5 years agoGit test commit, this time from Windows.
Thomas Dinges [Sat, 16 Nov 2013 15:13:18 +0000 (16:13 +0100)]
Git test commit, this time from Windows.

5 years agoAttempt to fix compilation error on windows
Sergey Sharybin [Sat, 16 Nov 2013 14:39:36 +0000 (20:39 +0600)]
Attempt to fix compilation error on windows

Seems simply need to add .exe to msgfmt executable when
adding custom command to SCons.

Don't have windows by hand atm, so can not test for sure.

5 years agoMore fix for this stupid uppercase-to-lowercase convertion code... Thanks to Lockal...
Bastien Montagne [Sat, 16 Nov 2013 14:31:00 +0000 (15:31 +0100)]
More fix for this stupid uppercase-to-lowercase convertion code... Thanks to Lockal for noting this!

5 years agoRemove nested check for .git folder exists in buildinfo.cmake
Sergey Sharybin [Sat, 16 Nov 2013 13:40:26 +0000 (19:40 +0600)]
Remove nested check for .git folder exists in buildinfo.cmake

Title says it all :)

Pointer by IRIE Shinsuke in the ML, thanks!

5 years agoFix T37476 Bevel modifier got weight from wrong edge
Howard Trickey [Sat, 16 Nov 2013 13:07:03 +0000 (08:07 -0500)]
Fix T37476 Bevel modifier got weight from wrong edge

The bevel modifier with 'weight' activated was reading
the weights from the wrong edges.

5 years agoAnd here is the first git (testing) commit by me!
Ton Roosendaal [Sat, 16 Nov 2013 10:57:40 +0000 (11:57 +0100)]
And here is the first git (testing) commit by me!
Welcome to a brave new GPL3 world... well if commit works :)

5 years agoFix T37442: Disabled uiList would "freeze" Blender when trying to drag-resize it.
mont29 [Sat, 16 Nov 2013 10:30:48 +0000 (11:30 +0100)]
Fix T37442: Disabled uiList would "freeze" Blender when trying to drag-resize it.

For now, simply disable drag-resize for disabled uiLists!

5 years agoTest
Thomas Dinges [Sat, 16 Nov 2013 09:59:08 +0000 (10:59 +0100)]

5 years agoThe tittle for "Maximum Draw Type" had capitalization inconsistent with the rest...
Daniel Salazar [Sat, 16 Nov 2013 08:50:01 +0000 (02:50 -0600)]
The tittle for "Maximum Draw Type" had capitalization inconsistent with the rest of Blender

5 years agorangetree: support for building on freebsd-10
Campbell Barton [Sat, 16 Nov 2013 07:31:13 +0000 (18:31 +1100)]
rangetree: support for building on freebsd-10

5 years agoHelp menu: report a bug now links to the new tracker at
Brecht Van Lommel [Sat, 16 Nov 2013 02:26:04 +0000 (03:26 +0100)]
Help menu: report a bug now links to the new tracker at

There will be a redirect from the website as well, but
might as well update the link here.

5 years agoAddons: remove tracker URL from addon template, as this is no longer used,
Brecht Van Lömmel [Sat, 16 Nov 2013 02:08:51 +0000 (03:08 +0100)]
Addons: remove tracker URL from addon template, as this is no longer used,
and link to the new Addons bug reporting page.

5 years agoOSX/cmake: rollback ca5bbe0e419b04d16ee7eb12fa7eaffa0266815a cause OSX versions can...
Jens Verwiebe [Fri, 15 Nov 2013 22:20:16 +0000 (23:20 +0100)]
OSX/cmake: rollback ca5bbe0e419b04d16ee7eb12fa7eaffa0266815a cause OSX versions can have not subversion thus sw_vers -productVersion is not usable for sdk setting

5 years agoCode cleanup, cont. A small omission here.
Antony Riakiotakis [Fri, 15 Nov 2013 21:04:18 +0000 (23:04 +0200)]
Code cleanup, cont. A small omission here.

5 years agoCode cleanup: Use different redraw options for sculpt mask operators.
Antony Riakiotakis [Fri, 15 Nov 2013 21:00:15 +0000 (23:00 +0200)]
Code cleanup: Use different redraw options for sculpt mask operators.

Current redraw options also did an unnecessary normal recalculation on
updated nodes.

