Documentation: clarify track/no-track option.
[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
79c96c57 5command-line skills, but no previous knowledge of git.
d19fbc3c 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 ...
04483524 87$ git checkout -b masterwork origin/master
ef89f701
BF
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]
04483524
JN
158 name = Your Name Comes Here
159 email = you@yourdomain.example.com
ef89f701
BF
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
04483524 198$ git am mbox # import patches from the mailbox "mbox"
ef89f701
BF
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-----------------------------------------------
04e50e94 241$ git fsck
b181d57f
BF
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
81b6c950
BF
291A single git repository may contain multiple branches. It keeps track
292of them by keeping a list of <<def_head,heads>> which reference the
293latest version on each branch; the gitlink:git-branch[1] command shows
294you the list of branch heads:
d19fbc3c
BF
295
296------------------------------------------------
297$ git branch
298* master
299------------------------------------------------
300
81b6c950
BF
301A freshly cloned repository contains a single branch head, named
302"master", and working directory is initialized to the state of
303the project referred to by "master".
d19fbc3c 304
81b6c950
BF
305Most projects also use <<def_tag,tags>>. Tags, like heads, are
306references into the project's history, and can be listed using the
d19fbc3c
BF
307gitlink:git-tag[1] command:
308
309------------------------------------------------
310$ git tag -l
311v2.6.11
312v2.6.11-tree
313v2.6.12
314v2.6.12-rc2
315v2.6.12-rc3
316v2.6.12-rc4
317v2.6.12-rc5
318v2.6.12-rc6
319v2.6.13
320...
321------------------------------------------------
322
fe4b3e59 323Tags are expected to always point at the same version of a project,
81b6c950 324while heads are expected to advance as development progresses.
fe4b3e59 325
81b6c950 326Create a new branch head pointing to one of these versions and check it
d19fbc3c
BF
327out using gitlink:git-checkout[1]:
328
329------------------------------------------------
330$ git checkout -b new v2.6.13
331------------------------------------------------
332
333The working directory then reflects the contents that the project had
334when it was tagged v2.6.13, and gitlink:git-branch[1] shows two
335branches, with an asterisk marking the currently checked-out branch:
336
337------------------------------------------------
338$ git branch
339 master
340* new
341------------------------------------------------
342
343If you decide that you'd rather see version 2.6.17, you can modify
344the current branch to point at v2.6.17 instead, with
345
346------------------------------------------------
347$ git reset --hard v2.6.17
348------------------------------------------------
349
81b6c950 350Note that if the current branch head was your only reference to a
d19fbc3c 351particular point in history, then resetting that branch may leave you
81b6c950
BF
352with no way to find the history it used to point to; so use this command
353carefully.
d19fbc3c
BF
354
355Understanding History: Commits
356------------------------------
357
358Every change in the history of a project is represented by a commit.
359The gitlink:git-show[1] command shows the most recent commit on the
360current branch:
361
362------------------------------------------------
363$ git show
364commit 2b5f6dcce5bf94b9b119e9ed8d537098ec61c3d2
365Author: Jamal Hadi Salim <hadi@cyberus.ca>
366Date: Sat Dec 2 22:22:25 2006 -0800
367
368 [XFRM]: Fix aevent structuring to be more complete.
369
370 aevents can not uniquely identify an SA. We break the ABI with this
371 patch, but consensus is that since it is not yet utilized by any
372 (known) application then it is fine (better do it now than later).
373
374 Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
375 Signed-off-by: David S. Miller <davem@davemloft.net>
376
377diff --git a/Documentation/networking/xfrm_sync.txt b/Documentation/networking/xfrm_sync.txt
378index 8be626f..d7aac9d 100644
379--- a/Documentation/networking/xfrm_sync.txt
380+++ b/Documentation/networking/xfrm_sync.txt
381@@ -47,10 +47,13 @@ aevent_id structure looks like:
382
383 struct xfrm_aevent_id {
384 struct xfrm_usersa_id sa_id;
385+ xfrm_address_t saddr;
386 __u32 flags;
387+ __u32 reqid;
388 };
389...
390------------------------------------------------
391
392As you can see, a commit shows who made the latest change, what they
393did, and why.
394
35121930
BF
395Every commit has a 40-hexdigit id, sometimes called the "object name" or the
396"SHA1 id", shown on the first line of the "git show" output. You can usually
397refer to a commit by a shorter name, such as a tag or a branch name, but this
398longer name can also be useful. Most importantly, it is a globally unique
399name for this commit: so if you tell somebody else the object name (for
400example in email), then you are guaranteed that name will refer to the same
401commit in their repository that it does in yours (assuming their repository
402has that commit at all). Since the object name is computed as a hash over the
403contents of the commit, you are guaranteed that the commit can never change
404without its name also changing.
405
406In fact, in <<git-internals>> we shall see that everything stored in git
407history, including file data and directory contents, is stored in an object
408with a name that is a hash of its contents.
d19fbc3c
BF
409
410Understanding history: commits, parents, and reachability
411~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
412
413Every commit (except the very first commit in a project) also has a
414parent commit which shows what happened before this commit.
415Following the chain of parents will eventually take you back to the
416beginning of the project.
417
418However, the commits do not form a simple list; git allows lines of
419development to diverge and then reconverge, and the point where two
420lines of development reconverge is called a "merge". The commit
421representing a merge can therefore have more than one parent, with
422each parent representing the most recent commit on one of the lines
423of development leading to that point.
424
425The best way to see how this works is using the gitlink:gitk[1]
426command; running gitk now on a git repository and looking for merge
427commits will help understand how the git organizes history.
428
429In the following, we say that commit X is "reachable" from commit Y
430if commit X is an ancestor of commit Y. Equivalently, you could say
431that Y is a descendent of X, or that there is a chain of parents
432leading from commit Y to commit X.
433
3dff5379
PR
434Understanding history: History diagrams
435~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
d19fbc3c
BF
436
437We will sometimes represent git history using diagrams like the one
438below. Commits are shown as "o", and the links between them with
439lines drawn with - / and \. Time goes left to right:
440
1dc71a91
BF
441
442................................................
d19fbc3c
BF
443 o--o--o <-- Branch A
444 /
445 o--o--o <-- master
446 \
447 o--o--o <-- Branch B
1dc71a91 448................................................
d19fbc3c
BF
449
450If we need to talk about a particular commit, the character "o" may
451be replaced with another letter or number.
452
453Understanding history: What is a branch?
454~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
455
81b6c950
BF
456When we need to be precise, we will use the word "branch" to mean a line
457of development, and "branch head" (or just "head") to mean a reference
458to the most recent commit on a branch. In the example above, the branch
459head named "A" is a pointer to one particular commit, but we refer to
460the line of three commits leading up to that point as all being part of
d19fbc3c
BF
461"branch A".
462
81b6c950
BF
463However, when no confusion will result, we often just use the term
464"branch" both for branches and for branch heads.
d19fbc3c
BF
465
466Manipulating branches
467---------------------
468
469Creating, deleting, and modifying branches is quick and easy; here's
470a summary of the commands:
471
472git branch::
473 list all branches
474git branch <branch>::
475 create a new branch named <branch>, referencing the same
476 point in history as the current branch
477git branch <branch> <start-point>::
478 create a new branch named <branch>, referencing
479 <start-point>, which may be specified any way you like,
480 including using a branch name or a tag name
481git branch -d <branch>::
482 delete the branch <branch>; if the branch you are deleting
483 points to a commit which is not reachable from this branch,
484 this command will fail with a warning.
485git branch -D <branch>::
486 even if the branch points to a commit not reachable
487 from the current branch, you may know that that commit
488 is still reachable from some other branch or tag. In that
489 case it is safe to use this command to force git to delete
490 the branch.
491git checkout <branch>::
492 make the current branch <branch>, updating the working
493 directory to reflect the version referenced by <branch>
494git checkout -b <new> <start-point>::
495 create a new branch <new> referencing <start-point>, and
496 check it out.
497
498It is also useful to know that the special symbol "HEAD" can always
499be used to refer to the current branch.
500
501Examining branches from a remote repository
502-------------------------------------------
503
504The "master" branch that was created at the time you cloned is a copy
505of the HEAD in the repository that you cloned from. That repository
506may also have had other branches, though, and your local repository
507keeps branches which track each of those remote branches, which you
508can view using the "-r" option to gitlink:git-branch[1]:
509
510------------------------------------------------
511$ git branch -r
512 origin/HEAD
513 origin/html
514 origin/maint
515 origin/man
516 origin/master
517 origin/next
518 origin/pu
519 origin/todo
520------------------------------------------------
521
522You cannot check out these remote-tracking branches, but you can
523examine them on a branch of your own, just as you would a tag:
524
525------------------------------------------------
526$ git checkout -b my-todo-copy origin/todo
527------------------------------------------------
528
529Note that the name "origin" is just the name that git uses by default
530to refer to the repository that you cloned from.
531
532[[how-git-stores-references]]
f60b9642
BF
533Naming branches, tags, and other references
534-------------------------------------------
d19fbc3c
BF
535
536Branches, remote-tracking branches, and tags are all references to
f60b9642
BF
537commits. All references are named with a slash-separated path name
538starting with "refs"; the names we've been using so far are actually
539shorthand:
d19fbc3c 540
f60b9642
BF
541 - The branch "test" is short for "refs/heads/test".
542 - The tag "v2.6.18" is short for "refs/tags/v2.6.18".
543 - "origin/master" is short for "refs/remotes/origin/master".
d19fbc3c 544
f60b9642
BF
545The full name is occasionally useful if, for example, there ever
546exists a tag and a branch with the same name.
d19fbc3c 547
f60b9642
BF
548As another useful shortcut, if the repository "origin" posesses only
549a single branch, you can refer to that branch as just "origin".
d19fbc3c 550
f60b9642
BF
551More generally, if you have defined a remote repository named
552"example", you can refer to the branch in that repository as
553"example". And for a repository with multiple branches, this will
554refer to the branch designated as the "HEAD" branch.
d19fbc3c
BF
555
556For the complete list of paths which git checks for references, and
f60b9642
BF
557the order it uses to decide which to choose when there are multiple
558references with the same shorthand name, see the "SPECIFYING
559REVISIONS" section of gitlink:git-rev-parse[1].
d19fbc3c
BF
560
561[[Updating-a-repository-with-git-fetch]]
562Updating a repository with git fetch
563------------------------------------
564
565Eventually the developer cloned from will do additional work in her
566repository, creating new commits and advancing the branches to point
567at the new commits.
568
569The command "git fetch", with no arguments, will update all of the
570remote-tracking branches to the latest version found in her
571repository. It will not touch any of your own branches--not even the
572"master" branch that was created for you on clone.
573
d5cd5de4
BF
574Fetching branches from other repositories
575-----------------------------------------
576
577You can also track branches from repositories other than the one you
578cloned from, using gitlink:git-remote[1]:
579
580-------------------------------------------------
581$ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
04483524 582$ git fetch linux-nfs
d5cd5de4
BF
583* refs/remotes/linux-nfs/master: storing branch 'master' ...
584 commit: bf81b46
585-------------------------------------------------
586
587New remote-tracking branches will be stored under the shorthand name
588that you gave "git remote add", in this case linux-nfs:
589
590-------------------------------------------------
591$ git branch -r
592linux-nfs/master
593origin/master
594-------------------------------------------------
595
596If you run "git fetch <remote>" later, the tracking branches for the
597named <remote> will be updated.
598
599If you examine the file .git/config, you will see that git has added
600a new stanza:
601
602-------------------------------------------------
603$ cat .git/config
604...
605[remote "linux-nfs"]
923642fe
BF
606 url = git://linux-nfs.org/pub/nfs-2.6.git
607 fetch = +refs/heads/*:refs/remotes/linux-nfs/*
d5cd5de4
BF
608...
609-------------------------------------------------
610
fc90c536
BF
611This is what causes git to track the remote's branches; you may modify
612or delete these configuration options by editing .git/config with a
613text editor. (See the "CONFIGURATION FILE" section of
614gitlink:git-config[1] for details.)
d5cd5de4 615
d19fbc3c
BF
616Exploring git history
617=====================
618
619Git is best thought of as a tool for storing the history of a
620collection of files. It does this by storing compressed snapshots of
621the contents of a file heirarchy, together with "commits" which show
622the relationships between these snapshots.
623
624Git provides extremely flexible and fast tools for exploring the
625history of a project.
