6 years agoInterface:
Thomas Dinges [Mon, 4 Nov 2013 22:49:49 +0000 (22:49 +0000)]
* Move the "Add" menu from the Info header into the 3D View header.

Patch by Andrew Buttery (axb), with small tweaks by myself. (Patch ID #37241).

Approved by Brecht and Jonathan.

6 years ago* More fixes to make scons / Windows happy. :)
Thomas Dinges [Mon, 4 Nov 2013 22:33:02 +0000 (22:33 +0000)]
* More fixes to make scons / Windows happy. :)

6 years agoFix for recent BGE commits, when building with c++ guardedalloc.
Bastien Montagne [Mon, 4 Nov 2013 22:22:54 +0000 (22:22 +0000)]
Fix for recent BGE commits, when building with c++ guardedalloc.

6 years ago* Fix blenfont scons include after recent GE changes.
Thomas Dinges [Mon, 4 Nov 2013 22:14:04 +0000 (22:14 +0000)]
* Fix blenfont scons include after recent GE changes.

Mitchell: Please check if these includes also need to be added in scons:

6 years agoBGE Physics Cleanup: Fixing an error that GCC didn't catch.
Mitchell Stokes [Mon, 4 Nov 2013 19:23:00 +0000 (19:23 +0000)]
BGE Physics Cleanup: Fixing an error that GCC didn't catch.

6 years agoBGE: Fixing up a small issue from merging Rasterizer cleanup code.
Mitchell Stokes [Mon, 4 Nov 2013 19:22:56 +0000 (19:22 +0000)]
BGE: Fixing up a small issue from merging Rasterizer cleanup code.

6 years agoBGE: The recent physics cleanup was using KX_GameObject::GetParent() with out calling...
Mitchell Stokes [Mon, 4 Nov 2013 19:22:52 +0000 (19:22 +0000)]
BGE: The recent physics cleanup was using KX_GameObject::GetParent() with out calling parent->Release(). Since GetParent() does an AddRef(), this was causing a leak, which resulted in Zombie Object errors.

6 years agoBGE: Cleaning up the BGE's physics code and removing KX_IPhysicsController and KX_Bul...
Mitchell Stokes [Mon, 4 Nov 2013 19:22:47 +0000 (19:22 +0000)]
BGE: Cleaning up the BGE's physics code and removing KX_IPhysicsController and KX_BulletPhysicsController. Instead, we just use PHY_IPhysicsController, which removes a lot of duplicate code.

This is a squashed commit of the following:
    BGE Physics Cleanup: Fix crashes with LibLoading and replication. Also fixing some memory leaks.
    BGE Physics Cleanup: Removing KX_IPhysicsController and KX_BulletPhysicsController.
    BGE Physics Cleanup: Moving the replication code outside of KX_BlenderBulletController and switching KX_ConvertPhysicsObjects to create a CcdPhysicsController instead of a KX_BlenderBulletController.
    BGE Physics Cleanup: Getting rid of an unsued KX_BulletPhysicsController.h include in KX_Scene.cpp.
    BGE Physics Cleanup: Removing unused KX_IPhysicsController and KX_BulletPhysicsController includes.
    BGE Physics Cleanup: Removing m_pPhysicsController1 and GetPhysicsController1() from KX_GameObject.
    BGE Physics Cleanup: Remove SetRigidBody() from KX_IPhysicsController and remove GetName() from CcdPhysicsController.
    BGE Physics Cleanup: Moving Add/RemoveCompoundChild() from KX_IPhysicsController to PHY_IPhysicsController.
    BGE Physics Cleanup: Removing GetLocalInertia() from KX_IPhysicsController.
    BGE Physics Cleanup: Making BlenderBulletCharacterController derive from PHY_ICharacter and removing CharacterWrapper from CcdPhysicsEnvironment.cpp. Also removing the character functions from KX_IPhysicsController.
    BGE Physics Cleanup: Removing GetOrientation(), SetOrientation(), SetPosition(), SetScaling(), and GetRadius() from KX_IPhysicsController.
    BGE Physics Cleanup: Removing GetReactionForce() since all implementations returned (0, 0, 0). The Python interface for KX_GameObject still has reaction force code, but it still also returns (0, 0, 0). This can probably be removed as well, but removing it can break scripts, so I'll leave it for now.
    BGE Physics Cleanup: Removing Get/SetLinVelocityMin() and Get/SetLinVelocityMax() from KX_IPhysicsController.
    BGE Physics Cleanup: Removing SetMargin(), RelativeTranslate(), and RelativeRotate() from KX_IPhysicsController.
    BGE Physics Cleanup: Using constant references for function arguments in PHY_IPhysicsController where appropriate.
    BGE Physics Cleanup: Removing ApplyImpulse() from KX_IPhysicsController.
    BGE Physics Cleanup: Removing ResolveCombinedVelocities() from KX_IPhysicsController.
    BGE Physics Cleanup: Accidently removed a return when cleaning up KX_GameObject::PyGetVelocity().
    BGE Physics Cleanup: Remove GetLinearVelocity(), GetAngularVelocity() and GetVelocity() from KX_IPhysicsController. The corresponding PHY_IPhysicsController functions now also take Moto types instead of scalars to match the KX_IPhysicsController interface.
    BGE Physics Cleanup: Moving SuspendDynamics, RestoreDynamics, SetMass, GetMass, and SetTransform from KX_IPhysicsController to PHY_IPhysicsController.
    BGE Physics Cleanup: PHY_IPhysicsEnvironment and derived classes now use the same naming scheme as PHY_IController.
    BGE Physics Cleanup: PHY_IMotionState and derived classes now use the same naming convention as PHY_IController.
    BGE Phsyics Cleanup: Making PHY_IController and its derived classes follow a consistent naming scheme for member functions. They now all start with capital letters (e.g., setWorldOrientation becomes SetWorldOrientation).
    BGE Physics Cleanup: Getting rid of KX_GameObject::SuspendDynamics() and KX_GameObject::RestoreDynamics(). Instead, use the functions from the physics controller.
    BGE: Some first steps in trying to cleanup the KX_IPhysicsController mess. KX_GameObject now has a GetPhysicsController() and a GetPhysicsController1(). The former returns a PHY_IPhysicsController* while the latter returns a KX_IPhysicsController. The goal is to get everything using GetPhysicsController() instead of GetPhysicsController1().

