Print the real filename that we failed to open.
[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. For a command such as "git clone", just use
22
23 ------------------------------------------------
24 $ man git-clone
25 ------------------------------------------------
26
27 See also <<git-quick-start>> for a brief overview of git commands,
28 without any explanation.
29
30 Finally, see <<todo>> for ways that you can help make this manual more
31 complete.
32
33
34 [[repositories-and-branches]]
35 Repositories and Branches
36 =========================
37
38 [[how-to-get-a-git-repository]]
39 How to get a git repository
40 ---------------------------
41
42 It will be useful to have a git repository to experiment with as you
43 read this manual.
44
45 The best way to get one is by using the gitlink:git-clone[1] command to
46 download a copy of an existing repository. If you don't already have a
47 project in mind, here are some interesting examples:
48
49 ------------------------------------------------
50 # git itself (approx. 10MB download):
51 $ git clone git://git.kernel.org/pub/scm/git/git.git
52 # the linux kernel (approx. 150MB download):
53 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
54 ------------------------------------------------
55
56 The initial clone may be time-consuming for a large project, but you
57 will only need to clone once.
58
59 The clone command creates a new directory named after the project
60 ("git" or "linux-2.6" in the examples above). After you cd into this
61 directory, you will see that it contains a copy of the project files,
62 together with a special top-level directory named ".git", which
63 contains all the information about the history of the project.
64
65 [[how-to-check-out]]
66 How to check out a different version of a project
67 -------------------------------------------------
68
69 Git is best thought of as a tool for storing the history of a collection
70 of files. It stores the history as a compressed collection of
71 interrelated snapshots of the project's contents. In git each such
72 version is called a <<def_commit,commit>>.
73
74 A single git repository may contain multiple branches. It keeps track
75 of them by keeping a list of <<def_head,heads>> which reference the
76 latest commit on each branch; the gitlink:git-branch[1] command shows
77 you the list of branch heads:
78
79 ------------------------------------------------
80 $ git branch
81 * master
82 ------------------------------------------------
83
84 A freshly cloned repository contains a single branch head, by default
85 named "master", with the working directory initialized to the state of
86 the project referred to by that branch head.
87
88 Most projects also use <<def_tag,tags>>. Tags, like heads, are
89 references into the project's history, and can be listed using the
90 gitlink:git-tag[1] command:
91
92 ------------------------------------------------
93 $ git tag -l
94 v2.6.11
95 v2.6.11-tree
96 v2.6.12
97 v2.6.12-rc2
98 v2.6.12-rc3
99 v2.6.12-rc4
100 v2.6.12-rc5
101 v2.6.12-rc6
102 v2.6.13
103 ...
104 ------------------------------------------------
105
106 Tags are expected to always point at the same version of a project,
107 while heads are expected to advance as development progresses.
108
109 Create a new branch head pointing to one of these versions and check it
110 out using gitlink:git-checkout[1]:
111
112 ------------------------------------------------
113 $ git checkout -b new v2.6.13
114 ------------------------------------------------
115
116 The working directory then reflects the contents that the project had
117 when it was tagged v2.6.13, and gitlink:git-branch[1] shows two
118 branches, with an asterisk marking the currently checked-out branch:
119
120 ------------------------------------------------
121 $ git branch
122 master
123 * new
124 ------------------------------------------------
125
126 If you decide that you'd rather see version 2.6.17, you can modify
127 the current branch to point at v2.6.17 instead, with
128
129 ------------------------------------------------
130 $ git reset --hard v2.6.17
131 ------------------------------------------------
132
133 Note that if the current branch head was your only reference to a
134 particular point in history, then resetting that branch may leave you
135 with no way to find the history it used to point to; so use this command
136 carefully.
137
138 [[understanding-commits]]
139 Understanding History: Commits
140 ------------------------------
141
142 Every change in the history of a project is represented by a commit.
143 The gitlink:git-show[1] command shows the most recent commit on the
144 current branch:
145
146 ------------------------------------------------
147 $ git show
148 commit 17cf781661e6d38f737f15f53ab552f1e95960d7
149 Author: Linus Torvalds <torvalds@ppc970.osdl.org.(none)>
150 Date: Tue Apr 19 14:11:06 2005 -0700
151
152 Remove duplicate getenv(DB_ENVIRONMENT) call
153
154 Noted by Tony Luck.
155
156 diff --git a/init-db.c b/init-db.c
157 index 65898fa..b002dc6 100644
158 --- a/init-db.c
159 +++ b/init-db.c
160 @@ -7,7 +7,7 @@
161
162 int main(int argc, char **argv)
163 {
164 - char *sha1_dir = getenv(DB_ENVIRONMENT), *path;
165 + char *sha1_dir, *path;
166 int len, i;
167
168 if (mkdir(".git", 0755) < 0) {
169 ------------------------------------------------
170
171 As you can see, a commit shows who made the latest change, what they
172 did, and why.
173
174 Every commit has a 40-hexdigit id, sometimes called the "object name" or the
175 "SHA1 id", shown on the first line of the "git show" output. You can usually
176 refer to a commit by a shorter name, such as a tag or a branch name, but this
177 longer name can also be useful. Most importantly, it is a globally unique
178 name for this commit: so if you tell somebody else the object name (for
179 example in email), then you are guaranteed that name will refer to the same
180 commit in their repository that it does in yours (assuming their repository
181 has that commit at all). Since the object name is computed as a hash over the
182 contents of the commit, you are guaranteed that the commit can never change
183 without its name also changing.
184
185 In fact, in <<git-concepts>> we shall see that everything stored in git
186 history, including file data and directory contents, is stored in an object
187 with a name that is a hash of its contents.
188
189 [[understanding-reachability]]
190 Understanding history: commits, parents, and reachability
191 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
192
193 Every commit (except the very first commit in a project) also has a
194 parent commit which shows what happened before this commit.
195 Following the chain of parents will eventually take you back to the
196 beginning of the project.
197
198 However, the commits do not form a simple list; git allows lines of
199 development to diverge and then reconverge, and the point where two
200 lines of development reconverge is called a "merge". The commit
201 representing a merge can therefore have more than one parent, with
202 each parent representing the most recent commit on one of the lines
203 of development leading to that point.
204
205 The best way to see how this works is using the gitlink:gitk[1]
206 command; running gitk now on a git repository and looking for merge
207 commits will help understand how the git organizes history.
208
209 In the following, we say that commit X is "reachable" from commit Y
210 if commit X is an ancestor of commit Y. Equivalently, you could say
211 that Y is a descendant of X, or that there is a chain of parents
212 leading from commit Y to commit X.
213
214 [[history-diagrams]]
215 Understanding history: History diagrams
216 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
217
218 We will sometimes represent git history using diagrams like the one
219 below. Commits are shown as "o", and the links between them with
220 lines drawn with - / and \. Time goes left to right:
221
222
223 ................................................
224 o--o--o <-- Branch A
225 /
226 o--o--o <-- master
227 \
228 o--o--o <-- Branch B
229 ................................................
230
231 If we need to talk about a particular commit, the character "o" may
232 be replaced with another letter or number.
233
234 [[what-is-a-branch]]
235 Understanding history: What is a branch?
236 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
237
238 When we need to be precise, we will use the word "branch" to mean a line
239 of development, and "branch head" (or just "head") to mean a reference
240 to the most recent commit on a branch. In the example above, the branch
241 head named "A" is a pointer to one particular commit, but we refer to
242 the line of three commits leading up to that point as all being part of
243 "branch A".
244
245 However, when no confusion will result, we often just use the term
246 "branch" both for branches and for branch heads.
247
248 [[manipulating-branches]]
249 Manipulating branches
250 ---------------------
251
252 Creating, deleting, and modifying branches is quick and easy; here's
253 a summary of the commands:
254
255 git branch::
256 list all branches
257 git branch <branch>::
258 create a new branch named <branch>, referencing the same
259 point in history as the current branch
260 git branch <branch> <start-point>::
261 create a new branch named <branch>, referencing
262 <start-point>, which may be specified any way you like,
263 including using a branch name or a tag name
264 git branch -d <branch>::
265 delete the branch <branch>; if the branch you are deleting
266 points to a commit which is not reachable from the current
267 branch, this command will fail with a warning.
268 git branch -D <branch>::
269 even if the branch points to a commit not reachable
270 from the current branch, you may know that that commit
271 is still reachable from some other branch or tag. In that
272 case it is safe to use this command to force git to delete
273 the branch.
274 git checkout <branch>::
275 make the current branch <branch>, updating the working
276 directory to reflect the version referenced by <branch>
277 git checkout -b <new> <start-point>::
278 create a new branch <new> referencing <start-point>, and
279 check it out.
280
281 The special symbol "HEAD" can always be used to refer to the current
282 branch. In fact, git uses a file named "HEAD" in the .git directory to
283 remember which branch is current:
284
285 ------------------------------------------------
286 $ cat .git/HEAD
287 ref: refs/heads/master
288 ------------------------------------------------
289
290 [[detached-head]]
291 Examining an old version without creating a new branch
292 ------------------------------------------------------
293
294 The git-checkout command normally expects a branch head, but will also
295 accept an arbitrary commit; for example, you can check out the commit
296 referenced by a tag:
297
298 ------------------------------------------------
299 $ git checkout v2.6.17
300 Note: moving to "v2.6.17" which isn't a local branch
301 If you want to create a new branch from this checkout, you may do so
302 (now or later) by using -b with the checkout command again. Example:
303 git checkout -b <new_branch_name>
304 HEAD is now at 427abfa... Linux v2.6.17
305 ------------------------------------------------
306
307 The HEAD then refers to the SHA1 of the commit instead of to a branch,
308 and git branch shows that you are no longer on a branch:
309
310 ------------------------------------------------
311 $ cat .git/HEAD
312 427abfa28afedffadfca9dd8b067eb6d36bac53f
313 $ git branch
314 * (no branch)
315 master
316 ------------------------------------------------
317
318 In this case we say that the HEAD is "detached".
319
320 This is an easy way to check out a particular version without having to
321 make up a name for the new branch. You can still create a new branch
322 (or tag) for this version later if you decide to.
323
324 [[examining-remote-branches]]
325 Examining branches from a remote repository
326 -------------------------------------------
327
328 The "master" branch that was created at the time you cloned is a copy
329 of the HEAD in the repository that you cloned from. That repository
330 may also have had other branches, though, and your local repository
331 keeps branches which track each of those remote branches, which you
332 can view using the "-r" option to gitlink:git-branch[1]:
333
334 ------------------------------------------------
335 $ git branch -r
336 origin/HEAD
337 origin/html
338 origin/maint
339 origin/man
340 origin/master
341 origin/next
342 origin/pu
343 origin/todo
344 ------------------------------------------------
345
346 You cannot check out these remote-tracking branches, but you can
347 examine them on a branch of your own, just as you would a tag:
348
349 ------------------------------------------------
350 $ git checkout -b my-todo-copy origin/todo
351 ------------------------------------------------
352
353 Note that the name "origin" is just the name that git uses by default
354 to refer to the repository that you cloned from.
355
356 [[how-git-stores-references]]
357 Naming branches, tags, and other references
358 -------------------------------------------
359
360 Branches, remote-tracking branches, and tags are all references to
361 commits. All references are named with a slash-separated path name
362 starting with "refs"; the names we've been using so far are actually
363 shorthand:
364
365 - The branch "test" is short for "refs/heads/test".
366 - The tag "v2.6.18" is short for "refs/tags/v2.6.18".
367 - "origin/master" is short for "refs/remotes/origin/master".
368
369 The full name is occasionally useful if, for example, there ever
370 exists a tag and a branch with the same name.
371
372 (Newly created refs are actually stored in the .git/refs directory,
373 under the path given by their name. However, for efficiency reasons
374 they may also be packed together in a single file; see
375 gitlink:git-pack-refs[1]).
376
377 As another useful shortcut, the "HEAD" of a repository can be referred
378 to just using the name of that repository. So, for example, "origin"
379 is usually a shortcut for the HEAD branch in the repository "origin".
380
381 For the complete list of paths which git checks for references, and
382 the order it uses to decide which to choose when there are multiple
383 references with the same shorthand name, see the "SPECIFYING
384 REVISIONS" section of gitlink:git-rev-parse[1].
385
386 [[Updating-a-repository-with-git-fetch]]
387 Updating a repository with git fetch
388 ------------------------------------
389
390 Eventually the developer cloned from will do additional work in her
391 repository, creating new commits and advancing the branches to point
392 at the new commits.
393
394 The command "git fetch", with no arguments, will update all of the
395 remote-tracking branches to the latest version found in her
396 repository. It will not touch any of your own branches--not even the
397 "master" branch that was created for you on clone.
398
399 [[fetching-branches]]
400 Fetching branches from other repositories
401 -----------------------------------------
402
403 You can also track branches from repositories other than the one you
404 cloned from, using gitlink:git-remote[1]:
405
406 -------------------------------------------------
407 $ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
408 $ git fetch linux-nfs
409 * refs/remotes/linux-nfs/master: storing branch 'master' ...
410 commit: bf81b46
411 -------------------------------------------------
412
413 New remote-tracking branches will be stored under the shorthand name
414 that you gave "git remote add", in this case linux-nfs:
415
416 -------------------------------------------------
417 $ git branch -r
418 linux-nfs/master
419 origin/master
420 -------------------------------------------------
421
422 If you run "git fetch <remote>" later, the tracking branches for the
423 named <remote> will be updated.
424
425 If you examine the file .git/config, you will see that git has added
426 a new stanza:
427
428 -------------------------------------------------
429 $ cat .git/config
430 ...
431 [remote "linux-nfs"]
432 url = git://linux-nfs.org/pub/nfs-2.6.git
433 fetch = +refs/heads/*:refs/remotes/linux-nfs/*
434 ...
435 -------------------------------------------------
436
437 This is what causes git to track the remote's branches; you may modify
438 or delete these configuration options by editing .git/config with a
439 text editor. (See the "CONFIGURATION FILE" section of
440 gitlink:git-config[1] for details.)
441
442 [[exploring-git-history]]
443 Exploring git history
444 =====================
445
446 Git is best thought of as a tool for storing the history of a
447 collection of files. It does this by storing compressed snapshots of
448 the contents of a file hierarchy, together with "commits" which show
449 the relationships between these snapshots.
450
451 Git provides extremely flexible and fast tools for exploring the
452 history of a project.
453
454 We start with one specialized tool that is useful for finding the
455 commit that introduced a bug into a project.
456
457 [[using-bisect]]
458 How to use bisect to find a regression
459 --------------------------------------
460
461 Suppose version 2.6.18 of your project worked, but the version at
462 "master" crashes. Sometimes the best way to find the cause of such a
463 regression is to perform a brute-force search through the project's
464 history to find the particular commit that caused the problem. The
465 gitlink:git-bisect[1] command can help you do this:
466
467 -------------------------------------------------
468 $ git bisect start
469 $ git bisect good v2.6.18
470 $ git bisect bad master
471 Bisecting: 3537 revisions left to test after this
472 [65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
473 -------------------------------------------------
474
475 If you run "git branch" at this point, you'll see that git has
476 temporarily moved you to a new branch named "bisect". This branch
477 points to a commit (with commit id 65934...) that is reachable from
478 "master" but not from v2.6.18. Compile and test it, and see whether
479 it crashes. Assume it does crash. Then:
480
481 -------------------------------------------------
482 $ git bisect bad
483 Bisecting: 1769 revisions left to test after this
484 [7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
485 -------------------------------------------------
486
487 checks out an older version. Continue like this, telling git at each
488 stage whether the version it gives you is good or bad, and notice
489 that the number of revisions left to test is cut approximately in
490 half each time.
491
492 After about 13 tests (in this case), it will output the commit id of
493 the guilty commit. You can then examine the commit with
494 gitlink:git-show[1], find out who wrote it, and mail them your bug
495 report with the commit id. Finally, run
496
497 -------------------------------------------------
498 $ git bisect reset
499 -------------------------------------------------
500
501 to return you to the branch you were on before and delete the
502 temporary "bisect" branch.
503
504 Note that the version which git-bisect checks out for you at each
505 point is just a suggestion, and you're free to try a different
506 version if you think it would be a good idea. For example,
507 occasionally you may land on a commit that broke something unrelated;
508 run
509
510 -------------------------------------------------
511 $ git bisect visualize
512 -------------------------------------------------
513
514 which will run gitk and label the commit it chose with a marker that
515 says "bisect". Chose a safe-looking commit nearby, note its commit
516 id, and check it out with:
517
518 -------------------------------------------------
519 $ git reset --hard fb47ddb2db...
520 -------------------------------------------------
521
522 then test, run "bisect good" or "bisect bad" as appropriate, and
523 continue.
524
525 [[naming-commits]]
526 Naming commits
527 --------------
528
529 We have seen several ways of naming commits already:
530
531 - 40-hexdigit object name
532 - branch name: refers to the commit at the head of the given
533 branch
534 - tag name: refers to the commit pointed to by the given tag
535 (we've seen branches and tags are special cases of
536 <<how-git-stores-references,references>>).
537 - HEAD: refers to the head of the current branch
538
539 There are many more; see the "SPECIFYING REVISIONS" section of the
540 gitlink:git-rev-parse[1] man page for the complete list of ways to
541 name revisions. Some examples:
542
543 -------------------------------------------------
544 $ git show fb47ddb2 # the first few characters of the object name
545 # are usually enough to specify it uniquely
546 $ git show HEAD^ # the parent of the HEAD commit
547 $ git show HEAD^^ # the grandparent
548 $ git show HEAD~4 # the great-great-grandparent
549 -------------------------------------------------
550
551 Recall that merge commits may have more than one parent; by default,
552 ^ and ~ follow the first parent listed in the commit, but you can
553 also choose:
554
555 -------------------------------------------------
556 $ git show HEAD^1 # show the first parent of HEAD
557 $ git show HEAD^2 # show the second parent of HEAD
558 -------------------------------------------------
559
560 In addition to HEAD, there are several other special names for
561 commits:
562
563 Merges (to be discussed later), as well as operations such as
564 git-reset, which change the currently checked-out commit, generally
565 set ORIG_HEAD to the value HEAD had before the current operation.
566
567 The git-fetch operation always stores the head of the last fetched
568 branch in FETCH_HEAD. For example, if you run git fetch without
569 specifying a local branch as the target of the operation
570
571 -------------------------------------------------
572 $ git fetch git://example.com/proj.git theirbranch
573 -------------------------------------------------
574
575 the fetched commits will still be available from FETCH_HEAD.
576
577 When we discuss merges we'll also see the special name MERGE_HEAD,
578 which refers to the other branch that we're merging in to the current
579 branch.
580
581 The gitlink:git-rev-parse[1] command is a low-level command that is
582 occasionally useful for translating some name for a commit to the object
583 name for that commit:
584
585 -------------------------------------------------
586 $ git rev-parse origin
587 e05db0fd4f31dde7005f075a84f96b360d05984b
588 -------------------------------------------------
589
590 [[creating-tags]]
591 Creating tags
592 -------------
593
594 We can also create a tag to refer to a particular commit; after
595 running
596
597 -------------------------------------------------
598 $ git tag stable-1 1b2e1d63ff
599 -------------------------------------------------
600
601 You can use stable-1 to refer to the commit 1b2e1d63ff.
602
603 This creates a "lightweight" tag. If you would also like to include a
604 comment with the tag, and possibly sign it cryptographically, then you
605 should create a tag object instead; see the gitlink:git-tag[1] man page
606 for details.
