An object that is supplied as an option value on instantiation will not end up on the instance (but a clone will)

[timmolendijk]

If you do:

var somedata = {x: 1};

$('#element').mywidget({data: somedata});

You would expect:

somedata === $('#element').mywidget('option', 'data')

Yet this is not the case!

I suspect this is caused by the deep-copy at jquery.ui.widget.js:57:

	$[ namespace ][ name ].prototype = $.extend( true, basePrototype, {
		namespace: namespace,
		widgetName: name,
		widgetEventPrefix: $[ namespace ][ name ].prototype.widgetEventPrefix || name,
		widgetBaseClass: fullName
	}, prototype );

P.S. Line 57 is where this code occurs in 1.8rc1, but the problem applies to 1.8rc3 as well.

[timmolendijk]

On second thought I think the cause is the deep-copy at jquery.ui.widget.js:73:

		// allow multiple hashes to be passed on init
		options = !isMethodCall && args.length ?
			$.extend.apply( null, [ true, options ].concat(args) ) :
			options;

[timmolendijk]

On third thought I'm pretty sure the deep-copy at jquery.ui.widget.js:129 is causing this problem:

		this.options = $.extend( true, {},
			this.options,
			$.metadata && $.metadata.get( element )[ this.widgetName ],
			options );

[Scott González]

We don't claim anywhere that the options hash originally supplied would be the options hash that actually gets used. This is not a bug.

[timmolendijk]

1. I'm not talking about the options hash, I'm talking about the values of options.

2. Wouldn't you agree that it is to be expected, or at least to be desired that if you supply an object x as the value of an option, this exact same object x ends up as the value of that option after instantiation?

[timmolendijk]

See previous comment

[Scott González]

No. Deep copying is the intended behavior. Please do not re-open this ticket.