allow user.url to be accessed without the server running

Reduces the number of different ways we need to build the same URLs.

jupyterhub/handlers/login.py

@@ -43,7 +43,7 @@ class LoginHandler(BaseHandler):
         if user:
             if not next_url:
                 if user.running:
-                    next_url = user.server.base_url
+                    next_url = user.url
                 else:
                     next_url = self.hub.server.base_url
             # set new login cookie

jupyterhub/handlers/pages.py

@@ -75,10 +75,7 @@ class SpawnHandler(BaseHandler):
             self.finish(self._render_form())
         else:
             # not running, no form. Trigger spawn.
-            # Creating the URL manually since the server does not
-            # exist yet
-            url = url_path_join(self.base_url, 'user', quote(user.name))
-            self.redirect(url)
+            self.redirect(user.url)
 
     @web.authenticated
     @gen.coroutine

jupyterhub/user.py

@@ -186,10 +186,10 @@ class User(HasTraits):
         if self.settings.get('subdomain_host'):
             return '{host}{path}'.format(
                 host=self.host,
-                path=self.server.base_url,
+                path=self.base_url,
             )
         else:
-            return self.server.base_url
+            return self.base_url
     
     @gen.coroutine
     def spawn(self, options=None):

share/jupyter/hub/templates/home.html

@@ -10,7 +10,7 @@
       {% endif %}
       <a id="start" class="btn btn-lg btn-success"
       {% if user.running %}
-        href="{{ user.server.base_url }}"
+        href="{{ user.url }}"
       {% else %}
         href="{{base_url}}spawn"
       {% endif %}