6 years agoBGE Rasterizer Cleanup: The Blenderplayer now loads the monospace font so it can...
Mitchell Stokes [Mon, 4 Nov 2013 19:22:15 +0000 (19:22 +0000)]
BGE Rasterizer Cleanup: The Blenderplayer now loads the monospace font so it can properly draw the framerate and profile display.

6 years agoBGE Rasterizer Cleanup: Cleaning up some includes.
Mitchell Stokes [Mon, 4 Nov 2013 19:22:10 +0000 (19:22 +0000)]
BGE Rasterizer Cleanup: Cleaning up some includes.

6 years agoBGE Rasterizer Cleanup: Removing the Singletexture material mode. More conversion...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:50 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Removing the Singletexture material mode. More conversion code will probably be needed.

6 years agoBGE Rasterizer Cleanup: Moving the RAS_OpenGLRasterizer::ApplyLights() code into...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:37 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Moving the RAS_OpenGLRasterizer::ApplyLights() code into RAS_OpenGLRasterizer::ProcessLighting().

6 years agoBGE Rasterizer Cleanup: Removing the need to reference KX_BlenderMaterial or KX_Polyg...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:32 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Removing the need to reference KX_BlenderMaterial or KX_PolygonMaterial in RAS_OpenGLRasterizer.

6 years agoBGE Rasterizer Cleanup: Getting rid of the BL_Material.h include in RAS_OpenGLRasteri...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:25 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Getting rid of the BL_Material.h include in RAS_OpenGLRasterizer.cpp.

6 years agoBGE Rasterizer Cleanup: Getting rid of RAS_IRasterizer::RenderText(), and just adding...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:21 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Getting rid of RAS_IRasterizer::RenderText(), and just adding the code to IndexPrimitives_3DText(), which is the only function that uses RenderText().

6 years agoBGE Rasterizer Cleanup: Removing KX_BlenderGL since it was mostly one-line functions...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:16 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Removing KX_BlenderGL since it was mostly one-line functions used by KX_BlenderCanvas. KX_BlenderCanvas now just calls those functions directly.

6 years agoBGE Rasterizer Cleanup: Removing RAS_IRenderTools and moving the functionality to...
Mitchell Stokes [Mon, 4 Nov 2013 19:21:07 +0000 (19:21 +0000)]
BGE Rasterizer Cleanup: Removing RAS_IRenderTools and moving the functionality to RAS_IRasterizer. RAS_OpenGLRasterizer is a bit of a mess now with references to Ketsji and other modules it shouldn't be accessing.

