6 gittutorial-2 - A tutorial introduction to git: part two
15 You should work through linkgit:gittutorial[7] before reading this tutorial.
17 The goal of this tutorial is to introduce two fundamental pieces of
18 git's architecture--the object database and the index file--and to
19 provide the reader with everything necessary to understand the rest
20 of the git documentation.
22 The git object database
23 -----------------------
25 Let's start a new project and create a small amount of history:
27 ------------------------------------------------
31 Initialized empty Git repository in .git/
32 $ echo 'hello world' > file.txt
34 $ git commit -a -m "initial commit"
35 Created root-commit 54196cc (initial commit) on master
36 create mode 100644 file.txt
37 $ echo 'hello world!' >file.txt
38 $ git commit -a -m "add emphasis"
39 Created c4d59f3 (add emphasis) on master
40 ------------------------------------------------
42 What are the 7 digits of hex that git responded to the commit with?
44 We saw in part one of the tutorial that commits have names like this.
45 It turns out that every object in the git history is stored under
46 a 40-digit hex name. That name is the SHA1 hash of the object's
47 contents; among other things, this ensures that git will never store
48 the same data twice (since identical data is given an identical SHA1
49 name), and that the contents of a git object will never change (since
50 that would change the object's name as well). The 7 char hex strings
51 here are simply the abbreviation of such 40 character long strings.
52 Abbreviations can be used everywhere where the 40 character strings
53 can be used, so long as they are unambiguous.
55 It is expected that the content of the commit object you created while
56 following the example above generates a different SHA1 hash than
57 the one shown above because the commit object records the time when
58 it was created and the name of the person performing the commit.
60 We can ask git about this particular object with the `cat-file`
61 command. Don't copy the 40 hex digits from this example but use those
62 from your own version. Note that you can shorten it to only a few
63 characters to save yourself typing all 40 hex digits:
65 ------------------------------------------------
66 $ git cat-file -t 54196cc2
68 $ git cat-file commit 54196cc2
69 tree 92b8b694ffb1675e5975148e1121810081dbdffe
70 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
71 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
74 ------------------------------------------------
76 A tree can refer to one or more "blob" objects, each corresponding to
77 a file. In addition, a tree can also refer to other tree objects,
78 thus creating a directory hierarchy. You can examine the contents of
79 any tree using ls-tree (remember that a long enough initial portion
80 of the SHA1 will also work):
82 ------------------------------------------------
83 $ git ls-tree 92b8b694
84 100644 blob 3b18e512dba79e4c8300dd08aeb37f8e728b8dad file.txt
85 ------------------------------------------------
87 Thus we see that this tree has one file in it. The SHA1 hash is a
88 reference to that file's data:
90 ------------------------------------------------
91 $ git cat-file -t 3b18e512
93 ------------------------------------------------
95 A "blob" is just file data, which we can also examine with cat-file:
97 ------------------------------------------------
98 $ git cat-file blob 3b18e512
100 ------------------------------------------------
102 Note that this is the old file data; so the object that git named in
103 its response to the initial tree was a tree with a snapshot of the
104 directory state that was recorded by the first commit.
106 All of these objects are stored under their SHA1 names inside the git
109 ------------------------------------------------
115 .git/objects/3b/18e512dba79e4c8300dd08aeb37f8e728b8dad
117 .git/objects/92/b8b694ffb1675e5975148e1121810081dbdffe
119 .git/objects/54/196cc2703dc165cbd373a65a4dcf22d50ae7f7
121 .git/objects/a0/423896973644771497bdc03eb99d5281615b51
123 .git/objects/d0/492b368b66bdabf2ac1fd8c92b39d3db916e59
125 .git/objects/c4/d59f390b9cfd4318117afde11d601c1085f241
126 ------------------------------------------------
128 and the contents of these files is just the compressed data plus a
129 header identifying their length and their type. The type is either a
130 blob, a tree, a commit, or a tag.
