Add a small regression test to -wrap-flow property testing

[zhorvath]

We need to extend the wrap-flow testing with a small regression test to get the accurate result. Testing only the property parsing is not enough anymore.

[stucox]

Do we need this? Unless there's proof of a browser giving a false positive with the standard detect, we try to go for the simplest (and hence fastest) detect we can.

[paulirish]

if it IS necessary then lets document in the code why. Thanks for your help on this detect, @zhorvath

[zhorvath]

Unfortunately, yes. We parse the property in WebKit, but we haven't got the implementation for the wrap-flow yet, so the original test case is false+ for WebKit.

@paulirish: I'm going to add a comment to the code.

[paulirish]

@zhorvath how does CSS.supports() fair with wrapFlow:end ?

[zhorvath]

@paulirish: sorry, I don't know what CSS.supports() is. Could you send me a pointer to it?

[stucox]

CSS.supports = native feature detection — spec

It's only implemented in FirefoxAurora and Opera at the moment, neither of which support wrap-flow.

[myakura]

CSS.supports() is implemented in WebKit, but both Safari and Chromium hasn't enabled the build flag...

[zhorvath]

So, do you like this pull request? :)

[stucox]

Yep, cheers. I take it it's still true in IE10?

[zhorvath]

Yes. I tested with the latest public IE10.

[stucox]

Nice one, thanks. LGTM.