user-manual: git-fsck, dangling objects
[git/git.git] / Documentation / user-manual.txt
CommitLineData
d19fbc3c
BF
1Git User's Manual
2_________________
3
4This manual is designed to be readable by someone with basic unix
5commandline skills, but no previous knowledge of git.
6
ef89f701 7Chapter 1 gives a brief overview of git commands, without any
b181d57f 8explanation; you may prefer to skip to chapter 2 on a first reading.
ef89f701
BF
9
10Chapters 2 and 3 explain how to fetch and study a project using
d5cd5de4
BF
11git--the tools you'd need to build and test a particular version of a
12software project, to search for regressions, and so on.
6bd9b682 13
ef89f701 14Chapter 4 explains how to do development with git, and chapter 5 how
d5cd5de4 15to share that development with others.
6bd9b682
BF
16
17Further chapters cover more specialized topics.
18
d19fbc3c
BF
19Comprehensive reference documentation is available through the man
20pages. For a command such as "git clone", just use
21
22------------------------------------------------
23$ man git-clone
24------------------------------------------------
25
ef89f701
BF
26Git Quick Start
27===============
28
29This is a quick summary of the major commands; the following chapters
30will explain how these work in more detail.
31
32Creating a new repository
33-------------------------
34
35From a tarball:
36
37-----------------------------------------------
38$ tar xzf project.tar.gz
39$ cd project
40$ git init
41Initialized empty Git repository in .git/
42$ git add .
43$ git commit
44-----------------------------------------------
45
46From a remote repository:
47
48-----------------------------------------------
49$ git clone git://example.com/pub/project.git
50$ cd project
51-----------------------------------------------
52
53Managing branches
54-----------------
55
56-----------------------------------------------
57$ git branch # list all branches in this repo
58$ git checkout test # switch working directory to branch "test"
59$ git branch new # create branch "new" starting at current HEAD
60$ git branch -d new # delete branch "new"
61-----------------------------------------------
62
63Instead of basing new branch on current HEAD (the default), use:
64
65-----------------------------------------------
66$ git branch new test # branch named "test"
67$ git branch new v2.6.15 # tag named v2.6.15
68$ git branch new HEAD^ # commit before the most recent
69$ git branch new HEAD^^ # commit before that
70$ git branch new test~10 # ten commits before tip of branch "test"
71-----------------------------------------------
72
73Create and switch to a new branch at the same time:
74
75-----------------------------------------------
76$ git checkout -b new v2.6.15
77-----------------------------------------------
78
79Update and examine branches from the repository you cloned from:
80
81-----------------------------------------------
82$ git fetch # update
83$ git branch -r # list
84 origin/master
85 origin/next
86 ...
87$ git branch checkout -b masterwork origin/master
88-----------------------------------------------
89
90Fetch a branch from a different repository, and give it a new
91name in your repository:
92
93-----------------------------------------------
94$ git fetch git://example.com/project.git theirbranch:mybranch
95$ git fetch git://example.com/project.git v2.6.15:mybranch
96-----------------------------------------------
97
98Keep a list of repositories you work with regularly:
99
100-----------------------------------------------
101$ git remote add example git://example.com/project.git
b181d57f 102$ git remote # list remote repositories
ef89f701
BF
103example
104origin
b181d57f 105$ git remote show example # get details
ef89f701
BF
106* remote example
107 URL: git://example.com/project.git
108 Tracked remote branches
109 master next ...
b181d57f
BF
110$ git fetch example # update branches from example
111$ git branch -r # list all remote branches
ef89f701
BF
112-----------------------------------------------
113
114
115Exploring history
116-----------------
117
118-----------------------------------------------
119$ gitk # visualize and browse history
120$ git log # list all commits
121$ git log src/ # ...modifying src/
122$ git log v2.6.15..v2.6.16 # ...in v2.6.16, not in v2.6.15
123$ git log master..test # ...in branch test, not in branch master
124$ git log test..master # ...in branch master, but not in test
125$ git log test...master # ...in one branch, not in both
126$ git log -S'foo()' # ...where difference contain "foo()"
127$ git log --since="2 weeks ago"
128$ git log -p # show patches as well
129$ git show # most recent commit
130$ git diff v2.6.15..v2.6.16 # diff between two tagged versions
131$ git diff v2.6.15..HEAD # diff with current head
132$ git grep "foo()" # search working directory for "foo()"
133$ git grep v2.6.15 "foo()" # search old tree for "foo()"
134$ git show v2.6.15:a.txt # look at old version of a.txt
135-----------------------------------------------
136
b181d57f 137Search for regressions:
ef89f701
BF
138
139-----------------------------------------------
140$ git bisect start
141$ git bisect bad # current version is bad
142$ git bisect good v2.6.13-rc2 # last known good revision
143Bisecting: 675 revisions left to test after this
144 # test here, then:
145$ git bisect good # if this revision is good, or
146$ git bisect bad # if this revision is bad.
147 # repeat until done.
148-----------------------------------------------
149
150Making changes
151--------------
152
153Make sure git knows who to blame:
154
155------------------------------------------------
156$ cat >~/.gitconfig <<\EOF
157[user]
158name = Your Name Comes Here
159email = you@yourdomain.example.com
160EOF
161------------------------------------------------
162
163Select file contents to include in the next commit, then make the
164commit:
165
166-----------------------------------------------
167$ git add a.txt # updated file
168$ git add b.txt # new file
169$ git rm c.txt # old file
170$ git commit
171-----------------------------------------------
172
173Or, prepare and create the commit in one step:
174
175-----------------------------------------------
b181d57f 176$ git commit d.txt # use latest content only of d.txt
ef89f701
BF
177$ git commit -a # use latest content of all tracked files
178-----------------------------------------------
179
180Merging
181-------
182
183-----------------------------------------------
184$ git merge test # merge branch "test" into the current branch
185$ git pull git://example.com/project.git master
186 # fetch and merge in remote branch
187$ git pull . test # equivalent to git merge test
188-----------------------------------------------
189
e4add70c
BF
190Sharing your changes
191--------------------
ef89f701
BF
192
193Importing or exporting patches:
194
195-----------------------------------------------
196$ git format-patch origin..HEAD # format a patch for each commit
197 # in HEAD but not in origin
198$ git-am mbox # import patches from the mailbox "mbox"
199-----------------------------------------------
200
ef89f701
BF
201Fetch a branch in a different git repository, then merge into the
202current branch:
203
204-----------------------------------------------
205$ git pull git://example.com/project.git theirbranch
206-----------------------------------------------
207
208Store the fetched branch into a local branch before merging into the
209current branch:
210
211-----------------------------------------------
212$ git pull git://example.com/project.git theirbranch:mybranch
213-----------------------------------------------
214
e4add70c
BF
215After creating commits on a local branch, update the remote
216branch with your commits:
217
218-----------------------------------------------
219$ git push ssh://example.com/project.git mybranch:theirbranch
220-----------------------------------------------
221
222When remote and local branch are both named "test":
223
224-----------------------------------------------
225$ git push ssh://example.com/project.git test
226-----------------------------------------------
227
228Shortcut version for a frequently used remote repository:
229
230-----------------------------------------------
231$ git remote add example ssh://example.com/project.git
232$ git push example test
233-----------------------------------------------
234
b181d57f
BF
235Repository maintenance
236----------------------
237
238Check for corruption:
239
240-----------------------------------------------
241$ git fsck-objects
242-----------------------------------------------
243
244Recompress, remove unused cruft:
245
246-----------------------------------------------
247$ git gc
248-----------------------------------------------
249
d19fbc3c
BF
250Repositories and Branches
251=========================
252
253How to get a git repository
254---------------------------
255
256It will be useful to have a git repository to experiment with as you
257read this manual.
258
259The best way to get one is by using the gitlink:git-clone[1] command
260to download a copy of an existing repository for a project that you
261are interested in. If you don't already have a project in mind, here
262are some interesting examples:
263
264------------------------------------------------
265 # git itself (approx. 10MB download):
266$ git clone git://git.kernel.org/pub/scm/git/git.git
267 # the linux kernel (approx. 150MB download):
268$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
269------------------------------------------------
270
271The initial clone may be time-consuming for a large project, but you
272will only need to clone once.
273
274The clone command creates a new directory named after the project
275("git" or "linux-2.6" in the examples above). After you cd into this
276directory, you will see that it contains a copy of the project files,
277together with a special top-level directory named ".git", which
278contains all the information about the history of the project.
279
d5cd5de4 280In most of the following, examples will be taken from one of the two
d19fbc3c
BF
281repositories above.
282
283How to check out a different version of a project
284-------------------------------------------------
285
286Git is best thought of as a tool for storing the history of a
287collection of files. It stores the history as a compressed
288collection of interrelated snapshots (versions) of the project's
289contents.
290
291A single git repository may contain multiple branches. Each branch
292is a bookmark referencing a particular point in the project history.
293The gitlink:git-branch[1] command shows you the list of branches:
294
295------------------------------------------------
296$ git branch
297* master
298------------------------------------------------
299
300A freshly cloned repository contains a single branch, named "master",
301and the working directory contains the version of the project
302referred to by the master branch.
303
304Most projects also use tags. Tags, like branches, are references
305into the project's history, and can be listed using the
306gitlink:git-tag[1] command:
307
308------------------------------------------------
309$ git tag -l
310v2.6.11
311v2.6.11-tree
312v2.6.12
313v2.6.12-rc2
314v2.6.12-rc3
315v2.6.12-rc4
316v2.6.12-rc5
317v2.6.12-rc6
318v2.6.13
319...
320------------------------------------------------
321
fe4b3e59
BF
322Tags are expected to always point at the same version of a project,
323while branches are expected to advance as development progresses.
324
d19fbc3c
BF
325Create a new branch pointing to one of these versions and check it
326out using gitlink:git-checkout[1]:
327
328------------------------------------------------
329$ git checkout -b new v2.6.13
330------------------------------------------------
331
332The working directory then reflects the contents that the project had
333when it was tagged v2.6.13, and gitlink:git-branch[1] shows two
334branches, with an asterisk marking the currently checked-out branch:
335
336------------------------------------------------
337$ git branch
338 master
339* new
340------------------------------------------------
341
342If you decide that you'd rather see version 2.6.17, you can modify
343the current branch to point at v2.6.17 instead, with
344
345------------------------------------------------
346$ git reset --hard v2.6.17
347------------------------------------------------
348
349Note that if the current branch was your only reference to a
350particular point in history, then resetting that branch may leave you
351with no way to find the history it used to point to; so use this
352command carefully.
353
354Understanding History: Commits
355------------------------------
356
357Every change in the history of a project is represented by a commit.
358The gitlink:git-show[1] command shows the most recent commit on the
359current branch:
360
361------------------------------------------------
362$ git show
363commit 2b5f6dcce5bf94b9b119e9ed8d537098ec61c3d2
364Author: Jamal Hadi Salim <hadi@cyberus.ca>
365Date: Sat Dec 2 22:22:25 2006 -0800
366
367 [XFRM]: Fix aevent structuring to be more complete.
368
369 aevents can not uniquely identify an SA. We break the ABI with this
370 patch, but consensus is that since it is not yet utilized by any
371 (known) application then it is fine (better do it now than later).
372
373 Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
374 Signed-off-by: David S. Miller <davem@davemloft.net>
375
376diff --git a/Documentation/networking/xfrm_sync.txt b/Documentation/networking/xfrm_sync.txt
377index 8be626f..d7aac9d 100644
378--- a/Documentation/networking/xfrm_sync.txt
379+++ b/Documentation/networking/xfrm_sync.txt
380@@ -47,10 +47,13 @@ aevent_id structure looks like:
381
382 struct xfrm_aevent_id {
383 struct xfrm_usersa_id sa_id;
384+ xfrm_address_t saddr;
385 __u32 flags;
386+ __u32 reqid;
387 };
388...
389------------------------------------------------
390
391As you can see, a commit shows who made the latest change, what they
392did, and why.
393
b181d57f
BF
394Every commit has a 40-hexdigit id, sometimes called the "object name"
395or the "SHA1 id", shown on the first line of the "git show" output.
