blender.git
5 years agoLibmv: Code cleanup, avoid C11 feature with structure re-definition
Sergey Sharybin [Fri, 31 Oct 2014 22:18:40 +0000 (03:18 +0500)]
Libmv: Code cleanup, avoid C11 feature with structure re-definition

5 years agoCycles: Code cleanup, proper class vs. struct declaration
Sergey Sharybin [Fri, 31 Oct 2014 22:09:46 +0000 (03:09 +0500)]
Cycles: Code cleanup, proper class vs. struct declaration

5 years agoLibmv: Code cleanup, mixed class/struct in declaration/definition
Sergey Sharybin [Fri, 31 Oct 2014 21:53:36 +0000 (02:53 +0500)]
Libmv: Code cleanup, mixed class/struct in declaration/definition

5 years agoCleanup: Remove unused light function.
Thomas Dinges [Fri, 31 Oct 2014 21:54:05 +0000 (22:54 +0100)]
Cleanup: Remove unused light function.

5 years agoOptimize vertex parent for edit mode without modifiers
Sergey Sharybin [Fri, 31 Oct 2014 19:15:32 +0000 (20:15 +0100)]
Optimize vertex parent for edit mode without modifiers

No need to run lookup on the origindex in this case at all.

5 years agoOptimize vertex parent in cases there are only deform and SS modifiers
Sergey Sharybin [Fri, 31 Oct 2014 19:06:19 +0000 (20:06 +0100)]
Optimize vertex parent in cases there are only deform and SS modifiers

In cases when the subsurf modifier is the last in the stack and there
are only deformation modifiers before it we can skip doing full orig
vertex lookup.

This is rather common situation here in animatic.

5 years agoSCons: Expose missing command line arguments controlling GLEW
Sergey Sharybin [Fri, 31 Oct 2014 17:34:25 +0000 (22:34 +0500)]
SCons: Expose missing command line arguments controlling GLEW

5 years agoFix T42349: Windows Render command line Cycles Crash
Sergey Sharybin [Fri, 31 Oct 2014 17:28:39 +0000 (22:28 +0500)]
Fix T42349: Windows Render command line Cycles Crash

The issue was caused by GLEW MX enabled in SCons by default so
basically previous commit already fixed the crash. But we need
to be safe here.

For now the fix is simple and not that clean, just check if
there's an OpenGL context available and if not we don't do any
GLSL magic.

This is to be cleaned up after some discussion with the viewport
project guys.

5 years agoDisable GLEW_MX in scons
Sergey Sharybin [Fri, 31 Oct 2014 17:12:21 +0000 (22:12 +0500)]
Disable GLEW_MX in scons

This was already disabled in CMake and we need to keep build systems
in 100% synchronized state.

P.S. Would it be a real fix to drop SCons?

5 years agoFollowup to rB154af70f2b5b7b: always 'pack in' ongoing versionning code when bumping...
Bastien Montagne [Fri, 31 Oct 2014 17:26:58 +0000 (18:26 +0100)]
Followup to rB154af70f2b5b7b: always 'pack in' ongoing versionning code when bumping version.

5 years agoThemes auto-update from py script.
Bastien Montagne [Fri, 31 Oct 2014 17:14:20 +0000 (18:14 +0100)]
Themes auto-update from py script.

5 years agoFix T42454: Themes: Some colors were missed somehow in an update two years ago.
Bastien Montagne [Fri, 31 Oct 2014 17:12:28 +0000 (18:12 +0100)]
Fix T42454: Themes: Some colors were missed somehow in an update two years ago.

Had to fix it by hand (restting them to default values), theme maintainers,
please check your theme from time to time and track such ugly 'pitch black' fields.

And we devs should run theme update script much more often too - will do in next commit!

5 years agoFix T42432: Correct BLENDER_SUBVERSION
julianeisel [Fri, 31 Oct 2014 16:54:18 +0000 (17:54 +0100)]
Fix T42432: Correct BLENDER_SUBVERSION

Bump up BLENDER_SUBVERSION to 2 to allow saving of widget emboss.

