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 checkout and
98 git merge run. They also affect how Git stores the contents you prepare
99 in the working tree in the repository upon git add and git commit.
100 text
102 This attribute enables and controls end-of-line normalization. When
103 a text file is normalized, its line endings are converted to LF in
104 the repository. To control what line ending style is used in the
105 working directory, use the eol attribute for a single file and the
106 core.eol configuration variable for all text files. Note that
107 core.autocrlf overrides core.eol
108 Set
110 Setting the text attribute on a path enables end-of-line
111 normalization and marks the path as a text file. End-of-line
112 conversion takes place without guessing the content type.
113 Unset
115 Unsetting the text attribute on a path tells Git not to attempt
116 any end-of-line conversion upon checkin or checkout.
117 Set to string value "auto"
119 When text is set to "auto", the path is marked for automatic
120 end-of-line conversion. If Git decides that the content is
121 text, its line endings are converted to LF on checkin. When the
122 file has been committed with CRLF, no conversion is done.
123 Unspecified
125 If the text attribute is unspecified, Git uses the
126 core.autocrlf configuration variable to determine if the file
127 should be converted.
128 Any other value causes Git to act as if text has been left
130 unspecified.
131 eol
133 This attribute sets a specific line-ending style to be used in the
134 working directory. It enables end-of-line conversion without any
135 content checks, effectively setting the text attribute. Note that
136 setting this attribute on paths which are in the index with CRLF
137 line endings may make the paths to be considered dirty. Adding the
138 path to the index again will normalize the line endings in the
139 index.
140 Set to string value "crlf"
142 This setting forces Git to normalize line endings for this file
143 on checkin and convert them to CRLF when the file is checked
144 out.
145 Set to string value "lf"
147 This setting forces Git to normalize line endings to LF on
148 checkin and prevents conversion to CRLF when the file is
149 checked out.
150 Backwards compatibility with crlf attribute
152 For backwards compatibility, the crlf attribute is interpreted as
153 follows:
154 crlf text
156 -crlf -text
157 crlf=input eol=lf
158 End-of-line conversion
161 While Git normally leaves file contents alone, it can be configured
162 to normalize line endings to LF in the repository and, optionally,
163 to convert them to CRLF when files are checked out.
164 If you simply want to have CRLF line endings in your working
166 directory regardless of the repository you are working with, you
167 can set the config variable "core.autocrlf" without using any
168 attributes.
169 [core]
171 autocrlf = true
172 This does not force normalization of text files, but does ensure
175 that text files that you introduce to the repository have their
176 line endings normalized to LF when they are added, and that files
177 that are already normalized in the repository stay normalized.
178 If you want to ensure that text files that any contributor
180 introduces to the repository have their line endings normalized,
181 you can set the text attribute to "auto" for all files.
182 * text=auto
184 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,
213 unset their text attribute before running git add -u.
214 manual.pdf -text
216 Conversely, text files that Git does not detect can have
219 normalization enabled manually.
220 weirdchars.txt text
222 If core.safecrlf is set to "true" or "warn", Git verifies if the
225 conversion is reversible for the current setting of core.autocrlf.
226 For "true", Git rejects irreversible conversions; for "warn", Git
227 only prints a warning but accepts an irreversible conversion. The
228 safety triggers to prevent such a conversion done to the files in
229 the work tree, but there are a few exceptions. Even though...
230 · git add itself does not touch the files in the work tree, the
232 next checkout would, so the safety triggers;
233 · git apply to update a text file with a patch does touch the
235 files in the work tree, but the operation is about text files
236 and CRLF conversion is about fixing the line ending
237 inconsistencies, so the safety does not trigger;
238 · git diff itself does not touch the files in the work tree, it
240 is often run to inspect the changes you intend to next git add.
241 To catch potential problems early, safety triggers.
242 working-tree-encoding
244 Git recognizes files encoded in ASCII or one of its supersets (e.g.
245 UTF-8, ISO-8859-1, ...) as text files. Files encoded in certain
246 other encodings (e.g. UTF-16) are interpreted as binary and
247 consequently built-in Git text processing tools (e.g. git diff) as
248 well as most Git web front ends do not visualize the contents of
249 these files by default.
