blender.git
7 years agoFixed compilation error with msvc caused by recent libmv commit
Sergey Sharybin [Mon, 14 Nov 2011 10:47:34 +0000 (10:47 +0000)]
Fixed compilation error with msvc caused by recent libmv commit

7 years agomingw compiles again after OceanSim stuff
Joshua Leung [Mon, 14 Nov 2011 10:35:11 +0000 (10:35 +0000)]
mingw compiles again after OceanSim stuff

7 years agofix for shiftx/y for recently added VIEW3D_OT_camera_to_view_selected operator.
Campbell Barton [Mon, 14 Nov 2011 08:43:09 +0000 (08:43 +0000)]
fix for shiftx/y for recently added VIEW3D_OT_camera_to_view_selected operator.

7 years agoFix for [#29247] Incorrect black color with dynamic paint
Miika Hamalainen [Mon, 14 Nov 2011 08:18:04 +0000 (08:18 +0000)]
Fix for [#29247] Incorrect black color with dynamic paint

7 years agoFix for #29241, ocean sim modifier crash when removing all vertices.
Lukas Toenne [Mon, 14 Nov 2011 08:06:24 +0000 (08:06 +0000)]
Fix for #29241, ocean sim modifier crash when removing all vertices.

The object draw method did not check for valid data when mapping back to original face, constructive modifier on empty mesh does that.

7 years agomake ocean rna more consistent with existing rna names
Campbell Barton [Mon, 14 Nov 2011 07:18:32 +0000 (07:18 +0000)]
make ocean rna more consistent with existing rna names

7 years agomake rna naming for dynamic paint consistent with existing names & conventions.
Campbell Barton [Mon, 14 Nov 2011 07:07:59 +0000 (07:07 +0000)]
make rna naming for dynamic paint consistent with existing names & conventions.

7 years agoDPAINT_OT_output_toggle operator was using an index option for what was really a...
Campbell Barton [Mon, 14 Nov 2011 06:46:07 +0000 (06:46 +0000)]
DPAINT_OT_output_toggle operator was using an index option for what was really a toggle between 2 values, changed its index option to an enum.

if a value other than 1/0 was given it would use an uninitialized pointer too (compiler warning, review should pick up this stuff).

also renamed some RNA attrs:
 output_name --> output_name_a
 output_name2 --> output_name_b
 do_output1 --> use_output_a
 do_output2 --> use_output_b
 do_smudge --> use_smudge
 max_velocity --> velocity_max

7 years agoCamera tracking: interface cleanup and small buttons renaming
Sergey Sharybin [Mon, 14 Nov 2011 06:41:42 +0000 (06:41 +0000)]
Camera tracking: interface cleanup and small buttons renaming

- Move tracking-related constraints to own section in list
  Currently there are only two constraints, so can look a bit odd,
  but it'll be other constraints like "Object Solver" and so.
- Move motion-tracking parameters from 3D viewport Display panel
  to it's own panel.
- Get rid of "Bundle" in 3d viewport. It's quite obvious that it's
  a 3D representation of tracks is used in 3D viewport and it shouldn't
  be so confusing for artists now.
- Also get rid of "Bundle" in Follow Track constraint.
  Old files can change a bit because of changes in DNA.
- Also get rid of "Bundles" in operator which creates vertices cloud
  from 3D position of tracks.
- Rename "Principal Point" to "Optical Center" in the interface.

7 years agoCamera tracking: code cleanup
Sergey Sharybin [Mon, 14 Nov 2011 06:41:32 +0000 (06:41 +0000)]
Camera tracking: code cleanup

7 years agoAssorted camera tracker improvements
Sergey Sharybin [Mon, 14 Nov 2011 06:41:23 +0000 (06:41 +0000)]
Assorted camera tracker improvements

- Add support for refining the camera's intrinsic parameters
  during a solve. Currently, refining supports only the following
  combinations of intrinsic parameters:

    f
    f, cx, cy
    f, cx, cy, k1, k2
    f, k1
    f, k1, k2

  This is not the same as autocalibration, since the user must
  still make a reasonable initial guess about the focal length and
  other parameters, whereas true autocalibration would eliminate
  the need for the user specify intrinsic parameters at all.

  However, the solver works well with only rough guesses for the
  focal length, so perhaps full autocalibation is not that
  important.

