user-manual: Update for receive.denyCurrentBranch=refuse
[git/git.git] / Documentation / user-manual.txt
1 Git User's Manual (for version 1.5.3 or newer)
2 ______________________________________________
3
4
5 Git is a fast distributed revision control system.
6
7 This manual is designed to be readable by someone with basic UNIX
8 command-line skills, but no previous knowledge of git.
9
10 <<repositories-and-branches>> and <<exploring-git-history>> explain how
11 to fetch and study a project using git--read these chapters to learn how
12 to build and test a particular version of a software project, search for
13 regressions, and so on.
14
15 People needing to do actual development will also want to read
16 <<Developing-With-git>> and <<sharing-development>>.
17
18 Further chapters cover more specialized topics.
19
20 Comprehensive reference documentation is available through the man
21 pages, or linkgit:git-help[1] command. For example, for the command
22 "git clone <repo>", you can either use:
23
24 ------------------------------------------------
25 $ man git-clone
26 ------------------------------------------------
27
28 or:
29
30 ------------------------------------------------
31 $ git help clone
32 ------------------------------------------------
33
34 With the latter, you can use the manual viewer of your choice; see
35 linkgit:git-help[1] for more information.
36
37 See also <<git-quick-start>> for a brief overview of git commands,
38 without any explanation.
39
40 Finally, see <<todo>> for ways that you can help make this manual more
41 complete.
42
43
44 [[repositories-and-branches]]
45 Repositories and Branches
46 =========================
47
48 [[how-to-get-a-git-repository]]
49 How to get a git repository
50 ---------------------------
51
52 It will be useful to have a git repository to experiment with as you
53 read this manual.
54
55 The best way to get one is by using the linkgit:git-clone[1] command to
56 download a copy of an existing repository. If you don't already have a
57 project in mind, here are some interesting examples:
58
59 ------------------------------------------------
60 # git itself (approx. 10MB download):
61 $ git clone git://git.kernel.org/pub/scm/git/git.git
62 # the Linux kernel (approx. 150MB download):
63 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
64 ------------------------------------------------
65
66 The initial clone may be time-consuming for a large project, but you
67 will only need to clone once.
68
69 The clone command creates a new directory named after the project ("git"
70 or "linux-2.6" in the examples above). After you cd into this
71 directory, you will see that it contains a copy of the project files,
72 called the <<def_working_tree,working tree>>, together with a special
73 top-level directory named ".git", which contains all the information
74 about the history of the project.
75
76 [[how-to-check-out]]
77 How to check out a different version of a project
78 -------------------------------------------------
79
80 Git is best thought of as a tool for storing the history of a collection
81 of files. It stores the history as a compressed collection of
82 interrelated snapshots of the project's contents. In git each such
83 version is called a <<def_commit,commit>>.
84
85 Those snapshots aren't necessarily all arranged in a single line from
86 oldest to newest; instead, work may simultaneously proceed along
87 parallel lines of development, called <<def_branch,branches>>, which may
88 merge and diverge.
89
90 A single git repository can track development on multiple branches. It
91 does this by keeping a list of <<def_head,heads>> which reference the
92 latest commit on each branch; the linkgit:git-branch[1] command shows
93 you the list of branch heads:
94
95 ------------------------------------------------
96 $ git branch
97 * master
98 ------------------------------------------------
99
100 A freshly cloned repository contains a single branch head, by default
101 named "master", with the working directory initialized to the state of
102 the project referred to by that branch head.
103
104 Most projects also use <<def_tag,tags>>. Tags, like heads, are
105 references into the project's history, and can be listed using the
106 linkgit:git-tag[1] command:
107
108 ------------------------------------------------
109 $ git tag -l
110 v2.6.11
111 v2.6.11-tree
112 v2.6.12
113 v2.6.12-rc2
114 v2.6.12-rc3
115 v2.6.12-rc4
116 v2.6.12-rc5
117 v2.6.12-rc6
118 v2.6.13
119 ...
120 ------------------------------------------------
121
122 Tags are expected to always point at the same version of a project,
123 while heads are expected to advance as development progresses.
124
125 Create a new branch head pointing to one of these versions and check it
126 out using linkgit:git-checkout[1]:
127
128 ------------------------------------------------
129 $ git checkout -b new v2.6.13
130 ------------------------------------------------
131
132 The working directory then reflects the contents that the project had
133 when it was tagged v2.6.13, and linkgit:git-branch[1] shows two
134 branches, with an asterisk marking the currently checked-out branch:
135
136 ------------------------------------------------
137 $ git branch
138 master
139 * new
140 ------------------------------------------------
141
142 If you decide that you'd rather see version 2.6.17, you can modify
143 the current branch to point at v2.6.17 instead, with
144
145 ------------------------------------------------
146 $ git reset --hard v2.6.17
147 ------------------------------------------------
148
149 Note that if the current branch head was your only reference to a
150 particular point in history, then resetting that branch may leave you
151 with no way to find the history it used to point to; so use this command
152 carefully.
153
154 [[understanding-commits]]
155 Understanding History: Commits
156 ------------------------------
157
158 Every change in the history of a project is represented by a commit.
159 The linkgit:git-show[1] command shows the most recent commit on the
160 current branch:
161
162 ------------------------------------------------
163 $ git show
164 commit 17cf781661e6d38f737f15f53ab552f1e95960d7
165 Author: Linus Torvalds <torvalds@ppc970.osdl.org.(none)>
166 Date: Tue Apr 19 14:11:06 2005 -0700
167
168 Remove duplicate getenv(DB_ENVIRONMENT) call
169
170 Noted by Tony Luck.
171
172 diff --git a/init-db.c b/init-db.c
173 index 65898fa..b002dc6 100644
174 --- a/init-db.c
175 +++ b/init-db.c
176 @@ -7,7 +7,7 @@
177
178 int main(int argc, char **argv)
179 {
180 - char *sha1_dir = getenv(DB_ENVIRONMENT), *path;
181 + char *sha1_dir, *path;
182 int len, i;
183
184 if (mkdir(".git", 0755) < 0) {
185 ------------------------------------------------
186
187 As you can see, a commit shows who made the latest change, what they
188 did, and why.
189
190 Every commit has a 40-hexdigit id, sometimes called the "object name" or the
191 "SHA-1 id", shown on the first line of the "git show" output. You can usually
192 refer to a commit by a shorter name, such as a tag or a branch name, but this
193 longer name can also be useful. Most importantly, it is a globally unique
194 name for this commit: so if you tell somebody else the object name (for
195 example in email), then you are guaranteed that name will refer to the same
196 commit in their repository that it does in yours (assuming their repository
197 has that commit at all). Since the object name is computed as a hash over the
198 contents of the commit, you are guaranteed that the commit can never change
199 without its name also changing.
200
201 In fact, in <<git-concepts>> we shall see that everything stored in git
202 history, including file data and directory contents, is stored in an object
203 with a name that is a hash of its contents.
204
205 [[understanding-reachability]]
206 Understanding history: commits, parents, and reachability
207 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
208
209 Every commit (except the very first commit in a project) also has a
210 parent commit which shows what happened before this commit.
211 Following the chain of parents will eventually take you back to the
212 beginning of the project.
213
214 However, the commits do not form a simple list; git allows lines of
215 development to diverge and then reconverge, and the point where two
216 lines of development reconverge is called a "merge". The commit
217 representing a merge can therefore have more than one parent, with
218 each parent representing the most recent commit on one of the lines
219 of development leading to that point.
220
221 The best way to see how this works is using the linkgit:gitk[1]
222 command; running gitk now on a git repository and looking for merge
223 commits will help understand how the git organizes history.
224
225 In the following, we say that commit X is "reachable" from commit Y
226 if commit X is an ancestor of commit Y. Equivalently, you could say
227 that Y is a descendant of X, or that there is a chain of parents
228 leading from commit Y to commit X.
229
230 [[history-diagrams]]
231 Understanding history: History diagrams
232 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
233
234 We will sometimes represent git history using diagrams like the one
235 below. Commits are shown as "o", and the links between them with
236 lines drawn with - / and \. Time goes left to right:
237
238
239 ................................................
240 o--o--o <-- Branch A
241 /
242 o--o--o <-- master
243 \
244 o--o--o <-- Branch B
245 ................................................
246
247 If we need to talk about a particular commit, the character "o" may
248 be replaced with another letter or number.
249
250 [[what-is-a-branch]]
251 Understanding history: What is a branch?
252 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
253
254 When we need to be precise, we will use the word "branch" to mean a line
255 of development, and "branch head" (or just "head") to mean a reference
256 to the most recent commit on a branch. In the example above, the branch
257 head named "A" is a pointer to one particular commit, but we refer to
258 the line of three commits leading up to that point as all being part of
259 "branch A".
260
261 However, when no confusion will result, we often just use the term
262 "branch" both for branches and for branch heads.
263
264 [[manipulating-branches]]
265 Manipulating branches
266 ---------------------
267
268 Creating, deleting, and modifying branches is quick and easy; here's
269 a summary of the commands:
270
271 git branch::
272 list all branches
273 git branch <branch>::
274 create a new branch named <branch>, referencing the same
275 point in history as the current branch
276 git branch <branch> <start-point>::
277 create a new branch named <branch>, referencing
278 <start-point>, which may be specified any way you like,
279 including using a branch name or a tag name
280 git branch -d <branch>::
281 delete the branch <branch>; if the branch you are deleting
282 points to a commit which is not reachable from the current
283 branch, this command will fail with a warning.
284 git branch -D <branch>::
285 even if the branch points to a commit not reachable
286 from the current branch, you may know that that commit
287 is still reachable from some other branch or tag. In that
288 case it is safe to use this command to force git to delete
289 the branch.
290 git checkout <branch>::
291 make the current branch <branch>, updating the working
292 directory to reflect the version referenced by <branch>
293 git checkout -b <new> <start-point>::
294 create a new branch <new> referencing <start-point>, and
295 check it out.
296
297 The special symbol "HEAD" can always be used to refer to the current
298 branch. In fact, git uses a file named "HEAD" in the .git directory to
299 remember which branch is current:
300
301 ------------------------------------------------
302 $ cat .git/HEAD
303 ref: refs/heads/master
304 ------------------------------------------------
305
306 [[detached-head]]
307 Examining an old version without creating a new branch
308 ------------------------------------------------------
309
310 The `git checkout` command normally expects a branch head, but will also
311 accept an arbitrary commit; for example, you can check out the commit
312 referenced by a tag:
313
314 ------------------------------------------------
315 $ git checkout v2.6.17
316 Note: moving to "v2.6.17" which isn't a local branch
317 If you want to create a new branch from this checkout, you may do so
318 (now or later) by using -b with the checkout command again. Example:
319 git checkout -b <new_branch_name>
320 HEAD is now at 427abfa... Linux v2.6.17
321 ------------------------------------------------
322
323 The HEAD then refers to the SHA-1 of the commit instead of to a branch,
324 and git branch shows that you are no longer on a branch:
325
326 ------------------------------------------------
327 $ cat .git/HEAD
328 427abfa28afedffadfca9dd8b067eb6d36bac53f
329 $ git branch
330 * (no branch)
331 master
332 ------------------------------------------------
333
334 In this case we say that the HEAD is "detached".
335
336 This is an easy way to check out a particular version without having to
337 make up a name for the new branch. You can still create a new branch
338 (or tag) for this version later if you decide to.
339
340 [[examining-remote-branches]]
341 Examining branches from a remote repository
342 -------------------------------------------
343
344 The "master" branch that was created at the time you cloned is a copy
345 of the HEAD in the repository that you cloned from. That repository
346 may also have had other branches, though, and your local repository
347 keeps branches which track each of those remote branches, called
348 remote-tracking branches, which you
349 can view using the "-r" option to linkgit:git-branch[1]:
350
351 ------------------------------------------------
352 $ git branch -r
353 origin/HEAD
354 origin/html
355 origin/maint
356 origin/man
357 origin/master
358 origin/next
359 origin/pu
360 origin/todo
361 ------------------------------------------------
362
363 In this example, "origin" is called a remote repository, or "remote"
364 for short. The branches of this repository are called "remote
365 branches" from our point of view. The remote-tracking branches listed
366 above were created based on the remote branches at clone time and will
367 be updated by "git fetch" (hence "git pull") and "git push". See
368 <<Updating-a-repository-With-git-fetch>> for details.
369
370 You might want to build on one of these remote-tracking branches
371 on a branch of your own, just as you would for a tag:
372
373 ------------------------------------------------
374 $ git checkout -b my-todo-copy origin/todo
375 ------------------------------------------------
376
377 You can also check out "origin/todo" directly to examine it or
378 write a one-off patch. See <<detached-head,detached head>>.
379
380 Note that the name "origin" is just the name that git uses by default
381 to refer to the repository that you cloned from.
382
383 [[how-git-stores-references]]
384 Naming branches, tags, and other references
385 -------------------------------------------
386
387 Branches, remote-tracking branches, and tags are all references to
388 commits. All references are named with a slash-separated path name
389 starting with "refs"; the names we've been using so far are actually
390 shorthand:
391
392 - The branch "test" is short for "refs/heads/test".
393 - The tag "v2.6.18" is short for "refs/tags/v2.6.18".
394 - "origin/master" is short for "refs/remotes/origin/master".
395
396 The full name is occasionally useful if, for example, there ever
397 exists a tag and a branch with the same name.
398
399 (Newly created refs are actually stored in the .git/refs directory,
400 under the path given by their name. However, for efficiency reasons
401 they may also be packed together in a single file; see
402 linkgit:git-pack-refs[1]).
403
404 As another useful shortcut, the "HEAD" of a repository can be referred
405 to just using the name of that repository. So, for example, "origin"
406 is usually a shortcut for the HEAD branch in the repository "origin".
407
408 For the complete list of paths which git checks for references, and
409 the order it uses to decide which to choose when there are multiple
410 references with the same shorthand name, see the "SPECIFYING
411 REVISIONS" section of linkgit:gitrevisions[7].
412
413 [[Updating-a-repository-With-git-fetch]]
414 Updating a repository with git fetch
415 ------------------------------------
416
417 Eventually the developer cloned from will do additional work in her
418 repository, creating new commits and advancing the branches to point
419 at the new commits.
420
421 The command "git fetch", with no arguments, will update all of the
422 remote-tracking branches to the latest version found in her
423 repository. It will not touch any of your own branches--not even the
424 "master" branch that was created for you on clone.
425
426 [[fetching-branches]]
427 Fetching branches from other repositories
428 -----------------------------------------
429
430 You can also track branches from repositories other than the one you
431 cloned from, using linkgit:git-remote[1]:
432
433 -------------------------------------------------
434 $ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
435 $ git fetch linux-nfs
436 * refs/remotes/linux-nfs/master: storing branch 'master' ...
437 commit: bf81b46
438 -------------------------------------------------
439
440 New remote-tracking branches will be stored under the shorthand name
441 that you gave "git remote add", in this case linux-nfs:
442
443 -------------------------------------------------
444 $ git branch -r
445 linux-nfs/master
446 origin/master
447 -------------------------------------------------
448
449 If you run "git fetch <remote>" later, the remote-tracking branches for the
450 named <remote> will be updated.
451
452 If you examine the file .git/config, you will see that git has added
453 a new stanza:
454
455 -------------------------------------------------
456 $ cat .git/config
457 ...
458 [remote "linux-nfs"]
459 url = git://linux-nfs.org/pub/nfs-2.6.git
460 fetch = +refs/heads/*:refs/remotes/linux-nfs/*
461 ...
462 -------------------------------------------------
463
464 This is what causes git to track the remote's branches; you may modify
465 or delete these configuration options by editing .git/config with a
466 text editor. (See the "CONFIGURATION FILE" section of
467 linkgit:git-config[1] for details.)
468
469 [[exploring-git-history]]
470 Exploring git history
471 =====================
472
473 Git is best thought of as a tool for storing the history of a
474 collection of files. It does this by storing compressed snapshots of
475 the contents of a file hierarchy, together with "commits" which show
476 the relationships between these snapshots.
477
478 Git provides extremely flexible and fast tools for exploring the
479 history of a project.
480
481 We start with one specialized tool that is useful for finding the
482 commit that introduced a bug into a project.
483
484 [[using-bisect]]
485 How to use bisect to find a regression
486 --------------------------------------
487
488 Suppose version 2.6.18 of your project worked, but the version at
489 "master" crashes. Sometimes the best way to find the cause of such a
490 regression is to perform a brute-force search through the project's
491 history to find the particular commit that caused the problem. The
492 linkgit:git-bisect[1] command can help you do this:
493
494 -------------------------------------------------
495 $ git bisect start
496 $ git bisect good v2.6.18
497 $ git bisect bad master
498 Bisecting: 3537 revisions left to test after this
499 [65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
500 -------------------------------------------------
501
502 If you run "git branch" at this point, you'll see that git has
503 temporarily moved you in "(no branch)". HEAD is now detached from any
504 branch and points directly to a commit (with commit id 65934...) that
505 is reachable from "master" but not from v2.6.18. Compile and test it,
506 and see whether it crashes. Assume it does crash. Then:
507
508 -------------------------------------------------
509 $ git bisect bad
510 Bisecting: 1769 revisions left to test after this
511 [7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
512 -------------------------------------------------
513
514 checks out an older version. Continue like this, telling git at each
515 stage whether the version it gives you is good or bad, and notice
516 that the number of revisions left to test is cut approximately in
517 half each time.
518
519 After about 13 tests (in this case), it will output the commit id of
520 the guilty commit. You can then examine the commit with
521 linkgit:git-show[1], find out who wrote it, and mail them your bug
522 report with the commit id. Finally, run
523
524 -------------------------------------------------
525 $ git bisect reset
526 -------------------------------------------------
527
528 to return you to the branch you were on before.
529
530 Note that the version which `git bisect` checks out for you at each
531 point is just a suggestion, and you're free to try a different
532 version if you think it would be a good idea. For example,
533 occasionally you may land on a commit that broke something unrelated;
534 run
535
536 -------------------------------------------------
537 $ git bisect visualize
538 -------------------------------------------------
539
540 which will run gitk and label the commit it chose with a marker that
541 says "bisect". Choose a safe-looking commit nearby, note its commit
542 id, and check it out with:
543
544 -------------------------------------------------
545 $ git reset --hard fb47ddb2db...
546 -------------------------------------------------
547
548 then test, run "bisect good" or "bisect bad" as appropriate, and
549 continue.
550
551 Instead of "git bisect visualize" and then "git reset --hard
552 fb47ddb2db...", you might just want to tell git that you want to skip
553 the current commit:
554
555 -------------------------------------------------
556 $ git bisect skip
557 -------------------------------------------------
558
559 In this case, though, git may not eventually be able to tell the first
560 bad one between some first skipped commits and a later bad commit.
561
562 There are also ways to automate the bisecting process if you have a
563 test script that can tell a good from a bad commit. See
564 linkgit:git-bisect[1] for more information about this and other "git
565 bisect" features.
566
567 [[naming-commits]]
568 Naming commits
569 --------------
570
571 We have seen several ways of naming commits already:
572
573 - 40-hexdigit object name
574 - branch name: refers to the commit at the head of the given
575 branch
576 - tag name: refers to the commit pointed to by the given tag
577 (we've seen branches and tags are special cases of
578 <<how-git-stores-references,references>>).
579 - HEAD: refers to the head of the current branch
580
581 There are many more; see the "SPECIFYING REVISIONS" section of the
582 linkgit:gitrevisions[7] man page for the complete list of ways to
583 name revisions. Some examples:
584
585 -------------------------------------------------
586 $ git show fb47ddb2 # the first few characters of the object name
587 # are usually enough to specify it uniquely
588 $ git show HEAD^ # the parent of the HEAD commit
589 $ git show HEAD^^ # the grandparent
590 $ git show HEAD~4 # the great-great-grandparent
591 -------------------------------------------------
592
593 Recall that merge commits may have more than one parent; by default,
594 ^ and ~ follow the first parent listed in the commit, but you can
595 also choose:
596
597 -------------------------------------------------
598 $ git show HEAD^1 # show the first parent of HEAD
599 $ git show HEAD^2 # show the second parent of HEAD
600 -------------------------------------------------
601
602 In addition to HEAD, there are several other special names for
603 commits:
604
605 Merges (to be discussed later), as well as operations such as
606 `git reset`, which change the currently checked-out commit, generally
607 set ORIG_HEAD to the value HEAD had before the current operation.
