Get Started!

WITH_BF_OPENMP = True
WITH_BF_FFTW3 = True
WITH_BF_GAMEENGINE = True
WITH_BF_OPENEXR = True
WITH_BF_FFMPEG = True
WITH_BF_OPENAL = True
WITH_BF_JACK = True
WITH_BF_SNDFILE = True
WITH_BF_SDL = True
WITH_BF_DDS = True
WITH_BF_BULLET = True
WITH_BF_FREETYPE = True
WITH_BF_COLLADA = True
WITH_BF_PLAYER = True
WITH_BF_RAYOPTIMIZATION = True

build on Ubuntu 11.10

tested on Ubuntu 10.10 (maverick), Ubuntu 11.04 (natty), Ubuntu 11.10 (ocelot), Ubuntu 12.04 (pangolin), Fedora 16

http://twitter.com/derfish

---------------------------------------------------------

changes:

Commit by campbellbarton :: r46603
fix geometry to origin in mesh editmode - was broken since bmesh merge.

Commit by campbellbarton :: r46602
code cleanup: use vector math function minmax_v3v3_v3() and other minor vector function edits.

Commit by dingto :: r46601
Sequencer / Clip Editor:
* Make it more clear that the Memory Cache Limit is used by the Clip Editor as well.

Commit by dingto :: r46600
More work for [#31429] Curve menu polish.
* Curve Extrude and Duplicate from Toolbar also moves the points now, same as for Mesh and in alignment with the menu.
* Fixed yet another tooltip.

Commit by campbellbarton :: r46595
code cleanup: use static functions for transform operators

Commit by campbellbarton :: r46594
add access to bevel weight modal operator (only available in the Shift+E edge menu currently)

Commit by campbellbarton :: r46593
patch [#31434] IDMaterials.pop/append() UI update from console
from Dan Eicher (dna)

Commit by campbellbarton :: r46592
code cleanup: missed some min/max use of float[3], also found case of DO_MIN/MAX re-reading same value from array
4 times when it can do once (use minf rather then MIN2, same for maxf)

Commit by campbellbarton :: r46591
code cleanup: minor improvements to float/vector usage.

Commit by jesterking :: r46590
Disable sm_13 CUDA kernel building for now, as the build fails with out of memory.

Commit by campbellbarton :: r46589
previous commit didnt work right, the flag is still needed, wrap access with functions so the NULL check is
always done.

Commit by jesterking :: r46588
Enable setting of CUDA kernels to compile on command-line:
python scons/scons.py BF_CYCLES_CUDA_BINARIES_ARCH=sm_20,sm21

Commit by campbellbarton :: r46587
replace GP_PAINTFLAG_STROKEADDED with a NULL check, saves worrying about keeping the flag correct after undo.

Commit by campbellbarton :: r46586
fix [#31427] Crash when undoing until no point in grease pencil poly mode

Commit by campbellbarton :: r46585
fix [#31428] bpy.ops.wm.path_open only works once

Commit by campbellbarton :: r46584
fix for own bad logic with polygon normal calculation, was reading one past the loop array (reported as bug #31431).

Commit by dingto :: r46583
3D View:
* Changed Duplicate and Exture Operator fpr Curves/Surfaces to Duplicate/Extrude and Move, like for Mesh objects,
for consistency. This is also consistent with Shift+D and E shortcut.
This fixes [#31429] Curve menu polish.
* Fixed a wrong tooltip for metaball duplication.

Commit by campbellbarton :: r46582
code cleanup: header cleanup and remove some duplicate defines.

Commit by campbellbarton :: r46581
style cleanup: blenkernel

Commit by mont29 :: r46579
Fix [#31426] Unwrapping, First unwrap unwraps with LSCM, not with Angle Based
When operator method property is not set, the operator uses the one from current scene settings. We must update
the operator property accordingly...
Also updated default startup.blend file, to match default AngleBased method (was on Conformal one).

Commit by campbellbarton :: r46578
code cleanup: some style and use math defines, also small speedup for dynamicpaint, dont call acosf twice.

Commit by campbellbarton :: r46577
style cleanup: whitespace

Commit by campbellbarton :: r46576
style cleanup: remaining BLI files.

Commit by campbellbarton :: r46575
style cleanup: whitespace, bli & makesdna

Commit by campbellbarton :: r46574
optimization for select similar
- use angle_normalized_v3v3() where both vectors are known to be normalized.
- remove needless radian to degrees conversions.
- move checks for customdata layers outside the inner loop (for bevel and crease).

Commit by campbellbarton :: r46573
add bevel to select similar edges operator

Commit by campbellbarton :: r46571
style cleanup: mostly whitespace in rna

Commit by campbellbarton :: r46570
fix for building without libmv

Commit by halley :: r46569
reworded tooltip for Tripod Motion button; fixing spelling and making more clear

Commit by campbellbarton :: r46560
fix [#30905] Operator preset save doesn't serialize PropertyGroups
fix based on patch by Julien Rivaud (frnchfrgg), but re-written a bit differently not to be operator specific.

Commit by campbellbarton :: r46559
don't stat bookmark files on load, can make blender hang on slow networks (eg, the blender institutes gigabit
connection)... or if the server is busy.

Commit by nazgul :: r46557
Added verbosity command line argument. Currently used for libmv only.
Usage: ./blender --verbose 
Also fixed some crashes when not passing number to --threads argument
or not passing file format to -F argument.

Commit by jesterking :: r46556
Enable OpenCOLLADA support for MingW64 builds by default.
Note: don't forget to update your local checkout of lib/mingw64

Commit by lukastoenne :: r46554
Fix for UI layout alignment of fixed-size items (including those with width < min. width). Such layout items were
ignoring the center and right alignment in layouts before.

Commit by campbellbarton :: r46553
style cleanup: whitespace

Commit by campbellbarton :: r46551
add missing NULL of scene->ed

Commit by campbellbarton :: r46550
dont use sdl include when building headless

Commit by campbellbarton :: r46547
edits to recent sequencer api additions not to do scene lookups, the scene is already known.

