Merge branch 'ks/pack-objects-bitmap'
[git/git.git] / Documentation / gitattributes.txt
1 gitattributes(5)
2 ================
3
4 NAME
5 ----
6 gitattributes - defining attributes per path
7
8 SYNOPSIS
9 --------
10 $GIT_DIR/info/attributes, .gitattributes
11
12
13 DESCRIPTION
14 -----------
15
16 A `gitattributes` file is a simple text file that gives
17 `attributes` to pathnames.
18
19 Each line in `gitattributes` file is of form:
20
21 pattern attr1 attr2 ...
22
23 That is, a pattern followed by an attributes list,
24 separated by whitespaces. When the pattern matches the
25 path in question, the attributes listed on the line are given to
26 the path.
27
28 Each attribute can be in one of these states for a given path:
29
30 Set::
31
32 The path has the attribute with special value "true";
33 this is specified by listing only the name of the
34 attribute in the attribute list.
35
36 Unset::
37
38 The path has the attribute with special value "false";
39 this is specified by listing the name of the attribute
40 prefixed with a dash `-` in the attribute list.
41
42 Set to a value::
43
44 The path has the attribute with specified string value;
45 this is specified by listing the name of the attribute
46 followed by an equal sign `=` and its value in the
47 attribute list.
48
49 Unspecified::
50
51 No pattern matches the path, and nothing says if
52 the path has or does not have the attribute, the
53 attribute for the path is said to be Unspecified.
54
55 When more than one pattern matches the path, a later line
56 overrides an earlier line. This overriding is done per
57 attribute. The rules how the pattern matches paths are the
58 same as in `.gitignore` files; see linkgit:gitignore[5].
59 Unlike `.gitignore`, negative patterns are forbidden.
60
61 When deciding what attributes are assigned to a path, Git
62 consults `$GIT_DIR/info/attributes` file (which has the highest
63 precedence), `.gitattributes` file in the same directory as the
64 path in question, and its parent directories up to the toplevel of the
65 work tree (the further the directory that contains `.gitattributes`
66 is from the path in question, the lower its precedence). Finally
67 global and system-wide files are considered (they have the lowest
68 precedence).
69
70 When the `.gitattributes` file is missing from the work tree, the
71 path in the index is used as a fall-back. During checkout process,
72 `.gitattributes` in the index is used and then the file in the
73 working tree is used as a fall-back.
74
75 If you wish to affect only a single repository (i.e., to assign
76 attributes to files that are particular to
77 one user's workflow for that repository), then
78 attributes should be placed in the `$GIT_DIR/info/attributes` file.
79 Attributes which should be version-controlled and distributed to other
80 repositories (i.e., attributes of interest to all users) should go into
81 `.gitattributes` files. Attributes that should affect all repositories
82 for a single user should be placed in a file specified by the
83 `core.attributesFile` configuration option (see linkgit:git-config[1]).
84 Its default value is $XDG_CONFIG_HOME/git/attributes. If $XDG_CONFIG_HOME
85 is either not set or empty, $HOME/.config/git/attributes is used instead.
86 Attributes for all users on a system should be placed in the
87 `$(prefix)/etc/gitattributes` file.
88
89 Sometimes you would need to override an setting of an attribute
90 for a path to `Unspecified` state. This can be done by listing
91 the name of the attribute prefixed with an exclamation point `!`.
92
93
94 EFFECTS
95 -------
96
97 Certain operations by Git can be influenced by assigning
98 particular attributes to a path. Currently, the following
99 operations are attributes-aware.
100
101 Checking-out and checking-in
102 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
103
104 These attributes affect how the contents stored in the
105 repository are copied to the working tree files when commands
106 such as 'git checkout' and 'git merge' run. They also affect how
107 Git stores the contents you prepare in the working tree in the
108 repository upon 'git add' and 'git commit'.
109
110 `text`
111 ^^^^^^
112
113 This attribute enables and controls end-of-line normalization. When a
114 text file is normalized, its line endings are converted to LF in the
115 repository. To control what line ending style is used in the working
116 directory, use the `eol` attribute for a single file and the
117 `core.eol` configuration variable for all text files.