396You can usually refer to a commit by a shorter name, such as a tag or a
397branch name, but this longer name can also be useful. Most
398importantly, it is a globally unique name for this commit: so if you
399tell somebody else the object name (for example in email), then you are
400guaranteed that name will refer to the same commit in their repository
401that you it does in yours (assuming their repository has that commit at
402all).
d19fbc3c
BF
403
404Understanding history: commits, parents, and reachability
405~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
406
407Every commit (except the very first commit in a project) also has a
408parent commit which shows what happened before this commit.
409Following the chain of parents will eventually take you back to the
410beginning of the project.
411
412However, the commits do not form a simple list; git allows lines of
413development to diverge and then reconverge, and the point where two
414lines of development reconverge is called a "merge". The commit
415representing a merge can therefore have more than one parent, with
416each parent representing the most recent commit on one of the lines
417of development leading to that point.
418
419The best way to see how this works is using the gitlink:gitk[1]
420command; running gitk now on a git repository and looking for merge
421commits will help understand how the git organizes history.
422
423In the following, we say that commit X is "reachable" from commit Y
424if commit X is an ancestor of commit Y. Equivalently, you could say
425that Y is a descendent of X, or that there is a chain of parents
426leading from commit Y to commit X.
427
428Undestanding history: History diagrams
429~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
430
431We will sometimes represent git history using diagrams like the one
432below. Commits are shown as "o", and the links between them with
433lines drawn with - / and \. Time goes left to right:
434
435 o--o--o <-- Branch A
436 /
437 o--o--o <-- master
438 \
439 o--o--o <-- Branch B
440
441If we need to talk about a particular commit, the character "o" may
442be replaced with another letter or number.
443
444Understanding history: What is a branch?
445~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
446
447Though we've been using the word "branch" to mean a kind of reference
448to a particular commit, the word branch is also commonly used to
449refer to the line of commits leading up to that point. In the
450example above, git may think of the branch named "A" as just a
451pointer to one particular commit, but we may refer informally to the
452line of three commits leading up to that point as all being part of
453"branch A".
454
455If we need to make it clear that we're just talking about the most
456recent commit on the branch, we may refer to that commit as the
457"head" of the branch.
458
459Manipulating branches
460---------------------
461
462Creating, deleting, and modifying branches is quick and easy; here's
463a summary of the commands:
464
465git branch::
466 list all branches
467git branch <branch>::
468 create a new branch named <branch>, referencing the same
469 point in history as the current branch
470git branch <branch> <start-point>::
471 create a new branch named <branch>, referencing
472 <start-point>, which may be specified any way you like,
473 including using a branch name or a tag name
474git branch -d <branch>::
475 delete the branch <branch>; if the branch you are deleting
476 points to a commit which is not reachable from this branch,
477 this command will fail with a warning.
478git branch -D <branch>::
479 even if the branch points to a commit not reachable
480 from the current branch, you may know that that commit
481 is still reachable from some other branch or tag. In that
482 case it is safe to use this command to force git to delete
483 the branch.
484git checkout <branch>::
485 make the current branch <branch>, updating the working
486 directory to reflect the version referenced by <branch>
487git checkout -b <new> <start-point>::
488 create a new branch <new> referencing <start-point>, and
489 check it out.
490
491It is also useful to know that the special symbol "HEAD" can always
492be used to refer to the current branch.
493
494Examining branches from a remote repository
495-------------------------------------------
496
497The "master" branch that was created at the time you cloned is a copy
498of the HEAD in the repository that you cloned from. That repository
499may also have had other branches, though, and your local repository
500keeps branches which track each of those remote branches, which you
501can view using the "-r" option to gitlink:git-branch[1]:
502
503------------------------------------------------
504$ git branch -r
505 origin/HEAD
506 origin/html
507 origin/maint
508 origin/man
509 origin/master
510 origin/next
511 origin/pu
512 origin/todo
513------------------------------------------------
514
515You cannot check out these remote-tracking branches, but you can
516examine them on a branch of your own, just as you would a tag:
517
518------------------------------------------------
519$ git checkout -b my-todo-copy origin/todo
520------------------------------------------------
521
522Note that the name "origin" is just the name that git uses by default
523to refer to the repository that you cloned from.
524
525[[how-git-stores-references]]
f60b9642
BF
526Naming branches, tags, and other references
527-------------------------------------------
d19fbc3c
BF
528
529Branches, remote-tracking branches, and tags are all references to
f60b9642
BF
530commits. All references are named with a slash-separated path name
531starting with "refs"; the names we've been using so far are actually
532shorthand:
d19fbc3c 533
f60b9642
BF
534 - The branch "test" is short for "refs/heads/test".
535 - The tag "v2.6.18" is short for "refs/tags/v2.6.18".
536 - "origin/master" is short for "refs/remotes/origin/master".
d19fbc3c 537
f60b9642
BF
538The full name is occasionally useful if, for example, there ever
539exists a tag and a branch with the same name.
d19fbc3c 540
f60b9642
BF
541As another useful shortcut, if the repository "origin" posesses only
542a single branch, you can refer to that branch as just "origin".
d19fbc3c 543
f60b9642
BF
544More generally, if you have defined a remote repository named
545"example", you can refer to the branch in that repository as
546"example". And for a repository with multiple branches, this will
547refer to the branch designated as the "HEAD" branch.
d19fbc3c
BF
548
549For the complete list of paths which git checks for references, and
f60b9642
BF
550the order it uses to decide which to choose when there are multiple
551references with the same shorthand name, see the "SPECIFYING
552REVISIONS" section of gitlink:git-rev-parse[1].
d19fbc3c
BF
553
554[[Updating-a-repository-with-git-fetch]]
555Updating a repository with git fetch
556------------------------------------
557
558Eventually the developer cloned from will do additional work in her
559repository, creating new commits and advancing the branches to point
560at the new commits.
561
562The command "git fetch", with no arguments, will update all of the
563remote-tracking branches to the latest version found in her
564repository. It will not touch any of your own branches--not even the
565"master" branch that was created for you on clone.
566
d5cd5de4
BF
567Fetching branches from other repositories
568-----------------------------------------
569
570You can also track branches from repositories other than the one you
571cloned from, using gitlink:git-remote[1]:
572
573-------------------------------------------------
574$ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
575$ git fetch
576* refs/remotes/linux-nfs/master: storing branch 'master' ...
577 commit: bf81b46
578-------------------------------------------------
579
580New remote-tracking branches will be stored under the shorthand name
581that you gave "git remote add", in this case linux-nfs:
582
583-------------------------------------------------
584$ git branch -r
585linux-nfs/master
586origin/master
587-------------------------------------------------
588
589If you run "git fetch <remote>" later, the tracking branches for the
590named <remote> will be updated.
591
592If you examine the file .git/config, you will see that git has added
593a new stanza:
594
595-------------------------------------------------
596$ cat .git/config
597...
598[remote "linux-nfs"]
599 url = git://linux-nfs.org/~bfields/git.git
600 fetch = +refs/heads/*:refs/remotes/linux-nfs-read/*
601...
602-------------------------------------------------
603
604This is what causes git to track the remote's branches; you may
605modify or delete these configuration options by editing .git/config
606with a text editor.
607
d19fbc3c
BF
608Exploring git history
609=====================
610
611Git is best thought of as a tool for storing the history of a
612collection of files. It does this by storing compressed snapshots of
613the contents of a file heirarchy, together with "commits" which show
614the relationships between these snapshots.
615
616Git provides extremely flexible and fast tools for exploring the
617history of a project.
618
619We start with one specialized tool which is useful for finding the
620commit that introduced a bug into a project.
621
622How to use bisect to find a regression
623--------------------------------------
624
625Suppose version 2.6.18 of your project worked, but the version at
626"master" crashes. Sometimes the best way to find the cause of such a
627regression is to perform a brute-force search through the project's
628history to find the particular commit that caused the problem. The
629gitlink:git-bisect[1] command can help you do this:
630
631-------------------------------------------------
632$ git bisect start
633$ git bisect good v2.6.18
634$ git bisect bad master
635Bisecting: 3537 revisions left to test after this
636[65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
637-------------------------------------------------
638
639If you run "git branch" at this point, you'll see that git has
640temporarily moved you to a new branch named "bisect". This branch
641points to a commit (with commit id 65934...) that is reachable from
642v2.6.19 but not from v2.6.18. Compile and test it, and see whether
643it crashes. Assume it does crash. Then:
644
645-------------------------------------------------
646$ git bisect bad
647Bisecting: 1769 revisions left to test after this
648[7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
649-------------------------------------------------
650
651checks out an older version. Continue like this, telling git at each
652stage whether the version it gives you is good or bad, and notice
653that the number of revisions left to test is cut approximately in
654half each time.
655
656After about 13 tests (in this case), it will output the commit id of
657the guilty commit. You can then examine the commit with
658gitlink:git-show[1], find out who wrote it, and mail them your bug
659report with the commit id. Finally, run
660
661-------------------------------------------------
662$ git bisect reset
663-------------------------------------------------
664
665to return you to the branch you were on before and delete the
666temporary "bisect" branch.
667
668Note that the version which git-bisect checks out for you at each
669point is just a suggestion, and you're free to try a different
670version if you think it would be a good idea. For example,
671occasionally you may land on a commit that broke something unrelated;
672run
673
674-------------------------------------------------
675$ git bisect-visualize
676-------------------------------------------------
677
678which will run gitk and label the commit it chose with a marker that
679says "bisect". Chose a safe-looking commit nearby, note its commit
680id, and check it out with:
681
682-------------------------------------------------
683$ git reset --hard fb47ddb2db...
684-------------------------------------------------
685
686then test, run "bisect good" or "bisect bad" as appropriate, and
687continue.
688
689Naming commits
690--------------
691
692We have seen several ways of naming commits already:
693
eb6ae7f4 694 - 40-hexdigit SHA1 id
d19fbc3c
BF
695 - branch name: refers to the commit at the head of the given
696 branch
697 - tag name: refers to the commit pointed to by the given tag
698 (we've seen branches and tags are special cases of
699 <<how-git-stores-references,references>>).
700 - HEAD: refers to the head of the current branch
701
eb6ae7f4 702There are many more; see the "SPECIFYING REVISIONS" section of the
aec053bb 703gitlink:git-rev-parse[1] man page for the complete list of ways to
d19fbc3c
BF
704name revisions. Some examples:
705
706-------------------------------------------------
707$ git show fb47ddb2 # the first few characters of the SHA1 id
708 # are usually enough to specify it uniquely
709$ git show HEAD^ # the parent of the HEAD commit
710$ git show HEAD^^ # the grandparent
711$ git show HEAD~4 # the great-great-grandparent
712-------------------------------------------------
713
714Recall that merge commits may have more than one parent; by default,
715^ and ~ follow the first parent listed in the commit, but you can
716also choose:
717
718-------------------------------------------------
719$ git show HEAD^1 # show the first parent of HEAD
720$ git show HEAD^2 # show the second parent of HEAD
721-------------------------------------------------
722
723In addition to HEAD, there are several other special names for
724commits:
725
726Merges (to be discussed later), as well as operations such as
727git-reset, which change the currently checked-out commit, generally
728set ORIG_HEAD to the value HEAD had before the current operation.
729
730The git-fetch operation always stores the head of the last fetched
731branch in FETCH_HEAD. For example, if you run git fetch without
732specifying a local branch as the target of the operation
733
734-------------------------------------------------
735$ git fetch git://example.com/proj.git theirbranch
736-------------------------------------------------
737
738the fetched commits will still be available from FETCH_HEAD.
739
740When we discuss merges we'll also see the special name MERGE_HEAD,
741which refers to the other branch that we're merging in to the current
742branch.
