When xrefitem, or its specializations, is used some extra XML elements
are generated which were not properly parsed; this commit adds the extra
parsing required.

* properly parses a variable list element, which consists of a set of
  varlistentry + listitem elements
* docTitleTypeSub was extended to to parse the "ref" and "anchor"
  elements which exist inside a varlistentry.

Signed-off-by: Fabio Utzig <[email protected]>

When xrefitem or any of its specializations is used, Doxygen generates a
new file of kind="page"; this is very similar to a "group" or
"namespace" file.

This commit adds proper filter handling and a new directive, doxygenpage,
that can be used to include documentation from a page block.

Signed-off-by: Fabio Utzig <[email protected]>

Signed-off-by: Fabio Utzig <[email protected]>

When a content block was not found, the message would imply it is of
type "namespace"; fix it so that the proper "kind" is displayed
("namespace", or "group", or "page").

Signed-off-by: Fabio Utzig <[email protected]>

The elements "docvariablelist", "docvarlistentry" and "docanchor" are
used whenever an xrefitem is used in a project, and the proper compound
of kind="page" is generated.

This commits adds rendering of the elements in a definition block, and
properly adds a target to anchor elements so they can be referenced
later.

Signed-off-by: Fabio Utzig <[email protected]>

Mimic Doxygen html output, by adding a pending_xref to xrefitem titles,
which can later be resolved to the generated xrefsect page, if it was
added to the project.

Signed-off-by: Fabio Utzig <[email protected]>

Add initial documentation with "doxygenpage" usage relying on previously
added xrefsect examples.

Signed-off-by: Fabio Utzig <[email protected]>