  Adding support for the last two combinations, (f, k1) and (f,
  k1, k2) required changes to the library libmv depends on for
  bundle adjustment, SSBA. These changes should get ported
  upstream not just to libmv but to SSBA as well.

- Improved the region of convergence for bundle adjustment by
  increasing the number of Levenberg-Marquardt iterations from 50
  to 500. This way, the solver is able to crawl out of the bad
  local minima it gets stuck in when changing from, for example,
  bundling k1 and k2 to just k1 and resetting k2 to 0.

- Add several new region tracker implementations. A region tracker
  is a libmv concept, which refers to tracking a template image
  pattern through frames. The impact to end users is that tracking
  should "just work better". I am reserving a more detailed
  writeup, and maybe a paper, for later.

- Other libmv tweaks, such as detecting that a tracker is headed
  outside of the image bounds.

This includes several changes made directly to the libmv extern
code rather expecting to get those changes through normal libmv
channels, because I, the libmv BDFL, decided it was faster to work
on libmv directly in Blender, then later reverse-port the libmv
changes from Blender back into libmv trunk. The interesting part
is that I added a full Levenberg-Marquardt loop to the region
tracking code, which should lead to a more stable solutions. I
also added a hacky implementation of "Efficient Second-Order
Minimization" for tracking, which works nicely. A more detailed
quantitative evaluation will follow.

Original patch by Keir, cleaned a bit by myself.

7 years agofix uninitialized memory use when an object has modifiers but no ocean modifier.
Campbell Barton [Mon, 14 Nov 2011 06:11:40 +0000 (06:11 +0000)]
fix uninitialized memory use when an object has modifiers but no ocean modifier.

7 years agoOcean baking was using uninitialized memory, but further investigation it was calcula...
Campbell Barton [Mon, 14 Nov 2011 05:55:50 +0000 (05:55 +0000)]
Ocean baking was using uninitialized memory, but further investigation it was calculating foam values when they were not used.

avoid calculating foam and allocating memory when its not needed.

7 years agoFix #29245: BuilderBot can't build anymore
Sergey Sharybin [Mon, 14 Nov 2011 05:52:06 +0000 (05:52 +0000)]
Fix #29245: BuilderBot can't build anymore

Buildbot is using different settings for blender and player,
enable fftw library in rules for player now.

7 years agoVIEW3D_OT_camera_to_view_selected operator to move the camera to frame
Campbell Barton [Mon, 14 Nov 2011 03:54:23 +0000 (03:54 +0000)]
VIEW3D_OT_camera_to_view_selected operator to move the camera to frame
all selected, renderable objects.

7 years agoMoving all node angle-type values to radians. This also fixes [#29151] rotate node...
Bastien Montagne [Sun, 13 Nov 2011 18:03:27 +0000 (18:03 +0000)]
Moving all node angle-type values to radians. This also fixes [#29151] rotate node wrong input (mixing up radians and degrees).

Warning!

Angles in nodes have just been moved to consistant Radians values (ANGLE subtype of RNA Float property). You will still see them as degrees in the GUI, though, unless you chose otherwise in Scene properties, Units panel.

Conversion from degrees to radians for old files is obviously done at loading time, but if you use a mixed pipeline of trunk and releases, be carefull!

Loading a 2.60.4 file (or higher) into any previous version of Blender, your angles in nodes will have odd values (well, radians interpreted as degrees)!

And if you save such file in a pre-2.60.4 version, the angle node values will be converted again when loaded in Blender 2.60.4 or higher...

This affects following nodes:
* Compo: Rotate, Defocus, ChromaMatte, Glare and DirectionalBlur
* Shader: Mapping
And all future code using the TexMapping struct’s rotation part (its rot memember is now in radians).

7 years agoFix #29240: multi-res bake broken in 2.60
Sergey Sharybin [Sun, 13 Nov 2011 17:45:42 +0000 (17:45 +0000)]
Fix #29240: multi-res bake broken in 2.60

Use preview subdivision level even when in sculpt mode. Makes more sense
than baking against sculpting subdivision level.

7 years agoCamera solving: fixed incorrect warnings about failure of solving some frames
Sergey Sharybin [Sun, 13 Nov 2011 17:16:04 +0000 (17:16 +0000)]
Camera solving: fixed incorrect warnings about failure of solving some frames

Error was caused y not very accurate calculating which frames should be solved.

