The wait_until option is documented to default to load, and it did for URLs. For html, though, the API didn’t wait for anything unless you set wait_until.
So HTML that loaded scripts, styles or images from other URLs could be captured before they were ready. A request with scripts that called the page’s own code could also run before that code was loaded and fail: the API then waited for wait_for_selector until it timed out.
Now HTML waits for the load event by default, like a URL. If the page doesn’t finish loading within the navigation timeout, the API renders it as it is instead of failing, as it does for URLs.
HTML that loads slow external resources may now take longer to render. If you don’t need to wait for them, set wait_until=domcontentloaded.