250 In these cases you can tell Git the encoding of a file in the
252 working directory with the working-tree-encoding attribute. If a
253 file with this attribute is added to Git, then Git reencodes the
254 content from the specified encoding to UTF-8. Finally, Git stores
255 the UTF-8 encoded content in its internal data structure (called
256 "the index"). On checkout the content is reencoded back to the
257 specified encoding.
258 Please note that using the working-tree-encoding attribute may have
260 a number of pitfalls:
261 · Alternative Git implementations (e.g. JGit or libgit2) and
263 older Git versions (as of March 2018) do not support the
264 working-tree-encoding attribute. If you decide to use the
265 working-tree-encoding attribute in your repository, then it is
266 strongly recommended to ensure that all clients working with
267 the repository support it.
268 For example, Microsoft Visual Studio resources files (*.rc) or
270 PowerShell script files (*.ps1) are sometimes encoded in
271 UTF-16. If you declare *.ps1 as files as UTF-16 and you add
272 foo.ps1 with a working-tree-encoding enabled Git client, then
273 foo.ps1 will be stored as UTF-8 internally. A client without
274 working-tree-encoding support will checkout foo.ps1 as UTF-8
275 encoded file. This will typically cause trouble for the users
276 of this file.
277 If a Git client, that does not support the
279 working-tree-encoding attribute, adds a new file bar.ps1, then
280 bar.ps1 will be stored "as-is" internally (in this example
281 probably as UTF-16). A client with working-tree-encoding
282 support will interpret the internal contents as UTF-8 and try
283 to convert it to UTF-16 on checkout. That operation will fail
284 and cause an error.
285 · Reencoding content to non-UTF encodings can cause errors as the
287 conversion might not be UTF-8 round trip safe. If you suspect
288 your encoding to not be round trip safe, then add it to
289 core.checkRoundtripEncoding to make Git check the round trip
290 encoding (see git-config(1)). SHIFT-JIS (Japanese character
291 set) is known to have round trip issues with UTF-8 and is
292 checked by default.
293 · Reencoding content requires resources that might slow down
295 certain Git operations (e.g git checkout or git add).
296 Use the working-tree-encoding attribute only if you cannot store a
298 file in UTF-8 encoding and if you want Git to be able to process
299 the content as text.
300 As an example, use the following attributes if your *.ps1 files are
302 UTF-16 encoded with byte order mark (BOM) and you want Git to
303 perform automatic line ending conversion based on your platform.
304 *.ps1 text working-tree-encoding=UTF-16
306 Use the following attributes if your *.ps1 files are UTF-16 little
309 endian encoded without BOM and you want Git to use Windows line
310 endings in the working directory. Please note, it is highly
311 recommended to explicitly define the line endings with eol if the
312 working-tree-encoding attribute is used to avoid ambiguity.
313 *.ps1 text working-tree-encoding=UTF-16LE eol=CRLF
315 You can get a list of all available encodings on your platform with
318 the following command:
319 iconv --list
321 If you do not know the encoding of a file, then you can use the
324 file command to guess the encoding:
325 file foo.ps1
327 ident
330 When the attribute ident is set for a path, Git replaces $Id$ in
331 the blob object with $Id:, followed by the 40-character hexadecimal
332 blob object name, followed by a dollar sign $ upon checkout. Any
333 byte sequence that begins with $Id: and ends with $ in the worktree
334 file is replaced with $Id$ upon check-in.
335 filter
337 A filter attribute can be set to a string value that names a filter
338 driver specified in the configuration.
339 A filter driver consists of a clean command and a smudge command,
341 either of which can be left unspecified. Upon checkout, when the
342 smudge command is specified, the command is fed the blob object
343 from its standard input, and its standard output is used to update
344 the worktree file. Similarly, the clean command is used to convert
345 the contents of worktree file upon checkin. By default these
346 commands process only a single blob and terminate. If a long
347 running process filter is used in place of clean and/or smudge
348 filters, then Git can process all blobs with a single filter
349 command invocation for the entire life of a single Git command, for
350 example git add --all. If a long running process filter is
351 configured then it always takes precedence over a configured single
352 blob filter. See section below for the description of the protocol
353 used to communicate with a process filter.
354 One use of the content filtering is to massage the content into a
356 shape that is more convenient for the platform, filesystem, and the
357 user to use. For this mode of operation, the key phrase here is
358 "more convenient" and not "turning something unusable into usable".
359 In other words, the intent is that if someone unsets the filter
360 driver definition, or does not have the appropriate filter program,
361 the project should still be usable.