5 years agoCleanup: naming
Campbell Barton [Fri, 31 Oct 2014 14:32:59 +0000 (15:32 +0100)]
Cleanup: naming

5 years agoCode cleanup: synchronize const qualifiers between declaration and definition
Sergey Sharybin [Fri, 31 Oct 2014 15:17:02 +0000 (16:17 +0100)]
Code cleanup: synchronize const qualifiers between declaration and definition

5 years agoCode cleanup: Remove unused variable
Sergey Sharybin [Fri, 31 Oct 2014 15:13:13 +0000 (16:13 +0100)]
Code cleanup: Remove unused variable

5 years agoCompositor: Don't let HSV correction to output negative colors
Sergey Sharybin [Fri, 31 Oct 2014 13:51:59 +0000 (14:51 +0100)]
Compositor: Don't let HSV correction to output negative colors

5 years agoFix T42391: HSV correction shader node gives negative values
Sergey Sharybin [Fri, 31 Oct 2014 13:42:48 +0000 (14:42 +0100)]
Fix T42391: HSV correction shader node gives negative values

This mainly happens when over-saturating already saturated color.
After some discussion with Campbell and loads of tests we decided
to clamp the result RGB color. As an alternative we might want to
clamp corrected HSV values instead, but that would lead to some
larger changes in the render results.

TODO: The same is to be done for compositor nodes.

5 years agoTexture Paint Add Simple UVs:
Antony Riakiotakis [Fri, 31 Oct 2014 13:37:36 +0000 (14:37 +0100)]
Texture Paint Add Simple UVs:

Add simple uvs now does a cube unwrap and pack operation. Result is not
optimal by far but it should not result in crashes and it will be quite
usable for simple cases.

5 years agoFix T42277: Apply track's mask on displaying preview not working
Sergey Sharybin [Fri, 31 Oct 2014 11:59:55 +0000 (12:59 +0100)]
Fix T42277: Apply track's mask on displaying preview not working

5 years agoFix T42344: EWA filter produces blured results
Sergey Sharybin [Fri, 31 Oct 2014 11:35:20 +0000 (12:35 +0100)]
Fix T42344: EWA filter produces blured results

Derivatives variable names are swapped in the old EWA filter code,
need to adjust for that.

TODO: Make naming fore clear in there.

5 years agoFix T42401: Gaussian blur node is visibly squared-off at larger sizes
Sergey Sharybin [Fri, 31 Oct 2014 11:15:38 +0000 (12:15 +0100)]
Fix T42401: Gaussian blur node is visibly squared-off at larger sizes

The root of the issue comes to the way how we sample the gaussian filter in
RE_filter_value(). We need to scale x to -3*sigma..3*sigma segment in order
to get the whole bell.

The old code tried to do it, but failed dramatically, plus it used some weird
gaussian sampling formula. Replaced it with much more clear one, which gives
proper blur now.

There's no visible different in AA sampling in BI render tho.

Other filters like Mitchell still tends to give wrong square shaped blurs,
but they're much more difficult to resolve because they're just wrong in
the code -- for some reason smaller kernel size means more blur. Let's solve
this later.

5 years agoFix T42447: Shrinkwrap constraint: mismatch in handling sclaing in projection case.
Bastien Montagne [Fri, 31 Oct 2014 10:07:52 +0000 (11:07 +0100)]
Fix T42447: Shrinkwrap constraint: mismatch in handling sclaing in projection case.

Constraint space conversion ignores object scale, which is OK in most cases. But here,
we are converting a normal from world to local space, and when later converting it
into target space to actually do the BVH raycast, we use TransformSpace which
does applies objects' scaling to normals, as expected.

Best solution here is to also take object's scale into account when converting
from local to world space.

5 years agoRevert "Scons/funstuff: notify when the binaries are compiled"
Campbell Barton [Fri, 31 Oct 2014 09:35:33 +0000 (10:35 +0100)]
Revert "Scons/funstuff: notify when the binaries are compiled"

This reverts commit cdf53701599ace0b3410cbb75b0313bd969c417c.

