6 git-update-index - Register file contents in the working tree to the index
13 [--add] [--remove | --force-remove] [--replace]
14 [--refresh] [-q] [--unmerged] [--ignore-missing]
15 [(--cacheinfo <mode>,<object>,<file>)...]
17 [--[no-]assume-unchanged]
18 [--[no-]skip-worktree]
19 [--[no-]ignore-skip-worktree-entries]
20 [--[no-]fsmonitor-valid]
23 [--[no-|test-|force-]untracked-cache]
25 [--really-refresh] [--unresolve] [--again | -g]
26 [--info-only] [--index-info]
27 [-z] [--stdin] [--index-version <n>]
33 Modifies the index. Each file mentioned is updated into the index and
34 any 'unmerged' or 'needs updating' state is cleared.
36 See also linkgit:git-add[1] for a more user-friendly way to do some of
37 the most common operations on the index.
39 The way 'git update-index' handles files it is told about can be modified
40 using the various options:
45 If a specified file isn't in the index already then it's
47 Default behaviour is to ignore new files.
50 If a specified file is in the index but is missing then it's
52 Default behavior is to ignore removed files.
55 Looks at the current index and checks to see if merges or
56 updates are needed by checking stat() information.
59 Quiet. If --refresh finds that the index needs an update, the
60 default behavior is to error out. This option makes
61 'git update-index' continue anyway.
64 Do not try to update submodules. This option is only respected
65 when passed before --refresh.
68 If --refresh finds unmerged changes in the index, the default
69 behavior is to error out. This option makes 'git update-index'
73 Ignores missing files during a --refresh
75 --cacheinfo <mode>,<object>,<path>::
76 --cacheinfo <mode> <object> <path>::
77 Directly insert the specified info into the index. For
78 backward compatibility, you can also give these three
79 arguments as three separate parameters, but new users are
80 encouraged to use a single-parameter form.
83 Read index information from stdin.
86 Set the execute permissions on the updated files.
88 --[no-]assume-unchanged::
89 When this flag is specified, the object names recorded
90 for the paths are not updated. Instead, this option
91 sets/unsets the "assume unchanged" bit for the
92 paths. When the "assume unchanged" bit is on, the user
93 promises not to change the file and allows Git to assume
94 that the working tree file matches what is recorded in
95 the index. If you want to change the working tree file,
96 you need to unset the bit to tell Git. This is
97 sometimes helpful when working with a big project on a
98 filesystem that has a very slow lstat(2) system call
101 Git will fail (gracefully) in case it needs to modify this file
102 in the index e.g. when merging in a commit;
103 thus, in case the assumed-untracked file is changed upstream,
104 you will need to handle the situation manually.
107 Like `--refresh`, but checks stat information unconditionally,
108 without regard to the "assume unchanged" setting.
110 --[no-]skip-worktree::
111 When one of these flags is specified, the object names recorded
112 for the paths are not updated. Instead, these options
113 set and unset the "skip-worktree" bit for the paths. See
114 section "Skip-worktree bit" below for more information.
117 --[no-]ignore-skip-worktree-entries::
118 Do not remove skip-worktree (AKA "index-only") entries even when
119 the `--remove` option was specified.
121 --[no-]fsmonitor-valid::
122 When one of these flags is specified, the object names recorded
123 for the paths are not updated. Instead, these options
124 set and unset the "fsmonitor valid" bit for the paths. See
125 section "File System Monitor" below for more information.
129 Runs 'git update-index' itself on the paths whose index
130 entries are different from those of the `HEAD` commit.
133 Restores the 'unmerged' or 'needs updating' state of a
134 file during a merge if it was cleared by accident.
137 Do not create objects in the object database for all
138 <file> arguments that follow this flag; just insert
139 their object IDs into the index.
142 Remove the file from the index even when the working directory
143 still has such a file. (Implies --remove.)
146 By default, when a file `path` exists in the index,
147 'git update-index' refuses an attempt to add `path/file`.
148 Similarly if a file `path/file` exists, a file `path`
149 cannot be added. With --replace flag, existing entries
150 that conflict with the entry being added are
151 automatically removed with warning messages.
