fix: widget masks overlap when navigating between screens.

[thisames]

📜 Description

At first I thought it would be necessary to implement a way to scrape widgets by context, as I did in posthog, but I realized that your code was already correct, it just needs one adjustment.

applying debugVisitOnstageChildren to your children does exactly what we were looking for. a way to "separate the elements that are hidden and not interactive from the active elements in the current context of the screen"

💡 Motivation and Context

widget masks overlap when navigating between screens.

#2309

💚 How did you test it?

I just did several tests with different navigations and applications, checking and debugging the widgets

2024-12-11.21-50-19.mp4

📝 Checklist

• I reviewed submitted code
• I added tests to verify changes
• No new PII added or SDK only sends newly added PII if sendDefaultPii is enabled
• I updated the docs if needed
• All tests passing
• No breaking changes

🔮 Next steps

[vaind]

This does not seem to be 100% reliable. So far in my testing, I've seen one occasion when masks were off. Never seen that before. I'll check the implementations of debugVisitOnstageChildren().

In the following screenshot (and the correct one for reference), it seems way off. Note: this has been taken while scrolling rapidly.

[thisames]

What app are you testing with? I would like to reproduce the same time, but besides that, in my tests if I scroll quickly with visitChildElements the same thing will happen (masks were off).

try comparing the test you just did using visitChildElements and debugVisitOnstageChildren.

[vaind]

What app are you testing with?

https://github.com/vaind/spotube/tree/sentry

in my tests if I scroll quickly with visitChildElements the same thing will happen (masks were off).
try comparing the test you just did using visitChildElements and debugVisitOnstageChildren.

Interesting, I've never noticed that before. If you have a good repro you could shere for this, I'd like to take a look.