Merge branch 'maint'
[git/git.git] / Documentation / core-tutorial.txt
1 A git core tutorial for developers
2 ==================================
3
4 Introduction
5 ------------
6
7 This is trying to be a short tutorial on setting up and using a git
8 repository, mainly because being hands-on and using explicit examples is
9 often the best way of explaining what is going on.
10
11 In normal life, most people wouldn't use the "core" git programs
12 directly, but rather script around them to make them more palatable.
13 Understanding the core git stuff may help some people get those scripts
14 done, though, and it may also be instructive in helping people
15 understand what it is that the higher-level helper scripts are actually
16 doing.
17
18 The core git is often called "plumbing", with the prettier user
19 interfaces on top of it called "porcelain". You may not want to use the
20 plumbing directly very often, but it can be good to know what the
21 plumbing does for when the porcelain isn't flushing.
22
23 The material presented here often goes deep describing how things
24 work internally. If you are mostly interested in using git as a
25 SCM, you can skip them during your first pass.
26
27 [NOTE]
28 And those "too deep" descriptions are often marked as Note.
29
30 [NOTE]
31 If you are already familiar with another version control system,
32 like CVS, you may want to take a look at
33 link:everyday.html[Everyday GIT in 20 commands or so] first
34 before reading this.
35
36
37 Creating a git repository
38 -------------------------
39
40 Creating a new git repository couldn't be easier: all git repositories start
41 out empty, and the only thing you need to do is find yourself a
42 subdirectory that you want to use as a working tree - either an empty
43 one for a totally new project, or an existing working tree that you want
44 to import into git.
45
46 For our first example, we're going to start a totally new repository from
47 scratch, with no pre-existing files, and we'll call it `git-tutorial`.
48 To start up, create a subdirectory for it, change into that
49 subdirectory, and initialize the git infrastructure with `git-init-db`:
50
51 ------------------------------------------------
52 $ mkdir git-tutorial
53 $ cd git-tutorial
54 $ git-init-db
55 ------------------------------------------------
56
57 to which git will reply
58
59 ----------------
60 Initialized empty Git repository in .git/
61 ----------------
62
63 which is just git's way of saying that you haven't been doing anything
64 strange, and that it will have created a local `.git` directory setup for
65 your new project. You will now have a `.git` directory, and you can
66 inspect that with `ls`. For your new empty project, it should show you
67 three entries, among other things:
68
69 - a file called `HEAD`, that has `ref: refs/heads/master` in it.
70 This is similar to a symbolic link and points at
71 `refs/heads/master` relative to the `HEAD` file.
72 +
73 Don't worry about the fact that the file that the `HEAD` link points to
74 doesn't even exist yet -- you haven't created the commit that will
75 start your `HEAD` development branch yet.
76
77 - a subdirectory called `objects`, which will contain all the
78 objects of your project. You should never have any real reason to
79 look at the objects directly, but you might want to know that these
80 objects are what contains all the real 'data' in your repository.
81
82 - a subdirectory called `refs`, which contains references to objects.
83
84 In particular, the `refs` subdirectory will contain two other
85 subdirectories, named `heads` and `tags` respectively. They do
86 exactly what their names imply: they contain references to any number
87 of different 'heads' of development (aka 'branches'), and to any
88 'tags' that you have created to name specific versions in your
89 repository.
90
91 One note: the special `master` head is the default branch, which is
92 why the `.git/HEAD` file was created points to it even if it
93 doesn't yet exist. Basically, the `HEAD` link is supposed to always
94 point to the branch you are working on right now, and you always
95 start out expecting to work on the `master` branch.
96
97 However, this is only a convention, and you can name your branches
98 anything you want, and don't have to ever even 'have' a `master`
99 branch. A number of the git tools will assume that `.git/HEAD` is
100 valid, though.
101
102 [NOTE]
103 An 'object' is identified by its 160-bit SHA1 hash, aka 'object name',
104 and a reference to an object is always the 40-byte hex
105 representation of that SHA1 name. The files in the `refs`
106 subdirectory are expected to contain these hex references
107 (usually with a final `\'\n\'` at the end), and you should thus
108 expect to see a number of 41-byte files containing these
109 references in these `refs` subdirectories when you actually start
110 populating your tree.
111
112 [NOTE]
113 An advanced user may want to take a look at the
114 link:repository-layout.html[repository layout] document
115 after finishing this tutorial.
116
117 You have now created your first git repository. Of course, since it's
118 empty, that's not very useful, so let's start populating it with data.
119
120
121 Populating a git repository
122 ---------------------------
123
124 We'll keep this simple and stupid, so we'll start off with populating a
125 few trivial files just to get a feel for it.
126
127 Start off with just creating any random files that you want to maintain
128 in your git repository. We'll start off with a few bad examples, just to
129 get a feel for how this works:
130
131 ------------------------------------------------
132 $ echo "Hello World" >hello
133 $ echo "Silly example" >example
134 ------------------------------------------------
135
136 you have now created two files in your working tree (aka 'working directory'),
137 but to actually check in your hard work, you will have to go through two steps:
138
139 - fill in the 'index' file (aka 'cache') with the information about your
140 working tree state.
141
142 - commit that index file as an object.
143
144 The first step is trivial: when you want to tell git about any changes
145 to your working tree, you use the `git-update-index` program. That
146 program normally just takes a list of filenames you want to update, but
147 to avoid trivial mistakes, it refuses to add new entries to the index
148 (or remove existing ones) unless you explicitly tell it that you're
149 adding a new entry with the `\--add` flag (or removing an entry with the
150 `\--remove`) flag.
151
152 So to populate the index with the two files you just created, you can do
153
154 ------------------------------------------------
155 $ git-update-index --add hello example
156 ------------------------------------------------
157
158 and you have now told git to track those two files.
159
160 In fact, as you did that, if you now look into your object directory,
161 you'll notice that git will have added two new objects to the object
162 database. If you did exactly the steps above, you should now be able to do
163
164
165 ----------------
166 $ ls .git/objects/??/*
167 ----------------
168
169 and see two files:
170
171 ----------------
172 .git/objects/55/7db03de997c86a4a028e1ebd3a1ceb225be238
173 .git/objects/f2/4c74a2e500f5ee1332c86b94199f52b1d1d962
174 ----------------
175
176 which correspond with the objects with names of `557db...` and
177 `f24c7...` respectively.
178
179 If you want to, you can use `git-cat-file` to look at those objects, but
180 you'll have to use the object name, not the filename of the object:
181
182 ----------------
183 $ git-cat-file -t 557db03de997c86a4a028e1ebd3a1ceb225be238
184 ----------------
185
186 where the `-t` tells `git-cat-file` to tell you what the "type" of the
187 object is. git will tell you that you have a "blob" object (i.e., just a
188 regular file), and you can see the contents with
189
190 ----------------
191 $ git-cat-file "blob" 557db03
192 ----------------
193
194 which will print out "Hello World". The object `557db03` is nothing
195 more than the contents of your file `hello`.
196
197 [NOTE]
198 Don't confuse that object with the file `hello` itself. The
199 object is literally just those specific *contents* of the file, and
200 however much you later change the contents in file `hello`, the object
201 we just looked at will never change. Objects are immutable.
202
203 [NOTE]
204 The second example demonstrates that you can
205 abbreviate the object name to only the first several
206 hexadecimal digits in most places.
207
208 Anyway, as we mentioned previously, you normally never actually take a
209 look at the objects themselves, and typing long 40-character hex
210 names is not something you'd normally want to do. The above digression
211 was just to show that `git-update-index` did something magical, and
212 actually saved away the contents of your files into the git object
213 database.
