Blender Git Commits

style cleanyp: split > 120 width lines.

fix for incorrect check to find largest axis (was checking Y twice, not Z).

replace MIN2 / MAX2 with minf / maxf to avoid calling functions multiple times.

fix for issue with bevel when the 2 edges form a straight line. this isnt really fixing the bigger problems but the error did show up baddly as an artifact.

svn merge ^/trunk/blender -r44213:44235 --- fixes bmesh shading bug [#30125]

Change texface draw to always use glShadeModel(GL_SMOOTH), with texface draw.

without this each face would get a solid color, this is the same method used in object mode.

also copy BLI_array.h fix from bmesh branch.

skip per face texface poly lookup while drawing when the layer isnt found.

also remove some unused code.

left print in by mistake

edit-derived-mesh had a table for vert/edge/face lookups - that were never used.

removed these and replaced them with calls to BM_vert_at_index, BM_edge_at_index - which are not ideal but same as trunk does.

Fix rest of #27022, collada export: add bone parenting of objects

- SceneExporter collects a list of child-objects for armature-object and passes it onto ArmatureExporter
- SceneExporter's writeNodes is then called from ArmatureExporter for matching child-objects for bone.
- ArmatureExporter removes written child-objects from list, objects not exported as being bone parented are exported as direct children of the armature-node.
- Should play nice with current Second Life-compatibility.

A nicer implementation would require some design changes, will have to wait.

Wring scaling and precision error in previous commit.

Proxies: FFmpeg proxy builder wasn't taking image quality into account at all
which made it using default quality settings which are really bad for camera
tracking (and perhaps for CSE too).

haven't found Jpeg quality setting for FFmpeg which will behave in the same way
as quality setting for image sequence, but seems that mapping image quality
from 1..100 UI range to 31..1 range of qmin/qmax gives expected result.

Collada export: changes to joints/weights in skincontroller

- Collecting joints/weights in one place, it's easier to exclude zero weights or vertexgroups with no matching bone than trying to match same logic in multiple places.
- Still not exporting -1 joints for vertices without weights, but also not outputting -1 joint + weight for each vertexgroup without a matching bone.
- The exported weights are now normalized.

Last I tested this patch stopped 3ds Max crashing on import of file from #29465 (opencollada / internal .dae).

Merging r44140 through r44226 from trunk into soc-2011-tomato

Minor change collada export; don't even run the parentinv-test when ob->parent is NULL

BGE bug #30173: ImageRender not working when initialized on frame 1. This was because the canvas display area was initialized on render frame 1, which happens after logic frame 1. Fixed in blender and player by initializing the area on canvas creation.

minor bmesh api changes.
- use SMOOTH flag for edges (removed sharp flag and inverted), gives us an extra flag slot available.
- rename BM_edge_is_nonmanifold --> BM_edge_is_manifold and invert result, BM_vert_is_nonmanifold too.

Remesh modifier: extensive refactoring of the Octree class.

The changes mostly center around two new structures, InternalNode and
LeafNode. These provide an explicit representation of the Octree
nodes, which formerly were manipulated as opaque byte arrays.

A fair amount of commented out/unused code was also removed. This
includes the "CINDY" code, which may yet be useful, easy to bring back
if so.

There should be no difference in the output of the remesh modifier,
but memory usage may be slightly different. The flood fill bytes are
no longer optional; they will be allocated whether or not the 'remove
disconnect components' flag is set. The leaf node is probably not as
tightly packed due to alignment issues; this could be fixed with the
__attribute__((packed)) flag in gcc (probably there's an MSVC
equivalent), but not sure it's worth it. The internal nodes should
take up less space on 32-bit systems, allocating sizeof(pointer) now
rather than constant eight bytes.

These changes were made in persuit of bug #30158 (remesh crashes on
PowerPC). There's still a fair amount of bitwise stuff in the Octree,
so may still be endian issues and not yet sure if this fixes the bug,
but should be much easier to track down problems now.

Merged changes in the trunk up to revision 44221.

Conflicts resolved:
release/datafiles/blender_icons.png
source/blender/blenkernel/intern/anim_sys.c
source/blender/blenloader/intern/readfile.c

#collada animation import, bugfix #29082

Fix skeletal animation import for <rotation>, <scale> and <translate> transform types.

Tested for correctness visually using Seymouranim2.dae from http://collada.org/owl and animated creatures from DKGamesModels folder (from private section of the same site). The results match the results in FXComposer-2.5.

Since this is a fix towards correct reading of collada, it shouldn't break existing compatibility with Second Life.