7 years agoQuiting a warning in Ocean code (commenting out var).
Bastien Montagne [Sun, 13 Nov 2011 16:38:14 +0000 (16:38 +0000)]
Quiting a warning in Ocean code (commenting out var).

7 years agonew math utility function isect_plane_plane_v3
Campbell Barton [Sun, 13 Nov 2011 16:28:52 +0000 (16:28 +0000)]
new math utility function isect_plane_plane_v3

7 years agoRelease cycle:
Thomas Dinges [Sun, 13 Nov 2011 16:24:15 +0000 (16:24 +0000)]
Release cycle:
Moving on to BCon 3: beta

7 years agoremove double promotions and some formatting edits (tabs & spaces mixed)
Campbell Barton [Sun, 13 Nov 2011 16:10:01 +0000 (16:10 +0000)]
remove double promotions and some formatting edits (tabs & spaces mixed)

7 years agoremove Id:'s from shader node headers
Campbell Barton [Sun, 13 Nov 2011 16:08:01 +0000 (16:08 +0000)]
remove Id:'s from shader node headers

7 years agoMake Ocean modifier compile again! Also removed it’s $ID$ heading stuff.
Bastien Montagne [Sun, 13 Nov 2011 15:41:40 +0000 (15:41 +0000)]
Make Ocean modifier compile again! Also removed it’s $ID$ heading stuff.

7 years agoensure that the path and directory are joined correctly for ocean cache (assumed...
Campbell Barton [Sun, 13 Nov 2011 15:17:24 +0000 (15:17 +0000)]
ensure that the path and directory are joined correctly for ocean cache (assumed path ended with a '/')

7 years agoScons_buildsystem: add WITH_BF_OCEANSIM = True to all configs
Jens Verwiebe [Sun, 13 Nov 2011 15:10:54 +0000 (15:10 +0000)]
Scons_buildsystem: add WITH_BF_OCEANSIM = True to all configs

7 years agoOcean Sim: sort alphabetically in modifier list, fix python error trying to
Brecht Van Lommel [Sun, 13 Nov 2011 14:54:11 +0000 (14:54 +0000)]
Ocean Sim: sort alphabetically in modifier list, fix python error trying to
show "Built without OceanSim modifier" message.

7 years agoadd the ocean modifier to bpath visitor
Campbell Barton [Sun, 13 Nov 2011 14:50:19 +0000 (14:50 +0000)]
add the ocean modifier to bpath visitor

7 years agoStub data in Ocean modifier when disabled, C standard does not allow empty structs.
Lukas Toenne [Sun, 13 Nov 2011 14:39:10 +0000 (14:39 +0000)]
Stub data in Ocean modifier when disabled, C standard does not allow empty structs.

7 years agopep8 edits
Campbell Barton [Sun, 13 Nov 2011 14:38:00 +0000 (14:38 +0000)]
pep8 edits

7 years agoallow passing vars to cmake from our makefile
Campbell Barton [Sun, 13 Nov 2011 14:29:17 +0000 (14:29 +0000)]
allow passing vars to cmake from our makefile

7 years agorename cmake WITH_OCEANSIM --> WITH_MOD_OCEANSIM and tag CYCLES_CUDA_BINARIES_ARCH...
Campbell Barton [Sun, 13 Nov 2011 14:16:43 +0000 (14:16 +0000)]
rename cmake WITH_OCEANSIM --> WITH_MOD_OCEANSIM and tag CYCLES_CUDA_BINARIES_ARCH as advanced.

7 years agoadd missing break in ocean_texture switch statement & quiet some warnings.
Campbell Barton [Sun, 13 Nov 2011 14:11:02 +0000 (14:11 +0000)]
add missing break in ocean_texture switch statement & quiet some warnings.

7 years agofix another buffer overrun error, also replace BLI_snprintf with BLI_strncpy since...
Campbell Barton [Sun, 13 Nov 2011 13:56:40 +0000 (13:56 +0000)]
fix another buffer overrun error, also replace BLI_snprintf with BLI_strncpy since no formatting is needed in this case.

7 years agofix for buffer overrun when grease pencil drawing in the clip editor.
Campbell Barton [Sun, 13 Nov 2011 13:52:15 +0000 (13:52 +0000)]
fix for buffer overrun when grease pencil drawing in the clip editor.

