Merge branch 'master' into texture_nodes_refactor

Conflicts:
source/blender/nodes/texture/nodes/node_texture_math.c

Merge branch 'master' into texture_nodes_refactor

Port some more nodes

Now only node group remained to be ported, all the rest of the nodes
cant' be actually supported and will be removed.

As a replacement for them we'll bring some more control by adding
some mapping sockets to the input nodes.

Port procedural texture nodes to new framework

Support some more nodes

Port several more nodes to the new system

Need to switch to some other work now, maybe someone will want to play
around with the code meanwhile.

Bring back image input, checker, viewer and output nodes

This more like a prove-of-concept change because well, we've got
a dependency solver in the nodes already which seems to work just
fine for the shading nodes. So want to see if this is something we
can use for texture nodes now.

This isn't that far away from the original idea of having some
SVM-like machine for texture nodes.

Hell of a lot of nodes are not gonna to work yet,

Port texture nodes to the exec pool in renderer

This only makes it so renderer uses the proper API to deal with the
tree evaluation. Tree nodes still does nothing because all the delegate
system has been just removed.

Painting and sculpting will also fail and abort .

Revert "Cycles: Implement an area preserving parameterization sampling for area lamps"

This reverts commit c4235ce820b0a9c3d019e31e7d4360162a0282a5.

Bring material nodes back to the render engine

The idea is to make make it so per-tree execution data is owned by
the renderer and being passed to the tree evaluation routines.

Get rid of old legacy stuff happening all over the place

Basically get rid of texture nodes deligates and exec data stored
in the DNA node tree structure.

Blender is pretty much unusable now, because there are whole areas
to be ported over.

Cycles: Implement an area preserving parameterization sampling for area lamps

Replace old code for area lamps which was more like incorrect with more correct
one using the following paper as a reference:

Carlos Urena et al.
An Area-Preserving Parametrization for Spherical Rectangles.
https://www.solidangle.com/research/egsr2013_spherical_rectangle.pdf

Implementation is straight from the paper, currently the rectangle contants are
calculated for each of the samples. Ideally we need to pre-calculate them.

The old PDF is still used, which makes difference real subtle. This is to be
corrected before final commit.

Reviewers: brecht, juicyfruit

Subscribers: dingto, ton

Differential Revision: https://developer.blender.org/D823