]>
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 | 23 | That is, a pattern followed by an attributes list, |
860a74d9 NTND |
24 | separated by whitespaces. Leading and trailing whitespaces are |
25 | ignored. Lines that begin with '#' are ignored. Patterns | |
26 | that begin with a double quote are quoted in C style. | |
27 | When the pattern matches the path in question, the attributes | |
28 | listed on the line are given to the path. | |
88e7fdf2 JH |
29 | |
30 | Each attribute can be in one of these states for a given path: | |
31 | ||
32 | Set:: | |
33 | ||
34 | The path has the attribute with special value "true"; | |
35 | this is specified by listing only the name of the | |
36 | attribute in the attribute list. | |
37 | ||
38 | Unset:: | |
39 | ||
40 | The path has the attribute with special value "false"; | |
41 | this is specified by listing the name of the attribute | |
42 | prefixed with a dash `-` in the attribute list. | |
43 | ||
44 | Set to a value:: | |
45 | ||
46 | The path has the attribute with specified string value; | |
47 | this is specified by listing the name of the attribute | |
48 | followed by an equal sign `=` and its value in the | |
49 | attribute list. | |
50 | ||
51 | Unspecified:: | |
52 | ||
3f74c8e8 | 53 | No pattern matches the path, and nothing says if |
b9d14ffb JH |
54 | the path has or does not have the attribute, the |
55 | attribute for the path is said to be Unspecified. | |
88e7fdf2 | 56 | |
3f74c8e8 | 57 | When more than one pattern matches the path, a later line |
b9d14ffb | 58 | overrides an earlier line. This overriding is done per |
3f74c8e8 JS |
59 | attribute. The rules how the pattern matches paths are the |
60 | same as in `.gitignore` files; see linkgit:gitignore[5]. | |
82dce998 | 61 | Unlike `.gitignore`, negative patterns are forbidden. |
88e7fdf2 | 62 | |
2de9b711 | 63 | When deciding what attributes are assigned to a path, Git |
88e7fdf2 JH |
64 | consults `$GIT_DIR/info/attributes` file (which has the highest |
65 | precedence), `.gitattributes` file in the same directory as the | |
20ff3ec2 JM |
66 | path in question, and its parent directories up to the toplevel of the |
67 | work tree (the further the directory that contains `.gitattributes` | |
6df42ab9 PO |
68 | is from the path in question, the lower its precedence). Finally |
69 | global and system-wide files are considered (they have the lowest | |
70 | precedence). | |
88e7fdf2 | 71 | |
40701adb NTND |
72 | When the `.gitattributes` file is missing from the work tree, the |
73 | path in the index is used as a fall-back. During checkout process, | |
74 | `.gitattributes` in the index is used and then the file in the | |
75 | working tree is used as a fall-back. | |
76 | ||
90b22907 | 77 | If you wish to affect only a single repository (i.e., to assign |
6df42ab9 PO |
78 | attributes to files that are particular to |
79 | one user's workflow for that repository), then | |
90b22907 JK |
80 | attributes should be placed in the `$GIT_DIR/info/attributes` file. |
81 | Attributes which should be version-controlled and distributed to other | |
82 | repositories (i.e., attributes of interest to all users) should go into | |
6df42ab9 PO |
83 | `.gitattributes` files. Attributes that should affect all repositories |
84 | for a single user should be placed in a file specified by the | |
da0005b8 | 85 | `core.attributesFile` configuration option (see linkgit:git-config[1]). |
684e40f6 HKNN |
86 | Its default value is $XDG_CONFIG_HOME/git/attributes. If $XDG_CONFIG_HOME |
87 | is either not set or empty, $HOME/.config/git/attributes is used instead. | |
6df42ab9 PO |
88 | Attributes for all users on a system should be placed in the |
89 | `$(prefix)/etc/gitattributes` file. | |
90b22907 | 90 | |
faa4e8ce | 91 | Sometimes you would need to override a setting of an attribute |
0922570c | 92 | for a path to `Unspecified` state. This can be done by listing |
88e7fdf2 JH |
93 | the name of the attribute prefixed with an exclamation point `!`. |
94 | ||
95 | ||
96 | EFFECTS | |
97 | ------- | |
98 | ||
2de9b711 | 99 | Certain operations by Git can be influenced by assigning |
ae7aa499 JH |
100 | particular attributes to a path. Currently, the following |
101 | operations are attributes-aware. | |
88e7fdf2 JH |
102 | |
103 | Checking-out and checking-in | |
104 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
105 | ||
3fed15f5 | 106 | These attributes affect how the contents stored in the |
88e7fdf2 | 107 | repository are copied to the working tree files when commands |
0b444cdb | 108 | such as 'git checkout' and 'git merge' run. They also affect how |
2de9b711 | 109 | Git stores the contents you prepare in the working tree in the |
0b444cdb | 110 | repository upon 'git add' and 'git commit'. |
88e7fdf2 | 111 | |
5ec3e670 | 112 | `text` |
3fed15f5 JH |
113 | ^^^^^^ |
114 | ||
fd6cce9e EB |
115 | This attribute enables and controls end-of-line normalization. When a |
116 | text file is normalized, its line endings are converted to LF in the | |
117 | repository. To control what line ending style is used in the working | |
118 | directory, use the `eol` attribute for a single file and the | |
942e7747 | 119 | `core.eol` configuration variable for all text files. |
65237284 | 120 | Note that `core.autocrlf` overrides `core.eol` |
3fed15f5 | 121 | |
88e7fdf2 JH |
122 | Set:: |
123 | ||
5ec3e670 | 124 | Setting the `text` attribute on a path enables end-of-line |
fd6cce9e EB |
125 | normalization and marks the path as a text file. End-of-line |
126 | conversion takes place without guessing the content type. | |
88e7fdf2 JH |
127 | |
128 | Unset:: | |
129 | ||
2de9b711 | 130 | Unsetting the `text` attribute on a path tells Git not to |
bbb896d8 | 131 | attempt any end-of-line conversion upon checkin or checkout. |
88e7fdf2 | 132 | |
fd6cce9e | 133 | Set to string value "auto":: |
88e7fdf2 | 134 | |
5ec3e670 | 135 | When `text` is set to "auto", the path is marked for automatic |
65237284 TB |
136 | end-of-line conversion. If Git decides that the content is |
137 | text, its line endings are converted to LF on checkin. | |
2e3a16b2 | 138 | When the file has been committed with CRLF, no conversion is done. |
88e7fdf2 | 139 | |
88e7fdf2 JH |
140 | Unspecified:: |
141 | ||
2de9b711 | 142 | If the `text` attribute is unspecified, Git uses the |
942e7747 EB |
143 | `core.autocrlf` configuration variable to determine if the |
144 | file should be converted. | |
88e7fdf2 | 145 | |
2de9b711 | 146 | Any other value causes Git to act as if `text` has been left |
fd6cce9e | 147 | unspecified. |
88e7fdf2 | 148 | |
fd6cce9e EB |
149 | `eol` |
150 | ^^^^^ | |
88e7fdf2 | 151 | |
fd6cce9e | 152 | This attribute sets a specific line-ending style to be used in the |
65237284 | 153 | working directory. It enables end-of-line conversion without any |
3bc4b8f7 BB |
154 | content checks, effectively setting the `text` attribute. Note that |
155 | setting this attribute on paths which are in the index with CRLF line | |
156 | endings may make the paths to be considered dirty. Adding the path to | |
157 | the index again will normalize the line endings in the index. | |
88e7fdf2 | 158 | |
fd6cce9e | 159 | Set to string value "crlf":: |
88e7fdf2 | 160 | |
2de9b711 | 161 | This setting forces Git to normalize line endings for this |