743
aec053bb
BF
744The gitlink:git-rev-parse[1] command is a low-level command that is
745occasionally useful for translating some name for a commit to the SHA1 id for
746that commit:
747
748-------------------------------------------------
749$ git rev-parse origin
750e05db0fd4f31dde7005f075a84f96b360d05984b
751-------------------------------------------------
752
d19fbc3c
BF
753Creating tags
754-------------
755
756We can also create a tag to refer to a particular commit; after
757running
758
759-------------------------------------------------
760$ git-tag stable-1 1b2e1d63ff
761-------------------------------------------------
762
763You can use stable-1 to refer to the commit 1b2e1d63ff.
764
765This creates a "lightweight" tag. If the tag is a tag you wish to
766share with others, and possibly sign cryptographically, then you
767should create a tag object instead; see the gitlink:git-tag[1] man
768page for details.
769
770Browsing revisions
771------------------
772
773The gitlink:git-log[1] command can show lists of commits. On its
774own, it shows all commits reachable from the parent commit; but you
775can also make more specific requests:
776
777-------------------------------------------------
778$ git log v2.5.. # commits since (not reachable from) v2.5
779$ git log test..master # commits reachable from master but not test
780$ git log master..test # ...reachable from test but not master
781$ git log master...test # ...reachable from either test or master,
782 # but not both
783$ git log --since="2 weeks ago" # commits from the last 2 weeks
784$ git log Makefile # commits which modify Makefile
785$ git log fs/ # ... which modify any file under fs/
786$ git log -S'foo()' # commits which add or remove any file data
787 # matching the string 'foo()'
788-------------------------------------------------
789
790And of course you can combine all of these; the following finds
791commits since v2.5 which touch the Makefile or any file under fs:
792
793-------------------------------------------------
794$ git log v2.5.. Makefile fs/
795-------------------------------------------------
796
797You can also ask git log to show patches:
798
799-------------------------------------------------
800$ git log -p
801-------------------------------------------------
802
803See the "--pretty" option in the gitlink:git-log[1] man page for more
804display options.
805
806Note that git log starts with the most recent commit and works
807backwards through the parents; however, since git history can contain
808multiple independant lines of development, the particular order that
809commits are listed in may be somewhat arbitrary.
810
811Generating diffs
812----------------
813
814You can generate diffs between any two versions using
815gitlink:git-diff[1]:
816
817-------------------------------------------------
818$ git diff master..test
819-------------------------------------------------
820
821Sometimes what you want instead is a set of patches:
822
823-------------------------------------------------
824$ git format-patch master..test
825-------------------------------------------------
826
827will generate a file with a patch for each commit reachable from test
828but not from master. Note that if master also has commits which are
829not reachable from test, then the combined result of these patches
830will not be the same as the diff produced by the git-diff example.
831
832Viewing old file versions
833-------------------------
834
835You can always view an old version of a file by just checking out the
836correct revision first. But sometimes it is more convenient to be
837able to view an old version of a single file without checking
838anything out; this command does that:
839
840-------------------------------------------------
841$ git show v2.5:fs/locks.c
842-------------------------------------------------
843
844Before the colon may be anything that names a commit, and after it
845may be any path to a file tracked by git.
846
aec053bb
BF
847Examples
848--------
849
850Check whether two branches point at the same history
2f99710c 851~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
aec053bb
BF
852
853Suppose you want to check whether two branches point at the same point
854in history.
855
856-------------------------------------------------
857$ git diff origin..master
858-------------------------------------------------
859
69f7ad73
BF
860will tell you whether the contents of the project are the same at the
861two branches; in theory, however, it's possible that the same project
862contents could have been arrived at by two different historical
863routes. You could compare the SHA1 id's:
aec053bb
BF
864
865-------------------------------------------------
866$ git rev-list origin
867e05db0fd4f31dde7005f075a84f96b360d05984b
868$ git rev-list master
869e05db0fd4f31dde7005f075a84f96b360d05984b
870-------------------------------------------------
871
69f7ad73
BF
872Or you could recall that the ... operator selects all commits
873contained reachable from either one reference or the other but not
874both: so
aec053bb
BF
875
876-------------------------------------------------
877$ git log origin...master
878-------------------------------------------------
879
880will return no commits when the two branches are equal.
881
b181d57f
BF
882Find first tagged version including a given fix
883~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
aec053bb 884
69f7ad73
BF
885Suppose you know that the commit e05db0fd fixed a certain problem.
886You'd like to find the earliest tagged release that contains that
887fix.
888
889Of course, there may be more than one answer--if the history branched
890after commit e05db0fd, then there could be multiple "earliest" tagged
891releases.
892
893You could just visually inspect the commits since e05db0fd:
894
895-------------------------------------------------
896$ gitk e05db0fd..
897-------------------------------------------------
898
b181d57f
BF
899Or you can use gitlink:git-name-rev[1], which will give the commit a
900name based on any tag it finds pointing to one of the commit's
901descendants:
902
903-------------------------------------------------
904$ git name-rev e05db0fd
905e05db0fd tags/v1.5.0-rc1^0~23
906-------------------------------------------------
907
908The gitlink:git-describe[1] command does the opposite, naming the
909revision using a tag on which the given commit is based:
910
911-------------------------------------------------
912$ git describe e05db0fd
913v1.5.0-rc0-ge05db0f
914-------------------------------------------------
915
916but that may sometimes help you guess which tags might come after the
917given commit.
918
919If you just want to verify whether a given tagged version contains a
920given commit, you could use gitlink:git-merge-base[1]:
921
922-------------------------------------------------
923$ git merge-base e05db0fd v1.5.0-rc1
924e05db0fd4f31dde7005f075a84f96b360d05984b
925-------------------------------------------------
926
927The merge-base command finds a common ancestor of the given commits,
928and always returns one or the other in the case where one is a
929descendant of the other; so the above output shows that e05db0fd
930actually is an ancestor of v1.5.0-rc1.
931
932Alternatively, note that
933
934-------------------------------------------------
935$ git log v1.5.0-rc1..305db0fd
936-------------------------------------------------
937
938will produce empty output if and only if v1.5.0-rc1 includes 305db0fd,
939because it outputs only commits that are not reachable from v1.5.0-rc1.
aec053bb 940
d19fbc3c
BF
941Developing with git
942===================
943
944Telling git your name
945---------------------
946
947Before creating any commits, you should introduce yourself to git. The
948easiest way to do so is:
949
950------------------------------------------------
951$ cat >~/.gitconfig <<\EOF
952[user]
953 name = Your Name Comes Here
954 email = you@yourdomain.example.com
955EOF
956------------------------------------------------
957
958
959Creating a new repository
960-------------------------
961
962Creating a new repository from scratch is very easy:
963
964-------------------------------------------------
965$ mkdir project
966$ cd project
f1d2b477 967$ git init
d19fbc3c
BF
968-------------------------------------------------
969
970If you have some initial content (say, a tarball):
971
972-------------------------------------------------
973$ tar -xzvf project.tar.gz
974$ cd project
f1d2b477 975$ git init
d19fbc3c
BF
976$ git add . # include everything below ./ in the first commit:
977$ git commit
978-------------------------------------------------
979
980[[how-to-make-a-commit]]
981how to make a commit
982--------------------
983
984Creating a new commit takes three steps:
985
986 1. Making some changes to the working directory using your
987 favorite editor.
988 2. Telling git about your changes.
989 3. Creating the commit using the content you told git about
990 in step 2.
991
992In practice, you can interleave and repeat steps 1 and 2 as many
993times as you want: in order to keep track of what you want committed
994at step 3, git maintains a snapshot of the tree's contents in a
995special staging area called "the index."
996
01997b4a
BF
997At the beginning, the content of the index will be identical to
998that of the HEAD. The command "git diff --cached", which shows
999the difference between the HEAD and the index, should therefore
1000produce no output at that point.
eb6ae7f4 1001
d19fbc3c
BF
1002Modifying the index is easy:
1003
1004To update the index with the new contents of a modified file, use
1005
1006-------------------------------------------------
1007$ git add path/to/file
1008-------------------------------------------------
1009
1010To add the contents of a new file to the index, use
1011
1012-------------------------------------------------
1013$ git add path/to/file
1014-------------------------------------------------
1015
eb6ae7f4 1016To remove a file from the index and from the working tree,
d19fbc3c
BF
1017
1018-------------------------------------------------
1019$ git rm path/to/file
1020-------------------------------------------------
1021
1022After each step you can verify that
1023
1024-------------------------------------------------
1025$ git diff --cached
1026-------------------------------------------------
1027
1028always shows the difference between the HEAD and the index file--this
1029is what you'd commit if you created the commit now--and that
1030
1031-------------------------------------------------
1032$ git diff
1033-------------------------------------------------
1034
1035shows the difference between the working tree and the index file.
1036
1037Note that "git add" always adds just the current contents of a file
1038to the index; further changes to the same file will be ignored unless
1039you run git-add on the file again.
1040
1041When you're ready, just run
1042
1043-------------------------------------------------
1044$ git commit
1045-------------------------------------------------
1046
1047and git will prompt you for a commit message and then create the new
1048commmit. Check to make sure it looks like what you expected with
1049
1050-------------------------------------------------
1051$ git show
1052-------------------------------------------------
1053
1054As a special shortcut,
1055
1056-------------------------------------------------
1057$ git commit -a
1058-------------------------------------------------
1059
1060will update the index with any files that you've modified or removed
1061and create a commit, all in one step.
1062
1063A number of commands are useful for keeping track of what you're
1064about to commit:
1065
1066-------------------------------------------------
1067$ git diff --cached # difference between HEAD and the index; what
1068 # would be commited if you ran "commit" now.
1069$ git diff # difference between the index file and your
1070 # working directory; changes that would not
1071 # be included if you ran "commit" now.
1072$ git status # a brief per-file summary of the above.
1073-------------------------------------------------
1074
1075creating good commit messages
1076-----------------------------
1077
1078Though not required, it's a good idea to begin the commit message
1079with a single short (less than 50 character) line summarizing the
1080change, followed by a blank line and then a more thorough
1081description. Tools that turn commits into email, for example, use
1082the first line on the Subject line and the rest of the commit in the
1083body.
1084
1085how to merge
1086------------
1087
1088You can rejoin two diverging branches of development using
1089gitlink:git-merge[1]:
1090
1091-------------------------------------------------
1092$ git merge branchname
1093-------------------------------------------------
1094
1095merges the development in the branch "branchname" into the current
1096branch. If there are conflicts--for example, if the same file is
1097modified in two different ways in the remote branch and the local
1098branch--then you are warned; the output may look something like this:
1099
1100-------------------------------------------------
1101$ git pull . next
1102Trying really trivial in-index merge...
1103fatal: Merge requires file-level merging
1104Nope.
1105Merging HEAD with 77976da35a11db4580b80ae27e8d65caf5208086
1106Merging:
110715e2162 world
110877976da goodbye
1109found 1 common ancestor(s):
1110d122ed4 initial
1111Auto-merging file.txt
1112CONFLICT (content): Merge conflict in file.txt
1113Automatic merge failed; fix conflicts and then commit the result.
1114-------------------------------------------------
1115
1116Conflict markers are left in the problematic files, and after
1117you resolve the conflicts manually, you can update the index
1118with the contents and run git commit, as you normally would when
1119creating a new file.
1120
1121If you examine the resulting commit using gitk, you will see that it
1122has two parents, one pointing to the top of the current branch, and
1123one to the top of the other branch.
1124
1125In more detail:
1126
1127[[resolving-a-merge]]
1128Resolving a merge
1129-----------------
1130
1131When a merge isn't resolved automatically, git leaves the index and
1132the working tree in a special state that gives you all the
1133information you need to help resolve the merge.
1134
1135Files with conflicts are marked specially in the index, so until you
1136resolve the problem and update the index, git commit will fail:
1137
1138-------------------------------------------------
1139$ git commit
1140file.txt: needs merge
1141-------------------------------------------------
1142
1143Also, git status will list those files as "unmerged".
