Fix #32398: Mirror modifier with "Merge" enabled producing pairs of faces
sharing the same set of vertices.

Modified the CDDM_merge_verts function (currently only used by the Mirror
modifier) to skip faces using all merged vertices.

Fix for [#17421] Issue with Displace modifier not updating with
animated textures

I have added a dependsOnTime function for the Displace modifier which checks
if the displacement texture has IPOs, is a plugin, or uses an animated image.

Fix for [#17291] Array modifier merge caps leaves a whole
This code was missing a test_index_face call after merging, so it was making
quads with v4 == 0.

Fix for [#17376] Applying mirror modifier does not respect UV pinning.

The custom data system wasn't swapping pinning or selection flags for MTFace
data when face vertices were reordered.

Fixed bug #14510 - wave modifier crash

CDDM_apply_vert_coords needs a CDDerivedMesh, but it was getting a
CCGDerivedMesh from the preceding Subsurf modifier. This fix just makes
a CDDerivedMesh copy of the supplied DerivedMesh rather than using it directly.

Patch #8882 - Falloff in the wave modifier
This patch adds the ability to specify a falloff radius in the Wave modifier.
Currently only linear falloff is supported.

Thanks to Michael Fox for the patch!

Patch #8613 - Vertex groups in the mirror modifier.
This patch re-assigns the mirrored data to use vertex groups with "mirrored"
names (e.g. L_arm -> R_arm, Leg.R -> Leg.L etc.). Vertex groups with the
"mirrored" names must already exist in the base mesh.

This means that it is no longer necessary to apply the mirror modifier in order
to rig the mirrored data independently.

Thanks to Michael Fox for the patch!

Fix for edgesplit crashes reported in bugs 6695 and 7142. The problem, as
found by Riku Palomäki, was that "bridge" vertices (which connect two or more
faces that don't share edges) were not being included in the maximum final
vertex count calculation (used for memory allocation, hence the crashes).

Unfortunately Riku's patch actually stopped bridge vertices from being split
correctly, so I have fixed the problem by adding the maximum number of vertices
that can be generated from bridge vertices to the maximum final vertex count.

Thanks to Riku Palomäki for finding the cause!

Fixed editmode post-modifier normal calculation for the case where there are
only deform modifiers in the stack. This was an old bug: a typo caused quad
face normals to be calculated using the same vertex twice, giving bad normals.

Fixed the "Enable modifier during editmode" button for the case when there are
only deform modifiers in the stack. The code that copied the cage mesh to the
final mesh didn't check for vertex coordinates; I have added this check.