154 Instead of taking a list of paths from the command line,
155 read a list of paths from the standard input. Paths are
156 separated by LF (i.e. one path per line) by default.
159 Report what is being added and removed from the index.
161 --index-version <n>::
162 Write the resulting index out in the named on-disk format version.
163 Supported versions are 2, 3, and 4. The current default version is 2
164 or 3, depending on whether extra features are used, such as
165 `git add -N`. With `--verbose`, also report the version the index
166 file uses before and after this command.
168 Version 4 performs a simple pathname compression that reduces index
169 size by 30%-50% on large repositories, which results in faster load
170 time. Git supports it since version 1.8.0, released in October 2012,
171 and support for it was added to libgit2 in 2016 and to JGit in 2020.
172 Older versions of this manual page called it "relatively young", but
173 it should be considered mature technology these days.
175 --show-index-version::
176 Report the index format version used by the on-disk index file.
177 See `--index-version` above.
180 Only meaningful with `--stdin` or `--index-info`; paths are
181 separated with NUL character instead of LF.
185 Enable or disable split index mode. If split-index mode is
186 already enabled and `--split-index` is given again, all
187 changes in $GIT_DIR/index are pushed back to the shared index
190 These options take effect whatever the value of the `core.splitIndex`
191 configuration variable (see linkgit:git-config[1]). But a warning is
192 emitted when the change goes against the configured value, as the
193 configured value will take effect next time the index is read and this
194 will remove the intended effect of the option.
197 --no-untracked-cache::
198 Enable or disable untracked cache feature. Please use
199 `--test-untracked-cache` before enabling it.
201 These options take effect whatever the value of the `core.untrackedCache`
202 configuration variable (see linkgit:git-config[1]). But a warning is
203 emitted when the change goes against the configured value, as the
204 configured value will take effect next time the index is read and this
205 will remove the intended effect of the option.
207 --test-untracked-cache::
208 Only perform tests on the working directory to make sure
209 untracked cache can be used. You have to manually enable
210 untracked cache using `--untracked-cache` or
211 `--force-untracked-cache` or the `core.untrackedCache`
212 configuration variable afterwards if you really want to use
213 it. If a test fails the exit code is 1 and a message
214 explains what is not working as needed, otherwise the exit
215 code is 0 and OK is printed.
217 --force-untracked-cache::
218 Same as `--untracked-cache`. Provided for backwards
219 compatibility with older versions of Git where
220 `--untracked-cache` used to imply `--test-untracked-cache` but
221 this option would enable the extension unconditionally.
225 Enable or disable files system monitor feature. These options
226 take effect whatever the value of the `core.fsmonitor`
227 configuration variable (see linkgit:git-config[1]). But a warning
228 is emitted when the change goes against the configured value, as
229 the configured value will take effect next time the index is
230 read and this will remove the intended effect of the option.
233 Do not interpret any more arguments as options.
237 Note that files beginning with '.' are discarded. This includes
238 `./file` and `dir/./file`. If you don't want this, then use
240 The same applies to directories ending '/' and paths with '//'
244 `--refresh` does not calculate a new sha1 file or bring the index
245 up to date for mode/content changes. But what it *does* do is to
246 "re-match" the stat information of a file with the index, so that you
247 can refresh the index for a file that hasn't been changed but where
248 the stat entry is out of date.
250 For example, you'd want to do this after doing a 'git read-tree', to link
251 up the stat index details with the proper files.
253 USING --CACHEINFO OR --INFO-ONLY
254 --------------------------------
255 `--cacheinfo` is used to register a file that is not in the
256 current working directory. This is useful for minimum-checkout
259 To pretend you have a file at path with mode and sha1, say:
262 $ git update-index --add --cacheinfo <mode>,<sha1>,<path>
265 `--info-only` is used to register files without placing them in the object
266 database. This is useful for status-only repositories.
