prevent sending invalid XML chars in chat / groupchat and status. Fixes #3296

src/common/helpers.py

@@ -645,6 +645,11 @@ def ensure_utf8_string(string):
 		pass
 	return string
 
+def remove_invalid_xml_chars(string):
+	if string:
+		string = re.sub(gajim.interface.invalid_XML_chars_re, '', string)
+	return string
+
 def get_windows_reg_env(varname, default=''):
 	'''asks for paths commonly used but not exposed as ENVs
 	in english Windows 2003 those are:

src/dialogs.py

@@ -608,6 +608,7 @@ class ChangeStatusMessageDialog:
 			beg, end = self.message_buffer.get_bounds()
 			message = self.message_buffer.get_text(beg, end).decode('utf-8')\
 				.strip()
+			message = helpers.remove_invalid_xml_chars(message)
 			msg = helpers.to_one_line(message)
 			if self.show:
 				gajim.config.set('last_status_msg_' + self.show, msg)

src/gajim.py

@@ -2416,6 +2416,10 @@ class Interface:
 		# at least one character in 3 parts (before @, after @, after .)
 		self.sth_at_sth_dot_sth_re = re.compile(r'\S+@\S+\.\S*[^\s)?]')
 
+		# Invalid XML chars
+		invalid_XML_chars = u'[\x00-\x08]|[\x0b-\x0c]|[\x0e-\x19]|[\ud800-\udfff]|[\ufffe-\uffff]'
+		self.invalid_XML_chars_re = re.compile(invalid_XML_chars)
+
 		re.purge() # clear the regular expression cache
 
 	def on_emoticon_sort(self, emot1, emot2):

src/message_control.py

@@ -20,6 +20,7 @@
 import gtkgui_helpers
 
 from common import gajim
+from common import helpers
 
 # Derived types MUST register their type IDs here if custom behavor is required
 TYPE_CHAT = 'chat'
@@ -154,6 +155,9 @@ class MessageControl:
 		# Send the given message to the active tab.
 		# Doesn't return None if error
 		jid = self.contact.jid
+
+		message = helpers.remove_invalid_xml_chars(message)
+
 		original_message = message
 		conn = gajim.connections[self.account]