]>
Commit | Line | Data |
---|---|---|
1 | git-read-tree(1) | |
2 | ================ | |
3 | ||
4 | NAME | |
5 | ---- | |
6 | git-read-tree - Reads tree information into the index | |
7 | ||
8 | ||
9 | SYNOPSIS | |
10 | -------- | |
11 | [verse] | |
12 | 'git read-tree' [[-m [--trivial] [--aggressive] | --reset | --prefix=<prefix>] | |
13 | [-u [--exclude-per-directory=<gitignore>] | -i]] | |
14 | [--index-output=<file>] [--no-sparse-checkout] | |
15 | (--empty | <tree-ish1> [<tree-ish2> [<tree-ish3>]]) | |
16 | ||
17 | ||
18 | DESCRIPTION | |
19 | ----------- | |
20 | Reads the tree information given by <tree-ish> into the index, | |
21 | but does not actually *update* any of the files it "caches". (see: | |
22 | linkgit:git-checkout-index[1]) | |
23 | ||
24 | Optionally, it can merge a tree into the index, perform a | |
25 | fast-forward (i.e. 2-way) merge, or a 3-way merge, with the `-m` | |
26 | flag. When used with `-m`, the `-u` flag causes it to also update | |
27 | the files in the work tree with the result of the merge. | |
28 | ||
29 | Trivial merges are done by 'git read-tree' itself. Only conflicting paths | |
30 | will be in unmerged state when 'git read-tree' returns. | |
31 | ||
32 | OPTIONS | |
33 | ------- | |
34 | -m:: | |
35 | Perform a merge, not just a read. The command will | |
36 | refuse to run if your index file has unmerged entries, | |
37 | indicating that you have not finished previous merge you | |
38 | started. | |
39 | ||
40 | --reset:: | |
41 | Same as -m, except that unmerged entries are discarded instead | |
42 | of failing. When used with `-u`, updates leading to loss of | |
43 | working tree changes will not abort the operation. | |
44 | ||
45 | -u:: | |
46 | After a successful merge, update the files in the work | |
47 | tree with the result of the merge. | |
48 | ||
49 | -i:: | |
50 | Usually a merge requires the index file as well as the | |
51 | files in the working tree to be up to date with the | |
52 | current head commit, in order not to lose local | |
53 | changes. This flag disables the check with the working | |
54 | tree and is meant to be used when creating a merge of | |
55 | trees that are not directly related to the current | |
56 | working tree status into a temporary index file. | |
57 | ||
58 | -n:: | |
59 | --dry-run:: | |
60 | Check if the command would error out, without updating the index | |
61 | or the files in the working tree for real. | |
62 | ||
63 | -v:: | |
64 | Show the progress of checking files out. | |
65 | ||
66 | --trivial:: | |
67 | Restrict three-way merge by 'git read-tree' to happen | |
68 | only if there is no file-level merging required, instead | |
69 | of resolving merge for trivial cases and leaving | |
70 | conflicting files unresolved in the index. | |
71 | ||
72 | --aggressive:: | |
73 | Usually a three-way merge by 'git read-tree' resolves | |
74 | the merge for really trivial cases and leaves other | |
75 | cases unresolved in the index, so that porcelains can | |
76 | implement different merge policies. This flag makes the | |
77 | command resolve a few more cases internally: | |
78 | + | |
79 | * when one side removes a path and the other side leaves the path | |
80 | unmodified. The resolution is to remove that path. | |
81 | * when both sides remove a path. The resolution is to remove that path. | |
82 | * when both sides add a path identically. The resolution | |
83 | is to add that path. | |
84 | ||
85 | --prefix=<prefix>:: | |
86 | Keep the current index contents, and read the contents | |
87 | of the named tree-ish under the directory at `<prefix>`. | |
88 | The command will refuse to overwrite entries that already | |
89 | existed in the original index file. | |
90 | ||
91 | --exclude-per-directory=<gitignore>:: | |
92 | When running the command with `-u` and `-m` options, the | |
93 | merge result may need to overwrite paths that are not | |
94 | tracked in the current branch. The command usually | |
95 | refuses to proceed with the merge to avoid losing such a | |
96 | path. However this safety valve sometimes gets in the | |