1144
1145All of the changes that git was able to merge automatically are
1146already added to the index file, so gitlink:git-diff[1] shows only
1147the conflicts. Also, it uses a somewhat unusual syntax:
1148
1149-------------------------------------------------
1150$ git diff
1151diff --cc file.txt
1152index 802992c,2b60207..0000000
1153--- a/file.txt
1154+++ b/file.txt
1155@@@ -1,1 -1,1 +1,5 @@@
1156++<<<<<<< HEAD:file.txt
1157 +Hello world
1158++=======
1159+ Goodbye
1160++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1161-------------------------------------------------
1162
1163Recall that the commit which will be commited after we resolve this
1164conflict will have two parents instead of the usual one: one parent
1165will be HEAD, the tip of the current branch; the other will be the
1166tip of the other branch, which is stored temporarily in MERGE_HEAD.
1167
1168The diff above shows the differences between the working-tree version
1169of file.txt and two previous version: one version from HEAD, and one
1170from MERGE_HEAD. So instead of preceding each line by a single "+"
1171or "-", it now uses two columns: the first column is used for
1172differences between the first parent and the working directory copy,
1173and the second for differences between the second parent and the
1174working directory copy. Thus after resolving the conflict in the
1175obvious way, the diff will look like:
1176
1177-------------------------------------------------
1178$ git diff
1179diff --cc file.txt
1180index 802992c,2b60207..0000000
1181--- a/file.txt
1182+++ b/file.txt
1183@@@ -1,1 -1,1 +1,1 @@@
1184- Hello world
1185 -Goodbye
1186++Goodbye world
1187-------------------------------------------------
1188
1189This shows that our resolved version deleted "Hello world" from the
1190first parent, deleted "Goodbye" from the second parent, and added
1191"Goodbye world", which was previously absent from both.
1192
1193The gitlink:git-log[1] command also provides special help for merges:
1194
1195-------------------------------------------------
1196$ git log --merge
1197-------------------------------------------------
1198
1199This will list all commits which exist only on HEAD or on MERGE_HEAD,
1200and which touch an unmerged file.
1201
1202We can now add the resolved version to the index and commit:
1203
1204-------------------------------------------------
1205$ git add file.txt
1206$ git commit
1207-------------------------------------------------
1208
1209Note that the commit message will already be filled in for you with
1210some information about the merge. Normally you can just use this
1211default message unchanged, but you may add additional commentary of
1212your own if desired.
1213
1214[[undoing-a-merge]]
1215undoing a merge
1216---------------
1217
1218If you get stuck and decide to just give up and throw the whole mess
1219away, you can always return to the pre-merge state with
1220
1221-------------------------------------------------
1222$ git reset --hard HEAD
1223-------------------------------------------------
1224
1225Or, if you've already commited the merge that you want to throw away,
1226
1227-------------------------------------------------
1228$ git reset --hard HEAD^
1229-------------------------------------------------
1230
1231However, this last command can be dangerous in some cases--never
1232throw away a commit you have already committed if that commit may
1233itself have been merged into another branch, as doing so may confuse
1234further merges.
1235
1236Fast-forward merges
1237-------------------
1238
1239There is one special case not mentioned above, which is treated
1240differently. Normally, a merge results in a merge commit, with two
1241parents, one pointing at each of the two lines of development that
1242were merged.
1243
1244However, if one of the two lines of development is completely
1245contained within the other--so every commit present in the one is
1246already contained in the other--then git just performs a
1247<<fast-forwards,fast forward>>; the head of the current branch is
1248moved forward to point at the head of the merged-in branch, without
1249any new commits being created.
1250
b684f830
BF
1251Fixing mistakes
1252---------------
1253
1254If you've messed up the working tree, but haven't yet committed your
1255mistake, you can return the entire working tree to the last committed
1256state with
1257
1258-------------------------------------------------
1259$ git reset --hard HEAD
1260-------------------------------------------------
1261
1262If you make a commit that you later wish you hadn't, there are two
1263fundamentally different ways to fix the problem:
1264
1265 1. You can create a new commit that undoes whatever was done
1266 by the previous commit. This is the correct thing if your
1267 mistake has already been made public.
1268
1269 2. You can go back and modify the old commit. You should
1270 never do this if you have already made the history public;
1271 git does not normally expect the "history" of a project to
1272 change, and cannot correctly perform repeated merges from
1273 a branch that has had its history changed.
1274
1275Fixing a mistake with a new commit
1276~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1277
1278Creating a new commit that reverts an earlier change is very easy;
1279just pass the gitlink:git-revert[1] command a reference to the bad
1280commit; for example, to revert the most recent commit:
1281
1282-------------------------------------------------
1283$ git revert HEAD
1284-------------------------------------------------
1285
1286This will create a new commit which undoes the change in HEAD. You
1287will be given a chance to edit the commit message for the new commit.
1288
1289You can also revert an earlier change, for example, the next-to-last:
1290
1291-------------------------------------------------
1292$ git revert HEAD^
1293-------------------------------------------------
1294
1295In this case git will attempt to undo the old change while leaving
1296intact any changes made since then. If more recent changes overlap
1297with the changes to be reverted, then you will be asked to fix
1298conflicts manually, just as in the case of <<resolving-a-merge,
1299resolving a merge>>.
1300
1301Fixing a mistake by editing history
1302~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1303
1304If the problematic commit is the most recent commit, and you have not
1305yet made that commit public, then you may just
1306<<undoing-a-merge,destroy it using git-reset>>.
1307
1308Alternatively, you
1309can edit the working directory and update the index to fix your
1310mistake, just as if you were going to <<how-to-make-a-commit,create a
1311new commit>>, then run
1312
1313-------------------------------------------------
1314$ git commit --amend
1315-------------------------------------------------
1316
1317which will replace the old commit by a new commit incorporating your
1318changes, giving you a chance to edit the old commit message first.
1319
1320Again, you should never do this to a commit that may already have
1321been merged into another branch; use gitlink:git-revert[1] instead in
1322that case.
1323
1324It is also possible to edit commits further back in the history, but
1325this is an advanced topic to be left for
1326<<cleaning-up-history,another chapter>>.
1327
1328Checking out an old version of a file
1329~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1330
1331In the process of undoing a previous bad change, you may find it
1332useful to check out an older version of a particular file using
1333gitlink:git-checkout[1]. We've used git checkout before to switch
1334branches, but it has quite different behavior if it is given a path
1335name: the command
1336
1337-------------------------------------------------
1338$ git checkout HEAD^ path/to/file
1339-------------------------------------------------
1340
1341replaces path/to/file by the contents it had in the commit HEAD^, and
1342also updates the index to match. It does not change branches.
1343
1344If you just want to look at an old version of the file, without
1345modifying the working directory, you can do that with
1346gitlink:git-show[1]:
1347
1348-------------------------------------------------
1349$ git show HEAD^ path/to/file
1350-------------------------------------------------
1351
1352which will display the given version of the file.
1353
d19fbc3c
BF
1354Ensuring good performance
1355-------------------------
1356
1357On large repositories, git depends on compression to keep the history
1358information from taking up to much space on disk or in memory.
1359
1360This compression is not performed automatically. Therefore you
17217090 1361should occasionally run gitlink:git-gc[1]:
d19fbc3c
BF
1362
1363-------------------------------------------------
1364$ git gc
1365-------------------------------------------------
1366
17217090
BF
1367to recompress the archive. This can be very time-consuming, so
1368you may prefer to run git-gc when you are not doing other work.
d19fbc3c 1369
11e016a3
BF
1370Ensuring reliability
1371--------------------
1372
1373Checking the repository for corruption
1374~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1375
21dcb3b7
BF
1376The gitlink:git-fsck-objects[1] command runs a number of self-consistency
1377checks on the repository, and reports on any problems. This may take some
1378time. The most common warning by far is about "dangling" objects:
1379
1380-------------------------------------------------
1381$ git fsck-objects
1382dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1383dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1384dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1385dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1386dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1387dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1388dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1389dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1390...
1391-------------------------------------------------
1392
1393Dangling objects are objects that are harmless, but also unnecessary; you can
1394remove them at any time with gitlink:git-prune[1] or the --prune option to
1395gitlink:git-gc[1]:
1396
1397-------------------------------------------------
1398$ git gc --prune
1399-------------------------------------------------
1400
1401This may be time-consuming. Unlike most other git operations (including git-gc
1402when run without any options), it is not safe to prune while other git
1403operations are in progress in the same repository.
1404
1405For more about dangling merges, see <<dangling-merges>>.
1406
11e016a3
BF
1407
1408Recovering lost changes
1409~~~~~~~~~~~~~~~~~~~~~~~
1410
1411TODO:
1412 reflog
1413 git-fsck
1414 low-level examination of objects
1415
d19fbc3c 1416Sharing development with others
b684f830 1417===============================
d19fbc3c
BF
1418
1419[[getting-updates-with-git-pull]]
1420Getting updates with git pull
b684f830 1421-----------------------------
d19fbc3c
BF
1422
1423After you clone a repository and make a few changes of your own, you
1424may wish to check the original repository for updates and merge them
1425into your own work.
1426
1427We have already seen <<Updating-a-repository-with-git-fetch,how to
1428keep remote tracking branches up to date>> with gitlink:git-fetch[1],
1429and how to merge two branches. So you can merge in changes from the
1430original repository's master branch with:
1431
1432-------------------------------------------------
1433$ git fetch
1434$ git merge origin/master
1435-------------------------------------------------
1436
1437However, the gitlink:git-pull[1] command provides a way to do this in
1438one step:
1439
1440-------------------------------------------------
1441$ git pull origin master
1442-------------------------------------------------
1443
1444In fact, "origin" is normally the default repository to pull from,
1445and the default branch is normally the HEAD of the remote repository,
1446so often you can accomplish the above with just
1447
1448-------------------------------------------------
1449$ git pull
1450-------------------------------------------------
1451
1452See the descriptions of the branch.<name>.remote and
1453branch.<name>.merge options in gitlink:git-repo-config[1] to learn
1454how to control these defaults depending on the current branch.
1455
1456In addition to saving you keystrokes, "git pull" also helps you by
1457producing a default commit message documenting the branch and
1458repository that you pulled from.
1459
1460(But note that no such commit will be created in the case of a
1461<<fast-forwards,fast forward>>; instead, your branch will just be
1462updated to point to the latest commit from the upstream branch).
1463
4c63ff45
BF
1464The git-pull command can also be given "." as the "remote" repository, in
1465which case it just merges in a branch from the current repository; so
1466the commands
1467
1468-------------------------------------------------
1469$ git pull . branch
1470$ git merge branch
1471-------------------------------------------------
1472
1473are roughly equivalent. The former is actually very commonly used.
1474
d19fbc3c 1475Submitting patches to a project
b684f830 1476-------------------------------
d19fbc3c
BF
1477
1478If you just have a few changes, the simplest way to submit them may
1479just be to send them as patches in email:
1480
1481First, use gitlink:git-format-patches[1]; for example:
1482
1483-------------------------------------------------
eb6ae7f4 1484$ git format-patch origin
d19fbc3c
BF
1485-------------------------------------------------
1486
1487will produce a numbered series of files in the current directory, one
1488for each patch in the current branch but not in origin/HEAD.
1489
1490You can then import these into your mail client and send them by
1491hand. However, if you have a lot to send at once, you may prefer to
1492use the gitlink:git-send-email[1] script to automate the process.
1493Consult the mailing list for your project first to determine how they
1494prefer such patches be handled.
1495
1496Importing patches to a project
b684f830 1497------------------------------
d19fbc3c
BF
1498
1499Git also provides a tool called gitlink:git-am[1] (am stands for
1500"apply mailbox"), for importing such an emailed series of patches.
1501Just save all of the patch-containing messages, in order, into a
1502single mailbox file, say "patches.mbox", then run
1503
1504-------------------------------------------------
eb6ae7f4 1505$ git am -3 patches.mbox
d19fbc3c
BF
1506-------------------------------------------------
1507
1508Git will apply each patch in order; if any conflicts are found, it
1509will stop, and you can fix the conflicts as described in
01997b4a
BF
1510"<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1511git to perform a merge; if you would prefer it just to abort and
1512leave your tree and index untouched, you may omit that option.)
1513
1514Once the index is updated with the results of the conflict
1515resolution, instead of creating a new commit, just run
d19fbc3c
BF
1516
1517-------------------------------------------------
1518$ git am --resolved
1519-------------------------------------------------
1520
1521and git will create the commit for you and continue applying the
1522remaining patches from the mailbox.
