Overlays are not removed when app is unmounted about single-spa-inspector HOT 8 CLOSED

Maybe using MutationObserver is worthwhile for this? basically, observe for if/when the childList is empty. Could also possibly done with ResizeObserver since it may not just be because children were removed... Just thinking.

from single-spa-inspector.

I don't think we need MutationObserver here. Can't we just hook into the click event for the Mount and Unmount buttons (here) and call a prop function to setState inside of apps.component in order to trigger this effect to clean up?

Seems like since we're the ones controlling the Unmount button that we should be able to manage the state needed to get the overlay unmounted.

from single-spa-inspector.

My thought process is that the devtools should display if the app has a visible size, maybe even using some threshold to determine if it should be shown. Any number of things can cause an app to become unmounted to so just hooking into the buttons doesn't seem ideal, but if there is a way to hook into the activity functions that would work well too.

from single-spa-inspector.

Single-spa’s app change event (https://single-spa.js.org/docs/api.html#app-change-event) might be useful then? I’m very averse to introducing a mutation observer when we have 100% control over all of the code involved

from single-spa-inspector.

Ah, hmm, it may be as easy as changing the event we're listening to from the routing event to the app change event then?

Since that event I believe causes a rerender, which should in turn add/remove overlays? or am I wrong there?

from single-spa-inspector.

Yea, app-change looks like exactly what we want to use here. No MutationObserver then!

from single-spa-inspector.

Wait but single-spa:routing-event fires every time that app-change is fired, plus more. I did not realize that we were already listening to it. Listening to app-change won't help then. The bug must be something else?

from single-spa-inspector.

Hmm possibly. I think this is more like a missed feature, not quite a bug. I'll look into whether still using routing-event is enough. I personally haven't had time to dig into it.

from single-spa-inspector.