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. When the pattern matches the path in question, the
21 attributes listed on the line are given to the path.
22 Each attribute can be in one of these states for a given path:
24 Set
26 The path has the attribute with special value "true"; this is
27 specified by listing only the name of the attribute in the
28 attribute list.
29 Unset
31 The path has the attribute with special value "false"; this is
32 specified by listing the name of the attribute prefixed with a dash
33 - in the attribute list.
34 Set to a value
36 The path has the attribute with specified string value; this is
37 specified by listing the name of the attribute followed by an equal
38 sign = and its value in the attribute list.
39 Unspecified
41 No pattern matches the path, and nothing says if the path has or
42 does not have the attribute, the attribute for the path is said to
43 be Unspecified.
44 When more than one pattern matches the path, a later line overrides an
46 earlier line. This overriding is done per attribute. The rules how the
47 pattern matches paths are the same as in .gitignore files; see
48 gitignore(5). Unlike .gitignore, negative patterns are forbidden.
49 When deciding what attributes are assigned to a path, Git consults
51 $GIT_DIR/info/attributes file (which has the highest precedence),
52 .gitattributes file in the same directory as the path in question, and
53 its parent directories up to the toplevel of the work tree (the further
54 the directory that contains .gitattributes is from the path in
55 question, the lower its precedence). Finally global and system-wide
56 files are considered (they have the lowest precedence).
57 When the .gitattributes file is missing from the work tree, the path in
59 the index is used as a fall-back. During checkout process,
60 .gitattributes in the index is used and then the file in the working
61 tree is used as a fall-back.
62 If you wish to affect only a single repository (i.e., to assign
64 attributes to files that are particular to one user’s workflow for that
65 repository), then attributes should be placed in the
66 $GIT_DIR/info/attributes file. Attributes which should be
67 version-controlled and distributed to other repositories (i.e.,
68 attributes of interest to all users) should go into .gitattributes
69 files. Attributes that should affect all repositories for a single user
70 should be placed in a file specified by the core.attributesfile
71 configuration option (see git-config(1)). Its default value is
72 $XDG_CONFIG_HOME/git/attributes. If $XDG_CONFIG_HOME is either not set
73 or empty, $HOME/.config/git/attributes is used instead. Attributes for
74 all users on a system should be placed in the
75 $(prefix)/etc/gitattributes file.
76 Sometimes you would need to override an setting of an attribute for a
78 path to Unspecified state. This can be done by listing the name of the
79 attribute prefixed with an exclamation point !.
80
82 Certain operations by Git can be influenced by assigning particular
83 attributes to a path. Currently, the following operations are
84 attributes-aware.
85 Checking-out and checking-in
87 These attributes affect how the contents stored in the repository are
88 copied to the working tree files when commands such as git checkout and
89 git merge run. They also affect how Git stores the contents you prepare
90 in the working tree in the repository upon git add and git commit.
91 text
93 This attribute enables and controls end-of-line normalization. When
94 a text file is normalized, its line endings are converted to LF in
95 the repository. To control what line ending style is used in the
96 working directory, use the eol attribute for a single file and the
97 core.eol configuration variable for all text files.
98 Set
100 Setting the text attribute on a path enables end-of-line
101 normalization and marks the path as a text file. End-of-line
102 conversion takes place without guessing the content type.
103 Unset
105 Unsetting the text attribute on a path tells Git not to attempt
106 any end-of-line conversion upon checkin or checkout.
107 Set to string value "auto"
109 When text is set to "auto", the path is marked for automatic
110 end-of-line normalization. If Git decides that the content is
111 text, its line endings are normalized to LF on checkin.
112 Unspecified
114 If the text attribute is unspecified, Git uses the
115 core.autocrlf configuration variable to determine if the file
116 should be converted.
117 Any other value causes Git to act as if text has been left
119 unspecified.
120 eol
122 This attribute sets a specific line-ending style to be used in the
123 working directory. It enables end-of-line normalization without any
124 content checks, effectively setting the text attribute.
