1GIT-PACK-OBJECTS(1) Git Manual GIT-PACK-OBJECTS(1)
2
3
4
6 git-pack-objects - Create a packed archive of objects
7
9 git pack-objects [-q | --progress | --all-progress] [--all-progress-implied]
10 [--no-reuse-delta] [--delta-base-offset] [--non-empty]
11 [--local] [--incremental] [--window=<n>] [--depth=<n>]
12 [--revs [--unpacked | --all]] [--keep-pack=<pack-name>]
13 [--stdout [--filter=<filter-spec>] | base-name]
14 [--shallow] [--keep-true-parents] [--sparse] < object-list
15
16
18 Reads list of objects from the standard input, and writes either one or
19 more packed archives with the specified base-name to disk, or a packed
20 archive to the standard output.
21
22 A packed archive is an efficient way to transfer a set of objects
23 between two repositories as well as an access efficient archival
24 format. In a packed archive, an object is either stored as a compressed
25 whole or as a difference from some other object. The latter is often
26 called a delta.
27
28 The packed archive format (.pack) is designed to be self-contained so
29 that it can be unpacked without any further information. Therefore,
30 each object that a delta depends upon must be present within the pack.
31
32 A pack index file (.idx) is generated for fast, random access to the
33 objects in the pack. Placing both the index file (.idx) and the packed
34 archive (.pack) in the pack/ subdirectory of $GIT_OBJECT_DIRECTORY (or
35 any of the directories on $GIT_ALTERNATE_OBJECT_DIRECTORIES) enables
36 Git to read from the pack archive.
37
38 The git unpack-objects command can read the packed archive and expand
39 the objects contained in the pack into "one-file one-object" format;
40 this is typically done by the smart-pull commands when a pack is
41 created on-the-fly for efficient network transport by their peers.
42
44 base-name
45 Write into pairs of files (.pack and .idx), using <base-name> to
46 determine the name of the created file. When this option is used,
47 the two files in a pair are written in
48 <base-name>-<SHA-1>.{pack,idx} files. <SHA-1> is a hash based on
49 the pack content and is written to the standard output of the
50 command.
51
52 --stdout
53 Write the pack contents (what would have been written to .pack
54 file) out to the standard output.
55
56 --revs
57 Read the revision arguments from the standard input, instead of
58 individual object names. The revision arguments are processed the
59 same way as git rev-list with the --objects flag uses its commit
60 arguments to build the list of objects it outputs. The objects on
61 the resulting list are packed. Besides revisions, --not or
62 --shallow <SHA-1> lines are also accepted.
63
64 --unpacked
65 This implies --revs. When processing the list of revision arguments
66 read from the standard input, limit the objects packed to those
67 that are not already packed.
68
69 --all
70 This implies --revs. In addition to the list of revision arguments
71 read from the standard input, pretend as if all refs under refs/
72 are specified to be included.
73
74 --include-tag
75 Include unasked-for annotated tags if the object they reference was
76 included in the resulting packfile. This can be useful to send new
77 tags to native Git clients.
78
79 --window=<n>, --depth=<n>
80 These two options affect how the objects contained in the pack are
81 stored using delta compression. The objects are first internally
82 sorted by type, size and optionally names and compared against the
83 other objects within --window to see if using delta compression
84 saves space. --depth limits the maximum delta depth; making it too
85 deep affects the performance on the unpacker side, because delta
86 data needs to be applied that many times to get to the necessary
87 object.
88
89 The default value for --window is 10 and --depth is 50. The maximum
90 depth is 4095.
91
92 --window-memory=<n>
93 This option provides an additional limit on top of --window; the
94 window size will dynamically scale down so as to not take up more
95 than <n> bytes in memory. This is useful in repositories with a mix
96 of large and small objects to not run out of memory with a large
97 window, but still be able to take advantage of the large window for
98 the smaller objects. The size can be suffixed with "k", "m", or
99 "g". --window-memory=0 makes memory usage unlimited. The default
100 is taken from the pack.windowMemory configuration variable.
101
102 --max-pack-size=<n>
103 In unusual scenarios, you may not be able to create files larger
104 than a certain size on your filesystem, and this option can be used
105 to tell the command to split the output packfile into multiple
106 independent packfiles, each not larger than the given size. The
107 size can be suffixed with "k", "m", or "g". The minimum size
108 allowed is limited to 1 MiB. This option prevents the creation of a
109 bitmap index. The default is unlimited, unless the config variable
110 pack.packSizeLimit is set.
111
112 --honor-pack-keep
113 This flag causes an object already in a local pack that has a .keep
114 file to be ignored, even if it would have otherwise been packed.
115
116 --keep-pack=<pack-name>
117 This flag causes an object already in the given pack to be ignored,
118 even if it would have otherwise been packed. <pack-name> is the
119 pack file name without leading directory (e.g. pack-123.pack). The
120 option could be specified multiple times to keep multiple packs.
121
122 --incremental
123 This flag causes an object already in a pack to be ignored even if
124 it would have otherwise been packed.