942e7747 EB |
162 | file on checkin and convert them to CRLF when the file is |
163 | checked out. | |
fd6cce9e EB |
164 | |
165 | Set to string value "lf":: | |
166 | ||
2de9b711 | 167 | This setting forces Git to normalize line endings to LF on |
fd6cce9e | 168 | checkin and prevents conversion to CRLF when the file is |
942e7747 | 169 | checked out. |
5ec3e670 EB |
170 | |
171 | Backwards compatibility with `crlf` attribute | |
172 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
173 | ||
174 | For backwards compatibility, the `crlf` attribute is interpreted as | |
175 | follows: | |
176 | ||
177 | ------------------------ | |
178 | crlf text | |
179 | -crlf -text | |
180 | crlf=input eol=lf | |
181 | ------------------------ | |
fd6cce9e EB |
182 | |
183 | End-of-line conversion | |
184 | ^^^^^^^^^^^^^^^^^^^^^^ | |
185 | ||
2de9b711 | 186 | While Git normally leaves file contents alone, it can be configured to |
fd6cce9e EB |
187 | normalize line endings to LF in the repository and, optionally, to |
188 | convert them to CRLF when files are checked out. | |
189 | ||
fd6cce9e EB |
190 | If you simply want to have CRLF line endings in your working directory |
191 | regardless of the repository you are working with, you can set the | |
65237284 | 192 | config variable "core.autocrlf" without using any attributes. |
fd6cce9e EB |
193 | |
194 | ------------------------ | |
195 | [core] | |
196 | autocrlf = true | |
197 | ------------------------ | |
198 | ||
e28eae31 | 199 | This does not force normalization of text files, but does ensure |
fd6cce9e EB |
200 | that text files that you introduce to the repository have their line |
201 | endings normalized to LF when they are added, and that files that are | |
942e7747 | 202 | already normalized in the repository stay normalized. |
fd6cce9e | 203 | |
e28eae31 TB |
204 | If you want to ensure that text files that any contributor introduces to |
205 | the repository have their line endings normalized, you can set the | |
206 | `text` attribute to "auto" for _all_ files. | |
88e7fdf2 | 207 | |
fd6cce9e | 208 | ------------------------ |
5ec3e670 | 209 | * text=auto |
fd6cce9e EB |
210 | ------------------------ |
211 | ||
e28eae31 TB |
212 | The attributes allow a fine-grained control, how the line endings |
213 | are converted. | |
214 | Here is an example that will make Git normalize .txt, .vcproj and .sh | |
215 | files, ensure that .vcproj files have CRLF and .sh files have LF in | |
216 | the working directory, and prevent .jpg files from being normalized | |
217 | regardless of their content. | |
218 | ||
219 | ------------------------ | |
220 | * text=auto | |
221 | *.txt text | |
222 | *.vcproj text eol=crlf | |
223 | *.sh text eol=lf | |
224 | *.jpg -text | |
225 | ------------------------ | |
226 | ||
227 | NOTE: When `text=auto` conversion is enabled in a cross-platform | |
228 | project using push and pull to a central repository the text files | |
229 | containing CRLFs should be normalized. | |
fd6cce9e | 230 | |
e28eae31 | 231 | From a clean working directory: |
fd6cce9e EB |
232 | |
233 | ------------------------------------------------- | |
e28eae31 | 234 | $ echo "* text=auto" >.gitattributes |
773a8891 | 235 | $ git read-tree --empty # Clean index, force re-scan of working directory |
85999743 | 236 | $ git add . |
fd6cce9e | 237 | $ git status # Show files that will be normalized |
fd6cce9e EB |
238 | $ git commit -m "Introduce end-of-line normalization" |
239 | ------------------------------------------------- | |
240 | ||
241 | If any files that should not be normalized show up in 'git status', | |
5ec3e670 | 242 | unset their `text` attribute before running 'git add -u'. |
fd6cce9e EB |
243 | |
244 | ------------------------ | |
5ec3e670 | 245 | manual.pdf -text |
fd6cce9e | 246 | ------------------------ |
88e7fdf2 | 247 | |
2de9b711 | 248 | Conversely, text files that Git does not detect can have normalization |
fd6cce9e | 249 | enabled manually. |
88e7fdf2 | 250 | |
fd6cce9e | 251 | ------------------------ |
5ec3e670 | 252 | weirdchars.txt text |
fd6cce9e | 253 | ------------------------ |
88e7fdf2 | 254 | |
2de9b711 | 255 | If `core.safecrlf` is set to "true" or "warn", Git verifies if |
21e5ad50 | 256 | the conversion is reversible for the current setting of |
2de9b711 TA |
257 | `core.autocrlf`. For "true", Git rejects irreversible |
258 | conversions; for "warn", Git only prints a warning but accepts | |
21e5ad50 SP |
259 | an irreversible conversion. The safety triggers to prevent such |
260 | a conversion done to the files in the work tree, but there are a | |
261 | few exceptions. Even though... | |
262 | ||
0b444cdb | 263 | - 'git add' itself does not touch the files in the work tree, the |
21e5ad50 SP |
264 | next checkout would, so the safety triggers; |
265 | ||
0b444cdb | 266 | - 'git apply' to update a text file with a patch does touch the files |
21e5ad50 SP |
267 | in the work tree, but the operation is about text files and CRLF |
268 | conversion is about fixing the line ending inconsistencies, so the | |
269 | safety does not trigger; | |
270 | ||
0b444cdb TR |
271 | - 'git diff' itself does not touch the files in the work tree, it is |
272 | often run to inspect the changes you intend to next 'git add'. To | |
21e5ad50 SP |
273 | catch potential problems early, safety triggers. |
274 | ||
88e7fdf2 | 275 | |
3fed15f5 JH |
276 | `ident` |
277 | ^^^^^^^ | |
278 | ||
2de9b711 | 279 | When the attribute `ident` is set for a path, Git replaces |
2c850f12 | 280 | `$Id$` in the blob object with `$Id:`, followed by the |
3fed15f5 JH |
281 | 40-character hexadecimal blob object name, followed by a dollar |
282 | sign `$` upon checkout. Any byte sequence that begins with | |
af9b54bb AP |
283 | `$Id:` and ends with `$` in the worktree file is replaced |
284 | with `$Id$` upon check-in. | |
3fed15f5 JH |
285 | |
286 | ||
aa4ed402 JH |
287 | `filter` |
288 | ^^^^^^^^ | |
289 | ||
c05ef938 | 290 | A `filter` attribute can be set to a string value that names a |
aa4ed402 JH |
291 | filter driver specified in the configuration. |
292 | ||
c05ef938 | 293 | A filter driver consists of a `clean` command and a `smudge` |
aa4ed402 | 294 | command, either of which can be left unspecified. Upon |
c05ef938 WC |
295 | checkout, when the `smudge` command is specified, the command is |
296 | fed the blob object from its standard input, and its standard | |
297 | output is used to update the worktree file. Similarly, the | |
298 | `clean` command is used to convert the contents of worktree file | |
edcc8581 LS |
299 | upon checkin. By default these commands process only a single |
300 | blob and terminate. If a long running `process` filter is used | |
301 | in place of `clean` and/or `smudge` filters, then Git can process | |
302 | all blobs with a single filter command invocation for the entire | |
303 | life of a single Git command, for example `git add --all`. If a | |
304 | long running `process` filter is configured then it always takes | |
305 | precedence over a configured single blob filter. See section | |
306 | below for the description of the protocol used to communicate with | |
307 | a `process` filter. | |
aa4ed402 | 308 | |
36daaaca JB |