Please always keep build-systems simple, boring & functional,
anyone who liked notification popups can write a simple build script.

5 years agoView3d - walk: expose gravity toggle to keymap
Dalai Felinto [Fri, 31 Oct 2014 00:16:43 +0000 (22:16 -0200)]
View3d - walk: expose gravity toggle to keymap

5 years agoCleanup: use function to calc shell thickness
Campbell Barton [Thu, 30 Oct 2014 21:31:03 +0000 (22:31 +0100)]
Cleanup: use function to calc shell thickness

5 years agoRemove leftover error prints.
Antony Riakiotakis [Thu, 30 Oct 2014 20:48:44 +0000 (21:48 +0100)]
Remove leftover error prints.

5 years agoFix T42431 lagging on texture painting.
Antony Riakiotakis [Thu, 30 Oct 2014 20:47:25 +0000 (21:47 +0100)]
Fix T42431 lagging on texture painting.

Issue here is that seam outset was being calculated wrong and as a
result we filled too many unneeded pixels. Code can be improved here by
clamping perhaps but left it as is for now. Thanks to Campbell for the
help resolving this!

5 years agoLibmv: Support disabled color channels in tracking settings
Sergey Sharybin [Thu, 30 Oct 2014 18:13:53 +0000 (23:13 +0500)]
Libmv: Support disabled color channels in tracking settings

This was never ported to a new tracking pipeline and now it's done using
FrameAccessor::Transform routines. Quite striaghtforward, but i've changed
order of grayscale conversion in blender side with call of transform callback.

This way it's much easier to perform rescaling in libmv side.

5 years agoLibmv: Replace region tracker with autotracker in Blender
Sergey Sharybin [Sun, 22 Jun 2014 14:19:48 +0000 (20:19 +0600)]
Libmv: Replace region tracker with autotracker in Blender

The title actually tells it all, this commit switches Blender to use the new
autotrack API from Libmv.

From the user point of view it means that prediction model is now used when
tracking which gives really nice results.

All the other changes are not really visible for users, those are just frame
accessors, caches and so for the new API.

5 years agoLibmv: Add autotrack API to the C-API
Sergey Sharybin [Sun, 22 Jun 2014 08:53:21 +0000 (14:53 +0600)]
Libmv: Add autotrack API to the C-API

Pretty much straightforward changes, nothing to be mentioned specially.

5 years agoLibmv: Initial commit of unfinished AutoTrack API
Keir Mierle [Sun, 22 Jun 2014 08:34:23 +0000 (14:34 +0600)]
Libmv: Initial commit of unfinished AutoTrack API

This starts the creating the new AutoTrack API. The new API will
make it possible for libmv to do full autotracking, including
predictive tracking and also support multiple motion models (3D
planes etc).

The first goal (not in this patch) is to convert Blender to use
the new API without adding any new functionality.

This API currently contanins:

- Frame accessor to access frames which are stored in Blender side.
- New Tracks implementation
- New Reconstruction implementation

Currently this API only tested on doing the same frame-to-frame
tracking as the old API allowed to do. But it also supports now
predictive tracking which is based on the Kalman filter.

5 years agoFix project-file generators (didn't close files)
Campbell Barton [Thu, 30 Oct 2014 17:24:57 +0000 (18:24 +0100)]
Fix project-file generators (didn't close files)

5 years agoAdd a few more warnings so coders hopefully will take note that event
Antony Riakiotakis [Thu, 30 Oct 2014 15:16:29 +0000 (16:16 +0100)]
Add a few more warnings so coders hopefully will take note that event
codes get stored in keymaps and should not be changed

5 years agoComplementary fix for rB8054372: Follow the common naming scheme by
Lukas Tönne [Thu, 30 Oct 2014 14:26:41 +0000 (15:26 +0100)]
Complementary fix for rB8054372: Follow the common naming scheme by
using negate_mat3_m4 instead of negate_m4.

This avoids changing the behavior and only flips the 3x3 part of the
matrix.

5 years agoFix issue in fullscreen patch.
Antony Riakiotakis [Thu, 30 Oct 2014 14:24:56 +0000 (15:24 +0100)]
Fix issue in fullscreen patch.

Don't change old event codes! Keymap diffing depends on those being the
same across versions to work.

5 years agoMath Lib: pseudoinverse_m4_m4 changed input matrix
Campbell Barton [Thu, 30 Oct 2014 13:49:57 +0000 (14:49 +0100)]
Math Lib: pseudoinverse_m4_m4 changed input matrix

5 years agoTransform: snapping: normal-align: Cleanup!
Bastien Montagne [Thu, 30 Oct 2014 11:54:05 +0000 (12:54 +0100)]
Transform: snapping: normal-align: Cleanup!

Use BLI's `rotation_between_vecs_to_mat3` helper instead of own custom code,
both simplifies the code and fixes wrong handling when snapped normal was exactly
opposed to org one (i.e. 180° rot case).

5 years agoMath Lib: add transpose_m3_m3, m3_m4, m4_m4
Campbell Barton [Thu, 30 Oct 2014 11:13:00 +0000 (12:13 +0100)]
Math Lib: add transpose_m3_m3, m3_m4, m4_m4

5 years agoCycles / CUDA: Better fix for missing sm_52 kernel, in case user compiles himself.
Thomas Dinges [Thu, 30 Oct 2014 10:42:59 +0000 (11:42 +0100)]
Cycles / CUDA: Better fix for missing sm_52 kernel, in case user compiles himself.

5 years agoCycles / OSL: Support microfacet() closure color function from OSL 1.5
Thomas Dinges [Thu, 30 Oct 2014 10:33:27 +0000 (11:33 +0100)]
Cycles / OSL: Support microfacet() closure color function from OSL 1.5

This is basically just a wrapper class, which maps the generic call from the OSL spec to our closures.

Example usage:

shader microfacet_osl(
    color Color = color(0.8),
    int Distribution = 0,
    normal Normal = N,
    vector Tangent = normalize(dPdu),
    float RoughnessU = 0.0,
    float RoughnessV = 0.0,
    float IOR = 1.4,
    int Refract = 0,
    output closure color BSDF = 0)
{
    if (Distribution == 0)
        BSDF = Color * microfacet("ggx", Normal, Tangent, RoughnessU, RoughnessV, IOR, Refract);
    else
        BSDF = Color * microfacet("beckmann", Normal, Tangent, RoughnessU, RoughnessV, IOR, Refract);
}

5 years agoFix for mat3_to_rot_size modifying input matrix
Campbell Barton [Thu, 30 Oct 2014 09:37:27 +0000 (10:37 +0100)]
Fix for mat3_to_rot_size modifying input matrix

5 years agoFix negate_m3 (taking 4x4 matrix)
Campbell Barton [Thu, 30 Oct 2014 09:26:22 +0000 (10:26 +0100)]
Fix negate_m3 (taking 4x4 matrix)

Cycles bake used incorrectly.

5 years agoDeduplicate some code by using a function pointer to the real kernel
Martijn Berger [Thu, 30 Oct 2014 09:17:42 +0000 (10:17 +0100)]
Deduplicate some code by using a function pointer to the real kernel
This has no performance impact what so ever and is already used in the adaptive sampling patch

5 years agoFix T42444: Neg scale rotated object flips 2x axis
Campbell Barton [Thu, 30 Oct 2014 09:15:43 +0000 (10:15 +0100)]
Fix T42444: Neg scale rotated object flips 2x axis

This is a very old bug and a few attempts have been made to fix it before,
Old code was checking worldspace axis flipping, instead of axis flipping WRT the axis-alignment of the object.

5 years agoFix T42441: Smoke doesn't render in 3DView if domain object's maximum draw type is...
Bastien Montagne [Thu, 30 Oct 2014 08:34:17 +0000 (09:34 +0100)]
Fix T42441: Smoke doesn't render in 3DView if domain object's maximum draw type is set to "wire"

Added an exception in convertblender.c's is_object_hidden(), so that an object with active
smoke modifier is never considered hidden.

