Release 1.6.1 to dev

- Add extra job statuses that Slurm parser accepts as they represent
  jobs that will still have used resrouces.
- Add matching tests for all possible job statuses.

- Add NODE_FAIL to accepted Slurm statuses as it is also a "terminated"
  status.
- Update tests and add comments and error messages to them.

- Remove negative wallduration check from SLRUM parser as parse_time
  function prevents that from happening.
- Update tests to match.

Extra SLURM statuses

Correct files that have CRLF line endings to LF

Remove references to specific LSF versions as that restriction has been
removed from the code a while ago (but the docs weren't updated).

Remove references to specific LSF versions

- the test captures what happens as of 1.6.1 and
  helps convince us we aren't breaking anything
  when making any future chnages to default
  CloudRecord values.

- Helps convince us we aren't breaking anything
  when making any future changes to default
  CloudRecord values

- when it is ommited, NULL is written as the CpuCount in the CloudRecord
  table, which causes problems at summariser time as
  NULL CpuCounts aren't allowed in the CloudSummaries table

- Cleans up any existing cloud data with a NULL CpuCount,
  which causes warnings when we try and insert cloud
  summarises with NULL CpuCounts into the cloud summary table

Fix summariser CpuCount warnings for cloud accounting

Simplify update script inclusion in SPEC file by matching on wildcard.
This will help avoid ommitting files in future (as was almost the case
here where the 1.6.2 script was missing from the file section).