https://github.com/behdad/harfbuzz/issues/538

Fixes https://github.com/behdad/harfbuzz/issues/538

This is a temporary hack just to avoid incorrect shaping results, pending proper support
for caching shape-plans with variation coordinates.

* Fix missing xscale in rtl graphite unpacking

* Oops didn't need to mess with yscale

Test with, eg.:
$ LC_ALL=fr_FR.utf-8 ./hb-view NotoSansArabic-VF.ttf بهداد --variations wght=1.2

-o, -O, -u, -v, -V

If buffers are empty, content type should be ignored.

This fixes last of the failing tests: fuzzed.tests.  Green again!

Fixes tests/shaping/tests/cluster.tests

Fixes https://github.com/behdad/harfbuzz/issues/154

Fixes https://github.com/behdad/harfbuzz/issues/528
"Kannada JIHVAMULIYA and UPADHMANIYA insert dotted circles"

Fixes tests/automatic-fractions.tests

If *I* get this wrong in 2017, I have no idea how others get their head around
cluster math...

Fixes tests/arabic-fallback-shaping.tests

I like to have a mode where CONTAINS_NOTDEF and CONTAINS_DOTTEDCIRCLE are not
returned.  Abused a value of -1 for that.  hb-shape now uses it.  Fixes two
of the six tests failing with --verify in test/shaping/run-tests.sh.

Disabled for now.  Will enable and fix failures after next release.

New API.

Change hb_buffer_diff to explicitly cast result of abs to unsigned when
comparing with position_fuzz to avoid unsafe signed/unsigned comparions
warnings on windows.

Based on patch from Jonathan Kew.

Needs more cleaning up and documentation.

New API:
hb_buffer_diff_flags_t
hb_buffer_diff()