309 | One use of the content filtering is to massage the content into a shape |
310 | that is more convenient for the platform, filesystem, and the user to use. | |
311 | For this mode of operation, the key phrase here is "more convenient" and | |
312 | not "turning something unusable into usable". In other words, the intent | |
313 | is that if someone unsets the filter driver definition, or does not have | |
314 | the appropriate filter program, the project should still be usable. | |
315 | ||
316 | Another use of the content filtering is to store the content that cannot | |
317 | be directly used in the repository (e.g. a UUID that refers to the true | |
2de9b711 | 318 | content stored outside Git, or an encrypted content) and turn it into a |
36daaaca JB |
319 | usable form upon checkout (e.g. download the external content, or decrypt |
320 | the encrypted content). | |
321 | ||
322 | These two filters behave differently, and by default, a filter is taken as | |
323 | the former, massaging the contents into more convenient shape. A missing | |
324 | filter driver definition in the config, or a filter driver that exits with | |
325 | a non-zero status, is not an error but makes the filter a no-op passthru. | |
326 | ||
327 | You can declare that a filter turns a content that by itself is unusable | |
328 | into a usable content by setting the filter.<driver>.required configuration | |
329 | variable to `true`. | |
aa4ed402 | 330 | |
d79f5d17 NS |
331 | For example, in .gitattributes, you would assign the `filter` |
332 | attribute for paths. | |
333 | ||
334 | ------------------------ | |
335 | *.c filter=indent | |
336 | ------------------------ | |
337 | ||
338 | Then you would define a "filter.indent.clean" and "filter.indent.smudge" | |
339 | configuration in your .git/config to specify a pair of commands to | |
340 | modify the contents of C programs when the source files are checked | |
341 | in ("clean" is run) and checked out (no change is made because the | |
342 | command is "cat"). | |
343 | ||
344 | ------------------------ | |
345 | [filter "indent"] | |
346 | clean = indent | |
347 | smudge = cat | |
348 | ------------------------ | |
349 | ||
f217f0e8 EB |
350 | For best results, `clean` should not alter its output further if it is |
351 | run twice ("clean->clean" should be equivalent to "clean"), and | |
352 | multiple `smudge` commands should not alter `clean`'s output | |
353 | ("smudge->smudge->clean" should be equivalent to "clean"). See the | |
354 | section on merging below. | |
355 | ||
356 | The "indent" filter is well-behaved in this regard: it will not modify | |
357 | input that is already correctly indented. In this case, the lack of a | |
358 | smudge filter means that the clean filter _must_ accept its own output | |
359 | without modifying it. | |
360 | ||
36daaaca JB |
361 | If a filter _must_ succeed in order to make the stored contents usable, |
362 | you can declare that the filter is `required`, in the configuration: | |
363 | ||
364 | ------------------------ | |
365 | [filter "crypt"] | |
366 | clean = openssl enc ... | |
367 | smudge = openssl enc -d ... | |
368 | required | |
369 | ------------------------ | |
370 | ||
a2b665de PW |
371 | Sequence "%f" on the filter command line is replaced with the name of |
372 | the file the filter is working on. A filter might use this in keyword | |
373 | substitution. For example: | |
374 | ||
375 | ------------------------ | |
376 | [filter "p4"] | |
377 | clean = git-p4-filter --clean %f | |
378 | smudge = git-p4-filter --smudge %f | |
379 | ------------------------ | |
380 | ||
52db4b04 JH |
381 | Note that "%f" is the name of the path that is being worked on. Depending |
382 | on the version that is being filtered, the corresponding file on disk may | |
383 | not exist, or may have different contents. So, smudge and clean commands | |
384 | should not try to access the file on disk, but only act as filters on the | |
385 | content provided to them on standard input. | |
aa4ed402 | 386 | |
edcc8581 LS |
387 | Long Running Filter Process |
388 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
389 | ||
390 | If the filter command (a string value) is defined via | |
391 | `filter.<driver>.process` then Git can process all blobs with a | |
392 | single filter invocation for the entire life of a single Git | |
393 | command. This is achieved by using a packet format (pkt-line, | |
394 | see technical/protocol-common.txt) based protocol over standard | |
395 | input and standard output as follows. All packets, except for the | |
396 | "*CONTENT" packets and the "0000" flush packet, are considered | |
397 | text and therefore are terminated by a LF. | |
398 | ||
399 | Git starts the filter when it encounters the first file | |
400 | that needs to be cleaned or smudged. After the filter started | |
401 | Git sends a welcome message ("git-filter-client"), a list of supported | |
402 | protocol version numbers, and a flush packet. Git expects to read a welcome | |
403 | response message ("git-filter-server"), exactly one protocol version number | |
404 | from the previously sent list, and a flush packet. All further | |
405 | communication will be based on the selected version. The remaining | |
406 | protocol description below documents "version=2". Please note that | |
407 | "version=42" in the example below does not exist and is only there | |
408 | to illustrate how the protocol would look like with more than one | |
409 | version. | |
410 | ||
411 | After the version negotiation Git sends a list of all capabilities that | |
412 | it supports and a flush packet. Git expects to read a list of desired | |
413 | capabilities, which must be a subset of the supported capabilities list, | |
414 | and a flush packet as response: | |
415 | ------------------------ | |
416 | packet: git> git-filter-client | |
417 | packet: git> version=2 | |
418 | packet: git> version=42 | |
419 | packet: git> 0000 | |
420 | packet: git< git-filter-server | |
421 | packet: git< version=2 | |
422 | packet: git< 0000 | |
423 | packet: git> capability=clean | |
424 | packet: git> capability=smudge | |
425 | packet: git> capability=not-yet-invented | |
426 | packet: git> 0000 | |
427 | packet: git< capability=clean | |
428 | packet: git< capability=smudge | |
429 | packet: git< 0000 | |
430 | ------------------------ | |
2841e8f8 LS |
431 | Supported filter capabilities in version 2 are "clean", "smudge", |
432 | and "delay". | |
edcc8581 LS |
433 | |
434 | Afterwards Git sends a list of "key=value" pairs terminated with | |
435 | a flush packet. The list will contain at least the filter command | |
436 | (based on the supported capabilities) and the pathname of the file | |
437 | to filter relative to the repository root. Right after the flush packet | |
438 | Git sends the content split in zero or more pkt-line packets and a | |
439 | flush packet to terminate content. Please note, that the filter | |
440 | must not send any response before it received the content and the | |
c6b0831c LS |
441 | final flush packet. Also note that the "value" of a "key=value" pair |
442 | can contain the "=" character whereas the key would never contain | |
443 | that character. | |
edcc8581 LS |
444 | ------------------------ |
445 | packet: git> command=smudge | |