214
215 Updating the index did something else too: it created a `.git/index`
216 file. This is the index that describes your current working tree, and
217 something you should be very aware of. Again, you normally never worry
218 about the index file itself, but you should be aware of the fact that
219 you have not actually really "checked in" your files into git so far,
220 you've only *told* git about them.
221
222 However, since git knows about them, you can now start using some of the
223 most basic git commands to manipulate the files or look at their status.
224
225 In particular, let's not even check in the two files into git yet, we'll
226 start off by adding another line to `hello` first:
227
228 ------------------------------------------------
229 $ echo "It's a new day for git" >>hello
230 ------------------------------------------------
231
232 and you can now, since you told git about the previous state of `hello`, ask
233 git what has changed in the tree compared to your old index, using the
234 `git-diff-files` command:
235
236 ------------
237 $ git-diff-files
238 ------------
239
240 Oops. That wasn't very readable. It just spit out its own internal
241 version of a `diff`, but that internal version really just tells you
242 that it has noticed that "hello" has been modified, and that the old object
243 contents it had have been replaced with something else.
244
245 To make it readable, we can tell git-diff-files to output the
246 differences as a patch, using the `-p` flag:
247
248 ------------
249 $ git-diff-files -p
250 diff --git a/hello b/hello
251 index 557db03..263414f 100644
252 --- a/hello
253 +++ b/hello
254 @@ -1 +1,2 @@
255 Hello World
256 +It's a new day for git
257 ----
258
259 i.e. the diff of the change we caused by adding another line to `hello`.
260
261 In other words, `git-diff-files` always shows us the difference between
262 what is recorded in the index, and what is currently in the working
263 tree. That's very useful.
264
265 A common shorthand for `git-diff-files -p` is to just write `git
266 diff`, which will do the same thing.
267
268 ------------
269 $ git diff
270 diff --git a/hello b/hello
271 index 557db03..263414f 100644
272 --- a/hello
273 +++ b/hello
274 @@ -1 +1,2 @@
275 Hello World
276 +It's a new day for git
277 ------------
278
279
280 Committing git state
281 --------------------
282
283 Now, we want to go to the next stage in git, which is to take the files
284 that git knows about in the index, and commit them as a real tree. We do
285 that in two phases: creating a 'tree' object, and committing that 'tree'
286 object as a 'commit' object together with an explanation of what the
287 tree was all about, along with information of how we came to that state.
288
289 Creating a tree object is trivial, and is done with `git-write-tree`.
290 There are no options or other input: git-write-tree will take the
291 current index state, and write an object that describes that whole
292 index. In other words, we're now tying together all the different
293 filenames with their contents (and their permissions), and we're
294 creating the equivalent of a git "directory" object:
295
296 ------------------------------------------------
297 $ git-write-tree
298 ------------------------------------------------
299
300 and this will just output the name of the resulting tree, in this case
301 (if you have done exactly as I've described) it should be
302
303 ----------------
304 8988da15d077d4829fc51d8544c097def6644dbb
305 ----------------
306
307 which is another incomprehensible object name. Again, if you want to,
308 you can use `git-cat-file -t 8988d\...` to see that this time the object
309 is not a "blob" object, but a "tree" object (you can also use
310 `git-cat-file` to actually output the raw object contents, but you'll see
311 mainly a binary mess, so that's less interesting).
312
313 However -- normally you'd never use `git-write-tree` on its own, because
314 normally you always commit a tree into a commit object using the
315 `git-commit-tree` command. In fact, it's easier to not actually use
316 `git-write-tree` on its own at all, but to just pass its result in as an
317 argument to `git-commit-tree`.
318
319 `git-commit-tree` normally takes several arguments -- it wants to know
320 what the 'parent' of a commit was, but since this is the first commit
321 ever in this new repository, and it has no parents, we only need to pass in
322 the object name of the tree. However, `git-commit-tree`
323 also wants to get a commit message
324 on its standard input, and it will write out the resulting object name for the
325 commit to its standard output.
326
327 And this is where we create the `.git/refs/heads/master` file
328 which is pointed at by `HEAD`. This file is supposed to contain
329 the reference to the top-of-tree of the master branch, and since
330 that's exactly what `git-commit-tree` spits out, we can do this
331 all with a sequence of simple shell commands:
332
333 ------------------------------------------------
334 $ tree=$(git-write-tree)
335 $ commit=$(echo 'Initial commit' | git-commit-tree $tree)
336 $ git-update-ref HEAD $commit
337 ------------------------------------------------
338
339 In this case this creates a totally new commit that is not related to
340 anything else. Normally you do this only *once* for a project ever, and
341 all later commits will be parented on top of an earlier commit.
342
343 Again, normally you'd never actually do this by hand. There is a
344 helpful script called `git commit` that will do all of this for you. So
345 you could have just written `git commit`
346 instead, and it would have done the above magic scripting for you.
347
348
349 Making a change
350 ---------------
351
352 Remember how we did the `git-update-index` on file `hello` and then we
353 changed `hello` afterward, and could compare the new state of `hello` with the
354 state we saved in the index file?
355
356 Further, remember how I said that `git-write-tree` writes the contents
357 of the *index* file to the tree, and thus what we just committed was in
358 fact the *original* contents of the file `hello`, not the new ones. We did
359 that on purpose, to show the difference between the index state, and the
360 state in the working tree, and how they don't have to match, even
361 when we commit things.
362
363 As before, if we do `git-diff-files -p` in our git-tutorial project,
364 we'll still see the same difference we saw last time: the index file
365 hasn't changed by the act of committing anything. However, now that we
366 have committed something, we can also learn to use a new command:
367 `git-diff-index`.
368
369 Unlike `git-diff-files`, which showed the difference between the index
370 file and the working tree, `git-diff-index` shows the differences
371 between a committed *tree* and either the index file or the working
372 tree. In other words, `git-diff-index` wants a tree to be diffed
373 against, and before we did the commit, we couldn't do that, because we
374 didn't have anything to diff against.
375
376 But now we can do
377
378 ----------------
379 $ git-diff-index -p HEAD
380 ----------------
381
382 (where `-p` has the same meaning as it did in `git-diff-files`), and it
383 will show us the same difference, but for a totally different reason.
384 Now we're comparing the working tree not against the index file,
385 but against the tree we just wrote. It just so happens that those two
386 are obviously the same, so we get the same result.
387
388 Again, because this is a common operation, you can also just shorthand
389 it with
390
391 ----------------
392 $ git diff HEAD
393 ----------------
394
395 which ends up doing the above for you.
396
397 In other words, `git-diff-index` normally compares a tree against the
398 working tree, but when given the `\--cached` flag, it is told to
399 instead compare against just the index cache contents, and ignore the
400 current working tree state entirely. Since we just wrote the index
401 file to HEAD, doing `git-diff-index \--cached -p HEAD` should thus return
402 an empty set of differences, and that's exactly what it does.
403
404 [NOTE]
405 ================
406 `git-diff-index` really always uses the index for its
407 comparisons, and saying that it compares a tree against the working
408 tree is thus not strictly accurate. In particular, the list of
409 files to compare (the "meta-data") *always* comes from the index file,
410 regardless of whether the `\--cached` flag is used or not. The `\--cached`
411 flag really only determines whether the file *contents* to be compared
412 come from the working tree or not.
413
414 This is not hard to understand, as soon as you realize that git simply
415 never knows (or cares) about files that it is not told about
416 explicitly. git will never go *looking* for files to compare, it
417 expects you to tell it what the files are, and that's what the index
418 is there for.
419 ================
420
421 However, our next step is to commit the *change* we did, and again, to
422 understand what's going on, keep in mind the difference between "working
423 tree contents", "index file" and "committed tree". We have changes
424 in the working tree that we want to commit, and we always have to
425 work through the index file, so the first thing we need to do is to
426 update the index cache:
427
428 ------------------------------------------------
429 $ git-update-index hello
430 ------------------------------------------------
431
432 (note how we didn't need the `\--add` flag this time, since git knew
433 about the file already).