132 The simplest commit to find is the HEAD commit, which we can find
135 ------------------------------------------------
137 ref: refs/heads/master
138 ------------------------------------------------
140 As you can see, this tells us which branch we're currently on, and it
141 tells us this by naming a file under the .git directory, which itself
142 contains a SHA1 name referring to a commit object, which we can
143 examine with cat-file:
145 ------------------------------------------------
146 $ cat .git/refs/heads/master
147 c4d59f390b9cfd4318117afde11d601c1085f241
148 $ git cat-file -t c4d59f39
150 $ git cat-file commit c4d59f39
151 tree d0492b368b66bdabf2ac1fd8c92b39d3db916e59
152 parent 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
153 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
154 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
157 ------------------------------------------------
159 The "tree" object here refers to the new state of the tree:
161 ------------------------------------------------
162 $ git ls-tree d0492b36
163 100644 blob a0423896973644771497bdc03eb99d5281615b51 file.txt
164 $ git cat-file blob a0423896
166 ------------------------------------------------
168 and the "parent" object refers to the previous commit:
170 ------------------------------------------------
171 $ git cat-file commit 54196cc2
172 tree 92b8b694ffb1675e5975148e1121810081dbdffe
173 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
174 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
177 ------------------------------------------------
179 The tree object is the tree we examined first, and this commit is
180 unusual in that it lacks any parent.
182 Most commits have only one parent, but it is also common for a commit
183 to have multiple parents. In that case the commit represents a
184 merge, with the parent references pointing to the heads of the merged
187 Besides blobs, trees, and commits, the only remaining type of object
188 is a "tag", which we won't discuss here; refer to linkgit:git-tag[1]
191 So now we know how git uses the object database to represent a
194 * "commit" objects refer to "tree" objects representing the
195 snapshot of a directory tree at a particular point in the
196 history, and refer to "parent" commits to show how they're
197 connected into the project history.
198 * "tree" objects represent the state of a single directory,
199 associating directory names to "blob" objects containing file
200 data and "tree" objects containing subdirectory information.
201 * "blob" objects contain file data without any other structure.
202 * References to commit objects at the head of each branch are
203 stored in files under .git/refs/heads/.
204 * The name of the current branch is stored in .git/HEAD.
206 Note, by the way, that lots of commands take a tree as an argument.
207 But as we can see above, a tree can be referred to in many different
208 ways--by the SHA1 name for that tree, by the name of a commit that
209 refers to the tree, by the name of a branch whose head refers to that
210 tree, etc.--and most such commands can accept any of these names.
212 In command synopses, the word "tree-ish" is sometimes used to
213 designate such an argument.
218 The primary tool we've been using to create commits is `git-commit
219 -a`, which creates a commit including every change you've made to
220 your working tree. But what if you want to commit changes only to
221 certain files? Or only certain changes to certain files?
223 If we look at the way commits are created under the cover, we'll see
224 that there are more flexible ways creating commits.
226 Continuing with our test-project, let's modify file.txt again:
228 ------------------------------------------------
229 $ echo "hello world, again" >>file.txt
230 ------------------------------------------------
232 but this time instead of immediately making the commit, let's take an
233 intermediate step, and ask for diffs along the way to keep track of
236 ------------------------------------------------
245 ------------------------------------------------
247 The last diff is empty, but no new commits have been made, and the
248 head still doesn't contain the new line:
