]>
Commit | Line | Data |
---|---|---|
88e7fdf2 JH |
1 | gitattributes(5) |
2 | ================ | |
3 | ||
4 | NAME | |
5 | ---- | |
6 | gitattributes - defining attributes per path | |
7 | ||
8 | SYNOPSIS | |
9 | -------- | |
e5b5c1d2 | 10 | $GIT_DIR/info/attributes, .gitattributes |
88e7fdf2 JH |
11 | |
12 | ||
13 | DESCRIPTION | |
14 | ----------- | |
15 | ||
16 | A `gitattributes` file is a simple text file that gives | |
17 | `attributes` to pathnames. | |
18 | ||
19 | Each line in `gitattributes` file is of form: | |
20 | ||
3f74c8e8 | 21 | pattern attr1 attr2 ... |
88e7fdf2 | 22 | |
3f74c8e8 JS |
23 | That is, a pattern followed by an attributes list, |
24 | separated by whitespaces. When the pattern matches the | |
88e7fdf2 JH |
25 | path in question, the attributes listed on the line are given to |
26 | the path. | |
27 | ||
28 | Each attribute can be in one of these states for a given path: | |
29 | ||
30 | Set:: | |
31 | ||
32 | The path has the attribute with special value "true"; | |
33 | this is specified by listing only the name of the | |
34 | attribute in the attribute list. | |
35 | ||
36 | Unset:: | |
37 | ||
38 | The path has the attribute with special value "false"; | |
39 | this is specified by listing the name of the attribute | |
40 | prefixed with a dash `-` in the attribute list. | |
41 | ||
42 | Set to a value:: | |
43 | ||
44 | The path has the attribute with specified string value; | |
45 | this is specified by listing the name of the attribute | |
46 | followed by an equal sign `=` and its value in the | |
47 | attribute list. | |
48 | ||
49 | Unspecified:: | |
50 | ||
3f74c8e8 | 51 | No pattern matches the path, and nothing says if |
b9d14ffb JH |
52 | the path has or does not have the attribute, the |
53 | attribute for the path is said to be Unspecified. | |
88e7fdf2 | 54 | |
3f74c8e8 | 55 | When more than one pattern matches the path, a later line |
b9d14ffb | 56 | overrides an earlier line. This overriding is done per |
3f74c8e8 JS |
57 | attribute. The rules how the pattern matches paths are the |
58 | same as in `.gitignore` files; see linkgit:gitignore[5]. | |
88e7fdf2 JH |
59 | |
60 | When deciding what attributes are assigned to a path, git | |
61 | consults `$GIT_DIR/info/attributes` file (which has the highest | |
62 | precedence), `.gitattributes` file in the same directory as the | |
20ff3ec2 JM |
63 | path in question, and its parent directories up to the toplevel of the |
64 | work tree (the further the directory that contains `.gitattributes` | |
65 | is from the path in question, the lower its precedence). | |
88e7fdf2 | 66 | |
90b22907 JK |
67 | If you wish to affect only a single repository (i.e., to assign |
68 | attributes to files that are particular to one user's workflow), then | |
69 | attributes should be placed in the `$GIT_DIR/info/attributes` file. | |
70 | Attributes which should be version-controlled and distributed to other | |
71 | repositories (i.e., attributes of interest to all users) should go into | |
72 | `.gitattributes` files. | |
73 | ||
88e7fdf2 JH |
74 | Sometimes you would need to override an setting of an attribute |
75 | for a path to `unspecified` state. This can be done by listing | |
76 | the name of the attribute prefixed with an exclamation point `!`. | |
77 | ||
78 | ||
79 | EFFECTS | |
80 | ------- | |
81 | ||
82 | Certain operations by git can be influenced by assigning | |
ae7aa499 JH |
83 | particular attributes to a path. Currently, the following |
84 | operations are attributes-aware. | |
88e7fdf2 JH |
85 | |
86 | Checking-out and checking-in | |
87 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
88 | ||
3fed15f5 | 89 | These attributes affect how the contents stored in the |
88e7fdf2 | 90 | repository are copied to the working tree files when commands |
0b444cdb | 91 | such as 'git checkout' and 'git merge' run. They also affect how |
88e7fdf2 | 92 | git stores the contents you prepare in the working tree in the |
0b444cdb | 93 | repository upon 'git add' and 'git commit'. |
88e7fdf2 | 94 | |
5ec3e670 | 95 | `text` |
3fed15f5 JH |
96 | ^^^^^^ |
97 | ||
fd6cce9e EB |
98 | This attribute enables and controls end-of-line normalization. When a |
99 | text file is normalized, its line endings are converted to LF in the | |
100 | repository. To control what line ending style is used in the working | |
101 | directory, use the `eol` attribute for a single file and the | |
942e7747 | 102 | `core.eol` configuration variable for all text files. |
3fed15f5 | 103 | |
88e7fdf2 JH |
104 | Set:: |
105 | ||
5ec3e670 | 106 | Setting the `text` attribute on a path enables end-of-line |
fd6cce9e EB |
107 | normalization and marks the path as a text file. End-of-line |
108 | conversion takes place without guessing the content type. | |
88e7fdf2 JH |
109 | |
110 | Unset:: | |
111 | ||
5ec3e670 | 112 | Unsetting the `text` attribute on a path tells git not to |