7 years agoOSX: go back to former zoombehaviour with magicmouse in Lion, no time to recode to...
Jens Verwiebe [Sun, 13 Nov 2011 13:25:22 +0000 (13:25 +0000)]
OSX: go back to former zoombehaviour with magicmouse in Lion, no time to recode to NSTouch in the near time

7 years agoFix #29041: parenting problem with tree IK for iTaSC and iksolver, where it
Brecht Van Lommel [Sun, 13 Nov 2011 13:08:15 +0000 (13:08 +0000)]
Fix #29041: parenting problem with tree IK for iTaSC and iksolver, where it
would use the wrong bone as parent on brancing. Patch by Juha Maki-Kanto.

7 years agoOcean Modifer:
Thomas Dinges [Sun, 13 Nov 2011 13:04:46 +0000 (13:04 +0000)]
Ocean Modifer:
* Reuse the placeholder "eModifierType_EmptySlot" for the new Ocean modifier.

WARNING:
DO NOT use Ocean modifier in trunk with a revision below this commit! Files won't load then!!

7 years agoOcean Sim:
Thomas Dinges [Sun, 13 Nov 2011 12:45:47 +0000 (12:45 +0000)]
Ocean Sim:
* Changed the user interface for the Ocean modifier, to use less space and look better.
* Changed rna name cachepath to filepath for consistency (fluid cache path also uses "filepath")

7 years agoFix #29048: iTaSC solver crash on certain compilers/platforms, due to memory
Brecht Van Lommel [Sun, 13 Nov 2011 12:25:14 +0000 (12:25 +0000)]
Fix #29048: iTaSC solver crash on certain compilers/platforms, due to memory
alignment issues with Eigen. Patch by Tobias Oelgarte.

7 years agoOcean Sim modifier patch
Lukas Toenne [Sun, 13 Nov 2011 12:17:27 +0000 (12:17 +0000)]
Ocean Sim modifier patch
by Matt Ebb, Hamed Zaghaghi

This adds a new Modifier "Ocean" to simulate large-scale wave motion.
Details can be found in the wiki documentation [1], the project homepage [2] and the patch tracker [3]

The modifier is disabled by default for now. To enable it, the WITH_OCEANSIM (cmake) / WITH_BF_OCEANSIM (scons) flags have to be set. The code depends on fftw3, so this also has to be enabled.

[1]
http://wiki.blender.org/index.php/Doc:2.6/Manual/Modifiers/Simulation/Ocean

[2]
http://www.savetheoceansim.com

[3]
http://projects.blender.org/tracker/?group_id=9&atid=127&func=detail&aid=28338

7 years agoCycles:
Thomas Dinges [Sun, 13 Nov 2011 11:40:35 +0000 (11:40 +0000)]
Cycles:
* Typo fix, patch by David on the mailing list.

7 years agoSpeedup of multires baker
Sergey Sharybin [Sun, 13 Nov 2011 10:48:46 +0000 (10:48 +0000)]
Speedup of multires baker

Issue was caused by how CCGDM handles ORIGINDEX cystom layer:
It runs cycle through all faces to fill origindex array on each
call of dm->getFaceDataArray(dm, CD_ORIGINDEX)

Solved issue by obtaining origindex array once on baker data initialization and
using this stored array when interpolating multires grid data.

7 years agoFix crash when multires-baking in sculpt mode when sculpt mode is active and sculpt...
Sergey Sharybin [Sun, 13 Nov 2011 10:09:34 +0000 (10:09 +0000)]
Fix crash when multires-baking in sculpt mode when sculpt mode is active and sculpt level is set to 0.

7 years agoCycles Addon:
Thomas Dinges [Sun, 13 Nov 2011 10:05:07 +0000 (10:05 +0000)]
Cycles Addon:
* Added URL to wiki
* Marked as official
* api/blender bump

7 years agoformatting edits only - no functional changes
Campbell Barton [Sun, 13 Nov 2011 09:38:53 +0000 (09:38 +0000)]
formatting edits only - no functional changes

7 years agoinclude invalid type name in mathutils error messages.
Campbell Barton [Sun, 13 Nov 2011 09:20:04 +0000 (09:20 +0000)]
include invalid type name in mathutils error messages.