362 Another use of the content filtering is to store the content that
364 cannot be directly used in the repository (e.g. a UUID that refers
365 to the true content stored outside Git, or an encrypted content)
366 and turn it into a usable form upon checkout (e.g. download the
367 external content, or decrypt the encrypted content).
368 These two filters behave differently, and by default, a filter is
370 taken as the former, massaging the contents into more convenient
371 shape. A missing filter driver definition in the config, or a
372 filter driver that exits with a non-zero status, is not an error
373 but makes the filter a no-op passthru.
374 You can declare that a filter turns a content that by itself is
376 unusable into a usable content by setting the
377 filter.<driver>.required configuration variable to true.
378 Note: Whenever the clean filter is changed, the repo should be
380 renormalized: $ git add --renormalize .
381 For example, in .gitattributes, you would assign the filter
383 attribute for paths.
384 *.c filter=indent
386 Then you would define a "filter.indent.clean" and
389 "filter.indent.smudge" configuration in your .git/config to specify
390 a pair of commands to modify the contents of C programs when the
391 source files are checked in ("clean" is run) and checked out (no
392 change is made because the command is "cat").
393 [filter "indent"]
395 clean = indent
396 smudge = cat
397 For best results, clean should not alter its output further if it
400 is run twice ("clean→clean" should be equivalent to "clean"), and
401 multiple smudge commands should not alter clean's output
402 ("smudge→smudge→clean" should be equivalent to "clean"). See the
403 section on merging below.
404 The "indent" filter is well-behaved in this regard: it will not
406 modify input that is already correctly indented. In this case, the
407 lack of a smudge filter means that the clean filter must accept its
408 own output without modifying it.
409 If a filter must succeed in order to make the stored contents
411 usable, you can declare that the filter is required, in the
412 configuration:
413 [filter "crypt"]
415 clean = openssl enc ...
416 smudge = openssl enc -d ...
417 required
418 Sequence "%f" on the filter command line is replaced with the name
421 of the file the filter is working on. A filter might use this in
422 keyword substitution. For example:
423 [filter "p4"]
425 clean = git-p4-filter --clean %f
426 smudge = git-p4-filter --smudge %f
427 Note that "%f" is the name of the path that is being worked on.
430 Depending on the version that is being filtered, the corresponding
431 file on disk may not exist, or may have different contents. So,
432 smudge and clean commands should not try to access the file on
433 disk, but only act as filters on the content provided to them on
434 standard input.
435 Long Running Filter Process
437 If the filter command (a string value) is defined via
438 filter.<driver>.process then Git can process all blobs with a
439 single filter invocation for the entire life of a single Git
440 command. This is achieved by using the long-running process
441 protocol (described in
442 technical/long-running-process-protocol.txt).
443 When Git encounters the first file that needs to be cleaned or
445 smudged, it starts the filter and performs the handshake. In the
446 handshake, the welcome message sent by Git is "git-filter-client",
447 only version 2 is suppported, and the supported capabilities are
448 "clean", "smudge", and "delay".
449 Afterwards Git sends a list of "key=value" pairs terminated with a
451 flush packet. The list will contain at least the filter command
452 (based on the supported capabilities) and the pathname of the file
453 to filter relative to the repository root. Right after the flush
454 packet Git sends the content split in zero or more pkt-line packets
455 and a flush packet to terminate content. Please note, that the
456 filter must not send any response before it received the content
457 and the final flush packet. Also note that the "value" of a
458 "key=value" pair can contain the "=" character whereas the key
459 would never contain that character.
460 packet: git> command=smudge
462 packet: git> pathname=path/testfile.dat
463 packet: git> 0000
464 packet: git> CONTENT
465 packet: git> 0000
466 The filter is expected to respond with a list of "key=value" pairs
469 terminated with a flush packet. If the filter does not experience
470 problems then the list must contain a "success" status. Right after
471 these packets the filter is expected to send the content in zero or
472 more pkt-line packets and a flush packet at the end. Finally, a
473 second list of "key=value" pairs terminated with a flush packet is
474 expected. The filter can change the status in the second list or
475 keep the status as is with an empty list. Please note that the
476 empty list must be terminated with a flush packet regardless.
477 packet: git< status=success
479 packet: git< 0000
480 packet: git< SMUDGED_CONTENT
481 packet: git< 0000
482 packet: git< 0000 # empty list, keep "status=success" unchanged!
