77e73bf22cf9cb809399b3935d78aaa56c1d762b
1 Git User's Manual (for version 1.5.1 or newer)
2 ______________________________________________
5 Git is a fast distributed revision control system.
7 This manual is designed to be readable by someone with basic unix
8 command-line skills, but no previous knowledge of git.
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.
15 People needing to do actual development will also want to read
16 <<Developing-with-git>> and <<sharing-development>>.
18 Further chapters cover more specialized topics.
20 Comprehensive reference documentation is available through the man
21 pages. For a command such as "git clone", just use
23 ------------------------------------------------
24 $ man git-clone
25 ------------------------------------------------
27 See also <<git-quick-start>> for a brief overview of git commands,
28 without any explanation.
30 Finally, see <<todo>> for ways that you can help make this manual more
31 complete.
34 [[repositories-and-branches]]
35 Repositories and Branches
36 =========================
38 [[how-to-get-a-git-repository]]
39 How to get a git repository
40 ---------------------------
42 It will be useful to have a git repository to experiment with as you
43 read this manual.
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:
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 ------------------------------------------------
56 The initial clone may be time-consuming for a large project, but you
57 will only need to clone once.
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.
65 [[how-to-check-out]]
66 How to check out a different version of a project
67 -------------------------------------------------
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>>.
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:
79 ------------------------------------------------
80 $ git branch
81 * master
82 ------------------------------------------------
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.
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:
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 ------------------------------------------------
106 Tags are expected to always point at the same version of a project,
107 while heads are expected to advance as development progresses.
109 Create a new branch head pointing to one of these versions and check it
110 out using gitlink:git-checkout[1]:
112 ------------------------------------------------
113 $ git checkout -b new v2.6.13
114 ------------------------------------------------
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:
120 ------------------------------------------------
121 $ git branch
122 master
123 * new
124 ------------------------------------------------
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
129 ------------------------------------------------
130 $ git reset --hard v2.6.17
131 ------------------------------------------------
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.
138 [[understanding-commits]]
139 Understanding History: Commits
140 ------------------------------
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:
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
152 Remove duplicate getenv(DB_ENVIRONMENT) call
154 Noted by Tony Luck.
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 @@
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;
168 if (mkdir(".git", 0755) < 0) {
169 ------------------------------------------------
171 As you can see, a commit shows who made the latest change, what they
172 did, and why.
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.
185 In fact, in <<git-internals>> 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.
189 [[understanding-reachability]]
190 Understanding history: commits, parents, and reachability
191 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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.
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.
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.
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 descendent of X, or that there is a chain of parents
212 leading from commit Y to commit X.
214 [[history-diagrams]]
215 Understanding history: History diagrams
216 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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:
223 ................................................
224 o--o--o <-- Branch A
225 /
226 o--o--o <-- master
227 \
228 o--o--o <-- Branch B
229 ................................................
231 If we need to talk about a particular commit, the character "o" may
232 be replaced with another letter or number.
234 [[what-is-a-branch]]
235 Understanding history: What is a branch?
236 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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".
245 However, when no confusion will result, we often just use the term
246 "branch" both for branches and for branch heads.
248 [[manipulating-branches]]
249 Manipulating branches
250 ---------------------
252 Creating, deleting, and modifying branches is quick and easy; here's
253 a summary of the commands:
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.
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:
285 ------------------------------------------------
286 $ cat .git/HEAD
287 ref: refs/heads/master
288 ------------------------------------------------
290 [[detached-head]]
291 Examining an old version without creating a new branch
292 ------------------------------------------------------
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:
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 ------------------------------------------------
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:
310 ------------------------------------------------
311 $ cat .git/HEAD
312 427abfa28afedffadfca9dd8b067eb6d36bac53f
313 $ git branch
314 * (no branch)
315 master
316 ------------------------------------------------
318 In this case we say that the HEAD is "detached".
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.
324 [[examining-remote-branches]]
325 Examining branches from a remote repository
326 -------------------------------------------
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]:
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 ------------------------------------------------
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:
349 ------------------------------------------------
350 $ git checkout -b my-todo-copy origin/todo
351 ------------------------------------------------
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.
356 [[how-git-stores-references]]
357 Naming branches, tags, and other references
358 -------------------------------------------
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:
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".
369 The full name is occasionally useful if, for example, there ever
370 exists a tag and a branch with the same name.
372 As another useful shortcut, the "HEAD" of a repository can be referred
373 to just using the name of that repository. So, for example, "origin"
374 is usually a shortcut for the HEAD branch in the repository "origin".
376 For the complete list of paths which git checks for references, and
377 the order it uses to decide which to choose when there are multiple
378 references with the same shorthand name, see the "SPECIFYING
379 REVISIONS" section of gitlink:git-rev-parse[1].
381 [[Updating-a-repository-with-git-fetch]]
382 Updating a repository with git fetch
383 ------------------------------------
385 Eventually the developer cloned from will do additional work in her
386 repository, creating new commits and advancing the branches to point
387 at the new commits.
389 The command "git fetch", with no arguments, will update all of the
390 remote-tracking branches to the latest version found in her
391 repository. It will not touch any of your own branches--not even the
392 "master" branch that was created for you on clone.
394 [[fetching-branches]]
395 Fetching branches from other repositories
396 -----------------------------------------
398 You can also track branches from repositories other than the one you
399 cloned from, using gitlink:git-remote[1]:
401 -------------------------------------------------
402 $ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
403 $ git fetch linux-nfs
404 * refs/remotes/linux-nfs/master: storing branch 'master' ...
405 commit: bf81b46
406 -------------------------------------------------
408 New remote-tracking branches will be stored under the shorthand name
409 that you gave "git remote add", in this case linux-nfs:
411 -------------------------------------------------
412 $ git branch -r
413 linux-nfs/master
414 origin/master
415 -------------------------------------------------
417 If you run "git fetch <remote>" later, the tracking branches for the
418 named <remote> will be updated.
420 If you examine the file .git/config, you will see that git has added
421 a new stanza:
423 -------------------------------------------------
424 $ cat .git/config
425 ...
426 [remote "linux-nfs"]
427 url = git://linux-nfs.org/pub/nfs-2.6.git
428 fetch = +refs/heads/*:refs/remotes/linux-nfs/*
429 ...
430 -------------------------------------------------
432 This is what causes git to track the remote's branches; you may modify
433 or delete these configuration options by editing .git/config with a
434 text editor. (See the "CONFIGURATION FILE" section of
435 gitlink:git-config[1] for details.)
437 [[exploring-git-history]]
438 Exploring git history
439 =====================
441 Git is best thought of as a tool for storing the history of a
442 collection of files. It does this by storing compressed snapshots of
443 the contents of a file heirarchy, together with "commits" which show
444 the relationships between these snapshots.
446 Git provides extremely flexible and fast tools for exploring the
447 history of a project.
449 We start with one specialized tool that is useful for finding the
450 commit that introduced a bug into a project.
452 [[using-bisect]]
453 How to use bisect to find a regression
454 --------------------------------------
456 Suppose version 2.6.18 of your project worked, but the version at
457 "master" crashes. Sometimes the best way to find the cause of such a
458 regression is to perform a brute-force search through the project's
459 history to find the particular commit that caused the problem. The
460 gitlink:git-bisect[1] command can help you do this:
462 -------------------------------------------------
463 $ git bisect start
464 $ git bisect good v2.6.18
465 $ git bisect bad master
466 Bisecting: 3537 revisions left to test after this
467 [65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
468 -------------------------------------------------
470 If you run "git branch" at this point, you'll see that git has
471 temporarily moved you to a new branch named "bisect". This branch
472 points to a commit (with commit id 65934...) that is reachable from
473 v2.6.19 but not from v2.6.18. Compile and test it, and see whether
474 it crashes. Assume it does crash. Then:
476 -------------------------------------------------
477 $ git bisect bad
478 Bisecting: 1769 revisions left to test after this
479 [7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
480 -------------------------------------------------
482 checks out an older version. Continue like this, telling git at each
483 stage whether the version it gives you is good or bad, and notice
484 that the number of revisions left to test is cut approximately in
485 half each time.
487 After about 13 tests (in this case), it will output the commit id of
488 the guilty commit. You can then examine the commit with
489 gitlink:git-show[1], find out who wrote it, and mail them your bug
490 report with the commit id. Finally, run
492 -------------------------------------------------
493 $ git bisect reset
494 -------------------------------------------------
496 to return you to the branch you were on before and delete the
497 temporary "bisect" branch.
499 Note that the version which git-bisect checks out for you at each
500 point is just a suggestion, and you're free to try a different
501 version if you think it would be a good idea. For example,
502 occasionally you may land on a commit that broke something unrelated;
503 run
505 -------------------------------------------------
506 $ git bisect visualize
507 -------------------------------------------------
509 which will run gitk and label the commit it chose with a marker that
510 says "bisect". Chose a safe-looking commit nearby, note its commit
511 id, and check it out with:
513 -------------------------------------------------
514 $ git reset --hard fb47ddb2db...
515 -------------------------------------------------
517 then test, run "bisect good" or "bisect bad" as appropriate, and
518 continue.
520 [[naming-commits]]
521 Naming commits
522 --------------
524 We have seen several ways of naming commits already:
526 - 40-hexdigit object name
527 - branch name: refers to the commit at the head of the given
528 branch
529 - tag name: refers to the commit pointed to by the given tag
530 (we've seen branches and tags are special cases of
531 <<how-git-stores-references,references>>).
532 - HEAD: refers to the head of the current branch
534 There are many more; see the "SPECIFYING REVISIONS" section of the
535 gitlink:git-rev-parse[1] man page for the complete list of ways to
536 name revisions. Some examples:
538 -------------------------------------------------
539 $ git show fb47ddb2 # the first few characters of the object name
540 # are usually enough to specify it uniquely
541 $ git show HEAD^ # the parent of the HEAD commit
542 $ git show HEAD^^ # the grandparent
543 $ git show HEAD~4 # the great-great-grandparent
544 -------------------------------------------------
546 Recall that merge commits may have more than one parent; by default,
547 ^ and ~ follow the first parent listed in the commit, but you can
548 also choose:
550 -------------------------------------------------
551 $ git show HEAD^1 # show the first parent of HEAD
552 $ git show HEAD^2 # show the second parent of HEAD
553 -------------------------------------------------
555 In addition to HEAD, there are several other special names for
556 commits:
558 Merges (to be discussed later), as well as operations such as
559 git-reset, which change the currently checked-out commit, generally
560 set ORIG_HEAD to the value HEAD had before the current operation.
562 The git-fetch operation always stores the head of the last fetched
563 branch in FETCH_HEAD. For example, if you run git fetch without
564 specifying a local branch as the target of the operation
566 -------------------------------------------------
567 $ git fetch git://example.com/proj.git theirbranch
568 -------------------------------------------------
570 the fetched commits will still be available from FETCH_HEAD.
572 When we discuss merges we'll also see the special name MERGE_HEAD,
573 which refers to the other branch that we're merging in to the current
574 branch.
576 The gitlink:git-rev-parse[1] command is a low-level command that is
577 occasionally useful for translating some name for a commit to the object
578 name for that commit:
580 -------------------------------------------------
581 $ git rev-parse origin
582 e05db0fd4f31dde7005f075a84f96b360d05984b
583 -------------------------------------------------
585 [[creating-tags]]
586 Creating tags
587 -------------
589 We can also create a tag to refer to a particular commit; after
590 running
592 -------------------------------------------------
593 $ git tag stable-1 1b2e1d63ff
594 -------------------------------------------------
596 You can use stable-1 to refer to the commit 1b2e1d63ff.
598 This creates a "lightweight" tag. If you would also like to include a
599 comment with the tag, and possibly sign it cryptographically, then you
600 should create a tag object instead; see the gitlink:git-tag[1] man page
601 for details.
603 [[browsing-revisions]]
604 Browsing revisions
605 ------------------
607 The gitlink:git-log[1] command can show lists of commits. On its
608 own, it shows all commits reachable from the parent commit; but you
609 can also make more specific requests:
611 -------------------------------------------------
612 $ git log v2.5.. # commits since (not reachable from) v2.5
613 $ git log test..master # commits reachable from master but not test
614 $ git log master..test # ...reachable from test but not master
615 $ git log master...test # ...reachable from either test or master,
616 # but not both
617 $ git log --since="2 weeks ago" # commits from the last 2 weeks
618 $ git log Makefile # commits which modify Makefile
619 $ git log fs/ # ... which modify any file under fs/
620 $ git log -S'foo()' # commits which add or remove any file data
621 # matching the string 'foo()'
622 -------------------------------------------------
624 And of course you can combine all of these; the following finds
625 commits since v2.5 which touch the Makefile or any file under fs:
627 -------------------------------------------------
628 $ git log v2.5.. Makefile fs/
629 -------------------------------------------------
631 You can also ask git log to show patches:
633 -------------------------------------------------
634 $ git log -p
635 -------------------------------------------------
637 See the "--pretty" option in the gitlink:git-log[1] man page for more
638 display options.
640 Note that git log starts with the most recent commit and works
641 backwards through the parents; however, since git history can contain
642 multiple independent lines of development, the particular order that
643 commits are listed in may be somewhat arbitrary.
645 [[generating-diffs]]
646 Generating diffs
647 ----------------
649 You can generate diffs between any two versions using
650 gitlink:git-diff[1]:
652 -------------------------------------------------
653 $ git diff master..test
654 -------------------------------------------------
656 Sometimes what you want instead is a set of patches:
658 -------------------------------------------------
659 $ git format-patch master..test
660 -------------------------------------------------
662 will generate a file with a patch for each commit reachable from test
663 but not from master. Note that if master also has commits which are
664 not reachable from test, then the combined result of these patches
665 will not be the same as the diff produced by the git-diff example.
667 [[viewing-old-file-versions]]
668 Viewing old file versions
669 -------------------------
671 You can always view an old version of a file by just checking out the
672 correct revision first. But sometimes it is more convenient to be
673 able to view an old version of a single file without checking
674 anything out; this command does that:
676 -------------------------------------------------
677 $ git show v2.5:fs/locks.c
678 -------------------------------------------------
680 Before the colon may be anything that names a commit, and after it
681 may be any path to a file tracked by git.
683 [[history-examples]]
684 Examples
685 --------
687 [[counting-commits-on-a-branch]]
688 Counting the number of commits on a branch
689 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
691 Suppose you want to know how many commits you've made on "mybranch"
692 since it diverged from "origin":
694 -------------------------------------------------
695 $ git log --pretty=oneline origin..mybranch | wc -l
696 -------------------------------------------------
698 Alternatively, you may often see this sort of thing done with the
699 lower-level command gitlink:git-rev-list[1], which just lists the SHA1's
700 of all the given commits:
702 -------------------------------------------------
703 $ git rev-list origin..mybranch | wc -l
704 -------------------------------------------------
706 [[checking-for-equal-branches]]
707 Check whether two branches point at the same history
708 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
710 Suppose you want to check whether two branches point at the same point
711 in history.
713 -------------------------------------------------
714 $ git diff origin..master
715 -------------------------------------------------
717 will tell you whether the contents of the project are the same at the
718 two branches; in theory, however, it's possible that the same project
719 contents could have been arrived at by two different historical
720 routes. You could compare the object names:
722 -------------------------------------------------
723 $ git rev-list origin
724 e05db0fd4f31dde7005f075a84f96b360d05984b
725 $ git rev-list master
726 e05db0fd4f31dde7005f075a84f96b360d05984b
727 -------------------------------------------------
729 Or you could recall that the ... operator selects all commits
730 contained reachable from either one reference or the other but not
731 both: so
733 -------------------------------------------------
734 $ git log origin...master
735 -------------------------------------------------
737 will return no commits when the two branches are equal.