250 ------------------------------------------------
252 diff --git a/file.txt b/file.txt
253 index a042389..513feba 100644
259 ------------------------------------------------
261 So 'git-diff' is comparing against something other than the head.
262 The thing that it's comparing against is actually the index file,
263 which is stored in .git/index in a binary format, but whose contents
264 we can examine with ls-files:
266 ------------------------------------------------
267 $ git ls-files --stage
268 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
269 $ git cat-file -t 513feba2
271 $ git cat-file blob 513feba2
274 ------------------------------------------------
276 So what our 'git-add' did was store a new blob and then put
277 a reference to it in the index file. If we modify the file again,
278 we'll see that the new modifications are reflected in the 'git-diff'
281 ------------------------------------------------
282 $ echo 'again?' >>file.txt
284 index 513feba..ba3da7b 100644
291 ------------------------------------------------
293 With the right arguments, 'git-diff' can also show us the difference
294 between the working directory and the last commit, or between the
295 index and the last commit:
297 ------------------------------------------------
299 diff --git a/file.txt b/file.txt
300 index a042389..ba3da7b 100644
308 diff --git a/file.txt b/file.txt
309 index a042389..513feba 100644
315 ------------------------------------------------
317 At any time, we can create a new commit using 'git-commit' (without
318 the "-a" option), and verify that the state committed only includes the
319 changes stored in the index file, not the additional change that is
320 still only in our working tree:
322 ------------------------------------------------
323 $ git commit -m "repeat"
325 diff --git a/file.txt b/file.txt
326 index 513feba..ba3da7b 100644
333 ------------------------------------------------
335 So by default 'git-commit' uses the index to create the commit, not
336 the working tree; the "-a" option to commit tells it to first update
337 the index with all changes in the working tree.
339 Finally, it's worth looking at the effect of 'git-add' on the index
342 ------------------------------------------------
343 $ echo "goodbye, world" >closing.txt
344 $ git add closing.txt
345 ------------------------------------------------
347 The effect of the 'git-add' was to add one entry to the index file:
349 ------------------------------------------------
350 $ git ls-files --stage
351 100644 8b9743b20d4b15be3955fc8d5cd2b09cd2336138 0 closing.txt
352 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
353 ------------------------------------------------
355 And, as you can see with cat-file, this new entry refers to the
356 current contents of the file:
358 ------------------------------------------------
359 $ git cat-file blob 8b9743b2
361 ------------------------------------------------
363 The "status" command is a useful way to get a quick summary of the
366 ------------------------------------------------
369 # Changes to be committed:
370 # (use "git reset HEAD <file>..." to unstage)
372 # new file: closing.txt
374 # Changed but not updated:
375 # (use "git add <file>..." to update what will be committed)
379 ------------------------------------------------
381 Since the current state of closing.txt is cached in the index file,
382 it is listed as "Changes to be committed". Since file.txt has
383 changes in the working directory that aren't reflected in the index,
384 it is marked "changed but not updated". At this point, running "git
385 commit" would create a commit that added closing.txt (with its new
386 contents), but that didn't modify file.txt.
388 Also, note that a bare `git diff` shows the changes to file.txt, but
389 not the addition of closing.txt, because the version of closing.txt
390 in the index file is identical to the one in the working directory.
392 In addition to being the staging area for new commits, the index file
393 is also populated from the object database when checking out a
394 branch, and is used to hold the trees involved in a merge operation.
395 See linkgit:gitcore-tutorial[7] and the relevant man
401 At this point you should know everything necessary to read the man
402 pages for any of the git commands; one good place to start would be
403 with the commands mentioned in link:everyday.html[Everyday git]. You
404 should be able to find any unknown jargon in linkgit:gitglossary[7].
406 The link:user-manual.html[Git User's Manual] provides a more
407 comprehensive introduction to git.
409 linkgit:gitcvs-migration[7] explains how to
410 import a CVS repository into git, and shows how to use git in a
413 For some interesting examples of git use, see the
414 link:howto-index.html[howtos].
416 For git developers, linkgit:gitcore-tutorial[7] goes
417 into detail on the lower-level git mechanisms involved in, for
418 example, creating a new commit.
422 linkgit:gittutorial[7],
423 linkgit:gitcvs-migration[7],
424 linkgit:gitcore-tutorial[7],
425 linkgit:gitglossary[7],
426 link:everyday.html[Everyday git],
427 link:user-manual.html[The Git User's Manual]
431 Part of the linkgit:git[1] suite.