4 years agoMinor tweaks to make fill and invert support gpixel operations
Antony Riakiotakis [Thu, 30 Apr 2015 11:52:25 +0000 (13:52 +0200)]
Minor tweaks to make fill and invert support gpixel operations

4 years agoCode Cleanup: Simplified insert_keyframe_button and delete_keyframe_button
Joshua Leung [Thu, 30 Apr 2015 10:58:53 +0000 (22:58 +1200)]
Code Cleanup: Simplified insert_keyframe_button and delete_keyframe_button

As a followup for the previous commit, do the same thing for the insert/delete
keyframe button operators as is done for the clear keyframes op. There really isn't
much need/reason for conducting the looping there, as those functions natively
handle this themselves already.

4 years agoCode Cleanup: Simplify Clear Keyframes operator's code
Joshua Leung [Thu, 30 Apr 2015 10:50:31 +0000 (22:50 +1200)]
Code Cleanup: Simplify Clear Keyframes operator's code

On second thought, the previous commit was just adding additional complexity which
wasn't needed, as the operator was wasting effort by doing this looping itself.

4 years agoFix T44558 - "Clear Keyframes" complains when operating on an array property and...
Joshua Leung [Thu, 30 Apr 2015 10:43:48 +0000 (22:43 +1200)]
Fix T44558 - "Clear Keyframes" complains when operating on an array property and it had deleted the action in the process

4 years agoCycles: Fix possible uninitialized XML read state which might cause crashes
Sergey Sharybin [Thu, 30 Apr 2015 10:46:09 +0000 (15:46 +0500)]
Cycles: Fix possible uninitialized XML read state which might cause crashes

4 years agoFix T44541 aka gigapixel image render support in blender.
Antony Riakiotakis [Thu, 30 Apr 2015 10:10:58 +0000 (12:10 +0200)]
Fix T44541 aka gigapixel image render support in blender.

Moral of the story: Make sure that size_t is used whenever pointer
arithmetic is involved. For images, that basically means whenever any
squared dimensions are involved. Casting an operand to size_t early in
the operation is usually sufficient to force the entire operation to

There might still be places lurking where we don't support this
correctly. This has been tested with render pipeline, quite a few image
functions (meaning we can paint on such images now, albeit somewhat
slowly ;) ) and export to jpeg. Too many places in code to check so I
guess we'll be handling cases as they come.

Don't try this at home unless you have an immense ammount of RAM.
First GPixel render of suzanne in the multiverse can be found here:

Can be viewed from blender (takes about 3.3 GB after loading but may
take more during loading so 8GB might be more safe to try this).

4 years agoFix T44556: Crash on the World Properties > Preview
Sergey Sharybin [Thu, 30 Apr 2015 09:52:59 +0000 (14:52 +0500)]
Fix T44556: Crash on the World Properties > Preview

Issue was caused by de-referencing NULL pointer, rres did not have any
views because nothing was rendered yet.

Needs more closer look about where else such a de-reference could happen.

4 years agoFix T44548: Cycles Tube Mapping off / not compatible with BI
Sergey Sharybin [Thu, 30 Apr 2015 09:21:32 +0000 (14:21 +0500)]
Fix T44548: Cycles Tube Mapping off / not compatible with BI

Was a typo in original implementation, probably a result of some code reshuffle
happened for optimization reasons.

4 years agoView3D: expose size as a distance in units
Campbell Barton [Thu, 30 Apr 2015 08:45:47 +0000 (18:45 +1000)]
View3D: expose size as a distance in units

also note that size/rotation doesn't work for camera views. see T44511

4 years agoCMake: Solve compilation error from path with a space
Sergey Sharybin [Thu, 30 Apr 2015 07:55:52 +0000 (12:55 +0500)]
CMake: Solve compilation error from path with a space

4 years agoCorrect docstring
Campbell Barton [Wed, 29 Apr 2015 23:55:54 +0000 (09:55 +1000)]
Correct docstring

4 years agoCorrect missing break
Campbell Barton [Wed, 29 Apr 2015 22:18:32 +0000 (08:18 +1000)]
Correct missing break

4 years agoCleanup: style
Campbell Barton [Wed, 29 Apr 2015 22:07:15 +0000 (08:07 +1000)]
Cleanup: style

4 years agoCleanup: Update some comments and add ToDo.
Thomas Dinges [Wed, 29 Apr 2015 21:56:46 +0000 (23:56 +0200)]
Cleanup: Update some comments and add ToDo.

