* Draw/Inflate/Layer now keep working on the original mesh coordinates and
normals from when the stroke started. This helps avoid the mesh blowing
up, but can still be better. The old behavior is still available as
"Accumulate" in the UI.
* This requires some more memory usage for the BVH, would like to find a
way to avoid that.
* Smooth falloff is now the default.
* Spacing is now enabled by default, with a value of 7.5.
* Anchored now stores normals per node to save some memory.
* Smooth: vert-face map is now only created when this tool is used, would be
best to also avoid using it here to avoid a sudden increase in memory, but
is not trivial.
* Grab: now no longer uses active verts list and loops over nodes like other
tools.
* Layer: uses original coordinates from undo now to save memory when not
using persistent layer.
* Anchored: this option works again now, though is still quite slow as it
loops over all verts/faces.
Smooth, layer tools and the anchored option could still be improved to use
less memory and/or work faster by only doing things per node.
that it is able to store changes in the mesh more compact than global undo.
It doesn't integrate well with multires yet, will tackle that when I start
looking into multires, for now still focusing on sculpt on regular meshes.
The weak point now is the thread-safe atomic access to normals from multiple
threads, did not seem to be a bottleneck in my tests but I don't really trust
it to be fast.
Recoded the way that Spline-IK computes the x+z axes of the bones so that flipping artifacts are minimised, and the rotation of individual bones can be used to affect the results of the solution, as per requests from Cessen.
The bone matrices are now computed normally, and then made to conform to the orientation + scaling imposed by the splines, using the Damped-Track method. Previously, the axes of the bones were calculated without regarding the prior orientation of other bones in the chain, which lead to "z-twists".
Notes for further investigation:
- There appears to be some shearing that gets introduced now. Unforunately, I can't seem to isolate the cause of this, but I hope it's not going to become too much of a problem in general.
- Maybe inverse corrections for rotation will now be necessary when using transform tools?
Non-ID pointers in DNA can only point to data from own ID block, so
now instead it uses an index into the particle system list, but still
exposed as a pointer through RNA.
* Sculpting, normal update and bounding box code is now multithreaded
using OpenMP.
* Fix a number of update issues: normals on node boundaries, outdated
bounding boxes, partial redraw, .. . There's probably still a few
left, but should be better now.
* Clicking once now does a single paint instead of two (was also
painting on mouse up event).
* Smooth shading now is enabled for the full mesh when the first face
uses it (so it can be tested at least).
Implementation Notes:
* PBVH search can now be done either using a callback or bt gathering the
nodes in an array. The latter makes multithreading with OpenMP easier.
* Normals update code is now inside PBVH, was doing it per node before but
should do all faces first and only then vertices.
* Instead of using search modes + 1 modified flag, now nodes get 4 flags
to indicate what needs to be updated for them, found that this makes it
easier for me to understand the code and fix update bugs.
* PBVHNode is now exposed as an abstract type, I think this makes it more
clear what is happening than having it's data passed as part of callback
functions.
* Active_verts list was replaced by looping over nodes and the vertices
inside them. However the grab brush still uses the active_verts system,
will fix that later.
* Some micro-optimizations, like avoiding a few multiplications/divisions,
using local variables instead of pointers, or looping over fewer vertices
to update the bounding boxes.
1) "Even Divisions" - This option ignores the length of bones when considering how they should fit along the curve. This is useful for getting a smoother curve fit without having to worry about getting the bone lengths spot on. By default, this is disabled.
2) "Keep Max Length" - This option prevents the bone chain from extending past its natural length when the spline is stretched beyond that length. When the spline length is substatially shorter though, this bones get scaled to zero; making this option possibly useful for doing "growing tips".
This is essentially a 'no scale' option, although the behaviour when the curve is shorter is really a compromise since the curve cannot be accurately satisfied + left intact without some scaling being applied due to the way this works.
3) "Radius to Thickness" - The average radius of the spline between at the head+tail of each bone determines the x+z scaling of the bone.
* Fixed crash when reloading a file with Spline IK and/or Damped Track constraints. The targets for these constraints weren't getting relinked.