bbb896d8 | 113 | attempt any end-of-line conversion upon checkin or checkout. |
88e7fdf2 | 114 | |
fd6cce9e | 115 | Set to string value "auto":: |
88e7fdf2 | 116 | |
5ec3e670 | 117 | When `text` is set to "auto", the path is marked for automatic |
fd6cce9e EB |
118 | end-of-line normalization. If git decides that the content is |
119 | text, its line endings are normalized to LF on checkin. | |
88e7fdf2 | 120 | |
88e7fdf2 JH |
121 | Unspecified:: |
122 | ||
942e7747 EB |
123 | If the `text` attribute is unspecified, git uses the |
124 | `core.autocrlf` configuration variable to determine if the | |
125 | file should be converted. | |
88e7fdf2 | 126 | |
5ec3e670 | 127 | Any other value causes git to act as if `text` has been left |
fd6cce9e | 128 | unspecified. |
88e7fdf2 | 129 | |
fd6cce9e EB |
130 | `eol` |
131 | ^^^^^ | |
88e7fdf2 | 132 | |
fd6cce9e EB |
133 | This attribute sets a specific line-ending style to be used in the |
134 | working directory. It enables end-of-line normalization without any | |
942e7747 | 135 | content checks, effectively setting the `text` attribute. |
88e7fdf2 | 136 | |
fd6cce9e | 137 | Set to string value "crlf":: |
88e7fdf2 | 138 | |
942e7747 EB |
139 | This setting forces git to normalize line endings for this |
140 | file on checkin and convert them to CRLF when the file is | |
141 | checked out. | |
fd6cce9e EB |
142 | |
143 | Set to string value "lf":: | |
144 | ||
145 | This setting forces git to normalize line endings to LF on | |
146 | checkin and prevents conversion to CRLF when the file is | |
942e7747 | 147 | checked out. |
5ec3e670 EB |
148 | |
149 | Backwards compatibility with `crlf` attribute | |
150 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
151 | ||
152 | For backwards compatibility, the `crlf` attribute is interpreted as | |
153 | follows: | |
154 | ||
155 | ------------------------ | |
156 | crlf text | |
157 | -crlf -text | |
158 | crlf=input eol=lf | |
159 | ------------------------ | |
fd6cce9e EB |
160 | |
161 | End-of-line conversion | |
162 | ^^^^^^^^^^^^^^^^^^^^^^ | |
163 | ||
164 | While git normally leaves file contents alone, it can be configured to | |
165 | normalize line endings to LF in the repository and, optionally, to | |
166 | convert them to CRLF when files are checked out. | |
167 | ||
168 | Here is an example that will make git normalize .txt, .vcproj and .sh | |
169 | files, ensure that .vcproj files have CRLF and .sh files have LF in | |
170 | the working directory, and prevent .jpg files from being normalized | |
171 | regardless of their content. | |
172 | ||
173 | ------------------------ | |
5ec3e670 | 174 | *.txt text |
fd6cce9e EB |
175 | *.vcproj eol=crlf |
176 | *.sh eol=lf | |
5ec3e670 | 177 | *.jpg -text |
fd6cce9e EB |
178 | ------------------------ |
179 | ||
180 | Other source code management systems normalize all text files in their | |
181 | repositories, and there are two ways to enable similar automatic | |
182 | normalization in git. | |
183 | ||
184 | If you simply want to have CRLF line endings in your working directory | |
185 | regardless of the repository you are working with, you can set the | |
186 | config variable "core.autocrlf" without changing any attributes. | |
187 | ||
188 | ------------------------ | |
189 | [core] | |
190 | autocrlf = true | |
191 | ------------------------ | |
192 | ||
193 | This does not force normalization of all text files, but does ensure | |
194 | that text files that you introduce to the repository have their line | |
195 | endings normalized to LF when they are added, and that files that are | |
942e7747 | 196 | already normalized in the repository stay normalized. |
fd6cce9e EB |
197 | |
198 | If you want to interoperate with a source code management system that | |
199 | enforces end-of-line normalization, or you simply want all text files | |
5ec3e670 | 200 | in your repository to be normalized, you should instead set the `text` |
fd6cce9e | 201 | attribute to "auto" for _all_ files. |
88e7fdf2 | 202 | |
fd6cce9e | 203 | ------------------------ |
5ec3e670 | 204 | * text=auto |
fd6cce9e EB |
205 | ------------------------ |
206 | ||
207 | This ensures that all files that git considers to be text will have | |
942e7747 EB |
208 | normalized (LF) line endings in the repository. The `core.eol` |
209 | configuration variable controls which line endings git will use for | |
210 | normalized files in your working directory; the default is to use the | |
211 | native line ending for your platform, or CRLF if `core.autocrlf` is | |
212 | set. | |
fd6cce9e | 213 | |
5ec3e670 | 214 | NOTE: When `text=auto` normalization is enabled in an existing |
fd6cce9e EB |
215 | repository, any text files containing CRLFs should be normalized. If |
216 | they are not they will be normalized the next time someone tries to | |
217 | change them, causing unfortunate misattribution. From a clean working | |
218 | directory: | |
219 | ||
220 | ------------------------------------------------- | |
