path: root/BUILDING.txt

*******************************************************************************
**     Building TigerVNC
*******************************************************************************


================================
Build Requirements (All Systems)
================================

-- CMake (http://www.cmake.org) v2.8 or later

-- FLTK 1.3.0
   * Must be patched using fltk/extensions.patch (in the TigerVNC source
     directory) to get full functionality
   * See "Building FLTK" below.

-- If building TLS support:
   * GnuTLS and its dependencies (libgcrypt, libtasn1, libgpg-error)

-- If building native language support (NLS):
   * Gnu gettext 0.14.4 or later
   * See "Building Native Language Support" below.

-- libjpeg-turbo SDK
   * If your system does not include libjpeg-turbo, then you can download the
     SDK from http://sourceforge.net/projects/libjpeg-turbo/files/


=========================
Build Requirements (Unix)
=========================

-- Non-Mac platforms:
   * X11 development kit

-- If building Xvnc:
   * autoconf 2.57 or later
   * automake 1.7 or later
   * libtool 1.4 or later
   * OpenSSL v0.9.7 or later
   * Python v2.3 or later
   * zlib v1.2 or later

   * pkgconfig 0.20 or later
     If you are using an older RPM-based system which does not meet this
     minimum requirement (such as Red Hat Enterprise Linux 4), then you can
     build pkgconfig 0.20 from the Fedora 5 source RPM:
     http://archive.fedoraproject.org/pub/archive/fedora/linux/core/5/source/SRPMS/pkgconfig-0.20-2.2.1.src.rpm

   * gettext 0.14.4 or later
     If you are using an older RPM-based system which does not meet this
     minimum requirement (such as Red Hat Enterprise Linux 4), then you can
     build gettext 0.14.5 from the Fedora 5 source RPM:
     http://archive.fedoraproject.org/pub/archive/fedora/linux/core/5/source/SRPMS/gettext-0.14.5-3.src.rpm
     NOTE: gcj and libgcj are required when building the above source RPM.


============================
Build Requirements (Windows)
============================

-- Microsoft Visual C++ 2005 or later

   If you don't already have Visual C++, then the easiest way to get it is by
   installing the Windows SDK:

   http://msdn.microsoft.com/en-us/windows/bb980924.aspx

   The Windows SDK includes both 32-bit and 64-bit Visual C++ compilers and
   everything necessary to build TigerVNC.

   * For 32-bit builds, you can also use Microsoft Visual C++ Express
     Edition.  Visual C++ Express Edition is a free download.
   * If you intend to build TigerVNC from the command line, then add the
     appropriate compiler and SDK directories to the INCLUDE, LIB, and PATH
     environment variables.  This is generally accomplished by executing
     vcvars32.bat or vcvars64.bat and SetEnv.cmd.  vcvars32.bat and
     vcvars64.bat are part of Visual C++ and are located in the same directory
     as the compiler.  SetEnv.cmd is part of the Windows SDK.  You can pass
     optional arguments to SetEnv.cmd to specify a 32-bit or 64-bit build
     environment.

... OR ...

-- MinGW

-- Inno Setup (needed to build the TigerVNC installer)
   Inno Setup can be downloaded from http://www.jrsoftware.org/isinfo.php.
   You also need the Inno Setup Preprocessor, which is available in the
   Inno Setup QuickStart Pack.

   Add the directory containing iscc.exe (for instance, 
   C:\Program Files\Inno Setup 5) to the system or user PATH environment
   variable prior to building TigerVNC.


=============
Building FLTK
=============

Currently, TigerVNC requires functionality that is not yet included in the
upstream version of FLTK.  Thus, it will be necessary to build FLTK from
source.  To do this:

1. Download the FLTK 1.3.0 source tarball from http://www.fltk.org/
2. Extract the source tarball
3. cd fltk-1.3.0
4. patch -p1 < {tigervnc_source_directory}/fltk/extensions.patch
5. Use CMake to build FLTK using the same procedures described below for
   building TigerVNC.  The recipes in the "Build Recipes" section also apply.
6. (optional) Use 'make install' to install FLTK into a directory of your
   choosing.
7. When building TigerVNC, set the FLTK_FLUID_EXECUTABLE CMake variable to the
   location of the fluid executable that was built in Step 5 or installed in
   Step 6.  This gives CMake a hint as to where to find the FLTK library.
8. If you did not install FLTK, then set the FLTK_INCLUDE_DIR CMake variable to
   the location of the FLTK source directory.