6 years agoUI messages fixes...
Bastien Montagne [Mon, 4 Nov 2013 18:58:22 +0000 (18:58 +0000)]
UI messages fixes...

6 years agoMissing convert button on particle modifier panel. Cache pathcache check
Antony Riakiotakis [Mon, 4 Nov 2013 18:40:45 +0000 (18:40 +0000)]
Missing convert button on particle modifier panel. Cache pathcache check
should be only on path display mode.

6 years agoUpdate i18n tools for new build_hash...
Bastien Montagne [Mon, 4 Nov 2013 18:26:56 +0000 (18:26 +0000)]
Update i18n tools for new build_hash...

6 years agobullet: Remove patch applied upstream
Sergej Reich [Mon, 4 Nov 2013 17:02:01 +0000 (17:02 +0000)]
bullet: Remove patch applied upstream

6 years agoMade buildinfo aware of builds from GIT
Sergey Sharybin [Mon, 4 Nov 2013 13:21:39 +0000 (13:21 +0000)]
Made buildinfo aware of builds from GIT

- Use commit number since last annotated tag as a
  revision number replacement. It'll eb followed
  by 'M' symbol if there're local modification in
  the source tree.

- Commit short SHA1 is included. Helps getting
  information about commit used to build blender
  with much faster.

- If build is not done from master branch, this also
  will be noticed in the splash screen.

This commit also replaces revision stored in the
files with git-specific fields (change and hash).
This is kind of breaks compatibility, meaning
files which were saved before this change wouldn't
display any information about which revision they
were saved with. When we'll finally switch to git,
we'll see proper hash and change number since
previous release in the files, for until then
svn version will be used as a change number and
hash will be empty.

Not a huge deal, since this field was only used
by developers to help torubleshooting things and
isn't needed for blender itself.

Some additional tweaks are probably needed :)

6 years agoadd CDDM_lower_num_loops(), for completeness (currently unused).
Campbell Barton [Mon, 4 Nov 2013 12:01:46 +0000 (12:01 +0000)]
add CDDM_lower_num_loops(), for completeness (currently unused).

6 years agocode cleanup: typo in function name
Campbell Barton [Mon, 4 Nov 2013 11:27:11 +0000 (11:27 +0000)]
code cleanup: typo in function name

6 years agoBugfix [#36950] Restrict Frame Range not being obeyed on Driver F-Curves
Joshua Leung [Mon, 4 Nov 2013 04:18:28 +0000 (04:18 +0000)]
Bugfix [#36950] Restrict Frame Range not being obeyed on Driver F-Curves

By default, drivers (used to) automatically map the driver result (i.e.
"evaltime" or the x-coordinates in the driver graphs) to results. This evaltime
=> cvalue mapping is necessary when there are absolutely no keyframes or
modifiers on a driver F-Curve, or else nothing would happen.

However, when there are modifiers on these driver F-Curves, and these modifiers
only work within certain ranges, there would be confusing and unwanted
situations where even if you clamped the modifiers to only generating a curve
within certain frame ranges, the final driver output would still ignore the
results of the curve due to the underlying 1-1 mapping.

This commit introduces a check to ensure that this automatic mapping won't
happen during such invalid ranges.

6 years agoFix for uninitialised var
Joshua Leung [Mon, 4 Nov 2013 01:18:33 +0000 (01:18 +0000)]
Fix for uninitialised var

6 years agocode cleanup: remove unused operator (select all handles this case now)
Campbell Barton [Mon, 4 Nov 2013 00:55:00 +0000 (00:55 +0000)]
code cleanup: remove unused operator (select all handles this case now)

6 years agoFix [#37297] Crash when Adding and Removing Fluid Particlesystem.
Bastien Montagne [Sun, 3 Nov 2013 21:11:27 +0000 (21:11 +0000)]
Fix [#37297] Crash when Adding and Removing Fluid Particlesystem.

Using freed mem...

6 years agoFix [#37275] can't import 2.68a keymap
Bastien Montagne [Sun, 3 Nov 2013 18:04:45 +0000 (18:04 +0000)]
Fix [#37275] can't import 2.68a keymap

Now exported keymaps will still be usable accross versions, even if some operator properties disappear (write a warning in console in this case, instead of "crashing").

Also factorized a bit of code here!

6 years agoMinor code improvements: avoid unnecessary Python object allocations in Freestyle...
Tamito Kajiyama [Sun, 3 Nov 2013 14:25:37 +0000 (14:25 +0000)]
Minor code improvements: avoid unnecessary Python object allocations in Freestyle color blending.

