1GITATTRIBUTES(5) Git Manual GITATTRIBUTES(5)
2
6 gitattributes - Defining attributes per path
7
9 $GIT_DIR/info/attributes, .gitattributes
10
12 A gitattributes file is a simple text file that gives attributes to
13 pathnames.
14 Each line in gitattributes file is of form:
16 pattern attr1 attr2 ...
18 That is, a pattern followed by an attributes list, separated by
20 whitespaces. Leading and trailing whitespaces are ignored. Lines that
21 begin with # are ignored. Patterns that begin with a double quote are
22 quoted in C style. When the pattern matches the path in question, the
23 attributes listed on the line are given to the path.
24 Each attribute can be in one of these states for a given path:
26 Set
28 The path has the attribute with special value "true"; this is
29 specified by listing only the name of the attribute in the
30 attribute list.
31 Unset
33 The path has the attribute with special value "false"; this is
34 specified by listing the name of the attribute prefixed with a dash
35 - in the attribute list.
36 Set to a value
38 The path has the attribute with specified string value; this is
39 specified by listing the name of the attribute followed by an equal
40 sign = and its value in the attribute list.
41 Unspecified
43 No pattern matches the path, and nothing says if the path has or
44 does not have the attribute, the attribute for the path is said to
45 be Unspecified.
46 When more than one pattern matches the path, a later line overrides an
48 earlier line. This overriding is done per attribute.
49 The rules by which the pattern matches paths are the same as in
51 .gitignore files (see gitignore(5)), with a few exceptions:
52 • negative patterns are forbidden
54 • patterns that match a directory do not recursively match paths
56 inside that directory (so using the trailing-slash path/ syntax is
57 pointless in an attributes file; use path/** instead)
58 When deciding what attributes are assigned to a path, Git consults
60 $GIT_DIR/info/attributes file (which has the highest precedence),
61 .gitattributes file in the same directory as the path in question, and
62 its parent directories up to the toplevel of the work tree (the further
63 the directory that contains .gitattributes is from the path in
64 question, the lower its precedence). Finally global and system-wide
65 files are considered (they have the lowest precedence).
66 When the .gitattributes file is missing from the work tree, the path in
68 the index is used as a fall-back. During checkout process,
69 .gitattributes in the index is used and then the file in the working
70 tree is used as a fall-back.
71 If you wish to affect only a single repository (i.e., to assign
73 attributes to files that are particular to one user’s workflow for that
74 repository), then attributes should be placed in the
75 $GIT_DIR/info/attributes file. Attributes which should be
76 version-controlled and distributed to other repositories (i.e.,
77 attributes of interest to all users) should go into .gitattributes
78 files. Attributes that should affect all repositories for a single user
79 should be placed in a file specified by the core.attributesFile
80 configuration option (see git-config(1)). Its default value is
81 $XDG_CONFIG_HOME/git/attributes. If $XDG_CONFIG_HOME is either not set
82 or empty, $HOME/.config/git/attributes is used instead. Attributes for
83 all users on a system should be placed in the
84 $(prefix)/etc/gitattributes file.
85 Sometimes you would need to override a setting of an attribute for a
87 path to Unspecified state. This can be done by listing the name of the
88 attribute prefixed with an exclamation point !.
89
91 Certain operations by Git can be influenced by assigning particular
92 attributes to a path. Currently, the following operations are
93 attributes-aware.
94 Checking-out and checking-in
96 These attributes affect how the contents stored in the repository are
97 copied to the working tree files when commands such as git switch, git
98 checkout and git merge run. They also affect how Git stores the
99 contents you prepare in the working tree in the repository upon git add
100 and git commit.
101 text
103 This attribute enables and controls end-of-line normalization. When
104 a text file is normalized, its line endings are converted to LF in
105 the repository. To control what line ending style is used in the
106 working directory, use the eol attribute for a single file and the
107 core.eol configuration variable for all text files. Note that
108 setting core.autocrlf to true or input overrides core.eol (see the
109 definitions of those options in git-config(1)).
110 Set
112 Setting the text attribute on a path enables end-of-line
113 normalization and marks the path as a text file. End-of-line
114 conversion takes place without guessing the content type.
115 Unset
117 Unsetting the text attribute on a path tells Git not to attempt
118 any end-of-line conversion upon checkin or checkout.
119 Set to string value "auto"
121 When text is set to "auto", the path is marked for automatic
122 end-of-line conversion. If Git decides that the content is
123 text, its line endings are converted to LF on checkin. When the
124 file has been committed with CRLF, no conversion is done.
125 Unspecified
127 If the text attribute is unspecified, Git uses the
128 core.autocrlf configuration variable to determine if the file
129 should be converted.
130 Any other value causes Git to act as if text has been left
132 unspecified.
133 eol
135 This attribute sets a specific line-ending style to be used in the
136 working directory. This attribute has effect only if the text
137 attribute is set or unspecified, or if it is set to auto, the file
138 is detected as text, and it is stored with LF endings in the index.
