Documentation / git-read-tree.txton commit rebase: Allow merge strategies to be used when rebasing (58634db)
   1git-read-tree(1)
   2================
   3
   4NAME
   5----
   6git-read-tree - Reads tree information into the index
   7
   8
   9SYNOPSIS
  10--------
  11'git-read-tree' (<tree-ish> | [[-m [--aggressive] | --reset | --prefix=<prefix>] [-u | -i]] <tree-ish1> [<tree-ish2> [<tree-ish3>]])
  12
  13
  14DESCRIPTION
  15-----------
  16Reads the tree information given by <tree-ish> into the index,
  17but does not actually *update* any of the files it "caches". (see:
  18gitlink:git-checkout-index[1])
  19
  20Optionally, it can merge a tree into the index, perform a
  21fast-forward (i.e. 2-way) merge, or a 3-way merge, with the `-m`
  22flag.  When used with `-m`, the `-u` flag causes it to also update
  23the files in the work tree with the result of the merge.
  24
  25Trivial merges are done by `git-read-tree` itself.  Only conflicting paths
  26will be in unmerged state when `git-read-tree` returns.
  27
  28OPTIONS
  29-------
  30-m::
  31        Perform a merge, not just a read.  The command will
  32        refuse to run if your index file has unmerged entries,
  33        indicating that you have not finished previous merge you
  34        started.
  35
  36--reset::
  37        Same as -m, except that unmerged entries are discarded
  38        instead of failing.
  39
  40-u::
  41        After a successful merge, update the files in the work
  42        tree with the result of the merge.
  43
  44-i::
  45        Usually a merge requires the index file as well as the
  46        files in the working tree are up to date with the
  47        current head commit, in order not to lose local
  48        changes.  This flag disables the check with the working
  49        tree and is meant to be used when creating a merge of
  50        trees that are not directly related to the current
  51        working tree status into a temporary index file.
  52
  53--aggressive::
  54        Usually a three-way merge by `git-read-tree` resolves
  55        the merge for really trivial cases and leaves other
  56        cases unresolved in the index, so that Porcelains can
  57        implement different merge policies.  This flag makes the
  58        command to resolve a few more cases internally:
  59+
  60* when one side removes a path and the other side leaves the path
  61  unmodified.  The resolution is to remove that path.
  62* when both sides remove a path.  The resolution is to remove that path.
  63* when both sides adds a path identically.  The resolution
  64  is to add that path.
  65
  66--prefix=<prefix>/::
  67        Keep the current index contents, and read the contents
  68        of named tree-ish under directory at `<prefix>`.  The
  69        original index file cannot have anything at the path
  70        `<prefix>` itself, and have nothing in `<prefix>/`
  71        directory.  Note that the `<prefix>/` value must end
  72        with a slash.
  73
  74
  75<tree-ish#>::
  76        The id of the tree object(s) to be read/merged.
  77
  78
  79Merging
  80-------
  81If `-m` is specified, `git-read-tree` can perform 3 kinds of
  82merge, a single tree merge if only 1 tree is given, a
  83fast-forward merge with 2 trees, or a 3-way merge if 3 trees are
  84provided.
  85
  86
  87Single Tree Merge
  88~~~~~~~~~~~~~~~~~
  89If only 1 tree is specified, git-read-tree operates as if the user did not
  90specify `-m`, except that if the original index has an entry for a
  91given pathname, and the contents of the path matches with the tree
  92being read, the stat info from the index is used. (In other words, the
  93index's stat()s take precedence over the merged tree's).
  94
  95That means that if you do a `git-read-tree -m <newtree>` followed by a
  96`git-checkout-index -f -u -a`, the `git-checkout-index` only checks out
  97the stuff that really changed.
  98
  99This is used to avoid unnecessary false hits when `git-diff-files` is
 100run after `git-read-tree`.
 101
 102
 103Two Tree Merge
 104~~~~~~~~~~~~~~
 105
 106Typically, this is invoked as `git-read-tree -m $H $M`, where $H
 107is the head commit of the current repository, and $M is the head
 108of a foreign tree, which is simply ahead of $H (i.e. we are in a
 109fast forward situation).
 110
 111When two trees are specified, the user is telling git-read-tree
 112the following:
 113
 114     1. The current index and work tree is derived from $H, but
 115        the user may have local changes in them since $H;
 116
 117     2. The user wants to fast-forward to $M.
