effect for a render engine using new shading nodes. In short:
* No longer uses images assigned to faces in the uv layer, rather the active
image texture node is what is edited/painted/drawn.
* Textured draw type now shows the active image texture node, with solid
lighting.
* Material draw mode shows GLSL shader of a simplified material node tree,
using solid lighting.
* Textures for modifiers, brushes, etc, are now available from a dropdown in
the texture tab in the properties editor. These do not use new shading nodes
yet.
http://wiki.blender.org/index.php/Dev:2.6/Source/Render/TextureWorkflow
- Assume all areas are using valid MovieClipUser when using functions from movieclip.c
- Set active scene clip to movie-related nodes in compositor rather than checking
if there's only one clip in datablocks
- Fixed users count issue when solving camera motion
ColorMapping struct for color manipulation of textures. These will be the
standard built-in texture node options for manipulating the incoming texture
coordinate and outgoing color.
===========================
Rest of changes from camera tracking gsoc project.
This commit includes:
- New compositor nodes:
* Movie Clip input node
* Movie Undistortion node
* Transformation node
* 2D stabilization node
- Slight changes in existing node to prevent code duplication
Curve Map exec function (for RGB curves) used a broken test condition for constant factor=1.0f eval. This must take the factor buffer into account, checking the factor vec for default value is not sufficient. These kinds of test should use a function interface, there are probably a lot more bugs like this hidden in the code.
* Renamed one of the two "File is Saved" entries, as having two
entries with the same name in the Datablocks viewer was confusing
* Edited the tooltip text for "speed" option for dupliframes to
clarify what it does (or what it's supposed to do)
* Scene.use_shading_nodes property to check if RenderEngine is using new shading
nodes system, and RenderEngine.bl_use_shading_nodes to set this.
* Add mechanism for tagging nodes as being compatible with the old/new system.
- removed AVG2, was only used once.
- remove unused LONGCOPY define.
- removed BLI_STRUCT_OFFSET, was only used once, replce with offsetof
- formatting edits, split some macros over multiple lines.
This prevents access to non-existent typeinfo during type initialization,
when node types have been removed and such nodes are deleted from older files.
All blenkernel functions now only set the node->update flag instead of directly
calling the update function. All operators, etc. calling blenkernel functions
to modify nodes should make a ntreeUpdate call afterward (they already did that
anyway).
Editor/RNA/renderer/etc. high-level functions still can do immediate updates by
using nodeUpdate and nodeUpdateID (replacing NodeTagChanged/NodeTagIDChanged
respectively). These old functions were previously used only for setting
compositor node needexec flags and clearing cached data, but have become generic
update functions that require type-specific functionality (i.e. a valid typeinfo
struct).
The cause is an unknown node socket type in node groups. Ideally the node system
should handle this better and remove the unknown sockets from groups, but this
is a bit of a risky fix to do now, so instead the shader socket type has been
added, since this is a simple change and the code has been tested well.
The image node output is the default value when no image is selected. In pre-2.60 this was always initialized to 0 alpha, the defaults written in the node socket templates were completely ignored for outputs (except for value and RGB input nodes, which use these as button values). Now the stack values are initialized with the template defaults, which are all 0 by default. This patch changes alpha to 0 for image and render layer outputs too.
The problem was that all outputs got tagged indiscriminately (esp. hidden render layer sockets), leading to full recalculation every time. This was caused by erroneous tagging of bNodeStacks with hasinput/hasoutput flags. This patch restores the old behaviour of tagging all non-static stacks as input values and all outputs that are connected to some input. Only difference is in node groups, where the hasoutput flag is no longer abused for tagging internal buffers, here the is_copy flag is used instead.