125 Set to string value "crlf"
127 This setting forces Git to normalize line endings for this file
128 on checkin and convert them to CRLF when the file is checked
129 out.
130 Set to string value "lf"
132 This setting forces Git to normalize line endings to LF on
133 checkin and prevents conversion to CRLF when the file is
134 checked out.
135 Backwards compatibility with crlf attribute
137 For backwards compatibility, the crlf attribute is interpreted as
138 follows:
139 crlf text
141 -crlf -text
142 crlf=input eol=lf
143 End-of-line conversion
146 While Git normally leaves file contents alone, it can be configured
147 to normalize line endings to LF in the repository and, optionally,
148 to convert them to CRLF when files are checked out.
149 Here is an example that will make Git normalize .txt, .vcproj and
151 .sh files, ensure that .vcproj files have CRLF and .sh files have
152 LF in the working directory, and prevent .jpg files from being
153 normalized regardless of their content.
154 *.txt text
156 *.vcproj eol=crlf
157 *.sh eol=lf
158 *.jpg -text
159 Other source code management systems normalize all text files in
162 their repositories, and there are two ways to enable similar
163 automatic normalization in Git.
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 changing any
168 attributes.
169 [core]
171 autocrlf = true
172 This does not force normalization of all text files, but does
175 ensure that text files that you introduce to the repository have
176 their line endings normalized to LF when they are added, and that
177 files that are already normalized in the repository stay
178 normalized.
179 If you want to interoperate with a source code management system
181 that enforces end-of-line normalization, or you simply want all
182 text files in your repository to be normalized, you should instead
183 set the text attribute to "auto" for all files.
184 * text=auto
186 This ensures that all files that Git considers to be text will have
189 normalized (LF) line endings in the repository. The core.eol
190 configuration variable controls which line endings Git will use for
191 normalized files in your working directory; the default is to use
192 the native line ending for your platform, or CRLF if core.autocrlf
193 is set.
194 Note
196 When text=auto normalization is enabled in an existing
197 repository, any text files containing CRLFs should be
198 normalized. If they are not they will be normalized the next
199 time someone tries to change them, causing unfortunate
200 misattribution. From a clean working directory:
201 $ echo "* text=auto" >>.gitattributes
203 $ rm .git/index # Remove the index to force Git to
204 $ git reset # re-scan the working directory
205 $ git status # Show files that will be normalized
206 $ git add -u
207 $ git add .gitattributes
208 $ git commit -m "Introduce end-of-line normalization"
209 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
218 normalization enabled manually.
219 weirdchars.txt text
221 If core.safecrlf is set to "true" or "warn", Git verifies if the
224 conversion is reversible for the current setting of core.autocrlf.
225 For "true", Git rejects irreversible conversions; for "warn", Git
226 only prints a warning but accepts an irreversible conversion. The
227 safety triggers to prevent such a conversion done to the files in
228 the work tree, but there are a few exceptions. Even though...
229 · git add itself does not touch the files in the work tree, the
231 next checkout would, so the safety triggers;
232 · git apply to update a text file with a patch does touch the
234 files in the work tree, but the operation is about text files
235 and CRLF conversion is about fixing the line ending
236 inconsistencies, so the safety does not trigger;
237 · git diff itself does not touch the files in the work tree, it
239 is often run to inspect the changes you intend to next git add.
240 To catch potential problems early, safety triggers.
241 ident
243 When the attribute ident is set for a path, Git replaces $Id$ in
244 the blob object with $Id:, followed by the 40-character hexadecimal
245 blob object name, followed by a dollar sign $ upon checkout. Any
246 byte sequence that begins with $Id: and ends with $ in the worktree
247 file is replaced with $Id$ upon check-in.
248 filter
250 A filter attribute can be set to a string value that names a filter
251 driver specified in the configuration.