4 years agoCycles: Record all possible volume intersections for SSS and camera checks
Thomas Dinges [Wed, 29 Apr 2015 21:21:05 +0000 (23:21 +0200)]
Cycles: Record all possible volume intersections for SSS and camera checks

This replaces sequential ray moving followed with scene intersection with
single BVH traversal, which gives us all possible intersections.

Only implemented for CPU, due to qsort and a bigger memory usage on GPU
which we rather avoid. GPU still uses the regular bvh volume intersection code, while CPU now uses the new code.

This improves render performance for scenes with:
a) Camera inside volume mesh
b) SSS mesh intersecting a volume mesh/domain

In simple volume files (not much geometry) performance is roughly the same
(slightly faster). In files with a lot of geometry, the performance
increase is larger. bmps.blend with a volume shader and camera inside the
mesh, it renders ~10% faster here.

Patch by Sergey and myself.

Differential Revision:

4 years agoFix T44484: Edge-split corrupts mesh
Campbell Barton [Wed, 29 Apr 2015 21:22:18 +0000 (07:22 +1000)]
Fix T44484: Edge-split corrupts mesh

Splitting non-manifold edges could produce duplicate edges.

4 years agoBMesh: use BM_face_loop_separate_multi for rip
Campbell Barton [Wed, 29 Apr 2015 20:23:01 +0000 (06:23 +1000)]
BMesh: use BM_face_loop_separate_multi for rip

Resolves bug over-splitting non-manifold connected edges.

4 years agoBMesh: add BM_face_loop_separate_multi
Campbell Barton [Wed, 29 Apr 2015 19:52:48 +0000 (05:52 +1000)]
BMesh: add BM_face_loop_separate_multi

New utility function to handle splitting off multiple loops from a face at once.

4 years agoBMesh: simplify/optimize loop splitting logic
Campbell Barton [Wed, 29 Apr 2015 16:37:21 +0000 (02:37 +1000)]
BMesh: simplify/optimize loop splitting logic

To split off a single loop, was splitting all fans off the vertex, then merging back together (except for one).

Now simply splits off one loop.

4 years agoBMesh: refactor edge-vert swapping into API call
Campbell Barton [Wed, 29 Apr 2015 16:25:32 +0000 (02:25 +1000)]
BMesh: refactor edge-vert swapping into API call

4 years agoBMesh: minor change to swap-vert api
Campbell Barton [Wed, 29 Apr 2015 15:45:20 +0000 (01:45 +1000)]
BMesh: minor change to swap-vert api

- assert if the verts not in the edge (all callers assume success)
- rename to bmesh_disk_vert_swap
- swap src/dst arg order.

4 years agoCycles: Fix wrong termination criteria in SSS volume stack update
Sergey Sharybin [Wed, 29 Apr 2015 20:20:17 +0000 (01:20 +0500)]
Cycles: Fix wrong termination criteria in SSS volume stack update

Another issue spotted with Thomas.

4 years agoCycles: Fix wrong order in object flags calculations
Sergey Sharybin [Wed, 29 Apr 2015 20:07:38 +0000 (01:07 +0500)]
Cycles: Fix wrong order in object flags calculations

Object flags are depending on bounding box which is only available after
mesh synchronization.

This was broken since 7fd4c44 which happened quite close to the release
and oddly enough was not sopped by anyone. Render test is coming for this.

Was spotted by Thomas Dinges while working on another patch.

4 years agoFix filename has "new render view" suffix reported by Jason van Gumster (Fweeb)
Dalai Felinto [Wed, 29 Apr 2015 19:35:19 +0000 (16:35 -0300)]
Fix filename has "new render view" suffix reported by Jason van Gumster (Fweeb)

Bug introduced in f8540d7fd5a47bc9d1d676d5aaaa0de379c71637

4 years agoFile browser - change thumbnails size with a slider
Bastien Montagne [Wed, 29 Apr 2015 19:25:34 +0000 (21:25 +0200)]
File browser - change thumbnails size with a slider

We can now scale from 32px up to 256px (default has been upgraded to 128px).
Thumbnails are now generated as 'large', i.e. 256px.

Previews are scaled up if necessary, unlike icons (for folders or files without preview images).

Note that .blend thumbnails themselves remain in 128px for now (they are embeded in .blend files,
not quite sure we want to make them four times bigger...).

Patch by DMS (Yaron Dames), with final edits by myself.

