Oskar Wickström

OK, I can reproduce this in GHCi (and `cabal v2-repl`) locally on the `Hello` example. Haven't had time to dig into it yet, but I'll have a closer look this...

Switching the propagation, if I'm not mistaken (it's been a while since I hacked on this), has the unfortunate effect of the special `Exit` event not being emitted. Maybe this...

I've also confirmed it hangs in GHCi for me. Can't say say for certain, but I think it has worked before (maybe an older GHCi or haskell-gi package version).

I don't have much time right now to do the relevant Stackage maintenance work, but if someone would like to, feel free to!

Oh, I just accepted https://github.com/owickstrom/gi-gtk-declarative/pull/59, which relaxes the upper bounds for those packages. I haven't yet released new versions of the gi-gtk-declarative* packages, but will try to get to it...

@ix 0.6.0 published to Hackage now, both packages.

@juhp I think this release solves the version bound problems you had.

I haven't used `ComboBoxText` before, but I had a quick look at it. I think we can go either of the following two routes: 1. Use `CustomWidget` and implement a...

This is strange. I'm not too familiar with the `realize` signal and how it should work. I'm thinking that it might be some ordering problem with the showing of widgets...

You're right, it's not explicitly supported right now. You can hack it in by including `Window` bins in an existing container, but the problem is that the container grows: ```haskell...