607
608 [[browsing-revisions]]
609 Browsing revisions
610 ------------------
611
612 The gitlink:git-log[1] command can show lists of commits. On its
613 own, it shows all commits reachable from the parent commit; but you
614 can also make more specific requests:
615
616 -------------------------------------------------
617 $ git log v2.5.. # commits since (not reachable from) v2.5
618 $ git log test..master # commits reachable from master but not test
619 $ git log master..test # ...reachable from test but not master
620 $ git log master...test # ...reachable from either test or master,
621 # but not both
622 $ git log --since="2 weeks ago" # commits from the last 2 weeks
623 $ git log Makefile # commits which modify Makefile
624 $ git log fs/ # ... which modify any file under fs/
625 $ git log -S'foo()' # commits which add or remove any file data
626 # matching the string 'foo()'
627 -------------------------------------------------
628
629 And of course you can combine all of these; the following finds
630 commits since v2.5 which touch the Makefile or any file under fs:
631
632 -------------------------------------------------
633 $ git log v2.5.. Makefile fs/
634 -------------------------------------------------
635
636 You can also ask git log to show patches:
637
638 -------------------------------------------------
639 $ git log -p
640 -------------------------------------------------
641
642 See the "--pretty" option in the gitlink:git-log[1] man page for more
643 display options.
644
645 Note that git log starts with the most recent commit and works
646 backwards through the parents; however, since git history can contain
647 multiple independent lines of development, the particular order that
648 commits are listed in may be somewhat arbitrary.
649
650 [[generating-diffs]]
651 Generating diffs
652 ----------------
653
654 You can generate diffs between any two versions using
655 gitlink:git-diff[1]:
656
657 -------------------------------------------------
658 $ git diff master..test
659 -------------------------------------------------
660
661 That will produce the diff between the tips of the two branches. If
662 you'd prefer to find the diff from their common ancestor to test, you
663 can use three dots instead of two:
664
665 -------------------------------------------------
666 $ git diff master...test
667 -------------------------------------------------
668
669 Sometimes what you want instead is a set of patches; for this you can
670 use gitlink:git-format-patch[1]:
671
672 -------------------------------------------------
673 $ git format-patch master..test
674 -------------------------------------------------
675
676 will generate a file with a patch for each commit reachable from test
677 but not from master.
678
679 [[viewing-old-file-versions]]
680 Viewing old file versions
681 -------------------------
682
683 You can always view an old version of a file by just checking out the
684 correct revision first. But sometimes it is more convenient to be
685 able to view an old version of a single file without checking
686 anything out; this command does that:
687
688 -------------------------------------------------
689 $ git show v2.5:fs/locks.c
690 -------------------------------------------------
691
692 Before the colon may be anything that names a commit, and after it
693 may be any path to a file tracked by git.
694
695 [[history-examples]]
696 Examples
697 --------
698
699 [[counting-commits-on-a-branch]]
700 Counting the number of commits on a branch
701 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
702
703 Suppose you want to know how many commits you've made on "mybranch"
704 since it diverged from "origin":
705
706 -------------------------------------------------
707 $ git log --pretty=oneline origin..mybranch | wc -l
708 -------------------------------------------------
709
710 Alternatively, you may often see this sort of thing done with the
711 lower-level command gitlink:git-rev-list[1], which just lists the SHA1's
712 of all the given commits:
713
714 -------------------------------------------------
715 $ git rev-list origin..mybranch | wc -l
716 -------------------------------------------------
717
718 [[checking-for-equal-branches]]
719 Check whether two branches point at the same history
720 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
721
722 Suppose you want to check whether two branches point at the same point
723 in history.
724
725 -------------------------------------------------
726 $ git diff origin..master
727 -------------------------------------------------
728
729 will tell you whether the contents of the project are the same at the
730 two branches; in theory, however, it's possible that the same project
731 contents could have been arrived at by two different historical
732 routes. You could compare the object names:
733
734 -------------------------------------------------
735 $ git rev-list origin
736 e05db0fd4f31dde7005f075a84f96b360d05984b
737 $ git rev-list master
738 e05db0fd4f31dde7005f075a84f96b360d05984b
739 -------------------------------------------------
740
741 Or you could recall that the ... operator selects all commits
742 contained reachable from either one reference or the other but not
743 both: so
744
745 -------------------------------------------------
746 $ git log origin...master
747 -------------------------------------------------
748
749 will return no commits when the two branches are equal.
750
751 [[finding-tagged-descendants]]
752 Find first tagged version including a given fix
753 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
754
755 Suppose you know that the commit e05db0fd fixed a certain problem.
756 You'd like to find the earliest tagged release that contains that
757 fix.
758
759 Of course, there may be more than one answer--if the history branched
760 after commit e05db0fd, then there could be multiple "earliest" tagged
761 releases.
762
763 You could just visually inspect the commits since e05db0fd:
764
765 -------------------------------------------------
766 $ gitk e05db0fd..
767 -------------------------------------------------
768
769 Or you can use gitlink:git-name-rev[1], which will give the commit a
770 name based on any tag it finds pointing to one of the commit's
771 descendants:
772
773 -------------------------------------------------
774 $ git name-rev --tags e05db0fd
775 e05db0fd tags/v1.5.0-rc1^0~23
776 -------------------------------------------------
777
778 The gitlink:git-describe[1] command does the opposite, naming the
779 revision using a tag on which the given commit is based:
780
781 -------------------------------------------------
782 $ git describe e05db0fd
783 v1.5.0-rc0-260-ge05db0f
784 -------------------------------------------------
785
786 but that may sometimes help you guess which tags might come after the
787 given commit.
788
789 If you just want to verify whether a given tagged version contains a
790 given commit, you could use gitlink:git-merge-base[1]:
791
792 -------------------------------------------------
793 $ git merge-base e05db0fd v1.5.0-rc1
794 e05db0fd4f31dde7005f075a84f96b360d05984b
795 -------------------------------------------------
796
797 The merge-base command finds a common ancestor of the given commits,
798 and always returns one or the other in the case where one is a
799 descendant of the other; so the above output shows that e05db0fd
800 actually is an ancestor of v1.5.0-rc1.
801
802 Alternatively, note that
803
804 -------------------------------------------------
805 $ git log v1.5.0-rc1..e05db0fd
806 -------------------------------------------------
807
808 will produce empty output if and only if v1.5.0-rc1 includes e05db0fd,
809 because it outputs only commits that are not reachable from v1.5.0-rc1.
810
811 As yet another alternative, the gitlink:git-show-branch[1] command lists
812 the commits reachable from its arguments with a display on the left-hand
813 side that indicates which arguments that commit is reachable from. So,
814 you can run something like
815
816 -------------------------------------------------
817 $ git show-branch e05db0fd v1.5.0-rc0 v1.5.0-rc1 v1.5.0-rc2
818 ! [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
819 available
820 ! [v1.5.0-rc0] GIT v1.5.0 preview
821 ! [v1.5.0-rc1] GIT v1.5.0-rc1
822 ! [v1.5.0-rc2] GIT v1.5.0-rc2
823 ...
824 -------------------------------------------------
825
826 then search for a line that looks like
827
828 -------------------------------------------------
829 + ++ [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
830 available
831 -------------------------------------------------
832
833 Which shows that e05db0fd is reachable from itself, from v1.5.0-rc1, and
834 from v1.5.0-rc2, but not from v1.5.0-rc0.
835
836 [[showing-commits-unique-to-a-branch]]
837 Showing commits unique to a given branch
838 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
839
840 Suppose you would like to see all the commits reachable from the branch
841 head named "master" but not from any other head in your repository.
842
843 We can list all the heads in this repository with
844 gitlink:git-show-ref[1]:
845
846 -------------------------------------------------
847 $ git show-ref --heads
848 bf62196b5e363d73353a9dcf094c59595f3153b7 refs/heads/core-tutorial
849 db768d5504c1bb46f63ee9d6e1772bd047e05bf9 refs/heads/maint
850 a07157ac624b2524a059a3414e99f6f44bebc1e7 refs/heads/master
851 24dbc180ea14dc1aebe09f14c8ecf32010690627 refs/heads/tutorial-2
852 1e87486ae06626c2f31eaa63d26fc0fd646c8af2 refs/heads/tutorial-fixes
853 -------------------------------------------------
854
855 We can get just the branch-head names, and remove "master", with
856 the help of the standard utilities cut and grep:
857
858 -------------------------------------------------
859 $ git show-ref --heads | cut -d' ' -f2 | grep -v '^refs/heads/master'
860 refs/heads/core-tutorial
861 refs/heads/maint
862 refs/heads/tutorial-2
863 refs/heads/tutorial-fixes
864 -------------------------------------------------
865
866 And then we can ask to see all the commits reachable from master
867 but not from these other heads:
868
869 -------------------------------------------------
870 $ gitk master --not $( git show-ref --heads | cut -d' ' -f2 |
871 grep -v '^refs/heads/master' )
872 -------------------------------------------------
873
874 Obviously, endless variations are possible; for example, to see all
875 commits reachable from some head but not from any tag in the repository:
876
877 -------------------------------------------------
878 $ gitk $( git show-ref --heads ) --not $( git show-ref --tags )
879 -------------------------------------------------
880
881 (See gitlink:git-rev-parse[1] for explanations of commit-selecting
882 syntax such as `--not`.)
883
884 [[making-a-release]]
885 Creating a changelog and tarball for a software release
886 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
887
888 The gitlink:git-archive[1] command can create a tar or zip archive from
889 any version of a project; for example:
890
891 -------------------------------------------------
892 $ git archive --format=tar --prefix=project/ HEAD | gzip >latest.tar.gz
893 -------------------------------------------------
894
895 will use HEAD to produce a tar archive in which each filename is
896 preceded by "project/".
897
898 If you're releasing a new version of a software project, you may want
899 to simultaneously make a changelog to include in the release
900 announcement.
901
902 Linus Torvalds, for example, makes new kernel releases by tagging them,
903 then running:
904
905 -------------------------------------------------
906 $ release-script 2.6.12 2.6.13-rc6 2.6.13-rc7
907 -------------------------------------------------
908
909 where release-script is a shell script that looks like:
910
911 -------------------------------------------------
912 #!/bin/sh
913 stable="$1"
914 last="$2"
915 new="$3"
916 echo "# git tag v$new"
917 echo "git archive --prefix=linux-$new/ v$new | gzip -9 > ../linux-$new.tar.gz"
918 echo "git diff v$stable v$new | gzip -9 > ../patch-$new.gz"
919 echo "git log --no-merges v$new ^v$last > ../ChangeLog-$new"
920 echo "git shortlog --no-merges v$new ^v$last > ../ShortLog"
921 echo "git diff --stat --summary -M v$last v$new > ../diffstat-$new"
922 -------------------------------------------------
923
924 and then he just cut-and-pastes the output commands after verifying that
925 they look OK.
926
927 [[Finding-comments-with-given-content]]
928 Finding commits referencing a file with given content
929 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
930
931 Somebody hands you a copy of a file, and asks which commits modified a
932 file such that it contained the given content either before or after the
933 commit. You can find out with this:
934
935 -------------------------------------------------
936 $ git log --raw --abbrev=40 --pretty=oneline |
937 grep -B 1 `git hash-object filename`
938 -------------------------------------------------
939
940 Figuring out why this works is left as an exercise to the (advanced)
941 student. The gitlink:git-log[1], gitlink:git-diff-tree[1], and
942 gitlink:git-hash-object[1] man pages may prove helpful.
943
944 [[Developing-with-git]]
945 Developing with git
946 ===================
947
948 [[telling-git-your-name]]
949 Telling git your name
950 ---------------------
951
952 Before creating any commits, you should introduce yourself to git. The
953 easiest way to do so is to make sure the following lines appear in a
954 file named .gitconfig in your home directory:
955
956 ------------------------------------------------
957 [user]
958 name = Your Name Comes Here
959 email = you@yourdomain.example.com
960 ------------------------------------------------
961
962 (See the "CONFIGURATION FILE" section of gitlink:git-config[1] for
963 details on the configuration file.)
964
965
966 [[creating-a-new-repository]]
967 Creating a new repository
968 -------------------------
969
970 Creating a new repository from scratch is very easy:
971
972 -------------------------------------------------
973 $ mkdir project
974 $ cd project
975 $ git init
976 -------------------------------------------------
977
978 If you have some initial content (say, a tarball):
979
980 -------------------------------------------------
981 $ tar -xzvf project.tar.gz
982 $ cd project
983 $ git init
984 $ git add . # include everything below ./ in the first commit:
985 $ git commit
986 -------------------------------------------------
987
988 [[how-to-make-a-commit]]
989 How to make a commit
990 --------------------
991
992 Creating a new commit takes three steps:
993
994 1. Making some changes to the working directory using your
995 favorite editor.
996 2. Telling git about your changes.
997 3. Creating the commit using the content you told git about
998 in step 2.
999
1000 In practice, you can interleave and repeat steps 1 and 2 as many
1001 times as you want: in order to keep track of what you want committed
1002 at step 3, git maintains a snapshot of the tree's contents in a
1003 special staging area called "the index."
1004
1005 At the beginning, the content of the index will be identical to
1006 that of the HEAD. The command "git diff --cached", which shows
1007 the difference between the HEAD and the index, should therefore
1008 produce no output at that point.
1009
1010 Modifying the index is easy:
1011
1012 To update the index with the new contents of a modified file, use
1013
1014 -------------------------------------------------
1015 $ git add path/to/file
1016 -------------------------------------------------
1017
1018 To add the contents of a new file to the index, use
1019
1020 -------------------------------------------------
1021 $ git add path/to/file
1022 -------------------------------------------------
1023
1024 To remove a file from the index and from the working tree,
1025
1026 -------------------------------------------------
1027 $ git rm path/to/file
1028 -------------------------------------------------
1029
1030 After each step you can verify that
1031
1032 -------------------------------------------------
1033 $ git diff --cached
1034 -------------------------------------------------
1035
1036 always shows the difference between the HEAD and the index file--this
1037 is what you'd commit if you created the commit now--and that
1038
1039 -------------------------------------------------
1040 $ git diff
1041 -------------------------------------------------
1042
1043 shows the difference between the working tree and the index file.
1044
1045 Note that "git add" always adds just the current contents of a file
1046 to the index; further changes to the same file will be ignored unless
1047 you run git-add on the file again.
1048
1049 When you're ready, just run
1050
1051 -------------------------------------------------
1052 $ git commit
1053 -------------------------------------------------
1054
1055 and git will prompt you for a commit message and then create the new
1056 commit. Check to make sure it looks like what you expected with
1057
1058 -------------------------------------------------
1059 $ git show
1060 -------------------------------------------------
1061
1062 As a special shortcut,
1063
1064 -------------------------------------------------
1065 $ git commit -a
1066 -------------------------------------------------
1067
1068 will update the index with any files that you've modified or removed
1069 and create a commit, all in one step.
1070
1071 A number of commands are useful for keeping track of what you're
1072 about to commit:
1073
1074 -------------------------------------------------
1075 $ git diff --cached # difference between HEAD and the index; what
1076 # would be committed if you ran "commit" now.
1077 $ git diff # difference between the index file and your
1078 # working directory; changes that would not
1079 # be included if you ran "commit" now.
1080 $ git diff HEAD # difference between HEAD and working tree; what
1081 # would be committed if you ran "commit -a" now.
1082 $ git status # a brief per-file summary of the above.
1083 -------------------------------------------------
1084
1085 You can also use gitlink:git-gui[1] to create commits, view changes in
1086 the index and the working tree files, and individually select diff hunks
1087 for inclusion in the index (by right-clicking on the diff hunk and
1088 choosing "Stage Hunk For Commit").
1089
1090 [[creating-good-commit-messages]]
1091 Creating good commit messages
1092 -----------------------------
1093
1094 Though not required, it's a good idea to begin the commit message
1095 with a single short (less than 50 character) line summarizing the
1096 change, followed by a blank line and then a more thorough
1097 description. Tools that turn commits into email, for example, use
1098 the first line on the Subject line and the rest of the commit in the
1099 body.
1100
1101 [[ignoring-files]]
1102 Ignoring files
1103 --------------
1104
1105 A project will often generate files that you do 'not' want to track with git.
1106 This typically includes files generated by a build process or temporary
1107 backup files made by your editor. Of course, 'not' tracking files with git
1108 is just a matter of 'not' calling "`git add`" on them. But it quickly becomes
1109 annoying to have these untracked files lying around; e.g. they make
1110 "`git add .`" and "`git commit -a`" practically useless, and they keep
1111 showing up in the output of "`git status`".
1112
1113 You can tell git to ignore certain files by creating a file called .gitignore
1114 in the top level of your working directory, with contents such as:
1115
1116 -------------------------------------------------
1117 # Lines starting with '#' are considered comments.
1118 # Ignore any file named foo.txt.
1119 foo.txt
1120 # Ignore (generated) html files,
1121 *.html
1122 # except foo.html which is maintained by hand.
1123 !foo.html
1124 # Ignore objects and archives.
1125 *.[oa]
1126 -------------------------------------------------
1127
1128 See gitlink:gitignore[5] for a detailed explanation of the syntax. You can
1129 also place .gitignore files in other directories in your working tree, and they
1130 will apply to those directories and their subdirectories. The `.gitignore`
1131 files can be added to your repository like any other files (just run `git add
1132 .gitignore` and `git commit`, as usual), which is convenient when the exclude
1133 patterns (such as patterns matching build output files) would also make sense
1134 for other users who clone your repository.
1135
1136 If you wish the exclude patterns to affect only certain repositories
1137 (instead of every repository for a given project), you may instead put
1138 them in a file in your repository named .git/info/exclude, or in any file
1139 specified by the `core.excludesfile` configuration variable. Some git
1140 commands can also take exclude patterns directly on the command line.
1141 See gitlink:gitignore[5] for the details.
1142
1143 [[how-to-merge]]
1144 How to merge
1145 ------------
1146
1147 You can rejoin two diverging branches of development using
1148 gitlink:git-merge[1]:
1149
1150 -------------------------------------------------
1151 $ git merge branchname
1152 -------------------------------------------------
1153
1154 merges the development in the branch "branchname" into the current
1155 branch. If there are conflicts--for example, if the same file is
1156 modified in two different ways in the remote branch and the local
1157 branch--then you are warned; the output may look something like this:
1158
1159 -------------------------------------------------
1160 $ git merge next
1161 100% (4/4) done
1162 Auto-merged file.txt
1163 CONFLICT (content): Merge conflict in file.txt
1164 Automatic merge failed; fix conflicts and then commit the result.
1165 -------------------------------------------------
1166
1167 Conflict markers are left in the problematic files, and after
1168 you resolve the conflicts manually, you can update the index
1169 with the contents and run git commit, as you normally would when
1170 creating a new file.