5 years agoUI: Fix tooltip text color not resetting correctly
julianeisel [Thu, 30 Oct 2014 00:50:48 +0000 (01:50 +0100)]
UI: Fix tooltip text color not resetting correctly

The "Reset to Default Theme" operator didn't reset the tooltip text color correctly.

5 years agoUI: let multi-num-drag change different ID values
Campbell Barton [Wed, 29 Oct 2014 21:35:09 +0000 (22:35 +0100)]
UI: let multi-num-drag change different ID values

Handy for property-chart addon.

5 years agoFix T42395: Stored "render data" in particles mismatches the particle
Lukas Tönne [Wed, 29 Oct 2014 17:47:11 +0000 (18:47 +0100)]
Fix T42395: Stored "render data" in particles mismatches the particle
amount and crashes after Cycles render.

This is a hack to fix this, but at this point the system is hopelessly
broken anyway and no good fix other than total rewriting is possible.

5 years agoCorrect stub
Campbell Barton [Wed, 29 Oct 2014 16:27:01 +0000 (17:27 +0100)]
Correct stub

5 years agoFix T42294: Bullet rigidbody point cache reading was using uninitialized
Lukas Tönne [Wed, 29 Oct 2014 13:59:29 +0000 (14:59 +0100)]
Fix T42294: Bullet rigidbody point cache reading was using uninitialized
key values for velocity.

5 years agoCleanup: warnings, typos
Campbell Barton [Wed, 29 Oct 2014 13:11:19 +0000 (14:11 +0100)]
Cleanup: warnings, typos

5 years agoFix for Ctrl+Del/Backspace not setting text dirty
Campbell Barton [Wed, 29 Oct 2014 12:51:27 +0000 (13:51 +0100)]
Fix for Ctrl+Del/Backspace not setting text dirty

5 years agoFix T42394: Copy Rigid Body Tools would not work as expected if dest ob had no rigidb...
Bastien Montagne [Wed, 29 Oct 2014 12:06:08 +0000 (13:06 +0100)]
Fix T42394: Copy Rigid Body Tools would not work as expected if dest ob had no rigidbody yet.

Adding new object to RigidBodyWorld obgroup is not a good way to do that, since it only
takes effect (create rigid_body for new objects) when you change current frame.

Better to use rigidbody.object_add() operator here!

5 years agoFix T42389: Removed outdated comment from custom_nodes.py example script.
Lukas Tönne [Wed, 29 Oct 2014 12:05:21 +0000 (13:05 +0100)]
Fix T42389: Removed outdated comment from custom_nodes.py example script.

There was some confusion about the bl_icon attribute being optional. In
fact the functionality for hiding a tree type in the editor header is
now controlled by writing a poll function for the tree type instead,
which is more flexible and clean.

http://www.blender.org/documentation/blender_python_api_2_71_6/bpy.types.NodeTree.html#bpy.types.NodeTree.poll

5 years agoCycles: Optimize math node without links to a single value node
Sergey Sharybin [Sun, 26 Oct 2014 14:40:04 +0000 (19:40 +0500)]
Cycles: Optimize math node without links to a single value node

Pretty straightforward implementation. Just needed to move some functions
around to make them available at shader compile time.

5 years agoCycles: Add a soft min/max UI value for volume step size, usually a range from 0...
Thomas Dinges [Wed, 29 Oct 2014 11:15:29 +0000 (12:15 +0100)]
Cycles: Add a soft min/max UI value for volume step size, usually a range from 0.01 to 1.0 is fine.

5 years agoRevert "Fix T42222"
Campbell Barton [Wed, 29 Oct 2014 10:10:07 +0000 (11:10 +0100)]
Revert "Fix T42222"

This reverts commit 507712db3fd7aa7bb903f6860f5a4eb29aa2be02.

Error was quite an old compiler, which had further warnings/errors.

Old compilers can just have this defined in BLI_math

5 years agoCleanup: Style fixes for closures, mainly bitflags and conditions.
Thomas Dinges [Wed, 29 Oct 2014 08:56:21 +0000 (09:56 +0100)]
Cleanup: Style fixes for closures, mainly bitflags and conditions.