252 A filter driver consists of a clean command and a smudge command,
254 either of which can be left unspecified. Upon checkout, when the
255 smudge command is specified, the command is fed the blob object
256 from its standard input, and its standard output is used to update
257 the worktree file. Similarly, the clean command is used to convert
258 the contents of worktree file upon checkin.
259 One use of the content filtering is to massage the content into a
261 shape that is more convenient for the platform, filesystem, and the
262 user to use. For this mode of operation, the key phrase here is
263 "more convenient" and not "turning something unusable into usable".
264 In other words, the intent is that if someone unsets the filter
265 driver definition, or does not have the appropriate filter program,
266 the project should still be usable.
267 Another use of the content filtering is to store the content that
269 cannot be directly used in the repository (e.g. a UUID that refers
270 to the true content stored outside Git, or an encrypted content)
271 and turn it into a usable form upon checkout (e.g. download the
272 external content, or decrypt the encrypted content).
273 These two filters behave differently, and by default, a filter is
275 taken as the former, massaging the contents into more convenient
276 shape. A missing filter driver definition in the config, or a
277 filter driver that exits with a non-zero status, is not an error
278 but makes the filter a no-op passthru.
279 You can declare that a filter turns a content that by itself is
281 unusable into a usable content by setting the
282 filter.<driver>.required configuration variable to true.
283 For example, in .gitattributes, you would assign the filter
285 attribute for paths.
286 *.c filter=indent
288 Then you would define a "filter.indent.clean" and
291 "filter.indent.smudge" configuration in your .git/config to specify
292 a pair of commands to modify the contents of C programs when the
293 source files are checked in ("clean" is run) and checked out (no
294 change is made because the command is "cat").
295 [filter "indent"]
297 clean = indent
298 smudge = cat
299 For best results, clean should not alter its output further if it
302 is run twice ("clean→clean" should be equivalent to "clean"), and
303 multiple smudge commands should not alter clean's output
304 ("smudge→smudge→clean" should be equivalent to "clean"). See the
305 section on merging below.
306 The "indent" filter is well-behaved in this regard: it will not
308 modify input that is already correctly indented. In this case, the
309 lack of a smudge filter means that the clean filter must accept its
310 own output without modifying it.
311 If a filter must succeed in order to make the stored contents
313 usable, you can declare that the filter is required, in the
314 configuration:
315 [filter "crypt"]
317 clean = openssl enc ...
318 smudge = openssl enc -d ...
319 required
320 Sequence "%f" on the filter command line is replaced with the name
323 of the file the filter is working on. A filter might use this in
324 keyword substitution. For example:
325 [filter "p4"]
327 clean = git-p4-filter --clean %f
328 smudge = git-p4-filter --smudge %f
329 Interaction between checkin/checkout attributes
332 In the check-in codepath, the worktree file is first converted with
333 filter driver (if specified and corresponding driver defined), then
334 the result is processed with ident (if specified), and then finally
335 with text (again, if specified and applicable).
336 In the check-out codepath, the blob content is first converted with
338 text, and then ident and fed to filter.
339 Merging branches with differing checkin/checkout attributes
341 If you have added attributes to a file that cause the canonical
342 repository format for that file to change, such as adding a
343 clean/smudge filter or text/eol/ident attributes, merging anything
344 where the attribute is not in place would normally cause merge
345 conflicts.
346 To prevent these unnecessary merge conflicts, Git can be told to
348 run a virtual check-out and check-in of all three stages of a file
349 when resolving a three-way merge by setting the merge.renormalize
350 configuration variable. This prevents changes caused by check-in
351 conversion from causing spurious merge conflicts when a converted
352 file is merged with an unconverted file.
353 As long as a "smudge→clean" results in the same output as a "clean"
355 even on files that are already smudged, this strategy will
356 automatically resolve all filter-related conflicts. Filters that do
357 not act in this way may cause additional merge conflicts that must
358 be resolved manually.
359 Generating diff text
361 diff
362 The attribute diff affects how Git generates diffs for particular
363 files. It can tell Git whether to generate a textual patch for the
364 path or to treat the path as a binary file. It can also affect what
365 line is shown on the hunk header @@ -k,l +n,m @@ line, tell Git to
366 use an external command to generate the diff, or ask Git to convert
367 binary files to a text format before generating the diff.