118 Note that `core.autocrlf` overrides `core.eol`
119
120 Set::
121
122 Setting the `text` attribute on a path enables end-of-line
123 normalization and marks the path as a text file. End-of-line
124 conversion takes place without guessing the content type.
125
126 Unset::
127
128 Unsetting the `text` attribute on a path tells Git not to
129 attempt any end-of-line conversion upon checkin or checkout.
130
131 Set to string value "auto"::
132
133 When `text` is set to "auto", the path is marked for automatic
134 end-of-line conversion. If Git decides that the content is
135 text, its line endings are converted to LF on checkin.
136 When the file has been committed with CRLF, no conversion is done.
137
138 Unspecified::
139
140 If the `text` attribute is unspecified, Git uses the
141 `core.autocrlf` configuration variable to determine if the
142 file should be converted.
143
144 Any other value causes Git to act as if `text` has been left
145 unspecified.
146
147 `eol`
148 ^^^^^
149
150 This attribute sets a specific line-ending style to be used in the
151 working directory. It enables end-of-line conversion without any
152 content checks, effectively setting the `text` attribute.
153
154 Set to string value "crlf"::
155
156 This setting forces Git to normalize line endings for this
157 file on checkin and convert them to CRLF when the file is
158 checked out.
159
160 Set to string value "lf"::
161
162 This setting forces Git to normalize line endings to LF on
163 checkin and prevents conversion to CRLF when the file is
164 checked out.
165
166 Backwards compatibility with `crlf` attribute
167 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
168
169 For backwards compatibility, the `crlf` attribute is interpreted as
170 follows:
171
172 ------------------------
173 crlf text
174 -crlf -text
175 crlf=input eol=lf
176 ------------------------
177
178 End-of-line conversion
179 ^^^^^^^^^^^^^^^^^^^^^^
180
181 While Git normally leaves file contents alone, it can be configured to
182 normalize line endings to LF in the repository and, optionally, to
183 convert them to CRLF when files are checked out.
184
185 If you simply want to have CRLF line endings in your working directory
186 regardless of the repository you are working with, you can set the
187 config variable "core.autocrlf" without using any attributes.
188
189 ------------------------
190 [core]
191 autocrlf = true
192 ------------------------
193
194 This does not force normalization of text files, but does ensure
195 that text files that you introduce to the repository have their line
196 endings normalized to LF when they are added, and that files that are
197 already normalized in the repository stay normalized.
198
199 If you want to ensure that text files that any contributor introduces to
200 the repository have their line endings normalized, you can set the
201 `text` attribute to "auto" for _all_ files.
202
203 ------------------------
204 * text=auto
205 ------------------------
206
207 The attributes allow a fine-grained control, how the line endings
208 are converted.
209 Here is an example that will make Git normalize .txt, .vcproj and .sh
210 files, ensure that .vcproj files have CRLF and .sh files have LF in
211 the working directory, and prevent .jpg files from being normalized
212 regardless of their content.
213
214 ------------------------
215 * text=auto
216 *.txt text
217 *.vcproj text eol=crlf
218 *.sh text eol=lf
219 *.jpg -text
220 ------------------------
221
222 NOTE: When `text=auto` conversion is enabled in a cross-platform
223 project using push and pull to a central repository the text files
224 containing CRLFs should be normalized.
225
226 From a clean working directory:
227
228 -------------------------------------------------
229 $ echo "* text=auto" >.gitattributes
230 $ rm .git/index # Remove the index to force Git to
231 $ git reset # re-scan the working directory
232 $ git status # Show files that will be normalized
233 $ git add -u
234 $ git add .gitattributes
235 $ git commit -m "Introduce end-of-line normalization"
236 -------------------------------------------------
237
238 If any files that should not be normalized show up in 'git status',
239 unset their `text` attribute before running 'git add -u'.
240
241 ------------------------
242 manual.pdf -text
243 ------------------------
244
245 Conversely, text files that Git does not detect can have normalization
246 enabled manually.
247
248 ------------------------
249 weirdchars.txt text
250 ------------------------
251
252 If `core.safecrlf` is set to "true" or "warn", Git verifies if
253 the conversion is reversible for the current setting of
254 `core.autocrlf`. For "true", Git rejects irreversible
255 conversions; for "warn", Git only prints a warning but accepts
256 an irreversible conversion. The safety triggers to prevent such
257 a conversion done to the files in the work tree, but there are a
258 few exceptions. Even though...
