]>
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` | |
6df42ab9 PO |
65 | is from the path in question, the lower its precedence). Finally |
66 | global and system-wide files are considered (they have the lowest | |
67 | precedence). | |
88e7fdf2 | 68 | |
90b22907 | 69 | If you wish to affect only a single repository (i.e., to assign |
6df42ab9 PO |
70 | attributes to files that are particular to |
71 | one user's workflow for that repository), then | |
90b22907 JK |
72 | attributes should be placed in the `$GIT_DIR/info/attributes` file. |
73 | Attributes which should be version-controlled and distributed to other | |
74 | repositories (i.e., attributes of interest to all users) should go into | |
6df42ab9 PO |
75 | `.gitattributes` files. Attributes that should affect all repositories |
76 | for a single user should be placed in a file specified by the | |
77 | `core.attributesfile` configuration option (see linkgit:git-config[1]). | |
78 | Attributes for all users on a system should be placed in the | |
79 | `$(prefix)/etc/gitattributes` file. | |
90b22907 | 80 | |
88e7fdf2 JH |
81 | Sometimes you would need to override an setting of an attribute |
82 | for a path to `unspecified` state. This can be done by listing | |
83 | the name of the attribute prefixed with an exclamation point `!`. | |
84 | ||
85 | ||
86 | EFFECTS | |
87 | ------- | |
88 | ||
89 | Certain operations by git can be influenced by assigning | |
ae7aa499 JH |
90 | particular attributes to a path. Currently, the following |
91 | operations are attributes-aware. | |
88e7fdf2 JH |
92 | |
93 | Checking-out and checking-in | |
94 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
95 | ||
3fed15f5 | 96 | These attributes affect how the contents stored in the |
88e7fdf2 | 97 | repository are copied to the working tree files when commands |
0b444cdb | 98 | such as 'git checkout' and 'git merge' run. They also affect how |
88e7fdf2 | 99 | git stores the contents you prepare in the working tree in the |
0b444cdb | 100 | repository upon 'git add' and 'git commit'. |
88e7fdf2 | 101 | |
5ec3e670 | 102 | `text` |
3fed15f5 JH |
103 | ^^^^^^ |
104 | ||
fd6cce9e EB |
105 | This attribute enables and controls end-of-line normalization. When a |
106 | text file is normalized, its line endings are converted to LF in the | |
107 | repository. To control what line ending style is used in the working | |
108 | directory, use the `eol` attribute for a single file and the | |
942e7747 | 109 | `core.eol` configuration variable for all text files. |
3fed15f5 | 110 | |
88e7fdf2 JH |
111 | Set:: |
112 | ||
5ec3e670 | 113 | Setting the `text` attribute on a path enables end-of-line |
fd6cce9e EB |
114 | normalization and marks the path as a text file. End-of-line |
115 | conversion takes place without guessing the content type. | |
88e7fdf2 JH |
116 | |
117 | Unset:: | |
118 | ||
5ec3e670 | 119 | Unsetting the `text` attribute on a path tells git not to |
bbb896d8 | 120 | attempt any end-of-line conversion upon checkin or checkout. |
88e7fdf2 | 121 | |
fd6cce9e | 122 | Set to string value "auto":: |
88e7fdf2 | 123 | |
5ec3e670 | 124 | When `text` is set to "auto", the path is marked for automatic |
fd6cce9e EB |
125 | end-of-line normalization. If git decides that the content is |
126 | text, its line endings are normalized to LF on checkin. | |
88e7fdf2 | 127 | |
88e7fdf2 JH |
128 | Unspecified:: |
129 | ||
942e7747 EB |
130 | If the `text` attribute is unspecified, git uses the |
131 | `core.autocrlf` configuration variable to determine if the | |
132 | file should be converted. | |
88e7fdf2 | 133 | |
5ec3e670 | 134 | Any other value causes git to act as if `text` has been left |
fd6cce9e | 135 | unspecified. |
88e7fdf2 | 136 | |
fd6cce9e EB |
137 | `eol` |
138 | ^^^^^ | |
88e7fdf2 | 139 | |
fd6cce9e EB |
140 | This attribute sets a specific line-ending style to be used in the |
141 | working directory. It enables end-of-line normalization without any | |
942e7747 | 142 | content checks, effectively setting the `text` attribute. |
88e7fdf2 | 143 | |
fd6cce9e | 144 | Set to string value "crlf":: |
88e7fdf2 | 145 | |
942e7747 EB |
146 | This setting forces git to normalize line endings for this |
147 | file on checkin and convert them to CRLF when the file is | |
148 | checked out. | |
fd6cce9e EB |
149 | |
150 | Set to string value "lf":: | |
151 | ||
152 | This setting forces git to normalize line endings to LF on | |
153 | checkin and prevents conversion to CRLF when the file is | |
942e7747 | 154 | checked out. |
5ec3e670 EB |
155 | |
156 | Backwards compatibility with `crlf` attribute | |
157 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
158 | ||
159 | For backwards compatibility, the `crlf` attribute is interpreted as | |
160 | follows: | |
161 | ||
162 | ------------------------ | |
163 | crlf text | |
164 | -crlf -text | |
165 | crlf=input eol=lf | |
166 | ------------------------ | |
fd6cce9e EB |
167 | |
168 | End-of-line conversion | |