739 [[finding-tagged-descendants]]
740 Find first tagged version including a given fix
741 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
743 Suppose you know that the commit e05db0fd fixed a certain problem.
744 You'd like to find the earliest tagged release that contains that
745 fix.
747 Of course, there may be more than one answer--if the history branched
748 after commit e05db0fd, then there could be multiple "earliest" tagged
749 releases.
751 You could just visually inspect the commits since e05db0fd:
753 -------------------------------------------------
754 $ gitk e05db0fd..
755 -------------------------------------------------
757 Or you can use gitlink:git-name-rev[1], which will give the commit a
758 name based on any tag it finds pointing to one of the commit's
759 descendants:
761 -------------------------------------------------
762 $ git name-rev --tags e05db0fd
763 e05db0fd tags/v1.5.0-rc1^0~23
764 -------------------------------------------------
766 The gitlink:git-describe[1] command does the opposite, naming the
767 revision using a tag on which the given commit is based:
769 -------------------------------------------------
770 $ git describe e05db0fd
771 v1.5.0-rc0-260-ge05db0f
772 -------------------------------------------------
774 but that may sometimes help you guess which tags might come after the
775 given commit.
777 If you just want to verify whether a given tagged version contains a
778 given commit, you could use gitlink:git-merge-base[1]:
780 -------------------------------------------------
781 $ git merge-base e05db0fd v1.5.0-rc1
782 e05db0fd4f31dde7005f075a84f96b360d05984b
783 -------------------------------------------------
785 The merge-base command finds a common ancestor of the given commits,
786 and always returns one or the other in the case where one is a
787 descendant of the other; so the above output shows that e05db0fd
788 actually is an ancestor of v1.5.0-rc1.
790 Alternatively, note that
792 -------------------------------------------------
793 $ git log v1.5.0-rc1..e05db0fd
794 -------------------------------------------------
796 will produce empty output if and only if v1.5.0-rc1 includes e05db0fd,
797 because it outputs only commits that are not reachable from v1.5.0-rc1.
799 As yet another alternative, the gitlink:git-show-branch[1] command lists
800 the commits reachable from its arguments with a display on the left-hand
801 side that indicates which arguments that commit is reachable from. So,
802 you can run something like
804 -------------------------------------------------
805 $ git show-branch e05db0fd v1.5.0-rc0 v1.5.0-rc1 v1.5.0-rc2
806 ! [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
807 available
808 ! [v1.5.0-rc0] GIT v1.5.0 preview
809 ! [v1.5.0-rc1] GIT v1.5.0-rc1
810 ! [v1.5.0-rc2] GIT v1.5.0-rc2
811 ...
812 -------------------------------------------------
814 then search for a line that looks like
816 -------------------------------------------------
817 + ++ [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
818 available
819 -------------------------------------------------
821 Which shows that e05db0fd is reachable from itself, from v1.5.0-rc1, and
822 from v1.5.0-rc2, but not from v1.5.0-rc0.
824 [[showing-commits-unique-to-a-branch]]
825 Showing commits unique to a given branch
826 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
828 Suppose you would like to see all the commits reachable from the branch
829 head named "master" but not from any other head in your repository.
831 We can list all the heads in this repository with
832 gitlink:git-show-ref[1]:
834 -------------------------------------------------
835 $ git show-ref --heads
836 bf62196b5e363d73353a9dcf094c59595f3153b7 refs/heads/core-tutorial
837 db768d5504c1bb46f63ee9d6e1772bd047e05bf9 refs/heads/maint
838 a07157ac624b2524a059a3414e99f6f44bebc1e7 refs/heads/master
839 24dbc180ea14dc1aebe09f14c8ecf32010690627 refs/heads/tutorial-2
840 1e87486ae06626c2f31eaa63d26fc0fd646c8af2 refs/heads/tutorial-fixes
841 -------------------------------------------------
843 We can get just the branch-head names, and remove "master", with
844 the help of the standard utilities cut and grep:
846 -------------------------------------------------
847 $ git show-ref --heads | cut -d' ' -f2 | grep -v '^refs/heads/master'
848 refs/heads/core-tutorial
849 refs/heads/maint
850 refs/heads/tutorial-2
851 refs/heads/tutorial-fixes
852 -------------------------------------------------
854 And then we can ask to see all the commits reachable from master
855 but not from these other heads:
857 -------------------------------------------------
858 $ gitk master --not $( git show-ref --heads | cut -d' ' -f2 |
859 grep -v '^refs/heads/master' )
860 -------------------------------------------------
862 Obviously, endless variations are possible; for example, to see all
863 commits reachable from some head but not from any tag in the repository:
865 -------------------------------------------------
866 $ gitk $( git show-ref --heads ) --not $( git show-ref --tags )
867 -------------------------------------------------
869 (See gitlink:git-rev-parse[1] for explanations of commit-selecting
870 syntax such as `--not`.)
872 [[making-a-release]]
873 Creating a changelog and tarball for a software release
874 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
876 The gitlink:git-archive[1] command can create a tar or zip archive from
877 any version of a project; for example:
879 -------------------------------------------------
880 $ git archive --format=tar --prefix=project/ HEAD | gzip >latest.tar.gz
881 -------------------------------------------------
883 will use HEAD to produce a tar archive in which each filename is
884 preceded by "project/".
886 If you're releasing a new version of a software project, you may want
887 to simultaneously make a changelog to include in the release
888 announcement.
890 Linus Torvalds, for example, makes new kernel releases by tagging them,
891 then running:
893 -------------------------------------------------
894 $ release-script 2.6.12 2.6.13-rc6 2.6.13-rc7
895 -------------------------------------------------
897 where release-script is a shell script that looks like:
899 -------------------------------------------------
900 #!/bin/sh
901 stable="$1"
902 last="$2"
903 new="$3"
904 echo "# git tag v$new"
905 echo "git archive --prefix=linux-$new/ v$new | gzip -9 > ../linux-$new.tar.gz"
906 echo "git diff v$stable v$new | gzip -9 > ../patch-$new.gz"
907 echo "git log --no-merges v$new ^v$last > ../ChangeLog-$new"
908 echo "git shortlog --no-merges v$new ^v$last > ../ShortLog"
909 echo "git diff --stat --summary -M v$last v$new > ../diffstat-$new"
910 -------------------------------------------------
912 and then he just cut-and-pastes the output commands after verifying that
913 they look OK.
915 [[Finding-comments-with-given-content]]
916 Finding commits referencing a file with given content
917 -----------------------------------------------------
919 Somebody hands you a copy of a file, and asks which commits modified a
920 file such that it contained the given content either before or after the
921 commit. You can find out with this:
923 -------------------------------------------------
924 $ git log --raw -r --abbrev=40 --pretty=oneline -- filename |
925 grep -B 1 `git hash-object filename`
926 -------------------------------------------------
928 Figuring out why this works is left as an exercise to the (advanced)
929 student. The gitlink:git-log[1], gitlink:git-diff-tree[1], and
930 gitlink:git-hash-object[1] man pages may prove helpful.
932 [[Developing-with-git]]
933 Developing with git
934 ===================
936 [[telling-git-your-name]]
937 Telling git your name
938 ---------------------
940 Before creating any commits, you should introduce yourself to git. The
941 easiest way to do so is to make sure the following lines appear in a
942 file named .gitconfig in your home directory:
944 ------------------------------------------------
945 [user]
946 name = Your Name Comes Here
947 email = you@yourdomain.example.com
948 ------------------------------------------------
950 (See the "CONFIGURATION FILE" section of gitlink:git-config[1] for
951 details on the configuration file.)
954 [[creating-a-new-repository]]
955 Creating a new repository
956 -------------------------
958 Creating a new repository from scratch is very easy:
960 -------------------------------------------------
961 $ mkdir project
962 $ cd project
963 $ git init
964 -------------------------------------------------
966 If you have some initial content (say, a tarball):
968 -------------------------------------------------
969 $ tar -xzvf project.tar.gz
970 $ cd project
971 $ git init
972 $ git add . # include everything below ./ in the first commit:
973 $ git commit
974 -------------------------------------------------
976 [[how-to-make-a-commit]]
977 How to make a commit
978 --------------------
980 Creating a new commit takes three steps:
982 1. Making some changes to the working directory using your
983 favorite editor.
984 2. Telling git about your changes.
985 3. Creating the commit using the content you told git about
986 in step 2.
988 In practice, you can interleave and repeat steps 1 and 2 as many
989 times as you want: in order to keep track of what you want committed
990 at step 3, git maintains a snapshot of the tree's contents in a
991 special staging area called "the index."
993 At the beginning, the content of the index will be identical to
994 that of the HEAD. The command "git diff --cached", which shows
995 the difference between the HEAD and the index, should therefore
996 produce no output at that point.
998 Modifying the index is easy:
1000 To update the index with the new contents of a modified file, use
1002 -------------------------------------------------
1003 $ git add path/to/file
1004 -------------------------------------------------
1006 To add the contents of a new file to the index, use
1008 -------------------------------------------------
1009 $ git add path/to/file
1010 -------------------------------------------------
1012 To remove a file from the index and from the working tree,
1014 -------------------------------------------------
1015 $ git rm path/to/file
1016 -------------------------------------------------
1018 After each step you can verify that
1020 -------------------------------------------------
1021 $ git diff --cached
1022 -------------------------------------------------
1024 always shows the difference between the HEAD and the index file--this
1025 is what you'd commit if you created the commit now--and that
1027 -------------------------------------------------
1028 $ git diff
1029 -------------------------------------------------
1031 shows the difference between the working tree and the index file.
1033 Note that "git add" always adds just the current contents of a file
1034 to the index; further changes to the same file will be ignored unless
1035 you run git-add on the file again.
1037 When you're ready, just run
1039 -------------------------------------------------
1040 $ git commit
1041 -------------------------------------------------
1043 and git will prompt you for a commit message and then create the new
1044 commit. Check to make sure it looks like what you expected with
1046 -------------------------------------------------
1047 $ git show
1048 -------------------------------------------------
1050 As a special shortcut,
1052 -------------------------------------------------
1053 $ git commit -a
1054 -------------------------------------------------
1056 will update the index with any files that you've modified or removed
1057 and create a commit, all in one step.
1059 A number of commands are useful for keeping track of what you're
1060 about to commit:
1062 -------------------------------------------------
1063 $ git diff --cached # difference between HEAD and the index; what
1064 # would be commited if you ran "commit" now.
1065 $ git diff # difference between the index file and your
1066 # working directory; changes that would not
1067 # be included if you ran "commit" now.
1068 $ git diff HEAD # difference between HEAD and working tree; what
1069 # would be committed if you ran "commit -a" now.
1070 $ git status # a brief per-file summary of the above.
1071 -------------------------------------------------
1073 [[creating-good-commit-messages]]
1074 Creating good commit messages
1075 -----------------------------
1077 Though not required, it's a good idea to begin the commit message
1078 with a single short (less than 50 character) line summarizing the
1079 change, followed by a blank line and then a more thorough
1080 description. Tools that turn commits into email, for example, use
1081 the first line on the Subject line and the rest of the commit in the
1082 body.
1084 [[ignoring-files]]
1085 Ignoring files
1086 --------------
1088 A project will often generate files that you do 'not' want to track with git.
1089 This typically includes files generated by a build process or temporary
1090 backup files made by your editor. Of course, 'not' tracking files with git
1091 is just a matter of 'not' calling "`git add`" on them. But it quickly becomes
1092 annoying to have these untracked files lying around; e.g. they make
1093 "`git add .`" and "`git commit -a`" practically useless, and they keep
1094 showing up in the output of "`git status`".
1096 You can tell git to ignore certain files by creating a file called .gitignore
1097 in the top level of your working directory, with contents such as:
1099 -------------------------------------------------
1100 # Lines starting with '#' are considered comments.
1101 # Ignore any file named foo.txt.
1102 foo.txt
1103 # Ignore (generated) html files,
1104 *.html
1105 # except foo.html which is maintained by hand.
1106 !foo.html
1107 # Ignore objects and archives.
1108 *.[oa]
1109 -------------------------------------------------
1111 See gitlink:gitignore[5] for a detailed explanation of the syntax. You can
1112 also place .gitignore files in other directories in your working tree, and they
1113 will apply to those directories and their subdirectories. The `.gitignore`
1114 files can be added to your repository like any other files (just run `git add
1115 .gitignore` and `git commit`, as usual), which is convenient when the exclude
1116 patterns (such as patterns matching build output files) would also make sense
1117 for other users who clone your repository.
1119 If you wish the exclude patterns to affect only certain repositories
1120 (instead of every repository for a given project), you may instead put
1121 them in a file in your repository named .git/info/exclude, or in any file
1122 specified by the `core.excludesfile` configuration variable. Some git
1123 commands can also take exclude patterns directly on the command line.
1124 See gitlink:gitignore[5] for the details.
1126 [[how-to-merge]]
1127 How to merge
1128 ------------
1130 You can rejoin two diverging branches of development using
1131 gitlink:git-merge[1]:
1133 -------------------------------------------------
1134 $ git merge branchname
1135 -------------------------------------------------
1137 merges the development in the branch "branchname" into the current
1138 branch. If there are conflicts--for example, if the same file is
1139 modified in two different ways in the remote branch and the local
1140 branch--then you are warned; the output may look something like this:
1142 -------------------------------------------------
1143 $ git merge next
1144 100% (4/4) done
1145 Auto-merged file.txt
1146 CONFLICT (content): Merge conflict in file.txt
1147 Automatic merge failed; fix conflicts and then commit the result.
1148 -------------------------------------------------
1150 Conflict markers are left in the problematic files, and after
1151 you resolve the conflicts manually, you can update the index
1152 with the contents and run git commit, as you normally would when
1153 creating a new file.
1155 If you examine the resulting commit using gitk, you will see that it
1156 has two parents, one pointing to the top of the current branch, and
1157 one to the top of the other branch.
1159 [[resolving-a-merge]]
1160 Resolving a merge
1161 -----------------
1163 When a merge isn't resolved automatically, git leaves the index and
1164 the working tree in a special state that gives you all the
1165 information you need to help resolve the merge.
1167 Files with conflicts are marked specially in the index, so until you
1168 resolve the problem and update the index, gitlink:git-commit[1] will
1169 fail:
1171 -------------------------------------------------
1172 $ git commit
1173 file.txt: needs merge
1174 -------------------------------------------------
1176 Also, gitlink:git-status[1] will list those files as "unmerged", and the
1177 files with conflicts will have conflict markers added, like this:
1179 -------------------------------------------------
1180 <<<<<<< HEAD:file.txt
1181 Hello world
1182 =======
1183 Goodbye
1184 >>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1185 -------------------------------------------------
1187 All you need to do is edit the files to resolve the conflicts, and then
1189 -------------------------------------------------
1190 $ git add file.txt
1191 $ git commit
1192 -------------------------------------------------
1194 Note that the commit message will already be filled in for you with
1195 some information about the merge. Normally you can just use this
1196 default message unchanged, but you may add additional commentary of
1197 your own if desired.
