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>] [--] <paths>...
10 git reset (--patch | -p) [<tree-ish>] [--] [<paths>...]
11 git reset [--soft | --mixed [-N] | --hard | --merge | --keep] [-q] [<commit>]
12
15 In the first and second form, copy entries from <tree-ish> to the
16 index. In the third form, set the current branch head (HEAD) to
17 <commit>, 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>] [--] <paths>...
21 This form resets the index entries for all <paths> to their state
22 at <tree-ish>. (It does not affect the working tree or the current
23 branch.)
24 This means that git reset <paths> is the opposite of git add
26 <paths>.
27 After running git reset <paths> to update the index entry, you can
29 use git-checkout(1) to check the contents out of the index to the
30 working tree. Alternatively, using git-checkout(1) and specifying a
31 commit, you can copy the contents of a path out of a commit to the
32 index and to the working tree in one go.
33 git reset (--patch | -p) [<tree-ish>] [--] [<paths>...]
35 Interactively select hunks in the difference between the index and
36 <tree-ish> (defaults to HEAD). The chosen hunks are applied in
37 reverse to the index.
38 This means that git reset -p is the opposite of git add -p, i.e.
40 you can use it to selectively reset hunks. See the “Interactive
41 Mode” section of git-add(1) to learn how to operate the --patch
42 mode.
43 git reset [<mode>] [<commit>]
45 This form resets the current branch head to <commit> and possibly
46 updates the index (resetting it to the tree of <commit>) and the
47 working tree depending on <mode>. If <mode> is omitted, defaults to
48 "--mixed". The <mode> must be one of the following:
49 --soft
51 Does not touch the index file or the working tree at all (but
52 resets the head to <commit>, just like all modes do). This
53 leaves all your changed files "Changes to be committed", as git
54 status would put it.
55 --mixed
57 Resets the index but not the working tree (i.e., the changed
58 files are preserved but not marked for commit) and reports what
59 has not been updated. This is the default action.
60 If -N is specified, removed paths are marked as intent-to-add
62 (see git-add(1)).
63 --hard
65 Resets the index and working tree. Any changes to tracked files
66 in the working tree since <commit> are discarded.
67 --merge
69 Resets the index and updates the files in the working tree that
70 are different between <commit> and HEAD, but keeps those which
71 are different between the index and working tree (i.e. which
72 have changes which have not been added). If a file that is
73 different between <commit> and the index has unstaged changes,
74 reset is aborted.
75 In other words, --merge does something like a git read-tree -u
77 -m <commit>, but carries forward unmerged index entries.
78 --keep
80 Resets index entries and updates files in the working tree that
81 are different between <commit> and HEAD. If a file that is
82 different between <commit> and HEAD has local changes, reset is
83 aborted.
84 If you want to undo a commit other than the latest on a branch, git-
86 revert(1) is your friend.
87
89 -q, --quiet
90 Be quiet, only report errors.
91
93 Undo add
94 $ edit [1m(1)
96 $ git add frotz.c filfre.c
97 $ mailx [1m(2)
98 $ git reset [1m(3)
99 $ git pull git://info.example.com/ nitfol [1m(4)
100 1. You are happily working on something, and find the changes in
102 these files are in good order. You do not want to see them when you
103 run "git diff", because you plan to work on other files and changes
104 with these files are distracting.
105 2. Somebody asks you to pull, and the changes sound worthy of
106 merging.
107 3. However, you already dirtied the index (i.e. your index does not
108 match the HEAD commit). But you know the pull you are going to make
109 does not affect frotz.c or filfre.c, so you revert the index
110 changes for these two files. Your changes in working tree remain
111 there.
112 4. Then you can pull and merge, leaving frotz.c and filfre.c
113 changes still in the working tree.
114 Undo a commit and redo
116 $ git commit ...
118 $ git reset --soft HEAD^ [1m(1)
119 $ edit [1m(2)
120 $ git commit -a -c ORIG_HEAD [1m(3)
121 1. This is most often done when you remembered what you just
123 committed is incomplete, or you misspelled your commit message, or
124 both. Leaves working tree as it was before "reset".
125 2. Make corrections to working tree files.
126 3. "reset" copies the old head to .git/ORIG_HEAD; redo the commit
127 by starting with its log message. If you do not need to edit the
128 message further, you can give -C option instead.
129 See also the --amend option to git-commit(1).
131 Undo a commit, making it a topic branch
133 $ git branch topic/wip [1m(1)
135 $ git reset --hard HEAD~3 [1m(2)
136 $ git checkout topic/wip [1m(3)
137 1. You have made some commits, but realize they were premature to
139 be in the "master" branch. You want to continue polishing them in a
140 topic branch, so create "topic/wip" branch off of the current HEAD.
141 2. Rewind the master branch to get rid of those three commits.
142 3. Switch to "topic/wip" branch and keep working.
143 Undo commits permanently
145 $ git commit ...
