| blob | 52a1052e40e1999923595ab9fb79b4d48c719936 |
1 '\" t
2 .\" Title: gitformat-pack
3 .\" Author: [FIXME: author] [see http://www.docbook.org/tdg5/en/html/author]
4 .\" Generator: DocBook XSL Stylesheets v1.79.2 <http://docbook.sf.net/>
5 .\" Date: 2024-08-19
6 .\" Manual: Git Manual
7 .\" Source: Git 2.46.0.267.gbb9c16bd4f
8 .\" Language: English
9 .\"
10 .TH "GITFORMAT\-PACK" "5" "2024-08-19" "Git 2\&.46\&.0\&.267\&.gbb9c16" "Git Manual"
11 .\" -----------------------------------------------------------------
12 .\" * Define some portability stuff
13 .\" -----------------------------------------------------------------
14 .\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
15 .\" http://bugs.debian.org/507673
16 .\" http://lists.gnu.org/archive/html/groff/2009-02/msg00013.html
17 .\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
18 .ie \n(.g .ds Aq \(aq
19 .el .ds Aq '
20 .\" -----------------------------------------------------------------
21 .\" * set default formatting
22 .\" -----------------------------------------------------------------
23 .\" disable hyphenation
24 .nh
25 .\" disable justification (adjust text to left margin only)
26 .ad l
27 .\" -----------------------------------------------------------------
28 .\" * MAIN CONTENT STARTS HERE *
29 .\" -----------------------------------------------------------------
30 .SH "NAME"
31 gitformat-pack \- Git pack format
32 .SH "SYNOPSIS"
33 .sp
34 .nf
35 $GIT_DIR/objects/pack/pack\-\fB\&.{pack,idx}
36 $GIT_DIR/objects/pack/pack\-\fR\&.rev
37 $GIT_DIR/objects/pack/pack\-*\&.mtimes
38 $GIT_DIR/objects/pack/multi\-pack\-index
39 .fi
40 .SH "DESCRIPTION"
41 .sp
42 The Git pack format is how Git stores most of its primary repository data\&. Over the lifetime of a repository, loose objects (if any) and smaller packs are consolidated into larger pack(s)\&. See \fBgit-gc\fR(1) and \fBgit-pack-objects\fR(1)\&.
43 .sp
44 The pack format is also used over\-the\-wire, see e\&.g\&. \fBgitprotocol-v2\fR(5), as well as being a part of other container formats in the case of \fBgitformat-bundle\fR(5)\&.
45 .SH "CHECKSUMS AND OBJECT IDS"
46 .sp
47 In a repository using the traditional SHA\-1, pack checksums, index checksums, and object IDs (object names) mentioned below are all computed using SHA\-1\&. Similarly, in SHA\-256 repositories, these values are computed using SHA\-256\&.
48 .SH "PACK\-*\&.PACK FILES HAVE THE FOLLOWING FORMAT:"
49 .sp
50 .RS 4
51 .ie n \{\
52 \h'-04'\(bu\h'+03'\c
53 .\}
54 .el \{\
55 .sp -1
56 .IP \(bu 2.3
57 .\}
58 A header appears at the beginning and consists of the following:
59 .sp
60 .if n \{\
61 .RS 4
62 .\}
63 .nf
64 4\-byte signature:
65 The signature is: {\*(AqP\*(Aq, \*(AqA\*(Aq, \*(AqC\*(Aq, \*(AqK\*(Aq}
66 .fi
67 .if n \{\
68 .RE
69 .\}
70 .sp
71 .if n \{\
72 .RS 4
73 .\}
74 .nf
75 4\-byte version number (network byte order):
76 Git currently accepts version number 2 or 3 but
77 generates version 2 only\&.
78 .fi
79 .if n \{\
80 .RE
81 .\}
82 .sp
83 .if n \{\
84 .RS 4
85 .\}
86 .nf
87 4\-byte number of objects contained in the pack (network byte order)
88 .fi
89 .if n \{\
90 .RE
91 .\}
92 .sp
93 .if n \{\
94 .RS 4
95 .\}
96 .nf
97 Observation: we cannot have more than 4G versions ;\-) and
98 more than 4G objects in a pack\&.
99 .fi
100 .if n \{\
101 .RE
102 .\}
103 .RE
104 .sp
105 .RS 4
106 .ie n \{\
107 \h'-04'\(bu\h'+03'\c
108 .\}
109 .el \{\
110 .sp -1
111 .IP \(bu 2.3
112 .\}
113 The header is followed by a number of object entries, each of which looks like this:
114 .sp
115 .if n \{\
116 .RS 4
117 .\}
118 .nf
119 (undeltified representation)
120 n\-byte type and length (3\-bit type, (n\-1)*7+4\-bit length)
121 compressed data
122 .fi
123 .if n \{\
124 .RE
125 .\}
126 .sp
127 .if n \{\
128 .RS 4
129 .\}
130 .nf
131 (deltified representation)
132 n\-byte type and length (3\-bit type, (n\-1)*7+4\-bit length)
133 base object name if OBJ_REF_DELTA or a negative relative
134 offset from the delta object\*(Aqs position in the pack if this
135 is an OBJ_OFS_DELTA object
136 compressed delta data
137 .fi
138 .if n \{\
139 .RE
140 .\}
141 .sp
142 .if n \{\
143 .RS 4
144 .\}
145 .nf
146 Observation: the length of each object is encoded in a variable
147 length format and is not constrained to 32\-bit or anything\&.