139 Note that setting this attribute on paths which are in the index
140 with CRLF line endings may make the paths to be considered dirty
141 unless text=auto is set. Adding the path to the index again will
142 normalize the line endings in the index.
143 Set to string value "crlf"
145 This setting forces Git to normalize line endings for this file
146 on checkin and convert them to CRLF when the file is checked
147 out.
148 Set to string value "lf"
150 This setting forces Git to normalize line endings to LF on
151 checkin and prevents conversion to CRLF when the file is
152 checked out.
153 Backwards compatibility with crlf attribute
155 For backwards compatibility, the crlf attribute is interpreted as
156 follows:
157 crlf text
159 -crlf -text
160 crlf=input eol=lf
161 End-of-line conversion
163 While Git normally leaves file contents alone, it can be configured
164 to normalize line endings to LF in the repository and, optionally,
165 to convert them to CRLF when files are checked out.
166 If you simply want to have CRLF line endings in your working
168 directory regardless of the repository you are working with, you
169 can set the config variable "core.autocrlf" without using any
170 attributes.
171 [core]
173 autocrlf = true
174 This does not force normalization of text files, but does ensure
176 that text files that you introduce to the repository have their
177 line endings normalized to LF when they are added, and that files
178 that are already normalized in the repository stay normalized.
179 If you want to ensure that text files that any contributor
181 introduces to the repository have their line endings normalized,
182 you can set the text attribute to "auto" for all files.
183 * text=auto
185 The attributes allow a fine-grained control, how the line endings
187 are converted. Here is an example that will make Git normalize
188 .txt, .vcproj and .sh files, ensure that .vcproj files have CRLF
189 and .sh files have LF in the working directory, and prevent .jpg
190 files from being normalized regardless of their content.
191 * text=auto
193 *.txt text
194 *.vcproj text eol=crlf
195 *.sh text eol=lf
196 *.jpg -text
197 Note
200 When text=auto conversion is enabled in a cross-platform
201 project using push and pull to a central repository the text
202 files containing CRLFs should be normalized.
203 From a clean working directory:
205 $ echo "* text=auto" >.gitattributes
207 $ git add --renormalize .
208 $ git status # Show files that will be normalized
209 $ git commit -m "Introduce end-of-line normalization"
210 If any files that should not be normalized show up in git status,
212 unset their text attribute before running git add -u.
213 manual.pdf -text
215 Conversely, text files that Git does not detect can have
217 normalization enabled manually.
218 weirdchars.txt text
220 If core.safecrlf is set to "true" or "warn", Git verifies if the
222 conversion is reversible for the current setting of core.autocrlf.
223 For "true", Git rejects irreversible conversions; for "warn", Git
224 only prints a warning but accepts an irreversible conversion. The
225 safety triggers to prevent such a conversion done to the files in
226 the work tree, but there are a few exceptions. Even though...
227 • git add itself does not touch the files in the work tree, the
229 next checkout would, so the safety triggers;
230 • git apply to update a text file with a patch does touch the
232 files in the work tree, but the operation is about text files
233 and CRLF conversion is about fixing the line ending
234 inconsistencies, so the safety does not trigger;
235 • git diff itself does not touch the files in the work tree, it
237 is often run to inspect the changes you intend to next git add.
238 To catch potential problems early, safety triggers.
239 working-tree-encoding
241 Git recognizes files encoded in ASCII or one of its supersets (e.g.
242 UTF-8, ISO-8859-1, ...) as text files. Files encoded in certain
243 other encodings (e.g. UTF-16) are interpreted as binary and
244 consequently built-in Git text processing tools (e.g. git diff) as
245 well as most Git web front ends do not visualize the contents of
246 these files by default.
247 In these cases you can tell Git the encoding of a file in the
249 working directory with the working-tree-encoding attribute. If a
250 file with this attribute is added to Git, then Git re-encodes the
251 content from the specified encoding to UTF-8. Finally, Git stores
252 the UTF-8 encoded content in its internal data structure (called
253 "the index"). On checkout the content is re-encoded back to the
254 specified encoding.
255 Please note that using the working-tree-encoding attribute may have
257 a number of pitfalls:
258 • Alternative Git implementations (e.g. JGit or libgit2) and
260 older Git versions (as of March 2018) do not support the
261 working-tree-encoding attribute. If you decide to use the
262 working-tree-encoding attribute in your repository, then it is
263 strongly recommended to ensure that all clients working with
264 the repository support it.
265 For example, Microsoft Visual Studio resources files (*.rc) or
267 PowerShell script files (*.ps1) are sometimes encoded in
268 UTF-16. If you declare *.ps1 as files as UTF-16 and you add
269 foo.ps1 with a working-tree-encoding enabled Git client, then
270 foo.ps1 will be stored as UTF-8 internally. A client without
271 working-tree-encoding support will checkout foo.ps1 as UTF-8
272 encoded file. This will typically cause trouble for the users
273 of this file.
