Make CFLint output engine consistent across all output formats

[TheRealAgentK]

Currently different output formats use different streams to create their format. Some of them use result marshallers, others manually create XML and then throw XSLT at it for Findbugs or HTML and JSON/TXT are written directly as a manual process.

I think the output engine needs a re-think for 2.0. I like the marshaller approach and I think we could probably pull all output systems into that architecture - which would clean up the current flow quite a bit.

[KamasamaK]

For the dev branch, the JSON format includes multiple new objects at the end of the same array as each issue. That does not appear to match the XML output's change, which does not use the issue tag but rather a new counts tag for that information. That should constitute a new object after the array, but that would also mean that the entire JSON output would be wrapped in an object now. I think that too would be consistent with the XML output.

[TheRealAgentK]

Answered in #323 - let's leave this ticket alone and move over there, #331 was really meant for me to internally rebuild the various output engines, not the formats of XML/JSON etc. I think the current discussion is better had in #323.

[TheRealAgentK]

TODO here, too:

Make sure we remove the limitation of only being able to create one of the XML-based formats only at the same time.