treewide: correct several "up-to-date" to "up to date"
[git/git.git] / Documentation / gitcore-tutorial.txt
1 gitcore-tutorial(7)
2 ===================
3
4 NAME
5 ----
6 gitcore-tutorial - A Git core tutorial for developers
7
8 SYNOPSIS
9 --------
10 git *
11
12 DESCRIPTION
13 -----------
14
15 This tutorial explains how to use the "core" Git commands to set up and
16 work with a Git repository.
17
18 If you just need to use Git as a revision control system you may prefer
19 to start with "A Tutorial Introduction to Git" (linkgit:gittutorial[7]) or
20 link:user-manual.html[the Git User Manual].
21
22 However, an understanding of these low-level tools can be helpful if
23 you want to understand Git's internals.
24
25 The core Git is often called "plumbing", with the prettier user
26 interfaces on top of it called "porcelain". You may not want to use the
27 plumbing directly very often, but it can be good to know what the
28 plumbing does when the porcelain isn't flushing.
29
30 Back when this document was originally written, many porcelain
31 commands were shell scripts. For simplicity, it still uses them as
32 examples to illustrate how plumbing is fit together to form the
33 porcelain commands. The source tree includes some of these scripts in
34 contrib/examples/ for reference. Although these are not implemented as
35 shell scripts anymore, the description of what the plumbing layer
36 commands do is still valid.
37
38 [NOTE]
39 Deeper technical details are often marked as Notes, which you can
40 skip on your first reading.
41
42
43 Creating a Git repository
44 -------------------------
45
46 Creating a new Git repository couldn't be easier: all Git repositories start
47 out empty, and the only thing you need to do is find yourself a
48 subdirectory that you want to use as a working tree - either an empty
49 one for a totally new project, or an existing working tree that you want
50 to import into Git.
51
52 For our first example, we're going to start a totally new repository from
53 scratch, with no pre-existing files, and we'll call it 'git-tutorial'.
54 To start up, create a subdirectory for it, change into that
55 subdirectory, and initialize the Git infrastructure with 'git init':
56
57 ------------------------------------------------
58 $ mkdir git-tutorial
59 $ cd git-tutorial
60 $ git init
61 ------------------------------------------------
62
63 to which Git will reply
64
65 ----------------
66 Initialized empty Git repository in .git/
67 ----------------
68
69 which is just Git's way of saying that you haven't been doing anything
70 strange, and that it will have created a local `.git` directory setup for
71 your new project. You will now have a `.git` directory, and you can
72 inspect that with 'ls'. For your new empty project, it should show you
73 three entries, among other things:
74
75 - a file called `HEAD`, that has `ref: refs/heads/master` in it.
76 This is similar to a symbolic link and points at
77 `refs/heads/master` relative to the `HEAD` file.
78 +
79 Don't worry about the fact that the file that the `HEAD` link points to
80 doesn't even exist yet -- you haven't created the commit that will
81 start your `HEAD` development branch yet.
82
83 - a subdirectory called `objects`, which will contain all the
84 objects of your project. You should never have any real reason to
85 look at the objects directly, but you might want to know that these
86 objects are what contains all the real 'data' in your repository.
87
88 - a subdirectory called `refs`, which contains references to objects.
89
90 In particular, the `refs` subdirectory will contain two other
91 subdirectories, named `heads` and `tags` respectively. They do
92 exactly what their names imply: they contain references to any number
93 of different 'heads' of development (aka 'branches'), and to any
94 'tags' that you have created to name specific versions in your
95 repository.
96
97 One note: the special `master` head is the default branch, which is
98 why the `.git/HEAD` file was created points to it even if it
99 doesn't yet exist. Basically, the `HEAD` link is supposed to always
100 point to the branch you are working on right now, and you always
101 start out expecting to work on the `master` branch.
102
103 However, this is only a convention, and you can name your branches
104 anything you want, and don't have to ever even 'have' a `master`
105 branch. A number of the Git tools will assume that `.git/HEAD` is
106 valid, though.
107
108 [NOTE]
109 An 'object' is identified by its 160-bit SHA-1 hash, aka 'object name',
110 and a reference to an object is always the 40-byte hex
111 representation of that SHA-1 name. The files in the `refs`
112 subdirectory are expected to contain these hex references
113 (usually with a final `\n` at the end), and you should thus
114 expect to see a number of 41-byte files containing these
115 references in these `refs` subdirectories when you actually start
116 populating your tree.
117
118 [NOTE]
119 An advanced user may want to take a look at linkgit:gitrepository-layout[5]
120 after finishing this tutorial.
121
122 You have now created your first Git repository. Of course, since it's
123 empty, that's not very useful, so let's start populating it with data.
124
125
126 Populating a Git repository
127 ---------------------------
128
129 We'll keep this simple and stupid, so we'll start off with populating a
130 few trivial files just to get a feel for it.
131
132 Start off with just creating any random files that you want to maintain
133 in your Git repository. We'll start off with a few bad examples, just to
134 get a feel for how this works:
135
136 ------------------------------------------------
137 $ echo "Hello World" >hello
138 $ echo "Silly example" >example
139 ------------------------------------------------
140
141 you have now created two files in your working tree (aka 'working directory'),
142 but to actually check in your hard work, you will have to go through two steps:
143
144 - fill in the 'index' file (aka 'cache') with the information about your
145 working tree state.
146
147 - commit that index file as an object.
148
149 The first step is trivial: when you want to tell Git about any changes
150 to your working tree, you use the 'git update-index' program. That
151 program normally just takes a list of filenames you want to update, but
152 to avoid trivial mistakes, it refuses to add new entries to the index
153 (or remove existing ones) unless you explicitly tell it that you're
154 adding a new entry with the `--add` flag (or removing an entry with the
155 `--remove`) flag.
156
157 So to populate the index with the two files you just created, you can do
158
159 ------------------------------------------------
160 $ git update-index --add hello example
161 ------------------------------------------------
162
163 and you have now told Git to track those two files.
164
165 In fact, as you did that, if you now look into your object directory,
166 you'll notice that Git will have added two new objects to the object
167 database. If you did exactly the steps above, you should now be able to do
168
169
170 ----------------
171 $ ls .git/objects/??/*
172 ----------------
173
174 and see two files:
175
176 ----------------
177 .git/objects/55/7db03de997c86a4a028e1ebd3a1ceb225be238
178 .git/objects/f2/4c74a2e500f5ee1332c86b94199f52b1d1d962
179 ----------------
180
181 which correspond with the objects with names of `557db...` and
182 `f24c7...` respectively.
183
184 If you want to, you can use 'git cat-file' to look at those objects, but
185 you'll have to use the object name, not the filename of the object:
186
187 ----------------
188 $ git cat-file -t 557db03de997c86a4a028e1ebd3a1ceb225be238
189 ----------------
190
191 where the `-t` tells 'git cat-file' to tell you what the "type" of the
192 object is. Git will tell you that you have a "blob" object (i.e., just a
193 regular file), and you can see the contents with
194
195 ----------------
196 $ git cat-file blob 557db03
197 ----------------
198
199 which will print out "Hello World". The object `557db03` is nothing
200 more than the contents of your file `hello`.
201
202 [NOTE]
203 Don't confuse that object with the file `hello` itself. The
204 object is literally just those specific *contents* of the file, and
205 however much you later change the contents in file `hello`, the object
206 we just looked at will never change. Objects are immutable.
207
208 [NOTE]
209 The second example demonstrates that you can
210 abbreviate the object name to only the first several
211 hexadecimal digits in most places.