1199 The above is all you need to know to resolve a simple merge. But git
1200 also provides more information to help resolve conflicts:
1202 [[conflict-resolution]]
1203 Getting conflict-resolution help during a merge
1204 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1206 All of the changes that git was able to merge automatically are
1207 already added to the index file, so gitlink:git-diff[1] shows only
1208 the conflicts. It uses an unusual syntax:
1210 -------------------------------------------------
1211 $ git diff
1212 diff --cc file.txt
1213 index 802992c,2b60207..0000000
1214 --- a/file.txt
1215 +++ b/file.txt
1216 @@@ -1,1 -1,1 +1,5 @@@
1217 ++<<<<<<< HEAD:file.txt
1218 +Hello world
1219 ++=======
1220 + Goodbye
1221 ++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1222 -------------------------------------------------
1224 Recall that the commit which will be commited after we resolve this
1225 conflict will have two parents instead of the usual one: one parent
1226 will be HEAD, the tip of the current branch; the other will be the
1227 tip of the other branch, which is stored temporarily in MERGE_HEAD.
1229 During the merge, the index holds three versions of each file. Each of
1230 these three "file stages" represents a different version of the file:
1232 -------------------------------------------------
1233 $ git show :1:file.txt # the file in a common ancestor of both branches
1234 $ git show :2:file.txt # the version from HEAD, but including any
1235 # nonconflicting changes from MERGE_HEAD
1236 $ git show :3:file.txt # the version from MERGE_HEAD, but including any
1237 # nonconflicting changes from HEAD.
1238 -------------------------------------------------
1240 Since the stage 2 and stage 3 versions have already been updated with
1241 nonconflicting changes, the only remaining differences between them are
1242 the important ones; thus gitlink:git-diff[1] can use the information in
1243 the index to show only those conflicts.
1245 The diff above shows the differences between the working-tree version of
1246 file.txt and the stage 2 and stage 3 versions. So instead of preceding
1247 each line by a single "+" or "-", it now uses two columns: the first
1248 column is used for differences between the first parent and the working
1249 directory copy, and the second for differences between the second parent
1250 and the working directory copy. (See the "COMBINED DIFF FORMAT" section
1251 of gitlink:git-diff-files[1] for a details of the format.)
1253 After resolving the conflict in the obvious way (but before updating the
1254 index), the diff will look like:
1256 -------------------------------------------------
1257 $ git diff
1258 diff --cc file.txt
1259 index 802992c,2b60207..0000000
1260 --- a/file.txt
1261 +++ b/file.txt
1262 @@@ -1,1 -1,1 +1,1 @@@
1263 - Hello world
1264 -Goodbye
1265 ++Goodbye world
1266 -------------------------------------------------
1268 This shows that our resolved version deleted "Hello world" from the
1269 first parent, deleted "Goodbye" from the second parent, and added
1270 "Goodbye world", which was previously absent from both.
1272 Some special diff options allow diffing the working directory against
1273 any of these stages:
1275 -------------------------------------------------
1276 $ git diff -1 file.txt # diff against stage 1
1277 $ git diff --base file.txt # same as the above
1278 $ git diff -2 file.txt # diff against stage 2
1279 $ git diff --ours file.txt # same as the above
1280 $ git diff -3 file.txt # diff against stage 3
1281 $ git diff --theirs file.txt # same as the above.
1282 -------------------------------------------------
1284 The gitlink:git-log[1] and gitk[1] commands also provide special help
1285 for merges:
1287 -------------------------------------------------
1288 $ git log --merge
1289 $ gitk --merge
1290 -------------------------------------------------
1292 These will display all commits which exist only on HEAD or on
1293 MERGE_HEAD, and which touch an unmerged file.
1295 You may also use gitlink:git-mergetool[1], which lets you merge the
1296 unmerged files using external tools such as emacs or kdiff3.
1298 Each time you resolve the conflicts in a file and update the index:
1300 -------------------------------------------------
1301 $ git add file.txt
1302 -------------------------------------------------
1304 the different stages of that file will be "collapsed", after which
1305 git-diff will (by default) no longer show diffs for that file.
1307 [[undoing-a-merge]]
1308 Undoing a merge
1309 ---------------
1311 If you get stuck and decide to just give up and throw the whole mess
1312 away, you can always return to the pre-merge state with
1314 -------------------------------------------------
1315 $ git reset --hard HEAD
1316 -------------------------------------------------
1318 Or, if you've already commited the merge that you want to throw away,
1320 -------------------------------------------------
1321 $ git reset --hard ORIG_HEAD
1322 -------------------------------------------------
1324 However, this last command can be dangerous in some cases--never
1325 throw away a commit you have already committed if that commit may
1326 itself have been merged into another branch, as doing so may confuse
1327 further merges.
1329 [[fast-forwards]]
1330 Fast-forward merges
1331 -------------------
1333 There is one special case not mentioned above, which is treated
1334 differently. Normally, a merge results in a merge commit, with two
1335 parents, one pointing at each of the two lines of development that
1336 were merged.
1338 However, if the current branch is a descendant of the other--so every
1339 commit present in the one is already contained in the other--then git
1340 just performs a "fast forward"; the head of the current branch is moved
1341 forward to point at the head of the merged-in branch, without any new
1342 commits being created.
1344 [[fixing-mistakes]]
1345 Fixing mistakes
1346 ---------------
1348 If you've messed up the working tree, but haven't yet committed your
1349 mistake, you can return the entire working tree to the last committed
1350 state with
1352 -------------------------------------------------
1353 $ git reset --hard HEAD
1354 -------------------------------------------------
1356 If you make a commit that you later wish you hadn't, there are two
1357 fundamentally different ways to fix the problem:
1359 1. You can create a new commit that undoes whatever was done
1360 by the previous commit. This is the correct thing if your
1361 mistake has already been made public.
1363 2. You can go back and modify the old commit. You should
1364 never do this if you have already made the history public;
1365 git does not normally expect the "history" of a project to
1366 change, and cannot correctly perform repeated merges from
1367 a branch that has had its history changed.
1369 [[reverting-a-commit]]
1370 Fixing a mistake with a new commit
1371 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1373 Creating a new commit that reverts an earlier change is very easy;
1374 just pass the gitlink:git-revert[1] command a reference to the bad
1375 commit; for example, to revert the most recent commit:
1377 -------------------------------------------------
1378 $ git revert HEAD
1379 -------------------------------------------------
1381 This will create a new commit which undoes the change in HEAD. You
1382 will be given a chance to edit the commit message for the new commit.
1384 You can also revert an earlier change, for example, the next-to-last:
1386 -------------------------------------------------
1387 $ git revert HEAD^
1388 -------------------------------------------------
1390 In this case git will attempt to undo the old change while leaving
1391 intact any changes made since then. If more recent changes overlap
1392 with the changes to be reverted, then you will be asked to fix
1393 conflicts manually, just as in the case of <<resolving-a-merge,
1394 resolving a merge>>.
1396 [[fixing-a-mistake-by-editing-history]]
1397 Fixing a mistake by editing history
1398 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1400 If the problematic commit is the most recent commit, and you have not
1401 yet made that commit public, then you may just
1402 <<undoing-a-merge,destroy it using git-reset>>.
1404 Alternatively, you
1405 can edit the working directory and update the index to fix your
1406 mistake, just as if you were going to <<how-to-make-a-commit,create a
1407 new commit>>, then run
1409 -------------------------------------------------
1410 $ git commit --amend
1411 -------------------------------------------------
1413 which will replace the old commit by a new commit incorporating your
1414 changes, giving you a chance to edit the old commit message first.
1416 Again, you should never do this to a commit that may already have
1417 been merged into another branch; use gitlink:git-revert[1] instead in
1418 that case.
1420 It is also possible to edit commits further back in the history, but
1421 this is an advanced topic to be left for
1422 <<cleaning-up-history,another chapter>>.
1424 [[checkout-of-path]]
1425 Checking out an old version of a file
1426 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1428 In the process of undoing a previous bad change, you may find it
1429 useful to check out an older version of a particular file using
1430 gitlink:git-checkout[1]. We've used git checkout before to switch
1431 branches, but it has quite different behavior if it is given a path
1432 name: the command
1434 -------------------------------------------------
1435 $ git checkout HEAD^ path/to/file
1436 -------------------------------------------------
1438 replaces path/to/file by the contents it had in the commit HEAD^, and
1439 also updates the index to match. It does not change branches.
1441 If you just want to look at an old version of the file, without
1442 modifying the working directory, you can do that with
1443 gitlink:git-show[1]:
1445 -------------------------------------------------
1446 $ git show HEAD^:path/to/file
1447 -------------------------------------------------
1449 which will display the given version of the file.
1451 [[ensuring-good-performance]]
1452 Ensuring good performance
1453 -------------------------
1455 On large repositories, git depends on compression to keep the history
1456 information from taking up to much space on disk or in memory.
1458 This compression is not performed automatically. Therefore you
1459 should occasionally run gitlink:git-gc[1]:
1461 -------------------------------------------------
1462 $ git gc
1463 -------------------------------------------------
1465 to recompress the archive. This can be very time-consuming, so
1466 you may prefer to run git-gc when you are not doing other work.
1469 [[ensuring-reliability]]
1470 Ensuring reliability
1471 --------------------
1473 [[checking-for-corruption]]
1474 Checking the repository for corruption
1475 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1477 The gitlink:git-fsck[1] command runs a number of self-consistency checks
1478 on the repository, and reports on any problems. This may take some
1479 time. The most common warning by far is about "dangling" objects:
1481 -------------------------------------------------
1482 $ git fsck
1483 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1484 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1485 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1486 dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1487 dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1488 dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1489 dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1490 dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1491 ...
1492 -------------------------------------------------
1494 Dangling objects are not a problem. At worst they may take up a little
1495 extra disk space. They can sometimes provide a last-resort method for
1496 recovering lost work--see <<dangling-objects>> for details. However, if
1497 you wish, you can remove them with gitlink:git-prune[1] or the --prune
1498 option to gitlink:git-gc[1]:
1500 -------------------------------------------------
1501 $ git gc --prune
1502 -------------------------------------------------
1504 This may be time-consuming. Unlike most other git operations (including
1505 git-gc when run without any options), it is not safe to prune while
1506 other git operations are in progress in the same repository.
1508 [[recovering-lost-changes]]
1509 Recovering lost changes
1510 ~~~~~~~~~~~~~~~~~~~~~~~
1512 [[reflogs]]
1513 Reflogs
1514 ^^^^^^^
1516 Say you modify a branch with gitlink:git-reset[1] --hard, and then
1517 realize that the branch was the only reference you had to that point in
1518 history.
1520 Fortunately, git also keeps a log, called a "reflog", of all the
1521 previous values of each branch. So in this case you can still find the
1522 old history using, for example,
1524 -------------------------------------------------
1525 $ git log master@{1}
1526 -------------------------------------------------
1528 This lists the commits reachable from the previous version of the head.
1529 This syntax can be used to with any git command that accepts a commit,
1530 not just with git log. Some other examples:
1532 -------------------------------------------------
1533 $ git show master@{2} # See where the branch pointed 2,
1534 $ git show master@{3} # 3, ... changes ago.
1535 $ gitk master@{yesterday} # See where it pointed yesterday,
1536 $ gitk master@{"1 week ago"} # ... or last week
1537 $ git log --walk-reflogs master # show reflog entries for master
1538 -------------------------------------------------
1540 A separate reflog is kept for the HEAD, so
1542 -------------------------------------------------
1543 $ git show HEAD@{"1 week ago"}
1544 -------------------------------------------------
1546 will show what HEAD pointed to one week ago, not what the current branch
1547 pointed to one week ago. This allows you to see the history of what
1548 you've checked out.
1550 The reflogs are kept by default for 30 days, after which they may be
1551 pruned. See gitlink:git-reflog[1] and gitlink:git-gc[1] to learn
1552 how to control this pruning, and see the "SPECIFYING REVISIONS"
1553 section of gitlink:git-rev-parse[1] for details.
1555 Note that the reflog history is very different from normal git history.
1556 While normal history is shared by every repository that works on the
1557 same project, the reflog history is not shared: it tells you only about
1558 how the branches in your local repository have changed over time.
1560 [[dangling-object-recovery]]
1561 Examining dangling objects
1562 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1564 In some situations the reflog may not be able to save you. For example,
1565 suppose you delete a branch, then realize you need the history it
1566 contained. The reflog is also deleted; however, if you have not yet
1567 pruned the repository, then you may still be able to find the lost
1568 commits in the dangling objects that git-fsck reports. See
1569 <<dangling-objects>> for the details.
1571 -------------------------------------------------
1572 $ git fsck
1573 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1574 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1575 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1576 ...
1577 -------------------------------------------------
1579 You can examine
1580 one of those dangling commits with, for example,
1582 ------------------------------------------------
1583 $ gitk 7281251ddd --not --all
1584 ------------------------------------------------
1586 which does what it sounds like: it says that you want to see the commit
1587 history that is described by the dangling commit(s), but not the
1588 history that is described by all your existing branches and tags. Thus
1589 you get exactly the history reachable from that commit that is lost.
1590 (And notice that it might not be just one commit: we only report the
1591 "tip of the line" as being dangling, but there might be a whole deep
1592 and complex commit history that was dropped.)
1594 If you decide you want the history back, you can always create a new
1595 reference pointing to it, for example, a new branch:
1597 ------------------------------------------------
1598 $ git branch recovered-branch 7281251ddd
1599 ------------------------------------------------
1601 Other types of dangling objects (blobs and trees) are also possible, and
1602 dangling objects can arise in other situations.
1605 [[sharing-development]]
1606 Sharing development with others
1607 ===============================
1609 [[getting-updates-with-git-pull]]
1610 Getting updates with git pull
1611 -----------------------------
1613 After you clone a repository and make a few changes of your own, you
1614 may wish to check the original repository for updates and merge them
1615 into your own work.
1617 We have already seen <<Updating-a-repository-with-git-fetch,how to
1618 keep remote tracking branches up to date>> with gitlink:git-fetch[1],
1619 and how to merge two branches. So you can merge in changes from the
1620 original repository's master branch with:
1622 -------------------------------------------------
1623 $ git fetch
1624 $ git merge origin/master
1625 -------------------------------------------------
1627 However, the gitlink:git-pull[1] command provides a way to do this in
1628 one step:
1630 -------------------------------------------------
1631 $ git pull origin master
1632 -------------------------------------------------
1634 In fact, "origin" is normally the default repository to pull from,
1635 and the default branch is normally the HEAD of the remote repository,
1636 so often you can accomplish the above with just
1638 -------------------------------------------------
1639 $ git pull
1640 -------------------------------------------------
1642 See the descriptions of the branch.<name>.remote and branch.<name>.merge
1643 options in gitlink:git-config[1] to learn how to control these defaults
1644 depending on the current branch. Also note that the --track option to
1645 gitlink:git-branch[1] and gitlink:git-checkout[1] can be used to
1646 automatically set the default remote branch to pull from at the time
1647 that a branch is created:
1649 -------------------------------------------------
1650 $ git checkout --track -b maint origin/maint
1651 -------------------------------------------------
1653 In addition to saving you keystrokes, "git pull" also helps you by
1654 producing a default commit message documenting the branch and
1655 repository that you pulled from.
1657 (But note that no such commit will be created in the case of a
1658 <<fast-forwards,fast forward>>; instead, your branch will just be
1659 updated to point to the latest commit from the upstream branch.)