148 .fi
149 .if n \{\
150 .RE
151 .\}
152 .RE
153 .sp
154 .RS 4
155 .ie n \{\
156 \h'-04'\(bu\h'+03'\c
157 .\}
158 .el \{\
159 .sp -1
160 .IP \(bu 2.3
161 .\}
162 The trailer records a pack checksum of all of the above\&.
163 .RE
164 .SS "Object types"
165 .sp
166 Valid object types are:
167 .sp
168 .RS 4
169 .ie n \{\
170 \h'-04'\(bu\h'+03'\c
171 .\}
172 .el \{\
173 .sp -1
174 .IP \(bu 2.3
175 .\}
176 OBJ_COMMIT (1)
177 .RE
178 .sp
179 .RS 4
180 .ie n \{\
181 \h'-04'\(bu\h'+03'\c
182 .\}
183 .el \{\
184 .sp -1
185 .IP \(bu 2.3
186 .\}
187 OBJ_TREE (2)
188 .RE
189 .sp
190 .RS 4
191 .ie n \{\
192 \h'-04'\(bu\h'+03'\c
193 .\}
194 .el \{\
195 .sp -1
196 .IP \(bu 2.3
197 .\}
198 OBJ_BLOB (3)
199 .RE
200 .sp
201 .RS 4
202 .ie n \{\
203 \h'-04'\(bu\h'+03'\c
204 .\}
205 .el \{\
206 .sp -1
207 .IP \(bu 2.3
208 .\}
209 OBJ_TAG (4)
210 .RE
211 .sp
212 .RS 4
213 .ie n \{\
214 \h'-04'\(bu\h'+03'\c
215 .\}
216 .el \{\
217 .sp -1
218 .IP \(bu 2.3
219 .\}
220 OBJ_OFS_DELTA (6)
221 .RE
222 .sp
223 .RS 4
224 .ie n \{\
225 \h'-04'\(bu\h'+03'\c
226 .\}
227 .el \{\
228 .sp -1
229 .IP \(bu 2.3
230 .\}
231 OBJ_REF_DELTA (7)
232 .RE
233 .sp
234 Type 5 is reserved for future expansion\&. Type 0 is invalid\&.
235 .SS "Size encoding"
236 .sp
237 This document uses the following "size encoding" of non\-negative integers: From each byte, the seven least significant bits are used to form the resulting integer\&. As long as the most significant bit is 1, this process continues; the byte with MSB 0 provides the last seven bits\&. The seven\-bit chunks are concatenated\&. Later values are more significant\&.
238 .sp
239 This size encoding should not be confused with the "offset encoding", which is also used in this document\&.
240 .SS "Deltified representation"
241 .sp
242 Conceptually there are only four object types: commit, tree, tag and blob\&. However to save space, an object could be stored as a "delta" of another "base" object\&. These representations are assigned new types ofs\-delta and ref\-delta, which is only valid in a pack file\&.
243 .sp
244 Both ofs\-delta and ref\-delta store the "delta" to be applied to another object (called \fIbase object\fR) to reconstruct the object\&. The difference between them is, ref\-delta directly encodes base object name\&. If the base object is in the same pack, ofs\-delta encodes the offset of the base object in the pack instead\&.
245 .sp
246 The base object could also be deltified if it\(cqs in the same pack\&. Ref\-delta can also refer to an object outside the pack (i\&.e\&. the so\-called "thin pack")\&. When stored on disk however, the pack should be self contained to avoid cyclic dependency\&.
247 .sp
248 The delta data starts with the size of the base object and the size of the object to be reconstructed\&. These sizes are encoded using the size encoding from above\&. The remainder of the delta data is a sequence of instructions to reconstruct the object from the base object\&. If the base object is deltified, it must be converted to canonical form first\&. Each instruction appends more and more data to the target object until it\(cqs complete\&. There are two supported instructions so far: one for copying a byte range from the source object and one for inserting new data embedded in the instruction itself\&.
249 .sp
250 Each instruction has variable length\&. Instruction type is determined by the seventh bit of the first octet\&. The following diagrams follow the convention in RFC 1951 (Deflate compressed data format)\&.
251 .sp
252 .it 1 an-trap
253 .nr an-no-space-flag 1
254 .nr an-break-flag 1
255 .br
256 .ps +1
257 \fBInstruction to copy from base object\fR
258 .RS 4
259 .sp
260 .if n \{\
261 .RS 4
262 .\}
263 .nf
264 +\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+
265 | 1xxxxxxx | offset1 | offset2 | offset3 | offset4 | size1 | size2 | size3 |
266 +\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+
267 .fi
268 .if n \{\
269 .RE
270 .\}
271 .sp
272 This is the instruction format to copy a byte range from the source object\&. It encodes the offset to copy from and the number of bytes to copy\&. Offset and size are in little\-endian order\&.
273 .sp
274 All offset and size bytes are optional\&. This is to reduce the instruction size when encoding small offsets or sizes\&. The first seven bits in the first octet determine which of the next seven octets is present\&. If bit zero is set, offset1 is present\&. If bit one is set offset2 is present and so on\&.
275 .sp
276 Note that a more compact instruction does not change offset and size encoding\&. For example, if only offset2 is omitted like below, offset3 still contains bits 16\-23\&. It does not become offset2 and contains bits 8\-15 even if it\(cqs right next to offset1\&.
277 .sp
278 .if n \{\
279 .RS 4
280 .\}
281 .nf
282 +\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+
283 | 10000101 | offset1 | offset3 |
284 +\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-+
285 .fi
286 .if n \{\
287 .RE
288 .\}
289 .sp
290 In its most compact form, this instruction only takes up one byte (0x80) with both offset and size omitted, which will have default values zero\&. There is another exception: size zero is automatically converted to 0x10000\&.
291 .RE
292 .sp
293 .it 1 an-trap
294 .nr an-no-space-flag 1
295 .nr an-break-flag 1
296 .br
297 .ps +1
298 \fBInstruction to add new data\fR
299 .RS 4
300 .sp
301 .if n \{\
302 .RS 4
303 .\}
304 .nf
305 +\-\-\-\-\-\-\-\-\-\-+============+
306 | 0xxxxxxx | data |
307 +\-\-\-\-\-\-\-\-\-\-+============+
308 .fi
309 .if n \{\
310 .RE
311 .\}
312 .sp
313 This is the instruction to construct the target object without the base object\&. The following data is appended to the target object\&. The first seven bits of the first octet determine the size of data in bytes\&. The size must be non\-zero\&.