212
213 Anyway, as we mentioned previously, you normally never actually take a
214 look at the objects themselves, and typing long 40-character hex
215 names is not something you'd normally want to do. The above digression
216 was just to show that 'git update-index' did something magical, and
217 actually saved away the contents of your files into the Git object
218 database.
219
220 Updating the index did something else too: it created a `.git/index`
221 file. This is the index that describes your current working tree, and
222 something you should be very aware of. Again, you normally never worry
223 about the index file itself, but you should be aware of the fact that
224 you have not actually really "checked in" your files into Git so far,
225 you've only *told* Git about them.
226
227 However, since Git knows about them, you can now start using some of the
228 most basic Git commands to manipulate the files or look at their status.
229
230 In particular, let's not even check in the two files into Git yet, we'll
231 start off by adding another line to `hello` first:
232
233 ------------------------------------------------
234 $ echo "It's a new day for git" >>hello
235 ------------------------------------------------
236
237 and you can now, since you told Git about the previous state of `hello`, ask
238 Git what has changed in the tree compared to your old index, using the
239 'git diff-files' command:
240
241 ------------
242 $ git diff-files
243 ------------
244
245 Oops. That wasn't very readable. It just spit out its own internal
246 version of a 'diff', but that internal version really just tells you
247 that it has noticed that "hello" has been modified, and that the old object
248 contents it had have been replaced with something else.
249
250 To make it readable, we can tell 'git diff-files' to output the
251 differences as a patch, using the `-p` flag:
252
253 ------------
254 $ git diff-files -p
255 diff --git a/hello b/hello
256 index 557db03..263414f 100644
257 --- a/hello
258 +++ b/hello
259 @@ -1 +1,2 @@
260 Hello World
261 +It's a new day for git
262 ------------
263
264 i.e. the diff of the change we caused by adding another line to `hello`.
265
266 In other words, 'git diff-files' always shows us the difference between
267 what is recorded in the index, and what is currently in the working
268 tree. That's very useful.
269
270 A common shorthand for `git diff-files -p` is to just write `git
271 diff`, which will do the same thing.
272
273 ------------
274 $ git diff
275 diff --git a/hello b/hello
276 index 557db03..263414f 100644
277 --- a/hello
278 +++ b/hello
279 @@ -1 +1,2 @@
280 Hello World
281 +It's a new day for git
282 ------------
283
284
285 Committing Git state
286 --------------------
287
288 Now, we want to go to the next stage in Git, which is to take the files
289 that Git knows about in the index, and commit them as a real tree. We do
290 that in two phases: creating a 'tree' object, and committing that 'tree'
291 object as a 'commit' object together with an explanation of what the
292 tree was all about, along with information of how we came to that state.
293
294 Creating a tree object is trivial, and is done with 'git write-tree'.
295 There are no options or other input: `git write-tree` will take the
296 current index state, and write an object that describes that whole
297 index. In other words, we're now tying together all the different
298 filenames with their contents (and their permissions), and we're
299 creating the equivalent of a Git "directory" object:
300
301 ------------------------------------------------
302 $ git write-tree
303 ------------------------------------------------
304
305 and this will just output the name of the resulting tree, in this case
306 (if you have done exactly as I've described) it should be
307
308 ----------------
309 8988da15d077d4829fc51d8544c097def6644dbb
310 ----------------
311
312 which is another incomprehensible object name. Again, if you want to,
313 you can use `git cat-file -t 8988d...` to see that this time the object
314 is not a "blob" object, but a "tree" object (you can also use
315 `git cat-file` to actually output the raw object contents, but you'll see
316 mainly a binary mess, so that's less interesting).
317
318 However -- normally you'd never use 'git write-tree' on its own, because
319 normally you always commit a tree into a commit object using the
320 'git commit-tree' command. In fact, it's easier to not actually use
321 'git write-tree' on its own at all, but to just pass its result in as an
322 argument to 'git commit-tree'.
323
324 'git commit-tree' normally takes several arguments -- it wants to know
325 what the 'parent' of a commit was, but since this is the first commit
326 ever in this new repository, and it has no parents, we only need to pass in
327 the object name of the tree. However, 'git commit-tree' also wants to get a
328 commit message on its standard input, and it will write out the resulting
329 object name for the commit to its standard output.
330
331 And this is where we create the `.git/refs/heads/master` file
332 which is pointed at by `HEAD`. This file is supposed to contain
333 the reference to the top-of-tree of the master branch, and since
334 that's exactly what 'git commit-tree' spits out, we can do this
335 all with a sequence of simple shell commands:
336
337 ------------------------------------------------
338 $ tree=$(git write-tree)
339 $ commit=$(echo 'Initial commit' | git commit-tree $tree)
340 $ git update-ref HEAD $commit
341 ------------------------------------------------
342
343 In this case this creates a totally new commit that is not related to
344 anything else. Normally you do this only *once* for a project ever, and
345 all later commits will be parented on top of an earlier commit.
346
347 Again, normally you'd never actually do this by hand. There is a
348 helpful script called `git commit` that will do all of this for you. So
349 you could have just written `git commit`
350 instead, and it would have done the above magic scripting for you.
351
352
353 Making a change
354 ---------------
355
356 Remember how we did the 'git update-index' on file `hello` and then we
357 changed `hello` afterward, and could compare the new state of `hello` with the
358 state we saved in the index file?
359
360 Further, remember how I said that 'git write-tree' writes the contents
361 of the *index* file to the tree, and thus what we just committed was in
362 fact the *original* contents of the file `hello`, not the new ones. We did
363 that on purpose, to show the difference between the index state, and the
364 state in the working tree, and how they don't have to match, even
365 when we commit things.
366
367 As before, if we do `git diff-files -p` in our git-tutorial project,
368 we'll still see the same difference we saw last time: the index file
369 hasn't changed by the act of committing anything. However, now that we
370 have committed something, we can also learn to use a new command:
371 'git diff-index'.
372
373 Unlike 'git diff-files', which showed the difference between the index
374 file and the working tree, 'git diff-index' shows the differences
375 between a committed *tree* and either the index file or the working
376 tree. In other words, 'git diff-index' wants a tree to be diffed
377 against, and before we did the commit, we couldn't do that, because we
378 didn't have anything to diff against.
379
380 But now we can do
381
382 ----------------
383 $ git diff-index -p HEAD
384 ----------------
385
386 (where `-p` has the same meaning as it did in 'git diff-files'), and it
387 will show us the same difference, but for a totally different reason.
388 Now we're comparing the working tree not against the index file,
389 but against the tree we just wrote. It just so happens that those two
390 are obviously the same, so we get the same result.
391
392 Again, because this is a common operation, you can also just shorthand
393 it with
394
395 ----------------
396 $ git diff HEAD
397 ----------------
398
399 which ends up doing the above for you.
400
401 In other words, 'git diff-index' normally compares a tree against the
402 working tree, but when given the `--cached` flag, it is told to
403 instead compare against just the index cache contents, and ignore the
404 current working tree state entirely. Since we just wrote the index
405 file to HEAD, doing `git diff-index --cached -p HEAD` should thus return
406 an empty set of differences, and that's exactly what it does.
407
408 [NOTE]
409 ================
410 'git diff-index' really always uses the index for its
411 comparisons, and saying that it compares a tree against the working
412 tree is thus not strictly accurate. In particular, the list of
413 files to compare (the "meta-data") *always* comes from the index file,
414 regardless of whether the `--cached` flag is used or not. The `--cached`
415 flag really only determines whether the file *contents* to be compared
416 come from the working tree or not.
