commentjson raises the wrong error type on trailing comma

[EliRibble]

For example, let's say commentjson is told to load the following:

{
  // there is a trailing comma below
  "c": 1,
}

This will be caught on commentjson.py:170 as a lark.exceptions.UnexpectedCharacters exception. That seems reasonable. Unfortunately, commentjson.py:205 will then catch this as a generic Exception and translate it into a JSONLibraryException(e). This is bad because it breaks API-compatibility with the Python json library, which would just raise a ValueError

[vaidik]

Not fixing this directly. Instead of that support for trailing commas is being added to expand use-cases for commentjson.