This is a bunch of hacks for now.

Improves Bengali a bit.

Althought IndicMatraCategory.txt classifies it as Top_And_Right matra,
it does not have Unicode decomposition, and Uniscribe does not do
anything special about it either.

Gujarati failures down from 0.672% to 0.0130966%.

Following what the spec says.

Brings down Telugu failures from 40% to 3.75%, and Kannada failures from
44% to 10%.  Does NOT affect other scripts' test results.

This is a hack for now.  Will be fixed when we do complex-shaper-driven
normalization properly.

The results with or without decomposition are the same, but Uniscribe
does not normalize, so this matches better.

Makes no visual difference.

Fixes most of the failures.  Down from 15% to 1.3%!

This is a hack for now, to get us going with Khmer.  This will be
refactored properly later to move the complex logic into complex
shapers.

We need this to set data on objects safely without worrying that some
other thread unsets it by setting it at the same time.

Where causing issues for people with MSVC.

Adjust glib fallback implementation.

The tests are not hooked up for ICU yet.

Add compose() and decompose() unicode funcs.  These implement
pair-wise canonical composition/decomposition.

The glib/icu implementations are lacking for now.  We are adding
API for this to glib, but I cannot find any useful API in ICU.
May end of implementing these in-house.

Changed all unicode_funcs callback names to remove the "_get" part.
Eg, hb_unicode_get_script_func_t is now hb_unicode_script_func_t,
and hb_unicode_get_script() is hb_unicode_script() now.

In preparation for adding more advanced unicode funcs.

Mutex (and Windows.h by extension) are fairly isolated now.

Makes way for adding arbitrary user_data support.

So we can liberally use the simple features of C++ that parts of the
codebase is already using.

This was a bizzare piece of API that I inherited from cairo.  It has
been wrong adding them to cairo in the first place.  Remove them before
someone uses them!

Unicode data providers can now be subclassed, including support for
chain-up.  The interface should now be nicely bindable, as well.

Also fix glib unicode funcs that where broken after hb_script_t
changes.  Nicely caught by the test-unicode.c added in this commit.

Also adds --script support to hb-view.

If a script tag is not known to us, we pass the ISO 15924 tag around.
Right now, the OT layer ignores that, but we can fix it to blindly
convert that to an OT script tag.

- Rename HB_SCRIPT_INVALID_CODE to HB_SCRIPT_INVALID

- Add HB_DIRECTION_INVALID

- Make hb_script_get_horizontal_direction() public

- Make hb_shape() guess script from buffer text (first non-common
  non-inherit script) if buffer script is set to HB_SCRIPT_INVALID (this
  is NOT the default.)

- Make hb_shape() guess direction from buffer script if buffer direction
is set to HB_DIRECTION_INVALID (this is NOT the default.)

- Make hb-view.c set INVALID script and direction on the buffer.

The above changes are meant to make hb-view fairly useful for uni-script
uni-direction text.  The guessing behavior however is NOT the default of
hb_shape() and must be asked for explicitly.  This is intended, because
the guess is not a suitable substitute to full-fledged bidi and script
segmentation.  It's just a testing tool.

Correspond to the make_mutable() ones.

One in particular is not a straight getter: hb_font_unset_funcs() is
special because of the specific needs of the lifecycle management of
the user_data object.