5 years agoCleanup: Remove unused function in Translucent BSDF.
Thomas Dinges [Wed, 29 Oct 2014 08:42:19 +0000 (09:42 +0100)]
Cleanup: Remove unused function in Translucent BSDF.

5 years agoPlayer: fix conflicting type introduced in 2f0bdcb306999
Jens Verwiebe [Tue, 28 Oct 2014 20:39:54 +0000 (21:39 +0100)]
Player: fix conflicting type introduced in 2f0bdcb306999

5 years agoOSX/GHOST: fix T42305, appswitching not reliable, proposed by Fabio Arnold
Jens Verwiebe [Tue, 28 Oct 2014 19:43:52 +0000 (20:43 +0100)]
OSX/GHOST: fix T42305, appswitching not reliable, proposed by Fabio Arnold

5 years agoMarkers: optimize drawing
Campbell Barton [Tue, 28 Oct 2014 19:02:44 +0000 (20:02 +0100)]
Markers: optimize drawing

- avoid 2x glScalef per marker
- skip markers outside the view
- merge drawing into a single loop

5 years agoCleanup: rename draw_markers_time -> ED_markers_draw
Campbell Barton [Tue, 28 Oct 2014 18:03:13 +0000 (19:03 +0100)]
Cleanup: rename draw_markers_time -> ED_markers_draw

5 years agoMarkers: show the area that handles marker events
Campbell Barton [Tue, 28 Oct 2014 17:51:28 +0000 (18:51 +0100)]
Markers: show the area that handles marker events

5 years agoCleanup: remove redundant 'object' parent class
Campbell Barton [Tue, 28 Oct 2014 17:42:06 +0000 (18:42 +0100)]
Cleanup: remove redundant 'object' parent class

5 years agoCleanup: use SELECT flag
Campbell Barton [Tue, 28 Oct 2014 17:39:43 +0000 (18:39 +0100)]
Cleanup: use SELECT flag

5 years agoFix T41041: 'Delete keyframe' removes markers too
Campbell Barton [Tue, 28 Oct 2014 16:51:06 +0000 (17:51 +0100)]
Fix T41041: 'Delete keyframe' removes markers too

Operators that trigger UI events (but nothing else)
were using 'CANCELLED' making it impossible to tell if an invoke
function failed, or opened a menu.

5 years agoLattice: add-object radius should never scale data
Campbell Barton [Tue, 28 Oct 2014 14:53:31 +0000 (15:53 +0100)]
Lattice: add-object radius should never scale data

own error, lattice assumes rest state is unscaled data,
scaling needs to be done in object mode.

5 years agoWM: clear operator memory on file load
Campbell Barton [Tue, 28 Oct 2014 14:47:51 +0000 (15:47 +0100)]
WM: clear operator memory on file load

Was causing problems when opening scenes with different scale set.

5 years agoWM: unneeded alloc the operator-type iterator
Campbell Barton [Tue, 28 Oct 2014 14:41:05 +0000 (15:41 +0100)]
WM: unneeded alloc the operator-type iterator

5 years agoKeymap: PKey only runs BGE when engine is set
Campbell Barton [Tue, 28 Oct 2014 12:34:11 +0000 (13:34 +0100)]
Keymap: PKey only runs BGE when engine is set

5 years agoCleanup: de-duplicate engine-id's
Campbell Barton [Tue, 28 Oct 2014 11:49:04 +0000 (12:49 +0100)]
Cleanup: de-duplicate engine-id's

5 years agoRemove redundant BLI_exists call
Campbell Barton [Tue, 28 Oct 2014 10:33:05 +0000 (11:33 +0100)]
Remove redundant BLI_exists call

5 years agoBMesh Py API: Fast index lookups for vert/edge/faces
Campbell Barton [Tue, 28 Oct 2014 08:49:02 +0000 (09:49 +0100)]
BMesh Py API: Fast index lookups for vert/edge/faces

This changes the Py API to use array lookup table.
Previously this could be very slow since it would loop over all elements.