Also, for the box and lasso mask only push an undo node if any vertex
has actually been influenced.

5 years agoSupport for symmetrical box masking in sculpt mode.
Antony Riakiotakis [Fri, 15 Nov 2013 20:20:51 +0000 (22:20 +0200)]
Support for symmetrical box masking in sculpt mode.

Now box masking will take the symmetry options into account.

5 years agoOSX/cmake: use CMAKE_XCODE_ATTRIBUTE rather than CMAKE_OSX_SYSROOT and set min deploy...
Jens Verwiebe [Fri, 15 Nov 2013 19:05:23 +0000 (20:05 +0100)]
OSX/cmake: use CMAKE_XCODE_ATTRIBUTE rather than CMAKE_OSX_SYSROOT and set min deployment target to 10.6 now

5 years agoOSX/cmake: simplify system-version detection
Jens Verwiebe [Fri, 15 Nov 2013 18:26:43 +0000 (19:26 +0100)]
OSX/cmake: simplify system-version detection

5 years agoFix collada after git switch ( undefined build_commit_time )
Jens [Fri, 15 Nov 2013 17:37:57 +0000 (18:37 +0100)]
Fix collada after git switch ( undefined build_commit_time )

5 years agoFix for own mistake in r61178: bNodeTree->links ListBase was being modified while...
Lukas Tönne [Fri, 15 Nov 2013 15:54:05 +0000 (16:54 +0100)]
Fix for own mistake in r61178: bNodeTree->links ListBase was being modified while iterating ...

5 years agoFix Bevel artifacts bug T37053
howardt [Fri, 15 Nov 2013 15:19:38 +0000 (10:19 -0500)]
Fix Bevel artifacts bug T37053

In the case that there are two beveled edges with one unbeveled
one in between, and the widths don't allow them to magically
line up, it is better to slide along unbeveled edge.

Sometimes bevel widths are uneven (this was the case before)
and it is a followup TODO to do a width cleanup pass afterwards
to even the edges up as much as possible.

5 years agoCmake: little typo breaking player
jensverwiebe [Fri, 15 Nov 2013 15:30:49 +0000 (16:30 +0100)]
Cmake: little typo breaking player

5 years agoOSX/scons: make another gcc conditional futureproof
Jens Verwiebe [Fri, 15 Nov 2013 09:46:10 +0000 (10:46 +0100)]
OSX/scons: make another gcc conditional futureproof

5 years agoAdded the wrong file. Now addin .arcconfig for real.
Sergey Sharybin [Fri, 15 Nov 2013 14:40:43 +0000 (20:40 +0600)]
Added the wrong file. Now addin .arcconfig for real.

Summary is still:

5 years agoAdd .arcconfig for Phabricator Arcanist
Sergey Sharybin [Fri, 15 Nov 2013 14:38:39 +0000 (20:38 +0600)]
Add .arcconfig for Phabricator Arcanist


5 years agoFurther tweaks to buildinfo
Sergey Sharybin [Fri, 15 Nov 2013 11:11:59 +0000 (17:11 +0600)]
Further tweaks to buildinfo

Old idea with changes since previous release tag
didn't work good enough. In most of the cases tag
was done in a branch hence not actually reachable
from the master branch.

Now change since release is gone, and date of
the latest commit is used instead.

The date is displayed in format YYYY-MM-DD HH:mm
in the splash.

New fields:

- build_commit_timestamp is an unix timestamp of
  the commit blender was build from.
- build_commit_date is a date of that commit.
- build_commit_time is a time of that commit.

Reviewers: campbellbarton

Differential Revision:

5 years agosupport for paths with spaces in makefile
Campbell Barton [Fri, 15 Nov 2013 07:59:50 +0000 (18:59 +1100)]
support for paths with spaces in makefile

5 years agoadd gitignore
Campbell Barton [Fri, 15 Nov 2013 07:51:31 +0000 (18:51 +1100)]
add gitignore

5 years agoAdd functional and cctype headers to
Sergey Sharybin [Thu, 14 Nov 2013 21:51:17 +0000 (03:51 +0600)]
Add functional and cctype headers to

Fixes compilation with MSVC compilers.

Patch by leszekswirski (Lech Swirski) with some own tweaks.

