This changes the current script.max_size_in_bytes to be dynamic so it can be 
set through the cluster settings API. This setting is also applied to inline scripts 
in the compile method of ScriptService to prevent excessively long inline 
scripts from being compiled. The script length limit is removed from Painless as 
this is no longer necessary with the protection in compile.

Workaround for: elastic/elasticsearch#35184

Workaround for: elastic/elasticsearch#35184

…better error handling, --cacert, multiple time fields (#239)

* run_test.py: Improve/cleanup and add YAML support for input files

* Cleanup shell scripts according to ShellCheck recommendations

* run_test.py: Cleanup Indices after test to not pollute tests env

* run_test.py: More useful error message if logging action did not run

* run_test.py: Refactor

* run_test.py: Use load_file() for ES scripts as well to support YAML

* run_test.py: Implement --no-execute-watch needed for deployment

Needed for: elastic/elasticsearch#30112 (comment)

> I already have a workaround in place for this which consists of automatically deploying as many watches as I need to send different mails. Those watches are derived from my watch definition. For this, I extended [run_test.py](https://github.com/elastic/examples/blob/master/Alerting/Sample%20Watches/run_test.py) to inject Python code after the watch definition is read. Not ideal, but it is maintainable.

* run_test.py: Support to inject Python code, useful for deployment

Needed for: elastic/elasticsearch#30112 (comment)

> I already have a workaround in place for this which consists of automatically deploying as many watches as I need to send different mails. Those watches are derived from my watch definition. For this, I extended [run_test.py](https://github.com/elastic/examples/blob/master/Alerting/Sample%20Watches/run_test.py) to inject Python code after the watch definition is read. Not ideal, but it is maintainable.

* run_test.py: Implement --no-test-index needed for deployment

Needed for: elastic/elasticsearch#30112 (comment)

> I already have a workaround in place for this which consists of automatically deploying as many watches as I need to send different mails. Those watches are derived from my watch definition. For this, I extended [run_test.py](https://github.com/elastic/examples/blob/master/Alerting/Sample%20Watches/run_test.py) to inject Python code after the watch definition is read. Not ideal, but it is maintainable.

* run_test.py: Add --metadata-git-commit switch to augment watch metadata

* run_test.py: Add --cacert parameter

* run_test.py: More useful error message if logging action did not run

* run_test.py: Use `git rev-parse --short HEAD` for --metadata-git-commit

* run_test.py: More useful error message if transform failed

* run_test.py: Implement --minify-scripts

Workaround for: elastic/elasticsearch#35184

* "Scripts may be no longer than 16384 characters." is in ES<v6.6 not >6.6

* run_test.py: Improve compatibility with ES 7.0.x and index templates

* run_test.py: Better error message if expected_response is not defined

* run_test.py: Show watch exception on execution failure

* run_test.py: In case a transform fails the transform input is relevant

* run_test.py: Support multiple time fields

Useful when you have two time fields that in reality should be very
close so in testing it is enough to set them to the same value.

* [run_test.py] ES 7 support. Update to Py3 and drop elasticsearch_xpack.

* [run_test.py] Add --verbose parameter to debug ES responses

* [run_test.py] Comply with Python Enhancement Proposals

* [run_test.py] Comply with reuse.software

* [run_test.py] Avoid `not` in condition to make it easier to understand

* [run_test.py] Use str.format instead of "%s" % for consistency

* [run_test.py] Fix ./run_all_tests.sh test run. All passing again.

* [run_test.py] Support nested fields in time_fields test parameter

Example:

```yaml
time_fields:
  - '@timestamp'
  - 'event.created'
```

* [run_test.py] Use dict.get shortcut