The test framework used by javadoc does not sanitize the output, resulting in malformed
unicode escapes causing jtreg's junit converter to fail.

parent.document.title="Overview (Testing \\"Window \\'Title\\'\\" with a \\\\ backslash and a / forward slash and a \\u00E8 unicode char also a tab and also a \\t special character another \\u0002 unicode))"; in 8016675/overview-summary.html$
$
Running subtest #2... Passed$
not found:$
parent.document.title="Overview (Testing "Window 'Title'" with a \\ backslash and a / forward slash and a \u00e8 unicode char also a tab and also a ^I special character another \u0002 unicode))"; in 8016675/overview-summary.html$

Note in the second case we have a single \ instead of \\.

Attached is the jtr, in order to test this to make sure that there are no illegal
characters, run jtreg -xml:verify <all-the-other-usual-options>