Documentation / gittutorial-2.txton commit Merge branch 'maint-1.5.6' into maint-1.6.0 (141201d)
   1gittutorial-2(7)
   2================
   3
   4NAME
   5----
   6gittutorial-2 - A tutorial introduction to git: part two
   7
   8SYNOPSIS
   9--------
  10git *
  11
  12DESCRIPTION
  13-----------
  14
  15You should work through linkgit:gittutorial[7] before reading this tutorial.
  16
  17The goal of this tutorial is to introduce two fundamental pieces of
  18git's architecture--the object database and the index file--and to
  19provide the reader with everything necessary to understand the rest
  20of the git documentation.
  21
  22The git object database
  23-----------------------
  24
  25Let's start a new project and create a small amount of history:
  26
  27------------------------------------------------
  28$ mkdir test-project
  29$ cd test-project
  30$ git init
  31Initialized empty Git repository in .git/
  32$ echo 'hello world' > file.txt
  33$ git add .
  34$ git commit -a -m "initial commit"
  35Created initial commit 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
  36 create mode 100644 file.txt
  37$ echo 'hello world!' >file.txt
  38$ git commit -a -m "add emphasis"
  39Created commit c4d59f390b9cfd4318117afde11d601c1085f241
  40------------------------------------------------
  41
  42What are the 40 digits of hex that git responded to the commit with?
  43
  44We saw in part one of the tutorial that commits have names like this.
  45It turns out that every object in the git history is stored under
  46such a 40-digit hex name.  That name is the SHA1 hash of the object's
  47contents; among other things, this ensures that git will never store
  48the same data twice (since identical data is given an identical SHA1
  49name), and that the contents of a git object will never change (since
  50that would change the object's name as well).
  51
  52It is expected that the content of the commit object you created while
  53following the example above generates a different SHA1 hash than
  54the one shown above because the commit object records the time when
  55it was created and the name of the person performing the commit.
  56
  57We can ask git about this particular object with the `cat-file`
  58command. Don't copy the 40 hex digits from this example but use those
  59from your own version. Note that you can shorten it to only a few
  60characters to save yourself typing all 40 hex digits:
  61
  62------------------------------------------------
  63$ git cat-file -t 54196cc2
  64commit
  65$ git cat-file commit 54196cc2
  66tree 92b8b694ffb1675e5975148e1121810081dbdffe
  67author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
  68committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
  69
  70initial commit
  71------------------------------------------------
  72
  73A tree can refer to one or more "blob" objects, each corresponding to
  74a file.  In addition, a tree can also refer to other tree objects,
  75thus creating a directory hierarchy.  You can examine the contents of
  76any tree using ls-tree (remember that a long enough initial portion
  77of the SHA1 will also work):
  78
  79------------------------------------------------
  80$ git ls-tree 92b8b694
  81100644 blob 3b18e512dba79e4c8300dd08aeb37f8e728b8dad    file.txt
  82------------------------------------------------
  83
  84Thus we see that this tree has one file in it.  The SHA1 hash is a
  85reference to that file's data:
  86
  87------------------------------------------------
  88$ git cat-file -t 3b18e512
  89blob
  90------------------------------------------------
  91
  92A "blob" is just file data, which we can also examine with cat-file:
  93
  94------------------------------------------------
  95$ git cat-file blob 3b18e512
  96hello world
  97------------------------------------------------
  98
  99Note that this is the old file data; so the object that git named in
 100its response to the initial tree was a tree with a snapshot of the
 101directory state that was recorded by the first commit.
 102
 103All of these objects are stored under their SHA1 names inside the git
 104directory:
 105
 106------------------------------------------------
 107$ find .git/objects/
 108.git/objects/
 109.git/objects/pack
 110.git/objects/info
 111.git/objects/3b
 112.git/objects/3b/18e512dba79e4c8300dd08aeb37f8e728b8dad
 113.git/objects/92
 114.git/objects/92/b8b694ffb1675e5975148e1121810081dbdffe
 115.git/objects/54
 116.git/objects/54/196cc2703dc165cbd373a65a4dcf22d50ae7f7
 117.git/objects/a0
 118.git/objects/a0/423896973644771497bdc03eb99d5281615b51
 119.git/objects/d0
 120.git/objects/d0/492b368b66bdabf2ac1fd8c92b39d3db916e59
 121.git/objects/c4
 122.git/objects/c4/d59f390b9cfd4318117afde11d601c1085f241
 123------------------------------------------------
 124
 125and the contents of these files is just the compressed data plus a
 126header identifying their length and their type.  The type is either a
 127blob, a tree, a commit, or a tag.
