diff --git a/src/z3c/rml/attr.py b/src/z3c/rml/attr.py
index e09d440..f67b7ce 100644
--- a/src/z3c/rml/attr.py
+++ b/src/z3c/rml/attr.py
@@ -17,6 +17,7 @@
 """
 __docformat__ = "reStructuredText"
 import cStringIO
+import os
 import re
 import reportlab
 import reportlab.lib.colors
@@ -164,17 +165,35 @@ def convert(self, value, context=None):
 
 class Image(Text):
 
-    open = urllib.urlopen
+    open = staticmethod(urllib.urlopen)
+    packageExtract = re.compile('^\[([0-9A-z_.]*)\]/(.*)$')
 
     def __init__(self, name=None, default=DEFAULT, onlyOpen=False):
         super(Image, self).__init__(name, default)
         self.onlyOpen = onlyOpen
 
     def convert(self, value, context=None):
+        # Check whether the value is of the form:
+        #    [<module.path>]/rel/path/image.gif"
+        if value.startswith('['):
+            result = self.packageExtract.match(value)
+            if result is None:
+                raise ValueError(
+                    'The package-path-pair you specified was incorrect')
+            modulepath, path = result.groups()
+            module = __import__(modulepath, {}, {}, (modulepath))
+            value = os.path.join(os.path.dirname(module.__file__), path)
+        # Open/Download the file
         fileObj = self.open(value)
         if self.onlyOpen:
             return fileObj
-        return reportlab.lib.utils.ImageReader(fileObj)
+        # ImageReader wants to be able to seek, but URL info objects can only
+        # be read, so we make a string IO object out of it
+        sio = cStringIO.StringIO()
+        sio.write(fileObj.read())
+        fileObj.close()
+        sio.seek(0)
+        return reportlab.lib.utils.ImageReader(sio)
 
 
 class Color(Text):
@@ -184,12 +203,12 @@ def convert(self, value, context=None):
         if value in ALL_COLORS:
             return ALL_COLORS[value]
         # Decimal triplet
-        rgb = value.split(',')
+        rgb = Sequence(valueType=Float(), length=3).convert(value)
         if len(rgb) == 3:
-            return (float(num) for num in rgb)
+            return [float(num) for num in rgb]
         # Hexdecimal triplet
         if value.startswith('#'):
-            return (float(int(value[i:i+1], 16)) for i in range(1, 7, 2))
+            return [float(int(value[i:i+1], 16)) for i in range(1, 7, 2)]
         raise ValueError('%r not a valid color.' %value)
 
 
@@ -267,6 +286,8 @@ def __init__(self):
         super(TextNode, self).__init__('TEXT')
 
     def get(self, element, default=DEFAULT, context=None):
+        if element.text is None:
+            return u''
         return unicode(element.text).strip()
 
 class FirstLevelTextNode(TextNode):
diff --git a/src/z3c/rml/canvas.py b/src/z3c/rml/canvas.py
index 659d0ef..6174e46 100644
--- a/src/z3c/rml/canvas.py
+++ b/src/z3c/rml/canvas.py
@@ -247,9 +247,8 @@ def process(self):
             if subElement.tail is not None:
                 self.processPoints(subElement.tail)
 
-        if kw.get('close', False):
+        if kw.pop('close', False):
             self.path.close()
-            del kw['close']
 
         self.context.drawPath(self.path, **kw)
 
diff --git a/src/z3c/rml/chart.py b/src/z3c/rml/chart.py
index 2987d84..2aca97a 100644
--- a/src/z3c/rml/chart.py
+++ b/src/z3c/rml/chart.py
@@ -267,7 +267,7 @@ class ValueAxis(Axis):
         attr.Bool('forceZero'), # TODO: Support 'near'
         attr.Measurement('minimumTickSpacing'),
         attr.Int('maximumTicks'),
-        attr.Text('labelTextFormat'),
+        attr.Attribute('labelTextFormat'),
         attr.Text('labelTextPostFormat'),
         attr.Float('labelTextScale'),
         attr.Float('valueMin'),
diff --git a/src/z3c/rml/rml2pdf.py b/src/z3c/rml/rml2pdf.py
index 825c5a5..f120bb5 100644
--- a/src/z3c/rml/rml2pdf.py
+++ b/src/z3c/rml/rml2pdf.py
@@ -16,6 +16,7 @@
 $Id$
 """
 __docformat__ = "reStructuredText"
+import cStringIO
 import os
 import sys
 import zope.interface
@@ -25,6 +26,15 @@
 zope.interface.moduleProvides(interfaces.IRML2PDF)
 
 
+def parseString(xml):
+    root = etree.fromstring(xml)
+    doc = document.Document(root)
+    output = cStringIO.StringIO()
+    doc.process(output)
+    output.seek(0)
+    return output
+
+
 def go(xmlInputName, outputFileName=None, outDir=None, dtdDir=None):
     if dtdDir is not None:
         sys.stderr.write('The ``dtdDir`` option is not yet supported.')