Falko Schindler

Thanks everyone for the investigation! Indeed it looks like a uvicorn issue which we can't do much about. It needs to be fixed upstream. Let's close this issue.

> Compare for > 0 is not necessary because the code is only reached if we have processed a NEMA message with a time stamp. Oh ok. I would have...

Yes, processing all these messages takes a while, which freezes the browser for a short while. In PR #3022 I'm experimenting with another promising approach to speed up the initialization...

I'm still undecided about this PR. The open questions in https://github.com/zauberzeug/nicegui/pull/3017#issue-2279112168 are still standing. There might be side effects (like running into message size limits) that cause trouble as well....

Hi @h0uter, Somehow I can't reproduce the problem with Chrome, Firefox and Safari on Mac. So it's hard to tell what's going on. Can anyone else reproduce it?

Interesting news, @kyloe! So as far as I understand, they suggest to traverse the tabs once they are mounted. We could wrap the Quasar tab components in order to add...

Thanks for bringing this up, @v479038280! This bug very strange. I tried to reproduce it with Quasar and Tailwind, but it works as expected: ```html Hello Hello const app =...

Oh great, @python-and-fiction! Changing the Tailwind version in my example from https://github.com/zauberzeug/nicegui/issues/3176#issuecomment-2151463506 to 3.2.0 (which is currently used by NiceGUI 1.4.x) breaks indeed the layout. ```html ``` Since we usually...

@xaptronic You mean moving `ui.zeitline` into a separate repo/package? Sure, this might be reasonable.

> > 2. Why do we need to call get_slot_stack() in clear()? I guess it's for some important but slightly unintuitive reason. > > Otherwise app.reset crashes during the tests...