Differential Revision:

5 years agoFix for msgfmt leaving context from previous message string
Sergey Sharybin [Wed, 13 Nov 2013 11:02:28 +0000 (17:02 +0600)]
Fix for msgfmt leaving context from previous message string

5 years agoRemove debug-only print from cmake macros
Sergey Sharybin [Wed, 13 Nov 2013 10:42:31 +0000 (16:42 +0600)]
Remove debug-only print from cmake macros

Left it there by an accident.

5 years agoInitialize git submodules for addons, locales and scons
Sergey Sharybin [Fri, 15 Nov 2013 06:19:08 +0000 (12:19 +0600)]
Initialize git submodules for addons, locales and scons

Uses relative paths to repositories, so this is expected
to work fine for any protocol we support (git, ssh and http).

Uses ignore=all for all the submodules, so updating them
to latest remote hash does not tags blender repository
as changes. But it is still possible to make changes to
submodules and commit them from their path.

5 years agoFix for [#37362] Audio strips sometimes are evaluated incorrectly.
Joerg Mueller [Tue, 12 Nov 2013 18:29:08 +0000 (18:29 +0000)]
Fix for [#37362] Audio strips sometimes are evaluated incorrectly.

For details see bug comments. The problem was that blender's animation system didn't update the audio animation system anymore due to an optimization. Fixed this in a complex but proper way in the audio animation system, so that it can handle gaps of missing values.

5 years agoSmall example for dynamic node labels in the template script.
Lukas Toenne [Tue, 12 Nov 2013 18:18:06 +0000 (18:18 +0000)]
Small example for dynamic node labels in the template script.

5 years agoMake dynamic node labels possible as a registerable function 'draw_label' (simple...
Lukas Toenne [Tue, 12 Nov 2013 18:18:04 +0000 (18:18 +0000)]
Make dynamic node labels possible as a registerable function 'draw_label' (simple 'label' identifier is already in use, need to avoid API breakage). This should simply return a string. The dynamic label can still be overridden by the user-defined node.label string.

5 years agoInstead of requiring a const char* return from the (optional) node label callback...
Lukas Toenne [Tue, 12 Nov 2013 18:17:58 +0000 (18:17 +0000)]
Instead of requiring a const char* return from the (optional) node label callback function, let it write into a mutable string buffer. This will allow actual dynamic labels for nodes using the python

5 years agoChange the behavior of AO pass in Blender internal's shader/render node tree so that...
Irie Shinsuke [Tue, 12 Nov 2013 10:59:40 +0000 (10:59 +0000)]
Change the behavior of AO pass in Blender internal's shader/render node tree so that it becomes (1.0, 1.0, 1.0) when AO is disabled.

For materials using AO pass, this makes the material preview and the GLSL
preview more accurate, but shouldn't affect final rendering in most cases
because we usually enable AO when using the AO pass in node tree.

Thanks to Brecht for code review.

5 years agoFix [#37409] Missing Buttons in Edit Strip Sub-Menu for Sound Strips in the VSE
Bastien Montagne [Tue, 12 Nov 2013 07:02:04 +0000 (07:02 +0000)]
Fix [#37409] Missing Buttons in Edit Strip Sub-Menu for Sound Strips in the VSE

Add back mute & lock buttons to audio strips (these should not have been removed, own fault).

5 years agoFix #37092 and #37381: crashes in the .object() method of Freestyle iterators.
Tamito Kajiyama [Tue, 12 Nov 2013 02:28:26 +0000 (02:28 +0000)]
Fix #37092 and #37381: crashes in the .object() method of Freestyle iterators.
Now the method checks if the iterator is at the end, and returns None if that is the case.

5 years agoFix [#37380] vertex paint colors don't render.
Bastien Montagne [Mon, 11 Nov 2013 20:37:19 +0000 (20:37 +0000)]
Fix [#37380] vertex paint colors don't render.

Another Evil Typo (r) one, you could add much more than the 8 allowed VCol layers!

Note: added some (warning-only) checks in mesh validate functions, but we still have a big issue with new cdlayer merge function, which could generate more than 8 layers of UVs or VCol... Don't know yet how to handle this situation. :(

5 years agoFix [#37388] Grid fill crashes blender in specific situation.
Bastien Montagne [Mon, 11 Nov 2013 14:29:01 +0000 (14:29 +0000)]
Fix [#37388] Grid fill crashes blender in specific situation.