608
609 The `git fetch` operation always stores the head of the last fetched
610 branch in FETCH_HEAD. For example, if you run `git fetch` without
611 specifying a local branch as the target of the operation
612
613 -------------------------------------------------
614 $ git fetch git://example.com/proj.git theirbranch
615 -------------------------------------------------
616
617 the fetched commits will still be available from FETCH_HEAD.
618
619 When we discuss merges we'll also see the special name MERGE_HEAD,
620 which refers to the other branch that we're merging in to the current
621 branch.
622
623 The linkgit:git-rev-parse[1] command is a low-level command that is
624 occasionally useful for translating some name for a commit to the object
625 name for that commit:
626
627 -------------------------------------------------
628 $ git rev-parse origin
629 e05db0fd4f31dde7005f075a84f96b360d05984b
630 -------------------------------------------------
631
632 [[creating-tags]]
633 Creating tags
634 -------------
635
636 We can also create a tag to refer to a particular commit; after
637 running
638
639 -------------------------------------------------
640 $ git tag stable-1 1b2e1d63ff
641 -------------------------------------------------
642
643 You can use stable-1 to refer to the commit 1b2e1d63ff.
644
645 This creates a "lightweight" tag. If you would also like to include a
646 comment with the tag, and possibly sign it cryptographically, then you
647 should create a tag object instead; see the linkgit:git-tag[1] man page
648 for details.
649
650 [[browsing-revisions]]
651 Browsing revisions
652 ------------------
653
654 The linkgit:git-log[1] command can show lists of commits. On its
655 own, it shows all commits reachable from the parent commit; but you
656 can also make more specific requests:
657
658 -------------------------------------------------
659 $ git log v2.5.. # commits since (not reachable from) v2.5
660 $ git log test..master # commits reachable from master but not test
661 $ git log master..test # ...reachable from test but not master
662 $ git log master...test # ...reachable from either test or master,
663 # but not both
664 $ git log --since="2 weeks ago" # commits from the last 2 weeks
665 $ git log Makefile # commits which modify Makefile
666 $ git log fs/ # ... which modify any file under fs/
667 $ git log -S'foo()' # commits which add or remove any file data
668 # matching the string 'foo()'
669 -------------------------------------------------
670
671 And of course you can combine all of these; the following finds
672 commits since v2.5 which touch the Makefile or any file under fs:
673
674 -------------------------------------------------
675 $ git log v2.5.. Makefile fs/
676 -------------------------------------------------
677
678 You can also ask git log to show patches:
679
680 -------------------------------------------------
681 $ git log -p
682 -------------------------------------------------
683
684 See the "--pretty" option in the linkgit:git-log[1] man page for more
685 display options.
686
687 Note that git log starts with the most recent commit and works
688 backwards through the parents; however, since git history can contain
689 multiple independent lines of development, the particular order that
690 commits are listed in may be somewhat arbitrary.
691
692 [[generating-diffs]]
693 Generating diffs
694 ----------------
695
696 You can generate diffs between any two versions using
697 linkgit:git-diff[1]:
698
699 -------------------------------------------------
700 $ git diff master..test
701 -------------------------------------------------
702
703 That will produce the diff between the tips of the two branches. If
704 you'd prefer to find the diff from their common ancestor to test, you
705 can use three dots instead of two:
706
707 -------------------------------------------------
708 $ git diff master...test
709 -------------------------------------------------
710
711 Sometimes what you want instead is a set of patches; for this you can
712 use linkgit:git-format-patch[1]:
713
714 -------------------------------------------------
715 $ git format-patch master..test
716 -------------------------------------------------
717
718 will generate a file with a patch for each commit reachable from test
719 but not from master.
720
721 [[viewing-old-file-versions]]
722 Viewing old file versions
723 -------------------------
724
725 You can always view an old version of a file by just checking out the
726 correct revision first. But sometimes it is more convenient to be
727 able to view an old version of a single file without checking
728 anything out; this command does that:
729
730 -------------------------------------------------
731 $ git show v2.5:fs/locks.c
732 -------------------------------------------------
733
734 Before the colon may be anything that names a commit, and after it
735 may be any path to a file tracked by git.
736
737 [[history-examples]]
738 Examples
739 --------
740
741 [[counting-commits-on-a-branch]]
742 Counting the number of commits on a branch
743 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
744
745 Suppose you want to know how many commits you've made on "mybranch"
746 since it diverged from "origin":
747
748 -------------------------------------------------
749 $ git log --pretty=oneline origin..mybranch | wc -l
750 -------------------------------------------------
751
752 Alternatively, you may often see this sort of thing done with the
753 lower-level command linkgit:git-rev-list[1], which just lists the SHA-1's
754 of all the given commits:
755
756 -------------------------------------------------
757 $ git rev-list origin..mybranch | wc -l
758 -------------------------------------------------
759
760 [[checking-for-equal-branches]]
761 Check whether two branches point at the same history
762 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
763
764 Suppose you want to check whether two branches point at the same point
765 in history.
766
767 -------------------------------------------------
768 $ git diff origin..master
769 -------------------------------------------------
770
771 will tell you whether the contents of the project are the same at the
772 two branches; in theory, however, it's possible that the same project
773 contents could have been arrived at by two different historical
774 routes. You could compare the object names:
775
776 -------------------------------------------------
777 $ git rev-list origin
778 e05db0fd4f31dde7005f075a84f96b360d05984b
779 $ git rev-list master
780 e05db0fd4f31dde7005f075a84f96b360d05984b
781 -------------------------------------------------
782
783 Or you could recall that the ... operator selects all commits
784 contained reachable from either one reference or the other but not
785 both: so
786
787 -------------------------------------------------
788 $ git log origin...master
789 -------------------------------------------------
790
791 will return no commits when the two branches are equal.
792
793 [[finding-tagged-descendants]]
794 Find first tagged version including a given fix
795 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
796
797 Suppose you know that the commit e05db0fd fixed a certain problem.
798 You'd like to find the earliest tagged release that contains that
799 fix.
800
801 Of course, there may be more than one answer--if the history branched
802 after commit e05db0fd, then there could be multiple "earliest" tagged
803 releases.
804
805 You could just visually inspect the commits since e05db0fd:
806
807 -------------------------------------------------
808 $ gitk e05db0fd..
809 -------------------------------------------------
810
811 Or you can use linkgit:git-name-rev[1], which will give the commit a
812 name based on any tag it finds pointing to one of the commit's
813 descendants:
814
815 -------------------------------------------------
816 $ git name-rev --tags e05db0fd
817 e05db0fd tags/v1.5.0-rc1^0~23
818 -------------------------------------------------
819
820 The linkgit:git-describe[1] command does the opposite, naming the
821 revision using a tag on which the given commit is based:
822
823 -------------------------------------------------
824 $ git describe e05db0fd
825 v1.5.0-rc0-260-ge05db0f
826 -------------------------------------------------
827
828 but that may sometimes help you guess which tags might come after the
829 given commit.
830
831 If you just want to verify whether a given tagged version contains a
832 given commit, you could use linkgit:git-merge-base[1]:
833
834 -------------------------------------------------
835 $ git merge-base e05db0fd v1.5.0-rc1
836 e05db0fd4f31dde7005f075a84f96b360d05984b
837 -------------------------------------------------
838
839 The merge-base command finds a common ancestor of the given commits,
840 and always returns one or the other in the case where one is a
841 descendant of the other; so the above output shows that e05db0fd
842 actually is an ancestor of v1.5.0-rc1.
843
844 Alternatively, note that
845
846 -------------------------------------------------
847 $ git log v1.5.0-rc1..e05db0fd
848 -------------------------------------------------
849
850 will produce empty output if and only if v1.5.0-rc1 includes e05db0fd,
851 because it outputs only commits that are not reachable from v1.5.0-rc1.
852
853 As yet another alternative, the linkgit:git-show-branch[1] command lists
854 the commits reachable from its arguments with a display on the left-hand
855 side that indicates which arguments that commit is reachable from. So,
856 you can run something like
857
858 -------------------------------------------------
859 $ git show-branch e05db0fd v1.5.0-rc0 v1.5.0-rc1 v1.5.0-rc2
860 ! [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
861 available
862 ! [v1.5.0-rc0] GIT v1.5.0 preview
863 ! [v1.5.0-rc1] GIT v1.5.0-rc1
864 ! [v1.5.0-rc2] GIT v1.5.0-rc2
865 ...
866 -------------------------------------------------
867
868 then search for a line that looks like
869
870 -------------------------------------------------
871 + ++ [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
872 available
873 -------------------------------------------------
874
875 Which shows that e05db0fd is reachable from itself, from v1.5.0-rc1, and
876 from v1.5.0-rc2, but not from v1.5.0-rc0.
877
878 [[showing-commits-unique-to-a-branch]]
879 Showing commits unique to a given branch
880 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
881
882 Suppose you would like to see all the commits reachable from the branch
883 head named "master" but not from any other head in your repository.
884
885 We can list all the heads in this repository with
886 linkgit:git-show-ref[1]:
887
888 -------------------------------------------------
889 $ git show-ref --heads
890 bf62196b5e363d73353a9dcf094c59595f3153b7 refs/heads/core-tutorial
891 db768d5504c1bb46f63ee9d6e1772bd047e05bf9 refs/heads/maint
892 a07157ac624b2524a059a3414e99f6f44bebc1e7 refs/heads/master
893 24dbc180ea14dc1aebe09f14c8ecf32010690627 refs/heads/tutorial-2
894 1e87486ae06626c2f31eaa63d26fc0fd646c8af2 refs/heads/tutorial-fixes
895 -------------------------------------------------
896
897 We can get just the branch-head names, and remove "master", with
898 the help of the standard utilities cut and grep:
899
900 -------------------------------------------------
901 $ git show-ref --heads | cut -d' ' -f2 | grep -v '^refs/heads/master'
902 refs/heads/core-tutorial
903 refs/heads/maint
904 refs/heads/tutorial-2
905 refs/heads/tutorial-fixes
906 -------------------------------------------------
907
908 And then we can ask to see all the commits reachable from master
909 but not from these other heads:
910
911 -------------------------------------------------
912 $ gitk master --not $( git show-ref --heads | cut -d' ' -f2 |
913 grep -v '^refs/heads/master' )
914 -------------------------------------------------
915
916 Obviously, endless variations are possible; for example, to see all
917 commits reachable from some head but not from any tag in the repository:
918
919 -------------------------------------------------
920 $ gitk $( git show-ref --heads ) --not $( git show-ref --tags )
921 -------------------------------------------------
922
923 (See linkgit:gitrevisions[7] for explanations of commit-selecting
924 syntax such as `--not`.)
925
926 [[making-a-release]]
927 Creating a changelog and tarball for a software release
928 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
929
930 The linkgit:git-archive[1] command can create a tar or zip archive from
931 any version of a project; for example:
932
933 -------------------------------------------------
934 $ git archive --format=tar --prefix=project/ HEAD | gzip >latest.tar.gz
935 -------------------------------------------------
936
937 will use HEAD to produce a tar archive in which each filename is
938 preceded by "project/".
939
940 If you're releasing a new version of a software project, you may want
941 to simultaneously make a changelog to include in the release
942 announcement.
943
944 Linus Torvalds, for example, makes new kernel releases by tagging them,
945 then running:
946
947 -------------------------------------------------
948 $ release-script 2.6.12 2.6.13-rc6 2.6.13-rc7
949 -------------------------------------------------
950
951 where release-script is a shell script that looks like:
952
953 -------------------------------------------------
954 #!/bin/sh
955 stable="$1"
956 last="$2"
957 new="$3"
958 echo "# git tag v$new"
959 echo "git archive --prefix=linux-$new/ v$new | gzip -9 > ../linux-$new.tar.gz"
960 echo "git diff v$stable v$new | gzip -9 > ../patch-$new.gz"
961 echo "git log --no-merges v$new ^v$last > ../ChangeLog-$new"
962 echo "git shortlog --no-merges v$new ^v$last > ../ShortLog"
963 echo "git diff --stat --summary -M v$last v$new > ../diffstat-$new"
964 -------------------------------------------------
965
966 and then he just cut-and-pastes the output commands after verifying that
967 they look OK.
968
969 [[Finding-commits-With-given-Content]]
970 Finding commits referencing a file with given content
971 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
972
973 Somebody hands you a copy of a file, and asks which commits modified a
974 file such that it contained the given content either before or after the
975 commit. You can find out with this:
976
977 -------------------------------------------------
978 $ git log --raw --abbrev=40 --pretty=oneline |
979 grep -B 1 `git hash-object filename`
980 -------------------------------------------------
981
982 Figuring out why this works is left as an exercise to the (advanced)
983 student. The linkgit:git-log[1], linkgit:git-diff-tree[1], and
984 linkgit:git-hash-object[1] man pages may prove helpful.
985
986 [[Developing-With-git]]
987 Developing with git
988 ===================
989
990 [[telling-git-your-name]]
991 Telling git your name
992 ---------------------
993
994 Before creating any commits, you should introduce yourself to git. The
995 easiest way to do so is to make sure the following lines appear in a
996 file named .gitconfig in your home directory:
997
998 ------------------------------------------------
999 [user]
1000 name = Your Name Comes Here
1001 email = you@yourdomain.example.com
1002 ------------------------------------------------
1003
1004 (See the "CONFIGURATION FILE" section of linkgit:git-config[1] for
1005 details on the configuration file.)
1006
1007
1008 [[creating-a-new-repository]]
1009 Creating a new repository
1010 -------------------------
1011
1012 Creating a new repository from scratch is very easy:
1013
1014 -------------------------------------------------
1015 $ mkdir project
1016 $ cd project
1017 $ git init
1018 -------------------------------------------------
1019
1020 If you have some initial content (say, a tarball):
1021
1022 -------------------------------------------------
1023 $ tar xzvf project.tar.gz
1024 $ cd project
1025 $ git init
1026 $ git add . # include everything below ./ in the first commit:
1027 $ git commit
1028 -------------------------------------------------
1029
1030 [[how-to-make-a-commit]]
1031 How to make a commit
1032 --------------------
1033
1034 Creating a new commit takes three steps:
1035
1036 1. Making some changes to the working directory using your
1037 favorite editor.
1038 2. Telling git about your changes.
1039 3. Creating the commit using the content you told git about
1040 in step 2.
1041
1042 In practice, you can interleave and repeat steps 1 and 2 as many
1043 times as you want: in order to keep track of what you want committed
1044 at step 3, git maintains a snapshot of the tree's contents in a
1045 special staging area called "the index."
1046
1047 At the beginning, the content of the index will be identical to
1048 that of the HEAD. The command "git diff --cached", which shows
1049 the difference between the HEAD and the index, should therefore
1050 produce no output at that point.
1051
1052 Modifying the index is easy:
1053
1054 To update the index with the new contents of a modified file, use
1055
1056 -------------------------------------------------
1057 $ git add path/to/file
1058 -------------------------------------------------
1059
1060 To add the contents of a new file to the index, use
1061
1062 -------------------------------------------------
1063 $ git add path/to/file
1064 -------------------------------------------------
1065
1066 To remove a file from the index and from the working tree,
1067
1068 -------------------------------------------------
1069 $ git rm path/to/file
1070 -------------------------------------------------
1071
1072 After each step you can verify that
1073
1074 -------------------------------------------------
1075 $ git diff --cached
1076 -------------------------------------------------
1077
1078 always shows the difference between the HEAD and the index file--this
1079 is what you'd commit if you created the commit now--and that
1080
1081 -------------------------------------------------
1082 $ git diff
1083 -------------------------------------------------
1084
1085 shows the difference between the working tree and the index file.
1086
1087 Note that "git add" always adds just the current contents of a file
1088 to the index; further changes to the same file will be ignored unless
1089 you run `git add` on the file again.
1090
1091 When you're ready, just run
1092
1093 -------------------------------------------------
1094 $ git commit
1095 -------------------------------------------------
1096
1097 and git will prompt you for a commit message and then create the new
1098 commit. Check to make sure it looks like what you expected with
1099
1100 -------------------------------------------------
1101 $ git show
1102 -------------------------------------------------
1103
1104 As a special shortcut,
1105
1106 -------------------------------------------------
1107 $ git commit -a
1108 -------------------------------------------------
1109
1110 will update the index with any files that you've modified or removed
1111 and create a commit, all in one step.
1112
1113 A number of commands are useful for keeping track of what you're
1114 about to commit:
1115
1116 -------------------------------------------------
1117 $ git diff --cached # difference between HEAD and the index; what
1118 # would be committed if you ran "commit" now.
1119 $ git diff # difference between the index file and your
1120 # working directory; changes that would not
1121 # be included if you ran "commit" now.
1122 $ git diff HEAD # difference between HEAD and working tree; what
1123 # would be committed if you ran "commit -a" now.
1124 $ git status # a brief per-file summary of the above.
1125 -------------------------------------------------
1126
1127 You can also use linkgit:git-gui[1] to create commits, view changes in
1128 the index and the working tree files, and individually select diff hunks
1129 for inclusion in the index (by right-clicking on the diff hunk and
1130 choosing "Stage Hunk For Commit").
1131
1132 [[creating-good-commit-messages]]
1133 Creating good commit messages
1134 -----------------------------
1135
1136 Though not required, it's a good idea to begin the commit message
1137 with a single short (less than 50 character) line summarizing the
1138 change, followed by a blank line and then a more thorough
1139 description. The text up to the first blank line in a commit
1140 message is treated as the commit title, and that title is used
1141 throughout git. For example, linkgit:git-format-patch[1] turns a
1142 commit into email, and it uses the title on the Subject line and the
1143 rest of the commit in the body.
1144
1145
1146 [[ignoring-files]]
1147 Ignoring files
1148 --------------
1149
1150 A project will often generate files that you do 'not' want to track with git.
1151 This typically includes files generated by a build process or temporary
1152 backup files made by your editor. Of course, 'not' tracking files with git
1153 is just a matter of 'not' calling `git add` on them. But it quickly becomes
1154 annoying to have these untracked files lying around; e.g. they make
1155 `git add .` practically useless, and they keep showing up in the output of
1156 `git status`.
1157
1158 You can tell git to ignore certain files by creating a file called .gitignore
1159 in the top level of your working directory, with contents such as:
1160
1161 -------------------------------------------------
1162 # Lines starting with '#' are considered comments.
1163 # Ignore any file named foo.txt.
1164 foo.txt
1165 # Ignore (generated) html files,
1166 *.html
1167 # except foo.html which is maintained by hand.
1168 !foo.html
1169 # Ignore objects and archives.
1170 *.[oa]
1171 -------------------------------------------------
1172
1173 See linkgit:gitignore[5] for a detailed explanation of the syntax. You can
1174 also place .gitignore files in other directories in your working tree, and they
1175 will apply to those directories and their subdirectories. The `.gitignore`
1176 files can be added to your repository like any other files (just run `git add
1177 .gitignore` and `git commit`, as usual), which is convenient when the exclude
1178 patterns (such as patterns matching build output files) would also make sense
1179 for other users who clone your repository.
1180
1181 If you wish the exclude patterns to affect only certain repositories
1182 (instead of every repository for a given project), you may instead put
1183 them in a file in your repository named .git/info/exclude, or in any file
1184 specified by the `core.excludesfile` configuration variable. Some git
1185 commands can also take exclude patterns directly on the command line.
1186 See linkgit:gitignore[5] for the details.
1187
1188 [[how-to-merge]]
1189 How to merge
1190 ------------
1191
1192 You can rejoin two diverging branches of development using
1193 linkgit:git-merge[1]:
1194
1195 -------------------------------------------------
1196 $ git merge branchname
1197 -------------------------------------------------
1198
1199 merges the development in the branch "branchname" into the current
1200 branch.
1201
1202 A merge is made by combining the changes made in "branchname" and the
1203 changes made up to the latest commit in your current branch since
1204 their histories forked. The work tree is overwritten by the result of
1205 the merge when this combining is done cleanly, or overwritten by a
1206 half-merged results when this combining results in conflicts.
1207 Therefore, if you have uncommitted changes touching the same files as
1208 the ones impacted by the merge, Git will refuse to proceed. Most of
1209 the time, you will want to commit your changes before you can merge,
1210 and if you don't, then linkgit:git-stash[1] can take these changes
1211 away while you're doing the merge, and reapply them afterwards.
