Grit gives you object oriented read/write access to Git repositories via Ruby.
Перейти к файлу
Jeff King 13d5737ee6 fsync loose objects before moving into place
When we write a loose object to disk, we simply close the
file object before moving it into place. If the machine
crashes shortly after our write, the contents may not have
been committed to disk (depending your filesystem, usually
the metadata is, and you end up with a corrupt, zero-length
loose object file).

This is especially bad because we report that the object is
successfully written, which means we may have updated refs
to point to it. A corrupt object at that point means not
only does the operation fail, but the repository is left in
a corrupted and unusable state.

We can fix this by calling fsync on the object file before
linking it into place. Between this and the previous commit,
our object writing should now behave exactly like git's
internal routines.
2013-02-23 14:07:24 -08:00
examples Update from GitHub. 2010-08-03 17:05:19 -06:00
lib fsync loose objects before moving into place 2013-02-23 14:07:24 -08:00
test Missed some stuff 2013-02-23 14:06:38 -08:00
.gitignore Added Repo#commit_deltas_from as a (fairly expensive and lazy) way of getting 2009-02-14 11:49:16 +01:00
API.txt added some simple write ops : add, remove, commit 2008-05-29 11:33:09 -07:00
History.txt Release 2.5.0 2012-04-22 00:02:35 -07:00
LICENSE convert readme to markdown 2009-02-28 15:14:10 -08:00
PURE_TODO added some simple write ops : add, remove, commit 2008-05-29 11:33:09 -07:00
README.md Update Readme. 2010-08-31 15:23:03 -07:00
Rakefile Ruby 1.9 compatibility 2010-08-30 17:37:38 -07:00
benchmarks.rb added some simple write ops : add, remove, commit 2008-05-29 11:33:09 -07:00
benchmarks.txt added some simple write ops : add, remove, commit 2008-05-29 11:33:09 -07:00
grit.gemspec Where github/github actual is 2013-02-23 14:04:37 -08:00

README.md

Grit

Grit gives you object oriented read/write access to Git repositories via Ruby. The main goals are stability and performance. To this end, some of the interactions with Git repositories are done by shelling out to the system's git command, and other interactions are done with pure Ruby reimplementations of core Git functionality. This choice, however, is transparent to end users, and you need not know which method is being used.

This software was developed to power GitHub, and should be considered production ready. An extensive test suite is provided to verify its correctness.

Grit is maintained by Tom Preston-Werner, Scott Chacon, Chris Wanstrath, and PJ Hyett.

This documentation is accurate as of Grit 2.3.

Requirements

Install

Easiest install is via RubyGems:

$ gem install grit

Source

Grit's Git repo is available on GitHub, which can be browsed at:

http://github.com/mojombo/grit

and cloned with:

git clone git://github.com/mojombo/grit.git

Development

You will need these gems to get tests to pass:

  • mocha

Contributing

If you'd like to hack on Grit, follow these instructions. To get all of the dependencies, install the gem first.

  1. Fork the project to your own account
  2. Clone down your fork
  3. Create a thoughtfully named topic branch to contain your change
  4. Hack away
  5. Add tests and make sure everything still passes by running rake
  6. If you are adding new functionality, document it in README.md
  7. Do not change the version number, I will do that on my end
  8. If necessary, rebase your commits into logical chunks, without errors
  9. Push the branch up to GitHub
  10. Send a pull request for your branch

Usage

Grit gives you object model access to your Git repositories. Once you have created a Repo object, you can traverse it to find parent commits, trees, blobs, etc.

Initialize a Repo object

The first step is to create a Grit::Repo object to represent your repo. In this documentation I include the Grit module to reduce typing.

require 'grit'
repo = Grit::Repo.new("/Users/tom/dev/grit")

In the above example, the directory /Users/tom/dev/grit is my working directory and contains the .git directory. You can also initialize Grit with a bare repo.

repo = Repo.new("/var/git/grit.git")

Getting a list of commits

From the Repo object, you can get a list of commits as an array of Commit objects.