Commit by campbellbarton :: r46546
no need to clamp python values twice when assigning.

Commit by campbellbarton :: r46545
sequencer bug, was possible to add strips to negative channels, the channel clamping was ignored when calling
RNA_property_int_set() when setting operator int/floats.

Commit by campbellbarton :: r46544
- dont create scene.sequence_editor on read, means data-browser or autocomplete will allocate a sequencer.
... instead add scene.sequencer_editor_create / clear, these match id.animation_data_* functions.
- refactor for names, for scene level functions call them BKE_sequencer_*

Commit by lukastoenne :: r46542
A couple more changes to the file and image nodes to improve access to layers that don't follow Blender's rlayer.rpass
naming scheme.
--- Changes to File Output node ---
* Flat layer names in EXR multilayer files.
 For a socket with name "AAA" the previous resulting EXR layer name would be "AAA.AAA", i.e. the render layer as
 well as render pass would use the socket name.
 Now the "render_layer.render_pass" scheme is ignored in multilayer files, socket names are directly written to EXR
 layers (EXR layer name is "AAA" in this example). If sockets should have a notion of "render layer" this can still
 be achieved by explicitly adding a separator, e.g. "AAA.BBB". When loading such layers into a Blender Image struct,
 the name is interpreted as a "render_layer.render_pass" again (although the image node does not care about it,
 see below).
* Socket sub-paths (for singlelayer) or layer names (for multilayer) are stored in dedicated string variables
in the socket storage data. This way the RNA can define precise string subtypes (PROP_FILEPATH) and length. The
file/layer slots are defined as separate structs with own name properties in the RNA as well, so they can be used
nicely with the list template.
* Ensure unique socket paths/layer names to prevent overwriting of files and layers respectively.
--- Changes to Image node ---
* Loading multilayer OpenEXR files has improved layer name splitting into render layer + render pass names now. This
properly supports arbitrary EXR layer names now.
  Example:
 OpenEXR layer name: AAA.BBB.CCC
 is split into
 Render layer name: AAA.BBB
 Render pass name: CCC
 If the layer name has no '.' separators the render layer name is empty.
* Image node ignores the selected render layer in the image user data. Instead all existing layers are displayed
at the same time by combining the render layer names with render pass names again, to reconstruct the original
EXR layer name. This avoids the problem that render layers with empty name are not selectetable in the dropdown
and allows using all image layers at the same time without duplicating the node.

Commit by campbellbarton :: r46541
style cleanup: mainly sculpt/whitespace

Commit by nazgul :: r46537
Compilation error fix: abort() wasn't declared in BKE_ccg.h before

Commit by gsrb3d :: r46535
SVN maintenance.

Commit by blendix :: r46534
Cycles: reviewed the task scheduler code and fixed (hopefully all) windows threading problems.

Commit by nicholasbishop :: r46532
Add mask-drawing support to GPU_Buffers.
This is the last commit of the sculpt masking merge. Documentation:
http://wiki.blender.org/index.php/User:Nicholasbishop/PaintMasks
Thanks to Brecht for reviewing!
* For VBO, add color to the VertexBufferFormat structure as three
  unsigned bytes. Since mask elements are scalar the three color
  components are identical to eachother, but the fixed-function OpenGL
  pipeline requires colors to be either three or four components.
* For the same reason, multires VBO drawing now copies into the
  VertexBufferFormat format as well.
* Regression: material colors will not show up correctly now, masks
  colors are overriding. Not sure how to fix this nicely (would be
  much easier to fix if drawing with vertex shaders.)
* Also, masks will only draw PBVH drawing, so only 'solid' drawing
  will work correctly with masks.

Commit by nicholasbishop :: r46531
Use VertexBufferFormat for multires VBO.

Commit by nicholasbishop :: r46530
Add keymap and menu entries for masking.
* Add CTRL+IKEY to invert the mask.
* Add ALT+MKEY to clear the mask.
* Change the 'Hide' menu in sculpt mode to 'Hide/Mask', adds entires
  for clearing, filling, and inverting the mask, as well as hiding
  masked regions.

Commit by nicholasbishop :: r46529
Add support for hiding masked regions.
Add a new mode, PARTIALVIS_MASKED, to the PAINT_OT_hide_show operator.

Commit by nicholasbishop :: r46528
Add a paint mask operator to clear, fill, or invert the mask.

Commit by nicholasbishop :: r46527
Update the keymap for the mask brush.
* Add MKEY as a toggle for the mask brush. We could use ALT similar to
  SHIFT toggling the smooth brush, but it would conflict with MMB
  emulation (not to mention many window managers.)
* When the mask brush is active, SHIFT toggles it into smooth mode.

Commit by nicholasbishop :: r46526
Add mask brush for sculpt mode.
The mask brush currently has two modes, 'draw' and 'smooth'.

Commit by nicholasbishop :: r46525
Add new mask-brush icon from Julio Iglesias.

Commit by nicholasbishop :: r46524
Use paint mask when calculating sculpt strength.

Commit by nicholasbishop :: r46523
Add undo/redo support for paint masks.

Commit by nicholasbishop :: r46522
Ensure mask layers are always present in sculpt mode.

Commit by nicholasbishop :: r46521
Copy GridPaintMask to vertex paint mask when applying multires.
Adds new subsurf_copy_grid_paint_mask() function similar to
subsurf_copy_grid_hidden().

