in an int instead of a short (truncation was causing strange results,
   especially when ORed with NODE_NO_RESPOND flag).
 o No longer bother processing the '*' at end of node state strings.
   sinfo -t now shows both responding and non-responding nodes by
   default --responding or --dead can be used to choose one or the other.

switch (was setting DRAINED nodes in state DRAINING).

option. Use SHOW_ALL as flag.

stack size to an appropriate size. System dies on AIX without this due
to too small of a stack.

Add bnr function descriptions directly to bnr.h.

responding in a timely fashion) then send a registration RPC
to slurmctld.

are allocated. Replace polling to test for allocation with processing of
allocation response RPC from slurmctld at allocation time.

 o Use job_desc_msg_create_from_options() to build job_desc_msg for
   batch jobs (less places to update in future)

   to specified user.
 o New srun option --no-shell
 o Changed uid/gid_from_name () to uid/gid_from_string () and allow
   either user/group name or numerical user/group ID.
 o updated srun(1) documentation with new --no-shell option and
   changed semantics for --uid.

registration (in particular after changing state from UNKNOWN to DRAINING,
an inactive node was not registering to reset its state to DRAINED)