Browserurlchange

[jeffbcross]

Pushing for Travis...you know the drill

DUPE of #8785

[jeffbcross]

@tbosch Travis is happy now.

For the test that was causing full page reload in IE, I just wrapped the whole test block in a condition to make it not execute if !history.replaceState (since the test was for history.replaceState which isn't supported). Something was causing the page reload when constructing the $browser in that test.

I also added a separate test to test that changing the hash would fail before and succeed after this fix.

[jeffbcross]

@tbosch I'm closing this PR; do what you will with my commit.

[tbosch]

Great, thanks for the investigation!
Thought about the test a little more and moved it into browserSpec.js, as there are FakeWindow and FakeDocument that it can use, so we don't need a special case for IE9.