Reviewers: mont29

Subscribers: Severin, mont29

Differential Revision:

4 years agoFix T44132: Crash after open EXR format
Sergey Sharybin [Wed, 29 Apr 2015 18:44:57 +0000 (23:44 +0500)]
Fix T44132: Crash after open EXR format

Was own mistake on adding Alpha socket for Combined pass.

4 years agoOSX: fix T44533, system_bookmarks not showing utf standard ( umlauts etc. )
Jens Verwiebe [Wed, 29 Apr 2015 18:07:38 +0000 (20:07 +0200)]
OSX: fix T44533, system_bookmarks not showing utf standard ( umlauts etc. )

4 years agoFix typo using interp_v3_v3v3 over float[2] variables...
Bastien Montagne [Wed, 29 Apr 2015 17:38:19 +0000 (19:38 +0200)]
Fix typo using interp_v3_v3v3 over float[2] variables...

Found by asan! ;)

4 years agoMake sure integer calculation uses large precision
Antony Riakiotakis [Wed, 29 Apr 2015 16:49:38 +0000 (18:49 +0200)]
Make sure integer calculation uses large precision

4 years agoRemove unused stubs
Sergey Sharybin [Wed, 29 Apr 2015 16:47:30 +0000 (21:47 +0500)]
Remove unused stubs

Noticed by TristanPorteries in IRC.

4 years agoCleanup: make grumpy asan happy and do not use `new []` to allocate data freed by...
Bastien Montagne [Wed, 29 Apr 2015 15:46:03 +0000 (17:46 +0200)]
Cleanup: make grumpy asan happy and do not use `new []` to allocate data freed by `free()`.

Probably nothing crucial, but asan build would crash on that stupid glitch... Annoying.

4 years agofix for double call to update callback in uiTemplateIconView
Ines Almeida [Wed, 29 Apr 2015 15:25:54 +0000 (16:25 +0100)]
fix for double call to update callback in uiTemplateIconView

4 years agocleanup for uiTemplateIconView
Ines Almeida [Wed, 29 Apr 2015 15:22:04 +0000 (16:22 +0100)]
cleanup for uiTemplateIconView

4 years agoFix T44542: 'extend selection' editmode tool would select hidden elements.
Bastien Montagne [Wed, 29 Apr 2015 15:14:40 +0000 (17:14 +0200)]
Fix T44542: 'extend selection' editmode tool would select hidden elements.

4 years agoMulti-View cleanup: using RenderResult->rect* only for temporary RenderResults
Dalai Felinto [Wed, 29 Apr 2015 14:26:30 +0000 (11:26 -0300)]
Multi-View cleanup: using RenderResult->rect* only for temporary RenderResults

Originally I wanted to get rid of RenderResult->rect* entirely, but it's
convenient to have for temporary structs.

This patch makes sure they are used only when really needed, which
should help clearing the code out.

(they are needed when using RE_AcquireResultImage() - which produces a
RenderResult with no RenderView)

Reviewers: sergey

Differential Revision:

4 years agoDon't check against scene notifier twice
Antony Riakiotakis [Wed, 29 Apr 2015 14:28:20 +0000 (16:28 +0200)]
Don't check against scene notifier twice

4 years agoMulti-View: new util functions RE_RenderViewGetById() and RE_RenderViewGetByName()
Dalai Felinto [Wed, 29 Apr 2015 14:18:18 +0000 (11:18 -0300)]
Multi-View: new util functions RE_RenderViewGetById() and RE_RenderViewGetByName()

Both functions try to find a valid RenderView and if they can't they
fallback to the first RenderView of the RenderResult

4 years agoAdd macro BLI_SMALLSTACK_AS_TABLE
Campbell Barton [Wed, 29 Apr 2015 13:54:32 +0000 (23:54 +1000)]

Use for edge-split (a little less overhead compare to popping each item).

4 years agoCycles: Fix crashes when loading cache created with pre-leaf split builds
Sergey Sharybin [Wed, 29 Apr 2015 10:48:49 +0000 (15:48 +0500)]
Cycles: Fix crashes when loading cache created with pre-leaf split builds

4 years agoFix T44544: Cached BVH is broken since BVH leaf split
Sergey Sharybin [Wed, 29 Apr 2015 10:29:42 +0000 (15:29 +0500)]
Fix T44544: Cached BVH is broken since BVH leaf split

Still need to solve issues with reading old cache with new builds.