314 .RE
315 .sp
316 .it 1 an-trap
317 .nr an-no-space-flag 1
318 .nr an-break-flag 1
319 .br
320 .ps +1
321 \fBReserved instruction\fR
322 .RS 4
323 .sp
324 .if n \{\
325 .RS 4
326 .\}
327 .nf
328 +\-\-\-\-\-\-\-\-\-\-+============
329 | 00000000 |
330 +\-\-\-\-\-\-\-\-\-\-+============
331 .fi
332 .if n \{\
333 .RE
334 .\}
335 .sp
336 This is the instruction reserved for future expansion\&.
337 .RE
338 .SH "ORIGINAL (VERSION 1) PACK\-*\&.IDX FILES HAVE THE FOLLOWING FORMAT:"
339 .sp
340 .RS 4
341 .ie n \{\
342 \h'-04'\(bu\h'+03'\c
343 .\}
344 .el \{\
345 .sp -1
346 .IP \(bu 2.3
347 .\}
348 The header consists of 256 4\-byte network byte order integers\&. N\-th entry of this table records the number of objects in the corresponding pack, the first byte of whose object name is less than or equal to N\&. This is called the
349 \fIfirst\-level fan\-out\fR
350 table\&.
351 .RE
352 .sp
353 .RS 4
354 .ie n \{\
355 \h'-04'\(bu\h'+03'\c
356 .\}
357 .el \{\
358 .sp -1
359 .IP \(bu 2.3
360 .\}
361 The header is followed by sorted 24\-byte entries, one entry per object in the pack\&. Each entry is:
362 .sp
363 .if n \{\
364 .RS 4
365 .\}
366 .nf
367 4\-byte network byte order integer, recording where the
368 object is stored in the packfile as the offset from the
369 beginning\&.
370 .fi
371 .if n \{\
372 .RE
373 .\}
374 .sp
375 .if n \{\
376 .RS 4
377 .\}
378 .nf
379 one object name of the appropriate size\&.
380 .fi
381 .if n \{\
382 .RE
383 .\}
384 .RE
385 .sp
386 .RS 4
387 .ie n \{\
388 \h'-04'\(bu\h'+03'\c
389 .\}
390 .el \{\
391 .sp -1
392 .IP \(bu 2.3
393 .\}
394 The file is concluded with a trailer:
395 .sp
396 .if n \{\
397 .RS 4
398 .\}
399 .nf
400 A copy of the pack checksum at the end of the corresponding
401 packfile\&.
402 .fi
403 .if n \{\
404 .RE
405 .\}
406 .sp
407 .if n \{\
408 .RS 4
409 .\}
410 .nf
411 Index checksum of all of the above\&.
412 .fi
413 .if n \{\
414 .RE
415 .\}
416 .RE
417 .sp
418 Pack Idx file:
419 .sp
420 .if n \{\
421 .RS 4
422 .\}
423 .nf
424 \-\- +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
425 fanout | fanout[0] = 2 (for example) |\-\&.
426 table +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+ |
427 | fanout[1] | |
428 +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+ |
429 | fanout[2] | |
430 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
431 | fanout[255] = total objects |\-\-\-\&.
432 \-\- +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+ | |
433 main | offset | | |
434 index | object name 00XXXXXXXXXXXXXXXX | | |
435 table +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+ | |
436 | offset | | |
437 | object name 00XXXXXXXXXXXXXXXX | | |
438 +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+<+ |
439 \&.\-| offset | |
440 | | object name 01XXXXXXXXXXXXXXXX | |
441 | +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+ |
442 | | offset | |
443 | | object name 01XXXXXXXXXXXXXXXX | |
444 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
445 | | offset | |
446 | | object name FFXXXXXXXXXXXXXXXX | |
447 \-\-| +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+<\-\-+
448 trailer | | packfile checksum |
449 | +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
450 | | idxfile checksum |
451 | +\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
452 \&.\-\-\-\-\-\-\-\&.
453 |
454 Pack file entry: <+
455 .fi
456 .if n \{\
457 .RE
458 .\}
459 .sp
460 .if n \{\
461 .RS 4
462 .\}
463 .nf
464 packed object header:
465 1\-byte size extension bit (MSB)
466 type (next 3 bit)
467 size0 (lower 4\-bit)
468 n\-byte sizeN (as long as MSB is set, each 7\-bit)
469 size0\&.\&.sizeN form 4+7+7+\&.\&.+7 bit integer, size0
470 is the least significant part, and sizeN is the
471 most significant part\&.
472 packed object data:
473 If it is not DELTA, then deflated bytes (the size above
474 is the size before compression)\&.
475 If it is REF_DELTA, then
476 base object name (the size above is the
477 size of the delta data that follows)\&.
478 delta data, deflated\&.
479 If it is OFS_DELTA, then
480 n\-byte offset (see below) interpreted as a negative
481 offset from the type\-byte of the header of the
482 ofs\-delta entry (the size above is the size of
483 the delta data that follows)\&.
484 delta data, deflated\&.
485 .fi
486 .if n \{\
487 .RE
488 .\}
489 .sp
490 .if n \{\
491 .RS 4
492 .\}
493 .nf
494 offset encoding:
495 n bytes with MSB set in all but the last one\&.
496 The offset is then the number constructed by
497 concatenating the lower 7 bit of each byte, and
498 for n >= 2 adding 2^7 + 2^14 + \&.\&.\&. + 2^(7*(n\-1))
499 to the result\&.
500 .fi
501 .if n \{\
502 .RE
503 .\}
504 .SH "VERSION 2 PACK\-*\&.IDX FILES SUPPORT PACKS LARGER THAN 4 GIB, AND"
505 .sp
506 .if n \{\
507 .RS 4
508 .\}
509 .nf
510 have some other reorganizations\&. They have the format:
511 .fi
512 .if n \{\
513 .RE
514 .\}
515 .sp
516 .RS 4
517 .ie n \{\
518 \h'-04'\(bu\h'+03'\c
519 .\}
520 .el \{\
521 .sp -1
522 .IP \(bu 2.3
523 .\}
524 A 4\-byte magic number
525 \fI\e377tOc\fR
526 which is an unreasonable fanout[0] value\&.