* Fixed problems with removing Spline IK making some bones unable to be manipulated.
* Jotted down some comments in the Spline IK code noting places where additional tweaks will be added.
For now, this just assumes that the 'lens' parameter was animated (assuming a perspective lens was used). Unfortunately, this may not always be correct, but at least there's a path now that can lead to further tweaking.
At last, this commit introduces the Spline IK Constraint to Blender. Spline IK is a constraint that makes n bones follow the shape of a specified curve.
Simply add a chain of bones, add a curve, add a Spline IK Constraint to the tip bone and set the number of bones in the chain to make it work. Or, try the following test file:
http://download.blender.org/ftp/incoming/250_splineik_spine01.blend
Screenshots of this in action (as proof):
http://download.blender.org/ftp/incoming/b250_splineik_001_before.pnghttp://download.blender.org/ftp/incoming/b250_splineik_001_after.png
I've implemented this in a similar way to how standard IK solvers are done. However, this code is currently not an IK plugin, since I imagine that it would be useful to be able to combine the 2 types of IK. This can be easily changed though :)
Finally, a few notes on what to expect still:
* Constraint blending currently doesn't affect this. Getting that to work correctly will take a bit more work still.
* Options for not affecting the root joint (to make it easier to attach the chain to a stump or whatever), and non-uniform scaling options have yet to be added. I've marked the places where they can be added though
* Control over the twisting of the chain still needs investigation.
Have fun!
Limitations:
1) Parents and children of selected objects are excluded from the pool (siblings are ok) Making it work with that would required unparenting and reparenting after transform, that would turn nasty really quick.
2) Does not support Connected (this could be done through parent links, but see 3 first).
3) Parent relationships in affected objects aren't taken into account. When parent and children in the area of effect, remember that the children will also take the motion of the parents (with additive results). This could perhaps be fixed, but it could be nasty.
Other stuff:
New BASE_EDITABLE macro that checks if base is editable (like TESTBASELIB except it doesn't check for selection)
Add scene parameter to TESTBASELIB_BGMODE macro (using it from current scope is nasty)
This is effectively a C-port of Nathan Vegdahl's "No Twist" TrackTo PyConstraint, and has been added as a separate type of constraint to be consistent with the existing constraints (Locked Track, and Track To).
In general, this works considerably better than the existing "Track To" constraint, since it works by determining the smallest rotation necessary to get the current orientation of the owner to an orientation which would be tracking the target. It is also a much more straightforward approach than the weird old method the old Track To uses.
I've made a few tweaks to the code to deal with the (hopefully rare) cases where the target and the constrained are coincident. These don't appear to cause too much trouble in general.
TODO:
- Probably the naming of the constraints will change, to better convey their purposes. Naming suggestions welcome.
- undo stops all running jobs (operator redo was crashing with threaded render)
- adding new armatures was crashing if there was no valid view3d
- transform with an active hidden object would crash
The aim of this is to avoid having to set the selection each time before running an operator from python.
At the moment this is set as a python dictionary with string keys and rna values... eg.
C = {}
C["active_object"] = bpy.data.objects['SomeOb']
bpy.ops.object.game_property_new(C)
# ofcourse this works too..
bpy.ops.object.game_property_new({"active_object":ob})
# or...
C = {"main":bpy.data, "scene":bpy.data.scenes[0], "active_object":bpy.data.objects['SomeOb'], "selected_editable_objects":list(bpy.data.objects)}
bpy.ops.object.location_apply(C)
* There's an unresolved error in transform_conversions.c which I've flagged in this commit. I'm not quite sure what the exact intentions of that code were (i.e. was the "void_pointer = 1" really intended)
- Generated and uploaded api docs - http://www.blender.org/documentation/250PythonDoc
- Added Edit docs menu item & operators as discussed with Mindrones, Brecht, Stani & Letterip @ bconf, needs some web backend. python operator can aparently use xml/rpc to upload docstrings.
- Added operator invoke function - context.manager.invoke_props_popup(self.__operator__, event)
this calls a popup for invoke by default (which intern calls execute())
- Own recent commit to game framing applied to non-camera views too.
