Documentation / repository-layout.txton commit Documentation: illustrate send-pack pipeline. (b5ffa5c)
   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).
   7
   8objects::
   9        Object store associated with this repository.  Usually
  10        an object store is self sufficient (i.e. all the objects
  11        that are referred to by an object found in it are also
  12        found in it), but there are couple of ways to violate
  13        it.
  14+
  15. You could populate the repository by running a commit walker
  16without `-a` option.  Depending on which options are given, you
  17could have only commit objects without associated blobs and
  18trees this way, for example.  A repository with this kind of
  19incomplete object store is not suitable to be published to the
  20outside world but sometimes useful for private repository.
  21. You can be using `objects/info/alternates` mechanism, or
  22`$GIT_ALTERNATE_OBJECT_DIRECTORIES` mechanism to 'borrow'
  23objects from other object stores.  A repository with this kind
  24of incomplete object store is not suitable to be published for
  25use with dumb transports but otherwise is OK as long as
  26`objects/info/alternates` points at the right object stores
  27it borrows from.
  28
  29objects/[0-9a-f][0-9a-f]::
  30        Traditionally, each object is stored in its own file.
  31        They are split into 256 subdirectories using the first
  32        two letters from its object name to keep the number of
  33        directory entries `objects` directory itself needs to
  34        hold.  Objects found here are often called 'unpacked'
  35        objects.
  36
  37objects/pack::
  38        Packs (files that store many object in compressed form,
  39        along with index files to allow them to be randomly
  40        accessed) are found in this directory.
  41
  42objects/info::
  43        Additional information about the object store is
  44        recorded in this directory.
  45
  46objects/info/packs::
  47        This file is to help dumb transports discover what packs
  48        are available in this object store.  Whenever a pack is
  49        added or removed, `git update-server-info` should be run
  50        to keep this file up-to-date if the repository is
  51        published for dumb transports.  `git repack` does this
  52        by default.
  53
  54objects/info/alternates::
  55        This file records paths to alternate object stores that
  56        this object store borrows objects from, one pathname per
  57        line. Note that not only native Git tools use it locally,
  58        but the HTTP fetcher also tries to use it remotely; this
  59        will usually work if you have relative paths (relative
  60        to the object database, not to the repository!) in your
  61        alternates file, but it will not work if you use absolute
  62        paths unless the absolute path in filesystem and web URL
  63        is the same. See also 'objects/info/http-alternates'.
  64
  65objects/info/http-alternates::
  66        This file records URLs to alternate object stores that
  67        this object store borrows objects from, to be used when
  68        the repository is fetched over HTTP.
  69
  70refs::
  71        References are stored in subdirectories of this
  72        directory.  The `git prune` command knows to keep
  73        objects reachable from refs found in this directory and
  74        its subdirectories.
  75
  76refs/heads/`name`::
  77        records tip-of-the-tree commit objects of branch `name`
  78
  79refs/tags/`name`::
  80        records any object name (not necessarily a commit
  81        object, or a tag object that points at a commit object).
  82
  83HEAD::
  84        A symref (see glossary) to the `refs/heads/` namespace
  85        describing the currently active branch.  It does not mean
  86        much if the repository is not associated with any working tree
  87        (i.e. a 'bare' repository), but a valid git repository
  88        *must* have the HEAD file; some porcelains may use it to
  89        guess the designated "default" branch of the repository
  90        (usually 'master').  It is legal if the named branch
  91        'name' does not (yet) exist.  In some legacy setups, it is
  92        a symbolic link instead of a symref that points at the current
  93        branch.
  94
  95branches::
  96        A slightly deprecated way to store shorthands to be used
  97        to specify URL to `git fetch`, `git pull` and `git push`
  98        commands is to store a file in `branches/'name'` and
  99        give 'name' to these commands in place of 'repository'
 100        argument.
 101
 102hooks::
 103        Hooks are customization scripts used by various git
 104        commands.  A handful of sample hooks are installed when
 105        `git init-db` is run, but all of them are disabled by
 106        default.  To enable, they need to be made executable.
 107        Read link:hooks.html[hooks] for more details about
 108        each hook.
 109
 110index::
 111        The current index file for the repository.  It is
 112        usually not found in a bare repository.
 113
 114info::
 115        Additional information about the repository is recorded
 116        in this directory.
 117
 118info/refs::
 119        This file is to help dumb transports to discover what
 120        refs are available in this repository.  Whenever you
 121        create/delete a new branch or a new tag, `git
 122        update-server-info` should be run to keep this file
 123        up-to-date if the repository is published for dumb
 124        transports.  The `git-receive-pack` command, which is
 125        run on a remote repository when you `git push` into it,
 126        runs `hooks/update` hook to help you achieve this.
 127
 128info/grafts::
 129        This file records fake commit ancestry information, to
 130        pretend the set of parents a commit has is different
 131        from how the commit was actually created.  One record
 132        per line describes a commit and its fake parents by
 133        listing their 40-byte hexadecimal object names separated
 134        by a space and terminated by a newline.
 135
 136info/exclude::
 137        This file, by convention among Porcelains, stores the
 138        exclude pattern list. `.gitignore` is the per-directory
 139        ignore file.  `git status`, `git add`, `git rm` and `git
 140        clean` look at it but the core git commands do not look
 141        at it.  See also: gitlink:git-ls-files[1] `--exclude-from`
 142        and `--exclude-per-directory`.
 143
 144remotes::
 145        Stores shorthands to be used to give URL and default
 146        refnames to interact with remote repository to `git
 147        fetch`, `git pull` and `git push` commands.
 148
 149logs::
 150        Records of changes made to refs are stored in this
 151        directory.  See the documentation on git-update-ref
 152        for more information.
 153
 154logs/refs/heads/`name`::
 155        Records all changes made to the branch tip named `name`.
 156
 157logs/refs/tags/`name`::
 158        Records all changes made to the tag named `name`.