Now the python script is responsible for creating the internal lookup table (as with C code).

This will break some scripts.

5 years agoAdd Theme Option for UI Emboss
Jonathan Williamson [Mon, 27 Oct 2014 17:11:55 +0000 (18:11 +0100)]
Add Theme Option for UI Emboss

This adds a theme option for the embossing of UI widgets. By doing this users have much greater flexibility for creating nice themes. Previously many themes (particularly dark ones) looked quite bad due to the very obvious emboss. This made simpler, flat-style themes very challenging.

Closes T42228

Reviewed by @campbellbarton

5 years agoFix assert's, remove uiEndBlock calls
Campbell Barton [Mon, 27 Oct 2014 14:37:09 +0000 (15:37 +0100)]
Fix assert's, remove uiEndBlock calls

5 years agoFix T42393 crash in texture paint sampling in image mode when mesh did
Antony Riakiotakis [Mon, 27 Oct 2014 09:42:59 +0000 (10:42 +0100)]
Fix T42393 crash in texture paint sampling in image mode when mesh did
not have a material

5 years agoFix T42372: demo addon in doc was not handling keymaps correctly during (un)registration.
Bastien Montagne [Sun, 26 Oct 2014 09:01:03 +0000 (10:01 +0100)]
Fix T42372: demo addon in doc was not handling keymaps correctly during (un)registration.

First, you should unregister in reverse order you registered your operators, keymaps, etc.
Second, when registering keymaps you have to check keyconfigs are actually available (they are not in background mode).

5 years agoOSX/CLEW: silence hundreds of warnings: 'weak_import' attribute only applies to varia...
Jens Verwiebe [Fri, 24 Oct 2014 12:25:02 +0000 (14:25 +0200)]
OSX/CLEW: silence hundreds of warnings: 'weak_import' attribute only applies to variables and functions

5 years agoScons/funstuff: notify when the binaries are compiled
Jens Verwiebe [Thu, 23 Oct 2014 17:19:02 +0000 (19:19 +0200)]
Scons/funstuff: notify when the binaries are compiled

5 years agoTerminology Trim -> Slip tool. It's how it's mostly called in other
Antony Riakiotakis [Thu, 23 Oct 2014 14:48:34 +0000 (16:48 +0200)]
Terminology Trim -> Slip tool. It's how it's mostly called in other
software.

5 years agoFix T42367: Spline-ik offset evaluating curve
Campbell Barton [Thu, 23 Oct 2014 14:29:40 +0000 (16:29 +0200)]
Fix T42367: Spline-ik offset evaluating curve

5 years agoSpline IK: use malloc, arrays are filled instantly
Campbell Barton [Thu, 23 Oct 2014 14:12:47 +0000 (16:12 +0200)]
Spline IK: use malloc, arrays are filled instantly

5 years agoFix T42351: Freestyle will not render edges selected by Edge Type: Material Boundary...
Tamito Kajiyama [Thu, 23 Oct 2014 11:26:39 +0000 (20:26 +0900)]
Fix T42351: Freestyle will not render edges selected by Edge Type: Material Boundary, if the materials are different, but look identical.

5 years agoOSX/GHOST: more little cleanups
Jens Verwiebe [Thu, 23 Oct 2014 12:40:56 +0000 (14:40 +0200)]
OSX/GHOST: more little cleanups

5 years agoOSX/GHOST: some cleanups
Jens Verwiebe [Thu, 23 Oct 2014 12:34:56 +0000 (14:34 +0200)]
OSX/GHOST: some cleanups

5 years agoOSX: move notification into its own function
Jens Verwiebe [Thu, 23 Oct 2014 12:20:06 +0000 (14:20 +0200)]
OSX: move notification into its own function

5 years agoFix T42354 modal transform map not cycling through local/global
Antony Riakiotakis [Thu, 23 Oct 2014 12:16:36 +0000 (14:16 +0200)]
Fix T42354 modal transform map not cycling through local/global
orientation properly

Patch by Phillip Oeser, thanks!

