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]] [--exclude-per-directory=<gitignore>] [--index-output=<file>] <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--exclude-per-directory=<gitignore>:: 75 When running the command with `-u` and `-m` options, the 76 merge result may need to overwrite paths that are not 77 tracked in the current branch. The command usually 78 refuses to proceed with the merge to avoid losing such a 79 path. However this safety valve sometimes gets in the 80 way. For example, it often happens that the other 81 branch added a file that used to be a generated file in 82 your branch, and the safety valve triggers when you try 83 to switch to that branch after you ran `make` but before 84 running `make clean` to remove the generated file. This 85 option tells the command to read per-directory exclude 86 file (usually '.gitignore') and allows such an untracked 87 but explicitly ignored file to be overwritten. 88 89--index-output=<file>:: 90 Instead of writing the results out to `$GIT_INDEX_FILE`, 91 write the resulting index in the named file. While the 92 command is operating, the original index file is locked 93 with the same mechanism as usual. The file must allow 94 to be rename(2)ed into from a temporary file that is 95 created next to the usual index file; typically this 96 means it needs to be on the same filesystem as the index 97 file itself, and you need write permission to the 98 directories the index file and index output file are 99 located in. 100 101<tree-ish#>:: 102 The id of the tree object(s) to be read/merged. 103 104 105Merging 106------- 107If `-m` is specified, `git-read-tree` can perform 3 kinds of 108merge, a single tree merge if only 1 tree is given, a 109fast-forward merge with 2 trees, or a 3-way merge if 3 trees are 110provided. 111 112 113Single Tree Merge 114~~~~~~~~~~~~~~~~~ 115If only 1 tree is specified, git-read-tree operates as if the user did not 116specify `-m`, except that if the original index has an entry for a 117given pathname, and the contents of the path matches with the tree 118being read, the stat info from the index is used. (In other words, the 119index's stat()s take precedence over the merged tree's). 120 121That means that if you do a `git-read-tree -m <newtree>` followed by a 122`git-checkout-index -f -u -a`, the `git-checkout-index` only checks out 123the stuff that really changed. 124 125This is used to avoid unnecessary false hits when `git-diff-files` is 126run after `git-read-tree`. 127 128 129Two Tree Merge 130~~~~~~~~~~~~~~ 131 132Typically, this is invoked as `git-read-tree -m $H $M`, where $H 133is the head commit of the current repository, and $M is the head 134of a foreign tree, which is simply ahead of $H (i.e. we are in a 135fast forward situation). 136 137When two trees are specified, the user is telling git-read-tree 138the following: 139 140 1. The current index and work tree is derived from $H, but 141 the user may have local changes in them since $H; 142 143 2. The user wants to fast-forward to $M. 144 145In this case, the `git-read-tree -m $H $M` command makes sure 146that no local change is lost as the result of this "merge". 147Here are the "carry forward" rules: 148 149 I (index) H M Result 150 ------------------------------------------------------- 151 0 nothing nothing nothing (does not happen) 152 1 nothing nothing exists use M 153 2 nothing exists nothing remove path from index 154 3 nothing exists exists use M 155 156 clean I==H I==M 157 ------------------ 158 4 yes N/A N/A nothing nothing keep index 159 5 no N/A N/A nothing nothing keep index 160 161 6 yes N/A yes nothing exists keep index 162 7 no N/A yes nothing exists keep index 163 8 yes N/A no nothing exists fail 164 9 no N/A no nothing exists fail 165 166 10 yes yes N/A exists nothing remove path from index 167 11 no yes N/A exists nothing fail 168 12 yes no N/A exists nothing fail 169 13 no no N/A exists nothing fail 170 171 clean (H=M) 172 ------ 173 14 yes exists exists keep index 174 15 no exists exists keep index 175 176 clean I==H I==M (H!=M) 177 ------------------ 178 16 yes no no exists exists fail 179 17 no no no exists exists fail 180 18 yes no yes exists exists keep index 181 19 no no yes exists exists keep index 182 20 yes yes no exists exists use M 183 21 no yes no exists exists fail 184 185In all "keep index" cases, the index entry stays as in the 186original index file. If the entry were not up to date, 187git-read-tree keeps the copy in the work tree intact when 188operating under the -u flag. 189 190When this form of git-read-tree returns successfully, you can 191see what "local changes" you made are carried forward by running 192`git-diff-index --cached $M`. Note that this does not 193necessarily match `git-diff-index --cached $H` would have 194produced before such a two tree merge. This is because of cases 19518 and 19 --- if you already had the changes in $M (e.g. maybe 196you picked it up via e-mail in a patch form), `git-diff-index 197--cached $H` would have told you about the change before this 198merge, but it would not show in `git-diff-index --cached $M` 199output after two-tree merge. 200 201 2023-Way Merge 203~~~~~~~~~~~ 204Each "index" entry has two bits worth of "stage" state. stage 0 is the 205normal one, and is the only one you'd see in any kind of normal use. 206 207However, when you do `git-read-tree` with three trees, the "stage" 208starts out at 1. 209 210This means that you can do 211 212---------------- 213$ git-read-tree -m <tree1> <tree2> <tree3> 214---------------- 215 216and you will end up with an index with all of the <tree1> entries in 217"stage1", all of the <tree2> entries in "stage2" and all of the 218<tree3> entries in "stage3". When performing a merge of another 219branch into the current branch, we use the common ancestor tree 220as <tree1>, the current branch head as <tree2>, and the other 221branch head as <tree3>. 