274 If a Git client that does not support the working-tree-encoding
276 attribute adds a new file bar.ps1, then bar.ps1 will be stored
277 "as-is" internally (in this example probably as UTF-16). A
278 client with working-tree-encoding support will interpret the
279 internal contents as UTF-8 and try to convert it to UTF-16 on
280 checkout. That operation will fail and cause an error.
281 • Reencoding content to non-UTF encodings can cause errors as the
283 conversion might not be UTF-8 round trip safe. If you suspect
284 your encoding to not be round trip safe, then add it to
285 core.checkRoundtripEncoding to make Git check the round trip
286 encoding (see git-config(1)). SHIFT-JIS (Japanese character
287 set) is known to have round trip issues with UTF-8 and is
288 checked by default.
289 • Reencoding content requires resources that might slow down
291 certain Git operations (e.g git checkout or git add).
292 Use the working-tree-encoding attribute only if you cannot store a
294 file in UTF-8 encoding and if you want Git to be able to process
295 the content as text.
296 As an example, use the following attributes if your *.ps1 files are
298 UTF-16 encoded with byte order mark (BOM) and you want Git to
299 perform automatic line ending conversion based on your platform.
300 *.ps1 text working-tree-encoding=UTF-16
302 Use the following attributes if your *.ps1 files are UTF-16 little
304 endian encoded without BOM and you want Git to use Windows line
305 endings in the working directory (use UTF-16LE-BOM instead of
306 UTF-16LE if you want UTF-16 little endian with BOM). Please note,
307 it is highly recommended to explicitly define the line endings with
308 eol if the working-tree-encoding attribute is used to avoid
309 ambiguity.
310 *.ps1 text working-tree-encoding=UTF-16LE eol=CRLF
312 You can get a list of all available encodings on your platform with
314 the following command:
315 iconv --list
317 If you do not know the encoding of a file, then you can use the
319 file command to guess the encoding:
320 file foo.ps1
322 ident
324 When the attribute ident is set for a path, Git replaces $Id$ in
325 the blob object with $Id:, followed by the 40-character hexadecimal
326 blob object name, followed by a dollar sign $ upon checkout. Any
327 byte sequence that begins with $Id: and ends with $ in the worktree
328 file is replaced with $Id$ upon check-in.
329 filter
331 A filter attribute can be set to a string value that names a filter
332 driver specified in the configuration.
333 A filter driver consists of a clean command and a smudge command,
335 either of which can be left unspecified. Upon checkout, when the
336 smudge command is specified, the command is fed the blob object
337 from its standard input, and its standard output is used to update
338 the worktree file. Similarly, the clean command is used to convert
339 the contents of worktree file upon checkin. By default these
340 commands process only a single blob and terminate. If a long
341 running process filter is used in place of clean and/or smudge
342 filters, then Git can process all blobs with a single filter
343 command invocation for the entire life of a single Git command, for
344 example git add --all. If a long running process filter is
345 configured then it always takes precedence over a configured single
346 blob filter. See section below for the description of the protocol
347 used to communicate with a process filter.
348 One use of the content filtering is to massage the content into a
350 shape that is more convenient for the platform, filesystem, and the
351 user to use. For this mode of operation, the key phrase here is
352 "more convenient" and not "turning something unusable into usable".
353 In other words, the intent is that if someone unsets the filter
354 driver definition, or does not have the appropriate filter program,
355 the project should still be usable.
356 Another use of the content filtering is to store the content that
358 cannot be directly used in the repository (e.g. a UUID that refers
359 to the true content stored outside Git, or an encrypted content)
360 and turn it into a usable form upon checkout (e.g. download the
361 external content, or decrypt the encrypted content).
362 These two filters behave differently, and by default, a filter is
364 taken as the former, massaging the contents into more convenient
365 shape. A missing filter driver definition in the config, or a
366 filter driver that exits with a non-zero status, is not an error
367 but makes the filter a no-op passthru.
368 You can declare that a filter turns a content that by itself is
370 unusable into a usable content by setting the
371 filter.<driver>.required configuration variable to true.
372 Note: Whenever the clean filter is changed, the repo should be
374 renormalized: $ git add --renormalize .
375 For example, in .gitattributes, you would assign the filter
377 attribute for paths.
378 *.c filter=indent
380 Then you would define a "filter.indent.clean" and
382 "filter.indent.smudge" configuration in your .git/config to specify
383 a pair of commands to modify the contents of C programs when the
384 source files are checked in ("clean" is run) and checked out (no
385 change is made because the command is "cat").
386 [filter "indent"]
388 clean = indent
389 smudge = cat
390 For best results, clean should not alter its output further if it
392 is run twice ("clean→clean" should be equivalent to "clean"), and
393 multiple smudge commands should not alter clean's output
394 ("smudge→smudge→clean" should be equivalent to "clean"). See the
395 section on merging below.
396 The "indent" filter is well-behaved in this regard: it will not
398 modify input that is already correctly indented. In this case, the
399 lack of a smudge filter means that the clean filter must accept its
400 own output without modifying it.