1661 The git-pull command can also be given "." as the "remote" repository,
1662 in which case it just merges in a branch from the current repository; so
1663 the commands
1665 -------------------------------------------------
1666 $ git pull . branch
1667 $ git merge branch
1668 -------------------------------------------------
1670 are roughly equivalent. The former is actually very commonly used.
1672 [[submitting-patches]]
1673 Submitting patches to a project
1674 -------------------------------
1676 If you just have a few changes, the simplest way to submit them may
1677 just be to send them as patches in email:
1679 First, use gitlink:git-format-patch[1]; for example:
1681 -------------------------------------------------
1682 $ git format-patch origin
1683 -------------------------------------------------
1685 will produce a numbered series of files in the current directory, one
1686 for each patch in the current branch but not in origin/HEAD.
1688 You can then import these into your mail client and send them by
1689 hand. However, if you have a lot to send at once, you may prefer to
1690 use the gitlink:git-send-email[1] script to automate the process.
1691 Consult the mailing list for your project first to determine how they
1692 prefer such patches be handled.
1694 [[importing-patches]]
1695 Importing patches to a project
1696 ------------------------------
1698 Git also provides a tool called gitlink:git-am[1] (am stands for
1699 "apply mailbox"), for importing such an emailed series of patches.
1700 Just save all of the patch-containing messages, in order, into a
1701 single mailbox file, say "patches.mbox", then run
1703 -------------------------------------------------
1704 $ git am -3 patches.mbox
1705 -------------------------------------------------
1707 Git will apply each patch in order; if any conflicts are found, it
1708 will stop, and you can fix the conflicts as described in
1709 "<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1710 git to perform a merge; if you would prefer it just to abort and
1711 leave your tree and index untouched, you may omit that option.)
1713 Once the index is updated with the results of the conflict
1714 resolution, instead of creating a new commit, just run
1716 -------------------------------------------------
1717 $ git am --resolved
1718 -------------------------------------------------
1720 and git will create the commit for you and continue applying the
1721 remaining patches from the mailbox.
1723 The final result will be a series of commits, one for each patch in
1724 the original mailbox, with authorship and commit log message each
1725 taken from the message containing each patch.
1727 [[public-repositories]]
1728 Public git repositories
1729 -----------------------
1731 Another way to submit changes to a project is to tell the maintainer of
1732 that project to pull the changes from your repository using git-pull[1].
1733 In the section "<<getting-updates-with-git-pull, Getting updates with
1734 git pull>>" we described this as a way to get updates from the "main"
1735 repository, but it works just as well in the other direction.
1737 If you and the maintainer both have accounts on the same machine, then
1738 you can just pull changes from each other's repositories directly;
1739 commands that accept repository URLs as arguments will also accept a
1740 local directory name:
1742 -------------------------------------------------
1743 $ git clone /path/to/repository
1744 $ git pull /path/to/other/repository
1745 -------------------------------------------------
1747 or an ssh url:
1749 -------------------------------------------------
1750 $ git clone ssh://yourhost/~you/repository
1751 -------------------------------------------------
1753 For projects with few developers, or for synchronizing a few private
1754 repositories, this may be all you need.
1756 However, the more common way to do this is to maintain a separate public
1757 repository (usually on a different host) for others to pull changes
1758 from. This is usually more convenient, and allows you to cleanly
1759 separate private work in progress from publicly visible work.
1761 You will continue to do your day-to-day work in your personal
1762 repository, but periodically "push" changes from your personal
1763 repository into your public repository, allowing other developers to
1764 pull from that repository. So the flow of changes, in a situation
1765 where there is one other developer with a public repository, looks
1766 like this:
1768 you push
1769 your personal repo ------------------> your public repo
1770 ^ |
1771 | |
1772 | you pull | they pull
1773 | |
1774 | |
1775 | they push V
1776 their public repo <------------------- their repo
1778 We explain how to do this in the following sections.
1780 [[setting-up-a-public-repository]]
1781 Setting up a public repository
1782 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1784 Assume your personal repository is in the directory ~/proj. We
1785 first create a new clone of the repository and tell git-daemon that it
1786 is meant to be public:
1788 -------------------------------------------------
1789 $ git clone --bare ~/proj proj.git
1790 $ touch proj.git/git-daemon-export-ok
1791 -------------------------------------------------
1793 The resulting directory proj.git contains a "bare" git repository--it is
1794 just the contents of the ".git" directory, without any files checked out
1795 around it.
1797 Next, copy proj.git to the server where you plan to host the
1798 public repository. You can use scp, rsync, or whatever is most
1799 convenient.
1801 [[exporting-via-git]]
1802 Exporting a git repository via the git protocol
1803 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1805 This is the preferred method.
1807 If someone else administers the server, they should tell you what
1808 directory to put the repository in, and what git:// url it will appear
1809 at. You can then skip to the section
1810 "<<pushing-changes-to-a-public-repository,Pushing changes to a public
1811 repository>>", below.
1813 Otherwise, all you need to do is start gitlink:git-daemon[1]; it will
1814 listen on port 9418. By default, it will allow access to any directory
1815 that looks like a git directory and contains the magic file
1816 git-daemon-export-ok. Passing some directory paths as git-daemon
1817 arguments will further restrict the exports to those paths.
1819 You can also run git-daemon as an inetd service; see the
1820 gitlink:git-daemon[1] man page for details. (See especially the
1821 examples section.)
1823 [[exporting-via-http]]
1824 Exporting a git repository via http
1825 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1827 The git protocol gives better performance and reliability, but on a
1828 host with a web server set up, http exports may be simpler to set up.
1830 All you need to do is place the newly created bare git repository in
1831 a directory that is exported by the web server, and make some
1832 adjustments to give web clients some extra information they need:
1834 -------------------------------------------------
1835 $ mv proj.git /home/you/public_html/proj.git
1836 $ cd proj.git
1837 $ git --bare update-server-info
1838 $ chmod a+x hooks/post-update
1839 -------------------------------------------------
1841 (For an explanation of the last two lines, see
1842 gitlink:git-update-server-info[1], and the documentation
1843 link:hooks.html[Hooks used by git].)
1845 Advertise the url of proj.git. Anybody else should then be able to
1846 clone or pull from that url, for example with a commandline like:
1848 -------------------------------------------------
1849 $ git clone http://yourserver.com/~you/proj.git
1850 -------------------------------------------------
1852 (See also
1853 link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1854 for a slightly more sophisticated setup using WebDAV which also
1855 allows pushing over http.)
1857 [[pushing-changes-to-a-public-repository]]
1858 Pushing changes to a public repository
1859 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1861 Note that the two techniques outlined above (exporting via
1862 <<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1863 maintainers to fetch your latest changes, but they do not allow write
1864 access, which you will need to update the public repository with the
1865 latest changes created in your private repository.
1867 The simplest way to do this is using gitlink:git-push[1] and ssh; to
1868 update the remote branch named "master" with the latest state of your
1869 branch named "master", run
1871 -------------------------------------------------
1872 $ git push ssh://yourserver.com/~you/proj.git master:master
1873 -------------------------------------------------
1875 or just
1877 -------------------------------------------------
1878 $ git push ssh://yourserver.com/~you/proj.git master
1879 -------------------------------------------------
1881 As with git-fetch, git-push will complain if this does not result in
1882 a <<fast-forwards,fast forward>>. Normally this is a sign of
1883 something wrong. However, if you are sure you know what you're
1884 doing, you may force git-push to perform the update anyway by
1885 proceeding the branch name by a plus sign:
1887 -------------------------------------------------
1888 $ git push ssh://yourserver.com/~you/proj.git +master
1889 -------------------------------------------------
1891 Note that the target of a "push" is normally a
1892 <<def_bare_repository,bare>> repository. You can also push to a
1893 repository that has a checked-out working tree, but the working tree
1894 will not be updated by the push. This may lead to unexpected results if
1895 the branch you push to is the currently checked-out branch!
1897 As with git-fetch, you may also set up configuration options to
1898 save typing; so, for example, after
1900 -------------------------------------------------
1901 $ cat >>.git/config <<EOF
1902 [remote "public-repo"]
1903 url = ssh://yourserver.com/~you/proj.git
1904 EOF
1905 -------------------------------------------------
1907 you should be able to perform the above push with just
1909 -------------------------------------------------
1910 $ git push public-repo master
1911 -------------------------------------------------
1913 See the explanations of the remote.<name>.url, branch.<name>.remote,
1914 and remote.<name>.push options in gitlink:git-config[1] for
1915 details.
1917 [[setting-up-a-shared-repository]]
1918 Setting up a shared repository
1919 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1921 Another way to collaborate is by using a model similar to that
1922 commonly used in CVS, where several developers with special rights
1923 all push to and pull from a single shared repository. See
1924 link:cvs-migration.html[git for CVS users] for instructions on how to
1925 set this up.
1927 However, while there is nothing wrong with git's support for shared
1928 repositories, this mode of operation is not generally recommended,
1929 simply because the mode of collaboration that git supports--by
1930 exchanging patches and pulling from public repositories--has so many
1931 advantages over the central shared repository:
1933 - Git's ability to quickly import and merge patches allows a
1934 single maintainer to process incoming changes even at very
1935 high rates. And when that becomes too much, git-pull provides
1936 an easy way for that maintainer to delegate this job to other
1937 maintainers while still allowing optional review of incoming
1938 changes.
1939 - Since every developer's repository has the same complete copy
1940 of the project history, no repository is special, and it is
1941 trivial for another developer to take over maintenance of a
1942 project, either by mutual agreement, or because a maintainer
1943 becomes unresponsive or difficult to work with.
1944 - The lack of a central group of "committers" means there is
1945 less need for formal decisions about who is "in" and who is
1946 "out".
1948 [[setting-up-gitweb]]
1949 Allowing web browsing of a repository
1950 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1952 The gitweb cgi script provides users an easy way to browse your
1953 project's files and history without having to install git; see the file
1954 gitweb/INSTALL in the git source tree for instructions on setting it up.
1956 [[sharing-development-examples]]
1957 Examples
1958 --------
1960 [[maintaining-topic-branches]]
1961 Maintaining topic branches for a Linux subsystem maintainer
1962 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1964 This describes how Tony Luck uses git in his role as maintainer of the
1965 IA64 architecture for the Linux kernel.
1967 He uses two public branches:
1969 - A "test" tree into which patches are initially placed so that they
1970 can get some exposure when integrated with other ongoing development.
1971 This tree is available to Andrew for pulling into -mm whenever he
1972 wants.
1974 - A "release" tree into which tested patches are moved for final sanity
1975 checking, and as a vehicle to send them upstream to Linus (by sending
1976 him a "please pull" request.)
1978 He also uses a set of temporary branches ("topic branches"), each
1979 containing a logical grouping of patches.
1981 To set this up, first create your work tree by cloning Linus's public
1982 tree:
1984 -------------------------------------------------
1985 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git work
1986 $ cd work
1987 -------------------------------------------------
1989 Linus's tree will be stored in the remote branch named origin/master,
1990 and can be updated using gitlink:git-fetch[1]; you can track other
1991 public trees using gitlink:git-remote[1] to set up a "remote" and
1992 git-fetch[1] to keep them up-to-date; see <<repositories-and-branches>>.
1994 Now create the branches in which you are going to work; these start out
1995 at the current tip of origin/master branch, and should be set up (using
1996 the --track option to gitlink:git-branch[1]) to merge changes in from
1997 Linus by default.
1999 -------------------------------------------------
2000 $ git branch --track test origin/master
2001 $ git branch --track release origin/master
2002 -------------------------------------------------
2004 These can be easily kept up to date using gitlink:git-pull[1]
2006 -------------------------------------------------
2007 $ git checkout test && git pull
2008 $ git checkout release && git pull
2009 -------------------------------------------------
2011 Important note! If you have any local changes in these branches, then
2012 this merge will create a commit object in the history (with no local
2013 changes git will simply do a "Fast forward" merge). Many people dislike
2014 the "noise" that this creates in the Linux history, so you should avoid
2015 doing this capriciously in the "release" branch, as these noisy commits
2016 will become part of the permanent history when you ask Linus to pull
2017 from the release branch.
2019 A few configuration variables (see gitlink:git-config[1]) can
2020 make it easy to push both branches to your public tree. (See
2021 <<setting-up-a-public-repository>>.)
2023 -------------------------------------------------
2024 $ cat >> .git/config <<EOF
2025 [remote "mytree"]
2026 url = master.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6.git
2027 push = release
2028 push = test
2029 EOF
2030 -------------------------------------------------
2032 Then you can push both the test and release trees using
2033 gitlink:git-push[1]:
2035 -------------------------------------------------
2036 $ git push mytree
2037 -------------------------------------------------
2039 or push just one of the test and release branches using:
2041 -------------------------------------------------
2042 $ git push mytree test
2043 -------------------------------------------------
2045 or
2047 -------------------------------------------------
2048 $ git push mytree release
2049 -------------------------------------------------
2051 Now to apply some patches from the community. Think of a short
2052 snappy name for a branch to hold this patch (or related group of
2053 patches), and create a new branch from the current tip of Linus's
2054 branch:
2056 -------------------------------------------------
2057 $ git checkout -b speed-up-spinlocks origin
2058 -------------------------------------------------
2060 Now you apply the patch(es), run some tests, and commit the change(s). If
2061 the patch is a multi-part series, then you should apply each as a separate
2062 commit to this branch.
2064 -------------------------------------------------
2065 $ ... patch ... test ... commit [ ... patch ... test ... commit ]*
2066 -------------------------------------------------
2068 When you are happy with the state of this change, you can pull it into the
2069 "test" branch in preparation to make it public:
2071 -------------------------------------------------
2072 $ git checkout test && git pull . speed-up-spinlocks
2073 -------------------------------------------------
2075 It is unlikely that you would have any conflicts here ... but you might if you
2076 spent a while on this step and had also pulled new versions from upstream.
2078 Some time later when enough time has passed and testing done, you can pull the
2079 same branch into the "release" tree ready to go upstream. This is where you
2080 see the value of keeping each patch (or patch series) in its own branch. It
2081 means that the patches can be moved into the "release" tree in any order.
2083 -------------------------------------------------
2084 $ git checkout release && git pull . speed-up-spinlocks
2085 -------------------------------------------------
2087 After a while, you will have a number of branches, and despite the
2088 well chosen names you picked for each of them, you may forget what
2089 they are for, or what status they are in. To get a reminder of what
2090 changes are in a specific branch, use:
2092 -------------------------------------------------
2093 $ git log linux..branchname | git-shortlog
2094 -------------------------------------------------
2096 To see whether it has already been merged into the test or release branches
2097 use:
2099 -------------------------------------------------
2100 $ git log test..branchname
2101 -------------------------------------------------
2103 or
2105 -------------------------------------------------
2106 $ git log release..branchname
2107 -------------------------------------------------
2109 (If this branch has not yet been merged you will see some log entries.
2110 If it has been merged, then there will be no output.)
2112 Once a patch completes the great cycle (moving from test to release,
2113 then pulled by Linus, and finally coming back into your local
2114 "origin/master" branch) the branch for this change is no longer needed.
2115 You detect this when the output from:
2117 -------------------------------------------------
2118 $ git log origin..branchname
2119 -------------------------------------------------
2121 is empty. At this point the branch can be deleted:
2123 -------------------------------------------------
2124 $ git branch -d branchname
2125 -------------------------------------------------
2127 Some changes are so trivial that it is not necessary to create a separate
2128 branch and then merge into each of the test and release branches. For
2129 these changes, just apply directly to the "release" branch, and then
2130 merge that into the "test" branch.
