vs7c wrote:Can you do a regression test on a YouTube application?
Good question and the answer is, well, yes and no.
A typical YouTube video is "streamed" from the server into your browser, so it does not really have a fixed state that you can use as the basis for a regression test, so if you mean can you validate the content of the video the answer is no, not directly. You'd have to save the entire video and then use some non-browser comparison software or file differentiator to see if what you just viewed was the same as what you viewed yesterday.
But what we have done in the past is to record viewing the YouTube video by getting it started, and then set up a timer and a synchronization that waits for it to finish. The synchronization is usually on the last frame of the video, and we know from experience that the "last frame" will remain constant from day to day.
While the last frame doesn't have a DOM that you can interrogate, it can be used to do a screen-section checksum validation. That's what eValid calls a "validate and synchronize on screen rectangle", and it works very will to reconfirm completion and also the total playback times.