259
260 - 'git add' itself does not touch the files in the work tree, the
261 next checkout would, so the safety triggers;
262
263 - 'git apply' to update a text file with a patch does touch the files
264 in the work tree, but the operation is about text files and CRLF
265 conversion is about fixing the line ending inconsistencies, so the
266 safety does not trigger;
267
268 - 'git diff' itself does not touch the files in the work tree, it is
269 often run to inspect the changes you intend to next 'git add'. To
270 catch potential problems early, safety triggers.
271
272
273 `ident`
274 ^^^^^^^
275
276 When the attribute `ident` is set for a path, Git replaces
277 `$Id$` in the blob object with `$Id:`, followed by the
278 40-character hexadecimal blob object name, followed by a dollar
279 sign `$` upon checkout. Any byte sequence that begins with
280 `$Id:` and ends with `$` in the worktree file is replaced
281 with `$Id$` upon check-in.
282
283
284 `filter`
285 ^^^^^^^^
286
287 A `filter` attribute can be set to a string value that names a
288 filter driver specified in the configuration.
289
290 A filter driver consists of a `clean` command and a `smudge`
291 command, either of which can be left unspecified. Upon
292 checkout, when the `smudge` command is specified, the command is
293 fed the blob object from its standard input, and its standard
294 output is used to update the worktree file. Similarly, the
295 `clean` command is used to convert the contents of worktree file
296 upon checkin.
297
298 One use of the content filtering is to massage the content into a shape
299 that is more convenient for the platform, filesystem, and the user to use.
300 For this mode of operation, the key phrase here is "more convenient" and
301 not "turning something unusable into usable". In other words, the intent
302 is that if someone unsets the filter driver definition, or does not have
303 the appropriate filter program, the project should still be usable.
304
305 Another use of the content filtering is to store the content that cannot
306 be directly used in the repository (e.g. a UUID that refers to the true
307 content stored outside Git, or an encrypted content) and turn it into a
308 usable form upon checkout (e.g. download the external content, or decrypt
309 the encrypted content).
310
311 These two filters behave differently, and by default, a filter is taken as
312 the former, massaging the contents into more convenient shape. A missing
313 filter driver definition in the config, or a filter driver that exits with
314 a non-zero status, is not an error but makes the filter a no-op passthru.
315
316 You can declare that a filter turns a content that by itself is unusable
317 into a usable content by setting the filter.<driver>.required configuration
318 variable to `true`.
319
320 For example, in .gitattributes, you would assign the `filter`
321 attribute for paths.
322
323 ------------------------
324 *.c filter=indent
325 ------------------------
326
327 Then you would define a "filter.indent.clean" and "filter.indent.smudge"
328 configuration in your .git/config to specify a pair of commands to
329 modify the contents of C programs when the source files are checked
330 in ("clean" is run) and checked out (no change is made because the
331 command is "cat").
332
333 ------------------------
334 [filter "indent"]
335 clean = indent
336 smudge = cat
337 ------------------------
338
339 For best results, `clean` should not alter its output further if it is
340 run twice ("clean->clean" should be equivalent to "clean"), and
341 multiple `smudge` commands should not alter `clean`'s output
342 ("smudge->smudge->clean" should be equivalent to "clean"). See the
343 section on merging below.
344
345 The "indent" filter is well-behaved in this regard: it will not modify
346 input that is already correctly indented. In this case, the lack of a
347 smudge filter means that the clean filter _must_ accept its own output
348 without modifying it.
349
350 If a filter _must_ succeed in order to make the stored contents usable,
351 you can declare that the filter is `required`, in the configuration:
352
353 ------------------------
354 [filter "crypt"]
355 clean = openssl enc ...
356 smudge = openssl enc -d ...
357 required
358 ------------------------
359
360 Sequence "%f" on the filter command line is replaced with the name of
361 the file the filter is working on. A filter might use this in keyword
362 substitution. For example:
363
364 ------------------------
365 [filter "p4"]
366 clean = git-p4-filter --clean %f
367 smudge = git-p4-filter --smudge %f
368 ------------------------
369
370 Note that "%f" is the name of the path that is being worked on. Depending
371 on the version that is being filtered, the corresponding file on disk may
372 not exist, or may have different contents. So, smudge and clean commands
373 should not try to access the file on disk, but only act as filters on the
374 content provided to them on standard input.