2132 To create diffstat and shortlog summaries of changes to include in a "please
2133 pull" request to Linus you can use:
2135 -------------------------------------------------
2136 $ git diff --stat origin..release
2137 -------------------------------------------------
2139 and
2141 -------------------------------------------------
2142 $ git log -p origin..release | git shortlog
2143 -------------------------------------------------
2145 Here are some of the scripts that simplify all this even further.
2147 -------------------------------------------------
2148 ==== update script ====
2149 # Update a branch in my GIT tree. If the branch to be updated
2150 # is origin, then pull from kernel.org. Otherwise merge
2151 # origin/master branch into test|release branch
2153 case "$1" in
2154 test|release)
2155 git checkout $1 && git pull . origin
2156 ;;
2157 origin)
2158 before=$(cat .git/refs/remotes/origin/master)
2159 git fetch origin
2160 after=$(cat .git/refs/remotes/origin/master)
2161 if [ $before != $after ]
2162 then
2163 git log $before..$after | git shortlog
2164 fi
2165 ;;
2166 *)
2167 echo "Usage: $0 origin|test|release" 1>&2
2168 exit 1
2169 ;;
2170 esac
2171 -------------------------------------------------
2173 -------------------------------------------------
2174 ==== merge script ====
2175 # Merge a branch into either the test or release branch
2177 pname=$0
2179 usage()
2180 {
2181 echo "Usage: $pname branch test|release" 1>&2
2182 exit 1
2183 }
2185 if [ ! -f .git/refs/heads/"$1" ]
2186 then
2187 echo "Can't see branch <$1>" 1>&2
2188 usage
2189 fi
2191 case "$2" in
2192 test|release)
2193 if [ $(git log $2..$1 | wc -c) -eq 0 ]
2194 then
2195 echo $1 already merged into $2 1>&2
2196 exit 1
2197 fi
2198 git checkout $2 && git pull . $1
2199 ;;
2200 *)
2201 usage
2202 ;;
2203 esac
2204 -------------------------------------------------
2206 -------------------------------------------------
2207 ==== status script ====
2208 # report on status of my ia64 GIT tree
2210 gb=$(tput setab 2)
2211 rb=$(tput setab 1)
2212 restore=$(tput setab 9)
2214 if [ `git rev-list test..release | wc -c` -gt 0 ]
2215 then
2216 echo $rb Warning: commits in release that are not in test $restore
2217 git log test..release
2218 fi
2220 for branch in `ls .git/refs/heads`
2221 do
2222 if [ $branch = test -o $branch = release ]
2223 then
2224 continue
2225 fi
2227 echo -n $gb ======= $branch ====== $restore " "
2228 status=
2229 for ref in test release origin/master
2230 do
2231 if [ `git rev-list $ref..$branch | wc -c` -gt 0 ]
2232 then
2233 status=$status${ref:0:1}
2234 fi
2235 done
2236 case $status in
2237 trl)
2238 echo $rb Need to pull into test $restore
2239 ;;
2240 rl)
2241 echo "In test"
2242 ;;
2243 l)
2244 echo "Waiting for linus"
2245 ;;
2246 "")
2247 echo $rb All done $restore
2248 ;;
2249 *)
2250 echo $rb "<$status>" $restore
2251 ;;
2252 esac
2253 git log origin/master..$branch | git shortlog
2254 done
2255 -------------------------------------------------
2258 [[cleaning-up-history]]
2259 Rewriting history and maintaining patch series
2260 ==============================================
2262 Normally commits are only added to a project, never taken away or
2263 replaced. Git is designed with this assumption, and violating it will
2264 cause git's merge machinery (for example) to do the wrong thing.
2266 However, there is a situation in which it can be useful to violate this
2267 assumption.
2269 [[patch-series]]
2270 Creating the perfect patch series
2271 ---------------------------------
2273 Suppose you are a contributor to a large project, and you want to add a
2274 complicated feature, and to present it to the other developers in a way
2275 that makes it easy for them to read your changes, verify that they are
2276 correct, and understand why you made each change.
2278 If you present all of your changes as a single patch (or commit), they
2279 may find that it is too much to digest all at once.
2281 If you present them with the entire history of your work, complete with
2282 mistakes, corrections, and dead ends, they may be overwhelmed.
2284 So the ideal is usually to produce a series of patches such that:
2286 1. Each patch can be applied in order.
2288 2. Each patch includes a single logical change, together with a
2289 message explaining the change.
2291 3. No patch introduces a regression: after applying any initial
2292 part of the series, the resulting project still compiles and
2293 works, and has no bugs that it didn't have before.
2295 4. The complete series produces the same end result as your own
2296 (probably much messier!) development process did.
2298 We will introduce some tools that can help you do this, explain how to
2299 use them, and then explain some of the problems that can arise because
2300 you are rewriting history.
2302 [[using-git-rebase]]
2303 Keeping a patch series up to date using git-rebase
2304 --------------------------------------------------
2306 Suppose that you create a branch "mywork" on a remote-tracking branch
2307 "origin", and create some commits on top of it:
2309 -------------------------------------------------
2310 $ git checkout -b mywork origin
2311 $ vi file.txt
2312 $ git commit
2313 $ vi otherfile.txt
2314 $ git commit
2315 ...
2316 -------------------------------------------------
2318 You have performed no merges into mywork, so it is just a simple linear
2319 sequence of patches on top of "origin":
2321 ................................................
2322 o--o--o <-- origin
2323 \
2324 o--o--o <-- mywork
2325 ................................................
2327 Some more interesting work has been done in the upstream project, and
2328 "origin" has advanced:
2330 ................................................
2331 o--o--O--o--o--o <-- origin
2332 \
2333 a--b--c <-- mywork
2334 ................................................
2336 At this point, you could use "pull" to merge your changes back in;
2337 the result would create a new merge commit, like this:
2339 ................................................
2340 o--o--O--o--o--o <-- origin
2341 \ \
2342 a--b--c--m <-- mywork
2343 ................................................
2345 However, if you prefer to keep the history in mywork a simple series of
2346 commits without any merges, you may instead choose to use
2347 gitlink:git-rebase[1]:
2349 -------------------------------------------------
2350 $ git checkout mywork
2351 $ git rebase origin
2352 -------------------------------------------------
2354 This will remove each of your commits from mywork, temporarily saving
2355 them as patches (in a directory named ".dotest"), update mywork to
2356 point at the latest version of origin, then apply each of the saved
2357 patches to the new mywork. The result will look like:
2360 ................................................
2361 o--o--O--o--o--o <-- origin
2362 \
2363 a'--b'--c' <-- mywork
2364 ................................................
2366 In the process, it may discover conflicts. In that case it will stop
2367 and allow you to fix the conflicts; after fixing conflicts, use "git
2368 add" to update the index with those contents, and then, instead of
2369 running git-commit, just run
2371 -------------------------------------------------
2372 $ git rebase --continue
2373 -------------------------------------------------
2375 and git will continue applying the rest of the patches.
2377 At any point you may use the --abort option to abort this process and
2378 return mywork to the state it had before you started the rebase:
2380 -------------------------------------------------
2381 $ git rebase --abort
2382 -------------------------------------------------
2384 [[modifying-one-commit]]
2385 Modifying a single commit
2386 -------------------------
2388 We saw in <<fixing-a-mistake-by-editing-history>> that you can replace the
2389 most recent commit using
2391 -------------------------------------------------
2392 $ git commit --amend
2393 -------------------------------------------------
2395 which will replace the old commit by a new commit incorporating your
2396 changes, giving you a chance to edit the old commit message first.
2398 You can also use a combination of this and gitlink:git-rebase[1] to edit
2399 commits further back in your history. First, tag the problematic commit with
2401 -------------------------------------------------
2402 $ git tag bad mywork~5
2403 -------------------------------------------------
2405 (Either gitk or git-log may be useful for finding the commit.)
2407 Then check out that commit, edit it, and rebase the rest of the series
2408 on top of it (note that we could check out the commit on a temporary
2409 branch, but instead we're using a <<detached-head,detached head>>):
2411 -------------------------------------------------
2412 $ git checkout bad
2413 $ # make changes here and update the index
2414 $ git commit --amend
2415 $ git rebase --onto HEAD bad mywork
2416 -------------------------------------------------
2418 When you're done, you'll be left with mywork checked out, with the top
2419 patches on mywork reapplied on top of your modified commit. You can
2420 then clean up with
2422 -------------------------------------------------
2423 $ git tag -d bad
2424 -------------------------------------------------
2426 Note that the immutable nature of git history means that you haven't really
2427 "modified" existing commits; instead, you have replaced the old commits with
2428 new commits having new object names.
2430 [[reordering-patch-series]]
2431 Reordering or selecting from a patch series
2432 -------------------------------------------
2434 Given one existing commit, the gitlink:git-cherry-pick[1] command
2435 allows you to apply the change introduced by that commit and create a
2436 new commit that records it. So, for example, if "mywork" points to a
2437 series of patches on top of "origin", you might do something like:
2439 -------------------------------------------------
2440 $ git checkout -b mywork-new origin
2441 $ gitk origin..mywork &
2442 -------------------------------------------------
2444 And browse through the list of patches in the mywork branch using gitk,
2445 applying them (possibly in a different order) to mywork-new using
2446 cherry-pick, and possibly modifying them as you go using commit
2447 --amend.
2449 Another technique is to use git-format-patch to create a series of
2450 patches, then reset the state to before the patches:
2452 -------------------------------------------------
2453 $ git format-patch origin
2454 $ git reset --hard origin
2455 -------------------------------------------------
2457 Then modify, reorder, or eliminate patches as preferred before applying
2458 them again with gitlink:git-am[1].
2460 [[patch-series-tools]]
2461 Other tools
2462 -----------
2464 There are numerous other tools, such as stgit, which exist for the
2465 purpose of maintaining a patch series. These are outside of the scope of
2466 this manual.
2468 [[problems-with-rewriting-history]]
2469 Problems with rewriting history
2470 -------------------------------
2472 The primary problem with rewriting the history of a branch has to do
2473 with merging. Suppose somebody fetches your branch and merges it into
2474 their branch, with a result something like this:
2476 ................................................
2477 o--o--O--o--o--o <-- origin
2478 \ \
2479 t--t--t--m <-- their branch:
2480 ................................................
2482 Then suppose you modify the last three commits:
2484 ................................................
2485 o--o--o <-- new head of origin
2486 /
2487 o--o--O--o--o--o <-- old head of origin
2488 ................................................
2490 If we examined all this history together in one repository, it will
2491 look like:
2493 ................................................
2494 o--o--o <-- new head of origin
2495 /
2496 o--o--O--o--o--o <-- old head of origin
2497 \ \
2498 t--t--t--m <-- their branch:
2499 ................................................
2501 Git has no way of knowing that the new head is an updated version of
2502 the old head; it treats this situation exactly the same as it would if
2503 two developers had independently done the work on the old and new heads
2504 in parallel. At this point, if someone attempts to merge the new head
2505 in to their branch, git will attempt to merge together the two (old and
2506 new) lines of development, instead of trying to replace the old by the
2507 new. The results are likely to be unexpected.
2509 You may still choose to publish branches whose history is rewritten,
2510 and it may be useful for others to be able to fetch those branches in
2511 order to examine or test them, but they should not attempt to pull such
2512 branches into their own work.
2514 For true distributed development that supports proper merging,
2515 published branches should never be rewritten.
2517 [[advanced-branch-management]]
2518 Advanced branch management
2519 ==========================
2521 [[fetching-individual-branches]]
2522 Fetching individual branches
2523 ----------------------------
2525 Instead of using gitlink:git-remote[1], you can also choose just
2526 to update one branch at a time, and to store it locally under an
2527 arbitrary name:
2529 -------------------------------------------------
2530 $ git fetch origin todo:my-todo-work
2531 -------------------------------------------------
2533 The first argument, "origin", just tells git to fetch from the
2534 repository you originally cloned from. The second argument tells git
2535 to fetch the branch named "todo" from the remote repository, and to
2536 store it locally under the name refs/heads/my-todo-work.
2538 You can also fetch branches from other repositories; so
2540 -------------------------------------------------
2541 $ git fetch git://example.com/proj.git master:example-master
2542 -------------------------------------------------
2544 will create a new branch named "example-master" and store in it the
2545 branch named "master" from the repository at the given URL. If you
2546 already have a branch named example-master, it will attempt to
2547 <<fast-forwards,fast-forward>> to the commit given by example.com's
2548 master branch. In more detail:
2550 [[fetch-fast-forwards]]
2551 git fetch and fast-forwards
2552 ---------------------------
2554 In the previous example, when updating an existing branch, "git
2555 fetch" checks to make sure that the most recent commit on the remote
2556 branch is a descendant of the most recent commit on your copy of the
2557 branch before updating your copy of the branch to point at the new
2558 commit. Git calls this process a <<fast-forwards,fast forward>>.
2560 A fast forward looks something like this:
2562 ................................................
2563 o--o--o--o <-- old head of the branch
2564 \
2565 o--o--o <-- new head of the branch
2566 ................................................
2569 In some cases it is possible that the new head will *not* actually be
2570 a descendant of the old head. For example, the developer may have
2571 realized she made a serious mistake, and decided to backtrack,
2572 resulting in a situation like:
2574 ................................................
2575 o--o--o--o--a--b <-- old head of the branch
2576 \
2577 o--o--o <-- new head of the branch
2578 ................................................
2580 In this case, "git fetch" will fail, and print out a warning.
2582 In that case, you can still force git to update to the new head, as
2583 described in the following section. However, note that in the
2584 situation above this may mean losing the commits labeled "a" and "b",
2585 unless you've already created a reference of your own pointing to
2586 them.
2588 [[forcing-fetch]]
2589 Forcing git fetch to do non-fast-forward updates
2590 ------------------------------------------------
2592 If git fetch fails because the new head of a branch is not a
2593 descendant of the old head, you may force the update with:
2595 -------------------------------------------------
2596 $ git fetch git://example.com/proj.git +master:refs/remotes/example/master
2597 -------------------------------------------------
2599 Note the addition of the "+" sign. Alternatively, you can use the "-f"
2600 flag to force updates of all the fetched branches, as in:
2602 -------------------------------------------------
2603 $ git fetch -f origin
2604 -------------------------------------------------
2606 Be aware that commits that the old version of example/master pointed at
2607 may be lost, as we saw in the previous section.
2609 [[remote-branch-configuration]]
2610 Configuring remote branches
2611 ---------------------------
2613 We saw above that "origin" is just a shortcut to refer to the
2614 repository that you originally cloned from. This information is
2615 stored in git configuration variables, which you can see using
2616 gitlink:git-config[1]:
2618 -------------------------------------------------
2619 $ git config -l
2620 core.repositoryformatversion=0
2621 core.filemode=true
2622 core.logallrefupdates=true
2623 remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2624 remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2625 branch.master.remote=origin
2626 branch.master.merge=refs/heads/master
2627 -------------------------------------------------
2629 If there are other repositories that you also use frequently, you can
2630 create similar configuration options to save typing; for example,
2631 after
2633 -------------------------------------------------
2634 $ git config remote.example.url git://example.com/proj.git
2635 -------------------------------------------------
2637 then the following two commands will do the same thing:
2639 -------------------------------------------------
2640 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2641 $ git fetch example master:refs/remotes/example/master
2642 -------------------------------------------------
2644 Even better, if you add one more option:
2646 -------------------------------------------------
2647 $ git config remote.example.fetch master:refs/remotes/example/master
2648 -------------------------------------------------
2650 then the following commands will all do the same thing:
2652 -------------------------------------------------
2653 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2654 $ git fetch example master:refs/remotes/example/master
2655 $ git fetch example
2656 -------------------------------------------------
2658 You can also add a "+" to force the update each time:
2660 -------------------------------------------------
2661 $ git config remote.example.fetch +master:ref/remotes/example/master
2662 -------------------------------------------------
2664 Don't do this unless you're sure you won't mind "git fetch" possibly
2665 throwing away commits on mybranch.