527 .RE
528 .sp
529 .RS 4
530 .ie n \{\
531 \h'-04'\(bu\h'+03'\c
532 .\}
533 .el \{\
534 .sp -1
535 .IP \(bu 2.3
536 .\}
537 A 4\-byte version number (= 2)
538 .RE
539 .sp
540 .RS 4
541 .ie n \{\
542 \h'-04'\(bu\h'+03'\c
543 .\}
544 .el \{\
545 .sp -1
546 .IP \(bu 2.3
547 .\}
548 A 256\-entry fan\-out table just like v1\&.
549 .RE
550 .sp
551 .RS 4
552 .ie n \{\
553 \h'-04'\(bu\h'+03'\c
554 .\}
555 .el \{\
556 .sp -1
557 .IP \(bu 2.3
558 .\}
559 A table of sorted object names\&. These are packed together without offset values to reduce the cache footprint of the binary search for a specific object name\&.
560 .RE
561 .sp
562 .RS 4
563 .ie n \{\
564 \h'-04'\(bu\h'+03'\c
565 .\}
566 .el \{\
567 .sp -1
568 .IP \(bu 2.3
569 .\}
570 A table of 4\-byte CRC32 values of the packed object data\&. This is new in v2 so compressed data can be copied directly from pack to pack during repacking without undetected data corruption\&.
571 .RE
572 .sp
573 .RS 4
574 .ie n \{\
575 \h'-04'\(bu\h'+03'\c
576 .\}
577 .el \{\
578 .sp -1
579 .IP \(bu 2.3
580 .\}
581 A table of 4\-byte offset values (in network byte order)\&. These are usually 31\-bit pack file offsets, but large offsets are encoded as an index into the next table with the msbit set\&.
582 .RE
583 .sp
584 .RS 4
585 .ie n \{\
586 \h'-04'\(bu\h'+03'\c
587 .\}
588 .el \{\
589 .sp -1
590 .IP \(bu 2.3
591 .\}
592 A table of 8\-byte offset entries (empty for pack files less than 2 GiB)\&. Pack files are organized with heavily used objects toward the front, so most object references should not need to refer to this table\&.
593 .RE
594 .sp
595 .RS 4
596 .ie n \{\
597 \h'-04'\(bu\h'+03'\c
598 .\}
599 .el \{\
600 .sp -1
601 .IP \(bu 2.3
602 .\}
603 The same trailer as a v1 pack file:
604 .sp
605 .if n \{\
606 .RS 4
607 .\}
608 .nf
609 A copy of the pack checksum at the end of the
610 corresponding packfile\&.
611 .fi
612 .if n \{\
613 .RE
614 .\}
615 .sp
616 .if n \{\
617 .RS 4
618 .\}
619 .nf
620 Index checksum of all of the above\&.
621 .fi
622 .if n \{\
623 .RE
624 .\}
625 .RE
626 .SH "PACK\-*\&.REV FILES HAVE THE FORMAT:"
627 .sp
628 .RS 4
629 .ie n \{\
630 \h'-04'\(bu\h'+03'\c
631 .\}
632 .el \{\
633 .sp -1
634 .IP \(bu 2.3
635 .\}
636 A 4\-byte magic number
637 \fI0x52494458\fR
638 (\fIRIDX\fR)\&.
639 .RE
640 .sp
641 .RS 4
642 .ie n \{\
643 \h'-04'\(bu\h'+03'\c
644 .\}
645 .el \{\
646 .sp -1
647 .IP \(bu 2.3
648 .\}
649 A 4\-byte version identifier (= 1)\&.
650 .RE
651 .sp
652 .RS 4
653 .ie n \{\
654 \h'-04'\(bu\h'+03'\c
655 .\}
656 .el \{\
657 .sp -1
658 .IP \(bu 2.3
659 .\}
660 A 4\-byte hash function identifier (= 1 for SHA\-1, 2 for SHA\-256)\&.
661 .RE
662 .sp
663 .RS 4
664 .ie n \{\
665 \h'-04'\(bu\h'+03'\c
666 .\}
667 .el \{\
668 .sp -1
669 .IP \(bu 2.3
670 .\}
671 A table of index positions (one per packed object, num_objects in total, each a 4\-byte unsigned integer in network order), sorted by their corresponding offsets in the packfile\&.
672 .RE
673 .sp
674 .RS 4
675 .ie n \{\
676 \h'-04'\(bu\h'+03'\c
677 .\}
678 .el \{\
679 .sp -1
680 .IP \(bu 2.3
681 .\}
682 A trailer, containing a:
683 .sp
684 .if n \{\
685 .RS 4
686 .\}
687 .nf
688 checksum of the corresponding packfile, and
689 .fi
690 .if n \{\
691 .RE
692 .\}
693 .sp
694 .if n \{\
695 .RS 4
696 .\}
697 .nf
698 a checksum of all of the above\&.
699 .fi
700 .if n \{\
701 .RE
702 .\}
703 .RE
704 .sp
705 All 4\-byte numbers are in network order\&.
706 .SH "PACK\-*\&.MTIMES FILES HAVE THE FORMAT:"
707 .sp
708 All 4\-byte numbers are in network byte order\&.
709 .sp
710 .RS 4
711 .ie n \{\
712 \h'-04'\(bu\h'+03'\c
713 .\}
714 .el \{\
715 .sp -1
716 .IP \(bu 2.3
717 .\}
718 A 4\-byte magic number
719 \fI0x4d544d45\fR
720 (\fIMTME\fR)\&.
