HTML 5 tags are not supported by the PDF export

The PDF export currently doesn't support HTML 5 tags that could for example be introduced by the figure or HTML macro. They are removed using HTMLCleaner, but even if it didn't remove them they would still need to be supported by the transformation. The content of these tags is still exported, there is just no formatting supplied for figures for example.

This should be solved with the new client-side PDF export, see XWIKI-19268.