git *

This tutorial explains how to import a new project into Git, makechanges to it, and share changes with other developers.

If you are instead primarily interested in using Git to fetch a project,for example, to test the latest version, you may prefer to start withthe first two chapters ofThe Git User’s Manual.

First, note that you can get documentation for a command such asgitlog--graph with:

or:

With the latter, you can use the manual viewer of your choice; seegit-help(1) for more information.

It is a good idea to introduce yourself to Git with your name andpublic email address before doing any operation. The easiestway to do so is:

Assume you have a tarballproject.tar.gz with your initial work. Youcan place it under Git revision control as follows.

Git will reply

You’ve now initialized the working directory—​you may notice a newdirectory created, named .git.

Next, tell Git to take a snapshot of the contents of all files under thecurrent directory (note the .), withgitadd:

This snapshot is now stored in a temporary staging area which Git callsthe "index". You can permanently store the contents of the index in therepository withgitcommit:

This will prompt you for a commit message. You’ve now stored the firstversion of your project in Git.

Modify some files, then add their updated contents to the index:

You are now ready to commit. You can see what is about to be committedusinggitdiff with the--cached option:

(Without--cached,gitdiff will show you any changes thatyou’ve made but not yet added to the index.) You can also get a briefsummary of the situation withgitstatus:

If you need to make any further adjustments, do so now, and then add anynewly modified content to the index. Finally, commit your changes with:

This will again prompt you for a message describing the change, and thenrecord a new version of the project.

Alternatively, instead of runninggitadd beforehand, you can use

which will automatically notice any modified (but not new) files, addthem to the index, and commit, all in one step.

A note on commit messages: Though not required, it’s a good idea tobegin the commit message with a single short (no more than 50characters) line summarizing the change, followed by a blank line andthen a more thorough description. The text up to the first blank line ina commit message is treated as the commit title, and that title is usedthroughout Git. For example,git-format-patch(1) turns acommit into email, and it uses the title on the Subject line and therest of the commit in the body.

Many revision control systems provide anadd command that tells thesystem to start tracking changes to a new file. Git’sadd commanddoes something simpler and more powerful:gitadd is used both for newand newly modified files, and in both cases it takes a snapshot of thegiven files and stages that content in the index, ready for inclusion inthe next commit.

At any point you can view the history of your changes using

If you also want to see complete diffs at each step, use

Often the overview of the change is useful to get a feel ofeach step

A single Git repository can maintain multiple branches ofdevelopment. To create a new branch namedexperimental, use

If you now run

you’ll get a list of all existing branches:

Theexperimental branch is the one you just created, and themaster branch is a default branch that was created for youautomatically. The asterisk marks the branch you are currently on;type

to switch to theexperimental branch. Now edit a file, commit thechange, and switch back to themaster branch:

Check that the change you made is no longer visible, since it wasmade on theexperimental branch and you’re back on themaster branch.

You can make a different change on themaster branch:

at this point the two branches have diverged, with different changesmade in each. To merge the changes made inexperimental intomaster, run

If the changes don’t conflict, you’re done. If there are conflicts,markers will be left in the problematic files showing the conflict;

will show this. Once you’ve edited the files to resolve theconflicts,

will commit the result of the merge. Finally,

will show a nice graphical representation of the resulting history.

At this point you could delete theexperimental branch with

This command ensures that the changes in theexperimental branch arealready in the current branch.

If you develop on a branchcrazy-idea, then regret it, you can alwaysdelete the branch with

Branches are cheap and easy, so this is a good way to try somethingout.

Suppose that Alice has started a new project with a Git repository in/home/alice/project, and that Bob, who has a home directory on thesame machine, wants to contribute.

Bob begins with:

This creates a new directorymyrepo containing a clone of Alice’srepository. The clone is on an equal footing with the originalproject, possessing its own copy of the original project’s history.

Bob then makes some changes and commits them:

When he’s ready, he tells Alice to pull changes from the repositoryat/home/bob/myrepo. She does this with:

This merges the changes from Bob’smaster branch into Alice’scurrent branch. If Alice has made her own changes in the meantime,then she may need to manually fix any conflicts.

Thepull command thus performs two operations: it fetches changesfrom a remote branch, then merges them into the current branch.

Note that in general, Alice would want her local changes committed beforeinitiating thispull. If Bob’s work conflicts with what Alice did sincetheir histories forked, Alice will use her working tree and the index toresolve conflicts, and existing local changes will interfere with theconflict resolution process (Git will still perform the fetch but willrefuse to merge — Alice will have to get rid of her local changes insome way and pull again when this happens).

Alice can peek at what Bob did without merging first, using thefetchcommand; this allows Alice to inspect what Bob did, using a specialsymbolFETCH_HEAD, in order to determine if he has anything worthpulling, like this:

This operation is safe even if Alice has uncommitted local changes.The range notationHEAD..FETCH_HEAD means "show everything that is reachablefrom theFETCH_HEAD but exclude anything that is reachable fromHEAD".Alice already knows everything that leads to her current state (HEAD),and reviews what Bob has in his state (FETCH_HEAD) that she has notseen with this command.

If Alice wants to visualize what Bob did since their histories forkedshe can issue the following command:

This uses the same two-dot range notation we saw earlier withgitlog.

Alice may want to view what both of them did since they forked.She can use three-dot form instead of the two-dot form:

This means "show everything that is reachable from either one, butexclude anything that is reachable from both of them".

Please note that these range notations can be used with bothgitkandgitlog.

