[bugfix] Treat standalone carriage return as newline

[benjreinhart]

The existing code correctly handles CRLF, treating it as a single newline. However, it does not appear to treat a standalone carriage return as a newline. According to the spec, carriage returns should be treated as a newline.

Steps to reproduce:

> kdl.parse("foo\r\nbar")
{
  output: [
    { name: 'foo', properties: {}, values: [], children: [] },
    { name: 'bar', properties: {}, values: [], children: [] }
  ],
  errors: []
}

> kdl.parse("foo\rbar")
{
  output: undefined,
  errors: [
    EarlyExitException: Expecting: expecting at least one iteration which starts with one of these possible Token sequences::
      <[BOM] ,[WhiteSpace] ,[OpenMultiLineComment] ,[EscLine]>
    but found: 'bar'
        <...stacktrace...>  {
      token: [Object],
      resyncedTokens: [],
      previousToken: [Object],
      context: [Object]
    }
  ]
}

[larsgw]

Thank you for catching this! Could you also update the RegExp for the LineComment? It should match everything from // up to but not including the linespace, so the pattern should get a lot simpler (since it does not have to consider \r separately from \r\n now).

[larsgw]

Also, sorry to be pedantic but could you insert it after \x0C to keep it ordered?

[benjreinhart]

Yep, thought I had ordered it

[benjreinhart]

Updated from your feedback.

Might want to double check that regex update, but I think it looks good:

> "foo // line comment here\r\nbar".match(/\/\/[^\x0A\x0C\x0D\x85\u2028\u2029]*/)[0]
'// line comment here'

> "foo // line comment here\nbar".match(/\/\/[^\x0A\x0C\x0D\x85\u2028\u2029]*/)[0]
'// line comment here'

> "foo // line comment here\rbar".match(/\/\/[^\x0A\x0C\x0D\x85\u2028\u2029]*/)[0]
'// line comment here'