 118
 119In this case, the `git-read-tree -m $H $M` command makes sure
 120that no local change is lost as the result of this "merge".
 121Here are the "carry forward" rules:
 122
 123        I (index)           H        M        Result
 124       -------------------------------------------------------
 125      0 nothing             nothing  nothing  (does not happen)
 126      1 nothing             nothing  exists   use M
 127      2 nothing             exists   nothing  remove path from index
 128      3 nothing             exists   exists   use M
 129
 130        clean I==H  I==M
 131       ------------------
 132      4 yes   N/A   N/A     nothing  nothing  keep index
 133      5 no    N/A   N/A     nothing  nothing  keep index
 134
 135      6 yes   N/A   yes     nothing  exists   keep index
 136      7 no    N/A   yes     nothing  exists   keep index
 137      8 yes   N/A   no      nothing  exists   fail
 138      9 no    N/A   no      nothing  exists   fail
 139
 140     10 yes   yes   N/A     exists   nothing  remove path from index
 141     11 no    yes   N/A     exists   nothing  fail
 142     12 yes   no    N/A     exists   nothing  fail
 143     13 no    no    N/A     exists   nothing  fail
 144
 145        clean (H=M)
 146       ------
 147     14 yes                 exists   exists   keep index
 148     15 no                  exists   exists   keep index
 149
 150        clean I==H  I==M (H!=M)
 151       ------------------
 152     16 yes   no    no      exists   exists   fail
 153     17 no    no    no      exists   exists   fail
 154     18 yes   no    yes     exists   exists   keep index
 155     19 no    no    yes     exists   exists   keep index
 156     20 yes   yes   no      exists   exists   use M
 157     21 no    yes   no      exists   exists   fail
 158
 159In all "keep index" cases, the index entry stays as in the
 160original index file.  If the entry were not up to date,
 161git-read-tree keeps the copy in the work tree intact when
 162operating under the -u flag.
 163
 164When this form of git-read-tree returns successfully, you can
 165see what "local changes" you made are carried forward by running
 166`git-diff-index --cached $M`.  Note that this does not
 167necessarily match `git-diff-index --cached $H` would have
 168produced before such a two tree merge.  This is because of cases
 16918 and 19 --- if you already had the changes in $M (e.g. maybe
 170you picked it up via e-mail in a patch form), `git-diff-index
 171--cached $H` would have told you about the change before this
 172merge, but it would not show in `git-diff-index --cached $M`
 173output after two-tree merge.
 174
 175
 1763-Way Merge
 177~~~~~~~~~~~
 178Each "index" entry has two bits worth of "stage" state. stage 0 is the
 179normal one, and is the only one you'd see in any kind of normal use.
 180
 181However, when you do `git-read-tree` with three trees, the "stage"
 182starts out at 1.
 183
 184This means that you can do
 185
 186----------------
 187$ git-read-tree -m <tree1> <tree2> <tree3>
 188----------------
 189
 190and you will end up with an index with all of the <tree1> entries in
 191"stage1", all of the <tree2> entries in "stage2" and all of the
 192<tree3> entries in "stage3".  When performing a merge of another
 193branch into the current branch, we use the common ancestor tree
 194as <tree1>, the current branch head as <tree2>, and the other
 195branch head as <tree3>.
 196
 197Furthermore, `git-read-tree` has special-case logic that says: if you see
 198a file that matches in all respects in the following states, it
 199"collapses" back to "stage0":
 200
 201   - stage 2 and 3 are the same; take one or the other (it makes no
 202     difference - the same work has been done on our branch in
 203     stage 2 and their branch in stage 3)
 204
 205   - stage 1 and stage 2 are the same and stage 3 is different; take
 206     stage 3 (our branch in stage 2 did not do anything since the
 207     ancestor in stage 1 while their branch in stage 3 worked on
 208     it)
 209
 210   - stage 1 and stage 3 are the same and stage 2 is different take
 211     stage 2 (we did something while they did nothing)
 212
 213The `git-write-tree` command refuses to write a nonsensical tree, and it
 214will complain about unmerged entries if it sees a single entry that is not
 215stage 0.
 216
 217OK, this all sounds like a collection of totally nonsensical rules,
 218but it's actually exactly what you want in order to do a fast
 219merge. The different stages represent the "result tree" (stage 0, aka
 220"merged"), the original tree (stage 1, aka "orig"), and the two trees
 221you are trying to merge (stage 2 and 3 respectively).