417
418 This is not hard to understand, as soon as you realize that Git simply
419 never knows (or cares) about files that it is not told about
420 explicitly. Git will never go *looking* for files to compare, it
421 expects you to tell it what the files are, and that's what the index
422 is there for.
423 ================
424
425 However, our next step is to commit the *change* we did, and again, to
426 understand what's going on, keep in mind the difference between "working
427 tree contents", "index file" and "committed tree". We have changes
428 in the working tree that we want to commit, and we always have to
429 work through the index file, so the first thing we need to do is to
430 update the index cache:
431
432 ------------------------------------------------
433 $ git update-index hello
434 ------------------------------------------------
435
436 (note how we didn't need the `--add` flag this time, since Git knew
437 about the file already).
438
439 Note what happens to the different 'git diff-{asterisk}' versions here.
440 After we've updated `hello` in the index, `git diff-files -p` now shows no
441 differences, but `git diff-index -p HEAD` still *does* show that the
442 current state is different from the state we committed. In fact, now
443 'git diff-index' shows the same difference whether we use the `--cached`
444 flag or not, since now the index is coherent with the working tree.
445
446 Now, since we've updated `hello` in the index, we can commit the new
447 version. We could do it by writing the tree by hand again, and
448 committing the tree (this time we'd have to use the `-p HEAD` flag to
449 tell commit that the HEAD was the *parent* of the new commit, and that
450 this wasn't an initial commit any more), but you've done that once
451 already, so let's just use the helpful script this time:
452
453 ------------------------------------------------
454 $ git commit
455 ------------------------------------------------
456
457 which starts an editor for you to write the commit message and tells you
458 a bit about what you have done.
459
460 Write whatever message you want, and all the lines that start with '#'
461 will be pruned out, and the rest will be used as the commit message for
462 the change. If you decide you don't want to commit anything after all at
463 this point (you can continue to edit things and update the index), you
464 can just leave an empty message. Otherwise `git commit` will commit
465 the change for you.
466
467 You've now made your first real Git commit. And if you're interested in
468 looking at what `git commit` really does, feel free to investigate:
469 it's a few very simple shell scripts to generate the helpful (?) commit
470 message headers, and a few one-liners that actually do the
471 commit itself ('git commit').
472
473
474 Inspecting Changes
475 ------------------
476
477 While creating changes is useful, it's even more useful if you can tell
478 later what changed. The most useful command for this is another of the
479 'diff' family, namely 'git diff-tree'.
480
481 'git diff-tree' can be given two arbitrary trees, and it will tell you the
482 differences between them. Perhaps even more commonly, though, you can
483 give it just a single commit object, and it will figure out the parent
484 of that commit itself, and show the difference directly. Thus, to get
485 the same diff that we've already seen several times, we can now do
486
487 ----------------
488 $ git diff-tree -p HEAD
489 ----------------
490
491 (again, `-p` means to show the difference as a human-readable patch),
492 and it will show what the last commit (in `HEAD`) actually changed.
493
494 [NOTE]
495 ============
496 Here is an ASCII art by Jon Loeliger that illustrates how
497 various 'diff-{asterisk}' commands compare things.
498
499 diff-tree
500 +----+
501 | |
502 | |
503 V V
504 +-----------+
505 | Object DB |
506 | Backing |
507 | Store |
508 +-----------+
509 ^ ^
510 | |
511 | | diff-index --cached
512 | |
513 diff-index | V
514 | +-----------+
515 | | Index |
516 | | "cache" |
517 | +-----------+
518 | ^
519 | |
520 | | diff-files
521 | |
522 V V
523 +-----------+
524 | Working |
525 | Directory |
526 +-----------+
527 ============
528
529 More interestingly, you can also give 'git diff-tree' the `--pretty` flag,
530 which tells it to also show the commit message and author and date of the
531 commit, and you can tell it to show a whole series of diffs.
532 Alternatively, you can tell it to be "silent", and not show the diffs at
533 all, but just show the actual commit message.
534
535 In fact, together with the 'git rev-list' program (which generates a
536 list of revisions), 'git diff-tree' ends up being a veritable fount of
537 changes. You can emulate `git log`, `git log -p`, etc. with a trivial
538 script that pipes the output of `git rev-list` to `git diff-tree --stdin`,
539 which was exactly how early versions of `git log` were implemented.
540
541
542 Tagging a version
543 -----------------
544
545 In Git, there are two kinds of tags, a "light" one, and an "annotated tag".
546
547 A "light" tag is technically nothing more than a branch, except we put
548 it in the `.git/refs/tags/` subdirectory instead of calling it a `head`.
549 So the simplest form of tag involves nothing more than
550
551 ------------------------------------------------
552 $ git tag my-first-tag
553 ------------------------------------------------
554
555 which just writes the current `HEAD` into the `.git/refs/tags/my-first-tag`
556 file, after which point you can then use this symbolic name for that
557 particular state. You can, for example, do
558
559 ----------------
560 $ git diff my-first-tag
561 ----------------
562
563 to diff your current state against that tag which at this point will
564 obviously be an empty diff, but if you continue to develop and commit
565 stuff, you can use your tag as an "anchor-point" to see what has changed
566 since you tagged it.
567
568 An "annotated tag" is actually a real Git object, and contains not only a
569 pointer to the state you want to tag, but also a small tag name and
570 message, along with optionally a PGP signature that says that yes,
571 you really did
572 that tag. You create these annotated tags with either the `-a` or
573 `-s` flag to 'git tag':
574
575 ----------------
576 $ git tag -s <tagname>
577 ----------------
578
579 which will sign the current `HEAD` (but you can also give it another
580 argument that specifies the thing to tag, e.g., you could have tagged the
581 current `mybranch` point by using `git tag <tagname> mybranch`).
582
583 You normally only do signed tags for major releases or things
584 like that, while the light-weight tags are useful for any marking you
585 want to do -- any time you decide that you want to remember a certain
586 point, just create a private tag for it, and you have a nice symbolic
587 name for the state at that point.
588
589
590 Copying repositories
591 --------------------
592
593 Git repositories are normally totally self-sufficient and relocatable.
594 Unlike CVS, for example, there is no separate notion of
595 "repository" and "working tree". A Git repository normally *is* the
596 working tree, with the local Git information hidden in the `.git`
597 subdirectory. There is nothing else. What you see is what you got.
598
599 [NOTE]
600 You can tell Git to split the Git internal information from
601 the directory that it tracks, but we'll ignore that for now: it's not
602 how normal projects work, and it's really only meant for special uses.
603 So the mental model of "the Git information is always tied directly to
604 the working tree that it describes" may not be technically 100%
605 accurate, but it's a good model for all normal use.
606
607 This has two implications:
608
609 - if you grow bored with the tutorial repository you created (or you've
610 made a mistake and want to start all over), you can just do simple
611 +
612 ----------------
613 $ rm -rf git-tutorial
614 ----------------
615 +
616 and it will be gone. There's no external repository, and there's no
617 history outside the project you created.
618
619 - if you want to move or duplicate a Git repository, you can do so. There
620 is 'git clone' command, but if all you want to do is just to
621 create a copy of your repository (with all the full history that
622 went along with it), you can do so with a regular
623 `cp -a git-tutorial new-git-tutorial`.
624 +
625 Note that when you've moved or copied a Git repository, your Git index
626 file (which caches various information, notably some of the "stat"
627 information for the files involved) will likely need to be refreshed.
628 So after you do a `cp -a` to create a new copy, you'll want to do
629 +
630 ----------------
631 $ git update-index --refresh
632 ----------------
633 +
634 in the new repository to make sure that the index file is up to date.
