Ouch!

Sigh.

[subset] Have the subset fuzzer pull the text string to subset to from the end of the fuzzer provided data.

This reverts commit ff92de76.

Revert that and remove ASSERT_POD. Let's see which bots are
unhappy with this configuration...

Towards possibly using Null pool for some nil objects.

Null() and Crap() are only places that there's no space before '('...

Towards using Null object for nil objects.

Unused, but now that we have CrapPool, implement it.

Apparently our gcc-6.4.0 on Alpine Linux distribution doesn't like defining static const
inside a method, lets put that on outside the classes.

Trying to see if it fixes gcc-4.2 bots.

Don't know why I thought I should subtract one there...

llvm-gcc-4.2 bot had this problem:
hb-private.hh:812: error: initializer specified for non-virtual method 'void hb_vector_t<Type, StaticSize>::operator=(const hb_vector_t<Type, StaticSize>&) [with Type = hb_user_data_array_t::hb_user_data_item_t, unsigned int StaticSize = 1u]'

Removing the delete didn't work with a constructor. So, remove constructor.
Just disallow assignment.  Still better than nothing.

Clang build was failing with:

/usr/bin/ld: .libs/libharfbuzz_la-hb-blob.o: relocation R_X86_64_PC32 against undefined hidden symbol `_ZTH12_hb_CrapPool' can not be used when making a shared object

Instead of fighting it, just disable use of it.

This reverts commit face7cf5.

This reverts commit 7bf19801.