From: jbicha@debian.org 
  
 On Mon, Oct 27, 2025 at 5:52€€€AM Simon McVittie  wrote: 
 > Converting d-i to (a sensible threading model, and then) GTK 3 continues 
 > to be something I'm sporadically working on. Until that is something 
 > that the d-i team could realistically review and merge, I don't think 
 > removing GTK 2 is feasible, because the graphical installer is a 
 > requirement for Debian's support for languages that aren't feasible to 
 > implement on the Linux console (for example Arabic). 
 > 
 > (And as long as we need GTK 2's udebs, I would strongly prefer to 
 > continue to have its .deb as well, so that it's possible to test it 
 > without resorting to doing a whole d-i build.) 
 > 
 > Starting to raise more of the "relies on GTK 2" bugs to serious could 
 > make sense, particularly if you can point to specific functionality that 
 > is expected in 2025 applications but cannot be provided by GTK 2. But I 
 > predict that some Debian contributors will be vociferously opposed to 
 > its removal and I don't have the spoons to fight them over this. 
  
 Yes, as long as debian-installer uses GTK2, we must keep GTK2 in 
 Testing, and yes, keeping the library .deb around to support that make 
 sense. This is like what we have been doing with the old vte. 
  
 So my proposal is to try to have almost nothing else in Testing using 
 GTK2. And to make this a public goal now. So, gtk+2.0, vte, and the 
 cdebconf* packages would be at severity: important while the rest 
 would be serious. 
  
 We gave our reasons in the second paragraph of the initial MBF 
 https://lists.debian.org/debian-devel/2020/04/msg00405.html 
  
 Thank you, 
 Jeremy B€€cha 
  
 --- SoupGate-Win32 v1.05 
  * Origin: you cannot sedate... all the things you hate (1:229/2)