434
435 Note what happens to the different `git-diff-\*` versions here. After
436 we've updated `hello` in the index, `git-diff-files -p` now shows no
437 differences, but `git-diff-index -p HEAD` still *does* show that the
438 current state is different from the state we committed. In fact, now
439 `git-diff-index` shows the same difference whether we use the `--cached`
440 flag or not, since now the index is coherent with the working tree.
441
442 Now, since we've updated `hello` in the index, we can commit the new
443 version. We could do it by writing the tree by hand again, and
444 committing the tree (this time we'd have to use the `-p HEAD` flag to
445 tell commit that the HEAD was the *parent* of the new commit, and that
446 this wasn't an initial commit any more), but you've done that once
447 already, so let's just use the helpful script this time:
448
449 ------------------------------------------------
450 $ git commit
451 ------------------------------------------------
452
453 which starts an editor for you to write the commit message and tells you
454 a bit about what you have done.
455
456 Write whatever message you want, and all the lines that start with '#'
457 will be pruned out, and the rest will be used as the commit message for
458 the change. If you decide you don't want to commit anything after all at
459 this point (you can continue to edit things and update the index), you
460 can just leave an empty message. Otherwise `git commit` will commit
461 the change for you.
462
463 You've now made your first real git commit. And if you're interested in
464 looking at what `git commit` really does, feel free to investigate:
465 it's a few very simple shell scripts to generate the helpful (?) commit
466 message headers, and a few one-liners that actually do the
467 commit itself (`git-commit`).
468
469
470 Inspecting Changes
471 ------------------
472
473 While creating changes is useful, it's even more useful if you can tell
474 later what changed. The most useful command for this is another of the
475 `diff` family, namely `git-diff-tree`.
476
477 `git-diff-tree` can be given two arbitrary trees, and it will tell you the
478 differences between them. Perhaps even more commonly, though, you can
479 give it just a single commit object, and it will figure out the parent
480 of that commit itself, and show the difference directly. Thus, to get
481 the same diff that we've already seen several times, we can now do
482
483 ----------------
484 $ git-diff-tree -p HEAD
485 ----------------
486
487 (again, `-p` means to show the difference as a human-readable patch),
488 and it will show what the last commit (in `HEAD`) actually changed.
489
490 [NOTE]
491 ============
492 Here is an ASCII art by Jon Loeliger that illustrates how
493 various diff-\* commands compare things.
494
495 diff-tree
496 +----+
497 | |
498 | |
499 V V
500 +-----------+
501 | Object DB |
502 | Backing |
503 | Store |
504 +-----------+
505 ^ ^
506 | |
507 | | diff-index --cached
508 | |
509 diff-index | V
510 | +-----------+
511 | | Index |
512 | | "cache" |
513 | +-----------+
514 | ^
515 | |
516 | | diff-files
517 | |
518 V V
519 +-----------+
520 | Working |
521 | Directory |
522 +-----------+
523 ============
524
525 More interestingly, you can also give `git-diff-tree` the `--pretty` flag,
526 which tells it to also show the commit message and author and date of the
527 commit, and you can tell it to show a whole series of diffs.
528 Alternatively, you can tell it to be "silent", and not show the diffs at
529 all, but just show the actual commit message.
530
531 In fact, together with the `git-rev-list` program (which generates a
532 list of revisions), `git-diff-tree` ends up being a veritable fount of
533 changes. A trivial (but very useful) script called `git-whatchanged` is
534 included with git which does exactly this, and shows a log of recent
535 activities.
536
537 To see the whole history of our pitiful little git-tutorial project, you
538 can do
539
540 ----------------
541 $ git log
542 ----------------
543
544 which shows just the log messages, or if we want to see the log together
545 with the associated patches use the more complex (and much more
546 powerful)
547
548 ----------------
549 $ git-whatchanged -p --root
550 ----------------
551
552 and you will see exactly what has changed in the repository over its
553 short history.
554
555 [NOTE]
556 The `\--root` flag is a flag to `git-diff-tree` to tell it to
557 show the initial aka 'root' commit too. Normally you'd probably not
558 want to see the initial import diff, but since the tutorial project
559 was started from scratch and is so small, we use it to make the result
560 a bit more interesting.
561
562 With that, you should now be having some inkling of what git does, and
563 can explore on your own.
564
565 [NOTE]
566 Most likely, you are not directly using the core
567 git Plumbing commands, but using Porcelain like Cogito on top
568 of it. Cogito works a bit differently and you usually do not
569 have to run `git-update-index` yourself for changed files (you
570 do tell underlying git about additions and removals via
571 `cg-add` and `cg-rm` commands). Just before you make a commit
572 with `cg-commit`, Cogito figures out which files you modified,
573 and runs `git-update-index` on them for you.
574
575
576 Tagging a version
577 -----------------
578
579 In git, there are two kinds of tags, a "light" one, and an "annotated tag".
580
581 A "light" tag is technically nothing more than a branch, except we put
582 it in the `.git/refs/tags/` subdirectory instead of calling it a `head`.
583 So the simplest form of tag involves nothing more than
584
585 ------------------------------------------------
586 $ git tag my-first-tag
587 ------------------------------------------------
588
589 which just writes the current `HEAD` into the `.git/refs/tags/my-first-tag`
590 file, after which point you can then use this symbolic name for that
591 particular state. You can, for example, do
592
593 ----------------
594 $ git diff my-first-tag
595 ----------------
596
597 to diff your current state against that tag (which at this point will
598 obviously be an empty diff, but if you continue to develop and commit
599 stuff, you can use your tag as an "anchor-point" to see what has changed
600 since you tagged it.
601
602 An "annotated tag" is actually a real git object, and contains not only a
603 pointer to the state you want to tag, but also a small tag name and
604 message, along with optionally a PGP signature that says that yes,
605 you really did
606 that tag. You create these annotated tags with either the `-a` or
607 `-s` flag to `git tag`:
608
609 ----------------
610 $ git tag -s <tagname>
611 ----------------
612
613 which will sign the current `HEAD` (but you can also give it another
614 argument that specifies the thing to tag, i.e., you could have tagged the
615 current `mybranch` point by using `git tag <tagname> mybranch`).
616
617 You normally only do signed tags for major releases or things
618 like that, while the light-weight tags are useful for any marking you
619 want to do -- any time you decide that you want to remember a certain
620 point, just create a private tag for it, and you have a nice symbolic
621 name for the state at that point.
622
623
624 Copying repositories
625 --------------------
626
627 git repositories are normally totally self-sufficient and relocatable
628 Unlike CVS, for example, there is no separate notion of
629 "repository" and "working tree". A git repository normally *is* the
630 working tree, with the local git information hidden in the `.git`
631 subdirectory. There is nothing else. What you see is what you got.
632
633 [NOTE]
634 You can tell git to split the git internal information from
635 the directory that it tracks, but we'll ignore that for now: it's not
636 how normal projects work, and it's really only meant for special uses.
637 So the mental model of "the git information is always tied directly to
638 the working tree that it describes" may not be technically 100%
639 accurate, but it's a good model for all normal use.
640
641 This has two implications:
642
643 - if you grow bored with the tutorial repository you created (or you've
644 made a mistake and want to start all over), you can just do simple
645 +
646 ----------------
647 $ rm -rf git-tutorial
648 ----------------
649 +
650 and it will be gone. There's no external repository, and there's no
651 history outside the project you created.
