fix: replace limitted JSON.stringify with better formatter

[thecp]

Prior to this change expected and received data structures were printed
via using the JSON.stringify method to console. However this method is
limited by the fact that it doesn't support enhanced data types like Map,
Set, etc. except when using a replacer function.

[brandonroberts]

Thanks @thecp. Will you provide a snippet of what the changes would look like? Would like to see if its worth bringing in the extra dependency.

[thecp]

Hey Brandon, sure. Consider following example. You have a map (new Map([['name', 'Christoph']])) and you'd expect a different map (new Map([['name', 'John']])) instead.

With the old output you cannot see what is the problem as expected and received look exactly the same:

    Expected: (b|),
    Received: (?|),
        
    Expected:
    [{"frame":0,"notification":{"kind":"N","value":{}}},{"frame":0,"notification":{"kind":"C"}}]
        
    Received:
    [{"frame":0,"notification":{"kind":"N","value":{}}},{"frame":0,"notification":{"kind":"C"}}]

Whereas with the new output you actually see the difference now:

    Expected: (b|),
    Received: (?|),
        
    Expected:
    Array [
      Object {
        "frame": 0,
        "notification": Object {
          "error": undefined,
          "kind": "N",
          "value": Map {
            "name" => "John",
          },
        },
      },
      Object {
        "frame": 0,
        "notification": Object {
          "error": undefined,
          "kind": "C",
          "value": undefined,
        },
      },
    ]
        
    Received:
    Array [
      Object {
        "frame": 0,
        "notification": Object {
          "error": undefined,
          "kind": "N",
          "value": Map {
            "name" => "Christoph",
          },
        },
      },
      Object {
        "frame": 0,
        "notification": Object {
          "error": undefined,
          "kind": "C",
          "value": undefined,
        },
      },
    ]

Although the output is a bit bloated now, this is a big plus from a user perspective to really see what's causing the test to fail. I stumbled over this too, wasting a couple of hours looking for the issue...