What works:
The usual command line options for rendering.
All python scripts are loaded (which includes custom properties)
Render engines are loaded and can be used
-P to run scripts works partially: rna api works ok, not operators.
What doesn't:
Most operator calls in python. This is a problem with poll functions. (Brecht and Campbell are aware of this already)
Changes:
-d now also applied with -b (it was ignored before)
user file (.B25.blend) now also loaded in bg mode. This helps for custom paths and all.
wm is also initialized (it's needed for a lot of context calls)
Ghost, however, is not initialized.
This allows you to set and animate the values of socket inputs and outputs, for example the value node.
It's also a step on the way to manipulating node trees via python (i.e. linking node sockets to each other).
This fixes [#19841] RGB Node in compositor not working
* Convert all code to use new functions.
* Branch maintainers may want to skip this commit, and run this
conversion script instead, if they use a lot of math functions
in new code:
http://www.pasteall.org/9052/python
* Fix remaining issues before conversion.
* Inline various vector functions, currently enabled for all platforms.
I expect this to work in GCC/MSVC at least, if other platforms don't
support it, #ifdef's can be added.
this way the UI scripts are less likely to fail silently and wont let typos work ok.
also allow subclassing of the context, added a copy function,
bpy.context.copy(), returns the context as a python dict to be modified and used in python.
This also showed up an invalid brush member in the screen context.
The timestep used for the playback timer now gets adjusted accordingly when the frames-per-second setting gets changed during playback (i.e. in the "animation_step" operator). This is not as ideal as only updating this when the framerate setting is changed, but using an appropriate update function for this failed miserably.
Still a few quirks, including redraw issues on multilayer image input nodes, but it's pretty much there.
Would also be good to wrap the input/output sockets, too, will check on it.
This fixes bug [#19740] INPUT NODE: Cannot load images / motion pictures
initializing Freestyle and specifying contexts, making the API a bit messy.
Now FRS_initialize() is only for initialization, and contexts are specified
by new FRS_set_context() function just before starting rendering.
* Fixed the handling of the 'draw_active' flag for drawing of armatures. This is now cleared from bones in old files (so one bone always got represented as active in the viewport even when others were selected), and the flag is correctly set temporarily when drawing the bones (only one place had been done).
* Fixed typo with SplineIK that was making the root bone of the bone chains always be ignored. Similar functionality can come back at some point, but in a more useful form.
* Shortened the UI names for the F-Curve colouring modes to increase readability. The old ones were too long to be able to distinguish between entries in the UI.
- rna vars arm.bones.active & rna.edit_bones.active
- needed special undo support.
- readfile.c loads.
- duplicate and copy_armature support.
- keep the draw flag, renamed to BONE_DRAW_ACTIVE, only use for openGL drawing.
Note: it may be better to allow active/unselected as with objects.
- QuickTime export fixes.
Note that QuickTime export still crashes because it tries to open a "codec settings" dialog from the rendering background thread (and not the main/UI thread).
One quick fix may be to move the movie export initialization out of the render thread back into the operator function.
But a cleaner way would be to get rid of such a carbon/win32 dialog and place the codec settings inside blender interface
(additional fields in the output panel as it is currently the case for other file formats ?).
