So what's going on is the following:
* We run the `onloadstart` event on the media element (the `<audio>` element on the test-case). That triggers the `go()` function, all good so far.
* We set the `src` of an `<iframe>`, then call `setAttribute("ondurationchange", "x()")` on the media element.
* That goes through a doc update which does end up in `Document::MaybeInitializeFinalizeFrameLoaders` (as we have a frame loader to initialize).
* That ends up calling `nsDocShell::Stop` on the iframe with id="c", which flushes style from `DocLoaderIsEmpty()` in order to figure out whether we're done with loads.
* That flushes layout on the parent document, as we promote style flushes to layout flushes. That's also ~fine, though not great. The iframe document is `about:blank` at that point. It's a bit unfortunate that setting the src attribute of an `about:blank` iframe flushes the parent document layout, but that's just an observation... Anyhow at that point the stack looks like:
```
Bug 1578933 Comment 5 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Edit: clear because bugzilla tricked me before I finished the comment...