Can this script or the test that uses it be made to dynamically work with the mp4 (avc1/aac) or the ogv media, based on whether or not the runtime supports it? (Not all implementations of various builds of the implementations support the codecs in the mp4 test media, for instance.)

I suggest following the pattern I used more recently in the media-source/mediasource-changetype-play.html tests, where the top level test page dynamically:

1. in a distinct test within the page, checks that there is enough support in the implementation for the media (e.g., for this test, video CanPlayType(... mp4's mime vs ogv's mime vs etc) identifies at least 1 supported configuration, and
2. for each supported configuration, dynamically create a test that verifies the appropriate behavior. In this change, it looks like you might need to pass some sort of query parameter to the server to identify which test media is being requested; caveat I'm less familiar with using .py in wpt to serve dynamically the bytes requested, so there might be a better way.

What is the intent of seekTime? To be beyond currentTime? To be beyond duration? Note that the seeking spec constrains the seek target time to be in [earliest possible position, end of the media resource], then further constrains it by seekable ranges, etc. I suggest making this value defined in the async_test's scope, and calculated within videoLoaded(), with a comment describing the intent.

The intent here is to seek to an unbuffered position. The test file is 5 minutes long, and here we're seeking to 4 minutes. That will make the browser make an HTTP range request asking for data somewhere in the middle of the file. The server in this case is programmed to let these requests time out on purpose. This way, the test checks that timeupdate is fired regardless at the time of seek (which is the purpose of the test).

What's the intent behind removing the video from the DOM? If the goal is to reset the player, why not video.src = "";?

That it does not remain in the DOM after the test has run. It was created by the test, it may as well be removed by the test.

If the test fails (doesn't reach test.done()), then the video also won't be removed. The test.add_cleanup() function can be used to remove the video, right after it's inserted.

But when the test fails often you are interested in what state it was at that point so you can debug the issue.

Hmm yes, good point.