7 years agoOpenCL/Nvidia:
Thomas Dinges [Sat, 12 Nov 2011 22:22:00 +0000 (22:22 +0000)]
OpenCL/Nvidia:
* Enable OpenCL Full Shading on NVIDIA cards.

Notes:
It makes not much sense to use OpenCL on a nVidia card (as it is slower compared to CUDA), but as OpenCL comes without dependencies, it's an good alternative if you don't want to install the CUDA toolkit or the build comes without CUDA kernels.

7 years agoCycles:
Brecht Van Lommel [Sat, 12 Nov 2011 14:29:52 +0000 (14:29 +0000)]
Cycles:
* Fix object scaling update issue with interactive rendering + static BVH.
* Fix negative scaling issue with static BVH.
* Fix #29217: excessive fireflies in first sample.

7 years agomathutils.geometry.distance_point_to_plane(pt, plane_co, plane_no) - utility function...
Campbell Barton [Sat, 12 Nov 2011 10:06:56 +0000 (10:06 +0000)]
mathutils.geometry.distance_point_to_plane(pt, plane_co, plane_no) - utility function, BLI math version too.

7 years agoBugfix [#29212] Select before/after current frame deselects all channels in
Joshua Leung [Sat, 12 Nov 2011 05:00:24 +0000 (05:00 +0000)]
Bugfix [#29212] Select before/after current frame deselects all channels in
graph editor

No need to deselect all channels when doing the select left/right operator,
since we don't replace the selection afterwards with anything useful. This would
be a problem when the "Only Show Selected" option is enabled (as it is by
default), as this results in the curves being worked on disappearing from view.

F-Curves though will still get deselected as a result of the code to deselect
keyframes. Will need to review whether this is still a good idea (original
intention was so that only curves with keyframes selected would remain
selected).

7 years agoBugfix [#29125] Motion paths odd behaviour
Joshua Leung [Sat, 12 Nov 2011 04:40:53 +0000 (04:40 +0000)]
Bugfix [#29125] Motion paths odd behaviour

* Made it impossible to try to calculate/create new motion paths lasting 0
frames (i.e. 250 to 250) since we cannot allocate a zero-length array for these.
Start frame can now be at most end-frame - 1, and end frame at least start frame
+ 1

* If an invalid configuration does occur, warnings/reports will now be issued in
response to this instead of silently failing (as per this bugreport).

7 years agoHack-fix for crash-by-assertion on mingw builds on startup if the CUDA driver on
Joshua Leung [Sat, 12 Nov 2011 03:59:45 +0000 (03:59 +0000)]
Hack-fix for crash-by-assertion on mingw builds on startup if the CUDA driver on
the computer is too old (and probably can't be upgraded to a version that does)
to support the cuDriverGetVersion method.

7 years agoCycles: more linux build system tweaks, attempting to avoid using boost
Brecht Van Lommel [Fri, 11 Nov 2011 20:57:00 +0000 (20:57 +0000)]
Cycles: more linux build system tweaks, attempting to avoid using boost
system libraries by default if we have a lib/ directory.

7 years agoCycles: fix wrong openimageio path for lib/ directory.
Brecht Van Lommel [Fri, 11 Nov 2011 19:20:52 +0000 (19:20 +0000)]
Cycles: fix wrong openimageio path for lib/ directory.

7 years agoFix #29207: cancelling volume precache at high resolution was too slow, added
Brecht Van Lommel [Fri, 11 Nov 2011 16:01:13 +0000 (16:01 +0000)]
Fix #29207: cancelling volume precache at high resolution was too slow, added
more test_break calls.

7 years agoCMake/Scons: automatically use ffmpeg, boost and openimageio from the svn lib/
Brecht Van Lommel [Fri, 11 Nov 2011 14:52:24 +0000 (14:52 +0000)]
CMake/Scons: automatically use ffmpeg, boost and openimageio from the svn lib/
directory if available, which should make building with ffmpeg and cycles easier.