5ec3e670 | 221 | $ echo "* text=auto" >>.gitattributes |
fd6cce9e EB |
222 | $ rm .git/index # Remove the index to force git to |
223 | $ git reset # re-scan the working directory | |
224 | $ git status # Show files that will be normalized | |
225 | $ git add -u | |
226 | $ git add .gitattributes | |
227 | $ git commit -m "Introduce end-of-line normalization" | |
228 | ------------------------------------------------- | |
229 | ||
230 | If any files that should not be normalized show up in 'git status', | |
5ec3e670 | 231 | unset their `text` attribute before running 'git add -u'. |
fd6cce9e EB |
232 | |
233 | ------------------------ | |
5ec3e670 | 234 | manual.pdf -text |
fd6cce9e | 235 | ------------------------ |
88e7fdf2 | 236 | |
fd6cce9e EB |
237 | Conversely, text files that git does not detect can have normalization |
238 | enabled manually. | |
88e7fdf2 | 239 | |
fd6cce9e | 240 | ------------------------ |
5ec3e670 | 241 | weirdchars.txt text |
fd6cce9e | 242 | ------------------------ |
88e7fdf2 | 243 | |
21e5ad50 SP |
244 | If `core.safecrlf` is set to "true" or "warn", git verifies if |
245 | the conversion is reversible for the current setting of | |
246 | `core.autocrlf`. For "true", git rejects irreversible | |
247 | conversions; for "warn", git only prints a warning but accepts | |
248 | an irreversible conversion. The safety triggers to prevent such | |
249 | a conversion done to the files in the work tree, but there are a | |
250 | few exceptions. Even though... | |
251 | ||
0b444cdb | 252 | - 'git add' itself does not touch the files in the work tree, the |
21e5ad50 SP |
253 | next checkout would, so the safety triggers; |
254 | ||
0b444cdb | 255 | - 'git apply' to update a text file with a patch does touch the files |
21e5ad50 SP |
256 | in the work tree, but the operation is about text files and CRLF |
257 | conversion is about fixing the line ending inconsistencies, so the | |
258 | safety does not trigger; | |
259 | ||
0b444cdb TR |
260 | - 'git diff' itself does not touch the files in the work tree, it is |
261 | often run to inspect the changes you intend to next 'git add'. To | |
21e5ad50 SP |
262 | catch potential problems early, safety triggers. |
263 | ||
88e7fdf2 | 264 | |
3fed15f5 JH |
265 | `ident` |
266 | ^^^^^^^ | |
267 | ||
2c850f12 JK |
268 | When the attribute `ident` is set for a path, git replaces |
269 | `$Id$` in the blob object with `$Id:`, followed by the | |
3fed15f5 JH |
270 | 40-character hexadecimal blob object name, followed by a dollar |
271 | sign `$` upon checkout. Any byte sequence that begins with | |
af9b54bb AP |
272 | `$Id:` and ends with `$` in the worktree file is replaced |
273 | with `$Id$` upon check-in. | |
3fed15f5 JH |
274 | |
275 | ||
aa4ed402 JH |
276 | `filter` |
277 | ^^^^^^^^ | |
278 | ||
c05ef938 | 279 | A `filter` attribute can be set to a string value that names a |
aa4ed402 JH |
280 | filter driver specified in the configuration. |
281 | ||
c05ef938 | 282 | A filter driver consists of a `clean` command and a `smudge` |
aa4ed402 | 283 | command, either of which can be left unspecified. Upon |
c05ef938 WC |
284 | checkout, when the `smudge` command is specified, the command is |
285 | fed the blob object from its standard input, and its standard | |
286 | output is used to update the worktree file. Similarly, the | |
287 | `clean` command is used to convert the contents of worktree file | |
288 | upon checkin. | |
aa4ed402 | 289 | |
c05ef938 | 290 | A missing filter driver definition in the config is not an error |
aa4ed402 JH |
291 | but makes the filter a no-op passthru. |
292 | ||
293 | The content filtering is done to massage the content into a | |
294 | shape that is more convenient for the platform, filesystem, and | |
c05ef938 | 295 | the user to use. The key phrase here is "more convenient" and not |
4d84aff3 JS |
296 | "turning something unusable into usable". In other words, the |
297 | intent is that if someone unsets the filter driver definition, | |
298 | or does not have the appropriate filter program, the project | |
299 | should still be usable. | |
aa4ed402 | 300 | |
d79f5d17 NS |
301 | For example, in .gitattributes, you would assign the `filter` |
302 | attribute for paths. | |
303 | ||
304 | ------------------------ | |
305 | *.c filter=indent | |
306 | ------------------------ | |
307 | ||
308 | Then you would define a "filter.indent.clean" and "filter.indent.smudge" | |
309 | configuration in your .git/config to specify a pair of commands to | |
310 | modify the contents of C programs when the source files are checked | |
311 | in ("clean" is run) and checked out (no change is made because the | |
312 | command is "cat"). | |
313 | ||
314 | ------------------------ | |
315 | [filter "indent"] | |
316 | clean = indent | |
317 | smudge = cat | |
318 | ------------------------ | |
319 | ||
f217f0e8 EB |
320 | For best results, `clean` should not alter its output further if it is |