721 .RE
722 .sp
723 .RS 4
724 .ie n \{\
725 \h'-04'\(bu\h'+03'\c
726 .\}
727 .el \{\
728 .sp -1
729 .IP \(bu 2.3
730 .\}
731 A 4\-byte version identifier (= 1)\&.
732 .RE
733 .sp
734 .RS 4
735 .ie n \{\
736 \h'-04'\(bu\h'+03'\c
737 .\}
738 .el \{\
739 .sp -1
740 .IP \(bu 2.3
741 .\}
742 A 4\-byte hash function identifier (= 1 for SHA\-1, 2 for SHA\-256)\&.
743 .RE
744 .sp
745 .RS 4
746 .ie n \{\
747 \h'-04'\(bu\h'+03'\c
748 .\}
749 .el \{\
750 .sp -1
751 .IP \(bu 2.3
752 .\}
753 A table of 4\-byte unsigned integers\&. The ith value is the modification time (mtime) of the ith object in the corresponding pack by lexicographic (index) order\&. The mtimes count standard epoch seconds\&.
754 .RE
755 .sp
756 .RS 4
757 .ie n \{\
758 \h'-04'\(bu\h'+03'\c
759 .\}
760 .el \{\
761 .sp -1
762 .IP \(bu 2.3
763 .\}
764 A trailer, containing a checksum of the corresponding packfile, and a checksum of all of the above (each having length according to the specified hash function)\&.
765 .RE
766 .SH "MULTI\-PACK\-INDEX (MIDX) FILES HAVE THE FOLLOWING FORMAT:"
767 .sp
768 The multi\-pack\-index files refer to multiple pack\-files and loose objects\&.
769 .sp
770 In order to allow extensions that add extra data to the MIDX, we organize the body into "chunks" and provide a lookup table at the beginning of the body\&. The header includes certain length values, such as the number of packs, the number of base MIDX files, hash lengths and types\&.
771 .sp
772 All 4\-byte numbers are in network order\&.
773 .sp
774 HEADER:
775 .sp
776 .if n \{\
777 .RS 4
778 .\}
779 .nf
780 4\-byte signature:
781 The signature is: {\*(AqM\*(Aq, \*(AqI\*(Aq, \*(AqD\*(Aq, \*(AqX\*(Aq}
782 .fi
783 .if n \{\
784 .RE
785 .\}
786 .sp
787 .if n \{\
788 .RS 4
789 .\}
790 .nf
791 1\-byte version number:
792 Git only writes or recognizes version 1\&.
793 .fi
794 .if n \{\
795 .RE
796 .\}
797 .sp
798 .if n \{\
799 .RS 4
800 .\}
801 .nf
802 1\-byte Object Id Version
803 We infer the length of object IDs (OIDs) from this value:
804 1 => SHA\-1
805 2 => SHA\-256
806 If the hash type does not match the repository\*(Aqs hash algorithm,
807 the multi\-pack\-index file should be ignored with a warning
808 presented to the user\&.
809 .fi
810 .if n \{\
811 .RE
812 .\}
813 .sp
814 .if n \{\
815 .RS 4
816 .\}
817 .nf
818 1\-byte number of "chunks"
819 .fi
820 .if n \{\
821 .RE
822 .\}
823 .sp
824 .if n \{\
825 .RS 4
826 .\}
827 .nf
828 1\-byte number of base multi\-pack\-index files:
829 This value is currently always zero\&.
830 .fi
831 .if n \{\
832 .RE
833 .\}
834 .sp
835 .if n \{\
836 .RS 4
837 .\}
838 .nf
839 4\-byte number of pack files
840 .fi
841 .if n \{\
842 .RE
843 .\}
844 .sp
845 CHUNK LOOKUP:
846 .sp
847 .if n \{\
848 .RS 4
849 .\}
850 .nf
851 (C + 1) * 12 bytes providing the chunk offsets:
852 First 4 bytes describe chunk id\&. Value 0 is a terminating label\&.
853 Other 8 bytes provide offset in current file for chunk to start\&.
854 (Chunks are provided in file\-order, so you can infer the length
855 using the next chunk position if necessary\&.)
856 .fi
857 .if n \{\
858 .RE
859 .\}
860 .sp
861 .if n \{\
862 .RS 4
863 .\}
864 .nf
865 The CHUNK LOOKUP matches the table of contents from
866 the chunk\-based file format, see linkgit:gitformat\-chunk[5]\&.
867 .fi
868 .if n \{\
869 .RE
870 .\}
871 .sp
872 .if n \{\
873 .RS 4
874 .\}
875 .nf
876 The remaining data in the body is described one chunk at a time, and
877 these chunks may be given in any order\&. Chunks are required unless
878 otherwise specified\&.
879 .fi
880 .if n \{\
881 .RE
882 .\}
883 .sp
884 CHUNK DATA:
885 .sp
886 .if n \{\
887 .RS 4
888 .\}
889 .nf
890 Packfile Names (ID: {\*(AqP\*(Aq, \*(AqN\*(Aq, \*(AqA\*(Aq, \*(AqM\*(Aq})
891 Store the names of packfiles as a sequence of NUL\-terminated
892 strings\&. There is no extra padding between the filenames,
893 and they are listed in lexicographic order\&. The chunk itself
894 is padded at the end with between 0 and 3 NUL bytes to make the
895 chunk size a multiple of 4 bytes\&.
896 .fi
897 .if n \{\
898 .RE
899 .\}
900 .sp
901 .if n \{\
902 .RS 4
903 .\}
904 .nf
905 Bitmapped Packfiles (ID: {\*(AqB\*(Aq, \*(AqT\*(Aq, \*(AqM\*(Aq, \*(AqP\*(Aq})
906 Stores a table of two 4\-byte unsigned integers in network order\&.
907 Each table entry corresponds to a single pack (in the order that
908 they appear above in the `PNAM` chunk)\&. The values for each table
909 entry are as follows:
910 \- The first bit position (in pseudo\-pack order, see below) to
911 contain an object from that pack\&.
912 \- The number of bits whose objects are selected from that pack\&.
913 .fi
914 .if n \{\
915 .RE
916 .\}
917 .sp
918 .if n \{\
919 .RS 4
920 .\}
921 .nf
922 OID Fanout (ID: {\*(AqO\*(Aq, \*(AqI\*(Aq, \*(AqD\*(Aq, \*(AqF\*(Aq})
923 The ith entry, F[i], stores the number of OIDs with first
924 byte at most i\&. Thus F[255] stores the total
925 number of objects\&.