1212
1213 If the changes are independent enough, Git will automatically complete
1214 the merge and commit the result (or reuse an existing commit in case
1215 of <<fast-forwards,fast-forward>>, see below). On the other hand,
1216 if there are conflicts--for example, if the same file is
1217 modified in two different ways in the remote branch and the local
1218 branch--then you are warned; the output may look something like this:
1219
1220 -------------------------------------------------
1221 $ git merge next
1222 100% (4/4) done
1223 Auto-merged file.txt
1224 CONFLICT (content): Merge conflict in file.txt
1225 Automatic merge failed; fix conflicts and then commit the result.
1226 -------------------------------------------------
1227
1228 Conflict markers are left in the problematic files, and after
1229 you resolve the conflicts manually, you can update the index
1230 with the contents and run git commit, as you normally would when
1231 creating a new file.
1232
1233 If you examine the resulting commit using gitk, you will see that it
1234 has two parents, one pointing to the top of the current branch, and
1235 one to the top of the other branch.
1236
1237 [[resolving-a-merge]]
1238 Resolving a merge
1239 -----------------
1240
1241 When a merge isn't resolved automatically, git leaves the index and
1242 the working tree in a special state that gives you all the
1243 information you need to help resolve the merge.
1244
1245 Files with conflicts are marked specially in the index, so until you
1246 resolve the problem and update the index, linkgit:git-commit[1] will
1247 fail:
1248
1249 -------------------------------------------------
1250 $ git commit
1251 file.txt: needs merge
1252 -------------------------------------------------
1253
1254 Also, linkgit:git-status[1] will list those files as "unmerged", and the
1255 files with conflicts will have conflict markers added, like this:
1256
1257 -------------------------------------------------
1258 <<<<<<< HEAD:file.txt
1259 Hello world
1260 =======
1261 Goodbye
1262 >>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1263 -------------------------------------------------
1264
1265 All you need to do is edit the files to resolve the conflicts, and then
1266
1267 -------------------------------------------------
1268 $ git add file.txt
1269 $ git commit
1270 -------------------------------------------------
1271
1272 Note that the commit message will already be filled in for you with
1273 some information about the merge. Normally you can just use this
1274 default message unchanged, but you may add additional commentary of
1275 your own if desired.
1276
1277 The above is all you need to know to resolve a simple merge. But git
1278 also provides more information to help resolve conflicts:
1279
1280 [[conflict-resolution]]
1281 Getting conflict-resolution help during a merge
1282 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1283
1284 All of the changes that git was able to merge automatically are
1285 already added to the index file, so linkgit:git-diff[1] shows only
1286 the conflicts. It uses an unusual syntax:
1287
1288 -------------------------------------------------
1289 $ git diff
1290 diff --cc file.txt
1291 index 802992c,2b60207..0000000
1292 --- a/file.txt
1293 +++ b/file.txt
1294 @@@ -1,1 -1,1 +1,5 @@@
1295 ++<<<<<<< HEAD:file.txt
1296 +Hello world
1297 ++=======
1298 + Goodbye
1299 ++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1300 -------------------------------------------------
1301
1302 Recall that the commit which will be committed after we resolve this
1303 conflict will have two parents instead of the usual one: one parent
1304 will be HEAD, the tip of the current branch; the other will be the
1305 tip of the other branch, which is stored temporarily in MERGE_HEAD.
1306
1307 During the merge, the index holds three versions of each file. Each of
1308 these three "file stages" represents a different version of the file:
1309
1310 -------------------------------------------------
1311 $ git show :1:file.txt # the file in a common ancestor of both branches
1312 $ git show :2:file.txt # the version from HEAD.
1313 $ git show :3:file.txt # the version from MERGE_HEAD.
1314 -------------------------------------------------
1315
1316 When you ask linkgit:git-diff[1] to show the conflicts, it runs a
1317 three-way diff between the conflicted merge results in the work tree with
1318 stages 2 and 3 to show only hunks whose contents come from both sides,
1319 mixed (in other words, when a hunk's merge results come only from stage 2,
1320 that part is not conflicting and is not shown. Same for stage 3).
1321
1322 The diff above shows the differences between the working-tree version of
1323 file.txt and the stage 2 and stage 3 versions. So instead of preceding
1324 each line by a single "+" or "-", it now uses two columns: the first
1325 column is used for differences between the first parent and the working
1326 directory copy, and the second for differences between the second parent
1327 and the working directory copy. (See the "COMBINED DIFF FORMAT" section
1328 of linkgit:git-diff-files[1] for a details of the format.)
1329
1330 After resolving the conflict in the obvious way (but before updating the
1331 index), the diff will look like:
1332
1333 -------------------------------------------------
1334 $ git diff
1335 diff --cc file.txt
1336 index 802992c,2b60207..0000000
1337 --- a/file.txt
1338 +++ b/file.txt
1339 @@@ -1,1 -1,1 +1,1 @@@
1340 - Hello world
1341 -Goodbye
1342 ++Goodbye world
1343 -------------------------------------------------
1344
1345 This shows that our resolved version deleted "Hello world" from the
1346 first parent, deleted "Goodbye" from the second parent, and added
1347 "Goodbye world", which was previously absent from both.
1348
1349 Some special diff options allow diffing the working directory against
1350 any of these stages:
1351
1352 -------------------------------------------------
1353 $ git diff -1 file.txt # diff against stage 1
1354 $ git diff --base file.txt # same as the above
1355 $ git diff -2 file.txt # diff against stage 2
1356 $ git diff --ours file.txt # same as the above
1357 $ git diff -3 file.txt # diff against stage 3
1358 $ git diff --theirs file.txt # same as the above.
1359 -------------------------------------------------
1360
1361 The linkgit:git-log[1] and linkgit:gitk[1] commands also provide special help
1362 for merges:
1363
1364 -------------------------------------------------
1365 $ git log --merge
1366 $ gitk --merge
1367 -------------------------------------------------
1368
1369 These will display all commits which exist only on HEAD or on
1370 MERGE_HEAD, and which touch an unmerged file.
1371
1372 You may also use linkgit:git-mergetool[1], which lets you merge the
1373 unmerged files using external tools such as Emacs or kdiff3.
1374
1375 Each time you resolve the conflicts in a file and update the index:
1376
1377 -------------------------------------------------
1378 $ git add file.txt
1379 -------------------------------------------------
1380
1381 the different stages of that file will be "collapsed", after which
1382 `git diff` will (by default) no longer show diffs for that file.
1383
1384 [[undoing-a-merge]]
1385 Undoing a merge
1386 ---------------
1387
1388 If you get stuck and decide to just give up and throw the whole mess
1389 away, you can always return to the pre-merge state with
1390
1391 -------------------------------------------------
1392 $ git reset --hard HEAD
1393 -------------------------------------------------
1394
1395 Or, if you've already committed the merge that you want to throw away,
1396
1397 -------------------------------------------------
1398 $ git reset --hard ORIG_HEAD
1399 -------------------------------------------------
1400
1401 However, this last command can be dangerous in some cases--never
1402 throw away a commit you have already committed if that commit may
1403 itself have been merged into another branch, as doing so may confuse
1404 further merges.
1405
1406 [[fast-forwards]]
1407 Fast-forward merges
1408 -------------------
1409
1410 There is one special case not mentioned above, which is treated
1411 differently. Normally, a merge results in a merge commit, with two
1412 parents, one pointing at each of the two lines of development that
1413 were merged.
1414
1415 However, if the current branch is a descendant of the other--so every
1416 commit present in the one is already contained in the other--then git
1417 just performs a "fast-forward"; the head of the current branch is moved
1418 forward to point at the head of the merged-in branch, without any new
1419 commits being created.
1420
1421 [[fixing-mistakes]]
1422 Fixing mistakes
1423 ---------------
1424
1425 If you've messed up the working tree, but haven't yet committed your
1426 mistake, you can return the entire working tree to the last committed
1427 state with
1428
1429 -------------------------------------------------
1430 $ git reset --hard HEAD
1431 -------------------------------------------------
1432
1433 If you make a commit that you later wish you hadn't, there are two
1434 fundamentally different ways to fix the problem:
1435
1436 1. You can create a new commit that undoes whatever was done
1437 by the old commit. This is the correct thing if your
1438 mistake has already been made public.
1439
1440 2. You can go back and modify the old commit. You should
1441 never do this if you have already made the history public;
1442 git does not normally expect the "history" of a project to
1443 change, and cannot correctly perform repeated merges from
1444 a branch that has had its history changed.
1445
1446 [[reverting-a-commit]]
1447 Fixing a mistake with a new commit
1448 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1449
1450 Creating a new commit that reverts an earlier change is very easy;
1451 just pass the linkgit:git-revert[1] command a reference to the bad
1452 commit; for example, to revert the most recent commit:
1453
1454 -------------------------------------------------
1455 $ git revert HEAD
1456 -------------------------------------------------
1457
1458 This will create a new commit which undoes the change in HEAD. You
1459 will be given a chance to edit the commit message for the new commit.
1460
1461 You can also revert an earlier change, for example, the next-to-last:
1462
1463 -------------------------------------------------
1464 $ git revert HEAD^
1465 -------------------------------------------------
1466
1467 In this case git will attempt to undo the old change while leaving
1468 intact any changes made since then. If more recent changes overlap
1469 with the changes to be reverted, then you will be asked to fix
1470 conflicts manually, just as in the case of <<resolving-a-merge,
1471 resolving a merge>>.
1472
1473 [[fixing-a-mistake-by-rewriting-history]]
1474 Fixing a mistake by rewriting history
1475 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1476
1477 If the problematic commit is the most recent commit, and you have not
1478 yet made that commit public, then you may just
1479 <<undoing-a-merge,destroy it using `git reset`>>.
1480
1481 Alternatively, you
1482 can edit the working directory and update the index to fix your
1483 mistake, just as if you were going to <<how-to-make-a-commit,create a
1484 new commit>>, then run
1485
1486 -------------------------------------------------
1487 $ git commit --amend
1488 -------------------------------------------------
1489
1490 which will replace the old commit by a new commit incorporating your
1491 changes, giving you a chance to edit the old commit message first.
1492
1493 Again, you should never do this to a commit that may already have
1494 been merged into another branch; use linkgit:git-revert[1] instead in
1495 that case.
1496
1497 It is also possible to replace commits further back in the history, but
1498 this is an advanced topic to be left for
1499 <<cleaning-up-history,another chapter>>.
1500
1501 [[checkout-of-path]]
1502 Checking out an old version of a file
1503 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1504
1505 In the process of undoing a previous bad change, you may find it
1506 useful to check out an older version of a particular file using
1507 linkgit:git-checkout[1]. We've used `git checkout` before to switch
1508 branches, but it has quite different behavior if it is given a path
1509 name: the command
1510
1511 -------------------------------------------------
1512 $ git checkout HEAD^ path/to/file
1513 -------------------------------------------------
1514
1515 replaces path/to/file by the contents it had in the commit HEAD^, and
1516 also updates the index to match. It does not change branches.
1517
1518 If you just want to look at an old version of the file, without
1519 modifying the working directory, you can do that with
1520 linkgit:git-show[1]:
1521
1522 -------------------------------------------------
1523 $ git show HEAD^:path/to/file
1524 -------------------------------------------------
1525
1526 which will display the given version of the file.
1527
1528 [[interrupted-work]]
1529 Temporarily setting aside work in progress
1530 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1531
1532 While you are in the middle of working on something complicated, you
1533 find an unrelated but obvious and trivial bug. You would like to fix it
1534 before continuing. You can use linkgit:git-stash[1] to save the current
1535 state of your work, and after fixing the bug (or, optionally after doing
1536 so on a different branch and then coming back), unstash the
1537 work-in-progress changes.
1538
1539 ------------------------------------------------
1540 $ git stash save "work in progress for foo feature"
1541 ------------------------------------------------
1542
1543 This command will save your changes away to the `stash`, and
1544 reset your working tree and the index to match the tip of your
1545 current branch. Then you can make your fix as usual.
1546
1547 ------------------------------------------------
1548 ... edit and test ...
1549 $ git commit -a -m "blorpl: typofix"
1550 ------------------------------------------------
1551
1552 After that, you can go back to what you were working on with
1553 `git stash pop`:
1554
1555 ------------------------------------------------
1556 $ git stash pop
1557 ------------------------------------------------
1558
1559
1560 [[ensuring-good-performance]]
1561 Ensuring good performance
1562 -------------------------
1563
1564 On large repositories, git depends on compression to keep the history
1565 information from taking up too much space on disk or in memory.
1566
1567 This compression is not performed automatically. Therefore you
1568 should occasionally run linkgit:git-gc[1]:
1569
1570 -------------------------------------------------
1571 $ git gc
1572 -------------------------------------------------
1573
1574 to recompress the archive. This can be very time-consuming, so
1575 you may prefer to run `git gc` when you are not doing other work.
1576
1577
1578 [[ensuring-reliability]]
1579 Ensuring reliability
1580 --------------------
1581
1582 [[checking-for-corruption]]
1583 Checking the repository for corruption
1584 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1585
1586 The linkgit:git-fsck[1] command runs a number of self-consistency checks
1587 on the repository, and reports on any problems. This may take some
1588 time.
1589
1590 -------------------------------------------------
1591 $ git fsck
1592 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1593 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1594 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1595 dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1596 dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1597 dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1598 dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1599 dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1600 ...
1601 -------------------------------------------------
1602
1603 You will see informational messages on dangling objects. They are objects
1604 that still exist in the repository but are no longer referenced by any of
1605 your branches, and can (and will) be removed after a while with "gc".
1606 You can run `git fsck --no-dangling` to suppress these messages, and still
1607 view real errors.
1608
1609 [[recovering-lost-changes]]
1610 Recovering lost changes
1611 ~~~~~~~~~~~~~~~~~~~~~~~
1612
1613 [[reflogs]]
1614 Reflogs
1615 ^^^^^^^
1616
1617 Say you modify a branch with +linkgit:git-reset[1] \--hard+, and then
1618 realize that the branch was the only reference you had to that point in
1619 history.
1620
1621 Fortunately, git also keeps a log, called a "reflog", of all the
1622 previous values of each branch. So in this case you can still find the
1623 old history using, for example,
1624
1625 -------------------------------------------------
1626 $ git log master@{1}
1627 -------------------------------------------------
1628
1629 This lists the commits reachable from the previous version of the
1630 "master" branch head. This syntax can be used with any git command
1631 that accepts a commit, not just with git log. Some other examples:
1632
1633 -------------------------------------------------
1634 $ git show master@{2} # See where the branch pointed 2,
1635 $ git show master@{3} # 3, ... changes ago.
1636 $ gitk master@{yesterday} # See where it pointed yesterday,
1637 $ gitk master@{"1 week ago"} # ... or last week
1638 $ git log --walk-reflogs master # show reflog entries for master
1639 -------------------------------------------------
1640
1641 A separate reflog is kept for the HEAD, so
1642
1643 -------------------------------------------------
1644 $ git show HEAD@{"1 week ago"}
1645 -------------------------------------------------
1646
1647 will show what HEAD pointed to one week ago, not what the current branch
1648 pointed to one week ago. This allows you to see the history of what
1649 you've checked out.
1650
1651 The reflogs are kept by default for 30 days, after which they may be
1652 pruned. See linkgit:git-reflog[1] and linkgit:git-gc[1] to learn
1653 how to control this pruning, and see the "SPECIFYING REVISIONS"
1654 section of linkgit:gitrevisions[7] for details.
1655
1656 Note that the reflog history is very different from normal git history.
1657 While normal history is shared by every repository that works on the
1658 same project, the reflog history is not shared: it tells you only about
1659 how the branches in your local repository have changed over time.
1660
1661 [[dangling-object-recovery]]
1662 Examining dangling objects
1663 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1664
1665 In some situations the reflog may not be able to save you. For example,
1666 suppose you delete a branch, then realize you need the history it
1667 contained. The reflog is also deleted; however, if you have not yet
1668 pruned the repository, then you may still be able to find the lost
1669 commits in the dangling objects that `git fsck` reports. See
1670 <<dangling-objects>> for the details.
1671
1672 -------------------------------------------------
1673 $ git fsck
1674 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1675 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1676 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1677 ...
1678 -------------------------------------------------
1679
1680 You can examine
1681 one of those dangling commits with, for example,
1682
1683 ------------------------------------------------
1684 $ gitk 7281251ddd --not --all
1685 ------------------------------------------------
1686
1687 which does what it sounds like: it says that you want to see the commit
1688 history that is described by the dangling commit(s), but not the
1689 history that is described by all your existing branches and tags. Thus
1690 you get exactly the history reachable from that commit that is lost.
1691 (And notice that it might not be just one commit: we only report the
1692 "tip of the line" as being dangling, but there might be a whole deep
1693 and complex commit history that was dropped.)
1694
1695 If you decide you want the history back, you can always create a new
1696 reference pointing to it, for example, a new branch:
1697
1698 ------------------------------------------------
1699 $ git branch recovered-branch 7281251ddd
1700 ------------------------------------------------
1701
1702 Other types of dangling objects (blobs and trees) are also possible, and
1703 dangling objects can arise in other situations.
1704
1705
1706 [[sharing-development]]
1707 Sharing development with others
1708 ===============================
1709
1710 [[getting-updates-With-git-pull]]
1711 Getting updates with git pull
1712 -----------------------------
1713
1714 After you clone a repository and commit a few changes of your own, you
1715 may wish to check the original repository for updates and merge them
1716 into your own work.
1717
1718 We have already seen <<Updating-a-repository-With-git-fetch,how to
1719 keep remote-tracking branches up to date>> with linkgit:git-fetch[1],
1720 and how to merge two branches. So you can merge in changes from the
1721 original repository's master branch with:
1722
1723 -------------------------------------------------
1724 $ git fetch
1725 $ git merge origin/master
1726 -------------------------------------------------
1727
1728 However, the linkgit:git-pull[1] command provides a way to do this in
1729 one step:
1730
1731 -------------------------------------------------
1732 $ git pull origin master
1733 -------------------------------------------------
1734
1735 In fact, if you have "master" checked out, then this branch has been
1736 configured by "git clone" to get changes from the HEAD branch of the
1737 origin repository. So often you can
1738 accomplish the above with just a simple
1739
1740 -------------------------------------------------
1741 $ git pull
1742 -------------------------------------------------
1743
1744 This command will fetch changes from the remote branches to your
1745 remote-tracking branches `origin/*`, and merge the default branch into
1746 the current branch.
1747
1748 More generally, a branch that is created from a remote-tracking branch
1749 will pull
1750 by default from that branch. See the descriptions of the
1751 branch.<name>.remote and branch.<name>.merge options in
1752 linkgit:git-config[1], and the discussion of the `--track` option in
1753 linkgit:git-checkout[1], to learn how to control these defaults.
1754
1755 In addition to saving you keystrokes, "git pull" also helps you by
1756 producing a default commit message documenting the branch and
1757 repository that you pulled from.
1758
1759 (But note that no such commit will be created in the case of a
1760 <<fast-forwards,fast-forward>>; instead, your branch will just be
1761 updated to point to the latest commit from the upstream branch.)
1762
1763 The `git pull` command can also be given "." as the "remote" repository,
1764 in which case it just merges in a branch from the current repository; so
1765 the commands
1766
1767 -------------------------------------------------
1768 $ git pull . branch
1769 $ git merge branch
1770 -------------------------------------------------
1771
1772 are roughly equivalent. The former is actually very commonly used.
1773
1774 [[submitting-patches]]
1775 Submitting patches to a project
1776 -------------------------------
1777
1778 If you just have a few changes, the simplest way to submit them may
1779 just be to send them as patches in email:
1780
1781 First, use linkgit:git-format-patch[1]; for example:
1782
1783 -------------------------------------------------
1784 $ git format-patch origin
1785 -------------------------------------------------
1786
1787 will produce a numbered series of files in the current directory, one
1788 for each patch in the current branch but not in origin/HEAD.
1789
1790 You can then import these into your mail client and send them by
1791 hand. However, if you have a lot to send at once, you may prefer to
1792 use the linkgit:git-send-email[1] script to automate the process.
1793 Consult the mailing list for your project first to determine how they
1794 prefer such patches be handled.
1795
1796 [[importing-patches]]
1797 Importing patches to a project
1798 ------------------------------
1799
1800 Git also provides a tool called linkgit:git-am[1] (am stands for
1801 "apply mailbox"), for importing such an emailed series of patches.
1802 Just save all of the patch-containing messages, in order, into a
1803 single mailbox file, say "patches.mbox", then run
1804
1805 -------------------------------------------------
1806 $ git am -3 patches.mbox
1807 -------------------------------------------------
1808
1809 Git will apply each patch in order; if any conflicts are found, it
1810 will stop, and you can fix the conflicts as described in
1811 "<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1812 git to perform a merge; if you would prefer it just to abort and
1813 leave your tree and index untouched, you may omit that option.)