446 | packet: git> pathname=path/testfile.dat | |
447 | packet: git> 0000 | |
448 | packet: git> CONTENT | |
449 | packet: git> 0000 | |
450 | ------------------------ | |
451 | ||
452 | The filter is expected to respond with a list of "key=value" pairs | |
453 | terminated with a flush packet. If the filter does not experience | |
454 | problems then the list must contain a "success" status. Right after | |
455 | these packets the filter is expected to send the content in zero | |
456 | or more pkt-line packets and a flush packet at the end. Finally, a | |
457 | second list of "key=value" pairs terminated with a flush packet | |
458 | is expected. The filter can change the status in the second list | |
459 | or keep the status as is with an empty list. Please note that the | |
460 | empty list must be terminated with a flush packet regardless. | |
461 | ||
462 | ------------------------ | |
463 | packet: git< status=success | |
464 | packet: git< 0000 | |
465 | packet: git< SMUDGED_CONTENT | |
466 | packet: git< 0000 | |
467 | packet: git< 0000 # empty list, keep "status=success" unchanged! | |
468 | ------------------------ | |
469 | ||
470 | If the result content is empty then the filter is expected to respond | |
471 | with a "success" status and a flush packet to signal the empty content. | |
472 | ------------------------ | |
473 | packet: git< status=success | |
474 | packet: git< 0000 | |
475 | packet: git< 0000 # empty content! | |
476 | packet: git< 0000 # empty list, keep "status=success" unchanged! | |
477 | ------------------------ | |
478 | ||
479 | In case the filter cannot or does not want to process the content, | |
480 | it is expected to respond with an "error" status. | |
481 | ------------------------ | |
482 | packet: git< status=error | |
483 | packet: git< 0000 | |
484 | ------------------------ | |
485 | ||
486 | If the filter experiences an error during processing, then it can | |
487 | send the status "error" after the content was (partially or | |
488 | completely) sent. | |
489 | ------------------------ | |
490 | packet: git< status=success | |
491 | packet: git< 0000 | |
492 | packet: git< HALF_WRITTEN_ERRONEOUS_CONTENT | |
493 | packet: git< 0000 | |
494 | packet: git< status=error | |
495 | packet: git< 0000 | |
496 | ------------------------ | |
497 | ||
498 | In case the filter cannot or does not want to process the content | |
499 | as well as any future content for the lifetime of the Git process, | |
500 | then it is expected to respond with an "abort" status at any point | |
501 | in the protocol. | |
502 | ------------------------ | |
503 | packet: git< status=abort | |
504 | packet: git< 0000 | |
505 | ------------------------ | |
506 | ||
507 | Git neither stops nor restarts the filter process in case the | |
508 | "error"/"abort" status is set. However, Git sets its exit code | |
509 | according to the `filter.<driver>.required` flag, mimicking the | |
510 | behavior of the `filter.<driver>.clean` / `filter.<driver>.smudge` | |
511 | mechanism. | |
512 | ||
513 | If the filter dies during the communication or does not adhere to | |
514 | the protocol then Git will stop the filter process and restart it | |
515 | with the next file that needs to be processed. Depending on the | |
516 | `filter.<driver>.required` flag Git will interpret that as error. | |
517 | ||
2841e8f8 LS |
518 | After the filter has processed a command it is expected to wait for |
519 | a "key=value" list containing the next command. Git will close | |
edcc8581 LS |
520 | the command pipe on exit. The filter is expected to detect EOF |
521 | and exit gracefully on its own. Git will wait until the filter | |
522 | process has stopped. | |
523 | ||
2841e8f8 LS |
524 | Delay |
525 | ^^^^^ | |
526 | ||
527 | If the filter supports the "delay" capability, then Git can send the | |
528 | flag "can-delay" after the filter command and pathname. This flag | |
529 | denotes that the filter can delay filtering the current blob (e.g. to | |
530 | compensate network latencies) by responding with no content but with | |
531 | the status "delayed" and a flush packet. | |
532 | ------------------------ | |
533 | packet: git> command=smudge | |
534 | packet: git> pathname=path/testfile.dat | |
535 | packet: git> can-delay=1 | |
536 | packet: git> 0000 | |
537 | packet: git> CONTENT | |
538 | packet: git> 0000 | |
539 | packet: git< status=delayed | |
540 | packet: git< 0000 | |
541 | ------------------------ | |
542 | ||
543 | If the filter supports the "delay" capability then it must support the | |
544 | "list_available_blobs" command. If Git sends this command, then the | |
545 | filter is expected to return a list of pathnames representing blobs | |
546 | that have been delayed earlier and are now available. | |
547 | The list must be terminated with a flush packet followed | |
548 | by a "success" status that is also terminated with a flush packet. If | |
549 | no blobs for the delayed paths are available, yet, then the filter is | |
550 | expected to block the response until at least one blob becomes | |
551 | available. The filter can tell Git that it has no more delayed blobs | |
552 | by sending an empty list. As soon as the filter responds with an empty | |
553 | list, Git stops asking. All blobs that Git has not received at this | |
554 | point are considered missing and will result in an error. | |
555 | ||
556 | ------------------------ | |
557 | packet: git> command=list_available_blobs | |
558 | packet: git> 0000 | |
559 | packet: git< pathname=path/testfile.dat | |
560 | packet: git< pathname=path/otherfile.dat | |
561 | packet: git< 0000 | |
562 | packet: git< status=success | |
563 | packet: git< 0000 | |
564 | ------------------------ | |
565 | ||
566 | After Git received the pathnames, it will request the corresponding | |
567 | blobs again. These requests contain a pathname and an empty content | |
568 | section. The filter is expected to respond with the smudged content | |
569 | in the usual way as explained above. | |
570 | ------------------------ | |
571 | packet: git> command=smudge | |
572 | packet: git> pathname=path/testfile.dat | |
573 | packet: git> 0000 | |
574 | packet: git> 0000 # empty content! | |
575 | packet: git< status=success | |
576 | packet: git< 0000 | |
577 | packet: git< SMUDGED_CONTENT | |
578 | packet: git< 0000 | |
579 | packet: git< 0000 # empty list, keep "status=success" unchanged! | |
580 | ------------------------ | |
581 | ||
582 | Example | |
583 | ^^^^^^^ | |
584 | ||
0f71fa27 LS |
585 | A long running filter demo implementation can be found in |
586 | `contrib/long-running-filter/example.pl` located in the Git | |
587 | core repository. If you develop your own long running filter | |
edcc8581 LS |
588 | process then the `GIT_TRACE_PACKET` environment variables can be |
589 | very helpful for debugging (see linkgit:git[1]). | |
590 | ||
591 | Please note that you cannot use an existing `filter.<driver>.clean` | |
592 | or `filter.<driver>.smudge` command with `filter.<driver>.process` | |
593 | because the former two use a different inter process communication | |
594 | protocol than the latter one. | |
595 | ||
596 | ||
aa4ed402 JH |
597 | Interaction between checkin/checkout attributes |
598 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
