Bug 7260 - Allowing length prop to be used in jQuery.each

[jboesch]

With jQuery.map supporting objects in 1.6 (and works with props named 'length') I think we should consider adding 'length' prop support to jQuery.each. Here is Dan Heberden's isArray check (that I'm using) used in the jQuery.map commit: https://github.com/jquery/jquery/pull/299/files#L0R713

I discussed a little here: http://bugs.jquery.com/ticket/7260
Fiddle: http://jsfiddle.net/jboesch26/Ungsa/

This would allow looping over objects with length props. This newEach seems to be slightly faster as well (in all except IE): http://jsperf.com/each-object-supporting-length-key/6

I realize this is a touchy change, as jQuery.each is used absolutely everywhere... so any thoughts on this would be rockin' :)

[jboesch]

The reason I'm storing it in a var is because when I ran tests, it proved to be slightly more performant. Hey, every ms counts right!.. especially in jQuery.each.

[rwaldron]

I'm not sure I see the reason for adding another step to the loop stack. What does this look like with JUST support for a length property and no modifications to the rest of the logic?

[jboesch]

The reason I added another step to assign the variable, is because it proved to be slightly faster than doing a lookup on the object twice. In this perf test http://jsperf.com/each-object-supporting-length-key/6, newEach just contains the logic for allowing length prop to be used, in newEach2 it contains the logic for allowing length prop as well as the variable assignments in the loops.

[jboesch]

Reverted those local vars in the loop stack.

[jboesch]

Note to self: Add jQuery.isPlainObject as a last check
#356

[rkatic]

I started a discussion on isArrayLike thing: http://forum.jquery.com/topic/jquery-isarraylike-for-consistency

[jboesch]

Closing because this doesn't need to be addressed right away. Also, im so tired of talking about length prop support for jQuery.map and other ppl proposing isArrayLike stuff. Peace.