1GIT-RESET(1) Git Manual GIT-RESET(1)
2
6 git-reset - Reset current HEAD to the specified state
7
9 git reset [-q] [<tree-ish>] [--] <pathspec>...
10 git reset [-q] [--pathspec-from-file=<file> [--pathspec-file-nul]] [<tree-ish>]
11 git reset (--patch | -p) [<tree-ish>] [--] [<pathspec>...]
12 git reset [--soft | --mixed [-N] | --hard | --merge | --keep] [-q] [<commit>]
13
15 In the first three forms, copy entries from <tree-ish> to the index. In
16 the last form, set the current branch head (HEAD) to <commit>,
17 optionally modifying index and working tree to match. The
18 <tree-ish>/<commit> defaults to HEAD in all forms.
19 git reset [-q] [<tree-ish>] [--] <pathspec>..., git reset [-q]
21 [--pathspec-from-file=<file> [--pathspec-file-nul]] [<tree-ish>]
22 These forms reset the index entries for all paths that match the
23 <pathspec> to their state at <tree-ish>. (It does not affect the
24 working tree or the current branch.)
25 This means that git reset <pathspec> is the opposite of git add
27 <pathspec>. This command is equivalent to git restore
28 [--source=<tree-ish>] --staged <pathspec>....
29 After running git reset <pathspec> to update the index entry, you
31 can use git-restore(1) to check the contents out of the index to
32 the working tree. Alternatively, using git-restore(1) and
33 specifying a commit with --source, you can copy the contents of a
34 path out of a commit to the index and to the working tree in one
35 go.
36 git reset (--patch | -p) [<tree-ish>] [--] [<pathspec>...]
38 Interactively select hunks in the difference between the index and
39 <tree-ish> (defaults to HEAD). The chosen hunks are applied in
40 reverse to the index.
41 This means that git reset -p is the opposite of git add -p, i.e.
43 you can use it to selectively reset hunks. See the “Interactive
44 Mode” section of git-add(1) to learn how to operate the --patch
45 mode.
46 git reset [<mode>] [<commit>]
48 This form resets the current branch head to <commit> and possibly
49 updates the index (resetting it to the tree of <commit>) and the
50 working tree depending on <mode>. If <mode> is omitted, defaults to
51 --mixed. The <mode> must be one of the following:
52 --soft
54 Does not touch the index file or the working tree at all (but
55 resets the head to <commit>, just like all modes do). This
56 leaves all your changed files "Changes to be committed", as git
57 status would put it.
58 --mixed
60 Resets the index but not the working tree (i.e., the changed
61 files are preserved but not marked for commit) and reports what
62 has not been updated. This is the default action.
63 If -N is specified, removed paths are marked as intent-to-add
65 (see git-add(1)).
66 --hard
68 Resets the index and working tree. Any changes to tracked files
69 in the working tree since <commit> are discarded.
70 --merge
72 Resets the index and updates the files in the working tree that
73 are different between <commit> and HEAD, but keeps those which
74 are different between the index and working tree (i.e. which
75 have changes which have not been added). If a file that is
76 different between <commit> and the index has unstaged changes,
77 reset is aborted.
78 In other words, --merge does something like a git read-tree -u
80 -m <commit>, but carries forward unmerged index entries.
81 --keep
83 Resets index entries and updates files in the working tree that
84 are different between <commit> and HEAD. If a file that is
85 different between <commit> and HEAD has local changes, reset is
86 aborted.
87 See "Reset, restore and revert" in git(1) for the differences between
89 the three commands.
90
92 -q, --quiet, --no-quiet
93 Be quiet, only report errors. The default behavior is set by the
94 reset.quiet config option. --quiet and --no-quiet will override
95 the default behavior.
96 --pathspec-from-file=<file>
98 Pathspec is passed in <file> instead of commandline args. If <file>
99 is exactly - then standard input is used. Pathspec elements are
100 separated by LF or CR/LF. Pathspec elements can be quoted as
101 explained for the configuration variable core.quotePath (see git-
102 config(1)). See also --pathspec-file-nul and global
103 --literal-pathspecs.