599 | ||
600 | In the check-in codepath, the worktree file is first converted | |
601 | with `filter` driver (if specified and corresponding driver | |
602 | defined), then the result is processed with `ident` (if | |
5ec3e670 | 603 | specified), and then finally with `text` (again, if specified |
aa4ed402 JH |
604 | and applicable). |
605 | ||
606 | In the check-out codepath, the blob content is first converted | |
5ec3e670 | 607 | with `text`, and then `ident` and fed to `filter`. |
aa4ed402 JH |
608 | |
609 | ||
f217f0e8 EB |
610 | Merging branches with differing checkin/checkout attributes |
611 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
612 | ||
613 | If you have added attributes to a file that cause the canonical | |
614 | repository format for that file to change, such as adding a | |
615 | clean/smudge filter or text/eol/ident attributes, merging anything | |
616 | where the attribute is not in place would normally cause merge | |
617 | conflicts. | |
618 | ||
2de9b711 | 619 | To prevent these unnecessary merge conflicts, Git can be told to run a |
f217f0e8 EB |
620 | virtual check-out and check-in of all three stages of a file when |
621 | resolving a three-way merge by setting the `merge.renormalize` | |
622 | configuration variable. This prevents changes caused by check-in | |
623 | conversion from causing spurious merge conflicts when a converted file | |
624 | is merged with an unconverted file. | |
625 | ||
626 | As long as a "smudge->clean" results in the same output as a "clean" | |
627 | even on files that are already smudged, this strategy will | |
628 | automatically resolve all filter-related conflicts. Filters that do | |
629 | not act in this way may cause additional merge conflicts that must be | |
630 | resolved manually. | |
631 | ||
632 | ||
88e7fdf2 JH |
633 | Generating diff text |
634 | ~~~~~~~~~~~~~~~~~~~~ | |
635 | ||
4f73e240 JN |
636 | `diff` |
637 | ^^^^^^ | |
638 | ||
2de9b711 TA |
639 | The attribute `diff` affects how Git generates diffs for particular |
640 | files. It can tell Git whether to generate a textual patch for the path | |
678852d9 | 641 | or to treat the path as a binary file. It can also affect what line is |
2de9b711 TA |
642 | shown on the hunk header `@@ -k,l +n,m @@` line, tell Git to use an |
643 | external command to generate the diff, or ask Git to convert binary | |
678852d9 | 644 | files to a text format before generating the diff. |
88e7fdf2 JH |
645 | |
646 | Set:: | |
647 | ||
648 | A path to which the `diff` attribute is set is treated | |
649 | as text, even when they contain byte values that | |
650 | normally never appear in text files, such as NUL. | |
651 | ||
652 | Unset:: | |
653 | ||
654 | A path to which the `diff` attribute is unset will | |
678852d9 JK |
655 | generate `Binary files differ` (or a binary patch, if |
656 | binary patches are enabled). | |
88e7fdf2 JH |
657 | |
658 | Unspecified:: | |
659 | ||
660 | A path to which the `diff` attribute is unspecified | |
661 | first gets its contents inspected, and if it looks like | |
6bf3b813 NTND |
662 | text and is smaller than core.bigFileThreshold, it is treated |
663 | as text. Otherwise it would generate `Binary files differ`. | |
88e7fdf2 | 664 | |
2cc3167c JH |
665 | String:: |
666 | ||
678852d9 JK |
667 | Diff is shown using the specified diff driver. Each driver may |
668 | specify one or more options, as described in the following | |
669 | section. The options for the diff driver "foo" are defined | |
670 | by the configuration variables in the "diff.foo" section of the | |
2de9b711 | 671 | Git config file. |
2cc3167c JH |
672 | |
673 | ||
678852d9 JK |
674 | Defining an external diff driver |
675 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
2cc3167c JH |
676 | |
677 | The definition of a diff driver is done in `gitconfig`, not | |
678 | `gitattributes` file, so strictly speaking this manual page is a | |
679 | wrong place to talk about it. However... | |
680 | ||
678852d9 | 681 | To define an external diff driver `jcdiff`, add a section to your |
2cc3167c JH |
682 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: |
683 | ||
684 | ---------------------------------------------------------------- | |
685 | [diff "jcdiff"] | |
686 | command = j-c-diff | |
687 | ---------------------------------------------------------------- | |
688 | ||
2de9b711 | 689 | When Git needs to show you a diff for the path with `diff` |
2cc3167c JH |
690 | attribute set to `jcdiff`, it calls the command you specified |
691 | with the above configuration, i.e. `j-c-diff`, with 7 | |
692 | parameters, just like `GIT_EXTERNAL_DIFF` program is called. | |
9e1f0a85 | 693 | See linkgit:git[1] for details. |
88e7fdf2 JH |
694 | |
695 | ||
ae7aa499 JH |
696 | Defining a custom hunk-header |
697 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
698 | ||
c882c01e | 699 | Each group of changes (called a "hunk") in the textual diff output |
ae7aa499 JH |
700 | is prefixed with a line of the form: |
701 | ||
702 | @@ -k,l +n,m @@ TEXT | |
703 | ||
c882c01e GD |
704 | This is called a 'hunk header'. The "TEXT" portion is by default a line |
705 | that begins with an alphabet, an underscore or a dollar sign; this | |
706 | matches what GNU 'diff -p' output uses. This default selection however | |
707 | is not suited for some contents, and you can use a customized pattern | |
708 | to make a selection. | |
ae7aa499 | 709 | |
c882c01e | 710 | First, in .gitattributes, you would assign the `diff` attribute |
ae7aa499 JH |
711 | for paths. |
712 | ||
713 | ------------------------ | |
714 | *.tex diff=tex | |
715 | ------------------------ | |
716 | ||
edb7e82f | 717 | Then, you would define a "diff.tex.xfuncname" configuration to |
ae7aa499 | 718 | specify a regular expression that matches a line that you would |
c4c86d23 JK |
719 | want to appear as the hunk header "TEXT". Add a section to your |
720 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
ae7aa499 JH |
721 | |
722 | ------------------------ | |
723 | [diff "tex"] | |
45d9414f | 724 | xfuncname = "^(\\\\(sub)*section\\{.*)$" |
ae7aa499 JH |
725 | ------------------------ |
726 | ||
727 | Note. A single level of backslashes are eaten by the | |
728 | configuration file parser, so you would need to double the | |
729 | backslashes; the pattern above picks a line that begins with a | |
02783075 | 730 | backslash, and zero or more occurrences of `sub` followed by |
ae7aa499 JH |
731 | `section` followed by open brace, to the end of line. |
732 | ||
733 | There are a few built-in patterns to make this easier, and `tex` | |
734 | is one of them, so you do not have to write the above in your | |
735 | configuration file (you still need to enable this with the | |
d08ed6d6 GH |
736 | attribute mechanism, via `.gitattributes`). The following built in |
737 | patterns are available: | |
738 | ||
e90d065e AJ |
739 | - `ada` suitable for source code in the Ada language. |
740 | ||
23b5beb2 GH |
741 | - `bibtex` suitable for files with BibTeX coded references. |
742 | ||
80c49c3d TR |
743 | - `cpp` suitable for source code in the C and C++ languages. |
744 | ||
b221207d PO |
745 | - `csharp` suitable for source code in the C# language. |
746 | ||
0719f3ee WD |
747 | - `css` suitable for cascading style sheets. |
