Properly handle content type text/plain

planet/reconstitute.py

@@ -16,7 +16,6 @@ Todo:
 import re, time, md5, sgmllib
 from xml.sax.saxutils import escape
 from xml.dom import minidom, Node
-from BeautifulSoup import BeautifulSoup
 from planet.html5lib import liberalxmlparser, treebuilders
 import planet, config
 
@@ -139,9 +138,17 @@ def content(xentry, name, detail, bozo):
     xdiv = '<div xmlns="http://www.w3.org/1999/xhtml">%s</div>'
     xdoc = xentry.ownerDocument
     xcontent = xdoc.createElement(name)
+
     if isinstance(detail.value,unicode):
         detail.value=detail.value.encode('utf-8')
 
+    if not detail.has_key('type') or detail.type.lower().find('html')<0:
+        detail['value'] = escape(detail.value)
+        detail['type'] = 'text/html'
+
+    if detail.type.find('xhtml')>=0 and not bozo:
+        data = minidom.parseString(xdiv % detail.value).documentElement
+    else:
         parser = liberalxmlparser.XHTMLParser(tree=treebuilders.dom.TreeBuilder)
         html = parser.parse(xdiv % detail.value, encoding="utf-8")
         for body in html.documentElement.childNodes: