Fix bug that can permit someone to kill job array belonging to another user.

If a job is submitted with the option --mem=0, then once the job
is allocated resources, reset that value to the amount of memory
on the allocated node with the smallest amount of memory. This is
probably the best option for right now. The downside is that if
the job is requeued, its memory size will no longer be the total
memory on all nodes if the cluster is heterogeneous. On the upside,
some memory data is available for accounting purposes.
bug 1526

Fix support for --mem=0 (all memory of a node) with select/cons_res plugin.
Previously the job would not have it's memory constrained, but it would not
make that memory unavailable to other jobs either.
bug 1526

Conflicts:
	src/sbatch/opt.c

This was previously added to v14.11

Added new SchedulerParameters value of bf_min_age_reserve. The backfill
scheduler will not reserve resources for pending jobs until they have
been pending for at least the specified number of seconds. This can be
valuable if jobs lack time limits or all time limits have the same value.

Increase maximum MaxArraySize configuration parameter value from 1,000,001
to 4,000,001.
bug 1535

The data could be purged after job completion

Cray - Fix for launching batch step within an existing job allocation.
bug 1509

Change check of batch step ID from NO_VAL to SLURM_BATCH_SCRIPT
  for better consistency.

Conflicts:
	src/slurmd/slurmstepd/task.c
	src/squeue/print.c

This should not be a problem in v14.03, but the logic will be
  important in v14.11 due to larger job arrays being supported.

Make sure variables are initialized before passing a pointer to
  them as a function argument. The variable is set in the underlying
  function, so this is mostly cosmetic.

Partially revert commit 8d91ae22
The bug was introduced in version 14.11.0-pre4.
bug 1504

This is for bug 1514

before.

Bug 1507

fails

addition to commit 2c83bf4e

message comes in

Bug 1501

cluster(s) requested.

Bug 1492

For job running under a debugger, if the exec of the task fails, then
cancel its I/O and abort immediately rather than waiting 60 seconds for
I/O timeout.