* Updated GlGradientVolumeGenerator to generate GL_RGB16F textures to save memory

* Updated TensorDemo pipeline to show fused visualization of slice and glyph rendering

Changed log level of some warnings from DEBUG to WARNING, disabled some verbose debug messages in TGT

* Fixed TrackballNavigationEventListener not saving the current scene bounds -> resetting camera too often

* Implemented GeometryDataFactory::createCube() normal generation

Fixed GeometryRenderer rendering lines and points always with wireframe color and ignoring line width.

* Fixed SliceExtractor not rendering Y slice position correctly in XZ-plane mode
* Minor improvements in DataContainerInspector fragment shader and GlGradientVolumeGenerator

AutoEvaluationPipeline now also looks recursively for DataNameProperties in MetaProperties of its processors. Hence, it simulates ports also for nested processors. Furthermore, improved AutoEvaluationPipeline's documentation.

* RaycastingProcessor now has the p_targetImageID property and optionally uses a custom GLSL version for the shader
* Introducing PropertyCollection::getNestedProperty allowing to get nested properties with a single method call
* Fixed StringUtils::split() methods being broken with multi-character delimiters

C++ implicit conversion rules made ShaderManager::loadSeparate() ambiguous in some cases. So I decided to cut loose ends and completely refactored loading with standard version into ShaderManager::load() and loading with custom version into ShaderManager::loadWithCustomGlslVersion().
ShaderManager::loadSeparate() is no longer available!

* Fixed VolumeExplorer not always correctly invalidating sub-processors
* Updated ProxyGeometryGenerator to new Processor design (using INVALID_PROPERTIES invalidation level)

Fixed GlReduction, now working reliably for 2D and for 3D images. Added test pipeline GlReduction test.

* Extended GeometryRendererDemo to render two different geometries and combine the renderings using the RenderTargetCompositor
* Fixed point size in GeometryRenderer

* Introducing MultiIndexedGeometry data type
* Introducing GeometryDataFactory::createTeapot()

 * Supports different render modes: GL_POINTS, GL_LINES, GL_LINE_STRIP, GL_TRIANGLES, GL_TRIANGLE_STRIP, GL_TRIANGLE_FAN, GL_POLYGON
 * Supports optional wireframe rendering using Single-pass Wireframe Rendering technique
 * Supports calculation of per-face normals, if not present in geometry
 * Fixed shading

Refactoring AbstractProcessor::process() for clearer semantics and better and more uniform handling of invalidation levels:
 * AbstractProcessor::process() now calls updateShader(), updateProperties() and/or updateResult() with respect to the current invalidation level
 * each processor shall no longer override process() but the updateXYZ() methods, at minimum updateResult()
 * AbstractProcessor::process() takes care of (un)locking the processor itself (no need to do this from the outside anymore)

Further implicit changes:
 * Removed redundant HasPropertyCollection::updateProperties()

Completely moved the image data handle and histogram computation from the TF to the TF property. Histogram computation is now thread safe.

Started refactoring the TransferFunctionProperty: The optional ImageData handle will be stored in the TF property instead of the TF itself. Furthermore, the property now supports automatically fitting the TF window to the image data.

 * Introducing campvis::Tensor2<T> type
 * Updated MhdImageReader to support TensorImages