Blender Git Loki

Pre-merger changes.

Deprecated Freestyle-specific elements were removed from the code base,
in order to address all comments from branch code reviews.
https://codereview.appspot.com/7416049/

Backward compatibility won't be maintained in the following components:

- Freestyle edge/face marks in old .blend files are ignored.

Old .blend files can be converted by loading and saving the files using
a Freestyle branch build between revision 55581 and 55842.

- External style modules are no longer supported.

Instead text datablocks must be used to keep style module files within
.blend files. A branch build between revision 55741 and 55842 is useful
for identifying the style module file names that need reconfiguring.

svn merge ^/trunk/blender -r55815:55840

Node UI:
* After PyNodes merge, many node sockets had sliders, instead of the arrow buttons (as the PROP_FACTOR subtype now actually was used). Change those back now, after artists feedback.

fix: #34871 Cycles/CUDA/sm_35: Build problems probably due to issues with float3 operators (on windows?)

Fix: 34874 Weight Paint UI inconsistent: removed conflicting shortcut definition, added weight assign tools to specials menu

ctrl+a select all in button field

Minor white space fix.

Removed Freestyle-specific code from DerivedMesh.c.

The removed code was necessary due to a logical bug in the use of CD_ORIGINDEX
layer entries. Previously the index values were used to refer to the edges/polys
of the derived mesh, assuming that the derived mesh edges/polys consisted of those
from the original mesh plus newly added edges/polys. This assumption was not
correct, although it worked well for several modifiers such as Mirror and Solidify.
Now the index values are specifically used to reference the edge/polys of the
original mesh.

add relative offset option for poke tool.

patch [#34890] BMesh Poke Face.
by Francisco De La Cruz (xercesblue), with some of my own changes/improvements.

Converts faces to triangle-fans (useful to run on ngons).

To access select a group of faces and press "Alt+P" or alternatively select the operator from the Faces menu (Ctrl+F)

Fix #34880: cycles motion blur render issue with some compilers. Actually is a bigger
problem where accessing float4 members with [] stops working due to optimizer, will
check that later.

bmesh operator rename inset -> inset_region

patch [#34886] BMesh Individual Face Inset
from Francisco De La Cruz (xercesblue) with some simplifications to the patch.

fix [#34870] bmesh.ops.* parameter lists and descriptions don't show in PyConsole on auto-complete

more a feature request then a bug but nice to have __doc__ on bmesh operators.

Removed CD_MASK_FREESTYLE_EDGE and CD_MASK_FREESTYLE_FACE from CD_MASK_BAREMESH
and moved them to init_render_mesh() in convertblender.c where these CustomData
layers are specifically required.

add BM_face_calc_center_mean_weighted() gives much better result at cost of some speed.

Fix another part of #34877: cycles progress status text not showing correct with
per render layer samples in addition to the progress bar.

Also fixed job progress bar not working at all on high DPI / retina, was so small
the actual progress was not visible.

code cleanup: include order

Fix eternal compo refresh when having image editor with mask opened

This was a regression since 55719 and the reason is:

- image_refresh() will tag combo for re-calc when editing mask.
This is a feature, so you could have immediate feedback on masking
and compositing while editing mask. This is done from image_refresh().

- Color management sends NC_WINDOW notifier, so the whole window is
updated when changing color management settings. Image editor used
to refresh space when this notifier is sent.

The same notifier is sent by compositor job to redraw all possible
viewers.

Simple fix: don't refresh image space for NC_WINDOW, there's nothing
in image_refresh() which we would want to happen on NC_WINDOW event.

Bug fix #34866

Global undo/redo now clears the local undo editmode stack entirely.

Error goes back to the 2.3 days, when undo was added.

Global undo refreshes the entire internal database, so all ID pointers get invalid.
This cases editmode undo storage to fail, if it uses ID pointers as well.

The error was that for any Mesh undo stack, a single global undo would make the
mesh stack corrupt. Back in edit mode, on undo you'd lose assigned texture images,
or get bad crashes.

The downside is that people expect this to work... it's a fun feature to maintain
stacks separately. Having this instable is not acceptable though. Needs quite some
redesign work to solve it (like Dalai's kill-the-tface project :)