6 years agoFix for missing calls of BaseMath_ReadCallback() when accessing vector/color elements.
Tamito Kajiyama [Sun, 3 Nov 2013 14:24:02 +0000 (14:24 +0000)]
Fix for missing calls of BaseMath_ReadCallback() when accessing vector/color elements.
This bug was causing wrong color blending results in Freestyle color modifiers.
Problem report from Light BWK through personal communications, thanks!

6 years agoOSX: Fix an error in own 61065
Jens Verwiebe [Sun, 3 Nov 2013 13:10:18 +0000 (13:10 +0000)]
OSX: Fix an error in own 61065

6 years agoMotion tracking: use is_keyed to match naming in other areas of RNA
Sergey Sharybin [Sun, 3 Nov 2013 09:33:17 +0000 (09:33 +0000)]
Motion tracking: use is_keyed to match naming in other areas of RNA

6 years agocode cleanup: warnings
Campbell Barton [Sun, 3 Nov 2013 05:19:55 +0000 (05:19 +0000)]
code cleanup: warnings

6 years agoOSX: give application bundles an own icon to better differentiate vs. folders
Jens Verwiebe [Sat, 2 Nov 2013 17:58:53 +0000 (17:58 +0000)]
OSX: give application bundles an own icon to better differentiate vs. folders

6 years agoBugfix [#36687] Animation channels can't be grouped in action editor
Joshua Leung [Sat, 2 Nov 2013 13:11:06 +0000 (13:11 +0000)]
Bugfix [#36687] Animation channels can't be grouped in action editor

Internal filtering flags used to obtain AnimData blocks as result were not
working correctly in Action and ShapeKey modes. Instead, in these modes, they
were often returning F-Curves instead, which lead to the grouping operating
failing (and perhaps other unidentified bugs)

6 years agoQuick test commit - fixing what looks like a typo in a comment
Joshua Leung [Sat, 2 Nov 2013 13:08:48 +0000 (13:08 +0000)]
Quick test commit - fixing what looks like a typo in a comment

6 years agoFix: tooltip when adding a new 'Excluded path' for python execution
Andrea Weikert [Sat, 2 Nov 2013 10:48:34 +0000 (10:48 +0000)]
Fix: tooltip when adding a new 'Excluded path' for python execution

6 years agoFix: error message when removing an 'Excluded path' for python execution, small lefto...
Andrea Weikert [Sat, 2 Nov 2013 10:37:42 +0000 (10:37 +0000)]
Fix: error message when removing an 'Excluded path' for python execution, small leftover from copy/paste

6 years agofix for possible leak in the expression controller find identifier.
Campbell Barton [Sat, 2 Nov 2013 05:18:31 +0000 (05:18 +0000)]
fix for possible leak in the expression controller find identifier.

6 years agofix memory leak with navmesh
Campbell Barton [Sat, 2 Nov 2013 04:22:08 +0000 (04:22 +0000)]
fix memory leak with navmesh

6 years agocode cleanup: warnings
Campbell Barton [Sat, 2 Nov 2013 04:04:30 +0000 (04:04 +0000)]
code cleanup: warnings

6 years agoFix [#37290] Movie files filtering does not show *.ts files.
Bastien Montagne [Fri, 1 Nov 2013 18:20:21 +0000 (18:20 +0000)]
Fix [#37290] Movie files filtering does not show *.ts files.

6 years agomissing bit from last float precision update (patch by 'unknow' via irc)
Dalai Felinto [Fri, 1 Nov 2013 18:11:54 +0000 (18:11 +0000)]
missing bit from last float precision update (patch by 'unknow' via irc)

6 years agoFix blenderplayer build, stub for modifier_skin_customdata_ensure(...)
Thomas Beck [Fri, 1 Nov 2013 16:39:11 +0000 (16:39 +0000)]
Fix blenderplayer build, stub for modifier_skin_customdata_ensure(...)

6 years ago* Extrude options where available in object mode (search menu), resulting in error...
Thomas Dinges [Fri, 1 Nov 2013 13:14:17 +0000 (13:14 +0000)]
* Extrude options where available in object mode (search menu), resulting in error msg.
Added polls now, patch by Gottfried Hofmann (gottfried).

6 years agoAdd 'cut-through' option for Knife Project operator.
Howard Trickey [Fri, 1 Nov 2013 11:42:11 +0000 (11:42 +0000)]
Add 'cut-through' option for Knife Project operator.