4 years agoUse size_t to calculate size of tiles for rendering
Antony Riakiotakis [Wed, 29 Apr 2015 10:30:25 +0000 (12:30 +0200)]
Use size_t to calculate size of tiles for rendering

4 years agoCorrect comment
Campbell Barton [Wed, 29 Apr 2015 10:16:40 +0000 (20:16 +1000)]
Correct comment

4 years agoBuildbot: Attempt to fix cudakernels target screwing up autotest cache
Sergey Sharybin [Wed, 29 Apr 2015 09:51:50 +0000 (14:51 +0500)]
Buildbot: Attempt to fix cudakernels target screwing up autotest cache

4 years agoBMesh: optimize edge split
Campbell Barton [Wed, 29 Apr 2015 02:48:06 +0000 (12:48 +1000)]
BMesh: optimize edge split

Avoid hashing edges when splitting into fans,
Instead, walk & split fans until they're all done, gives approx 40% speedup.

4 years agoBMesh: replace smallhash flag for checking doubles
Campbell Barton [Wed, 29 Apr 2015 02:25:20 +0000 (12:25 +1000)]
BMesh: replace smallhash flag for checking doubles

4 years agoBMesh: avoid over-counting vert-edges
Campbell Barton [Wed, 29 Apr 2015 00:25:59 +0000 (10:25 +1000)]
BMesh: avoid over-counting vert-edges

4 years agoCleanup: headers
Campbell Barton [Wed, 29 Apr 2015 09:39:58 +0000 (19:39 +1000)]
Cleanup: headers

4 years agoSCons: Ignore .svn folder when installing site-packages
Sergey Sharybin [Wed, 29 Apr 2015 09:29:42 +0000 (14:29 +0500)]
SCons: Ignore .svn folder when installing site-packages

4 years agoRevert commit with perspective depth offset, makes simple cases such as
Antony Riakiotakis [Wed, 29 Apr 2015 09:19:23 +0000 (11:19 +0200)]
Revert commit with perspective depth offset, makes simple cases such as
loopcuts on cube have zfighting.

4 years agoSCons: Fix wrong flags usage after recent stdc89 changes
Sergey Sharybin [Wed, 29 Apr 2015 08:57:02 +0000 (13:57 +0500)]
SCons: Fix wrong flags usage after recent stdc89 changes

CCFLAGS are used for both C and C++ compilers and one is better not to pass
C-related flags to C++ compiler. C-compiler flags are to be passed via CFLAGS

4 years agoBuildbot: Fix typo in error message
Sergey Sharybin [Wed, 29 Apr 2015 08:47:47 +0000 (13:47 +0500)]
Buildbot: Fix typo in error message

4 years agoRenderResult should have a valid view whenever possible
Dalai Felinto [Tue, 28 Apr 2015 20:36:44 +0000 (17:36 -0300)]
RenderResult should have a valid view whenever possible

4 years agoRewind fix for T44505 (leave in ifdef'd)
Campbell Barton [Tue, 28 Apr 2015 20:08:45 +0000 (06:08 +1000)]
Rewind fix for T44505 (leave in ifdef'd)

This makes selection fail in simple cases,
default cube subdiv 10x for eg.

4 years agoError in last commit
Campbell Barton [Tue, 28 Apr 2015 19:36:39 +0000 (05:36 +1000)]
Error in last commit

4 years agoStore bit-depth for reuse, replace loop with shift
Campbell Barton [Tue, 28 Apr 2015 19:30:30 +0000 (05:30 +1000)]
Store bit-depth for reuse, replace loop with shift

4 years agoCompilation error fixes for strict compiler flags
Sergey Sharybin [Tue, 28 Apr 2015 19:23:52 +0000 (00:23 +0500)]
Compilation error fixes for strict compiler flags

4 years agoFix ortho part of T44505
Antony Riakiotakis [Tue, 28 Apr 2015 18:41:37 +0000 (20:41 +0200)]
Fix ortho part of T44505

In this case we can calculate an offset without worrying about
perspective correction. Unfortunately if looking from a camera we still
have depth issues here. There's no really general case that can fix this
so I'm leaving this as is.

4 years agoFix T44404: freestyle crashes blender.
Tamito Kajiyama [Tue, 28 Apr 2015 14:18:32 +0000 (23:18 +0900)]
Fix T44404: freestyle crashes blender.