 128
 129The simplest commit to find is the HEAD commit, which we can find
 130from .git/HEAD:
 131
 132------------------------------------------------
 133$ cat .git/HEAD
 134ref: refs/heads/master
 135------------------------------------------------
 136
 137As you can see, this tells us which branch we're currently on, and it
 138tells us this by naming a file under the .git directory, which itself
 139contains a SHA1 name referring to a commit object, which we can
 140examine with cat-file:
 141
 142------------------------------------------------
 143$ cat .git/refs/heads/master
 144c4d59f390b9cfd4318117afde11d601c1085f241
 145$ git cat-file -t c4d59f39
 146commit
 147$ git cat-file commit c4d59f39
 148tree d0492b368b66bdabf2ac1fd8c92b39d3db916e59
 149parent 54196cc2703dc165cbd373a65a4dcf22d50ae7f7
 150author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
 151committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143418702 -0500
 152
 153add emphasis
 154------------------------------------------------
 155
 156The "tree" object here refers to the new state of the tree:
 157
 158------------------------------------------------
 159$ git ls-tree d0492b36
 160100644 blob a0423896973644771497bdc03eb99d5281615b51    file.txt
 161$ git cat-file blob a0423896
 162hello world!
 163------------------------------------------------
 164
 165and the "parent" object refers to the previous commit:
 166
 167------------------------------------------------
 168$ git cat-file commit 54196cc2
 169tree 92b8b694ffb1675e5975148e1121810081dbdffe
 170author J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
 171committer J. Bruce Fields <bfields@puzzle.fieldses.org> 1143414668 -0500
 172
 173initial commit
 174------------------------------------------------
 175
 176The tree object is the tree we examined first, and this commit is
 177unusual in that it lacks any parent.
 178
 179Most commits have only one parent, but it is also common for a commit
 180to have multiple parents.   In that case the commit represents a
 181merge, with the parent references pointing to the heads of the merged
 182branches.
 183
 184Besides blobs, trees, and commits, the only remaining type of object
 185is a "tag", which we won't discuss here; refer to linkgit:git-tag[1]
 186for details.
 187
 188So now we know how git uses the object database to represent a
 189project's history:
 190
 191  * "commit" objects refer to "tree" objects representing the
 192    snapshot of a directory tree at a particular point in the
 193    history, and refer to "parent" commits to show how they're
 194    connected into the project history.
 195  * "tree" objects represent the state of a single directory,
 196    associating directory names to "blob" objects containing file
 197    data and "tree" objects containing subdirectory information.
 198  * "blob" objects contain file data without any other structure.
 199  * References to commit objects at the head of each branch are
 200    stored in files under .git/refs/heads/.
 201  * The name of the current branch is stored in .git/HEAD.
 202
 203Note, by the way, that lots of commands take a tree as an argument.
 204But as we can see above, a tree can be referred to in many different
 205ways--by the SHA1 name for that tree, by the name of a commit that
 206refers to the tree, by the name of a branch whose head refers to that
 207tree, etc.--and most such commands can accept any of these names.
 208
 209In command synopses, the word "tree-ish" is sometimes used to
 210designate such an argument.
 211
 212The index file
 213--------------
 214
 215The primary tool we've been using to create commits is `git-commit
 216-a`, which creates a commit including every change you've made to
 217your working tree.  But what if you want to commit changes only to
 218certain files?  Or only certain changes to certain files?
 219
 220If we look at the way commits are created under the cover, we'll see
 221that there are more flexible ways creating commits.
 222
 223Continuing with our test-project, let's modify file.txt again:
 224
 225------------------------------------------------
 226$ echo "hello world, again" >>file.txt
 227------------------------------------------------
 228
 229but this time instead of immediately making the commit, let's take an
 230intermediate step, and ask for diffs along the way to keep track of
 231what's happening:
 232
 233------------------------------------------------
 234$ git diff
 235--- a/file.txt
 236+++ b/file.txt
 237@@ -1 +1,2 @@
 238 hello world!
 239+hello world, again
 240$ git add file.txt
 241$ git diff
 242------------------------------------------------
 243
 244The last diff is empty, but no new commits have been made, and the
 245head still doesn't contain the new line:
 246
 247------------------------------------------------
 248$ git diff HEAD
 249diff --git a/file.txt b/file.txt
 250index a042389..513feba 100644
 251--- a/file.txt
 252+++ b/file.txt
 253@@ -1 +1,2 @@
 254 hello world!
 255+hello world, again
 256------------------------------------------------
 257
 258So 'git-diff' is comparing against something other than the head.
 259The thing that it's comparing against is actually the index file,
 260which is stored in .git/index in a binary format, but whose contents
 261we can examine with ls-files:
 262
 263------------------------------------------------
 264$ git ls-files --stage
 265100644 513feba2e53ebbd2532419ded848ba19de88ba00 0       file.txt
 266$ git cat-file -t 513feba2
 267blob
 268$ git cat-file blob 513feba2
 269hello world!
 270hello world, again
 271------------------------------------------------
 272
 273So what our 'git-add' did was store a new blob and then put
 274a reference to it in the index file.  If we modify the file again,
 275we'll see that the new modifications are reflected in the 'git-diff'
 276output:
 277
 278------------------------------------------------
 279$ echo 'again?' >>file.txt
 280$ git diff
 281index 513feba..ba3da7b 100644
 282--- a/file.txt
 283+++ b/file.txt
 284@@ -1,2 +1,3 @@
 285 hello world!
