CISST Libraries

On Linux (unix command line)

  1. Do the following commands to get the source from the cisst github repository

git clone https://github.com/jhu-cisst/cisst-saw.git --recursive

  1. Once you have downloaded the source, use CMake to configure cisst in a separate build directory

mkdir build
cd build
ccmake ../cisst-saw

  1. After creating your makefiles using CMake, build the cisst libraries

make

On MAC

Using XCode on Mac OS X Leopard

Notes:

  • Installation on more recent Mac OS X should be similar.

  • For cisstStereoVision only: If you plan to use OpenCV, you need to install it. We recommend using MacPort

  • For cisstStereoVision only: If you plan to use the svlImageWindow filter or other modules with X11 dependency, you need to install XQuartz (https://xquartz.macosforge.org).

  • You may use XCode for editing, compiling and debugging. However command line applications have to be executed from the Terminal because XCode does not open a command line for them by default. (Although there might be a way to force it somehow.)

  1. Do the following commands to get the source from the cisst github repository

git clone https://github.com/jhu-cisst/cisst-saw.git --recursive

  1. Once you have downloaded the source, use CMake to configure cisst in a separate build directory

mkdir build
cd build
ccmake -G Xcode ../cisst-saw

  1. After creating your makefiles using CMake, build the cisst libraries

make

  1. Open the project in XCode by double clicking on cisst-saw.xcodeproj in Mac OS Finder. When opening the project, XCode might ask for the directory to be used with the project. In that case just leave the default selection and continue.

On Windows

Reminder, you need a C++ compiler (Visual Studio) and CMake (http://www.cmake.org).

  1. Install a git client, see available clients from http://git-scm.com/downloads/guis (see also cisst Download)

  2. Checkout cisst from https://github.com/jhu-cisst/cisst-saw.git, make sure you set your git client’s flags to checkout recursively to get all the git submodules.

  3. Once you have downloaded everything into source, use CMake to configure cisst in the build directory

  4. Run CMake.

  • Where is the source code: cisst source directory

  • Where to build the binaries: custom build directory

  1. Click Configure and turn ON the desired libraries; e.g.:

  • CISST_cisstMultiTask = ON

  • CISST_cisstNumerical = ON

  • CISST_cisstOSAbstraction = ON

  • CISST_cisstParameterTypes = ON

  • CISST_HAS_CISSTNETLIB = ON

  1. Click Configure and set:

  • CISSTNETLIB_DOWNLOAD_NOW = ON

  1. Click Configure and click Generate

  2. Open cisst-saw.sln under cisst build directory.

  3. Build ALL_BUILD in debug and release configurations.

Optional External Dependencies

The following external dependencies are optional. The cisst libraries can be compiled without them, but with less functionality.

cisstNetLib

cisstNetLib is a customized binary distribution of some elements from the Netlib repository (which is Fortran code). The binary files can be downloaded and extracted (https://github.com/jhu-cisst/cisstNetlib), though it is more convenient to use CMake to download it, as follows (assuming you have an Internet connection):

  1. Select the BUILD_LIBS_cisstNumerical option, then press Configure

  2. Two new CMake entries will appear: CISSTNETLIB_DIR and CISSTNETLIB_DOWNLOAD_NOW

  3. Select CISSTNETLIB_DOWNLOAD_NOW and press Configure again

  4. Depending on your OS, you may be prompted for additional information, such as 32 vs. 64 bits

  5. CMake will automatically download the cisstNetLib libraries to your build tree and set the paths

Python

To wrap the cisst libraries for Python, the following steps are needed:

  1. Install Python on your system (Python 3.X is supported).

  2. Install numpy, because the cisst mathematical types (e.g., fixed size and dynamic vectors and matrices) are mapped to numpy arrays.

  3. Install SWIG, which automatically wraps C++ code for Python (as well as other languages)

To use the Interactive Research Environment (IRE):

  1. Select the CISST_cisstInteractive CMake option

  2. Install wxPython or iPython

XML

To use XML within cisst, you must turn ON the CISST_cisstCommonXML option in CMake. There are two options for XML support: LibXML2 or Qt. CMake will attempt to find these packages on its own. If both are present, the default is to use libXML2. This can be changed in CMake by modifying the CISST_XML_LIB variable to be either LibXml2 or QtXML. Note, however, that to use the Qt XML, your code must have a QApplication (i.e., it is not enough to just link with the Qt libraries). Also, if you are getting linker errors in external projects, be sure that cisst::cisstCommonXML is added to your target_link_libraries, i.e.:

target_link_libraries (yourProgram cisst::cisstCommon cisst::cisstCommonXML)

Windows

  • Go to http://www.zlatkovic.com/pub/libxml/ and download the zlib, libxml2, and iconv zipped folders. Extract these.