1523
1524The final result will be a series of commits, one for each patch in
1525the original mailbox, with authorship and commit log message each
1526taken from the message containing each patch.
1527
1528[[setting-up-a-public-repository]]
1529Setting up a public repository
b684f830 1530------------------------------
d19fbc3c
BF
1531
1532Another way to submit changes to a project is to simply tell the
1533maintainer of that project to pull from your repository, exactly as
1534you did in the section "<<getting-updates-with-git-pull, Getting
1535updates with git pull>>".
1536
1537If you and maintainer both have accounts on the same machine, then
1538then you can just pull changes from each other's repositories
1539directly; note that all of the command (gitlink:git-clone[1],
1540git-fetch[1], git-pull[1], etc.) which accept a URL as an argument
1541will also accept a local file patch; so, for example, you can
1542use
1543
1544-------------------------------------------------
1545$ git clone /path/to/repository
1546$ git pull /path/to/other/repository
1547-------------------------------------------------
1548
1549If this sort of setup is inconvenient or impossible, another (more
1550common) option is to set up a public repository on a public server.
1551This also allows you to cleanly separate private work in progress
1552from publicly visible work.
1553
1554You will continue to do your day-to-day work in your personal
1555repository, but periodically "push" changes from your personal
1556repository into your public repository, allowing other developers to
1557pull from that repository. So the flow of changes, in a situation
1558where there is one other developer with a public repository, looks
1559like this:
1560
1561 you push
1562 your personal repo ------------------> your public repo
1563 ^ |
1564 | |
1565 | you pull | they pull
1566 | |
1567 | |
1568 | they push V
1569 their public repo <------------------- their repo
1570
1571Now, assume your personal repository is in the directory ~/proj. We
1572first create a new clone of the repository:
1573
1574-------------------------------------------------
1575$ git clone --bare proj-clone.git
1576-------------------------------------------------
1577
1578The resulting directory proj-clone.git will contains a "bare" git
1579repository--it is just the contents of the ".git" directory, without
1580a checked-out copy of a working directory.
1581
1582Next, copy proj-clone.git to the server where you plan to host the
1583public repository. You can use scp, rsync, or whatever is most
1584convenient.
1585
1586If somebody else maintains the public server, they may already have
1587set up a git service for you, and you may skip to the section
1588"<<pushing-changes-to-a-public-repository,Pushing changes to a public
1589repository>>", below.
1590
1591Otherwise, the following sections explain how to export your newly
1592created public repository:
1593
1594[[exporting-via-http]]
1595Exporting a git repository via http
b684f830 1596-----------------------------------
d19fbc3c
BF
1597
1598The git protocol gives better performance and reliability, but on a
1599host with a web server set up, http exports may be simpler to set up.
1600
1601All you need to do is place the newly created bare git repository in
1602a directory that is exported by the web server, and make some
1603adjustments to give web clients some extra information they need:
1604
1605-------------------------------------------------
1606$ mv proj.git /home/you/public_html/proj.git
1607$ cd proj.git
1608$ git update-server-info
1609$ chmod a+x hooks/post-update
1610-------------------------------------------------
1611
1612(For an explanation of the last two lines, see
1613gitlink:git-update-server-info[1], and the documentation
1614link:hooks.txt[Hooks used by git].)
1615
1616Advertise the url of proj.git. Anybody else should then be able to
1617clone or pull from that url, for example with a commandline like:
1618
1619-------------------------------------------------
1620$ git clone http://yourserver.com/~you/proj.git
1621-------------------------------------------------
1622
1623(See also
1624link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1625for a slightly more sophisticated setup using WebDAV which also
1626allows pushing over http.)
1627
1628[[exporting-via-git]]
1629Exporting a git repository via the git protocol
b684f830 1630-----------------------------------------------
d19fbc3c
BF
1631
1632This is the preferred method.
1633
1634For now, we refer you to the gitlink:git-daemon[1] man page for
1635instructions. (See especially the examples section.)
1636
1637[[pushing-changes-to-a-public-repository]]
1638Pushing changes to a public repository
b684f830 1639--------------------------------------
d19fbc3c
BF
1640
1641Note that the two techniques outline above (exporting via
1642<<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1643maintainers to fetch your latest changes, but they do not allow write
1644access, which you will need to update the public repository with the
1645latest changes created in your private repository.
1646
1647The simplest way to do this is using gitlink:git-push[1] and ssh; to
1648update the remote branch named "master" with the latest state of your
1649branch named "master", run
1650
1651-------------------------------------------------
1652$ git push ssh://yourserver.com/~you/proj.git master:master
1653-------------------------------------------------
1654
1655or just
1656
1657-------------------------------------------------
1658$ git push ssh://yourserver.com/~you/proj.git master
1659-------------------------------------------------
1660
1661As with git-fetch, git-push will complain if this does not result in
1662a <<fast-forwards,fast forward>>. Normally this is a sign of
1663something wrong. However, if you are sure you know what you're
1664doing, you may force git-push to perform the update anyway by
1665proceeding the branch name by a plus sign:
1666
1667-------------------------------------------------
1668$ git push ssh://yourserver.com/~you/proj.git +master
1669-------------------------------------------------
1670
1671As with git-fetch, you may also set up configuration options to
1672save typing; so, for example, after
1673
1674-------------------------------------------------
1675$ cat >.git/config <<EOF
1676[remote "public-repo"]
1677 url = ssh://yourserver.com/~you/proj.git
1678EOF
1679-------------------------------------------------
1680
1681you should be able to perform the above push with just
1682
1683-------------------------------------------------
1684$ git push public-repo master
1685-------------------------------------------------
1686
1687See the explanations of the remote.<name>.url, branch.<name>.remote,
1688and remote.<name>.push options in gitlink:git-repo-config[1] for
1689details.
1690
1691Setting up a shared repository
b684f830 1692------------------------------
d19fbc3c
BF
1693
1694Another way to collaborate is by using a model similar to that
1695commonly used in CVS, where several developers with special rights
1696all push to and pull from a single shared repository. See
1697link:cvs-migration.txt[git for CVS users] for instructions on how to
1698set this up.
1699
b684f830
BF
1700Allow web browsing of a repository
1701----------------------------------
d19fbc3c 1702
b684f830 1703TODO: Brief setup-instructions for gitweb
d19fbc3c 1704
b684f830
BF
1705Examples
1706--------
d19fbc3c 1707
b684f830 1708TODO: topic branches, typical roles as in everyday.txt, ?
d19fbc3c 1709
d19fbc3c
BF
1710
1711Working with other version control systems
1712==========================================
1713
b181d57f 1714TODO: CVS, Subversion, series-of-release-tarballs, etc.
d19fbc3c
BF
1715
1716[[cleaning-up-history]]
4c63ff45
BF
1717Rewriting history and maintaining patch series
1718==============================================
1719
1720Normally commits are only added to a project, never taken away or
1721replaced. Git is designed with this assumption, and violating it will
1722cause git's merge machinery (for example) to do the wrong thing.
1723
1724However, there is a situation in which it can be useful to violate this
1725assumption.
1726
1727Creating the perfect patch series
1728---------------------------------
1729
1730Suppose you are a contributor to a large project, and you want to add a
1731complicated feature, and to present it to the other developers in a way
1732that makes it easy for them to read your changes, verify that they are
1733correct, and understand why you made each change.
1734
b181d57f
BF
1735If you present all of your changes as a single patch (or commit), they
1736may find it is too much to digest all at once.
4c63ff45
BF
1737
1738If you present them with the entire history of your work, complete with
1739mistakes, corrections, and dead ends, they may be overwhelmed.
1740
1741So the ideal is usually to produce a series of patches such that:
1742
1743 1. Each patch can be applied in order.
1744
1745 2. Each patch includes a single logical change, together with a
1746 message explaining the change.
1747
1748 3. No patch introduces a regression: after applying any initial
1749 part of the series, the resulting project still compiles and
1750 works, and has no bugs that it didn't have before.
1751
1752 4. The complete series produces the same end result as your own
1753 (probably much messier!) development process did.
1754
b181d57f
BF
1755We will introduce some tools that can help you do this, explain how to
1756use them, and then explain some of the problems that can arise because
1757you are rewriting history.
4c63ff45
BF
1758
1759Keeping a patch series up to date using git-rebase
1760--------------------------------------------------
1761
1762Suppose you have a series of commits in a branch "mywork", which
1763originally branched off from "origin".
1764
b181d57f
BF
1765Suppose you create a branch "mywork" on a remote-tracking branch
1766"origin", and created some commits on top of it:
4c63ff45
BF
1767
1768-------------------------------------------------
1769$ git checkout -b mywork origin
1770$ vi file.txt
1771$ git commit
1772$ vi otherfile.txt
1773$ git commit
1774...
1775-------------------------------------------------
1776
1777You have performed no merges into mywork, so it is just a simple linear
1778sequence of patches on top of "origin":
1779
1780
1781 o--o--o <-- origin
1782 \
1783 o--o--o <-- mywork
1784
1785Some more interesting work has been done in the upstream project, and
1786"origin" has advanced:
1787
1788 o--o--O--o--o--o <-- origin
1789 \
1790 a--b--c <-- mywork
1791
1792At this point, you could use "pull" to merge your changes back in;
1793the result would create a new merge commit, like this:
1794
1795
1796 o--o--O--o--o--o <-- origin
1797 \ \
1798 a--b--c--m <-- mywork
1799
1800However, if you prefer to keep the history in mywork a simple series of
1801commits without any merges, you may instead choose to use
1802gitlink:git-rebase[1]:
1803
1804-------------------------------------------------
1805$ git checkout mywork
1806$ git rebase origin
1807-------------------------------------------------
1808
b181d57f
BF
1809This will remove each of your commits from mywork, temporarily saving
1810them as patches (in a directory named ".dotest"), update mywork to
1811point at the latest version of origin, then apply each of the saved
1812patches to the new mywork. The result will look like:
4c63ff45
BF
1813
1814
1815 o--o--O--o--o--o <-- origin
1816 \
1817 a'--b'--c' <-- mywork
1818
b181d57f
BF
1819In the process, it may discover conflicts. In that case it will stop
1820and allow you to fix the conflicts; after fixing conflicts, use "git
1821add" to update the index with those contents, and then, instead of
1822running git-commit, just run
4c63ff45
BF
1823
1824-------------------------------------------------
1825$ git rebase --continue
1826-------------------------------------------------
1827
1828and git will continue applying the rest of the patches.
1829
1830At any point you may use the --abort option to abort this process and
1831return mywork to the state it had before you started the rebase:
1832
1833-------------------------------------------------
1834$ git rebase --abort
1835-------------------------------------------------
1836
1837Reordering or selecting from a patch series
1838-------------------------------------------
1839
b181d57f
BF
1840Given one existing commit, the gitlink:git-cherry-pick[1] command
1841allows you to apply the change introduced by that commit and create a
1842new commit that records it. So, for example, if "mywork" points to a
1843series of patches on top of "origin", you might do something like:
1844
1845-------------------------------------------------
1846$ git checkout -b mywork-new origin
1847$ gitk origin..mywork &
1848-------------------------------------------------
1849
1850And browse through the list of patches in the mywork branch using gitk,
1851applying them (possibly in a different order) to mywork-new using
1852cherry-pick, and possibly modifying them as you go using commit
1853--amend.
1854
1855Another technique is to use git-format-patch to create a series of
1856patches, then reset the state to before the patches:
4c63ff45 1857
b181d57f
BF
1858-------------------------------------------------
1859$ git format-patch origin
1860$ git reset --hard origin
1861-------------------------------------------------
4c63ff45 1862
b181d57f
BF
1863Then modify, reorder, or eliminate patches as preferred before applying
1864them again with gitlink:git-am[1].
4c63ff45
BF
1865
1866Other tools
1867-----------
1868
b181d57f
BF
1869There are numerous other tools, such as stgit, which exist for the
1870purpose of maintaining a patch series. These are out of the scope of
1871this manual.