321 | run twice ("clean->clean" should be equivalent to "clean"), and | |
322 | multiple `smudge` commands should not alter `clean`'s output | |
323 | ("smudge->smudge->clean" should be equivalent to "clean"). See the | |
324 | section on merging below. | |
325 | ||
326 | The "indent" filter is well-behaved in this regard: it will not modify | |
327 | input that is already correctly indented. In this case, the lack of a | |
328 | smudge filter means that the clean filter _must_ accept its own output | |
329 | without modifying it. | |
330 | ||
aa4ed402 JH |
331 | |
332 | Interaction between checkin/checkout attributes | |
333 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
334 | ||
335 | In the check-in codepath, the worktree file is first converted | |
336 | with `filter` driver (if specified and corresponding driver | |
337 | defined), then the result is processed with `ident` (if | |
5ec3e670 | 338 | specified), and then finally with `text` (again, if specified |
aa4ed402 JH |
339 | and applicable). |
340 | ||
341 | In the check-out codepath, the blob content is first converted | |
5ec3e670 | 342 | with `text`, and then `ident` and fed to `filter`. |
aa4ed402 JH |
343 | |
344 | ||
f217f0e8 EB |
345 | Merging branches with differing checkin/checkout attributes |
346 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
347 | ||
348 | If you have added attributes to a file that cause the canonical | |
349 | repository format for that file to change, such as adding a | |
350 | clean/smudge filter or text/eol/ident attributes, merging anything | |
351 | where the attribute is not in place would normally cause merge | |
352 | conflicts. | |
353 | ||
354 | To prevent these unnecessary merge conflicts, git can be told to run a | |
355 | virtual check-out and check-in of all three stages of a file when | |
356 | resolving a three-way merge by setting the `merge.renormalize` | |
357 | configuration variable. This prevents changes caused by check-in | |
358 | conversion from causing spurious merge conflicts when a converted file | |
359 | is merged with an unconverted file. | |
360 | ||
361 | As long as a "smudge->clean" results in the same output as a "clean" | |
362 | even on files that are already smudged, this strategy will | |
363 | automatically resolve all filter-related conflicts. Filters that do | |
364 | not act in this way may cause additional merge conflicts that must be | |
365 | resolved manually. | |
366 | ||
367 | ||
88e7fdf2 JH |
368 | Generating diff text |
369 | ~~~~~~~~~~~~~~~~~~~~ | |
370 | ||
4f73e240 JN |
371 | `diff` |
372 | ^^^^^^ | |
373 | ||
678852d9 JK |
374 | The attribute `diff` affects how 'git' generates diffs for particular |
375 | files. It can tell git whether to generate a textual patch for the path | |
376 | or to treat the path as a binary file. It can also affect what line is | |
377 | shown on the hunk header `@@ -k,l +n,m @@` line, tell git to use an | |
378 | external command to generate the diff, or ask git to convert binary | |
379 | files to a text format before generating the diff. | |
88e7fdf2 JH |
380 | |
381 | Set:: | |
382 | ||
383 | A path to which the `diff` attribute is set is treated | |
384 | as text, even when they contain byte values that | |
385 | normally never appear in text files, such as NUL. | |
386 | ||
387 | Unset:: | |
388 | ||
389 | A path to which the `diff` attribute is unset will | |
678852d9 JK |
390 | generate `Binary files differ` (or a binary patch, if |
391 | binary patches are enabled). | |
88e7fdf2 JH |
392 | |
393 | Unspecified:: | |
394 | ||
395 | A path to which the `diff` attribute is unspecified | |
396 | first gets its contents inspected, and if it looks like | |
397 | text, it is treated as text. Otherwise it would | |
398 | generate `Binary files differ`. | |
399 | ||
2cc3167c JH |
400 | String:: |
401 | ||
678852d9 JK |
402 | Diff is shown using the specified diff driver. Each driver may |
403 | specify one or more options, as described in the following | |
404 | section. The options for the diff driver "foo" are defined | |
405 | by the configuration variables in the "diff.foo" section of the | |
406 | git config file. | |
2cc3167c JH |
407 | |
408 | ||
678852d9 JK |
409 | Defining an external diff driver |
410 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
2cc3167c JH |
411 | |
412 | The definition of a diff driver is done in `gitconfig`, not | |
413 | `gitattributes` file, so strictly speaking this manual page is a | |
414 | wrong place to talk about it. However... | |
415 | ||
678852d9 | 416 | To define an external diff driver `jcdiff`, add a section to your |
2cc3167c JH |
417 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: |
418 | ||
419 | ---------------------------------------------------------------- | |
420 | [diff "jcdiff"] | |
421 | command = j-c-diff | |
422 | ---------------------------------------------------------------- | |