626
aacd404e 627We start with one specialized tool that is useful for finding the
d19fbc3c
BF
628commit that introduced a bug into a project.
629
630How to use bisect to find a regression
631--------------------------------------
632
633Suppose version 2.6.18 of your project worked, but the version at
634"master" crashes. Sometimes the best way to find the cause of such a
635regression is to perform a brute-force search through the project's
636history to find the particular commit that caused the problem. The
637gitlink:git-bisect[1] command can help you do this:
638
639-------------------------------------------------
640$ git bisect start
641$ git bisect good v2.6.18
642$ git bisect bad master
643Bisecting: 3537 revisions left to test after this
644[65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
645-------------------------------------------------
646
647If you run "git branch" at this point, you'll see that git has
648temporarily moved you to a new branch named "bisect". This branch
649points to a commit (with commit id 65934...) that is reachable from
650v2.6.19 but not from v2.6.18. Compile and test it, and see whether
651it crashes. Assume it does crash. Then:
652
653-------------------------------------------------
654$ git bisect bad
655Bisecting: 1769 revisions left to test after this
656[7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
657-------------------------------------------------
658
659checks out an older version. Continue like this, telling git at each
660stage whether the version it gives you is good or bad, and notice
661that the number of revisions left to test is cut approximately in
662half each time.
663
664After about 13 tests (in this case), it will output the commit id of
665the guilty commit. You can then examine the commit with
666gitlink:git-show[1], find out who wrote it, and mail them your bug
667report with the commit id. Finally, run
668
669-------------------------------------------------
670$ git bisect reset
671-------------------------------------------------
672
673to return you to the branch you were on before and delete the
674temporary "bisect" branch.
675
676Note that the version which git-bisect checks out for you at each
677point is just a suggestion, and you're free to try a different
678version if you think it would be a good idea. For example,
679occasionally you may land on a commit that broke something unrelated;
680run
681
682-------------------------------------------------
04483524 683$ git bisect visualize
d19fbc3c
BF
684-------------------------------------------------
685
686which will run gitk and label the commit it chose with a marker that
687says "bisect". Chose a safe-looking commit nearby, note its commit
688id, and check it out with:
689
690-------------------------------------------------
691$ git reset --hard fb47ddb2db...
692-------------------------------------------------
693
694then test, run "bisect good" or "bisect bad" as appropriate, and
695continue.
696
697Naming commits
698--------------
699
700We have seen several ways of naming commits already:
701
d55ae921 702 - 40-hexdigit object name
d19fbc3c
BF
703 - branch name: refers to the commit at the head of the given
704 branch
705 - tag name: refers to the commit pointed to by the given tag
706 (we've seen branches and tags are special cases of
707 <<how-git-stores-references,references>>).
708 - HEAD: refers to the head of the current branch
709
eb6ae7f4 710There are many more; see the "SPECIFYING REVISIONS" section of the
aec053bb 711gitlink:git-rev-parse[1] man page for the complete list of ways to
d19fbc3c
BF
712name revisions. Some examples:
713
714-------------------------------------------------
d55ae921 715$ git show fb47ddb2 # the first few characters of the object name
d19fbc3c
BF
716 # are usually enough to specify it uniquely
717$ git show HEAD^ # the parent of the HEAD commit
718$ git show HEAD^^ # the grandparent
719$ git show HEAD~4 # the great-great-grandparent
720-------------------------------------------------
721
722Recall that merge commits may have more than one parent; by default,
723^ and ~ follow the first parent listed in the commit, but you can
724also choose:
725
726-------------------------------------------------
727$ git show HEAD^1 # show the first parent of HEAD
728$ git show HEAD^2 # show the second parent of HEAD
729-------------------------------------------------
730
731In addition to HEAD, there are several other special names for
732commits:
733
734Merges (to be discussed later), as well as operations such as
735git-reset, which change the currently checked-out commit, generally
736set ORIG_HEAD to the value HEAD had before the current operation.
737
738The git-fetch operation always stores the head of the last fetched
739branch in FETCH_HEAD. For example, if you run git fetch without
740specifying a local branch as the target of the operation
741
742-------------------------------------------------
743$ git fetch git://example.com/proj.git theirbranch
744-------------------------------------------------
745
746the fetched commits will still be available from FETCH_HEAD.
747
748When we discuss merges we'll also see the special name MERGE_HEAD,
749which refers to the other branch that we're merging in to the current
750branch.
751
aec053bb 752The gitlink:git-rev-parse[1] command is a low-level command that is
d55ae921
BF
753occasionally useful for translating some name for a commit to the object
754name for that commit:
aec053bb
BF
755
756-------------------------------------------------
757$ git rev-parse origin
758e05db0fd4f31dde7005f075a84f96b360d05984b
759-------------------------------------------------
760
d19fbc3c
BF
761Creating tags
762-------------
763
764We can also create a tag to refer to a particular commit; after
765running
766
767-------------------------------------------------
04483524 768$ git tag stable-1 1b2e1d63ff
d19fbc3c
BF
769-------------------------------------------------
770
771You can use stable-1 to refer to the commit 1b2e1d63ff.
772
773This creates a "lightweight" tag. If the tag is a tag you wish to
774share with others, and possibly sign cryptographically, then you
775should create a tag object instead; see the gitlink:git-tag[1] man
776page for details.
777
778Browsing revisions
779------------------
780
781The gitlink:git-log[1] command can show lists of commits. On its
782own, it shows all commits reachable from the parent commit; but you
783can also make more specific requests:
784
785-------------------------------------------------
786$ git log v2.5.. # commits since (not reachable from) v2.5
787$ git log test..master # commits reachable from master but not test
788$ git log master..test # ...reachable from test but not master
789$ git log master...test # ...reachable from either test or master,
790 # but not both
791$ git log --since="2 weeks ago" # commits from the last 2 weeks
792$ git log Makefile # commits which modify Makefile
793$ git log fs/ # ... which modify any file under fs/
794$ git log -S'foo()' # commits which add or remove any file data
795 # matching the string 'foo()'
796-------------------------------------------------
797
798And of course you can combine all of these; the following finds
799commits since v2.5 which touch the Makefile or any file under fs:
800
801-------------------------------------------------
802$ git log v2.5.. Makefile fs/
803-------------------------------------------------
804
805You can also ask git log to show patches:
806
807-------------------------------------------------
808$ git log -p
809-------------------------------------------------
810
811See the "--pretty" option in the gitlink:git-log[1] man page for more
812display options.
813
814Note that git log starts with the most recent commit and works
815backwards through the parents; however, since git history can contain
3dff5379 816multiple independent lines of development, the particular order that
d19fbc3c
BF
817commits are listed in may be somewhat arbitrary.
818
819Generating diffs
820----------------
821
822You can generate diffs between any two versions using
823gitlink:git-diff[1]:
824
825-------------------------------------------------
826$ git diff master..test
827-------------------------------------------------
828
829Sometimes what you want instead is a set of patches:
830
831-------------------------------------------------
832$ git format-patch master..test
833-------------------------------------------------
834
835will generate a file with a patch for each commit reachable from test
836but not from master. Note that if master also has commits which are
837not reachable from test, then the combined result of these patches
838will not be the same as the diff produced by the git-diff example.
839
840Viewing old file versions
841-------------------------
842
843You can always view an old version of a file by just checking out the
844correct revision first. But sometimes it is more convenient to be
845able to view an old version of a single file without checking
846anything out; this command does that:
847
848-------------------------------------------------
849$ git show v2.5:fs/locks.c
850-------------------------------------------------
851
852Before the colon may be anything that names a commit, and after it
853may be any path to a file tracked by git.
854
aec053bb
BF
855Examples
856--------
857
858Check whether two branches point at the same history
2f99710c 859~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
aec053bb
BF
860
861Suppose you want to check whether two branches point at the same point
862in history.
863
864-------------------------------------------------
865$ git diff origin..master
866-------------------------------------------------
867
69f7ad73
BF
868will tell you whether the contents of the project are the same at the
869two branches; in theory, however, it's possible that the same project
870contents could have been arrived at by two different historical
d55ae921 871routes. You could compare the object names:
aec053bb
BF
872
873-------------------------------------------------
874$ git rev-list origin
875e05db0fd4f31dde7005f075a84f96b360d05984b
876$ git rev-list master
877e05db0fd4f31dde7005f075a84f96b360d05984b
878-------------------------------------------------
879
69f7ad73
BF
880Or you could recall that the ... operator selects all commits
881contained reachable from either one reference or the other but not
882both: so
aec053bb
BF
883
884-------------------------------------------------
885$ git log origin...master
886-------------------------------------------------
887
888will return no commits when the two branches are equal.
889
b181d57f
BF
890Find first tagged version including a given fix
891~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
aec053bb 892
69f7ad73
BF
893Suppose you know that the commit e05db0fd fixed a certain problem.
894You'd like to find the earliest tagged release that contains that
895fix.
896
897Of course, there may be more than one answer--if the history branched
898after commit e05db0fd, then there could be multiple "earliest" tagged
899releases.
900
901You could just visually inspect the commits since e05db0fd:
902
903-------------------------------------------------
904$ gitk e05db0fd..
905-------------------------------------------------
906
b181d57f
BF
907Or you can use gitlink:git-name-rev[1], which will give the commit a
908name based on any tag it finds pointing to one of the commit's
909descendants:
910
911-------------------------------------------------
04483524 912$ git name-rev --tags e05db0fd
b181d57f
BF
913e05db0fd tags/v1.5.0-rc1^0~23
914-------------------------------------------------
915
916The gitlink:git-describe[1] command does the opposite, naming the
917revision using a tag on which the given commit is based:
918
919-------------------------------------------------
920$ git describe e05db0fd
04483524 921v1.5.0-rc0-260-ge05db0f
b181d57f
BF
922-------------------------------------------------
923
924but that may sometimes help you guess which tags might come after the
925given commit.
926
927If you just want to verify whether a given tagged version contains a
928given commit, you could use gitlink:git-merge-base[1]:
929
930-------------------------------------------------
931$ git merge-base e05db0fd v1.5.0-rc1
932e05db0fd4f31dde7005f075a84f96b360d05984b
933-------------------------------------------------
934
935The merge-base command finds a common ancestor of the given commits,
936and always returns one or the other in the case where one is a
937descendant of the other; so the above output shows that e05db0fd
938actually is an ancestor of v1.5.0-rc1.
939
940Alternatively, note that
941
942-------------------------------------------------
4a7979ca 943$ git log v1.5.0-rc1..e05db0fd
b181d57f
BF
944-------------------------------------------------
945
4a7979ca 946will produce empty output if and only if v1.5.0-rc1 includes e05db0fd,
b181d57f 947because it outputs only commits that are not reachable from v1.5.0-rc1.
aec053bb 948
4a7979ca
BF
949As yet another alternative, the gitlink:git-show-branch[1] command lists
950the commits reachable from its arguments with a display on the left-hand
951side that indicates which arguments that commit is reachable from. So,
952you can run something like
953
954-------------------------------------------------
955$ git show-branch e05db0fd v1.5.0-rc0 v1.5.0-rc1 v1.5.0-rc2
956! [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
957available
958 ! [v1.5.0-rc0] GIT v1.5.0 preview
959 ! [v1.5.0-rc1] GIT v1.5.0-rc1
960 ! [v1.5.0-rc2] GIT v1.5.0-rc2
961...
962-------------------------------------------------
963
964then search for a line that looks like
965
966-------------------------------------------------
967+ ++ [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
968available
969-------------------------------------------------
970
971Which shows that e05db0fd is reachable from itself, from v1.5.0-rc1, and
972from v1.5.0-rc2, but not from v1.5.0-rc0.
973
974
d19fbc3c
BF
975Developing with git
976===================
977
978Telling git your name
979---------------------
980
981Before creating any commits, you should introduce yourself to git. The
982easiest way to do so is:
983
984------------------------------------------------
985$ cat >~/.gitconfig <<\EOF
986[user]
987 name = Your Name Comes Here
988 email = you@yourdomain.example.com
989EOF
990------------------------------------------------
991
fc90c536
BF
992(See the "CONFIGURATION FILE" section of gitlink:git-config[1] for
993details on the configuration file.)
