1gittutorial-2(7) 2================ 3 4NAME 5---- 6gittutorial-2 - A tutorial introduction to git: part two 7 8SYNOPSIS 9-------- 10[verse] 11git * 12 13DESCRIPTION 14----------- 15 16You should work through linkgit:gittutorial[7] before reading this tutorial. 17 18The goal of this tutorial is to introduce two fundamental pieces of 19git's architecture--the object database and the index file--and to 20provide the reader with everything necessary to understand the rest 21of the git documentation. 22 23The git object database 24----------------------- 25 26Let's start a new project and create a small amount of history: 27 28------------------------------------------------ 29$ mkdir test-project 30$ cd test-project 31$ git init 32Initialized empty Git repository in .git/ 33$ echo 'hello world' > file.txt 34$ git add . 35$ git commit -a -m "initial commit" 36[master (root-commit) 54196cc] initial commit 37 1 files changed, 1 insertions(+), 0 deletions(-) 38 create mode 100644 file.txt 39$ echo 'hello world!' >file.txt 40$ git commit -a -m "add emphasis" 41[master c4d59f3] add emphasis 42 1 files changed, 1 insertions(+), 1 deletions(-) 43------------------------------------------------ 44 45What are the 7 digits of hex that git responded to the commit with? 46 47We saw in part one of the tutorial that commits have names like this. 48It turns out that every object in the git history is stored under 49a 40-digit hex name. That name is the SHA1 hash of the object's 50contents; among other things, this ensures that git will never store 51the same data twice (since identical data is given an identical SHA1 52name), and that the contents of a git object will never change (since 53that would change the object's name as well). The 7 char hex strings 54here are simply the abbreviation of such 40 character long strings. 55Abbreviations can be used everywhere where the 40 character strings 56can be used, so long as they are unambiguous. 57 58It is expected that the content of the commit object you created while 59following the example above generates a different SHA1 hash than 60the one shown above because the commit object records the time when 61it was created and the name of the person performing the commit. 62 63We can ask git about this particular object with the `cat-file` 64command. Don't copy the 40 hex digits from this example but use those 65from your own version. Note that you can shorten it to only a few 66characters to save yourself typing all 40 hex digits: 67 68------------------------------------------------ 69$ git cat-file -t 54196cc2 70commit 71$ git cat-file commit 54196cc2 72tree 92b8b694ffb1675e5975148e1121810081dbdffe 73author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500 74committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500 75 76initial commit 77------------------------------------------------ 78 79A tree can refer to one or more "blob" objects, each corresponding to 80a file. In addition, a tree can also refer to other tree objects, 81thus creating a directory hierarchy. You can examine the contents of 82any tree using ls-tree (remember that a long enough initial portion 83of the SHA1 will also work): 84 85------------------------------------------------ 86$ git ls-tree 92b8b694 87100644 blob 3b18e512dba79e4c8300dd08aeb37f8e728b8dad file.txt 88------------------------------------------------ 89 90Thus we see that this tree has one file in it. The SHA1 hash is a 91reference to that file's data: 92 93------------------------------------------------ 94$ git cat-file -t 3b18e512 95blob 96------------------------------------------------ 97 98A "blob" is just file data, which we can also examine with cat-file: 99 100------------------------------------------------ 101$ git cat-file blob 3b18e512 102hello world 103------------------------------------------------ 104 105Note that this is the old file data; so the object that git named in 106its response to the initial tree was a tree with a snapshot of the 107directory state that was recorded by the first commit. 108 109All of these objects are stored under their SHA1 names inside the git 110directory: 111 112------------------------------------------------ 113$ find .git/objects/ 114.git/objects/ 115.git/objects/pack 116.git/objects/info 117.git/objects/3b 118.git/objects/3b/18e512dba79e4c8300dd08aeb37f8e728b8dad 119.git/objects/92 120.git/objects/92/b8b694ffb1675e5975148e1121810081dbdffe 121.git/objects/54 122.git/objects/54/196cc2703dc165cbd373a65a4dcf22d50ae7f7 123.git/objects/a0 124.git/objects/a0/423896973644771497bdc03eb99d5281615b51 125.git/objects/d0 126.git/objects/d0/492b368b66bdabf2ac1fd8c92b39d3db916e59 127.git/objects/c4 128.git/objects/c4/d59f390b9cfd4318117afde11d601c1085f241 129------------------------------------------------ 130 131and the contents of these files is just the compressed data plus a 132header identifying their length and their type. The type is either a 133blob, a tree, a commit, or a tag. 134 135The simplest commit to find is the HEAD commit, which we can find 136from .git/HEAD: 137 138------------------------------------------------ 139$ cat .git/HEAD 140ref: refs/heads/master 141------------------------------------------------ 142 143As you can see, this tells