423 | ||
424 | When git needs to show you a diff for the path with `diff` | |
425 | attribute set to `jcdiff`, it calls the command you specified | |
426 | with the above configuration, i.e. `j-c-diff`, with 7 | |
427 | parameters, just like `GIT_EXTERNAL_DIFF` program is called. | |
9e1f0a85 | 428 | See linkgit:git[1] for details. |
88e7fdf2 JH |
429 | |
430 | ||
ae7aa499 JH |
431 | Defining a custom hunk-header |
432 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
433 | ||
c882c01e | 434 | Each group of changes (called a "hunk") in the textual diff output |
ae7aa499 JH |
435 | is prefixed with a line of the form: |
436 | ||
437 | @@ -k,l +n,m @@ TEXT | |
438 | ||
c882c01e GD |
439 | This is called a 'hunk header'. The "TEXT" portion is by default a line |
440 | that begins with an alphabet, an underscore or a dollar sign; this | |
441 | matches what GNU 'diff -p' output uses. This default selection however | |
442 | is not suited for some contents, and you can use a customized pattern | |
443 | to make a selection. | |
ae7aa499 | 444 | |
c882c01e | 445 | First, in .gitattributes, you would assign the `diff` attribute |
ae7aa499 JH |
446 | for paths. |
447 | ||
448 | ------------------------ | |
449 | *.tex diff=tex | |
450 | ------------------------ | |
451 | ||
edb7e82f | 452 | Then, you would define a "diff.tex.xfuncname" configuration to |
ae7aa499 | 453 | specify a regular expression that matches a line that you would |
c4c86d23 JK |
454 | want to appear as the hunk header "TEXT". Add a section to your |
455 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
ae7aa499 JH |
456 | |
457 | ------------------------ | |
458 | [diff "tex"] | |
45d9414f | 459 | xfuncname = "^(\\\\(sub)*section\\{.*)$" |
ae7aa499 JH |
460 | ------------------------ |
461 | ||
462 | Note. A single level of backslashes are eaten by the | |
463 | configuration file parser, so you would need to double the | |
464 | backslashes; the pattern above picks a line that begins with a | |
02783075 | 465 | backslash, and zero or more occurrences of `sub` followed by |
ae7aa499 JH |
466 | `section` followed by open brace, to the end of line. |
467 | ||
468 | There are a few built-in patterns to make this easier, and `tex` | |
469 | is one of them, so you do not have to write the above in your | |
470 | configuration file (you still need to enable this with the | |
d08ed6d6 GH |
471 | attribute mechanism, via `.gitattributes`). The following built in |
472 | patterns are available: | |
473 | ||
23b5beb2 GH |
474 | - `bibtex` suitable for files with BibTeX coded references. |
475 | ||
80c49c3d TR |
476 | - `cpp` suitable for source code in the C and C++ languages. |
477 | ||
b221207d PO |
478 | - `csharp` suitable for source code in the C# language. |
479 | ||
909a5494 BC |
480 | - `fortran` suitable for source code in the Fortran language. |
481 | ||
af9ce1ff AE |
482 | - `html` suitable for HTML/XHTML documents. |
483 | ||
b66e00f1 | 484 | - `java` suitable for source code in the Java language. |
d08ed6d6 | 485 | |
5d1e958e JS |
486 | - `objc` suitable for source code in the Objective-C language. |
487 | ||
d08ed6d6 GH |
488 | - `pascal` suitable for source code in the Pascal/Delphi language. |
489 | ||
af9ce1ff AE |
490 | - `php` suitable for source code in the PHP language. |
491 | ||
7c17205b KS |
492 | - `python` suitable for source code in the Python language. |
493 | ||
d08ed6d6 GH |
494 | - `ruby` suitable for source code in the Ruby language. |
495 | ||
496 | - `tex` suitable for source code for LaTeX documents. | |
ae7aa499 JH |
497 | |
498 | ||
80c49c3d TR |
499 | Customizing word diff |
500 | ^^^^^^^^^^^^^^^^^^^^^ | |
501 | ||
882749a0 | 502 | You can customize the rules that `git diff --word-diff` uses to |
80c49c3d | 503 | split words in a line, by specifying an appropriate regular expression |
ae3b970a | 504 | in the "diff.*.wordRegex" configuration variable. For example, in TeX |
80c49c3d TR |
505 | a backslash followed by a sequence of letters forms a command, but |
506 | several such commands can be run together without intervening | |
c4c86d23 JK |
507 | whitespace. To separate them, use a regular expression in your |
508 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
80c49c3d TR |
509 | |
510 | ------------------------ | |
511 | [diff "tex"] | |
ae3b970a | 512 | wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+" |
80c49c3d TR |
513 | ------------------------ |
514 | ||
515 | A built-in pattern is provided for all languages listed in the | |
516 | previous section. | |
517 | ||
518 | ||
678852d9 JK |
519 | Performing text diffs of binary files |
520 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
521 | ||
522 | Sometimes it is desirable to see the diff of a text-converted | |
523 | version of some binary files. For example, a word processor | |
524 | document can be converted to an ASCII text representation, and | |
525 | the diff of the text shown. Even though this conversion loses | |