401 If a filter must succeed in order to make the stored contents
403 usable, you can declare that the filter is required, in the
404 configuration:
405 [filter "crypt"]
407 clean = openssl enc ...
408 smudge = openssl enc -d ...
409 required
410 Sequence "%f" on the filter command line is replaced with the name
412 of the file the filter is working on. A filter might use this in
413 keyword substitution. For example:
414 [filter "p4"]
416 clean = git-p4-filter --clean %f
417 smudge = git-p4-filter --smudge %f
418 Note that "%f" is the name of the path that is being worked on.
420 Depending on the version that is being filtered, the corresponding
421 file on disk may not exist, or may have different contents. So,
422 smudge and clean commands should not try to access the file on
423 disk, but only act as filters on the content provided to them on
424 standard input.
425 Long Running Filter Process
427 If the filter command (a string value) is defined via
428 filter.<driver>.process then Git can process all blobs with a
429 single filter invocation for the entire life of a single Git
430 command. This is achieved by using the long-running process
431 protocol (described in
432 technical/long-running-process-protocol.txt).
433 When Git encounters the first file that needs to be cleaned or
435 smudged, it starts the filter and performs the handshake. In the
436 handshake, the welcome message sent by Git is "git-filter-client",
437 only version 2 is supported, and the supported capabilities are
438 "clean", "smudge", and "delay".
439 Afterwards Git sends a list of "key=value" pairs terminated with a
441 flush packet. The list will contain at least the filter command
442 (based on the supported capabilities) and the pathname of the file
443 to filter relative to the repository root. Right after the flush
444 packet Git sends the content split in zero or more pkt-line packets
445 and a flush packet to terminate content. Please note, that the
446 filter must not send any response before it received the content
447 and the final flush packet. Also note that the "value" of a
448 "key=value" pair can contain the "=" character whereas the key
449 would never contain that character.
450 packet: git> command=smudge
452 packet: git> pathname=path/testfile.dat
453 packet: git> 0000
454 packet: git> CONTENT
455 packet: git> 0000
456 The filter is expected to respond with a list of "key=value" pairs
458 terminated with a flush packet. If the filter does not experience
459 problems then the list must contain a "success" status. Right after
460 these packets the filter is expected to send the content in zero or
461 more pkt-line packets and a flush packet at the end. Finally, a
462 second list of "key=value" pairs terminated with a flush packet is
463 expected. The filter can change the status in the second list or
464 keep the status as is with an empty list. Please note that the
465 empty list must be terminated with a flush packet regardless.
466 packet: git< status=success
468 packet: git< 0000
469 packet: git< SMUDGED_CONTENT
470 packet: git< 0000
471 packet: git< 0000 # empty list, keep "status=success" unchanged!
472 If the result content is empty then the filter is expected to
474 respond with a "success" status and a flush packet to signal the
475 empty content.
476 packet: git< status=success
478 packet: git< 0000
479 packet: git< 0000 # empty content!
480 packet: git< 0000 # empty list, keep "status=success" unchanged!
481 In case the filter cannot or does not want to process the content,
483 it is expected to respond with an "error" status.
484 packet: git< status=error
486 packet: git< 0000
487 If the filter experiences an error during processing, then it can
489 send the status "error" after the content was (partially or
490 completely) sent.
491 packet: git< status=success
493 packet: git< 0000
494 packet: git< HALF_WRITTEN_ERRONEOUS_CONTENT
495 packet: git< 0000
496 packet: git< status=error
497 packet: git< 0000
498 In case the filter cannot or does not want to process the content
500 as well as any future content for the lifetime of the Git process,
501 then it is expected to respond with an "abort" status at any point
502 in the protocol.
503 packet: git< status=abort
505 packet: git< 0000
506 Git neither stops nor restarts the filter process in case the
508 "error"/"abort" status is set. However, Git sets its exit code
509 according to the filter.<driver>.required flag, mimicking the
510 behavior of the filter.<driver>.clean / filter.<driver>.smudge
511 mechanism.
512 If the filter dies during the communication or does not adhere to
514 the protocol then Git will stop the filter process and restart it
515 with the next file that needs to be processed. Depending on the
516 filter.<driver>.required flag Git will interpret that as error.
517 Delay
519 If the filter supports the "delay" capability, then Git can send
520 the flag "can-delay" after the filter command and pathname. This
521 flag denotes that the filter can delay filtering the current blob
522 (e.g. to compensate network latencies) by responding with no
523 content but with the status "delayed" and a flush packet.
524 packet: git> command=smudge
526 packet: git> pathname=path/testfile.dat
527 packet: git> can-delay=1
528 packet: git> 0000
529 packet: git> CONTENT
530 packet: git> 0000
531 packet: git< status=delayed
532 packet: git< 0000
533 If the filter supports the "delay" capability then it must support
535 the "list_available_blobs" command. If Git sends this command, then
536 the filter is expected to return a list of pathnames representing
537 blobs that have been delayed earlier and are now available. The
538 list must be terminated with a flush packet followed by a "success"
539 status that is also terminated with a flush packet. If no blobs for
540 the delayed paths are available, yet, then the filter is expected
541 to block the response until at least one blob becomes available.
