiet, suppress feedback messages.
120
121 --[no-]progress::
122 Progress status is reported on the standard error stream
123 by default when it is attached to a terminal, unless `--quiet`
124 is specified. This flag enables progress reporting even if not
125 attached to a terminal, regardless of `--quiet`.
126
127 -f::
128 --force::
129 When switching branches, proceed even if the index or the
130 working tree differs from HEAD. This is used to throw away
131 local changes.
132 +
133 When checking out paths from the index, do not fail upon unmerged
134 entries; instead, unmerged entries are ignored.
135
136 --ours::
137 --theirs::
138 When checking out paths from the index, check out stage #2
139 ('ours') or #3 ('theirs') for unmerged paths.
140 +
141 Note that during `git rebase` and `git pull --rebase`, 'ours' and
142 'theirs' may appear swapped; `--ours` gives the version from the
143 branch the changes are rebased onto, while `--theirs` gives the
144 version from the branch that holds your work that is being rebased.
145 +
146 This is because `rebase` is used in a workflow that treats the
147 history at the remote as the shared canonical one, and treats the
148 work done on the branch you are rebasing as the third-party work to
149 be integrated, and you are temporarily assuming the role of the
150 keeper of the canonical history during the rebase. As the keeper of
151 the canonical history, you need to view the history from the remote
152 as `ours` (i.e. "our shared canonical history"), while what you did
153 on your side branch as `theirs` (i.e. "one contributor's work on top
154 of it").
155
156 -b <new_branch>::
157 Create a new branch named <new_branch> and start it at
158 <start_point>; see linkgit:git-branch[1] for details.
159
160 -B <new_branch>::
161 Creates the branch <new_branch> and start it at <start_point>;
162 if it already exists, then reset it to <start_point>. This is
163 equivalent to running "git branch" with "-f"; see
164 linkgit:git-branch[1] for details.
165
166 -t::
167 --track::
168 When creating a new branch, set up "upstream" configuration. See
169 "--track" in linkgit:git-branch[1] for details.
170 +
171 If no `-b` option is given, the name of the new branch will be
172 derived from the remote-tracking branch, by looking at the local part of
173 the refspec configured for the corresponding remote, and then stripping
174 the initial part up to the "*".
175 This would tell us to use "hack" as the local branch when branching
176 off of "origin/hack" (or "remotes/origin/hack", or even
177 "refs/remotes/origin/hack"). If the given name has no slash, or the above
178 guessing results in an empty name, the guessing is aborted. You can
179 explicitly give a name with `-b` in such a case.
180
181 --no-track::
182 Do not set up "upstream" configuration, even if the
183 branch.autoSetupMerge configuration variable is true.
184
185 -l::
186 Create the new branch's reflog; see linkgit:git-branch[1] for
187 details.
188
189 --detach::
190 Rather than checking out a branch to work on it, check out a
191 commit for inspection and discardable experiments.
192 This is the default behavior of "git checkout <commit>" when
193 <commit> is not a branch name. See the "DETACHED HEAD" section
194 below for details.
195
196 --orphan <new_branch>::
197 Create a new 'orphan' branch, named <new_branch>, started from
198 <start_point> and switch to it. The first commit made on this
199 new branch will have no parents and it will be the root of a new
200 history totally disconnected from all the other branches and
201 commits.
202 +
203 The index and the working tree are adjusted as if you had previously run
204 "git checkout <start_point>". This allows you to start a new history
205 that records a set of paths similar to <start_point> by easily running
206 "git commit -a" to make the root commit.
207 +
208 This can be useful when you want to publish the tree from a commit
209 without exposing its full history. You might want to do this to publish
210 an open source branch of a project whose current tree is "clean", but
211 whose full history contains proprietary or otherwise encumbered bits of
212 code.
213 +
214 If you want to start a disconnected history that records a set of paths
215 that is totally different from the one of <start_point>, then you should
216 clear the index and the working tree right after creating the orphan
217 branch by running "git rm -rf ." from the top level of the working tree.
218 Afterwards you will be ready to prepare your new files, repopulating the
219 working tree, by copying them from elsewhere, extracting a tarball, etc.
220
221 --ignore-skip-worktree-bits::
222 In sparse checkout mode, `git checkout -- <paths>` would
223 update only entries matched by <paths> and sparse patterns
224 in $GIT_DIR/info/sparse-checkout. This option ignores
225 the sparse patterns and adds back any files in <paths>.
226
227 -m::
228 --merge::
229 When switching branches,
230 if you have local modifications to one or more files that
231 are different between the current branch and the branch to
232 which you are switching, the command refuses to switch
233 branches in order to preserve your modifications in context.
234 However, with this option, a three-way merge between the current
235 branch, your working tree contents, and the new branch
236 is done, and you will be on the new branch.
237 +
238 When a merge conflict happens, the index entries for conflicting
239 paths are left unmerged, and you need to resolve the conflicts
240 and mark the resolved paths with `git add` (or `git rm` if the merge
241 should result in deletion of the path).