With some geometries, we can have a valid first path, without being able to find a valid second one, added needed check.

5 years agoUsed wrong RNA access routine in bevel amount commit.
Howard Trickey [Mon, 11 Nov 2013 12:19:39 +0000 (12:19 +0000)]
Used wrong RNA access routine in bevel amount commit.

5 years agoFix #37395: Rendering with a movie strip behind a scene shows inconsistent results
Sergey Sharybin [Mon, 11 Nov 2013 12:08:31 +0000 (12:08 +0000)]
Fix #37395: Rendering with a movie strip behind a scene shows inconsistent results

For now just make sure conversion to sequencer space will ensure imbuf's
color space names is set properly.

Might be some further changes needed to make colorspace flow more clear
in sequencer, but that's for later.

5 years agoFix [#37394] UV Map cannot be renamed.
Bastien Montagne [Sun, 10 Nov 2013 20:32:27 +0000 (20:32 +0000)]
Fix [#37394] UV Map cannot be renamed.

Own epic failure! CustomData_get_named_layer() returns a relative index, not an absolute one. :(

5 years agoFix [#37393] Input dimensions are case sensitive in the property panel, uppercase...
Bastien Montagne [Sun, 10 Nov 2013 19:51:44 +0000 (19:51 +0000)]
Fix [#37393] Input dimensions are case sensitive in the property panel, uppercase input is misinterpreted

Typo in lowercasing code!

5 years agomake IS_EQ and IS_EQF typecheck args
Campbell Barton [Sun, 10 Nov 2013 14:13:26 +0000 (14:13 +0000)]
make IS_EQ and IS_EQF typecheck args

5 years agoreplace IS_EQ -> IS_EQF for use with floats.
Campbell Barton [Sun, 10 Nov 2013 12:56:50 +0000 (12:56 +0000)]
replace IS_EQ -> IS_EQF for use with floats.

5 years agoAdded options for how bevel amount is measured.
Howard Trickey [Sun, 10 Nov 2013 12:31:57 +0000 (12:31 +0000)]
Added options for how bevel amount is measured.
Now there is an 'Offset Type' dropdown on tool
shelf with types:
Offset - current method, offset of new edge
  from old along sliding face
Width - width of new bevel face (if segments=1)
Depth - amount a chamfering plane moves down
  from original edge
Percent - percent of way sliding edges move
  along their adjacent edges

The different options mainly are useful when
beveling more than one edge at once.

Leaving as a TODO to put these in the modifier,
as doing that has more permanent effects so
want to let users shake out problems with this

5 years agomistake in last commit
Campbell Barton [Sun, 10 Nov 2013 12:29:15 +0000 (12:29 +0000)]
mistake in last commit

5 years agono need to call isnan() on unsigned ints (causes error on freebsd/clang)
Campbell Barton [Sun, 10 Nov 2013 12:26:03 +0000 (12:26 +0000)]
no need to call isnan() on unsigned ints (causes error on freebsd/clang)

5 years agoWork around for bug [#37376] wrong tooltips on few minus buttons
Bastien Montagne [Sun, 10 Nov 2013 08:15:24 +0000 (08:15 +0000)]
Work around for bug [#37376] wrong tooltips on few minus buttons

Until we have a way to customize tooltips based on operator parameters...

5 years agouiList: fix another bug, where active item labels in sublayouts would not get colored...
Bastien Montagne [Sat, 9 Nov 2013 18:44:37 +0000 (18:44 +0000)]
uiList: fix another bug, where active item labels in sublayouts would not get colored as active (need recursion in sublayouts!). Was affecting e.g. Group node inputs/outputs lists.

5 years agouiList: fix a bug in computing visual active index (afaict, in case of reordering...
Bastien Montagne [Sat, 9 Nov 2013 17:05:07 +0000 (17:05 +0000)]
uiList: fix a bug in computing visual active index (afaict, in case of reordering, was affecting "keep active visible" feature when resizing, and the "active label color" in custom themes).

