OpenSystemsDevelopment / qpdf

This reverts commit bb045907a043b5c6de9fb804ff11087333747329.

Move methods responsible for loading or keeping track of objects to
QPDF_objects.cc.

The split was part of the reverted #1297. Reintroducing it now makes it
easier/safer to keep the work to refactor the xref and object tables in
sync with main.

Figuring out which classes are part of the public API by using library
symbols is fragile (dependent on specific compiler optimizations) and
unreliable (misses some inline things). Instead, use castxml, a tool
that parses C++ to an abstract syntax tree and generates XML, to get a
reliable accounting of public classes and their sizes.

Replace data member oh with accessors.

Remove redundant overloads from QPDFObjGen::set.

Fix the logic to accept a top-level `zopfli.h` header, as that is
the location used by upstream's build system.

This requires a special build option.

This reverts commit c648b9a018105a3c30b3e7f3c5f8a058d3ddd92c, reversing
changes made to 12b67a3227df6b6df3a4f5f098e11cce173ff7d5.

Move methods responsible for loading or keeping track of objects to
QPDF_objects.cc.

Having it at compile time breaks cross-compilation and isn't really
right anyway.

Add temporary clone of SparseOHArray  to implement non-sparse mode.

* Define it even though previous experiments have shown it to be
  unnecessary since it seems like it may be necessary after all
* Add QPDF_DLL_CLASS to QPDFObjectHelper and QPDFDocumentHelper in
  case there's some future unknown reason why someone may want to have
  them and/or in case it helps with the weird
  QPDFNameTreeObjectHelper problem.

This was broken for cross-compilation and has probably been
unnecessary for several years now.

Also fix extraneous whitespace in related some tests.

Preparing to change the class name back to QPDFObject

When a QPDF is destroyed, changing indirect objects to direct nulls
makes them effectively disappear silently when they sneak into other
places. Instead, we should treat this as an error. Adding a destroyed
object type makes this possible.

This is in preparation for restoring a QPDFObject.hh to ease the
transition on qpdf_object_type_e.

This commit was created by
* Renaming QPDFObject.cc and QPDFObject.hh
* Replacing QPDFObject\b with QPDFValueProxy (where \b is word
  boundary)
* Running format-code
* Manually resorting files in libqpdf/CMakeLists.txt
* Manually refilling the comment in QPDF.hh near class Resolver

Allow QPDFObjectHandle::obj to be set prior resolving object.

ot_unresolved has been appended to the list object types in order to
preserve the output of existing test cases.

Report mem usage

Part of #729

Also, revert inlining of unparse and operator << from commit 4c6640c in
order to avoid exposing QUtil.

ABI breaking change

Also add stubs for top-level QPDF methods (createFromJSON,
updateFromJSON)

Bidirectional base64; will be used by JSON v2.

It was being run at configuration time, not build time.

* Change DLL_EXPORT to libqpdf_EXPORTS (internal to the build). The
  new name is cmake's default, is more conventional, and is less
  likely to clash with other symbols.
* Add QPDF_DLL_PRIVATE for non-Windows
* Make logic around when to define QPDF_DLL et al more explicit
* Add detailed comments