242 +
243 When checking out paths from the index, this option lets you recreate
244 the conflicted merge in the specified paths.
245 +
246 When switching branches with `--merge`, staged changes may be lost.
247
248 --conflict=<style>::
249 The same as --merge option above, but changes the way the
250 conflicting hunks are presented, overriding the
251 merge.conflictStyle configuration variable. Possible values are
252 "merge" (default) and "diff3" (in addition to what is shown by
253 "merge" style, shows the original contents).
254
255 -p::
256 --patch::
257 Interactively select hunks in the difference between the
258 <tree-ish> (or the index, if unspecified) and the working
259 tree. The chosen hunks are then applied in reverse to the
260 working tree (and if a <tree-ish> was specified, the index).
261 +
262 This means that you can use `git checkout -p` to selectively discard
263 edits from your current working tree. See the ``Interactive Mode''
264 section of linkgit:git-add[1] to learn how to operate the `--patch` mode.
265
266 --ignore-other-worktrees::
267 `git checkout` refuses when the wanted ref is already checked
268 out by another worktree. This option makes it check the ref
269 out anyway. In other words, the ref can be held by more than one
270 worktree.
271
272 --[no-]recurse-submodules::
273 Using --recurse-submodules will update the content of all initialized
274 submodules according to the commit recorded in the superproject. If
275 local modifications in a submodule would be overwritten the checkout
276 will fail unless `-f` is used. If nothing (or --no-recurse-submodules)
277 is used, the work trees of submodules will not be updated.
278 Just like linkgit:git-submodule[1], this will detach the
279 submodules HEAD.
280
281 --no-guess::
282 Do not attempt to create a branch if a remote tracking branch
283 of the same name exists.
284
285 <branch>::
286 Branch to checkout; if it refers to a branch (i.e., a name that,
287 when prepended with "refs/heads/", is a valid ref), then that
288 branch is checked out. Otherwise, if it refers to a valid
289 commit, your HEAD becomes "detached" and you are no longer on
290 any branch (see below for details).
291 +
292 You can use the `"@{-N}"` syntax to refer to the N-th last
293 branch/commit checked out using "git checkout" operation. You may
294 also specify `-` which is synonymous to `"@{-1}"`.
295 +
296 As a special case, you may use `"A...B"` as a shortcut for the
297 merge base of `A` and `B` if there is exactly one merge base. You can
298 leave out at most one of `A` and `B`, in which case it defaults to `HEAD`.
299
300 <new_branch>::
301 Name for the new branch.
302
303 <start_point>::
304 The name of a commit at which to start the new branch; see
305 linkgit:git-branch[1] for details. Defaults to HEAD.
306
307 <tree-ish>::
308 Tree to checkout from (when paths are given). If not specified,
309 the index will be used.
310
311
312
313 DETACHED HEAD
314 -------------
315 HEAD normally refers to a named branch (e.g. 'master'). Meanwhile, each
316 branch refers to a specific commit. Let's look at a repo with three
317 commits, one of them tagged, and with branch 'master' checked out:
318
319 ------------
320 HEAD (refers to branch 'master')
321 |
322 v
323 a---b---c branch 'master' (refers to commit 'c')
324 ^
325 |
326 tag 'v2.0' (refers to commit 'b')
327 ------------
328
329 When a commit is created in this state, the branch is updated to refer to
330 the new commit. Specifically, 'git commit' creates a new commit 'd', whose
331 parent is commit 'c', and then updates branch 'master' to refer to new
332 commit 'd'. HEAD still refers to branch 'master' and so indirectly now refers
333 to commit 'd':
334
335 ------------
336 $ edit; git add; git commit
337
338 HEAD (refers to branch 'master')
339 |
340 v
341 a---b---c---d branch 'master' (refers to commit 'd')
342 ^
343 |
344 tag 'v2.0' (refers to commit 'b')
345 ------------
346
347 It is sometimes useful to be able to checkout a commit that is not at
348 the tip of any named branch, or even to create a new commit that is not
349 referenced by a named branch. Let's look at what happens when we
350 checkout commit 'b' (here we show two ways this may be done):
351
352 ------------
353 $ git checkout v2.0 # or
354 $ git checkout master^^
355
356 HEAD (refers to commit 'b')
357 |
358 v
359 a---b---c---d branch 'master' (refers to commit 'd')
360 ^
361 |
362 tag 'v2.0' (refers to commit 'b')
363 ------------
364
365 Notice that regardless of which checkout command we use, HEAD now refers
366 directly to commit 'b'. This is known as being in detached HEAD state.