1171
1172 If you examine the resulting commit using gitk, you will see that it
1173 has two parents, one pointing to the top of the current branch, and
1174 one to the top of the other branch.
1175
1176 [[resolving-a-merge]]
1177 Resolving a merge
1178 -----------------
1179
1180 When a merge isn't resolved automatically, git leaves the index and
1181 the working tree in a special state that gives you all the
1182 information you need to help resolve the merge.
1183
1184 Files with conflicts are marked specially in the index, so until you
1185 resolve the problem and update the index, gitlink:git-commit[1] will
1186 fail:
1187
1188 -------------------------------------------------
1189 $ git commit
1190 file.txt: needs merge
1191 -------------------------------------------------
1192
1193 Also, gitlink:git-status[1] will list those files as "unmerged", and the
1194 files with conflicts will have conflict markers added, like this:
1195
1196 -------------------------------------------------
1197 <<<<<<< HEAD:file.txt
1198 Hello world
1199 =======
1200 Goodbye
1201 >>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1202 -------------------------------------------------
1203
1204 All you need to do is edit the files to resolve the conflicts, and then
1205
1206 -------------------------------------------------
1207 $ git add file.txt
1208 $ git commit
1209 -------------------------------------------------
1210
1211 Note that the commit message will already be filled in for you with
1212 some information about the merge. Normally you can just use this
1213 default message unchanged, but you may add additional commentary of
1214 your own if desired.
1215
1216 The above is all you need to know to resolve a simple merge. But git
1217 also provides more information to help resolve conflicts:
1218
1219 [[conflict-resolution]]
1220 Getting conflict-resolution help during a merge
1221 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1222
1223 All of the changes that git was able to merge automatically are
1224 already added to the index file, so gitlink:git-diff[1] shows only
1225 the conflicts. It uses an unusual syntax:
1226
1227 -------------------------------------------------
1228 $ git diff
1229 diff --cc file.txt
1230 index 802992c,2b60207..0000000
1231 --- a/file.txt
1232 +++ b/file.txt
1233 @@@ -1,1 -1,1 +1,5 @@@
1234 ++<<<<<<< HEAD:file.txt
1235 +Hello world
1236 ++=======
1237 + Goodbye
1238 ++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1239 -------------------------------------------------
1240
1241 Recall that the commit which will be committed after we resolve this
1242 conflict will have two parents instead of the usual one: one parent
1243 will be HEAD, the tip of the current branch; the other will be the
1244 tip of the other branch, which is stored temporarily in MERGE_HEAD.
1245
1246 During the merge, the index holds three versions of each file. Each of
1247 these three "file stages" represents a different version of the file:
1248
1249 -------------------------------------------------
1250 $ git show :1:file.txt # the file in a common ancestor of both branches
1251 $ git show :2:file.txt # the version from HEAD, but including any
1252 # nonconflicting changes from MERGE_HEAD
1253 $ git show :3:file.txt # the version from MERGE_HEAD, but including any
1254 # nonconflicting changes from HEAD.
1255 -------------------------------------------------
1256
1257 Since the stage 2 and stage 3 versions have already been updated with
1258 nonconflicting changes, the only remaining differences between them are
1259 the important ones; thus gitlink:git-diff[1] can use the information in
1260 the index to show only those conflicts.
1261
1262 The diff above shows the differences between the working-tree version of
1263 file.txt and the stage 2 and stage 3 versions. So instead of preceding
1264 each line by a single "+" or "-", it now uses two columns: the first
1265 column is used for differences between the first parent and the working
1266 directory copy, and the second for differences between the second parent
1267 and the working directory copy. (See the "COMBINED DIFF FORMAT" section
1268 of gitlink:git-diff-files[1] for a details of the format.)
1269
1270 After resolving the conflict in the obvious way (but before updating the
1271 index), the diff will look like:
1272
1273 -------------------------------------------------
1274 $ git diff
1275 diff --cc file.txt
1276 index 802992c,2b60207..0000000
1277 --- a/file.txt
1278 +++ b/file.txt
1279 @@@ -1,1 -1,1 +1,1 @@@
1280 - Hello world
1281 -Goodbye
1282 ++Goodbye world
1283 -------------------------------------------------
1284
1285 This shows that our resolved version deleted "Hello world" from the
1286 first parent, deleted "Goodbye" from the second parent, and added
1287 "Goodbye world", which was previously absent from both.
1288
1289 Some special diff options allow diffing the working directory against
1290 any of these stages:
1291
1292 -------------------------------------------------
1293 $ git diff -1 file.txt # diff against stage 1
1294 $ git diff --base file.txt # same as the above
1295 $ git diff -2 file.txt # diff against stage 2
1296 $ git diff --ours file.txt # same as the above
1297 $ git diff -3 file.txt # diff against stage 3
1298 $ git diff --theirs file.txt # same as the above.
1299 -------------------------------------------------
1300
1301 The gitlink:git-log[1] and gitk[1] commands also provide special help
1302 for merges:
1303
1304 -------------------------------------------------
1305 $ git log --merge
1306 $ gitk --merge
1307 -------------------------------------------------
1308
1309 These will display all commits which exist only on HEAD or on
1310 MERGE_HEAD, and which touch an unmerged file.
1311
1312 You may also use gitlink:git-mergetool[1], which lets you merge the
1313 unmerged files using external tools such as emacs or kdiff3.
1314
1315 Each time you resolve the conflicts in a file and update the index:
1316
1317 -------------------------------------------------
1318 $ git add file.txt
1319 -------------------------------------------------
1320
1321 the different stages of that file will be "collapsed", after which
1322 git-diff will (by default) no longer show diffs for that file.
1323
1324 [[undoing-a-merge]]
1325 Undoing a merge
1326 ---------------
1327
1328 If you get stuck and decide to just give up and throw the whole mess
1329 away, you can always return to the pre-merge state with
1330
1331 -------------------------------------------------
1332 $ git reset --hard HEAD
1333 -------------------------------------------------
1334
1335 Or, if you've already committed the merge that you want to throw away,
1336
1337 -------------------------------------------------
1338 $ git reset --hard ORIG_HEAD
1339 -------------------------------------------------
1340
1341 However, this last command can be dangerous in some cases--never
1342 throw away a commit you have already committed if that commit may
1343 itself have been merged into another branch, as doing so may confuse
1344 further merges.
1345
1346 [[fast-forwards]]
1347 Fast-forward merges
1348 -------------------
1349
1350 There is one special case not mentioned above, which is treated
1351 differently. Normally, a merge results in a merge commit, with two
1352 parents, one pointing at each of the two lines of development that
1353 were merged.
1354
1355 However, if the current branch is a descendant of the other--so every
1356 commit present in the one is already contained in the other--then git
1357 just performs a "fast forward"; the head of the current branch is moved
1358 forward to point at the head of the merged-in branch, without any new
1359 commits being created.
1360
1361 [[fixing-mistakes]]
1362 Fixing mistakes
1363 ---------------
1364
1365 If you've messed up the working tree, but haven't yet committed your
1366 mistake, you can return the entire working tree to the last committed
1367 state with
1368
1369 -------------------------------------------------
1370 $ git reset --hard HEAD
1371 -------------------------------------------------
1372
1373 If you make a commit that you later wish you hadn't, there are two
1374 fundamentally different ways to fix the problem:
1375
1376 1. You can create a new commit that undoes whatever was done
1377 by the old commit. This is the correct thing if your
1378 mistake has already been made public.
1379
1380 2. You can go back and modify the old commit. You should
1381 never do this if you have already made the history public;
1382 git does not normally expect the "history" of a project to
1383 change, and cannot correctly perform repeated merges from
1384 a branch that has had its history changed.
1385
1386 [[reverting-a-commit]]
1387 Fixing a mistake with a new commit
1388 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1389
1390 Creating a new commit that reverts an earlier change is very easy;
1391 just pass the gitlink:git-revert[1] command a reference to the bad
1392 commit; for example, to revert the most recent commit:
1393
1394 -------------------------------------------------
1395 $ git revert HEAD
1396 -------------------------------------------------
1397
1398 This will create a new commit which undoes the change in HEAD. You
1399 will be given a chance to edit the commit message for the new commit.
1400
1401 You can also revert an earlier change, for example, the next-to-last:
1402
1403 -------------------------------------------------
1404 $ git revert HEAD^
1405 -------------------------------------------------
1406
1407 In this case git will attempt to undo the old change while leaving
1408 intact any changes made since then. If more recent changes overlap
1409 with the changes to be reverted, then you will be asked to fix
1410 conflicts manually, just as in the case of <<resolving-a-merge,
1411 resolving a merge>>.
1412
1413 [[fixing-a-mistake-by-editing-history]]
1414 Fixing a mistake by editing history
1415 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1416
1417 If the problematic commit is the most recent commit, and you have not
1418 yet made that commit public, then you may just
1419 <<undoing-a-merge,destroy it using git-reset>>.
1420
1421 Alternatively, you
1422 can edit the working directory and update the index to fix your
1423 mistake, just as if you were going to <<how-to-make-a-commit,create a
1424 new commit>>, then run
1425
1426 -------------------------------------------------
1427 $ git commit --amend
1428 -------------------------------------------------
1429
1430 which will replace the old commit by a new commit incorporating your
1431 changes, giving you a chance to edit the old commit message first.
1432
1433 Again, you should never do this to a commit that may already have
1434 been merged into another branch; use gitlink:git-revert[1] instead in
1435 that case.
1436
1437 It is also possible to edit commits further back in the history, but
1438 this is an advanced topic to be left for
1439 <<cleaning-up-history,another chapter>>.
1440
1441 [[checkout-of-path]]
1442 Checking out an old version of a file
1443 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1444
1445 In the process of undoing a previous bad change, you may find it
1446 useful to check out an older version of a particular file using
1447 gitlink:git-checkout[1]. We've used git checkout before to switch
1448 branches, but it has quite different behavior if it is given a path
1449 name: the command
1450
1451 -------------------------------------------------
1452 $ git checkout HEAD^ path/to/file
1453 -------------------------------------------------
1454
1455 replaces path/to/file by the contents it had in the commit HEAD^, and
1456 also updates the index to match. It does not change branches.
1457
1458 If you just want to look at an old version of the file, without
1459 modifying the working directory, you can do that with
1460 gitlink:git-show[1]:
1461
1462 -------------------------------------------------
1463 $ git show HEAD^:path/to/file
1464 -------------------------------------------------
1465
1466 which will display the given version of the file.
1467
1468 [[interrupted-work]]
1469 Temporarily setting aside work in progress
1470 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1471
1472 While you are in the middle of working on something complicated, you
1473 find an unrelated but obvious and trivial bug. You would like to fix it
1474 before continuing. You can use gitlink:git-stash[1] to save the current
1475 state of your work, and after fixing the bug (or, optionally after doing
1476 so on a different branch and then coming back), unstash the
1477 work-in-progress changes.
1478
1479 ------------------------------------------------
1480 $ git stash "work in progress for foo feature"
1481 ------------------------------------------------
1482
1483 This command will save your changes away to the `stash`, and
1484 reset your working tree and the index to match the tip of your
1485 current branch. Then you can make your fix as usual.
1486
1487 ------------------------------------------------
1488 ... edit and test ...
1489 $ git commit -a -m "blorpl: typofix"
1490 ------------------------------------------------
1491
1492 After that, you can go back to what you were working on with
1493 `git stash apply`:
1494
1495 ------------------------------------------------
1496 $ git stash apply
1497 ------------------------------------------------
1498
1499
1500 [[ensuring-good-performance]]
1501 Ensuring good performance
1502 -------------------------
1503
1504 On large repositories, git depends on compression to keep the history
1505 information from taking up too much space on disk or in memory.
1506
1507 This compression is not performed automatically. Therefore you
1508 should occasionally run gitlink:git-gc[1]:
1509
1510 -------------------------------------------------
1511 $ git gc
1512 -------------------------------------------------
1513
1514 to recompress the archive. This can be very time-consuming, so
1515 you may prefer to run git-gc when you are not doing other work.
1516
1517
1518 [[ensuring-reliability]]
1519 Ensuring reliability
1520 --------------------
1521
1522 [[checking-for-corruption]]
1523 Checking the repository for corruption
1524 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1525
1526 The gitlink:git-fsck[1] command runs a number of self-consistency checks
1527 on the repository, and reports on any problems. This may take some
1528 time. The most common warning by far is about "dangling" objects:
1529
1530 -------------------------------------------------
1531 $ git fsck
1532 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1533 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1534 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1535 dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1536 dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1537 dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1538 dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1539 dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1540 ...
1541 -------------------------------------------------
1542
1543 Dangling objects are not a problem. At worst they may take up a little
1544 extra disk space. They can sometimes provide a last-resort method for
1545 recovering lost work--see <<dangling-objects>> for details. However, if
1546 you wish, you can remove them with gitlink:git-prune[1] or the `--prune`
1547 option to gitlink:git-gc[1]:
1548
1549 -------------------------------------------------
1550 $ git gc --prune
1551 -------------------------------------------------
1552
1553 This may be time-consuming. Unlike most other git operations (including
1554 git-gc when run without any options), it is not safe to prune while
1555 other git operations are in progress in the same repository.
1556
1557 [[recovering-lost-changes]]
1558 Recovering lost changes
1559 ~~~~~~~~~~~~~~~~~~~~~~~
1560
1561 [[reflogs]]
1562 Reflogs
1563 ^^^^^^^
1564
1565 Say you modify a branch with `gitlink:git-reset[1] --hard`, and then
1566 realize that the branch was the only reference you had to that point in
1567 history.
1568
1569 Fortunately, git also keeps a log, called a "reflog", of all the
1570 previous values of each branch. So in this case you can still find the
1571 old history using, for example,
1572
1573 -------------------------------------------------
1574 $ git log master@{1}
1575 -------------------------------------------------
1576
1577 This lists the commits reachable from the previous version of the
1578 "master" branch head. This syntax can be used with any git command
1579 that accepts a commit, not just with git log. Some other examples:
1580
1581 -------------------------------------------------
1582 $ git show master@{2} # See where the branch pointed 2,
1583 $ git show master@{3} # 3, ... changes ago.
1584 $ gitk master@{yesterday} # See where it pointed yesterday,
1585 $ gitk master@{"1 week ago"} # ... or last week
1586 $ git log --walk-reflogs master # show reflog entries for master
1587 -------------------------------------------------
1588
1589 A separate reflog is kept for the HEAD, so
1590
1591 -------------------------------------------------
1592 $ git show HEAD@{"1 week ago"}
1593 -------------------------------------------------
1594
1595 will show what HEAD pointed to one week ago, not what the current branch
1596 pointed to one week ago. This allows you to see the history of what
1597 you've checked out.
1598
1599 The reflogs are kept by default for 30 days, after which they may be
1600 pruned. See gitlink:git-reflog[1] and gitlink:git-gc[1] to learn
1601 how to control this pruning, and see the "SPECIFYING REVISIONS"
1602 section of gitlink:git-rev-parse[1] for details.
1603
1604 Note that the reflog history is very different from normal git history.
1605 While normal history is shared by every repository that works on the
1606 same project, the reflog history is not shared: it tells you only about
1607 how the branches in your local repository have changed over time.
1608
1609 [[dangling-object-recovery]]
1610 Examining dangling objects
1611 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1612
1613 In some situations the reflog may not be able to save you. For example,
1614 suppose you delete a branch, then realize you need the history it
1615 contained. The reflog is also deleted; however, if you have not yet
1616 pruned the repository, then you may still be able to find the lost
1617 commits in the dangling objects that git-fsck reports. See
1618 <<dangling-objects>> for the details.
1619
1620 -------------------------------------------------
1621 $ git fsck
1622 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1623 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1624 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1625 ...
1626 -------------------------------------------------
1627
1628 You can examine
1629 one of those dangling commits with, for example,
1630
1631 ------------------------------------------------
1632 $ gitk 7281251ddd --not --all
1633 ------------------------------------------------
1634
1635 which does what it sounds like: it says that you want to see the commit
1636 history that is described by the dangling commit(s), but not the
1637 history that is described by all your existing branches and tags. Thus
1638 you get exactly the history reachable from that commit that is lost.
1639 (And notice that it might not be just one commit: we only report the
1640 "tip of the line" as being dangling, but there might be a whole deep
1641 and complex commit history that was dropped.)
1642
1643 If you decide you want the history back, you can always create a new
1644 reference pointing to it, for example, a new branch:
1645
1646 ------------------------------------------------
1647 $ git branch recovered-branch 7281251ddd
1648 ------------------------------------------------
1649
1650 Other types of dangling objects (blobs and trees) are also possible, and
1651 dangling objects can arise in other situations.
1652
1653
1654 [[sharing-development]]
1655 Sharing development with others
1656 ===============================
1657
1658 [[getting-updates-with-git-pull]]
1659 Getting updates with git pull
1660 -----------------------------
1661
1662 After you clone a repository and make a few changes of your own, you
1663 may wish to check the original repository for updates and merge them
1664 into your own work.
1665
1666 We have already seen <<Updating-a-repository-with-git-fetch,how to
1667 keep remote tracking branches up to date>> with gitlink:git-fetch[1],
1668 and how to merge two branches. So you can merge in changes from the
1669 original repository's master branch with:
1670
1671 -------------------------------------------------
1672 $ git fetch
1673 $ git merge origin/master
1674 -------------------------------------------------
1675
1676 However, the gitlink:git-pull[1] command provides a way to do this in
1677 one step:
1678
1679 -------------------------------------------------
1680 $ git pull origin master
1681 -------------------------------------------------
1682
1683 In fact, if you have "master" checked out, then by default "git pull"
1684 merges from the HEAD branch of the origin repository. So often you can
1685 accomplish the above with just a simple
1686
1687 -------------------------------------------------
1688 $ git pull
1689 -------------------------------------------------
1690
1691 More generally, a branch that is created from a remote branch will pull
1692 by default from that branch. See the descriptions of the
1693 branch.<name>.remote and branch.<name>.merge options in
1694 gitlink:git-config[1], and the discussion of the `--track` option in
1695 gitlink:git-checkout[1], to learn how to control these defaults.
1696
1697 In addition to saving you keystrokes, "git pull" also helps you by
1698 producing a default commit message documenting the branch and
1699 repository that you pulled from.
1700
1701 (But note that no such commit will be created in the case of a
1702 <<fast-forwards,fast forward>>; instead, your branch will just be
1703 updated to point to the latest commit from the upstream branch.)
1704
1705 The git-pull command can also be given "." as the "remote" repository,
1706 in which case it just merges in a branch from the current repository; so
1707 the commands
1708
1709 -------------------------------------------------
1710 $ git pull . branch
1711 $ git merge branch
1712 -------------------------------------------------
1713
1714 are roughly equivalent. The former is actually very commonly used.
1715
1716 [[submitting-patches]]
1717 Submitting patches to a project
1718 -------------------------------
1719
1720 If you just have a few changes, the simplest way to submit them may
1721 just be to send them as patches in email:
1722
1723 First, use gitlink:git-format-patch[1]; for example:
1724
1725 -------------------------------------------------
1726 $ git format-patch origin
1727 -------------------------------------------------
1728
1729 will produce a numbered series of files in the current directory, one
1730 for each patch in the current branch but not in origin/HEAD.
