6 gittutorial-2 - A tutorial introduction to git: part two
15 You should work through linkgit:gittutorial[7][A tutorial introduction to
16 git] before reading this tutorial.
18 The goal of this tutorial is to introduce two fundamental pieces of
19 git's architecture--the object database and the index file--and to
20 provide the reader with everything necessary to understand the rest
21 of the git documentation.
23 The git object database
24 -----------------------
26 Let's start a new project and create a small amount of history:
28 ------------------------------------------------
32 Initialized empty Git repository in .git/
33 $ echo 'hello world' > file.txt
35 $ git commit -a -m "initial commit"
36 Created initial commit 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
37 create mode 100644 file.txt
38 $ echo 'hello world!' >file.txt
39 $ git commit -a -m "add emphasis"
40 Created commit c4d59f390b9cfd4318117afde11d601c1085f241
41 ------------------------------------------------
43 What are the 40 digits of hex that git responded to the commit with?
45 We saw in part one of the tutorial that commits have names like this.
46 It turns out that every object in the git history is stored under
47 such a 40-digit hex name. That name is the SHA1 hash of the object's
48 contents; among other things, this ensures that git will never store
49 the same data twice (since identical data is given an identical SHA1
50 name), and that the contents of a git object will never change (since
51 that would change the object's name as well).
53 It is expected that the content of the commit object you created while
54 following the example above generates a different SHA1 hash than
55 the one shown above because the commit object records the time when
56 it was created and the name of the person performing the commit.
58 We can ask git about this particular object with the cat-file
59 command. Don't copy the 40 hex digits from this example but use those
60 from your own version. Note that you can shorten it to only a few
61 characters to save yourself typing all 40 hex digits:
63 ------------------------------------------------
64 $ git-cat-file -t 54196cc2
66 $ git-cat-file commit 54196cc2
67 tree 92b8b694ffb1675e5975148e1121810081dbdffe
68 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
69 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
72 ------------------------------------------------
74 A tree can refer to one or more "blob" objects, each corresponding to
75 a file. In addition, a tree can also refer to other tree objects,
76 thus creating a directory hierarchy. You can examine the contents of
77 any tree using ls-tree (remember that a long enough initial portion
78 of the SHA1 will also work):
80 ------------------------------------------------
81 $ git ls-tree 92b8b694
82 100644 blob 3b18e512dba79e4c8300dd08aeb37f8e728b8dad file.txt
83 ------------------------------------------------
85 Thus we see that this tree has one file in it. The SHA1 hash is a
86 reference to that file's data:
88 ------------------------------------------------
89 $ git cat-file -t 3b18e512
91 ------------------------------------------------
93 A "blob" is just file data, which we can also examine with cat-file:
95 ------------------------------------------------
96 $ git cat-file blob 3b18e512
98 ------------------------------------------------
100 Note that this is the old file data; so the object that git named in
101 its response to the initial tree was a tree with a snapshot of the
102 directory state that was recorded by the first commit.
104 All of these objects are stored under their SHA1 names inside the git
107 ------------------------------------------------
113 .git/objects/3b/18e512dba79e4c8300dd08aeb37f8e728b8dad
115 .git/objects/92/b8b694ffb1675e5975148e1121810081dbdffe
117 .git/objects/54/196cc2703dc165cbd373a65a4dcf22d50ae7f7
119 .git/objects/a0/423896973644771497bdc03eb99d5281615b51
121 .git/objects/d0/492b368b66bdabf2ac1fd8c92b39d3db916e59
123 .git/objects/c4/d59f390b9cfd4318117afde11d601c1085f241
124 ------------------------------------------------
126 and the contents of these files is just the compressed data plus a
127 header identifying their length and their type. The type is either a
128 blob, a tree, a commit, or a tag.