994
d19fbc3c
BF
995
996Creating a new repository
997-------------------------
998
999Creating a new repository from scratch is very easy:
1000
1001-------------------------------------------------
1002$ mkdir project
1003$ cd project
f1d2b477 1004$ git init
d19fbc3c
BF
1005-------------------------------------------------
1006
1007If you have some initial content (say, a tarball):
1008
1009-------------------------------------------------
1010$ tar -xzvf project.tar.gz
1011$ cd project
f1d2b477 1012$ git init
d19fbc3c
BF
1013$ git add . # include everything below ./ in the first commit:
1014$ git commit
1015-------------------------------------------------
1016
1017[[how-to-make-a-commit]]
ae25c67a 1018How to make a commit
d19fbc3c
BF
1019--------------------
1020
1021Creating a new commit takes three steps:
1022
1023 1. Making some changes to the working directory using your
1024 favorite editor.
1025 2. Telling git about your changes.
1026 3. Creating the commit using the content you told git about
1027 in step 2.
1028
1029In practice, you can interleave and repeat steps 1 and 2 as many
1030times as you want: in order to keep track of what you want committed
1031at step 3, git maintains a snapshot of the tree's contents in a
1032special staging area called "the index."
1033
01997b4a
BF
1034At the beginning, the content of the index will be identical to
1035that of the HEAD. The command "git diff --cached", which shows
1036the difference between the HEAD and the index, should therefore
1037produce no output at that point.
eb6ae7f4 1038
d19fbc3c
BF
1039Modifying the index is easy:
1040
1041To update the index with the new contents of a modified file, use
1042
1043-------------------------------------------------
1044$ git add path/to/file
1045-------------------------------------------------
1046
1047To add the contents of a new file to the index, use
1048
1049-------------------------------------------------
1050$ git add path/to/file
1051-------------------------------------------------
1052
eb6ae7f4 1053To remove a file from the index and from the working tree,
d19fbc3c
BF
1054
1055-------------------------------------------------
1056$ git rm path/to/file
1057-------------------------------------------------
1058
1059After each step you can verify that
1060
1061-------------------------------------------------
1062$ git diff --cached
1063-------------------------------------------------
1064
1065always shows the difference between the HEAD and the index file--this
1066is what you'd commit if you created the commit now--and that
1067
1068-------------------------------------------------
1069$ git diff
1070-------------------------------------------------
1071
1072shows the difference between the working tree and the index file.
1073
1074Note that "git add" always adds just the current contents of a file
1075to the index; further changes to the same file will be ignored unless
1076you run git-add on the file again.
1077
1078When you're ready, just run
1079
1080-------------------------------------------------
1081$ git commit
1082-------------------------------------------------
1083
1084and git will prompt you for a commit message and then create the new
3dff5379 1085commit. Check to make sure it looks like what you expected with
d19fbc3c
BF
1086
1087-------------------------------------------------
1088$ git show
1089-------------------------------------------------
1090
1091As a special shortcut,
1092
1093-------------------------------------------------
1094$ git commit -a
1095-------------------------------------------------
1096
1097will update the index with any files that you've modified or removed
1098and create a commit, all in one step.
1099
1100A number of commands are useful for keeping track of what you're
1101about to commit:
1102
1103-------------------------------------------------
1104$ git diff --cached # difference between HEAD and the index; what
1105 # would be commited if you ran "commit" now.
1106$ git diff # difference between the index file and your
1107 # working directory; changes that would not
1108 # be included if you ran "commit" now.
1109$ git status # a brief per-file summary of the above.
1110-------------------------------------------------
1111
ae25c67a 1112Creating good commit messages
d19fbc3c
BF
1113-----------------------------
1114
1115Though not required, it's a good idea to begin the commit message
1116with a single short (less than 50 character) line summarizing the
1117change, followed by a blank line and then a more thorough
1118description. Tools that turn commits into email, for example, use
1119the first line on the Subject line and the rest of the commit in the
1120body.
1121
ae25c67a 1122How to merge
d19fbc3c
BF
1123------------
1124
1125You can rejoin two diverging branches of development using
1126gitlink:git-merge[1]:
1127
1128-------------------------------------------------
1129$ git merge branchname
1130-------------------------------------------------
1131
1132merges the development in the branch "branchname" into the current
1133branch. If there are conflicts--for example, if the same file is
1134modified in two different ways in the remote branch and the local
1135branch--then you are warned; the output may look something like this:
1136
1137-------------------------------------------------
fabbd8f6
BF
1138$ git merge next
1139 100% (4/4) done
1140Auto-merged file.txt
d19fbc3c
BF
1141CONFLICT (content): Merge conflict in file.txt
1142Automatic merge failed; fix conflicts and then commit the result.
1143-------------------------------------------------
1144
1145Conflict markers are left in the problematic files, and after
1146you resolve the conflicts manually, you can update the index
1147with the contents and run git commit, as you normally would when
1148creating a new file.
1149
1150If you examine the resulting commit using gitk, you will see that it
1151has two parents, one pointing to the top of the current branch, and
1152one to the top of the other branch.
1153
1154In more detail:
1155
1156[[resolving-a-merge]]
1157Resolving a merge
1158-----------------
1159
1160When a merge isn't resolved automatically, git leaves the index and
1161the working tree in a special state that gives you all the
1162information you need to help resolve the merge.
1163
1164Files with conflicts are marked specially in the index, so until you
ef561ac7
BF
1165resolve the problem and update the index, gitlink:git-commit[1] will
1166fail:
d19fbc3c
BF
1167
1168-------------------------------------------------
1169$ git commit
1170file.txt: needs merge
1171-------------------------------------------------
1172
ef561ac7
BF
1173Also, gitlink:git-status[1] will list those files as "unmerged", and the
1174files with conflicts will have conflict markers added, like this:
1175
1176-------------------------------------------------
1177<<<<<<< HEAD:file.txt
1178Hello world
1179=======
1180Goodbye
1181>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1182-------------------------------------------------
1183
1184All you need to do is edit the files to resolve the conflicts, and then
1185
1186-------------------------------------------------
1187$ git add file.txt
1188$ git commit
1189-------------------------------------------------
1190
1191Note that the commit message will already be filled in for you with
1192some information about the merge. Normally you can just use this
1193default message unchanged, but you may add additional commentary of
1194your own if desired.
1195
1196The above is all you need to know to resolve a simple merge. But git
1197also provides more information to help resolve conflicts:
1198
1199Getting conflict-resolution help during a merge
1200~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
d19fbc3c
BF
1201
1202All of the changes that git was able to merge automatically are
1203already added to the index file, so gitlink:git-diff[1] shows only
ef561ac7 1204the conflicts. It uses an unusual syntax:
d19fbc3c
BF
1205
1206-------------------------------------------------
1207$ git diff
1208diff --cc file.txt
1209index 802992c,2b60207..0000000
1210--- a/file.txt
1211+++ b/file.txt
1212@@@ -1,1 -1,1 +1,5 @@@
1213++<<<<<<< HEAD:file.txt
1214 +Hello world
1215++=======
1216+ Goodbye
1217++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1218-------------------------------------------------
1219
1220Recall that the commit which will be commited after we resolve this
1221conflict will have two parents instead of the usual one: one parent
1222will be HEAD, the tip of the current branch; the other will be the
1223tip of the other branch, which is stored temporarily in MERGE_HEAD.
1224
ef561ac7
BF
1225During the merge, the index holds three versions of each file. Each of
1226these three "file stages" represents a different version of the file:
1227
1228-------------------------------------------------
1229$ git show :1:file.txt # the file in a common ancestor of both branches
1230$ git show :2:file.txt # the version from HEAD, but including any
1231 # nonconflicting changes from MERGE_HEAD
1232$ git show :3:file.txt # the version from MERGE_HEAD, but including any
1233 # nonconflicting changes from HEAD.
1234-------------------------------------------------
1235
1236Since the stage 2 and stage 3 versions have already been updated with
1237nonconflicting changes, the only remaining differences between them are
1238the important ones; thus gitlink:git-diff[1] can use the information in
1239the index to show only those conflicts.
1240
1241The diff above shows the differences between the working-tree version of
1242file.txt and the stage 2 and stage 3 versions. So instead of preceding
1243each line by a single "+" or "-", it now uses two columns: the first
1244column is used for differences between the first parent and the working
1245directory copy, and the second for differences between the second parent
1246and the working directory copy. (See the "COMBINED DIFF FORMAT" section
1247of gitlink:git-diff-files[1] for a details of the format.)
1248
1249After resolving the conflict in the obvious way (but before updating the
1250index), the diff will look like:
d19fbc3c
BF
1251
1252-------------------------------------------------
1253$ git diff
1254diff --cc file.txt
1255index 802992c,2b60207..0000000
1256--- a/file.txt
1257+++ b/file.txt
1258@@@ -1,1 -1,1 +1,1 @@@
1259- Hello world
1260 -Goodbye
1261++Goodbye world
1262-------------------------------------------------
1263
1264This shows that our resolved version deleted "Hello world" from the
1265first parent, deleted "Goodbye" from the second parent, and added
1266"Goodbye world", which was previously absent from both.
1267
ef561ac7
BF
1268Some special diff options allow diffing the working directory against
1269any of these stages:
1270
1271-------------------------------------------------
1272$ git diff -1 file.txt # diff against stage 1
1273$ git diff --base file.txt # same as the above
1274$ git diff -2 file.txt # diff against stage 2
1275$ git diff --ours file.txt # same as the above
1276$ git diff -3 file.txt # diff against stage 3
1277$ git diff --theirs file.txt # same as the above.
1278-------------------------------------------------
1279
1280The gitlink:git-log[1] and gitk[1] commands also provide special help
1281for merges:
d19fbc3c
BF
1282
1283-------------------------------------------------
1284$ git log --merge
ef561ac7 1285$ gitk --merge
d19fbc3c
BF
1286-------------------------------------------------
1287
ef561ac7
BF
1288These will display all commits which exist only on HEAD or on
1289MERGE_HEAD, and which touch an unmerged file.
d19fbc3c 1290
ef561ac7 1291Each time you resolve the conflicts in a file and update the index:
d19fbc3c
BF
1292
1293-------------------------------------------------
1294$ git add file.txt
d19fbc3c
BF
1295-------------------------------------------------
1296
ef561ac7
BF
1297the different stages of that file will be "collapsed", after which
1298git-diff will (by default) no longer show diffs for that file.
d19fbc3c
BF
1299
1300[[undoing-a-merge]]
ae25c67a 1301Undoing a merge
d19fbc3c
BF
1302---------------
1303
1304If you get stuck and decide to just give up and throw the whole mess
1305away, you can always return to the pre-merge state with
1306
1307-------------------------------------------------
1308$ git reset --hard HEAD
1309-------------------------------------------------
1310
1311Or, if you've already commited the merge that you want to throw away,
1312
1313-------------------------------------------------
1c73bb0e 1314$ git reset --hard ORIG_HEAD
d19fbc3c
BF
1315-------------------------------------------------
1316
1317However, this last command can be dangerous in some cases--never
1318throw away a commit you have already committed if that commit may
1319itself have been merged into another branch, as doing so may confuse
1320further merges.
1321
1322Fast-forward merges
1323-------------------
1324
1325There is one special case not mentioned above, which is treated
1326differently. Normally, a merge results in a merge commit, with two
1327parents, one pointing at each of the two lines of development that
1328were merged.
1329
1330However, if one of the two lines of development is completely
1331contained within the other--so every commit present in the one is
1332already contained in the other--then git just performs a
1333<<fast-forwards,fast forward>>; the head of the current branch is
1334moved forward to point at the head of the merged-in branch, without
1335any new commits being created.
1336
b684f830
BF
1337Fixing mistakes
1338---------------
1339
1340If you've messed up the working tree, but haven't yet committed your
1341mistake, you can return the entire working tree to the last committed
1342state with
1343
1344-------------------------------------------------
1345$ git reset --hard HEAD
1346-------------------------------------------------
1347
1348If you make a commit that you later wish you hadn't, there are two
1349fundamentally different ways to fix the problem:
1350
1351 1. You can create a new commit that undoes whatever was done
1352 by the previous commit. This is the correct thing if your
1353 mistake has already been made public.
1354
1355 2. You can go back and modify the old commit. You should
1356 never do this if you have already made the history public;
1357 git does not normally expect the "history" of a project to
1358 change, and cannot correctly perform repeated merges from
1359 a branch that has had its history changed.
1360
1361Fixing a mistake with a new commit
1362~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1363
1364Creating a new commit that reverts an earlier change is very easy;
1365just pass the gitlink:git-revert[1] command a reference to the bad
1366commit; for example, to revert the most recent commit:
1367
1368-------------------------------------------------
1369$ git revert HEAD
1370-------------------------------------------------
1371
1372This will create a new commit which undoes the change in HEAD. You
1373will be given a chance to edit the commit message for the new commit.
