Add naïve loop to guess which encoding is used when email indicate wrong encoding

git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@24281 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5/Document/EmailDocument.py

@@ -129,6 +129,10 @@ class EmailDocument(File, TextDocument):
     for (name, value) in self._getMessage().items():
       for text, encoding in decode_header(value):
         if encoding is not None:
+          try:
+            text = text.decode(encoding).encode('utf-8')
+          except UnicodeDecodeError:
+            encoding = self._guessEncoding(text)
             text = text.decode(encoding).encode('utf-8')
         if name in result:
           result[name] = '%s %s' % (result[name], text)
@@ -596,6 +600,20 @@ class EmailDocument(File, TextDocument):
     """
     self.MailHost.send(message)
 
+  def _guessEncoding(self, string):
+    """
+    Some Email Clients indicate wrong encoding
+    This method try to guess which encoding is used.
+    """
+    from encodings.aliases import aliases
+    codec_list = set(aliases.values())
+    for codec in codec_list:
+      try:
+        string.decode(codec)
+      except (UnicodeDecodeError, IOError):
+        continue
+      return codec
+
 ## Compatibility layer
 #from Products.ERP5Type import Document
 #Document.MailMessage = EmailDocument