526 | some information, the resulting diff is useful for human | |
527 | viewing (but cannot be applied directly). | |
528 | ||
529 | The `textconv` config option is used to define a program for | |
530 | performing such a conversion. The program should take a single | |
531 | argument, the name of a file to convert, and produce the | |
532 | resulting text on stdout. | |
533 | ||
534 | For example, to show the diff of the exif information of a | |
535 | file instead of the binary information (assuming you have the | |
c4c86d23 JK |
536 | exif tool installed), add the following section to your |
537 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file): | |
678852d9 JK |
538 | |
539 | ------------------------ | |
540 | [diff "jpg"] | |
541 | textconv = exif | |
542 | ------------------------ | |
543 | ||
544 | NOTE: The text conversion is generally a one-way conversion; | |
545 | in this example, we lose the actual image contents and focus | |
546 | just on the text data. This means that diffs generated by | |
547 | textconv are _not_ suitable for applying. For this reason, | |
548 | only `git diff` and the `git log` family of commands (i.e., | |
549 | log, whatchanged, show) will perform text conversion. `git | |
550 | format-patch` will never generate this output. If you want to | |
551 | send somebody a text-converted diff of a binary file (e.g., | |
552 | because it quickly conveys the changes you have made), you | |
553 | should generate it separately and send it as a comment _in | |
554 | addition to_ the usual binary diff that you might send. | |
555 | ||
d9bae1a1 JK |
556 | Because text conversion can be slow, especially when doing a |
557 | large number of them with `git log -p`, git provides a mechanism | |
558 | to cache the output and use it in future diffs. To enable | |
559 | caching, set the "cachetextconv" variable in your diff driver's | |
560 | config. For example: | |
561 | ||
562 | ------------------------ | |
563 | [diff "jpg"] | |
564 | textconv = exif | |
565 | cachetextconv = true | |
566 | ------------------------ | |
567 | ||
568 | This will cache the result of running "exif" on each blob | |
569 | indefinitely. If you change the textconv config variable for a | |
570 | diff driver, git will automatically invalidate the cache entries | |
571 | and re-run the textconv filter. If you want to invalidate the | |
572 | cache manually (e.g., because your version of "exif" was updated | |
573 | and now produces better output), you can remove the cache | |
574 | manually with `git update-ref -d refs/notes/textconv/jpg` (where | |
575 | "jpg" is the name of the diff driver, as in the example above). | |
678852d9 | 576 | |
88e7fdf2 JH |
577 | Performing a three-way merge |
578 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
579 | ||
4f73e240 JN |
580 | `merge` |
581 | ^^^^^^^ | |
582 | ||
88e7fdf2 JH |
583 | The attribute `merge` affects how three versions of a file is |
584 | merged when a file-level merge is necessary during `git merge`, | |
57f6ec02 | 585 | and other commands such as `git revert` and `git cherry-pick`. |
88e7fdf2 JH |
586 | |
587 | Set:: | |
588 | ||
589 | Built-in 3-way merge driver is used to merge the | |
2fd02c92 | 590 | contents in a way similar to 'merge' command of `RCS` |
88e7fdf2 JH |
591 | suite. This is suitable for ordinary text files. |
592 | ||
593 | Unset:: | |
594 | ||
595 | Take the version from the current branch as the | |
596 | tentative merge result, and declare that the merge has | |
597 | conflicts. This is suitable for binary files that does | |
598 | not have a well-defined merge semantics. | |
599 | ||
600 | Unspecified:: | |
601 | ||
602 | By default, this uses the same built-in 3-way merge | |
603 | driver as is the case the `merge` attribute is set. | |
604 | However, `merge.default` configuration variable can name | |
605 | different merge driver to be used for paths to which the | |
606 | `merge` attribute is unspecified. | |
607 | ||
2cc3167c | 608 | String:: |
88e7fdf2 JH |
609 | |
610 | 3-way merge is performed using the specified custom | |
611 | merge driver. The built-in 3-way merge driver can be | |
612 | explicitly specified by asking for "text" driver; the | |
613 | built-in "take the current branch" driver can be | |
b9d14ffb | 614 | requested with "binary". |
88e7fdf2 JH |
615 | |
616 | ||
0e545f75 JH |
617 | Built-in merge drivers |
618 | ^^^^^^^^^^^^^^^^^^^^^^ | |
619 | ||
620 | There are a few built-in low-level merge drivers defined that | |
621 | can be asked for via the `merge` attribute. | |
622 | ||
623 | text:: | |
624 | ||
625 | Usual 3-way file level merge for text files. Conflicted | |
626 | regions are marked with conflict markers `<<<<<<<`, | |
627 | `=======` and `>>>>>>>`. The version from your branch | |
628 | appears before the `=======` marker, and the version | |
629 | from the merged branch appears after the `=======` | |