169 | ^^^^^^^^^^^^^^^^^^^^^^ | |
170 | ||
171 | While git normally leaves file contents alone, it can be configured to | |
172 | normalize line endings to LF in the repository and, optionally, to | |
173 | convert them to CRLF when files are checked out. | |
174 | ||
175 | Here is an example that will make git normalize .txt, .vcproj and .sh | |
176 | files, ensure that .vcproj files have CRLF and .sh files have LF in | |
177 | the working directory, and prevent .jpg files from being normalized | |
178 | regardless of their content. | |
179 | ||
180 | ------------------------ | |
5ec3e670 | 181 | *.txt text |
fd6cce9e EB |
182 | *.vcproj eol=crlf |
183 | *.sh eol=lf | |
5ec3e670 | 184 | *.jpg -text |
fd6cce9e EB |
185 | ------------------------ |
186 | ||
187 | Other source code management systems normalize all text files in their | |
188 | repositories, and there are two ways to enable similar automatic | |
189 | normalization in git. | |
190 | ||
191 | If you simply want to have CRLF line endings in your working directory | |
192 | regardless of the repository you are working with, you can set the | |
193 | config variable "core.autocrlf" without changing any attributes. | |
194 | ||
195 | ------------------------ | |
196 | [core] | |
197 | autocrlf = true | |
198 | ------------------------ | |
199 | ||
200 | This does not force normalization of all text files, but does ensure | |
201 | that text files that you introduce to the repository have their line | |
202 | endings normalized to LF when they are added, and that files that are | |
942e7747 | 203 | already normalized in the repository stay normalized. |
fd6cce9e EB |
204 | |
205 | If you want to interoperate with a source code management system that | |
206 | enforces end-of-line normalization, or you simply want all text files | |
5ec3e670 | 207 | in your repository to be normalized, you should instead set the `text` |
fd6cce9e | 208 | attribute to "auto" for _all_ files. |
88e7fdf2 | 209 | |
fd6cce9e | 210 | ------------------------ |
5ec3e670 | 211 | * text=auto |
fd6cce9e EB |
212 | ------------------------ |
213 | ||
214 | This ensures that all files that git considers to be text will have | |
942e7747 EB |
215 | normalized (LF) line endings in the repository. The `core.eol` |
216 | configuration variable controls which line endings git will use for | |
217 | normalized files in your working directory; the default is to use the | |
218 | native line ending for your platform, or CRLF if `core.autocrlf` is | |
219 | set. | |
fd6cce9e | 220 | |
5ec3e670 | 221 | NOTE: When `text=auto` normalization is enabled in an existing |
fd6cce9e EB |
222 | repository, any text files containing CRLFs should be normalized. If |
223 | they are not they will be normalized the next time someone tries to | |
224 | change them, causing unfortunate misattribution. From a clean working | |
225 | directory: | |
226 | ||
227 | ------------------------------------------------- | |
5ec3e670 | 228 | $ echo "* text=auto" >>.gitattributes |
fd6cce9e EB |
229 | $ rm .git/index # Remove the index to force git to |
230 | $ git reset # re-scan the working directory | |
231 | $ git status # Show files that will be normalized | |
232 | $ git add -u | |
233 | $ git add .gitattributes | |
234 | $ git commit -m "Introduce end-of-line normalization" | |
235 | ------------------------------------------------- | |
236 | ||
237 | If any files that should not be normalized show up in 'git status', | |
5ec3e670 | 238 | unset their `text` attribute before running 'git add -u'. |
fd6cce9e EB |
239 | |
240 | ------------------------ | |
5ec3e670 | 241 | manual.pdf -text |
fd6cce9e | 242 | ------------------------ |
88e7fdf2 | 243 | |
fd6cce9e EB |
244 | Conversely, text files that git does not detect can have normalization |
245 | enabled manually. | |
88e7fdf2 | 246 | |
fd6cce9e | 247 | ------------------------ |
5ec3e670 | 248 | weirdchars.txt text |
fd6cce9e | 249 | ------------------------ |
88e7fdf2 | 250 | |
21e5ad50 SP |
251 | If `core.safecrlf` is set to "true" or "warn", git verifies if |
252 | the conversion is reversible for the current setting of | |
253 | `core.autocrlf`. For "true", git rejects irreversible | |
254 | conversions; for "warn", git only prints a warning but accepts | |
255 | an irreversible conversion. The safety triggers to prevent such | |
256 | a conversion done to the files in the work tree, but there are a | |
257 | few exceptions. Even though... | |
258 | ||
0b444cdb | 259 | - 'git add' itself does not touch the files in the work tree, the |
21e5ad50 SP |
260 | next checkout would, so the safety triggers; |
261 | ||
0b444cdb | 262 | - 'git apply' to update a text file with a patch does touch the files |
21e5ad50 SP |
263 | in the work tree, but the operation is about text files and CRLF |
264 | conversion is about fixing the line ending inconsistencies, so the | |
265 | safety does not trigger; | |
266 | ||
0b444cdb TR |