97 | way. For example, it often happens that the other | |
98 | branch added a file that used to be a generated file in | |
99 | your branch, and the safety valve triggers when you try | |
100 | to switch to that branch after you ran `make` but before | |
101 | running `make clean` to remove the generated file. This | |
102 | option tells the command to read per-directory exclude | |
103 | file (usually '.gitignore') and allows such an untracked | |
104 | but explicitly ignored file to be overwritten. | |
105 | ||
106 | --index-output=<file>:: | |
107 | Instead of writing the results out to `$GIT_INDEX_FILE`, | |
108 | write the resulting index in the named file. While the | |
109 | command is operating, the original index file is locked | |
110 | with the same mechanism as usual. The file must allow | |
111 | to be rename(2)ed into from a temporary file that is | |
112 | created next to the usual index file; typically this | |
113 | means it needs to be on the same filesystem as the index | |
114 | file itself, and you need write permission to the | |
115 | directories the index file and index output file are | |
116 | located in. | |
117 | ||
118 | --[no-]recurse-submodules:: | |
119 | Using --recurse-submodules will update the content of all active | |
120 | submodules according to the commit recorded in the superproject by | |
121 | calling read-tree recursively, also setting the submodules' HEAD to be | |
122 | detached at that commit. | |
123 | ||
124 | --no-sparse-checkout:: | |
125 | Disable sparse checkout support even if `core.sparseCheckout` | |
126 | is true. | |
127 | ||
128 | --empty:: | |
129 | Instead of reading tree object(s) into the index, just empty | |
130 | it. | |
131 | ||
132 | -q:: | |
133 | --quiet:: | |
134 | Quiet, suppress feedback messages. | |
135 | ||
136 | <tree-ish#>:: | |
137 | The id of the tree object(s) to be read/merged. | |
138 | ||
139 | ||
140 | MERGING | |
141 | ------- | |
142 | If `-m` is specified, 'git read-tree' can perform 3 kinds of | |
143 | merge, a single tree merge if only 1 tree is given, a | |
144 | fast-forward merge with 2 trees, or a 3-way merge if 3 or more trees are | |
145 | provided. | |
146 | ||
147 | ||
148 | Single Tree Merge | |
149 | ~~~~~~~~~~~~~~~~~ | |
150 | If only 1 tree is specified, 'git read-tree' operates as if the user did not | |
151 | specify `-m`, except that if the original index has an entry for a | |
152 | given pathname, and the contents of the path match with the tree | |
153 | being read, the stat info from the index is used. (In other words, the | |
154 | index's stat()s take precedence over the merged tree's). | |
155 | ||
156 | That means that if you do a `git read-tree -m <newtree>` followed by a | |
157 | `git checkout-index -f -u -a`, the 'git checkout-index' only checks out | |
158 | the stuff that really changed. | |
159 | ||
160 | This is used to avoid unnecessary false hits when 'git diff-files' is | |
161 | run after 'git read-tree'. | |
162 | ||
163 | ||
164 | Two Tree Merge | |
165 | ~~~~~~~~~~~~~~ | |
166 | ||
167 | Typically, this is invoked as `git read-tree -m $H $M`, where $H | |
168 | is the head commit of the current repository, and $M is the head | |
169 | of a foreign tree, which is simply ahead of $H (i.e. we are in a | |
170 | fast-forward situation). | |
171 | ||
172 | When two trees are specified, the user is telling 'git read-tree' | |
173 | the following: | |
174 | ||
175 | 1. The current index and work tree is derived from $H, but | |
176 | the user may have local changes in them since $H. | |
177 | ||
178 | 2. The user wants to fast-forward to $M. | |
179 | ||
180 | In this case, the `git read-tree -m $H $M` command makes sure | |
181 | that no local change is lost as the result of this "merge". | |
182 | Here are the "carry forward" rules, where "I" denotes the index, | |
183 | "clean" means that index and work tree coincide, and "exists"/"nothing" | |
184 | refer to the presence of a path in the specified commit: | |
185 | ||
186 | .... | |
187 | I H M Result | |
188 | ------------------------------------------------------- | |
