| blob | 19cdcd03417dfa7ff7c744686466af733c367945 |
1 git-pack-objects(1)
2 ===================
4 NAME
5 ----
6 git-pack-objects - Create a packed archive of objects
9 SYNOPSIS
10 --------
11 [verse]
12 'git pack-objects' [-q | --progress | --all-progress] [--all-progress-implied]
13 [--no-reuse-delta] [--delta-base-offset] [--non-empty]
14 [--local] [--incremental] [--window=<n>] [--depth=<n>]
15 [--revs [--unpacked | --all]] [--stdout | base-name]
16 [--shallow] [--keep-true-parents] < object-list
19 DESCRIPTION
20 -----------
21 Reads list of objects from the standard input, and writes a packed
22 archive with specified base-name, or to the standard output.
24 A packed archive is an efficient way to transfer a set of objects
25 between two repositories as well as an access efficient archival
26 format. In a packed archive, an object is either stored as a
27 compressed whole or as a difference from some other object.
28 The latter is often called a delta.
30 The packed archive format (.pack) is designed to be self-contained
31 so that it can be unpacked without any further information. Therefore,
32 each object that a delta depends upon must be present within the pack.
34 A pack index file (.idx) is generated for fast, random access to the
35 objects in the pack. Placing both the index file (.idx) and the packed
36 archive (.pack) in the pack/ subdirectory of $GIT_OBJECT_DIRECTORY (or
37 any of the directories on $GIT_ALTERNATE_OBJECT_DIRECTORIES)
38 enables Git to read from the pack archive.
40 The 'git unpack-objects' command can read the packed archive and
41 expand the objects contained in the pack into "one-file
42 one-object" format; this is typically done by the smart-pull
43 commands when a pack is created on-the-fly for efficient network
44 transport by their peers.
47 OPTIONS
48 -------
49 base-name::
50 Write into a pair of files (.pack and .idx), using
51 <base-name> to determine the name of the created file.
52 When this option is used, the two files are written in
53 <base-name>-<SHA-1>.{pack,idx} files. <SHA-1> is a hash
54 based on the pack content and is written to the standard
55 output of the command.
57 --stdout::
58 Write the pack contents (what would have been written to
59 .pack file) out to the standard output.
61 --revs::
62 Read the revision arguments from the standard input, instead of
63 individual object names. The revision arguments are processed
64 the same way as 'git rev-list' with the `--objects` flag
65 uses its `commit` arguments to build the list of objects it
66 outputs. The objects on the resulting list are packed.
67 Besides revisions, `--not` or `--shallow <SHA-1>` lines are
68 also accepted.
70 --unpacked::
71 This implies `--revs`. When processing the list of
72 revision arguments read from the standard input, limit
73 the objects packed to those that are not already packed.
75 --all::
76 This implies `--revs`. In addition to the list of
77 revision arguments read from the standard input, pretend
78 as if all refs under `refs/` are specified to be
79 included.
81 --include-tag::
82 Include unasked-for annotated tags if the object they
83 reference was included in the resulting packfile. This
84 can be useful to send new tags to native Git clients.
86 --window=<n>::
87 --depth=<n>::
88 These two options affect how the objects contained in
89 the pack are stored using delta compression. The
90 objects are first internally sorted by type, size and
91 optionally names and compared against the other objects
92 within --window to see if using delta compression saves
93 space. --depth limits the maximum delta depth; making
94 it too deep affects the performance on the unpacker
95 side, because delta data needs to be applied that many
96 times to get to the necessary object.
97 The default value for --window is 10 and --depth is 50.
99 --window-memory=<n>::
100 This option provides an additional limit on top of `--window`;
101 the window size will dynamically scale down so as to not take
102 up more than '<n>' bytes in memory. This is useful in
103 repositories with a mix of large and small objects to not run
104 out of memory with a large window, but still be able to take
105 advantage of the large window for the smaller objects. The
106 size can be suffixed with "k", "m", or "g".
107 `--window-memory=0` makes memory usage unlimited, which is the
108 default.
110 --max-pack-size=<n>::
111 Maximum size of each output pack file. The size can be suffixed with
112 "k", "m", or "g". The minimum size allowed is limited to 1 MiB.
113 If specified, multiple packfiles may be created, which also
114 prevents the creation of a bitmap index.
115 The default is unlimited, unless the config variable
116 `pack.packSizeLimit` is set.
