61d4922928
Delta search was discarding discovered deltas if an object appeared near a type boundary in the delta search window. This has caused JGit to produce larger pack files than other implementations of the packing algorithm. Delta search works by pushing prior objects into a search window, an ordered list of objects to attempt to delta compress the next object against. (The window size is bounded, avoiding O(N^2) behavior.) For implementation reasons multiple object types can appear in the input list, and the window. PackWriter commonly passes both trees and blobs in the input list handed to the DeltaWindow algorithm. The pack file format requires an object to only delta compress against the same type, so the DeltaWindow algorithm must stop doing comparisions if a blob would be compared to a tree. Because the input list is sorted by object type and the window is recently considered prior objects, once a wrong type is discovered in the window the search algorithm stops and uses the current result. Unfortunately the termination condition was discarding any found delta by setting deltaBase and deltaBuf to null when it was trying to break the window search. When this bug occurs, the state of the DeltaWindow looks like this: current | \ / input list: tree0 tree1 blob1 blob2 window: blob1 tree1 tree0 / \ | res.prev As the loop iterates to the right across the window, it first finds that blob1 is a suitable delta base for blob2, and temporarily holds this in the bestDelta/deltaBuf fields. It then considers tree1, but tree1 has the wrong type (blob != tree), so the window loop must give up and fall through the remaining code. Moving the condition up and discarding the window contents allows the bestDelta/deltaBuf to be kept, letting the final file delta compress blob1 against blob0. The impact of this bug (and its fix) on real world repositories is likely minimal. The boundary from blob to tree happens approximately once in the search, as the input list is sorted by type. Only the first window size worth of blobs (e.g. 10 or 250) were failing to produce a delta in the final file. This bug fix does produce significantly different results for small test repositories created in the unit test suite, such as when a pack may contains 6 objects (2 commits, 2 trees, 2 blobs). Packing test cases can now better sample different output pack file sizes depending on delta compression and object reuse flags in PackConfig. Change-Id: Ibec09398d0305d4dbc0c66fce1daaf38eb71148f |
||
---|---|---|
.mvn | ||
lib | ||
org.eclipse.jgit | ||
org.eclipse.jgit.ant | ||
org.eclipse.jgit.ant.test | ||
org.eclipse.jgit.archive | ||
org.eclipse.jgit.http.apache | ||
org.eclipse.jgit.http.server | ||
org.eclipse.jgit.http.test | ||
org.eclipse.jgit.junit | ||
org.eclipse.jgit.junit.http | ||
org.eclipse.jgit.lfs | ||
org.eclipse.jgit.lfs.server | ||
org.eclipse.jgit.lfs.server.test | ||
org.eclipse.jgit.lfs.test | ||
org.eclipse.jgit.packaging | ||
org.eclipse.jgit.pgm | ||
org.eclipse.jgit.pgm.test | ||
org.eclipse.jgit.test | ||
org.eclipse.jgit.ui | ||
tools | ||
.buckconfig | ||
.buckversion | ||
.gitattributes | ||
.gitignore | ||
.mailmap | ||
BUCK | ||
BUILD | ||
CONTRIBUTING.md | ||
LICENSE | ||
README.md | ||
WORKSPACE | ||
pom.xml |
README.md
Java Git
An implementation of the Git version control system in pure Java.
This package is licensed under the EDL (Eclipse Distribution License).
JGit can be imported straight into Eclipse, built and tested from there, but the automated builds use Maven.
-
org.eclipse.jgit
A pure Java library capable of being run standalone, with no additional support libraries. It provides classes to read and write a Git repository and operate on a working directory.
All portions of JGit are covered by the EDL. Absolutely no GPL, LGPL or EPL contributions are accepted within this package.
-
org.eclipse.jgit.ant
Ant tasks based on JGit.
-
org.eclipse.jgit.archive
Support for exporting to various archive formats (zip etc).
-
org.eclipse.jgit.http.apache
Apache httpclient support
-
org.eclipse.jgit.http.server
Server for the smart and dumb Git HTTP protocol.
-
org.eclipse.jgit.pgm
Command-line interface Git commands implemented using JGit ("pgm" stands for program).
-
org.eclipse.jgit.packaging
Production of Eclipse features and p2 repository for JGit. See the JGit Wiki on why and how to use this module.
Tests
-
org.eclipse.jgit.junit
Helpers for unit testing
-
org.eclipse.jgit.test
Unit tests for org.eclipse.jgit
-
org.eclipse.jgit.ant.test
-
org.eclipse.jgit.pgm.test
-
org.eclipse.jgit.http.test
-
org.eclipse.jgit.junit.test
No further description needed
Warnings/Caveats
-
Native smbolic links are supported, provided the file system supports them. For Windows you must have Windows Vista/Windows 2008 or newer, use a non-administrator account and have the SeCreateSymbolicLinkPrivilege.
-
Only the timestamp of the index is used by jgit if the index is dirty.
-
JGit requires at least a Java 7 JDK.
-
CRLF conversion is performed depending on the core.autocrlf setting, however Git for Windows by default stores that setting during installation in the "system wide" configuration file. If Git is not installed, use the global or repository configuration for the core.autocrlf setting.
-
The system wide configuration file is located relative to where C Git is installed. Make sure Git can be found via the PATH environment variable. When installing Git for Windows check the "Run Git from the Windows Command Prompt" option. There are other options like Eclipse settings that can be used for pointing out where C Git is installed. Modifying PATH is the recommended option if C Git is installed.
-
We try to use the same notation of $HOME as C Git does. On Windows this is often not the same value as the user.home system property.
Package Features
-
org.eclipse.jgit/
-
Read loose and packed commits, trees, blobs, including deltafied objects.
-
Read objects from shared repositories
-
Write loose commits, trees, blobs.
-
Write blobs from local files or Java InputStreams.
-
Read blobs as Java InputStreams.
-
Copy trees to local directory, or local directory to a tree.
-
Lazily loads objects as necessary.
-
Read and write .git/config files.
-
Create a new repository.
-
Read and write refs, including walking through symrefs.
-
Read, update and write the Git index.
-
Checkout in dirty working directory if trivial.
-
Walk the history from a given set of commits looking for commits introducing changes in files under a specified path.
-
Object transport Fetch via ssh, git, http, Amazon S3 and bundles. Push via ssh, git and Amazon S3. JGit does not yet deltify the pushed packs so they may be a lot larger than C Git packs.
-
Garbage collection
-
Merge
-
Rebase
-
And much more
-
-
org.eclipse.jgit.pgm/
- Assorted set of command line utilities. Mostly for ad-hoc testing of jgit log, glog, fetch etc.
-
org.eclipse.jgit.ant/
- Ant tasks
-
org.eclipse.jgit.archive/
- Support for Zip/Tar and other formats
-
org.eclipse.http.*/
- HTTP client and server support
Missing Features
There are some missing features:
- gitattributes support
Support
Post question, comments or patches to the jgit-dev@eclipse.org mailing list. You need to be subscribed to post, see here:
https://dev.eclipse.org/mailman/listinfo/jgit-dev
Contributing
See the EGit Contributor Guide:
http://wiki.eclipse.org/EGit/Contributor_Guide
About Git
More information about Git, its repository format, and the canonical C based implementation can be obtained from the Git website: