XML_ERROR_JUNK_AFTER_DOC_ELEMENT - How to resolve

Hi,

I've just started working with expat.  I have xml files that are gzipped.  I
gzcat them and pipe them to my parser.
I am getting the XML_ERROR_JUNK_AFTER_DOC_ELEMENT error message and I'm
unsure how to resolve.  I was under the impression that it
is caused by extracter character after a document root close tag.  I tried
stripping the chars after that close tag but that doesnt seem to work.  Is
this caused by a new document starting immediately after the first one has
finished?

Does anyone have any suggestions?

Here is the error message and what was in the buffer:

Parse error:file:1:row:4:column:0:reason:junk after document element
BUFFER = nter></usage></dataSet></metrics>

<?xml version='1.0' encoding='UTF-8'?>
<metrics version="3.0" cr



My main loop where I read stdin and call the parser is below:

/***********************************************************************/
  /* Read stdin                                                          */
  /***********************************************************************/
  for (;;) {
    len = (int)fread(buff, 1, BUFFSIZE-1, stdin);
    if (ferror(stdin)) {
      fprintf(stderr,"Error reading stdin\n");
      exit(-2);
    }
    done = feof(stdin);
    //if nothing read then exit so AI doesnt blow up
    if ((len == 0) && (done) && (cur_file_num==0))
       break;

   if(XML_Parse(p, buff, strlen(buff), done) == XML_STATUS_ERROR){
          fprintf(stderr, "\nParse error at
host:%s:file:%d:row:%d:column:%d:reason:%s\n",
                 host, cur_file_num, XML_GetCurrentLineNumber(p),
                 XML_GetCurrentColumnNumber(p),
XML_ErrorString(XML_GetErrorCode(p)));
          fprintf(stderr,"BUFFER = %s\n",buff);
          exit(-3);
     }
    if(done)
      break;
    }

    /* Free memory used by the parser */
    if(p) {
      XML_ParserFree(p);
    }
  return 0;
}