125
126 --local
127 This flag causes an object that is borrowed from an alternate
128 object store to be ignored even if it would have otherwise been
129 packed.
130
131 --non-empty
132 Only create a packed archive if it would contain at least one
133 object.
134
135 --progress
136 Progress status is reported on the standard error stream by default
137 when it is attached to a terminal, unless -q is specified. This
138 flag forces progress status even if the standard error stream is
139 not directed to a terminal.
140
141 --all-progress
142 When --stdout is specified then progress report is displayed during
143 the object count and compression phases but inhibited during the
144 write-out phase. The reason is that in some cases the output stream
145 is directly linked to another command which may wish to display
146 progress status of its own as it processes incoming pack data. This
147 flag is like --progress except that it forces progress report for
148 the write-out phase as well even if --stdout is used.
149
150 --all-progress-implied
151 This is used to imply --all-progress whenever progress display is
152 activated. Unlike --all-progress this flag doesn’t actually force
153 any progress display by itself.
154
155 -q
156 This flag makes the command not to report its progress on the
157 standard error stream.
158
159 --no-reuse-delta
160 When creating a packed archive in a repository that has existing
161 packs, the command reuses existing deltas. This sometimes results
162 in a slightly suboptimal pack. This flag tells the command not to
163 reuse existing deltas but compute them from scratch.
164
165 --no-reuse-object
166 This flag tells the command not to reuse existing object data at
167 all, including non deltified object, forcing recompression of
168 everything. This implies --no-reuse-delta. Useful only in the
169 obscure case where wholesale enforcement of a different compression
170 level on the packed data is desired.
171
172 --compression=<n>
173 Specifies compression level for newly-compressed data in the
174 generated pack. If not specified, pack compression level is
175 determined first by pack.compression, then by core.compression, and
176 defaults to -1, the zlib default, if neither is set. Add
177 --no-reuse-object if you want to force a uniform compression level
178 on all data no matter the source.
179
180 --sparse
181 Use the "sparse" algorithm to determine which objects to include in
182 the pack, when combined with the "--revs" option. This algorithm
183 only walks trees that appear in paths that introduce new objects.
184 This can have significant performance benefits when computing a
185 pack to send a small change. However, it is possible that extra
186 objects are added to the pack-file if the included commits contain
187 certain types of direct renames.
188
189 --thin
190 Create a "thin" pack by omitting the common objects between a
191 sender and a receiver in order to reduce network transfer. This
192 option only makes sense in conjunction with --stdout.
193
194 Note: A thin pack violates the packed archive format by omitting
195 required objects and is thus unusable by Git without making it
196 self-contained. Use git index-pack --fix-thin (see git-index-
197 pack(1)) to restore the self-contained property.
198
199 --shallow
200 Optimize a pack that will be provided to a client with a shallow
201 repository. This option, combined with --thin, can result in a
202 smaller pack at the cost of speed.
203
204 --delta-base-offset
205 A packed archive can express the base object of a delta as either a
206 20-byte object name or as an offset in the stream, but ancient
207 versions of Git don’t understand the latter. By default, git
208 pack-objects only uses the former format for better compatibility.
209 This option allows the command to use the latter format for
210 compactness. Depending on the average delta chain length, this
211 option typically shrinks the resulting packfile by 3-5 per-cent.
212
213 Note: Porcelain commands such as git gc (see git-gc(1)), git repack
214 (see git-repack(1)) pass this option by default in modern Git when
215 they put objects in your repository into pack files. So does git
216 bundle (see git-bundle(1)) when it creates a bundle.
217
218 --threads=<n>
219 Specifies the number of threads to spawn when searching for best
220 delta matches. This requires that pack-objects be compiled with
221 pthreads otherwise this option is ignored with a warning. This is
222 meant to reduce packing time on multiprocessor machines. The
223 required amount of memory for the delta search window is however
224 multiplied by the number of threads. Specifying 0 will cause Git to
225 auto-detect the number of CPU’s and set the number of threads
226 accordingly.
227
228 --index-version=<version>[,<offset>]
229 This is intended to be used by the test suite only. It allows to
230 force the version for the generated pack index, and to force 64-bit
231 index entries on objects located above the given offset.
232
233 --keep-true-parents
234 With this option, parents that are hidden by grafts are packed
235 nevertheless.
236
237 --filter=<filter-spec>
238 Requires --stdout. Omits certain objects (usually blobs) from the
239 resulting packfile. See git-rev-list(1) for valid <filter-spec>
240 forms.
241
242 --no-filter
243 Turns off any previous --filter= argument.
244
245 --missing=<missing-action>
246 A debug option to help with future "partial clone" development.
247 This option specifies how missing objects are handled.
248
249 The form --missing=error requests that pack-objects stop with an
250 error if a missing object is encountered. This is the default
251 action.
252
253 The form --missing=allow-any will allow object traversal to
254 continue if a missing object is encountered. Missing objects will
255 silently be omitted from the results.