us which branch we're currently on, and it 144tells us this by naming a file under the .git directory, which itself 145contains a SHA1 name referring to a commit object, which we can 146examine with cat-file: 147 148------------------------------------------------ 149$ cat .git/refs/heads/master 150c4d59f390b9cfd4318117afde11d601c1085f241 151$ git cat-file -t c4d59f39 152commit 153$ git cat-file commit c4d59f39 154tree d0492b368b66bdabf2ac1fd8c92b39d3db916e59 155parent 54196cc2703dc165cbd373a65a4dcf22d50ae7f7 156author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500 157committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500 158 159add emphasis 160------------------------------------------------ 161 162The "tree" object here refers to the new state of the tree: 163 164------------------------------------------------ 165$ git ls-tree d0492b36 166100644 blob a0423896973644771497bdc03eb99d5281615b51 file.txt 167$ git cat-file blob a0423896 168hello world! 169------------------------------------------------ 170 171and the "parent" object refers to the previous commit: 172 173------------------------------------------------ 174$ git cat-file commit 54196cc2 175tree 92b8b694ffb1675e5975148e1121810081dbdffe 176author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500 177committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500 178 179initial commit 180------------------------------------------------ 181 182The tree object is the tree we examined first, and this commit is 183unusual in that it lacks any parent. 184 185Most commits have only one parent, but it is also common for a commit 186to have multiple parents. In that case the commit represents a 187merge, with the parent references pointing to the heads of the merged 188branches. 189 190Besides blobs, trees, and commits, the only remaining type of object 191is a "tag", which we won't discuss here; refer to linkgit:git-tag[1] 192for details. 193 194So now we know how git uses the object database to represent a 195project's history: 196 197 * "commit" objects refer to "tree" objects representing the 198 snapshot of a directory tree at a particular point in the 199 history, and refer to "parent" commits to show how they're 200 connected into the project history. 201 * "tree" objects represent the state of a single directory, 202 associating directory names to "blob" objects containing file 203 data and "tree" objects containing subdirectory information. 204 * "blob" objects contain file data without any other structure. 205 * References to commit objects at the head of each branch are 206 stored in files under .git/refs/heads/. 207 * The name of the current branch is stored in .git/HEAD. 208 209Note, by the way, that lots of commands take a tree as an argument. 210But as we can see above, a tree can be referred to in many different 211ways--by the SHA1 name for that tree, by the name of a commit that 212refers to the tree, by the name of a branch whose head refers to that 213tree, etc.--and most such commands can accept any of these names. 214 215In command synopses, the word "tree-ish" is sometimes used to 216designate such an argument. 217 218The index file 219-------------- 220 221The primary tool we've been using to create commits is `git-commit 222-a`, which creates a commit including every change you've made to 223your working tree. But what if you want to commit changes only to 224certain files? Or only certain changes to certain files? 225 226If we look at the way commits are created under the cover, we'll see 227that there are more flexible ways creating commits. 228 229Continuing with our test-project, let's modify file.txt again: 230 231------------------------------------------------ 232$ echo "hello world, again" >>file.txt 233------------------------------------------------ 234 235but this time instead of immediately making the commit, let's take an 236intermediate step, and ask for diffs along the way to keep track of 237what's happening: 238 239------------------------------------------------ 240$ git diff 241--- a/file.txt 242+++ b/file.txt 243@@ -1 +1,2 @@ 244 hello world! 245+hello world, again 246$ git add file.txt 247$ git diff 248------------------------------------------------ 249 250The last diff is empty, but no new commits have been made, and the 251head still doesn't contain the new line: 252 253------------------------------------------------ 254$ git diff HEAD 255diff --git a/file.txt b/file.txt 256index a042389..513feba 100644 257--- a/file.txt 258+++ b/file.txt 259@@ -1 +1,2 @@ 260 hello world! 261+hello world, again 262------------------------------------------------ 263 264So 'git diff' is comparing against something other than the head. 265The thing that it's comparing against is actually the index file, 266which is stored in .git/index in a binary format, but whose contents 267we can examine with ls-files: 268 269------------------------------------------------ 270$ git ls-files --stage 271100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt 272$ git cat-file -t 513feba2 273blob 274$ git cat-file blob 513feba2 275hello world! 