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