268 Both `--cacheinfo` and `--info-only` behave similarly: the index is updated
269 but the object database isn't. `--cacheinfo` is useful when the object is
270 in the database but the file isn't available locally. `--info-only` is
271 useful when the file is available, but you do not wish to update the
278 `--index-info` is a more powerful mechanism that lets you feed
279 multiple entry definitions from the standard input, and designed
280 specifically for scripts. It can take inputs of three formats:
282 . mode SP type SP sha1 TAB path
284 This format is to stuff `git ls-tree` output into the index.
286 . mode SP sha1 SP stage TAB path
288 This format is to put higher order stages into the
289 index file and matches 'git ls-files --stage' output.
291 . mode SP sha1 TAB path
293 This format is no longer produced by any Git command, but is
294 and will continue to be supported by `update-index --index-info`.
296 To place a higher stage entry to the index, the path should
297 first be removed by feeding a mode=0 entry for the path, and
298 then feeding necessary input lines in the third format.
300 For example, starting with this index:
304 100644 8a1218a1024a212bb3db30becd860315f9f3ac52 0 frotz
307 you can feed the following input to `--index-info`:
310 $ git update-index --index-info
311 0 0000000000000000000000000000000000000000 frotz
312 100644 8a1218a1024a212bb3db30becd860315f9f3ac52 1 frotz
313 100755 8a1218a1024a212bb3db30becd860315f9f3ac52 2 frotz
316 The first line of the input feeds 0 as the mode to remove the
317 path; the SHA-1 does not matter as long as it is well formatted.
318 Then the second and third line feeds stage 1 and stage 2 entries
319 for that path. After the above, we would end up with this:
323 100644 8a1218a1024a212bb3db30becd860315f9f3ac52 1 frotz
324 100755 8a1218a1024a212bb3db30becd860315f9f3ac52 2 frotz
328 USING ``ASSUME UNCHANGED'' BIT
329 ------------------------------
331 Many operations in Git depend on your filesystem to have an
332 efficient `lstat(2)` implementation, so that `st_mtime`
333 information for working tree files can be cheaply checked to see
334 if the file contents have changed from the version recorded in
335 the index file. Unfortunately, some filesystems have
336 inefficient `lstat(2)`. If your filesystem is one of them, you
337 can set "assume unchanged" bit to paths you have not changed to
338 cause Git not to do this check. Note that setting this bit on a
339 path does not mean Git will check the contents of the file to
340 see if it has changed -- it makes Git to omit any checking and
341 assume it has *not* changed. When you make changes to working
342 tree files, you have to explicitly tell Git about it by dropping
343 "assume unchanged" bit, either before or after you modify them.
345 In order to set "assume unchanged" bit, use `--assume-unchanged`
346 option. To unset, use `--no-assume-unchanged`. To see which files
347 have the "assume unchanged" bit set, use `git ls-files -v`
348 (see linkgit:git-ls-files[1]).
350 The command looks at `core.ignorestat` configuration variable. When
351 this is true, paths updated with `git update-index paths...` and
352 paths updated with other Git commands that update both index and
353 working tree (e.g. 'git apply --index', 'git checkout-index -u',
354 and 'git read-tree -u') are automatically marked as "assume
355 unchanged". Note that "assume unchanged" bit is *not* set if
356 `git update-index --refresh` finds the working tree file matches
357 the index (use `git update-index --really-refresh` if you want
358 to mark them as "assume unchanged").
360 Sometimes users confuse the assume-unchanged bit with the
361 skip-worktree bit. See the final paragraph in the "Skip-worktree bit"
362 section below for an explanation of the differences.
367 To update and refresh only the files already checked out:
370 $ git checkout-index -n -f -a && git update-index --ignore-missing --refresh
373 On an inefficient filesystem with `core.ignorestat` set::
376 $ git update-index --really-refresh <1>
377 $ git update-index --no-assume-unchanged foo.c <2>
378 $ git diff --name-only <3>
380 $ git diff --name-only <4>
382 $ git update-index foo.c <5>
383 $ git diff --name-only <6>
385 $ git diff --name-only <7>
386 $ git update-index --no-assume-unchanged foo.c <8>
387 $ git diff --name-only <9>
391 <1> forces lstat(2) to set "assume unchanged" bits for paths that match index.
392 <2> mark the path to be edited.
393 <3> this does lstat(2) and finds index matches the path.
394 <4> this does lstat(2) and finds index does *not* match the path.