147 $ git reset --hard HEAD~3 [1m(1)
148 1. The last three commits (HEAD, HEAD^, and HEAD~2) were bad and
150 you do not want to ever see them again. Do not do this if you have
151 already given these commits to somebody else. (See the "RECOVERING
152 FROM UPSTREAM REBASE" section in git-rebase(1) for the implications
153 of doing so.)
154 Undo a merge or pull
156 $ git pull [1m(1)
158 Auto-merging nitfol
159 CONFLICT (content): Merge conflict in nitfol
160 Automatic merge failed; fix conflicts and then commit the result.
161 $ git reset --hard [1m(2)
162 $ git pull . topic/branch [1m(3)
163 Updating from 41223... to 13134...
164 Fast-forward
165 $ git reset --hard ORIG_HEAD [1m(4)
166 1. Try to update from the upstream resulted in a lot of conflicts;
168 you were not ready to spend a lot of time merging right now, so you
169 decide to do that later.
170 2. "pull" has not made merge commit, so "git reset --hard" which is
171 a synonym for "git reset --hard HEAD" clears the mess from the
172 index file and the working tree.
173 3. Merge a topic branch into the current branch, which resulted in
174 a fast-forward.
175 4. But you decided that the topic branch is not ready for public
176 consumption yet. "pull" or "merge" always leaves the original tip
177 of the current branch in ORIG_HEAD, so resetting hard to it brings
178 your index file and the working tree back to that state, and resets
179 the tip of the branch to that commit.
180 Undo a merge or pull inside a dirty working tree
182 $ git pull [1m(1)
184 Auto-merging nitfol
185 Merge made by recursive.
186 nitfol | 20 +++++----
187 ...
188 $ git reset --merge ORIG_HEAD [1m(2)
189 1. Even if you may have local modifications in your working tree,
191 you can safely say "git pull" when you know that the change in the
192 other branch does not overlap with them.
193 2. After inspecting the result of the merge, you may find that the
194 change in the other branch is unsatisfactory. Running "git reset
195 --hard ORIG_HEAD" will let you go back to where you were, but it
196 will discard your local changes, which you do not want. "git reset
197 --merge" keeps your local changes.
198 Interrupted workflow
200 Suppose you are interrupted by an urgent fix request while you are
201 in the middle of a large change. The files in your working tree are
202 not in any shape to be committed yet, but you need to get to the
203 other branch for a quick bugfix.
204 $ git checkout feature ;# you were working in "feature" branch and
206 $ work work work ;# got interrupted
207 $ git commit -a -m "snapshot WIP" [1m(1)
208 $ git checkout master
209 $ fix fix fix
210 $ git commit ;# commit with real log
211 $ git checkout feature
212 $ git reset --soft HEAD^ ;# go back to WIP state [1m(2)
213 $ git reset [1m(3)
214 1. This commit will get blown away so a throw-away log message is
216 OK.
217 2. This removes the WIP commit from the commit history, and sets
218 your working tree to the state just before you made that snapshot.
219 3. At this point the index file still has all the WIP changes you
220 committed as snapshot WIP. This updates the index to show your WIP
221 files as uncommitted.
222 See also git-stash(1).
224 Reset a single file in the index
226 Suppose you have added a file to your index, but later decide you
227 do not want to add it to your commit. You can remove the file from
228 the index while keeping your changes with git reset.
229 $ git reset -- frotz.c [1m(1)
231 $ git commit -m "Commit files in index" [1m(2)
232 $ git add frotz.c [1m(3)
233 1. This removes the file from the index while keeping it in the
235 working directory.
236 2. This commits all other changes in the index.
237 3. Adds the file to the index again.
238 Keep changes in working tree while discarding some previous commits
240 Suppose you are working on something and you commit it, and then
241 you continue working a bit more, but now you think that what you
242 have in your working tree should be in another branch that has
243 nothing to do with what you committed previously. You can start a
244 new branch and reset it while keeping the changes in your working
245 tree.
246 $ git tag start
248 $ git checkout -b branch1
249 $ edit
250 $ git commit ... [1m(1)
251 $ edit
252 $ git checkout -b branch2 [1m(2)
253 $ git reset --keep start [1m(3)
254 1. This commits your first edits in branch1.
256 2. In the ideal world, you could have realized that the earlier
257 commit did not belong to the new topic when you created and
258 switched to branch2 (i.e. "git checkout -b branch2 start"), but
259 nobody is perfect.
260 3. But you can use "reset --keep" to remove the unwanted commit
261 after you switched to "branch2".
262 Split a commit apart into a sequence of commits
264 Suppose that you have created lots of logically separate changes
265 and committed them together. Then, later you decide that it might
266 be better to have each logical chunk associated with its own
267 commit. You can use git reset to rewind history without changing
268 the contents of your local files, and then successively use git add
269 -p to interactively select which hunks to include into each commit,
270 using git commit -c to pre-populate the commit message.
