Merge branch 'svn-crlf' of git:// into ew/svn-crlf
[git/git.git] / Documentation / git-pack-objects.txt
1 git-pack-objects(1)
2 ===================
5 ----
6 git-pack-objects - Create a packed archive of objects
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
20 -----------
21 Reads list of objects from the standard input, and writes either one or
22 more packed archives with the specified base-name to disk, or a packed
23 archive to the standard output.
25 A packed archive is an efficient way to transfer a set of objects
26 between two repositories as well as an access efficient archival
27 format. In a packed archive, an object is either stored as a
28 compressed whole or as a difference from some other object.
29 The latter is often called a delta.
31 The packed archive format (.pack) is designed to be self-contained
32 so that it can be unpacked without any further information. Therefore,
33 each object that a delta depends upon must be present within the pack.
35 A pack index file (.idx) is generated for fast, random access to the
36 objects in the pack. Placing both the index file (.idx) and the packed
37 archive (.pack) in the pack/ subdirectory of $GIT_OBJECT_DIRECTORY (or
38 any of the directories on $GIT_ALTERNATE_OBJECT_DIRECTORIES)
39 enables Git to read from the pack archive.
41 The 'git unpack-objects' command can read the packed archive and
42 expand the objects contained in the pack into "one-file
43 one-object" format; this is typically done by the smart-pull
44 commands when a pack is created on-the-fly for efficient network
45 transport by their peers.
49 -------
50 base-name::
51 Write into pairs of files (.pack and .idx), using
52 <base-name> to determine the name of the created file.
53 When this option is used, the two files in a pair are written in
54 <base-name>-<SHA-1>.{pack,idx} files. <SHA-1> is a hash
55 based on the pack content and is written to the standard
56 output of the command.
58 --stdout::
59 Write the pack contents (what would have been written to
60 .pack file) out to the standard output.
62 --revs::
63 Read the revision arguments from the standard input, instead of
64 individual object names. The revision arguments are processed
65 the same way as 'git rev-list' with the `--objects` flag
66 uses its `commit` arguments to build the list of objects it
67 outputs. The objects on the resulting list are packed.
68 Besides revisions, `--not` or `--shallow <SHA-1>` lines are
69 also accepted.
71 --unpacked::
72 This implies `--revs`. When processing the list of
73 revision arguments read from the standard input, limit
74 the objects packed to those that are not already packed.
76 --all::
77 This implies `--revs`. In addition to the list of
78 revision arguments read from the standard input, pretend
79 as if all refs under `refs/` are specified to be
80 included.
82 --include-tag::
83 Include unasked-for annotated tags if the object they
84 reference was included in the resulting packfile. This
85 can be useful to send new tags to native Git clients.
87 --window=<n>::
88 --depth=<n>::
89 These two options affect how the objects contained in
90 the pack are stored using delta compression. The
91 objects are first internally sorted by type, size and
92 optionally names and compared against the other objects
93 within --window to see if using delta compression saves
94 space. --depth limits the maximum delta depth; making
95 it too deep affects the performance on the unpacker
96 side, because delta data needs to be applied that many
97 times to get to the necessary object.
98 The default value for --window is 10 and --depth is 50.
100 --window-memory=<n>::
101 This option provides an additional limit on top of `--window`;
102 the window size will dynamically scale down so as to not take
103 up more than '<n>' bytes in memory. This is useful in
104 repositories with a mix of large and small objects to not run
105 out of memory with a large window, but still be able to take
106 advantage of the large window for the smaller objects. The
107 size can be suffixed with "k", "m", or "g".
108 `--window-memory=0` makes memory usage unlimited. The default
109 is taken from the `pack.windowMemory` configuration variable.
111 --max-pack-size=<n>::
112 In unusual scenarios, you may not be able to create files
113 larger than a certain size on your filesystem, and this option
114 can be used to tell the command to split the output packfile
115 into multiple independent packfiles, each not larger than the
116 given size. The size can be suffixed with
117 "k", "m", or "g". The minimum size allowed is limited to 1 MiB.
118 This option
119 prevents the creation of a bitmap index.