Commit by nicholasbishop :: r46520
Add mask support to CCGSubSurf and multires.
* Add new CCG function ccgSubSurf_setAllocMask(). Similar to to
  ccgSubSurf_setCalcVertexNormals(), it sets whether the CCG elements
  have a mask layer and what that layer's offset is. Unlike normals
  however, it doesn't change any behavior during CCG calculation; it's
  there only to give CCGKey information on the mask.
* Add a new flag to _getSubSurf(), CCG_ALLOC_MASK. If set, space for
  an extra layer is allocated, but the number of CCG layers is not set
  to include it. This is done because GridPaintMasks are absolute,
  rather than being relative to the subdivided output (as MDisp
  displacements are), so we skip subdividing paint masks here.
* Add a new flag to subsurf_make_derived_from_derived(),
  SUBSURF_ALLOC_PAINT_MASK. This controls whether CCG_ALLOC_MASK is
  set for _getSubSurf(). Related, masks are never loaded in during
  ss_sync_from_derivedmesh(). After subdivision is finished, if the
  alloc mask flag is set, the number of CCG layers is increase to 4
  with ccgSubSurf_setNumLayers().
* Add a new flag to multires_make_from_derived(),
  MULTIRES_ALLOC_PAINT_MASK. Not all multires functions need paint
  mask data (e.g. multiresModifier_base_apply.) This flag is always
  set in MOD_multires.c so that subdividing a mesh with a mask updates
  properly even when not in sculpt mode.
* Update multiresModifier_disp_run() to apply, calculate, and add mask
  elements. It's almost the same as the existing operations with xyz
  coordinates, but treats masks as absolute rather than displacements
  relative to subdivided values.
* Update multires_customdata_delete to free CD_GRID_PAINT_MASK in
  addition to CD_MDISPS.
* Update multires_del_higher() to call the new function
  multires_grid_paint_mask_downsample(), which allocates a
  lower-resolution paint mask grid and copies values over from the
  high-resolution grid.

Commit by nicholasbishop :: r46519
Add GridPaintMask accessor to paint.c.

Commit by nicholasbishop :: r46518
Add paint mask access to the PBVH vertex iterator.

Commit by nicholasbishop :: r46517
Add access to mesh vertex customdata to the PBVH.

Commit by nicholasbishop :: r46516
Add DNA and customdata entries for paint masks.
CD_PAINT_MASK is a layer of per-vertex floats for non-multires
meshes. Multires meshes use CD_GRID_PAINT_MASK, which is a layer of
per-loop GridPaintMask structures. GridPaintMask is similar to MDisp,
but contains an array of scalar floats.
Note: the GridPaintMask could be folded into MDisp, but this way
should be easier to add mask layers in the future (if we do fold
GridPaintMask into MDisp, the mask array should probably be an array
of arrays with a 'totmask' field so that mask layers can be easily
supported.)
Includes blenload read/write support for CD_PAINT_MASK and
CD_GRID_PAINT_MASK.

Commit by nicholasbishop :: r46515
Replace hardcoded DMGridData structure with CCGElem/CCGKey.
* Changes to DerivedMesh interface: DMGridData has been removed,
  getGridData() now returns an array of CCGElem pointers. Also added
  getGridKey() to initialize a CCGKey (implemented only by
  CCGDerivedMesh.)
* PBVH: added BLI_pbvh_get_grid_key().
* A lot of code is affected, but mainly is just replacing
  DMGridData.co, DMGridData.no, and sizeof(DMGridData) with the
  CCG_*_elem functions, removing the reliance on grid elements of
  exactly six floats.

Commit by nicholasbishop :: r46514
Add CCGKey/CCGElem for accessing CCGSubSurf elements.
CCGKey caches information about the CCGSubSurf element layout. This
data, along with the CCG_* inline functions, allows access to
CCGSubSurf elements with an arbitrary number of layers (as opposed to
the hardcoded DMGridData structure which assumes xyz coordinates
followed by three normal components.)
The CCGElem structure is declared but not defined anywhere, just used
as a convenient type.

Commit by nicholasbishop :: r46513
Modify CCGSubsurf to subdivide an arbitrary number of (float) layers.
The layout of vert data in CCGSubSurf is almost the same; previously
it was three floats (for xyz coordinate) optionally followed by three
floats for the normal. The only change is that the first three floats
can now be any number of floats.
* _getSubSurf takes a numLayers parameter to set the number of layers,
  stored in CCGMeshIFC.numLayers.
* All calls to _getSubSurf currently have numLayers set to 3, except
  for UV subsurf, where it is reduced to 2 (with a corresponding
  change when reading the results out to use float (*)[2] rather than
  float (*)[3].)
* The various VertData* macros in CCGSubSurf.c are now functions that
  take a CCGSubSurf pointer, which provides access to CCGMeshIFC,
  which has numLayers.
* Add ccgSubSurf_setNumLayers() to the API. Only changes the number of
  layers that get subdivided, doesn't change the amount of memory
  allocated. So if space for N layers is allocated, it's safe to set
  the number of layers to less than N, but not more.
* The rest of the changes are just adding the 'ss' parameter.

Commit by nicholasbishop :: r46512
Add new options to PAINT_OT_brush_select, toggle and create_missing.
The toggle option, if enabled, will toggle back and forth between two
brushes. (The first brush of the desired tool type will be toggled to,
running the toggle again switches back to the previously selected
brush.)
If no brush of the desired type is found, and the create_missing
option is enabled, a new brush of that type will be created and set.

Commit by nicholasbishop :: r46511
Add an RNA access function to get an enum item name from its value.
New function is RNA_enum_name_from_value.

Commit by nicholasbishop :: r46510
Code cleanup for multires_dm_create_from_derived().
Changed name to multires_make_derived_from_derived() and parameter
order to be more similar to subsurf_make_derived_from_derived().
Added MultiresFlags enum with flag values to replace the local_mmd and
useRenderParams parameters.

Commit by nicholasbishop :: r46509
Code cleanup for parameters of subsurf_make_derived_from_derived.
This is the first commit of the sculpt masking merge. Documentation:
http://wiki.blender.org/index.php/User:Nicholasbishop/PaintMasks
Thanks to Brecht for reviewing!
Replaced four boolean parameters with a single flag and a new enum,
SubsurfFlags.

Commit by blendix :: r46506
Fix #31379: cycles not recognizing 16 bit tiff as float image.

