I was just trying to get an automated test for Amazon Payment checkouts to put in place, however I failed in the end because Amazon is protecting (for a good reason) automated checkouts. I even failed when entering the captcha in the test run to record it once using VCR.
However it took me a little while to figure out how to switch the windows as its slightly different implemented than in the Phantom JS driver and there are no good reads on it on the web, so hopefully it will be helpful to other people. There are 2 ways to achieve it.
Switch driver to new window
#switch the focus
page.driver.browser.window_focus page.windows.last.handle
... do whatever you want
Execute an action or inspect the content of the page then continue on main window
page.driver.within_window page.windows.last.handle do
fill_in "email", with: "myemail@mailinator.com"
find(".signin-button").click
end
Inspect window handles hash
page.driver.browser.window_handles
=> [
[0] "{16262221-1070-4036-b4a8-f1d7af2d6dd8}",
[1] "{f37d0424-67c6-4e78-8a5f-77da37e27805}"
]