1731
1732 You can then import these into your mail client and send them by
1733 hand. However, if you have a lot to send at once, you may prefer to
1734 use the gitlink:git-send-email[1] script to automate the process.
1735 Consult the mailing list for your project first to determine how they
1736 prefer such patches be handled.
1737
1738 [[importing-patches]]
1739 Importing patches to a project
1740 ------------------------------
1741
1742 Git also provides a tool called gitlink:git-am[1] (am stands for
1743 "apply mailbox"), for importing such an emailed series of patches.
1744 Just save all of the patch-containing messages, in order, into a
1745 single mailbox file, say "patches.mbox", then run
1746
1747 -------------------------------------------------
1748 $ git am -3 patches.mbox
1749 -------------------------------------------------
1750
1751 Git will apply each patch in order; if any conflicts are found, it
1752 will stop, and you can fix the conflicts as described in
1753 "<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1754 git to perform a merge; if you would prefer it just to abort and
1755 leave your tree and index untouched, you may omit that option.)
1756
1757 Once the index is updated with the results of the conflict
1758 resolution, instead of creating a new commit, just run
1759
1760 -------------------------------------------------
1761 $ git am --resolved
1762 -------------------------------------------------
1763
1764 and git will create the commit for you and continue applying the
1765 remaining patches from the mailbox.
1766
1767 The final result will be a series of commits, one for each patch in
1768 the original mailbox, with authorship and commit log message each
1769 taken from the message containing each patch.
1770
1771 [[public-repositories]]
1772 Public git repositories
1773 -----------------------
1774
1775 Another way to submit changes to a project is to tell the maintainer
1776 of that project to pull the changes from your repository using
1777 gitlink:git-pull[1]. In the section "<<getting-updates-with-git-pull,
1778 Getting updates with git pull>>" we described this as a way to get
1779 updates from the "main" repository, but it works just as well in the
1780 other direction.
1781
1782 If you and the maintainer both have accounts on the same machine, then
1783 you can just pull changes from each other's repositories directly;
1784 commands that accept repository URLs as arguments will also accept a
1785 local directory name:
1786
1787 -------------------------------------------------
1788 $ git clone /path/to/repository
1789 $ git pull /path/to/other/repository
1790 -------------------------------------------------
1791
1792 or an ssh URL:
1793
1794 -------------------------------------------------
1795 $ git clone ssh://yourhost/~you/repository
1796 -------------------------------------------------
1797
1798 For projects with few developers, or for synchronizing a few private
1799 repositories, this may be all you need.
1800
1801 However, the more common way to do this is to maintain a separate public
1802 repository (usually on a different host) for others to pull changes
1803 from. This is usually more convenient, and allows you to cleanly
1804 separate private work in progress from publicly visible work.
1805
1806 You will continue to do your day-to-day work in your personal
1807 repository, but periodically "push" changes from your personal
1808 repository into your public repository, allowing other developers to
1809 pull from that repository. So the flow of changes, in a situation
1810 where there is one other developer with a public repository, looks
1811 like this:
1812
1813 you push
1814 your personal repo ------------------> your public repo
1815 ^ |
1816 | |
1817 | you pull | they pull
1818 | |
1819 | |
1820 | they push V
1821 their public repo <------------------- their repo
1822
1823 We explain how to do this in the following sections.
1824
1825 [[setting-up-a-public-repository]]
1826 Setting up a public repository
1827 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1828
1829 Assume your personal repository is in the directory ~/proj. We
1830 first create a new clone of the repository and tell git-daemon that it
1831 is meant to be public:
1832
1833 -------------------------------------------------
1834 $ git clone --bare ~/proj proj.git
1835 $ touch proj.git/git-daemon-export-ok
1836 -------------------------------------------------
1837
1838 The resulting directory proj.git contains a "bare" git repository--it is
1839 just the contents of the ".git" directory, without any files checked out
1840 around it.
1841
1842 Next, copy proj.git to the server where you plan to host the
1843 public repository. You can use scp, rsync, or whatever is most
1844 convenient.
1845
1846 [[exporting-via-git]]
1847 Exporting a git repository via the git protocol
1848 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1849
1850 This is the preferred method.
1851
1852 If someone else administers the server, they should tell you what
1853 directory to put the repository in, and what git:// URL it will appear
1854 at. You can then skip to the section
1855 "<<pushing-changes-to-a-public-repository,Pushing changes to a public
1856 repository>>", below.
1857
1858 Otherwise, all you need to do is start gitlink:git-daemon[1]; it will
1859 listen on port 9418. By default, it will allow access to any directory
1860 that looks like a git directory and contains the magic file
1861 git-daemon-export-ok. Passing some directory paths as git-daemon
1862 arguments will further restrict the exports to those paths.
1863
1864 You can also run git-daemon as an inetd service; see the
1865 gitlink:git-daemon[1] man page for details. (See especially the
1866 examples section.)
1867
1868 [[exporting-via-http]]
1869 Exporting a git repository via http
1870 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1871
1872 The git protocol gives better performance and reliability, but on a
1873 host with a web server set up, http exports may be simpler to set up.
1874
1875 All you need to do is place the newly created bare git repository in
1876 a directory that is exported by the web server, and make some
1877 adjustments to give web clients some extra information they need:
1878
1879 -------------------------------------------------
1880 $ mv proj.git /home/you/public_html/proj.git
1881 $ cd proj.git
1882 $ git --bare update-server-info
1883 $ chmod a+x hooks/post-update
1884 -------------------------------------------------
1885
1886 (For an explanation of the last two lines, see
1887 gitlink:git-update-server-info[1], and the documentation
1888 link:hooks.html[Hooks used by git].)
1889
1890 Advertise the URL of proj.git. Anybody else should then be able to
1891 clone or pull from that URL, for example with a command line like:
1892
1893 -------------------------------------------------
1894 $ git clone http://yourserver.com/~you/proj.git
1895 -------------------------------------------------
1896
1897 (See also
1898 link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1899 for a slightly more sophisticated setup using WebDAV which also
1900 allows pushing over http.)
1901
1902 [[pushing-changes-to-a-public-repository]]
1903 Pushing changes to a public repository
1904 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1905
1906 Note that the two techniques outlined above (exporting via
1907 <<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1908 maintainers to fetch your latest changes, but they do not allow write
1909 access, which you will need to update the public repository with the
1910 latest changes created in your private repository.
1911
1912 The simplest way to do this is using gitlink:git-push[1] and ssh; to
1913 update the remote branch named "master" with the latest state of your
1914 branch named "master", run
1915
1916 -------------------------------------------------
1917 $ git push ssh://yourserver.com/~you/proj.git master:master
1918 -------------------------------------------------
1919
1920 or just
1921
1922 -------------------------------------------------
1923 $ git push ssh://yourserver.com/~you/proj.git master
1924 -------------------------------------------------
1925
1926 As with git-fetch, git-push will complain if this does not result in
1927 a <<fast-forwards,fast forward>>. Normally this is a sign of
1928 something wrong. However, if you are sure you know what you're
1929 doing, you may force git-push to perform the update anyway by
1930 preceding the branch name by a plus sign:
1931
1932 -------------------------------------------------
1933 $ git push ssh://yourserver.com/~you/proj.git +master
1934 -------------------------------------------------
1935
1936 Note that the target of a "push" is normally a
1937 <<def_bare_repository,bare>> repository. You can also push to a
1938 repository that has a checked-out working tree, but the working tree
1939 will not be updated by the push. This may lead to unexpected results if
1940 the branch you push to is the currently checked-out branch!
1941
1942 As with git-fetch, you may also set up configuration options to
1943 save typing; so, for example, after
1944
1945 -------------------------------------------------
1946 $ cat >>.git/config <<EOF
1947 [remote "public-repo"]
1948 url = ssh://yourserver.com/~you/proj.git
1949 EOF
1950 -------------------------------------------------
1951
1952 you should be able to perform the above push with just
1953
1954 -------------------------------------------------
1955 $ git push public-repo master
1956 -------------------------------------------------
1957
1958 See the explanations of the remote.<name>.url, branch.<name>.remote,
1959 and remote.<name>.push options in gitlink:git-config[1] for
1960 details.
1961
1962 [[setting-up-a-shared-repository]]
1963 Setting up a shared repository
1964 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1965
1966 Another way to collaborate is by using a model similar to that
1967 commonly used in CVS, where several developers with special rights
1968 all push to and pull from a single shared repository. See
1969 link:cvs-migration.html[git for CVS users] for instructions on how to
1970 set this up.
1971
1972 However, while there is nothing wrong with git's support for shared
1973 repositories, this mode of operation is not generally recommended,
1974 simply because the mode of collaboration that git supports--by
1975 exchanging patches and pulling from public repositories--has so many
1976 advantages over the central shared repository:
1977
1978 - Git's ability to quickly import and merge patches allows a
1979 single maintainer to process incoming changes even at very
1980 high rates. And when that becomes too much, git-pull provides
1981 an easy way for that maintainer to delegate this job to other
1982 maintainers while still allowing optional review of incoming
1983 changes.
1984 - Since every developer's repository has the same complete copy
1985 of the project history, no repository is special, and it is
1986 trivial for another developer to take over maintenance of a
1987 project, either by mutual agreement, or because a maintainer
1988 becomes unresponsive or difficult to work with.
1989 - The lack of a central group of "committers" means there is
1990 less need for formal decisions about who is "in" and who is
1991 "out".
1992
1993 [[setting-up-gitweb]]
1994 Allowing web browsing of a repository
1995 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1996
1997 The gitweb cgi script provides users an easy way to browse your
1998 project's files and history without having to install git; see the file
1999 gitweb/INSTALL in the git source tree for instructions on setting it up.
2000
2001 [[sharing-development-examples]]
2002 Examples
2003 --------
2004
2005 [[maintaining-topic-branches]]
2006 Maintaining topic branches for a Linux subsystem maintainer
2007 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2008
2009 This describes how Tony Luck uses git in his role as maintainer of the
2010 IA64 architecture for the Linux kernel.
2011
2012 He uses two public branches:
2013
2014 - A "test" tree into which patches are initially placed so that they
2015 can get some exposure when integrated with other ongoing development.
2016 This tree is available to Andrew for pulling into -mm whenever he
2017 wants.
2018
2019 - A "release" tree into which tested patches are moved for final sanity
2020 checking, and as a vehicle to send them upstream to Linus (by sending
2021 him a "please pull" request.)
2022
2023 He also uses a set of temporary branches ("topic branches"), each
2024 containing a logical grouping of patches.
2025
2026 To set this up, first create your work tree by cloning Linus's public
2027 tree:
2028
2029 -------------------------------------------------
2030 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git work
2031 $ cd work
2032 -------------------------------------------------
2033
2034 Linus's tree will be stored in the remote branch named origin/master,
2035 and can be updated using gitlink:git-fetch[1]; you can track other
2036 public trees using gitlink:git-remote[1] to set up a "remote" and
2037 gitlink:git-fetch[1] to keep them up-to-date; see
2038 <<repositories-and-branches>>.
2039
2040 Now create the branches in which you are going to work; these start out
2041 at the current tip of origin/master branch, and should be set up (using
2042 the --track option to gitlink:git-branch[1]) to merge changes in from
2043 Linus by default.
2044
2045 -------------------------------------------------
2046 $ git branch --track test origin/master
2047 $ git branch --track release origin/master
2048 -------------------------------------------------
2049
2050 These can be easily kept up to date using gitlink:git-pull[1].
2051
2052 -------------------------------------------------
2053 $ git checkout test && git pull
2054 $ git checkout release && git pull
2055 -------------------------------------------------
2056
2057 Important note! If you have any local changes in these branches, then
2058 this merge will create a commit object in the history (with no local
2059 changes git will simply do a "Fast forward" merge). Many people dislike
2060 the "noise" that this creates in the Linux history, so you should avoid
2061 doing this capriciously in the "release" branch, as these noisy commits
2062 will become part of the permanent history when you ask Linus to pull
2063 from the release branch.
2064
2065 A few configuration variables (see gitlink:git-config[1]) can
2066 make it easy to push both branches to your public tree. (See
2067 <<setting-up-a-public-repository>>.)
2068
2069 -------------------------------------------------
2070 $ cat >> .git/config <<EOF
2071 [remote "mytree"]
2072 url = master.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6.git
2073 push = release
2074 push = test
2075 EOF
2076 -------------------------------------------------
2077
2078 Then you can push both the test and release trees using
2079 gitlink:git-push[1]:
2080
2081 -------------------------------------------------
2082 $ git push mytree
2083 -------------------------------------------------
2084
2085 or push just one of the test and release branches using:
2086
2087 -------------------------------------------------
2088 $ git push mytree test
2089 -------------------------------------------------
2090
2091 or
2092
2093 -------------------------------------------------
2094 $ git push mytree release
2095 -------------------------------------------------
2096
2097 Now to apply some patches from the community. Think of a short
2098 snappy name for a branch to hold this patch (or related group of
2099 patches), and create a new branch from the current tip of Linus's
2100 branch:
2101
2102 -------------------------------------------------
2103 $ git checkout -b speed-up-spinlocks origin
2104 -------------------------------------------------
2105
2106 Now you apply the patch(es), run some tests, and commit the change(s). If
2107 the patch is a multi-part series, then you should apply each as a separate
2108 commit to this branch.
2109
2110 -------------------------------------------------
2111 $ ... patch ... test ... commit [ ... patch ... test ... commit ]*
2112 -------------------------------------------------
2113
2114 When you are happy with the state of this change, you can pull it into the
2115 "test" branch in preparation to make it public:
2116
2117 -------------------------------------------------
2118 $ git checkout test && git pull . speed-up-spinlocks
2119 -------------------------------------------------
2120
2121 It is unlikely that you would have any conflicts here ... but you might if you
2122 spent a while on this step and had also pulled new versions from upstream.
2123
2124 Some time later when enough time has passed and testing done, you can pull the
2125 same branch into the "release" tree ready to go upstream. This is where you
2126 see the value of keeping each patch (or patch series) in its own branch. It
2127 means that the patches can be moved into the "release" tree in any order.
2128
2129 -------------------------------------------------
2130 $ git checkout release && git pull . speed-up-spinlocks
2131 -------------------------------------------------
2132
2133 After a while, you will have a number of branches, and despite the
2134 well chosen names you picked for each of them, you may forget what
2135 they are for, or what status they are in. To get a reminder of what
2136 changes are in a specific branch, use:
2137
2138 -------------------------------------------------
2139 $ git log linux..branchname | git-shortlog
2140 -------------------------------------------------
2141
2142 To see whether it has already been merged into the test or release branches,
2143 use:
2144
2145 -------------------------------------------------
2146 $ git log test..branchname
2147 -------------------------------------------------
2148
2149 or
2150
2151 -------------------------------------------------
2152 $ git log release..branchname
2153 -------------------------------------------------
2154
2155 (If this branch has not yet been merged, you will see some log entries.
2156 If it has been merged, then there will be no output.)
2157
2158 Once a patch completes the great cycle (moving from test to release,
2159 then pulled by Linus, and finally coming back into your local
2160 "origin/master" branch), the branch for this change is no longer needed.
2161 You detect this when the output from:
2162
2163 -------------------------------------------------
2164 $ git log origin..branchname
2165 -------------------------------------------------
2166
2167 is empty. At this point the branch can be deleted:
2168
2169 -------------------------------------------------
2170 $ git branch -d branchname
2171 -------------------------------------------------
2172
2173 Some changes are so trivial that it is not necessary to create a separate
2174 branch and then merge into each of the test and release branches. For
2175 these changes, just apply directly to the "release" branch, and then
2176 merge that into the "test" branch.
2177
2178 To create diffstat and shortlog summaries of changes to include in a "please
2179 pull" request to Linus you can use:
2180
2181 -------------------------------------------------
2182 $ git diff --stat origin..release
2183 -------------------------------------------------
2184
2185 and
2186
2187 -------------------------------------------------
2188 $ git log -p origin..release | git shortlog
2189 -------------------------------------------------
2190
2191 Here are some of the scripts that simplify all this even further.
2192
2193 -------------------------------------------------
2194 ==== update script ====
2195 # Update a branch in my GIT tree. If the branch to be updated
2196 # is origin, then pull from kernel.org. Otherwise merge
2197 # origin/master branch into test|release branch
2198
2199 case "$1" in
2200 test|release)
2201 git checkout $1 && git pull . origin
2202 ;;
2203 origin)
2204 before=$(git rev-parse refs/remotes/origin/master)
2205 git fetch origin
2206 after=$(git rev-parse refs/remotes/origin/master)
2207 if [ $before != $after ]
2208 then
2209 git log $before..$after | git shortlog
2210 fi
2211 ;;
2212 *)
2213 echo "Usage: $0 origin|test|release" 1>&2
2214 exit 1
2215 ;;
2216 esac
2217 -------------------------------------------------
2218
2219 -------------------------------------------------
2220 ==== merge script ====
2221 # Merge a branch into either the test or release branch
2222
2223 pname=$0
2224
2225 usage()
2226 {
2227 echo "Usage: $pname branch test|release" 1>&2
2228 exit 1
2229 }
2230
2231 git show-ref -q --verify -- refs/heads/"$1" || {
2232 echo "Can't see branch <$1>" 1>&2
2233 usage
2234 }
2235
2236 case "$2" in
2237 test|release)
2238 if [ $(git log $2..$1 | wc -c) -eq 0 ]
2239 then
2240 echo $1 already merged into $2 1>&2
2241 exit 1
2242 fi
2243 git checkout $2 && git pull . $1
2244 ;;
2245 *)
2246 usage
2247 ;;
2248 esac
2249 -------------------------------------------------
2250
2251 -------------------------------------------------
2252 ==== status script ====
2253 # report on status of my ia64 GIT tree
2254
2255 gb=$(tput setab 2)
2256 rb=$(tput setab 1)
2257 restore=$(tput setab 9)
2258
2259 if [ `git rev-list test..release | wc -c` -gt 0 ]
2260 then
2261 echo $rb Warning: commits in release that are not in test $restore
2262 git log test..release
2263 fi
2264
2265 for branch in `git show-ref --heads | sed 's|^.*/||'`
2266 do
2267 if [ $branch = test -o $branch = release ]
2268 then
2269 continue
2270 fi
2271
2272 echo -n $gb ======= $branch ====== $restore " "
2273 status=
2274 for ref in test release origin/master
2275 do
2276 if [ `git rev-list $ref..$branch | wc -c` -gt 0 ]
2277 then
2278 status=$status${ref:0:1}
2279 fi
2280 done
2281 case $status in
2282 trl)
2283 echo $rb Need to pull into test $restore
2284 ;;
2285 rl)
2286 echo "In test"
2287 ;;
2288 l)
2289 echo "Waiting for linus"
2290 ;;
2291 "")
2292 echo $rb All done $restore
2293 ;;
2294 *)
2295 echo $rb "<$status>" $restore
2296 ;;
2297 esac
2298 git log origin/master..$branch | git shortlog
2299 done
2300 -------------------------------------------------
2301
2302
2303 [[cleaning-up-history]]
2304 Rewriting history and maintaining patch series
2305 ==============================================
2306
2307 Normally commits are only added to a project, never taken away or
2308 replaced. Git is designed with this assumption, and violating it will
2309 cause git's merge machinery (for example) to do the wrong thing.