395 <5> registering the new version to index sets "assume unchanged" bit.
396 <6> and it is assumed unchanged.
397 <7> even after you edit it.
398 <8> you can tell about the change after the fact.
399 <9> now it checks with lstat(2) and finds it has been changed.
405 Skip-worktree bit can be defined in one (long) sentence: Tell git to
406 avoid writing the file to the working directory when reasonably
407 possible, and treat the file as unchanged when it is not
408 present in the working directory.
410 Note that not all git commands will pay attention to this bit, and
411 some only partially support it.
413 The update-index flags and the read-tree capabilities relating to the
414 skip-worktree bit predated the introduction of the
415 linkgit:git-sparse-checkout[1] command, which provides a much easier
416 way to configure and handle the skip-worktree bits. If you want to
417 reduce your working tree to only deal with a subset of the files in
418 the repository, we strongly encourage the use of
419 linkgit:git-sparse-checkout[1] in preference to the low-level
420 update-index and read-tree primitives.
422 The primary purpose of the skip-worktree bit is to enable sparse
423 checkouts, i.e. to have working directories with only a subset of
424 paths present. When the skip-worktree bit is set, Git commands (such
425 as `switch`, `pull`, `merge`) will avoid writing these files.
426 However, these commands will sometimes write these files anyway in
427 important cases such as conflicts during a merge or rebase. Git
428 commands will also avoid treating the lack of such files as an
429 intentional deletion; for example `git add -u` will not stage a
430 deletion for these files and `git commit -a` will not make a commit
431 deleting them either.
433 Although this bit looks similar to assume-unchanged bit, its goal is
434 different. The assume-unchanged bit is for leaving the file in the
435 working tree but having Git omit checking it for changes and presuming
436 that the file has not been changed (though if it can determine without
437 stat'ing the file that it has changed, it is free to record the
438 changes). skip-worktree tells Git to ignore the absence of the file,
439 avoid updating it when possible with commands that normally update
440 much of the working directory (e.g. `checkout`, `switch`, `pull`,
441 etc.), and not have its absence be recorded in commits. Note that in
442 sparse checkouts (setup by `git sparse-checkout` or by configuring
443 core.sparseCheckout to true), if a file is marked as skip-worktree in
444 the index but is found in the working tree, Git will clear the
445 skip-worktree bit for that file.
450 This mode is designed for repositories with very large indexes, and
451 aims at reducing the time it takes to repeatedly write these indexes.
453 In this mode, the index is split into two files, $GIT_DIR/index and
454 $GIT_DIR/sharedindex.<SHA-1>. Changes are accumulated in
455 $GIT_DIR/index, the split index, while the shared index file contains
456 all index entries and stays unchanged.
458 All changes in the split index are pushed back to the shared index
459 file when the number of entries in the split index reaches a level
460 specified by the splitIndex.maxPercentChange config variable (see
461 linkgit:git-config[1]).
463 Each time a new shared index file is created, the old shared index
464 files are deleted if their modification time is older than what is
465 specified by the splitIndex.sharedIndexExpire config variable (see
466 linkgit:git-config[1]).
468 To avoid deleting a shared index file that is still used, its
469 modification time is updated to the current time every time a new split
470 index based on the shared index file is either created or read from.
475 This cache is meant to speed up commands that involve determining
476 untracked files such as `git status`.
478 This feature works by recording the mtime of the working tree
479 directories and then omitting reading directories and stat calls
480 against files in those directories whose mtime hasn't changed. For
481 this to work the underlying operating system and file system must
482 change the `st_mtime` field of directories if files in the directory
483 are added, modified or deleted.
485 You can test whether the filesystem supports that with the
486 `--test-untracked-cache` option. The `--untracked-cache` option used
487 to implicitly perform that test in older versions of Git, but that's
490 If you want to enable (or disable) this feature, it is easier to use
491 the `core.untrackedCache` configuration variable (see
492 linkgit:git-config[1]) than using the `--untracked-cache` option to
493 `git update-index` in each repository, especially if you want to do so
494 across all repositories you use, because you can set the configuration
495 variable to `true` (or `false`) in your `$HOME/.gitconfig` just once
496 and have it affect all repositories you touch.