5 years agoFix T42330 game engine does not allow texture slots generation.
Antony Riakiotakis [Thu, 23 Oct 2014 11:39:45 +0000 (13:39 +0200)]
Fix T42330 game engine does not allow texture slots generation.

5 years agoCleanup: spelling
Campbell Barton [Thu, 23 Oct 2014 08:38:14 +0000 (10:38 +0200)]
Cleanup: spelling

5 years agoChecked each of my (jwilkins) XXX notes.
Jason Wilkins [Thu, 23 Oct 2014 01:03:25 +0000 (20:03 -0500)]
Checked each of my (jwilkins) XXX notes.

The ones in extern/glew-es have been changed to NOTE instead of XXX

GHOST_ContextEGL.cpp: It really does seem that it is not possible to query the swap interval using EGL

GHOST_WidnowCocoa.h: The comment referring to Carbon is clearly out of date, so I removed it.

math_geom.c: The node about not using tmax again is correct, but the code is kept for a future maintainer who will need to know how to compute it if they modify that code.

paint_image_proj.c (2698): The question about integer truncation does not appear to have been resolved.  It still seems to be an incorrectly implementation of rounding (I'd suggest using the round function instead of this hack).

5 years agoFix T42352.
Jonathan Williamson [Wed, 22 Oct 2014 21:25:31 +0000 (23:25 +0200)]
Fix T42352.

Add Recalcuate Normals to the Faces menu, next to other shading options.

Differential revision: https://developer.blender.org/D841
Signed-off-by: Thomas Dinges
5 years agoUsual UI messages fixes.
Bastien Montagne [Wed, 22 Oct 2014 20:26:45 +0000 (22:26 +0200)]
Usual UI messages fixes.

5 years agoOSX: as we cannot be sure to have ARC, better release allocated notification
Jens Verwiebe [Wed, 22 Oct 2014 18:17:04 +0000 (20:17 +0200)]
OSX: as we cannot be sure to have ARC, better release allocated notification

5 years agoOSX/GHOST: use notifications to inform user about a progress reached 100%
Jens Verwiebe [Wed, 22 Oct 2014 17:59:00 +0000 (19:59 +0200)]
OSX/GHOST: use notifications to inform user about a progress reached 100%
The occurance can be controlled in NotificationCenter, todo: move to own function ?

5 years agoRevert "Cycles: Implement Mitchell-Netravali pixel filter"
Sergey Sharybin [Wed, 22 Oct 2014 16:52:04 +0000 (22:52 +0600)]
Revert "Cycles: Implement Mitchell-Netravali pixel filter"

As it appears we can't really use mitchell filter together with the
current filter importance sampling,

This reverts commit 742911314322e5dae3a07469d0ca53b61427f978.

5 years agoCycles: Implement Mitchell-Netravali pixel filter
Sergey Sharybin [Wed, 22 Oct 2014 16:08:17 +0000 (18:08 +0200)]
Cycles: Implement Mitchell-Netravali pixel filter

It's the same filter which is used by default by Blender Internal renderer
and it gives crispier edges than gaussian filter.

Default filter for Cycles is unchanged because it's unclear if new filter
gives more noise or not. After some further real production tests we can
consider making Mitchell filter default for Cycles as well.

5 years agoCycles: Code cleanup -- use bitshifts instead of hardcoded constants
Sergey Sharybin [Wed, 22 Oct 2014 14:51:40 +0000 (20:51 +0600)]
Cycles: Code cleanup -- use bitshifts instead of hardcoded constants

This way it's easier to extend bitfields and see when we start running
out of free bits.

Plus added brief description of what SD_VOLUME_CUBIC flag means.

5 years agoCycles: Get rid of hardcoded enum values in the code
Sergey Sharybin [Wed, 22 Oct 2014 14:17:03 +0000 (16:17 +0200)]
Cycles: Get rid of hardcoded enum values in the code

Still need to keep enum definition in sync with the python code,
but the code itself is a bit more clear to understand now.