 222
 223The order of stages 1, 2 and 3 (hence the order of three
 224<tree-ish> command line arguments) are significant when you
 225start a 3-way merge with an index file that is already
 226populated.  Here is an outline of how the algorithm works:
 227
 228- if a file exists in identical format in all three trees, it will
 229  automatically collapse to "merged" state by git-read-tree.
 230
 231- a file that has _any_ difference what-so-ever in the three trees
 232  will stay as separate entries in the index. It's up to "porcelain
 233  policy" to determine how to remove the non-0 stages, and insert a
 234  merged version.
 235
 236- the index file saves and restores with all this information, so you
 237  can merge things incrementally, but as long as it has entries in
 238  stages 1/2/3 (i.e., "unmerged entries") you can't write the result. So
 239  now the merge algorithm ends up being really simple:
 240
 241  * you walk the index in order, and ignore all entries of stage 0,
 242    since they've already been done.
 243
 244  * if you find a "stage1", but no matching "stage2" or "stage3", you
 245    know it's been removed from both trees (it only existed in the
 246    original tree), and you remove that entry.
 247
 248  * if you find a matching "stage2" and "stage3" tree, you remove one
 249    of them, and turn the other into a "stage0" entry. Remove any
 250    matching "stage1" entry if it exists too.  .. all the normal
 251    trivial rules ..
 252
 253You would normally use `git-merge-index` with supplied
 254`git-merge-one-file` to do this last step.  The script updates
 255the files in the working tree as it merges each path and at the
 256end of a successful merge.
 257
 258When you start a 3-way merge with an index file that is already
 259populated, it is assumed that it represents the state of the
 260files in your work tree, and you can even have files with
 261changes unrecorded in the index file.  It is further assumed
 262that this state is "derived" from the stage 2 tree.  The 3-way
 263merge refuses to run if it finds an entry in the original index
 264file that does not match stage 2.
 265
 266This is done to prevent you from losing your work-in-progress
 267changes, and mixing your random changes in an unrelated merge
 268commit.  To illustrate, suppose you start from what has been
 269committed last to your repository:
 270
 271----------------
 272$ JC=`git-rev-parse --verify "HEAD^0"`
 273$ git-checkout-index -f -u -a $JC
 274----------------
 275
 276You do random edits, without running git-update-index.  And then
 277you notice that the tip of your "upstream" tree has advanced
 278since you pulled from him:
 279
 280----------------
 281$ git-fetch git://.... linus
 282$ LT=`cat .git/FETCH_HEAD`
 283----------------
 284
 285Your work tree is still based on your HEAD ($JC), but you have
 286some edits since.  Three-way merge makes sure that you have not
 287added or modified index entries since $JC, and if you haven't,
 288then does the right thing.  So with the following sequence:
 289
 290----------------
 291$ git-read-tree -m -u `git-merge-base $JC $LT` $JC $LT
 292$ git-merge-index git-merge-one-file -a
 293$ echo "Merge with Linus" | \
 294  git-commit-tree `git-write-tree` -p $JC -p $LT
 295----------------
 296
 297what you would commit is a pure merge between $JC and $LT without
 298your work-in-progress changes, and your work tree would be
 299updated to the result of the merge.
 300
 301However, if you have local changes in the working tree that
 302would be overwritten by this merge,`git-read-tree` will refuse
 303to run to prevent your changes from being lost.
 304
 305In other words, there is no need to worry about what exists only
 306in the working tree.  When you have local changes in a part of
 307the project that is not involved in the merge, your changes do
 308not interfere with the merge, and are kept intact.  When they
 309*do* interfere, the merge does not even start (`git-read-tree`
 310complains loudly and fails without modifying anything).  In such
 311a case, you can simply continue doing what you were in the
 312middle of doing, and when your working tree is ready (i.e. you
 313have finished your work-in-progress), attempt the merge again.
 314
 315
 316See Also
 317--------
 318gitlink:git-write-tree[1]; gitlink:git-ls-files[1]
 319
 320
 321Author
 322------
 323Written by Linus Torvalds <torvalds@osdl.org>
 324
 325Documentation
 326--------------
 327Documentation by David Greaves, Junio C Hamano and the git-list <git@vger.kernel.org>.
 328
 329GIT
 330---
 331Part of the gitlink:git[7] suite
 332