368 Set
370 A path to which the diff attribute is set is treated as text,
371 even when they contain byte values that normally never appear
372 in text files, such as NUL.
373 Unset
375 A path to which the diff attribute is unset will generate
376 Binary files differ (or a binary patch, if binary patches are
377 enabled).
378 Unspecified
380 A path to which the diff attribute is unspecified first gets
381 its contents inspected, and if it looks like text, it is
382 treated as text. Otherwise it would generate Binary files
383 differ.
384 String
386 Diff is shown using the specified diff driver. Each driver may
387 specify one or more options, as described in the following
388 section. The options for the diff driver "foo" are defined by
389 the configuration variables in the "diff.foo" section of the
390 Git config file.
391 Defining an external diff driver
393 The definition of a diff driver is done in gitconfig, not
394 gitattributes file, so strictly speaking this manual page is a
395 wrong place to talk about it. However...
396 To define an external diff driver jcdiff, add a section to your
398 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
399 [diff "jcdiff"]
401 command = j-c-diff
402 When Git needs to show you a diff for the path with diff attribute
405 set to jcdiff, it calls the command you specified with the above
406 configuration, i.e. j-c-diff, with 7 parameters, just like
407 GIT_EXTERNAL_DIFF program is called. See git(1) for details.
408 Defining a custom hunk-header
410 Each group of changes (called a "hunk") in the textual diff output
411 is prefixed with a line of the form:
412 @@ -k,l +n,m @@ TEXT
414 This is called a hunk header. The "TEXT" portion is by default a
416 line that begins with an alphabet, an underscore or a dollar sign;
417 this matches what GNU diff -p output uses. This default selection
418 however is not suited for some contents, and you can use a
419 customized pattern to make a selection.
420 First, in .gitattributes, you would assign the diff attribute for
422 paths.
423 *.tex diff=tex
425 Then, you would define a "diff.tex.xfuncname" configuration to
428 specify a regular expression that matches a line that you would
429 want to appear as the hunk header "TEXT". Add a section to your
430 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
431 [diff "tex"]
433 xfuncname = "^(\\\\(sub)*section\\{.*)$"
434 Note. A single level of backslashes are eaten by the configuration
437 file parser, so you would need to double the backslashes; the
438 pattern above picks a line that begins with a backslash, and zero
439 or more occurrences of sub followed by section followed by open
440 brace, to the end of line.
441 There are a few built-in patterns to make this easier, and tex is
443 one of them, so you do not have to write the above in your
444 configuration file (you still need to enable this with the
445 attribute mechanism, via .gitattributes). The following built in
446 patterns are available:
447 · ada suitable for source code in the Ada language.
449 · bibtex suitable for files with BibTeX coded references.
451 · cpp suitable for source code in the C and C++ languages.
453 · csharp suitable for source code in the C# language.
455 · fortran suitable for source code in the Fortran language.
457 · html suitable for HTML/XHTML documents.
459 · java suitable for source code in the Java language.
461 · matlab suitable for source code in the MATLAB language.
463 · objc suitable for source code in the Objective-C language.
465 · pascal suitable for source code in the Pascal/Delphi language.
467 · perl suitable for source code in the Perl language.
469 · php suitable for source code in the PHP language.
471 · python suitable for source code in the Python language.
473 · ruby suitable for source code in the Ruby language.
475 · tex suitable for source code for LaTeX documents.
477 Customizing word diff
479 You can customize the rules that git diff --word-diff uses to split
480 words in a line, by specifying an appropriate regular expression in
481 the "diff.*.wordRegex" configuration variable. For example, in TeX
482 a backslash followed by a sequence of letters forms a command, but
483 several such commands can be run together without intervening
484 whitespace. To separate them, use a regular expression in your
485 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
486 [diff "tex"]
488 wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+"
489 A built-in pattern is provided for all languages listed in the
492 previous section.
