Given a simple XML document like the following one:

<?xml version=\"1.1\"?>
<ROOT>
   <FILLER>ffffffff...fffff</FILLER>
   <TEST>content</TEST><TEST2>content2</TEST2>
   <FILLER>ffffffff...fffffffff</FILLER>
</ROOT>

the content of the TEST elements may be overwritten by content from the following elements (i.e. 'f's) if the content of the TEST elemts is at the border of an entity scanner buffer (XMLEntityScanner uses 8192 byte buffers) and there is still enough content to read after the TEST elements such that another buffer read is required.