635
636 Note that the second point is true even across machines. You can
637 duplicate a remote Git repository with *any* regular copy mechanism, be it
638 'scp', 'rsync' or 'wget'.
639
640 When copying a remote repository, you'll want to at a minimum update the
641 index cache when you do this, and especially with other peoples'
642 repositories you often want to make sure that the index cache is in some
643 known state (you don't know *what* they've done and not yet checked in),
644 so usually you'll precede the 'git update-index' with a
645
646 ----------------
647 $ git read-tree --reset HEAD
648 $ git update-index --refresh
649 ----------------
650
651 which will force a total index re-build from the tree pointed to by `HEAD`.
652 It resets the index contents to `HEAD`, and then the 'git update-index'
653 makes sure to match up all index entries with the checked-out files.
654 If the original repository had uncommitted changes in its
655 working tree, `git update-index --refresh` notices them and
656 tells you they need to be updated.
657
658 The above can also be written as simply
659
660 ----------------
661 $ git reset
662 ----------------
663
664 and in fact a lot of the common Git command combinations can be scripted
665 with the `git xyz` interfaces. You can learn things by just looking
666 at what the various git scripts do. For example, `git reset` used to be
667 the above two lines implemented in 'git reset', but some things like
668 'git status' and 'git commit' are slightly more complex scripts around
669 the basic Git commands.
670
671 Many (most?) public remote repositories will not contain any of
672 the checked out files or even an index file, and will *only* contain the
673 actual core Git files. Such a repository usually doesn't even have the
674 `.git` subdirectory, but has all the Git files directly in the
675 repository.
676
677 To create your own local live copy of such a "raw" Git repository, you'd
678 first create your own subdirectory for the project, and then copy the
679 raw repository contents into the `.git` directory. For example, to
680 create your own copy of the Git repository, you'd do the following
681
682 ----------------
683 $ mkdir my-git
684 $ cd my-git
685 $ rsync -rL rsync://rsync.kernel.org/pub/scm/git/git.git/ .git
686 ----------------
687
688 followed by
689
690 ----------------
691 $ git read-tree HEAD
692 ----------------
693
694 to populate the index. However, now you have populated the index, and
695 you have all the Git internal files, but you will notice that you don't
696 actually have any of the working tree files to work on. To get
697 those, you'd check them out with
698
699 ----------------
700 $ git checkout-index -u -a
701 ----------------
702
703 where the `-u` flag means that you want the checkout to keep the index
704 up to date (so that you don't have to refresh it afterward), and the
705 `-a` flag means "check out all files" (if you have a stale copy or an
706 older version of a checked out tree you may also need to add the `-f`
707 flag first, to tell 'git checkout-index' to *force* overwriting of any old
708 files).
709
710 Again, this can all be simplified with
711
712 ----------------
713 $ git clone git://git.kernel.org/pub/scm/git/git.git/ my-git
714 $ cd my-git
715 $ git checkout
716 ----------------
717
718 which will end up doing all of the above for you.
719
720 You have now successfully copied somebody else's (mine) remote
721 repository, and checked it out.
722
723
724 Creating a new branch
725 ---------------------
726
727 Branches in Git are really nothing more than pointers into the Git
728 object database from within the `.git/refs/` subdirectory, and as we
729 already discussed, the `HEAD` branch is nothing but a symlink to one of
730 these object pointers.
731
732 You can at any time create a new branch by just picking an arbitrary
733 point in the project history, and just writing the SHA-1 name of that
734 object into a file under `.git/refs/heads/`. You can use any filename you
735 want (and indeed, subdirectories), but the convention is that the
736 "normal" branch is called `master`. That's just a convention, though,
737 and nothing enforces it.
738
739 To show that as an example, let's go back to the git-tutorial repository we
740 used earlier, and create a branch in it. You do that by simply just
741 saying that you want to check out a new branch:
742
743 ------------
744 $ git checkout -b mybranch
745 ------------
746
747 will create a new branch based at the current `HEAD` position, and switch
748 to it.
749
750 [NOTE]
751 ================================================
752 If you make the decision to start your new branch at some
753 other point in the history than the current `HEAD`, you can do so by
754 just telling 'git checkout' what the base of the checkout would be.
755 In other words, if you have an earlier tag or branch, you'd just do
756
757 ------------
758 $ git checkout -b mybranch earlier-commit
759 ------------
760
761 and it would create the new branch `mybranch` at the earlier commit,
762 and check out the state at that time.
763 ================================================
764
765 You can always just jump back to your original `master` branch by doing
766
767 ------------
768 $ git checkout master
769 ------------
770
771 (or any other branch-name, for that matter) and if you forget which
772 branch you happen to be on, a simple
773
774 ------------
775 $ cat .git/HEAD
776 ------------
777
778 will tell you where it's pointing. To get the list of branches
779 you have, you can say
780
781 ------------
782 $ git branch
783 ------------
784
785 which used to be nothing more than a simple script around `ls .git/refs/heads`.
786 There will be an asterisk in front of the branch you are currently on.
787
788 Sometimes you may wish to create a new branch _without_ actually
789 checking it out and switching to it. If so, just use the command
790
791 ------------
792 $ git branch <branchname> [startingpoint]
793 ------------
794
795 which will simply _create_ the branch, but will not do anything further.
796 You can then later -- once you decide that you want to actually develop
797 on that branch -- switch to that branch with a regular 'git checkout'
798 with the branchname as the argument.
799
800
801 Merging two branches
802 --------------------
803
804 One of the ideas of having a branch is that you do some (possibly
805 experimental) work in it, and eventually merge it back to the main
806 branch. So assuming you created the above `mybranch` that started out
807 being the same as the original `master` branch, let's make sure we're in
808 that branch, and do some work there.
809
810 ------------------------------------------------
811 $ git checkout mybranch
812 $ echo "Work, work, work" >>hello
813 $ git commit -m "Some work." -i hello
814 ------------------------------------------------
815
816 Here, we just added another line to `hello`, and we used a shorthand for
817 doing both `git update-index hello` and `git commit` by just giving the
818 filename directly to `git commit`, with an `-i` flag (it tells
819 Git to 'include' that file in addition to what you have done to
820 the index file so far when making the commit). The `-m` flag is to give the
821 commit log message from the command line.
822
823 Now, to make it a bit more interesting, let's assume that somebody else
824 does some work in the original branch, and simulate that by going back
825 to the master branch, and editing the same file differently there:
826
827 ------------
828 $ git checkout master
829 ------------
830
831 Here, take a moment to look at the contents of `hello`, and notice how they
832 don't contain the work we just did in `mybranch` -- because that work
833 hasn't happened in the `master` branch at all. Then do
834
835 ------------
836 $ echo "Play, play, play" >>hello
837 $ echo "Lots of fun" >>example
838 $ git commit -m "Some fun." -i hello example
839 ------------
840
841 since the master branch is obviously in a much better mood.
842
843 Now, you've got two branches, and you decide that you want to merge the
844 work done. Before we do that, let's introduce a cool graphical tool that
845 helps you view what's going on:
846
847 ----------------
848 $ gitk --all
849 ----------------
850
851 will show you graphically both of your branches (that's what the `--all`
852 means: normally it will just show you your current `HEAD`) and their
853 histories. You can also see exactly how they came to be from a common
854 source.