1814
1815 Once the index is updated with the results of the conflict
1816 resolution, instead of creating a new commit, just run
1817
1818 -------------------------------------------------
1819 $ git am --resolved
1820 -------------------------------------------------
1821
1822 and git will create the commit for you and continue applying the
1823 remaining patches from the mailbox.
1824
1825 The final result will be a series of commits, one for each patch in
1826 the original mailbox, with authorship and commit log message each
1827 taken from the message containing each patch.
1828
1829 [[public-repositories]]
1830 Public git repositories
1831 -----------------------
1832
1833 Another way to submit changes to a project is to tell the maintainer
1834 of that project to pull the changes from your repository using
1835 linkgit:git-pull[1]. In the section "<<getting-updates-With-git-pull,
1836 Getting updates with `git pull`>>" we described this as a way to get
1837 updates from the "main" repository, but it works just as well in the
1838 other direction.
1839
1840 If you and the maintainer both have accounts on the same machine, then
1841 you can just pull changes from each other's repositories directly;
1842 commands that accept repository URLs as arguments will also accept a
1843 local directory name:
1844
1845 -------------------------------------------------
1846 $ git clone /path/to/repository
1847 $ git pull /path/to/other/repository
1848 -------------------------------------------------
1849
1850 or an ssh URL:
1851
1852 -------------------------------------------------
1853 $ git clone ssh://yourhost/~you/repository
1854 -------------------------------------------------
1855
1856 For projects with few developers, or for synchronizing a few private
1857 repositories, this may be all you need.
1858
1859 However, the more common way to do this is to maintain a separate public
1860 repository (usually on a different host) for others to pull changes
1861 from. This is usually more convenient, and allows you to cleanly
1862 separate private work in progress from publicly visible work.
1863
1864 You will continue to do your day-to-day work in your personal
1865 repository, but periodically "push" changes from your personal
1866 repository into your public repository, allowing other developers to
1867 pull from that repository. So the flow of changes, in a situation
1868 where there is one other developer with a public repository, looks
1869 like this:
1870
1871 you push
1872 your personal repo ------------------> your public repo
1873 ^ |
1874 | |
1875 | you pull | they pull
1876 | |
1877 | |
1878 | they push V
1879 their public repo <------------------- their repo
1880
1881 We explain how to do this in the following sections.
1882
1883 [[setting-up-a-public-repository]]
1884 Setting up a public repository
1885 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1886
1887 Assume your personal repository is in the directory ~/proj. We
1888 first create a new clone of the repository and tell `git daemon` that it
1889 is meant to be public:
1890
1891 -------------------------------------------------
1892 $ git clone --bare ~/proj proj.git
1893 $ touch proj.git/git-daemon-export-ok
1894 -------------------------------------------------
1895
1896 The resulting directory proj.git contains a "bare" git repository--it is
1897 just the contents of the ".git" directory, without any files checked out
1898 around it.
1899
1900 Next, copy proj.git to the server where you plan to host the
1901 public repository. You can use scp, rsync, or whatever is most
1902 convenient.
1903
1904 [[exporting-via-git]]
1905 Exporting a git repository via the git protocol
1906 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1907
1908 This is the preferred method.
1909
1910 If someone else administers the server, they should tell you what
1911 directory to put the repository in, and what git:// URL it will appear
1912 at. You can then skip to the section
1913 "<<pushing-changes-to-a-public-repository,Pushing changes to a public
1914 repository>>", below.
1915
1916 Otherwise, all you need to do is start linkgit:git-daemon[1]; it will
1917 listen on port 9418. By default, it will allow access to any directory
1918 that looks like a git directory and contains the magic file
1919 git-daemon-export-ok. Passing some directory paths as `git daemon`
1920 arguments will further restrict the exports to those paths.
1921
1922 You can also run `git daemon` as an inetd service; see the
1923 linkgit:git-daemon[1] man page for details. (See especially the
1924 examples section.)
1925
1926 [[exporting-via-http]]
1927 Exporting a git repository via http
1928 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1929
1930 The git protocol gives better performance and reliability, but on a
1931 host with a web server set up, http exports may be simpler to set up.
1932
1933 All you need to do is place the newly created bare git repository in
1934 a directory that is exported by the web server, and make some
1935 adjustments to give web clients some extra information they need:
1936
1937 -------------------------------------------------
1938 $ mv proj.git /home/you/public_html/proj.git
1939 $ cd proj.git
1940 $ git --bare update-server-info
1941 $ mv hooks/post-update.sample hooks/post-update
1942 -------------------------------------------------
1943
1944 (For an explanation of the last two lines, see
1945 linkgit:git-update-server-info[1] and linkgit:githooks[5].)
1946
1947 Advertise the URL of proj.git. Anybody else should then be able to
1948 clone or pull from that URL, for example with a command line like:
1949
1950 -------------------------------------------------
1951 $ git clone http://yourserver.com/~you/proj.git
1952 -------------------------------------------------
1953
1954 (See also
1955 link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1956 for a slightly more sophisticated setup using WebDAV which also
1957 allows pushing over http.)
1958
1959 [[pushing-changes-to-a-public-repository]]
1960 Pushing changes to a public repository
1961 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1962
1963 Note that the two techniques outlined above (exporting via
1964 <<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1965 maintainers to fetch your latest changes, but they do not allow write
1966 access, which you will need to update the public repository with the
1967 latest changes created in your private repository.
1968
1969 The simplest way to do this is using linkgit:git-push[1] and ssh; to
1970 update the remote branch named "master" with the latest state of your
1971 branch named "master", run
1972
1973 -------------------------------------------------
1974 $ git push ssh://yourserver.com/~you/proj.git master:master
1975 -------------------------------------------------
1976
1977 or just
1978
1979 -------------------------------------------------
1980 $ git push ssh://yourserver.com/~you/proj.git master
1981 -------------------------------------------------
1982
1983 As with `git fetch`, `git push` will complain if this does not result in a
1984 <<fast-forwards,fast-forward>>; see the following section for details on
1985 handling this case.
1986
1987 Note that the target of a "push" is normally a
1988 <<def_bare_repository,bare>> repository. You can also push to a
1989 repository that has a checked-out working tree, but a push to update the
1990 currently checked-out branch is denied by default to prevent confusion.
1991 See the description ofthe receive.denyCurrentBranch option
1992 in linkgit:git-config[1] for details.
1993
1994 As with `git fetch`, you may also set up configuration options to
1995 save typing; so, for example, after
1996
1997 -------------------------------------------------
1998 $ cat >>.git/config <<EOF
1999 [remote "public-repo"]
2000 url = ssh://yourserver.com/~you/proj.git
2001 EOF
2002 -------------------------------------------------
2003
2004 you should be able to perform the above push with just
2005
2006 -------------------------------------------------
2007 $ git push public-repo master
2008 -------------------------------------------------
2009
2010 See the explanations of the remote.<name>.url, branch.<name>.remote,
2011 and remote.<name>.push options in linkgit:git-config[1] for
2012 details.
2013
2014 [[forcing-push]]
2015 What to do when a push fails
2016 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2017
2018 If a push would not result in a <<fast-forwards,fast-forward>> of the
2019 remote branch, then it will fail with an error like:
2020
2021 -------------------------------------------------
2022 error: remote 'refs/heads/master' is not an ancestor of
2023 local 'refs/heads/master'.
2024 Maybe you are not up-to-date and need to pull first?
2025 error: failed to push to 'ssh://yourserver.com/~you/proj.git'
2026 -------------------------------------------------
2027
2028 This can happen, for example, if you:
2029
2030 - use `git reset --hard` to remove already-published commits, or
2031 - use `git commit --amend` to replace already-published commits
2032 (as in <<fixing-a-mistake-by-rewriting-history>>), or
2033 - use `git rebase` to rebase any already-published commits (as
2034 in <<using-git-rebase>>).
2035
2036 You may force `git push` to perform the update anyway by preceding the
2037 branch name with a plus sign:
2038
2039 -------------------------------------------------
2040 $ git push ssh://yourserver.com/~you/proj.git +master
2041 -------------------------------------------------
2042
2043 Normally whenever a branch head in a public repository is modified, it
2044 is modified to point to a descendant of the commit that it pointed to
2045 before. By forcing a push in this situation, you break that convention.
2046 (See <<problems-With-rewriting-history>>.)
2047
2048 Nevertheless, this is a common practice for people that need a simple
2049 way to publish a work-in-progress patch series, and it is an acceptable
2050 compromise as long as you warn other developers that this is how you
2051 intend to manage the branch.
2052
2053 It's also possible for a push to fail in this way when other people have
2054 the right to push to the same repository. In that case, the correct
2055 solution is to retry the push after first updating your work: either by a
2056 pull, or by a fetch followed by a rebase; see the
2057 <<setting-up-a-shared-repository,next section>> and
2058 linkgit:gitcvs-migration[7] for more.
2059
2060 [[setting-up-a-shared-repository]]
2061 Setting up a shared repository
2062 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2063
2064 Another way to collaborate is by using a model similar to that
2065 commonly used in CVS, where several developers with special rights
2066 all push to and pull from a single shared repository. See
2067 linkgit:gitcvs-migration[7] for instructions on how to
2068 set this up.
2069
2070 However, while there is nothing wrong with git's support for shared
2071 repositories, this mode of operation is not generally recommended,
2072 simply because the mode of collaboration that git supports--by
2073 exchanging patches and pulling from public repositories--has so many
2074 advantages over the central shared repository:
2075
2076 - Git's ability to quickly import and merge patches allows a
2077 single maintainer to process incoming changes even at very
2078 high rates. And when that becomes too much, `git pull` provides
2079 an easy way for that maintainer to delegate this job to other
2080 maintainers while still allowing optional review of incoming
2081 changes.
2082 - Since every developer's repository has the same complete copy
2083 of the project history, no repository is special, and it is
2084 trivial for another developer to take over maintenance of a
2085 project, either by mutual agreement, or because a maintainer
2086 becomes unresponsive or difficult to work with.
2087 - The lack of a central group of "committers" means there is
2088 less need for formal decisions about who is "in" and who is
2089 "out".
2090
2091 [[setting-up-gitweb]]
2092 Allowing web browsing of a repository
2093 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2094
2095 The gitweb cgi script provides users an easy way to browse your
2096 project's files and history without having to install git; see the file
2097 gitweb/INSTALL in the git source tree for instructions on setting it up.
2098
2099 [[sharing-development-examples]]
2100 Examples
2101 --------
2102
2103 [[maintaining-topic-branches]]
2104 Maintaining topic branches for a Linux subsystem maintainer
2105 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2106
2107 This describes how Tony Luck uses git in his role as maintainer of the
2108 IA64 architecture for the Linux kernel.
2109
2110 He uses two public branches:
2111
2112 - A "test" tree into which patches are initially placed so that they
2113 can get some exposure when integrated with other ongoing development.
2114 This tree is available to Andrew for pulling into -mm whenever he
2115 wants.
2116
2117 - A "release" tree into which tested patches are moved for final sanity
2118 checking, and as a vehicle to send them upstream to Linus (by sending
2119 him a "please pull" request.)
2120
2121 He also uses a set of temporary branches ("topic branches"), each
2122 containing a logical grouping of patches.
2123
2124 To set this up, first create your work tree by cloning Linus's public
2125 tree:
2126
2127 -------------------------------------------------
2128 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git work
2129 $ cd work
2130 -------------------------------------------------
2131
2132 Linus's tree will be stored in the remote-tracking branch named origin/master,
2133 and can be updated using linkgit:git-fetch[1]; you can track other
2134 public trees using linkgit:git-remote[1] to set up a "remote" and
2135 linkgit:git-fetch[1] to keep them up-to-date; see
2136 <<repositories-and-branches>>.
2137
2138 Now create the branches in which you are going to work; these start out
2139 at the current tip of origin/master branch, and should be set up (using
2140 the --track option to linkgit:git-branch[1]) to merge changes in from
2141 Linus by default.
2142
2143 -------------------------------------------------
2144 $ git branch --track test origin/master
2145 $ git branch --track release origin/master
2146 -------------------------------------------------
2147
2148 These can be easily kept up to date using linkgit:git-pull[1].
2149
2150 -------------------------------------------------
2151 $ git checkout test && git pull
2152 $ git checkout release && git pull
2153 -------------------------------------------------
2154
2155 Important note! If you have any local changes in these branches, then
2156 this merge will create a commit object in the history (with no local
2157 changes git will simply do a "fast-forward" merge). Many people dislike
2158 the "noise" that this creates in the Linux history, so you should avoid
2159 doing this capriciously in the "release" branch, as these noisy commits
2160 will become part of the permanent history when you ask Linus to pull
2161 from the release branch.
2162
2163 A few configuration variables (see linkgit:git-config[1]) can
2164 make it easy to push both branches to your public tree. (See
2165 <<setting-up-a-public-repository>>.)
2166
2167 -------------------------------------------------
2168 $ cat >> .git/config <<EOF
2169 [remote "mytree"]
2170 url = master.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6.git
2171 push = release
2172 push = test
2173 EOF
2174 -------------------------------------------------
2175
2176 Then you can push both the test and release trees using
2177 linkgit:git-push[1]:
2178
2179 -------------------------------------------------
2180 $ git push mytree
2181 -------------------------------------------------
2182
2183 or push just one of the test and release branches using:
2184
2185 -------------------------------------------------
2186 $ git push mytree test
2187 -------------------------------------------------
2188
2189 or
2190
2191 -------------------------------------------------
2192 $ git push mytree release
2193 -------------------------------------------------
2194
2195 Now to apply some patches from the community. Think of a short
2196 snappy name for a branch to hold this patch (or related group of
2197 patches), and create a new branch from a recent stable tag of
2198 Linus's branch. Picking a stable base for your branch will:
2199 1) help you: by avoiding inclusion of unrelated and perhaps lightly
2200 tested changes
2201 2) help future bug hunters that use "git bisect" to find problems
2202
2203 -------------------------------------------------
2204 $ git checkout -b speed-up-spinlocks v2.6.35
2205 -------------------------------------------------
2206
2207 Now you apply the patch(es), run some tests, and commit the change(s). If
2208 the patch is a multi-part series, then you should apply each as a separate
2209 commit to this branch.
2210
2211 -------------------------------------------------
2212 $ ... patch ... test ... commit [ ... patch ... test ... commit ]*
2213 -------------------------------------------------
2214
2215 When you are happy with the state of this change, you can pull it into the
2216 "test" branch in preparation to make it public:
2217
2218 -------------------------------------------------
2219 $ git checkout test && git pull . speed-up-spinlocks
2220 -------------------------------------------------
2221
2222 It is unlikely that you would have any conflicts here ... but you might if you
2223 spent a while on this step and had also pulled new versions from upstream.
2224
2225 Some time later when enough time has passed and testing done, you can pull the
2226 same branch into the "release" tree ready to go upstream. This is where you
2227 see the value of keeping each patch (or patch series) in its own branch. It
2228 means that the patches can be moved into the "release" tree in any order.
2229
2230 -------------------------------------------------
2231 $ git checkout release && git pull . speed-up-spinlocks
2232 -------------------------------------------------
2233
2234 After a while, you will have a number of branches, and despite the
2235 well chosen names you picked for each of them, you may forget what
2236 they are for, or what status they are in. To get a reminder of what
2237 changes are in a specific branch, use:
2238
2239 -------------------------------------------------
2240 $ git log linux..branchname | git shortlog
2241 -------------------------------------------------
2242
2243 To see whether it has already been merged into the test or release branches,
2244 use:
2245
2246 -------------------------------------------------
2247 $ git log test..branchname
2248 -------------------------------------------------
2249
2250 or
2251
2252 -------------------------------------------------
2253 $ git log release..branchname
2254 -------------------------------------------------
2255
2256 (If this branch has not yet been merged, you will see some log entries.
2257 If it has been merged, then there will be no output.)
2258
2259 Once a patch completes the great cycle (moving from test to release,
2260 then pulled by Linus, and finally coming back into your local
2261 "origin/master" branch), the branch for this change is no longer needed.
2262 You detect this when the output from:
2263
2264 -------------------------------------------------
2265 $ git log origin..branchname
2266 -------------------------------------------------
2267
2268 is empty. At this point the branch can be deleted:
2269
2270 -------------------------------------------------
2271 $ git branch -d branchname
2272 -------------------------------------------------
2273
2274 Some changes are so trivial that it is not necessary to create a separate
2275 branch and then merge into each of the test and release branches. For
2276 these changes, just apply directly to the "release" branch, and then
2277 merge that into the "test" branch.
2278
2279 To create diffstat and shortlog summaries of changes to include in a "please
2280 pull" request to Linus you can use:
2281
2282 -------------------------------------------------
2283 $ git diff --stat origin..release
2284 -------------------------------------------------
2285
2286 and
2287
2288 -------------------------------------------------
2289 $ git log -p origin..release | git shortlog
2290 -------------------------------------------------
2291
2292 Here are some of the scripts that simplify all this even further.
2293
2294 -------------------------------------------------
2295 ==== update script ====
2296 # Update a branch in my GIT tree. If the branch to be updated
2297 # is origin, then pull from kernel.org. Otherwise merge
2298 # origin/master branch into test|release branch
2299
2300 case "$1" in
2301 test|release)
2302 git checkout $1 && git pull . origin
2303 ;;
2304 origin)
2305 before=$(git rev-parse refs/remotes/origin/master)
2306 git fetch origin
2307 after=$(git rev-parse refs/remotes/origin/master)
2308 if [ $before != $after ]
2309 then
2310 git log $before..$after | git shortlog
2311 fi
2312 ;;
2313 *)
2314 echo "Usage: $0 origin|test|release" 1>&2
2315 exit 1
2316 ;;
2317 esac
2318 -------------------------------------------------
2319
2320 -------------------------------------------------
2321 ==== merge script ====
2322 # Merge a branch into either the test or release branch
2323
2324 pname=$0
2325
2326 usage()
2327 {
2328 echo "Usage: $pname branch test|release" 1>&2
2329 exit 1
2330 }
2331
2332 git show-ref -q --verify -- refs/heads/"$1" || {
2333 echo "Can't see branch <$1>" 1>&2
2334 usage
2335 }
2336
2337 case "$2" in
2338 test|release)
2339 if [ $(git log $2..$1 | wc -c) -eq 0 ]
2340 then
2341 echo $1 already merged into $2 1>&2
2342 exit 1
2343 fi
2344 git checkout $2 && git pull . $1
2345 ;;
2346 *)
2347 usage
2348 ;;
2349 esac
2350 -------------------------------------------------
2351
2352 -------------------------------------------------
2353 ==== status script ====
2354 # report on status of my ia64 GIT tree
2355
2356 gb=$(tput setab 2)
2357 rb=$(tput setab 1)
2358 restore=$(tput setab 9)
2359
2360 if [ `git rev-list test..release | wc -c` -gt 0 ]
2361 then
2362 echo $rb Warning: commits in release that are not in test $restore
2363 git log test..release
2364 fi
2365
2366 for branch in `git show-ref --heads | sed 's|^.*/||'`
2367 do
2368 if [ $branch = test -o $branch = release ]
2369 then
2370 continue
2371 fi
2372
2373 echo -n $gb ======= $branch ====== $restore " "
2374 status=
2375 for ref in test release origin/master
2376 do
2377 if [ `git rev-list $ref..$branch | wc -c` -gt 0 ]
2378 then
2379 status=$status${ref:0:1}
2380 fi
2381 done
2382 case $status in
2383 trl)
2384 echo $rb Need to pull into test $restore
2385 ;;
2386 rl)
2387 echo "In test"
2388 ;;
2389 l)
2390 echo "Waiting for linus"
2391 ;;
2392 "")
2393 echo $rb All done $restore
2394 ;;
2395 *)
2396 echo $rb "<$status>" $restore
2397 ;;
2398 esac
2399 git log origin/master..$branch | git shortlog
2400 done
2401 -------------------------------------------------
2402
2403
2404 [[cleaning-up-history]]
2405 Rewriting history and maintaining patch series
2406 ==============================================
2407
2408 Normally commits are only added to a project, never taken away or
2409 replaced. Git is designed with this assumption, and violating it will
2410 cause git's merge machinery (for example) to do the wrong thing.
2411
2412 However, there is a situation in which it can be useful to violate this
2413 assumption.
