--- a/cubicweb/hooks/__init__.py	Wed Mar 29 11:37:31 2017 +0200
+++ b/cubicweb/hooks/__init__.py	Wed Mar 29 11:45:19 2017 +0200
@@ -29,7 +29,7 @@
     events = ('server_startup',)
 
     def __call__(self):
-        if self.repo._scheduler is None:
+        if not self.repo.has_scheduler():
             return
         # XXX use named args and inner functions to avoid referencing globals
         # which may cause reloading pb
@@ -51,7 +51,7 @@
     events = ('server_startup',)
 
     def __call__(self):
-        if self.repo._scheduler is None:
+        if not self.repo.has_scheduler():
             return
         def update_feeds(repo):
             # take a list to avoid iterating on a dictionary whose size may
@@ -75,7 +75,7 @@
     events = ('server_startup',)
 
     def __call__(self):
-        if self.repo._scheduler is None:
+        if not self.repo.has_scheduler():
             return
         def expire_dataimports(repo=self.repo):
             for uri, source in repo.sources_by_uri.items():

--- a/cubicweb/server/repository.py	Wed Mar 29 11:37:31 2017 +0200
+++ b/cubicweb/server/repository.py	Wed Mar 29 11:45:19 2017 +0200
@@ -383,6 +383,12 @@
                 raise Exception('Is the database initialised ? (cause: %s)' % ex)
         return appschema
 
+    def has_scheduler(self):
+        """Return True if the repository has a scheduler attached and is able
+        to register looping tasks.
+        """
+        return self._scheduler is not None
+
     def run_scheduler(self):
         """Start repository scheduler after preparing the repository for that.
 
@@ -392,7 +398,7 @@
         XXX Other startup related stuffs are done elsewhere. In Repository
         XXX __init__ or in external codes (various server managers).
         """
-        assert self._scheduler is not None, \
+        assert self.has_scheduler(), \
             "This Repository is not intended to be used as a server"
         self.info(
             'starting repository scheduler with tasks: %s',
@@ -407,7 +413,7 @@
         """
         if self.config.repairing:
             return
-        if self._scheduler is None:
+        if not self.has_scheduler():
             self.warning(
                 'looping task %s will not run in this process where repository '
                 'has no scheduler; use "cubicweb-ctl scheduler <appid>" to '

--- a/doc/changes/3.25.rst	Wed Mar 29 11:37:31 2017 +0200
+++ b/doc/changes/3.25.rst	Wed Mar 29 11:45:19 2017 +0200
@@ -52,7 +52,11 @@
   scheduler, all cubicweb-ctl's "start" commands (i.e. ``start``, ``pyramid``,
   ``wsgi``) do not start repository *looping tasks manager* anymore, nor do
   they start the scheduler. Site administrators are thus expected to start
-  this scheduler as a separate process.
+  this scheduler as a separate process. Also, registering looping tasks (i.e.
+  calling ``repo.looping_tasks()``) is a no-op when the repository has no
+  scheduler set; a warning is issued in such cases. Application developers may
+  rely on repository's ``has_scheduler`` method to determine if they should
+  register a looping task or not.
 
 * In ``cubicweb.pyramid``, function ``make_cubicweb_application`` got renamed
   into ``config_from_cwconfig`` (950ce7d9f642).