483 If the result content is empty then the filter is expected to
486 respond with a "success" status and a flush packet to signal the
487 empty content.
488 packet: git< status=success
490 packet: git< 0000
491 packet: git< 0000 # empty content!
492 packet: git< 0000 # empty list, keep "status=success" unchanged!
493 In case the filter cannot or does not want to process the content,
496 it is expected to respond with an "error" status.
497 packet: git< status=error
499 packet: git< 0000
500 If the filter experiences an error during processing, then it can
503 send the status "error" after the content was (partially or
504 completely) sent.
505 packet: git< status=success
507 packet: git< 0000
508 packet: git< HALF_WRITTEN_ERRONEOUS_CONTENT
509 packet: git< 0000
510 packet: git< status=error
511 packet: git< 0000
512 In case the filter cannot or does not want to process the content
515 as well as any future content for the lifetime of the Git process,
516 then it is expected to respond with an "abort" status at any point
517 in the protocol.
518 packet: git< status=abort
520 packet: git< 0000
521 Git neither stops nor restarts the filter process in case the
524 "error"/"abort" status is set. However, Git sets its exit code
525 according to the filter.<driver>.required flag, mimicking the
526 behavior of the filter.<driver>.clean / filter.<driver>.smudge
527 mechanism.
528 If the filter dies during the communication or does not adhere to
530 the protocol then Git will stop the filter process and restart it
531 with the next file that needs to be processed. Depending on the
532 filter.<driver>.required flag Git will interpret that as error.
533 Delay
535 If the filter supports the "delay" capability, then Git can send
536 the flag "can-delay" after the filter command and pathname. This
537 flag denotes that the filter can delay filtering the current blob
538 (e.g. to compensate network latencies) by responding with no
539 content but with the status "delayed" and a flush packet.
540 packet: git> command=smudge
542 packet: git> pathname=path/testfile.dat
543 packet: git> can-delay=1
544 packet: git> 0000
545 packet: git> CONTENT
546 packet: git> 0000
547 packet: git< status=delayed
548 packet: git< 0000
549 If the filter supports the "delay" capability then it must support
552 the "list_available_blobs" command. If Git sends this command, then
553 the filter is expected to return a list of pathnames representing
554 blobs that have been delayed earlier and are now available. The
555 list must be terminated with a flush packet followed by a "success"
556 status that is also terminated with a flush packet. If no blobs for
557 the delayed paths are available, yet, then the filter is expected
558 to block the response until at least one blob becomes available.
559 The filter can tell Git that it has no more delayed blobs by
560 sending an empty list. As soon as the filter responds with an empty
561 list, Git stops asking. All blobs that Git has not received at this
562 point are considered missing and will result in an error.
563 packet: git> command=list_available_blobs
565 packet: git> 0000
566 packet: git< pathname=path/testfile.dat
567 packet: git< pathname=path/otherfile.dat
568 packet: git< 0000
569 packet: git< status=success
570 packet: git< 0000
571 After Git received the pathnames, it will request the corresponding
574 blobs again. These requests contain a pathname and an empty content
575 section. The filter is expected to respond with the smudged content
576 in the usual way as explained above.
577 packet: git> command=smudge
579 packet: git> pathname=path/testfile.dat
580 packet: git> 0000
581 packet: git> 0000 # empty content!
582 packet: git< status=success
583 packet: git< 0000
584 packet: git< SMUDGED_CONTENT
585 packet: git< 0000
586 packet: git< 0000 # empty list, keep "status=success" unchanged!
587 Example
590 A long running filter demo implementation can be found in
591 contrib/long-running-filter/example.pl located in the Git core
592 repository. If you develop your own long running filter process
593 then the GIT_TRACE_PACKET environment variables can be very helpful
594 for debugging (see git(1)).
595 Please note that you cannot use an existing filter.<driver>.clean
597 or filter.<driver>.smudge command with filter.<driver>.process
598 because the former two use a different inter process communication
599 protocol than the latter one.
600 Interaction between checkin/checkout attributes
602 In the check-in codepath, the worktree file is first converted with
603 filter driver (if specified and corresponding driver defined), then
604 the result is processed with ident (if specified), and then finally
605 with text (again, if specified and applicable).
606 In the check-out codepath, the blob content is first converted with
608 text, and then ident and fed to filter.