If this folder is not available or another path is specified in user-config.py
or CMakeCache.txt, this should have no effect. Wiki build instructions now show
how to use these.
http://wiki.blender.org/index.php/Dev:2.5/Doc/Building_Blender/Linux

7 years agocorrect indentation and some whitespace edits (no functional changes)
Campbell Barton [Fri, 11 Nov 2011 13:09:14 +0000 (13:09 +0000)]
correct indentation and some whitespace edits (no functional changes)

7 years agoquiet -Wdouble-promotion warnings
Campbell Barton [Fri, 11 Nov 2011 12:00:08 +0000 (12:00 +0000)]
quiet -Wdouble-promotion warnings

7 years agoFix #29220 & #29221: missing cycles updates when switching scenes.
Brecht Van Lommel [Fri, 11 Nov 2011 11:42:29 +0000 (11:42 +0000)]
Fix #29220 & #29221: missing cycles updates when switching scenes.

7 years agoDynamic Paint:
Miika Hamalainen [Fri, 11 Nov 2011 10:46:26 +0000 (10:46 +0000)]
Dynamic Paint:
* Renamed "Sharp" proximity falloff to "Constant".
* Added a new "Negate Volume" option for "Volume + Proximity" brush.
* Possible fix for random particle clipping errors.

7 years agofix [#29203] Camera can still move even when transforms locked
Campbell Barton [Fri, 11 Nov 2011 05:34:07 +0000 (05:34 +0000)]
fix [#29203] Camera can still move even when transforms locked

added object_tfm_protected_backup, object_tfm_protected_restore so its easier to transform the object and respect protected channels (otherwise you need checks everywhere for each channel which is verbose).

7 years agopep/style edits & quiet some warnings
Campbell Barton [Fri, 11 Nov 2011 03:28:46 +0000 (03:28 +0000)]
pep/style edits & quiet some warnings

7 years agoBugfix [#29210] ChildOf constraint Set Inverse
Joshua Leung [Fri, 11 Nov 2011 00:44:48 +0000 (00:44 +0000)]
Bugfix [#29210] ChildOf constraint Set Inverse

For Objects, Set Inverse was unreliable as the constraint didn't get reset
properly before being calculated

7 years agoDynamic Paint UI:
Thomas Dinges [Thu, 10 Nov 2011 20:31:50 +0000 (20:31 +0000)]
Dynamic Paint UI:
* Code cleanup, changed some splits to row's, much nicer when you only use 2 properties in a row.
* if > elif
* Removed some duplicated code for an if/elif/else check

Did some minor separator() changes too.

7 years ago2.6 Physics UI Files:
Thomas Dinges [Thu, 10 Nov 2011 19:10:23 +0000 (19:10 +0000)]
2.6 Physics UI Files:
* Remove redundant check for md, this is already done in the super class poll.

7 years agoSVN maintenance.
Guillermo S. Romero [Thu, 10 Nov 2011 17:43:37 +0000 (17:43 +0000)]
SVN maintenance.

7 years agoMark unused variable.
Sergey Sharybin [Thu, 10 Nov 2011 17:19:12 +0000 (17:19 +0000)]
Mark unused variable.

7 years agoCycles: fix missing particle panels and show emitter option not being respected.
Brecht Van Lommel [Thu, 10 Nov 2011 16:02:46 +0000 (16:02 +0000)]
Cycles: fix missing particle panels and show emitter option not being respected.

7 years agoCycles: attempt to fix build issue with mingw.
Brecht Van Lommel [Thu, 10 Nov 2011 14:53:56 +0000 (14:53 +0000)]
Cycles: attempt to fix build issue with mingw.

7 years agoCycles:
Brecht Van Lommel [Thu, 10 Nov 2011 14:32:16 +0000 (14:32 +0000)]
Cycles:
* Fix excessive fireflies in Velvet BSDF (patch by David).
* Disable some unused SSE code
* Remove RTTI disabling flags for now, this is giving some compile issues and
  was only needed of OSL which we're not using yet.

7 years agoCycles: attempted fix for error building with sse/mtune optimization flags.
Brecht Van Lommel [Thu, 10 Nov 2011 14:13:32 +0000 (14:13 +0000)]
Cycles: attempted fix for error building with sse/mtune optimization flags.

7 years agoCycles: fix crash in image assign/remove in image editor.
Brecht Van Lommel [Thu, 10 Nov 2011 13:39:25 +0000 (13:39 +0000)]
Cycles: fix crash in image assign/remove in image editor.

7 years agoNodes: fix crash loading old files that have node sockets changed, the version
Brecht Van Lommel [Thu, 10 Nov 2011 13:38:55 +0000 (13:38 +0000)]
Nodes: fix crash loading old files that have node sockets changed, the version
patch for the default value storage needs to be done before lib linking,
otherwise this can crash. Fixes problem opening some old cycles files.