4c63ff45
BF
1872
1873Problems with rewriting history
1874-------------------------------
1875
b181d57f
BF
1876The primary problem with rewriting the history of a branch has to do
1877with merging. Suppose somebody fetches your branch and merges it into
1878their branch, with a result something like this:
1879
1880 o--o--O--o--o--o <-- origin
1881 \ \
1882 t--t--t--m <-- their branch:
1883
1884Then suppose you modify the last three commits:
1885
1886 o--o--o <-- new head of origin
1887 /
1888 o--o--O--o--o--o <-- old head of origin
1889
1890If we examined all this history together in one repository, it will
1891look like:
1892
1893 o--o--o <-- new head of origin
1894 /
1895 o--o--O--o--o--o <-- old head of origin
1896 \ \
1897 t--t--t--m <-- their branch:
1898
1899Git has no way of knowing that the new head is an updated version of
1900the old head; it treats this situation exactly the same as it would if
1901two developers had independently done the work on the old and new heads
1902in parallel. At this point, if someone attempts to merge the new head
1903in to their branch, git will attempt to merge together the two (old and
1904new) lines of development, instead of trying to replace the old by the
1905new. The results are likely to be unexpected.
1906
1907You may still choose to publish branches whose history is rewritten,
1908and it may be useful for others to be able to fetch those branches in
1909order to examine or test them, but they should not attempt to pull such
1910branches into their own work.
1911
1912For true distributed development that supports proper merging,
1913published branches should never be rewritten.
1914
1915Advanced branch management
1916==========================
4c63ff45 1917
b181d57f
BF
1918Fetching individual branches
1919----------------------------
1920
1921Instead of using gitlink:git-remote[1], you can also choose just
1922to update one branch at a time, and to store it locally under an
1923arbitrary name:
1924
1925-------------------------------------------------
1926$ git fetch origin todo:my-todo-work
1927-------------------------------------------------
1928
1929The first argument, "origin", just tells git to fetch from the
1930repository you originally cloned from. The second argument tells git
1931to fetch the branch named "todo" from the remote repository, and to
1932store it locally under the name refs/heads/my-todo-work.
1933
1934You can also fetch branches from other repositories; so
1935
1936-------------------------------------------------
1937$ git fetch git://example.com/proj.git master:example-master
1938-------------------------------------------------
1939
1940will create a new branch named "example-master" and store in it the
1941branch named "master" from the repository at the given URL. If you
1942already have a branch named example-master, it will attempt to
1943"fast-forward" to the commit given by example.com's master branch. So
1944next we explain what a fast-forward is:
1945
1946[[fast-forwards]]
1947Understanding git history: fast-forwards
1948----------------------------------------
1949
1950In the previous example, when updating an existing branch, "git
1951fetch" checks to make sure that the most recent commit on the remote
1952branch is a descendant of the most recent commit on your copy of the
1953branch before updating your copy of the branch to point at the new
1954commit. Git calls this process a "fast forward".
1955
1956A fast forward looks something like this:
1957
1958 o--o--o--o <-- old head of the branch
1959 \
1960 o--o--o <-- new head of the branch
1961
1962
1963In some cases it is possible that the new head will *not* actually be
1964a descendant of the old head. For example, the developer may have
1965realized she made a serious mistake, and decided to backtrack,
1966resulting in a situation like:
1967
1968 o--o--o--o--a--b <-- old head of the branch
1969 \
1970 o--o--o <-- new head of the branch
1971
1972
1973
1974In this case, "git fetch" will fail, and print out a warning.
1975
1976In that case, you can still force git to update to the new head, as
1977described in the following section. However, note that in the
1978situation above this may mean losing the commits labeled "a" and "b",
1979unless you've already created a reference of your own pointing to
1980them.
1981
1982Forcing git fetch to do non-fast-forward updates
1983------------------------------------------------
1984
1985If git fetch fails because the new head of a branch is not a
1986descendant of the old head, you may force the update with:
1987
1988-------------------------------------------------
1989$ git fetch git://example.com/proj.git +master:refs/remotes/example/master
1990-------------------------------------------------
1991
1992Note the addition of the "+" sign. Be aware that commits which the
1993old version of example/master pointed at may be lost, as we saw in
1994the previous section.
1995
1996Configuring remote branches
1997---------------------------
1998
1999We saw above that "origin" is just a shortcut to refer to the
2000repository which you originally cloned from. This information is
2001stored in git configuration variables, which you can see using
2002gitlink:git-repo-config[1]:
2003
2004-------------------------------------------------
2005$ git-repo-config -l
2006core.repositoryformatversion=0
2007core.filemode=true
2008core.logallrefupdates=true
2009remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2010remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2011branch.master.remote=origin
2012branch.master.merge=refs/heads/master
2013-------------------------------------------------
2014
2015If there are other repositories that you also use frequently, you can
2016create similar configuration options to save typing; for example,
2017after
2018
2019-------------------------------------------------
2020$ git repo-config remote.example.url git://example.com/proj.git
2021-------------------------------------------------
2022
2023then the following two commands will do the same thing:
2024
2025-------------------------------------------------
2026$ git fetch git://example.com/proj.git master:refs/remotes/example/master
2027$ git fetch example master:refs/remotes/example/master
2028-------------------------------------------------
2029
2030Even better, if you add one more option:
2031
2032-------------------------------------------------
2033$ git repo-config remote.example.fetch master:refs/remotes/example/master
2034-------------------------------------------------
2035
2036then the following commands will all do the same thing:
2037
2038-------------------------------------------------
2039$ git fetch git://example.com/proj.git master:ref/remotes/example/master
2040$ git fetch example master:ref/remotes/example/master
2041$ git fetch example example/master
2042$ git fetch example
2043-------------------------------------------------
2044
2045You can also add a "+" to force the update each time:
2046
2047-------------------------------------------------
2048$ git repo-config remote.example.fetch +master:ref/remotes/example/master
2049-------------------------------------------------
2050
2051Don't do this unless you're sure you won't mind "git fetch" possibly
2052throwing away commits on mybranch.
2053
2054Also note that all of the above configuration can be performed by
2055directly editing the file .git/config instead of using
2056gitlink:git-repo-config[1].
2057
2058See gitlink:git-repo-config[1] for more details on the configuration
2059options mentioned above.
d19fbc3c 2060
d19fbc3c
BF
2061
2062Git internals
2063=============
2064
b181d57f
BF
2065There are two object abstractions: the "object database", and the
2066"current directory cache" aka "index".
2067
2068The Object Database
2069-------------------
2070
2071The object database is literally just a content-addressable collection
2072of objects. All objects are named by their content, which is
2073approximated by the SHA1 hash of the object itself. Objects may refer
2074to other objects (by referencing their SHA1 hash), and so you can
2075build up a hierarchy of objects.
2076
2077All objects have a statically determined "type" aka "tag", which is
2078determined at object creation time, and which identifies the format of
2079the object (i.e. how it is used, and how it can refer to other
2080objects). There are currently four different object types: "blob",
2081"tree", "commit" and "tag".
2082
2083A "blob" object cannot refer to any other object, and is, like the type
2084implies, a pure storage object containing some user data. It is used to
2085actually store the file data, i.e. a blob object is associated with some
2086particular version of some file.
2087
2088A "tree" object is an object that ties one or more "blob" objects into a
2089directory structure. In addition, a tree object can refer to other tree
2090objects, thus creating a directory hierarchy.
2091
2092A "commit" object ties such directory hierarchies together into
2093a DAG of revisions - each "commit" is associated with exactly one tree
2094(the directory hierarchy at the time of the commit). In addition, a
2095"commit" refers to one or more "parent" commit objects that describe the
2096history of how we arrived at that directory hierarchy.
2097
2098As a special case, a commit object with no parents is called the "root"
2099object, and is the point of an initial project commit. Each project
2100must have at least one root, and while you can tie several different
2101root objects together into one project by creating a commit object which
2102has two or more separate roots as its ultimate parents, that's probably
2103just going to confuse people. So aim for the notion of "one root object
2104per project", even if git itself does not enforce that.
2105
2106A "tag" object symbolically identifies and can be used to sign other
2107objects. It contains the identifier and type of another object, a
2108symbolic name (of course!) and, optionally, a signature.
2109
2110Regardless of object type, all objects share the following
2111characteristics: they are all deflated with zlib, and have a header
2112that not only specifies their type, but also provides size information
2113about the data in the object. It's worth noting that the SHA1 hash
2114that is used to name the object is the hash of the original data
2115plus this header, so `sha1sum` 'file' does not match the object name
2116for 'file'.
2117(Historical note: in the dawn of the age of git the hash
2118was the sha1 of the 'compressed' object.)
2119
2120As a result, the general consistency of an object can always be tested
2121independently of the contents or the type of the object: all objects can
2122be validated by verifying that (a) their hashes match the content of the
2123file and (b) the object successfully inflates to a stream of bytes that
2124forms a sequence of <ascii type without space> + <space> + <ascii decimal
2125size> + <byte\0> + <binary object data>.
2126
2127The structured objects can further have their structure and
2128connectivity to other objects verified. This is generally done with
2129the `git-fsck-objects` program, which generates a full dependency graph
2130of all objects, and verifies their internal consistency (in addition
2131to just verifying their superficial consistency through the hash).
2132
2133The object types in some more detail:
2134
2135Blob Object
2136-----------
2137
2138A "blob" object is nothing but a binary blob of data, and doesn't
2139refer to anything else. There is no signature or any other
2140verification of the data, so while the object is consistent (it 'is'
2141indexed by its sha1 hash, so the data itself is certainly correct), it
2142has absolutely no other attributes. No name associations, no
2143permissions. It is purely a blob of data (i.e. normally "file
2144contents").
2145
2146In particular, since the blob is entirely defined by its data, if two
2147files in a directory tree (or in multiple different versions of the
2148repository) have the same contents, they will share the same blob
2149object. The object is totally independent of its location in the
2150directory tree, and renaming a file does not change the object that
2151file is associated with in any way.
2152
2153A blob is typically created when gitlink:git-update-index[1]
2154is run, and its data can be accessed by gitlink:git-cat-file[1].
2155
2156Tree Object
2157-----------
2158
2159The next hierarchical object type is the "tree" object. A tree object
2160is a list of mode/name/blob data, sorted by name. Alternatively, the
2161mode data may specify a directory mode, in which case instead of
2162naming a blob, that name is associated with another TREE object.
2163
2164Like the "blob" object, a tree object is uniquely determined by the
2165set contents, and so two separate but identical trees will always
2166share the exact same object. This is true at all levels, i.e. it's
2167true for a "leaf" tree (which does not refer to any other trees, only
2168blobs) as well as for a whole subdirectory.
2169
2170For that reason a "tree" object is just a pure data abstraction: it
2171has no history, no signatures, no verification of validity, except
2172that since the contents are again protected by the hash itself, we can
2173trust that the tree is immutable and its contents never change.
2174
2175So you can trust the contents of a tree to be valid, the same way you
2176can trust the contents of a blob, but you don't know where those
2177contents 'came' from.
2178
2179Side note on trees: since a "tree" object is a sorted list of
2180"filename+content", you can create a diff between two trees without
2181actually having to unpack two trees. Just ignore all common parts,
2182and your diff will look right. In other words, you can effectively
2183(and efficiently) tell the difference between any two random trees by
2184O(n) where "n" is the size of the difference, rather than the size of
2185the tree.
2186
2187Side note 2 on trees: since the name of a "blob" depends entirely and
2188exclusively on its contents (i.e. there are no names or permissions
2189involved), you can see trivial renames or permission changes by
2190noticing that the blob stayed the same. However, renames with data
2191changes need a smarter "diff" implementation.
2192
2193A tree is created with gitlink:git-write-tree[1] and
2194its data can be accessed by gitlink:git-ls-tree[1].
2195Two trees can be compared with gitlink:git-diff-tree[1].
2196
2197Commit Object
2198-------------
2199
2200The "commit" object is an object that introduces the notion of
2201history into the picture. In contrast to the other objects, it
2202doesn't just describe the physical state of a tree, it describes how
2203we got there, and why.