267 | - 'git diff' itself does not touch the files in the work tree, it is |
268 | often run to inspect the changes you intend to next 'git add'. To | |
21e5ad50 SP |
269 | catch potential problems early, safety triggers. |
270 | ||
88e7fdf2 | 271 | |
3fed15f5 JH |
272 | `ident` |
273 | ^^^^^^^ | |
274 | ||
2c850f12 JK |
275 | When the attribute `ident` is set for a path, git replaces |
276 | `$Id$` in the blob object with `$Id:`, followed by the | |
3fed15f5 JH |
277 | 40-character hexadecimal blob object name, followed by a dollar |
278 | sign `$` upon checkout. Any byte sequence that begins with | |
af9b54bb AP |
279 | `$Id:` and ends with `$` in the worktree file is replaced |
280 | with `$Id$` upon check-in. | |
3fed15f5 JH |
281 | |
282 | ||
aa4ed402 JH |
283 | `filter` |
284 | ^^^^^^^^ | |
285 | ||
c05ef938 | 286 | A `filter` attribute can be set to a string value that names a |
aa4ed402 JH |
287 | filter driver specified in the configuration. |
288 | ||
c05ef938 | 289 | A filter driver consists of a `clean` command and a `smudge` |
aa4ed402 | 290 | command, either of which can be left unspecified. Upon |
c05ef938 WC |
291 | checkout, when the `smudge` command is specified, the command is |
292 | fed the blob object from its standard input, and its standard | |
293 | output is used to update the worktree file. Similarly, the | |
294 | `clean` command is used to convert the contents of worktree file | |
295 | upon checkin. | |
aa4ed402 | 296 | |
c05ef938 | 297 | A missing filter driver definition in the config is not an error |
aa4ed402 JH |
298 | but makes the filter a no-op passthru. |
299 | ||
300 | The content filtering is done to massage the content into a | |
301 | shape that is more convenient for the platform, filesystem, and | |
c05ef938 | 302 | the user to use. The key phrase here is "more convenient" and not |
4d84aff3 JS |
303 | "turning something unusable into usable". In other words, the |
304 | intent is that if someone unsets the filter driver definition, | |
305 | or does not have the appropriate filter program, the project | |
306 | should still be usable. | |
aa4ed402 | 307 | |
d79f5d17 NS |
308 | For example, in .gitattributes, you would assign the `filter` |
309 | attribute for paths. | |
310 | ||
311 | ------------------------ | |
312 | *.c filter=indent | |
313 | ------------------------ | |
314 | ||
315 | Then you would define a "filter.indent.clean" and "filter.indent.smudge" | |
316 | configuration in your .git/config to specify a pair of commands to | |
317 | modify the contents of C programs when the source files are checked | |
318 | in ("clean" is run) and checked out (no change is made because the | |
319 | command is "cat"). | |
320 | ||
321 | ------------------------ | |
322 | [filter "indent"] | |
323 | clean = indent | |
324 | smudge = cat | |
325 | ------------------------ | |
326 | ||
f217f0e8 EB |
327 | For best results, `clean` should not alter its output further if it is |
328 | run twice ("clean->clean" should be equivalent to "clean"), and | |
329 | multiple `smudge` commands should not alter `clean`'s output | |
330 | ("smudge->smudge->clean" should be equivalent to "clean"). See the | |
331 | section on merging below. | |
332 | ||
333 | The "indent" filter is well-behaved in this regard: it will not modify | |
334 | input that is already correctly indented. In this case, the lack of a | |
335 | smudge filter means that the clean filter _must_ accept its own output | |
336 | without modifying it. | |
337 | ||
a2b665de PW |
338 | Sequence "%f" on the filter command line is replaced with the name of |
339 | the file the filter is working on. A filter might use this in keyword | |
340 | substitution. For example: | |
341 | ||
342 | ------------------------ | |
343 | [filter "p4"] | |
344 | clean = git-p4-filter --clean %f | |
345 | smudge = git-p4-filter --smudge %f | |
346 | ------------------------ | |
347 | ||
aa4ed402 JH |
348 | |
349 | Interaction between checkin/checkout attributes | |
350 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
351 | ||
352 | In the check-in codepath, the worktree file is first converted | |
353 | with `filter` driver (if specified and corresponding driver | |
354 | defined), then the result is processed with `ident` (if | |
5ec3e670 | 355 | specified), and then finally with `text` (again, if specified |
aa4ed402 JH |
356 | and applicable). |
357 | ||
358 | In the check-out codepath, the blob content is first converted | |
5ec3e670 | 359 | with `text`, and then `ident` and fed to `filter`. |
aa4ed402 JH |
360 | |
361 | ||
f217f0e8 EB |
362 | Merging branches with differing checkin/checkout attributes |
363 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
364 | ||
365 | If you have added attributes to a file that cause the canonical | |
366 | repository format for that file to change, such as adding a | |
367 | clean/smudge filter or text/eol/ident attributes, merging anything | |
368 | where the attribute is not in place would normally cause merge | |