926 .fi
927 .if n \{\
928 .RE
929 .\}
930 .sp
931 .if n \{\
932 .RS 4
933 .\}
934 .nf
935 OID Lookup (ID: {\*(AqO\*(Aq, \*(AqI\*(Aq, \*(AqD\*(Aq, \*(AqL\*(Aq})
936 The OIDs for all objects in the MIDX are stored in lexicographic
937 order in this chunk\&.
938 .fi
939 .if n \{\
940 .RE
941 .\}
942 .sp
943 .if n \{\
944 .RS 4
945 .\}
946 .nf
947 Object Offsets (ID: {\*(AqO\*(Aq, \*(AqO\*(Aq, \*(AqF\*(Aq, \*(AqF\*(Aq})
948 Stores two 4\-byte values for every object\&.
949 1: The pack\-int\-id for the pack storing this object\&.
950 2: The offset within the pack\&.
951 If all offsets are less than 2^32, then the large offset chunk
952 will not exist and offsets are stored as in IDX v1\&.
953 If there is at least one offset value larger than 2^32\-1, then
954 the large offset chunk must exist, and offsets larger than
955 2^31\-1 must be stored in it instead\&. If the large offset chunk
956 exists and the 31st bit is on, then removing that bit reveals
957 the row in the large offsets containing the 8\-byte offset of
958 this object\&.
959 .fi
960 .if n \{\
961 .RE
962 .\}
963 .sp
964 .if n \{\
965 .RS 4
966 .\}
967 .nf
968 [Optional] Object Large Offsets (ID: {\*(AqL\*(Aq, \*(AqO\*(Aq, \*(AqF\*(Aq, \*(AqF\*(Aq})
969 8\-byte offsets into large packfiles\&.
970 .fi
971 .if n \{\
972 .RE
973 .\}
974 .sp
975 .if n \{\
976 .RS 4
977 .\}
978 .nf
979 [Optional] Bitmap pack order (ID: {\*(AqR\*(Aq, \*(AqI\*(Aq, \*(AqD\*(Aq, \*(AqX\*(Aq})
980 A list of MIDX positions (one per object in the MIDX, num_objects in
981 total, each a 4\-byte unsigned integer in network byte order), sorted
982 according to their relative bitmap/pseudo\-pack positions\&.
983 .fi
984 .if n \{\
985 .RE
986 .\}
987 .sp
988 TRAILER:
989 .sp
990 .if n \{\
991 .RS 4
992 .\}
993 .nf
994 Index checksum of the above contents\&.
995 .fi
996 .if n \{\
997 .RE
998 .\}
999 .SH "MULTI\-PACK\-INDEX REVERSE INDEXES"
1000 .sp
1001 Similar to the pack\-based reverse index, the multi\-pack index can also be used to generate a reverse index\&.
1002 .sp
1003 Instead of mapping between offset, pack\-, and index position, this reverse index maps between an object\(cqs position within the MIDX, and that object\(cqs position within a pseudo\-pack that the MIDX describes (i\&.e\&., the ith entry of the multi\-pack reverse index holds the MIDX position of ith object in pseudo\-pack order)\&.
1004 .sp
1005 To clarify the difference between these orderings, consider a multi\-pack reachability bitmap (which does not yet exist, but is what we are building towards here)\&. Each bit needs to correspond to an object in the MIDX, and so we need an efficient mapping from bit position to MIDX position\&.
1006 .sp
1007 One solution is to let bits occupy the same position in the oid\-sorted index stored by the MIDX\&. But because oids are effectively random, their resulting reachability bitmaps would have no locality, and thus compress poorly\&. (This is the reason that single\-pack bitmaps use the pack ordering, and not the \&.idx ordering, for the same purpose\&.)
1008 .sp
1009 So we\(cqd like to define an ordering for the whole MIDX based around pack ordering, which has far better locality (and thus compresses more efficiently)\&. We can think of a pseudo\-pack created by the concatenation of all of the packs in the MIDX\&. E\&.g\&., if we had a MIDX with three packs (a, b, c), with 10, 15, and 20 objects respectively, we can imagine an ordering of the objects like:
1010 .sp
1011 .if n \{\
1012 .RS 4
1013 .\}
1014 .nf
1015 |a,0|a,1|\&.\&.\&.|a,9|b,0|b,1|\&.\&.\&.|b,14|c,0|c,1|\&.\&.\&.|c,19|
1016 .fi
1017 .if n \{\
1018 .RE
1019 .\}
1020 .sp
1021 where the ordering of the packs is defined by the MIDX\(cqs pack list, and then the ordering of objects within each pack is the same as the order in the actual packfile\&.
1022 .sp
1023 Given the list of packs and their counts of objects, you can na\(:ively reconstruct that pseudo\-pack ordering (e\&.g\&., the object at position 27 must be (c,1) because packs "a" and "b" consumed 25 of the slots)\&. But there\(cqs a catch\&. Objects may be duplicated between packs, in which case the MIDX only stores one pointer to the object (and thus we\(cqd want only one slot in the bitmap)\&.
1024 .sp
1025 Callers could handle duplicates themselves by reading objects in order of their bit\-position, but that\(cqs linear in the number of objects, and much too expensive for ordinary bitmap lookups\&. Building a reverse index solves this, since it is the logical inverse of the index, and that index has already removed duplicates\&. But, building a reverse index on the fly can be expensive\&. Since we already have an on\-disk format for pack\-based reverse indexes, let\(cqs reuse it for the MIDX\(cqs pseudo\-pack, too\&.
1026 .sp
1027 Objects from the MIDX are ordered as follows to string together the pseudo\-pack\&. Let \fBpack(o)\fR return the pack from which \fBo\fR was selected by the MIDX, and define an ordering of packs based on their numeric ID (as stored by the MIDX)\&. Let \fBoffset(o)\fR return the object offset of \fBo\fR within \fBpack(o)\fR\&. Then, compare \fBo1\fR and \fBo2\fR as follows:
1028 .sp
1029 .RS 4
1030 .ie n \{\
1031 \h'-04'\(bu\h'+03'\c
1032 .\}
1033 .el \{\
1034 .sp -1
1035 .IP \(bu 2.3
1036 .\}
1037 If one of
1038 \fBpack(o1)\fR
1039 and
1040 \fBpack(o2)\fR
1041 is preferred and the other is not, then the preferred one sorts first\&.