2667 Also note that all of the above configuration can be performed by
2668 directly editing the file .git/config instead of using
2669 gitlink:git-config[1].
2671 See gitlink:git-config[1] for more details on the configuration
2672 options mentioned above.
2675 [[git-internals]]
2676 Git internals
2677 =============
2679 Git depends on two fundamental abstractions: the "object database", and
2680 the "current directory cache" aka "index".
2682 [[the-object-database]]
2683 The Object Database
2684 -------------------
2686 The object database is literally just a content-addressable collection
2687 of objects. All objects are named by their content, which is
2688 approximated by the SHA1 hash of the object itself. Objects may refer
2689 to other objects (by referencing their SHA1 hash), and so you can
2690 build up a hierarchy of objects.
2692 All objects have a statically determined "type" which is
2693 determined at object creation time, and which identifies the format of
2694 the object (i.e. how it is used, and how it can refer to other
2695 objects). There are currently four different object types: "blob",
2696 "tree", "commit", and "tag".
2698 A <<def_blob_object,"blob" object>> cannot refer to any other object,
2699 and is, as the name implies, a pure storage object containing some
2700 user data. It is used to actually store the file data, i.e. a blob
2701 object is associated with some particular version of some file.
2703 A <<def_tree_object,"tree" object>> is an object that ties one or more
2704 "blob" objects into a directory structure. In addition, a tree object
2705 can refer to other tree objects, thus creating a directory hierarchy.
2707 A <<def_commit_object,"commit" object>> ties such directory hierarchies
2708 together into a <<def_DAG,directed acyclic graph>> of revisions - each
2709 "commit" is associated with exactly one tree (the directory hierarchy at
2710 the time of the commit). In addition, a "commit" refers to one or more
2711 "parent" commit objects that describe the history of how we arrived at
2712 that directory hierarchy.
2714 As a special case, a commit object with no parents is called the "root"
2715 commit, and is the point of an initial project commit. Each project
2716 must have at least one root, and while you can tie several different
2717 root objects together into one project by creating a commit object which
2718 has two or more separate roots as its ultimate parents, that's probably
2719 just going to confuse people. So aim for the notion of "one root object
2720 per project", even if git itself does not enforce that.
2722 A <<def_tag_object,"tag" object>> symbolically identifies and can be
2723 used to sign other objects. It contains the identifier and type of
2724 another object, a symbolic name (of course!) and, optionally, a
2725 signature.
2727 Regardless of object type, all objects share the following
2728 characteristics: they are all deflated with zlib, and have a header
2729 that not only specifies their type, but also provides size information
2730 about the data in the object. It's worth noting that the SHA1 hash
2731 that is used to name the object is the hash of the original data
2732 plus this header, so `sha1sum` 'file' does not match the object name
2733 for 'file'.
2734 (Historical note: in the dawn of the age of git the hash
2735 was the sha1 of the 'compressed' object.)
2737 As a result, the general consistency of an object can always be tested
2738 independently of the contents or the type of the object: all objects can
2739 be validated by verifying that (a) their hashes match the content of the
2740 file and (b) the object successfully inflates to a stream of bytes that
2741 forms a sequence of <ascii type without space> {plus} <space> {plus} <ascii decimal
2742 size> {plus} <byte\0> {plus} <binary object data>.
2744 The structured objects can further have their structure and
2745 connectivity to other objects verified. This is generally done with
2746 the `git-fsck` program, which generates a full dependency graph
2747 of all objects, and verifies their internal consistency (in addition
2748 to just verifying their superficial consistency through the hash).
2750 The object types in some more detail:
2752 [[blob-object]]
2753 Blob Object
2754 -----------
2756 A "blob" object is nothing but a binary blob of data, and doesn't
2757 refer to anything else. There is no signature or any other
2758 verification of the data, so while the object is consistent (it 'is'
2759 indexed by its sha1 hash, so the data itself is certainly correct), it
2760 has absolutely no other attributes. No name associations, no
2761 permissions. It is purely a blob of data (i.e. normally "file
2762 contents").
2764 In particular, since the blob is entirely defined by its data, if two
2765 files in a directory tree (or in multiple different versions of the
2766 repository) have the same contents, they will share the same blob
2767 object. The object is totally independent of its location in the
2768 directory tree, and renaming a file does not change the object that
2769 file is associated with in any way.
2771 A blob is typically created when gitlink:git-update-index[1]
2772 is run, and its data can be accessed by gitlink:git-cat-file[1].
2774 [[tree-object]]
2775 Tree Object
2776 -----------
2778 The next hierarchical object type is the "tree" object. A tree object
2779 is a list of mode/name/blob data, sorted by name. Alternatively, the
2780 mode data may specify a directory mode, in which case instead of
2781 naming a blob, that name is associated with another TREE object.
2783 Like the "blob" object, a tree object is uniquely determined by the
2784 set contents, and so two separate but identical trees will always
2785 share the exact same object. This is true at all levels, i.e. it's
2786 true for a "leaf" tree (which does not refer to any other trees, only
2787 blobs) as well as for a whole subdirectory.
2789 For that reason a "tree" object is just a pure data abstraction: it
2790 has no history, no signatures, no verification of validity, except
2791 that since the contents are again protected by the hash itself, we can
2792 trust that the tree is immutable and its contents never change.
2794 So you can trust the contents of a tree to be valid, the same way you
2795 can trust the contents of a blob, but you don't know where those
2796 contents 'came' from.
2798 Side note on trees: since a "tree" object is a sorted list of
2799 "filename+content", you can create a diff between two trees without
2800 actually having to unpack two trees. Just ignore all common parts,
2801 and your diff will look right. In other words, you can effectively
2802 (and efficiently) tell the difference between any two random trees by
2803 O(n) where "n" is the size of the difference, rather than the size of
2804 the tree.
2806 Side note 2 on trees: since the name of a "blob" depends entirely and
2807 exclusively on its contents (i.e. there are no names or permissions
2808 involved), you can see trivial renames or permission changes by
2809 noticing that the blob stayed the same. However, renames with data
2810 changes need a smarter "diff" implementation.
2812 A tree is created with gitlink:git-write-tree[1] and
2813 its data can be accessed by gitlink:git-ls-tree[1].
2814 Two trees can be compared with gitlink:git-diff-tree[1].
2816 [[commit-object]]
2817 Commit Object
2818 -------------
2820 The "commit" object is an object that introduces the notion of
2821 history into the picture. In contrast to the other objects, it
2822 doesn't just describe the physical state of a tree, it describes how
2823 we got there, and why.
2825 A "commit" is defined by the tree-object that it results in, the
2826 parent commits (zero, one or more) that led up to that point, and a
2827 comment on what happened. Again, a commit is not trusted per se:
2828 the contents are well-defined and "safe" due to the cryptographically
2829 strong signatures at all levels, but there is no reason to believe
2830 that the tree is "good" or that the merge information makes sense.
2831 The parents do not have to actually have any relationship with the
2832 result, for example.
2834 Note on commits: unlike some SCM's, commits do not contain
2835 rename information or file mode change information. All of that is
2836 implicit in the trees involved (the result tree, and the result trees
2837 of the parents), and describing that makes no sense in this idiotic
2838 file manager.
2840 A commit is created with gitlink:git-commit-tree[1] and
2841 its data can be accessed by gitlink:git-cat-file[1].
2843 [[trust]]
2844 Trust
2845 -----
2847 An aside on the notion of "trust". Trust is really outside the scope
2848 of "git", but it's worth noting a few things. First off, since
2849 everything is hashed with SHA1, you 'can' trust that an object is
2850 intact and has not been messed with by external sources. So the name
2851 of an object uniquely identifies a known state - just not a state that
2852 you may want to trust.
2854 Furthermore, since the SHA1 signature of a commit refers to the
2855 SHA1 signatures of the tree it is associated with and the signatures
2856 of the parent, a single named commit specifies uniquely a whole set
2857 of history, with full contents. You can't later fake any step of the
2858 way once you have the name of a commit.
2860 So to introduce some real trust in the system, the only thing you need
2861 to do is to digitally sign just 'one' special note, which includes the
2862 name of a top-level commit. Your digital signature shows others
2863 that you trust that commit, and the immutability of the history of
2864 commits tells others that they can trust the whole history.
2866 In other words, you can easily validate a whole archive by just
2867 sending out a single email that tells the people the name (SHA1 hash)
2868 of the top commit, and digitally sign that email using something
2869 like GPG/PGP.
2871 To assist in this, git also provides the tag object...
2873 [[tag-object]]
2874 Tag Object
2875 ----------
2877 Git provides the "tag" object to simplify creating, managing and
2878 exchanging symbolic and signed tokens. The "tag" object at its
2879 simplest simply symbolically identifies another object by containing
2880 the sha1, type and symbolic name.
2882 However it can optionally contain additional signature information
2883 (which git doesn't care about as long as there's less than 8k of
2884 it). This can then be verified externally to git.
2886 Note that despite the tag features, "git" itself only handles content
2887 integrity; the trust framework (and signature provision and
2888 verification) has to come from outside.
2890 A tag is created with gitlink:git-mktag[1],
2891 its data can be accessed by gitlink:git-cat-file[1],
2892 and the signature can be verified by
2893 gitlink:git-verify-tag[1].
2896 [[the-index]]
2897 The "index" aka "Current Directory Cache"
2898 -----------------------------------------
2900 The index is a simple binary file, which contains an efficient
2901 representation of the contents of a virtual directory. It
2902 does so by a simple array that associates a set of names, dates,
2903 permissions and content (aka "blob") objects together. The cache is
2904 always kept ordered by name, and names are unique (with a few very
2905 specific rules) at any point in time, but the cache has no long-term
2906 meaning, and can be partially updated at any time.
2908 In particular, the index certainly does not need to be consistent with
2909 the current directory contents (in fact, most operations will depend on
2910 different ways to make the index 'not' be consistent with the directory
2911 hierarchy), but it has three very important attributes:
2913 '(a) it can re-generate the full state it caches (not just the
2914 directory structure: it contains pointers to the "blob" objects so
2915 that it can regenerate the data too)'
2917 As a special case, there is a clear and unambiguous one-way mapping
2918 from a current directory cache to a "tree object", which can be
2919 efficiently created from just the current directory cache without
2920 actually looking at any other data. So a directory cache at any one
2921 time uniquely specifies one and only one "tree" object (but has
2922 additional data to make it easy to match up that tree object with what
2923 has happened in the directory)
2925 '(b) it has efficient methods for finding inconsistencies between that
2926 cached state ("tree object waiting to be instantiated") and the
2927 current state.'
2929 '(c) it can additionally efficiently represent information about merge
2930 conflicts between different tree objects, allowing each pathname to be
2931 associated with sufficient information about the trees involved that
2932 you can create a three-way merge between them.'
2934 Those are the ONLY three things that the directory cache does. It's a
2935 cache, and the normal operation is to re-generate it completely from a
2936 known tree object, or update/compare it with a live tree that is being
2937 developed. If you blow the directory cache away entirely, you generally
2938 haven't lost any information as long as you have the name of the tree
2939 that it described.
2941 At the same time, the index is at the same time also the
2942 staging area for creating new trees, and creating a new tree always
2943 involves a controlled modification of the index file. In particular,
2944 the index file can have the representation of an intermediate tree that
2945 has not yet been instantiated. So the index can be thought of as a
2946 write-back cache, which can contain dirty information that has not yet
2947 been written back to the backing store.
2951 [[the-workflow]]
2952 The Workflow
2953 ------------
2955 Generally, all "git" operations work on the index file. Some operations
2956 work *purely* on the index file (showing the current state of the
2957 index), but most operations move data to and from the index file. Either
2958 from the database or from the working directory. Thus there are four
2959 main combinations:
2961 [[working-directory-to-index]]
2962 working directory -> index
2963 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2965 You update the index with information from the working directory with
2966 the gitlink:git-update-index[1] command. You
2967 generally update the index information by just specifying the filename
2968 you want to update, like so:
2970 -------------------------------------------------
2971 $ git-update-index filename
2972 -------------------------------------------------
2974 but to avoid common mistakes with filename globbing etc, the command
2975 will not normally add totally new entries or remove old entries,
2976 i.e. it will normally just update existing cache entries.
2978 To tell git that yes, you really do realize that certain files no
2979 longer exist, or that new files should be added, you
2980 should use the `--remove` and `--add` flags respectively.
2982 NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
2983 necessarily be removed: if the files still exist in your directory
2984 structure, the index will be updated with their new status, not
2985 removed. The only thing `--remove` means is that update-cache will be
2986 considering a removed file to be a valid thing, and if the file really
2987 does not exist any more, it will update the index accordingly.
2989 As a special case, you can also do `git-update-index --refresh`, which
2990 will refresh the "stat" information of each index to match the current
2991 stat information. It will 'not' update the object status itself, and
2992 it will only update the fields that are used to quickly test whether
2993 an object still matches its old backing store object.
2995 [[index-to-object-database]]
2996 index -> object database
2997 ~~~~~~~~~~~~~~~~~~~~~~~~
2999 You write your current index file to a "tree" object with the program
3001 -------------------------------------------------
3002 $ git-write-tree
3003 -------------------------------------------------
3005 that doesn't come with any options - it will just write out the
3006 current index into the set of tree objects that describe that state,
3007 and it will return the name of the resulting top-level tree. You can
3008 use that tree to re-generate the index at any time by going in the
3009 other direction:
3011 [[object-database-to-index]]
3012 object database -> index
3013 ~~~~~~~~~~~~~~~~~~~~~~~~
3015 You read a "tree" file from the object database, and use that to
3016 populate (and overwrite - don't do this if your index contains any
3017 unsaved state that you might want to restore later!) your current
3018 index. Normal operation is just
3020 -------------------------------------------------
3021 $ git-read-tree <sha1 of tree>
3022 -------------------------------------------------
3024 and your index file will now be equivalent to the tree that you saved
3025 earlier. However, that is only your 'index' file: your working
3026 directory contents have not been modified.
3028 [[index-to-working-directory]]
3029 index -> working directory
3030 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3032 You update your working directory from the index by "checking out"
3033 files. This is not a very common operation, since normally you'd just
3034 keep your files updated, and rather than write to your working
3035 directory, you'd tell the index files about the changes in your
3036 working directory (i.e. `git-update-index`).
3038 However, if you decide to jump to a new version, or check out somebody
3039 else's version, or just restore a previous tree, you'd populate your
3040 index file with read-tree, and then you need to check out the result
3041 with
3043 -------------------------------------------------
3044 $ git-checkout-index filename
3045 -------------------------------------------------
3047 or, if you want to check out all of the index, use `-a`.
