I'm pretty sure CrOS does, but I don't know about Windows.

We're using Aura for the Window management on both, so I would think the logic would be the same between Chrome OS and Windows. But, hopefully robliao@ or bsep@ can give you an answer, or point you to someone who can.

I'm currently not aware of any occlusion detection with respect to timers. A loose survey with @erikchen also reveals that it may have attempted to do this on Mac, but there were some issues with respect to assumptions made by webpages that depended on these timers.

Do we have cases where the foreground window is occluded by the background window? It seems that what we actually want might be that we want to throttle timers when the window is no longer foreground.

You're suggesting we should treat occluded tabs same as background tabs? Seems reasonable to me.

It's weird that this doesn't already work given those other events already fire when the window is occluded.

Just stumbled across this again.

Re comment 4: Yes, I'm suggesting that occluded tabs should behave the same as background tabs. The occlusion detection is already there since we fire visibilitychange-->hidden. It should be impossible to fire visibilitychange-->hidden and not act like a background tab.

This is likely trivial to fix once someone plumbs through where the logic error is.

Hmmm...I can't reproduce this anymore. Not sure what changed, but seems like it works.

Which platform were you testing with? There might still be issues on some platforms.