104 --pathspec-file-nul
106 Only meaningful with --pathspec-from-file. Pathspec elements are
107 separated with NUL character and all other characters are taken
108 literally (including newlines and quotes).
109 --
111 Do not interpret any more arguments as options.
112 <pathspec>...
114 Limits the paths affected by the operation.
115 For more details, see the pathspec entry in gitglossary(7).
117
119 Undo add
120 $ edit [1m(1)
122 $ git add frotz.c filfre.c
123 $ mailx [1m(2)
124 $ git reset [1m(3)
125 $ git pull git://info.example.com/ nitfol [1m(4)
126 1. You are happily working on something, and find the changes
128 in these files are in good order. You do not want to see
129 them when you run git diff, because you plan to work on
130 other files and changes with these files are distracting.
131 2. Somebody asks you to pull, and the changes sound worthy of
134 merging.
135 3. However, you already dirtied the index (i.e. your index
136 does not match the HEAD commit). But you know the pull you
137 are going to make does not affect frotz.c or filfre.c, so
138 you revert the index changes for these two files. Your
139 changes in working tree remain there.
140 4. Then you can pull and merge, leaving frotz.c and filfre.c
141 changes still in the working tree.
142 Undo a commit and redo
144 $ git commit ...
146 $ git reset --soft HEAD^ [1m(1)
147 $ edit [1m(2)
148 $ git commit -a -c ORIG_HEAD [1m(3)
149 1. This is most often done when you remembered what you just
152 committed is incomplete, or you misspelled your commit
153 message, or both. Leaves working tree as it was before
154 "reset".
155 2. Make corrections to working tree files.
156 3. "reset" copies the old head to .git/ORIG_HEAD; redo the
157 commit by starting with its log message. If you do not
158 need to edit the message further, you can give -C option
159 instead.
160 See also the --amend option to git-commit(1).
161 Undo a commit, making it a topic branch
163 $ git branch topic/wip [1m(1)
165 $ git reset --hard HEAD~3 [1m(2)
166 $ git switch topic/wip [1m(3)
167 1. You have made some commits, but realize they were
169 premature to be in the master branch. You want to continue
170 polishing them in a topic branch, so create topic/wip
171 branch off of the current HEAD.
172 2. Rewind the master branch to get rid of those three
173 commits.
174 3. Switch to topic/wip branch and keep working.
175 Undo commits permanently
177 $ git commit ...
179 $ git reset --hard HEAD~3 [1m(1)
180 1. The last three commits (HEAD, HEAD^, and HEAD~2) were bad
182 and you do not want to ever see them again. Do not do this
183 if you have already given these commits to somebody else.
184 (See the "RECOVERING FROM UPSTREAM REBASE" section in git-
185 rebase(1) for the implications of doing so.)
186 Undo a merge or pull
188 $ git pull [1m(1)
190 Auto-merging nitfol
191 CONFLICT (content): Merge conflict in nitfol
192 Automatic merge failed; fix conflicts and then commit the result.
193 $ git reset --hard [1m(2)
194 $ git pull . topic/branch [1m(3)
195 Updating from 41223... to 13134...
196 Fast-forward
197 $ git reset --hard ORIG_HEAD [1m(4)
198 1. Try to update from the upstream resulted in a lot of
200 conflicts; you were not ready to spend a lot of time
201 merging right now, so you decide to do that later.
202 2. "pull" has not made merge commit, so git reset --hard
203 which is a synonym for git reset --hard HEAD clears the
204 mess from the index file and the working tree.
205 3. Merge a topic branch into the current branch, which
206 resulted in a fast-forward.
207 4. But you decided that the topic branch is not ready for
208 public consumption yet. "pull" or "merge" always leaves
209 the original tip of the current branch in ORIG_HEAD, so
210 resetting hard to it brings your index file and the
211 working tree back to that state, and resets the tip of the
212 branch to that commit.
213 Undo a merge or pull inside a dirty working tree
215 $ git pull [1m(1)
217 Auto-merging nitfol
218 Merge made by recursive.
219 nitfol | 20 +++++----
220 ...