==================
Out-of-Tree Builds
==================

Binary objects, libraries, and executables are generated in the same directory
from which cmake was executed (the "binary directory"), and this directory need
not necessarily be the same as the TigerVNC source directory.  You can create
multiple independent binary directories, in which different versions of
TigerVNC can be built from the same source tree using different compilers or
settings.  In the sections below, {build_directory} refers to the binary
directory, whereas {source_directory} refers to the TigerVNC source directory.
For in-tree builds, these directories are the same.


=================
Building TigerVNC
=================


Unix/Mac
--------

The following procedure will build the TigerVNC Viewer on Linux and Unix
systems.  On 64-bit systems, this will build a 64-bit version of TigerVNC.  See
"Build Recipes" for specific build instructions for building a 32-bit version
of TigerVNC on 64-bit systems.

  cd {build_directory}
  cmake -G "Unix Makefiles" {source_directory} [additional CMake flags]
  make

Building the TigerVNC Server (Xvnc) is a bit trickier.  On newer systems
containing Xorg 7.4 or later (such as Fedora), Xvnc is typically built to use
the X11 shared libraries provided with the system.  The procedure for this is
very system-specific, since it requires specifying such things as font and X
resource directories.  Thus, it cannot be described in terms general enough for
this document.

Those using systems with older versions of Xorg must build a "legacy-friendly"
version of the TigerVNC Server.  This is accomplished by downloading and
building the more recent Xorg modules in a local directory and then building
Xvnc such that it links against the local build of these libraries, not the X11
libraries installed on the system.  The "build-xorg" script in the TigerVNC
source distribution (located under unix/) automates this process.

The following procedure will build both the TigerVNC Viewer and a
"legacy-friendly" version of the TigerVNC Server:

  cd {build_directory}
  sh {source_directory}/unix/build-xorg init
  sh {source_directory}/unix/build-xorg build [additional CMake flags]

build-xorg generates a version of Xvnc that has no external dependencies on the
X11 shared libraries or any other distribution-specific shared libraries.  This
version of Xvnc should be transportable across multiple O/S distributions.
build-xorg should work on Red Hat Enterprise 4, its contemporaries, and later
systems.  It probably will not work on older systems.  It has not been tested
on non-Linux systems (yet).

build-xorg can also be used to rebuild just the TigerVNC Server and Viewer,
once the X11 modules and other dependencies have been built for the first time.
This is convenient for testing changes that just apply to the TigerVNC source
code.  To accomplish this, run:

  sh {source_directory}/unix/build-xorg rebuild [additional make flags]

For instance,

  sh {source_directory}/unix/build-xorg rebuild clean

will clean both the Xvnc and vncviewer builds without destroying any of the
build configuration or module dependencies.


Visual C++ (Command Line)
-------------------------

  cd {build_directory}
  cmake -G "NMake Makefiles" -DCMAKE_BUILD_TYPE=Release {source_directory}
  nmake

This will build either a 32-bit or a 64-bit version of TigerVNC, depending
on which version of cl.exe is in the PATH.


Visual C++ (IDE)
----------------

Choose the appropriate CMake generator option for your version of Visual Studio
(run "cmake" with no arguments for a list of available generators.)  For
instance:

  cd {build_directory}
  cmake -G "Visual Studio 9 2008" {source_directory}

You can then open ALL_BUILD.vcproj in Visual Studio and build one of the
configurations in that project ("Debug", "Release", etc.) to generate a full
build of TigerVNC.


MinGW
-----

  cd {build_directory}
  cmake -G "MSYS Makefiles" {source_directory}
  make

This will generate only vncviewer.  Currently, Visual C++ must be used to build
WinVNC.


Debug Build
-----------

Add "-DCMAKE_BUILD_TYPE=Debug" to the CMake command line.  Or, if building with
NMake, remove "-DCMAKE_BUILD_TYPE=Release" (Debug builds are the default with
NMake.)