542 The filter can tell Git that it has no more delayed blobs by
543 sending an empty list. As soon as the filter responds with an empty
544 list, Git stops asking. All blobs that Git has not received at this
545 point are considered missing and will result in an error.
546 packet: git> command=list_available_blobs
548 packet: git> 0000
549 packet: git< pathname=path/testfile.dat
550 packet: git< pathname=path/otherfile.dat
551 packet: git< 0000
552 packet: git< status=success
553 packet: git< 0000
554 After Git received the pathnames, it will request the corresponding
556 blobs again. These requests contain a pathname and an empty content
557 section. The filter is expected to respond with the smudged content
558 in the usual way as explained above.
559 packet: git> command=smudge
561 packet: git> pathname=path/testfile.dat
562 packet: git> 0000
563 packet: git> 0000 # empty content!
564 packet: git< status=success
565 packet: git< 0000
566 packet: git< SMUDGED_CONTENT
567 packet: git< 0000
568 packet: git< 0000 # empty list, keep "status=success" unchanged!
569 Example
571 A long running filter demo implementation can be found in
572 contrib/long-running-filter/example.pl located in the Git core
573 repository. If you develop your own long running filter process
574 then the GIT_TRACE_PACKET environment variables can be very helpful
575 for debugging (see git(1)).
576 Please note that you cannot use an existing filter.<driver>.clean
578 or filter.<driver>.smudge command with filter.<driver>.process
579 because the former two use a different inter process communication
580 protocol than the latter one.
581 Interaction between checkin/checkout attributes
583 In the check-in codepath, the worktree file is first converted with
584 filter driver (if specified and corresponding driver defined), then
585 the result is processed with ident (if specified), and then finally
586 with text (again, if specified and applicable).
587 In the check-out codepath, the blob content is first converted with
589 text, and then ident and fed to filter.
590 Merging branches with differing checkin/checkout attributes
592 If you have added attributes to a file that cause the canonical
593 repository format for that file to change, such as adding a
594 clean/smudge filter or text/eol/ident attributes, merging anything
595 where the attribute is not in place would normally cause merge
596 conflicts.
597 To prevent these unnecessary merge conflicts, Git can be told to
599 run a virtual check-out and check-in of all three stages of a file
600 when resolving a three-way merge by setting the merge.renormalize
601 configuration variable. This prevents changes caused by check-in
602 conversion from causing spurious merge conflicts when a converted
603 file is merged with an unconverted file.
604 As long as a "smudge→clean" results in the same output as a "clean"
606 even on files that are already smudged, this strategy will
607 automatically resolve all filter-related conflicts. Filters that do
608 not act in this way may cause additional merge conflicts that must
609 be resolved manually.
610 Generating diff text
612 diff
613 The attribute diff affects how Git generates diffs for particular
614 files. It can tell Git whether to generate a textual patch for the
615 path or to treat the path as a binary file. It can also affect what
616 line is shown on the hunk header @@ -k,l +n,m @@ line, tell Git to
617 use an external command to generate the diff, or ask Git to convert
618 binary files to a text format before generating the diff.
619 Set
621 A path to which the diff attribute is set is treated as text,
622 even when they contain byte values that normally never appear
623 in text files, such as NUL.
624 Unset
626 A path to which the diff attribute is unset will generate
627 Binary files differ (or a binary patch, if binary patches are
628 enabled).
629 Unspecified
631 A path to which the diff attribute is unspecified first gets
632 its contents inspected, and if it looks like text and is
633 smaller than core.bigFileThreshold, it is treated as text.
634 Otherwise it would generate Binary files differ.
635 String
637 Diff is shown using the specified diff driver. Each driver may
638 specify one or more options, as described in the following
639 section. The options for the diff driver "foo" are defined by
640 the configuration variables in the "diff.foo" section of the
641 Git config file.
642 Defining an external diff driver
644 The definition of a diff driver is done in gitconfig, not
645 gitattributes file, so strictly speaking this manual page is a
646 wrong place to talk about it. However...
647 To define an external diff driver jcdiff, add a section to your
649 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
650 [diff "jcdiff"]
652 command = j-c-diff
653 When Git needs to show you a diff for the path with diff attribute
655 set to jcdiff, it calls the command you specified with the above
656 configuration, i.e. j-c-diff, with 7 parameters, just like
657 GIT_EXTERNAL_DIFF program is called. See git(1) for details.
658 Defining a custom hunk-header
660 Each group of changes (called a "hunk") in the textual diff output
661 is prefixed with a line of the form:
662 @@ -k,l +n,m @@ TEXT
664 This is called a hunk header. The "TEXT" portion is by default a
666 line that begins with an alphabet, an underscore or a dollar sign;
667 this matches what GNU diff -p output uses. This default selection
668 however is not suited for some contents, and you can use a
669 customized pattern to make a selection.
670 First, in .gitattributes, you would assign the diff attribute for
672 paths.
