Soliciting advice re german umlauts bug in XML_Feed_Parser

Greetings,

Many moons ago, this bug was reported
http://pear.php.net/bugs/bug.php?id=12916 against XML_Feed_Parser.

A patch was uploaded, and a few weeks ago it was applied to the code base
by Daniel O'Connor. However, this code still fails the tests as seen here:
http://test.pear.php.net:8080/job/XML_Feed_Parser/lastBuild/testReport/junit/(root)/Regressions/test_handlesGermanUmlauts/

The best assumption I have is that the sample xml file for this test is
crap because however I view that file, the target characters in the sample
look "wrong"; either silly boxen or "euro, cross, ellipsis, big funky s,
big funky y, small funky s, overlapping s".

https://github.com/pear/XML_Feed_Parser/blob/master/tests/regressions.php#L67is
the test,
https://github.com/pear/XML_Feed_Parser/blob/master/samples/bug12916.xml is
the sample.

How valid is the assumption? Suggestions for a better fix?

Thanks in advance.
-M