Test cases for PR #153 in JSON-Java

[johnjaylward]

This adds regression test cases for the Constructor and Put methods on JSONObject and JSONArray that should be run again JSON-Java commit stleary/JSON-java@48d31b7 or earlier as a base line.

Notes on running the tests:

• The baseline commit should still be compiled on java1.5 or higher so that generics are supported. The baseline is just the last commit before generics were used in the code itself.
• The tests will run unmodified on both the current master as well as on the PR branch
• When running against the baseline commit, you will need to clean out newer tests, like the BigNumber support, the fixes for JSONML, etc that didn't exist at the time of the baseline.
• The new regression tests I committed should pass for baseline and JSONObject and JSONArray initialization: JSON-java#153, but will fail on master

[stleary]

Will try redirecting this into a branch so that unit test and JSON-java master branches stay compatible.

[stleary]

Moved to branch https://github.com/stleary/JSON-Java-unit-test/tree/johnjaylward-FixCtorGenerics so that I could do some minor edits. I don't think any functionality is changed.

• Tested with failures against JSON-java master. This was expected.
• Tested successfully against JSONObject and JSONArray initialization: JSON-java#153.
• Tested with failures against limited ctor change for 153 JSON-java#164. Analysis in progress. If the failed tests are all due to previously unsupported functionality, I will probably not be too concerned.
• Tested against baseline. Ignored bigNumberOperations() and jsonObjectToStringSuppressWarningOnCastToMap(), which had compiler errors, and wrapObject(), which failed, I think due to testing a BigNumber. Other tests were all successful.

[johnjaylward]

I updated my notes in the description. I was unsure is github notified of description changes.

[johnjaylward]

For your failures on stleary/JSON-java#164, try making both Map generics ?.

i.e. Map<?,?>.

I believe your goal for pr 164 is to possibly replace 153. If that is the case, you will want all tests passing similarly to either BaseLine or pr153. No failures should be from "previous unsupported functionality" as pr164 is based off master.

[stleary]

I was unsure if github notified of description changes.

As far as I can tell, it does not.