Reject some() (and any()) if the input array contains not enough items

[jsor]

This fixes a bug where some() never resolves if the input array contains not enough items.

some([1, 2, 3], 4)->then(function() {
    // Never called
});

[jsor]

ping @cboden @clue @WyriHaximus

[clue]

The documentation seems to suggest otherwise. IMO it makes sense to reject this with a RangeException, but we should probably also add this special case to the README.

[jsor]

Added a note the the README (e29472f).

[clue]

See remark about documentation for any(), otherwise LGTM 👍

[clue]

README now LGTM :)

However, after further consideration, I think this should probably throw a subclass of LogicException rather than RuntimeException. After all this is in fact a logic error which could be detected at "compile time" (i.e. it does not really depend on a runtime decision). What's your view on this?

[jsor]

Good point. I've used RangeException after discovering this bug while looking through the bluebird code and extended the SPL RangeException (which in turn extends RuntimeException).

I think it would be strange to use the name RangeException and extend LogicException because this would differ from the SPL.

Maybe we should use another name...any suggestions?

[joshdifabio]

Would UnderflowException or LengthException, both of which extend LogicException, make sense here?

Edit: just realised that UnderflowException extends RuntimeException.

[jsor]

LengthException might be an option...

[jsor]

Renamed to LengthException in cd6bb56.

[jsor]

I'm planning to add this to 3.0.

[jsor]

Any objections against including this in a 2.4.0 release?

[cboden]

Technically isn't this an API break?

[jsor]

I think (and iirc we agreed on IRC) that this is a bug fix, because without this fix

some([1, 2, 3], 4)
    ->then(function() {
        // Never called
    })
    ->otherwise(function() {
        // Never called
    });

simply does nothing. It neither calls the fulfillment nor the rejection handler.

[cboden]

ah. 👍 for 2.4 then