Towards possibly using Null pool for some nil objects.

If a pointer type was passed to Null(), reinterpret_cast<> was
complaining about qualifiers being removed. Turns out I need the const on
the other side of "Type" to fix that.  Also remove unused const from
NullPool type.

Not sure if I like to keep it. For now, aim for correctness.

Common Regoin for Access Protection.  Like the NullPool, but writable.

We should hang those off somewhere else. For now, the unused ones can go.

Part of https://github.com/harfbuzz/harfbuzz/issues/1017

If we don't have visibility, make NullPool static.

If the `calloc` for `gsub_accels` or `gpos_accels` fails, then the
unlikely branch afterwards can be taken, which frees up the
`hb_ot_layout_t`, but since those fields can now be `nullptr`, then
we don't want to dereference them.

Untested, but compiles.

(And prevent it from happening in the future.)

Saves some more code size as well!

Saves 2k of .bss section.

Not used yet.

There are more broken versions of Tahoma out there on various Windows releases,
so we need to add them to our blacklist to avoid broken rendering.
See https://bugzilla.mozilla.org/show_bug.cgi?id=1279925 for details.

Patch from Phil Race.

In https://crbug.com/681813 another instance of Padauk was identified
triggering collapsed glyphs. Blacklist this version by patching
hb-ot-layout.cc to print out gdef, gsub, and gpos table length, then
adding those to the list of blacklisted versions.