855
856 Anyway, let's exit 'gitk' (`^Q` or the File menu), and decide that we want
857 to merge the work we did on the `mybranch` branch into the `master`
858 branch (which is currently our `HEAD` too). To do that, there's a nice
859 script called 'git merge', which wants to know which branches you want
860 to resolve and what the merge is all about:
861
862 ------------
863 $ git merge -m "Merge work in mybranch" mybranch
864 ------------
865
866 where the first argument is going to be used as the commit message if
867 the merge can be resolved automatically.
868
869 Now, in this case we've intentionally created a situation where the
870 merge will need to be fixed up by hand, though, so Git will do as much
871 of it as it can automatically (which in this case is just merge the `example`
872 file, which had no differences in the `mybranch` branch), and say:
873
874 ----------------
875 Auto-merging hello
876 CONFLICT (content): Merge conflict in hello
877 Automatic merge failed; fix conflicts and then commit the result.
878 ----------------
879
880 It tells you that it did an "Automatic merge", which
881 failed due to conflicts in `hello`.
882
883 Not to worry. It left the (trivial) conflict in `hello` in the same form you
884 should already be well used to if you've ever used CVS, so let's just
885 open `hello` in our editor (whatever that may be), and fix it up somehow.
886 I'd suggest just making it so that `hello` contains all four lines:
887
888 ------------
889 Hello World
890 It's a new day for git
891 Play, play, play
892 Work, work, work
893 ------------
894
895 and once you're happy with your manual merge, just do a
896
897 ------------
898 $ git commit -i hello
899 ------------
900
901 which will very loudly warn you that you're now committing a merge
902 (which is correct, so never mind), and you can write a small merge
903 message about your adventures in 'git merge'-land.
904
905 After you're done, start up `gitk --all` to see graphically what the
906 history looks like. Notice that `mybranch` still exists, and you can
907 switch to it, and continue to work with it if you want to. The
908 `mybranch` branch will not contain the merge, but next time you merge it
909 from the `master` branch, Git will know how you merged it, so you'll not
910 have to do _that_ merge again.
911
912 Another useful tool, especially if you do not always work in X-Window
913 environment, is `git show-branch`.
914
915 ------------------------------------------------
916 $ git show-branch --topo-order --more=1 master mybranch
917 * [master] Merge work in mybranch
918 ! [mybranch] Some work.
919 --
920 - [master] Merge work in mybranch
921 *+ [mybranch] Some work.
922 * [master^] Some fun.
923 ------------------------------------------------
924
925 The first two lines indicate that it is showing the two branches
926 with the titles of their top-of-the-tree commits, you are currently on
927 `master` branch (notice the asterisk `*` character), and the first
928 column for the later output lines is used to show commits contained in the
929 `master` branch, and the second column for the `mybranch`
930 branch. Three commits are shown along with their titles.
931 All of them have non blank characters in the first column (`*`
932 shows an ordinary commit on the current branch, `-` is a merge commit), which
933 means they are now part of the `master` branch. Only the "Some
934 work" commit has the plus `+` character in the second column,
935 because `mybranch` has not been merged to incorporate these
936 commits from the master branch. The string inside brackets
937 before the commit log message is a short name you can use to
938 name the commit. In the above example, 'master' and 'mybranch'
939 are branch heads. 'master^' is the first parent of 'master'
940 branch head. Please see linkgit:gitrevisions[7] if you want to
941 see more complex cases.
942
943 [NOTE]
944 Without the '--more=1' option, 'git show-branch' would not output the
945 '[master^]' commit, as '[mybranch]' commit is a common ancestor of
946 both 'master' and 'mybranch' tips. Please see linkgit:git-show-branch[1]
947 for details.
948
949 [NOTE]
950 If there were more commits on the 'master' branch after the merge, the
951 merge commit itself would not be shown by 'git show-branch' by
952 default. You would need to provide `--sparse` option to make the
953 merge commit visible in this case.
954
955 Now, let's pretend you are the one who did all the work in
956 `mybranch`, and the fruit of your hard work has finally been merged
957 to the `master` branch. Let's go back to `mybranch`, and run
958 'git merge' to get the "upstream changes" back to your branch.
959
960 ------------
961 $ git checkout mybranch
962 $ git merge -m "Merge upstream changes." master
963 ------------
964
965 This outputs something like this (the actual commit object names
966 would be different)
967
968 ----------------
969 Updating from ae3a2da... to a80b4aa....
970 Fast-forward (no commit created; -m option ignored)
971 example | 1 +
972 hello | 1 +
973 2 files changed, 2 insertions(+)
974 ----------------
975
976 Because your branch did not contain anything more than what had
977 already been merged into the `master` branch, the merge operation did
978 not actually do a merge. Instead, it just updated the top of
979 the tree of your branch to that of the `master` branch. This is
980 often called 'fast-forward' merge.
981
982 You can run `gitk --all` again to see how the commit ancestry
983 looks like, or run 'show-branch', which tells you this.
984
985 ------------------------------------------------
986 $ git show-branch master mybranch
987 ! [master] Merge work in mybranch
988 * [mybranch] Merge work in mybranch
989 --
990 -- [master] Merge work in mybranch
991 ------------------------------------------------
992
993
994 Merging external work
995 ---------------------
996
997 It's usually much more common that you merge with somebody else than
998 merging with your own branches, so it's worth pointing out that Git
999 makes that very easy too, and in fact, it's not that different from
1000 doing a 'git merge'. In fact, a remote merge ends up being nothing
1001 more than "fetch the work from a remote repository into a temporary tag"
1002 followed by a 'git merge'.
1003
1004 Fetching from a remote repository is done by, unsurprisingly,
1005 'git fetch':
1006
1007 ----------------
1008 $ git fetch <remote-repository>
1009 ----------------
1010
1011 One of the following transports can be used to name the
1012 repository to download from:
1013
1014 SSH::
1015 `remote.machine:/path/to/repo.git/` or
1016 +
1017 `ssh://remote.machine/path/to/repo.git/`
1018 +
1019 This transport can be used for both uploading and downloading,
1020 and requires you to have a log-in privilege over `ssh` to the
1021 remote machine. It finds out the set of objects the other side
1022 lacks by exchanging the head commits both ends have and
1023 transfers (close to) minimum set of objects. It is by far the
1024 most efficient way to exchange Git objects between repositories.
1025
1026 Local directory::
1027 `/path/to/repo.git/`
1028 +
1029 This transport is the same as SSH transport but uses 'sh' to run
1030 both ends on the local machine instead of running other end on
1031 the remote machine via 'ssh'.
1032
1033 Git Native::
1034 `git://remote.machine/path/to/repo.git/`
1035 +
1036 This transport was designed for anonymous downloading. Like SSH
1037 transport, it finds out the set of objects the downstream side
1038 lacks and transfers (close to) minimum set of objects.
1039
1040 HTTP(S)::
1041 `http://remote.machine/path/to/repo.git/`
1042 +
1043 Downloader from http and https URL
1044 first obtains the topmost commit object name from the remote site
1045 by looking at the specified refname under `repo.git/refs/` directory,
1046 and then tries to obtain the
1047 commit object by downloading from `repo.git/objects/xx/xxx...`
1048 using the object name of that commit object. Then it reads the
1049 commit object to find out its parent commits and the associate
1050 tree object; it repeats this process until it gets all the
1051 necessary objects. Because of this behavior, they are
1052 sometimes also called 'commit walkers'.
1053 +
1054 The 'commit walkers' are sometimes also called 'dumb
1055 transports', because they do not require any Git aware smart
1056 server like Git Native transport does. Any stock HTTP server
1057 that does not even support directory index would suffice. But
1058 you must prepare your repository with 'git update-server-info'
1059 to help dumb transport downloaders.