276hello world, again 277------------------------------------------------ 278 279So what our 'git add' did was store a new blob and then put 280a reference to it in the index file. If we modify the file again, 281we'll see that the new modifications are reflected in the 'git diff' 282output: 283 284------------------------------------------------ 285$ echo 'again?' >>file.txt 286$ git diff 287index 513feba..ba3da7b 100644 288--- a/file.txt 289+++ b/file.txt 290@@ -1,2 +1,3 @@ 291 hello world! 292 hello world, again 293+again? 294------------------------------------------------ 295 296With the right arguments, 'git diff' can also show us the difference 297between the working directory and the last commit, or between the 298index and the last commit: 299 300------------------------------------------------ 301$ git diff HEAD 302diff --git a/file.txt b/file.txt 303index a042389..ba3da7b 100644 304--- a/file.txt 305+++ b/file.txt 306@@ -1 +1,3 @@ 307 hello world! 308+hello world, again 309+again? 310$ git diff --cached 311diff --git a/file.txt b/file.txt 312index a042389..513feba 100644 313--- a/file.txt 314+++ b/file.txt 315@@ -1 +1,2 @@ 316 hello world! 317+hello world, again 318------------------------------------------------ 319 320At any time, we can create a new commit using 'git commit' (without 321the "-a" option), and verify that the state committed only includes the 322changes stored in the index file, not the additional change that is 323still only in our working tree: 324 325------------------------------------------------ 326$ git commit -m "repeat" 327$ git diff HEAD 328diff --git a/file.txt b/file.txt 329index 513feba..ba3da7b 100644 330--- a/file.txt 331+++ b/file.txt 332@@ -1,2 +1,3 @@ 333 hello world! 334 hello world, again 335+again? 336------------------------------------------------ 337 338So by default 'git commit' uses the index to create the commit, not 339the working tree; the "-a" option to commit tells it to first update 340the index with all changes in the working tree. 341 342Finally, it's worth looking at the effect of 'git add' on the index 343file: 344 345------------------------------------------------ 346$ echo "goodbye, world" >closing.txt 347$ git add closing.txt 348------------------------------------------------ 349 350The effect of the 'git add' was to add one entry to the index file: 351 352------------------------------------------------ 353$ git ls-files --stage 354100644 8b9743b20d4b15be3955fc8d5cd2b09cd2336138 0 closing.txt 355100644 513feba2e53ebbd2532419ded848ba19de88ba00 0 file.txt 356------------------------------------------------ 357 358And, as you can see with cat-file, this new entry refers to the 359current contents of the file: 360 361------------------------------------------------ 362$ git cat-file blob 8b9743b2 363goodbye, world 364------------------------------------------------ 365 366The "status" command is a useful way to get a quick summary of the 367situation: 368 369------------------------------------------------ 370$ git status 371# On branch master 372# Changes to be committed: 373# (use "git reset HEAD <file>..." to unstage) 374# 375# new file: closing.txt 376# 377# Changes not staged for commit: 378# (use "git add <file>..." to update what will be committed) 379# 380# modified: file.txt 381# 382------------------------------------------------ 383 384Since the current state of closing.txt is cached in the index file, 385it is listed as "Changes to be committed". Since file.txt has 386changes in the working directory that aren't reflected in the index, 387it is marked "changed but not updated". At this point, running "git 388commit" would create a commit that added closing.txt (with its new 389contents), but that didn't modify file.txt. 390 391Also, note that a bare `git diff` shows the changes to file.txt, but 392not the addition of closing.txt, because the version of closing.txt 393in the index file is identical to the one in the working directory. 394 395In addition to being the staging area for new commits, the index file 396is also populated from the object database when checking out a 397branch, and is used to hold the trees involved in a merge operation. 398See linkgit:gitcore-tutorial[7] and the relevant man 399pages for details. 400 401What next? 402---------- 403 404At this point you should know everything necessary to read the man 405pages for any of the git commands; one good place to start would be 406with the commands mentioned in link:everyday.html[Everyday git]. You 407should be able to find any unknown jargon in linkgit:gitglossary[7]. 408 409The link:user-manual.html[Git User's Manual] provides a more 410comprehensive introduction to git. 411 412linkgit:gitcvs-migration[7] explains how to 413import a CVS repository into git, and shows how to use git in a 414CVS-like way. 415 416For some interesting examples of git use, see the 417link:howto-index.html[howtos]. 418 419For git developers, linkgit:gitcore-tutorial[7] goes 420into detail on the lower-level git mechanisms involved in, for 421example, creating a new commit. 422 423SEE ALSO 424-------- 425linkgit:gittutorial[7], 426linkgit:gitcvs-migration[7], 427linkgit:gitcore-tutorial[7], 428linkgit:gitglossary[7], 429linkgit:git-help[1], 430link:everyday.html[Everyday git], 431link:user-manual.html[The Git User's Manual] 432 433GIT 434--- 435Part of the linkgit:git[1] suite.