1042 .sp
1043 (This is a detail that allows the MIDX bitmap to determine which pack should be used by the pack\-reuse mechanism, since it can ask the MIDX for the pack containing the object at bit position 0)\&.
1044 .RE
1045 .sp
1046 .RS 4
1047 .ie n \{\
1048 \h'-04'\(bu\h'+03'\c
1049 .\}
1050 .el \{\
1051 .sp -1
1052 .IP \(bu 2.3
1053 .\}
1054 If
1055 \fBpack(o1) \(!= pack(o2)\fR, then sort the two objects in descending order based on the pack ID\&.
1056 .RE
1057 .sp
1058 .RS 4
1059 .ie n \{\
1060 \h'-04'\(bu\h'+03'\c
1061 .\}
1062 .el \{\
1063 .sp -1
1064 .IP \(bu 2.3
1065 .\}
1066 Otherwise,
1067 \fBpack(o1) = pack(o2)\fR, and the objects are sorted in pack\-order (i\&.e\&.,
1068 \fBo1\fR
1069 sorts ahead of
1070 \fBo2\fR
1071 exactly when
1072 \fBoffset(o1) < offset(o2)\fR)\&.
1073 .RE
1074 .sp
1075 In short, a MIDX\(cqs pseudo\-pack is the de\-duplicated concatenation of objects in packs stored by the MIDX, laid out in pack order, and the packs arranged in MIDX order (with the preferred pack coming first)\&.
1076 .sp
1077 The MIDX\(cqs reverse index is stored in the optional \fIRIDX\fR chunk within the MIDX itself\&.
1078 .SS "\fBBTMP\fR chunk"
1079 .sp
1080 The Bitmapped Packfiles (\fBBTMP\fR) chunk encodes additional information about the objects in the multi\-pack index\(cqs reachability bitmap\&. Recall that objects from the MIDX are arranged in "pseudo\-pack" order (see above) for reachability bitmaps\&.
1081 .sp
1082 From the example above, suppose we have packs "a", "b", and "c", with 10, 15, and 20 objects, respectively\&. In pseudo\-pack order, those would be arranged as follows:
1083 .sp
1084 .if n \{\
1085 .RS 4
1086 .\}
1087 .nf
1088 |a,0|a,1|\&.\&.\&.|a,9|b,0|b,1|\&.\&.\&.|b,14|c,0|c,1|\&.\&.\&.|c,19|
1089 .fi
1090 .if n \{\
1091 .RE
1092 .\}
1093 .sp
1094 When working with single\-pack bitmaps (or, equivalently, multi\-pack reachability bitmaps with a preferred pack), \fBgit-pack-objects\fR(1) performs \(lqverbatim\(rq reuse, attempting to reuse chunks of the bitmapped or preferred packfile instead of adding objects to the packing list\&.
1095 .sp
1096 When a chunk of bytes is reused from an existing pack, any objects contained therein do not need to be added to the packing list, saving memory and CPU time\&. But a chunk from an existing packfile can only be reused when the following conditions are met:
1097 .sp
1098 .RS 4
1099 .ie n \{\
1100 \h'-04'\(bu\h'+03'\c
1101 .\}
1102 .el \{\
1103 .sp -1
1104 .IP \(bu 2.3
1105 .\}
1106 The chunk contains only objects which were requested by the caller (i\&.e\&. does not contain any objects which the caller didn\(cqt ask for explicitly or implicitly)\&.
1107 .RE
1108 .sp
1109 .RS 4
1110 .ie n \{\
1111 \h'-04'\(bu\h'+03'\c
1112 .\}
1113 .el \{\
1114 .sp -1
1115 .IP \(bu 2.3
1116 .\}
1117 All objects stored in non\-thin packs as offset\- or reference\-deltas also include their base object in the resulting pack\&.
1118 .RE
1119 .sp
1120 The \fBBTMP\fR chunk encodes the necessary information in order to implement multi\-pack reuse over a set of packfiles as described above\&. Specifically, the \fBBTMP\fR chunk encodes three pieces of information (all 32\-bit unsigned integers in network byte\-order) for each packfile \fBp\fR that is stored in the MIDX, as follows:
1121 .PP
1122 \fBbitmap_pos\fR
1123 .RS 4
1124 The first bit position (in pseudo\-pack order) in the multi\-pack index\(cqs reachability bitmap occupied by an object from
1125 \fBp\fR\&.
1126 .RE
1127 .PP
1128 \fBbitmap_nr\fR
1129 .RS 4
1130 The number of bit positions (including the one at
1131 \fBbitmap_pos\fR) that encode objects from that pack
1132 \fBp\fR\&.
1133 .RE
1134 .sp
1135 For example, the \fBBTMP\fR chunk corresponding to the above example (with packs \(lqa\(rq, \(lqb\(rq, and \(lqc\(rq) would look like:
1136 .TS
1137 allbox tab(:);
1138 ltB ltB ltB.
1139 T{
1140 \ \&
1141 T}:T{
1142 \fBbitmap_pos\fR
1143 T}:T{
1144 \fBbitmap_nr\fR
1145 T}
1146 .T&
1147 lt lt lt
1148 lt lt lt
1149 lt lt lt.
1150 T{
1151 .sp
1152 packfile \(lqa\(rq
1153 T}:T{
1154 .sp
1155 \fB0\fR
1156 T}:T{
1157 .sp
1158 \fB10\fR
1159 T}
1160 T{
1161 .sp
1162 packfile \(lqb\(rq
1163 T}:T{
1164 .sp
1165 \fB10\fR
1166 T}:T{
1167 .sp
1168 \fB15\fR
1169 T}
1170 T{
1171 .sp
1172 packfile \(lqc\(rq
1173 T}:T{
1174 .sp
1175 \fB25\fR
1176 T}:T{
1177 .sp
1178 \fB20\fR
1179 T}
1180 .TE
1181 .sp 1
1182 .sp
1183 With this information in place, we can treat each packfile as individually reusable in the same fashion as verbatim pack reuse is performed on individual packs prior to the implementation of the \fBBTMP\fR chunk\&.
