MAP_NORESERVE is not available on macOS for example so set the flag
to zero if not defined on the headers.

"writable" is not in use and probably we can go in the reverse direction
and remove the flags that are useful for enabled writable, this is just
for the sake of completeness however.

It removes io.h and other polyfills which we no longer need as 7e76d746

This makes it compatible with ARMCC which I had access in
a collaboration with @imgtec, thanks!

Basically hb_blob_create_from_file features three code paths,
mmap, Win32 and fallback.

We had fallback implementation even before this but it was relied
to "open" which is not available on some environments. This change
improved the situtation by using only fopen and friends for
fallback path.

Interestingly we could use "open" on Windows but in fact it was
emulated by MSVCRT so I've completely split that from Unix path
now that we have a distinct path for fallback path also.

Towards making blob more memory-allocation-friendly

https://bugzilla.gnome.org/show_bug.cgi?id=659212 for more context

New API:
- hb_blob_copy_writable_or_fail()

Defining _POSIX_C_SOURCE to an old version on FreeBSD can cause C99 to
be disabled in libc.

Part of fixing https://github.com/behdad/harfbuzz/pull/605

Fixes https://github.com/behdad/harfbuzz/issues/474

Fixes https://github.com/behdad/harfbuzz/issues/227

Fixes https://github.com/behdad/harfbuzz/issues/103

hb_blob_create() is considered C-only API.

Minimal shaping works now!

Fail if blob start plus length overflows; or if blob length
is greater than 2GB.  It takes a while for fonts to get to that
size.  In the mean time, it protects against bugs like this:

  http://www.icu-project.org/trac/ticket/11450

Also avoids some weird issues with 32bit vs 64bit systems
as we accept length as unsigned int.  As such, a length of
-1 will cause overflow on 32bit machines, but happily
accepted on a 64bit machine.  Avoid that.

Fixes https://github.com/behdad/harfbuzz/pull/45

Should move these out of the public header...

We're "clean" of introspection warnings now.  Remaining ones are about
graphite2 / freetype types not being introspectable.

Since our types are not associated with their methods, marking
constructors makes them inaccessible from bindings.  Undo for now.

This fixes a design bug with sanitize and sub-blobs that can
cause crashes.  Jonathan and I found and debugged this issue
when we tested a corrupt font with the md5sum / filename:
ea395483d37af0cb933f40689ff7b60a.  Two hours of intense
debugging we found out that the font has overlapping GSUB/GPOS
tables, and as such, sanitizing the second table can modify
the first one, which can cause all kinds of undefined behavior.

The correct way to fix this is to make sure sub-blobs are
always created readonly, since we consider the parent blob
to be a shared resource and can't modify it, even if it *is*
writable.

This essentially makes the READONLY_MAY_MAKE_WRITABLE mode
unused...  Maybe we should simply remove / deprecate it.

The OSAtomicAdd32Barrier operator returns the new value, we want the
old value.

We need this to set data on objects safely without worrying that some
other thread unsets it by setting it at the same time.

Where causing issues for people with MSVC.