3049 NOTE! git-checkout-index normally refuses to overwrite old files, so
3050 if you have an old version of the tree already checked out, you will
3051 need to use the "-f" flag ('before' the "-a" flag or the filename) to
3052 'force' the checkout.
3055 Finally, there are a few odds and ends which are not purely moving
3056 from one representation to the other:
3058 [[tying-it-all-together]]
3059 Tying it all together
3060 ~~~~~~~~~~~~~~~~~~~~~
3062 To commit a tree you have instantiated with "git-write-tree", you'd
3063 create a "commit" object that refers to that tree and the history
3064 behind it - most notably the "parent" commits that preceded it in
3065 history.
3067 Normally a "commit" has one parent: the previous state of the tree
3068 before a certain change was made. However, sometimes it can have two
3069 or more parent commits, in which case we call it a "merge", due to the
3070 fact that such a commit brings together ("merges") two or more
3071 previous states represented by other commits.
3073 In other words, while a "tree" represents a particular directory state
3074 of a working directory, a "commit" represents that state in "time",
3075 and explains how we got there.
3077 You create a commit object by giving it the tree that describes the
3078 state at the time of the commit, and a list of parents:
3080 -------------------------------------------------
3081 $ git-commit-tree <tree> -p <parent> [-p <parent2> ..]
3082 -------------------------------------------------
3084 and then giving the reason for the commit on stdin (either through
3085 redirection from a pipe or file, or by just typing it at the tty).
3087 git-commit-tree will return the name of the object that represents
3088 that commit, and you should save it away for later use. Normally,
3089 you'd commit a new `HEAD` state, and while git doesn't care where you
3090 save the note about that state, in practice we tend to just write the
3091 result to the file pointed at by `.git/HEAD`, so that we can always see
3092 what the last committed state was.
3094 Here is an ASCII art by Jon Loeliger that illustrates how
3095 various pieces fit together.
3097 ------------
3099 commit-tree
3100 commit obj
3101 +----+
3102 | |
3103 | |
3104 V V
3105 +-----------+
3106 | Object DB |
3107 | Backing |
3108 | Store |
3109 +-----------+
3110 ^
3111 write-tree | |
3112 tree obj | |
3113 | | read-tree
3114 | | tree obj
3115 V
3116 +-----------+
3117 | Index |
3118 | "cache" |
3119 +-----------+
3120 update-index ^
3121 blob obj | |
3122 | |
3123 checkout-index -u | | checkout-index
3124 stat | | blob obj
3125 V
3126 +-----------+
3127 | Working |
3128 | Directory |
3129 +-----------+
3131 ------------
3134 [[examining-the-data]]
3135 Examining the data
3136 ------------------
3138 You can examine the data represented in the object database and the
3139 index with various helper tools. For every object, you can use
3140 gitlink:git-cat-file[1] to examine details about the
3141 object:
3143 -------------------------------------------------
3144 $ git-cat-file -t <objectname>
3145 -------------------------------------------------
3147 shows the type of the object, and once you have the type (which is
3148 usually implicit in where you find the object), you can use
3150 -------------------------------------------------
3151 $ git-cat-file blob|tree|commit|tag <objectname>
3152 -------------------------------------------------
3154 to show its contents. NOTE! Trees have binary content, and as a result
3155 there is a special helper for showing that content, called
3156 `git-ls-tree`, which turns the binary content into a more easily
3157 readable form.
3159 It's especially instructive to look at "commit" objects, since those
3160 tend to be small and fairly self-explanatory. In particular, if you
3161 follow the convention of having the top commit name in `.git/HEAD`,
3162 you can do
3164 -------------------------------------------------
3165 $ git-cat-file commit HEAD
3166 -------------------------------------------------
3168 to see what the top commit was.
3170 [[merging-multiple-trees]]
3171 Merging multiple trees
3172 ----------------------
3174 Git helps you do a three-way merge, which you can expand to n-way by
3175 repeating the merge procedure arbitrary times until you finally
3176 "commit" the state. The normal situation is that you'd only do one
3177 three-way merge (two parents), and commit it, but if you like to, you
3178 can do multiple parents in one go.
3180 To do a three-way merge, you need the two sets of "commit" objects
3181 that you want to merge, use those to find the closest common parent (a
3182 third "commit" object), and then use those commit objects to find the
3183 state of the directory ("tree" object) at these points.
3185 To get the "base" for the merge, you first look up the common parent
3186 of two commits with
3188 -------------------------------------------------
3189 $ git-merge-base <commit1> <commit2>
3190 -------------------------------------------------
3192 which will return you the commit they are both based on. You should
3193 now look up the "tree" objects of those commits, which you can easily
3194 do with (for example)
3196 -------------------------------------------------
3197 $ git-cat-file commit <commitname> | head -1
3198 -------------------------------------------------
3200 since the tree object information is always the first line in a commit
3201 object.
3203 Once you know the three trees you are going to merge (the one "original"
3204 tree, aka the common tree, and the two "result" trees, aka the branches
3205 you want to merge), you do a "merge" read into the index. This will
3206 complain if it has to throw away your old index contents, so you should
3207 make sure that you've committed those - in fact you would normally
3208 always do a merge against your last commit (which should thus match what
3209 you have in your current index anyway).
3211 To do the merge, do
3213 -------------------------------------------------
3214 $ git-read-tree -m -u <origtree> <yourtree> <targettree>
3215 -------------------------------------------------
3217 which will do all trivial merge operations for you directly in the
3218 index file, and you can just write the result out with
3219 `git-write-tree`.
3222 [[merging-multiple-trees-2]]
3223 Merging multiple trees, continued
3224 ---------------------------------
3226 Sadly, many merges aren't trivial. If there are files that have
3227 been added.moved or removed, or if both branches have modified the
3228 same file, you will be left with an index tree that contains "merge
3229 entries" in it. Such an index tree can 'NOT' be written out to a tree
3230 object, and you will have to resolve any such merge clashes using
3231 other tools before you can write out the result.
3233 You can examine such index state with `git-ls-files --unmerged`
3234 command. An example:
3236 ------------------------------------------------
3237 $ git-read-tree -m $orig HEAD $target
3238 $ git-ls-files --unmerged
3239 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
3240 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
3241 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
3242 ------------------------------------------------
3244 Each line of the `git-ls-files --unmerged` output begins with
3245 the blob mode bits, blob SHA1, 'stage number', and the
3246 filename. The 'stage number' is git's way to say which tree it
3247 came from: stage 1 corresponds to `$orig` tree, stage 2 `HEAD`
3248 tree, and stage3 `$target` tree.
3250 Earlier we said that trivial merges are done inside
3251 `git-read-tree -m`. For example, if the file did not change
3252 from `$orig` to `HEAD` nor `$target`, or if the file changed
3253 from `$orig` to `HEAD` and `$orig` to `$target` the same way,
3254 obviously the final outcome is what is in `HEAD`. What the
3255 above example shows is that file `hello.c` was changed from
3256 `$orig` to `HEAD` and `$orig` to `$target` in a different way.
3257 You could resolve this by running your favorite 3-way merge
3258 program, e.g. `diff3`, `merge`, or git's own merge-file, on
3259 the blob objects from these three stages yourself, like this:
3261 ------------------------------------------------
3262 $ git-cat-file blob 263414f... >hello.c~1
3263 $ git-cat-file blob 06fa6a2... >hello.c~2
3264 $ git-cat-file blob cc44c73... >hello.c~3
3265 $ git merge-file hello.c~2 hello.c~1 hello.c~3
3266 ------------------------------------------------
3268 This would leave the merge result in `hello.c~2` file, along
3269 with conflict markers if there are conflicts. After verifying
3270 the merge result makes sense, you can tell git what the final
3271 merge result for this file is by:
3273 -------------------------------------------------
3274 $ mv -f hello.c~2 hello.c
3275 $ git-update-index hello.c
3276 -------------------------------------------------
3278 When a path is in unmerged state, running `git-update-index` for
3279 that path tells git to mark the path resolved.
3281 The above is the description of a git merge at the lowest level,
3282 to help you understand what conceptually happens under the hood.
3283 In practice, nobody, not even git itself, uses three `git-cat-file`
3284 for this. There is `git-merge-index` program that extracts the
3285 stages to temporary files and calls a "merge" script on it:
3287 -------------------------------------------------
3288 $ git-merge-index git-merge-one-file hello.c
3289 -------------------------------------------------
3291 and that is what higher level `git merge -s resolve` is implemented with.
3293 [[pack-files]]
3294 How git stores objects efficiently: pack files
3295 ----------------------------------------------
3297 We've seen how git stores each object in a file named after the
3298 object's SHA1 hash.
3300 Unfortunately this system becomes inefficient once a project has a
3301 lot of objects. Try this on an old project:
3303 ------------------------------------------------
3304 $ git count-objects
3305 6930 objects, 47620 kilobytes
3306 ------------------------------------------------
3308 The first number is the number of objects which are kept in
3309 individual files. The second is the amount of space taken up by
3310 those "loose" objects.
3312 You can save space and make git faster by moving these loose objects in
3313 to a "pack file", which stores a group of objects in an efficient
3314 compressed format; the details of how pack files are formatted can be
3315 found in link:technical/pack-format.txt[technical/pack-format.txt].
3317 To put the loose objects into a pack, just run git repack:
3319 ------------------------------------------------
3320 $ git repack
3321 Generating pack...
3322 Done counting 6020 objects.
3323 Deltifying 6020 objects.
3324 100% (6020/6020) done
3325 Writing 6020 objects.
3326 100% (6020/6020) done
3327 Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
3328 Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
3329 ------------------------------------------------
3331 You can then run
3333 ------------------------------------------------
3334 $ git prune
3335 ------------------------------------------------
3337 to remove any of the "loose" objects that are now contained in the
3338 pack. This will also remove any unreferenced objects (which may be
3339 created when, for example, you use "git reset" to remove a commit).
3340 You can verify that the loose objects are gone by looking at the
3341 .git/objects directory or by running
3343 ------------------------------------------------
3344 $ git count-objects
3345 0 objects, 0 kilobytes
3346 ------------------------------------------------
3348 Although the object files are gone, any commands that refer to those
3349 objects will work exactly as they did before.
3351 The gitlink:git-gc[1] command performs packing, pruning, and more for
3352 you, so is normally the only high-level command you need.
3354 [[dangling-objects]]
3355 Dangling objects
3356 ----------------
3358 The gitlink:git-fsck[1] command will sometimes complain about dangling
3359 objects. They are not a problem.
3361 The most common cause of dangling objects is that you've rebased a
3362 branch, or you have pulled from somebody else who rebased a branch--see
3363 <<cleaning-up-history>>. In that case, the old head of the original
3364 branch still exists, as does everything it pointed to. The branch
3365 pointer itself just doesn't, since you replaced it with another one.
3367 There are also other situations that cause dangling objects. For
3368 example, a "dangling blob" may arise because you did a "git add" of a
3369 file, but then, before you actually committed it and made it part of the
3370 bigger picture, you changed something else in that file and committed
3371 that *updated* thing - the old state that you added originally ends up
3372 not being pointed to by any commit or tree, so it's now a dangling blob
3373 object.
3375 Similarly, when the "recursive" merge strategy runs, and finds that
3376 there are criss-cross merges and thus more than one merge base (which is
3377 fairly unusual, but it does happen), it will generate one temporary
3378 midway tree (or possibly even more, if you had lots of criss-crossing
3379 merges and more than two merge bases) as a temporary internal merge
3380 base, and again, those are real objects, but the end result will not end
3381 up pointing to them, so they end up "dangling" in your repository.
3383 Generally, dangling objects aren't anything to worry about. They can
3384 even be very useful: if you screw something up, the dangling objects can
3385 be how you recover your old tree (say, you did a rebase, and realized
3386 that you really didn't want to - you can look at what dangling objects
3387 you have, and decide to reset your head to some old dangling state).
3389 For commits, you can just use:
3391 ------------------------------------------------
3392 $ gitk <dangling-commit-sha-goes-here> --not --all
3393 ------------------------------------------------
3395 This asks for all the history reachable from the given commit but not
3396 from any branch, tag, or other reference. If you decide it's something
3397 you want, you can always create a new reference to it, e.g.,
3399 ------------------------------------------------
3400 $ git branch recovered-branch <dangling-commit-sha-goes-here>
3401 ------------------------------------------------
3403 For blobs and trees, you can't do the same, but you can still examine
3404 them. You can just do
3406 ------------------------------------------------
3407 $ git show <dangling-blob/tree-sha-goes-here>
3408 ------------------------------------------------
3410 to show what the contents of the blob were (or, for a tree, basically
3411 what the "ls" for that directory was), and that may give you some idea
3412 of what the operation was that left that dangling object.
3414 Usually, dangling blobs and trees aren't very interesting. They're
3415 almost always the result of either being a half-way mergebase (the blob
3416 will often even have the conflict markers from a merge in it, if you
3417 have had conflicting merges that you fixed up by hand), or simply
3418 because you interrupted a "git fetch" with ^C or something like that,
3419 leaving _some_ of the new objects in the object database, but just
3420 dangling and useless.
3422 Anyway, once you are sure that you're not interested in any dangling
3423 state, you can just prune all unreachable objects:
3425 ------------------------------------------------
3426 $ git prune
3427 ------------------------------------------------
3429 and they'll be gone. But you should only run "git prune" on a quiescent
3430 repository - it's kind of like doing a filesystem fsck recovery: you
3431 don't want to do that while the filesystem is mounted.
3433 (The same is true of "git-fsck" itself, btw - but since
3434 git-fsck never actually *changes* the repository, it just reports
3435 on what it found, git-fsck itself is never "dangerous" to run.
3436 Running it while somebody is actually changing the repository can cause
3437 confusing and scary messages, but it won't actually do anything bad. In
3438 contrast, running "git prune" while somebody is actively changing the
3439 repository is a *BAD* idea).
3441 [[birdview-on-the-source-code]]
3442 A birds-eye view of Git's source code
3443 -------------------------------------
3445 It is not always easy for new developers to find their way through Git's
3446 source code. This section gives you a little guidance to show where to
3447 start.
3449 A good place to start is with the contents of the initial commit, with:
3451 ----------------------------------------------------
3452 $ git checkout e83c5163
3453 ----------------------------------------------------
3455 The initial revision lays the foundation for almost everything git has
3456 today, but is small enough to read in one sitting.
3458 Note that terminology has changed since that revision. For example, the
3459 README in that revision uses the word "changeset" to describe what we
3460 now call a <<def_commit_object,commit>>.
3462 Also, we do not call it "cache" any more, but "index", however, the
3463 file is still called `cache.h`. Remark: Not much reason to change it now,
3464 especially since there is no good single name for it anyway, because it is
3465 basically _the_ header file which is included by _all_ of Git's C sources.
3467 If you grasp the ideas in that initial commit, you should check out a
3468 more recent version and skim `cache.h`, `object.h` and `commit.h`.
3470 In the early days, Git (in the tradition of UNIX) was a bunch of programs
3471 which were extremely simple, and which you used in scripts, piping the
3472 output of one into another. This turned out to be good for initial
3473 development, since it was easier to test new things. However, recently
3474 many of these parts have become builtins, and some of the core has been
3475 "libified", i.e. put into libgit.a for performance, portability reasons,
3476 and to avoid code duplication.