652
653 - if you want to move or duplicate a git repository, you can do so. There
654 is `git clone` command, but if all you want to do is just to
655 create a copy of your repository (with all the full history that
656 went along with it), you can do so with a regular
657 `cp -a git-tutorial new-git-tutorial`.
658 +
659 Note that when you've moved or copied a git repository, your git index
660 file (which caches various information, notably some of the "stat"
661 information for the files involved) will likely need to be refreshed.
662 So after you do a `cp -a` to create a new copy, you'll want to do
663 +
664 ----------------
665 $ git-update-index --refresh
666 ----------------
667 +
668 in the new repository to make sure that the index file is up-to-date.
669
670 Note that the second point is true even across machines. You can
671 duplicate a remote git repository with *any* regular copy mechanism, be it
672 `scp`, `rsync` or `wget`.
673
674 When copying a remote repository, you'll want to at a minimum update the
675 index cache when you do this, and especially with other peoples'
676 repositories you often want to make sure that the index cache is in some
677 known state (you don't know *what* they've done and not yet checked in),
678 so usually you'll precede the `git-update-index` with a
679
680 ----------------
681 $ git-read-tree --reset HEAD
682 $ git-update-index --refresh
683 ----------------
684
685 which will force a total index re-build from the tree pointed to by `HEAD`.
686 It resets the index contents to `HEAD`, and then the `git-update-index`
687 makes sure to match up all index entries with the checked-out files.
688 If the original repository had uncommitted changes in its
689 working tree, `git-update-index --refresh` notices them and
690 tells you they need to be updated.
691
692 The above can also be written as simply
693
694 ----------------
695 $ git reset
696 ----------------
697
698 and in fact a lot of the common git command combinations can be scripted
699 with the `git xyz` interfaces. You can learn things by just looking
700 at what the various git scripts do. For example, `git reset` is the
701 above two lines implemented in `git-reset`, but some things like
702 `git status` and `git commit` are slightly more complex scripts around
703 the basic git commands.
704
705 Many (most?) public remote repositories will not contain any of
706 the checked out files or even an index file, and will *only* contain the
707 actual core git files. Such a repository usually doesn't even have the
708 `.git` subdirectory, but has all the git files directly in the
709 repository.
710
711 To create your own local live copy of such a "raw" git repository, you'd
712 first create your own subdirectory for the project, and then copy the
713 raw repository contents into the `.git` directory. For example, to
714 create your own copy of the git repository, you'd do the following
715
716 ----------------
717 $ mkdir my-git
718 $ cd my-git
719 $ rsync -rL rsync://rsync.kernel.org/pub/scm/git/git.git/ .git
720 ----------------
721
722 followed by
723
724 ----------------
725 $ git-read-tree HEAD
726 ----------------
727
728 to populate the index. However, now you have populated the index, and
729 you have all the git internal files, but you will notice that you don't
730 actually have any of the working tree files to work on. To get
731 those, you'd check them out with
732
733 ----------------
734 $ git-checkout-index -u -a
735 ----------------
736
737 where the `-u` flag means that you want the checkout to keep the index
738 up-to-date (so that you don't have to refresh it afterward), and the
739 `-a` flag means "check out all files" (if you have a stale copy or an
740 older version of a checked out tree you may also need to add the `-f`
741 flag first, to tell git-checkout-index to *force* overwriting of any old
742 files).
743
744 Again, this can all be simplified with
745
746 ----------------
747 $ git clone rsync://rsync.kernel.org/pub/scm/git/git.git/ my-git
748 $ cd my-git
749 $ git checkout
750 ----------------
751
752 which will end up doing all of the above for you.
753
754 You have now successfully copied somebody else's (mine) remote
755 repository, and checked it out.
756
757
758 Creating a new branch
759 ---------------------
760
761 Branches in git are really nothing more than pointers into the git
762 object database from within the `.git/refs/` subdirectory, and as we
763 already discussed, the `HEAD` branch is nothing but a symlink to one of
764 these object pointers.
765
766 You can at any time create a new branch by just picking an arbitrary
767 point in the project history, and just writing the SHA1 name of that
768 object into a file under `.git/refs/heads/`. You can use any filename you
769 want (and indeed, subdirectories), but the convention is that the
770 "normal" branch is called `master`. That's just a convention, though,
771 and nothing enforces it.
772
773 To show that as an example, let's go back to the git-tutorial repository we
774 used earlier, and create a branch in it. You do that by simply just
775 saying that you want to check out a new branch:
776
777 ------------
778 $ git checkout -b mybranch
779 ------------
780
781 will create a new branch based at the current `HEAD` position, and switch
782 to it.
783
784 [NOTE]
785 ================================================
786 If you make the decision to start your new branch at some
787 other point in the history than the current `HEAD`, you can do so by
788 just telling `git checkout` what the base of the checkout would be.
789 In other words, if you have an earlier tag or branch, you'd just do
790
791 ------------
792 $ git checkout -b mybranch earlier-commit
793 ------------
794
795 and it would create the new branch `mybranch` at the earlier commit,
796 and check out the state at that time.
797 ================================================
798
799 You can always just jump back to your original `master` branch by doing
800
801 ------------
802 $ git checkout master
803 ------------
804
805 (or any other branch-name, for that matter) and if you forget which
806 branch you happen to be on, a simple
807
808 ------------
809 $ cat .git/HEAD
810 ------------
811
812 will tell you where it's pointing. To get the list of branches
813 you have, you can say
814
815 ------------
816 $ git branch
817 ------------
818
819 which is nothing more than a simple script around `ls .git/refs/heads`.
820 There will be asterisk in front of the branch you are currently on.
821
822 Sometimes you may wish to create a new branch _without_ actually
823 checking it out and switching to it. If so, just use the command
824
825 ------------
826 $ git branch <branchname> [startingpoint]
827 ------------
828
829 which will simply _create_ the branch, but will not do anything further.
830 You can then later -- once you decide that you want to actually develop
831 on that branch -- switch to that branch with a regular `git checkout`
832 with the branchname as the argument.
833
834
835 Merging two branches
836 --------------------
837
838 One of the ideas of having a branch is that you do some (possibly
839 experimental) work in it, and eventually merge it back to the main
840 branch. So assuming you created the above `mybranch` that started out
841 being the same as the original `master` branch, let's make sure we're in
842 that branch, and do some work there.
843
844 ------------------------------------------------
845 $ git checkout mybranch
846 $ echo "Work, work, work" >>hello
847 $ git commit -m 'Some work.' -i hello
848 ------------------------------------------------
849
850 Here, we just added another line to `hello`, and we used a shorthand for
851 doing both `git-update-index hello` and `git commit` by just giving the
852 filename directly to `git commit`, with an `-i` flag (it tells
853 git to 'include' that file in addition to what you have done to
854 the index file so far when making the commit). The `-m` flag is to give the
855 commit log message from the command line.
856
857 Now, to make it a bit more interesting, let's assume that somebody else
858 does some work in the original branch, and simulate that by going back
859 to the master branch, and editing the same file differently there:
860
861 ------------
862 $ git checkout master
863 ------------
864
865 Here, take a moment to look at the contents of `hello`, and notice how they
866 don't contain the work we just did in `mybranch` -- because that work
867 hasn't happened in the `master` branch at all. Then do
868
869 ------------
870 $ echo "Play, play, play" >>hello
871 $ echo "Lots of fun" >>example
872 $ git commit -m 'Some fun.' -i hello example
873 ------------
874
875 since the master branch is obviously in a much better mood.
876
877 Now, you've got two branches, and you decide that you want to merge the
878 work done. Before we do that, let's introduce a cool graphical tool that
879 helps you view what's going on:
880
881 ----------------
882 $ gitk --all
883 ----------------
884
885 will show you graphically both of your branches (that's what the `\--all`
886 means: normally it will just show you your current `HEAD`) and their
887 histories. You can also see exactly how they came to be from a common
888 source.
