Rely on include_meta_content_type parameter of lxml.html.tostring API to setup correct

meta tag.
Previous implementation was not adding the content-type meta tag in all cases.


git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@45321 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5OOo/transforms/html_to_odt.py

@@ -32,17 +32,8 @@ class HTMLToOdt:
   def convert(self, orig, data, cache=None, filename=None, context=None, **kwargs):
     # Try to recover broken HTML documents, specially regarding encoding used
     html_node = etree.XML(orig, parser=html_parser)
-    html_tree = html_node.getroottree()
-    head = html_tree.find('head')
-    if head is None:
-      # This part of code is supposed to be useless
-      # lxml.html.tostring function with include_meta_content_type
-      # parameter to True, should do the same things. But it does not.
-      head = Element('head')
-      html_node.insert(0, head)
-      SubElement(head, 'meta', **{'http-equiv': 'Content-Type',
-                                  'content': 'text/html; charset=utf-8'})
-    orig = html.tostring(html_node, encoding='utf-8', method='xml')
+    orig = html.tostring(html_node, encoding='utf-8', method='xml',
+                         include_meta_content_type=True)
 
     # workaround a Bug in LibreOffice HTML import filter.
     # https://bugs.freedesktop.org/show_bug.cgi?id=36080

product/ERP5OOo/transforms/oood_commandtransform.py

@@ -194,7 +194,8 @@ class OOOdCommandTransform(commandtransform):
           parent_node.append(style_node)
           style_node.attrib.update({'type': 'text/css'})
           parent_node.remove(css_link_tag)
-    xml_output = html.tostring(xml_doc, encoding='utf-8', method='xml')
+    xml_output = html.tostring(xml_doc, encoding='utf-8', method='xml',
+                               include_meta_content_type=True)
     xml_output = xml_output.replace('<title/>', '<title></title>')
     return xml_output