bug: pgn headers write in double due to forgotten regex~ness of string

[jzacsh]

bug description copy/pasted from jzacsh@232e5d0

var chess = new Chess();
chess.move({from: 'e2', to: 'e4'});
var pgn = chess.pgn({max_width: 5, newline_char: '\n'});
// output:
// [White "carolyn"]
// [Black "hippo"]
// [Date "2014-1-18"]
// 
// 1. e4

var chessB = new Chess();
chessB.load_pgn(pgn);

// BUG HERE:
chessB.pgn({max_width: 5, newline_char: '\n'});
// outputs: [[White "carolyn"] [Black "hippo"] [Date "2014-1-18"] "[White "carolyn"] [Black "hippo"] [Date "2014-1-18"]"]  1. e4

test added that confirms this behavior in: jzacsh@fa55217

[jzacsh]

IMO: it would make more sense from an API/user perspective to pick a clear data type to be passed for the .split() call, and avoid using a string that looks strangely like (but not quite) a RegExp.

The string required (and internally defaulted to) is: '\r?\n'
Instead, the API should either:

• require a regex-valid string like it seems to be attempting, ie: user would pass: '\\r?\\n'
• require a RegExp object, ie: user would pass the result of /\r?\n/ or new RegExp('\r?\n')

As it is right now, there's just more scraping that has to take place (the mask function that tries to escape the API input -- which doesn't seem like a necessary pain).

[jzacsh]

If you want me to update the API per my last comment, feel free to reopen this issue and I'll happily send you another patch!

[ghost]

I know this is really old, but I was perusing this code for the first time and came across the mask function:
function mask(str) { return str.replace(/\\/g, '\\'); }
Please correct me if I'm wrong, but it looks like this function does absolutely nothing and may as well just return str itself. Is there any reason to keep this mask() function at all?