Blender Git Loki

Middle-clip tweaks:
* Now only middle-clip when both left and right parts of remaining strings are at least two times icon-width (thus most short labels will never be middle-clipped).
* Add a trailing ellipsis to right-clipped strings (unless it would take more than 20% of whole available width, in this case just clip chars, to keep as much data visible as possible!).

Armatures Editmode: option to recalculate roll from the angle to the parent

Text3d: underline offset on curved path wasnt working right

Fix T38075, crash from shader node add/replace tree view template.

The sanity check for copying socket default value was using the socket
idname for type compatibility checks, which is too strict. Subtypes
would not be recognized as copyable, but since only the plain data type
is needed below this is all that needs to be checked.

That alone would not cause crash (just missing default value copy),
but the tree view template was messing with the default_value DNA
directly by freeing it in advance, which is not necessary and should
be left to blenkernel when freeing the node. Otherwise this would
leave the node invalid without a default_value if the copy function
bails out.

Text3d: Add support for underline characters with text-on-path

Fix T38044: Filter Key Binding shows Incorrect result

Object Operators: make-dupliface and join-uvs weren't using UNDO.

also improve make-dupliface tip

Text3d: move text selection boxes from Curve to EditFont struct

resolves T38079

Fix bash/ksh specific test syntax

[ foo = bar ] has been the traditional compare operator, using == is supported by some shells, but not all. The latter offers no advantage over the former.

Differential Revision: https://developer.blender.org/D179

Code cleanup: defines for statfs were getting out of hand for BSD's.

add __DragonFly__ and internal defines to avoid copy-pasting checks.

also remove __CYGWIN32__ check, since cygwin is no longer supported.

OSX/scons: after discussion i was told cycles does decide to use sse4.1 kernel on the fly,
so autodetection is obsolete @buildtime. Leave in print for testing case

OSX/scons: autodetecting sse4.1 capability and set it to used

Fix to own previous commit for bevel vertex only case.

In separating out the adj mesh change from a profile change,
I'd forgotten that some profiles need to be flat (in particular,
for vertex-only bevels). This fixes that.

Added scons option to actually build the sse41 kernel. CMake already has it

Cycles: Don't include SIMD util file for OpenCL/CUDA, this fixes OpenCL compilation.

Take back last header copy, due it is for native only, must be a runtime solution, todo: do by definitions

Cmake: fix kernelcompile after introduction of util_simd.h

Cycles: Minor optimization (~1%) for texture access on CPU

Cleanup: use blend() in perlin noise (gives 12 less instructions on SSE4.1)

Bevel: use stack memory for interp_vmesh()

Rather then using the memarena and never freeing.