I don't know for other libs, but Qt has strong API stability guarantees, so you need to generate the bindings only once per major Qt version - that is, once every five or six years. Just because the binding hasn't been commited to for a few years doesn't mean that it won't work.
> It would just be so awesome if there were a viable cross-platform desktop GUI solution, for a modern statically typed language somewhere between the extremes of C++
What's wrong, with, as you mentioned, C++ and Qt ? It can be as high-level as you want it to. I've been successfully developing a large desktop app with it and it has been a pleasure from start to end.
For me that is JavaFX and Xamarin.Forms.
> The only thing that's "dead" is the notion of being able to get away with just writing a native Windows version, and ignoring Macs (or vice-versa)
Depending on the world region, you can still get away with targeting only Windows.
Not every country has lots of people able to pay for Apple devices, or even with Apple stores in every corner.
Some people from the GNOME and Rust communities are working on macro magic that allows you to create GObject classes from Rust [1].
The downside is that the applications do not look native on macOS or Windows. But so doesn't Electron ;).
Linux and BSD however can safely be ignored. (Writing this from an Ubuntu laptop.)
I maintain a Qt application for Linux/macOS/Windows, with in total ~70kloc the last time I looked. It started as a Linux application, but porting to Windows was not a lot of effort and the amount of platform specific code is very small. The painful part on Windows are some non-Qt dependencies, such as libxml2, libxslt, DB XML, etc.
The port to macOS required much more platform-specific code to implement macOS-specific behavior (such as keeping the application running when the last window closes), building working, signed application bundles, etc.
We used to talk about Windows/Linux for portability topics (conveniently ignoring BSD and others, I reckon). Somehow, Apple took over much of Linux's mindshare. And not just here. Many of my coworkers are asking for apple laptops. A huge fraction of keynote presenters use an apple laptop.
All that is a bit sad, considering Apple is even worse than Microsoft, freedom wise. They popularised the digital prison, that was originally confined mostly to game consoles. I'll never forgive that, no matter how much they contribute to open source.
It's all awesome, and this example here is certainly better than most (fantastic work!). The problem has traditionally been that most of these things are hobby projects from a single developer, and they go moribund when the developer gets bored and moves onto something else. Hell, if you Google around a bit, you might still find a "wxJava" binding on Sourceforge that I started and abandoned ~15 years ago.
To actually be fit for production use, one of these things needs to be either: (a) sponsored by some entity that keeps leadership continuity going, or (b) driven by an unusually large and unusually non-toxic volunteer group that is able to effectively mimic "a", or (c) created by someone who sees it as their crowning achievement and effectively dedicates their life to it. If this project really does have committed backing from the full KDE organization itself, then that's very encouraging!
I'd be interested to know if it's really an organization level commitment, or if they're just providing hosting for an experiment spearheaded by a single contributor (the Go team at Google has an "experimental" GUI library like this, but it's just one guy's noodling and no one really sees it as gaining traction outside of that). I'd also be curious what the cross-platform story is at this point (i.e. does it build and run on Windows?).
The only other examples I've seen so far that even come close to fitting the bill are in the Python community. And even that is a bit of a mess. There's confusion between two competing Qt bindings... and the wxWidgets binding has been going through a re-write for years, such that it's confusing whether to stay on the old version or adapt the new one (kinda like Python itself!).
But I still keep my eyes open, and look into this space every year or so. It would just be so awesome if there were a viable cross-platform desktop GUI solution, for a modern statically typed language somewhere between the extremes of C++ and JavaScript. I accept that desktop development is less important than it was 15 years ago... but I reject this notion that, "Desktop development is dead, just use Electron". There are still millions of use cases for true desktop applications. The only thing that's "dead" is the notion of being able to get away with just writing a native Windows version, and ignoring Macs (or vice-versa). And all of the cross-platform solutions have major problems, or are tied to languages or toolchains that people don't want to use, so that lack kinda sucks the oxygen out of the room for the open-source and hobbyist crowd on web forums.