Filter out non-ASCII characters on invalid UTF8

When an enumerate is done, it is possible that the iDRAC may return
invalid UTF8 that contains non-ASCII characters.  This causes an
XMLSyntaxError to be thrown.  This fix detects that situation and
filters out all non-ASCII characters to bypass the error.

See the following bug for further details:
https://bugs.launchpad.net/python-dracclient/+bug/1779412

Closes-Bug: #1779412
Change-Id: I5003785dee922920dcdd95c8d7e2a26e0bf97a7d

dracclient/tests/test_wsman.py

@@ -19,6 +19,7 @@ import lxml.objectify
 import mock
 import requests.exceptions
 import requests_mock
+import six
 
 from dracclient import exceptions
 from dracclient.tests import base
@@ -55,6 +56,14 @@ class ClientTestCase(base.BaseTest):
         self.assertRaises(exceptions.WSManInvalidResponse,
                           self.client.enumerate, 'resource')
 
+    @requests_mock.Mocker()
+    def test_enumerate_with_invalid_utf8(self, mock_requests):
+        mock_requests.post('https://1.2.3.4:443/wsman',
+                           content=six.b('<result>yay!\xC0</result>'))
+
+        resp = self.client.enumerate('resource')
+        self.assertEqual('yay!', resp.text)
+
     @requests_mock.Mocker()
     def test_enumerate_with_auto_pull(self, mock_requests):
         mock_requests.post(

dracclient/wsman.py

@@ -12,6 +12,8 @@
 #    under the License.
 
 import logging
+import re
+import six
 import time
 import uuid
 
@@ -157,7 +159,14 @@ class Client(object):
                                     filter_query, filter_dialect)
 
         resp = self._do_request(payload)
-        resp_xml = ElementTree.fromstring(resp.content)
+        try:
+            resp_xml = ElementTree.fromstring(resp.content)
+        except ElementTree.XMLSyntaxError:
+            LOG.warning('Received invalid content from iDRAC.  Filtering out '
+                        'non-ASCII characters: ' + repr(resp.content))
+            resp_xml = ElementTree.fromstring(re.sub(six.b('[^\x00-\x7f]'),
+                                                     six.b(''),
+                                                     resp.content))
 
         if auto_pull:
             # The first response returns "<wsman:Items>"