609 Merging branches with differing checkin/checkout attributes
611 If you have added attributes to a file that cause the canonical
612 repository format for that file to change, such as adding a
613 clean/smudge filter or text/eol/ident attributes, merging anything
614 where the attribute is not in place would normally cause merge
615 conflicts.
616 To prevent these unnecessary merge conflicts, Git can be told to
618 run a virtual check-out and check-in of all three stages of a file
619 when resolving a three-way merge by setting the merge.renormalize
620 configuration variable. This prevents changes caused by check-in
621 conversion from causing spurious merge conflicts when a converted
622 file is merged with an unconverted file.
623 As long as a "smudge→clean" results in the same output as a "clean"
625 even on files that are already smudged, this strategy will
626 automatically resolve all filter-related conflicts. Filters that do
627 not act in this way may cause additional merge conflicts that must
628 be resolved manually.
629 Generating diff text
631 diff
632 The attribute diff affects how Git generates diffs for particular
633 files. It can tell Git whether to generate a textual patch for the
634 path or to treat the path as a binary file. It can also affect what
635 line is shown on the hunk header @@ -k,l +n,m @@ line, tell Git to
636 use an external command to generate the diff, or ask Git to convert
637 binary files to a text format before generating the diff.
638 Set
640 A path to which the diff attribute is set is treated as text,
641 even when they contain byte values that normally never appear
642 in text files, such as NUL.
643 Unset
645 A path to which the diff attribute is unset will generate
646 Binary files differ (or a binary patch, if binary patches are
647 enabled).
648 Unspecified
650 A path to which the diff attribute is unspecified first gets
651 its contents inspected, and if it looks like text and is
652 smaller than core.bigFileThreshold, it is treated as text.
653 Otherwise it would generate Binary files differ.
654 String
656 Diff is shown using the specified diff driver. Each driver may
657 specify one or more options, as described in the following
658 section. The options for the diff driver "foo" are defined by
659 the configuration variables in the "diff.foo" section of the
660 Git config file.
661 Defining an external diff driver
663 The definition of a diff driver is done in gitconfig, not
664 gitattributes file, so strictly speaking this manual page is a
665 wrong place to talk about it. However...
666 To define an external diff driver jcdiff, add a section to your
668 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
669 [diff "jcdiff"]
671 command = j-c-diff
672 When Git needs to show you a diff for the path with diff attribute
675 set to jcdiff, it calls the command you specified with the above
676 configuration, i.e. j-c-diff, with 7 parameters, just like
677 GIT_EXTERNAL_DIFF program is called. See git(1) for details.
678 Defining a custom hunk-header
680 Each group of changes (called a "hunk") in the textual diff output
681 is prefixed with a line of the form:
682 @@ -k,l +n,m @@ TEXT
684 This is called a hunk header. The "TEXT" portion is by default a
686 line that begins with an alphabet, an underscore or a dollar sign;
687 this matches what GNU diff -p output uses. This default selection
688 however is not suited for some contents, and you can use a
689 customized pattern to make a selection.
690 First, in .gitattributes, you would assign the diff attribute for
692 paths.
693 *.tex diff=tex
695 Then, you would define a "diff.tex.xfuncname" configuration to
698 specify a regular expression that matches a line that you would
699 want to appear as the hunk header "TEXT". Add a section to your
700 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
701 [diff "tex"]
703 xfuncname = "^(\\\\(sub)*section\\{.*)$"
704 Note. A single level of backslashes are eaten by the configuration
707 file parser, so you would need to double the backslashes; the
708 pattern above picks a line that begins with a backslash, and zero
709 or more occurrences of sub followed by section followed by open
710 brace, to the end of line.
711 There are a few built-in patterns to make this easier, and tex is
713 one of them, so you do not have to write the above in your
714 configuration file (you still need to enable this with the
715 attribute mechanism, via .gitattributes). The following built in
716 patterns are available:
717 · ada suitable for source code in the Ada language.
719 · bibtex suitable for files with BibTeX coded references.
721 · cpp suitable for source code in the C and C++ languages.
723 · csharp suitable for source code in the C# language.
725 · css suitable for cascading style sheets.
727 · fortran suitable for source code in the Fortran language.
729 · fountain suitable for Fountain documents.
731 · golang suitable for source code in the Go language.
733 · html suitable for HTML/XHTML documents.
735 · java suitable for source code in the Java language.