2310
2311 However, there is a situation in which it can be useful to violate this
2312 assumption.
2313
2314 [[patch-series]]
2315 Creating the perfect patch series
2316 ---------------------------------
2317
2318 Suppose you are a contributor to a large project, and you want to add a
2319 complicated feature, and to present it to the other developers in a way
2320 that makes it easy for them to read your changes, verify that they are
2321 correct, and understand why you made each change.
2322
2323 If you present all of your changes as a single patch (or commit), they
2324 may find that it is too much to digest all at once.
2325
2326 If you present them with the entire history of your work, complete with
2327 mistakes, corrections, and dead ends, they may be overwhelmed.
2328
2329 So the ideal is usually to produce a series of patches such that:
2330
2331 1. Each patch can be applied in order.
2332
2333 2. Each patch includes a single logical change, together with a
2334 message explaining the change.
2335
2336 3. No patch introduces a regression: after applying any initial
2337 part of the series, the resulting project still compiles and
2338 works, and has no bugs that it didn't have before.
2339
2340 4. The complete series produces the same end result as your own
2341 (probably much messier!) development process did.
2342
2343 We will introduce some tools that can help you do this, explain how to
2344 use them, and then explain some of the problems that can arise because
2345 you are rewriting history.
2346
2347 [[using-git-rebase]]
2348 Keeping a patch series up to date using git-rebase
2349 --------------------------------------------------
2350
2351 Suppose that you create a branch "mywork" on a remote-tracking branch
2352 "origin", and create some commits on top of it:
2353
2354 -------------------------------------------------
2355 $ git checkout -b mywork origin
2356 $ vi file.txt
2357 $ git commit
2358 $ vi otherfile.txt
2359 $ git commit
2360 ...
2361 -------------------------------------------------
2362
2363 You have performed no merges into mywork, so it is just a simple linear
2364 sequence of patches on top of "origin":
2365
2366 ................................................
2367 o--o--o <-- origin
2368 \
2369 o--o--o <-- mywork
2370 ................................................
2371
2372 Some more interesting work has been done in the upstream project, and
2373 "origin" has advanced:
2374
2375 ................................................
2376 o--o--O--o--o--o <-- origin
2377 \
2378 a--b--c <-- mywork
2379 ................................................
2380
2381 At this point, you could use "pull" to merge your changes back in;
2382 the result would create a new merge commit, like this:
2383
2384 ................................................
2385 o--o--O--o--o--o <-- origin
2386 \ \
2387 a--b--c--m <-- mywork
2388 ................................................
2389
2390 However, if you prefer to keep the history in mywork a simple series of
2391 commits without any merges, you may instead choose to use
2392 gitlink:git-rebase[1]:
2393
2394 -------------------------------------------------
2395 $ git checkout mywork
2396 $ git rebase origin
2397 -------------------------------------------------
2398
2399 This will remove each of your commits from mywork, temporarily saving
2400 them as patches (in a directory named ".dotest"), update mywork to
2401 point at the latest version of origin, then apply each of the saved
2402 patches to the new mywork. The result will look like:
2403
2404
2405 ................................................
2406 o--o--O--o--o--o <-- origin
2407 \
2408 a'--b'--c' <-- mywork
2409 ................................................
2410
2411 In the process, it may discover conflicts. In that case it will stop
2412 and allow you to fix the conflicts; after fixing conflicts, use "git
2413 add" to update the index with those contents, and then, instead of
2414 running git-commit, just run
2415
2416 -------------------------------------------------
2417 $ git rebase --continue
2418 -------------------------------------------------
2419
2420 and git will continue applying the rest of the patches.
2421
2422 At any point you may use the `--abort` option to abort this process and
2423 return mywork to the state it had before you started the rebase:
2424
2425 -------------------------------------------------
2426 $ git rebase --abort
2427 -------------------------------------------------
2428
2429 [[modifying-one-commit]]
2430 Modifying a single commit
2431 -------------------------
2432
2433 We saw in <<fixing-a-mistake-by-editing-history>> that you can replace the
2434 most recent commit using
2435
2436 -------------------------------------------------
2437 $ git commit --amend
2438 -------------------------------------------------
2439
2440 which will replace the old commit by a new commit incorporating your
2441 changes, giving you a chance to edit the old commit message first.
2442
2443 You can also use a combination of this and gitlink:git-rebase[1] to edit
2444 commits further back in your history. First, tag the problematic commit with
2445
2446 -------------------------------------------------
2447 $ git tag bad mywork~5
2448 -------------------------------------------------
2449
2450 (Either gitk or git-log may be useful for finding the commit.)
2451
2452 Then check out that commit, edit it, and rebase the rest of the series
2453 on top of it (note that we could check out the commit on a temporary
2454 branch, but instead we're using a <<detached-head,detached head>>):
2455
2456 -------------------------------------------------
2457 $ git checkout bad
2458 $ # make changes here and update the index
2459 $ git commit --amend
2460 $ git rebase --onto HEAD bad mywork
2461 -------------------------------------------------
2462
2463 When you're done, you'll be left with mywork checked out, with the top
2464 patches on mywork reapplied on top of your modified commit. You can
2465 then clean up with
2466
2467 -------------------------------------------------
2468 $ git tag -d bad
2469 -------------------------------------------------
2470
2471 Note that the immutable nature of git history means that you haven't really
2472 "modified" existing commits; instead, you have replaced the old commits with
2473 new commits having new object names.
2474
2475 [[reordering-patch-series]]
2476 Reordering or selecting from a patch series
2477 -------------------------------------------
2478
2479 Given one existing commit, the gitlink:git-cherry-pick[1] command
2480 allows you to apply the change introduced by that commit and create a
2481 new commit that records it. So, for example, if "mywork" points to a
2482 series of patches on top of "origin", you might do something like:
2483
2484 -------------------------------------------------
2485 $ git checkout -b mywork-new origin
2486 $ gitk origin..mywork &
2487 -------------------------------------------------
2488
2489 and browse through the list of patches in the mywork branch using gitk,
2490 applying them (possibly in a different order) to mywork-new using
2491 cherry-pick, and possibly modifying them as you go using `commit --amend`.
2492 The gitlink:git-gui[1] command may also help as it allows you to
2493 individually select diff hunks for inclusion in the index (by
2494 right-clicking on the diff hunk and choosing "Stage Hunk for Commit").
2495
2496 Another technique is to use git-format-patch to create a series of
2497 patches, then reset the state to before the patches:
2498
2499 -------------------------------------------------
2500 $ git format-patch origin
2501 $ git reset --hard origin
2502 -------------------------------------------------
2503
2504 Then modify, reorder, or eliminate patches as preferred before applying
2505 them again with gitlink:git-am[1].
2506
2507 [[patch-series-tools]]
2508 Other tools
2509 -----------
2510
2511 There are numerous other tools, such as StGIT, which exist for the
2512 purpose of maintaining a patch series. These are outside of the scope of
2513 this manual.
2514
2515 [[problems-with-rewriting-history]]
2516 Problems with rewriting history
2517 -------------------------------
2518
2519 The primary problem with rewriting the history of a branch has to do
2520 with merging. Suppose somebody fetches your branch and merges it into
2521 their branch, with a result something like this:
2522
2523 ................................................
2524 o--o--O--o--o--o <-- origin
2525 \ \
2526 t--t--t--m <-- their branch:
2527 ................................................
2528
2529 Then suppose you modify the last three commits:
2530
2531 ................................................
2532 o--o--o <-- new head of origin
2533 /
2534 o--o--O--o--o--o <-- old head of origin
2535 ................................................
2536
2537 If we examined all this history together in one repository, it will
2538 look like:
2539
2540 ................................................
2541 o--o--o <-- new head of origin
2542 /
2543 o--o--O--o--o--o <-- old head of origin
2544 \ \
2545 t--t--t--m <-- their branch:
2546 ................................................
2547
2548 Git has no way of knowing that the new head is an updated version of
2549 the old head; it treats this situation exactly the same as it would if
2550 two developers had independently done the work on the old and new heads
2551 in parallel. At this point, if someone attempts to merge the new head
2552 in to their branch, git will attempt to merge together the two (old and
2553 new) lines of development, instead of trying to replace the old by the
2554 new. The results are likely to be unexpected.
2555
2556 You may still choose to publish branches whose history is rewritten,
2557 and it may be useful for others to be able to fetch those branches in
2558 order to examine or test them, but they should not attempt to pull such
2559 branches into their own work.
2560
2561 For true distributed development that supports proper merging,
2562 published branches should never be rewritten.
2563
2564 [[bisect-merges]]
2565 Why bisecting merge commits can be harder than bisecting linear history
2566 -----------------------------------------------------------------------
2567
2568 The gitlink:git-bisect[1] command correctly handles history that
2569 includes merge commits. However, when the commit that it finds is a
2570 merge commit, the user may need to work harder than usual to figure out
2571 why that commit introduced a problem.
2572
2573 Imagine this history:
2574
2575 ................................................
2576 ---Z---o---X---...---o---A---C---D
2577 \ /
2578 o---o---Y---...---o---B
2579 ................................................
2580
2581 Suppose that on the upper line of development, the meaning of one
2582 of the functions that exists at Z is changed at commit X. The
2583 commits from Z leading to A change both the function's
2584 implementation and all calling sites that exist at Z, as well
2585 as new calling sites they add, to be consistent. There is no
2586 bug at A.
2587
2588 Suppose that in the meantime on the lower line of development somebody
2589 adds a new calling site for that function at commit Y. The
2590 commits from Z leading to B all assume the old semantics of that
2591 function and the callers and the callee are consistent with each
2592 other. There is no bug at B, either.
2593
2594 Suppose further that the two development lines merge cleanly at C,
2595 so no conflict resolution is required.
2596
2597 Nevertheless, the code at C is broken, because the callers added
2598 on the lower line of development have not been converted to the new
2599 semantics introduced on the upper line of development. So if all
2600 you know is that D is bad, that Z is good, and that
2601 gitlink:git-bisect[1] identifies C as the culprit, how will you
2602 figure out that the problem is due to this change in semantics?
2603
2604 When the result of a git-bisect is a non-merge commit, you should
2605 normally be able to discover the problem by examining just that commit.
2606 Developers can make this easy by breaking their changes into small
2607 self-contained commits. That won't help in the case above, however,
2608 because the problem isn't obvious from examination of any single
2609 commit; instead, a global view of the development is required. To
2610 make matters worse, the change in semantics in the problematic
2611 function may be just one small part of the changes in the upper
2612 line of development.
2613
2614 On the other hand, if instead of merging at C you had rebased the
2615 history between Z to B on top of A, you would have gotten this
2616 linear history:
2617
2618 ................................................................
2619 ---Z---o---X--...---o---A---o---o---Y*--...---o---B*--D*
2620 ................................................................
2621
2622 Bisecting between Z and D* would hit a single culprit commit Y*,
2623 and understanding why Y* was broken would probably be easier.
2624
2625 Partly for this reason, many experienced git users, even when
2626 working on an otherwise merge-heavy project, keep the history
2627 linear by rebasing against the latest upstream version before
2628 publishing.
2629
2630 [[advanced-branch-management]]
2631 Advanced branch management
2632 ==========================
2633
2634 [[fetching-individual-branches]]
2635 Fetching individual branches
2636 ----------------------------
2637
2638 Instead of using gitlink:git-remote[1], you can also choose just
2639 to update one branch at a time, and to store it locally under an
2640 arbitrary name:
2641
2642 -------------------------------------------------
2643 $ git fetch origin todo:my-todo-work
2644 -------------------------------------------------
2645
2646 The first argument, "origin", just tells git to fetch from the
2647 repository you originally cloned from. The second argument tells git
2648 to fetch the branch named "todo" from the remote repository, and to
2649 store it locally under the name refs/heads/my-todo-work.
2650
2651 You can also fetch branches from other repositories; so
2652
2653 -------------------------------------------------
2654 $ git fetch git://example.com/proj.git master:example-master
2655 -------------------------------------------------
2656
2657 will create a new branch named "example-master" and store in it the
2658 branch named "master" from the repository at the given URL. If you
2659 already have a branch named example-master, it will attempt to
2660 <<fast-forwards,fast-forward>> to the commit given by example.com's
2661 master branch. In more detail:
2662
2663 [[fetch-fast-forwards]]
2664 git fetch and fast-forwards
2665 ---------------------------
2666
2667 In the previous example, when updating an existing branch, "git
2668 fetch" checks to make sure that the most recent commit on the remote
2669 branch is a descendant of the most recent commit on your copy of the
2670 branch before updating your copy of the branch to point at the new
2671 commit. Git calls this process a <<fast-forwards,fast forward>>.
2672
2673 A fast forward looks something like this:
2674
2675 ................................................
2676 o--o--o--o <-- old head of the branch
2677 \
2678 o--o--o <-- new head of the branch
2679 ................................................
2680
2681
2682 In some cases it is possible that the new head will *not* actually be
2683 a descendant of the old head. For example, the developer may have
2684 realized she made a serious mistake, and decided to backtrack,
2685 resulting in a situation like:
2686
2687 ................................................
2688 o--o--o--o--a--b <-- old head of the branch
2689 \
2690 o--o--o <-- new head of the branch
2691 ................................................
2692
2693 In this case, "git fetch" will fail, and print out a warning.
2694
2695 In that case, you can still force git to update to the new head, as
2696 described in the following section. However, note that in the
2697 situation above this may mean losing the commits labeled "a" and "b",
2698 unless you've already created a reference of your own pointing to
2699 them.
2700
2701 [[forcing-fetch]]
2702 Forcing git fetch to do non-fast-forward updates
2703 ------------------------------------------------
2704
2705 If git fetch fails because the new head of a branch is not a
2706 descendant of the old head, you may force the update with:
2707
2708 -------------------------------------------------
2709 $ git fetch git://example.com/proj.git +master:refs/remotes/example/master
2710 -------------------------------------------------
2711
2712 Note the addition of the "+" sign. Alternatively, you can use the "-f"
2713 flag to force updates of all the fetched branches, as in:
2714
2715 -------------------------------------------------
2716 $ git fetch -f origin
2717 -------------------------------------------------
2718
2719 Be aware that commits that the old version of example/master pointed at
2720 may be lost, as we saw in the previous section.
2721
2722 [[remote-branch-configuration]]
2723 Configuring remote branches
2724 ---------------------------
2725
2726 We saw above that "origin" is just a shortcut to refer to the
2727 repository that you originally cloned from. This information is
2728 stored in git configuration variables, which you can see using
2729 gitlink:git-config[1]:
2730
2731 -------------------------------------------------
2732 $ git config -l
2733 core.repositoryformatversion=0
2734 core.filemode=true
2735 core.logallrefupdates=true
2736 remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2737 remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2738 branch.master.remote=origin
2739 branch.master.merge=refs/heads/master
2740 -------------------------------------------------
2741
2742 If there are other repositories that you also use frequently, you can
2743 create similar configuration options to save typing; for example,
2744 after
2745
2746 -------------------------------------------------
2747 $ git config remote.example.url git://example.com/proj.git
2748 -------------------------------------------------
2749
2750 then the following two commands will do the same thing:
2751
2752 -------------------------------------------------
2753 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2754 $ git fetch example master:refs/remotes/example/master
2755 -------------------------------------------------
2756
2757 Even better, if you add one more option:
2758
2759 -------------------------------------------------
2760 $ git config remote.example.fetch master:refs/remotes/example/master
2761 -------------------------------------------------
2762
2763 then the following commands will all do the same thing:
2764
2765 -------------------------------------------------
2766 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2767 $ git fetch example master:refs/remotes/example/master
2768 $ git fetch example
2769 -------------------------------------------------
2770
2771 You can also add a "+" to force the update each time:
2772
2773 -------------------------------------------------
2774 $ git config remote.example.fetch +master:ref/remotes/example/master
2775 -------------------------------------------------
2776
2777 Don't do this unless you're sure you won't mind "git fetch" possibly
2778 throwing away commits on mybranch.
2779
2780 Also note that all of the above configuration can be performed by
2781 directly editing the file .git/config instead of using
2782 gitlink:git-config[1].
2783
2784 See gitlink:git-config[1] for more details on the configuration
2785 options mentioned above.
2786
2787
2788 [[git-concepts]]
2789 Git concepts
2790 ============
2791
2792 Git is built on a small number of simple but powerful ideas. While it
2793 is possible to get things done without understanding them, you will find
2794 git much more intuitive if you do.
2795
2796 We start with the most important, the <<def_object_database,object
2797 database>> and the <<def_index,index>>.
2798
2799 [[the-object-database]]
2800 The Object Database
2801 -------------------
2802
2803
2804 We already saw in <<understanding-commits>> that all commits are stored
2805 under a 40-digit "object name". In fact, all the information needed to
2806 represent the history of a project is stored in objects with such names.
2807 In each case the name is calculated by taking the SHA1 hash of the
2808 contents of the object. The SHA1 hash is a cryptographic hash function.
2809 What that means to us is that it is impossible to find two different
2810 objects with the same name. This has a number of advantages; among
2811 others:
2812
2813 - Git can quickly determine whether two objects are identical or not,
2814 just by comparing names.
2815 - Since object names are computed the same way in every repository, the
2816 same content stored in two repositories will always be stored under
2817 the same name.
2818 - Git can detect errors when it reads an object, by checking that the
2819 object's name is still the SHA1 hash of its contents.
2820
2821 (See <<object-details>> for the details of the object formatting and
2822 SHA1 calculation.)
2823
2824 There are four different types of objects: "blob", "tree", "commit", and
2825 "tag".
2826
2827 - A <<def_blob_object,"blob" object>> is used to store file data.
2828 - A <<def_tree_object,"tree" object>> is an object that ties one or more
2829 "blob" objects into a directory structure. In addition, a tree object
2830 can refer to other tree objects, thus creating a directory hierarchy.
2831 - A <<def_commit_object,"commit" object>> ties such directory hierarchies
2832 together into a <<def_DAG,directed acyclic graph>> of revisions--each
2833 commit contains the object name of exactly one tree designating the
2834 directory hierarchy at the time of the commit. In addition, a commit
2835 refers to "parent" commit objects that describe the history of how we
2836 arrived at that directory hierarchy.
2837 - A <<def_tag_object,"tag" object>> symbolically identifies and can be
2838 used to sign other objects. It contains the object name and type of
2839 another object, a symbolic name (of course!) and, optionally, a
2840 signature.