889
890 Anyway, let's exit `gitk` (`^Q` or the File menu), and decide that we want
891 to merge the work we did on the `mybranch` branch into the `master`
892 branch (which is currently our `HEAD` too). To do that, there's a nice
893 script called `git merge`, which wants to know which branches you want
894 to resolve and what the merge is all about:
895
896 ------------
897 $ git merge "Merge work in mybranch" HEAD mybranch
898 ------------
899
900 where the first argument is going to be used as the commit message if
901 the merge can be resolved automatically.
902
903 Now, in this case we've intentionally created a situation where the
904 merge will need to be fixed up by hand, though, so git will do as much
905 of it as it can automatically (which in this case is just merge the `example`
906 file, which had no differences in the `mybranch` branch), and say:
907
908 ----------------
909 Trying really trivial in-index merge...
910 fatal: Merge requires file-level merging
911 Nope.
912 ...
913 Auto-merging hello
914 CONFLICT (content): Merge conflict in hello
915 Automatic merge failed; fix up by hand
916 ----------------
917
918 which is way too verbose, but it basically tells you that it failed the
919 really trivial merge ("Simple merge") and did an "Automatic merge"
920 instead, but that too failed due to conflicts in `hello`.
921
922 Not to worry. It left the (trivial) conflict in `hello` in the same form you
923 should already be well used to if you've ever used CVS, so let's just
924 open `hello` in our editor (whatever that may be), and fix it up somehow.
925 I'd suggest just making it so that `hello` contains all four lines:
926
927 ------------
928 Hello World
929 It's a new day for git
930 Play, play, play
931 Work, work, work
932 ------------
933
934 and once you're happy with your manual merge, just do a
935
936 ------------
937 $ git commit -i hello
938 ------------
939
940 which will very loudly warn you that you're now committing a merge
941 (which is correct, so never mind), and you can write a small merge
942 message about your adventures in git-merge-land.
943
944 After you're done, start up `gitk \--all` to see graphically what the
945 history looks like. Notice that `mybranch` still exists, and you can
946 switch to it, and continue to work with it if you want to. The
947 `mybranch` branch will not contain the merge, but next time you merge it
948 from the `master` branch, git will know how you merged it, so you'll not
949 have to do _that_ merge again.
950
951 Another useful tool, especially if you do not always work in X-Window
952 environment, is `git show-branch`.
953
954 ------------------------------------------------
955 $ git show-branch --topo-order master mybranch
956 * [master] Merge work in mybranch
957 ! [mybranch] Some work.
958 --
959 - [master] Merge work in mybranch
960 *+ [mybranch] Some work.
961 ------------------------------------------------
962
963 The first two lines indicate that it is showing the two branches
964 and the first line of the commit log message from their
965 top-of-the-tree commits, you are currently on `master` branch
966 (notice the asterisk `\*` character), and the first column for
967 the later output lines is used to show commits contained in the
968 `master` branch, and the second column for the `mybranch`
969 branch. Three commits are shown along with their log messages.
970 All of them have non blank characters in the first column (`*`
971 shows an ordinary commit on the current branch, `.` is a merge commit), which
972 means they are now part of the `master` branch. Only the "Some
973 work" commit has the plus `+` character in the second column,
974 because `mybranch` has not been merged to incorporate these
975 commits from the master branch. The string inside brackets
976 before the commit log message is a short name you can use to
977 name the commit. In the above example, 'master' and 'mybranch'
978 are branch heads. 'master~1' is the first parent of 'master'
979 branch head. Please see 'git-rev-parse' documentation if you
980 see more complex cases.
981
982 Now, let's pretend you are the one who did all the work in
983 `mybranch`, and the fruit of your hard work has finally been merged
984 to the `master` branch. Let's go back to `mybranch`, and run
985 resolve to get the "upstream changes" back to your branch.
986
987 ------------
988 $ git checkout mybranch
989 $ git merge "Merge upstream changes." HEAD master
990 ------------
991
992 This outputs something like this (the actual commit object names
993 would be different)
994
995 ----------------
996 Updating from ae3a2da... to a80b4aa....
997 Fast forward
998 example | 1 +
999 hello | 1 +
1000 2 files changed, 2 insertions(+), 0 deletions(-)
1001 ----------------
1002
1003 Because your branch did not contain anything more than what are
1004 already merged into the `master` branch, the resolve operation did
1005 not actually do a merge. Instead, it just updated the top of
1006 the tree of your branch to that of the `master` branch. This is
1007 often called 'fast forward' merge.
1008
1009 You can run `gitk \--all` again to see how the commit ancestry
1010 looks like, or run `show-branch`, which tells you this.
1011
1012 ------------------------------------------------
1013 $ git show-branch master mybranch
1014 ! [master] Merge work in mybranch
1015 * [mybranch] Merge work in mybranch
1016 --
1017 -- [master] Merge work in mybranch
1018 ------------------------------------------------
1019
1020
1021 Merging external work
1022 ---------------------
1023
1024 It's usually much more common that you merge with somebody else than
1025 merging with your own branches, so it's worth pointing out that git
1026 makes that very easy too, and in fact, it's not that different from
1027 doing a `git merge`. In fact, a remote merge ends up being nothing
1028 more than "fetch the work from a remote repository into a temporary tag"
1029 followed by a `git merge`.
1030
1031 Fetching from a remote repository is done by, unsurprisingly,
1032 `git fetch`:
1033
1034 ----------------
1035 $ git fetch <remote-repository>
1036 ----------------
1037
1038 One of the following transports can be used to name the
1039 repository to download from:
1040
1041 Rsync::
1042 `rsync://remote.machine/path/to/repo.git/`
1043 +
1044 Rsync transport is usable for both uploading and downloading,
1045 but is completely unaware of what git does, and can produce
1046 unexpected results when you download from the public repository
1047 while the repository owner is uploading into it via `rsync`
1048 transport. Most notably, it could update the files under
1049 `refs/` which holds the object name of the topmost commits
1050 before uploading the files in `objects/` -- the downloader would
1051 obtain head commit object name while that object itself is still
1052 not available in the repository. For this reason, it is
1053 considered deprecated.
1054
1055 SSH::
1056 `remote.machine:/path/to/repo.git/` or
1057 +
1058 `ssh://remote.machine/path/to/repo.git/`
1059 +
1060 This transport can be used for both uploading and downloading,
1061 and requires you to have a log-in privilege over `ssh` to the
1062 remote machine. It finds out the set of objects the other side
1063 lacks by exchanging the head commits both ends have and
1064 transfers (close to) minimum set of objects. It is by far the
1065 most efficient way to exchange git objects between repositories.
1066
1067 Local directory::
1068 `/path/to/repo.git/`
1069 +
1070 This transport is the same as SSH transport but uses `sh` to run
1071 both ends on the local machine instead of running other end on
1072 the remote machine via `ssh`.
1073
1074 git Native::
1075 `git://remote.machine/path/to/repo.git/`
1076 +
1077 This transport was designed for anonymous downloading. Like SSH
1078 transport, it finds out the set of objects the downstream side
1079 lacks and transfers (close to) minimum set of objects.
1080
1081 HTTP(S)::
1082 `http://remote.machine/path/to/repo.git/`
1083 +
1084 Downloader from http and https URL
1085 first obtains the topmost commit object name from the remote site
1086 by looking at the specified refname under `repo.git/refs/` directory,
1087 and then tries to obtain the
1088 commit object by downloading from `repo.git/objects/xx/xxx\...`
1089 using the object name of that commit object. Then it reads the
1090 commit object to find out its parent commits and the associate
1091 tree object; it repeats this process until it gets all the
1092 necessary objects. Because of this behavior, they are
1093 sometimes also called 'commit walkers'.
