Refresh proxy settings on failure

1. When a user is connected to a network (ex: office network) using a laptop the nuxeo-drive is using the current network's proxy settings (thru system settings).
2. When the user leaves the current network and joins another network (home network) the nuxeo-drive stops working as it is still using the previous network proxy settings.
3. This is because the proxy settings are modified for the system but not in the nuxeo-drive settings.

Expected: nuxeo-drive should retrieve latest proxy settings from system and use it. I think if we do it when connection failure / network connectivity failure happens then nuxeo-drive can recover quickly after a network migration.

If we clear the remote document client, remote client caches in the engine whenever engine goes offline, then proxy server will be re-computed from system settings. This will enable nuxeo-drive to automatically recover faster. Example code change (egine.set_offline method):

diff --git a/nuxeo-drive-client/nxdrive/engine/engine.py b/nuxeo-drive-client/nxdrive/engine/engine.py
index 618e9bb..574c9a3 100644
--- a/nuxeo-drive-client/nxdrive/engine/engine.py
+++ b/nuxeo-drive-client/nxdrive/engine/engine.py
@@ -262,6 +262,11 @@ class Engine(QObject):
         if value:
             log.debug("Engine %s goes offline", self._uid)
             self._queue_manager.suspend()
+            proxy_settings = self._manager.get_proxy_settings()
+            if proxy_settings.config == 'System':
+                # Modification of system proxy settings, or changing network can cause engine to go offline. 
+                # So clear the client cache, to facilitate re-compute of proxy servers again.
+                self.invalidate_client_cache()
             self.offline.emit()
         else:
             log.debug("Engine %s goes online", self._uid)