Make "qsub -I" work (no arguments)
Some fixes for node/task count logic for interactive jobs

This bug in the new plugin would result in memory corruption.

Doing so causes confusion for Intel MPI

Add support for "on" and "before*" options

bugfixes and the new item extraction feature.

Set PBS_ACCOUNT, PBS_ENVIRONMENT, and PBS_QUEUE only for batch
jobs and only if the user submission sets the account and partition.

Add missing close quote

Also log discarded PBS options

No support for PBS "before" options

in qsub wrapper and #PBS options in sbatch scripts

-l accelerator=true|false	(GPU use)
-l mpiprocs=#	(processors per node)
-l naccelerators=#	(GPU count)
-l select=#		(node count)
-l ncpus=#		(task count)
-v key=value	(environment variable)
-W umask=#		(set job's umask)
Note: the -v option does NOT support quoted commas.

-l accelerator=true|false	(GPU use)
-l mpiprocs=#		(processors per node)
-l naccelerators=#	(GPU count)
-l select=#		(node count)
-l ncpus=#		(task count)
-v key=value		(environment variable)
-W umask=#		(set job's umask)

If a user job tries to load his environment remotely and the slurmd
is not running as root, the previous logic would generate a fatal
error, now it just logs the error and proceeds with the passed user
environment.

that ended in the OverTimeLimit interval.

The DEFAULT node/frontend/partition values will add to previous
default values and not reinitialize.
bug 397

someone.  They can still add or remove a user of any admin level.

Previously nodes in FUTURE state would have no name field sent,
resulting in siview core dumping

bug 403

bug 403

bug 403