375
376 Interaction between checkin/checkout attributes
377 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
378
379 In the check-in codepath, the worktree file is first converted
380 with `filter` driver (if specified and corresponding driver
381 defined), then the result is processed with `ident` (if
382 specified), and then finally with `text` (again, if specified
383 and applicable).
384
385 In the check-out codepath, the blob content is first converted
386 with `text`, and then `ident` and fed to `filter`.
387
388
389 Merging branches with differing checkin/checkout attributes
390 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
391
392 If you have added attributes to a file that cause the canonical
393 repository format for that file to change, such as adding a
394 clean/smudge filter or text/eol/ident attributes, merging anything
395 where the attribute is not in place would normally cause merge
396 conflicts.
397
398 To prevent these unnecessary merge conflicts, Git can be told to run a
399 virtual check-out and check-in of all three stages of a file when
400 resolving a three-way merge by setting the `merge.renormalize`
401 configuration variable. This prevents changes caused by check-in
402 conversion from causing spurious merge conflicts when a converted file
403 is merged with an unconverted file.
404
405 As long as a "smudge->clean" results in the same output as a "clean"
406 even on files that are already smudged, this strategy will
407 automatically resolve all filter-related conflicts. Filters that do
408 not act in this way may cause additional merge conflicts that must be
409 resolved manually.
410
411
412 Generating diff text
413 ~~~~~~~~~~~~~~~~~~~~
414
415 `diff`
416 ^^^^^^
417
418 The attribute `diff` affects how Git generates diffs for particular
419 files. It can tell Git whether to generate a textual patch for the path
420 or to treat the path as a binary file. It can also affect what line is
421 shown on the hunk header `@@ -k,l +n,m @@` line, tell Git to use an
422 external command to generate the diff, or ask Git to convert binary
423 files to a text format before generating the diff.
424
425 Set::
426
427 A path to which the `diff` attribute is set is treated
428 as text, even when they contain byte values that
429 normally never appear in text files, such as NUL.
430
431 Unset::
432
433 A path to which the `diff` attribute is unset will
434 generate `Binary files differ` (or a binary patch, if
435 binary patches are enabled).
436
437 Unspecified::
438
439 A path to which the `diff` attribute is unspecified
440 first gets its contents inspected, and if it looks like
441 text and is smaller than core.bigFileThreshold, it is treated
442 as text. Otherwise it would generate `Binary files differ`.
443
444 String::
445
446 Diff is shown using the specified diff driver. Each driver may
447 specify one or more options, as described in the following
448 section. The options for the diff driver "foo" are defined
449 by the configuration variables in the "diff.foo" section of the
450 Git config file.
451
452
453 Defining an external diff driver
454 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
455
456 The definition of a diff driver is done in `gitconfig`, not
457 `gitattributes` file, so strictly speaking this manual page is a
458 wrong place to talk about it. However...
459
460 To define an external diff driver `jcdiff`, add a section to your
461 `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this:
462
463 ----------------------------------------------------------------
464 [diff "jcdiff"]
465 command = j-c-diff
466 ----------------------------------------------------------------
467
468 When Git needs to show you a diff for the path with `diff`
469 attribute set to `jcdiff`, it calls the command you specified
470 with the above configuration, i.e. `j-c-diff`, with 7
471 parameters, just like `GIT_EXTERNAL_DIFF` program is called.
472 See linkgit:git[1] for details.
473
474
475 Defining a custom hunk-header
476 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
477
478 Each group of changes (called a "hunk") in the textual diff output
479 is prefixed with a line of the form:
480
481 @@ -k,l +n,m @@ TEXT
482
483 This is called a 'hunk header'. The "TEXT" portion is by default a line
484 that begins with an alphabet, an underscore or a dollar sign; this
485 matches what GNU 'diff -p' output uses. This default selection however
486 is not suited for some contents, and you can use a customized pattern
487 to make a selection.
488
489 First, in .gitattributes, you would assign the `diff` attribute
490 for paths.