118 --honor-pack-keep::
119 This flag causes an object already in a local pack that
120 has a .keep file to be ignored, even if it would have
121 otherwise been packed.
123 --incremental::
124 This flag causes an object already in a pack to be ignored
125 even if it would have otherwise been packed.
127 --local::
128 This flag causes an object that is borrowed from an alternate
129 object store to be ignored even if it would have otherwise been
130 packed.
132 --non-empty::
133 Only create a packed archive if it would contain at
134 least one object.
136 --progress::
137 Progress status is reported on the standard error stream
138 by default when it is attached to a terminal, unless -q
139 is specified. This flag forces progress status even if
140 the standard error stream is not directed to a terminal.
142 --all-progress::
143 When --stdout is specified then progress report is
144 displayed during the object count and compression phases
145 but inhibited during the write-out phase. The reason is
146 that in some cases the output stream is directly linked
147 to another command which may wish to display progress
148 status of its own as it processes incoming pack data.
149 This flag is like --progress except that it forces progress
150 report for the write-out phase as well even if --stdout is
151 used.
153 --all-progress-implied::
154 This is used to imply --all-progress whenever progress display
155 is activated. Unlike --all-progress this flag doesn't actually
156 force any progress display by itself.
158 -q::
159 This flag makes the command not to report its progress
160 on the standard error stream.
162 --no-reuse-delta::
163 When creating a packed archive in a repository that
164 has existing packs, the command reuses existing deltas.
165 This sometimes results in a slightly suboptimal pack.
166 This flag tells the command not to reuse existing deltas
167 but compute them from scratch.
169 --no-reuse-object::
170 This flag tells the command not to reuse existing object data at all,
171 including non deltified object, forcing recompression of everything.
172 This implies --no-reuse-delta. Useful only in the obscure case where
173 wholesale enforcement of a different compression level on the
174 packed data is desired.
176 --compression=<n>::
177 Specifies compression level for newly-compressed data in the
178 generated pack. If not specified, pack compression level is
179 determined first by pack.compression, then by core.compression,
180 and defaults to -1, the zlib default, if neither is set.
181 Add --no-reuse-object if you want to force a uniform compression
182 level on all data no matter the source.
184 --thin::
185 Create a "thin" pack by omitting the common objects between a
186 sender and a receiver in order to reduce network transfer. This
187 option only makes sense in conjunction with --stdout.
188 +
189 Note: A thin pack violates the packed archive format by omitting
190 required objects and is thus unusable by Git without making it
191 self-contained. Use `git index-pack --fix-thin`
192 (see linkgit:git-index-pack[1]) to restore the self-contained property.
194 --shallow::
195 Optimize a pack that will be provided to a client with a shallow
196 repository. This option, combined with --thin, can result in a
197 smaller pack at the cost of speed.
199 --delta-base-offset::
200 A packed archive can express the base object of a delta as
201 either a 20-byte object name or as an offset in the
202 stream, but ancient versions of Git don't understand the
203 latter. By default, 'git pack-objects' only uses the
204 former format for better compatibility. This option
205 allows the command to use the latter format for
206 compactness. Depending on the average delta chain
207 length, this option typically shrinks the resulting
208 packfile by 3-5 per-cent.
209 +
210 Note: Porcelain commands such as `git gc` (see linkgit:git-gc[1]),
211 `git repack` (see linkgit:git-repack[1]) pass this option by default
212 in modern Git when they put objects in your repository into pack files.
213 So does `git bundle` (see linkgit:git-bundle[1]) when it creates a bundle.
215 --threads=<n>::
216 Specifies the number of threads to spawn when searching for best
217 delta matches. This requires that pack-objects be compiled with
218 pthreads otherwise this option is ignored with a warning.
219 This is meant to reduce packing time on multiprocessor machines.
220 The required amount of memory for the delta search window is
221 however multiplied by the number of threads.
222 Specifying 0 will cause Git to auto-detect the number of CPU's
223 and set the number of threads accordingly.
225 --index-version=<version>[,<offset>]::
226 This is intended to be used by the test suite only. It allows
227 to force the version for the generated pack index, and to force
228 64-bit index entries on objects located above the given offset.
230 --keep-true-parents::
231 With this option, parents that are hidden by grafts are packed
232 nevertheless.
234 SEE ALSO
235 --------
236 linkgit:git-rev-list[1]
237 linkgit:git-repack[1]
238 linkgit:git-prune-packed[1]
240 GIT
241 ---
242 Part of the linkgit:git[1] suite