Hook AAT's lcar to _get_ligature_carets

A starting point, its sanitization is not tested however

I think it makes sense to accumulate it, but Ned tells me that's
what CoreText does.

Vastly improve kern/kerx tables, including cross-stream "kerning"

Let the fuzzing bots rip this code apart...

Finally:  Fixes https://github.com/harfbuzz/harfbuzz/issues/1356

Test case:
$ ./hb-shape GeezaPro.ttc -u U+0628,U+064A,U+064E,U+0651,U+0629
[u0629.final.tehMarbuta=4+713|u064e_u0651.shaddaFatha=1@0,-200+0|u064a.medial.yeh=1+656|u0628.initial.beh=0+656]

The mark positioning (kern table CrossStream kerning) only works if deleted
glyph (as result of ligation) is still in stream and pushed through the
state machine.

Happens even if the cross-stream kerning is for cursive attachment only.  Oh well..

Ouch!

Not needed now that we use GPOS attachment for cursive kerx.

Fixes https://github.com/harfbuzz/harfbuzz/issues/1365

Untested.

That's what the spec says for Backwards flag, only applicable to
formats 1 and 4.

Untested.