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' [[-m [--trivial] [--aggressive] | --reset | --prefix=<prefix>] 12 [-u [--exclude-per-directory=<gitignore>] | -i]] 13 [--index-output=<file>] 14 <tree-ish1> [<tree-ish2> [<tree-ish3>]] 15 16 17DESCRIPTION 18----------- 19Reads the tree information given by <tree-ish> into the index, 20but does not actually *update* any of the files it "caches". (see: 21linkgit:git-checkout-index[1]) 22 23Optionally, it can merge a tree into the index, perform a 24fast-forward (i.e. 2-way) merge, or a 3-way merge, with the `-m` 25flag. When used with `-m`, the `-u` flag causes it to also update 26the files in the work tree with the result of the merge. 27 28Trivial merges are done by 'git read-tree' itself. Only conflicting paths 29will be in unmerged state when 'git read-tree' returns. 30 31OPTIONS 32------- 33-m:: 34 Perform a merge, not just a read. The command will 35 refuse to run if your index file has unmerged entries, 36 indicating that you have not finished previous merge you 37 started. 38 39--reset:: 40 Same as -m, except that unmerged entries are discarded 41 instead of failing. 42 43-u:: 44 After a successful merge, update the files in the work 45 tree with the result of the merge. 46 47-i:: 48 Usually a merge requires the index file as well as the 49 files in the working tree are up to date with the 50 current head commit, in order not to lose local 51 changes. This flag disables the check with the working 52 tree and is meant to be used when creating a merge of 53 trees that are not directly related to the current 54 working tree status into a temporary index file. 55 56-v:: 57 Show the progress of checking files out. 58 59--trivial:: 60 Restrict three-way merge by 'git read-tree' to happen 61 only if there is no file-level merging required, instead 62 of resolving merge for trivial cases and leaving 63 conflicting files unresolved in the index. 64 65--aggressive:: 66 Usually a three-way merge by 'git read-tree' resolves 67 the merge for really trivial cases and leaves other 68 cases unresolved in the index, so that Porcelains can 69 implement different merge policies. This flag makes the 70 command to resolve a few more cases internally: 71+ 72* when one side removes a path and the other side leaves the path 73 unmodified. The resolution is to remove that path. 74* when both sides remove a path. The resolution is to remove that path. 75* when both sides adds a path identically. The resolution 76 is to add that path. 77 78--prefix=<prefix>/:: 79 Keep the current index contents, and read the contents 80 of named tree-ish under directory at `<prefix>`. The 81 original index file cannot have anything at the path 82 `<prefix>` itself, and have nothing in `<prefix>/` 83 directory. Note that the `<prefix>/` value must end 84 with a slash. 85 86--exclude-per-directory=<gitignore>:: 87 When running the command with `-u` and `-m` options, the 88 merge result may need to overwrite paths that are not 89 tracked in the current branch. The command usually 90 refuses to proceed with the merge to avoid losing such a 91 path. However this safety valve sometimes gets in the 92 way. For example, it often happens that the other 93 branch added a file that used to be a generated file in 94 your branch, and the safety valve triggers when you try 95 to switch to that branch after you ran `make` but before 96 running `make clean` to remove the generated file. This 97 option tells the command to read per-directory exclude 98 file (usually '.gitignore') and allows such an untracked 99 but explicitly ignored file to be overwritten. 100 101--index-output=<file>:: 102 Instead of writing the results out to `$GIT_INDEX_FILE`, 103 write the resulting index in the named file. While the 104 command is operating, the original index file is locked 105 with the same mechanism as usual. The file must allow 106 to be rename(2)ed into from a temporary file that is 107 created next to the usual index file; typically this 108 means it needs to be on the same filesystem as the index 109 file itself, and you need write permission to the 110 directories the index file and index output file are 111 located in. 112 113<tree-ish#>:: 114 The id of the tree object(s) to be read/merged. 115 116 117Merging 118------- 119If `-m` is specified, 'git read-tree' can perform 3 kinds of 120merge, a single tree merge if only 1 tree is given, a 121fast-forward merge with 2 trees, or a 3-way merge if 3 trees are 122provided. 123 124 125Single Tree Merge 126~~~~~~~~~~~~~~~~~ 127If only 1 tree is specified, 'git read-tree' operates as if the user did not 128specify `-m`, except that if the original index has an entry for a 129given pathname, and the contents of the path matches with the tree 130being read, the stat info from the index is used. (In other words, the 131index's stat()s take precedence over the merged tree's). 