491
492 ------------------------
493 *.tex diff=tex
494 ------------------------
495
496 Then, you would define a "diff.tex.xfuncname" configuration to
497 specify a regular expression that matches a line that you would
498 want to appear as the hunk header "TEXT". Add a section to your
499 `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this:
500
501 ------------------------
502 [diff "tex"]
503 xfuncname = "^(\\\\(sub)*section\\{.*)$"
504 ------------------------
505
506 Note. A single level of backslashes are eaten by the
507 configuration file parser, so you would need to double the
508 backslashes; the pattern above picks a line that begins with a
509 backslash, and zero or more occurrences of `sub` followed by
510 `section` followed by open brace, to the end of line.
511
512 There are a few built-in patterns to make this easier, and `tex`
513 is one of them, so you do not have to write the above in your
514 configuration file (you still need to enable this with the
515 attribute mechanism, via `.gitattributes`). The following built in
516 patterns are available:
517
518 - `ada` suitable for source code in the Ada language.
519
520 - `bibtex` suitable for files with BibTeX coded references.
521
522 - `cpp` suitable for source code in the C and C++ languages.
523
524 - `csharp` suitable for source code in the C# language.
525
526 - `css` suitable for cascading style sheets.
527
528 - `fortran` suitable for source code in the Fortran language.
529
530 - `fountain` suitable for Fountain documents.
531
532 - `html` suitable for HTML/XHTML documents.
533
534 - `java` suitable for source code in the Java language.
535
536 - `matlab` suitable for source code in the MATLAB language.
537
538 - `objc` suitable for source code in the Objective-C language.
539
540 - `pascal` suitable for source code in the Pascal/Delphi language.
541
542 - `perl` suitable for source code in the Perl language.
543
544 - `php` suitable for source code in the PHP language.
545
546 - `python` suitable for source code in the Python language.
547
548 - `ruby` suitable for source code in the Ruby language.
549
550 - `tex` suitable for source code for LaTeX documents.
551
552
553 Customizing word diff
554 ^^^^^^^^^^^^^^^^^^^^^
555
556 You can customize the rules that `git diff --word-diff` uses to
557 split words in a line, by specifying an appropriate regular expression
558 in the "diff.*.wordRegex" configuration variable. For example, in TeX
559 a backslash followed by a sequence of letters forms a command, but
560 several such commands can be run together without intervening
561 whitespace. To separate them, use a regular expression in your
562 `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this:
563
564 ------------------------
565 [diff "tex"]
566 wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+"
567 ------------------------
568
569 A built-in pattern is provided for all languages listed in the
570 previous section.
571
572
573 Performing text diffs of binary files
574 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
575
576 Sometimes it is desirable to see the diff of a text-converted
577 version of some binary files. For example, a word processor
578 document can be converted to an ASCII text representation, and
579 the diff of the text shown. Even though this conversion loses
580 some information, the resulting diff is useful for human
581 viewing (but cannot be applied directly).
582
583 The `textconv` config option is used to define a program for
584 performing such a conversion. The program should take a single
585 argument, the name of a file to convert, and produce the
586 resulting text on stdout.
587
588 For example, to show the diff of the exif information of a
589 file instead of the binary information (assuming you have the
590 exif tool installed), add the following section to your
591 `$GIT_DIR/config` file (or `$HOME/.gitconfig` file):
592
593 ------------------------
594 [diff "jpg"]
595 textconv = exif
596 ------------------------
597
598 NOTE: The text conversion is generally a one-way conversion;
599 in this example, we lose the actual image contents and focus
600 just on the text data. This means that diffs generated by
601 textconv are _not_ suitable for applying. For this reason,
602 only `git diff` and the `git log` family of commands (i.e.,
603 log, whatchanged, show) will perform text conversion. `git
604 format-patch` will never generate this output. If you want to
605 send somebody a text-converted diff of a binary file (e.g.,
606 because it quickly conveys the changes you have made), you
607 should generate it separately and send it as a comment _in
608 addition to_ the usual binary diff that you might send.