498 When the `core.untrackedCache` configuration variable is changed, the
499 untracked cache is added to or removed from the index the next time a
500 command reads the index; while when `--[no-|force-]untracked-cache`
501 are used, the untracked cache is immediately added to or removed from
504 Before 2.17, the untracked cache had a bug where replacing a directory
505 with a symlink to another directory could cause it to incorrectly show
506 files tracked by git as untracked. See the "status: add a failing test
507 showing a core.untrackedCache bug" commit to git.git. A workaround for
508 that is (and this might work for other undiscovered bugs in the
512 $ git -c core.untrackedCache=false status
515 This bug has also been shown to affect non-symlink cases of replacing
516 a directory with a file when it comes to the internal structures of
517 the untracked cache, but no case has been reported where this resulted in
518 wrong "git status" output.
520 There are also cases where existing indexes written by git versions
521 before 2.17 will reference directories that don't exist anymore,
522 potentially causing many "could not open directory" warnings to be
523 printed on "git status". These are new warnings for existing issues
524 that were previously silently discarded.
526 As with the bug described above the solution is to one-off do a "git
527 status" run with `core.untrackedCache=false` to flush out the leftover
533 This feature is intended to speed up git operations for repos that have
534 large working directories.
536 It enables git to work together with a file system monitor (see
537 linkgit:git-fsmonitor{litdd}daemon[1]
539 "fsmonitor-watchman" section of linkgit:githooks[5]) that can
540 inform it as to what files have been modified. This enables git to avoid
541 having to lstat() every file to find modified files.
543 When used in conjunction with the untracked cache, it can further improve
544 performance by avoiding the cost of scanning the entire working directory
545 looking for new files.
547 If you want to enable (or disable) this feature, it is easier to use
548 the `core.fsmonitor` configuration variable (see
549 linkgit:git-config[1]) than using the `--fsmonitor` option to `git
550 update-index` in each repository, especially if you want to do so
551 across all repositories you use, because you can set the configuration
552 variable in your `$HOME/.gitconfig` just once and have it affect all
553 repositories you touch.
555 When the `core.fsmonitor` configuration variable is changed, the
556 file system monitor is added to or removed from the index the next time
557 a command reads the index. When `--[no-]fsmonitor` are used, the file
558 system monitor is immediately added to or removed from the index.
563 The command honors `core.filemode` configuration variable. If
564 your repository is on a filesystem whose executable bits are
565 unreliable, this should be set to 'false' (see linkgit:git-config[1]).
566 This causes the command to ignore differences in file modes recorded
567 in the index and the file mode on the filesystem if they differ only on
568 executable bit. On such an unfortunate filesystem, you may
569 need to use 'git update-index --chmod='.
571 Quite similarly, if `core.symlinks` configuration variable is set
572 to 'false' (see linkgit:git-config[1]), symbolic links are checked out
573 as plain files, and this command does not modify a recorded file mode
574 from symbolic link to regular file.
576 The command looks at `core.ignorestat` configuration variable. See
577 'Using "assume unchanged" bit' section above.
579 The command also looks at `core.trustctime` configuration variable.
580 It can be useful when the inode change time is regularly modified by
581 something outside Git (file system crawlers and backup systems use
582 ctime for marking files processed) (see linkgit:git-config[1]).
584 The untracked cache extension can be enabled by the
585 `core.untrackedCache` configuration variable (see
586 linkgit:git-config[1]).
591 Users often try to use the assume-unchanged and skip-worktree bits
592 to tell Git to ignore changes to files that are tracked. This does not
593 work as expected, since Git may still check working tree files against
594 the index when performing certain operations. In general, Git does not
595 provide a way to ignore changes to tracked files, so alternate solutions
598 For example, if the file you want to change is some sort of config file,
599 the repository can include a sample config file that can then be copied
600 into the ignored name and modified. The repository can even include a
601 script to treat the sample file as a template, modifying and copying it
606 linkgit:git-config[1],
608 linkgit:git-ls-files[1]
612 Part of the linkgit:git[1] suite
projects / thirdparty / git.git / blob