2204
2205A "commit" is defined by the tree-object that it results in, the
2206parent commits (zero, one or more) that led up to that point, and a
2207comment on what happened. Again, a commit is not trusted per se:
2208the contents are well-defined and "safe" due to the cryptographically
2209strong signatures at all levels, but there is no reason to believe
2210that the tree is "good" or that the merge information makes sense.
2211The parents do not have to actually have any relationship with the
2212result, for example.
2213
2214Note on commits: unlike real SCM's, commits do not contain
2215rename information or file mode change information. All of that is
2216implicit in the trees involved (the result tree, and the result trees
2217of the parents), and describing that makes no sense in this idiotic
2218file manager.
2219
2220A commit is created with gitlink:git-commit-tree[1] and
2221its data can be accessed by gitlink:git-cat-file[1].
2222
2223Trust
2224-----
2225
2226An aside on the notion of "trust". Trust is really outside the scope
2227of "git", but it's worth noting a few things. First off, since
2228everything is hashed with SHA1, you 'can' trust that an object is
2229intact and has not been messed with by external sources. So the name
2230of an object uniquely identifies a known state - just not a state that
2231you may want to trust.
2232
2233Furthermore, since the SHA1 signature of a commit refers to the
2234SHA1 signatures of the tree it is associated with and the signatures
2235of the parent, a single named commit specifies uniquely a whole set
2236of history, with full contents. You can't later fake any step of the
2237way once you have the name of a commit.
2238
2239So to introduce some real trust in the system, the only thing you need
2240to do is to digitally sign just 'one' special note, which includes the
2241name of a top-level commit. Your digital signature shows others
2242that you trust that commit, and the immutability of the history of
2243commits tells others that they can trust the whole history.
2244
2245In other words, you can easily validate a whole archive by just
2246sending out a single email that tells the people the name (SHA1 hash)
2247of the top commit, and digitally sign that email using something
2248like GPG/PGP.
2249
2250To assist in this, git also provides the tag object...
2251
2252Tag Object
2253----------
2254
2255Git provides the "tag" object to simplify creating, managing and
2256exchanging symbolic and signed tokens. The "tag" object at its
2257simplest simply symbolically identifies another object by containing
2258the sha1, type and symbolic name.
2259
2260However it can optionally contain additional signature information
2261(which git doesn't care about as long as there's less than 8k of
2262it). This can then be verified externally to git.
2263
2264Note that despite the tag features, "git" itself only handles content
2265integrity; the trust framework (and signature provision and
2266verification) has to come from outside.
2267
2268A tag is created with gitlink:git-mktag[1],
2269its data can be accessed by gitlink:git-cat-file[1],
2270and the signature can be verified by
2271gitlink:git-verify-tag[1].
2272
2273
2274The "index" aka "Current Directory Cache"
2275-----------------------------------------
2276
2277The index is a simple binary file, which contains an efficient
2278representation of a virtual directory content at some random time. It
2279does so by a simple array that associates a set of names, dates,
2280permissions and content (aka "blob") objects together. The cache is
2281always kept ordered by name, and names are unique (with a few very
2282specific rules) at any point in time, but the cache has no long-term
2283meaning, and can be partially updated at any time.
2284
2285In particular, the index certainly does not need to be consistent with
2286the current directory contents (in fact, most operations will depend on
2287different ways to make the index 'not' be consistent with the directory
2288hierarchy), but it has three very important attributes:
2289
2290'(a) it can re-generate the full state it caches (not just the
2291directory structure: it contains pointers to the "blob" objects so
2292that it can regenerate the data too)'
2293
2294As a special case, there is a clear and unambiguous one-way mapping
2295from a current directory cache to a "tree object", which can be
2296efficiently created from just the current directory cache without
2297actually looking at any other data. So a directory cache at any one
2298time uniquely specifies one and only one "tree" object (but has
2299additional data to make it easy to match up that tree object with what
2300has happened in the directory)
2301
2302'(b) it has efficient methods for finding inconsistencies between that
2303cached state ("tree object waiting to be instantiated") and the
2304current state.'
2305
2306'(c) it can additionally efficiently represent information about merge
2307conflicts between different tree objects, allowing each pathname to be
2308associated with sufficient information about the trees involved that
2309you can create a three-way merge between them.'
2310
2311Those are the three ONLY things that the directory cache does. It's a
2312cache, and the normal operation is to re-generate it completely from a
2313known tree object, or update/compare it with a live tree that is being
2314developed. If you blow the directory cache away entirely, you generally
2315haven't lost any information as long as you have the name of the tree
2316that it described.
2317
2318At the same time, the index is at the same time also the
2319staging area for creating new trees, and creating a new tree always
2320involves a controlled modification of the index file. In particular,
2321the index file can have the representation of an intermediate tree that
2322has not yet been instantiated. So the index can be thought of as a
2323write-back cache, which can contain dirty information that has not yet
2324been written back to the backing store.
2325
2326
2327
2328The Workflow
2329------------
2330
2331Generally, all "git" operations work on the index file. Some operations
2332work *purely* on the index file (showing the current state of the
2333index), but most operations move data to and from the index file. Either
2334from the database or from the working directory. Thus there are four
2335main combinations:
2336
2337working directory -> index
2338~~~~~~~~~~~~~~~~~~~~~~~~~~
2339
2340You update the index with information from the working directory with
2341the gitlink:git-update-index[1] command. You
2342generally update the index information by just specifying the filename
2343you want to update, like so:
2344
2345-------------------------------------------------
2346$ git-update-index filename
2347-------------------------------------------------
2348
2349but to avoid common mistakes with filename globbing etc, the command
2350will not normally add totally new entries or remove old entries,
2351i.e. it will normally just update existing cache entries.
2352
2353To tell git that yes, you really do realize that certain files no
2354longer exist, or that new files should be added, you
2355should use the `--remove` and `--add` flags respectively.
2356
2357NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
2358necessarily be removed: if the files still exist in your directory
2359structure, the index will be updated with their new status, not
2360removed. The only thing `--remove` means is that update-cache will be
2361considering a removed file to be a valid thing, and if the file really
2362does not exist any more, it will update the index accordingly.
2363
2364As a special case, you can also do `git-update-index --refresh`, which
2365will refresh the "stat" information of each index to match the current
2366stat information. It will 'not' update the object status itself, and
2367it will only update the fields that are used to quickly test whether
2368an object still matches its old backing store object.
2369
2370index -> object database
2371~~~~~~~~~~~~~~~~~~~~~~~~
2372
2373You write your current index file to a "tree" object with the program
2374
2375-------------------------------------------------
2376$ git-write-tree
2377-------------------------------------------------
2378
2379that doesn't come with any options - it will just write out the
2380current index into the set of tree objects that describe that state,
2381and it will return the name of the resulting top-level tree. You can
2382use that tree to re-generate the index at any time by going in the
2383other direction:
2384
2385object database -> index
2386~~~~~~~~~~~~~~~~~~~~~~~~
2387
2388You read a "tree" file from the object database, and use that to
2389populate (and overwrite - don't do this if your index contains any
2390unsaved state that you might want to restore later!) your current
2391index. Normal operation is just
2392
2393-------------------------------------------------
2394$ git-read-tree <sha1 of tree>
2395-------------------------------------------------
2396
2397and your index file will now be equivalent to the tree that you saved
2398earlier. However, that is only your 'index' file: your working
2399directory contents have not been modified.
2400
2401index -> working directory
2402~~~~~~~~~~~~~~~~~~~~~~~~~~
2403
2404You update your working directory from the index by "checking out"
2405files. This is not a very common operation, since normally you'd just
2406keep your files updated, and rather than write to your working
2407directory, you'd tell the index files about the changes in your
2408working directory (i.e. `git-update-index`).
2409
2410However, if you decide to jump to a new version, or check out somebody
2411else's version, or just restore a previous tree, you'd populate your
2412index file with read-tree, and then you need to check out the result
2413with
2414
2415-------------------------------------------------
2416$ git-checkout-index filename
2417-------------------------------------------------
2418
2419or, if you want to check out all of the index, use `-a`.
2420
2421NOTE! git-checkout-index normally refuses to overwrite old files, so
2422if you have an old version of the tree already checked out, you will
2423need to use the "-f" flag ('before' the "-a" flag or the filename) to
2424'force' the checkout.
2425
2426
2427Finally, there are a few odds and ends which are not purely moving
2428from one representation to the other:
2429
2430Tying it all together
2431~~~~~~~~~~~~~~~~~~~~~
2432
2433To commit a tree you have instantiated with "git-write-tree", you'd
2434create a "commit" object that refers to that tree and the history
2435behind it - most notably the "parent" commits that preceded it in
2436history.
2437
2438Normally a "commit" has one parent: the previous state of the tree
2439before a certain change was made. However, sometimes it can have two
2440or more parent commits, in which case we call it a "merge", due to the
2441fact that such a commit brings together ("merges") two or more
2442previous states represented by other commits.
2443
2444In other words, while a "tree" represents a particular directory state
2445of a working directory, a "commit" represents that state in "time",
2446and explains how we got there.
2447
2448You create a commit object by giving it the tree that describes the
2449state at the time of the commit, and a list of parents:
2450
2451-------------------------------------------------
2452$ git-commit-tree <tree> -p <parent> [-p <parent2> ..]
2453-------------------------------------------------
2454
2455and then giving the reason for the commit on stdin (either through
2456redirection from a pipe or file, or by just typing it at the tty).
2457
2458git-commit-tree will return the name of the object that represents
2459that commit, and you should save it away for later use. Normally,
2460you'd commit a new `HEAD` state, and while git doesn't care where you
2461save the note about that state, in practice we tend to just write the
2462result to the file pointed at by `.git/HEAD`, so that we can always see
2463what the last committed state was.
2464
2465Here is an ASCII art by Jon Loeliger that illustrates how
2466various pieces fit together.
2467
2468------------
2469
2470 commit-tree
2471 commit obj
2472 +----+
2473 | |
2474 | |
2475 V V
2476 +-----------+
2477 | Object DB |
2478 | Backing |
2479 | Store |
2480 +-----------+
2481 ^
2482 write-tree | |
2483 tree obj | |
2484 | | read-tree
2485 | | tree obj
2486 V
2487 +-----------+
2488 | Index |
2489 | "cache" |
2490 +-----------+
2491 update-index ^
2492 blob obj | |
2493 | |
2494 checkout-index -u | | checkout-index
2495 stat | | blob obj
2496 V
2497 +-----------+
2498 | Working |
2499 | Directory |
2500 +-----------+
2501
2502------------
2503
2504
2505Examining the data
2506------------------
2507
2508You can examine the data represented in the object database and the
2509index with various helper tools. For every object, you can use
2510gitlink:git-cat-file[1] to examine details about the
2511object:
2512
2513-------------------------------------------------
2514$ git-cat-file -t <objectname>
2515-------------------------------------------------
2516
2517shows the type of the object, and once you have the type (which is
2518usually implicit in where you find the object), you can use
2519
2520-------------------------------------------------
2521$ git-cat-file blob|tree|commit|tag <objectname>
2522-------------------------------------------------
2523
2524to show its contents. NOTE! Trees have binary content, and as a result
2525there is a special helper for showing that content, called
2526`git-ls-tree`, which turns the binary content into a more easily
2527readable form.
2528
2529It's especially instructive to look at "commit" objects, since those
2530tend to be small and fairly self-explanatory. In particular, if you
2531follow the convention of having the top commit name in `.git/HEAD`,
2532you can do
2533
2534-------------------------------------------------
2535$ git-cat-file commit HEAD
2536-------------------------------------------------
2537
2538to see what the top commit was.
2539
2540Merging multiple trees
d19fbc3c
BF
2541----------------------
2542
b181d57f
BF
2543Git helps you do a three-way merge, which you can expand to n-way by
2544repeating the merge procedure arbitrary times until you finally
2545"commit" the state. The normal situation is that you'd only do one
2546three-way merge (two parents), and commit it, but if you like to, you
2547can do multiple parents in one go.