Self-Contained MinGW Build
--------------------------

If TigerVNC is built using MinGW, then it may depend on the MinGW libgcc DLL.
To eliminate this dependency, add

  -DCMAKE_C_FLAGS=-static-libgcc -DCMAKE_CXX_FLAGS=-static-libgcc

to the CMake command line.


======================================
Building Native Language Support (NLS)
======================================

NLS requires gettext, which is supplied with most Linux distributions but not
with most other operating systems.  Building gettext from source is
straightforward on Mac and other Unix variants, but on Windows it requires
MinGW.  Thus, it is not possible to enable NLS when building TigerVNC with
Visual C++.

You can override the ICONV_LIBRARIES and LIBINTL_LIBRARY CMake variables to
specify the locations of libiconv and libintl, respectively.  For instance,
adding

  -DLIBINTL_LIBRARY=/opt/gettext/lib/libintl.a

to the CMake command line would link TigerVNC against a static version of
libintl located under /opt/gettext.  Adding

  -DICONV_INCLUDE_DIR=/mingw/include \
    -DICONV_LIBRARIES=/mingw/lib/libiconv.a \
    -DGETTEXT_INCLUDE_DIR=/mingw/include \
    -DLIBINTL_LIBRARY=/mingw/lib/libintl.a

to the CMake command line would link TigerVNC against the static versions of
libiconv and libintl included in the MinGW Developer Toolkit.


===============================================
Building Transport Layer Security (TLS) support
===============================================

VeNCrypt (the TigerVNC security and authentication extensions) can be built
with TLS support, which provides built-in encryption for VNC sessions.  This
requires GnuTLS, which is readily available in many Linux distributions but not
as readily available in binary form on other types of systems.  This section
describes the issues associated with building a version of TigerVNC with TLS
support and how to work around those issues.


Unix/Mac
--------

In general, if you are building on a Unix-ish platform that has the GnuTLS
libraries and include files installed in the standard system locations, then
the TigerVNC build system should detect the system version of GnuTLS
automatically and link against it.  However, this produces a version of
TigerVNC that depends on the GnuTLS dynamic libraries, and thus the TigerVNC
binaries are not portable.

To build a fully portable, cross-compatible version of TigerVNC with TLS
support, it is necessary to link against the GnuTLS static library (as well
as the static libraries of its dependencies.)  If you are lucky enough, then
your O/S distribution may include pre-packaged versions of these static
libraries.  Otherwise, it will be necessary to build GnuTLS, libgcrypt,
libtasn1, and libgpg-error from source.

You can manipulate the GNUTLS_INCLUDE_DIR and GNUTLS_LIBRARY CMake variables to
build TigerVNC against a custom build of GnuTLS that is installed in a
non-system directory.  For instance, adding

  -DGNUTLS_INCLUDE_DIR=/opt/gnutls/include \
    -DGNUTLS_LIBRARY='/opt/gnutls/lib/libgnutls.a;/opt/gnutls/lib/libgcrypt.a;/opt/gnutls/lib/libgpg-error.a;/opt/gnutls/lib/libtasn1.a' \
    -DUSE_INCLUDED_ZLIB=1

to the CMake or 'build-xorg build' command line will cause TigerVNC to be
statically linked against a custom installation of GnuTLS that resides under
/opt/gnutls.  GnuTLS depends on zlib, so specifying -DUSE_INCLUDED_ZLIB=1 will
satisfy that dependency using TigerVNC's in-tree version of zlib, which
prevents TigerVNC from depending on the libz dynamic library.


MinGW
-----

An installer containing the GnuTLS header files, as well as static and dynamic
link libraries for 32-bit MinGW, can be downloaded from the following site:

http://josefsson.org/gnutls4win/

As of this writing, GnuTLS cannot be built cleanly with MinGW64 due to the fact
that portions of the code assume an LP64 data model (Windows uses LLP64.)
Thus, it is not possible at this time to produce a Win64 version of TigerVNC
with TLS support.

Whether you use the above installer or build GnuTLS from source, make sure that
you install the libraries and headers into a pathname that doesn't contain
spaces (the installer will try to install under c:\Program Files unless you
tell it otherwise.)  If the GnuTLS include path contains spaces, then the MinGW
resource compiler will barf when you try to build TigerVNC.