369 | conflicts. | |
370 | ||
371 | To prevent these unnecessary merge conflicts, git can be told to run a | |
372 | virtual check-out and check-in of all three stages of a file when | |
373 | resolving a three-way merge by setting the `merge.renormalize` | |
374 | configuration variable. This prevents changes caused by check-in | |
375 | conversion from causing spurious merge conflicts when a converted file | |
376 | is merged with an unconverted file. | |
377 | ||
378 | As long as a "smudge->clean" results in the same output as a "clean" | |
379 | even on files that are already smudged, this strategy will | |
380 | automatically resolve all filter-related conflicts. Filters that do | |
381 | not act in this way may cause additional merge conflicts that must be | |
382 | resolved manually. | |
383 | ||
384 | ||
88e7fdf2 JH |
385 | Generating diff text |
386 | ~~~~~~~~~~~~~~~~~~~~ | |
387 | ||
4f73e240 JN |
388 | `diff` |
389 | ^^^^^^ | |
390 | ||
678852d9 JK |
391 | The attribute `diff` affects how 'git' generates diffs for particular |
392 | files. It can tell git whether to generate a textual patch for the path | |
393 | or to treat the path as a binary file. It can also affect what line is | |
394 | shown on the hunk header `@@ -k,l +n,m @@` line, tell git to use an | |
395 | external command to generate the diff, or ask git to convert binary | |
396 | files to a text format before generating the diff. | |
88e7fdf2 JH |
397 | |
398 | Set:: | |
399 | ||
400 | A path to which the `diff` attribute is set is treated | |
401 | as text, even when they contain byte values that | |
402 | normally never appear in text files, such as NUL. | |
403 | ||
404 | Unset:: | |
405 | ||
406 | A path to which the `diff` attribute is unset will | |
678852d9 JK |
407 | generate `Binary files differ` (or a binary patch, if |
408 | binary patches are enabled). | |
88e7fdf2 JH |
409 | |
410 | Unspecified:: | |
411 | ||
412 | A path to which the `diff` attribute is unspecified | |
413 | first gets its contents inspected, and if it looks like | |
414 | text, it is treated as text. Otherwise it would | |
415 | generate `Binary files differ`. | |
416 | ||
2cc3167c JH |
417 | String:: |
418 | ||
678852d9 JK |
419 | Diff is shown using the specified diff driver. Each driver may |
420 | specify one or more options, as described in the following | |
421 | section. The options for the diff driver "foo" are defined | |
422 | by the configuration variables in the "diff.foo" section of the | |
423 | git config file. | |
2cc3167c JH |
424 | |
425 | ||
678852d9 JK |
426 | Defining an external diff driver |
427 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
2cc3167c JH |
428 | |
429 | The definition of a diff driver is done in `gitconfig`, not | |
430 | `gitattributes` file, so strictly speaking this manual page is a | |
431 | wrong place to talk about it. However... | |
432 | ||
678852d9 | 433 | To define an external diff driver `jcdiff`, add a section to your |
2cc3167c JH |
434 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: |
435 | ||
436 | ---------------------------------------------------------------- | |
437 | [diff "jcdiff"] | |
438 | command = j-c-diff | |
439 | ---------------------------------------------------------------- | |
440 | ||
441 | When git needs to show you a diff for the path with `diff` | |
442 | attribute set to `jcdiff`, it calls the command you specified | |
443 | with the above configuration, i.e. `j-c-diff`, with 7 | |
444 | parameters, just like `GIT_EXTERNAL_DIFF` program is called. | |
9e1f0a85 | 445 | See linkgit:git[1] for details. |
88e7fdf2 JH |
446 | |
447 | ||
ae7aa499 JH |
448 | Defining a custom hunk-header |
449 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
450 | ||
c882c01e | 451 | Each group of changes (called a "hunk") in the textual diff output |
ae7aa499 JH |
452 | is prefixed with a line of the form: |
453 | ||
454 | @@ -k,l +n,m @@ TEXT | |
455 | ||
c882c01e GD |
456 | This is called a 'hunk header'. The "TEXT" portion is by default a line |
457 | that begins with an alphabet, an underscore or a dollar sign; this | |
458 | matches what GNU 'diff -p' output uses. This default selection however | |
459 | is not suited for some contents, and you can use a customized pattern | |
460 | to make a selection. | |
ae7aa499 | 461 | |
c882c01e | 462 | First, in .gitattributes, you would assign the `diff` attribute |
ae7aa499 JH |
463 | for paths. |
464 | ||
465 | ------------------------ | |
466 | *.tex diff=tex | |
467 | ------------------------ | |
468 | ||
edb7e82f | 469 | Then, you would define a "diff.tex.xfuncname" configuration to |
ae7aa499 | 470 | specify a regular expression that matches a line that you would |
c4c86d23 JK |
471 | want to appear as the hunk header "TEXT". Add a section to your |
472 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
ae7aa499 JH |
473 | |
474 | ------------------------ | |
475 | [diff "tex"] | |
45d9414f | 476 | xfuncname = "^(\\\\(sub)*section\\{.*)$" |
ae7aa499 JH |
477 | ------------------------ |
478 | ||
479 | Note. A single level of backslashes are eaten by the | |
480 | configuration file parser, so you would need to double the | |
481 | backslashes; the pattern above picks a line that begins with a | |
02783075 | 482 | backslash, and zero or more occurrences of `sub` followed by |
ae7aa499 JH |
483 | `section` followed by open brace, to the end of line. |
484 | ||
485 | There are a few built-in patterns to make this easier, and `tex` | |
486 | is one of them, so you do not have to write the above in your | |
487 | configuration file (you still need to enable this with the | |
d08ed6d6 GH |
488 | attribute mechanism, via `.gitattributes`). The following built in |
489 | patterns are available: | |
490 | ||
23b5beb2 GH |
491 | - `bibtex` suitable for files with BibTeX coded references. |
492 | ||
80c49c3d TR |
493 | - `cpp` suitable for source code in the C and C++ languages. |
494 | ||
b221207d PO |
495 | - `csharp` suitable for source code in the C# language. |
496 | ||
909a5494 BC |
497 | - `fortran` suitable for source code in the Fortran language. |
498 | ||
af9ce1ff AE |
499 | - `html` suitable for HTML/XHTML documents. |
500 | ||
b66e00f1 | 501 | - `java` suitable for source code in the Java language. |
d08ed6d6 | 502 | |
5d1e958e JS |
503 | - `objc` suitable for source code in the Objective-C language. |
504 | ||
d08ed6d6 GH |
505 | - `pascal` suitable for source code in the Pascal/Delphi language. |
506 | ||
71a5d4bc JN |
507 | - `perl` suitable for source code in the Perl language. |
508 | ||
af9ce1ff AE |
509 | - `php` suitable for source code in the PHP language. |
510 | ||
7c17205b KS |
511 | - `python` suitable for source code in the Python language. |
512 | ||
d08ed6d6 GH |
513 | - `ruby` suitable for source code in the Ruby language. |
514 | ||
515 | - `tex` suitable for source code for LaTeX documents. | |
ae7aa499 JH |
516 | |
517 | ||
80c49c3d TR |
518 | Customizing word diff |
519 | ^^^^^^^^^^^^^^^^^^^^^ | |
520 | ||
882749a0 | 521 | You can customize the rules that `git diff --word-diff` uses to |
80c49c3d | 522 | split words in a line, by specifying an appropriate regular expression |
ae3b970a | 523 | in the "diff.*.wordRegex" configuration variable. For example, in TeX |
80c49c3d TR |
524 | a backslash followed by a sequence of letters forms a command, but |
525 | several such commands can be run together without intervening | |
c4c86d23 JK |
526 | whitespace. To separate them, use a regular expression in your |
527 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
80c49c3d TR |
528 | |
529 | ------------------------ | |
530 | [diff "tex"] | |
ae3b970a | 531 | wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+" |
80c49c3d TR |
532 | ------------------------ |
533 | ||
534 | A built-in pattern is provided for all languages listed in the | |
535 | previous section. | |
536 | ||
537 | ||
678852d9 JK |
538 | Performing text diffs of binary files |
539 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
540 | ||
541 | Sometimes it is desirable to see the diff of a text-converted | |
542 | version of some binary files. For example, a word processor | |
543 | document can be converted to an ASCII text representation, and | |
544 | the diff of the text shown. Even though this conversion loses | |
545 | some information, the resulting diff is useful for human | |
546 | viewing (but cannot be applied directly). | |
547 | ||
548 | The `textconv` config option is used to define a program for | |
549 | performing such a conversion. The program should take a single | |
550 | argument, the name of a file to convert, and produce the | |
551 | resulting text on stdout. | |
552 | ||
553 | For example, to show the diff of the exif information of a | |
554 | file instead of the binary information (assuming you have the | |
c4c86d23 JK |
555 | exif tool installed), add the following section to your |
556 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file): | |
678852d9 JK |
557 | |
558 | ------------------------ | |
559 | [diff "jpg"] | |
560 | textconv = exif | |
561 | ------------------------ | |
562 | ||
563 | NOTE: The text conversion is generally a one-way conversion; | |
564 | in this example, we lose the actual image contents and focus | |
565 | just on the text data. This means that diffs generated by | |
566 | textconv are _not_ suitable for applying. For this reason, | |
567 | only `git diff` and the `git log` family of commands (i.e., | |
568 | log, whatchanged, show) will perform text conversion. `git | |
569 | format-patch` will never generate this output. If you want to | |
570 | send somebody a text-converted diff of a binary file (e.g., | |
571 | because it quickly conveys the changes you have made), you | |
572 | should generate it separately and send it as a comment _in | |
573 | addition to_ the usual binary diff that you might send. | |
574 | ||
d9bae1a1 JK |
