The analysis above is correct as far as I can recall. The goal is to paint transparent once (to support underlay), and not have to repaint it again each frame.

This particularly invasive way to do it removes the need to synchronize the video underlay SurfaceView with the renderer SurfaveView when rotating, and so removes a lot of jank in that transition. We would need to solve the more fundamental Blink-fullscreen-jank issues to get the same result without the hack.

foolip@, could you add a test case for this? :)

This could be tested with a reftest with a video that goes fullscreen and this style:

video::backdrop {
  background: green;
}

The expected file would be solid green, but the actual results are going to be black. (Unless this has been fixed since filed.)

I wrote a test page here: http://mounirlamouri.github.io/sandbox/bugs/627850.html

Using browserstack, I found the current compat matrix:
- Firefox: working
- Edge: not working
- Safari: not working

Also, the behaviour seems to be the same on desktop (Linux) and mobile.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

I understand that the blink fullscreen transition code has improved since this bug was filed, we should re-test.