After inspecting what Bob did, if there is nothing urgent, Alice maydecide to continue working without pulling from Bob. If Bob’s historydoes have something Alice would immediately need, Alice may choose tostash her work-in-progress first, do apull, and then finally unstashher work-in-progress on top of the resulting history.

When you are working in a small closely knit group, it is notunusual to interact with the same repository over and overagain. By definingremote repository shorthand, you can makeit easier:

With this, Alice can perform the first part of thepull operationalone using thegitfetch command without merging them with her ownbranch, using:

Unlike the longhand form, when Alice fetches from Bob using aremote repository shorthand set up withgitremote, what wasfetched is stored in a remote-tracking branch, in this casebob/master. So after this:

shows a list of all the changes that Bob made since he branched fromAlice’smaster branch.

After examining those changes, Alicecould merge the changes into hermaster branch:

Thismerge can also be done bypulling from her own remote-trackingbranch, like this:

Note that git pull always merges into the current branch,regardless of what else is given on the command line.

Later, Bob can update his repo with Alice’s latest changes using

Note that he doesn’t need to give the path to Alice’s repository;when Bob cloned Alice’s repository, Git stored the location of herrepository in the repository configuration, and that location isused for pulls:

(The complete configuration created bygitclone is visible usinggitconfig-l, and thegit-config(1) man pageexplains the meaning of each option.)

Git also keeps a pristine copy of Alice’smaster branch under thenameorigin/master:

If Bob later decides to work from a different host, he can stillperform clones and pulls using the ssh protocol:

Alternatively, Git has a native protocol, or can use http;seegit-pull(1) for details.

Git can also be used in a CVS-like mode, with a central repositorythat various users push changes to; seegit-push(1) andgitcvs-migration(7).

Git history is represented as a series of interrelated commits. Wehave already seen that thegitlog command can list those commits.Note that first line of eachgitlog entry also gives a name for thecommit:

We can give this name togitshow to see the details about thiscommit.

But there are other ways to refer to commits. You can use any initialpart of the name that is long enough to uniquely identify the commit:

Every commit usually has one "parent" commitwhich points to the previous state of the project:

Note that merge commits may have more than one parent:

You can also give commits names of your own; after running

you can refer to1b2e1d63ff by the namev2.5. If you intend toshare this name with other people (for example, to identify a releaseversion), you should create a "tag" object, and perhaps sign it; seegit-tag(1) for details.

Any Git command that needs to know a commit can take any of thesenames. For example:

Be careful with that last command: in addition to losing any changesin the working directory, it will also remove all later commits fromthis branch. If this branch is the only branch containing thosecommits, they will be lost. Also, don’t usegitreset on apublicly-visible branch that other developers pull from, as it willforce needless merges on other developers to clean up the history.If you need to undo changes that you have pushed, usegitrevertinstead.

Thegitgrep command can search for strings in any version of yourproject, so

searches for all occurrences of "hello" inv2.5.

If you leave out the commit name,gitgrep will search any of thefiles it manages in your current directory. So

is a quick way to search just the files that are tracked by Git.

Many Git commands also take sets of commits, which can be specifiedin a number of ways. Here are some examples withgitlog:

You can also givegitlog a "range" of commits where the first is notnecessarily an ancestor of the second; for example, if the tips ofthe branchesstable andmaster diverged from a commoncommit some time ago, then

will list commits made in themaster branch but not in thestable branch, while

will show the list of commits made on the stable branch but notthemaster branch.

Thegitlog command has a weakness: it must present commits in alist. When the history has lines of development that diverged andthen merged back together, the order in whichgitlog presentsthose commits is meaningless.

Most projects with multiple contributors (such as the Linux kernel,or Git itself) have frequent merges, andgitk does a better job ofvisualizing their history. For example,

allows you to browse any commits from the last 2 weeks of commitsthat modified files under thedrivers directory. (Note: you canadjust gitk’s fonts by holding down the control key while pressing"-" or "+".)

Finally, most commands that take filenames will optionally allow youto precede any filename by a commit, to specify a particular versionof the file:

You can also usegitshow to see any such file:

This tutorial should be enough to perform basic distributed revisioncontrol for your projects. However, to fully understand the depthand power of Git you need to understand two simple ideas on which itis based:

• The object database is the rather elegant system used tostore the history of your project—​files, directories, andcommits.

• The index file is a cache of the state of a directory tree,used to create commits, check out working directories, andhold the various trees involved in a merge.

Part two of this tutorial explains the objectdatabase, the index file, and a few other odds and ends that you’llneed to make the most of Git. You can find it atgittutorial-2(7).

If you don’t want to continue with that right away, a few otherdigressions that may be interesting at this point are:

• git-format-patch(1),git-am(1): These convertseries of git commits into emailed patches, and vice versa,useful for projects such as the Linux kernel which rely heavilyon emailed patches.

• git-bisect(1): When there is a regression in yourproject, one way to track down the bug is by searching throughthe history to find the exact commit that’s to blame.gitbisectcan help you perform a binary search for that commit. It issmart enough to perform a close-to-optimal search even in thecase of complex non-linear history with lots of merged branches.

• gitworkflows(7): Gives an overview of recommendedworkflows.

• giteveryday(7): Everyday Git with 20 Commands Or So.

• gitcvs-migration(7): Git for CVS users.

gittutorial-2(7),gitcvs-migration(7),gitcore-tutorial(7),gitglossary(7),git-help(1),gitworkflows(7),giteveryday(7),The Git User’s Manual

Part of thegit(1) suite