fix(chips): safety check before getting length

[StefanFeederle]

If the model is not yet filled (ajax?) this functions fails with:
TypeError: Cannot read property 'length' of undefined

If we check the availability of "this.items" before accessing it we bypass this problem.

[googlebot]

Thanks for your pull request. It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please visit https://cla.developers.google.com/ to sign.

Once you've signed, please reply here (e.g. I signed it!) and we'll verify. Thanks.

---

• If you've already signed a CLA, it's possible we don't have your GitHub username or you're using a different email address. Check your existing CLA data and verify that your email is set on your git commits.
• If you signed the CLA as a corporation, please let us know the company's name.

[topherfangio]

@StefanFeederle Can you please add a quick unit test which demonstrates the issue. I'd like to make sure we have a test so we don't introduce a regression later.

Thanks!

[StefanFeederle]

Hey @topherfangio, here are two codepens to highlight this issue. I don't know how to write unit tests, sorry.

error example
fixed example

The only difference is that i added the code of this PR in the fixed one.

If databinding to a not existing object {{ctrl.form.name}} does not throw errors, md-chips should behave the same and safety-check the object. I know this is not the best coding style and could be circumvented in my code, but i think its also important that angular-material is inherently consistent. This error is really hard to track down for newbies like me.

[StefanFeederle]

I signed it!

[googlebot]

CLAs look good, thanks!

[googlebot]

CLAs look good, thanks!

[topherfangio]

@ThomasBurleson Would love a unit test to go along with this, but is a NPE check, so it's super-simple. Think we should go ahead and just merge it?