You can manipulate the GNUTLS_INCLUDE_DIR and GNUTLS_LIBRARY CMake variables to
specify the directory under which you installed GnuTLS.  For instance, adding

  -DGNUTLS_INCLUDE_DIR=/c/gnutls/include \
    -DGNUTLS_LIBRARY=/c/gnutls/lib/libgnutls.dll.a

to the CMake command line when using MinGW will cause TigerVNC to be linked
against GnuTLS DLLs that are installed under c:\gnutls.

Adding

  -DGNUTLS_INCLUDE_DIR=/c/gnutls/include \
    -DGNUTLS_LIBRARY='/c/gnutls/lib/libgnutls.a;/c/gnutls/lib/libgcrypt.a;/c/gnutls/lib/libtasn1.a;/c/gnutls/lib/libgpg-error.a'

to the CMake command line will cause TigerVNC to be statically linked against
GnuTLS libraries that are installed under c:\gnutls.

Note that the use of MinGW means that only the TigerVNC viewer can be built,
not the server.


Visual C++
----------

There is generally no sane way to build GnuTLS and its dependencies using
Visual C++.  Thus, it is necessary to either build the libraries with MinGW (or
download 32-bit versions of these from the link above), generate Visual C++
import libraries from the DLLs, then link TigerVNC against the Visual C++
import libraries.

In the instructions below, {gnutls_path} indicates the path under which GnuTLS
is installed (Example: c:\Program Files\GnuTLS-2.10.1).

To generate Visual C++ import libraries:

cd {gnutls_path}\lib
lib /def:..\bin\libgnutls-{version}.def /out:libgnutls.lib

Now, you can add the following arguments to the CMake command line:

  -DGNUTLS_INCLUDE_DIR={gnutls_path}\include \
    -DGNUTLS_LIBRARY={gnutls_path}\lib\libgnutls.lib

to build TigerVNC against the GnuTLS DLLs installed under {gnutls_path}.


===================
Installing TigerVNC
===================

You can use the build system to install TigerVNC into a directory of your
choosing.  To do this, add:

  -DCMAKE_INSTALL_PREFIX={install_directory}

to the CMake command line.  Then, after building TigerVNC, you can run
'make install' or 'nmake install' (or build the "install" target in the Visual
Studio IDE) to install it.

If you don't specify CMAKE_INSTALL_PREFIX, then the default is
c:\Program Files\TigerVNC on Windows and /usr/local on Unix.


=========================
Creating Release Packages
=========================

The following commands can be used to create various types of release packages:


Unix
----

(TODO: implement these in CMake.  OS X viewer also needs to be packaged into
an app instead of a package, since it can run stand-alone now.)

make tarball

  Create a binary tarball containing the TigerVNC Viewer

make servertarball

  Create a binary tarball containing both the TigerVNC Server and Viewer

make dmg

  Create Macintosh package/disk image.  This requires the PackageMaker
  application, which must be installed in /Developer/Applications/Utilities.

make udmg [BUILDDIR32={32-bit build directory}]

  On 64-bit OS X systems, this creates a version of the Macintosh package and
  disk image which contains universal i386/x86-64 binaries.  You should first
  configure a 32-bit out-of-tree build of TigerVNC, then configure a 64-bit
  out-of-tree build, then run 'make udmg' from the 64-bit build directory.  The
  build system will look for the 32-bit build under {source_directory}/osxx86
  by default, but you can override this by setting the BUILDDIR32 variable on
  the make command line as shown above.  Either the 64-bit or 32-bit build can
  be configured to be backward-compatible by using the instructions in the
  "Unix Build Recipes" section.


Windows
-------

If using NMake:

  cd {build_directory}
  nmake installer

If using MinGW:

  cd {build_directory}
  make installer

If using the Visual Studio IDE, build the "installer" project.

The installer package (TigerVNC[64].exe) will be located under
{build_directory}.  If building using the Visual Studio IDE, then the installer
package will be located in a subdirectory with the same name as the
configuration you built (such as {build_directory}\Debug\ or
{build_directory}\Release\).