221 $ git reset --merge ORIG_HEAD [1m(2)
222 1. Even if you may have local modifications in your working
224 tree, you can safely say git pull when you know that the
225 change in the other branch does not overlap with them.
226 2. After inspecting the result of the merge, you may find
227 that the change in the other branch is unsatisfactory.
228 Running git reset --hard ORIG_HEAD will let you go back to
229 where you were, but it will discard your local changes,
230 which you do not want. git reset --merge keeps your local
231 changes.
232 Interrupted workflow
234 Suppose you are interrupted by an urgent fix request while you are
235 in the middle of a large change. The files in your working tree are
236 not in any shape to be committed yet, but you need to get to the
237 other branch for a quick bugfix.
238 $ git switch feature ;# you were working in "feature" branch and
240 $ work work work ;# got interrupted
241 $ git commit -a -m "snapshot WIP" [1m(1)
242 $ git switch master
243 $ fix fix fix
244 $ git commit ;# commit with real log
245 $ git switch feature
246 $ git reset --soft HEAD^ ;# go back to WIP state [1m(2)
247 $ git reset [1m(3)
248 1. This commit will get blown away so a throw-away log
251 message is OK.
252 2. This removes the WIP commit from the commit history, and
253 sets your working tree to the state just before you made
254 that snapshot.
255 3. At this point the index file still has all the WIP changes
256 you committed as snapshot WIP. This updates the index to
257 show your WIP files as uncommitted.
258 See also git-stash(1).
259 Reset a single file in the index
261 Suppose you have added a file to your index, but later decide you
262 do not want to add it to your commit. You can remove the file from
263 the index while keeping your changes with git reset.
264 $ git reset -- frotz.c [1m(1)
266 $ git commit -m "Commit files in index" [1m(2)
267 $ git add frotz.c [1m(3)
268 1. This removes the file from the index while keeping it in
270 the working directory.
271 2. This commits all other changes in the index.
272 3. Adds the file to the index again.
273 Keep changes in working tree while discarding some previous commits
275 Suppose you are working on something and you commit it, and then
276 you continue working a bit more, but now you think that what you
277 have in your working tree should be in another branch that has
278 nothing to do with what you committed previously. You can start a
279 new branch and reset it while keeping the changes in your working
280 tree.
281 $ git tag start
283 $ git switch -c branch1
284 $ edit
285 $ git commit ... [1m(1)
286 $ edit
287 $ git switch -c branch2 [1m(2)
288 $ git reset --keep start [1m(3)
289 1. This commits your first edits in branch1.
291 2. In the ideal world, you could have realized that the
292 earlier commit did not belong to the new topic when you
293 created and switched to branch2 (i.e. git switch -c
294 branch2 start), but nobody is perfect.
295 3. But you can use reset --keep to remove the unwanted commit
296 after you switched to branch2.
297 Split a commit apart into a sequence of commits
299 Suppose that you have created lots of logically separate changes
300 and committed them together. Then, later you decide that it might
301 be better to have each logical chunk associated with its own
302 commit. You can use git reset to rewind history without changing
303 the contents of your local files, and then successively use git add
304 -p to interactively select which hunks to include into each commit,
305 using git commit -c to pre-populate the commit message.
306 $ git reset -N HEAD^ [1m(1)
308 $ git add -p [1m(2)
309 $ git diff --cached [1m(3)
310 $ git commit -c HEAD@{1} [1m(4)
311 ... [1m(5)
312 $ git add ... [1m(6)
313 $ git diff --cached [1m(7)
314 $ git commit ... [1m(8)
315 1. First, reset the history back one commit so that we remove
317 the original commit, but leave the working tree with all
318 the changes. The -N ensures that any new files added with
319 HEAD are still marked so that git add -p will find them.
320 2. Next, we interactively select diff hunks to add using the
321 git add -p facility. This will ask you about each diff
322 hunk in sequence and you can use simple commands such as
323 "yes, include this", "No don’t include this" or even the
324 very powerful "edit" facility.
325 3. Once satisfied with the hunks you want to include, you
326 should verify what has been prepared for the first commit
327 by using git diff --cached. This shows all the changes
328 that have been moved into the index and are about to be
329 committed.