Commit by nazgul :: r46505
Fix #31221: Multiple Monitors/Windows Double Click For Focus
Window might be focused by mouse click in configuration of window manager
when focus is not following mouse.
Click could have been done on a button and depending on window manager settings
click would be passed to blender or not, but in any case button under cursor
should be activated. Currently the easiest way to activate button is to
generate mouse move event.
Not sure about OSX, but on windows it worked fine because OS is generating
WM_MOUSEMOVE event "automatically" so nothing special from blender side was
needed.

Commit by campbellbarton :: r46504
minor changes to sequencer api
- rename seqelements.push(file) --> filename
- rename SoundSequence.waveform --> show_waveform
- dont pass context to sequencer functions. editing any scenes sequence can be done, irrespective of context.

Commit by nazgul :: r46503
Fix #31393: Shape Keys on curves with hooks
Own mistake when was fixing #30561 - handles ARE included in shape keys

Commit by campbellbarton :: r46502
style cleanup: sequencer and scene rna

Commit by campbellbarton :: r46501
patch [#30871] VSE py-api
from Dan Eicher (dna)
--- message from the tracker
Classes for all effect types with proper input attributes
Added new/delete functions for SequenceEditor.sequences.
push/pop functions for ImageSequence.elements to add/remove images
Moved waveform from the base class to SoundSequence (probably should be renamed use_waveform or show_waveform)
Fixed user count for scene and movie clip types
--- my own comments
- dont have blending mode argument from sequencer.new_*() functions. Better edit this after.
- dont change waveform attribute, seems unrelated change and should be kept for sound afaik.
- dont apply scene, clip usercount changes - Sergey dealt with these separately.

Commit by psy-fi :: r46496
Enable ffmpeg building with MinGW-w64. The libraries were compiled with gcc 4.7, however is has been claimed there
are issues with gcc > 4.4 so this is for testing purposes. Still it doesn't crash on startup anymore. Forgot to
mention on ffmpeg lib commit that committed ffmpeg version is 0.10.2

Commit by nazgul :: r46494
Style cleanup: use 2 spaces for indentation in bundle scripts

Commit by nazgul :: r46492
Carve and libmv bundling scripts: should work with svn checkout now

Commit by nazgul :: r46490
Bundle updated version of carve. Should be no functional changes, small code cleanup

Commit by nazgul :: r46489
Synchronize libmv with changes in git branch

Commit by campbellbarton :: r46486
fix [#31382] Loop Cut and Slide numpad (-)
investigation lead to finding 3 bugs here...
- transform key input handling didnt ignore minus key on an unsigned value as it should.
- not being able to set numcuts to 0 made typing in numbers not very useful.
- backspace would set the cuts to an unsigned value.

Commit by campbellbarton :: r46485
fix for fix, own fix for [#31368] didnt work right, now move shadeless & wire paint drawing into draw_mesh_paint().

Commit by nicholasbishop :: r46484
Potential fix for bug #31111, Remesh modifier generates artifacts on simple surfaces
Changed one of the intersection tests to use < rather than  pointer.

Commit by campbellbarton :: r46363
minor correction for last commit, use malloc rather then calloc (array is filled in later)

Commit by campbellbarton :: r46362
fix for limited dissolve (after sine intended fixes - not cleaning up before vertex dissolve would skip dissolving
some verts that should be dissolved).
now do this:
- edge dissolve
- cleanup (removing edges left over from dissolving faces)
  cleanup removes verts and NULL vertex input array
- dissolve verts which haven't been removed.

Commit by campbellbarton :: r46361
fix [#31197] Limited dissolve leaves faces/edges/verts behind
bad bug where vertices could be in a face more then once (which isn't allowed), now check for this when creating
a face.

Commit by campbellbarton :: r46360
style cleanup: blenkernel

Commit by mont29 :: r46358
Mesh elements sorting refactor.
Now only one operator. Same options for vertices, edges and faces (so adds edges sorting, and some options to
vertices sorting).
Face sorting should behave as previously. However, XSortVerts won’t pack anymore selected vertices at the begining of
the vert array (as it used to), if you want such behavior you’ll have to first run SortElements with Selected action.
Also added bug ref I forgot in r46354 (armature.c).

Commit by campbellbarton :: r46357
code cleanup: documentation commit

Commit by campbellbarton :: r46356
code cleanup: add getCursorGrabModeIsWarp(), makes the intent more clear where we check for mouse warping.

Commit by campbellbarton :: r46355
disable warping for tablet events (X11 only), added TODO's for OSX and win32.

Commit by mont29 :: r46354
Fix for [#31333] 2.63 Bone copy rotation becomes mad after entering/leaving armature edit mode
There is no real good solution to this problem, hopefully this threshold value will be a good compromize this
time... :(

Commit by blendix :: r46353
Code cleanup: patch by Thomas to fix PassType values to be consecutive.

Commit by campbellbarton :: r46352
style cleanup: BKE_*.c files which deal with library functions

Commit by mont29 :: r46351
Last part of fix for [#31157]: Some (actually, 172) operators have no tooltip.
Only remaining undocumented one is IMAGE_OT_record_composite (not sure what it actually does, nor even whether
it’s actually working or not...).
Note that I didn’t bother with operators flagged as OPTYPE_INTERNAL!

Commit by campbellbarton :: r46349
update for recent RNA change

Commit by campbellbarton :: r46348
style cleanup: modifiers

Commit by campbellbarton :: r46347
fix for same bug as r46346, incorrect code was copied.

Commit by campbellbarton :: r46346
fix for error in MOD_weightvgmix.c caused by confusion without braces.
CustomData_add_layer_named would run even when the dvert layer was found.

Commit by aligorith :: r46345
Bugfix [#31003] Animation data (e.g. actions, drivers) from the active material
of a material nodetree was not visible in the Animation Editors

Commit by campbellbarton :: r46343
add back pose specials menu, also add back the ability to clear user transforms on all bones (not just selected -
mango request - was possible in 2.4x).
- rename only_select op property to only_selected (both were used).
- only do mingw workaround when building with FREE_WINDOWS defined.

