This patch fixes Bug #1137, although it may not be the most straightforward way to do so.
I wasn't sure exactly what changed in the rendering of the HTML from Xplore, nor was I sure the exact format of the names that was supposed to be there before entry.clean() gets called; from what I could tell, the "author" matching pattern was stopping at a (new?) newline, before any of the authors' names were given ... these names were wrapped in HTML tags that were too long to be stripped by HTMLConverter, so I had to increase that MAX_LENGTH constant, too.
I didn't check with as many search terms as I should have, I guess. Some papers with very long names will require a larger MAX_TAG_LENGTH constant... see the search for
"Chintalapudi, Krishnakant Kant", for example.
This patch updates the previous one, with a MAX_TAG_LENGTH = 100.
Applied in the master branch