1GITTUTORIAL-2(7) Git Manual GITTUTORIAL-2(7)
2
6 gittutorial-2 - A tutorial introduction to Git: part two
7
9 git *
10
13 You should work through gittutorial(7) before reading this tutorial.
14 The goal of this tutorial is to introduce two fundamental pieces of
16 Git’s architecture—the object database and the index file—and to
17 provide the reader with everything necessary to understand the rest of
18 the Git documentation.
19
21 Let’s start a new project and create a small amount of history:
22 $ mkdir test-project
24 $ cd test-project
25 $ git init
26 Initialized empty Git repository in .git/
27 $ echo 'hello world' > file.txt
28 $ git add .
29 $ git commit -a -m "initial commit"
30 [master (root-commit) 54196cc] initial commit
31 1 file changed, 1 insertion(+)
32 create mode 100644 file.txt
33 $ echo 'hello world!' >file.txt
34 $ git commit -a -m "add emphasis"
35 [master c4d59f3] add emphasis
36 1 file changed, 1 insertion(+), 1 deletion(-)
37 What are the 7 digits of hex that Git responded to the commit with?
40 We saw in part one of the tutorial that commits have names like this.
42 It turns out that every object in the Git history is stored under a
43 40-digit hex name. That name is the SHA-1 hash of the object’s
44 contents; among other things, this ensures that Git will never store
45 the same data twice (since identical data is given an identical SHA-1
46 name), and that the contents of a Git object will never change (since
47 that would change the object’s name as well). The 7 char hex strings
48 here are simply the abbreviation of such 40 character long strings.
49 Abbreviations can be used everywhere where the 40 character strings can
50 be used, so long as they are unambiguous.
51 It is expected that the content of the commit object you created while
53 following the example above generates a different SHA-1 hash than the
54 one shown above because the commit object records the time when it was
55 created and the name of the person performing the commit.
56 We can ask Git about this particular object with the cat-file command.
58 Don’t copy the 40 hex digits from this example but use those from your
59 own version. Note that you can shorten it to only a few characters to
60 save yourself typing all 40 hex digits:
61 $ git cat-file -t 54196cc2
63 commit
64 $ git cat-file commit 54196cc2
65 tree 92b8b694ffb1675e5975148e1121810081dbdffe
66 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
67 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
68 initial commit
70 A tree can refer to one or more "blob" objects, each corresponding to a
73 file. In addition, a tree can also refer to other tree objects, thus
74 creating a directory hierarchy. You can examine the contents of any
75 tree using ls-tree (remember that a long enough initial portion of the
76 SHA-1 will also work):
77 $ git ls-tree 92b8b694
79 100644 blob 3b18e512dba79e4c8300dd08aeb37f8e728b8dad file.txt
80 Thus we see that this tree has one file in it. The SHA-1 hash is a
83 reference to that file’s data:
84 $ git cat-file -t 3b18e512
86 blob
87 A "blob" is just file data, which we can also examine with cat-file:
90 $ git cat-file blob 3b18e512
92 hello world
93 Note that this is the old file data; so the object that Git named in
96 its response to the initial tree was a tree with a snapshot of the
97 directory state that was recorded by the first commit.
98 All of these objects are stored under their SHA-1 names inside the Git
100 directory:
101 $ find .git/objects/
103 .git/objects/
104 .git/objects/pack
105 .git/objects/info
106 .git/objects/3b
107 .git/objects/3b/18e512dba79e4c8300dd08aeb37f8e728b8dad
108 .git/objects/92
109 .git/objects/92/b8b694ffb1675e5975148e1121810081dbdffe
110 .git/objects/54
111 .git/objects/54/196cc2703dc165cbd373a65a4dcf22d50ae7f7
112 .git/objects/a0
113 .git/objects/a0/423896973644771497bdc03eb99d5281615b51
114 .git/objects/d0
115 .git/objects/d0/492b368b66bdabf2ac1fd8c92b39d3db916e59
116 .git/objects/c4
117 .git/objects/c4/d59f390b9cfd4318117afde11d601c1085f241
118 and the contents of these files is just the compressed data plus a
121 header identifying their length and their type. The type is either a
122 blob, a tree, a commit, or a tag.