The reported crash case seems to be caused by freeing compiled Python
objects in a thread.  Now this issue is avoided by allocating a buffer to
store a Python script and using BPY_string_exec() to run the script.  This
makes it unnecessary to repeatedly create and destroy Text data blocks.

Many thanks to Campbell Barton for his help on the bug fix.

4 years agodisable ARRAY_SIZE fixed length check for Clang
Campbell Barton [Tue, 28 Apr 2015 15:36:14 +0000 (01:36 +1000)]
disable ARRAY_SIZE fixed length check for Clang

4 years agoAlleviate somewhat the issue of T44505
Antony Riakiotakis [Tue, 28 Apr 2015 15:18:32 +0000 (17:18 +0200)]
Alleviate somewhat the issue of T44505

Issue is zfighting with wire of mesh when parts of the mesh are close
together. We can make this slightly better by reducing the offset,
however this offset is calculated pre-perspective division and can vary
greatly with distance. Correct approach would be using polygon offset,
however we draw mesh wireframes as lines, (not polygons with polygon
mode line) so this approach will not work.
Alternatively, we could set an offset in a shader, however we don't have
code for that either.

4 years agoCurves: don't use 'charidx' for regular curves
Campbell Barton [Tue, 28 Apr 2015 14:19:34 +0000 (00:19 +1000)]
Curves: don't use 'charidx' for regular curves

Code attempted to sync them with materials,
but its not needed (and wasn't reliable).

4 years agoFix T44522: loop remapping between meshes when using 'nearest vert, best matching...
Bastien Montagne [Tue, 28 Apr 2015 13:57:11 +0000 (15:57 +0200)]
Fix T44522: loop remapping between meshes when using 'nearest vert, best matching normal' modes
would fail on coplanar faces (or smooth verts).

Loop remapping is really a tricky topic... For now, we enhance a bit more
our Frankenfunc by using distance between dest and source polygons as
fallback in case we have too much similar normals...

Probably not a perfect solution, but should be robust enough I hope.

One core question remains open though: do we want to stick to 'use only seams
to detect UV islands'? This makes things much simpler, but will obviously fail
in case of actual islands without matching seams. :/

4 years agoMinor cleanup in comments.
Bastien Montagne [Tue, 28 Apr 2015 12:15:27 +0000 (14:15 +0200)]
Minor cleanup in comments.

4 years agoFix T44516 grid lines obscuring selection outline.
Antony Riakiotakis [Tue, 28 Apr 2015 13:41:44 +0000 (15:41 +0200)]
Fix T44516 grid lines obscuring selection outline.

Was changed to draw after meshes without depth mask to make grid not
contribute to compositing effects. Now only draw it like this when we do
compositing (unfortunately can't have both).

4 years agoProject Paint: Add symmetry support
Campbell Barton [Tue, 28 Apr 2015 13:34:40 +0000 (23:34 +1000)]
Project Paint: Add symmetry support

- Access from symmetry panel (as with sculpt)
- Supports multiple axis at once.
- Supports all brush types including clone.

4 years agoCleanup: const correctness
Campbell Barton [Tue, 28 Apr 2015 11:17:48 +0000 (21:17 +1000)]
Cleanup: const correctness

4 years agoBLI_utildefines: add MEMSET_STRUCT_OFS macro
Campbell Barton [Tue, 28 Apr 2015 11:14:46 +0000 (21:14 +1000)]
BLI_utildefines: add MEMSET_STRUCT_OFS macro

4 years agoCleanup: Move Cycles volume stack update for subsurface into kernel_volume.h.
Thomas Dinges [Tue, 28 Apr 2015 09:11:39 +0000 (11:11 +0200)]
Cleanup: Move Cycles volume stack update for subsurface into kernel_volume.h.

4 years agoFix T44530 UV islands fail on subsurf after mirror modifier.
Antony Riakiotakis [Tue, 28 Apr 2015 09:12:47 +0000 (11:12 +0200)]
Fix T44530 UV islands fail on subsurf after mirror modifier.

Caused by own commit that changed island detection code. In the case of
modifiers we don't want to take winding information into account, but
left the code since there are use cases (like painting) which could use

4 years agoUse gnu89 for GCC
Campbell Barton [Tue, 28 Apr 2015 08:37:28 +0000 (18:37 +1000)]
Use gnu89 for GCC

GCC5 defaults to gnu11, which we may switch to later
but disable for now to prevent accidental use.