1374
1375You can also revert an earlier change, for example, the next-to-last:
1376
1377-------------------------------------------------
1378$ git revert HEAD^
1379-------------------------------------------------
1380
1381In this case git will attempt to undo the old change while leaving
1382intact any changes made since then. If more recent changes overlap
1383with the changes to be reverted, then you will be asked to fix
1384conflicts manually, just as in the case of <<resolving-a-merge,
1385resolving a merge>>.
1386
365aa199 1387[[fixing-a-mistake-by-editing-history]]
b684f830
BF
1388Fixing a mistake by editing history
1389~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1390
1391If the problematic commit is the most recent commit, and you have not
1392yet made that commit public, then you may just
1393<<undoing-a-merge,destroy it using git-reset>>.
1394
1395Alternatively, you
1396can edit the working directory and update the index to fix your
1397mistake, just as if you were going to <<how-to-make-a-commit,create a
1398new commit>>, then run
1399
1400-------------------------------------------------
1401$ git commit --amend
1402-------------------------------------------------
1403
1404which will replace the old commit by a new commit incorporating your
1405changes, giving you a chance to edit the old commit message first.
1406
1407Again, you should never do this to a commit that may already have
1408been merged into another branch; use gitlink:git-revert[1] instead in
1409that case.
1410
1411It is also possible to edit commits further back in the history, but
1412this is an advanced topic to be left for
1413<<cleaning-up-history,another chapter>>.
1414
1415Checking out an old version of a file
1416~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1417
1418In the process of undoing a previous bad change, you may find it
1419useful to check out an older version of a particular file using
1420gitlink:git-checkout[1]. We've used git checkout before to switch
1421branches, but it has quite different behavior if it is given a path
1422name: the command
1423
1424-------------------------------------------------
1425$ git checkout HEAD^ path/to/file
1426-------------------------------------------------
1427
1428replaces path/to/file by the contents it had in the commit HEAD^, and
1429also updates the index to match. It does not change branches.
1430
1431If you just want to look at an old version of the file, without
1432modifying the working directory, you can do that with
1433gitlink:git-show[1]:
1434
1435-------------------------------------------------
ed4eb0d8 1436$ git show HEAD^:path/to/file
b684f830
BF
1437-------------------------------------------------
1438
1439which will display the given version of the file.
1440
d19fbc3c
BF
1441Ensuring good performance
1442-------------------------
1443
1444On large repositories, git depends on compression to keep the history
1445information from taking up to much space on disk or in memory.
1446
1447This compression is not performed automatically. Therefore you
17217090 1448should occasionally run gitlink:git-gc[1]:
d19fbc3c
BF
1449
1450-------------------------------------------------
1451$ git gc
1452-------------------------------------------------
1453
17217090
BF
1454to recompress the archive. This can be very time-consuming, so
1455you may prefer to run git-gc when you are not doing other work.
d19fbc3c 1456
11e016a3
BF
1457Ensuring reliability
1458--------------------
1459
1460Checking the repository for corruption
1461~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1462
1191ee18
BF
1463The gitlink:git-fsck[1] command runs a number of self-consistency checks
1464on the repository, and reports on any problems. This may take some
21dcb3b7
BF
1465time. The most common warning by far is about "dangling" objects:
1466
1467-------------------------------------------------
04e50e94 1468$ git fsck
21dcb3b7
BF
1469dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1470dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1471dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1472dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1473dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1474dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1475dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1476dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1477...
1478-------------------------------------------------
1479
1191ee18
BF
1480Dangling objects are objects that are harmless, but also unnecessary;
1481you can remove them at any time with gitlink:git-prune[1] or the --prune
1482option to gitlink:git-gc[1]:
21dcb3b7
BF
1483
1484-------------------------------------------------
1485$ git gc --prune
1486-------------------------------------------------
1487
1191ee18
BF
1488This may be time-consuming. Unlike most other git operations (including
1489git-gc when run without any options), it is not safe to prune while
1490other git operations are in progress in the same repository.
21dcb3b7 1491
988b27d3 1492For more about dangling objects, see <<dangling-objects>>.
21dcb3b7 1493
11e016a3
BF
1494
1495Recovering lost changes
1496~~~~~~~~~~~~~~~~~~~~~~~
1497
559e4d7a
BF
1498Reflogs
1499^^^^^^^
1500
1501Say you modify a branch with gitlink:git-reset[1] --hard, and then
1502realize that the branch was the only reference you had to that point in
1503history.
1504
1505Fortunately, git also keeps a log, called a "reflog", of all the
1506previous values of each branch. So in this case you can still find the
1507old history using, for example,
1508
1509-------------------------------------------------
1510$ git log master@{1}
1511-------------------------------------------------
1512
1513This lists the commits reachable from the previous version of the head.
1514This syntax can be used to with any git command that accepts a commit,
1515not just with git log. Some other examples:
1516
1517-------------------------------------------------
1518$ git show master@{2} # See where the branch pointed 2,
1519$ git show master@{3} # 3, ... changes ago.
1520$ gitk master@{yesterday} # See where it pointed yesterday,
1521$ gitk master@{"1 week ago"} # ... or last week
1522-------------------------------------------------
1523
1524The reflogs are kept by default for 30 days, after which they may be
036be17e 1525pruned. See gitlink:git-reflog[1] and gitlink:git-gc[1] to learn
559e4d7a
BF
1526how to control this pruning, and see the "SPECIFYING REVISIONS"
1527section of gitlink:git-rev-parse[1] for details.
1528
1529Note that the reflog history is very different from normal git history.
1530While normal history is shared by every repository that works on the
1531same project, the reflog history is not shared: it tells you only about
1532how the branches in your local repository have changed over time.
1533
1534Examining dangling objects
1535^^^^^^^^^^^^^^^^^^^^^^^^^^
1536
1537In some situations the reflog may not be able to save you. For
1538example, suppose you delete a branch, then realize you need the history
79c96c57 1539it contained. The reflog is also deleted; however, if you have not
559e4d7a
BF
1540yet pruned the repository, then you may still be able to find
1541the lost commits; run git-fsck and watch for output that mentions
1542"dangling commits":
1543
1544-------------------------------------------------
1545$ git fsck
1546dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1547dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1548dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1549...
1550-------------------------------------------------
1551
aacd404e 1552You can examine
559e4d7a
BF
1553one of those dangling commits with, for example,
1554
1555------------------------------------------------
1556$ gitk 7281251ddd --not --all
1557------------------------------------------------
1558
1559which does what it sounds like: it says that you want to see the commit
1560history that is described by the dangling commit(s), but not the
1561history that is described by all your existing branches and tags. Thus
1562you get exactly the history reachable from that commit that is lost.
1563(And notice that it might not be just one commit: we only report the
1564"tip of the line" as being dangling, but there might be a whole deep
79c96c57 1565and complex commit history that was dropped.)
559e4d7a
BF
1566
1567If you decide you want the history back, you can always create a new
1568reference pointing to it, for example, a new branch:
1569
1570------------------------------------------------
1571$ git branch recovered-branch 7281251ddd
1572------------------------------------------------
1573
11e016a3 1574
d19fbc3c 1575Sharing development with others
b684f830 1576===============================
d19fbc3c
BF
1577
1578[[getting-updates-with-git-pull]]
1579Getting updates with git pull
b684f830 1580-----------------------------
d19fbc3c
BF
1581
1582After you clone a repository and make a few changes of your own, you
1583may wish to check the original repository for updates and merge them
1584into your own work.
1585
1586We have already seen <<Updating-a-repository-with-git-fetch,how to
1587keep remote tracking branches up to date>> with gitlink:git-fetch[1],
1588and how to merge two branches. So you can merge in changes from the
1589original repository's master branch with:
1590
1591-------------------------------------------------
1592$ git fetch
1593$ git merge origin/master
1594-------------------------------------------------
1595
1596However, the gitlink:git-pull[1] command provides a way to do this in
1597one step:
1598
1599-------------------------------------------------
1600$ git pull origin master
1601-------------------------------------------------
1602
1603In fact, "origin" is normally the default repository to pull from,
1604and the default branch is normally the HEAD of the remote repository,
1605so often you can accomplish the above with just
1606
1607-------------------------------------------------
1608$ git pull
1609-------------------------------------------------
1610
1611See the descriptions of the branch.<name>.remote and
9d13bda3 1612branch.<name>.merge options in gitlink:git-config[1] to learn
d19fbc3c
BF
1613how to control these defaults depending on the current branch.
1614
1615In addition to saving you keystrokes, "git pull" also helps you by
1616producing a default commit message documenting the branch and
1617repository that you pulled from.
1618
1619(But note that no such commit will be created in the case of a
1620<<fast-forwards,fast forward>>; instead, your branch will just be
79c96c57 1621updated to point to the latest commit from the upstream branch.)
d19fbc3c 1622
1191ee18
BF
1623The git-pull command can also be given "." as the "remote" repository,
1624in which case it just merges in a branch from the current repository; so
4c63ff45
BF
1625the commands
1626
1627-------------------------------------------------
1628$ git pull . branch
1629$ git merge branch
1630-------------------------------------------------
1631
1632are roughly equivalent. The former is actually very commonly used.
1633
d19fbc3c 1634Submitting patches to a project
b684f830 1635-------------------------------
d19fbc3c
BF
1636
1637If you just have a few changes, the simplest way to submit them may
1638just be to send them as patches in email:
1639
036be17e 1640First, use gitlink:git-format-patch[1]; for example:
d19fbc3c
BF
1641
1642-------------------------------------------------
eb6ae7f4 1643$ git format-patch origin
d19fbc3c
BF
1644-------------------------------------------------
1645
1646will produce a numbered series of files in the current directory, one
1647for each patch in the current branch but not in origin/HEAD.
1648
1649You can then import these into your mail client and send them by
1650hand. However, if you have a lot to send at once, you may prefer to
1651use the gitlink:git-send-email[1] script to automate the process.
1652Consult the mailing list for your project first to determine how they
1653prefer such patches be handled.
1654
1655Importing patches to a project
b684f830 1656------------------------------
d19fbc3c
BF
1657
1658Git also provides a tool called gitlink:git-am[1] (am stands for
1659"apply mailbox"), for importing such an emailed series of patches.
1660Just save all of the patch-containing messages, in order, into a
1661single mailbox file, say "patches.mbox", then run
1662
1663-------------------------------------------------
eb6ae7f4 1664$ git am -3 patches.mbox
d19fbc3c
BF
1665-------------------------------------------------
1666
1667Git will apply each patch in order; if any conflicts are found, it
1668will stop, and you can fix the conflicts as described in
01997b4a
BF
1669"<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1670git to perform a merge; if you would prefer it just to abort and
1671leave your tree and index untouched, you may omit that option.)
1672
1673Once the index is updated with the results of the conflict
1674resolution, instead of creating a new commit, just run
d19fbc3c
BF
1675
1676-------------------------------------------------
1677$ git am --resolved
1678-------------------------------------------------
1679
1680and git will create the commit for you and continue applying the
1681remaining patches from the mailbox.
1682
1683The final result will be a series of commits, one for each patch in
1684the original mailbox, with authorship and commit log message each
1685taken from the message containing each patch.
1686
1687[[setting-up-a-public-repository]]
1688Setting up a public repository
b684f830 1689------------------------------
d19fbc3c
BF
1690
1691Another way to submit changes to a project is to simply tell the
1692maintainer of that project to pull from your repository, exactly as
1693you did in the section "<<getting-updates-with-git-pull, Getting
1694updates with git pull>>".
1695
1696If you and maintainer both have accounts on the same machine, then
1697then you can just pull changes from each other's repositories
79c96c57
MC
1698directly; note that all of the commands (gitlink:git-clone[1],
1699git-fetch[1], git-pull[1], etc.) that accept a URL as an argument
21f13ee2 1700will also accept a local directory name; so, for example, you can
d19fbc3c
BF
1701use
1702
1703-------------------------------------------------
1704$ git clone /path/to/repository
1705$ git pull /path/to/other/repository
1706-------------------------------------------------
1707
1708If this sort of setup is inconvenient or impossible, another (more
1709common) option is to set up a public repository on a public server.
1710This also allows you to cleanly separate private work in progress
1711from publicly visible work.