132 133That means that if you do a `git read-tree -m <newtree>` followed by a 134`git checkout-index -f -u -a`, the 'git checkout-index' only checks out 135the stuff that really changed. 136 137This is used to avoid unnecessary false hits when 'git diff-files' is 138run after 'git read-tree'. 139 140 141Two Tree Merge 142~~~~~~~~~~~~~~ 143 144Typically, this is invoked as `git read-tree -m $H $M`, where $H 145is the head commit of the current repository, and $M is the head 146of a foreign tree, which is simply ahead of $H (i.e. we are in a 147fast-forward situation). 148 149When two trees are specified, the user is telling 'git read-tree' 150the following: 151 152 1. The current index and work tree is derived from $H, but 153 the user may have local changes in them since $H; 154 155 2. The user wants to fast-forward to $M. 156 157In this case, the `git read-tree -m $H $M` command makes sure 158that no local change is lost as the result of this "merge". 159Here are the "carry forward" rules: 160 161 I (index) H M Result 162 ------------------------------------------------------- 163 0 nothing nothing nothing (does not happen) 164 1 nothing nothing exists use M 165 2 nothing exists nothing remove path from index 166 3 nothing exists exists, use M if "initial checkout" 167 H == M keep index otherwise 168 exists fail 169 H != M 170 171 clean I==H I==M 172 ------------------ 173 4 yes N/A N/A nothing nothing keep index 174 5 no N/A N/A nothing nothing keep index 175 176 6 yes N/A yes nothing exists keep index 177 7 no N/A yes nothing exists keep index 178 8 yes N/A no nothing exists fail 179 9 no N/A no nothing exists fail 180 181 10 yes yes N/A exists nothing remove path from index 182 11 no yes N/A exists nothing fail 183 12 yes no N/A exists nothing fail 184 13 no no N/A exists nothing fail 185 186 clean (H=M) 187 ------ 188 14 yes exists exists keep index 189 15 no exists exists keep index 190 191 clean I==H I==M (H!=M) 192 ------------------ 193 16 yes no no exists exists fail 194 17 no no no exists exists fail 195 18 yes no yes exists exists keep index 196 19 no no yes exists exists keep index 197 20 yes yes no exists exists use M 198 21 no yes no exists exists fail 199 200In all "keep index" cases, the index entry stays as in the 201original index file. If the entry were not up to date, 202'git read-tree' keeps the copy in the work tree intact when 203operating under the -u flag. 204 205When this form of 'git read-tree' returns successfully, you can 206see what "local changes" you made are carried forward by running 207`git diff-index --cached $M`. Note that this does not 208necessarily match `git diff-index --cached $H` would have 209produced before such a two tree merge. This is because of cases 21018 and 19 --- if you already had the changes in $M (e.g. maybe 211you picked it up via e-mail in a patch form), `git diff-index 212--cached $H` would have told you about the change before this 213merge, but it would not show in `git diff-index --cached $M` 214output after two-tree merge. 215 216Case #3 is slightly tricky and needs explanation. The result from this 217rule logically should be to remove the path if the user staged the removal 218of the path and then switching to a new branch. That however will prevent 219the initial checkout from happening, so the rule is modified to use M (new 220tree) only when the contents of the index is empty. Otherwise the removal 221of the path is kept as long as $H and $M are the same. 222 2233-Way Merge 224~~~~~~~~~~~ 225Each "index" entry has two bits worth of "stage" state. stage 0 is the 226normal one, and is the only one you'd see in any kind of normal use. 227 228However, when you do 'git read-tree' with three trees, the "stage" 229starts out at 1. 230 231This means that you can do 232 233---------------- 234$ git read-tree -m <tree1> <tree2> <tree3> 235---------------- 236 237and you will end up with an index with all of the <tree1> entries in 238"stage1", all of the <tree2> entries in "stage2" and all of the 239<tree3> entries in "stage3". When performing a merge of another 240branch into the current branch, we use the common ancestor tree 241as <tree1>, the current branch head as <tree2>, and the other 242branch head as <tree3>. 243 244Furthermore, 'git read-tree' has special-case logic that says: if you see 245a file that matches in all respects in the following states, it 246"collapses" back to "stage0": 247 248 - stage 2 and 3 are the same; take one or the other (it makes no 249 difference - the same work has been done on our branch in 250 stage 2 and their branch in stage 3) 251 252 - stage 1 and stage 2 are the same and stage 3 is different; take 253 stage 3 (our branch in stage 2 did not do anything since the 254 ancestor in stage 1 while their branch in stage 3 worked on 255 it) 256 257 - stage 1 and stage 3 are the same and stage 2 is different take 258 stage 2 (we did something while they did nothing) 259 260The 'git write-tree' command refuses to write a nonsensical tree, and it 261will complain about unmerged entries if it sees a single entry that is not 262stage 0. 