748 | ||
909a5494 BC |
749 | - `fortran` suitable for source code in the Fortran language. |
750 | ||
69f9c87d ZB |
751 | - `fountain` suitable for Fountain documents. |
752 | ||
af9ce1ff AE |
753 | - `html` suitable for HTML/XHTML documents. |
754 | ||
b66e00f1 | 755 | - `java` suitable for source code in the Java language. |
d08ed6d6 | 756 | |
53b10a14 GH |
757 | - `matlab` suitable for source code in the MATLAB language. |
758 | ||
5d1e958e JS |
759 | - `objc` suitable for source code in the Objective-C language. |
760 | ||
d08ed6d6 GH |
761 | - `pascal` suitable for source code in the Pascal/Delphi language. |
762 | ||
71a5d4bc JN |
763 | - `perl` suitable for source code in the Perl language. |
764 | ||
af9ce1ff AE |
765 | - `php` suitable for source code in the PHP language. |
766 | ||
7c17205b KS |
767 | - `python` suitable for source code in the Python language. |
768 | ||
d08ed6d6 GH |
769 | - `ruby` suitable for source code in the Ruby language. |
770 | ||
771 | - `tex` suitable for source code for LaTeX documents. | |
ae7aa499 JH |
772 | |
773 | ||
80c49c3d TR |
774 | Customizing word diff |
775 | ^^^^^^^^^^^^^^^^^^^^^ | |
776 | ||
882749a0 | 777 | You can customize the rules that `git diff --word-diff` uses to |
80c49c3d | 778 | split words in a line, by specifying an appropriate regular expression |
ae3b970a | 779 | in the "diff.*.wordRegex" configuration variable. For example, in TeX |
80c49c3d TR |
780 | a backslash followed by a sequence of letters forms a command, but |
781 | several such commands can be run together without intervening | |
c4c86d23 JK |
782 | whitespace. To separate them, use a regular expression in your |
783 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
80c49c3d TR |
784 | |
785 | ------------------------ | |
786 | [diff "tex"] | |
ae3b970a | 787 | wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+" |
80c49c3d TR |
788 | ------------------------ |
789 | ||
790 | A built-in pattern is provided for all languages listed in the | |
791 | previous section. | |
792 | ||
793 | ||
678852d9 JK |
794 | Performing text diffs of binary files |
795 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
796 | ||
797 | Sometimes it is desirable to see the diff of a text-converted | |
798 | version of some binary files. For example, a word processor | |
799 | document can be converted to an ASCII text representation, and | |
800 | the diff of the text shown. Even though this conversion loses | |
801 | some information, the resulting diff is useful for human | |
802 | viewing (but cannot be applied directly). | |
803 | ||
804 | The `textconv` config option is used to define a program for | |
805 | performing such a conversion. The program should take a single | |
806 | argument, the name of a file to convert, and produce the | |
807 | resulting text on stdout. | |
808 | ||
809 | For example, to show the diff of the exif information of a | |
810 | file instead of the binary information (assuming you have the | |
c4c86d23 JK |
811 | exif tool installed), add the following section to your |
812 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file): | |
678852d9 JK |
813 | |
814 | ------------------------ | |
815 | [diff "jpg"] | |
816 | textconv = exif | |
817 | ------------------------ | |
818 | ||
819 | NOTE: The text conversion is generally a one-way conversion; | |
820 | in this example, we lose the actual image contents and focus | |
821 | just on the text data. This means that diffs generated by | |
822 | textconv are _not_ suitable for applying. For this reason, | |
823 | only `git diff` and the `git log` family of commands (i.e., | |
824 | log, whatchanged, show) will perform text conversion. `git | |
825 | format-patch` will never generate this output. If you want to | |
826 | send somebody a text-converted diff of a binary file (e.g., | |
827 | because it quickly conveys the changes you have made), you | |
828 | should generate it separately and send it as a comment _in | |
829 | addition to_ the usual binary diff that you might send. | |
830 | ||
d9bae1a1 | 831 | Because text conversion can be slow, especially when doing a |
2de9b711 | 832 | large number of them with `git log -p`, Git provides a mechanism |
d9bae1a1 JK |
833 | to cache the output and use it in future diffs. To enable |
834 | caching, set the "cachetextconv" variable in your diff driver's | |
835 | config. For example: | |
836 | ||
837 | ------------------------ | |
838 | [diff "jpg"] | |
839 | textconv = exif | |
840 | cachetextconv = true | |
841 | ------------------------ | |
842 | ||
843 | This will cache the result of running "exif" on each blob | |
844 | indefinitely. If you change the textconv config variable for a | |
2de9b711 | 845 | diff driver, Git will automatically invalidate the cache entries |
d9bae1a1 JK |
846 | and re-run the textconv filter. If you want to invalidate the |
847 | cache manually (e.g., because your version of "exif" was updated | |
848 | and now produces better output), you can remove the cache | |
849 | manually with `git update-ref -d refs/notes/textconv/jpg` (where | |
850 | "jpg" is the name of the diff driver, as in the example above). | |
678852d9 | 851 | |
55601c6a JK |
852 | Choosing textconv versus external diff |
853 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
854 | ||
855 | If you want to show differences between binary or specially-formatted | |
856 | blobs in your repository, you can choose to use either an external diff | |
857 | command, or to use textconv to convert them to a diff-able text format. | |
858 | Which method you choose depends on your exact situation. | |
859 | ||
860 | The advantage of using an external diff command is flexibility. You are | |
861 | not bound to find line-oriented changes, nor is it necessary for the | |
862 | output to resemble unified diff. You are free to locate and report | |
863 | changes in the most appropriate way for your data format. | |
864 | ||
865 | A textconv, by comparison, is much more limiting. You provide a | |
2de9b711 | 866 | transformation of the data into a line-oriented text format, and Git |
55601c6a JK |
867 | uses its regular diff tools to generate the output. There are several |
868 | advantages to choosing this method: | |
869 | ||
870 | 1. Ease of use. It is often much simpler to write a binary to text | |
871 | transformation than it is to perform your own diff. In many cases, | |
872 | existing programs can be used as textconv filters (e.g., exif, | |
873 | odt2txt). | |
874 | ||
875 | 2. Git diff features. By performing only the transformation step | |
2de9b711 | 876 | yourself, you can still utilize many of Git's diff features, |
55601c6a JK |
877 | including colorization, word-diff, and combined diffs for merges. |
878 | ||
879 | 3. Caching. Textconv caching can speed up repeated diffs, such as those | |
880 | you might trigger by running `git log -p`. | |
881 | ||
882 | ||
ab435611 JK |
883 | Marking files as binary |
884 | ^^^^^^^^^^^^^^^^^^^^^^^ | |
885 | ||
886 | Git usually guesses correctly whether a blob contains text or binary | |
887 | data by examining the beginning of the contents. However, sometimes you | |
888 | may want to override its decision, either because a blob contains binary | |
