Merge pull request #506 from garlick/sched_writer

rfc20, rfc40: define scheduling.writer

Author: mergify[bot]

spec_20.rst

@@ -223,8 +223,19 @@ R Format
 
 .. data:: scheduling
 
-  (*dictionary*, OPTIONAL) Scheduler-specific resource data which SHOULD NOT
-  be interpreted other Flux components.  When used, it SHALL ride along on the
+  (*dictionary*, OPTIONAL) Scheduler-specific resource data with the
+  following keys:
+
+  .. data:: writer
+
+    (*string*, OPTIONAL) If provided, a URI whose scheme identifies the
+    scheduler that produced the :data:`scheduling` key.  Remaining URI
+    components SHALL be interpreted by the named scheduler.  If not
+    provided, a value of ``fluxion`` SHALL be assumed.
+
+  Remaining scheduler-specific keys SHOULD be ignored by other Flux components.
+
+  When used, :data:`scheduling` SHALL ride along on the
   resource acquisition protocol (RFC 28) and resource allocation protocol
   (RFC 27) so that it may be included in static configuration, allocated to
   jobs, and passed down a Flux instance hierarchy.

spec_40.rst

@@ -56,6 +56,13 @@ scheduler within the ``scheduling`` key of *R*.
 Implementation
 **************
 
+The :data:`writer` key of the :data:`scheduling` dict SHOULD be set to
+to ``fluxion:jgf`` when this specification is used to create the extended
+resource data.  To support older scheduler data, RFC 20 states that
+if :data:`writer` is not not provided, a value of ``fluxion`` SHALL be
+assumed. This specification further specifies that if :data:`writer` or
+a URI path are not provided, the URI path SHALL be assumed to be ``jgf``.
+
 The Fluxion resource graph representation allows RFC4-compliant schedulers to
 serialize any subset of graph resource data into its value and later
 deserialize this value with no data loss. The ``scheduling`` key contains a