+ GitLFS client support
+ FilestoreModel now parses meta file
+ Read meta heading from cache if available
+ Authentication based on accept headers for browser view filestore login
+ PathModel & PathChangeModel now understands filestore items
+ Zip & Rar downloads contain include filestore items
+ Filestore servlet returns LFS JSON error only if accepted by client
+ DiffStat now knows repository to allow identification of filestore items
+ Filestore items identified and returned via view, raw & blob links on
blame, commitDiff, commit and Tree pages

Empty folders are automatically skipped when browsing repository tree  (similar to github "folder jumping" feature).

testRevlog was expecting 12 commits but getting 19. Looking at the
hello-world repository on github (on which the test executes) the 19
commit count appears to be correct.

These manager interfaces define how the GitBlit singleton will
eventually be split into smaller component managers.  The Wicket app and
all servlets have been updated to request the needed managers.  There
are _very_ few method signature changes - although there are a handful.

This is a surgical sharding of responsibility based on a proof of
concept refactor.  Instead of random references to  GittBlit.self()
there are now precise references to the manager interface required to
accomplish some task.  Some tasks may require references to multiple
managers.

The code is now littered with calls to GitBlit.getManager(class) and
those familiar with the code-base will no doubt notice the duplication
of methods from IUserService in IUserManager and the addition of
implementation methods in the GitBlit context class.  When the
GitBlit class is broken apart and the existing external authentication
user service classes are refactored to AuthenticationService classes,
this will again simplify and flatten.  But in order to safely and
cleanly modularize the stable code-base we will have to live with a
little duplication for a short while.

Change-Id: I7314ec8acaab2dcc6092785ed4434cc09fdbbe16

Change-Id: I172bd4d493eee1013ec542ab97f1be29040f816d

The graph is generated server-side and therefore requires that the
commit table row height be fixed and match the row height of the
servlet.  There will be layout misalignment if remotes refs are
displayed.  Perhaps this can be improved in the future.

Change-Id: I39d0ffc7b1c3679976ce8c198c772ff86238f1a5

Implement adjusting file and directory permissions for shared
repositories in JGitUtils.

Implement the methods getFilemode and setFilemode
in JnaUtils. Not using the libc names as we don't
necessarily use JNA and because it is not necessarily
a one to one mapping.

This is a massive commit which reorganizes the entire project structure
(although it is still monolithic), removes the Build classes, and
switches to Moxie, a smarter Ant build tookit based on the original
Gitblit Build classes.

The Ant build script will likely require additional fine-tuning, but
this is big step forward.

Implemented discrete repository access permissions to replace the
really primitive course-grained permissions used to this point.  This
implementation allows for finer-grained access control, but still
falls short of integrated, branch-based permissions sought by some.

Access permissions follow the conventions established by Gitosis and
Gitolite so they should feel immediately comfortable to experienced
users.  This permissions infrastructure is complete and works exactly as
expected.  Unfortunately, there is no ui in this commit to change
permissions, that will be forthcoming.  In the meantime, Gitblit
hot-reloads users.conf so the permissions can be manipulated at runtime
with a text editor.

The following per-repository permissions are now supported:
- V (view in web ui, RSS feeds, download zip)
- R (clone)
- RW (clone and push)
- RWC (clone and push with ref creation)
- RWD (clone and push with ref creation, deletion)
- RW+ (clone and push with ref creation, deletion, rewind)

And a users.conf entry looks something like this:

	[user "hannibal"]
		password = bossman
		repository = RWD:topsecret.git

Migrated to JUnit4-style tests.
Replaced assertTrue with assertEquals where appropriate.
Start Gitblit instance with dedicated unit testing settings file and
user service.
Integrated RPC, Federation, and Git servlet unit tests into suite.

If an object id was not specified Gitblit used HEAD to perform the
operation.  This breaks under some conditions like working with a
repository that does not have any commits on master but does have
commits on a vcs-import branch.

The new approach is to centralize the resolution of unspecified object
ids to a common method which resolves HEAD first but uses the most
recently modified branch if HEAD points to nothing.

This commit also includes a non-functional method for creating an empty
branch. I couldn't figure out how to make JGit create an orphaned
branch.

* Build script overhaul including building & publishing GO, WAR, Docs,
and Site.
* Restored JGit 0.12.1 dependency and backported Blame.  Got tired of
waiting for JGit 1.0.0 Maven artifacts.
* Changed Summary Page layout
* Optional cookie authentication
* Added icons for log, tags, and branches panels.
* Show last commit author and short message on branches panel.
* Unit testing.
* Documentation.