889 | data later in the file, or because the content, while technically | |
890 | composed of text characters, is opaque to a human reader. For example, | |
f745acb0 | 891 | many postscript files contain only ASCII characters, but produce noisy |
ab435611 JK |
892 | and meaningless diffs. |
893 | ||
894 | The simplest way to mark a file as binary is to unset the diff | |
895 | attribute in the `.gitattributes` file: | |
896 | ||
897 | ------------------------ | |
898 | *.ps -diff | |
899 | ------------------------ | |
900 | ||
2de9b711 | 901 | This will cause Git to generate `Binary files differ` (or a binary |
ab435611 JK |
902 | patch, if binary patches are enabled) instead of a regular diff. |
903 | ||
904 | However, one may also want to specify other diff driver attributes. For | |
905 | example, you might want to use `textconv` to convert postscript files to | |
f745acb0 | 906 | an ASCII representation for human viewing, but otherwise treat them as |
ab435611 JK |
907 | binary files. You cannot specify both `-diff` and `diff=ps` attributes. |
908 | The solution is to use the `diff.*.binary` config option: | |
909 | ||
910 | ------------------------ | |
911 | [diff "ps"] | |
912 | textconv = ps2ascii | |
913 | binary = true | |
914 | ------------------------ | |
915 | ||
88e7fdf2 JH |
916 | Performing a three-way merge |
917 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
918 | ||
4f73e240 JN |
919 | `merge` |
920 | ^^^^^^^ | |
921 | ||
b547ce0b | 922 | The attribute `merge` affects how three versions of a file are |
88e7fdf2 | 923 | merged when a file-level merge is necessary during `git merge`, |
57f6ec02 | 924 | and other commands such as `git revert` and `git cherry-pick`. |
88e7fdf2 JH |
925 | |
926 | Set:: | |
927 | ||
928 | Built-in 3-way merge driver is used to merge the | |
2fd02c92 | 929 | contents in a way similar to 'merge' command of `RCS` |
88e7fdf2 JH |
930 | suite. This is suitable for ordinary text files. |
931 | ||
932 | Unset:: | |
933 | ||
934 | Take the version from the current branch as the | |
935 | tentative merge result, and declare that the merge has | |
b547ce0b | 936 | conflicts. This is suitable for binary files that do |
88e7fdf2 JH |
937 | not have a well-defined merge semantics. |
938 | ||
939 | Unspecified:: | |
940 | ||
941 | By default, this uses the same built-in 3-way merge | |
b547ce0b AS |
942 | driver as is the case when the `merge` attribute is set. |
943 | However, the `merge.default` configuration variable can name | |
944 | different merge driver to be used with paths for which the | |
88e7fdf2 JH |
945 | `merge` attribute is unspecified. |
946 | ||
2cc3167c | 947 | String:: |
88e7fdf2 JH |
948 | |
949 | 3-way merge is performed using the specified custom | |
950 | merge driver. The built-in 3-way merge driver can be | |
951 | explicitly specified by asking for "text" driver; the | |
952 | built-in "take the current branch" driver can be | |
b9d14ffb | 953 | requested with "binary". |
88e7fdf2 JH |
954 | |
955 | ||
0e545f75 JH |
956 | Built-in merge drivers |
957 | ^^^^^^^^^^^^^^^^^^^^^^ | |
958 | ||
959 | There are a few built-in low-level merge drivers defined that | |
960 | can be asked for via the `merge` attribute. | |
961 | ||
962 | text:: | |
963 | ||
964 | Usual 3-way file level merge for text files. Conflicted | |
965 | regions are marked with conflict markers `<<<<<<<`, | |
966 | `=======` and `>>>>>>>`. The version from your branch | |
967 | appears before the `=======` marker, and the version | |
968 | from the merged branch appears after the `=======` | |
969 | marker. | |
970 | ||
971 | binary:: | |
972 | ||
973 | Keep the version from your branch in the work tree, but | |
974 | leave the path in the conflicted state for the user to | |
975 | sort out. | |
976 | ||
977 | union:: | |
978 | ||
979 | Run 3-way file level merge for text files, but take | |
980 | lines from both versions, instead of leaving conflict | |
981 | markers. This tends to leave the added lines in the | |
982 | resulting file in random order and the user should | |
983 | verify the result. Do not use this if you do not | |
984 | understand the implications. | |
985 | ||
986 | ||
88e7fdf2 JH |
987 | Defining a custom merge driver |
988 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
989 | ||
0e545f75 JH |
990 | The definition of a merge driver is done in the `.git/config` |
991 | file, not in the `gitattributes` file, so strictly speaking this | |
992 | manual page is a wrong place to talk about it. However... | |
88e7fdf2 JH |
993 | |
994 | To define a custom merge driver `filfre`, add a section to your | |
995 | `$GIT_DIR/config` file (or `$HOME/.gitconfig` file) like this: | |
996 | ||
997 | ---------------------------------------------------------------- | |
998 | [merge "filfre"] | |
999 | name = feel-free merge driver | |
ef45bb1f | 1000 | driver = filfre %O %A %B %L %P |
88e7fdf2 JH |
1001 | recursive = binary |
1002 | ---------------------------------------------------------------- | |
1003 | ||
1004 | The `merge.*.name` variable gives the driver a human-readable | |
1005 | name. | |
1006 | ||
1007 | The `merge.*.driver` variable's value is used to construct a | |
1008 | command to run to merge ancestor's version (`%O`), current | |
1009 | version (`%A`) and the other branches' version (`%B`). These | |
1010 | three tokens are replaced with the names of temporary files that | |
1011 | hold the contents of these versions when the command line is | |
16758621 BW |
1012 | built. Additionally, %L will be replaced with the conflict marker |
1013 | size (see below). | |
88e7fdf2 JH |
1014 | |
1015 | The merge driver is expected to leave the result of the merge in | |
1016 | the file named with `%A` by overwriting it, and exit with zero | |
1017 | status if it managed to merge them cleanly, or non-zero if there | |
1018 | were conflicts. | |
1019 | ||
1020 | The `merge.*.recursive` variable specifies what other merge | |
1021 | driver to use when the merge driver is called for an internal | |
1022 | merge between common ancestors, when there are more than one. | |
1023 | When left unspecified, the driver itself is used for both | |
1024 | internal merge and the final merge. | |
1025 | ||
ef45bb1f JH |
1026 | The merge driver can learn the pathname in which the merged result |
1027 | will be stored via placeholder `%P`. | |
1028 | ||
88e7fdf2 | 1029 | |
4c734803 JH |
1030 | `conflict-marker-size` |
1031 | ^^^^^^^^^^^^^^^^^^^^^^ | |
1032 | ||
1033 | This attribute controls the length of conflict markers left in | |
1034 | the work tree file during a conflicted merge. Only setting to | |
1035 | the value to a positive integer has any meaningful effect. | |
1036 | ||
1037 | For example, this line in `.gitattributes` can be used to tell the merge | |
1038 | machinery to leave much longer (instead of the usual 7-character-long) | |
1039 | conflict markers when merging the file `Documentation/git-merge.txt` | |
1040 | results in a conflict. | |
1041 | ||
1042 | ------------------------ | |
1043 | Documentation/git-merge.txt conflict-marker-size=32 | |
1044 | ------------------------ | |
1045 | ||
1046 | ||
cf1b7869 JH |
1047 | Checking whitespace errors |
1048 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
1049 | ||
1050 | `whitespace` | |
1051 | ^^^^^^^^^^^^ | |
1052 | ||
1053 | The `core.whitespace` configuration variable allows you to define what | |
2fd02c92 | 1054 | 'diff' and 'apply' should consider whitespace errors for all paths in |
5162e697 | 1055 | the project (See linkgit:git-config[1]). This attribute gives you finer |
cf1b7869 JH |
1056 | control per path. |
1057 | ||
1058 | Set:: | |
1059 | ||
2de9b711 | 1060 | Notice all types of potential whitespace errors known to Git. |
f4b05a49 JS |
1061 | The tab width is taken from the value of the `core.whitespace` |
1062 | configuration variable. | |
cf1b7869 JH |
1063 | |
1064 | Unset:: | |
1065 | ||
1066 | Do not notice anything as error. | |
1067 | ||
1068 | Unspecified:: | |
1069 | ||
f4b05a49 | 1070 | Use the value of the `core.whitespace` configuration variable to |
cf1b7869 JH |
1071 | decide what to notice as error. |
1072 | ||
1073 | String:: | |
1074 | ||
1075 | Specify a comma separate list of common whitespace problems to | |
f4b05a49 | 1076 | notice in the same format as the `core.whitespace` configuration |
cf1b7869 JH |
1077 | variable. |
1078 | ||
1079 | ||
8a33dd8b JH |
1080 | Creating an archive |
1081 | ~~~~~~~~~~~~~~~~~~~ | |
1082 | ||
08b51f51 JH |
1083 | `export-ignore` |
1084 | ^^^^^^^^^^^^^^^ | |
1085 | ||
1086 | Files and directories with the attribute `export-ignore` won't be added to | |
1087 | archive files. | |
1088 | ||
8a33dd8b JH |
1089 | `export-subst` |
1090 | ^^^^^^^^^^^^^^ | |
1091 | ||
2de9b711 | 1092 | If the attribute `export-subst` is set for a file then Git will expand |
8a33dd8b | 1093 | several placeholders when adding this file to an archive. The |
08b51f51 | 1094 | expansion depends on the availability of a commit ID, i.e., if |
8a33dd8b JH |
1095 | linkgit:git-archive[1] has been given a tree instead of a commit or a |
1096 | tag then no replacement will be done. The placeholders are the same | |
1097 | as those for the option `--pretty=format:` of linkgit:git-log[1], | |
1098 | except that they need to be wrapped like this: `$Format:PLACEHOLDERS$` | |
1099 | in the file. E.g. the string `$Format:%H$` will be replaced by the | |
1100 | commit hash. | |
1101 | ||
1102 | ||
975457f1 NG |
1103 | Packing objects |
1104 | ~~~~~~~~~~~~~~~ | |
1105 | ||
1106 | `delta` | |
1107 | ^^^^^^^ | |
1108 | ||
1109 | Delta compression will not be attempted for blobs for paths with the | |
1110 | attribute `delta` set to false. | |
1111 | ||
1112 | ||
a2df1fb2 AG |
1113 | Viewing files in GUI tools |
1114 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
1115 | ||
1116 | `encoding` | |
1117 | ^^^^^^^^^^ | |
1118 | ||
1119 | The value of this attribute specifies the character encoding that should | |
1120 | be used by GUI tools (e.g. linkgit:gitk[1] and linkgit:git-gui[1]) to | |
1121 | display the contents of the relevant file. Note that due to performance | |
1122 | considerations linkgit:gitk[1] does not use this attribute unless you | |
1123 | manually enable per-file encodings in its options. | |
1124 | ||
1125 | If this attribute is not set or has an invalid value, the value of the | |
1126 | `gui.encoding` configuration variable is used instead | |
1127 | (See linkgit:git-config[1]). | |
1128 | ||
1129 | ||
0922570c | 1130 | USING MACRO ATTRIBUTES |
bbb896d8 JH |
1131 | ---------------------- |
1132 | ||
1133 | You do not want any end-of-line conversions applied to, nor textual diffs | |
1134 | produced for, any binary file you track. You would need to specify e.g. | |
1135 | ||
1136 | ------------ | |
5ec3e670 | 1137 | *.jpg -text -diff |
bbb896d8 JH |
1138 | ------------ |
1139 | ||
1140 | but that may become cumbersome, when you have many attributes. Using | |
0922570c | 1141 | macro attributes, you can define an attribute that, when set, also |
98e84066 | 1142 | sets or unsets a number of other attributes at the same time. The |
0922570c | 1143 | system knows a built-in macro attribute, `binary`: |
bbb896d8 JH |
1144 | |
1145 | ------------ | |
1146 | *.jpg binary | |
1147 | ------------ | |
1148 | ||
98e84066 | 1149 | Setting the "binary" attribute also unsets the "text" and "diff" |
0922570c | 1150 | attributes as above. Note that macro attributes can only be "Set", |
98e84066 MH |
1151 | though setting one might have the effect of setting or unsetting other |
1152 | attributes or even returning other attributes to the "Unspecified" | |
1153 | state. | |
bbb896d8 JH |
1154 | |
1155 | ||
0922570c | 1156 | DEFINING MACRO ATTRIBUTES |
bbb896d8 JH |
1157 | ------------------------- |
1158 | ||
e78e6967 MH |
1159 | Custom macro attributes can be defined only in top-level gitattributes |
1160 | files (`$GIT_DIR/info/attributes`, the `.gitattributes` file at the | |
1161 | top level of the working tree, or the global or system-wide | |
1162 | gitattributes files), not in `.gitattributes` files in working tree | |
1163 | subdirectories. The built-in macro attribute "binary" is equivalent | |
1164 | to: | |
bbb896d8 JH |
1165 | |
1166 | ------------ | |
155a4b71 | 1167 | [attr]binary -diff -merge -text |
bbb896d8 JH |
1168 | ------------ |
1169 | ||
1170 | ||
88e7fdf2 JH |
1171 | EXAMPLE |
1172 | ------- | |
1173 | ||
1174 | If you have these three `gitattributes` file: | |
1175 | ||
1176 | ---------------------------------------------------------------- | |
1177 | (in $GIT_DIR/info/attributes) | |
1178 | ||
1179 | a* foo !bar -baz | |
1180 | ||
1181 | (in .gitattributes) | |
1182 | abc foo bar baz | |
1183 | ||
1184 | (in t/.gitattributes) | |
1185 | ab* merge=filfre | |
1186 | abc -foo -bar | |
1187 | *.c frotz | |
1188 | ---------------------------------------------------------------- | |
1189 | ||
1190 | the attributes given to path `t/abc` are computed as follows: | |
1191 | ||
1192 | 1. By examining `t/.gitattributes` (which is in the same | |
2de9b711 | 1193 | directory as the path in question), Git finds that the first |
88e7fdf2 JH |
1194 | line matches. `merge` attribute is set. It also finds that |
1195 | the second line matches, and attributes `foo` and `bar` | |
1196 | are unset. | |
1197 | ||
1198 | 2. Then it examines `.gitattributes` (which is in the parent | |
1199 | directory), and finds that the first line matches, but | |
1200 | `t/.gitattributes` file already decided how `merge`, `foo` | |
1201 | and `bar` attributes should be given to this path, so it | |
1202 | leaves `foo` and `bar` unset. Attribute `baz` is set. | |
1203 | ||
5c759f96 | 1204 | 3. Finally it examines `$GIT_DIR/info/attributes`. This file |
88e7fdf2 JH |
1205 | is used to override the in-tree settings. The first line is |
1206 | a match, and `foo` is set, `bar` is reverted to unspecified | |
1207 | state, and `baz` is unset. | |
1208 | ||
02783075 | 1209 | As the result, the attributes assignment to `t/abc` becomes: |
88e7fdf2 JH |
1210 | |
1211 | ---------------------------------------------------------------- | |
1212 | foo set to true | |
1213 | bar unspecified | |
1214 | baz set to false | |
1215 | merge set to string value "filfre" | |
1216 | frotz unspecified | |
1217 | ---------------------------------------------------------------- | |
1218 | ||
1219 | ||
cde15181 MH |
1220 | SEE ALSO |
1221 | -------- | |
1222 | linkgit:git-check-attr[1]. | |
8460b2fc | 1223 | |
88e7fdf2 JH |
1224 | GIT |
1225 | --- | |
9e1f0a85 | 1226 | Part of the linkgit:git[1] suite |