The first case here is even more fun... let's just let
xstrfmtcat() sort this all out instead of manually determing
the required buffer size.

Functionality is still available if required, but has been deprecated
since 17.11, and has been removed on master before 19.05.

Un-document is now lest people start trying to use it.

Bug 5712.

bugs 5682 and 4584

No change in logic

The output of "scontrol show node" changed with respect to GPU bound
sockets. There was no error reported by test20.4, but the output was
wrong without this commit.

# Conflicts:
#	NEWS

support.

Bug 4871

is removed from the system.

Bug 5665

burst buffer.

Bug 5665

block distributions are failing with some configurations
ompi 4.0.x (head as of 10 sept 2018)
slurm 18.08
PMIx 3.0
The test with a block distribution is hanging. This will let the
remaining 3 MPI steps in test1.18 run to completion while we wait
for a fix from ompi.
Note that Alex and Danny are not seeing this failure, but they
might not be working off the latest ompi code.

test39.12 was failing with some configurations by double counting
available cores. Moving some code fixes that problem and speeds
up the logic as well.

Specifically log Cores/CPUs related errors using proper name,
not logging bad CPUs information as bad Core information.

Log SLURMD_HOSTNAME on remote host rather than "hostname", which
is more meaningful on a multi-slurmd configuration

or a database admin.

Bug 5679

Make tests work properly if GPU count exceeds CPU count on a node

The GRES information now contains socket binding information,
necessitating this change to the parsing function

Bug 5359

Bug 5679

regression from combination of 037bb3cb and 5a6a26f6

Bug 4479

node_registration_status_msg.

Bug 5672

Bug 5672

Coverity CID 164845