And include it in hb_shape_plan_t itself.

Fixes https://oss-fuzz.com/v2/testcase-detail/5754863779053568

Use normal, production, shared libraries.

Fixes https://github.com/harfbuzz/harfbuzz/issues/1237

Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11367

This is the root cause of bug worked around in 19e77e01.

Still no shape plan caching for variations though.

No other shaper will need shape_plan_data, by definition.  So, remove
abstraction layer and always create hb_ot_shape_plan_t as part of
hb_shape_plan_t.

This reverts commit 442a72d9.

Doesn't make sense.  No idea how my local compilers where happy with it!

Ouch :).

Although didn't need it...

Fix hb_blob_ptr_t::get_length () as well.

To remove custom null object next..

We never rely on that being equal to readonly.  Just not being
writable.  Maybe not even that given that the object is inert.

In prep for next commit, using default null pool.

Use in a couple of places.  Push to bots to see how many unhappy before
I convert the rest.

Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11312
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11305