1094 +
1095 The 'commit walkers' are sometimes also called 'dumb
1096 transports', because they do not require any git aware smart
1097 server like git Native transport does. Any stock HTTP server
1098 that does not even support directory index would suffice. But
1099 you must prepare your repository with `git-update-server-info`
1100 to help dumb transport downloaders.
1101 +
1102 There are (confusingly enough) `git-ssh-fetch` and `git-ssh-upload`
1103 programs, which are 'commit walkers'; they outlived their
1104 usefulness when git Native and SSH transports were introduced,
1105 and not used by `git pull` or `git push` scripts.
1106
1107 Once you fetch from the remote repository, you `resolve` that
1108 with your current branch.
1109
1110 However -- it's such a common thing to `fetch` and then
1111 immediately `resolve`, that it's called `git pull`, and you can
1112 simply do
1113
1114 ----------------
1115 $ git pull <remote-repository>
1116 ----------------
1117
1118 and optionally give a branch-name for the remote end as a second
1119 argument.
1120
1121 [NOTE]
1122 You could do without using any branches at all, by
1123 keeping as many local repositories as you would like to have
1124 branches, and merging between them with `git pull`, just like
1125 you merge between branches. The advantage of this approach is
1126 that it lets you keep set of files for each `branch` checked
1127 out and you may find it easier to switch back and forth if you
1128 juggle multiple lines of development simultaneously. Of
1129 course, you will pay the price of more disk usage to hold
1130 multiple working trees, but disk space is cheap these days.
1131
1132 [NOTE]
1133 You could even pull from your own repository by
1134 giving '.' as <remote-repository> parameter to `git pull`. This
1135 is useful when you want to merge a local branch (or more, if you
1136 are making an Octopus) into the current branch.
1137
1138 It is likely that you will be pulling from the same remote
1139 repository from time to time. As a short hand, you can store
1140 the remote repository URL in a file under .git/remotes/
1141 directory, like this:
1142
1143 ------------------------------------------------
1144 $ mkdir -p .git/remotes/
1145 $ cat >.git/remotes/linus <<\EOF
1146 URL: http://www.kernel.org/pub/scm/git/git.git/
1147 EOF
1148 ------------------------------------------------
1149
1150 and use the filename to `git pull` instead of the full URL.
1151 The URL specified in such file can even be a prefix
1152 of a full URL, like this:
1153
1154 ------------------------------------------------
1155 $ cat >.git/remotes/jgarzik <<\EOF
1156 URL: http://www.kernel.org/pub/scm/linux/git/jgarzik/
1157 EOF
1158 ------------------------------------------------
1159
1160
1161 Examples.
1162
1163 . `git pull linus`
1164 . `git pull linus tag v0.99.1`
1165 . `git pull jgarzik/netdev-2.6.git/ e100`
1166
1167 the above are equivalent to:
1168
1169 . `git pull http://www.kernel.org/pub/scm/git/git.git/ HEAD`
1170 . `git pull http://www.kernel.org/pub/scm/git/git.git/ tag v0.99.1`
1171 . `git pull http://www.kernel.org/pub/.../jgarzik/netdev-2.6.git e100`
1172
1173
1174 How does the merge work?
1175 ------------------------
1176
1177 We said this tutorial shows what plumbing does to help you cope
1178 with the porcelain that isn't flushing, but we so far did not
1179 talk about how the merge really works. If you are following
1180 this tutorial the first time, I'd suggest to skip to "Publishing
1181 your work" section and come back here later.
1182
1183 OK, still with me? To give us an example to look at, let's go
1184 back to the earlier repository with "hello" and "example" file,
1185 and bring ourselves back to the pre-merge state:
1186
1187 ------------
1188 $ git show-branch --more=3 master mybranch
1189 ! [master] Merge work in mybranch
1190 * [mybranch] Merge work in mybranch
1191 --
1192 -- [master] Merge work in mybranch
1193 +* [master^2] Some work.
1194 +* [master^] Some fun.
1195 ------------
1196
1197 Remember, before running `git merge`, our `master` head was at
1198 "Some fun." commit, while our `mybranch` head was at "Some
1199 work." commit.
1200
1201 ------------
1202 $ git checkout mybranch
1203 $ git reset --hard master^2
1204 $ git checkout master
1205 $ git reset --hard master^
1206 ------------
1207
1208 After rewinding, the commit structure should look like this:
1209
1210 ------------
1211 $ git show-branch
1212 * [master] Some fun.
1213 ! [mybranch] Some work.
1214 --
1215 + [mybranch] Some work.
1216 * [master] Some fun.
1217 *+ [mybranch^] New day.
1218 ------------
1219
1220 Now we are ready to experiment with the merge by hand.
1221
1222 `git merge` command, when merging two branches, uses 3-way merge
1223 algorithm. First, it finds the common ancestor between them.
1224 The command it uses is `git-merge-base`:
1225
1226 ------------
1227 $ mb=$(git-merge-base HEAD mybranch)
1228 ------------
1229
1230 The command writes the commit object name of the common ancestor
1231 to the standard output, so we captured its output to a variable,
1232 because we will be using it in the next step. BTW, the common
1233 ancestor commit is the "New day." commit in this case. You can
1234 tell it by:
1235
1236 ------------
1237 $ git-name-rev $mb
1238 my-first-tag
1239 ------------
1240
1241 After finding out a common ancestor commit, the second step is
1242 this:
1243
1244 ------------
1245 $ git-read-tree -m -u $mb HEAD mybranch
1246 ------------
1247
1248 This is the same `git-read-tree` command we have already seen,
1249 but it takes three trees, unlike previous examples. This reads
1250 the contents of each tree into different 'stage' in the index
1251 file (the first tree goes to stage 1, the second stage 2,
1252 etc.). After reading three trees into three stages, the paths
1253 that are the same in all three stages are 'collapsed' into stage
1254 0. Also paths that are the same in two of three stages are
1255 collapsed into stage 0, taking the SHA1 from either stage 2 or
1256 stage 3, whichever is different from stage 1 (i.e. only one side
1257 changed from the common ancestor).
1258
1259 After 'collapsing' operation, paths that are different in three
1260 trees are left in non-zero stages. At this point, you can
1261 inspect the index file with this command:
1262
1263 ------------
1264 $ git-ls-files --stage
1265 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example
1266 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello
1267 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello
1268 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1269 ------------
1270
1271 In our example of only two files, we did not have unchanged
1272 files so only 'example' resulted in collapsing, but in real-life
1273 large projects, only small number of files change in one commit,
1274 and this 'collapsing' tends to trivially merge most of the paths
1275 fairly quickly, leaving only a handful the real changes in non-zero
1276 stages.
1277
1278 To look at only non-zero stages, use `\--unmerged` flag:
1279
1280 ------------
1281 $ git-ls-files --unmerged
1282 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello
1283 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello
1284 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1285 ------------
1286
1287 The next step of merging is to merge these three versions of the
1288 file, using 3-way merge. This is done by giving
1289 `git-merge-one-file` command as one of the arguments to
1290 `git-merge-index` command:
1291
1292 ------------
1293 $ git-merge-index git-merge-one-file hello
1294 Auto-merging hello.
1295 merge: warning: conflicts during merge
1296 ERROR: Merge conflict in hello.
1297 fatal: merge program failed
1298 ------------
1299
1300 `git-merge-one-file` script is called with parameters to
1301 describe those three versions, and is responsible to leave the
1302 merge results in the working tree.
