get_glyph_from_name() coming soon.

Fixes https://github.com/behdad/harfbuzz/issues/585

Such that client cannot accidentally destroy them, even though that
will be a bug in their code...

Oops.

Saves some more code size as well!

Even 1 is too many but putting 0 breaks compile.  Saves 3k in .so

Saves 2k of .bss section.

In ten years we never used them...

Fixes https://github.com/behdad/harfbuzz/issues/571

Fixes https://github.com/behdad/harfbuzz/issues/585

Part of fixing https://github.com/behdad/harfbuzz/issues/585

Part of fixing https://github.com/behdad/harfbuzz/issues/585

https://github.com/behdad/harfbuzz/issues/585

Used to build static string arrays that use no relocation.

https://github.com/behdad/harfbuzz/commit/49a41dc75931cc2d2f7e74b7801f8cc327039e8e#commitcomment-25203194

Fixes https://github.com/behdad/harfbuzz/issues/579.

This adds a couple extra entries to the set used, such that iteration
would fail in 1.6.2 due to the bug in hb_set_t::page_t::next().

https://ci.appveyor.com/project/behdad/harfbuzz/build/1.0.668/job/ikfj8peowgrp059t

The get_glyph_from_name() is incorrect; needs fixing.

Turns out we already have support for “post” table, it just needed to be
activated and put in use.

https://github.com/behdad/harfbuzz/issues/579

Fixes https://bugs.freedesktop.org/show_bug.cgi?id=103346

Trying to figure out why this is failing on some Debian bots:
https://bugs.freedesktop.org/show_bug.cgi?id=103346

Maybe I should check for GNU grep specifically now that I'm using -n as well.