Updates javadoc to match actual exceptions thrown.

[johnjaylward]

What problem does this code solve?

• Corrects Javadoc for Minor - JSONObject.put(String Object) throws wrong Exception #402
• Some API methods are added/modified
• Updates some some boxing calls to use the newer <Number>.valueOf methods instead of new <Number>
• Updates calls to return sooner. i.e

this.put(value);
return this;

changes to:

return this.put(value);

Risks
Low. Changes are either documentation only, or semantic to reduce stack by a single dword by reusing the return values. Changes to API methods are not expected to affect existing applications.

Changes to the API?

• Adds new API methods:
  JSONArray put(float) throws JSONException
  JSONArray put(int, float) throws JSONException
• Modifies behavior of existing API methods
  JSONArray put(double) : Removes validity check (because put(Object) does)
  JSONArray put(Object): Tests validity before assignment
  JSONArray put(int, Object): tests validity if in-list, add instead of put NULL after end of current list.
  JSONObject(Map): Throws NPE if key is null

Will this require a new release?
No

Should the documentation be updated?
This PR updates the documentation. If some is missed, please note it in the comments or Code Review sections.

Does it break the unit tests?
No. New unit tests were added to reinforce the API

Was any code refactored in this commit?
Yes. As mentioned above some boxing calls were updated to the newer java5+ <Number>.valueOf instead of using the constructors as recommended by java code practices. Also, some returns were optimized which should not affect execution beyond decreasing stack size by a single dword.

Review status
ACCEPTED

[stleary]

Looks good, starting 3 day comment window.

[data-enabler]

JSONObject(Map): Throws NPE if key is null

A bit late here, but I would consider throwing an exception for an input that used to be accepted a breaking API change. In my company's codebase we had some tests that verified this behavior that were broken as a result, and a currently unknown amount of code that could actually be depending on this behavior.

Obviously this ship has already sailed (and rejecting null keys makes much more sense), but can we at least point out the breaking change in the release history documentation? Would be very useful as a heads-up for anyone trying to upgrade from an old version.

[stleary]

Moving this comment to new issues