Commit by aligorith :: r46342
Bugfix [#31330] Shape Key Editor does not switch to assigned Key Action

Commit by mont29 :: r46341
Fix for [#31327] Text editor menu incorrect (Patch attached)
Patch by Justin Dailey, thx!

Commit by jwilkins :: r46340
Cleaned up glreport.py (pardon my dust)
See the README.txt for details of how to read the generated report.
Consolidated the 600+ extension files into a set of JSON style Python files.
Not much of a space reduction, but definitely a reduction in clutter.
People are free to edit the .gl files.  The warning in the files against editing is for people who are generating
new .gl files from other sources.

Commit by aligorith :: r46339
Timeline Header: Only show the "use active keyingset only" toggle (beside the
autokey button) if autokeying is enabled. This was only useful when the latter
was actually enabled.

Commit by aligorith :: r46338
Bugfix [#31297] Transform menu often showed operators which didn't make sense
for the contexts they appeared in
The VIEW3D_MT_transform menu was being used everywhere, although not all the
operators there applied to the current mode (e.g. the pivot point, animated
transforms to deltas, and a few others were Object-mode only). Split off the
offending parts into separate menus for the specific modes, and made these
extend the basic menu type.

Commit by jwilkins :: r46337
typo, agle -> angle

Commit by aligorith :: r46336
Bugfix:
While testing the other my previous commit (removing deformflag stuff), I
noticed that trying to change set a parent object for the current object (via
the Object Properties) would cause the current object to "blow up", and for
cyclic dependency warnings to be spewed to the console.
Adding a dummy usage of one of the vars here seems to solve it. Perhaps a case
of weirdo compiler optimisations?

Commit by aligorith :: r46335
Removing the old armature 'deform' settings (use vertex groups/envelopes/quats).
As far as I could tell, these were really only still used for "virtual
modifiers", though we really don't use these anymore. Instead, most of the time,
people need to set these settings in armature modifiers directly (these didn't
even get copied over in that case). This was a source of confusion and
redundancy, so removing these now.
This change can be reverted if these were actually of some use out there...

Commit by aligorith :: r46334
Bugfixes for Motion Path drawing/updating in light of the recent changes:
* Added proper "update" operators in place of the abuse of the calculate
operators, so now the display ranges won't get overwritten everytime (with the
default values) you go to update the paths.
* Display range settings in properties editor now actually work. Before, the "In
Range" mode only displayed the entire paths.

Commit by aligorith :: r46333
Simplifying constraints code - removed relink_data() callbacks in favour of a
more generic system which goes through id_looper(), reducing the maintainance
burden for new constraint authors

Commit by aligorith :: r46332
Deleting action constraints (and a few others) now adjusts the usercounts of
their referenced data correctly

Commit by campbellbarton :: r46331
code cleanup: naming - BKE_mesh_*

Commit by blendix :: r46329
Cycles: minor refactoring of fisheye code to fit code style.

Commit by blendix :: r46328
Cycles: threading optimizations
* Multithreaded image loading, each thread can load a separate image.
* Better multithreading for multiple instanced meshes, different threads can now
  build BVH's for different meshes, rather than all cooperating on the same mesh.
  Especially noticeable for dynamic BVH building for the viewport, gave about
  2x faster build on 8 core in fairly complex scene with many objects.
* The main thread waiting for worker threads can now also work itself, so
  (num_cores + 1) threads will be working, this supposedly gives better
  performance on some operating systems, but did not measure performance for
  this very detailed yet.

Commit by blendix :: r46327
Fix #31318: cycles preview missing update for other materials using same node
group with certain operations.

Commit by mont29 :: r46325
Fix related to [#31157]: Tips (descriptions) of macro operators were not set into underlying RNA struct, hence
did not show up in UI.

Commit by mont29 :: r46322
Fix [#31322] Assign to Active Group adds new group

Commit by nazgul :: r46321
- Fixed issue with printing file version triple times when loading file in debug mode
- Remove commented out headers from versioning module

Commit by mont29 :: r46320
Second part of fix for [#31157]: Some (actually, 172) operators have no tooltip.
About 30 undocumented ops remaining...

Commit by campbellbarton :: r46319
code cleanup: BKE_libblock_find_name() now takes an ID constant rather then a string.

Commit by campbellbarton :: r46318
feature request from VenomGfx- lock to active as an operator since its tedious setting the object and bone manually
(especially if you have it right in front of you)
uses keys
- Shift+PadPeriod --- to set
- Alt+PadPeriod --- to clear (also clears cursor and camera locking)

Commit by campbellbarton :: r46317
code cleanup: naming - pose/armature/image
also use ..._find_name(..., name) rather then ..._find_named(..., name) --- both were used.

Commit by aligorith :: r46316
Bugfixes for various ID-block references (Constraints, NLA)
* ID-blocks referenced by Constraints but not being used as the target objects
(such as Actions in the Action Constraint, or Text Blocks in PyConstraints) now
get usercounts for being referenced in this way. This should fix ancient bugs
such as [#19205] and [#8593]. More tests still needed to verify that this
does now play nicely with proxies.
* Changing actions used by NLA strips should now update the usercounts
accordingly

Commit by campbellbarton :: r46315
code cleanup: BKE_ naming, also make bpy.data.images.load() always load a new image. (not use existing one)

Commit by campbellbarton :: r46314
code cleanup: BKE_scene api naming.
also stop numpy from being found in /usr/include with cmake.

Commit by campbellbarton :: r46313
code cleanup: function naming, use BKE_*type* prefix.

Commit by blendix :: r46311
Fix related to #31317: viewport render layer rendering now uses render visibility
rather than viewport visibility, is more useful that way.

Commit by blendix :: r46310
Fix #31302: UV unwrap no longer remember last conformal/angle based unwrapper
choice with bmesh.

