Macro recording is extremely buggy, and generates un-readable or
debuggable macros, keep it available, but only in experimental mode.

Signed-off-by: Sebastian Spaeth <Sebastian@SSpaeth.de>

Some custom shapes can have shaded parts, like for example 3d can,
the bevelled buttons etc. Those shaded colors are calculated
internally, and have been way off at times. Now using HSV color
space & the originally documented luminance modifications in steps
of 10 percent. Compared to MSO, still no 100 percent match, but
that seems due to gamma correction there.
Additionally, starting with MSO12, gradients on those shaded surfaces
look much better; adapted code to display gradients equally nice.

Note that most of this patch also applies to ooxml import; note as
well that customshapes from *all* kind of input files (including ODF
docs) now look different than before; no real way of changing this
in a backward-compatible way, since behaviour of custom shapes is
mandated (mostly) by internal tables, and not stored in a file.

Applies patches/dev300/ppt-customshape-shading-fix (much of it was
accepted at OOo already, via i#102797)

Applies patches/dev300/ppt-customshape-shading-fix.diff: fixed prob
with line arrows - the extra-added single point polygons lead to
extra arrows randomly around the custom shape. i#105654

With 3.0, Impress numbering was modified because of the odf1.2
changes in list level handling; we differ in the way we
consider paragraphs to belong to the 'same' numbering group,
compared to ppt. now, as long as the para has the same outline
level (and no 'restart numbering' flag), it is considered to
belong to the same numbering group, i.e. numbering will mono-
tonically increase. Compares favorable to PPT. Additionally,
fixed a repaint bug that prevented following paragraph's numbers
to change, when certain modifications were made on the current one.

Applies patches/dev300/sd-ruler-fix.diff

Much easier to interact with shapes - screen dpi is only getting
higher these days

Thanks to Joseph Power for finding this.

In practice the changeset is 'inspired' by vosremoval-mutex.diff
but was essentially redone manually