If enabled, it makes knife project act as the
cut-through (Shift-K) version of knife.
This option will soon be more useful when a better
cut-though Knife change is submitted, allowing
this to work for cuts within faces in addition
to cuts across them.

6 years agoCycles:
Thomas Dinges [Fri, 1 Nov 2013 09:37:42 +0000 (09:37 +0000)]
* Change the default Light Path settings.

* Diffuse/Glossy bounces are now set to 4, to give a bit faster renders in default scenes. More bounces are often not needed (especially in animation).
* Transmission bounces have been increased to 12, to not run into problems with dark glass too quickly.
* Max/Min bounces are now 12/3.

6 years agoFix [#37266] Skin modifier can't be copied.
Bastien Montagne [Fri, 1 Nov 2013 08:31:36 +0000 (08:31 +0000)]
Fix [#37266] Skin modifier can't be copied.

Patch by Martin Felke, many thanks.

When copying that modifier across objects, we also have to ensure that a skin CDLayer is present in dest objects (just as when adding it).

6 years agoFix [#37224] Icons: tiny interface issue with certain DPI settings.
Bastien Montagne [Fri, 1 Nov 2013 08:03:25 +0000 (08:03 +0000)]
Fix [#37224] Icons: tiny interface issue with certain DPI settings.

For some reason (consistant margin?), the whole icon stuff is shifted by two vertically in the svg file - but the bottom row of numbers was not, comming in touch with the icon area, and at some DPI values it showed in icons...

6 years agoFix #37270: Blender crashes in several situations
Sergey Sharybin [Fri, 1 Nov 2013 04:24:29 +0000 (04:24 +0000)]
Fix #37270: Blender crashes in several situations

Issue was caused by cloth modifier used for particle
system dynamic have an invalid error field pointer.
Seems at some point cloth failed to apply and set an

This commit only fixes crash with existing file,
will look further into why exactly such situation

6 years agoExpose is_keyframed flag for tracking marker
Sergey Sharybin [Fri, 1 Nov 2013 04:06:01 +0000 (04:06 +0000)]
Expose is_keyframed flag for tracking marker

So now it's possible to know from python whether
marker is a keyframe or tracked one.

6 years agofix for weight-paint crash picking the weight with the mask modifier (or any modifier...
Campbell Barton [Fri, 1 Nov 2013 01:14:36 +0000 (01:14 +0000)]
fix for weight-paint crash picking the weight with the mask modifier (or any modifier which removes geometry).
was incorrectly mixing indices from the DerivedMesh and the original mesh.

6 years agocode cleanup: spelling
Campbell Barton [Thu, 31 Oct 2013 23:52:44 +0000 (23:52 +0000)]
code cleanup: spelling

6 years agoFIX: [#37199] .ogg file is not recognized as a movie in File Browser
Andrea Weikert [Thu, 31 Oct 2013 21:00:55 +0000 (21:00 +0000)]
FIX: [#37199] .ogg file is not recognized as a movie in File Browser

CHANGE: writing an animation in the ogg movie format now defaults to .ogv (ogg/video) which is recommended by for video
CHANGE: for .ogg files a check is added whether Blender can read it as a movie (is avi or ffmpeg movie), otherwise assume audio
CHANGE: the anim player now filters for the same extensions as the file browser

6 years agoFix #37261 Rendering a Render Layer from another scene doesn't update.
Lukas Toenne [Thu, 31 Oct 2013 17:20:31 +0000 (17:20 +0000)]
Fix #37261 Rendering a Render Layer from another scene doesn't update.

The scene pointer used for looking up the appropriate source of render result images in the image editor was always taken from context. This means that render results for a different scene would never be
displayed in the image editor.

To give feedback on running renders, try to get the running render job's scene pointer in the image editor for render result type images. This only happens during rendering, apart from that the regular
context scene result is displayed.

6 years agocode cleanup: spelling
Campbell Barton [Thu, 31 Oct 2013 14:10:01 +0000 (14:10 +0000)]
code cleanup: spelling

6 years agoadd atomic_ops.h to cmake's source code listing.
Campbell Barton [Thu, 31 Oct 2013 14:09:01 +0000 (14:09 +0000)]
add atomic_ops.h to cmake's source code listing.

6 years agoMinor update to r60953.
Bastien Montagne [Thu, 31 Oct 2013 13:11:01 +0000 (13:11 +0000)]
Minor update to r60953.