256
257 The form --missing=allow-promisor is like allow-any, but will only
258 allow object traversal to continue for EXPECTED promisor missing
259 objects. Unexpected missing object will raise an error.
260
261 --exclude-promisor-objects
262 Omit objects that are known to be in the promisor remote. (This
263 option has the purpose of operating only on locally created
264 objects, so that when we repack, we still maintain a distinction
265 between locally created objects [without .promisor] and objects
266 from the promisor remote [with .promisor].) This is used with
267 partial clone.
268
269 --keep-unreachable
270 Objects unreachable from the refs in packs named with --unpacked=
271 option are added to the resulting pack, in addition to the
272 reachable objects that are not in packs marked with *.keep files.
273 This implies --revs.
274
275 --pack-loose-unreachable
276 Pack unreachable loose objects (and their loose counterparts
277 removed). This implies --revs.
278
279 --unpack-unreachable
280 Keep unreachable objects in loose form. This implies --revs.
281
282 --delta-islands
283 Restrict delta matches based on "islands". See DELTA ISLANDS below.
284
286 When possible, pack-objects tries to reuse existing on-disk deltas to
287 avoid having to search for new ones on the fly. This is an important
288 optimization for serving fetches, because it means the server can avoid
289 inflating most objects at all and just send the bytes directly from
290 disk. This optimization can’t work when an object is stored as a delta
291 against a base which the receiver does not have (and which we are not
292 already sending). In that case the server "breaks" the delta and has to
293 find a new one, which has a high CPU cost. Therefore it’s important for
294 performance that the set of objects in on-disk delta relationships
295 match what a client would fetch.
296
297 In a normal repository, this tends to work automatically. The objects
298 are mostly reachable from the branches and tags, and that’s what
299 clients fetch. Any deltas we find on the server are likely to be
300 between objects the client has or will have.
301
302 But in some repository setups, you may have several related but
303 separate groups of ref tips, with clients tending to fetch those groups
304 independently. For example, imagine that you are hosting several
305 "forks" of a repository in a single shared object store, and letting
306 clients view them as separate repositories through GIT_NAMESPACE or
307 separate repos using the alternates mechanism. A naive repack may find
308 that the optimal delta for an object is against a base that is only
309 found in another fork. But when a client fetches, they will not have
310 the base object, and we’ll have to find a new delta on the fly.
311
312 A similar situation may exist if you have many refs outside of
313 refs/heads/ and refs/tags/ that point to related objects (e.g.,
314 refs/pull or refs/changes used by some hosting providers). By default,
315 clients fetch only heads and tags, and deltas against objects found
316 only in those other groups cannot be sent as-is.
317
318 Delta islands solve this problem by allowing you to group your refs
319 into distinct "islands". Pack-objects computes which objects are
320 reachable from which islands, and refuses to make a delta from an
321 object A against a base which is not present in all of A's islands.
322 This results in slightly larger packs (because we miss some delta
323 opportunities), but guarantees that a fetch of one island will not have
324 to recompute deltas on the fly due to crossing island boundaries.
325
326 When repacking with delta islands the delta window tends to get clogged
327 with candidates that are forbidden by the config. Repacking with a big
328 --window helps (and doesn’t take as long as it otherwise might because
329 we can reject some object pairs based on islands before doing any
330 computation on the content).
331
332 Islands are configured via the pack.island option, which can be
333 specified multiple times. Each value is a left-anchored regular
334 expressions matching refnames. For example:
335
336 [pack]
337 island = refs/heads/
338 island = refs/tags/
339
340
341 puts heads and tags into an island (whose name is the empty string; see
342 below for more on naming). Any refs which do not match those regular
343 expressions (e.g., refs/pull/123) is not in any island. Any object
344 which is reachable only from refs/pull/ (but not heads or tags) is
345 therefore not a candidate to be used as a base for refs/heads/.
346
347 Refs are grouped into islands based on their "names", and two regexes
348 that produce the same name are considered to be in the same island. The
349 names are computed from the regexes by concatenating any capture groups
350 from the regex, with a - dash in between. (And if there are no capture
351 groups, then the name is the empty string, as in the above example.)
352 This allows you to create arbitrary numbers of islands. Only up to 14
353 such capture groups are supported though.
354
355 For example, imagine you store the refs for each fork in
356 refs/virtual/ID, where ID is a numeric identifier. You might then
357 configure:
358
359 [pack]
360 island = refs/virtual/([0-9]+)/heads/
361 island = refs/virtual/([0-9]+)/tags/
362 island = refs/virtual/([0-9]+)/(pull)/
363
364
365 That puts the heads and tags for each fork in their own island (named
366 "1234" or similar), and the pull refs for each go into their own
367 "1234-pull".
368
369 Note that we pick a single island for each regex to go into, using
370 "last one wins" ordering (which allows repo-specific config to take
371 precedence over user-wide config, and so forth).
372
374 git-rev-list(1) git-repack(1) git-prune-packed(1)
375
377 Part of the git(1) suite
378
379
380
381Git 2.24.1 12/10/2019 GIT-PACK-OBJECTS(1)