575 | Because text conversion can be slow, especially when doing a |
576 | large number of them with `git log -p`, git provides a mechanism | |
577 | to cache the output and use it in future diffs. To enable | |
578 | caching, set the "cachetextconv" variable in your diff driver's | |
579 | config. For example: | |
580 | ||
581 | ------------------------ | |
582 | [diff "jpg"] | |
583 | textconv = exif | |
584 | cachetextconv = true | |
585 | ------------------------ | |
586 | ||
587 | This will cache the result of running "exif" on each blob | |
588 | indefinitely. If you change the textconv config variable for a | |
589 | diff driver, git will automatically invalidate the cache entries | |
590 | and re-run the textconv filter. If you want to invalidate the | |
591 | cache manually (e.g., because your version of "exif" was updated | |
592 | and now produces better output), you can remove the cache | |
593 | manually with `git update-ref -d refs/notes/textconv/jpg` (where | |
594 | "jpg" is the name of the diff driver, as in the example above). | |
678852d9 | 595 | |
ab435611 JK |
596 | Marking files as binary |
597 | ^^^^^^^^^^^^^^^^^^^^^^^ | |
598 | ||
599 | Git usually guesses correctly whether a blob contains text or binary | |
600 | data by examining the beginning of the contents. However, sometimes you | |
601 | may want to override its decision, either because a blob contains binary | |
602 | data later in the file, or because the content, while technically | |
603 | composed of text characters, is opaque to a human reader. For example, | |
604 | many postscript files contain only ascii characters, but produce noisy | |
605 | and meaningless diffs. | |
606 | ||
607 | The simplest way to mark a file as binary is to unset the diff | |
608 | attribute in the `.gitattributes` file: | |
609 | ||
610 | ------------------------ | |
611 | *.ps -diff | |
612 | ------------------------ | |
613 | ||
614 | This will cause git to generate `Binary files differ` (or a binary | |
615 | patch, if binary patches are enabled) instead of a regular diff. | |
616 | ||
617 | However, one may also want to specify other diff driver attributes. For | |
618 | example, you might want to use `textconv` to convert postscript files to | |
619 | an ascii representation for human viewing, but otherwise treat them as | |
620 | binary files. You cannot specify both `-diff` and `diff=ps` attributes. | |
621 | The solution is to use the `diff.*.binary` config option: | |
622 | ||
623 | ------------------------ | |
624 | [diff "ps"] | |
625 | textconv = ps2ascii | |
626 | binary = true | |
627 | ------------------------ | |
628 | ||
88e7fdf2 JH |
629 | Performing a three-way merge |
630 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
631 | ||
4f73e240 JN |
632 | `merge` |
633 | ^^^^^^^ | |
634 | ||
88e7fdf2 JH |
635 | The attribute `merge` affects how three versions of a file is |
636 | merged when a file-level merge is necessary during `git merge`, | |
57f6ec02 | 637 | and other commands such as `git revert` and `git cherry-pick`. |
88e7fdf2 JH |
638 | |
639 | Set:: | |
640 | ||
641 | Built-in 3-way merge driver is used to merge the | |
2fd02c92 | 642 | contents in a way similar to 'merge' command of `RCS` |
88e7fdf2 JH |
643 | suite. This is suitable for ordinary text files. |
644 | ||
645 | Unset:: | |
646 | ||
647 | Take the version from the current branch as the | |
648 | tentative merge result, and declare that the merge has | |
649 | conflicts. This is suitable for binary files that does | |
650 | not have a well-defined merge semantics. | |
651 | ||
652 | Unspecified:: | |
653 | ||
654 | By default, this uses the same built-in 3-way merge | |
655 | driver as is the case the `merge` attribute is set. | |
656 | However, `merge.default` configuration variable can name | |
657 | different merge driver to be used for paths to which the | |
658 | `merge` attribute is unspecified. | |
659 | ||
2cc3167c | 660 | String:: |
88e7fdf2 JH |
661 | |
662 | 3-way merge is performed using the specified custom | |
663 | merge driver. The built-in 3-way merge driver can be | |
664 | explicitly specified by asking for "text" driver; the | |
665 | built-in "take the current branch" driver can be | |
b9d14ffb | 666 | requested with "binary". |
88e7fdf2 JH |
667 | |
668 | ||
0e545f75 JH |
669 | Built-in merge drivers |
670 | ^^^^^^^^^^^^^^^^^^^^^^ | |
671 | ||
672 | There are a few built-in low-level merge drivers defined that | |
673 | can be asked for via the `merge` attribute. | |
674 | ||
675 | text:: | |
676 | ||
677 | Usual 3-way file level merge for text files. Conflicted | |
678 | regions are marked with conflict markers `<<<<<<<`, | |
679 | `=======` and `>>>>>>>`. The version from your branch | |
680 | appears before the `=======` marker, and the version | |
681 | from the merged branch appears after the `=======` | |
682 | marker. | |
683 | ||
684 | binary:: | |
685 | ||
686 | Keep the version from your branch in the work tree, but | |