5 years agoOSX/scons: user request, make options succes transparent by printing it
Jens Verwiebe [Sat, 9 Nov 2013 16:32:10 +0000 (16:32 +0000)]
OSX/scons: user request, make options succes transparent by printing  it

5 years agoOSX/scons: spelling and cleanup comments
Jens Verwiebe [Sat, 9 Nov 2013 16:00:08 +0000 (16:00 +0000)]
OSX/scons: spelling and cleanup comments

5 years agoOSX/scons: Only cosmetical, a nicer formatting for available sdk listing
Jens Verwiebe [Sat, 9 Nov 2013 13:31:20 +0000 (13:31 +0000)]
OSX/scons: Only cosmetical, a nicer formatting for available sdk listing

5 years agoCycles / Fresnel Node:
Thomas Dinges [Sat, 9 Nov 2013 13:14:00 +0000 (13:14 +0000)]
Cycles / Fresnel Node:
* Add a "Normal" Input to the Fresnel node.
* Fix for the Fresnel GLSL code (normalize the Incoming vector).

Patch #37384 by Philipp Oeser (lichtwerk) , thanks!

5 years agoMore fix for [#37327] Inconsistent numeric input conversion.
Bastien Montagne [Sat, 9 Nov 2013 10:35:32 +0000 (10:35 +0000)]
More fix for [#37327] Inconsistent numeric input conversion.

When a single element is involved, apply directly the values instead of using the diff from init values. This avoids glitches when going from huge values to very small ones (due to float precision).

Note we should probably switch to doubles here, ultimately, but would leave that for later.

Manys thanks to Armin Zingler, who did all the investigation work on this point!

5 years agoBe ready for changes in bf-translations repository
Sergey Sharybin [Fri, 8 Nov 2013 20:44:48 +0000 (20:44 +0000)]
Be ready for changes in bf-translations repository

Made it so if there's release/datafiles/locale/po
folder, then all the .po files will be converted
to .mo at blender compile time and installed to
an appropriate location.

Uses small own implementation msgfmt which is
based on from Python project, but also
supports contexts.

There's no functional changes for until we've
switched to use source .po files instead of
pre-compiled .mo.

P.S. Well, there's one change which is
     being compiled even if it's not used, but
     would rather not clutter code with checks
     since pretty soon we'll use this program

5 years agoknife tool: Remove unused 'extend' variable
Sergej Reich [Fri, 8 Nov 2013 18:01:07 +0000 (18:01 +0000)]
knife tool: Remove unused 'extend' variable

5 years agoknife tool: Clean up line hits when starting a new cut
Sergej Reich [Fri, 8 Nov 2013 18:01:05 +0000 (18:01 +0000)]
knife tool: Clean up line hits when starting a new cut

Fixes intersection points of the previous cut still being drawn when
starting a new cut.

5 years agouse assert to check for incorrect use of BLF_draw_ascii()
Campbell Barton [Fri, 8 Nov 2013 15:01:43 +0000 (15:01 +0000)]
use assert to check for incorrect use of BLF_draw_ascii()

5 years agoOSX/scons: make MACOSX_SDK_CHECK a local var is sconstruct and remove from env, also...
Jens Verwiebe [Fri, 8 Nov 2013 13:44:10 +0000 (13:44 +0000)]
OSX/scons: make MACOSX_SDK_CHECK a local var is sconstruct and remove from env, also fix a longstanding misuse of MACOSX_SDK_CHECK in ceres, use MACOSX_SDK instead

5 years agoOSX/scons: remove never used MAC_PROC and minor cleanups
Jens Verwiebe [Fri, 8 Nov 2013 13:19:51 +0000 (13:19 +0000)]
OSX/scons: remove never used MAC_PROC and minor cleanups

5 years ago[#37327] Inconsistent numeric input conversion.
Bastien Montagne [Fri, 8 Nov 2013 11:25:50 +0000 (11:25 +0000)]
[#37327] Inconsistent numeric input conversion.

Issue was actually that micrometer was not drawing correctly (from r58165), reverted that fix and instead use utf8 drawing for editmode metrics when using a unit system (we already had a similar hack for surfaces and volumes, anyway).

5 years agofix for crash when deleting from an id property, with a non-string key.
Campbell Barton [Fri, 8 Nov 2013 08:46:43 +0000 (08:46 +0000)]
fix for crash when deleting from an id property, with a non-string key.

