Regression with chaining multiple delay + animation calls

[cnanney]

In upgrading to 1.5, my  flip counter script stopped working. I've created a test case here:

 http://jsfiddle.net/cnanney/3QsYY/

You can see the animation works in 1.4.4, but when you switch to 1.5 and run it, nothing happens.

[cnanney]

I've reduced this further to a single frame of animation:

 http://jsfiddle.net/cnanney/3QsYY/1/

Works in 1.4.4, not in 1.5. Seems an animation will not fire if chained to a delay.

[rwaldron]

Confirmed - but you're second "reduced" test doesn't work at all.

[cnanney]

The second example was only one frame of the animation, which was kind of hard to see, but you'll notice the top of the digit has started to fall. This one is easier to see:  http://jsfiddle.net/cnanney/3QsYY/2/

[jitter]

This isn't a bug. The  animate() documentation reads:

All animated properties should be animated to a single numeric value

background-position isn't a property that consists of a single numeric value.

[jitter]

Duplicate of #8160.

[cnanney]

@jitter, thanks for the quick response. Fixed by separating into backgroundPositionX and Y animations.

[jitter]

Replying to cnanney:

@jitter, thanks for the quick response. Fixed by separating into backgroundPositionX and Y animations.

np. But you might need to be careful there as (afaik) not all browsers support backgroundPositionX/Y .

[cnanney]

You're right... works fine on Webkit and IE, but not Firefox and Opera. Of course.

So jQuery lost the ability to animate background positions without the use of an external plugin (cssHooks)? Interesting.

[dmethvin]

If the background-position animation worked at any point, it was by luck and not by design. The docs have always said that "All animated properties should be animated to a single numeric value."

[cnanney]

Thanks for all the information. In my situation using animate() in this way wasn't even the most efficient solution anyway.

However, I did notice some strange behavior with regards to background-position animation: When supplying two values with 'px' (20px 30px), it worked in all browsers in 1.4.4, and none in 1.5. But when only the first has px (20px 30), it is the opposite, works in 1.5, but not for 1.4.4:

 http://jsfiddle.net/cnanney/3QsYY/5/

The change in the rfxnum regex doesn't seem to support this change in behavior. I guess if it was never officially supported anyway, I'm glad to have stopped using it. Thanks again.