Should have no visible effect.

Fixes 200 failures out of previous 4290 cases in the OO.o Indic
dictionary (of ~16000 entries).

In preparation for Hangul shaper.

Based on patch from Jonathan Kew.

Where causing issues for people with MSVC.

Read comment.

Not sure where this is documented, but I remember this being the desired
behavior.

test-shape-complex failures are down from 48 to 46.  Meh.

Not quite working just yet.  False alarm re 10 failures.  It was
crashing.  Ouch!  Back to 48 failures.

Not applied yet.

Brings test-shape-complex failures down from 52 to 10!

I hereby declare harfbuzz-ng supporting Indic!

But I'm not sure...  Down from 54 failures to 52.

In old-style Indic OT standards, the post-base Halants are moved after
their base.  Emulate that by moving first post-base Halant to
post-last-consonant.

Brings test-shape-complex failures down from 88 to 54.  Getting there!

Brings test-shape-complex failures down from 104 to 92.  Way to go!

Is the categorization correct?  I don't know.

Number of failing shape-complex tests goes from 125 down to 94.

Next: Add Ra handling and it's fair to say we kinda support Indic :).

Find the base consonant and apply basic Indic features accordingly.
Nothing complete, but does something for now.  Specifically:
no Ra handling right now, and no ZWJ/ZWNJ.

Number of failing shape-complex tests goes from 174 down to 125.

Next: reorder matras.

This allows the Indic shaper to request decomposed characters.  This will
handle split matra for free.  Other shapers prefer precomposed
characters.

Instead of always applying those two features before the complex shaper,
let the complex shaper decide whether they should be applied first.

Also add stub for Indic's final_reordering().

Still not much there.

Copied form HarfBuzz.old Indic data.  These are below and post
consonants.  This is temporary.  Read the comment in the patch.

Using an incredible tool called Ragel.

Matches OT types now.