1712
1713You will continue to do your day-to-day work in your personal
1714repository, but periodically "push" changes from your personal
1715repository into your public repository, allowing other developers to
1716pull from that repository. So the flow of changes, in a situation
1717where there is one other developer with a public repository, looks
1718like this:
1719
1720 you push
1721 your personal repo ------------------> your public repo
1722 ^ |
1723 | |
1724 | you pull | they pull
1725 | |
1726 | |
1727 | they push V
1728 their public repo <------------------- their repo
1729
1730Now, assume your personal repository is in the directory ~/proj. We
1731first create a new clone of the repository:
1732
1733-------------------------------------------------
1734$ git clone --bare proj-clone.git
1735-------------------------------------------------
1736
1737The resulting directory proj-clone.git will contains a "bare" git
1738repository--it is just the contents of the ".git" directory, without
1739a checked-out copy of a working directory.
1740
1741Next, copy proj-clone.git to the server where you plan to host the
1742public repository. You can use scp, rsync, or whatever is most
1743convenient.
1744
1745If somebody else maintains the public server, they may already have
1746set up a git service for you, and you may skip to the section
1747"<<pushing-changes-to-a-public-repository,Pushing changes to a public
1748repository>>", below.
1749
1750Otherwise, the following sections explain how to export your newly
1751created public repository:
1752
1753[[exporting-via-http]]
1754Exporting a git repository via http
b684f830 1755-----------------------------------
d19fbc3c
BF
1756
1757The git protocol gives better performance and reliability, but on a
1758host with a web server set up, http exports may be simpler to set up.
1759
1760All you need to do is place the newly created bare git repository in
1761a directory that is exported by the web server, and make some
1762adjustments to give web clients some extra information they need:
1763
1764-------------------------------------------------
1765$ mv proj.git /home/you/public_html/proj.git
1766$ cd proj.git
1767$ git update-server-info
1768$ chmod a+x hooks/post-update
1769-------------------------------------------------
1770
1771(For an explanation of the last two lines, see
1772gitlink:git-update-server-info[1], and the documentation
1773link:hooks.txt[Hooks used by git].)
1774
1775Advertise the url of proj.git. Anybody else should then be able to
1776clone or pull from that url, for example with a commandline like:
1777
1778-------------------------------------------------
1779$ git clone http://yourserver.com/~you/proj.git
1780-------------------------------------------------
1781
1782(See also
1783link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1784for a slightly more sophisticated setup using WebDAV which also
1785allows pushing over http.)
1786
1787[[exporting-via-git]]
1788Exporting a git repository via the git protocol
b684f830 1789-----------------------------------------------
d19fbc3c
BF
1790
1791This is the preferred method.
1792
1793For now, we refer you to the gitlink:git-daemon[1] man page for
1794instructions. (See especially the examples section.)
1795
1796[[pushing-changes-to-a-public-repository]]
1797Pushing changes to a public repository
b684f830 1798--------------------------------------
d19fbc3c
BF
1799
1800Note that the two techniques outline above (exporting via
1801<<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1802maintainers to fetch your latest changes, but they do not allow write
1803access, which you will need to update the public repository with the
1804latest changes created in your private repository.
1805
1806The simplest way to do this is using gitlink:git-push[1] and ssh; to
1807update the remote branch named "master" with the latest state of your
1808branch named "master", run
1809
1810-------------------------------------------------
1811$ git push ssh://yourserver.com/~you/proj.git master:master
1812-------------------------------------------------
1813
1814or just
1815
1816-------------------------------------------------
1817$ git push ssh://yourserver.com/~you/proj.git master
1818-------------------------------------------------
1819
1820As with git-fetch, git-push will complain if this does not result in
1821a <<fast-forwards,fast forward>>. Normally this is a sign of
1822something wrong. However, if you are sure you know what you're
1823doing, you may force git-push to perform the update anyway by
1824proceeding the branch name by a plus sign:
1825
1826-------------------------------------------------
1827$ git push ssh://yourserver.com/~you/proj.git +master
1828-------------------------------------------------
1829
1830As with git-fetch, you may also set up configuration options to
1831save typing; so, for example, after
1832
1833-------------------------------------------------
1834$ cat >.git/config <<EOF
1835[remote "public-repo"]
1836 url = ssh://yourserver.com/~you/proj.git
1837EOF
1838-------------------------------------------------
1839
1840you should be able to perform the above push with just
1841
1842-------------------------------------------------
1843$ git push public-repo master
1844-------------------------------------------------
1845
1846See the explanations of the remote.<name>.url, branch.<name>.remote,
9d13bda3 1847and remote.<name>.push options in gitlink:git-config[1] for
d19fbc3c
BF
1848details.
1849
1850Setting up a shared repository
b684f830 1851------------------------------
d19fbc3c
BF
1852
1853Another way to collaborate is by using a model similar to that
1854commonly used in CVS, where several developers with special rights
1855all push to and pull from a single shared repository. See
1856link:cvs-migration.txt[git for CVS users] for instructions on how to
1857set this up.
1858
b684f830
BF
1859Allow web browsing of a repository
1860----------------------------------
d19fbc3c 1861
a8cd1402
BF
1862The gitweb cgi script provides users an easy way to browse your
1863project's files and history without having to install git; see the file
04483524 1864gitweb/INSTALL in the git source tree for instructions on setting it up.
d19fbc3c 1865
b684f830
BF
1866Examples
1867--------
d19fbc3c 1868
b684f830 1869TODO: topic branches, typical roles as in everyday.txt, ?
d19fbc3c 1870
d19fbc3c 1871
d19fbc3c 1872[[cleaning-up-history]]
4c63ff45
BF
1873Rewriting history and maintaining patch series
1874==============================================
1875
1876Normally commits are only added to a project, never taken away or
1877replaced. Git is designed with this assumption, and violating it will
1878cause git's merge machinery (for example) to do the wrong thing.
1879
1880However, there is a situation in which it can be useful to violate this
1881assumption.
1882
1883Creating the perfect patch series
1884---------------------------------
1885
1886Suppose you are a contributor to a large project, and you want to add a
1887complicated feature, and to present it to the other developers in a way
1888that makes it easy for them to read your changes, verify that they are
1889correct, and understand why you made each change.
1890
b181d57f 1891If you present all of your changes as a single patch (or commit), they
79c96c57 1892may find that it is too much to digest all at once.
4c63ff45
BF
1893
1894If you present them with the entire history of your work, complete with
1895mistakes, corrections, and dead ends, they may be overwhelmed.
1896
1897So the ideal is usually to produce a series of patches such that:
1898
1899 1. Each patch can be applied in order.
1900
1901 2. Each patch includes a single logical change, together with a
1902 message explaining the change.
1903
1904 3. No patch introduces a regression: after applying any initial
1905 part of the series, the resulting project still compiles and
1906 works, and has no bugs that it didn't have before.
1907
1908 4. The complete series produces the same end result as your own
1909 (probably much messier!) development process did.
1910
b181d57f
BF
1911We will introduce some tools that can help you do this, explain how to
1912use them, and then explain some of the problems that can arise because
1913you are rewriting history.
4c63ff45
BF
1914
1915Keeping a patch series up to date using git-rebase
1916--------------------------------------------------
1917
79c96c57
MC
1918Suppose that you create a branch "mywork" on a remote-tracking branch
1919"origin", and create some commits on top of it:
4c63ff45
BF
1920
1921-------------------------------------------------
1922$ git checkout -b mywork origin
1923$ vi file.txt
1924$ git commit
1925$ vi otherfile.txt
1926$ git commit
1927...
1928-------------------------------------------------
1929
1930You have performed no merges into mywork, so it is just a simple linear
1931sequence of patches on top of "origin":
1932
1dc71a91 1933................................................
4c63ff45
BF
1934 o--o--o <-- origin
1935 \
1936 o--o--o <-- mywork
1dc71a91 1937................................................
4c63ff45
BF
1938
1939Some more interesting work has been done in the upstream project, and
1940"origin" has advanced:
1941
1dc71a91 1942................................................
4c63ff45
BF
1943 o--o--O--o--o--o <-- origin
1944 \
1945 a--b--c <-- mywork
1dc71a91 1946................................................
4c63ff45
BF
1947
1948At this point, you could use "pull" to merge your changes back in;
1949the result would create a new merge commit, like this:
1950
1dc71a91 1951................................................
4c63ff45
BF
1952 o--o--O--o--o--o <-- origin
1953 \ \
1954 a--b--c--m <-- mywork
1dc71a91 1955................................................
4c63ff45
BF
1956
1957However, if you prefer to keep the history in mywork a simple series of
1958commits without any merges, you may instead choose to use
1959gitlink:git-rebase[1]:
1960
1961-------------------------------------------------
1962$ git checkout mywork
1963$ git rebase origin
1964-------------------------------------------------
1965
b181d57f
BF
1966This will remove each of your commits from mywork, temporarily saving
1967them as patches (in a directory named ".dotest"), update mywork to
1968point at the latest version of origin, then apply each of the saved
1969patches to the new mywork. The result will look like:
4c63ff45
BF
1970
1971
1dc71a91 1972................................................
4c63ff45
BF
1973 o--o--O--o--o--o <-- origin
1974 \
1975 a'--b'--c' <-- mywork
1dc71a91 1976................................................
4c63ff45 1977
b181d57f
BF
1978In the process, it may discover conflicts. In that case it will stop
1979and allow you to fix the conflicts; after fixing conflicts, use "git
1980add" to update the index with those contents, and then, instead of
1981running git-commit, just run
4c63ff45
BF
1982
1983-------------------------------------------------
1984$ git rebase --continue
1985-------------------------------------------------
1986
1987and git will continue applying the rest of the patches.
1988
1989At any point you may use the --abort option to abort this process and
1990return mywork to the state it had before you started the rebase:
1991
1992-------------------------------------------------
1993$ git rebase --abort
1994-------------------------------------------------
1995
365aa199
BF
1996Modifying a single commit
1997-------------------------
1998
1999We saw in <<fixing-a-mistake-by-editing-history>> that you can replace the
2000most recent commit using
2001
2002-------------------------------------------------
2003$ git commit --amend
2004-------------------------------------------------
2005
2006which will replace the old commit by a new commit incorporating your
2007changes, giving you a chance to edit the old commit message first.
2008
2009You can also use a combination of this and gitlink:git-rebase[1] to edit
2010commits further back in your history. First, tag the problematic commit with
2011
2012-------------------------------------------------
2013$ git tag bad mywork~5
2014-------------------------------------------------
2015
2016(Either gitk or git-log may be useful for finding the commit.)
2017
2018Then check out a new branch at that commit, edit it, and rebase the rest of
2019the series on top of it:
2020
2021-------------------------------------------------
2022$ git checkout -b TMP bad
2023$ # make changes here and update the index
2024$ git commit --amend
2025$ git rebase --onto TMP bad mywork
2026-------------------------------------------------
2027
2028When you're done, you'll be left with mywork checked out, with the top patches
2029on mywork reapplied on top of the modified commit you created in TMP. You can
2030then clean up with
2031
2032-------------------------------------------------
2033$ git branch -d TMP
2034$ git tag -d bad
2035-------------------------------------------------
2036
2037Note that the immutable nature of git history means that you haven't really
2038"modified" existing commits; instead, you have replaced the old commits with
2039new commits having new object names.
2040
4c63ff45
BF
2041Reordering or selecting from a patch series
2042-------------------------------------------
2043
b181d57f
BF
2044Given one existing commit, the gitlink:git-cherry-pick[1] command
2045allows you to apply the change introduced by that commit and create a
2046new commit that records it. So, for example, if "mywork" points to a
2047series of patches on top of "origin", you might do something like:
2048
2049-------------------------------------------------
2050$ git checkout -b mywork-new origin
2051$ gitk origin..mywork &
2052-------------------------------------------------
2053
2054And browse through the list of patches in the mywork branch using gitk,
2055applying them (possibly in a different order) to mywork-new using
2056cherry-pick, and possibly modifying them as you go using commit
2057--amend.
2058
2059Another technique is to use git-format-patch to create a series of
2060patches, then reset the state to before the patches:
4c63ff45 2061
b181d57f
BF
2062-------------------------------------------------
2063$ git format-patch origin
2064$ git reset --hard origin
2065-------------------------------------------------
4c63ff45 2066
b181d57f
BF
2067Then modify, reorder, or eliminate patches as preferred before applying
2068them again with gitlink:git-am[1].