737 · matlab suitable for source code in the MATLAB language.
739 · objc suitable for source code in the Objective-C language.
741 · pascal suitable for source code in the Pascal/Delphi language.
743 · perl suitable for source code in the Perl language.
745 · php suitable for source code in the PHP language.
747 · python suitable for source code in the Python language.
749 · ruby suitable for source code in the Ruby language.
751 · tex suitable for source code for LaTeX documents.
753 Customizing word diff
755 You can customize the rules that git diff --word-diff uses to split
756 words in a line, by specifying an appropriate regular expression in
757 the "diff.*.wordRegex" configuration variable. For example, in TeX
758 a backslash followed by a sequence of letters forms a command, but
759 several such commands can be run together without intervening
760 whitespace. To separate them, use a regular expression in your
761 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
762 [diff "tex"]
764 wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+"
765 A built-in pattern is provided for all languages listed in the
768 previous section.
769 Performing text diffs of binary files
771 Sometimes it is desirable to see the diff of a text-converted
772 version of some binary files. For example, a word processor
773 document can be converted to an ASCII text representation, and the
774 diff of the text shown. Even though this conversion loses some
775 information, the resulting diff is useful for human viewing (but
776 cannot be applied directly).
777 The textconv config option is used to define a program for
779 performing such a conversion. The program should take a single
780 argument, the name of a file to convert, and produce the resulting
781 text on stdout.
782 For example, to show the diff of the exif information of a file
784 instead of the binary information (assuming you have the exif tool
785 installed), add the following section to your $GIT_DIR/config file
786 (or $HOME/.gitconfig file):
787 [diff "jpg"]
789 textconv = exif
790 Note
793 The text conversion is generally a one-way conversion; in this
794 example, we lose the actual image contents and focus just on
795 the text data. This means that diffs generated by textconv are
796 not suitable for applying. For this reason, only git diff and
797 the git log family of commands (i.e., log, whatchanged, show)
798 will perform text conversion. git format-patch will never
799 generate this output. If you want to send somebody a
800 text-converted diff of a binary file (e.g., because it quickly
801 conveys the changes you have made), you should generate it
802 separately and send it as a comment in addition to the usual
803 binary diff that you might send.
804 Because text conversion can be slow, especially when doing a large
806 number of them with git log -p, Git provides a mechanism to cache
807 the output and use it in future diffs. To enable caching, set the
808 "cachetextconv" variable in your diff driver’s config. For example:
809 [diff "jpg"]
811 textconv = exif
812 cachetextconv = true
813 This will cache the result of running "exif" on each blob
816 indefinitely. If you change the textconv config variable for a diff
817 driver, Git will automatically invalidate the cache entries and
818 re-run the textconv filter. If you want to invalidate the cache
819 manually (e.g., because your version of "exif" was updated and now
820 produces better output), you can remove the cache manually with git
821 update-ref -d refs/notes/textconv/jpg (where "jpg" is the name of
822 the diff driver, as in the example above).
823 Choosing textconv versus external diff
825 If you want to show differences between binary or
826 specially-formatted blobs in your repository, you can choose to use
827 either an external diff command, or to use textconv to convert them
828 to a diff-able text format. Which method you choose depends on your
829 exact situation.
830 The advantage of using an external diff command is flexibility. You
832 are not bound to find line-oriented changes, nor is it necessary
833 for the output to resemble unified diff. You are free to locate and
834 report changes in the most appropriate way for your data format.
835 A textconv, by comparison, is much more limiting. You provide a
837 transformation of the data into a line-oriented text format, and
838 Git uses its regular diff tools to generate the output. There are
839 several advantages to choosing this method:
840 1. Ease of use. It is often much simpler to write a binary to text
842 transformation than it is to perform your own diff. In many
843 cases, existing programs can be used as textconv filters (e.g.,
844 exif, odt2txt).
845 2. Git diff features. By performing only the transformation step
847 yourself, you can still utilize many of Git’s diff features,
848 including colorization, word-diff, and combined diffs for
849 merges.
850 3. Caching. Textconv caching can speed up repeated diffs, such as
852 those you might trigger by running git log -p.
