I tend to agree, can't really see why it's needed. Perhaps there was some perf reason? Can you put together a jsperf to see if that's the case? I'd be interested in seeing whether the "special fast case" is really faster in modern browsers as well.

I'd be interested in seeing whether the "special fast case" is really faster in modern browsers as well.

My thought exactly.

Hi,
sure, here's some sample jsperf tests.

http://jsperf.com/each-samples

http://jsperf.com/jquery-each-modified-1
http://jsperf.com/jquery-each-original-1

Interesting! I ran some more browsers through, the one marked Other is IE11.

My feeling is that the "special fast case" isn't worth the extra code, it's a hair faster but if this loop overhead was truly an issue then people would be screaming that Firefox is unusable given its bad performance against the others. And if perf is truly an issue you shouldn't be using .each() or Array#forEach you should use a loop.

Anyone else?

The results in firefox are frightening but since they're also consistent between the two paths in the most recent version, I'd be in favour of simplifying the code.

And if perf is truly an issue you shouldn't be using .each() or Array#forEach you should use a loop.

👍
These all look close enough, and therefore probably not worth the size and complexity.

Edit: Except that the two modes are more different than they appear on first inspection. 😦
Still worth a look, though.

Yeah they are not exactly the same, and removing this block will cause 2 ajax tests to fail.

Thanks again for the contribution! I'm ready to land this for 1.12/2.2, but you first need to sign our CLA with your GitHub email address.

In FF33 the fast path seems to be 2x faster then the usual path, see jsperfs
(and FF33 seems the fastest browser over all now, for the fast-case)

Thanks, I have signed the CLA already.

Where does this stand? It sounded like @gibson042 was ready to land but it stalled.

Sorry about the delay!

No worries, thanks for merging it :)