Let me know if this breaks anyone's anything.

This reverts commit 10f96462.

See discussion with Khaled Hosny on mailing list.  In short, since
integers here can be negative, and int division is "round towards
zero", proper rounding should take sign into account.  Just skip
doing it again, has been serving us well before.

This was introduced in 029f46bd
and was in fact a bug in gtk-doc master:

  https://bugzilla.gnome.org/show_bug.cgi?id=707426

I was under the impression that EXTRA_DISTs will happen regardless of
automake conditionals.  Apparently I was wrong.

Even though this is not what autoconf recommends, it reduces doc
rebuilds when version doesn't change but configure changes.

They've been disabled for a while and no one cared.  We're past
the point to need them for testing, and if we ever need to
resurrect them again, well, they're in git graveyard somewhere.

It's copied by gtk-docize.

Also enable bootstrapping without them.

Move it to new harfbuzz-gobject library.  Implement enums and
boxed types for object-types.  Still have to do boxed types for
value types.

https://github.com/blinkseb/harfbuzz/commit/f65dafa4138e1fb4395bf646fa33bb01a86a7e9a

Add deprecated alias for old name.

Makes documentation / introspection easier.

https://bugzilla.redhat.com/show_bug.cgi?id=1001645

That moment that you have to accept that you made API mistakes...