123 The simplest commit to find is the HEAD commit, which we can find from
125 .git/HEAD:
126 $ cat .git/HEAD
128 ref: refs/heads/master
129 As you can see, this tells us which branch we’re currently on, and it
132 tells us this by naming a file under the .git directory, which itself
133 contains a SHA-1 name referring to a commit object, which we can
134 examine with cat-file:
135 $ cat .git/refs/heads/master
137 c4d59f390b9cfd4318117afde11d601c1085f241
138 $ git cat-file -t c4d59f39
139 commit
140 $ git cat-file commit c4d59f39
141 tree d0492b368b66bdabf2ac1fd8c92b39d3db916e59
142 parent 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
143 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
144 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
145 add emphasis
147 The "tree" object here refers to the new state of the tree:
150 $ git ls-tree d0492b36
152 100644 blob a0423896973644771497bdc03eb99d5281615b51 file.txt
153 $ git cat-file blob a0423896
154 hello world!
155 and the "parent" object refers to the previous commit:
158 $ git cat-file commit 54196cc2
160 tree 92b8b694ffb1675e5975148e1121810081dbdffe
161 author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
162 committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
163 initial commit
165 The tree object is the tree we examined first, and this commit is
168 unusual in that it lacks any parent.
169 Most commits have only one parent, but it is also common for a commit
171 to have multiple parents. In that case the commit represents a merge,
172 with the parent references pointing to the heads of the merged
173 branches.
174 Besides blobs, trees, and commits, the only remaining type of object is
176 a "tag", which we won’t discuss here; refer to git-tag(1) for details.
177 So now we know how Git uses the object database to represent a
179 project’s history:
180 · "commit" objects refer to "tree" objects representing the snapshot
182 of a directory tree at a particular point in the history, and refer
183 to "parent" commits to show how they’re connected into the project
184 history.
185 · "tree" objects represent the state of a single directory,
187 associating directory names to "blob" objects containing file data
188 and "tree" objects containing subdirectory information.
189 · "blob" objects contain file data without any other structure.
191 · References to commit objects at the head of each branch are stored
193 in files under .git/refs/heads/.
194 · The name of the current branch is stored in .git/HEAD.
196 Note, by the way, that lots of commands take a tree as an argument. But
198 as we can see above, a tree can be referred to in many different ways—
199 by the SHA-1 name for that tree, by the name of a commit that refers to
200 the tree, by the name of a branch whose head refers to that tree,
201 etc.--and most such commands can accept any of these names.
202 In command synopses, the word "tree-ish" is sometimes used to designate
204 such an argument.
205
207 The primary tool we’ve been using to create commits is git-commit -a,
208 which creates a commit including every change you’ve made to your
209 working tree. But what if you want to commit changes only to certain
210 files? Or only certain changes to certain files?
211 If we look at the way commits are created under the cover, we’ll see
213 that there are more flexible ways creating commits.
214 Continuing with our test-project, let’s modify file.txt again:
216 $ echo "hello world, again" >>file.txt
218 but this time instead of immediately making the commit, let’s take an
221 intermediate step, and ask for diffs along the way to keep track of
222 what’s happening:
223 $ git diff
225 --- a/file.txt
226 +++ b/file.txt
227 @@ -1 +1,2 @@
228 hello world!
229 +hello world, again
230 $ git add file.txt
231 $ git diff
232 The last diff is empty, but no new commits have been made, and the head
235 still doesn’t contain the new line:
236 $ git diff HEAD
238 diff --git a/file.txt b/file.txt
239 index a042389..513feba 100644
240 --- a/file.txt
241 +++ b/file.txt
242 @@ -1 +1,2 @@
243 hello world!
244 +hello world, again
245 So git diff is comparing against something other than the head. The
248 thing that it’s comparing against is actually the index file, which is
249 stored in .git/index in a binary format, but whose contents we can
250 examine with ls-files:
251 $ git ls-files --stage
253 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
254 $ git cat-file -t 513feba2
255 blob
256 $ git cat-file blob 513feba2
257 hello world!
258 hello world, again
259 So what our git add did was store a new blob and then put a reference
262 to it in the index file. If we modify the file again, we’ll see that
263 the new modifications are reflected in the git diff output:
264 $ echo 'again?' >>file.txt
266 $ git diff
267 index 513feba..ba3da7b 100644
268 --- a/file.txt
269 +++ b/file.txt
270 @@ -1,2 +1,3 @@
271 hello world!