7 years agoFix [#29214] build modifier calculates wrong fraction of faces/verts/edges to draw
Andrew Wiggin [Thu, 10 Nov 2011 13:28:58 +0000 (13:28 +0000)]
Fix [#29214] build modifier calculates wrong fraction of faces/verts/edges to draw

7 years agoCycles: when material/world/lamp doesn't use nodes, show color in the UI.
Brecht Van Lommel [Thu, 10 Nov 2011 13:00:53 +0000 (13:00 +0000)]
Cycles: when material/world/lamp doesn't use nodes, show color in the UI.

7 years agoDynamic Paint:
Miika Hamalainen [Thu, 10 Nov 2011 13:00:27 +0000 (13:00 +0000)]
Dynamic Paint:
* Added BKE_dynamicpaint.h to CMake list.
* Changed some unsigned integers back to signed, as OpenMP can't iterate unsigned values anyway.
* Fixed more warnings.

7 years agoCycles: many of the new nodes were not showing properties/buttons in nodes in
Brecht Van Lommel [Thu, 10 Nov 2011 12:53:40 +0000 (12:53 +0000)]
Cycles: many of the new nodes were not showing properties/buttons in nodes in
the node editor.

7 years agoCycles:
Brecht Van Lommel [Thu, 10 Nov 2011 12:52:17 +0000 (12:52 +0000)]
Cycles:
* Add back option to bundle CUDA kernel binaries with builds.
* Disable runtime CUDA kernel compilation on Windows, couldn't get this working,
  since it seems to depend on visual studio being installed, even though for
  this particular case it shouldn't be needed. CMake only at the moment.
* Runtime compilation on linux/mac should now work if nvcc is not installed in
  the default location, but available in PATH.

7 years agoquiet some warnings from recent merge
Campbell Barton [Thu, 10 Nov 2011 12:28:26 +0000 (12:28 +0000)]
quiet some warnings from recent merge

7 years agoFixed possible issue drawing undistorted clip in clip editor.
Sergey Sharybin [Thu, 10 Nov 2011 12:15:09 +0000 (12:15 +0000)]
Fixed possible issue drawing undistorted clip in clip editor.
Check for marker's visibility was a bit odd.

7 years agoRestored GPL header of NOD_composite.h (funny s&r ?!?).
Bastien Montagne [Thu, 10 Nov 2011 11:33:57 +0000 (11:33 +0000)]
Restored GPL header of NOD_composite.h (funny s&r ?!?).

7 years agoWhitespace tweaks
Joshua Leung [Thu, 10 Nov 2011 10:34:26 +0000 (10:34 +0000)]
Whitespace tweaks

7 years agoDynamic Paint merge:
Miika Hamalainen [Thu, 10 Nov 2011 10:24:34 +0000 (10:24 +0000)]
Dynamic Paint merge:
Commit Dynamic Paint from "soc-2011-carrot" branch into trunk.

End-user documentation:
http://wiki.blender.org/index.php/Doc:2.6/Manual/Modifiers/Simulation/Dynamic_Paint

GSoC wiki page:
http://wiki.blender.org/index.php/User:MiikaH/GSoC-2011-DynamicPaint

7 years agoWorkaround #29205: Crash while using Edge Split modifier (while rotating vertices...
Sergey Sharybin [Thu, 10 Nov 2011 08:29:22 +0000 (08:29 +0000)]
Workaround #29205: Crash while using Edge Split modifier (while rotating vertices in edit mode with edge split viewable)

This bug is caused by exactly the same reason as #26316: differences in how new vertices/edges
are getting calculated first and how they're adding later. In some cases extra vertices are
creating which weren't counted before.

This patch prevents crash in such situations, but result mesh can be a bit wrong.
This should work fine in bmesh, so think it's acceptable to have such workaround
before actual fix coming with bmesh.

7 years ago- possible fix for vertex paint 'fill' crash.
Campbell Barton [Thu, 10 Nov 2011 06:06:35 +0000 (06:06 +0000)]
- possible fix for vertex paint 'fill' crash.
- remove unneeded call to defvert_find_index() when weight painting.