6 years agoFix for Freestyle rendering errors with color blend modes SOFT_LIGHT and LINEAR_LIGHT.
Tamito Kajiyama [Thu, 31 Oct 2013 07:41:40 +0000 (07:41 +0000)]
Fix for Freestyle rendering errors with color blend modes SOFT_LIGHT and LINEAR_LIGHT.

6 years agocode cleanup: warnings
Campbell Barton [Thu, 31 Oct 2013 03:30:20 +0000 (03:30 +0000)]
code cleanup: warnings

6 years agoPython ui: increase interface float precision limit from 6 to 7
Dalai Felinto [Thu, 31 Oct 2013 03:21:20 +0000 (03:21 +0000)]
Python ui: increase interface float precision limit from 6 to 7

6 years agocancelling an operator popup now calls the operators cancel callback.
Campbell Barton [Wed, 30 Oct 2013 23:15:27 +0000 (23:15 +0000)]
cancelling an operator popup now calls the operators cancel callback.

6 years agoremove return argument from wmOperatorType->cancel, was only ever returning OPERATOR_...
Campbell Barton [Wed, 30 Oct 2013 23:08:53 +0000 (23:08 +0000)]
remove return argument from wmOperatorType->cancel, was only ever returning OPERATOR_CANCELLED.

6 years agofix for uninitialized value use in newly added fcurve normalized view.
Campbell Barton [Wed, 30 Oct 2013 22:13:28 +0000 (22:13 +0000)]
fix for uninitialized value use in newly added fcurve normalized view.
also quiet warning without openmp.

6 years agoAdd .webm to the list of extensions recognized as video
Andrea Weikert [Wed, 30 Oct 2013 21:16:13 +0000 (21:16 +0000)]
Add .webm to the list of extensions recognized as video
(Patch by Philipp Oeser, submitted in [#37199], many thanks)

6 years agoBugfix [#37185] Maya Config: Move and Scale Don't Work In Dope Sheet
Joshua Leung [Wed, 30 Oct 2013 11:33:29 +0000 (11:33 +0000)]
Bugfix [#37185] Maya Config: Move and Scale Don't Work In Dope Sheet

Maya keymap was activating wrong transform modes for translation/scaling - in
DopeSheet and NLA, these use the TIME_TRANSLATE/TIME_SCALE modes instead

6 years agoCode cleanup: typo-fix
Sergey Sharybin [Wed, 30 Oct 2013 11:33:11 +0000 (11:33 +0000)]
Code cleanup: typo-fix

6 years agorevert own commit r60607, caused bug [#37253]
Campbell Barton [Wed, 30 Oct 2013 11:32:30 +0000 (11:32 +0000)]
revert own commit r60607, caused bug [#37253]

6 years agoFix #37194, OSL script crashes blender. The lookup functions for finding Cycles shade...
Lukas Toenne [Wed, 30 Oct 2013 11:21:31 +0000 (11:21 +0000)]
Fix #37194, OSL script crashes blender. The lookup functions for finding Cycles shader inputs/outputs based on socket names are using a few modifications on the Blender socket names. But these only apply
to standard nodes where the Blender socket names can differ from associated Cycles names and may require additional indices to make them unique. Script node sockets are already unique and exact due to
being generated from the script function parameters.

6 years agoFix #37252: Mask modifier doesn't copy settings on object copy
Sergey Sharybin [Wed, 30 Oct 2013 09:44:29 +0000 (09:44 +0000)]
Fix #37252: Mask modifier doesn't copy settings on object copy

6 years agoAdd some update notifiers for a few boolean paint properties (probably a
Antony Riakiotakis [Wed, 30 Oct 2013 04:56:32 +0000 (04:56 +0000)]
Add some update notifiers for a few boolean paint properties (probably a
lot more needed) so that they get properly updated in UI when the
property is edited through python or a custom key binding to context
toggle operator.

6 years agoOne more optimization for lasso, clip PBVH against object space planes
Antony Riakiotakis [Wed, 30 Oct 2013 01:58:19 +0000 (01:58 +0000)]
One more optimization for lasso, clip PBVH against object space planes
of lasso enclosing rectangle.

6 years agoOptimization of lasso masking using scanfill.
Antony Riakiotakis [Wed, 30 Oct 2013 01:20:08 +0000 (01:20 +0000)]
Optimization of lasso masking using scanfill.

Thanks to Campbell for the advice!