367 It means simply that HEAD refers to a specific commit, as opposed to
368 referring to a named branch. Let's see what happens when we create a commit:
369
370 ------------
371 $ edit; git add; git commit
372
373 HEAD (refers to commit 'e')
374 |
375 v
376 e
377 /
378 a---b---c---d branch 'master' (refers to commit 'd')
379 ^
380 |
381 tag 'v2.0' (refers to commit 'b')
382 ------------
383
384 There is now a new commit 'e', but it is referenced only by HEAD. We can
385 of course add yet another commit in this state:
386
387 ------------
388 $ edit; git add; git commit
389
390 HEAD (refers to commit 'f')
391 |
392 v
393 e---f
394 /
395 a---b---c---d branch 'master' (refers to commit 'd')
396 ^
397 |
398 tag 'v2.0' (refers to commit 'b')
399 ------------
400
401 In fact, we can perform all the normal Git operations. But, let's look
402 at what happens when we then checkout master:
403
404 ------------
405 $ git checkout master
406
407 HEAD (refers to branch 'master')
408 e---f |
409 / v
410 a---b---c---d branch 'master' (refers to commit 'd')
411 ^
412 |
413 tag 'v2.0' (refers to commit 'b')
414 ------------
415
416 It is important to realize that at this point nothing refers to commit
417 'f'. Eventually commit 'f' (and by extension commit 'e') will be deleted
418 by the routine Git garbage collection process, unless we create a reference
419 before that happens. If we have not yet moved away from commit 'f',
420 any of these will create a reference to it:
421
422 ------------
423 $ git checkout -b foo <1>
424 $ git branch foo <2>
425 $ git tag foo <3>
426 ------------
427
428 <1> creates a new branch 'foo', which refers to commit 'f', and then
429 updates HEAD to refer to branch 'foo'. In other words, we'll no longer
430 be in detached HEAD state after this command.
431
432 <2> similarly creates a new branch 'foo', which refers to commit 'f',
433 but leaves HEAD detached.
434
435 <3> creates a new tag 'foo', which refers to commit 'f',
436 leaving HEAD detached.
437
438 If we have moved away from commit 'f', then we must first recover its object
439 name (typically by using git reflog), and then we can create a reference to
440 it. For example, to see the last two commits to which HEAD referred, we
441 can use either of these commands:
442
443 ------------
444 $ git reflog -2 HEAD # or
445 $ git log -g -2 HEAD
446 ------------
447
448 ARGUMENT DISAMBIGUATION
449 -----------------------
450
451 When there is only one argument given and it is not `--` (e.g. "git
452 checkout abc"), and when the argument is both a valid `<tree-ish>`
453 (e.g. a branch "abc" exists) and a valid `<pathspec>` (e.g. a file
454 or a directory whose name is "abc" exists), Git would usually ask
455 you to disambiguate. Because checking out a branch is so common an
456 operation, however, "git checkout abc" takes "abc" as a `<tree-ish>`
457 in such a situation. Use `git checkout -- <pathspec>` if you want
458 to checkout these paths out of the index.
459
460 EXAMPLES
461 --------
462
463 . The following sequence checks out the `master` branch, reverts
464 the `Makefile` to two revisions back, deletes hello.c by
465 mistake, and gets it back from the index.
466 +
467 ------------
468 $ git checkout master <1>
469 $ git checkout master~2 Makefile <2>
470 $ rm -f hello.c
471 $ git checkout hello.c <3>
472 ------------
473 +
474 <1> switch branch
475 <2> take a file out of another commit
476 <3> restore hello.c from the index
477 +
478 If you want to check out _all_ C source files out of the index,
479 you can say
480 +
481 ------------
482 $ git checkout -- '*.c'
483 ------------
484 +
485 Note the quotes around `*.c`. The file `hello.c` will also be
486 checked out, even though it is no longer in the working tree,
487 because the file globbing is used to match entries in the index
488 (not in the working tree by the shell).
489 +
490 If you have an unfortunate branch that is named `hello.c`, this
491 step would be confused as an instruction to switch to that branch.
492 You should instead write:
493 +
494 ------------
495 $ git checkout -- hello.c
496 ------------
497
498 . After working in the wrong branch, switching to the correct
499 branch would be done using:
500 +
501 ------------
502 $ git checkout mytopic
503 ------------
504 +
505 However, your "wrong" branch and correct "mytopic" branch may
506 differ in files that you have modified locally, in which case
507 the above checkout would fail like this:
508 +
509 ------------
510 $ git checkout mytopic
511 error: You have local changes to 'frotz'; not switching branches.
512 ------------
513 +
514 You can give the `-m` flag to the command, which would try a
515 three-way merge:
516 +
517 ------------
518 $ git checkout -m mytopic
519 Auto-merging frotz
520 ------------
521 +
522 After this three-way merge, the local modifications are _not_
523 registered in your index file, so `git diff` would show you what
524 changes you made since the tip of the new branch.
525
526 . When a merge conflict happens during switching branches with
527 the `-m` option, you would see something like this:
528 +
529 ------------
530 $ git checkout -m mytopic
531 Auto-merging frotz
532 ERROR: Merge conflict in frotz
533 fatal: merge program failed
534 ------------
535 +
536 At this point, `git diff` shows the changes cleanly merged as in
537 the previous example, as well as the changes in the conflicted
538 files. Edit and resolve the conflict and mark it resolved with
539 `git add` as usual:
540 +
541 ------------
542 $ edit frotz
543 $ git add frotz
544 ------------
545
546 GIT
547 ---
548 Part of the linkgit:git[1] suite