493 Performing text diffs of binary files
495 Sometimes it is desirable to see the diff of a text-converted
496 version of some binary files. For example, a word processor
497 document can be converted to an ASCII text representation, and the
498 diff of the text shown. Even though this conversion loses some
499 information, the resulting diff is useful for human viewing (but
500 cannot be applied directly).
501 The textconv config option is used to define a program for
503 performing such a conversion. The program should take a single
504 argument, the name of a file to convert, and produce the resulting
505 text on stdout.
506 For example, to show the diff of the exif information of a file
508 instead of the binary information (assuming you have the exif tool
509 installed), add the following section to your $GIT_DIR/config file
510 (or $HOME/.gitconfig file):
511 [diff "jpg"]
513 textconv = exif
514 Note
517 The text conversion is generally a one-way conversion; in this
518 example, we lose the actual image contents and focus just on
519 the text data. This means that diffs generated by textconv are
520 not suitable for applying. For this reason, only git diff and
521 the git log family of commands (i.e., log, whatchanged, show)
522 will perform text conversion. git format-patch will never
523 generate this output. If you want to send somebody a
524 text-converted diff of a binary file (e.g., because it quickly
525 conveys the changes you have made), you should generate it
526 separately and send it as a comment in addition to the usual
527 binary diff that you might send.
528 Because text conversion can be slow, especially when doing a large
530 number of them with git log -p, Git provides a mechanism to cache
531 the output and use it in future diffs. To enable caching, set the
532 "cachetextconv" variable in your diff driver’s config. For example:
533 [diff "jpg"]
535 textconv = exif
536 cachetextconv = true
537 This will cache the result of running "exif" on each blob
540 indefinitely. If you change the textconv config variable for a diff
541 driver, Git will automatically invalidate the cache entries and
542 re-run the textconv filter. If you want to invalidate the cache
543 manually (e.g., because your version of "exif" was updated and now
544 produces better output), you can remove the cache manually with git
545 update-ref -d refs/notes/textconv/jpg (where "jpg" is the name of
546 the diff driver, as in the example above).
547 Choosing textconv versus external diff
549 If you want to show differences between binary or
550 specially-formatted blobs in your repository, you can choose to use
551 either an external diff command, or to use textconv to convert them
552 to a diff-able text format. Which method you choose depends on your
553 exact situation.
554 The advantage of using an external diff command is flexibility. You
556 are not bound to find line-oriented changes, nor is it necessary
557 for the output to resemble unified diff. You are free to locate and
558 report changes in the most appropriate way for your data format.
559 A textconv, by comparison, is much more limiting. You provide a
561 transformation of the data into a line-oriented text format, and
562 Git uses its regular diff tools to generate the output. There are
563 several advantages to choosing this method:
564 1. Ease of use. It is often much simpler to write a binary to text
566 transformation than it is to perform your own diff. In many
567 cases, existing programs can be used as textconv filters (e.g.,
568 exif, odt2txt).
569 2. Git diff features. By performing only the transformation step
571 yourself, you can still utilize many of Git’s diff features,
572 including colorization, word-diff, and combined diffs for
573 merges.
574 3. Caching. Textconv caching can speed up repeated diffs, such as
576 those you might trigger by running git log -p.
577 Marking files as binary
579 Git usually guesses correctly whether a blob contains text or
580 binary data by examining the beginning of the contents. However,
581 sometimes you may want to override its decision, either because a
582 blob contains binary data later in the file, or because the
583 content, while technically composed of text characters, is opaque
584 to a human reader. For example, many postscript files contain only
585 ascii characters, but produce noisy and meaningless diffs.
586 The simplest way to mark a file as binary is to unset the diff
588 attribute in the .gitattributes file:
589 *.ps -diff
591 This will cause Git to generate Binary files differ (or a binary
594 patch, if binary patches are enabled) instead of a regular diff.
