runPlayers doesn't take a second argument. It sounds like you're suggesting one unofficially but I'd rather we not have this bite us later if we add one without looking at this. The lack of specificity won't hurt because elements already played won't get played again.

This is a guard against attempting all players again. The internal player guard is not working well in many cases. We also need to look at the refresh calls in the Context Vue component someday, it causes some of the issues I observe @boutell

See apos.util.runPlayers = function (el, newPlayer) {. We end up with players being called back to back and breaking a lot of stuff. Again, now we have a lot more room for racing conditions and it bites us (mostly in edit mode, when we have also ajax dom updates from what it seems and manipulate the dom while players are attempting to run).

Should be newPlayers

I still think we're fighting ghosts here, the real problem is players that replace the element completely, which we should document you should not do because it prevents tracking if players have been run (or your player must check that on its own if it insists on replacing the element)

But fine with this optimization if we name the parameter better

I tested on players that are not replacing the element, but mounting (inserting child nodes). The player guard itself is mostly not working because the rapid event fire and DOM changes racing conditions. I spent a good amount of time debugging the whole. I think with this change, players are doing good on itself, eliminating the same tick executions during initialization. The next thing we might want to insect someday is the Vue context component refresh calls, that in some cases are spamming the refresh event - it should be probably better computed and/or slightly throttled.

I came up with a better strategy I think. I like it more because it gives no "freedom" to pass any player outside of already registered, but still does improve the performance and guards against "player spamming".

Did some more testing and debugging... and reverted the whole new players thing. The performance gain is questionable, we need a more clear strategy for improvement here.