687 | leave the path in the conflicted state for the user to | |
688 | sort out. | |
689 | ||
690 | union:: | |
691 | ||
692 | Run 3-way file level merge for text files, but take | |
693 | lines from both versions, instead of leaving conflict | |
694 | markers. This tends to leave the added lines in the | |
695 | resulting file in random order and the user should | |
696 | verify the result. Do not use this if you do not | |
697 | understand the implications. | |
698 | ||
699 | ||
88e7fdf2 JH |
700 | Defining a custom merge driver |
701 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
702 | ||
0e545f75 JH |
703 | The definition of a merge driver is done in the `.git/config` |
704 | file, not in the `gitattributes` file, so strictly speaking this | |
705 | manual page is a wrong place to talk about it. However... | |
88e7fdf2 JH |
706 | |
707 | To define a custom merge driver `filfre`, add a section to your | |
708 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
709 | ||
710 | ---------------------------------------------------------------- | |
711 | [merge "filfre"] | |
712 | name = feel-free merge driver | |
713 | driver = filfre %O %A %B | |
714 | recursive = binary | |
715 | ---------------------------------------------------------------- | |
716 | ||
717 | The `merge.*.name` variable gives the driver a human-readable | |
718 | name. | |
719 | ||
720 | The `merge.*.driver` variable's value is used to construct a | |
721 | command to run to merge ancestor's version (`%O`), current | |
722 | version (`%A`) and the other branches' version (`%B`). These | |
723 | three tokens are replaced with the names of temporary files that | |
724 | hold the contents of these versions when the command line is | |
16758621 BW |
725 | built. Additionally, %L will be replaced with the conflict marker |
726 | size (see below). | |
88e7fdf2 JH |
727 | |
728 | The merge driver is expected to leave the result of the merge in | |
729 | the file named with `%A` by overwriting it, and exit with zero | |
730 | status if it managed to merge them cleanly, or non-zero if there | |
731 | were conflicts. | |
732 | ||
733 | The `merge.*.recursive` variable specifies what other merge | |
734 | driver to use when the merge driver is called for an internal | |
735 | merge between common ancestors, when there are more than one. | |
736 | When left unspecified, the driver itself is used for both | |
737 | internal merge and the final merge. | |
738 | ||
739 | ||
4c734803 JH |
740 | `conflict-marker-size` |
741 | ^^^^^^^^^^^^^^^^^^^^^^ | |
742 | ||
743 | This attribute controls the length of conflict markers left in | |
744 | the work tree file during a conflicted merge. Only setting to | |
745 | the value to a positive integer has any meaningful effect. | |
746 | ||
747 | For example, this line in `.gitattributes` can be used to tell the merge | |
748 | machinery to leave much longer (instead of the usual 7-character-long) | |
749 | conflict markers when merging the file `Documentation/git-merge.txt` | |
750 | results in a conflict. | |
751 | ||
752 | ------------------------ | |
753 | Documentation/git-merge.txt conflict-marker-size=32 | |
754 | ------------------------ | |
755 | ||
756 | ||
cf1b7869 JH |
757 | Checking whitespace errors |
758 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
759 | ||
760 | `whitespace` | |
761 | ^^^^^^^^^^^^ | |
762 | ||
763 | The `core.whitespace` configuration variable allows you to define what | |
2fd02c92 | 764 | 'diff' and 'apply' should consider whitespace errors for all paths in |
5162e697 | 765 | the project (See linkgit:git-config[1]). This attribute gives you finer |
cf1b7869 JH |
766 | control per path. |
767 | ||
768 | Set:: | |
769 | ||
770 | Notice all types of potential whitespace errors known to git. | |
f4b05a49 JS |
771 | The tab width is taken from the value of the `core.whitespace` |
772 | configuration variable. | |
cf1b7869 JH |
773 | |
774 | Unset:: | |
775 | ||
776 | Do not notice anything as error. | |
777 | ||
778 | Unspecified:: | |
779 | ||
f4b05a49 | 780 | Use the value of the `core.whitespace` configuration variable to |
cf1b7869 JH |
781 | decide what to notice as error. |
782 | ||
783 | String:: | |
784 | ||
785 | Specify a comma separate list of common whitespace problems to | |
f4b05a49 | 786 | notice in the same format as the `core.whitespace` configuration |
cf1b7869 JH |
787 | variable. |
788 | ||
789 | ||
8a33dd8b JH |
790 | Creating an archive |
791 | ~~~~~~~~~~~~~~~~~~~ | |
792 | ||
08b51f51 JH |
793 | `export-ignore` |
794 | ^^^^^^^^^^^^^^^ | |
795 | ||
796 | Files and directories with the attribute `export-ignore` won't be added to | |
797 | archive files. | |
798 | ||
8a33dd8b JH |
799 | `export-subst` |
800 | ^^^^^^^^^^^^^^ | |
801 | ||
802 | If the attribute `export-subst` is set for a file then git will expand | |
803 | several placeholders when adding this file to an archive. The | |
08b51f51 | 804 | expansion depends on the availability of a commit ID, i.e., if |