595 However, one may also want to specify other diff driver attributes.
597 For example, you might want to use textconv to convert postscript
598 files to an ascii representation for human viewing, but otherwise
599 treat them as binary files. You cannot specify both -diff and
600 diff=ps attributes. The solution is to use the diff.*.binary config
601 option:
602 [diff "ps"]
604 textconv = ps2ascii
605 binary = true
606 Performing a three-way merge
609 merge
610 The attribute merge affects how three versions of a file are merged
611 when a file-level merge is necessary during git merge, and other
612 commands such as git revert and git cherry-pick.
613 Set
615 Built-in 3-way merge driver is used to merge the contents in a
616 way similar to merge command of RCS suite. This is suitable for
617 ordinary text files.
618 Unset
620 Take the version from the current branch as the tentative merge
621 result, and declare that the merge has conflicts. This is
622 suitable for binary files that do not have a well-defined merge
623 semantics.
624 Unspecified
626 By default, this uses the same built-in 3-way merge driver as
627 is the case when the merge attribute is set. However, the
628 merge.default configuration variable can name different merge
629 driver to be used with paths for which the merge attribute is
630 unspecified.
631 String
633 3-way merge is performed using the specified custom merge
634 driver. The built-in 3-way merge driver can be explicitly
635 specified by asking for "text" driver; the built-in "take the
636 current branch" driver can be requested with "binary".
637 Built-in merge drivers
639 There are a few built-in low-level merge drivers defined that can
640 be asked for via the merge attribute.
641 text
643 Usual 3-way file level merge for text files. Conflicted regions
644 are marked with conflict markers <<<<<<<, ======= and >>>>>>>.
645 The version from your branch appears before the ======= marker,
646 and the version from the merged branch appears after the
647 ======= marker.
648 binary
650 Keep the version from your branch in the work tree, but leave
651 the path in the conflicted state for the user to sort out.
652 union
654 Run 3-way file level merge for text files, but take lines from
655 both versions, instead of leaving conflict markers. This tends
656 to leave the added lines in the resulting file in random order
657 and the user should verify the result. Do not use this if you
658 do not understand the implications.
659 Defining a custom merge driver
661 The definition of a merge driver is done in the .git/config file,
662 not in the gitattributes file, so strictly speaking this manual
663 page is a wrong place to talk about it. However...
664 To define a custom merge driver filfre, add a section to your
666 $GIT_DIR/config file (or $HOME/.gitconfig file) like this:
667 [merge "filfre"]
669 name = feel-free merge driver
670 driver = filfre %O %A %B
671 recursive = binary
672 The merge.*.name variable gives the driver a human-readable name.
675 The ‘merge.*.driver` variable’s value is used to construct a
677 command to run to merge ancestor’s version (%O), current version
678 (%A) and the other branches’ version (%B). These three tokens are
679 replaced with the names of temporary files that hold the contents
680 of these versions when the command line is built. Additionally, %L
681 will be replaced with the conflict marker size (see below).
682 The merge driver is expected to leave the result of the merge in
684 the file named with %A by overwriting it, and exit with zero status
685 if it managed to merge them cleanly, or non-zero if there were
686 conflicts.
687 The merge.*.recursive variable specifies what other merge driver to
689 use when the merge driver is called for an internal merge between
690 common ancestors, when there are more than one. When left
691 unspecified, the driver itself is used for both internal merge and
692 the final merge.
693 conflict-marker-size
695 This attribute controls the length of conflict markers left in the
696 work tree file during a conflicted merge. Only setting to the value
697 to a positive integer has any meaningful effect.
698 For example, this line in .gitattributes can be used to tell the
700 merge machinery to leave much longer (instead of the usual
701 7-character-long) conflict markers when merging the file
702 Documentation/git-merge.txt results in a conflict.
703 Documentation/git-merge.txt conflict-marker-size=32
705 Checking whitespace errors
708 whitespace
709 The core.whitespace configuration variable allows you to define
710 what diff and apply should consider whitespace errors for all paths
711 in the project (See git-config(1)). This attribute gives you finer
712 control per path.