Commit by aligorith :: r46309
Style Cleanup: Wrapping with parens for safety and whitespace edits

Commit by aligorith :: r46308
Mango Bugfix - Actions being tweaked in NLA Editor were not being played back if
the track that they belonged to was currently being played back "solo"

Commit by campbellbarton :: r46307
code cleanup: brush/camera naming

Commit by campbellbarton :: r46306
code cleanup: function naming for BLI functions.

Commit by nexyon :: r46305
Patch [#31240] Fix audaspace deadlock
fix for [#31097] glibc error when playing sound using BGE
by Wander Lairson Costa
Note: This deadlock fix makes the code non-threadsafe again, a proper solution has to be found still.

Commit by nexyon :: r46304
Fixes for
* [#31285] VSE: audio pitch change delays audio
* [#31260] VSE Trimmed audio plays when overlaped

Commit by nazgul :: r46297
Clip editor: fixed crash when opening dopesheet view without having clip opened

Commit by campbellbarton :: r46294
fix [#31136] Save All Edited only works for Saved external image, not New or Packed image (bpy.ops.image.save_dirty)

Commit by dingto :: r46293
Windows:
* Fix compile for recent do_versions() splitting.

Commit by nazgul :: r46291
Added start and end bevel factor for curves, so now it's possible to make
a bevelled curve which isn't fully covered with a bevel.

Commit by campbellbarton :: r46290
fix [#31205] Loop cut don't slide when hidden faces between

Commit by blendix :: r46289
Fix (harmless) uninitialized memory usage in BVH binning.
Fix unneeded warnings with c++ guardedalloc, delete NULL is valid.

Commit by dfelinto :: r46288
small fix for equisolid fisheye (cycles)
the FOV formular is: R = 2 * lens * sin (theta / 2)
in this case theta is fov/2 already, thus the fix

Commit by dfelinto :: r46287
Fisheye Camera for Cycles
For sample images see:
http://www.dalaifelinto.com/?p=399 (equisolid)
http://www.dalaifelinto.com/?p=389 (equidistant)
The 'use_panorama' option is now part of a new Camera type: 'Panorama'.
Created two other panorama cameras:
- Equisolid: most of lens in the market simulate this lens - e.g. Nikon, Canon, ...)
             this works as a real lens up to an extent. The final result takes the
             sensor dimensions into account also.
             .:. to simulate a Nikon DX2S with a 10.5mm lens do:
                 sensor: 23.7 x 15.7
                 fisheye lens: 10.5
                 fisheye fov: 180
                 render dimensions: 4288 x 2848
- Equidistant: this is not a real lens model. Although the old equidistant lens simulate
               this lens. The result is always as a circular fisheye that takes the whole sensor
               (in other words, it doesn't take the sensor into consideration).
               This is perfect for fulldomes ;)
               For the UI we have 10 to 360 as soft values and 10 to 3600 as hard values (because we can).
Reference material:
http://www.hdrlabs.com/tutorials/downloads_files/HDRI%20for%20CGI.pdf
http://www.bobatkins.com/photography/technical/field_of_view.html
Note, this is not a real simulation of the light path through the lens.
The ideal solution would be this:
https://graphics.stanford.edu/wikis/cs348b-11/Assignment3
http://www.graphics.stanford.edu/papers/camera/
Thanks Brecht for the fix, suggestions and code review.
Kudos for the dome community for keeping me stimulated on the topic since 2009 ;)
Patch partly implemented during lab time at VisGraf, IMPA - Rio de Janeiro.

Commit by campbellbarton :: r46286
make text move up/down into a single operator with a direction property

Commit by nazgul :: r46285
Split do_versions into separate files for pre-2.50 versions and 2.5x versions
This should make it easier to navigate through readfile.c and also hopefully
will prevent corereview to fail parsing this file.

Commit by campbellbarton :: r46282
fix [#31235] Limited Dissolve problems
this is in fact 2 bugs.
- unselected edges between 2 faces that were joined didnt get removed.
- in face mode, edges and verts at the boundary of the selection would get incorrectly dissolved.
also quiet float/double promotion warning.

Commit by mont29 :: r46281
First part of fix for [#31157]: Some (actually, 172) operators have no tooltip.
Notes:
* This commit adds about a third of missing tips (a few are rather dummy, as name already says everything, but
better that than "(undocumented operator)" showing in UI!
* There is a problem with macros, their tips are not registered in RNA. Got a patch for this, will submit it to
campbo asap.

Commit by aligorith :: r46280
Patch [#31279] clarifiy a python error-string (when incorrectly specifying enum
items from python)
Thanks Philipp Oeser (lichtwerk)

Commit by aligorith :: r46279
Patch [#30654] Wiki Quick Hack: Text editor move lines up/down
Submitted by: Justin Dailey (dail)
Patch allows the current line (or selected lines) to be moved up and down with
Ctrl+Shift+Up and Ctrl+Shift+Down. Has undo/redo support and operators in python
menu.

Commit by campbellbarton :: r46278
own mistake in recent commit CustomData_layertype_is_singleton()
was stopping bmesh python api adding multiple layers.

Commit by campbellbarton :: r46277
small improvement to last commit and quiet double promotion warnings.

Commit by campbellbarton :: r46276
code cleanup: use much simpler (and likely faster) polygon normal calculation in object mode.

Commit by lukastoenne :: r46275
Removed the automatic socket selection in the modal node linking operator.
This was originally reimplemented as part of the socket selection feature, but since selecting a socket necessarily
also selected the owning node, it messes with the manual user selection of nodes too much (and doesn't add any
additional usability).

Commit by blendix :: r46273
Fix: forgot to commit these files as part of strand bugfix.

Commit by jwilkins :: r46270
Renamed "fake" OpenGL identifiers.
Any identifier that looks like an OpenGL identifier, but isn't, causes a false alarm by the glreport.py tool.
Most of these were in comments so I just rephrased the comments.  There were a couple of static functions/macros
that were easy enough to rename.  Only the glTexco and glIndex fields of the DMVertexAttribs struct was public
and had non-local uses.

