svn revision 4100
Major new features:
* Support for vector node types: Line, PolyLine, Curve, Circle, Polygon,
Rectangle. Vector nodes can be textured, with the closed vectors taking two
textures: one for the outline, one for the fill. Vector drawing is GPU-based.
* Plugin support. Plugins written in C++ can define complete new node types.
* Powerful (but still experimental) new animation framework.
* Unified & much more stable camera drivers. Consistent support for all
resolutions & framerates we could test. Support for more than one camera on all
platforms. fw800 is supported.
* Uniform support for aggregate types in node interfaces (node.pos, node.pivot,
...)
Other improvements:
* gcc 4.4 and ffmpeg 0.5.0 compatibility.
* Added letterspacing support to words node.
* Threaded videos are now the default.
* Improved video seek and loop performance.
* New Node::getElementByPos() method that returns the node at that point on the
screen.
* Added avg.svnrevision and avg.svndate properties so the avg version can be
queried.
* New avg_showfont.py utility to help font selection.
* New avg_showcamera.py utility to help camera setup.
* New avg.getMemoryUsage function that returns the resident set size.
* New Words.addFontDir().
* Added support for mipmapping node textures, including videos.
* rasternode.maskhref to support static alpha masks for any videos, word nodes,
etc.
Syntax and semantic changes:
* In the course of adding aggregate types for node attributes, several attributes
have changed:
- words.size is now called words.fontsize.
- pivotx and pivoty have been replaced by pivot.
* right-aligned and centered words nodes are positioned differently.
* avgtrackerrc syntax has changed significantly; see src/avgtrackerrc.minimal.
* Support for firewire cameras using libdc1394 ver. 1 has been phased out. Please
install ver. 2.x.
* The old python-based anim framework is marked deprecated. You need to explicitly
use the namespace anim to access it.