Blender Git Loki

OSX: revert default compile to i386, for it breakes QTKit-condition for some unknown reason

Updating render layer and image output sockets was not working due to changes in snode_set_context. Before the r39941 this used a hack to 'test for first drawing', but since the snode->nodetree was set to NULL each time at the start of that function the condition would always evaluate true anyway. Simply removed that check to restore previous behavior.

bmesh: proper sculpting on non-multires object

Problem was caused by unneeded call of CDDM_calc_normals when
creating derived mesh for object which duplicated mvert array.

It shouldn't happen unless object has got shapekeys or deformation
modifiers on the top of stack.

Fix for crash when duplicating image or render layer nodes.

These nodes use cached buffers (in sock->cache), which need to be cleared when duplicating the node, otherwise the buffer will be freed twice.

bmesh: corrected fix for svulpting on non-multires mesh

Previous fix wasn't correct. It made things working, but it removed
advantage of using PBVH for speed because derivedMesh used to be
recreated on each stroke step.

This commit fixes non-multires sculpting for Solid shading,
wireframe and render wouldn't work correct -- it'll require
a bit more tricky things because DM created for base mesh
doesn't points to mesh->mverts anymore, so sculpting and
drawing modules are using different verts arrays.

bmesh: small fixes for compilation and subdivide operator

Fix #28541: crash with operator reports in some cases, caused by my recent fix
for missing reports, thanks to Sergey for tracking this down.

Fixed offset for socket remove buttons ('x') in node groups.

Apply patch [#28415] 3d mouse orbit mode individual axes enhancement
Submitted by Rainer Wahler

This patch enables individual axis inversion in trackball mode.

Code for setting/getting/freeing executable nodes in compositor was still using the plain nodetree->nodes list, which is not dependency sorted any more. This caused missing updates for viewers and other nodes. Instead this code now uses the execution data generated before, which has a correctly sorted list of bNodeExecData.

Initialize the exec data in for composite with the node tree pointer. Doesn't matter now since composites are not executed in parallel, but just to be safe.

Patch 28537 from Andrew Wiggin, fixes second issue in 28492 (edge ring deselect not working

Camera tracking integration
===========================

- Fix for pixel aspect: solver supposed claibration was made against
aspected image, manual calibration worked in non-aspected image.
- Fixed incorrect usage of sensor size when creating projection matrix
for tracking structure.

Camera tracking integration
===========================

- Added python method MovieClip.tracking.tracs.add() to add new track.
Marker for this track is getting created on position (0, 0) and it's
keyframed to specified frame number.

- Implemented rotation 2D stabilization. It's supposed to be used in the
following workflow:

* Set tracks which defines location stabilization.
* Choose one track which defines orientation.
* Vector which connects median point of location and track used for
rotation would have the same orientation along all frames.

2D stabilization with rotation can be a bit slow.

- Some internal refactor and code clean-up.

Merging r39984 through r39997 from trunk into soc-2011-tomato

Merging r39717 through r39983 from trunk into soc-2011-tomato

interface + naming improvements to vertex wright modifier
- WeightVG -> Vertex Weight
- mapping_mode -> falloff_type
- nicer layout for VertexWeightModifiers add/remove options

Another fix for value nodes, those need full template definitions for their output sockets, since they use their range limits for buttons.

SVN maintenance.

Attempt to fix compile error on OSX