- trying to determine bone orientation from joint matrix
Looks like DAE joint matrix orientation cannot be relied upon as a source for bone direction. It can be axis-aligned or pointing in
totally different direction:
http://img75.imageshack.us/img75/8778/flufty.png
It seems to me the best solution is to add a bone with custom draw type that has no pointing direction, sphere for example. Will code
this after merge!
For each object in scene linked with armature writing the following information:
- bone names
- vertex weights
- bone inverse bind matrices (fake values for now)
Assuming objects don't share Armatures for now.
- splitted code into separate classes: ArmatureImporter, MeshImporter
- commented out texture-to-mesh binding code which was difficult to comprehend, Chingiz will re-think/re-write
When the mesh field is left blank and Physics option is enabled, it reinstances the physics mesh from the existing mesh.
like calling gameOb.reinstancePhysicsMesh() from python.
Add support back for reinstancePhysics mesh, a frequently requested feature in the BGE forums.
from what I can tell Sumo supported this but bullet never did.
Currently only accessible via python at the moment.
- rigid body, dynamic, static types work.
- instanced physics meshes are modified too.
- compound shapes are not supported.
Physics mesh can be re-instanced from...
* shape keys & armature deformations
* subsurf (any other modifiers too)
* RAS_TexVert's (can be modified from python)
Moved the reinstancePhysicsMesh functions from RAS_MeshObject into KX_GameObject since the physics data is stored here.
video and blend file demo.
http://www.graphicall.org/ftp/ideasman42/reinstance.ogvhttp://www.graphicall.org/ftp/ideasman42/reinstance_demo.blend
Since in DAE only a matrix is stored per bone, some tricks have to be applied to build an armature from it:
- direction for "leaf" bones is hard-coded to Y vector. This will be changed to bone matrix orientation, somehow code will need to
derive the bone "forward" axis
- for "leaf" bones length cannot be determined, currently it is hard-coded (second screenshot), but I think changing it to the average
length of all non-leaf bones is ok
- parent bones get their tail at last child's head, this is not good (first screenshot). For bones with more than one child Chingiz
suggested not to link it with any child, I think it's a good idea. Length will be the same as for leaf bones.
Screenshots:
http://img232.imageshack.us/img232/8905/screenfhy.pnghttp://img406.imageshack.us/img406/3004/tobaggan.png
This adds a RenderEngine type to RNA, which can be subclassed
in python (c++ will follow once we support subclassing there).
It's very basic, but plugs into the pipeline nicely. Two example
scripts:
http://www.pasteall.org/6635/pythonhttp://www.pasteall.org/6636/python
Issues:
* Render runs in a separate thread, and there is unrestricted
access, so it's possible to crash blender with unsafe access.
* Save buffers and full sample are not supported yet.
* ID blocks can now get RNA properties defined from python, e.g.:
bpy.types.Scene.BoolProperty(..)
* RNA structs/functions/properties can now get pointers duplicated
(mostly strings), since we can't point to some static string then.
* Added ExtensionRNA struct to add into *Type structs for subclassing,
is a bit more compact than defining the 4 variables each time.
Only disadvantage is it requires including RNA in more places.
* Windows fixes for texture filter & bump patches, thanks
Jean-Michel Soler for noting.
* Added sqrtf/sinf/fabsf/... fallback #ifdefs in BLI_arithb.h,
those should be safe to use now. Replacing the double for the
float version throughout the code can be done once, but would
need proper testing.
Patch by Alfredo de Greef. Considerably improves the quality of bump
mapping, and texture filtering for displacement and warp too. Mainly
this is achieved by getting the texture derivatives just right in
various cases, many thanks to Alfredo for figuring this one out, works
great.
This is enabled by default now, but disabled still for existing
textures to preserve backwards compatibility. Can be enabled with
the "New Bump" option in the material texture slot in the outliner.
Also, I made the range for the normal factor a bit smaller since this
gives stronger effects, but note that you can still type in larger
values than the slider allows.
Patch by Alfredo de Greef with high quality image texture filters.
This adds 3 new filters:
* SAT: Summed Area Tables. This is like mipmaps, but using somewhat
more memory avoids some artifacts.
* EWA: Ellipitical Weighted Average, anisotropic filter.
* FELINE: Fast elliptical lines for anisotropic texture mapping.
The one change I made to this was to try to fix an alpha/premul
problem, hopefully I didn't break anything, it looks compatible
with the existing filter now for me.
