Double-counts lines under certain (Windows) line-ending conditions

[cmstead]

I discovered while traversing a produced AST that the node positions extended beyond the length of the file, nearly doubling the line length. After poking around and testing a hypothesis, I discovered that by normalizing lines endings to "\n", the line count returned to normal.

Operating System: Windows (works as expected on Mac)
Line ending: Windows standard (\r\n)

JSON source in question:

{
    "Snippet snippet": {
        "prefix": "snippet",
        "body": [
            "\"$1\": {",
            "\t\"prefix\": \"$2\",",
            "\t\"body\": [",
            "\t\t$3",
            "\t],",
            "\t\"description\": \"$4\"",
            "},"
        ],
        "description": "A snippet for creating new snippets"
    }
}

The normalization fix I applied in my parser abstraction is as follows:

function normalizeSourceLineEndings(source) {
    return source.split(/\r?\n/).join('\n');
}

I know this project is old, and potentially unmaintained, but I figured this issue would, at least, help someone else if they encountered a similar issue.

Thank you for building this.

Cheers!