2414
2415 [[patch-series]]
2416 Creating the perfect patch series
2417 ---------------------------------
2418
2419 Suppose you are a contributor to a large project, and you want to add a
2420 complicated feature, and to present it to the other developers in a way
2421 that makes it easy for them to read your changes, verify that they are
2422 correct, and understand why you made each change.
2423
2424 If you present all of your changes as a single patch (or commit), they
2425 may find that it is too much to digest all at once.
2426
2427 If you present them with the entire history of your work, complete with
2428 mistakes, corrections, and dead ends, they may be overwhelmed.
2429
2430 So the ideal is usually to produce a series of patches such that:
2431
2432 1. Each patch can be applied in order.
2433
2434 2. Each patch includes a single logical change, together with a
2435 message explaining the change.
2436
2437 3. No patch introduces a regression: after applying any initial
2438 part of the series, the resulting project still compiles and
2439 works, and has no bugs that it didn't have before.
2440
2441 4. The complete series produces the same end result as your own
2442 (probably much messier!) development process did.
2443
2444 We will introduce some tools that can help you do this, explain how to
2445 use them, and then explain some of the problems that can arise because
2446 you are rewriting history.
2447
2448 [[using-git-rebase]]
2449 Keeping a patch series up to date using git rebase
2450 --------------------------------------------------
2451
2452 Suppose that you create a branch "mywork" on a remote-tracking branch
2453 "origin", and create some commits on top of it:
2454
2455 -------------------------------------------------
2456 $ git checkout -b mywork origin
2457 $ vi file.txt
2458 $ git commit
2459 $ vi otherfile.txt
2460 $ git commit
2461 ...
2462 -------------------------------------------------
2463
2464 You have performed no merges into mywork, so it is just a simple linear
2465 sequence of patches on top of "origin":
2466
2467 ................................................
2468 o--o--O <-- origin
2469 \
2470 a--b--c <-- mywork
2471 ................................................
2472
2473 Some more interesting work has been done in the upstream project, and
2474 "origin" has advanced:
2475
2476 ................................................
2477 o--o--O--o--o--o <-- origin
2478 \
2479 a--b--c <-- mywork
2480 ................................................
2481
2482 At this point, you could use "pull" to merge your changes back in;
2483 the result would create a new merge commit, like this:
2484
2485 ................................................
2486 o--o--O--o--o--o <-- origin
2487 \ \
2488 a--b--c--m <-- mywork
2489 ................................................
2490
2491 However, if you prefer to keep the history in mywork a simple series of
2492 commits without any merges, you may instead choose to use
2493 linkgit:git-rebase[1]:
2494
2495 -------------------------------------------------
2496 $ git checkout mywork
2497 $ git rebase origin
2498 -------------------------------------------------
2499
2500 This will remove each of your commits from mywork, temporarily saving
2501 them as patches (in a directory named ".git/rebase-apply"), update mywork to
2502 point at the latest version of origin, then apply each of the saved
2503 patches to the new mywork. The result will look like:
2504
2505
2506 ................................................
2507 o--o--O--o--o--o <-- origin
2508 \
2509 a'--b'--c' <-- mywork
2510 ................................................
2511
2512 In the process, it may discover conflicts. In that case it will stop
2513 and allow you to fix the conflicts; after fixing conflicts, use `git add`
2514 to update the index with those contents, and then, instead of
2515 running `git commit`, just run
2516
2517 -------------------------------------------------
2518 $ git rebase --continue
2519 -------------------------------------------------
2520
2521 and git will continue applying the rest of the patches.
2522
2523 At any point you may use the `--abort` option to abort this process and
2524 return mywork to the state it had before you started the rebase:
2525
2526 -------------------------------------------------
2527 $ git rebase --abort
2528 -------------------------------------------------
2529
2530 [[rewriting-one-commit]]
2531 Rewriting a single commit
2532 -------------------------
2533
2534 We saw in <<fixing-a-mistake-by-rewriting-history>> that you can replace the
2535 most recent commit using
2536
2537 -------------------------------------------------
2538 $ git commit --amend
2539 -------------------------------------------------
2540
2541 which will replace the old commit by a new commit incorporating your
2542 changes, giving you a chance to edit the old commit message first.
2543
2544 You can also use a combination of this and linkgit:git-rebase[1] to
2545 replace a commit further back in your history and recreate the
2546 intervening changes on top of it. First, tag the problematic commit
2547 with
2548
2549 -------------------------------------------------
2550 $ git tag bad mywork~5
2551 -------------------------------------------------
2552
2553 (Either gitk or `git log` may be useful for finding the commit.)
2554
2555 Then check out that commit, edit it, and rebase the rest of the series
2556 on top of it (note that we could check out the commit on a temporary
2557 branch, but instead we're using a <<detached-head,detached head>>):
2558
2559 -------------------------------------------------
2560 $ git checkout bad
2561 $ # make changes here and update the index
2562 $ git commit --amend
2563 $ git rebase --onto HEAD bad mywork
2564 -------------------------------------------------
2565
2566 When you're done, you'll be left with mywork checked out, with the top
2567 patches on mywork reapplied on top of your modified commit. You can
2568 then clean up with
2569
2570 -------------------------------------------------
2571 $ git tag -d bad
2572 -------------------------------------------------
2573
2574 Note that the immutable nature of git history means that you haven't really
2575 "modified" existing commits; instead, you have replaced the old commits with
2576 new commits having new object names.
2577
2578 [[reordering-patch-series]]
2579 Reordering or selecting from a patch series
2580 -------------------------------------------
2581
2582 Given one existing commit, the linkgit:git-cherry-pick[1] command
2583 allows you to apply the change introduced by that commit and create a
2584 new commit that records it. So, for example, if "mywork" points to a
2585 series of patches on top of "origin", you might do something like:
2586
2587 -------------------------------------------------
2588 $ git checkout -b mywork-new origin
2589 $ gitk origin..mywork &
2590 -------------------------------------------------
2591
2592 and browse through the list of patches in the mywork branch using gitk,
2593 applying them (possibly in a different order) to mywork-new using
2594 cherry-pick, and possibly modifying them as you go using `git commit --amend`.
2595 The linkgit:git-gui[1] command may also help as it allows you to
2596 individually select diff hunks for inclusion in the index (by
2597 right-clicking on the diff hunk and choosing "Stage Hunk for Commit").
2598
2599 Another technique is to use `git format-patch` to create a series of
2600 patches, then reset the state to before the patches:
2601
2602 -------------------------------------------------
2603 $ git format-patch origin
2604 $ git reset --hard origin
2605 -------------------------------------------------
2606
2607 Then modify, reorder, or eliminate patches as preferred before applying
2608 them again with linkgit:git-am[1].
2609
2610 [[patch-series-tools]]
2611 Other tools
2612 -----------
2613
2614 There are numerous other tools, such as StGit, which exist for the
2615 purpose of maintaining a patch series. These are outside of the scope of
2616 this manual.
2617
2618 [[problems-With-rewriting-history]]
2619 Problems with rewriting history
2620 -------------------------------
2621
2622 The primary problem with rewriting the history of a branch has to do
2623 with merging. Suppose somebody fetches your branch and merges it into
2624 their branch, with a result something like this:
2625
2626 ................................................
2627 o--o--O--o--o--o <-- origin
2628 \ \
2629 t--t--t--m <-- their branch:
2630 ................................................
2631
2632 Then suppose you modify the last three commits:
2633
2634 ................................................
2635 o--o--o <-- new head of origin
2636 /
2637 o--o--O--o--o--o <-- old head of origin
2638 ................................................
2639
2640 If we examined all this history together in one repository, it will
2641 look like:
2642
2643 ................................................
2644 o--o--o <-- new head of origin
2645 /
2646 o--o--O--o--o--o <-- old head of origin
2647 \ \
2648 t--t--t--m <-- their branch:
2649 ................................................
2650
2651 Git has no way of knowing that the new head is an updated version of
2652 the old head; it treats this situation exactly the same as it would if
2653 two developers had independently done the work on the old and new heads
2654 in parallel. At this point, if someone attempts to merge the new head
2655 in to their branch, git will attempt to merge together the two (old and
2656 new) lines of development, instead of trying to replace the old by the
2657 new. The results are likely to be unexpected.
2658
2659 You may still choose to publish branches whose history is rewritten,
2660 and it may be useful for others to be able to fetch those branches in
2661 order to examine or test them, but they should not attempt to pull such
2662 branches into their own work.
2663
2664 For true distributed development that supports proper merging,
2665 published branches should never be rewritten.
2666
2667 [[bisect-merges]]
2668 Why bisecting merge commits can be harder than bisecting linear history
2669 -----------------------------------------------------------------------
2670
2671 The linkgit:git-bisect[1] command correctly handles history that
2672 includes merge commits. However, when the commit that it finds is a
2673 merge commit, the user may need to work harder than usual to figure out
2674 why that commit introduced a problem.
2675
2676 Imagine this history:
2677
2678 ................................................
2679 ---Z---o---X---...---o---A---C---D
2680 \ /
2681 o---o---Y---...---o---B
2682 ................................................
2683
2684 Suppose that on the upper line of development, the meaning of one
2685 of the functions that exists at Z is changed at commit X. The
2686 commits from Z leading to A change both the function's
2687 implementation and all calling sites that exist at Z, as well
2688 as new calling sites they add, to be consistent. There is no
2689 bug at A.
2690
2691 Suppose that in the meantime on the lower line of development somebody
2692 adds a new calling site for that function at commit Y. The
2693 commits from Z leading to B all assume the old semantics of that
2694 function and the callers and the callee are consistent with each
2695 other. There is no bug at B, either.
2696
2697 Suppose further that the two development lines merge cleanly at C,
2698 so no conflict resolution is required.
2699
2700 Nevertheless, the code at C is broken, because the callers added
2701 on the lower line of development have not been converted to the new
2702 semantics introduced on the upper line of development. So if all
2703 you know is that D is bad, that Z is good, and that
2704 linkgit:git-bisect[1] identifies C as the culprit, how will you
2705 figure out that the problem is due to this change in semantics?
2706
2707 When the result of a `git bisect` is a non-merge commit, you should
2708 normally be able to discover the problem by examining just that commit.
2709 Developers can make this easy by breaking their changes into small
2710 self-contained commits. That won't help in the case above, however,
2711 because the problem isn't obvious from examination of any single
2712 commit; instead, a global view of the development is required. To
2713 make matters worse, the change in semantics in the problematic
2714 function may be just one small part of the changes in the upper
2715 line of development.
2716
2717 On the other hand, if instead of merging at C you had rebased the
2718 history between Z to B on top of A, you would have gotten this
2719 linear history:
2720
2721 ................................................................
2722 ---Z---o---X--...---o---A---o---o---Y*--...---o---B*--D*
2723 ................................................................
2724
2725 Bisecting between Z and D* would hit a single culprit commit Y*,
2726 and understanding why Y* was broken would probably be easier.
2727
2728 Partly for this reason, many experienced git users, even when
2729 working on an otherwise merge-heavy project, keep the history
2730 linear by rebasing against the latest upstream version before
2731 publishing.
2732
2733 [[advanced-branch-management]]
2734 Advanced branch management
2735 ==========================
2736
2737 [[fetching-individual-branches]]
2738 Fetching individual branches
2739 ----------------------------
2740
2741 Instead of using linkgit:git-remote[1], you can also choose just
2742 to update one branch at a time, and to store it locally under an
2743 arbitrary name:
2744
2745 -------------------------------------------------
2746 $ git fetch origin todo:my-todo-work
2747 -------------------------------------------------
2748
2749 The first argument, "origin", just tells git to fetch from the
2750 repository you originally cloned from. The second argument tells git
2751 to fetch the branch named "todo" from the remote repository, and to
2752 store it locally under the name refs/heads/my-todo-work.
2753
2754 You can also fetch branches from other repositories; so
2755
2756 -------------------------------------------------
2757 $ git fetch git://example.com/proj.git master:example-master
2758 -------------------------------------------------
2759
2760 will create a new branch named "example-master" and store in it the
2761 branch named "master" from the repository at the given URL. If you
2762 already have a branch named example-master, it will attempt to
2763 <<fast-forwards,fast-forward>> to the commit given by example.com's
2764 master branch. In more detail:
2765
2766 [[fetch-fast-forwards]]
2767 git fetch and fast-forwards
2768 ---------------------------
2769
2770 In the previous example, when updating an existing branch, "git fetch"
2771 checks to make sure that the most recent commit on the remote
2772 branch is a descendant of the most recent commit on your copy of the
2773 branch before updating your copy of the branch to point at the new
2774 commit. Git calls this process a <<fast-forwards,fast-forward>>.
2775
2776 A fast-forward looks something like this:
2777
2778 ................................................
2779 o--o--o--o <-- old head of the branch
2780 \
2781 o--o--o <-- new head of the branch
2782 ................................................
2783
2784
2785 In some cases it is possible that the new head will *not* actually be
2786 a descendant of the old head. For example, the developer may have
2787 realized she made a serious mistake, and decided to backtrack,
2788 resulting in a situation like:
2789
2790 ................................................
2791 o--o--o--o--a--b <-- old head of the branch
2792 \
2793 o--o--o <-- new head of the branch
2794 ................................................
2795
2796 In this case, "git fetch" will fail, and print out a warning.
2797
2798 In that case, you can still force git to update to the new head, as
2799 described in the following section. However, note that in the
2800 situation above this may mean losing the commits labeled "a" and "b",
2801 unless you've already created a reference of your own pointing to
2802 them.
2803
2804 [[forcing-fetch]]
2805 Forcing git fetch to do non-fast-forward updates
2806 ------------------------------------------------
2807
2808 If git fetch fails because the new head of a branch is not a
2809 descendant of the old head, you may force the update with:
2810
2811 -------------------------------------------------
2812 $ git fetch git://example.com/proj.git +master:refs/remotes/example/master
2813 -------------------------------------------------
2814
2815 Note the addition of the "+" sign. Alternatively, you can use the "-f"
2816 flag to force updates of all the fetched branches, as in:
2817
2818 -------------------------------------------------
2819 $ git fetch -f origin
2820 -------------------------------------------------
2821
2822 Be aware that commits that the old version of example/master pointed at
2823 may be lost, as we saw in the previous section.
2824
2825 [[remote-branch-configuration]]
2826 Configuring remote-tracking branches
2827 ------------------------------------
2828
2829 We saw above that "origin" is just a shortcut to refer to the
2830 repository that you originally cloned from. This information is
2831 stored in git configuration variables, which you can see using
2832 linkgit:git-config[1]:
2833
2834 -------------------------------------------------
2835 $ git config -l
2836 core.repositoryformatversion=0
2837 core.filemode=true
2838 core.logallrefupdates=true
2839 remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2840 remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2841 branch.master.remote=origin
2842 branch.master.merge=refs/heads/master
2843 -------------------------------------------------
2844
2845 If there are other repositories that you also use frequently, you can
2846 create similar configuration options to save typing; for example,
2847 after
2848
2849 -------------------------------------------------
2850 $ git config remote.example.url git://example.com/proj.git
2851 -------------------------------------------------
2852
2853 then the following two commands will do the same thing:
2854
2855 -------------------------------------------------
2856 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2857 $ git fetch example master:refs/remotes/example/master
2858 -------------------------------------------------
2859
2860 Even better, if you add one more option:
2861
2862 -------------------------------------------------
2863 $ git config remote.example.fetch master:refs/remotes/example/master
2864 -------------------------------------------------
2865
2866 then the following commands will all do the same thing:
2867
2868 -------------------------------------------------
2869 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2870 $ git fetch example master:refs/remotes/example/master
2871 $ git fetch example
2872 -------------------------------------------------
2873
2874 You can also add a "+" to force the update each time:
2875
2876 -------------------------------------------------
2877 $ git config remote.example.fetch +master:refs/remotes/example/master
2878 -------------------------------------------------
2879
2880 Don't do this unless you're sure you won't mind "git fetch" possibly
2881 throwing away commits on 'example/master'.
2882
2883 Also note that all of the above configuration can be performed by
2884 directly editing the file .git/config instead of using
2885 linkgit:git-config[1].
2886
2887 See linkgit:git-config[1] for more details on the configuration
2888 options mentioned above.
2889
2890
2891 [[git-concepts]]
2892 Git concepts
2893 ============
2894
2895 Git is built on a small number of simple but powerful ideas. While it
2896 is possible to get things done without understanding them, you will find
2897 git much more intuitive if you do.
2898
2899 We start with the most important, the <<def_object_database,object
2900 database>> and the <<def_index,index>>.
2901
2902 [[the-object-database]]
2903 The Object Database
2904 -------------------
2905
2906
2907 We already saw in <<understanding-commits>> that all commits are stored
2908 under a 40-digit "object name". In fact, all the information needed to
2909 represent the history of a project is stored in objects with such names.
2910 In each case the name is calculated by taking the SHA-1 hash of the
2911 contents of the object. The SHA-1 hash is a cryptographic hash function.
2912 What that means to us is that it is impossible to find two different
2913 objects with the same name. This has a number of advantages; among
2914 others:
2915
2916 - Git can quickly determine whether two objects are identical or not,
2917 just by comparing names.
2918 - Since object names are computed the same way in every repository, the
2919 same content stored in two repositories will always be stored under
2920 the same name.
2921 - Git can detect errors when it reads an object, by checking that the
2922 object's name is still the SHA-1 hash of its contents.
2923
2924 (See <<object-details>> for the details of the object formatting and
2925 SHA-1 calculation.)
2926
2927 There are four different types of objects: "blob", "tree", "commit", and
2928 "tag".
2929
2930 - A <<def_blob_object,"blob" object>> is used to store file data.
2931 - A <<def_tree_object,"tree" object>> ties one or more
2932 "blob" objects into a directory structure. In addition, a tree object
2933 can refer to other tree objects, thus creating a directory hierarchy.
2934 - A <<def_commit_object,"commit" object>> ties such directory hierarchies
2935 together into a <<def_DAG,directed acyclic graph>> of revisions--each
2936 commit contains the object name of exactly one tree designating the
2937 directory hierarchy at the time of the commit. In addition, a commit
2938 refers to "parent" commit objects that describe the history of how we
2939 arrived at that directory hierarchy.
2940 - A <<def_tag_object,"tag" object>> symbolically identifies and can be
2941 used to sign other objects. It contains the object name and type of
2942 another object, a symbolic name (of course!) and, optionally, a
2943 signature.
2944
2945 The object types in some more detail:
2946
2947 [[commit-object]]
2948 Commit Object
2949 ~~~~~~~~~~~~~
2950
2951 The "commit" object links a physical state of a tree with a description
2952 of how we got there and why. Use the --pretty=raw option to
2953 linkgit:git-show[1] or linkgit:git-log[1] to examine your favorite
2954 commit:
2955
2956 ------------------------------------------------
2957 $ git show -s --pretty=raw 2be7fcb476
2958 commit 2be7fcb4764f2dbcee52635b91fedb1b3dcf7ab4
2959 tree fb3a8bdd0ceddd019615af4d57a53f43d8cee2bf
2960 parent 257a84d9d02e90447b149af58b271c19405edb6a
2961 author Dave Watson <dwatson@mimvista.com> 1187576872 -0400
2962 committer Junio C Hamano <gitster@pobox.com> 1187591163 -0700
2963
2964 Fix misspelling of 'suppress' in docs
2965
2966 Signed-off-by: Junio C Hamano <gitster@pobox.com>
2967 ------------------------------------------------
2968
2969 As you can see, a commit is defined by:
2970
2971 - a tree: The SHA-1 name of a tree object (as defined below), representing
2972 the contents of a directory at a certain point in time.
2973 - parent(s): The SHA-1 name(s) of some number of commits which represent the
2974 immediately previous step(s) in the history of the project. The
2975 example above has one parent; merge commits may have more than
2976 one. A commit with no parents is called a "root" commit, and
2977 represents the initial revision of a project. Each project must have
2978 at least one root. A project can also have multiple roots, though
2979 that isn't common (or necessarily a good idea).
2980 - an author: The name of the person responsible for this change, together
2981 with its date.
2982 - a committer: The name of the person who actually created the commit,
2983 with the date it was done. This may be different from the author, for
2984 example, if the author was someone who wrote a patch and emailed it
2985 to the person who used it to create the commit.
2986 - a comment describing this commit.
2987
2988 Note that a commit does not itself contain any information about what
2989 actually changed; all changes are calculated by comparing the contents
2990 of the tree referred to by this commit with the trees associated with
2991 its parents. In particular, git does not attempt to record file renames
2992 explicitly, though it can identify cases where the existence of the same
2993 file data at changing paths suggests a rename. (See, for example, the
2994 -M option to linkgit:git-diff[1]).
