Documentation / git-read-tree.txton commit [PATCH] Docs - Makefile update (b2bf34d)
   1git-read-tree(1)
   2================
   3v0.1, May 2005
   4
   5NAME
   6----
   7git-read-tree - Reads tree information into the directory cache
   8
   9
  10SYNOPSIS
  11--------
  12'git-read-tree' (<tree-ish> | -m <tree-ish1> [<tree-ish2> <tree-ish3>])"
  13
  14DESCRIPTION
  15-----------
  16Reads the tree information given by <tree> into the directory cache,
  17but does not actually *update* any of the files it "caches". (see:
  18git-checkout-cache)
  19
  20Optionally, it can merge a tree into the cache or perform a 3-way
  21merge.
  22
  23Trivial merges are done by "git-read-tree" itself.  Only conflicting paths
  24will be in unmerged state when "git-read-tree" returns.
  25
  26OPTIONS
  27-------
  28-m::
  29        Perform a merge, not just a read
  30
  31<tree-ish#>::
  32        The id of the tree object(s) to be read/merged.
  33
  34
  35Merging
  36-------
  37If '-m' is specified, "git-read-tree" performs 2 kinds of merge, a single tree
  38merge if only 1 tree is given or a 3-way merge if 3 trees are
  39provided.
  40
  41Single Tree Merge
  42~~~~~~~~~~~~~~~~~
  43If only 1 tree is specified, git-read-tree operates as if the user did not
  44specify '-m', except that if the original cache has an entry for a
  45given pathname; and the contents of the path matches with the tree
  46being read, the stat info from the cache is used. (In other words, the
  47cache's stat()s take precedence over the merged tree's)
  48
  49That means that if you do a "git-read-tree -m <newtree>" followed by a
  50"git-checkout-cache -f -a", the "git-checkout-cache" only checks out
  51the stuff that really changed.
  52
  53This is used to avoid unnecessary false hits when "git-diff-files" is
  54run after git-read-tree.
  55
  563-Way Merge
  57~~~~~~~~~~~
  58Each "index" entry has two bits worth of "stage" state. stage 0 is the
  59normal one, and is the only one you'd see in any kind of normal use.
  60
  61However, when you do "git-read-tree" with three trees, the "stage"
  62starts out at 1.
  63
  64This means that you can do
  65
  66        git-read-tree -m <tree1> <tree2> <tree3>
  67
  68and you will end up with an index with all of the <tree1> entries in
  69"stage1", all of the <tree2> entries in "stage2" and all of the
  70<tree3> entries in "stage3".
  71
  72Furthermore, "git-read-tree" has special-case logic that says: if you see
  73a file that matches in all respects in the following states, it
  74"collapses" back to "stage0":
  75
  76   - stage 2 and 3 are the same; take one or the other (it makes no
  77     difference - the same work has been done on stage 2 and 3)
  78
  79   - stage 1 and stage 2 are the same and stage 3 is different; take
  80     stage 3 (some work has been done on stage 3)
  81
  82   - stage 1 and stage 3 are the same and stage 2 is different take
  83     stage 2 (some work has been done on stage 2)
  84
  85The "git-write-tree" command refuses to write a nonsensical tree, and it
  86will complain about unmerged entries if it sees a single entry that is not
  87stage 0.
  88
  89Ok, this all sounds like a collection of totally nonsensical rules,
  90but it's actually exactly what you want in order to do a fast
  91merge. The different stages represent the "result tree" (stage 0, aka
  92"merged"), the original tree (stage 1, aka "orig"), and the two trees
  93you are trying to merge (stage 2 and 3 respectively).
  94
  95In fact, the way "git-read-tree" works, it's entirely agnostic about how
  96you assign the stages, and you could really assign them any which way,
  97and the above is just a suggested way to do it (except since
  98"git-write-tree" refuses to write anything but stage0 entries, it makes
  99sense to always consider stage 0 to be the "full merge" state).
 100
 101So what happens? Try it out. Select the original tree, and two trees
 102to merge, and look how it works:
 103
 104- if a file exists in identical format in all three trees, it will
 105  automatically collapse to "merged" state by the new git-read-tree.
 106
 107- a file that has _any_ difference what-so-ever in the three trees
 108  will stay as separate entries in the index. It's up to "script
 109  policy" to determine how to remove the non-0 stages, and insert a
 110  merged version.  But since the index is always sorted, they're easy
 111  to find: they'll be clustered together.
 112
 113- the index file saves and restores with all this information, so you
 114  can merge things incrementally, but as long as it has entries in
 115  stages 1/2/3 (ie "unmerged entries") you can't write the result. So
 116  now the merge algorithm ends up being really simple:
 117
 118  * you walk the index in order, and ignore all entries of stage 0,
 119    since they've already been done.
 120
 121  * if you find a "stage1", but no matching "stage2" or "stage3", you
 122    know it's been removed from both trees (it only existed in the
 123    original tree), and you remove that entry.
 124
 125  * if you find a matching "stage2" and "stage3" tree, you remove one
 126    of them, and turn the other into a "stage0" entry. Remove any
 127    matching "stage1" entry if it exists too.  .. all the normal
 128    trivial rules ..
 129
 130Incidentally - it also means that you don't even have to have a
 131separate subdirectory for this. All the information literally is in
 132the index file, which is a temporary thing anyway. There is no need to
 133worry about what is in the working directory, since it is never shown
 134and never used.
 135
 136See Also
 137--------
 138link:git-write-tree.html[git-write-tree]; link:git-ls-files.html[git-ls-files]
 139
 140
 141Author
 142------
 143Written by Linus Torvalds <torvalds@osdl.org>
 144
 145Documentation
 146--------------
 147Documentation by David Greaves, Junio C Hamano and the git-list <git@vger.kernel.org>.
 148
 149GIT
 150---
 151Part of the link:git.html[git] suite
 152