- PstreamBuffers for parallel VTK output.
  - avoids MPI overflows for larger meshes

- new vtk::lineWriter for writing edges

- vtk::fileWriter::writeUniform now also supports processor-specific
  uniform values instead of assuming that they are identical everywhere.

- reduces common boilerplate, ensures that label fields are
  more easily handled.

- reduces code, simplifies creation of new, specialized polyData
  writers.

- new templated vtk::GenericPatchWriter, which adds support for
  writing both uindirectPrimitivePatch + indirectPrimitivePatch types.

- handle geometric fields separately from regular fields with
    * vtk:GenericPatchGeoFieldsWriter
    * vtk:indirectPatchGeoFieldsWriter
    * vtk:uindirectPatchGeoFieldsWriter

- returns a range of `int` values that can be iterated across.
  For example,

      for (const int proci : Pstream::allProcs()) { ... }

  instead of

      for (label proci = 0; proci < Pstream::nProcs(); ++proci) { ... }

- new vtk::internalMeshWriter, vtk::patchMeshWriter
  intermediate classes without finiteVolume dependencies.

  Enables direct use with a polyMesh.
  Makes vtk::internalWriter, vtk::patchWriter header/template only.

- An identity is often useful when generating connectivity and offset
  information.

- The optional repeat value for vtk::write() allows it to also be used
  as a fill method.

- can be useful for diagnosing mesh internals and the locations of
  decomposed cells.

- could be triggered if running in parallel, but requesting procIDs
  to be written with a non-parallel version of the writer.

- the problem occurred when running the writers in a parallel solver
  or utility but requesting output on the master only.

  Adjusted the logic to avoid globalIndex for these cases.
  Previously the if (parallel_) checks were happening later, after the
  globalIndex had already been created.

- could be triggered if running in parallel, but requesting procIDs
  to be written with a non-parallel version of the writer.

- removed vtk::writeField templates. No longer used.

- parallel list output for foamVtkOutput

- simplified '.series' file output

- beginDataArray() method instead of openDataArray() + closeTag()
  since this seems to be the most common use anyhow.
  With an optional argument for leaving the tag open, this works the
  same as openDataArray() which may be deprecated in the future.

- begin/end methods for CellData, PointData, FieldData (commonly used)

- templating parameters for file headers, content version,
  legacy fields. This improves coding robustness and convenience of use.

- use formatter and higher-level methods for legacy output

- attribute quoting character now part of the formatter itself
  instead of as an argument for xmlAttr().
  Toggle with quoting() method.

- pair-wise processing of xml attributes, which also allows them to be
  passed as optional entries when creating an xml tag.

- xmlComment with multiple arguments

- elminate the foamVtkFormatter operator() in favour of xmlAttr.
  Improves readability and the purpose is clearer.

- with the xml append format it is possible to write raw binary
  (instead of base64), but the writer becomes more complicated.
  Either needs two passes to create, or need to allocate a block
  of space for the header information (like VTK itself does) and
  write later.

    * internalWriter
    * patchWriter
    * surfaceMeshWriter
    * lagrangianWriter

 Also these special purpose ones:
    * foamVtkWriteSurfFields

- this shifts responsibility away from caller to the individual writers
  for knowing which file formats are supported and which file ending is
  appropriate. When the writer receives the output format request,
  it can elect to downgrade or otherwise adjust it to what it can
  actually manage (eg, legacy vs xml vs xml-append).

  But currently still just with legacy format backends.