6 years agoSome comment fixes, add new sculpt masking operators to menus
Antony Riakiotakis [Wed, 30 Oct 2013 00:54:41 +0000 (00:54 +0000)]
Some comment fixes, add new sculpt masking operators to menus

6 years agoLasso select tool for masking in sculpting. Initial code, non optimized
Antony Riakiotakis [Wed, 30 Oct 2013 00:37:13 +0000 (00:37 +0000)]
Lasso select tool for masking in sculpting. Initial code, non optimized
for now.
Used to be ultra terrible but with threading (openmp) there is slightly
better performance
and is ready for testing. To use press shift-ctrl-lclick. Still no
ability to
remove mask. Coming soon.

Also make box selection threaded (openmp) and comment fix.

6 years agoCode cleanup: more int->bool conversions mask module
Sergey Sharybin [Tue, 29 Oct 2013 18:56:21 +0000 (18:56 +0000)]
Code cleanup: more int->bool conversions mask module

6 years agoCode cleanup: use bool instead of int in mask module
Sergey Sharybin [Tue, 29 Oct 2013 18:46:45 +0000 (18:46 +0000)]
Code cleanup: use bool instead of int in mask module

6 years agoProject Pampa request: FCurves normalized display
Sergey Sharybin [Tue, 29 Oct 2013 18:10:52 +0000 (18:10 +0000)]
Project Pampa request: FCurves normalized display

Added two options to a header of FCurve editor:

- Normalize which makes it so every individual
  curve is fit into -1..1 space.

- Auto-normalize, which probably is to be called
  "Lock" which "locks" curve normalization scale.
  This is useful to prevent curves from jumping
  around when tweaking it.

It's debatable whether it need to be a button to
normalize curves n purpose only, and it's fully
depends on animator's workflow.

Here during Project Pampa we've got Francesco
who get used to auto-renormalization and Hjalti
who prefers locked behavior.

Docs are to be ready soon by Francesco.
Thanks Brecht for the review!

6 years agoMake anim system safer for threading
Sergey Sharybin [Tue, 29 Oct 2013 18:10:48 +0000 (18:10 +0000)]
Make anim system safer for threading

Remove usages of ANIM_unit_mapping_apply_fcurve in favor of
runtime scale factor apply.

There're still calls to ANIM_nla_mapping_apply_fcurve are
hanging around, they're the next t be cleaned up!

6 years agoFix #37192, Rendered preview causes crash when deleting a material node in shader...
Lukas Toenne [Tue, 29 Oct 2013 17:46:01 +0000 (17:46 +0000)]
Fix #37192, Rendered preview causes crash when deleting a material node in shader node editor. The 'free' callback for node execution data was accessed from the node->typeinfo, but this pointer can
become invalid because the render database is not immediately freed after the job finishes. To avoid access to dangling node pointers, store the function callback in the exec data itself. The node
pointer must not be accessed in the free function (wasn't used before either), these functions are purely for the execution data.

6 years agofix [#37250] dynotopo initial triangulation
Dalai Felinto [Tue, 29 Oct 2013 17:14:43 +0000 (17:14 +0000)]
fix [#37250] dynotopo initial triangulation

hmrf arguments order was wrong ... but int/bool casting made it pass
through the compiler

6 years agoCode cleanup: remove unused block from multire baker
Sergey Sharybin [Tue, 29 Oct 2013 16:43:58 +0000 (16:43 +0000)]
Code cleanup: remove unused block from multire baker

Mapping from grid on ll 0 was not used and was incorrect
since bmesh merge anyway.

6 years agoEnable multires bake to level 0
Sergey Sharybin [Tue, 29 Oct 2013 16:43:56 +0000 (16:43 +0000)]
Enable multires bake to level 0

Uses trick with crating simple subdivided mesh with
level 1 and bakes against it.

From quick tests seems to be working correct, more
tests are welcome.

6 years agoFix #37246: Dynamic paint end frame is limited to 9999
Sergey Sharybin [Tue, 29 Oct 2013 16:06:21 +0000 (16:06 +0000)]
Fix #37246: Dynamic paint end frame is limited to 9999

Not sure what's actual reason for this, increased it to MAXFRAMRF.

6 years agoProject Pampa fix: animplayer will use preview range if preview is enabled
Sergey Sharybin [Tue, 29 Oct 2013 15:49:01 +0000 (15:49 +0000)]
Project Pampa fix: animplayer will use preview range if preview is enabled

6 years agoFix for particle texture influence bug causing undefined (nan) values: When using...
Lukas Toenne [Tue, 29 Oct 2013 14:48:25 +0000 (14:48 +0000)]
Fix for particle texture influence bug causing undefined (nan) values: When using a texture with "Particles/Strands" coordinates that in turn has a "Lifetime" influence on the particles, the texture eval
accesses the particle lifetime/dietime data before it is actually defined! This is a design flaw, but to avoid corrupted data for now just initialize the lifetime/dietime values for particles in advance
before evaluating the texture.