2995
2996 A commit is usually created by linkgit:git-commit[1], which creates a
2997 commit whose parent is normally the current HEAD, and whose tree is
2998 taken from the content currently stored in the index.
2999
3000 [[tree-object]]
3001 Tree Object
3002 ~~~~~~~~~~~
3003
3004 The ever-versatile linkgit:git-show[1] command can also be used to
3005 examine tree objects, but linkgit:git-ls-tree[1] will give you more
3006 details:
3007
3008 ------------------------------------------------
3009 $ git ls-tree fb3a8bdd0ce
3010 100644 blob 63c918c667fa005ff12ad89437f2fdc80926e21c .gitignore
3011 100644 blob 5529b198e8d14decbe4ad99db3f7fb632de0439d .mailmap
3012 100644 blob 6ff87c4664981e4397625791c8ea3bbb5f2279a3 COPYING
3013 040000 tree 2fb783e477100ce076f6bf57e4a6f026013dc745 Documentation
3014 100755 blob 3c0032cec592a765692234f1cba47dfdcc3a9200 GIT-VERSION-GEN
3015 100644 blob 289b046a443c0647624607d471289b2c7dcd470b INSTALL
3016 100644 blob 4eb463797adc693dc168b926b6932ff53f17d0b1 Makefile
3017 100644 blob 548142c327a6790ff8821d67c2ee1eff7a656b52 README
3018 ...
3019 ------------------------------------------------
3020
3021 As you can see, a tree object contains a list of entries, each with a
3022 mode, object type, SHA-1 name, and name, sorted by name. It represents
3023 the contents of a single directory tree.
3024
3025 The object type may be a blob, representing the contents of a file, or
3026 another tree, representing the contents of a subdirectory. Since trees
3027 and blobs, like all other objects, are named by the SHA-1 hash of their
3028 contents, two trees have the same SHA-1 name if and only if their
3029 contents (including, recursively, the contents of all subdirectories)
3030 are identical. This allows git to quickly determine the differences
3031 between two related tree objects, since it can ignore any entries with
3032 identical object names.
3033
3034 (Note: in the presence of submodules, trees may also have commits as
3035 entries. See <<submodules>> for documentation.)
3036
3037 Note that the files all have mode 644 or 755: git actually only pays
3038 attention to the executable bit.
3039
3040 [[blob-object]]
3041 Blob Object
3042 ~~~~~~~~~~~
3043
3044 You can use linkgit:git-show[1] to examine the contents of a blob; take,
3045 for example, the blob in the entry for "COPYING" from the tree above:
3046
3047 ------------------------------------------------
3048 $ git show 6ff87c4664
3049
3050 Note that the only valid version of the GPL as far as this project
3051 is concerned is _this_ particular version of the license (ie v2, not
3052 v2.2 or v3.x or whatever), unless explicitly otherwise stated.
3053 ...
3054 ------------------------------------------------
3055
3056 A "blob" object is nothing but a binary blob of data. It doesn't refer
3057 to anything else or have attributes of any kind.
3058
3059 Since the blob is entirely defined by its data, if two files in a
3060 directory tree (or in multiple different versions of the repository)
3061 have the same contents, they will share the same blob object. The object
3062 is totally independent of its location in the directory tree, and
3063 renaming a file does not change the object that file is associated with.
3064
3065 Note that any tree or blob object can be examined using
3066 linkgit:git-show[1] with the <revision>:<path> syntax. This can
3067 sometimes be useful for browsing the contents of a tree that is not
3068 currently checked out.
3069
3070 [[trust]]
3071 Trust
3072 ~~~~~
3073
3074 If you receive the SHA-1 name of a blob from one source, and its contents
3075 from another (possibly untrusted) source, you can still trust that those
3076 contents are correct as long as the SHA-1 name agrees. This is because
3077 the SHA-1 is designed so that it is infeasible to find different contents
3078 that produce the same hash.
3079
3080 Similarly, you need only trust the SHA-1 name of a top-level tree object
3081 to trust the contents of the entire directory that it refers to, and if
3082 you receive the SHA-1 name of a commit from a trusted source, then you
3083 can easily verify the entire history of commits reachable through
3084 parents of that commit, and all of those contents of the trees referred
3085 to by those commits.
3086
3087 So to introduce some real trust in the system, the only thing you need
3088 to do is to digitally sign just 'one' special note, which includes the
3089 name of a top-level commit. Your digital signature shows others
3090 that you trust that commit, and the immutability of the history of
3091 commits tells others that they can trust the whole history.
3092
3093 In other words, you can easily validate a whole archive by just
3094 sending out a single email that tells the people the name (SHA-1 hash)
3095 of the top commit, and digitally sign that email using something
3096 like GPG/PGP.
3097
3098 To assist in this, git also provides the tag object...
3099
3100 [[tag-object]]
3101 Tag Object
3102 ~~~~~~~~~~
3103
3104 A tag object contains an object, object type, tag name, the name of the
3105 person ("tagger") who created the tag, and a message, which may contain
3106 a signature, as can be seen using linkgit:git-cat-file[1]:
3107
3108 ------------------------------------------------
3109 $ git cat-file tag v1.5.0
3110 object 437b1b20df4b356c9342dac8d38849f24ef44f27
3111 type commit
3112 tag v1.5.0
3113 tagger Junio C Hamano <junkio@cox.net> 1171411200 +0000
3114
3115 GIT 1.5.0
3116 -----BEGIN PGP SIGNATURE-----
3117 Version: GnuPG v1.4.6 (GNU/Linux)
3118
3119 iD8DBQBF0lGqwMbZpPMRm5oRAuRiAJ9ohBLd7s2kqjkKlq1qqC57SbnmzQCdG4ui
3120 nLE/L9aUXdWeTFPron96DLA=
3121 =2E+0
3122 -----END PGP SIGNATURE-----
3123 ------------------------------------------------
3124
3125 See the linkgit:git-tag[1] command to learn how to create and verify tag
3126 objects. (Note that linkgit:git-tag[1] can also be used to create
3127 "lightweight tags", which are not tag objects at all, but just simple
3128 references whose names begin with "refs/tags/").
3129
3130 [[pack-files]]
3131 How git stores objects efficiently: pack files
3132 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3133
3134 Newly created objects are initially created in a file named after the
3135 object's SHA-1 hash (stored in .git/objects).
3136
3137 Unfortunately this system becomes inefficient once a project has a
3138 lot of objects. Try this on an old project:
3139
3140 ------------------------------------------------
3141 $ git count-objects
3142 6930 objects, 47620 kilobytes
3143 ------------------------------------------------
3144
3145 The first number is the number of objects which are kept in
3146 individual files. The second is the amount of space taken up by
3147 those "loose" objects.
3148
3149 You can save space and make git faster by moving these loose objects in
3150 to a "pack file", which stores a group of objects in an efficient
3151 compressed format; the details of how pack files are formatted can be
3152 found in link:technical/pack-format.txt[technical/pack-format.txt].
3153
3154 To put the loose objects into a pack, just run git repack:
3155
3156 ------------------------------------------------
3157 $ git repack
3158 Generating pack...
3159 Done counting 6020 objects.
3160 Deltifying 6020 objects.
3161 100% (6020/6020) done
3162 Writing 6020 objects.
3163 100% (6020/6020) done
3164 Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
3165 Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
3166 ------------------------------------------------
3167
3168 You can then run
3169
3170 ------------------------------------------------
3171 $ git prune
3172 ------------------------------------------------
3173
3174 to remove any of the "loose" objects that are now contained in the
3175 pack. This will also remove any unreferenced objects (which may be
3176 created when, for example, you use "git reset" to remove a commit).
3177 You can verify that the loose objects are gone by looking at the
3178 .git/objects directory or by running
3179
3180 ------------------------------------------------
3181 $ git count-objects
3182 0 objects, 0 kilobytes
3183 ------------------------------------------------
3184
3185 Although the object files are gone, any commands that refer to those
3186 objects will work exactly as they did before.
3187
3188 The linkgit:git-gc[1] command performs packing, pruning, and more for
3189 you, so is normally the only high-level command you need.
3190
3191 [[dangling-objects]]
3192 Dangling objects
3193 ~~~~~~~~~~~~~~~~
3194
3195 The linkgit:git-fsck[1] command will sometimes complain about dangling
3196 objects. They are not a problem.
3197
3198 The most common cause of dangling objects is that you've rebased a
3199 branch, or you have pulled from somebody else who rebased a branch--see
3200 <<cleaning-up-history>>. In that case, the old head of the original
3201 branch still exists, as does everything it pointed to. The branch
3202 pointer itself just doesn't, since you replaced it with another one.
3203
3204 There are also other situations that cause dangling objects. For
3205 example, a "dangling blob" may arise because you did a "git add" of a
3206 file, but then, before you actually committed it and made it part of the
3207 bigger picture, you changed something else in that file and committed
3208 that *updated* thing--the old state that you added originally ends up
3209 not being pointed to by any commit or tree, so it's now a dangling blob
3210 object.
3211
3212 Similarly, when the "recursive" merge strategy runs, and finds that
3213 there are criss-cross merges and thus more than one merge base (which is
3214 fairly unusual, but it does happen), it will generate one temporary
3215 midway tree (or possibly even more, if you had lots of criss-crossing
3216 merges and more than two merge bases) as a temporary internal merge
3217 base, and again, those are real objects, but the end result will not end
3218 up pointing to them, so they end up "dangling" in your repository.
3219
3220 Generally, dangling objects aren't anything to worry about. They can
3221 even be very useful: if you screw something up, the dangling objects can
3222 be how you recover your old tree (say, you did a rebase, and realized
3223 that you really didn't want to--you can look at what dangling objects
3224 you have, and decide to reset your head to some old dangling state).
3225
3226 For commits, you can just use:
3227
3228 ------------------------------------------------
3229 $ gitk <dangling-commit-sha-goes-here> --not --all
3230 ------------------------------------------------
3231
3232 This asks for all the history reachable from the given commit but not
3233 from any branch, tag, or other reference. If you decide it's something
3234 you want, you can always create a new reference to it, e.g.,
3235
3236 ------------------------------------------------
3237 $ git branch recovered-branch <dangling-commit-sha-goes-here>
3238 ------------------------------------------------
3239
3240 For blobs and trees, you can't do the same, but you can still examine
3241 them. You can just do
3242
3243 ------------------------------------------------
3244 $ git show <dangling-blob/tree-sha-goes-here>
3245 ------------------------------------------------
3246
3247 to show what the contents of the blob were (or, for a tree, basically
3248 what the "ls" for that directory was), and that may give you some idea
3249 of what the operation was that left that dangling object.
3250
3251 Usually, dangling blobs and trees aren't very interesting. They're
3252 almost always the result of either being a half-way mergebase (the blob
3253 will often even have the conflict markers from a merge in it, if you
3254 have had conflicting merges that you fixed up by hand), or simply
3255 because you interrupted a "git fetch" with ^C or something like that,
3256 leaving _some_ of the new objects in the object database, but just
3257 dangling and useless.
3258
3259 Anyway, once you are sure that you're not interested in any dangling
3260 state, you can just prune all unreachable objects:
3261
3262 ------------------------------------------------
3263 $ git prune
3264 ------------------------------------------------
3265
3266 and they'll be gone. But you should only run "git prune" on a quiescent
3267 repository--it's kind of like doing a filesystem fsck recovery: you
3268 don't want to do that while the filesystem is mounted.
3269
3270 (The same is true of "git fsck" itself, btw, but since
3271 `git fsck` never actually *changes* the repository, it just reports
3272 on what it found, `git fsck` itself is never 'dangerous' to run.
3273 Running it while somebody is actually changing the repository can cause
3274 confusing and scary messages, but it won't actually do anything bad. In
3275 contrast, running "git prune" while somebody is actively changing the
3276 repository is a *BAD* idea).
3277
3278 [[recovering-from-repository-corruption]]
3279 Recovering from repository corruption
3280 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3281
3282 By design, git treats data trusted to it with caution. However, even in
3283 the absence of bugs in git itself, it is still possible that hardware or
3284 operating system errors could corrupt data.
3285
3286 The first defense against such problems is backups. You can back up a
3287 git directory using clone, or just using cp, tar, or any other backup
3288 mechanism.
3289
3290 As a last resort, you can search for the corrupted objects and attempt
3291 to replace them by hand. Back up your repository before attempting this
3292 in case you corrupt things even more in the process.
3293
3294 We'll assume that the problem is a single missing or corrupted blob,
3295 which is sometimes a solvable problem. (Recovering missing trees and
3296 especially commits is *much* harder).
3297
3298 Before starting, verify that there is corruption, and figure out where
3299 it is with linkgit:git-fsck[1]; this may be time-consuming.
3300
3301 Assume the output looks like this:
3302
3303 ------------------------------------------------
3304 $ git fsck --full --no-dangling
3305 broken link from tree 2d9263c6d23595e7cb2a21e5ebbb53655278dff8
3306 to blob 4b9458b3786228369c63936db65827de3cc06200
3307 missing blob 4b9458b3786228369c63936db65827de3cc06200
3308 ------------------------------------------------
3309
3310 Now you know that blob 4b9458b3 is missing, and that the tree 2d9263c6
3311 points to it. If you could find just one copy of that missing blob
3312 object, possibly in some other repository, you could move it into
3313 .git/objects/4b/9458b3... and be done. Suppose you can't. You can
3314 still examine the tree that pointed to it with linkgit:git-ls-tree[1],
3315 which might output something like:
3316
3317 ------------------------------------------------
3318 $ git ls-tree 2d9263c6d23595e7cb2a21e5ebbb53655278dff8
3319 100644 blob 8d14531846b95bfa3564b58ccfb7913a034323b8 .gitignore
3320 100644 blob ebf9bf84da0aab5ed944264a5db2a65fe3a3e883 .mailmap
3321 100644 blob ca442d313d86dc67e0a2e5d584b465bd382cbf5c COPYING
3322 ...
3323 100644 blob 4b9458b3786228369c63936db65827de3cc06200 myfile
3324 ...
3325 ------------------------------------------------
3326
3327 So now you know that the missing blob was the data for a file named
3328 "myfile". And chances are you can also identify the directory--let's
3329 say it's in "somedirectory". If you're lucky the missing copy might be
3330 the same as the copy you have checked out in your working tree at
3331 "somedirectory/myfile"; you can test whether that's right with
3332 linkgit:git-hash-object[1]:
3333
3334 ------------------------------------------------
3335 $ git hash-object -w somedirectory/myfile
3336 ------------------------------------------------
3337
3338 which will create and store a blob object with the contents of
3339 somedirectory/myfile, and output the SHA-1 of that object. if you're
3340 extremely lucky it might be 4b9458b3786228369c63936db65827de3cc06200, in
3341 which case you've guessed right, and the corruption is fixed!
3342
3343 Otherwise, you need more information. How do you tell which version of
3344 the file has been lost?
3345
3346 The easiest way to do this is with:
3347
3348 ------------------------------------------------
3349 $ git log --raw --all --full-history -- somedirectory/myfile
3350 ------------------------------------------------
3351
3352 Because you're asking for raw output, you'll now get something like
3353
3354 ------------------------------------------------
3355 commit abc
3356 Author:
3357 Date:
3358 ...
3359 :100644 100644 4b9458b... newsha... M somedirectory/myfile
3360
3361
3362 commit xyz
3363 Author:
3364 Date:
3365
3366 ...
3367 :100644 100644 oldsha... 4b9458b... M somedirectory/myfile
3368 ------------------------------------------------
3369
3370 This tells you that the immediately following version of the file was
3371 "newsha", and that the immediately preceding version was "oldsha".
3372 You also know the commit messages that went with the change from oldsha
3373 to 4b9458b and with the change from 4b9458b to newsha.
3374
3375 If you've been committing small enough changes, you may now have a good
3376 shot at reconstructing the contents of the in-between state 4b9458b.
3377
3378 If you can do that, you can now recreate the missing object with
3379
3380 ------------------------------------------------
3381 $ git hash-object -w <recreated-file>
3382 ------------------------------------------------
3383
3384 and your repository is good again!
3385
3386 (Btw, you could have ignored the fsck, and started with doing a
3387
3388 ------------------------------------------------
3389 $ git log --raw --all
3390 ------------------------------------------------
3391
3392 and just looked for the sha of the missing object (4b9458b..) in that
3393 whole thing. It's up to you - git does *have* a lot of information, it is
3394 just missing one particular blob version.
3395
3396 [[the-index]]
3397 The index
3398 -----------
3399
3400 The index is a binary file (generally kept in .git/index) containing a
3401 sorted list of path names, each with permissions and the SHA-1 of a blob
3402 object; linkgit:git-ls-files[1] can show you the contents of the index:
3403
3404 -------------------------------------------------
3405 $ git ls-files --stage
3406 100644 63c918c667fa005ff12ad89437f2fdc80926e21c 0 .gitignore
3407 100644 5529b198e8d14decbe4ad99db3f7fb632de0439d 0 .mailmap
3408 100644 6ff87c4664981e4397625791c8ea3bbb5f2279a3 0 COPYING
3409 100644 a37b2152bd26be2c2289e1f57a292534a51a93c7 0 Documentation/.gitignore
3410 100644 fbefe9a45b00a54b58d94d06eca48b03d40a50e0 0 Documentation/Makefile
3411 ...
3412 100644 2511aef8d89ab52be5ec6a5e46236b4b6bcd07ea 0 xdiff/xtypes.h
3413 100644 2ade97b2574a9f77e7ae4002a4e07a6a38e46d07 0 xdiff/xutils.c
3414 100644 d5de8292e05e7c36c4b68857c1cf9855e3d2f70a 0 xdiff/xutils.h
3415 -------------------------------------------------
3416
3417 Note that in older documentation you may see the index called the
3418 "current directory cache" or just the "cache". It has three important
3419 properties:
3420
3421 1. The index contains all the information necessary to generate a single
3422 (uniquely determined) tree object.
3423 +
3424 For example, running linkgit:git-commit[1] generates this tree object
3425 from the index, stores it in the object database, and uses it as the
3426 tree object associated with the new commit.
3427
3428 2. The index enables fast comparisons between the tree object it defines
3429 and the working tree.
3430 +
3431 It does this by storing some additional data for each entry (such as
3432 the last modified time). This data is not displayed above, and is not
3433 stored in the created tree object, but it can be used to determine
3434 quickly which files in the working directory differ from what was
3435 stored in the index, and thus save git from having to read all of the
3436 data from such files to look for changes.
3437
3438 3. It can efficiently represent information about merge conflicts
3439 between different tree objects, allowing each pathname to be
3440 associated with sufficient information about the trees involved that
3441 you can create a three-way merge between them.
3442 +
3443 We saw in <<conflict-resolution>> that during a merge the index can
3444 store multiple versions of a single file (called "stages"). The third
3445 column in the linkgit:git-ls-files[1] output above is the stage
3446 number, and will take on values other than 0 for files with merge
3447 conflicts.
3448
3449 The index is thus a sort of temporary staging area, which is filled with
3450 a tree which you are in the process of working on.
3451
3452 If you blow the index away entirely, you generally haven't lost any
3453 information as long as you have the name of the tree that it described.
3454
3455 [[submodules]]
3456 Submodules
3457 ==========
3458
3459 Large projects are often composed of smaller, self-contained modules. For
3460 example, an embedded Linux distribution's source tree would include every
3461 piece of software in the distribution with some local modifications; a movie
3462 player might need to build against a specific, known-working version of a
3463 decompression library; several independent programs might all share the same
3464 build scripts.
3465
3466 With centralized revision control systems this is often accomplished by
3467 including every module in one single repository. Developers can check out
3468 all modules or only the modules they need to work with. They can even modify
3469 files across several modules in a single commit while moving things around
3470 or updating APIs and translations.
3471
3472 Git does not allow partial checkouts, so duplicating this approach in Git
3473 would force developers to keep a local copy of modules they are not
3474 interested in touching. Commits in an enormous checkout would be slower
3475 than you'd expect as Git would have to scan every directory for changes.
3476 If modules have a lot of local history, clones would take forever.
3477
3478 On the plus side, distributed revision control systems can much better
3479 integrate with external sources. In a centralized model, a single arbitrary
3480 snapshot of the external project is exported from its own revision control
3481 and then imported into the local revision control on a vendor branch. All
3482 the history is hidden. With distributed revision control you can clone the
3483 entire external history and much more easily follow development and re-merge
3484 local changes.
