ConfluenceXHTMLParser does not repair CDATA containing square brackets

The Confluence XML Export creates nested CDATA sections and handles them by "damaging" the inner CDATA sections by inserting a space between the closing ]] and > .

When importing, the ConfluenceXHTMLParser tries to make up for that, but the attempt to repair fails if the text inside the CDATA section contains one or several closing square brackets.

(As far as I can see, the export simply adds a space behind every two consecutive closing square brackets. This might even make the job of repairing them simpler.)