5 years agoFix for [#37293] Audio does not Scrub, plays fully instead
Joerg Mueller [Thu, 7 Nov 2013 23:40:06 +0000 (23:40 +0000)]
Fix for [#37293] Audio does not Scrub, plays fully instead

5 years agoOSX/buildbot: after scons refactor we only need the vars to change
Jens Verwiebe [Thu, 7 Nov 2013 15:25:11 +0000 (15:25 +0000)]
OSX/buildbot: after scons refactor we only need the vars to change

5 years agoOSX/scons: make gcc version detection more futureproof
Jens Verwiebe [Thu, 7 Nov 2013 15:05:59 +0000 (15:05 +0000)]
OSX/scons: make gcc version detection more futureproof

5 years agoTweak for ui range of track size in 3d viewport
Sergey Sharybin [Thu, 7 Nov 2013 14:07:13 +0000 (14:07 +0000)]
Tweak for ui range of track size in 3d viewport

Patch by Sebastian Koenig, thanks!

5 years agoOSX/scons: some cleanups
Jens Verwiebe [Thu, 7 Nov 2013 14:06:21 +0000 (14:06 +0000)]
OSX/scons: some cleanups

5 years agoOSX/scons: ARCH_FLAGS where not prperly added to CPPFLAGS
Jens Verwiebe [Thu, 7 Nov 2013 13:56:40 +0000 (13:56 +0000)]
OSX/scons: ARCH_FLAGS where not prperly added to CPPFLAGS

5 years ago"Render Audio" cleanup:
Bastien Montagne [Thu, 7 Nov 2013 13:32:32 +0000 (13:32 +0000)]
"Render Audio" cleanup:
* Removed audio-only options from ffmpeg render settings (added some versionning code too)!
* Moved the Mixdon button from the Scene->Audio pannel to the Render->Render panel.

5 years agoFix the small mess in read versionning code (we had two "pending versioning" blocks...
Bastien Montagne [Thu, 7 Nov 2013 13:14:33 +0000 (13:14 +0000)]
Fix the small mess in read versionning code (we had two "pending versioning" blocks, and a triangulate one that was outside of any block, plus I think existing "pending versionning" code was not added when 2.69.2 sub-version was created...).

Moved all non-yet-versioned stuff back into a single block at the end, let’s try not to forget to integrate it for 2.69.3! ;)

5 years agoFix for [#36819] *AFTER 2.69* Audio Only Sequencer wont render
Joerg Mueller [Thu, 7 Nov 2013 12:33:35 +0000 (12:33 +0000)]
Fix for [#36819] *AFTER 2.69* Audio Only Sequencer wont render

The original description is not a bug, but the mixdown options were missing as scons was lacking the required defines during compile time.

5 years agoFix for node transform offsets when adding new nodes: The transform operator was...
Lukas Toenne [Thu, 7 Nov 2013 09:15:29 +0000 (09:15 +0000)]
Fix for node transform offsets when adding new nodes: The transform operator was using the node->totr rect for defining the initial location of the node - but this is only defined after the node has been
drawn at least once, so nodes would jump to (0,0) after adding ... Use the locx/locy instead (including DPI factor), which is more reliable.

5 years agoFix for node link highlighting not getting reset if a transform operator is cancelled...
Lukas Toenne [Thu, 7 Nov 2013 09:02:30 +0000 (09:02 +0000)]
Fix for node link highlighting not getting reset if a transform operator is cancelled and the node removed. Pass the 'test' option in node link insert testing on to the conditions function, so that at least the highlight flags get cleared before exiting the link insert function, even if there is no actual selected node to insert.

5 years agoCleanup: nice bool return value from ed_node_link_conditions function instead of...
Lukas Toenne [Thu, 7 Nov 2013 09:02:29 +0000 (09:02 +0000)]
Cleanup: nice bool return value from ed_node_link_conditions function instead of using the space pointer for this.

5 years agoFix #37338: Add objects from addon work in a Linked Scene
Sergey Sharybin [Thu, 7 Nov 2013 08:40:47 +0000 (08:40 +0000)]
Fix #37338: Add objects from addon work in a Linked Scene

Added a poll function to add object helper which checks
whether scene is linked or not.

All the primitives which are delivered from this helper
will work properly for linked scenes. If there're still
primitives which are not delivered from this class, well
nothing i can do now.

