XML-invalid characters in Lingbuzz article metadata bleed through to the feed, making it unparseable #2
Description
Currently the feed contains broken XML that fails to parse in readers, because it pulls through the U+0002 character included in the keywords of this LingBuzz article, and that isn't a valid character for XML 1.0. (If you look at the page source, the "whitespace" in the middle of the keyword 'indefinite classifier' actually contains U+0002).
I'm surprised that this isn't taken care of by underlying 'rss' or 'xml' Node package, but it seems that the invalid character bleeds through.