4 years agoFix for Clang type check
Campbell Barton [Tue, 28 Apr 2015 08:48:41 +0000 (18:48 +1000)]
Fix for Clang type check

4 years agoAdded missing update tag to recalculate data after permutating materials
Gaia Clary [Tue, 28 Apr 2015 08:52:19 +0000 (10:52 +0200)]
Added missing update tag to recalculate data after permutating materials

4 years agoCleanup: redundant checks
Campbell Barton [Tue, 28 Apr 2015 06:47:31 +0000 (16:47 +1000)]
Cleanup: redundant checks

4 years agoAdd material slot reorder buttons
Gaia Clary [Mon, 27 Apr 2015 21:24:56 +0000 (07:24 +1000)]
Add material slot reorder buttons

4 years agoCycles: Initialize portal variable directly, so we can avoid the one NULL check.
Thomas Dinges [Mon, 27 Apr 2015 21:12:53 +0000 (23:12 +0200)]
Cycles: Initialize portal variable directly, so we can avoid the one NULL check.

4 years agoMath Lib: add range_vn_u
Campbell Barton [Mon, 27 Apr 2015 20:54:17 +0000 (06:54 +1000)]
Math Lib: add range_vn_u

4 years agoCycles: Added support for light portals
Lukas Stockner [Mon, 27 Apr 2015 19:51:55 +0000 (00:51 +0500)]
Cycles: Added support for light portals

This patch adds support for light portals: objects that help sampling the
environment light, therefore improving convergence. Using them tor other
lights in a unidirectional pathtracer is virtually useless.

The sampling is done with the area-preserving code already used for area lamps.
MIS is used both for combination of different portals and for combining portal-
and envmap-sampling.

The direction of portals is considered, they aren't used if the sampling point
is behind them.

Reviewers: sergey, dingto, #cycles

Reviewed By: dingto, #cycles

Subscribers: Lapineige, nutel, jtheninja, dsisco11, januz, vitorbalbio, candreacchio, TARDISMaker, lichtwerk, ace_dragon, marcog, mib2berlin, Tunge, lopataasdf, lordodin, sergey, dingto

Differential Revision:

4 years agoCleanup: use strict flags
Campbell Barton [Mon, 27 Apr 2015 18:30:07 +0000 (04:30 +1000)]
Cleanup: use strict flags

4 years agoCycles: Use native saturate function for CUDA
Sergey Sharybin [Mon, 27 Apr 2015 19:13:03 +0000 (00:13 +0500)]
Cycles: Use native saturate function for CUDA

This more a workaround for CUDA optimizer which can't optimize clamp(x, 0, 1)
into a single instruction and uses 4 instructions instead.

Original patch by @lockal with own modification:

  Don't make changes outside of the kernel. They don't make any difference
  anyway and term saturate() has a bit different meaning outside of kernel.

This gives around 2% of speedup in Barcelona file, but in more complex shader
setups with lots of math nodes with clamping speedup could be much nicer.

Subscribers: dingto

Projects: #cycles

Differential Revision:

4 years agoBLI_array: add permute utility function
Campbell Barton [Mon, 27 Apr 2015 18:10:31 +0000 (04:10 +1000)]
BLI_array: add permute utility function

4 years agoUsual UI messages fixes...
Bastien Montagne [Mon, 27 Apr 2015 18:10:32 +0000 (20:10 +0200)]
Usual UI messages fixes...

4 years agoFix double-free on copied Text compiled py-code
Campbell Barton [Mon, 27 Apr 2015 14:10:34 +0000 (00:10 +1000)]
Fix double-free on copied Text compiled py-code

4 years agoFix T44464: Viewport mipmaps no longer toggle off
Campbell Barton [Mon, 27 Apr 2015 15:14:48 +0000 (01:14 +1000)]
Fix T44464: Viewport mipmaps no longer toggle off

4 years agoCleanup: use bool for mipmap args
Campbell Barton [Mon, 27 Apr 2015 15:03:28 +0000 (01:03 +1000)]
Cleanup: use bool for mipmap args

4 years agoCleanup: use mul_v3_mat3_m4v3
Campbell Barton [Mon, 27 Apr 2015 13:23:04 +0000 (23:23 +1000)]
Cleanup: use mul_v3_mat3_m4v3

4 years agoMulti-View: Code cleanup
Dalai Felinto [Mon, 27 Apr 2015 15:09:04 +0000 (12:09 -0300)]
Multi-View: Code cleanup

I finally put the time into understanding what was going on here.
Basically RE_AcquireResultImage() produces RenderResults without
RenderViews. That will be fine for now since I'm planning to refactor
RenderResult soon.

