lots of cleanup

Changed node_mgr.c to pack config information (rather than actual node
data) iff FAST_SCHEDULE flag is set. The get node info call will return
the data actually used for scheduling purposes, which may not be the
data associated with each individual node.

specified at submit time start with a value of 1<<16 and DPCS uses
values 1 to 0xffffffff. This preclude effective use of a max job_id
in search function, which was able to quickly indicate that a job_id
requested by DPCS was not already in use. I added a job_id hash table
to restore quick job find operations.
Added check of packstr function to insure string would not go over
64k bytes, which is the limit for the protocol used.

specified at submit time start with a value of 1<<16 and DPCS uses
values 1 to 0xffffffff. This preclude effective use of a max job_id
in search function, which was able to quickly indicate that a job_id
requested by DPCS was not already in use. I added a job_id hash table
to restore quick job find operations.

Function was renamed.

Added a few extra helper funtions

Revised job API descriptions per feedback.

Added support for looping on datagram receives until all the data that is expect on the encapsulated stream has arrived.
This should have been in the original implementation.

Implemented slurm_send_controller_buffer

NOTE:  primary and secondary controller addresses are currently #defined
this needs to be read from a config file.  SLURM_PORT is also #define to 7000

inclusion of autoconf socket #ifdefs

renamed set_slurm_addr_hton to slurm_set_addr_uint

adjusted arithmetic for packing wrapper routines

   slurmctld.h and api/slurm.h respectively.

   most of the stuff in these files did not belong in common, it has been
   moved elsewhere.

   these are the slurmctld-specific portions of what used to be
   src/common/bits_bytes.h

   what used to be src/common/bits_bytes.c