1060
1061 Once you fetch from the remote repository, you `merge` that
1062 with your current branch.
1063
1064 However -- it's such a common thing to `fetch` and then
1065 immediately `merge`, that it's called `git pull`, and you can
1066 simply do
1067
1068 ----------------
1069 $ git pull <remote-repository>
1070 ----------------
1071
1072 and optionally give a branch-name for the remote end as a second
1073 argument.
1074
1075 [NOTE]
1076 You could do without using any branches at all, by
1077 keeping as many local repositories as you would like to have
1078 branches, and merging between them with 'git pull', just like
1079 you merge between branches. The advantage of this approach is
1080 that it lets you keep a set of files for each `branch` checked
1081 out and you may find it easier to switch back and forth if you
1082 juggle multiple lines of development simultaneously. Of
1083 course, you will pay the price of more disk usage to hold
1084 multiple working trees, but disk space is cheap these days.
1085
1086 It is likely that you will be pulling from the same remote
1087 repository from time to time. As a short hand, you can store
1088 the remote repository URL in the local repository's config file
1089 like this:
1090
1091 ------------------------------------------------
1092 $ git config remote.linus.url http://www.kernel.org/pub/scm/git/git.git/
1093 ------------------------------------------------
1094
1095 and use the "linus" keyword with 'git pull' instead of the full URL.
1096
1097 Examples.
1098
1099 . `git pull linus`
1100 . `git pull linus tag v0.99.1`
1101
1102 the above are equivalent to:
1103
1104 . `git pull http://www.kernel.org/pub/scm/git/git.git/ HEAD`
1105 . `git pull http://www.kernel.org/pub/scm/git/git.git/ tag v0.99.1`
1106
1107
1108 How does the merge work?
1109 ------------------------
1110
1111 We said this tutorial shows what plumbing does to help you cope
1112 with the porcelain that isn't flushing, but we so far did not
1113 talk about how the merge really works. If you are following
1114 this tutorial the first time, I'd suggest to skip to "Publishing
1115 your work" section and come back here later.
1116
1117 OK, still with me? To give us an example to look at, let's go
1118 back to the earlier repository with "hello" and "example" file,
1119 and bring ourselves back to the pre-merge state:
1120
1121 ------------
1122 $ git show-branch --more=2 master mybranch
1123 ! [master] Merge work in mybranch
1124 * [mybranch] Merge work in mybranch
1125 --
1126 -- [master] Merge work in mybranch
1127 +* [master^2] Some work.
1128 +* [master^] Some fun.
1129 ------------
1130
1131 Remember, before running 'git merge', our `master` head was at
1132 "Some fun." commit, while our `mybranch` head was at "Some
1133 work." commit.
1134
1135 ------------
1136 $ git checkout mybranch
1137 $ git reset --hard master^2
1138 $ git checkout master
1139 $ git reset --hard master^
1140 ------------
1141
1142 After rewinding, the commit structure should look like this:
1143
1144 ------------
1145 $ git show-branch
1146 * [master] Some fun.
1147 ! [mybranch] Some work.
1148 --
1149 * [master] Some fun.
1150 + [mybranch] Some work.
1151 *+ [master^] Initial commit
1152 ------------
1153
1154 Now we are ready to experiment with the merge by hand.
1155
1156 `git merge` command, when merging two branches, uses 3-way merge
1157 algorithm. First, it finds the common ancestor between them.
1158 The command it uses is 'git merge-base':
1159
1160 ------------
1161 $ mb=$(git merge-base HEAD mybranch)
1162 ------------
1163
1164 The command writes the commit object name of the common ancestor
1165 to the standard output, so we captured its output to a variable,
1166 because we will be using it in the next step. By the way, the common
1167 ancestor commit is the "Initial commit" commit in this case. You can
1168 tell it by:
1169
1170 ------------
1171 $ git name-rev --name-only --tags $mb
1172 my-first-tag
1173 ------------
1174
1175 After finding out a common ancestor commit, the second step is
1176 this:
1177
1178 ------------
1179 $ git read-tree -m -u $mb HEAD mybranch
1180 ------------
1181
1182 This is the same 'git read-tree' command we have already seen,
1183 but it takes three trees, unlike previous examples. This reads
1184 the contents of each tree into different 'stage' in the index
1185 file (the first tree goes to stage 1, the second to stage 2,
1186 etc.). After reading three trees into three stages, the paths
1187 that are the same in all three stages are 'collapsed' into stage
1188 0. Also paths that are the same in two of three stages are
1189 collapsed into stage 0, taking the SHA-1 from either stage 2 or
1190 stage 3, whichever is different from stage 1 (i.e. only one side
1191 changed from the common ancestor).
1192
1193 After 'collapsing' operation, paths that are different in three
1194 trees are left in non-zero stages. At this point, you can
1195 inspect the index file with this command:
1196
1197 ------------
1198 $ git ls-files --stage
1199 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example
1200 100644 557db03de997c86a4a028e1ebd3a1ceb225be238 1 hello
1201 100644 ba42a2a96e3027f3333e13ede4ccf4498c3ae942 2 hello
1202 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1203 ------------
1204
1205 In our example of only two files, we did not have unchanged
1206 files so only 'example' resulted in collapsing. But in real-life
1207 large projects, when only a small number of files change in one commit,
1208 this 'collapsing' tends to trivially merge most of the paths
1209 fairly quickly, leaving only a handful of real changes in non-zero
1210 stages.
1211
1212 To look at only non-zero stages, use `--unmerged` flag:
1213
1214 ------------
1215 $ git ls-files --unmerged
1216 100644 557db03de997c86a4a028e1ebd3a1ceb225be238 1 hello
1217 100644 ba42a2a96e3027f3333e13ede4ccf4498c3ae942 2 hello
1218 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1219 ------------
1220
1221 The next step of merging is to merge these three versions of the
1222 file, using 3-way merge. This is done by giving
1223 'git merge-one-file' command as one of the arguments to
1224 'git merge-index' command:
1225
1226 ------------
1227 $ git merge-index git-merge-one-file hello
1228 Auto-merging hello
1229 ERROR: Merge conflict in hello
1230 fatal: merge program failed
1231 ------------
1232
1233 'git merge-one-file' script is called with parameters to
1234 describe those three versions, and is responsible to leave the
1235 merge results in the working tree.
1236 It is a fairly straightforward shell script, and
1237 eventually calls 'merge' program from RCS suite to perform a
1238 file-level 3-way merge. In this case, 'merge' detects
1239 conflicts, and the merge result with conflict marks is left in
1240 the working tree.. This can be seen if you run `ls-files
1241 --stage` again at this point:
1242
1243 ------------
1244 $ git ls-files --stage
1245 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example
1246 100644 557db03de997c86a4a028e1ebd3a1ceb225be238 1 hello
1247 100644 ba42a2a96e3027f3333e13ede4ccf4498c3ae942 2 hello
1248 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1249 ------------
1250
1251 This is the state of the index file and the working file after
1252 'git merge' returns control back to you, leaving the conflicting
1253 merge for you to resolve. Notice that the path `hello` is still
1254 unmerged, and what you see with 'git diff' at this point is
1255 differences since stage 2 (i.e. your version).
1256
1257
1258 Publishing your work
1259 --------------------
1260
1261 So, we can use somebody else's work from a remote repository, but
1262 how can *you* prepare a repository to let other people pull from
1263 it?
1264
1265 You do your real work in your working tree that has your
1266 primary repository hanging under it as its `.git` subdirectory.