* Buttons in header now use operators too. The paste-flipped button needs attention though, since the flipped argument isn't set yet
* Assigned Ctrl-C, Ctrl-V, and Ctrl-Shift-V to Copy/Paste/Paste-Flipped respectively for now.
* Auto-Keying for this doesn't work again yet. On todo for later...
---
* Also, new armatures now get the flag to show custom bone colours enabled by default.
* World and Lamp previews now working here too.
* Experiment with list template, showing only icons. Unfortunately
texture icon render crashes combined with preview render so it
shows all icons the same.
* Influence panels updated, with slider for each option. The values
are still linked though, will fix that later.
* Image texture controls a bit more complete, still WIP.
* Color ramp back.
* Added a new UI Template for the 3-colour picker used to visualise + select the custom colours for a bone group.
* Finished wrapping the colour properties for Bone Groups in RNA. Although changing the colour-set used will change the displayed/cached colours, changing the colours via the colour wells will not change the colour set to 'custom' (as per 2.4x) yet. This needs a nice solution...
* Fixed context-related bugs with the Assign/Remove operators for bone groups. These were using context-iterators for selected posechannels, but that was only defined/valid for the 3d view (but not for the buttons window), hence a failure in that case.
Modal keymaps.
I've tried to make it as simple as possible, yet still using sufficient facilities to enable self-documenting UIs, saving/reading in files, and proper Python support.
The simplicity is: the 'modal keymap' just checks an event, uses event matching similarly to other keymap matching, and if there's a match it changes the event type, and sets the event value to what the modal keymap has defined. The event values are being defined using EnumPropertyItem structs, so the UI will be able to show all options in self-documenting way.
This system also allows to still handle hardcoded own events.
Tech doc:
1) define keymap
- Create map with unique name, WM_modalkeymap_add()
- Give map property definitions (EnumPropertyItem *)
This only for UI, so user can get information on available options
2) items
- WM_modalkeymap_add_item(): give it an enum value for events
3) activate
- In keymap definition code, assign the modal keymap to operatortype
WM_modalkeymap_assign()
4) event manager
- The event handler will check for modal keymap, if so:
- If the modal map has a match:
- Sets event->type to EVT_MODAL_MAP
- Sets event->val to the enum value
5) modal handler
- If event type is EVT_MODAL_MAP:
- Check event->val, handle it
- Other events can just be handled still
Two examples added in the code:
editors/transform/transform.c: transform_modal_keymap()
editors/screen/screen_ops.c: keymap_modal_set()
Also: to support 'key release' the define KM_RELEASE now is officially
used in event manager, this is not '0', so don't check key events with
the old convention if(event->val) but use if(event->val==KM_PRESS)
* Added Bone Groups UI to 'Armature' context buttons for now. Later, it may be more convenient to have these with bones instead?
* Added operators for the operations that can be performed on these groups. Moved the core adding/removing functions to blenkernel so that they can be used elsewhere in future if need be.
* Properly wrapped bone groups in RNA. Copied the way that Vertex Groups are wrapped, since they share some similarities. Setting colours for bone groups still needs more work though.
for a while py2.x will work but eventually be dropped when most OS's support it, so Id recommend upgrading.
The following instructions are only needed if you don't use python3.1 installed in the default location.
For releases users wont have to worry about this.
# in python3.1 source dir, build and install into your own dir, /opt/py31 is just an example.
./configure --prefix="/opt/py31"; make; make install
# In the scons user-config.py...
BF_PYTHON = "/opt/py31"
# ... now build ...
#
# Blender now needs 2 things to run. ./lib/libpython3.1.so and the python modules.
# Symlink (or copy) python modules, blender sets this path for modules on startup if it is found.
ln -s /opt/py31/lib/python3.1 ~/.blender/python
# Currently static linking is not working without hacks because of limitations in scons.
# for releases we can workaround, but for now its easier to set an environment variable.
# To start blender so it can find libpython3.1.so make this into a shell script to save yourself typing it in all the time.
export LD_LIBRARY_PATH="/opt/py31/lib/"
./blender
Moved Shade Smooth/Flat from Mesh obdata panel to tools area. These kinds of operator tools aren't really allowed in the buttons window anymore - whole point of new tools area :)
The only operators that are allowed in buttons window are things that act on the RNA fields, like add/remove buttons for adding vertex groups etc.