8a33dd8b JH |
805 | linkgit:git-archive[1] has been given a tree instead of a commit or a |
806 | tag then no replacement will be done. The placeholders are the same | |
807 | as those for the option `--pretty=format:` of linkgit:git-log[1], | |
808 | except that they need to be wrapped like this: `$Format:PLACEHOLDERS$` | |
809 | in the file. E.g. the string `$Format:%H$` will be replaced by the | |
810 | commit hash. | |
811 | ||
812 | ||
975457f1 NG |
813 | Packing objects |
814 | ~~~~~~~~~~~~~~~ | |
815 | ||
816 | `delta` | |
817 | ^^^^^^^ | |
818 | ||
819 | Delta compression will not be attempted for blobs for paths with the | |
820 | attribute `delta` set to false. | |
821 | ||
822 | ||
a2df1fb2 AG |
823 | Viewing files in GUI tools |
824 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
825 | ||
826 | `encoding` | |
827 | ^^^^^^^^^^ | |
828 | ||
829 | The value of this attribute specifies the character encoding that should | |
830 | be used by GUI tools (e.g. linkgit:gitk[1] and linkgit:git-gui[1]) to | |
831 | display the contents of the relevant file. Note that due to performance | |
832 | considerations linkgit:gitk[1] does not use this attribute unless you | |
833 | manually enable per-file encodings in its options. | |
834 | ||
835 | If this attribute is not set or has an invalid value, the value of the | |
836 | `gui.encoding` configuration variable is used instead | |
837 | (See linkgit:git-config[1]). | |
838 | ||
839 | ||
bbb896d8 JH |
840 | USING ATTRIBUTE MACROS |
841 | ---------------------- | |
842 | ||
843 | You do not want any end-of-line conversions applied to, nor textual diffs | |
844 | produced for, any binary file you track. You would need to specify e.g. | |
845 | ||
846 | ------------ | |
5ec3e670 | 847 | *.jpg -text -diff |
bbb896d8 JH |
848 | ------------ |
849 | ||
850 | but that may become cumbersome, when you have many attributes. Using | |
851 | attribute macros, you can specify groups of attributes set or unset at | |
852 | the same time. The system knows a built-in attribute macro, `binary`: | |
853 | ||
854 | ------------ | |
855 | *.jpg binary | |
856 | ------------ | |
857 | ||
858 | which is equivalent to the above. Note that the attribute macros can only | |
859 | be "Set" (see the above example that sets "binary" macro as if it were an | |
5ec3e670 | 860 | ordinary attribute --- setting it in turn unsets "text" and "diff"). |
bbb896d8 JH |
861 | |
862 | ||
863 | DEFINING ATTRIBUTE MACROS | |
864 | ------------------------- | |
865 | ||
866 | Custom attribute macros can be defined only in the `.gitattributes` file | |
867 | at the toplevel (i.e. not in any subdirectory). The built-in attribute | |
868 | macro "binary" is equivalent to: | |
869 | ||
870 | ------------ | |
5ec3e670 | 871 | [attr]binary -diff -text |
bbb896d8 JH |
872 | ------------ |
873 | ||
874 | ||
88e7fdf2 JH |
875 | EXAMPLE |
876 | ------- | |
877 | ||
878 | If you have these three `gitattributes` file: | |
879 | ||
880 | ---------------------------------------------------------------- | |
881 | (in $GIT_DIR/info/attributes) | |
882 | ||
883 | a* foo !bar -baz | |
884 | ||
885 | (in .gitattributes) | |
886 | abc foo bar baz | |
887 | ||
888 | (in t/.gitattributes) | |
889 | ab* merge=filfre | |
890 | abc -foo -bar | |
891 | *.c frotz | |
892 | ---------------------------------------------------------------- | |
893 | ||
894 | the attributes given to path `t/abc` are computed as follows: | |
895 | ||
896 | 1. By examining `t/.gitattributes` (which is in the same | |
02783075 | 897 | directory as the path in question), git finds that the first |
88e7fdf2 JH |
898 | line matches. `merge` attribute is set. It also finds that |
899 | the second line matches, and attributes `foo` and `bar` | |
900 | are unset. | |
901 | ||
902 | 2. Then it examines `.gitattributes` (which is in the parent | |
903 | directory), and finds that the first line matches, but | |
904 | `t/.gitattributes` file already decided how `merge`, `foo` | |
905 | and `bar` attributes should be given to this path, so it | |
906 | leaves `foo` and `bar` unset. Attribute `baz` is set. | |
907 | ||
5c759f96 | 908 | 3. Finally it examines `$GIT_DIR/info/attributes`. This file |
88e7fdf2 JH |
909 | is used to override the in-tree settings. The first line is |
910 | a match, and `foo` is set, `bar` is reverted to unspecified | |
911 | state, and `baz` is unset. | |
912 | ||
02783075 | 913 | As the result, the attributes assignment to `t/abc` becomes: |
88e7fdf2 JH |
914 | |
915 | ---------------------------------------------------------------- | |
916 | foo set to true | |
917 | bar unspecified | |
918 | baz set to false | |
919 | merge set to string value "filfre" | |
920 | frotz unspecified | |
921 | ---------------------------------------------------------------- | |
922 | ||
923 | ||
8460b2fc | 924 | |
88e7fdf2 JH |
925 | GIT |
926 | --- | |
9e1f0a85 | 927 | Part of the linkgit:git[1] suite |