4c63ff45
BF
2069
2070Other tools
2071-----------
2072
b181d57f 2073There are numerous other tools, such as stgit, which exist for the
79c96c57 2074purpose of maintaining a patch series. These are outside of the scope of
b181d57f 2075this manual.
4c63ff45
BF
2076
2077Problems with rewriting history
2078-------------------------------
2079
b181d57f
BF
2080The primary problem with rewriting the history of a branch has to do
2081with merging. Suppose somebody fetches your branch and merges it into
2082their branch, with a result something like this:
2083
1dc71a91 2084................................................
b181d57f
BF
2085 o--o--O--o--o--o <-- origin
2086 \ \
2087 t--t--t--m <-- their branch:
1dc71a91 2088................................................
b181d57f
BF
2089
2090Then suppose you modify the last three commits:
2091
1dc71a91 2092................................................
b181d57f
BF
2093 o--o--o <-- new head of origin
2094 /
2095 o--o--O--o--o--o <-- old head of origin
1dc71a91 2096................................................
b181d57f
BF
2097
2098If we examined all this history together in one repository, it will
2099look like:
2100
1dc71a91 2101................................................
b181d57f
BF
2102 o--o--o <-- new head of origin
2103 /
2104 o--o--O--o--o--o <-- old head of origin
2105 \ \
2106 t--t--t--m <-- their branch:
1dc71a91 2107................................................
b181d57f
BF
2108
2109Git has no way of knowing that the new head is an updated version of
2110the old head; it treats this situation exactly the same as it would if
2111two developers had independently done the work on the old and new heads
2112in parallel. At this point, if someone attempts to merge the new head
2113in to their branch, git will attempt to merge together the two (old and
2114new) lines of development, instead of trying to replace the old by the
2115new. The results are likely to be unexpected.
2116
2117You may still choose to publish branches whose history is rewritten,
2118and it may be useful for others to be able to fetch those branches in
2119order to examine or test them, but they should not attempt to pull such
2120branches into their own work.
2121
2122For true distributed development that supports proper merging,
2123published branches should never be rewritten.
2124
2125Advanced branch management
2126==========================
4c63ff45 2127
b181d57f
BF
2128Fetching individual branches
2129----------------------------
2130
2131Instead of using gitlink:git-remote[1], you can also choose just
2132to update one branch at a time, and to store it locally under an
2133arbitrary name:
2134
2135-------------------------------------------------
2136$ git fetch origin todo:my-todo-work
2137-------------------------------------------------
2138
2139The first argument, "origin", just tells git to fetch from the
2140repository you originally cloned from. The second argument tells git
2141to fetch the branch named "todo" from the remote repository, and to
2142store it locally under the name refs/heads/my-todo-work.
2143
2144You can also fetch branches from other repositories; so
2145
2146-------------------------------------------------
2147$ git fetch git://example.com/proj.git master:example-master
2148-------------------------------------------------
2149
2150will create a new branch named "example-master" and store in it the
2151branch named "master" from the repository at the given URL. If you
2152already have a branch named example-master, it will attempt to
2153"fast-forward" to the commit given by example.com's master branch. So
2154next we explain what a fast-forward is:
2155
2156[[fast-forwards]]
2157Understanding git history: fast-forwards
2158----------------------------------------
2159
2160In the previous example, when updating an existing branch, "git
2161fetch" checks to make sure that the most recent commit on the remote
2162branch is a descendant of the most recent commit on your copy of the
2163branch before updating your copy of the branch to point at the new
2164commit. Git calls this process a "fast forward".
2165
2166A fast forward looks something like this:
2167
1dc71a91 2168................................................
b181d57f
BF
2169 o--o--o--o <-- old head of the branch
2170 \
2171 o--o--o <-- new head of the branch
1dc71a91 2172................................................
b181d57f
BF
2173
2174
2175In some cases it is possible that the new head will *not* actually be
2176a descendant of the old head. For example, the developer may have
2177realized she made a serious mistake, and decided to backtrack,
2178resulting in a situation like:
2179
1dc71a91 2180................................................
b181d57f
BF
2181 o--o--o--o--a--b <-- old head of the branch
2182 \
2183 o--o--o <-- new head of the branch
1dc71a91 2184................................................
b181d57f
BF
2185
2186In this case, "git fetch" will fail, and print out a warning.
2187
2188In that case, you can still force git to update to the new head, as
2189described in the following section. However, note that in the
2190situation above this may mean losing the commits labeled "a" and "b",
2191unless you've already created a reference of your own pointing to
2192them.
2193
2194Forcing git fetch to do non-fast-forward updates
2195------------------------------------------------
2196
2197If git fetch fails because the new head of a branch is not a
2198descendant of the old head, you may force the update with:
2199
2200-------------------------------------------------
2201$ git fetch git://example.com/proj.git +master:refs/remotes/example/master
2202-------------------------------------------------
2203
79c96c57 2204Note the addition of the "+" sign. Be aware that commits that the
b181d57f
BF
2205old version of example/master pointed at may be lost, as we saw in
2206the previous section.
2207
2208Configuring remote branches
2209---------------------------
2210
2211We saw above that "origin" is just a shortcut to refer to the
79c96c57 2212repository that you originally cloned from. This information is
b181d57f 2213stored in git configuration variables, which you can see using
9d13bda3 2214gitlink:git-config[1]:
b181d57f
BF
2215
2216-------------------------------------------------
9d13bda3 2217$ git config -l
b181d57f
BF
2218core.repositoryformatversion=0
2219core.filemode=true
2220core.logallrefupdates=true
2221remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2222remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2223branch.master.remote=origin
2224branch.master.merge=refs/heads/master
2225-------------------------------------------------
2226
2227If there are other repositories that you also use frequently, you can
2228create similar configuration options to save typing; for example,
2229after
2230
2231-------------------------------------------------
9d13bda3 2232$ git config remote.example.url git://example.com/proj.git
b181d57f
BF
2233-------------------------------------------------
2234
2235then the following two commands will do the same thing:
2236
2237-------------------------------------------------
2238$ git fetch git://example.com/proj.git master:refs/remotes/example/master
2239$ git fetch example master:refs/remotes/example/master
2240-------------------------------------------------
2241
2242Even better, if you add one more option:
2243
2244-------------------------------------------------
9d13bda3 2245$ git config remote.example.fetch master:refs/remotes/example/master
b181d57f
BF
2246-------------------------------------------------
2247
2248then the following commands will all do the same thing:
2249
2250-------------------------------------------------
2251$ git fetch git://example.com/proj.git master:ref/remotes/example/master
2252$ git fetch example master:ref/remotes/example/master
2253$ git fetch example example/master
2254$ git fetch example
2255-------------------------------------------------
2256
2257You can also add a "+" to force the update each time:
2258
2259-------------------------------------------------
9d13bda3 2260$ git config remote.example.fetch +master:ref/remotes/example/master
b181d57f
BF
2261-------------------------------------------------
2262
2263Don't do this unless you're sure you won't mind "git fetch" possibly
2264throwing away commits on mybranch.
2265
2266Also note that all of the above configuration can be performed by
2267directly editing the file .git/config instead of using
9d13bda3 2268gitlink:git-config[1].
b181d57f 2269
9d13bda3 2270See gitlink:git-config[1] for more details on the configuration
b181d57f 2271options mentioned above.
d19fbc3c 2272
d19fbc3c 2273
35121930 2274[[git-internals]]
d19fbc3c
BF
2275Git internals
2276=============
2277
b181d57f
BF
2278There are two object abstractions: the "object database", and the
2279"current directory cache" aka "index".
2280
2281The Object Database
2282-------------------
2283
2284The object database is literally just a content-addressable collection
2285of objects. All objects are named by their content, which is
2286approximated by the SHA1 hash of the object itself. Objects may refer
2287to other objects (by referencing their SHA1 hash), and so you can
2288build up a hierarchy of objects.
2289
2290All objects have a statically determined "type" aka "tag", which is
2291determined at object creation time, and which identifies the format of
2292the object (i.e. how it is used, and how it can refer to other
2293objects). There are currently four different object types: "blob",
2294"tree", "commit" and "tag".
2295
2296A "blob" object cannot refer to any other object, and is, like the type
2297implies, a pure storage object containing some user data. It is used to
2298actually store the file data, i.e. a blob object is associated with some
2299particular version of some file.
2300
2301A "tree" object is an object that ties one or more "blob" objects into a
2302directory structure. In addition, a tree object can refer to other tree
2303objects, thus creating a directory hierarchy.
2304
2305A "commit" object ties such directory hierarchies together into
2306a DAG of revisions - each "commit" is associated with exactly one tree
2307(the directory hierarchy at the time of the commit). In addition, a
2308"commit" refers to one or more "parent" commit objects that describe the
2309history of how we arrived at that directory hierarchy.
2310
2311As a special case, a commit object with no parents is called the "root"
2312object, and is the point of an initial project commit. Each project
2313must have at least one root, and while you can tie several different
2314root objects together into one project by creating a commit object which
2315has two or more separate roots as its ultimate parents, that's probably
2316just going to confuse people. So aim for the notion of "one root object
2317per project", even if git itself does not enforce that.
2318
2319A "tag" object symbolically identifies and can be used to sign other
2320objects. It contains the identifier and type of another object, a
2321symbolic name (of course!) and, optionally, a signature.
2322
2323Regardless of object type, all objects share the following
2324characteristics: they are all deflated with zlib, and have a header
2325that not only specifies their type, but also provides size information
2326about the data in the object. It's worth noting that the SHA1 hash
2327that is used to name the object is the hash of the original data
2328plus this header, so `sha1sum` 'file' does not match the object name
2329for 'file'.
2330(Historical note: in the dawn of the age of git the hash
2331was the sha1 of the 'compressed' object.)
2332
2333As a result, the general consistency of an object can always be tested
2334independently of the contents or the type of the object: all objects can
2335be validated by verifying that (a) their hashes match the content of the
2336file and (b) the object successfully inflates to a stream of bytes that
2337forms a sequence of <ascii type without space> + <space> + <ascii decimal
2338size> + <byte\0> + <binary object data>.
2339
2340The structured objects can further have their structure and
2341connectivity to other objects verified. This is generally done with
04e50e94 2342the `git-fsck` program, which generates a full dependency graph
b181d57f
BF
2343of all objects, and verifies their internal consistency (in addition
2344to just verifying their superficial consistency through the hash).
2345
2346The object types in some more detail:
2347
2348Blob Object
2349-----------
2350
2351A "blob" object is nothing but a binary blob of data, and doesn't
2352refer to anything else. There is no signature or any other
2353verification of the data, so while the object is consistent (it 'is'
2354indexed by its sha1 hash, so the data itself is certainly correct), it
2355has absolutely no other attributes. No name associations, no
2356permissions. It is purely a blob of data (i.e. normally "file
2357contents").
2358
2359In particular, since the blob is entirely defined by its data, if two
2360files in a directory tree (or in multiple different versions of the
2361repository) have the same contents, they will share the same blob
2362object. The object is totally independent of its location in the
2363directory tree, and renaming a file does not change the object that
2364file is associated with in any way.
2365
2366A blob is typically created when gitlink:git-update-index[1]
2367is run, and its data can be accessed by gitlink:git-cat-file[1].
2368
2369Tree Object
2370-----------
2371
2372The next hierarchical object type is the "tree" object. A tree object
2373is a list of mode/name/blob data, sorted by name. Alternatively, the
2374mode data may specify a directory mode, in which case instead of
2375naming a blob, that name is associated with another TREE object.
2376
2377Like the "blob" object, a tree object is uniquely determined by the
2378set contents, and so two separate but identical trees will always
2379share the exact same object. This is true at all levels, i.e. it's
2380true for a "leaf" tree (which does not refer to any other trees, only
2381blobs) as well as for a whole subdirectory.
2382
2383For that reason a "tree" object is just a pure data abstraction: it
2384has no history, no signatures, no verification of validity, except
2385that since the contents are again protected by the hash itself, we can
2386trust that the tree is immutable and its contents never change.
2387
2388So you can trust the contents of a tree to be valid, the same way you
2389can trust the contents of a blob, but you don't know where those
2390contents 'came' from.
2391
2392Side note on trees: since a "tree" object is a sorted list of
2393"filename+content", you can create a diff between two trees without
2394actually having to unpack two trees. Just ignore all common parts,
2395and your diff will look right. In other words, you can effectively
2396(and efficiently) tell the difference between any two random trees by
2397O(n) where "n" is the size of the difference, rather than the size of
2398the tree.
