[xml] Converting xmlDocPtr back to text loses PI's

[Shane Adams <adamsch1 yahoo com>]

Hello - 
 
 I'm processing an xml document that contains <?xml-stylesheet ...> entries amongst others.  I've parsed this 
document and manipulated the tree to do some other processing (cutting out some subchildren).  Anyway I'm 
using xmlNodeDumpOutput to regenerate the document.
 
 The problem I'm encountering is that PI's are not present in the resulting document.  In case it helps the 
document is parsed using 
 xmlSAXParseMemoryWithData.
 
 Any pointers?  I've looked in the usual places and am posting here as my fallback
 
 
 Cheers,
   Shane