1267 You *could* make that repository accessible remotely and ask
1268 people to pull from it, but in practice that is not the way
1269 things are usually done. A recommended way is to have a public
1270 repository, make it reachable by other people, and when the
1271 changes you made in your primary working tree are in good shape,
1272 update the public repository from it. This is often called
1273 'pushing'.
1274
1275 [NOTE]
1276 This public repository could further be mirrored, and that is
1277 how Git repositories at `kernel.org` are managed.
1278
1279 Publishing the changes from your local (private) repository to
1280 your remote (public) repository requires a write privilege on
1281 the remote machine. You need to have an SSH account there to
1282 run a single command, 'git-receive-pack'.
1283
1284 First, you need to create an empty repository on the remote
1285 machine that will house your public repository. This empty
1286 repository will be populated and be kept up to date by pushing
1287 into it later. Obviously, this repository creation needs to be
1288 done only once.
1289
1290 [NOTE]
1291 'git push' uses a pair of commands,
1292 'git send-pack' on your local machine, and 'git-receive-pack'
1293 on the remote machine. The communication between the two over
1294 the network internally uses an SSH connection.
1295
1296 Your private repository's Git directory is usually `.git`, but
1297 your public repository is often named after the project name,
1298 i.e. `<project>.git`. Let's create such a public repository for
1299 project `my-git`. After logging into the remote machine, create
1300 an empty directory:
1301
1302 ------------
1303 $ mkdir my-git.git
1304 ------------
1305
1306 Then, make that directory into a Git repository by running
1307 'git init', but this time, since its name is not the usual
1308 `.git`, we do things slightly differently:
1309
1310 ------------
1311 $ GIT_DIR=my-git.git git init
1312 ------------
1313
1314 Make sure this directory is available for others you want your
1315 changes to be pulled via the transport of your choice. Also
1316 you need to make sure that you have the 'git-receive-pack'
1317 program on the `$PATH`.
1318
1319 [NOTE]
1320 Many installations of sshd do not invoke your shell as the login
1321 shell when you directly run programs; what this means is that if
1322 your login shell is 'bash', only `.bashrc` is read and not
1323 `.bash_profile`. As a workaround, make sure `.bashrc` sets up
1324 `$PATH` so that you can run 'git-receive-pack' program.
1325
1326 [NOTE]
1327 If you plan to publish this repository to be accessed over http,
1328 you should do `mv my-git.git/hooks/post-update.sample
1329 my-git.git/hooks/post-update` at this point.
1330 This makes sure that every time you push into this
1331 repository, `git update-server-info` is run.
1332
1333 Your "public repository" is now ready to accept your changes.
1334 Come back to the machine you have your private repository. From
1335 there, run this command:
1336
1337 ------------
1338 $ git push <public-host>:/path/to/my-git.git master
1339 ------------
1340
1341 This synchronizes your public repository to match the named
1342 branch head (i.e. `master` in this case) and objects reachable
1343 from them in your current repository.
1344
1345 As a real example, this is how I update my public Git
1346 repository. Kernel.org mirror network takes care of the
1347 propagation to other publicly visible machines:
1348
1349 ------------
1350 $ git push master.kernel.org:/pub/scm/git/git.git/
1351 ------------
1352
1353
1354 Packing your repository
1355 -----------------------
1356
1357 Earlier, we saw that one file under `.git/objects/??/` directory
1358 is stored for each Git object you create. This representation
1359 is efficient to create atomically and safely, but
1360 not so convenient to transport over the network. Since Git objects are
1361 immutable once they are created, there is a way to optimize the
1362 storage by "packing them together". The command
1363
1364 ------------
1365 $ git repack
1366 ------------
1367
1368 will do it for you. If you followed the tutorial examples, you
1369 would have accumulated about 17 objects in `.git/objects/??/`
1370 directories by now. 'git repack' tells you how many objects it
1371 packed, and stores the packed file in the `.git/objects/pack`
1372 directory.
1373
1374 [NOTE]
1375 You will see two files, `pack-*.pack` and `pack-*.idx`,
1376 in `.git/objects/pack` directory. They are closely related to
1377 each other, and if you ever copy them by hand to a different
1378 repository for whatever reason, you should make sure you copy
1379 them together. The former holds all the data from the objects
1380 in the pack, and the latter holds the index for random
1381 access.
1382
1383 If you are paranoid, running 'git verify-pack' command would
1384 detect if you have a corrupt pack, but do not worry too much.
1385 Our programs are always perfect ;-).
1386
1387 Once you have packed objects, you do not need to leave the
1388 unpacked objects that are contained in the pack file anymore.
1389
1390 ------------
1391 $ git prune-packed
1392 ------------
1393
1394 would remove them for you.
1395
1396 You can try running `find .git/objects -type f` before and after
1397 you run `git prune-packed` if you are curious. Also `git
1398 count-objects` would tell you how many unpacked objects are in
1399 your repository and how much space they are consuming.
1400
1401 [NOTE]
1402 `git pull` is slightly cumbersome for HTTP transport, as a
1403 packed repository may contain relatively few objects in a
1404 relatively large pack. If you expect many HTTP pulls from your
1405 public repository you might want to repack & prune often, or
1406 never.
1407
1408 If you run `git repack` again at this point, it will say
1409 "Nothing new to pack.". Once you continue your development and
1410 accumulate the changes, running `git repack` again will create a
1411 new pack, that contains objects created since you packed your
1412 repository the last time. We recommend that you pack your project
1413 soon after the initial import (unless you are starting your
1414 project from scratch), and then run `git repack` every once in a
1415 while, depending on how active your project is.
1416
1417 When a repository is synchronized via `git push` and `git pull`
1418 objects packed in the source repository are usually stored
1419 unpacked in the destination.
1420 While this allows you to use different packing strategies on
1421 both ends, it also means you may need to repack both
1422 repositories every once in a while.
1423
1424
1425 Working with Others
1426 -------------------
1427
1428 Although Git is a truly distributed system, it is often
1429 convenient to organize your project with an informal hierarchy
1430 of developers. Linux kernel development is run this way. There
1431 is a nice illustration (page 17, "Merges to Mainline") in
1432 https://web.archive.org/web/20120915203609/http://www.xenotime.net/linux/mentor/linux-mentoring-2006.pdf[Randy Dunlap's presentation].
1433
1434 It should be stressed that this hierarchy is purely *informal*.
1435 There is nothing fundamental in Git that enforces the "chain of
1436 patch flow" this hierarchy implies. You do not have to pull
1437 from only one remote repository.
1438
1439 A recommended workflow for a "project lead" goes like this:
1440
1441 1. Prepare your primary repository on your local machine. Your
1442 work is done there.
1443
1444 2. Prepare a public repository accessible to others.
1445 +
1446 If other people are pulling from your repository over dumb
1447 transport protocols (HTTP), you need to keep this repository
1448 'dumb transport friendly'. After `git init`,
1449 `$GIT_DIR/hooks/post-update.sample` copied from the standard templates
1450 would contain a call to 'git update-server-info'
1451 but you need to manually enable the hook with
1452 `mv post-update.sample post-update`. This makes sure
1453 'git update-server-info' keeps the necessary files up to date.
1454
1455 3. Push into the public repository from your primary
1456 repository.
1457
1458 4. 'git repack' the public repository. This establishes a big
1459 pack that contains the initial set of objects as the
1460 baseline, and possibly 'git prune' if the transport
1461 used for pulling from your repository supports packed
1462 repositories.