271 $ git reset -N HEAD^ [1m(1)
273 $ git add -p [1m(2)
274 $ git diff --cached [1m(3)
275 $ git commit -c HEAD@{1} [1m(4)
276 ... [1m(5)
277 $ git add ... [1m(6)
278 $ git diff --cached [1m(7)
279 $ git commit ... [1m(8)
280 1. First, reset the history back one commit so that we remove the
282 original commit, but leave the working tree with all the changes.
283 The -N ensures that any new files added with HEAD are still marked
284 so that git add -p will find them.
285 2. Next, we interactively select diff hunks to add using the git
286 add -p facility. This will ask you about each diff hunk in sequence
287 and you can use simple commands such as "yes, include this", "No
288 don’t include this" or even the very powerful "edit" facility.
289 3. Once satisfied with the hunks you want to include, you should
290 verify what has been prepared for the first commit by using git
291 diff --cached. This shows all the changes that have been moved into
292 the index and are about to be committed.
293 4. Next, commit the changes stored in the index. The -c option
294 specifies to pre-populate the commit message from the original
295 message that you started is a special notation for the commit that
296 HEAD used to be at prior to the original reset commit (1 change
297 ago). See git-reflog(1) for more details. You may also use any
298 other valid commit reference.
299 5. You can repeat steps 2-4 multiple times to break the original
300 code into any number of commits.
301 6. Now you’ve split out many of the changes into their own commits,
302 and might no longer use the patch mode of git add, in order to
303 select all remaining uncommitted changes.
304 7. Once again, check to verify that you’ve included what you want
305 to. You may also wish to verify that git diff doesn’t show any
306 remaining changes to be committed later.
307 8. And finally create the final commit.
308
310 The tables below show what happens when running:
311 git reset --option target
313 to reset the HEAD to another commit (target) with the different reset
316 options depending on the state of the files.
317 In these tables, A, B, C and D are some different states of a file. For
319 example, the first line of the first table means that if a file is in
320 state A in the working tree, in state B in the index, in state C in
321 HEAD and in state D in the target, then "git reset --soft target" will
322 leave the file in the working tree in state A and in the index in state
323 B. It resets (i.e. moves) the HEAD (i.e. the tip of the current branch,
324 if you are on one) to "target" (which has the file in state D).
325 working index HEAD target working index HEAD
327 ----------------------------------------------------
328 A B C D --soft A B D
329 --mixed A D D
330 --hard D D D
331 --merge (disallowed)
332 --keep (disallowed)
333 working index HEAD target working index HEAD
335 ----------------------------------------------------
336 A B C C --soft A B C
337 --mixed A C C
338 --hard C C C
339 --merge (disallowed)
340 --keep A C C
341 working index HEAD target working index HEAD
343 ----------------------------------------------------
344 B B C D --soft B B D
345 --mixed B D D
346 --hard D D D
347 --merge D D D
348 --keep (disallowed)
349 working index HEAD target working index HEAD
351 ----------------------------------------------------
352 B B C C --soft B B C
353 --mixed B C C
354 --hard C C C
355 --merge C C C
356 --keep B C C
357 working index HEAD target working index HEAD
359 ----------------------------------------------------
360 B C C D --soft B C D
361 --mixed B D D
362 --hard D D D
363 --merge (disallowed)
364 --keep (disallowed)
365 working index HEAD target working index HEAD
367 ----------------------------------------------------
368 B C C C --soft B C C
369 --mixed B C C
370 --hard C C C
371 --merge B C C
372 --keep B C C
373 "reset --merge" is meant to be used when resetting out of a conflicted
375 merge. Any mergy operation guarantees that the working tree file that
376 is involved in the merge does not have local change wrt the index
377 before it starts, and that it writes the result out to the working
378 tree. So if we see some difference between the index and the target and
379 also between the index and the working tree, then it means that we are
380 not resetting out from a state that a mergy operation left after
381 failing with a conflict. That is why we disallow --merge option in this
382 case.
383 "reset --keep" is meant to be used when removing some of the last
385 commits in the current branch while keeping changes in the working
386 tree. If there could be conflicts between the changes in the commit we
387 want to remove and the changes in the working tree we want to keep, the
388 reset is disallowed. That’s why it is disallowed if there are both
389 changes between the working tree and HEAD, and between HEAD and the
390 target. To be safe, it is also disallowed when there are unmerged
391 entries.
392 The following tables show what happens when there are unmerged entries:
394 working index HEAD target working index HEAD
396 ----------------------------------------------------
397 X U A B --soft (disallowed)
398 --mixed X B B
399 --hard B B B
400 --merge B B B
401 --keep (disallowed)
402 working index HEAD target working index HEAD
404 ----------------------------------------------------
405 X U A A --soft (disallowed)
406 --mixed X A A
407 --hard A A A
408 --merge A A A
409 --keep (disallowed)
410 X means any state and U means an unmerged index.
412
414 Part of the git(1) suite
415Git 2.18.1 05/14/2019 GIT-RESET(1)