713 Set
715 Notice all types of potential whitespace errors known to Git.
716 The tab width is taken from the value of the core.whitespace
717 configuration variable.
718 Unset
720 Do not notice anything as error.
721 Unspecified
723 Use the value of the core.whitespace configuration variable to
724 decide what to notice as error.
725 String
727 Specify a comma separate list of common whitespace problems to
728 notice in the same format as the core.whitespace configuration
729 variable.
730 Creating an archive
732 export-ignore
733 Files and directories with the attribute export-ignore won’t be
734 added to archive files.
735 export-subst
737 If the attribute export-subst is set for a file then Git will
738 expand several placeholders when adding this file to an archive.
739 The expansion depends on the availability of a commit ID, i.e., if
740 git-archive(1) has been given a tree instead of a commit or a tag
741 then no replacement will be done. The placeholders are the same as
742 those for the option --pretty=format: of git-log(1), except that
743 they need to be wrapped like this: $Format:PLACEHOLDERS$ in the
744 file. E.g. the string $Format:%H$ will be replaced by the commit
745 hash.
746 Packing objects
748 delta
749 Delta compression will not be attempted for blobs for paths with
750 the attribute delta set to false.
751 Viewing files in GUI tools
753 encoding
754 The value of this attribute specifies the character encoding that
755 should be used by GUI tools (e.g. gitk(1) and git-gui(1)) to
756 display the contents of the relevant file. Note that due to
757 performance considerations gitk(1) does not use this attribute
758 unless you manually enable per-file encodings in its options.
759 If this attribute is not set or has an invalid value, the value of
761 the gui.encoding configuration variable is used instead (See git-
762 config(1)).
763
765 You do not want any end-of-line conversions applied to, nor textual
766 diffs produced for, any binary file you track. You would need to
767 specify e.g.
768 *.jpg -text -diff
770 but that may become cumbersome, when you have many attributes. Using
773 macro attributes, you can define an attribute that, when set, also sets
774 or unsets a number of other attributes at the same time. The system
775 knows a built-in macro attribute, binary:
776 *.jpg binary
778 Setting the "binary" attribute also unsets the "text" and "diff"
781 attributes as above. Note that macro attributes can only be "Set",
782 though setting one might have the effect of setting or unsetting other
783 attributes or even returning other attributes to the "Unspecified"
784 state.
785
787 Custom macro attributes can be defined only in the .gitattributes file
788 at the toplevel (i.e. not in any subdirectory). The built-in macro
789 attribute "binary" is equivalent to:
790 [attr]binary -diff -merge -text
792
795 If you have these three gitattributes file:
796 (in $GIT_DIR/info/attributes)
798 a* foo !bar -baz
800 (in .gitattributes)
802 abc foo bar baz
803 (in t/.gitattributes)
805 ab* merge=filfre
806 abc -foo -bar
807 *.c frotz
808 the attributes given to path t/abc are computed as follows:
811 1. By examining t/.gitattributes (which is in the same directory as
813 the path in question), Git finds that the first line matches.
814 merge attribute is set. It also finds that the second line matches,
815 and attributes foo and bar are unset.
816 2. Then it examines .gitattributes (which is in the parent directory),
818 and finds that the first line matches, but t/.gitattributes file
819 already decided how merge, foo and bar attributes should be given
820 to this path, so it leaves foo and bar unset. Attribute baz is set.
821 3. Finally it examines $GIT_DIR/info/attributes. This file is used to
823 override the in-tree settings. The first line is a match, and foo
824 is set, bar is reverted to unspecified state, and baz is unset.
825 As the result, the attributes assignment to t/abc becomes:
827 foo set to true
829 bar unspecified
830 baz set to false
831 merge set to string value "filfre"
832 frotz unspecified
833
836 git-check-attr(1).
837
839 Part of the git(1) suite
840Git 1.8.3.1 11/19/2018 GITATTRIBUTES(5)