 286 hello world, again
 287+again?
 288------------------------------------------------
 289
 290With the right arguments, 'git-diff' can also show us the difference
 291between the working directory and the last commit, or between the
 292index and the last commit:
 293
 294------------------------------------------------
 295$ git diff HEAD
 296diff --git a/file.txt b/file.txt
 297index a042389..ba3da7b 100644
 298--- a/file.txt
 299+++ b/file.txt
 300@@ -1 +1,3 @@
 301 hello world!
 302+hello world, again
 303+again?
 304$ git diff --cached
 305diff --git a/file.txt b/file.txt
 306index a042389..513feba 100644
 307--- a/file.txt
 308+++ b/file.txt
 309@@ -1 +1,2 @@
 310 hello world!
 311+hello world, again
 312------------------------------------------------
 313
 314At any time, we can create a new commit using 'git-commit' (without
 315the "-a" option), and verify that the state committed only includes the
 316changes stored in the index file, not the additional change that is
 317still only in our working tree:
 318
 319------------------------------------------------
 320$ git commit -m "repeat"
 321$ git diff HEAD
 322diff --git a/file.txt b/file.txt
 323index 513feba..ba3da7b 100644
 324--- a/file.txt
 325+++ b/file.txt
 326@@ -1,2 +1,3 @@
 327 hello world!
 328 hello world, again
 329+again?
 330------------------------------------------------
 331
 332So by default 'git-commit' uses the index to create the commit, not
 333the working tree; the "-a" option to commit tells it to first update
 334the index with all changes in the working tree.
 335
 336Finally, it's worth looking at the effect of 'git-add' on the index
 337file:
 338
 339------------------------------------------------
 340$ echo "goodbye, world" >closing.txt
 341$ git add closing.txt
 342------------------------------------------------
 343
 344The effect of the 'git-add' was to add one entry to the index file:
 345
 346------------------------------------------------
 347$ git ls-files --stage
 348100644 8b9743b20d4b15be3955fc8d5cd2b09cd2336138 0       closing.txt
 349100644 513feba2e53ebbd2532419ded848ba19de88ba00 0       file.txt
 350------------------------------------------------
 351
 352And, as you can see with cat-file, this new entry refers to the
 353current contents of the file:
 354
 355------------------------------------------------
 356$ git cat-file blob 8b9743b2
 357goodbye, world
 358------------------------------------------------
 359
 360The "status" command is a useful way to get a quick summary of the
 361situation:
 362
 363------------------------------------------------
 364$ git status
 365# On branch master
 366# Changes to be committed:
 367#   (use "git reset HEAD <file>..." to unstage)
 368#
 369#       new file: closing.txt
 370#
 371# Changed but not updated:
 372#   (use "git add <file>..." to update what will be committed)
 373#
 374#       modified: file.txt
 375#
 376------------------------------------------------
 377
 378Since the current state of closing.txt is cached in the index file,
 379it is listed as "Changes to be committed".  Since file.txt has
 380changes in the working directory that aren't reflected in the index,
 381it is marked "changed but not updated".  At this point, running "git
 382commit" would create a commit that added closing.txt (with its new
 383contents), but that didn't modify file.txt.
 384
 385Also, note that a bare `git diff` shows the changes to file.txt, but
 386not the addition of closing.txt, because the version of closing.txt
 387in the index file is identical to the one in the working directory.
 388
 389In addition to being the staging area for new commits, the index file
 390is also populated from the object database when checking out a
 391branch, and is used to hold the trees involved in a merge operation.
 392See linkgit:gitcore-tutorial[7] and the relevant man
 393pages for details.
 394
 395What next?
 396----------
 397
 398At this point you should know everything necessary to read the man
 399pages for any of the git commands; one good place to start would be
 400with the commands mentioned in link:everyday.html[Everyday git].  You
 401should be able to find any unknown jargon in linkgit:gitglossary[7].
 402
 403The link:user-manual.html[Git User's Manual] provides a more
 404comprehensive introduction to git.
 405
 406linkgit:gitcvs-migration[7] explains how to
 407import a CVS repository into git, and shows how to use git in a
 408CVS-like way.
 409
 410For some interesting examples of git use, see the
 411link:howto-index.html[howtos].
 412
 413For git developers, linkgit:gitcore-tutorial[7] goes
 414into detail on the lower-level git mechanisms involved in, for
 415example, creating a new commit.
 416
 417SEE ALSO
 418--------
 419linkgit:gittutorial[7],
 420linkgit:gitcvs-migration[7],
 421linkgit:gitcore-tutorial[7],
 422linkgit:gitglossary[7],
 423link:everyday.html[Everyday git],
 424link:user-manual.html[The Git User's Manual]
 425
 426GIT
 427---
 428Part of the linkgit:git[1] suite.