1303 It is a fairly straightforward shell script, and
1304 eventually calls `merge` program from RCS suite to perform a
1305 file-level 3-way merge. In this case, `merge` detects
1306 conflicts, and the merge result with conflict marks is left in
1307 the working tree.. This can be seen if you run `ls-files
1308 --stage` again at this point:
1309
1310 ------------
1311 $ git-ls-files --stage
1312 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example
1313 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello
1314 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello
1315 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello
1316 ------------
1317
1318 This is the state of the index file and the working file after
1319 `git merge` returns control back to you, leaving the conflicting
1320 merge for you to resolve. Notice that the path `hello` is still
1321 unmerged, and what you see with `git diff` at this point is
1322 differences since stage 2 (i.e. your version).
1323
1324
1325 Publishing your work
1326 --------------------
1327
1328 So we can use somebody else's work from a remote repository; but
1329 how can *you* prepare a repository to let other people pull from
1330 it?
1331
1332 Your do your real work in your working tree that has your
1333 primary repository hanging under it as its `.git` subdirectory.
1334 You *could* make that repository accessible remotely and ask
1335 people to pull from it, but in practice that is not the way
1336 things are usually done. A recommended way is to have a public
1337 repository, make it reachable by other people, and when the
1338 changes you made in your primary working tree are in good shape,
1339 update the public repository from it. This is often called
1340 'pushing'.
1341
1342 [NOTE]
1343 This public repository could further be mirrored, and that is
1344 how git repositories at `kernel.org` are managed.
1345
1346 Publishing the changes from your local (private) repository to
1347 your remote (public) repository requires a write privilege on
1348 the remote machine. You need to have an SSH account there to
1349 run a single command, `git-receive-pack`.
1350
1351 First, you need to create an empty repository on the remote
1352 machine that will house your public repository. This empty
1353 repository will be populated and be kept up-to-date by pushing
1354 into it later. Obviously, this repository creation needs to be
1355 done only once.
1356
1357 [NOTE]
1358 `git push` uses a pair of programs,
1359 `git-send-pack` on your local machine, and `git-receive-pack`
1360 on the remote machine. The communication between the two over
1361 the network internally uses an SSH connection.
1362
1363 Your private repository's git directory is usually `.git`, but
1364 your public repository is often named after the project name,
1365 i.e. `<project>.git`. Let's create such a public repository for
1366 project `my-git`. After logging into the remote machine, create
1367 an empty directory:
1368
1369 ------------
1370 $ mkdir my-git.git
1371 ------------
1372
1373 Then, make that directory into a git repository by running
1374 `git init-db`, but this time, since its name is not the usual
1375 `.git`, we do things slightly differently:
1376
1377 ------------
1378 $ GIT_DIR=my-git.git git-init-db
1379 ------------
1380
1381 Make sure this directory is available for others you want your
1382 changes to be pulled by via the transport of your choice. Also
1383 you need to make sure that you have the `git-receive-pack`
1384 program on the `$PATH`.
1385
1386 [NOTE]
1387 Many installations of sshd do not invoke your shell as the login
1388 shell when you directly run programs; what this means is that if
1389 your login shell is `bash`, only `.bashrc` is read and not
1390 `.bash_profile`. As a workaround, make sure `.bashrc` sets up
1391 `$PATH` so that you can run `git-receive-pack` program.
1392
1393 [NOTE]
1394 If you plan to publish this repository to be accessed over http,
1395 you should do `chmod +x my-git.git/hooks/post-update` at this
1396 point. This makes sure that every time you push into this
1397 repository, `git-update-server-info` is run.
1398
1399 Your "public repository" is now ready to accept your changes.
1400 Come back to the machine you have your private repository. From
1401 there, run this command:
1402
1403 ------------
1404 $ git push <public-host>:/path/to/my-git.git master
1405 ------------
1406
1407 This synchronizes your public repository to match the named
1408 branch head (i.e. `master` in this case) and objects reachable
1409 from them in your current repository.
1410
1411 As a real example, this is how I update my public git
1412 repository. Kernel.org mirror network takes care of the
1413 propagation to other publicly visible machines:
1414
1415 ------------
1416 $ git push master.kernel.org:/pub/scm/git/git.git/
1417 ------------
1418
1419
1420 Packing your repository
1421 -----------------------
1422
1423 Earlier, we saw that one file under `.git/objects/??/` directory
1424 is stored for each git object you create. This representation
1425 is efficient to create atomically and safely, but
1426 not so convenient to transport over the network. Since git objects are
1427 immutable once they are created, there is a way to optimize the
1428 storage by "packing them together". The command
1429
1430 ------------
1431 $ git repack
1432 ------------
1433
1434 will do it for you. If you followed the tutorial examples, you
1435 would have accumulated about 17 objects in `.git/objects/??/`
1436 directories by now. `git repack` tells you how many objects it
1437 packed, and stores the packed file in `.git/objects/pack`
1438 directory.
1439
1440 [NOTE]
1441 You will see two files, `pack-\*.pack` and `pack-\*.idx`,
1442 in `.git/objects/pack` directory. They are closely related to
1443 each other, and if you ever copy them by hand to a different
1444 repository for whatever reason, you should make sure you copy
1445 them together. The former holds all the data from the objects
1446 in the pack, and the latter holds the index for random
1447 access.
1448
1449 If you are paranoid, running `git-verify-pack` command would
1450 detect if you have a corrupt pack, but do not worry too much.
1451 Our programs are always perfect ;-).
1452
1453 Once you have packed objects, you do not need to leave the
1454 unpacked objects that are contained in the pack file anymore.
1455
1456 ------------
1457 $ git prune-packed
1458 ------------
1459
1460 would remove them for you.
1461
1462 You can try running `find .git/objects -type f` before and after
1463 you run `git prune-packed` if you are curious. Also `git
1464 count-objects` would tell you how many unpacked objects are in
1465 your repository and how much space they are consuming.
1466
1467 [NOTE]
1468 `git pull` is slightly cumbersome for HTTP transport, as a
1469 packed repository may contain relatively few objects in a
1470 relatively large pack. If you expect many HTTP pulls from your
1471 public repository you might want to repack & prune often, or
1472 never.
1473
1474 If you run `git repack` again at this point, it will say
1475 "Nothing to pack". Once you continue your development and
1476 accumulate the changes, running `git repack` again will create a
1477 new pack, that contains objects created since you packed your
1478 repository the last time. We recommend that you pack your project
1479 soon after the initial import (unless you are starting your
1480 project from scratch), and then run `git repack` every once in a
1481 while, depending on how active your project is.
1482
1483 When a repository is synchronized via `git push` and `git pull`
1484 objects packed in the source repository are usually stored
1485 unpacked in the destination, unless rsync transport is used.
1486 While this allows you to use different packing strategies on
1487 both ends, it also means you may need to repack both
1488 repositories every once in a while.
1489
1490
1491 Working with Others
1492 -------------------
1493
1494 Although git is a truly distributed system, it is often
1495 convenient to organize your project with an informal hierarchy
1496 of developers. Linux kernel development is run this way. There
1497 is a nice illustration (page 17, "Merges to Mainline") in Randy
1498 Dunlap's presentation (`http://tinyurl.com/a2jdg`).
1499
1500 It should be stressed that this hierarchy is purely *informal*.
1501 There is nothing fundamental in git that enforces the "chain of
1502 patch flow" this hierarchy implies. You do not have to pull
1503 from only one remote repository.
1504
1505 A recommended workflow for a "project lead" goes like this:
1506
1507 1. Prepare your primary repository on your local machine. Your
1508 work is done there.
1509
1510 2. Prepare a public repository accessible to others.
1511 +
1512 If other people are pulling from your repository over dumb
1513 transport protocols (HTTP), you need to keep this repository
1514 'dumb transport friendly'. After `git init-db`,
1515 `$GIT_DIR/hooks/post-update` copied from the standard templates
1516 would contain a call to `git-update-server-info` but the
1517 `post-update` hook itself is disabled by default -- enable it
1518 with `chmod +x post-update`. This makes sure `git-update-server-info`
1519 keeps the necessary files up-to-date.