1463
1464 5. Keep working in your primary repository. Your changes
1465 include modifications of your own, patches you receive via
1466 e-mails, and merges resulting from pulling the "public"
1467 repositories of your "subsystem maintainers".
1468 +
1469 You can repack this private repository whenever you feel like.
1470
1471 6. Push your changes to the public repository, and announce it
1472 to the public.
1473
1474 7. Every once in a while, 'git repack' the public repository.
1475 Go back to step 5. and continue working.
1476
1477
1478 A recommended work cycle for a "subsystem maintainer" who works
1479 on that project and has an own "public repository" goes like this:
1480
1481 1. Prepare your work repository, by running 'git clone' on the public
1482 repository of the "project lead". The URL used for the
1483 initial cloning is stored in the remote.origin.url
1484 configuration variable.
1485
1486 2. Prepare a public repository accessible to others, just like
1487 the "project lead" person does.
1488
1489 3. Copy over the packed files from "project lead" public
1490 repository to your public repository, unless the "project
1491 lead" repository lives on the same machine as yours. In the
1492 latter case, you can use `objects/info/alternates` file to
1493 point at the repository you are borrowing from.
1494
1495 4. Push into the public repository from your primary
1496 repository. Run 'git repack', and possibly 'git prune' if the
1497 transport used for pulling from your repository supports
1498 packed repositories.
1499
1500 5. Keep working in your primary repository. Your changes
1501 include modifications of your own, patches you receive via
1502 e-mails, and merges resulting from pulling the "public"
1503 repositories of your "project lead" and possibly your
1504 "sub-subsystem maintainers".
1505 +
1506 You can repack this private repository whenever you feel
1507 like.
1508
1509 6. Push your changes to your public repository, and ask your
1510 "project lead" and possibly your "sub-subsystem
1511 maintainers" to pull from it.
1512
1513 7. Every once in a while, 'git repack' the public repository.
1514 Go back to step 5. and continue working.
1515
1516
1517 A recommended work cycle for an "individual developer" who does
1518 not have a "public" repository is somewhat different. It goes
1519 like this:
1520
1521 1. Prepare your work repository, by 'git clone' the public
1522 repository of the "project lead" (or a "subsystem
1523 maintainer", if you work on a subsystem). The URL used for
1524 the initial cloning is stored in the remote.origin.url
1525 configuration variable.
1526
1527 2. Do your work in your repository on 'master' branch.
1528
1529 3. Run `git fetch origin` from the public repository of your
1530 upstream every once in a while. This does only the first
1531 half of `git pull` but does not merge. The head of the
1532 public repository is stored in `.git/refs/remotes/origin/master`.
1533
1534 4. Use `git cherry origin` to see which ones of your patches
1535 were accepted, and/or use `git rebase origin` to port your
1536 unmerged changes forward to the updated upstream.
1537
1538 5. Use `git format-patch origin` to prepare patches for e-mail
1539 submission to your upstream and send it out. Go back to
1540 step 2. and continue.
1541
1542
1543 Working with Others, Shared Repository Style
1544 --------------------------------------------
1545
1546 If you are coming from a CVS background, the style of cooperation
1547 suggested in the previous section may be new to you. You do not
1548 have to worry. Git supports the "shared public repository" style of
1549 cooperation you are probably more familiar with as well.
1550
1551 See linkgit:gitcvs-migration[7] for the details.
1552
1553 Bundling your work together
1554 ---------------------------
1555
1556 It is likely that you will be working on more than one thing at
1557 a time. It is easy to manage those more-or-less independent tasks
1558 using branches with Git.
1559
1560 We have already seen how branches work previously,
1561 with "fun and work" example using two branches. The idea is the
1562 same if there are more than two branches. Let's say you started
1563 out from "master" head, and have some new code in the "master"
1564 branch, and two independent fixes in the "commit-fix" and
1565 "diff-fix" branches:
1566
1567 ------------
1568 $ git show-branch
1569 ! [commit-fix] Fix commit message normalization.
1570 ! [diff-fix] Fix rename detection.
1571 * [master] Release candidate #1
1572 ---
1573 + [diff-fix] Fix rename detection.
1574 + [diff-fix~1] Better common substring algorithm.
1575 + [commit-fix] Fix commit message normalization.
1576 * [master] Release candidate #1
1577 ++* [diff-fix~2] Pretty-print messages.
1578 ------------
1579
1580 Both fixes are tested well, and at this point, you want to merge
1581 in both of them. You could merge in 'diff-fix' first and then
1582 'commit-fix' next, like this:
1583
1584 ------------
1585 $ git merge -m "Merge fix in diff-fix" diff-fix
1586 $ git merge -m "Merge fix in commit-fix" commit-fix
1587 ------------
1588
1589 Which would result in:
1590
1591 ------------
1592 $ git show-branch
1593 ! [commit-fix] Fix commit message normalization.
1594 ! [diff-fix] Fix rename detection.
1595 * [master] Merge fix in commit-fix
1596 ---
1597 - [master] Merge fix in commit-fix
1598 + * [commit-fix] Fix commit message normalization.
1599 - [master~1] Merge fix in diff-fix
1600 +* [diff-fix] Fix rename detection.
1601 +* [diff-fix~1] Better common substring algorithm.
1602 * [master~2] Release candidate #1
1603 ++* [master~3] Pretty-print messages.
1604 ------------
1605
1606 However, there is no particular reason to merge in one branch
1607 first and the other next, when what you have are a set of truly
1608 independent changes (if the order mattered, then they are not
1609 independent by definition). You could instead merge those two
1610 branches into the current branch at once. First let's undo what
1611 we just did and start over. We would want to get the master
1612 branch before these two merges by resetting it to 'master~2':
1613
1614 ------------
1615 $ git reset --hard master~2
1616 ------------
1617
1618 You can make sure `git show-branch` matches the state before
1619 those two 'git merge' you just did. Then, instead of running
1620 two 'git merge' commands in a row, you would merge these two
1621 branch heads (this is known as 'making an Octopus'):
1622
1623 ------------
1624 $ git merge commit-fix diff-fix
1625 $ git show-branch
1626 ! [commit-fix] Fix commit message normalization.
1627 ! [diff-fix] Fix rename detection.
1628 * [master] Octopus merge of branches 'diff-fix' and 'commit-fix'
1629 ---
1630 - [master] Octopus merge of branches 'diff-fix' and 'commit-fix'
1631 + * [commit-fix] Fix commit message normalization.
1632 +* [diff-fix] Fix rename detection.
1633 +* [diff-fix~1] Better common substring algorithm.
1634 * [master~1] Release candidate #1
1635 ++* [master~2] Pretty-print messages.
1636 ------------
1637
1638 Note that you should not do Octopus just because you can. An octopus
1639 is a valid thing to do and often makes it easier to view the
1640 commit history if you are merging more than two independent
1641 changes at the same time. However, if you have merge conflicts
1642 with any of the branches you are merging in and need to hand
1643 resolve, that is an indication that the development happened in
1644 those branches were not independent after all, and you should
1645 merge two at a time, documenting how you resolved the conflicts,
1646 and the reason why you preferred changes made in one side over
1647 the other. Otherwise it would make the project history harder
1648 to follow, not easier.
1649
1650 SEE ALSO
1651 --------
1652 linkgit:gittutorial[7],
1653 linkgit:gittutorial-2[7],
1654 linkgit:gitcvs-migration[7],
1655 linkgit:git-help[1],
1656 linkgit:giteveryday[7],
1657 link:user-manual.html[The Git User's Manual]
1658
1659 GIT
1660 ---
1661 Part of the linkgit:git[1] suite