Pisu Vam proto, protoze tyto vase vypady jsou opakovene, zjednodusene a za me zbytecne dehonestujici. Myslim, ze i Vam bude neprijemne, kdyz nekdo bude (opakovane) psat Pirati "nezklamou", kdy jednotlivec neco provede. Dekuji za pochopeni a zvazeni, JS

Dobry den, vase opakujici se vypady oproti celku jsou mi velmi neprijemne. Za cirkev se povazuji vsichni verici a vy nas nalepkujete a odsuzujete vsechny, protoze nekdo jiny neco udelal. A ikdyby tam snad patril jen kler, tak porad absoltuni vetsina nic neudelala a neublizila.

@p-y.wtf just got a heart attack

yeah. see the predictive back handler in my sample. I changed my impl of nav back to trigger back press instead of remove last. though this is extremely hacky.

I've just found there was already a ticket: issuetracker.google.com/issues/44055...

Filled:
- issuetracker.google.com/issues/45534...
- issuetracker.google.com/issues/45534...

I just found the code in NavDisplay that renders OverlayScene. No transition support at all. I guess that's a bit unexpected -> maybe that's what you mean by "more structured API between an OverlaScene and NavDisplay" :)

Thanks a lot for those answers.

Attaching my implementation. Does anybody have an idea how to animate the backStack.removeLast()? cc @ianlake.bsky.social

But... how can I handle "back stack" removeLast() animation? Basically, NavDisplay stops rendering my scene, and that's it. I may fake it via artificial backPress dispatch instead of removeLast, but that doesn't seem the be "the proper way".

Putting a PredictiveBackHandler there also enables handling this animation/eases the suspending finish animation.

The initial animation is a LaunchedEffect, also not a problem.

That, along with synthetic back stacks and different handling whether you are on your own task stack or another apps' task stack (which is the only time when back and up differ), is going to be the next PR after this first one.

Sadly, it's not covering the "Up button ves Back Button" case I'm most interested in.