2841
2842 The object types in some more detail:
2843
2844 [[commit-object]]
2845 Commit Object
2846 ~~~~~~~~~~~~~
2847
2848 The "commit" object links a physical state of a tree with a description
2849 of how we got there and why. Use the --pretty=raw option to
2850 gitlink:git-show[1] or gitlink:git-log[1] to examine your favorite
2851 commit:
2852
2853 ------------------------------------------------
2854 $ git show -s --pretty=raw 2be7fcb476
2855 commit 2be7fcb4764f2dbcee52635b91fedb1b3dcf7ab4
2856 tree fb3a8bdd0ceddd019615af4d57a53f43d8cee2bf
2857 parent 257a84d9d02e90447b149af58b271c19405edb6a
2858 author Dave Watson <dwatson@mimvista.com> 1187576872 -0400
2859 committer Junio C Hamano <gitster@pobox.com> 1187591163 -0700
2860
2861 Fix misspelling of 'suppress' in docs
2862
2863 Signed-off-by: Junio C Hamano <gitster@pobox.com>
2864 ------------------------------------------------
2865
2866 As you can see, a commit is defined by:
2867
2868 - a tree: The SHA1 name of a tree object (as defined below), representing
2869 the contents of a directory at a certain point in time.
2870 - parent(s): The SHA1 name of some number of commits which represent the
2871 immediately prevoius step(s) in the history of the project. The
2872 example above has one parent; merge commits may have more than
2873 one. A commit with no parents is called a "root" commit, and
2874 represents the initial revision of a project. Each project must have
2875 at least one root. A project can also have multiple roots, though
2876 that isn't common (or necessarily a good idea).
2877 - an author: The name of the person responsible for this change, together
2878 with its date.
2879 - a committer: The name of the person who actually created the commit,
2880 with the date it was done. This may be different from the author, for
2881 example, if the author was someone who wrote a patch and emailed it
2882 to the person who used it to create the commit.
2883 - a comment describing this commit.
2884
2885 Note that a commit does not itself contain any information about what
2886 actually changed; all changes are calculated by comparing the contents
2887 of the tree referred to by this commit with the trees associated with
2888 its parents. In particular, git does not attempt to record file renames
2889 explicitly, though it can identify cases where the existence of the same
2890 file data at changing paths suggests a rename. (See, for example, the
2891 -M option to gitlink:git-diff[1]).
2892
2893 A commit is usually created by gitlink:git-commit[1], which creates a
2894 commit whose parent is normally the current HEAD, and whose tree is
2895 taken from the content currently stored in the index.
2896
2897 [[tree-object]]
2898 Tree Object
2899 ~~~~~~~~~~~
2900
2901 The ever-versatile gitlink:git-show[1] command can also be used to
2902 examine tree objects, but gitlink:git-ls-tree[1] will give you more
2903 details:
2904
2905 ------------------------------------------------
2906 $ git ls-tree fb3a8bdd0ce
2907 100644 blob 63c918c667fa005ff12ad89437f2fdc80926e21c .gitignore
2908 100644 blob 5529b198e8d14decbe4ad99db3f7fb632de0439d .mailmap
2909 100644 blob 6ff87c4664981e4397625791c8ea3bbb5f2279a3 COPYING
2910 040000 tree 2fb783e477100ce076f6bf57e4a6f026013dc745 Documentation
2911 100755 blob 3c0032cec592a765692234f1cba47dfdcc3a9200 GIT-VERSION-GEN
2912 100644 blob 289b046a443c0647624607d471289b2c7dcd470b INSTALL
2913 100644 blob 4eb463797adc693dc168b926b6932ff53f17d0b1 Makefile
2914 100644 blob 548142c327a6790ff8821d67c2ee1eff7a656b52 README
2915 ...
2916 ------------------------------------------------
2917
2918 As you can see, a tree object contains a list of entries, each with a
2919 mode, object type, SHA1 name, and name, sorted by name. It represents
2920 the contents of a single directory tree.
2921
2922 The object type may be a blob, representing the contents of a file, or
2923 another tree, representing the contents of a subdirectory. Since trees
2924 and blobs, like all other objects, are named by the SHA1 hash of their
2925 contents, two trees have the same SHA1 name if and only if their
2926 contents (including, recursively, the contents of all subdirectories)
2927 are identical. This allows git to quickly determine the differences
2928 between two related tree objects, since it can ignore any entries with
2929 identical object names.
2930
2931 (Note: in the presence of submodules, trees may also have commits as
2932 entries. See <<submodules>> for documentation.)
2933
2934 Note that the files all have mode 644 or 755: git actually only pays
2935 attention to the executable bit.
2936
2937 [[blob-object]]
2938 Blob Object
2939 ~~~~~~~~~~~
2940
2941 You can use gitlink:git-show[1] to examine the contents of a blob; take,
2942 for example, the blob in the entry for "COPYING" from the tree above:
2943
2944 ------------------------------------------------
2945 $ git show 6ff87c4664
2946
2947 Note that the only valid version of the GPL as far as this project
2948 is concerned is _this_ particular version of the license (ie v2, not
2949 v2.2 or v3.x or whatever), unless explicitly otherwise stated.
2950 ...
2951 ------------------------------------------------
2952
2953 A "blob" object is nothing but a binary blob of data. It doesn't refer
2954 to anything else or have attributes of any kind.
2955
2956 Since the blob is entirely defined by its data, if two files in a
2957 directory tree (or in multiple different versions of the repository)
2958 have the same contents, they will share the same blob object. The object
2959 is totally independent of its location in the directory tree, and
2960 renaming a file does not change the object that file is associated with.
2961
2962 Note that any tree or blob object can be examined using
2963 gitlink:git-show[1] with the <revision>:<path> syntax. This can
2964 sometimes be useful for browsing the contents of a tree that is not
2965 currently checked out.
2966
2967 [[trust]]
2968 Trust
2969 ~~~~~
2970
2971 If you receive the SHA1 name of a blob from one source, and its contents
2972 from another (possibly untrusted) source, you can still trust that those
2973 contents are correct as long as the SHA1 name agrees. This is because
2974 the SHA1 is designed so that it is infeasible to find different contents
2975 that produce the same hash.
2976
2977 Similarly, you need only trust the SHA1 name of a top-level tree object
2978 to trust the contents of the entire directory that it refers to, and if
2979 you receive the SHA1 name of a commit from a trusted source, then you
2980 can easily verify the entire history of commits reachable through
2981 parents of that commit, and all of those contents of the trees referred
2982 to by those commits.
2983
2984 So to introduce some real trust in the system, the only thing you need
2985 to do is to digitally sign just 'one' special note, which includes the
2986 name of a top-level commit. Your digital signature shows others
2987 that you trust that commit, and the immutability of the history of
2988 commits tells others that they can trust the whole history.
2989
2990 In other words, you can easily validate a whole archive by just
2991 sending out a single email that tells the people the name (SHA1 hash)
2992 of the top commit, and digitally sign that email using something
2993 like GPG/PGP.
2994
2995 To assist in this, git also provides the tag object...
2996
2997 [[tag-object]]
2998 Tag Object
2999 ~~~~~~~~~~
3000
3001 A tag object contains an object, object type, tag name, the name of the
3002 person ("tagger") who created the tag, and a message, which may contain
3003 a signature, as can be seen using the gitlink:git-cat-file[1]:
3004
3005 ------------------------------------------------
3006 $ git cat-file tag v1.5.0
3007 object 437b1b20df4b356c9342dac8d38849f24ef44f27
3008 type commit
3009 tag v1.5.0
3010 tagger Junio C Hamano <junkio@cox.net> 1171411200 +0000
3011
3012 GIT 1.5.0
3013 -----BEGIN PGP SIGNATURE-----
3014 Version: GnuPG v1.4.6 (GNU/Linux)
3015
3016 iD8DBQBF0lGqwMbZpPMRm5oRAuRiAJ9ohBLd7s2kqjkKlq1qqC57SbnmzQCdG4ui
3017 nLE/L9aUXdWeTFPron96DLA=
3018 =2E+0
3019 -----END PGP SIGNATURE-----
3020 ------------------------------------------------
3021
3022 See the gitlink:git-tag[1] command to learn how to create and verify tag
3023 objects. (Note that gitlink:git-tag[1] can also be used to create
3024 "lightweight tags", which are not tag objects at all, but just simple
3025 references whose names begin with "refs/tags/").
3026
3027 [[pack-files]]
3028 How git stores objects efficiently: pack files
3029 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3030
3031 Newly created objects are initially created in a file named after the
3032 object's SHA1 hash (stored in .git/objects).
3033
3034 Unfortunately this system becomes inefficient once a project has a
3035 lot of objects. Try this on an old project:
3036
3037 ------------------------------------------------
3038 $ git count-objects
3039 6930 objects, 47620 kilobytes
3040 ------------------------------------------------
3041
3042 The first number is the number of objects which are kept in
3043 individual files. The second is the amount of space taken up by
3044 those "loose" objects.
3045
3046 You can save space and make git faster by moving these loose objects in
3047 to a "pack file", which stores a group of objects in an efficient
3048 compressed format; the details of how pack files are formatted can be
3049 found in link:technical/pack-format.txt[technical/pack-format.txt].
3050
3051 To put the loose objects into a pack, just run git repack:
3052
3053 ------------------------------------------------
3054 $ git repack
3055 Generating pack...
3056 Done counting 6020 objects.
3057 Deltifying 6020 objects.
3058 100% (6020/6020) done
3059 Writing 6020 objects.
3060 100% (6020/6020) done
3061 Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
3062 Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
3063 ------------------------------------------------
3064
3065 You can then run
3066
3067 ------------------------------------------------
3068 $ git prune
3069 ------------------------------------------------
3070
3071 to remove any of the "loose" objects that are now contained in the
3072 pack. This will also remove any unreferenced objects (which may be
3073 created when, for example, you use "git reset" to remove a commit).
3074 You can verify that the loose objects are gone by looking at the
3075 .git/objects directory or by running
3076
3077 ------------------------------------------------
3078 $ git count-objects
3079 0 objects, 0 kilobytes
3080 ------------------------------------------------
3081
3082 Although the object files are gone, any commands that refer to those
3083 objects will work exactly as they did before.
3084
3085 The gitlink:git-gc[1] command performs packing, pruning, and more for
3086 you, so is normally the only high-level command you need.
3087
3088 [[dangling-objects]]
3089 Dangling objects
3090 ~~~~~~~~~~~~~~~~
3091
3092 The gitlink:git-fsck[1] command will sometimes complain about dangling
3093 objects. They are not a problem.
3094
3095 The most common cause of dangling objects is that you've rebased a
3096 branch, or you have pulled from somebody else who rebased a branch--see
3097 <<cleaning-up-history>>. In that case, the old head of the original
3098 branch still exists, as does everything it pointed to. The branch
3099 pointer itself just doesn't, since you replaced it with another one.
3100
3101 There are also other situations that cause dangling objects. For
3102 example, a "dangling blob" may arise because you did a "git add" of a
3103 file, but then, before you actually committed it and made it part of the
3104 bigger picture, you changed something else in that file and committed
3105 that *updated* thing--the old state that you added originally ends up
3106 not being pointed to by any commit or tree, so it's now a dangling blob
3107 object.
3108
3109 Similarly, when the "recursive" merge strategy runs, and finds that
3110 there are criss-cross merges and thus more than one merge base (which is
3111 fairly unusual, but it does happen), it will generate one temporary
3112 midway tree (or possibly even more, if you had lots of criss-crossing
3113 merges and more than two merge bases) as a temporary internal merge
3114 base, and again, those are real objects, but the end result will not end
3115 up pointing to them, so they end up "dangling" in your repository.
3116
3117 Generally, dangling objects aren't anything to worry about. They can
3118 even be very useful: if you screw something up, the dangling objects can
3119 be how you recover your old tree (say, you did a rebase, and realized
3120 that you really didn't want to--you can look at what dangling objects
3121 you have, and decide to reset your head to some old dangling state).
3122
3123 For commits, you can just use:
3124
3125 ------------------------------------------------
3126 $ gitk <dangling-commit-sha-goes-here> --not --all
3127 ------------------------------------------------
3128
3129 This asks for all the history reachable from the given commit but not
3130 from any branch, tag, or other reference. If you decide it's something
3131 you want, you can always create a new reference to it, e.g.,
3132
3133 ------------------------------------------------
3134 $ git branch recovered-branch <dangling-commit-sha-goes-here>
3135 ------------------------------------------------
3136
3137 For blobs and trees, you can't do the same, but you can still examine
3138 them. You can just do
3139
3140 ------------------------------------------------
3141 $ git show <dangling-blob/tree-sha-goes-here>
3142 ------------------------------------------------
3143
3144 to show what the contents of the blob were (or, for a tree, basically
3145 what the "ls" for that directory was), and that may give you some idea
3146 of what the operation was that left that dangling object.
3147
3148 Usually, dangling blobs and trees aren't very interesting. They're
3149 almost always the result of either being a half-way mergebase (the blob
3150 will often even have the conflict markers from a merge in it, if you
3151 have had conflicting merges that you fixed up by hand), or simply
3152 because you interrupted a "git fetch" with ^C or something like that,
3153 leaving _some_ of the new objects in the object database, but just
3154 dangling and useless.
3155
3156 Anyway, once you are sure that you're not interested in any dangling
3157 state, you can just prune all unreachable objects:
3158
3159 ------------------------------------------------
3160 $ git prune
3161 ------------------------------------------------
3162
3163 and they'll be gone. But you should only run "git prune" on a quiescent
3164 repository--it's kind of like doing a filesystem fsck recovery: you
3165 don't want to do that while the filesystem is mounted.
3166
3167 (The same is true of "git-fsck" itself, btw, but since
3168 git-fsck never actually *changes* the repository, it just reports
3169 on what it found, git-fsck itself is never "dangerous" to run.
3170 Running it while somebody is actually changing the repository can cause
3171 confusing and scary messages, but it won't actually do anything bad. In
3172 contrast, running "git prune" while somebody is actively changing the
3173 repository is a *BAD* idea).
3174
3175 [[the-index]]
3176 The index
3177 -----------
3178
3179 The index is a binary file (generally kept in .git/index) containing a
3180 sorted list of path names, each with permissions and the SHA1 of a blob
3181 object; gitlink:git-ls-files[1] can show you the contents of the index:
3182
3183 -------------------------------------------------
3184 $ git ls-files --stage
3185 100644 63c918c667fa005ff12ad89437f2fdc80926e21c 0 .gitignore
3186 100644 5529b198e8d14decbe4ad99db3f7fb632de0439d 0 .mailmap
3187 100644 6ff87c4664981e4397625791c8ea3bbb5f2279a3 0 COPYING
3188 100644 a37b2152bd26be2c2289e1f57a292534a51a93c7 0 Documentation/.gitignore
3189 100644 fbefe9a45b00a54b58d94d06eca48b03d40a50e0 0 Documentation/Makefile
3190 ...
3191 100644 2511aef8d89ab52be5ec6a5e46236b4b6bcd07ea 0 xdiff/xtypes.h
3192 100644 2ade97b2574a9f77e7ae4002a4e07a6a38e46d07 0 xdiff/xutils.c
3193 100644 d5de8292e05e7c36c4b68857c1cf9855e3d2f70a 0 xdiff/xutils.h
3194 -------------------------------------------------
3195
3196 Note that in older documentation you may see the index called the
3197 "current directory cache" or just the "cache". It has three important
3198 properties:
3199
3200 1. The index contains all the information necessary to generate a single
3201 (uniquely determined) tree object.
3202 +
3203 For example, running gitlink:git-commit[1] generates this tree object
3204 from the index, stores it in the object database, and uses it as the
3205 tree object associated with the new commit.
3206
3207 2. The index enables fast comparisons between the tree object it defines
3208 and the working tree.
3209 +
3210 It does this by storing some additional data for each entry (such as
3211 the last modified time). This data is not displayed above, and is not
3212 stored in the created tree object, but it can be used to determine
3213 quickly which files in the working directory differ from what was
3214 stored in the index, and thus save git from having to read all of the
3215 data from such files to look for changes.
3216
3217 3. It can efficiently represent information about merge conflicts
3218 between different tree objects, allowing each pathname to be
3219 associated with sufficient information about the trees involved that
3220 you can create a three-way merge between them.
3221 +
3222 We saw in <<conflict-resolution>> that during a merge the index can
3223 store multiple versions of a single file (called "stages"). The third
3224 column in the gitlink:git-ls-files[1] output above is the stage
3225 number, and will take on values other than 0 for files with merge
3226 conflicts.
3227
3228 The index is thus a sort of temporary staging area, which is filled with
3229 a tree which you are in the process of working on.
3230
3231 If you blow the index away entirely, you generally haven't lost any
3232 information as long as you have the name of the tree that it described.
3233
3234 [[submodules]]
3235 Submodules
3236 ==========
3237
3238 Large projects are often composed of smaller, self-contained modules. For
3239 example, an embedded Linux distribution's source tree would include every
3240 piece of software in the distribution with some local modifications; a movie
3241 player might need to build against a specific, known-working version of a
3242 decompression library; several independent programs might all share the same
3243 build scripts.
3244
3245 With centralized revision control systems this is often accomplished by
3246 including every module in one single repository. Developers can check out
3247 all modules or only the modules they need to work with. They can even modify
3248 files across several modules in a single commit while moving things around
3249 or updating APIs and translations.
3250
3251 Git does not allow partial checkouts, so duplicating this approach in Git
3252 would force developers to keep a local copy of modules they are not
3253 interested in touching. Commits in an enormous checkout would be slower
3254 than you'd expect as Git would have to scan every directory for changes.
3255 If modules have a lot of local history, clones would take forever.
3256
3257 On the plus side, distributed revision control systems can much better
3258 integrate with external sources. In a centralized model, a single arbitrary
3259 snapshot of the external project is exported from its own revision control
3260 and then imported into the local revision control on a vendor branch. All
3261 the history is hidden. With distributed revision control you can clone the
3262 entire external history and much more easily follow development and re-merge
3263 local changes.
3264
3265 Git's submodule support allows a repository to contain, as a subdirectory, a
3266 checkout of an external project. Submodules maintain their own identity;
3267 the submodule support just stores the submodule repository location and
3268 commit ID, so other developers who clone the containing project
3269 ("superproject") can easily clone all the submodules at the same revision.
3270 Partial checkouts of the superproject are possible: you can tell Git to
3271 clone none, some or all of the submodules.
3272
3273 The gitlink:git-submodule[1] command is available since Git 1.5.3. Users
3274 with Git 1.5.2 can look up the submodule commits in the repository and
3275 manually check them out; earlier versions won't recognize the submodules at
3276 all.
3277
3278 To see how submodule support works, create (for example) four example
3279 repositories that can be used later as a submodule:
3280
3281 -------------------------------------------------
3282 $ mkdir ~/git
3283 $ cd ~/git
3284 $ for i in a b c d
3285 do
3286 mkdir $i
3287 cd $i
3288 git init
3289 echo "module $i" > $i.txt
3290 git add $i.txt
3291 git commit -m "Initial commit, submodule $i"
3292 cd ..
3293 done
3294 -------------------------------------------------
3295
3296 Now create the superproject and add all the submodules:
3297
3298 -------------------------------------------------
3299 $ mkdir super
3300 $ cd super
3301 $ git init
3302 $ for i in a b c d
3303 do
3304 git submodule add ~/git/$i
3305 done
3306 -------------------------------------------------
3307
3308 NOTE: Do not use local URLs here if you plan to publish your superproject!
3309
3310 See what files `git submodule` created:
3311
3312 -------------------------------------------------
3313 $ ls -a
3314 . .. .git .gitmodules a b c d
3315 -------------------------------------------------
3316
3317 The `git submodule add` command does a couple of things:
3318
3319 - It clones the submodule under the current directory and by default checks out
3320 the master branch.
