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