Respect character references in the xref target
Since AsciiDoc reserves numerous characters for formatting, it's sometimes necessary to escape those characters using a character reference (e.g., +
). A prime example is ++
, which must be escaped using {pp}
. However, when {pp}
is used in the target of an xref, Antora gets tripped up by the #
in the character reference, mistakingly interpreting it as the start of the URL fragment. Antora should decode these character references first, then resolve the target, then restore HTML special characters (so as not to break the HTML).