609
610 Because text conversion can be slow, especially when doing a
611 large number of them with `git log -p`, Git provides a mechanism
612 to cache the output and use it in future diffs. To enable
613 caching, set the "cachetextconv" variable in your diff driver's
614 config. For example:
615
616 ------------------------
617 [diff "jpg"]
618 textconv = exif
619 cachetextconv = true
620 ------------------------
621
622 This will cache the result of running "exif" on each blob
623 indefinitely. If you change the textconv config variable for a
624 diff driver, Git will automatically invalidate the cache entries
625 and re-run the textconv filter. If you want to invalidate the
626 cache manually (e.g., because your version of "exif" was updated
627 and now produces better output), you can remove the cache
628 manually with `git update-ref -d refs/notes/textconv/jpg` (where
629 "jpg" is the name of the diff driver, as in the example above).
630
631 Choosing textconv versus external diff
632 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
633
634 If you want to show differences between binary or specially-formatted
635 blobs in your repository, you can choose to use either an external diff
636 command, or to use textconv to convert them to a diff-able text format.
637 Which method you choose depends on your exact situation.
638
639 The advantage of using an external diff command is flexibility. You are
640 not bound to find line-oriented changes, nor is it necessary for the
641 output to resemble unified diff. You are free to locate and report
642 changes in the most appropriate way for your data format.
643
644 A textconv, by comparison, is much more limiting. You provide a
645 transformation of the data into a line-oriented text format, and Git
646 uses its regular diff tools to generate the output. There are several
647 advantages to choosing this method:
648
649 1. Ease of use. It is often much simpler to write a binary to text
650 transformation than it is to perform your own diff. In many cases,
651 existing programs can be used as textconv filters (e.g., exif,
652 odt2txt).
653
654 2. Git diff features. By performing only the transformation step
655 yourself, you can still utilize many of Git's diff features,
656 including colorization, word-diff, and combined diffs for merges.
657
658 3. Caching. Textconv caching can speed up repeated diffs, such as those
659 you might trigger by running `git log -p`.
660
661
662 Marking files as binary
663 ^^^^^^^^^^^^^^^^^^^^^^^
664
665 Git usually guesses correctly whether a blob contains text or binary
666 data by examining the beginning of the contents. However, sometimes you
667 may want to override its decision, either because a blob contains binary
668 data later in the file, or because the content, while technically
669 composed of text characters, is opaque to a human reader. For example,
670 many postscript files contain only ASCII characters, but produce noisy
671 and meaningless diffs.
672
673 The simplest way to mark a file as binary is to unset the diff
674 attribute in the `.gitattributes` file:
675
676 ------------------------
677 *.ps -diff
678 ------------------------
679
680 This will cause Git to generate `Binary files differ` (or a binary
681 patch, if binary patches are enabled) instead of a regular diff.
682
683 However, one may also want to specify other diff driver attributes. For
684 example, you might want to use `textconv` to convert postscript files to
685 an ASCII representation for human viewing, but otherwise treat them as
686 binary files. You cannot specify both `-diff` and `diff=ps` attributes.
687 The solution is to use the `diff.*.binary` config option:
688
689 ------------------------
690 [diff "ps"]
691 textconv = ps2ascii
692 binary = true
693 ------------------------
694
695 Performing a three-way merge
696 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
697
698 `merge`
699 ^^^^^^^
700
701 The attribute `merge` affects how three versions of a file are
702 merged when a file-level merge is necessary during `git merge`,
703 and other commands such as `git revert` and `git cherry-pick`.
704
705 Set::
706
707 Built-in 3-way merge driver is used to merge the
708 contents in a way similar to 'merge' command of `RCS`
709 suite. This is suitable for ordinary text files.
710
711 Unset::
712
713 Take the version from the current branch as the
714 tentative merge result, and declare that the merge has
715 conflicts. This is suitable for binary files that do
716 not have a well-defined merge semantics.
717
718 Unspecified::
719
720 By default, this uses the same built-in 3-way merge
721 driver as is the case when the `merge` attribute is set.
722 However, the `merge.default` configuration variable can name
723 different merge driver to be used with paths for which the
724 `merge` attribute is unspecified.
725
726 String::
727
728 3-way merge is performed using the specified custom
729 merge driver. The built-in 3-way merge driver can be
730 explicitly specified by asking for "text" driver; the
731 built-in "take the current branch" driver can be
732 requested with "binary".
733
734
735 Built-in merge drivers
736 ^^^^^^^^^^^^^^^^^^^^^^
737
738 There are a few built-in low-level merge drivers defined that
739 can be asked for via the `merge` attribute.
740
741 text::
742
743 Usual 3-way file level merge for text files. Conflicted
744 regions are marked with conflict markers `<<<<<<<`,
745 `=======` and `>>>>>>>`. The version from your branch
746 appears before the `=======` marker, and the version
747 from the merged branch appears after the `=======`
748 marker.
749
750 binary::
751
752 Keep the version from your branch in the work tree, but
753 leave the path in the conflicted state for the user to
754 sort out.
755
756 union::
757
758 Run 3-way file level merge for text files, but take
759 lines from both versions, instead of leaving conflict
760 markers. This tends to leave the added lines in the
761 resulting file in random order and the user should
762 verify the result. Do not use this if you do not
763 understand the implications.
764
765
766 Defining a custom merge driver
767 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
768
769 The definition of a merge driver is done in the `.git/config`
770 file, not in the `gitattributes` file, so strictly speaking this
771 manual page is a wrong place to talk about it. However...
772
773 To define a custom merge driver `filfre`, add a section to your
774 `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this:
775
776 ----------------------------------------------------------------
777 [merge "filfre"]
778 name = feel-free merge driver
779 driver = filfre %O %A %B %L %P
780 recursive = binary
781 ----------------------------------------------------------------
782
783 The `merge.*.name` variable gives the driver a human-readable
784 name.
785
786 The `merge.*.driver` variable's value is used to construct a
787 command to run to merge ancestor's version (`%O`), current
788 version (`%A`) and the other branches' version (`%B`). These
789 three tokens are replaced with the names of temporary files that
790 hold the contents of these versions when the command line is
791 built. Additionally, %L will be replaced with the conflict marker
792 size (see below).
793
794 The merge driver is expected to leave the result of the merge in
795 the file named with `%A` by overwriting it, and exit with zero
796 status if it managed to merge them cleanly, or non-zero if there
797 were conflicts.
798
799 The `merge.*.recursive` variable specifies what other merge
800 driver to use when the merge driver is called for an internal
801 merge between common ancestors, when there are more than one.
802 When left unspecified, the driver itself is used for both
803 internal merge and the final merge.
804
805 The merge driver can learn the pathname in which the merged result
806 will be stored via placeholder `%P`.
807
808
809 `conflict-marker-size`
810 ^^^^^^^^^^^^^^^^^^^^^^
811
812 This attribute controls the length of conflict markers left in
813 the work tree file during a conflicted merge. Only setting to
814 the value to a positive integer has any meaningful effect.
815
816 For example, this line in `.gitattributes` can be used to tell the merge
817 machinery to leave much longer (instead of the usual 7-character-long)
818 conflict markers when merging the file `Documentation/git-merge.txt`
819 results in a conflict.
820
821 ------------------------
822 Documentation/git-merge.txt conflict-marker-size=32
823 ------------------------
824
825
826 Checking whitespace errors
827 ~~~~~~~~~~~~~~~~~~~~~~~~~~
828
829 `whitespace`
830 ^^^^^^^^^^^^
831
832 The `core.whitespace` configuration variable allows you to define what
833 'diff' and 'apply' should consider whitespace errors for all paths in
834 the project (See linkgit:git-config[1]). This attribute gives you finer
835 control per path.
836
837 Set::
838
839 Notice all types of potential whitespace errors known to Git.
840 The tab width is taken from the value of the `core.whitespace`
841 configuration variable.
842
843 Unset::
844
845 Do not notice anything as error.
846
847 Unspecified::
848
849 Use the value of the `core.whitespace` configuration variable to
850 decide what to notice as error.
851
852 String::
853
854 Specify a comma separate list of common whitespace problems to
855 notice in the same format as the `core.whitespace` configuration
856 variable.
857
858
859 Creating an archive
860 ~~~~~~~~~~~~~~~~~~~
861
862 `export-ignore`
863 ^^^^^^^^^^^^^^^
864
865 Files and directories with the attribute `export-ignore` won't be added to
866 archive files.
867
868 `export-subst`
869 ^^^^^^^^^^^^^^
870
871 If the attribute `export-subst` is set for a file then Git will expand
872 several placeholders when adding this file to an archive. The
873 expansion depends on the availability of a commit ID, i.e., if
874 linkgit:git-archive[1] has been given a tree instead of a commit or a
875 tag then no replacement will be done. The placeholders are the same
876 as those for the option `--pretty=format:` of linkgit:git-log[1],
877 except that they need to be wrapped like this: `$Format:PLACEHOLDERS$`
878 in the file. E.g. the string `$Format:%H$` will be replaced by the
879 commit hash.
