Improve error message when stop is called.

[wparad]

Spent way to long trying to figure why errors were being thrown. Also helps the myriad of other libraries using shifty:

kimmobrunfeldt/progressbar.js#255

[jeremyckahn]

Hi @wparad, sorry for all the trouble you had tracking this down. Thank you for this PR! This all looks pretty good to me. I will test it out and get it released just as soon as I have the time to!

[wparad]

Hi @wparad, sorry for all the trouble you had tracking this down. Thank you for this PR! This all looks pretty good to me. I will test it out and get it released just as soon as I have the time to!

Awesome to hear. I actually wasn't sure about the _attachment property being passed to the resolve and reject functions as a second parameter. I'm not aware of the native promises or promise libraries doing anything with that. But perhaps I am missing something.

[jeremyckahn]

Awesome to hear. I actually wasn't sure about the _attachment property being passed to the resolve and reject functions as a second parameter. I'm not aware of the native promises or promise libraries doing anything with that. But perhaps I am missing something.

Oh, I just realized that this is going to change how tween promises get rejected a tiny bit. I'm okay with that, but I think it warrants a minor version bump when I release this!

In that case, I'd like to roll the _attachment data into the first parameter. I can take care of that after merging this!

[jeremyckahn]

I had to make a few changes to the PR, but I will merge this now. Thank you for the fix @wparad!

[jeremyckahn]

This is in 2.9.0.

[kimmobrunfeldt]

This improved error message is great, but I thought it would be perfectly fine to call the stop method while an animation is running without throwing an error.

[jeremyckahn]

You raise a good point @kimmobrunfeldt. It's fair to say that it's not an error to prematurely stop a tween. I like the catch/reject Promise semantics to differentiate between "tween completed" and "tween canceled," though I agree that calling all canceled tweens an error might be a little heavy-handed.

I don't have much time to do significant development on Shifty at the moment, but I'll think on a more robust API for this and roll that into 2.10.0 when I can!

[jeremyckahn]

FYI @kimmobrunfeldt @wparad, I just released Shifty 2.10.0 which removes the reject logic entirely (d18b278). After giving it some thought, it didn't seem necessary and was just creating problems. Hopefully this makes Shifty easier to use in your projects!