-#define indic_offset_total 4416
+#define indic_offset_total 3816

-}; /* Table occupancy: 60% */
+}; /* Table occupancy: 69% */

No functional change.

HB_VERSION_CHECK's comparison was originally written wrongly
by mistake.  When API tests were written, they were also written
wrongly to pass given the wrong implementation... Sigh.

Given the purpose of this API, there's no point in fixing it
without renaming it.  As such, rename.

API changes:

  HB_VERSION_CHECK -> HB_VERSION_ATLEAST
  hb_version_check -> hb_version_atleast

https://bugzilla.mozilla.org/show_bug.cgi?id=986802
Fixes https://github.com/behdad/harfbuzz/pull/39

API Change:

-hb_ot_layout_language_get_required_feature_index
+hb_ot_layout_language_get_required_feature

New API takes an extra pointer argument.  Pass NULL in to get
behavior of previous API.

Reworked by behdad

Such that the complexity of the algorithm used is clear at
call site.

In an effort to make the algorithm used clear.

Still needs update to joining table to fully work.

...so that they overstrike preceding glyph.

https://github.com/behdad/harfbuzz/pull/43

If pre-base reordering Ra is NOT formed (or formed and then
broken up), we should consider that Ra as base.  This is
observable when there's a left matra or dotreph that positions
before base.

Now, it might be that we shouldn't do this if the Ra happend
to form a below form.  We can't quite deduce that right now...

Micro test added.  Also at:

https://code.google.com/a/google.com/p/noto-alpha/issues/detail?id=186#c29

Sometimes font designers form half/pref/etc consonant forms
unconditionally and then undo that conditionally.  Try to
recover the OT_H classification in those cases.

No test number changes expected.

No test numbers change.

Previous commit was not compiling with clang.

As in building for Android / iPhone.  Only set it if
struct{char} alignment is not 1.

NOT tested on an actual Arm architecture.  Guess we'll know
when this makes it to people's build bots.

Normally if you want to, say, conditionally prevent a 'pref', you
would use blocking contextual matching.  Some designers instead
form the 'pref' form, then undo it in context.  To detect that
we now also remember glyphs that went through MultipleSubst.

In the only place that this is used, Uniscribe seems to only care
about the "last" transformation between Ligature and Multiple
substitions.  Ie. if you ligate, expand, and ligate again, it
moves the pref, but if you ligate and expand it doesn't.  That's
why we clear the MULTIPLIED bit when setting LIGATED.

Micro-test added.  Test: U+0D2F,0D4D,0D30 with font from:

[1]
https://code.google.com/a/google.com/p/noto-alpha/issues/detail?id=186#c29

Roboto was hitting this.  FreeType also has pretty much the
same code for this, in ttcmap.c:tt_cmap4_validate():

    /* in certain fonts, the `length' field is invalid and goes */
    /* out of bound.  We try to correct this here...            */
    if ( table + length > valid->limit )
    {
      if ( valid->level >= FT_VALIDATE_TIGHT )
        FT_INVALID_TOO_SHORT;

      length = (FT_UInt)( valid->limit - table );
    }

Can happen either in broken fonts, or as a result of sanitize().

Bug 77800 - cppcheck reports

Apparently they can intermix with other OT_A.

Test: U+0915,A8E2,1CD0

Patch from Martin Hosken.  I expect this to fix the following bugs:

https://bugs.freedesktop.org/show_bug.cgi?id=75076
https://bugzilla.gnome.org/show_bug.cgi?id=723582
https://bugzilla.redhat.com/show_bug.cgi?id=998812

Tests: U+0967,0951 U+0031,093F