OOo after refresh can decide to cleanup text:span elements

Take into account this behaviour.

Before refresh:
  <text:reference-mark-start text:name="week"/>
      <text:span>Monday</text:span>
    <text:reference-mark-end text:name="week"/
After refresh:
  <text:reference-mark-start text:name="week"/>Monday<text:reference-mark-end text:name="week"/>


git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@41625 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5OOo/tests/testFormPrintoutAsODT.py

@@ -259,10 +259,16 @@ class TestFormPrintoutAsODT(TestFormPrintoutMixin):
     content_tree = etree.XML(content_xml)
     span_list = content_tree.xpath('//text:reference-mark-start[@text:name="week"]/following-sibling::text:span',
                                    namespaces=content_tree.nsmap)
-    self.assertEquals(1, len(span_list))
-    span = span_list[0]
-    self.assertEquals('line1', span.text)
-    self.assertEquals('line2', span[0].tail)
+    if span_list:
+      self.assertEquals(1, len(span_list))
+      span = span_list[0]
+      self.assertEquals('line1', span.text)
+      self.assertEquals('line2', span[0].tail)
+    else:
+      reference_mark_node = content_tree.xpath('//text:reference-mark-start[@text:name="week"][1]',
+                                               namespaces=content_tree.nsmap)[0]
+      self.assertEquals('line1', reference_mark_node.tail)
+      self.assertEquals('line2', reference_mark_node.getnext().tail)
     self._validate(odf_document)
 
   def test_01_Paragraph_08_Field_Format(self):