880
881
882 Packing objects
883 ~~~~~~~~~~~~~~~
884
885 `delta`
886 ^^^^^^^
887
888 Delta compression will not be attempted for blobs for paths with the
889 attribute `delta` set to false.
890
891
892 Viewing files in GUI tools
893 ~~~~~~~~~~~~~~~~~~~~~~~~~~
894
895 `encoding`
896 ^^^^^^^^^^
897
898 The value of this attribute specifies the character encoding that should
899 be used by GUI tools (e.g. linkgit:gitk[1] and linkgit:git-gui[1]) to
900 display the contents of the relevant file. Note that due to performance
901 considerations linkgit:gitk[1] does not use this attribute unless you
902 manually enable per-file encodings in its options.
903
904 If this attribute is not set or has an invalid value, the value of the
905 `gui.encoding` configuration variable is used instead
906 (See linkgit:git-config[1]).
907
908
909 USING MACRO ATTRIBUTES
910 ----------------------
911
912 You do not want any end-of-line conversions applied to, nor textual diffs
913 produced for, any binary file you track. You would need to specify e.g.
914
915 ------------
916 *.jpg -text -diff
917 ------------
918
919 but that may become cumbersome, when you have many attributes. Using
920 macro attributes, you can define an attribute that, when set, also
921 sets or unsets a number of other attributes at the same time. The
922 system knows a built-in macro attribute, `binary`:
923
924 ------------
925 *.jpg binary
926 ------------
927
928 Setting the "binary" attribute also unsets the "text" and "diff"
929 attributes as above. Note that macro attributes can only be "Set",
930 though setting one might have the effect of setting or unsetting other
931 attributes or even returning other attributes to the "Unspecified"
932 state.
933
934
935 DEFINING MACRO ATTRIBUTES
936 -------------------------
937
938 Custom macro attributes can be defined only in top-level gitattributes
939 files (`$GIT_DIR/info/attributes`, the `.gitattributes` file at the
940 top level of the working tree, or the global or system-wide
941 gitattributes files), not in `.gitattributes` files in working tree
942 subdirectories. The built-in macro attribute "binary" is equivalent
943 to:
944
945 ------------
946 [attr]binary -diff -merge -text
947 ------------
948
949
950 EXAMPLE
951 -------
952
953 If you have these three `gitattributes` file:
954
955 ----------------------------------------------------------------
956 (in $GIT_DIR/info/attributes)
957
958 a* foo !bar -baz
959
960 (in .gitattributes)
961 abc foo bar baz
962
963 (in t/.gitattributes)
964 ab* merge=filfre
965 abc -foo -bar
966 *.c frotz
967 ----------------------------------------------------------------
968
969 the attributes given to path `t/abc` are computed as follows:
970
971 1. By examining `t/.gitattributes` (which is in the same
972 directory as the path in question), Git finds that the first
973 line matches. `merge` attribute is set. It also finds that
974 the second line matches, and attributes `foo` and `bar`
975 are unset.
976
977 2. Then it examines `.gitattributes` (which is in the parent
978 directory), and finds that the first line matches, but
979 `t/.gitattributes` file already decided how `merge`, `foo`
980 and `bar` attributes should be given to this path, so it
981 leaves `foo` and `bar` unset. Attribute `baz` is set.
982
983 3. Finally it examines `$GIT_DIR/info/attributes`. This file
984 is used to override the in-tree settings. The first line is
985 a match, and `foo` is set, `bar` is reverted to unspecified
986 state, and `baz` is unset.
987
988 As the result, the attributes assignment to `t/abc` becomes:
989
990 ----------------------------------------------------------------
991 foo set to true
992 bar unspecified
993 baz set to false
994 merge set to string value "filfre"
995 frotz unspecified
996 ----------------------------------------------------------------
997
998
999 SEE ALSO
1000 --------
1001 linkgit:git-check-attr[1].
1002
1003 GIT
1004 ---
1005 Part of the linkgit:git[1] suite