853 Marking files as binary
855 Git usually guesses correctly whether a blob contains text or
856 binary data by examining the beginning of the contents. However,
857 sometimes you may want to override its decision, either because a
858 blob contains binary data later in the file, or because the
859 content, while technically composed of text characters, is opaque
860 to a human reader. For example, many postscript files contain only
861 ASCII characters, but produce noisy and meaningless diffs.
862 The simplest way to mark a file as binary is to unset the diff
864 attribute in the .gitattributes file:
865 *.ps -diff
867 This will cause Git to generate Binary files differ (or a binary
870 patch, if binary patches are enabled) instead of a regular diff.
871 However, one may also want to specify other diff driver attributes.
873 For example, you might want to use textconv to convert postscript
874 files to an ASCII representation for human viewing, but otherwise
875 treat them as binary files. You cannot specify both -diff and
876 diff=ps attributes. The solution is to use the diff.*.binary config
877 option:
878 [diff "ps"]
880 textconv = ps2ascii
881 binary = true
882 Performing a three-way merge
885 merge
886 The attribute merge affects how three versions of a file are merged
887 when a file-level merge is necessary during git merge, and other
888 commands such as git revert and git cherry-pick.
889 Set
891 Built-in 3-way merge driver is used to merge the contents in a
892 way similar to merge command of RCS suite. This is suitable for
893 ordinary text files.
894 Unset
896 Take the version from the current branch as the tentative merge
897 result, and declare that the merge has conflicts. This is
898 suitable for binary files that do not have a well-defined merge
899 semantics.
900 Unspecified
902 By default, this uses the same built-in 3-way merge driver as
903 is the case when the merge attribute is set. However, the
904 merge.default configuration variable can name different merge
905 driver to be used with paths for which the merge attribute is
906 unspecified.
907 String
909 3-way merge is performed using the specified custom merge
910 driver. The built-in 3-way merge driver can be explicitly
911 specified by asking for "text" driver; the built-in "take the
912 current branch" driver can be requested with "binary".
913 Built-in merge drivers
915 There are a few built-in low-level merge drivers defined that can
916 be asked for via the merge attribute.
917 text
919 Usual 3-way file level merge for text files. Conflicted regions
920 are marked with conflict markers <<<<<<<, ======= and >>>>>>>.
921 The version from your branch appears before the ======= marker,
922 and the version from the merged branch appears after the
923 ======= marker.
924 binary
926 Keep the version from your branch in the work tree, but leave
927 the path in the conflicted state for the user to sort out.
928 union
930 Run 3-way file level merge for text files, but take lines from
931 both versions, instead of leaving conflict markers. This tends
932 to leave the added lines in the resulting file in random order
933 and the user should verify the result. Do not use this if you
934 do not understand the implications.
935 Defining a custom merge driver
937 The definition of a merge driver is done in the .git/config file,
938 not in the gitattributes file, so strictly speaking this manual
939 page is a wrong place to talk about it. However...
940 To define a custom merge driver filfre, add a section to your
942 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
943 [merge "filfre"]
945 name = feel-free merge driver
946 driver = filfre %O %A %B %L %P
947 recursive = binary
948 The merge.*.name variable gives the driver a human-readable name.
951 The ‘merge.*.driver` variable’s value is used to construct a
953 command to run to merge ancestor’s version (%O), current version
954 (%A) and the other branches’ version (%B). These three tokens are
955 replaced with the names of temporary files that hold the contents
956 of these versions when the command line is built. Additionally, %L
957 will be replaced with the conflict marker size (see below).
958 The merge driver is expected to leave the result of the merge in
960 the file named with %A by overwriting it, and exit with zero status
961 if it managed to merge them cleanly, or non-zero if there were
962 conflicts.
963 The merge.*.recursive variable specifies what other merge driver to
965 use when the merge driver is called for an internal merge between
966 common ancestors, when there are more than one. When left
967 unspecified, the driver itself is used for both internal merge and
968 the final merge.
969 The merge driver can learn the pathname in which the merged result
971 will be stored via placeholder %P.
972 conflict-marker-size
974 This attribute controls the length of conflict markers left in the
975 work tree file during a conflicted merge. Only setting to the value
976 to a positive integer has any meaningful effect.
977 For example, this line in .gitattributes can be used to tell the
979 merge machinery to leave much longer (instead of the usual
980 7-character-long) conflict markers when merging the file
981 Documentation/git-merge.txt results in a conflict.
982 Documentation/git-merge.txt conflict-marker-size=32
984 Checking whitespace errors
987 whitespace
988 The core.whitespace configuration variable allows you to define
989 what diff and apply should consider whitespace errors for all paths
990 in the project (See git-config(1)). This attribute gives you finer
991 control per path.