3485
3486 Git's submodule support allows a repository to contain, as a subdirectory, a
3487 checkout of an external project. Submodules maintain their own identity;
3488 the submodule support just stores the submodule repository location and
3489 commit ID, so other developers who clone the containing project
3490 ("superproject") can easily clone all the submodules at the same revision.
3491 Partial checkouts of the superproject are possible: you can tell Git to
3492 clone none, some or all of the submodules.
3493
3494 The linkgit:git-submodule[1] command is available since Git 1.5.3. Users
3495 with Git 1.5.2 can look up the submodule commits in the repository and
3496 manually check them out; earlier versions won't recognize the submodules at
3497 all.
3498
3499 To see how submodule support works, create (for example) four example
3500 repositories that can be used later as a submodule:
3501
3502 -------------------------------------------------
3503 $ mkdir ~/git
3504 $ cd ~/git
3505 $ for i in a b c d
3506 do
3507 mkdir $i
3508 cd $i
3509 git init
3510 echo "module $i" > $i.txt
3511 git add $i.txt
3512 git commit -m "Initial commit, submodule $i"
3513 cd ..
3514 done
3515 -------------------------------------------------
3516
3517 Now create the superproject and add all the submodules:
3518
3519 -------------------------------------------------
3520 $ mkdir super
3521 $ cd super
3522 $ git init
3523 $ for i in a b c d
3524 do
3525 git submodule add ~/git/$i $i
3526 done
3527 -------------------------------------------------
3528
3529 NOTE: Do not use local URLs here if you plan to publish your superproject!
3530
3531 See what files `git submodule` created:
3532
3533 -------------------------------------------------
3534 $ ls -a
3535 . .. .git .gitmodules a b c d
3536 -------------------------------------------------
3537
3538 The `git submodule add <repo> <path>` command does a couple of things:
3539
3540 - It clones the submodule from <repo> to the given <path> under the
3541 current directory and by default checks out the master branch.
3542 - It adds the submodule's clone path to the linkgit:gitmodules[5] file and
3543 adds this file to the index, ready to be committed.
3544 - It adds the submodule's current commit ID to the index, ready to be
3545 committed.
3546
3547 Commit the superproject:
3548
3549 -------------------------------------------------
3550 $ git commit -m "Add submodules a, b, c and d."
3551 -------------------------------------------------
3552
3553 Now clone the superproject:
3554
3555 -------------------------------------------------
3556 $ cd ..
3557 $ git clone super cloned
3558 $ cd cloned
3559 -------------------------------------------------
3560
3561 The submodule directories are there, but they're empty:
3562
3563 -------------------------------------------------
3564 $ ls -a a
3565 . ..
3566 $ git submodule status
3567 -d266b9873ad50488163457f025db7cdd9683d88b a
3568 -e81d457da15309b4fef4249aba9b50187999670d b
3569 -c1536a972b9affea0f16e0680ba87332dc059146 c
3570 -d96249ff5d57de5de093e6baff9e0aafa5276a74 d
3571 -------------------------------------------------
3572
3573 NOTE: The commit object names shown above would be different for you, but they
3574 should match the HEAD commit object names of your repositories. You can check
3575 it by running `git ls-remote ../a`.
3576
3577 Pulling down the submodules is a two-step process. First run `git submodule
3578 init` to add the submodule repository URLs to `.git/config`:
3579
3580 -------------------------------------------------
3581 $ git submodule init
3582 -------------------------------------------------
3583
3584 Now use `git submodule update` to clone the repositories and check out the
3585 commits specified in the superproject:
3586
3587 -------------------------------------------------
3588 $ git submodule update
3589 $ cd a
3590 $ ls -a
3591 . .. .git a.txt
3592 -------------------------------------------------
3593
3594 One major difference between `git submodule update` and `git submodule add` is
3595 that `git submodule update` checks out a specific commit, rather than the tip
3596 of a branch. It's like checking out a tag: the head is detached, so you're not
3597 working on a branch.
3598
3599 -------------------------------------------------
3600 $ git branch
3601 * (no branch)
3602 master
3603 -------------------------------------------------
3604
3605 If you want to make a change within a submodule and you have a detached head,
3606 then you should create or checkout a branch, make your changes, publish the
3607 change within the submodule, and then update the superproject to reference the
3608 new commit:
3609
3610 -------------------------------------------------
3611 $ git checkout master
3612 -------------------------------------------------
3613
3614 or
3615
3616 -------------------------------------------------
3617 $ git checkout -b fix-up
3618 -------------------------------------------------
3619
3620 then
3621
3622 -------------------------------------------------
3623 $ echo "adding a line again" >> a.txt
3624 $ git commit -a -m "Updated the submodule from within the superproject."
3625 $ git push
3626 $ cd ..
3627 $ git diff
3628 diff --git a/a b/a
3629 index d266b98..261dfac 160000
3630 --- a/a
3631 +++ b/a
3632 @@ -1 +1 @@
3633 -Subproject commit d266b9873ad50488163457f025db7cdd9683d88b
3634 +Subproject commit 261dfac35cb99d380eb966e102c1197139f7fa24
3635 $ git add a
3636 $ git commit -m "Updated submodule a."
3637 $ git push
3638 -------------------------------------------------
3639
3640 You have to run `git submodule update` after `git pull` if you want to update
3641 submodules, too.
3642
3643 Pitfalls with submodules
3644 ------------------------
3645
3646 Always publish the submodule change before publishing the change to the
3647 superproject that references it. If you forget to publish the submodule change,
3648 others won't be able to clone the repository:
3649
3650 -------------------------------------------------
3651 $ cd ~/git/super/a
3652 $ echo i added another line to this file >> a.txt
3653 $ git commit -a -m "doing it wrong this time"
3654 $ cd ..
3655 $ git add a
3656 $ git commit -m "Updated submodule a again."
3657 $ git push
3658 $ cd ~/git/cloned
3659 $ git pull
3660 $ git submodule update
3661 error: pathspec '261dfac35cb99d380eb966e102c1197139f7fa24' did not match any file(s) known to git.
3662 Did you forget to 'git add'?
3663 Unable to checkout '261dfac35cb99d380eb966e102c1197139f7fa24' in submodule path 'a'
3664 -------------------------------------------------
3665
3666 In older git versions it could be easily forgotten to commit new or modified
3667 files in a submodule, which silently leads to similar problems as not pushing
3668 the submodule changes. Starting with git 1.7.0 both "git status" and "git diff"
3669 in the superproject show submodules as modified when they contain new or
3670 modified files to protect against accidentally committing such a state. "git
3671 diff" will also add a "-dirty" to the work tree side when generating patch
3672 output or used with the --submodule option:
3673
3674 -------------------------------------------------
3675 $ git diff
3676 diff --git a/sub b/sub
3677 --- a/sub
3678 +++ b/sub
3679 @@ -1 +1 @@
3680 -Subproject commit 3f356705649b5d566d97ff843cf193359229a453
3681 +Subproject commit 3f356705649b5d566d97ff843cf193359229a453-dirty
3682 $ git diff --submodule
3683 Submodule sub 3f35670..3f35670-dirty:
3684 -------------------------------------------------
3685
3686 You also should not rewind branches in a submodule beyond commits that were
3687 ever recorded in any superproject.
3688
3689 It's not safe to run `git submodule update` if you've made and committed
3690 changes within a submodule without checking out a branch first. They will be
3691 silently overwritten:
3692
3693 -------------------------------------------------
3694 $ cat a.txt
3695 module a
3696 $ echo line added from private2 >> a.txt
3697 $ git commit -a -m "line added inside private2"
3698 $ cd ..
3699 $ git submodule update
3700 Submodule path 'a': checked out 'd266b9873ad50488163457f025db7cdd9683d88b'
3701 $ cd a
3702 $ cat a.txt
3703 module a
3704 -------------------------------------------------
3705
3706 NOTE: The changes are still visible in the submodule's reflog.
3707
3708 This is not the case if you did not commit your changes.
3709
3710 [[low-level-operations]]
3711 Low-level git operations
3712 ========================
3713
3714 Many of the higher-level commands were originally implemented as shell
3715 scripts using a smaller core of low-level git commands. These can still
3716 be useful when doing unusual things with git, or just as a way to
3717 understand its inner workings.
3718
3719 [[object-manipulation]]
3720 Object access and manipulation
3721 ------------------------------
3722
3723 The linkgit:git-cat-file[1] command can show the contents of any object,
3724 though the higher-level linkgit:git-show[1] is usually more useful.
3725
3726 The linkgit:git-commit-tree[1] command allows constructing commits with
3727 arbitrary parents and trees.
3728
3729 A tree can be created with linkgit:git-write-tree[1] and its data can be
3730 accessed by linkgit:git-ls-tree[1]. Two trees can be compared with
3731 linkgit:git-diff-tree[1].
3732
3733 A tag is created with linkgit:git-mktag[1], and the signature can be
3734 verified by linkgit:git-verify-tag[1], though it is normally simpler to
3735 use linkgit:git-tag[1] for both.
3736
3737 [[the-workflow]]
3738 The Workflow
3739 ------------
3740
3741 High-level operations such as linkgit:git-commit[1],
3742 linkgit:git-checkout[1] and linkgit:git-reset[1] work by moving data
3743 between the working tree, the index, and the object database. Git
3744 provides low-level operations which perform each of these steps
3745 individually.
3746
3747 Generally, all "git" operations work on the index file. Some operations
3748 work *purely* on the index file (showing the current state of the
3749 index), but most operations move data between the index file and either
3750 the database or the working directory. Thus there are four main
3751 combinations:
3752
3753 [[working-directory-to-index]]
3754 working directory -> index
3755 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3756
3757 The linkgit:git-update-index[1] command updates the index with
3758 information from the working directory. You generally update the
3759 index information by just specifying the filename you want to update,
3760 like so:
3761
3762 -------------------------------------------------
3763 $ git update-index filename
3764 -------------------------------------------------
3765
3766 but to avoid common mistakes with filename globbing etc, the command
3767 will not normally add totally new entries or remove old entries,
3768 i.e. it will normally just update existing cache entries.
3769
3770 To tell git that yes, you really do realize that certain files no
3771 longer exist, or that new files should be added, you
3772 should use the `--remove` and `--add` flags respectively.
3773
3774 NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
3775 necessarily be removed: if the files still exist in your directory
3776 structure, the index will be updated with their new status, not
3777 removed. The only thing `--remove` means is that update-index will be
3778 considering a removed file to be a valid thing, and if the file really
3779 does not exist any more, it will update the index accordingly.
3780
3781 As a special case, you can also do `git update-index --refresh`, which
3782 will refresh the "stat" information of each index to match the current
3783 stat information. It will 'not' update the object status itself, and
3784 it will only update the fields that are used to quickly test whether
3785 an object still matches its old backing store object.
3786
3787 The previously introduced linkgit:git-add[1] is just a wrapper for
3788 linkgit:git-update-index[1].
3789
3790 [[index-to-object-database]]
3791 index -> object database
3792 ~~~~~~~~~~~~~~~~~~~~~~~~
3793
3794 You write your current index file to a "tree" object with the program
3795
3796 -------------------------------------------------
3797 $ git write-tree
3798 -------------------------------------------------
3799
3800 that doesn't come with any options--it will just write out the
3801 current index into the set of tree objects that describe that state,
3802 and it will return the name of the resulting top-level tree. You can
3803 use that tree to re-generate the index at any time by going in the
3804 other direction:
3805
3806 [[object-database-to-index]]
3807 object database -> index
3808 ~~~~~~~~~~~~~~~~~~~~~~~~
3809
3810 You read a "tree" file from the object database, and use that to
3811 populate (and overwrite--don't do this if your index contains any
3812 unsaved state that you might want to restore later!) your current
3813 index. Normal operation is just
3814
3815 -------------------------------------------------
3816 $ git read-tree <SHA-1 of tree>
3817 -------------------------------------------------
3818
3819 and your index file will now be equivalent to the tree that you saved
3820 earlier. However, that is only your 'index' file: your working
3821 directory contents have not been modified.
3822
3823 [[index-to-working-directory]]
3824 index -> working directory
3825 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3826
3827 You update your working directory from the index by "checking out"
3828 files. This is not a very common operation, since normally you'd just
3829 keep your files updated, and rather than write to your working
3830 directory, you'd tell the index files about the changes in your
3831 working directory (i.e. `git update-index`).
3832
3833 However, if you decide to jump to a new version, or check out somebody
3834 else's version, or just restore a previous tree, you'd populate your
3835 index file with read-tree, and then you need to check out the result
3836 with
3837
3838 -------------------------------------------------
3839 $ git checkout-index filename
3840 -------------------------------------------------
3841
3842 or, if you want to check out all of the index, use `-a`.
3843
3844 NOTE! `git checkout-index` normally refuses to overwrite old files, so
3845 if you have an old version of the tree already checked out, you will
3846 need to use the "-f" flag ('before' the "-a" flag or the filename) to
3847 'force' the checkout.
3848
3849
3850 Finally, there are a few odds and ends which are not purely moving
3851 from one representation to the other:
3852
3853 [[tying-it-all-together]]
3854 Tying it all together
3855 ~~~~~~~~~~~~~~~~~~~~~
3856
3857 To commit a tree you have instantiated with "git write-tree", you'd
3858 create a "commit" object that refers to that tree and the history
3859 behind it--most notably the "parent" commits that preceded it in
3860 history.
3861
3862 Normally a "commit" has one parent: the previous state of the tree
3863 before a certain change was made. However, sometimes it can have two
3864 or more parent commits, in which case we call it a "merge", due to the
3865 fact that such a commit brings together ("merges") two or more
3866 previous states represented by other commits.
3867
3868 In other words, while a "tree" represents a particular directory state
3869 of a working directory, a "commit" represents that state in "time",
3870 and explains how we got there.
3871
3872 You create a commit object by giving it the tree that describes the
3873 state at the time of the commit, and a list of parents:
3874
3875 -------------------------------------------------
3876 $ git commit-tree <tree> -p <parent> [(-p <parent2>)...]
3877 -------------------------------------------------
3878
3879 and then giving the reason for the commit on stdin (either through
3880 redirection from a pipe or file, or by just typing it at the tty).
3881
3882 `git commit-tree` will return the name of the object that represents
3883 that commit, and you should save it away for later use. Normally,
3884 you'd commit a new `HEAD` state, and while git doesn't care where you
3885 save the note about that state, in practice we tend to just write the
3886 result to the file pointed at by `.git/HEAD`, so that we can always see
3887 what the last committed state was.
3888
3889 Here is an ASCII art by Jon Loeliger that illustrates how
3890 various pieces fit together.
3891
3892 ------------
3893
3894 commit-tree
3895 commit obj
3896 +----+
3897 | |
3898 | |
3899 V V
3900 +-----------+
3901 | Object DB |
3902 | Backing |
3903 | Store |
3904 +-----------+
3905 ^
3906 write-tree | |
3907 tree obj | |
3908 | | read-tree
3909 | | tree obj
3910 V
3911 +-----------+
3912 | Index |
3913 | "cache" |
3914 +-----------+
3915 update-index ^
3916 blob obj | |
3917 | |
3918 checkout-index -u | | checkout-index
3919 stat | | blob obj
3920 V
3921 +-----------+
3922 | Working |
3923 | Directory |
3924 +-----------+
3925
3926 ------------
3927
3928
3929 [[examining-the-data]]
3930 Examining the data
3931 ------------------
3932
3933 You can examine the data represented in the object database and the
3934 index with various helper tools. For every object, you can use
3935 linkgit:git-cat-file[1] to examine details about the
3936 object:
3937
3938 -------------------------------------------------
3939 $ git cat-file -t <objectname>
3940 -------------------------------------------------
3941
3942 shows the type of the object, and once you have the type (which is
3943 usually implicit in where you find the object), you can use
3944
3945 -------------------------------------------------
3946 $ git cat-file blob|tree|commit|tag <objectname>
3947 -------------------------------------------------
3948
3949 to show its contents. NOTE! Trees have binary content, and as a result
3950 there is a special helper for showing that content, called
3951 `git ls-tree`, which turns the binary content into a more easily
3952 readable form.
3953
3954 It's especially instructive to look at "commit" objects, since those
3955 tend to be small and fairly self-explanatory. In particular, if you
3956 follow the convention of having the top commit name in `.git/HEAD`,
3957 you can do
3958
3959 -------------------------------------------------
3960 $ git cat-file commit HEAD
3961 -------------------------------------------------
3962
3963 to see what the top commit was.
3964
3965 [[merging-multiple-trees]]
3966 Merging multiple trees
3967 ----------------------
3968
3969 Git helps you do a three-way merge, which you can expand to n-way by
3970 repeating the merge procedure arbitrary times until you finally
3971 "commit" the state. The normal situation is that you'd only do one
3972 three-way merge (two parents), and commit it, but if you like to, you
3973 can do multiple parents in one go.
3974
3975 To do a three-way merge, you need the two sets of "commit" objects
3976 that you want to merge, use those to find the closest common parent (a
3977 third "commit" object), and then use those commit objects to find the
3978 state of the directory ("tree" object) at these points.
3979
3980 To get the "base" for the merge, you first look up the common parent
3981 of two commits with
3982
3983 -------------------------------------------------
3984 $ git merge-base <commit1> <commit2>
3985 -------------------------------------------------
3986
3987 which will return you the commit they are both based on. You should
3988 now look up the "tree" objects of those commits, which you can easily
3989 do with (for example)
3990
3991 -------------------------------------------------
3992 $ git cat-file commit <commitname> | head -1
3993 -------------------------------------------------
3994
3995 since the tree object information is always the first line in a commit
3996 object.
3997
3998 Once you know the three trees you are going to merge (the one "original"
3999 tree, aka the common tree, and the two "result" trees, aka the branches
4000 you want to merge), you do a "merge" read into the index. This will
4001 complain if it has to throw away your old index contents, so you should
4002 make sure that you've committed those--in fact you would normally
4003 always do a merge against your last commit (which should thus match what
4004 you have in your current index anyway).
4005
4006 To do the merge, do
4007
4008 -------------------------------------------------
4009 $ git read-tree -m -u <origtree> <yourtree> <targettree>
4010 -------------------------------------------------
4011
4012 which will do all trivial merge operations for you directly in the
4013 index file, and you can just write the result out with
4014 `git write-tree`.
4015
4016
4017 [[merging-multiple-trees-2]]
4018 Merging multiple trees, continued
4019 ---------------------------------
4020
4021 Sadly, many merges aren't trivial. If there are files that have
4022 been added, moved or removed, or if both branches have modified the
4023 same file, you will be left with an index tree that contains "merge
4024 entries" in it. Such an index tree can 'NOT' be written out to a tree
4025 object, and you will have to resolve any such merge clashes using
4026 other tools before you can write out the result.
4027
4028 You can examine such index state with `git ls-files --unmerged`
4029 command. An example:
4030
4031 ------------------------------------------------
4032 $ git read-tree -m $orig HEAD $target
4033 $ git ls-files --unmerged
4034 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
4035 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
4036 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
4037 ------------------------------------------------
4038
4039 Each line of the `git ls-files --unmerged` output begins with
4040 the blob mode bits, blob SHA-1, 'stage number', and the
4041 filename. The 'stage number' is git's way to say which tree it
4042 came from: stage 1 corresponds to the `$orig` tree, stage 2 to
4043 the `HEAD` tree, and stage 3 to the `$target` tree.
4044
4045 Earlier we said that trivial merges are done inside
4046 `git read-tree -m`. For example, if the file did not change
4047 from `$orig` to `HEAD` nor `$target`, or if the file changed
4048 from `$orig` to `HEAD` and `$orig` to `$target` the same way,
4049 obviously the final outcome is what is in `HEAD`. What the
4050 above example shows is that file `hello.c` was changed from
4051 `$orig` to `HEAD` and `$orig` to `$target` in a different way.
4052 You could resolve this by running your favorite 3-way merge
4053 program, e.g. `diff3`, `merge`, or git's own merge-file, on
4054 the blob objects from these three stages yourself, like this:
4055
4056 ------------------------------------------------
4057 $ git cat-file blob 263414f... >hello.c~1
4058 $ git cat-file blob 06fa6a2... >hello.c~2
4059 $ git cat-file blob cc44c73... >hello.c~3
4060 $ git merge-file hello.c~2 hello.c~1 hello.c~3
4061 ------------------------------------------------
4062
4063 This would leave the merge result in `hello.c~2` file, along
4064 with conflict markers if there are conflicts. After verifying
4065 the merge result makes sense, you can tell git what the final
4066 merge result for this file is by:
4067
4068 -------------------------------------------------
4069 $ mv -f hello.c~2 hello.c
4070 $ git update-index hello.c
4071 -------------------------------------------------
4072
4073 When a path is in the "unmerged" state, running `git update-index` for
4074 that path tells git to mark the path resolved.