222 223Furthermore, `git-read-tree` has special-case logic that says: if you see 224a file that matches in all respects in the following states, it 225"collapses" back to "stage0": 226 227 - stage 2 and 3 are the same; take one or the other (it makes no 228 difference - the same work has been done on our branch in 229 stage 2 and their branch in stage 3) 230 231 - stage 1 and stage 2 are the same and stage 3 is different; take 232 stage 3 (our branch in stage 2 did not do anything since the 233 ancestor in stage 1 while their branch in stage 3 worked on 234 it) 235 236 - stage 1 and stage 3 are the same and stage 2 is different take 237 stage 2 (we did something while they did nothing) 238 239The `git-write-tree` command refuses to write a nonsensical tree, and it 240will complain about unmerged entries if it sees a single entry that is not 241stage 0. 242 243OK, this all sounds like a collection of totally nonsensical rules, 244but it's actually exactly what you want in order to do a fast 245merge. The different stages represent the "result tree" (stage 0, aka 246"merged"), the original tree (stage 1, aka "orig"), and the two trees 247you are trying to merge (stage 2 and 3 respectively). 248 249The order of stages 1, 2 and 3 (hence the order of three 250<tree-ish> command line arguments) are significant when you 251start a 3-way merge with an index file that is already 252populated. Here is an outline of how the algorithm works: 253 254- if a file exists in identical format in all three trees, it will 255 automatically collapse to "merged" state by git-read-tree. 256 257- a file that has _any_ difference what-so-ever in the three trees 258 will stay as separate entries in the index. It's up to "porcelain 259 policy" to determine how to remove the non-0 stages, and insert a 260 merged version. 261 262- the index file saves and restores with all this information, so you 263 can merge things incrementally, but as long as it has entries in 264 stages 1/2/3 (i.e., "unmerged entries") you can't write the result. So 265 now the merge algorithm ends up being really simple: 266 267 * you walk the index in order, and ignore all entries of stage 0, 268 since they've already been done. 269 270 * if you find a "stage1", but no matching "stage2" or "stage3", you 271 know it's been removed from both trees (it only existed in the 272 original tree), and you remove that entry. 273 274 * if you find a matching "stage2" and "stage3" tree, you remove one 275 of them, and turn the other into a "stage0" entry. Remove any 276 matching "stage1" entry if it exists too. .. all the normal 277 trivial rules .. 278 279You would normally use `git-merge-index` with supplied 280`git-merge-one-file` to do this last step. The script updates 281the files in the working tree as it merges each path and at the 282end of a successful merge. 283 284When you start a 3-way merge with an index file that is already 285populated, it is assumed that it represents the state of the 286files in your work tree, and you can even have files with 287changes unrecorded in the index file. It is further assumed 288that this state is "derived" from the stage 2 tree. The 3-way 289merge refuses to run if it finds an entry in the original index 290file that does not match stage 2. 291 292This is done to prevent you from losing your work-in-progress 293changes, and mixing your random changes in an unrelated merge 294commit. To illustrate, suppose you start from what has been 295committed last to your repository: 296 297---------------- 298$ JC=`git-rev-parse --verify "HEAD^0"` 299$ git-checkout-index -f -u -a $JC 300---------------- 301 302You do random edits, without running git-update-index. And then 303you notice that the tip of your "upstream" tree has advanced 304since you pulled from him: 305 306---------------- 307$ git-fetch git://.... linus 308$ LT=`cat .git/FETCH_HEAD` 309---------------- 310 311Your work tree is still based on your HEAD ($JC), but you have 312some edits since. Three-way merge makes sure that you have not 313added or modified index entries since $JC, and if you haven't, 314then does the right thing. So with the following sequence: 315 316---------------- 317$ git-read-tree -m -u `git-merge-base $JC $LT` $JC $LT 318$ git-merge-index git-merge-one-file -a 319$ echo "Merge with Linus" | \ 320 git-commit-tree `git-write-tree` -p $JC -p $LT 321---------------- 322 323what you would commit is a pure merge between $JC and $LT without 324your work-in-progress changes, and your work tree would be 325updated to the result of the merge. 326 327However, if you have local changes in the working tree that 328would be overwritten by this merge,`git-read-tree` will refuse 329to run to prevent your changes from being lost. 330 331In other words, there is no need to worry about what exists only 332in the working tree. When you have local changes in a part of 333the project that is not involved in the merge, your changes do 334not interfere with the merge, and are kept intact. When they 335*do* interfere, the merge does not even start (`git-read-tree` 336complains loudly and fails without modifying anything). In such 337a case, you can simply continue doing what you were in the 338middle of doing, and when your working tree is ready (i.e. you 339have finished your work-in-progress), attempt the merge again. 340 341 342See Also 343-------- 344gitlink:git-write-tree[1]; gitlink:git-ls-files[1]; 345gitlink:gitignore[5] 346 347 348Author 349------ 350Written by Linus Torvalds <torvalds@osdl.org> 351 352Documentation 353-------------- 354Documentation by David Greaves, Junio C Hamano and the git-list <git@vger.kernel.org>. 355 356GIT 357--- 358Part of the gitlink:git[7] suite 359