2399
2400Side note 2 on trees: since the name of a "blob" depends entirely and
2401exclusively on its contents (i.e. there are no names or permissions
2402involved), you can see trivial renames or permission changes by
2403noticing that the blob stayed the same. However, renames with data
2404changes need a smarter "diff" implementation.
2405
2406A tree is created with gitlink:git-write-tree[1] and
2407its data can be accessed by gitlink:git-ls-tree[1].
2408Two trees can be compared with gitlink:git-diff-tree[1].
2409
2410Commit Object
2411-------------
2412
2413The "commit" object is an object that introduces the notion of
2414history into the picture. In contrast to the other objects, it
2415doesn't just describe the physical state of a tree, it describes how
2416we got there, and why.
2417
2418A "commit" is defined by the tree-object that it results in, the
2419parent commits (zero, one or more) that led up to that point, and a
2420comment on what happened. Again, a commit is not trusted per se:
2421the contents are well-defined and "safe" due to the cryptographically
2422strong signatures at all levels, but there is no reason to believe
2423that the tree is "good" or that the merge information makes sense.
2424The parents do not have to actually have any relationship with the
2425result, for example.
2426
2427Note on commits: unlike real SCM's, commits do not contain
2428rename information or file mode change information. All of that is
2429implicit in the trees involved (the result tree, and the result trees
2430of the parents), and describing that makes no sense in this idiotic
2431file manager.
2432
2433A commit is created with gitlink:git-commit-tree[1] and
2434its data can be accessed by gitlink:git-cat-file[1].
2435
2436Trust
2437-----
2438
2439An aside on the notion of "trust". Trust is really outside the scope
2440of "git", but it's worth noting a few things. First off, since
2441everything is hashed with SHA1, you 'can' trust that an object is
2442intact and has not been messed with by external sources. So the name
2443of an object uniquely identifies a known state - just not a state that
2444you may want to trust.
2445
2446Furthermore, since the SHA1 signature of a commit refers to the
2447SHA1 signatures of the tree it is associated with and the signatures
2448of the parent, a single named commit specifies uniquely a whole set
2449of history, with full contents. You can't later fake any step of the
2450way once you have the name of a commit.
2451
2452So to introduce some real trust in the system, the only thing you need
2453to do is to digitally sign just 'one' special note, which includes the
2454name of a top-level commit. Your digital signature shows others
2455that you trust that commit, and the immutability of the history of
2456commits tells others that they can trust the whole history.
2457
2458In other words, you can easily validate a whole archive by just
2459sending out a single email that tells the people the name (SHA1 hash)
2460of the top commit, and digitally sign that email using something
2461like GPG/PGP.
2462
2463To assist in this, git also provides the tag object...
2464
2465Tag Object
2466----------
2467
2468Git provides the "tag" object to simplify creating, managing and
2469exchanging symbolic and signed tokens. The "tag" object at its
2470simplest simply symbolically identifies another object by containing
2471the sha1, type and symbolic name.
2472
2473However it can optionally contain additional signature information
2474(which git doesn't care about as long as there's less than 8k of
2475it). This can then be verified externally to git.
2476
2477Note that despite the tag features, "git" itself only handles content
2478integrity; the trust framework (and signature provision and
2479verification) has to come from outside.
2480
2481A tag is created with gitlink:git-mktag[1],
2482its data can be accessed by gitlink:git-cat-file[1],
2483and the signature can be verified by
2484gitlink:git-verify-tag[1].
2485
2486
2487The "index" aka "Current Directory Cache"
2488-----------------------------------------
2489
2490The index is a simple binary file, which contains an efficient
2491representation of a virtual directory content at some random time. It
2492does so by a simple array that associates a set of names, dates,
2493permissions and content (aka "blob") objects together. The cache is
2494always kept ordered by name, and names are unique (with a few very
2495specific rules) at any point in time, but the cache has no long-term
2496meaning, and can be partially updated at any time.
2497
2498In particular, the index certainly does not need to be consistent with
2499the current directory contents (in fact, most operations will depend on
2500different ways to make the index 'not' be consistent with the directory
2501hierarchy), but it has three very important attributes:
2502
2503'(a) it can re-generate the full state it caches (not just the
2504directory structure: it contains pointers to the "blob" objects so
2505that it can regenerate the data too)'
2506
2507As a special case, there is a clear and unambiguous one-way mapping
2508from a current directory cache to a "tree object", which can be
2509efficiently created from just the current directory cache without
2510actually looking at any other data. So a directory cache at any one
2511time uniquely specifies one and only one "tree" object (but has
2512additional data to make it easy to match up that tree object with what
2513has happened in the directory)
2514
2515'(b) it has efficient methods for finding inconsistencies between that
2516cached state ("tree object waiting to be instantiated") and the
2517current state.'
2518
2519'(c) it can additionally efficiently represent information about merge
2520conflicts between different tree objects, allowing each pathname to be
2521associated with sufficient information about the trees involved that
2522you can create a three-way merge between them.'
2523
79c96c57 2524Those are the ONLY three things that the directory cache does. It's a
b181d57f
BF
2525cache, and the normal operation is to re-generate it completely from a
2526known tree object, or update/compare it with a live tree that is being
2527developed. If you blow the directory cache away entirely, you generally
2528haven't lost any information as long as you have the name of the tree
2529that it described.
2530
2531At the same time, the index is at the same time also the
2532staging area for creating new trees, and creating a new tree always
2533involves a controlled modification of the index file. In particular,
2534the index file can have the representation of an intermediate tree that
2535has not yet been instantiated. So the index can be thought of as a
2536write-back cache, which can contain dirty information that has not yet
2537been written back to the backing store.
2538
2539
2540
2541The Workflow
2542------------
2543
2544Generally, all "git" operations work on the index file. Some operations
2545work *purely* on the index file (showing the current state of the
2546index), but most operations move data to and from the index file. Either
2547from the database or from the working directory. Thus there are four
2548main combinations:
2549
2550working directory -> index
2551~~~~~~~~~~~~~~~~~~~~~~~~~~
2552
2553You update the index with information from the working directory with
2554the gitlink:git-update-index[1] command. You
2555generally update the index information by just specifying the filename
2556you want to update, like so:
2557
2558-------------------------------------------------
2559$ git-update-index filename
2560-------------------------------------------------
2561
2562but to avoid common mistakes with filename globbing etc, the command
2563will not normally add totally new entries or remove old entries,
2564i.e. it will normally just update existing cache entries.
2565
2566To tell git that yes, you really do realize that certain files no
2567longer exist, or that new files should be added, you
2568should use the `--remove` and `--add` flags respectively.
2569
2570NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
2571necessarily be removed: if the files still exist in your directory
2572structure, the index will be updated with their new status, not
2573removed. The only thing `--remove` means is that update-cache will be
2574considering a removed file to be a valid thing, and if the file really
2575does not exist any more, it will update the index accordingly.
2576
2577As a special case, you can also do `git-update-index --refresh`, which
2578will refresh the "stat" information of each index to match the current
2579stat information. It will 'not' update the object status itself, and
2580it will only update the fields that are used to quickly test whether
2581an object still matches its old backing store object.
2582
2583index -> object database
2584~~~~~~~~~~~~~~~~~~~~~~~~
2585
2586You write your current index file to a "tree" object with the program
2587
2588-------------------------------------------------
2589$ git-write-tree
2590-------------------------------------------------
2591
2592that doesn't come with any options - it will just write out the
2593current index into the set of tree objects that describe that state,
2594and it will return the name of the resulting top-level tree. You can
2595use that tree to re-generate the index at any time by going in the
2596other direction:
2597
2598object database -> index
2599~~~~~~~~~~~~~~~~~~~~~~~~
2600
2601You read a "tree" file from the object database, and use that to
2602populate (and overwrite - don't do this if your index contains any
2603unsaved state that you might want to restore later!) your current
2604index. Normal operation is just
2605
2606-------------------------------------------------
2607$ git-read-tree <sha1 of tree>
2608-------------------------------------------------
2609
2610and your index file will now be equivalent to the tree that you saved
2611earlier. However, that is only your 'index' file: your working
2612directory contents have not been modified.
2613
2614index -> working directory
2615~~~~~~~~~~~~~~~~~~~~~~~~~~
2616
2617You update your working directory from the index by "checking out"
2618files. This is not a very common operation, since normally you'd just
2619keep your files updated, and rather than write to your working
2620directory, you'd tell the index files about the changes in your
2621working directory (i.e. `git-update-index`).
2622
2623However, if you decide to jump to a new version, or check out somebody
2624else's version, or just restore a previous tree, you'd populate your
2625index file with read-tree, and then you need to check out the result
2626with
2627
2628-------------------------------------------------
2629$ git-checkout-index filename
2630-------------------------------------------------
2631
2632or, if you want to check out all of the index, use `-a`.
2633
2634NOTE! git-checkout-index normally refuses to overwrite old files, so
2635if you have an old version of the tree already checked out, you will
2636need to use the "-f" flag ('before' the "-a" flag or the filename) to
2637'force' the checkout.
2638
2639
2640Finally, there are a few odds and ends which are not purely moving
2641from one representation to the other:
2642
2643Tying it all together
2644~~~~~~~~~~~~~~~~~~~~~
2645
2646To commit a tree you have instantiated with "git-write-tree", you'd
2647create a "commit" object that refers to that tree and the history
2648behind it - most notably the "parent" commits that preceded it in
2649history.
2650
2651Normally a "commit" has one parent: the previous state of the tree
2652before a certain change was made. However, sometimes it can have two
2653or more parent commits, in which case we call it a "merge", due to the
2654fact that such a commit brings together ("merges") two or more
2655previous states represented by other commits.
2656
2657In other words, while a "tree" represents a particular directory state
2658of a working directory, a "commit" represents that state in "time",
2659and explains how we got there.
2660
2661You create a commit object by giving it the tree that describes the
2662state at the time of the commit, and a list of parents:
2663
2664-------------------------------------------------
2665$ git-commit-tree <tree> -p <parent> [-p <parent2> ..]
2666-------------------------------------------------
2667
2668and then giving the reason for the commit on stdin (either through
2669redirection from a pipe or file, or by just typing it at the tty).
2670
2671git-commit-tree will return the name of the object that represents
2672that commit, and you should save it away for later use. Normally,
2673you'd commit a new `HEAD` state, and while git doesn't care where you
2674save the note about that state, in practice we tend to just write the
2675result to the file pointed at by `.git/HEAD`, so that we can always see
2676what the last committed state was.
2677
2678Here is an ASCII art by Jon Loeliger that illustrates how
2679various pieces fit together.
2680
2681------------
2682
2683 commit-tree
2684 commit obj
2685 +----+
2686 | |
2687 | |
2688 V V
2689 +-----------+
2690 | Object DB |
2691 | Backing |
2692 | Store |
2693 +-----------+
2694 ^
2695 write-tree | |
2696 tree obj | |
2697 | | read-tree
2698 | | tree obj
2699 V
2700 +-----------+
2701 | Index |
2702 | "cache" |
2703 +-----------+
2704 update-index ^
2705 blob obj | |
2706 | |
2707 checkout-index -u | | checkout-index
2708 stat | | blob obj
2709 V
2710 +-----------+
2711 | Working |
2712 | Directory |
2713 +-----------+
2714
2715------------
2716
2717
2718Examining the data
2719------------------
2720
2721You can examine the data represented in the object database and the
2722index with various helper tools. For every object, you can use
2723gitlink:git-cat-file[1] to examine details about the
2724object:
2725
2726-------------------------------------------------
2727$ git-cat-file -t <objectname>
2728-------------------------------------------------
2729
2730shows the type of the object, and once you have the type (which is
2731usually implicit in where you find the object), you can use
2732
2733-------------------------------------------------
2734$ git-cat-file blob|tree|commit|tag <objectname>
2735-------------------------------------------------
2736
2737to show its contents. NOTE! Trees have binary content, and as a result
2738there is a special helper for showing that content, called
2739`git-ls-tree`, which turns the binary content into a more easily
2740readable form.
2741
2742It's especially instructive to look at "commit" objects, since those
2743tend to be small and fairly self-explanatory. In particular, if you
2744follow the convention of having the top commit name in `.git/HEAD`,
2745you can do
2746
2747-------------------------------------------------
2748$ git-cat-file commit HEAD
2749-------------------------------------------------
2750
2751to see what the top commit was.