263 264OK, this all sounds like a collection of totally nonsensical rules, 265but it's actually exactly what you want in order to do a fast 266merge. The different stages represent the "result tree" (stage 0, aka 267"merged"), the original tree (stage 1, aka "orig"), and the two trees 268you are trying to merge (stage 2 and 3 respectively). 269 270The order of stages 1, 2 and 3 (hence the order of three 271<tree-ish> command line arguments) are significant when you 272start a 3-way merge with an index file that is already 273populated. Here is an outline of how the algorithm works: 274 275- if a file exists in identical format in all three trees, it will 276 automatically collapse to "merged" state by 'git read-tree'. 277 278- a file that has _any_ difference what-so-ever in the three trees 279 will stay as separate entries in the index. It's up to "porcelain 280 policy" to determine how to remove the non-0 stages, and insert a 281 merged version. 282 283- the index file saves and restores with all this information, so you 284 can merge things incrementally, but as long as it has entries in 285 stages 1/2/3 (i.e., "unmerged entries") you can't write the result. So 286 now the merge algorithm ends up being really simple: 287 288 * you walk the index in order, and ignore all entries of stage 0, 289 since they've already been done. 290 291 * if you find a "stage1", but no matching "stage2" or "stage3", you 292 know it's been removed from both trees (it only existed in the 293 original tree), and you remove that entry. 294 295 * if you find a matching "stage2" and "stage3" tree, you remove one 296 of them, and turn the other into a "stage0" entry. Remove any 297 matching "stage1" entry if it exists too. .. all the normal 298 trivial rules .. 299 300You would normally use 'git merge-index' with supplied 301'git merge-one-file' to do this last step. The script updates 302the files in the working tree as it merges each path and at the 303end of a successful merge. 304 305When you start a 3-way merge with an index file that is already 306populated, it is assumed that it represents the state of the 307files in your work tree, and you can even have files with 308changes unrecorded in the index file. It is further assumed 309that this state is "derived" from the stage 2 tree. The 3-way 310merge refuses to run if it finds an entry in the original index 311file that does not match stage 2. 312 313This is done to prevent you from losing your work-in-progress 314changes, and mixing your random changes in an unrelated merge 315commit. To illustrate, suppose you start from what has been 316committed last to your repository: 317 318---------------- 319$ JC=`git rev-parse --verify "HEAD^0"` 320$ git checkout-index -f -u -a $JC 321---------------- 322 323You do random edits, without running 'git update-index'. And then 324you notice that the tip of your "upstream" tree has advanced 325since you pulled from him: 326 327---------------- 328$ git fetch git://.... linus 329$ LT=`cat .git/FETCH_HEAD` 330---------------- 331 332Your work tree is still based on your HEAD ($JC), but you have 333some edits since. Three-way merge makes sure that you have not 334added or modified index entries since $JC, and if you haven't, 335then does the right thing. So with the following sequence: 336 337---------------- 338$ git read-tree -m -u `git merge-base $JC $LT` $JC $LT 339$ git merge-index git-merge-one-file -a 340$ echo "Merge with Linus" | \ 341 git commit-tree `git write-tree` -p $JC -p $LT 342---------------- 343 344what you would commit is a pure merge between $JC and $LT without 345your work-in-progress changes, and your work tree would be 346updated to the result of the merge. 347 348However, if you have local changes in the working tree that 349would be overwritten by this merge, 'git read-tree' will refuse 350to run to prevent your changes from being lost. 351 352In other words, there is no need to worry about what exists only 353in the working tree. When you have local changes in a part of 354the project that is not involved in the merge, your changes do 355not interfere with the merge, and are kept intact. When they 356*do* interfere, the merge does not even start ('git read-tree' 357complains loudly and fails without modifying anything). In such 358a case, you can simply continue doing what you were in the 359middle of doing, and when your working tree is ready (i.e. you 360have finished your work-in-progress), attempt the merge again. 361 362 363SEE ALSO 364-------- 365linkgit:git-write-tree[1]; linkgit:git-ls-files[1]; 366linkgit:gitignore[5] 367 368 369Author 370------ 371Written by Linus Torvalds <torvalds@osdl.org> 372 373Documentation 374-------------- 375Documentation by David Greaves, Junio C Hamano and the git-list <git@vger.kernel.org>. 376 377GIT 378--- 379Part of the linkgit:git[1] suite