330 4. Next, commit the changes stored in the index. The -c
332 option specifies to pre-populate the commit message from
333 the original message that you started with in the first
334 commit. This is helpful to avoid retyping it. The HEAD@{1}
335 is a special notation for the commit that HEAD used to be
336 at prior to the original reset commit (1 change ago). See
337 git-reflog(1) for more details. You may also use any other
338 valid commit reference.
339 5. You can repeat steps 2-4 multiple times to break the
340 original code into any number of commits.
341 6. Now you’ve split out many of the changes into their own
342 commits, and might no longer use the patch mode of git
343 add, in order to select all remaining uncommitted changes.
344 7. Once again, check to verify that you’ve included what you
345 want to. You may also wish to verify that git diff doesn’t
346 show any remaining changes to be committed later.
347 8. And finally create the final commit.
348
350 The tables below show what happens when running:
351 git reset --option target
353 to reset the HEAD to another commit (target) with the different reset
355 options depending on the state of the files.
356 In these tables, A, B, C and D are some different states of a file. For
358 example, the first line of the first table means that if a file is in
359 state A in the working tree, in state B in the index, in state C in
360 HEAD and in state D in the target, then git reset --soft target will
361 leave the file in the working tree in state A and in the index in state
362 B. It resets (i.e. moves) the HEAD (i.e. the tip of the current branch,
363 if you are on one) to target (which has the file in state D).
364 working index HEAD target working index HEAD
366 ----------------------------------------------------
367 A B C D --soft A B D
368 --mixed A D D
369 --hard D D D
370 --merge (disallowed)
371 --keep (disallowed)
372 working index HEAD target working index HEAD
374 ----------------------------------------------------
375 A B C C --soft A B C
376 --mixed A C C
377 --hard C C C
378 --merge (disallowed)
379 --keep A C C
380 working index HEAD target working index HEAD
382 ----------------------------------------------------
383 B B C D --soft B B D
384 --mixed B D D
385 --hard D D D
386 --merge D D D
387 --keep (disallowed)
388 working index HEAD target working index HEAD
390 ----------------------------------------------------
391 B B C C --soft B B C
392 --mixed B C C
393 --hard C C C
394 --merge C C C
395 --keep B C C
396 working index HEAD target working index HEAD
398 ----------------------------------------------------
399 B C C D --soft B C D
400 --mixed B D D
401 --hard D D D
402 --merge (disallowed)
403 --keep (disallowed)
404 working index HEAD target working index HEAD
406 ----------------------------------------------------
407 B C C C --soft B C C
408 --mixed B C C
409 --hard C C C
410 --merge B C C
411 --keep B C C
412 reset --merge is meant to be used when resetting out of a conflicted
414 merge. Any mergy operation guarantees that the working tree file that
415 is involved in the merge does not have a local change with respect to
416 the index before it starts, and that it writes the result out to the
417 working tree. So if we see some difference between the index and the
418 target and also between the index and the working tree, then it means
419 that we are not resetting out from a state that a mergy operation left
420 after failing with a conflict. That is why we disallow --merge option
421 in this case.
422 reset --keep is meant to be used when removing some of the last commits
424 in the current branch while keeping changes in the working tree. If
425 there could be conflicts between the changes in the commit we want to
426 remove and the changes in the working tree we want to keep, the reset
427 is disallowed. That’s why it is disallowed if there are both changes
428 between the working tree and HEAD, and between HEAD and the target. To
429 be safe, it is also disallowed when there are unmerged entries.
430 The following tables show what happens when there are unmerged entries:
432 working index HEAD target working index HEAD
434 ----------------------------------------------------
435 X U A B --soft (disallowed)
436 --mixed X B B
437 --hard B B B
438 --merge B B B
439 --keep (disallowed)
440 working index HEAD target working index HEAD
442 ----------------------------------------------------
443 X U A A --soft (disallowed)
444 --mixed X A A
445 --hard A A A
446 --merge A A A
447 --keep (disallowed)
448 X means any state and U means an unmerged index.
450
452 Part of the git(1) suite
453Git 2.26.2 2020-04-20 GIT-RESET(1)