1184 .SH "CRUFT PACKS"
1185 .sp
1186 The cruft packs feature offer an alternative to Git\(cqs traditional mechanism of removing unreachable objects\&. This document provides an overview of Git\(cqs pruning mechanism, and how a cruft pack can be used instead to accomplish the same\&.
1187 .SS "Background"
1188 .sp
1189 To remove unreachable objects from your repository, Git offers \fBgit repack \-Ad\fR (see \fBgit-repack\fR(1))\&. Quoting from the documentation:
1190 .sp
1191 .if n \{\
1192 .RS 4
1193 .\}
1194 .nf
1195 [\&.\&.\&.] unreachable objects in a previous pack become loose, unpacked objects,
1196 instead of being left in the old pack\&. [\&.\&.\&.] loose unreachable objects will be
1197 pruned according to normal expiry rules with the next \*(Aqgit gc\*(Aq invocation\&.
1198 .fi
1199 .if n \{\
1200 .RE
1201 .\}
1202 .sp
1203 Unreachable objects aren\(cqt removed immediately, since doing so could race with an incoming push which may reference an object which is about to be deleted\&. Instead, those unreachable objects are stored as loose objects and stay that way until they are older than the expiration window, at which point they are removed by \fBgit-prune\fR(1)\&.
1204 .sp
1205 Git must store these unreachable objects loose in order to keep track of their per\-object mtimes\&. If these unreachable objects were written into one big pack, then either freshening that pack (because an object contained within it was re\-written) or creating a new pack of unreachable objects would cause the pack\(cqs mtime to get updated, and the objects within it would never leave the expiration window\&. Instead, objects are stored loose in order to keep track of the individual object mtimes and avoid a situation where all cruft objects are freshened at once\&.
1206 .sp
1207 This can lead to undesirable situations when a repository contains many unreachable objects which have not yet left the grace period\&. Having large directories in the shards of \fB\&.git/objects\fR can lead to decreased performance in the repository\&. But given enough unreachable objects, this can lead to inode starvation and degrade the performance of the whole system\&. Since we can never pack those objects, these repositories often take up a large amount of disk space, since we can only zlib compress them, but not store them in delta chains\&.
1208 .SS "Cruft packs"
1209 .sp
1210 A cruft pack eliminates the need for storing unreachable objects in a loose state by including the per\-object mtimes in a separate file alongside a single pack containing all loose objects\&.
1211 .sp
1212 A cruft pack is written by \fBgit repack \-\-cruft\fR when generating a new pack\&. \fBgit-pack-objects\fR(1)\*(Aqs \fB\-\-cruft\fR option\&. Note that \fBgit repack \-\-cruft\fR is a classic all\-into\-one repack, meaning that everything in the resulting pack is reachable, and everything else is unreachable\&. Once written, the \fB\-\-cruft\fR option instructs \fBgit repack\fR to generate another pack containing only objects not packed in the previous step (which equates to packing all unreachable objects together)\&. This progresses as follows:
1213 .sp
1214 .RS 4
1215 .ie n \{\
1216 \h'-04' 1.\h'+01'\c
1217 .\}
1218 .el \{\
1219 .sp -1
1220 .IP " 1." 4.2
1221 .\}
1222 Enumerate every object, marking any object which is (a) not contained in a kept\-pack, and (b) whose mtime is within the grace period as a traversal tip\&.
1223 .RE
1224 .sp
1225 .RS 4
1226 .ie n \{\
1227 \h'-04' 2.\h'+01'\c
1228 .\}
1229 .el \{\
1230 .sp -1
1231 .IP " 2." 4.2
1232 .\}
1233 Perform a reachability traversal based on the tips gathered in the previous step, adding every object along the way to the pack\&.
1234 .RE
1235 .sp
1236 .RS 4
1237 .ie n \{\
1238 \h'-04' 3.\h'+01'\c
1239 .\}
1240 .el \{\
1241 .sp -1
1242 .IP " 3." 4.2
1243 .\}
1244 Write the pack out, along with a
1245 \fB\&.mtimes\fR
1246 file that records the per\-object timestamps\&.
1247 .RE
1248 .sp
1249 This mode is invoked internally by \fBgit-repack\fR(1) when instructed to write a cruft pack\&. Crucially, the set of in\-core kept packs is exactly the set of packs which will not be deleted by the repack; in other words, they contain all of the repository\(cqs reachable objects\&.
1250 .sp
1251 When a repository already has a cruft pack, \fBgit repack \-\-cruft\fR typically only adds objects to it\&. An exception to this is when \fBgit repack\fR is given the \fB\-\-cruft\-expiration\fR option, which allows the generated cruft pack to omit expired objects instead of waiting for \fBgit-gc\fR(1) to expire those objects later on\&.
1252 .sp
1253 It is \fBgit-gc\fR(1) that is typically responsible for removing expired unreachable objects\&.
1254 .SS "Alternatives"
1255 .sp
1256 Notable alternatives to this design include:
1257 .sp
1258 .RS 4
1259 .ie n \{\
1260 \h'-04'\(bu\h'+03'\c
1261 .\}
1262 .el \{\
1263 .sp -1
1264 .IP \(bu 2.3
1265 .\}
1266 The location of the per\-object mtime data\&.
1267 .RE
1268 .sp
1269 On the location of mtime data, a new auxiliary file tied to the pack was chosen to avoid complicating the \fB\&.idx\fR format\&. If the \fB\&.idx\fR format were ever to gain support for optional chunks of data, it may make sense to consolidate the \fB\&.mtimes\fR format into the \fB\&.idx\fR itself\&.
1270 .SH "GIT"
1271 .sp
1272 Part of the \fBgit\fR(1) suite