272 hello world, again
273 +again?
274 With the right arguments, git diff can also show us the difference
277 between the working directory and the last commit, or between the index
278 and the last commit:
279 $ git diff HEAD
281 diff --git a/file.txt b/file.txt
282 index a042389..ba3da7b 100644
283 --- a/file.txt
284 +++ b/file.txt
285 @@ -1 +1,3 @@
286 hello world!
287 +hello world, again
288 +again?
289 $ git diff --cached
290 diff --git a/file.txt b/file.txt
291 index a042389..513feba 100644
292 --- a/file.txt
293 +++ b/file.txt
294 @@ -1 +1,2 @@
295 hello world!
296 +hello world, again
297 At any time, we can create a new commit using git commit (without the
300 "-a" option), and verify that the state committed only includes the
301 changes stored in the index file, not the additional change that is
302 still only in our working tree:
303 $ git commit -m "repeat"
305 $ git diff HEAD
306 diff --git a/file.txt b/file.txt
307 index 513feba..ba3da7b 100644
308 --- a/file.txt
309 +++ b/file.txt
310 @@ -1,2 +1,3 @@
311 hello world!
312 hello world, again
313 +again?
314 So by default git commit uses the index to create the commit, not the
317 working tree; the "-a" option to commit tells it to first update the
318 index with all changes in the working tree.
319 Finally, it’s worth looking at the effect of git add on the index file:
321 $ echo "goodbye, world" >closing.txt
323 $ git add closing.txt
324 The effect of the git add was to add one entry to the index file:
327 $ git ls-files --stage
329 100644 8b9743b20d4b15be3955fc8d5cd2b09cd2336138 0 closing.txt
330 100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt
331 And, as you can see with cat-file, this new entry refers to the current
334 contents of the file:
335 $ git cat-file blob 8b9743b2
337 goodbye, world
338 The "status" command is a useful way to get a quick summary of the
341 situation:
342 $ git status
344 On branch master
345 Changes to be committed:
346 (use "git restore --staged <file>..." to unstage)
347 new file: closing.txt
349 Changes not staged for commit:
351 (use "git add <file>..." to update what will be committed)
352 (use "git restore <file>..." to discard changes in working directory)
353 modified: file.txt
355 Since the current state of closing.txt is cached in the index file, it
358 is listed as "Changes to be committed". Since file.txt has changes in
359 the working directory that aren’t reflected in the index, it is marked
360 "changed but not updated". At this point, running "git commit" would
361 create a commit that added closing.txt (with its new contents), but
362 that didn’t modify file.txt.
363 Also, note that a bare git diff shows the changes to file.txt, but not
365 the addition of closing.txt, because the version of closing.txt in the
366 index file is identical to the one in the working directory.
367 In addition to being the staging area for new commits, the index file
369 is also populated from the object database when checking out a branch,
370 and is used to hold the trees involved in a merge operation. See
371 gitcore-tutorial(7) and the relevant man pages for details.
372
374 At this point you should know everything necessary to read the man
375 pages for any of the git commands; one good place to start would be
376 with the commands mentioned in giteveryday(7). You should be able to
377 find any unknown jargon in gitglossary(7).
378 The Git User’s Manual[1] provides a more comprehensive introduction to
380 Git.
381 gitcvs-migration(7) explains how to import a CVS repository into Git,
383 and shows how to use Git in a CVS-like way.
384 For some interesting examples of Git use, see the howtos[2].
386 For Git developers, gitcore-tutorial(7) goes into detail on the
388 lower-level Git mechanisms involved in, for example, creating a new
389 commit.
390
392 gittutorial(7), gitcvs-migration(7), gitcore-tutorial(7),
393 gitglossary(7), git-help(1), giteveryday(7), The Git User’s Manual[1]
394
396 Part of the git(1) suite
397
399 1. Git User’s Manual
400 file:///usr/share/doc/git/user-manual.html
401 2. howtos
403 file:///usr/share/doc/git/howto-index.html
404Git 2.24.1 12/10/2019 GITTUTORIAL-2(7)