630 | marker. | |
631 | ||
632 | binary:: | |
633 | ||
634 | Keep the version from your branch in the work tree, but | |
635 | leave the path in the conflicted state for the user to | |
636 | sort out. | |
637 | ||
638 | union:: | |
639 | ||
640 | Run 3-way file level merge for text files, but take | |
641 | lines from both versions, instead of leaving conflict | |
642 | markers. This tends to leave the added lines in the | |
643 | resulting file in random order and the user should | |
644 | verify the result. Do not use this if you do not | |
645 | understand the implications. | |
646 | ||
647 | ||
88e7fdf2 JH |
648 | Defining a custom merge driver |
649 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
650 | ||
0e545f75 JH |
651 | The definition of a merge driver is done in the `.git/config` |
652 | file, not in the `gitattributes` file, so strictly speaking this | |
653 | manual page is a wrong place to talk about it. However... | |
88e7fdf2 JH |
654 | |
655 | To define a custom merge driver `filfre`, add a section to your | |
656 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
657 | ||
658 | ---------------------------------------------------------------- | |
659 | [merge "filfre"] | |
660 | name = feel-free merge driver | |
661 | driver = filfre %O %A %B | |
662 | recursive = binary | |
663 | ---------------------------------------------------------------- | |
664 | ||
665 | The `merge.*.name` variable gives the driver a human-readable | |
666 | name. | |
667 | ||
668 | The `merge.*.driver` variable's value is used to construct a | |
669 | command to run to merge ancestor's version (`%O`), current | |
670 | version (`%A`) and the other branches' version (`%B`). These | |
671 | three tokens are replaced with the names of temporary files that | |
672 | hold the contents of these versions when the command line is | |
16758621 BW |
673 | built. Additionally, %L will be replaced with the conflict marker |
674 | size (see below). | |
88e7fdf2 JH |
675 | |
676 | The merge driver is expected to leave the result of the merge in | |
677 | the file named with `%A` by overwriting it, and exit with zero | |
678 | status if it managed to merge them cleanly, or non-zero if there | |
679 | were conflicts. | |
680 | ||
681 | The `merge.*.recursive` variable specifies what other merge | |
682 | driver to use when the merge driver is called for an internal | |
683 | merge between common ancestors, when there are more than one. | |
684 | When left unspecified, the driver itself is used for both | |
685 | internal merge and the final merge. | |
686 | ||
687 | ||
4c734803 JH |
688 | `conflict-marker-size` |
689 | ^^^^^^^^^^^^^^^^^^^^^^ | |
690 | ||
691 | This attribute controls the length of conflict markers left in | |
692 | the work tree file during a conflicted merge. Only setting to | |
693 | the value to a positive integer has any meaningful effect. | |
694 | ||
695 | For example, this line in `.gitattributes` can be used to tell the merge | |
696 | machinery to leave much longer (instead of the usual 7-character-long) | |
697 | conflict markers when merging the file `Documentation/git-merge.txt` | |
698 | results in a conflict. | |
699 | ||
700 | ------------------------ | |
701 | Documentation/git-merge.txt conflict-marker-size=32 | |
702 | ------------------------ | |
703 | ||
704 | ||
cf1b7869 JH |
705 | Checking whitespace errors |
706 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
707 | ||
708 | `whitespace` | |
709 | ^^^^^^^^^^^^ | |
710 | ||
711 | The `core.whitespace` configuration variable allows you to define what | |
2fd02c92 | 712 | 'diff' and 'apply' should consider whitespace errors for all paths in |
5162e697 | 713 | the project (See linkgit:git-config[1]). This attribute gives you finer |
cf1b7869 JH |
714 | control per path. |
715 | ||
716 | Set:: | |
717 | ||
718 | Notice all types of potential whitespace errors known to git. | |
719 | ||
720 | Unset:: | |
721 | ||
722 | Do not notice anything as error. | |
723 | ||
724 | Unspecified:: | |
725 | ||
726 | Use the value of `core.whitespace` configuration variable to | |
727 | decide what to notice as error. | |
728 | ||
729 | String:: | |
730 | ||
731 | Specify a comma separate list of common whitespace problems to | |
732 | notice in the same format as `core.whitespace` configuration | |
733 | variable. | |
734 | ||
735 | ||
8a33dd8b JH |
736 | Creating an archive |
737 | ~~~~~~~~~~~~~~~~~~~ | |
738 | ||
08b51f51 JH |
739 | `export-ignore` |
740 | ^^^^^^^^^^^^^^^ | |
741 | ||
742 | Files and directories with the attribute `export-ignore` won't be added to | |
743 | archive files. | |
744 | ||
8a33dd8b JH |
745 | `export-subst` |
746 | ^^^^^^^^^^^^^^ | |
747 | ||
748 | If the attribute `export-subst` is set for a file then git will expand | |
749 | several placeholders when adding this file to an archive. The | |
08b51f51 | 750 | expansion depends on the availability of a commit ID, i.e., if |