Commit by nazgul :: r46269
Fix #31288: Blender crash when select a material
Quite the same issue as render crash on missed shader groups, needed a NULL-check
in node exec function as well.

Commit by blendix :: r46265
Fix: issue in strand render + instancing bugfix, also optimized it a bit to avoid
unnecessary memory allocations.

Commit by blendix :: r46264
Fix #31286: saving 16 bit BW tiff could crash due to invalid memory access. Also
found that 16 bit RGBA saving was not working, fixed as well.

Commit by jwilkins :: r46263
GL Tool Improvements
The glreport.py tool now recognizes OpenGL 1.0 and OpenGL 1.1 enumerations.
Renamed some files so they have more consistent names.

Commit by nazgul :: r46262
Throw an error and prevent compilation if there're unknown structures detected by DNA

Commit by blendix :: r46261
Cycles: tweak ATI OpenGL/CUDA fix more with extra error check.

Commit by blendix :: r46260
Fix related to #31118: ensure tesselated faces are available when exporting
to fluid sim. Is not actually the cause of the bug.

Commit by jwilkins :: r46259
There is no GLenum called LIGHT_POSITION
There is an FAQ that mentions a mythical GL_LIGHT_POSITION, and lots of programmers speak of it, but this mythical
creature does not exist!
The correct symbol is GL_POSITION

Commit by campbellbarton :: r46258
style checker tool - white space and line length checks (only check line lengths ending with ;)

Commit by nicholasbishop :: r46257
Small hull bmop fix, distance check from plane needs absolute value.

Commit by jwilkins :: r46256
GL Tools
glreport.py now lists several more categories
        Toolkits like GLUT and GLU
        System librarys GLX, AGL, and WGL
        ES 1.1 (Common and Common-Light)
        ES 2.0
        Deprecated 1.x functions
Several other bug fixes.
OpenGL 1.0 and 1.1 enumerations have not been added to VERSION_1_0 and VERSION_1_1 files yet.
Set eol-style to native

Commit by gsrb3d :: r46253
SVN maintenance.

Commit by nazgul :: r46252
Style cleanup: spaces around operators

Commit by nazgul :: r46251
Camera tracking: show disabled active track even if Show Disabled is switched off

Commit by blendix :: r46250
Possible fix for #31054: cycles viewport rendering not working with CUDA for
computation and ATI card for OpenGL.

Commit by nazgul :: r46249
Camera tracking: switch dopesheet information to lazy calculation
All operators which changes tracking data now just tags dopsheet as outdated,
actual re-calculaiton of happens only when this information is actually needed
(like on dopesheet draw).
This makes things a bit faster when there's no dopesheet visible in current
screen and also makes it much easier to update dopesheet using dependency
graph.
Also renamed dopesheet_sort_order to dopesheet_sort_method in rna and internal
stuff which makes much more sense and also correlated with naming in
file browser.

Commit by campbellbarton :: r46247
code cleanup: double promotion & some style cleanup

Commit by blendix :: r46246
Fix #31258: collada now selects newly added objects after import.

Commit by jwilkins :: r46245
svn eol properties and a shebang for glreport.py

Commit by campbellbarton :: r46244
code cleanup:
- replace iterators with macros
- move vertexCos checks outside the for loops (use 2 for loops).
- style cleanup

Commit by campbellbarton :: r46243
fix for last commit, (warning was without openmp only)

Commit by jwilkins :: r46242
Set the executable property on glreport.py

Commit by blendix :: r46241
Fix strand render + instancing render bug, gave tile artifacts.

Commit by campbellbarton :: r46240
style cleanup: formatting and some float/double promotion

Commit by campbellbarton :: r46239
code cleanup: minor changes to get trunk compiling with strict warnings.

Commit by nazgul :: r46238
Clip editor: sort order for dopesheet channels
Supported sorting by name, longest tracked segment and total tracked frames.
Internally tracks are stored in Tracking datablock, but sort order is
a clip editor space property and sorting happens on clip editor draw.
If there's no dopesheet opened with different sort orders it's not
a problem due to re-sorting wouldn't happen.
Also fixed draw issue of tracked segments introduced in previous commit.

Commit by nazgul :: r46237
Camera tracking: pre-calculate tracked segments for dopesheet channels

Commit by nazgul :: r46236
Camera tracking: clear clipboard path before copying tracks to it.

Commit by nazgul :: r46235
Camera tracking: apparently one of tracking presets was never merged from tomato

Commit by blendix :: r46234
Fix #31234: vertex normals not update after knife cut.

Commit by blendix :: r46233
Fix #31275: mesh draw issues after adding hook to mesh in edit mode.

Commit by blendix :: r46232
Fix #31273: videotexture ImageRender not working in blenderplayer.

Commit by blendix :: r46231
Fix #31259: particle grid distribution not working.

Commit by nazgul :: r46230
Fix #31272: Blender Crashes when press render
Issue was caused by missed libraries. Seems compositor nodes already does check
in such cases, added the same check to texture and shader nodes.

Commit by nazgul :: r46229
Fix #31254: OpenGL-Rendering dark without Effect-Strip
Issue was caused by the fact, that sequencer is working in sRGB space, but
when there's only image input strips we need to make sure conversion from
byte to float buffer would keep float buffer in sRGB space and wouldn't
make it linear as it's supposed to be in other areas.

Commit by psy-fi :: r46228
Fix #31178, the cause is hidden and selected faces would return NULL as UVElements. To avoid checking the whole
contingency of UV synch selection + face selection + face hidden, added a NULL check.

Commit by nazgul :: r46227
Fix #31266: Track preview is shifted by half a pixel

Commit by blendix :: r46226
Fix #31139: fractal mesh subdivide was only working along normal where previously
it would displace in all directions. Now there's an operator option to control this.

Commit by blendix :: r46225
Fix #31257: tiff reader not reading 16 bit grayscale images correctly.