3321 - It adds the submodule's clone path to the gitlink:gitmodules[5] file and
3322 adds this file to the index, ready to be committed.
3323 - It adds the submodule's current commit ID to the index, ready to be
3324 committed.
3325
3326 Commit the superproject:
3327
3328 -------------------------------------------------
3329 $ git commit -m "Add submodules a, b, c and d."
3330 -------------------------------------------------
3331
3332 Now clone the superproject:
3333
3334 -------------------------------------------------
3335 $ cd ..
3336 $ git clone super cloned
3337 $ cd cloned
3338 -------------------------------------------------
3339
3340 The submodule directories are there, but they're empty:
3341
3342 -------------------------------------------------
3343 $ ls -a a
3344 . ..
3345 $ git submodule status
3346 -d266b9873ad50488163457f025db7cdd9683d88b a
3347 -e81d457da15309b4fef4249aba9b50187999670d b
3348 -c1536a972b9affea0f16e0680ba87332dc059146 c
3349 -d96249ff5d57de5de093e6baff9e0aafa5276a74 d
3350 -------------------------------------------------
3351
3352 NOTE: The commit object names shown above would be different for you, but they
3353 should match the HEAD commit object names of your repositories. You can check
3354 it by running `git ls-remote ../a`.
3355
3356 Pulling down the submodules is a two-step process. First run `git submodule
3357 init` to add the submodule repository URLs to `.git/config`:
3358
3359 -------------------------------------------------
3360 $ git submodule init
3361 -------------------------------------------------
3362
3363 Now use `git submodule update` to clone the repositories and check out the
3364 commits specified in the superproject:
3365
3366 -------------------------------------------------
3367 $ git submodule update
3368 $ cd a
3369 $ ls -a
3370 . .. .git a.txt
3371 -------------------------------------------------
3372
3373 One major difference between `git submodule update` and `git submodule add` is
3374 that `git submodule update` checks out a specific commit, rather than the tip
3375 of a branch. It's like checking out a tag: the head is detached, so you're not
3376 working on a branch.
3377
3378 -------------------------------------------------
3379 $ git branch
3380 * (no branch)
3381 master
3382 -------------------------------------------------
3383
3384 If you want to make a change within a submodule and you have a detached head,
3385 then you should create or checkout a branch, make your changes, publish the
3386 change within the submodule, and then update the superproject to reference the
3387 new commit:
3388
3389 -------------------------------------------------
3390 $ git checkout master
3391 -------------------------------------------------
3392
3393 or
3394
3395 -------------------------------------------------
3396 $ git checkout -b fix-up
3397 -------------------------------------------------
3398
3399 then
3400
3401 -------------------------------------------------
3402 $ echo "adding a line again" >> a.txt
3403 $ git commit -a -m "Updated the submodule from within the superproject."
3404 $ git push
3405 $ cd ..
3406 $ git diff
3407 diff --git a/a b/a
3408 index d266b98..261dfac 160000
3409 --- a/a
3410 +++ b/a
3411 @@ -1 +1 @@
3412 -Subproject commit d266b9873ad50488163457f025db7cdd9683d88b
3413 +Subproject commit 261dfac35cb99d380eb966e102c1197139f7fa24
3414 $ git add a
3415 $ git commit -m "Updated submodule a."
3416 $ git push
3417 -------------------------------------------------
3418
3419 You have to run `git submodule update` after `git pull` if you want to update
3420 submodules, too.
3421
3422 Pitfalls with submodules
3423 ------------------------
3424
3425 Always publish the submodule change before publishing the change to the
3426 superproject that references it. If you forget to publish the submodule change,
3427 others won't be able to clone the repository:
3428
3429 -------------------------------------------------
3430 $ cd ~/git/super/a
3431 $ echo i added another line to this file >> a.txt
3432 $ git commit -a -m "doing it wrong this time"
3433 $ cd ..
3434 $ git add a
3435 $ git commit -m "Updated submodule a again."
3436 $ git push
3437 $ cd ~/git/cloned
3438 $ git pull
3439 $ git submodule update
3440 error: pathspec '261dfac35cb99d380eb966e102c1197139f7fa24' did not match any file(s) known to git.
3441 Did you forget to 'git add'?
3442 Unable to checkout '261dfac35cb99d380eb966e102c1197139f7fa24' in submodule path 'a'
3443 -------------------------------------------------
3444
3445 You also should not rewind branches in a submodule beyond commits that were
3446 ever recorded in any superproject.
3447
3448 It's not safe to run `git submodule update` if you've made and committed
3449 changes within a submodule without checking out a branch first. They will be
3450 silently overwritten:
3451
3452 -------------------------------------------------
3453 $ cat a.txt
3454 module a
3455 $ echo line added from private2 >> a.txt
3456 $ git commit -a -m "line added inside private2"
3457 $ cd ..
3458 $ git submodule update
3459 Submodule path 'a': checked out 'd266b9873ad50488163457f025db7cdd9683d88b'
3460 $ cd a
3461 $ cat a.txt
3462 module a
3463 -------------------------------------------------
3464
3465 NOTE: The changes are still visible in the submodule's reflog.
3466
3467 This is not the case if you did not commit your changes.
3468
3469 [[low-level-operations]]
3470 Low-level git operations
3471 ========================
3472
3473 Many of the higher-level commands were originally implemented as shell
3474 scripts using a smaller core of low-level git commands. These can still
3475 be useful when doing unusual things with git, or just as a way to
3476 understand its inner workings.
3477
3478 [[object-manipulation]]
3479 Object access and manipulation
3480 ------------------------------
3481
3482 The gitlink:git-cat-file[1] command can show the contents of any object,
3483 though the higher-level gitlink:git-show[1] is usually more useful.
3484
3485 The gitlink:git-commit-tree[1] command allows constructing commits with
3486 arbitrary parents and trees.
3487
3488 A tree can be created with gitlink:git-write-tree[1] and its data can be
3489 accessed by gitlink:git-ls-tree[1]. Two trees can be compared with
3490 gitlink:git-diff-tree[1].
3491
3492 A tag is created with gitlink:git-mktag[1], and the signature can be
3493 verified by gitlink:git-verify-tag[1], though it is normally simpler to
3494 use gitlink:git-tag[1] for both.
3495
3496 [[the-workflow]]
3497 The Workflow
3498 ------------
3499
3500 High-level operations such as gitlink:git-commit[1],
3501 gitlink:git-checkout[1] and gitlink:git-reset[1] work by moving data
3502 between the working tree, the index, and the object database. Git
3503 provides low-level operations which perform each of these steps
3504 individually.
3505
3506 Generally, all "git" operations work on the index file. Some operations
3507 work *purely* on the index file (showing the current state of the
3508 index), but most operations move data between the index file and either
3509 the database or the working directory. Thus there are four main
3510 combinations:
3511
3512 [[working-directory-to-index]]
3513 working directory -> index
3514 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3515
3516 The gitlink:git-update-index[1] command updates the index with
3517 information from the working directory. You generally update the
3518 index information by just specifying the filename you want to update,
3519 like so:
3520
3521 -------------------------------------------------
3522 $ git update-index filename
3523 -------------------------------------------------
3524
3525 but to avoid common mistakes with filename globbing etc, the command
3526 will not normally add totally new entries or remove old entries,
3527 i.e. it will normally just update existing cache entries.
3528
3529 To tell git that yes, you really do realize that certain files no
3530 longer exist, or that new files should be added, you
3531 should use the `--remove` and `--add` flags respectively.
3532
3533 NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
3534 necessarily be removed: if the files still exist in your directory
3535 structure, the index will be updated with their new status, not
3536 removed. The only thing `--remove` means is that update-cache will be
3537 considering a removed file to be a valid thing, and if the file really
3538 does not exist any more, it will update the index accordingly.
3539
3540 As a special case, you can also do `git-update-index --refresh`, which
3541 will refresh the "stat" information of each index to match the current
3542 stat information. It will 'not' update the object status itself, and
3543 it will only update the fields that are used to quickly test whether
3544 an object still matches its old backing store object.
3545
3546 The previously introduced gitlink:git-add[1] is just a wrapper for
3547 gitlink:git-update-index[1].
3548
3549 [[index-to-object-database]]
3550 index -> object database
3551 ~~~~~~~~~~~~~~~~~~~~~~~~
3552
3553 You write your current index file to a "tree" object with the program
3554
3555 -------------------------------------------------
3556 $ git write-tree
3557 -------------------------------------------------
3558
3559 that doesn't come with any options--it will just write out the
3560 current index into the set of tree objects that describe that state,
3561 and it will return the name of the resulting top-level tree. You can
3562 use that tree to re-generate the index at any time by going in the
3563 other direction:
3564
3565 [[object-database-to-index]]
3566 object database -> index
3567 ~~~~~~~~~~~~~~~~~~~~~~~~
3568
3569 You read a "tree" file from the object database, and use that to
3570 populate (and overwrite--don't do this if your index contains any
3571 unsaved state that you might want to restore later!) your current
3572 index. Normal operation is just
3573
3574 -------------------------------------------------
3575 $ git-read-tree <sha1 of tree>
3576 -------------------------------------------------
3577
3578 and your index file will now be equivalent to the tree that you saved
3579 earlier. However, that is only your 'index' file: your working
3580 directory contents have not been modified.
3581
3582 [[index-to-working-directory]]
3583 index -> working directory
3584 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3585
3586 You update your working directory from the index by "checking out"
3587 files. This is not a very common operation, since normally you'd just
3588 keep your files updated, and rather than write to your working
3589 directory, you'd tell the index files about the changes in your
3590 working directory (i.e. `git-update-index`).
3591
3592 However, if you decide to jump to a new version, or check out somebody
3593 else's version, or just restore a previous tree, you'd populate your
3594 index file with read-tree, and then you need to check out the result
3595 with
3596
3597 -------------------------------------------------
3598 $ git-checkout-index filename
3599 -------------------------------------------------
3600
3601 or, if you want to check out all of the index, use `-a`.
3602
3603 NOTE! git-checkout-index normally refuses to overwrite old files, so
3604 if you have an old version of the tree already checked out, you will
3605 need to use the "-f" flag ('before' the "-a" flag or the filename) to
3606 'force' the checkout.
3607
3608
3609 Finally, there are a few odds and ends which are not purely moving
3610 from one representation to the other:
3611
3612 [[tying-it-all-together]]
3613 Tying it all together
3614 ~~~~~~~~~~~~~~~~~~~~~
3615
3616 To commit a tree you have instantiated with "git-write-tree", you'd
3617 create a "commit" object that refers to that tree and the history
3618 behind it--most notably the "parent" commits that preceded it in
3619 history.
3620
3621 Normally a "commit" has one parent: the previous state of the tree
3622 before a certain change was made. However, sometimes it can have two
3623 or more parent commits, in which case we call it a "merge", due to the
3624 fact that such a commit brings together ("merges") two or more
3625 previous states represented by other commits.
3626
3627 In other words, while a "tree" represents a particular directory state
3628 of a working directory, a "commit" represents that state in "time",
3629 and explains how we got there.
3630
3631 You create a commit object by giving it the tree that describes the
3632 state at the time of the commit, and a list of parents:
3633
3634 -------------------------------------------------
3635 $ git-commit-tree <tree> -p <parent> [-p <parent2> ..]
3636 -------------------------------------------------
3637
3638 and then giving the reason for the commit on stdin (either through
3639 redirection from a pipe or file, or by just typing it at the tty).
3640
3641 git-commit-tree will return the name of the object that represents
3642 that commit, and you should save it away for later use. Normally,
3643 you'd commit a new `HEAD` state, and while git doesn't care where you
3644 save the note about that state, in practice we tend to just write the
3645 result to the file pointed at by `.git/HEAD`, so that we can always see
3646 what the last committed state was.
3647
3648 Here is an ASCII art by Jon Loeliger that illustrates how
3649 various pieces fit together.
3650
3651 ------------
3652
3653 commit-tree
3654 commit obj
3655 +----+
3656 | |
3657 | |
3658 V V
3659 +-----------+
3660 | Object DB |
3661 | Backing |
3662 | Store |
3663 +-----------+
3664 ^
3665 write-tree | |
3666 tree obj | |
3667 | | read-tree
3668 | | tree obj
3669 V
3670 +-----------+
3671 | Index |
3672 | "cache" |
3673 +-----------+
3674 update-index ^
3675 blob obj | |
3676 | |
3677 checkout-index -u | | checkout-index
3678 stat | | blob obj
3679 V
3680 +-----------+
3681 | Working |
3682 | Directory |
3683 +-----------+
3684
3685 ------------
3686
3687
3688 [[examining-the-data]]
3689 Examining the data
3690 ------------------
3691
3692 You can examine the data represented in the object database and the
3693 index with various helper tools. For every object, you can use
3694 gitlink:git-cat-file[1] to examine details about the
3695 object:
3696
3697 -------------------------------------------------
3698 $ git-cat-file -t <objectname>
3699 -------------------------------------------------
3700
3701 shows the type of the object, and once you have the type (which is
3702 usually implicit in where you find the object), you can use
3703
3704 -------------------------------------------------
3705 $ git-cat-file blob|tree|commit|tag <objectname>
3706 -------------------------------------------------
3707
3708 to show its contents. NOTE! Trees have binary content, and as a result
3709 there is a special helper for showing that content, called
3710 `git-ls-tree`, which turns the binary content into a more easily
3711 readable form.
3712
3713 It's especially instructive to look at "commit" objects, since those
3714 tend to be small and fairly self-explanatory. In particular, if you
3715 follow the convention of having the top commit name in `.git/HEAD`,
3716 you can do
3717
3718 -------------------------------------------------
3719 $ git-cat-file commit HEAD
3720 -------------------------------------------------
3721
3722 to see what the top commit was.
3723
3724 [[merging-multiple-trees]]
3725 Merging multiple trees
3726 ----------------------
3727
3728 Git helps you do a three-way merge, which you can expand to n-way by
3729 repeating the merge procedure arbitrary times until you finally
3730 "commit" the state. The normal situation is that you'd only do one
3731 three-way merge (two parents), and commit it, but if you like to, you
3732 can do multiple parents in one go.
3733
3734 To do a three-way merge, you need the two sets of "commit" objects
3735 that you want to merge, use those to find the closest common parent (a
3736 third "commit" object), and then use those commit objects to find the
3737 state of the directory ("tree" object) at these points.
3738
3739 To get the "base" for the merge, you first look up the common parent
3740 of two commits with
3741
3742 -------------------------------------------------
3743 $ git-merge-base <commit1> <commit2>
3744 -------------------------------------------------
3745
3746 which will return you the commit they are both based on. You should
3747 now look up the "tree" objects of those commits, which you can easily
3748 do with (for example)
3749
3750 -------------------------------------------------
3751 $ git-cat-file commit <commitname> | head -1
3752 -------------------------------------------------
3753
3754 since the tree object information is always the first line in a commit
3755 object.
3756
3757 Once you know the three trees you are going to merge (the one "original"
3758 tree, aka the common tree, and the two "result" trees, aka the branches
3759 you want to merge), you do a "merge" read into the index. This will
3760 complain if it has to throw away your old index contents, so you should
3761 make sure that you've committed those--in fact you would normally
3762 always do a merge against your last commit (which should thus match what
3763 you have in your current index anyway).
3764
3765 To do the merge, do
3766
3767 -------------------------------------------------
3768 $ git-read-tree -m -u <origtree> <yourtree> <targettree>
3769 -------------------------------------------------
3770
3771 which will do all trivial merge operations for you directly in the
3772 index file, and you can just write the result out with
3773 `git-write-tree`.
3774
3775
3776 [[merging-multiple-trees-2]]
3777 Merging multiple trees, continued
3778 ---------------------------------
3779
3780 Sadly, many merges aren't trivial. If there are files that have
3781 been added, moved or removed, or if both branches have modified the
3782 same file, you will be left with an index tree that contains "merge
3783 entries" in it. Such an index tree can 'NOT' be written out to a tree
3784 object, and you will have to resolve any such merge clashes using
3785 other tools before you can write out the result.
3786
3787 You can examine such index state with `git-ls-files --unmerged`
3788 command. An example:
3789
3790 ------------------------------------------------
3791 $ git-read-tree -m $orig HEAD $target
3792 $ git-ls-files --unmerged
3793 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
3794 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
3795 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
3796 ------------------------------------------------
3797
3798 Each line of the `git-ls-files --unmerged` output begins with
3799 the blob mode bits, blob SHA1, 'stage number', and the
3800 filename. The 'stage number' is git's way to say which tree it
3801 came from: stage 1 corresponds to `$orig` tree, stage 2 `HEAD`
3802 tree, and stage3 `$target` tree.
3803
3804 Earlier we said that trivial merges are done inside
3805 `git-read-tree -m`. For example, if the file did not change
3806 from `$orig` to `HEAD` nor `$target`, or if the file changed
3807 from `$orig` to `HEAD` and `$orig` to `$target` the same way,
3808 obviously the final outcome is what is in `HEAD`. What the
3809 above example shows is that file `hello.c` was changed from
3810 `$orig` to `HEAD` and `$orig` to `$target` in a different way.
3811 You could resolve this by running your favorite 3-way merge
3812 program, e.g. `diff3`, `merge`, or git's own merge-file, on
3813 the blob objects from these three stages yourself, like this:
3814
3815 ------------------------------------------------
3816 $ git-cat-file blob 263414f... >hello.c~1
3817 $ git-cat-file blob 06fa6a2... >hello.c~2
3818 $ git-cat-file blob cc44c73... >hello.c~3
3819 $ git merge-file hello.c~2 hello.c~1 hello.c~3
3820 ------------------------------------------------
3821
3822 This would leave the merge result in `hello.c~2` file, along
3823 with conflict markers if there are conflicts. After verifying
3824 the merge result makes sense, you can tell git what the final
3825 merge result for this file is by:
3826
3827 -------------------------------------------------
3828 $ mv -f hello.c~2 hello.c
3829 $ git-update-index hello.c
3830 -------------------------------------------------
3831
3832 When a path is in unmerged state, running `git-update-index` for
3833 that path tells git to mark the path resolved.
3834
3835 The above is the description of a git merge at the lowest level,
3836 to help you understand what conceptually happens under the hood.
3837 In practice, nobody, not even git itself, uses three `git-cat-file`
3838 for this. There is `git-merge-index` program that extracts the
3839 stages to temporary files and calls a "merge" script on it:
3840
3841 -------------------------------------------------
3842 $ git-merge-index git-merge-one-file hello.c
3843 -------------------------------------------------
3844
3845 and that is what higher level `git merge -s resolve` is implemented with.
3846
3847 [[hacking-git]]
3848 Hacking git
3849 ===========
3850
3851 This chapter covers internal details of the git implementation which
3852 probably only git developers need to understand.