673 *.tex diff=tex
675 Then, you would define a "diff.tex.xfuncname" configuration to
677 specify a regular expression that matches a line that you would
678 want to appear as the hunk header "TEXT". Add a section to your
679 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
680 [diff "tex"]
682 xfuncname = "^(\\\\(sub)*section\\{.*)$"
683 Note. A single level of backslashes are eaten by the configuration
685 file parser, so you would need to double the backslashes; the
686 pattern above picks a line that begins with a backslash, and zero
687 or more occurrences of sub followed by section followed by open
688 brace, to the end of line.
689 There are a few built-in patterns to make this easier, and tex is
691 one of them, so you do not have to write the above in your
692 configuration file (you still need to enable this with the
693 attribute mechanism, via .gitattributes). The following built in
694 patterns are available:
695 • ada suitable for source code in the Ada language.
697 • bash suitable for source code in the Bourne-Again SHell
699 language. Covers a superset of POSIX shell function
700 definitions.
701 • bibtex suitable for files with BibTeX coded references.
703 • cpp suitable for source code in the C and C++ languages.
705 • csharp suitable for source code in the C# language.
707 • css suitable for cascading style sheets.
709 • dts suitable for devicetree (DTS) files.
711 • elixir suitable for source code in the Elixir language.
713 • fortran suitable for source code in the Fortran language.
715 • fountain suitable for Fountain documents.
717 • golang suitable for source code in the Go language.
719 • html suitable for HTML/XHTML documents.
721 • java suitable for source code in the Java language.
723 • kotlin suitable for source code in the Kotlin language.
725 • markdown suitable for Markdown documents.
727 • matlab suitable for source code in the MATLAB and Octave
729 languages.
730 • objc suitable for source code in the Objective-C language.
732 • pascal suitable for source code in the Pascal/Delphi language.
734 • perl suitable for source code in the Perl language.
736 • php suitable for source code in the PHP language.
738 • python suitable for source code in the Python language.
740 • ruby suitable for source code in the Ruby language.
742 • rust suitable for source code in the Rust language.
744 • scheme suitable for source code in the Scheme language.
746 • tex suitable for source code for LaTeX documents.
748 Customizing word diff
750 You can customize the rules that git diff --word-diff uses to split
751 words in a line, by specifying an appropriate regular expression in
752 the "diff.*.wordRegex" configuration variable. For example, in TeX
753 a backslash followed by a sequence of letters forms a command, but
754 several such commands can be run together without intervening
755 whitespace. To separate them, use a regular expression in your
756 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
757 [diff "tex"]
759 wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+"
760 A built-in pattern is provided for all languages listed in the
762 previous section.
763 Performing text diffs of binary files
765 Sometimes it is desirable to see the diff of a text-converted
766 version of some binary files. For example, a word processor
767 document can be converted to an ASCII text representation, and the
768 diff of the text shown. Even though this conversion loses some
769 information, the resulting diff is useful for human viewing (but
770 cannot be applied directly).
771 The textconv config option is used to define a program for
773 performing such a conversion. The program should take a single
774 argument, the name of a file to convert, and produce the resulting
775 text on stdout.
776 For example, to show the diff of the exif information of a file
778 instead of the binary information (assuming you have the exif tool
779 installed), add the following section to your $GIT_DIR/config file
780 (or $HOME/.gitconfig file):
781 [diff "jpg"]
783 textconv = exif
784 Note
787 The text conversion is generally a one-way conversion; in this
788 example, we lose the actual image contents and focus just on
789 the text data. This means that diffs generated by textconv are
790 not suitable for applying. For this reason, only git diff and
791 the git log family of commands (i.e., log, whatchanged, show)
792 will perform text conversion. git format-patch will never
793 generate this output. If you want to send somebody a
794 text-converted diff of a binary file (e.g., because it quickly
795 conveys the changes you have made), you should generate it
796 separately and send it as a comment in addition to the usual
797 binary diff that you might send.
798 Because text conversion can be slow, especially when doing a large
800 number of them with git log -p, Git provides a mechanism to cache
801 the output and use it in future diffs. To enable caching, set the
802 "cachetextconv" variable in your diff driver’s config. For example:
803 [diff "jpg"]
805 textconv = exif
806 cachetextconv = true
807 This will cache the result of running "exif" on each blob
809 indefinitely. If you change the textconv config variable for a diff
810 driver, Git will automatically invalidate the cache entries and
811 re-run the textconv filter. If you want to invalidate the cache
812 manually (e.g., because your version of "exif" was updated and now
813 produces better output), you can remove the cache manually with git
814 update-ref -d refs/notes/textconv/jpg (where "jpg" is the name of
815 the diff driver, as in the example above).
816 Choosing textconv versus external diff
818 If you want to show differences between binary or
819 specially-formatted blobs in your repository, you can choose to use
820 either an external diff command, or to use textconv to convert them
821 to a diff-able text format. Which method you choose depends on your
822 exact situation.