2548
2549To do a three-way merge, you need the two sets of "commit" objects
2550that you want to merge, use those to find the closest common parent (a
2551third "commit" object), and then use those commit objects to find the
2552state of the directory ("tree" object) at these points.
2553
2554To get the "base" for the merge, you first look up the common parent
2555of two commits with
2556
2557-------------------------------------------------
2558$ git-merge-base <commit1> <commit2>
2559-------------------------------------------------
2560
2561which will return you the commit they are both based on. You should
2562now look up the "tree" objects of those commits, which you can easily
2563do with (for example)
2564
2565-------------------------------------------------
2566$ git-cat-file commit <commitname> | head -1
2567-------------------------------------------------
2568
2569since the tree object information is always the first line in a commit
2570object.
2571
2572Once you know the three trees you are going to merge (the one
2573"original" tree, aka the common case, and the two "result" trees, aka
2574the branches you want to merge), you do a "merge" read into the
2575index. This will complain if it has to throw away your old index contents, so you should
2576make sure that you've committed those - in fact you would normally
2577always do a merge against your last commit (which should thus match
2578what you have in your current index anyway).
2579
2580To do the merge, do
2581
2582-------------------------------------------------
2583$ git-read-tree -m -u <origtree> <yourtree> <targettree>
2584-------------------------------------------------
2585
2586which will do all trivial merge operations for you directly in the
2587index file, and you can just write the result out with
2588`git-write-tree`.
2589
2590
2591Merging multiple trees, continued
2592---------------------------------
2593
2594Sadly, many merges aren't trivial. If there are files that have
2595been added.moved or removed, or if both branches have modified the
2596same file, you will be left with an index tree that contains "merge
2597entries" in it. Such an index tree can 'NOT' be written out to a tree
2598object, and you will have to resolve any such merge clashes using
2599other tools before you can write out the result.
2600
2601You can examine such index state with `git-ls-files --unmerged`
2602command. An example:
2603
2604------------------------------------------------
2605$ git-read-tree -m $orig HEAD $target
2606$ git-ls-files --unmerged
2607100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
2608100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
2609100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
2610------------------------------------------------
2611
2612Each line of the `git-ls-files --unmerged` output begins with
2613the blob mode bits, blob SHA1, 'stage number', and the
2614filename. The 'stage number' is git's way to say which tree it
2615came from: stage 1 corresponds to `$orig` tree, stage 2 `HEAD`
2616tree, and stage3 `$target` tree.
2617
2618Earlier we said that trivial merges are done inside
2619`git-read-tree -m`. For example, if the file did not change
2620from `$orig` to `HEAD` nor `$target`, or if the file changed
2621from `$orig` to `HEAD` and `$orig` to `$target` the same way,
2622obviously the final outcome is what is in `HEAD`. What the
2623above example shows is that file `hello.c` was changed from
2624`$orig` to `HEAD` and `$orig` to `$target` in a different way.
2625You could resolve this by running your favorite 3-way merge
2626program, e.g. `diff3` or `merge`, on the blob objects from
2627these three stages yourself, like this:
2628
2629------------------------------------------------
2630$ git-cat-file blob 263414f... >hello.c~1
2631$ git-cat-file blob 06fa6a2... >hello.c~2
2632$ git-cat-file blob cc44c73... >hello.c~3
2633$ merge hello.c~2 hello.c~1 hello.c~3
2634------------------------------------------------
2635
2636This would leave the merge result in `hello.c~2` file, along
2637with conflict markers if there are conflicts. After verifying
2638the merge result makes sense, you can tell git what the final
2639merge result for this file is by:
2640
2641-------------------------------------------------
2642$ mv -f hello.c~2 hello.c
2643$ git-update-index hello.c
2644-------------------------------------------------
2645
2646When a path is in unmerged state, running `git-update-index` for
2647that path tells git to mark the path resolved.
2648
2649The above is the description of a git merge at the lowest level,
2650to help you understand what conceptually happens under the hood.
2651In practice, nobody, not even git itself, uses three `git-cat-file`
2652for this. There is `git-merge-index` program that extracts the
2653stages to temporary files and calls a "merge" script on it:
2654
2655-------------------------------------------------
2656$ git-merge-index git-merge-one-file hello.c
2657-------------------------------------------------
2658
2659and that is what higher level `git resolve` is implemented with.
2660
2661How git stores objects efficiently: pack files
2662----------------------------------------------
2663
2664We've seen how git stores each object in a file named after the
2665object's SHA1 hash.
2666
2667Unfortunately this system becomes inefficient once a project has a
2668lot of objects. Try this on an old project:
2669
2670------------------------------------------------
2671$ git count-objects
26726930 objects, 47620 kilobytes
2673------------------------------------------------
2674
2675The first number is the number of objects which are kept in
2676individual files. The second is the amount of space taken up by
2677those "loose" objects.
2678
2679You can save space and make git faster by moving these loose objects in
2680to a "pack file", which stores a group of objects in an efficient
2681compressed format; the details of how pack files are formatted can be
2682found in link:technical/pack-format.txt[technical/pack-format.txt].
2683
2684To put the loose objects into a pack, just run git repack:
2685
2686------------------------------------------------
2687$ git repack
2688Generating pack...
2689Done counting 6020 objects.
2690Deltifying 6020 objects.
2691 100% (6020/6020) done
2692Writing 6020 objects.
2693 100% (6020/6020) done
2694Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
2695Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
2696------------------------------------------------
2697
2698You can then run
2699
2700------------------------------------------------
2701$ git prune
2702------------------------------------------------
2703
2704to remove any of the "loose" objects that are now contained in the
2705pack. This will also remove any unreferenced objects (which may be
2706created when, for example, you use "git reset" to remove a commit).
2707You can verify that the loose objects are gone by looking at the
2708.git/objects directory or by running
2709
2710------------------------------------------------
2711$ git count-objects
27120 objects, 0 kilobytes
2713------------------------------------------------
2714
2715Although the object files are gone, any commands that refer to those
2716objects will work exactly as they did before.
2717
2718The gitlink:git-gc[1] command performs packing, pruning, and more for
2719you, so is normally the only high-level command you need.
d19fbc3c 2720
21dcb3b7
BF
2721[[dangling-objects]]
2722Dangling objects
2723^^^^^^^^^^^^^^^^
2724
2725The gitlink:git-fsck-objects[1] command will sometimes complain about dangling
2726objects. They are not a problem.
2727
2728The most common cause of dangling objects is that you've rebased a branch, or
2729you have pulled from somebody else who rebased a branch--see
2730<<cleaning-up-history>>. In that case, the old head of the original branch
2731still exists, as does obviously everything it pointed to. The branch pointer
2732itself just doesn't, since you replaced it with another one.
2733
2734There are also other situations too that cause dangling objects. For example, a
2735"dangling blob" may arise because you did a "git add" of a file, but then,
2736before you actually committed it and made it part of the bigger picture, you
2737changed something else in that file and committed that *updated* thing - the
2738old state that you added originally ends up not being pointed to by any
2739commit or tree, so it's now a dangling blob object.
2740
2741Similarly, when the "recursive" merge strategy runs, and finds that there
2742are criss-cross merges and thus more than one merge base (which is fairly
2743unusual, but it does happen), it will generate one temporary midway tree
2744(or possibly even more, if you had lots of criss-crossing merges and
2745more than two merge bases) as a temporary internal merge base, and again,
2746those are real objects, but the end result will not end up pointing to
2747them, so they end up "dangling" in your repository.
2748
2749Generally, dangling objects aren't anything to worry about. They can even
2750be very useful: if you screw something up, the dangling objects can be how
2751you recover your old tree (say, you did a rebase, and realized that you
2752really didn't want to - you can look at what dangling objects you have,
2753and decide to reset your head to some old dangling state).
2754
2755For commits, the most useful thing to do with dangling objects tends to be
2756to do a simple
2757
2758------------------------------------------------
2759$ gitk <dangling-commit-sha-goes-here> --not --all
2760------------------------------------------------
2761
2762which means exactly what it sounds like: it says that you want to see the
2763commit history that is described by the dangling commit(s), but you do NOT
2764want to see the history that is described by all your branches and tags
2765(which are the things you normally reach). That basically shows you in a
2766nice way what the dangling commit was (and notice that it might not be
2767just one commit: we only report the "tip of the line" as being dangling,
2768but there might be a whole deep and complex commit history that has gotten
2769dropped - rebasing will do that).
2770
2771For blobs and trees, you can't do the same, but you can examine them. You
2772can just do
2773
2774------------------------------------------------
2775$ git show <dangling-blob/tree-sha-goes-here>
2776------------------------------------------------
2777
2778to show what the contents of the blob were (or, for a tree, basically what
2779the "ls" for that directory was), and that may give you some idea of what
2780the operation was that left that dangling object.
2781
2782Usually, dangling blobs and trees aren't very interesting. They're almost
2783always the result of either being a half-way mergebase (the blob will
2784often even have the conflict markers from a merge in it, if you have had
2785conflicting merges that you fixed up by hand), or simply because you
2786interrupted a "git fetch" with ^C or something like that, leaving _some_
2787of the new objects in the object database, but just dangling and useless.
2788
2789Anyway, once you are sure that you're not interested in any dangling
2790state, you can just prune all unreachable objects:
2791
2792------------------------------------------------
2793$ git prune
2794------------------------------------------------
2795
2796and they'll be gone. But you should only run "git prune" on a quiescent
2797repository - it's kind of like doing a filesystem fsck recovery: you don't
2798want to do that while the filesystem is mounted.
2799
2800(The same is true of "git-fsck-objects" itself, btw - but since
2801git-fsck-objects never actually *changes* the repository, it just reports
2802on what it found, git-fsck-objects itself is never "dangerous" to run.
2803Running it while somebody is actually changing the repository can cause
2804confusing and scary messages, but it won't actually do anything bad. In
2805contrast, running "git prune" while somebody is actively changing the
2806repository is a *BAD* idea).
2807
d19fbc3c
BF
2808Glossary of git terms
2809=====================
2810
2811include::glossary.txt[]
2812
6bd9b682
BF
2813Notes and todo list for this manual
2814===================================
2815
2816This is a work in progress.
2817
2818The basic requirements:
2f99710c
BF
2819 - It must be readable in order, from beginning to end, by
2820 someone intelligent with a basic grasp of the unix
2821 commandline, but without any special knowledge of git. If
2822 necessary, any other prerequisites should be specifically
2823 mentioned as they arise.
2824 - Whenever possible, section headings should clearly describe
2825 the task they explain how to do, in language that requires
2826 no more knowledge than necessary: for example, "importing
2827 patches into a project" rather than "the git-am command"
6bd9b682 2828
d5cd5de4
BF
2829Think about how to create a clear chapter dependency graph that will
2830allow people to get to important topics without necessarily reading
2831everything in between.
d19fbc3c
BF
2832
2833Scan Documentation/ for other stuff left out; in particular:
2834 howto's
2835 README
2836 some of technical/?
2837 hooks
2838 etc.
2839
2840Scan email archives for other stuff left out
2841
2842Scan man pages to see if any assume more background than this manual
2843provides.
2844
2f99710c 2845Simplify beginning by suggesting disconnected head instead of
b181d57f 2846temporary branch creation?
d19fbc3c
BF
2847
2848Explain how to refer to file stages in the "how to resolve a merge"
e9c0390a 2849section: diff -1, -2, -3, --ours, --theirs :1:/path notation. The
2f99710c 2850"git ls-files --unmerged --stage" thing is sorta useful too,
b181d57f 2851actually. And note gitk --merge.
e9c0390a 2852
2f99710c
BF
2853Add more good examples. Entire sections of just cookbook examples
2854might be a good idea; maybe make an "advanced examples" section a
2855standard end-of-chapter section?
d19fbc3c
BF
2856
2857Include cross-references to the glossary, where appropriate.
2858
e9c0390a
BF
2859To document:
2860 reflogs, git reflog expire
2861 shallow clones?? See draft 1.5.0 release notes for some documentation.