6 years agoFix #37152: Knife tool doesn't work properly with Maya preset (Zooming and Panning)
Sergey Sharybin [Tue, 29 Oct 2013 11:59:03 +0000 (11:59 +0000)]
Fix #37152: Knife tool doesn't work properly with Maya preset (Zooming and Panning)

6 years ago(Due to a change in function call) Added default triangulation method for Collada...
Gaia Clary [Tue, 29 Oct 2013 10:49:44 +0000 (10:49 +0000)]
(Due to a change in function call) Added default triangulation method for Collada exporter

6 years agoFix collada compile, too less arguments
Jens Verwiebe [Tue, 29 Oct 2013 10:13:49 +0000 (10:13 +0000)]
Fix collada compile, too less arguments

6 years agostyle cleanup
Campbell Barton [Tue, 29 Oct 2013 03:04:46 +0000 (03:04 +0000)]
style cleanup

6 years agoTriangulate Modifier: using different ngon and quad methods
Dalai Felinto [Tue, 29 Oct 2013 02:42:51 +0000 (02:42 +0000)]
Triangulate Modifier: using different ngon and quad methods

Quads: Beauty, Fixed, Fixed Alternate, Shortest Diagonal
Ngons: Beauty, Scanfill

* Shortest Diagonal is the default method in the modifier (popular
  elsewhere), but beauty is the default in Ctrl+T).

* Remove the need for output slot and beauty operator to be called
after Clt+T

Patch with collaborations and reviewed by Campbell Barton

6 years agoEagerly attempt to refine a track before doing a brute search
Keir Mierle [Tue, 29 Oct 2013 01:06:50 +0000 (01:06 +0000)]
Eagerly attempt to refine a track before doing a brute search

Before the refinement phase of tracking, a brute force SAD search
is run across the search area. This works well but is slow;
especially if the guess for the track's location is accurate.

This patch runs a refinement phase before running a brute force
search, hoping that the guessed position (in x2, y2) is close to
the best answer. If it is, then no brute search is done. If it is
not, then a normal brute force search followed by refinement is

In some cases this may produce worse tracks than before; the
regressions will need investigation. The predictive motion model
(to be implemented) will reduce the probability of that happening.

6 years agopatch [#37218] Split operator for armatures
Campbell Barton [Tue, 29 Oct 2013 00:10:03 +0000 (00:10 +0000)]
patch [#37218] Split operator for armatures
from Henrik Aarnio (hjaarnio)

6 years agopatch [#37217] Make WM_gesture_lines draw the lasso unfilled. (Adding reroute nodes...
Campbell Barton [Tue, 29 Oct 2013 00:05:03 +0000 (00:05 +0000)]
patch [#37217] Make WM_gesture_lines draw the lasso unfilled. (Adding reroute nodes, cutting node links)
by Henrik Aarnio (hjaarnio)

6 years agopatch [#37219] RenderParts.partsdone is not a boolean, but a counter
Campbell Barton [Mon, 28 Oct 2013 23:59:28 +0000 (23:59 +0000)]
patch [#37219] RenderParts.partsdone is not a boolean, but a counter
from Lawrence D'Oliveiro (ldo)

6 years agopatch [#37229] Save process slot in blender-softwaregl script
Campbell Barton [Mon, 28 Oct 2013 23:48:53 +0000 (23:48 +0000)]
patch [#37229] Save process slot in blender-softwaregl script
from Lawrence D'Oliveiro (ldo)

6 years agopatch [#37231] Bone fill sets bone to connected, even if neither of its points is...
Campbell Barton [Mon, 28 Oct 2013 23:45:36 +0000 (23:45 +0000)]
patch [#37231] Bone fill sets bone to connected, even if neither of its points is at the tail point of points it's connecting.
from Henrik Aarnio (hjaarnio)

6 years agocorrect error in recent commit.
Campbell Barton [Mon, 28 Oct 2013 23:35:25 +0000 (23:35 +0000)]
correct error in recent commit.

6 years agoMake it possible to track specified clip
Sergey Sharybin [Mon, 28 Oct 2013 21:49:49 +0000 (21:49 +0000)]
Make it possible to track specified clip

Added a property to track_marker operator to be
able to render clip which name was passed to the
operator instead of clip from current clip editor

Very much useful for automatic tests.