4075
4076 The above is the description of a git merge at the lowest level,
4077 to help you understand what conceptually happens under the hood.
4078 In practice, nobody, not even git itself, runs `git cat-file` three times
4079 for this. There is a `git merge-index` program that extracts the
4080 stages to temporary files and calls a "merge" script on it:
4081
4082 -------------------------------------------------
4083 $ git merge-index git-merge-one-file hello.c
4084 -------------------------------------------------
4085
4086 and that is what higher level `git merge -s resolve` is implemented with.
4087
4088 [[hacking-git]]
4089 Hacking git
4090 ===========
4091
4092 This chapter covers internal details of the git implementation which
4093 probably only git developers need to understand.
4094
4095 [[object-details]]
4096 Object storage format
4097 ---------------------
4098
4099 All objects have a statically determined "type" which identifies the
4100 format of the object (i.e. how it is used, and how it can refer to other
4101 objects). There are currently four different object types: "blob",
4102 "tree", "commit", and "tag".
4103
4104 Regardless of object type, all objects share the following
4105 characteristics: they are all deflated with zlib, and have a header
4106 that not only specifies their type, but also provides size information
4107 about the data in the object. It's worth noting that the SHA-1 hash
4108 that is used to name the object is the hash of the original data
4109 plus this header, so `sha1sum` 'file' does not match the object name
4110 for 'file'.
4111 (Historical note: in the dawn of the age of git the hash
4112 was the SHA-1 of the 'compressed' object.)
4113
4114 As a result, the general consistency of an object can always be tested
4115 independently of the contents or the type of the object: all objects can
4116 be validated by verifying that (a) their hashes match the content of the
4117 file and (b) the object successfully inflates to a stream of bytes that
4118 forms a sequence of <ascii type without space> {plus} <space> {plus} <ascii decimal
4119 size> {plus} <byte\0> {plus} <binary object data>.
4120
4121 The structured objects can further have their structure and
4122 connectivity to other objects verified. This is generally done with
4123 the `git fsck` program, which generates a full dependency graph
4124 of all objects, and verifies their internal consistency (in addition
4125 to just verifying their superficial consistency through the hash).
4126
4127 [[birdview-on-the-source-code]]
4128 A birds-eye view of Git's source code
4129 -------------------------------------
4130
4131 It is not always easy for new developers to find their way through Git's
4132 source code. This section gives you a little guidance to show where to
4133 start.
4134
4135 A good place to start is with the contents of the initial commit, with:
4136
4137 ----------------------------------------------------
4138 $ git checkout e83c5163
4139 ----------------------------------------------------
4140
4141 The initial revision lays the foundation for almost everything git has
4142 today, but is small enough to read in one sitting.
4143
4144 Note that terminology has changed since that revision. For example, the
4145 README in that revision uses the word "changeset" to describe what we
4146 now call a <<def_commit_object,commit>>.
4147
4148 Also, we do not call it "cache" any more, but rather "index"; however, the
4149 file is still called `cache.h`. Remark: Not much reason to change it now,
4150 especially since there is no good single name for it anyway, because it is
4151 basically _the_ header file which is included by _all_ of Git's C sources.
4152
4153 If you grasp the ideas in that initial commit, you should check out a
4154 more recent version and skim `cache.h`, `object.h` and `commit.h`.
4155
4156 In the early days, Git (in the tradition of UNIX) was a bunch of programs
4157 which were extremely simple, and which you used in scripts, piping the
4158 output of one into another. This turned out to be good for initial
4159 development, since it was easier to test new things. However, recently
4160 many of these parts have become builtins, and some of the core has been
4161 "libified", i.e. put into libgit.a for performance, portability reasons,
4162 and to avoid code duplication.
4163
4164 By now, you know what the index is (and find the corresponding data
4165 structures in `cache.h`), and that there are just a couple of object types
4166 (blobs, trees, commits and tags) which inherit their common structure from
4167 `struct object`, which is their first member (and thus, you can cast e.g.
4168 `(struct object *)commit` to achieve the _same_ as `&commit->object`, i.e.
4169 get at the object name and flags).
4170
4171 Now is a good point to take a break to let this information sink in.
4172
4173 Next step: get familiar with the object naming. Read <<naming-commits>>.
4174 There are quite a few ways to name an object (and not only revisions!).
4175 All of these are handled in `sha1_name.c`. Just have a quick look at
4176 the function `get_sha1()`. A lot of the special handling is done by
4177 functions like `get_sha1_basic()` or the likes.
4178
4179 This is just to get you into the groove for the most libified part of Git:
4180 the revision walker.
4181
4182 Basically, the initial version of `git log` was a shell script:
4183
4184 ----------------------------------------------------------------
4185 $ git-rev-list --pretty $(git-rev-parse --default HEAD "$@") | \
4186 LESS=-S ${PAGER:-less}
4187 ----------------------------------------------------------------
4188
4189 What does this mean?
4190
4191 `git rev-list` is the original version of the revision walker, which
4192 _always_ printed a list of revisions to stdout. It is still functional,
4193 and needs to, since most new Git commands start out as scripts using
4194 `git rev-list`.
4195
4196 `git rev-parse` is not as important any more; it was only used to filter out
4197 options that were relevant for the different plumbing commands that were
4198 called by the script.
4199
4200 Most of what `git rev-list` did is contained in `revision.c` and
4201 `revision.h`. It wraps the options in a struct named `rev_info`, which
4202 controls how and what revisions are walked, and more.
4203
4204 The original job of `git rev-parse` is now taken by the function
4205 `setup_revisions()`, which parses the revisions and the common command line
4206 options for the revision walker. This information is stored in the struct
4207 `rev_info` for later consumption. You can do your own command line option
4208 parsing after calling `setup_revisions()`. After that, you have to call
4209 `prepare_revision_walk()` for initialization, and then you can get the
4210 commits one by one with the function `get_revision()`.
4211
4212 If you are interested in more details of the revision walking process,
4213 just have a look at the first implementation of `cmd_log()`; call
4214 `git show v1.3.0~155^2~4` and scroll down to that function (note that you
4215 no longer need to call `setup_pager()` directly).
4216
4217 Nowadays, `git log` is a builtin, which means that it is _contained_ in the
4218 command `git`. The source side of a builtin is
4219
4220 - a function called `cmd_<bla>`, typically defined in `builtin-<bla>.c`,
4221 and declared in `builtin.h`,
4222
4223 - an entry in the `commands[]` array in `git.c`, and
4224
4225 - an entry in `BUILTIN_OBJECTS` in the `Makefile`.
4226
4227 Sometimes, more than one builtin is contained in one source file. For
4228 example, `cmd_whatchanged()` and `cmd_log()` both reside in `builtin-log.c`,
4229 since they share quite a bit of code. In that case, the commands which are
4230 _not_ named like the `.c` file in which they live have to be listed in
4231 `BUILT_INS` in the `Makefile`.
4232
4233 `git log` looks more complicated in C than it does in the original script,
4234 but that allows for a much greater flexibility and performance.
4235
4236 Here again it is a good point to take a pause.
4237
4238 Lesson three is: study the code. Really, it is the best way to learn about
4239 the organization of Git (after you know the basic concepts).
4240
4241 So, think about something which you are interested in, say, "how can I
4242 access a blob just knowing the object name of it?". The first step is to
4243 find a Git command with which you can do it. In this example, it is either
4244 `git show` or `git cat-file`.
4245
4246 For the sake of clarity, let's stay with `git cat-file`, because it
4247
4248 - is plumbing, and
4249
4250 - was around even in the initial commit (it literally went only through
4251 some 20 revisions as `cat-file.c`, was renamed to `builtin-cat-file.c`
4252 when made a builtin, and then saw less than 10 versions).
4253
4254 So, look into `builtin-cat-file.c`, search for `cmd_cat_file()` and look what
4255 it does.
4256
4257 ------------------------------------------------------------------
4258 git_config(git_default_config);
4259 if (argc != 3)
4260 usage("git cat-file [-t|-s|-e|-p|<type>] <sha1>");
4261 if (get_sha1(argv[2], sha1))
4262 die("Not a valid object name %s", argv[2]);
4263 ------------------------------------------------------------------
4264
4265 Let's skip over the obvious details; the only really interesting part
4266 here is the call to `get_sha1()`. It tries to interpret `argv[2]` as an
4267 object name, and if it refers to an object which is present in the current
4268 repository, it writes the resulting SHA-1 into the variable `sha1`.
4269
4270 Two things are interesting here:
4271
4272 - `get_sha1()` returns 0 on _success_. This might surprise some new
4273 Git hackers, but there is a long tradition in UNIX to return different
4274 negative numbers in case of different errors--and 0 on success.
4275
4276 - the variable `sha1` in the function signature of `get_sha1()` is `unsigned
4277 char *`, but is actually expected to be a pointer to `unsigned
4278 char[20]`. This variable will contain the 160-bit SHA-1 of the given
4279 commit. Note that whenever a SHA-1 is passed as `unsigned char *`, it
4280 is the binary representation, as opposed to the ASCII representation in
4281 hex characters, which is passed as `char *`.
4282
4283 You will see both of these things throughout the code.
4284
4285 Now, for the meat:
4286
4287 -----------------------------------------------------------------------------
4288 case 0:
4289 buf = read_object_with_reference(sha1, argv[1], &size, NULL);
4290 -----------------------------------------------------------------------------
4291
4292 This is how you read a blob (actually, not only a blob, but any type of
4293 object). To know how the function `read_object_with_reference()` actually
4294 works, find the source code for it (something like `git grep
4295 read_object_with | grep ":[a-z]"` in the git repository), and read
4296 the source.
4297
4298 To find out how the result can be used, just read on in `cmd_cat_file()`:
4299
4300 -----------------------------------
4301 write_or_die(1, buf, size);
4302 -----------------------------------
4303
4304 Sometimes, you do not know where to look for a feature. In many such cases,
4305 it helps to search through the output of `git log`, and then `git show` the
4306 corresponding commit.
4307
4308 Example: If you know that there was some test case for `git bundle`, but
4309 do not remember where it was (yes, you _could_ `git grep bundle t/`, but that
4310 does not illustrate the point!):
4311
4312 ------------------------
4313 $ git log --no-merges t/
4314 ------------------------
4315
4316 In the pager (`less`), just search for "bundle", go a few lines back,
4317 and see that it is in commit 18449ab0... Now just copy this object name,
4318 and paste it into the command line
4319
4320 -------------------
4321 $ git show 18449ab0
4322 -------------------
4323
4324 Voila.
4325
4326 Another example: Find out what to do in order to make some script a
4327 builtin:
4328
4329 -------------------------------------------------
4330 $ git log --no-merges --diff-filter=A builtin-*.c
4331 -------------------------------------------------
4332
4333 You see, Git is actually the best tool to find out about the source of Git
4334 itself!
4335
4336 [[glossary]]
4337 Git Glossary
4338 ============
4339
4340 include::glossary-content.txt[]
4341
4342 [[git-quick-start]]
4343 Appendix A: Git Quick Reference
4344 ===============================
4345
4346 This is a quick summary of the major commands; the previous chapters
4347 explain how these work in more detail.
4348
4349 [[quick-creating-a-new-repository]]
4350 Creating a new repository
4351 -------------------------
4352
4353 From a tarball:
4354
4355 -----------------------------------------------
4356 $ tar xzf project.tar.gz
4357 $ cd project
4358 $ git init
4359 Initialized empty Git repository in .git/
4360 $ git add .
4361 $ git commit
4362 -----------------------------------------------
4363
4364 From a remote repository:
4365
4366 -----------------------------------------------
4367 $ git clone git://example.com/pub/project.git
4368 $ cd project
4369 -----------------------------------------------
4370
4371 [[managing-branches]]
4372 Managing branches
4373 -----------------
4374
4375 -----------------------------------------------
4376 $ git branch # list all local branches in this repo
4377 $ git checkout test # switch working directory to branch "test"
4378 $ git branch new # create branch "new" starting at current HEAD
4379 $ git branch -d new # delete branch "new"
4380 -----------------------------------------------
4381
4382 Instead of basing a new branch on current HEAD (the default), use:
4383
4384 -----------------------------------------------
4385 $ git branch new test # branch named "test"
4386 $ git branch new v2.6.15 # tag named v2.6.15
4387 $ git branch new HEAD^ # commit before the most recent
4388 $ git branch new HEAD^^ # commit before that
4389 $ git branch new test~10 # ten commits before tip of branch "test"
4390 -----------------------------------------------
4391
4392 Create and switch to a new branch at the same time:
4393
4394 -----------------------------------------------
4395 $ git checkout -b new v2.6.15
4396 -----------------------------------------------
4397
4398 Update and examine branches from the repository you cloned from:
4399
4400 -----------------------------------------------
4401 $ git fetch # update
4402 $ git branch -r # list
4403 origin/master
4404 origin/next
4405 ...
4406 $ git checkout -b masterwork origin/master
4407 -----------------------------------------------
4408
4409 Fetch a branch from a different repository, and give it a new
4410 name in your repository:
4411
4412 -----------------------------------------------
4413 $ git fetch git://example.com/project.git theirbranch:mybranch
4414 $ git fetch git://example.com/project.git v2.6.15:mybranch
4415 -----------------------------------------------
4416
4417 Keep a list of repositories you work with regularly:
4418
4419 -----------------------------------------------
4420 $ git remote add example git://example.com/project.git
4421 $ git remote # list remote repositories
4422 example
4423 origin
4424 $ git remote show example # get details
4425 * remote example
4426 URL: git://example.com/project.git
4427 Tracked remote branches
4428 master
4429 next
4430 ...
4431 $ git fetch example # update branches from example
4432 $ git branch -r # list all remote branches
4433 -----------------------------------------------
4434
4435
4436 [[exploring-history]]
4437 Exploring history
4438 -----------------
4439
4440 -----------------------------------------------
4441 $ gitk # visualize and browse history
4442 $ git log # list all commits
4443 $ git log src/ # ...modifying src/
4444 $ git log v2.6.15..v2.6.16 # ...in v2.6.16, not in v2.6.15
4445 $ git log master..test # ...in branch test, not in branch master
4446 $ git log test..master # ...in branch master, but not in test
4447 $ git log test...master # ...in one branch, not in both
4448 $ git log -S'foo()' # ...where difference contain "foo()"
4449 $ git log --since="2 weeks ago"
4450 $ git log -p # show patches as well
4451 $ git show # most recent commit
4452 $ git diff v2.6.15..v2.6.16 # diff between two tagged versions
4453 $ git diff v2.6.15..HEAD # diff with current head
4454 $ git grep "foo()" # search working directory for "foo()"
4455 $ git grep v2.6.15 "foo()" # search old tree for "foo()"
4456 $ git show v2.6.15:a.txt # look at old version of a.txt
4457 -----------------------------------------------
4458
4459 Search for regressions:
4460
4461 -----------------------------------------------
4462 $ git bisect start
4463 $ git bisect bad # current version is bad
4464 $ git bisect good v2.6.13-rc2 # last known good revision
4465 Bisecting: 675 revisions left to test after this
4466 # test here, then:
4467 $ git bisect good # if this revision is good, or
4468 $ git bisect bad # if this revision is bad.
4469 # repeat until done.
4470 -----------------------------------------------
4471
4472 [[making-changes]]
4473 Making changes
4474 --------------
4475
4476 Make sure git knows who to blame:
4477
4478 ------------------------------------------------
4479 $ cat >>~/.gitconfig <<\EOF
4480 [user]
4481 name = Your Name Comes Here
4482 email = you@yourdomain.example.com
4483 EOF
4484 ------------------------------------------------
4485
4486 Select file contents to include in the next commit, then make the
4487 commit:
4488
4489 -----------------------------------------------
4490 $ git add a.txt # updated file
4491 $ git add b.txt # new file
4492 $ git rm c.txt # old file
4493 $ git commit
4494 -----------------------------------------------
4495
4496 Or, prepare and create the commit in one step:
4497
4498 -----------------------------------------------
4499 $ git commit d.txt # use latest content only of d.txt
4500 $ git commit -a # use latest content of all tracked files
4501 -----------------------------------------------
4502
4503 [[merging]]
4504 Merging
4505 -------
4506
4507 -----------------------------------------------
4508 $ git merge test # merge branch "test" into the current branch
4509 $ git pull git://example.com/project.git master
4510 # fetch and merge in remote branch
4511 $ git pull . test # equivalent to git merge test
4512 -----------------------------------------------
4513
4514 [[sharing-your-changes]]
4515 Sharing your changes
4516 --------------------
4517
4518 Importing or exporting patches:
4519
4520 -----------------------------------------------
4521 $ git format-patch origin..HEAD # format a patch for each commit
4522 # in HEAD but not in origin
4523 $ git am mbox # import patches from the mailbox "mbox"
4524 -----------------------------------------------
4525
4526 Fetch a branch in a different git repository, then merge into the
4527 current branch:
4528
4529 -----------------------------------------------
4530 $ git pull git://example.com/project.git theirbranch
4531 -----------------------------------------------
4532
4533 Store the fetched branch into a local branch before merging into the
4534 current branch:
4535
4536 -----------------------------------------------
4537 $ git pull git://example.com/project.git theirbranch:mybranch
4538 -----------------------------------------------
4539
4540 After creating commits on a local branch, update the remote
4541 branch with your commits:
4542
4543 -----------------------------------------------
4544 $ git push ssh://example.com/project.git mybranch:theirbranch
4545 -----------------------------------------------
4546
4547 When remote and local branch are both named "test":
4548
4549 -----------------------------------------------
4550 $ git push ssh://example.com/project.git test
4551 -----------------------------------------------
4552
4553 Shortcut version for a frequently used remote repository:
4554
4555 -----------------------------------------------
4556 $ git remote add example ssh://example.com/project.git
4557 $ git push example test
4558 -----------------------------------------------
4559
4560 [[repository-maintenance]]
4561 Repository maintenance
4562 ----------------------
4563
4564 Check for corruption:
4565
4566 -----------------------------------------------
4567 $ git fsck
4568 -----------------------------------------------
4569
4570 Recompress, remove unused cruft:
4571
4572 -----------------------------------------------
4573 $ git gc
4574 -----------------------------------------------
4575
4576
4577 [[todo]]
4578 Appendix B: Notes and todo list for this manual
4579 ===============================================
4580
4581 This is a work in progress.
4582
4583 The basic requirements:
4584
4585 - It must be readable in order, from beginning to end, by someone
4586 intelligent with a basic grasp of the UNIX command line, but without
4587 any special knowledge of git. If necessary, any other prerequisites
4588 should be specifically mentioned as they arise.
4589 - Whenever possible, section headings should clearly describe the task
4590 they explain how to do, in language that requires no more knowledge
4591 than necessary: for example, "importing patches into a project" rather
4592 than "the `git am` command"
4593
4594 Think about how to create a clear chapter dependency graph that will
4595 allow people to get to important topics without necessarily reading
4596 everything in between.
4597
4598 Scan Documentation/ for other stuff left out; in particular:
4599
4600 - howto's
4601 - some of technical/?
4602 - hooks
4603 - list of commands in linkgit:git[1]
4604
4605 Scan email archives for other stuff left out
4606
4607 Scan man pages to see if any assume more background than this manual
4608 provides.
4609
4610 Simplify beginning by suggesting disconnected head instead of
4611 temporary branch creation?
4612
4613 Add more good examples. Entire sections of just cookbook examples
4614 might be a good idea; maybe make an "advanced examples" section a
4615 standard end-of-chapter section?
4616
4617 Include cross-references to the glossary, where appropriate.
4618
4619 Document shallow clones? See draft 1.5.0 release notes for some
4620 documentation.
4621
4622 Add a section on working with other version control systems, including
4623 CVS, Subversion, and just imports of series of release tarballs.
4624
4625 More details on gitweb?
4626
4627 Write a chapter on using plumbing and writing scripts.
4628
4629 Alternates, clone -reference, etc.
4630
4631 More on recovery from repository corruption. See:
4632 http://marc.theaimsgroup.com/?l=git&m=117263864820799&w=2
4633 http://marc.theaimsgroup.com/?l=git&m=117147855503798&w=2