There might be stuff lingering in the buffer simply because flush()
hasn't been called in a while, rather than because the transport is
congested.

We now get notifications when the output buffer empties, and we
already caught incoming RTT pongs, meaning we can now react at the
proper time to retry a congested update rather than use a timer.

We need to proper respect the timeouts set up by the core RFB code
and not just the polling scheduler.

This makes sure we keep the socket busy even if one rect takes some
time to encode.

What to do when a socket is writeable should be handled in the
RFB core code as there may be other events we want to fire off
when this happens.

The logic was a bit confusing and superfluous.

We haven't supported such an old version of Windows for some time.

GetMaster() wasn't exported until Xorg 1.18 so it is unsafe to use
from libvnc.so for older versions.

The package is based on the contrib ubuntu-trusty debian package. The
following modifications have been made:

The current tigervnc sources can be used unmodified as the .orig
tarball for the debian package, so get-orig-source.sh is gone.

The ubuntu xorg-server-sources package provides the xorg sources used
to build tigervnc. The two small patches which where in the trusty
quilt series are now applied to the xorg sources after they have been
extracted. There is no quilt series anymore at this point.

Fltk 1.3.3 is available in the xenial repositories and it is used to build
xtigervncviewer. No need to get the fltk sources. patch_fltk.sh is gone.

Fixed some dependencies to reflect changes between trusty and xenial.

Removed configure options not recognized by configure.

The install target is not called a second time for the viewer package,
but the viewer files are moved from the server installation. (Eliminate
file conflicts during binary package installation.)

Note: Some dependencies (including libfltk1.3 and xorg-server-source) are
found in the xenial universe repository only.

xserver118.patch is identical to xserver117.patch by content.

cmake_policy() by itself raises the minimum requirement on CMake.
We want it to be optional though, so add the necessary if() around
it.

Patch taken directly from f22 nettle package

Fixes numerous problems with the SSH tunneling implementation
on the java viewer and adds many significant new SSH-related
features.  SSH tunneling is now highly configurable via the
both the command line and GUI. The embedded client can parse
openssh config files.  An external client may be specified,
along with a custom arguments template.

We haven't supported them in a long time so we shouldn't mention
them in the interface and documentation.

Previously the incoming clipboard was unconditionally set to both
the PRIMARY and CLIPBOARD selection. This isn't always what the
user want, so make it configurable.

This is the simplest approach, with setPrimaryCB having the same
relationship to acceptClipboard as sendPrimaryCB has to
sendClipboard. This does not change vncviewer, which should probably
have something similar on UNIX platforms only.

Pierre Ossman: Changed naming from accept to set

The builtin causes errors in the test, and we're also trying to
determine if it is provided by libc or not.

We don't modify the link interface directly so these warnings are
just useless noise. We could in theory use the new naming instead,
but at least CMake 2.8.12 is buggy and screws up the link when that
policy is active.

It just makes CMake error out rather than continue without NLS
support.

Allows to specify which configuration parameters can be modified on runtime.

NASM 2.01 can be rebuilt from el6 spec file

change the Copyright year

Various fixes for Java viewer UI

Apple changed the implementation of KeyEvent.getKeyText() to
return the unicode for the key glyph instead of a textual
description.

Significant improvements to the Java viewer UI to
make it more robust and tolerant of new tabs, fonts,
etc.  TextFields and editable ComboBoxes that are
specific to integer values now only accept integer
input.

Additional change need for dynamic build of libX11

Building libX11 statically causes problems with libpthread, however
the version of libX11 that ships with el5 is too old to build libXext.
So we have to build a dynamic version of libX11 and link against that
while building.  The resulting binary will actually be linked against
the system libX11, which should be ABI compatible as far as Xvnc is
concerned.

6262bac introduced a regression that resulted in a 1 pixel border
around the window on some platforms.  The underlying issue was
actually caused by overriding the getPreferredSize() method in
DesktopWindow and returning the scaled width & height, which
might not have been set to valid values.

Add workaround for Vino's VeNCrypt implementation to Java viewer