follow Python conventions for boolean values True and False might be singletons and the "default" values for "boolean" expressions, but "all" values in Python has a boolean value and should be evaluated as such. Checking with 'is True' and 'is False' is thus confusing, error prone and unnessarily complex. If we anywhere rely and nullable boolean fields from the database layer and don't want the null value to be treated as False then we should check explicitly for null with 'is None'.

--- a/rhodecode/lib/celerylib/__init__.py	Thu Mar 28 01:10:45 2013 +0100
+++ b/rhodecode/lib/celerylib/__init__.py	Thu Mar 28 01:10:45 2013 +0100
@@ -124,7 +124,7 @@
             ret = func(*fargs, **fkwargs)
             return ret
         finally:
-            if CELERY_ON and CELERY_EAGER is False:
+            if CELERY_ON and not CELERY_EAGER:
                 meta.Session.remove()
 
     return decorator(__wrapper, func)