What's holding you back from upgrading to 1.8 at this point?

This resonated a lot, thank you so much for sharing 💚

That's more of an architectural question vs a Compose question. But yes, you'll definitely want to re-use your player instance(s).
Have you checked the new PreloadManager? That was built specifically with short form video use cases in mind? See github.com/android/soci...

Ready to dive in? Check out the updated docs for details & examples: developer.android.com/media/media3...
Feedback? Let us know on GitHub: github.com/androidx/media

#Android #VideoPlayback

Why this approach? We're giving you the flexibility you need. Instead of a pre-built UI, you get robust, tested state logic. This ensures your custom controls are perfectly synced with playback. Reliability of Media3 + creative freedom of Compose = 🤩

This release contains:

- PlayerSurface: Clean wrapper for Surface/Texture for easy video rendering.
- PresentationState: Smoother transitions between playlist items.
- Composable State Classes: Build your own UI controls! Play/Pause, Next/Prev, Shuffle, Repeat, and more to come!

Got you! Accidental clickbait haha 😆

I'll raise this to see if we can change the order here somehow. I know the pain :)

Ahh that's cool! So let me try to understand; you registered your broadcast receiver in your manifest, launched a foreground service when it was called, and then in the foreground service you render Compose UI?
Did you encounter any strange effects while using Compose to draw on top of other apps?

Did you write any code using Compose for that use case? In general broadcast receivers and foreground services are a very complex area, I'm coming in from the Compose side trying to understand where they overlap.

Followed up anyway and indeed Android Studio / Intellij doesn't use type information for auto-complete as it requires too much processing and would make auto-complete too slow. Using smart completion will do what you need.

Can you give a specific example so I can raise this internally?