1gitrepository-layout(5) 2======================= 3 4NAME 5---- 6gitrepository-layout - Git Repository Layout 7 8SYNOPSIS 9-------- 10$GIT_DIR/* 11 12DESCRIPTION 13----------- 14 15You may find these things in your git repository (`.git` 16directory for a repository associated with your working tree, or 17`<project>.git` directory for a public 'bare' repository. It is 18also possible to have a working tree where `.git` is a plain 19ascii file containing `gitdir: <path>`, i.e. the path to the 20real git repository). 21 22objects:: 23 Object store associated with this repository. Usually 24 an object store is self sufficient (i.e. all the objects 25 that are referred to by an object found in it are also 26 found in it), but there are couple of ways to violate 27 it. 28+ 29. You could populate the repository by running a commit walker 30without `-a` option. Depending on which options are given, you 31could have only commit objects without associated blobs and 32trees this way, for example. A repository with this kind of 33incomplete object store is not suitable to be published to the 34outside world but sometimes useful for private repository. 35. You also could have an incomplete but locally usable repository 36by cloning shallowly. See linkgit:git-clone[1]. 37. You can be using `objects/info/alternates` mechanism, or 38`$GIT_ALTERNATE_OBJECT_DIRECTORIES` mechanism to 'borrow' 39objects from other object stores. A repository with this kind 40of incomplete object store is not suitable to be published for 41use with dumb transports but otherwise is OK as long as 42`objects/info/alternates` points at the right object stores 43it borrows from. 44 45objects/[0-9a-f][0-9a-f]:: 46 Traditionally, each object is stored in its own file. 47 They are split into 256 subdirectories using the first 48 two letters from its object name to keep the number of 49 directory entries `objects` directory itself needs to 50 hold. Objects found here are often called 'unpacked' 51 (or 'loose') objects. 52 53objects/pack:: 54 Packs (files that store many object in compressed form, 55 along with index files to allow them to be randomly 56 accessed) are found in this directory. 57 58objects/info:: 59 Additional information about the object store is 60 recorded in this directory. 61 62objects/info/packs:: 63 This file is to help dumb transports discover what packs 64 are available in this object store. Whenever a pack is 65 added or removed, `git update-server-info` should be run 66 to keep this file up-to-date if the repository is 67 published for dumb transports. 'git repack' does this 68 by default. 69 70objects/info/alternates:: 71 This file records paths to alternate object stores that 72 this object store borrows objects from, one pathname per 73 line. Note that not only native Git tools use it locally, 74 but the HTTP fetcher also tries to use it remotely; this 75 will usually work if you have relative paths (relative 76 to the object database, not to the repository!) in your 77 alternates file, but it will not work if you use absolute 78 paths unless the absolute path in filesystem and web URL 79 is the same. See also 'objects/info/http-alternates'. 80 81objects/info/http-alternates:: 82 This file records URLs to alternate object stores that 83 this object store borrows objects from, to be used when 84 the repository is fetched over HTTP. 85 86refs:: 87 References are stored in subdirectories of this 88 directory. The 'git prune' command knows to keep 89 objects reachable from refs found in this directory and 90 its subdirectories. 91 92refs/heads/`name`:: 93 records tip-of-the-tree commit objects of branch `name` 94 95refs/tags/`name`:: 96 records any object name (not necessarily a commit 97 object, or a tag object that points at a commit object). 98 99refs/remotes/`name`:: 100 records tip-of-the-tree commit objects of branches copied 101 from a remote repository. 102 103packed-refs:: 104 records the same information as refs/heads/, refs/tags/, 105 and friends record in a more efficient way. See 106 linkgit:git-pack-refs[1]. 107 108HEAD:: 109 A symref (see glossary) to the `refs/heads/` namespace 110 describing the currently active branch. It does not mean 111 much if the repository is not associated with any working tree 112 (i.e. a 'bare' repository), but a valid git repository 113 *must* have the HEAD file; some porcelains may use it to 114 guess the designated "default" branch of the repository 115 (usually 'master'). It is legal if the named branch 116 'name' does not (yet) exist. In some legacy setups, it is 117 a symbolic link instead of a symref that points at the current 118 branch. 119+ 120HEAD can also record a specific commit directly, instead of 121being a symref to point at the current branch. Such a state 122is often called 'detached HEAD', and almost all commands work 123identically as normal. See linkgit:git-checkout[1] for 124details. 125 126branches:: 127 A slightly deprecated way to store shorthands to be used 128 to specify URL to 'git fetch', 'git pull' and 'git push' 129 commands is to store a file in `branches/<name>` and 130 give 'name' to these commands in place of 'repository' 131 argument. 132 133hooks:: 134 Hooks are customization scripts used by various git 135 commands. A handful of sample hooks are installed when 136 'git init' is run, but all of them are disabled by 137 default. To enable, the `.sample` suffix has to be 138 removed from the filename by renaming. 139 Read linkgit:githooks[5] for more details about 140 each hook. 141 142index:: 143 The current index file for the repository. It is 144 usually not found in a bare repository. 145 146info:: 147 Additional information about the repository is recorded 148 in this directory. 149 150info/refs:: 151 This file helps dumb transports discover what refs are 152 available in this repository. If the repository is 153 published for dumb transports, this file should be 154 regenerated by 'git update-server-info' every time a tag 155 or branch is created or modified. This is normally done 156 from the `hooks/update` hook, which is run by the 157 'git-receive-pack' command when you 'git push' into the 158 repository. 159 160info/grafts:: 161 This file records fake commit ancestry information, to 162 pretend the set of parents a commit has is different 163 from how the commit was actually created. One record 164 per line describes a commit and its fake parents by 165 listing their 40-byte hexadecimal object names separated 166 by a space and terminated by a newline. 167 168info/exclude:: 169 This file, by convention among Porcelains, stores the 170 exclude pattern list. `.gitignore` is the per-directory 171 ignore file. 'git status', 'git add', 'git rm' and 172 'git clean' look at it but the core git commands do not look 173 at it. See also: linkgit:gitignore[5]. 174 175remotes:: 176 Stores shorthands to be used to give URL and default 177 refnames to interact with remote repository to 178 'git fetch', 'git pull' and 'git push' commands. 179 180logs:: 181 Records of changes made to refs are stored in this 182 directory. See linkgit:git-update-ref[1] 183 for more information. 184 185logs/refs/heads/`name`:: 186 Records all changes made to the branch tip named `name`. 187 188logs/refs/tags/`name`:: 189 Records all changes made to the tag named `name`. 190 191shallow:: 192 This is similar to `info/grafts` but is internally used 193 and maintained by shallow clone mechanism. See `--depth` 194 option to linkgit:git-clone[1] and linkgit:git-fetch[1]. 195 196SEE ALSO 197-------- 198linkgit:git-init[1], 199linkgit:git-clone[1], 200linkgit:git-fetch[1], 201linkgit:git-pack-refs[1], 202linkgit:git-gc[1], 203linkgit:git-checkout[1], 204linkgit:gitglossary[7], 205link:user-manual.html[The Git User's Manual] 206 207GIT 208--- 209Part of the linkgit:git[1] suite.