feat(ngModelOptions): support for easily debouncing `default` only

[gkalpak]

Do you want to request a feature or report a bug?
Feature.

What is the current behavior?
The delay specified in an ngModelOptions' debounce.default property will be used as a "catch-all" value, debouncing all updateOn events that don't explicitly specifiy a debounce delay.

If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://plnkr.co or similar (template: http://plnkr.co/edit/tpl:yBpEi4).
Demo

What is the expected behavior?
debounce.default should only apply to the default event.
For a "catch-all" delay, we could use the '*' property, which is similar to how it is used for defining inheritance in ngModelOptions. E.g.

{
  updateOn: 'default blur keyup keydown',
  debounce: {default: 100, blur: 0, '*': 333}
}

What is the motivation / use case for changing the behavior?
The current behavior can be confusing, since updateOn supports a special event, default which means the default input event for that input/browser. With the proposed change, it will be still possible to provide a "catch-all" delay (via '*') and debounce.default will only affect the default event, which is more intuitive. Additionally, it will be easier to apply debouncing on the default event only, without having to explicitly set delays for all other supported events.

Which versions of Angular, and which browser / OS are affected by this issue? Did this work in previous versions of Angular? Please also test with the latest stable and snapshot (https://code.angularjs.org/snapshot/) versions.
Probably 1.3+.

[petebacondarwin]

I believe we can do this without a BC if we fallback on default if there is no "*"

[gkalpak]

I believe we can do this without a BC if we fallback on default if there is no "*"

This would fix most of the issue (i.e you would still need and extra '*': 0 if you only want to debounce default), but would make the API even less intuitive imo. I would prefer to leave it as is (and maybe document it better), than introducing more compexity into the rules.