823 The advantage of using an external diff command is flexibility. You
825 are not bound to find line-oriented changes, nor is it necessary
826 for the output to resemble unified diff. You are free to locate and
827 report changes in the most appropriate way for your data format.
828 A textconv, by comparison, is much more limiting. You provide a
830 transformation of the data into a line-oriented text format, and
831 Git uses its regular diff tools to generate the output. There are
832 several advantages to choosing this method:
833 1. Ease of use. It is often much simpler to write a binary to text
835 transformation than it is to perform your own diff. In many
836 cases, existing programs can be used as textconv filters (e.g.,
837 exif, odt2txt).
838 2. Git diff features. By performing only the transformation step
840 yourself, you can still utilize many of Git’s diff features,
841 including colorization, word-diff, and combined diffs for
842 merges.
843 3. Caching. Textconv caching can speed up repeated diffs, such as
845 those you might trigger by running git log -p.
846 Marking files as binary
848 Git usually guesses correctly whether a blob contains text or
849 binary data by examining the beginning of the contents. However,
850 sometimes you may want to override its decision, either because a
851 blob contains binary data later in the file, or because the
852 content, while technically composed of text characters, is opaque
853 to a human reader. For example, many postscript files contain only
854 ASCII characters, but produce noisy and meaningless diffs.
855 The simplest way to mark a file as binary is to unset the diff
857 attribute in the .gitattributes file:
858 *.ps -diff
860 This will cause Git to generate Binary files differ (or a binary
862 patch, if binary patches are enabled) instead of a regular diff.
863 However, one may also want to specify other diff driver attributes.
865 For example, you might want to use textconv to convert postscript
866 files to an ASCII representation for human viewing, but otherwise
867 treat them as binary files. You cannot specify both -diff and
868 diff=ps attributes. The solution is to use the diff.*.binary config
869 option:
870 [diff "ps"]
872 textconv = ps2ascii
873 binary = true
874 Performing a three-way merge
876 merge
877 The attribute merge affects how three versions of a file are merged
878 when a file-level merge is necessary during git merge, and other
879 commands such as git revert and git cherry-pick.
880 Set
882 Built-in 3-way merge driver is used to merge the contents in a
883 way similar to merge command of RCS suite. This is suitable for
884 ordinary text files.
885 Unset
887 Take the version from the current branch as the tentative merge
888 result, and declare that the merge has conflicts. This is
889 suitable for binary files that do not have a well-defined merge
890 semantics.
891 Unspecified
893 By default, this uses the same built-in 3-way merge driver as
894 is the case when the merge attribute is set. However, the
895 merge.default configuration variable can name different merge
896 driver to be used with paths for which the merge attribute is
897 unspecified.
898 String
900 3-way merge is performed using the specified custom merge
901 driver. The built-in 3-way merge driver can be explicitly
902 specified by asking for "text" driver; the built-in "take the
903 current branch" driver can be requested with "binary".
904 Built-in merge drivers
906 There are a few built-in low-level merge drivers defined that can
907 be asked for via the merge attribute.
908 text
910 Usual 3-way file level merge for text files. Conflicted regions
911 are marked with conflict markers <<<<<<<, ======= and >>>>>>>.
912 The version from your branch appears before the ======= marker,
913 and the version from the merged branch appears after the
914 ======= marker.
915 binary
917 Keep the version from your branch in the work tree, but leave
918 the path in the conflicted state for the user to sort out.
919 union
921 Run 3-way file level merge for text files, but take lines from
922 both versions, instead of leaving conflict markers. This tends
923 to leave the added lines in the resulting file in random order
924 and the user should verify the result. Do not use this if you
925 do not understand the implications.
926 Defining a custom merge driver
928 The definition of a merge driver is done in the .git/config file,
929 not in the gitattributes file, so strictly speaking this manual
930 page is a wrong place to talk about it. However...
931 To define a custom merge driver filfre, add a section to your
933 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
934 [merge "filfre"]
936 name = feel-free merge driver
937 driver = filfre %O %A %B %L %P
938 recursive = binary
939 The merge.*.name variable gives the driver a human-readable name.
941 The merge.*.driver variable’s value is used to construct a command
943 to run to merge ancestor’s version (%O), current version (%A) and
944 the other branches' version (%B). These three tokens are replaced
945 with the names of temporary files that hold the contents of these
946 versions when the command line is built. Additionally, %L will be
947 replaced with the conflict marker size (see below).
948 The merge driver is expected to leave the result of the merge in
950 the file named with %A by overwriting it, and exit with zero status
951 if it managed to merge them cleanly, or non-zero if there were
952 conflicts.
953 The merge.*.recursive variable specifies what other merge driver to
955 use when the merge driver is called for an internal merge between
956 common ancestors, when there are more than one. When left
957 unspecified, the driver itself is used for both internal merge and
958 the final merge.
959 The merge driver can learn the pathname in which the merged result
961 will be stored via placeholder %P.
962 conflict-marker-size
964 This attribute controls the length of conflict markers left in the
965 work tree file during a conflicted merge. Only setting to the value
966 to a positive integer has any meaningful effect.
