Until now, Jenkins.overallNodeProvisioner was tracking an inconsistent
pair of numbers. On one hand, it was tracking all the executors in the
system, but then it was also only tracking the # of free-roaming
label-unassigned jobs as the queue length.

So if all your slaves are set with Node.Mode.EXCLUSIVE and you have some
free-roaming jobs, then it'll never tickle NodePrivisioner.

In other words, the # of executors weren't reflecting the actual # of
executors that can execute what's counted as the queue length.

See the problem report at
https://groups.google.com/forum/?fromgroups#!topic/jenkinsci-dev/bUwGEgOwv4Q

To fix this, I introduced another LoadStatistics that only counts
the # of executors that can execute free-roaming jobs, and # of
free-roaming jobs as the queue length. In this way, two pairs of numbers
are consistent.

This now allows us to use overallLoadStatistics for really the entire
system, including all executors and the total queue length. This is
primarily for administrators to see the resource utilization, and it is
not useful for NodeProvisioner input because it's mixing too many
different things.

The semantics change in OverallLoadStatistics.queueLength allows us to
deprecate its totalQueueLength field.

Fall back in non-JNA case wasn't working.

Backward compatibility is preventing the View.READ permission to apply
correctly. It actually overrides the View.READ instead of complementing
it.

This change only applies default READ right if the View.READ is not
available, and the user has View.CONFIGURE + the view is not empty.

Include the file system path that was checked in the low disk space
error message

The actual fix was in c86aa679

The previous implementation was always appending the per-configuration unique suffix, making it impossible for different configuration builds to share workspaces. In this fix, we introduce a secondary field to control the workspace of sub-builds (which can be either absolute or relative to the matrix head workspace.)

exposed data about installed plugins, installable plugins, update center jobs, etc., to the REST API

Previously, the untar code tries to set the last modified time and mode
on every untarred file.  However, if the tar contains a broken symlink,
or a symlink that points to a file that has not been untarred yet, the
time/mode setting would fail on the broken symlink.

Symlinks don't have meaningful modified times or modes of their own, so
only set these values on non-symlinks.

Rename the file "a" in the test to expose the bug.

If a resource with 'Set-Cookie' header is cached (either by intermediary
like HTTP proxy and reverse proxy, or by the browser), it'll cause
identity swap / session mix-up as discussed in this ticket.

I suspect this was caused by HttpSessionContextIntegrationFilter2, which
is the only code path that attempts to create a session when a request
to a static resource is made.

So I'm disabling the creation of session in
HttpSessionContextIntegrationFilter2. This in turn requires that we
have sessions already created when the authentication was successful and
people need to login (or else the login will have no effect.)

We already do so in layout.jelly, so any request that renders a Jenkins
page would have a session, but I've also added it in
AuthenticationProcessingFilter2, which ensures that a successful login
does have a session.

When a client uses the buildWithParameters API and request a json output, return the job definition as a json string instead of doing a redirect on the job page.

Conflicts:

	changelog.html

(cherry picked from commit 2e5a7bf4)