Adobe Reader fails to decode that URL but we did it in 2.0.7. Tests by Tim showed 38 ocurrences of "http://www.nap.edu/catalog/11140.html" lost in file 534948.pdf (attached file is reduced) in the last regression test. This is because of the changes in
PDFBOX-3913. I'll fix that and support UTF-16 (LE) as well and add a test.