This changes the default load_flags of fonts created using
hb_ft_font_create() from NO_HINTING to DEFAULT.  Hope that doesn't
break too much client code.

Code calling hb_ft_font_set_funcs() is unaffected.

User can define hb_malloc_impl, etc, to name of custom allocator functions
that have the same signature as malloc.

Makes for a better default and avoids nasty inheritance issues.
See mailing list thread "Default hb_font_t scale".

We don't rely on that.  However, whenever hb_font_make_immutable()
is called, it makes its parenting chain immutable.

Part of https://github.com/behdad/harfbuzz/issues/136

Move the empty-buffer check from hb_shape_full() to hb_shape_plan_execute().

Reported by Simon Cozens.

Not functional change (expected!).

Fixes possible invalid read of two bytes.

Reported by Behzad Najjarpour Jabbari, Secunia Research.

As measured / improved by Benson Limketkai.

The BCP-47 registry defines a variant subtag "fonipa" that can be used
in combination with arbitrary other language tags. For example,
"rm-CH-fonipa-sursilv" indicates the Sursilvan dialect of Romansh
as used in Switzerland, transcribed used the International Phonetic
Alphabet.

http://www.iana.org/assignments/language-subtag-registry/language-subtag-registry

Fixes https://github.com/behdad/harfbuzz/issues/103

A while back we marked every API as version:1.0.  We should fix them all
to reflect real version they were introduced.  This is a start.

Patch from Nikolay Sivov.

This can possibly produce more granular clusters.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=541608
and cluster test.

Currently fails.
https://bugzilla.gnome.org/show_bug.cgi?id=541608

Needed for upcoming merge-clusters fix.

Use gtk-doc to build user's manual as well as reference

First two chapters. More to follow.

See previous commit.

Previously we were setting refcount of freed objects to the inert value, which
was harmful because it caused further destroy()s of the freed object to NOT
call free() and hence hide the bug.  Indeed, after eb0bf3ae test-object
was double-free'ing objects and this was never caught on Linux.  It only was
caught as crashing on Mac.

Now we poison refcount upon freeing and check that it's valid whenever reading
it.  Makes test-object fail now.

See thread "Issue with cursive attachment" started by Khaled.
Turned out fixing this wasn't as bad as I had assumed.  I like the
new code better; we now have a theoretical model of cursive
connections that is easier to reason about.