189 | 0 nothing nothing nothing (does not happen) | |
190 | 1 nothing nothing exists use M | |
191 | 2 nothing exists nothing remove path from index | |
192 | 3 nothing exists exists, use M if "initial checkout", | |
193 | H == M keep index otherwise | |
194 | exists, fail | |
195 | H != M | |
196 | ||
197 | clean I==H I==M | |
198 | ------------------ | |
199 | 4 yes N/A N/A nothing nothing keep index | |
200 | 5 no N/A N/A nothing nothing keep index | |
201 | ||
202 | 6 yes N/A yes nothing exists keep index | |
203 | 7 no N/A yes nothing exists keep index | |
204 | 8 yes N/A no nothing exists fail | |
205 | 9 no N/A no nothing exists fail | |
206 | ||
207 | 10 yes yes N/A exists nothing remove path from index | |
208 | 11 no yes N/A exists nothing fail | |
209 | 12 yes no N/A exists nothing fail | |
210 | 13 no no N/A exists nothing fail | |
211 | ||
212 | clean (H==M) | |
213 | ------ | |
214 | 14 yes exists exists keep index | |
215 | 15 no exists exists keep index | |
216 | ||
217 | clean I==H I==M (H!=M) | |
218 | ------------------ | |
219 | 16 yes no no exists exists fail | |
220 | 17 no no no exists exists fail | |
221 | 18 yes no yes exists exists keep index | |
222 | 19 no no yes exists exists keep index | |
223 | 20 yes yes no exists exists use M | |
224 | 21 no yes no exists exists fail | |
225 | .... | |
226 | ||
227 | In all "keep index" cases, the index entry stays as in the | |
228 | original index file. If the entry is not up to date, | |
229 | 'git read-tree' keeps the copy in the work tree intact when | |
230 | operating under the -u flag. | |
231 | ||
232 | When this form of 'git read-tree' returns successfully, you can | |
233 | see which of the "local changes" that you made were carried forward by running | |
234 | `git diff-index --cached $M`. Note that this does not | |
235 | necessarily match what `git diff-index --cached $H` would have | |
236 | produced before such a two tree merge. This is because of cases | |
237 | 18 and 19 --- if you already had the changes in $M (e.g. maybe | |
238 | you picked it up via e-mail in a patch form), `git diff-index | |
239 | --cached $H` would have told you about the change before this | |
240 | merge, but it would not show in `git diff-index --cached $M` | |
241 | output after the two-tree merge. | |
242 | ||
243 | Case 3 is slightly tricky and needs explanation. The result from this | |
244 | rule logically should be to remove the path if the user staged the removal | |
245 | of the path and then switching to a new branch. That however will prevent | |
246 | the initial checkout from happening, so the rule is modified to use M (new | |
247 | tree) only when the content of the index is empty. Otherwise the removal | |
248 | of the path is kept as long as $H and $M are the same. | |
249 | ||
250 | 3-Way Merge | |
251 | ~~~~~~~~~~~ | |
252 | Each "index" entry has two bits worth of "stage" state. stage 0 is the | |
253 | normal one, and is the only one you'd see in any kind of normal use. | |
254 | ||
255 | However, when you do 'git read-tree' with three trees, the "stage" | |
256 | starts out at 1. | |
257 | ||
258 | This means that you can do | |
259 | ||
260 | ---------------- | |
261 | $ git read-tree -m <tree1> <tree2> <tree3> | |
262 | ---------------- | |
263 | ||
264 | and you will end up with an index with all of the <tree1> entries in | |
265 | "stage1", all of the <tree2> entries in "stage2" and all of the | |
266 | <tree3> entries in "stage3". When performing a merge of another | |
267 | branch into the current branch, we use the common ancestor tree | |
268 | as <tree1>, the current branch head as <tree2>, and the other | |
269 | branch head as <tree3>. | |
270 | ||
271 | Furthermore, 'git read-tree' has special-case logic that says: if you see | |
272 | a file that matches in all respects in the following states, it | |
273 | "collapses" back to "stage0": | |
274 | ||
275 | - stage 2 and 3 are the same; take one or the other (it makes no | |
276 | difference - the same work has been done on our branch in | |
277 | stage 2 and their branch in stage 3) | |
278 | ||
