Remove non-ascii named sample file.

It is causing problems with the Travis tests and while releasing,
at least when using check-manifest.

Products/feedfeeder/doc/feedconsuming.txt

@@ -171,9 +171,9 @@ kept.
   >>> duplicate2.text
   u'<div>\nwoo hoo, a party!\n</div>'
   >>> len(duplicate2.enclosures)
-  3
+  2
   >>> sorted([x.id for x in duplicate2.enclosures.values()])
-  [u'test2.doc', u'test3.xls', u't\xe8st4.odt']
+  [u'test2.doc', u'test3.xls']
 
 Get our item that talks about the thesis defence of Reinout.
 

Products/feedfeeder/tests/samples/samplefeed1.xml

@@ -42,7 +42,6 @@
   <id>http://myplace.com</id>
   <link type="application/pdf" href="file:test2.doc" rel="enclosure" length="737" />
   <link type="application/pdf" href="file:test3.xls" rel="enclosure" length="693" />
-  <link type="application/pdf" href="file:tèst4.odt" rel="enclosure" length="699" />
   <content type="xhtml">
 <div xmlns="http://www.w3.org/1999/xhtml">
 woo hoo, a party!