7 years agoadd some missing headers to cmake, also add some files as comments since it seems...
Campbell Barton [Thu, 10 Nov 2011 06:05:22 +0000 (06:05 +0000)]
add some missing headers to cmake, also add some files as comments since it seems they should be added but evidently work fine without.

7 years agoquiet warning
Campbell Barton [Thu, 10 Nov 2011 04:03:08 +0000 (04:03 +0000)]
quiet warning

7 years agopresets now work from non-redo popups,
Campbell Barton [Thu, 10 Nov 2011 03:44:50 +0000 (03:44 +0000)]
presets now work from non-redo popups,

ended up having to add a new pointer into the uiBlock (which I'd rather have avoided), but setting the uiLayoutSetContextPointer(..) was complicated to properly use for submenus and popus.

7 years agoWhile reviewing this part of the code for a Blender Proceedings article, I noticed...
Jason Hays [Thu, 10 Nov 2011 02:09:58 +0000 (02:09 +0000)]
While reviewing this part of the code for a Blender Proceedings article, I noticed that the lock enforcer was taking several improper steps to redistribute evenly with multi-paint.

Beforehand, a few problems were in view: some of the indexing was changed towards the end to use the more efficient stack structure, but needed to use the correct def group indices.

Also, the designated selected group would use its own value to acquire the standard to base change distributed to the others.

Lastly, the total_change was used rather than -left_overs in the formula to compute the new designated weight within the means of the locks' allowed changes.

Now, while maintaining the ratios of the selection, it correctly returns left over change that could not be redistributed to the unlocked groups.

7 years agoOne library from OpenEXR is needed for oiio. Hope buildbot is working fine now.
Sergey Sharybin [Wed, 9 Nov 2011 19:25:16 +0000 (19:25 +0000)]
One library from OpenEXR is needed for oiio. Hope buildbot is working fine now.

7 years agoDisable libmv for blenderplayer rules used by buildbot
Sergey Sharybin [Wed, 9 Nov 2011 19:04:17 +0000 (19:04 +0000)]
Disable libmv for blenderplayer rules used by buildbot

7 years ago- Added options to link staticly against oiio/boost
Sergey Sharybin [Wed, 9 Nov 2011 19:02:42 +0000 (19:02 +0000)]
- Added options to link staticly against oiio/boost
- Also changed buildbot settings so now linking should happe staticly

7 years agoFixes for MinGW cmake debug build:
Antony Riakiotakis [Wed, 9 Nov 2011 18:50:48 +0000 (18:50 +0000)]
Fixes for MinGW cmake debug build:

* OpenEXR doesn't need debug suffix
* Fix libmv template issue when linking by removing duplicate libmv inclusion. I wonder how this never turned up in release builds as well.

Important: Since OpenImageIO went into trunk, OpenEXR, possibly along with other image libraries will need to be turned on too because OIIO depends on them.

7 years agoFix ${} var usage. Was missing BF_OIIO and BF_BOOST init.
Nathan Letwory [Wed, 9 Nov 2011 17:22:55 +0000 (17:22 +0000)]
Fix ${} var usage. Was missing BF_OIIO and BF_BOOST init.

Noticed still missing: OSX file copying (cycles .py files, kernel files, licenses).

7 years agoEnable Cycles for buildbot
Sergey Sharybin [Wed, 9 Nov 2011 17:18:37 +0000 (17:18 +0000)]
Enable Cycles for buildbot

7 years agoMerge with trunk r41701 soc-2011-carrot
Miika Hamalainen [Wed, 9 Nov 2011 15:46:53 +0000 (15:46 +0000)]
Merge with trunk r41701

7 years agoFixed/restored translation of node classes (in foreach_nodeclass funcs)
Bastien Montagne [Wed, 9 Nov 2011 15:00:11 +0000 (15:00 +0000)]
Fixed/restored translation of node classes (in foreach_nodeclass funcs)

7 years agoMoving i18n code to BLF_translation.h
Bastien Montagne [Wed, 9 Nov 2011 14:13:17 +0000 (14:13 +0000)]
Moving i18n code to BLF_translation.h

This way, we will be able to use it in non-UI parts of the code too.

7 years agoNo tabs in scons! Sorry for that..
Antony Riakiotakis [Wed, 9 Nov 2011 14:00:24 +0000 (14:00 +0000)]
No tabs in scons! Sorry for that..