4 years agoMuti-View: viewport crashes when empty is active camera - reported over IRC by Campbe...
Dalai Felinto [Mon, 27 Apr 2015 13:39:09 +0000 (10:39 -0300)]
Muti-View: viewport crashes when empty is active camera - reported over IRC by Campbell Barton

4 years agoFix T44514 Crash under cycles rendering using F12, based on patch from Gaia Clary
Dalai Felinto [Mon, 27 Apr 2015 13:33:18 +0000 (10:33 -0300)]
Fix T44514 Crash under cycles rendering using F12, based on patch from Gaia Clary

4 years agoFix T44512: Cycles fails normal bake /w scale flip
Campbell Barton [Mon, 27 Apr 2015 12:43:01 +0000 (22:43 +1000)]
Fix T44512: Cycles fails normal bake /w scale flip

4 years agoMath Lib: add mul_v3_mat3_m4v3
Campbell Barton [Mon, 27 Apr 2015 11:50:38 +0000 (21:50 +1000)]
Math Lib: add mul_v3_mat3_m4v3

4 years agoFix T44011: Ruler/Knife/Loop-cut fail in quad-view
Campbell Barton [Mon, 27 Apr 2015 08:53:45 +0000 (18:53 +1000)]
Fix T44011: Ruler/Knife/Loop-cut fail in quad-view

This is a kind of sloppy-focus,
resolving long standing bug with loop-cut/knife/ruler /w quad-view.

Where activating a tool would lock onto one of quad-views,
especially problematic when activating from the toolbar or menus.

4 years agoPy API: expose operator runtime flags
Campbell Barton [Mon, 27 Apr 2015 08:45:48 +0000 (18:45 +1000)]
Py API: expose operator runtime flags

eg from operator invoke/execute:

4 years agoCleanup: rename GRAB_POINTER -> GRAB_CURSOR
Campbell Barton [Mon, 27 Apr 2015 08:44:27 +0000 (18:44 +1000)]

Term pointer is overloaded already.

4 years agoSculpt: symmetrize was leaving edge-tags dirty
Campbell Barton [Sun, 26 Apr 2015 17:35:16 +0000 (03:35 +1000)]
Sculpt: symmetrize was leaving edge-tags dirty

This is used for the edge-queue

4 years agoCorrect own error in recent sculpt changes
Campbell Barton [Sun, 26 Apr 2015 17:34:13 +0000 (03:34 +1000)]
Correct own error in recent sculpt changes

mixed up squared nonsquared length, also remove invalid verify check.

4 years agoUI: 'View Online Manual' Shortcut
Julian Eisel [Sun, 26 Apr 2015 15:17:51 +0000 (01:17 +1000)]
UI: 'View Online Manual' Shortcut

D1031, implement proposal T37478 to give easy access to the online manual.

Use Alt+F1 while hovering over a button/setting.

4 years agoBGE : addObject in python without reference object.
Porteries Tristan [Sun, 26 Apr 2015 14:29:43 +0000 (16:29 +0200)]
BGE : addObject in python without reference object.

Making the reference argument optional for the addObject function.

This allows to keep the rotation, scale and position of the original object.
To avoid layer problems with lights if the reference arguments is None, the new object have the same layer than the active layers in scene.

Reviewers: lordloki, moguri, hg1, sybren

Reviewed By: hg1, sybren

Subscribers: agoose77

Projects: #game_engine

Differential Revision:

4 years agoBMesh: increase subdiv smooth strength
Campbell Barton [Sun, 26 Apr 2015 11:20:25 +0000 (21:20 +1000)]
BMesh: increase subdiv smooth strength

Now 1.0 gives near spherical output

4 years agoCleanup: bmesh subdivide
Campbell Barton [Sun, 26 Apr 2015 10:32:08 +0000 (20:32 +1000)]
Cleanup: bmesh subdivide

var names were a bit cryptic, set return arg last.

4 years agoBMesh: subdiv smooth, use simpler even calculation
Campbell Barton [Sun, 26 Apr 2015 09:19:05 +0000 (19:19 +1000)]
BMesh: subdiv smooth, use simpler even calculation

Was checking all vertices adjacent faces,
now just compare the difference between normal angles.

Also default to inverse-square for loopcut-subdiv falloff.