3478 By now, you know what the index is (and find the corresponding data
3479 structures in `cache.h`), and that there are just a couple of object types
3480 (blobs, trees, commits and tags) which inherit their common structure from
3481 `struct object`, which is their first member (and thus, you can cast e.g.
3482 `(struct object *)commit` to achieve the _same_ as `&commit->object`, i.e.
3483 get at the object name and flags).
3485 Now is a good point to take a break to let this information sink in.
3487 Next step: get familiar with the object naming. Read <<naming-commits>>.
3488 There are quite a few ways to name an object (and not only revisions!).
3489 All of these are handled in `sha1_name.c`. Just have a quick look at
3490 the function `get_sha1()`. A lot of the special handling is done by
3491 functions like `get_sha1_basic()` or the likes.
3493 This is just to get you into the groove for the most libified part of Git:
3494 the revision walker.
3496 Basically, the initial version of `git log` was a shell script:
3498 ----------------------------------------------------------------
3499 $ git-rev-list --pretty $(git-rev-parse --default HEAD "$@") | \
3500 LESS=-S ${PAGER:-less}
3501 ----------------------------------------------------------------
3503 What does this mean?
3505 `git-rev-list` is the original version of the revision walker, which
3506 _always_ printed a list of revisions to stdout. It is still functional,
3507 and needs to, since most new Git programs start out as scripts using
3508 `git-rev-list`.
3510 `git-rev-parse` is not as important any more; it was only used to filter out
3511 options that were relevant for the different plumbing commands that were
3512 called by the script.
3514 Most of what `git-rev-list` did is contained in `revision.c` and
3515 `revision.h`. It wraps the options in a struct named `rev_info`, which
3516 controls how and what revisions are walked, and more.
3518 The original job of `git-rev-parse` is now taken by the function
3519 `setup_revisions()`, which parses the revisions and the common command line
3520 options for the revision walker. This information is stored in the struct
3521 `rev_info` for later consumption. You can do your own command line option
3522 parsing after calling `setup_revisions()`. After that, you have to call
3523 `prepare_revision_walk()` for initialization, and then you can get the
3524 commits one by one with the function `get_revision()`.
3526 If you are interested in more details of the revision walking process,
3527 just have a look at the first implementation of `cmd_log()`; call
3528 `git-show v1.3.0~155^2~4` and scroll down to that function (note that you
3529 no longer need to call `setup_pager()` directly).
3531 Nowadays, `git log` is a builtin, which means that it is _contained_ in the
3532 command `git`. The source side of a builtin is
3534 - a function called `cmd_<bla>`, typically defined in `builtin-<bla>.c`,
3535 and declared in `builtin.h`,
3537 - an entry in the `commands[]` array in `git.c`, and
3539 - an entry in `BUILTIN_OBJECTS` in the `Makefile`.
3541 Sometimes, more than one builtin is contained in one source file. For
3542 example, `cmd_whatchanged()` and `cmd_log()` both reside in `builtin-log.c`,
3543 since they share quite a bit of code. In that case, the commands which are
3544 _not_ named like the `.c` file in which they live have to be listed in
3545 `BUILT_INS` in the `Makefile`.
3547 `git log` looks more complicated in C than it does in the original script,
3548 but that allows for a much greater flexibility and performance.
3550 Here again it is a good point to take a pause.
3552 Lesson three is: study the code. Really, it is the best way to learn about
3553 the organization of Git (after you know the basic concepts).
3555 So, think about something which you are interested in, say, "how can I
3556 access a blob just knowing the object name of it?". The first step is to
3557 find a Git command with which you can do it. In this example, it is either
3558 `git show` or `git cat-file`.
3560 For the sake of clarity, let's stay with `git cat-file`, because it
3562 - is plumbing, and
3564 - was around even in the initial commit (it literally went only through
3565 some 20 revisions as `cat-file.c`, was renamed to `builtin-cat-file.c`
3566 when made a builtin, and then saw less than 10 versions).
3568 So, look into `builtin-cat-file.c`, search for `cmd_cat_file()` and look what
3569 it does.
3571 ------------------------------------------------------------------
3572 git_config(git_default_config);
3573 if (argc != 3)
3574 usage("git-cat-file [-t|-s|-e|-p|<type>] <sha1>");
3575 if (get_sha1(argv[2], sha1))
3576 die("Not a valid object name %s", argv[2]);
3577 ------------------------------------------------------------------
3579 Let's skip over the obvious details; the only really interesting part
3580 here is the call to `get_sha1()`. It tries to interpret `argv[2]` as an
3581 object name, and if it refers to an object which is present in the current
3582 repository, it writes the resulting SHA-1 into the variable `sha1`.
3584 Two things are interesting here:
3586 - `get_sha1()` returns 0 on _success_. This might surprise some new
3587 Git hackers, but there is a long tradition in UNIX to return different
3588 negative numbers in case of different errors -- and 0 on success.
3590 - the variable `sha1` in the function signature of `get_sha1()` is `unsigned
3591 char \*`, but is actually expected to be a pointer to `unsigned
3592 char[20]`. This variable will contain the 160-bit SHA-1 of the given
3593 commit. Note that whenever a SHA-1 is passed as `unsigned char \*`, it
3594 is the binary representation, as opposed to the ASCII representation in
3595 hex characters, which is passed as `char *`.
3597 You will see both of these things throughout the code.
3599 Now, for the meat:
3601 -----------------------------------------------------------------------------
3602 case 0:
3603 buf = read_object_with_reference(sha1, argv[1], &size, NULL);
3604 -----------------------------------------------------------------------------
3606 This is how you read a blob (actually, not only a blob, but any type of
3607 object). To know how the function `read_object_with_reference()` actually
3608 works, find the source code for it (something like `git grep
3609 read_object_with | grep ":[a-z]"` in the git repository), and read
3610 the source.
3612 To find out how the result can be used, just read on in `cmd_cat_file()`:
3614 -----------------------------------
3615 write_or_die(1, buf, size);
3616 -----------------------------------
3618 Sometimes, you do not know where to look for a feature. In many such cases,
3619 it helps to search through the output of `git log`, and then `git show` the
3620 corresponding commit.
3622 Example: If you know that there was some test case for `git bundle`, but
3623 do not remember where it was (yes, you _could_ `git grep bundle t/`, but that
3624 does not illustrate the point!):
3626 ------------------------
3627 $ git log --no-merges t/
3628 ------------------------
3630 In the pager (`less`), just search for "bundle", go a few lines back,
3631 and see that it is in commit 18449ab0... Now just copy this object name,
3632 and paste it into the command line
3634 -------------------
3635 $ git show 18449ab0
3636 -------------------
3638 Voila.
3640 Another example: Find out what to do in order to make some script a
3641 builtin:
3643 -------------------------------------------------
3644 $ git log --no-merges --diff-filter=A builtin-*.c
3645 -------------------------------------------------
3647 You see, Git is actually the best tool to find out about the source of Git
3648 itself!
3650 [[glossary]]
3651 include::glossary.txt[]
3653 [[git-quick-start]]
3654 Appendix A: Git Quick Reference
3655 ===============================
3657 This is a quick summary of the major commands; the previous chapters
3658 explain how these work in more detail.
3660 [[quick-creating-a-new-repository]]
3661 Creating a new repository
3662 -------------------------
3664 From a tarball:
3666 -----------------------------------------------
3667 $ tar xzf project.tar.gz
3668 $ cd project
3669 $ git init
3670 Initialized empty Git repository in .git/
3671 $ git add .
3672 $ git commit
3673 -----------------------------------------------
3675 From a remote repository:
3677 -----------------------------------------------
3678 $ git clone git://example.com/pub/project.git
3679 $ cd project
3680 -----------------------------------------------
3682 [[managing-branches]]
3683 Managing branches
3684 -----------------
3686 -----------------------------------------------
3687 $ git branch # list all local branches in this repo
3688 $ git checkout test # switch working directory to branch "test"
3689 $ git branch new # create branch "new" starting at current HEAD
3690 $ git branch -d new # delete branch "new"
3691 -----------------------------------------------
3693 Instead of basing new branch on current HEAD (the default), use:
3695 -----------------------------------------------
3696 $ git branch new test # branch named "test"
3697 $ git branch new v2.6.15 # tag named v2.6.15
3698 $ git branch new HEAD^ # commit before the most recent
3699 $ git branch new HEAD^^ # commit before that
3700 $ git branch new test~10 # ten commits before tip of branch "test"
3701 -----------------------------------------------
3703 Create and switch to a new branch at the same time:
3705 -----------------------------------------------
3706 $ git checkout -b new v2.6.15
3707 -----------------------------------------------
3709 Update and examine branches from the repository you cloned from:
3711 -----------------------------------------------
3712 $ git fetch # update
3713 $ git branch -r # list
3714 origin/master
3715 origin/next
3716 ...
3717 $ git checkout -b masterwork origin/master
3718 -----------------------------------------------
3720 Fetch a branch from a different repository, and give it a new
3721 name in your repository:
3723 -----------------------------------------------
3724 $ git fetch git://example.com/project.git theirbranch:mybranch
3725 $ git fetch git://example.com/project.git v2.6.15:mybranch
3726 -----------------------------------------------
3728 Keep a list of repositories you work with regularly:
3730 -----------------------------------------------
3731 $ git remote add example git://example.com/project.git
3732 $ git remote # list remote repositories
3733 example
3734 origin
3735 $ git remote show example # get details
3736 * remote example
3737 URL: git://example.com/project.git
3738 Tracked remote branches
3739 master next ...
3740 $ git fetch example # update branches from example
3741 $ git branch -r # list all remote branches
3742 -----------------------------------------------
3745 [[exploring-history]]
3746 Exploring history
3747 -----------------
3749 -----------------------------------------------
3750 $ gitk # visualize and browse history
3751 $ git log # list all commits
3752 $ git log src/ # ...modifying src/
3753 $ git log v2.6.15..v2.6.16 # ...in v2.6.16, not in v2.6.15
3754 $ git log master..test # ...in branch test, not in branch master
3755 $ git log test..master # ...in branch master, but not in test
3756 $ git log test...master # ...in one branch, not in both
3757 $ git log -S'foo()' # ...where difference contain "foo()"
3758 $ git log --since="2 weeks ago"
3759 $ git log -p # show patches as well
3760 $ git show # most recent commit
3761 $ git diff v2.6.15..v2.6.16 # diff between two tagged versions
3762 $ git diff v2.6.15..HEAD # diff with current head
3763 $ git grep "foo()" # search working directory for "foo()"
3764 $ git grep v2.6.15 "foo()" # search old tree for "foo()"
3765 $ git show v2.6.15:a.txt # look at old version of a.txt
3766 -----------------------------------------------
3768 Search for regressions:
3770 -----------------------------------------------
3771 $ git bisect start
3772 $ git bisect bad # current version is bad
3773 $ git bisect good v2.6.13-rc2 # last known good revision
3774 Bisecting: 675 revisions left to test after this
3775 # test here, then:
3776 $ git bisect good # if this revision is good, or
3777 $ git bisect bad # if this revision is bad.
3778 # repeat until done.
3779 -----------------------------------------------
3781 [[making-changes]]
3782 Making changes
3783 --------------
3785 Make sure git knows who to blame:
3787 ------------------------------------------------
3788 $ cat >>~/.gitconfig <<\EOF
3789 [user]
3790 name = Your Name Comes Here
3791 email = you@yourdomain.example.com
3792 EOF
3793 ------------------------------------------------
3795 Select file contents to include in the next commit, then make the
3796 commit:
3798 -----------------------------------------------
3799 $ git add a.txt # updated file
3800 $ git add b.txt # new file
3801 $ git rm c.txt # old file
3802 $ git commit
3803 -----------------------------------------------
3805 Or, prepare and create the commit in one step:
3807 -----------------------------------------------
3808 $ git commit d.txt # use latest content only of d.txt
3809 $ git commit -a # use latest content of all tracked files
3810 -----------------------------------------------
3812 [[merging]]
3813 Merging
3814 -------
3816 -----------------------------------------------
3817 $ git merge test # merge branch "test" into the current branch
3818 $ git pull git://example.com/project.git master
3819 # fetch and merge in remote branch
3820 $ git pull . test # equivalent to git merge test
3821 -----------------------------------------------
3823 [[sharing-your-changes]]
3824 Sharing your changes
3825 --------------------
3827 Importing or exporting patches:
3829 -----------------------------------------------
3830 $ git format-patch origin..HEAD # format a patch for each commit
3831 # in HEAD but not in origin
3832 $ git am mbox # import patches from the mailbox "mbox"
3833 -----------------------------------------------
3835 Fetch a branch in a different git repository, then merge into the
3836 current branch:
3838 -----------------------------------------------
3839 $ git pull git://example.com/project.git theirbranch
3840 -----------------------------------------------
3842 Store the fetched branch into a local branch before merging into the
3843 current branch:
3845 -----------------------------------------------
3846 $ git pull git://example.com/project.git theirbranch:mybranch
3847 -----------------------------------------------
3849 After creating commits on a local branch, update the remote
3850 branch with your commits:
3852 -----------------------------------------------
3853 $ git push ssh://example.com/project.git mybranch:theirbranch
3854 -----------------------------------------------
3856 When remote and local branch are both named "test":
3858 -----------------------------------------------
3859 $ git push ssh://example.com/project.git test
3860 -----------------------------------------------
3862 Shortcut version for a frequently used remote repository:
3864 -----------------------------------------------
3865 $ git remote add example ssh://example.com/project.git
3866 $ git push example test
3867 -----------------------------------------------
3869 [[repository-maintenance]]
3870 Repository maintenance
3871 ----------------------
3873 Check for corruption:
3875 -----------------------------------------------
3876 $ git fsck
3877 -----------------------------------------------
3879 Recompress, remove unused cruft:
3881 -----------------------------------------------
3882 $ git gc
3883 -----------------------------------------------
3886 [[todo]]
3887 Appendix B: Notes and todo list for this manual
3888 ===============================================
3890 This is a work in progress.
3892 The basic requirements:
3893 - It must be readable in order, from beginning to end, by
3894 someone intelligent with a basic grasp of the unix
3895 commandline, but without any special knowledge of git. If
3896 necessary, any other prerequisites should be specifically
3897 mentioned as they arise.
3898 - Whenever possible, section headings should clearly describe
3899 the task they explain how to do, in language that requires
3900 no more knowledge than necessary: for example, "importing
3901 patches into a project" rather than "the git-am command"
3903 Think about how to create a clear chapter dependency graph that will
3904 allow people to get to important topics without necessarily reading
3905 everything in between.
3907 Scan Documentation/ for other stuff left out; in particular:
3908 howto's
3909 some of technical/?
3910 hooks
3911 list of commands in gitlink:git[1]
3913 Scan email archives for other stuff left out
3915 Scan man pages to see if any assume more background than this manual
3916 provides.
3918 Simplify beginning by suggesting disconnected head instead of
3919 temporary branch creation?
3921 Add more good examples. Entire sections of just cookbook examples
3922 might be a good idea; maybe make an "advanced examples" section a
3923 standard end-of-chapter section?
3925 Include cross-references to the glossary, where appropriate.
3927 Document shallow clones? See draft 1.5.0 release notes for some
3928 documentation.
3930 Add a section on working with other version control systems, including
3931 CVS, Subversion, and just imports of series of release tarballs.
3933 More details on gitweb?
3935 Write a chapter on using plumbing and writing scripts.
3937 Alternates, clone -reference, etc.
3939 git unpack-objects -r for recovery