120 The default is unlimited, unless the config variable
121 `pack.packSizeLimit` is set.
123 --honor-pack-keep::
124 This flag causes an object already in a local pack that
125 has a .keep file to be ignored, even if it would have
126 otherwise been packed.
128 --incremental::
129 This flag causes an object already in a pack to be ignored
130 even if it would have otherwise been packed.
132 --local::
133 This flag causes an object that is borrowed from an alternate
134 object store to be ignored even if it would have otherwise been
135 packed.
137 --non-empty::
138 Only create a packed archive if it would contain at
139 least one object.
141 --progress::
142 Progress status is reported on the standard error stream
143 by default when it is attached to a terminal, unless -q
144 is specified. This flag forces progress status even if
145 the standard error stream is not directed to a terminal.
147 --all-progress::
148 When --stdout is specified then progress report is
149 displayed during the object count and compression phases
150 but inhibited during the write-out phase. The reason is
151 that in some cases the output stream is directly linked
152 to another command which may wish to display progress
153 status of its own as it processes incoming pack data.
154 This flag is like --progress except that it forces progress
155 report for the write-out phase as well even if --stdout is
156 used.
158 --all-progress-implied::
159 This is used to imply --all-progress whenever progress display
160 is activated. Unlike --all-progress this flag doesn't actually
161 force any progress display by itself.
163 -q::
164 This flag makes the command not to report its progress
165 on the standard error stream.
167 --no-reuse-delta::
168 When creating a packed archive in a repository that
169 has existing packs, the command reuses existing deltas.
170 This sometimes results in a slightly suboptimal pack.
171 This flag tells the command not to reuse existing deltas
172 but compute them from scratch.
174 --no-reuse-object::
175 This flag tells the command not to reuse existing object data at all,
176 including non deltified object, forcing recompression of everything.
177 This implies --no-reuse-delta. Useful only in the obscure case where
178 wholesale enforcement of a different compression level on the
179 packed data is desired.
181 --compression=<n>::
182 Specifies compression level for newly-compressed data in the
183 generated pack. If not specified, pack compression level is
184 determined first by pack.compression, then by core.compression,
185 and defaults to -1, the zlib default, if neither is set.
186 Add --no-reuse-object if you want to force a uniform compression
187 level on all data no matter the source.
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`
197 (see linkgit:git-index-pack[1]) to restore the self-contained property.
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.
204 --delta-base-offset::
205 A packed archive can express the base object of a delta as
206 either a 20-byte object name or as an offset in the
207 stream, but ancient versions of Git don't understand the
208 latter. By default, 'git pack-objects' only uses the
209 former format for better compatibility. This option
210 allows the command to use the latter format for
211 compactness. Depending on the average delta chain
212 length, this option typically shrinks the resulting
213 packfile by 3-5 per-cent.
214 +
215 Note: Porcelain commands such as `git gc` (see linkgit:git-gc[1]),
216 `git repack` (see linkgit:git-repack[1]) pass this option by default
217 in modern Git when they put objects in your repository into pack files.
218 So does `git bundle` (see linkgit:git-bundle[1]) when it creates a bundle.
220 --threads=<n>::
221 Specifies the number of threads to spawn when searching for best
222 delta matches. This requires that pack-objects be compiled with
223 pthreads otherwise this option is ignored with a warning.
224 This is meant to reduce packing time on multiprocessor machines.
225 The required amount of memory for the delta search window is
226 however multiplied by the number of threads.
227 Specifying 0 will cause Git to auto-detect the number of CPU's
228 and set the number of threads accordingly.
230 --index-version=<version>[,<offset>]::
231 This is intended to be used by the test suite only. It allows
232 to force the version for the generated pack index, and to force
233 64-bit index entries on objects located above the given offset.
235 --keep-true-parents::
236 With this option, parents that are hidden by grafts are packed
237 nevertheless.
240 --------
241 linkgit:git-rev-list[1]
242 linkgit:git-repack[1]
243 linkgit:git-prune-packed[1]
245 GIT
246 ---
247 Part of the linkgit:git[1] suite