Accessor signatures of $.fn.data() have a side effect

[mike@…]

When used as a getter, $.fn.data will "fall back" to the DOM data- attribute. Additionally, it will save this value to the internal data lookup table for that element. This means that $.fn.data is an accessor with a side effect.

var $el = $( "<div>" );
$el.attr( "data-john", 23 );
$el.data( "john" );			// 23
$el.attr( "data-john", 45 );
$el.data( "john" );			// 23

This behavior is surprising--querying the state of the element changes it.

Granted, ​the API documentation contains a caveat about this detail:

The data- attributes are pulled in the first time the data property is accessed and then are no longer accessed or mutated (all data values are then stored internally in jQuery).

...but one could argue that the need for such documentation is a sign of an unintuitive interface.

Modifying this behavior would certainly be a breaking change, but I believe it is in jQuery's interest to implement an API that is consistent and easy to reason about.

[Rick Waldron]

jQuery.fn.data only _reads_ from data-* attributes one time and never again, as described in the documentation: ​http://api.jquery.com/data/#data-html5

The data- attributes are pulled in the first time the data property is accessed and then are no longer accessed or mutated (all data values are then stored internally in jQuery).

EDIT: Note there is no side effect of accessing a value: ​http://jsfiddle.net/rwaldron/5RLb7/ What you observed was jQuery's refusal to re-read that element's data-* attributes.

[mike@…]

Actually, jQuery's refusal to re-read the data-* attribute is the side effect I was referring to.

[Rick Waldron]

Replying to mike@…:

Actually, jQuery's refusal to re-read the data-* attribute is the side effect I was referring to.

In that case, you're absolutely right. Apologies for the misinterpretation—I thought you meant that data was overriding the attr value.