repo.commits
# => [#<Grit::Commit "e80bbd2ce67651aa18e57fb0b43618ad4baf7750">,
      #<Grit::Commit "91169e1f5fa4de2eaea3f176461f5dc784796769">,
      #<Grit::Commit "038af8c329ef7c1bae4568b98bd5c58510465493">,
      #<Grit::Commit "40d3057d09a7a4d61059bca9dca5ae698de58cbe">,
      #<Grit::Commit "4ea50f4754937bf19461af58ce3b3d24c77311d9">]

Called without arguments, Repo#commits returns a list of up to ten commits reachable by the master branch (starting at the latest commit). You can ask for commits beginning at a different branch, commit, tag, etc.

repo.commits('mybranch')
repo.commits('40d3057d09a7a4d61059bca9dca5ae698de58cbe')
repo.commits('v0.1')

You can specify the maximum number of commits to return.

repo.commits('master', 100)

If you need paging, you can specify a number of commits to skip.

repo.commits('master', 10, 20)

The above will return commits 21-30 from the commit list.

The Commit object

Commit objects contain information about that commit.

head = repo.commits.first

head.id
# => "e80bbd2ce67651aa18e57fb0b43618ad4baf7750"

head.parents
# => [#<Grit::Commit "91169e1f5fa4de2eaea3f176461f5dc784796769">]

head.tree
# => #<Grit::Tree "3536eb9abac69c3e4db583ad38f3d30f8db4771f">

head.author
# => #<Grit::Actor "Tom Preston-Werner <tom@mojombo.com>">

head.authored_date
# => Wed Oct 24 22:02:31 -0700 2007

head.committer
# => #<Grit::Actor "Tom Preston-Werner <tom@mojombo.com>">

head.committed_date
# => Wed Oct 24 22:02:31 -0700 2007

head.message
# => "add Actor inspect"

You can traverse a commit's ancestry by chaining calls to #parents.

repo.commits.first.parents[0].parents[0].parents[0]

The above corresponds to master^^^ or master~3 in Git parlance.

The Tree object

A tree records pointers to the contents of a directory. Let's say you want the root tree of the latest commit on the master branch.

tree = repo.commits.first.tree
# => #<Grit::Tree "3536eb9abac69c3e4db583ad38f3d30f8db4771f">

tree.id
# => "3536eb9abac69c3e4db583ad38f3d30f8db4771f"

Once you have a tree, you can get the contents.

contents = tree.contents
# => [#<Grit::Blob "4ebc8aea50e0a67e000ba29a30809d0a7b9b2666">,
      #<Grit::Blob "81d2c27608b352814cbe979a6acd678d30219678">,
      #<Grit::Tree "c3d07b0083f01a6e1ac969a0f32b8d06f20c62e5">,
      #<Grit::Tree "4d00fe177a8407dbbc64a24dbfc564762c0922d8">]

This tree contains two Blob objects and two Tree objects. The trees are subdirectories and the blobs are files. Trees below the root have additional attributes.

contents.last.name
# => "lib"

contents.last.mode
# => "040000"

There is a convenience method that allows you to get a named sub-object from a tree.

tree / "lib"
# => #<Grit::Tree "e74893a3d8a25cbb1367cf241cc741bfd503c4b2">

You can also get a tree directly from the repo if you know its name.

repo.tree
# => #<Grit::Tree "master">

repo.tree("91169e1f5fa4de2eaea3f176461f5dc784796769")
# => #<Grit::Tree "91169e1f5fa4de2eaea3f176461f5dc784796769">

The Blob object

A blob represents a file. Trees often contain blobs.

blob = tree.contents.first
# => #<Grit::Blob "4ebc8aea50e0a67e000ba29a30809d0a7b9b2666">

A blob has certain attributes.

blob.id
# => "4ebc8aea50e0a67e000ba29a30809d0a7b9b2666"

blob.name
# => "README.txt"

blob.mode
# => "100644"

blob.size
# => 7726

You can get the data of a blob as a string.

blob.data
# => "Grit is a library to ..."

You can also get a blob directly from the repo if you know its name.

repo.blob("4ebc8aea50e0a67e000ba29a30809d0a7b9b2666")
# => #<Grit::Blob "4ebc8aea50e0a67e000ba29a30809d0a7b9b2666">

Other

There are many more API methods available that are not documented here. Please reference the code for more functionality.

Copyright (c) 2010 Tom Preston-Werner. See LICENSE for details.