279 | - stage 1 and stage 2 are the same and stage 3 is different; take | |
280 | stage 3 (our branch in stage 2 did not do anything since the | |
281 | ancestor in stage 1 while their branch in stage 3 worked on | |
282 | it) | |
283 | ||
284 | - stage 1 and stage 3 are the same and stage 2 is different take | |
285 | stage 2 (we did something while they did nothing) | |
286 | ||
287 | The 'git write-tree' command refuses to write a nonsensical tree, and it | |
288 | will complain about unmerged entries if it sees a single entry that is not | |
289 | stage 0. | |
290 | ||
291 | OK, this all sounds like a collection of totally nonsensical rules, | |
292 | but it's actually exactly what you want in order to do a fast | |
293 | merge. The different stages represent the "result tree" (stage 0, aka | |
294 | "merged"), the original tree (stage 1, aka "orig"), and the two trees | |
295 | you are trying to merge (stage 2 and 3 respectively). | |
296 | ||
297 | The order of stages 1, 2 and 3 (hence the order of three | |
298 | <tree-ish> command-line arguments) are significant when you | |
299 | start a 3-way merge with an index file that is already | |
300 | populated. Here is an outline of how the algorithm works: | |
301 | ||
302 | - if a file exists in identical format in all three trees, it will | |
303 | automatically collapse to "merged" state by 'git read-tree'. | |
304 | ||
305 | - a file that has _any_ difference what-so-ever in the three trees | |
306 | will stay as separate entries in the index. It's up to "porcelain | |
307 | policy" to determine how to remove the non-0 stages, and insert a | |
308 | merged version. | |
309 | ||
310 | - the index file saves and restores with all this information, so you | |
311 | can merge things incrementally, but as long as it has entries in | |
312 | stages 1/2/3 (i.e., "unmerged entries") you can't write the result. So | |
313 | now the merge algorithm ends up being really simple: | |
314 | ||
315 | * you walk the index in order, and ignore all entries of stage 0, | |
316 | since they've already been done. | |
317 | ||
318 | * if you find a "stage1", but no matching "stage2" or "stage3", you | |
319 | know it's been removed from both trees (it only existed in the | |
320 | original tree), and you remove that entry. | |
321 | ||
322 | * if you find a matching "stage2" and "stage3" tree, you remove one | |
323 | of them, and turn the other into a "stage0" entry. Remove any | |
324 | matching "stage1" entry if it exists too. .. all the normal | |
325 | trivial rules .. | |
326 | ||
327 | You would normally use 'git merge-index' with supplied | |
328 | 'git merge-one-file' to do this last step. The script updates | |
329 | the files in the working tree as it merges each path and at the | |
330 | end of a successful merge. | |
331 | ||
332 | When you start a 3-way merge with an index file that is already | |
333 | populated, it is assumed that it represents the state of the | |
334 | files in your work tree, and you can even have files with | |
335 | changes unrecorded in the index file. It is further assumed | |
336 | that this state is "derived" from the stage 2 tree. The 3-way | |
337 | merge refuses to run if it finds an entry in the original index | |
338 | file that does not match stage 2. | |
339 | ||
340 | This is done to prevent you from losing your work-in-progress | |
341 | changes, and mixing your random changes in an unrelated merge | |
342 | commit. To illustrate, suppose you start from what has been | |
343 | committed last to your repository: | |
344 | ||
345 | ---------------- | |
346 | $ JC=`git rev-parse --verify "HEAD^0"` | |
347 | $ git checkout-index -f -u -a $JC | |
348 | ---------------- | |
349 | ||
350 | You do random edits, without running 'git update-index'. And then | |
351 | you notice that the tip of your "upstream" tree has advanced | |
352 | since you pulled from him: | |
353 | ||
354 | ---------------- | |
355 | $ git fetch git://.... linus | |
356 | $ LT=`git rev-parse FETCH_HEAD` | |
357 | ---------------- | |
358 | ||
359 | Your work tree is still based on your HEAD ($JC), but you have | |