- v3d->persp is deprecated but still used in some places.
- Transforming strips could overlap 1 frame if moving them below frame 0
- Transforming overlapping strips could go into an eternal loop (though overlapping strips should not exist)
* #19727: Noise modifier does nothing with size 1.0
When the 'Size' and 'Phase' parameters were both 1.0 exactly, and evaltime was an integer (as is the case when doing animation evaluation but not for Graph Editor drawing), the noise calculation function was bailing out. Now, the 'z' component supplied to this function is a decimal value (hardcoded to 0.1 after experimentation) to try and avoid this situation.
* Graph Editor 'Bake' operator was using wrong poll callback, making it useless when trying to use it on a F-Curve that only has modifiers on it (i.e. the main use case of the operator!)
Blended shape keys can now be displayed & edited in edit mode. This
is much like showing an armature modifier in edit mode, and shape keys
now are a applied as a virtual modifier (for mesh & lattice only, curve
doesn't fit in the stack well due to tilt).
The main thing missing still is being able to switch between the active
shape key in edit mode, that's more complicated.. but the weights of
other shapes can be edited while in edit mode.
One thing to be careful about is that this does automatic crazyspace
correction, which means that if you edit a shape key with a low value,
the actual vertices will be moved to correct for that and actually move
a (potentially much) longer distance.
Also includes some UI tweaks, mainly placing some buttons horizontally
since the vertical list was getting too long.
Internal change to not apply the shape keys to the Mesh vertex coordinates,
but rather use it as part of the derivedmesh/displist evaluation. This only
has one practical advantage right now, which is that you can now make a
linked duplicate and pin it's shape key to a different shape than the first
object.
Further, this makes shape keys correctly fit into the modifier stack design,
which will help implement some other features later. Also it means the mesh
vertex coordinates are now really the orco's.
* RNA Path fixing when renaming data now checks if a path in question cannot be resolved before trying to fix it. This should reduce the number of misindentified cases I hope.
* Silenced compiler warnings for EdgeSlide stuff that mingw was making about unused variables.
* Transform code was not properly fixed to work with the new way that axis-angle data was stored
* The order of the args for the conversion function when switching rotation representations was wrong, causing problems when switching from quaternion to axis angle (i.e. these occurred for newly created bones).
Auto save is now working again in 2.5. It will also remember now what
the location of the original file was when recovering it, so that
library links still work and saving the restored file does not save to
the temp directory. There is also a new Recover Auto Save operator
which will open the filebrowser in the temp directory and show the
auto saved .blends.
Implemenation Notes:
* Timer storage was moved from window to windowmanager, so we can have
windowmanager level timers too now, doesn't make sense to have
autosave timer attached to a particular window.
* FileGlobal now has a filename field storing where the file was saved.
Note that this is only used when loading a file through the recover
operators, regular file read doesn't use it, so copying the quit.blend
manually over the original file will still work as expected.
* Jobs timer no longer uses operator now, this seems more like an
internal thing, changing keymaps should not make it possible to break
the jobs manager.
* Autosave is postponed by 10 seconds when a modal operator is running,
e.g. transform or file browsing.
* Moved setting G.sce in setup_app_data before depsgraph updates, these
can use the filename for pointcaches.
* Now the old/new names get tagged with [" "] before the search and replace operation, which should alleviate problems with searching for 'bone' and ending up with all instances of 'boney' 'boney.r' etc. also getting renamed.
* Cleaned up some compiler warnings, and removed an unused function from an earlier attempt at this work.
RNA Paths used in F-Curve, Drivers, etc. now get renamed when some data that they use gets renamed. This only works when things like Bones, Constraints, Shape Keys, and Modifiers get renamed, but other cases can get added easily.
The code here only performs simple string replacements, so there is the potential for problems when several sets of data with the same names are present. For example, if there are multiple armatures with bones that have the same names, renaming a bone on one armature (with a bone on another armature having the same name) will break all the drivers on the other one, even though they aren't really connected. However, I don't expect the aforementioned scenario to really be a problem in most production scenarios.