2752
2753Merging multiple trees
d19fbc3c
BF
2754----------------------
2755
b181d57f
BF
2756Git helps you do a three-way merge, which you can expand to n-way by
2757repeating the merge procedure arbitrary times until you finally
2758"commit" the state. The normal situation is that you'd only do one
2759three-way merge (two parents), and commit it, but if you like to, you
2760can do multiple parents in one go.
2761
2762To do a three-way merge, you need the two sets of "commit" objects
2763that you want to merge, use those to find the closest common parent (a
2764third "commit" object), and then use those commit objects to find the
2765state of the directory ("tree" object) at these points.
2766
2767To get the "base" for the merge, you first look up the common parent
2768of two commits with
2769
2770-------------------------------------------------
2771$ git-merge-base <commit1> <commit2>
2772-------------------------------------------------
2773
2774which will return you the commit they are both based on. You should
2775now look up the "tree" objects of those commits, which you can easily
2776do with (for example)
2777
2778-------------------------------------------------
2779$ git-cat-file commit <commitname> | head -1
2780-------------------------------------------------
2781
2782since the tree object information is always the first line in a commit
2783object.
2784
1191ee18
BF
2785Once you know the three trees you are going to merge (the one "original"
2786tree, aka the common case, and the two "result" trees, aka the branches
2787you want to merge), you do a "merge" read into the index. This will
2788complain if it has to throw away your old index contents, so you should
b181d57f 2789make sure that you've committed those - in fact you would normally
1191ee18
BF
2790always do a merge against your last commit (which should thus match what
2791you have in your current index anyway).
b181d57f
BF
2792
2793To do the merge, do
2794
2795-------------------------------------------------
2796$ git-read-tree -m -u <origtree> <yourtree> <targettree>
2797-------------------------------------------------
2798
2799which will do all trivial merge operations for you directly in the
2800index file, and you can just write the result out with
2801`git-write-tree`.
2802
2803
2804Merging multiple trees, continued
2805---------------------------------
2806
2807Sadly, many merges aren't trivial. If there are files that have
2808been added.moved or removed, or if both branches have modified the
2809same file, you will be left with an index tree that contains "merge
2810entries" in it. Such an index tree can 'NOT' be written out to a tree
2811object, and you will have to resolve any such merge clashes using
2812other tools before you can write out the result.
2813
2814You can examine such index state with `git-ls-files --unmerged`
2815command. An example:
2816
2817------------------------------------------------
2818$ git-read-tree -m $orig HEAD $target
2819$ git-ls-files --unmerged
2820100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
2821100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
2822100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
2823------------------------------------------------
2824
2825Each line of the `git-ls-files --unmerged` output begins with
2826the blob mode bits, blob SHA1, 'stage number', and the
2827filename. The 'stage number' is git's way to say which tree it
2828came from: stage 1 corresponds to `$orig` tree, stage 2 `HEAD`
2829tree, and stage3 `$target` tree.
2830
2831Earlier we said that trivial merges are done inside
2832`git-read-tree -m`. For example, if the file did not change
2833from `$orig` to `HEAD` nor `$target`, or if the file changed
2834from `$orig` to `HEAD` and `$orig` to `$target` the same way,
2835obviously the final outcome is what is in `HEAD`. What the
2836above example shows is that file `hello.c` was changed from
2837`$orig` to `HEAD` and `$orig` to `$target` in a different way.
2838You could resolve this by running your favorite 3-way merge
2839program, e.g. `diff3` or `merge`, on the blob objects from
2840these three stages yourself, like this:
2841
2842------------------------------------------------
2843$ git-cat-file blob 263414f... >hello.c~1
2844$ git-cat-file blob 06fa6a2... >hello.c~2
2845$ git-cat-file blob cc44c73... >hello.c~3
2846$ merge hello.c~2 hello.c~1 hello.c~3
2847------------------------------------------------
2848
2849This would leave the merge result in `hello.c~2` file, along
2850with conflict markers if there are conflicts. After verifying
2851the merge result makes sense, you can tell git what the final
2852merge result for this file is by:
2853
2854-------------------------------------------------
2855$ mv -f hello.c~2 hello.c
2856$ git-update-index hello.c
2857-------------------------------------------------
2858
2859When a path is in unmerged state, running `git-update-index` for
2860that path tells git to mark the path resolved.
2861
2862The above is the description of a git merge at the lowest level,
2863to help you understand what conceptually happens under the hood.
2864In practice, nobody, not even git itself, uses three `git-cat-file`
2865for this. There is `git-merge-index` program that extracts the
2866stages to temporary files and calls a "merge" script on it:
2867
2868-------------------------------------------------
2869$ git-merge-index git-merge-one-file hello.c
2870-------------------------------------------------
2871
207dfa07 2872and that is what higher level `git merge -s resolve` is implemented with.
b181d57f
BF
2873
2874How git stores objects efficiently: pack files
2875----------------------------------------------
2876
2877We've seen how git stores each object in a file named after the
2878object's SHA1 hash.
2879
2880Unfortunately this system becomes inefficient once a project has a
2881lot of objects. Try this on an old project:
2882
2883------------------------------------------------
2884$ git count-objects
28856930 objects, 47620 kilobytes
2886------------------------------------------------
2887
2888The first number is the number of objects which are kept in
2889individual files. The second is the amount of space taken up by
2890those "loose" objects.
2891
2892You can save space and make git faster by moving these loose objects in
2893to a "pack file", which stores a group of objects in an efficient
2894compressed format; the details of how pack files are formatted can be
2895found in link:technical/pack-format.txt[technical/pack-format.txt].
2896
2897To put the loose objects into a pack, just run git repack:
2898
2899------------------------------------------------
2900$ git repack
2901Generating pack...
2902Done counting 6020 objects.
2903Deltifying 6020 objects.
2904 100% (6020/6020) done
2905Writing 6020 objects.
2906 100% (6020/6020) done
2907Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
2908Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
2909------------------------------------------------
2910
2911You can then run
2912
2913------------------------------------------------
2914$ git prune
2915------------------------------------------------
2916
2917to remove any of the "loose" objects that are now contained in the
2918pack. This will also remove any unreferenced objects (which may be
2919created when, for example, you use "git reset" to remove a commit).
2920You can verify that the loose objects are gone by looking at the
2921.git/objects directory or by running
2922
2923------------------------------------------------
2924$ git count-objects
29250 objects, 0 kilobytes
2926------------------------------------------------
2927
2928Although the object files are gone, any commands that refer to those
2929objects will work exactly as they did before.
2930
2931The gitlink:git-gc[1] command performs packing, pruning, and more for
2932you, so is normally the only high-level command you need.
d19fbc3c 2933
21dcb3b7
BF
2934[[dangling-objects]]
2935Dangling objects
61b41790 2936----------------
21dcb3b7 2937
04e50e94 2938The gitlink:git-fsck[1] command will sometimes complain about dangling
21dcb3b7
BF
2939objects. They are not a problem.
2940
1191ee18
BF
2941The most common cause of dangling objects is that you've rebased a
2942branch, or you have pulled from somebody else who rebased a branch--see
2943<<cleaning-up-history>>. In that case, the old head of the original
2944branch still exists, as does obviously everything it pointed to. The
2945branch pointer itself just doesn't, since you replaced it with another
2946one.
2947
2948There are also other situations too that cause dangling objects. For
2949example, a "dangling blob" may arise because you did a "git add" of a
2950file, but then, before you actually committed it and made it part of the
2951bigger picture, you changed something else in that file and committed
2952that *updated* thing - the old state that you added originally ends up
2953not being pointed to by any commit or tree, so it's now a dangling blob
2954object.
2955
2956Similarly, when the "recursive" merge strategy runs, and finds that
2957there are criss-cross merges and thus more than one merge base (which is
2958fairly unusual, but it does happen), it will generate one temporary
2959midway tree (or possibly even more, if you had lots of criss-crossing
2960merges and more than two merge bases) as a temporary internal merge
2961base, and again, those are real objects, but the end result will not end
2962up pointing to them, so they end up "dangling" in your repository.
2963
2964Generally, dangling objects aren't anything to worry about. They can
2965even be very useful: if you screw something up, the dangling objects can
2966be how you recover your old tree (say, you did a rebase, and realized
2967that you really didn't want to - you can look at what dangling objects
2968you have, and decide to reset your head to some old dangling state).
21dcb3b7 2969
559e4d7a
BF
2970For commits, the most useful thing to do with dangling objects tends to
2971be to do a simple
21dcb3b7
BF
2972
2973------------------------------------------------
2974$ gitk <dangling-commit-sha-goes-here> --not --all
2975------------------------------------------------
2976
1191ee18
BF
2977For blobs and trees, you can't do the same, but you can examine them.
2978You can just do
21dcb3b7
BF
2979
2980------------------------------------------------
2981$ git show <dangling-blob/tree-sha-goes-here>
2982------------------------------------------------
2983
1191ee18
BF
2984to show what the contents of the blob were (or, for a tree, basically
2985what the "ls" for that directory was), and that may give you some idea
2986of what the operation was that left that dangling object.
21dcb3b7 2987
1191ee18
BF
2988Usually, dangling blobs and trees aren't very interesting. They're
2989almost always the result of either being a half-way mergebase (the blob
2990will often even have the conflict markers from a merge in it, if you
2991have had conflicting merges that you fixed up by hand), or simply
2992because you interrupted a "git fetch" with ^C or something like that,
2993leaving _some_ of the new objects in the object database, but just
2994dangling and useless.
21dcb3b7
BF
2995
2996Anyway, once you are sure that you're not interested in any dangling
2997state, you can just prune all unreachable objects:
2998
2999------------------------------------------------
3000$ git prune
3001------------------------------------------------
3002
1191ee18
BF
3003and they'll be gone. But you should only run "git prune" on a quiescent
3004repository - it's kind of like doing a filesystem fsck recovery: you
3005don't want to do that while the filesystem is mounted.
21dcb3b7 3006
04e50e94
BF
3007(The same is true of "git-fsck" itself, btw - but since
3008git-fsck never actually *changes* the repository, it just reports
3009on what it found, git-fsck itself is never "dangerous" to run.
21dcb3b7
BF
3010Running it while somebody is actually changing the repository can cause
3011confusing and scary messages, but it won't actually do anything bad. In
3012contrast, running "git prune" while somebody is actively changing the
3013repository is a *BAD* idea).
3014
d19fbc3c
BF
3015include::glossary.txt[]
3016
6bd9b682
BF
3017Notes and todo list for this manual
3018===================================
3019
3020This is a work in progress.
3021
3022The basic requirements:
2f99710c
BF
3023 - It must be readable in order, from beginning to end, by
3024 someone intelligent with a basic grasp of the unix
3025 commandline, but without any special knowledge of git. If
3026 necessary, any other prerequisites should be specifically
3027 mentioned as they arise.
3028 - Whenever possible, section headings should clearly describe
3029 the task they explain how to do, in language that requires
3030 no more knowledge than necessary: for example, "importing
3031 patches into a project" rather than "the git-am command"
6bd9b682 3032
d5cd5de4
BF
3033Think about how to create a clear chapter dependency graph that will
3034allow people to get to important topics without necessarily reading
3035everything in between.
d19fbc3c 3036
aacd404e
MC
3037Say something about .gitignore.
3038
d19fbc3c
BF
3039Scan Documentation/ for other stuff left out; in particular:
3040 howto's
d19fbc3c
BF
3041 some of technical/?
3042 hooks
0b375ab0 3043 list of commands in gitlink:git[1]
d19fbc3c
BF
3044
3045Scan email archives for other stuff left out
3046
3047Scan man pages to see if any assume more background than this manual
3048provides.
3049
2f99710c 3050Simplify beginning by suggesting disconnected head instead of
b181d57f 3051temporary branch creation?
d19fbc3c 3052
2f99710c
BF
3053Add more good examples. Entire sections of just cookbook examples
3054might be a good idea; maybe make an "advanced examples" section a
3055standard end-of-chapter section?
d19fbc3c
BF
3056
3057Include cross-references to the glossary, where appropriate.
3058
9a241220
BF
3059Document shallow clones? See draft 1.5.0 release notes for some
3060documentation.
3061
3dff5379 3062Add a section on working with other version control systems, including
9a241220
BF
3063CVS, Subversion, and just imports of series of release tarballs.
3064
a8cd1402 3065More details on gitweb?
0b375ab0
BF
3066
3067Write a chapter on using plumbing and writing scripts.