NOTE:  If TigerVNC is built with TLS support, then the build system will
attempt to package the GnuTLS DLLs into the Windows installer.  It looks for
these DLLs in a directory called "bin" one level up from GNUTLS_INCLUDE_DIR.


=============
Build Recipes
=============


32-bit Build on 64-bit Linux/Unix (including OS X)
--------------------------------------------------

Set the following environment variables before running CMake or
'build-xorg build':

  CFLAGS='-O3 -m32'
  CXXFLAGS='-O3 -m32'
  LDFLAGS=-m32


64-bit Backward-Compatible Build on 64-bit OS X
-----------------------------------------------

Add

  -DCMAKE_OSX_SYSROOT=/Developer/SDKs/MacOSX10.5.sdk \
    -DCMAKE_OSX_DEPLOYMENT_TARGET=10.5 

to the CMake command line.  The OS X 10.5 SDK must be installed.


32-bit Backward-Compatible Build on 64-bit OS X
-----------------------------------------------

Set the following environment variables:

  CC=gcc-4.0
  CXX=g++-4.0
  CFLAGS='-O3 -m32'
  CXXFLAGS='-O3 -m32'
  LDFLAGS=-m32

and add

  -DCMAKE_OSX_SYSROOT=/Developer/SDKs/MacOSX10.4u.sdk \
    -DCMAKE_OSX_DEPLOYMENT_TARGET=10.4

to the CMake command line.  The OS X 10.4 SDK must be installed.


64-bit MinGW Build on Cygwin
----------------------------

  cd {build_directory}
  CC=/usr/bin/x86_64-w64-mingw32-gcc CXX=/usr/bin/x86_64-w64-mingw32-g++ \
    RC=/usr/bin/x86_64-w64-mingw32-windres \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DCMAKE_AR=/usr/bin/x86_64-w64-mingw32-ar \
    -DCMAKE_RANLIB=/usr/bin/x86_64-w64-mingw32-ranlib {source_directory}
  make

This produces a 64-bit build of TigerVNC that does not depend on cygwin1.dll or
other Cygwin DLL's.  The mingw64-x86_64-gcc-core and mingw64-x86_64-gcc-g++
packages (and their dependencies) must be installed.


32-bit MinGW Build on Cygwin
----------------------------

  cd {build_directory}
  CC=/usr/bin/i686-w64-mingw32-gcc CXX=/usr/bin/i686-w64-mingw32-g++ \
    RC=/usr/bin/i686-w64-mingw32-windres \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DDCMAKE_AR=/usr/bin/i686-w64-mingw32-ar \
    -DCMAKE_RANLIB=/usr/bin/i686-w64-mingw32-ranlib {source_directory}
  make

This produces a 32-bit build of TigerVNC that does not depend on cygwin1.dll or
other Cygwin DLL's.  The mingw64-i686-gcc-core and mingw64-i686-gcc-g++
packages (and their dependencies) must be installed.


MinGW-w64 Build on Windows
--------------------------

This produces a 64-bit build of TigerVNC using the "native" MinGW-w64 toolchain
(which is faster than the Cygwin version):

  cd {build_directory}
  CC={mingw-w64_binary_path}/x86_64-w64-mingw32-gcc \
    CXX={mingw-w64_binary_path}/x86_64-w64-mingw32-g++ \
    RC={mingw-w64_binary_path}/x86_64-w64-mingw32-windres \
    cmake -G "MSYS Makefiles" \
    -DCMAKE_AR={mingw-w64_binary_path}/x86_64-w64-mingw32-ar \
    -DCMAKE_RANLIB={mingw-w64_binary_path}/x86_64-w64-mingw32-ranlib \
    {source_directory}
  make


MinGW Build on Linux
--------------------

  cd {build_directory}
  CC={mingw_binary_path}/i386-mingw32-gcc \
    CXX={mingw_binary_path}/i386-mingw32-g++ \
    RC={mingw_binary_path}/i386-mingw32-windres \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DCMAKE_AR={mingw_binary_path}/i386-mingw32-ar \
    -DCMAKE_RANLIB={mingw_binary_path}/i386-mingw32-ranlib \
    {source_directory}
  make