360 | some edits since. Three-way merge makes sure that you have not | |
361 | added or modified index entries since $JC, and if you haven't, | |
362 | then does the right thing. So with the following sequence: | |
363 | ||
364 | ---------------- | |
365 | $ git read-tree -m -u `git merge-base $JC $LT` $JC $LT | |
366 | $ git merge-index git-merge-one-file -a | |
367 | $ echo "Merge with Linus" | \ | |
368 | git commit-tree `git write-tree` -p $JC -p $LT | |
369 | ---------------- | |
370 | ||
371 | what you would commit is a pure merge between $JC and $LT without | |
372 | your work-in-progress changes, and your work tree would be | |
373 | updated to the result of the merge. | |
374 | ||
375 | However, if you have local changes in the working tree that | |
376 | would be overwritten by this merge, 'git read-tree' will refuse | |
377 | to run to prevent your changes from being lost. | |
378 | ||
379 | In other words, there is no need to worry about what exists only | |
380 | in the working tree. When you have local changes in a part of | |
381 | the project that is not involved in the merge, your changes do | |
382 | not interfere with the merge, and are kept intact. When they | |
383 | *do* interfere, the merge does not even start ('git read-tree' | |
384 | complains loudly and fails without modifying anything). In such | |
385 | a case, you can simply continue doing what you were in the | |
386 | middle of doing, and when your working tree is ready (i.e. you | |
387 | have finished your work-in-progress), attempt the merge again. | |
388 | ||
389 | ||
390 | SPARSE CHECKOUT | |
391 | --------------- | |
392 | ||
393 | "Sparse checkout" allows populating the working directory sparsely. | |
394 | It uses the skip-worktree bit (see linkgit:git-update-index[1]) to tell | |
395 | Git whether a file in the working directory is worth looking at. | |
396 | ||
397 | 'git read-tree' and other merge-based commands ('git merge', 'git | |
398 | checkout'...) can help maintaining the skip-worktree bitmap and working | |
399 | directory update. `$GIT_DIR/info/sparse-checkout` is used to | |
400 | define the skip-worktree reference bitmap. When 'git read-tree' needs | |
401 | to update the working directory, it resets the skip-worktree bit in the index | |
402 | based on this file, which uses the same syntax as .gitignore files. | |
403 | If an entry matches a pattern in this file, skip-worktree will not be | |
404 | set on that entry. Otherwise, skip-worktree will be set. | |
405 | ||
406 | Then it compares the new skip-worktree value with the previous one. If | |
407 | skip-worktree turns from set to unset, it will add the corresponding | |
408 | file back. If it turns from unset to set, that file will be removed. | |
409 | ||
410 | While `$GIT_DIR/info/sparse-checkout` is usually used to specify what | |
411 | files are in, you can also specify what files are _not_ in, using | |
412 | negate patterns. For example, to remove the file `unwanted`: | |
413 | ||
414 | ---------------- | |
415 | /* | |
416 | !unwanted | |
417 | ---------------- | |
418 | ||
419 | Another tricky thing is fully repopulating the working directory when you | |
420 | no longer want sparse checkout. You cannot just disable "sparse | |
421 | checkout" because skip-worktree bits are still in the index and your working | |
422 | directory is still sparsely populated. You should re-populate the working | |
423 | directory with the `$GIT_DIR/info/sparse-checkout` file content as | |
424 | follows: | |
425 | ||
426 | ---------------- | |
427 | /* | |
428 | ---------------- | |
429 | ||
430 | Then you can disable sparse checkout. Sparse checkout support in 'git | |
431 | read-tree' and similar commands is disabled by default. You need to | |
432 | turn `core.sparseCheckout` on in order to have sparse checkout | |
433 | support. | |
434 | ||
435 | ||
436 | SEE ALSO | |
437 | -------- | |
438 | linkgit:git-write-tree[1]; linkgit:git-ls-files[1]; | |
439 | linkgit:gitignore[5]; linkgit:git-sparse-checkout[1]; | |
440 | ||
441 | GIT | |
442 | --- | |
443 | Part of the linkgit:git[1] suite |