967 For example, this line in .gitattributes can be used to tell the
969 merge machinery to leave much longer (instead of the usual
970 7-character-long) conflict markers when merging the file
971 Documentation/git-merge.txt results in a conflict.
972 Documentation/git-merge.txt conflict-marker-size=32
974 Checking whitespace errors
976 whitespace
977 The core.whitespace configuration variable allows you to define
978 what diff and apply should consider whitespace errors for all paths
979 in the project (See git-config(1)). This attribute gives you finer
980 control per path.
981 Set
983 Notice all types of potential whitespace errors known to Git.
984 The tab width is taken from the value of the core.whitespace
985 configuration variable.
986 Unset
988 Do not notice anything as error.
989 Unspecified
991 Use the value of the core.whitespace configuration variable to
992 decide what to notice as error.
993 String
995 Specify a comma separate list of common whitespace problems to
996 notice in the same format as the core.whitespace configuration
997 variable.
998 Creating an archive
1000 export-ignore
1001 Files and directories with the attribute export-ignore won’t be
1002 added to archive files.
1003 export-subst
1005 If the attribute export-subst is set for a file then Git will
1006 expand several placeholders when adding this file to an archive.
1007 The expansion depends on the availability of a commit ID, i.e., if
1008 git-archive(1) has been given a tree instead of a commit or a tag
1009 then no replacement will be done. The placeholders are the same as
1010 those for the option --pretty=format: of git-log(1), except that
1011 they need to be wrapped like this: $Format:PLACEHOLDERS$ in the
1012 file. E.g. the string $Format:%H$ will be replaced by the commit
1013 hash. However, only one %(describe) placeholder is expanded per
1014 archive to avoid denial-of-service attacks.
1015 Packing objects
1017 delta
1018 Delta compression will not be attempted for blobs for paths with
1019 the attribute delta set to false.
1020 Viewing files in GUI tools
1022 encoding
1023 The value of this attribute specifies the character encoding that
1024 should be used by GUI tools (e.g. gitk(1) and git-gui(1)) to
1025 display the contents of the relevant file. Note that due to
1026 performance considerations gitk(1) does not use this attribute
1027 unless you manually enable per-file encodings in its options.
1028 If this attribute is not set or has an invalid value, the value of
1030 the gui.encoding configuration variable is used instead (See git-
1031 config(1)).
1032
1034 You do not want any end-of-line conversions applied to, nor textual
1035 diffs produced for, any binary file you track. You would need to
1036 specify e.g.
1037 *.jpg -text -diff
1039 but that may become cumbersome, when you have many attributes. Using
1041 macro attributes, you can define an attribute that, when set, also sets
1042 or unsets a number of other attributes at the same time. The system
1043 knows a built-in macro attribute, binary:
1044 *.jpg binary
1046 Setting the "binary" attribute also unsets the "text" and "diff"
1048 attributes as above. Note that macro attributes can only be "Set",
1049 though setting one might have the effect of setting or unsetting other
1050 attributes or even returning other attributes to the "Unspecified"
1051 state.
1052
1054 Custom macro attributes can be defined only in top-level gitattributes
1055 files ($GIT_DIR/info/attributes, the .gitattributes file at the top
1056 level of the working tree, or the global or system-wide gitattributes
1057 files), not in .gitattributes files in working tree subdirectories. The
1058 built-in macro attribute "binary" is equivalent to:
1059 [attr]binary -diff -merge -text
1061
1063 Git does not follow symbolic links when accessing a .gitattributes file
1064 in the working tree. This keeps behavior consistent when the file is
1065 accessed from the index or a tree versus from the filesystem.
1066
1068 If you have these three gitattributes file:
1069 (in $GIT_DIR/info/attributes)
1071 a* foo !bar -baz
1073 (in .gitattributes)
1075 abc foo bar baz
1076 (in t/.gitattributes)
1078 ab* merge=filfre
1079 abc -foo -bar
1080 *.c frotz
1081 the attributes given to path t/abc are computed as follows:
1083 1. By examining t/.gitattributes (which is in the same directory as
1085 the path in question), Git finds that the first line matches.
1086 merge attribute is set. It also finds that the second line matches,
1087 and attributes foo and bar are unset.
1088 2. Then it examines .gitattributes (which is in the parent directory),
1090 and finds that the first line matches, but t/.gitattributes file
1091 already decided how merge, foo and bar attributes should be given
1092 to this path, so it leaves foo and bar unset. Attribute baz is set.
1093 3. Finally it examines $GIT_DIR/info/attributes. This file is used to
1095 override the in-tree settings. The first line is a match, and foo
1096 is set, bar is reverted to unspecified state, and baz is unset.
1097 As the result, the attributes assignment to t/abc becomes:
1099 foo set to true
1101 bar unspecified
1102 baz set to false
1103 merge set to string value "filfre"
1104 frotz unspecified
1105
1107 git-check-attr(1).
1108
1110 Part of the git(1) suite
1111Git 2.39.1 2023-01-13 GITATTRIBUTES(5)