5 years agoFix #37348, Different behaviour in Node editor. Adding nodes from the search operator...
Lukas Toenne [Thu, 7 Nov 2013 08:33:09 +0000 (08:33 +0000)]
Fix #37348, Different behaviour in Node editor. Adding nodes from the search operator now also removes them again if the subsequent transform is cancelled, like with the regular node shift+A menu

5 years agoOSX/scons: for completeness, added WITH_BF_OPENMP switch
Jens Verwiebe [Wed, 6 Nov 2013 23:19:01 +0000 (23:19 +0000)]
OSX/scons: for completeness, added WITH_BF_OPENMP switch

5 years agoRemoved the automatic "link swapping" feature from the node link operator: When a...
Lukas Toenne [Wed, 6 Nov 2013 21:21:37 +0000 (21:21 +0000)]
Removed the automatic "link swapping" feature from the node link operator: When a link was being dragged to an already connected input, the existing links were shifted to the next free socket. This was
originally intended as a way to speed up workflow for math and mix nodes, but more often than not it just gets in the way. Most binary (or n-ary) functions are not even commutative, i.e. changing the
order of sockets does not usually produce the correct result. Also this includes the more common case where one actually wants to replace a socket, which then requires a second click to remove the
shifted connection. All in all this is not a helpful feature.

5 years agoFix for error in r61159: the new gpencil_new_layer_col in UserDef is supposed to...
Lukas Toenne [Wed, 6 Nov 2013 21:08:16 +0000 (21:08 +0000)]
Fix for error in r61159: the new gpencil_new_layer_col in UserDef is supposed to be a 4 float RGBA color, but has only 3 floats. This was overwriting tweak_threshold in UserDef and breaking transform
operators ...

5 years agoOSX/scons: huge change, now you can set a just with the desired variab...
Jens Verwiebe [Wed, 6 Nov 2013 20:59:04 +0000 (20:59 +0000)]
OSX/scons: huge change, now you can set a just with the desired variables changed

5 years agoFix [#34675] *AFTER 2.69* Info view shows duplicate operators with incorrect values...
Bastien Montagne [Wed, 6 Nov 2013 20:56:18 +0000 (20:56 +0000)]
Fix [#34675] *AFTER 2.69* Info view shows duplicate operators with incorrect values of args

Refactored a bit WM api to generate operator's pystring, now it can also handle correctly macro operators. Thanks to Campbell for the review!

5 years agoBGE: Fix for #37335 "Moving the camera with a key (after the recent BGE cleanup commi...
Mitchell Stokes [Wed, 6 Nov 2013 19:40:37 +0000 (19:40 +0000)]
BGE: Fix for #37335 "Moving the camera with a key (after the recent BGE cleanup commits) now crashes the game" reported by Ace Dragon.

CcdPhysicsEnvironment->GetCharacterController(); was missing a NULL check.

5 years agoPatch #37274: Circle select for node editor, by Henrik Aarnio (hjaarnio).
Lukas Toenne [Wed, 6 Nov 2013 19:21:42 +0000 (19:21 +0000)]
Patch #37274: Circle select for node editor, by Henrik Aarnio (hjaarnio).

Circle select was missing from node editor, and C key was assigned to now defunct "show cyclic dependencies". This patch remaps the key and adds circle select operator.
Functions to check intersection between rctf/rcti and a circle were also added to rct.c for code cleanliness and consistency.

5 years agoRemoved the "Show Cyclic Dependencies" operator for nodes. This operator is an old...
Lukas Toenne [Wed, 6 Nov 2013 18:56:51 +0000 (18:56 +0000)]
Removed the "Show Cyclic Dependencies" operator for nodes. This operator is an old relic implemented for showing cyclic node connections as red links. This is not necessary any more, the operator simply
called the generic node tree update function, which is happening anyway after all relevant node operators (if it doesn't that has to be considered a bug).

It has been suggested to better use the C key for circle select, this remains to be discussed.

5 years agoTrue grid snapping for nodes: This snaps nodes to the actual background grid instead...
Lukas Toenne [Wed, 6 Nov 2013 17:46:32 +0000 (17:46 +0000)]
True grid snapping for nodes: This snaps nodes to the actual background grid instead of using incremental offset (which is not useful for nodes). Increment snapping has been disabled for nodes to avoid
confusion, grid snap is now the default as it seems to be the most wanted and easy to use mode.