130 The simplest commit to find is the HEAD commit, which we can find
133 ------------------------------------------------
135 ref: refs/heads/master
136 ------------------------------------------------
138 As you can see, this tells us which branch we're currently on, and it
139 tells us this by naming a file under the .git directory, which itself
140 contains a SHA1 name referring to a commit object, which we can
141 examine with cat-file:
143 ------------------------------------------------
144 $ cat .git/refs/heads/master
145 c4d59f390b9cfd4318117afde11d601c1085f241
146 $ git cat-file -t c4d59f39
148 $ git cat-file commit c4d59f39
149 tree d0492b368b66bdabf2ac1fd8c92b39d3db916e59
150 parent 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
151 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
152 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
155 ------------------------------------------------
157 The "tree" object here refers to the new state of the tree:
159 ------------------------------------------------
160 $ git ls-tree d0492b36
161 100644 blob a0423896973644771497bdc03eb99d5281615b51 file.txt
162 $ git cat-file blob a0423896
164 ------------------------------------------------
166 and the "parent" object refers to the previous commit:
168 ------------------------------------------------
169 $ git-cat-file commit 54196cc2
170 tree 92b8b694ffb1675e5975148e1121810081dbdffe
171 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
172 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
175 ------------------------------------------------
177 The tree object is the tree we examined first, and this commit is
178 unusual in that it lacks any parent.
180 Most commits have only one parent, but it is also common for a commit
181 to have multiple parents. In that case the commit represents a
182 merge, with the parent references pointing to the heads of the merged
185 Besides blobs, trees, and commits, the only remaining type of object
186 is a "tag", which we won't discuss here; refer to linkgit:git-tag[1]
189 So now we know how git uses the object database to represent a
192 * "commit" objects refer to "tree" objects representing the
193 snapshot of a directory tree at a particular point in the
194 history, and refer to "parent" commits to show how they're
195 connected into the project history.
196 * "tree" objects represent the state of a single directory,
197 associating directory names to "blob" objects containing file
198 data and "tree" objects containing subdirectory information.
199 * "blob" objects contain file data without any other structure.
200 * References to commit objects at the head of each branch are
201 stored in files under .git/refs/heads/.
202 * The name of the current branch is stored in .git/HEAD.
204 Note, by the way, that lots of commands take a tree as an argument.
205 But as we can see above, a tree can be referred to in many different
206 ways--by the SHA1 name for that tree, by the name of a commit that
207 refers to the tree, by the name of a branch whose head refers to that
208 tree, etc.--and most such commands can accept any of these names.
210 In command synopses, the word "tree-ish" is sometimes used to
211 designate such an argument.
216 The primary tool we've been using to create commits is "git commit
217 -a", which creates a commit including every change you've made to
218 your working tree. But what if you want to commit changes only to
219 certain files? Or only certain changes to certain files?
221 If we look at the way commits are created under the cover, we'll see
222 that there are more flexible ways creating commits.
224 Continuing with our test-project, let's modify file.txt again:
226 ------------------------------------------------
227 $ echo "hello world, again" >>file.txt
228 ------------------------------------------------
230 but this time instead of immediately making the commit, let's take an
231 intermediate step, and ask for diffs along the way to keep track of
234 ------------------------------------------------
243 ------------------------------------------------
245 The last diff is empty, but no new commits have been made, and the
246 head still doesn't contain the new line:
248 ------------------------------------------------
250 diff --git a/file.txt b/file.txt
251 index a042389..513feba 100644
257 ------------------------------------------------
259 So "git diff" is comparing against something other than the head.
260 The thing that it's comparing against is actually the index file,
261 which is stored in .git/index in a binary format, but whose contents
262 we can examine with ls-files:
264 ------------------------------------------------
265 $ git ls-files --stage
266 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
267 $ git cat-file -t 513feba2
269 $ git cat-file blob 513feba2
272 ------------------------------------------------
274 So what our "git add" did was store a new blob and then put
275 a reference to it in the index file. If we modify the file again,
276 we'll see that the new modifications are reflected in the "git-diff"
279 ------------------------------------------------
280 $ echo 'again?' >>file.txt
282 index 513feba..ba3da7b 100644
289 ------------------------------------------------
291 With the right arguments, git diff can also show us the difference
292 between the working directory and the last commit, or between the
293 index and the last commit:
295 ------------------------------------------------
297 diff --git a/file.txt b/file.txt
298 index a042389..ba3da7b 100644
306 diff --git a/file.txt b/file.txt
307 index a042389..513feba 100644
313 ------------------------------------------------
315 At any time, we can create a new commit using "git commit" (without
316 the -a option), and verify that the state committed only includes the
317 changes stored in the index file, not the additional change that is
318 still only in our working tree:
320 ------------------------------------------------
321 $ git commit -m "repeat"
323 diff --git a/file.txt b/file.txt
324 index 513feba..ba3da7b 100644
331 ------------------------------------------------
333 So by default "git commit" uses the index to create the commit, not
334 the working tree; the -a option to commit tells it to first update
335 the index with all changes in the working tree.
337 Finally, it's worth looking at the effect of "git add" on the index
340 ------------------------------------------------
341 $ echo "goodbye, world" >closing.txt
342 $ git add closing.txt
343 ------------------------------------------------
345 The effect of the "git add" was to add one entry to the index file:
347 ------------------------------------------------
348 $ git ls-files --stage
349 100644 8b9743b20d4b15be3955fc8d5cd2b09cd2336138 0 closing.txt
350 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
351 ------------------------------------------------
353 And, as you can see with cat-file, this new entry refers to the
354 current contents of the file:
356 ------------------------------------------------
357 $ git cat-file blob 8b9743b2
359 ------------------------------------------------
361 The "status" command is a useful way to get a quick summary of the
364 ------------------------------------------------
367 # Changes to be committed:
368 # (use "git reset HEAD <file>..." to unstage)
370 # new file: closing.txt
372 # Changed but not updated:
373 # (use "git add <file>..." to update what will be committed)
377 ------------------------------------------------
379 Since the current state of closing.txt is cached in the index file,
380 it is listed as "Changes to be committed". Since file.txt has
381 changes in the working directory that aren't reflected in the index,
382 it is marked "changed but not updated". At this point, running "git
383 commit" would create a commit that added closing.txt (with its new
384 contents), but that didn't modify file.txt.
386 Also, note that a bare "git diff" shows the changes to file.txt, but
387 not the addition of closing.txt, because the version of closing.txt
388 in the index file is identical to the one in the working directory.
390 In addition to being the staging area for new commits, the index file
391 is also populated from the object database when checking out a
392 branch, and is used to hold the trees involved in a merge operation.
393 See the linkgit:gitcore-tutorial[7][core tutorial] and the relevant man
399 At this point you should know everything necessary to read the man
400 pages for any of the git commands; one good place to start would be
401 with the commands mentioned in link:everyday.html[Everyday git]. You
402 should be able to find any unknown jargon in the
403 linkgit:gitglossary[7][Glossary].
405 The link:user-manual.html[Git User's Manual] provides a more
406 comprehensive introduction to git.
408 The linkgit:gitcvs-migration[7][CVS migration] document explains how to
409 import a CVS repository into git, and shows how to use git in a
412 For some interesting examples of git use, see the
413 link:howto-index.html[howtos].
415 For git developers, the linkgit:gitcore-tutorial[7][Core tutorial] goes
416 into detail on the lower-level git mechanisms involved in, for
417 example, creating a new commit.
421 linkgit:gittutorial[7],
422 linkgit:gitcvs-migration[7],
423 linkgit:gitcore-tutorial[7],
424 linkgit:gitglossary[7],
425 link:everyday.html[Everyday git],
426 link:user-manual.html[The Git User's Manual]
430 Part of the linkgit:git[1] suite.