There are zero plans to ever move to GTK3 or GTK4 or GTKN.
There are dreams & aspirations to drop our use of GTK and fall back on GDK (just a window & event abstraction), but this is not on any timetable because it means reimplementing text-entry, menus, tree/list views and the file browser.
By "zero plans" do you mean you have evaluated them and decided they are incompatible with Ardour's goals, or just that there is nothing in the pipe or on the roadmap to do the port?
One major feature that gtk3+ (or SDL2) brings is Wayland compatibility.
Porting a 600k line codebase to a different GUI is a fool's game (the vast majority of Ardour's code and complexity is in the GUI, not the backend).
We did it once before, from GTK1 to GTK2, which should have been easy. And indeed, the actual GUI toolkit switchover was relatively easy. The problem is that along the way, you keep running into other issues and ideas. We likely could have done GTK1 to GTK2 in a month, but it took nearly a year.
We have no need of traditional desktop GUI toolkits (or SDL) except for the 4 widget classes I mentioned. We have our own internal canvas-based "toolkit" that is Cairo-based, and thus only requires a window+rgba-buffer+event abstraction.
There are zero plans to ever move to GTK3 or GTK4 or GTKN.
There are dreams & aspirations to drop our use of GTK and fall back on GDK (just a window & event abstraction), but this is not on any timetable because it means reimplementing text-entry, menus, tree/list views and the file browser.