992 Set
994 Notice all types of potential whitespace errors known to Git.
995 The tab width is taken from the value of the core.whitespace
996 configuration variable.
997 Unset
999 Do not notice anything as error.
1000 Unspecified
1002 Use the value of the core.whitespace configuration variable to
1003 decide what to notice as error.
1004 String
1006 Specify a comma separate list of common whitespace problems to
1007 notice in the same format as the core.whitespace configuration
1008 variable.
1009 Creating an archive
1011 export-ignore
1012 Files and directories with the attribute export-ignore won’t be
1013 added to archive files.
1014 export-subst
1016 If the attribute export-subst is set for a file then Git will
1017 expand several placeholders when adding this file to an archive.
1018 The expansion depends on the availability of a commit ID, i.e., if
1019 git-archive(1) has been given a tree instead of a commit or a tag
1020 then no replacement will be done. The placeholders are the same as
1021 those for the option --pretty=format: of git-log(1), except that
1022 they need to be wrapped like this: $Format:PLACEHOLDERS$ in the
1023 file. E.g. the string $Format:%H$ will be replaced by the commit
1024 hash.
1025 Packing objects
1027 delta
1028 Delta compression will not be attempted for blobs for paths with
1029 the attribute delta set to false.
1030 Viewing files in GUI tools
1032 encoding
1033 The value of this attribute specifies the character encoding that
1034 should be used by GUI tools (e.g. gitk(1) and git-gui(1)) to
1035 display the contents of the relevant file. Note that due to
1036 performance considerations gitk(1) does not use this attribute
1037 unless you manually enable per-file encodings in its options.
1038 If this attribute is not set or has an invalid value, the value of
1040 the gui.encoding configuration variable is used instead (See git-
1041 config(1)).
1042
1044 You do not want any end-of-line conversions applied to, nor textual
1045 diffs produced for, any binary file you track. You would need to
1046 specify e.g.
1047 *.jpg -text -diff
1049 but that may become cumbersome, when you have many attributes. Using
1052 macro attributes, you can define an attribute that, when set, also sets
1053 or unsets a number of other attributes at the same time. The system
1054 knows a built-in macro attribute, binary:
1055 *.jpg binary
1057 Setting the "binary" attribute also unsets the "text" and "diff"
1060 attributes as above. Note that macro attributes can only be "Set",
1061 though setting one might have the effect of setting or unsetting other
1062 attributes or even returning other attributes to the "Unspecified"
1063 state.
1064
1066 Custom macro attributes can be defined only in top-level gitattributes
1067 files ($GIT_DIR/info/attributes, the .gitattributes file at the top
1068 level of the working tree, or the global or system-wide gitattributes
1069 files), not in .gitattributes files in working tree subdirectories. The
1070 built-in macro attribute "binary" is equivalent to:
1071 [attr]binary -diff -merge -text
1073
1076 If you have these three gitattributes file:
1077 (in $GIT_DIR/info/attributes)
1079 a* foo !bar -baz
1081 (in .gitattributes)
1083 abc foo bar baz
1084 (in t/.gitattributes)
1086 ab* merge=filfre
1087 abc -foo -bar
1088 *.c frotz
1089 the attributes given to path t/abc are computed as follows:
1092 1. By examining t/.gitattributes (which is in the same directory as
1094 the path in question), Git finds that the first line matches.
1095 merge attribute is set. It also finds that the second line matches,
1096 and attributes foo and bar are unset.
1097 2. Then it examines .gitattributes (which is in the parent directory),
1099 and finds that the first line matches, but t/.gitattributes file
1100 already decided how merge, foo and bar attributes should be given
1101 to this path, so it leaves foo and bar unset. Attribute baz is set.
1102 3. Finally it examines $GIT_DIR/info/attributes. This file is used to
1104 override the in-tree settings. The first line is a match, and foo
1105 is set, bar is reverted to unspecified state, and baz is unset.
1106 As the result, the attributes assignment to t/abc becomes:
1108 foo set to true
1110 bar unspecified
1111 baz set to false
1112 merge set to string value "filfre"
1113 frotz unspecified
1114
1117 git-check-attr(1).
1118
1120 Part of the git(1) suite
1121Git 2.20.1 12/15/2018 GITATTRIBUTES(5)