Commit by blendix :: r46224
Fix #31108: entering local view with glsl shows objects black. Now keeps lamps
affecting the material even if they are not part of the local view.

Commit by lukastoenne :: r46222
Enum property items were missing a terminator item.

Commit by mont29 :: r46221
Grr! Fix for own error in r46193.

Commit by jwilkins :: r46220
GL Tool - glreport.py
glreport.py is a script that generates a report detailing how Blender uses OpenGL versions and extensions.
Other OpenGL related tools are planned and the intent is for all of the tools to live in the same directory.

Commit by jwilkins :: r46219
Patch [#30965] Cancel Sculpt Stroke w/ ESCAPE
If the RMB has not been released after starting a sculpt stroke, then hitting escape will cancel the stroke in
progress and undo any changes to the mesh.
This is a slightly faster work-flow than using undo, is a feature available in other paint programs, and also puts
in place the infrastructure to add other keys later that could tweak strokes in different ways.

Commit by jwilkins :: r46218
Patch [#30255] Center Last Stroke
Enables the NUMPERIOD view centering operator to work in sculpt mode.  Hitting NUMPERIOD while in sculpt mode will
center the view on the end of the last sculpting stroke made by the user.
This is useful for quickly refocusing on the current work area without fussing with the view controls.
It does not zoom into the stroke (slide only).
It does nothing if there have been no strokes.

Commit by gsrb3d :: r46217
SVN maintenance.

Commit by jwilkins :: r46216
ED_object_exit_editmode needs ED_object.h

Commit by blendix :: r46215
Fix #31256: collada export while in edit mode does not export edit mode changes,
now uses same call as render to flush mesh without leaving edit mode.

Commit by blendix :: r46214
Fix #31250, #31248: wrong vertex normals after transform apply, collada import,
sculpt shape key switch. All cases that called this function needed parameter
only_face_normals set to false, so changed it now.
Also fixed wrong user count for imported mesh from collada and simplified
previous fix for tesselated faces to polygons conversion.

Commit by blendix :: r46213
Fix #31253: collada export default file name now is no longer always
untitled.dae, but rather uses .blend file name like other exporters.
Patch by Gaia Clary.

Commit by nicholasbishop :: r46212
Initialize an input in bmo_hull.

Commit by blendix :: r46206
Fix #31247: cycles crash after recent bugfix.

Commit by blendix :: r46205
Fix #31089: collada file crashing on importing file with unknown/unsupported animation data.

Commit by nazgul :: r46202
Style cleanup: spaces around operators

Commit by nazgul :: r46201
Clip editor: cleanup of View menu - do not show operators which doesn't make sense in dopesheet/curve view

Commit by blendix :: r46200
Cycles: add Ray Length output to Light Path node. This gives the distance travelled
by the last light ray. One use case for this might be to do absorption.
Patch #31232 by Agustin benavidez, see this blog post for details:
http://agus3d.blogspot.com.ar/2012/05/blender-cycles-ray-length-node-output.html

Commit by nazgul :: r46199
Clip editor: remove Z-key shortcut for curve view. It doesn't make many sense anymore.

Commit by nazgul :: r46198
Clip editor: added drag-n-drop support so files can be opened in clip editor
by drag-n-dropping them from an external application.

Commit by kanttori :: r46197
Fix #31176: Collada model shows weird textures in editmode
Issue with multimaterial meshes and mface to mpoly conversion being before assignment of materials (which is done
on meshobject instancing?).
Added explicit bmesh conversion to MeshImporter which is called from DocumentImporter::import.

Commit by blendix :: r46196
Fix #31190: mirror modifier caused non-planar quads to be split differently on
the other side, now keep the first vertex of the polygon the same to avoid this.

Commit by blendix :: r46195
Fix wrong unified weight paint value version patch, was doing incorrect version check.
Fix #31209: weight paint sample & fill not using correct brush/unified value.

Commit by blendix :: r46194
Fix #31236: linking a socket of a different type to an existing node group output
would crash, did not convert default value storage correctly.

Commit by mont29 :: r46193
Small enhancement to Fast Gaussian compo blur node: do not compute when size is below 0.001!
(Was already checked/done for other blur algos, can be annoying when you animate the blur size to apply it only
on a few frames.)

Commit by blendix :: r46192
Fix: --debug-python command option was not working correct.

Commit by blendix :: r46191
Fixes for recent cycles and python/context commits.

Commit by mont29 :: r46190
CMake: Fix to install numpy (source dir was "hardcoded", rather use PYTHON_NUMPY_PATH), was failing under Debian
testing.
Note that there is still a problem, destination ("site-packages") is not in blender's python path, so you have to
edit sys.path before being able to import numpy... but at least it installs again.

Commit by blendix :: r46188
Python/context: tweak code further so that when you set screen/area/region
from python, it actually gets data context from there as well.

Commit by blendix :: r46187
Python: documentation about overriding context members.

Commit by blendix :: r46186
Fix crash setting area.type from a context that not include the right window,
ideally this function should become context free, for now just trick it to
execute in the right context.

Commit by blendix :: r46185
Python/context: allow overriding window/screen/area/region context for running operators
from python, this is useful to run an operator in a particular place in the UI.

Commit by campbellbarton :: r46184
bmesh: use fixed stack size for face flipping array.
also quiet unused var warnings

Commit by blendix :: r46183
Fix #31199 & #31112: cycles not working well with vertex/weight paint selection
mask drawing.
Now refactored the code a bit so that in no longer calls textured mesh drawing
for the face mask drawing, just handle it as part of regular paint color drawing.
Should also make the blender internal behavior more logical where it would start
showing textures in solid mode when enabling face masking.

Commit by blendix :: r46181
Fix 31220: glsl sun lamp shadows did not apparently respect shadow flag when
working in blender render rather than blender game mode. Tweaked the flags a
bit to make it compatible more, but for full configuration you need to be in
blender game mode still.

Commit by blendix :: r46180
Fix #31102: changing scene.use_color_man