  • Each folder has a lib, include, and bin subfolder. Put the contents of the lib, include, and bin files of the zlib and iconv into their respective places (subfolders of the same name) in the libxml2 folder. The end result should be that each folder within libxml2 should contain the following:

  • bin

  • libxml2.dll

  • minigzip.exe

  • zlib1.dll

  • iconv.exe

  • iconv.dll

  • xmlcatalog.exe

  • xmllint.exe

  • include

  • libxml folder

  • iconv.h

  • zconf.h

  • zlib.h

  • lib

  • iconv_a.lib

  • zdll.lib

  • zlib.lib

  • iconv.lib

  • libxml2.lib

  • libxml2_a.lib

  • libxml2_a_dll.lib

  • You can discard the original zlib and iconv folders, which should now be empty.

  • Now that this is done, run cmake on your CISST folder and turn on CISST_HAS_XML, it will ask you the location of xmllint.exe, include, and libxml2.lib.

  • Remember to recompile CISST afterwards.

uDrawGraph

The cisst libraries include a component viewer(mtsComponentViewer class in cisstMultiTask) that displays an interactive graphical visualization of the components in the system, whether a single process or distributed between different processes. This component viewer relies on the uDrawGraph package that can be obtained from http://www.informatik.uni-bremen.de/uDrawGraph/en/home.html

After installing it, be sure to add the binary directory to your path.

Linux

  • Download LINUX package from http://www.informatik.uni-bremen.de/uDrawGraph/en/download/download.html

  • Extract uDraw: tar xfz uDrawGraph-3.1.1-0-linux-i386.tar.gz

  • Copy to /usr/local/: sudo mv uDrawGraph-3.1 /usr/local/

  • Add the binary directory to path and $UDG_HOME (add the following block to .bashrc file)

    # uDrawGraph
if [ -f /usr/local/uDrawGraph-3.1/README.txt ]; then
    export UDG_HOME=/usr/local/uDrawGraph-3.1
    export PATH=$PATH:/usr/local/uDrawGraph-3.1/bin
fi

GLUT

The OpenGL Utility Toolkit (GLUT) is used by sawGLUTSimulator.

Windows

  • Download and install GLUT (http://www.xmission.com/~nate/glut.html). For detailed installation process, please refer to README-win32.txt file (which is a bit dated) and/or the following:

  • Copy glut32.dll to C:\Windows\System32 (on a 64-bit machine, this should be C:\Windows\SysWOW64).

  • Copy glut.h to your C++ include directory; by convention this should be in a GL subdirectory

  • For Visual Studio 2008, this could be C:\Program Files\Microsoft Visual Studio 9.0\VC\include\GL (create the GL subdirectory if it does not exist, or put the file in C:\Program Files\Microsoft Visual Studio 9.0\VC\include)

  • For Visual Studio 2003, this could be C:\Program Files\Microsoft Visual Studio .NET 2003\Vc7\PlatformSDK\Include\gl

  • Copy glut32.lib to your C++ library directory

  • For Visual Studio 2008, this could be C:\Program Files\Microsoft Visual Studio 9.0\VC\lib

  • For Visual Studio 2003, this could be C:\Program Files\Microsoft Visual Studio .NET 2003\Vc7\PlatformSDK\Lib

Linux

  • Install the freeglut package

  • On Fedora, “yum install glut-devel”

  • On Ubuntu, “apt-get install freeglut3-dev”

Hardware Drivers

Galil Motion Controller

CppUnit

We now use CMake External Project to download and patch CppUnit. This is convenient on computers that don’t already have CppUnit. For Linux and Mac OS, we still recommend to use the native package managers to install the development versions of CppUnit (using apt-get, port, …). On other computers (Windows mostly), you can now use the cisst CMake configuration option CISST_USE_EXTERNAL. When CISST_USE_EXTERNAL is turned on and you turn on the option to build the cisst unit tests (CISST_BUILD_TESTS), CppUnit sources will downloaded and compiled along cisst when you are building cisst. Please note that this requires a network connection for the first build.