8a33dd8b JH |
751 | linkgit:git-archive[1] has been given a tree instead of a commit or a |
752 | tag then no replacement will be done. The placeholders are the same | |
753 | as those for the option `--pretty=format:` of linkgit:git-log[1], | |
754 | except that they need to be wrapped like this: `$Format:PLACEHOLDERS$` | |
755 | in the file. E.g. the string `$Format:%H$` will be replaced by the | |
756 | commit hash. | |
757 | ||
758 | ||
975457f1 NG |
759 | Packing objects |
760 | ~~~~~~~~~~~~~~~ | |
761 | ||
762 | `delta` | |
763 | ^^^^^^^ | |
764 | ||
765 | Delta compression will not be attempted for blobs for paths with the | |
766 | attribute `delta` set to false. | |
767 | ||
768 | ||
a2df1fb2 AG |
769 | Viewing files in GUI tools |
770 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
771 | ||
772 | `encoding` | |
773 | ^^^^^^^^^^ | |
774 | ||
775 | The value of this attribute specifies the character encoding that should | |
776 | be used by GUI tools (e.g. linkgit:gitk[1] and linkgit:git-gui[1]) to | |
777 | display the contents of the relevant file. Note that due to performance | |
778 | considerations linkgit:gitk[1] does not use this attribute unless you | |
779 | manually enable per-file encodings in its options. | |
780 | ||
781 | If this attribute is not set or has an invalid value, the value of the | |
782 | `gui.encoding` configuration variable is used instead | |
783 | (See linkgit:git-config[1]). | |
784 | ||
785 | ||
bbb896d8 JH |
786 | USING ATTRIBUTE MACROS |
787 | ---------------------- | |
788 | ||
789 | You do not want any end-of-line conversions applied to, nor textual diffs | |
790 | produced for, any binary file you track. You would need to specify e.g. | |
791 | ||
792 | ------------ | |
5ec3e670 | 793 | *.jpg -text -diff |
bbb896d8 JH |
794 | ------------ |
795 | ||
796 | but that may become cumbersome, when you have many attributes. Using | |
797 | attribute macros, you can specify groups of attributes set or unset at | |
798 | the same time. The system knows a built-in attribute macro, `binary`: | |
799 | ||
800 | ------------ | |
801 | *.jpg binary | |
802 | ------------ | |
803 | ||
804 | which is equivalent to the above. Note that the attribute macros can only | |
805 | be "Set" (see the above example that sets "binary" macro as if it were an | |
5ec3e670 | 806 | ordinary attribute --- setting it in turn unsets "text" and "diff"). |
bbb896d8 JH |
807 | |
808 | ||
809 | DEFINING ATTRIBUTE MACROS | |
810 | ------------------------- | |
811 | ||
812 | Custom attribute macros can be defined only in the `.gitattributes` file | |
813 | at the toplevel (i.e. not in any subdirectory). The built-in attribute | |
814 | macro "binary" is equivalent to: | |
815 | ||
816 | ------------ | |
5ec3e670 | 817 | [attr]binary -diff -text |
bbb896d8 JH |
818 | ------------ |
819 | ||
820 | ||
88e7fdf2 JH |
821 | EXAMPLE |
822 | ------- | |
823 | ||
824 | If you have these three `gitattributes` file: | |
825 | ||
826 | ---------------------------------------------------------------- | |
827 | (in $GIT_DIR/info/attributes) | |
828 | ||
829 | a* foo !bar -baz | |
830 | ||
831 | (in .gitattributes) | |
832 | abc foo bar baz | |
833 | ||
834 | (in t/.gitattributes) | |
835 | ab* merge=filfre | |
836 | abc -foo -bar | |
837 | *.c frotz | |
838 | ---------------------------------------------------------------- | |
839 | ||
840 | the attributes given to path `t/abc` are computed as follows: | |
841 | ||
842 | 1. By examining `t/.gitattributes` (which is in the same | |
02783075 | 843 | directory as the path in question), git finds that the first |
88e7fdf2 JH |
844 | line matches. `merge` attribute is set. It also finds that |
845 | the second line matches, and attributes `foo` and `bar` | |
846 | are unset. | |
847 | ||
848 | 2. Then it examines `.gitattributes` (which is in the parent | |
849 | directory), and finds that the first line matches, but | |
850 | `t/.gitattributes` file already decided how `merge`, `foo` | |
851 | and `bar` attributes should be given to this path, so it | |
852 | leaves `foo` and `bar` unset. Attribute `baz` is set. | |
853 | ||
5c759f96 | 854 | 3. Finally it examines `$GIT_DIR/info/attributes`. This file |
88e7fdf2 JH |
855 | is used to override the in-tree settings. The first line is |
856 | a match, and `foo` is set, `bar` is reverted to unspecified | |
857 | state, and `baz` is unset. | |
858 | ||
02783075 | 859 | As the result, the attributes assignment to `t/abc` becomes: |
88e7fdf2 JH |
860 | |
861 | ---------------------------------------------------------------- | |
862 | foo set to true | |
863 | bar unspecified | |
864 | baz set to false | |
865 | merge set to string value "filfre" | |
866 | frotz unspecified | |
867 | ---------------------------------------------------------------- | |
868 | ||
869 | ||
8460b2fc | 870 | |
88e7fdf2 JH |
871 | GIT |
872 | --- | |
9e1f0a85 | 873 | Part of the linkgit:git[1] suite |