3853
3854 [[object-details]]
3855 Object storage format
3856 ---------------------
3857
3858 All objects have a statically determined "type" which identifies the
3859 format of the object (i.e. how it is used, and how it can refer to other
3860 objects). There are currently four different object types: "blob",
3861 "tree", "commit", and "tag".
3862
3863 Regardless of object type, all objects share the following
3864 characteristics: they are all deflated with zlib, and have a header
3865 that not only specifies their type, but also provides size information
3866 about the data in the object. It's worth noting that the SHA1 hash
3867 that is used to name the object is the hash of the original data
3868 plus this header, so `sha1sum` 'file' does not match the object name
3869 for 'file'.
3870 (Historical note: in the dawn of the age of git the hash
3871 was the sha1 of the 'compressed' object.)
3872
3873 As a result, the general consistency of an object can always be tested
3874 independently of the contents or the type of the object: all objects can
3875 be validated by verifying that (a) their hashes match the content of the
3876 file and (b) the object successfully inflates to a stream of bytes that
3877 forms a sequence of <ascii type without space> {plus} <space> {plus} <ascii decimal
3878 size> {plus} <byte\0> {plus} <binary object data>.
3879
3880 The structured objects can further have their structure and
3881 connectivity to other objects verified. This is generally done with
3882 the `git-fsck` program, which generates a full dependency graph
3883 of all objects, and verifies their internal consistency (in addition
3884 to just verifying their superficial consistency through the hash).
3885
3886 [[birdview-on-the-source-code]]
3887 A birds-eye view of Git's source code
3888 -------------------------------------
3889
3890 It is not always easy for new developers to find their way through Git's
3891 source code. This section gives you a little guidance to show where to
3892 start.
3893
3894 A good place to start is with the contents of the initial commit, with:
3895
3896 ----------------------------------------------------
3897 $ git checkout e83c5163
3898 ----------------------------------------------------
3899
3900 The initial revision lays the foundation for almost everything git has
3901 today, but is small enough to read in one sitting.
3902
3903 Note that terminology has changed since that revision. For example, the
3904 README in that revision uses the word "changeset" to describe what we
3905 now call a <<def_commit_object,commit>>.
3906
3907 Also, we do not call it "cache" any more, but "index", however, the
3908 file is still called `cache.h`. Remark: Not much reason to change it now,
3909 especially since there is no good single name for it anyway, because it is
3910 basically _the_ header file which is included by _all_ of Git's C sources.
3911
3912 If you grasp the ideas in that initial commit, you should check out a
3913 more recent version and skim `cache.h`, `object.h` and `commit.h`.
3914
3915 In the early days, Git (in the tradition of UNIX) was a bunch of programs
3916 which were extremely simple, and which you used in scripts, piping the
3917 output of one into another. This turned out to be good for initial
3918 development, since it was easier to test new things. However, recently
3919 many of these parts have become builtins, and some of the core has been
3920 "libified", i.e. put into libgit.a for performance, portability reasons,
3921 and to avoid code duplication.
3922
3923 By now, you know what the index is (and find the corresponding data
3924 structures in `cache.h`), and that there are just a couple of object types
3925 (blobs, trees, commits and tags) which inherit their common structure from
3926 `struct object`, which is their first member (and thus, you can cast e.g.
3927 `(struct object *)commit` to achieve the _same_ as `&commit->object`, i.e.
3928 get at the object name and flags).
3929
3930 Now is a good point to take a break to let this information sink in.
3931
3932 Next step: get familiar with the object naming. Read <<naming-commits>>.
3933 There are quite a few ways to name an object (and not only revisions!).
3934 All of these are handled in `sha1_name.c`. Just have a quick look at
3935 the function `get_sha1()`. A lot of the special handling is done by
3936 functions like `get_sha1_basic()` or the likes.
3937
3938 This is just to get you into the groove for the most libified part of Git:
3939 the revision walker.
3940
3941 Basically, the initial version of `git log` was a shell script:
3942
3943 ----------------------------------------------------------------
3944 $ git-rev-list --pretty $(git-rev-parse --default HEAD "$@") | \
3945 LESS=-S ${PAGER:-less}
3946 ----------------------------------------------------------------
3947
3948 What does this mean?
3949
3950 `git-rev-list` is the original version of the revision walker, which
3951 _always_ printed a list of revisions to stdout. It is still functional,
3952 and needs to, since most new Git programs start out as scripts using
3953 `git-rev-list`.
3954
3955 `git-rev-parse` is not as important any more; it was only used to filter out
3956 options that were relevant for the different plumbing commands that were
3957 called by the script.
3958
3959 Most of what `git-rev-list` did is contained in `revision.c` and
3960 `revision.h`. It wraps the options in a struct named `rev_info`, which
3961 controls how and what revisions are walked, and more.
3962
3963 The original job of `git-rev-parse` is now taken by the function
3964 `setup_revisions()`, which parses the revisions and the common command line
3965 options for the revision walker. This information is stored in the struct
3966 `rev_info` for later consumption. You can do your own command line option
3967 parsing after calling `setup_revisions()`. After that, you have to call
3968 `prepare_revision_walk()` for initialization, and then you can get the
3969 commits one by one with the function `get_revision()`.
3970
3971 If you are interested in more details of the revision walking process,
3972 just have a look at the first implementation of `cmd_log()`; call
3973 `git-show v1.3.0~155^2~4` and scroll down to that function (note that you
3974 no longer need to call `setup_pager()` directly).
3975
3976 Nowadays, `git log` is a builtin, which means that it is _contained_ in the
3977 command `git`. The source side of a builtin is
3978
3979 - a function called `cmd_<bla>`, typically defined in `builtin-<bla>.c`,
3980 and declared in `builtin.h`,
3981
3982 - an entry in the `commands[]` array in `git.c`, and
3983
3984 - an entry in `BUILTIN_OBJECTS` in the `Makefile`.
3985
3986 Sometimes, more than one builtin is contained in one source file. For
3987 example, `cmd_whatchanged()` and `cmd_log()` both reside in `builtin-log.c`,
3988 since they share quite a bit of code. In that case, the commands which are
3989 _not_ named like the `.c` file in which they live have to be listed in
3990 `BUILT_INS` in the `Makefile`.
3991
3992 `git log` looks more complicated in C than it does in the original script,
3993 but that allows for a much greater flexibility and performance.
3994
3995 Here again it is a good point to take a pause.
3996
3997 Lesson three is: study the code. Really, it is the best way to learn about
3998 the organization of Git (after you know the basic concepts).
3999
4000 So, think about something which you are interested in, say, "how can I
4001 access a blob just knowing the object name of it?". The first step is to
4002 find a Git command with which you can do it. In this example, it is either
4003 `git show` or `git cat-file`.
4004
4005 For the sake of clarity, let's stay with `git cat-file`, because it
4006
4007 - is plumbing, and
4008
4009 - was around even in the initial commit (it literally went only through
4010 some 20 revisions as `cat-file.c`, was renamed to `builtin-cat-file.c`
4011 when made a builtin, and then saw less than 10 versions).
4012
4013 So, look into `builtin-cat-file.c`, search for `cmd_cat_file()` and look what
4014 it does.
4015
4016 ------------------------------------------------------------------
4017 git_config(git_default_config);
4018 if (argc != 3)
4019 usage("git-cat-file [-t|-s|-e|-p|<type>] <sha1>");
4020 if (get_sha1(argv[2], sha1))
4021 die("Not a valid object name %s", argv[2]);
4022 ------------------------------------------------------------------
4023
4024 Let's skip over the obvious details; the only really interesting part
4025 here is the call to `get_sha1()`. It tries to interpret `argv[2]` as an
4026 object name, and if it refers to an object which is present in the current
4027 repository, it writes the resulting SHA-1 into the variable `sha1`.
4028
4029 Two things are interesting here:
4030
4031 - `get_sha1()` returns 0 on _success_. This might surprise some new
4032 Git hackers, but there is a long tradition in UNIX to return different
4033 negative numbers in case of different errors--and 0 on success.
4034
4035 - the variable `sha1` in the function signature of `get_sha1()` is `unsigned
4036 char \*`, but is actually expected to be a pointer to `unsigned
4037 char[20]`. This variable will contain the 160-bit SHA-1 of the given
4038 commit. Note that whenever a SHA-1 is passed as `unsigned char \*`, it
4039 is the binary representation, as opposed to the ASCII representation in
4040 hex characters, which is passed as `char *`.
4041
4042 You will see both of these things throughout the code.
4043
4044 Now, for the meat:
4045
4046 -----------------------------------------------------------------------------
4047 case 0:
4048 buf = read_object_with_reference(sha1, argv[1], &size, NULL);
4049 -----------------------------------------------------------------------------
4050
4051 This is how you read a blob (actually, not only a blob, but any type of
4052 object). To know how the function `read_object_with_reference()` actually
4053 works, find the source code for it (something like `git grep
4054 read_object_with | grep ":[a-z]"` in the git repository), and read
4055 the source.
4056
4057 To find out how the result can be used, just read on in `cmd_cat_file()`:
4058
4059 -----------------------------------
4060 write_or_die(1, buf, size);
4061 -----------------------------------
4062
4063 Sometimes, you do not know where to look for a feature. In many such cases,
4064 it helps to search through the output of `git log`, and then `git show` the
4065 corresponding commit.
4066
4067 Example: If you know that there was some test case for `git bundle`, but
4068 do not remember where it was (yes, you _could_ `git grep bundle t/`, but that
4069 does not illustrate the point!):
4070
4071 ------------------------
4072 $ git log --no-merges t/
4073 ------------------------
4074
4075 In the pager (`less`), just search for "bundle", go a few lines back,
4076 and see that it is in commit 18449ab0... Now just copy this object name,
4077 and paste it into the command line
4078
4079 -------------------
4080 $ git show 18449ab0
4081 -------------------
4082
4083 Voila.
4084
4085 Another example: Find out what to do in order to make some script a
4086 builtin:
4087
4088 -------------------------------------------------
4089 $ git log --no-merges --diff-filter=A builtin-*.c
4090 -------------------------------------------------
4091
4092 You see, Git is actually the best tool to find out about the source of Git
4093 itself!
4094
4095 [[glossary]]
4096 include::glossary.txt[]
4097
4098 [[git-quick-start]]
4099 Appendix A: Git Quick Reference
4100 ===============================
4101
4102 This is a quick summary of the major commands; the previous chapters
4103 explain how these work in more detail.
4104
4105 [[quick-creating-a-new-repository]]
4106 Creating a new repository
4107 -------------------------
4108
4109 From a tarball:
4110
4111 -----------------------------------------------
4112 $ tar xzf project.tar.gz
4113 $ cd project
4114 $ git init
4115 Initialized empty Git repository in .git/
4116 $ git add .
4117 $ git commit
4118 -----------------------------------------------
4119
4120 From a remote repository:
4121
4122 -----------------------------------------------
4123 $ git clone git://example.com/pub/project.git
4124 $ cd project
4125 -----------------------------------------------
4126
4127 [[managing-branches]]
4128 Managing branches
4129 -----------------
4130
4131 -----------------------------------------------
4132 $ git branch # list all local branches in this repo
4133 $ git checkout test # switch working directory to branch "test"
4134 $ git branch new # create branch "new" starting at current HEAD
4135 $ git branch -d new # delete branch "new"
4136 -----------------------------------------------
4137
4138 Instead of basing a new branch on current HEAD (the default), use:
4139
4140 -----------------------------------------------
4141 $ git branch new test # branch named "test"
4142 $ git branch new v2.6.15 # tag named v2.6.15
4143 $ git branch new HEAD^ # commit before the most recent
4144 $ git branch new HEAD^^ # commit before that
4145 $ git branch new test~10 # ten commits before tip of branch "test"
4146 -----------------------------------------------
4147
4148 Create and switch to a new branch at the same time:
4149
4150 -----------------------------------------------
4151 $ git checkout -b new v2.6.15
4152 -----------------------------------------------
4153
4154 Update and examine branches from the repository you cloned from:
4155
4156 -----------------------------------------------
4157 $ git fetch # update
4158 $ git branch -r # list
4159 origin/master
4160 origin/next
4161 ...
4162 $ git checkout -b masterwork origin/master
4163 -----------------------------------------------
4164
4165 Fetch a branch from a different repository, and give it a new
4166 name in your repository:
4167
4168 -----------------------------------------------
4169 $ git fetch git://example.com/project.git theirbranch:mybranch
4170 $ git fetch git://example.com/project.git v2.6.15:mybranch
4171 -----------------------------------------------
4172
4173 Keep a list of repositories you work with regularly:
4174
4175 -----------------------------------------------
4176 $ git remote add example git://example.com/project.git
4177 $ git remote # list remote repositories
4178 example
4179 origin
4180 $ git remote show example # get details
4181 * remote example
4182 URL: git://example.com/project.git
4183 Tracked remote branches
4184 master next ...
4185 $ git fetch example # update branches from example
4186 $ git branch -r # list all remote branches
4187 -----------------------------------------------
4188
4189
4190 [[exploring-history]]
4191 Exploring history
4192 -----------------
4193
4194 -----------------------------------------------
4195 $ gitk # visualize and browse history
4196 $ git log # list all commits
4197 $ git log src/ # ...modifying src/
4198 $ git log v2.6.15..v2.6.16 # ...in v2.6.16, not in v2.6.15
4199 $ git log master..test # ...in branch test, not in branch master
4200 $ git log test..master # ...in branch master, but not in test
4201 $ git log test...master # ...in one branch, not in both
4202 $ git log -S'foo()' # ...where difference contain "foo()"
4203 $ git log --since="2 weeks ago"
4204 $ git log -p # show patches as well
4205 $ git show # most recent commit
4206 $ git diff v2.6.15..v2.6.16 # diff between two tagged versions
4207 $ git diff v2.6.15..HEAD # diff with current head
4208 $ git grep "foo()" # search working directory for "foo()"
4209 $ git grep v2.6.15 "foo()" # search old tree for "foo()"
4210 $ git show v2.6.15:a.txt # look at old version of a.txt
4211 -----------------------------------------------
4212
4213 Search for regressions:
4214
4215 -----------------------------------------------
4216 $ git bisect start
4217 $ git bisect bad # current version is bad
4218 $ git bisect good v2.6.13-rc2 # last known good revision
4219 Bisecting: 675 revisions left to test after this
4220 # test here, then:
4221 $ git bisect good # if this revision is good, or
4222 $ git bisect bad # if this revision is bad.
4223 # repeat until done.
4224 -----------------------------------------------
4225
4226 [[making-changes]]
4227 Making changes
4228 --------------
4229
4230 Make sure git knows who to blame:
4231
4232 ------------------------------------------------
4233 $ cat >>~/.gitconfig <<\EOF
4234 [user]
4235 name = Your Name Comes Here
4236 email = you@yourdomain.example.com
4237 EOF
4238 ------------------------------------------------
4239
4240 Select file contents to include in the next commit, then make the
4241 commit:
4242
4243 -----------------------------------------------
4244 $ git add a.txt # updated file
4245 $ git add b.txt # new file
4246 $ git rm c.txt # old file
4247 $ git commit
4248 -----------------------------------------------
4249
4250 Or, prepare and create the commit in one step:
4251
4252 -----------------------------------------------
4253 $ git commit d.txt # use latest content only of d.txt
4254 $ git commit -a # use latest content of all tracked files
4255 -----------------------------------------------
4256
4257 [[merging]]
4258 Merging
4259 -------
4260
4261 -----------------------------------------------
4262 $ git merge test # merge branch "test" into the current branch
4263 $ git pull git://example.com/project.git master
4264 # fetch and merge in remote branch
4265 $ git pull . test # equivalent to git merge test
4266 -----------------------------------------------
4267
4268 [[sharing-your-changes]]
4269 Sharing your changes
4270 --------------------
4271
4272 Importing or exporting patches:
4273
4274 -----------------------------------------------
4275 $ git format-patch origin..HEAD # format a patch for each commit
4276 # in HEAD but not in origin
4277 $ git am mbox # import patches from the mailbox "mbox"
4278 -----------------------------------------------
4279
4280 Fetch a branch in a different git repository, then merge into the
4281 current branch:
4282
4283 -----------------------------------------------
4284 $ git pull git://example.com/project.git theirbranch
4285 -----------------------------------------------
4286
4287 Store the fetched branch into a local branch before merging into the
4288 current branch:
4289
4290 -----------------------------------------------
4291 $ git pull git://example.com/project.git theirbranch:mybranch
4292 -----------------------------------------------
4293
4294 After creating commits on a local branch, update the remote
4295 branch with your commits:
4296
4297 -----------------------------------------------
4298 $ git push ssh://example.com/project.git mybranch:theirbranch
4299 -----------------------------------------------
4300
4301 When remote and local branch are both named "test":
4302
4303 -----------------------------------------------
4304 $ git push ssh://example.com/project.git test
4305 -----------------------------------------------
4306
4307 Shortcut version for a frequently used remote repository:
4308
4309 -----------------------------------------------
4310 $ git remote add example ssh://example.com/project.git
4311 $ git push example test
4312 -----------------------------------------------
4313
4314 [[repository-maintenance]]
4315 Repository maintenance
4316 ----------------------
4317
4318 Check for corruption:
4319
4320 -----------------------------------------------
4321 $ git fsck
4322 -----------------------------------------------
4323
4324 Recompress, remove unused cruft:
4325
4326 -----------------------------------------------
4327 $ git gc
4328 -----------------------------------------------
4329
4330
4331 [[todo]]
4332 Appendix B: Notes and todo list for this manual
4333 ===============================================
4334
4335 This is a work in progress.
4336
4337 The basic requirements:
4338
4339 - It must be readable in order, from beginning to end, by someone
4340 intelligent with a basic grasp of the UNIX command line, but without
4341 any special knowledge of git. If necessary, any other prerequisites
4342 should be specifically mentioned as they arise.
4343 - Whenever possible, section headings should clearly describe the task
4344 they explain how to do, in language that requires no more knowledge
4345 than necessary: for example, "importing patches into a project" rather
4346 than "the git-am command"
4347
4348 Think about how to create a clear chapter dependency graph that will
4349 allow people to get to important topics without necessarily reading
4350 everything in between.
4351
4352 Scan Documentation/ for other stuff left out; in particular:
4353
4354 - howto's
4355 - some of technical/?
4356 - hooks
4357 - list of commands in gitlink:git[1]
4358
4359 Scan email archives for other stuff left out
4360
4361 Scan man pages to see if any assume more background than this manual
4362 provides.
4363
4364 Simplify beginning by suggesting disconnected head instead of
4365 temporary branch creation?
4366
4367 Add more good examples. Entire sections of just cookbook examples
4368 might be a good idea; maybe make an "advanced examples" section a
4369 standard end-of-chapter section?
4370
4371 Include cross-references to the glossary, where appropriate.
4372
4373 Document shallow clones? See draft 1.5.0 release notes for some
4374 documentation.
4375
4376 Add a section on working with other version control systems, including
4377 CVS, Subversion, and just imports of series of release tarballs.
4378
4379 More details on gitweb?
4380
4381 Write a chapter on using plumbing and writing scripts.
4382
4383 Alternates, clone -reference, etc.
4384
4385 git unpack-objects -r for recovery