Absolute grid snapping happens in a somewhat generic function 'applyGridAbsolute', which could also be used for objects and other transforms later on. It is conceptually similar to the 'project' snapping
option, in that it calculates a delta vector for each element on top of the overall transform, which places each node on the grid.

Node transform now uses the top-left node corner for TransformData->loc. The transform center is still the average of node centers, so that scaling and rotation works nicely.

snapGrid*** functions have been renamed to snapGridIncrement*** to distinguish better between incremental and absolute grid snapping.

5 years agoOSX/scons: remove obsolete compile flag
Jens Verwiebe [Wed, 6 Nov 2013 15:46:57 +0000 (15:46 +0000)]
OSX/scons: remove obsolete compile flag

5 years agoOSX/scons: remove obsolete compile flags and silence warnings same time
Jens Verwiebe [Wed, 6 Nov 2013 15:42:01 +0000 (15:42 +0000)]
OSX/scons: remove obsolete compile flags and silence warnings same time

5 years agoOSX/scons: remove another outdated conditional
Jens Verwiebe [Wed, 6 Nov 2013 15:37:14 +0000 (15:37 +0000)]
OSX/scons: remove another outdated conditional

5 years agoFix knife bug exposed by valgrind.
Howard Trickey [Wed, 6 Nov 2013 15:27:19 +0000 (15:27 +0000)]
Fix knife bug exposed by valgrind.
Was reading cage coordinate from those of existing
BMVerts even for newly created verts that don't
have cage coordinates there.

5 years agoOSX/scons: remove local MAC_MIN_VERS var and use the always identical MACOSX_DEPLOYME...
Jens Verwiebe [Wed, 6 Nov 2013 14:03:51 +0000 (14:03 +0000)]
OSX/scons: remove local MAC_MIN_VERS var and use the always identical MACOSX_DEPLOYMENT_TARGET instead

5 years agoOSX/scons: more cleanup of redundant vars, preparation to remove local vars to not...
Jens Verwiebe [Wed, 6 Nov 2013 13:56:11 +0000 (13:56 +0000)]
OSX/scons: more cleanup of redundant vars, preparation to remove local vars to not need to cp whole config as user-config later

5 years agoOSX/scons: simplification for getting the system version ( major, minor ), use sw_ver...
Jens Verwiebe [Wed, 6 Nov 2013 12:52:09 +0000 (12:52 +0000)]
OSX/scons: simplification for getting the system version ( major, minor ), use sw_vers -productVersion instead of uname -r, we must not redine it then from darwin version -> osx version

5 years agoSyncing methods for Color Balance node LGG and ASC-CDL modes. The settings for either...
Lukas Toenne [Wed, 6 Nov 2013 12:44:54 +0000 (12:44 +0000)]
Syncing methods for Color Balance node LGG and ASC-CDL modes. The settings for either mode are converted into equivalent settings of the other. This keeps the result of both modes roughly the same and
mimics the previous behavior when settings were shared by both modes (but not equivalent).
NOTE: Due to the use of additional sRGB conversion in the LGG mode the result is not entirely accurate, this should perhaps be fixed.

Settings for each mode are kept in their own color values nevertheless, this avoids potential problems with float precision.

5 years agoFix #37333: Bad default value in Color Balance. Use independent offset/power/slope...
Lukas Toenne [Wed, 6 Nov 2013 12:44:51 +0000 (12:44 +0000)]
Fix #37333: Bad default value in Color Balance. Use independent offset/power/slope variables for the CDL mode in color balance node. This avoids stupid default values in particular for offset, which would be 1 when just using the lift value for it.

5 years agoRemoved the DNA storage for LGG lift and inverse gamma in the color balance node...
Lukas Toenne [Wed, 6 Nov 2013 12:44:49 +0000 (12:44 +0000)]
Removed the DNA storage for LGG lift and inverse gamma in the color balance node. These values were always calculated at execution time, so there is no need to keep them around in DNA data and no forward compatibility break either. Only reason they were stored in DNA before is that the old compositor had no other means of keeping precomputed values around for every pixel than storing the DNA node data, with new compositor this is no longer necessary (values are stored in operations).

5 years agoGrease Pencil: User-Pref for setting the default colour of newly created layers
Joshua Leung [Wed, 6 Nov 2013 10:59:05 +0000 (10:59 +0000)]
Grease Pencil: User-Pref for setting the default colour of newly created layers