1520
1521 3. Push into the public repository from your primary
1522 repository.
1523
1524 4. `git repack` the public repository. This establishes a big
1525 pack that contains the initial set of objects as the
1526 baseline, and possibly `git prune` if the transport
1527 used for pulling from your repository supports packed
1528 repositories.
1529
1530 5. Keep working in your primary repository. Your changes
1531 include modifications of your own, patches you receive via
1532 e-mails, and merges resulting from pulling the "public"
1533 repositories of your "subsystem maintainers".
1534 +
1535 You can repack this private repository whenever you feel like.
1536
1537 6. Push your changes to the public repository, and announce it
1538 to the public.
1539
1540 7. Every once in a while, "git repack" the public repository.
1541 Go back to step 5. and continue working.
1542
1543
1544 A recommended work cycle for a "subsystem maintainer" who works
1545 on that project and has an own "public repository" goes like this:
1546
1547 1. Prepare your work repository, by `git clone` the public
1548 repository of the "project lead". The URL used for the
1549 initial cloning is stored in `.git/remotes/origin`.
1550
1551 2. Prepare a public repository accessible to others, just like
1552 the "project lead" person does.
1553
1554 3. Copy over the packed files from "project lead" public
1555 repository to your public repository, unless the "project
1556 lead" repository lives on the same machine as yours. In the
1557 latter case, you can use `objects/info/alternates` file to
1558 point at the repository you are borrowing from.
1559
1560 4. Push into the public repository from your primary
1561 repository. Run `git repack`, and possibly `git prune` if the
1562 transport used for pulling from your repository supports
1563 packed repositories.
1564
1565 5. Keep working in your primary repository. Your changes
1566 include modifications of your own, patches you receive via
1567 e-mails, and merges resulting from pulling the "public"
1568 repositories of your "project lead" and possibly your
1569 "sub-subsystem maintainers".
1570 +
1571 You can repack this private repository whenever you feel
1572 like.
1573
1574 6. Push your changes to your public repository, and ask your
1575 "project lead" and possibly your "sub-subsystem
1576 maintainers" to pull from it.
1577
1578 7. Every once in a while, `git repack` the public repository.
1579 Go back to step 5. and continue working.
1580
1581
1582 A recommended work cycle for an "individual developer" who does
1583 not have a "public" repository is somewhat different. It goes
1584 like this:
1585
1586 1. Prepare your work repository, by `git clone` the public
1587 repository of the "project lead" (or a "subsystem
1588 maintainer", if you work on a subsystem). The URL used for
1589 the initial cloning is stored in `.git/remotes/origin`.
1590
1591 2. Do your work in your repository on 'master' branch.
1592
1593 3. Run `git fetch origin` from the public repository of your
1594 upstream every once in a while. This does only the first
1595 half of `git pull` but does not merge. The head of the
1596 public repository is stored in `.git/refs/heads/origin`.
1597
1598 4. Use `git cherry origin` to see which ones of your patches
1599 were accepted, and/or use `git rebase origin` to port your
1600 unmerged changes forward to the updated upstream.
1601
1602 5. Use `git format-patch origin` to prepare patches for e-mail
1603 submission to your upstream and send it out. Go back to
1604 step 2. and continue.
1605
1606
1607 Working with Others, Shared Repository Style
1608 --------------------------------------------
1609
1610 If you are coming from CVS background, the style of cooperation
1611 suggested in the previous section may be new to you. You do not
1612 have to worry. git supports "shared public repository" style of
1613 cooperation you are probably more familiar with as well.
1614
1615 See link:cvs-migration.html[git for CVS users] for the details.
1616
1617 Bundling your work together
1618 ---------------------------
1619
1620 It is likely that you will be working on more than one thing at
1621 a time. It is easy to manage those more-or-less independent tasks
1622 using branches with git.
1623
1624 We have already seen how branches work previously,
1625 with "fun and work" example using two branches. The idea is the
1626 same if there are more than two branches. Let's say you started
1627 out from "master" head, and have some new code in the "master"
1628 branch, and two independent fixes in the "commit-fix" and
1629 "diff-fix" branches:
1630
1631 ------------
1632 $ git show-branch
1633 ! [commit-fix] Fix commit message normalization.
1634 ! [diff-fix] Fix rename detection.
1635 * [master] Release candidate #1
1636 ---
1637 + [diff-fix] Fix rename detection.
1638 + [diff-fix~1] Better common substring algorithm.
1639 + [commit-fix] Fix commit message normalization.
1640 * [master] Release candidate #1
1641 ++* [diff-fix~2] Pretty-print messages.
1642 ------------
1643
1644 Both fixes are tested well, and at this point, you want to merge
1645 in both of them. You could merge in 'diff-fix' first and then
1646 'commit-fix' next, like this:
1647
1648 ------------
1649 $ git merge 'Merge fix in diff-fix' master diff-fix
1650 $ git merge 'Merge fix in commit-fix' master commit-fix
1651 ------------
1652
1653 Which would result in:
1654
1655 ------------
1656 $ git show-branch
1657 ! [commit-fix] Fix commit message normalization.
1658 ! [diff-fix] Fix rename detection.
1659 * [master] Merge fix in commit-fix
1660 ---
1661 - [master] Merge fix in commit-fix
1662 + * [commit-fix] Fix commit message normalization.
1663 - [master~1] Merge fix in diff-fix
1664 +* [diff-fix] Fix rename detection.
1665 +* [diff-fix~1] Better common substring algorithm.
1666 * [master~2] Release candidate #1
1667 ++* [master~3] Pretty-print messages.
1668 ------------
1669
1670 However, there is no particular reason to merge in one branch
1671 first and the other next, when what you have are a set of truly
1672 independent changes (if the order mattered, then they are not
1673 independent by definition). You could instead merge those two
1674 branches into the current branch at once. First let's undo what
1675 we just did and start over. We would want to get the master
1676 branch before these two merges by resetting it to 'master~2':
1677
1678 ------------
1679 $ git reset --hard master~2
1680 ------------
1681
1682 You can make sure 'git show-branch' matches the state before
1683 those two 'git merge' you just did. Then, instead of running
1684 two 'git merge' commands in a row, you would pull these two
1685 branch heads (this is known as 'making an Octopus'):
1686
1687 ------------
1688 $ git pull . commit-fix diff-fix
1689 $ git show-branch
1690 ! [commit-fix] Fix commit message normalization.
1691 ! [diff-fix] Fix rename detection.
1692 * [master] Octopus merge of branches 'diff-fix' and 'commit-fix'
1693 ---
1694 - [master] Octopus merge of branches 'diff-fix' and 'commit-fix'
1695 + * [commit-fix] Fix commit message normalization.
1696 +* [diff-fix] Fix rename detection.
1697 +* [diff-fix~1] Better common substring algorithm.
1698 * [master~1] Release candidate #1
1699 ++* [master~2] Pretty-print messages.
1700 ------------
1701
1702 Note that you should not do Octopus because you can. An octopus
1703 is a valid thing to do and often makes it easier to view the
1704 commit history if you are pulling more than two independent
1705 changes at the same time. However, if you have merge conflicts
1706 with any of the branches you are merging in and need to hand
1707 resolve, that is an indication that the development happened in
1708 those branches were not independent after all, and you should
1709 merge two at a time, documenting how you resolved the conflicts,
1710 and the reason why you preferred changes made in one side over
1711 the other. Otherwise it would make the project history harder
1712 to follow, not easier.
1713
1714 [ to be continued.. cvsimports ]